US3826625A

US3826625A - Method and apparatus for growing crystalline bodies from the melt using a porous die member

Info

Publication number: US3826625A
Application number: US00196450A
Authority: US
Inventors: J Bailey
Original assignee: Tyco Laboratories Inc
Current assignee: Saint Gobain Ceramics and Plastics Inc
Priority date: 1971-11-08
Filing date: 1971-11-08
Publication date: 1974-07-30
Anticipated expiration: 1991-07-30

Abstract

THIS INVENTION IS AN IMPROVEMENT OVER THE METHOD DESCRIBED IN U.S. PAT. NO. 3,591,348 FOR GROWING CRYSTALLINE BODIES FROM THE MELT. THE IMPROVEMENT ESSENTIALLY CONSISTS OF USING A POROUS FORMING MEMBER (ALSO CALLED A DIE) THAT IS CHARACTERIZED BY AN INTERCONNECTING NETWORK OF PORES OR CELLS OF CAPILLARY PROPORTIONS.

Description

W WM J. S'BAILEY METHOD AND APPARATUS FOR GROWING CRYSTALLINE BODIES FROM THE MELT USING A POROUS DIE MEMBER 2 Sheets-Sheet 1 Filed Nov. 8, 1971 PULLING MECHANISM INVENTOR JOHN S. BAILEY BY -S2Ai/[er pancliacio AT TORNEYS y 3U, EM J. 5. BAILEY METHOD AND APPARATUS FOR GROWING GRYSTALLINE BODIES; FROM THE MELT USING A POROUS DIE MEMBER 2 Sheets-Sheet 2 Filed Nov. 8, 1971 H wm INVENTOR.

JOHN S. BAILEY pana L'Jcio ATTORNEYS mv l k United Sttes Patent 3,826,625 METHOD AND APPARATUS FOR GROWING CRYSTALLINE BODIES FROM THE MELT USING A POROUS DEE MEMBER John S. Bailey, (Iharlestown, Mass, assiguor to Tyco Laboratories, inc, Waltham, Mass. lFiled Nov. 8, 1971, Ser. No. 196,450 lint. Cl. B011} 17/18 TLS. Cl. 23-301 S1 14 (llaims ABSTRACT OF THE DISCLOSURE The invention is an improvement over the method described in U.S. Pat. No. 3,591,348 for growing crystalline bodies from the melt. The improvement essentially consists of using a porous forming member (also called a die) that is characterized by an interconnecting network of pores or cells of capillary proportions.

This invention relates to the art of crystal growth and more particularly to growth of substantially monocrystalline bodies having cross-sections of arbitrary and predetermined configurations.

Various methods have been developed for growing monocrystalline bodies from a melt. The present invention pertains to an improvement in growing crystalline bodies from a melt according to what is called the edgedefined, film fed growth technique (also known as the EFG process). Details of this process are described in United States Letters Pat. No. 3,591,348, issued July 6, 1971 to Harold E. LaBelle, In, for Method of Growing Crystalline Materials.

In the EFG process the shape of the crystalline body is determined by the external or edge configuration of the end surface of a forming member which for want of a better name is called a die. An advantage of the process is that bodies of selected shapes can be produced commencing with the simplest of seed crystal geometries-namely, a round small diameter seed crystal. The process involves growth on a seed from a liquid film of feed material sandwiched between the growing body and the end surface of the die, with the liquid in the film being continuously replenished from a suitable melt reservoir via one or more capillaries in the die member. The die is made of a heat resistant material that is wetted by the feed material. By appropriately controlling the pulling speed of the growing body and the temperature of the liquid film the film can be made to spread (under the influence of the surface tension at its periphery) across the full expanse of the end surface of the die until it reaches the perimeter or perimeters thereof formed by intersection of that surface with the side surface or surfaces of the die. The angle of intersection of the aforesaid surfaces of the die is such relative to the contact angle of the liquid film that the liquids surface tension will prevent it from overrunning the edge or edges of the dies end surface. Preferably the angle of intersection is a right angle which is simplest to achieve and thus most practical to have. The growing body grows to the shape of the film which conforms to the edge configuration of the dies end surface. Since the liquid film has no way of discriminating between an outside edge and an inside edge of the dies end surface, a continuous hole may be grown in the crystalline body by providing in that surface a blind hole of the same shape as the hole desired in the growing body, provided, however, that any such hole in the dies end surface is made large enough so that surface tension will not cause the film around the hole to fill in over the hole. From the foregoing brief description it is believed clear that the term edge-defined, film fed growth denotes the essential feature of the EFG process the shape of the growing crystalline body is defined by the edge configuration of the die and growth takes place from a film of liquid which is constantly replenished.

Thus it is possible to grow bodies of various shapes, e.g. round tubes, fiat ribbons, and filaments, to substantially close tolerances on a continuous or semi-continuous basis. One important use of the EFG process is to grow monocrystalline tubes of selected ceramic materials such as alumina for use as envelopes for high intensity vapor lamps. However, as heretofore practiced, the EFG process has been burdened by the relatively high cost of manufacturing the forming members or dies. One reason for this high cost is the difiiculty of forming one or more capillaries in the die. Since the usual capillaries are of relatively small size, eg a diameter in the order of 0.010 inch, they can be formed only with sophisticated techniques and equipment. This is not only time consuming but costly. Furthermore, in order to maximize the rate of crystal growth and also facilitate establishing and maintaining the liquid film from which crystal growth occurs, it is often necessary to provide two or more capillaries in the forming member, with the several capillaries located at strategic points calculated to provide an even distribution of melt to the growth film.

OBJECTS AND SUMMARY OF THE INVENTION Accordingly, a primary object of this invention is to improve upon the aforesaid EFG process so as to facilitate growth of substantially monocrystalline bodies of selected and predetermined cross-sectional configurations.

Another primary object of this invention is to provide improved apparatus for growing crystalline bodies of selected cross-sectional shapes from a thin film of melt.

A specific object is to provide an improved forming member for the purpose described which is less expensive to manufacture and provides faster and more uniform dis tribution of melt to replenish the .film from which growth occurs.

Described briefly the improvement of this invention comprises providing a forming member which is made of a heat resistant foam material having interconnecting open cells with the interstices (i.e., the cells and their connecting passageways) being of capillary proportions so as to permit melt to rise within the forming member by capillary actlon.

Other features and many of the attendant advantages of this invention are set forth or rendered obvious in the following detailed description which is to be considered together with the drawings.

THE DRAWINGS FIG. 1 is a fragmentary elevati-onal view, partly in section, of apparatus comprising a furnace, crucible, and a die assembly used in practicing the invention to grow a tubular body;

FIG. 2 is an enlarged view of a portion of the apparatus of FIG. 1 showing details of the die assembly and the initial stage of the process of growing a tubular body;

FIGS. 3 and 4 are fragmentary views of the die assembly of FIG. 2 showing successive stages in the growth of a tubular monocrystalline body;

FIG. 5 is a view similar to FIG. 2 of a die assembly used to grow a solid rod;

FIGS. 6 and 7 show a modification of the die of FIG. 2; and

FIG. 8 is another modified form of die for growing a tubular body.

Identical or corresponding parts: of the same or different embodiments of the apparatus are referred to by the same numerals in the several figures.

The present invention may be used to produce integral substantially monocrystalline bodies made of any one of a variety of materials that melt congruently (i.e., compounds that melt to a liquid of the same composition at an invarient temperature) and solidify in identifiable crystal lattices. By way of example, the material may be alumina, barium titanate, lithium niobate and yttrium aluminum garnet. The invention is also applicable to other materials that do not melt congruently, e.g., alloys or materials that form mixed crystals. The following detailed description of the invention and the apparatus used in practicing the same is directed to growing monocrystalline alumina bodies of selected geometry.

FIG. 1 shows one form of furnace that may be used to practice the invention. The furnace consists of a vertically movable horizontal bed 2 which engages a stationary furnace enclosure consisting of two concentric-spaced quartz tubes 4 and 6. At its bottom end the inner tube 4 is positioned in a gasket 5 in the bed. Surrounding tube 4 is a sleeve 8 that screws into a collar 10. Between sleeve 8 and color 10 is an O-ring 12 and a spacer 13. The O- ring is compressed against tube 4 to form a seal. The upper end of sleeve 8 is spaced from tube 4 so as to accommodate the bottom end of tube 6. The bottom end of tube 6 is secured in place by an O-ring 14 and a spacer 15 compressure between a collar 16 that screws onto sleeve 8. Sleeve 8 is provided with an inlet port fitted with a flexible pipe 20. The upper ends of tubes 4 and 6 are secured in a head 22 so that they remain stationary when the bed is lowered. Head 22 has an outlet port with a flexible pipe 24. Although not shown it is to be understood that head 22 includes means similar to sleeve 8, O-

rings

12 and 14, and collars 10 and 16 for holding the two tubes in concentric sealed relation. Pipes and 24 are connected to a pump (not shown) that continuously circulates cooling water through the space between the two quartz tubes. The interior of the furnace enclosure is connected by a pipe 28 to a vacuum pump or to a regulated source of inert gas such as argon or helium. The furnace enclosure also is surrounded by an R.F. heating coil 30 that is coupled to a controllable 500 kc. power supply (not shown) of conventional construction. The heating coil may be moved up or down along the length of the furnace enclosure and means (not shown) are provided for supporting the coil in any selected elevation. At this point it is to be noted that the circulating water not only keeps the inner quartz tube at a safe temperature but also absorbs most of the infrared energy and thereby makes visual observation of crystal growth more comfortable to the observer.

The head 22 is adapted to provide entry into the furnace enclosure of an elongate pulling rod 32 that is connected to and forms part of a conventional crystal pulling mechanism represented schematically at 34. It is to be noted that the type of crystal-pulling mechanism is not critical to the invention and that the construction thereof may be varied substantially. However, it is preferable to employ a crystal pulling mechanism that is hydraulically controlled since it otters the advantage of being vibration-free and providing a uniform pulling speed. Regardless of its exact construction which is not required to be described in detail, it is to be understood that the pulling mechanism 34 is adapted to move pulling rod 32 axially at a controlled rate. Pulling rod 32 is disposed coaxially with the quartz tubes 4 and 6 and its lower end has an extension in the form of a metal holder 36 that is adapted to releasably hold a monocrystalline seed or body 38 on which an integral monocrystalline extension is to be grown as hereafter described.

Located within the furnace enclosure is a cylindrical heat susceptor 40 made of carbon. The top end of .susceptor 40 is open but its bottom end is closed off by an end wall. The susceptor is supported on a tungsten rod 42 that is mounted in bed 2. Supported within susceptor 40 on a short tungsten rod 44 is a crucible 46 adapted to contain a melt 48 of the material to be grown onto the tube 38. The crucible is made of a material that will withstand the operating temperatures and will not react with or dissolve in the melt. With an alumina melt, the crucible is made of molybdenum, but it also may be made of tungsten, iridium or some other material with similar properties with respect to molten alumina. Where a molybdenum crucible is used, it must be spaced from the susceptor since there is a eutectic reaction between carbon and molybdenum at about 2200 C. The inside of the crucible is of suitable size and shape, preferably with a constant diameter. To help obtain the high operating temperatures necessary for the process, a cylindrical radiation shield 50 made of carbon cloth may be wrapped around the carbon susceptor. The carbon cloth greatly reduces the heat loss from the carbon susceptor.

Referring now to FIGS. 1 and 2, mounted in crucible 46 is a die assembly 56 comprising a cylindrical rod 58 that is aifixed (e.g. by a press fit) to a supporting disc 60 that rests on a shoulder 62 formed in the side wall of the crucible at the upper end thereof. The rod 58 is made of open-celled molybdenum foam. The open cells or pores of the foam are identified by the numeral 64. At least some of these cells are sized to function as capillaries for molten alumina. The upper end of rod 58 terminates in a flat horizontal surface 66 which intersects the rods outer surface at a right angle. It is to be noted that rod 58 projects above disc 60 so as to be visible to the operator and has a relatively large coaxial blind hole or cavity 68 extending down from its upper surface. Cavity 68 is large enough in diameter so that surface tension will not cause the film (hereinafter described) to fill in over it. The length of the rod 58 and the size of the open cells or pores 64 are such that molten alumina can rise in an fully till the open cells by action of capillary rise so long as the level of the melt in the crucible is above a predetermined level, preferably so long as the bottom of the rod is wetted by melt. In this connection it is to be noted that the inerconnected network of pores may be viewed as providing a plurality of vertical passageways or channels of capillary dimension, whereby in each passageway a column of melt can rise to a height determined by the equation:

where h is the distance in cm. that the column will rise; T is the surface tension of the melt in dynes/cm.; 0 is the contact angle; a' is the density of the liquid; r is the internal radius of the passageway in cm.; and g is the gravitational constant in cm./sec. By way of example, in a capillary of 0.75 mm. diameter in a molybdenum memher, a column of molten alumina may be expected to rise more than 11 cm. by capillary action.

FIGS. 2, 3, and 4 illustrate how a monocrystalline tube of a ceramic material may be grown on a monocrystalline ceramic seed tube 38. The seed tube 38 is of the same composition as the melt and may have been grown using a porous die as herein described or by using a non-porous die in accordance with the method described in said US. Pat. No. 3,591,348. The tube 38 may have the same dimensions as the tube to be grown or may have different dimensions, e.g. the tube may have the same CD. but a smaller ID. as shown in FIGS. 2-4. Initially the tube is inserted vertically and is disposed in axial alignment with the die assembly. Then with the pores of the rod 58 filled with melt by action of capillary rise and the power input to coil 30 adjusted so that the upper surface 66 of the die assembly is preferably at least about 1040 C. higher than the melting point of the tube 38, the tube is lowered into contact with the surface 68 and held there long enough for a portion of the end of the tube to melt and form a liquid film 70 that connects with the melt in the topmost pores of the die. It is to be noted that the pores are shown empty in FIGS. 2-4 (and also 6-8) in order to render them more distinct to the reader and that in fact they are filled with melt as above described. Further it is to be understood with reference to FIG. 2 that before the end of tube 38 is melted to form film 70, the melt in each port that opens into the surface 66 has a concave meniscus with the edge of the meniscus being substantially flush with surface 66. The temperature gradient along the length of the tube and the temperature of surface 66 are factors influencing how much of the tube melts and the thickness of film 70. In this connection it is to be noted that the tube functions as a heat sink and the temperature of the tube at successively higher points thereon is affected by the height of coil 30 and susceptor 40 and also the power input to the coil. In practice these parameters are adjusted so that the initial film 70 has a thickness in the order of 0.1 mm.

Once the film 70 has connected with melt in the pores of the rod 58, the pulling mechanism 34 is actuated to pull tube 38 upwardly away from surface 66. The pulling speed is set so that the film adhering to the tube because of surface tension will crystallize due to a drop in temperature at the solid tube-liquid film interface which occurs as a result of the pulling. The pulling speed also must be such that surface tension will cause the film to spread inwardly toward the boundary of cavity 68 (see FIG. 3). As the film spreads inwardly, crystal growth will occur at all points along the horizontal expanse of the film with the result that a tubular monocrystalline extension is formed on the tube which has a constant outside diameter but a progressively decreasing inside diameter. The film consumed by the crystal growth is replaced by additional melt which is supplied by inflow via the pores 64. During the initial phase of crystal growth, the film continues to spread inwardly until it reaches the boundary of cavity 68, where it stops spreading. Consequently the initial crystal growth on the tube appears to form a tapered inside flange (FIG. 3). Thereafter the crystal growths stop spreading inwardly and continues longitudinally, with the result that a tubular extension 72 is produced that has the same OzD. as the rod 58 and an ID. corresponding substantially to the diameter of cavity 68. Growth is continued until supply of melt in the crucible has been substantially depleted or until the pulling mechanism has reached the limit of its pulling stroke. In the latter case as the end of the pulling stroke is reached, the pulling speed is increased enough to cause the crystal growth to pull free of film 70.

It is to be noted that the pulling speed and the temperature of the film may be varied during crystal growth. However, during crystal growth the pulling speed should not be so great nor the temperature so high as to cause the tube to pull free of the melt film. In growing alpha-alumina, it is preferred to have an initial pulling speed of about 0.1 in/min. and to increase the speed to about 0.2 in./ min. after the film has expanded enough to fully cover the end surface 66 of the die assembly heated as above described. The pulling speed of the tube and the temperature of the film control the film thickness which controls the rate of film spreading. Increasing the temperature of surface 66 (and hence the temperature of the film) and increasing the pulling speed each have the effect of increasing the film thickness.

FIG. 5 shows a modification of the invention for growing a round monocrystalline rod or filament (the distinction between a rod and a filament is a difference in outside diameter, i.e., O.D.). The die assembly for growing a rod or filament is the same as that for growing a tube except that no blind hole or cavity is provided in the upper surface of the rod 58. In growing a rod or filament, the seed may be a rod or filament or tube or other available monocrystalline body, e.g., a ribbon of rectangular cross-section. In this case the seed is a round hollow tube 38. The upper surface of rod 58 may be flat as in FIG. 2 or may be concave as shown in FIG. 5. The growth process is essentially as described above in connection with FIG. 2-4. Initially the seed is brought into engagement with the upper surface 66A of the rod long enough to melt and form a film 7tl. Enough of the seed may be melted to form a film of annular shape as shown in FIG. 2 in which case the film must be made to spread out so as to fully cover surface 66A. Preferably enough of the seed is melted so that the film fully covers the surface 66A as shown in FIG. 5. The film 70 connects with melt in the pores of the die,

so that as growth proceeds additional melt flows up through the die to replenish the film. The initial pulling speed is set according to whether or not it is necessary to cause the film to spread out to cover all of the surface 66A. As pulling of the seed proceeds, crystal growth will occur at all points along the film so that with the film fully covering surface 66A, the crystal growth as shown at 73 will be a solid monocrystalline rod having substantially the same CD. as rod 58.

FIG. 6 shows still another modification of the invention. In this case the die assembly is the same as that of FIG. 2 except that the cavity 8 is filled with a material 74 that meets the following requirements: (a) is not wetted by the melt and (b) will not react with the melt, the foam of which the die is made, the disc 60 or the crucible 46. The material 74 may be either a liquid or a solid at the operating temperature of the process. However, if it is a liquid, it must not vaporize at the operating temperature and pressure of the system. By way of example, if the melt is alumina and the die assembly is made of molybdenum, the material 74 may be tin. Further by way of example, if the melt is germanium and the rod 58 is made of tungsten foam, the material 74 may be a solid plug of carbon. Similarly, if the melt material is copper and the die is made of molybdenum, the material 74 may be a solid plug of alumina. Tungsten and molybdenum are wetted by germanium and copper respectively, but carbon and alumina are not wetted by germanium and copper respectively. Use of non-wetting material 74 is advantageous in that it prevents the cavity from being accidently filled with melt as can occur, for example, when the seed is melted for the purpose of forming film 70 if the seed is initially contacted with the inside edge of surface 66 at a temperature so high that a substantial portion of the seed melts and flows directly into the cavity. If the cavity is filled with melt, the film 70 will close over it and the crystal body that is grown will be a solid rod instead of a tube. This problem is obviated by prefilling the cavity with the nonwetting material 74. Since the latter is not wetted by the melt, it assists in defining the inner edge of surface 66, with the result that film 70 will extend over the entire surface 68 and a tubular monocrystalline body is grown on the end of the seed as shown in FIG. 7.

It also is contemplated to clad the porous die member so as to render its vertical surfaces non-porous, this embodiment is shown in FIG. 8. In this figure the rod 58 is made of molybdenum foam as above described and the sides and bottom surfaces defining the cavity 68 are cladded with a layer 78 of non-porous molybdenum. If desired, the exterior surface of rod 58 may also be cladded with a layer '80 of non-porous molybdenum. Cladding the exterior surface of the rod 58 and the side surface of the cavity 68 is beneficial in preventing damage to the die and minimizing introduction of impurities into the pores of the die. More emphatically the cladding assures that the inside and outside edges of the top surface 68 are sharp and also smooth and uninterrupted. Uninterrupted edges are essential if the body to be grown is to have a smooth surface, while shape edges are desirable to assure that the film of melt will not overrun the edges of the top end surface. Cladding the side and bottom surfaces of cavity 68 is advantageous where the pores in such surfaces are oversized to the extent that the capillary forces are so so small that melt will tend to run out of the pores into the cavity, in which case, if suflicient melt accumulates in the cavity, a solid rod rather than a tubular body will be grown for the reasons described above in connection with FIG. 6. Cladding the vertical surfaces of the rod 58 also has the effect of strengthening it. This is particularly beneficial where the wall thickness of that section of the rod having the cavity 68 is relatively small. In practice it is preferred to finish machine the upper surface 66 of the rod 58 after it has been cladded, so that the upper end surfaces of the cladding are flush with and form extensions of surface 66. Thus in practicing the process the film 70 will extend over the upper end surfaces of the cladding and the grown crystal will have an OD. substantially the same as the CD. of cladding 80 and its ID. will be substantially the same as the diameter of cladding 78. It is to be noted that the cladding 80 need not extend for the full length of rod 58 and, for example, may extend from the top end surface 66 only down as far as disc 60. Similarly, the cladding 78 may extend from the top end surface 66 only part way down the side surface of cavity '68.

Of course the invention is applicable to dies for growing crystal bodies of other geometric shapes, e.g. bodies of rectangular or triangular cross-sections as disclosed in said U.S. Pat. No. 3,591,348. In each case the die may be cladded as above described. Of course the cladding should consist of a material that will be wetted by the melt and will not react with the melt, the crucible or the component parts of the die assembly. Preferably the cladding consists of the same composition as the die. Thus if the die is made of a porous ceramic, the cladding should consist of a dense non-porous coating of the same ceramic. The cladding may consist of a preformed member that is force fitted to the die. Thus cladding 80 may be a molybdenum sleeve that is press-fited over rod 58. Alternatively, the cladding may be formed in situ by dipping, vapor deposition, or other suitable technique known to persons skilled in the art.

It is to be noted that a number of materials are commercially available in porous form, e.g., open-cell foam or sponge. Among these are molybdenum and titanium sponge. Porous ceramics are also available commercially. They also may be made by infiltrating an open cell metal foam material, e.g., molybdenum sponge with a molten ceramic material, e.g. alumina, solidifying the ceramic in the foam, and then burning or vaporizing or leaching out the metal so as to leave a porous ceramic. Another mode of forming an open cell porous ceramic is to ad mix particles of ceramics and a material having a relatively low burning point, compressing the composition to a suitable density, and then sintering the compressed mass to burn out the low-burning point material.

As used herein, the term open-celled foam (sometimes also identified in the art as open-celled sponge) refers to a body of a selected material which may be a metal or metal alloy or a ceramic material, having interconnected open cells. The foam can be visualized as being made up of solid particles of relatively uniform size held together in random fashion at corners or along edges, leaving holes or cells between them that form an interconnecting network of pores. The holes or cells may be of a size comparable to the size of the solid particles and may range in size from micropores to macropores. As used herein the term micropore is a pore that is small enough to fill and hold a given liquid (i.e., melt material) by capillary action whereas a macropore is one that is large enough to exhibit substantially little or no capillary attraction for the same liquid, whereby the liquid will drain from the macropore by gravity. Of course, what is a macropore for one liquid may be a micropore for another liquid; it depends on the surface tension of the liquid that wets the foam. Furthermore, the optimum or preferred size of the pores depends upon the height to which themelt must rise in the die by capillary action. Preferably, the overall height of the die and the size of the pores are such that melt can rise to the top of the die by capillary action so long as the bottom of the die is trapped by (i.e. submerged in) or wetted by melt in the crucible. A minimum requirement is that at least some of the cells or pores at all levels of the die be small enough to function as capillaries for the melt material so that the melt material can rise to the top end of the die by capillary action. Preferably at least 50% of the pores are of micropore size and more preferably substantially all of the cells are of micropore size and are of substantially uniform size and shape. The presence of some macropores is helpful in accelerating infiltration of the submerged portion of the die by the melt material.

Accordingly, the density of the foam may vary but the density should not be so low as to render the foam too fragile to withstand handling and machining. Foam density is conveniently measured in percent density of the standard density of the material of which the foam is made. Thus a 40% density modlybdenum foam means a foam whose density is 40% of that of molybdenum under standard conditions (which is 10.2 grams per cubic centimeter).

The following example illustrates a preferred mode of practicing the invention. A molybdenum crucible having an internal diameter of about 1% inch, a wall thickness of about inch, an an internal depth of about inch is positioned in the furnace in the manner shown in FIG. 1. Disposed in the crucible is a die assembly constructed generally as shown in FIG. 2. The disc 60 is made of molybdenum and the rod 58 is made of porous molybdenum having open interconnected cells with an average diameter of 0.0004 inch. Rod 58 is round and its dimensions are as follows: a rod diameter of about inch and a rod length such that its flat upper end projects about inch above the crucible. The cavity 68 has a depth of about inch and a diameter of about inch. The crucible is filled with substantially pure polycrystalline alpha-alumina and a monocrystalline alpa-alumina tube 38 grown previously by the EFG technique is mounted in holder 38. Tube 38 is cylindrical and was grown so that the c-axis of its crystal lattice extends parallel to its geometric axis. Additonally, tube 38 has an outside diameter identical to the diameter of rod 58 and a wall thickness of about 0.03 inch. Tube 38 is mounted on holder 36 so that it is aligned axially with rod 58. Access to seed holder 36 and the susceptor 40 is achieved by lowering bed 2' away from the furnace enclosure and lowering the seed holder below the bottom end of furnace tube 4. With the bed restored to the position of FIG. 1, cooling water is introducd between the two quartz tubes, and the enclosure is evacuated and filled with argon to a pressure of about one atmosphere which is maintained during the growth period. Then the RF. coil 30 is energized and operated so that the alumina in the crucible is brought to a molten condition (alumina has a melting point in the vicinity of 2050 C.) and the surface 68 reaches a temperature of about 2070 C. As the solid alumina is converted to the melt 48, the melt will infiltrate rod 58 and fill the pores, i.e. cells, of its submerged portion. The infiltrating melt will also rise in the rod by capillary action until substantially all of the cells are fully filled. After affording time for temperature equilibrium to be established, the pulling mechanism is actuated and operated so that the tube 38 is moved into contact with the upper surface 66 of the die assembly and allowed to rest in that position to allow the bottom end of the tube to melt and form film 70 as shown in FIG. 2. After about 60 seconds, the tube is withdrawn vertically at the rate of about 0.1-0.2 inch per minute. As the tube is withdrawn, crystal growth will occur on the seed and the film of melt will begin to spread over the surface 66 due to its affinity with the newly grown material on the tube and the films surface tension. The latter force also causes additional melt t0 flow out of the capillaries and add to the total volume of film.

Since the film functions as a growth pool of melt, as the film spreads out over the surface 66, the growth also expands horizontally. At a pulling speed of 0.l-0.2 inch/ minute growth will propagate vertically throughout the entire horizontal expanse of the film with the result that the growing crystal will also begin to grow radially inward as shown in FIG. 3 until after about 3 minutes it will conform in cross-sectional area and shape to surface 66. As growth continues (see FIG. 4), the growing body will be found to have a circular symmetry with an OD. substantially the same as that of the surface 66 and an ID. substantially the same as the diameter of cavity 68 Growth is continued at the same pulling speed and temperature until the supply of melt in the crucible is substantially exhausted, whereupon the pulling speed is immediately increased to about 1.0 inch/minute, whereby the growing body pulls free of film 70. Thereafter the furnace is cooled and tube 38 retrieved from holder 36. The extension grown on the tube is found to have smooth inner and outer surfaces, and a wall thickness of about .06 inch. The grown crystal is found to be substantially monocrystalline and a crystallographic extension of the crystal lattice of tube 38.

If growth occur on the tube but the film does not immediately begin to spread, steps are taken to force the melt to spread as desired. This can be accomplished by adjusting the temperature of the film or by adjusting the pulling speed. Preferably, the temperature of the surface 66 is held constant and the pulling speed adjusted until spreading of the film is observed.

It is to be noted that after the film has fully covered surface 66, if the operating temperature (as determined by the average temperature of film 70) is held constant close to but slightly above the melting point of the material to be grown, the pulling speed may be varied within limits (depending upon the operating temperature) without any substantial change in the cross-section of the grown crystal. Similarly if the pulling speed is held constant, the operating temperature may be varied substantially (e.g., a change of as much as 15-30 degrees with respect to the melting point of alumina) without any substantial change in the cross-section of the grown crystal.

The fact that the grown crystalline extension has substantially the same shape and size as the surface 66 confirms that the film 70 comprises a growth zone which is substantially isothermal in a direction parallel to surface 66. It is to be noted that the film has a depth in the order of about 0.1 mm. under usual growth conditions and has a vertical temperature gradient. Surface 66 functions substantially as an isothermal heater. Where the tube has a relatively large OD. and wall thickness, it may be necessary to increase the rate of heating slightly so that the temperature of the upper end surface 66 of the die assembly before it is contacted by the tube is greater than that normally required to be maintained for satisfactory growth. This higher temperature offsets the heat sink effect of the tube which may cause the growth pool, i.e., the film of melt, to have a lower average temperature than expected. Unless this heat sink effect is offset by an increase in the rate of heating, the tube may not melt, or the film may not spread rapidly over the surface 66, unless the pulling speed is adjusted to compensate for the heat sink effect.

The procedure described in the foregoing example may be used in growing substantially monocrystalline alumina extensions of other cross-sectional shapes, e.g., rectangular, triangular, square, etc., on seeds of the same or different cross-sections. Thus by using a die assembly with an upper film-supporting surface that is square and has no cavity as shown at 68, it is possible to grow a solid extension of square cross-section onto a round or square tube or rod. If the square film-supporting surface has a cavity that is round or square, the crystal body that is produced will be tubular with a round or square bore. The procedure described in the foregoing example may also be used in growing substantially monocrystalline alumina bodies with die assemblies of the type shown in FIGS. 5, 6, and 8. In growing substantially monocrystalline tubes of alumina with the die of FIG. 6, it is preferred that the die be made of porous molybdenum and the cavity 68 be filled with tin.

As noted above, an important advantage of the invention is that it is applicable to crystalline materials other than alumina. It is not limited to congruently melting materials and, for example, encompasses growth of materials that solidify in cubic, rhombohedral, hexagonal and tetragonal crystal structures, including spinel, ruby, beryllia, barium titanate, yttrium aluminum garnet, lithium niobate, germanium and copper mentioned above. With respect to these and other materials, the process is essentially the same as that described above for alpha-alumina, except that it requires different operating temperatures because of different melting points. Additionally, certain changes may be required in the apparatus, e.g., different crucible materials in order to avoid reaction between the melt and the crucible, different die materials, and (in the case of the die of FIG. 6) different non-wetting materials. For example, spinel (Al O -MgO) is grown with a crucible and die assembly made of molybdenum or iridium; for beryllia, the crucible and die assembly are made of tungsten.

It is to be noted also that the porous member of the die assembly may be made of two or more parts that are secured together to form one unitary die member. This is convenient, for example, when forming a die having a U-shaped film supporting surface for use in growing a crystal body in the form of a U-shaped channel. A further variation is to make only the top section of the die of porous material, with the bottom section being non-porous but having one or more vertical bores of capillary size formed therein for leading melt up to the porous section. The top end of the porous member also may be concave instead of fiat.

Laue X-ray back reflection photographs of crystal growth produced according to the foregoing invention reveals that the crystal growth usually comprises one or two, and in some cases three or four, crystals growing together longitudinally separated by a low angle (usually within 4 of the c-direction) grain boundary. Therefore, for convenience and in the interest of avoiding any suggestion that the crystal growth is polycrystalline in character, it is preferred to describe it as substantially monocrystalline, it being understood this term is intended to embrace a crystalline body that is comprised of a single crystal or two or more crystals, 'e.g., a bicrystal or tricrystal, growing together longitudinally but separated by a relatively small angle (i.e., less than about 4) grain boundary. The same term is used to denote the crystallographic nature of the seed tube.

With respect to the die assembly, it is to be understood that in the following claims the term end surface is intended to cover the effective film-supporting surface of the die, whether made as a single piece or comprising two or more parts.

I claim:

1. Apparatus for growing a substantially monocrystalline body of a first selected material from a thin liquid film of said material comprising a crucible for holding a supply of said material, and a solid forming member in said crucible having an upper end surface for supporting said liquid film, at least the upper end of said member being porous and comprising a network of interconnected open cells forming a plurality of passageways that function as capillaries for said first selected material in molten form, at least some of said passageways opening in said end surface and connecting with openings near the bottom end of said member whereby a melt of said material in said crucible can rise to said surface via said passageways by action of capillary rise, said member being made of a second selected material from the class consisting of metals, metal alloys and ceramics that (1) is wetted by said first material, (2) has a higher melting point than said first material, and (3) will not react with said first selected material at a temperature in the order of the melting point of said first material and means for pulling a substantially monocrystalline body from said liquid film at the upper surface of said forming member 2. Apparatus according to claim 1 wherein said member is a sponge of said second material.

3. Apparatus according to claim 1 wherein said first material is alumina and said second material is molybdenum, tungsten or iridium.

4. Apparatus according to claim 1 wherein the outside edge of said top end surface is claid with a non-porous solid material that is wetted by and does not react with said melt.

5. Apparatus according to claim 1 wherein said top end surface has an inside edge and said inside edge is clad with a non-porous solid material that is wetted by and does not react with said melt.

6. Apparatus according to claim 1 wherein said top end surface has a cavity and said cavity is filled with a material that is not wettable by said melt.

7. Apparatus according to claim 6 wherein said forming member is made of molybdenum and said cavity is filled with tin.

-8. Apparatus according to claim 1 wherein some of said cells are of micropore size and some of said cells are of macropore size.

9. Apparatus according to claim 1 wherein said forming member is made of molybdenum sponge.

10. Apparatus according to claim 1 wherein said forming member is made of a ceramic material.

11. The method of growing a crystalline body of a selected material comprising providing a molten pool of said material about a solid die member that is made of an open celled foam material from the class consisting of metals, metal alloys, and ceramics and has a substantially horizontal top end surface so that the cells of said die member are in communication with said pool and said top end surface is exposed, infiltrating cells at the bottom of said die member with melt from said pool and filling other higher cells with said melt by action of capillary rise, establishing a thin liquid film of said material on said top end surface so that said film substantially fully covers said top end surface and is joined with the melt in said cells and growing and pulling a crystalline body of said material from the full horizontal expanse of said film, and continually feeding melt from said pool to said surface via said open cells at a rate consistent with the rate of crystal growth so as to replenish said film and maintain said crystal growth.

12. Method of growing a substantially monocrystalline body of a first selected material comprising:

surrounding a porous solid member made of a material from the class consisting of metals, metal alloys, and ceramics and having a substantially horizontal top end surface with a liquid pool of said first material so that the said pool is below said top end surface;

infiltrating said member with said liquid and filling the pores of the upper end of said member with said liquid by action of capillary rise; establishing a liquid film of said material on said top end surface; and growing a substantially monocrystalline body of said material from said film; and

feeding additional liquid to said top end surface via said pores to replace the material consumed by crystal growth from said film.

13. Method according to claim 12 wherein said member is cladded with a layer of a solid non-porous material that surrounds and is flush with the edge of said top end surface, and further establishing said liquid film so that it covers said layer.

14. Method according to claim 13 wherein said first material is alumina and said member and said non-porous material are made of molybdenum.

References Cited UNITED STATES PATENTS 3.236,768 2/1966 Litt 210-23 3,243,267 3/1966 Piper 23273 3,338,761 8/1967 Cheney et al. 23273 3,525,594 8/1970 Barrett 23273 3,591,348 7/1971 LaBelle, Jr. 23301 3,627,499 12/1971 LeDuc et al. 23273 3,687,633 8/1972 LaBelle et al. 23273 FOREIGN PATENTS 1,246,828 10/1960 France 23273 NORMAN YUDKOFF, Primary Examiner R. T. FOSTER, Assistant Examiner US. Cl. X.R. 23273 SP