λ = 5 10 p (i.e. 5 cm at 10 3 Torr, or ( )

Similar documents
MOLECULAR SURFACE PUMPING: CRYOPUMPING

Non-Evaporable Getters

Vacuum for Accelerators

Industrial Surfaces Behaviour Related to the Adsorption and Desorption of Hydrogen at Cryogenic Temperature

NEXTorr HV 100 HIGHLIGHTS

ION Pumps for UHV Systems, Synchrotrons & Particle Accelerators. Mauro Audi, Academic, Government & Research Marketing Manager

Experience from the LEP Vacuum System

A SIMULATION STUDY OF THE ELECTRON CLOUD IN THE EXPERIMENTAL REGIONS OF LHC

UHV - Technology. Oswald Gröbner

Vacuum II. G. Franchetti CAS - Bilbao. 30/5/2011 G. Franchetti 1

Vacuum Pumps. Two general classes exist: Gas transfer physical removal of matter. Mechanical, diffusion, turbomolecular

Repetition: Physical Deposition Processes

Modelling of the Target Voltage Behaviour in Reactive Sputtering R. De Gryse*, D. Depla University Ghent, Krijgslaan 281/S1, B-9000 GENT, Belgium

SORB-AC Getter Wafer Modules and Panels. We support your innovation

DEPOSITION OF THIN TiO 2 FILMS BY DC MAGNETRON SPUTTERING METHOD

Chemistry B11 Chapter 6 Gases, Liquids, and Solids

Lecture 4. Ultrahigh Vacuum Science and Technology

A Vacuum point of view

Sputter Ion Pump (Ion Pump) By Biswajit

CapaciTorr HV Pumps. making innovation happen,together

Cold / sticky system. Vincent Baglin. CERN AT-VAC, Geneva

Monte Carlo Simulations of Synchrotron Radiation and Vacuum Performance of the MAX IV Light Sources

Lecture 1: Vapour Growth Techniques

Vacuum Technology for Particle Accelerators

Adsorption of gases on solids (focus on physisorption)

Plasma Deposition (Overview) Lecture 1

Ultra-High Vacuum Technology. Sputter Ion Pumps l/s

- intermolecular forces forces that exist between molecules

CapaciTorr HV 200 HIGHLIGHTS

Hydrogen Sorption in Zirconium and Relevant Surface Phenomena

NEXTorr Z 100 HIGHLIGHTS

Modern Methods in Heterogeneous Catalysis Research: Preparation of Model Systems by Physical Methods

Chemistry: The Central Science

Influence of Amorphous Carbon Deposition on the Probability for Recombination of Neutral Oxygen Atoms on Aluminium Surfaces

LECTURE 5 SUMMARY OF KEY IDEAS

Introduction to Thin Film Processing

Optimized Annular Triode Ion Pump for Experimental Areas in the LHC

Vacuum. Residual pressure can thwart the best cryogenic design. Each gas molecule collision carries ~kt from the hot exterior to the cold interior.

Pumping Speed in the Drydown Zone

Thin Film Deposition. Reading Assignments: Plummer, Chap 9.1~9.4

Vacuum. Kai Schwarzwälder, Institut für Physik Universität Basel October 6 th 2006

To move a particle in a (straight) line over a large distance

Ch. 11 States of matter

Module 5: "Adsoption" Lecture 25: The Lecture Contains: Definition. Applications. How does Adsorption occur? Physisorption Chemisorption.

AND LIMITATIONS K. BALOGH

Vacuum Technology and film growth. Diffusion Resistor

Light-Induced Atom Desorption in Alkali Vapor Cells

ADSORPTION ON SURFACES. Kinetics of small molecule binding to solid surfaces

CHEMISTRY Matter and Change. Chapter 12: States of Matter

6.5 Optical-Coating-Deposition Technologies

CHAPTER 13. States of Matter. Kinetic = motion. Polar vs. Nonpolar. Gases. Hon Chem 13.notebook

The Liquid and Solid States

Table 1: Residence time (τ) in seconds for adsorbed molecules

CHEM 116 Phase Changes and Phase Diagrams

( KS A ) (1) , vapour, vapor (USA) , saturation vapour pressure. , standard reference conditions for gases. , degree of saturation

Adsorption Processes. Ali Ahmadpour Chemical Eng. Dept. Ferdowsi University of Mashhad

An Investigation on NEG Thick Film for Vacuum packaging of MEMS

Hydrogen in the LCLS2 Beamline Vacuum

VACUUM PUMPING METHODS

1. SYNCHROTRON RADIATION-INDUCED DESORPTION

Chapter 11. Liquids and Intermolecular Forces

2. As gas P increases and/or T is lowered, intermolecular forces become significant, and deviations from ideal gas laws occur (van der Waal equation).

Lecture 10 Thin Film Growth

Evidence for structure sensitivity in the high pressure CO NO reaction over Pd(111) and Pd(100)

Chemistry 101 Chapter 14 Liquids & Solids

The Liquid and Solid States

Metal Deposition. Filament Evaporation E-beam Evaporation Sputter Deposition

Slide 1 / A gas at a pressure of 10.0 Pa exerts a force of N on an area of 5.5 m 2 A 55 B 0.55 C 5.5 D 1.8 E 18

ORGANIC CHEMISTRY. Serkan SAYINER, DVM PhD, Assist. Prof.

Supporting Information

compared to gases. They are incompressible. Their density doesn t change with temperature. These similarities are due

All chemical bonding is based on the following relationships of electrostatics: 2. Each period on the periodic table

Revision Sheet Final Exam Term

Part III: Theoretical Surface Science Adsorption at Surfaces

Supporting Information. for. Angew. Chem. Int. Ed. Z Wiley-VCH 2003

SEY and Surface Analysis Measurements on FNAL Main Injector Ring S/S Beam Chamber Material

Etching Issues - Anisotropy. Dry Etching. Dry Etching Overview. Etching Issues - Selectivity

M10/4/CHEMI/SPM/ENG/TZ1/XX+ CHEMISTRY. Wednesday 12 May 2010 (afternoon) 45 minutes INSTRUCTIONS TO CANDIDATES

Effect of Spiral Microwave Antenna Configuration on the Production of Nano-crystalline Film by Chemical Sputtering in ECR Plasma

The fundamental difference between. particles.

Gases. Chapter 5. Copyright The McGraw-Hill Companies, Inc. Permission required for reproduction or display.

Wafer holders. Mo- or Ta- made holders Bonding: In (Ga), or In-free (clamped) Quick and easy transfer

Atmospheric Analysis Gases. Sampling and analysis of gaseous compounds

Chemistry Instrumental Analysis Lecture 34. Chem 4631

CHEMISTRY LTF DIAGNOSTIC TEST STATES OF MATTER TEST CODE:

OUTLINE. States of Matter, Forces of Attraction Phase Changes Gases The Ideal Gas Law Gas Stoichiometry

Gases, Liquids and Solids

TMT4320 Nanomaterials November 10 th, Thin films by physical/chemical methods (From chapter 24 and 25)

Beam induced desorption

Some Getter Numbers. There are 2 SAES getters we are considering: ST707 [1] and ST122 [2, 3].

EE 527 MICROFABRICATION. Lecture 24 Tai-Chang Chen University of Washington

The Vacuum Case for KATRIN

4 States of matter. N Goalby chemrevise.org 1. Ideal Gas Equation

Surface Chemistry & States of Matter

High-Pressure Volumetric Analyzer

CHAPTER 6: Etching. Chapter 6 1

Chapter 11. Freedom of Motion. Comparisons of the States of Matter. Liquids, Solids, and Intermolecular Forces

Surface processes during thin-film growth

Designing Graphene for Hydrogen Storage

2. As gas P increases and/or T is lowered, intermolecular forces become significant, and deviations from ideal gas laws occur (van der Waal equation).

Transcription:

Getter pumping C. Benvenuti R&B Energy Research, Geneva, Switzerland Abstract A surface may provide a useful pumping action when able to retain adsorbed gas molecules for the duration of a given experiment. To fulfil this condition at room temperature, strong binding forces, as those resulting from chemical reactions, are required. Materials able to react with gases to form stable chemical compounds are called getters. The two main families of getters (evaporable and non-evaporable, or NEG) are presented and discussed. Special emphasis is placed on the NEG strips currently used for the vacuum systems of particle accelerators, and on the newly developed NEG thin-film coatings, in view of their possible future applications. 1 Generalities 1.1 Molecular mean free path λ 7.3T λ = cm 10 r p 20 2 π 0 where T is the absolute temperature (K), p the pressure (Torr) and (cm 2 2 15 ). For N 2, π r 0 = 4.26 10 cm 2. At 295 K, at 10 9 Torr). 2 π r 0 the collision cross-section 3 λ = 5 10 p (i.e. 5 cm at 10 3 Torr, or Only molecular flow ( λ >> wall distance) will be considered hereafter. 1.2 Conductance and surface pumping The conductance C of an orifice (zero wall thickness approximation) may be expressed as ( ) 12 1 2 T M C = 3.64 s cm 6 5 10 cm with T = absolute temperature (K), M = molecular weight. If all molecules impinging on a surface are captured, C represents the specific pumping speed S of the surface. More generally where α is the sticking probability ( 0 α 1) ( ) 12 1 2 T M S = α 3.64 s cm. 1 2 1 2 For α = 1, S 44 s cm for H and S 12 s cm for N at room temperature. 1.3 Surface adsorption capacity 2 2 The monolayer capacity of an atomically flat surface is of the order of 5 10 14 molecules cm -2. If this monolayer of gas is released in a spherical volume of 1 (surface area ~ 500 cm 2 ) the pressure 313

C. BENVENUTI increase would be 7 10 3 Torr. Therefore surface degassing is the main obstacle to achieving UHV conditions. Conversely under UHV conditions, surface pumping may be adopted. 1.4 Surface saturation time (monolayer formation) Combining surface pumping speed and capacity, the surface saturation time t sat becomes 6 sat 10 t p seconds (p = pressure in Torr) when assuming sticking probability = 1, atomically flat surface, ambient temperature, N 2 gas (M = 28). Therefore, t sat 1 second at 10 6 Torr under these assumptions. 1.5 Sojourn time, desorption probability For a molecule adsorbed with a binding energy E on a surface at a temperature T the probability of escape ν is E 0e RT ν = ν (Frenkel, 1924) where ν 0 is an attempt frequency, of the order of the vibrational energy of the adsorbed molecules ( ν 10 13 s 1 ) 0 and 3 R = 1.98 10 kcal mole 1 K 1. It is often more useful to use the reciprocal of this term, t = 1/ ν, known as mean sojourn time or mean surface lifetime E RT τ = τ 0 e. Obviously, a surface provides a good pumping action when τ is much larger than the duration of the experiment. Typical values of τ are given in Table 1 [1]. Large τ values imply large E. However, surfaces providing low E values may display a useful pumping action when reducing the working temperature. The Frenkel equation is often used in its logarithmic form. By plotting the logarithm of the measured pressure versus 1/T, a straight line is obtained the slope of which defines the energy of the considered adsorption process. Table 1: Mean stay time for adsorbed molecules at 300 K for various values of the adsorption energy, assuming τ 10 13 0 = s Energy Typical cases τ a (kcal/mol) (s) 0.1 He 1.2 10 13 1.5 H 2 physisorbed 1.3 10 12 3.5 4 Ar, CO, N 2, CO 2, (physisorbed) 1 10 11 10 15 Weak chemisorption 3 10 6 Organics physisorbed 2 10 2 20 H 2 chemisorbed 100 25 6 10 5 (1 week) 30 CO chemisorbed on Ni 4 10 9 (> 100 yr) 40 1 10 17 ( age of the Earth) 150 O chemisorbed on W 10 1100 ( 10 1090 centuries) 314

GETTER PUMPING 1.6 Gas-surface binding energies Two types of forces may bind a gas molecule to a surface: Chemical forces involving electrons, e.g., hydrogen bonding, covalent or metallic bonding, characterized by binding energies typically in the range of some ev per molecule or larger than 10 kcal/mole (0.4 ev/molecule 10 kcal/mole). van der Waals forces of electrostatic nature, e.g., dispersion or polar forces, characterized by smaller binding energies, lower than 0.4 ev/molecule or 10 kcal/mole. In the first case (chemical adsorption or chemisorption) long sojourn times are possible at room temperature (getter pumping). In the second case (physical adsorption or physisorption) pumping requires surface cooling (cryopumping) because the mean sojourn time at room temperature is too short (see Table 1). 1.7 Definition of getters and getter pump types Getters are materials able to fix gas molecules on their surface in the form of stable chemical compounds. To do so, their surface must be clean. There are two ways of producing a clean gettering surface by in situ deposition of a fresh getter film, by heating an oxidized getter to a temperature high enough to diffuse oxygen from the surface into the getter bulk. In the first case we speak about evaporable getters, in the second case about non-evaporable getters (NEG) and the required heating temperature is called activation temperature. 2 Evaporable getters The two materials most widely used as evaporable getters are barium and titanium [2]. Barium is usually sublimated from a BaAl 4 alloy by heating at ~ 900 C. It is used for pumping vacuum sealed devices (electron tubes) and the sublimation is done in one single process before sealing. It is not used for UHV applications and will not be discussed here. For more details see Ref. [3]. Titanium is the most widely used evaporable getter for UHV applications. It is usually sublimated from filaments made of Ti alloys (with Mo or Ta) heated up to 1500 C, temperature at which the Ti vapour pressure is about 10 3 Torr. Titanium films provide sticking probabilities of 1 5 10 2 for H 2 and 0.4 0.6 for CO at room temperature. Cooling to liquid N 2 temperature enhances these values to 0.1 0.3 for H 2 and about 1 for CO [4]. Other materials (Ta, Nb, V, Zr, Mo) have also been used, but their behaviour is not as good for various reasons [5]. Warning: Sticking probabilities reported in the literature for a given gas-getter combination present a large spread. Besides possible experimental errors, this spread has two distinct causes, namely: surface roughness which may result in a large number of molecule surface interactions, i.e. higher capture probability. initial surface contamination which may reduce the adsorption site density available for pumping; this effect is particularly important for experimental systems which provide a base pressure higher than about 10 9 Torr and/or when the sticking probability is measured at high pressures (10 7 10 6 Torr). 315

C. BENVENUTI At room temperature all gases but H 2 adsorbed by a Ti film remain on the surface, resulting in a progressive reduction of pumping speed (surface blocking). On the contrary, H 2 diffuses and its pumping speed is not affected by the pumped amount. On the other hand, the high (> 30 kcal/mole) binding energies prevent the desorption of gases adsorbed on Ti at practically allowed temperatures. Again H 2 represents an exception, since its lower binding energy (~ 20 kcal/mole) allows desorption by heating (see Table 1). The initial pumping speed of a Ti sublimation pump may be restored by a further sublimation process. The total pumping capacity is therefore very large and depends on the available amount of Ti in the filament. The ultimate pressure of a Ti sublimation pump is in principle not limited; in practice it may be spoiled by the presence of rare gases and methane if an adequate pumping for these gases is not foreseen. In conclusion, Ti sublimation pumps provide the following advantages and disadvantages: Advantages: large pumping speed, large pumping capacity, unlimited ultimate pressure, compact, inexpensive, easily operated. Disadvantages: no pumping for inert gases, CH 4 production (?) and localized pumping (not very suitable for conductance-limited vacuum systems, as those of particle accelerators). 3 Non-evaporable getters (NEG) NEGs are usually produced by fixing a powder of the getter material to a (metal) substrate by pressing, sintering or cathaphoresis [6]. After insertion in the system to be pumped, activation is carried out by heating. Usually NEGs are alloys of the elements of the IV B column of the Periodic Table, to which some of the actinides and rare earths may be added. Also Al is often added to increase the diffusivity of the adsorbed gases when heating. For more details on this subject see Ref. [7]. The performance of a given NEG is characterized by activation temperature, sticking probability, surface capacity, total pumping capacity (for H 2 and for heavier gases) and particulate loss. Since NEGs are highly porous (and the small grains are in poor contact with the substrate), the danger of pyrophoricity imposes a lower limit to the activation temperature (350 400 C). If compared to Ti sublimation pumping, NEG pumping presents the risk of powder peel-off (excessive heating or H 2 embrittlement) and a lower pumping capacity; however, NEGs may provide linear pumping and passive activation. If the activation temperature is compatible with the baking temperature of the chamber where the NEG is inserted, the getter may be activated during bakeout. This feature is particularly attractive because it removes the need of electric feedthroughs and powering/control systems and allows increasing the NEG surface and consequently its pumping speed. All NEGs available on the market are produced by SAES Getters. We will consider here only the types St 101 and the St 707 (Table 2), which are the best suited for UHV applications. Table 2 St 101 [8 10] St 707 [11, 12] Composition Zr 84%, Al 16% (at. %) Zr 70%, V 24.6%, Fe 5.4% Getter layer thickness ~ 0.1 mm ~ 0.1 mm Activation 750 C for about 30' 400 C for about 1 hour Porosity ~ 10% ~ 10% Substrate steel or constantan steel or constantan H 2 dissociation pressure *) log( p H ) = 4.28 + 2log(q) 7000/T log( p 2 H 2 ) = 5.14 + 2log(q) 6250/T *) Pressure p in mbar, H 2 quantity q in mbar g 1, T in degrees Kelvin. 316

GETTER PUMPING Note that the equilibrium pressure of H 2 over the St 707 (same temperature and H 2 concentration) is about two orders of magnitude higher compared to St 101. Warning: Owing to its low activation temperature, the St 707 may be ignited by spot welding. Both the St 101 and the St 707 have been extensively studied at CERN. The St 101 was selected to provide the main pumping for the Large Electron Positron Collider (LEP) [13] (see Fig. 1). The main results are the following. Ultimate pressure Fig. 1: Cross-section of the LEP dipole vacuum chamber A LEP chamber 12 m long, made of Al alloy, baked at 150 C, pumped by St 101 and a sputter-ion pump of about 30 s 1 speed reaches an ultimate pressure of about 2 10 12 Torr, mainly due to Ar and CH 4, gases not pumped by the getter. By adding six additional sputter-ion pumps, a pressure of about 5 10 13 Torr (mainly H 2 ) is obtained [14] (see Figs. 2 and 3). Making use of the St 707 fully covering the inner walls of a 3 m long stainless-steel chamber of 160 mm diameter (see Fig. 4), pressures in the low 10 14 Torr have been achieved after passive activation during a 350 C bakeout [15]. Fig. 2: Variation of the total pressure measured on a 12-m long LEP chamber equipped with a NEG pump and seven sputter-ion pumps, as a function of the number of sputter-ion pumps ignited Fig. 3: Same variation as in Fig. 2, for the argon and methane partial pressures 317

C. BENVENUTI Fig. 4: Schematic view of the St 707 total NEG pump Pumping speed variation as a function of the amount of pumped gas The pumping curves obtained at room temperature for the St 101 [16], represented in Fig. 5, indicate that, after saturation of the external surfaces, the pumping speed is limited by the conductance to the internal pores of the getter coating. A mathematical model of the pumping process points out that, for a given gas load in the coverage range of practical interest, the NEG pumping speed is proportional to the square of the coating porosity. This model allows one to estimate the porosity from the S(Q) curves, in good agreement with the values obtained by direct porosity measurements. The developed model also predicts that CO is not dissociated on the St 101 (single-site adsorption) while H 2 is dissociated (two-site adsorption) as well as N 2, which, however, occupies six to eight adsorption sites. When gas mixtures are pumped [17], CO was found to inhibit the pumping of other gases, while N 2 has a small surface blocking effect and H 2 does not produce any blocking at all. Fig. 5: Pumping speed variation of the St 101 NEG strip (30 mm wide) for H 2, CO, N 2 as a function of the amount of pumped gas 318

GETTER PUMPING 4 Thin-film getter coatings Thin-film getter coatings, produced by sputtering, have been found to recover their chemical reactivity after exposure to ambient air followed by in situ baking [18, 19]. These coatings represent the last step of a process during which the pump has been moved progressively closer to the vacuum chamber walls. The getter thin-film coating actually transforms the vacuum chamber from a gas source into a pump. Many metal coatings were studied at CERN, namely Ti, Zr, Hf, Nb, V and some of their binary and ternary alloys [18, 19]. The main goal was to reduce the activation temperature so as to allow activation also when using aluminium vacuum chambers, which cannot be baked at temperatures higher than about 200 C. This goal was achieved with a TiZrV alloy [20]. This alloy would be highly pyrophoric when used as a powder, but does not present this risk in the form of a thin film, thanks to the thermal stabilization provided by the much thicker substrate. The main results achieved using this coating are shown in Figs. 6 and 7. This study has been completed and the TiZrV coating has been applied to the vacuum chambers of many accelerators, and in particular to the warm sectors of the LHC at CERN. Fig. 6: Comparison of the ultimate pressures achieved in stainless-steel chambers coated with TiZr and TiZrV thin films. The chamber is 2 m long, its diameter is 100 mm, and it is connected at one extremity to a sputter-ion/titanium sublimation pumping station via an orifice of 25 s 1 conductance for H 2. The pressure is measured at room temperature at the opposite chamber extremity after a 24-hour baking at the indicated temperatures (without air exposure between bakeouts). Fig. 7: Comparison of the H 2 sticking factors under the same conditions as described for Fig. 6 319

C. BENVENUTI Compared to traditional NEG strips, the thin-film getter coatings provide the following advantages: reduced degassing lower activation temperature (200 C) wider choice of substrate possible materials may be produced which could not be produced otherwise (metastable and amorphous alloys) no space is required (coatings are a few microns thick) very low secondary electron emission. In spite of being thinner than the powder coatings of traditional NEG strips by almost two orders of magnitude, thin-film getter coatings may provide pumping speeds and total surface capacities very similar to those of the St 707 NEG. They may also withstand many activation/air venting cycles without excessive performance deterioration. A TiZr film has undergone over 20 such cycles without losing more than 50% of its original pumping speed for H 2. References [1] J.M. Lafferty, Foundations of Vacuum Science and Technology (Wiley, New York, 1998), p. 551. [2] Ibid., p. 275. [3] Ibid., pp. 276 290. [4] P.A. Redhead, J.P. Hobson and E.V. Kornelsen, The Physical Basis of Ultrahigh Vacuum (American Institute of Physics, New York, 1993), p. 400. [5] Ibid., p. 402. [6] G.L. Saksaganskii, Getter and Getter-Ion Vacuum Pumps (Harwood Academic Publishers, Chur, 1994), p. 258. [7] J.M. Lafferty, Foundations of Vacuum Science and Technology (Wiley, New York, 1998), p. 297. [8] T.A. Giorgi, Jpn J. Appl. Phys. Suppl. 2, 1 (1974) 53. [9] P. Della Porta, T.A. Giorgi, S. Origlio and F. Ricca, in Trans. of the 8th Nat. Vac. Symp. (Pergamon, New York, 1962), p. 229. [10] D. Ferrario, A. Barosi, M. Borghi and T.A. Giorgi, in Proc. of the 9th Symp. on Fusion Technology (Pergamon, New York, 1976), p. 51. [11] C. Boffito, B. Ferrario, P. Della Porta and L. Rosai, J. Vac. Sci. Technol. 18 (1981) 1117. [12] C. Benvenuti and P. Chiggiato, J. Vac. Sci. Technol. A 14 (1996) 3278. [13] C. Benvenuti, Nucl. Instrum. Methods 205 (1983) 391. [14] C. Benvenuti, J.P. Bojon, P. Chiggiato and G. Losch, Vacuum 44 (1993) 507. [15] C. Benvenuti and P. Chiggiato, Vacuum 44 (1993) 511. [16] C. Benvenuti and F. Francia, J. Vac. Sci. Technol. A 6 (1988) 2528. [17] C. Benvenuti and F. Francia, J. Vac. Sci. Technol. A 8 (1990) 3864. [18] C. Benvenuti, P. Chiggiato, F. Cicoira and Y. L Aminot, J. Vac. Sci. Technol. A 16 (1998) 148. [19] C. Benvenuti, P. Chiggiato, F. Cicoira and V. Ruzinov, Vacuum 50 (1998) 57. [20] C. Benvenuti, J.M. Cazeneuve, P. Chiggiato, F. Cicoira, A. Escudeiro Santana, V. Johanek, V. Ruzinov and J. Fraxedas, IVC14, Birmingham, 1998, Eds. J.P. Coad et al., Vacuum 53 (1999) 219 225. 320