Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 20061
Cold / sticky system
Vincent Baglin
CERN AT-VAC, Geneva
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 20062
Cold / sticky system ??
Vacuum ??
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 20063
Eureka ??
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 20064
Cold / sticky systemOutline
1. Cryopumping
2. Adsorption isotherms
3. Cryosorbers in cold systems
4. He leaks in cold systems
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 20065
1. Cryopumping
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 20066
Desorption probability
• The desorption probability is a function of the binding energy, E and the temperature, T (first order desorption, Frenkel 1924). The surface coverage, θ, varies like :
• The desorption process is caracterised by the sojourn time, τ :
• For large E and small T, molecules remains onto the surface : CRYOPUMPING
• For some combination of E and T, the molecule is desorbed (bake out)
Tk E-
0 e - νθθ=
dtd
e 0
Tk E
ντ =
(ν0 ~ 1013 Hz, k = 86.17 10-6 eV/K)
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 20067
Sojourn timeChemisorbed molecules
1.E-04
1.E-03
1.E-02
1.E-01
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
0 50 100 150 200 250 300 350 400
Degres Celcius
Sojo
urn
time
(h)
H2O_AluminumH2_Stainless steelCO_Stainless steelCO_Stainless steel
~ 1 Year
~ 1 s
0.9 eV1.1 eV
1.2 eV
1.7 eV
• At room temperature :
H2O dominates
> 100 ºC to remove H2O
• A bake out up to 300 ºCremoves molecules with large binding energies which could be desorbed by stimulated desorption
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 20068
Cryopumping regimes
Physisorption• Sub-monolayer coverage : attractive force (van der Waals) between a gas molecules and a material• Binding energy for physical adsorption• H2 from 20 to 85 meV for smooth and porous materials resp.• 1 h sojourn time at 5.2 K and 26 K for smooth and porous materials resp.
Condensation
• For thick gas coverage, only forces between gas molecules• Energy of vaporisation 9 to 175 meV for H2 and CO2 resp.• 1 h sojourn time at 2.8 K and 53.4 K for H2 and CO2 resp.
sub-monolayers quantities of gas can be physisorbed at their boiling temperature (ex : H2 boils at 20.3 K and a bake-out above 100 ºC removes water)
Cryotrapping
• Use of a easily condensable carrier (e.g. Ar) to trap molecules withhigh vapor pressure (e.g. He, H2)
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 20069
Sticking probability/coefficient
• Probability : 0 < σ < 1ν collision rate (molecules.s-1.cm-2)
• Function of gas, surface, surface coverage, temperature of gas and surface temperature
• Pumping speed
i.e : σ times the conductance of a surface
incident
sticking
incident
departingincident -
νν
ννν
σ ==
vA 41 vA
PP - 1
41 S
sat
σσ ≈⎟⎟⎠
⎞⎜⎜⎝
⎛=
[ ]MT 3.63 cm.l.s S 2-1- σ=
0.6
0.7
0.8
0.9
1
0 100 200 300 400 500 600 700 800 900 1000
Gas temperature (K)
Con
dens
atio
n co
effic
ient
3,5 K3,7 K3,77 K3,9 K
J.N. Chubb et al. Vacuum/vol 15/number 10/491-496
J.N. Chubb et al. J. Phys. D, 1968, vol 1, 361
H2 at 300 K incident onto a surface at 3.1 K
• H2 and CO at 4.2 K :SH2 = 5.3 l.s-1.cm-2
SCO = 1.4 l.s-1.cm-2
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200610
Capture factor, Cf
• Takes into account the geometry of the system :
Baffle in a cryopump Holes in the electron shield of the LHC beam screen
σ 1 2 3 4
0.1 0.48 0.26 0.39 0.43
1 0.68 0.36 0.51 0.57
A.A. Krasnov. Vacuum 73 (2004) 195-199
Cf ~ 0.3
R. Haefer. J. Phys. E : Sci. Instrum., Vol 14, 1981, 273-288
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200611
Thermal transpiration
• Vacuum gauges are located at room temperature to reduce heat load• For small aperture, the collision rate, ν, is conserved at the cold / warm transition
• Since the average velocity scales like √T
_vn
41 =ν
Bayart-Alpert gauge
Capacitance gauge
IGaz injection
H2 tank
Turbomolecular pump
Injection tube
Thermal screens
Screens
Liquid helium level
Injection holes
Leak valve
Extractor gauge
Screen
Cryostat
P1, T1
P2, T2
1
2
2
1
2
1
2
1
TT
nn
TT
PP
=
=
T (K) 4.2 77
P1/P2 8 2
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200612
Experimental evidence of thermal transpirationStatic conditions
1E-14
1E-13
1E-12
1E-11
1E-10
1E-09
1E-08
1E-07
1E-06
1E-05
0 1E+15 2E+15 3E+15 4E+15 5E+15 6E+15 7E+15
Surface coverage (H2.cm-2)
H2 P
ress
ure
(Tor
r at
4,2
K)
Pressure measured at room temperature with the Bayard-Alpert gauge and corrected from thermal transpiration
Pressure measured with the extractor gauge
located at 4.2 KBayart-Alpert gauge
Capacitance gauge
IGaz injection
H2 tank
Turbomolecular pump
Injection tube
Thermal screens
Screens
Liquid helium level
Injection holes
Leak valve
Extractor gauge
Screen
Cryostat
P1, T1
P2, T2
V. Baglin et al. CERN Vacuum Technical Note 1995
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200613
2. Adsorption isotherms
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200614
Adsorption isotherm
• Measurement, at constant temperature, of the equilibrium pressure for a given gas coverage, θ• Varies with:
molecular speciessurface temperature (under 20 K only H2 and He)surface naturegas composition inside the chamber...
• Models :Henry’s law for low surface coverage
DRK (Dubinin, Radushkevich and Kaganer) for metalic, glass and porous substrate. Valid at low pressure. Good prediction with temperature variation
BET (Brunauer, Emmet and Teller). Multi-monolayer description
θ = c P
( ) ( ) ln = ln - D kT lnPPmSatθ θ ⎛
⎝⎜⎞⎠⎟
⎛⎝⎜
⎞⎠⎟
2
( )( )P
P P = 1
+
- 1
PPSat m m Satθ α θ
αα θ−
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200615
Saturated vapour pressure
• Pressure over liquid or gas phase (many monolayers condensed)• Clausius-Clapeyron equation : Log Psat = A – B/T
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200616
C. Benvenuti et al. J.Vac.Sci. 13(6), Nov/Dec 1976, 1172-1182
H2 adsorption isotherm on stainless steel
A monolayer
• Condensation cryopumps allows to pump large quantities of H2
• CERN ISR condensation cryopump operated with liquid He at 2.3 K (50 Torr on the He bath)
C. Benvenuti et al. Vacuum, 29, 11-12, (1974) 591
H2 adsorption isotherm at 4.2 K on CO2 condensat
A = 1 1016 CO2.cm-2
B = 2 1016 CO2.cm-2
C = 0 CO2.cm-2
θ
CO2 condensate
E. Wallén, JVSTA 14(5), 2916, Sep./Oct. 1996
• Packing growth for CO2 films
⇒ Porous layer
• DRK adsorption capacity :0.3 H2/CO2
• Electroplated Cu
• Reduction of the saturated vapour pressure by orders of magnitude
⇒ Cryotrapping
• Electroplated Cu
• In cryopumps CO2 is admitted to enhance the pumping of H2 and He
H2 adsorption isotherm at 4.2 K in co-adsorption with CO2
CO2 concentration
θE. Wallén, JVSTA 14(5), 2916, Sep./Oct. 1996
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200619
He adsorption isotherm from 1.9 to 4.2 K
E. Wallén. J.Vac.Sci.A 15(2), Mar/Apr 1997, 265-274.
• Sub-monolayer range
• Approach of Henry’s law at low coverage
• The isotherms are well described by the DRK model
• θm ~ 1.3 1015 H2/cm2
• Stainless steel
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200620
Cryosorbing materials
• Large capacity• Large pumping speed • Large temperature working range (up to ~ 30 K)
e.g. Activated Charcoal used for cryopumps (see cryopumps talk)Capacity ~ 1022 H2/g i.e. 1021 monolayers (P. Redhead, Physical basis of UHV,1968)
Sticking coefficient ~ 30 % at 30 K (T. Satake, Fus. Tech. Vol 6., Sept. 1984)
20 K cryopanels
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200621
B.E.T surface area – Roughness factor
• Multi-monolayers theory
• Valid for 0.01<P/Psat<0.3
• BET monolayer = θm
• α = exp (ΔE/kT) >>1
1.E-07
1.E-06
1.E-05
1.E-04
1.E-03
1.E-02
1.0E+13 1.0E+14 1.0E+15 1.0E+16 1.0E+17 1.0E+18
Xenon/cm2Xe
non
pres
sure
at 3
00 K
(Tor
r)
304 L & Cuunbaked and baked
Aluminium unbaked and baked
Sealed anodisedaluminium unbaked
Unbaked NEG
Baked NEG
Unsealedanodised
aluminium baked
Unsealed anodised aluminium unbaked
10-2
10-3
10-4
10-5
10-6
10-7
1013 1014 1015 1016 1017 1018
Technical surface Unbaked Baked at 150 ºC
Copper Cu-DHP acid etched 1,4 1,9
Stainless steel 304 L vacuum fired 1,3 1,5 (at 300 ºC)
Aluminium degreased 3,5 3,5
Sealed anodised aluminium 12 V 24,9 not measured
Unsealed anodised aluminium 12 V 537,5 556,0
NEG St 707 70,3 156,3
( )( )P
P P =
1
+ - 1
PP
P
P
Sat m m Sat m Satθ α θαα θ θ−
≈1
R =AA
A A
R
G
m
G
=× θ
0
1
2
3
4
5
6
7
8
9
10
0 0.1 0.2 0.3 0.4 0.5P / PSat
P /
[ ⎝(P
Sat-P
)] [T
orr.
l]-1
V. Baglin. CERN Vacuum Technical Note 1997
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200622
H2 adsorption isotherm on cryosorbersWoven carbon fiber developed by BINP
• Capacity :
1018 H2/cm2 at 6 K1017 H2/cm2 at 30 K
Capacity
V. Baglin et al. EPAC’04, Luzern 2004.
X 25
X 10 000
V. Anashin et al. Vacuum 75 (2004) 293-299
R ~ 103 RCu
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200623
3. Cryosorbers in cold systems.
Case of the LHC superconducting magnets operating at 4.5 K
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200624
The CERN Large Hadron Collider (LHC)
• 26.7 km circumference
• 8 arcs of 2.8 km
• 8 long straight sections of 575 m
• 4 experiments
• 7 TeV
• 1st beam in summer 2007
ATLAS
CMSALICE
LHCb
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200625
LHC dipole vacuum system
• Cold bore (CB) at 1.9 K• Beam screen (BS) at 5-20 K (intercept thermal loads)
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200626
LHC vacuum system principle
• Molecular desorption stimulated by photon, electron and ion bombardment (see beam-vacuum talks)• Desorbed molecules are pumped on the beam vacuum chamber• 100 h beam life time (nuclear scattering) equivalent to ~ 10-8 Torr H2 at 300 K
In room temperature elements
• Molecular chemisorption on NEG coating and in sputter ion pump (see getter talks)
In cryogenic elements
• Molecular physisorption onto cryogenic surfaces (weak binding energy)• Molecules with a low recycling yield are first physisorbed onto the beam screen (CH4, H2O, CO, CO2)and then onto the cold bore• H2 is physisorbed onto the cold bore
Dipole cold bore at 1.9 KDia. 50/53 mm
Beam screen5 - 20 KDia. 46.4/48.5 mm
Cooling tubesDia. 3.7/4.8 mm
Photons
Holepumping
Wallpumping
Desorbed molecules
Electrons stripes
36.8
mm
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200627
LHC dynamic pressure with perforated beam screen
End Window
Ti
Ti
CollimatorValve
`Photon Stop'
ConductanceConductance
DryTMP
DryScrollPump
RotaryPump
TMP
IonPump
Valve`Isolation'
IonPump
TMP
RotaryPump
e-
BA 6BA 4
BA 5
BA 3BA 2 BA 1
RGA 2RGA 3
Extractorgauge
EPA
Capacitancegauge
ControlVolume
Gaz supply
Cold bore
Beam screen(liner) Valve VVS 3
`DownstreamIsolation'
• EC = 194 eV• 3 1016 photons.s-1.m-1
• BS ~ 5 - 100 K• CB ~ 2.5 - 4.2 K
0.E+00
2.E-10
4.E-10
6.E-10
8.E-10
1.E-09
0.0E+00 5.0E+20 1.0E+21 1.5E+21 2.0E+21 2.5E+21
Dose (Photons/m )
RG
A 3
Par
tial P
ress
ure
Incr
ease
(Tor
r)
No holes
~ 2% holes
Parasitic pumping (CB ~ 3.5 K)BS ~ 5 K
BS ~ 19 KCB ~ 4 K
Magnet cold bore at 1.9 KDia. 50/53 mm
Beam screen5 - 20 KDia. 46.4/48.6 mm
Cooling tubesDia. 3.7/4.76 mm
36.8
mm
Photon
Desorbed molecules
Wallpumping
Holepumping
V. Baglin et al. EPAC’00, Vienna 2000.
• Equilibrium pressure
• Equilibrium coverage
n = C
eqη &Γ
CS = m'
0
θηησθ ⎟⎟⎠
⎞⎜⎜⎝
⎛eq
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200628
LHC Long straight section vacuum system
• Dynamic vacuum ⇒ perforated beam screens
• 1.9 K cold bore (~660 m, arc beam screen technology)• ~ 4.5 K cold bore ( ~ 740 m) ⇒ cryosorbers
• Required performances (for installation of 200 cm2/m):• Operates from 5 to 20 K• Capacity larger than 1018 H2/cm2
• Capture coefficient larger than 15 %
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200629
Why cryosorbers ?
Design requires > 100 h life time with 4.5 K cold bore and thick surface coverage⇒ porous surface
Saturated vapour pressure from Honig and Hook (1960)
1E-13
1E-12
1E-11
1E-10
1E-09
1E-08
1E-07
1E-06
1E-05
1E-04
1E-03
1E-02
1E-01
1E+00
1E+01
1E+02
1E+03
1 10 100 1000
Temperature (K)
Pres
sure
(Tor
r, 30
0 K
)
He
H2
CH4
H2O
N2
CO
O2
Ar
CO2
Beam screen
100 h beam life time
4.5 K1.9 K
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200630
Performance with perforated cryosorbing screens
• ~ 2 % holes• Charcoal cryosorber• 206 cm2.m-1
• OFE Cu• 2.2 m• ID 47 mm
Strips with charcoal
4
547 / 50
113
10
6
1.E-11
1.E-10
1.E-09
1.E-08
1.E-07
0 5 10 15 20 25 30Time (h)
CO
LDEX
Par
tial P
ress
ure
(Tor
r)
10
15
20
25
30
35
40
45
50
55
Tem
pera
ture
(K)
Η2CH4COCO2
• 100 monolayers of H2 condensed onto the beam screen prior SR irradiation
• 2 1019 H2/cm2 condensed onto the cryosorber (20 x required capacity)
• CB > 70 K
• BS ~15 and 20 K
• Below design pressure (10-8 Torr)
V. Baglin et al. EPAC’02, Paris 2002.
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200631
LSS cryosorbers regenerationPrinciple of operation
• The cryosorbers installed onto the back of the BS provide the required capacity and pumpingspeed for H2. They are located in cryoelements operating with 4.5 K cold bores
• A regeneration during the annual shutdown for removing the H2 is required
• The active charcoal is regenerated at ~ 80 K
• While regenerating, the beam is OFF and the BS should be warmed up to more than 80 K and the CB held at more than 20 K (emptying cold mass)
• While the H2 is desorbed from the cryosorbers, it is pumped by an external pumped system.
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200632
4. He leaks in cold systems.
LHC beam tube case
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200633
LHC : Superconducting technology
• Air leak or He leaks could appear in the beam tube during operation :the consequences are risk of magnet quench, pressure bump and radiation dose
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200634
Pressure at the level of the leak
Pressure73.5 m
away from the leak
Quench limit :4 10-7 Torr
E. Wallén, JVST A 15(6), Nov/Dec 1997
Example : LHC Test stringLeak rate 6 10-5 Torr.l/sDistance 75.3 m
Description of a He leak
• A He pressure wave is developed with time along the beam vacuum chamber• The He wave can span over several tens of meter without being detected• The local pressure bump gives a local proton loss (risk of quench)
2 Q XF - XF
QQHe front
D
1 mPressure
Distance from leak
xF(t1)x
Po,t2
Po,t1
xF(t2)
PxF
P. Hobson et al. J.Vac.Sci. A. 11(4), Jul/Aug 1993, 1566-1573
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200635
Time to provoke a quench
0
30
60
90
120
150
1.E-07 1.E-06 1.E-05
He leak rate at 300 K (Torr.l/s)
Tim
e in
pre
senc
e of
He
leak
(day
s) Nominal1/3 of nominal1/10 of nominal
He leak rate with risk of quench
• Lower leak rate :Require a pumping of the beam tube on the yearly basis (cold bore >~4K)
• Larger leak rate will provoke a magnet quench within :30 to 100 days beam operation for He leak rate of 10-6 Torr.l/sA day of beam operation for He leak rate of 10-5 Torr.l/s
• 1 year of operation ~ 150 days
Helium leak rate above 5 10-7 Torr.l/s shall be
detected to avoid the risk of a quench
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200636
Vacuum gauge diagnostic
• By design, a vacuum gauge is placed every 3 or 4 cells (320-428 m) i.e. 6 to 8 gauges per arc
• Small probability to detect a He leak in the arcs
• At nominal beam current, leak rates below 2 10-7 Torr.l/s are systematically detected by the arc vacuum gauges before a quench occurs
• So, one needs to find other diagnostic means
The vacuum gauges does not allow a systematic on line detection of the He
leaks in the LHC arc !
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200637
Beam loss monitor based diagnostic
• BLM provide a measurement every 53 m at each arc quadrupole
• All quench due a He wave half-length larger than 26 m are detected by the BLM
At nominal beam current, leak rates below 1.3 10-6 Torr.l/s are systematically detected by the BLM system before a quench
Maximum He leak rate which can be detected by the BLM system before a quench occurs
1.0E-06
1.0E-05
1.0E-04
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Beam current (A)
He
leak
rat
e at
300
K (T
orr.
l/s)
Nominal1st year
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200638
Cryogenic based diagnostic
• Per cell (102 m), a power of ~ 1 W/m can be measured in the cold masses
• Requires stable operation to integrate over a long time (~ 6h)
At nominal beam current, leak rates below 1 10-5 Torr.l/s are systematically detected by the
cryogenic system before a quench
Remaining time, after the dissipation of 1 W/m in the cold mass, before a quench occurs
0
50
100
150
1.E-07 1.E-06 1.E-05 1.E-04
He leak rate at 300 K (Torr.l/s)
Tim
e be
fore
que
nch
(day
s)
Nominal1/3 nominal1/10 nominal
0
6
12
18
24
1.E-06 1.E-05 1.E-04
He leak rate at 300 K (Torr.l/s)
Tim
e be
fore
que
nch
(h)
Nominal1/3 nominal1/10 nominal
Remaining time, after the dissipation of 1 W/m in the cold mass, before a quench occurs
0
50
100
150
1.E-07 1.E-06 1.E-05 1.E-04
He leak rate at 300 K (Torr.l/s)
Tim
e be
fore
que
nch
(day
s)
Nominal1/3 nominal1/10 nominal
0
6
12
18
24
1.E-06 1.E-05 1.E-04
He leak rate at 300 K (Torr.l/s)
Tim
e be
fore
que
nch
(h)
Nominal1/3 nominal1/10 nominal
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200639
Suspected leak in a given area
(1)• If possible, the cold mass temperature shall be temporarily increased to 4 K to move the He fronttowards the nearest short straight section
• Vacuum gauge or residual gas analyser can be locally installed, at this short straight section, in the tunnel for a leak detection
• Mobile radiation monitors can be installed to monitor a “radiation front” while the beam is circulating
(2)
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200640
In case of a quench
• After a quench, the faulty magnet is identified by the triggered diode.
• The cold bore is warmed up to more than 30-40 K during the quench
• The He is flushed to the nearest unquenched magnet and condensed over ~ 10 m
• The He shall be flush to the nearest SSS to perform a leak detection with a RGA
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200641
Repair
• Warm up of the cold bore to > 4 K and pump out of the He every month allows to operate with leak rates up to:
- 2 10-6 Torr.l/s and 1/3 of nominal beam current- 4 10-6 Torr.l/s and 1/10 of nominal beam current
• Time estimate ~ 1 day
• Leak rates larger than 4 10-6 Torr.l/s require an exchange of the magnet
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200642
Exchange of a magnet
Time estimate ~ few weeks !BE SURE that the OBSERVED QUENCH is due to a He LEAK at THE GIVEN MAGNET
… otherwise …
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200643
Exchange of a magnetBE SURE that the OBSERVED QUENCH is due to a He LEAK at THE GIVEN MAGNET
… otherwise …
Lyn EvansLHC Project Leader
Vincent Baglin CERN Accelerator School - Platja d'Aro - 16/24 May 200644
Thank you for your attention !!!
The guy responsible of the magnet exchange will be …