DE3010925A1 - Vacuum vapour deposition appts. using electron beam evaporator - where rotary plate contg. evapn. crucibles is cooled by molten gallium resting on water:cooled box - Google Patents

Abstract

The appts. contains a rotating copper plate (I), in which is a circular row of holes forming crucibles contg. materials to be evaporated by the electron beam. Below plate (I) is a stationary and water-cooled copper plate (II); and heat is transferred from plate (I) to plate (II) by a layer of molten gallium. The layer of molten gallium is pref. conical in shape, so that any gas bubbles in the Ga can rapidly escape upwards. Instead of Ga, a gallium alloy contg. 13% indium, and with m.pt. 17 deg.C., can be used. Used in the mfr. of thin film transistors or similar structures, where several different materials are successively evaporated. With similar conventional appts. using only water cooling, water vapour enters the vacuum system and also the deposit, causing serious problems. These problems are eliminated by the invention.

Description

Description :

When vacuum evaporation of thin layers of refractory materials one uses electron beam heating of the material because one does not due to the harmful reaction of the vapor deposition material with the lying material is limited.

Water-cooled copper bowls are used as crucibles, so they stay cold. The vapor deposition material is heated from the surface. Often one wants to vaporize several substances for thin-film structures one after the other with the same electron beam without breaking the vacuum. It is easier to keep the electron beam stationary and to move the crucibles below it than the other way around. The crucibles are milled into a circular copper plate as bowls, which is water-cooled from below. Since this crucible plate must be rotated, the highly vacuum-tight supply and discharge of the cooling water is a difficult problem. Despite known leaks, the vacuum molds use metal-rubber seals similar to the one shown in Fig. 1, and mostly do not mention the fact that a surge of water vapor enters the vacuum system with every rotation of the crucible. If, in addition, the cooling water is not distilled but tap water, the solid residues or evaporating water skins remain under the rubber as crystalline scale deposits, which then act as emery and abrade the metal shaft, which leads to further leaks.

In the course of our attempts to improve thin-film transistors, this lack of vapor deposition turned out to be intolerable. Because with each rotation of the crucible plate, so much water vapor entered the vacuum system that more or less thick water-covering layers formed on the vapor-deposition layers. At higher seal wear also aer pressure of the crucible plate increased in the system at each turning so high (> Ιϋ ~ ^ψ nib), in that the electron gun is abschaltete for safety reasons. After replacing the O-rings and the rotating shaft, the old condition was restored after a short time.

This led us to the construction of the new system shown in Figure 2, v / elche eliminated this problem.

In Figure 2:

1: the electron beam of 10,000 volts, which is magnetically bent by 270 °, max. I amps

2: The evaporating material No. 1

3: The cold evaporation material No. 6th

4: The steam jet

5: Turntable copper plate with 6 wells

6: Stationary copper base plate, water-cooled

7: Lower closing plate, steel, stationary

8: Liquid gallium or GaIlium-indium metal layer for heat dissipation from 5 to 6

9: Overflow edge of the gallium basin

10: Spacer pins

11: Cooling water for Toil 6

12: Inflow pipe (welded)

13: Drain pipe (welded)

130040/0289 BATHROOM GRIGaSNiAL

-Jr-

14: Central guide pin of turntable 5, copper (iron is alloyed) 15: Stationary Viton Α seal between 6 and 7

16: Groove for drive link chain to rotate 5 (engages in 4 pegs along the circumference}

17: Drive link chain

18: Link chain drive wheel

19: Accidental gas bubble in the gallium metal, escaping

20: Groove to catch the overflowing gallium.

The design is based on the knowledge that none of the known low-vapor pressure fluids, such as vacuum pump oils, have sufficiently high thermal conductivity. Mercury has too high a vapor pressure at room temperature and is poisonous. Gallium melts at 29 C, a Ga-13 % In alloy even at 17 C, and the vapor pressure of these metals at 100 C is immeasurably small. Since their thermal conductivity is proportional to their electronic conductivity (Wiedemann - Franz's law), they are about 10 times less thermally conductive than copper (whereby pure gallium is better than the Ga-In alloy), but still better than a gas-filled or evacuated or Column filled with high vacuum pump oil.

Since there is a possibility that when heated in a vacuum in liquid Metal forms a gas bubble, it is advantageous to use the GaIliumschient as after cone surface facing upwards so that the bladder ascends quickly can. If it did not do this, the copper crucible would overheat locally 5, with catastrophic meltdown, the result.

It is therefore of great importance that the GaIlium-Metal before being filled into the system is thoroughly degassed in a vacuum at at least 500 C, as well as the copper parts 5 and 6. In addition, the copper parts 5 and 6 must be assembled before assembly have been wetted with gallium, which can be done with a small glass bristle brush (gallium wets pure copper eagerly). The gallium layer thickness can be adjusted by means of the pins 10. It should not be less than 1 mm, since with thinner layers any gas bubbles have difficulty escaping (surface tension). Driving the crucible plate 5 into the various positions does not cause any problems, as shown in FIG.

100 g gallium were required in the experimental facility, at a price of DM 150.00. The purity of 99.9 % is completely sufficient.

As a further result it should be mentioned that the furrow 20, which about catches emerging gallium, is von Mutzen. During prolonged operation, the gallium surface that comes into contact with air is formed at 9 there is an oxide skin floating there, but which has no negative effect.

Although the melting point of gallium is given as 29.8 C, we found that this liquid can be subcooled by 10 C before it solidifies. As soon as this has taken place, the crucible disk 5 can of course no longer be rotated. To avoid this, we have connected the cooling water supply line 12 to a cold water tap in the laboratory and keep the cooling water lukewarm (approx. 35 C). This is not necessary with the Ga-In alloy (MP 17 C), not even with water at 10 ^° C.

1 30040/0289 ORIGINAL! NSPECTED

Claims (3)

Patent claims : Vacuum evaporation with electron beam evaporation using a copper turntable with several crucibles, characterized in that that the heat dissipation from the radiantly heated copper turntable to the water-cooled stationary copper base plate is made by a layer of liquid gallium. 2) vial electron beam vacuum evaporation system according to claim 1, characterized characterized in that the gallium layer is inclined conically upwards, so that gas bubbles in the gallium can soak upwards quickly. 3) Much crucible electron beam vacuum evaporation system according to claim 1-2, characterized in that instead of gallium, an alloy of gallium-13 % ^{indium which melts at 17 0} C is used. 130040/0289