Title of Invention	METHOD OF GROWING SINGLE CRYSTALS FROM MELT
Abstract	A method for producing single crystals by growing from melt, which comprises fusing starting material and pulling a single crystal by crystallization of the melt on a seed crystal with controlled removal of crystallization heat and with using independent heat sources which constitute thermal zones, characterized in that the independent heat sources constitute two equal-sized coaxially arranged thermal zones which make up a united thermal area for the melt and the single crystal being grown, said zones being separated by the melt surface, fusing of the starting material being carried out in two stages: first by heating the upper thermal zone with supplying to the upper heater 30-50% of power reguired for obtaining the melt, until in the upper thermal zone maximum temperature is reached, which secures stable state of the solid phase of the seed crystal; then the remaining power is supplied to the lower thermal zone to the lower heater with maintaining constant temperature of the upper thermal zone till complete melting of the charge; the process of the single crystal enlargement and growing is conducted with controlled lowering of temperature in the upper thermal zone, the amount of power supplied to the lower thermal zone being preserved constant.

Title of Invention

METHOD OF GROWING SINGLE CRYSTALS FROM MELT

Abstract

A method for producing single crystals by growing from melt, which comprises fusing starting material and pulling a single crystal by crystallization of the melt on a seed crystal with controlled removal of crystallization heat and with using independent heat sources which constitute thermal zones, characterized in that the independent heat sources constitute two equal-sized coaxially arranged thermal zones which make up a united thermal area for the melt and the single crystal being grown, said zones being separated by the melt surface, fusing of the starting material being carried out in two stages: first by heating the upper thermal zone with supplying to the upper heater 30-50% of power reguired for obtaining the melt, until in the upper thermal zone maximum temperature is reached, which secures stable state of the solid phase of the seed crystal; then the remaining power is supplied to the lower thermal zone to the lower heater with maintaining constant temperature of the upper thermal zone till complete melting of the charge; the process of the single crystal enlargement and growing is conducted with controlled lowering of temperature in the upper thermal zone, the amount of power supplied to the lower thermal zone being preserved constant.

Full Text	Description METHOD OF GROWING SINGLE CRYSTALS FROM MELT The present invention relates to the technology of growing single crystals from melts on a seed crystal. The technical problem to be solved by the present invention is to provide a universal method of growing single crystals • of various chemical composition, for example, of A2B6 and A2B5 type, and also single crystals of refractory oxides, such as sapphire. Monocrystalline materials of A2B5 and A2B6 type and those based on oxides are used as optical materials. The development of instrumentation on the basis of these materials substantially adds to the demand for them and raises the guality, efficiency and cost price requirements. Known in the art is a method of growing single crystals of leucosapphire from melt on a seed crystal, which comprises the presence of temperature gradients ranging within 0.05 to 1.0°C/mm and the ratio of the deviation of vertical temperature gradients to radial ones > 1, vacuum melting of the starting charge, introducing a seed crystal and pulling a single crystal from the cooling melt (see RU 2056463, C30B 29/20, publ. 1996). The essence of the method consists in determining the seed crystal introducing temperature from the appearance of a 1-3 mm single crystal on the surface of the melt being cooled and in growing a single crystal with the pulling rate being varied stepwise from 0.1 mm/hr at the beginning of crystallization to 1.0 mm/hr in the final stage of the process with simultaneous lowering of the melt temperature with a sped of 25-50°C/hr. The process of growing is terminated by cooling the obtained single crystal with the rate of 25-50°C/hr. Pulling with a small rate in the initial stage of the process makes it possible to provide formation of a regular lattice, to rule out the appearance of dislocations and blocks, and the formation of bubbles. Pulling with a 10-fold increase of the rate in the final stage provides a reduction of the process time. This method contemplates growing a crystal from an "overcooled" melt at the crystallization front. Since the radial temperature gradient at the center of the melt, where the seed crystal is located, is always equal to zero, even a slight lowering of the heater temperature and then of the melt temperature creates an area at the crystallization front, wherein the temperature is lower than the crystallization temperature. For the removal of the crystallization heat evolving in the course of growing the crystal, in the method under discussion the crystal growth rate is lowered and thereby the amount of heat evolved during crystallization is reduced and time is provided for the heat to be removed along the crystal due to the heat conduction of the material of the crystal being grown. This method is disadvantageous in a low performance which rules out its use in the production of large lots of single crystals. A method is known of growing gallium arsenide single crystals from melt on a seed crystal, in which the seed crystal is brought in contact with the melt placed in a crucible under a layer of liquid flux, followed by crystallization of the entire volume of the melt with liquid encapsulation with flux (see RU 2054495, C30B 17/.00, publ. 1996). The method is intended for growing single crystals of gallium arsenide for manufacturing integrated circuit substrates. Therefore the thickness of the melt layer is selected to be equal to the thickness of the substrate. This method cannot be used for growing bulk single crystals of gallium arsenide. A method is known of growing optical single crystals from melt by the Czochralski process with the use of three heaters (high-frequency heating of a platinum crucible with a melt, a heater for heating the crucible bottom, and a controlled active-resistance conical heater for heating the seed crystal, its holder and rod), which comprises melting down starting strongly ground oxides and growing a single crystal from melt on a rotating seed crystal, in which method growing is carried out when a thermal equilibrium has been established and a flat or slightly convex surface of the melt-crystal interface has been reached. For this purpose the entire process is conducted with additional heating of the seed crystal, the seed holder and rod to a temperature exceeding the crystal growing temperature by a value providing the ratio of IR radiation in the melt and solid phase, i.e., λmelt / λcrystal =0.25. In the stage of enlargement of the crystal being pulled heating is not carried out (see DD 290226, A5, C03B 15/22, publ. 1991). The essence of the method consists in that a temperature field is created, with which the temperature of the seed crystal with the holder and rod is equal to the temperature ~Tmelt of the crystal being grown. Seeding, enlargement and further growth of the crystal is based on the difference in the heat transmission of IR radiation from the crystallization front through the crystal and through the melt. Since the transmission coefficient of the crystal is greater than that of the melt, the crystal, naturally, has a somewhat lower temperature than the melt. As a result, crystal growing on the seed takes place. This thermodynamic equilibrium is maintained automatically throughout the crystal growth period by additional heating of the seed, seed holder and rod. This method has been adopted as the most relevant prior art. Said method suffers from a number of essential disadvantages which do not allow its use for a considerable range of materials grown by Czochralski or Kiropulos methods: 1. The method cannot be used for decomposable materials, e.g., A3B5, in which the vapor pressure of one of the components at Tmelt reaches 40 atm and over. 2. The method cannot be used for vaporable materials, e.g., A2B6, in which the vapor pressure of both components reaches 3 atm and over at Tmelt of the compounds. 3. The method cannot be used for a number of materials having a high plasticity at Tmelt, because there takes place plastic deformation of the growing crystal as such under its own weight (e.g., α-Al2O3) , when plastic deformation is observed already at the temperature of 1600°C. 4. The method cannot be used for a number of materials having small values of melt overcooling (for instance, for CdTe ∆Tover = 1°C), when fluctuations of the heater temperature as small as ± 0.5°C lead either to melting of the seed crystal or to spontaneous crystallization of the melt in the crucible. 5. The method cannot be used for a number of materials in which the emissivity factor of the melt and of the crystal, as well as the absorption of IR rays from the heater are close, e.g., in the case of Ge, Si, InSb, etc. 6. The difference and amount of dark heat passed through the melt and the crystal are incommensurably smaller than the evolving crystallization heat. For example, α-Al2O3 = 255 cal/g, therefore the method can be realized only at very small crystallization rates, i.e., it is industrially inapplicable. The technical result of the claimed invention is its universality with regard to the material of the single crystal to be grown, enhanced performance, and improved structural finish of resulting single crystals by ruling out melt overcooling in the coursed of growing. The technical result is attained by that in the method of growing single crystals from melt, which comprises fusing the starting material and pulling a single crystal by crystallization of the melt on a seed crystal with controlled removal of the crystallization heat and using independent heating sources constituting thermal zones, according to the invention, independent heating sources constitute two equal- sized coaxially arranged thermal zones, which make up a united thermal area for the melt and the single crystal being grown and are separated by the melt surface, fusing of the starting material being carried out in two stages: first by heating the upper thermal zone with supplying to the upper heater 30-50% of power required for obtaining the melt, until in the upper thermal zone maximum temperature is reached, which secures stable state of the solid phase of the seed crystal; then the remaining power is supplied to the lower thermal zone to the lower heater with maintaining constant temperature of the upper thermal zone till complete melting of the charge; the process of the single crystal enlargement and growing is conducted with controlled lowering of temperature in the upper thermal zone, the amount of power supplied to the lower thermal zone being preserved constant. Besides, the removal of crystallization heat in the stage of single crystal* enlargement and growing is carried out with the rate of single crystal crystallization, calculated from the formula: in the crystal growing zone, deg./cm; Cp is the specific heat of the crystal, cal/gdeg.; The essence of the invention is as follows. For growing an "ideal" single crystal in terms of its structure, in the known methods it is necessary to find "the golden mean" between the temperature gradient created by t° of the melt and t° in the zone of the growing crystal for removing the crystallization heat. In RU 2056463, for removing crystallization heat, the crystallization rate is decreased, so as to provide for the heat removal along the crystal at the expense of the heat conduction of the seed material. Or, else, the power supply to the heater in the zone of the crucible with the melt is decreased, the melt temperature being thereby lowered (RU 2054495). The use of these techniques has negative aspects. Lowering the crystallization rate sharply reduces the process efficiency. Lowering the melt temperature at the expense of lowering the heating power leads to overcooling the melt at the crystallization front and, as a consequence, to the origination of structural defects (low-angle boundaries, polycrystalline structure). In the claimed invention the evolving crystallization heat is removed along the crystal by increasing the axial temperature gradient in the zone of growing crystal from its minimal value. In the present invention a principally novel technique is used for attaining this effect. In the reaction zone, where the starting material and the seed crystal with a holder and rod are located, two equal- sized thermal zones are created by two independent heaters, these zones being disposed one above the other and constituting a united thermal area. For fusing the starting charge, first the upper thermal zone is heated by supplying to the upper heater a part (30-50%) of the power required for fusing the starting material. Such amount of power makes it possible to heat the upper zone to the maximum temperature which is capable of preserving the seed crystal in practically stable solid-phase state (Tcrit) . Thus, for example, for A3B5 compounds this is practical absence of dissociation, for A2B6 compounds this is the absence of evaporation, for refractory oxides α-Al2O3 this is the absence of plastic deformation. Then the remaining power is supplied to the lower heater for heating the lower thermal zone. The temperature of the upper heater, equal to Tcrit, is maintained constant till complete melting of the starting charge and attaining dynamic equilibrium between the liquid phase (melt) and the solid phase (seed crystal). The dynamic equilibrium having been reached, stabilization of the power of the lower heater is effected. The temperature difference between Tmelt of the melt and Tcrit creates minimal axial gradients above the melt and in the melt in the united thermal area constituted by the upper and lower thermal zones. The enlargement and growing of the single crystal are carried out with varying the axial temperature gradient above the melt by lowering the power supplied to the upper heater and thereby lowering the temperature of the upper thermal zone. Such temperature lowering is effected while the power supplied to the lower heater remains preserved. Growing of the crystal is achieved owing to the removal the crystallization heat, performed by increasing the temperature gradient above the melt by lowering the temperature of the upper heater. The crystallization rate (g/sec.) is calculated from formula (1). As the temperature of the upper heater lowers and the temperature of the upper thermal zone lowers accordingly, the axial temperature gradient increases, and the crystal grows. Concurrently therewith, the amount of power supplied to the lower heater being preserved throughout the process of growing there takes place lowering of the crystal-melt temperature from the more heated body to the less heated body according to the scheme: lower heater → crucible → melt periphery→ melt centers crystal → upper heater. The temperature of the melt lowers via the crystallization zone in proportion with the lowering of the temperature of the upper heater, precluding the melt overcooling at the crystallization front. Therefore the crystal being grown always grows only toward the area of the more heated melt, the heat removal always goes through the center of the melt in the direction of the growing crystal. In all the known methods lowering of the crystal-melt temperature goes in an opposite direction. Carrying out the process by the claimed techniques ensures the absence of low-angle boundaries in the grown crystals and a low dislocation density. Further, it is very important that the process efficiency grows in proportion to the lowering of the temperature of the upper heater, since overcooling of the melt is absent till the moment the upfper heater is switched off. The claimed techniques characterize a fundamentally new single crystal growing technology, according to which single crystals are obtained not from "overcooled" melts but from "superheated" melts, because the heat flow always goes along the above-described pathway. The united thermal area constituted by the equal-sized thermal zones disposed one above the other and separated by the melt surface comprises the zone of the crucible zoned and the zone of the to-be crystal. Melting of the charge is effected by the total supply of power to both thermal zones in such a manner that lowering of the power supplied to one of the thermal zones should lead to the crystallization of the melt in the crucible. Creation of the united thermaJ area determines the creation of united axial and radial temperature gradients minimum possible for a particular material. Since in the course of crystal enlargement and .growth variation of the temperature gradients does not lead to overcooling of the melt, the method makes it possible to grow single crystals in which the proper value of the melt overcooling temperature can be within the range of from 70°C to ~ 0°C. Furthermore, the creation of the united thermal area constituted by the equal-sized thermal zones makes it possible to cool the grown crystal down to room temperature under isothermal conditions by leveling the temperatures in the thermal zones by way of simultaneous lowering of power of the upper and lower heaters. Since in carrying out the method overcooling of the melt at the crystallization front is absent, using this method it is possible to grow single crystals from materials in which the faculty of the melt for overcooling is close to zero and which up till now could not be grown by the Czochralski method (CdTe, a-Al2O3 in the direction [0001]; GaAs in the direction [001] or had definite structural variances. Exemplary Embodiment of the Method Power is supplied to the resistance heater disposed above the crucible with a charge, and the temperature is brought close to T critical for the given crystallizing material. Critical temperatures for materials are such temperatures, at which there occur irreversible and uncontrolled processes on the surface of the solid phase of the crystal: processes of dissociation, evaporation, plastic deformation, etc., when further practical application of the crystallization process makes no sense. Thus, for instance, the temperature at which there is observed noticeable dissociation of growing GaP crystal above the flux is - 1300°C when Tmelt - 1467°C; the temperature at which there is observed noticeable evaporation of growing CdTe crystal above the flux is ~ 700°C when Tmelt = 1092°C; the temperature at which there is observed noticeable plastic deformation of growing α-Al2O3 crystal is ~ 1600°C, and plastic deformation of Si crystal is > 1100°C when Tmelt = 1420°C, etc. After the temperature of the upper resistance heater becomes close to critical, stabilization of the temperature of the upper thermal zone is carried out. The temperature sensing unit is a thermocouple mounted in the upper part of the upper heater for minimizing the influence of the lower resistance heater on the thermocouple readings. After the entire mass of the inner outfit of the furnace is heated, power is supplied to the lower heater which serves for fusing the charge in the crucible. The temperature of the upper heater remains constant and equal to ≤ Tcrit. As the charge becomes gradually melted and the melt temperature stabilized, a critical (minimal) axial temperature gradient or the given material is created automatically. Then seeding is carried out. As stable dynamic equilibrium is reached gradually between the solid phase (the seed crystal) and the liquid phase (the melt) (the presence of a constant bright halo around the seed), the temperature of the upper heater is lowered, thereby increasing the axial, and, consequently, the radial temperature gradients of the seed crystal and of the melt, respectively, with the power of the lower heater stabilized. In this way conditions are created, under which the crystal grows from the superheated melt. The areas of melt overheating at the crystallization front are absent throughout the process of crystallization .Since the crystal grows continuously from the "superheated" but not from the "overcooled" melt, this makes it possible not only to eliminate the undesirable formation of structural defects on the crystallization front, but gives an opportunity to obtain by Czochralski method such materials which could not be produced or were difficult to produce heretofore, e.g., CdTe, GaAs single crystals in the [100] orientation, Al2O3 in the [0001] orientation, and the like. When the whole melt has crystallized in the form of crystal grown on the seed, the temperature of the lower heater is lowered down to the temperature of the upper heater, and then the power values of the heaters are lowered synchronously till room temperature is reached, isothermal conditions being thus created for removing residual thermal stresses within the entire bulk of the single crystal. So, when growing an α-Al2O3 single crystal from a "superheated" melt, after preparing the smelting chamber for the process (loading the charge, mounting round-shaped or profiled heaters, mounting an appropriate seed crystal, oriented in accordance with the heater profile, degassing and creating a definite atmosphere in the smelting chamber, and other operations), the upper heater is switched on and brought to a temperature according to the thermocouple in the range of Tcrit ~ 1600°C. Tcrit having been reached, all the outfit in the smelting chamber is heated for several hours till the heat exchange becomes stable. Then the lower heater is switched on, its power is adjusted to bring the heater to the temperature equal to Tmelt Al2O3 = 2050°C and maintained to the stable state which is determined visually from the behavior of the melt surface. By manipulating the power of the lower heater, seeding is carried out. The stable state of the system is determined from a stable halo around the seed crystal on the liquid-solid phase interface. After the dynamic equilibrium is reached, stabilization of the power of the lower heater is carried out, and stability is preserved throughout the crystallization process . By lowering the temperature in accordance with the thermocouple readings with an accuracy of ± 0.5°C enlargement of the crystal is carried out. Removal of the crystallization heat occurs along the crystal as a result of the temperature gradient increasing with time. Since the lower heater is power-stabilized, the temperature of the crucible with the melt will also lower accordingly, but it will always remain higher than Tmelt, i.e., the melt at the crystallization front is always "superheated". Maintaining the diameter of the growing crystal is accomplished in accordance with the increment of its weight per unit of time, i.e., in accordance with the crystallization rate which is precalculated from relation (1), and downloaded into the program. On completion of the single crystal growing process, the power of the lower heater is lowered down to leveling the temperatures in the crucible and in the zone of the grown single crystal. The crystal is cooled to room temperature under isothermal conditions by simultaneous lowering of the power values of the two heaters. An example of particular embodiment of the method: The meaning of the formula ∆m/∆τ - as mentioned and already referred to above - consists in the determination of the temperature conditions for growing single crystals from "overheated" melts by Amosov method with maximum permissible rates. When using this formula, it should be taken into account that in the process of enlargement the radius of the crystal being grown increases continuously with the course of time: Rl = Ro + ∆T/A, R2 = Rl + ∆T/A, etc. As the radius of the crystal increases, approaching the radius of the crucible (the stage of completion of the enlargement), owing to an increase of the radial gradient at the walls of the crucible ∆T/A → 0 , and further lowering of the temperature on the upper heater practically does not lead to an increase of the crystal radius. The crystal continues growing owing to an increase of the axial temperature gradient along the crystal. Starting parameters: We substitute these values into formula (1) of the crystallization rate (∆m/∆x) . Then we set the lowering of the temperature of the upper heater, e.g., ∆T = -5°C. In the stage of enlargement the increment of the crystal weight will be Am/Ax =8.02 g/hr. The process of the crystal enlargement is continued by further lowering of the temperature of the upper heater till a prescribed diameter of the crystal, e.g., 120 mm, is obtained. The process of growing is carried out at a constant value of the crystal weight increment per unit of time and with lowering the temperature of the upper heater. With the diameter of the crystal being grown 120 mm, ∆m/∆τ is 1149 g/hr. The process of growing a 30 kg crystal lasts for ~ 26 hours. Thermal stresses, low-angle boundaries and structural defects are completely absent in the grown crystal. So, the claimed method makes it possible to grow with a high efficiency bulky single crystals without limitations as to their chemical composition and noted for a perfect structure. WE CLAIM: 1. A method for producing single crystals by growing from melt, which comprises fusing starting material and pulling a single crystal by crystallization of the melt on a seed crystal with controlled removal of crystallization heat and with using independent heat sources which constitute thermal zones, characterized in that the independent heat sources constitute two equal-sized coaxially arranged thermal zones which make up a united thermal area for the melt and the single crystal being grown, said zones being separated by the melt surface, fusing of the starting material being carried out in two stages: first by heating the upper thermal zone with supplying to the upper heater 30-50% of power required for obtaining the melt, until in the upper thermal zone maximum temperature is reached, which secures stable state of the solid phase of the seed crystal; then the remaining power is supplied to the lower thermal zone to the lower heater with maintaining constant temperature of the upper thermal zone till complete melting of the charge; the process of the single crystal enlargement and growing is conducted with controlled lowering of temperature in the upper thermal zone, the amount of power supplied to the lower thermal zone being preserved constant. 2. The method according to claim 1, characterized in that the removal of crystallization heat in the stage of single crystal enlargement and growing is carried out with the crystallization rate calculated from the formula: ∆m is the weight of the crystal, g; ∆τ is the time of weight increment (∆m), sec; Tmelt is the melting temperature of the starting material, deg.; Tcrit is the maximum temperature of the stable state of the solid phase of the seed crystal, deg.; ∆T is the temperature variation on the upper heater during the process, deg.; ∆Hmaelt is the .specific heat of melting, cal/g; p is the pressure, const.; R is the radius of the crystal, cm; is the radial temperature gradient at the crystallization front, deg./cm; is the initial axial temperature gradient in the crystal growing zone, deg./cm; Cp is the specific heat of the crystal", cal/g deg.; λ is the heat conduction of the crystal, cal/cm sec deg. A method for producing single crystals by growing from melt, which comprises fusing starting material and pulling a single crystal by crystallization of the melt on a seed crystal with controlled removal of crystallization heat and with using independent heat sources which constitute thermal zones, characterized in that the independent heat sources constitute two equal-sized coaxially arranged thermal zones which make up a united thermal area for the melt and the single crystal being grown, said zones being separated by the melt surface, fusing of the starting material being carried out in two stages: first by heating the upper thermal zone with supplying to the upper heater 30-50% of power reguired for obtaining the melt, until in the upper thermal zone maximum temperature is reached, which secures stable state of the solid phase of the seed crystal; then the remaining power is supplied to the lower thermal zone to the lower heater with maintaining constant temperature of the upper thermal zone till complete melting of the charge; the process of the single crystal enlargement and growing is conducted with controlled lowering of temperature in the upper thermal zone, the amount of power supplied to the lower thermal zone being preserved constant.

Full Text

Description
METHOD OF GROWING SINGLE CRYSTALS FROM MELT
The present invention relates to the technology of growing
single crystals from melts on a seed crystal.
The technical problem to be solved by the present invention
is to provide a universal method of growing single crystals •
of various chemical composition, for example, of A2B6 and A2B5
type, and also single crystals of refractory oxides, such as
sapphire.
Monocrystalline materials of A2B5 and A2B6 type and those
based on oxides are used as optical materials. The
development of instrumentation on the basis of these
materials substantially adds to the demand for them and
raises the guality, efficiency and cost price requirements.
Known in the art is a method of growing single crystals of
leucosapphire from melt on a seed crystal, which comprises
the presence of temperature gradients ranging within 0.05 to
1.0°C/mm and the ratio of the deviation of vertical
temperature gradients to radial ones > 1, vacuum melting of
the starting charge, introducing a seed crystal and pulling a
single crystal from the cooling melt (see RU 2056463, C30B
29/20, publ. 1996).
The essence of the method consists in determining the seed
crystal introducing temperature from the appearance of a 1-3
mm single crystal on the surface of the melt being cooled and
in growing a single crystal with the pulling rate being
varied stepwise from 0.1 mm/hr at the beginning of
crystallization to 1.0 mm/hr in the final stage of the

process with simultaneous lowering of the melt temperature
with a sped of 25-50°C/hr. The process of growing is
terminated by cooling the obtained single crystal with the
rate of 25-50°C/hr. Pulling with a small rate in the initial
stage of the process makes it possible to provide formation
of a regular lattice, to rule out the appearance of
dislocations and blocks, and the formation of bubbles.
Pulling with a 10-fold increase of the rate in the final
stage provides a reduction of the process time.
This method contemplates growing a crystal from an
"overcooled" melt at the crystallization front. Since the
radial temperature gradient at the center of the melt, where
the seed crystal is located, is always equal to zero, even a
slight lowering of the heater temperature and then of the
melt temperature creates an area at the crystallization
front, wherein the temperature is lower than the
crystallization temperature. For the removal of the
crystallization heat evolving in the course of growing the
crystal, in the method under discussion the crystal growth
rate is lowered and thereby the amount of heat evolved
during crystallization is reduced and time is provided for
the heat to be removed along the crystal due to the heat
conduction of the material of the crystal being grown.
This method is disadvantageous in a low performance which
rules out its use in the production of large lots of single
crystals.
A method is known of growing gallium arsenide single crystals
from melt on a seed crystal, in which the seed crystal is
brought in contact with the melt placed in a crucible under a
layer of liquid flux, followed by crystallization of the
entire volume of the melt with liquid encapsulation with flux

(see RU 2054495, C30B 17/.00, publ. 1996). The method is
intended for growing single crystals of gallium arsenide for
manufacturing integrated circuit substrates. Therefore the
thickness of the melt layer is selected to be equal to the
thickness of the substrate. This method cannot be used for
growing bulk single crystals of gallium arsenide.
A method is known of growing optical single crystals from
melt by the Czochralski process with the use of three heaters
(high-frequency heating of a platinum crucible with a melt, a
heater for heating the crucible bottom, and a controlled
active-resistance conical heater for heating the seed
crystal, its holder and rod), which comprises melting down
starting strongly ground oxides and growing a single crystal
from melt on a rotating seed crystal, in which method growing
is carried out when a thermal equilibrium has been
established and a flat or slightly convex surface of the
melt-crystal interface has been reached. For this purpose the
entire process is conducted with additional heating of the
seed crystal, the seed holder and rod to a temperature
exceeding the crystal growing temperature by a value
providing the ratio of IR radiation in the melt and solid
phase, i.e., λmelt / λcrystal =0.25. In the stage of
enlargement of the crystal being pulled heating is not
carried out (see DD 290226, A5, C03B 15/22, publ. 1991).
The essence of the method consists in that a temperature
field is created, with which the temperature of the seed
crystal with the holder and rod is equal to the temperature
~Tmelt of the crystal being grown. Seeding, enlargement and
further growth of the crystal is based on the difference in
the heat transmission of IR radiation from the
crystallization front through the crystal and through the
melt. Since the transmission coefficient of the crystal is

greater than that of the melt, the crystal, naturally, has a
somewhat lower temperature than the melt. As a result,
crystal growing on the seed takes place. This thermodynamic
equilibrium is maintained automatically throughout the
crystal growth period by additional heating of the seed, seed
holder and rod. This method has been adopted as the most
relevant prior art.
Said method suffers from a number of essential disadvantages
which do not allow its use for a considerable range of
materials grown by Czochralski or Kiropulos methods:
1. The method cannot be used for decomposable materials,
e.g., A3B5, in which the vapor pressure of one of the
components at Tmelt reaches 40 atm and over.
2. The method cannot be used for vaporable materials,
e.g., A2B6, in which the vapor pressure of both components
reaches 3 atm and over at Tmelt of the compounds.
3. The method cannot be used for a number of materials
having a high plasticity at Tmelt, because there takes place
plastic deformation of the growing crystal as such under its
own weight (e.g., α-Al2O3) , when plastic deformation is
observed already at the temperature of 1600°C.
4. The method cannot be used for a number of materials
having small values of melt overcooling (for instance, for
CdTe ∆Tover = 1°C), when fluctuations of the heater
temperature as small as ± 0.5°C lead either to melting of the
seed crystal or to spontaneous crystallization of the melt in
the crucible.

5. The method cannot be used for a number of materials
in which the emissivity factor of the melt and of the
crystal, as well as the absorption of IR rays from the heater
are close, e.g., in the case of Ge, Si, InSb, etc.
6. The difference and amount of dark heat passed through
the melt and the crystal are incommensurably smaller than the
evolving crystallization heat. For example, α-Al2O3 = 255
cal/g, therefore the method can be realized only at very
small crystallization rates, i.e., it is industrially
inapplicable.
The technical result of the claimed invention is its
universality with regard to the material of the single
crystal to be grown, enhanced performance, and improved
structural finish of resulting single crystals by ruling out
melt overcooling in the coursed of growing.
The technical result is attained by that in the method of
growing single crystals from melt, which comprises fusing the
starting material and pulling a single crystal by
crystallization of the melt on a seed crystal with controlled
removal of the crystallization heat and using independent
heating sources constituting thermal zones, according to the
invention, independent heating sources constitute two equal-
sized coaxially arranged thermal zones, which make up a
united thermal area for the melt and the single crystal being
grown and are separated by the melt surface, fusing of the
starting material being carried out in two stages: first by
heating the upper thermal zone with supplying to the upper
heater 30-50% of power required for obtaining the melt, until
in the upper thermal zone maximum temperature is reached,
which secures stable state of the solid phase of the seed
crystal; then the remaining power is supplied to the lower

thermal zone to the lower heater with maintaining constant
temperature of the upper thermal zone till complete melting
of the charge; the process of the single crystal enlargement
and growing is conducted with controlled lowering of
temperature in the upper thermal zone, the amount of power
supplied to the lower thermal zone being preserved constant.
Besides, the removal of crystallization heat in the stage of
single crystal* enlargement and growing is carried out with
the rate of single crystal crystallization, calculated from
the formula:

in the crystal growing zone, deg./cm;
Cp is the specific heat of the crystal, cal/gdeg.;

The essence of the invention is as follows. For growing an
"ideal" single crystal in terms of its structure, in the
known methods it is necessary to find "the golden mean"
between the temperature gradient created by t° of the melt
and t° in the zone of the growing crystal for removing the
crystallization heat.
In RU 2056463, for removing crystallization heat, the
crystallization rate is decreased, so as to provide for the
heat removal along the crystal at the expense of the heat
conduction of the seed material.
Or, else, the power supply to the heater in the zone of the
crucible with the melt is decreased, the melt temperature
being thereby lowered (RU 2054495).
The use of these techniques has negative aspects. Lowering
the crystallization rate sharply reduces the process
efficiency. Lowering the melt temperature at the expense of
lowering the heating power leads to overcooling the melt at
the crystallization front and, as a consequence, to the
origination of structural defects (low-angle boundaries,
polycrystalline structure).
In the claimed invention the evolving crystallization heat is
removed along the crystal by increasing the axial temperature
gradient in the zone of growing crystal from its minimal
value.
In the present invention a principally novel technique is
used for attaining this effect.

In the reaction zone, where the starting material and the
seed crystal with a holder and rod are located, two equal-
sized thermal zones are created by two independent heaters,
these zones being disposed one above the other and
constituting a united thermal area. For fusing the starting
charge, first the upper thermal zone is heated by supplying
to the upper heater a part (30-50%) of the power required for
fusing the starting material. Such amount of power makes it
possible to heat the upper zone to the maximum temperature
which is capable of preserving the seed crystal in
practically stable solid-phase state (Tcrit) . Thus, for
example, for A3B5 compounds this is practical absence of
dissociation, for A2B6 compounds this is the absence of
evaporation, for refractory oxides α-Al2O3 this is the
absence of plastic deformation. Then the remaining power is
supplied to the lower heater for heating the lower thermal
zone. The temperature of the upper heater, equal to Tcrit, is
maintained constant till complete melting of the starting
charge and attaining dynamic equilibrium between the liquid
phase (melt) and the solid phase (seed crystal).
The dynamic equilibrium having been reached, stabilization of
the power of the lower heater is effected.
The temperature difference between Tmelt of the melt and Tcrit
creates minimal axial gradients above the melt and in the
melt in the united thermal area constituted by the upper and
lower thermal zones. The enlargement and growing of the
single crystal are carried out with varying the axial
temperature gradient above the melt by lowering the power
supplied to the upper heater and thereby lowering the
temperature of the upper thermal zone. Such temperature

lowering is effected while the power supplied to the lower
heater remains preserved.
Growing of the crystal is achieved owing to the removal the
crystallization heat, performed by increasing the temperature
gradient above the melt by lowering the temperature of the
upper heater. The crystallization rate (g/sec.) is calculated
from formula (1).
As the temperature of the upper heater lowers and the
temperature of the upper thermal zone lowers accordingly, the
axial temperature gradient increases, and the crystal grows.
Concurrently therewith, the amount of power supplied to the
lower heater being preserved throughout the process of
growing there takes place lowering of the crystal-melt
temperature from the more heated body to the less heated body
according to the scheme: lower heater → crucible → melt
periphery→ melt centers crystal → upper heater.
The temperature of the melt lowers via the crystallization
zone in proportion with the lowering of the temperature of
the upper heater, precluding the melt overcooling at the
crystallization front. Therefore the crystal being grown
always grows only toward the area of the more heated melt,
the heat removal always goes through the center of the melt
in the direction of the growing crystal.
In all the known methods lowering of the crystal-melt
temperature goes in an opposite direction.
Carrying out the process by the claimed techniques ensures
the absence of low-angle boundaries in the grown crystals and
a low dislocation density. Further, it is very important that

the process efficiency grows in proportion to the lowering of
the temperature of the upper heater, since overcooling of the
melt is absent till the moment the upfper heater is switched
off.
The claimed techniques characterize a fundamentally new
single crystal growing technology, according to which single
crystals are obtained not from "overcooled" melts but from
"superheated" melts, because the heat flow always goes along
the above-described pathway.
The united thermal area constituted by the equal-sized
thermal zones disposed one above the other and separated by
the melt surface comprises the zone of the crucible zoned and
the zone of the to-be crystal. Melting of the charge is
effected by the total supply of power to both thermal zones
in such a manner that lowering of the power supplied to one
of the thermal zones should lead to the crystallization of
the melt in the crucible. Creation of the united thermaJ area
determines the creation of united axial and radial
temperature gradients minimum possible for a particular
material. Since in the course of crystal enlargement and
.growth variation of the temperature gradients does not lead
to overcooling of the melt, the method makes it possible to
grow single crystals in which the proper value of the melt
overcooling temperature can be within the range of from 70°C
to ~ 0°C.
Furthermore, the creation of the united thermal area
constituted by the equal-sized thermal zones makes it
possible to cool the grown crystal down to room temperature
under isothermal conditions by leveling the temperatures in
the thermal zones by way of simultaneous lowering of power
of the upper and lower heaters.

Since in carrying out the method overcooling of the melt at
the crystallization front is absent, using this method it is
possible to grow single crystals from materials in which the
faculty of the melt for overcooling is close to zero and
which up till now could not be grown by the Czochralski
method (CdTe, a-Al2O3 in the direction [0001]; GaAs in the
direction [001] or had definite structural variances.
Exemplary Embodiment of the Method
Power is supplied to the resistance heater disposed above the
crucible with a charge, and the temperature is brought close
to T critical for the given crystallizing material.
Critical temperatures for materials are such temperatures, at
which there occur irreversible and uncontrolled processes on
the surface of the solid phase of the crystal: processes of
dissociation, evaporation, plastic deformation, etc., when
further practical application of the crystallization process
makes no sense. Thus, for instance, the temperature at which
there is observed noticeable dissociation of growing GaP
crystal above the flux is - 1300°C when Tmelt - 1467°C; the
temperature at which there is observed noticeable evaporation
of growing CdTe crystal above the flux is ~ 700°C when Tmelt =
1092°C; the temperature at which there is observed noticeable
plastic deformation of growing α-Al2O3 crystal is ~ 1600°C,
and plastic deformation of Si crystal is > 1100°C when Tmelt =
1420°C, etc.
After the temperature of the upper resistance heater becomes
close to critical, stabilization of the temperature of the
upper thermal zone is carried out. The temperature sensing
unit is a thermocouple mounted in the upper part of the upper

heater for minimizing the influence of the lower resistance
heater on the thermocouple readings.
After the entire mass of the inner outfit of the furnace is
heated, power is supplied to the lower heater which serves
for fusing the charge in the crucible. The temperature of the
upper heater remains constant and equal to ≤ Tcrit. As the
charge becomes gradually melted and the melt temperature
stabilized, a critical (minimal) axial temperature gradient
or the given material is created automatically. Then seeding
is carried out. As stable dynamic equilibrium is reached
gradually between the solid phase (the seed crystal) and the
liquid phase (the melt) (the presence of a constant bright
halo around the seed), the temperature of the upper heater is
lowered, thereby increasing the axial, and, consequently, the
radial temperature gradients of the seed crystal and of the
melt, respectively, with the power of the lower heater
stabilized.
In this way conditions are created, under which the crystal
grows from the superheated melt. The areas of melt
overheating at the crystallization front are absent
throughout the process of crystallization .Since the crystal
grows continuously from the "superheated" but not from the
"overcooled" melt, this makes it possible not only to
eliminate the undesirable formation of structural defects on
the crystallization front, but gives an opportunity to obtain
by Czochralski method such materials which could not be
produced or were difficult to produce heretofore, e.g., CdTe,
GaAs single crystals in the [100] orientation, Al2O3 in the
[0001] orientation, and the like.
When the whole melt has crystallized in the form of crystal
grown on the seed, the temperature of the lower heater is

lowered down to the temperature of the upper heater, and then
the power values of the heaters are lowered synchronously
till room temperature is reached, isothermal conditions being
thus created for removing residual thermal stresses within
the entire bulk of the single crystal.
So, when growing an α-Al2O3 single crystal from a
"superheated" melt, after preparing the smelting chamber for
the process (loading the charge, mounting round-shaped or
profiled heaters, mounting an appropriate seed crystal,
oriented in accordance with the heater profile, degassing and
creating a definite atmosphere in the smelting chamber, and
other operations), the upper heater is switched on and
brought to a temperature according to the thermocouple in the
range of Tcrit ~ 1600°C. Tcrit having been reached, all the
outfit in the smelting chamber is heated for several hours
till the heat exchange becomes stable. Then the lower heater
is switched on, its power is adjusted to bring the heater to
the temperature equal to Tmelt Al2O3 = 2050°C and maintained to
the stable state which is determined visually from the
behavior of the melt surface. By manipulating the power of
the lower heater, seeding is carried out. The stable state of
the system is determined from a stable halo around the seed
crystal on the liquid-solid phase interface. After the
dynamic equilibrium is reached, stabilization of the power of
the lower heater is carried out, and stability is preserved
throughout the crystallization process . By lowering the
temperature in accordance with the thermocouple readings with
an accuracy of ± 0.5°C enlargement of the crystal is carried
out. Removal of the crystallization heat occurs along the
crystal as a result of the temperature gradient increasing
with time.

Since the lower heater is power-stabilized, the temperature
of the crucible with the melt will also lower accordingly,
but it will always remain higher than Tmelt, i.e., the melt at
the crystallization front is always "superheated".
Maintaining the diameter of the growing crystal is
accomplished in accordance with the increment of its weight
per unit of time, i.e., in accordance with the
crystallization rate which is precalculated from relation
(1), and downloaded into the program. On completion of the
single crystal growing process, the power of the lower heater
is lowered down to leveling the temperatures in the crucible
and in the zone of the grown single crystal. The crystal is
cooled to room temperature under isothermal conditions by
simultaneous lowering of the power values of the two heaters.
An example of particular embodiment of the method:
The meaning of the formula ∆m/∆τ - as mentioned and already
referred to above - consists in the determination of the
temperature conditions for growing single crystals from
"overheated" melts by Amosov method with maximum permissible
rates. When using this formula, it should be taken into
account that in the process of enlargement the radius of the
crystal being grown increases continuously with the course of
time: Rl = Ro + ∆T/A, R2 = Rl + ∆T/A, etc.
As the radius of the crystal increases, approaching the
radius of the crucible (the stage of completion of the
enlargement), owing to an increase of the radial gradient at
the walls of the crucible ∆T/A → 0 , and further lowering of
the temperature on the upper heater practically does not lead
to an increase of the crystal radius. The crystal continues
growing owing to an increase of the axial temperature
gradient along the crystal.

Starting parameters:

We substitute these values into formula (1) of the
crystallization rate (∆m/∆x) . Then we set the lowering of the
temperature of the upper heater, e.g., ∆T = -5°C. In the
stage of enlargement the increment of the crystal weight will
be Am/Ax =8.02 g/hr.
The process of the crystal enlargement is continued by
further lowering of the temperature of the upper heater till
a prescribed diameter of the crystal, e.g., 120 mm, is
obtained.
The process of growing is carried out at a constant value of
the crystal weight increment per unit of time and with
lowering the temperature of the upper heater. With the
diameter of the crystal being grown 120 mm, ∆m/∆τ is 1149
g/hr.
The process of growing a 30 kg crystal lasts for ~ 26 hours.

Thermal stresses, low-angle boundaries and structural defects
are completely absent in the grown crystal.
So, the claimed method makes it possible to grow with a high
efficiency bulky single crystals without limitations as to
their chemical composition and noted for a perfect structure.

WE CLAIM:
1. A method for producing single crystals by growing from
melt, which comprises fusing starting material and pulling a
single crystal by crystallization of the melt on a seed
crystal with controlled removal of crystallization heat and
with using independent heat sources which constitute thermal
zones, characterized in that the independent heat sources
constitute two equal-sized coaxially arranged thermal zones
which make up a united thermal area for the melt and the
single crystal being grown, said zones being separated by the
melt surface, fusing of the starting material being carried
out in two stages: first by heating the upper thermal zone
with supplying to the upper heater 30-50% of power required
for obtaining the melt, until in the upper thermal zone
maximum temperature is reached, which secures stable state of
the solid phase of the seed crystal; then the remaining power
is supplied to the lower thermal zone to the lower heater
with maintaining constant temperature of the upper thermal
zone till complete melting of the charge; the process of the
single crystal enlargement and growing is conducted with
controlled lowering of temperature in the upper thermal zone,
the amount of power supplied to the lower thermal zone being
preserved constant.
2. The method according to claim 1, characterized in that the
removal of crystallization heat in the stage of single
crystal enlargement and growing is carried out with the
crystallization rate calculated from the formula:

∆m is the weight of the crystal, g;

∆τ is the time of weight increment (∆m), sec;
Tmelt is the melting temperature of the starting
material, deg.;
Tcrit is the maximum temperature of the stable state of
the solid phase of the seed crystal, deg.;
∆T is the temperature variation on the upper heater
during the process, deg.;
∆Hmaelt is the .specific heat of melting, cal/g;
p is the pressure, const.;
R is the radius of the crystal, cm;
is the radial temperature gradient at the
crystallization front, deg./cm;
is the initial axial temperature gradient
in the crystal growing zone, deg./cm;
Cp is the specific heat of the crystal", cal/g deg.;
λ is the heat conduction of the crystal, cal/cm sec deg.

A method for producing single crystals by growing from
melt, which comprises fusing starting material and pulling a
single crystal by crystallization of the melt on a seed
crystal with controlled removal of crystallization heat and
with using independent heat sources which constitute thermal
zones, characterized in that the independent heat sources
constitute two equal-sized coaxially arranged thermal zones
which make up a united thermal area for the melt and the
single crystal being grown, said zones being separated by the
melt surface, fusing of the starting material being carried
out in two stages: first by heating the upper thermal zone
with supplying to the upper heater 30-50% of power reguired
for obtaining the melt, until in the upper thermal zone
maximum temperature is reached, which secures stable state of
the solid phase of the seed crystal; then the remaining power
is supplied to the lower thermal zone to the lower heater
with maintaining constant temperature of the upper thermal
zone till complete melting of the charge; the process of the
single crystal enlargement and growing is conducted with
controlled lowering of temperature in the upper thermal zone,
the amount of power supplied to the lower thermal zone being
preserved constant.

METHOD OF GROWING SINGLE CRYSTALS FROM MELT

Documents:

Inventors:

PCT Conventions: