24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

1

MESURES de DEGAZAGE

G.Vandoni, CERN

Definitions, UnitsSome theory and phenomenology with some tentative explanationMethods of measurement:

Throughputconductance modulationtwo path

Pressure rate-of-rise (accumulation)Coupling throughput and accumulationTotal mass lossTPD-TDS

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

Definitions

2

“[Intrinsic] Outgassing rate is the instantaneous net amount of gas leaving the material per unit time.”

AVS 9.1-1964 Reporting of Outgassing Data

Measured outgassing rate is the difference between intrinsic outgassing rate and the rate of readsorption of the surfaces in the test chamber.

Depending on sticking s, the desorbed gas may re-absorb on the vessel or the sample itself after multiple collisions.

sample

vacuum vessel

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

3

Definitions

fr. Dégazage

eng. Outgassing: spontaneous, thermal

eng. Degassing: (particle or thermally) induced

Dégazage thermique

Dégazage induit

fr. eng. Désorption: release of adsorbed molecules as the result of the impact of particles or by thermal energy

Outgassing: spontaneous evolution of gas from a solid or a liquid in a vacuum

Outgassing flux is the quantity of gas leaving the surface per unit time at a specified time t after the start of evacuation

Degassing: deliberate removal of gas from a solid or liquid in vacuum as the result of the impact of molecules, electrons, ions, photons, or by heating.

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

Glossary

4

p : pressure po : pressure at time to pu: pressure at time ∞

S : pumping speed Seff : effective (at the vessel entrance) pumping speed

A : surface area, effective (macroscopic)

V : volume of the vacuum vessel

Q: outgassing rate, total, i.e., gas released per unit time (should read ) Q

q: surface specific outgassing rate, total, i.e., gas released per unit time and unit surface (should read ) q

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

5

Units of gas quantity

p.V= energy! To use [pV] as a gas quantity, one must declare at which T.Q may be measured in [Watt]

kB [kB units] (1/kBT) at 296 K [1/kBT units]

1.38 x 10-23 Pa.m3 K-1 2.45 x 1020 (Pa.m3)-1

1.38 x 10-22 mbar.l.K-1 2.45 x 1019 (mbar.l)-1

1.04 x 10-22 torr.l.K-1 3.3 x 1019 (torr.l)-1

R [R units] RT at 296 K [RT units]

8.31 Pa m3 K-1 mol-1 2.46 x 103 Pa m3 mol-1

83.1 mbar l K-1 mol-1 2.46 x 104 mbar l mol-1

62.4 torr l K-1 mol-1 1.85 x 104 torr l mol-1

mol

ecul

esm

oles

TRNTkNVp molB

N number of moleculesNmol number of moles, i.e., of batches of 6.02 x 1023 molecules ABNkR

RTdt

dNTkdtdN

dtpVdQ mol

B )(

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

6

Units of outgassing rates

To From

10-3 7.5 x 10-4 2.5 x 1016 4.1 x 10-8

103 0.75 2.5 x 1019 4.1 x 10-5

1330 1.33 3.3 x 1019 5.4 x 10-5

4.1 x 10-17 4.1 x 10-20 3.1 x 10-20 1.66 x 10-24

2.4 x 107 2.4 x 104 1.9 x 104 6.02 x 1023

Conversion from pV to N or Nmol at 296K*

* standard AMBIENT temperature (and pressure) SATP: 296K ± 3Kstandard temperature and pressure STP: 273.15K

NA

timeATRN

timeATkN

timeAVp

AQq molB

q surface specific outgassing rate

Q total outgassing rateA sample surface

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

7

Gas quantity to mass

Gas flux [mbar.l. s-1] mass loss rate [g s-1]

If gas species not known, Mmol=28g and outgassing data in nitrogen equivalent.

Nmol

Mmol : molecular weight of the gasmolMRTpVM

molMRT

dtdM

dtpVdQ )(

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

8

Why measure outgassing rates

In absence of leaks, the total outgassing rate Q at time t and the applied pumping speed S determine the pressure in the system at time t.

S can be varied by ~ 3 orders of magnitude: 1 to 1000 l.s-1

Q may vary by > 10 orders of magnitude: 10-5 to 10-15 mbar.l.s-1cm-2

Sound design of a vacuum system requires knowledge and limitation of Q

pu ultimate pressure (at ∞ time)

*Strictly, (1) is valid only in a volume where gas evolves from surfaces and flows due to pumping: not where large T differences exist.

upSQp

upStQtp )()(

(1)*

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

9

Typical outgassing ratesMaterial mbar l s-1 cm-2 molecules s-1 cm-2

Neoprene, non-baked, @10h 10-5 (H2O) 3.3 x 1014

Viton, non-baked, @10h 10-7 (H2O) 3.3 1012

Viton, baked 5 x 10-10 (H2O) 1 x 1010

Unbaked stainless steel @10h 3 x 10-10 (H2O) 6.6 x 109

Baked stainless steel (150oC, 24h) 3 x 10-12 (H2) 6.6 x 107

Vacuum fired stainless steel, 4 cycles 300oC, 24h

10-14 (H2) 6.6 x 105

Baked OFS Copper (200oC, 24h) 10-14 (H2) 6.6 x 105

Ti-Zr-V activated at 180oC (24h) <10-16 (CH4) <3.3 x 103

Ti-Zr-V activated at 180oC (24h) ≈10-18 (Kr) ≈ 30

mbar l s-1 molecules s-1 Bayard-Alpert gauge, SVT, W filament ≈ 10-9 (CO, CO2,H2) ≈ 3 x 1010

B-A gauge SVT, thoriated filament ≈ 10-10 (CO, CO2,H2) ≈ 3 x 109

QMS Inficon 10-8 ÷ 10-10 3 x 1011 ÷ 3 x 109

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

10

Gas specific outgassing rates

Gas Al Cu St.steel Be

H2 7 x 10-12 1.4 x 10-11 7 x 10-12 1.4 x 10-11

CH4 5 x 10-13 5 x 10-13 5 x 10-13 1 x 10-12

CO 5 x 10-12 1 x 10-12 5 x 10-12 1 x 10-11

CO2 5 x 10-13 2.5 x 10-13 5 x 10-13 1 x 10-12

H2O 3 x 10-10 3 x 10-10 3 x 10-10 6 x 10-10

Gas Al Cu St.steel Be

H2 5 x 10-13 1 x 10-12 5 x 10-13 1 x 10-12

CH4 5 x 10-15 5 x 10-15 5 x 10-15 1 x 10-14

CO 1 x 10-14 1 x 10-14 1 x 10-14 2 x 10-14

CO2 1 x 10-14 5 x 10-15 1 x 10-14 2 x 10-14

H2O 1 x 10-14 <1 x 10-15 1 x 10-14 2 x 10-14

Metals, unbaked, at 10h [torr.l.s-1.cm-2]

Metals, baked, at 50h [torr.l.s-1.cm-2]

CAUTION: Design values circulating at CERN, no reference

Dominant

Dominant3 ordersof magnitude

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

11

Pumpdown: effect of outgassing on p(t)In a “perfect” vacuum vessel of volume V, pumped with constant pumping speed S, if p(t) is the pressure at time t, without leaks, permeation or outgassing:

3

Pres

sure

time

10-4

10-5

10-6

1 2 4 5 6 7

ln p versus t

pumpedQSpdtdpV

Volume depletion t

VSpp oideal lnln

tnkpreal lnln

with 0.5n1.2 n

real ktp

tVS

oideal epp

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

12

Pumpdown: effect of outgassing on p(t)In a “perfect” vacuum vessel of volume V, pumped with constant pumping speed S, if p(t) is the pressure at time t, without leaks, permeation or outgassing:

pumpedQSpdtdpV

Volume depletion t

VSpp oideal lnln

tnkpreal lnln

nreal ktp

tVS

oideal epp

Pres

sure

time

10-4

10-5

10-6

1 10 100 1000

ln(p) versus ln(t)

ntqtq 1)(

At constant pumping speed S, this is also the time dependence of q

where q1 is the specific outgassing rate after 1 unit time (1h) pumping

n ~1 for metallic unbaked surfaces~0.5 for elastomers Knowing q1h (or q10h)extrapolate

with 0.5n1.2

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

13

Outgassing rate time evolution

Type I, unbaked metals

Type II, plastics

Type III, plastics ottetq /)(

Let’s go for some insight into outgassing time evolution…

2/1)( ttq n=1/2

1)( ttq n=1

10-9

10-8

10-7

10-6

10-5

1 10

NB: For baked metals, q(t) goes with n~0.5 on a short timescale and with exp(-t/to) on a long timescale.Of course, on a very different absolute range…

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

14

Phenomena behind outgassing

Surface:Adsorption

chemisorption/ physisorptionDesorption

Bulk:(Dissociation and) PermeationDiffusion

Molecules deposited while exposed at atmospheric pressure desorb under vacuum.(Dissociation) and diffusion of atomic species in the near surface, through microcracks and the oxyde layer.Diffusion from the bulk of atomic H, trapped during manufacturing process, recombination at the surface and desorption as H2.Permeation through thin walls and elastomeric seals.

Outgassing:Combination of all these!

All relatively well understood and modeled:It’s their combination which is not easy to understand!

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

15

Surface bondingE

zEb

zo

Eb

E

zzo

PhysisorptionVan der Waals

Eb~6-40kJ/mole, 0.06-0.4 eV/moleculezo~2-4ÅCoverage: ~ multilayer

ChemisorptionElectron sharing

Eb~40-1000kJ/mole, 0.4-10eV/moleculezo~1.5ÅCoverage ~ 1 monolayer

Eth=kT=0.026eV at 300K

Real surface: multitude of adsorption sites, hence of binding energies

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

16

Dynamical description

E

zEb

zo

Simplified treatment (Frenkel 1924):In a potential well of depth Eb, , the molecule vibrates with characteristic frequency ~ n=kT/h~1013 s-1 . Frequency of attempts to overcome the potential barrier and break free from the surface.

Probability that fluctuation in thermal energy is ≥Eb: exp(-Eb/kT)

Boltzmann factor

Probability per unit time that a molecule will desorb:

)/exp(1)/exp(0

0 kTEkTE bb

n

Frequency of attempts

Probability to get free in 1 attempt

Average time of stay:

)/exp(10/1 13 kTEb

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

17

Time of stay: binding energy

Eb [kJ/mol]

Eb [eV]

a

6 0.06 1.2 x 10-12 s

15 0.16 5 x 10-11 s

40 0.41 1.2 x 10-6 s

60 0.62 5.5 x0-3 s

80 0.83 15 s

90 0.93 15 min

95 0.98 2 h

100 1.04 8 x 104 s ~ 1 d

120 1.24 2 x 108 s ~ 10 years

150 1.55 7 x 1013 s ~ 20,000 years

Dependence of stay time on binding energy at room temperature

1eV~97kJ/mole

)/exp(10/1 13 kTEba The stronger Eb, or the lower T, the longer

the average time of stay (sojourn time)

kT at room temperature: 0.025eV

physisorption

weak chemisorption

strong chemisorption

.

.

.

.

Neither very short nor very long times bother us.

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

18

Time of stay: temperature

)/exp(10/1 13 kTEba

T [K]

a

77 1041 s

295 15 s

300 1 ms

Typical binding energy of H2O: 80kJ/mole, 0.8eV

Effect of temperature on stay time

We still don’t know why q(t)~t-n, but we may recognize several features of desorption, which then determine changes in outgassing rates:

Effect of temperature: cryopumping, bakeoutQuantities and coverages: q ≤ 1ML (by stay time)Binding energies

Bridge between microscopic view and macroscopic observation

TPD-TDS

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

19

Unbaked metals: t-1

Starting from the surface physics vision, several models have been attempted to explain the t-1 behaviour and relative independence from treatments, history, composition: superposition of isotherms, surface and sub-surface diffusion through micropores in the oxyde layer…

2cmslmbar

][103)(

9

httqfor standard rugosityA

B Dominated by H2O , chemisorbed at the first 1÷2 layers~ binding energy 85-90 kJ/mol, stay time [min-hr]

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

20

Diffusion limited desorption

d

Some initial concentration of gas is trapped in a slab of finite dimension d.

The concentration of this gas is zero at the border of the slab.

The gas will get out of the solid progressively, because of concentration gradient and gradually empty the slab.

Vacuum

Vacuum

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

21

Diffusion: polymers and baked metalsFick’s laws

dxdcDq

Flux ~concentration gradient

ttxc

xtxcD

),(),(2

2

Conservation of mass

Limit at small timesDdt /025.0 2

Limit at long timesDdt /025.0 2

Concentration decays quickly close to the boundaries of the solid, remaining fairly constant in the bulk of the slab.Outgassing rate:

Concentration may be considered as uniform in the slab of material. It decreases exponentially with time, with time constant d2/p2DOutgassing rate:

ttq 121

dx

tddcDq

),2

(deg

ottq /exp “Volume depletion”

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

2222

Water solubility: 0.1 to 0.5 wt.% (4.4 to 22x1019 molecules/cm3)

10 to 50 times larger than the H total content in as

produced austenitic stainless steel

Water diffusivity at RT: 5 x 10-9 cm2 s-1

2000 times larger than that reported for H in austenitic

stainless steel.

* After G.Mensitieri et al., J.Appl.Polym.Sci., 37, 381, (1989)

10-9

10-8

10-7

10-6

2.6 2.8 3.0 3.2 3.4 3.6 3.8

H2O diffusivity in PEEK *

Diffu

sivity

[cm

2 s-1

]

1000/T [K]

RT

Outgassing rate of polymers is known to be much higher than that of metals. Two reasons explain this phenomenon: a polymer contains much more gases than a metal, and the gas mobility in polymers is orders of magnitude larger than in metals. Example

O COO

[ ]n

PEEK

Courtesy P.Chiggiato

Outgassing of Polymers

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

23

Outgassing rate of materials

Type I, metals1)( ttq

Type II, plastics2/1)( ttq

Type III, plastics ottetq /)(

Time dependence1)( ttq Determined by H2O desorption2/1)( ttq Bulk diffusion rate limited

ottetq /)( Volume depletion

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

25

Outgassing measurement methods

Methode par debit Methode par accumulation

Methode par double pesee

p

C

Sp

Q p Qb

Qs

22 21

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

26

Throughput method

p

C

Sp

Q sample

known conductance

pump

Total or partial pressure gauge

10-7

10-6

10-5

0.0001

100 1000 104 105

Out

gass

ing

(mba

r.l.s

-1)

Time (s)

BlankQ231Q244

y = 0.14 * x (̂-1.02)y = 0.12 * x (̂-1.00)y = 0.09 * x (̂-1.08)

Qv determined by a blank run, with pv(t) the blank pumpdown curve.

In steady state (after an equilibration time):

effSQp

effStQtp )()(

Qs sample outgassingQv vessel outgassingvs QQQ

Outgassing rate:Calculated from raw data, by subtracting pressure readings at corresponding times:

)()()( tptptp v

AS sample’s surface

Sample outgassing rates

effh A

hpSq

)10(10

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

27

Throughput method: choice of COrifice: chamfered edge, conical section transmission a=1±0.00001 (see K.F.Poulter 1973)Conductance is then:

80o

A

The area of C is chosen as a function of the expected outgassing rate and taking the error sources into account.

A

Sp

Q

pout

pIf C«Sp, i.e. Sp>50C, then Seff≈C

B If C too small w.r. to s.AS, readsorption on the vessel’s walls will fake the measurement and result in estimated Q too small.

AS sample area, s sticking coeff for evolved gases

2

2cm inAslMRTAC

mol

]/[8.112

effout pSpC(pQ )

C C >10x gauge pumping speed

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

28

Sensitivity limits -1-

Detection limit of outgassing rate: >25% of background

Baked, stainless steel spherical vessel 2112103 cmslmbarqv

lV 10 19106 slmbarQv

Gauge outgassing, W filament

1910 slmbarQgauge

Background pressure

mbarpbkg10107 19107 slmbarQbkg

Bkg outgassing

19min 108.1 slmbarQ

H2 outgassing

110 slSeff

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

29

Sensitivity limits -2-

Detection limit of outgassing rate: >25% of background

Baked, vacuum fired stainless steel spherical vessel

211410 cmslmbarqv

Gauge outgassing, thoriated filament

11010 slmbarQgauge

110102.1 slmbarQbkgBkg outgassing

111min 103 slmbarQ

Gauge outgassing dominates

H2 outgassing

lV 10 111102.2 slmbarQv

110 slSeff

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

30

Effect of reasdorption

The vessel and the sample itself, readsorb the outgassed flux.

Measured outgassing rate is not intrinsic outgassing grate.

im QQ

cminAslsAm

RTsASA ]/[8.112

SA re-pumping speed

A re-pumping surface s sticking coefficient

Due to readsorption, measurement will “miss” a part of the intrinsic outgassing flux.

Remember:H2O sticking s=10-3 at room temperature

H2O outgassing

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

31

Reasdorption in the throughput method

2cmin]/[8.112

AslsAm

RTsASA

SA re-pumping speed

Seff effective applied pumping speed, typically 10l/s

Relevant parameterSeff / SA p

C=Seff

Q

The gas released by the sample is not entirely evacuated. A fraction of it is reabsorbed by sample or vessel

eff

AeffeffA S

SpSSSpQ 1)(int

effm pSQ

eff

A

m SS

QQ

1int

Sticking s=10-3 (H2O) at 20oCC=1cm2

3int mQ

Q

A A/Seff

[cm2] 1 10-2 10-3 10-4

1 1 2 1 1 110 10 11 1.1 1 1

100 100 101 2 1.1 12345 2345 2346 24 3 1.2 10l vessel +100cm2 sample

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

32

Error introduced by gauge

Outgassing of gauge:Contribution to Q

Pumping effect of gauge:Contribution to S

thoriated W filament

W filamentIonic pumping Ions are trapped at the collector and are removed from the gas phase as neutrals; and may be buried by implantation. Affects all gases including noble gases. ~10-2 l/s

Chemical pumpingReaction or cracking at the hot filament. Dissociation of H2O, CnH2n, H2. Does not affect monoatomic gases

Thoria coated filaments have lower temperature, hence lower S. ~1 l/s

Error relevant for H2, when vessel background becomes comparable

-1slmbar 910gaugeQ

-1slmbar 1010gaugeQ

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

33

Fisher-Mommsen dome

Fisher-Mommsen dome, used for Pumping speed measurements

more generally permits a precise quantification of gas flux via pressure difference measurement:

E. Fischer, H. Mommsen, Vacuum 17, 309 (1967)

The Fisher Mommsen dome has geometry, dimensions and position of the gauge such that measured pressures are identical to those measurable in the ideal case.

Thermodynamically, p is defined only in an enclosed system at equilibrium. To approach isotropy, we need a “infinite volume” dome, and gauges far from the pump. The gauge should then measure a isotropic, Maxwellian gas distribution. The inlet C and pump S should have a negligible effect on the distribution.

)( 12 ppCQ

…or how to reduce error due to pressure anisotropy

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

34

Throughput - Conductance modulationA method to get rid of the measurement error in S and the differential error of 2 gauges

p

Q, Sp should not vary while the conductance is varied (Pp should remain constant).

Involves ratios and not absolute valuesIndependent of Sp

21,CCC 21,SSS 21, ppp SppQ pump

CSS p

111

constCQp

p

1/Cp1-p2

Slope Q

21

11CC

K.Terada et al, J.Vac.Sci.Technol. A7 (189) 2397

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

35

Throughput - Conductance modulation

R.P.Henry, Le Vide, No 82 (1959) 226; Le Vide No 144 (1969) 316

Variable stop of a aerial photography camera, embedded in a gate valve:C varies between 3.5 and 173 l/s/

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

36

Two path method

QVb

Va Q1

Q2p1

p2

G1

to pump

CG2

sample chamber

Va open Vb closed

conventional throughput method

Va closed Vb open

A

B

Needs only G1.G2 used just for concomitant, conventional throughput method

CPPQ AAA )( 21

CPPQ BBB )( 21

P2A = P2B constant! determined by Q+Q1+Q2 and S.

CPPCPPCPPQQQ BABBAABA )()()( 112121

K.Saito, Y.Sato et al. Vacuum 47, (1996) 749

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

37

Two path method

Vb

Va Qa

Qbp1

p2

G1

to pump

CG2

Systematic errorsSmall difference between 2 pumping paths is seeked: outgassing rate of apparatus outside test chamber must be kept as low as possible. • Va, Vb all metal• Connection pipes to Va, Vb: short• Gas flux qp from valves and connecting pipes measured by blank, with 2-path method.

blank flanges

Single gauge method!Used for very low outgassing rates: by differential measurement, X-ray limit and outgassing of gauge are cancelled out.

cancels constant systematic errorBA PP 11

C=6 l/s, Qmin10-14 mbar.l/s.cm2

K.Saito, Y.Sato et al. Vacuum 47, (1996) 749

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

38

Two path method

K.Saito, Y.Sato et al. Vacuum 47, (1996) 749

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

39

Pressure(-rate-of)-rise methodThe system is pumped down to po. At time to, the system is suddenly isolated from the pumps. Pressure begins to rise.

The rate of rise dp/dt is highest at the instant of isolation.It will gradually decrease to zero when outgassing rate and surface readsorption + gauge pumping balance each other.

Pumpdown may be resumed by opening the isolation valve, before the pressure has increased significantly (e.g. less than x 3).

sb

Bo

tt

QQQ

dtdP

TkVtQ

dtdN

o

)(

VtQtQTk

dtdp osob

Btt o

)()(

p Qb

Qs p

tto

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

40

Error sources

Pumping action of the gaugeNon-pumping gauges: spinning rotor, capacitance diaphragmIf a hot-cathode ionization gauge is used, reduce electron current and choose a low-temperature filamentSwitch gauge on and off intermittently

Background outgassing from the vacuum vessel, the gauges, the isolation valveLimit background by choosing all-metal isolation valveApply a reproducible venting procedure, take measurement at the same time

Readsorption on the sample and the vessel’s walls

To be significant, QS ≥ 0.25 × QB

Example:Baked spherical vessel, V=10l, qv=3.10-12 mbar l /s.cm2 Area=2245cm2, → QB = 6.7.10-9 mbar.l/s→ QS ≥ 1.7.10-9 mbar.l/s → with a sample of 100cm2, qS ≥ 1.7.10-11 mbar l /s.cm2

Very practical if the outgassing sample is the vessel itself

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

Effect of readsorption: pressure rise method

41

p

t

tVQTkptp Bo )(

p Qb

Qs

Pressure rise may be followed for a long period of time (sometimes even ~30days). To minimize gauges’ pumping, gauges can be switched on/off intermittently. If Q=constant, then p(t)=linear without repumping

no re

pum

ping

repumping

Measurement of H2 outgassing in a baked and vacuum fired stainless steel vessel show a linear increase of p(t): This demonstrates that H2 does not reabsorb on stainless steel.

P.A.Redhead, J.Vac.Sci.Technol.A 14(4) 1996

In the case of H2 outgassing, the error is then only due to gauge pumping.

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

42

Accumulation + throughput methodExtends the limits of the throughput method to lowest outgassing rates

Limits the effect of the gaugeLimits the effect of the system’s outgassing

The sample is in a separate vessel. After pumpdown, the sample vessel is isolated and gas is accumulated for a time ta.At time ta the leak valve is opened. Pressure evolves as shown below.

measuring dome

ta t

p

tt

teffbsaa

a

dttpSGGtG )()(

G total evolved gas quantity (not rate!)

ta may be very long!

QbQsp

S

Seff

small volume no gauge variable leak

valve

accumulation dome

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

43

Accumulation + throughput method

ta

I(ta)

measured for different ta

tt

t aa

a

tIdttp )()(

Linear regression of I(ta) versus ta: slope gives Q

tt

teffaaa

a

dttpStQtG )()(

Limitation: repumping during ta. For gases which are not reabsorbed and when the test vessel is the sample, very high sensitivity may be reached.Ex. ~30 molecules.s-1.cm-2, Kr trapped during production by sputtering and evolving from NEG

measuring dome

QbQsp

S

Seff

accumulation dome

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

44

Mass (or Weight) loss method

sample

cooling platecollector chamber

separation plate

heated sample chamber and plug (Cu)

125 oC

collector plate25 oC

10-6mbar Simultaneous measure of

Total Mass Loss

Collected Volatile Condensable MaterialsQuantity outgassed at 125oC and condensed on collector at 25oCWater Vapor RegainedMass of H2O vapor regained after reconditioning, upon exposure to 50% humidity.

oo

o mm

mmTML %

On a laboratory microbalance, a sample and its containing boat (Cu) is weighed.A collector plate is also weighed.The sample is put in a vacuum vessel, evacuated at 10-6mbar, then heated at 125oC for 24hAfter this time, the sample and the collector are weighed again.

ASTM standard E-595-07 NASA database outgassing.nasa.gov

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

45

Mass (or Weight) loss methodASTM standard test method E-595-07, based on apparatus developed at NASA

PROCEDURE:Conditioning (50% humidity, 23oC, 24h)Weighing of sample / Weighing of collectorEvacuation at 10-6mbar, keep at 125oC for 24h/ Venting with dry inert gasRapid (<2min) weighing of sample/Weighing of collector

QUANTITIES and PrecisionSample weight ~ 100 to 300mgBalance precision: ±1mgTheoretical limit precision: 0.002 %

Possible bias and mitigationCleanliness of manipulationInhomogeneity of sample Control collectors are used to check cross contamination or poor technique: it these detect more than 50mg, the measure is discarded

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

46

TML system at CERN

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

47

Mass loss method

At NASA, rate of weight loss was measured continuously, by weighing under vacuum

Podlaseck et al, Trans 9th AVS Symp (1962) 320

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

48

Temperature Programmed Desorption

Complementary technique to overcome sensitivity limitations of previous methods.Adapted to fundamental studies of interaction between gases and surfaces.

Thermal Desorption (Spectroscopy)

Temperature Programmed Desorption or Thermal Desorption Spectroscopy is a standard surface science technique which has been used to provide information on the binding energies of atomic and molecular species adsorbed on a solid surface.

Requires knowledge or assumptions on the underlying physical modelVery useful on “simplified” systems, as usual in surface physics: monocrystals, few absorbed species, well-controlled preparation conditions.

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

49

Temperature Pgrogrammed Desorption

sam

ple

QMS

Partial pressures are measured versus T

Provides information on binding energy of atoms and molecules on surfaces

Probability of desorption~ frequency of attempts no, Boltzmann factor, function of coverage f(q). dqm/dt is rate of desorption=probability of desorption x number of molecules in monolayer nm

Assumption: desorption rate of an adsorbate is to its measured partial pressure. This condition is rigorous only if pumping speed > rate of desorption.

Ed desorption activation energy

TkE

dtd

B

dno

m expqnq

tCdtdT

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

50

TPD-TDS

Analysis: rewrite Polanyi-Wigner equation

Arrhenius plot: Plot ln(dqm/dt ) versus 1/T, obtain a straight line with

slope -Ed/kB intercept ln(no) +n.ln(qm)

TkEn

dtd

B

dmo

m qnq lnlnln

TkE

dtd

B

dno

m expqnq

Leading edge analysis: use only <5% of desorption trace, such that qm is ~ constant.

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

51

Conclusion

Pressure rate-of-rise

Throughput

Mass loss

Gauge pumping speed, readsorption, gauge and background outgassing, are all sources of error in outgassing rate measurements.

Care must be taken and case-by-case analysis, together with some knowledge of the adsorption, diffusion, outgassing physics is required to chose the best configuration.

Two-path

Conductance modulation

Accumulation + throughput

Temperature Programmed Desorption – Thermal Desorption Spectroscopy

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

52

Bibliography and Acknowledgements

I am heartily grateful to P.Chiggiato for guiding me through the meanders of outgassing phenomenology, theory and measurement

Recommended practices for measuring and reporting outgassing dataP.A.Redhead, J.Vac.Sci.Technol. A20, 1667 (2002)

Outgassing of vacuum materials IR.J.Elsey, Vacuum 25 (7) 299 (1975)

Outgassing of vacuum materials IIR.J.Elsey, Vacuum 25 (8) 347 (1975)

P.Chiggiato, CAS 2006 Lecture, in pdf.

24. O

ctob

re 2

012

– Mes

ures

de

dega

zage

- G.V

ando

ni

53

Further reading

Measurement system for low outgassing materials by switching between two pumping pathsK.Saito et al, Vacuum 47, 749 (1996)

Accuracy of the conductance modulation method for the measurement of pumping speed and outgassing rateY.Tuzi, Vacuum 41, 2004 (1990)

Effects of readsorption on outgassing rate measurementsP.A.Redhead, J.Vac.Sci.Technol. A14, 2599 (1996)

Monte-Carlo computation on molecular flow in pumping speed test domesE.Fischer, H.Mommsen, ISR-VAC (1966)

Outgassing database NASA:http://www.outgassing.nasa.gov

Outgassing database ESA:http://esmat.esa.int/Services/outgassing_data/outgassing_data.html