JP2786534B2 - Crystal formation method - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for forming a crystal, and more particularly to a method for forming a crystal in which a single crystal is selectively grown on a substrate having a surface having a low nucleation density. About.

The crystal forming method of the present invention is suitably used for electronic devices such as semiconductor integrated circuits, optical integrated circuits, and magnetic circuits, optical devices, magnetic devices, piezoelectric devices, surface acoustic devices, and the like.

[Prior Art] Conventionally, a single crystal film used for a semiconductor element, an optical element, or the like has been formed by epitaxially growing a single crystal on a single crystal substrate.

For example, on a Si single crystal substrate (silicon wafer)
It is known that Si, Ge, GaAs, etc. are epitaxially grown from a liquid phase, a gas phase, or a solid phase, and that single crystals such as GaAs, GaAlAs are epitaxially grown on a GaAs single crystal substrate. ing.

Using the semiconductor film formed in this manner, a semiconductor element and an integrated circuit, a light emitting element such as a semiconductor laser and an LED, and the like are manufactured.

In recent years, research and development of ultra-high-speed transistors using two-dimensional electron gas and super-lattice devices using quantum wells have been actively pursued.
MBE (molecular beam epitaxy) and MOCVD using ultra-high vacuum
(High-precision epitaxial technology such as metal organic chemical vapor deposition).

In such epitaxial growth on a single crystal substrate,
It is necessary to match the lattice constant and the coefficient of thermal expansion between the single crystal material of the base and the epitaxial growth layer.

For example, Si on sapphire, an insulator single crystal substrate,
Although it is possible to grow a single crystal film epitaxially, crystal lattice defects at the interface due to a shift in the lattice constant and diffusion of aluminum, which is a component of sapphire, into the epitaxial layer, etc., pose problems in application to electronic devices and circuits. Has become.

Thus, it can be seen that the conventional method for forming a single crystal film by epitaxial growth largely depends on the base material. Mathews and others are exploring combinations of substrate materials and epitaxial growth layers (EPITAXIAL GROWTH.Aoade
mic Press, New York, 1975 edited by JWMathews).

Also, the size of the substrate is currently about 6 inches for a Si wafer, and the enlargement of GaAs and sapphire substrates is further delayed.

Further, since the single crystal substrate has a high manufacturing cost, the cost per chip increases.

As described above, in order to form a single crystal layer on which a high-quality device can be manufactured by the conventional method, there is a problem that the type of the base material is limited to an extremely narrow range.

On the other hand, in recent years, research and development of a three-dimensional integrated circuit that achieves high integration and multifunctionality by laminating semiconductor elements in a normal direction of a base has been actively performed.

In addition, research and development of large-area semiconductor devices such as solar cells and liquid crystal pixel switching transistors in which elements are arranged in an array on inexpensive glass are becoming active every year.

What is common to both of them is that a technique of forming a semiconductor film on an amorphous insulator and forming an electronic element such as a transistor thereon is required. In particular, a technique for forming a high-quality single crystal semiconductor over an amorphous insulator is desired.

Generally, when a thin film is deposited on an amorphous insulator substrate such as SiO ₂ , the lack of long-range order of the substrate material results in
The crystalline structure of the deposited film is amorphous or polycrystalline. (Here, an amorphous film is one in which the short-range order of the nearest atom is preserved, but there is no longer long-range order.) And a polycrystalline film is a collection of single crystal grains having no specific crystal orientation separated by grain boundaries).

For example, when Si is formed on SiO ₂ by a CVD method, amorphous silicon is formed at a deposition temperature of about 600 ° C. or lower, and polycrystalline with a grain size of several hundred to several thousand で あ れ ば at a higher temperature. It becomes silicon. However, the grain size of polycrystalline silicon greatly changes depending on the forming conditions.

Furthermore, by melting and solidifying the amorphous or polycrystalline film with an energy beam such as a laser or a rod heater,
A polycrystalline film having a large grain size of about a micron or millimeter has been obtained (Single-Crystal silicon on non-si
ngle-crystal insulators.Journal of crystal Growth
vo1.63, No.3, October, 1983 edited by GWCullen).

When a transistor is formed on the thin film of each crystal structure formed in this way and the electron mobility is measured from the characteristics, amorphous silicon has a particle size of ~ 0.1 cm ² / V The polycrystalline silicon has a mobility of 1 to 10 cm ² / V · sec, and the polycrystalline silicon having a large grain size by fusion and solidification has the same mobility as that of the single crystal silicon.

From this result, it can be seen that there is a large difference in the electrical characteristics between the element formed in the single crystal region in the crystal grain and the element formed over the grain boundary.

That is, the deposited film on the amorphous has an amorphous or polycrystalline structure, and the device manufactured there has much lower performance than the device manufactured on the single crystal layer.
Therefore, as a use, a simple switching element,
Limited to solar cells, photoelectric conversion elements, etc.

[Problems to be Solved by the Invention] As described above, since the long-range order does not exist on the surface of the amorphous substrate unlike the single-crystal surface and only the short-range order is maintained, The structure of the thin film of (1) is polycrystalline in which the positions of the grain boundaries are disordered at best, and the amorphous substrate surface has not only long-range order but also the crystal orientation (normal direction of the substrate and in-plane orientation). Since there is no specified anisotropy, it was impossible to control the crystal orientation of the upper layer.

That is, a problem in depositing a single crystal on an amorphous substrate is to establish a technique for controlling a grain boundary position and controlling a crystal orientation.

In the following, a description will be given of a conventional technique relating to control of the position of a grain boundary and control of a crystal orientation and its problems.

As for the control of the grain boundary position, it has been shown that the grain boundary position can be determined by artificially defining the nucleation position in advance (Japanese Patent Laid-Open No. 63-107016).
elective Nucleation based Epitaxy) (T.Yonehara, Y.Nishigaki, H.Mizutani, S.Kondoh, K.Yam
agata, T.Noma and T.Ishikawa, Applied Physics Letter
s vo1.12, pp.1231,1988). This technology uses Si ₃ N ₄ on SiO ₂
Are localized, and the nucleation site is formed. A single crystal of Si grows and collides with a crystal grown from an adjacent site to form a grain boundary and determine a grain position.

In addition, instead of the "nucleus" spontaneously generated at the nucleation position, a non-single-crystal original seed material is patterned in advance, and this is transformed into a single crystal by utilizing the aggregation phenomenon. A technique for forming a “seed crystal” was disclosed (Japanese Patent Application Laid-Open No. 1-132117).

Plane orientation of these crystals (crystal orientation perpendicular to substrate)
Is determined by factors such as stabilization of interfacial energy with the amorphous interface that becomes a substrate at the time of nucleation or aggregation of original seeds, stabilization of surface energy, relaxation of internal stress, etc.
It is extremely difficult to obtain a completely single orientation, and a plurality of orientations are often mixed in addition to the main surface orientation. In particular, the in-plane crystal orientation is not determined solely because the nucleation plane or the aggregation plane is amorphous and has no anisotropy.

On the other hand, in 1987, HISmith showed for the first time that it was possible to control the crystal orientation of KCl deposited on an amorphous substrate surface by artificially imparting anisotropy by lithography to the surface of the amorphous substrate, and named it Graphepitaxy ( H.
I. Smith and DCFlanders, Applied Physics Letters v
o1.32, pp.349,1978) (HISmith USPatent No.4,33
3,792,1982).

After that, the grain growth of Ge thin film (T.Yonehara, HISmith, C.
V. Thompson and JEPalmer, Applied Physics Letters
vo1.45, pp.631,1984), initial growth of Sn (LSDarken and
DHLowndere, Applied Physics Letters vo1.40, pp.95
4, 1987), it was confirmed that the artificial relief pattern on the substrate surface affected the crystal orientation.

However, KCl and Sn were found to have an effect on the orientation of each of the separated crystals in the initial stage of their deposition in Graphepitaxy. For continuous layers, crystals were grown by laser annealing after depositing Si (MWCeis, DAFla
nders and HISmith, Applied Physics Letters vol.3
5, pp71,1979) and solid phase growth of Ga (T.Yonehara, HISmit)
h, CVThompson and JEPalmer, Applied Physics Lett
ers vol.45, pp631, 1984).

However, even in the case of Si and Ge, although the orientation is controlled to some extent, crystal groups are arranged in a mosaic shape, and furthermore, the crystals have grain boundaries with crystals having slightly different crystal orientations. The position was disordered, and a single crystal was not obtained uniformly over a large area.

This is because, in addition to the fact that the three-dimensional crystal orientations of the individual crystals do not completely match, the surface relay pattern has an uncontrolled nucleation position.

It is an object of the present invention to provide a method for forming a crystal in which the position of a grain boundary is controlled and a single crystal having a plane orientation and an in-plane orientation is grown.

[Means for Solving the Problems] The method of forming a crystal according to the present invention comprises the steps of:
All the side walls and the bottom surface are arranged in a rectangular parallelepiped or cubic shape in contact with and surrounded by the insulator, and the original seeds are arranged so that the volume thereof is small enough to singly aggregate. The heat treatment is performed to produce a single crystal seed crystal having a controlled plane orientation and in-plane orientation, and a crystal growth treatment is performed to selectively grow the single crystal starting from the seed crystal. I do.

 Hereinafter, the principle of the present invention will be described.

As described above, the plane orientation of the aggregated crystal is determined by minimizing the free surface energy of the crystal and the interfacial energy with the substrate as the driving force. For example, when polycrystalline Si is agglomerated on SiO ₂ , the surface energy of Si is minimized on the (111) plane, and the plane orientation is attempted to be aligned to (111). On the other hand, Si-SiO ₂
Since the interface energy of (100) is stable, the interfacial energy tries to be equal to (100). Other polycrystalline Si
It is generally (110), depending on its deposition temperature.
Since it has an orientation, factors in its initial state are also added. As a result, Si agglomerated on SiO ₂ is relatively strongly oriented to (111), and (110) exists at about 1/2 to 1/3 of the intensity of (111) in terms of X-ray diffraction reflection intensity. I do. In addition, weak peaks such as (100), (311), (331), and (422) are observed. The reason that the (111) orientation is dominant is that the agglomerated Si becomes hemispherical crystals, whose surface area is much larger than the interface area, and that the surface energy is stable due to the factors that determine the orientation. Is considered to be the strongest factor.

However, even if the material on the surface of the substrate is changed or the conditions for aggregation are changed, it is extremely difficult to completely control one orientation, because many factors influence as described above. Furthermore, it is almost impossible to control the in-plane orientation by agglomeration on a two-dimensional plane.

Therefore, in the present invention, the inventors first consider a configuration in which stabilization of interface energy is the main factor for determining the orientation in order to control one orientation, and then consider a configuration in which the in-plane orientation is also controlled. Was. It is the configuration of the present invention that satisfies these configurations at one time, that is, the side and bottom surfaces of the cuboid or cubic original seed are surrounded by contact with the insulator, and are sufficient to be singly aggregated. It is obtained by forming a small volume and aggregating it.

Specifically, it is sufficient to consider an original seed as shown in FIG. 1 (A). However, this cuboid (or cubic) is assumed to be surrounded by all the remaining surfaces in contact with the insulator except for one surface (the upper surface of the xy plane).

This means that the stabilization of the interfacial energy at the time of aggregation has been affected only by the bottom surface of the original seed and the insulating surface of the base, that is, only one surface. In other words, since five planes contribute to stabilization of interfacial energy, and all five planes are orthogonal to adjacent planes,
It has the advantage that it can act as a plane equivalent to the stable orientation structure of crystals formed by aggregation. At the same time, since the exposed area is smaller than before, the effect of reducing the influence of the structural change due to the stabilization of the surface energy at the time of aggregation is accompanied. As a result, the control of orientation (including in-plane), such as that seen in graphoepitaxy or the example of Ge grain growth that occurs in a surface relay structure, is reproduced in a system of aggregation of a single original seed. It is done.

Hereinafter, specific embodiments will be described with reference to FIGS. 1 (A) to (D) and FIGS. 2 (A) to (D).

(Example of embodiment) First, the shape and size (size) of the most important original seed in the present invention will be described.

The shape of the original seed must be a rectangular parallelepiped or a cube as already described, but strictly speaking, the object of the present invention can be achieved even if it is not a perfect rectangular parallelepiped or a cube. Conversely, assuming that there is a complete cuboid original seed, even if it is surrounded by an insulator and agglomerated, the single crystal seed crystal can be identified with no deviation of several seconds in the direction perpendicular to the insulator surface. Orientation is difficult. If crystals grown from such a seed crystal collide with each other, a grain boundary is necessarily generated there. Therefore, in the present application, the position of the grain boundary is controlled, and it suffices that the position is within a range that does not cause variation in device characteristics due to variation in orientation during device fabrication. Specifically, the normal photolinography process and the pattern edge "drow" generated in the etching process, deviation from the design value due to over-etching,
That is, an angle within several degrees is an allowable range.

Further, “a rectangular or cubic original seed whose side wall and bottom surface are in contact with and surrounded by an insulator” is, for example, the first seed.
The shape is as shown in FIGS. (B), (C) and (D). FIGS. 1B and 1C show a so-called “embedded type”. This is a type in which a part of the surface of the insulator substrate 2 is patterned in advance into a square or a rectangle and etched to form a rectangular parallelepiped or cubic space, in which original seeds are embedded. In this case, the original seed 1 to be embedded has the exposed surface flush with the surface of the insulator as shown in FIG. 1B, that is, the entire surface is flat as shown in FIG. 1C. May be located at a position lower than the surface of the insulator, that is, the height of the original seed (the value of c shown in FIG. 1 (A)) may be smaller than the etching depth. FIG. 4D shows a “sidewall forming type”. This is done by first patterning the seeds into a rectangular parallelepiped or cube on an insulating substrate or a substrate whose surface layer is made of insulating parts, and then forming an insulating material only on the side walls, leaving the top surface, and surrounding it. It is of the type to do.

In any case, the material of the wall surface is not necessarily
It does not have to match with the material of. For example, after forming a rectangular parallelepiped or cubic space in the substrate in the “embedded type”, an insulating film may be formed in this space, and the surface of the insulating film may be used as a wall surface of the insulator.

Next, the size of the original seed will be described. In FIG. 1 (A), the size of the original seed is preferably a, b ≦ 2 μm (a = b may be satisfied), more preferably 1 μm or less, and most preferably about 0.5 μm. This is because when the sizes of a and b exceed 2 μm, the original seeds are not aggregated into a single body but are divided into a plurality and aggregated. However, whether or not the film is divided largely depends on the thickness of the film, that is, the value of c. When the aspect ratio (aspect ratio) between a or b and c is large, the film is easily divided. The optimal value of a and b is 0.5 μm
The reason is that if it is designed to be smaller than 0.5 μm, it becomes difficult to apply a general semiconductor photo process. On the other hand, the thickness of the original seed (= c) is preferably 0.05 μm or more and 0.5 μm or less, more preferably 0.1 μm or more.
0.4 μm or less. The optimum value is determined by the size of a and b. For the value of c, in fact, the thinner the thinner, the easier the agglomeration occurs. The values of a and b must also be reduced, which makes it difficult to apply a general photo process and reduces the influence of the side wall for controlling the in-plane orientation as described above. Occurs.
When the value of c is larger than 0.5 μm, aggregation hardly occurs. Therefore, the optimal value of the thickness of the original seed is 0.5 ≧ c ≧
Determined within the range of 0.05 (μm).

The material of the original seed to be agglomerated next may be a semiconductor element such as Si, Ge, Sn or the like, a metal such as Au, Ag, Cu, Pt, Pd, or an alloy, a compound, or a mixture.

Insulator comprising a non-nucleating surface SiO _2, Si _x H _y, amorphous insulating material such as SiON is used. Also, the composition of the insulator on the non-nucleus forming surface and the insulator on the portion in contact with the original seed do not necessarily have to be the same.

Next, FIGS. 2 (A) to 2 (A) show a process chart of an embodiment of the present invention.
This will be described with reference to FIG.

First, as shown in FIG. 2 (A), a minute concave portion of the insulator substrate 22 or a substrate whose surface layer is covered with an insulator.
The substances to be the original seeds 21a and 21b are embedded in 23a and 23b, respectively. The recesses 23a and 23b are cuboids or cubes. Note that, as described above, the side walls of the original seeds 21a and 21b may be in contact with and surrounded by the insulator as shown in FIG.

As described above, the material of the original seeds 21a and 21b is Si, Ge, or the like, but may be polycrystalline or amorphous.

Next, as shown in FIG. 2 (B), the original seeds 21a and 21b are heat-treated in a hydrogen atmosphere to cause agglomeration, and
Single-crystal seed crystals 24a and 24b having uniform in-plane orientation are formed. The condition of the heat treatment depends on the material of the original seed and its volume, but when the original seed material is a semiconductor element such as Si or Ge, the temperature is usually about 700 to 1100 ° C. The pressure may be normal pressure, but aggregation is more likely to occur when the pressure is reduced (several Torr to 200 Torr). In addition, if the original seed is doped with a large amount of impurities (phosphorus, boric acid, arsenic, etc.), there is an effect of lowering the temperature at which aggregation starts.

Next, as shown in FIG. 2 (C), crystals are selectively grown around the aggregated single crystal seed crystal. At this time, the growth
This is performed by a CVD method or the like. For example, when growing a Si crystal, a silane-based source gas such as SiH ₄ or Si ₂ H ₆ or SiH ₂ C
l, SiHCl _3, chlorosilane-based source gas of SiCl _4, etc., SiH ₂
A chlorosilane-based source gas such as F ₂ or SiF ₄ is used,
This can be achieved by mixing this with an etching gas such as HCl and an H ₂ diluting gas and growing at a temperature of 800 to 1200 ° C.

Similarly, the crystal 25a that has continued to grow from the seed crystal 24a has a grain boundary (Grain Bounda) at an intermediate position between the crystal 25b that has grown from the adjacent seed crystal 24b and the first patterned seed crystal.
ry) yields 26. Originally, if two crystals 25a and 25b have exactly the same crystallographic orientation including the in-plane direction, no grain boundary will be formed and a continuous crystal film will be formed. Due to the misalignment, grain boundaries are formed in most cases.

Since the grown crystals 25a and 25b are single crystals, they have crystal faces (facets) unique to the crystals. Therefore, when an electronic element or the like is formed on the grown crystal, the crystal is polished as necessary as shown in FIG.
The surface may be flattened.

Example An example of the present invention will be described below with reference to the drawings.

(Example 1) FIGS. 3 (A) to 3 (D) are process diagrams showing a step of forming an original seed of a first example of the method for forming a crystal of the present invention.

The step of aggregating the original seed to form a seed crystal and growing a single crystal starting from the seed crystal is shown in FIGS.
Since the process is the same as that described with reference to FIG.
To (D) and the same reference numerals.

First, as shown in FIG. 3A, a 4-inch Si wafer 31 was prepared as a substrate, and the wafer surface was oxidized by 1 μm to form an SiO ₂ layer 32. After that, a part of the surface of the SiO ₂ layer 32 was patterned into a square by a normal photo process, and the square part was etched to the middle of the SiO ₂ layer by RIE (reactive ion etching). At this time, one side of the square was a = b = 0.8 μm, and the etching depth c was 0.4 μm. A plurality of similar concave portions were formed in a matrix with a vertical and horizontal interval 1 = 60 μm (only the concave portions 33a and 33b are shown in FIG. 3A).

Next, as shown in FIG. 3 (B), LPCV
0.4 μm of polycrystalline Si34 was deposited by using the D method. Phosphorus ( ³¹ P ⁺ ) was implanted into polycrystalline Si by an ion implantation method so as to have a concentration of 8 × 10 ²⁰ cm ⁻³ . Polycrystalline further doped with phosphorus
A resist 35 was coated on the Si 34 so that the surface became flat.

Next, as shown in FIG. 3 (C), under the condition that the etching rate of the resist and that of polycrystalline Si are the same by RIE, an etch bag is performed on both layers until the surface of the SiO ₂ layer 32 is exposed. Only polycrystalline Si34a and 34b were left in the inside. These polycrystalline Si34a and 34b were used as original seeds. still,
At this time, the etch-back condition is that OFPR5000 is used for the resist, and the gas used is C ₂ F ₆ / O ₂ = 80/35 (s) as the RIE condition.
ccm), pressure 50pa. output 2kW.

Next, as shown in FIG. 2 (B), the polycrystalline Si 34a, 34b embedded in the rectangular
Heat treatment was performed in a hydrogen atmosphere at a pressure of 100 Torr and a temperature of 1050 ° C. for 10 minutes. Then, the polycrystalline Si34a, 34b aggregated in the solid phase, and the surface became hemispherical and changed to single crystals 24a, 24b.

Next, as shown in FIG. 2 (C), the aggregated single crystal Si
Si crystals were selectively grown using 24a and 24b as seed crystals. The growth conditions were as follows: SiH ₂ Cl ₂ as a source gas, HCl as an additive gas for etching, and H ₂ as a carrier gas.
The substrate was placed in a gas mixture of 3, 1.6, and 100 (l / min) at a temperature of 1030 ° C., a pressure of 80 Torr, and a growth time of 90 minutes. As a result, when the Si single crystal 25a grew 30 μm in the lateral direction, it collided with the adjacent single crystal 25b to form a grain boundary 26. However, the facet shape and facet direction of the appearance were almost the same for all crystals.

The obtained crystal was oriented in the direction (100) perpendicular to the substrate, and other directions could not be observed when measured by X-ray diffraction. Regarding the in-plane orientation, ECP (Electron Chan
neling Pattern), the patterning direction of the seed crystal was ± 100 relative to the (100) equivalent direction.
He was still turning within 5mm.

Next, as shown in FIG. 2 (D), the crystal was planarized in order to form an element such as a transistor on the crystal thus obtained. The polishing was performed in exactly the same process as that for polishing a Si wafer, that is, using a weak alkaline solution in which abrasive grains were suspended, followed by rough polishing and finish polishing in the order of 1 μm in thickness.

Example 2 FIGS. 4 (A) to 4 (D) are process diagrams showing an original seed forming step of a second example of the crystal forming method of the present invention.

The step of aggregating the original seed to form a seed crystal and growing a single crystal starting from the seed crystal is shown in FIGS.
Since the process is the same as that described with reference to FIG.
To (D) and the same reference numerals.

First, as shown in FIG. 4A, 0.1 μm of polycrystalline Si is deposited on a 4-inch diameter fused quartz substrate 41 by LPCVD.
Further, a Si ₃ N ₄ film 43 was deposited thereon by 0.05 μm by LPCVD.

Next, as shown in FIG. 4 (B), an island region of 1 × 1 μm in length and width is patterned at an interval of 50 μm, and an Si ₃ N ₄ film and a polycrystalline Si film of another portion are left except for the island region. All were etched by RIE.

Next, as shown in FIG. 4 (C), the polycrystalline Sis 42a and 42b separated into islands were oxidized with the caps of the Si ₃ N ₄ films 43a and 43b attached. Square polycrystalline Si viewed from above the substrate
Is oxidized, and the polycrystalline Si 42a, 42b is surrounded by SiO ₂ walls 44a, 44b.

Next, as shown in FIG. 4 (D), hot phosphoric acid (H ₃ P
O ₄ ) was used to etch only the Si ₃ N ₄ layers 43 a and 43 b on the polycrystalline Si 42 ′.

After that, according to the process diagrams shown in FIGS. 2 (A) to 2 (D),
Under the same conditions as in the first embodiment, polycrystalline Si42a and 42b were aggregated and further grown selectively. Flattening was performed in the same manner.

Example 3 FIGS. 5A to 5C are process diagrams showing forming steps of a third example of the crystal forming method of the present invention.

First, as shown in FIG. 5 (A), a fused quartz substrate 51 having a diameter of 4 inches is prepared, and the surface thereof is 40 μm in length and width and 0.8 μm in depth.
The μm region was etched by an ordinary etching process using photolithography and RIE, and this was used as a crystal growth region 53. Note that this crystal growth region was formed in a lattice pattern at intervals of 60 μm as the distance between the centers of the squares.

Next, an additional 0.7 × is placed at the center of the concave bottom of this crystal growth region.
The 0.7 μm ² square region was etched to a depth of 0.2 μm, and this was used as a seed crystal forming region.

Ge is sputter-deposited on the entire surface of the quartz substrate 51 by 0.2.
Then, Ge deposited outside the seed crystal forming region was removed by etch-back in the same manner as in the first embodiment. The original seed 52 of Ge was left in the seed crystal forming region.

Next, as shown in FIG. 5 (B), the substrate on which the Ge original seed 52 has been applied is heat-treated in a hydrogen atmosphere under the conditions of 80 Torr and 750 ° C. to aggregate the original seed to form a single crystal Ge. Seed crystal
54.

Next, starting from the single crystal Ge, a Si crystal was selectively grown from the gas phase, and a Si single crystal 55 having a size of about 50 μm was grown. The growth conditions at this time were the same as in the first embodiment.

The flattening of the obtained crystal is performed in the same manner as in the first embodiment.
The polishing was performed according to the polishing process of the Si wafer. However, at this time, colloidal silica was used as the abrasive grains, and the polishing speed was reduced immediately before the polished surface of the crystal reached the surface of the quartz substrate 51, so that the Si crystal surface was made to coincide with the surface of the quartz substrate 51. Then polishing was finished.

Since the crystal was grown and polished in this manner, the Si single crystal 55 'could be obtained in a completely separated form and completely flat on the entire surface.

The plane orientation of the crystals was measured for each individual crystal by the micro X-ray diffraction method.
Extremely good orientation with a deviation of 3 ° or less was shown. In addition, when the whole X-ray diffraction was observed, no peak other than (100) appeared. When the in-plane orientation was observed by the ECP method, the orientation was (100) with respect to the patterning direction of the seed crystal (the lattice pattern direction of the crystal growth region), and the rotation was within ± 4 ° from the equivalent orientation.

[Effects of the Invention] As described above, according to the method for forming a crystal of the present invention, a single crystal is selectively grown starting from a single crystal seed crystal having a controlled plane orientation and in-plane orientation. Therefore, the position of the grain boundary is controlled on the insulating substrate, and a single crystal having a uniform plane orientation and in-plane orientation can be formed.

[Brief description of the drawings]

1 (A) to 1 (D) are explanatory views showing the shape of an original seed and the form of an insulator surrounding the original seed. 2 (A) to 2 (D) are process diagrams of one embodiment of the method for forming a crystal of the present invention. FIGS. 3 (A) to 3 (C) are process diagrams showing a process of forming an original seed of the first embodiment of the method of forming a crystal of the present invention. FIGS. 4 (A) to 4 (D) are process diagrams showing a step of forming original seeds according to a second embodiment of the method of forming crystals of the present invention. 5 (A) to 5 (C) are process diagrams showing a forming process of a third embodiment of the crystal forming method of the present invention. 1 ... original seed, 2 ... insulating substrate, 31 ... Si wafer,
32: SiO ₂ layer, 33a, 33b: recess, 34, 34a, 34b: polycrystalline Si, 35: resist, 41: fused quartz substrate, 42a, 42b
…… Polycrystalline Si, 43a, 43b …… Si ₃ N ₄ film, 44a, 44b …… SiO ₂
Wall, 51: fused quartz substrate, 52: original seed, 53: crystal growth region, 54: seed crystal, 55, 55 ': Si single crystal.

Claims (4)

(57) [Claims] A cubic or cubic shape in which all side walls and bottom surfaces are in contact with an insulator and are surrounded on a surface of a substrate having a surface having a small nucleation density or in a recess formed in the substrate. And, the original seeds whose volume is small enough to be aggregated singly are arranged, and a heat treatment for causing the original seeds to cause aggregation is performed to obtain a monocrystalline seed crystal with a controlled plane orientation and in-plane orientation, A method for forming a crystal in which a single crystal is selectively grown starting from the seed crystal by performing a crystal growth treatment. 2. The method according to claim 1, wherein the surface having a low nucleation density is an amorphous insulating surface. 3. The method according to claim 1, wherein the heat treatment temperature is lower than the melting point of the original seed. 4. The method according to claim 1, wherein the heat treatment of the original seed is performed in a hydrogen atmosphere.