Search

Surface contamination of silicon produced by ion implantation is examined by ESCA and IMA, and following results are disclosed. The contaminant film thickness can be evaluated by using ESCA. Deposition rate of the film increases with increase of ion current density and partial pressure of hydrocarbons in the vacuum system, and decrease of ion energy and substrate temperature. The contaminant film is very tenacious. Heat treatment in air can remove the film, but heat treatment in vacuum cannot. Surface contamination produced by ion implantation is formed by polymerization of hydrocarbon molecules, mainly originated from diffusion pump oil, caused by high energy ion bombardment.

A high vacuum evaporation apparatus for semicontinuous preparation of ferromagnetic thin films is described, which consists of a operation chamber and its associated vacuum system. The vacuum chamber, 300mm (inner dia.) ×800m.rn. mild steel Tee, has a door flange on each side and contains the demountable operating system, having a store-box for non-treated samples, driving mechanism, magnets, ovens-for outgassing, heating on evaporation and annealing samples-, reservoir for finished samples, a evaporation source, and etc. on a base frame provided with four wheels.And, all operations are manipulated by vacuum sealed driving-shafts inserted in each adaptor.The glass substrate fixed on a holder is put on the chain conveyer from the store-box and moved place by place and subjected to the serial treatment processes, i. e., outgassing, deposition, and annealing at each position and finally sent into the sample reservoir. Ten substrate holders or more can be charged in the store-box and two samples can be prepared at the same time under same conditions in every case.The pressure during each operation except for the evaporation can be maintained at 57×10-6 Torr.The pressure during evaporation is usually kept between 13×10-5Torr and can be reduced further to 56×10-6Torr by applying a liquid air cold trap.The demountable operating system and its each part is described in detail.The pumping system consists of a rotary pump (500l/min) and a three-stages oil diffusion pump with water-cooling baffle.

The factors that determine the ultimate vacuum in the system of oil diffusion pumps and sealed off tube are studied.In the case of two diffusion pump system, the ultimate vacuum of 1 ×10-10mmHg (measured value) without liquid nitrogen trap is obtained by adjusting in put power of boiler heater.In the one diffusion pump (Hickman pump), the ultimate vacuum of 1. 5×10-10mmHg is obtained. But input power of boiler in the first case is larger than later case. In order to produce and maintain ultra high vacuum in the sealed off tube with B A ionization gauge, heating loop and parts must be made specially and thus pressure of 1. 5×10-10mmHg can be maintained while electrodes are heated to 2800°K.

The mechanism of the oil contamination in a DC-705 oil diffusion-pumped uhv system has been investigated using a high sensitive residual gas analyzer. The system was equipped with a rotary pump and a 4-in. diffusion pump with a water-cooled baffle. It was found that main contaminants in the system come from the cracking of the rotary pump fluid in the diffusion pump boiler. The degradation of the ultimate pressure is proportional to the product of the concentration of the rotary pump fluid and the cracking rate in the hot boiler. In order to prevent the inflow of the fluid into the boiler, it is effective to make a temperature gradient along the foreline tube of the diffusion pump and to evaporate the high vapor pressure fluid on the wall surface warmed up to 70-100°C. For the system being out of operation, we have proposed to place a bakable valve on the foreline which is closed except when both the foreline tube and the valve are warming up with the diffusion pump operation.

The substitute sensitivity factor (φ) for the single ionization gauge method was obtained experimentally for propane and helium gas which had been used for the leak hunting at the laboratory with three fixed leaks and an oil diffusion pump system.A series of fixed leak was made by flattening an annealed copper tube with inner diameter of about 1 mm and a flattened length of all the leaks was in 14 mm. The flow rates of those leaks were stably 3.0×10-3, 2.1×10-4, and 1.3×10-5 mm.Hg·1·sec-1 for air from an atmosphere to vacuum respectively.When a probe gas was covering over the leak or removing from it, transitional changes of an ionization gauge, which electron currents was controlled automatically at a constant value by electric circuit, were recorded on an electrical self-recorder.As to the φ values of the probe gases on three fixed leaks fixed values were not obtained just as the reference showed for other gases, but these values spreaded in a wide range as follows : Probe gas leak rates of fixed leaks for air in mm Hg·l·sec-13.0×10-3 2.1×10-4 1.3×10-5Propane 2613.8 13.64.8 6.83.3Helium - (0.960.92) - (0.930.89) - (0.880.85) In order to anaiyze these experimental data, flow rates of the gases were directly measured by the constant pressure method having a small dead volume. The characteristics of flow ratio to air are given as changing from viscous flow to molecular one as the leak rate of air becomes smaller.The spread of φ is thought to be due to the fact that the pumping speed of oil diffusion pump system is unstable at transitional pressure rather than the ionization cross section of probe gases to electrons varies in the gauge.

An ultrahigh vacuum system which is composed of a Hickman type oil diffusion pump, a sorption pump and a non- refrigerated isolation trap using molecular sieve 13x as active material was tested. This combination has advantages of protecting the evacuation system from dangers by electricity failure. suspension of water-supply and of avoiding the trouble for a steady supply of refrigerant.The results obtained are summarized as follows.1) Amounts of air sorbed on molecular sieve 13x at 14°C and -183°C were about 2.5 l·μHg /g and 91446 l·μHg/g. respectively (at an apparent critical backing pressure=85μ) after bake -out at 400°C 26 for hours.2) Maintaining time of ultimate pressure depends on the activity of molecular sieveWhen a sorption pump with new molecular sieve was operated the system could be kept at an ultimate pressure below 5×10-9 mmHg for about 4 days. but with fairly used up one it was below 7×10-9mmHg for about 2 days.

A TiC getter pump with evaporation source of TiC heater was made in order to evacuate a coating apparatus of mirror of 2.0m diameter.According to previous report the TiC bars were inserted into the ordinary alumina crucible which was a reservoir of molton metal. This time Ti is used in a wire form and is fed constantly onto a TiC heater. The conclusions are as follows.1) TiC is useful as Ti evaporating heater.2) When an oil diffusion pump was operated as a fore pump, the compound rate of exhaustion was, greatly affected by back diffusion of oil vapour. When the valve between the oil diffusion pump and the getter pump was appropriately opened, the rate of exhaustion was increased to a value 25 times as great as the rate when the valve was fully opened.3) A getter pump is not only useful for very high vacuum apparatuses with metal seals, but also, the conventional vacuum apparatus with synthetic rubber seals.

A modified Nottingham gauge described in this paper has a thin gold film deposited on glass envelope as the screen electrode. Both ends of the electron collecting electrode, a cylindrical grid type, are electronically closed by a pair of wire net. The gauge sensitivity was stable and no effect of Balkhausen-Kurz oscillation was observed in the operating range while in Bayard-Alpert gauge the effect sometimes caused trouble. As this result operation of the gauge at a larger electron emission current than that for a conventional B-A gauge was proved feasible, which in turn results in practically higher sensitivity of the gauge.Electron bombarding power required for outgassing of electrode is about 40 watt d.c.. Glass wall temperature was less than 40°C for normal operation of 0.5 milliampere electron emission. The sensitivity for nitrogen is 20.4±3, the upper limit of linearity is 4μA in ion current for nitrogen. Limit of measurement caused by soft X-ray effect was estimated to be 7×10-11 Torr of nitrogen equivalent pressure. The ultimate pressure of zeolite trap and diffusion pump system used in this study was 2× 10-10 Torr. Some behaviours of the system near the ultimate vacuum are also described.

Two series of residual gas analysis were made by using a quadrupole mass filter for two independent vacuum systems. One of the systems comprises a oil diffusion pump (4″)using silicone DC-705 as a working fluid, and the other two oil diffusion pumps (2.5″) using naphthyl derivative (mono eicosyl). The results of the analysis indicate the existence of an unnegligible amount of the diffusion oil vapor in the vacuum chamber. The two diffusion pump oils gave their own characteristic mass peaks, and tentative assignment was made for the structures of these fragment ions. It was found from the residual. gas analysis that the vacuum system of the naphthyl derivative diffusion pumps is more favorable for carrying out some kind of electron impact experiment.

A space chamber was made for testing the electron bombardment type ion rocket engines. The chamber was evacuated with a 26-inch oil diffusion pump followed by a 6-inch one. The ion beam was accelerated up to 5 kV and the beam current was in the range of 0-100 mA. The gas desorption induced by ion bombardment on the LN cooled wall is mainly discussed. The experimental result is also described on the other problem whether or not the oil diffusion pump is available to use for evacuation of mercury vapor. With cryopumping of the LN cooled wall, ion rocket engine could be operated in the pressure range of 10-6 Torr.

The chevron baffle with the two-stage type thermo-module could be kept at -47°C when the hot side of the module was cooled to 11°C by water. The final pressure was reduced by about one order and reached ca. 1 × 10-8 Torr with this baffle. The main residual gas component was the vapor of diffusion pump oil, DC-704. When the baffle was cooled by water with-out thermo-electrical cooling, the vapor of rotary pump oil was predominant. Radio-frequency mass spectrometer was used for the residual gas analysis.

Secondary electron emission (SEE) coefficient a and inelastic reflection coefficient η of Octoil-S, DC-704, and DC-705 have been studied by a pulsed-beam method in the primary range 502000 eV. For these three specimens, the maximum values of a are 2.8, 2.0, and 2.0, respectively, and the primary energies corresponding to the maximum a are equally 200 eV. The maximum values of η are 0.14, 0.17, and 0.17, respectively. SEE properties of DC-705 are analyzed by using δ-η diagram (δ =σ-η) and the effective maximum emission depth of true secondaries is determined to be 2030A. The results are compared with those for other molecular substances. It is also shown that the observation of SEE gives a very sensitive method for detecting organic contamination in vacuum systems.

The conventional evacuating system : combination of oil diffusion pump, booster pump and mechanical rotary pump, has serious faults in the case of pumping the chemically unstable matter. For instance, in the degassing stage of the molecular distillation plant, the gas is composed of air, volatile matter, water vapor and mist of unstable compound. These components are decomposed or polymerized in a evacuating system, and damage the normal opperation of pumps, especially in oil rotary pump. This is caused by less circulation of oil, and moving parts would be damaged.To conquer this difficulty, we tried to select the new pump composion ; diffusion-booster, steam ejector and water jet. This system has not any moving parts, so we can avoid adove difficulty. The performances of these pumps are as follows : Diffusion booster : Pumping speed (800900 l/sec at 10-310-2 mmHg) Critial back pressure (57mmHg) Steam ejector : Pumping speed (20 l/sec at 1 mmHg), Back pressure (40 mmHg) Water jet : Pumping speed (0. 95 l/sec at 5 mmHg), Back pressure (atmosphere).The overall performance shows very satisfactory result, and this system is now under continious operation for over a year.

The adsorption time of oil molecules on the vacuum chamber wall plays an important role on evacuation process of a vacuum system with oil diffusion pump. As the adsorption time of pump fluid molecules is considerably long, it takes a very long time until vapor of pump fluids establishes stationary flow through piping. This is the reason why ultimate pressure of a well-designed vacuum system evacuated by oil diffusion pump can be much less than the vapor pressure of pump fluids. Furthermore, such a pressure is kept for a long time, even when liquid nitrogen trap is not used, as was shown by D. Alpert.By measuring the time lag in setting up the stationary flow of oil vapor (di-2-ethyl-hexyl phthalate ; DOP), the adsorption time is obtained at 50 to 85°C on borosillicate glass. The activation energy of desorption calculated from the adsorption time is about 22 kcal/mole for DOP, which is almost equal to the heat of vaporization.Conditions for measurements are also discussed in detail.

Residual gases in a vacuum system made of stainless steel were analysed by a mass spectrometer having a partial pressure sensitivity of 5×10-13 mmHg and having a resolving power of about 150. The pumping system consists of a 4″ oil diffusion pump (Apezon oil C) followed by dry ice and liquid nitrogen cold traps. The mass spectrometer is 90° sector type and its mean ion radius is 200 mm. Its main parts are made of stainless steel and gaskets of aluminium are used for vacuum shielding. After about 20 hours evacuation and baking at 150°200°C, the final pressure of 12× 10-8 mmHg was obtained. Decreasing manner of peak heights during degassing of the ion source was observed. At the final state of the evacuating process, predominant components of the residual gases were air (N2, O2, Ar etc.), H2, H2O, CO, and CO2. Partial pressures of these gases were of the order of 10-9 mmHg. As for hydrocarbon gases, the partial pressures were less than 10-10 mmHg and it was observed that heavier components had smaller partial pressures. From these results it was supposed that hydrogen and carbon monoxide would be left as main components of residual gases in a well baked, leakless vacuum system.

Since the titanium getter pump was investigated by Prof. Herb in 1954, it has been recognized that it is a powerful tool to obtain “clean vacuum”.In Japan, valuable improvements were done by the group associated under Prof. Kumagai (J. Vac. Soc. Japan 1 22, July 1958), subsequent work has followed in Japan Vacuum Engineering Co. Ltd., and applications have been found both in nuclear machines and in some industrial high vacuum instrument such as vacuum sintering furnaces, metal vapor coating units and small electron beam furnaces.Many difficulties were experienced in manufacturing and operating the continuous evaporation type (Herb type) titanium getter pumps. In this paper cautions to be paid for troubles to be encountered in continuous running, effect of back-pumping by means of a diffusion pump on boosting a pumping speed, and ultimate pressure, are described.Conclusions for Herb type pump are as follows ; in the present step, it is not suitable for ultra-high vacuum application, while its ultimate pressure is limited in the region between 10-7 to 5 × 10-8 mmHg. The occluded gas in the titanium wire released during continuous feeding and evaporation may be a gas source which limit the ultimate pressure. But it may be useful for obtaining 10-6-10-7 mmHg in short pumping time in cooperation with the diffusion pump. It is better to appreciate the titanium getter pump as one of good trap for active gases, but poor one for inert gases and organic vapors.Flashing type absorber pump is very simple in construction, but is suitable for metal ultra-high vacuum system when it is used with a L - N trap and a diffusion pump.Triode type pump such as Huber's, and Alpert's, or Hall's modyfication of Penning type pump are suitable for glass ultrahigh vacuum system, because they have appreciable pumping speed for helium. Desorption of the occluded gases, however, will be a problem.

The IPCR 160 cm cyclotron has a vacuum chamber, i. e the accelerating chamber, of which the volume is 1×104l and the surface areas in vacuum is about 140m2. Many parts, such as dees, ion source and surface shim coils and other undesirable outgassing sources are included in the accelerating chamber. Therefore, vacuum pumps are required to obtain high vacuum in the presence of much outgassing and to have small pump-down time from atmosphere, and also to be firm from troubles.The pumping system is composed with 32″, 14″ and 6″oil diffusion pumps in series connection aims to promote self purification of oil. Experiments were carried out to confirm this purpose. Several data with regard to 32″ oil diffusion pump, including pumping speed for varying jet gaps and heater wattage, are given in this paper. In normal operation the pumping speeds of 32″ oil diffusion pump are 2×104l/sec and at the connection between the pumping system and the accelerating chamber are about 1×104 l/sec with the baffle. Rough pumping system consists of 1500l/min rotary pump and 300M3/h roots pump with 650l/min rotary backing pump. When there are much outgassing, the roots pump is connected with the 14″ oil diffusion pump in order to evacuate with large throughput. Outgassing throughput during pumping down was measured under several conditions.In early stage, the base pressure in the accelerating chamber was 1×10-5 Torr and operating pressure 12×10-5 Torr. But its base pressure has been improved to 12×10-6 Torr and operating pressure to 25×10-6 Torr after replacing rubber water pipes in vacuum with metallic flexible one.The amount of leakage of the accelerating chamber is less than 0.3 lus and total throughput about 2 lus including outgassing.

In an oil diffusion pump system having a booster and a main diffusion pump in series, readings of a BA gauge equipped at high vacuum have been observed to be increased by operating an ionization gauge placed between the two pumps. This effect appears even when the filament of the latter gauge is turned on without applying the grid voltage. The effect is largely increased with increasing forepressure of the main pump and with decreasing the heater input of this pump, and is also proportional to electron current of the BA gauge. The effect is not affected by cooling a trap, with liquid nitrogen, equipped between the main pump and the BA gauge, but diminishes by cooling a trap, with liquid nitrogen, equipped between the ionization gauge and a foreline. This effect can be explained neither by outgassing of the ionization gauge nor by escape of electrons or ions from the gauge. The most probable mechanism is that some of residual gases are decomposed on the hot filament surface and that hydrogen produced by the decomposition back-diffuses through the main pump.

This paper is a technical report of synthetic phosphate lubricants, Cellulube 220 and 300, in certain aspects of their use as diffusion pump fluids. The vapour pressures, the ultimate pressures, the stabilities, and the chemical properties have been examined for these fluids. The ultimate pressure obtained experimentally by using the fluid in the metal pump only with' an unrefrigerated chevron baffle is 1.0 × 10-8Torr or lower. It is presumed that the chemical species is not tricresyl phosphate, but is probably trixylenyl phosphate, although it is written to be triaryl phosphate in the commercial literature of the manufacturer.

The definitions of “true speed” and the influence of the shape of the test dome on the “measured speed” of diffusion pumps are reviewed. Equations are given for the measured speed to be expected from an “ideal pump” (for which no molecules return to the test dome) and for the molecular conductance (as measured with a tubulated gauge) of a pipe directly connected to a pump of the same diameter. Approximate equations are also given for the influence on the reading of a nude ionization gauge of the average gas temperature due to molecules evaporating or scattered from the cold surface of a cryopump or from the hot surfaces of a diffusion pump.

An evaporating system using a liquid helium cryopump was constructed and its pumpdown curves and pressure change during evaporation were measured.Using this system, the internal stress and the electrical resistivity of deposited aluminum films were investigated. They are compared with the properties about the film deposited in the diffusion pump system and the difference between them is discussed.

The characteristics of bakable metal ultra-high vacuum system using two oil diffusion pumps in series are presented. The construction of each parts in vacuum system : indium valve, liquid nitrogen trap with chevron baffle, freon baffle, metal sealed flange and etc are described in part I. The back diffusion of gas molecules against oil vapour streams from low vacuum side to high vacuum one is studied and effects of the degassing on final vacuum are discussed in part II. As the back diffusion using helium gas, the compression ratio Pf/PO (Pf : fore vacuum pressure, PO : high vacuum pressure) at different boiler in puts 430W, 540W are respectively 4.3 × 103 and 3 × 104. The ultra-high vacuum system has attained to the pressure of 10-10mmHg.

Gaede graduated at the age of 23, he demonstrated his rotary mercury pump at 26, invented the molecular pump at 33 and the diffusion pump at 35—truly remarkable achievements, to be followed 20 years later by the introduction of the gas ballast principle into vacuum pump design. A list of his outstanding publications is appended.