Getter Pumping Research Articles

Vacuum pumping system for the multi-aperture low-energy continuously-operated electron accelerator with a high-density beam current

The vacuum pumping system for an electron accelerator has been developed on the basis of arc sources of getter film and ion getter pumps. The contradictory problem has been solved, that is, it was necessary to provide, on the one hand, the maximum efficiency of the pumping speed and, on the other, a minimal return flow of the metallic plasma shunting the accelerating gap. The program and procedure have been elaborated for the numerical simulation of getter pumping taking into account the mass-transfer non-uniformity. The experimentally obtained vacuum characteristics of the developed pumping system for the accelerator are presented.

Investigations of a method for determining pumping speed and sorption capacity of nonevaporable getters based on in situ calibrated throughput

This paper presents a gas-sorption method, based on an in situ calibrated throughput, for characterizing nonevaporable getters (NEGs) of Ti- and Zr-based alloys. The main characteristics of the method are the adjustment of a constant injected gas flow, the prompt calibration of a throughput, and the measurement of an increasing sorption pressure. It is essential that inert vacuum gauges be selected for measurements of total pressure to prevent contamination of a test gas and poisoning of an NEG under test. A capacitance diaphragm gauge is used for preparation of the injected gas flow by filling the gas reservoir with a test gas and adjusting the conductance of a variable leak valve. A spinning rotor gauge (SRG) is used for in situ calibration of the throughput by means of an integrated gas-flow calibration facility and the rate-of-pressure-rise method, and by continuous measurement of the sorption pressure in a test chamber. A getter pumping speed and a sorbed quantity are calculated while considering the throughput, the sorption pressure, and the sorption time. The throughput must be selected in accordance with both the sorption characteristics of a particular NEG type and the measurement capabilities of the SRG. At room temperature, porous thick-film NEGs and structured thin-film NEGs exhibit an initial getter pumping speed on the order of 1 and 0.1 l s−1 cm−2, respectively. The corresponding sorption capacity of a few times 10−4 mbar l cm−2 is comparable for all chemically active gases except for H2, which amounts to a few times 100 × 10−3 mbar l cm−2 in the case of porous NEGs and ∼1 × 10−3 mbar l cm−2 in the case of structured NEGs. Detailed analysis of uncertainties of SRG pressure measurements below 1 × 10−3 mbar gives an expanded uncertainty of the getter pumping speed and the sorbed quantity of 1.9% and 3.5%, respectively. While temporally interrupting the throughput, the setup enables discrete measurements of the background pressure, which increases because of the accumulation of nongetterable gases. After completion of the gas-sorption test, a qualitative determination of the accumulated inert gas composition is performed by the gas-burst method using a quadrupole mass spectrometer (QMS) mounted on a vacuum system. The quantitative analysis is enabled by the use of a QMS calibrated in situ by known amounts of a certain gas such as H2, CH4, CO, Ar and Kr.

Outgassing from stainless steel and the effects of the gauges

The true outgassing from surfaces is often masked by the production of gases and pumping within the measuring gauge. Quadrupole mass analysers have been used to measure the outgassing from carefully prepared samples of stainless steel. By using an isolation pressure rise technique coupled to non-evaporable getter pumping, the problem of production of methane by the gauge is overcome and it has been possible to measure the outgassing of methane to very low levels. The specific outgassing rate of methane from 316 L stainless steel is found to be ⩽5×10 −22 mbar l s −1 cm −2. The outgassing of the other gases has been measured using several gauges. The specific outgassing of hydrogen is found to be 4.5×10 −15 mbar l s −1 cm −2 and the total of all other gasses amounted to no more than 10% of this figure. Although the measurements of the other gases are more ambiguous than those for methane, it is concluded that, in all probability, only hydrogen is outgassed from stainless steel.

Getter requirements for a cathode ray tube with a diamond coated field emitter electron source

A single diamond coated field emitter (FE) tip could be used as a replacement for the thermionic cathode in small electron beam devices offering better optoelectronic performance. The most important parameter limiting the use of field emitting sources is their short lifetime, which is closely related to the pressure level within the pinched off device. Aside from the initial evacuation of the tube and bakeout, getter selection and the activation procedure would seem to play the crucial role. In the present study, the pressure of miniature cathode ray tubes (CRTs) was measured by a spinning rotor gauge (SRG) after the initial evacuation and bakeout procedures, again after activation of the getters, and finally during operation under typical conditions. After the evaporation of barium getters by the prescribed procedures, the pressure did not drop, but increased slowly. The accumulated gas was shown to be methane. In CRTs with a thermionic electron source, getter pumping action could be triggered by switching on the cathode heater, which efficiently cracked the generated methane. In CRTs with a diamond coated FE tip, there was no way to decrease the residual methane pressure to the desired (i.e., safe) level. The emitters, previously approved and characterized in ultrahigh vacuum, were thus irreversibly damaged ab initio. Ba getters consequently cannot be applied in devices applying diamond coated FE tips just because of geometrical constraints, as usually stated, but mainly because of methane generation. In CRTs with nonevaporable getters, a much lower pressure increase was registered by the SRG during several days. This means that they could maintain the residual atmosphere at the level that may provide a means for long-term operation of devices with diamond coated FE electron sources.

Design of the vacuum system for the elliptical multipole wiggler at the Advanced Photon Source

A vacuum system for the Advanced Photon Source elliptical multipole wiggler (EMW) that will operate at a pressure of 10−9 Torr with a storage ring current of 100 mA at 7.0 GeV has been designed and is being fabricated. The major part of the system is a stainless steel chamber with a 66.6 mm by 19.6 mm rectangular cross section. The length of the vacuum chamber is 3100 mm, and the wall thickness is 1.2 mm. Two versions of the vacuum chamber will be produced: with and without distributed nonevaporable getter (NEG) pumping. The version with NEG pumping will have slides on the top and bottom walls to accommodate sintered plates available from SAES. To activate these plates, the entire vacuum chamber will be baked from the outside up to a temperature of 350° C–450 °C. Provision for the baking is included in the design of the vacuum system, its support, and in the EMW itself. The complexity introduced into the design by the need for external activation of the NEG plates is eliminated in the design of the second version of the chamber. In this chamber, a sufficiently low outgassing rate may be achieved by extensive surface cleaning and baking in a vacuum furnace (10−6 Torr) up to a temperature of 950 ° C as has been achieved at the ESRF. Both versions are being pursued in parallel.

Some important developments in capture pumping technology in the last 40 years

Highlights of the history of capture pumping technology are given. The subjects covered include some of the important chronological developments in sputter-ion, electrostatic-ion, sublimation, nonevaporable getter, and cryogenic pumping in the last 40 years. All of these forms of pumps proved to be essential tools for the attainment of ultrahigh vacuum.

Approach to estimate gettering effects in Ti–O2 reactive sputtering process

In reactive sputtering gettering of reactive gas at the chamber wall plays an important role in process condition changes. In this article, reactive gas mass balance and sputtering yield changes in Ti–O2 reactive sputtering have been analyzed by pressure change observation and film composition analysis to examine gettering effects quantitatively. The obtained results show the drastic change in gettering effects at the mode transition which takes place at a critical O2 partial pressure. Until the mode transition, more than 90% of the introduced O2 is consumed by gettering. The amount of gettered O2 displays a maximum just before target poisoning occurs. At this point the calculated getter pumping speed reaches about ten times larger than the physical pumping speed. After the transition, getter pumping speed ranges between 1/5–1/10 of the physical pumping speed. The sputtering yield also drops by a factor of 17 at the transition. The adsorbed/incident O2 flux ratio is found to decrease from 0.36 to 0.0010 with increasing O2 partial pressure. These results quantitatively indicate the role of gettering in condition changes.

An investigation of hysteresis effects as a function of pumping speed, sputtering current, and O2/Ar ratio, in Ti-O2 reactive sputtering processes

Understanding of mode transition and hysteresis behavior in reactive sputtering is necessary to widen its applications. The effects of physical pumping speed, sputtering current, and reactive/inert gas ratio on mode transitions and hysteresis in Ti-O2 dc reactive sputtering are examined. As O2 gas flow rate is increased or decreased, mode transition takes place as a consequence of the target poisoning or cleaning. The results obtained explain the important roles of pumping speed and sputtering current on hysteresis. Also, it is found that as pumping speed increases hysteresis width decreases. In addition, for smaller sputtering currents and constant pumping speed, hysteresis width shrinks. The change in transition points as a function of pumping speed and sputtering current are successfully explained by considering getter pumping capacity or getter pumping speed. Elimination of hysteresis is not investigated in this experiment. From the discussion it is suggested that hysteresis is a result of difference in sputtering yields from a poisoned target and from a metallic target, which provides a difference in gettering capacity.

Sensitive High-Vacuum Monitoring for Philips Electron Microscopes Which Employ ION Getter Pumping

A very sensitive vacuum monitor may be easily and inexpensively added to any Philips CM or 400 series electron microscope. This digital, auto-ranging volt meter will also operate as an excellent leak detector.The newer Philips computerized transmission electron microscopes have no separate high vacuum gauge and sense high vacuum the same way as it is done on the 400 series instruments. The high vacuum monitoring technique for these microscopes is accomplished by relating the electrical current required by the Ion Getter Pump to the quality of the vacuum (i.e. the better the vacuum, the lower the current). The system for vacuum monitoring measures the voltage across a low value resistor which is placed in series with the Getter Pump and the pump power supply and is usually located near ground or earth potential. The power supply delivers about 7000 V.D.C. to the pump and the voltage across the low value resistor is proportional to the current being drawn by the pump. This voltage is measured, amplified and then compressed by a logarithmic amplifier so one meter scale or a simple digital-to-analog converter can provide a reading equivalent to the high vacuum value. Because of the logarithmic amplifier, small changes cannot be seen; only the general vacuum quality can be monitored.

A Vacuum System for the Argonne 6 GeV Synchrotron Light Source

The ANL vacuum system for the 6 GeV light source storage ring features non-evaporable strip getter pumps for uniform pumping around the ring within a gas desorption antechamber, and it also features lumped getter pumping directly under and above crotch radiation absorbers that are positioned after each bending magnet. Based on experiments at ANL in 1980 and by others, the technical and economical advantages have been established for the use of the distributed NeG pumps of non-magnetic strips coated with a nonevaporable Zr Al getter matrix. The NeG strip pump lifetime approaches ten years. The antechamber improves the isolation of the gas desorption process from the main beam chamber and beam. The combination oE these vacuum techniques; the NeG strip getter pumps, the gas desorption antechambers, and the lumped ion and lumped getter pumping provide a unique and reliable system for maintaining long beam lifetime.

Tandem-Mirror Experiment-Upgrade neutral pressure measurement diagnostic systems

The Tandem-Mirror Experiment-Upgrade (TMX-U) has a large and complex system of Bayard–Alpert, magnetron, and Penning gauges, in addition to mass spectrometers (RGA), all of which measure neutral pressures in the many internal regions of TMX-U. These pressure measurements are used as part of the confinement physics data base as well as for management of the TMX-U vacuum system. Dynamic pressures are modeled by a coupled-volumes simulation code, which includes wall reflux, getter pumping, and plasma pumping.

Technique for i n v a c u o passivation of ZrAl alloy bulk getters

We have developed a method for passivating previously activated ZrAl alloy bulk getters. Such passivation is essential for avoiding embrittlement of the getters during discharge cleaning operations in tokamak fusion devices. The method consists of exposing the getters (at room temperature) to a controlled amount of oxygen, followed by pulse discharge cleaning to restore the vacuum vessel walls to clean conditions. We have tested this method on the Tokamak Fusion Test Reactor (TFTR) and present measurements of the decrease in the hydrogen and oxygen pumping speeds as functions of the quantity of oxygen adsorbed. These measurements are in good agreement with previous laboratory results. Using this method followed by 20 h of pulse discharge cleaning, we have reduced the getter pumping speed by more than a factor of 2×103 and have recovered the same clean vacuum conditions as existed prior to the oxygen exposure.

Decreases in deuterium pumping by St707 getter alloy caused by carbon dioxide preexposure

Intentional passivation of the deuterium pumping of the solid getter alloy St707 has been attempted by exposing samples of St707 to carbon dioxide at different pressures, temperatures and exposure times relevant for application to the getter modules in the ALT-I pump limiter. It was found that one of the most effective treatments examined was a 30 min, 1 Torr exposure at approximately 100 °C. This preexposure kept the getter pumping speed less than 0.001 of its rated value for about 3 min when exposed to deuterium at 1 Torr and 30 °C. After this ‘‘incubation’’ period, the getter speed increases to values greater than ∼1% of fully activated values. If left under high deuterium pressure, the getter eventually flakes off the substrate. Video observations of the flaking process indicate that individual particles leave the constantan getter substrate at velocities of 1 m/s. Attempts at passivating the getter using oxygen and carbon monoxide were found to be no more effective than using carbon dioxide, suggesting that there is no way to completely passivate the getter with these gases at pressures low enough for application to in situ getter arrays used in tokamaks.

Tokamak pump limiters

Experiments with pump limiters on several operating tokamaks have established them as efficient collectors of particles. The gas pressure rise within the chamber behind the limiters has been as high as 50 mTorr when there is no internal chamber pumping. Observations of the plasma power distribution over the front face of these limiter modules yield estimates for the scale length of radial power decay consistent with predictions of relatively simple theory. Interaction of the in-flowing plasma with recycling neutral gas near the limiter deflector plate is predicted to become important when the effective ionization mean free path is comparable to or less than the neutral atom mean path length within the throat structure of the limiter.Recent experiments with a scoop limiter without active internal pumping have been carried out in the PDX tokamak with up to 6 MW of auxiliary neutral beam heating. Experiments have also been performed with a rotating head pump limiter in the PLT tokamak in conjunction with RF plasma heating. Extensive experiments have been done in the ISX-B tokamak and first experiments have been completed with the ALT-I limiter in TEXTOR. The pump limiter modules in these latter two machines have internal getter pumping. Experiments in ISX-B are with ohmic and auxiliary neutral beam heating. The results in ISX-B and TEXTOR show that active density control and particle removal is achieved with pump limiters. In ISX-B, the boundary layer (or scape-off layer) plasma partially screens the core plasma from gas injection. In both ISX-B and TEXTOR, the pressure internal to the module scales linearly with plasma density but in ISX-B, with neutral beam injection, a nonlinear increase is observed at the highest densities studied. Plasma plugging is the suspected cause. Results from PDX suggest that a regime may exist in which core plasma energy confinement improves using a pump limiter during neutral beam injection. Asymmetric radial profiles and an increased edge electron temperature are observed in discharges with improved confinement. The injection of small amounts of neon into ISX-B has more clearly shown an improved electron core energy confinement during neutral beam injection. While carried out with a regular limiter, this “Z-mode” of operation is ideal for use with pump limiters and should be a way to achieve energy confinement times similar to values for H-mode tokamak plasmas. The implication of all these results for the design of a reactor pump limiter is described.

Characterization of titanium films in the Tandem Mirror Experiment Upgrade (TMX-U)

Titanium gettering is used extensively in the Tandem Mirror Experiment Upgrade (TMX-U) to achieve good vacuum conditions for the production of a sustained energetic plasma. The gettering films under different operating conditions were analyzed by thermal desorption spectroscopy and Auger electron spectroscopy. A sputter profile of the films showed a Ti deposition rate of ≂1 nm per evaporation sequence. Large quantities of protium (H/Ti≊0.2), oxygen (O/Ti≊0.3), and carbon (C/Ti≊0.1) were detected in the films regardless of their location and the plasma isotope to which they were exposed. While the Ti films that were exposed to deuterium plasmas showed a substantial increase in deuterium content, deuterium still remained a small fraction of the overall hydrogen concentration (D/H≊0.1). Analyses of the results suggest that getter pumping of H2O from the background gas in the (TMX-U) vacuum system is the primary source for protium and oxygen measured in the films.

Getter pumping speed measurements in the range 10−2 to 10−7 l per s

A procedure for measurement of getter pumping speeds several orders of magnitude below the lowest reported rates in the literature is described. The method used was found to be useful in evaluating ambient hydrogen pumping characteristics for small getters in the presence of trace contaminant gases (less than 0.13 μmol). Poisoning effects are described for the action of carbon monoxide and carbon dioxide on a zirconium–titanium nickel alloy bulk getter. Results obtained indicated that the poisoning effect is much less severe in the case of barium flash getters.

Pioneer Venus Sounder Probe Neutral Gas Mass Spectrometer

A neutral gas mass spectrometer was flown to Venus as part of the Pioneer Venus Multiprobe to measure the composition of its lower atmosphere. Encounter occurred on December 9, 1978. The instrument, mounted in the Sounder Probe, was activated after the probe entered the top of the atmosphere, and it obtained data during the descent from 62 km to the surface. Atmospheric gases were sampled through a pair of microleaks, the effluent from which was pumped by a combination of ion and getter pumping. A pneumatically operated valve, controlled by the ambient atmospheric pressure, maintained the ion source pressure at a nearly constant value during descent while the atmospheric pressure varied by 3 orders of magnitude. This action preserved the more than 6-decade dynamic range of the instrument throughout the descent. A single focusing magnetic sector field mass spectrometer with mass resolution sufficient to reasonably separate argon from C3H4 at 40 amu provided the mass analysis and relative abundance measurements. A microprocessor controlled the operation of the mass spectrometer through a highly efficient peak-top stepping routine and data compression algorithm that effected a scan of the mass spectrum from 1 to 208 amu in 64 s while requiring an information rate of only 40 b/s to return the data to earth. A subscale height altitude resolution was thus obtained. Weight, size, and power requirements were minimized to be consistent with interplanetary flight constraints.

Getter pumping to allow economical sputtering with xenon

Increases in sputter rate, under equivalent commercial operating conditions, can often be achieved by using inert gases heavier than Ar. High atomic weight targets and high sputter voltages result in Kr and Xe having higher sputter yields than Ar. The larger ionization cross sections of Kr and Xe also result in a plasma enhancement. The high consumption rate of sputter gas has previously prevented the economical utilization of Kr and Xe. A d.c.-powered Ti sputter target was added to an r.f. planar diode sputtering system. The sputtered titanium was deposited directly on the vacuum chamber walls to form a large room temperature getter pumping surface. Calculations showed an increase of two orders of magnitude in the pumping speeds of O 2, N 2, H 2O and H 2 over that of a conventional throttled 10 cm diffusion pump. This pumping speed allowed improved sputter cleaning of substrates and the deposition of higher purity films. Operating the Ti getter pump with the diffusion pump closed allowed sputtering with very little inert gas consumption. R.f. sputtering of a Ti target with 750 W at 8 mTorr gas pressure resulted in a sputter gas consumption rate of 3 × 10 -3 cm 3 min -1 for both Ar and Xe. This consumption rate corresponds to a Xe cost of less than 1 ¢ h -1.

Method for Determining Getter Activity in Vacuum Devices

A commonly accepted method for determining the activity of a barium getter film within vacuum devices requires the measurement of pressure over the film. The getter pumping speed may be calculated from this measurement when the gas inlet rate and manifold pressure are known. In this work an attempt is made to evaluate the utility of determining getter activity as a function of emission decay of an oxide cathode. Mass spectrometer and emission decay data obtained from experiments carried out using either new or operated color picture tubes indicate that decay to an arbitrary level of 80%, induced by slow oxygen-poisioning, occurs rapidly within a narrow pressure range. This indicates that determination of oxygen getter capacity based on oxide-cathode emission decay may serve as an alternate method for comparing relative getter film activities.

Paper 9: Considerations in the Design of Ultra High Vacuum Systems Using Ion Getter and Sublimation Pumping

Paper 9: Considerations in the Design of Ultra High Vacuum Systems Using Ion Getter and Sublimation Pumping