US3039071A

US3039071A - Electrical resistance-type heater

Info

Publication number: US3039071A
Application number: US43975A
Authority: US
Inventors: William M Ford
Original assignee: Individual
Current assignee: Individual
Priority date: 1959-07-06
Filing date: 1960-07-08
Publication date: 1962-06-12
Anticipated expiration: 1979-06-12

Description

June 12, 19 2 w. M. FORD 3,03 0

ELECTRICAL RESISTANCE-TYPE HEATER Original Filed July 6, 1959 4 Sheets-Sheet 1 INVENTOR. WILL/AM FORD June 12, 1962 W. M. FORD ELECTRICAL RESISTANCE-TYPE HEATER Original Filed July 6, 1959 4 Sheets-Sheet 2 IN V EN TOR.

WILL/AM M. F0120 June 12, 196 w. M. FORD 3,039,071

ELECTRICAL RESISTANCE-TYPE HEATER Original Filed July 6, 1959 4 Sheets-Sheet 3 IN VEN TOR.

WILLIAM M. FORD June 12, 1962 w. M. FORD ELECTRICAL RESISTANCE-TYPE HEATER 4 Sheets-Sheet 4 Original Filed July 6, 1959 R M V w. m i 1 M m 0 WM MA 9 8 (1 u e n A r o C BY WILLIAM M. FORD m fli ATTORNEY United States Patent 6 Claims. (Cl. 333-294) This invention relates to an electrical resistance-type heater used in apparatus for growing large crystals by a modified Czochralski technique. This application is a division of application Serial Number 824,975, filed July 6, 1959.

Heretofore, the Czochralski technique gave satisfactory results for growing crystals up to about two inches in diameter and, with some modifications, up to about five inches in diameter. The present invention makes it possible to grow crystals from six to twenty inches in diameter and larger.

For those unfamiliar with the Czochralski technique, it can be briefly summarized. A seed crystal, attached to one end of a rod, is dipped into a crucible of molten material and is then retracted slowly, pulling some of the material up with it. The crucible is heated warmer than the atmosphere above it, which is below the freezing or solidification point of the molten material. :In fact, the sides and bottom of the crucible are, when pulling begins, above the melting point of the material, and the center of the upper surface of the molten material is between the melting point and the freezing point. As a result the material that is pulled'up crystallizes and pulls up the material below it, which also crystallizes in turn. The process is continued until all the material has been crystallized in this manner. An important fact is that this kind of crystallization often produces a single large crystal and, if properly done, always produces crystals of single-crystal density. I

For small crystals, the technique is not difiicult to practice. Special furnaces, with thermocouples or devices sensitive to infrared radiation and recorder and control devices for the crucible heaters, have enabled its use for such difiicult materials as silicon and for crystals somewhat larger than two inches. But heretofore, the crystal size has been limited to a maximum of less than six inches in diameter and even then the formation of the larger crystals has been difi'icult and expensive.

The invention is applicable to many, if not all, types of crystals. Large crystals of silicon, germanium, aluminum, iron, magnesium, molybdenum, gallium arsenide, gallium antimonide, magnesium dioxide, calcium fluoride, lead selenide, indium antimonide, indium arsenide, aluminurn phosphide, and aluminum-antimonide can be made by this invention, but so can others. No restriction of the invention is intended by using in the following discussion, silicon as an example. Silicon is one of the'most useful crystal materials, one of the hardest to make-in large sizes, and one on which much work. has been done over many years without being able to get large, crystals. In

addition to its infrared uses, silicon (and some other materials) are-useful as semiconductors, for solar batteries, photovoltaic detectors, photoconductive detectors, transistors, rectifiers, and Peltier-junction materials, The present invention makes large crystals of silicon not only possible but also feasible and economical.

The present invention provides a novel electrical control of the thermal gradient in the molten material regulating the liquid-solid state equilibrium, and thereby makes possible the manufacture of single crystalline and dense polycrystalline material of diameters limited only by the power supply and the size of the crucible.

Other problems, as well as other objects andadvantages ice of the invention, are dealt with later on in this specification or will be quite apparent from a careful reading thereof.

In the drawings:

FIG. 1 is a view in front elevation of a furnace assembly embodying the principles of the invention.

FIG. 2 is an enlarged view in elevation and in section A of the furnace of FIG. 1.

FIG. 3 is a bottom plan view in section taken along the line 3--3 in FIG. 2.

FIG. 4 is a view in elevation and in section of a novel contact assembly that supports and supplies electrical energy to the heaters.

FIG- 5 is an electrical circuit diagram of the furnace heaters and their control devices.

As shown in FIG. 1, a furnace 25 embodying the principles of this invention may be supported by a liquid cooled plate 26 on standards 27 above a base 28. A seed rod 30 is mounted axially with respect to the furnace 25 and enters it from above. A motor 31 may be supported on the base 23 for raising and lowering the seed rod 30 through gears 32, shaft 33, gear boxes 34, rotating worms 35, and geared blocks 36. The blocks 36 carry a plate 37 that, in addition to serving as a bearinged support for the seed rod 30, also supports a motor 38 that rotates the seed rod 30. A similar motor 39, carried by the furnace 25 below the plate 26, rotates a crucible inside the furnace 25.

The furnace 25 (see also FIG. 2) has a metal shell or housing 40 enclosing a chamber 41. The shell 40 is made from metal capable of withstanding high temperatures and is cooled exteriorly, both by air and by cold water flowing through a coil '42. The chamber 41 is gastight, the shell 40 being sealed to the bottom plate 26 and to a top plate or cover 43, all openings through the plates 26 and 43 being tightly sealed. The plates 26 and 43 are preferably water cooled. All air is flushed from the chamber 41, and it is filled with an inert atmosphere, such as argon, through a port 44.

The top plate or cover 43 has a central sealed opening 45, through which the seed rod 30 passes, being both rotatable and reciprocable with respect thereto and sealed. Windows 46 enable viewing and a prism 47 enables measurement of the diameter of the crystal being grown. Ports 48 lead to the water-cooling chamber inside the cover 43. V

A very important part of this invention is the structure 1 of the electrical-resistance type heating elements. These comprise a peripheral heater 50 and a separate bottom heater 51, and the peripheral heater 50 constitutes a plurality of segments capable of separate adjustment.

In a silicon furnace 25 both the

heaters

50 and 51 are preferably constructed of graphite. Molybdenum disili- 'cide and other suitable materials may be used if desired.

Graphite has a very low vapor pressure at the high melting. temperature of silicon, and it has the necessary conductivity-resistance characteristics lacking in most other materials that do not vaporize at high temperatures. Thus it does not tend to adversely affect the purity of the silicon, as do materials that produce contaminating vapors at the operating temperature of the furnace, e.g., above 1420 C. for a silicon furnace. V

In essence, the heater 50 may be considered as a graphite tube 52 with a radially inset portion 53 and a shoulder 54 at its lower end. Notches 55 extend from the bottom to near the top where they end at bridges 56. Notches 57 extend from the top to the bottom, except for bridges 58 at the inner periphery of the portion 53. Thus a complete circuit of graphite bar is described, the graphite havingza substantially constant cross-sectional area throughout, to give a constant resistance through any section;

At the bottom and substantially coplanar with the shoulder 54- is the annular bottom heater 51. Notches 60 extend radially from the inner periphery 61 to outer peripheral bridges 62, while notches 63 extend in radially from the outerperiphery 64 to inner peripheral bridges 65. Again there is structure providing a constant-cross sectional-area path of graphite around 360, so that the heater 51 will heat evenly.

As shown in FIGS. 1 and 2 a pedestal housing 70 depends from the plate 26. The motor 39 drives a worm gear71 .which drives an annular pinion gear 72. The gear 72 may rest on a suitable bearing 73 which, in turn, may rest on a shoulder 74 of the housing 70, and a hollow graphite pedestal 75 extends upwardly therefrom into the chamber 4-1 and through the opening 61 in the heater 51. Since the upper end of the pedestal 75 is very hot, the housing 7 is preferably water cooled.

At the upper end of the pedestal 75 is threaded an annular graphite flange 76, On this sits a cup-shaped graphite susceptor 77, held in place by an integral depending ring 78 on its lower surface. Inside the susceptor 77 is a crucible 80 of interiorly glazed quartz. I have discovered that transparent fused quartz is not necessary; opaque quartz, if interiorly glazed, gives as good results and costs only a fraction as much. This discovery is quite important, for it had been thought heretofore that opaque quartz crucibles could not be used in making silicon crystals, as indeed they cannot unless the interior surface is well self-glazed. But once so glazed, there is as little tendency of the opaque quartz to combine with the silicon and form silicon monoxide as with the fused transparent quartz. Since the crucibles are almost invariably broken, and since opaque quartz costs less than one-tenth as much as transparent quartz, the economics 'of the process is considerably alfected by this discovery.

Moreover, transparent quartz is not commercially available in sizes greater than about 6 inches in diameter, whereas opaque quartz is available from stock up to 48 inches in diameter.

As will be apparent from the drawings and the foregoing description, the motor 39, through the gears 71 and 72, drives the pedestal 75 and therefore rotates the susceptor 77 and crucible 80 together. The rotation is slow (usually about 3-20 r.p.m.), serving principally to help avoid the formation of uneven temperature zones. By itself, rotation is not sufficient, however, for reasons that will soon become apparent.

At the bottom of the hollow pedestal 75 is a thermopile 81, such as a Rayo tube. This tube looks through the tubular passage 82 in the pedestal 75 up to a spot 83 at the center of the bottom of the susceptor 77 and detects small changes in temperature and then actuates a control device to take corrective steps. The arrangement is conventional and so will be described only briefly. The thermopile 81 (see FIG. is connected to a DC. amplifier 84 through a bucking circuit 85, the bucking circuit 85 having a battery 86, a load resistance 87 and a manually set potentiometer 88. Thus, once the potentiometer 88 is set, departures from the corresponding temperature on the spot 83 cause electric current to be applied to the amplifier 84. This amplified difference then goes to a suitable control unit that takes corrective action, such as a Leeds and Northrup Speedomax H recorder-controller 90, which is connected into the heating circuit, as by a lead 91.

Thus, the control device 90 keeps the susceptor spot 83 at a temperature set by the potentiometer 88. This temperature level may be raised and lowered by adjusting the potentiometer 88, the control 99 acting both on the peripheral heater 50 and the bottom heater 51.

The side wall 40 of the furnace is shielded from the heaters by a tubular shield 92 of opaque quartz, which is glazed on its interior surface and reflects the heat back to the

heaters

50 and 51. Its reflectiveness is afiected,

however, during operation by small condensations of tend to result in letting one segment of the heater 50 get cooler than the other segments. This problem, and others causing cool areas on the peripheral heater 50, are solved by the present invention, in which the unitary heater 50 is divided electrically into at least three arcuate cylindrical segments, four

such segments

93, 94, 95, and 96 being shown in the device illustrated here. Each of these segments can be separately controlled, and all of them are also controlled together by the aforementioned control system.

As shown in FIG. 5, each

segment

93, 94, 95, and 96 and the bottom heater 51 is located across a respective secondary 97 of a separate transformer 98, there being five transformers 98or as many as there are heating units. The primary 99 of each transformer 98 is separately coupled through a saturable reactor 100 to a source of primary power by lines 101 and 162. Typically this source supplies 220 volts, single phase at 600 amperes. The saturation coil of each reactor 100 is energized by the output of a magnetic amplifier 103. Each magnetic amplifier 103 is connected to a rheostat 104, and all the rheostats 104 are connected to the control line 91.

Thus, by setting each rheostat 104 separately, the

segments

93, 94, 95, and 96 of the heater 50 can be adjusted relatively to each other and to the heater 51. Then, the potentiometer 88 can beused to raise and lower the current to all the heaters and their segments while leaving the separate relative adjustments undisturbed. This control is what enables me to grow large crystals where others had failed.

An important factor in the present invention is that the necessary separate adjustment of peripheral segments, which is essential to growing large crystals, cannot be obtained from radio-frequency induction-coil furnaces. There, the annular passage of each turn of the coil prevents such adjustment. The prior-art furnaces all use induction coils, which are incapable of the necessary delicate adjustments achieved by this invention. The use of electrical-resistance type heaters is therefore an important feature of the invention. Note however, that segmentation of the heater 50 does not mean four separate lengths of graphite. All it requires is four terminals 105, coupled to the transformers 98 with the proper phase balance, as shown in FIG. 5, with alternate secondaries 97 out of phase,

The prior art had always turned away from high-heat furnaces using resistance heaters, because ordinary resistance heaters cannot transmit enough power attemperatures of 1500" C., as is required with silicon and as my

graphite heaters

50 and 51 can, and because of the difiiculty of connecting power to such heaters. This invention has also solved that difiicult problem.

The problem arises for several reasons: (1) copper cables cannot be connected directly to the

heaters

50, 51 for the copper would melt; (2) graphite and metal have greatly different temperature coeflicients of expansion and this tends to cause either element breakage or loosecontacts, with resultant erratic behavior; (3) graphite-tometal contacts have a high surface resistance, especially at low contact pressures; (4) copper oxidizes rapidly at high temperatures, resulting in high contact resistances; (5) hot graphite oxidizes rapidly in air; and (6) the high heat concerned is diflicult to dissipate. a

The invention solves this by (1) making graphite-tographite contact at the heaters, so that there is no problem of melting, uneven expansion, loose contacts, or surface resistance; (2) providing graphite to copper (silicon bronze) contact over extended areas and in such a way that the graphite is hot and the copper cool, thereby creating hi'gh contact pressures that reduce the surface resistance; (3) cooling this contact area at a point outside the furnace; and (4) keeping all the graphite and the graphiteto-copper contact inside an inert gas atmosphere, where oxidation cannot take place.

In the present invention the heater 50 has four graphite terminals 105 with sockets 106 supported by four contact assemblies 110 and the heater 51 has two stepped graphite terminals 107 with sockets 108 supported by two contact assemblies 111. Except for minor geometry and size considerations, these assemblies 110 and 111 are identical; so the following description of an assembly 110 applies to the assembly 111 as well. Reference will be mainly to FIG. 4.

The assembly 110 is made up of a graphite post 112, a silicon-bronze screw 113, and a silicon-bronze receptacle 114.

The post 112 is solid graphite. It includes a thick column 115 that has a shoulder 116. The shoulder rests on a support-insulator ring 117, preferably of Transite (asbestos) that in turn rests in an annular recess 119 in the plate 26 and insulates the post 112 from the plate 26, while centering it in an opening 129 through the plate 26. An upper end portion 118 of the graphite post 112 threads into the heater socket 105 and provides the graphite-to-graphite contact :at that point. An interiorly and exteriorly threaded portion 120 of the post 112 extends down below the shoulder 116 and through and below the plate 26, and into it is threaded the silicon-bronze screw 113 to provide considerable contact area, only a head 121 projecting out at the bottom.

The silicon-bronze receptacle 114 has an interiorly threaded passage 121 which is threaded around the portion 120 of the graphite post 112 to provide considerably more contact area and enclosure of the graphite and of the graphite-to-copper contact. Its upper end 122 is provided with an annular groove 123 in which a sealing and insulating O-ring 124 (it may be made from silicone rubber) is seated. The O-ring 124 bears tightly against the bottom of the plate 26 and insulates the receptacle 114 from it and also prevents air from reaching the graphite post 112. For one problem was that graphite oxidizes readily and to give it access to air at the high temperatures involved, would mean rapid consumption of the post 112. This structure prevents that by encasing it in the siliconbronze receptacle 114, where the post 112 has access only to the inert atmosphere of the chamber 41.

Cooling is obtained by an aluminum tube 125 which surrounds the receptacle i114. and provides a chamber 126 closed by O-rings 127 except for ports 128 through which cooling water enters, circulating around almost the whole length of the receptacle 114 and the depending portion 120 of the post 112. By cooling the silicon bronze receptacle 114, the hot graphite 120 exerts considerable contact pressure on the receptacle threads 121 increasing conductivity.

The bottom of the receptacle 114 is provided with a threaded binding post 130 to which the electrical cables are attached. Contact is made by engagement of the silicon-bronze screw 113 interiorly as well as by the exterior sleeve portion 121 of the receptacle 114. The screw head 121 fits snugly in a socket 131 to give current transfer Without the possibility of exposing any graphite to air.

Operation begins'by placing silicon chunks in the crucible 80, inserting the crucible 801into1the susceptor '77 and in the furnace 25, closing the top 43 tightly, and putting an inert atmosphere under pressure into the chamber 41 through the ports 44. Then, maximum power is applied to all the

heater segments

93, 94, 95, and 96 and to the heater 51 while also rotating the crucible 80, by the motor 39 rotating the pedestal 75. The seed rod 30 is kept up high at this stage or even outside the chamber 41 to protect the seed. The chamber 41 can be kept closed except during insertion of the seed rod 30; external gases do not enter the atmosphere within the chamber 41, during insertion, since the gas inside tentiometer 88. The motor 39 is stopped, so that the crucible 80 is stationary. Then the rheostats 104 are each separately adjusted to obtain an even temperature from all

segments

93, 94, 95, and 96, this being dope by observing the silicon while cooling the crucible 80 to the solidification point of silicon and making sure that icing occurs evenly over the surface, adjusting the separate segments until it does. Observation is through the windows 46. The adjustment of the

segments

93, 94, 95, and 96 is then complete, and no further adjustment in the rheostats 104 is made during that particular crystal unless random solidification is observed, in which instance readjustment is made.

In the third step, the seed rod 30, carrying a seed crystal is lowered into the molten silicon. Before this, the motor 39 is again started to rotate the crucible 80 slowly, while the motor 38 rotates the seed rod 30 in the opposite direction at about the same speed. The temperature of the upper surface of the liquid has now to be adjusted, so that its periphery is above the melting point so that random solidification will not occur at the periphery. The center of the upper surface is kept below the melting point and above the freezing or solidification point of the silicon. To do this, the heaters and 51 are ini-tally separately adjusted, the heater 50 generally being at this time a few degrees cooler than the heater 51. In other words, there is a temperature differential. In the entire process, the bottom vand sides of the crucible 80 have to be kept above the melting point of the silicon to prevent the growing crystal from adhering to any point on the crucible 80.

With the temperature properly adjusted, the seed crystal is lowered until it just enters the liquid. Then it is raised, forming a meniscus and crystal growth begins. At this stage, the potentiometer 88 is turned to lower the power to all segments of the heater 50 and to the heater 51, to lower the temperature rapidly, while the rod 30 is pulled very slowly, to get the maximum growth rate.

As the crystal approaches the desired diameter-a value slightly less than the inside diameter of the crucible 80 the pull rate on the rod 30 is increased. The temperature is maintained, and the ingot shoulders off.

Once the maximum diameter is achieved, the pull rate and the temperatures of the heaters are varied to maintain the minimum variations in that diameter. As the liquid level drops, the potentiometer 88 is turned to lower gradually the heat to both

heaters

50 and 51, and also the temperature differential is lowered by adjust- 1 ment of the heater 51 through the r-heos-tat 104. The

pulling continues in this manner. If freezing occurs at the crucible wall, the peripheral heat is increased and the frozen portion remelted; if freezing occurs beneath the ingot (stopping rotation of the crucible the central heat is increased to remelt that frozen portion. The gradual downward temperature adjustment continues with or without these interruptions until all the silicon is gone. The'power then is turned off, the pulling stopped and, after cooling, the crystal ingot is withdrawn.

The present invention makes it possible to grow different kinds of crystals and to make adjustments in each.

kind of crystal. A typical ingot of the type just discussed has a sloping upper end, but a substantially flat upper end can be provided by taking care in the fourth step to have an initialpull rate that is nearly zero to get the maximum radial growth with minimum vertical growth. Then, when the silicon reaches its maximum desired diameter, both the pull rate and the temperature are increased to shoulder off the crystal. Then the pull rate is kept substantially uniform and the temperature gradually decreased so as to hold this diameter, and pulling is continued until all the silicon is exhausted.

By varying the type of seed and the shape of the crucible, other shapes can also be made in a manner now quite apparent from the foregoing explanations.

7 c To those skilled in the art to which this invention relates, many changes in construction and widely differing embodiments and applications of the invention will suggest themselves without departing from the spirit and scope of the invention. The disclosures and the description herein are purely illustrative and are not intended to be in any sense limiting.

I claim:

1. An electrical-resistance heater comprising a unitary tube of graphite with a radially inwardly extending flange at its lower end terminating in an inner periphery, said tube having a series of first evenly spaced axially extending slots extending through said tube from the lower end up to a distance short of the upper end, leaving bridges there, and, midway in between said first slots, a series of second evenly spaced axially extending slots extending through said tube from the upper end down through said lower end except for bridges at the inner periphery of said flange, so that there is a continuous graphite bar of great length.

2. The apparatus of claim 1 wherein said bar has a substantially constant cross-sectional area at all points.

3. The apparatus of claim 2 wherein heater segments are provided by electrical contact members evenly spaced around the lower end of said tube.

4. An electrical-resistance type heater comprising a graphite ring having an inner periphery and an outer periphery with a first series of evenly disposed radiallyextending notches extending from said inner periphery through said graphite and out to outer peripheral bridges and a second series of evenly disposed radially extending notches, each midway between two notches of said first series, extending from said outer periphery in to inner peripheral bridges, to make a continuous graphite bar.

5. The apparatus of claim 4 wherein the cross'sectional area of said bar is constant.

6. The apparatus of claim 5 having two contact members exactly 180 apart.

References Cited in the file of this patent UNITED STATES PATENTS