VACUUM SYSTEM FOR THIN FILM DEPOSITION

A. A. KHERUMENKO

(Presented by the scientific seminar of the Research Institute of High Voltages)

One of the most widespread methods at present for depositing thin films of solid substances is the method of vacuum evaporation. There are many different types of vacuum systems for depositing thin films on various substrates [1]. However, many vacuum systems have a number of significant shortcomings. Thus, for example, in systems in which fore-vacuum and oil-vapor diffusion pumps are used for evacuating the system, the possibility is not excluded that vacuum oil may enter the structure of the deposited film, even when various types of oil-freezing and oil-trapping devices are used; this substantially affects the results of physical investigations of films. Holland [2] investigated contaminated ultrahigh-vacuum systems and found that, with two nitrogen traps connected in series and a vacuum in the system of \(10^{-9}\) torr, oil vapors condensed on a cold glass substrate at a rate of \(5\ \mathring{A}/\text{hour}\). With one nitrogen trap and a vacuum in the system of \(10^{-9}\) torr, the rate of film formation varied from 38 to \(160\ \mathring{A}/\text{hour}\).

Eliminating the entry of oil vapors into the structure of the deposited film by freezing out the entire system, by placing it in liquid nitrogen or hydrogen, is a rather complicated technical problem.

From the standpoint of simplicity and convenience in operation, oil-free pumping systems are of known interest; these use titanium getter-ion pumps and active gas absorbers, for example zeolite cooled with liquid nitrogen.

The vacuum system consists of a glass bell jar with a diameter and height of 200 mm, with a metal rim glued to the bell jar with epoxy resin. The rim is mounted through a vacuum seal on a metal plate of stainless steel, to which, from below, a titanium pump and cylinders with zeolite are connected through valves. For degassing the titanium pump at the first start-up, and for convenience in operation, a fore-vacuum pump RVN-20 is installed, connected to the system through a valve, and an oil-vapor pumping unit VA-05-4. These pumps were used to evacuate the system in order to investigate the effect of oil vapors entering the film structure on its electrophysical properties. The use of a cylindrical deep plate as the base for the glass bell jar instead of a flat plate has fund—

The main advantage is that radiant heat from the evaporators does not heat the glass walls of the bell jar or the sealing material, since the plate is cooled by running water. Multiampere vacuum-tight electrodes with collet clamps, cooled by water, are introduced through the lateral cylindrical surface of the plate. Fluoroplastic-4 was used as the sealing material, since vacuum rubber releases a large amount of contaminants into the system [3]. All vacuum seals are structurally made in the form of an annular rectangular shim fitting into a groove. In addition, provision is made for the possibility of replacing all fluoroplastic seals with metal ones, with clamping of the metal ring between two annular projections of trapezoidal profile. For vertical movement of the substrate and for positioning it above the various evaporators, the simplest and most effective method of displacement was used, carried out by means of an electromagnet outside the bell jar. The substrate was heated by means of a spiral serpentine heater. A thermocouple was used to measure the substrate temperature.

In the studies of Berger, Greenhouse, and Nicholas [4] it was shown that, when ordinary evaporators in the form of spirals and boats were used for evaporating silicon monoxide and nichrome, ejection of macroscopic particles from the evaporator was visually observed; on reaching the substrate, these particles had sufficient kinetic energy to pierce through the already deposited layer.

However, their evaporator design, which eliminates bombardment of the substrate surface by macroparticles of the evaporated substance, cannot be considered technological, although the evaporator itself appears original.

For evaporating silicon monoxide, York [5] proposed another system for separating the particle flux of the evaporated substance. A round cover with a hole in the center was placed on the upper part of a cylindrical crucible made of tantalum sheet and heated by a cylindrical spiral. Inside the crucible, somewhat below the cover, a tantalum disk was mounted coaxially; its diameter was larger than the diameter of the hole in the cover. Thus, direct exit of macroparticles of the evaporated substance that could reach the substrate was prevented.

In the installation under consideration, two types of evaporators were used. Metals were evaporated from cylindrical quartz, alundum, and graphite crucibles heated by a cylindrical spiral wound over the crucible from molybdenum wire of diameter 1 mm. To reduce thermal radiation, a quartz tube was placed over the molybdenum spiral.

To remove macroparticles from the evaporated flux, a diaphragm in the form of a truncated cone without a bottom was inserted into the upper part of the crucible; its base fitted tightly against the inner surface of the crucible, while the smaller-diameter part lowered into the crucible was spaced at some distance from this surface. Below the diaphragm a disk was fastened, the diameter of which was somewhat larger than the smaller opening of the conical diaphragm, but there was a sufficient gap between this disk and the inner cylindrical surface of the crucible. Since the entire separating device was located inside the crucible, the diaphragm and disk were heated to a temperature preventing condensation of the evaporated metal on them.

For evaporating dielectrics, alundum and quartz crucibles were used; these were placed inside a tantalum cylinder made by contact welding from sheet, through which current was passed. To reduce thermal radiation, over the heating cylinder …

...a second cylinder of larger diameter was fitted over it. The upper ends of the cylinders were welded by contact welding, and the lower ends were connected to the current-carrying electrodes. Thus, the current flowed through the inner cylinder and then through the outer one. In addition, the entire heating element was enclosed by two cylindrical coaxial cylinders of larger diameters.

Such a design of the evaporator provided sufficiently uniform heating of the crucible, which is very important, since the dielectric must evaporate without chemical decomposition, something more difficult to achieve when ordinary evaporators are used. The design of the separator was the same as for metal evaporators.

Depending on the tasks at hand, the installation made it possible to obtain both individual films of various substances and capacitors in the form of a system of metal–dielectric films without breaking the vacuum. Without heating the entire vacuum system and without using special traps, the titanium pump evacuated the system to a vacuum of \(10^{-7}\) torr with all other pumps switched off.

REFERENCES

1. Holland. Deposition of Thin Films in Vacuum. GEI, Moscow–Leningrad, 1963.
2. Holland, “Rev. Seient. Jnstrum,” 33, No. 4, pp. 377–382, 1963.
3. Holland, “Vacuum,” 13, No. 3, pp. 173–180, 1963.
4. Vergara, Greenhause and Nichkolas, “Rev. Seient. Jnstrum,” 34, No. 5, pp. 520–522, 1963.
5. York, “J. Electrocken Soc,” 110, No. 4, pp. 271–275, 1963.