US2188693A - Manufacture of silicon carbide resistors - Google Patents

Description

Jan. 30, 1940. A. J... THOMPSON MANUFACTURE 3F SILICON CARBIDE RESISTORS Filed July 26, 1938 TO no 'on-ms IWSAT I00 c.

03 51 .5 .32 zu wzro mmw Jooo INVENIYOR. ALMER- :r. THOMPSON.

ATTORNEY.

Paten tedJan. 3b, 7 a

PATENT OFFICE MANUFACTURE OF SILICbN CARBIDE BEBISTOBB Almer- J. Thompson, Niagara Falls, N. Y., assiznor \to The Carborundum Compa y, Niagara Falls,

N. E, a corporation of Delaware Application July 28, 1938, Serial No. 221,318

8Claims.

This invention relates to the manufacture of silicon carbide resistors, and more particularly to a method of manufacture adapted for continuous production of recrystallized resistors" having desirable electrical properties. The present application is a continuation-in-part of my co-pending application, SerlalNo. 677,294, filed June 23, 1933.

Recrystallized silicon carbide articles are ordi- 10 narily manufactured by forming a mass of silicon carbide grains and a temporary binder to shape, embedding the formed mass in a protective mixture of sand and carbon, and heating the mixture and the articles embedded there- 1 in to a very high temperature in order to produce recrystallization of the silicon carbide. The heating time of such a process is necessarily very prolonged because of the extremely low thermal conductivity of the sand-carbon mixture, and in the usual commercial furnaces the time of heating is from twenty-four to thirtysix hours. An embedding mixture of sand and carbon has heretofore been considered necessary in order to generate the proper silicon vapor atmosphere to produce recrystallization.

The electrical properties of recrystallized silicon carbide as made by the above procedure make it entirely unsuited for use as an electrical heating element. The variation in reao sistance with changes in temperature is extremely great and the resistance at room temperature may be more than one hundred times that of the material when it is heated to temperatures from 1000 C. to 1500 C. The electrical resistance of the silicon carbide is also very diificult to duplicate and even in the same furnace run, resistors may be obtained which vary in resistance by as much as several hundred times.

It is known that silicon carbide resistors having very satisfactory electrical properties can be made by curing the resistors at a high temperature if the heating current is forced directly through the resistor itself during curing. Re-

slstors made by such a process ordinarily have a positive temperature coefllclent of electrical resistance. The process, however, is somewhat difficult to carry out as the unburned silicon carbide mix, from which the resistors are formed,

is practically an electrical insulator and a very high voltage is required to force a current through a resistor of even relatively short length. The number of resistors which can be burned in a single operation is thus limited. The process can not be operated continuously for with every batch of resistors, it is necessary to embed them in the burning bed or trough, allow them to cool after the resistors are raised to the required temperature, and replace a fresh batch of resistors in the burning bed for a second op- 5 eration.

In many instances a positive temperature coeflicient of resistance for a silicon carbide resistor is not absolutely essential. The present invention is concerned with a method of pro- 10 ducing resistors in large quantities in which the resistors have electrical characteristics entirely different from the usual recrystallized silicon carbide and in which the properties approach those obtained by passing the heating current 15 through the resistor during the curing process.

We have found that if the resistors, after being formed from a mix of silicon carbide grains and a temporary binder, are subjected to heat from a highly heated chamber furnace, 20 instead of being embedded in a sand-carbon mixture as has heretofore been considered neeessary to eflect satisfactory recrystallization, recrystallization can be efiected almost instantaneously and the electrical properties of the ma- 25 terial are improved to a point where the resistors can be used satisfactorily for many electrical heating purposes. Preferably, a temperature of at least 2200 C, should be maintained in the furnace and the temperature may so be as high as 26il0 C. I have found that the electrical properties of the resistor are to a large extent a function of the temperature used for recrystallization, and if the curing temperature is maintained at a value fairly close to the decomposition temperature of silicon carbide, (2600 C.) a resistor having properties which are satisfactory for most purposes can be produced without applying eithera sand-carbon embedding mixture or passing the current an through the resistor during cg.

The time required to produce a satisfactory resistor will, of course, depend upon the term perature of the furnace but is usually not more than a very few minutes for an element of small 45 diameter, as, for example, one-half inch or less.

At temperatures of from 2500" C..to 2600 C. recrystallization of an element of this diameter can be eifected in from two to five minutes. It is desirable that the temperature of the fur- 50 nace be regulated so that recrystallization takes place in a period of time which is less than thirty minutes. I have fouhd that with lower temperatures where the time of recrystallization is unduly prolonged, as, for example, at tem- 55 peratures from 2000 C. to 2200" C., the properties of the resistor tend to revert to those which are characteristic of the usual recrystallized silicon carb de.

The temperatures required. for recrystallized silicon carbide are ordinarily very diflicult to obtain in a furnace of the chamber type as these temperatures are far above those ordinarily encountered in electrical resistance furnaces. I

have found that if a carbonaceous container isused for the furnace chamber and the current electrically induced into the container or the wall of the furnace, preferably by high frequency induction, the temperatures required for recrystallization can be easily and practicably obtained. One of the features of my process is that the furnace can be constructed so that the resistors can be passed through the heating chamber continuously. For this purpose a tunnel kiln type of furnace in which the heating zone is heated by high frequency induction can be employed,

In the drawing, which is intended to merely assist in illustrating the invention and not to limit the same,

Figure l is a longitudinal sectional view, partly diagrammatic, showing a continuous type of furnace which may be employed in recrystallizing resistors in accordance with my improved process, and

Figure 2 is a diagram showing the resistance temperature characteristics of several silicon carbide resistors, contrasting the characteristics of a resistor recrystallized in accordance with the 01d procedure with the characteristics of a resistor recrystallized in accordance with my improved" procedure.

Considering Figure 2 in detail, curve A repre sents the change in resistance with temperature of a silicon carbide resistor recrystallized by embedding the unburned material in a sandcarbon mix and passing the current through a conducting core also buried within the mix, this being the former procedure referred to above. Curves B, C and D show the changes in resistance with temperature of resistors recrystallized in accordance with my process at temperatures of 2300 C., 2400 C. and 2500 C., respectively. It will be observed that when the resistors are recrystallized in a high temperature chamber, furnace, kiln, or the like, the high electrical resistance of the recrystallized material at room temperature is remarkably decreased over that which obtains when the silicon carbide is recrystallized in accordance with the old procedure. As the temperature of the recrystallization process is increased to the point where it becomes very close to the decomposition temperature of silicon carbide, the resistance at room temperature does not greatly exceed the resistance of the element when heated to the operating temperature. At high temperatures the resistance temperature coefilcient tends to become substantially zero.

A resistor having these properties can be operated very satisfactorily from a source of current in which the voltage is maintained at a constant value. With resistors having the properties shown in curve vA, which represents-the resistance-temperature characteristic of ordinary recrystallized silicon carbide, the resistance at room temperature is so great that a very high voltage is required to force an initial heating current through the resistor; and, this voltage must be continually regulated throughout the heating process, as otherwise the sudden increase in current with rising temperature will completely disintegrate the resistor.

I have shown in Figure 1 a type of furnace in which the rods may be subjected directly to elevated temperatures. The furnace shown is of the induction type and may comprise a carbonaceous container such as the tube' 5, this container preferably being made from artificial graphitefsuch as that known in the trade as Acheson graphite), which, as iswell known in the art, oxidizes slowly producing carbon monoxide. The container 5 is considerably longer than its diameter and therefore a reducing atmosphere is maintained inside the furnace at all times. This carbonaceous container, as shown in Figure 1, is embedded in suitable thermal insulating material 6 such as powdered carbon (preferably lamp black). The furnace is heated by electrical induction from the coils I which surround the furnace, being mounted on an electrically insulating refractory tube 9. I prefer to form the carbonaceous container in three sections, as shown in Figure 1, the central section 5 being inserted between end sections Ill. The sections are slightly spaced so as to assist in preventing the migration of heat to the ends of the furnace, and all three sections are embedded in the suitable thermal insulating material 6. The coils 1 surround the central section 5 so as to concentrate the heat of the furnace within the section.

The furnace is first heated to the temperature necessary to impart the desired electrical properties to the resistors, preferably a temperature of 2200 C. or above. The rods are introduced into the hot zone of the furnace and maintained there for the short period of time necessary to effect recrystallization.

If desired, the resistors may be mounted in suitable carriers for guidingthe resistors through the carbonaceous container. For instance, each end of the resistor may be centered in a carrier which slidably fits the interior of the tubes 5 and i 0. If a plurality of resistors are to be passed through the furnace simultaneously, the resistors should preferably be distributed symmetrically about the axis of the carriers. As above indicated, the rate of progress of the resistors through the furnace should be controlled so that the proper length of time required for recrystallization is allowed. This period will generally be less than fifteen minutes and periods in excess of thirty minutes should be avoided, as the silicon carbide tends to revert to the usual recrystallized form having undesirable electrical properties, if the heat treatment is substantially prolonged.

In carrying out my process, a suitable mix which will produce a satisfactory recrystallized silicon carbide resistor by the methods heretofore used can be employed. I may, for example, use chemically purified grain, or I may use the dense or compact variety of silicon carbide which is sometimes used for producing a resistor having a positive temperature coefficient of resistance when the heating current is forced directly through the resistor itself during curing. Itis desirable to add a small quantity of carbon, as' .for example, from two to three percent, to the mix aheaecs Other temporary binders can be used, but soself consists of a graphite tube of considerable 7 length compared to its diameter. Because of this, a reducing atmosphere is maintained inside the furnace at all times. It is, however, sometimes desirable before introducing the formed resistors into the furnace, to give them a protective coating so as to make sure that no oxidation of the silicon carbide will occur at the elevated temperatures prevailing in the furnace. This coating may be either finely divided carbonaceous material such as coke, or may be a mixture of finely divided sand and carbon containing from 50% to 75% silicon. The protec tive coating is preferably applied in the form of a wet slurry having such a consistency that the material will adhere to the rod to form a coating of about thick when the rod is dipped into the slurry and then removed. A

coating of this thickness offers very little resistance to the direct transfer of heat to the body of the resistor, and the rod can be heated quickly to the recrystallization temperature by direct radiation from the highly heated furnace walls. When the coating is applied as a slurry and subsequently dried, it is very compact, and the thermal conductivity is substantially the same as that of the rod itself, whereas when the resistors are embedded in a loose protective mix, the thermal conductivity of the mix is very low and a relatively large quantity of mix is required to secure adequate protection from oxidation, so that a prolonged heating time is necessary.

I have also used an atmosphere in the furnace to prevent oxidation such as carbon monox ide. If carbon monoxide is used, the furnace temperature must be raised sumciently to take care oi the heat losses resulting from the introduction of cool gas into the furnace. I have also found that an inert gas such as argon, helium or the like may be introduced into the carbonaceous container to produce non-oidng conditions.

One of the chief advantages of the method of heat treating herein described is that it permits the rapid curing of a large number of resistors by a batch heating processor by the continuous passage of the resistors through a furnace designed forcontinuous operation. In the method of burning silicon carbide resistors by passing the current through them, the number of resistors which can be treated in a single operation is limited, and the voltage required in passing the current through the resistors is approximately 500 volts or even greater per linear foot of burning bed, so that only a relatively small number of resistors can be burned with equipment of any reasonable voltage. With my process recrystallization can be carried on continuously and the number of resistors which canbe produced in a single operation is practically unlimited.

In the above description I have employed the term recrystallization. It has been known for some time that when molded shapes of silicon carbide are heated to a sufliciently high temperature, the crystals grow together (apparently by evaporation and redeposition) to form a ;o-

.herent mass which retains its strength even at very high temperatures. in such a process, the crystals are self-bonded and no fusible bonding material is used with the exception of a temporary agglutinant. In employing the term recrystallization in the specification and claims I do so in the sense thus well known in the art.

While I have shown and described certain embodiments which I prefer to use, it will be understood that my invention is not so limited. but may be otherwise embodied and practiced within the scope of the iollog claims:

1. The method of recrystallizing a silicon carbide resistor so as to produce a resistor having temporary binder, applying to the molded article a relatively thin coating or protective material containing a carbonaceous ingredient to prevent oxidation, and heat treating the coated article within a chamber of a Min, furnace. or the like to a temperature oi at least 2200 C.

to produce self-bonding of the grains.

3. The steps in the process of making a silicon carbide resistor which comprise molding the resister from a mix oi silicon carbide and a temporary binder, applying to the molded article a relatively t coating of protectivematerial to prevent oxidation during heat treatment, and recrystallizing the coated article within a chamber of a kiln, furnace, or the like at 2200 C. or above.

a. in the method oi heat treating silicon car= bide resistors to produce self-bonding of the rains, the step of subjecting the resistors to externally applied heat within the chamber of a kiln, furnace, or the like operating at a temperature between 2200 C. and 26il0 .C., while protecting the resistors from omdation by means of an inert gas.

5. in the method of heat treating silicon carbide resistors to produce self-bonding ofthe grains, the step or subjecting the resistors to externally applied heating within a chamber of a kiln, furnace, or the like for not more than thirty minutes at a temperature between 2200 C. and 2600 C. while protecting the resistors from oxidation by means oi a carbonaceous atmosphere,

6. The method oi making a silicon carbide resistor so as to produce a resistor having a relatively low resistance/at room temperature, which comprises forming the resistor element 'from a mix consisting principally of silicon cardill resistance at room temperature, which com-1 prises forming a heating element from a mix consisting principally oi silicon carbide grains and a temporary binder, and. recrystallizing said grains by subjecting the formed element to externally applied radiant heat at a temperature of atleast 2200 C. within the highly heated walls of a kiln, furnace or the like for a. period not exceeding thirty minutes. 4 8. The method of making silicon carbide resistors having relatively low resistances at room temperature which comprises forming the resisters from a mix consisting principally of silicon carbide, maintaining a heating chamber at a temperatures! 2200 C. or above, and subjecting the formed resistors to the radiant heat from said heating chamber for a period not exceeding 30 minutes by passing them through said heating chamber while maintaining a carbonaceous atmosphere surrounding the resistors Y I whereby the silicon carbide grains composing the resistor are self-bonded."

ALMER THOMPON.

Images