US2217422A - Glass-metal seal - Google Patents

Description

Patented Oct. 8, 1940 UNITED STATES PATENT OFFICE 2,217,422 GLASS-BIETAL SEAL No Drawing. Original application July 5, 1929, Serial No. 376,291, now Patent No. 2,062,335,

dated December 1, 1936.

Divided and this ar plication April 24, 1936, Serial No. 76,199

4 Claims.

My invention relates to alloys and particularly to alloys of metals in the iron group of the periodic table.

This application is a division of my copending application Serial No. 376,291, filed July 5, 1929, which application issued on December 1, 1936, as Patent 2,062,335, and is assigned to the same assignee as this application.

One object of my invention is to provide an alloy having a predetermined coefficient of expansion over a temperature range from substantially room temperature to the annealing tem perature of certain glasses, particularly ordinary lead glass.

Another object of my invention is to provide an alloy having a smaller coefilcient of expansion than the nickel-steel alloys.

Another object of my invention is to provide an alloy adapted to form a vacuum-tight seal through lead glass.

Another object of my invention is to provide an improved method of sealing electrical con ductors through glass;

A further object of my invention is to provide a method for producing a vacuum-tight seal of electrical conductors, having only a limited range of expansivity substantially equal to that of ordinary glasses, through the latter without causing breakage of the seals in cooling to room temperature.

Other objects of my invention will become apparent on reading the following specification.

In connection with the manufacture of vacuum-tight electrical-discharge devices in glass containers, the problem arises of making vacuum-tight seals of metals through the glass walls. In forming such seals, or small hole is melted in the glass wall of the container; the metal lead is inserted and the molten glass squeezed around the lead and fused thoroughly in contact therewith. In order that the resulting seal shall not crack or leak after it has cooled, it is necessary that the metal and the glass shall have substantially the same expansivity with temperature over the range from the annealing temperature of glass down to room temperature;

and preferably that the metal be one which the glass will wet. Platinum metal has the proper expansivity when used in combination with ordinary lead glass, and the first successful seals through glass were made with platinum leads.

Platinum is an extremely expensive metal and, consequently, composite conductors having cores of nickel-steel covered by thin jackets of copper were developed, and thus cheaper, but equally satisfactory, seals through ordinary lead glass were obtained. The coefficient of expansion of copper is higher than that of lead glass; and, since the jacket of copper is relatively thin, the nickel-steel alloy employed is one which has a coefiicient of expansion nearly equal to that of lead glass. Copper is a metal which lead glass wets readily; and the expansivity of nickelsteel is slightly less than that of glass over the range of approximately 400 C. between the temperature at which strains readily relieve themselves in lead glass and room temperature.

I find that, by employing certain other nickelcobalt-iron alloys, and even certain nickel-iron alloys, seals may be made through ordinary lead glass; and this without the necessity for copperjacketing the alloy.

In accordance with the foregoing principles, I

have found that substitution of cobalt for nickel in nickel-iron alloys results in a decrease of expansivlty of approximately 0.5x 10- cm. per cm. per degree centigrade for each percent of cobalt thus substituted. Ordinary lead glass has an expansivity of approximately 9 10 cm. per cm. per degree C. while bore-silicate glasses have expansivities between 3.0 and 'l.0 l0-- cm. per cm. per degree C. In general, I have found that, in the case of alloys of low expansivity containing nickel and iron, whether or not they also contain cobalt, it is desirable that the relationship expressed by the. following formula should hold.

% nickel+2.5 manganese)+18 carbon) iron 0.50 to 0.60 the best value for the fraction being about 0.55. It also appears desirable, for most purposes, to minimize the amount of manganese, the maximum percentage 01' this element being preferably below 0.2%. Carbon, however, appears, in most instances, to be desirable up to the amount of 0.3%, as it permits a higher cobalt content than would otherwise be desirable and does not seriously affect expansivity.

As previously stated, I have found certain nickel-cobalt-iron alloys to seal successfully into ordinary lead glass having a coefllcient or expansion in the neighborhood of 9X 10- cm. per

cm. per degree C. For instance, alloys compris-.

ing 30% to 45% nickel, 5% to 25% cobalt and 0 to 1% manganese, form satisfactory seals with lead glass having an expansivity between 6x 10- and 10x10-' cm. per cm. per degree C.

I have found that the preparation 01' vacuumtight seals between metals and glass is rendered more certain if the metals are annealed in vacuo or in a bath of a fused salt which readily dissolves the oxides of the metal. For example, sodium nitrate is one such salt, many others being known to skilled chemists.

I have also found that a certain amount of trouble is caused in making vacuum-tight seals because of gases released at a high temperature, and largely results from a reaction between carbon in the metal and oxide formed on the surface of the metal during heating in air.

This source of gas may be eliminated either by removing carbon from the metal or by coating the metal with another metal which is impervious to oxygen and contains no carbon.

Small amounts of carbon cannot be avoided in iron-base alloys without elaborate precautions. Consequently, it is more convenient to remove the carbon from the alloythan to prevent its introduction therein. Carbon may be removed from the molten alloy by maintaining it in a vacuum in contact with refractory oxides or by passing hydrogen through the metal. This treatment is preferable when the finished piece is bulky. When, however; the metal is finished to a small cross-section, the carbon may be removed conveniently from the solid metal. The treatment required is heating at an appropriate temperature in hydrogen or in a fused salt bath containing no great amount of carbon compounds. A bath comprising 15% sodium chloride and 85% barium chloride at from 700 to 1200" centigrade is one suitable bath.

Coating to avoid gas generation at the surface may be readily accomplished by' electrolytic deposition of a thin layer of chromium or other carbon-free metal provided that the oxygencarbon relations within the alloy are not such as to cause severe gas release at working temperatures. In the event that gas is evolved without surface oxidation, at treatment, such as that just described, may first be applied to prevent injury to the coating by the gas.

for high-temperature service.

The temperature at which strains are readily 'relieved in glass is, of course, not an absolutely fixed quantity, even for glass of a given chemical This treatment is particularly applicable to plated metals intended At very high temperatures, the

fieotion temperature, that is-to say, the maximum temperature, at which coefficient of expansion of the above-described alloys remains low, decreases as the coeiiicient of expansion itself is decreased, it is desirable to form seals with as low an annealing temperature as possible. I have found that alloys, in which the inflection temperature is relatively low, may be employed to produce satisfactory seals if the completed seal is maintained at a temperature of about'50 C. above the inflection temperature for a considerable period of time. Thus, while the strains are relieved slowly, they do so progressively and with certainty, and a strong vacuum-tight seal results in the end.

As an example of the softer glasses, referred to above as lead glasses, the following may be taken as typical in composition:

63.1% SiOz; 20.2% PbO; 0.28% A1203; 0.94% CaO; 7.6% NaO; 5.5% K20; 0.88% M11304.

The above subject-matter is also described and referred to in my copending application Serial No. 376,292, filed July 5, 1929, for Alloys, now Patent 1,942,260, issued January 2, 1934.

While I have, in the foregoing, described certain particular embodiments of my invention, it will be understood that these are for purposes of illustration only, and that the broad principles may be otherwise utilized, as will be readily apparent to those skilled in the art.

I claim as my invention:

1. A vacuum-tight seal between an alloy consisting of iron, nickel, cobalt andmanganese and a glass containing substantially 63.1% 810:; 20.2% PbO; 0.28% A1203; 0.94% 09.0; 7.6% NaO; 5.5% K20 and 0.88%-Mn3O4.

2. A vacuum-tight seal between a glass containing substantially 63.1% S102; 20.2% PbO; 0.28% A1203; 0.94% 021.0; 7.6% NaO; 5.5% K20 and 0.88% MnaOi and'an alloy comprising 30 to 45% nickel, 5 to 25% cobalt, from a trace to 1% manganese and the remainder iron.

3. A vacuum-tight seal between a vitreous material having an expansivlty of 6.5 to x10- cm. per cm. per C., and an alloy comprising 43 to 45% nickel, 10 to cobalt, from a trace to less than 1% manganese and the remainder iron.

4. A vacuum-tight seal between a vitreous material containing substantially 63.1% S102;

20.2% PbO; 0.28% A1203; 0.94% (39.0; 7.6% NaO; 5.5% K and 0.88% M11304 and an alloy comprising 43 to 45% nickel, 10 to 15% cobalt, from a trace. to less than 1% manganese and the remainder iron.

HOWARD SCO'I'I'.