US 3886097 A
A method of making low avalanche metal oxide varistors is described. Metal oxide powders are isostatically pressed to form a rod. The rod is fired, and then sliced into a plurality of discs. Electrodes are applied to the major faces of the discs, to produce discrete varistors. Thinner discs can be made by this technique than can be made by uniaxially pressing discrete discs, to get lower avalanche voltages. Also, comparative data is presented showing that varistors from slices of this isostatically pressed rod inherently have lower avalanche voltages than uniaxially pressed discrete disc varistors.
Claims available in
Description (OCR text may contain errors)
United States Patent 1191 Hossenlopp 1 51 May 27, 1975 METHOD FOR MAKING A LOW AVALANCHE VOLTAGE METAL OXIDE VARISTOR  Inventor: Arthur M. Hossenlopp, Kokomo,
 'Assignee: General Motors Corporation,
221 Filed: Nov. 12, 1973 21 Appl. No.: 414,814
 US. Cl. 252/519; 29/621; 252/518  Int. Cl. ..H01c 7/10; HOlc 17/00  Field of Search 252/518, 519; 29/610, 621
3,805,114 4/1974 Matsuoka et al. 252/518 X Primary Examiner-T. l-l. Tubbesing Attorney, Agent, or FirmRobert J. Wallace [5 7 ABSTRACT A method of making low avalanche metal oxide varistors is described. Metal oxide powders are isostatically pressed to form a rod. The rod is fired, and then sliced into a plurality of discs. Electrodes are applied to the major faces of the discs, to produce discrete varistors. Thinner discs can be made by this technique than can be made by uniaxially pressing discrete discs, to get lower avalanche voltages. Also, comparative data is presented showing that varistors from slices of this isostatically pressed rod inherently have lower avalanche voltages than uniaxially pressed discrete disc varistors.
5 Claims, No Drawings METHOD FOR MAKING A LOW AVALANCHE VOLTAGE METAL OXIDE VARISTOR BACKGROUND OF THE INVENTION This invention relates to metal oxide varistors and more particularly to an improved method that is particularly useful in making metal oxide varistors that avalanche at lower voltages.
Varistors are resistors having markedly nonlinear voltage-ampere characteristics. Resistivity varies with current and voltage. At some voltage, for any given current level, the varistor becomes highly conductive. This voltage is referred to herein as the avalanche voltage. In addition to current, avalanche voltage is also known to be affected by metal oxide composition and thickness of the varistor. The oxide firing process, of course, can affect it too. However, in general there is a distinct increase in avalanche voltage per unit increase in varistor thickness for each distinct oxide composition used.
It is not especially difficult to make metal oxide varis ,tors which avalanche above 100 volts. Such devices are conventionally made by uniaxially cold pressing a selected mixture of metal oxide powders into a relatively thick disc, firing the disc, grinding or lapping the disc to obtain the desired degree of flatness and parallelism on the opposite major surfaces of the disc, and then applying electrodes to these ground or lapped major surfaces. Each disc is thus separately formed and ground to a predetermined thickness, and then electrodes applied.
On the other hand, considerable difficulty lies in making varistors that avalanche at less than 50 volts. Even with improved metal oxide mixtures, the discs must be made extremely thin. For example, disc thickness must be reduced to thicknesses below 0.05 inch and even below 0.03 inch. In some instances the discs simply cannot be made thin enough to obtain the desired low avalanche voltage characteristic. It is extremely difficult to obtain uniform distribution of only a small quantity of powdered materials over a relatively large area in a pressing die. This produces an uneven density in the resultant thin disc that is formed, which in turn produces uneven avalanche characteristics across the face of the finished product. Analogously, poor surface flatness and parallelism can result in the disc after firing. In addition, the thinner the disc is, the more fragile it is. The fragile discs must be handled both before and after firing, resulting in considerable waste due to breakage.
As a result, effort has been directed toward developing metal oxide compositions which inherently avalanche at lower voltages per unit thickness. With such compositions the varistor discs can be made in larger thicknesses, which reduce the fabrication problems previously referred to. However, these improved compositions are still not good enough to eliminate them. Such compositions are described in United States Pat. No. 3,663,348 Masuyama et al., as well as in the Journal of the ElectrochemicalSociety, Vol. 114, No. 8, pages 833-842, in an article entitled Polycrystalline Zinc Oxide Dielectrics, by Delaney et al.
On the other hand, I have found a new method of using these compositions, which avoids the fabrication problems referred to, and permits one to readily and consistently obtain varistors with low avalanche voltages.
OBJECTS AND SUMMARY OF THE INVENTION It is, therefore, a principal object of this invention to provide an improved method of making metal oxide varistors, which is particularly useful in making metal oxide varistors that avalanche at low voltages.
This object and other objects, features and advantages of this invention are obtained by isostatically pressing a selected mixture of metal oxide powders into an elongated body, firing the elongated body to form a coherent mass having a predetermined nonlinear voltagc-ampere characteristic, transversely sawing the elongated body into a plurality of thin slices having substantially flat and parallel opposite major faces as formed, and attaching electrodes to the opposite major faces of the slices.
DESCRIPTION OF THE PREFERRED EMBODIMENTS In this invention, the variator discs are made by transversely slicing a tired bar, not by uniaxial cold pressing. The slices can readily be made much thinner than one can make discs by cold pressing. Moreover, as formed, the slices are flat and parallel. Satisfactory devices can be made without extensive grinding or lapping, and in most instances it is unnecessary. Discs considerably thinner than 0.050 inch can readily be made. Hence, varistors having low avalanche voltages can readily be made. in general, this technique is most advantageous in making varistor discs below 0.05 inch in thickness, and more particularly below about 0.03 inch. Varistor discs of about 0.01 4 0.02 inch in thickness can readily be made using equipment such as used in sawing slices from monocrystalline semiconductor ingots. Even the thinnest discs are substantially flat and parallel as formed.
However, in addition to these and other benefits I have unexpectedly found still. another benefit. Testing has shown that relatively thick varistors produced by my method inherently exhibit a lower avalanche voltage than equally thick varistors produced by uniaxial pressing. As will subsequently be shown by comparative data, there is a significant difference. Presumably, this advantage is present in my thin varistors too. Thus, a double advantage is apparently obtained when making thin varistors by my method. Firstly, one can make thinner varistor discs, to inherently obtain lower avalanche voltages. Secondly, the resultant thin discs apparently exhibit a lesser increase in avalanche voltage per unit increase in thickness than uniaxially pressed discs.
The oxide mixture for my method is prepared for pressing in the same way an oxide mixture is prepared for pressing in the conventional uniaxial cold pressing technique. Any of the known and accepted oxide pow der compositions and preparation techniques can be used. However, my method is directed to the manufacture of low avalanche voltage varistors. I prefer to use those metal oxide compositions which are particularly formulated to produce lower avalanche voltages. Zinc oxide mixtures containing a small amount of bismuth oxide are known to produce this effect, and are generally preferred. It is all preferred that the mixtures have a particle size which is less than about 20 microns. One specific oxide composition which can be used includes,
in mole percent, ZnO 98 percent, Bi O 0.5 percent, C 0 0.5 percent, MnO 0.5 percent, and Cr O 0.5 percent.
ln an example of my process, 100.72 grams of the above specific oxide composition was placed with 300 cc. of methanol in a container, and ball milled for three hours. Ball milling, rather than simple mixing, insures that the oxide particle size will be below about 20 microns and that the mixture will be uniform. After milling, the batch was emptied onto a sheet of aluminum foil and air dried. It was then passed through a 20 mesh screen, which separated the mixture from the balls used during ball milling. About grams of an aqueous solution containing 2 percent by weight polyvinyl alcohol inch and fired identically as described above for the isostatically pressed rod. After firing, the discs were ground and lapped in the normal and accepted manner, which reduced them to a thickness of about 0. 145 inch. Electrodes were then applied to their major faces in exactly the same way electrodes were applied to the aforementioned slices from the isostatically pressed rod.
The following table shows the results obtained on testing the varistors thus made, with varistors 1 4 being the uniaxially pressed discs, and varistors 5 7 being slices from the isostatically pressed rod. The avalanche voltage is shown in volts to the right of each of the seven varistors for each of the current levels noted.
Varis- Microamperes Milliamperes tor .1 .2 .5 l 2 3 5 1O 40 was then added to the mixture and uniformly blended with it. The resultant powder blend was then passed through a 20 mesh screen and dried overnight at 65 C.
The powder blend was then placed in an elongated rubber isostatic pressing mold, and the mold sealed. The interior of the mold was generally cylindrical, having a diameter of about /2 inch and a length of about 3 inches. The mold was lowered into a pressing chamber of an isostatic pressing apparatus wherein the mold was surrounded by a pressing fluid. The pressing chamber was then sealed, and the fluid pressurized to about 19,400 pounds per square inch, with a 5 minute dwell at that pressure. The pressure was then released, the pressing chamber opened, and the rubber mold removed. A compacted rod of the oxide blend was produced.
The rod was removed from the mold and heated in an oven at 220 C. for about 40 hours. It was then placed in a cool furnace and furnace heating started. When the furnace reached a soak temperature of 2,300 F., it was held there for about three hours. Furnace heating was then discontinued and the furnace allowed to cool overnight, about 15 hours, to a temperature of about 300 F. 400 F. During the first hour the furnace cooled about 300 F. 500 F. The rod was then removed from the furnace and allowed to cool to room temperature. It was then transversely sliced into a plurality of discs about 0.145 inch in thickness, with the major surfaces of the slices being substantially flat and parallel as sliced.
The slices were then ready for application of electrodes to their major surfaces without any further operations such as grinding or lapping. The electrodes were applied by hand brushing a commercially available silver-palladium-glass cermet mixture onto their major surfaces, and then fusing the mixture to bond it to the slice.
For comparison, an identical powder blend was prepared as described above and used to make 0.58 inch diameter circular discs by uniaxial cold pressing. The discs were individually pressed to a thickness of 0.156
All seven of the above varistors were nominally of the same size. However, as can be seen from the above table, the uniaxially pressed varistors consistently avalanched at higher voltages than varistors made from slices of the isostatically pressed rod. There is a difference at a current level of one milliampere of about one volt per 0.001 inch in varistor thickness. Alpha, the change in voltage with current, was generally higher in varistors made from the isostatically pressed bar.
As mentioned, in my method the oxide powder is prepared for pressing in the known and accepted manner. For example, if the particle size of the powdered mixture is not of a small enough size, it is initially ground to a smaller particle size. A temporary binder and lubricant is normally included in the mixture. In the examples described above, polyvinyl alcohol was used both as a temporary binder and lubricant. About 2 5 percent by weight of the 2 percent aqueous solution can be used. However, other substances can be used for such a combined function, or separate substances can be used for each.
Similarly, after blending with additives, the powdered mixture is screened and dried to form agglomerates of appropriate size for pressing. The size of the agglomerates used, dryness, and manner of formation are all variables, as they are in uniaxial cold pressing.
My isostatically pressed rod can be fired in the known and accepted way for uniaxially pressed discs. By this 1 mean it should be preheated, raised to a high temperature, and slowly cooled. However, the precise preheating and slow cooling used in my method is more critical than it is in firing uniaxially pressed discs. For example, the rod can be initially slowly heated, or heated to some constant lower temperature for a period of time, to drive off volatiles such as water, the temporary binder, and lubricant. Then it can be heated to the maximum firing, or soak, temperature. Analogously, it is not necessary to place the pressed rod in a cool furnace and then start heating to the desired soak temperature. However, such practice will insure that any remaining volatiles can escape without adversely affecting the pressed rod. After soaking at the desired temperature the pressed part is then slowly cooled. Furnace cooling is a typical technique by which this can be done.
As is known, the specific soak temperature and time for soaking used during the firing will affect the electrical properties and their resultant product. I prefer to use a soak temperature of approximately 2,000 F. 23,00 F. for about 1 3 hours. At the higher temperature I prefer to soak for only about one hour, while at the lower temperature I prefer to soak about three hours. However, it is to be understood that the particular time and temperature used will depend upon the particular electrical characteristics desired.
The particular form of isostatic pressing which can be used is not especially significant. The technique previously described herein, a wet bag" technique, can be employed as well as the dry bag technique which is disclosed in US. Pat. No. 3,034,191 Schaefer. The pressure used during the isostatic pressing is preferably about 15,000 to 20,000 pounds per square inch. However, in some instances one may desire to use a pressure as low as only 10,000 pounds per square inch.
The fired bar is sliced in my method using a diamond saw to obtain the narrowest saw width, and least kerf waste. However, other saws such as a silicon carbide saw could be used. The thickness of the slice sawed will depend upon the selected avalanche voltage which is desired. The general, it is most advantageous to use my method in making disc varistors less than about 0.5 inch thick.
1. A method of making nonlinear metal oxide resistors with low avalanche voltage characteristics comprising the steps of:
forming a uniform mixture of nonlinear resistor metal oxide powders for pressing into a predetermined shape,
isostatically pressing said mixture under a pressure of about 10,000 20,000 pounds per square inch to form an elongated body,
firing said elongated body at a temperature of about 2000 C. to 2300 C. for about 1 3 hours to produce a coherent mass having a predetermined nonlinear voltage-ampere characteristic,
transversely sawing said elongated body into a plurality of thin discs, each of which has substantially flat and parallel opposite major faces as formed that are suitable for attachment of electrodes and has a thickness less than about 0.05 inch commensurate with a predetermined avalanche voltage desired, and
attaching electrodes to said opposite major faces of said discs to form a plurality of discrete nonlinear metal oxide resistors that avalanche at a predetermined voltage below about 50 volts. 2. A method of making nonlinear metal oxide resistors with low avalanche voltage characteristics comprising the steps of:
uniformly mixing a plurality of nonlinear resistor metal oxides having a particle size less than about microns,
isostatically pressing said oxides under a pressure of about 15,000 20,000 pounds per square inch to form an elongated body,
firing said elongated body at a temperature of about 2000 C. 2300 C. for about 1 3 hours,
sawing said body into aplurality of discs having substantially flat and parallel opposite major surfaces that as formed are suitable for attachment of electrodes and having a thickness less than about 0.05 inch commensurate with a predetermined low avalanche voltage desired, and
attaching electrodes to said opposite major faces of said discs to form a plurality of discrete nonlinear metal oxide resistors that avalanche at a predetermined voltage below about 50 volts.
3. A method of making nonlinear metal oxide resistors with low avalanche voltage characteristics comprising the steps of:
uniformly mixing zinc oxide and a small proportion of other metal oxides with a temporary binder, said zinc oxide and other oxides having a particle size less than about 20 microns,
isostatically pressing said uniform mixture under a pressure of about 15,000 20,000 pounds per square inch to form an elongated body,
firing said elongated body at a temperature of about 2,000 C. 2,300 C. for about 1 3 hours to produce a coherent mass having a predetermined nonlinear voltage-ampere characteristic, transversely sawing said body into a plurality of discs having substantially flat and parallel opposite major faces as formed that are suitable for attachment of electrodes and having a thickness of less than 0.03 inch commensurate with a predetermined low avalanche voltage desired, and
attaching electrodes to said opposite major faces of said discs to form a plurality of discrete nonlinear metal oxide resistors that avalanche at a predetermined voltage below about 50 volts. 4. A method of consistently and reliably making metal oxide varistors with extremely low avalanche voltages comprising the steps of:
forming a uniform mixture of a plurality of metal oxide powders having a particle size less than about 20 microns and 2 5 percent by weight of an aqueous solution containing 2 percent by volume polyvinyl alcohol, said plurality of metal oxide powders consisting essentially of, by mole percent, about 98 percent ZnO. 0.5 percent Bi O 0.5 percent C0 0 0.5 percent MnO, and 0.5 percent Cr O passing said uniform mixture through a screen to obtain mixture agglomerates of desired size, drying the screened agglomerates, isostatically pressing the dry powder agglomerates into an elongated generally cylindrical body at about 15,000 20,000 pounds per square inch,
preheating said body to remove said polyvinyl alcohol from said mixture,
heating said body to a temperature of about 2,000
C. 2,300 C. and maintaining said body at said temperature for about 1 3 hours and then allowing said body to slowly cool,
transversely sawing said body into a plurality of discs having substantially flat and parallel opposite major faces as formed that are suitable for attachment of electrodes and having a thickness less than about 0.03 inch commensurate with a predetermined desired nonlinear voltage-amphere characteristic, and
attaching an electrode to each of said opposite major faces of said discs to form a plurality of discrete resistor discs having a nonlinear voltage-ampere characteristic with an avalanche voltage below about 50 volts.
5. A metal oxide varistor made in accordance with the method as recited in claim 1.