S = —— dV dtlgonchar/courses/p9826/Vacuum_fundamentals.pdf · ture, turbomolecular pumps with magnetic bearings, the A-Series vacuum gauges, the TRANSPECTOR and XPR mass spectrometer

Fundamentals of Vacuum Technology

FUNDAMENTALS

dVS = ——dt

D00 E 19.06.2001 21:35 Uhr Seite 1

GeneralFundamentals of Vacuum Technology

D00.02 LEYBOLD VACUUM PRODUCTS AND REFERENCE BOOK 2001/2002


revised and compiled byDr. Walter Umrathwith contributions from

Dr. Hermann Adam †, Alfred Bolz, Hermann Boy, Heinz Dohmen, Karl Gogol, Dr. Wolfgang Jorisch, Walter Mönning, Dr. Hans-Jürgen Mundinger, Hans-Dieter Otten, Willi Scheer, Helmut Seiger, Dr. Wolfgang Schwarz, Klaus Stepputat, Dieter Urban, Heinz-Josef Wirtzfeld, Heinz-Joachim Zenker

Preface

A great deal has transpired since the final reprint of the previous edition of Fundamen-tals of Vacuum Technology appeared in 1987. LEYBOLD has in the meantime introduceda number of new developments in the field. These include the dry-running ALL·ex chemicals pump, the COOLVAC-FIRST cryopump systems with quick regeneration fea-ture, turbomolecular pumps with magnetic bearings, the A-Series vacuum gauges, theTRANSPECTOR and XPR mass spectrometer transmitters, leak detectors in the ULseries, and the ECOTEC 500 leak detector for refrigerants and many other gases. Moreo-ver, the present edition of the “Fundamentals” goes into much greater detail on sometopics. Among these are residual gas analyses at low pressures, measurement of lowpressures, pressure monitoring, open- and closed-loop pressure control, and leaks andtheir detection. Included for the first time are the sections covering the devices used tomeasure and control the application of coatings and uses for vacuum technology in thecoating process. Naturally LEYBOLD’s “Vacuum Technology Training Center” at Colognewas dependent on the invaluable support of numerous associates in collating the litera-ture on hand and preparing new sections; I would like to expressly thank all those indi-viduals at this juncture. A special word of appreciation is due the CommunicationsDepartment for its professional contribution in preparing this document for publishing.Regrettably Dr. Hermann Adam, who at one time compiled the very first version of the“Fundamentals”, did not live to see the publication of this edition. Though he had beenin retirement for many years, he nonetheless labored over the corrections and editing ofthis current edition until shortly before his death.I hope that this volume will enjoy a response as favorable as the previous version.

Dr. Walter UmrathCologne, August, 1998

D00 E 19.06.2001 21:35 Uhr Seite 2

Contents Fundamentals of Vacuum Technology

D00.03LEYBOLD VACUUM PRODUCTS AND REFERENCE BOOK 2001/2002

D00

1. Vacuum physics Quantities,their symbols, units of measure and definitions . . .7

1.1 Basic terms and concepts in vacuum technology . . . . . .7

1.2 Atmospheric air . . . . . . . . .101.3 Gas laws and models . . . . .111.3.1 Continuum theory . . . . . . . .111.3.2 Kinetic gas theory . . . . . . . .111.4 The pressure ranges in

vacuum technology and their characterization . . . . . 12

1.5 Types of flow and conductance . . . . . . . . . . . .12

1.5.1 Types of flow . . . . . . . . . . .121.5.2 Calculating conductance

values . . . . . . . . . . . . . . . . .131.5.3 Conductance for piping

and openings . . . . . . . . . . .141.5.4 Conductance values for

other elements . . . . . . . . . .15

2. Vacuum Generation . . . . .162.1 Vacuum pumps: A survey .162.1.1 Oscillation displacement

vacuum pumps . . . . . . . . . .172.1.1.1 Diaphragm pumps . . . . . . .172.1.2 Liquid sealed rotary

displacement pumps . . . . . .172.1.2.1 Liquid ring pumps . . . . . . . .172.1.2.2 Oil sealed rotary

displacement pumps . . . . . .182.1.2.2.1 Rotary vane pumps

(TRIVAC A, TRIVAC B, TRIVAC E, SOGEVAC) . . . . .18

2.1.2.2.2 Rotary plunger pumps (E-Pumps) . . . . . . . . . . . . .20

2.1.2.2.3 Trochoid pumps . . . . . . . . .212.1.2.2.4 The gas ballast . . . . . . . . . .212.1.3 Dry compressing rotary

displacement pumps . . . . . .242.1.3.1 Roots pumps . . . . . . . . . . .242.1.3.2 Claw pumps . . . . . . . . . . . .272.1.3.2.1 Claw pumps with internal

compression for the semiconductor industry(“DRYVAC Series”) . . . . . . .28

2.1.3.2.2 Claw pump without internalcompression for chemistry applications (“ALL·ex”) . . . .31

2.1.4 Accessories for oil-sealed rotary displacement pumps . . . . . . . . . . . . . . . .33

2.1.5 Condensers . . . . . . . . . . . .332.1.6 Fluid-entrainment pumps . .352.1.6.1 (Oil) Diffusion pumps . . . . .362.1.6.2 Oil vapor ejector pumps . . .382.1.6.3 Pump fluids . . . . . . . . . . . .39

2.1.6.4 Pump fluid backstreaming and its suppression (Vapor barriers, baffles) . . .39

2.1.6.5 Water jet pumps and steam ejectors . . . . . . . . . .40

2.1.7 Turbomolecular pumps . . . .412.1.8 Sorption pumps . . . . . . . . .452.1.8.1 Adsorption pumps . . . . . . .452.1.8.2 Sublimation pumps . . . . . . .462.1.8.3 Sputter-ion pumps . . . . . . .462.1.8.4 Non evaporable getter

pumps (NEG pumps) . . . . .482.1.9 Cryopumps . . . . . . . . . . . . .492.1.9.1 Types of cryopump . . . . . . .492.1.9.2 The cold head and its

operating principle . . . . . . .502.1.9.3 The refrigerator cryopump .512.1.9.4 Bonding of gases to cold

surfaces . . . . . . . . . . . . . . .512.1.9.5 Pumping speed and position

of the cryopanels . . . . . . . .522.1.9.6 Characteristic quantities of a

cryopump . . . . . . . . . . . . . .532.2 Choice of pumping

process . . . . . . . . . . . . . . .562.2.1 Survey of the most usual

pumping processes . . . . . . .562.2.2 Pumping of gases

(dry processes) . . . . . . . . .572.2.3 Pumping of gases and

vapors (wet processes) . . . .582.2.4 Drying processes . . . . . . . .602.2.5 Production of an oil-free

(hydrocarbon-free) vacuum . . . . . . . . . . . . . . .60

2.2.6 Ultrahigh vacuum working Techniques . . . . . . . . . . . . .61

2.3 Evacuation of a vacuum chamber and determination of pump sizes . . . . . . . . . . .62

2.3.1 Evacuation of a vacuum chamber (without additionalsources of gas or vapor) . . .62

2.3.1.1 Evacuation of a chamber in the rough vacuum region . . . . . . . . . . . . . . . . .62

2.3.1.2 Evacuation of a chamber in the high vacuum region . . . . . . . . . . . . . . . . .63

2.3.1.3 Evacuation of a chamber in the medium vacuum region . . . . . . . . . . . . . . . . .63

2.3.2 Determination of a suitablebacking pump . . . . . . . . . . .64

2.3.3 Determination of pump-down time from nomograms . . . . . . . . . . . .65

2.3.4 Evacuation of a chamber where gases and vapors are evolved . . . . . . . . . . . . .66

2.3.5 Selection of pumps for drying processes . . . . . . . .66

2.3.6 Flanges and their seals . . . .672.3.7 Choice of suitable valves . . .682.3.8 Gas locks and seal-off

fittings . . . . . . . . . . . . . . . .69

3. Vacuum measurement, monitoring, control and regulation . . . . . . . . . . . .70

3.1 Fundamentals of low-pressure measurement . . . .70

3.2 Vacuum gauges with pressure reading that is independent of the type of gas . . . . . . . . . . . . . . . . .72

3.2.1 Bourdon vacuum gauges . . .723.2.2 Diaphragm vacuum

gauges . . . . . . . . . . . . . . . .723.2.2.1 Capsule vacuum gauges . . .723.2.2.2 DIAVAC diaphragm vacuum

gauge . . . . . . . . . . . . . . . . .723.2.2.3 Precision diaphragm

vacuum gauges . . . . . . . . . .723.2.2.4 Capacitance diaphragm

gauges . . . . . . . . . . . . . . . .733.2.3 Liquid-filled (mercury)

vacuum gauges . . . . . . . . . .743.2.3.1 U-tube vacuum gauges . . . .743.2.3.2 Compression vacuum

gauges (according to McLeod) . . . . . . . . . . . . . . .74

3.3 Vacuum gauges with gas-dependent pressure reading . . . . . . . . . . . . . . . .75

3.3.1 Spinning rotor gauge (SRG) (VISCOVAC) . . . . . . .75

3.3.2 Thermal conductivity vacuum gauges . . . . . . . . . .76

3.3.3 Ionization vacuum gauges . .773.3.3.1 Cold-cathode ionization

vacuum gauges (Penning vacuum gauges) . . . . . . . . .77

3.3.3.2 Hot-cathode ionization vacuum gauges . . . . . . . . . .78

3.4 Adjustment and calibration; DKD, PTB national standards . . . . . . . . . . . . . .81

3.4.1 Examples of fundamental pressure measurement methods (as standard methods for calibrating vacuum gauges . . . . . . . . . .81

3.5 Pressure monitoring, control and regulation in vacuumsystems . . . . . . . . . . . . . . .83

D00 E 19.06.2001 21:35 Uhr Seite 3

ContentsFundamentals of Vacuum Technology


3.5.1 Fundamentals of pressure monitoring and control . . . .83

3.5.2 Automatic protection, monitoring and control of vacuum systems . . . . . . .83

3.5.3 Pressure regulation and control in rough and medium vacuum systems . . . . . . . . . . . . . . .84

3.5.4 Pressure regulation in high and ultrahigh vacuum systems . . . . . . . . .87

3.5.5 Examples of applications with diaphragm controllers . . . . . . . . . . . . .88

4. Analysis of gas at low pressures using mass spectrometry . . . . . . . . . .89

4.1 General . . . . . . . . . . . . . . . .894.2 A historical review . . . . . . . .894.3 The quadrupole mass

spectrometer (TRANSPECTOR) . . . . . . . .89

4.3.1 Design of the sensor . . . . . .904.3.1.1 The normal (open)

ion source . . . . . . . . . . . . .904.3.1.2 The quadrupole

separation system . . . . . . . .914.3.1.3 The measurement

system (detector) . . . . . . . .924.4 Gas admission and

pressure adaptation . . . . . .934.4.1 Metering valve . . . . . . . . . .934.4.2 Pressure converter . . . . . . .934.4.3 Closed ion source (CIS) . . .934.4.4 Aggressive gas monitor

(AGM) . . . . . . . . . . . . . . . .944.5 Descriptive values in

mass spectrometry (specifications) . . . . . . . . . .94

4.5.1 Line width (resolution) . . . .944.5.2 Mass range . . . . . . . . . . . . .954.5.3 Sensitivity . . . . . . . . . . . . . .954.5.4 Smallest detectable

partial pressure . . . . . . . . . .954.5.5 Smallest detectable

partial pressure ratio (concentration) . . . . . . . . . .95

4.5.6 Linearity range . . . . . . . . . .954.5.7 Information on surfaces

and amenability to bake-out . . . . . . . . . . . . . . .95

4.6 Evaluating spectra . . . . . . . .964.6.1 Ionization and fundamental

problems in gas analysis . . .964.6.2 Partial pressure

measurement . . . . . . . . . .1004.6.3 Qualitative gas analysis . . .100

4.6.4 Quantitative gas analysis . .1014.7 Software . . . . . . . . . . . . . .1024.7.1 Standard SQX software

(DOS) for stand-alone operation(1 MS plus, 1 PC, RS 232) . . . . . . . . . .102

4.7.2 Multiplex/DOS software MQX (1 to 8 MS plus 1 PC, RS 485) . . . . . . . . . .102

4.7.3 Process-oriented software –Transpector-Ware for Windows . . . . . . . . . . . . .102

4.7.4 Development software TranspectorView . . . . . . . .102

4.8 Partial pressure regulation . . . . . . . . . . . . .102

4.9 Maintenance . . . . . . . . . . .103

5. Leaks and their detection . . . . . . . . . . .104

5.1 Types of leaks . . . . . . . . . .1045.2 Leak rate, leak size,

mass flow . . . . . . . . . . . . .1045.2.1 The standard helium

leak rate . . . . . . . . . . . . . .1065.2.2 Conversion equations . . . .1065.3 Terms and definitions . . . .1065.4 Leak detection methods

without a leak detector unit . . . . . . . . . . . . . . . . . .107

5.4.1 Pressure rise test . . . . . . .1075.4.2 Pressure drop test . . . . . .1085.4.3 Leak test using vacuum

gauges which are sensitive to the type of gas . . . . . . .108

5.4.4 Bubble immersion test . . .1095.4.5 Foam-spray test . . . . . . . .1095.4.6 Vacuum box check bubble .1095.4.7 Krypton 85 test . . . . . . . . .1095.4.8 High-frequency vacuum

test . . . . . . . . . . . . . . . . . .1095.4.9 Testing with chemical

reactions and dye penetration . . . . . . . . . . . .110

5.5 Leak detectors and how they work . . . . . . . . . . . . .110

5.5.1 Halogen leak detectors (HLD 4000, D-Tek) . . . . . .110

5.5.2 Leak detectors withmass spectrometers (MS) .110

5.5.2.1 The operating principle for a MSLD . . . . . . . . . . . .111

5.5.2.2 Detection limit, background, gas storage in oil (gas ballast),floating zero-point suppression . . . . . . . . . . .111

5.5.2.3 Calibrating leak detectors; test leaks . . . . . . . . . . . . .112

5.5.2.4 Leak detectors with quadrupole mass spectrometer (ECOTEC II) . . . . . . . . . . .113

5.5.2.5 Helium leak detectors with 180° sector mass spectrometer (UL 200, UL 500) . . . . . . .114

5.5.2.6 Direct-flow and counter-flow leakdetectors . . . . . . . . . . . . .115

5.5.2.7 Partial flow operation . . . .1155.5.2.8 Connection to vacuum

systems . . . . . . . . . . . . . .1165.5.2.9 Time constants . . . . . . . .1165.6 Limit values / Specifications

for the leak detector . . . . .1175.7 Leak detection techniques

using helium leak detectors . . . . . . . . . . . . .117

5.7.1 Spray technique (local leak test) . . . . . . . . .117

5.7.2 Sniffer technology (local leak testing using the positive pressure method) . . . . . . . . . . . . . .118

5.7.3 Vacuum envelope test (integral leak test) . . . . . . .118

5.7.3.1 Envelope test – test specimen pressurized with helium . . . . . . . . . . . .118a) Envelope test with

concentration measurement and subsequent leak rate calculation . . .118

b) Direct measurement of the leak rate with the leak detector (rigid envelope) . . . . . .118

5.7.3.2 Envelope test with test specimen evacuated . . . . .118a) Envelope =

“plastic tent” . . . . . . . .118b) Rigid envelope . . . . . . .119

5.7.4 “Bombing” test, “Storage under pressure” .119

5.8 Industrial leak testing . . . .119

6. Thin film controllers and control units with quartz oscillators . . . . . . . . . . .120

6.1 Introduction . . . . . . . . . . .1206.2 Basic principles of coating

thickness measurement with quartz oscillators . . . .120

6.3 The shape of quartz oscillator crystals . . . . . . .121

6.4 Period measurement . . . . .1226.5 The Z match technique . . .122

D00 E 19.06.2001 21:35 Uhr Seite 4

Contents Fundamentals of Vacuum Technology


6.6 The active oscillator . . . . .1226.7 The mode-lock oscillator . .1236.8 Auto Z match technique . .1246.9 Coating thickness

regulation . . . . . . . . . . . . .1256.10 INFICON instrument

variants . . . . . . . . . . . . . .127

7. Application of vacuum technology for coating techniques . . . . . . . . . .128

7.1 Vacuum coating technique . . . . . . . . . . . . .128

7.2 Coating sources . . . . . . . .1287.2.1 Thermal evaporators

(boats, wires etc.) . . . . . . .1287.2.2 Electron beam

evaporators (electron guns) . . . . . . . . .129

7.2.3 Cathode sputtering . . . . . .1297.2.4 Chemical vapor

deposition . . . . . . . . . . . . .1297.3 Vacuum coating

technology/coating systems . . . . . . . . . . . . . .130

7.3.1 Coating of parts . . . . . . . .1307.3.2 Web coating . . . . . . . . . . .1307.3.3 Optical coatings . . . . . . . .1317.3.4 Glass coating . . . . . . . . . .1327.3.5 Systems for producing

data storage disks . . . . . . .132

8. Instructions for vacuum equipment operation . . . .134

8.1 Causes of faults where the desired ultimate pressure is not achieved or is achieved too slowly . . . . .134

8.2 Contamination of vacuum vessels and eliminating contamination . . . . . . . . . .134

8.3 General operating infor-mation for vacuum pumps . . . . . . . . . . . . . . .134

8.3.1 Oil-sealed rotary vacuum pumps (Rotary vane pumps and rotary pistonpumps) . . . . . . . . . . . . . . .135

8.3.1.1 Oil consumption, oil contamination, oil change . . . . . . . . . . . . . . .135

8.3.1.2 Selection of the pump oil when handling aggressive vapors . . . . . . .135

8.3.1.3 Measures when pumping various chemical substances . . . . . . . . . . . .136

8.3.1.4 Operating defects while pumping with gas ballast –potential sources of error where the required ultimate pressure is not achieved . .137

8.3.2 Roots pumps . . . . . . . . . .1378.3.2.1 General operating

instructions, installation and commissioning . . . . . .137

8.3.2.2 Oil change, maintenance work . . . . . . . . . . . . . . . . .137

8.3.2.3 Actions in case of operational disturbances . .138

8.3.3 Turbomolecular pumps . . .1388.3.3.1 General operating

instructions . . . . . . . . . . . .1388.3.3.2 Maintenance . . . . . . . . . . .1388.3.4 Diffusion and vapor-jet

vacuum pumps . . . . . . . . .1398.3.4.1 Changing the pump fluid

and cleaning the pump . . .1398.3.4.2 Operating errors with

diffusion and vapor-jet pumps . . . . . . . . . . . . . . .139

8.3.5 Adsorption pumps . . . . . .1398.3.5.1 Reduction of adsorption

capacity . . . . . . . . . . . . . .1398.3.5.2 Changing the molecular

sieve . . . . . . . . . . . . . . . . .1398.3.6 Titanium sublimation

pumps . . . . . . . . . . . . . . .1408.3.7 Sputter-ion pumps . . . . . .1408.4 Information on working

with vacuum gauges . . . . .1408.4.1 Information on installing

vacuum sensors . . . . . . . .1408.4.2 Contamination at the

measurement system and its removal . . . . . . . . . . . .141

8.4.3 The influence of magnetic and electrical fields . . . . . .141

8.4.4 Connectors, power pack, measurement systems . . .141

9. Tables, formulas, nomograms, diagrams and symbols . . . . . . . . .142

Tab I Permissible pressure units including the torr and its conversion . . . . . . . . . . . .142

Tab II Conversion of pressure units . . . . . . . . . . . . . . . . .142

Tab III Mean free path . . . . . . . . .142Tab IV Compilation of important

formulas pertaining to the kinetic theory of gases . . .143

Tab V Important values . . . . . . . .143

Tab VI Conversion of pumping speed (volume flow rate) units . . . . . . . . . . . . . . . . .144

Tab VII Conversion of throughput(a,b) QpV units; leak rate units . . . . . . . . . . . . . . . . .144

Tab VIII Composition of atmospheric air . . . . . . . . .145

Tab IX Pressure ranges used in vacuum technology and theircharacteristics . . . . . . . . . .145

Tab X Outgassing rate of materials . . . . . . . . . . . . . .145

Tab XI Nominal internal diameters (DN) and internal diameters of tubes, pipes and apertureswith circular cross-section(according to PNEUROP). .146

Tab XII Important data for common solvents . . . . . . .146

Tab XIII Saturation pressure and density of water . . . . . . . .147

Tab XIV Hazard classificationof fluids . . . . . . . . . . . . . . . .148

Tab XV Chemical resistance of commonly used elastomer gaskets and sealingmaterials . . . . . . . . . . . . .149

Tab XVI Symbols used invacuum technology . . . . . . . . . . . .152

Tab XVII Temperature comparison and conversion table . . . . .154

Fig. 9.1 Variation of mean free path l (cm) with pressure for various gases . . . . . . . . . .154

Fig. 9.2 Diagram of kinetics of gases for air at 20 °C . . . .154

Fig. 9.3 Decrease in air pressure and change in temperature as a function of altitude . . . . . .155

Fig. 9.4 Change in gas composition of the atmosphere as a function of altitude . . . . . .155

Fig. 9.5 Conductance values for piping of commonly used nominal internal diameters with circular cross-section for molecular flow . . . . . . .155

Fig. 9.6 Conductance values for piping of commonly used nominal internal diameters with circular cross-section for molecular flow . . . . . . .155

Fig. 9.7 Nomogram for determination of pump-down time tp of a vessel in the rough vacuum pressure range . . .156

D00

D00 E 19.06.2001 21:35 Uhr Seite 5

ContentsFundamentals of Vacuum Technology


Fig. 9.8 Nomogram for determination of the conductance of tubes with a circular cross-section for air at 20 °C in the region of molecular flow . .157

Fig. 9.9 Nomogram for determination of conductance of tubes in the entire pressure range .158

Fig. 9.10 Determination of pump-down time in the medium vacuum range taking into account the evolution of gas from the walls . . . . . . . . . . . . . .159

Fig.9.11 Saturation vapor pressure of various substances . . . . . . . . . . . .160

Fig. 9.12 Saturation vapor pressure of pump fluids for oil and mercury fluid entrainment pumps . . . . . . . . . . . . . . .160

Fig. 9.13 Saturation vapor pressure of major metals used in vacuum technology . . . . . .160

Fig. 9.14 Vapor pressure of nonmetallic sealing materials (the vapor pressure curve for fluoro rubber lies between the curves for silicone rubber and Teflon). . . . . . .161

Fig. 9.15 Saturation vapor pressure ps of various substances relevant for cryogenic technology in a temperaturerange of T = 2 – 80 K. . . . . . . . . . . .161

Fig. 9.16 Common working ranges of vacuum pumps . . . . . . .161

Fig. 9.16a Measurement ranges of common vacuum gauges . . . . . . . . . . . . . . .162

Fig. 9.17 Specific volume of saturated water vapor . . . .163

Fig. 9.18 Breakdown voltage between electrodes for air(Paschen curve) . . . . . . . .163

Fig 9.19 Phase diagram of water . . .164

10. The statutory units used invacuum technology . . . . .165

10.1 Introduction . . . . . . . . . . .16510.2 Alphabetical list of variables,

symbols and units frequentlyused in vacuum technology and its applications . . . . . .165

10.3 Remarks on alphabetical list in Section 10.2 . . . . . .168

10.4 Tables . . . . . . . . . . . . . . . .17010.4.1 Basic SI units . . . . . . . . . .17010.4.2 Derived coherent SI units

with special names and symbols . . . . . . . . . . .170

10.4.3 Atomic units . . . . . . . . . .17010.4.4 Derived noncoherent SI

units with special names and symbols . . . . . . . . . . .170

11. National and internationalstandards and recommendations particularly relevant to vacuum technology . . . . .171

11.1 National and international standards and recommendations of special relevance to vacuum technology . . . . . .171

12. References . . . . . . . . . .174

13. Index . . . . . . . . . . . . . .184

D00 E 19.06.2001 21:35 Uhr Seite 6

Vacuum Physics Fundamentals of Vacuum Technology


1. Quantities,their symbols,units of measure anddefinitions

(cf. DIN 28 400, Part 1, 1990, DIN 1314 and DIN 28 402)

1.1 Basic terms andconcepts in vacuum technology

Pressure p (mbar)of fluids (gases and liquids). (Quantity:pressure; symbol: p; unit of measure:millibar; abbreviation: mbar.) Pressure isdefined in DIN Standard 1314 as thequotient of standardized force applied to asurface and the extent of this surface(force referenced to the surface area).Even though the Torr is no longer used asa unit for measuring pressure (see Section10), it is nonetheless useful in the interestof “transparency” to mention this pressureunit: 1 Torr is that gas pressure which isable to raise a column of mercury by 1 mmat 0 °C. (Standard atmospheric pressure is760 Torr or 760 mm Hg.) Pressure p canbe more closely defined by way of sub-scripts:

Absolute pressure pabsAbsolute pressure is always specified invacuum technology so that the “abs” indexcan normally be omitted.

Total pressure ptThe total pressure in a vessel is the sum ofthe partial pressures for all the gases andvapors within the vessel.

Partial pressure piThe partial pressure of a certain gas orvapor is the pressure which that gas orvapor would exert if it alone were presentin the vessel.

Important note: Particularly in roughvacuum technology, partial pressure in amix of gas and vapor is often understoodto be the sum of the partial pressures for

all the non-condensable componentspresent in the mix – in case of the “parti-al ultimate pressure” at a rotary vanepump, for example.

Saturation vapor pressure psThe pressure of the saturated vapor isreferred to as saturation vapor pressureps. ps will be a function of temperature forany given substance.

Vapor pressure pdPartial pressure of those vapors which canbe liquefied at the temperature of liquidnitrogen (LN2).

Standard pressure pnStandard pressure pn is defined in DIN1343 as a pressure of pn = 1013.25 mbar.

Ultimate pressure pendThe lowest pressure which can be achie-ved in a vacuum vessel. The socalled ulti-mate pressure pend depends not only onthe pump’s suction speed but also uponthe vapor pressure pd for the lubricants,sealants and propellants used in the pump.If a container is evacuated simply with anoil-sealed rotary (positive displacement)vacuum pump, then the ultimate pressurewhich can be attained will be determinedprimarily by the vapor pressure of thepump oil being used and, depending onthe cleanliness of the vessel, also on thevapors released from the vessel walls and,of course, on the leak tightness of thevacuum vessel itself.

Ambient pressure pambor (absolute) atmospheric pressure

Overpressure pe or gauge pressure(Index symbol from “excess”)

pe = pabs – pamb

Here positive values for pe will indicateoverpressure or gauge pressure; negativevalues will characterize a vacuum.

Working pressure pWDuring evacuation the gases and/or vaporsare removed from a vessel. Gases areunderstood to be matter in a gaseous statewhich will not, however, condense atworking or operating temperature. Vaporis also matter in a gaseous state but it maybe liquefied at prevailing temperatures byincreasing pressure. Finally, saturatedvapor is matter which at the prevailing

temperature is gas in equilibrium with theliquid phase of the same substance. Astrict differentiation between gases andvapors will be made in the commentswhich follow only where necessary forcomplete understanding.

Particle number density n (cm-3)According to the kinetic gas theory thenumber n of the gas molecules, referencedto the volume, is dependent on pressure pand thermodynamic temperature T asexpressed in the following:

p = n · k · T (1.1)

n = particle number densityk = Boltzmann’s constant

At a certain temperature, therefore, thepressure exerted by a gas depends only onthe particle number density and not on thenature of the gas. The nature of a gaseousparticle is characterized, among otherfactors, by its mass mT.

Gas density ρ (kg · m-3, g · cm-3)The product of the particle number densi-ty n and the particle mass mT is the gasdensity ρ:

ρ = n · mT (1.2)

The ideal gas lawThe relationship between the mass mT of agas molecule and the molar mass M of thisgas is as follows:

M = NA · mT (1.3)

Avogadro’s number (or constant) NAindicates how many gas particles will becontained in a mole of gas. In addition tothis, it is the proportionality factor betweenthe gas constant R and Boltzmann’sconstant k:

R = NA · k (1.4)

Derivable directly from the aboveequations (1.1) to (1.4) is the correlationbetween the pressure ρ and the gasdensity ρ of an ideal gas.

R · Tp = ρ ⋅ ––––– (1.5)

M

In practice we will often consider a certainenclosed volume V in which the gas ispresent at a certain pressure p. If m is themass of the gas present within thatvolume, then

D00

D00 E 19.06.2001 21:35 Uhr Seite 7

ture this is always equivalent to reducingthe gas pressure p. Explicit attention mustat this point be drawn to the fact that areduction in pressure (maintaining thevolume) can be achieved not only byreducing the particle number density n butalso (in accordance with Equation 1.5) byreducing temperature T at constant gasdensity. This important phenomenon willalways have to be taken into accountwhere the temperature is not uniformthroughout volume V.

The following important terms andconcepts are often used in vacuumtechnology:

Volume V (l, m3, cm3)The term volume is used to designatea) the purely geometric, usually predeter-

mined, volumetric content of a vacuumchamber or a complete vacuum systemincluding all the piping and connectingspaces (this volume can be calculated);

b) the pressure-dependent volume of agas or vapor which, for example, ismoved by a pump or absorbed by anadsorption agent.

Volumetric flow (flow volume) qv(l/s, m3/h, cm3/s)The term “flow volume” designates thevolume of the gas which flows through apiping element within a unit of time, at thepressure and temperature prevailing at theparticular moment. Here one must realizethat, although volumetric flow may beidentical, the number of molecules movedmay differ, depending on the pressure andtemperature.

Pumping speed S (l/s, m3/h, cm3/s)The pumping speed is the volumetric flowthrough the pump’s intake port.

dVS = –––– (1.8a)

dt

If S remains constant during the pumpingprocess, then one can use the differencequotient instead of the differential quoti-ent:

∆VS = ––– (1.8b)

∆t

(A conversion table for the various units ofmeasure used in conjunction with pumping speed is provided in Section 9,Table VI).

mρ = ––– (1.6)

V

The ideal gas law then follows directlyfrom equation (1.5):

mp ⋅ V = ––– ⋅ R ⋅ T = ν ⋅ R ⋅ T (1.7)

M

Here the quotient m / M is the number ofmoles u present in volume V.

The simpler form applies for m / M = 1, i.e.for 1 mole:

p · V = R · T (1.7a)

The following numerical example isintended to illustrate the correlationbetween the mass of the gas and pressurefor gases with differing molar masses,drawing here on the numerical values inTable IV (Chapter 9). Contained in a 10-liter volume, at 20 °C, will be

a) 1g of heliumb) 1g of nitrogen

When using the equation (1.7) thereresults then at V = 10 l , m = 1 g, R = 83.14 mbar·l·mol-1·K-1, T = 293 K (20 °C)

In case a) where M = 4 g · mole-1

(monatomic gas):

In case b), with M = 28 ≠ g mole-1 (diato-mic gas):

The result, though appearing to beparadoxical, is that a certain mass of alight gas exerts a greater pressure than thesame mass of a heavier gas. If one takesinto account, however, that at the samegas density (see Equation 1.2) moreparticles of a lighter gas (large n, small m)will be present than for the heavier gas(small n, large m), the results becomemore understandable since only theparticle number density n is determinantfor the pressure level, assuming equaltemperature (see Equation 1.1).

The main task of vacuum technology is toreduce the particle number density n insi-de a given volume V. At constant tempera-

Vacuum PhysicsFundamentals of Vacuum Technology

D00.08

Quantity of gas (pV value), (mbar ⋅ l)The quantity of a gas can be indicated byway of its mass or its weight in the units ofmeasure normally used for mass orweight. In practice, however, the productof p · V is often more interesting in vacu-um technology than the mass or weight ofa quantity of gas. The value embraces anenergy dimension and is specified inmillibar·liters (mbar·l) (Equation 1.7).Where the nature of the gas and itstemperature are known, it is possible touse Equation 1.7b to calculate the mass mfor the quantity of gas on the basis of theproduct of p · V:

(1.7)

(1.7b)

Although it is not absolutely correct,reference is often made in practice to the“quantity of gas” p · V for a certain gas.This specification is incomplete; thetemperature of the gas T, usually roomtemperature (293 K), is normally implicitlyassumed to be known.

Example:The mass of 100 mbar · l of nitrogen (N2)at room temperature (approx. 300 K) is:

Analogous to this, at T = 300 K:

1 mbar · l O2 = 1.28 · 10-3 g O2

70 mbar · l Ar = 1.31 · 10-1 g Ar

The quantity of gas flowing through apiping element during a unit of time – inaccordance with the two concepts for gasquantity described above – can be indica-ted in either of two ways, these being:

Mass flow qm (kg/h, g/s),this is the quantity of a gas which flowsthrough a piping element, referenced to time

mqm = ––– or as

t

mmbar g mol

mbar mol K K=

· · ·

· · · ·

−

− −100 28

83 300

1

1 1`

`

g g=·

=

=

2800300 83

0 113.

m p V MR T

= · ··

p V mM

R T· = · ·

mbar=

=

609

pg mbar mol K K

K g mol=

·· · · ·

·· · ·

– –

–

1 83 14 293

10 4

1 1

1

. · · ·

·

`

`

mbar=

=

87

pg mbar mol K K

K g mol=

·· · · ·

·· · ·

– –

–

1 83 14 293

10 28

1 1

1

. · · ·

·

`

`

LEYBOLD VACUUM PRODUCTS AND REFERENCE BOOK 2001/2002

D00 E 19.06.2001 21:35 Uhr Seite 8



pV flow qpV (mbar·l·s-1).pV flow is the product of the pressure andvolume of a quantity of gas flowingthrough a piping element, divided by time,i.e.:

p · V d (p · V)qpV = –––––– = ––––––––

t dt

pV flow is a measure of the mass flow ofthe gas; the temperature to be indicatedhere.

Pump throughput qpVThe pumping capacity (throughput) for apump is equal either to the mass flowthrough the pump intake port:

mqm = –––– (1.9)

t

or to the pV flow through the pump’s inta-ke port:

p · VqpV = ––––––

t

It is normally specified in mbar·l·s-1. Herep is the pressure on the intake side of thepump. If p and V are constant at the intakeside of the pump, the throughput of thispump can be expressed with the simpleequation

qpV = p · S (1.10a)

where S is the pumping speed of the pumpat intake pressure of p.

(The throughput of a pump is often indica-ted with Q, as well.)

The concept of pump throughput is ofmajor significance in practice and shouldnot be confused with the pumping speed!The pump throughput is the quantity ofgas moved by the pump over a unit oftime, expressed in mbar ≠ l/s; the pum-ping speed is the “transportation capacity”which the pump makes available within aspecific unit of time, measured in m3/h or l/s.The throughput value is important in deter-mining the size of the backing pump inrelationship to the size of a high vacuumpump with which it is connected in seriesin order to ensure that the backing pumpwill be able to “take off” the gas moved bythe high vacuum pump (see Section 2.32).

Conductance C (l · s-1)The pV flow through any desired pipingelement, i.e. pipe or hose, valves, nozzles,openings in a wall between two vessels,etc., is indicated with

qpV = C(p1 – p2) = ∆p · C (1.11)

Here ∆p = (p1 – p2) is the differential bet-ween the pressures at the inlet and outletends of the piping element. The proportio-nality factor C is designated as the con-ductance value or simply “conductance”. Itis affected by the geometry of the pipingelement and can even be calculated forsome simpler configurations (see Section1.5).

In the high and ultrahigh vacuum ranges,C is a constant which is independent ofpressure; in the rough and medium-highregimes it is, by contrast, dependent onpressure. As a consequence, the calculati-on of C for the piping elements must becarried out separately for the individualpressure ranges (see Section 1.5 for moredetailed information).

From the definition of the volumetric flowit is also possible to state that: Theconductance value C is the flow volumethrough a piping element. The equation(1.11) could be thought of as “Ohm’s lawfor vacuum technology”, in which qpVcorresponds to current, ∆p the voltage andC the electrical conductance value. Analo-gous to Ohm’s law in the science of elec-tricity, the resistance to flow

1R = –––

C

has been introduced as the reciprocalvalue to the conductance value. The equation (1.11) can then be re-written as:

1qpV = —— ⋅ ∆p (1.12)

R

The following applies directly for connec-tion in series:

R∑ = R1 + R2 + R3 . . . (1.13)

When connected in parallel, the followingapplies:

1 1 1 1––– = –– + –– + –– + ⋅ ⋅ ⋅ ⋅ (1.13a)R∑ R1 R2 R3

Leak rate qL (mbar·l·s-1)According to the definition formulatedabove it is easy to understand that the sizeof a gas leak, i.e. movement through unde-sired passages or “pipe” elements, willalso be given in mbar·l·s-1. A leak rate isoften measured or indicated with atmos-pheric pressure prevailing on the one sideof the barrier and a vacuum at the otherside (p < 1 mbar). If helium (which may beused as a tracer gas, for example) is pas-sed through the leak under exactly theseconditions, then one refers to “standardhelium conditions”.

Outgassing (mbar·l)The term outgassing refers to the liberati-on of gases and vapors from the walls of avacuum chamber or other components onthe inside of a vacuum system. This quan-tity of gas is also characterized by the pro-duct of p · V, where V is the volume of thevessel into which the gases are liberated,and by p, or better ∆p, the increase inpressure resulting from the introduction ofgases into this volume.

Outgassing rate (mbar·l·s-1)This is the outgassing through a period oftime, expressed in mbar·l·s-1.

Outgassing rate (mbar·l·s-1 · cm-2)(referenced to surface area)In order to estimate the amount of gaswhich will have to be extracted, knowledgeof the size of the interior surface area, itsmaterial and the surface characteristics,their outgassing rate referenced to the sur-face area and their progress through timeare important.

Mean free path of the molecules λ (cm)and collision rate z (s-1)The concept that a gas comprises a largenumber of distinct particles between which– aside from the collisions – there are noeffective forces, has led to a number oftheoretical considerations which we sum-marize today under the designation “kine-tic theory of gases”.

One of the first and at the same time mostbeneficial results of this theory was thecalculation of gas pressure p as a functionof gas density and the mean square ofvelocity c2 for the individual gas moleculesin the mass of molecules mT: D00

D00 E 19.06.2001 21:35 Uhr Seite 9



1 --- 1 ---p = –– ρ ⋅ c2 = –– ⋅ n ⋅ mT ⋅ c2 (1.14)

3 3

where

k · Tc2 = 3 ⋅ ——– (1.15)

mT

The gas molecules fly about and amongeach other, at every possible velocity, andbombard both the vessel walls and collide(elastically) with each other. This motion ofthe gas molecules is described numerical-ly with the assistance of the kinetic theoryof gases. A molecule’s average number ofcollisions over a given period of time, theso-called collision index z, and the meanpath distance which each gas moleculecovers between two collisions with othermolecules, the so-called mean free pathlength λ, are described as shown below asa function of the mean molecule velocity c-

the molecule diameter 2r and the particlenumber density molecules n – as a verygood approximation:

CZ = (1.16)

λ

where

(1.17)

and

(1.18)

Thus the mean free path length λ for theparticle number density n is, in accordan-ce with equation (1.1), inversely proportio-nal to pressure p. Thus the following rela-tionship holds, at constant temperature T,for every gas

λ ⋅ p = const (1.19)

Used to calculate the mean free pathlength λ for any arbitrary pressures andvarious gases are Table III and Fig. 9.1 inChapter 9. The equations in gas kineticswhich are most important for vacuumtechnology are also summarized (Table IV)in chapter 9.

( )λ

π=

· · ·1

2 2 2n r

c k Tm

R TMT

= · ··

= · ··

8 8π π

Impingement rate zA (cm-2 ⋅ s-1) andmonolayer formation time τ (s)A technique frequently used to characteri-ze the pressure state in the high vacuumregime is the calculation of the time requi-red to form a monomolecular or monoato-mic layer on a gas-free surface, on theassumption that every molecule will stickto the surface. This monolayer formationtime is closely related with the so-calledimpingement rate zA. With a gas at rest theimpingement rate will indicate the numberof molecules which collide with the surfa-ce inside the vacuum vessel per unit oftime and surface area:

(1.20)

If a is the number of spaces, per unit ofsurface area, which can accept a specificgas, then the monolayer formation time is

(1.21)

Collision frequency zv (cm-3 · s-1)This is the product of the collision rate zand the half of the particle number densityn, since the collision of two molecules is tobe counted as only one collision:

(1.21a)z n zV = ·2

τ = = ··

az

an cA

4

z n cA = ·

4

1.2 Atmospheric airPrior to evacuation, every vacuum systemon earth contains air and it will always besurrounded by air during operation. Thismakes it necessary to be familiar with thephysical and chemical properties ofatmospheric air.

The atmosphere is made up of a number ofgases and, near the earth’s surface, watervapor as well. The pressure exerted byatmospheric air is referenced to sea level.Average atmospheric pressure is 1013 mbar (equivalent to the “atmos-phere”, a unit of measure used earlier).Table VIII in Chapter 9 shows the compo-sition of the standard atmosphere at relati-ve humidity of 50 % and temperature of 20 °C. In terms of vacuum technology thefollowing points should be noted in regardto the composition of the air:a) The water vapor contained in the air,

varying according to the humidity level,plays an important part when evacua-ting a vacuum plant (see Section 2.2.3).

b) The considerable amount of the inertgas argon should be taken into accountin evacuation procedures using sorpti-on pumps (see Section 2.1.8).

c) In spite of the very low content of heli-um in the atmosphere, only about 5ppm (parts per million), this inert gasmakes itself particularly obvious inultrahigh vacuum systems which aresealed with Viton or which incorporateglass or quartz components. Helium isable to permeate these substances to ameasurable extent.

The pressure of atmospheric air falls withrising altitude above the earth’s surface(see Fig. 9.3 in Chapter 9). High vacuumprevails at an altitude of about 100 km andultrahigh vacuum above 400 km. The com-position of the air also changes with thedistance to the surface of the earth (seeFig. 9.4 in Chapter 9).

D00 E 19.06.2001 21:35 Uhr Seite 10



1.3 Gas laws andmodels

1.3.1 Continuum theory

Model concept: Gas is “pourable” (fluid)and flows in a way similar to a liquid. Thecontinuum theory and the summarizationof the gas laws which follows are based onexperience and can explain all the proces-ses in gases near atmospheric pressure.Only after it became possible using everbetter vacuum pumps to dilute the air tothe extent that the mean free path rose farbeyond the dimensions of the vessel weremore far-reaching assumptions necessary;these culminated in the kinetic gas theory.The kinetic gas theory applies throughoutthe entire pressure range; the continuumtheory represents the (historically older)special case in the gas laws where atmos-pheric conditions prevail.

Summary of the most important gas laws(continuum theory)

Boyle-Mariotte Law

p ⋅ V = const.

for T = constant (isotherm)

Gay-Lussac’s Law (Charles’ Law)

for p = constant (isobar)

Amonton’s Law

for V = constant (isochor)

Dalton’s Law

Poisson’s Law

p ⋅ Vk = const (adiabatic)

Avogadro’s LawmV

mV

M M1

1

2

21 2: :=

p ptotalii=∑

p p t= + ·0 1( )γ

V V t= + ·0 1( )β

Ideal gas Law

Also: Equation of state for ideal gases(from the continuum theory)

van der Waals’ Equation

a, b = constants (internal pressure, covolumes)

Vm = Molar volumeAlso: Equation of state for real gases

Clausius-Clapeyron Equation

L = Enthalpy of evaporation,T = Evaporation temperature,Vm,v, Vm,l = Molar volumes of vapor

or liquid

1.3.2 Kinetic gas theory

With the acceptance of the atomic view ofthe world – accompanied by the necessityto explain reactions in extremely dilutegases (where the continuum theory fails) –the “kinetic gas theory” was developed.Using this it is possible not only to derivethe ideal gas law in another manner butalso to calculate many other quantitiesinvolved with the kinetics of gases – suchas collision rates, mean free path lengths,monolayer formation time, diffusion con-stants and many other quantities.

Model concepts and basic assumptions:1. Atoms/molecules are points.2. Forces are transmitted from one to

another only by collision.3. The collisions are elastic.4. Molecular disorder (randomness)

prevails.

A very much simplified model was develo-ped by Krönig. Located in a cube are Nparticles, one-sixth of which are movingtoward any given surface of the cube. Ifthe edge of the cube is 1 cm long, then itwill contain n particles (particle numberdensity); within a unit of time n · c · ∆t/6molecules will reach each wall where thechange of pulse per molecule, due to the

L T dpdT

V Vm v m l

= · · −( ), ,

( ) ( )p aV

V b R Tm

m+ · − = ·2

p V mM

R T R T· = · · = · ·ν

change of direction through 180 °, will beequal to 2 · mT · c. The sum of the pulsechanges for all the molecules impinging onthe wall will result in a force effective onthis wall or the pressure acting on the wall,per unit of surface area.

where

Derived from this is

Ideal gas law (derived from the kineticgas theory)If one replaces c2 with c2– then a compari-son of these two “general” gas equationswill show:

or

The expression in brackets on the left-hand side is the Boltzmann constant k; thaton the right-hand side a measure of themolecules’ mean kinetic energy:

Boltzmann constant

Mean kinetic energy of the molecules

thus

In this form the gas equation provides agas-kinetic indication of the temperature!

The mass of the molecules is

m MN

Mass molMolecules molT

A

= = //

p V N k T N E kin· = · · = · ·23

E kin

mT c=

· 2

2

km

TR

MJK

=·

= −138 10 23. ·

p V NmT R

MT N

mT c· = ·

·· = · ·

·( ) ( )2

3 2

2

p V mM

R T N mT c· = · · = · · ·13

2

p V N mT c· = · · ·13

2

n NV

=

n c mT c n c mT p6

2 13

2· · · · = · · · =

D00

D00 E 19.06.2001 21:35 Uhr Seite 11



where NA is Avogadro’s number (previously: Loschmidt number).Avogadro constant

NA = 6.022 ⋅ 1023 mol–1

For 1 mole, and

V = Vm = 22.414 l (molar volume);

Thus from the ideal gas law at standardconditions(Tn = 273.15 K and pn = 1013.25 mbar):

For the general gas constant:

1.4 The pressure ranges in vacuumtechnology and theircharacterization

(See also Table IX in Chapter 9.) It is com-mon in vacuum technology to subdivide itswide overall pressure range – which spansmore than 16 powers of ten – into smallerindividual regimes. These are generallydefined as follows:

Rough vacuum (RV) 1000 – 1 mbarMedium vacuum (MV) 1 – 10-3.mbarHigh vacuum (HV) 10-3 – 10-7 mbarUltrahigh vacuum (UHV)

10-7 – (10-14) mbar

This division is, naturally, somewhat arbit-rary. Chemists in particular may refer tothe spectrum of greatest interest to them,lying between 100 and 1 mbar, as “inter-mediate vacuum”. Some engineers maynot refer to vacuum at all but instead speakof “low pressure” or even “negative pres-sure”. The pressure regimes listed abovecan, however, be delineated quite satisfac-torily from an observation of the gas-kine-tic situation and the nature of gas flow. Theoperating technologies in the various ran-ges will differ, as well.

R mbar molK

mbarmol K

= · ·

=

=

··

−1013 25 22 4273 15

8314

1. .

.

.

`

`

p V mM

R T· = · ·

z nV = ·

2

1.5 Types of flow andconductance

Three types of flow are mainly encounte-red in vacuum technology: viscous or con-tinuous flow, molecular flow and – at thetransition between these two – the so-cal-led Knudsen flow.

1.5.1 Types of flow

Viscous or continuum flowThis will be found almost exclusively in therough vacuum range. The character of thistype of flow is determined by the interac-tion of the molecules. Consequently inter-nal friction, the viscosity of the flowingsubstance, is a major factor. If vortex moti-on appears in the streaming process, onespeaks of turbulent flow. If various layersof the flowing medium slide one over theother, then the term laminar flow or layerflux may be applied.

Laminar flow in circular tubes with para-bolic velocity distribution is known as Poi-seuille flow. This special case is foundfrequently in vacuum technology. Viscousflow will generally be found where themolecules’ mean free path is considerab-ly shorter than the diameter of the pipe:λ « d.

A characteristic quantity describing theviscous flow state is the dimensionlessReynolds number Re.

Re is the product of the pipe diameter, flowvelocity, density and reciprocal value of theviscosity (internal friction) of the gaswhich is flowing. Flow is turbulent whereRe > 2200, laminar where Re < 2200.

The phenomenon of choked flow may alsobe observed in the viscous flow situation.It plays a part when venting and evacua-ting a vacuum vessel and where there areleaks.

Gas will always flow where there is a diffe-rence in pressure ∆p = (p1 – p2) > 0. Theintensity of the gas flow, i.e. the quantity ofgas flowing over a period of time, riseswith the pressure differential. In the caseof viscous flow, however, this will be thecase only until the flow velocity, which alsorises, reaches the speed of sound. This isalways the case at a certain pressure diffe-rential and this value may be characterizedas “critical”:

(1.22)

A further rise in ∆p > ∆pcrit would notresult in any further rise in gas flow; anyincrease is inhibited. For air at 20,°C thegas dynamics theory reveals a criticalvalue of

(1.23)

The chart in Fig. 1.1 represents sche-matically the venting (or airing) of an eva-cuated container through an opening in theenvelope (venting valve), allowing ambientair at p = 1000 mbar to enter. In accordan-ce with the information given above, theresultant critical pressure is ∆pcrit = 1000 ⋅ (1–0.528) mbar ≈ 470 mbar;i.e. where ∆p > 470 mbar the flow rate willbe choked; where ∆p < 470 mbar the gasflow will decline.

Molecular flowMolecular flow prevails in the high andultrahigh vacuum ranges. In these regimesthe molecules can move freely, withoutany mutual interference. Molecular flow ispresent where the mean free path lengthfor a particle is very much larger than thediameter of the pipe: λ >> d.

pp crit

2

10528

= .

∆p pppcrit

crit= −

1

2

11

∆pmbar

qm%

1

2

1000qm∆p

470

0

25

50

75

100

s

venting time (t)

1 Gas flow rate qm choked = constant (maximum value)

2 Gas flow not impeded, qm drops to ∆p = 0

Fig. 1.1 Schematic representation of venting anevacuated vessel

D00 E 19.06.2001 21:35 Uhr Seite 12



Knudsen flowThe transitional range between viscousflow and molecular flow is known as Knud-sen flow. It is prevalent in the mediumvacuum range: λ ≈ d.

The product of pressure p and pipe diame-ter d for a particular gas at a certain tem-perature can serve as a characterizingquantity for the various types of flow.Using the numerical values provided inTable III, Chapter 9, the following equiva-lent relationships exist for air at 20 °C:

Rough vacuum – Viscous flow

⇔ p ⋅ d > 6.0 ⋅ 10-1 mbar ⋅ cm

Medium vacuum – Knudsen flow

⇔

⇔ 6 ⋅ 10-1 > p ⋅ d > 1.3 ⋅ 10-2 mbar ⋅ cm

High and ultrahigh vacuum – Molecularflow

⇔ p ⋅ d < 1.3 ⋅ 10-2 mbar · cm

In the viscous flow range the preferredspeed direction for all the gas moleculeswill be identical to the macroscopic direc-tion of flow for the gas. This alignment iscompelled by the fact that the gas particlesare densely packed and will collide withone another far more often than with theboundary walls of the apparatus. Themacroscopic speed of the gas is a “groupvelocity” and is not identical with the “ther-mal velocity” of the gas molecules.

In the molecular flow range, on the otherhand, impact of the particles with the wallspredominates. As a result of reflection (butalso of desorption following a certain resi-dence period on the container walls) a gasparticle can move in any arbitrary directionin a high vacuum; it is no longer possibleto speak of “flow” in the macroscopicsense.

It would make little sense to attempt todetermine the vacuum pressure ranges asa function of the geometric operatingsituation in each case. The limits for theindividual pressure regimes (see Table IXin Chapter 9) were selected in such a waythat when working with normal-sized labo-

λ > d2

d d100 2

< <λ

λ < d100

ratory equipment the collisions of the gasparticles among each other will predomi-nate in the rough vacuum range whereasin the high and ultrahigh vacuum rangesimpact of the gas particles on the contai-ner walls will predominate.

In the high and ultrahigh vacuum rangesthe properties of the vacuum containerwall will be of decisive importance sincebelow 10-3 mbar there will be more gasmolecules on the surfaces than in thechamber itself. If one assumes a mono-molecular adsorbed layer on the insidewall of an evacuated sphere with 1 l volu-me, then the ratio of the number of adsorbed particles to the number of freemolecules in the space will be as follows:

at 1 mbar 10-2

at 10-6 mbar 10+4

at 10-11 mbar 10+9

For this reason the monolayer formationtime t (see Section 1.1) is used to charac-terize ultrahigh vacuum and to distinguishthis regime from the high vacuum range.The monolayer formation time τ is only afraction of a second in the high vacuumrange while in the ultrahigh vacuum rangeit extends over a period of minutes orhours. Surfaces free of gases can therefo-re be achieved (and maintained over lon-ger periods of time) only under ultrahighvacuum conditions.

Further physical properties change aspressure changes. For example, the ther-mal conductivity and the internal friction ofgases in the medium vacuum range arehighly sensitive to pressure. In the roughand high vacuum regimes, in contrast,these two properties are virtually indepen-dent of pressure.

Thus, not only will the pumps needed toachieve these pressures in the variousvacuum ranges differ, but also differentvacuum gauges will be required. A cleararrangement of pumps and measurementinstruments for the individual pressureranges is shown in Figures 9.16 and 9.16ain Chapter 9.

1.5.2 Calculating conductance values

The effective pumping speed required toevacuate a vessel or to carry out a processinside a vacuum system will correspond tothe inlet speed of a particular pump (or thepump system) only if the pump is joineddirectly to the vessel or system. Practicallyspeaking, this is possible only in rare situa-tions. It is almost always necessary toinclude an intermediate piping system com-prising valves, separators, cold traps andthe like. All this represents an resistance toflow, the consequence of which is that theeffective pumping speed Seff is always lessthan the pumping speed S of the pump orthe pumping system alone. Thus to ensurea certain effective pumping speed at thevacuum vessel it is necessary to select apump with greater pumping speed. Thecorrelation between S and Seff is indicatedby the following basic equation:

(1.24)

Here C is the total conductance value forthe pipe system, made up of the individualvalues for the various components whichare connected in series (valves, baffles,separators, etc.):

(1.25)

Equation (1.24) tells us that only in thesituation where C = ∞ (meaning that theflow resistance is equal to 0) will S = Seff.A number of helpful equations is availableto the vacuum technologist for calculatingthe conductance value C for piping sec-tions. The conductance values for valves,cold traps, separators and vapor barrierswill, as a rule, have to be determined empi-rically.It should be noted that in general that theconductance in a vacuum component isnot a constant value which is independentof prevailing vacuum levels, but ratherdepends strongly on the nature of the flow(continuum or molecular flow; see below)and thus on pressure. When using con-ductance indices in vacuum technologycalculations, therefore, it is always neces-sary to pay attention to the fact that onlythe conductance values applicable to a cer-tain pressure regime may be applied inthat regime.

1 1 1 1 1

1 2 3C C C C Cn= + + + . . .

1 1 1S S Ceff

= +

D00

D00 E 19.06.2001 21:35 Uhr Seite 13



1.5.3 Conductance for pipingand orifices

Conductance values will depend not onlyon the pressure and the nature of the gaswhich is flowing, but also on the sectionalshape of the conducting element (e.g. cir-cular or elliptical cross section). Other fac-tors are the length and whether the ele-ment is straight or curved. The result isthat various equations are required to takeinto account practical situations. Each ofthese equations is valid only for a particu-lar pressure range. This is always to beconsidered in calculations.a) Conductance for a straight pipe, which

is not too short, of length l, with a cir-cular cross section of diameter d for thelaminar, Knudsen and molecular flowranges, valid for air at 20 °C (Knudsenequation):

(1.26)

where

d = Pipe inside diameter in cml = Pipe length in cm (l ≥ 10 d)p1 = Pressure at start of pipe

(along the direction of flow) in mbarp2 = Pressure at end of pipe

(along the direction of flow) in mbar

If one rewrites the second term in (1.26) inthe following form

(1.26a)

with

(1.27)

it is possible to derive the two importantlimits from the course of the function f (d · p–):

Limit for laminar flow(d · p– > 6 · 10-1 mbar · cm):

(1.28a)

Limit for molecular flow(d · p– < 6 · 10-2 mbar · cm):

Cdl p s= · ·1354

` /

( )f d pd p d p

d p· = + · · + · · ·

+ · ·1 203 2 10

1 237

3 2 2.78

( )Cdl f d p= · · ·12 13

.

pp p

=+1 2

2

C dl

p= +1354 d

ld pd p

s· + · ·+ · ·

12 1 1 1921 237

3. /`

(1.28b)

In the molecular flow region the conduc-tance value is independent of pressure!The complete Knudsen equation (1.26) willhave to be used in the transitional area10–2 < d · p– < 6 · 10-1 mbar · cm. Conduc-tance values for straight pipes of standardnominal diameters are shown in Figure 9.5(laminar flow) and Figure 9.6 (molecularflow) in Chapter 9. Additional nomogramsfor conductance determination will also befound in Chapter 9 (Figures 9.8 and 9.9).

b) Conductance value C for an orifice A(A in cm2): For continuum flow (viscousflow) the following equations (afterPrandtl) apply to air at 20 °C wherep2/p1 = δ:

for δ ≥ 0.528 (1.29)

for δ ≤ 0.528 (1.29a)

and for δ ≤ 0,03 (1.29b)

δ = 0.528 is the critical pressure situation for air

Flow is choked at δ < 0.528; gas flow isthus constant. In the case of molecularflow (high vacuum) the following will applyfor air:

Cmol = 11,6 · A · l · s-1 (A in cm2) (1.30)

Given in addition in Figure 1.3 are the

pp

crit

2

1

C A svisc = ·20`

C Asvisc = ·

−20

1 δ`

CA

svisc = · · − · −76 6 1 10.712 0.288. δ δ δ

`

Cdl

s= ·12 13

. /`pumping speeds S*visc and S*mol refer-enced to the area A of the opening and asa function of δ = p2/p1. The equationsgiven apply to air at 20 °C. The molar mas-ses for the flowing gas are taken into con-sideration in the general equations, notshown here.

When working with other gases it will benecessary to multiply the conductancevalues specified for air by the factorsshown in Table 1.1.

Fig. 1.2 Flow of a gas through an opening (A) at highpressures (viscous flow)

l á sÐ1 á cmÐ2

Fig. 1.3 Conductance values relative to the area,C*visc, C*mol, and pumping speed S*viscand S*mol for an orifice A, depending on thepressure relationship p2/p1 for air at 20 °C.

Gas (20 °C) Molecular flow Laminar flowAir 1.00 1.00

Oxygen 0.947 0.91Neon 1.013 1.05

Helium 2.64 0.92Hydrogen 3.77 2.07

Carbon dioxide 0.808 1.26Water vapor 1.263 1.7

Table 1.1 Conversion factors (see text)

D00 E 19.06.2001 21:35 Uhr Seite 14



Nomographic determination of conduc-tance valuesThe conductance values for piping andopenings through which air and othergases pass can be determined with nomo-graphic methods. It is possible not only todetermine the conductance value forpiping at specified values for diameter,length and pressure, but also the size ofthe pipe diameter required when a pum-ping set is to achieve a certain effectivepumping speed at a given pressure andgiven length of the line. It is also possibleto establish the maximum permissible pipelength where the other parameters areknown. The values obtained naturally donot apply to turbulent flows. In doubtfulsituations, the Reynolds number Re (seeSection 1.5.) should be estimated usingthe relationship which is approximatedbelow

(1.31)

Here qpV = S · p is the flow output in mbarl/s, d the diameter of the pipe in cm. A compilation of nomograms which haveproved to be useful in practice will befound in Chapter 9.

1.5.4 Conductance values for other elements

Where the line contains elbows or othercurves (such as in right-angle valves),these can be taken into account by assu-ming a greater effective length leff of theline. This can be estimated as follows:

(1.32)

Wherelaxial : axial length of the line (in cm)leff : Effective length of the line (in cm)d : Inside diameter of the line (in cm)θ : Angle of the elbow

(degrees of angle)

l l deff axial= + ·°·133

180. θ

Re = ·15 qdpV

The technical data in the Leybold catalogstates the conductance values for vaporbarriers, cold traps, adsorption traps andvalves for the molecular flow range. At hig-her pressures, e.g. in the Knudsen andlaminar flow ranges, valves will have aboutthe same conductance values as pipes ofcorresponding nominal diameters andaxial lengths. In regard to right-angle val-ves the conductance calculation for anelbow must be applied.

In the case of dust filters which are used toprotect gas ballast pumps and rootspumps, the percentage restriction valuefor the various pressure levels are listed inthe catalog. Other components, namely thecondensate separators and condensers,are designed so that they will not reducepumping speed to any appreciable extent.

The following may be used as a rule ofthumb for dimensioning vacuum lines:The lines should be as short and as wideas possible. They must exhibit at least thesame cross-section as the intake port atthe pump. If particular circumstances pre-vent shortening the suction line, then it isadvisable, whenever this is justifiable fromthe engineering and economic points ofview, to include a roots pump in the suc-tion line. This then acts as a gas entrain-ment pump which reduces line impedance.

D00Axial length

D00 E 19.06.2001 21:35 Uhr Seite 15

Vacuum GenerationFundamentals of Vacuum Technology


2. Vacuumgeneration

2.1. Vacuum pumps: Asurvey

Vacuum pumps are used to reduce the gaspressure in a certain volume and thus thegas density (see equation 1.5). Conse-quently consider the gas particles need tobe removed from the volume. Basicallydifferentiation is made between twoclasses of vacuum pumps:a) Vacuum pumps where – via one or

several compression stages – the gasparticles are removed from the volumewhich is to be pumped and ejected into

the atmosphere (compression pumps).The gas particles are pumped by meansof displacement or pulse transfer.

b) Vacuum pumps where the gas particleswhich are to be removed condense onor are bonded by other means (e.g.chemically) to a solid surface, whichoften is part of the boundary formingvolume itself.

A classification which is more in line withthe state-of-the-art and practical applica-tions makes a difference between thefollowing types of pumps, of which thefirst three classes belong to the compres-sion pumps and where the two remainingclasses belong to the condensation andgetter pumps:1. Pumps which operate with periodically

increasing and decreasing pump cham-ber volumes (rotary vane and rotaryplunger pumps; also trochoid pumps)

2. Pumps which transport quantities ofgas from the low pressure side to thehigh pressure side without changingthe volume of the pumping chamber(Roots pumps, turbomolecular pumps)

3. Pumps where the pumping effect isbased mainly on the diffusion of gasesinto a gas-free high speed vapor jet(vapor pumps)

4. Pumps which pump vapors by means ofcondensation (condensers) and pumpswhich pump permanent gases by wayof condensation at very low tempera-tures (cryopumps)

5. Pumps which bond or incorporate ga-ses by adsorption or absorption tosurfaces which are substantially free ofgases (sorption pumps).

A survey on these classes of vacuumpumps is given in the diagram of Table 2.1.

Adsorptionpump

Fluid entrainmentvacuum pump

Ejectorvacuum pump

Liquid jetvacuum pump

Gas jetvacuum pump

Vapor jetvacuum pump

Diffusion pump

Self-purifyingdiffusion pump

Fractionatingdiffusion pump

Diffusion ejectorpump

Dragvacuum pump

Gaseous ringvacuum pump

Turbinevacuum pump

Axial flow vacuum pump

Radial flow vacuum pump

Molecular dragvacuum pump

Turbomolecular pump

Rotaryvacuum pump

Liquid sealedvacuum pump

Liquid ring vacuum pump

Rotary vanevacuum pump

Multiple vanevacuum pump

Rotary pistonvacuum pump

Rotary plungervacuum pump

Dry compressingvacuum pump

Roots vacuum pump

Clawvacuum pump

Scroll pump

Reciprocatingpositive displacement

vacuum pump

Diaphragm vacuum pump

Piston vacuum pump

Ion transfervacuum pump

Getter pump

Bulk getter pump

Sublimation pump

Getter ion pump

Evaporation ion pump

Sputter-ion pump

Cryopump

Condenser

Vacuum pump(Operating principle)

Kinetic vacuum pump

Gas transfer vacuum pump

Positive displacementvacuum pump

Entrapmentvacuum pump

Table 2.1 Classification of vacuum pumps

D00 E 19.06.2001 21:35 Uhr Seite 16

Vacuum Generation Fundamentals of Vacuum Technology


2.1.1 Oscillation displacementvacuum pumps

2.1.1.1 Diaphragm pumps

Recently, diaphragm pumps havebecoming ever more important, mainly forenvironmental reasons. They are alterna-tives to water jet vacuum pumps, sincediaphragm pumps do not produce anywaste water. Overall, a diaphragm vacuumpump can save up to 90 % of the operatingcosts compared to a water jet pump.Compared to rotary vane pumps, thepumping chamber of diaphragm pumpsare entirely free of oil. By design, no oilimmersed shaft seals are required.Diaphragm vacuum pumps are single ormulti-stage dry compressing vacuumpumps (diaphragm pumps having up tofour stages are being manufactured). Herethe circumference of a diaphragm istensioned between a pump head and thecasing wall (Fig. 2.1). It is moved in anoscillating way by means of a connectingrod and an eccentric. The pumping orcompression chamber, the volume ofwhich increases and decreases periodi-cally, effects the pumping action. Thevalves are arranged in such a way thatduring the phase where the volume of thepumping chamber increases it is open tothe intake line. During compression, thepumping chamber is linked to the exhaustline. The diaphragm provides a hermeticseal between the gear chamber and the

pumping chamber so that it remains freeof oil and lubricants (dry compressingvacuum pump). Diaphragm and valves arethe only components in contact with themedium which is to be pumped. Whencoating the diaphragm with PTFE (Teflon)and when manufacturing the inlet andexhaust valves of a highly fluorinatedelastomer as in the case of the DIVAC fromLEYBOLD, it is then possible to pumpaggressive vapors and gases. It is thuswell suited for vacuum applications in thechemistry lab.

Due to the limited elastic deformability ofthe diaphragm only a comparatively lowpumping speed is obtained. In the case ofthis pumping principle a volume remainsat the upper dead center – the so called“dead space” – from where the gases cannot be moved to the exhaust line. Thequantity of gas which remains at the ex-haust pressure expands into the expandingpumping chamber during the subsequentsuction stroke thereby filling it, so that asthe intake pressure reduces the quantity ofinflowing new gas reduces more andmore. Thus volumetric efficiency worsenscontinuously for this reason. Diaphragmvacuum pumps are not capable ofattaining a higher compression ratio thanthe ratio between “dead space” andmaximum volume of the pumping cham-ber. In the case of single-stage diaphragmvacuum pumps the attainable ultimatepressure amounts to approximately 80

mbar. Two-stage pumps such as theDIVAC from LEYBOLD can attain about 10 mbar (see Fig. 2.2), three-stage pumpscan attain about 2 mbar and four-stagediaphragm pumps can reach about 5·10-1 mbar.

Diaphragm pumps offering such a lowultimate pressure are suited as backingpumps for turbomolecular pumps with fullyintegrated Scroll stages (compound or widerange turbomolecular pumps, such as theTURBOVAC 55 from LEYBOLD). In this waya pump system is obtained which is ab-solutely free of oil, this being of great im-portance to measurement arrangementsinvolving mass spectrometer systems andleak detectors. In contrast to rotary vanepumps this combination of pumps for leakdetectors offers the advantage thatnaturally no helium is dissolved in thediaphragm pump thereby entirely avoidinga possible build up of a helium background.

2.1.2 Liquid sealed rotarydisplacement pumps

2.1.2.1 Liquid ring pumps

Due to the pumping principle and thesimple design, liquid ring vacuum pumpsare particularly suited to pumping gasesand vapors which may also contain smallamounts of liquid. Air, saturated with watervapors or other gases containing conden-

D00(1) Casing lid(2) Valves(3) Lid(4) Diaphragm disk

(5) Diaphragm(6) Diaphragm support disk(7) Connecting rod(8) Eccentric disk

Fig. 2.1 Schematic on the design of a diaphragm pump stage (Vacuubrand) Fig. 2.2 Principle of operation for a two-stage diaphragm pump (Vacuubrand)

Opening and closing of the valves, path and pumping mechanismduring four subsequent phases of a turn of the connecting rod (a-d)

a)

b)

c)

d)

1st stage 2nd stage

IN EX

D00 E 19.06.2001 21:35 Uhr Seite 17



sable constituents, may be pumpedwithout problems. By design, liquid ringpumps are insensitive to any contamina-tion which may be present in the gas flow.The attainable intake pressures are in theregion between atmospheric pressure andthe vapor pressure of the operating liquidused. For water at 15 °C it is possible toattain an operating pressure of 33 mbar. Atypical application of water ring vacuumpumps is venting of steam turbines inpower plants. Liquid ring vacuum pumps(Fig. 2.3) are rotary displacement pumpswhich require an operating liquid whichrotates during operation to pump the gas.

The blade wheel is arranged eccentricallyin a cylindrical casing. When not in opera-tion, approximately half of the pump isfilled with the operating fluid. In the axialdirection the cells formed by the bladewheel are limited and sealed off by“control discs”. These control discs areequipped with suction and ejection slotswhich lead to the corresponding ports ofthe pump. After having switched on such apump the blade wheel runs eccentricallywithin the casing; thus a concentricallyrotating liquid ring is created which at thenarrowest point fully fills the spacebetween the casing and the blade wheeland which retracts from the chambers asthe rotation continues. The gas is suckedin as the chambers empty and compres-sion is obtained by subsequent filling. Thelimits for the intake or discharge processare set by the geometry of the openings inthe control discs.

In addition to the task of compression, theoperating fluid fulfills three furtherimportant tasks:

1. Removal of the heat produced by thecompression process.

2. Uptake of liquids and vapors(condensate).

3. Providing the seal between the bladewheel and the casing.

2.1.2.2 Oil sealed rotary displacementpumps

A displacement vacuum pump is generallya vacuum pump in which the gas which isto be pumped is sucked in with the aid ofpistons, rotors, vanes and valves orsimilar, possibly compressed and thendischarged. The pumping process iseffected by the rotary motion of the pistoninside the pump. Differentiation should bemade between oiled and dry compressingdisplacement pumps. By the use of sealingoil it is possible to attain in a single-stagehigh compression ratios of up to about105. Without oil, “inner leakiness” isconsiderably greater and the attainablecompression ratio is correspondingly less,about 10.As shown in the classification Table 2.1,the oil sealed displacement pumps includerotary vane and rotary plunger pumps ofsingle and two-stage design as well assingle-stage trochoid pumps which todayare only of historic interest. Such pumpsare all equipped with a gas ballast facilitywhich was described in detail (for detailssee 2.1.2.2.4) for the first time by Gaede(1935). Within specified engineeringlimits, the gas ballast facility permitspumping of vapors (water vapor inparticular) without condensation of thevapors in the pump.

2.1.2.2.1 Rotary vane pumps (TRIVAC A, TRIVAC B, TRIVAC E, SOGEVAC)

Rotary vane pumps (see also Figs. 2.5 and2.6) consist of a cylindrical housing(pump-ing ring) (1) in which an eccentri-cally suspended and slotted rotor (2) turnsin the direction of the arrow. The rotor hasvanes (16) which are forced outwardsusually by centrifugal force but also bysprings so that the vanes slide inside thehousing. Gas entering through the intake(4) is pushed along by the vanes and isfinally ejected from the pump by the oilsealed exhaust valve (12).

The older range of TRIVAC A pumps (Fig.2.5) from LEYBOLD has three radial vanesoffset by 120°. The TRIVAC B range (Fig.2.6) has only two vanes offset by 180°. In

both cases the vanes are forced outwardsby the centrifugal forces without the use ofsprings. At low ambient temperatures thispossibly requires the use of a thinner oil.The A-Series is lubricated through thearising pressure difference whereas the B-Series pumps have a geared oil pump forpressure lubrication. The TRIVAC B-Seriesis equipped with a particularly reliable anti-suckback valve; a horizontal or vertical

Fig. 2.3 Liquid ring vacuum pump, schematic (Siemens)

1 Rotor2 Rotor shaft3 Casing

4 Intake channel5 Liquid ring6 Flex. discharge channel

Fig. 2.4 Arrangement of the sealing passage in rotaryvane pumps also known as “duo seal”

Constant, minimum clearance a for the entire sealing passage b

Fig. 2.5 Cross section of a single-stage rotary vanepump (TRIVAC A)

1 Pump housing2 Rotor3 Oil-level sight glass4 Suction duct5 Anti-suckback valve6 Dirt trap7 Intake port8 Lid of gas ballast

valve

9 Exhaust port10 Air inlet silencer11 Oil filter12 Exhaust valve13 Exhaust duct14 Gas ballast duct15 Oil injection16 Vane

D00 E 19.06.2001 21:35 Uhr Seite 18



arrangement for the intake and exhaustports. The oil level sight glass and the gasballast actuator are all on the same side of

the oil box (user friendly design). Incombination with the TRIVAC BCS systemit may be equipped with a very com-prehensive range of accessories, designedchiefly for semiconductor applications.The oil reservoir of the rotary vane pumpand also that of the other oil sealeddisplacement pumps serves the purposeof lubrication and sealing, and also to filldead spaces and slots. It removes the heatof gas compression, i.e. for coolingpurposes. The oil provides a seal betweenrotor and pump ring. These parts are“almost” in contact along a straight line(cylinder jacket line). In order to increasethe oil sealed surface area a so-calledsealing passage is integrated into thepumping ring (see Fig. 2.4). This providesa better seal and allows a higher com-pression ratio or a lower ultimate pres-sure. LEYBOLD manufactures threedifferent ranges of rotary vane pumpswhich are specially adapted to differentapplications such as high intake pressure,low ultimate pressure or applications inthe semiconductor industry. A summary ofthe more important characteristics ofthese ranges is given in Table 2.2. TheTRIVAC rotary vane pumps are producedas single-stage (TRIVAC S) and two-stage(TRIVAC D) pumps (see Fig. 2.7). With the

two-stage oil sealed pumps it is possibleto attain lower operating and ultimatepressures ompared to the correspondingsingle-stage pumps. The reason for this isthat in the case of single-stage pumps, oilis unavoidably in contact with theatmosphere outside, from where gas istaken up which partially escapes to thevacuum side thereby restricting theattainable ultimate pressure. In the oilsealed two-stage displacement pumpsmanufactured by LEYBOLD, oil which hasalready been degassed is supplied to thestage on the side of the vacuum (stage 1 inFig. 2.7): the ultimate pressure lies almost

D00

TRIVAC A TRIVAC B TRIVAC BCS TRIVAC E SOGEVAC

Vanes per stage 3 2 2 2 3 (tangential)

1 – 1.5 1.6 – – 16 – 25Pumping speed 2 – 4 4 – 8 16 – 25 2.5 40 – 100

[m3/h] 8 – 16 16 – 25 40 – 65 – 180 – 28030 – 60 40 – 65 – – 585 – 1200

Sealing passage yes yes yes yes no

Ultimate pressure, < 2 · 10–2 < 2 · 10–2 < 2 · 10–2 – < 5 · 10–1

single-stage [mbar]

Ultimate pressure < 2.5 · 10–4 < 1 · 10–4 < 1 · 10–4 < 1 · 10–4 –two-stage [mbar]

Oil supply Pressure difference Gear pump Gear pump Eccentric pump Pressure difference

Slots Comparable for all types: about 0.01 to 0.05 mm

Bearing/lubrication Axial face / oil Axial face / oil Axial face / oil Ball / grease Ball / oil

Special – Hydropneumatic Coated parts in Many Cost-effectivecharacteristics anti-suckback valve contact with medium accessories

Media No ammonia Clean to Aggressive and Clean to Cleanlight particles corrosive light particles

Main areas of Multi- Multi- Semiconductor Multi- Packagingapplication purpose purpose industry purpose industry

Fig. 2.6 Cross section of a single-stage rotary vanepump (TRIVAC B)

1 Intake port2 Dirt trap3 Anti-suckback

valve4 Intake duct5 Vane6 Pumping cham-

ber7 Rotor

8 Orifice, connec-tion for inert gasballast

9 Exhaust duct10 Exhaust valve11 Demister12 Spring13 Orifice; connec-

tion for oil filter

Leafspringof thevalve

I High vacuum stageII Second forevacuum stage

Fig. 2.7 Cross section of a two-stage rotary vanepump, schematic

Valve stop

Table 2.2 Rotary vacuum pump ranges

D00 E 19.06.2001 21:35 Uhr Seite 19



in the high vacuum range, the lowestoperating pressures lie in the rangebetween medium vacuum / high vacuum.Note: operating the so called high vacuumstage (stage 1) with only very little oil orno oil at all will – in spite of the very lowultimate pressure – in practice lead toconsiderable difficulties and will signifi-cantly impair operation of the pump.

2.1.2.2.2 Rotary plunger pumps (E-Pumps)

Shown in Fig. 2.9 is a sectional view of arotary plunger pump of the single blocktype. Here a piston (2) which is movedalong by an eccentric (3) turning in the

direction of the arrow moves along thechamber wall. The gas which is to bepumped flows into the pump through theintake port (11), passes through the intakechannel of the slide valve (12) into thepumping chamber (14). The slide valveforms a unit with the piston and slides toand fro between the rotatable valve guidein the casing (hinge bar 13). The gasdrawn into the pump finally enters thecompression chamber (4). While turning,the piston compresses this quantity of gasuntil it is ejected through the oil sealedvalve (5). As in the case of rotary vanepumps, the oil reservoir is used forlubrication, sealing, filling of dead spacesand cooling. Since the pumping chamberis divided by the piston into two spaces,

each turn completes an operating cycle(see Fig. 2.10). Rotary plunger pumps aremanufactured as single and two-stagepumps. In many vacuum processescombining a Roots pump with a single-stage rotary plunger pump may offer moreadvantages than a two-stage rotaryplunger pump alone. If such a combinationor a two-stage pump is inadequate, theuse of a Roots pump in connection with atwo-stage pump is recommended. Thisdoes not apply to combinations involvingrotary vane pumps and Roots pumps.

Motor powerThe motors supplied with the rotary vaneand rotary plunger pumps deliver enoughpower at ambient temperatures of 12 °Cand when using our special oils to coverthe maximum power requirement (at about400 mbar). Within the actual operatingrange of the pump, the drive system of thewarmed up pump needs to supply onlyabout one third of the installed motorpower (see Fig. 2.11).

Fig. 2.8a Cross section of a two-stage rotary vanepump (TRIVAC E)

Fig. 2.8b SOGEVAC pump SV 300 with three tangen-tial vanes

Fig. 2.9 Cross section of a single-stage rotary plun-ger pump (monoblock design)

1 Casing2 Cylindrical piston3 Eccentric4 Compression

chamber5 Oil sealed pressure

valve6 Oil-level sight glass7 Gas ballast channel

8 Exhaust pot9 Gas ballast valve10 Dirt trap11 Intake port12 Slide valve13 Hinge bar14 Pumping chamber

(air is flowing in)

Fig. 2.10 Operating cycle of a rotary plunger pump (for positions 1 to 9 of the plunger)

1 Upper dead point2 Slot in suction channel of

slide valve is freed – begin-ning of suction period

3 Lower dead point – slot insuction channel is quite free,and pumped-in gas (arrow)enters freely into the pum-ping chamber (shown sha-ded)

4 Slot in suction channel is clo-sed again by swivelling hingebar – end of suction period

5 Upper dead point – maximumspace between rotatingpiston and stator

6 Shortly before beginning ofcompression period, the frontsurface of the rotating plun-ger frees gas ballast opening

– commencement of gas bal-last inlet

7 Gas ballast opening is quitefree

8 End of gas ballast inlet9 End of pumping period.

D00 E 19.06.2001 21:35 Uhr Seite 20



2.1.2.2.3 Trochoid pumps

Trochoid pumps belong to the class of socalled rotary piston pumps, which (seeoverview of Table 2.1) in turn belong to thegroup of rotary pumps. With rotary pistonpumps, the piston’s center of gravity runsalong a circular path about the rotationalaxis (hence rotary piston machines). Arotary piston pump can – in contrast to therotary plunger pump – be completelybalanced dynamically. This offers theadvantage that larger pumps can operatewithout vibration so that they can beinstalled without needing foundations.Moreover, such pumps may be operated athigher speed, compared to rotary plungerpumps (see below). The volume of thepumping chamber with respect to thevolume of the entire pump – the so calledspecific volume – is, in the case oftrochoid pumps, approximately twice ofthat of rotary plunger pumps. Largerrotary plunger pumps run at speeds of 500 rpm. The trochoid pump may run at1000 rpm and this applies also to largerdesigns. It is thus about four times smallercompared to a rotary plunger pump havingthe same pumping speed and runs withoutproducing any vibrations. Unfortunatelythe advantages in the area of engineeringare combined with great disadvantages in

the area of manufacturing, so that todayLEYBOLD does not produce trochoidpumps any more. Operation of such apump is shown in the sectional diagram ofFig. 2.12.

2.1.2.2.4 The gas ballast

The gas ballast facility as used in the rotaryvane, rotary plunger and trochoid pumpspermits not only pumping of permanentgases but also even larger quantities ofcondensable gases.

The gas ballast facility (see Fig. 2.13)prevents condensation of vapors in thepump chamber of the pump. Whenpumping vapors these may only becompressed up to their saturation vaporpressure at the temperature of the pump.If pumping water vapor, for example, at apump temperature of 70 °C, the vapor mayonly be compressed to 312 mbar(saturation vapor pressure of water at 70 °C (see Table XIII in Section 9)). Whencompressing further, the water vaporcondenses without increasing thepressure. No overpressure is created inthe pump and the exhaust valve is notopened. Instead the water vapor remainsas water in the pump and emulsifies with

the pump’s oil. This very rapidly impairsthe lubricating properties of the oil and thepump may even seize when it has taken uptoo much water. The gas ballast facilitydeveloped in 1935 by Wolfgang Gaedeinhibits the occurrence of condensation ofthe vapor in the pump as follows. Beforethe actual compression process begins(see Fig. 2.13), a precisely defined quantityof air (“the gas ballast”) is admitted intothe pumping chamber of the pump. Thequantity is such that the compression ratioof the pump is reduced to 10:1 max. Nowvapors which have been taken in by thepump may be compressed together withthe gas ballast, before reaching theircondensation point and ejected from thepump. The partial pressure of the vaporswhich are taken in may however notexceed a certain value. It must be so lowthat in the case of a compression by afactor of 10, the vapors can not condenseat the operating temperature of the pump.When pumping water vapor this criticalvalue is termed the “water vaportolerance”.

Shown schematically in Fig. 2.14 is thepumping process with and without gasballast as it takes place in a rotary vanepump when pumping condensable vapors.

D00

Fig. 2.11 Motor power of a rotary plunger pump (pum-ping speed 60 m3/h) as a function of intakepressure and operating temperature. The cur-ves for gas ballast pumps of other sizes aresimilar.

1 Operating temp.curve 1 32°C,




5 Theoretic curve foradiabatic compres-sion

6 Theoretic curve forisotherm compres-sion

Pressure [mbar]

Fig. 2.12 Cross section of a trochoid pump

1 Toothed wheel fixed to the driving shaft2 Toothed wheel fixed to the piston3 Elliptic piston4 Inner surface of the pump5 Driving shaft6 Eccentric

Pow

er o

f the

driv

e m

otor

[Wat

t]

Suction

Com

pres

sion

Gas

bal

last

inle

t

Gas ballastinlet

Positionof theleadingvane

Discharge

Fig. 2.13 Working process within a rotary vane pumpwith gas ballast

1–2 Suction2–5 Compression3–4 Gas ballast inlet5–6 Discharge

D00 E 19.06.2001 21:35 Uhr Seite 21



Two requirements must be met whenpumping vapors:1) the pump must be at operating

temperature.2) the gas ballast valve must be open.

(With the gas ballast valve open thetemperature of the pump increases byabout 10 °C. Before pumping vapors thepump should be operated for half an hourwith the gas ballast valve open).

Simultaneous pumping of gases andvaporsWhen simultaneously pumping permanentgases and condensable vapors from avacuum system, the quantity of permanentgas will often suffice to prevent anycondensation of the vapors inside thepump. The quantity of vapor which may bepumped without condensation in thepump can be calculated as follows:

(2.1)

Where:pvapor = is the partial pressure of

vapor at the intake of thepump

pperm = is the total pressure of allpumped permanent gasesat the intake of the pump

pvapor, sat = is the saturation pressureof the pumped vapor,depending on temperatu-re (see Fig. 2.15)

psum = pexhaust + ∆pvalve +∆pexhaust filter

∆pvalve = is the pressure differenceacross the exhaust valvewhich amounts depen-ding on type of pump andoperating conditions to0.2 ... 0.4 bar

pp p

ppperm

vapour sat

sum

vapour

vapour + < ,

∆pexhaust filter = is the pressure differenceacross the exhaust filteramounting to 0 ... 0.5 bar

Example 1:With a rotary vane pump with an externaloil mist filter in series, a mixture of watervapor and air is being pumped. The follo-wing values are used for applying eq. (2.1):

pexhaust = 1 bar

∆pvalve + ∆pexhaust filter = 0.35 bar,temperature of the pump 70 °C

Hence:

psum = 1.35 bar; pvapor sat (H2O) = 312 mbar

(see Table XIII in chapter 9)

Using eq. (2.1) follows:

The pressure of the water vapor in theair/water vapor mixture must not exceed23 % of the total pressure of the mixture.

Example 2:Ethanoic acid is to be pumped with arotary plunger pump.

pexhaust = 1.1 bar (taking into consideration the flow resistance of the pipes)

∆pvalve = 0.25 bar

∆pexhaust filter = 0.15 bar (pressure loss in the oilmist trap)

Hence:

psum = 1.5 bar.

By controlled cooling the pump and oiltemperature is set at 100 °C. The saturati-on pressure of the acid therefore is – see Fig. 2.15 – pvapor, sat = 500 mbar.

From eq. (2.1) follows:

Returning to the question of pumpingwater vapor in a mixture with air, the ratio3 parts of permanent gases to 1 part ofwater vapor, as indicated in example 1, canbe for guidance only. In actual practice it is

pp p

vapour, acid

vapour, acid air

.

.+= < =0 5

1 513

pp p

vapour , H2O

vapour , H2O air.312

13500 23

+ < =

Fig. 2.14 Diagram of pumping process in a rotary vanepump without (left) and with (right) gas bal-last device when pumping condensable sub-stances

a) Without gas bal-last

1) Pump is connectedto the vessel, whichis already almostempty of air (70 mbar) – it mustthus transportmostly vapor particles

2) Pump chamber isseparated from thevessel – compres-sion begins

3) Content of pumpchamber is alreadyso far compressedthat the vapor con-denses to formdroplets – over-pressure is not yetreached

4) Residual air onlynow produces therequired overpres-sure and opens thedischarge valve,but the vapor hasalready condensedand the dropletsare precipitated inthe pump.

b) With gas ballast1) Pump is connected

to the vessel, whichis already almostempty of air (70 mbar) – it mustthus transportmostly vapor particles

2) Pump chamber isseparated from thevessel – now thegas ballast valve,through which thepump chamber isfilled with additio-nal air from outsi-de, opens – thisadditional air is cal-led gas ballast

3) Discharge valve ispressed open, andparticles of vaporand gas are pushedout – the overpres-sure required forthis to occur is rea-ched very earlybecause of the sup-plementary gasballast air, as at thebeginning the enti-re pumping pro-cess condensationcannot occur

4) The pump dischar-ges further air andvapor

Fig. 2.15 Saturation vapor pressures

Temperature [° C]

Satu

ratio

n va

por p

ress

ure

[Tor

r]

Satu

ratio

n va

por p

ress

ure

[mba

r]

D00 E 19.06.2001 21:35 Uhr Seite 22



recommended to run up a rotary pump ofthe types described hitherto always withthe gas ballast valve open, because it takessome time until the pump has reached itsfinal working temperature.

From eq. (2.1) follows for the permissiblepartial pressure pvapor of the pumpedvapor the relation

(2.2)

This relation shows that with pperm = 0 novapors can be pumped without condensa-tion in the pump, unless the gas-ballastconcept is applied. The corresponding for-mula is:

(2.3)

Where:B = is the volume of air at

1013 mbar which is admittedto the pump chamber per unittime, called in brief the “gasballast”

S = is the nominal speed of thepump (volume flow rate)

psum = is the pressure at the discharge outlet of the pump,assumed to be a maximum at1330 mbar

pvapor, sat = Saturation vapor pressure ofthe vapor at the pump’s exhaust port

pvapor, g.b. = is the partial pressure of anyvapor that might be presentin the gas used as gas ballast(e.g. water vapor contained inthe atmospheric air whenused as gas ballast)

pperm = is the total pressure of all per-manent gases at the inlet portof the pump

Eq. (2.3) shows that when using gasballast (B ≠ 0) vapors can also be pumpedwithout condensation if no gas is presentat the intake of the pump. The gas ballastmay also be a mixture of non-condensablegas and condensable vapor as long as thepartial pressure of this vapor (pvapor, g.b.) isless than the saturation pressure pvapor, satof the pumped vapor at the temperature ofthe pump.

p pBS

(p pp p

pp p p

vapour sumVa pour, sat vapour , g.b.

sum vapour , sat

vapour sat

sum vapour satperm

≤ ·−

−

+

+

−

)

pp

p ppvapour

vapour, sat

sum vapour, satperm≤

− ·

Water vapor toleranceAn important special case in the generalconsiderations made above relating to thetopic of vapor tolerance is that of pumpingwater vapor. According to PNEUROP watervapor tolerance is defined as follows:

“Water vapor tolerance is the highestpressure at which a vacuum pump, undernormal ambient temperatures andpressure conditions (20 °C, 1013 mbar),can continuously take in and transportpure water vapor. It is quoted in mbar”. Itis designated as pW,O.

Applying eq. (2.3) to this special casemeans:

pperm = 0 and pvapor, sat = ps (H2O), thus:

(2.4)

If for the gas ballast gas atmospheric air of50 % humidity is used, then pvapor, g.b. = 13 mbar; with B/S = 0.10 – a usual figure in practice – and psum(total exhaust pressure) = 1330 mbar, thewater vapor tolerance pW,0 as function ofthe temperature of the pump is represen-ted by the lowest curve in diagram Fig.2.16. The other curves correspond to thepumping of water vapor-air mixtures,hence pperm = pair ≠ O), indicated by thesymbol pL in millibar. In these cases a hig-her amount of water vapor partial pressu-re pw can be pumped as shown in the dia-gram. The figures for pW,0 given in thecatalogue therefore refer to the lower limitand are on the safe side.

According to equation 2.4 an increase in

p pBS

p H O pp p H OW sumO

s vapour, g.b.

sum s,

( )

( )=

−−

2

2

the gas ballast B would result in an increa-sed water vapor tolerance pW,0. In prac-tice, an increase in B, especially in the caseof single-stage gas ballast pumps isrestricted by the fact that the attainableultimate vacuum for a gas ballast pumpoperated with the gas ballast valve openbecomes worse as the gas ballast Bincreases. Similar considerations alsoapply to the general equation 2.3 for thevapor tolerance pvapor.

At the beginning of a pump down process,the gas ballast pump should always beoperated with the gas ballast valve open.In almost all cases a thin layer of water willbe present on the wall of a vessel, whichonly evaporates gradually. In order toattain low ultimate pressures the gasballast valve should only be closed afterthe vapor has been pumped out. LEYBOLDpumps generally offer a water vapor tole-rance of between 33 and 66 mbar. Two-stage pumps may offer other levels ofwater vapor tolerance corresponding tothe compression ratio between theirstages – provided they have pumpingchamber of different sizes.

Other gases as ballastGenerally atmospheric air is used as thegas ballast medium. In special cases,when pumping explosive or toxic gases,for example, other permanent gases likenoble gases or nitrogen, may be used (seeSection 8.3.1.3).

D00Fig. 2.16 Partial pressure pW of water vapor that can

be pumped with the gas ballast valve openwithout condensation in the pump, as a fun-ction of the pump temperature for variouspartial pressures pL of air. The lowest curvecorresponds to the water vapor tolerancepW, O of the pump

Temperature of the pump

p W[m

bar]

pL [mbar]

D00 E 19.06.2001 21:35 Uhr Seite 23



2.1.3 Dry compressing rotarydisplacement pumps

2.1.3.1 Roots pumps

The design principle of the Roots pumpswas already invented in 1848 by IsaiahDavies, but it was 20 years later before itwas implemented in practice by theAmericans Francis and Philander Roots.Initially such pumps were used as blowersfor combustion motors. Later, by invertingthe drive arrangement, the principle wasemployed in gas meters. Only since 1954has this principle been employed invacuum engineering. Roots pumps areused in pump combinations together withbacking pumps (rotary vane- and rotaryplunger pumps) and extend their operatingrange well into the medium vacuum range.With two stage Roots pumps this extendsinto the high vacuum range. The operatingprinciple of Roots pumps permits theassembly of units having very highpumping speeds (over 100,000 m3/h)which often are more economical tooperate than steam ejector pumps runningin the same operating range.A Roots vacuum pump (see Fig. 2.17) is arotary positive-displacement type of pumpwhere two symmetrically-shaped impel-lers rotate inside the pump casing pasteach other in close proximity. The tworotors have a cross section resemblingapproximately the shape of a figure 8 andare synchronized by a toothed gear. Theclearance between the rotors and thecasing wall as well as between the rotorsthemselves amounts only to a few tenthsof a millimeter. For this reason Rootspumps may be operated at high speeds

without mechanical wear. In contrast torotary vane and rotary plunger pumps,Roots pumps are not oil sealed, so that theinternal leakage of dry compressingpumps by design results in the fact thatcompression ratios only in the range 10 –100 can be attained. The internal leakageof Roots pumps, and also other drycompressing pumps for that matter, ismainly based on the fact that owing to theoperating principle certain surface areas ofthe pump chamber are assigned to theintake side and the compression side ofthe pump in alternating fashion. During thecompression phase these surface areas(rotors and casing) are loaded with gas(boundary layer); during the suction phasethis gas is released. The thickness of thetraveling gas layer depends on theclearance between the two rotors andbetween the rotors and the casing wall.Due to the relatively complex thermalconditions within the Roots pump it is notpossible to base one’s consideration onthe cold state. The smallest clearances andthus the lowest back flows are attained atoperating pressures in the region of 1 mbar. Subsequently it is possible toattain in this region the highest com-pression ratios, but this pressure range isalso most critical in view of contactsbetween the rotors and the casing.

Characteristic quantities of roots pumpsThe quantity of gas Qeff effectively pumpedby a Roots pump is calculated from thetheoretically pumped quantity of gas Qthand the internal leakage QiR (as thequantity of gas which is lost) as:

Qeff = Qth – QiR (2.5)

The following applies to the theoreticallypumped quantity of gas:

Qth = pa · Sth (2.6)

where pa is the intake pressure and Sth isthe theoretical pumping speed. This in turnis the product of the pumping volume VSand the speed n:

Sth = n · VS (2.7)

Similarly the internal leakage QiR iscalculated as:

QiR = n · ViR (2.8)

where pV is the forevacuum pressure(pressure on the forevacuum side) and SiRis a (notional) “reflow” pumping speedwith

SiR = n · ViR (2.9)

i.e. the product of speed n and internal lea-kage volume ViR.

Volumetric efficiency of a Roots pumps isgiven by

(2.10)

By using equations 2.5, 2.6, 2.7 and 2.8one obtains

(2.11)

When designating the compression pv/paas k one obtains

(2.11a)

Maximum compression is attained at zerothroughput (see PNEUROP and DIN 28 426, Part 2). It is designated as k0:

(2.12)

k0 is a characteristic quantity for the Rootspump which usually is stated as a functionof the forevacuum pressure pV (see Fig.2.18). k0 also depends (slightly) on thetype of gas.For the efficiency of the Roots pump, the

generally valid equation applies:

(2.13)

Normally a Roots pump will be operated inconnection with a downstream roughvacuum pump having a nominal pumpingspeed SV. The continuity equation gives:

SV · pV = Seff · pa = η · Sth · pa (2.14)

η = −1 kko

kSS

th

iR0 0= =( )η

η = −1 kSS

iR

th

η = − ·1pp

SS

V

a

iR

th

η =QQ

eff

th

Fig. 2.17 Schematic cross section of a Roots pump

12 2

35

4

1 Intake flange2 Rotors3 Chamber

4 Exhaust flange5 Casing

Fig. 2.18 Maximum compression k0 of the Roots pumpRUVAC WA 2001 as a function of fore vacu-um pressure pV

D00 E 19.06.2001 21:36 Uhr Seite 24



From this

(2.15)

The ratio Sth/SV (theoretical pumpingspeed of the Roots pump / pumping speedof the backing pump) is termed thegradation kth. From (2.15) one obtains

k = η · kth (2.16)

Equation (2.16) implies that the compres-sion k attainable with a Roots pump mustalways be less than the grading kthbetween Roots pump and backing pumpsince volumetric efficiency is always < 1.When combining equations (2.13) and(2.16) one obtains for the efficiency thewell known expression

(2.17)

The characteristic quantities to be found inequation 2.17 are only for the combinationof the Roots pump and the backing pump,namely maximum compression k0 of theRoots pump and gradation kth betweenRoots pump and backing pump.With the aid of the above equations thepumping speed curve of a given combi-nation of Roots pump and backing pumpmay be calculated. For this the followingmust be known:a) the theoretical pumping speed of the

Roots pump: Sth

η = +k

k ko th

0

kpp

SS

V

a

th

V= = ·η

b) the max. compression as a function ofthe fore vacuum pressure: k0 (pV)

c) the pumping speed characteristic of thebacking pump SV (pV)

The way in which the calculation is carriedout can be seen in Table 2.3 giving the datafor the combination of a Roots pumpRUVAC WA 2001 / E 250 (single-stage ro-tary plunger pump, operated without gasballast). In this the following is taken forSth:

Sth = 2050 – 2.5 % = 2000 m3/h

The method outlined above may also beapplied to arrangements which consist of arotary pump as the backing pump andseveral Roots pumps connected in series,for example. Initially one determines – inline with an iteration method – thepumping characteristic of the backingpump plus the first Roots pump and thenconsiders this combination as the backingpump for the second Roots pump and soon. Of course it is required that thetheoretical pumping speed of all pumps ofthe arrangement be known and that thecompression at zero throughput k0 as afunction of the backing pressure is alsoknown. As already stated, it depends onthe vacuum process which grading will bemost suitable. It may be an advantagewhen backing pump and Roots pump bothhave the same pumping speed in therough vacuum range.

Power requirement of a roots pumpCompression in a Roots pump is perfor-med by way of external compression andis termed as isochoric compression.Experience shows that the following equa-tion holds approximately:

Ncompression = Sth (pv – pa) (2.18)

In order to determine the total power (so-called shaft output) of the pump, mecha-nical power losses NV (for example in thebearing seals) must be considered:

Ntot = Ncompression + ∑ NV (2.19)

The power losses summarized in NV are –as shown by experience – approximatelyproportional to Sth, i.e.:

∑ Nv = const · Sth (2.20)

Depending on the type of pump and itsdesign the value of the constant rangesbetween 0.5 and 2 Wh / m3 .The total power is thus:

Ntot = Sth (pv – pa + const.)

The corresponding numerical value equa-tion which is useful for calculations is:

Ntot = Sth (pv – pa + const.) · 3 · 10-2 Watt(2.21)

with pv, pa in mbar, Sth in m3 / h andthe constant “const.” being between 18 and 72 mbar.

D00

Forevacuum Pumping speed kth = Sth / Sv k0 (pv) of η = k0 / k0+kth Intake pressurepressure Sv of the = 2001/Sv the RUVAC (Volumetric. Seff = η Sth pa = pv · Sv / Seff

Pv E 250 WA 2001 efficiency) (equation 2.14)

100 250 8.0 12.5 0.61 1.220 21

40 250 8.0 18 0.69 1.380 7.2

10 250 8.0 33 0.8 1.600 1.6

5 250 8.0 42 0.84 1.680 0.75

1 250 8.0 41 0.84 1.680 0.15

5 · 10–1 220 9.1 35 0.79 1.580 7 · 10–2

1 · 10–1 120 16.6 23 0.6 1.200 1 · 10–2

4 · 10–2 30 67 18 0.21 420 3 · 10–3

↓ ↓

Pumping speed characteristicfor the combination

WA 2001 / E250

Table 2.3

The values taken from the two right-hand columns give point by point the pumping speed curve forthe combination WA 2001/E250 (see Fig. 2.19, topmost curve)

D00 E 19.06.2001 21:36 Uhr Seite 25



Load rating of a roots pumpThe amount of power drawn by the pumpdetermines its temperature. If the tempe-rature increases over a certain level,determined by the maximum permissiblepressure difference pV – pa, the dangerexists that the rotors may seize in thecasing due to their thermal expansion. Themaximum permissible pressure difference∆pmax is influenced by the following fac-tors: forevacuum or compression pressurepV, pumping speed of the backing pumpSV, speed of the Roots pump n, gradationkth and the adiabatic exponent κ of thepumped gas. ∆pmax increases when pVand SV increase and decreases when n andkth increase. Thus the maximum differencebetween forevacuum pressure and intakepressure, pV – pa must – during conti-nuous operation – not exceed a certainvalue depending on the type of pump.Such values are in the range between 130and 50 mbar. However, the maximumpermissible pressure difference for conti-nuous operation may be exceed for briefperiods. In the case of special designs,which use gas cooling, for example, highpressure differences are also permissibleduring continuous operation.

Types of motors used with roots pumpsStandard flange-mounted motors are usedas the drive. The shaft feedthroughs aresealed by two oil sealed radial shaft sealsrunning on a wear resistant bushing inorder to protect the drive shaft. Flangemotors of any protection class, voltage orfrequency may be used.

Integral leak tightness of this version is < 10-4 mbar·l·s-1.

In the case of better leak tightnessrequirements of < 10-5 mbar·l·s-1 theRoots pump is equipped with a cannedmotor. The rotor is seated in the vacuumon the drive shaft of the pump and isseparated from the stator by a vacuum-tight non-magnetic tube. The stator coilsare cooled by a fan having its own drivemotor. Thus shaft seals which might besubject to wear are no longer required. Theuse of Roots pumps equipped with cannedmotors is especially recommended whenpumping high purity-, toxic- or radioactivegases and vapors.

Maintaining the allowed pressuredifferenceIn the case of standard Roots pumps,measures must be introduced to ensurethat the maximum permissible pressuredifference between intake and exhaust portdue to design constraints is not exceeded.This is done either by a pressure switch,which cuts the Roots pump in and outdepending on the intake pressure, or byusing a pressure difference or overflowvalve in the bypass of the Roots pumps(Fig. 2.20 and 2.21). The use of anoverflow valve in the bypass of the Rootspump is the better and more reliablesolution. The weight and spring loadedvalve is set to the maximum permissible

Fig. 2.19 Pumping speed curves for different pump combinations with the corresponding backing pumps

Intake pressure pa →

Pum

ping

spe

ed S

WAU 2001

SOGEVAC SV 1200

Fig. 2.20 Cross section of a Roots pumps with bypass line

Fig. 2.21 Vacuum diagram – Roots pump with integra-ted bypass line and backing pump

D00 E 19.06.2001 21:36 Uhr Seite 26



pressure difference of the particular pump.This ensures that the Roots pump is notoverloaded and that it may be operated inany pressure range. In practice this meansthat the Roots pump can be switched on,together with the backing pump, atatmospheric pressure. In the process anypressure increases will not adversely affectcombined operation, i.e. the Roots pumpis not switched off in such circumstances.

Pre-admission cooling (Fig. 2.22)In the case of Roots pumps with pre-admission cooling, the compression pro-cess basically is the same as that of a nor-mal Roots pump. Since greater pressuredifferences are allowed more installedpower is needed, which at the given speedand the pressure difference between inletand discharge port is directly proportionaland is composed of the theoretical workdone on compression and various powerlosses. The compression process endsnormally after opening of the pumpingchamber in the direction of the dischargeport. At this moment warmed gas at higherpressure flows into the pumping chamberand compresses the transported volume ofgas. This compression process is perfor-med in advance in the case of pre-admis-sion cooling. Before the rotor opens thepumping chamber in the direction of thedischarge port, compressed and cooledgas flows into the pumping chamber viathe pre-admission channel. Finally therotors eject the pumped medium via thedischarge port. The cooled gas, which in

the case of single-stage compression istaken from the atmosphere and admittedfrom the pre-admission cooler, and whichin the case of multi-stage pump systems istaken from downstream gas coolers,performs a pre-compression and removesby “inner cooling” the heat of compressionat the point of time it occurs.

2.1.3.2 Claw pumps

Like Roots pumps, claw pumps belong tothe group of dry compressing rotarypiston vacuum pumps (or rotary vacuumpumps). These pumps may have severalstages; their rotors have the shape ofclaws.

The design principle of a claw pump isexplained by first using an example of afour-stage design. The cross section insidethe pump’s casing has the shape of twopartly overlapping cylinders (Fig. 2.23).Within these cylinders there are two freelyrotating rotors in each pump stage: (1)with their claws and the matching recessesrotating in opposing directions about theirvertical axes. The rotors are synchronizedby a gear just like a Roots pump. In orderto attain an optimum seal, the clearancebetween the rotor at the center of thecasing and the amount of clearance withrespect to the inside casing wall is verysmall; both are in the order of magnitudeof a few 0.01 mm. The rotors periodicallyopen and close the intake and dischargeslots (5) and (4). At the beginning of the

work cycle in position a, the right rotor justopens the intake slot (5). Gas now flowsinto the continually increasing intake space(3) in position b until the right rotor sealsoff the intake slot (5) in position c. Afterboth claws have passed through the centerposition, the gas which has entered is thencompressed in the compression chamber(2) (position a) so long until the left rotorreleases the discharge slot (4) (position b)thereby discharging the gas. Immediatelyafter the compression process has started(position a) the intake slot (5) is openedsimultaneously and gas again flows intothe forming intake space (3) (position b).Influx and discharge of the gas isperformed during two half periods. Eachrotor turns twice during a full work cycle.Located between the pumping stages areintermediate discs with flow channelswhich run from the discharge side of theupper stage to the intake side of the nextstage, so that all inlet or exhaust sides arearranged vertically above each other (Fig.2.24). Whereas in a Roots pump the in-coming gas is pumped through the pumpat a constant volume and compression isonly performed in the forevacuum line(see Section 2.1.3.1), the claw pumpcompresses the gas already within thepumping chamber until the rotor releasesthe discharge slot. Shown in Fig. 2.25 arethe average pressure conditions in theindividual pumping stages of a DRYVAC atan intake pressure of 1 mbar. In order tomeet widely differing requirementsLEYBOLD manufactures two different

D00

Fig. 2.22 Diagram of a Roots pump with pre-admis-sion cooling

1 Intake port2 Discharge port3 Gas cooler4 Flow of cold gas

1

2

3

44

Fig. 2.23 Principle of operation

a

b

c

1 Rotors2 Compression

chamber3 Intake space

4 Exhaust slot5 Intake slot6 Intermediate stage purge

gas

P

Z

PZ

PZ

P

Fig. 2.24 Arrangement of the pumps and guiding ofthe gas flow. P = Pump stage Z = Interme-diate ring

D00 E 19.06.2001 21:36 Uhr Seite 27



series of claw pumps, which chiefly differin the type of compression process used: 1) Pumps with internal compression,

multi-stage for the semiconductorindustry (DRYVAC Series) and

2) Pumps without internal compression,two-stage for chemistry applications(“ALL·ex”).

Figs. 2.26 and 2.27 demonstrate thedifferences in design. Shown is the courseof the pressure as a function of the volumeof the pumping chamber by way of a pVdiagram.

Fig. 2.26 shows the (polytropic) course ofthe compression for pumps with internalcompression. The pressure increases untilthe discharge slot is opened. If at thatpoint the exhaust back pressure has notbeen reached, the compression space issuddenly vented with hot exhaust gas. Asthe volume is reduced further, the gas atexhaust pressure is ejected. The workdone on compression is represented bythe area under the p-v curve 1-2-3-4. It isalmost completely converted into heat. Inthe case of dry compressing vacuumpumps not much of this heat can be lost tothe cooled casing due to the low density ofthe gas. This results in high gas tempe-ratures within the pump. Experiences withclaw pumps show that the highest tem-peratures occur at the rotors.

Shown in Fig. 2.27 is the principle ofisochoric compression in a p-v diagram.Here the compression is not performed byreducing the volume of the pumpingchamber, but by venting with cold gas

which is applied from the outside aftercompletion of the intake process. This issimilar to the admission of a gas ballastwhen opening the gas ballast valve aftercompletion of the intake phase. From thediagram it is apparent that in the case ofisochoric compression the work done oncompression must be increased, but coldgas instead of hot exhaust gas is used forventing. This method of direct gas coolingresults in considerably reduced rotortemperatures. Pumps of this kind are dis-cussed as “ALL·ex” in Section 2.1.3.2.2.

2.1.3.2.1 Claw pumps with internalcompression for thesemiconductor industry(“DRYVAC series”)

Design of DRYVAC PumpsDue to the work done on compression inthe individual pumping stages, multi-stageclaw pumps require water cooling for thefour stages to remove the compressionheat. Whereas the pumping chamber ofthe pump is free of sealants and lubri-cants, the gear and the lower pump shaftare lubricated with perfluoropolyether(PFPE). The gear box is virtually herme-tically sealed from the pumping chamberby piston rings and a radial shaft seal. Thebearings in the upper end disk arelubricated with PFPE grease. In order toprotect the bearings and shaft sealsagainst aggressive substances, a barriergas facility is provided. A controlled watercooling system allows the control of thecasing temperature over a wide range asthe pump is subjected to differing gasloads coming from the process. The fourstage design is available in severalpumping speed and equipment grades of25, 50 and 100 m3/h DRYVAC pumps:a) as the basic version for non-aggressive

clean processes: DRYVAC 25 B, 50 Band 100 B (Fig. 2.29a)

b) as a version for semiconductorprocesses: DRYVAC 25 P, 50 P and 100 P (Fig. 2.29b)

c) as a system version with integrated selfmonitoring: DRYVAC 50 S and 100 S

d) as a system version with integrated self

Fig. 2.25 Pressures in pump stages 1 to 4

Stage 1

Stage 2

Stage 3

Stage 4

IN

IN

IN

IN

OUT

OUT

OUT

OUT

1 mbar

3 mbar

150 mbar

15 mbar

3 mbar 15 mbar

150 mbar

1000 mbar

P

V

W Compr.

Fig. 2.26 Compression curve for a claw pump withinternal compression

1 2

3

4

P

V

W Compr.

Fig. 2.27 Compression curve for a claw pump withoutinternal compression (“isochoric compression”)

1 2

3

Fig. 2.28 DRYVAC pump

1 Intake port 2 Front panel / electronics 3 Main switch

1

2

3

D00 E 19.06.2001 21:36 Uhr Seite 28



monitoring offering an increasedpumping speed in the lower pressurerange: DRYVAC 251 S and 501 S (Fig.2.29c)

The ultimate pressure attainable with theDRYVAC 251 S or 501 S is – compared tothe versions without integrated Rootspump – by approximately one order ofmagnitude lower (from 2·10-2 mbar to 3·10-3 mbar) and the attainable throughputis also considerably increased. It is ofcourse possible to directly flange mountLEYBOLD RUVAC pumps on to theDRYVAC models (in the case of semi-conductor processes also mostly with aPFPE oil filling for the bearing chambers).

The pumps of the DRYVAC family are theclassic dry compressing claw vacuumpumps that are preferably used in the se-miconductor industry, whereby the pumps

need to meet a variety of special require-ments. In semiconductor processes, as inmany other vacuum applications, theformation of particles and dusts during theprocess and/or in the course of compres-sing the pumped substances to atmos-pheric pressure within the pump, is una-voidable. In the case of vacuum pumpsoperating on the claw principle it ispossible to convey particles through thepump by means of so called “pneumaticconveying”. This prevents the depositionof particles and thus the formation oflayers within the pump and reduces therisk that the claw rotor may seize. Caremust be taken to ensure that the velocity ofthe gas flow within the individual pumpingstages is at all times greater than thesettling speed of the particles entrained inthe gas flow. As can be seen in Fig. 2.31,the settling speed of the particles depends

strongly on their size. The mean velocity(VGas) of the flowing gas during thecompression phase is given by thefollowing equation:

(2.22)

qpV = gas throughputp = pressureA = surface area

One can see that with increasing pressurethe velocity of the pumped gas flow slowsdown and attains the order of magnitudeof the settling speed of the particles in thegas flow (Fig. 2.32). This means that therisk of depositing particles in the operatingchamber of the pump and the resulting

vGas

qp A

mbar smbar cm

qp A

ms

pV

pV

= =·

·· ··

= ··

·

−` 1

2

10

D00

Fig. 2.29a Vacuum diagram for the DRYVAC B

Coolingwater

Exhaustline

Intake line

Casingsuction

Fig. 2.29b Vacuum diagram for the DRYVAC P

Inertgas

Coolingwater

Exhaustline

Intake line

Casingsuction

100 P only

Fig. 2.29c Vacuum diagram for the DRYVAC S

TSH

PSL

PSH

FSL

MPS

PT 100

CS

EPS

To be provided by the customer

Temperature switch

Pressure switch

Pressure switch

Flow switch

Motor protection switch

Temperature sensor

For DRYVAC with LIMS

Current sensor

Exhaust pressure sensor

Fig. 2.30 Key to Figures 2.29a – 2.29cFig. 2.31 Settling speed as a function of pressure p.

Parameter: particle size

Pressure

Particle size

Lim

it sp

eed

Fig. 2.32 Mean gas velocity vg during compressionwithout purge gas (left) and with purge gas(right) in stages 2, 3 and 4

Throughput cross sec-tion 6.5 cm2, constant

Throughput inmbar·l/s

Stage 22 500 mbar·l/s



Gas

velo

city

Pressure

D00 E 19.06.2001 21:36 Uhr Seite 29



impairment increases with increasingpressure. In parallel to this the potentialfor the formation of particles from thegaseous phase increases at increasingcompression levels. In order to keep thesize of the forming particles small and thustheir settling speed low and to maintain ahigh velocity for the gas, an additionalquantity of gas is supplied into the pumpvia the individual intermediate discs (purgegas). For this, the inflowing quantity ofpurge gas is matched to the pressureconditions prevailing in the individualpumping stages (see top right part of Fig.2.32). This keeps the velocity of the gasflow high enough within the entire pumpby so-called pneumatic pumping. Throughthe way in which the gas is lead within thepump, i.e. from the intake through the fourpumping stages with the related intermedi-

ate discs to the exhaust, it is possible toreduce the influence of the purge gas onthe ultimate pressure to a minimum. Testresults (Fig. 2.33) indicate that the in-fluence of purge gas in the fourth stage is– as to be expected – of the lowest levelsince there are located between this stageand the intake side the three other pum-ping stages. The admission of purge gasvia the second and third stages (Fig. 2.33)has a comparatively small influence on theultimate pressure as can be seen from thepumping speed curve in Fig. 2.34. Finally itcan be said that the formation of particlesis to be expected in most CVD processes.When using dry compressing clawsvacuum pumps, the controlled admissionof purge gas via the individual intermediatediscs is the best approach to avoid theformation of layers. When applying this

method several effects can be noted:• The admitted purge gas dilutes the

pumped mixture of substances, par-ticle-forming reactions will not occur, orare at least delayed.

• The risk of an explosion through self-igniting substances is significantlyreduced.

• Particles which have formed are con-veyed pneumatically through the pump

• Losses in pumping speed and a reduc-tion in ultimate pressure can be keptvery small due to the special way inwhich the gas is made to pass throughthe pump.

Fig. 2.33 Ultimate pressure of the DRYVAC 100S as a function of pure gas flowin stages 2 – 4

mbar

Ultim

ate

pres

sure

Fig. 2.34 Pumping speed with and without purge gas

Stage 2Stage 3Stage 4

Pressure

Pum

ping

spe

ed

2 500 mbar · l/s8 300 mbar · l/s20 000 mbar · l/s

Fig. 2.35 Simple arrangement of the dry compression “ALL·ex” pump

Intake port

Claws1st stage

2nd stage

Axial face seals

Motor

Coupling

Gear, completewith shafts andbearings

Purge gas flow

D00 E 19.06.2001 21:36 Uhr Seite 30



2.1.3.2.2 Claw pumps without internalcompression for chemistryapplications (“ALL·ex” )

The chemical industry requires vacuumpumps which are highly reliable and whichdo not produce waste materials such ascontaminated waste oil or waste water. Ifthis can be done, the operating costs ofsuch a vacuum pump are low in view ofthe measures otherwise required for pro-tecting the environment (disposal of wasteoil and water, for example). For operationof the simple and rugged “ALL·ex” pumpfrom LEYBOLD there are no restrictions asto the vapor flow or the pressure rangeduring continuous operation. The “ALL·ex”may be operated within the entire pressurerange from 5 to 1000 mbar withoutrestrictions.

Design of the “ALL·ex” pumpThe design of the two-stage “ALL·ex” isshown schematically in Fig. 2.35. The gasflows from top to bottom through thevertically arranged pumping stages inorder to facilitate the ejection of con-densates and rinsing liquids which mayhave formed. The casing of the pump iswater cooled and permits cooling of thefirst stage. There is no sealed link betweengas chamber and cooling channel so thatthe entry of cooling water into thepumping chamber can be excluded. Thepressure-burst resistant design of theentire unit underlines the safety concept in

view of protection against internal explo-sions, something which was also takeninto account by direct cooling with coldgas (see operating principle). A specialfeature of the “ALL·ex” is that both shaftshave their bearings exclusively in the gear.On the pumping side, the shafts are free(cantilevered). This simple design allowsthe user to quickly disassemble the pumpfor cleaning and servicing without theneed for special tools.

In order to ensure a proper seal againstthe process medium in the pumpingchamber the shaft seal is of the axial faceseal type – a sealing concept well provenin chemistry applications. This type of sealis capable of sealing liquids againstliquids, so that the pump becomesrinseable and insensitive to formingcondensate. Fig. 2.36 shows the compo-nents supplied with the “ALL·ex”, togetherwith a gas cooler and a receiver.

Operating principleIsochoric compression, which also servesthe purpose of limiting the temperatureultimately attained during compression,especially in the stage on the side of theatmosphere, and which ensures protectionagainst internal explosions, is performedby venting the pumping chamber with coldgas from a closed refrigerating gas cycle(Fig. 2.37). Fig. 2.38-1 indicates the start ofthe intake process by opening the intakeslot through the control edge of the right

rotor. The process gas then flows into theintake chamber which increases in size. Theintake process is caused by the pressuregradient produced by increasing thevolume of the pumping chamber. Themaximum volume is attained after 3/4 of arevolution of the rotors (Fig. 2.38-2). Afterthe end of the intake process, the controledge of the left rotor opens the cold gasinlet and at the same time the control edgeof the right rotor opens the intake slot (Fig.2.38-3) once more. In Fig. 2.38-4 thecontrol edge of the left rotor terminates thedischarge of the gas which has beencompressed to 1000 mbar with the coldgas; at the same time the control edge ofthe right rotor completes an intake processagain.

The total emissions from the system are

D00

Fig. 2.37 Circulation of the cold gas in the “ALL·ex”with cooler / condenser

Cooling gas

Exhaust gas

Fig. 2.36 “ALL·ex” pump

3

4

5

6

1

2

1 Motor2 Pump 3 Intake port4 Exhaust port

5 Exhaust cooler6 Cooling water

connection7 Cooling receiver

7

m h3 -1.

2

10

2

864

468

1011

100

1000

Ansaugdruckmbar100 1000

Sau

gver

mˆg

en

Fig. 2.39 Pumping speed characteristic of an ALL·ex” 250

Intake pressure

Pum

ping

spe

ed

D00 E 19.06.2001 21:36 Uhr Seite 31



not increased by the large quantities of coldgas, since a closed refrigerating cycle ismaintained by way of an externally arran-ged gas cooler and condenser (Fig. 2.37).The hot exhaust gas is made to passthrough the cooler and is partly returned inthe form of cold gas for pre-admissioncooling into the pump. The pump takes inthe quantity of cold process gas needed forventing the pumping chamber back into thecompression space on its own. This pro-cess, however, has no influence on thepumping speed of the “ALL·ex” because theintake process has already ended when theventing process starts. Designing thecooler as a condenser allows for simple

solvent recovery. The method of direct gascooling, i.e. venting of the pumping cham-ber with cold gas supplied from outside(instead of hot exhaust gas) results in thecase of the “ALL·ex” in rotor temperatureswhich are so low that mixtures of substa-nces rated as ExT3 can be pumped reliablyunder all operating conditions. The“ALL·ex” thus fully meets the requirementsof the chemical industry concerning theprotection against internal explosions. Acertain degree of liquid compatibility makesthe “ALL·ex” rinseable, thus avoiding theformation of layers in the pump, for examp-le, or the capability of dissolving layerswhich may already have formed respective-

ly. The rinsing liquids are usually applied tothe pump after completion of the connec-ted process (batch operation) or while theprocess is in progress during brief blockingphases. Even while the “ALL·ex” is atstandstill and while the pumping chamberis completely filled with liquid it is possibleto start this pump up. Shown in Fig. 2.39 isthe pumping speed characteristic of an“ALL·ex” 250. This pump has a nominalpumping speed of 250 m3/h and anultimate pressure of < 10 mbar. At 10 mbar it still has a pumping speed of 100 m3/h. The continuous operatingpressure of the pump may be as high as1000 mbar; it consumes 13.5 kW ofelectric power.

Vmax

P

P

P

P

Vmin

100 1000(10) (100)

Vmax

Vmin

100 1000(10) (100)

Vmax

Vmin

100 1000(10) (100)

Vmax

Vmin

100 1000(10) (100)

Cold gas inlet

Exhaust slot

Exhaust slot

Cold gas inlet

Cold gas inlet

Beginning of admitting cold gas

IntakeslotVolume of the pump

chamber starts to increase

Suction

Volume of the pumpchamber at maximum

End of suction

Volume of the pump chamberstars to decrease (withoutcompression). Pressureincrease to 1000 mbaronly by admitting cold gas.

Ejection of the mixturecomposed of suckedin gas and cold gas.

mbar

mbar

mbar

mbar

4

3

2

1

Fig. 2.38 Diagrams illustrating the pumping principle of the ALL·ex” pump (claw pump without inner compression)

D00 E 19.06.2001 21:36 Uhr Seite 32



2.1.4 Accessories for oil-sealed rotarydisplacement pumps

During a vacuum process, substancesharmful to rotary pumps can be present ina vacuum chamber.

Elimination of water vaporWater vapor arises in wet vacuum proces-ses. This can cause water to be depositedin the inlet line. If this condensate reachesthe inlet port of the pump, contaminationof the pump oil can result. The pumpingperformance of oil-sealed pumps can besignificantly impaired in this way. More-over, water vapor discharged through theoutlet valve of the pump can condense inthe discharge outlet line. The condensatecan, if the outlet line is not correctly ar-ranged, run down and reach the interior ofthe pump through the discharge outletvalve. Therefore, in the presence of watervapor and other vapors, the use ofcondensate traps is strongly recommen-ded. If no discharge outlet line is connec-ted to the gas ballast pump (e.g., withsmaller rotary vane pumps), the use ofdischarge filters is recommended. Thesecatch the oil mist discharged from thepump.

Some pumps have easily exchangeable fil-ter cartridges that not only hold back oilmist, but clean the circulating pump oil.Whenever the amount of water vapor pre-sent is greater than the water vapor tole-rance of the pump, a condenser shouldalways be installed between the vessel andthe pump. (For further details, see Section2.1.5)

Elimination of dustSolid impurities, such as dust and grit,significantly increase the wear on thepistons and the surfaces in the interior ofthe pump housing. If there is a danger thatsuch impurities can enter the pump, a dustseparator or a dust filter should be in-stalled in the inlet line of the pump. Todaynot only conventional filters having fairlylarge casings and matching filter insertsare available, but also fine mesh filterswhich are mounted in the centering ring ofthe small flange. If required, it is recom-mended to widen the cross section with KFadaptors.

Elimination of oil vapor The attainable ultimate pressure with oil-sealed rotary pumps is strongly influencedby water vapor and hydrocarbons from thepump oil. Even with two-stage rotary vanepumps, a small amount of back-streamingof these molecules from the pump interiorinto the vacuum chamber cannot beavoided. For the production of hydrocar-bon-free high and ultrahigh vacuum, forexample, with sputter-ion or turbomolecu-lar pumps, a vacuum as free as possible ofoil is also necessary on the forevacuumside of these pumps. To obtain this, me-dium vacuum adsorption traps (see Fig.2.40) filled with a suitable adsorptionmaterial (e.g., LINDE molecular sieve 13X)are installed in the inlet line of such oil-sealed forepumps. The mode of action of asorption trap is similar to that of an ad-sorption pump. For further details, seeSection 2.1.8. If foreline adsorption trapsare installed in the inlet line of oil-sealedrotary vane pumps in continuous opera-tion, two adsorption traps in parallel arerecommended, each separated by valves.Experience shows that the zeolite used asthe adsorption material loses much of itsadsorption capacity after about 10 – 14days of running time, after which theother, now-regenerated, adsorption trapcan be utilized; hence the process cancontinue uninterrupted. By heating theadsorption trap, which is now not connec-ted in the pumping line, the vaporsescaping from the surface of the zeolitecan be most conveniently pumped awaywith an auxiliary pump. In operation, pum-ping by the gas ballast pump generallyleads to a covering of the zeolite in theother, unheated adsorption trap and thusto a premature reduction of the adsorptioncapacity of this trap.

Reduction of the effective pumpingspeedAll filters, separators, condensers, andvalves in the inlet line reduce the effectivepumping speed of the pump. On the basisof the values of the conductances orresistances normally supplied bymanufacturers, the actual pumping speedof the pump can be calculated. For furtherdetails, see Section 1.5.2.

2.1.5 CondensersFor pumping larger quantities of watervapor, the condenser is the most econo-mical pump. As a rule, the condenser iscooled with water of such temperature thatthe condenser temperature lies sufficientlybelow the dew point of the water vapor andan economical condensation or pumpingaction is guaranteed. For cooling, however,media such as brine and refrigerants (NH3,Freon) can also be used.When pumping water vapor in a largeindustrial plant, a certain quantity of air isalways involved, which is either containedin the vapor or originates from leaks in theplant (the following considerations for airand water vapor obviously apply also ingeneral for vapors other than water vapor).Therefore, the condenser must be backed

D00

Fig. 2.40 Cross section of a medium vacuum adsorption trap

1 Housing2 Basket holding the sieve3 Molecular sieve (filling) 4 Sealing flanges5 Intake port with small flange6 Upper section7 Vessel for the heater or refrigerant8 Connection on the side of the pump with

small flange

D00 E 19.06.2001 21:36 Uhr Seite 33



by a gas ballast pump (see Fig. 2.41) andhence always works – like the Roots pump– in a combination. The gas ballast pumphas the function of pumping the fraction ofair, which is often only a small part of thewater-vapor mixture concerned, withoutsimultaneously pumping much watervapor. It is, therefore, understandable that,within the combination of condenser andgas ballast pump in the stationary con-dition, the ratios of flow, which occur in theregion of rough vacuum, are not easilyassessed without further consideration.The simple application of the continuityequation is not adequate because one is nolonger concerned with a source or sink-freefield of flow (the condenser is, on the basisof condensation processes, a sink). This isemphasized especially at this point. In apractical case of “non-functioning” of thecondenser – gas ballast pump combination,it might be unjustifiable to blame thecondenser for the failure.

In sizing the combination of condenserand gas ballast pump, the following pointsmust be considered:a) the fraction of permanent gases (air)

pumped simultaneously with the watervapor should not be too great. At partialpressures of air that are more thanabout 5 % of the total pressure at theexit of the condenser, a marked accu-mulation of air is produced in front ofthe condenser surfaces. The condenserthen cannot reach its full capacity (Seealso the account in Section 2.2.3 on thesimultaneous pumping of gases andvapors).

b) the water vapor pressure at thecondenser exit – that is, at the inlet side

of the gas ballast pump – should not(when the quantity of permanent gasdescribed in more detail in Section2.2.3 is not pumped simultaneously) begreater than the water vapor tolerancefor the gas ballast pump involved. If –as cannot always be avoided in practice– a higher water vapor partial pressureis to be expected at the condenser exit,it is convenient to insert a throttle bet-ween the condenser exit and the inletport of the gas ballast pump. Theconductance of this throttle should bevariable and regulated (see Section1.5.2) so that, with full throttling, thepressure at the inlet port of the gasballast pump cannot become higherthan the water vapor tolerance. Also,the use of other refrigerants or a de-crease of the cooling water temperaturemay often permit the water vapor pres-sure to fall below the required value.

For a mathematical evaluation of thecombination of condenser and gas ballastpump, it can be assumed that no loss ofpressure occurs in the condenser, that thetotal pressure at the condenser entranceptot 1, is equal to the total pressure at thecondenser exit, ptot 2:

ptot 1 = ptot 2 (2.23)

The total pressure consists of the sum ofthe partial pressure portions of the air ppand the water vapor pv:

pp1 + pv1 = pp2 + pv2 (2.23a)

As a consequence of the action of the

condenser, the water vapor pressure pD2 atthe exit of the condenser is always lowerthan that at the entrance; for (2.23) to befulfilled, the partial pressure of air pp2 atthe exit must be higher than at theentrance pp1, (see Fig. 2.43), even whenno throttle is present.

The higher air partial pressure pp2 at thecondenser exit is produced by an accumu-lation of air, which, as long as it is presentat the exit, results in a stationary flowequilibrium. From this accumulation of air,the (eventually throttled) gas ballast pumpin equilibrium removes just so much asstreams from the entrance (1) through thecondenser.

All calculations are based on (2.23a) forwhich, however, information on thequantity of pumped vapors and permanentgases, the composition, and the pressureshould be available. The size of thecondenser and gas ballast pump can becalculated, where these two quantities are,indeed, not mutually independent. Fig. 2.42represents the result of such a calculationas an example of a condenser having acondensation surface of 1 m2, and at aninlet pressure pv1, of 40 mbar, acondensation capacity that amounts to 15 kg / h of pure water vapor if the fractionof the permanent gases is very small. 1 m3

of cooling water is used per hour, at a lineoverpressure of 3 bar and a temperature of12 °C. The necessary pumping speed of thegas ballast pump depends on the existingoperating conditions, particularly the size ofthe condenser. Depending on the efficiency

Fig. 2.42 Condensation capacity of the condenser(surface area available to condensation 1 m2)as a function of intake pressure pD1 of thewater vapor. Curve a: Cooling water tempera-ture 12 °C. Curve b: Temperature 25 °C.Consumption in both cases 1 m3/h at 3 baroverpressure

Intake pressure pD1

Cond

ensa

tion

capa

city

[kg

· h-1

]

Fig. 2.43 Schematic representation of the pressure dis-tribution in the condenser. The full lines corre-spond to the conditions in a condenser inwhich a small pressure drop takes place (ptot 2 < ptot 1). The dashed lines are those foran ideal condenser (ptot 2 ≈ ptot 1). pD: Partialpressure of the water vapor, pL: Partial pressu-re of the air

Pv1

P’v2

Pp1

Pv2

Pp2P’p2

Fig. 2.41 Condenser (I) with downstream gas ballastpump (II) for pumping of large quantities ofwater vapor in the rough vacuum range (III)– adjustable throttle

1 Inlet of the condenser 2 Discharge of the condenser 3 See text

D00 E 19.06.2001 21:36 Uhr Seite 34



of the condenser, the water vapor partialpressure pv2 lies more or less above thesaturation pressure pS which correspondsto the temperature of the refrigerant. (Bycooling with water at 12 °C, pS, would be15 mbar (see Table XIII in Section 9)).Correspondingly, the partial air pressurepp2 that prevails at the condenser exit alsovaries. With a large condenser, pv2 ≈ pS, the air partial pressure pp,2 is thuslarge, and because pp · V = const, thevolume of air involved is small. Therefore,only a relatively small gas ballast pump isnecessary. However, if the condenser issmall, the opposite case arises:pv2 > pS · pp2, is small. Here a relativelylarge gas ballast pump is required. Sincethe quantity of air involved during apumping process that uses condensers isnot necessarily constant but alternateswithin more or less wide limits, theconsiderations to be made are moredifficult. Therefore, it is necessary that thepumping speed of the gas ballast pumpeffective at the condenser can be regulatedwithin certain limits. In practice, the following measures areusual:a) A throttle section is placed between the

gas ballast pump and the condenser,which can be short-circuited duringrough pumping. The flow resistance ofthe throttle section must be adjustableso that the effective speed of the pumpcan be reduced to the required value.This value can be calculated using theequations given in Section 2.2.3.

b) Next to the large pump for rough pum-ping a holding pump with low speed isinstalled, which is of a size corres-ponding to the minimum prevailing gasquantity. The objective of this holdingpump is merely to maintain optimumoperating pressure during the process.

c) The necessary quantity of air is ad-mitted into the inlet line of the pumpthrough a variable-leak valve. Thisadditional quantity of air acts like anenlarged gas ballast, increasing thewater vapor tolerance of the pump.However, this measure usually resultsin reduced condenser capacity. More-over, the additional admitted quantity ofair means additional power consump-tion and (see Section 8.3.1.1) increasedoil consumption. As the efficiency of thecondenser deteriorates with too great apartial pressure of air in the condenser,the admission of air should not be in

front, but generally only behind thecondenser.

If the starting time of a process is shorterthan the total running time, technically thesimplest method – the roughing and theholding pump – is used. Processes withstrongly varying conditions require anadjustable throttle section and, if needed,an adjustable air admittance.

On the inlet side of the gas ballast pump awater vapor partial pressure pv2 is alwayspresent, which is at least as large as thesaturation vapor pressure of water at thecoolant temperature. This ideal case isrealizable in practice only with a very largecondenser (see above).With a view to practice and from the statedfundamental rules, consider the twofollowing cases:

1. Pumping of permanent gases withsmall amounts of water vapor. Here thesize of the condenser – gas ballast pumpcombination is decided on the basis of thepumped-off permanent gas quantity. Thecondenser function is merely to reduce thewater vapor pressure at the inlet port ofthe gas ballast pump to a value below thewater vapor tolerance.

2. Pumping of water vapor with smallamounts of permanent gases. Here, tomake the condenser highly effective, assmall as possible a partial pressure of thepermanent gases in the condenser issought. Even if the water vapor partialpressure in the condenser should begreater than the water vapor tolerance ofthe gas ballast pump, a relatively small gasballast pump is, in general, sufficient withthe then required throttling to pump awaythe prevailing permanent gases.

Important note: During the process, if thepressure in the condenser drops below thesaturation vapor pressure of the conden-sate (dependent on the cooling watertemperature), the condenser must beblocked out or at least the collectedcondensate isolated. If this is not done, thegas ballast pump again will pump out thevapor previously condensed in thecondenser

2.1.6 Fluid-entrainment pumpsBasically, a distinction is made betweenejector pumps such as water jet pumps(17 mbar < p < 1013 mbar), vapor ejectorvacuum pumps (10-3 mbar < p < 10-1 mbar) and diffusionpumps (p < 10-3 mbar). Ejector vacuumpumps are used mainly for the productionof medium vacuum. Diffusion pumpsproduce high and ultrahigh vacuum. Bothtypes operate with a fast-moving stream ofpump fluid in vapor or liquid form (waterjet as well as water vapor, oil or mercuryvapor). The pumping mechanism of allfluid-entrainment pumps is basically thesame. The pumped gas molecules areremoved from the vessel and enter into thepump fluid stream which expands afterpassing through a nozzle. The moleculesof the pump fluid stream transfer by wayof impact impulses to the gas molecules inthe direction of the flow. Thus the gaswhich is to be pumped is moved to a spacehaving a higher pressure.In fluid-entrainment pumps correspondingvapor pressures arise during operationdepending on the type of pump fluid andthe temperature as well as the design ofthe nozzle. In the case of oil diffusionpumps this may amount to 1 mbar in theboiling chamber. The backing pressure inthe pump must be low enough to allow thevapor to flow out. To ensure this, suchpumps require corresponding backingpumps, mostly of the mechanical type. Thevapor jet cannot enter the vessel since itcondenses at the cooled outer walls of thepump after having been ejected throughthe nozzle.

Wolfgang Gaede was the first to realizethat gases at comparatively low pressurecan be pumped with the aid of a pumpfluid stream of essentially higher pressureand that, therefore, the gas moleculesfrom a region of low total pressure moveinto a region of high total pressure. Thisapparently paradoxical state of affairsdevelops as the vapor stream is initiallyentirely free of gas, so that the gases froma region of higher partial gas pressure (thevessel) can diffuse into a region of lowerpartial gas pressure (the vapor stream).This basic Gaede concept was used byLangmuir (1915) in the construction of thefirst modern diffusion pump. The firstdiffusion pumps were mercury diffusionpumps made of glass, later of metal. In theSixties, mercury as the medium was

D00

D00 E 19.06.2001 21:36 Uhr Seite 35



almost completely replaced by oil. Toobtain as high a vapor stream velocity aspossible, he allowed the vapor stream toemanate from a nozzle with supersonicspeed. The pump fluid vapor, which con-stitutes the vapor jet, is condensed at thecooled wall of the pump housing, whereasthe transported gas is further compressed,usually in one or more succeeding stages,before it is removed by the backing pump.The compression ratios, which can beobtained with fluid entrainment pumps,are very high: if there is a pressure of 10-9

mbar at the inlet port of the fluid-entrainment pump and a backing pressureof 10-2 mbar, the pumped gas is com-pressed by a factor of 107!Basically the ultimate pressure of fluidentrainment pumps is restricted by thevalue for the partial pressure of the fluidused at the operating temperature of thepump. In practice one tries to improve thisby introducing baffles or cold traps. Theseare “condensers” between fluid entrain-ment pump and vacuum chamber, so thatthe ultimate pressure which can beattained in the vacuum chamber is nowonly limited by the partial pressure of thefluid at the temperature of the baffle.

The various types of fluid entrainmentpumps are essentially distinguished by thedensity of the pump fluid at the exit of thetop nozzle facing the high vacuum side ofthe pump: 1. Low vapor density:

Diffusion pumps

Oil diffusion pumps (Series: LEYBODIFF, DI and DIP)Mercury diffusion pumps

2. High vapor density: Vapor jet pumps

Water vapor pumpsOil vapor jet pumpsMercury vapor jet pumps

3. Combined oil diffusion/ vapor jet pumps

4. Water jet pumps

Cooling of fluid entrainment pumpsThe heater power that is continuouslysupplied for vaporizing the pump fluid influid-entrainment pumps must be dissipa-ted by efficient cooling. The energyrequired for pumping the gases andvapors is minimal. The outside walls of thecasing of diffusion pumps are cooled,

generally with water. Smaller oil diffusionpumps can, however, also be cooled withan air stream because a low wall tempe-rature is not so decisive to the efficiency asfor mercury diffusion pumps. Oil diffusionpumps can operate well with wall tempera-tures of 30 °C, whereas the walls of mer-cury diffusion pumps must be cooled to 15 °C. To protect the pumps from thedanger of failure of the cooling water –insofar as the cooling-water coil is notcontrolled by thermally operated protec-tive switching – a water circulation moni-tor should be installed in the cooling watercircuit; hence, evaporation of the pumpfluid from the pump walls is avoided.

2.1.6.1 (Oil) Diffusion pumps

These pumps consist basically (see Fig.2.44) of a pump body (3) with a cooledwall (4) and a three- or four-stage nozzlesystem (A – D). The oil serving as pumpfluid is in the boiler (2) and is vaporizedfrom here by electrical heating (1). Thepump fluid vapor streams through theriser tubes and emerges with supersonicspeed from the ring-shaped nozzles (A –D). Thereafter the jet so-formed widenslike an umbrella and reaches the wallwhere condensation of the pump fluidoccurs. The liquid condensate flowsdownward as a thin film along the wall andfinally returns into the boiler. Because ofthis spreading of the jet, the vapor densityis relatively low. The diffusion of air or anypumped gases (or vapors) into the jet is sorapid that despite its high velocity the jet

Fig. 2.44 Mode of operation of a diffusion pump

1 Heater2 Boiler3 Pump body4 Cooling coil

5 High vacuum flange6 Gas molecules7 Vapor jet8 Backing vacuum connection

ABC NozzlesDH

D00 E 19.06.2001 21:36 Uhr Seite 36



becomes virtually completely saturatedwith the pumped medium. Therefore, overa wide pressure range diffusion pumpshave a high pumping speed. This ispractically constant over the entireworking region of the diffusion pump (≤ 10-3 mbar) because the air at these lowpressures cannot influence the jet, so itscourse remains undisturbed. At higherinlet pressures, the course of the jet isaltered. As a result, the pumping speeddecreases until, at about 10-1 mbar, itbecomes immeasurably small.The forevacuum pressure also influencesthe vapor jet and becomes detrimental ifits value exceeds a certain critical limit.This limit is called maximum backingpressure or critical forepressure. Thecapacity of the chosen backing pump mustbe such (see 2.3.2) that the amount of gasdischarged from the diffusion pump ispumped off without building up a backingpressure that is near the maximumbacking pressure or even exceeding it.

The attainable ultimate pressure dependson the construction of the pump, the vaporpressure of the pump fluid used, themaximum possible condensation of thepump fluid, and the cleanliness of thevessel. Moreover, backstreaming of thepump fluid into the vessel should bereduced as far as possible by suitablebaffles or cold traps (see Section 2.1.6.4).

Degassing of the pump oilIn oil diffusion pumps it is necessary forthe pump fluid to be degassed before it isreturned to the boiler. On heating of thepump oil, decomposition products canarise in the pump. Contamination from thevessel can get into the pump or becontained in the pump in the first place.These constituents of the pump fluid cansignificantly worsen the ultimate pressureattainable by a diffusion pump, if they arenot kept away from the vessel. Therefore,the pump fluid must be freed of theseimpurities and from absorbed gases.

This is the function of the degassingsection, through which the circulating oilpasses shortly before re-entry into theboiler. In the degassing section, the mostvolatile impurities escape. Degassing isobtained by the carefully controlledtemperature distribution in the pump. Thecondensed pump fluid, which runs downthe cooled walls as a thin film, is raised toa temperature of about 130 °C below the

lowest diffusion stage, to allow the volatilecomponents to evaporate and be removedby the backing pump. Therefore, the re-evaporating pump fluid consists of onlythe less volatile components of the pumpoil.

Pumping speedThe magnitude of the specific pumpingspeed S of a diffusion pump – that is, thepumping speed per unit of area of theactual inlet surface – depends on severalparameters, including the position anddimensions of the high vacuum stage, thevelocity of the pump fluid vapor, and themean molecular velocity c- of the gas beingpumped (see equation 1.17 in Section1.1). With the aid of the kinetic theory ofgases, the maximum attainable specificpumping speed at room temperature onpumping air is calculated to Smax = 11.6 l · s-1 · cm-2. This is the spe-cific (molecular) flow conductance of theintake area of the pump, resembling anaperture of the same surface area (seeequation 1.30 in Section 1.5.3). Quitegenerally, diffusion pumps have a higherpumping speed for lighter gases comparedto heavier gases.

To characterize the effectiveness of adiffusion pump, the so called HO factor isdefined. This is the ratio of the actuallyobtained specific pumping speed to thetheoretical maximum possible specificpumping speed. In the case of diffusionpumps from LEYBOLD optimum valuesare attained (of 0.3 for the smallest and upto 0.55 for the larger pumps).

The various oil diffusion pumps manufac-tured by LEYBOLD differ in the followingdesign features (see Fig. 2.45 a and b).

a) LEYBODIFF seriesThis series of pumps is equipped with afractionating device. The various consti-tuents of the pump fluid are selected sothat the high vacuum nozzle is suppliedonly by the fraction of the pump fluid thathas the lowest vapor pressure. Thisassures a particularly low ultimate pres-sure. Fractionating occurs because thedegassed oil first enters the outer part ofthe boiler, which serves the nozzle on thebacking vacuum side. Here a part of themore volatile constituents evaporates.Hence the already purified pump fluidreaches the intermediate part of the boiler,which serves the intermediate nozzle. Here

D00Fig. 2.45 Diagram showing the basic differences in LEYBOLD oil diffusion pumps

1

2

a) LEYBODIFF-Pump with fractionatingfacility1 Center2 Middle section3 Outer part of the boiler

(fractionation)

b) DI pump; side view on to theinternal heater1 Thermostat sensor2 Heating cartridge

D00 E 19.06.2001 21:37 Uhr Seite 37



the lighter constituents are evaporated ingreater quantities than the heavier con-stituents. When the oil enters the centralregion of the boiler, which serves the highvacuum nozzle, it has already been freed ofthe light volatile constituents.

b) DI seriesIn these pumps an evaporation process forthe pump fluid which is essentially free ofbursts is attained by the exceptional heaterdesign resulting in a highly constantpumping speed over time. The heater is ofthe internal type and consists of heatingcartridges into which tubes with solderedon thermal conductivity panels areintroduced. The tubes made of stainlesssteel are welded horizontally into thepump’s body and are located above the oillevel. The thermal conductivity panelsmade of copper are only in part immersedin the pump fluid. Those parts of thethermal conductivity panels are so ratedthat the pump fluid can evaporateintensively but without any retardation of

boiling. Those parts of the thermalconductivity panels above the oil levelsupply additional energy to the vapor.Owing to the special design of the heatingsystem, the heater cartridges may beexchanged also while the pump is still hot.

2.1.6.2 Oil vapor ejector pumps

The pumping action of a vapor ejectorstage is explained with the aid of Fig. 2.46.The pump fluid enters under high pressurep1 the nozzle (1), constructed as a Lavalnozzle. There it is expanded to the inletpressure p2. On this expansion, thesudden change of energy is accompaniedby an increase of the velocity. Theconsequently accelerated pump fluidvapor jet streams through the mixer region(3), which is connected to the vessel (4)being evacuated. Here the gas moleculesemerging from the vessel are draggedalong with the vapor jet. The mixture,pump fluid vapor – gas, now enters thediffuser nozzle constructed as a Venturinozzle (2). Here the vapor – gas mixture iscompressed to the backing pressure p3with simultaneous diminution of thevelocity. The pump fluid vapor is then con-densed at the pump walls, whereas theentrained gas is removed by the backingpump. Oil vapor ejector pumps are ideallysuited for the pumping of larger quantitiesof gas or vapor in the pressure regionbetween 1 and 10-3 mbar. The higherdensity of the vapor stream in the nozzlesensures that the diffusion of the pumpedgas in the vapor stream takes place muchmore slowly than in diffusion pumps, sothat only the outer layers of the vaporstream are permeated by gas. Moreover,the surface through which the diffusionoccurs is much smaller because of thespecial construction of the nozzles. Thespecific pumping speed of the vaporejector pumps is, therefore, smaller thanthat of the diffusion pumps. As the pum-ped gas in the neighborhood of the jetunder the essentially higher inlet pressuredecisively influences the course of the flowlines, optimum conditions are obtainedonly at certain inlet pressures. Therefore,the pumping speed does not remain con-stant toward low inlet pressures. As aconsequence of the high vapor streamvelocity and density, oil vapor ejectorpumps can transport gases against a rela-tively high backing pressure. Their criticalbacking pressure lies at a few millibars.

The oil vapor ejector pumps used inpresent-day vacuum technology have, ingeneral, one or more diffusion stages andseveral subsequent ejector stages. Thenozzle system of the booster is construc-ted from two diffusion stages and twoejector stages in cascade (see Fig. 2.47).The diffusion stages provide the highpumping speed between 10-4 and 10-3

mbar (see Fig. 2.48), the ejector stages,the high gas throughput at high pressures(see Fig. 2.49) and the high critical backingpressure. Insensitivity to dust and vaporsdissolved in the pump fluid is obtained bya spacious boiler and a large pump fluid

Fig. 2.46 Operation of a vapor jet pump

1 Nozzle (Laval) 2 Diffuser nozzle (Venturi) 3 Mixing chamber4 Connection to the vacuum chamber

1 High vacuum port2 Diffusion stages3 Ejector stages

Fig. 2.47 Diagram of an oil jet (booster) pump

Fig. 2.48 Pumping speed of various vapor pumps as afunction of intake pressure related to a nomi-nal pumping speed of 1000 l/s.End of the working range of oil vapor ejectorpumps (A) and diffusion pumps (B)

Intake pressure pa

Pum

ping

spe

ed [l

· s–1

]

Fig. 2.49 Pumping speed of various vapor pumps (derived from Fig. 2.48)

Intake pressure pa

Pum

ping

spe

ed [T

orr ·

l · s

–1]

Pum

ping

spe

ed [m

bar ·

l · s

–1]

D00 E 19.06.2001 21:37 Uhr Seite 38



reservoir. Large quantities of impuritiescan be contained in the boiler without de-terioration of the pumping characteristics.

2.1.6.3 Pump fluids

a) OilsThe suitable pump fluids for oil diffusionpumps are mineral oils, silicone oils, andoils based on the polyphenyl ethers.Severe demands are placed on such oilswhich are met only by special fluids. Theproperties of these, such as vapor pres-sure, thermal and chemical resistance,particularly against air, determine thechoice of oil to be used in a given type ofpump or to attain a given ultimate vacuum.The vapor pressure of the oils used invapor pumps is lower than that of mer-cury. Organic pump fluids are more sen-sitive in operation than mercury, becausethe oils can be decomposed by long-termadmission of air. Silicone oils, however,withstand longer lasting frequent admis-sions of air into the operational pump.

Typical mineral oils are DIFFELEN light,normal and ultra. The different types ofDIFFELEN are close tolerance fractions of ahigh quality base product (see our cata-log).

Silicone oils (DC 704, DC 705, for examp-le) are uniform chemical compounds(organic polymers). They are highlyresistant to oxidation in the case of airinrushes and offer special thermal stabilitycharacteristics.

DC 705 has an extremely low vapor pres-sure and is thus suited for use in diffusionpumps which are used to attain extremelylow ultimate pressures of < 10-10 mbar.ULTRALEN is a polyphenylether. This fluidis recommended in all those cases where aparticularly oxidation-resistant pump fluidmust be used and where silicone oilswould interfere with the process.

APIEZON AP 201 is an oil of exceptionalthermal and chemical resistance capableof delivering the required high pumpingspeed in connection with oil vapor ejectorpumps operating in the medium vacuumrange. The attainable ultimate total pres-sure amounts to about 10-4 mbar.

b) MercuryMercury is a very suitable pump fluid. It isa chemical element that during vaporiza-

tion neither decomposes nor becomesstrongly oxidized when air is admitted.However, at room temperature it has acomparatively high vapor pressure of 10-3 mbar. If lower ultimate total pressuresare to be reached, cold traps with liquidnitrogen are needed. With their aid,ultimate total pressures of 10-10 mbar canbe obtained with mercury diffusion pumps.Because mercury is toxic, as alreadymentioned, and because it presents ahazard to the environment, it is nowadayshardly ever used as a pump fluid.LEYBOLD supplies pumps with mercury asthe pump fluid only on request. The vaporpressure curves of pump fluids are given inFig. 9.12, Section 9.

2.1.6.4 Pump fluid backstreaming andits suppression (vaporbarriers, baffles)

In the vapor stream from the topmostnozzle of a diffusion pump, pump fluidmolecules not only travel in the directionof streaming to the cooled pump wall, butalso have backward components ofvelocity because of intermolecular colli-sions. They can thus stream in the direc-tion of the vessel. In the case ofLEYBODIFF and DI pumps, the oil-back-streaming amounts to a few microgramsper minute for each square centimeter ofinlet cross-sectional area. To reduce thisbackstreaming as much as possible,various measures must be undertakensimultaneously:a) the high vacuum-side nozzle and the

shape of the part of the pump bodysurrounding this nozzle must beconstructed so that as few as possiblevapor molecules emerge sideways inthe path of the vapor stream from thenozzle exit to the cooled pump wall.

b) the method for cooling the pump wallmust allow as complete as possiblecondensation of the pump fluid vaporand, after condensation, the fluid mustbe able to drain away readily.

c) one or more pump-fluid traps, baffles,or cold traps must be inserted betweenthe pump and the vessel, depending onthe ultimate pressure that is required.

Two chief requirements must be met in theconstruction of baffles or cold traps for oildiffusion pumps. First, as far as possible,all backstreaming pump-fluid vapor mole-cules should remain attached to (conden-

sed at) the inner cooled surfaces of thesedevices. Second, the condensation sur-faces must be so constructed and geome-trically arranged that the flow conductanceof the baffles or cold traps is as large aspossible for the pumped gas. These tworequirements are summarized by the term“optically opaque”. This means that theparticles cannot enter the baffle withouthitting the wall, although the baffle has ahigh conductance. The implementation ofthis idea has resulted in a variety ofdesigns that take into account one or theother requirement.

A cold cap baffle is constructed so that itcan be mounted immediately above thehigh vacuum nozzle. The cold cap baffle ismade of metal of high thermal conductivityin good thermal contact with the cooledpump wall, so that in practice it is main-tained at the cooling-water temperature or,with air-cooled diffusion pumps, at ambi-ent temperature. In larger types of pumps,the cold cap baffle is water cooled andpermanently attached to the pump body.The effective pumping speed of a diffusionpump is reduced by about 10 % oninstallation of the cold cap baffle, but theoil backstreaming is reduced by about 90to 95 %.

Shell baffles consist of concentricallyarranged shells and a central baffle plate.With appropriate cooling by water orrefrigeration, almost entirely oil vapor-freevacua can be produced by this means. Theeffective pumping speed of the diffusionpump remains at least at 50 %, althoughshell baffles are optically opaque. This typeof baffle has been developed by LEYBOLDin two different forms: with a stainless-steel cooling coil or – in the so-calledAstrotorus baffles – with cooling inserts ofcopper. The casing of the former type ismade entirely of stainless steel.

For the smaller air-cooled, oil diffusionpumps, plate baffles are used. The air-cooled arrangement consists of a copperplate with copper webs to the housingwall. The temperature of the plate baffleremains nearly ambient during theoperation of the diffusion pump.

Hydrocarbon-free vacuumIf extreme demands are made on freedomfrom oil vapor with vacuum produced byoil diffusion pumps, cold traps should beused that are cooled with liquid nitrogen

D00

D00 E 19.06.2001 21:37 Uhr Seite 39



so that they are maintained at a tempera-ture of -196 °C.

Low-temperature baffles or cold trapsshould always be used with a cold cap inplace. On this the greatest part of thebackstreaming oil is condensed, so thatthe inevitable loss of pump fluid from thecondensation of the pump fluid on the low-temperature surface is kept at a minimum.With longer-term operation it is always ad-visable to install, in place of the cold cap, awater-cooled shell or chevron bafflebetween the diffusion pump and the low-temperature baffle or cold trap (see Fig.2.50).

LEYBOLD manufactures cold traps madeof metal so called LN2 cold traps. Thesecold traps are to be used in those caseswhere a cold trap is to be operated forprolonged periods of time without re-quiring a filling facility for liquid nitrogen.

The temperature increase at the vesselcontaining the refrigerant is so slight overthe operating time that – as the liquid leveldrops – no significant desorption of thecondensate takes place. Located on thepumping side is an impact panel made ofcopper. The low temperature of this panelensures that the greater part of thecondensed pump fluid remains in theliquid state and may drip back into thepump. Today the oils used to operate dif-fusion pumps have a very low vapor pres-sure at room temperature (for exampleDIFFELEN light, 2 · 10-8 mbar; DC 705, 4 · 10-10 mbar). The specified provisionswith a liquid-nitrogen-cooled baffle or coldtrap would enable an absolutely oil-freevacuum to be produced. In practice, however, complete suppres-sion of oil-backstreaming is never at-tained. There are always a few pump-fluidmolecules that, as a result of collisionswith one another, reach the vessel withouthaving hit one of the cooled surfaces of thebaffle or the cold trap. Moreover, there arealways a few highly volatile components ofthe pump fluid that do not remain attachedto the very low temperature surfaces. Thetemperature and the vapor moleculesadsorbed at the surface of the vesseldetermine exactly the pressure in thevessel. If, the surfaces are not fully cove-red with adsorbed pump-fluid moleculesafter a bake-out process, their vaporpressure contributes only insignificantly tothe pressure in the vessel.

After a certain time, the “stay-down time”,a continuous layer of oil molecules buildsup, and the ultimate pressure is practicallydetermined by the vapor pressure of thepump fluid at the temperature of the vesselwalls. This “stay-down” time can evenamount to several hours, indeed even todays, with the use of low-temperaturebaffles.

Oil can reach the vessel not only as vapor,but also as a liquid film, because oil wetsreadily and thus creeps up the wall.

By installation of an anticreep barrier (seeFig. 2.50) made of Teflon polymer, a mate-rial that is not wetted by oil and can standa bake-out temperature up to 200 °C,further creeping of the oil can be effecti-vely prevented. It is most appropriate toarrange the anticreep barrier above theupper baffle (see Fig. 2.50).

Note:It must be noted that data on back-streaming as specified in catalogs applyonly to continuously-operated oil diffusionpumps. Shortly after starting a pump theuppermost nozzle will not eject a welldirected vapor jet. Instead oil vaporspreads in all directions for several se-conds and the backstreaming effect isstrong. When switching a diffusion pumpon and off frequently the degree of oilbrackstreaming will be greater.

2.1.6.5 Water jet pumps and steamejectors

Included in the class of fluid-entrainmentpumps are not only pumps that use a fast-streaming vapor as the pump fluid, butalso liquid jet pumps. The simplest andcheapest vacuum pumps are water jetpumps. As in a vapor pump (see Fig. 2.46or 2.51), the liquid stream is first releasedfrom a nozzle and then, because ofturbulence, mixes with the pumped gas inthe mixing chamber. Finally, the movementof the water – gas mixture is slowed downin a Venturi tube. The ultimate totalpressure in a container that is pumped bya water jet pump is determined by thevapor pressure of the water and, forexample, at a water temperature of 15 °Camounts to about 17 mbar.

Essentially higher pumping speeds andlower ultimate pressures are produced bysteam ejector pumps. The sectionthrough one stage is shown in Fig. 2.51.The markings correspond with thoseshown in Fig. 2.46. In practice, severalpumping stages are usually mounted incascade. For laboratory work, two-stagepump combinations are suitable andconsist of a steam ejector stage and awater jet (backing) stage, both made ofglass. The water jet backing stage enablesoperation without other backing pumps.With the help of a vapor stream atoverpressure, the vacuum chamber can beevacuated to an ultimate pressure of about3 mbar. The condensate from the steam isled off through the drain attachment. Thewater jet stage of this pump is cooled withwater to increase its efficiency. Steamejector pumps are especially suitable forwork in laboratories, particularly if veryaggressive vapors are to be pumped.Steam ejector pumps, which will operateat a pressure of a few millibars, are

Fig. 2.50 Schematic arrangement of baffle, anticreepbarrier and cold trap above a diffusion pump

1 Diffusion pump with cold cap baffle (cooled by contact),

2 Shell or chevron baffle 3 Anticreep barrier4 Sealing gasket5 Bearing ring6 LN2 cold trap, 7 Vacuum chamber

D00 E 19.06.2001 21:37 Uhr Seite 40

2.1.7 Turbomolecular pumps

The principle of the molecular pump – wellknown since 1913 – is that the gas par-ticles to be pumped receive, throughimpact with the rapidly moving surfaces ofa rotor, an impulse in a required flow direc-tion. The surfaces of the rotor – usuallydisk-shaped – form, with the stationarysurfaces of a stator, intervening spaces inwhich the gas is transported to the backingport. In the original Gaede molecularpump and its modifications, the inter-vening spaces (transport channels) werevery narrow, which led to constructionaldifficulties and a high degree of sus-ceptibility to mechanical contamination.

At the end of the Fifties, it became possible– through a turbine-like design and bymodification of the ideas of Gaede – toproduce a technically viable pump the so-called “turbomolecular pump”. Thespaces between the stator and the rotordisks were made in the order ofmillimeters, so that essentially largertolerances could be obtained. Thereby,greater security in operation was achieved.However, a pumping effect of anysignificance is only attained when thecircumferential velocity (at the outsiderim) of the rotor blades reaches the orderof magnitude of the average thermalvelocity of the molecules which are to bepumped. Kinetic gas theory supplies for c-

of the equation 1.17:

in which the dependency on the type of gasas a function of molar mass M is contained.The calculation involving cgs-units (where R = 83.14 · 106 mbar · cm3/mol · K)results in the following Table:

Gas Molar Mean thermalMass M velocity (m/s)

————————————–————H2 2 1761He 4 1245

H2O 18 587Ne 20 557CO 28 471N2 28 471Air 28.96 463O2 32 440Ar 40 394

CO2 44 375CC13F (F11) 134.78 68Table 2.4 c- as a function of molar mass M

c R TM= · ·

·8π



especially recommended for pumpinglaboratory distillation apparatus andsimilar plants when the pressure from asimple water jet pump is insufficient. Inthis instance, the use of rotary pumpswould not be economical.

Even in spite of their low investment costswater jet pumps and steam ejectors arebeing replaced in the laboratories moreand more by diaphragm pumps because ofthe environmental problems of using wateras the pump fluid. Solvent entering thewater can only be removed again throughcomplex cleaning methods (distillation).

Whereas the dependence of the pumpingspeed on the type of gas is fairly low (S ∼ c

_∼ 1 / EE

__M ), the dependence of the

compression k0 at zero throughput andthus also the compression k, because of k0 ∼ eEM log k0 ∼ E

__M , is greater as

shown by the experimentally-determinedrelationship in Fig. 2.55.Example:

from theory it follows that

this with k0 (He) = 3 · 103 from Fig. 2.55results in:

log k0 (N2) = 2.65 · log (3 · 103) = 9.21

or k0 (N2) = 1.6 · 109.

This agrees – as expected – well (order ofmagnitude) with the experimentallydetermined value for k0 (N2) = 2.0 · 108

from Fig. 2.55. In view of the optimizationsfor the individual rotor stages commontoday, this consideration is no longercorrect for the entire pump. Shown in Fig.2.56 are the values as measured for amodern TURBOVAC 340 M.In order to meet the condition, a circumfe-rential velocity for the rotor of the sameorder of magnitude as c- high rotor speedsare required for turbomolecular pumps.They range from about 36,000 rpm forpumps having a large diameter rotor(TURBOVAC 1000) to 72,000 rpm in thecase of smaller rotor diameters(TURBOVAC 35 / 55). Such high speedsnaturally raise questions as to a reliablebearing concept. LEYBOLD offers threeconcepts, the advantages and disadvan-tages of which are detailed in the follo-wing:• Oil lubrication / steel ball bearings

+ Good compatibility with particles bycirculating oil lubricant

- Can only be installed vertically+ Low maintenance

• Grease lubrication / hybrid bearings+ Installation in any orientation+ Suited for mobile systems± Air cooling will do for many applica-

tions+ Lubricated for life (of the bearings)

log ( )log ( ) .

k Hek N

0

0 2

428

17

12 65= = =

log ( ) . log ( )k N k He0 02 652⇒ = ·

D00

Fig. 2.51 Schematic representation of the operationof a steam ejector pump

1 Steam inlet 2 Jet nozzle3 Diffuser4 Mixing region5 Connection to the vacuum chamber

D00 E 19.06.2001 21:37 Uhr Seite 41



• Free of lubricants / magnetic suspen-sion+ No wear+ No maintenance+ Absolutely free of hydrocarbons+ Low noise and vibration levels+ Installation in any orientation

Steel ball bearings / hybrid ball bearings(ceramic ball bearings): Even a brief tearin the thin lubricating film between theballs and the races can – if the same typeof material is used – result in microweldingat the points of contact. This severely re-duces the service life of the bearings. Byusing dissimilar materials in so called hy-brid bearings (races: steel, balls: ceramics)the effect of microwelding is avoided.

The most elegant bearing concept is thatof the magnetic suspension. As early as1976 LEYBOLD delivered magneticallysuspended turbomolecular pumps – thelegendary series 550M and 560M. At thattime a purely active magnetic suspension(i.e. with electromagnets) was used. Ad-vances in electronics and the use of per-manent magnets (passive magneticsuspension) based on the “System KFA

Jülich” permitted the magnetic suspen-sion concept to spread widely. In thissystem the rotor is maintained in a stableposition without contact during operation,by magnetic forces. Absolutely no lubri-cants are required. So-called touch downbearings are integrated for shutdown.

Fig. 2.52 shows a sectional drawing of atypical turbomolecular pump. The pump isan axial flow compressor of verticaldesign, the active or pumping part ofwhich consists of a rotor (6) and a stator(2). Turbine blades are located around thecircumferences of the stator and the rotor.Each rotor – stator pair of circular bladerows forms one stage, so that theassembly is composed of a multitude ofstages mounted in series. The gas to bepumped arrives directly through theaperture of the inlet flange (1), that is,without any loss of conductance, at theactive pumping area of the top blades ofthe rotor – stator assembly. This is equip-ped with blades of especially large radialspan to allow a large annular inlet area.The gas captured by these stages istransferred to the lower compression

stages, whose blades have shorter radialspans, where the gas is compressed tobacking pressure or rough vacuum pres-sure. The turbine rotor (6) is mounted onthe drive shaft, which is supported by twoprecision ball bearings (8 and 11), accom-modated in the motor housing. The rotorshaft is directly driven by a medium-fre-quency motor housed in the forevacuumspace within the rotor, so that no rotaryshaft lead-through to the outside atmos-phere is necessary. This motor is poweredand automatically controlled by an externalfrequency converter, normally a solid-statefrequency converter that ensures a verylow noise level. For special applications,for example, in areas exposed to radiation,motor generator frequency converters areused.

The vertical rotor – stator configurationprovides optimum flow conditions of thegas at the inlet.

To ensure vibration-free running at highrotational speeds, the turbine is dyna-mically balanced at two levels during itsassembly.

Fig. 2.52 Schematic diagram of a grease lubricated TURBOVAC 151turbomolecular pump

7 Pump casing8 Ball bearings9 Cooling water connection10 3-phase motor11 Ball bearings

1 High vacuum inlet flange2 Stator pack3 Venting flange4 Forevacuum flange5 Splinter guard6 Rotor

Fig. 2.52a Cross section of a HY.CONE turbomolecular pump

Turbomole-cularpump stage

Siegbahnstage

1

2

3

4

5

6

8

7

5 Bearing6 Motor7 Fan8 Bearing

1 Vacuum port 2 High vacuum flange3 Rotor4 Stator

D00 E 19.06.2001 21:37 Uhr Seite 42



The pumping speed (volume flow rate)characteristics of turbomolecular pumpsare shown in Fig. 2.53. The pumpingspeed remains constant over the entireworking pressure range. It decreases atintake pressures above 10-3 mbar, as thisthreshold value marks the transition fromthe region of molecular flow to the regionof laminar viscous flow of gases. Fig. 2.54shows also that the pumping speeddepends on the type of gas.

The compression ratio (often also simplytermed compression) of turbomolecularpumps is the ratio between the partialpressure of one gas component at the

forevacuum flange of the pump and that atthe high vacuum flange: maximum com-pression k0 is to be found at zerothroughput. For physical reasons, thecompression ratio of turbomolecularpumps is very high for heavy moleculesbut considerably lower for light molecules.The relationship between compressionand molecular mass is shown in Fig. 2.55.Shown in Fig. 2.56 are the compressioncurves of a TURBOVAC 340 M for N2, Heand H2 as a function of the backing pres-sure. Because of the high compressionratio for heavy hydrocarbon molecules,turbomolecular pumps can be directlyconnected to a vacuum chamber without

the aid of one or more cooled baffles ortraps and without the risk of a measurablepartial pressure for hydrocarbons in thevacuum chamber (hydrocarbon-freevacuum! – see also Fig. 2.57: residual gasspectrum above a TURBOVAC 361). As thehydrogen partial pressure attained by therotary backing pump is very low, theturbomolecular pump is capable ofattaining ultimate pressures in the 10-11

mbar range in spite of its rather moderatecompression for H2. To produce such ex-tremely low pressures, it will, of course, benecessary to strictly observe the generalrules of UHV technology: the vacuumchamber and the upper part of the turbo-

D00

10ñ62 4 6 8

10ñ5 10ñ4 p 10ñ3 10ñ2 10ñ1

10000

1000

500

200

100

10

S

mbar

340M

50/55

361

600

151

1000/1000 MC

Fig. 2.53 Pumping speed for air of different turbomolecular pumps

l · s–1

Fig. 2.54 Pumping speed curves of a TURBOVAC 600 for H2, He, N2 and Ar

l · s–1

Fig. 2.55 TURBOVAC 450 – Maximum compressionk0 as a function of molar mass M

k0

√MFig. 2.56 Maximum compression k0 of a turbomolecular

pump TURBOVAC 340 M for H2, He and N2 asa function of backing pressure Fig. 2.57 Spectrum above a TURBOVAC 361

M = Mass number = Relative molar mass at an ionization 1

I = Ion current

D00 E 19.06.2001 21:37 Uhr Seite 43



molecular pump must be baked out, andmetal seals must be used. At very lowpressures the residual gas is composedmainly of H2 originating from the metalwalls of the chamber. The spectrum in Fig.2.57 shows the residual gas compositionin front of the inlet of a turbomolecularpump at an ultimate pressure of 7 · 10-10

mbar nitrogen equivalent. It appears thatthe portion of H2 in the total quantity ofgas amounts to approximately 90 to 95 %.The fraction of “heavier” molecules is con-siderably reduced and masses greater than44 were not detected. An important crite-rion in the assessment of the quality of aresidual gas spectrum are the measurablehydrocarbons from the lubricants used inthe vacuum pump system. Of course an“absolutely hydrocarbon-free vacuum”can only be produced with pump systemswhich are free of lubricants, i.e. for exam-ple with magnetically-suspended turbo-molecular pumps and dry compressingbacking pumps. When operated correctly(venting at any kind of standstill) no hy-drocarbons are detectable also in the spec-trum of normal turbomolecular pumps.A further development of the turbomo-lecular pump is the hybrid or compoundturbomolecular pump. This is actually twopumps on a common shaft in a singlecasing. The high vacuum stage for the mo-lecular flow region is a classic turbo-molecular pump, the second pump for theviscous flow range is a molecular drag orfriction pump.

LEYBOLD manufactures pumps such asthe TURBOVAC 55 with an integratedHolweck stage (screw-type compressor)and, for example, the HY.CONE 60 orHY.CONE 200 with an integrated Siegbahnstage (spiral compressor). The requiredbacking pressure then amounts to a fewmbar so that the backing pump is onlyrequired to compress from about 5 to 10mbar to atmospheric pressure. A sectionalview of a HY.CONE is shown in Fig. 2.52a.

Information on the operation of turbomo-lecular pumps

StartingAs a rule turbomolecular pumps shouldgenerally be started together with thebacking pump in order to reduce anybackstreaming of oil from the backingpump into the vacuum chamber. A delayedstart of the turbomolecular pump, makessense in the case of rather small backing

pump sets and large vacuum chambers. Ata known pumping speed for the backingpump SV (m3/h) and a known volume forthe vacuum chamber (m3) it is possible toestimate the cut-in pressure for theturbomolecular pump:Simultaneous start when

and delayed start when

at a cut-in pressure of:

(2.24)

When evacuating larger volumes the cut-inpressure for turbomolecular pumps mayalso be determined with the aid of thediagram of Fig. 2.58.

VentingAfter switching off or in the event of apower failure, turbomolecular pumpsshould always be vented in order toprevent any backdiffusion of hydrocarbonsfrom the forevacuum side into the vacuumchamber. After switching off the pump thecooling water supply should also beswitched off to prevent the possiblecondensation of water vapor. In order toprotect the rotor it is recommended tocomply with the (minimum) venting timesstated in the operating instructions. The

p e mbarV Start

SVV

,·=

6

SV hv < −40 1

SV hv > −40 1

pump should be vented (except in the caseof operation with a barrier gas) via theventing flange which already contains asintered metal throttle, so that venting maybe performed using a normal valve or apower failure venting valve.

Barrier gas operationIn the case of pumps equipped with abarrier gas facility, inert gas – such as drynitrogen – may be applied through aspecial flange so as to protect the motorspace and the bearings against aggressivemedia. A special barrier gas and ventingvalve meters the necessary quantity ofbarrier gas and may also serve as a ven-ting valve.

Decoupling of vibrationsTURBOVAC pumps are precisely balancedand may generally be connected directly tothe apparatus. Only in the case of highlysensitive instruments, such as electronmicroscopes, is it recommended to installvibration absorbers which reduce thepresent vibrations to a minimum. For mag-netically suspended pumps a directconnection to the vacuum apparatus willusually do because of the extremely lowvibrations produced by such pumps.

For special applications such as opera-tion in strong magnetic fields, radiationhazard areas or in a tritium atmosphere,please contact our Technical Sales Depart-ment which has the necessary experienceand which is available to you at any time.

Fig. 2.58 Determination of the cut-in pressure forturbomolecular pumps when evacuatinglarge vessels

D00 E 19.06.2001 21:37 Uhr Seite 44



2.1.8 Sorption pumps

The term “sorption pumps” includes allarrangements for the removal of gases andvapors from a space by sorption means.The pumped gas particles are therebybound at the surfaces or in the interior ofthese agents, by either physical tempera-ture-dependent adsorption forces (van derWaals forces), chemisorption, absorption,or by becoming embedded during thecourse of the continuous formation of newsorbing surfaces. By comparing their ope-rating principles, we can distinguishbetween adsorption pumps, in which thesorption of gases takes place simply bytemperature-controlled adsorption pro-cesses, and getter pumps, in which thesorption and retention of gases areessentially caused by the formation ofchemical compounds. Gettering is thebonding of gases to pure, mostly metallicsurfaces, which are not covered by oxideor carbide layers. Such surfaces alwaysform during manufacture, installation orwhile venting the system. The mostlymetallic highest purity getter surfaces arecontinuously generated either directly inthe vacuum by evaporation (evaporatorpumps) or by sputtering (sputter pumps)or the passivating surface layer of thegetter (metal) is removed by degassing thevacuum, so that the pure material isexposed to the vacuum. This step is calledactivation (NEG pumps NEG = Non Eva-porable Getter).

2.1.8.1 Adsorption pumps

Adsorption pumps (see Fig. 2.59) workaccording to the principle of the physicaladsorption of gases at the surface ofmolecular sieves or other adsorptionmaterials (e.g. activated Al2O3). Zeolite13X is frequently used as an adsorptionmaterial. This alkali aluminosilicatepossesses for a mass of the material anextraordinarily large surface area, about1000 m2/g of solid substance. Corres-pondingly, its ability to take up gas is con-siderable.

The pore diameter of zeolite 13X is about13 Å, which is within the order of size ofwater vapor, oil vapor, and larger gasmolecules (about 10 Å). Assuming that themean molecular diameter is half this value,5 · 10-8 cm, about 5 · 1018 molecules areadsorbed in a monolayer on a surface of

1 m2. For nitrogen molecules with arelative molecular mass Mr = 28 thatcorresponds to about 2 · 10-4 g or 0.20mbar · l (see also Section 1.1). Therefore,an adsorption surface of 1000 m2 iscapable of adsorbing a monomolecularlayer in which more than 133 mbar · l ofgas is bound. Hydrogen and light noble gases, such ashelium and neon, have a relatively small

particle diameter compared with the poresize of 13 Å for zeolite 13X. These gasesare, therefore, very poorly adsorbed.

The adsorption of gases at surfaces isdependent not only on the temperature,but more importantly on the pressureabove the adsorption surface. Thedependence is represented graphically fora few gases by the adsorption isothermsgiven in Fig. 2.60. In practice, adsorptionpumps are connected through a valve tothe vessel to be evacuated. It is onimmersing the body of the pump in liquidnitrogen that the sorption effect is madetechnically useful. Because of the differentadsorption properties, the pumping speedand ultimate pressure of an adsorptionpump are different for the various gasmolecules: the best values are achieved fornitrogen, carbon dioxide, water vapor, andhydrocarbon vapors. Light noble gases arehardly pumped at all because the diameterof the particles is small compared to thepores of the zeolite. As the sorption effectdecreases with increased coverage of thezeolite surfaces, the pumping speed fallsoff with an increasing number of theparticles already adsorbed. The pumpingspeed of an adsorption pump is, therefore,dependent on the quantity of gas alreadypumped and so is not constant with time.

The ultimate pressure attainable withadsorption pumps is determined in the

D00

Fig. 2.59 Cross section of an adsorption pump

5 Thermal conductingvanes

6 Adsorption material(e.g. Zeolith)

1 Inlet port2 Degassing port3 Support4 Pump body

Fig. 2.60 Adsorption isotherms of zeolite 13X for nitrogen at -195 °C and 20 °C, as well as for helium andneon at -195 °C

Pressure [Torr]

Adso

rbed

qua

ntity

of g

as p

er q

uant

ity o

f ads

orbe

nt [m

bar ·

l · g

–1]

Pressure [mbar]

D00 E 19.06.2001 21:37 Uhr Seite 45



first instance by those gases that prevail inthe vessel at the beginning of the pumpingprocess and are poorly or not at alladsorbed (e.g. neon or helium) at thezeolite surface. In atmospheric air, a fewparts per million of these gases arepresent. Therefore, pressures < 10-2 mbarcan be obtained.

If pressures below 10-3 mbar exclusivelyare to be produced with adsorption pumps,as far as possible no neon or heliumshould be present in the gas mixture.After a pumping process, the pump mustbe warmed only to room temperature forthe adsorbed gases to be given off and thezeolite is regenerated for reuse. If air (ordamp gas) containing a great deal of watervapor has been pumped, it is recom-mended to bake out the pump completelydry for a few hours at 200 °C or above.

To pump out larger vessels, severaladsorption pumps are used in parallel or inseries. First, the pressure is reduced fromatmospheric pressure to a few millibars bythe first stage in order to “capture” manynoble gas molecules of helium and neon.After the pumps of this stage have beensaturated, the valves to these pumps areclosed and a previously closed valve to afurther adsorption pump still containingclean adsorbent is opened so that thispump may pump down the vacuumchamber to the next lower pressure level.This procedure can be continued until theultimate pressure cannot be furtherimproved by adding further clean adsorp-tion pumps.

2.1.8.2 Sublimation pumps

Sublimation pumps are sorption pumps inwhich a getter material is evaporated anddeposited on a cold inner wall as a getterfilm. On the surface of such a getter filmthe gas molecules form stable com-pounds, which have an immeasurably lowvapor pressure. The active getter film isrenewed by subsequent evaporations.Generally titanium is used in sublimationpumps as the getter. The titanium isevaporated from a wire made of a specialalloy of a high titanium content which isheated by an electric current. Although theoptimum sorption capacity (about onenitrogen atom for each evaporated tita-nium atom) can scarcely be obtained inpractice, titanium sublimation pumps have

an extraordinarily high pumping speed foractive gases, which, particularly on star-ting processes or on the sudden evolutionof greater quantities of gas, can be rapidlypumped away. As sublimation pumpsfunction as auxiliary pumps (boosters) tosputter-ion pumps and turbomolecularpumps, their installation is often indi-spensable (like the “boosters” in vaporejector pumps; see Section 2.1.6.2).

2.1.8.3 Sputter-ion pumps

The pumping action of sputter-ion pumpsis based on sorption processes that areinitiated by ionized gas particles in a Pen-ning discharge (cold cathode discharge).By means of “paralleling many individualPenning cells” the sputter ion pumpattains a sufficiently high pumping speedfor the individual gases.

Operation of sputter-ion pumpsThe ions impinge upon the cathode of thecold cathode discharge electrode systemand sputter the cathode material (titanium).The titanium deposited at other locationsacts as a getter film and adsorbs reactivegas particles (e.g., nitrogen, oxygen, hydro-gen). The energy of the ionized gas par-ticles is not only high enough to sputter thecathode material but also to let theimpinging ions penetrate deeply into thecathode material (ion implantation). Thissorption process “pumps” ions of all types,including ions of gases which do notchemically react with the sputtered titaniumfilm, i.e. mainly noble gases.

The following arrangement is used toproduce the ions: stainless-steel, cylindri-cal anodes are closely arranged between,with their axes perpendicular to, twoparallel cathodes (see Fig. 2.61). Thecathodes are at negative potential (a fewkilovolts) against the anode. The entireelectrode system is maintained in a strong,homogeneous magnetic field of a fluxdensity of B = 0.1 T, (T = Tesla = 104

Gauss) produced by a permanent magnetattached to the outside of the pump’scasing. The gas discharge profduced bythe high tension contains electrons andions. Under the influence of the magneticfield the electrons travel along long spiraltracks (see Fig. 2.61) until they impinge onthe anode cylinder of the correspondingcell. The long track increases ion yield,which even at low gas densities (pres-

sures) is sufficient to maintain a self-sustained gas discharge. A supply of elec-trons from a hot cathode is not required.Because of their great mass, themovement of the ions is unaffected by themagnetic field of the given order of mag-nitude; they flow off along the shortestpath and bombard the cathode.

The discharge current i is proportional tothe number density of neutral particles n0,the electron density n-, and the length l ofthe total discharge path:

i = n0 · n– · σ · l (2.25)

The effective cross section s for ionizingcollisions depends on the type of gas.According to (2.25), the discharge currenti is a function of the number particledensity n0, as in a Penning gauge, and itcan be used as a measure of the pressurein the range from 10-4 to 10-8 mbar. Atlower pressures the measurements are notreproducible due to interferences fromfield emission effects.In diode-type, sputter-ion pumps, with anelectrode system configuration as shownin Fig. 2.62, the getter films are formed onthe anode surfaces and between thesputtering regions of the oppositecathode. The ions are buried in thecathode surfaces. As cathode sputteringproceeds, the buried gas particles are setfree again. Therefore, the pumping actionfor noble gases that can be pumped onlyby ion burial will vanish after some time

Fig. 2.61 Operating principle of a sputter-ion pump

PZ

← ⊕ Direction of motion of the ionized gasmolecules

• → Direction of motion of the sputteredtitanium

- – – - Spiral tracks of the electronsPZ Penning cells

D00 E 19.06.2001 21:37 Uhr Seite 46



and a “memory effect” will occur.

Unlike diode-type pumps, triode sputter-ion pumps exhibit excellent stability intheir pumping speed for noble gasesbecause sputtering and film formingsurfaces are separated. Fig. 2.63 showsthe electrode configuration of triodesputter-ion pumps. Their greater efficiencyfor pumping noble gases is explained asfollows: the geometry of the system favorsgrazing incidence of the ions on thetitanium bars of the cathode grid, wherebythe sputtering rate is considerably higherthan with perpendicular incidence. Thesputtered titanium moves in about thesame direction as the incident ions. Thegetter films form preferentially on the thirdelectrode, the target plate, which is theactual wall of the pump housing. There isan increasing yield of ionized particles thatare grazingly incident on the cathode gridwhere they are neutralized and reflected

and from which they travel to the targetplate at an energy still considerably higherthan the thermal energy 1/2 · k · T of thegas particles. The energetic neutralparticles can penetrate into the targetsurface layer, but their sputtering effect isonly negligible. These buried or implantedparticles are finally covered by freshtitanium layers. As the target is at positivepotential, any positive ions arriving thereare repelled and cannot sputter the targetlayers. Hence the buried noble gas atomsare not set free again. The pumping speedof triode sputter-ion pumps for noblegases does not decrease during theoperation of the pump.

The pumping speed of sputter-ion pumpsdepends on the pressure and the type ofgas. It is measured according to themethods stated in DIN 28 429 andPNEUROP 5615. The pumping speedcurve S(p) has a maximum. The nominal

pumping speed Sn is given by the maxi-mum of the pumping speed curve for airwhereby the corresponding pressure mustbe stated.For air, nitrogen, carbon dioxide and watervapor, the pumping speed is practically thesame. Compared with the pumping speedfor air, the pumping speeds of sputter-ionpumps for other gases amount toapproximately:Hydrogen 150 to 200 %Methane 100 %Other light hydrocarbons 80 to 120 % Oxygen 80 % Argon 30 % Helium 28 %

Sputter-ion pumps of the triode type excelin contrast to the diode-type pumps inhigh-noble gas stability. Argon is pumpedstably even at an inlet pressure of 1 · 10-5 mbar. The pumps can be startedwithout difficulties at pressures higher than1 · 10-2 mbar and can operate continuouslyat an air inlet producing a constant airpressure of 5 · 10-5 mbar. A new kind ofdesign for the electrodes extends theservice life of the cathodes by 50 %.

Influence on processes in the vacuumchamber by magnetic stray fields andstray ions from the sputter-ion pump.The high-magnetic-field strength requiredfor the pumping action leads inevitably tostray magnetic fields in the neighborhoodof the magnets. As a result, processes inthe vacuum chamber can be disturbed insome cases, so the sputter-ion pumpconcerned should be provided with ascreening arrangement. The forms andkinds of such a screening arrangement canbe regarded as at an optimum if theprocesses taking place in the vacuumchamber are disturbed by no more thanthe earth’s magnetic field which is presentin any case.

Fig. 2.64 shows the magnetic stray field atthe plane of the intake flange of a sputter-ion pump IZ 270 and also at a parallelplane 150 mm above. If stray ions from thedischarge region are to be prevented fromreaching the vacuum chamber, a suitablescreen can be set up by a metal sieve atopposite potential in the inlet opening ofthe sputter-ion pump (ion barrier). This,however, reduces the pumping speed ofthe sputter-ion pump depending on themesh size of the selected metal sieve.

D00

Fig. 2.62 Electrode configuration in a diode sputter-ion pump

• Titanium atomsÖ Gas molecules⊕ Ionsü ElectronsB Magnetic field

Fig. 2.63 Electrode configuration in a triode sputter-ion pump

• Titanium atomsÖ Gas molecules⊕ Ionsü ElectronsA Anode cylinder

(same as in the diodepump)

B Magnetic fieldF Target plate

(pump housing)as the third electrode

K Cathode grid

D00 E 19.06.2001 21:37 Uhr Seite 47



2.1.8.4 Non evaporable getter pumps(NEG Pumps)

The non evaporable getter pump operateswith a non evaporable, compact gettermaterial, the structure of which is porousat the atomic level so that it can take uplarge quantities of gas. The gas moleculesadsorbed on the surface of the gettermaterial diffuse rapidly inside the materialthereby making place for further gasmolecules impinging on the surface. Thenon evaporable getter pump contains aheating element which is used to heat thegetter material to an optimum temperaturedepending on the type of gas which ispreferably to be pumped. At a highertemperature the getter material which hasbeen saturated with the gas is regenerated(activated). As the getter material, mostlyzirconium-aluminum alloys are used in theform of strips. The special properties ofNEG pumps are:• constant pumping speed in the HV and

UHV ranges

• no pressure restrictions up to about 12mbar

• particularly high pumping speed forhydrogen and its isotopes

• after activation the pump can oftenoperate at room temperature and will

then need no electrical energy

• no interference by magnetic fields

• hydrocarbon-free vacuum

• free of vibrations

• low weight

NEG pumps are mostly used incombination with other UHV pumps (tur-bomolecular and cryopumps). Such com-binations are especially useful whenwanting to further reduce the ultimatepressure of UHV systems, since hydrogencontributes mostly to the ultimate pres-sure in an UHV system, and for which NEGpumps have a particularly high pumpingspeed, whereas the pumping effect for H2of other pumps is low. Some typicalexamples for applications in which NEGpumps are used are particle acceleratorsand similar research systems, surfaceanalysis instruments, SEM columns andsputtering systems. NEG pumps aremanufactured offering pumping speeds ofseveral l/s to about 1000 l/s. Custompumps are capable of attaining a pumpingspeed for hydrogen which is by severalorders of magnitude higher.

Fig. 2.64 Stray magnetic field of a sputter-ion pumpin two places parallel to the inlet flange(inserts) curves show lines of constantmagnetic induction B in Gauss.1 Gauss = 1 · 10–4 Tesla

D00 E 19.06.2001 21:37 Uhr Seite 48



2.1.9 Cryopumps

As you may have observed watercondenses on cold water mains orwindows and ice forms on the evaporatorunit in your refrigerator. This effect ofcondensation of gases and vapors on coldsurfaces, water vapor in particular, as it isknown in every day life, occurs not only atatmospheric pressure but also in vacuum.

This effect has been utilized for a long timein condensers (see 2.1.5) mainly inconnection with chemical processes;previously the baffle on diffusion pumpsused to be cooled with refrigeratingmachines. Also in a sealed space (vacuumchamber) the formation of condensate ona cold surface means that a large numberof gas molecules are removed from thevolume: they remain located on the coldsurface and do not take part any longer inthe hectic gas atmosphere within thevacuum chamber. We then say that theparticles have been pumped and talk ofcryopumps when the “pumping effect” isattained by means of cold surfaces.

Cryo engineering differs from refrigeration

engineering in that the temperaturesinvolved in cryo engineering are in therange below 120 K (< -153 °C). Here weare dealing with two questions:a) What cooling principle is used in cryo

engineering or in cryopumps and howis the thermal load of the cold surfacelead away or reduced?

b) What are the operating principles of thecryopumps?

2.1.9.1 Types of cryopump

Depending on the cooling principle adifference is made between• Bath cryostats• Continuous flow cryopumps• Refrigerator cryopumps

In the case of bath cryostats – in the mostsimple case a cold trap filled with LN2 (li-quid nitrogen) – the pumping surface iscooled by direct contact with a liquefiedgas. On a surface cooled with LN2

(T ≈ 77 K) H2O and CO2 are able tocondense. On a surface cooled to ≈ 10 Kall gases except He and Ne may be

pumped by way of condensation. Asurface cooled with liquid helium (T ≈ 4.2 K) is capable of condensing allgases except helium.In continuous flow cryopumps the coldsurface is designed to operate as a heatexchanger. Liquid helium in sufficientquantity is pumped by an auxiliary pumpfrom a reservoir into the evaporator in orderto attain a sufficiently low temperature atthe cold surface (cryopanel).

The liquid helium evaporates in the heatexchanger and thus cools down thecryopanel. The waste gas which isgenerated (He) is used in a second heatexchanger to cool the baffle of a thermalradiation shield which protects the systemfrom thermal radiation coming from theoutside. The cold helium exhaust gasejected by the helium pump is supplied toa helium recovery unit. The temperature atthe cryopanels can be controlled bycontrolling the helium flow.

Today refrigerator cryopumps are beingused almost exclusively (cold upondemand). These pumps operate basicallymuch in the same way as a commonhousehold refrigerator, whereby thefollowing thermodynamic cycles usinghelium as the refrigerant may be employed:• Gifford-McMahon process• Stirling process• Brayton process• Claude process

The Gifford-McMahon process is mostlyused today and this process is that whichhas been developed furthest. It offers thepossibility of separating the locations forthe large compressor unit and theexpansion unit in which the refrigerationprocess takes place. Thus a compact andlow vibration cold source can be designed.The cryopumps series-manufactured byLEYBOLD operate with two-stage coldheads according to the Gifford-McMahonprocess which is discussed in detail in thefollowing.

The entire scope of a refrigerator cryopumpis shown in Fig. 2.65 and consists of thecompressor unit (1) which is linked viaflexible pressure lines (2) – and thusvibration-free – to the cryopump (3). Thecryopump itself consists of the pumpcasing and the cold head within. Helium isused as the refrigerant which circulates in aclosed cycle with the aid of the compressor.

D00Fig. 2.65 All items of a refrigerator cryopump

1 Compressor unit2 Flexible pressure

lines

3 Cold head (withoutcondensation surfaces)

3

1

2

D00 E 19.06.2001 21:37 Uhr Seite 49



2.1.9.2 The cold head and its opera-ting principle (Fig. 2.66)

Within the cold head, a cylinder is dividedinto two working spaces V1 and V2 by adisplacer. During operation the right spaceV1 is warm and the left space V2 is cold. Ata displacer frequency f the refrigeratingpower W of the refrigerator is:

W = (V2,max – V2,min) · (pH – pN) · f (2.26)

The displacer is moved to and fro pneuma-tically so that the gas is forced through thedisplacer and thus through the regenerator

located inside the displacer. The regenera-tor is a heat accumulator having a largeheat exchanging surface and capacity,which operates as a heat exchanger withinthe cycle. Outlined in Fig. 2.66 are the fourphases of refrigeration in a single-stagerefrigerator cold head operating accordingto the Gifford-McMahon principle.

The two-stage cold headThe series-manufactured refrigeratorcryopumps from LEYBOLD use a two-stage cold head operating according to theGifford-McMahon principle (see Fig. 2.67).

In two series connected stages thetemperature of the helium is reduced toabout 30 K in the first stage and further toabout 10 K in the second stage. Theattainable low temperatures dependamong other things on the type ofregenerator. Commonly copperbronze isused in the regenerator of the first stageand lead in the second stage. Othermaterials are available as regenerators forspecial applications like cryostats forextremely low temperatures (T < 10 K).The design of a two-stage cold head isshown schematically in Fig. 2.67. By

Fig. 2.66 Refrigerating phases using a single-stage cold head operating according to the Gifford-McMahon process

Phase 1:

The displacer is at the left dead center;V2 where the cold is produced has itsminimum size. Valve N remains closed,H is opened. Gas at the pressure pHflows through the regenerator into V2.There the gas warms up by the pressureincrease in V1.

Phase 2:

Valve H remains open, valve N closed:the displacer moves to the right andejects the gas from V1 through the rege-nerator to V2 where it cools down at thecold regenerator.; V2 has its maximumvolume.

Phase 3:

Valve H is closed and the valve N to thelow pressure reservoir is opened. Thegas expands from pH to pN and therebycools down. This removes heat from thevicinity and it is transported with theexpanding gas to the compressor.

Phase 4:

With valve N open the displacer movesto the left; the gas from V2,max flowsthrough the regenerator, cooling it downand then flows into the volume V1 andinto the low pressure reservoir. Thiscompletes the cycle.

V2 (cold) V1 (warm)

Regenerator Displacer

V2 (cold) V1 (warm)


V2 (cold) V1 (warm)


V2 (cold) V1


D00 E 19.06.2001 21:37 Uhr Seite 50



means of a control mechanism with amotor driven control valve (18) withcontrol disk (17) and control holes first thepressure in the control volume (16) ischanged which causes the displacers (6)of the first stage and the second stage (11)to move; immediately thereafter thepressure in the entire volume of thecylinder is equalized by the controlmechanism. The cold head is linked viaflexible pressure lines to the compressor.

2.1.9.3 The refrigerator cryopump

Fig. 2.68 shows the design of a cryopump.It is cooled by a two-stage cold head. Thethermal radiation shield (5) with the baffle(6) is closely linked thermally to the firststage (9) of the cold head. For pressuresbelow 10-3 mbar the thermal load iscaused mostly by thermal radiation. Forthis reason the second stage (7) with thecondensation and cryosorption panels (8)

is surrounded by the thermal radiationshield (5) which is black on the inside andpolished as well as nickel plated on theoutside. Under no-load conditions thebaffle and the thermal radiation shield(first stage) attain a temperature rangingbetween 50 to 80 K at the cryopanels andabout 10 K at the second stage. Thesurface temperatures of these cryopanelsare decisive to the actual pumpingprocess. These surface temperaturesdepend on the refrigerating powersupplied by the cold head, and the thermalconduction properties in the direction ofthe pump’s casing. During operation of thecryopump, loading caused by the gas andthe heat of condensation results in furtherwarming of the cryopanels. The surfacetemperature does not only depend on thetemperature of the cryopanel, but also onthe temperature of the gas which hasalready been frozen on to the cryopanel.The cryopanels (8) attached to the second

stage (7) of the cold head are coated withactivated charcoal on the inside in order tobe able to pump gases which do not easilycondense and which can only be pumpedby cryosorption (see 2.1.9.4).

2.1.9.4 Bonding of gases to coldsurfaces

The thermal conductivity of the condensed(solid) gases depends very much on theirstructure and thus on the way in which thecondensate is produced. Variations inthermal conductivity over several orders ofmagnitude are possible! As the condensa-te increases in thickness, thermal resis-tance and thus the surface temperatureincrease subsequently reducing the pum-ping speed. The maximum pumping speedof a newly regenerated pump is stated asits nominal pumping speed. The bondingprocess for the various gases in thecryopump is performed in three steps: first

D00

Fig. 2.67 Two-stage cold head

1 Electric connections and currentfeedthrough for the motor in thecold head

2 He high pressure connection 3 He low pressure connection4 Cylinder, 1st stage 5 Displacer, 1st stage 6 Regenerator, 1st stage7 Expansion volume, 1st stage8 1st (cooling) stage (copper

flange)9 Cylinder, 2nd stage10 Displacer, 2nd stage

11 Regenerator, 2nd stage 12 Expansion volume, 2nd stage13 2nd (cooling) stage

(copper flange) 14 Measurement chamber for the

vapor pressure15 Control piston16 Control volume17 Control disk18 Control valve19 Gauge for the hydrogen vapor

pressure thermometer20 Motor in the cold head

Fig. 2.68 Design of a refrigerator cryopump (schematic)

1 High vacuum flange2 Pump casing 3 Forevacuum flange4 Safety valve for gas discharge5 Thermal radiation shield6 Baffle7 2nd stage of the cold head

(≈10 K); 8 Cryopanels

9 1st stage of the cold head (≈ 50 – 80 K)

10 Gauge for the hydrogen vaporpressure thermometer

11 Helium gas connections12 Motor of the cold head

with casing and electricconnections

D00 E 19.06.2001 21:37 Uhr Seite 51



the mixture of different gases and vaporsmeets the baffle which is at a temperatureof about 80 K. Here mostly H2O and CO2are condensed. The remaining gasespenetrate the baffle and impinge in theoutside of the cryopanel of the secondstage which is cooled to about 10 K. Heregases like N2, O2 or Ar will condense. OnlyH2, He and Ne will remain. These gasescan not be pumped by the cryopanels andthese pass after several impacts with thethermal radiation shield to the inside ofthese panels which are coated with anadsorbent (cryosorption panels) wherethey are bonded by cryosorption. Thus forthe purpose of considering a cryopumpthe gases are divided into three groupsdepending at which temperatures withinthe cryopump their partial pressure dropsbelow 10-9 mbar:

1st group:ps < 10-9 mbar at T ≈ 77K (LN2): H2O, CO2

2nd group:ps < 10-9 mbar at T ≈ 20K: N2, O2, Ar

3rd group:ps < 10-9 mbar at T < 4.2K: H2, He, Ne

A difference is made between the differentbonding mechanisms as follows:

Cryocondensation is the physical and re-versible bonding of gas molecules throughVan der Waals forces on sufficiently coldsurfaces of the same material. The bondenergy is equal to the energy of vaporiza-tion of the solid gas bonded to the surfaceand thus decreases as the thickness of thecondensate increases as does the vapor

pressure. Cryosorption is the physical andreversible bonding of gas moleculesthrough Van der Waals forces on suffici-ently cold surfaces of other materials. Thebond energy is equal to the heat ofadsorption which is greater than the heatof vaporization. As soon as a monolayerhas been formed, the following moleculesimpinge on a surface of the same kind(sorbent) and the process transforms intocryocondensation. The higher bond energyfor cryocondensation prevents the furthergrowth of the condensate layer therebyrestricting the capacity for the adsorbedgases. However, the adsorbents used, likeactivated charcoal, silica gel, alumina geland molecular sieve, have a porousstructure with very large specific surfaceareas of about 106 m2/kg. Cryotrapping isunderstood as the inclusion of a lowboiling point gas which is difficult to pumpsuch as hydrogen, in the matrix of a gashaving a higher boiling point and whichcan be pumped easily such as Ar, CH4 orCO2. At the same temperature thecondensate mixture has a saturation vaporpressure which is by several orders ofmagnitude lower than the pure condensateof the gas with the lower boiling point.

2.1.9.5 Pumping speed and position ofthe cryopanels

Considering the position of the cryopanelsin the cryopump, the conductance fromthe vacuum flange to this surface and alsothe subtractive pumping sequence (whathas already condensed at the baffle can

not arrive at the second stage andconsume capacity there), the situationarises as shown in Fig. 2.69.

The gas molecules entering the pump pro-duce the area related theoretical pumpingspeed according the equation 2.29a with T= 293 K. The different pumping speedshave been combined for three representa-tive gases H2, N2 and H20 taken from eachof the aforementioned groups. Since watervapor is pumped on the entire entry area ofthe cryopump, the pumping speedmeasured for water vapor correspondsalmost exactly to the theoretical pumpingspeed calculated for the intake flange ofthe cryopump. N2 on the other hand mustfirst overcome the baffle before it can bebonded on to the cryocondensation panel.Depending on the design of the baffle, 30to 50 percent of all N2 molecules arereflected.

H2 arrives at the cryosorption panels afterfurther collisions and thus cooling of thegas. In the case of optimally designedcryopanels and a good contact with the ac-tive charcoal up to 50 percent of the H2which has overcome the baffle can bebonded. Due to the restrictions regardingaccess to the pumping surfaces andcooling of the gas by collisions with thewalls inside the pump before the gasreaches the pumping surface, the measu-red pumping speed for these two gasesamounts only to a fraction of the theoreti-cal pumping speed. The part which is notpumped is reflected chiefly by the baffle.Moreover, the adsorption probability forH2 differs between the various adsorbentsand is < 1, whereas the probabilities for thecondensation of water vapor and N2 ≈ 1.

Three differing capacities of a pump for thegases which can be pumped result fromthe size of the three surfaces (baffle,condensation surface at the outside of thesecond stage and sorption surface at theinside of the second stage). In the designof a cryopump, a mean gas composition(air) is assumed which naturally does notapply to all vacuum processes (sputteringprocesses, for example. See 2.1.9.6“Partial Regeneration”).

Fig. 2.69 Cryopanels – temperature and position define the efficiency in the cryopump

Hydrogen Water vapor NitrogenArea related conductance of the intake flange in l / s · cm2:

43.9 14.7 11.8Area related pumping speed of the cryopump in l / s · cm2:

13.2 14.6 7.1Ratio between pumping speed and conductance:

30 % 99 % 60 %

D00 E 19.06.2001 21:37 Uhr Seite 52



2.1.9.6 Characteristic quantities of acryopump

The characteristic quantities of a cryo-pump are as follows (in no particularorder):

• Cooldown time

• Crossover value

• Ultimate pressure

• Capacity

• Refrigerating power and net refrigerating power

• Regeneration time

• Throughput and maximum pV flow

• Pumping speed

• Service life / duration of operation

• Starting pressure

Cooldown time: The cooldown time ofcryopumps is the time span from start-upuntil the pumping effect sets in. In the caseof refrigerator cryopumps the cooldowntime is stated as the time it takes for thesecond stage of the cold head to cooldown from 293 K to 20 K.

Crossover value: The crossover value is acharacteristic quantity of an already coldrefrigerator cryopump. It is of significancewhen the pump is connected to a vacuumchamber via an HV / UHV valve. Thecrossover value is that quantity of gas withrespect to Tn = 293 K which the vacuumchamber may maximally contain so thatthe temperature of the cryopanels doesnot increase above 20 K due to the gasburst when opening the valve. Thecrossover value is usually stated as a pVvalue in in mbar · l.The crossover value and the chambervolume V result in the crossover pressurepc to which the vacuum chamber must beevacuated first before opening the valveleading to the cryopump. The followingmay serve as a guide:

(2.27)

V = Volume of the vacuum chamber (l),

Q.2(20K) = Net refrigerating capacity in

Watts, available at the second stage of thecold head at 20 K.

pc V

Q K mbar≤ ·35 202·

( )

Ultimate pressure pend: For the case ofcryocondensation (see Section 2.1.9.4)the ultimate pressure can be calculated by:

(2.28)

pS is the saturation vapor pressure of thegas or gases which are to be pumped atthe temperature TK of the cryopanel and TGis the gas temperature (wall temperature inthe vicinity of the cryopanel).

Example: With the aid of the vaporpressure curves in Fig. 9.15 for H2 and N2the ultimate pressures summarized inTable 2.6 at TG = 300 K result.

The Table shows that for hydrogen attemperatures T < 3 K at a gas temperatureof TG= 300 K (i.e. when the cryopanel isexposed to the thermal radiation of thewall) sufficiently low ultimate pressurescan be attained. Due to a number of inter-fering factors like desorption from the walland leaks, the theoretical ultimate pres-sures are not attained in practice.

Capacity C (mbar · l): The capacity of acryopump for a certain gas is that quantityof gas (pV value at Tn = 293 K) which canbe bonded by the cryopanels before thepumping speed for this type of gas Gdrops to below 50 % of its initial value.The capacity for gases which are pumpedby means of cryosorption depends on thequantity and properties of the sorptionagent; it is pressure dependent andgenerally by several orders of magnitudelower compared to the pressureindependent capacity for gases which arepumped by means of cryocondensation.

Refrigerating power Q.

(W): The refrige-rating power of a refrigeration source at atemperature T gives the amount of heatthat can be extracted by the refrigeratingsource whilst still maintaining thistemperature. In the case of refrigerators ithas been agreed to state for single-stagecold heads the refrigerating power at 80 Kand for two-stage cold heads the

( )p p T TTend s K

G

K= ·

refrigerating power for the first stage at 80K and for the second stage at 20 K whensimultaneously loading both stagesthermally. During the measurement ofrefrigerating power the thermal load isgenerated by electric heaters. The refrige-rating power is greatest at room tempe-rature and is lowest (Zero) at ultimate tem-perature.

Net refrigerating power Q.

(W): In thecase of refrigerator cryopumps the netrefrigerating power available at the usualoperating temperatures (T1 < 80 K, T2 < 20 K) substantially definesthe throughput and the crossover value.The net. refrigerating power is – depen-ding on the configuration of the pump –much lower than the refrigerating power ofthe cold head used without the pump.

pV flow see 1.1

Regeneration time: As a gas trappingdevice, the cryopump must be regeneratedafter a certain period of operation. Rege-neration involves the removal of condens-ed and adsorbed gases from the cryopa-nels by heating. The regeneration can berun fully or only partially and mainly differsby the way in which the cryopanels areheated.

In the case of total regeneration a diffe-rence is made between:1. Natural warming: after switching off the

compressor, the cryopanels at firstwarm up only very slowly by thermalconduction and then in additionthrough the released gases.

2. Purge gas method: the cryopump iswarmed up by admitting warm purginggas.

3. Electric heaters: the cryopanels of thecryopump are warmed up by heaters atthe first and second stages. Thereleased gases are discharged eitherthrough an overpressure valve (purgegas method) or by mechanical backingpumps. Depending on the size of thepump one will have to expect aregeneration time of several hours.

D00

TK (K) Ultimate pressure Ult. press. (mbar) Ult. press.(mbar)(according to equ. 2.28) H2 N2

2.5 10.95 · ps 3.28 · 10–14 immeasurably low4.2 8.66 · ps 4.33 · 10–9 immeasurably low20 3.87 · ps 3.87 · 10+3 3.87 · 10–11

Table 2.6 Ultimate temperatures at a wall temperature of 300 K

D00 E 19.06.2001 21:37 Uhr Seite 53



Partial regeneration: Since the limitationin the service life of a cryopump dependsin most applications on the capacity limitfor the gases nitrogen, argon andhydrogen pumped by the second stage, itwill often be required to regenerate onlythis stage. Water vapor is retained duringpartial regeneration by the baffle. For this,the temperature of the first stage must bemaintained below 140 K or otherwise thepartial pressure of the water vapor wouldbecome so high that water moleculeswould contaminate the adsorbent on thesecond stage.

In 1992, LEYBOLD was the first manufac-turer of cryopumps to develop a methodpermitting such a partial regeneration.This fast regeneration process is micro-processor controlled and permits a partialregeneration of the cryopump in about 40minutes compared to 6 hours needed for atotal regeneration based on the purge gasmethod. A comparison between the typicalcycles for total and partial regeneration isshown in Fig. 2.70. The time saved by theFast Regeneration System is apparent. In aproduction environment for typical sput-tering processes one will have to expectone total regeneration after 24 partialregenerations.

Throughput and maximum pV flow:(mbar l/s): the throughput of a cryopumpfor a certain gas depends on the pV flow ofthe gas G through the intake opening ofthe pump:

QG = qpV,G; the following equationapplies

QG = pG · SG with

pG = intake pressure,

SG = pumping capacity for the gas G

The maximum pV flow at which thecryopanels are warmed up to T ≈ 20 K inthe case of continuous operation, de-pends on the net refrigerating power of thepump at this temperature and the type ofgas. For refrigerator cryopumps andcondensable gases the following may betaken as a guide:

Qmax = 2.3 Q.

2 (20 K) mbar · l/s

Q.

2 (20 K) is the net refrigerating power inWatts available at the second stage of thecold heat at 20 K. In the case of inter-mittent operation, a higher pV flow is per-missible (see crossover value).

Pumping speed Sth: The following appliesto the (theoretical) pumping speed of acryopump:

(2.29)

AK Size of the cryopanels

SA Surface area related pumping speed(area related impact rate according toequations 1.17 and 1.20, proportionalto the mean velocity of the gasmolecules in the direction of thecryopanel)

a Probability of condensation(pumping)

pend Ultimate pressure (see above)

p Pressure in the vacuum chamberThe equation (2.29) applies to a cryopanelbuilt into the vacuum chamber, the surfacearea of which is small compared to thesurface of the vacuum chamber. Atsufficiently low temperatures α = 1 for allgases. The equation (2.29) shows that for

S A Sppth K Aend= · · · −

α 1

p >> pend the expression in bracketsapproaches 1 so that in the oversaturatedcase

p >> pend > Ps so that:

Sth = AK · SA (2.29a)

with

TG Gas temperature in K

M Molar mass

Given in Table 2.7 is the surface area-related pumping speed SA in l · s-1 · cm-2

for some gases at two different gastemperatures TG in K determined accor-ding to equation 2.29a. The values statedin the Table are limit values. In practice thecondition of an almost undisturbedequilibrium (small cryopanels comparedto a large wall surface) is often not true,because large cryopanels are required toattain short pumpdown times and a goodend vacuum. Deviations also result whenthe cryopanels are surrounded by a cooledbaffle at which the velocity of thepenetrating molecules is already reducedby cooling.

Service life / duration of operation top (s):The duration of operation of the cryopumpfor a particular gas depends on theequation:

with

CG = Capacity of the cryopump for thegas G

QG(t) =Throughput of the cryopump forthe gas at the point of time t

If the constant mean over time for thethroughput QG

__is known, the following

applies:

(2.30)

After the period of operation top,G haselapsed the cryopump must be rege-nerated with respect to the type of gas G.

Starting pressure po: Basically it ispossible to start a cryopump at atmosphe-ric pressure. However, this is not desirablefor several reasons. As long as the meanfree path of the gas molecules is smaller

t CQ

Cp Sop G

G

G

G

G G, = =

·

C Q t dtG G

top G

= ∫ ( ),

0

S R TM

c TM

s cmAG G= ·

· ·== ·

24365 2

π. /

Fig. 2.70 Comparison between total (1) and partial (2) regeneration

Time

(2) (1)

about 6 hours40 minutes

Tem

pera

ture

(K)

2 1

D00 E 19.06.2001 21:37 Uhr Seite 54



than the dimensions of the vacuum cham-ber (p > 10-3 mbar), thermal conductivityof the gas is so high that an unacceptablylarge amount of heat is transferred to thecryopanels. Further, a relatively thick layerof condensate would form on the cryo-panel during starting. This would markedlyreduce the capacity of the cryopumpavailable to the actual operating phase.Gas (usually air) would be bonded to theadsorbent, since the bonding energy forthis is lower than that for the condensationsurfaces. This would further reduce thealready limited capacity for hydrogen. It isrecommended that cryopumps in the highvacuum or ultrahigh vacuum range arestarted with the aid of a backing pump atpressures of p < 5 · 10-2 mbar. As soon asthe starting pressure has been attained thebacking pump may be switched off.

D00

M SA SA TS Triple pointSymbol Gas Molar at 293 K at 80 K Boiling point (= melting point)

mass gas temp. gas temp. 1013 mbar Tt pt

g/mol l/s · cm2 l/s · cm2 K K mbar

H2 Hydrogen 2.016 43.88 22.93 20.27 13.80 70.4

He Helium 4.003 31.14 16.27 4.222 2.173 50.52

CH4 Methane 4.003 15.56 8.13 111.67 90.67 116.7

H2O Water 18.015 14.68 – 373.15 273.16 6.09

Ne Neon 20.183 13.87 7.25 27.102 24.559 433.0

CO Carbon monoxide 28.000 11.77 6.15 81.67 68.09 153.7

N2 Nitrogen 28.013 11.77 6.15 77.348 63.148 126.1

Air 28.96 11.58 6.05 ≈ 80.5 ≈ 58.5 –

O2 Oxygen 31.999 11.01 5.76 90.188 54.361 1.52

Ar Argon 39.948 9.86 5.15 87.26 83.82 687.5

Kr Krypton 83.80 6.81 3.56 119.4 115.94 713.9

Xe Xenon 131.3 5.44 2.84 165.2 161.4

Table 2.7 Surface-related pumping speeds for some gases

D00 E 19.06.2001 21:37 Uhr Seite 55



2.2 Choice of pumpingprocess

2.2.1 Survey of the most usualpumping processes

Vacuum technology has undergone rapiddevelopment since the Fifties. In researchand in most branches of industry today, itis indispensable.

Corresponding to the many areas of ap-plication, the number of technical proce-dures in vacuum processes is extraordi-narily large. These cannot be describedwithin the scope of this section, becausethe basic calculations in this section covermainly the pumping process, not theprocess taking place in the vessel. Asurvey of the most important processes invacuum technology and the pressureregions in which these processes arechiefly carried out is given in the diagrams(Figures 2.71 and 2.72).

Generally, the pumping operation for theseprocesses can be divided into twocategories – dry- and wet – vacuum proce-dures, that is, into processes in which nosignificant amounts of vapor have to bepumped and those in which vapors(mostly water or organic) arise.

Distinctions between the two categoriesare described briefly:

Dry processes work primarily in a narrowand limited pressure region.The system is usually evacuated to asuitable characteristic pressure before theactual working process begins. Thishappens, for example, in plants for evapo-rative coating, electron-beam welding, andcrystal pulling; in particle accelerators,mass spectrometers, electron micro-scopes; and others.

Further, there are dry processes in whichdegassing in vacuum is the actualtechnical process. These include work ininduction- and arc furnaces, steel degas-sing plants, and plants for the manufactureof pure metals and electron tubes.

Wet processes are undertaken primarilyin a prescribed working operation thatcovers a wider pressure region. This is especially important in the drying ofsolid materials. If, for instance, work isundertaken prematurely at too low apressure, the outer surfaces dry out tooquickly. As a result, the thermal contact tothe moisture to be evaporated is impairedand the drying time is considerablyincreased. Predominantly processes thatare carried out in drying, impregnating,and freeze-drying plants belong in thiscategory.

In the removal of water vapor from liquidsor in their distillation, particularly indegassing columns, vacuum filling, and

resin-casting plants, as well as inmolecular distillation, the production of aslarge a liquid surface as possible isimportant. In all wet processes the provi-sion of the necessary heat for evapo-ration of the moisture is of great impor-tance.

Basic pumping procedures are given in thefollowing paragraphs.

If you have specific questions, you shouldget in touch with a specialist department inLEYBOLD where experts are available toyou who can draw on many years ofexperience.

Classifications of typical vacuum proces-ses and plants according to the pressureregions.Rough vacuum 1013 mbar – 1 mbar• Drying, distillation, and steel degassing.Medium vacuum 1 -10-3 mbar• Molecular distillation, freeze-drying,

impregnation, melting and castingfurnaces, and arc furnaces.

High vacuum 10-3 – 10-7 mbar• Evaporative coating, crystal pulling,

mass spectrometers, tube production,electron microscopes, electron beamplants, and particle accelerators.

Ultrahigh vacuum: < 10-7 mbar• Nuclear fusion, storage rings for

accelerators, space research, andsurface physics.

Rough vacuumMedium vacuumHigh vacuumUltrahigh vacuum

Mass spectrometers

Molecular beam apparatus

Ion sources

Particle accelerators

Electron microscopes

Electron diffraction apparatus

Vacuum spectographs

Low-temperature research

Production of thin films

Surface physics

Plasma research

Nuclear fusion apparatus

Space simulation

Material research

Preparations forelectron microscopy

10–13 10–10 10–7 10–3 100 103

Pressure [mbar]

Fig. 2.71 Pressure ranges (p < 1000 mbar) of physical and chemical analytical methods

D00 E 19.06.2001 21:37 Uhr Seite 56



2.2.2 Pumping of gases (dry processes)

For dry processes in which a non conden-sable gas mixture (e.g., air) is to bepumped, the pump to be used is clearlycharacterized by the required workingpressure and the quantity of gas to bepumped away. The choice of the requiredworking pressure is considered in thissection. The choice of the required pumpis dealt with in Section 2.3.

Each of the various pumps has a charac-teristic working range in which it has aparticularly high efficiency. Therefore, themost suitable pumps for use in the follo-wing individual pressure regions are des-cribed. For every dry-vacuum process, thevessel must first be evacuated. It is quitepossible that the pumps used for this maybe different from those that are theoptimum choices for a process that isundertaken at definite working pressures.In every case the choice should be madewith particular consideration for thepressure region in which the workingprocess predominantly occurs.

a) Rough vacuum (1013 – 1 mbar)The usual working region of the rotarypumps described in Section 2 lies below80 mbar. At higher pressures these pumpshave a very high power consumption (seeFig. 2.11) and a high oil consumption (seeSection 8.3.1.1). Therefore, if gases are to

be pumped above 80 mbar over longperiods, one should use, particularly oneconomic grounds, jet pumps, water ringpumps or dry running, multi-vane pumps.Rotary vane and rotary piston pumps areespecially suitable for pumping downvessels from atmospheric pressure topressures below 80 mbar, so that they canwork continuously at low pressures. Iflarge quantities of gas arise at inletpressures below 40 mbar, the connectionin series of a Roots pump is recommen-ded. Then, for the backing pump speedrequired for the process concerned, amuch smaller rotary vane or piston pumpcan be used.

b) Medium vacuum (1 – 10-3 mbar)If a vacuum vessel is merely to be eva-cuated to pressures in the medium vacu-um region, perhaps to that of the requiredbacking pressure for diffusion or sputter-ion pumps, single- and two-stage rotarypumps are adequate for pressures down to10-1 and 10-3 mbar, respectively. It is es-sentially more difficult to select thesuitable type of pump if medium vacuumprocesses are concerned in which gasesor vapors are evolved continuously andmust be pumped away. An important hintmay be given at this point. Close to theattainable ultimate pressure, the pumpingspeed of all rotary pumps falls off rapidly.Therefore, the lowest limit for the normalworking pressure region of these pumpsshould be that at which the pumping speed

still amounts to about 50 % of the nominalpumping speed.

Between 1 and 10-2 mbar at the onset oflarge quantities of gas, Roots pumps withrotary pumps as backing pumps haveoptimum pumping properties (see Section2.1.3.1). For this pressure range, a single-stage rotary pump is sufficient, if the chiefworking region lies above 10-1 mbar. If itlies between 10-1 and 10-2 mbar, a two-stage backing pump is recommended.Below 10-2 mbar the pumping speed ofsingle-stage Roots pumps in combinationwith two-stage rotary pumps as backingpumps decreases. However, between 10-2

and 10-4 mbar, two-stage Roots pumps (ortwo single-stage Roots pumps in series)with two-stage rotary pumps as backingpumps still have a very high pumpingspeed. Conversely, this pressure region isthe usual working region for vapor ejectorpumps. For work in this pressure region,they are the most economical pumps topurchase. As backing pumps, single-stagerotary positive displacement pumps aresuitable. If very little maintenance and val-veless operation are convenient (i.e., smallvessels in short operation cycles are to bepumped to about 10-4 mbar or largevessels are to be maintained at this pres-sure unattended for weeks), the previouslymentioned two-stage Roots pumps withtwo-stage rotary pumps as backing pumpsare the suitable combinations. Allthough,such a combination does not work as

D00


10–10 10–7 10–3 100 103

Pressure [mbar]

Annealing of metalsMelting of metals

Degassing of molten metalsSteel degassing

Electron-beam meltingElectron-beam welding

Evaporation coatingSputtering of metals

Zone melting and crystal growing in high-vacuumMolecular distillationDegassing of liquids

SublimationCasting of resins and lacquers

Drying of plasticsDrying of insulating papers

Freeze-drying of mass materialsFreeze-drying of pharmaceutical products

Production of incandescent lampsProduction of electron tubes

Production of gas-discharge tubes

Fig. 2.72 Pressure ranges of industrial vacuum processes

D00 E 19.06.2001 21:37 Uhr Seite 57



economically as the corresponding vaporejector pump, it can operate for a muchlonger time without maintenance.

c) High vacuum (10-3 to 10-7 mbar)Diffusion, sputter-ion, and turbomolecularpumps typically operate in the pressureregion below 10-3 mbar. If the workingregion varies during a process, differentpumping systems must be fitted to thevessel. There are also special diffusionpumps that combine the typical propertiesof a diffusion pump (low ultimate pres-sure, high pumping speed in the highvacuum region) with the outstandingproperties of a vapor ejector pump (highthroughput in the medium vacuum region,high critical backing pressure). If theworking region lies between 10-2 and 10-6

mbar, these diffusion pumps are, ingeneral, specially recommended.

d) Ultrahigh vacuum (< 10-7 mbar)For the production of pressures in theultrahigh vacuum region, sputter-ion, andsublimation pumps, as well as turbomole-cular pumps and cryopumps, are used incombination with suitable forepumps. Thepump best suited to a particular UHVprocess depends on various conditions(for further details, see Section 2.5).

2.2.3 Pumping of gases andvapors (wet processes)

When vapors must be pumped, in additionto the factors working pressure andpumping speed, a third determining factoris added namely the vapor partial pressure– which may vary considerably in thecourse of a process. This factor is decisivein determining the pumping arrangementto be installed. In this regard, the con-densers described in Section 2.15 are veryimportant adjuncts to rotary displacementpumps. They have a particularly highpumping speed when pumping vapors.The next section covers pumping of watervapor (the most frequent case). Theconsiderations apply similarly to othernon-aggressive vapors.

Pumping of Water Vapor Water vapor is frequently removed bypumps that operate with water or steam asa pump fluid, for example, water ringpumps or steam ejector pumps. This de-pends considerably on circumstances,however, because the economy of steamejector pumps at low pressures is gene-rally far inferior to that of rotary pumps.For pumping a vapor – gas mixture inwhich the vapor portion is large but the airportion is small, the vapor can be pumpedby condensers and the permanent gases,by relatively small gas ballast pumps (seeSection 2.1.5).

Comparatively, then, a pump set con-

sisting of a Roots pump, condenser, andbacking pump, which can transport 100kg/h of vapor and 18 kg/h of air at an inletpressure of 50 mbar, has a powerrequirement of 4 – 10 kW (depending onthe quantity of air involved). A steamejector pump of the same performancerequires about 60 kW without altering thequantity of air involved.

For the pumping of water vapor, gas ballastpumps and combinations of gas ballastpumps, Roots pumps, and condensers areespecially suitable.

Pumping of water vapor with gas ballastpumps The ratio of vapor partial pressure pv to airpartial pressure pp is decisive in theevaluation of the correct arrangement ofgas ballast pumps, as shown previously byequations 2.2 and 2.3. Therefore, if thewater vapor tolerance of the gas ballastpump is known, graphs may be obtainedthat clearly give the correct use of gasballast pumps for pumping water vapor(see Fig. 2.73). Large single-stage rotaryplunger pumps have, in general, anoperating temperature of about 77 °C andhence a water vapor tolerance of about 60mbar. This value is used to determine thedifferent operating regions in Fig. 2.73. Inaddition it is assumed that the pressure atthe discharge outlet port of the gas ballastpump can increase to a maximum of 1330mbar until the discharge outlet valveopens.

Region A: Single-stage, rotary plungerpumps without gas ballast inlet. At a saturation vapor pressure pS of 419mbar at 77 °C, according to equation 2.2,the requirement is given that pv < 0.46 pp,where pv is the water vapor partial pressure pp is the partial pressure of airpv + pp = ptot total pressure

This requirement is valid in the wholeworking region of the single-stage rotaryplunger pump – hence, at total pressuresbetween 10-1 and 1013 mbar.

Region B: Single-stage rotary plungerpumps with gas ballast and an inletcondenser. In this region the water vapor pressureexceeds the admissible partial pressure atthe inlet. The gas ballast pump must,Fig. 2.73 Areas of application for gas ballast pumps and condensers pumping water vapor (o.G. = without gas ballast)

Wat

er v

apor

par

tial p

ress

ure

p v

Air partial pressure pp

pv/pp

= 0.4

6

D00 E 19.06.2001 21:37 Uhr Seite 58



therefore, have a condenser inserted at theinlet, which is so rated that the water vaporpartial pressure at the inlet port of therotary pump does not exceed theadmissible value. The correct dimensionsof the condenser are selected dependingon the quantity of water vapor involved.For further details, see Section 2.1.5. At awater vapor tolerance of 60 mbar, thelower limit of this region is

pv > 6O + 0.46 pp mbar

Region C: Single-stage rotary plungerpumps with a gas ballast. The lower limit of region C is characterizedby the lower limit of the working region ofthis pump. It lies, therefore, at about ptot = 1 mbar. If large quantities of vaporarise in this region, it is often moreeconomical to insert a condenser: 20 kg ofvapor at 28 mbar results in a volume ofabout 1000 m3. It is not sensible to pumpthis volume with a rotary pump. As a ruleof thumb:

A condenser should always be inserted atthe pump’s inlet if saturated water vaporarises for a considerable time. As a precaution, therefore, a Roots pumpshould always be inserted in front of thecondenser at low inlet pressures so thatthe condensation capacity is essentiallyenhanced. The condensation capacity doesnot depend only on the vapor pressure,but also on the refrigerant temperature. Atlow vapor pressures, therefore, effectivecondensation can be obtained only if therefrigerant temperature is correspondinglylow. At vapor pressures below 6.5 mbar,for example, the insertion of a condenseris sensible only if the refrigerant tempera-ture is less than 0 °C. Often at low pres-sures a gas – vapor mixture withunsaturated water vapor is pumped (forfurther details, see Section 2.1.5). Ingeneral, then, one can dispense with thecondenser.

Region D: Two-stage gas ballast pumps,Roots pumps, and vapor ejector pumps,always according to the total pressureconcerned in the process. It must again be noted that the water vaportolerance of two-stage gas ballast pumpsis frequently lower than that of correspon-ding single-stage pumps.

Pumping of water vapor with roots pumpsNormally, Roots pumps are not aseconomical as gas ballast pumps forcontinuous operation at pressures above40 mbar. With very large pump sets, whichwork with very specialized gear ratios andare provided with bypass lines, however,the specific energy consumption is indeedmore favorable. If Roots pumps areinstalled to pump vapors, as in the case ofgas ballast pumps, a chart can be giventhat includes all possible cases (see Fig.2.74).

Region A: A Roots pump with a single-stage rotary plunger pump without gasballast. As there is merely a compression betweenthe Roots pump and the rotary plungerpump, the following applies here too:

pv < 0.46 pp

The requirement is valid over the entireworking region of the pump combinationand, therefore, for total pressures between10-2 and 40 mbar (or 1013 mbar for Rootspumps with a bypass line).

Region B: A main condenser, a Rootspump with a bypass line, an intermediatecondenser, and a gas ballast pump. This combination is economical only iflarge water vapor quantities are to bepumped continuously at inlet pressuresabove about 40 mbar. The size of the maincondenser depends on the quantity ofvapor involved. The intermediate con-

denser must decrease the vapor partialpressure below 60 mbar. Hence, the gasballast pump should be large enough onlyto prevent the air partial pressure behindthe intermediate condenser from ex-ceeding a certain value; for example, if thetotal pressure behind the Roots pump(which is always equal to the totalpressure behind the intermediate conden-ser) is 133 mbar, the gas ballast pumpmust pump at least at a partial air pressureof 73 mbar, the quantity of air transportedto it by the Roots pump. Otherwise, it musttake in more water vapor than it cantolerate. This is a basic requirement: theuse of gas ballast pumps is wise only ifair is also pumped!

With an ideally leak-free vessel, the gasballast pump should be isolated after therequired operating pressure is reached andpumping continued with the condenseronly. Section 2.1.5 explains the bestpossible combination of pumps andcondensers.

Region C: A Roots pump, an intermediatecondenser, and a gas ballast pump. The lower limit of the water vapor partialpressure is determined through thecompression ratio of the Roots pump atthe backing pressure, which is determinedby the saturation vapor pressure of thecondensed water. Also, in this region theintermediate condenser must be able toreduce the vapor partial pressure to atleast 60 mbar. The stated arrangement is

D00Fig. 2.74 Areas of application for Roots pumps and condensers pumping water vapor (o.G. = without gas ballast)


Wat

er v

apor

par

tial p

ress

ure

p v


p v/p p

= 0.46

D00 E 19.06.2001 21:38 Uhr Seite 59



suitable – when cooling the condenserwith water at 15 °C – for water vaporpressures between about 4 and 40 mbar.

Region D: A roots pump and a gas ballastpump. In this region D the limits also dependessentially on the stages and ratios ofsizes of the pumps. In general, however,this combination can always be usedbetween the previously discussed limits –therefore, between 10-2 and 4 mbar.

2.2.4 Drying processes

Often a vacuum process covers several ofthe regions quoted here. In batch dryingthe process can, for example (see Fig.2.74), begin in region A (evacuation of theempty vessel) and then move throughregions B, C, and D in steps. Then thecourse of the process would be as follows:

A. Evacuating the vessel by a gas ballastpump and a Roots pump with a bypassline.

B. Connecting the two condensersbecause of the increasing vapor pressureproduced by heating the material. The choice of the pumping System isdecided by the highest vapor partialpressure occurring and the lowest airpartial pressure at the inlet.

C. Bypassing the main condenserIt will now not have an effect. Instead itwould only be pumped empty by thepumping system with a further drop invapor pressure.

D. Bypassing the intermediate condenserRoots pumps and gas ballast pumps alonecan now continue pumping. With short-term drying, the separation of thecondenser filled with condensed water isparticularly important, because the gasballast pump would continue to pumpfrom the condenser the previouslycondensed water vapor at the saturationvapor pressure of water.

With longer-term drying processes, itsuffices to shut off the condensatecollector from the condenser. Then onlythe remaining condensate film on thecooling tubes can reevaporate. Dependingon the size of the gas ballast pump, thisreevaporation ensues in 30 – 60 min.

E. If the drying process should terminateat still lower pressures When a pressure below 10-2 mbar isreached a previously bypassed oil vaporejector pump should be switched on inaddition.

Drying of solid substances As previously indicated, the drying of solidsubstances brings about a series of furtherproblems. It no longer suffices that onesimply pumps out a vessel and then waitsuntil the water vapor diffuses from thesolid substance. This method is indeedtechnically possible, but it would intorablyincrease the drying time.

It is not a simple technical procedure tokeep the drying time as short as possible.Both the water content and the layer thick-ness of the drying substance are im-portant. Only the principles can be statedhere. In case of special questions weadvise you to contact our experts in ourCologne factory.

The moisture content E of a material to bedried of which the diffusion coefficientdepends on the moisture content (e.g. withplastics) as a function of the drying time tis given in close approximation by thefollowing equation:

(2.31)

E0 where E is the moisture content beforedrying

q is the temperature-dependent coef-ficient. Thus equation (2.31) servesonly for the temperature at which qwas determined

K is a factor that depends on thetemperature, the water vapor partialpressure in the vicinity of the material,the dimensions, and the properties ofthe material.

With the aid of this approximate equation,the drying characteristics of manysubstances can be assessed. If K and qhave been determined for varioustemperatures and water vapor partialpressures, the values for othertemperatures are easily interpolated, sothat the course of the drying process canbe calculated under all operatingconditions. With the aid of a similaritytransformation, one can further comparethe course of drying process of a material

( )E

E

K t q=+ ·

0

1%

with known properties with that of amaterial with different properties.

Fundamentally, in the drying of a material, a few rules are noteworthy:Experience has shown that shorter dryingtimes are obtained if the water vaporpartial pressure at the surface of thematerial is relatively high, that is, if thesurface of the material to be dried is notyet fully free of moisture. This is possiblebecause the heat conduction between thesource of heat and the material is greaterat higher pressures and the resistance todiffusion in a moist surface layer is smallerthan in a dry one. To fulfill the conditionsof a moist surface, the pressure in thedrying chamber is controlled. If the neces-sary relatively high water vapor partialpressure cannot be maintained perma-nently, the operation of the condenser istemporarily discontinued. The pressure inthe chamber then increases and thesurface of the material becomes moistagain. To reduce the water vapor partialpressure in the vessel in a controlled way,it may be possible to regulate the re-frigerant temperature in the condenser. Inthis way, the condenser temperatureattains preset values, and the water vaporpartial pressure can be reduced in a con-trolled manner.

2.2.5 Production of an oil-free(hydrocarbon-free)vacuum

Backstreaming vapor pump fluids, vaporsof oils, rotary pump lubricants, and theircracking products can significantly disturbvarious working processes in vacuum.Therefore, it is recommended that certainapplications use pumps and devices thatreliably exclude the presence of hydro-carbon vapors.

a) Rough vacuum region (1013 to 1 mbar)

Instead of rotary pumps, large water jet,steam ejector, or water ring pumps can beused. For batch evacuation, and the pro-duction of hydrocarbon-free fore vacuumfor sputter-ion pumps, adsorption pumps(see Section 2.1.8.1) are suitable. If theuse of oil-sealed rotary vane pumpscannot be avoided, basically two-stage

D00 E 19.06.2001 21:38 Uhr Seite 60



rotary vane pumps should be used. Thesmall amount of oil vapor that back-streams out of the inlet ports of thesepumps can be almost completely removedby a sorption trap (see Section 2.1.4)inserted in the pumping line.

b) Medium vacuum region (1 to 10-3 mbar)

For the pumping of large quantities of gasin this pressure region, vapor ejectorpumps are by far the most suitable. Withmercury vapor ejector pumps, completelyoil-free vacua can be produced. As a pre-caution, the insertion of a cold trap chilledwith liquid nitrogen is recommended sothat the harmful mercury vapor does notenter the vessel. With the medium vacuumsorption traps described under a), it ispossible with two-stage rotary vanepumps to produce almost oil-free vacuadown to below 10-4 mbar.

Absolutely oil-free vacua may be producedin the medium vacuum region withadsorption pumps. Since the pumpingaction of these pumps for the light noblegases is only small, vessels initially filledwith air can only be evacuated by them toabout 10-2 mbar. Pressures of 10-3 mbaror lower can then be produced withadsorption pumps only if neither neon norhelium is present in the gas mixture to bepumped. In such cases it can be useful toexpel the air in the vessel by first floodingwith nitrogen and then pumping it away.

c) High- and ultrahigh vacuum region (< 10-3 mbar)

When there is significant evolution of gasin the pressure regions that must be pum-ped, turbomolecular pumps, or cryo-pumps should be used. A sputter-ionpump is especially suitable for maintainingthe lowest possible pressure for longperiods in a sealed system where theprocess does not release large quantitiesof gas. Magnetically-suspended turbomo-lecular pumps also guarantee hydrocar-bon-free vacua. However, while thesepumps are switched off, oil vapors canenter the vessel through the pump. Bysuitable means (e.g., using an isolatingvalve or venting the vessel with argon),contamination of the vessel walls can beimpeded when the pump is stationary. Ifthe emphasis is on generating a “hydroc-arbon-free vacuum” with turbomolecularpumps, then hybrid turbomolecular

pumps with diaphragm pumps or classicturbomolecular pumps combined withscroll pumps should be used as oil-freebacking pumps.

2.2.6 Ultrahigh vacuumworking Techniques

The boundary between the high and ultra-high vacuum region cannot be preciselydefined with regard to the working me-thods. In practice, a border between thetwo regions is brought about becausepressures in the high vacuum region maybe obtained by the usual pumps, valves,seals, and other components, whereas forpressures in the UHV region, anothertechnology and differently constructedcomponents are generally required. The“border” lies at a few 10-8 mbar. Therefore,pressures below 10-7 mbar should ge-nerally be associated with the UHV region.

The gas density is very small in the UHVregion and is significantly influenced byoutgasing rate of the vessel walls and bythe tiniest leakages at joints. Moreover, inconnection with a series of importanttechnical applications to characterize theUHV region, generally the monolayer time(see also equation 1.21) has becomeimportant. This is understood as the time tthat elapses before a monomolecular ormonatomic layer forms on an initiallyideally cleaned surface that is exposed tothe gas particles. Assuming that every gasparticle that arrives at the surface finds afree place and remains there, a convenientformula for τ is

(p in mbar)

Therefore, in UHV (p < 10-7 mbar) themonolayer formation time is of the orderof minutes to hours or longer and thus ofthe same length of time as that needed forexperiments and processes in vacuum.The practical requirements that arise havebecome particularly significant in solid-state physics, such as for the study of thinfilms or electron tube technology. A UHVsystem is different from the usual highvacuum system for the following reasons: a) the leak rate is extremely small (use of

metallic seals), b) the gas evolution of the inner surfaces

τ = · –3 210 6.

p s

of the vacuum vessel and of theattached components (e.g., connectingtubulation; valves, seals) can be madeextremely small,

c) suitable means (cold traps, baffles) areprovided to prevent gases or vapors ortheir reaction products that have origi-nated from the pumps used fromreaching the vacuum vessel (nobackstreaming).

To fulfill these conditions, the individualcomponents used in UHV apparatus mustbe bakeable and extremely leaktight.Stainless steel is the preferred material forUHV components.

The construction, start-up, and operationof an UHV system also demands specialcare, cleanliness, and, above all, time. Theassembly must be appropriate; that is, theindividual components must not be in theleast damaged (i.e. by scratches onprecision-worked sealing surfaces). Fun-damentally, every newly-assembled UHVapparatus must be tested for leaks with ahelium leak detector before it is operated.Especially important here is the testing ofdemountable joints (flange connections),glass seals, and welded or brazed joints.After testing, the UHV apparatus must bebaked out. This is necessary for glass aswell as for metal apparatus. The bake-outextends not only over the vacuum vessel,but frequently also to the attached parts,particularly the vacuum gauges. Theindividual stages of the bake-out, whichcan last many hours for a larger system,and the bake-out temperature are arrangedaccording to the kind of plant and theultimate pressure required. If, after theapparatus has been cooled and the othernecessary measures undertaken (e.g.,cooling down cold traps or baffles), theultimate pressure is apparently notobtained, a repeated leak test with ahelium leak detector is recommended.Details on the components, sealingmethods and vacuum gauges are providedin our catalog.

D00

D00 E 19.06.2001 21:38 Uhr Seite 61



2.3. Evacuation of avacuum chamberand determinationof pump sizes

Basically, two independent questionsarise concerning the size of a vacuumsystem:1. What effective pumping speed must the

pump arrangement maintain to reducethe pressure in a given vessel over agiven time to a desired value?

2. What effective pumping speed must thepump arrangement reach during avacuum process so that gases andvapors released into the vessel can bequickly pumped away while a givenpressure (the operating pressure) in thevessel, is maintained and notexceeded?

During the pumping-out procedure ofcertain processes (e.g., drying and hea-ting), vapors are produced that were notoriginally present in the vacuum chamber,so that a third question arises:

3. What effective pumping speed must thepump arrangement reach so that theprocess can be completed within acertain time?

The effective pumping speed of a pumparrangement is understood as the actualpumping speed of the entire pump ar-rangement that prevails at the vessel.The nominal pumping speed of the pumpcan then be determined from the effectivepumping speed if the flow resistance(conductances) of the baffles, cold traps,filters, valves, and tubulations installedbetween the pump and the vessel areknown (see Sections 1.5.2 to 1.5.4). In thedetermination of the required nominalpumping speed it is further assumed thatthe vacuum system is leaktight; therefore,the leak rate must be so small that gasesflowing in from outside are immediatelyremoved by the connected pump arrange-ment and the pressure in the vessel doesnot alter (for further details, see Section5). The questions listed above under 1., 2.and 3. are characteristic for the three mostessential exercises of vacuum technology

1. Evacuation of the vessel to reach aspecified pressure.

2. Pumping of continuously evolving

quantities of gas and vapor at a certainpressure.

3. Pumping of the gases and vaporsproduced during a process by variationof temperature and pressure.

Initial evacuation of a vacuum chamber isinfluenced in the medium-, high-, andultrahigh vacuum regions by continuallyevolving quantities of gas, because inthese regions the escape of gases andvapors from the walls of the vessel is sosignificant that they alone determine thedimensions and layout of the vacuumsystem.

2.3.1 Evacuation of a vacuumchamber (withoutadditional sources of gasor vapor)

Because of the factors described above, anassessment of the pump-down time mustbe basically different for the evacuation ofa container in the rough vacuum regionfrom evacuation in the medium- and highvacuum regions.

2.3.1.1 Evacuation of a chamber in therough vacuum region

In this case the required effective pumpingspeed Seff, of a vacuum pump assembly isdependent only on the required pressurep, the volume V of the container, and thepump-down time t.

With constant pumping speed Seff andassuming that the ultimate pressure pendattainable with the pump arrangement issuch that pend << p, the decrease with timeof the pressure p(t) in a chamber is givenby the equation:

(2.32)

Beginning at 1013 mbar at time t = 0, theeffective pumping speed is calculateddepending on the pump-down time t fromequation (2.32) as follows:

(2.33a)

(2.33b)`np Seff

Vt

1013= − ·

dpp

p Seff

V t1013

∫ = − ·

− = ·dpdt

Seff

Vp

(2.34)

Introducing the dimensionless factor

(2.34a)

into equation (2.34), the relationshipbetween the effective pumping speed Seff,and the pump-down time t is given by

(2.35)

The ratio V/Seff is generally designated asa time constant τ. Thus the pump-downtime of a vacuum chamber fromatmospheric pressure to a pressure p isgiven by:

t = τ · s (2.36)

with

and

The dependence of the factor from thedesired pressure is shown in Fig. 2.75. Itshould be noted that the pumping speed ofsingle-stage rotary vane and rotary pistonpumps decreases below 10 mbar with gasballast and below 1 mbar without gasballast. This fundamental behavior isdifferent for pumps of various sizes and

σ = n p1013

τ = VSeff

SeffVt

= ·σ

σ= = ·`n p p1013 2 3 1013. log

SeffVt

np

Vt p= · = · ·`

1013 2 3 1013. log

Fig. 2.75 Dependency of the dimensionless factor s forcalculation of pumpdown time t according toequation 2.36. The broken line applies tosingle-stage pumps where the pumpingspeed decreases below 10 mbar

Dimensionless factor σ

Pres

sure

→

D00 E 19.06.2001 21:38 Uhr Seite 62



types but should not be ignored in thedetermination of the dependence of thepump-down time on pump size. It must bepointed out that the equations (2.32 to2.36) as well Fig. 2.75 only apply when theultimate pressure attained with the pumpused is by several orders of magnitudelower than the desired pressure.

Example: A vacuum chamber having avolume of 500 l shall be pumped down to1 mbar within 10 minutes. What effectivepumping speed is required?

500 l = 0.5 m3; 10 min = 1/6 h

According to equation (2.34) it followsthat:

For the example given above one reads offthe value of 7 from the straight line in Fig.2.75. However, from the broken line avalue of 8 is read off. According toequation (2.35) the following is obtained:

under consideration of the fact that thepumping speed reduces below 10 mbar.The required effective pumping speed thusamounts to about 24 m3/h.

2.3.1.2 Evacuation of a chamber in thehigh vacuum region

It is considerably more difficult to givegeneral formulas for use in the high vacu-um region. Since the pumping time toreach a given high vacuum pressuredepends essentially on the gas evolutionfrom the chamber’s inner surfaces, thecondition and pre-treatment of thesesurfaces are of great significance invacuum technology. Under no circums-tances should the material used exhibitporous regions or – particularly withregard to bake-out – contain cavities; theinner surfaces must be as smooth aspossible (true surface = geometric sur-face) and thoroughly cleaned (anddegreased). Gas evolution varies greatlywith the choice of material and the surface

Seff m h or= · =0 516

7 21 3. /

Seff m h= · =0 516

8 24 3. /

Seff

m h

= · ·

= · · =

0 5 2 3 10131

3 2 3 3 01 20 8 31 6

. . log

. . . //

condition. Useful data are collected inTable X (Section 9). The gas evolution canbe determined experimentally only fromcase to case by the pressure-rise method:the system is evacuated as thoroughly aspossible, and finally the pump and thechamber are isolated by a valve. Now thetime is measured for the pressure withinthe chamber (volume V) to rise by acertain amount, for example, a power of10. The gas quantity Q that arises per unittime is calculated from:

(2.37)

(∆p = measured pressure rise )

The gas quantity Q consists of the sum ofall the gas evolution and all leaks possiblypresent. Whether it is from gas evolutionor leakage may be determined by thefollowing method:

The gas quantity arising from gas evolu-tion must become smaller with time, thequantity of gas entering the system fromleakage remains constant with time.Experimentally, this distinction is notalways easily made, since it often takes aconsiderable length of time – with puregas evolution – before the measured pres-sure-time curve approaches a constant (oralmost a constant) final value; thus thebeginning of this curve follows a straightline for long times and so simulatesleakage (see Section 5, Leaks and LeakDetection).

If the gas evolution Q and the requiredpressure pend are known, it is easy to de-termine the necessary effective pumpingspeed:

(2.38)

Example: A vacuum chamber of 500 l mayhave a total surface area (including allsystems) of about 5 m2. A steady gasevolution of 2 · 10-4 mbar · l/s is assumedper m2 of surface area. This is a levelwhich is to be expected when valves orrotary feedthroughs, for example areconnected to the vacuum chamber. Inorder to maintain in the system a pressureof 1 · 10-5 mbar, the pump must have apumping speed of

Seffmbar smbar

s= · · – ·· – =5 2 10 4

1 10 5100`

`/ /

SeffQ

pend=

Q p Vt

= ·∆

A pumping speed of 100 l/s alone isrequired to continuously pump away thequantity of gas flowing in through theleaks or evolving from the chamber walls.Here the evacuation process is similar tothe examples given in Sections 2.3.1.1.However, in the case of a diffusion pumpthe pumping process does not begin atatmospheric pressure but at the forevacu-um pressure pV instead. Then equation(2.34) transforms into:

At a backing pressure of pV = 2 · 10-3 mbar“compression” K is in our example:

In order to attain an ultimate pressure of 1 · 10-5 mbar within 5 minutes afterstarting to pump with the diffusion pumpan effective pumping speed of

is required. This is much less compared tothe effective pumping speed needed tomaintain the ultimate pressure. Pump-down time and ultimate vacuum in thehigh vacuum and ultrahigh vacuum rangesdepends mostly on the gas evolution rateand the leak rates. The underlying mathe-matical rules can not be covered here. Forthese please refer to books specializing onthat topic.

2.3.1.3 Evacuation of a chamber in themedium vacuum region

In the rough vacuum region, the volume ofthe vessel is decisive for the time involvedin the pumping process. In the high andultrahigh vacuum regions, however, thegas evolution from the walls plays asignificant role. In the medium vacuumregion, the pumping process is influencedby both quantities. Moreover, in themedium vacuum region, particularly withrotary pumps, the ultimate pressure pendattainable is no longer negligible. If thequantity of gas entering the chamber isknown to be at a rate Q (in millibars literper second) from gas evolution from thewalls and leakage, the differential equation(2.32) for the pumping process becomes

Seff s=·

· · ≈5005 60

2 3 200 9. log `

K = ⋅ –

⋅ – =2 10 3

1 10 5 200

SeffVt p

Vt KpV= · = ·`n `n

D00

D00 E 19.06.2001 21:38 Uhr Seite 63



(2.39)

Integration of this equation leads to

(2.40)

wherep0 is the pressure at the beginning of the

pumping processp is the desired pressure

In contrast to equation 2.33b this equationdoes not permit a definite solution for Seff,therefore, the effective pumping speed fora known gas evolution cannot be deter-mined from the time – pressure curvewithout further information.In practice, therefore, the followingmethod will determine a pump withsufficiently high pumping speed:a) The pumping speed is calculated from

equation 2.34 as a result of the volumeof the chamber without gas evolutionand the desired pump-down time.

b The quotient of the gas evolution rateand this pumping speed is found. Thisquotient must be smaller than therequired pressure; for safety, it must beabout ten times lower. If this conditionis not fulfilled, a pump with corres-pondingly higher pumping speed mustbe chosen.

2.3.2 Determination of asuitable backing pump

The gas or vapor quantity transportedthrough a high vacuum pump must also behandled by the backing pump. Moreover,in the operation of the high vacuum pump(diffusion pump, turbomolecular pump),the maximum permissible backingpressure must never, even for a short time,be exceeded. If Q is the effective quantityof gas or vapor, which is pumped by thehigh vacuum pump with an effectivepumping speed Seff at an inlet pressure pA,this gas quantity must certainly be trans-ported by the backing pump at a pumpingspeed of SV at the backing pressure pV. Forthe effective throughput Q, the continuityequation applies:

Q = pA · Seff = pv · SV (2.41)

t VS

eff

po

pend Q Seff

p pend

Q Seff

=− −

− −

n/

/

dpdt

Seff p pend

Q

V= −

− −

The required pumping speed of thebacking pump is calculated from:

(2.41a)

Example: In the case of a diffusion pumphaving a pumping speed of 400 l/s theeffective pumping speed is 50 % of thevalue stated in the catalog when using ashell baffle. The max. permissible backingpressure is 2 · 10-1 mbar. The pumpingspeed required as a minimum for thebacking pump depends on the intakepressure pA according to equation 2.41a.

At an intake pressure of pA = 1 · 10-2 mbarthe pumping speed for the high vacuumpump as stated in the catalog is about 100l/s, subsequently 50 % of this is 50 l/s.Therefore the pumping speed of thebacking pump must amount to at least

At an intake pressure of pA = 1 · 10-3 mbarthe pump has already reached its nominalpumping speed of 400 l/s; the effectivepumping speed is now Seff = 200 l/s; thusthe required pumping speed for thebacking pump amounts to

If the high vacuum pump is to be used forpumping of vapors between 10-3 and 10-2 mbar, then a backing pump offering anominal pumping speed of 12 m3/h mustbe used, which in any case must have apumping speed of 9 m3/h at a pressure of2 · 10-1 mbar. If no vapors are to bepumped, a single-stage rotary vane pumpoperated without gas ballast will do inmost cases. If (even slight) components ofvapor are also to be pumped, one shouldin any case use a two-stage gas ballastpump as the backing pump which offers –also with gas ballast – the requiredpumping speed at 2 · 10-1 mbar.

If the high vacuum pump is only to beused at intake pressures below 10-3 mbar,a smaller backing pump will do; in the caseof the example given this will be a pumpoffering a pumping speed of 6 m3/h. If thecontinuous intake pressures are evenlower, below 10-4 mbar, for example, therequired pumping speed for the backingpump can be calculated from equation2.41a as:

SV s m h= · –

· – · = =1 10 3

2 10 1200 1 3 6 3/ . /

hSV s m= · –

· – · = =1 10 2

2 10 150 2 5 9 3. / /`

SV

pApV

Seff= · Theoretically in this case a smaller backingpump having a pumping speed of about 1m3/h could be used. But in practice alarger backing pump should be installedbecause, especially when starting up avacuum system, large amounts of gas mayoccur for brief periods. Operation of thehigh vacuum pump is endangered if thequantities of gas can not be pumped awayimmediately by the backing pump. If oneworks permanently at very low inlet pres-sures, the installation of a ballast volume(backing-line vessel or surge vessel)between the high vacuum pump and thebacking pump is recommended. Thebacking pump then should be operated forshort times only. The maximum admis-sible backing pressure, however, mustnever be exceeded. The size of the ballast volume depends onthe total quantity of gas to be pumped perunit of time. If this rate is very low, the ruleof thumb indicates that 0.5 l of ballastvolume allows 1 min of pumping time withthe backing pump isolated.

For finding the most adequate size ofbacking pump, a graphical method may beused in many cases. In this case thestarting point is the pumping speed cha-racteristic of the pumps according toequation 2.41.

The pumping speed characteristic of apump is easily derived from the measuredpumping speed (volume flow rate)characteristic of the pump as shown for a6000 l/s diffusion pump (see curve S inFig. 2.76). To arrive at the throughputcharacteristic (curve Q in Fig. 2.76), onemust multiply each ordinate value of S byits corresponding pA value and plottedagainst this value. If it is assumed that theinlet pressure of the diffusion pump doesnot exceed 10-2 mbar, the maximumthroughput is 9.5 mbar · l/s

Hence, the size of the backing pump mustbe such that this throughput can behandled by the pump at an intake pressure(of the backing pump) that is equal to orpreferably lower than the maximumpermissible backing pressure of the dif-fusion pump; that is, 4 · 10-1 mbar for the6000 l/s diffusion pump.

After accounting for the pumping speed

SV s m h= · –

· – · = =1 10 4

2 10 1200 0 1 0 36 3. / . /`

D00 E 19.06.2001 21:38 Uhr Seite 64



characteristics of commercially availabletwo-stage rotary plunger pumps, thethroughput characteristic for each pump iscalculated in a manner similar to that usedto find the Q curve for the diffusion pumpin Fig. 2.76 a. The result is the group of Qcurves numbered 1 – 4 in Fig. 2.76 b,whereby four 2-stage rotary-plungerpumps were considered, whose nominalspeeds were 200, 100, 50, and 25 m3/h,respectively. The critical backing pressureof the 6000 l/s diffusion pump is marked asV.B. (p = 4 · 10-1 mbar). Now the maximumthroughput Q = 9.5 mbar · l/s is shown ashorizontal line a. This line intersects thefour throughput curves. Counting fromright to left, the first point of intersectionthat corresponds to an intake pressurebelow the critical backing pressure of 4 · 10-1 mbar is made with throughputcharacteristic 2. This corresponds to thetwo-stage rotary plunger pump with anominal pumping speed of 100 m3/h.Therefore, this pump is the correct backingpump for the 6000 l/s diffusion pumpunder the preceding assumption.

However, if the pumping process is suchthat the maximum throughput of 9.5 mbar · l/s is unlikely, a smaller backingpump can, of course, be used. This is self-explanatory, for example, from line b inFig. 2.76 b, which corresponds to amaximum throughput of only 2 mbar l/s.In this case a 25 m3/h two-stage rotary-plunger pump would be sufficient.

2.3.3 Determination of pump-down time fromnomograms

In practice, for instance, when estimatingthe cost of a planned vacuum plant, cal-culation of the pump-down time from theeffective pumping speed Seff, the requiredpressure p, and the chamber volume V byformulas presented would be too trouble-some and time-consuming. Nomogramsare very helpful here. By using the nomo-gram in Fig. 9.7 in Section 9, one canquickly estimate the pump-down time forvacuum plants evacuated with rotarypumps, if the pumping speed of the pumpconcerned is fairly constant through thepressure region involved. By studying theexamples presented, one can easily under-stand the application of the nomogram. The pump-down times of rotary vane androtary piston pumps, insofar as the pum-ping speed of the pump concerned is con-stant down to the required pressure, canbe determined by reference to example 1.

In general, Roots pumps do not haveconstant pumping speeds in the workingregion involved. For the evaluation of thepump-down time, it usually suffices toassume the mean pumping speed.Examples 2 and 3 of the nomogram show,in this context, that for Roots pumps, thecompression ratio K refers not to theatmospheric pressure (1013 mbar), but to

the pressure at which the Roots pump isswitched on.

In the medium vacuum region, the gasevolution or the leak rate becomessignificantly evident. From the nomogram9.10 in Section 9, the correspondingcalculations of the pump-down time in thisvacuum region can be approximated.

In many applications it is expedient torelate the attainable pressures at any giventime to the pump-down time. This is easilypossible with reference to the nomogram9.7 in Section 9.

As a first example the pump-down cha-racteristic – that is, the relationship pres-sure p (denoted as desired pressure pend)versus pumping time tp – is derived fromthe nomogram for evacuating a vessel of 5 m3 volume by the single-stage rotaryplunger pump E 250 with an effectivepumping speed of Seff = 250 m3/h and anultimate pressure pend,p = 3 · 10-1 mbarwhen operated with a gas ballast and atpend,p = 3 · 10-2 mbar without a gas ballast.The time constant τ = V / Seff (see equation2.36) is the same in both cases andamounts as per nomogram 9.7 to about 70 s (column 3). For any given value ofpend > pend,p the straight line connectingthe “70 s point” on column 3 with the (pend – pend,p) value on the right-handscale of column 5 gives the correspondingtp value. The results of this procedure areshown as curves a and b in Fig. 2.77.

D00Fig. 2.76 Diagram for graphically determining a suitable backing pump

Intake pressure pa [mbar]

Thro

ughp

ut Q

[mba

r·l·

s–1]

Q [m

bar·

l·s–1

]

Pum

ping

spe

ed S

[l·s

–1]

a) Pumping speed characteristic of a 6000 l/s diffusion pump b) Series of throughput curves for two-stage rotaryplunger pumps (V.B. = critical forevacuum pressure)

D00 E 19.06.2001 21:38 Uhr Seite 65



It is somewhat more tedious to determinethe (pend,tp) relationship for a combinationof pumps. The second example discussedin the following deals with evacuating avessel of 5 m3 volume by the pumpcombination Roots pump WA 1001 andthe backing pump E 250 (as in thepreceding example). Pumping starts withthe E 250 pump operated without gasballast alone, until the Roots pump isswitched on at the pressure of 10 mbar. Asthe pumping speed characteristic of thecombination WA 1001/ E 250 – in contrastto the characteristic of the E 250 – is nolonger a horizontal straight line over thebest part of the pressure range (comparethis to the corresponding course of thecharacteristic for the combination WA

2001 / E 250 in Fig. 2.19), one introduces,as an approximation, average values ofSeff, related to defined pressure ranges. Inthe case of the WA 1001/ E 250combination the following average figuresapply:

Seff = 800 m3/h in the range 10 – 1 mbar,

Seff = 900 m3/h in the range 1 mbar to 5 · 10-2 mbar,

Seff = 500 m3/h in the range 5 · 10-2 to 5 · 10-3 mbar

The ultimate pressure of the combinationWA 1001 / E 250 is: Pend,p = 3 · 10-3 mbar.From these figures the corresponding timeconstants in the nomogram can bedetermined; from there, the pump-downtime tp can be found by calculating thepressure reduction R on the left side ofcolumn 5. The result is curve c in Fig. 2.77.

Computer aided calculations at LEYBOLDOf course calculations for our industrialsystems are performed by computerprograms. These require high per-formance computers and are thus usuallynot available for simple initial calculations.

2.3.4 Evacuation of a chamberwhere gases and vaporsare evolved

The preceding observations about thepump-down time are significantly altered ifvapors and gases arise during theevacuation process. With bake-out pro-cesses particularly, large quantities ofvapor can arise when the surfaces of thechamber are cleared of contamination. Theresulting necessary pump-down timedepends on very different parameters.Increased heating of the chamber walls isaccompanied by increased desorption ofgases and vapors from the walls. However,because the higher temperatures result inan accelerated escape of gases and vaporsfrom the walls, the rate at which they canbe removed from the chamber is also in-creased.

The magnitude of the allowable tempera-ture for the bake-out process in questionwill, indeed, be determined essentially bythe material in the chamber. Precise pump-down times can then be estimated bycalculation only if the quantity of the

evolving and pumped vapors is known.However, since this is seldom the caseexcept with drying processes, a quantita-tive consideration of this question isabandoned within the scope of thispublication.

2.3.5 Selection of pumps fordrying processes

Fundamentally, we must distinguish bet-ween short-term drying and drying pro-cesses that can require several hours oreven days. Independently of the durationof drying, all drying processes proceedapproximately as in Section 2.24

As an example of an application, the dryingof salt (short-term drying) is described,this being an already well-proven dryingprocess.

Drying of salt First, 400 kg of finely divided salt with awater content of about 8 % by mass is tobe dried in the shortest possible time(about 1 h) until the water content is lessthan 1 % by mass. The expected waterevolution amounts to about 28 kg. The saltin the chamber is continuously agitatedduring the drying process and heated toabout 80 °C. The vacuum system isschematically drawn in Fig. 2.78.

During the first quarter of drying time farmore than half the quantity of water vaporis evolved. Then the condenser is theactual main pump. Because of the highwater vapor temperature and, at thebeginning of the drying, the very highwater vapor pressure, the condensationefficiency of the condenser is significantlyincreased. In Fig. 2.78 it is understood thattwo parallel condensers each of 1 m2 con-densation surface can together condenseabout 15 l of water at an inlet pressure of100 mbar in 15 min. However, during thisinitial process, it must be ensured that thewater vapor pressure at the inlet port ofthe rotary piston pump does not exceed 60mbar (see Section 2.15 for further details).Since the backing pump has only to pumpaway the small part of the noncondensablegases at this stage, a single-stage rotarypiston pump TRIVAC S 65 B will suffice.With increasing process time, the watervapor evolution decreases, as does the

Fig. 2.77 Pumpdown time, tp, of a 5 m3 vessel using arotary plunger pump E 250 having a nominalpumping speed of 250 m3/h with (a) andwithout (b) gas ballast, as well asRoots/rotary plunger pump combination WA1001 / E250 for a cut-in pressure of 10 mbarfor the WA 1001 (e)

D00 E 19.06.2001 21:38 Uhr Seite 66



water vapor pressure in the condenser.After the water pressure in the chamberfalls below 27 mbar, the Roots pump (say,a Roots pump RUVAC WA 501) isswitched in. Thereby the water vapor ispumped more rapidly out of the chamber,the pressure increases in the condensers,and their condensation efficiency againincreases. The condensers are isolated bya valve when their water vapor reaches itssaturation vapor pressure. At this point,there is a water vapor pressure in thechamber of only about 4 mbar, andpumping is accomplished by the Rootspump with a gas ballast backing pumpuntil the water vapor pressure reachesabout 0.65 mbar. From experience it canbe assumed that the salt has now reachedthe desired degree of dryness.

Drying of paper If the pumps are to be of the correct sizefor a longer process run, it is expedient tobreak down the process run into charac-teristic sections. As an example, paperdrying is explained in the following wherethe paper has an initial moisture content of8 %, and the vessel has the volume V.

1. EvacuationThe backing pump must be suitably ratedwith regard to the volume of the vessel and

the desired pump-down time. This pump-down time is arranged according to thedesired process duration: if the process isto be finished after 12 – 15 h, the pump-down time should not last longer than 1 h.The size of the backing pump may beeasily calculated according to Section2.3.1.

2. PredryingDuring predrying – depending on thepressure region in which the work iscarried out – about 75 % of the moisture isdrawn off. This predrying should occupythe first third of the drying time. The rate atwhich predrying proceeds depends almostexclusively on the sufficiency of the heatsupply. For predrying 1 ton of paper in 5 h,60 kg of water must be evaporated; that is,an energy expenditure of about 40 kWh isneeded to evaporate water. Since the papermust be heated to a temperature of about120 °C at the same time, an average ofabout 20 kW must be provided. The meanvapor evolution per hour amounts to 12 kg. Therefore, a condenser with a capa-city of 15 kg/h should be sufficient. If thepaper is sufficiently preheated (perhaps byair-circulation drying) before evacuation,in the first hour of drying, double vaporevolution must be anticipated.

3. Main dryingIf, in the second stage, the pressure in afurther 5 h is to fall from 20 to about 5.3 mbar and 75 % of the total moisture(i.e., 19 % of the total moisture of 15 kg)is to be drawn off, the pump must, accor-ding to equations (2.37) and (2.38), have apumping speed of

According to equation 1.7, 15 kg of watervapor corresponds at 15 °C to a quantity ofwater vapor of

Hence the Roots pump RUVAC WA 1001would be the suitable pump. Thepermissible remaining moisture in theproduct determines the attainable ultimate

S m heff =·

=200005 5 3

750 3.

/

V p m R TM

mbar m subsequently

· = · · = · ·

≈

≈∆ 15 83 14 28818

20000 3

.

·

SeffV p

t p= ·

·∆

pressure. The relationship between theultimate pressure and the remainingmoisture is fixed for every product but dif-ferent from product to product. LEYBOLDhas many years of experience to its recordregarding applications in this area.Assume that a 0.1 % residual moisturecontent is required, for which the neces-sary ultimate pressure is 6 · 10-2 mbar.During the last 5 h the remaining 6 % ofthe moisture content, or 5 kg of water, isremoved. At a mean pressure of about0.65 mbar, 2000 m3/h of vapor is evolved.Two possibilities are offered: a) One continues working with the above-

mentioned Roots pump WA 1001. Theultimate total pressure settles at a valueaccording to the water vapor quantityevolving. One waits until a pressure ofabout 6.5 · 10-2 mbar is reached, whichnaturally takes a longer time.

b From the beginning, a somewhat largerRoots pump is chosen (e.g., the RUVAC WA 2001 with a pumping speedof 2000 m3/h is suitable). For largerquantities of paper (5000 kg, forexample) such a pumping system willbe suitable which at a pumping speedfor water vapor of up to 20,000 m3/hautomatically lowers the pressure from27 to 10-2 mbar. The entire time needfor drying is significantly reduced whenusing such pumps.

2.3.6 Flanges and their seals

In general, demountable joints in metallicvacuum components, pumps, valves,tubulations, and so on are provided withflanges. Vacuum components for rough,medium, and high vacuum from LEYBOLDare equipped with the following stan-dardized flange systems:

• Small flanges (KF) (quick-action con-nections to DIN 28 403) of nominalwidths 10, 16, 20, 25, 32, 40 and 50mm. The values 10, 16, 25 and 40 arepreferred widths according to thePNEUROP recommendations and theISO recommendations of the technicalcommittee ISO/TC 112 (see also Sec-tion 11). For a complete connection oftwo identical flanges one clamping ringand one centering ring are required.

• Clamp flanges (ISO-K) of nominalwidths 65, 100, 160, 250, 320, 400,

D00Fig. 2.78 Vacuum diagram for drying of salt. Pumpcombination consisting of Roots pump, con-denser and rotary plunger pump for stepwiseswitching of the pumping process (see text)

1 Vacuum chamber with salt filling2 RUVAC WA 5013 Condensers4 Throttle valve5 Rotary plunger pump

D00 E 19.06.2001 21:38 Uhr Seite 67



500 and 630 mm. Also, these flangescorrespond to the nominal widths andconstruction of the PNEUROP andISO/TC 112 recommendations. Clampflanges are joined together by clampsor collar rings. Centering rings or gas-kets are needed for sealing.

• Bolted flanges (ISO-F) for the samenominal widths as above (according toPNEUROP and ISO/TC 112). In specialcases bolted flanges having a smallernominal width are used. Clamp flangesand bolted flanges are in accordancewith DIN 28 404.

The nominal width is approximately equalto the free inner diameter of the flange inmillimeters; greater deviations are excep-tions, so the clamp flange DN 63 has aninner diameter of 70 mm. See also Table XIin Section 9).

High vacuum components are made ofaluminum or stainless steel. Stainlesssteel is slightly more expensive but offersa variety of advantages: lower degassingrate, corrosion resistant, can be degassedat temperatures up to 200 °C, metal sealsare possible and stainless steel is muchmore resistant to scratching compared toaluminum.

Ultrahigh vacuum components are madeof stainless steel and have CF flangesbakeable to a high temperature. Thesecomponents, including the flanges, aremanufactured in a series production,starting with a nominal width of 16 up to250 mm. CF flanges are available as fixedflanges or also with rotatable collarflanges. They may be linked with CONFLATflanges from almost all manufacturers.Copper gaskets are used for sealingpurposes.

Basically, the flanges should not besmaller than the connecting tubes and thecomponents that are joined to them. Whenno aggressive gases and vapors arepumped and the vacuum system is notexposed to a temperature above 80 °C,sealing with NBR (Perbunan) or CR (Neo-prene) flange O-rings is satisfactory forwork in the rough-, medium-, and highvacuum regions. This is often the casewhen testing the operation of vacuumsystems before they are finally assembled.

All stainless steel flanges may be degassedat temperatures up to 200 °C without

impairment. However, then Perbunansealing material is not suitable as a flangesealant. Rather, VITILAN® (a special FPM)sealing rings and also aluminum seals,which allow heating processes up to 150 °C and 200 °C respectively, should beused. After such degassing, pressuresdown to 10-8 mbar, i.e. down to the UHVrange, can be attained in vacuum systems.

Generating pressures below 10-8 mbarrequires higher bake-out temperatures. Asexplained above (see Section 2.2.6) workin the UHV range requires a basicallydifferent approach and the use of CFflanges fitted with metallic sealing rings.

2.3.7 Choice of suitable valves

Vacuum technology puts great demandson the functioning and reliability of thevalves, which are often needed in largenumbers in a plant. The demands arefulfilled only if correct shut-off devices areinstalled for each application, dependingon the method of construction, method ofoperation, and size. Moreover, in theconstruction and operation of vacuumplants, factors such as the flowconductance and leak-tightness of valvesare of great importance.

Valves are constructed so that they will notthrottle pumping speed. Hence, whenopened fully, their conductance in therough and medium vacuum regions equalsthat of corresponding tube components.For example, the conductance of a right-angle valve will equal the conductance of abent tube of the same nominal bore andangle. Similarly, the conductance of thevalve for molecular flow (i. e., in the highand ultrahigh vacuum regions), is so highthat no significant throttling occurs. Actualvalues for the conductance of variouscomponents are given in the catalog.

To meet stringent leak-tightness demands,high-quality vacuum valves are designedso that gas molecules adhering to thesurface of the valve shaft are nottransferred from the outer atmosphere intothe vacuum during operation. Such valvesare therefore equipped with metal bellowsfor isolating the valve shaft from theatmosphere, or alternatively, they are fullyencapsulated, that is, only static seals existbetween atmosphere and vacuum. Thisgroup is comprised of all medium and highvacuum valves from LEYBOLD that areoperated either manually or electro-pneumatically (Fig. 2.80) and (Fig. 2.79).The leak rate of these valves is less than10-9 mbar · l/s.

Valves sealed with oil or grease can be

Fig. 2.79 Right angle vacuum valve with solenoid actuator

1 Casing2 Valve disk

Fig. 2.80 Right angle vacuum valve with electropneumatic actuator

4 Compressed air supply5 Piston

3 Compression spring4 Solenoid coil

1 Casing2 Valve disk3 Bellows

D00 E 19.06.2001 21:38 Uhr Seite 68



used for highly stringent demands. Theirleakage rate is also about 10-9 mbar · l/s.However, a special case is the pendulum-type gate valve. Despite its grease-coveredseal, the leak rate between vacuum andexternal atmosphere is virtually the sameas for bellows-sealed valves because whenthe valve is in operation the shaft carriesout only a rotary motion so that no gasmolecules are transferred into the vacuum.Pendulum-type gate valves are not manu-factured by LEYBOLD.

For working pressures down to 10-7 mbar,valves of standard design suffice becausetheir seals and the housing materials aresuch that permeation and outgassing areinsignificant to the actual process. Ifpressures down to 10-9 mbar are required,baking up to 200 °C is usually necessary,which requires heat resistant sealingmaterials (e.g., VITILAN®) and materialsof high mechanical strength, with prepared(inner) surfaces and a low outgassing rate.Such valves are usually made of stainlesssteel. Flange connections are sealed withaluminum gaskets, so permeation pro-blems of elastomer seals are avoided. Inthe UHV range these issues are of specialsignificance so that mainly metallic sealsmust be used. The gas molecules bondedto the surface of the materials have, atpressures below 10-9 mbar, a very greatinfluence. They can only be pumped awaywithin a reasonable period of time bysimultaneous degassing. Degassingtemperatures up to 500 °C required in UHVsystems, pose special requirements on thesealing materials and the entire sealinggeometry. Gaskets made of gold or coppermust be used. The various applications require valveswith different drives, that is, valves that aremanually operated, electropneumatically-or magnetically-operated, and motordriven, such as variable-leak valves. Thevariety is even more enhanced by thevarious housing designs. In addition to thevarious materials used, right-angle andstraight-through valves are required.Depending on their nominal width andintended application, flanges fitted tovalves may be small (KF), clamp (ISO-K),bolted (ISO-F), or UHV (CF).

In addition to the vacuum valves, whichperform solely an isolation function (fullyopen – fully closed position), specialvalves are needed for special functions.

Typical are variable leak valves, whichcover the leakage range from 10-10 cm3/s(NTP) up to 1.6 · 103 cm3/s (NTP). Thesevalves are usually motor driven andsuitable for remote control and when theyare connected to a pressure gauge, theprocess pressures can be set andmaintained. Other special valves fulfillsafety functions, such as rapid, automaticcut-off of diffusion pumps or vacuumsystems in the event of a power failure. Forexample, SECUVAC valves belong to thisgroup. In the event of a power failure, theycut off the vacuum system from thepumping system and vent the forevacuumsystem. The vacuum system is enabledonly after a certain minimum pressure(about 200 mbar) has been attained oncethe power has been restored.

When aggressive gases or vapors have tobe pumped, valves made of stainless steeland sealed with VITILAN® sealant areusually used. For nuclear technology,valves have been developed that are sealedwith special elastomer or metal gaskets.We will be pleased to provide furtherdesign information for your area ofapplication upon request.

2.3.8 Gas locks and seal-offfittings

In many cases it is desirable not only to beable to seal off gas-filled or evacuated ves-sels, but also to be in a position to checkthe pressure or the vacuum in thesevessels at some later time and to post-eva-cuate or supplement or exchange the gasfilling.

This can be done quite easily with a seal-offfitting from LEYBOLD which is actuated viaa corresponding gas lock. The small flangeconnection of the evacuated or gas-filledvessel is hermetically sealed off within thetube by a small closure piece which formsthe actual valve. The gas lock required foractuation is removed after evacuation orfilling with gas. Thus one gas lock will do toactuate any number of seal-off fittings.Shown in Fig. 2.81 is a sectional view ofsuch an arrangement. Gas locks and seal-off fittings are manufactured by LEYBOLDhaving a nominal width of DN 16 KF, DN 25KF and DN 40 KF. They are made of stain-less steel. The leak rate of the seal-off fit-tings is less than 1 · 10-9 mbar l/s. Theycan sustain overpressures up to 2.5 bar,are temperature resistant up to 150 °C andmay be protected against dirt by a standardblank flange.Typical application examples are double-walled vessels with an insulating vacuum,like Dewar vessels, liquid gas vessels(tanks) or long distance energy pipelinesand many more. They are also used forevacuation or post-evacuation of referenceand support vacua in scientific instrumentsseal-off fittings with gas locks are oftenused. Previously it was necessary to have apump permanently connected in order topost-evacuate as required. Through theuse of gas locks with seal-off fittings avacuum-tight seal is provided for thevessel and the pump is only required fromtime to time for checking or post-evacuation.

D00

DN

a

h

a

h1 h2

d

Fig. 2.81 Gas lock with centering ring and seal-off fitting, sectional view

D00 E 19.06.2001 21:38 Uhr Seite 69

Vacuum MeasurementFundamentals of Vacuum Technology


3. Vacuum measurement,monitoring,control andregulation

The pressures measured in vacuum tech-nology today cover a range from 1013mbar to 10-12 mbar, i.e. over 15 orders ofmagnitude. The enormous dynamics invol-ved here can be shown through an analo-gy analysis of vacuum pressure measure-ment and length measurement, as depic-ted in Table 3.1.Measuring instruments designated asvacuum gauges are used for measurementin this broad pressure range. Since it isimpossible for physical reasons to build avacuum gauge which can carry out quanti-tative measurements in the entire vacuumrange, a series of vacuum gauges is avai-lable, each of which has a characteristicmeasuring range that usually extends overseveral orders of magnitude (see Fig.9.16a). In order to be able to allocate thelargest possible measuring ranges to theindividual types of vacuum gauges, oneaccepts the fact that the measurementuncertainty rises very rapidly, by up to 100 % in some cases, at the upper andlower range limits. This interrelationship isshown in Fig. 3.1 using the example of theVISCOVAC. Therefore, a distinction mustbe made between the measuring range asstated in the catalogue and the measuringrange for “precise” measurement. Themeasuring ranges of the individual vacu-um gauges are limited in the upper andlower range by physical effects.

3.1 Fundamentals oflow-pressure measurement

Vacuum gauges are devices for measuringgas pressures below atmospheric pressu-re (DIN 28 400, Part 3, 1992 issue). Inmany cases the pressure indicationdepends on the nature of the gas. Withcompression vacuum gauges it should benoted that if vapors are present, condensa-tion may occur due to the compression, asa result of which the pressure indication isfalsified. Compression vacuum gaugesmeasure the sum of the partial pressuresof all gas components that do not conden-se during the measurement procedure. Inthe case of mechanically compressingpumps, the final partial pressure can bemeasured in this way (see 1.1). Anotherway of measuring this pressure, is to free-ze out the condensable components in anLN2 cold trap. Exact measurement of par-tial pressures of certain gases or vapors iscarried out with the aid of partial pressuremeasuring instruments which operate onthe mass spectrometer principle (see sec-tion 4).

Dependence of the pressure indicationon the type of gasA distinction must be made between thefollowing vacuum gauges:1. Instruments that by definition measure

the pressure as the force which acts onan area, the so-called direct or absolu-te vacuum gauges. According to thekinetic theory of gases, this force,which the particles exert through theirimpact on the wall, depends only on thenumber of gas molecules per unit volu-me (number density of molecules n)and their temperature, but not on theirmolar mass. The reading of the measu-ring instrument is independent of thetype of gas. Such units include liquid-filled vacuum gauges and mechanicalvacuum gauges.

2. Instruments with indirect pressuremeasurement. In this case, the pressu-re is determined as a function of a pres-sure-dependent (or more accurately,density-dependent) property (thermalconductivity, ionization probability,electrical conductivity) of the gas.These properties are dependent on themolar mass as well as on the pressure.

Analogy analysisDetermination by Absolute

means of pressure Length

empirical worldof human beings 1 bar 1 m

simple measuring methods > 1 mbar > 1 mm

mechanicalmeasuring methods > 10–3 mbar > 1 mm

indirect methods 10–9 mbar ≈ 1/100

atom ∅

extreme indirect ≈ 0.18methods 10–12 mbar electron ∅

Table 3.1

“favorable measuring range”

(relative measurement uncertainty < 5%)

20

15

10

5

110–6 10–5 10–4 10–3 10–2 10–1 1

Rela

tive

mea

sure

men

t unc

erta

inty

(%)

Pressure (mbar)

Fig. 3.1 Measurement uncertainty distribution over the measuring range: VISCOVAC

D00 E 19.06.2001 21:38 Uhr Seite 70

Vacuum Measurement Fundamentals of Vacuum Technology


The pressure reading of the measuringinstrument depends on the type of gas.

The scales of these pressure measuringinstruments are always based on air ornitrogen as the test gas. For other gases orvapors correction factors, usually basedon air or nitrogen, must be given (seeTable 3.2). For precise pressure measure-ment with indirectly measuring vacuumgauges that determine the number densitythrough the application of electrical energy(indirect pressure measurement), it isimportant to know the gas composition. Inpractice, the gas composition is knownonly as a rough approximation. In manycases, however, it is sufficient to knowwhether light or heavy molecules predomi-nate in the gas mixture whose pressure isto be measured (e.g. hydrogen or pumpfluid vapor molecules).

Example: If the pressure of a gas essenti-ally consisting of pump fluid molecules ismeasured with an ionization vacuumgauge, then the pressure reading (applyingto air or N2), as shown in Table 3.2, is toohigh by a factor of about 10.

Measurement of pressures in the roughvacuum range can be carried out relativelyprecisely by means of vacuum gauges withdirect pressure measurement. Measure-ment of lower pressures, on the otherhand, is almost always subject to a num-ber of fundamental errors that limit themeasuring accuracy right from the start sothat it is not comparable at all to thedegree of accuracy usually achieved withmeasuring instruments. In order to mea-sure pressure in the medium and high

vacuum ranges with a measurementuncertainty of less than 50 %, the personconducting the experiment must proceedwith extreme care. Pressure measure-ments that need to be accurate to a fewpercent require great effort and, in general,the deployment of special measuringinstruments. This applies particularly to allpressure measurements in the ultrahighvacuum range (p < 10-7 mbar).

To be able to make a meaningful statementabout a pressure indicated by a vacuumgauge, one first has to take into account atwhat location and in what way the measu-ring system is connected. In all pressureareas where laminar flows prevail (1013 > p > 10-1 mbar), note must betaken of pressure gradients caused bypumping. Immediately in front of thepump (as seen from the vessel), a lowerpressure is created than in the vessel. Evencomponents having a high conductancemay create such a pressure gradient.Finally, the conductance of the connectingline between the vacuum system and themeasuring system must not be too smallbecause the line will otherwise be evacua-ted too slowly in the pressure region oflaminar flow so that the indicated pressureis too high.

The situation is more complicated in thecase of high and ultrahigh vacuum. According to the specific installation fea-tures, an excessively high pressure or, inthe case of well-degassed measuringtubes, an excessively low pressure may berecorded due to outgassing of the walls ofthe vacuum gauge or inadequate degas-sing of the measuring system. In high andultrahigh vacuum, pressure equalizationbetween the vacuum system and the mea-suring tubes may take a long time. If pos-sible, so-called nude gauges are used. Thelatter are inserted directly in the vacuumsystem, flange-mounted, without aconnecting line or an envelope. Specialconsideration must always be given to theinfluence of the measuring process itselfon the pressure measurement. For exam-ple, in ionization vacuum gauges that workwith a hot cathode, gas particles, especial-ly those of the higher hydrocarbons, arethermally broken down. This alters the gascomposition. Such effects play a role inconnection with pressure measurement inthe ultrahigh vacuum range. The sameapplies to gas clean-up in ionization vacu-um gauges, in particular Penning gauges

(of the order of 10-2 to 10-1 l/s). Contami-nation of the measuring system, interfe-ring electrical and magnetic fields, insula-tion errors and inadmissibly high ambienttemperatures falsify pressure measure-ment. The consequences of these avoida-ble errors and the necessary remedies areindicated in the discussion of the individu-al measuring systems and in summaryform in section 8.4.

Selection of vacuum gaugesThe desired pressure range is not the onlyfactor considered when selecting a suita-ble measuring instrument. The operatingconditions under which the gauge worksalso play an important role. If measure-ments are to be carried out under difficultoperating conditions, i.e. if there is a highrisk of contamination, vibrations in thetubes cannot be ruled out, air bursts canbe expected, etc., then the measuringinstrument must be robust. In industrialfacilities, Bourdon gauges, diaphragmvacuum gauges, thermal conductivityvacuum gauges, hot cathode ionizationvacuum gauges and Penning vacuum gau-ges are used. Some of these measuringinstruments are sensitive to adverse ope-rating conditions. They should and canonly be used successfully if the abovementioned sources of errors are excludedas far as possible and the operatinginstructions are followed.

D00

Given the presence Correction factor of predominantly based on N2

(type of gas) (nitrogen = 1)He 6.9Ne 4.35Ar 0.83Kr 0.59Xe 0.33Hg 0.303H2 2.4CO 0.92CO2 0.69CH4 0.8

higher hydrocarbons 0.1 – 0.4

Table 3.2 Correction factors

D00 E 19.06.2001 21:38 Uhr Seite 71



3.2 Vacuum gaugeswith pressure reading that isindependent of thetype of gas

Mechanical vacuum gauges measure thepressure directly by recording the forcewhich the particles (molecules and atoms)in a gas-filled space exert on a surface byvirtue of their thermal velocity.

3.2.1 Bourdon vacuum gauges

The interior of a tube bent into a circulararc (so-called Bourdon tube) (3) isconnected to the vessel to be evacuated(Fig. 3.2). Through the effect of the exter-nal air pressure the end of the tube isdeflected to a greater or lesser extentduring evacuation and the attached pointermechanism (4) and (2) is actuated. Sincethe pressure reading depends on the exter-nal atmospheric pressure, it is accurateonly to approximately 10 mbar, providedthat the change in the ambient atmosphe-ric pressure is not corrected.

3.2.2 Diaphragm vacuum gauges

3.2.2.1 Capsule vacuum gauges

The best-known design of a diaphragmvacuum gauge is a barometer with ananeroid capsule as the measuring system.It contains a hermetically sealed, evacu-ated, thin-walled diaphragm capsule madeof a copper-beryllium alloy. As the pressu-re drops, the capsule diaphragm expands.This movement is transmitted to a point bya lever system. The capsule vacuumgauge, designed according to this princi-ple, indicates the pressure on a linearscale, independent of the external atmos-pheric pressure.

3.2.2.2 DIAVAC diaphragm vacuumgauge

The most accurate pressure reading possi-ble is frequently required for levels below50 mbar. In this case, a different diaphragmvacuum gauge is more suitable, i.e. theDIAVAC, whose pressure scale is consider-ably extended between 1 and 100 mbar.The section of the interior in which thelever system (2) of the gauge head is loca-ted (see Fig. 3.3) is evacuated to a referen-ce pressure pref of less than 10-3 mbar.

The closure to the vessel is in the form of acorrugated diaphragm (4) of special steel.As long as the vessel is not evacuated, thisdiaphragm is pressed firmly against thewall (1). As evacuation increases, the diffe-rence between the pressure to be measu-red px and the reference pressure decrea-ses. The diaphragm bends only slightly atfirst, but then below 100 mbar to a greaterdegree. With the DIAVAC the diaphragmdeflection is again transmitted to a pointer(9). In particular the measuring range bet-ween 1 and 20 mbar is considerably exten-ded so that the pressure can be read quiteaccurately (to about 0.3 mbar). The sensi-tivity to vibration of this instrument issomewhat higher than for the capsulevacuum gauge.

3.2.2.3 Precision diaphragm vacuumgauges

A significantly higher measuring accuracythan that of the capsule vacuum gauge andthe DIAVAC is achieved by the precisiondiaphragm vacuum gauge. The design ofthese vacuum gauges resembles that ofcapsule vacuum gauges. The scale is line-ar. The obtainable degree of precision isthe maximum possible with present-daystate-of-the-art equipment. These instru-ments permit measurement of 10-1 mbarwith a full-scale deflection of 20 mbar. Thegreater degree of precision also means ahigher sensitivity to vibration.Capsule vacuum gauges measure pressu-re accurately to 10 mbar (due to the linearscale, they are least accurate at the lowpressure end of the scale). If only pressu-res below 30 mbar are to be measured, theDIAVAC is recommended because its rea-ding (see above) is considerably moreaccurate. For extremely precise measuringaccuracy requirements precision dia-phragm vacuum gauges should be used. Iflow pressures have to be measured accu-rately and for this reason a measuringrange of, for example, up to 20 mbar isselected, higher pressures can no longerbe measured since these gauges have alinear scale. All mechanical vacuum gau-ges are sensitive to vibration to someextent. Small vibrations, such as those thatarise in the case of direct connection to abacking pump, are generally not detrimen-tal.

Fig. 3.2 Cross-section of a Bourdon gauge

1 Connecting tube to connection flange

2 Pointer3 Bourdon tube4 Lever system

Fig. 3.3 Cross-section of DIAVAC DV 1000 diaphragm vacuum gauge

1 Base plate 2 Lever system 3 Connecting flange4 Diaphragm 5 Reference

pressure pref, 6 Pinch-off end

7 Mirror sheet8 Plexiglass sheet9 Pointer 10 Glass bett11 Mounting plate12 Housing

D00 E 19.06.2001 21:38 Uhr Seite 72



3.2.2.4 Capacitance diaphragm gauges

Deflection of a diaphragm can also be elec-trically measured as “strain” or as a chan-ge in capacitance. In the past, four straingauges, which change their resistancewhen the diaphragm is deflected, i.e.under tensile load, were mounted on ametallic diaphragm in a bridge circuit. AtLEYBOLD such instruments have beengiven a special designation, i.e. MEMBRANOVAC. Later, silicon dia-phragms that contained four such “strainresistances” directly on their surface wereused. The electrical arrangement againconsisted of a bridge circuit, and a con-stant current was fed in at two oppositecorner points while a linear voltage signalproportional to the pressure was picked upat the two other corner points. Fig. 3.4 illu-strates the principle of this arrangement.Such instruments were designated as PIEZOVAC units and are still in use inmany cases. Today the deflection of thediaphragm is measured as the change in

capacitance of a plate capacitor: one elec-trode is fixed, the other is formed by thediaphragm. When the diaphragm is deflec-ted, the distance between the electrodesand thus capacitance of the capacitor isaltered. Fig. 3.5 illustrates the principle ofthis arrangement. A distinction is madebetween sensors with metallic and thosewith ceramic diaphragms. The structure ofthe two types is similar and is shown onthe basis of two examples in Fig. 3.6.Capacitance diaphragm gauges are usedfrom atmospheric pressure to 1·10-3 mbar(from 10-4 mbar the measurement uncer-tainty rises rapidly). To ensure sufficientdeflection of the diaphragms at such lowpressures, diaphragms of varying thickn-esses are used for the various pressurelevels. In each case, the pressure can bemeasured with the sensors to an accuracyof 3 powers of ten:1013 to 1 mbar100 to 10–1 mbar10 to 10–2 mbar und1 to 10–3 mbar.

If the pressures to be measured exceedthese range limits, it is recommended thata multichannel unit with two or three sen-sors be used, possibly with automaticchannel changeover. The capacitance dia-phragm gauge thus represents, for allpractical purposes, the only absolute pres-sure measuring instrument that is inde-pendent of the type of gas and designedfor pressures under 1 mbar. Today twotypes of capacitive sensors are available:1) Sensors DI 200 and DI 2000 with

aluminum oxide diaphragms, which areparticularly overload-free, with theMEMBRANOVAC DM 11 and DM 12units.

2) Sensors with Inconel diaphragms CM 1,DM 10, CM 100, CM 1000 with extre-mely high resolution, with the DM 21and DM 22 units.

D00

Fig. 3.4 Piezoelectric sensor (basic diagram) Fig. 3.5 Capacitive sensor (basic diagram)

1 2

p1 p2

C1 C2

Fig. 3.6 Capacitive sensors (basic diagram)

Diaphragm(Inconel) Reference chamber closure

C ~ A/dC = capacitanceA = aread = distance C ~ A/d

C = capacitanceA = aread = distance

Diaphragm(ceramic)

Grid + reference chamberclosureCapacitor plates

Signal converter + amplifier

24 V DC, 4 – 20 mA

Sensor body (ceramic)

Electrode

Amplifier + 15 V DC

– 15 V DC

0 – 10 V

left: CAPACITRON (Inconel diaphragm) right: MEMBRANOVAC (aluminum oxide diaphragm)

D00 E 19.06.2001 21:38 Uhr Seite 73



3.2.3 Liquid-filled (mercury)vacuum gauges

3.2.3.1 U-tube vacuum gauges

U-tube vacuum gauges filled with mercuryare the simplest and most exact instru-ments for measuring pressure in the roughvacuum range (1013 to a few mbar).Unfortunately their use in technical plantsis limited because of their size and pro-neness to breakage (see 3.4.1a).

In the evacuated limb of the U-tube vacu-um gauge a constant pressure is maintai-ned equal to the vapor pressure of mercu-ry at room temperature (about 10-3 mbar).The other limb is connected to the volumein which the pressure is to be measured.From the difference in the levels of the twocolumns, the pressure to be measured canbe determined on the mbar scale provided.The reading is independent of the atmos-pheric pressure.

3.2.3.2 Compression vacuum gauges(according to McLeod)

The compression vacuum gauge develo-ped by McLeod in 1874 is a very rarelyused type of vacuum gauge today. In itsrefined form the instrument can be usedfor absolute pressure measurement in thehigh vacuum range down to 10-5 mbar. Inthe past it was frequently used as a refe-rence instrument for the calibration ofmedium and sometimes also of high vacu-um gauges. For such measurements,however, numerous precautionary ruleshad to be taken into account before it waspossible to assess the measuring accu-

racy. The pressure is measured by com-pressing a quantity of gas that initiallyoccupies a large volume into a smallervolume by raising a mercury level. Theincreased pressure obtained in this man-ner can be measured in the same way as ina U-tube manometer and from it the origi-nal pressure is calculated (see equationsbelow).

According to the type of scale division, adistinction is made between two forms ofcompression vacuum gauges: those with alinear scale (see Fig. 3.7) and those with asquare-law scale (see Fig. 3.8). In the caseof the compression vacuum gauges of theMcLeod linear-scale type, the ratio of theenclosed residual volume Vc to the totalvolume V must be known for each heightof the mercury level in the measurementcapillary; this ratio is shown on the scaleprovided with the instrument. In the caseof compression vacuum gauges with asquare-law scale, the total volume and thecapillary diameter d must be known.

Nowadays a “shortened” McLeod typecompression vacuum gauge according toKammerer is used to measure the “partialfinal pressure” of mechanically com-pressing pumps. Through the high degreeof compression the condensable gas com-ponents (vapors) are discharged as liquid(the volume of the same mass is thensmaller by a factor of around 105 and canbe neglected in the measurement) so thatonly the pressure of the permanentlygaseous components is measured (this iswhere the expression permanent gasescomes from).

Principle of measurement with compres-sion vacuum gaugesIf h is the difference in the mercury levelbetween the measurement capillary andthe reference capillary (measured in mm),then it follows from the Boyle-Mariottelaw:

p · V = (p + h) · Vc (3.1)

(3.1a)

p measured in mm of mercury (= torr). If Vc << V, then:

(3.1b)

Vc and V must be known, h is read off(linear scale).

These relationships remain unchanged ifthe difference in level is read off a scalewith mbar division. The pressure is thenobtained in mbar:

h in mm (3.1c)

If during measurement the mercury levelin the measurement capillary is always setso that the mercury level in the referencecapillary corresponds to the upper end ofthe measurement capillary (see Figs. 3.7and 3.8), the volume Vc is always given by:

(3.1d)

h ....difference in level, see Fig. 3.5d ....inside diameter of measurement

capillary

Vc h d= ⋅ ⋅π4

2

p h VcV

= ⋅ ⋅43

p h VcV

= ⋅

p h VcV Vc

= ⋅ −

Fig. 3.7 McLeod compression vacuum gauge with linear scale (equation 3.1b)

Volume V [cm3]

torr

Pres

sure

p mea

surin

gra

nge

Lower limit for:Vcmin. = 5 · 10–3 cm3

hmin. = 1 mm

Upper limit for:Vcmax. = 0.1 cm3

hmax. = 100 mm

Upper limit for:Vcmax. = 1 cm3

hmax. = 100 mm

mea

surin

gra

nge

Fig. 3.8 McLeod compression vacuum gauge with square-law scale (equation 3.1f)

Volume V [cm3]

torr

Lower limit for:d = 2.5 mm

Lower limit for:d = 1 mm

Upper limit for:d = 2.5 mm

Upper limit for:d = 1 mm

mea

surin

gra

nge

mea

surin

gra

nge

D00 E 19.06.2001 21:38 Uhr Seite 74



If this term is substituted for Vc in equati-on (3.1b), the result is:

(3.1e)

that is, a square-law scale in mm (torr) if dand V are measured in mm or mm3. If thescale is to be divided into mbar, then theequation is:

(3.1f)

where h in mmd in mm

and V in mm3

Compression vacuum gauges ensure areading of the sum of all partial pressuresof the permanent gases, provided that novapors are present that condense duringthe compression procedure.

The measuring range between the top andbottom ends is limited by the maximumand minimum ratios of the capillary volu-me to the total volume (see Figs. 3.7 and3.8). The accuracy of the pressure mea-surement depends to a great extent on thereading accuracy. By using a vernier andmirror, pressure measurements with anaccuracy of ± 2 % can be achieved. In thelow pressure range, where h is very small,this accuracy is no longer attainable, chief-ly because small geometric deviationshave a very noticeable effect at the closedend of the capillary (systematic error).

The presence of vapors that may conden-se during compression influences themeasurement, often in an indefinite man-ner. One can easily determine whethervapors having a pressure that is not negli-gible are present. This can be done by set-ting different heights h in the measure-ment capillary under constant pressurewhile using the linear scale and then cal-culating p according to equation 3.1b. If novapors are present, or only those whosepressure is negligible at room temperature(such as mercury), then the same value ofp must result for each h.

The scale of compression vacuum gaugescan be calculated from the geometricdimensions. This is why they were used inthe past by official calibration stations asnormal pressure (see equation 3.4.1a).

p h dV

= ⋅ ⋅23

2π

p h dV

= ⋅ ⋅24

2π

3.3 Vacuum gaugeswith gas-dependentpressure reading

This type of vacuum gauge does not mea-sure the pressure directly as an area-rela-ted force, but indirectly by means of otherphysical variables that are proportional tothe number density of particles and thus tothe pressure. The vacuum gauges withgas-dependent pressure reading include:the decrement gauge (3.3.1), the thermalconductivity vacuum gauge (3.3.2) and theionization vacuum gauge having differentdesigns (3.3.3).

The instruments consist of the actual sen-sor (gauge head, sensor) and the controlunit required to operate it. The pressurescales or digital displays are usually basedon nitrogen pressures; if the true pressurepT of a gas (or vapor) has to be determi-ned, the indicated pressure pI must bemultiplied by a factor that is characteristicfor this gas. These factors differ, depen-ding on the type of instrument, and are eit-her given in tabular form as factors inde-pendent of pressure (see Table 3.2) or, ifthey depend on the pressure, must bedetermined on the basis of a diagram (seeFig. 3.11).In general, the following applies:

True pressure pT = indicated pressure pI · correction factor

If the pressure is read off a “nitrogenscale” but not corrected, one refers to“nitrogen equivalent” values.

In all electrical vacuum gauges (they inclu-de vacuum gauges that are dependent onthe type of gas) the increasing use of com-puters has led to the wish to display thepressure directly on the screen, e.g. to ins-ert it at the appropriate place in a processflow diagram. To be able to use the moststandardized computer interfaces possi-ble, so-called transmitters (signal conver-ters with standardized current outputs) arebuilt instead of a sensor and display unit(e.g. THERMOVAC transmitter, Penningtransmitter, IONIVAC transmitter). Trans-mitters require a supply voltage (e.g. +24volts) and deliver a pressure-dependentcurrent signal that is linear over the entiremeasuring range from 4 to 20 mA or 0 – 10 V. The pressure reading is not pro-

vided until after supply of this signal to thecomputer and processing by the appro-priate software and is then displayed direc-tly on the screen.

3.3.1 Spinning rotor gauge(SRG) (VISCOVAC)

Pressure-dependent gas friction at low gaspressures can be utilized to measure pres-sures in the medium and high vacuumrange. In technical instruments of this kinda steel ball that has a diameter of severalmillimeters and is suspended withoutcontact in a magnetic field (see Fig. 3.9) isused as the measuring element. The ball isset into rotation through an electro-magnetic rotating field: after reaching astarting speed (around 425 Hz), the ball isleft to itself. The speed then declines at arate that depends on the prevailing pressu-re under the influence of the pressure-dependent gas friction. The gas pressureis derived from the relative decline of thespeed f (slowing down) using the follo-wing equation:

(3.2)

p = gas pressurer = radius of the ball ρ = density of the ball materialc- = mean speed of the gas particles,

dependent on type of gasσ = coefficient of friction of the ball, inde-

pendent of the type of gas, nearly 1.

−− ⋅ = ⋅ ⋅⋅ ⋅

fdfdt

pc r

10π

σρ

D00Fig. 3.9 Cross-section of the gauge head of a

VISCOVAC VM 212 spinning rotor gauge(SRG)

1 Ball2 Measuring tube,

closed at one end,welded into connection flange 7

3 Permanent magnets4 Stabilization coils5 4 drive coils6 Bubble level7 Connection flange

D00 E 19.06.2001 21:38 Uhr Seite 75



As long as a measurement uncertainty of 3 % is sufficient, which is usually the case,one can apply σ = 1 so that the sensitivityof the spinning rotor gauge (SRG) withrotating steel ball is given by the calculablephysical size of the ball, i.e. the productradius x density r · ρ (see equation 3.2).Once a ball has been “calibrated”, it is sui-table for use as a “transfer standard”, i.e.as a reference device for calibrating ano-ther vacuum gauge through comparison,and is characterized by high long-term sta-bility. Measurements with the VISCOVACare not limited to measurement of thepressure, however. Other variables invol-ved in the kinetic theory of gases, such asmean free path, monolayer time, particlenumber density and impingement rate, canalso be measured. The instrument permitsstorage of 10 programs and easy change-over between these programs. The measu-ring time per slowing-down operation isbetween 5 seconds for high pressures andabout 40 seconds for lower pressures. Themeasurement sequence of the instrumentis controlled fully automatically by amicroprocessor so that a new value is dis-played after every measurement (slowing-down procedure). The programs additio-nally enable calculation of a number of sta-tistical variables (arithmetic mean, stan-dard deviation) after a previously specifiednumber of measurements.

While in the case of the kinetic theory ofgases with VISCOVAC the counting of par-ticles directly represents the measuringprinciple (transferring the particle pulsesto the rotating ball, which is thus sloweddown). With other electrical measuringmethods that are dependent on the type ofgas, the particle number density is measu-red indirectly by means of the amount ofheat lost through the particles (thermalconductivity vacuum gauge) or by meansof the number of ions formed (ionizationvacuum gauge).

3.3.2 Thermal conductivityvacuum gauges

Classical physics teaches and providesexperimental confirmation that the thermalconductivity of a static gas is independentof the pressure at higher pressures (par-ticle number density), p > 1 mbar. At lowerpressures, p < 1 mbar, however, the ther-mal conductivity is pressure-dependent(approximately proportional 1 / ADM). It de-creases in the medium vacuum range star-ting from approx. 1 mbar proportionally tothe pressure and reaches a value of zero inthe high vacuum range. This pressuredependence is utilized in the thermal con-ductivity vacuum gauge and enables preci-se measurement (dependent on the type ofgas) of pressures in the medium vacuumrange.

The most widespread measuring instru-ment of this kind is the Pirani vacuumgauge. A current-carrying filament with aradius of r1 heated up to around 100 to150 °C (Fig. 3.10) gives off the heat gene-rated in it to the gas surrounding itthrough radiation and thermal conduction(as well as, of course, to the supports atthe filament ends). In the rough vacuumrange the thermal conduction through gasconvection is virtually independent ofpressure (see Fig. 3.10). If, however, at afew mbar, the mean free path of the gas isof the same order of magnitude as the fila-ment diameter, this type of heat transferdeclines more and more, becoming depen-

dent on the density and thus on the pres-sure. Below 10-3 mbar the mean free pathof a gas roughly corresponds to the size ofradius r2 of the measuring tubes. The sen-sing filament in the gauge head forms abranch of a Wheatstone bridge. In theTHERMOTRON thermal conductivity gau-ges with variable resistance which werecommonly used in the past, the sensingfilament was heated with a constant cur-rent. As gas pressure increases, the tem-perature of the filament decreases becauseof the greater thermal conduction throughthe gas so that the bridge becomes out ofbalance. The bridge current serves as ameasure for the gas pressure, which isindicated on a measuring scale. In theTHERMOVAC thermal conductivity gau-ges with constant resistance which arealmost exclusively built today, the sensingfilament is also a branch of a Wheatstonebridge. The heating voltage applied to thisbridge is regulated so that the resistanceand therefore the temperature of the fila-ment remain constant, regardless of theheat loss. This means that the bridge isalways balanced. This mode of regulationinvolves a time constant of a few millise-conds so that such instruments, in con-trast to those with variable resistance, res-pond very quickly to pressure changes.The voltage applied to the bridge is a mea-sure of the pressure. The measuring volta-ge is corrected electronically such that anapproximately logarithmic scale is obtai-ned over the entire measuring range. Ther-

l ‹‹ r1

-l r# 2

-

l ›› r1

-

r1

r2 l ‹‹ r – r2 1

-

I II III

10–5 10–4 10–3 10–2 10–1

Druck [mbar]

Abge

führ

ter W

ärm

eflu

ß

100101

Fig. 3.10 Dependence of the amount heat dissipated by a heated filament (radius r1) in a tube (radius r2) at a constanttemperature difference on the gas pressure (schematic diagram)

Pressure [mbar]

Heat

loss

I Thermal dissipation due to radiation and conduction in the metallic endsII Thermal dissipation due to the gas, pressure-dependentIII Thermal dissipation due to radiation and convection

D00 E 19.06.2001 21:38 Uhr Seite 76



mal conductivity vacuum gauges with con-stant resistance have a measuring rangefrom 10-4 to 1013 mbar. Due to the veryshort response time, they are particularlysuitable for controlling and pressure moni-toring applications (see section 3.5). Themeasurement uncertainty varies in the dif-ferent pressure ranges. The maximumerror at full-scale deflection is about 1 to 2 %. In the most sensitive range, i.e. bet-ween 10-3 and 1 mbar, this corresponds toaround 10 % of the pressure reading. Themeasurement uncertainty is significantlygreater outside this range.

As in all vacuum gauges dependent on thetype of gas, the scales of the indicatinginstruments and digital displays in thecase of thermal conductivity vacuum gau-ges also apply to nitrogen and air. Withinthe limits of error, the pressure of gaseswith similar molecular masses, i.e. O2, COand others, can be read off directly. Cali-bration curves for a series of gases areshown in Fig. 3.11.

An extreme example of the discrepancybetween true pressure pT and indicatedpressure pI in pressure measurement isthe admission of air to a vacuum systemwith argon from a pressure cylinder toavoid moisture (pumping time). Accordingto Fig. 3.11, one would obtain a pI readingof only 40 mbar on reaching an “Ar atmos-

pheric pressure” pT with a THERMOVAC asa pressure measuring instrument. Argonmight escape from the vessel (coveropens, bell jar rises). For such and similarapplications, pressure switches or vacuumgauges that are independent of the type ofgas must be used (see section 3.2).

3.3.3 Ionization vacuum gauges

Ionization vacuum gauges are the mostimportant instruments for measuring gaspressures in the high and ultrahigh vacu-um ranges. They measure the pressure interms of the number density of particlesproportional to the pressure. The gaswhose pressure is to be measured entersthe gauge heads of the instruments and ispartially ionized with the help of an electricfield. Ionization takes place when electronsare accelerated in the electric field andattain sufficient energy to form positiveions on impact with gas molecules. Theseions transmit their charge to a measuringelectrode (ion collector) in the system. Theion current, generated in this manner (or,more precisely, the electron current in thefeed line of the measuring electrode that isrequired to neutralize these ions) is a mea-sure of the pressure because the ion yield

is proportional to the particle number den-sity and thus to the pressure.

The formation of ions is a consequence ofeither a discharge at a high electric fieldstrength (so-called cold-cathode or Pen-ning discharge, see 3.3.3.1) or the impactof electrons that are emitted from a hotcathode (see 3.3.3.2).

Under otherwise constant conditions, theion yield and thus the ion current dependon the type of gas since some gases areeasier to ionize than others. As all vacuumgauges with a pressure reading that isdependent on the type of gas, ionizationvacuum gauges are calibrated with nitro-gen as the reference gas (nitrogen equiva-lent pressure, see 3.3). To obtain the truepressure for gases other than nitrogen, theread-off pressure must be multiplied bythe correction factor given in Table 3.2 forthe gas concerned. The factors stated inTable 3.2 are assumed to be independentof the pressure, though they dependsomewhat on the geometry of the electro-de system. Therefore, they are to be regar-ded as average values for various types ofionization vacuum gauges (see Fig. 3.16).

3.3.3.1 Cold-cathode ionization vacuum gauges (Penningvacuum gauges)

Ionization vacuum gauges which operatewith cold discharge are called cold-catho-de- or Penning vacuum gauges. Thedischarge process in a measuring tube is,in principle, the same as in the electrodesystem of a sputter ion pump (see section2.1.8.3). A common feature of all types ofcold-cathode ionization vacuum gauges isthat they contain just two unheatedelectrodes, a cathode and an anode, bet-ween which a so-called cold discharge isinitiated and maintained by means of a d.c.voltage (of around 2 kV) so that thedischarge continues at very low pressures.This is achieved by using a magnetic fieldto make the paths of the electrons longenough so that the rate of their collisionwith gas molecules is sufficiently large toform the number of charge carriers requi-red to maintain the discharge. The magne-tic field (see Fig. 3.12) is arranged suchthat the magnetic field lines of force crossthe electric field lines. In this way the elec-trons are confined to a spiral path. Thepositive and negative charge carriers pro-

D00Fig. 3.11 Calibration curves of THERMOVAC gauges for various gases, based on nitrogen equivalent reading

Indicated pressure [mbar] pI

True

pre

ssur

e [m

bar]

pW

pI > pT

pI < pT

N2, O2 air – pI = pT

D00 E 19.06.2001 21:38 Uhr Seite 77



duced by collision move to the correspon-ding electrodes and form the pressure-dependent discharge current, which isindicated on the meter. The reading inmbar depends on the type of gas. Theupper limit of the measuring range is givenby the fact that above a level of several 10-2 mbar the Penning discharge changesto a glow discharge with intense light out-put in which the current (at constant volta-ge) depends only to a small extent on thepressure and is therefore not suitable formeasurement purposes. In all Penninggauges there is considerably higher gassorption than in ionization vacuum gaugesthat operate with a hot cathode. A Penningmeasuring tube pumps gases similarly to asputter ion pump (S ≈ 10-2 l/s). Here againthe ions produced in the discharge areaccelerated towards the cathode wherethey are partly retained and partly causesputtering of the cathode material. Thesputtered cathode material forms a gette-ring surface film on the walls of the gaugetube. In spite of these disadvantages,which result in a relatively high degree ofinaccuracy in the pressure reading (up toaround 50 %), the cold-cathode ionizationgauge has three very outstanding advanta-ges. First, it is the least expensive of allhigh vacuum measuring instruments.Second, the measuring system is insensi-

tive to the sudden admission of air and tovibrations; and third, the instrument iseasy to operate.

3.3.3.2 Hot-cathode ionization vacuumgauges

Generally speaking, such gauges refer tomeasuring systems consisting of threeelectrodes (cathode, anode and ion collec-tor) where the cathode is a hot cathode.Cathodes used to be made of tungsten butare now usually made of oxide-coated iri-dium (Th2O3, Y2O3) to reduce the electronoutput work and make them more resi-stant to oxygen. Ionization vacuum gaugesof this type work with low voltages andwithout an external magnetic field. The hotcathode is a very high-yield source of elec-trons. The electrons are accelerated in theelectric field (see Fig. 3.13) and receivesufficient energy from the field to ionizethe gas in which the electrode system islocated. The positive gas ions formed aretransported to the ion collector, which isnegative with respect to the cathode, andgive up their charge there. The ion currentthereby generated is a measure of the gasdensity and thus of the gas pressure. If i-is the electron current emitted by the hotcathode, the pressure-proportional current

i+ produced in the measuring system isdefined by:

i+ = C · i– · p und (3.3)

(3.3a)

The variable C is the vacuum gauge con-stant of the measuring system. For nitro-gen this variable is generally around 10 mbar-1. With a constant electron cur-rent the sensitivity S of a gauge head isdefined as the quotient of the ion currentand the pressure. For an electron currentof 1 mA and C = 10 mbar-1, therefore, thesensitivity S of the gauge head is:

S = i+ / p = C · i- = 10 mbar-1 · 1 mA = 10 mbar-1 · 10-3 A = 1 · 10-2 A/mbar.

Hot-cathode ionization vacuum gaugesalso exhibit gas sorption (pumpingaction), which, however, is considerablysmaller than with Penning systems, i.e.approx. 10-3 l/s. Essentially this gas sorp-tion takes place on the glass wall of thegauge head and, to a lesser extent, at theion collector. Here use is made of nudegauges that are easy to operate because anexternal magnet is not needed. The upperlimit of the measuring range of the hot-cathode ionization gauge is around 10-2 mbar (with the exception of specialdesigns). It is basically defined by thescatter processes of ions at gas moleculesdue to the shorter free path at higher pres-sures (the ions no longer reach the ioncollector = lower ion yield). Moreover,uncontrollable glow or arc discharges mayform at higher pressures and electrostaticdischarges can occur in glass tubes. Inthese cases the indicated pressure pI maydeviate substantially from the true pressu-re pT.

At low pressures the measuring range islimited by two effects: by the X-ray effectand by the ion desorption effect. Theseeffects results in loss of the strict propor-tionality between the pressure and the ioncurrent and produce a low pressure thres-hold that apparently cannot be crossed(see Fig. 3.14).

The X-ray effect (see Fig. 3.15)The electrons emitted from the cathodeimpinge on the anode, releasing photons(soft X-rays). These photons, in turn,trigger photoelectrons from surfaces they

p ii C

=⋅

+

−

Fig. 3.12 Cross-section of PENNINGVAC PR 35 gauge

1 Small flange DN 25 KF;DN 40 KF

2 Housing3 Ring anode with

ignition pin

4 Ceramic washer5 Current leadthrough6 Connecting bush7 Anode pin8 Cathode plate

Fig. 3.13 Schematic diagram and potential curve in ahot-cathode ionization vacuum gauge

i+: ion currenti-: electron current

Anode

UC(+ 50V)

UA(+ 200V)

UA

UC

Cathode Ion collector

D00 E 19.06.2001 21:38 Uhr Seite 78



strike. The photoelectrons released fromthe ion collector flow to the anode, i.e. theion collector emits an electron current,which is indicated in the same manner asa positive ion current flowing to the ioncollector. This photocurrent simulates apressure. This effect is called the positiveX-ray effect, and it depends on the anodevoltage as well as on the size of the surfa-ce of the ion collector.

Under certain circumstances, however,there is also a negative X-ray effect. Pho-tons which impinge on the wall surroun-ding the gauge head release photoelec-trons there, which again flow towards theanode, and since the anode is a grid struc-ture, they also flow into the space withinthe anode. If the surrounding wall has thesame potential as the ion collector, e.g.ground potential, a portion of the electronsreleased at the wall can reach the ioncollector. This results in the flow of anelectron current to the ion collector, i.e. anegative current flows which can compen-sate the positive ion current. This negativeX-ray effect depends on the potential of theouter wall of the gauge head.

The ion desorption effectAdsorbed gases can be desorbed from asurface by electron impact. For an ionizati-on gauge this means that, if there is a layerof adsorbed gas on the anode, these gasesare partly desorbed as ions by the impin-ging electrons. The ions reach the ion

collector and lead to a pressure indicationthat is initially independent of the pressurebut rises as the electron current increases.If such a small electron current is used sothat the number of electrons incident at thesurface is small compared to the number ofadsorbed gas particles, every electron willbe able to desorb positive ions. If the elec-tron current is then increased, desorptionwill initially increase because more elec-trons impinge on the surface. This finallyleads to a reduction in adsorbed gas par-ticles at the surface. The reading falls againand generally reaches values that may beconsiderably lower than the pressure rea-ding observed with a small electron cur-rent. As a consequence of this effect inpractice, one must ascertain whether thepressure reading has been influenced by adesorption current. This can be done mostsimply by temporarily altering the electroncurrent by a factor of 10 or 100. The rea-ding for the larger electron current is themore precise pressure value.

In addition to the conventional ionizationgauge, whose electrode structure resemb-les that of a common triode, there arevarious ionization vacuum gauge systems(Bayard-Alpert system, Bayard-Alpertsystem with modulator, extractor system)which more or less suppress the twoeffects, depending on the design, and aretherefore used for measurement in thehigh and ultrahigh vacuum range. Today

the Bayard-Alpert system is usually thestandard system.

a) The conventional ionization vacuumgauge

A triode of conventional design (see Fig.3.16 a) is used as the gauge head, but it isslightly modified so that the outer electro-de serves as the ion collector and the gridwithin it as the anode. With this arrange-ment the electrons are forced to take verylong paths (oscillating around the gridwires of the anode) so that the probabilityof ionizing collisions and thus the sensiti-vity of the gauge are relatively high. Becau-se the triode system can generally only beused in high vacuum on account of itsstrong X-ray effect, the gas sorption (pum-ping) effect and the gas content of theelectrode system have only a slight effecton the pressure measurement.

D00Fig. 3.14 Apparent low pressure limit due to X-ray

effect in a normal ionization vacuum gauge

Actual pressure mbar

Indi

cate

d pr

essu

re m

bar

Fig. 3.15 Explanation of the X-ray effect in a conventio-nal ionization gauge. The electrons e- emittedby the cathode C collide with anode A andtrigger a soft X-ray radiation (photons) there.This radiation strikes, in part, the ion collec-tor and generates photoelectrons eÐ

s there

C CathodeA AnodeI Ion collector

I Pressure reading without X-ray effectII Apparent low pressure limit due to X-ray effectIII Sum of I and II

Fig. 3.16 Schematic drawing of the electrode arrangement of various ionization vacuumgauge measuring systems

I ion collectorSc screenM modulator

A anodeC cathodeR reflector

a)

b)

c)

d)

e)

Conventionalionization vacuum gaugesystem

ionization vacu-um gaugesystem for higherpressures (up to1 mbar)

Bayard-Alpert ionization vacuum gaugesystem

Bayard-Alpert ionization vacuum gaugesystem withmodulator

extractor ionization vacuum gaugesystem

D00 E 19.06.2001 21:38 Uhr Seite 79



b) The high-pressure ionization vacuumgauge (up to 1 mbar)

A triode is again used as the electrodesystem (see Fig. 3.16 b), but this time withan unmodified conventional design. Sincethe gauge is designed to allow pressuremeasurements up to 1 mbar, the cathodemust be resistant to relatively high oxygenpressure. Therefore, it is designed as a so-called non-burnout cathode, consisting ofan yttria-coated iridium ribbon. To obtain arectilinear characteristic (ion current as alinear function of the pressure) up to apressure of 1 mbar, a high-ohmic resistoris installed in the anode circuit.

c) Bayard-Alpert ionization vacuumgauge (the standard measuringsystem used today)

To ensure linearity between the gas pres-sure and the ion current over as large apressure range as possible, the X-rayeffect must be suppressed as far as possi-ble. In the electrode arrangement develo-ped by Bayard and Alpert, this is achievedby virtue of the fact that the hot cathode islocated outside the anode and the ioncollector is a thin wire forming the axis ofthe electrode system (see Fig. 3.16 c). TheX-ray effect is reduced by two to threeorders of magnitude due to the greatreduction in the surface area of the ioncollector. When pressures in the ultrahighvacuum range are measured, the innersurfaces of the gauge head and theconnections to the vessel affect the pres-sure reading. The various effects of ad-sorption, desorption, dissociation and flowphenomena cannot be dealt with in thiscontext. By using Bayard-Alpert systemsas nude gauge systems that are placeddirectly in the vessel, errors in measure-ment can be extensively avoided becauseof the above mentioned effects.

d) Bayard-Alpert ionization vacuum gaugewith modulator

The Bayard-Alpert system with modulator(see Fig. 3.16 d), introduced by Redhead,offers pressure measurement in whicherrors due to X-ray and ion desorptioneffects can be quantitatively taken intoaccount. In this arrangement there is asecond thin wire, the modulator, near theanode in addition to the ion collector insi-de the anode. If this modulator is set at theanode potential, it does not influence themeasurement. If, on the other hand, thesame potential is applied to the modulator

as that on the ion collector, part of the ioncurrent formed flows to the modulator andthe current that flows to the ion collectorbecomes smaller. The indicated pressurepI of the ionization gauge with modulatorset to the anode potential consists of theportion due to the gas pressure pg and thatdue to the X-ray effect pγ:

pA = pg + pγ (3.4)

After switching the modulator from theanode potential over to the ion collectorpotential, the modulated pressure readingpM is lower than the pI reading because aportion of the ions now reaches the modu-lator. Hence:

pM = α · pg + pγ (3.5)

with α < 1.

The pg share of the X-ray effect is the samein both cases. After determining the diffe-rence between (3.4) and (3.5), we obtainthe equation for the gas pressure pg:

(3.6)

α can immediately be determined by expe-riment at a higher pressure (around 10-6 mbar) at which the X-ray effect andthus pγ are negligible. The pressure corre-sponding to the two modulator potentialsare read off and their ratio is formed. Thismodulation method has the additionaladvantage that the ion desorption effect isdetermined in this way. It permits pressu-re measurements up to the 10-11 mbarrange with relatively little effort.

e) Extractor ionization vacuum gaugeDisruptive effects that influence pressuremeasurement can also be extensively eli-minated by means of an ion-optical systemfirst suggested by Redhead. With thisextractor system (see Fig. 3.16 e) the ionsfrom the anode cylinder are focused on avery thin and short ion collector. The ioncollector is set up in a space, the rear wallof which is formed by a cup-shaped elec-trode that is maintained at the anodepotential so that it cannot be reached byions emanating from the gas space. Due tothe geometry of the system as well as thepotential of the of individual electrodes,the disruptive influences through X-rayeffects and ion desorption are almost com-pletely excluded without the need of amodulator. The extractor system measurespressures between 10-4 and 10-12 mbar.

pg

pA

pM=

−−1 α

Another advantage is that the measuringsystem is designed as a nude gauge with adiameter of only 35 mm so that it can beinstalled in small apparatus.

D00 E 19.06.2001 21:38 Uhr Seite 80



3.4 Adjustment andcalibration; DKD,PTB national standards

Definition of terms: Since these terms areoften confused in daily usage, a clear defi-nition of them will first be provided:

Adjustment or tuning refers to the correctsetting of an instrument. For example, set-ting 0 and 100 % in THERMOVACs or set-ting the mass spectrometer to mass 4 inthe helium leak detector.

Calibration inspection refers to com-parison with a standard in accordance withcertain statutory regulations by speciallyauthorized personnel (Bureau of Stand-ards). If the outcome of this regular ins-pection is positive, an operating permit forthe next operation period (e.g. three years)is made visible for outsiders by means of asticker or lead seal. If the outcome is nega-tive, the instrument is withdrawn fromoperation.

Calibration refers to comparison with astandard in accordance with certain statutory regulations by specially authori-zed personnel (calibration facility). Theresult of this procedure is a calibration cer-tificate which contains the deviations ofthe readings of the instrument being cali-brated from the standard.

Calibration facilities carry out this calibra-tion work. One problem that arises is thequestion of how good the standards areand where they are calibrated. Such stan-dards are calibrated in calibration facilitiesof the German Calibration Service (DKD).The German Calibration Service is mana-ged by the Federal Physical-TechnicalInstitute (PTB). Its function is to ensurethat measuring and testing equipmentused for industrial measurement purposesis subjected to official standards. Calibrati-on of vacuum gauges and test leaks withinthe framework of the DKD has been assig-ned to LEYBOLD, as well as other compa-nies, by the PTB. The required calibrationpump bench was set up in accordancewith DIN 28 418 (see Table 11.1) and theninspected and accepted by the PTB. Thestandards of the DKD facilities, so-calledtransfer standards (reference vacuum

gauges), are calibrated directly by the PTBat regular intervals. Vacuum gauges of allmakes are calibrated on an impartial basisby LEYBOLD in Cologne according tocustomer order. A DKD calibration certifi-cate is issued with all characteristic dataon the calibration. The standards of theFederal Physical-Technical Institute are theso-called national standards. To be able toguarantee adequate measuring accuracyor as little measurement uncertainty aspossible in its calibrations, the PTB largelycarries out its measurements through theapplication of fundamental methods. Thismeans, for example, that one attempts todescribe the calibration pressures throughthe measurement of force and area or bythinning the gases in strict accordancewith physical laws. The chain of the recali-bration of standard instruments carriedout once a year at the next higher qualifiedcalibration facility up to the PTB is called“resetting to national standards”. In othercountries as well, similar methods are car-ried out by the national standards institu-tes as those applied by the Federal Physi-cal-Technical Institute (PTB) in Germany.Fig. 3.17 shows the pressure scale of thePTB. Calibration guidelines are specified inDIN standards (DIN 28 416) and ISO pro-posals.

3.4.1 Examples of fundamentalpressure measurementmethods (as standardmethods for calibratingvacuum gauges)

a) Measuring pressure with a referencegauge

An example of such an instrument is theU-tube vacuum gauge, with which themeasurement of the pressure in the mea-surement capillary is based on a measure-ment of the weight over the length of themercury column.

In the past the McLeod vacuum gauge wasalso used for calibration purposes. With aprecision-made McLeod and carefully exe-cuted measurements, taking into accountall possible sources of error, pressuresdown to 10-4 mbar can be measured withconsiderable accuracy by means of suchan instrument.Another reference gauge is the VISCOVACdecrement gauge with rotating ball (see3.3.1) as well as the capacitance dia-phragm gauge (see 3.2.2.4).

b) Generation of a known pressure; staticexpansion method

On the basis of a certain quantity of gaswhose parameters p, V and T are knownexactly – p lies within the measuring rangeof a reference gauge such as a U-tube or

D0010–12 10–9 10–6 10–3 100 103

0.1

0.3

1

3

10

30

Rela

tive

unce

rtain

ly o

f the

pre

ssur

e de

term

inat

ion

[%]

Pressure [mbar]

Dynamicexpansion

Molecularbeam Static expansion

U-Tube

Fig. 3.17 Pressure scale of Federal Physical-Technical Institute (PTB), Berlin, (status as at August 1984) for inert gases,nitrogen and methane

D00 E 19.06.2001 21:38 Uhr Seite 81



McLeod vacuum gauge – a lower pressurewithin the working range of ionization gau-ges is reached via expansion in severalstages.

If the gas having volume V1 is expanded toa volume (V1 + V2), and from V2 to (V2 + V3), etc., one obtains, after n stagesof expansion:

(3.7)

p1 = initial pressure measured directly inmbar

pn = calibration pressure

The volumes here must be known as pre-cisely as possible (see Fig. 3.18) and thetemperature has to remain constant. Thismethod requires that the apparatus usedbe kept very clean and reaches its limit atpressures where the gas quantity can bealtered by desorption or adsorption effectsbeyond the permissible limits of error.According to experience, this lower limit isaround 5 · 10-7 mbar. This method is cal-led the static expansion method becausethe pressure and volume of the gas at restare the decisive variables.

pn

pV

V V

V

V V

Vn

Vn Vn= ⋅

+⋅

+⋅⋅ ⋅⋅ −

− +11

1 2

2

2 3

1

1

c) Dynamic expansion method (see Fig. 3.19)

According to this method, the calibrationpressure p is produced by admitting gas ata constant throughput rate Q into a va-cuum chamber while gas is simultaneous-ly pumped out of the chamber by a pumpunit with a constant pumping speed S. Atequilibrium the following applies accor-ding to equation 1.10 a:

p = Q/S

Q is obtained either from the quantity ofgas that flows into the calibration chamberfrom a supply vessel in which constantpressure prevails or from the quantity ofgas flowing into the calibration chamber ata measured pressure through a knownconductance. The pressure in front of theinlet valve must be high enough so that itcan be measured with a reference gauge.The inlet apertures of the valve (smallcapillaries, sintered bodies) must be sosmall that the condition d << λ is met, i.e.a molecular flow and hence a constantconductance of the inlet valve are obtained(see Section 1.5). The quantity of gas isthen defined by p1 · L1, where p1 = pressure in front of the inlet valve and

L1 = conductance of the valve. The pum-ping system consists of a precisely mea-sured aperture with the conductance L2 ina wall that is as thin as possible (screenconductance) and a pump with a pumpingspeed of PSp:

(3.8)

and thus

(3.9)

This method has the advantage that, afterreaching a state of equilibrium, sorptioneffects can be ignored and this procedurecan therefore be used for calibrating gau-ges at very low pressures.

p2

p p )(1 +L

S

L L

L S= =⋅ ⋅ ⋅

1 11 1 2

2 p

SL Sp

L Sp

L

LSp

=⋅+

=+

2

2

2

1 2

+

++ + +

+

+

IM IM

13000 cm3V =4

p4

p3

p2

p1

V = 25 cm33V = 25 cm1

3

V = 1000 cm23

Fig. 3.18 Generation of low pressures through static expansion Fig. 3.19 Apparatus for calibration according to the dynamic expansion method

p2 = p1 (Sp >> L2)L1S

1 Volume 12 Volume 23 Inlet valve

(conductance L1)4 Aperture with

conductance L25 Valve6 to pump system7 Valve8 to gas reservoir9 Valve

10 LN2 cold trap11 to pump system12 U-tube vacuum

gauge13 McLeod vacuum

gauge14 Valve15 Calibrated

ionization gaugetube

16 to pump(pumping speedPSp)

17 Gas inlet18 Mass spectrometer19,20 Gauges to be

calibrated21 Nude gauge to be

calibrated22 Bake-out furnace

D00 E 19.06.2001 21:38 Uhr Seite 82



3.5 Pressure monitoring, controland regulation in vacuum systems

3.5.1 Fundamentals of pressuremonitoring and control

In all vacuum processes the pressure inthe system must be constantly checkedand, if necessary, regulated. Modern plantcontrol additionally requires that all mea-sured values which are important formonitoring a plant are transmitted to cen-tral stations, monitoring and control cen-ters and compiled in a clear manner. Pres-sure changes are frequently recorded overtime by recording equipment. This meansthat additional demands are placed onvacuum gauges:a) continuous indication of measured

values, analog and digital as far as pos-sible

b) clear and convenient reading of themeasured values

c) recorder output to connect a recordinginstrument or control or regulationequipment

d) built-in computer interface (e.g. RS 232)e) facility for triggering switching operati-

ons through built-in trigger points

These demands are generally met by allvacuum gauges that have an electric mea-sured value display, with the exception ofBourdon, precision diaphragm and liquid-filled vacuum gauges. The respective con-trol units are equipped with recorder out-puts that supply continuous voltages bet-ween 0 and 10 V, depending on the pres-sure reading on the meter scale, so thatthe pressure values can be recorded overtime by means of a recording instrument.If a pressure switching unit is connected tothe recorder output of the gauge, swit-ching operations can be triggered whenthe values go over or below specified set-points. The setpoints or switch thresholdvalues for triggering switching operationsdirectly in the gauges are called triggervalues. Apart from vacuum gauges, thereare diaphragm pressure switches that trig-ger a switching operation (without displayof a measured value) via a contact ampli-fier when a certain pressure is reached.Valves, for example, can also be controlledthrough such switching operations.

3.5.2 Automatic protection,monitoring and control ofvacuum systems

Protection of a vacuum system againstmalfunctions is extremely important. Inthe event of failure, very high materialvalues may be at risk, whether throughloss of the entire system or major compo-nents of it, due to loss of the batch ofmaterial to be processed or due to furtherproduction down time. Adequate operatio-nal control and protection should therefo-re be provided for, particularly in the caseof large production plants. The individualfactors to be taken into account in thisconnection are best illustrated on the basisof an example: Fig. 3.20 shows the sche-matic diagram of a high vacuum pumpsystem. The vessel (11) can be evacuatedby means of a Roots pump (14) or a diffu-sion pump (15), both of which operate inconjunction with the backing pump (1).The Roots pump is used in the mediumvacuum range and the diffusion pump inthe high vacuum range. The valves (3), (8)and (16) are operated electropneuma-tically. The individual components are

actuated from a control panel with push-buttons. The pump system is to be protec-ted against the following malfunctions:a) power failureb) drop in pressure in the compressed air

networkc) failure of cooling water to the diffusion

pumpd) fault in diffusion pump heating systeme) failure of backing pumpf) pressure rise in the vessel above a

maximum permissible valueg) pressure rise above a maximum

backing pressure (critical forepressureof the diffusion pump)

The measures to be taken in order to fore-stall such malfunctions will be discussedin the same order:a) Measures in the event of power failure:

All valves are closed so as to preventadmission of air to the vacuum vesseland protect the diffusion pump againstdamage.

b) Protection in the event of a drop inpressure in the compressed air net-work: The compressed air is monitoredby a pressure monitoring device (5). If

D00Fig. 3.20 Schematic diagram of a high vacuum pump system with optional operation of a Roots pump or

a diffusion pump

1 Backing pump2 Backing pressure

monitoring device3 Electropneumatic valve4 Compressed air connection5 Pressure monitoring device

6 Temperature monitoring device7 Cooling water monitoring device8 Electropneumatic valve9 Recorder10 High-vacuum monitoring device11 Vessel

12 High-vacuum gauge13 Limit switches14 Roots pump15 Diffusion pump16 Electropneumatic valve17 Venting valve

D00 E 19.06.2001 21:39 Uhr Seite 83



the pressure falls under a specifiedvalue, a signal can initially be emitted orthe valves can be automatically closed.In this case, a sufficient reserve supplyof compressed air is necessary (notshown in Fig. 3.20), which allows allvalves to be actuated at least once.

c) Measures in the event of failure of coo-ling water to the diffusion pump: Thecooling water is monitored by a flow ortemperature monitoring device (6) and(7). If the flow of cooling water is ina-dequate, the heater of the diffusionpump is switched off and a signal isgiven; the valve (8) closes.

d) Protection against failure of the diffusi-on pump heater: Interruption of the dif-fusion pump heating system can bemonitored by a relay. If the temperaturerises above a maximum permissiblevalue, a temperature monitoring device(6) responds. In both cases the valve(8) closes and a signal is given.

e) Protection in the event of backing pumpfailure: Belt-driven backing pumpsmust have a centrifugal switch whichshuts down the entire system in theevent of belt breakage or anothermalfunction. Monoblock pumps forwhich the drive is mounted directly onthe shaft can be monitored by currentrelays and the like.

f) Protection against a pressure rise in thevessel above a certain limit value: Thehigh vacuum monitoring device (10)emits a signal when a specified pressu-re is exceeded.

g) Ensuring the critical forepressure of thediffusion pump: When a certain backingpressure is exceeded, all valves are clo-sed by the backing pressure monitoringdevice (2), the pumps are switched offand again a signal is given. The positionof the valves (3), (8) and (16) is indica-ted on the control panel by means oflimit switches (13). The pressure in thevessel is measured with a high vacuumgauge (12) and recorded with a recor-der (9). Protection against operatingerrors can be provided by interlockingthe individual switches so that they canonly be actuated in a predeterminedsequence. The diffusion pump, forexample, may not be switched on whenthe backing pump is not running or therequired backing pressure is not main-tained or the cooling water circulation isnot functioning.

In principle, it is not a big step from a

system protected against all malfunctionsto a fully automatic, program-controlledplant, though the complexity of the electri-cal circuits, of course, increases signifi-cantly.

3.5.3 Pressure regulation andcontrol in rough andmedium vacuum systems

Control and regulation have the function ofgiving a physical variable – in this case thepressure in the vacuum system – a certainvalue. The common feature is the actuatorwhich changes the energy supply to thephysical variable and thus the variable its-elf. Control refers to influencing a systemor unit through commands. In this casethe actuator and hence the actual value ofthe physical variable is changed directlywith a manipulated variable. Example:Actuation of a valve by means of a pressu-re-dependent switch. The actual value maychange in an undesirable way due to addi-tional external influences. The controlledunit cannot react to the control unit. Forthis reason control systems are said tohave an open operating sequence. In thecase of regulation, the actual value of thephysical variable is constantly comparedto the specified setpoint and regulated ifthere is any deviation so that it completelyapproximates the setpoint as far as possi-ble. For all practical purposes regulationalways requires control. The main differen-

ce is the controller in which the setpointand the actual value are compared. Thetotality of all elements involved in the con-trol process forms the control circuit. Theterms and characteristic variables fordescribing control processes are stipula-ted in DIN 19226.

Generally a distinction is made betweendiscontinuous control (e.g. two-step orthree-step control) with specification of apressure window, within which the pressu-re may vary, and continuous control (e.g.PID control) with a specified pressure set-point, which should be maintained as pre-cisely as possible. We have two possibleways of adjusting the pressure in a vacu-um system: first, by changing the pumpingspeed (altering the speed of the pump orthrottling by closing a valve); second,through admission of gas (opening avalve). This results in a total of 4 procedu-res.

Discontinuous pressure regulationAlthough continuous regulation undoub-tedly represents the more elegant proce-dure, in many cases two-step or three-stepregulation is fully adequate in all vacuumranges. To specify the pressure window,two or three variable, pressure-dependentswitch contacts are necessary. It does notmatter here whether the switch contactsare installed in a gauge with display or in adownstream unit or whether it is a pressu-re switch without display. Fig. 3.21 illust-rates the difference between two-stepregulation through pumping speed thrott-ling, two-point regulation through gas

Time TimeTime

Pressure PressurePressure

pAtm pAtmpAtm

pmax

pmax

pmin pmin

pmittepmin

pumping speed throttlingand gas admission

pumping speed throttling

Two-step regulation through Three-step regulation through

gas admission

Fig. 3.21 Schematic diagram of two-step and three-step regulation

D00 E 19.06.2001 21:39 Uhr Seite 84



admission and three-point regulationthrough a combination of pumping speedthrottling and gas admission. Figures 3.22and 3.23 show the circuit and structure ofthe two two-step regulation systems. Inthe case of two-step regulation throughpumping speed throttling (Fig. 3.22), vol-tage is supplied to pump valve 4, i.e. it isopen when the relay contacts are in therelease condition. At a level below theupper switching point the valve remainsopen because of the self-holding functionof the auxiliary relay. Only at a level belowthe lower switching point is the relay lat-ching released. If the pressure subse-quently rises, the valve is opened again atthe upper switching point.

In the case of two-step regulation throughgas admission, the inlet valve is initiallyclosed. If the upper pressure switchingpoint is not reached, nothing changes;only when the pressure falls below thelower switching point, do the “makecontacts” open the gas inlet valve andactuate the auxiliary relay with self-holdingfunction simultaneously. Return to the idlestate with closing of the gas inlet valve isnot effected until after the upper switchingpoint is exceeded due to the release of therelay self-holding function.

Fig. 3.24 shows the corresponding three-step regulation system which was createdwith the two components just described.

As the name indicates, two switchingpoints, the lower switching point of theregulation system through pumping speedthrottling and the upper switching point ofthe gas inlet regulation system, were com-bined.

To avoid the complicated installation withauxiliary relays, many units offer a facilityfor changing the type of function of thebuilt-in trigger values via software. Initiallyone can choose between individual swit-ching points (so-called “level triggers”)and interlinked switching points (“intervaltriggers”). These functions are explainedin Fig. 3.25. With interval triggers one canalso select the size of the hysteresis and

D00

Fig. 3.22 Two-step regulation through pumping speed throttling

➀ Gauge with two switching points

➁ Throttle valve➂ Vacuum pump➃ Pump valve➄ Vacuum vessel

Fu FuseR, Mp Mains connection 220 V/50 HzSmax Switching point for maximum valueSmin Switching point for minimum valuePV Pump valveR1 Auxiliary relay for pump valveK1 Relay contact of R1M Measuring and switching device

Fig. 3.23 Two-step regulation through gas admission

➀ Gauge with two switching points

➁ Variable-leak valve➂ Inlet valve➃ Gas supply➄ Throttle valve➅ Vacuum pump➆ Vacuum vessel

Fu FuseR, Mp Mains connection 220 V/50 HzSmax Switching point for maximum valueSmin Switching point for minimum valueEV Inlet valveR2 Auxiliary relay for inlet valveK2 Relay contact of R2M Measuring and switching device

Fig. 3.24 Three-step regulation system

➀ Gauge with three switching points

➁ Variable-leak valve➂ Variable-leak valve➃ Inlet valve➄ Gas supply➅ Throttle valve➆ Vacuum pump➇ Pump valve➈ Vacuum vessel

Fu FuseR, Mp Mains connection 220 V/50 HzSmax Switching point for maximum valueSmitte Switching point for mean valueSmin Switching point for minimum valueT TORROSTAT® S 020PV Pump valveEV Inlet valveR1 Auxiliary relay for pump intervalR2 Auxiliary relay for inlet intervalK1 Relay contact of R1K2 Relay contact of R2M Measuring and switching device

Fig. 3.25 Diagram of level triggers and interval triggers

D00 E 19.06.2001 21:39 Uhr Seite 85



the type of setpoint specification, i.e. eitherfixed setting in the unit or specificationthrough an external voltage, e.g. from 0 –10 volts. A three-step regulation system(without auxiliary relay), for example, canbe set up with the LEYBOLD MEMBRANOVACof the A series. Fig. 3.26 shows differentunits of the new LEYBOLD A series, which,although they function according to diffe-rent measuring methods, all display a uni-form appearance.

Continuous pressure regulationWe have to make a distinction here bet-ween electric controllers (e.g. PID con-trollers) with a proportional valve as actua-tor and mechanical diaphragm controllers.In a regulation system with electric con-trollers the coordination between control-ler and actuator (piezoelectric gas inletvalve, inlet valve with motor drive, butterf-ly control valve, throttle valve) is difficultbecause of the very different boundaryconditions (volume of the vessel, effectivepumping speed at the vessel, pressurecontrol range). Such control circuits tendto vibrate easily when process mal-functions occur. It is virtually impossible tospecify generally valid standard values.

Many control problems can be better sol-ved with a diaphragm controller. The func-tion of the diaphragm controller (see Fig.3.27) can be easily derived from that of adiaphragm vacuum gauge: the blunt end ofa tube or pipe is either closed off by meansof an elastic rubber diaphragm (for refe-

rence pressure > process pressure) orreleased (for reference pressure < processpressure) so that in the latter case, aconnection is established between the pro-cess side and the vacuum pump. This ele-gant and more or less “automatic” regula-tion system has excellent control characte-ristics (see Fig. 3.28).

To achieve higher flow rates, several dia-phragm controllers can be connected inparallel. This means that the processchambers and the reference chambers arealso connected in parallel. Fig. 3.29 shows

such a connection of 3 MR 50 diaphragmcontrollers.

To control a vacuum process, it is fre-quently necessary to modify the pressurein individual process steps. With a dia-phragm controller this can be done eithermanually or via electric control of the refe-rence pressure.

Electric control of the reference pressureof a diaphragm controller is relatively easybecause of the small reference volume thatalways remains constant. Fig. 3.31 showssuch an arrangement on the left as a pic-

Fig. 3.26 LEYBOLD-A series, equipment with level and interval triggers Fig. 3.27 Principle of a diaphragm controller

reference pressureadjustment valve

pump connection

control chamber

controller seat

reference chamber

diaphragm

measuringconnection

for referencechamber

measuringconnectionfor process pressure

process chamberconnection

Fig. 3.28 Control characteristics of a diaphragm controller

P1 = process pressure, P2 = pressure in pump, Pref = reference pressure

D00 E 19.06.2001 21:39 Uhr Seite 86



ture and on the right schematically, see3.5.5 for application examples with diaphragm controllers.

To be able to change the reference pressu-re and thus the process pressure towardshigher pressures, a gas inlet valve mustadditionally be installed at the processchamber. This valve is opened by means ofa differential pressure switch (not shownin Fig. 3.31) when the desired higher pro-cess pressure exceeds the current processpressure by more than the pressure diffe-rence set on the differential pressureswitch.

3.5.4 Pressure regulation inhigh and ultrahigh vacu-um systems

If the pressure is to be kept constant wit-hin certain limits, an equilibrium must beestablished between the gas admitted tothe vacuum vessel and the gas simulta-neously removed by the pump with the aidof valves or throttling devices. This is notvery difficult in rough and medium vacu-um systems because desorption of adsor-bed gases from the walls is generally neg-ligible in comparison to the quantity of gasflowing through the system. Pressureregulation can be carried out through gasinlet or pumping speed regulation. Howe-ver, the use of diaphragm controllers isonly possible between atmospheric pres-sure and about 10 mbar.

In the high and ultrahigh vacuum range,on the other hand, the gas evolution fromthe vessel walls has a decisive influence onthe pressure. Setting of specific pressurevalues in the high and ultrahigh vacuumrange, therefore, is only possible if the gasevolution from the walls is negligible inrelation to the controlled admission of gasby means of the pressure-regulating unit.For this reason, pressure regulation in thisrange is usually effected as gas admissionregulation with an electric PID controller.Piezoelectric or servomotor-controlled

variable-leak valves are used as actuators.Only bakeable all-metal gas inlet valvesshould be used for pressure regulationbelow 10-6 mbar.

D00

Fig. 3.29 Triple connection of diaphragm controllers

Fig. 3.30 Control of vacuum drying processes by regu-lation of the intake pressure of the vacuumpump according to the water vapor tolerance

DC Diaphragm controllerP Vacuum pumpM Measuring and switching devicePS Pressure sensorV1 Pump valveV2 Gas inlet valveTH ThrottleRC Reference chamberPC Process chamberCV Internal reference pressure control valve

Fig. 3.31 Diaphragm controller with external automatic reference pressure regulation

DC Diaphragm controllerPS Process pressure sensorRS Reference pressure sensorV1 Gas inlet valveV2 Pump valveV3 Gas inlet variable-leak valveTH ThrottleM Measuring and switching devicePP Process pumpRC Reference chamberPC Process chamberAP Auxiliary pumpCV Internal reference pressure control valve

D00 E 19.06.2001 21:39 Uhr Seite 87



3.5.5 Examples of applicationswith diaphragm controllers

1) Regulation of a drying/distillation pro-cess, taking into account the maximumwater vapor tolerance of a vane typerotary pump

In a drying process it is frequently desira-ble to carry out drying solely by means ofvacuum pumps without inserting conden-sers. In view of the limited water vaportolerance of vacuum pumps – approx. 30 mbar as a rule – this would result incondensation of the vapors produced wit-hin the vacuum pump, given non-throttledor non-regulated pumping speed. One canavoid this through process-dependentremote control of a diaphragm controllerwith auxiliary control valves and a measu-ring and switching device with a pressuresensor at the inlet connection of the vacu-um pump if the intake pressure is adaptedto the pumps water vapor tolerancethrough automatic monitoring of the inta-ke pressure of the vacuum pump and bythrottling the pumping speed. Fig. 3.30shows the principle of this arrangement.

Mode of operation: Starting from atmos-pheric pressure with the process heatingswitched off, valve V1 is initially open(maximum switching point exceeded) sothat atmospheric pressure also prevails inthe reference chamber.

The diaphragm controller is therefore clo-sed. When the system is started up, theconnecting line between the vacuum pumpand pump valve V2 is first evacuated. Assoon as the pressure drops below themaximum switching point, valve V1 clo-ses. When the pressure falls below theminimum switching point, valve V2 opens.

In this manner the pressure in the referen-ce chamber is slowly lowered, the thrott-ling of the diaphragm controller is reducedaccordingly and thus the process pressureis lowered until the quantity of process gasis greater than the quantity conveyed bythe pump so that the minimum switchingpoint is again exceeded. Valve V2 closesagain. This interaction repeats itself untilthe pressure in the process chamber hasdropped below the minimum switchingpoint. After that, valve V2 remains open sothat the process can be brought down tothe required final pressure with a comple-tely open diaphragm controller.

The material to be dried is usually heatedto intensify and speed up the drying pro-cess. If a certain amount of water vapor isproduced, the intake pressure rises abovethe two switching points. As a result, valveV2 first closes and V1 opens. Throughincoming air or protective gas the pressu-re in the reference chamber is raised andthe throughput at the diaphragm controllerthus throttled until the intake pressure ofthe vacuum pump has dropped below theset maximum switching point again. Thenvalve V1 closes.

Depending on the quantity of vapor thataccumulates, the throughput of the diaphragm controller is set by increasingor decreasing the reference pressure ineach case so that the maximum permissi-ble partial water vapor pressure at the inletconnection of the vacuum pump is neverexceeded.

As soon as the pressure in the processchamber drops below the set minimumswitching point towards the end of the dry-ing process, valve V2 opens and remainsopen. In this way the unthrottled cross-section of the diaphragm controller is avai-lable again for rapid final drying. At thesame time the final drying procedure canbe monitored by means of the pressuresensor PS.

2) Pressure regulation by means of dia-phragm controller with external automatic reference pressure adjust-ment (see Fig. 3.31)

For automatic vacuum processes withregulated process pressure, presetting ofthe desired set pressure must often func-tion and be monitored automatically. If adiaphragm controller is used, this can bedone by equipping the reference chamberwith a measuring and switching device anda control valve block at the referencechamber. The principle of this arrange-ment is shown in Fig. 3.31.

Mode of operation: Starting with atmos-pheric pressure, gas inlet valve V1 is clo-sed at the beginning of the process. Pumpvalve V2 opens. The process chamber isnow evacuated until the set pressure,which is preset at the measuring and swit-ching device, is reached in the processchamber and in the reference chamber.When the pressure falls below the set swit-ching threshold, pump valve V2 closes. Asa result, the pressure value attained is

“caught” as the reference pressure in thereference chamber (RC) of the diaphragmcontroller (DC). Now the process pressureis automatically maintained at a constantlevel according to the set reference pres-sure by means of the diaphragm controller(DC). If the reference pressure should risein the course of the process due to a leak,this is automatically detected by the mea-suring and switching device and correctedby briefly opening pump valve V2. Thisadditional control function enhances theoperational reliability and extends therange of application. Correcting the increa-sed reference pressure to the originally setvalue is of special interest for regulatedhelium circuits because the pressure risein the reference chamber (RC) of the dia-phragm controller can be compensated forthrough this arrangement as a consequen-ce of the unavoidable helium permeabilityof the controller diaphragm of FPM.

To be able to change the reference pressu-re and thus increase the process pressureto higher pressures, a gas inlet valve mustbe additionally installed at the processchamber. This valve is opened by means ofa differential pressure switch (not shownin Fig. 3.31) when the desired higher pro-cess pressure exceeds the current processpressure by more than the pressure diffe-rential set at the differential pressureswitch.

D00 E 19.06.2001 21:39 Uhr Seite 88

D00

Mass Spectrometry Fundamentals of Vacuum Technology


4. Analysis of gasat low pressuresusing massspectrometry

4.1 GeneralAnalyses of gases at low pressures areuseful not only when analyzing the residu-al gases from a vacuum pump, leak testingat a flange connection or for supply lines(compressed air, water) in a vacuum. Theyare also essential in the broader fields ofvacuum technology applications and pro-cesses. For example in the analysis of pro-cess gases used in applying thin layers ofcoatings to substrates. The equipmentused for qualitative and/or quantitativeanalyses of gases includes specially deve-loped mass spectrometers with extremelysmall dimensions which, like any othervacuum gauge, can be connected directlyto the vacuum system. Their size distin-guishes these measurement instrumentsfrom other mass spectrometers such asthose used for the chemical analyses ofgases. The latter devices are poorly suited,for example, for use as partial pressuremeasurement units since they are toolarge, require a long connector line to thevacuum chamber and cannot be baked outwith the vacuum chamber itself. The in-vestment for an analytical mass spectro-meter would be unjustifiably great since,for example, the requirements as to reso-lution are far less stringent for partial pres-sure measurements. Partial pressure isunderstood to be that pressure exerted bya certain type of gas within a mix of gases.The total of the partial pressures for all thetypes of gas gives the total pressure. Thedistinction among the various types ofgases is essentially on the basis of theirmolar masses. The primary purpose ofanalysis is therefore to register qualita-tively the proportions of gas within asystem as regards the molar masses anddetermine quantitatively the amount of theindividual types of gases associated withthe various atomic numbers.

Partial pressure measurement devices

which are in common use comprise themeasurement system proper (the sensor)and the control device required for its ope-ration. The sensor contains the ion source,the separation system and the ion trap.The separation of ions differing in massesand charges is often effected by utilizingphenomena which cause the ions to reso-nate in electrical and magnetic fields.

Initially, the control units were quite clum-sy and offered uncountable manipulationoptions. It was often the case that onlyphysicists were able to handle and usethem. With the introduction of PCs therequirements in regard to the control unitsbecame ever greater. At first, they were fit-ted with interfaces for linkage to the com-puter. Attempts were made later to equip aPC with an additional measurement circuitboard for sensor operation. Today’s sen-sors are in fact transmitters equipped withan electrical power supply unit attacheddirect at the atmosphere side; communica-tion with a PC from that point is via thestandard computer ports (RS 232, RS485). Operating convenience is achievedby the software which runs on the PC.

4.2 A historical reviewFollowing Thomson’s first attempt in 1897to determine the ratio of charge to masse/m for the electron, it was quite sometime (into the 1950s) before a large num-ber and variety of analysis systems cameinto use in vacuum technology. Theseincluded the Omegatron, the Topatron andultimately the quadrupole mass spectro-meter proposed by Paul and Steinwedel in1958, available from INFICON in its stan-dard version as the TRANSPECTOR (seeFig. 4.1). The first uses of mass spectro-

metry in vacuum-assisted process techno-logy applications presumably date back toBackus’ work in the years 1943 / 44. Hecarried out studies at the RadiographicLaboratories at the University of California.Seeking to separate uranium isotopes, heused a 180° sector field spectrometer afterDempster (1918), which he referred to asa “vacuum analyzer”. Even today a similarterm, namely the “residual gas analyzer”(RGA), is frequently used in the U.S.A. andthe U.K. instead of “mass spectrometer”.Today’s applications in process monitoringare found above all in the production ofsemiconductor components.

4.3 The quadrupolemass spectrometer(TRANSPECTOR)

The ion beam extracted from the electronimpact ion source is diverted into a qua-drupole separation system containing fourrod-shaped electrodes. The cross sectionsof the four rods form the circle of curvatu-re for a hyperbola so that the surroundingelectrical field is nearly hyperbolic. Each ofthe two opposing rods exhibits equalpotential, this being a DC voltage and asuperimposed high-frequency AC voltage

a: High-performance sensor with Channeltronb: Compact sensor with Micro-Channelplatec: High-performance sensor with Faraday cup

ab

c

Fig. 4.1a TRANSPECTOR sensors

Fig. 4.1b TRANSPECTOR XPR sensor

D00 E 19.06.2001 21:39 Uhr Seite 89

Mass SpectrometryFundamentals of Vacuum Technology


(Fig. 4.2). The voltages applied inducetransverse oscillations in the ions traver-sing the center, between the rods. Theamplitudes of almost all oscillationsescalate so that ultimately the ions willmake contact with the rods; only in thecase of ions with a certain ratio of mass tocharge m/e is the resonance conditionwhich allows passage through the systemsatisfied. Once they have escaped from theseparation system the ions move to the iontrap (detector, a Faraday cup) which mayalso take the form of a secondary electronmultiplier pick-up (SEMP).

The length of the sensor and the sepa-ration system is about 15 cm. To ensurethat the ions can travel unhindered fromthe ion source to the ion trap, the meanfree path length inside the sensor must beconsiderably greater than 15 cm. For airand nitrogen, the value is about p · λ = 6 · 10–3 mbar · cm. At p = 1 · 10-4

bar this corresponds to a mean free pathlength of λ = 60 cm. This pressure is gene-rally taken to be the minimum vacuum formass spectrometers. The emergency shut-down feature for the cathode (respondingto excessive pressure) is almost alwaysset for about 5 · 10-4 mbar. The desire tobe able to use quadrupole spectrometersat higher pressures too, without specialpressure convertors, led to the develop-ment of the XPR sensor at Inficon (XPRstanding for extended pressure range). Toenable direct measurement in the range ofabout 2 · 10-2 mbar, so important for sput-ter processes, the rod system was reducedfrom 12 cm to a length of 2 cm. To ensurethat the ions can execute the number oftransverse oscillations required for sharpmass separation, this number being about100, the frequency of the current in the

XPR sensor had to be raised from about 2MHz to approximately 6 times that value,namely to 13 MHz. In spite of the reductionin the length of the rod system, ion yield isstill reduced due to dispersion processesat such high pressures. Additional electro-nic correction is required to achieve per-fect depiction of the spectrum. The dimen-sions of the XPR sensor are so small thatit can “hide” entirely inside the tubulationof the connection flange (DN 40, CF) andthus occupies no space in the vacuumchamber proper. Fig. 4.1a shows the sizecomparison for the normal high-perfor-mance sensors with and without the Chan-neltron SEMP, the normal sensor withchannel-plate SEMP. Fig. 4.1b shows theXPR sensor. The high vacuum required forthe sensor is often generated with a TURBOVAC 50 turbomolecular pump anda D 1.6 B rotary vane pump. With its greatcompression capacity, a further advantageof the turbomolecular pump whenhandling high molar mass gases is that thesensor and its cathode are ideally pro-tected from contamination from thedirection of the forepump.

4.3.1 Design of the sensor

One can think of the sensor as having beenderived from an extractor measurementsystem (see Fig. 4.3), whereby the separation system was inserted betweenthe ion source and the ion trap.

4.3.1.1 The normal (open) ion source

The ion source comprises an arrangement

of the cathode, anode and several baffles.The electron emission, kept constant, cau-ses partial ionization of the residual gas,into which the ion source is “immersed” ascompletely as possible. The vacuum in thevicinity of the sensor will naturally be influ-enced by baking the walls or the cathode.The ions will be extracted through the baf-fles along the direction of the separationsystem. One of the baffles is connected toa separate amplifier and – entirely inde-pendent of ion separation – provides con-tinuous total pressure measurement (seeFig. 4.4). The cathodes are made of iridiumwire and have a thorium oxide coating toreduce the work associated with electrondischarge. (For some time now the thori-um oxide has gradually been replaced byyttrium oxide.) These coatings reduce theelectron discharge work function so thatthe desired emission flow will be achievedeven at lower cathode temperatures. Avai-lable for special applications are tungsten cathodes (insensitive to hydrocarbons butsensitive to oxygen) or rhenium cathodes(insensitive to oxygen and hydrocarbonsbut evaporate slowly during operation dueto the high vapor pressure).

Shielding

CathodeAnode

Focussing plate(extractor diaphragm) Ion source exit

diaphragm(total pressure measurement)

Quadrupole exitdiaphragm

Ion detectorQuadrupole separation systemIon source

Fig. 4.2 Schematic for quadrupole mass spectrometer Fig. 4.3 Quadrupole mass spectrometer – Extractor ionization vacuum gauge

Extractor measurement system

Transpector measurement head

Anode

Cathode

Ref lec-tor

Ion trap

Fig. 4.4 Open ion source

Cathode

Shielding

Extractor diaphragm

Total pressure diaphragm

Anode

D00 E 19.06.2001 21:39 Uhr Seite 90



4.3.1.2 The quadrupole separationsystem

Here the ions are separated on the basis oftheir mass-to-charge ratio. We know fromphysics that the deflection of electricallycharged particles (ions) from their trajec-tory is possible only in accordance withtheir ratio of mass to charge, since theattraction of the particles is proportional tothe charge while the inertia (which resistschange) is proportional to its mass. Theseparation system comprises four cylin-drical metal rods, set up in parallel and iso-lated one from the other; the two opposingrods are charged with identical potential.Fig. 4.2 shows schematically the arrange-ment of the rods and their power supply.The electrical field Φ inside the separationsystem is generated by superimposing aDC voltage and a high-frequency AC volta-ge:

Φ = (U + V ⋅ cos ωt) · (x2 – y2) / r02

r0 = radius of the cylinder which can beinscribed inside the system of rods

Exerting an effect on a single charged ionmoving near and parallel to the center lineinside the separation system and perpen-dicular to its movement are the forces

The mathematical treatment of these equa-tions of motion uses Mathieu’s differentialequations. It is demonstrated that thereare stable and unstable ion paths. With thestable paths, the distance of the ions fromthe separation system center line alwaysremains less than ro (passage condition).With unstable paths, the distance from theaxis will grow until the ion ultimately colli-des with a rod surface. The ion will bedischarged (neutralized), thus becomingunavailable to the detector (blocking con-dition).Even without solving the differential equa-tion, it is possible to arrive at a purely phe-nomenological explanation which leads toan understanding of the most importantcharacteristics of the quadrupole sepa-ration system.

If we imagine that we cut open the se-

F er

x tx = − ⋅ ⋅ ⋅2

02

cos ( )ω

F er

y ty = − ⋅ ⋅ ⋅2

02

cos ( )ω

Fz = 0

paration system and observe the deflec-tion of a singly ionized, positive ion withatomic number M, moving in two planes,which are perpendicular one to the otherand each passing through the centers oftwo opposing rods. We proceed step-by-step and first observe the xz plane (Fig.4.5, left) and then the yz plane (Fig.4.5,right):1. Only DC potential U at the rods:

xz plane (left): Positive potential of +Uat the rod, with a repellant effect on theion, keeping it centered; it reaches thecollector (→ passage).

yz plane (right): Negative potential onthe rod -U, meaning that at even the tiniest deviations from the center axisthe ion will be drawn toward the nearestrod and neutralized there; it does notreach the collector (→ blocking).

2. Superimposition of high-frequencyvoltage V · cos ω t:

xz plane (left): Rod potential +U + V · cos ω t. Withrising AC voltage amplitude V the ionwill be excited to execute transverseoscillations with ever greater amplitu-des until it makes contact with a rodand is neutralized. The separationsystem remains blocked for very largevalues of V.

yz plane (right): Rod potential -U -V · cos ω t. Here againsuperimposition induces an additionalforce so that as of a certain value for V

the amplitude of the transverse oscilla-tions will be smaller than the clearancebetween the rods and the ion can passto the collector at very large V.

3. Ion emission i+ = i+ (V) for a fixed massof M:

xz plane (left): For voltages of V < V1 thedeflection which leads to an escalationof the oscillations is smaller than V1, i.e.still in the “pass” range. Where V > V1the deflection will be sufficient to induce escalation and thus blockage.

yz plane (right): For voltages of V < V1the deflection which leads to the dam-ping of the oscillations is smaller thanV1, i.e. still in the “block” range. WhereV > V1 the damping will be sufficient tosettle oscillations, allowing passage.

4. Ion flow i+ = i+ (M) for a fixed ratio of U / V:

Here the relationships are exactly oppo-site to those for i+ = i+ (V) since theinfluence of V on light masses is great-er than on heavy masses.

xz plane: For masses of M < M1 thedeflection which results in escalation ofthe oscillations is greater than at M1,which means that the ions will beblocked. At M > M1 the deflection is nolonger sufficient for escalation, so thatthe ion can pass.

yz plane: For masses of M < M1 thedeflection which results in damping ofthe oscillations is greater than at M1,which means that the ion will pass. AtM > M1 the damping is not sufficient tocalm the system and so the ion isblocked.

5. Combination of the xz and yz planes. Inthe superimposition of the ion currentsi+ = i+ (M) for both pairs of rods (U / Vbeing fixed) there are three importantranges:

Range I : No passage for M due to theblocking behavior of the xz pair of rods.

Range II : The pass factor of the rodsystems for mass M is determined bythe U/V ratio (other ions will not pass).We see that great permeability (corre-sponding to high sensitivity) is boughtat the price of low selectivity (= resolu-tion, see Section 4.5). Ideal adjustmentof the separation system thus requiresa compromise between these two prop-

D00

xz plane yz plane+

+

+

+

+

+

-

-

+

-

-

+

Superimposition of the xy and yz planes

Rod:+U

Transmission:full

Rod:+U+V, cos ω

Transmission:low-pass

i+ i+

i+

VV1 V1 V

M

i+

MM1 M1

i+

M

I IIIII

yz xz

Rod:–U

Transmission:none

Rod:–U–V · cos ωTransmission:

high-pass

( )UV fixed

UV

.. Selectivity (resolution) Sensitivity

1

2

3

4

5

Fig. 4.5 Phenomenological explanation of the separation system

D00 E 19.06.2001 21:39 Uhr Seite 91



erties. To achieve constant resolution,the U/V ratio will remain constant overthe entire measurement range. The“atomic number” M (see 4.6.1) of theions which can pass through the sepa-ration system must satisfy this condi-tion:

V = High-frequency amplitude, rO = Quadrupole inscribed radiusf = High-frequencyAs a result of this linear dependencythere results a mass spectrum with li-near mass scale due to simultaneous,proportional modification of U and V.

Range III : M cannot pass, due to theblocking characteristics of the yz pair ofrods.

4.3.1.3 The measurement system(detector)

Once they have left the separation systemthe ions will meet the ion trap or detectorwhich, in the simplest instance, will be inthe form of a Faraday cage (Faraday cup).In any case the ions which impinge on thedetector will be neutralized by electronsfrom the ion trap. Shown, after electricalamplification, as the measurement signalitself is the corresponding “ion emissionstream”. To achieve greater sensitivity, asecondary electron multiplier pickup(SEMP) can be employed in place of theFaraday cup.

me

MV

f ro≈ =

⋅ ⋅14.438 2 2

Channeltrons or Channelplates can beused as SEMPs. SEMPs are virtually iner-tia-free amplifiers with gain of about 10+6

at the outset; this will indeed drop offduring the initial use phase but will thenbecome virtually constant over a long peri-od of time. Fig. 4.6 shows at the left thebasic configuration of a Faraday ion trapand, on the right, a section through aChanneltron. When recording spectra thescanning period per mass line t0 and thetime constants of the amplifier t shouldsatisfy the condition that t0 = 10 τ. Inmodern devices such as the

TRANSPECTOR the otherwise unlimitedselection of the scanning period and theamplifier time constants will be restrictedby microprocessor control to logical pairsof values.

Amplifier

Amplifier

Connectionto front endof the insidesurface

Collector

Positive ion

Negative high voltageResistance ≈ 108 Ω

Resistance of the inner surface

R ≈ 4 · 106 Ω

Separation system output

Electron suppressor

Faraday cup

Fig. 4.6 Left: principle of the Faraday cup; right: configuration of the Channeltron

SS compensates for L No segregationeff 2 →

Massspectrometer

L molecular d2 L→ Àλ2

S L S ~eff 1 eff→À

L2 ~ 1CDM

L1 ~ 1

1CD

CD

M

M

Non-segregating gas inlet system

Stage B Stage A

p = 1 ... 10 mbar p = 10 ... 1000 mbarp 10 mbar≤ –4

Capillary

NosegregationLaminar flowMolecular flow

L2 L3

L1Seff QPumping

Q QHV Pumping¿

QHV

_

_

(Transition laminar/molecular)

Fig. 4.7 Principle of the pressure converter (stage B only in the single-stage version and stages A and B in two-stage units)

D00 E 19.06.2001 21:39 Uhr Seite 92



4.4 Gas admission andpressure adaptation

4.4.1 Metering valveThe simplest way to adapt a classical massspectrometer to pressures exceeding 1 · 10-4 mbar is by way of a meteringvalve. The inherent disadvantage is, howe-ver, that since the flow properties are notunequivocally defined, a deviation from theoriginal gas composition might result.

4.4.2 Pressure converterIn order to examine a gas mix at total pres-sure exceeding 1 · 10-4 mbar it is neces-sary to use pressure converters which willnot segregate the gases. Figure 4.7 is usedto help explain how such a pressure con-verter works:

a. Process pressure < 1 mbar: Single-stage pressure converter. Gas is allowed to pass out of the vacuumvessel in molecular flow, through a dia-phragm with conductance value L2 andinto the “sensor chamber” (with its ownhigh vacuum system). Molecular flow cau-ses segregation but this will be indepen-dent of the pressure level (see Section1.5). A second diaphragm with molecularflow, located between the sensor chamberand the turbomolecular pump, will com-pensate for the segregation occurring atL2.

b. Process pressure > 1 mbar: Two-stagepressure converter. Using a small (rotaryvane) pump a laminar stream of gas isdiverted from the rough vacuum areathrough a capillary or diaphragm (conduc-tance value L3). Prior to entry into thepump, at a pressure of about 1 mbar, asmall part of this flow is again allowed toenter the sensor chamber through the dia-phragm with conductance value L2, againas molecular flow.

A falsification of the gas compositionresulting from adsorption and condensati-on can be avoided by heating the pressureconverter and the capillary.

To evaluate the influence on the gas com-position by the measurement unit itself,information on the heating temperature,the materials and surface areas for themetallic, glass and ceramic componentswill be needed along with specifications onthe material and dimensions of the catho-de (and ultimately regarding the electronimpact energy for the ion source as well).

4.4.3 Closed ion source (CIS)In order to curb – or avoid entirely – influ-ences which could stem from the sensorchamber or the cathode (e.g. disturbanceof the CO-CO2 equilibrium by heating thecathode) a closed ion source (CIS) will beused in many cases (see Fig 4.8).

D00

Process: 10-2 mbar Process: 10-2 mbar

Impact chamberCathode chamber„Exit diaphragm“

10-5 10-3

10-5 10-5

Exit diaphragm

PumpPump

10-5 10-5

10-5 10-5

(Elastomer) (Metal)ValveInlet diaphragm

Example of the sputterprocessTo be detected is 1 ppm N ascontamination in argon,the working gas

2

1 ppm N 2 at the inlet:

10-6 ·10-5 mbar = 10-11 mbar

Background:Residual gas (valve closed) 10-6 mbar totaltotal, including 1% by mass 28 : 10-8 mbar

Background noise

Signal at 1‰ of backgroundcannot be detected

1 ppm N 2 at the inlet:

10-6 ·10-3 mbar = 10-9 mbar

Background:Residual gas (valve closed) 10-7 mbar totaltotal, including 1% by mass 28 : 10-9 mbar

1 ppm

Background noiseSignal twice the background noiseamplitude; can just be clearly detected

Fig. 4.8 Open ion source (left) and closed ion source (right)

D00 E 19.06.2001 21:39 Uhr Seite 93



The CIS is divided into two sections: acathode chamber where the electrons areemitted, and an impact chamber, wherethe impact ionization of the gas particlestakes place. The two chambers are pum-ped differentially: the pressure in thecathode chamber comes to about 10-5 mbar, that in the impact room about10-3 mbar. The gas from the vacuumchamber is allowed to pass into the impactchamber by way of a metal-sealed, bakea-ble valve (pressure converter, ultrahighvacuum technology). There high-yieldionization takes place at about 10-3 mbar.The electrons exerting the impact are emit-ted in the cathode chamber at about 10-5 mbar and pass through small ope-nings from there into the impact chamber.The signal-to-noise ratio (residual gas) viaà vis the open ion source will be increasedoverall by a factor of 10+3 or more. Figure4.8 shows the fundamental difference bet-ween the configurations for open and clo-sed ion sources for a typical application insputter technology. With the modifieddesign of the CIS compared with the openion source in regard to both the geometryand the electron energy (open ion source102 eV, CIS 75 or 35 eV), different frag-ment distribution patterns may be foundwhere a lower electron energy level is sel-ected. For example, the argon36++ isotopeat mass of 18 cannot be detected at elec-tron energy of less than 43.5 eV and cantherefore not falsify the detection of H2O+

at mass 18 in the sputter processes usingargon as the working gas – processeswhich are of great importance in industry.

4.4.4 Aggressive gas monitor(AGM)

In many cases the process gas to beexamined is so aggressive that the catho-de would survive for only a short period oftime. The AGM uses the property of lami-nar flow by way of which there is no“reverse” flow of any kind. Controlled witha separate AGM valve, a part of the wor-king gas fed to the processes is introducedas “purging gas”, ahead of the pressureconverter, to the TRANSPECTOR; this setsup a flow toward the vacuum chamber.Thus process gas can reach the TRANS-PECTOR only with the AGM valve closed.When the valve is open the TRANSPEC-TOR sees only pure working gas. Fig. 4.9shows the AGM principle.

4.5 Descriptive valuesin mass spectrometry (specifications)

A partial pressure measurement unit ischaracterized essentially by the followingproperties (DIN 28 410):

4.5.1 Line width (resolution)

The line width is a measure of the degreeto which differentiation can be made bet-ween two adjacent lines of the sameheight. The resolution is normally indica-ted. It is defined as: R = M / ∆M and is con-stant for the quadrupole spectrometeracross the entire mass range, slightlygreater than 1 or ∆M < 1.

Often an expression such as “unit resoluti-on with 15% valley” is used. This meansthat the “bottom of the valley” betweentwo adjacent peaks of identical heightcomes to 15 % of the height of the peak or,put another way, at 7.5 % of its peakheight the line width DM measured acrossan individual peak equals 1 amu (atomicmass unit); see in this context the sche-matic drawing in Fig. 4.10.

Process:e. g. 50 mbar

Impact chamberCathode chamber

“Exit diaphragms”

10-5

10-3 10-5

10-5 PumpDiaphragm

Working gas for the process (Ar)

AGM protective gas valve

Fig. 4.9 Principle behind the aggressive gas monitor (AGM)

D00 E 19.06.2001 21:39 Uhr Seite 94



4.5.2 Mass range

The mass range is characterized by theatomic numbers for the lightest and hea-viest ions with a single charge which aredetected with the unit.

4.5.3 Sensitivity

Sensitivity E is the quotient of the measu-red ion flow and the associated partialpressure; it is normally specified for argonor nitrogen:

(4.1)

Typical values are:

Faraday cup:

SEM:

4.5.4 Smallest detectable partial pressure

The smallest detectable partial pressure isdefined as a ratio of noise amplitude tosensitivity:

Æ á i+R

= Noise amplitude

Example (from Fig. 4.11):

Sensitivity E =

Noise amplitude Æ á i+R

= 4 · 10–14 A

1 10 4⋅ – Ambar

Pi

E mbarRmin

⋅( )=

+∆

E Ambar

= ⋅ +1 10 2

E Ambar

= ⋅ –1 10 4

Eip

AmbarG

=

+

4.5.5 Smallest detectable partial pressure ratio(concentration)

The definition is:

SDPPR = pmin / pΣ (ppm)

This definition, which is somewhat “clum-sy” for practical use, is to be explainedusing the detection of argon36 in the air asthe example: Air contains 0.93 % argon byvolume; the relative isotope frequency ofAr40 to Ar36 is 99.6 % to 0.337 %. Thusthe share of Ar36 in the air can be calcula-ted as follows:

0.93 · 10–2 · 0.337 · 10–2 = 3.13 · 10–5 = 31.3 ppm

Figure 4.11 shows the screen print-out forthe measurement. The peak height for Ar36in the illustration is determined to be 1.5 · 10-13 A and noise amplitude ∆ · i+

Rto

be 4 · 10-14 A. The minimum detectableconcentration is that concentration at whichthe height of the peak is equal to the noiseamplitude. This results in the smallest mea-surable peak height being 1.5 · 10-13 A/2.4 · 10-14 A = 1.875. Thesmallest detectable concentration is thenderived from this by calculation to arrive at:

31.3 · 10–6 / 1.875 = 16.69 · 10–6

= 16.69 ppm.

4.5.6 Linearity rangeThe linearity range is that pressure rangefor the reference gas (N2, Ar) in which sen-sitivity remains constant within limits

p FCA

A mbarmbarmin( )

/=

⋅⋅

= ⋅–

––4 10

1 104 10

14

410

which are to be specified (± 10 % for par-tial pressure measurement devices).

In the range below 1 · 10-6 mbar the rela-tionship between the ion flow and partialpressure is strictly linear. Between 1 · 10-6

mbar and 1 · 10-4 mbar there are minordeviations from linear characteristics.Above 1 · 10-4 mbar these deviations growuntil, ultimately, in a range above 10-2

mbar the ions for the dense gas atmos-phere will no longer be able to reach theion trap. The emergency shut-down for thecathode (at excessive pressure) is almostalways set for 5 · 10-4 mbar. Depending onthe information required, there will be dif-fering upper limits for use.

In analytical applications, 1 · 10-6 mbarshould not be exceeded if at all possible.The range from 1 · 10-6 mbar to 1 · 10-4

mbar is still suitable for clear depictions ofthe gas composition and partial pressureregulation (see Fig. 4.12).

4.5.7 Information on surfacesand amenability to bake-out

Additional information required to evaluatea sensor includes specifications on thebake-out temperature (during measure-ment or with the cathode or SEMP swit-ched off), materials used and surfaceareas of the metal, glass and ceramic com-ponents and the material and dimensionsfor the cathode; data is also needed on theelectron impact energy at the ion source(and on whether it is adjustable). Thesevalues are critical to uninterrupted operati-on and to any influence on the gas compo-sition by the sensor itself.

D00

M

∆M

1 amu

M + 1 Atomic number

7,5%

15%

i+

100%

Fig. 4.10 Line width – 15 % valley Fig. 4.11 Detection of Argon36

10–8 10–7 10–6 10–5 10–4 10–3 log P

Automatic shut-down:5 · 10≈ –4

log i+

Exact measurementrange

Regulation

Fig. 4.12 Qualitative linearity curve

D00 E 19.06.2001 21:39 Uhr Seite 95



4.6 Evaluating spectra

4.6.1 Ionization and fundamen-tal problems in gas ana-lysis

Continuous change in the voltages appliedto the electrodes in the separation system(“scanning”) gives rise to a relationshipbetween the ion flow I+ and the “atomicnumber” which is proportional to the m/eratio and expressed as:

(4.2)

(Mr = relative molar mass, ne = number of elementary charges e)

This is the so-called mass spectrum, i+ = i+(M). The spectrum thus shows thepeaks i+ as ordinates, plotted against theatomic number M along the abscissa. Oneof the difficulties in interpreting a massspectrum such as this is due to the factthat one and the same mass as per theequation (4.2) may be associated withvarious ions. Typical examples, amongmany others, are: The atomic number M = 16 corresponds to CH4

+ and O2++;

M = 28 for CO+, N2+ and C2H+! Particular

attention must therefore be paid to the fol-lowing points when evaluating spectra:

1) In the case of isotopes we are dealingwith differing positron counts in the nucle-us (mass) of the ion at identical nuclearcharge numbers (gas type). Some valuesfor relative isotope frequency are compiledin Table 4.2.

2) Depending on the energy of the impac-ting electrons (equalling the potential diffe-rential, cathode – anode), ions may be eit-her singly or multiply ionized. For example,one finds Ar+ at mass of 40, Ar++ at massof 20 and Ar+++ at mass of 13.3. At massof 20 one will, however, also find neon,Ne+. There are threshold energy levels forthe impacting electrons for all ionizationstates for every type of gas, i.e., each typeof ion can be formed only above the asso-ciated energy threshold. This is shown forAr in Fig. 4.13.

3) Specific ionization of the variousgases Sgas, this being the number of ionsformed, per cm and mbar, by collisionswith electrons; this will vary from one typeof gas to the next. For most gases the ion

M Mn

r

e=

yield is greatest at an electron energy levelbetween about 80 and 110 eV; see Fig.4.14.

In practice the differing ionization rates forthe individual gases will be taken intoaccount by standardization against nitro-gen; relative ionization probabilities(RIP) in relationship to nitrogen will beindicated (Table 4.3).

4) Finally, gas molecules are often brokendown into fragments by ionization. Thefragment distribution patterns thus crea-ted are the so-called characteristic spectra(fingerprint, cracking pattern).

Important: In the tables the individual frag-ments specified are standardized eitheragainst the maximum peak (in % or ‰ ofthe highest peak) or against the total of allpeaks (see the examples in Table 4.4).

Element Ordinal- Atomic Relativenumber number frequency

H 1 1 99.985 2 0.015

He 2 3 0.000134 ≈ 100.0

B 5 10 19.7811 80.22

C 6 12 98.89213 1.108

N 7 14 99.6315 0.37

O 8 16 99.75917 0.037418 0.2039

F 9 19 100.0Ne 10 20 90.92

21 0.25722 8.82

Na 11 23 100.0Al 13 27 100.0Si 14 28 92.27

29 4.6830 3.05

P 15 31 100.0

Table 4.2 Relative frequency of isotopes

0

2

4

6

8

10

12

100 200 300 400 500

Ar +

Electron energy (eV)

Ions

form

ed p

er c

m ·

mba

r

Fig. 4.13 Number of the various Ar ions produced, as a factor of electron energy level

Ar4+ 200 eV

Element Ordinal- Atomic Relativenumber number frequency

S 16 32 95.0633 0.7434 4.1836 0.016

Cl 17 35 75.437 24.6

Ar 18 36 0.33738 0.06340 99.60

Kr 36 78 0.35480 2.2782 11.5683 11.5584 56.9086 17.37

Xe 54 124 0.096126 0.090128 1.919129 26.44130 4.08131 21.18132 26.89134 10.44136 8.87

Table 4.2 Relative frequency of isotopes

Thresholdenergy for argon ions

Ar+ 15,7 eV Ar++ 43,5 eV Ar3+ 85,0 eV

D00 E 19.06.2001 21:39 Uhr Seite 96



Both the nature of the fragments createdand the possibility for multiple ionizationwill depend on the geometry (differing ionnumber, depending on the length of theionization path) and on the energy of theimpacting electrons (threshold energy forcertain types of ions). Table values arealways referenced to a certain ion sourcewith a certain electron energy level. This iswhy it is difficult to compare the results

obtained using devices made by differentmanufacturers.

Often the probable partial pressure for oneof the masses involved will be estimatedthrough critical analysis of the spectrum.Thus the presence of air in the vacuumvessel (which may indicate a leak) is mani-fested by the detection of a quantity of O2

+

(with mass of 32) which is about one-

quarter of the share of N2+ with its mass of

28. If, on the other hand, no oxygen isdetected in the spectrum, then the peak atatomic number 28 would indicate carbonmonoxide. In so far as the peak at atomicnumber 28 reflects the CO+ fragment ofCO2 (atomic number 44), this share is 11 % of the value measured for atomicnumber 44 (Table 4.5). On the other hand,in all cases where nitrogen is present, ato-mic number 14 (N2

++) will always be foundin the spectrum in addition to the atomicnumber 28 (N2

+); in the case of carbonmonoxide, on the other hand, there willalways appear – in addition to CO+ – thefragmentary masses of 12 (C+) and 16 (O2

++)).

Figure 4.15 uses a simplified example of a“model spectrum” with superimpositionsof hydrogen, nitrogen, oxygen, watervapor, carbon monoxide, carbon dioxide,neon and argon to demonstrate the diffi-culties involved in evaluating spectra.

D00

0 5 10 15 20 25 30 35 40 45 50

HydrogenNitrogen

OxygenWater

Carbon dioxideNeon

ArgonCarbon monoxide

H+

H+

H2+

H2+

Ne++

C+

C+

13C+

N+

O+

O+

O+

O+

H3O+

OH+

H2O+

Ne+

Ar++

22Ne+

N2+

CO+

CO+

13CO+

14N15N+

O2+

16O18O+ 36Ar+

Ar+ CO2+

13C16O2+

Fig. 4.15 Model spectrum

Evaluation problems: The peak at atomic number 16 may, for example, be due to oxygen fragments resulting from O2, H2O, CO2 and CO; the peak at ato-mic number 28 from contributions by N2 as well as by CO and CO as a fragment of CO2; the peak at atomic number 20 could result from singly ionizedNe and double-ionized Ar

Fig. 4.14 Specific ionization S for various gases by electrons exhibiting energy level E

Electron energy (eV)

Ions

form

ed p

er c

m ·

mba

r

D00 E 19.06.2001 21:39 Uhr Seite 97



Type of gas Symbol RIP Type of gas Symbol RIPAcetone (Propanone) (CH3)2CO 3.6 Hydrogen chloride HCl 1.6Air 1.0 Hydrogen fluoride HF 1.4Ammonia NH3 1.3 Hydrogen iodide HI 3.1Argon Ar 1.2 Hydrogen sulfide H2S 2.2Benzene C6H6 5.9 Iodine I2Benzoic acid C6H5COOH 5.5 Krypton Kr 1.7Bromine Br 3.8 Lithium Li 1.9Butane C4H10 4.9 Methane CH4 1.6Carbon dioxide CO2 1.4 Methanol CH3OH 1.8Carbon disulfide CS2 4.8 Neon Ne 0.23Carbon monoxide CO 1.05 Nitrogen N2 1.0Carbon tetrachloride CCl4 6.0 Nitrogen oxide NO 1.2Chlorobenzene C6H4Cl 7.0 Nitrogen dioxide N2O 1.7Chloroethane C2H3Cl 4.0 Oxygen O2 1.0Chloroform CHCl3 4.8 n-pentane C5H17 6.0Chlormethane CH3Cl 3.1 Phenol C6H5OH 6.2Cyclohexene C6H12 6.4 Phosphine PH3 2.6Deuterium D2 0.35 Propane C3H8 3.7Dichlorodifluoromethane CCl2F2 2.7 Silver perchlorate AgClO4 3.6Dichloromethane CH2Cl2 7.8 Tin iodide Snl4 6.7Dinitrobenzene C6H4(NO2)2 7.8 Sulfur dioxide SO2 2.1Ethane C2H6 2.6 Sulfur hexafluoride SF6 2.3Ethanol C2H5OH 3.6 Toluene C6H5CH3 6.8Ethylene oxide (CH2)2O 2.5 Trinitrobenzene C6H3(NO2)3 9.0Helium He 0.14 Water vapor H2O 11.0Hexane C6H14 6.6 Xenon Xe 3.0Hydrogen H2 0.44 Xylols C6H4(CH3)2 7.8

Table 4.3 Relative ionization probabilities (RIP) vis à vis nitrogen, electron energy 102 eV

Electron energy : 75 eV (PGA 100) 102 eV (Transpector)Gas Symbol Mass Σ = 100 % Greatest peak = 100 % Σ = 100 % Greatest peak = 100 %Argon Ar 40 74.9 100 90.9 100

20 24.7 33.1 9.1 1036 0.3

Carbon dioxide CO2 45 0.95 1.3 0.8 144 72.7 100 84 10028 8.3 11.5 9.2 1116 11.7 16.1 7.6 912 6.15 8.4 5 6

Carbon monoxide CO 29 1.89 2.0 0.9 128 91.3 100 92.6 10016 1.1 1.2 1.9 214 1.7 1.9 0.812 3.5 3.8 4.6 5

Neon Ne 22 9.2 10.2 0.9 1120 89.6 100 90.1 10010 0.84 0.93 9 4

Oxygen O2 34 0.45 0.5332 84.2 100 90.1 10016 15.0 17.8 9.9 11

Nitrogen N2 29 0.7 0.8 0.9 128 86.3 100 92.6 10014 12.8 15 6.5 12

Water vapor H2O 19 1.4 2.318 60 100 74.1 10017 16.1 27 18.5 2516 1.9 3.2 1.5 22 5.0 8.4 1.5 21 15.5 20 4.4 6

Table 4.4 Fragment distribution for certain gases at 75 eV and 102 eV

D00 E 19.06.2001 21:39 Uhr Seite 98



D00

No Gas Symbol 1 = 100 2 3 4 5 6

1 Acetone (CH3)2CO 43/100 15/42 58/20 14/10 27/19 42/8

2 Air 28/100 32/27 14/6 16/3 40/1 -

3 Ammonia NH3 17/100 16/80 15/8 14/2 - -

4 Argon Ar 40/100 20/10 - - - -

5 Benzene C6H6 78/100 77/22 51/18 50/17 52/15 39/10

6 Carbon dioxide CO2 44/100 28/11 16/9 12/6 45/1 22/1

7 Carbon monoxide CO 28/100 12/5 16/2 29/1 - -

8 Carbon tetrachloride CCl4 117/100 119/91 47/51 82/42 35/39 121/29

9 Carbon tetrafluoride CF4 69/100 50/12 31/5 19/4 - -

10 Diff. pump oil, DC 705 78/100 76/83 39/73 43/59 91/32 -

11 Diff. pump oil, Fomblin 69/100 20/28 16/16 31/9 97/8 47/8

12 Diff. pump oil, PPE 50/100 77/89 63/29 62/27 64/21 38/7

13 Ethanol CH3CH2OH 31/100 45/34 27/24 29/23 46/17 26/8

14 Halocarbon 11 CCl3F 101/100 103/60 35/16 66/15 47/12 31/10

15 Halocarbon 12 CCl2F2 85/100 87/32 50/16 35/12 - -

16 Halocarbon 13 CClF3 69/100 85/15 50/14 31/9 35/7 87/5

17 Halocarbon 14 CF4 69/100 12/7 19/6 31/5 50/8 -

18 Halocarbon 23 CHF3 51/100 31/58 69/40 50/19 52/1 21/1

19 Halocarbon 113 C2C13F3 101/100 103/62 85/55 31/50 151/41 153/25

20 Helium He 4/100 - - - - -

21 Heptane C7H16 43/100 41/62 29/49 27/40 57/34 71/28

22 Hexane C6H14 41/100 43/92 57/85 29/84 27/65 56/50

23 Hydrogen H2 2/100 1/5 - - - -

24 Hydrogen sulfide H2S 34/100 32/44 33/42 36/4 35/2 -

25 Isopropyl alcohol C3H8O 45/100 43/16 27/16 29/10 41/7 39/6

26 Krypton Kr 84/100 86/31 83/20 82/20 80/4 -

27 Methane CH4 16/100 15/85 14/16 13/8 1/4 12/2

28 Mehtyl alcohol CH3OH 31/100 29/74 32/67 15/50 28/16 2/16

29 Methyl ethyl ketone C4H8O 43/100 29/25 72/16 27/16 57/6 42/5

30 Mechanical pump oil 43/100 41/91 57/73 55/64 71/20 39/19

31 Neon Ne 20/100 22/10 10/1 - - -

32 Nitrogen N2 28/100 14/7 29/1 - - -

33 Oxygen O2 32/100 16/11 - - - -

34 Perfluorokerosene 69/100 119/17 51/12 131/11 100/5 31/4

35 Perfluor-tributylamine C12F27N 69/100 131/18 31/6 51/5 50/3 114/2

36 Silane SiH4 30/100 31/80 29/31 28/28 32/8 33/2

37 Silicon tetrafluoride SiF4 85/100 87/12 28/12 33/10 86/5 47/5

38 Toluene C6H5CH3 91/100 92/62 39/12 65/6 45.5/4 51/4

39 Trichloroethane C2H3Cl3 97/100 61/87 99/61 26/43 27/31 63/27

40 Trichloroethylene C2HCl3 95/100 130/90 132/85 97/64 60/57 35/31

41 Trifluoromethane CHF3 69/100 51/91 31/49 50/42 12/4 -

42 Turbomolecular pump oil 43/100 57/88 41/76 55/73 71/52 69/49

43 Water vapor H2O 18/100 17/25 1/6 16/2 2/2 -

44 Xenon Xe 132/100 129/98 131/79 134/39 136/33 130/15

Table 4.5 Spectrum library of the 6 highest peaks for the TRANSPECTOR

D00 E 19.06.2001 21:39 Uhr Seite 99



4.6.2 Partial pressure measurement

The number of ions i+gas produced from agas in the ion source is proportional to theemission current i–, to the specific ioni-zation Sgas, to a geometry factor f repre-senting the ionization path inside theionization source, to the relative ionizationprobability RIPgas, and to the partial pres-sure pgas. This number of ions producedis, by definition, made equal to the sensi-tivity Egas times the partial pressure pgas:

i+gas (produced) = i- · Sgas · f · RIPgas · pgas= Egas · pgas

due to RIPN2 = 1

EN2 is equal to i– · SN2 · f

and Egas is equal to EN2 · RIPgas

Almost all gases form fragments duringionization. To achieve quantitative evalu-ation one must either add the ion flows atthe appropriate peaks or measure (with aknown fragment factor [FF]) one peak andcalculate the overall ion flow on that basis:

In order to maintain the number of ionsarriving at the ion trap, it is necessary tomultiply the number above with the trans-mission factor TF(m), which will be depen-dent on mass, in order to take into accountthe permeability of the separation systemfor atomic number m (analogous to this,there is the detection factor for the SEMP;it, however, is often already contained inTF). The transmission factor (also: ion-optical transmission) is thus the quotientof the ions measured and the ions produ-ced.

Thus

and with

Egas = EN2 · RIPgas

the ultimate result is:

( )p

i producedBF Egas

gas m

gas m gas=

⋅

+,

,

2

2

⇒( )

pi measuredBF E TF mgas

gas m

gas m gas=

⋅ ⋅

+,

, ( )2

2

i produced i iiBF

iBF

E p

gas gas m gas mgas m

gas mgas m

gas mgas gas

+ + ++

+= + + =

= = = ⋅

( ) ....

....

, ,,

,,

,

1 21

12

2

(4.3)

The partial pressure is calculated from theion flow measured for a certain fragmentby multiplication with two factors. The firstfactor will depend only on the nitrogensensitivity of the detector and thus is aconstant for the device. The second willdepend only on the specific ion properties.

These factors will have to be entered sepa-rately for units with direct partial pressureindication (at least for less common typesof ions).

4.6.3 Qualitative gas analysis

The analysis of spectra assumes certainworking hypotheses:

1. Every type of molecule produces a cer-tain, constant mass spectrum or frag-ment spectrum which is characteristicfor this type of molecule (fingerprint,cracking pattern).

2. The spectrum of every mixture of gasesis the same as would be found throughlinear superimposition of the spectra ofthe individual gases. The height of thepeaks will depend on the gas pressure.

3. The ion flow for each peak is proportio-nal to the partial pressure of each com-ponent responsible for the peak. Sincethe ion flow is proportional to the parti-al pressure, the constant of proportio-nality (sensitivity) varies from one gasto the next.

Although these assumptions are notalways correct (see Robertson: MassSpectrometry) they do represent a usefulworking hypothesis.

In qualitative analysis, the unknown spec-trum is compared with a known spectrumin a library. Each gas is “definitively deter-mined” by its spectrum. The comparisonwith library data is a simple pattern reco-gnition process. Depending on the availa-bility, the comparison may be made usingany of a number ancillary aids. So, forexample, in accordance with the position,size and sequence of the five or ten highestpeaks. Naturally, comparison is possible

( )p i measured

E BF RIP TF m

gas gas m

N gas m gas

= ⋅

⋅⋅ ⋅

+,

, ( )

2

2 2

1 1

only after the spectrum has been standar-dized, by setting the height of the highestline equal to 100 or 1000 (see Table 4.5 asan example).

The comparison can be made manually onthe basis of collections of tables (forexample, A. Cornu & R. Massot: Compil-ation of Mass Spectral Data) or may beeffected with computer assistance; largedatabases can be used (e.g. Mass SpectralData Base, Royal Society of Chemistry,Cambridge).

When making comparisons with libraryinformation, it is necessary to pay attenti-on to whether identical ion sources or atleast identical electron impact energieswere used.

All these capabilities are, however, gene-rally too elaborate for the problemsencountered in vacuum technology. Manycommercial mass spectrometers can showa number of library spectra in the screenso that the user can see immediatelywhether the “library substance” might becontained in the substance measured.Usually the measured spectrum was theresult of a mix of gases and it is particu-larly convenient if the screen offers thecapacity for subtracting (by way of trial)the spectra of individual (or several) gasesfrom the measured spectrum. The gas canbe present only when the subtraction doesnot yield any negative values for the majorpeaks. Figure 4.16 shows such a step-by-step subtraction procedure using theTranspector-Ware software.

Regardless of how the qualitative analysisis prepared, the result is always just a“suggestion”, i.e. an assumption as towhich gases the mixture might contain.This suggestion will have still to be exami-ned, e.g. by considering the likelihood thata certain substance would be contained inthe spectrum. In addition, recording a newspectrum for this substance can help toachieve clarity.

D00 E 19.06.2001 21:39 Uhr Seite 100



4.6.4 Quantitative gas analysis

Particular difficulties are encounteredwhen interpreting the spectrum of an unknown mixture of gases. The proporti-ons of ion flow from various sources canbe offset one against the other only afterall the sources have been identified. Inmany applications in vacuum technologyone will be dealing with mixtures of a fewsimple gases of known identity, with ato-mic numbers less of than 50 (whereby theprocess-related gases can representexceptions). In the normal, more compli-cated case there will be a spectrum with amultitude of superimpositions in a com-pletely unknown mixture of many gascomponents; here a qualitative analysiswill have to be made before attemptingquantitative analysis. The degree of diffi-culty encountered will depend on the num-ber of superimpositions (individual/a few/many).

In the case of individual superimpositions,mutual, balancing of the ion flows duringmeasurement of one and the same type ofgas for several atomic numbers can oftenbe productive.

Where there is a larger number of super-impositions and a limited number of gasesoverall, tabular evaluation using correctionfactors vis à vis the spectrum of a calibra-tion gas of known composition can oftenbe helpful.

In the most general case a plurality ofgases will make a greater or lesser contri-bution to the ion flow for all the masses.The share of a gas g in each case for theatomic number m will be expressed by thefragment factor Ffm,g. In order to simplifycalculation, the fragment factor Ffm,g willalso contain the transmission factor TFand the detection factor DF. Then the ioncurrent to mass m, as a function of theoverall ion currents of all the gases invol-ved, in matrix notation, is:

The ion current vector for the atomic num-bers m (resulting from the contributionsby the fragments of the individual gases) is

i

i

i

BF BF

FF

BF BF

I

I

I

j

m

u

j k j o

m g

u k u o

k

g

o

+

+

+

+

+

+

⋅⋅

⋅⋅

=

⋅ ⋅ ⋅ ⋅ ⋅⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅⋅ ⋅ ⋅ ⋅ ⋅ ⋅⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅

⋅ ⋅ ⋅ ⋅ ⋅

·

⋅⋅

⋅⋅

, ,

,

, , D00

A Parentrange

A

Parent spectrum

Parent spectrum Assumption:Groups1 Kr=

=

=

ypton+

2 Krypton++

2 1

Library spectrum:Krypton

Parent spectrum without krypton Assumption:3 Argon+

4 Argon++

Library spectrum:Argon

Parent spectrum without argon Assumption:5 Neon+

Library spectrum:Neon

Parent spectrum after detectionof krypton, argon and neon

34

5

Fig. 4.16 Subtracting spectra contained in libraries

D00 E 19.06.2001 21:39 Uhr Seite 101


LEYBOLD VACUUM PRODUCTS AND REFERENCE BOOK 2001/2002

equal to the fragment matrix times the vec-tor of the sum of the flows for the indivi-dual gases.

or:

(in simplified notation: i = FF · I)

where im+ = ion flow vector for the atomic

numbers, resulting from contributions offragments of various individual gases

= fragment matrix

Ig+ = Vector of the sum of the flows for the

individual gases or:

One sees that the ion flow caused by a gasis proportional to the partial pressure. Thelinear equation system can be solved onlyfor the special instance where m = g(square matrix); it is over-identified for m > g. Due to unavoidable measurementerror (noise, etc.) there is no set of overallion flow I+g (partial pressures or concentra-tions) which satisfies the equation systemexactly. Among all the conceivable soluti-ons it is now necessary to identify set I+*gwhich after inverse calculation to the parti-al ion flows I+*m will exhibit the smallestsquared deviation from the partial ion cur-rents im

+ actually measured. Thus:

This minimization problem is mathematic-ally identical to the solution of anotherequation system

FFT · i = FFT · FF · I

which can be evaluated direct by the com-puter. The ion current vector for the indivi-dual gases is then:

[ ] [ ][ ]

IFF i FF BF

FF BF

T T

T=⋅ ⋅ ⋅

⋅

–1

det

( )i im m− =∑ *+ + min2

Ff m, g6444474448

i p E RIP FF TFm g N g m m+ = ⋅ ⋅ ⋅ ⋅∑

2

Transmission factorfor the mass m

Fragment factorfor the gas to mass m

Relative ionization probabilityfor the gas

Nitrogen sensitivity (equipment constant)Partial pressure of the gas

Ion current for atomic number m

BFm gg k

,=∑0

i BF Im m g gg k

+ +

== ·∑ ,

0

4.7 Software

4.7.1 Standard SQX software(DOS) for stand-aloneoperation (1 MS plus 1 PC, RS 232)

The conventional software package (SQX)contains the standard routines for theoperation of the mass spectrometer(MS)– various spectra depictions, queriesof individual channels with the correspon-ding screen displays as tables or barcharts, partial pressure conversion, trenddisplays, comparison with spectra libra-ries (with the capability for trial subtrac-tion of library spectra), leak testing modeetc. – and for sensor balancing, as well.Using PCs as the computer and displayunit naturally makes available all the usualfunctions including storing and retrievingdata, printing, etc. Characteristic of theconventional software package is that spe-cific individual spectra will be measured,even though the measurement is fullyautomated and takes place at a point intime which is specified in advance. Aspectrum of this type can thus be only a“snapshot” of a process in progress.

4.7.2 Multiplex/DOS softwareMQX (1 to 8 MS plus 1 PC, RS 485)

The first step toward process-orientedsoftware by INFICON is the MQX. It makespossible simultaneous monitoring of amaximum of eight sensors and you canapply all the SQX functions at each sensor.

4.7.3 Process-oriented soft-ware –Transpector-Warefor Windows

Transpector-Ware is based on an entirelynew philosophy. During the course of theprocess (and using settings – the “recipe”– determined beforehand) data will berecorded continuously – like the individualframes in a video. These data can be sto-red or otherwise evaluated. It is possible

in particular to analyze interesting processsections exactly, both during the processand retroactively, once the process has runto completion, without having to interruptthe measurement operations which arerunning in the background. Whereongoing monitoring of identical processesis undertaken the program can generatestatistics (calculating mean values andstandard deviations) from which a band-width for “favorable process operation”can be derived. Error reports are issuedwhere limit values are exceeded.

4.7.4 Development software –TranspectorView

This software used to for develop customsoftware versions for special situations. Itis based on the LabView developmentpackage and includes the drivers requiredto operate the Transpector.

4.8 Partial pressureregulation

Some processes, such as reactive sputterprocesses, require the most constant pos-sible incidence rates for the reacting gasmolecules on the substrate being coated.

The “incidence rate” is the same as the“impingement rate” discussed in Chapter1; it is directly proportional to the partialpressure. The simplest attempt to keep thepartial pressure for a gas component con-stant is throughput by regulating with aflow controller; it does have the disadvan-tage that the regulator cannot determinewhether, when and where the gas con-sumption or the composition of the gas inthe vacuum chamber changes. The farsuperior and more effective option is parti-al pressure control using a mass spectro-meter via gas inlet valves. Here the signifi-cant peaks of the gases being consideredare assigned to channels in the mass spec-trometer. Suitable regulators compare theanalog output signals for these channelswith set-point values and derive from thedifference between the target and actualvalues for each channel the appropriateactuation signal for the gas inlet valve for

D00.102

D00 E 19.06.2001 21:39 Uhr Seite 102



the channel. A configuration of this kindhas been realized to control six channels inthe QUADREX PPC. Gas inlet valves mat-ching the unit can also be delivered.

The gas used to measure the impingementrate (partial pressure) must naturally bedrawn from a representative point in thevacuum chamber. When evaluating thetime constant for a regulation circuit of thistype it is important to take into account allthe time aspects and not just the electricalsignal propagation and the processing inthe mass spectrometer, but also the vacu-um-technology time constants and flowvelocities, as illustrated in Figure 4.17.Pressure converters or unfavorably instal-led gas inlet lines joining the control valveand the vacuum vessel will make particu-larly large contributions to the overall timeconstant. It is generally better to establisha favorable S/N ratio with a large signal (i.e.through an inlet diaphragm with a largeopening) rather than with long integrationperiods at the individual channels. Contra-sted in Figure 4.18 are the effects of boo-sting pressure and lengthening the integra-tion time on signal detectability. In depic-tions a, b and c only the integration periodwas raised, from 0.1 to 1.0 and 10 seconds(thus by an overall factor of 100), respec-tively. By comparison, in the sequence a-d-e-f, at constant integration time, the totalpressure was raised in three steps, from7.2 · 10-6 mbar to 7.2 · 10-5 mbar (or by afactor of just 10 overall).

4.9 Maintenance(Cathode service life, sensor balancing,cleaning the ion source and rod system)

The service life of the cathode will dependgreatly on the nature of the loading. Expe-rience has shown that the product of ope-rating period multiplied by the operatingpressure can serve as a measure for theloading. Higher operating pressures (in arange of 1 · 10-4 to 1 · 10-3 mbar) have aparticularly deleterious effect on servicelife, as do certain chemical influences suchas refrigerants, for example. Changing outthe cathode is quite easy, thanks to thesimple design of the sensor. It is advisab-le, however, to take this opportunity tochange out or at least clean the entire ionsource.

Sensor balancing at the mass axis (oftenerroneously referred to as calibration) isdone today in a very easy fashion with thesoftware (e.g. SQX, Transpector-Ware)and can be observed directly in the screen.Naturally, not only the arrangement alongthe mass axis will be determined here, butalso the shape of the lines, i.e. resolutionand sensitivity (see Section 4.5).

It will be necessary to clean the sensoronly in exceptional cases where it is heavi-ly soiled. It is usually entirely sufficient toclean the ion source, which can be easilydismantled and cleaned. The rod systemcan be cleaned in an ultrasonic bath onceit has been removed from the config-uration. If dismantling the system is una-voidable due to particularly stubborn

grime, then the adjustment of the rodswhich will be required afterwards will haveto be carried out at the factory.

D00

TMP50CF

Sensor

Pressure stage

Vacuum vessel

Mass spectrometer

Regulation valvet3t2

t1

t6

t5

t4

efee

ef_

Fig. 4.17 Partial shares for overall time constants

Fig. 4.18 Improving the signal-to-noise ratio by increasing the pressure or extending the integration time

a b c

d fe

D00 E 19.06.2001 21:39 Uhr Seite 103

Leak DetectionFundamentals of Vacuum Technology


5. Leaks and theirdetection

Apart from the vacuum systems them-selves and the individual components usedin their construction (vacuum chambers,piping, valves, detachable [flange] connec-tions, measurement instruments, etc.),there are large numbers of other systemsand products found in industry and rese-arch which must meet stringent require-ments in regard to leaks or creating a so-called “hermetic” seal. Among these aremany assemblies and processes in theautomotive and refrigeration industries inparticular but also in many other branchesof industry. Working pressure in this caseis often above ambient pressure. Here“hermetically sealed” is defined only as arelative “absence of leaks”. Generalizedstatements often made, such as “no detec-table leaks” or “leak rate zero”, do notrepresent an adequate basis for accep-tance testing. Every experienced engineerknows that properly formulated accep-tance specifications will indicate a certainleak rate (see Section 5.2) under definedconditions. Which leak rate is acceptable isalso determined by the application itself.

5.1 Types of leaksDifferentiation is made among the follo-wing leaks, depending on the nature of thematerial or joining fault:

• Leaks in detachable connections: Flan-ges, ground mating surfaces, covers

• Leaks in permanent connections:Solderand welding seams, glued joints

• Leaks due to porosity: particularly follo-wing mechanical deformation (ben-ding!) or thermal processing of poly-crystalline materials and cast compo-nents

• Thermal leaks (reversible): opening upat extreme temperature loading (heat/cold), above all at solder joints

• Apparent (virtual) leaks: quantities ofgas will be liberated from hollows andcavities inside cast parts, blind holesand joints (also due to the evaporationof liquids)

• Indirect leaks: leaking supply lines invacuum systems or furnaces (water,compressed air, brine)

• “Serial leaks”: this is the leak at the endof several “spaces connected in series”,e.g. a leak in the oil-filled section of theoil pan in a rotary vane pump

• “One-way leaks”: these will allow gas topass in one direction but are tight in theother direction (very seldom)

An area which is not gas-tight but which isnot leaky in the sense that a defect is pre-sent would be the

• Permeation (naturally permeability) ofgas through materials such as rubberhoses, elastomer seals, etc. (unlessthese parts have become brittle andthus “leaky”).

5.2 Leak rate, leaksize, mass flow

No vacuum device or system can ever beabsolutely vacuum-tight and it does notactually need to be. The simple essential isthat the leak rate be low enough that therequired operating pressure, gas balanceand ultimate pressure in the vacuum con-tainer are not influenced. It follows that therequirements in regard to the gas-tightn-ess of an apparatus are the more stringentthe lower the required pressure level is. Inorder to be able to register leaks quantita-tively, the concept of the “leak rate” withthe symbol QL was introduced; it is mea-sured with mbar·l·s-1 or cm3/s (STP) asthe unit of measure. A leak rate of QL = 1 mbar·l·s-1 is present when in anenclosed, evacuated vessel with a volumeof 1 l the pressure rises by 1 mbar persecond or, where there is positive pressu-re in the container, pressure drops by 1mbar. The leak rate QL defined as a mea-sure of leakiness is normally specified inthe unit of measure mbar·l·s-1. With theassistance of the status equation (1.7) onecan calculate QL when giving the tempera-ture T and the type of gas M, registeringthis quantitatively as mass flow, e.g. in theg/s unit of measure. The appropriaterelationship is then:

(5.1)( )Q p V

tR TM

mtL = ⋅ = ⋅ ⋅

∆∆

∆∆

where R = 83.14 mbar · l/mol · K, T = tem-perature in K; M = molar mass in g/mole;∆m for the mass in g; ∆t is the time periodin seconds. Equation 5.1 is then used a) to determine the mass flow ∆m / ∆t at a

known pV gas flow of ∆p · V/∆t (see inthis context the example at 5.4.1) or

b) to determine the pV leak gas flow wherethe mass flow is known (see the follo-wing example).

Example for case b) above:

A refrigeration system using Freon (R 12)exhibits refrigerant loss of 1 g of Freon peryear (at 25 °C). How large is the leak gasflow QL? According to equation 5.1 forM(R12) = 121 g/mole:

Thus the Freon loss comes to QL = 6.5 · 10–6 mbar·l·s-1. According to the“rule of thumb” for high vacuum systemsgiven below, the refrigeration system men-tioned in this example may be deemed tobe very tight. Additional conversions forQL are shown in Tables VIIa and VIIb inChapter 9.

The following rule of thumb for quantita-tive characterization of high vacuumequipment may be applied:

Total leak rate < 10-6 mbar·l·s-1:Equipment is very tight

Total leak rate 10-5 mbar·l·s-1:Equipment is sufficiently tight

Total leak rate > 10-4 mbar·l·s-1:Equipment is leaky

A leak can in fact be “overcome” by apump of sufficient capacity because it istrue that (for example at ultimate pressurepend and disregarding the gas liberatedfrom the interior surfaces):

(5.2)

(QL Leak rate, Seff the effective pumpingspeed at the pressure vessel)

pend

QL

Seff

=

( )Q p Vt

mbar K gmol K g mol year

mbars

L =⋅

= ⋅ ⋅ ⋅⋅ ⋅ ⋅ ⋅

= ⋅ ⋅ ⋅⋅

⋅ ⋅⋅

–∆

∆8314 298 1

121 1

8314 2 10 1121 1 315 10

1

2

7

.

. .98.

`

`

= ⋅ ⋅

⋅ ⋅⋅ ⋅ ⋅

= ⋅ ⋅ ⋅–

8314 2 101 10 315

10

65 10

2

2–7

7

. .98.21 .

mbars

mbars

`

`

D00 E 19.06.2001 21:39 Uhr Seite 104

Concept / criterion Comment QL [mbar · l/s] Relevant particle size

Water-tight *) Droplets QL < 10–2

Vapor-tight “Sweating” QL < 10–3

Bacteria-tight *)

(cocci) QL < 10–4 Avg. ≈ 1 µm

(rod-shaped) Avg. ≈ 0.5-1 µm, 2–10 µm longOil-tight QL < 10–5

Virus-tight *)

(vaccines such as pox) QL < 10–6 Ø≈ 3 · 10–7 m(smallest viruses,bacteriophages) QL < 10–8 Ø 3 · 10–8 m

(viroids, RNA) QL < 10–10 Ø ª≈ 1 · 10–9 m (thread-like)

Gas-tight QL < 10–7

“Absolutely tight” Technical QL < 10–10

*) As opposed to vapor, it is necessary to differentiate between hydrophilic and hydrophobic solids. This also applies to bacteria and viruses since they are transported primarily in solutions.

Leak Detection Fundamentals of Vacuum Technology


Where Seff is sufficiently great it is possi-ble – regardless of the value for the leakrate QL – always to achieve a pre-determi-ned ultimate pressure of pend. In practice,however, an infinite increase of Seff will runup against economic and engineering limi-tations (such as the space required by thesystem).

Whenever it is not possible to achieve thedesired ultimate pressure in an apparatusthere are usually two causes which can becited: The presence of leaks and/or gasbeing liberated from the container wallsand sealants.

Partial pressure analysis using a massspectrometer or the pressure rise methodmay be used to differentiate between thesetwo causes. Since the pressure risemethod will only prove the presence of aleak without indicating its location in theapparatus, it is advisable to use a heliumleak detector with which leaks can, ingeneral, also be located much more quick-ly.

In order to achieve an overview of the cor-relation between the geometric size of thehole and the associated leak rate it is pos-sible to operate on the basis of the follo-wing, rough estimate: A circular hole 1 cmin diameter in the wall of a vacuum vesselis closed with a gate valve. Atmosphericpressure prevails outside, a vacuum insi-de. When the valve is suddenly opened allthe air molecules in a cylinder 1 cm in dia-meter and 330 m high would within a 1-second period of time “fall into” the holeat the speed of sound (330 m/s). Thequantity flowing into the vessel eachsecond will be 1013 mbar times the cylin-der volume (see Fig. 5.1). The result is thatfor a hole 1 cm in diameter QL (air) will be2.6 · 104 mbar·l·s-1. If all other conditionsare kept identical and helium is allowed toflow into the hole at its speed of sound of970 m/s, then in analogous fashion the QL(helium) will come to 7.7 · 10+4 mbar·l·s-1,or a pV leaking gas current which is largerby a factor of 970 / 330 = 2.94. This grea-ter “sensitivity” for helium is used in leakdetection practice and has resulted in thedevelopment and mass production of hig-hly sensitive helium-based leak detectors(see Section 5.5.2).

Shown in Figure 5.1 is the correlation bet-ween the leak rate and hole size for air, withthe approximation value of QL (air) of 10+4 mbar·l·s-1 for the “1 cm hole”. The

table shows that when the hole diameter isreduced to 1 µm (= 0.001 mm) the leak ratewill come to 10-4 mbar·l·s-1, a value whichin vacuum technology already represents amajor leak (see the rule of thumb above). Aleak rate of 10-12 mbar·l·s-1 corresponds tohole diameter of 1 Å; this is the lowerdetection limit for modern helium leakdetectors. Since the grid constants formany solids amount to several Å and thediameter of smaller molecules and atoms(H2, He) are about 1 Å, inherent permeati-on by solids can be registered metrologi-cally using helium leak detectors. This has

led to the development of calibrated refe-rence leaks with very small leak rates (seeSection 5.5.2.3). This is a measurable “lackof tightness” but not a “leak” in the senseof being a defect in the material or joint.Estimates or measurements of the sizes ofatoms, molecules, viruses, bacteria, etc.have often given rise to everyday termssuch as “watertight” or “bacteria-tight”;see Table 5.1.

Compiled in Figure 5.2 are the nature anddetection limits of frequently used leakdetection methods.

D00

∆p = 1013 mbar, Hole diameter d = 1 cm

Gas speed = Speed of sound = 330 ms

mbar ·s

`

mbar ·s

``s

`s

cms

3ms

1 ·4

2 πVolume/second:

Quantity/second:

Diameter cm Leak rate

1013 mbar · 25.95 = 2.63 · 10 10+4 +4P

330 · · cm = 25.95 · 10 = 25.952 +3

10 m= 1.0 cm 1010 m= 1.0 mm 1010 m= 0.1 mm 10 (= 1)10 m= 0.01 mm 1010 m= 1.0 m 1010 m= 0.1 m 1010 m= 0.01 m 1010 m= 1.0 nm 1010 m= 1.0 Angstrom 10 (Detection limit, He leak detector)

–2 +4

–3 +2

–4 0

–5 –2

–6 –4

–7 –6

–8 –8

–9 –10

–10 –12

µµµ

Fig. 5.1 Correlation between leak rate and hole size

Table 5.1 Estimating borderline leak rates

D00 E 19.06.2001 21:39 Uhr Seite 105

Range Laminar Molecular

Pressure

Gas



5.2.1 The standard helium leak rate

Required for unequivocal definition of aleak are, first, specifications for the pressures prevailing on either side of thepartition and, secondly, the nature of themedium passing through that partition(viscosity) or its molar mass. The designa-tion “helium standard leak” (He Std) hasbecome customary to designate a situati-on frequently found in practice, wheretesting is carried out using helium at 1 bardifferential between (external) atmosphe-ric pressure and the vacuum inside asystem (internal, p < 1 mbar), the designa-tion “helium standard leak rate” has beco-me customary. In order to indicate therejection rate for a test using helium understandard helium conditions it is necessaryfirst to convert the real conditions of use tohelium standard conditions (see Section5.2.2). Some examples of such conversi-ons are shown in Figure 5.3.

5.2.2 Conversion equationsWhen calculating pressure relationshipsand types of gas (viscosity) it is necessaryto keep in mind that different equations areapplicable to laminar and molecular flow;the boundary between these areas is verydifficult to ascertain. As a guideline onemay assume that laminar flow is present atleak rates where QL > 10-5 mbar · l/s andmolecular flow at leak rates where QL < 10-7 mbar · l/s. In the intermediate

range the manufacturer (who is liableunder the guarantee terms) must assumevalues on the safe side. The equations arelisted in Table 5.2.Here indices “I” and “II” refer to the one orthe other pressure ratio and indices “1”and “2” reference the inside and outside ofthe leak point, respectively.

5.3 Terms and definitions

When searching for leaks one will general-ly have to distinguish between two tasks:

1. Locating leaks and2. Measuring the leak rate.

In addition, we distinguish, based on thedirection of flow for the fluid, between the

a. vacuum method (sometimes known asan “outside-in leak”), where the direc-tion of flow is into the test specimen(pressure inside the specimen beingless than ambient pressure), and the

b. positive pressure method (often refer-

red to as the “inside-out leak”), wherethe fluid passes from inside the testspecimen outward (pressure inside thespecimen being greater than ambientpressure).

The specimens should wherever possiblebe examined in a configuration corre-sponding to their later application – com-ponents for vacuum applications using thevacuum method and using the positivepressure method for parts which will bepressurized on the inside.

When measuring leak rates we differentia-te between registering

a. individual leaks (local measurement) –sketches b and d in Figure 5.4, and regi-stering

b. the total of all leaks in the test specimen(integral measurement) – sketches aand c in Figure 5.4.

The leak rate which is no longer tolerablein accordance with the acceptance specifi-cations is known as the rejection rate. Itscalculation is based on the condition thatthe test specimen may not fail during itsplanned utilization period due to faultscaused by leaks, and this to a certain

Vac

uum

met

hod

Ove

rpre

ssur

e m

etho

d

Helium leak detector ULTRATEST UL 200 dry/UL 500

ULTRATEST with helium sniffer

Halogen sniffer HLD4000A

Bubble test

Pressure drop test

103................100 10-1 10-2 10-3 10-4 10-5 10-6 10-7 10-8 10-9 10-10 10-11 10-12 mbar · l · s-1

Ecotec II / Protec

Contura Z

Pressure rise

Helium leak detector ULTRATEST UL 200/UL 500 dry/Modul 200/LDS 1000

Fig. 5.2 Leak rate ranges for various leak detection processes and devices

Leak <----> HoleQ ... Leak rate,In short: Leak

Substance quantity hole unit of timetrhough per Helium standard leak rate:p = 1 bar, p < 1 mbar ( p = 1 bar)Test gas = Helium

1 2 ∆

Familiar leaks: Quantity escaping: Standard He leak rate:

Dripping water faucet 34

10

3.18 · 10

4.19 · 10

430 Frigen

–2

–4

–3 –3

–2

–5

–2

–5–3

3

Water Air He Std

Air

He Std

He Std

He Std

He StdF12

Air

Air

= 6.45 0.17

= 4.24 · 10

0.9 · 10

4.3 · 10

1.88 · 10

4.33 · 10= 2.8 · 10

mg

Ncm

g

mbar · ø mbar · ø

mbar · ø

mbar · ø

mbar · ø

mbar · ø

mbar · ømbar · ø

mbar · ø

mbar · ø

s

s

a

s s

s

s

s

s

ss

s

s

4 mm diam., 1 Hz, p = 4 bar∆

Hair on a gasket

Bicycle tube in water(bubble test)2 mm diam., 1 Hz, p = 0.1 bar∆

Car tire loses air25 l, 6 Mo: 1.8 --> 1.6 bar

Small refrigerant cylinderempties in 1 year430 g refrigerant R12, 25°C

Definition: Q = ∆ (p · V)∆t

1

1

1

2

2

2

Fig. 5.3 Examples for conversion into helium standard leak rates

Table 5.2 Conversion formulae for changes of pressure and gas type

( ) ( )Q p p Q p pI II II I⋅ − = ⋅ −12

22

12

22 ( ) ( )Q p p Q p pI II II I⋅ − = ⋅ −1 2 1 2

Q M Q Mgas A gas A gas B gas B⋅ = ⋅Qgas A á ηgas A = Qgas B á ηgas B

D00 E 19.06.2001 21:39 Uhr Seite 106



degree of certainty. Often it is not the leakrate for the test specimen under normaloperating conditions which is determined,but rather the throughput rate of a test gas– primarily helium – under test conditions.The values thus found will have to be con-verted to correspond to the actual applica-tion situation in regard to the pressuresinside and outside the test specimen andthe type of gas (or liquid) being handled.

Where a vacuum is present inside the testspecimen (p < 1 mbar), atmospheric pres-sure outside, and helium is used at the testgas, one refers to standard helium condi-tions. Standard helium conditions arealways present during helium leak detec-tion for a high vacuum system when thesystem is connected to a leak detector andis sprayed with helium (spray technique).If the specimen is evacuated solely by theleak detector, then one would say that theleak detector is operating in the direct-flow mode. If the specimen is itself a com-plete vacuum system with its own vacuumpump and if the leak detector is operated

in parallel to the system’s pumps, then onerefers to partial-flow mode. One alsorefers to partial stream mode when a sepa-rate auxiliary pump is used parallel to theleak detector.

When using the positive pressure methodit is sometimes either impractical or in factimpossible to measure the leakage ratedirectly while it could certainly be sensedin an envelope which encloses the testspecimen. The measurement can be madeby connecting that envelope to the leakdetector or by accumulation (increasingthe concentration) of the test gas insidethe envelope. The “bombing test” is aspecial version of the accumulation test(see Section 5.7.4). In the so-called sniffertechnique, another variation of the of thepositive pressure technique, the (test) gasissuing from leaks is collected (extracted)by a special apparatus and fed to the leakdetector. This procedure can be carried outusing either helium or refrigerants or SF6as the test gas.

5.4 Leak detectionmethods without aleak detector unit

The most sensible differentiation betweenthe test methods used is differentiation asto whether or not special leak detectionequipment is used.

In the simplest case a leak can be determi-ned qualitatively and, when using certaintest techniques, quantitatively as well (thisbeing the leak rate) without the assistanceof a special leak detector. Thus the quanti-ty of water dripping from a leaking waterfaucet can be determined, through a cer-tain period of time, using a graduatedcylinder but one would hardly refer to thisas a leak detector unit. In those caseswhere the leak rate can be determinedduring the course of the search for the leakwithout using a leak detector (see, forexample, Sect. 5.4.1), this will often beconverted to the helium standard leak rate(Sect. 5.2.1). This standard leak rate valueis frequently needed when issuing accept-ance certificates but can also be of servicewhen comparing leak rate values deter-mined by helium leak detector devices.

In spite of careful inspection of the indivi-dual engineering components, leaks mayalso be present in an apparatus followingits assembly – be it due to poorly seatedseals or damaged sealing surfaces. Theprocesses used to examine an apparatuswill depend on the size of the leaks and onthe degree of tightness being targeted andalso on whether the apparatus is made ofmetal or glass or other materials. Someleak detection techniques are sketched outbelow. They will be selected for use inaccordance with the particular applicationsituations; economic factors may play animportant part here.

5.4.1 Pressure rise test

This leak testing method capitalizes on thefact that a leak will allow a quantity of gas– remaining uniform through a period oftime – to enter a sufficiently evacuateddevice (impeded gas flow, see Fig. 1.1). Incontrast, the quantity of gas liberated fromcontainer walls and from the materialsused for sealing (if these are not suffi-

D00

Fig. 5.4 Leak test techniques and terminology

Helium

Helium

Helium

c: Integral leak detection (test gas enrichmentinside the enclosure); pressurized test gas insidespecimen

d: Local leak detection; pressurized test gas inside the specimen

a: Integral leak detection; vacuum insidespecimen

b: Local leak detection; vacuum inside specimen

D00 E 19.06.2001 21:39 Uhr Seite 107



ciently free of outgassing) will declinethrough time since these will practicallyalways be condensable vapors for whichan equilibrium pressure is reached atsome time (see Fig. 5.5). The valve at thepump end of the evacuated vacuum vesselwill be closed in preparation for pressurerise measurements. Then the time is mea-sured during which the pressure rises by acertain amount (by one power of ten, forexample). The valve is opened again andthe pump is run again for some time, fol-lowing which the process will be repeated.If the time noted for this same amount ofpressure rise remains constant, then a leakis present, assuming that the waiting peri-od between the two pressure rise trialswas long enough. The length of an appro-priate waiting period will depend on thenature and size of the device. If the pres-sure rise is more moderate during thesecond phase, then the rise may be assu-med to result from gases liberated fromthe inner surfaces of the vessel. One mayalso attempt to differentiate between leaksand contamination by interpreting thecurve depicting the rise in pressure. Plot-ted on a graph with linear scales, the curvefor the rise in pressure must be a straightline where a leak is present, even at higherpressures. If the pressure rise is due togas being liberated from the walls (owingultimately to contamination), then thepressure rise will gradually taper off andwill approach a final and stable value. Inmost cases both phenomena will occursimultaneously so that separating the twocauses is often difficult if not impossible.These relationships are shown schematic-

ally in Figure 5.5. Once it has become clearthat the rise in pressure is due solely to areal leak, then the leak rate can be deter-mined quantitatively from the pressurerise, plotted against time, in accordancewith the following equation:

(5.3)

Example: Once the vacuum vessel with avolume of 20 l has been isolated from thepump, the pressure in the apparatus risesfrom 1·10-4 mbar to 1·10-3 mbar in 300 s.Thus, in accordance with equation 5.2, theleak rate will be

The leak rate, expressed as mass flow ∆m / ∆t, is derived from equation 5.1 at QL = 6 · 10-5 mbar · l/s, T = 20 °C and themolar mass for air (M = 29 g/mole) at

If the container is evacuated with a TURBOVAC 50 turbomolecular pump, forexample (S = 50 l/s), which is attached tothe vacuum vessel by way of a shut-offvalve, then one may expect an effectivepumping speed of about Seff = 30 l/s. Thusthe ultimate pressure will be

Naturally it is possible to improve this ulti-mate pressure, should it be insufficient, byusing a larger-capacity pump (e.g. theTURBOVAC 151) and at the same time toreduce the pump-down time required toreach ultimate pressure.

Today leak tests for vacuum systems areusually carried out with helium leak detec-tors and the vacuum method (see Section5.7.1). The apparatus is evacuated and atest gas is sprayed around the outside. Inthis case it must be possible to detect (onthe basis of samplings inside the appara-

pend

QL

Seff

mbar ss

mbar

= =– ⋅⋅ ⋅ –

⋅ –

= ⋅ –

6 10 5 1

30 1

2 10 6

`

`

Q mt

mbars

gmol

mol Kmbar K

gs

L = = ⋅ ⋅ ⋅ ⋅ ⋅

⋅ ⋅⋅ ⋅ ⋅

= ⋅

–

–

∆∆

6 10 29

8314 293 107 10

5

28

`

`.

QL

mbars

=⋅ – − ⋅ – ⋅

= ⋅ – ⋅ = – ⋅

1 10 3 1 10 4 20

300

9 10 4 20300

6 10 5⋅ `

QLp V

t= ⋅∆

∆

tus) the test gas which has passed throughleaks and into the apparatus. Another opti-on is to use the positive-pressure leak test.A test gas (helium) is used to fill the appa-ratus being inspected and to build up aslight positive pressure; the test gas willpass to the outside through the leaks andwill be detected outside the device. Theleaks are located with leak sprays (or soapsuds, 5.4.5) or – when using He or H2 asthe test gas – with a leak detector and snif-fer unit (5.7.2).

5.4.2 Pressure drop test

The thinking here is analogous to that forthe pressure rise method (Section 5.4.1).The method is, however, used only rarelyto check for leaks in vacuum systems. Ifthis is nonetheless done, then gauge pres-sure should not exceed 1 bar since theflange connectors used in vacuum techno-logy will as a rule not tolerate higher pres-sures. Positive pressure testing is, on theother hand, a technique commonlyemployed in tank engineering. When dea-ling with large containers and the long testperiods they require for the pressure dropthere it may under certain circumstancesbe necessary to consider the effects oftemperature changes. As a consequence itmay happen, for example, that the systemcools to below the saturation pressure forwater vapor, causing water to condense;this will have to be taken into accountwhen assessing the pressure decline.

5.4.3 Leak test using vacuum gauges which are sensitive to the typeof gas

The fact that the pressure reading at vacu-um gauges (see Section 3.3) is sensitive tothe type of gas involved can, to a certainextent, be utilized for leak detection purpo-ses. Thus it is possible to brush or spraysuspected leaks with alcohol. The alcoholvapors which flow into the device – thethermal conductivity and ionizablity ofwhich will vary greatly from the same pro-perties for air – will affect and changepressure indication to a greater or lesser

Fig. 5.5 Pressure rise within a vessel after the pump isswitched off

1 Leak2 Gas evolved from the container walls3 Leak + gas evolution

Time

Pres

sure

D00 E 19.06.2001 21:39 Uhr Seite 108



extent. The availability of more precise,easy-to-use helium leak detectors has,however, rendered this method almostcompletely obsolete.

5.4.4 Bubble immersion testThe pressurized test specimen is sub-merged in a liquid bath. Rising gas bub-bles indicate the leak. Leak detection willdepend greatly on the attentiveness of theperson conducting the test and involvesthe temptation to increase the “sensitivity”by using ever higher temperatures, wher-ein the applicable safety regulations aresometimes disregarded. This method isvery time-consuming for smaller leaks, asTable 5.3 shows. It references leak testingon a refrigeration system using type R12refrigerant. Here the leak rate is specifiedin grams of refrigerant loss per year (g/a).Water is used as a test liquid (which maybe heated or to which a surfactant may beadded) or petroleum-based oils. The surfa-ce tension should not exceed 75 dyn/cm (1 dyn = 10-5 N).

5.4.5 Foam-spray test

In many cases pressurized containers orgas lines (including the gas supply linesfor vacuum systems) can be checked quiteconveniently for leaks by brushing orspraying a surfactant solution on them.Corresponding leak detection sprays arealso available commercially. Escaping gasforms “soap bubbles” at the leak points.

Here, again, the detection of smaller leaksis time-consuming and will depend greatlyon the attentiveness of the inspector. Thehydrogen gas cooling systems used inpower plant generators represent a specialcase. These are indeed sometimes testedin the fashion described above but theycan be examined much better and at muchhigher sensitivity by “sniffing” the hydro-gen escaping at leaks using a helium leakdetector which has been adjusted to res-pond to H2 (see Section 5.7.2).

5.4.6 Vacuum box check bubble

As a variation on the spray technique men-tioned above, in which the escaping gascauses the bubbles, it is possible to placea so-called “vacuum box” with a seal(something like a diver’s goggles) on thesurface being examined once it has beensprayed with a soap solution. This box isthen evacuated with a vacuum pump. Airentering from the outside through leakswill cause bubbles inside the box, whichcan be observed through a glass windowin the box. In this way it is also possible,for example, to examine flat sheet metalplates for leaks. Vacuum boxes are availa-ble for a variety of applications, made tosuit a wide range of surface contours.

5.4.7 Krypton 85 test

When dealing with small, hermetically sea-led parts where the enclosure is leaky,

krypton 85, a gaseous, radioactive isotope,can first be forced into the device by app-lying pressure from the outside. Once anexactly measured holding period has elap-sed the pressure will be relieved, the com-ponent flushed and the activity of the “gascharge” will be measured. In the same wayit is also possible to use helium as the testgas (see Section 5.7.4, bombing test).

5.4.8 High-frequency vacuumtest

The so-called high-frequency vacuumtester can be used not only to check thepressure in glass equipment but also tolocate porous areas in plastic or paint coa-tings on metals. This comprises a hand-held unit with a brush-like high-frequencyelectrode and a power pack. The shapeand color of the electrical gas dischargecan serve as a rough indicator for the pres-sure prevailing inside glass equipment. Inthe case of the vacuum tester – whichcomprises primarily a tesla transformer(which delivers a high-voltage, high-fre-quency AC current) – the corona electrodeapproaching the apparatus will trigger anelectrode-free discharge inside the appara-tus. The intensity and color of this dischar-ge will depend on the pressure and thetype of gas. The luminous discharge phe-nomenon allows us to draw conclusionsregarding the approximate value for thepressure prevailing inside the apparatus.The discharge luminosity will disappear athigh and low pressures.

When searching for leaks in glass equip-ment the suspect sections will be scannedor traced with the high-frequency vacuumtester electrode. Where there is a leak anarc will strike through to the pore in theglass wall, tracing a brightly lit dischargetrail. Small pores can be enlarged by thesesparks! The corona discharge of the vacu-um tester can also penetrate thin areas inthe glass particularly at weld points andtransitional areas between intermediatecomponents. Equipment which was ori-ginally leak-free can become leaky in thisfashion! In contrast to the actual leakdetector units, the high-frequency vacuumtester is highly limited in its functioning. D00

Freon F12 loss Time taken to form Equivalent Detection time using per year a gas bubble leak rate helium leak detector

(g/a) (s) (cm3[STP]/s) (s)

280 13.3 1.8 · 10–3 a few seconds

84 40 5.4 · 10–4 a few seconds

28 145 1.8 · 10–4 a few seconds

14 290 9.0· 10–5 a few seconds

2.8 24 min 1.8 · 10–5 a few seconds

0.28 * 6 h 1.8 · 10–6 a few seconds*) This leak rate represents the detection limit for good halogen leak detectors (≈ 0,1 g/a).

Table 5.3 Comparison of bubble test method (immersion technique) wit helium leak detector

D00 E 19.06.2001 21:39 Uhr Seite 109



5.4.9 Test with chemical reactions and dye penetration

Occasionally leaks can also be located ordetected by means of chemical reactionswhich result in a discoloration or by pene-tration of a dye solution into fine openings.The discoloration of a flame due to halo-gen gas escaping through leaks was usedearlier to locate leaks in solder joints forrefrigeration units.

A less frequently employed example of achemical effect would be that of escapingammonia when it makes contact with oza-lid paper (blueprint paper) or with othermaterials suitably prepared and wrappedaround the outside of the specimen. Leaksare then detected based on the discolorati-on of the paper.

An example of a dye penetration test is theinspection of the tightness of rubber plugsor plungers in glass tubes, used for exam-ple in testing materials suitability for dis-posable syringes or pharmaceutical packa-ges. When evaluating tiny leaks for liquidsit will be necessary to consider the wetabi-

lity of the surface of the solid and the capil-lary action; see also Table 5.1. Some wide-ly used leak detection methods are shown –together with the test gas, application rangeand their particular features – in Table 5.4.

5.5 Leak detectors andhow they work

Most leak testing today is carried out usingspecial leak detection devices. These candetect far smaller leak rates than techni-ques which do not use special equipment.These methods are all based on using spe-cific gases for testing purposes. The diffe-rences in the physical properties of thesetest gases and the gases used in real-lifeapplications or those surrounding the testconfiguration will be measured by the leakdetectors. This could, for example, be thediffering thermal conductivity of the testgas and surrounding air. The most widelyused method today, however, is the detec-tion of helium used as the test gas.

The function of most leak detectors isbased on the fact that testing is conductedwith a special test gas, i.e. with a mediumother than the one used in normal operati-on. The leak test may, for example, be car-ried out using helium, which is detectedusing a mass spectrometer, even thoughthe component being tested might, forexample, be a cardiac pacemaker whoseinterior components are to be protectedagainst the ingress of bodily fluids duringnormal operation. This example alonemakes it clear that the varying flow pro-perties of the test and the working medianeed to be taken into consideration.

5.5.1 Halogen leak detectors (HLD4000, D-Tek)

Gaseous chemical compounds whosemolecules contain chlorine and/or fluorine– such as refrigerants R12, R22 andR134a – will influence the emissions ofalkali ions from a surface impregnatedwith a mixture of KOH and Iron(III)hydro-xide and maintained at 800 °C to 900 °C byan external Pt heater. The released ionsflow to a cathode where the ion current ismeasured and then amplified (halogendiode principle). This effect is so great thatpartial pressures for halogens can be mea-sured down to 10-7 mbar.

Whereas such devices were used in thepast for leak testing in accordance with thevacuum method, today – because of theproblems associated with the CFCs – moresniffer units are being built. The attainabledetection limit is about 1 · 10-6 mbar · l/sfor all the devisces. Equipment operating inaccordance with the halogen diode princi-ple can also detect SF6. Consequently thesesniffer units are used to determine whetherrefrigerants are escaping from a refrigera-tion unit or from an SF6 type switch box (fil-led with arc suppression gas).

5.5.2 Leak detectors with massspectrometers (MSLD)

The detection of a test gas using massspectrometers is far and away the mostsensitive leak detection method and the onemost widely used in industry. The MS leak

Method Test gas Pressure rangeSmallest detectableleak rate

Quantitativemeasurement

mbar /s` g/a R 134 a

Positive pressure

Positive pressure

Positive pressure

Vacuum

Positive pressure

Positive pressure

Positive pressure

Positive pressureand vacuum

Positive pressure(vacuum)

10–4

10–4

7

70

10–5

10(10 )

–6

–5

1010

–12

–7

10 – 10–3 –5

10–2

10–3

10–2

70

7 · 10–1

7 · 10–1

7 · 10–1

7 · 10(10 )

–3

–1

7 · 10–3

7 · 107 · 10

–9

–4

10 – 7–1

10–4

Air

Air and othergases

Water

Air and othergases

Helium

Refrigerants,helium andother gases

Substancescontaininghalogens

Gases otherthan air

Air and others

Air and others

Pressurerise test

Pressuredrop test

Water pressuretest

Bubble test

Heliumleak detection

Universalsnifferleak detector

Halogenleak detection

Thermal conducti-vity leak detector

Ultrasonicmicrophone

Foamingliquids

Vacuum,positive pressure

No

No

No

Yes

Yes

Yes

Yes

No

No

Withlimitations

Table 5.4 Comparison of leak detection methods

D00 E 19.06.2001 21:39 Uhr Seite 110

Test gase.g. He

Test specimen

Leak detector

Exhaust

Test connection

Q = p · SHe He effHe



detectors developed for this purpose makepossible quantitative measurement of leakrates in a range extending across manypowers of ten (see Section 5.2) wherebythe lower limit ≈ 10-12 mbar · l/s, thusmaking it possible to demonstrate the inhe-rent gas permeability of solids where heli-um is used as the test gas. It is actuallypossible in principle to detect all gasesusing mass spectrometry. Of all the availa-ble options, the use of helium as a tracergas has proved to be especially practical.The detection of helium using the massspectrometer is absolutely (!) unequivocal.Helium is chemically inert, non-explosive,non-toxic, is present in normal air in a con-centration of only 5 ppm and is quite eco-nomical. Two types of mass spectrometerare used in commercially available MSLD’s:a) The quadrupole mass spectrometer, alt-

hough this is used less frequently dueto the more elaborate and complexdesign (above all due to the electricalsupply for the sensor), or

b) the 180° magnetic sector field massspectrometer, primarily due to the rela-tively simple design.

Regardless of the functional principleemployed, every mass spectrometer com-prises three physically important sub-systems: the ion source, separationsystem and ion trap. The ions must be ableto travel along the path from the ion sour-ce and through the separation system tothe ion trap, to the greatest possible extentwithout colliding with gas molecules. Thispath amounts to about 15 cm for all typesof spectrometers and thus requires amedium free path length of at least 60 cm,corresponding to pressure of about 1 · 10-4 mbar; in other words, a massspectrometer will operate only in a vacu-um. Due to the minimum vacuum level of1 · 10-4 mbar, a high vacuum will be requi-red. Turbomolecular pumps and suitableroughing pumps are used in modern leakdetectors. Associated with the individualcomponent groups are the required electri-cal- and electronic supply systems andsoftware which, via a microprocessor,allow for the greatest possible degree ofautomation in the operating sequence,including all adjustment and calibrationroutines and measured value display.

5.5.2.1 The operating principle for aMSLD

The basic function of a leak detector andthe difference between a leak detector andmass spectrometer can be explained usingFigure 5.6. This sketch shows the mostcommonly found configuration for leakdetection using the helium spray method(see Section 5.7.1) at a vacuum compo-nent. When the sprayed helium is drawninto the component through a leak it ispumped thorough the interior of the leakdetector to the exhaust, where it again lea-ves the detector. Assuming that the detec-tor itself is free of leaks, the amount of gasflowing through each pipe section (at anydesired point) per unit of time will remainconstant regardless of the cross sectionand the routing of the piping. The follo-wing applies for the entry into the pumpingport at the vacuum pump:

Q = p · S (5.4)

At all other points

Q = p · Seff (5.4a)

applies, taking the line losses into account.

The equation applies to all gases which arepumped through the piping and thus alsofor helium.

QHe = pHe · Seff, He (5.4b)

In this case the gas quantity per unit oftime is the leak rate being sought; the totalpressure may not be used, but only theshare for helium or the partial pressure forhelium. This signal is delivered by themass spectrometer when it is set for ato-mic number 4 (helium). The value for Seffis a constant for every series of leak detec-tors, making it possible to use a micropro-cessor to multiply the signal arriving fromthe mass spectrometer by a numericalconstant and to have the leak rate dis-played direct.

5.5.2.2 Detection limit, background,gas storage in oil (gas ballast), floating zero-pointsuppression

The smallest detectable leak rate is dicta-ted by the natural background level for thegas to be detected. Even with the testconnector at the leak detector closed,every gas will pass – counter to the pum-ping direction – through the exhaust andthrough the pumps (but will be reducedaccordingly by their compression) throughto the spectrometer and will be detectedthere if the electronic means are adequateto do so. The signal generated representsthe detection limit. The high vacuumsystem used to evacuate the mass spec-

D00Fig. 5.6 Basic operating principle for a leak detector

D00 E 19.06.2001 21:39 Uhr Seite 111



trometer will normally comprise a turbo-molecular pump and an oil-sealed rotaryvane pump. (Diffusion pumps were usedearlier instead of the turbomolecularpumps.) Like every liquid, the sealing oil inthe rotary vane pump has the capability ofdissolving gases until equilibrium is rea-ched between the gas dissolved in the oiland the gas outside the oil. When thepump is warm (operating temperature)this equilibrium state represents the detec-tion limit for the leak detector. The heliumstored in the oil thus influences the detec-tion limit for the leak detector. It is possi-ble for test gas to enter not only throughthe test connection and into the leak detec-tor; improper installation or inept handlingof the test gas can allow test gas to enterthrough the exhaust and the airing or gasballast valve and into the interior of the

detector, to increase the helium level in theoil and the elastomer seals there and thusto induce a background signal in the massspectrometer which is well above the nor-mal detection limit. When the device isproperly installed (see Fig. 5.7) the gasballast valve and the airing valve will beconnected to fresh air and the dischargeline (oil filter!) should at least be routed tooutside the room where the leak test takesplace.

An increased test gas (helium) backgroundlevel can be lowered by opening the gasballast valve and introducing gas which isfree of the test gas (helium-free gas, freshair). The dissolved helium will be flushedout, so to speak. Since the effect alwaysaffects only the part of the oil present inthe pump body at the particular moment,the flushing procedure will have to be con-

tinued until all the oil from the pump’s oilpan has been recirculated several times.This period of time will usually be 20 to 30minutes.

In order to spare the user the trouble ofalways having to keep an eye on the back-ground level, what has been dubbed floa-ting zero-point suppression has been inte-grated into the automatic operating con-cepts of all INFICON leak detectors (Sec-tion 5.5.2.5). Here the background levelmeasured after the inlet valve has beenclosed is placed in storage; when the valveis then opened again this value will auto-matically be deducted from subsequentmeasurements. Only at a relatively highthreshold level will the display panel showa warning indicating that the backgroundnoise level is too high. Figure 5.8 is provi-ded to illustrate the process followed inzero point suppression. Chart on the left.The signal is clearly larger than the back-ground. Center chart: the background hasrisen considerably; the signal can hardlybe discerned. Chart on the right: the back-ground is suppressed electrically; the sig-nal can again be clearly identified.

Independent of this floating zero-pointsuppression, all the leak detectors offerthe capability for manual zero point shif-ting. Here the display for the leak detectorat the particular moment will be “reset tozero” so that only rises in the leak ratefrom that point on will be shown. This ser-ves only to facilitate the evaluation of a dis-play but can, of course, not influence itsaccuracy.

Modern leak detectors are being more fre-quently equipped with oil-free vacuumsystems, the so-called “dry leak detectors”(UL 200 dry, UL 500 dry). Here the pro-blem of gas being dissolved in oil does notoccur but similar purging techniques willnonetheless be employed.

5.5.2.3 Calibrating leak detectors; test leaks

Calibrating a leak detector is to be under-stood as matching the display at a leakdetector unit, to which a test leak is atta-ched, with the value shown on the “label”or calibration certificate. The prerequisitefor this is correct adjustment of the ionpaths in the spectrometer, also known astuning. Often the distinction is not madequite so carefully and both procedures

Fig. 5.7 Correct set-up for a MSLD

Test connection

Venting valve

Mass spectrometer

Turbomolecular pump

Roughing pump

Gas ballast valve

Exhaust

Fig. 5.8 Example of zero-point suppression

10–6

10–7

10–8

10–9

10–10

10–11

caution

Equipment background level: < 2 · 10–10 1 · 10–8 1 · 10–10 (suppressed)Leak: 2 · 10–8 2 · 10–8 2 · 10–8

Display: 2 · 10–8 3 · 10–8 2 · 10–8

prec

fine

D00 E 19.06.2001 21:39 Uhr Seite 112

QMA 200

externalparticlefilter

flowlimiter 1

internalparticlefilter

flow divider 1

flow divider 2

flow meter

flowlimiter 2

flowlimiter 3

pv

1

2



together are referred to as calibration.

In the calibration process proper thestraight-line curve representing the nume-rically correct, linear correlation betweenthe gas flow per unit of time and the leakrate is defined by two points: the zero point(no display where no emissions are detec-ted) and the value shown with the test leak(correct display for a known leak).

In vacuum operations (spray technique,see Section 5.7.1) one must differentiatebetween two types of calibration: with aninternal or external test leak. When using atest leak built into the leak detector the unitcan itself be calibrated but it can only cali-brate itself. When using an external testleak not just the device but also a comple-te configuration, such as a partial flowarrangement, can be included. Internal testleaks are permanently installed and cannotbe misplaced. At present all the leak detec-tors being distributed by INFICON are fit-ted with an automatic calibration routine.

Sniffer units or configurations will as a rulehave to be calibrated with special, externaltest leaks in which there is a guarantee thaton the one hand all the test gas issuingfrom the test leak reaches the tip of theprobe and on the other hand that the gasflow in the sniffer unit is not hindered bycalibration. When making measurementsusing the sniffer technique (see Section5.7.2) it is also necessary to take intoaccount the distance from the probe tip tothe surface of the specimen and the scan-ning speed; these must be included as apart of the calibration. In the special casewhere helium concentration is being mea-sured, calibration can be made using thehelium content in the air, which is a uni-form 5 ppm world-wide.

Test leaks (also known as standard leaksor reference leaks) normally comprise agas supply, a choke with a defined con-ductance value, and a valve. The configu-ration will be in accordance with the testleak rate required. Figure 5.9 showsvarious test leaks. Permeation leaks areusually used for leak rates of 10-10 < QL < 10-7, capillaries, between 10-8 and 10-4 and, for very large leak ratesin a range from 10 to 1000 mbar · l/s, pipesections or orifice plates with exactly defined conductance values (dimensions).Test leaks used with a refrigerant chargerepresent a special situation since the ref-rigerants are liquid at room temperature.Such test leaks have a supply space forliquid from which, through a shut-offvalve, the space filled only with the refrige-rant vapor (saturation vapor pressure) canbe reached, ahead of the capillary leak.One technological problem which is diffi-cult to solve is posed by the fact that allrefrigerants are also very good solvents foroil and grease and thus are often seriouslycontaminated so that it is difficult to fill thetest leaks with pure refrigerant. Decisivehere is not only the chemical compositionbut above all dissolved particles which canrepeatedly clog the fine capillaries.

5.5.2.4 Leak detectors with quadrupole mass spectrometer (Ecotec II)

INFICON builds leak detectors with qua-drupole mass spectrometers to registermasses greater than helium. Apart fromspecial cases, these will be refrigerants.These devices thus serve to examine thetightness of refrigeration units, particular-ly those for refrigerators and air conditio-ning equipment.

Figure 4.2 shows a functional diagram fora quadrupole mass spectrometer. Of thefour rods in the separation system, the twopairs of opposing rods will have identicalpotential and excite the ions passingthrough along the center line so that theyoscillate transversely. Only when theamplitude of these oscillations remainssmaller than the distance between the rodscan the appropriate ion pass through thesystem of rods and ultimately reach theion trap, where it will discharge and thusbe counted. The flow of electrons thuscreated in the line forms the measurementsignal proper. The other ions come intocontact with one of the rods and will beneutralized there.

Figure 5.10 shows the vacuum schematicfor an Ecotec II. The mass spectrometer

D00Fig. 5.9 Examples for the construction of test leaks

a b c d e

a Reference leak without gas supply, TL4, TL6b Reference leak for sniffer and vacuum

applications, TL4-6c (Internal) capillary test leak TL7 d Permeation (diffusion) reference leak, TL8e Refrigerant calibrated leak

Fig. 5.10 Vacuum schematic for the Ecotec II

1 Diaphragm pump2 Piezo-resistive

pressure sensor3 Turbomolecular

pump4 Quadrupole mass

spectrometer5 Sniffer line6 Gas flow limiter7 Gas flow limiter8 Gas flow meter

1

2

3

4

6

7

8

5

5

D00 E 19.06.2001 21:39 Uhr Seite 113



(4) only operates under high vacuum con-ditions, i.e. the pressure here must alwaysremain below 10-4 mbar. This vacuum isgenerated by the turbomolecular pump (3)with the support of the diaphragm pump(1). The pressure PV between the twopumps is measured with a piezo resistivemeasuring system (2) and this pressurelies in the range between 1 to 4 mbar whilein the measurement mode. This pressuremust not exceed a value of 10 mbar asotherwise the turbomolecular pump willnot be capable of maintaining the vacuumin the mass spectrometer. The unit caneasily be switched over at the control unitfrom helium to any of various refrigerants,some of which may be selected as desired.Naturally the unit must be calibrated sepa-rately for each of these masses. Once set,however, the values remain available instorage so that after calibration has beeneffected for all the gases (and a separatereference leak is required for each gas!) it

is possible to switch directly from one gasto another.

5.5.2.5 Helium leak detectors with180° sector mass spectrometer (UL 200, UL 500)

These units are the most sensitive and alsoprovide the greatest degree of certainty.Here “certain” is intended to mean thatthere is no other method with which onecan, with greater reliability and better sta-bility, locate leaks and measure themquantitatively. For this reason helium leakdetectors, even though the purchase priceis relatively high, are often far more eco-nomical in the long run since much lesstime is required for the leak detection pro-cedure itself.

A helium leak detector comprises basicallytwo sub-systems in portable units andthree in stationary units. These are:

1. the mass spectrometer

2. the high vacuum pump and

3. the auxiliary roughing pump system instationary units.

The mass spectrometer (see Fig. 5.11)comprises the ion source (1–4) and thedeflection system (5–9). The ion beam isextracted through the orifice plate (5) andenters the magnetic field (8) at a certainenergy level. Inside the magnetic field theions move along circular paths wherebythe radius for a low mass is smaller thanthat for higher masses. With the correctsetting of the acceleration voltage duringtuning one can achieve a situation in whichthe ions describe a circular arc with a defi-ned curvature radius. Where mass 4 (heli-um) is involved, they pass thorough theaperture (9) to the ion trap (13). In somedevices the discharge current for the ionsimpinging upon the total pressure electro-des will be measured and evaluated as atotal pressure signal. Ions with masseswhich are too small or too great should notbe allowed to reach the ion trap (13) at all,but some of these ions will do so in spiteof this, either because they are deflectedby collisions with neutral gas particles orbecause their initial energy deviates too farfrom the required energy level. These ionsare then sorted out by the suppressor (11)so that only ions exhibiting a mass of 4(helium) can reach the ion detector (13).The electron energy at the ion source is 80eV. It is kept this low so that componentswith a specific mass of 4 and higher –such as multi-ionized carbon or quadruplyionized oxygen – cannot be created. Theion sources for the mass spectrometer aresimple, rugged and easy to replace. Theyare heated continuously during operationand are thus sensitive to contamination.The two selectable yttrium oxide coated iri-dium cathodes have a long service life.These cathodes are largely insensitive toair ingress, i.e. the quick-acting safety cut-out will keep them from burning out evenif air enters. However, prolonged use of theion source may eventually lead to cathodeembrittlement and can cause the cathodeto splinter if exposed to vibrations orshock.

Depending on the way in which the inlet isconnected to the mass spectrometer, onecan differentiate between two types ofMSLD.

1

2345

6

7

89

10

11

12

13

14

Fig. 5.11 Configuration of the 180° sector mass spectrometer

1 Ion source flange2 Cathode

(2 cathodes, Ir + Y2O3)3 Anode4 Shielding of the ion

source with discharge orifice

10 Magnetic field 11 Suppressor12 Shielding of the ion trap13 Ion trap14 Flange for ion trap with

preamplifier

5 Extractor6 Ion traces for M > 47 Total pressure electrode8 Ion traces for M = 49 Intermediate orifice

plate

D00 E 19.06.2001 21:39 Uhr Seite 114



5.5.2.6 Direct-flow and counter-flowleak detectors

Figure 5.12 shows the vacuum schematicfor the two leak detector types. In bothcases the mass spectrometer is evacuatedby the high vacuum pumping system com-prising a turbomolecular pump and arotary vane pump. The diagram on the leftshows a direct-flow leak detector. Gasfrom the inlet port is admitted to the spec-trometer via a cold trap. It is actually equi-valent to a cryopump in which all thevapors and other contaminants condense.(The cold trap in the past also providedeffective protection against the oil vaporsof the diffusion pumps used at that time).The auxiliary roughing pump system ser-ves to pre-evacuate the components to betested or the connector line between theleak detector and the system to be tested.Once the relatively low inlet pressure(pumping time!) has been reached, thevalve between the auxiliary pumpingsystem and the cold trap will be opened forthe measurement. The Seff used in equati-on 5.4b is the pumping speed of the tur-bomolecular pump at the ion source loca-tion:

QHe = pHe · Seff,turbomolecular pump ion source(5.5a)

In the case of direct-flow leak detectors, anincrease in the sensitivity can be achievedby reducing the pumping speed, for exam-ple by installing a throttle between the tur-bomolecular pump and the cold trap. Thisis also employed to achieve maximumsensitivity. To take an example:

The smallest detectable partial pressure

for helium is pmin,He = 1 · 10-12 mbar. The pumpingspeed for helium would be SHe = 10 l/s.Then the smallest detectable leak rate is Qmin = 1 · 10-12 mbar · 10 l/s = 1 · 10-11 mbar · l/s. If the pumping speedis now reduced to 1/s, then one will achie-ve the smallest detectable leak rate of 1 · 10-12 mbar · l/s. One must keep inmind, however, that with the increase inthe sensitivity the time constant for achie-ving a stable test gas pressure in the testspecimen will be correspondingly larger(see Section 5.5.2.9).

In Figure 5.12 the right hand diagramshows the schematic for the counter-flowleak detector. The mass spectrometer, thehigh vacuum system and also the auxiliaryroughing pump system correspond exact-ly to the configuration for the direct-flowarrangement. The feed of the gas to beexamined is however connected betweenthe roughing pump and the turbomolecu-lar pump. Helium which reaches thisbranch point after the valve is opened willcause an increase in the helium pressurein the turbomolecular pump and in themass spectrometer. The pumping speedSeff inserted in equation 5.4b is the pum-ping speed for the rotary vane pump at thebranch point. The partial helium pressureestablished there, reduced by the heliumcompression factor for the turbomolecularpump, is measured at the mass spectro-meter. The speed of the turbomolecularpump in the counter-flow leak detectors isregulated so that pump compression alsoremains constant. Equation 5.5b is derivedfrom equation 5.5a:

QHe = pHe · Seff · K (5.5b)

Seff = effective pumping speed at the rotary vane pump at the

branching pointK = Helium compression factor at

the turbomolecular pump

The counter-flow leak detector is a particu-lar benefit for automatic vacuum unitssince there is a clearly measurable pressu-re at which the valve can be opened, name-ly the roughing vacuum pressure at theturbomolecular pump. Since the turbo-molecular pump has a very large compres-sion capacity for high masses, heavy mo-lecules in comparison to the light test gas,helium (M = 4), can in practice not reachthe mass spectrometer. The turbomolecu-lar pump thus provides ideal protection forthe mass spectrometer and thus elimina-tes the need for an LN2 cold tap, which iscertainly the greatest advantage for theuser. Historically, counter-flow leak detec-tors were developed later. This was due inpart to inadequate pumping speed stabili-ty, which for a long time was not sufficientwith the rotary vane pumps used here. Forboth types of leak detector, stationaryunits use a built-in auxiliary pump to assistin the evacuation of the test port. With por-table leak detectors, it may be necessary toprovide a separate, external pump, thisbeing for weight reasons.

5.5.2.7 Partial flow operation

Where the size of the vacuum vessel or theleak makes it impossible to evacuate thetest specimen to the necessary inlet pres-sure, or where this would simply take toolong, then supplementary pumps will haveto be used. In this case the helium leakdetector is operated in accordance with theso-called “partial flow” concept. Thismeans that usually the larger part of thegas extracted from the test object will beremoved by an additional, suitably dimen-sioned pump system, so that only a part ofthe gas stream reaches the helium leakdetector (see Fig. 5.13). The splitting ofthe gas flow is effected in accordance withthe pumping speed prevailing at the bran-ching point. The following then applies:

QVacuum vessel = γ · DisplayLeak detector (5.6)

where g is characterized as the partial flowratio, i.e. that fraction of the overall leakcurrent which is displayed at the detector.

D00

Test specimen Test specimen

Test gas stream Test gas stream

ppLN

p

p High vacuum pump

High vacuum pump

Roughing pumpRoughing pump

Cold trap:S = 6.1 /s · cmFl 1000 cmS = 6,100 /s

`

`

2

2≈

Auxiliary pumpAuxiliary pump

< 10 mbar–4

< 10 mbar–4

HeHe2

TOT

TOT

MS MS

Fig. 5.12 Full-flow and counter-flow leak detector

Solution 1: Direct-flow leak detector Solution 2: Counter-flow leak detector

D00 E 19.06.2001 21:39 Uhr Seite 115



Where the partial flow ratio is unknown, gcan be determined with a reference leakattached at the vacuum vessel:

Display at the leak detectorγ = ——————————— (5.7)

QL for the reference leak

5.5.2.8 Connection to vacuum systems

The partial flow concept is usually used inmaking the connection of a helium leakdetector to vacuum systems with multi-stage vacuum pump sets. When conside-ring where to best make the connection, itmust be kept in mind that these are usual-ly small, portable units which have only alow pumping speed at the connection flan-ge (often less than 1 l/s). This makes it allthe more important to estimate – based on

the partial flow ratio to be expected vis àvis a diffusion pump with pumping speedof 12000 l/s, for example – which leakrates can be detected at all. In systemswith high vacuum- and Roots pumps, thesurest option is to connect the leak detec-tor between the rotary vane pump and theroots pump or between the roots pumpand the high vacuum pump. If the pressu-re there is greater than the permissibleinlet pressure for the leak detector, thenthe leak detector will have to be connectedby way of a metering (variable leak) valve.Naturally one will have to have a suitableconnector flange available. It is also advis-able to install a valve at this point from theoutset so that, when needed, the leakdetector can quickly be coupled (with thesystem running) and leak detection cancommence immediately after opening the

valve. In order to avoid this valve beingopened inadvertently, it should be sealedoff with a blank flange during normal vacu-um system operation.

A second method for coupling to largersystems, for example, those used forremoving the air from the turbines in po-wer generating stations, is to couple at thedischarge. A sniffer unit is inserted in thesystem where it discharges to atmosphe-re. One then sniffs the increase in the heli-um concentration in the exhaust. Withouta tight coupling to the exhaust, however,the detection limit for this application willbe limited to 5 ppm, the natural heliumcontent in the air. In power plants it is suf-ficient to insert the tip of the probe at anangle of about 45° from the top into thedischarge line (usually pointing upward) ofthe (water ring) pump.

5.5.2.9 Time constants

The time constant for a vacuum system isset by

(5.8)

τ = Time constantV = Volume of the containerSeff = Effective pumping speed, at the

test object

Figure 5.14 shows the course of the signalafter spraying a leak in a test specimenattached to a leak detector, for three diffe-rent configurations:1. Center: The specimen with volume of V

is joined directly with the leak detectorLD (effective pumping speed of S).

2. Left: In addition to 1, a partial flowpump with the same effective pumpingspeed, S’ = S, is attached to the testspecimen.

3. Right: As at 1, but S is throttled down to0.5 · S.

The signals can be interpreted as follow:

1: Following a “dead period” (or “delaytime”) up to a discernible signal level, thesignal, which is proportional to the partialpressure for helium, will rise to its fullvalue of pHe = Q/Seff in accordance withequation 5.9:

(5.9 )p QS

eHeeff

t

= ⋅ −

–

1 τ

τ = VSeff

mbar ·s

`

mbar ·s

`

15024.66

`m

VSeff

mbar ·s

`

mbar ·s

`

mbar ·s

`

8(8 + 16.66)

16,66(8 + 16.66)

8

mbar ·s

`

`s

`s

`s

`s

`s

`s

`s

`s

ms

3

Partial flow principle (example)

V = 150 `

S = 60 = 16.66 Partial flow pump (PFP)PFP

S = S + Seff PFP LD S = 8 Leak detector (LD)LD→

Q = 3 · 10 (Leak rate)He–5

A)

Signal to Leak detector: 3 · 10 · = 9.73 · 10–5 –6

Signal to partial flowpump: 3 · 10 · = 2.02 · 10–5 –5

Check: Overall signal Q = Q + Q = 3.00 · 10He LD PFP–5

γ ... Partial flow ratio

Overal pumping speed: S = S + S = 8 + 16.66 = 24.66eff LD PFP

Signal amplitude:Splitting of the gas flow (also of the testgas!) in accordance with the effective pumpingspeed at the partial flow branch point

H

HPartial flow ratio = Fraction of the overall flow to the leak detector

B) Response time: t = 3 · = 3 · = 18.25 s95%

Estimate: Value for S, V and are uncertain certain: calibrate with reference leakγ →

γ = = =

γ = =or

QQ

LD

He

QQ + Q

LD

LD PFP

SS + S

LD

LD PFP

11 + nnn

11 + nnn

QQ

PFP

LD

SS

PFP

LD

Q = · QLD Heγ

Display Leak rate

Fig. 5.13 Partial flow principle

D00 E 19.06.2001 21:39 Uhr Seite 116



The signal will attain a prortion of its ulti-mate value after

t = 1 τ . . 63.3 % t = 2 τ . . 86.5 %t = 3 τ . . 95.0 % t = 4 τ . . 98.2 %t = 5 τ . . 99.3 % t = 6 τ . . 99.8 %

The period required to reach 95 % of theultimate value is normally referred to asthe response time.

2: With the installation of the partial flowpump both the time constant and the sig-nal amplitude will be reduced by a factor of2; that means a quicker rise but a signalwhich is only half as great. A small timeconstant means quick changes and thusquick display and, in turn, short leak detec-tion times.

3: The throttling of the pumping speed to0.5 S, increases both the time constantand the signal amplitude by a factor of 2. Alarge value for t thus increases the timerequired appropriately. Great sensitivity,achieved by reducing the pumping speed,is always associated with greater timerequirements and thus by no means isalways of advantage.

An estimate of the overall time constantsfor several volumes connected one behindto another and to the associated pumpscan be made in an initial approximation byadding the individual time constants.

5.6 Limit values / Specifications forthe leak detector

1. The smallest detectable leak rate.

2. The effective pumping speed at thetest connection.

3. The maximum permissible pressureinside the test specimen (also themaximum permissible inlet pressure).This pressure pmax will be about 10-1

for LDs with classical PFPs and about 2to 10 mbar for LDs with compoundPFPs. The product of this maximumpermissible operating pressure and thepumping speed S of the pump systemat the detector’s test connection is themaximum permissible throughput:Qmax = pmax · Seff, connector (5.10)

This equation shows that it is by no meansadvantageous to attain high sensitivity bythrottling down the pumping speed. Themaximum permissible throughput wouldotherwise be too small. The unit is not fun-ctional when – either due to one large leakor several smaller leaks – more gas flowsinto the unit than the maximum permissi-ble throughput rate for the leak detector.

5.7 Leak detectiontechniques usinghelium leak detectors

5.7.1 Spray technique (local leak test)

The test specimen, connected to the heli-um leak detector, is slowly traced with avery fine stream of helium from the spraypistol, aimed at likely leakage points (wel-ding seams, flange connectors, fusedjoints), bearing in mind the time constantof the system as per Equation 5.8 (see Fig.5.14). The volume sprayed must be adju-sted to suit the leak rate to be detected andthe size and accessibility of the objectbeing tested. Although helium is lighterthan air and therefore will collect beneaththe ceiling of the room, it will be so welldistributed by drafts and turbulence indu-ced by movements within the room thatone need not assume that helium will befound primarily (or only) at the top of theroom during search for leaks. In spite ofthis, it is advisable, particularly when dea-ling with larger components, to start thesearch for leaks at the top.

In order to avoid a surge of helium whenthe spray valve is opened (as this would“contaminate” the entire environment) it isadvisable to install a choke valve to adjustthe helium quantity, directly before or afterthe spray pistol (see Fig. 5.15). The correctquantity can be determined easiest by sub-merging the outlet opening in a containerof water and setting the valve on the basisof the rising bubbles. Variable-area flow-meters are indeed available for the requi-red small flow quantities but are actuallytoo expensive. In addition, it is easy to usethe water-filled container at any time todetermine whether helium is still flowing.

The helium content of the air can also bedetected with helium leak detectors wherelarge leaks allow so much air to enter thevessel that the 5 ppm share of helium inthe air is sufficient for detection purposes.The leak rate is then:

Display (pure He) Display (atmosph. He)———————– = ——————————

1 5 · 10-6

QL = Display (pure He) (5.11)

= 2 · 10+5 · Display (atmospheric He)

D00

3

2

1

100%

100%1,0

2,0

0,5

0

100%

95%

95%

95%

toDead time

3 · 3 · 3 · =

( = ... Time constant)τTime

2 (3 · )(3 · )=

S 2/

Signalamplitude

QS

Q = 2p

= p

Signal riseQS

12

VS

VS

V VS

VS

VS

P 2/Q

S + S’=

S 2/V

S + S’

Compensation period, e.g. t = 3 · = 3 ·95% τ

Slower, more sensitiveFaster, less sensitive normal

32 1

S’S S S 2/

MS MS MS

Throttle

Q

V V V

Q Q

LD LD LD

Fig. 5.14 Signal responses and pumping speed

D00 E 19.06.2001 21:39 Uhr Seite 117



5.7.2 Sniffer technology (local leak test using thepositive pressuremethod)

Here the points suspected of leaking at thepressurized test specimen (see Fig. 5.4, d)are carefully traced with a test gas probewhich is connected with the leak detectorby way of a hose. Either helium or hydro-gen can be detected with the INFICON heli-um leak detectors. The sensitivity of themethod and the accuracy of locating leakypoints will depend on the nature of thesniffer used and the response time for theleak detector to which it is connected. Inaddition, it will depend on the speed atwhich the probe is passed by the leakpoints and the distance between the tip ofthe probe and the surface of the test spe-cimen. The many parameters which play apart here make it more difficult to determi-ne the leak rates quantitatively. Using snif-fer processes it is possible, virtually inde-pendent of the type of gas, to detect leakrates of about 10-7 mbar · l/s. The limitati-on of sensitivity in the detection of heliumis due primarily to the helium in the atmos-phere (see Chapter 9, Table VIII). In regardto quantitative measurements, the leak

detector and sniffer unit will have to becalibrated together. Here the distance fromthe specimen and the tracing speed willhave to be included in calibration, too.

5.7.3 Vacuum envelope test (integral leak test)

Vacuum envelope tests are integral leaktests using helium as the test gas, in whichthe test specimen is enclosed either in arigid (usually metal) enclosure or in a lightplastic envelope. The helium which entersor leaves (depending on the nature of thetest) the test specimen is passed to a heli-um leak detector, where it is measured.Envelope tests are made either with thetest specimen pressurized with helium(Fig. 5.4c) or with the test specimen eva-cuated (Fig. 5.4a). In both cases it may benecessary to convert the helium enrich-ment figure (accumulation) to the heliumstandard leak rate.

5.7.3.1 Envelope test – test specimenpressurized with helium

a) Envelope test with concentration measurement and subsequent leakrate calculation

To determine overall leakiness of a testobject pressurized with helium the objectshall be enclosed in an envelope which iseither rigid or deformable (plastic). Thetest gas leaving the leaks accumulates sothat the helium concentration in the enve-lope rises. Following an enrichment periodto be determined (operating period) thechange in concentration inside the envelo-pe will be measured with a sniffer connec-ted to the helium detection unit. The over-all leak rate (integral leak rate) can be cal-culated following calibration of the testconfiguration with a reference concen-tration, e.g. atmospheric air. This methodmakes it possible to detect even the smal-lest overall leakiness and is suitable in par-ticular for automated industrial leaktesting. Due to gas accumulation, thelimits for normal sniffer techniques areshifted toward lower leak rates and theambient conditions such as temperature,air flow and sniffer tracing speed loseinfluence. When using plastic envelopes itis necessary to take into account heliumpermeation through the plastic envelopeduring long enrichment periods.

b) Direct measurement of the leak ratewith the leak detector (rigid envelope)

When the test specimen, pressurized withhelium, is placed in a rigid vacuum cham-ber, connected to a helium leak detector,the integral leak rate can be read directly atthe leak detector.

5.7.3.2 Envelope test with test specimen evacuated

a) Envelope = “plastic tent”The evacuated test specimen is surroun-ded by a light-weight (plastic) enclosureand this is then filled with helium once theatmospheric air has been removed. Whenusing a plastic bag as the envelope, thebag should be pressed against the testspecimen before filling it with helium inorder to expel as much air as possible andto make the measurement with the puresthelium charge possible. The entire outsidesurface of the test object is in contact withthe test gas. If test gas passes throughleaks and into the test specimen, then the

Fig. 5.15 Helium spray equipment

Avoiding the “helium surge” when the pistol valve is openeda) Throttle hose orb) Adjustable throttle valve ahead of the spray pistolMinimum helium quantity for correct display: Changing the setting for the throttle shall not affectindication.The minimum quantity is always much smaller than one would set without a flowmeter (e.g. bylistening for flow or letting the helium flow across moistened lips). The simplest check without aflowmeter: Letting gas bubble through water.

D00 E 19.06.2001 21:39 Uhr Seite 118



integral leak rate will be indicated, regard-less of the number of leaks. In addition, itis necessary to observe when repeatingtesting in enclosed areas that the heliumcontent of the room will rise quite rapidlywhen the envelope is removed. Using pla-stic bags is thus more advisable for “one-off” testing of large plants. The plasticenvelope used here is often referred to asa “tent”.

b) Rigid envelopeThe use of a solid vacuum vessel as therigid envelope, on the other hand, is betterfor repetitive testing where an integral testis to be made. When solid envelopes areused it is also possible to recover the heli-um once the test has been completed.

5.7.4 “Bombing” test, “Storage under pressure”

The “bombing” test is use to check thetightness of components which are alrea-dy hermetically sealed and which exhibit agas-filled, internal cavity. The componentsto be examined (e.g. transistors, IC hou-sings, dry-reed relays, reed contact swit-ches, quartz oscillators, laser diodes andthe like) are placed in a pressure vesselwhich is filled with helium. Operating withthe test gas at relatively high pressure (5to 10 bar) and leaving the system standingover several hours the test gas (helium)will collect inside the leaking specimens.This procedure is the actual “bombing”. Tomake the leak test, then, the specimens areplaced in a vacuum chamber following“bombing”, in the same way as describedfor the vacuum envelope test. The overallleak rate is then determined. Specimenswith large leaks will, however, lose theirtest gas concentration even as the vacuumchamber is being evacuated, so that theywill not be recognized as leaky during theactual leak test using the detector. It is forthis reason that another test to registervery large leaks will have to be made priorto the leak test in the vacuum chamber.

5.8 Industrial leaktesting

Industrial leak testing using helium as thetest gas is characterized above all by the factthat the leak detection equipment is fullyintegrated into the manufacturing line. Thedesign and construction of such test unitswill naturally take into account the task to becarried out in each case (e.g. leak testingvehicle rims made of aluminum or leaktesting for metal drums). Mass-produced,standardized component modules will beused wherever possible. The parts to beexamined are fed to the leak testing system(envelope test with rigid envelope and positi-ve pressure [5.7.3.1b] or vacuum [5.7.3.2b]inside the specimen) by way of a conveyorsystem. There they will be examined indivi-dually using the integral methods and auto-matically moved on. Specimens found to beleaking will be shunted to the side.

The advantages of the helium test method,seen from the industrial point of view, maybe summarized as follows:

• The leak rates which can be detectedwith this process go far beyond all prac-tical requirements.

• The integral leak test, i.e. the total leakrate for all individual leaks, facilitatesthe detection of microscopic and spon-ge-like distributed leaks which alto-gether result in leakage losses similarto those for a larger individual leak.

• The testing procedure and sequencecan be fully automated.

• The cyclical, automatic test systemcheck (self-monitoring) of the deviceensures great testing reliability.

• Helium is non-toxic and non-hazardous(no maximum allowable concentrationsneed be observed).

• Testing can be easily documented, indi-cating the parameters and results, on aprinter.

Use of the helium test method will result inconsiderable increases in efficiency (cyc-ling times being only a matter of secondsin length) and lead to a considerableincrease in testing reliability. As a result ofthis and due to the EN/ISO 9000 require-ments, traditional industrial test methods(water bath, soap bubble test, etc.) willnow largely be abandoned.

D00

D00 E 19.06.2001 21:39 Uhr Seite 119

Thin Film Controllers / Control UnitsFundamentals of Vacuum Technology


6 Thin film controllers andcontrol unitswith quartz oscillators

6.1 IntroductionIt took a long time to go from the coatingof quartz crystals for frequency finetuning, which has long been in practice, toutilization of frequency change to determi-ne the mass per unit area as a microbalan-ce with the present-day degree of precisi-on. In 1880 two brothers, J. and P. Curie,discovered the piezoelectric effect. Undermechanical loads on certain quartz crystalsurfaces, electrical charges occur that arecaused by the asymmetrical crystallinestructure of SiO2. Conversely, in a piezo-crystal deformations appear in an electri-cal field and mechanical oscillations occurin an alternating field. A distinction ismade between bending oscillations, thick-ness shear mode and thickness shearoscillations. Depending on the orientationof the cut plane to the crystal lattice, anumber of different cuts are distinguished,of which only the so-called AT cut with acut angle of 35°10" is used in thin film con-trollers because the frequency has a verylow temperature dependence in the rangebetween 0 and 50 °C with this cut. Accor-dingly, an attempt must be made not toexceed this temperature range during coa-ting (water cooling of crystal holder).Since there is still a problem with “quartzcapacity” (i.e. the maximum possible coa-ting thickness of the quartz at which it still

oscillates reliably) despite refined techno-logy, a number of approaches have beendeveloped to expand this capacity:

1. The use of several crystals, one behindthe other, in a multiple crystal holderwith automatic change and data upda-ting in the event of imminent failure of aquartz: CrystalSix.

2. The RateWatcher function, in which thequartz is alternately exposed to the coa-ting beam for a short time until all mea-surements and regulation have beencarried out and then remains coveredby a shutter for a longer period of time.

The selection of the “right” crystal holderthus plays an important role in all measu-rements with quartz oscillators. Variouscrystal holder designs are recommendedfor the different applications: with or with-out shutter, bakeable for UHV, doublecrystal holder or crystal six as well as spe-cial versions for sputter applications. Inaddition to these important and more“mechanical” aspects, the advances inmeasuring and control technology andequipment features will be discussed inthe following.

6.2 Basic principles ofcoating thicknessmeasurement withquartz oscillators

The quartz oscillator coating thicknessgauge (thin film controller) utilizes the pie-zoelectric sensitivity of a quartz oscillator(monitor crystal) to the supplied mass.This property is utilized to monitor thecoating rate and final thickness duringvacuum coating.

A very sharp electromechanical resonanceoccurs at certain discrete frequencies ofthe voltage applied. If mass is added to thesurface of the quartz crystal oscillating inresonance, this resonance frequency isdiminished. This frequency shift is veryreproducible and is now understood preci-sely for various oscillation modes ofquartz. Today this phenomenon, which iseasy to understand in heuristic terms, isan indispensable measuring and processcontrol tool, with which a coating increase

of less than one atomic layer can be detec-ted.

In the late 1950s Sauerbrey and Lostis dis-covered that the frequency shift connectedwith the coating of the quartz crystal is afunction of the change in mass due to thecoating material in the following way:

(6.1)

Mf mass of the coatingMq mass of the quartz prior to coatingFq frequency prior to coatingFc frequency after coating∆F = Ff – Fc ... frequency shift due to coa-

tingIf the following are now applied: Mf = (Mc – Mq) = Df · ρf · A and

Mq = Dq · rq · A, where T = the coatingthickness, ρ = density and A stands forarea while the index q stands for the stateof the “uncoated quartz” and c for the stateafter “frequency shift due to coating”, thefollowing results are obtained for the coa-ting thickness:

with

where

N = Fq · Dq is the frequency constant (forthe AT cut NAT = 166100 Hz · cm) and ρq = 2.649 g/cm3 is the density of thequartz. The coating thickness is thus pro-portional to the frequency shift ∆F andinversely proportional to the density ρf ofthe coating material. The equation

for the coating thickness was used in thefirst coating thickness measuring unitswith “frequency measurement” ever used.According to this equation, a crystal with astarting frequency of 6.0 MHz displays adecline in frequency of 2.27 Hz after coa-ting with 1Å of aluminum (de = 2.77 g/cm3). In this way the growthof a fixed coating due to evaporation orsputtering can be monitored through pre-cise measurement of the frequency shift ofthe crystal. It was only when knowledge ofthe quantitative interrelationship of this

D ·K Ff

f= ρ

∆

K· · ·D F

F

N

Fq q q

q

AT q

q

= =ρ ρ

2 2

D · · · ··

FF

D FF

K Ff

q

qq q

q f f= =ρ

ρ ρ∆ ∆

M MM

M ·or withF F

F Fqq

ff

q q= =∆ ∆

Fig. 6.1 Natural frequency as a function of temperaturein an AT cut quartz crystal

Temperature (°C)

rel.

frequ

ency

cha

nge

(ppm

)

D00 E 19.06.2001 21:40 Uhr Seite 120

Thin Film Controllers / Control Units Fundamentals of Vacuum Technology


effect was acquired that it became possibleto determine precisely the quantity ofmaterial that is deposited on a substrate ina vacuum. Previously this had been prac-tically impossible.

6.3 The shape ofquartz oscillatorcrystals

Regardless of how sophisticated the elec-tronic environment is, the main compo-nent for coating measurement remains themonitor quartz crystal. Originally monitorquartzes had a square shape. Fig. 6.4shows the resonance spectrum of a quartzresonator with the design used today (Fig.6.3). The lowest resonance frequency isinitially given by a thickness shear oscilla-tion, which is called the fundamental wave.

The characteristic motions of the thicknessshear oscillation are parallel to the maincrystal boundary surfaces. In other words:the surfaces are shift antinodes, see Fig.6.2. The resonance frequencies slightlyabove the basic frequency are called“anharmonic” and are a combination ofthickness shear and thickness rotationoscillation forms. The resonance frequen-cy at around three times the value for thefundamental wave is called “quasi-harmo-nic”. Near the quasi-harmonic there arealso a number of anharmonics with aslightly higher frequency.

The design of the monitor crystals usednowadays (see Fig. 6.3) displays a numberof significant improvements over the origi-nal square crystals. The first improvementwas the use of round crystals. The enlar-ged symmetry greatly reduced the numberof possible oscillation modes. A secondgroup of improvements involved providingone of the surfaces with a contour andmaking the excitation electrode smaller.

The two together ensure that the acousticenergy is recorded. Reducing the electrodediameter limits the excitation to the midd-le area. The surface contour consumes theenergy of the moving acoustic waves befo-re they reach the crystal edge. It is notreflected into the center where it couldinterfere with new incoming waves.

Such a small crystal behaves like an infini-tely expanded crystal. However, if thecrystal vibrations remain restricted to thecenter, one can clamp the outer edge to acrystal holder, without engendering unde-sired side effects. Moreover, contouringreduces the resonance intensity of undesi-red anharmonics. This limits the capacityof the resonator to maintain these oscilla-tions considerably.

Use of an adhesive coating has enhancedthe adhesion of the quartz electrode. Eventhe rate spikes occurring with increasingfilm stress (strain) and caused by micro-tears in the coating were reduced. Coatingmaterial remains at these micro-tearswithout adhesion and therefore cannotoscillate. These open areas are not registe-red and thus an incorrect thickness is indi-cated.

Fig. 6.4 shows the frequency behavior of aquartz crystal shaped as in Fig. 6.3. Theordinate represents the amplitude of theoscillation or also the current flowingthrough the crystal as a function of the fre-quency on the abscissa.

Usually an AT cut is chosen for the coatingthickness measurement because throughthe selection of the cut angle the frequen-cy has a very small temperature coefficientat room temperature.

Since one cannot distinguish between

D00

Fig. 6.2 Thickness shear oscillations

NodeE→

↔

Fig. 6.3 Shape of INFICON quartz crystals

Fig. 6.4 Frequency resonance spectrumFrequency (MHz)

log

Rela

tive

inte

nsy

D00 E 19.06.2001 21:40 Uhr Seite 121



• coating: frequency reduction = negati-ve influence

• temperature change:negative or positive influence

• temperature gradients on the crystal,positive or negative

• stresses caused by the coating

it is important to minimize the temperatu-re influence. This is the only way to mea-sure small differences in mass.

6.4 Period measurement

Although the instruments that functionedaccording to equation 6.2 were very use-ful, it soon became obvious that for thedesired accuracy their area of applicationwas typically limited to ∆F < 0.02 Fq. Evenat a relative frequency change of (Fq – Fc) / Fq < 2 %, errors of around 2 %occurred in the coating thickness measu-rement so that the "usable service life" ofthe coating in the case of a 6-MHz monitorcrystal was about 120 kHz.In 1961 Behrndt discovered that:

with (6.3)

Tc = 1 / Fc ... oscillation period, coatedTq = 1 / Fq ... oscillation period, uncoated

The period measurement (measurement ofthe oscillation duration) was the result ofthe introduction of digital time measure-ment and the discovery of the proportiona-lity of crystal thickness Dq and oscillationduration Tq. The necessary precision ofthickness measurement permits applicati-on of equation 6.3 up to about ∆F < 0.05 Fq.

In period measurement a second crystaloscillator is essentially used as a referenceoscillator that is not coated and usuallyoscillates at a much higher frequency thanthe monitor crystal. The reference oscilla-tor generates small precision time inter-vals, with which the oscillation duration ofthe monitor crystal is determined. This isdone by means of two pulse counters: thefirst counts a fixed number of monitoroscillations m. The second is started si-

MM

T TT

FF

f

q

c q

q c=

−=

( ) ∆

multaneously with the first and counts theoscillations of the reference crystal duringm oscillations of the monitor crystal.Because the reference frequency Fr isknown and stable, the time for m monitoroscillations can be determined accuratelyto ± 2/Fr. The monitor oscillation period isthen

where n is the reading of the referencecounter. The accuracy of the measurementis determined by the frequency of the refe-rence oscillator and the length of the coun-ting time that is specified through the sizeof m.

For low coating rates, small densities ofthe coating material and fast measure-ments (that require short counting times),it is important to have a reference oscilla-tor with a high frequency. All of this requi-res great time precision so that the smallcoating-related frequency shifts can beresolved. If the frequency shift of the mo-nitor crystal decreases between two mea-surements on the order of magnitude ofthe frequency measurement accuracy,good rate regulation becomes impossible(rate regulation: regulation of the energysupply to the coating source so that a spe-cified coating thickness growth per timeunit is maintained). The great measure-ment uncertainty then causes more noisein the closed loop, which can only becountered with longer time constants. Thisin turn makes the corrections due tosystem deviation slow so that relativelylong deviations from the desired rateresult. This may not be important for sim-ple coatings, but for critical coatings, as inthe case of optical filters or very thin, slowly growing single-crystal coatings,errors may result. In many cases, the desi-red properties of such coatings are lost ifthe rate deviations are more than one ortwo percent. Finally, frequency and stabili-ty of the reference oscillator determine theprecision of the measurement.

·n

Fr m

6.5 The Z match technique

Miller and Bolef (1968) treated the quartzoscillator and coating system as a single-dimensional, coherent acoustic resonator.Lu and Lewis (1972) developed the simpli-fied Z match equation on that basis. Simul-taneous advances in electronics, particu-larly the microprocessor, made it possibleto solve the Z match equation in real time.Most coating process control units soldtoday use this sophisticated equation,which takes into account the acoustic pro-perties of the quartz oscillator/coatingsystem:

(6.4)

acoustic impedance ratio

Uq = shear module, quartzUf = shear module, film

This led to basic understanding of the con-version of frequency shift into thicknesswhich enabled correct results in a practicaltime frame for process control. To achievethis high degree of accuracy, the usermust only enter an additional materialparameter Zf for the coating material. Thevalidity of the equation was confirmed formany materials and it applies to frequencyshifts up to ∆F < 0.4 Fq! Note that equati-on 6.2 was only valid up to ∆F < 0.02 Fq.And equation 6.3 only up to ∆F < 0.05 Fq.

6.6 The active oscillator

All units developed up to now are based onuse of an active oscillator, as shown sche-matically in Fig. 6.5. This circuit keeps thecrystal actively in resonance so that anytype of oscillation duration or frequencymeasurement can be carried out. In thistype of circuit the oscillation is maintainedas long as sufficient energy is provided bythe amplifier to compensate for losses inthe crystal oscillation circuit and thecrystal can effect the necessary phaseshift. The basic stability of the crystal

Zd U

d Uq q

f f=

⋅⋅

( )TN dd F Z

arctg Z tgF FFf

AT q

f c

q c

q=

⋅⋅ ⋅ ⋅

⋅ ⋅

⋅ −

π

π

D00 E 19.06.2001 21:40 Uhr Seite 122



oscillator is created through the suddenphase change that takes place near theseries resonance point even with a smallchange in crystal frequency, see Fig. 6.6.

Normally an oscillator circuit is designedsuch that the crystal requires a phase shiftof 0 degrees to permit work at the seriesresonance point. Long- and short-termfrequency stability are properties of crystaloscillators because very small frequencydifferences are needed to maintain thephase shift necessary for the oscillation.The frequency stability is ensured throughthe quartz crystal, even if there are long-term shifts in the electrical values that arecaused by “phase jitter” due to temperatu-re, ageing or short-term noise. If mass isadded to the crystal, its electrical pro-perties change.

Fig. 6.7 shows the same graph as Fig 6.6,but for a thickly coated crystal. It has lostthe steep slope displayed in Fig. 6.6.Because the phase rise is less steep, any

noise in the oscillator circuit leads to a lar-ger frequency shift than would be the casewith a new crystal. In extreme cases, theoriginal phase/frequency curve shape isnot retained; the crystal is not able to carryout a full 90° phase shift.

The impedance “Z” can increase to veryhigh values. If this happens, the oscillatorprefers to oscillate in resonance with ananharmonic frequency. Sometimes thiscondition is met for only a short time andthe oscillator oscillation jumps back andforth between a basic and an anharmonicoscillation or it remains as an anharmonicoscillation. This phenomenon is wellknown as “mode hopping”. In addition tothe noise of the rate signal created, thismay also lead to incorrect termination of acoating because of the phase jump. It isimportant here that, nevertheless, the con-troller frequently continues to work underthese conditions. Whether this has occur-red can only be ascertained by noting thatthe coating thickness is suddenly signifi-

cantly smaller, i.e. by the amount of thefrequency difference between the funda-mental wave and the anharmonic adoptedby the oscillation.

6.7 The mode-lockoscillator

INFICON has developed a new technologyfor overcoming these constraints on theactive oscillator. The new system con-stantly analyzes the response of the crystalto an applied frequency: not only to deter-mine the (series) resonance frequency, butalso to ensure that the quartz oscillates inthe desired mode. The new system isinsensitive to mode hopping and the resul-tant inaccuracy. It is fast and precise. Thecrystal frequency is determined 10 times asecond with an accuracy to less than0.0005 Hz.

The ability of the system to initially identifyand then measure a certain mode opensup new opportunities thanks to the advan-tages of the additional information contentof these modes. This new, “intelligent”measuring device makes use of thephase/frequency properties of the quartzcrystal to determine the resonance fre-quency. It works by applying a synthesizedsinus wave of a certain frequency to thecrystal and measuring the phase differen-ce between the applied signal voltage andthe current flowing through the crystal. Inthe case of series resonance, this differen-ce is exactly zero degrees; then the crystalbehaves like an ohmic resistance. By dis-connecting the applied voltage and thecurrent that returns from the crystal, one

D00

Fig. 6.5 Circuit of the active oscillator

crystal

amplifieroutput

Fig. 6.6 Crystal frequencies near the series resonance point

Frequency (MHz)

phas

e (d

egre

es)

log

.Z. (

Ohm

) | impedance |

series resonance

phase

Fig. 6.7 Oscillations of a thickly coated crystal

Frequency (MHz)

phas

e (d

egre

ss)

log

.Z. (

Ohm

)

| impedance |

series resonance

phase

D00 E 19.06.2001 21:40 Uhr Seite 123



can determine with a phase comparatorwhether the applied frequency is higher orlower than the crystal resonance point.

The crystal impedance is capacitive at fre-quencies below the fundamental wave andinductive at frequencies above the reso-nance. This information is useful if theresonance frequency of a crystal is unkno-wn. A brief frequency sweep is carried outuntil the phase comparator changes overand thus marks the resonance. For ATquartzes we know that the lowest usablefrequency is the fundamental wave. Theanharmonics are slightly above that. Thisinformation is not only important for thebeginning, but also in the rare case that theinstrument loses “track” of the fundamen-tal wave. Once the frequency spectrum ofthe crystal is determined, the instrumentmust track the shift in resonance frequen-cy, constantly carry out frequency measu-rements and then convert them intothickness.

Use of the “intelligent” measuring systemhas a number of obvious advantages overthe earlier generation of active oscillators,primarily insensitivity to mode hopping aswell as speed and accuracy of measure-ment. This technique also enables theintroduction of sophisticated propertieswhich were not even conceivable with anactive oscillator setup. The same devicethat permits the new technology to identifythe fundamental wave with one sweep canalso be used to identify other oscillationmodes, such as the anharmonics or quasi-harmonics. The unit not only has a devicefor constantly tracking the fundamentalwave, but can also be employed to jumpback and forth between two or moremodes. This query of different modes cantake place for two modes with 10 Hz on thesame crystal.

6.8 Auto Z match technique

The only catch in the use of equation 6.4 isthat the acoustic impedance must beknown. There are a number of cases wherea compromise has to be made with accu-racy due to incomplete or restricted know-ledge of the material constants of the coa-ting material:

1) The Z values of the solid material oftendeviate from those of a coating. Thincoatings are very sensitive to processparameters, especially in a sputter envi-ronment. As a result, the existing valuesfor solid material are not adequate.

2) For many exotic substances, includingalloys, the Z value is not known and noteasy to determine.

3) It is repeatedly necessary to carry out aprecise coating thickness measurementfor multiple coating with the samecrystal sensor. This applies in particularto optical multiple and semi-conductorcoatings with a high temperature coeffi-cient TC. However, the effective Z valueof the mixture of multiple coatings isunknown.

In such a case, therefore, the only effectivemethod is to assume a Z value of 1, i.e. toignore reality with respect to wave propa-gation in multi-substance systems. Thisincorrect assumption causes errors in theprediction of thickness and rate. Themagnitude of the error depends on thecoating thickness and the amount of devia-tion from the actual Z value.

In 1989 A. Wajid invented the mode-lockoscillator. He presumed that a connectionexisted between the fundamental wave andone of the anharmonics, similar to thatascertained by Benes between the funda-mental oscillation and the third quasi-har-monic oscillation. The frequencies of thefundamental and the anharmonic oscillati-ons are very similar and they solve theproblem of the capacity of long cables. Hefound the necessary considerations forestablishing this connection in works byWilson (1954) as well as Tiersten andSmythe (1979).

The contour of the crystal, i.e. the spheri-cal shape of one side, has the effect ofseparating the individual modes furtherfrom each other and preventing energytransfer from one mode to another. Theusual method of identification is to desig-nate the fundamental oscillation as (100),the lowest anharmonic frequency as (102)and the next higher anharmonic as (120).These three indices of the mode nomen-clature are based on the number of phasereversals in the wave motion along thethree crystal axes. The above mentionedworks by Wilson, Tiersten and Smytheexamine the properties of the modes bystudying the influence of the radius of the

cut on the position of the anharmonic inrelation to the fundamental oscillation.

If one side of the quartz is coated withmaterial, the spectrum of the resonancesis shifted to lower frequencies. It has beenobserved that the three above mentionedmodes have a somewhat differing masssensitivity and thus experience somewhatdifferent frequency shifts. This differenceis utilized to determine the Z value of thematerial. By using the equations for theindividual modes and observing the fre-quencies for the (100) and the (102)mode, one can calculate the ratio of thetwo elastic constants C60 and C55. Thesetwo elastic constants are based on theshear motion. The key element in Wajid’stheory is the following equation:

(6.5)

with

M ... area mass/density ratio (ratio of coa-ting mass to quartz mass per areaunit)

Z ... Z value

It is a fortunate coincidence that the pro-duct M π Z also appears in the Lu-Lewisequation (equation 6.4). It can be used toassess the effective Z value from the follo-wing equations:

(6.6)

or

Here Fq and Fc are the frequencies of thenon-coated or coated quartz in the (100)mode of the fundamental wave. Because ofthe ambiguity of the mathematical func-tions used, the Z value calculated in thisway is not always a positively definedvariable. This has no consequences of anysignificance because M is determined inanother way by assessing Z and measuringthe frequency shift. Therefore, the thickn-ess and rate of the coating are calculatedone after the other from the known M.One must be aware of the limits of thistechnique. Since the assessment of Zdepends on frequency shifts of two mo-

Ztg M Z F

F

tg FF

c

q

c

q

= −

⋅ ⋅ ⋅

⋅

π

π

tg M Z FF

Z tg FF

c

q

c

q⋅ ⋅ ⋅

+ ⋅ ⋅

=π π 0

( )( / )( / )

C CC C M Z

coated

uncoated

55 66

55 66

11

≈+ ⋅

D00 E 19.06.2001 21:40 Uhr Seite 124



des, any minimal shift leads to errors dueto substantial mechanical or thermal stres-ses. It is not necessary to mention thatunder such circumstances the Z matchtechnique, too, leads to similar errors.Nevertheless, the automatic Z value deter-mination of the Z match technique issomewhat more reliable regarding occur-rence of errors because the amplitude dis-tribution of the (102) mode is asymmetricover the active crystal surface and that ofthe (100) mode is symmetric.

According to our experience, coating-rela-ted stresses have the most unfavorableeffect on the crystal. This effect is particu-larly pronounced in the presence of gas,e.g. in sputter processes or reactive vacu-um coating or sputter processes. If the Zvalue for solid material is known, it is bet-ter to use it than to carry out automaticdetermination of the “auto Z ratio”. Incases of parallel coating and coatingsequences, however, automatic Z determi-nation is significantly better.

6.9 Coating thicknessregulation

The last point to be treated here is thetheory of the closed loop for coatingthickness measuring units to effect coatinggrowth at a controlled (constant) growthrate. The measuring advantages of theinstruments, such as speed, precision andreliability, would not be completely exploi-ted if this information were not inputtedinto an improved process monitoringsystem. For a coating process this meansthe coating rate should be kept as closeand stable as possible to a setpoint. Thepurpose of the closed loop is to make useof the information flow of the measuringsystem in order to regulate the capacity fora special evaporation source in an appro-priately adapted way. When the systemfunctions correctly, the controller transla-tes small deviations of the controlled para-meter (the rate) from the setpoint into cor-rection values of the re-adjusted evaporati-on capacity parameter. The ability of thecontroller to measure quickly and precise-ly keeps the process from deviating signi-ficantly from the setpoint.

The most widespread type of controller isthe PID controller. Here P stands for pro-portional, I for integral and D for differen-tial control function. In the following someof the properties of this controller aredescribed in detail. Information on thesystem behavior is gained through a stepresponse to a control fault in certain con-troller settings.

This response is recorded, and thenimproved control parameters for a newtest are estimated. This procedure is con-tinued until a satisfactory result is achie-ved. At the end the controller is optimizedso that its parameters exactly match thecharacteristics of the evaporator source.

It is a long and frustrating process toadjust a controller to an evaporation sour-ce, requiring several minutes for stabiliza-tion and hours to obtain satisfactoryresults. Often the parameters selected fora certain rate are not suitable for an alteredrate. Thus, a controller should ideallyadjust itself, as the new controllers in INFICON coating measuring units do. Atthe beginning of installation and connec-tion the user has the unit measure the cha-racteristics of the evaporation source. Either a PID controller is used as the basisfor slow sources or another type of con-troller for fast sources without significantdead time.

In relevant literature a distinction is madebetween three different ways of settingcontrollers. Depending on which data are

used for the setting, a distinction is madebetween the closed loop, open loop andresonance response method.

Due to the simplicity with which the expe-rimental data can be obtained, we prefer-red the open loop method. Moreover,application of this technique permitsextensive elimination of the trial and errormethod.

The Auto Control Tune function developedby Inficon characterizes a process on thebasis of its step responses. After a step-by-step change in the power the resultingchanges in the rate as a function of time aresmoothed and stored. The important stepresponses are determined, see Fig. 6.8.

In general, it is not possible to characteri-ze all processes exactly, so several appro-ximations have to be made. Normally oneassumes that the dynamic characteristiccan be reproduced by a process of the firstorder plus dead time. The Laplace trans-formation for this assumption (transfer tothe s plane) is approximated:

with (6.8)

Kp = amplification in stationary stateL = dead timeτ = time constant

These three parameters are determinedthrough the response curve of the pro-cess. An attempt has been made by means

OutputInput

Kτ sp

Ls

=⋅

⋅ +

−

101

D00

L0 t Time t(0.632)

point ofmaximumrise

0.0632 Kp

1.00 Kp

T = t – L1 (0.632)

K = (change in output signal)/(change in control signal)p

Fig. 6.8 Process response to a step change with t = 0 (open loop, control signal amplified)

D00 E 19.06.2001 21:40 Uhr Seite 125



of several methods to calculate the requi-red parameters of the system responsefrom curves, as shown in Fig. 6.8. Thisresults in a 1-point accordance at 63.2 %of the transition (a time constant), anexponential accordance at two points andan exponential accordance weightedaccording to the method of the smallestsquares. A process is sufficiently characte-rized by this information so that the con-troller algorithm can be applied. Equation6.9 shows the Laplace transformation forthe very often used PID controller:

(6.9)

with

M(s)= controlled variable or powerKc = Control amplification

(the proportional term)Ti = integration timeTd = differentiation timeE(s) = process deviation

Fig. 6.9 shows the control algorithm and aprocess with a phase shift of the first orderand a dead time. The dynamics of the mea-suring device and the control elements (inour case the evaporator and the powersupply) are implicitly contained in the pro-cess block. R(s) represents the rate set-point. The return mechanism is the devia-tion created between the measured preci-pitation rate C(s) and the rate setpointR(s).

The key to use of any control system is toselect the correct values for Kc, Td and Ti.The “optimum control” is a somewhatsubjective term that is made clear by thepresence of different mathematical defini-tions:Usually the smallest square error ISE(Integral Square Error) is used as a mea-sure of the quality of the control:

(6.10)

Here e is the error (the deviation): e = rate

ISE e t dt= ⋅∫ 2( )

M s K ST

T S E sci

d( ) ( )= + +

1· · ·

setpoint minus measured rate. ISE is rela-tively insensitive to small deviations, butlarge deviations contribute substantially tothe value of the integral. The result is small“overshoots”, but long ripple times becau-se deviations occurring late contribute litt-le to the integral.

The integral of the absolute value of thedeviation IAE (Integral Absolute Error)was also proposed as a measure for con-trol quality:

(6.11)

This is more sensitive for small deviations,but less sensitive for large deviations thanISE.

Graham and Lanthrop introduced the inte-gral over time, multiplied by the absoluteerror ITAE (Integral Time Absolute Error),as a measure for control quality:

(6.12)

The ITAE is sensitive to initial and, to acertain extent, unavoidable deviations.Optimum control responses definedthrough ITAE consequently have shortresponse times and larger “overshoots”than in the case of the other two criteria.However, ITAE has proven to be very use-ful for evaluating the regulation of coatingprocesses.

INFICON’S Auto Control Tune is based onmeasurements of the system responsewith an open loop. The characteristic ofthe system response is calculated on thebasis of a step change in the control sig-nal. It is determined experimentallythrough two kinds of curve accordance attwo points. This can be done either quick-ly with a random rate or more preciselywith a rate close to the desired setpoint.Since the process response depends onthe position of the system (in our case thecoating growth rate), it is best measurednear the desired work point. The processinformation measured in this way (process

ITAE t e t dt= ⋅ ⋅∫ ( )

IAE e t dt= ⋅∫ ( )

amplification Kp, time constant T1 anddead time L) are used to generate the mostappropriate PID control parameters.The best results in evaluating coating con-trol units are achieved with ITAE. There areovershoots, but the reaction is fast and theripple time short. Controller setting condi-tions have been worked up for all integralevaluation criteria just mentioned so as tominimize the related deviations. With amanual input as well as with experimentaldetermination of the process responsecoefficients, the ideal PID coefficients forthe ITAE evaluation can easily be calcula-ted from equations 6.13, 6.14 and 6.15:

(6.13)

(6.14)

(6.15)

For slow systems the time interval bet-ween the forced changes in control voltageis extended to avoid “hanging” the control-ler (hanging = rapid growth of the controlsignal without the system being able torespond to the altered signal). This makesa response to the previous change in thecontroller setting and “powerful” control-ler settings possible. Another advantage isthe greater insensitivity to process noisebecause the data used for control do notcome from merely one measurement, butfrom several, so that the mass-integratingnature of the quartz crystal is utilized.

In processes with short response times(short time constants) and small tounmeasurable dead times, the PID control-ler often has difficulties with the noise ofthe coating process (beam deflection,rapid thermal short-circuits between meltand evaporator, etc.). In these cases a con-trol algorithm of the integral reset type isused with success. This controller alwaysintegrates the deviation and presses thesystem towards zero deviation. This tech-nique works well with small or completelyimperceptible dead times. However, if it isused with a noticeable phase shift or deadtime, the controller tends to generateoscillations because it overcompensatesthe controller signal before the system hasa chance to respond. Auto Control Tune

T T LTd = ⋅ ⋅

( . )

.

0381 11

0995

TT

LTi =

⋅

1191 1

0738.

.

KK

LTc

p=

⋅

–

1361

0947.

.

[process] [controller]

precipitation rateC(s)K · e

T1s + 1p aaa

K (1 + + T * s)c d

sTi

( )Σ+

–

setpointR(s)

deviationE(s)

–Ls

Fig. 6.9 Block diagram of the PID controller

D00 E 19.06.2001 21:40 Uhr Seite 126



recognizes the properties of these fastsystems during the measurement of a stepresponse and utilizes the information tocalculate the control amplification for anon-PID control algorithm.

6.10 INFICONinstrument variants

The instrument models available differboth in hardware and software equipment:the simplest unit, the XTM/2, is purely ameasuring or display device that cannotcontrol vacuum coating.

The XTC/2 and XTC/C group can controlvacuum coating sources and up to threedifferent coatings of a process (not to beconfused with nine different coating pro-grams). In the case of XTM/2, XTC/2 andXTC/C units, the AutoZero and AutoTunefunctions are not available, and measure-ment with several sensors simultaneouslyas well as simultaneous control of twovacuum coating sources are not possible.

However, the IC/5 offers all comfort func-tions available today: measurement withup to eight sensors with AutoZero andAutoTune as well as capability of simulta-neous control of two evaporator sources.Moreover, it offers 24 material programs,with which 250 coatings in 50 processescan be programmed. To simplify operationand avoid errors, the unit also has a dis-kette drive. All types of crystal holders canbe connected here. The thickness resoluti-on is around 1 Å, the rate resolution forrates between 0 and 99.9 Å/s around 0.1Å/s and for rates between 100 and 999 Å/saround 1 Å/s. A particularly attractive opti-on offered by the IC/5 is a microbalanceboard with a highly stable referencequartz. This oscillator is 50 times morestable than the standard oscillator; long-term stability and accuracy are then 2 ppmover the entire temperature range. Thisoption is specially designed for coatings ofmaterial with low density and at low coa-ting rates. This is important for space con-tamination and sorption studies, for exam-ple.

D00

D00 E 19.06.2001 21:40 Uhr Seite 127

Applications of Vacuum TechnologyFundamentals of Vacuum Technology


7. Application ofvacuum technology forcoating techniques

7.1 Vacuum coatingtechnique

Vacuum technology has been increasinglyused in industrial production processesduring the last two decades. Some ofthese processes and their typical workingpressure ranges are shown in Fig. 7.1.

Since a discussion of all processes isbeyond the scope of this brochure, thissection will be restricted to a discussion ofseveral examples of applications in theimportant field of coating technology.

Deposition of thin films is used to changethe surface properties of the base material,the substrate. For example, optical pro-perties such as transmission or reflectionof lenses and other glass products, can beadjusted by applying suitable coating layersystems. Metal coatings on plastic web

produce conductive coatings for film capa-citors. Polymer layers on metals enhancethe corrosion resistance of the substrate.

Through the use of vacuum it is possibleto create coatings with a high degree ofuniform thickness ranging from severalnanometers to more than 100 mm whilestill achieving very good reproducibility ofthe coating properties. Flat substrates,web and strip, as well as complex molded-plastic parts can be coated with virtuallyno restrictions as to the substrate materi-al. For example, metals, alloys, glass, cera-mics, plastics and paper can be coated.The variety of coating materials is alsovery large. In addition to metal and alloycoatings, layers may be produced fromvarious chemical compounds or layers ofdifferent materials applied in sandwichform. A significant advantage of vacuumcoating over other methods is that manyspecial coating properties desired, such asstructure, hardness, electrical conductivityor refractive index, are obtained merely byselecting a specific coating method andthe process paramaters for a certain coa-ting material.

7.2 Coating sourcesIn all vacuum coating methods layers areformed by deposition of material from thegas phase. The coating material may beformed by physical processes such as eva-poration and sputtering, or by chemicalreaction. Therefore, a distinction is madebetween physical and chemical vapordeposition:

• physical vapor deposition = PVD

• chemical vapor deposition = CVD.

7.2.1 Thermal evaporators (boats, wires etc.)

In the evaporation process the material tobe deposited is heated to a temperaturehigh enough to reach a sufficiently highvapor pressure and the desired evaporati-on or condensation rate is set. The sim-plest sources used in evaporation consistof wire filaments, boats of sheet metal orelectrically conductive ceramics that areheated by passing an electrical currentthrough them (Fig. 7.2). However, thereare restrictions regarding the type of mate-rial to be heated. In some cases it is notpossible to achieve the necessary evapora-


10–10 10–7 10–3 100 103

Pressure [mbar]

Annealing of metals

Degassing of melts

Electron beam melting

Electron beam welding

Evaporation

Sputtering of metals

Casting of resins and lacquers

Drying of plastics

Drying of insulating papers

Freeze-drying of bulk goods

Freeze-drying of pharmaceutical products

Fig. 7.1 Pressure ranges for various industrial processes

D00 E 19.06.2001 21:40 Uhr Seite 128

Applications of Vacuum Technology Fundamentals of Vacuum Technology


tor temperatures without significantly eva-porating the source holder and thus con-taminating the coating. Furthermore che-mical reactions between the holder and thematerial to be evaporated can occur resul-ting in either a reduction of the lifetime ofthe evaporator or contamination of thecoating.

7.2.2 Electron beam evaporators (electron guns)

To evaporate coating material using anelectron beam gun, the material, which iskept in a water-cooled crucible, is bombar-ded by a focused electron beam and there-by heated. Since the crucible remains cold,in principle, contamination of the coatingby crucible material is avoided and a highdegree of coating purity is achieved. Withthe focused electron beam, very high tem-peratures of the material to be evaporatedcan be obtained and thus very high evapo-ration rates. Consequently, high-meltingpoint compounds such as oxides can beevaporated in addition to metals andalloys. By changing the power of the elec-tron beam the evaporation rate is easilyand rapidly controlled.

7.2.3 Cathode sputtering

In the cathode sputtering process, the tar-get, a solid, is bombarded with high ener-gy ions in a gas discharge (Fig. 7.3). Theimpinging ions transfer their momentumto the atoms in the target material,knocking them off. These displaced atoms– the sputtered particles – condense onthe substrate facing the target. Comparedto evaporated particles, sputtered particleshave considerably higher kinetic energy.Therefore, the conditions for condensationand layer growth are very different in thetwo processes. Sputtered layers usuallyhave higher adhesive strength and a den-ser coating structure than evaporatedones. Sputter cathodes are available inmany different geometric shapes and sizesas well as electrical circuits configurations.What all sputter cathodes have in commonis a large particle source area compared toevaporators and the capability to coat largesubstrates with a high degree of uniformi-ty. In this type of process metals, alloys ofany composition as well as oxides can beused as coating materials.

7.2.4 Chemical vapor deposition

In contrast to PVD methods, where thesubstance to be deposited is either solid orliquid, in chemical vapor deposition thesubstance is already in the vapor phasewhen admitted to the vacuum system. Todeposit it, the substance must be thermal-ly excited, i.e. by means of appropriatehigh temperatures or with a plasma. Gene-rally, in this type of process, a large num-ber of chemical reactions take place, someof which are taken advantage of to controlthe desired composition and properties ofthe coating. For example, using silicon-hydrogen monomers, soft Si-H polymercoatings, hard silicon coatings or – byaddition of oxygen – quartz coatings canbe created by controlling process parame-ters.

D00

Fig. 7.2 Various thermal evaporators

Hairpin-shaped evaporator made of twisted tungsten wire

Spiral evaporatormade of twisted tung-sten wire

Trough-shapedevaporator

Trough-shapedevaporator withceramic coating

Basket-shapedevaporator

Evaporator made ofelectrically conductiveceramics

Boat-shapedevaporator

Evaporatorwith ceramicinsert

Fig. 7.3 Schematic diagram of a high-performance cathode sputter arrangement

5

4

3

2

1

6

7

8

9

1 Substrates2 Sputtered

atoms 3 Anode4 Electrons 5 Target

6 Cathode 7 Magnetic

field lines 8 Argon ions9 Substrate

D00 E 19.06.2001 21:40 Uhr Seite 129



7.3 Vacuum coatingtechnology/ coating systems

7.3.1 Coating of parts

For molded-plastic parts, vacuum coatingtechniques are increasingly replacing con-ventional coating methods, such as elec-troplating. For example, using vacuumcoating methods, automobile reflectorsobtain a mirror-like surface, plastic articlesin the furniture, decoration, clock andwatch as well as electronics industry aremetal-coated and optical effects are crea-ted on articles in the decoration industry.

Fig. 7.4 shows a type of vacuum system inwhich large batches of molded-plasticparts can be coated simultaneously. Thesubstrates are placed on a cage that rota-tes past the coating source, a sputtercathode in this example. In some applica-tions, by using a glow discharge treat-ment, the substrates are cleaned and thesurface is activated prior to the coatingprocess. This enhances the adhesivestrength and reproducibility of the coatingproperties. A corrosion protection coatingcan be applied after sputtering. In thiscase, a monomer vapor is admitted intothe system and a high-frequency plasmadischarge ignited. The monomer is actived

in the plasma and deposits on the substra-tes as a polymer coating. In this type ofsystem there may be plastic substrateswith a surface area of several 10 m2 on thecage, causing a correspondingly highdesorption gas flow. The vacuum systemmust be able to attain the required pressu-res reliably despite these high gas loads.In the example shown, the system is eva-cuated with a combination of a backingand Roots pump. A diffusion pump alongwith a cold surface forms the high vacuumpump system. The cold surfaces pump alarge portion of the vapor and volatile sub-stances emitted by the plastic parts whilethe diffusion pump basically removes thenon-condensable gases as well as thenoble gas required for the sputter process.A completely different concept for thesame process steps is shown in Fig. 7.5.The system consists of four separate stati-ons made up of a drum rotating around thevertical axis with four substrate chambersand process stations mounted in the vacu-um chamber. During rotation, a substratechamber moves from the loading andunloading station to the pretreatment sta-tion, to the metallization station, to theprotective coating station and then back tothe initial position. Since each station hasits own pumping system, all four proces-ses can run simultaneously with entirelyindependent adjustable process parame-ters. The vacuum system comprises of

turbomolecular pumps and backing pumpsets consisting of Roots and rotary vanepumps.

7.3.2 Web coating

Metal-coated plastic webs and papers playan important role in food packaging. Theypreserve food longer according to storageand transport logistics requirements andgive packaging an attractive appearence.Another important area of application ofmetal-coated web is the production of filmcapacitors for electrical and electronicsapplications. Metal-coating is carried outin vacuum web coating systems. Fig. 7.6shows a typical scheme. The unit consistsof two chambers, the winding chamberwith the roll of web to be coated and thewinding system, as well as the coatingchamber, where the evaporators are loca-ted. The two chambers are sealed fromeach other, except for two slits throughwhich the web runs. This makes it possibleto pump high gas loads from the web rollusing a relatively small pumping set. Thepressure in the winding chamber may bemore than a factor of 100 higher than thepressure simultaneously established in thecoating chamber. The pump set for thewinding chamber usually consists of acombination of Roots and rotary vanepumps.

Fig. 7.4 Diagramm of a batch system for coating partsFig. 7.5 Multi-chamber parts-coating unit (rotationally symmetric in-line system

DynaMet 4V)

1 Vacuum chamber2 High-performance cathode3 Substrate holder4 Substrates5 Diffusion pump6 Roots pump

7 Rotary piston pump8 Cold trap9 High vacuum valve10 Valve for bypass line11 Foreline valve12 Venting valve

D00 E 19.06.2001 21:40 Uhr Seite 130



With strongly degassing rolls of paper, itmay be necessary to install a cold surfacein the winding chamber to act as a watervapor pump. The rolls of the plastic web orpaper typically have diameters between400 and 1000 mm and a width of 400 to3000 mm. A precise, electronically con-trolled winding system is required for win-ding and unwinding as well as web gui-dance.

During the coating process the web, at aspeed of more than 10 m/s, passes a groupof evaporators consisting of ceramic boats,from which aluminium is evaporated. Toachieve the necessary Al-coating thicknessat these high web speeds, very high evapo-ration rates are required. The evaporatorsmust be run at temperatures in excess of1400 °C. The thermal radiation of the eva-porators, together with the heat of conden-sation of the growing layer, yields a consi-derable thermal load for the web. With thehelp of cooled rollers, the foil is cooledduring and after coating so that it is notdamaged during coating and has cooleddown sufficiently prior to winding.

During the entire coating process the coa-ting thickness is continuously monitoredwith an optical measuring system or bymeans of electrical resistance measure-ment devices. The measured values arecompared with the coating thickness set-points in the system and the evaporatorpower is thus automatically controlled.

7.3.3 Optical coatings

Vacuum coatings have a broad range ofapplications in production of ophthalmicoptics, lenses for cameras and other opti-cal instruments as well as a wide variety ofoptical filters and special mirrors. Toobtain the desired transmission or reflec-tion properties, at least three, but someti-mes up to 50 coatings are applied to theglass or plastic substrates. The coatingproperties, such as thickness and refrac-tive index of the individual coatings, mustbe controlled very precisely and matchedto each other. Most of these coatings areproduced using electron beam evapora-tors in single-chamber units (Fig. 7.7). Theevaporators are installed at the bottom ofthe chamber, usually with automaticallyoperated crucibles, in which there areseveral different materials. The substratesare mounted on a rotating calotte abovethe evaporators. Application of suitableshieldings combined with relative move-ment between evaporators and substrates,results in a very high degree of coatinguniformity. With the help of quartz coatingthickness monitors (see Section 6) anddirect measurement of the attained opticalproperties of the coating system duringcoating, the coating process is fully con-trolled automatically.

One of the key requirements of coatings isthat they retain their properties underusual ambient conditions over long

periods of time. This requires to producethe densest coatings possible, into whichneither oxygen nor water can penetrate.Using glass lenses, this is achieved bykeeping the substrates at temperatures upto 300 °C during coating by means ofradiation heaters. However, plastic lenses,as those used in eyeglass optics, are notallowed to be heated above 80 °C. Toobtain dense, stable coatings these sub-strates are bombarded with Ar ions froman ion source during coating. Through theion bombardement the right amount ofenergy is applied to the growing layer sothat the coated particles are arranged onthe energetically most favorable latticesites, without the substrate temperaturereaching unacceptably high values. At thesame time oxygen can be added to theargon. The resulting oxygen ions are veryreactive and ensure that the oxygen isincluded in the growing layer as desired.

The vacuum system of such a coating unitusually consists of a backing pump setcomprising a rotary vane pump and Rootspump as well as a high vacuum pumpingsystem. Depending on the requirements,diffusion pumps, cryo pumps or turbo-molecular pumps are used here, in mostcases in connection with large refrigerator-cooled cold surfaces. The pumps must beinstalled and protected by shieldings in away that no coating material can enter thepumps and the heaters in the system do

D00Fig. 7.6 Schematic diagram of a vacuum web coating system

1 Unwinder2 Coating source3 Coating roller

4 Drawing roller5 Take-up reel

High-perfor-manceplasmasourceElectron

beamevaporator

Monomer

Ar

O2

Fig. 7.7 Coating unit for optical coating systems

D00 E 19.06.2001 21:40 Uhr Seite 131



not thermally overload them. Since shiel-ding always reduces the effective pumpingspeed, the system manufacturer must finda suitable compromise between shieldingeffect and reduction of pumping speed.

7.3.4 Glass coating

Coated glass plays a major role in a num-ber of applications: window panes inmoderate and cold climate zones are pro-vided with heat-reflecting coating systemsto lower heating costs; in countries withhigh intensity solar radiation, solar protec-tion coatings are used that reduce air conditioning costs; coated car windowsreduce the heating-up of the interior andmirrors are used both in the furniture andthe automobile industry. Most of thesecoatings are produced in large in-linevacuum systems. Fig. 7.8 shows a typicalsystem. The individual glass panes aretransported into a entrance chamber atatmospheric pressure. After the entrancevalve is closed, the chamber is evacuatedwith a forepump set. As soon as the pres-sure is low enough, the valve to the eva-cuated transfer chamber can be opened.The glass pane is moved into the transferchamber and from there at constant speedto the process chambers, where coating iscarried out by means of sputter cathodes.On the exit side there is, in analogy to theentrance side, a transfer chamber in whichthe pane is parked until it can be transfer-red out through the exit chamber.

Most of the coatings consist of a stack ofalternative layers of metal and oxide. Sincethe metal layers may not be contaminated

with oxygen, the individual process stati-ons have to be vacuum-isolated from eachother and from the transfer stations. Uti-lization of valves for separating processchambers is unsatisfactory because itincreases plant dimensions. To avoid fre-quent and undesirable starting and stop-ping of the glass panes, the process cham-bers are vacuum-separated through so-called “slit locks”, i.e. constantly open slitscombined with an intermediate chamberwith its own vacuum pump (Fig. 7.9). Thegaps in the slits are kept as small as tech-nically possible to minimize clearance andtherefore conductance as the glass panesare transported through them. The pum-ping speed at the intermediate chamber iskept as high as possible in order to achie-ve a considerably lower pressure in theintermediate chamber than in the processchambers. This lower pressure greatlyreduces the gas flow from a processchamber via the intermediate chamber tothe adjacent process chamber. For verystringent separation requirements it maybe necessary to place several intermediatechambers between two process chambers.

The glass coating process requires highgas flows for the sputter processes as wellas low hydrocarbon concentration. Theonly vacuum pump which satisfies theserequirements as well as high pumpingspeed stability over time are turbo-mole-cular pumps which are used almost exclu-sively.

While the transfer and process chambersare constantly evacuated, the entrance andexit chambers must be periodically ventedand then evacuated again. Due to the largevolumes of these chambers and the shortcycle times, a high pumping speed is

required. It is provided by combinations ofrotary vane pumps and Roots pumps. Forparticularly short cycle times gas cooledRoots pumps are also used.

All major functions of a plant, such asglass transport, control of the sputter pro-cesses and pump control, are carried outfully automatically. This is the only way toensure high productivity along with highproduct quality.

7.3.5 Systems for producingdata storage disks

Coatings for magnetic- or magneto-opticdata storage media usually consist ofseveral functional coatings that are appliedto mechanically finished disks. If severalplates are placed on one common carrier,the coating processes can be carried out ina system using a similar principle to thatused for glass coating. However, mostdisks must be coated on both sides andthere are substantially greater low particlecontamination requirements as comparedto glass coating. Therefore, in-linesystems for data memories use a verticalcarrier that runs through the system (Fig.7.10). The sputter cathodes in the processstations are mounted on both sides of thecarrier so that the front and back side ofthe disk can be coated simultaneously.

An entirely different concept is applied forcoating of single disks. In this case the dif-ferent process stations are arranged in acircle in a vacuum chamber (Fig. 7.11).The disks are transferred individually froma magazine to a star-shaped transport arm.

SZ

L1Z LZ2

Process chamber1

Intermediate chamber

Slits

Process chamber2

S2S1

← →

Fig. 7.9 Principle of chamber separation through pressure stages

L1Z, LZ2 = conductance between intermediate chamber and process chamber 1 or 2

SZ = pumping speed at intermediate chamberS1, S2 = pumping speed at process chamber 1 or 2

Entrance chamber

Sputter chambers

Exit chamberTransfer chamber 1

Transfer chamber 2

to backing pumps

Fig. 7.8 Plant for coating glass panes – 3-chamber in-line system, throughput up to3,600,000 m2 / year

D00 E 19.06.2001 21:40 Uhr Seite 132



The transport arm cycles one station furt-her after each process step and in this waytransports the substrates from one pro-cess station to the next. During cycling allprocesses are switched off and the stati-ons are vacuum-linked to each other. Assoon as the arm has reached the processposition, the individual stations are sepe-

rated from each other by closing seals.Each station is pumped by means of itsown turbomolecular pump and the indivi-dual processes are started. As many pro-cess stations as there are in the system asmany processes can be performed in par-allel. By sealing off the process stations,excellent vacuum separation of the indivi-

dual processes can be achieved. However,since the slowest process step determinesthe cycle interval, two process stationsmay have to be dedicated for particularlytimeconsuming processes.

D00

Fig. 7.10 Plant for coating data storage disks with carrier transport system

Fig. 7.11 Plant for individual coating of data storage disks

D00 E 19.06.2001 21:40 Uhr Seite 133

Instruction for Equipment OperationFundamentals of Vacuum Technology


8. Instructions forvacuum equip-ment operation

8.1 Causes of faultswhere the desiredultimate pressureis not achieved oris achieved tooslowly

If the desired ultimate pressure is not rea-ched at all in vacuum equipment or if it isattained only after an excessively longpumping period, then the following pro-blems may be the reason:If the desired ultimate pressure is not reached,then

• the apparatus may be leaky or dirty,

• the pump may dirty or damaged,

• the vacuum gauge may be defective.

If the desired ultimate pressure is reachedonly after a very long running time, then

• the apparatus may be dirty,

• the pumping line may be restricted,

• the pump may be dirty or of too small acapacity,

• the pumping speed may be restricteddue to other causes.

In order to locate the fault, one normallyproceeds by separating the evacuated ves-sel from the pump system (where this ispossible) and checking the vessel alone forleaks and contamination using the pressu-re rise method, for example. If it has beenfound that the vessel is free of defects inthis regard, then the measurement systemwill be checked for cleanliness (see Sec-tion 8.38) and ultimately – if required – thepump or the pumping system itself will beexamined.

8.2 Contamination ofvacuum vesselsand eliminatingcontamination

In addition to the pressure rise method(Section 6.1) there is a further method fordetecting contamination, based on the factthat condensable vapors will generallyaccount for the major share of the gas mixin dirty vessels: here the pressure readingat a compression vacuum gauge (partialpressure for the non-condensable gases) iscompared with that at an electrical vacuumgauge, e.g. a thermal conductivity orionization vacuum gauge (measuring totalpressure). These two vacuum gaugesmust, however, be clean themselves.Where vapors are present the compressionvacuum gauge will indicate much betterpressure than the electrical vacuum gauge.This is a sure sign that the vessel walls arecontaminated, usually with oil. Anothercommonly used procedure is to comparethe pressure indication of one and thesame vacuum gauge (not a compressionvacuum gauge) with and without a coldtrap inserted in the line: Filling the cold trapwith liquid nitrogen will cause the pressureto drop abruptly, by one power of ten ormore, if the container is contaminatedsince the vapors will freeze out in the trap.

Eliminating contamination for glassequipmentIf the contaminants exhibit a high vaporpressure, then they can be pumped outrelatively quickly. If this is not successful,then the apparatus will have to be cleaned.Contaminated glass components will firstbe cleaned with chromic-sulfuric acid mix-ture or – if this is necessary – with dilutehydrofluoric acid (1:30). They are then rin-sed with distilled water. If this does notbring about the desired results, then anorganic solvent can be tried. Then theglass components will again have to berinsed with methanol and distilled water.(Do not use denatured alcohol!)

Eliminating contamination at metallicequipmentMetal components will usually exhibit traces of contamination by oil and greases.If these cannot be readily removed by pum-ping down the vessel, then an appropriateorganic solvent (denatured alcohol is

unsuitable in all cases) will have to be usedfor cleaning. Maximum cleanliness can beachieved with vapor baths such as thosecommonly found in industry. If one desiresto get down to extremely low pressure ran-ges (< 10-7 mbar), then – after cleaning –the metal components will have to be bakedout at temperatures of up to 200 °C.Seriously contaminated metal componentswill first have to be cleaned by cutting awayor sandblasting the top surface. Thesemethods suffer the disadvantage that thesurface area for the surface thus treatedwill be increased through roughening andactive centers may potentially be formedwhich would readily adsorb vapor molecu-les. Supplementary cleaning in the vaporbath (see above) is advisable. In somecases electrolytic pickling of the surfacemay be beneficial. In the case of high vacu-um components it is necessary to payattention to ensuring that the pickling doesnot turn into etching, which would serious-ly increase the surface area. Polishing sur-faces which have been sandblasted is notnecessary when working in the rough andmedium vacuum ranges since the surfaceplays only a subordinate role in these pres-sure regimes.

8.3 General operatinginformation forvacuum pumps

If no defects are found in the vacuum ves-sels and at the measurement tubes or ifthe apparatus still does not operate satis-factorily after the faults have been rec-tified, then one should first check the flan-ge seals at the pump end of the systemand possibly the shut-off valve. Flangeseals are known to be places at whichleaks can appear the most easily, resultingfrom slight scratches and mechanicaldamage which initially appears insignifi-cant. If no defect can be discerned here,either, then it is advisable to check to seewhether the pumps have been maintainedin accordance with the operating instruc-tions.

Given initially in this section are generalinstructions on pump maintenance, to befollowed in order to avoid such defectsfrom the very outset. In addition, potentialerrors and their causes are discussed.

D00 E 19.06.2001 21:40 Uhr Seite 134

Instruction for Equipment Operation Fundamentals of Vacuum Technology


8.3.1 Oil-sealed rotary vacuumpumps (Rotary vane pumps and rotary piston pumps)

8.3.1.1 Oil consumption, oil contamination, oil change

The oil serves to:• lubricate moving parts,• seal moving parts against atmospheric

pressure,• seal the valve,• fill the dead space below the valve,• to seal the various operational spaces

one from another.

In all pumps it is possible to check the oilcharge during operation using the built-inoil level sight glass. During continuousoperation in particular it is necessary toensure that the oil charge never falls below

the minimum level. During a pumping pro-cess oil-sealed rotary pumps will emit oilvapors from the discharge port, this beingdue to the high operating temperature.This leads to oil loss to an extent whichwill depend on the quantity of gas or vaporwhich is drawn into the pump. Larger oildroplets can be retained by installing acoarse separator in the discharge line. Thiswill reduce oil loss considerably. The fineoil mist filter installed in some pumps willalso retain the finest oil droplets (fineseparation) so that no oil at all will leavethe outlet of the pump and oil loss is redu-ced practically to zero since – as in coarseseparation – the oil which is separated outis returned to the pump. With pumps with-out an integral fine separator this device isoffered as an optional extra.

If an oil-filled rotary pump is operatedwithout an oil separation and return devi-ce, then it will be necessary to expect acertain amount of oil consumption, the

extent of which will depend on the size ofthe pump and the nature of the operations.In the worst case this can amount to about2 cm3 for every cubic meter of air pumped(at STP and including the gas ballast alsodrawn in). Figure 8.1 makes it possible topredict the amount of oil loss to be expec-ted in practical situations. The exampledemonstrates that greater oil losses mustbe expected when operating the pump withgas ballast. This situation, which is gene-rally valid, is always to be taken intoaccount in practice.If the pump oil has become unusabledueto exposure to the vapors or contaminantswhich are encountered in the process,then the oil will have to be replaced. It isimpossible to formulate any hard-and-fastrules as to when an oil change will berequired since the nature of the operationswill determine how long the oil will remaingood. Under clean conditions (e.g. backingpumps for diffusion pumps in electro-nuclear accelerators) rotary vacuumpumps can run for years without an oilchange. Under extremely “dirty” conditi-ons (e.g. during impregnation) it may benecessary to change the oil daily. The oilwill have to be replaced when its originallight brown color, has turned dark brownor black due to ageing or has becomecloudy because liquid (such as water) hasentered the pump. An oil change is alsonecessary when flakes form in corrosionprotection oil, indicating that the corrosionprotection agent is exhausted.

Changing the oilThe oil change should always be carriedout with the pump switched off but at ope-rating temperature. The oil drain (or fill)opening provided for each pump is to beused for this purpose. Where the pump ismore seriously contaminated, then itshould be cleaned. The applicable opera-ting instructions are to be observed in thiscase.

8.3.1.2 Selection of the pump oilwhen handling aggressivevapors

If corrosive vapors (e.g. the vapors formedby acids) are to be pumped, then a PROTELEN® corrosion protection oilshould be used in place of the normalpump oil (N 62). These types of vapors willthen react with the basic (alkaline) corrosi-on protection agent in the oil. The conti-

D00Fig. 8.1 Oil loss for oil-sealed pumps (referenced to an approximate maximum value of 2 cm3 oil loss per cubic meter

of air drawn in [STP])

Example: Oil losses for a TRIVAC S 16 A at pressure of 1 mbar:a) Without gas ballast: in accordance with the pumping speed curve S = 15 m3 / h; according to diagram:

oil loss = 0.03 cm3 / h (line a)b) With gas ballast: S = 9 m3/h at 1 mbar; oil loss = 0.018 cm3 / h (line b1), plus additional loss due to

gas ballast quantity (0.1 times the 1.6 m3 / h rated pumping speed); that is: Chart on the horizontalline b2 up to atmospheric pressure: additional loss: 3 cm3 / h (diagonal line). Overall loss during gasballast operation 0.018 + 3 = 3.018 cm3 / h

Intake pressure [mbar]

Pum

ping

spe

ed [m

3· h

–1]∏

D00 E 19.06.2001 21:40 Uhr Seite 135



nuous neutralization reactions will exhaustthe corrosion protection agent at a ratedepending on the quantity and acidity ofthe vapors. The oil will have to be changedmore frequently, in accordance with thesefactors. Corrosion protection oils are eit-her very hygroscopic or they easily formemulsions with water. Consequently apump which is filled with corrosion pro-tection oil will absorb moisture from theair if it is out of service for an extendedperiod of time. In no case should one everuse a pump filled with corrosion protectionoil in order to pump water vapor since thelubricating and corrosion inhibition pro-perties of the oil would be adversely affec-ted. Once the oil has absorbed water it willno longer be possible for such pumps toachieve the ultimate pressures whichwould be the case with fresh corrosionprotection oil or standard pump oil (N 62).Oil-filled pumps should, under normaloperating con-ditions, not be filled withcorrosion protection oil. N 62 oil is prefer-red when pumping air, water vapor andnon-corrosive organic vapors in so far asthere is positive protection against thevapors condensing inside the pump.

8.3.1.3 Measures when pumpingvarious chemical substances

This discussion cannot provide exhaustivecoverage of the many and varied applicati-on fields for oil-filled vacuum pumps in thechemicals industry. Our many years ofexperience with the most difficult of che-micals applications can be used to solveyour particular problems. Three aspectsshould, however, be mentioned briefly:pumping explosive gas mixes, condensab-le vapors, and corrosive vapors and gases.

Explosion protectionApplicable safety and environmental pro-tection regulations shall be observed whenplanning and engineering vacuumsystems. The operator must be familiarwith the substances which the system willbe pumping and take into account not onlynormal operating conditions but alsoabnormal situations, operating outsidenormal parameters. The most importantaids to avoiding explosive mixtures are –in addition to inertization by adding pro-tective gases – maintaining the explosionlimit values with the aid of condensers,adsorption traps and gas scrubbers.

Protection against condensationLEYBOLD pumps offer three options forkeeping vapors from condensing in thepumps:

• The gas ballast principle (See Fig. 2.14).This increases considerably the amountof vapor which the pump can tolerate.

• Increased pump temperature. The rug-ged design of our pumps makes it pos-sible to run them at temperatures of upto 120 °C. Thus the tolerance for purewater vapor, for example, will rise by afactor of five when compared with nor-mal gas ballast operation.

• Using vacuum condensers (see Section2.15). These act as selective pumps andshould be sized so that the downstreamgas ballast pump will not receive morevapor than the amount correspondingto the appropriate vapor tolerance.

Corrosion protectionOil-sealed pumps are already quite satis-factorily protected against corrosion dueto the oil film which will be present on allthe component surfaces. Corrosion is defi-ned here as the electrochemical dis-solution of metals, i.e. the release of elec-trons by the metal atom and their accep-tance by the oxidation agent (corrosivegas). A metal atom which is susceptible tocorrosion must therefore be exposed to anactive atom of the oxidation agent.

This makes clear how the oil-sealed pumpis protected against corrosion; the concen-tration of the oxidation agent in the oil isnegligible and thus the opportunity for themetal to release electrons is equally small.This also makes it clear that the use of so-called “non-rusting” or “stainless” steelsdoes not make sense since oxidation isnecessary for the passivation of thesesteels, in order to reach the so-called pas-sive region for these steel compounds.The critical passivation current density willnormally not appear in oil-sealed pumps.

a) AcidsOur pumps are fundamentally suited topumping acids. In special situations problems with the oil and with auxiliaryequipment attached at the intake and/ordischarge end may occur. Our engineers inCologne are available to assist in solvingsuch problems.

b) AnhydridesCO (carbon monoxide) is a strong redu-

cing agent. When CO is being pumped it istherefore important that air not be used asthe gas ballast but rather that inert gasesbe used at the very outside (e.g. Ar or N2).Inert gas ballast should also be used whenpumping SO2, SO3, and H2S. A corrosioninhibiting oil is also to be used when hand-ling these three anhydrides. Carbon dioxi-de (CO2) can be pumped without makingany special arrangements.

c) Alkaline solutionsNormal N 62 pump oil is to be used topump basic (alkaline) solutions. Sodiumhydroxide and caustic potash solutionsshould not be pumped in their concen-trated form. Ammonia is highly amenableto pumping with the gas ballast valve clo-sed. Alkaline organic media such asmethylamine and dimethylamine can alsobe pumped satisfactorily, but with the gasballast valve open.

d) Elementary gasesPumping nitrogen and inert gases requiresno special measures.

When handling hydrogen it is necessary tomake note of the hazard of creating anexplosive mixture. The gas ballast valvemay in no case be opened when dealingwith hydrogen. The motors driving thepumps must be of explosion-proof design.

Oxygen: Particular caution is requiredwhen pumping pure oxygen! Speciallyformulated pump oils must be used forthis purpose. We can supply these,accompanied by an approval certificateissued by the German Federal MaterialsTesting Authority (BAM), following consul-tation.

e) AlkanesThe low molecular weight alkanes such asmethane and butane can be pumped withthe gas ballast valve closed or using inertgas as the gas ballast and/or at increasedtemperature of the pump. But important –Increased explosion hazard!

f) AlcoholsOnce operating temperature has been rea-ched, methanol and ethanol can be extrac-ted without using gas ballast (N 62 pumpoil). To pump higher molecular weightalcohols (e.g. butanol) the gas ballastvalve will have to be opened or other pro-tective measures will have to be imple-mented to prevent condensation.

D00 E 19.06.2001 21:40 Uhr Seite 136



g) SolventsAcetone: Can be extracted without difficul-ty; wait until normal operating temperatu-re is reached.

Benzene: Caution – vapors are highlyflammable. Ultimate pressure is degradedby dilution of the pump oil. Mixtures of airand benzene are explosive and flammable.Work without a gas ballast! Inert gasesmay possibly be used as ballast gas.

Carbon tetrachloride and trichlorethy-lene: Amenable to pumping; non-flam-mable and non-explosive but will be dis-solved in oil and thus increase the ultima-te pressure; wait until normal operatingtemperature is achieved. Keep the gas bal-last open when pumping carbon tet andother non-flammable solvents. Use N 62pump oil.

Toluene: Pump through the low-temper-ature baffle and without gas ballast. Useinert gas, not air, as the gas ballast.

8.3.1.4 Operating defects while pumping with gas ballast –Potential sources of errorwhere the required ultimatepressure is not achieved

a) The pump oil is contaminated (parti-cularly with dissolved vapors). Checkthe color and properties. Remedy:Allow the pump to run for an extendedperiod of time with the vacuum vesselisolated and the gas ballast valve open.In case of heavy contamination an oilchange is advisable. The pump shouldnever be allowed to stand for a longerperiod of time when it contains conta-minated oil.

b) The sliding components in the pumpare worn or damaged. Under clean con-ditions our pumps can run for manyyears without any particular mainte-nance effort. Where the pump has beenallowed to run for a longer period oftime with dirty oil, however, the bea-rings and the gate valves may exhibitmechanical damage. This must alwaysbe assumed when the pump no longerachieves the ultimate pressure specifiedin the catalog even though the oil hasbeen changed. In this case the pumpshould be sent in for repair or ourcustomer service department should becontacted.

c) The measurement instrument is soiled(see Section 8.4.2).

Potential sources of error when the pumpno longer turns• Check the pump electrical supply.• The pump has stood for a long time

containing contaminated or resinousoil.

• The pump is colder than 10°C and theoil is stiff. Heat the pump.

• There is a mechanical error. Please con-tact our customer service department.

Oil exits at the shaftIf oil is discharged at the shaft, then theSeeger rotary shaft circlip in the drive bea-ring will have to be inspected and possiblyreplaced. The design of the pumps makesit possible to replace this ring easily, follo-wing the operating instructions providedwith the unit.

8.3.2 Roots pumps

8.3.2.1 General operating instruc-tions, installation and com-missioning

Roots pumps must be exactly level. Whenattaching the pump it is necessary to ensu-re that the pump is not under any tensionor strain whatsoever. Any strains in thepump casing caused by the connectionlines shall be avoided. Any strain to whichthe pump is subjected will endanger thepump since the clearances inside the rootspump are very small.

Roots pumps are connected to the linepower supply via the motor terminal strip;a motor protection switch (overload/overheating) shall be provided as requiredby local codes.

The direction of motor rotation shall bechecked with the intake and outlet portsopen prior to installing the pump. Thedrive shaft, seen from the motor end, mustrotate counter-clockwise. Note the arrowon the motor indicating the direction ofrotation! If the roots pump runs in thewrong direction, then it is reversed byinterchanging two of the phases at themotor connection cord.

The roots pump may be switched on onlyafter the roughing pump has evacuated the

vacuum vessel down to the cut-in pressu-re.

Permissible cut-in pressure PE will dependon the reduction ratios of the roots pumpas against the roughing pump and is cal-culated by dividing the permissible pressu-re differential ∆pmax by the compressionratio, reduced by the value of 1:

If the pump is protected using a dia-phragm-type pressure switch, then thepump will be switched on automatically. Ifa combination of roots pump and roughingpump is to convey highly volatile sub-stances such as liquids with a low boilingpoint, then it is advisable to use a rootspump which is equipped with an integralbypass line and a valve which will respondto a pre-set pressure. Example: Rootsvacuum pumps RUVAC WAU/WSU.

Roots pumps from the RUVAC WAU/WSUseries, being equipped with bypass lines,can generally be switched on together withthe forepump. The bypass protects theseroots pumps against overloading.

8.3.2.2 Oil change, maintenance work

Under clean operating conditions the oil inthe roots pump will be loaded only as aresult of the natural wear in the bearingsand in the gearing. We neverthelessrecommend making the first oil changeafter about 500 hours in service in order toremove any metal particles which mighthave been created by abrasion during therun-in period. Otherwise it will be suffi-cient to change the oil every 3000 hours inoperation. When extracting gases contai-ning dust or where other contaminants arepresent, it will be necessary to change theoil more frequently. If the pumps have torun at high ambient temperatures, then theoil in the sealing ring chamber shall alsobe replaced at each oil change.

We recommend using our specially formu-lated N 62 oil.

Under “dirty” operating conditions it ispossible for dust deposits to form a“crust” in the pump chamber. These conta-minants will deposit in part in the pumping

p whereE

pkth

=–

∆ max1

kthTheoretical pumping speed for the roots pump

Nominal pumping speed for the roughing pump=

D00

D00 E 19.06.2001 21:40 Uhr Seite 137



chamber and in part on the pump’s impel-lers. They may be removed, once the twoconnection lines have been detached, eit-her by blowing out the system with drycompressed air or by rinsing with a suita-ble cleaning agent, such as petroleumether (naphtha).

The oil in the roots pump will then have tobe changed. The rotor may be turned onlyby hand during cleaning since, due to thehigh speed when the motor is running, thedeposits could damage the pump as theydislodge from the surfaces.

Grime which cannot be eliminated bywashing can be removed mechanicallyusing a wire brush, metallic scrubber orscraper.

Important!The dislodged residues may not remain inthe pump chamber. After cleaning is com-pleted check the pump for operability byslowly turning the impellers by hand.There may be no resistance to rotation. Itis generally not necessary to dismantle theroots pump. If this should nevertheless berequired due to heavy soiling, then it ishighly advisable to have this done by themanufacturer.

8.3.2.3 Actions in case of operationaldisturbances

1. Pump becomes too warm: (maximumoperating temperature 100 to 115 °C)

Possible causes:

• Overloading: Excessive heat of com-pression due to an excessively highpressure ratio. Check the pressurevalue adjustments and the tightness ofthe vacuum chamber!

• Incorrect clearances: The distancesbetween the rotors and the housingshave been narrowed due to dirt ormechanical strain.

• Soiled bearings: Excessive friction inthe bearings

• Improper oil level: If the oil level is toohigh, the gears will touch the oil, causing friction resistance. Where theoil level is too low the system will not belubricated.

• Incorrect oil type: An SAE 30 grade oilmust be used for the pump.

2. Excessive power consumption: All thefactors which can lead to elevated tem-peratures can also cause excessisveamounts of power to be drawn. Themotor is defective if excessive powerrequirements are not accompanied by arise in the temperature at the pump.

3. Oiling at the pump chamber:

Possible causes:

• Oil level too high: Oil is subjected toexcessive thermal loading. Oil foam isswept along.

• Oil mixed with the product: Azeotropicdegasification of the oil.

• Pump leaking: Air ingress through theoil drain or filler screw will cause a largestream of air and conveyance of oil intothe pumping chamber.

4. Abnormal running noises:

Possible causes:

• Grime at the impeller• Bearing or gearing damage• Impellers are touching the housing

In the case of bearing or gearing damageor where the impeller scrapes the housingthe pump should be repaired only by themanufacturer.

8.3.3 Turbomolecular pumps

8.3.3.1 General operating instructions

In spite of the relatively large gap betweenthe pump rotor and the stator, the turbo-molecular pumps should be protectedagainst foreign objects entering throughthe intake port. It is for this reason that thepump should never be operated withoutthe supplied wire-mesh splinter guard. Inaddition, sharp shock to the pump whenrunning and sudden changes in attitudeare to be avoided.

Over and above this, and particularly forthe larger types with long rotor blades,airing the pump to atmospheric pressurewhile the impellers are rotating may bedone only when observing exactly therules given in the operating instructions.Under certain circumstances it is possibleto operate turbomolecular pumps underexceptional conditions.

The turbomolecular pump may, for exam-ple, be used unprotected inside magneticfields if the magnetic induction at the sur-face of the pump casing does not exceed B = 3 · 10-3 T when radially impinging andB = 15 · 10-3 T at axial impingement.

In a radioactive environment standard tur-bomolecular pumps can be used withouthazard at dose rates of from 104 to 105

rad. If higher dose rates are encountered,then certain materials in the pump can bemodified in order to withstand the greaterloading. The electronic frequency conver-ters in such cases are to be set up outsidethe radioactive areas since the semicon-ductors used inside them can tolerate adose rate of only about 103 rad. The use ofmotor-driven frequency converters whichcan withstand up to 108 rad representsanother option.

Roughing (backing) pumps are requiredfor the operation of turbomolecularpumps. Depending on the size of the ves-sel to be evacuated, the turbomolecularpumps and forepumps may be switchedon simultaneously. If the time required topump the vessel down to about 1 mbarusing the particular fore-pump is longerthan the run-up time for the pump (seeoperating instructions), then it is advisab-le to delay switching on the turbomolecu-lar pump. Bypass lines are advisable whenusing turbomolecular pumps in systemsset up for batch (cyclical) operations inorder to save the run-up time for thepump. Opening the high vacuum valve isnot dangerous at pressures of about 10-1 mbar.

8.3.3.2 Maintenance

Turbomolecular pumps and frequencyconverters are nearly maintenance-free. Inthe case of oil-lubricated pumps it isnecessary to replace the bearing lubricantat certain intervals (between 1500 and2500 hours in operation, depending on thetype). This is not required in the case ofgrease-lubricated pumps (lifetime lubri-cation). If it should become necessary toclean the pump’s turbine unit, then thiscan easily be done by the customer, obser-ving the procedures described in the ope-rating instructions.

D00 E 19.06.2001 21:40 Uhr Seite 138



8.3.4 Diffusion and vapor-jetvacuum pumps

8.3.4.1 Changing the pump fluid andcleaning the pump

Changing the pump fluid is always ne-cessary whenever the pump no longerachieves the required ultimate vacuum orwhen its pumping speed falls off. The ser-vice life of the pump fluid will as a rulecome to several weeks or months – oreven years – and will depend largely on theoperating conditions for the pump. It isreduced by frequent pumping at high pres-sures, by exposure to aggressive vaporsand by air ingress of longer duration (thisdoes not apply to silicone oil and mer-cury).

In the case of oil diffusion and vapor-jetpumps the danger presented to the pumpfluid where there is air ingress with thepump hot is often overestimated. Whereair ingress (up to atmospheric pressure) isencountered only occasionally and forshort periods of time then silicone oil willnot be attacked at all and the DIFFELENpump fluid will only barely be affected. Theproducts with considerably higher vaporpressures which can be created throughoxidation are removed again in a shortperiod of time by the fractionation anddegassing equipment in the pump (seeSection 2.1.6.1). Even though the pumpfluid which was originally light in color hasbeen turned brown by air ingress, thisneed not necessarily mean that the medi-um has become unusable. If, on the otherhand, the pump fluid has turned cloudyand has become more viscous, as well(which may be the case following periodsof air ingress lasting for several minutes)then the medium will have to be replaced.It is possible that under certain circum-stances cracking products from the pumpfluid may make the oil in the forepumpunserviceable so that here, too, an oilchange will have to be made.

Mercury diffusion and vapor-jet pumps areless sensitive to air ingress than oil diffusi-on pumps. The oxidation of the hot mer-cury caused by the air ingress is negligiblein regard to the operating characteristicsof the pump when compared with the mer-cury loss in the forepump line.

Changing the pump fluid: The interior sec-tion will be extracted from the pump and

the contaminated pump fluid poured out.The interior section and the pump body arethen cleaned with pure petroleum ether(naphtha). The interior section and pumpbody of mercury pumps should have beencleaned beforehand with a clean brush;use a bottle brush for the nozzle bores.Ensure that all the nozzle orifices are pro-perly cleaned. It is advantageous to evapo-rate all solvent residues in a drying kiln.Then the inside section is inserted onceagain and the fresh pump fluid is installedthrough the forevacuum port. It is neces-sary to ensure that the upper nozzle coveris not moistened with pump fluid! Do notinstall too much pump fluid!

8.3.4.2 Operating errors with diffusionand vapor-jet pumps

Potential sources of defects when thedesired ultimate pressure is not reached• Coolant temperature is too high; inade-

quate water throughput. The coolantflow should always be monitored by awater flowmeter in order to protect thepump from damage. Remedy: Measurethe exit temperature of the coolantwater (it should not exceed 30 °C).Increase the coolant flow-through rate.The cooling coils at the pump may haveto be decalcified.

• Forevacuum pressure too high: This ispossible particularly where vaporswhich are either evacuated from thevessel or are created as cracking pro-ducts from the driving medium (e.g.following air ingress) get into the roug-hing pump. Check the forevacuumpressure with the oil diffusion pumpdisconnected. Remedy: Run the fore-vacuum pump for an extended period oftime with gas ballast. If this is not suffi-cient, then the oil in the forepump willhave to be changed.

• Pump fluid at the diffusion pump spentor unserviceable: Replace the drivingmedium.

• Heating is incorrect: Check the heatingoutput and check for good thermalcontact between the heating plate andthe bottom of the boiler section. Repla-ce the heating unit if necessary.

• Leaks, contamination: Remedy: if thepump has been contaminated by vaporsit may help to use a metering valve topass air through the apparatus forsome period of time; here the pressure

should not exceed 10-2 mbar when DIFFELEN is being used.

• Measurement system old or soiled (seeSection 8.4.2).

Potential sources of error where there isinsufficient pumping speed:• Forevacuum pressure is too high:

Check the forevacuum; allow the gasballast pump to run for a longer periodof time with the gas ballast valve open.It may be necessary to change the oil inthe forepump.

• The pump fluid in the diffusion pumphas become unserviceable: Drivingmedium will have to be replaced.

• Nozzles at the diffusion pump are clog-ged: Clean the diffusion pump.

• Heating is too weak: Check heating out-put; examine heating plate for goodthermal contact with the bottom of theboiler chamber.

• Substances are present in the vacuumvessel which have a higher vapor pres-sure than the driving medium beingused: among these are, for example,mercury, which is particularly hazard-ous because the mercury vapors willform amalgams with the nonferrousmetals in the oil diffusion pump andthus make it impossible to achieve per-fect vacuums.

8.3.5 Adsorption pumps

8.3.5.1 Reduction of adsorption capacity

A considerable reduction in pumpingspeed and failure to reach the ultimatepressure which is normally attainable inspite of thermal regeneration having beencarried out indicates that the zeolite beingused has become contaminated by outsidesubstances. It does not make good senseto attempt to rejuvenate the contaminatedzeolite with special thermal processes. Thezeolite should simply be replaced.

8.3.5.2 Changing the molecular sieve

It will be necessary to rinse the adsorptionpump thoroughly with solvents beforeinstalling the new zeolite charge. Beforeputting the adsorption pump which has

D00

D00 E 19.06.2001 21:40 Uhr Seite 139



been charged with fresh zeolite into servi-ce it is also necessary to bake out the newzeolite charge, under vacuum and usingthe heating element associated with thepump, for a period of several hours so thatcontaminants which might have collectedduring the storage period can dissipate.

8.3.6 Titanium sublimationpumps

Each of the turns in the titanium sublimati-on pump contains approximately 1.2 g ofuseable titanium supply. At a heating cur-rent of 50 A the surface temperature comesto about 1850 K, the sublimation rateapproximately 0.12 g/h, i.e. a turn can beoperated continuously for about 10 hours.Since at pressures below 1 · 10-6 mbar sub-limation is not continuous but rather only atintervals which – at low pressures (below 5 · 10-8 mbar) and low gas volumes – arealready more than ten times the actual sub-limation period, one may assume a pum-ping period of almost one month at a wor-king pressure of 10-10 mbar per turn.

The effective pumping speed of a titaniumsublimation pump will depend on the get-ter screen surface and the geometry of thesuction opening. The pumping speed, refe-renced to the surface area of the gettersurface, will be dependent on the type ofgas and the getter screen temperature.Since inert gases, for example, cannot bepumped at all, titanium sublimation pumpsshould always be used only with an auxili-ary pump (sputter-ion pump, turbomole-cular pump) used to pump these gas com-ponents. The supplementary pump can bemuch smaller than the titanium sublimati-on pump. Only in a few special cases canone do without the additional pump.

The selection of the coolant is dictated bythe working conditions and the require-ments in terms of ultimate pressure. Athigh pressures, above 1 · 10-6 mbar, morethermal energy is applied to the getterscreen by frequent sublimation cycles. It isfor this reason that cooling with liquidnitrogen is more favorable. At low pressu-res water cooling may be sufficient. Thegetter screen should if at all possible beheated to room temperature before airingthe pump as otherwise the humidity in theair would condense out on the surface.

8.3.7 Sputter-ion pumps

Sputter-ion pumps use high-voltage cur-rent. Installation and connection should becarried out only under the supervision of aqualified specialist. The operating instruc-tions shall be observed!

The service life of sputter-ion pumpsdepends linearly on the pump’s operatingpressure. In the case of pumps manufac-tured by LEYBOLD, the following applies:

p · T = 45 · 10-3 mbar · h(p = operating pressure, T = service life)

This means that for operating pressure of10-3 mbar

the service life is 45 hours10-6 mbar

the service life is 45,000 hours10-9 mbar

the service life is 45,000,000 hours

If a triode pump is not needed over anextended period of time it can either beoperated continuously at low pressure –with practically no influence on the servicelife – or it can be aired, removed from thepump and packed in a dust-tight container.The starting properties of the sputter-ion(triode) pumps manufactured by LEYBOLD are so good that no problemswill be encountered when returning theunits to service, even after a longer periodin storage.

When the sputter-ion pumps are installedone should ensure that the magnetic fieldswill not interfere with the operation ofother devices (ionization vacuum gauges,partial pressure measurement units, etc.).Mounting devices for the sputter-ionpumps may not short circuit the induc-tance flow and thus weaken the air gapinductance and pumping speed.

If the ultimate pressure which can be attai-ned is not satisfactory even though theapparatus is properly sealed, then it willusually be sufficient to bake out the atta-ched equipment at about 200 to 250 °C. Ifthe pressure here rises to about 1 · 10-5 mbar when this is done, then thesputter ion pump will become so hot afterevacuating the gases for two hours that itwill not be necessary to heat it any furtherin addition. It is also possible to heat thepump by allowing air to enter for 2 hoursat 10-5 mbar before the other apparatus isthen subsequently baked out. If the ultima-te pressure is still not satisfactory, then the

pump itself will have to be baked out for afew hours at 250 to 300 °C (but not higherthan 350 °C!). The pump should withoutfail remain in operation throughout thisperiod! If the pressure rises above 5 · 10-5 mbar it will be necessary either toheat more slowly or to connect an auxiliarypump. Before airing one should allow thehot sputter-ion pump enough time to cooldown at least to 150 °C.

8.4 Information onworking with vacuum gauges

8.4.1 Information on installingvacuum sensors

Here both the external situation in theimmediate vicinity of the vacuum appara-tus and the operating conditions within theapparatus (e.g. working pressure, compo-sition of the gas content) will be important.It is initially necessary to ascertain whetherthe measurement system being installed issensitive in regard to its physical attitude.Sensors should only be installed verticallywith the vacuum flange at the bottom tokeep condensates, metal flakes and filingsfrom collecting in the sensor or even smallcomponents such as tiny screws and thelike from falling into the sensorand themeasurement system. The hot incandes-cent filaments could also bend and deformimproperly and cause electrical shortsinside the measurement system. This isthe reason behind the following generalrule: If at all possible, install sensorsvertically and open to the bottom. It isalso very important to install measurementsystems if at all possible at those points inthe vacuum system which will remain freeof vibration during operation.

The outside temperature must be takeninto account and above all it is necessaryto avoid hot kilns, furnaces or stoves orother sources of intense radiation whichgenerate an ambient temperature aroundthe measurement system which lies abovethe specific acceptable value. Excessiveambient temperatures will result in falsepressure indications in thermal conductivi-ty vacuum sensors.

D00 E 19.06.2001 21:40 Uhr Seite 140



8.4.2 Contamination at themeasurement system andits removal

The vacuum gauges used in vacuum tech-nology for pressure measurement will cer-tainly work under “dirty” conditions. Thisis quite understandable since a vacuumdevice or system does not serve simply toproduce low pressures but rather and pri-marily have to run processes in chemistry,metallurgy or nuclear physics at low pres-sures. Here, depending on the nature ofthe process, considerable quantities ofgases or vapors will be liberated eithercontinuously or intermittently; these canpass into the measurement systems provi-ded for pressure measurement and instal-led in the vacuum system and – due tosurface reactions or through simple depo-sits – can falsify the pressure measure-ments considerably. This is true for alltypes of vacuum gauges whereby, of cour-se, high-sensitivity, high-accuracy measu-rement systems are particularly suscep-tible to soiling resulting from the causesnamed. One can attempt to protect themeasurement systems against contam-ination by providing suitable shielding.This, however, will often lead to the pres-sure registered by the measurementsystem – which is indeed clean – deviatingconsiderably from the pressure actuallyprevailing in the system.

It is not fundamentally possible to keep themeasurement system in a vacuum gaugefrom becoming soiled. Thus it is necessaryto ensure that

• the influence of the contamination onpressure measurement remains assmall as possible and that

• the measurement system can readily becleaned.

These two conditions are not easy to satis-fy by most vacuum gauges in practice.

Dirt in a compression vacuum gauge willcause an incorrect and uncontrollablepressure indication. Dirty THERMOVACsensors will show a pressure which is toohigh in the lower measurement rangesince the surface of the hot wire has chan-ged. In Penning vacuum gauges contami-nation will induce pressure readings whichare far too low since the discharge cur-rents will become smaller. In the case ofionization vacuum gauges with hot catho-

des, electrodes and the tube walls can besoiled which, under certain circumstances,will result in a reduction of dielectricstrengths. Here, however, the measure-ment systems can usually be baked outand degassed by passing a currentthrough or by electron bombardment,quite aside from the fact that ionizationvacuum gauges are often used in the ultra-high vacuum range where it is necessaryto ensure clean operating conditions forother reasons.

8.4.3 The influence of magnetic and electricalfields

In all those measurement instrumentswhich use the ionization of gas moleculesas the measurement principle (cold-catho-de and hot-cathode ionization vacuumgauges), strong magnetic leakage fields orelectrical potentials can have a major influ-ence on the pressure indication. At lowpressures it is also possible for wall poten-tials which deviate from the cathodepotential to influence the ion trap current.

In vacuum measurement systems used inthe high and ultrahigh regimes it is neces-sary to ensure in particular that the requi-red high insulation values for the high-vol-tage electrodes and ion traps also be main-tained during operation and sometimeseven during bake-out procedures. Insulati-on defects may occur both in the externalfeed line and inside the measurementsystem itself. Insufficient insulation at thetrap (detector) lead may allow creep cur-rents – at low pressures – to stimulateoverly high pressure value readings. Thevery low ion trap currents make it neces-sary for this lead to be particularly wellinsulated. Inside the measurement sen-sors, too, creep currents can occur if thetrap is not effectively shielded against theother electrodes.

An error frequently made when connectingmeasurement sensors to the vacuumsystem is the use of connector pipingwhich is unacceptably long and narrow.The conductance value must in all casesbe kept as large as possible. The mostfavorable solution is to use integratedmeasurement systems. Whenever connec-tor lines of lower conductance values are

used the pressure indication, dependingon the cleanliness of the measurementsensors and the connector line, may beeither too high or too low. Here measure-ment errors by more than one completeorder of magnitude are possible! Wheresystems can be baked out it is necessaryto ensure that the connector line can alsobe heated.

8.4.4 Connectors, power pack,measurement systems

The measurement cables (connectorcables between the sensor and the vacuumgauge control unit) are normally 2 m long.If longer measurement cables must beused – for installation in control panels, forexample – then it will be necessary toexamine the situation to determinewhether the pressure reading might be fal-sified. Information on the options for usingover-length cables can be obtained fromour technical consulting department.

D00

D00 E 19.06.2001 21:40 Uhr Seite 141

Tables, Formulas, DiagramsFundamentals of Vacuum Technology


9. Tables, formulas, nomograms and symbols

Unit N · m–2, Pa 2) mbar bar Torr

1 N · m–2 (= 1 Pa) 1 1 · 10–2 1 · 10–5 7.5 · 10–3

1 mbar 100 1 1 · 10–3 0.75

1 bar 1 · 105 1 · 103 1 750

1 Torr 3) 133 1.33 1.33 · 10–3 1

1) The torr is included in the table only to facilitate the transition fromthis familiar unit to the statutory units N · m-2, mbar and bar. In futu-re the pressure units torr, mm water column, mm mercury column(mm Hg), % vacuum, technical atmosphere (at), physicalatmosphe-re (atm), atmosphere absolute (ata), pressure above atmosphericand pressure below atmospheric may no longer be used. Referenceis made to DIN 1314 in this context.

2) The unit Newton divided by square meters (N · m-2) is also designated asPascal (Pa): 1 N · m-2= 1 Pa.Newton divided by square meters or Pascal is the SI unit for the pres-sure of fluids.

3) 1 torr = 4/3 mbar; fl torr = 1 mbar.

Table I: Permissible pressure units including the torr 1) and its conversion

Abbrev. Gas C* = λ · p[cm · mbar]

H2 Hydrogen 12.00 · 10–3

He Helium 18.00 · 10–3

Ne Neon 12.30 · 10–3

Ar Argon 6.40 · 10–3

Kr Krypton 4.80 · 10–3

Xe Xenon 3.60 · 10–3

Hg Mercury 3.05 · 10–3

O2 Oxygen 6.50 · 10–3

N2 Nitrogen 6.10 · 10–3

HCl Hydrochloric acid 4.35 · 10–3

CO2 Carbon dioxide 3.95 · 10–3

H2O Water vapor 3.95 · 10–3

NH3 Ammonia 4.60 · 10–3

C2H5OH Ethanol 2.10 · 10–3

Cl2 Chlorine 3.05 · 10–3

Air Air 6.67 · 10–3

Table III: Mean free path lValues of the product c* of the mean free path λ ( and pressure p for variousgases at 20 °C (see also Fig. 9.1)

1 ↓ = → mbar Pa dyn · cm–2 atm Torr inch Micron cm kp · cm–2 lb · in–2 lb · ft–2

(N/m3) (µbar) (phys.) (mm Hg) Hg (µ) H2O (at tech.) (psi)

mbar 1 102 103 9.87 · 10–4 0.75 2.953 · 10–2 7.5 · 102 1.02 1.02 · 10–3 1.45 · 10–2 2.089

Pa 10–2 1 10 9.87 · 10–6 7.5 · 10–3 2.953 · 10–4 7.5 1.02 · 10–2 1.02 · 10–5 1.45 · 10–4 2.089 · 10–2

µbar 10–3 0.1 1 9.87 · 10–7 7.5 · 10–4 2.953 · 10–5 7.5 · 10–1 1.02 · 10–3 1.02 · 10–6 1.45 · 10–5 2.089 · 10–3

atm 1013 1.01 · 105 1.01 · 106 1 760 29.92 7.6 · 105 1.03 · 103 1.033 14.697 2116.4

Torr 1.33 1.33 · 102 1.33 · 103 1.316 · 10–3 1 3.937 · 10–2 103 1.3595 1.36 · 10–3 1.934 · 10–2 2.7847

in Hg 33.86 33.9 · 102 33.9 · 103 3.342 · 10–2 25.4 1 2.54 · 104 34.53 3.453 · 10–2 0.48115 70.731

µ 1.33 · 10–3 1.33 · 10–1 1.333 1.316 · 10–6 10–3 3.937 · 10–5 1 1.36 · 10–3 1.36 · 10–6 1.934 · 10–5 2.785 · 10–3

cm H2O 0.9807 98.07 980.7 9.678 · 10–4 0.7356 2.896 · 10–2 7.36 · 102 1 10–3 1.422 · 10–2 2.0483

at 9.81 · 102 9.81 · 104 9.81 · 105 0.968 7.36 · 102 28.96 7.36 · 105 103 1 14.22 2048.3

psi 68.95 68.95 · 102 68.95 · 103 6.804 · 10–2 51.71 2.036 51.71 · 103 70.31 7.03 · 10–2 1 1.44 · 102

lb · ft–2 0.4788 47.88 478.8 4.725 · 10–4 0. 3591 1.414 · 10–2 359.1 0.488 4.88 · 10–4 6.94 · 10–3 1

Normal conditions: 0 °C and sea level, i.e. p = 1013 mbar = 760 mm Hg = 760 torr = 1 atmin Hg = inches of mercury; 1 mtorr (millitorr) = 10-3 torr = 1 µ (micron … µm Hg column)Pounds per square inch = lb · in-2 = lb / sqin = psi (psig = psi gauge … pressure above atmospheric, pressure gauge reading; psia = psi absolute … absolute pressure)Pounds per square foot = lb / sqft = lb / ft2; kgf/sqcm2 = kg force per square cm = kp / cm2 = at; analogously also: lbf / squin = psi1 dyn · cm-2 (cgs) = 1 µbar (microbar) = 1 barye; 1 bar = 0.1 Mpa; 1 cm water column (cm water column = g / cm2 at 4 °C) = 1 Ger (Geryk)atm … physical atmosphere – at … technical atmosphere; 100 - (x mbar / 10.13) = y % vacuum

Table II: Conversion of pressure units

D00 E 19.06.2001 21:40 Uhr Seite 142

Tables, Formulas, Diagrams Fundamentals of Vacuum Technology


D00

Table IV: Compilation of important formulas pertaining to the kinetic theory of gases

2 ⋅ R ⋅ TABBBMB

VARIABLE

Most probable speedof particles cw

Mean velocityof particles c–

Mean square of velocityof particles c2–

Gas pressure p of particles

Number density of particles n

Area-related impingement ZA

Volume collision rate ZV

Equation of state of ideal gas

Area-related mass flow rate qm, A

c* = λ · p in cm · mbar (see Tab. III)k Boltzmann constant in mbar · l · K–1

λ mean free path in cmM molar mass in g · mol–1

General formula

p = n ⋅ k ⋅ T1p = ⋅ n ⋅ mT ⋅ c2–31p = ⋅ % ⋅ c2–3

1ZA = ⋅p2c*

p = 13.80 ⋅ 10–20 ⋅ n ⋅ T [mbar] p = 4.04 ⋅ 10–17 ⋅ n [mbar] (applies to all gases)

p = 2.5 ⋅ 1016 ⋅ p [cm–3] (applies to all gases)

ZA = 2.85 ⋅ 1020 ⋅ p [cm–2 s–1] (see Fig. 78.2)

ZV = 8.6 ⋅ 1022 ⋅ p2 [cm–3 s–1] (see Fig. 78.2)

p ⋅ V = 2.44 ⋅ 104 ν [mbar ⋅ ` ] (for all gases)p ⋅ V = 83.14 ⋅ ν ⋅ T [mbar ⋅ ` ]

Qm, A = 4.377 ⋅ 10–2 ⋅ p [g cm–2 s–1]

p ⋅ V = ν ⋅ R ⋅ T

qm, A = ZA ⋅ mT = ⋅ p

n = p/kT

qm, A = 1.38 ⋅ 10–2 ⋅ p g [cm–2 s–1]

For easy calculation Value for air at 20 °C

cw = cw = 410 [m/s]

c– =

c2– = c2– = 25.16 ⋅ 104

c– = 464 [m/s]

cm2

s23 ⋅ R ⋅ T

M

MABTMABBBB2 ⋅ π⋅ k ⋅ T ⋅ NA

NAABBBB2 ⋅ π ⋅ M ⋅ k ⋅ T

2 ⋅ NAABBBBπ⋅ M ⋅ k ⋅ T

pn = 7.25 ⋅ 1018 ⋅ [cm–3]T

pZA = 2.63 · 1022 · · p [cm–2 s–1]ABBM · TBB

p2ZV = 5.27 · 1022 · [cm–3 s–1]

c* · ABBBBMBB· T

1ZA = · n · c–4

ZA = · p

[ ]

8 ⋅ R ⋅ TABBBπ⋅ MB

1ZV = ⋅ n ⋅ c–

2 λ

cw = 1.29 ⋅ 104 T cmABM s[ ]

c2– = 2.49 ⋅ 108 T cm2

M s2[ ]

T cmABM sc– = 1.46 ⋅ 104 [ ]

mT particle mass in gNA Avogadro constant in mol–1

n number density of particles in cm–3

υ amount of substance in mol

p gas pressure in mbarR molar gas constant in mbar · l · mol–1 K–1

T thermodynamic temperature in KV volume in l

Table V: Important values

Designation, Symbol Value and Remarksalphabetically unit

Atomic mass unit mu 1.6605 · 10–27 kgAvogadro constant NA 6.0225 · 1023 mol–1 Number of particles per mol,

formerly: Loschmidt numberBoltzmann constant k 1.3805 · 10–23 J · K–1

mbar · l13.805 · 10–23K

Electron rest mass me 9.1091 · 10–31 kgElementary charge e 1.6021 · 10–19 A · sMolar gas constant R 8.314 J · mol–1 K–1

mbar · l= 83.14 R = NA · kmol · KMolar volume of 22.414 m3 kmol–1 DIN 1343; formerly: molar volumethe ideal gas Vo 22.414 l · mol–1 at 0 °C and 1013 mbarStandard acceleration of free fall gn 9.8066 m · s–2

Planck constant h 6.6256 · 10–34 J · sW Stefan-Boltzmann constant s 5.669 · 10–8 also: unit conductance, radiation constant

m2 K4

– e A · sSpecific electron charge – 1.7588 · 1011

me kg Speed of light in vacuum c 2.9979 · 108 m · s–1

Standard reference density ln kg · m–3 Density at ϑ = 0 °C and pn = 1013 mbarof a gasStandard reference pressure pn 101.325 Pa = 1013 mbar DIN 1343 (Nov. 75)Standard reference temperature Tn Tn = 273.15 K, J = 0 °C DIN 1343 (Nov. 75)

D00 E 19.06.2001 21:40 Uhr Seite 143



Unit l · s–1 m3 · h–1 cm3 · s–1 cuft · min–1

1 l · s–1 1 3.6 1000 2.121 m3 · h–1 0.2778 1 277.8 0.5891 cm3 · s–1 10–3 3.6 · 10–3 1 2.1 · 10–3

1 cuft · min–1 0.4719 1.699 471.95 1Table VI: Conversion of pumping speed (volume flow rate) units

Table VIIa: Conversion of throughput (QpV) units; (leak rate) units

1↓ = ... → mbar · l/s kg · h–1 (20°C) kg · h–1 (0°C) cm3/h (NTP) cm3/s (NTP) Torr · l/s g/a (F12. 20 °C) g/a (F12. 25 °C) µ · cfm lusec Pa · l/s slpmmbar · l/s 1 4.28 · 10–3 4.59 · 10–3 3554 0.987 0.75 1.56 · 105 1.54 · 105 1593 7.52 · 102 100 59.2 · 10–3

kg · h–1 (20 °C) 234 1 1.073 8.31 · 105 231 175 – – 37.2 · 104 1.75 · 105 23.4 · 103 13.86kg · h–1 (0 °C) 218 0.932 1 7.74 · 105 215 163 – – 34.6 · 104 1.63 · 105 21.8 · 103 12.91cm3/h (NTP) 2.81 · 10–4 1.20 · 10–6 1.29 · 10–6 1 2.78 · 10–4 2.11 · 10–4 44 – 44.7 · 10–2 2.11 · 10–1 2.81 · 10–2 1.66 · 10–5

cm3/s (NTP) 1.013 4.33 · 10–3 4.65 · 10–3 3600 1 0.760 1.58 · 105 – 1611 760 101 6 · 10–2

Torr · l/s 1.33 5.70 · 10–3 6.12 · 10–3 4727 1.32 1 2.08 · 105 2.05 · 105 2119 1 · 103 133 78.8 · 10–3

g/a (F12. 20 °C) 6.39 · 10–6 – – 2.27 · 10–2 6.31 · 10–6 4.80 · 10–6 1 – 10.2 · 10–3 4.8 · 10–3 6.39 · 10–4 37.9 · 10–8

g/a (F12. 25 °C) 6.50 · 10–6 – – – – 4.88 · 10–6 – 1 10.4 · 10–3 4.89 · 10–3 6.5 · 10–4 38.5 · 10–8

µ · cfm 6.28 · 10–4 2.69 · 10–6 2.89 · 10–6 2.24 6.21 · 10–4 4.72 · 10–4 98.16 96.58 1 0.472 6.28 · 10–2 37.2 · 10–6

lusec 1.33 · 10–3 5.70 · 10–6 6.12 · 10–6 4.737 1.32 · 10–3 1 · 10–3 208 205 2.12 1 13.3 · 10–2 78.8 · 10–6

Pa · l/s 1 · 10–2 4.28 · 10–5 4.59 · 10–5 35.54 9.87 · 10–3 7.5 · 10–3 1.56 · 103 1.54 · 103 15.93 7.50 1 59.2 · 10–5

slpm 16.88 72.15 · 10–3 77.45 · 10–3 60.08 · 103 16.67 12.69 2.64 · 106 2.60 · 106 26.9 · 103 12.7 · 103 16.9 · 102 1

1 cm3 (NTP) = 1 cm3 under normal conditions (T = 273.15 K; p = 1013.25 mbar) NTP = normal temperature and pressure (1 atm; 0 °C) R = 83.14 mbar · l · mol–1 · K–1

1 cm3 (NTP) · h–1 = 1 atm · cm3 · h–1 = 1 Ncm3 · h–1 = 1 std cch 1 sccm = 10–3 slpm = 10–3 N · l · min–1 = 60 sccsSI coherent: 1 Pa · m3 · s–1 = 10 mbar · l · s–1; R = 8.314 Pa · m3 · mol–1 · K–1; M in kg / mol1 cm3 (NTP) · s–1 = 1 sccs = 60 cm3 (NTP) · min–1 60 sccm = 60 std ccm = 60 Ncm3 · min–1

1 lusec = 1 l · µ · s–1 1 · µ = 1 micron = 10–3 Torr 1 lusec = 10–3 Torr · l · s–1

Freon F 12 (CCl2F2) M = 120.92 g · mol–1; air M = 28.96 g · mol–1

Note: Anglo-American units are not abbreviated nonuniformly! Example: Standard cubic centimeter per minute → sccm = sccpm = std ccm = std ccpm

1↓= ... → mbar · l/s cm3/s **) Torr · l/s Pa · m3/s g/a *) oz/yr *) lb/yr *) atm · ft3/min µ · l/s µ · ft3/h µ · ft3/minmbar · l/s 1 0.987 0.75 10–1 1.56 · 105 5.5 · 103 3.4 · 102 2.10 · 10–3 7.52 · 102 9.56 · 104 1593cm3/s **) 1.013 1 0.76 1.01 · 10–1 1.58 · 105 5.6 · 103 3.44 · 102 2.12 · 10–3 760 96.6 · 103 1614Torr · l/s 1.33 1.32 1 1.33 · 10–1 2.08 · 105 7.3 · 103 4.52 · 102 2.79 · 10–3 103 1.27 · 105 2119Pa · m3/s 10 9.9 7.5 1 1.56 · 106 5.51 · 104 3.4 · 103 2.09 · 10–2 7.5 · 103 9.54 · 105 15.9 · 103

g/a *) 6.39 · 10–6 6.31 · 10–6 4.80 · 10–6 6.41 · 10–7 1 3.5 · 10–2 2.17 · 10–3 1.34 · 10–8 4.8 · 10–3 0.612 10.2 · 10–3

oz/yr *) 1.82 · 10–4 1.79 · 10–4 1.36 · 10–4 1.82 · 10–5 28.33 1 6.18 ·· 10–2 3.80 · 10–7 0.136 17.34 0.289lb/yr *) 2.94 · 10–3 2.86 · 10–3 2.17 · 10–3 2.94 · 10–4 4.57 · 102 16 1 6.17 · 10–6 2.18 280 4.68

atm · ft3/min 4.77 · 102 4.72 · 102 3.58 · 102 47.7 7.46 · 107 2.63 · 106 1.62 · 105 1 3.58 · 105 4.55 · 107 7.60 · 105

µ · l/s 1.33 · 10–3 1.32 · 10–3 10–3 1.33 · 10–4 208 7.34 4.52 · 10–1 2.79 · 10–6 1 127 2.12µ · ft3/h 1.05 · 10–5 1.04 · 10–5 7.87 · 10–6 1.05 · 10–6 1.63 5.77 · 10–2 3.57 · 10–3 2.20 · 10–8 7.86 · 10–3 1 1.67 · 10–2

µ · ft3/min 6.28 · 10–4 6.20 · 10–4 4.72 · 10–4 6.28 · 10–5 98 3.46 2.14 · 10–1 1.32 · 10–6 0.472 60 1

1 · µ · ft3 · h–1 = 1.04 · 10–5 stsd cc per second 1 micron cubic foot per hour = 0.0079 micron liter per second 1 kg = 2.2046 pounds (lb)1 cm3 · s–1 (NTP) = 1 atm · cm3 · s–1 = 1 scc · s–1 = 1 sccss 1 micron liter per second = 0.0013 std cc per second = 1 lusec 1 cubic foot (cfut. cf) =

28.3168 dm3

1 atm · ft3 · min–1 = 1 cfm (NTP) 1 micron cubic foot per minute = 1 µ · ft3 · min–1 = 1 µ · cuft · min–1 = 1 µ · cfm1 lb = 16 ounces (oz)

1 Pa · m3/s = 1 Pa · m3/s (anglo-am.) = 103 Pa · l/s 1 standard cc per second = 96.600 micron cubic feet per hour 1 lusec = 1 µ · l · s–1

1 µ · l · s–1 = 127 µ · ft3 · h–1 = 0.0013 std cc per second = 1 lusec 1 std cc/sec = 760 µ · l · s–1

*) F12 (20 °C) C.Cl2F2 M = 120.92 h/mol **) (NTP) normal temperature and pressure 1 atm und 0 °C

Table VII b: Conversion of throughput (QpV) units; (leak rate) units

D00 E 19.06.2001 21:40 Uhr Seite 144



D00

% by weight % by volume Partial pressure mbarN2 75.51 78.1 792O2 23.01 20.93 212Ar 1.29 0.93 9.47CO2 0.04 0.03 0.31Ne 1.2 · 10–3 1.8 · 10–3 1.9 · 10–2

He 7 · 10–5 7 · 10–5 5.3 · 10–3

CH4 2 · 10–4 2 · 10–4 2 · 10–3

Kr 3 · 10–4 1.1 · 10–4 1.1 · 10–3

N2O 6 · 10–5 5 · 10–5 5 · 10–4

H2 5 · 10–6 5 · 10–5 5 · 10–4

Xe 4 · 10–5 8.7 · 10–6 9 · 10–5

O3 9 · 10–6 7 · 10–6 7 · 10–5

Σ 100 % Σ 100 % Σ 101350 % RH at 20 °C 1.6 1.15 11.7

Note: In the composition of atmospheric air the relative humidity (RH) is indicated separately alongwith the temperature.At the given relative humidity, therefore, the air pressure read on the barometer is 1024 mbar

Table VIII: Composition of atmospheric air

Rough vacuum Medium vacuum High vacuum Ultrahigh vacuum

Pressure p [mbar] 1013 – 1 1 – 10–3 10–3 – 10–7 < 10–7

Particle number density n [cm–3] 1019 – 1016 1016 – 1013 1013 – 109 < 109

Mean free path λ [cm] < 10–2 10–2 – 10 10 – 105 > 105

Impingement rate Za [cm–2 · s–1] 1023 – 1020 1020 – 1017 1017 – 1013 < 1013

Vol.-related collision rate ZV [cm–3 · s–1] 1029 – 1023 1023 – 1017 1017 – 109 < 109

Monolayer time τ [s] < 10–5 10–5 – 10–2 10–2 – 100 > 100Type of gas flow Viscous flow Knudsen flow Molecular flow Molecular flow

Other special features Convection dependent Significant change in Significant reduction- Particles on the on pressure thermal conductivity in volume surfaces dominate

of a gas related collision rate to a great extend in relation to particles in gaseous space

Table IX: Pressure ranges used in vacuum technology and their characteristics (numbers rounded off to whole power of ten)

At room temperature

Standard values1 Metals Nonmetals(mbar · l · s–1 · cm–2) 10–9 ... · 10–7 10–7 ... · 10–5

Outgassing rates (standard values) as a function of time

Examples: 1/2 hr. 1 hr. 3 hr. 5 hr. Examples: 1/2 hr. 1 hr. 3 hr. 5 hr.Ag 1.5 · 10–8 1.1 · 10–8 2 · 10–9 Silicone 1.5 · 10–5 8 · 10–6 3.5 · 10–6 1.5 · 10–6

Al 2 · 10–8 6 · 10–9 NBR 4 · 10–6 3 · 10–6 1.5 · 10–6 1 · 10–6

Cu 4 · 10–8 2 · 10–8 6 · 10–9 3.5 · 10–9 Acrylic glass 1.5 · 10–6 1.2 · 10–6 8 · 10–7 5 · 10–7

Stainless steel 9 · 10–8 3.5 · 10–8 2.5 · 10–8 FPM, FKM 7 · 10–7 4 · 10–7 2 · 10–7 1.5 · 10–7

1 All values depend largely on pretreatment!

Table X: Outgassing rate of materials in mbar · l · s–1 · cm–2

D00 E 19.06.2001 21:40 Uhr Seite 145



Nominal width (DN) Internal diam. (mm)

Series R5 R10

10 1016 16

20 2125 24

32 3440 41

50 5163 70

80 83100 102

125 127160 153

200 213250 261

320 318400 400

500 501630 651

800 8001000 1000

The nominal internal diameters correspond approximately to theinternal diameters of the pipeline components” (DIN 2402 - Feb.1976). The left-hand column of the nominal internal diameter seriesis preferred in practice.

Table XI: Nominal internal diameters (DN) and internal diameters of tubes, pipes andapertures with circular cross-section (according to PNEUROP)

Solvent Relative Density Melting Boiling Maximum admissiblemolecular g / cm3 point point concentration (MAC)mass (20 °C) °C °C cm3 / m3

Acetone 58 0.798 56Benzene (solution) 78 0.8788 5.49 80.2 25Petrol (light) 0.68 ... 0.72 > 100Carbon tetrachloride 153.8 1.592 – 22.9 76.7 25Chloroform 119.4 1.48 – 63.5 61 50Diethyl ether 46 0.7967 –114.5 78 1000Ethyl alcohol 74 0.713 – 116.4 34.6 400Hexane 86 0.66 – 93.5 71 500Isopropanol 60.1 0.785 – 89.5 82.4 400Methanol 32 0.795 – 97.9 64.7 200 (toxic!)Methylene chloride 85 1.328 41Nitromethane 61 1.138 – 29.2 101.75 100Petroleum ether mixture 0.64 – 40 ... 60Trichlorethylene (“Tri”) 131.4 1.47 55Water 18.02 0.998 0.00 100.0 –

Table XII: Important data (characteristic figures) for common solvents

D00 E 19.06.2001 21:40 Uhr Seite 146



D00

t ps %D°C mbar g/m3

– 100 1.403 · 10–5 1.756 · 10–5

– 99 1.719 2.139– 98 2.101 2.599– 97 2.561 3.150– 96 3.117 3.812– 95 3.784 · 10–5 4.602 · 10–5

– 94 4.584 5.544– 93 5.542 6.665– 92 6.685 7.996– 91 8.049 9.574– 90 9.672 · 10–5 11.44 · 10–5

– 89 11.60 13.65– 88 13.88 16.24– 87 16.58 19.30– 86 19.77 22.89– 85 23.53 · 10–5 27.10 · 10–5

– 84 27.96 32.03– 83 33.16 37.78– 82 39.25 44.49– 81 46.38 52.30– 80 0.5473 · 10–3 0.6138 · 10–3

– 79 0.6444 0.7191– 78 0.7577 0.8413– 77 0.8894 0.9824– 76 1.042 1.145– 75 1.220 · 10–3 1.334 · 10–3

– 74 1.425 1.550– 73 1.662 1.799– 72 1.936 2.085– 71 2.252 2.414– 70 2.615 · 10–3 2.789 · 10–3

– 69 3.032 3.218– 68 3.511 3.708– 67 4.060 4.267– 66 4.688 4.903– 65 5.406 · 10–3 5.627 · 10–3

– 64 6.225 6.449– 63 7.159 7.381– 62 8.223 8.438– 61 9.432 9.633– 60 10.80 · 10–3 10.98 · 10–3

– 59 12.36 12.51– 58 14.13 14.23– 57 16.12 16.16– 56 18.38 18.34– 55 20.92 · 10–3 20.78 · 10–3

– 54 23.80 23.53– 53 27.03 26.60– 52 30.67 30.05– 51 34.76 33.90– 50 39.35 · 10–3 38.21 · 10–3

– 49 44.49 43.01– 48 50.26 48.37– 47 56.71 54.33– 46 63.93 60.98– 45 71.98 · 10–3 68.36 · 10–3

– 44 80.97 76.56– 43 90.98 85.65– 42 102.1 95.70– 41 114.5 · 10–3 106.9 · 10–3

– 40 0.1283 0.1192– 39 0.1436 0.1329– 38 0.1606 0.1480– 37 0.1794 0.1646– 36 0.2002 0.1829


– 35 0.2233 0.2032– 34 0.2488 0.2254– 33 0.2769 0.2498– 32 0.3079 0.2767– 31 0.3421 0.3061– 30 0.3798 0.3385– 29 0.4213 0.3739– 28 0.4669 0.4127– 27 0.5170 0.4551– 26 0.5720 0.5015– 25 0.6323 0.5521– 24 0.6985 0.6075– 23 0.7709 0.6678– 22 0.8502 0.7336– 21 0.9370 0.8053– 20 1.032 0.8835– 19 1.135 0.9678– 18 1.248 1.060– 17 1.371 1.160– 16 1.506 1.269– 15 1.652 1.387– 14 1.811 1.515– 13 1.984 1.653– 12 2.172 1.803– 11 2.376 1.964– 10 2.597 2.139– 9 2.837 2.328– 8 3.097 2.532– 7 3.379 2.752– 6 3.685 2.990– 5 4.015 3.246– 4 4.372 3.521– 3 4.757 3.817– 2 5.173 4.136– 1 5.623 4.479

0 6.108 4.8471 6.566 5.1922 7.055 5.5593 7.575 5.9474 8.129 6.3605 8.719 6.7976 9.347 7.2607 10.01 7.7508 10.72 8.2709 11.47 8.819

10 12.27 9.39911 13.12 10.0112 14.02 10.6613 14.97 11.3514 15.98 12.0715 17.04 12.8316 18.17 13.6317 19.37 14.4818 20.63 15.3719 21.96 16.3120 23.37 17.3021 24.86 18.3422 26.43 19.4323 28.09 20.5824 29.83 21.7825 31.67 23.0526 33.61 24.3827 35.65 25.78

28 37.80 27.2429 40.06 28.78


30 42.43 30.3831 44.93 32.0732 47.55 33.8333 50.31 35.6834 53.20 37.6135 56.24 39.6336 59.42 41.7537 62.76 43.9638 66.26 46.2639 69.93 48.6740 73.78 51.1941 77.80 53.8242 82.02 56.5643 86.42 59.4144 91.03 62.3945 95.86 65.5046 100.9 68.7347 106.2 72.1048 111.7 75.6149 117.4 79.2650 123.4 83.0651 129.7 87.0152 136.2 91.1253 143.0 95.3954 150.1 99.8355 157.5 104.456 165.2 109.257 173.2 114.258 181.5 119.459 190.2 124.760 199.2 130.261 208.6 135.962 218.4 141.963 228.5 148.164 293.1 154.565 250.1 161.266 261.5 168.167 273.3 175.268 285.6 182.669 298.4 190.270 311.6 198.171 325.3 206.372 339.6 214.773 354.3 223.574 369.6 232.575 385.5 241.876 401.9 251.577 418.9 261.478 436.5 271.779 454.7 282.380 473.6 293.381 493.1 304.682 513.3 316.383 534.2 328.384 555.7 340.785 578.0 353.586 601.0 366.687 624.9 380.288 649.5 394.289 674.9 408.690 701.1 423.591 728.2 438.892 756.1 454.593 784.9 470.794 814.6 487.4


95 845.3 504.596 876.9 522.197 909.4 540.398 943.0 558.999 977.6 578.1100 1013.2 597.8101 1050 618.0102 1088 638.8103 1127 660.2104 1167 682.2105 1208 704.7106 1250 727.8107 1294 751.6108 1339 776.0109 1385 801.0110 1433 826.7111 1481 853.0112 1532 880.0113 1583 907.7114 1636 936.1115 1691 965.2116 1746 995.0117 1804 1026118 1863 1057119 1923 1089120 1985 1122121 2049 1156122 2114 1190123 2182 1225124 2250 1262125 2321 1299126 2393 1337127 2467 1375128 2543 1415129 2621 1456130 2701 1497131 2783 1540132 2867 1583133 2953 1627134 3041 1673135 3131 1719136 3223 1767137 3317 1815138 3414 1865139 3512 1915140 3614 1967

Sources: Smithsonian Meteorological Tables 6th. ed. (1971) and VDI vapor tables 6th ed (1963).

Table XIII: Saturation pressure ps and vapor density %D of water in a temperature range from –100 °C to +140 °C

D00 E 19.06.2001 21:40 Uhr Seite 147



Group A 3)

Methane c

Ethane c

Propane c

Butane c

Pentane c

Hexane c

Heptane c

Octane a

Cyclohexane c

Propylene (propene) a

Styrene (s) b

Benzene (s) c

Toluene (s) –

Xylene a

Naphthalene –

Methanol (s) c

Ethanol (s) c

Propyl alcohol (propanol) c

Butyl alcohol (butanol) a

Phenol –

Acetaldehyde (ethanal) a

Acetone (s) (propanone) c

Methyl ethyl ketone (s) (propan-2-one) c

Ethyl acetate (s) a

Butyl acetate (s) c

Amyl acetate (s) –

Ethyl methacrylate –

Acetic acid (ethanoic acid) b

Methyl chloride (s) a

Methylene chloride (s) (dichlormethane) –

Ammonia a

acetonitrile a

Aniline –

Pyridine –

Group B 3)

Ethylene c

Buta-1,3-diene c

Acrylonitrile c

hydrogen cyanide a

Dethyl ether (s) c

Ethylene oxide (oxiran) c

1.4 Dioxan a

Tetrahydrofuran a

Tetrafluoroethylene a

Group C 3)

Hydrogen c

Acetylene (ethyne) c

Carbon bisulfide c

1) Minimum Electrical Spark Gap2) Minimum Ignition Current

The ratio is based on the MIC value for laboratory methane3) Group allocation:

a – according to MESG valueb – according to MIC ratioc – according to both MESG value and MIC ratio(s) – solvent

Table XIV: Hazard classification of fluids according to their MESG1 and/or MIC2 values. (Extract from European Standard EN 50.014)

Legend

MESG 1)

MIC 2) ratio

Group A

> 0.9 mm

> 0.8 mm

Group B

0.5 ... 0.9 mm

0.45 ... 0.8 mm

Group C

< 0.5 mm

< 0.45 mm

D00 E 19.06.2001 21:40 Uhr Seite 148



D00

Nitri

le-b

utad

iene

rubb

er (N

BR) P

erbu

nan

Chlo

ropr

ene

rubb

er (C

R) N

eopr

ene

Silic

one

rubb

erFl

uoro

rubb

er (F

PM, F

KM) V

iton

Teflo

n (P

TFE)

EPDM

Medium

x = resistant– = conditionally resistanto = not resistant

Acetaldehyde o o o o x xAcetic acid (crystalline), pure – o – o x –Acetic acid, industrial x xAcetic acid vapors x x – xAcetic acid, 20 % x x x

Acetic acid, 50 % o x x xAcetic acid, 80 % o o xAcetic anhydride – x x xAceto-acetic ester o x –Acetone o – o x x

Acetophenone o x xAcetylene x – x xAcrylnitrile – o xAir, clean x x x x x xAir, oily x x x x x o

Ammonia liquid x x – – x xAmmonia gas x x o xAmyl acetate o o o x xAmyl alcohol – – o o x xAniline o o x x x

Anthracene oil o o x xASTM oil No. 1 x x x x x oASTM oil No. 2 x x x x x oASTM oil No. 3 – – – o x oBenzaldehyde 100 % x x

Benzene o o x x oBenzene bromide o o x oBenzoic acid x xBitumen xBlast furnace gas x x x x –

Boron trifluoride x x xBromine o o x xButadiene x oButane x – x x oButyl acetate o o o x –

Butyl alcohol – x – o x xButyl glykol x x xButyraldehyde o o x x –Carbolineum o o x x x oCarbon bisulfide o o – x x o

Nitri

le-b

utad

iene

rubb

er (N

BR) P

erbu

nan

Chlo

ropr

ene

rubb

er (C

R) N

eopr

ene

Silic

one

rubb

erFl

uoro

rubb

er (F

PM, F

KM) V

iton

Teflo

n (P

TFE)

EPDM

Medium


Carbon dioxide, dry x x x x xCarbon dioxide, wet x x x xCarbon tetrachloride o o o x x oChloracetic acid o o o x xChlorinated solvents o x

Clorine, dry x xChlorine water – x o x xChlorine, wet o o x x xChlorobromomethane o x –Chlorobenzene o o o x o

Chloroform o o x x oChloromethyl o – x x oCitrus oils o o xCoke furnace gas o o x oCopra oil acid – o – x

Cottonseed oil x – x x –Cresol x x oCrude petroleum – x x oCyclohexane x x x oCyclohexanone o o o x o

Cyclohexylamine x oDecahydronaphtalene x x oDesmodur T o o x oDesmophene 2000 x xDibutylphthalate o o x o x x

Dichlorethylene xDichlorethane x x x oDiethylamine – o x xDiethylene glycol x x x xDiethyl ether o – o x o

Diethyl ether o o x –Diethyl sebazate o o x –Dichlorbenzene o o x oDichlorbutylene x x oDiesel oil o – x x o

Di-isopropyl ketone o x xDimethyl ether o x xDimethylaniline o o xDimethyl formamide o o o x –Dioctylphthalate o o x x –

Table XV: Chemical resistance of commonly used elastomer gaskets and sealingmaterials


D00 E 19.06.2001 21:40 Uhr Seite 149



Nitri

le-b

utad

iene

rubb

er (N

BR) P

erbu

nan

Chlo

ropr

ene

rubb

er (C

R) N

eopr

ene

Silic

one

rubb

erFl

uoro

rubb

er (F

PM, F

KM) V

iton

Teflo

n (P

TFE)

EPDM

Medium


Dioxan o o x –Diphenyl o o – x x oDiphenyloxyd o x xEdenol 888 x xEssential oils o o – x o

Ethyl acetate (acetic ether) o o x xEthane x – o x x oEthyl acetate o o o o x oEthyl acrylate o xEthyl alcohol, denatured – – – x

Ethyl alcohol, pure – – – x xEthyl cloride x – o x –Ethylene bromide o o xEthylene chloride x –Ethylene dichloride o o – x

Ethylene glycol x x o x x xEthyl ether o o o x –Ethyl silicate x x xEthyl acrylate xFatty acids – – x

Fatty alcohol x x x x –Fir leaf oil x o xFluorbenzene x o x oHydrofluoric acid, cold, 5 % x x xHydrofluoric acid, cold, pure – x

Formaldehyde x – x x xFormalin, 55 % x x x xFormic acid – – x –Formic acid methyl ester o – – xFreon 11 x x o x

Freon 12 x x – x –Freon 22 o x o x xFreon 113 x x x xFurane o o x oFurfurol o o o o x

Gas oill x – x x oGenerator gas x – x xGlycerine x x x x x xGlycol x x x x x xHalowax oil o o x

Nitri

le-b

utad

iene

rubb

er (N

BR) P

erbu

nan

Chlo

ropr

ene

rubb

er (C

R) N

eopr

ene

Silic

one

rubb

erFl

uoro

rubb

er (F

PM, F

KM) V

iton

Teflo

n (P

TFE)

EPDM

Medium


Heating fuel oil (coal base) o o x xHeating fuel oil (petroleum crude base) x – x x oHeptane x – o x x oHexaldehyde o o x xHexane x – o x x o

Hydraulic fluidsHydraulic oils DIN 51524 x – – x x oPhosphoric ester HFD o o o xPolyglycol water HFC x – x x x x

Hydrobromic acid o o x x

Hydrobromic crystalline acid x xHydrocyanic acid – – x xHydrogen bromide xHydrogen gas 20 x x – xHydrogen sulfide x x x x

Isobutyl alcohol – x x xIsopropyl acetate o o x –Isopropyl alcohol – x x xIsopropyl chloride o o x oIsopropyl ether o x

Kerosene – – x xKerosine x – x x oLighting gas – – x x oMaleic anhydride xMercury x x x x

Methane x – x –Methane (pit gas) x x x x –Methylene chloride o o o x oMethyl acrylate o x oMethyl alcohol (methanol) – – x o x x

Methyl ethyl ketone o o o x –Methyl isobutyl ketone o o x –Methyl methacrylate o o o x oMethyl salicylate o o x –Naphtalene o o x o

Natural gas – x – x x –Nitrobenzene o o o x oNitrous oxide x x xOleic acid x x x oOrange oil o o x x



D00 E 19.06.2001 21:40 Uhr Seite 150



D00

Nitri

le-b

utad

iene

rubb

er (N

BR) P

erbu

nan

Chlo

ropr

ene

rubb

er (C

R) N

eopr

ene

Silic

one

rubb

erFl

uoro

rubb

er (F

PM, F

KM) V

iton

Teflo

n (P

TFE)

EPDM

Medium


Oxygen x x x xOzone o – x x x xPalmitic acid x x x oPalm oil acid – o x xParaffin x x x x o

Paraffin oil x x x oPentachlordiphenyl o x x oPentane x x o x x oPerchloroethylene o o x x oPetrol x – o x o

Petrol alcohol 3:1 – o x xPetrol benzene 4:1 x o o x x oPetrol benzene 7:3 o o o x x oPetrol benzene 3:2 o o o x x oPetrol benzene 1:1 o o o x x o

Petrol benzene 3:7 o o o x xPetrol benzene spirit 5:3:2 o o o x oPhenol o o x x x oPhenyl ethyl ether o o x oPhenylic acid (phenol) o o – – x –

Phosphorous chloride o – x xPhthalic anhydride x x x xPiperidine o o xPolyglycol x xPropane, gas x x x x x

Propylene oxyde o x –Propyl alcohol x x x xPydraul F-9 o o x x xPydraul AC o x x xPydraul A 150 o x x

Pydraul A 200 o x x xPyridine o x –Salicylic acid x x x xSkydrol 500 x o x xSkydrol 7000 x x x

Stearic acid – x xStyrene o o x x oSulfur – x x x x xSulfur dioxide o o x x xSulfur trioxide, dry o – x x –

Nitri

le-b

utad

iene

rubb

er (N

BR) P

erbu

nan

Chlo

ropr

ene

rubb

er (C

R) N

eopr

ene

Silic

one

rubb

erFl

uoro

rubb

er (F

PM, F

KM) V

iton

Teflo

n (P

TFE)

EPDM

Medium


Tar oil o x oTetrachlorethylene x x oTetrahydrofurane o o x oTetraline o o o x oToluene o o o x x o

Transformer oil x x x x oTrain oil o x x –Triethanolamine o o x –Tributoxyethyl phosphate o o x oTributyl phosphate o o o x o

Trichloroethane o o x xTrichloroethylene o o x x oTrichloroethyl phosphate 20 – x xTrichloroethyl phosphate 80 o x xTrichloracetic acid 60 x x –

Tricresyl phosphate o x x x –Turpentine – – x x oTurpentine oil, pure x x x oVinyl acetate o o xVinylaceto-acetic acid 3:2 o – o o x

Vinyl chloride, liquid x x xWater 50 x x x x xWater 100 x – x x xWood oil – xXylamon o o x x o

Xylene o o o x x o



D00 E 19.06.2001 21:40 Uhr Seite 151



Vacuum symbolsAll symbols with the exception of thosemarked with *) do not depend on the posi-tion.

*) These symbols may only be used in theposition shown here (tip of the anglepointing down)

The symbols for vacuum pumps shouldalways be arranged such that the side withthe constriction is allocated to the higherpressure

Vacuum pumps

Piston vacuum pump


Vacuum pump, general

Rotary positive displacement pump *)

Rotary plunger vacuum pump *)

Sliding vane rotary vacuum pump *)

Rotary piston vacuum pump *)

Liquid ring vacuum pump *)

Roots vacuum pump *)

Turbine vacuum pump, general

Radial flow vacuum pump

Axial flow vacuum pump

Turbomolecular pump

Table XVI: Symbols used in vacuum technology (extract from DIN 28401)

Diffusion pump *)

Adsorption pump *)

Ejector vacuum pump *)

Getter pump

Sputter-ion pump

Cryopump

Scroll pump *)

Evaporation pump

Condensate trap, general

Condensate trap with heatexchanger (e.g. cooled)

Gas filter, general

Accessories

Baffle, general

Cooled baffle

Filtering apparatus, general

Cold trap, general

Cold trap with coolant reservoir

Sorption trap

Throttling

Vacuum chamber

Vacuum bell jar

Vacuum chambers

Shut-off device, general

Shut-off devices

Shut-off valve, straight-through valve

Right-angle valve

Stop cock

Three-way stop cock

Gate valve

Butterfly valve

Right-angle stop cock

Nonreturn valve

Safety shut-off valve

Manual operation

Variable leak valve

Electromagnetic operation

Modes of operation

D00 E 19.06.2001 21:40 Uhr Seite 152



D00Table XVI: Symbols used in vacuum technology (extract from DIN 28401)

Hydraulic or pneumatic operation

Electric motor drive

Weight-operated

Flange connection, general

Bolted flange connection

Small flange connection

Connections andpiping

Clamped flange connection

Threaded tube connection

Ball-and-socket joint

Spigot-and-socket joint

Taper ground joint connection

Intersection of two lines withconnection

Intersection of two lines without connection

Branch-off point

Combination of ducts

Flexible connection (e.g. bellows, flexible tubing)

Linear-motion leadthrough,flange-mounted

Linear-motion leadthrough,without flange

Leadthrough for transmissionof rotary and linear motion

Rotary transmission leadthrough

Electric current leadthrough

General symbol for vacuum *)

Vacuum measurement, vacu-um gauge head *

Vacuum gauge, operating anddisplay unit for vacuum gaugehead *)

Vacuum gauge, recording *)

Vacuum gauge with analogmeasured-value display *)

Vacuum gauge with digitalmeasured-value display *)

Measurement of throughput

Measurement and gauges

D00 E 19.06.2001 21:40 Uhr Seite 153



Kelvin Celsius Réaumur Fahrenheit Rankine

Boiling point H2O 373 100 80 212 672

Body temperature 37 °C 310 37 30 99 559

Room temperature 293 20 16 68 527

Freezing point H2O 273 0 0 32 492

NaCl/H2O 50:50 255 –18 –14 0 460

Freezing point Hg 34 –39 –31 –39 422

CO2 (dry ice) 195 –78 –63 –109 352

Boiling point LN2 77 –196 –157 –321 170

Absolute zero point 0 –273 –219 –460 0

Table XVII: Temperature comparison and conversion table (rounded off to whole degrees)

KKelvin

°CCelsius

°CRéaumur

°FFahrenheit

°RRankine

K °C °R °F °RKelvin Celsius Réaumur Fahrenheit Rankine

1 K – 273 (K – 273) (K – 273) + 32 K = 1,8 K

°C + 273 1 · °C · °C + 32 (°C + 273)

· °R + 273 · °R 1 · °R + 32 (°R + 273)

(°F – 32) + 273 (°F – 32) (°F – 32) 1 °F + 460

(°R) (°R – 273) (°R – 273) °R – 460 1

45

95

95

95

95

45

54

54

59

59

59

59

49

45

59

94

59

54

[ ]

[ ]

Conversion in

Fig. 9.1: Variation of mean free path λ (cm) with pressure for various gases

Pressure p [mbar]

Mea

n fre

e pa

th λ

[cm

]

Fig. 9.2: Diagram of kinetics of gases for air at 20 °C

λ : mean free path in cm (λ ~ 1/p)n : particle number density in cm–3 (n ~ p)ZA : area-related impingement rate in cm–3 · s–1 (ZA ~ p2)ZV : volume-related collision rate in cm–3 · s–1 (ZV ~ p2)

1λ ∼p

Pressure p [mbar]

Z A– p

2 Z V–

p2

D00 E 19.06.2001 21:41 Uhr Seite 154



D00

Fig. 9.3: Decrease in air pressure (1) and change in temperature (2) as a function ofaltitude

Pres

sure

[mba

r]

Altitude [km]

Tem

pera

ture

(K)

Fig. 9.5: Conductance values for piping of commonly used nominal width with circu-lar cross-section for laminar flow (p = 1 mbar) according to equation 53a.(Thick lines refer to preferred DN) Flow medium: air (d, l in cm!)

Pipe length l [cm]

106

105

104

103

102

101

100

86

4

2

101 102 103 1042 4 6 8

cond

ucta

nce

[l · c

m–1

]

Fig. 9.4: Change in gas composition of the atmosphere as a function of altitude

Molecules/atoms [cm–3]

Altit

ude

(km

)

Fig. 9.6: Conductance values for piping of commonly used nominal width with cir-cular cross-section for molecular flow according to equation 53b. (Thicklines refer to preferred DN) Flow medium: air (d, l in cm!)

Pipe length l [cm]

cond

ucta

nce

[l · s

–1]

D00 E 19.06.2001 21:41 Uhr Seite 155



Fig. 9.7: Nomogram for determination of pump-down time tp of a vessel in the rough vacuum pressure range

pEND – pend, pmbar

pSTART < 1013 mbarpSTART = 1013 mbar

pSTART – pend, P

pEND – pend, PR =

Column ➀: Vessel volume V in litersColumn ➁: Maximum effective pumping speed Seff,max

at the vessel in (left) liters per second or(right) cubic meters per hour.

Column ➂: Pump-down time tp in (top right) secondsor (center left) minutes or (bottom right)hours.

Column ➃: Right:Pressure pEND in millibar at the END of thepump-down time if the atmosphericpressure pSTART ( pn = 1013 prevailed at theSTART of the pump-down time. The desiredpressure pEND is to be reduced by theultimate pressure of the pump pult,p and thedifferential value is to be used in the co-lumns. If there is inflow qpV,in, the valuepend – pult,p – qpV,in / Seff,max is to be usedin the columns.

Left:Pressure reduction ratio R = (pSTART – pult,p– qpV,in / Seff,max)/(pend – pult,p – qpV,in /Seff,max), if the pressure pSTART prevails atthe beginning of the pumping operationand the pressure is to be lowered to pENDby pumping down.The pressure dependence of the pumpingspeed is taken into account in the nomo-gram and is expressed in column ➄ bypult,p. If the pump pressure pult,p is small inrelation to the pressure pend which isdesired at the end of the pump-downoperation, this corresponds to a constantpumping speed S or Seff during the entirepumping process.

Example 1 with regard to nomogram 9.7:

A vessel with the volume V = 2000 l is to be pumpeddown from a pressure of pSTART = 1000 mbar(atmospheric pressure) to a pressure of pEND = 10-2

mbar by means of a rotary plunger pump with aneffective pumping speed at the vessel of Seff,max = 60 m3/h = 16.7 l · s-1. The pump-down timecan be obtained from the nomogram in two steps:

1) Determination of τ: A straight line is drawnthrough V = 2000 l (column ➀) and Seff = 60 m3/h-1

= 16.7 l · s-1 (column ➁) and the value t = 120 s = 2 min is read off at the intersection of these straightlines with column ➂ (note that the uncertainty of thisprocedure is around ∆τ = ± 10 s so that the relativeuncertainty is about 10 %).

2) Determination of tp: The ultimate pressure of therotary pump is pult,p = 3 · 10-2 mbar, the apparatus isclean and leakage negligible (set qpV,in = 0); this ispSTART – pult,p = 10-1 mbar – 3 · 10-2 mbar = 7 · 10-2

mbar. Now a straight line is drawn through the pointfound under 1) t = 120 s (column ➂) and the pointpEND – pult,p = 7 · 10-2 mbar (column ➄) and theintersection of these straight lines with column ➃

tp = 1100 s = 18.5 min is read off. (Again the relativeuncertainty of the procedure is around 10 % so thatthe relative uncertainty of tp is about 15 %.) Takinginto account an additional safety factor of 20 %, one can assume a pump-down time of tp = 18.5 min · (1 + 15 % + 20 %)

= 18.5 min · 1.35 = 25 min.

Example 2 with regard to nomogram 9.7:A clean and dry vacuum system (qpV,in = 0) with V = 2000 l (as in example 1) is to be pumped downto a pressure of pEND = 10-2 mbar. Since this pres-sure is smaller than the ultimate pressure of therotary piston pump (Seff,max = 60 m3/h = 16.7 l ( s-1 = 3 · 10-2 mbar), a Roots pump must be usedin connection with a rotary piston pump. The formerhas a starting pressure of p1 = 20 mbar, a pumpingspeed of Seff,max = 200 m3/h – 55 l · s-1 as well aspult,p – 4 · 10-3 mbar. From pstart = 1000 mbar to p =20 mbar one works with the rotary piston pump andthen connects the Roots pump from p1 = 20 mbar topEND = 10-2 mbar, where the rotary piston pump actsas a backing pump. For the first pumping step oneobtains the time constant τ = 120 s = 2 min from thenomogram as in example 1 (straight line through V = 2000 l, Seff = 16.7 l · s-1). If this point in column ➂ is connected with the point p1 - pult,p = 20 mbar – 3 · 10-2 mbar = 20 mbar (pult,pis ignored here, i.e. the rotary piston pump has aconstant pumping speed over the entire range from1000 mbar to 20 mbar) in column 5, one obtains tp,1= 7.7 min. The Roots pump must reduce thepressure from p1 = 20 mbar to pEND = 10-2 mbar, i.e.the pressure reduction ratio R = (20 mbar – 4 · 10-3

mbar) / (10-2 mbar-4 · 10-3) = 20/6 · 10-3 mbar =3300.The time constant is obtained (straight line V = 2000l in column ➀, Seff = 55 l · s–1 in column ➁) at = 37s (in column ➂). If this point in column ➂ isconnected to R = 3300 in column ➄, then oneobtains in column ➃ tp, 2 = 290 s = 4.8 min. If onetakes into account tu = 1 minfor the changeovertime, this results in a pump-down time of tp = tp1 + tu + tp2 = 7.7 min + 1 min + 4.8 min = 13.5 min.

D00 E 19.06.2001 21:41 Uhr Seite 156



D00Fig. 9.8: Nomogram for determination of the conductance of tubes with a circular cross-section for air at 20 °C in the region of molecular flow (according to J. DELAFOSSE and

G. MONGODIN: Les calculs de la Technique du Vide, special issue “Le Vide”, 1961)

Tube

leng

th

Corr

ectio

n fa

ctor

for s

hort

tube

s

Cond

ucta

nce

for m

olec

ular

flow

Tube

dia

met

er

C

Example: What diameter d must a 1.5-m-long pipe have sothat it has a conductance of about C = 1000 l / sec in theregion of molecular flow? The points l = 1.5 m and C = 1000l/sec are joined by a straight line which is extended to inters-ect the scale for the diameter d. The value d = 24 cm is obtai-ned. The input conductance of the tube, which depends onthe ratio d / l and must not be neglected in the case of short

tubes, is taken into account by means of a correction factorα. For d / l < 0.1, α can be set equal to 1. In our example d/l = 0.16 and α = 0.83 (intersection point of the straight linewith the a scale). Hence, the effective conductance of thepipeline is reduced to C · α = 1000 · 0.83 = 830 l/sec. If d is increased to 25 cm, one obtains a conductance of 1200 · 0.82 = 985 l / sec (dashed straight line).

D00 E 19.06.2001 21:41 Uhr Seite 157



Fig. 9.9: Nomogram for determination of conductance of tubes (air, 20 °C) in the entire pressure range

Tube

leng

th [m

eter

s]

Pres

sure

[mba

r]

Mol

. flo

wKn

udse

n flo

wLa

min

ar fl

ow

Clau

sing

fact

or

Unco

rrec

ted

cond

ucta

nce

for m

ol. f

low

[m3

· h–1

]

Tube

inte

rnal

dia

met

er [c

m]

Air 20 °C

Corr

ectio

n fa

ctor

for h

ighe

r pre

ssur

es

CPr

essu

re [m

bar]

cond

ucta

nce

[ l ·

s–1]

Procedure: For a given length (l) and internal dia-meter (d), the conductance Cm, which is indepen-dent of pressure, must be determined in the mole-cular flow region. To find the conductance C* inthe laminar flow or Knudsen flow region with agiven mean pressure of p in the tube, the conduc-tance value previously calculated for Cm has to bemultiplied by the correction factor a determined inthe nomogram: C* = Cm · α.

Example: A tube with a length of 1 m and an internaldiameter of 5 cm has an (uncorrected) conductanceC of around 17 l/s in the molecular flow region, asdetermined using the appropriate connecting linesbetween the “l” scale and the “d” scale. The conduc-tance C found in this manner must be multiplied by the clausing factor γ = 0.963 (intersection ofconnecting line with the γ scale) to obtain the true conductance Cm in the molecular flow region: Cm · γ = 17 · 0.963 = 16.37 l/s.

In a tube with a length of 1 m and an internal dia-meter of 5 cm a molecular flow prevails if the meanpressure p in the tube is < 2.7 · 10-3 mbar.To determine the conductance C* at higher pressures than 2.7 · 10-3 mbar, at 8 · 10-2 mbar (= 6 · 10-2 torr), for example, the correspondingpoint on the p scale is connected with the point d = 5 cm on the “d” scale. This connecting lineintersects the “a“ scale at the point α = 5.5. The conductance C* at p = 8 · 10-2 mbar is: C* = Cm · α = 16.37 · 5.5 = 90 l/s.

D00 E 19.06.2001 21:41 Uhr Seite 158



D00Fig. 9.10: Determination of pump-down time in the medium vacuum range taking into account the outgasing from the walls

Pum

p-do

wn

time

t [m

in]

Volu

me

V [m

in3 ]

[m3

· h–1

] Pum

ping

spe

ed

Gas

evol

utio

n [m

bar ·

l · s

–1· m

–2]

Example 2

Example 1

weak

normal

strong

Area F [m2]

The nomogram indicates the relationship betweenthe nominal pumping speed of the pump, the cham-ber volume, size and nature of the inner surface aswell as the time required to reduce the pressurefrom 10 mbar to 10-3 mbar.

Example 1: A given chamber has a volume of 70 m3

and an inner surface area of 100 m2; a substantialgas evolution of 2 · 10-3 mbar · l · s-1 · m-2 is assu-med. The first question is to decide whether a pumpwith a nominal pumping speed of 1300 m3/h isgenerally suitable in this case. The coordinates for

the surface area concerned of 100 m2 and a gas evo-lution of 2 · 10-3 mbar · l · s-1 · m-2 result in an inters-ection point A, which is joined to point B by anupward sloping line and then connected via a verticalline to the curve that is based on the pumping speedof the pump of 1300 m3/h (D). If the projection to thecurve is within the marked curve area (F), the pum-ping speed of the pump is adequate for gas evoluti-on. The relevant pump-down time (reduction of pres-sure from 10 mbar to 10-3 mbar) is then given as 30 min on the basis of the line connecting the point 1300 m3/h on the pumping speed scale to the point

70 m3 (C) on the volume scale: the extension resultsin the intersection point at 30 min (E) on the timescale.

In example 2 one has to determine what pumping speed the pump must have if the vessel (volume = approx. 3 m3) with a surface area of 16 m2 and a low gas evolution of 8 · 10-5 mbar · l · s-1 · m-2 is to be evacuated from10 mbar to 10-3 mbar within a time of 10 min. Thenomogram shows that in this case a pump with anominal pumping speed of 150 m3/h is appropriate.

D00 E 19.06.2001 21:41 Uhr Seite 159



Fig. 9.11: Saturation vapor pressure of various substances

Halocarbon 11Halocarbon 12

Halocarbon 13

Halocarbon 22

Trichorethylene

Acetone

Temperature [°C]

Vapo

r pre

ssur

e [m

bar]

Temperature (°C)

1,00E-12

1,00E-11

1,00E-10

1,00E-09

1,00E-08

1,00E-07

1,00E-06

1,00E-05

1,00E-04

1,00E-03

1,00E-02

1,00E-01

1,00E+00

1,00E+01

1,00E+02

1,00E+031000

100

10

1

10-1

10-2

10-3

10-4

10-5

10-6

10-7

10-8

10-9

10-10

10-11

10-12

0 25 50 75 100 150 200 250

Mercury

Aziepon 201

Santovac 5(similar to Ultralen)

Diffelennormal

Diffelen ultra

DC 705

DC 704

Diffelenlight

Vapo

r pre

ssur

e (m

bar)

Fig. 9.12: Saturation vapor pressure of pump fluids for oil and mercury fluid entrainment pumps

Fig. 9.13: Saturation vapor pressure of major metals used in vacuum technology

Tem

pera

ture

[K]

Vapor pressure

Tem

pera

ture

[°C]

D00 E 19.06.2001 21:41 Uhr Seite 160



D00

Fig. 9.14: Vapor pressure of nonmetallic sealing materials (the vapor pressure curve for fluoro rubber lies between the curves for silicone rubber and Teflon)

1 NBR (Perbunan) 2 Silicone rubber 3 Teflon

Vapo

r pre

ssur

e [m

bar]

Temperature [°C]

Fig. 9.15: Saturation vapor pressure ps of various substances relevant for cryogenic technology in a temperaturerange of T = 2 – 80 K

p critical pointP melting point

Rough vacuum1 to approx. 103 mbar102 to approx. 105 Pa

Medium vacuum10–3 to 1 mbar10–1 to 102 Pa

High vacuum10–7 to 10–3 mbar

10–5 to 10–1 Pa

Ultrahigh vacuum<10–7 mbar

<10–5 Pa

Fig. 9.16: Common working ranges of vacuum pumps

10–13 10–12 10–11 10–10 10–9 10–8 10–7 10–6 10–5 10–4 10–3 10–2 10–1 100 101 102

Piston vacuum pump

Turbine vacuum pump

Liquid jet vacuum pump

Turbomolecular pump

Diffusion pump

Submilation pump

Sputter-ion pump

Cryopump

10310–14

p in mbar →Working range for special model or special operating data Normal working range

Liquid-ring vacuum pump

Multiple-vane rotary vacuum pump

Trochoide vacuum pump


Sliding-vane rotary vacuum pump

Roots vacuum pump

Rotary plunger vacuum pump

Gaseous-ring vacuum pump

Vapor jet vacuum pump

Diffusion ejector pump

Adsorption pump

D00 E 19.06.2001 21:41 Uhr Seite 161



Ultrahigh vacuum<10–7 mbar

<10–5 Pa

High vacuum10–7 to 10–3 mbar

10–5 to 10–1 Pa

Medium vacuum10–3 to 1 mbar10–1 to 102 Pa

Rough vacuum1 to approx. 103 mbar102 to approx. 105 Pa

Fig. 9.16a: Measurement ranges of common vacuum gauges

10–13 10–12 10–11 10–10 10–9 10–8 10–7 10–6 10–5 10–4 10–3 10–2 10–1 100 101 102

Bourdon vacuum gauge

Liquid level vacuum gauge

(McLeod vacuum gauge)

Decrement vacuum gauge

Pirani vacuum gauge

Thermal conductivity vacuum gauge

Compression vacuum gauge

Cold-cathode ionization vacuum gauge

Penning ionization vacuum gauge

Hot-cathode ionization vacuum gauge

10310–14

p in mbar →

Working range for special models or special operating data

Thermocouple vacuum gauge

Bimetallic vacuum gauge

Thermistor vacuum gauge

Triode ionization vacuum gauge for medium vacuum

Triode ionization vacuum gauge for high vacuum

Bayard-Alpert ionization vacuum gauge

Bayard-Alpert ionization vacuum gauge with modulator

Extractor vacuum gauge

Partial pressure vacuum gauge

The customary limits are indicated in the diagram.

Spring element vacuum gauge

Capacitance diaphragm vacuum gauge

Diaphragm vacuum gauge

Pressure balance

Piezoelectric vacuum gauge

U-tube vacuum gauge

Magnetron gauge

D00 E 19.06.2001 21:41 Uhr Seite 162



D00

Fig. 9.17: Specific volume Vsp of saturated water vapor in m3/kg within a range of 0.013 to 133 mbar

Saturated vapor

Fig. 9.18: Breakdown voltage U between parallel electrodes in a homogeneous electrical field as a function of gas pressure p distance between electrodes d (in mm) (Paschen curve), for air

D00 E 19.06.2001 21:41 Uhr Seite 163



Fig. 9.19: Phase diagram of water

liquid

Melting

Sublimation

Wat

er v

apor

pre

ssur

e [m

bar]

GASEOUS

Temperature [°C]

Triple point(0.01°C, 6.09 mbar)

Evaporation

SOLID

D00 E 19.06.2001 21:41 Uhr Seite 164

Statutory Units Fundamentals of Vacuum Technology


10. The statutoryunits used invacuum technology

10.1 IntroductionTwo federal German laws and the relatedimplementing provisions stipulate whichunits must be used for measurementstoday (generally since January 1, 1978) inbusiness and official documents and com-munications. The provisions resulted in anumber of fundamental changes that alsohave to be taken into account in vacuumtechnology. Many of the units commonlyused in the past, such as torr, gauss, stan-dard cubic meter, atmosphere, poise, kilo-calorie, kilogram-force, etc., are no longerpermissible. Instead, other units are to beused, some of which are new while otherswere previously used in other fields. Thealphabetical list in Section 10.2 containsthe major variables relevant for vacuumtechnology along with their symbols and

the units now to be used, including the SIunits (see below) and legally permissibleunits derived from them. The list is follo-wed by a number of remarks in Section10.3. The purpose of the remarks is, on theone hand, to establish a connection withprevious practice wherever this is neces-sary and, on the other hand, to provideexplanations on practical use of the con-tent of the alphabetical list.

The statutory units for measurements arebased on the seven basic SI units of theSystème International (SI).

Statutory units are:

a) the basic SI units (Table 10.4.1)

b) units derived from the basic SI units, insome cases with special names and unitsymbols (Tables 10.4.2 and 10.4.4)

c) units used in atomic physics (Table10.4.3)

d) decimal multiples and decimal parts ofunits, some with special names

Examples: 105 N ( m-2 = 1 bar)1 dm3 = 1 l (liter)103 kg = 1 t (ton)

Detailed descriptions are provided in publi-cations by W. Haeder and E. Gärtner (DIN),by IUPAP 1987 and by S. German, P. Draht(PTB). These should always be referred toif the present summary tailored to vacuumtechnology leaves any questions open.

10.2 Alphabetical list 1)

of variables, symbols and unitsfrequently used invacuum technologyand its applications (seealso DIN 28 402)

1) The list is based on work done by Prof. Dr. I.Lückert, for which we would like to express ourgratitude

D00

No. Variable Symbol SI- Preferred statutory No. of remark Notesunit units in Section 10.3

1 Activity (of a radioactive substance) A s–1 (Bq) s–1 3/12 (General gas constant) – see no. 733 Work W J J, kJ, kWh, Ws4 Atomic mass mu kg kg, mg see Table V in Sect. 95 Avogadro constant NA mol–1 mol–1

6 Acceleration a m · s–2 m · s–2, cm · s–2

7 Boltzmann constant k J · K–1 j · K–1, mbar · l · K–1 see Table V in Sect. 98 Celsius temperature ϑ (theta) – °C 3/29 Vapor pressure pv N · m–2, Pa mbar, bar 3/3 Pa = Pascal10 Time t s s, min, h see Table 10.4.411 Density (gas density) ρ (ro) kg · m–3 kg · m–3, g · cm–3 3/612 Dielectric constant ε (epsilon) F · m–1 F · m–1, As · V–1 · m–1 F = Farad13 Diffusion coefficient D m2 · s–1 m2 · s–1, cm2 · s–1

14 Moment of momentum L N · s · m N · s · m15 Torque M N · m N · m, kN · m16 Rotational speed,

rotational frequency n, f s–1 s–1, min–1

17 Pressure in fluids p N · m–2, Pa bar, mbar 3/3 Pa = Pascal18 Pressure as mechanical stress p N · m–2, Pa N · mm–2 3/419 Diameter d m cm, mm20 Dynamic viscosity η (eta) Pa · s mPa · s 3/521 Effective pressure pe N · m–2, Pa mbar 3/3 see also no. 12622 Electric field strength E V · m–1 V · m–1

23 Electrical capacitance C F F, µF, pF F = Farad24 Electrical conductivity σ (sigma) S · m–1 S · m–1

D00 E 19.06.2001 21:41 Uhr Seite 165

Statutory UnitsFundamentals of Vacuum Technology


No. Variable Symbol SI unit Preferred statutory No. of remark Notesunits in Section 10.3

25 Electrical conductance G S S S = Siemens26 Electrical voltage U V V, mV, kV27 Electric current density S A · m–2 a · m–2, A · cm–2

28 Electric current intensity I A A, mA, µA29 Electrical resistance R Ω (ohm) Ω , kΩ , MΩ 30 Quantity of electricity (electric charge) Q C C, As C = Coulomb31 Electron rest mass me kg kg, g see Table V in Sect. 932 Elementary charge e C C, As33 Ultimate pressure pult N · m–2, Pa mbar34 Energy E J J, kJ, kWh, eV J = Joule35 Energy dose D J · k–1 3/5 a36 Acceleration of free fall g m · s–2 m · s–2 see Table V in Sect. 937 Area A m2 m2, cm2

38 Area-related impingement rate ZA m–2 · s–1 m–2 · s–1; cm–2 · s–1

39 Frequency f Hz Hz, kHz, MHz40 Gas permeability Qperm m3 (NTP) cm3 (NTP) 3/19 d =day (see Tab. 10.4.4–––––––––– ––––––––––

m2 · s · Pa m2 · d · bar see no. 73 and no. 103)41 Gas constant R42 Velocity v m · s–1 m · s–1, mm · s–1, km · h–1

43 Weight (mass) m kg kg, g, mg 3/644 Weight (force) G N N, kN 3/745 Height h m m, cm, mm46 Lift s m cm see also no. 13947 Ion dose J C · kg–1 c · kg–1, C · g–1 3/848 Pulse p (b) N · s N · s49 Inductance L H H, mH H = Henry50 Isentropic exponent κ (kappa) – – κ = cp · cv

–1

51 Isobaric molar heat capacity Cmp J · mol–1 · K–1 J · mol–1 · K–1

52 Isobaric specific heat capacity cp J · kg–1 · K–1 J · kg–1 · K–1

53 Isochore molar heat capacity Cmv J · mol–1 · K–1

54 Isochore specific heat capacity cv J · kg–1 · K–1 J · kg–1 · K–1

55 Kinematic viscosity ν (nü) m2 · s–1 mm2 · s–1, cm2 · s–1 3/956 Kinetic energy EK J J57 Force F N N, kN, mN 3/10 N = Newton58 Length l m m, cm, mm 3/1159 Linear expansion coefficient α (alpha) m m––––– –––––––– ; K–1

m · K m · K60 Leak rate QL N · m · s–1 mbar · l ; cm3 3/12––––––– –––– (NTP)

s s61 Power P W W, kW, mW62 Magnetic field strength H A · m–1 A · m–1 3/1363 Magnetic flux density B T T 3/14 T = Tesla64 Magnetic flux Φ (phi) Wb, V · s V · s 3/15 Wb = Weber65 Magnetic induction B T T see no. 6366 Mass m kg kg, g, mg 3/667 Mass flow rate qm kg · s–1 kg · s–1, kg · h–1, g · s–1

68 Mass content wi kg · kg–1 %, o/oo, ppm ppm = parts per million69 Mass concentration ρi (ro-i) kg · m–3 kg · m–3, g · m–3, g · cm–3

70 Moment of inertia J kg · m2 kg · m2

71 Mean free path λ m m, cm72 Molality bi mol · kg–1 mol · kg–1

73 Molar gas constant R J mbar · l see Table V in Sect. 9–––––––––– ––––––––––mol · K mol · K

74 Molar mass (quantity-related mass) M kg mol–1 kg · kmol–1, g · mol–1

D00 E 19.06.2001 21:41 Uhr Seite 166




75 Molar volume Vm m3 · mol–1 m3 · mol–1, l · mol–1

76 Molar volume, standard Vmn m3 · mol–1 m3 · mol–1 (NTP) see Table V in Sect. 9l · mol–1 (NTP)

77 Molecular mass m kg g78 Normal stress (mech.) σ (sigma) N · m–2 N · mm–2

79 Standard density of a gas ρn (ro-en) kg · m–3 kg · m–3, g · cm–3

80 Standard pressure pn N · m–2, Pa mbar see Table V in Sect. 981 Standard volume Vn m3 m3 (NTP), cm3 (NTP) 3/1682 Partial pressure pi N · m–2, Pa mbar 3/1783 Period T s s, ms, µs84 Permeation coefficient P m3 · m cm2 3/18––––––––– ––––––––

s · m2 · bar s · mbar85 Planck constant h J · s J · s see Table V in Sect. 986 pV throughput qpV N · m · s–1 mbar · l · s–1 3/1987 pV value pV N · m mbar · l 3/1988 Radius (also molecular radius) r m cm, mm, µm89 Space charge density ρ (ro) C · m–3 C · m–3, As · m–3

90 Solid angle Ω (omega) sr sr sr = steradian91 Relative atomic mass AT – – 3/20 nondimensional variab.92 Relative molecular mass Mr – – 3/21 nondimensional variab.93 Relative particle mass Mr – – nondimensional variab.94 Residual vapor pressure prd N · m–2, Pa mbar95 Residual gas pressure prg N · m–2, Pa mbar96 Residual (total) pressure pr N · m–2, Pa mbar97 Reynold number nondimensional Re – – nondimensional variab.

variable98 Saturation vapor pressure ps N · m–2, Pa mbar99 Throughput (of a pump) qpV, Q N · m · s–1 mbar · l · s–1

100 Pumping speed S m3 · s–1 m3 · h–1, l · s–1 see no. 132101 Stress (mech.) ρ, σ, τ N · m–2 N · m–2, N · mm–2 3/4 see no. 18

(ro,sigma, tau)

102 Specific electron charge –e · me–1 C · kg–1 C · kg–1, As · kg–1 see Table V in Sect. 9

103 Specific gas constant Ri J · kg–1 · K–1 mbar · l 3/22––––––––kg · K

104 Specific ion charge e · m–1 C · kg–1 C · kg–1, As · kg–1

105 Specific electrical resistance ρ (ro) Ω · m Ω · cm, Ω · mm2 · m–1

106 Specific volume v m3 · kg–1 m3 · kg–1; cm3 · g–1

107 Specific heat capacity c J · kg–1 · K–1 J · kg–1 · K–1, J · g–1 · K–1 3/23108 Stefan-Boltzmann constant s (sigma) W W see Table V in Sect. 9–––––– ––––––

m2 · K4 m2 · K4

109 Quantity of substance ν (nü) mol mol, kmol110 Throughput of substance qv mol · s–1 mol · s–1

111 Concentration of substance ci mol · m–3 mol · m–3, mol · l–1 for substance “i”112 Collision rate Z s–1 s–1

113 Conductance C, German: L m3 · s–1 m3 · s–1, l · s–1

114 Flow resistance R s · m–3 s · m–3, s · l–1

115 Number of particles N – – nondimensional variab.116 Particle number density (volume-related) n m–3 cm–3

117 Particle number density (time-related) qN s–1 s–1 see no. 120118 Particle throughput density jN m–2 · s–1 m–2 · s–1, cm–2 · s–1 see no. 121119 Particle mass m kg kg, g120 Particle flux qN s–1 s–1 see no. 117

D00

D00 E 19.06.2001 21:41 Uhr Seite 167




121 Particle flux density jN m–2 · s–1 m–2 · s–1, cm–2 · s–1 see no. 118122 Thermodyn. temperature T K K, mK123 Temperature difference ∆T, ∆ϑ K K, °C 3/24124 Temperature conductivity a m2 · s–1 a = λ · ρ–1 · cp125 Total pressure pt N · m–2, Pa mbar 3/3126 Overpressure pe N · m–2, Pa mbar 3/3127 Ambient pressure Pamb N · m–2, Pa mbar, bar 3/3128 Speed of light in vacuum c m · s–1 m · s–1, km · s–1 see Table V in Sect. 9129 Evaporation heat Ld J kJ130 Viscosity, dynamic η (eta) Pa · s mPa · s see no. 20131 Volume V m3 m3, l, cm3

132 Volume throughput (volumetric flow) qv m3 · s–1 m3 · h–1, l · s–1

133 Volume concentration σi m3 · m–3 l · l–1, %, o/oo, ppm ppm = parts per million(sigma-i)

134 Volume-related collision rate Zv s–1 · m–3 s–1 · m–3, s–1 · cm–3

135 Quantity of heat Q J J, kJ, kWh, Ws 3/25136 Heat capacity C J · K–1 J · K–1, kJ · K–1

137 Thermal conductivity λ W W––––– ––––––(lambda) K · m K · m

138 Heat transfer coefficient α (alpha) W––––––K · m2

139 Path length s m m, cm140 Wave length λ (lambda) m nm 3/11141 Angle (plane) α, β, γ rad rad rad, °, ‘, ‘’ 3/26 rad = radian

(alpha,beta, gamma)

142 Angular acceleration α (alpha) rad · s–2 rad · s–2

143 Angular velocity ω (omega) rad · s–1 rad · s–1

144 Efficiency η (eta) – – nondimensional variab.145 Time t s s, min, h, nn, mn see Table 10.44146 Period of time t, ∆t s s, min, h see Table 10.44

10.3 Remarks on alphabetical list in Section 10.2

3/1: ActivityThe unit previously used was curie (Ci).

1 Ci = 3.7 · 1010 · s-1 = 37 ns-1

3/2: (°C) Celsius temperatureThe term degrees Celsius is a special namefor the SI unit kelvin (K) [see no. 122] forindicating Celsius temperatures. The termdegrees Celsius is legally approved.

3/3: PressureThe revised version of DIN 1314 must be complied with. The specifications ofthis standard primarily apply to fluids(liquids, gases, vapors). In DIN 1314,

bar (1 bar = 0.1 MPA = 105 Pa) is stated in addition to the (derived) SI unit, 1 Pa = 1 N · m-2, as a special name for onetenth of a megapascal (Mpa). This is inaccordance with ISO/1000 (11/92), p. 7.Accordingly the millibar (mbar), a veryuseful unit for vacuum technology, is alsopermissible: 1 mbar = 102 Pa = 0.75 torr.The unit “torr” is no longer permissible.

Special noteExclusively absolute pressures are measu-red and used for calculations in vacuumtechnology.

In applications involving high pressures,frequently pressures are used that arebased on the respective atmospheric pres-sure (ambient pressure) pamb. Accordingto DIN 1314, the difference between apressure p and the respective atmosphericpressure (ambient pressure) pamb is desi-

gnated as overpressure pe: pe = p – pamb.The overpressure can have positive ornegative values.

Conversions1 kg · cm-2 = 980.665 mbar = 981 mbar

1 at (technical atmosphere) = 980.665 mbar = 981 mbar

1 atm (physical atmosphere) = 1013.25 mbar = 1013 mbar

1 atmosphere above atmospheric pressure(atmospheric overpressure) = 2026.50 mbar = 2 bar

1 atm1 torr = 1 mm Hg = –––––– = 1

760

33.322 Pa= 1.333 mbar

1 meter head of water = 9806.65 Pa = 98 mbar

D00 E 19.06.2001 21:41 Uhr Seite 168



1 mm Hg = 133.332 Pa = 1.333 mbar = 4/3 mbar

The pressure as mechanical stress(strength) is generally given in pascal (Pa)and in N · nm–2.

Conversions:1 Pa = 1 N · m–2 = 10–6 N · mm–2

1 kg · cm–2 = 98,100 Pa = 0.981 N · mm–2

= 0,1 N mm–2

1 kg · mm–2 = 9,810,000 Pa = 9.81 N · mm–2 = 10 N · mm–2

3/5: Dynamic viscosityThe unit previously used was poise (P).

1 P = 0.1 Pa · s = 1 g · cm–1 · s–1

3/5a: Energy doseRad (rd) is no longer permissible.

11 rd = –––– J · kg–1

100

3/6: WeightDIN 1305 is to be complied with in thiscontext. Because of its previous ambiva-lence, the word weight should only beused to designate a variable of the natureof a mass as a weighing result for indica-ting quantities of goods.

The designations “specific weight” and“specific gravity” should no longer beused. Instead, one should say density.

3/7: Weight forceSee DIN 1305. The previous units pond (p)and kilopond, i.e. kilogram-force, (kp) aswell as other decimal multiples of p are nolonger used.

1 kp = 9.81 N

3/8: Ion doseThe previously used unit was the Röntgen(R). 1 R = 2,58 · 10–4 C · kg–1

3/9: Kinematic viscosityThe previously used unit was stokes (St).

1 St = 1 cm2 · s–1; 1 cSt = 1 mm2 · s–1

3/10: ForceThe dyne, the CGS unit for force, is no lon-ger used.

1 dyne = 10-5 N

3/11: Length/wavelengthThe unit Ångström (Å) (e.g. for wave-length) will no longer be used in future(see Table 4.6).

1 Å = 10-8 cm = 0.1 nm

3/12: Leak rateIn DIN 40.046 sheet 102 (draft of August1973 issue), the unit mbar · dm3 · s-1

(= mbar · l · s-1) is used for the leak rate.Note that the leak rate corresponding to the unit 1 mbar · l · s-1 at 20 °C is practically the same as the leak rate 1 cm3 · s-1 (NTP). (See also 3/17)

3/13: Magnetic field strengthThe previously used unit was the oersted(Oe).

1 Oe = 79.577 A · m-1

3/14: Magnetic flux densityThe previously used unit was the gauss(G).

1 G = 10-4 Vs · m-2 = 10-4 T (T = Tesla)

3/15: Magnetic fluxThe previously used unit was the maxwell(M).

1 M = 10-8 Wb (Weber)

3/16: Standard volumeDIN 1343 must be complied with.

The designation m3 (NTP) or m3 (pn, Tn) is proposed, though the expression inparentheses does not belong to the unitsymbol m3 but points out that it refers tothe volume of a gas in its normal state (Tn = 273 K, pn = 1013 mbar).

3/17: Partial pressureThe index “i” indicates that it is the partialpressure of the “i-th” gas that is containedin a gas mixture.

3/18: Gas permeabilityThe permeation coefficient is defined asthe gas flow m3 · s-1 (volumetric flow pV)that goes through a fixed test unit of agiven area (m2) and thickness (m) at agiven pressure difference (bar).

According to DIN 53.380 and DIN 7740,Sheet 1, supplement, the gas permeability(see no. 40) is defined as “the volume of agas, converted to 0 °C and 760 torr, whichgoes through 1 m2 of the product to betested at a certain temperature and a cer-tain pressure differential during a day (= 24 hours)”.

3/19: pV throughput/pV valueDIN 28.400, Sheet 1 is to be taken intoaccount here. No. 86 and no. 87 have aquantitative physical significance only ifthe temperature is indicated in each case.

3/20: Relative atomic massMisleadingly called “atomic weight” in thepast!

3/21: Relative molecular massMisleadingly called “molecular weight” inthe past!

3/22: Specific gas constantAs mass-related gas constant of the substance “i”. Ri = Rm ( Mi-1; Mi molarmass (no. 74) of the substance “i”. Seealso DIN 1345.

3/23: Specific heat capacityAlso called specific heat:

Specific heat (capacity) at constant pressure: cp.

Specific heat (capacity) at constant volume: cV.

3/24: Temperature differenceTemperature differences are given in K, but can also be expressed in °C. The designation degrees (deg) is no longerpermissible.

D00

D00 E 19.06.2001 21:41 Uhr Seite 169



3/25: Quantity of heatThe units calorie (cal) and kilocalorie (kcal)are no longer be used.

1 kcal = 4.2 kJ

3/26: Angle1 radian (rad) is equal to the plane anglewhich, as the central angle of a circle, cutsout an arc having a length of 1 m from thecircle. See also DIN 1315 (8/82).

π1° = ––––– rad: 1’ = 1°/60; 1’’ = 1’/60.

180

360°1 rad = ––––– · 60°

2π

10.4 Tables

10.4.1 Basic SI units

Basic unit Symbol Variable————————————————meter m lengthkilogramm kg masssecond s time, period;durationampere A electric currentkelvin K thermodyn.

temperaturemole mol quantity of

substancecandela cd luminous

intensity

10.4.2 Derived coherent 1) SIunits with special namesand symbols (alphabetical)

Name Symbol Variable Relationshipof unit————————————————————

coulomb C quantity of electricity or electric charge 1 C = 1 A · s

farad F electrical capacitance 1 F = 1 A · s · V–1

henry H inductance 1 H = 1 V · s · A–1

hertz Hz frequency 1 Hz = 1 · s–1

joule J energy, work, 1 J = 1 N · m quantity = Wsof heat

lumen lm luminous flux 1 lm = cd · sr

lux lx illuminance 1 lx = 1 lm · m–2

newton N force 1 N = 1 kgm · s–2

ohm Ω electrical 1 Ω = 1 V · A–1

resistance

pascal Pa pressure, 1 Pa = 1 N · m–2

mechanical stress

radian rad 2) angle, 1 rad = 1 m · m–1

plane angle

siemens S electrical conductance 1 S = 1 · Ω–1

steradian sr 2) solid angle 1 sr = 1 m2 · m–2

tesla T magnetic 1 T = 1 Wb · m–2

flux density or induction

volt V electrical 1 V = 1 W · A–1

voltageor potential difference

watt W power, 1 W = 1 J · s–1

energy flux,heat

flux

weber Wb magnetic 1 Wb = 1 V · sflux 1 Wb = 1 V · s

1) Formed with numeric factor 1; e.g. 1 C= 1 As, 1 Pa = 1 N · m–2

2) Additional SI unit

10.4.3 Atomic units

Basic unit Symbol Vartiable————————————————Atomic Mass for

indication ofmass mu particle mass;unit 1 mu = 1/12 mass

of 12Calso amu (atomic massunit).

Electron eV energyvolt

10.4.4 Derived noncoherent SI units with specialnames and symbols

Basic unit Symbol Definition————————————————Day d 1 d = 86.400 s

Hour h 1 h = 3.600 s

Minute min 1 min = 60 s

Round angle – 2 p rad

πDegree (°) –––– rad

180

πMinute (‘) –––––– rad

10.800

1(= ––– grad)

60

p Second (‘’) ––––––– rad

648.000

1(= –––– minute)

60

D00 E 19.06.2001 21:41 Uhr Seite 170

Vacuum Technology Standards Fundamentals of Vacuum Technology


11. National andinternationalstandards andrecommendati-ons particular-ly relevant tovacuum technology

For around 20 years now, numerous stan-dards and recommendations have beendrawn up at national and international leveland revised, whenever necessary, inaccordance with the state of the art. Thesestandards and recommendations must beobserved whenever use is made of vacu-um equipment (pumps, gauges, valves,etc.) and vacuum apparatus, systems andplants are assembled. They not only con-tain specifications applying specially tovacuum technology, but also go beyondthis specific field and involve, for example,physical units, formulas, noise protectionregulations, etc.

National standards are primarily DIN stan-dards, particularly those relating to thearea of vacuum technology in the DINStandards Committee on MechanicalEngineering (NAM). International stand-ards and recommendations are drawn upand issued

a) by the International StandardizationOrganization (ISO), in particular by ISOCommittee TC 112 (vacuum technolo-gy)

b) by the European Committee of Manu-facturers of Compressors, vacuumpumps and compressed air tools(PNEUROP), in particular by PNEUROPSubcommittee C5 (vacuum technology)

c) by the European Committee for Stan-dardization (CEN), in particular by Tech-nical Committee TC 138 (nondestruc-tive testing) and Technical CommitteeTC 318.

The content of the documents drawn up bythe international organizations in a) to c),

with German participation (also by LEYBOLD), has been extensively incorpo-rated into DIN standards, as reflected indesignations such as DIN / ISO or DIN /EN.

The most important standards to be com-plied with are listed in Table 11.1 below.

Abbreviations used:D = draftCD = Committee Draft

11.1 National and international standards and recommendationsof special relevance to vacuum technology

A) National agreements, Part 1: DIN

DIN Title Issue

1301-8 E UnitsPart 1 – Names of units,

symbols 1993Part 2 – Parts and multiples

generally used 2/78Part 3 – Conversions for units

no longer used 6/79

1304 General symbolsPart 1 – General symbols 3/94Part 2 – Symbols for

meteorology and geophysics 9/89

Part 3 – Symbols for electrical energy supply 3/89

Part 5 – Symbols for flow mechanics 9/89

Part 6 – Symbols for electrical communicationstechnology 5/92

Part 7 – Symbols for meteorology and geophysics

Part 7 – Symbols forelectric machines 1/91

1305 Mass; weighed value, force, weight force, weightload, definitions 1/88

1306 Density; definitions 6/84

1313 Physical variables and equations, definitions, spelling 4/78

1314 Pressure; basic definitions, units 2/77

DIN Title Issue

1319 Basic definitions for measurement technologyPart 1 – Basic definitions 1/95Part 2 – Definitions for the

use of gauges 1/80Part 3 – Definitions for

measurement uncertainty andevaluation of gauges

and measuring equipment 8/83

Part 4 – Treatment of use of measurements 12/85

A) National agreements, Part 1: DIN(cont.)

DIN Title Issue

1343 Normal state, reference state 1/90

1345 Thermodynamics; basic definitions

12/93

1952 Flow measurement with screens, nozzles, etc. 7/82

2402 Pipelines; nominal internal diameters, definitions,classification 2/76

3535 Seals for gas supply – Part 6 4/94

8964 Circuit parts for refrigeration systems with hermetic and semi-hermetic compressorsPart 1: Tests 3/96Part 2: Requirements 9/86

(E 12/95)

16005 Overpressure gauges with elastic measuring element for general use Requirements and testing 2/87

16006 Overpressure gauges with Bourdon tube Safety-related requirements and testing 2/87

19226 – 1 Control and instrumentation technology; control and regulation technology; definitions – general principles 2/94

– 4 Control and instrumentation technology; control and regulation technology; definitions for control and regulation systems 2/94

– 5 Control and instrumentation technology; control and regulation technology; functional definitions 2/94

25436 Integral leak rate test of safety vessel with absolute pressure method 7/80

D00

D00 E 19.06.2001 21:41 Uhr Seite 171

Vacuum Technology StandardsFundamentals of Vacuum Technology


DIN Title Issue

28090 Static seals for flange connectionsPart 1 - Characteristic values

for seals and testing methods 9/95

Part 2 - Seals made of sealing plate; special testing methods for quality assurance 9/95

28400 Vacuum technology; designations and definitionsPart 1 - Basic definitions, units,

ranges, characteristic variables and basic principles 5/90

Part 2 - Vacuum pumps 10/80Part 3 - Vacuum plants;

characteristic variables and gauges 6/92

Part 4 - Vacuum coating technology 3/76

Part 5 - Vacuum drying and vacuum freeze-drying 3/81

Part 6 - Analytical techniques for surface technology 10/80

Part 7 - Vacuum metallurgy 7/78Part 8 - Vacuum systems,

components and accessories 10/80

(E 7/91)

28401 Vacuum technology; symbols – overview 11/76

28402 Vacuum technology: variables, symbols, units - overview 12/76

28403 Vacuum technology; quick connections, small flange connections 9/86

28404 Vacuum technology: flanges, dimensions 10/86

28410 Vacuum technology; mass spectrometer partial pressure gauges, definitions, characteristic variables, operating conditions 11/86

28411 Vacuum technology; acceptance specifications for mass spectrometer leak detection devices, definitions 3/76

28416 Vacuum technology; calibration of vacuum gauges ina range from 10-3 to 10-7 mbar. General methods; pressure reduction through constant flow 3/76

28417 Vacuum technology; measuring pV mass flow according to volumetric method at constant pressure 3/76

28418 Vacuum technology; standard method for calibrating vacuum gauges through direct compa-rison with a reference devicePart 1 – General principles 5/76Part 2 – Ionization vacuum

gauges 9/78Part 3 – Thermal conductivity vacuum gauges 8/80

DIN Title Issue

28426 Vacuum technology; acceptance specifications or frotary piston vacuum pumpsPart 1 – Rotary piston and

vane type rotary vacuum pumps in rough and medium vacuum range

Part 2 – Roots vacuum pumps in medium vacuum range

28427 Vacuum technology; acceptance specifications for diffusion pumps and ejector vacuum pumps for pump fluid vapor pressures of less than 1 mbar 2/83

28428 Vacuum technology; acceptance specifications for turbo-molecular pumps 11/78

28429 Vacuum technology; acceptance specifications for getter-ion pumps 8/85

28430 Vacuum technology; measuring specifications for ejector vacuum pumps and ejector compressors. Pump fluid: water vapor 11/84

28431 Acceptance specifications for liquid ring vacuum pumps 1/87

28432 Acceptance specifications for diaphragm vacuum pumps E 5/95

53380 Testing of plastic foils, determination of gas permeability 6/69

(E 10/83)

45635 Noise measurement at machines: measurement of airborne noise, enveloping surface methods.Part 13 – Compressors,

including vacuumpumps, positive displacement, turbo and steam ejectors 2/77

55350 Definitions of quality assurance and statisticsPart 11 – Basic definitions of

quality assurance 8/95Part 18 – Definitions regarding

certification of results of quality tests/quality test certificates 7/87

66038 Torr – millibar; millibar – torr conversion tables 4/71

– Thesaurus Vacui (definition of terms) 1969

A) European/national agreements, EN,DIN/EN, CEN

DIN/EN Title Issue

EN 473 Training and certification of personnel for nondestructive testing (including leak test) 7/93

837-1 Pressure gauges, Part 1: Pressure gauges with Bourdon tubes, dimensions, measurement technology, requirements and testing 2/97

837-2 Pressure gauges, Part 2: Selection and installation recommendations for pressure gauges 1/95

837-3 Pressure gauges, Part 3: Pressure gauges with plate and capsule elements, dimensions, measurement technology, requirements and testing 2/97

1330-8 E Nondestructive testing - definitions for leak test – terminology 6/94

1779 E Nondestructive testing – leak test. Instructions for selection of a testing method 3/95

1338-8 E Nondestructive testing – leak test. Terminology on leak test 1994

1518 E Nondestructive testing – determination of characteristic variables for mass spectrometer leak detectors

1593 E Nondestructive testing – bubble type testing method 12/94

NMP 826 Calibration of gaseous reference leaks, CD 9/95

Nr. 09–95

B) International agreements, ISO, EN/ISO

DIN Title Issue

1000 SI units and recommendations for the use of their multiples and of certain other units 11/92

1607 / 1 Positive displacement vacuum pumps. Measurement of performance characteristics. 12/93Part 1 - Measurement of volume rate of flow(pumping speed)

1607 / 2 Positive displacement vacuum pumps. Measurement of performance characteristics. 11/89Part 2 - Measurement of ultimate pressure

D00 E 19.06.2001 21:41 Uhr Seite 172

Vacuum Technology Standards Fundamentals of Vacuum Technology


DIN Title Issue

1608 / 1 Vapor vacuum pumps. Part 1 Measurement of volume

rate of flow 12/93

1608 / 2 Vapor vacuum pumps. Part 2: Measurement of critical

backing pressure 12/89

1609 Vacuum technology. Flange dimensions 3/86

DIN/ISO Standard atmosphere 12/792533

2861 / 1 Quick release couplings. Dimensions 8/74Part 1 - Clamped Type

2861 / 2 Quick release couplings. Dimensions 8/80Part 2 - Screwed type

3529 / 1 Vacuum Technology Vocabulary 12/81Part 1 - General terms

3529 / 2 Vacuum Technology Vocabulary 12/81Part 2 - Vacuum pumps and

related terms

3529 / 3 Vacuum Technology Vocabulary 12/81Part 3 - Vacuum gauges

3556 / 1 Measurement of performance characteristics. 1992Part 1 - Sputter ion pumps (E)

3567 Vacuum gauges. Calibration by direct comparison with a reference gauge (CD) 2/91

3568 Ionisation vacuum gauge. Calibration by direct comparison with a reference gauge (CD) 2/91

3570 / 1 Vacuum gauges – standard methods for calibration 2/91Part 1 - Pressure reduction

by continuous flow in the pressure range 10–1 ... 10–5 Pa

3669 Vacuum Technology. Bakable flanges, dimensions. 2/86Part 1: Clamped Type

EN/ISO Rubber and plastic hoses and 4080 hose lines – determination of

gas permeability 4/95

5167 Measurement of fluid flow by means of orifice plates, nozzles etc. 1980

5300 Vacuum gauges of the thermal conductivity type. 2/91Calibration by direct comparison with a reference gauge (CD)

9803 Pipeline fittings-mounting, dimensions (E) 2/93

DIN Title Issue

DIN/ISO Requirements placed on quality 10012 assurance for measuring

equipment 8/92Part 1 – Confirmation system

for measuring equipment

C) PNEUROP/C5 (6.93)

Number Title/remark identical Issueto DIN

5607 Vacuum pumps; acceptance specifications 28427 1972Part II: (Fluid entrainment pumps)

5608 Vacuum pumps; acceptance specifications 28428 1973Part III: (Turbomolecular pumps)

5615 Vacuum pumps; acceptance specifications 28429 1976Part IV: (Getter-ion pumps)

6601 Measurement of performance of ejector vacuum pumps and ejector compressors 28430 5/78

6602 Vacuum pumps; acceptance specificationsPart I: (Oil-sealed rotary pumps and Roots pumps) 28426 1979

6606 Vacuum flanges and connections; 28403 1985dimensions and

28404

PN5ASR Vacuum pumps, CC/5 acceptance specifications

refrigerator cooledcryopumps 7/89

D00

D00 E 19.06.2001 21:41 Uhr Seite 173

ReferencesFundamentals of Vacuum Technology


12. References

1. Overview, definitions andhistory

K. Diels, R. JaekelLeybold Vacuum HandbookPergamon Press1st Ed. 1966

W. Haeder, E. GärtnerDie gesetzlichen Einheiten in der TechnikBeuth-Vertrieb GmbH, 5. Aufl. 1980, Berlin 30, Köln, Frankfurt (Main)

H. EbertVakuum-Chronik, A documentation onworks concerning vacuum that were published before 1928PTB-Bericht ATWD-11, September 1977

M. Dunkel„Gedenken an Wolfgang Gaede“Physikalische Blätter Nr. 34 (1978), Heft 5, Pages 228-232 as well asVakuum-Technik, 27. Jahrgang, No. 4, Pages 99-101

IUPAP (SUNANCO Commission)Symbols, Units etc.Document 25, 1987

Leybold AGVademekum, 93 pages, 1988

M. Wutz, H. Adam, W. WalcherTheory and Practice of Vacuum Technology5. Aufl., 696 pages, 1992, Friedrich Vieweg u. Sohn, Braunschweig/ Wiesbaden

A. Guthrie and R. K. WakerlingVakuum Equipment and Techniques264 pages, 1949, McGraw-Hill, New York/London/Toronto

D. J. HucknallVacuum Technology and Applications1st Ed., 319 pages, 1991Butterworth-Heinemann, Oxford

C. M. van AttaVacuum Science and Engineering459 pages, 1965, McGraw-HillNew York/San Francisco/Toronto/London/SydneyJ. M. Lafferty et. al.Foundations of Vacuum Science andTechnology704 pages, 1998, Wiley 1998

A. SchubertNormen und Empfehlungen für die Vakuum-TechnikVakuum in der Praxis, Vol. 3, 1991, 211-217

H. Scharmann Vakuum – Gestern und HeuteVakuum in der Praxis, Vol. 2, 1990, 276-281

M. AuwärterDas Vakuum und W. GaedeVakuum-Technik, Vol. 32, 1983, 234-247

J. F. O’HanlonA User’s Guide to Vacuum Technology3nd Ed., 402 pages, Wiley 1989, New York

G. ReichWolfgang Gaede – Einige Gedanken zuseinem 50. Todestag aus heutiger SichtVakuum in der Praxis, 7th year, 1995, 136-140

S. German, P. DrahtHandbuch SI EinheitenVieweg Braunschweig/Wiesbaden, 1979, 460 pages

“Gesetz über Einheiten im Meßwesen”vom 2. Juli 1969“Gesetz zur Änderung des Gesetzes überEinheiten im Meßwesen” vom 6. Juli1973“Ausführungsverordnungen” vom 26.Juni 1970

In Vakuum-Technik Vol. 35, 1986:Th. MulderOtto von GuerickePages 101-110

P. SchneiderZur Entwicklung der Luftpumpen-Initiationen und erste Reife bis 1730 Pages 111-123

L. FabelPhysik in der 2. Hälfte des 19. Jahrhunderts und die vakuumtechnische Entwicklung bis GaedePages 128-138H.-B. Bürger: G. Ch. Lichtenberg und die VakuumtechnikPages 124-127

G. Reich:Gaede und seine Zeit Pages 139-145

H. AdamVakuum-Technik in der Zeit nach Gaede (1945 to the present);Pages 146-147

G. ReichDie Entwicklung der Gasreibungspumpenvon Gaede, über Holweck, Siegbahn biszu Pfleiderer und Becker (mit zahlreichenLiteraturangaben)Vakuum-Technik in der Praxis, Vol. 4,1992, 206-213

G. ReichCarl Hoffman (1844-1910), der Erfinderder DrehschieberpumpeVakuum in der Praxis, 1994, 205-208

Th. MulderBlaise Pascal und der Puy de Dôme –Große Männer der Vakuum-TechnikVakuum in der Praxis, 1994, 283-289

W. Pupp und H. K. HartmannVakuum-Technik, Grundlagen und Anwen-dungenC. Hanser, München, 1991, Wien,

2. Vacuum pumps

2.1 Positive displacementpumps, condensers

W. GaedeDemonstration einer rotierenden QuecksilberpumpePhysikalische Zeitschrift, 6, 1905, 758-760

D00 E 19.06.2001 21:41 Uhr Seite 174

W. GaedeGasballastpumpenZeitschrift für Naturforschung, 2a, 1947, 233-238

W. Armbruster und A. LorenzDas maximale Kompressionsverhältnisund der volumetrische Wirkungsgrad vonVakuumpumpen nach dem RootsprinzipVakuum-Technik, 7, 1958, 81-85

W. Armbruster und A. LorenzDie Kombination Rootspumpe-Wasser-ringpumpeVakuum-Technik, 7, 1958, 85-88

H. ReylanderÜber die Wasserdampfverträglichkeit vonGasballastpumpenVakuum-Technik, 7, 1958, 78-81

F. FauserCharakteristik von Pumpsystemen fürgrößere Wasserdampfmengen unter Vakuum und unter Anwendung von Kon-densation und Kompression des Wasserdampfes1965 Transactions of the Third Internatio-nal Vacuum Congress, Stuttgart, Bd. 2/II,393-395, Pergamon Press, Oxford 1966

M. WutzDas Abpumpen von Dämpfen mit gekühl-ten KondensatorenVakuum-Technik, 16, 1967, 53-56

H. HamacherKennfeldberechnung für RootspumpenDLR FB 69-88, 1969

H. HamacherBeitrag zur Berechnung des Saugvermögens von RootspumpenVakuum-Technik, 19, 1970, 215-221

H. HamacherExperimentelle Untersuchungen an Nachkühlern von Rootspumpen Vakuum-Technik, 23, 1974, 129-135

M. RannowÖlgedichtete Vakuumpumpen in der ChemieChemie-Technik, No. 7, 1978, 39-41

H. P. Berges et al.TRIVAC-B, ein neues Vakuumpumpen-Konzept für universelle AnwendungenVakuum-Technik, 31, 1982, 168-171

References Fundamentals of Vacuum Technology


H. LangVakuumpumpen in der chemischen Indu-strie – WälzkolbenpumpenVakuum-Technik, 1980, 72-82

H. F. WeberVakuumpumpen in der chemischen Industrie – ölgedichteteRotationsvakuumpumpenVakuum-Technik, 1980, 98-104

D. BartelsVakuumpumpen in der chemischen IndustrieFlüssigkeitsring-Vakuumpumpen – A Vakuum-Technik, 1980, 131-140

R. W. Adam und C. DahmlosFlüssigkeitsring-Vakuumpumpen – BVakuum-Technik, 1980, 141-148

U. SeegebrechtFörderung trockener Luft und vongesättigtem Luft-Wasserdampfgemischmit Flüssigkeitsring-Vakuumpumpen Vakuum-Technik, 1980, 246-252

H.-D. BürgerFortschritte beim Betrieb von WälzkolbenpumpenVakuum-Technik 1983, 140-147

U. SeegebrechtEinfluß der Temperatur des Fördermittelsauf das Saugvermögen von Flüssigkeits-ring-Vakuumpumpen bei der Förderungvon trockener Luft Vakuum-Technik, 1985, 10-14

P. Bachmann und H.-P. BergerSicherheitsaspekte beim Einsatz von ölgedichteten Drehschiebervakuumpum-pen in CVD-AnwendungenVakuum-Technik, 1987, 41-47

U. FusselTrockenlaufende Vakuumpumpen in derchemischen IndustrieVakuum in der Praxis, 1994, 85-88

L. RipperExplosionsschutz-Maßnahmen an Vaku-umpumpen (with numerous references torelevant literature)Vakuum in der Praxis, 1994, 91-100

K. P. MüllerTrockenlaufende Drehschiebervakuum-pumpen in einer Vielzweck-Produktions-anlageVakuum in der Praxis, 1994, 109-112

F. J. Eckle, W. Jorisch, R. LachenmannVakuum-Technik im ChemielaborVakuum in der Praxis, 1991, 126-133

P. Bachmann und M. KuhnEinsatz von Vorpumpen im Al-Ätzprozeß.Erprobung trockenverdichtender Klauen-pumpen und ölgedichteter Drehschieber-Vakuumpumpen im VergleichVakuum in der Praxis, 1990, 15 – 21

U. GottschlichVakuumpumpen im ChemielaborVakuum in der Praxis, 1990, 257-260

M. H. HablanianAufbau und Eigenschaften verschiedenerölfreier Vakuumpumpen für den Grob-und Feinvakuumbereich (wichtigeLiteraturangaben)Vakuum in der Praxis, 1990, 96-102

B. W. Wenkebach und J. A. WickholdVakuumerzeugung mit Flüssigkeitsring-VakuumpumpenVakuum in der Praxis, 1989, 303-310

U. Gottschlich und W. JorischMechanische Vakuumpumpen im ChemieeinsatzVakuum in Forschung und Praxis, 1989,113-116

W. JorischNeue Wege bei der Vakuumerzeugung inder chemischen VerfahrenstechnikVakuum in der Praxis, 1995, 115-118

D. LamprechtTrockenlaufende VakuumpumpenVakuum in der Praxis, 1993, 255-259

P. Deckert et al.Die Membranvakuumpumpe – Entwick-lung und technischer StandVakuum in der Praxis, 1993, 165-171

W. Jorisch und U. GottschlichFrischölschmierung – Umlaufschmierung,Gegensätze oder Ergänzung?Vakuum in der Praxis, 1992, 115-118 D00

D00 E 19.06.2001 21:41 Uhr Seite 175



W. Jitschin et al.Das Saugvermögen von Pumpen: Unter-suchung verschiedener Meßverfahren imGrobvakuumbereichVakuum in Forschung und Praxis, 7,(1995) 183 -193

H.P. Berges and M. KuhnHandling of Particles in ForevacuumpumpsVacuum, Vol. 41, 1990, 1828-1832

M. H. HablanianThe emerging technologies of oil-freevacuum pumpsJ. Vac. Sci. Technol. A6 (3), 1988, 1177-1182

E. Zakrzewski, P. L. May and B. S. EmslieDevelopments in vacuum Pumpingsystems based on mechanical pumpswith an oil free swept volumeVacuum, 38, 968, 757-760

H. WycliffeMechanical high-vacuum pumps with anoil-free swept volumeJ. Vac. Sci. Technol. A5 (4) 1987, 2608-2611

A. P. Troup and D. TurellDry pumps operating under harsh condictions in the semiconductor industryJ. Vac. Sci. Technol. A7 (3), 1989, 2381-2386

P. Bachmann and M. KuhnEvaluation of dry pumps vs rotary vanepumps in aluminium etchingVacuum 41, 1990, 1825-1827

H. P. Berges and D. GötzOil-free vacuum pumps of compactdesignVacuum, Vol. 38, 1988, 761-763

2.2 Turbomolecular pumps

W. GaedeDie MolekularluftpumpeAnnalen der Physik, 41, 1913, 337-380

W. BeckerEine neue MolekularpumpeVakuum-Technik, 7, 1958, 149-152

W. ArmbrusterVakuumpumpenkombinationen für Labor,Technikum und ProduktionChemiker-Zeitung / Chemische Apparatur,88, 1964, 895-899

W. Becker Die Turbo-MolekularpumpeVakuum-Technik, 15, 1966, 211-218 and254-260

R. Frank et al.Leistungsdaten von Turbo-Molekularpum-pen des Typs TURBOVAC mit senkrechtangeordnetem AxialkompressorVakuum-Technik, 24, 1975, 78 -85

W. BeckerEine gegenüberstellende Betrachtung vonDiffusionspumpen und MolekularpumpenErgebnisse europäischer Ultrahochvaku-umforschungLeybold-Heraeus GmbH u. Co., in its ownpublishing house, Cologne 1968, 41-48

R. Frank, E. UsselmannKohlenwasserstoffreier Betrieb mit Turbo-Molekularpumpen des Typs TURBOVACVakuum-Technik, 25, 1976, 48-51

R. Frank, E. UsselmannMagnetgelagerte Turbo-Molekularpumpendes Typs TURBOVACVakuum-Technik, 25, 1976, 141-145

H.-H. Henning und G. KnorrNeue luftgekühlte, lageunabhängigeTurbo-Molekularpumpen für Industrie undForschung Vakuum-Technik, 30, 1981, 98-101

H.-H. Henning und H. P. CasparWälzlagerungen in Turbo-Molekularpum-penVakuum-Technik, 1982, 109-113

E. Kellner et al.Einsatz von Turbo-Molekularpumpen beiAuspumpvorgängen im Grob- und FeinvakuumbereichVakuum-Technik, 1983, 136-139

D. E. Götz und H.-H. HenningNeue Turbo-Molekularpumpe für überwiegend industrielle Anwendungen Vakuum-Technik, 1988, 130-135

J. Henning30 Jahre Turbo-MolekularpumpeVakuum-Technik, 1988, 134-141

P. Duval et. al.Die SpiromolekularpumpeVakuum-Technik, 1988, 142-148

G. ReichBerechnung und Messung der Abhängig-keit des Saugvermögens von Turbo-Mole-kularpumpen von der GasartVakuum-Technik, 1989, 3-8

J. HenningDie Entwicklung der Turbo-Molekular-pumpeVakuum in der Praxis, 1991, 28-30

D. UrbanModerne Bildröhrenfertigung mit Turbo-MolekularpumpenVakuum in der Praxis, 1991, 196-198

O. Ganschow et al.Zuverlässigkeit von Turbo-Molekular-pumpenVakuum in der Praxis, 1993, 90-96

M. H. HablanianKonstruktion und Eigenschaften vonturbinenartigen Hochvakuumpumpen Vakuum in der Praxis, 1994, 20-26

J. H. Fremerey und H.-P. KabelitzTurbo-Molekularpumpe mit einer neuartigen MagnetlagerungVakuum-Technik, 1989, 18-22

H. P. Kabelitz and J.K. FremereyTurbomolecular vacuum pumps with anew magnetic bearing conceptVacuum 38, 1988, 673-676

E. Tazioukow et al.Theoretical and experimental investigationof rarefied gas flow in molecular pumpsVakuum in Forschung und Praxis, 7,1995, 53-56

D00 E 19.06.2001 21:41 Uhr Seite 176

2.3 Fluid entrainment pumps

W. GaedeDie Diffusion der Gase durch Quecksilber-dampf bei niederen Drücken und die Dif-fusionspumpeAnnalen der Physik, 46, 1915, 357-392

W. GaedeDie ÖldiffusionspumpeZ. techn. Physik, 13, 1932, 210-212R. Jaeckel, H. G. Nöller und H. KutscherDie physikalischen Vorgänge in Diffusi-ons- und DampfstrahlpumpenVakuum-Technik, 3, 1954, 1-15

W. Bächler und H. G. NöllerFraktionierung und Entgasung in Öl-DiffusionspumpenZ. angew. Physik einschl. Nukleonik, 9,1957, 612-616

H. G. NöllerWeshalb sind systematische Fehler beiSaugvermögensmessungen besondersgroß für Hochvakuumpumpen großer Leistung ?Vakuum-Technik, 12, 1963, 291-293

W. Bächler und H.-J. ForthDie wichtigsten Einflußgrößen bei der Entwicklung von Diffusionspumpen Vakuum-Technik, 13, 1964, 71-75

W. ReicheltBemerkungen zur Arbeitsweise modernerDiffusionpumpenVakuum-Technik, 13, 1964, 148-152

H. G. NöllerTheory of Vacuum Diffusion PumpsHandbook of Physics, Vol.1, Part 6, (pp. 323...419) Ed. A. H. Beck, PergamonPress Ltd., London, W.I. (1966)

G. HerklotzEnddruckversuche mit Diffusionspumpenhohen Saugvermögens und RestgasspektrenVakuum-Technik, 20, 1971, 11 – 14

H. G. NöllerDie Bedeutung von Knudsenzahlen undÄhnlichkeitsgesetzen in Diffusions- undDampfstrahlpumpenVakuum-Technik, 26, 1977, 72-78



R. GöslingTreibmittelpumpenVakuum-Technik, 1980, 163-168

M. WutzGrundlagen zur Bestimmung der charakteristischen Daten von Dampfstrahl-EjektorpumpenVakuum-Technik, 1982, 146-153

H. BayerDampfstrahlpumpenVakuum-Technik, 1980, 169-178

H. BayerVakuumerzeugung durch Dampfstrahl-Vakuumpumpen Vakuum in der Praxis, 1989, 127-135

F. HinrichsAufbau, Betriebsverhalten und Regelbarkeit von Dampfstrahl-Vakuumpumpen Vakuum in der Praxis, 1991, 102-108

2.4 Sorption pumps

G. KienelZur Desorption von Gasen in Getter-Ionenpumpen in „Physik und Technik vonSorptions- und Desorptionsvorgängen beiniederen Drücken“Rudolf A. Lange Verlag, 1963, Esch/Tau-nus, 266-270

W. BächlerIonen-Zerstäuberpumen, ihre Wirkungsweise und AnwendungLeybold-Heraeus GmbH u. Co., in its ownpublishing house, Cologne 1966W. EspeZur Adsorption von Gasen und Dämpfenan MolekularsiebenFeinwerktechnik, 70, 1966, 269-273

G. KienelVakuumerzeugung durch Kondensationund durch SorptionChemikerzeitung / Chem. Apparatur 91,1967, 83-89 und 155-161

H. HochErzeugung von kohlenwasserstoffreiemUltrahochvakuumVakuum-Technik, 16, 1967, 156-158

W. Bächler und H. HenningNeuere Untersuchungen über den Edelgas-Pumpmechnismus von Ionenzerstäuberpumpen des Diodentyps Proc. of the Forth Intern. Vacuum Congress 1968, I. 365-368,Inst. of Physics, Conference Series No. 5,London

H. HenningDer Erinnerungseffekt für Argon bei Trioden-IonenzerstäuberpumpenVakuum-Technik, 24, 1975, 37-43

2.5 Cryopumps and cryoengineering

R. A. HaeferCryo-Pumping456 pages, 1989 Oxford University Press,Oxford

H. Frey und R-A. HaeferTieftemperaturtechnologie, 560 pages,VDI-Verlag, Düsseldorf, 1981

G. Klipping und W. MascherVakuumerzeugung durch Kondensationan tiefgekühlten Flächen, I. KryopumpenVakuum-Technik, 11, 1962, 81-85

W. Bächler, G. Klipping und W. Mascher Cryopump System operating down to 2,5K, 1962 Trans. Ninth National VacuumSymposium, American Vacuum Society,216-219, The Macmillan Company, New York

G. KlippingKryotechnik – Experimentieren bei tiefenTemperaturenChemie-Ingenieur-Technik, 36, 1964, 430-441

M. SchinkmannMesssen und Regeln tiefer Temperaturen,Teil I: Thermodynamische Verfahren Meßtechnik, 81, 1973, 175-181

G. Schäfer, M. SchinkmannMesssen und Regeln tiefer Temperaturen,Teil II: Elektrische Verfahren, Meßtechnik, 82, 1974, 31-38 D00

D00 E 19.06.2001 21:41 Uhr Seite 177



R. Frank et al.Entwicklung von Refrigeratoren für denEinbau in KryopumpenVakuum-Technik, 30, 1981, 134-137

J. J. Scheer und J. VisserAnwendungen von Kryopumpen in derindustriellen Vakuumtechnik Vakuum-Technik, 31, 1982, 34-45

P. DuvalDiffusionspumpen, Turbo-Molekularpum-pen oder Kryopumpen ? – Auswahlkriteri-en für HochvakuumpumpenVakuum-Technik, 31, 1982, 99-105

H. Henning und H.-H. KleinPumpen von Helium mit Refrigerator-Kryopumpen Vakuum-Technik, 34, 1985, 181-184

H.-H. Klein et al.Einsatz von Kryopumpen in ProduktionsanlagenVakuum-Technik, 34, 1986, 203-211

D. Müller und M. SydowKryopumpen im Vergleich mit anderenHochvakuumpumpenVakuum in der Praxis, 2, 1990, 270-274

G. Kiese und G. VoßKryopumpen mit neuartiger Regenerationstechnik Vakuum in der Praxis, 4, 1992, 189-192

2.6 Oil backstreaming

G. LevinA quantitativ appraisal of the backstreaming of forepump oil vaporJ. Vac. Sci. Technol. A 3 (6), 1985, 2212-2213

M. A. Baker and L. LaurensonA quartz crystal microbalance holder forlow Temperature use in vacuumVacuum Vol. 17, (12), 647-648, 1967(Letters to the Editor)

M. A. Baker and W. SteckelmacherThe Measurement of Contamination inVacuum SystemsVuoto, scienza e technologia, Bd.3 , (1/2), 3-17, 1970

J. P. Deville, L. Holland and L. LaurensonMeasurement of the rate of evaporation ofPump oils using a crystal vibrator3rd. Internat. Vac. Congr Stuttgart 153-160, Pergamon Press, Oxford, 1965

L. Laurenson, S. Hickman and R. G. LiveseyRotary pump backstreaming: An analytical appriasal of practical resultsand the factors affecting themJ. Vac. Sci. Technol. A 6 (2), 238-242, 1988

B. D. Power, A. M. I. Mech, E. Crawleyand D. J. CrawleySources, Measurement and Control ofBackstreaming in Oil Vapour VacuumPumpsVacuum, Vol. 4 (4), 415-437, 1957

M. A. BakerA cooled quartz crystal microbalancemethode for measuring diffusion pumpbackstreamingJournal of Scientific Instruments (Journalof Physics E), Series 2, Volume 1, 774-776, 1968

N. S. HarrisDiffusion pump back-streamingVacuum, Vol. 27 (9), 519-530, 1977

M. A. BakerVapour and Gas Measurements in Vacu-um with the Quartz Crystal Microbalancein Vol.1, Proceedings of the ninth Confe-rence on Vacuum Microbalance Techni-ques, „Progress in Vacuum MicrobalanceTechniques“

Th. Gast and E. Robens ed., Heyden & Son Ldt., London, New York,Rheine, 1970

M. A. Baker and L. LaurensonThe use of a quartz crystal microbalancefor measuring vapour backstreamingfrom mechanical pumpsVacuum, Volume 16 (11), 633-637, 1966

R. D. Oswald and D. J. CrawleyA method of measuring back migration ofoil through a baffleVacuum, Vol. 16 (11), 623-624, 1966

M. H. HablanianBackstreaming Measurements above Liquid-Nitrogen Traps Vac. Sci. Tech., Vol. 6, 265-268, 1969

Z. Hulek, Z. Cespiro, R. Salomonovic, M. Setvak and J. VoltrMeasurement of oil deposit resulting frombackstreaming in a diffusion pumpsystem by proton elastic scatteringVacuum, Vol. 41 (7-9), 1853-1855, 1990

M. H. HablanianElimination of backstreaming frommechanical vacuum pumpsJ. Vac. Sci. Technol. A5 (4), 1987, 2612-2615

3. Ultrahigh vacuumtechnology

G. KienelProbleme und neuere Entwicklungen aufdem Ultrahochvakuum-Gebiet VDI-Zeitschrift, 106, 1964, 777-786

G. Kienel und E. WanetzkyEine mehrmals verwendbare Metalldichtung für ausheizbare Uktrahochvakuum-Ventile und FlanschdichtungenVakuum-Technik, 15, 1966, 59-61

H. G. NöllerPhysikalische und technische Vorausset-zungen für die Herstellung und Anwen-dung von UHV-Geräten.“Ergebnisse europäischer Ultrahochvaku-um Forschung”LEYBOLD-HERAEUS GmbH u. Co., in itsown publishing house, Cologne 1968, 49-58

W. BächlerProbleme bei der Erzeugung von Ultrahochvakuum mit modernen Vakuumpumpen. „Ergebnisse europäischer Ultrahochvakuum Forschung“

Leybold-Heraeus GmbH u. Co., in its ownpublishing house, Cologne 1968, 139-148

D00 E 19.06.2001 21:41 Uhr Seite 178



P. Readhead, J. P. Hobson und E. V. KornelsenThe Physical Basis of Ultrahigh VacuumChapman and Hall, London, 1968

E. Bergandt und H. HenningMethoden zur Erzeugung von UltrahochvakuumVakuum-Technik, 25,1970, 131-140

H. WahlDas Hochvakuumsystem der CERN am450 GeV Supersynchrotron und Speichering (SPS)Vakuum in der Praxis, 1989, 43-51

F. GrotelüschenDas UHV-System bei DESY. 1. TeilVakuum in der Praxis, 4, 1991, 266-273

D. TrinesDas Strahlrohrvakuumsystem des Hera-ProtonenringesVakuum in der Praxis, 2, 1992, 91-99

G. Schröder et al.COSV- eine neue Forschungsanlage mitUHV-TechnologieVakuum in der Praxis, 5, 1993, 229-235

W. JacobiDas Vakuumsystem der GSI-Beschleuni-geranlageVakuum in der Praxis, 6, 1994, 273-281

4. Conductances,flanges, valves,etc.

M. KnudsenGesetze der Molekularströmung und derinneren Reibungsströmung der Gasedurch RöhrenAnnalen der Physik, 4th issue, 28, 1909,75-130

P. ClausingÜber die Strömung sehr verdünnter Gasedurch Röhren von beliebiger Länge Annalen der Physik, 5th issue, 12, 1932, 961-989

W. RöllingerDie Verwendung von Klammerflanschen inder VakuumtechnikVakuum-Technik, 13, 1964, 42-45

H. HochAusheizbare Verbindungen an Hoch-vakuum-ApparaturenVakuum-Technik, 10, 1961, 235-238

W. Bächler und I. WikbergDual Seal Bakable Section Valves of theCERN Intersection Storage RingVacuum, 21, 1971, 457-459

K. TeutenbergUHV-Ganzmetallventile großer NennweiteVakuum-Technik, 21, 1972, 169-174

H. HenningThe approximate calculation of transmission probabilitiesVacuum, 28, 1978, No. 3, Seite 151

G. KühnGasströme durch Spalte im GrobvakuumVakuum-Technik, 33, 1984, 171-175

R. Haberland und B. VogtUHV-Ventil für extrem viele SchließzyklenVakuum-Technik, 34, 1985, 184-185

A. SeleVakuum-Ventile (VAT)Vakuum in der Praxis, 1, 1989, 206-212

L. FikesBerechnung von Auspumpkurven mitHilfe der Analogie von Gasstrom undelektrischem StromVakuum in der Praxis, 4, 1992, 265-268

W. HerzZuverlässige Flanschverbindung imAnwendungsgebiet der Tieftemperatur-und VakuumtechnikVakuum-Technik, 29, 1980, 67-68

5. Measurement oflow pressures

C. Meinke und G. ReichVermeidung von Fehlmessungen mit demSystem McLeod-KühlfalleVakuum-Technik, 12, 1963, 79-82

P. A. Readhead and J. P. HobsonTotal Pressure Measurem. below 10–10Torr with Nonmagnetic Ionisation GaugeBrit. J. Appl. Phys., 16, 1965, 1555-1556

C. Meinke und G. ReichComparison of Static and Dymanic Calibration Methods for Ionisation GaugesJ. Vac. Sci. Techn., 4, 1967, 356-359

G. Reich und W. SchulzProbleme bei der Verwendung von Ionisations-Vakuummetern im Druck-bereich oberhalb 10–2 TorrProc. of the Fourth Intern. Vacuum Congress, 1968, II. Inst. of Physics Conference Series No.6, London, 661-665

G. ReichProbleme bei der Messung sehr niedrigerTotal- und Partialdrücke„Ergebnisse europäischer Ultrahoch-vakuum Forschung“Leybold-Heraeus GmbH u. Co., in its ownpublishing house, Cologne 1968, 99-106

A. Barz and P. Kocian Extractor Gauge as a Nude SystemJ. Vac. Sci Techn. 7, 1970, 1, 200-203

U. Beeck and G. ReichComparison of the Pressure Indication of aBayard-Alpert and an Extractor GaugeJ. Vac. Sci. and Techn. 9, 1972, 1,126-128

U. BeeckUntersuchungen über die Druckmessun-gen mit Glühkathoden-Inisations-Vakuum-metern im Bereich größer als 10–3 TorrVakuum-Technik, 22, 1973, 16-20

G. ReichÜber die Möglichkeiten der Messung sehrniedriger DrückeMeßtechnik, 2, 1973, 46-52

D00

D00 E 19.06.2001 21:41 Uhr Seite 179



G. ReichSpinning rotor viscosity gauge; a transferstandard for the laboratory or an accurategauge for vacuum process controlJ. Vac. Sci. Technol., 20 (4), 1982, 1148-1152

G. ReichDas Gasreibungs-Vakuummeter VISCOVAC VM 210Vakuum-Technik, 31, 1982, 172-178

G. Grosse and G. MesserCalibration of Vacuum Gauges at Pressures below 10–9 mbar with a molecular beam methodVakuum-Technik, 30, 1981, 226-231

Chr. Edelmann et al.: Möglichkeiten derMeßbereichserweiterung bei Glühkatho-den-Ionisationsmanometern (numerousreferences to relevant literature)Vakuum-Technik, 31, 1982, 2-10

Chr. EdelmannStand und Entwicklungstendenzen derTotaldruckmessung in der Vakuum-Tech-nikVakuum-Technik, 33, 1984, 162-180

J. K. FremereyDas GasreibungsvakuummeterVakuum-Technik, 36, 1987, 205-209

G. MesserKalibrierung von VakuummeternVakuum-Technik, 36, 1987, 185-192

G. Messer und W. GrosseEntwicklung der Vakuum-Metrologie in derPTB (numerous references to relevant li-terature)Vakuum-Technik, 36, 1987, 173-184

G. ReichIndustrielle VakuummeßtechnikVakuum-Technik, 36, 1987, 193-197

L. Schmidt und E. EichlerDie Praxis einer DKD-KalibrierstelleVakuum-Technik, 36, 1987, 78-82

C. KündigVakuummeßgeräte für TotaldruckVakuum in der Praxis, 2, 1990, 167-176

Chr. EdelmannGlühkahtoden-Ionisationsmanometer fürhohe Drücke im Vakuumbereich

Vakuum in der Praxis, 3, 1991, 290-296

M. Ruschitzka and W. JitschinPhysikalische Grundlagen des WärmeleitungsvakuummetersVakuum in der Praxis, 4, 1992, 37-43

T. KoopmannNeue Trends in der Vakuum-MeßtechnikVakuum in der Praxis, 5, 1993, 249-254

Chr. EdelmannDie Entwicklung der Totaldruckmessungim UHV- und ExtremvakuumbereichVakuum in der Praxis, 6, 1994, 213-219

W. JitschinKalibrierung, Abnahme und Zertifizierung(with numerous references to relevant li-terature)Vakuum in der Praxis, 6, 1994, 193-204

W. JitschinObere Meßbereichsgrenze von Glühkato-den-IonisationsvakuummeternVakuum in Forschung und Praxis, 7, 1995, 47-48

F. Mertens et al.Einfluß von Gasadsorbaten auf die Eigen-schaften eines Glühkatoden-Ionisations-vakuummeters mit axialer Emission nachChen und SuenVakuum in der Praxis, 7, 1995, 145-149

6. Pressure monitoring, controland regulation

K. G. MüllerBetriebsüberwachung, Steuerung undAutomatisierung von VakuumanlagenChemie-Ingenieur-Technik, 35, 1963, 73-77

G. KienelElektrische Schaltgeräte der Vakuumtech-nikElektro-Technik, 50, 1968, 5-6

A. Bolz, H. Dohmen und H.-J. SchubertProzeßdruckregelung in der VakuumtechnikLeybold Firmendruckschrift 179.54.01

H. DohmenVakuumdruckmessung und -Regelung inder chemischen VerfahrenstechnikVakuum in der Praxis, 6,1994, 113-115

N. PöchheimDruckregelung in VakuumsystemenVakuum in Forschung und Praxis, 7, 1995, 39-46

R. Heinen und W. SchwarzDruckregelung bei Vakuumprozessendurch umrichtergespeiste RootspumpenVakuum-Technik, 35, 1986, 231-236

7. Mass spectrometergas analysis at lowpressures

H. HochTotal- und Partialdruckmessungen beiDrücken zwischen 2 · 10–10 und 2 · 10–2 TorrVakuum-Technik, 16, 1967, 8-13

H. JungePartialdruckmessung und Partialdruck-meßgeräteG-I-T May 1967, 389-394 and June 1967, 533-538

A. KlugeEin neues Quadrupolmassenspektrometermit massenunabhängiger Empfindlichkeit Vakuum-Technik, 23, 1974, 168-171

S. BurzynskiMicroprocessor controlled quadrupolemass spectrometerVacuum, 32, 1982, 163-168

W. Große BleyQuantitative Gasanalyse mit dem Quadrupol MassenspektrometerVakuum-Technik, 38, 1989, 9-17

A. J. B. RobertsonMass SpectrometryMethuen & Co, Ltd., London, 1954

C. Brunee und H. VoshageMassenspektrometrieKarl Thiemig Verlag, München, 1964

D00 E 19.06.2001 21:41 Uhr Seite 180



A. Cornu and R. MassotCompilation of Mass Spectral DataHeyden and Son Ltd., London, 1966

P. DawsonQuadrupole Mass SpectroscopyElsevier, Amsterdam, 1976

J. BackusChap. 11 in “Characteristics of ElectricalDischarges in Magnetic Fields”National Nuclear Energy Series, Div. I,Vol. 5, McGraw-Hill Book Company Inc.,New York, 1949

J. BackusUniversity of California Radiation Laboratory Report, RL 20.6.36, Mar. 1945.

8. Leaks and Leakdetection

8.1 Mass spectrometer leakdetection

G. KienelLecksuche an Vakuumanlagen auf elektrischem WegeElektrotechnik, 49, 1967, 592-594

U. BeeckMöglichkeiten und Grenzen der automati-schen Lecksuche im Bereich unter 10–8 Torr. l/sVakuum-Technik, 23, 1974, 77-80Lecksuche an ChemieanlagenDechema Monographien (Ed. H. E. Bühlerand K. Steiger), Vol. 89, Verlag Chemie,Weinheim / New York

W. JansenGrundlagen der Dichtheitsprüfung mitHilfe von TestgasenVakuum-Technik, 29, 1980, 105-113

K. PaascheLecksuche an ChemieanlagenVakuum-Technik, 29, 1980, 227-231

H. B. Bürger Lecksuche an Chemieanlagen mit He-Massenspektrometer-LecksuchernVakuum-Technik, 29, 1980, 232-245

Chr. FallandEin neuer Universal-Lecksucher mit luftgekühlter Turbo-MolekularpumpeVakuum-Technik, 29, 1980, 205-208

W. JansenGrundlagen der Dichtheitsprüfung mitHilfe von technischen GasenVakuum-Technik, 29, 1980, 105-113

H. MennengaDichtheitsprüfung von KleinteilenVakuum-Technik, 29, 1980, 195-200

Chr. FallandEntwicklung von He-Lecksuchtechnikenfür UHV-Systeme großer Beschleuniger-und SpeicherringeVakuum-Technik, 30, 1981, 41-44

W. Engelhardt et al.Lecksuchanlagen in der IndustrieVakuum-Technik, 33, 1984, 238-241

G. Sänger et al.Über die Lecksuche bei RaumfahrzeugenVakuum-Technik, 33, 1984, 42-47

W. Jitschin et al.He-Diffusionslecks als sekundäre Normalefür den GasdurchflußVakuum-Technik, 36, 1987, 230-233

W. Große BleyModerne He-Leckdetektoren unterschiedlicher Prinzipien im prakti-schen EinsatzVakuum in der Praxis, 1, 1989, 201-205

H. D. BürgerLecksucher (with references to relevantliterature)Vakuum in der Praxis, 2, 1990, 56-58

W. FuhrmannEinführung in die industrielle Dichtheits-prüftechnikVakuum in der Praxis, 3, 1991, 188-195

W. FuhrmannIndustrielle Dichtheitsprüfung – ohneTestgas nach dem Massenspektrometrie-verfahrenVakuum in Forschung und Praxis, 7, 1995, 179 -182

8.2 Leak detection with halogen leak detectors

H. Moesta und P. SchuffÜber den thermionischen HalogendetektorBerichte der Bunsengesellschaft für physikaische Chemie, Bd. 69, 895-900, 1965Verlag Chemie, GmbH, Weinheim, Bergstraße

J. C. Leh and Chih-shun LuUS Patent Nr. 3,751,968Solid State Sensor

9 Film thicknessmeasurement andcontrol

G. Z. SauerbreyPhys. Verhandl. 8, 113, 1957

G. Z. SauerbreyVerwendung von Schwingquarzen zurWägung dünner Schichten und zur MikrowägungZeitschrift für Physik 155, 206-222, 1959L. Holland, L. Laurenson and J. P. DevilleUse of a Quartz Crystal Vibrator in Vacu-um Destillation InvstigationsNature, 206 (4987), 883-885, 1965

R. BechmannÜber die Temperaturabhängigkeit der Frequenz von AT- und BT-QuarzresonatorenArchiv für Elektronik und Übertragungs-technik, Bd. 9, 513-518, 1955

K. H. Behrndt and R. W. LoveAutomatic control of Film Deposition Ratewith the crystal oscillator for preparationof alloy films.Vacuum 12 ,1-9, 1962

P. LostisAutomatic Control of Film Deposition Ratewith the Crystal Oscillator for Preparationfo Alloy Films.Rev. Opt. 38, 1 (1959)

D00

D00 E 19.06.2001 21:41 Uhr Seite 181



K. H. BehrndtLongterm operation of crystal oscillatorsin thin film deposition J. Vac. Sci. Technol. 8, 622 (1971)

L. Wimmer, S. Hertl, J. Hemetsberger andE. BenesNew method of measuring vibrationamplitudes of quartz crystals. Rev. Sci. Instruments 55 (4) , 608, 1984

P. J. Cumpson and M. P. SeahMeas. Sci. Technol., 1, 548, 1990

J. G. Miller and D. I. BolefSensitivity Enhancement by the use ofAcoustic Resonators in cw UltrasonicSpectroscopy.J. Appl. Phys. 39, 4589, (1968)

J. G. Miller and D. I. BolefAcoustic Wave Analysis of the Operationof Quartz Crystal Film Thickness Moni-tors.J. Appl. Phys. 39, 5815, (1968)

C. Lu and O. LewisInvestigation of Film thickness determination by oscillating quartz resonators with large mass load.J. Appl. Phys. 43, 4385 (1972)

C. LuMas determination with piezoelectricquartz crystal resonators.J. Vac. Sci. Technol. Vol. 12 (1), 581-582, 1975

A. WajidU.S. Patent No. 505,112,642 (May 12, 1992)

C. HurdU.S. Patent No. 5,117,192 (May 26,1992)

E. BenesImproved Qartz Crystal MicrobalanceTechniqueJ. Appl. Phys. 56, (3), 608-626 (1984)

C. J. WilsonVibration modes of AT-cut convex quartzresonators.J. Phys. d 7, 2449, (1974)

H. F. Tiersten and R. C. SmytheAn analysis of contowced crystal resonators operating in overtones ofcoupled thickness shear and thickness

twist.J. Acoustic Soc. Am. 65, (6) 1455, 1979

R. E. Bennett, C. Rutkoeski and L. A. TaylorProceedings of the Thirteenth AnnualSymposium on Frequency Controll, 479, 1959

Chih-shun LuImproving the accuracy of Quartz csystalmonitorsResearch/Development, Vol. 25, 45-50,1974, Technical Publishing Company

A. WajidImproving the accuracy of a quartz crystalmicrobalance with automatic determinati-on of acoustic impedance ratio.Rev. Sci. Instruments, Vol. 62 (8), 2026-2033, 1991

D. Graham and R. C. LanthropThe Synthesis fo Optimum Transient Response: Criteria and Standard FormsTransactions IEEE, Vol. 72 pt. II, Nov. 1953

A. M. Lopez, J. A. Miller, C. L. Smith andP. W. MurrillTuning Controllers with Error-Integral CriteriaInstrumentation Technology, Nov. 1969

C. L. Smith and P. W. MurrilA More Precise Method for Tuning ControllersISA Journal, May 1966

G. H. Cohen and G. A. CoonTheoretical considerations of RetardedControlTaylor Technical Data Sheet Taylor Instrument Companies, Rochester, New York

J. G. Ziegler and N. B. NicholsOptimum Settings for Automatic Controllers Taylor Technical Data Sheet No. TDS10A100, Taylor Instrument Companies,Rochester, New York

C. Lu and A. W. CzandernaApplication of Piezoelectric Quarz CrystalMicrobalances (Vol.7 of: Methodes andPhenomena, Their Applications in Sienceand Technology)Elesvier, Amsterdam, Oxford, New York,Tokio, 1984

G. Simmons and H. WangSingle Crystal Elastic Constants and Calculated Aggregate Properties – A HandbookThe MIT Press, Cambridge, Massachu-setts, 1971

C. D. Stockbridgein Vol. 5 “Vacuum Microbalance Techniques” K. Behrndt, editor, PlenumPress, Inc., New York, 1966

S. SotierSchwingquarz-SchichtdickenmessungVakuum in der Praxis 1992, 182-188

10. Materials andmaterial processing

W. EspeWerkstoffkunde der HochvakuumtechnikVol. 1 1959, Vol. 2 1960, Vol. 3 1961, VEB Deutscher Verlag der Wissenschaf-ten, Berlin

W. EspeWerkstoffe für trennbare metallische Ver-bindungen der UltrahochvakuumtechnikFeinwerktechnik, 68, 1964, 131-140

W. EspeSynthetische Zeolithe und ihre Verwen-dung in der HochvakuumtechnikExperimentelle Technik der Physik, XII,1964, 293-308

H. AdamAllgemeiner Überblick über die Werkstoffeder Vakuumtechnik und deren AuswahlHaus der Technik Vortragsveröffentlichun-gen “Werkstoffe und Werkstoffverbindun-gen in der Vakuumtechnik” H. 172, Vul-kan-Verlag, Dr. W. Classen, Essen, 1968,4 – 13

D00 E 19.06.2001 21:41 Uhr Seite 182



K. VerfußBessere Oberflächenvergütung durchElektropolieren – am Beispiel der Vaku-um-TechnikVDI-Berichte, 183, 1972, 29-34

K. VerfußSchweißen und HartlötenHaus der Technik, Vortragsveröffentli-chungen “Werkstoffe und Werkstoffver-bindungen in der Vakuumtechnik”, H. 172Vulkan-Verlag Dr. W. Classen, Essen,1968, Seiten 39 -49

Chr. EdelmannGasabgabe von Festkörpern im VakuumVakuum-Technik, 38, 1989, 223-243

R. FritschBesonderheiten vakuumdichrter SchweißverbindungenVakuum-Technik, 38, 1989, 94-102

H. HenningVakuumgerechte Werkstoffe und Verbindungstechnik, Part 1Vakuum in der Praxis, 2, 1990, 30-34

R. FritschVakuumgerechte Werkstoffe und Verbindungstechnik, Part 2Vakuum in der Praxis, 2, 1990, 104-112

M. MühlloffVakuumgerechte Werkstoffe und Verbindungstechnik, Part 3Vakuum in der Praxis, 2, 1990, 179-184

11. Dictionaries

F. WeberElsevier’s Dictionary of High Vacuum Science and Technology (German,English, French, Spanish, Italian, Russian)Elsevier Verlag 1968

Hurrle / Jablonski / RothTechnical Dictionary of Vacuum Physicsand Vacuum Technology (German, English, French, Russian)Pergamon Press Verlag, Oxford, 1972

D00

D00 E 19.06.2001 21:41 Uhr Seite 183

IndexFundamentals of Vacuum Technology


13. Index Absolute pressure . . . . . . . . . . . . . . . .7Absorption isotherms . . . . . . . . . . . . .45Absorption pumps . . . . . . . . . . .45, 139Absorption traps . . . . . . . . . . . . . . . .33Accessories for rotary displacement pumps . . . . . . . . . . . . .33Active oscillator . . . . . . . . . . . .122, 123Adjustment and calibration of vacuum gauges . . . . . . . . . . . . . . . . .88Adsorption pumps, Instructions for operation . . . . .139, 140Aggressive vapors . . . . . . . . . . . . . .135AGM (aggressive gas monitor) . . . . .94Air, atmospheric . . . . . . . . . . . . . . . . .10ALL·ex pumps . . . . . . . . . . . . . . .28, 30Ambient pressure . . . . . . . . . . . . . . . . .7Amonton's law . . . . . . . . . . . . . . . . . .11Anticreep barrier . . . . . . . . . . . . . . . .40Anti-suckback valve . . . . . . . . . . . . . .19APIEZON AP 201 . . . . . . . . . . . .39, 160Atmospheric air . . . . . . . . . . . . . . . . .10Atmospheric air, composition . . . . . .145Atmospheric pressure . . . . . . . . . . . . .7Atomic units . . . . . . . . . . . . . . . . . .170Autoc ontrol tune . . . . . . . . . .125, 126Automatic protection, monitoring and control of vacuum systems . . . . .83Auto-Z match technique . . . . . . . . . .122Avogadro's law . . . . . . . . . . . . . . . . .11Avogadro's number (Loschmidt number) . . . . . . . . . . . . . . . . . . .12, 143Backing line vessel . . . . . . . . . . . . . . .64Baffle . . . . . . . . . . . . . . . . . . . . . . . . .36Baffles (vapor barriers) . . . . . .36, 37, 39Baking (degassing) . . . . . . . .61, 68, 141Barrier gas operation . . . . . . . . . . . . .44Basis SI units . . . . . . . . . . . . . . . . . .170Bath cryostats . . . . . . . . . . . . . . . . . .49Bayard-Alpert gauge . . . . . . . . . . . . . .80Boat (thermal evaporator) . . . . . . . . .128Boltzmann constant . . . . . . . . . .11, 143Bombing-Test (storing under pressure) . . . . . . . . . . . . . . . . . . . . .119Booster (oil-jet) pump . . . . . . . . .38, 46Bourdon vacuum gauge . . . . . . . . . . .72Boyle-Mariotte law . . . . . . . . . . . . . . .11Break down voltage (Paschen curve for air) . . . . . . . . . . .163Bubble (immersion) test . . . . . . . . .109Calibration curves of THERMOVAC gauges . . . . . . . . . . . . .77Calibration inspection . . . . . . . . . . . . .81Coating sources . . . . . . . . . . . . . . . .128Capacitance diaphragm gauges . . . . .73Capsule vacuum gauge . . . . . . . . . . .772

Causes of faults if/when the desired ultimate pressure is not achieved . . . . . . . . . . . . . . . . .134Ceramic ball bearings (hybrid ball bearings) . . . . . . . . . . . . . . . . . . . . . .42CF-flange (conflathflange) . . . . . . . . .68Changing the molecular sieve . . . . . .139Charles' law (Gay-Lussac's law) . . . . .11Chemical resistance of elastomer gaskets . . . . . . . . . . . . . . .149, 150, 151Chemical vapor deposition (CVD) . . .129Choked flow, critical pressuredifference . . . . . . . . . . . . . . . . . . . . . .12CIS (closed ion source) . . . . . . . . . . .93Clamp flange . . . . . . . . . . . . . . . . . . .67Classification of vacuum pumps . . . . .16Clausius-Clapeyron equation . . . . . . .11Claw pump . . . . . . . . . . . . . . . . . . . .27Closed ion source, (CIS) . . . . . . . . . . .93Coating of parts . . . . . . . . . . . . . . . .130Coating thickness regulation . . . . . . .125Cold cap baffle . . . . . . . . . . . . . . . . . .39Cold cathode ionization vacuum gauge . . . . . . . . . . . . . . . . . . . . . . . .77Cold head . . . . . . . . . . . . . . . . . . . . .50Cold surfaces, bonding of gases to . . . . . . . . . . . . . . . . . . . . . . .51Cold traps . . . . . . . . . . . . . . . . . . . . .39Collision frequency . . . . . . . . . . . . . . .10Collision rate . . . . . . . . . . . . . . . . . . .10Common solvents . . . . . . . . . . . . . .146Compression . . . . . . . . . . . . . . . . . . .43Compression vacuum gauges . . . . . . .74Condensate traps . . . . . . . . . . . . . . . .33Condensers . . . . . . . . . . . . . . . .33, 174Conductance . . . . . . . . . . . . . . . . .9, 13Conductance of openings . . . . . . . . .179Conductance of piping . . . . . . . . . . . . .14, 155, 161, 179Conductance, nomographic determination . . . . . . . . . . . . . . . . . . .14Conductances, calculation of . . . . . . .13Connection of leak detectors to vacuum systems . . . . . . . . . . . . . . .116Contamination of vacuum sensors . .139Contamination of vacuum vessels . . .134Continous flow . . . . . . . . . . . . . . . . . .12Continous flow cryopumps . . . . . . . . .49Continuum theory . . . . . . . . . . . . . . .11Conversion of leak rate units . . . . . .106Conversion of leak rate units . . . . . .144Conversion of pressure units . . . . . .142Conversion of pV-throughput units . .144Corrosion protection . . . . . . . . .135, 136Counter-flow leak detector . . . . . . . .115Cracking pattern . . . . . . . . . . . . . . .136Critical pressure difference (choked flow) . . . . . . . . . . . . . . . . . . .12Crossover value . . . . . . . . . . . . . . . . .53

Cryocondensation . . . . . . . . . . . . . . .52Cryopumps . . . . . . . . . . . . . . . .49, 177Cryosorption . . . . . . . . . . . . . . . . . . .52Cryotrapping . . . . . . . . . . . . . . . . . . .52Crystal Six . . . . . . . . . . . . . . . . . . .120Cut in (start) pressure . . . . . .44, 54, 55CVD (chemical vapor deposition) . . . . . . . . . . . . . . . .128, 129Dalton's law . . . . . . . . . . . . . . . . . . . .11Danger classes of fluids . . . . . . . . . .148Data storage coating . . . . . . . . . . . .132DC 704, DC 705 (Silicone oils) . . . . . .39Degassing of the pump oil . . . . . . . . .37Derived coherent and not coherent SI units with special names and symbols . . . . . . . . . . . . .170Detection limit (leak detectors) . . . . .111Determination of a suitable backing pump . . . . . . . . . . . . . . . . . .64Determination of pump down time from Nomograms . . . . . . . . . . . .65Determination of pump sizes . . . . . . .62DI series diffusion pumps . . . . . . . . .38Diaphragm contoller, examples of application . . . . . . . . . . . . . . . .86, 87Diaphragm vacuum gauges . . . . . . . .72Diaphragm vacuum pumps . . . . . . . . .17DIAVAC diaphragm vacuum gauge . . .72DIFFELEN, light, normal, ultra . . .39, 160Diffusion / vapor-jet pumps, Instructions for operation . . . . . . . . .139Diffusion pumps . . . . . . . . . . . . . . . .36Diode-type sputter ion pumps . . . . . .46Direct-flow leak detector . . . . . . . . . .115Discharge filters . . . . . . . . . . . . . . . . .33Displacement pumps . . . . . . . . .16, 174DIVAC vacuum pump . . . . . . . . . . . . .17DKD (Deutscher Kalibrierdienst) German calibration service . . . . . . . . .81Dry compressing rotary displacement pumps . . . . . . . . . . . . .24Dry processes . . . . . . . . . . . . . . . . . .57Drying of paper . . . . . . . . . . . . . . . . .67Drying processes . . . . . . . . . . . . . . . .60Drying processes, selection of pumps for . . . . . . . . . . . . . . . . . . . . .66DRYVAC-Pumps . . . . . . . . . . . . . . . .28D-Tek . . . . . . . . . . . . . . . . . . . . . . .110Duo seal (sealing passage) . . .17, 18, 19Dust separator (dust filter) . . . . . . . . .33Dynamic expansion method . . . . . . . .82Ecotec II . . . . . . . . . . . . . . . . . . . . .113Effective pumping speed . . . . . . . .33, 62Elastomer gaskets . . . .69, 149, 150, 151Electrical break down voltage(Paschen curve air) . . . . . . . . . . . . .163Electron beam evaporators (electron guns) . . . . . . . . . . . . . . . .129Envelope test . . . . . . . . . . . . . .118, 119

D00 E 19.06.2001 21:41 Uhr Seite 184

Index Fundamentals of Vacuum Technology


Envelope test (concentration measurement) . . . . . . . . . . . . . . . . .118Evacuation in the rough / medium / high vacuum region . . . . . . . .62, 63, 64Evacuation of gases / vapors . . . . . . .66Evaluating spectra . . . . . . . . . . . . . . .96Expansion method static / dynamic . . . . . . . . . . . . . . . . . . . .81, 82Extractor ionization vacuum gauge . . . . . . . . . . . . . . . . . . . . . . . .80Fast regeneration (partial regeneration) . . . . . . . . . . . . . . . . . . .54Fingerprint . . . . . . . . . . . . . . . . . . . . .96Flanges and their seals . . . . . . . .67, 179Floating zero-point suppression . . . .111Fluid entrainment pumps . . . . . .35, 177Foam spray leak test . . . . . . . . . . . .109Fractionation of pump fluids . . . . . . . .37Fragment distribution pattern . . . . . . .98Fundamental pressure measurement methods . . . . . . . . . . . .81Gas analysis . . . . . . . .89, 100, 101, 180Gas ballast . . . . . . . . . . . . . .21, 22, 111Gas composition as a function of altitude . . . . . . . . . . . . . . . . . . . .155Gas constant, general (molar) .7, 11, 143Gas density . . . . . . . . . . . . . . . . . . . . .7Gas dependent pressure reading, vacuum gauges with . . . . . . . . . . . . .75Gas discharge . . . . . . . . . . . . . . .46, 78Gas independent pressure reading, vacuum gauges with . . . . . . . . . . . . .72Gas laws . . . . . . . . . . . . . . . . . . . . . .11Gas locks . . . . . . . . . . . . . . . . . . . . . .69Gas sorption (pumping) of vacuum gauges . . . . . . . . . . . . . .78, 79Gas storage in the oil of rotary vane pumps . . . . . . . . . . . . . . . . . . .111Gaskets . . . . . . . . . . .68, 149, 150, 151Gay-Lussac's law . . . . . . . . . . . . . . . .11General gas constant (Molar gas constant) . . . . . . .7, 11, 143Getter pumps . . . . . . . . . . . . . . . . . . .45Glass coating . . . . . . . . . . . . . . . . . .132Halogen leak detector . . . . . . . .110, 181Helium leak detectors with 180°sector mass spectrometer . . . . . . . .114Helium spray equipment . . . . . . . . . .118Helium standard leak rate . . . . . . . .106High frequency vacuum test . . . . . . .109High pressure ionization vacuum gauge . . . . . . . . . . . . . . . . . . . . . . . .80High vacuum range . . . . . . . . . . .62, 63HLD 4000 . . . . . . . . . . . . . . . . . . . .110HO-factor (diffusion pumps) . . . . . . .37Hot cathode ionization vacuum gauge . . . . . . . . . . . . . . . . . . . . . . . .78HY.CONE pumps . . . . . . . . . . . . . . . .44

Hybrid ball bearings (Ceramic ball bearings) . . . . . . . . . . . . . . . . . . . . .42Hydrocarbon-free vacuum . . . . . .39, 60IC 5 . . . . . . . . . . . . . . . . . . . . . . . .127Ideal gas law . . . . . . . . . . . . . . . . . 7, 11 Impingement rate . . . . . . . . . . . . . . . .10Industrial leak testing . . . . . . . . . . . .119Influence of magnetic / electrical fields . . . . . . . . . . . . . . . . .141Internal compression (claw pumps) . .28Inside-out leak . . . . . . . . . . . . . . . . .107Integral leak rate . . . . . . . . . . . . . . .107Internal reflow (roots pumps) . . . . . . .24Ion desorption effect . . . . . . . . . . . . .79Ion sputter pumps . . . . . . . . . . . .45, 46Ionization vacuum gauge for higher pressures up to 1 mbar . . . . . .80Ionization vacuum gauges . . . . . . . . .77Ionization, specific (gas analysis) . . . .96Isotopes . . . . . . . . . . . . . . . . . . . . . .96Kammerer compression vacuum gauge . . . . . . . . . . . . . . . . . . . . . . . .74Kinetic gas theory . . . . . . . . . . . . . . .11Kinetic of gases, diagram of . . . . . . .154Kinetic of gases, formulas . . . . . . . .143Knudsen flow . . . . . . . . . . . . . . . . . . .13Krypton 85 test . . . . . . . . . . . . . . . .109Laminar flow . . . . . . . . . . . . . . . . . . .12Langmuir-Taylor-effect . . . . . . . . . .111Laval nozzle . . . . . . . . . . . . . . . . . . . .38Leak detection . . . . . . . . . . . . .104, 181Leak detection using Helium leak detectors . . . . . . . . . . . . . . . . .117Leak detection without leak detector .107Leak detection, leak test . . . . . . . . . .104Leak detectors with 180° sector mass spectrometer . . . . . . . . . . . . .114Leak detectors with mass spectrometer . . . . . . . . . . . . . .110, 181Leak detectors with quadrupole mass spectrometer . . . . . . . . . . . . .113Leak detectors, how they work . . . . .110Leak rate, hole size, conversion . . . . . . . . . .9, 104, 105, 106Leak test (chemical reactions, dye penetration) . . . . . . . . . . . . . . . . . . .110Leak test, using vacuum gauges sensitive to the type of gas . . . . . . . .108LEYBODIFF-Pumps . . . . . . . . . . . . . .37INFICON Quartz crystal controllers . . . . . . . . . . . . . . . . . . . .127Line width . . . . . . . . . . . . . . . . . . . . .94Linearity range of quadrupole gauges . . . . . . . . . . . . . . . . . . . . . . .102Liquid ring pumps . . . . . . . . . . . . . . .17Liquid sealed rotary displacement pumps . . . . . . . . . . . . . . . . . . . . . . . .17Liquid-filled vacuum gauges . . . . . . . .74Literature references . . . . . . . .174 – 183

LN2 cold traps . . . . . . . . . . . . . . . . . .40Local leak rate . . . . . . . . . . . . . . . . .106Loschmidt's number (Avogadro constant) . . . . . . . . . .11, 143Magnetic suspension (bearings) . . . . .42Mass flow . . . . . . . . . . . . . . . . . .8, 104Mass flow (leak detection) . . . . . . . .104Mass range . . . . . . . . . . . . . . . . . . .195Mass spectrometer, general, historical . . . . . . . . . . . . . . . . . .89, 180Maximum backing pressure (critical forevacuum pressure) . . . . . .37McLeod vacuum gauge . . . . . . . . . . .74Mean free path . . . . . . . . . . .9, 142, 159Measuring range of vacuum gauges . . . . . . . . . . . . . . . . . . . . . . .161Measuring range, favorable . . . . . . . .70Measuring ranges of vacuum gauges . . . . . . . . . . . . . . . . . . . . . . .162Measuring vacuum, vacuum gauges . . . . . . . . . . . . . . . . . . . .70, 179Medium vacuum adsorption trap . . . .33MEMBRANOVAC . . . . . . . . . . . . . . . .73Mercury (pump fluid) . . . . . . .36, 39,160Mode-lock oscillator . . . . . . . . . . . .123Molar gas constant . . . . . . . . .7, 11, 143Molar mass (molecular weight) . . . . . . . . . .7, 10, 11Molecular flow . . . . . . . . . . . . . . .12, 13Molecular sieve . . . . . . . . . . . . .45, 139Monolayer . . . . . . . . . . . . . . . . . . . . .10Monolayer formation time . . . .10, 13, 61National standards, resetting to . . . . .81NEG pumps (non evaporable getter pumps) . . . . . . . . . . . . . . .45, 48Neoprene . . . . . . . . . .68, 149, 150, 151Nitrogen equivalent . . . . . . . . . . .71, 78Nominal internal diameter and internal diameter of tubes . . . . . . . . .146Nomogram . . . . . . . . . . . . . . . . . . . .65Nomogram: conductance of tubes / entire pressure range . . . . . .158Nomogram: conductance of tubes / laminar flow range . . . . . . . .155Nomogram: conductance of tubes / molecular flow range . . .155, 162Nomogram: pump down time / medium vacuum, taking in account the outgasing from the walls . . . . . . . . . . . . . . . . . . . . . . . .159Nomogram: pump down time / rough vacuum . . . . . . . . . . . . . . . . .156Non evaporable getter (NEG) pumps . . . . . . . . . . . . . . . . . . . . .45, 48Non gas-tight area . . . . . . . . . . . . .1104Nude gauge (nude system) . . . . . . . . .72Oil backstreaming . . . . . . . . . . . .39, 178Oil change . . . . . . . . . . . . . . . . . . . .137Oil consumption . . . .135, 136, 137, 138

D00

D00 E 19.06.2001 21:41 Uhr Seite 185

IndexFundamentals of Vacuum Technology


Oil contamination . . . . . . . . . . . . . . .135Oil diffusion pumps . . . . . . . . . . . . . .36Oil sealed rotary displacement pumps . . . . . . . . . . . . . . . . . . . . . . . .18Oil vapor ejector vacuum pumps . . . .36Oil-free (hydrocarbon-free) vacuum . . . . . . . . . . . . . . . . . . . .39, 60Oils (pump fluids) . . . . . . . . . . . . . . .39Open (normal) ion source . . . . . . . . . .90Optical coatings . . . . . . . . . . . . . . . .131Oscillation displacement pumps . . . . .17Oscillator, ( active, mode-lock) .122, 123Outgasing of materials . . . . . . . . . . .145Outgasing rate (referred to surface area) . . . . . . . . . . . . . . . . .9, 61Outside-in leak . . . . . . . . . . . . . . . . .107Overpressure . . . . . . . . . . . . . . . . . . . .7Oxide-coated cathodes . . . . . . . . .78, 90Partial final pressure . . . . . . . . . . . . .74Partial flow opeartion . . . . . . . . . . . .115Partial flow ratio . . . . . . . . . . . . . . . .116Partial pressure . . . . . . . . . . . . . . . . . .7Partial pressure measurement . . . . .100Partial pressure regulation . . . . . . . .101Particle number density . . . . . . . . . . . .7Paschen curve . . . . . . . . . . . . . . . . .163Penning vacuum gauges . . . . . . . . . .77Perbunan . . . . . . . . .68, 149 - 151, 161Period measurement . . . . . . . . . . . .122Permissible pressure units . . . . . . . .142Phase diagram of water . . . . . . . . . .164Photons . . . . . . . . . . . . . . . . . . . . . . .79PIEZOVAC . . . . . . . . . . . . . . . . . . . . .73Pirani vacuum gauge . . . . . . . . . . . . .76Plastic tent (envelope) . . . . . . . . . . .118Plate baffle . . . . . . . . . . . . . . . . . . . . .39PNEUROP . . . . . . . . . . . . . . .165 – 168PNEUROP flanges . . . . . . . . . . . . . . .68Poiseuille flow . . . . . . . . . . . . . . . . . .12Poisson's law . . . . . . . . . . . . . . . . . .11Positive pressure methode (leak detection) . . . . . . . . . . . . . . . .106Pre-admission cooling (roots pumps) . . . . . . . . . . . . . . . . . . . . . . .27Precision diaphragm vacuum gauge . .72Pressure . . . . . . . . . . . . . . . . . . .7, 168Pressure and temperature as function of altitude . . . . . . . . . . . . . .155Pressure converter . . . . . . . . . . . . . . .92Pressure dependence of the mean free path . . . . . . . . . . . . .145, 154Pressure difference oil supply . . . . . . .19Pressure lubrication by geared oil pump . . . . . . . . . . . . . . . . . . . . . .19Pressure measurement direct / indirect . . . . . . . . . .71, 179, 180Pressure measurement, depending on /independent of the type of gas . . . . . .71

Pressure ranges in vacuumtechnology . . . . . . . . . . .12, 56, 57, 145Pressure regulation / control . . .84, 180Pressure regulation / control rough and medium vacuum systems . . . . . .84Pressure regulation in high and ultra high vacuum systems . . . . . . . .87Pressure regulation, continuous / discontinuous . . . . . . . . . . . . . . .84, 86Pressure rise / drop (leak) test .107, 108Pressure units . . . . . . . . . . . . . . .7, 142PTB (Federal physical-technical institute) . . . . . . . . . . . . . . . . . . . . . .81Pumpdown time . . . . . . . . . . . . .62 - 66Pump fluid . . . . . . . . . . . . . . . . . . . . .39Pump fluid backstreaming . . . . . . . . .39Pump fluid change cleaning (diffusion pumps) . . . . . . . . . . . . . .139Pump oil, selection when handlingaggressive vapors . . . . . . . . . . . . . .135Pump throughput . . . . . . . . . . . . . . . . .9Pumping (gas sorption) of vacuum gauges . . . . . . . . . . . . . .77, 78Pumping chamber . . . . . . . . . . . . . . .16Pumping of gases . . . . . . . . . . . . . . .57Pumping of gases and vapors . . . . .20, 21, 34, 52, 57, 58, 135Pumping speed . . . . . . . . . . . . . . . . . .8Pumping speed units, conversion of . . . . . . . . . . . . . . . . . .144Pumping various chemical substances . . . . . . . . . . . . . . . . . . .136Purge gas . . . . . . . . . . . . . . . . . . . . .29PVD (physical vapor deposition) . . .128pV-flow . . . . . . . . . . . . . . . . . . . . . . . .9pV-value . . . . . . . . . . . . . . . . . . . . . . .8Quadrupole mass spectrometer . . . . .89Quadrupole, design of the sensor . . . .90Quadrupole, gas admission / pressure adaptation . . . . . . . . . . . . . .93Quadrupole, measurement system (detector) . . . . . . . . . . . . . . . . . . . . . .92Quadrupole, separating system . . . . .91Quadrupole, specifications . . . . . . . . .94Qualitative gas analysis . . . . . . . . . .100Quantitative gas analysis . . . . . . . . .101Quantity of gas (pV value) . . . . . . . . . .8Quartz crystals, shape of . . . . . . . . .121Rate watcher . . . . . . . . . . . . . . . . . .120Reduction of adsorption capacity . . .139Reduction ratio . . . . . . . . . . .24, 25, 137Refrigerator cryopump . . . . . . . . .49, 51Regeneration time . . . . . . . . . . . . . . .53Relative ionization probability (RIP) . .98Residual gas composition (spectrum) . . . . . . . . . . . . . . . . . .43, 44Response time of leak detectors . . . .117Reynold's number . . . . . . . . . . . . . . .12Rigid envelope . . . . . . . . . . . . .118, 119

Roots pumps . . . . . . . . . . . . . . . . . . .24Roots pumps, Instructions for operation . . . . . . . . . . . . . . . . . . . . .137Rotary displacement pumps . . . . . . . .18Rotary plunger pumps . . . . . . . . . . . .20Rotary vane / piston pumps, Instructions for operation . . . . . . . . .135Rotary vane pumps . . . . . . . . . . . . . .18Salt, drying of . . . . . . . . . . . . . . . . . .66Saturation vapor pressure (nonmetallic gaskets) . . . . . . . . . . . .161Saturation vapor pressure . . . . . . .7, 22Saturation vapor pressure (cryogenic technology) . . . . . . . . . . .161Saturation vapor pressure (metals) . .160Saturation vapor pressure (pump fluids) . . . . . . . . . . . . . . . . . .160Saturation vapor pressure (solvents) . . . . . . . . . . . . . . . . . . . .160Saturation vapor pressure and vapor density of water . . . . . . .147, 164Sealing passage . . . . . . . . . . . . . .18, 19Seal-off fitting . . . . . . . . . . . . . . . . . .69Selection of pumps . . . . . . . . . . . . . .56Selection of pumps for drying processes . . . . . . . . . . . . . . . . . . . . .66Sensitivity of quadrupole sensors . . . .95Sensitivity of vacuum gauges . . . . . . .78Separating system of mass spectrometers . . . . . . . . . . . . . .90Shell baffle . . . . . . . . . . . . . . . . . . . . .39Silicone oils, DC 704, DC 705 . .39, 160Small flange . . . . . . . . . . . . . . . . . . . .67Smallest detectable concentration . . . .95Smallest detectable partial pressure . .95Smallest detectable partial pressure ratio . . . . . . . . . . . . . . . . . .95Sniffer technology . . . . . . . . . . . . . .118Software for TRANSPECTOR . . . . . .102SOGEVAC pumps . . . . . . . . . . . . . . .18Solvents . . . . . . . . . . . . . . . . . . . . . .146Sorption pumps . . . . . . . . . . . . .45, 177Specific volume of water vapor .147, 163Spinning rotor gauge (SRG) . . . . . . .75Spray technique (Helium) . . . . . . . . .117Sputter ion pumps, Instructions for operation . . . . . . . . . . . . . . . . . .140Sputter pumps . . . . . . . . . . . . . . . . . .46Sputtering . . . . . . . . . . . . . . . . . . . .129Sputtering (cathode sputtering) . . . .129Sputter-ion pumps . . . . . . . . . . . .46, 47SRG (spinning rotor gauge), VISCOVAC . . . . . . . . . . . . . . . . . . . . .75Stability for noble gases (sputter ionpumps) . . . . . . . . . . . . . . . . .46, 47, 48Standard pressure . . . . . . . . . . . . . . . .7Standards in vacuum technology . . . . . . . .171 – 173Static expansion method . . . . . . .81, 82

D00 E 19.06.2001 21:41 Uhr Seite 186

Index Fundamentals of Vacuum Technology


Steam ejector pumps . . . . . . . . . . . . .40Storage under pressure(bombing test) . . . . . . . . . . . . . . . . .119Stray magnetic field . . . . . . . . . . . . . .47Stray magnetic field (sputter ion pumps) . . . . . . . . . . . . . .47Sublimation pumps . . . . . . . . . . .46, 47Symbols and units, alphabetical list . . . . . . . . . . .165 – 170Symbols used in vacuum technology . . . . . . . . . . . . . . .1572, 153Temperature comparison and conversion table . . . . . . . . . . . . . . . .154Temperature in the atmosphere . . . .155Terms and definitions (leak detection) . . . . . . . . . . . . . . . .106Test gas accumulation . . . . . . . . . . .119Test leaks . . . . . . . . . . . . . . . . . . . . .113Thermal conductivity vacuum gauge, constant / variable resistance . . . . . . .76Thermal conductivity vacuum gauges . . . . . . . . . . . . . . . . . . . . . . . .76Thermal evaporator (boat) . . . . . . . .128THERMOVAC . . . . . . . . . . . . . . . . . .75Thickness control with quartz oscillators . . . . . . . . . . . . . . . . . . . .120Thickness measurement . . . . . . . . . .120Thin film controllers . . . . .120, 127, 181Throtteling of pumping speed when using condensers . . . . . . . .34, 35Time constant . . . . . . . . . . . . . .62, 116Titanium sublimation pump . . . . . . . .46Titanium sublimation pumps, Instructions for operation . . . . . . . . .140Torr and its conversion . . . . . . . . . . .142Total pressure . . . . . . . . . . . . . . . . . . .7Transfer standard . . . . . . . . . . . . . . . .81Transmitter . . . . . . . . . . . . . . . . . . . .75TRANSPECTOR . . . . . . . . . . . . . . . . .89Triode sputter ion pumps . . . . . . . . . .47TRIVAC pumps . . . . . . . . . . . . . . . . .19Trochoid pumps . . . . . . . . . . . . . . . .21Tuning / adjustment and calibration of leak detectors . . . . . . .112Turbomolecular pumps . . . . . . . .41, 176Turbomolecular pumps, Instructions for operation . . . . . . . . .138TURBOVAC pumps . . . . . . . . . . . . . .43Turbulent flow . . . . . . . . . . . . . . . . . .12Types of leak . . . . . . . . . . . . . . . . . .104Types pV flow . . . . . . . . . . . . . . . . . .12UL 200, UL 500 . . . . . . . . . . . . . . .114Ultimate pressure . . . . . . . . . . . . . . . . .7Ultra high vacuum . .11, 13, 60, 61, 178ULTRALEN . . . . . . . . . . . . . . . .39, 161 Units, symbols . . . . . . . . . . . .165 – 170U-tube vacuum gauge . . . . . . . . . . . .74

Vacumm meters, instructions on installing . . . . . . . . . . . . . . . . . . . . .140Vacuum coating techniques . . . . . . .128Vacuum control . . . . . . . . . . . . . . . . .70Vacuum equipment, Instructions for operation . . . . . . . . . . . . . . . . . . . . .134Vacuum gauge contant . . . . . . . . . . . .78Vacuum method (leak detection) . . .107Vacuum physics . . . . . . . . . . . . . . . . . .7Vacuum pumps, literature references . . . . . . . . . . . . . . . . . . . .174Vacuum pumps, survey, classification . . . . . . . . . . . . . . . .16, 17Vacuum ranges (Pressure ranges) . . . . . . . . . . . .13, 56, 57, 145, 161, 162Vacuum regulation . . . . . . . . . . . . . . .70Vacuum symbols . . . . . . . . . . .152, 153Vacuum coating technology . . . . . . .128Values of important physical constants . . . . . . . . . . . . . . . . . . . . .143Valves . . . . . . . . . . . . . . . . . . . .67, 178Van der Waals' equation . . . . . . . . . . .11Vapor density of water . . . . . . .147, 164Vapor pressure . . . .7, 39, 160, 161, 164Vapor-jet pumps . . . .38, 39, 40, 41, 139Venturi nozzle . . . . . . . . . . . . . . . . . .38Viscous (continuum) flow . . . . . . . . .12VISCOVAC vacuum gauge . . . . . . . .775Vitilan, Viton . . . . . . . . . . .68, 149, 161Volume . . . . . . . . . . . . . . . . . . . . . . . .8Volumetric efficiency (roots pumps) . .24Volumetric flow . . . . . . . . . . . . . . . . . .8Water jet pumps . . . . . . . . . . . . . . . .40Water ring pumps . . . . . . . . . . . . . . .18Water vapor tolerance . . . . . . . . . . . .23Web coating . . . . . . . . . . . . . . . . . . .130Wet processes . . . . . . . . . . . . . . . . . .58Working pressure . . . . . . . . . . . . . . . .7Working ranges of vacuum pumps . .161X-ray effect . . . . . . . . . . . . . . . . . . . .78XTC, XTM . . . . . . . . . . . . . . . . . . . .127Zeolith . . . . . . . . . . . . . . . . . . . . . . .45Z-Match technique . . . . . . . . . . . . .122

D00

D00 E 19.06.2001 21:41 Uhr Seite 187



D00 E 19.06.2001 21:41 Uhr Seite 188

Documents

S = —— dV dtlgonchar/courses/p9826/Vacuum_fundamentals.pdf · ture, turbomolecular pumps with magnetic bearings, the A-Series vacuum gauges, the TRANSPECTOR and XPR mass spectrometer