Non-Evaporable Getters

Size: px

Start display at page:

Download "Non-Evaporable Getters"

June Adams
6 years ago
Views:

1 Non-Evaporable Getters Dr. Oleg B. Malyshev Senior Vacuum Scientist ASTeC Vacuum Science Group, STFC Daresbury Laboratory, UK VS-2, October 2011, Ricoh Arena, Coventry 1

2 Sorption and diffusion Gas atoms or molecules (adsorbate) Solid surface (adsorbent) Gas, adsorbate Sticking probability s 1 Physisorption (dipole or van der Waals forces) Chemisorption (covalent linkage) adsorbate desorption surface diffusion Binding energy Reflecting probability (1 s) Solid adsorbent Handbook of vacuum technology. Ed. K. Jousten, Weley-VCH, Weinheim, 2008, Chapters 6,11 diffusion absorbed molecules VS-2, October 2011, Ricoh Arena, Coventry 2/40

3 Classification of sorption pumps Sorption pumps Adsorption pumps (physi- and cryosorption ) Getter pumps (physi- and chemisorption) With ionisation Without ionisation Orbitron pumps Ion getter pump Evaporation pumps Bulk getter pumps NEG pumps NEG coating VS-2, October 2011, Ricoh Arena, Coventry 3/40

4 Getters Gas accumulation in getter materials due to: Adsorption at the surface Absorption (i.e. solvation) in the bulk Due to diffusion, reversible process Chemical binding Irreversible process Evaporation pumps Adsorption mainly Chemical binding at surface Covering with a fresh material after saturation Bulk (non-evaporable) getter pumps Bulk getter not only adsorb gases at the surface but also employs an effect of gas diffusion into a getter material Re-activation by heating to an activation temperature VS-2, October 2011, Ricoh Arena, Coventry 4

5 Complications in studying and using NEG Complexity of getter material and processes: Oxide layers Surface roughness Bulk structure and morphology Temperature (activation and operation) History of using the getter Gas-solid combinations Optimum condition of operation depends on applications VS-2, October 2011, Ricoh Arena, Coventry 5/40

6 NEG structure Zr, V, Ti as active metals mixed with Fe, Al Developed surface area and boundaries between grains Protective oxide layer before activation NEG powder Metal substrate VS-2, October 2011, Ricoh Arena, Coventry 6/40/

7 How adsorption works Sticking coefficient Local parameter, depends on incident angle and surface orientation Sticking probability Statistically averaged value for the surface Capture probability β Statistical value for a pump with a given geometry Q β Cannel with sorbing walls L/d=10 β α α α α VS-2, October 2011, Ricoh Arena, Coventry 7/40

Types of NEGs (SAES getters) St707 Strip The

micrometers on each side of the strip Compressed

Getter Wafer Modules (based on St101 or St707

8 Types of NEGs (SAES getters) St707 Strip The thickness of the getter layer is about 70 micrometers on each side of the strip Compressed getter pills and washers for various applications Getter Wafer Modules (based on St101 or St707 getter strips) are specifically designed to maximize hydrogen pumping speed. Cartridge pumps can be mounted on CF flanges VS-2, October 2011, Ricoh Arena, Coventry 8/40/

9 Pumping What can be pumped by NEG M + O 2 MO M + N 2 MN M + CO 2 CO + MO MC + MO M + CO MC + MO M + H 2 O H + MO MO + H (bulk) M + H 2 M + H (bulk) M + C x H y MC + H (bulk) at T > 500 C What can t be pumped Hydrocarbons, C x H y, etc. at T < 500 C He, Ne, Ar, Kr, Xe (inert gases) VS-2, October 2011, Ricoh Arena, Coventry 9/40

Activation H 2 sorption regeneration cycle First activation consiredations: If capacity Q=5000 mbar l Activation pressure P a = 10-4 mbar Pumping speed during activation S eff = 100 l/s Duration

10 Activation H 2 sorption regeneration cycle First activation consiredations: If capacity Q=5000 mbar l Activation pressure P a = 10-4 mbar Pumping speed during activation S eff = 100 l/s Duration of activation t: P a = 10-2 mbar and S eff = 100 l/s t P a = 10-3 mbar and S eff = 10 l/s t P Q S act 5 10 eff s 140hr s 1.4 hr 6days VS-2, October 2011, Ricoh Arena, Coventry 10/40

Activation temperature and duration (SAES Getters) Short (tens minutes) activation used for independent activation T 300 C allows activation during

11 Activation temperature and duration (SAES Getters) Short (tens minutes) activation used for independent activation T 300 C allows activation during vacuum system bakeout (without special arrangement) NEG partially activated during bakeout at T 250 C VS-2, October 2011, Ricoh Arena, Coventry 11/40

12 Problems and limitations Limited capacity for pumping Aging (limited number of activation cycles before full saturation) Hydrogen enbrittlement (2 x pumping capacity) Thermal fatigue (number of activation cycles before initial peel-off) Storage (in vacuum, N 2 or inert gas atmosphere) VS-2, October 2011, Ricoh Arena, Coventry 12/40

Operation at high temperature Below activation temperature High temperature increases the diffusion Continues re-activation Operation at

13 Operation at high temperature Below activation temperature High temperature increases the diffusion Continues re-activation Operation at activation temperature for high pumping speed and capacity. Requires another pump for H 2 pumping VS-2, October 2011, Ricoh Arena, Coventry 13/40

14 Use of NEGs High gas load UHV/XHV application In combination with a pump for noble gases and hydrocarbons, ex. such as TMP and SIP Purification of noble gases The only gases that can t be pumped by NEG Low power consumption Power used for activation only Hydrogen pressure regulator in UHV system By changing the NEG temperature VS-2, October 2011, Ricoh Arena, Coventry 14/40

15 NEG application in accelerators VS-2, October 2011, Ricoh Arena, Coventry 15/40

16 Usual accelerator vacuum chamber Long tube with length L >> a, where a - transversal dimension Average pressure depends on vacuum conductance u(l,a) of the beam vacuum chamber P L 1 P ql kbt 12u 2S eff z L WS-63, September 2010, Ávila, Spain 16

17 Vacuum chamber cross sections Beam pipe Circular or elliptical 4 mm d, a, b 200 mm Vacuum chamber with an antechamber for larger vacuum conductance, U b d a In dipole magnetic field Distributed pumping With NEG strips (LEP in CERN) WS-63, September 2010, Ávila, Spain 17/

18 Two concepts of the ideal vacuum chamber Traditional: surface which outgasses as little as possible ( nil ideally) surface which does not pump otherwise that surface is contaminated over time Results in Surface cleaning, conditioning, coatings Vacuum firing, ex-situ baling Baking in-situ to up to 300 C Separate pumps New (C. Benvenuti, CERN, ~1998): surface which outgasses as little as possible ( nil ideally) a surface which does pump, however, will not be contaminated due to a very low outgassing rate Results in NEG coated surface There should be no un-coated parts Activating (baking) in-situ at C Small pumps for C x H y and noble gases VS-2, October 2011, Ricoh Arena, Coventry 18/40

19 Stainless steel vs NEG coated vacuum chamber under SR V.V. Anashin et al. Vacuum 75 (2004), p EVC-11, September 2010, Salamanca, Spain 19/40

20 NEG coating for accelerators First used in the ESRF (France); ELETTRA (Italy), Diamond LS (UK); Soleil (France) first fully NEG coated; LHC (Switzerland) longest NEG coated vacuum chamber; SIS-18 (Germany) and many others. NEG film capacity for CO and CO 2 is ~1ML: If P = 10-9 mbar then 1 ML can be sorbed just in ~ s; Lab measurements of different NEG coatings often don t repeat CERN s data on sticking probability and capacity; However, NEG coated parts of accelerators work well. EVC-11, September 2010, Salamanca, Spain 20/40

21 NEG coating for accelerators (2) What is required: Input data for accelerator design: (D,E,T a ), (M, T a ), pumping capacity; Better understanding: what and why; practical do s and don t s; Further development of this coating: lower, T a, SEY; higher (M), pumping capacity; optimising for an application. EVC-11, September 2010, Salamanca, Spain 21/40

22 What NEG coating does Reduces gas desorption: A pure metal film ~1- m thick without contaminants. A barrier for molecules from the bulk of vacuum chamber. Vacuum NEG Subsurface Bulk Coating Layers Increases distributed pumping speed, S: where A sorbing surface on whole vacuum chamber surface S = A v/4; sticking probability, A surface area, v mean molecular velocity EVC-11, September 2010, Salamanca, Spain 22/40

23 Deposition method Cylindrical magnetron deposition Planar magnetron deposition Details are in the talk A3.TM.OR.13 by Dr. R. Valizadeh EVC-11, September 2010, Salamanca, Spain 23/40

24 ASTeC activation procedure O.B. Malyshev, K.J. Middleman, J.S. Colligon and R. Valizadeh. J. Vac. Sci. Technol. A 27 (2009), p WS-63, September 2010, Ávila, Spain 24

25 NEG pumping properties Pressure ratio P1/P2 measured during gas injection is used to estimate: initial sticking probability and sorption capacity WS-63, September 2010, Ávila, Spain 25

Hexagonal lattice structure Ti film V film Hexagonal lattice

26 Films deposited on Si test sample from a single metal wire Zr film. Average grain size nm. Hexagonal lattice structure Ti film Average grain size nm. V film Average grain size nm. Hexagonal lattice structure WS-63, September 2010, Ávila, Spain 26

27 CO pumping capacity [ML] CO sticking probability H2 sticking probability Single metal pumping properties Ti Zr V Hf Activation temperature [ C] Ti Zr V Hf Ti Zr V Hf WS-63, September 2010, Ávila, Spain Zr is best: Lowest activation Temp. and highest capacity Hf Ti V has highest activation temperature 27/40 27

Binary films deposited on Si test sample from twisted wires. Ti-Zr film. Average grain size 50 100 nm. Hexagonal lattice structure.

28 Binary films deposited on Si test sample from twisted wires. Ti-Zr film. Average grain size nm. Hexagonal lattice structure. Zr-V Average grain size nm Ti-V film. Average grain size nm. Hexagonal lattice structure. WS-63, September 2010, Ávila, Spain 28

29 CO pumping capacity [ML] CO sticking probability H2 sticking probability Binary alloy pumping properties Ti-Zr Ti-V Zr-V Ti-Zr Ti-V Zr-V Activation temperature [ C] Zr-V is best Ti-Zr activation temperature is lower than for Ti-V 0.01 Ti-Zr Ti-V Zr-V WS-63, September 2010, Ávila, Spain Zr-Hf was not studied 29/40

30 Ternary NEG film deposited on Si test sample from twisted Ti, V, Zr, and Hf wires and TiZrV alloy wire Ti-Hf-Zr twisted wire V-Hf-Zr twisted wire Ti-Zr-V alloy wire Ti-Zr-V twisted wire Cylindrical Magnetron: Power = 60 W, P Kr = 10-2 mbar, deposition rate = 0.12 nm/s, T = 120 C. 30 Average grain size 5 nm. Hexagonal lattice structure.

31 CO pumping capacity CO sticking probability H2 sticking probability Ternary alloy pumping properties Ti-Zr-V Hf-Zr-V Ti-Zr-Hf Ti-Hf-V Activation temperature [ C] 1 Hf-Zr-V, Ti-Zr-Hf and Ti-Hf-V are comparable Ti-Zr-V has the highest activation temperature WS-63, September 2010, Ávila, Spain 31

32 CO sticking probability Twisted wires vs. alloy target: reducing Ta 1.E+00 1.E-01 TiZrV(twisted wires) TiZrV (alloy wire) TiZrV (alloy wire) 1.E Activation temperature, o C R. Valizadeh, O.B. Malyshev, J.S. Colligon, V. Vishnyakov. Accepted by J. Vac. Sci. Technol. Aug WS-63, September 2010, Ávila, Spain 32

33 Quaternary NEG alloy film deposited on Si test sample from twisted Ti, V, Zr, and Hf wires. Cylindrical Magnetron: Power = 60 W, P Kr = 10-2 mbar, deposition rate = 0.12 nm/s, T = 120 C. Very glassy structure. 33

CO pumping capacity CO sticking probability H2 sticking probability Quaternary alloy pumping properties 1 0.

34 CO pumping capacity CO sticking probability H2 sticking probability Quaternary alloy pumping properties Ti-Zr-Hf-V Hf-Zr-V Ti-Zr-Hf Ti-Hf-V Ti-Zr-V Ti-Zr Zr-V Zr Activation temperature [ C] Ti-Zr-Hf-V is the best Hf-Zr-V, Ti-Zr-Hf, Ti-Hf-V and Zr are comparable Ti-Zr-V is lower Zr-V (best binary alloy) has the lowest activation temperature WS-63, September 2010, Ávila, Spain 34

35 Pressure in the accelerator vacuum chamber P where - desorption yield - sticking probability Improving pumping properties is limited: < H2 < < CO < < CO2 < 0.6 Reducing the desorption yields. in orders of magnitude is a realistic task WS-63, September 2010, Ávila, Spain 35

36 Reducing the gas desorption from the NEG coatings Main gases in the NEG coated vacuum chamber are H 2 and CH 4 Only H 2 can diffuse through the NEG film under bombardment or heat CH 4 is most likely created on the NEG surface from diffused H 2 and C (originally from sorbed CO and CO 2 ) Therefore the H 2 diffusion must be suppressed Where H 2 come from? WS-63, September 2010, Ávila, Spain 36

37 Reducing the gas desorption from the NEG coatings Gas molecules are contained on the NEG coating surface after exposure to air Vacuum inside the NEG coating trapped during deposition NEG Coating in subsurface substrate layer in the substrate bulk Subsurface Layers Bulk WS-63, September 2010, Ávila, Spain 37

38 Reducing the gas desorption from the NEG coatings Gas molecules are contained on the NEG coating surface after exposure to air minimise exposure to air inside the NEG coating trapped during deposition purity of discharge gas background pressure in subsurface substrate layer substrate bakeout before NEG deposition in the substrate bulk vacuum firing WS-63, September 2010, Ávila, Spain Vacuum NEG Coating Subsurface Layers 38 Bulk

Malyshev, R. Valizadeh, J.S. Colligon et al. J. Vac. Sci.

39 SEM images of films (film morphology ) columnar Best for pumping dense A first candidate for a barrier O.B. Malyshev, R. Valizadeh, J.S. Colligon et al. J. Vac. Sci. Technol. A 27 (2009), p WS-63, September 2010, Ávila, Spain 39

40 Electron stimulated desorption Modified NEG pumping properties evaluation rig: To measure sticking probability To measure electron stimulated gas desorption as a function of Electron energy Dose Wall temperature ( C) Activation/bakeout temperature Can be used for samples with: NEG coating Low desorption coating No coatings WS-63, September 2010, Ávila, Spain 40

41 Electron Stimulated Desorption (ESD) studies programme ESD as a function of Activation/bakeout temperature Electron energy Electron dose Coating density, morphology and structure Deposition conditions Substrate WS-63, September 2010, Ávila, Spain 41

42 yield [molecules/electron] yield [molecules/electron] ESD: stainless steel vs activated NEG coated vacuum chamber H2 CH4 H2O CO O2 Ar CO dose [electrons/m2] Baked to 250 C for 24 hrs dose [electrons/m2] Activated to 180 C for 24 hrs O.B. Malyshev, A. Smith, R. Valizadeh, A. Hannah. Accepted by J. Vac. Sci. Technol., Aug WS-63, September 2010, Ávila, Spain 42

43 Electron stimulated NEG activation P1 P2 Activated at 180 C Electron bombardment 1 Electron bombardment Non-activated NEG 10 G1 G The electron stimulated NEG activation efficiency estimated as < 1 < Surface coverage -3 [CO/e [monolayers] - ] WS-63, September 2010, Ávila, Spain CO 1 e D 43 CO

44 (E e- ) for different gases for NEG coating EVC-11, September 2010, Salamanca, Spain 44

45 NEG cartridges vs films Bulk NEG NEG film Thickness 50 m 5 mm Materials Zr-V-Fe (St 707) Zr-Al (St 101) Ti-Zr-V (CERN) Ti-Zr-Hf-V (ASTeC) Activation T C C Pumping capacity Large 1 ML for CO, CO 2 No. activation cycles before full staturations A few hundreds < 100 Working range HV, UHV, XHV UHV, XHV Activation by photon or electron bombamdment n/a possible Main purpose pumping Barrier coating to reduce outgassing + pumping VS-2, October 2011, Ricoh Arena, Coventry 45/40/

46 Conclusions NEGs pay and increasingly important role in vacuum technology including UHV/XHV They have wide range of applications, including particle accelerators NEG can be activated in vacuum by heat, NEG films can also be activated by photon or electron bombardment Ones activated no power or controllers required for operation No noise and cooling channels In some applications no cables, feedthroughs To use them some special knowledge of NEG installation, activation operation are essential VS-2, October 2011, Ricoh Arena, Coventry 46/40

Beam induced desorption

Accelerators in a new light Beam induced desorption Dr. Oleg B. Malyshev, ASTeC Vacuum Science Group, STFC Daresbury Laboratory, UK oleg.malyshev@stfc.ac.uk CAS on Vacuum for Particle Accelerators 2017