US2874448A

US2874448A - Method for stabilizing semi-conductor rectifiers

Info

Publication number: US2874448A
Application number: US336892A
Authority: US
Inventors: William F Haldeman
Original assignee: Individual
Current assignee: Individual
Priority date: 1953-02-13
Filing date: 1953-02-13
Publication date: 1959-02-24
Anticipated expiration: 1976-02-24

Description

Feb. 24; 1959 w. F. HALDE'MAN METHOD FOR STABILIZING SEMI-CONDUCTOR RECTIFIERS Filed Feb. 15. 1953 FIG 2 FIG ROOM TEMP .OOI AMPS FORWARD CURRENT 4 @320 n moadh mwm OmdBmOu CYCLEQ F ORWARD G URRENT, I- MILLIAMPERS HOT- COLD FIG 3 F ORWARD CURRENT, 1""ILLIAMPERS INVENTOR. WILLIAM F. HALDEMAN nited States Pa 'fifo r 2,874,448 NIETHOD FOR STABILIZING SEMI-CONDUCTOR RECTIFIERS William'F. I Ialdeman, Indianapolis, Ind. I Application February 13, 1953, Serial No. 536,892 9 Claims. c1. 29-253 (Granted under Title 35, U.;s. Code (1952 sec. 266) The invention described herein may be manufactured and used by or for the Government of the United States of America for governmental purposes without the payment of any royalties thereon or therefor.

.My invention relates to semiconductor-type rectifiers and is,particularly directed to methods and means for stabilizing the resistance-current characteristics of said rectifiers. p

It has long been known that two dissimilar metals placed in contact have unidirectional current carrying characteristics and. in recent years the metal germanium has beenextensively used as one of these metals, the

second metal being a cat whisker made of steel ma nonferrous metal spring-pressed into contact with the germanium. In an effort to accentuate the unidirectionality of such a rectifier and to standardize the current-resistance characteristics, the germanium is melted with a fraction ofa percentage of an impurity and' permitted to cool slowly. In spite of care in preparing the germanium ingot there is found to be radical variations in electrical characteristics among different ingots and among difierent parts of each ingot. Worse, the germanium changes its characteristics with aging, thus upsetting circuitry in which it hasbeen incorporated and calibrated. I I have found that commercial semiconductor-type rectifiers have resistance-current characteristics that may be defined by the expression where I is current of the order of a few milliam'peres, Ris an effective resistance in ohms, hereinafter defined, anda and b are constants. The exponent a determines the'curvature of-the exponentialv resistance curve of any one semiconductor rectifier at one temperature. It is variations of these constants, a and b, after manufacture that is particularly troublesome. In my co-pending application: entitled Testing Semi-Conductor Rectifiers, Serial.l;\l9.-336 ,893, filed February 13,1953, issued on April24, 1956, as'U. Sglatent No. 2,743,420, I describe in detail how rectifiers may be tested and grouped ac- Other objects of my invention and its true scope will appear as the following description of one embodiment 2,874,448 Patented *Feb. 24, 195? ice 2 of my invention proceeds. In the accompanying drawmgs:

Fig. 1 shows the effective forward resistance characteristic of typical semiconductor rectifiers,

Fig. 2 is a graph showing how the effective resistance of a rectifier varies with current for different values of the constant b, and

Fig. 3is a graph showing how the effective resistance of a rectifier varies with current for different values of the exponent a.

I have found that the resistance of a semiconductor rectifier varies with the forward current through the rectifier. The typical effective resistance-current characteristic of commercial rectifiers are represented by the curves of Figs. 2 and3. Unfortunately, the current does not vary linearly with resistance but varies exponentially and suggests that the current-to-eifective resistance rela tionship may be defined by the expression where I and R are, respectively, current and effective resistance and where a and b are constants. If the value of the exponent a varies in use the curvature of the characteristic curve will change as suggested in Fig. 3, in which case circuitry containing the rectifier becomes unbalanced and decalibrated. It is particularly troublesome having dissimilar a values cross as in Fig. 3.

Effective resistance is defined to be that linear or ohmic resistance which will pass the same average current during a specified voltage pulse as will pass a non-linear rectifier when subjected to the same voltage pulse. It is like wise desirable that the constant b be stable during life so that the static parameters of the rectifier will not change and upset the circuitry in which it is incorporated.

It has been found that the conductivity of germanium may be changed when the crystals are bombarded by highenergy particles from a cyclotron. Bombarding the crystals appears to produce lattice dislocations and perhaps otherwise affect the electrical properties. I assumed, then, that the crystal structure and the position of the impurity atom might be changed with thermal-agitation, and I found this to be true when the crystals were heated. Although the electrical changes were, small when the heatinduced energy of vibration was low, large changes occurredover long periods of time. The energy of vibration of an atom is increased as the temperature of the crystal is increased. It then became clear that exposing these germanium crystals to temperature fluctuations might accelerate the process of impurity migration to more stable locations. i

On the assumption that alternately heating and cooling a germanium crystal containing suitable impurity molecules would tend toshift the molecules to more stable positions, I measured the effective forward resist; ance of a number of commercially available germanium rectifiers; Germanium rectifiers having effective forward resistances of the values indicated in Fig. 1 by curves 10 and 11, had 'an effective forward resistance of about 330 ohms when received from the manufacturer. As these particular rectifiers alternately were chilled to 69" C." and heated to 0., forwardresistance continued to increase to 450 and 415' ohms, respectively, after four cycles of freezing and heating. Rectifiers with such widely variable forward resistances as these are obviously an: suited'for commercial uses where the operating tempera; tures may be extreme. Other rectifiers such as those hav: ing the forward resistances indicated by

curves

12,13 and I4, on the other hand, decrease from an initial 440 "ohms to values below 400 ,ohms' as the crystal was subjected alternately to extreme high and low temperatures. Surprisingly, as the cycling of the crystal through extreme temperatures continued it was found that the forward resistance changed less and less, percentage-wise. It

seems reasonable to assume from these results that alternate heating and cooling of the semiconductor-type rectifier will stabilize its effective forward resistance-current characteristics.

In determining the range of temperatures to which the crystal must be subjected for adequate stabilization, expermentation indicated that the range may be uite narrow or quite wide. It was found that by heating and cooling to temperatures, respectively, above and below the range of temperatures to which the rectifiers would be subjected in use, the stabilization process could be accelerated. For some military purposes these ranges may vary from the temperatures of the tropics to the sub-zero temperatures of the polar regions. Norrower temperature operating ranges are of course usual. The maximum temperature to which a manufactured semiconductor rectifier can be subjected appears to be limited only by the temperature to which the glass-to-metal seals and other mechanical parts of the rectifier may be carried without damage. The minimum temperature to which the crystals may be frozen likewise is determined by the mechanical characteristics ofthe rectifiers. It was found, for example, that with some commercial rectifiers the difference in coefiicient of expansion of the various parts may actually cause the contact electrode such as a cat whisker to separate from or shift on the crystal, or cause a fracture at the seals.

The number of heating-cooling cycles necessary to stabilize the forward resistance characteristic of the rectifier is likewise individual to each rectifier. The number of cycles in turn is dependent upon the magnitude of the temperature ranges employed. In the case of the rectifiers graphed in Fig. 1, about 18 cycles, using temperature extremes of 60 C. and +100 C., was considered satisfactory for sufficiently stabilizing the rectifiers for use in critically balanced circuits. No appreciable drift in the forward resistance of rectifiers thus processed has been noted after several months of use. The number of cycles of course may vary widely, the greater the number the greater being the invariability of the forward resistance.

Good results have been obtained when the crystals are heated to +100 C. and heldat that temperature for 60 minutes and then refrigerated to 60 C. for about 60 minutes each, the crystals being held at room temperature for about minutes between each heating and cooling excursion. As suggested above, when commercially available germanium rectifiers are thus treated for at least four cycles, the constants a and b in the resistancecurrent expression become su1ficiently stable for many circuit applications. The constants a and b fluctuate less and less as the number of heating and cooling cycles increase.

Many modifications of my novel heating-cooling cycl ing may be made without departing from the scope of my invention. The time-temperature constant of each heat treatment, and the time-temperature constant of each cooling treatment should be sufficiently long, and the number of heating and cooling cycles should be sufficiently large that they render subs antially invariable the constants a and b in the relationship where I and R are the forward current and etfective resistance, respectively, of the rectifier in the operating temperature range of the rectifier, and a and b are said constants.

Although the temperature cycling described and hereinafter claimed very probably produces the etfects suggested relative to impurity location stabilization, it is to be understood that the utility of the method and the beneficial results which ensue may, in fact, be found in the future to be the result of other changes; for example, changes at therectifying junction of the semiconductor, not clearly understood at the present state of the art.

I claim:

l. The method of treating a semiconductor-type rectifier having dissimilar metals in contact, said method comprising heating said rectifier to a temperature above the normal operating temperature of the rectifier, and then cooling said rectifier to a temperature below the normal operating temperature of the rectifier.

2. The method of treating a rectifier defined in claim 1 further comprising the step of holding the rectifier at sub stantially room temperature after the said heating excursion until the temperature of the rectifier substantially returns to room temperature.

3. The method of treating an assembled semiconduc tor-type rectifier to stabilize the effective forward resistance-current charactertistics of the rectifier comprising alternately heating and cooling said rectifier to temperatures outside the normal operating temperature range of said rectifier.

4. The method of accelerating the aging and resultant stabilization of the forward resistance-current characteristics of semiconductor rectifiers, comprising heating said rectifiers to approximately degrees centigrade above zero and then cooling said rectifiers to approximately 60 degrees below zero, and repeating the mentioned heating and cooling steps a plurality of times.

5. In the method defined in claim 4, holding the rectifiers at each of the high and the low temperatures for approximately one hour.

6. In the method defined in claim 4, holding the rectifiers at room temperature for approximately one quarter hour between the heating and cooling excursions.

7. The method of stabilizing the resistance-current characteristic ofan assembled semiconductor-type rectifier said characteristic being defined by the relationship, I- =bR, where R is resistance, I is current and a and b are constants, said process comprising alternately heating and cooling said rectifier to uncritical temperature extremes of about +l00 C. and about --60 C. maintaining the said rectifier for about one hour at the temperature extreme of each of the said heating and cooling excursions, and continuing the heating and cooling excursions until said constant a reaches the desired degree of stabilization within a predetermined range of operating temperatures between the temperature extremes of the aforesaid heating and cooling excursions;

8. The method of stabilizing the resistance-current characteristic of an assembled semiconductor-type rectifier, said characteristic being defined by the relationship, l- =bR, where R is resistance, I is current and a and b are constants, said process comprising alternately heating and cooling said rectifier between about +l00 .C. and about -60 C. for about one hour for'each heating and cooling excursion, and continuing the heatingand cooling excursions for at least four cycles.

9. The method of stabilizing the resistance-current characteristic of an assembled semiconductor-type rectifier, said characteristic being defined by the relationship, I" =bR, where R is resistance, I is current and a and b are constants, said process comprising heating said rectifier to about +100 C., then cooling'said rectifier to about -60 C., and repeating the heating and cooling until said constants reach a desired degree of stabilization for a given temperature.

References Cited in the file of this patent UNITED STATES PATENTS 2,395,259 Ellis et al. Feb. 19, .1946 2,464,066 Addink et al. Mar. 8, 1949 2,602,763 Scaff et al. July 8, 1952