JP5370393B2 - Compound semiconductor single crystal substrate - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a substrate of a group III-V compound semiconductor single crystal having excellent quality. <P>SOLUTION: An n-type gallium arsenide substrate has an average dislocation density of less than 100 cm<SP>-2</SP>and a silicon concentration of 5&times;10<SP>16</SP>cm<SP>-3</SP>or more and less than 5&times;10<SP>17</SP>cm<SP>-3</SP>. A semi-insulating gallium arsenide substrate has an average dislocation density of less than 1,000 cm<SP>-2</SP>, a silicon concentration of less than 5&times;10<SP>15</SP>cm<SP>-3</SP>, and a specific resistance of 1&times;10<SP>3</SP>&Omega;cm or more. An n-type indium phosphide substrate has an average dislocation density of less than 100 cm<SP>-2</SP>, a sulfur concentration of 1&times;10<SP>17</SP>cm<SP>-3</SP>or more and less than 3&times;10<SP>18</SP>cm<SP>-3</SP>, or has an average dislocation density of less than 1,000 cm<SP>-2</SP>and a tin concentration of 1&times;10<SP>17</SP>cm<SP>-3</SP>or more and less than 5&times;10<SP>18</SP>cm<SP>-3</SP>. A semi-insulating indium phosphide substrate has a dislocation density of less than 1,000 cm<SP>-2</SP>and a specific resistance of 1&times;10<SP>3</SP>&Omega;cm or more, and is doped with iron. <P>COPYRIGHT: (C)2011,JPO&amp;INPIT

Description

  The present invention relates to a group III-V compound semiconductor single crystal such as GaAs, InP, GaP, InAs, GaSb, and InSb used in the optoelectronic field such as a light emitting diode (LED) and a laser diode (LD) and the electronic field such as a transistor. Or a substrate of a II-VI group compound semiconductor single crystal such as CdTe or ZnSe.

  Examples of compound semiconductor single crystals include III-V group compound semiconductor single crystals such as gallium arsenide (GaAs), gallium phosphide (GaP), indium arsenide (InAs), indium phosphide (InP), and tellurium II-VI group compound semiconductor single crystals such as cadmium (CdTe) and zinc selenide (ZnSe) can be given.

  These compound semiconductor single crystals have been conventionally used in the horizontal Bridgman method (HB method), the liquid sealing pulling method (LEC method), the vertical Bridgman method (VB method), the vertical temperature gradient solidification method (VGF method), and the VB method. And various industrial methods such as a method using the VGF method together. Among these, the VB method and the VGF method (including methods using them in combination) have attracted particular attention in recent years because they can produce crystals with a low dislocation density that cannot be produced by other production methods.

  In the VB method, for example, as described in Non-Patent Document 1, a crucible containing a compound semiconductor raw material is placed in a furnace having a heating means such as a heater and the raw material is melted, and then the heating means is raised. This is a method for growing a single crystal by lowering the crucible and solidifying the melt from the seed crystal side installed at the lower end of the crucible. A material such as pyrolytic boron nitride (pBN) or quartz is generally used for the crucible. On the other hand, in the VGF method, as described in Patent Document 1, for example, the positional relationship between the heating means and the crucible is fixed, and the temperature profile of the heating means is changed to lower the temperature from the seed crystal side. This is a method for growing a single crystal.

III-V compound semiconductors such as gallium arsenide (GaAs), gallium phosphide (GaP), indium phosphide (InP), indium arsenide (InAs), cadmium telluride (CdTe), zinc selenide (ZnSe), etc. In the single crystal growth of II-VI group compound semiconductors by the VB method or the VGF method, it is necessary to prevent dissociative evaporation of high vapor pressure components. Therefore, the crucible containing the raw material is sealed in a quartz ampule to control the vapor pressure in the ampule, or the surface of the raw material melt in a high pressure stainless steel chamber is made of a liquid such as boron oxide (B ₂ O ₃ ). A method of suppressing dissociation and evaporation of a high vapor pressure component from a raw material melt by covering with a sealing agent and further pressure-sealing with argon gas or nitrogen gas is performed.

Special Table 2002-540051 Japanese Patent Laid-Open No. 5-124877

Akira Hikawa, Semiconductor Research 35, Crystal Growth and Evaluation of Compound Semiconductors 6 (Industry Research Committee, published in 1991) T. Kawase, M. Tatsumi and T. Nishida, Crystal Growth Technology, Edited by H. J. Scheel and T. Fukuda, 2003 John Wiley & Sons, Ltd., p364-365.

  As in the VB method and the VGF method, a compound semiconductor seed crystal and raw material are accommodated in a vertical single crystal growth vessel, and a part of the raw material and seed crystal are melted by heating means to substantially achieve stoichiometry. In the manufacturing method of a single crystal in which a raw material melt adjusted to the composition is prepared and crystal growth is performed from the unmelted portion of the seed crystal toward the raw material melt side, the raw material melt is set to a temperature higher than the melting point, and the melting point temperature position is set. A temperature gradient is formed such that the temperature decreases from the solid-liquid interface formed along the seed crystal toward the seed crystal.

In Non-Patent Document 2, the solidification latent heat generated at the solid-liquid interface is Q _L , the heat flux flowing from the solid-liquid interface to the crystal side is Q _OUT , and the heat flux flowing from the melt side to the solid-liquid interface is Q _IN , when representing the heat flux flowing from the crystal side to the crystal in the liquid near the interface with Q _R, the shape of the solid-liquid interface is dependent on the heat flux balance equation (1), the solid-liquid interface shape of Q _R> 0 It is sometimes described as convex toward the melt side and concave when Q _R <0. In the crystal growth of compound semiconductors, it is known that crystal defects such as lineage (dislocation arrays) occur if the shape of the solid-liquid interface becomes concave. In Non-Patent Document 2, it is preferable to make the shape of the solid-liquid interface as flat as possible. By satisfying the relationship of the following formula (2), the solid-liquid interface is convex toward the melt side (concave Q _L , Q _IN , and Q _OUT need to be adjusted. Also, reducing the solidified latent heat Q _L by reducing the "growth rate is effective to realize a solid-liquid interface of the convex, it has the disadvantage of reducing the productivity of crystals. Therefore, Q _IN By adjusting _QOUT and QOUT, it is preferable to realize a convex solid-liquid interface. " Specifically, it is necessary to reduce Q _IN or increase Q _OUT .
Q _R = Q _OUT − (Q _L + Q _IN ) (1)
Q _R = Q _OUT − (Q _L + Q _IN )> 0 (2)
In the conventional method in which crystal growth is performed while maintaining the temperature of the melt at a temperature higher than the melting point, there is a limit to reducing Q _IN , so that Q _OUT is increased by increasing the temperature gradient on the crystal side, It is necessary to satisfy the relationship of the formula (2). Non-Patent Document 2 shows the heat flux balance in the LEC method. However, in the VB method and the VGF method, only the arrangement of the solid and the liquid is opposite to each other, and the idea regarding the heat flux balance is the same.

  Lowering the dislocation density of compound semiconductor crystals is an important issue in terms of crystal quality. In addition, since the higher the growth rate of the crystal, the lower the manufacturing cost can be. Therefore, an increase in the growth rate is also an important issue in crystal production. However, in the conventional manufacturing method, it was impossible to satisfy the two conditions of lowering the dislocation density and increasing the growth rate at the same time. The reason is considered as follows.

That is, as the crystal growth rate is increased, the amount of solidification latent heat Q _L generated per unit time increases, so that it is necessary to increase Q _OUT in order to grow crystals while maintaining an appropriate solid-liquid interface shape. Become. In order to increase Q _OUT , it is necessary to increase the temperature gradient in the longitudinal direction of the crystal, which causes a large thermal stress inside the crystal. Since the thermal stress generated in the crystal generates or propagates dislocations and increases the dislocation density of the crystal, it is impossible to reduce the dislocation density. Also, since there is a limit to the increase in Q _OUT , it is impossible to increase the crystal growth rate while maintaining an appropriate solid-liquid interface shape. Furthermore, it is considered that such restrictions on dislocation density and growth rate are particularly noticeable in large-diameter crystal growth with a large amount of latent heat generation per unit time. In relation to the growth rate and dislocation density of a GaAs crystal having a diameter of 100 mm, for example, the dislocation density at a growth rate of 4 mm / h is 2,000 to 5,000 cm ⁻² , but the dislocation density at a growth rate of 8 mm / h. Is from 20,000 to 50,000 cm ⁻² . Further, regarding the relationship between the growth rate and dislocation density of a GaAs crystal having a diameter of 150 mm, for example, the dislocation density is 3,000 to 10,000 cm ⁻² when the growth rate is 3 mm / h, but the growth rate is 6 mm / h. The dislocation density is 30,000 to 100,000 cm ^-2 .

  The inventor stores the seed crystal and raw material of the compound semiconductor in a vertical single crystal growth vessel by the VB method or the VGF method, and substantially melts the raw material and a part of the seed crystal by heating means. In a method for producing a single crystal in which a raw material melt adjusted to a stoichiometric composition is prepared and crystal growth is performed from the unmelted portion of the seed crystal toward the raw material melt side, at least a part of the raw material melt is brought into a melt state. It was found that a new solidified phase does not occur in the melt even when the temperature is lower than the melting point while keeping it. Furthermore, when a part of the raw material melt was kept in the melt state and the temperature was lower than the melting point, and crystal growth proceeded from the unmelted portion of the seed crystal toward the melt side, new isolation was found in the melt. It has been found that crystal growth proceeds in one direction from the seed crystal to the top of the melt without generating a solidified phase, and a complete single crystal without a grain boundary can be grown.

  The present inventor can also grow a crystal with a smaller temperature gradient in the direction of the crystal growth axis than before, by lowering the temperature from the melting point while maintaining at least a part of the raw material melt in the melt state. Further, as a result, it was found that crystals with a very low dislocation density that could not be realized by the conventional production method can be grown. Further, it has been found that even when the temperature gradient is reduced in this way, a single crystal can be grown at a faster rate than before. In other words, even if the temperature gradient on the crystal side is reduced and the heat radiation to the crystal side is reduced, the solid-liquid interface shape can be maintained in an appropriate shape, and at a high growth rate without causing crystal defects such as lineage. They found that single crystals could be grown.

  Generally, it is generally considered that if at least a part of the raw material melt substantially adjusted to the stoichiometric composition is brought to a temperature below the melting point, the part below the melting point is solidified. The state where the melt becomes lower than the melting point is called “supercooled state”. The difference between the temperature and the melting point of the melt in the supercooled state is called “the degree of supercooling”. Usually, the degree of supercooling is close to zero, and the solid-liquid interface may be considered to be approximately equal to the melting point temperature. However, the present inventors have found that compound semiconductors, such as III-V group compound semiconductors such as gallium arsenide (GaAs), gallium phosphide (GaP), indium arsenide (InAs), indium phosphide (InP), and tellurium In the single crystal growth of II-VI group compound semiconductors such as cadmium iodide (CdTe) and zinc selenide (ZnSe), it was found that a solidified phase does not easily occur even if the raw material melt portion is supercooled. . Thus, the reason why a new solidified phase does not occur in the melt even when the temperature is lower than the melting point while maintaining the state of the melt is considered as follows.

  In order for the solidified phase to appear in the melt, it is necessary to satisfy the conditions necessary for generating crystal nuclei in the melt. One of the conditions is the temperature of the melt, and the lower the melt temperature, the more easily crystal nuclei are generated. However, even if the embryo, which is the previous stage of the crystal nucleus, is accidentally generated, most of it disappears due to its large surface energy. Even if the raw material melt is in a supercooled state, the fact that a new solidified phase is not easily generated in the melt is considered to be closely related to such a mechanism of crystal nucleus generation. In the case where a solid phase already exists, for example, when a single crystal is grown using a seed crystal, it is not necessary to generate a new crystal nucleus, so that it is considered that a large supercooling state does not occur.

  In the melt, crystal nuclei usually tend to be generated on the inner wall of the container containing the melt. Compared with the case where crystal nuclei are generated three-dimensionally in the melt, the surface energy of the crystal nuclei is smaller when generated two-dimensionally so as to adhere to the vessel wall, and the nuclei are more likely to grow stably. Because. Therefore, it is considered that the degree of supercooling when crystal nuclei are generated on the container wall largely depends on the state of the container wall. The present inventor, when using a pyrolytic boron nitride (pBN) container as a vertical single crystal growth container, it is difficult to generate crystal nuclei on the container wall, and it is easy to realize a large degree of supercooling. It has been found that a greater degree of supercooling can be achieved when a boron oxide coating is interposed between the vessel and the melt. In the experiment conducted by the present inventors, when a pBN container was used, no new solidified phase was generated in the melt even when the temperature of the melt was lowered by 50 ° C. or more from the melting point. In addition, when a boron oxide film is interposed between the pBN container and the raw material melt, a new solidified phase is formed in the melt even if the temperature of the melt is lowered by 100 ° C. or more from the melting point. Did not occur. Thus, the present inventor has found the condition that does not generate a solidified phase even when the raw material melt of the compound semiconductor is supercooled, and has reached the present invention.

Considering the heat flux balance in the vicinity of the solid-liquid interface in the present invention as in Non-Patent Document 2, it can be understood as follows. As described above, in the conventional method, the shape of the solid-liquid interface depends on the heat flux balance of the equation (1), and the relationship of the following equation (2a) obtained by modifying the equation (2) or (2): Q _L , Q _OUT , and Q _IN need to be controlled to satisfy.
Q _OUT > Q _L + Q _IN (2a)
Here, conventionally, the temperature of the melt is higher than the melting point, and since heat flows from the melt side toward the solid-liquid interface, Q _IN > 0. If the crystal growth rate is increased, the value of the solidification latent heat Q _L generated per unit time increases, so the value of Q _OUT needs to be increased. In order to increase the Q _OUT value, it is necessary to increase the temperature gradient in the longitudinal direction of the crystal, so that a large thermal stress is generated inside the crystal and the dislocation density increases.

In the present invention, at least a part of the melt is brought into a supercooled state. Since the temperature of such a supercooled melt is lower than the melting point, it acts to remove heat from the vicinity of the solid-liquid interface. That is, in the method of the present invention, as the heat flux flowing out from the solid-liquid interface, in addition to the heat flux Q _{OUT (SOLID)} > 0 flowing out to the crystal side, the heat flux Q _{OUT (LIQUID)} flowing out to the melt side > 0 exists. Since solidification latent heat Q _L is generated at the solid-liquid interface, a thermal balance as shown in the following formula (3) is established near the solid-liquid interface.
Q _R = Q _{OUT (SOLID)} + Q _{OUT (LIQUID)} – Q _L (3)
Conditions for realizing a stable solid-liquid interface shape are as shown in the following equation (4).
Q _R = Q _{OUT (SOLID)} + Q _{OUT (LIQUID)} -Q _L > 0 (4)
The following equation (4a) obtained by transforming equation (4) is a necessary condition for maintaining an appropriate solid-liquid interface shape.
Q _{OUT (SOLID)} + Q _{OUT (LIQUID)} > Q _L (4a)
As is clear from the comparison between the formula (2a) and the formula (4a), the conventional method only takes the latent heat of solidification to the crystal side, but the method of the present invention not only absorbs the latent heat of solidification to the crystal side. Can also take away. The thermal conductivity of the melt is several times larger than the thermal conductivity of the crystal, and heat can be efficiently transferred to the low temperature part by natural convection or forced convection, so that a much larger heat dissipation effect than that of the crystal side can be obtained. In the present invention, by utilizing such an effective heat dissipation effect on the melt side, the heat flux Q _{OUT (SOLID)} on the crystal side can be kept small, so that it is much smaller than the conventional method. It is considered that crystals can be grown under a temperature gradient. The actual heat flux balance is more complicated and cannot be expressed by simple equations such as equations (1), (2), (2a), (3), (4), and (4a). Qualitatively, it can be explained as above.

  Also, by utilizing such a large heat dissipation effect on the melt side, high-quality crystals can be grown while maintaining an appropriate solid-liquid interface shape even at a large crystal growth rate that could not be considered by conventional common sense. It became possible to make it. Furthermore, it has been clarified that an ultra-large compound semiconductor crystal having a diameter of 200 mm, which was considered impossible by the conventional method, can be reduced in dislocation density and increased in growth rate.

By the way, Patent Document 2 proposes a method for suppressing convection in the melt. The 13th to 17th lines in the paragraph [0011] include “the solid-liquid interface position control heater 1 and the melt. A gap 14 having a dimension l = about 20 mm is provided between the forming heater 2 and a valley T _{L of} temperature distribution (= about 1233 to 1237 ° C.) is formed ”. The conceptual diagram is shown in FIG. As described above, Patent Document 2 has a description as if the temperature of the GaAs melt is lower than the melting point (mp = 1238 ° C.), but this does not mean that. A part of the description of Patent Document 2 is cited below.

“[0003] [Problems to be Solved by the Invention] As described above, in the conventional vertical Bridgman method, the temperature distribution in the high-temperature region 21 above the solid-liquid interface 10 is gradually higher than that in the solid-liquid interface 10. 3 is used, the convection of the melt 6 (indicated by an arrow) affects the position of the solid-liquid interface 10. Accordingly, as shown in FIG. It tends to be concave with respect to the melt direction (arrow a) opposite to 8. As a result, (1) defects are easily taken from the crucible wall and the single crystal yield is poor, and (2) the growth rate must be slowed down. "[0004] The object of the present invention is to provide a buffer zone that prevents the convection of the melt from affecting the solid-liquid interface." By eliminating the disadvantages of the prior art described above, ... ", [0009] [Action] If the gap between the interface heating part and the high-temperature heating part is widened to provide a gap or a cooling means for forming a temperature distribution valley between them, the convection in the melt is blocked vertically. In addition, since the temperature gradient becomes 0 ° C./cm, the solid-liquid interface shape becomes convex in the melt direction, and the single crystal yield is improved. [0010] The distance L between the valley of the temperature distribution and the solid-liquid interface is L ≧ 0.5d because L <0. In the case of 5d, it is not effective because the convection of the melt reaches the solid-liquid interface even if the valley of the temperature distribution is provided, and L ≦ 2d is set when L> 2d. Because the melt height is too high, the temperature scent of the melt in the meantime is kept at 0 ° C / cm. Occurrence of problems such as the formation of new convection in the middle of the temperature valley and the solid-liquid interface, or too low temperature in the temperature valley, causing the melt to solidify in that area. It is preferable that L = about d, which is practically stable .... "
As is clear from the above description, Patent Document 2 does not suggest that the melt is supercooled. Patent Document 2, Paragraph [0010], Lines 8 to 10: "... Probability of occurrence of such a problem that the temperature of the valley of the temperature is lowered too much and the melt is solidified at that part. It is clear that the description of “...” clearly indicates that if the temperature of the melt is lowered too much, the temperature of the melt becomes lower than the melting point, and the melt is solidified at that portion. The probability that such troubles will occur is high. "

Further, _TL in Patent Document 2 is the heater temperature. In FIG. 1, the furnace temperature distribution 18 (FIG. 1B) is shown by the length of the heater, that is, the seed crystal 8. It is clear from the fact that the temperature distribution is also shown in the lower space portion. Thus, even if the temperature of the heater is partly below the melting point, the temperature of the melt does not become below the melting point. Since the temperature of the part where the crucible 4 is installed receives strong radiant heat from the part where the temperature is high, the temperature is generally higher than the melting point.

Regarding the temperature of T _{L in} Patent Document 2, in the case of “L = about d” (the last line in the paragraph [0011]) in Example 1 of the paragraph [0011], the temperature is changed to the 16th line in the paragraph [0011]. Although described as “valley T _L (= about 1233 to 1237 ° C.)”, Example 3 in paragraph [0014] was grown under the same conditions as Example 1 except that “L = about 2d”. As a result, if the temperature of T _L is not set to around 1237 ° C., the melt in the vicinity tends to solidify and it is often difficult to continue the growth, so that T _L is raised to 1238 to 1239 ° C. As a result of the growth, it was found that the convection of the melt became large and the growth conditions became unstable, but a single crystal was obtained when the melt did not solidify at T _L = about 1237 ° C. obtained. is described as ... ", the T _L = about 1237 ℃ even melt Of it it is stated that occurs. Since Patent Document 2, paragraph [0011], line 11 describes “melting point (mp = 1238 ° C.)”, “ _TL = about 1237 ° C.” is only 1 ° C. lower than the melting point. . It is common knowledge of those skilled in the art that when the valley of the temperature distribution of the heater in FIG. 1B is set around 1237 ° C., the temperature of the melt is 1238 ° C. or higher. Such a fact and Patent Document 2, paragraph [0014], lines 2 to 3, “As a result, if the temperature of T _L is not set to around 1237 ° C., the melt in the vicinity tends to solidify and grow. In view of the description that “it was often difficult to continue”, Patent Document 2 describes that if the melt is brought to a temperature below the melting point, the melt is solidified and a single crystal cannot be obtained. Should not be below the melting point. "

On the other hand, in the present invention, a raw material melt in which a seed crystal and a raw material are housed in a vertical single crystal growth vessel and a part of the raw material and the seed crystal are melted by a heating means to be substantially adjusted to a stoichiometric composition. In the method for producing a compound semiconductor single crystal in which crystal growth is performed from the unmelted portion of the seed crystal toward the raw material melt side, at least a part of the raw material melt is kept at a temperature lower than the melting point while keeping the molten state. The crystal growth proceeds from the unmelted portion of the seed crystal toward the melt side. In the present invention, the uppermost part of the raw material melt can be made lower than the melting point, and crystal growth can proceed from the unmelted portion of the seed crystal toward the melt side. In this way, the low-temperature melt cooled at the top of the melt is carried by convection to increase the heat flux Q _{OUT (LIQUID)} flowing out to the melt side, effectively depriving the solidification latent heat Q _L. However, when the melt depth is very large, the melt convection may be disturbed and the crystal growth may become unstable. Therefore, in the present invention, a temperature distribution is formed so that the temperature of the raw material melt is substantially uniform, and solidification proceeds from the unmelted portion of the seed crystal toward the melt side. Thus, by making the temperature of the melt substantially uniform, stable crystal growth can be realized even in a deep melt.

In the method of the present invention, the greater the amount of decrease in the melt temperature from the melting point, that is, the greater the degree of supercooling, the greater the heat flux Q _{OUT (LIQUID)} flowing out to the melt side. A big effect is acquired. Preferably, at least a part of the raw material melt is made 10 ° C. or more, more preferably 30 ° C. or more, more preferably 50 ° C. or more lower than the melting point, and solidification proceeds from the unmelted portion of the seed crystal toward the melt side. .

In this way, while the raw material melt is brought into a supercooled state, the temperature of the melt in contact with the solid is maintained at a temperature higher than the melting point by heating the vicinity of the interface between the solid and the liquid with a heater or the like. It is preferable that the temperature of the melt positioned at the upper part is lower than the melting point, and solidification proceeds from the unmelted portion of the seed crystal toward the melt side. Although latent heat is generated when the melt solidifies, the sum of the heat flux Q _{OUT (SOLID)} flowing out to the crystal side and the heat flux Q _{OUT (LIQUID)} flowing out to the melt side Q _{OUT (SOLID)} + Q _{OUT ( If LIQUID)} is too large, the solidification rate will be too high, and excessive solidification latent heat will be generated in a short time, making the temperature distribution in the melt non-uniform and making the crystal growth unstable. . Therefore, by maintaining the temperature of the melt in contact with the solid at a temperature higher than the melting point by heating the vicinity of the interface between the solid and the liquid with a heater, the temperature distribution in the vicinity of the solid-liquid interface is properly maintained and the solid-liquid interface is stable. Crystals can be grown while maintaining the shape and growth rate.

  Even when the temperature of the outer wall surface of the crucible heated by the heater is equal to or lower than the melting point temperature, latent heat of solidification is generated, so that the temperature of the melt contacting the solid can be kept higher than the melting point. However, in order to stably realize such a temperature distribution, it is preferable to heat the outer wall surface of the crucible to a temperature higher than the melting point. On the other hand, as the volume of the melt having a temperature lower than the melting point is increased, a larger supercooling effect is obtained. Therefore, it is preferable to minimize the region where the crucible outer wall surface is heated to a temperature higher than the melting point. Specifically, the region is preferably 120 mm or less in the vertical direction of the vertical crystal growth vessel, and more preferably 60 mm or less.

  Further, it is preferable that the vertical single crystal growth vessel containing the seed crystal and the raw material is moved downward relative to the heating means to advance the solidification from below to above. In this way, the temperature distribution in the vicinity of the solid-liquid interface can be stably maintained, so that crystals can be grown while maintaining a stable solid-liquid interface shape and growth rate.

  The method of the present invention is particularly effective in the growth of single crystals having a body having a diameter larger than 50 mm. That is, the larger the diameter of the body portion of the single crystal, the larger the amount of latent heat of solidification generated per unit time during growth. Therefore, the method according to the present invention can obtain a large amount of heat dissipation as a whole by dissipating heat to the melt side. Is particularly effective compared to the conventional method. Further, by applying the present invention, the crystal body can be grown at a crystal growth rate (that is, a moving speed of the solid-liquid interface) larger than 5 mm / h, which was considered impossible by the conventional method. it can. Furthermore, in the present invention, a single crystal having good characteristics can be obtained even when the crystal body is grown at a speed higher than 10 mm / h.

  Usually, the cross-sectional area of the seed crystal is about 0.5 to 5% of the body cross-sectional area of the single crystal to be grown. The reason for using a seed crystal having such a small cross-sectional area is to reduce the cost of the seed crystal as much as possible. The seed crystal is usually collected by cutting from a single crystal body. Since a larger number of seed crystals having a smaller diameter can be collected than a seed crystal having a larger diameter, the cost per seed crystal can be reduced. However, when growing a crystal having a thick body by increasing the diameter from such a small seed crystal, the amount of latent heat of solidification generated per unit time increases as the crystal growth proceeds in the increased diameter portion. At this time, in order to grow crystals while maintaining a stable solid-liquid interface shape and growth rate, it is necessary to increase the amount of heat released in accordance with the increase in the amount of latent heat of solidification. Such complicated control of the heat radiation amount is poor in reproducibility, and thus becomes a big problem when producing crystals. If crystal growth is attempted at a speed greater than 5 mm / h, particularly greater than 10 mm / h, the amount of change in time of solidification latent heat becomes very large. Therefore, in the method of the present invention, the seed crystal preferably has a cross-sectional area of 25% or more of the grown single crystal body, more preferably has a cross-sectional area of 50% or more, and has a cross-sectional area of 75% or more. More preferably.

  Further, the inclination angle of the increased diameter portion from the seed crystal to the grown single crystal body portion with respect to the crystal central axis is preferably 30 ° or less, more preferably 22.5 ° or less, and further preferably 15 ° or less. If the inclination angle of the increased diameter portion is reduced in this way, the growth time until reaching the single crystal body where the wafer product can be collected becomes longer, and the proportion of the portion that is less than the wafer product diameter increases. The yield of is reduced. Therefore, if a seed crystal having a large cross-sectional area is used, even when the inclination angle of the increased diameter portion is reduced, the growth time until reaching the single crystal body portion can be shortened, and the proportion of the portion that does not satisfy the wafer product diameter can also be reduced. As described above, by gradually changing the diameter from the seed crystal to the grown crystal body, it is possible to reduce the amount of change in latent heat of solidification, even at a growth rate higher than 5 mm / h or 10 mm / h. It is possible to grow crystals while maintaining a stable solid-liquid interface shape and growth rate.

In the method of the present invention, since the heat flux Q _{OUT (SOLID)} to the crystal side can be kept small, the crystal can be grown under a temperature gradient much smaller than that of the conventional method. More specifically, the temperature gradient in the crystal growth axis direction on the outer wall surface of the crucible, that is, the longitudinal direction is preferably less than 5 ° C./cm, more preferably less than 3.5 ° C./cm, and still more preferably less than 2 ° C./cm. Then, solidification of the grown single crystal body is advanced. In addition, from the start to the end of crystal growth, the temperature of the grown single crystal should be maintained at a temperature higher than 0.84 times, more preferably 0.89 times, more preferably 0.94 times the melting point temperature in absolute temperature. Is preferred. Thus, by reducing the temperature gradient in the growth axis direction on the outer wall surface of the crucible during single crystal body growth, the temperature of the grown single crystal is maintained at a predetermined temperature or higher from the start to the end of crystal growth. By doing so, it is possible to suppress the generation of thermal stress during crystal growth, and in particular, it is possible to reduce the dislocation density of the single crystal body corresponding to the wafer product sampling portion.

  The vertical single crystal growth vessel is preferably made of pyrolytic boron nitride (pBN), and a boron oxide film is preferably interposed between the vessel and the melt. As described above, crystal nuclei are unlikely to form on the wall surface of the pyrolytic boron nitride (pBN) container, and if a boron oxide film is interposed between the container and the melt, Crystal nuclei are less likely to occur. And since it can prevent reliably that a new crystal nucleus is generated when a melt is supercooled, the method of this invention can be performed reliably. In order to insert a boron oxide film between the single crystal growth vessel and the melt, the vessel may contain solid boron oxide together with the seed crystal and the raw material, but it contains the seed crystal and the raw material. Prior to this, it is more effective to form a boron oxide film on at least a portion of the inner wall of the container that is supposed to come into contact with the melt.

  The method of the present invention is particularly effective when applied to the growth of compound semiconductor single crystals such as gallium arsenide (GaAs), indium phosphide (InP), and gallium phosphide (GaP). Generally, since the crystal of these compound semiconductors has a low thermal conductivity, the amount of heat released to the crystal side cannot be increased. In addition, compound semiconductor crystals are more easily plastically deformed than silicon (Si) from which dislocation-free crystals can be obtained relatively easily. Therefore, it is extremely difficult to achieve an increase in the growth rate of the compound semiconductor single crystal and a reduction in dislocation density with the conventional method, and the method of the present invention is effective. In addition, these compound semiconductor melts are suitable as targets for carrying out the method of the present invention because a solidified phase hardly occurs even when the degree of supercooling is increased.

  Further, according to the present invention, a vertical single crystal growth container for containing a compound semiconductor seed crystal and a raw material, and a raw material fusion in which a part of the raw material and the seed crystal are melted and adjusted to a substantially stoichiometric composition. In a compound semiconductor single crystal manufacturing apparatus for growing a crystal from an unmelted portion of a seed crystal toward a raw material melt side. Including at least a second heater disposed below and relatively lower the temperature of the first heater, and lowering the temperature of the first heater to be lower than the temperature of the second heater, so that at least a part of the raw material melt is in a molten state An apparatus for manufacturing a compound semiconductor single crystal is proposed in which the crystal growth is allowed to proceed from the lower side to the upper side while maintaining the temperature at a temperature lower than the melting point. Such an apparatus makes it easy to bring the raw material melt into a supercooled state.

  In such an apparatus, the high-temperature gas in the furnace heated by the high-temperature heater disposed below becomes a rising air flow to increase the temperature in the vicinity of the low-temperature heater disposed above, May cause fluctuations. On the other hand, a low temperature gas in the vicinity of the heater disposed above becomes a descending airflow, which may lower the temperature in the vicinity of the high temperature heater and cause a large temperature fluctuation. As described above, when a heater having a high temperature is arranged below and a heater having a low temperature is arranged above, it may be difficult to realize a desired temperature distribution. In addition, since a large temperature fluctuation occurs due to strong gas convection, it may be difficult to stabilize the temperature of the crystal or the melt. Furthermore, since the upper supercooling region is heated by radiation from the lower high temperature region, it may be difficult to realize supercooling.

  Therefore, at least one heat is provided between the second heater disposed below and set to a high temperature and the first heater disposed above and set to a temperature lower than that of the second heater. It is preferable to arrange a shielding plate or a heat insulating plate. Thus, by providing a heat shielding plate or a heat insulating plate between the first heater and the second heater, thermal interference between the high temperature portion and the low temperature portion due to radiation can be suppressed. Further, since gas convection is also suppressed, temperature fluctuation is reduced, and a supercooled state can be realized with good reproducibility. By disposing at least one heater further above the first heater, the melt temperature can be controlled more precisely. As a result, not only can the supercooled state be realized with high reproducibility, but also the convection of the supercooled melt can be suppressed by controlling the temperature distribution.

  Further, it is preferable to dispose a third heater having a temperature lower than that of the second heater below the second heater, and at least one heat shield is provided between the second heater and the third heater. It is more preferable to arrange a plate or a heat insulating plate. Furthermore, it is more preferable to dispose at least one heater further below the third heater. As described above, the third heater whose temperature is set lower than that of the second heater on the crystal side and at least one heater below the third heater are arranged, and at least 1 is provided between the second heater and the third heater. By arranging two heat shielding plates or heat insulating plates, the temperature distribution on the crystal side can be precisely controlled. As a result, it becomes easy to grow a crystal having a low dislocation density by reducing the thermal stress generated in the crystal.

  The second heater forming the high temperature part is preferably composed of a plurality of sub heaters, and more preferably at least one heat shielding plate or heat insulating plate is arranged between the sub heaters. The temperature distribution of the high temperature portion formed by the second heater greatly affects the solid-liquid interface shape that affects the quality of the crystal. In this way, the second heater that forms the high-temperature part is configured by a plurality of sub-heaters, and the temperature distribution in the high-temperature part is precisely controlled by arranging at least one heat shielding plate or heat insulating plate between the sub-heaters. Easy to do. As a result, the shape of the solid-liquid interface can be precisely controlled, and a greater effect can be obtained with respect to an increase in growth rate and a reduction in dislocation density.

The present invention provides an n-type gallium arsenide substrate having an average dislocation density of less than 100 cm ⁻² and a silicon concentration of 5 × 10 ¹⁶ cm ⁻³ or more and less than 5 × 10 ¹⁷ cm ⁻³ . Silicon is added to the n-type gallium arsenide substrate having electrical conductivity as a dopant for controlling desired electrical characteristics. The n-type gallium arsenide substrate is mainly used for optical devices such as LEDs and LDs. In an optical device, since dislocations adversely affect device characteristics, a low dislocation density is required. If silicon is added to the gallium arsenide crystal, the dislocation density of the gallium arsenide crystal is reduced by the solid solution strengthening action. However, in the case of a low silicon concentration of less than 5 × 10 ¹⁷ cm ^−3, the effect of solid solution strengthening cannot be sufficiently obtained. Conventionally, a low dislocation density gallium arsenide crystal having an average dislocation density of less than 100 cm ⁻² is used. It was difficult to train.

In the present invention, by growing the gallium arsenide crystal in a supercooled state, the average dislocation density is less than 100 cm ⁻² even at a low silicon concentration of less than 5 × 10 ¹⁷ cm ^−3. Under these conditions, an n-type gallium arsenide substrate having a low dislocation density of less than 50 cm ⁻² is obtained. Even for large-diameter substrates having a diameter of 100 mm or more, more than 125 mm or more, which are more difficult to achieve a low dislocation density, with a low silicon concentration of less than 5 × 10 ¹⁷ cm ⁻³ and an average dislocation density of less than 100 cm ⁻² , even better In this case, an n-type gallium arsenide substrate having a very low dislocation density of less than 50 cm ⁻² and most preferably less than 10 cm ⁻² can be obtained.

When growing an indium phosphide crystal under supercooled condition, the average dislocation density is less than 100 cm ⁻² even under a low sulfur concentration of less than 3 × 10 ¹⁸ cm ⁻³ , and 50 cm under more precisely controlled conditions. An n-type indium phosphide substrate having a low dislocation density of less than ⁻² is obtained. Even for large-diameter substrates having a diameter of 100 mm or more, more than 125 mm or more, which are more difficult to achieve a low dislocation density, a low sulfur concentration of less than 3 × 10 ¹⁸ cm ⁻³ and an average dislocation density of less than 100 cm ⁻² are even better. In this case, an n-type indium phosphide substrate having a very low dislocation density of less than 50 cm ⁻² and most preferably less than 10 cm ⁻² is obtained.

  The present inventor cannot expect the crystal obtained by the conventional method by using the low dislocation density n-type gallium arsenide substrate or n-type indium phosphide substrate thus obtained as a substrate for optical devices. It has been found that not only good device characteristics can be obtained, but also high yields can be obtained stably. This is considered to be due not only to the direct influence on the device characteristics due to the low dislocation density, but also to the characteristics inherent to the crystal grown in a supercooled state, particularly microscopic changes.

The present invention further provides a semi-insulating gallium arsenide substrate having an average dislocation density of less than 1000 cm ⁻² , a silicon concentration of less than 5 × 10 ¹⁶ cm ^−3, and a specific resistance of 1 × 10 ³ Ωcm or more. . A semi-insulating gallium arsenide substrate having a specific resistance of 1 × 10 ³ Ωcm or more is generally added with a very small amount of chromium or carbon as a dopant for controlling electrical characteristics. Even when these dopants are added to the gallium arsenide crystal, the dislocation density of the gallium arsenide crystal is not significantly reduced by the solid solution strengthening action as in the case of silicon addition. Therefore, conventionally, it has been difficult to grow a low dislocation density crystal having an average dislocation density in the substrate plane of less than 1000 cm ⁻² .

In the present invention, by growing the gallium arsenide crystal in a supercooled state in the melt, the average dislocation density is less than 1000 cm ⁻² , better is less than 500 cm ⁻² , and the best is less than 300 cm ^−2. Thus, a semi-insulating gallium arsenide substrate having a low dislocation density can be obtained. Further, even in a large-diameter substrate having a diameter of 150 mm or more, further having a diameter of 200 mm or more, for which it is difficult to reduce the dislocation density, the average dislocation density is less than 1000 cm ⁻² , more preferable is less than 500 cm ⁻² , and most preferable is 300 cm ^−. A semi-insulating gallium arsenide substrate having a low dislocation density of less than ² can be obtained.

The present invention further provides a semi-insulating indium phosphide substrate having an average dislocation density of less than 1000 cm ⁻² and a specific resistance of 1 × 10 ³ Ωcm or more. A semi-insulating indium phosphide substrate having a specific resistance of 1 × 10 ³ Ωcm or more is added with a very small amount of iron as a dopant for controlling electrical characteristics. However, even if iron is added to the indium phosphide crystal, the dislocation density of the indium phosphide crystal is not significantly reduced by the solid solution strengthening action as in the case of sulfur addition. Therefore, conventionally, it has been difficult to grow a low dislocation density crystal having an average dislocation density of less than 1000 cm ⁻² .

In the present invention, by growing the indium phosphide crystal under supercooled condition, the average dislocation density is less than 1000 cm ⁻² , even better, less than 500 cm ⁻² , and the best is less than 300 cm ^−2. Thus, an iron-doped semi-insulating indium phosphide substrate having a low dislocation density and a specific resistance of 1 × 10 ³ Ωcm or more can be obtained. Further, even in a large-diameter substrate having a diameter of 150 mm or more, further having a diameter of 200 mm or more, for which it is difficult to reduce the dislocation density, the average dislocation density is less than 1000 cm ⁻² , more preferable is less than 500 cm ⁻² , and the best is 300 cm ⁻ An iron-doped semi-insulating indium phosphide substrate having a low dislocation density of less than ² and a specific resistance of 1 × 10 ³ Ωcm or more can be obtained.

The present inventor uses the low-dislocation-density semi-insulating gallium arsenide substrate and semi-insulating indium phosphide substrate thus obtained, particularly for substrates for optoelectronic devices such as OEICs (optoelectronic integrated circuits). Thus, it has been found that not only good device characteristics that could not be expected from crystals obtained by the conventional method can be obtained, but also high yields can be stably obtained. The OEIC is a device in which an optical device and an electronic device are manufactured on one chip. As described above, dislocations adversely affect the characteristics of optical devices. The reason why good substrate characteristics and high yield can be obtained by the substrate of the present invention is not only the direct influence on the device characteristics due to the low dislocation density, but also the inherent characteristics of crystals grown in a supercooled state of the melt, This is thought to be because, in particular, changes in micro characteristics contribute. Such good device characteristics and high yield were remarkably observed in the substrate to which carbon was added at 1 × 10 ¹⁴ cm ⁻³ or more. As a result of growing the crystal in the supercooled state, the result of lowering the dislocation density, or the change in crystal characteristics due to its combined effect, this contributes to good device characteristics and high yield. Conceivable.

(A) is typical sectional drawing of an example of the apparatus which manufactures the n-type gallium arsenide single crystal in this invention, (B) is a graph which shows an example of the temperature distribution in the crucible outer wall of the apparatus. In the method of Example 1, it is typical sectional drawing which shows an example of the method of accommodating a gallium arsenide polycrystalline raw material, boron oxide, and a dopant (solid silicon) in a pBN crucible. 2 is a schematic front view showing an appearance of a single crystal grown by the method of Example 1 and names of respective parts. FIG. In Example 2, it is typical sectional drawing which shows an example of the apparatus which manufactures a semi-insulating gallium arsenide single crystal. In Example 3, it is typical sectional drawing which shows an example of the apparatus which manufactures an n-type indium phosphide single crystal. In Example 4, it is typical sectional drawing which shows an example of the apparatus which manufactures a semi-insulating indium phosphide single crystal.

Example 1
The method of the present invention is applied to single crystal growth of n-type gallium arsenide (GaAs). 1A and 1B respectively show a schematic cross-sectional view of a single crystal manufacturing apparatus used in the present invention and an example of a temperature distribution graph of the apparatus. For crystal growth, heaters 3 and 4 for forming the upper supercooling region, heaters 5 and 6 for heating the vicinity of the solid-liquid interface, and heaters 7 and 8 for controlling the temperature distribution of the grown crystal Is provided. Further, consideration is given to the thermal insulation between the respective regions by the heat insulating plate 9 provided between the heaters 4 and 5 and the heat insulating plate 10 provided between the heaters 6 and 7. Further, a heat shield plate 11 is provided between the heaters 5 and 6 in order to precisely control the temperature distribution near the solid-liquid interface. The heaters 3 to 8 and the stainless steel chamber 1 are insulated by the heat insulating material 2.

As shown in the schematic cross-sectional view of FIG. 2, a pBN crucible 14 having an inner diameter of about 105 mm, the inner surface of which has been previously coated with a boron oxide film 17 a by oxidation treatment, is placed on the lower shaft 12 in the stainless steel chamber 1. It is installed on the provided crucible base 13. A seed crystal 15 having a diameter of 95 mm is accommodated in the seed crystal accommodating portion at the lower end of the crucible 14, and further about 15 kg of the pre-synthesized gallium arsenide polycrystal raw material 16 and about boron oxide (B ₂ O ₃ ) 17 700 g and further contains solid silicon 20 as a dopant. This solid silicon is obtained by crushing a high-purity silicon wafer so that the silicon concentration at the shoulder portion (boundary between the increased diameter portion and the body portion) of the grown gallium arsenide crystal is 5 × 10 ¹⁶ cm ^−3. The amount is weighed and used.

  Thereafter, the stainless steel chamber 1 is quickly sealed, and the inside of the chamber is evacuated by a vacuum pump. Thereafter, nitrogen gas is introduced into the chamber so that the pressure during crystal growth is 0.5 MPa as a gauge pressure, and heating is performed by increasing the temperature of the heater. In the process of heating and heating, the boron oxide 17 is first softened and melted to completely seal the gallium arsenide polycrystal raw material 16 and the seed crystal 15. Thereafter, the temperature is further raised and the temperature of the heaters 3 to 8 is adjusted, while the lower part of the seed crystal 15 is maintained at a temperature below its melting point, while the polycrystalline raw material 16 and the upper part of the seed crystal 15 are higher than the melting point. To form a raw material melt 18 (see FIG. 1A). By maintaining in this state for a certain period of time, the composition and temperature of the raw material melt are stabilized, and then the temperature of the heaters 3 and 4 is slowly lowered to heat the outer wall surface of the crucible 14 to a temperature equal to or higher than the melting point. Is adjusted to a temperature distribution so that the temperature of the melt is about 30 ° C. lower than the melting point so that the temperature is about 60 mm in the vertical direction (see FIG. 1B). To form. Then, the silicon-doped gallium arsenide single crystal 19 having a diameter of about 105 mm is grown by lowering the lower shaft 12 at a speed of 5.5 mm / h while rotating at 3 rpm (see FIG. 1A). At this time, the temperature gradient in the crystal growth axis direction on the outer wall surface of the crucible 14 is 1.5 ° C./cm.

In FIG. 3, the appearance of the single crystal thus grown and the names of the respective parts are shown. From the shoulder portion of the single crystal, an n-type gallium arsenide substrate having a silicon concentration of 5 × 10 ¹⁶ cm ⁻³ and an average dislocation density of about 90 cm ⁻² can be obtained. Also, an n-type gallium arsenide substrate having a silicon concentration of 1 × 10 ¹⁷ cm ⁻³ and an average dislocation density of about 40 cm ⁻² can be obtained from the tail of the same crystal.

(Example 2)
In Example 2, a single crystal of semi-insulating gallium arsenide (GaAs) is grown using the same apparatus as in Example 1 as shown in FIG. A pBN crucible 34 having an inner diameter of about 155 mm whose inner surface is previously coated with a boron oxide film 17a by oxidation treatment is placed on a crucible base 33 provided on the upper portion of the lower shaft 32 in the stainless steel chamber 21. A seed crystal 35 having a diameter of 145 mm is accommodated in the seed crystal accommodating portion at the lower end of the crucible 34, and further about 30 kg of a pre-synthesized gallium arsenide polycrystal raw material and boron oxide (B ₂ O ₃ ) 37 about 1500 g. And solid carbon (not shown) as a dopant. This solid carbon is used by weighing the amount so that the concentration of carbon in the shoulder portion of the grown gallium arsenide crystal is 5 × 10 ¹⁵ cm ⁻³ .

  Thereafter, the stainless steel chamber 21 is quickly sealed, and the inside of the chamber is evacuated by a vacuum pump. Thereafter, nitrogen gas is introduced into the chamber so that the pressure during crystal growth is 0.5 MPa as a gauge pressure, and heating is performed by increasing the temperature of the heater. In the process of heating and heating, the boron oxide 37 is first softened and melted to completely seal the gallium arsenide polycrystal raw material and the seed crystal 35. Thereafter, the temperature is further increased and the temperature of the heaters 23 to 28 is adjusted, while the lower part of the seed crystal 35 is maintained at a temperature below the melting point, while the polycrystalline raw material and the upper part of the seed crystal 35 are heated to a temperature higher than the melting point. Thus, the raw material melt 38 is formed. Further, by maintaining this state for a certain period of time, the composition and temperature of the raw material melt 38 are stabilized, and then the temperatures of the heaters 23 and 24 are slowly lowered to heat the outer wall surface of the crucible 34 to a temperature equal to or higher than the melting point. The supercooled melt is formed by adjusting the temperature distribution so that the region becomes a region of about 60 mm in the vertical direction and the temperature of the melt is lower than the melting point by about 10 ° C. Thereafter, the semi-insulating gallium arsenide single crystal 39 having a diameter of about 155 mm is grown by lowering the lower shaft 32 at a speed of 8 mm / h while rotating at 3 rpm. In this case, the temperature gradient in the crystal growth axis direction on the outer wall surface of the crucible is 4.5 ° C./cm.

From the gallium arsenide single crystal thus obtained, a low dislocation density substrate having a carbon concentration of 5 × 10 ¹⁵ cm ⁻³ and an average dislocation density of about 900 cm ⁻² can be obtained. Note that the concentration of impurity silicon in this crystal is less than 5 × 10 ¹⁵ cm ⁻³ , and does not contribute to the reduction of dislocation density.

(Example 3)
In Example 3, a single crystal of n-type indium phosphide (InP) is grown using the same apparatus as in Example 1 as shown in FIG. A pBN crucible 54 having an inner diameter of about 105 mm, the inner surface of which has been previously coated with a boron oxide film 17a by oxidation treatment, is placed on a crucible base 53 provided on the upper portion of the lower shaft 52 in a stainless steel chamber 41. A seed crystal 55 having a diameter of 70 mm is accommodated in a seed crystal accommodating portion at the lower end of the crucible 54, and about 10 kg of a pre-synthesized indium phosphide polycrystal raw material and boron oxide (B ₂ O ₃ ) 57 about 700 g. And indium sulfide (not shown) as a dopant. This indium sulfide is used by weighing the amount so that the concentration of sulfur in the shoulder of the grown indium phosphide crystal is 1 × 10 ¹⁸ cm ⁻³ .

  Thereafter, the stainless steel chamber 41 is quickly sealed, and the inside of the chamber is evacuated by a vacuum pump. Thereafter, nitrogen gas is introduced into the chamber so that the pressure during crystal growth is 5 MPa as a gauge pressure, and heating is performed by raising the temperature of the heater. In the process of heating and heating, the boron oxide 57 is first softened and melted to completely seal the indium phosphide polycrystal raw material and the seed crystal 55. Thereafter, the temperature is further raised and the temperature of the heaters 43 to 48 is adjusted, while the lower part of the seed crystal is maintained at a temperature below the melting point, while the polycrystalline raw material and the upper part of the seed crystal 55 are heated to a temperature higher than the melting point. Then, the raw material melt 58 is formed. Further, the composition and temperature of the raw material melt 58 are stabilized by maintaining in this state for a certain period of time, and then the temperature of the heaters 43 and 44 is slowly lowered to heat the crucible outer wall surface to a temperature higher than the melting point. The temperature distribution is adjusted so that the temperature of the melt is about 120 mm in the vertical direction and the temperature of the melt is lower than the melting point by about 40 ° C. to form a supercooled melt. Thereafter, the lower shaft 52 is lowered at a speed of 10.5 mm / h to grow an n-type indium phosphide single crystal 59 having a diameter of about 105 mm. In this case, the temperature gradient in the crystal growth axis direction on the outer wall surface of the crucible 54 is 2.0 ° C./cm.

From the shoulder portion of the indium phosphide single crystal thus obtained, a low dislocation density substrate having a sulfur concentration of 1 × 10 ¹⁸ cm ⁻³ and an average dislocation density of about 80 cm ⁻² can be obtained.

Example 4
In Example 4, a single crystal of semi-insulating indium phosphide (InP) is grown using the same apparatus as in Example 1 as shown in FIG. A pBN crucible 74 having an inner diameter of about 155 mm whose inner surface is previously coated with a boron oxide film 17a by an oxidation treatment is placed on a crucible base 73 provided on the upper portion of the lower shaft 72 in a stainless steel chamber 61. A seed crystal 75 having a diameter of 110 mm is accommodated in the seed crystal accommodating portion at the lower end of the crucible 74, and further about 20 kg of a pre-synthesized indium phosphide polycrystal raw material and about 1500 g of boron oxide (B ₂ O ₃ ) 77. And high-purity iron (not shown) as a dopant. This high-purity iron is used by weighing the amount so that the concentration of iron in the shoulder portion of the grown indium phosphide crystal is 1 × 10 ¹⁶ cm ⁻³ .

  Thereafter, the stainless steel chamber 61 is quickly sealed, and the inside of the chamber is evacuated by a vacuum pump. Thereafter, nitrogen gas is introduced into the chamber so that the pressure during crystal growth is 5 MPa in gauge pressure, and heating is performed by raising the temperature of the heater. In the process of heating and heating, the boron oxide 77 is first softened and melted to completely seal the indium phosphide polycrystal raw material and the seed crystal 75. Thereafter, the temperature is further raised and the temperature of the heaters 63 to 68 is adjusted to keep the lower part of the seed crystal at a temperature below the melting point, while heating the polycrystalline raw material and the upper part of the seed crystal 77 to a temperature higher than the melting point. A raw material melt 78 is formed. Further, the composition and temperature of the raw material melt 78 are stabilized by maintaining this state for a certain period of time, and then the temperature of the heaters 63 and 64 is slowly lowered to heat the crucible outer wall surface to a temperature equal to or higher than the melting point. However, the supercooled melt is formed by adjusting the temperature distribution so that the melt is in a region of about 120 mm in the vertical direction and the temperature of the melt is lower than the melting point by about 50 ° C. Thereafter, the lower shaft 72 is lowered at a speed of 10.5 mm / h to grow a semi-insulating indium phosphide single crystal 79 having a diameter of about 155 mm. In this case, the temperature gradient in the crystal growth axis direction on the outer wall surface of the crucible 74 is 3.0 ° C./cm.

From the shoulder portion of the indium phosphide single crystal thus obtained, a low dislocation density substrate with an iron concentration of 1 × 10 ¹⁶ cm ⁻³ and an average dislocation density of about 600 cm ⁻² is obtained.

DESCRIPTION OF SYMBOLS 1 Stainless steel chamber, 2 Thermal insulation material, 3, 4, 5, 6, 7, 8 Heater, 9, 10 Thermal insulation board, 11 Thermal shielding board, 12 Lower shaft, 13 Crucible stand, 14 pBN crucible, 15 Seed crystal, 16 Gallium arsenide polycrystal raw material, 17 Boron oxide (B ₂ O ₃ ), 17a Boron oxide film, 18 Raw material melt, 19 Gallium arsenide single crystal, 20 Solid silicon, 21 Stainless steel chamber, 22 Thermal insulation, 23, 24, 25, 26, 27, 28 Heater, 29, 30 Heat insulation plate, 31 Heat shield plate, 32 Lower shaft, 33 Crucible base, 34 pBN crucible, 35 Seed crystal, 37 Boron oxide (B ₂ O ₃ ), 38 Raw material melt, 39 Gallium arsenide single crystal, 41 Stainless steel chamber, 42 Heat insulation material, 43, 44, 45, 46, 47, 48 Heater, 49, 50 Heat insulation plate, 51 Heat shield plate, 52 Lower shaft, 53 Crucible Stand, 5 4 pBN crucible, 55 seed crystal, 57 boron oxide (B ₂ O ₃ ), 58 raw material melt, 59 indium phosphide single crystal, 61 stainless steel chamber, 62 heat insulating material, 63, 64, 65, 66, 67, 68 heater, 69, 70 heat insulating plate, 71 heat shielding plate, 72 lower shaft, 73 crucible base, 74 pBN crucible, 75 seed crystal, 77 boron oxide (B ₂ O ₃ ), 78 raw material melt, 79 indium phosphide Single crystal.

Claims (9)

An n-type having an average dislocation density of 40 cm ⁻² or more and less than 100 cm ⁻² , a silicon concentration of 5 × 10 ¹⁶ cm ⁻³ or more and less than 5 × 10 ¹⁷ cm ⁻³ , and a diameter of 125 mm or less. Gallium arsenide substrate. The n-type gallium arsenide substrate according to claim 1, having a diameter of 100 mm or more. It has an average dislocation density of 300 cm ⁻² or more and less than 1000 cm ^−2, a silicon concentration of less than 5 × 10 ¹⁵ cm ⁻³ , a specific resistance of 1 × 10 ³ Ωcm or more, and a diameter of 200 mm or less. Semi-insulating gallium arsenide substrate. The semi-insulating gallium arsenide substrate according to claim 3 , wherein carbon of 1 × 10 ¹⁴ cm ⁻³ or more is added. The semi-insulating gallium arsenide substrate according to claim 3 or 4 , having a diameter of 150 mm or more. An n-type having an average dislocation density of 10 cm ⁻² or more and less than 100 cm ⁻² , a sulfur concentration of 1 × 10 ¹⁷ cm ⁻³ or more and less than 3 × 10 ¹⁸ cm ⁻³ , and a diameter of 125 mm or less. Indium phosphide substrate. Semi-insulating indium phosphide having a dislocation density of 300 cm ⁻² or more and less than 1000 cm ^−2, a specific resistance of 1 × 10 ³ Ωcm or more, a diameter of 100 mm or more and 200 mm or less, and doped with iron substrate. The semi-insulating indium phosphide substrate according to claim 7 , wherein iron is added in an amount of 1 × 10 ¹⁵ cm ⁻³ or more. The semi-insulating indium phosphide substrate according to claim 7 or 8 , which has a diameter of 150 mm or more.

Images