JP2013177256A - Periodic table group 13 metal nitride substrate - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a periodic table group 13 metal nitride substrate high in crystallinity near a crystal surface.SOLUTION: An as-grown crystal is sliced and then subjected to grinding and polishing so that (FWMH)/(FWMH)≤1.25 is satisfied when the minimum half width obtained when the half widths of an X-ray rocking curve are measured by varying the penetration depth in the direction from the substrate principal plane to the substrate thickness is defined as (FWMH), and the minimum half width calculated from the theory of dynamical diffraction of a periodic table group 13 metal nitride perfect catalyst is defined as (FWMH). Thus, the difference between the metal nitride and the perfect crystal at the minimum half-width of the X-ray rocking curve present in a position about 112 nm deep from the substrate surface is reduced.

Description

  The present invention relates to a periodic table group 13 metal nitride substrate, and more particularly to a periodic table group 13 metal nitride substrate having an ideal low strain surface layer.

  Periodic table Group 13 metal nitrides typified by gallium nitride are high-frequency and high-power electrons such as light-emitting devices such as light-emitting diodes and laser diodes, high electron mobility transistors (HEMT), and heterojunction bipolar transistors (HBT). Useful as a material applied to devices. In particular, when used as a light emitting device, efficient light emission is required due to the recent trend of energy saving, and for this purpose, a periodic table group 13 metal nitride crystal having high crystallinity is required.

  The nitride crystal having high crystallinity is, for example, a plane spacing at a depth of 0.3 μm in a plane spacing of a specific crystal lattice plane obtained from an X-ray diffraction measurement that changes the X-ray penetration depth from the crystal surface. And the uniform distortion of the surface layer of the crystal obtained from the interplanar spacing at a depth of 5 μm is disclosed (see Patent Document 1).

JP 2007-5526 A

K. Akimoto and T. Emoto, Journal of Physics: Condensed Matter, 22, 473001, 2010.

When the present inventors examined the above method, the information obtained by the conventional X-ray diffraction method was obtained only for the entire crystal, and could not be obtained by limiting the vicinity of the surface. Therefore, the present inventors have found that even a nitride crystal satisfying the parameters described in Patent Document 1 may have strain and crystal orientation distribution on the crystal surface. That is, a periodic table group 13 metal nitride substrate having high crystallinity on the crystal surface has not been studied.
An object of the present invention is to obtain a periodic table group 13 metal nitride substrate having high crystallinity in the vicinity of the crystal surface.

The inventors of the present invention have studied to obtain a periodic table group 13 metal nitride substrate having high crystallinity in the vicinity of the crystal surface. By using the “extremely asymmetric X-ray diffraction method”, the vicinity of the crystal surface is obtained. It has been found that this situation can be accurately grasped, and that the half width of the X-ray rocking curve in a complete crystal can be calculated by this method (see Non-Patent Document 1).
Under such circumstances, the present inventors pay attention to the fact that the full width at half maximum of the X-ray rocking curve in the perfect crystal shows a minimum value in the vicinity of the total reflection critical angle, and the actual periodic table group 13 metal nitride substrate X Measure the full width at half maximum of the line rocking curve and compare the minimum value with the minimum value in the complete crystal. If both values are approximate, the crystallinity near the crystal surface is high, and it is used as a light-emitting device. The present invention was completed by discovering that light was efficiently emitted when the light was emitted.

That is, the present invention is as follows.
(1) The minimum of the half-value width obtained when the half-value width of the X-ray rocking curve is measured by changing the penetration depth from the substrate main surface to the substrate thickness direction with respect to the periodic table group 13 metal nitride substrate. Define the value as (FWMH) ⁱ _m ,
When the minimum value of the half width calculated from the dynamic diffraction theory of the group 13 metal nitride perfect crystal of the periodic table is defined as (FWMH) ⁱ _t ,
(FWMH) ⁱ _m / (FWMH) ⁱ _t ≦ 1.25
The periodic table group 13 metal nitride substrate satisfying
(2) Obtained when the half-value width of the X-ray rocking curve is measured by changing the penetration depth from 500 nm to 1000 nm from the substrate main surface to the substrate thickness direction with respect to the periodic table group 13 metal nitride substrate. The periodic table group 13 metal nitride substrate according to (1), wherein an average value of the half widths is 200 arcsec or less.
(3) X-ray rocking curve measurement was performed on the periodic table group 13 metal nitride substrate at a penetration depth of 300 nm from the substrate main surface in the substrate thickness direction, and the dispersion of the obtained crystal lattice plane inclination Δθ was zero. The periodic table group 13 metal nitride substrate according to (1) or (2), which is .050 ° or less.
(4) X-ray rocking curve measurement is performed on the periodic table group 13 metal nitride substrate at a penetration depth of 400 nm from the main surface of the substrate in the thickness direction of the substrate, and the dispersion of the obtained crystal lattice plane inclination Δθ is 0 The periodic table group 13 metal nitride substrate according to any one of (1) to (3), which is 0.040 ° or less.
(5) An X-ray rocking curve measurement at a penetration depth of 500 nm from the substrate main surface to the substrate thickness direction is performed on the group 13 metal nitride substrate of the periodic table, and the dispersion of the obtained crystal lattice plane inclination Δθ is 0 The periodic table group 13 metal nitride substrate according to any one of (1) to (4), which is 0.030 ° or less.
(6) The X-ray rocking curve measurement was performed on the periodic table group 13 metal nitride substrate by changing the penetration depth from 300 nm to 400 nm from the substrate main surface to the substrate thickness direction, and the horizontal axis represents the penetration depth. The period according to any one of (1) to (5), wherein the slope of the linear line formed when the vertical axis is plotted as the dispersion of the slope Δθ of the crystal lattice plane is −0.01 ° / 100 nm or more. Table Group 13 metal nitride substrate.

  The periodic table group 13 metal nitride substrate of the present invention has high crystallinity on the substrate surface and emits light efficiently when used as a light emitting device.

It is a conceptual diagram which shows the aspect of the apparatus which manufactures the as-grown crystal of this invention. It is a conceptual diagram which shows the aspect of the apparatus which performs a X-ray rocking curve measurement. It is the graph which plotted the half value width of the GaN board | substrate obtained in the Example of this invention, and the half value width of a GaN perfect crystal. It is a graph which shows the fitting in the process in which the dispersion | distribution of inclination (DELTA) (theta) of a crystal lattice plane in this invention is calculated. It is a graph which shows the ratio of (DELTA) (theta) in the process which calculates dispersion | distribution of inclination (DELTA) (theta) of a crystal lattice plane in this invention. It is the graph which plotted the dispersion | distribution of inclination (DELTA) (theta) of a crystal lattice plane in the board | substrate obtained in the Example of this invention.

  The periodic table group 13 metal nitride substrate of the present invention will be described in detail below. The description of the constituent elements may be based on typical embodiments of the present invention, but the present invention is not limited to such embodiments.

In the present invention, the periodic table group 13 metal nitride substrate means a semiconductor substrate made of a periodic table group 13 metal nitride crystal, and the periodic table group 13 metal nitride crystal is, for example, Ga _x Al _y In ₁₋ It is represented by an _xy N crystal (where 0 ≦ x ≦ 1, 0 ≦ y ≦ 1), and specifically includes gallium nitride, aluminum nitride, indium nitride, or a mixed crystal thereof. In the present invention, the main surface means a surface on which a device is to be formed or a widest surface in the structure. In the nitride substrate of the present invention, the surface index of the main surface is not particularly limited, and may be any of a C surface that is a polar surface, an A surface that is a nonpolar surface, an M surface, or a semipolar surface.

In the present invention, a crystal having a surface on which an epitaxial layer is formed in order to form a semiconductor element is referred to as a substrate. In general, morphological processing such as slicing and grinding is performed from an as-grown crystal to obtain a substrate. In the present invention, a substrate in the middle of processing is not called a substrate but called a crystal.
Grinding is generally a process mainly for the purpose of form processing and refers to a process using fixed abrasive grains. The method for grinding is not particularly limited, but is a fixed abrasive in which diamond abrasive grains, silicon carbide abrasive grains, etc. are fixed with a binder, and the abrasive grains having a relatively large particle diameter are used for rough cutting. The method can be used.
On the other hand, polishing is generally a process aimed mainly at reducing distortion caused by surface processing. In the present invention, mechanical polishing and chemical mechanical polishing are collectively referred to as polishing. A method for performing polishing is not particularly limited, but a method of adjusting a fine surface state using loose abrasive grains having a relatively small particle diameter can be used. Of course, such surface polishing is possible even when fixed abrasive grains having a relatively small particle diameter are used, and even when fixed abrasive grains are used, they can be included in the range of polishing.
In the present invention, the term “plate” refers to a plate to which a crystal is attached in order to attach the crystal to an apparatus during grinding or polishing. The plate surface to which the crystal is attached is preferably flat in order to obtain a substrate having a uniform thickness after grinding and polishing.

The periodic table group 13 metal nitride substrate of the present invention measures the half-value width of the X-ray rocking curve with respect to the periodic table group 13 metal nitride substrate by changing the penetration depth from the substrate main surface to the substrate thickness direction. When the minimum value of the full width at half maximum is defined as (FWMH) ⁱ _m, and the minimum value of the full width at half maximum of the group 13 metal nitride perfect crystal of the periodic table is defined as (FWMH) ⁱ _t , (FWMH) ⁱ _m / (FWMH) ⁱ _t ≦ 1.25, a periodic table group 13 metal nitride substrate.
As described above, the information obtained by the conventional X-ray diffraction method is only information on the entire crystal, and information cannot be obtained by limiting the vicinity of the crystal surface. For this reason, it is not possible to fully recognize the existence of strain and crystal orientation distribution existing on the crystal surface, and methods for reducing the existence of such strain and crystal orientation distribution existing on the crystal surface have been studied. There wasn't.

The present inventors can accurately grasp the situation near the crystal surface by the “extremely asymmetric X-ray diffraction method”, and calculate the half-value width of the X-ray rocking curve in the complete crystal by this method. It was found that this is possible (see Non-Patent Document 1). Then, paying attention to the fact that the full width at half maximum of the X-ray rocking curve in the complete crystal shows a minimum value near the total reflection critical angle, the half width of the X-ray rocking curve was measured for an actual substrate, and the minimum value and the complete crystal In comparison with the local minimum value, the crystallinity was high when both values were approximated, and it was found that light was efficiently emitted when used as a light-emitting device.
In addition, when the half width of the (103) reflection X-ray rocking curve is measured for the GaN substrate having the C plane as the main surface, the minimum value is shown at the penetration depth of about 112 nm from the C plane, which is the total reflection. The local minimum is near the critical angle. When the main surface has an off angle, the penetration depth that gives the minimum value varies depending on the magnitude of the off angle.

The value of (FWMH) ⁱ _m / (FWMH) ⁱ _t is preferably 1.23 or less, more preferably 1.21 or less, still more preferably 1.18 or less. 1
Particularly preferred is 5 or less. The closer this value is to 1, it can be said that it has a surface similar to a perfect crystal and has good crystallinity. The value of the _{^{(FWMH) i m / (FWMH}} ) i t is usually 1.00 or more.
Further, in the GaN substrate having the C plane as the main surface, the minimum value of the half-value width of the X-ray rocking curve in the complete crystal is 44.5 arcsec as described later, but the periodic table group 13 metal nitride of the present invention is used. The minimum value of the half width of the substrate is preferably 55.0 arcsec or less, more preferably 54.0 arcsec or less, further preferably 53.0 arcsec or less, and particularly preferably 52.0 arcsec or less. preferable.

In the present invention, (FWMH) ⁱ _m of the periodic table Group 13 metal nitride substrate can be extremely measured by asymmetrical X-ray diffraction method. This method measures the rocking curve by making X-rays enter near the total reflection critical angle, for example, by setting the incident wave of X-rays at an angle of about 0 ° to 2 ° with respect to the crystal surface.

In this case, it is preferable to use an X-ray source having a high intensity, for example, synchrotron radiation.
Hereinafter, a measurement method when a gallium nitride substrate is used as the Group 13 metal nitride substrate of the periodic table will be described with reference to FIG.
FIG. 2 shows a measuring instrument that can be used in the present invention. X-rays emitted from the radiation source 1 are monochromatized and collimated by the spectral crystal 2. A beam having a divergence angle by a spectral crystal of 10 arcsec or less and a monochromaticity of ΔE / E = 4.0 × 10 ⁻⁴ or less is used. The X-rays monochromatized and collimated by the spectroscopic crystal 2 then pass through the ion chamber 3 and enter the periodic table group 13 metal nitride substrate disposed on the goniometer 4. The incident angle at this time is 0 ° to 2 ° from the nitride substrate surface. Then, the rocking curve is measured by measuring the diffracted wave from the nitride substrate surface with the detector 5.

The full width at half maximum of the X-ray rocking curve of the group 13 metal nitride perfect crystal of the periodic table can be calculated by the method using the dynamic diffraction theory described in Non-Patent Document 1. Specifically, since an extremely asymmetric X-ray diffraction method is used, it is necessary to consider specular reflection waves, and the diffraction intensity can be obtained from the following basic equation using the theory of three-wave approximation.

The half width of the periodic crystal of Group 13 metal nitride perfect crystal thus calculated gives a local minimum value near the total reflection critical angle. As a result of studies by the present inventors, the X-ray incident at the total reflection critical angle, that is, the angle at which the X-ray is totally reflected, contains information that best represents the quality of the crystal surface. Therefore, even in the periodic table group 13 metal nitride substrate actually manufactured, the crystallinity of the substrate surface is best represented, and a crystal having a close value in comparison with a perfect crystal using the minimum value of the half width is a crystal. When it is made into a light emitting device, it emits light efficiently.
There are some periodic table group 13 metal nitride substrates that do not have a minimum value in the half-value width in the X-ray rocking curve measurement. In such a case, the full width at half maximum at the penetration depth that gives the minimum value near the critical angle near the total reflection in the complete crystal is regarded as the minimum value, and the ratio of the full width at half maximum in the complete crystal is calculated. And

In comparison with such a perfect crystal, the method for providing a periodic table group 13 metal nitride substrate showing a close minimum value in X-ray rocking curve measurement is grinding and polishing performed after slicing from an as-grown crystal. It may be performed appropriately.
For example, if chipping occurs at the ends of the crystal due to polishing, the chip may cause scratches on the crystal surface, and a substrate with good crystallinity cannot be obtained. In order to prevent such chipping of the crystal end due to polishing, it is possible to chamfer the end of the crystal before chamfering. Further, in the polishing, mechanical polishing is performed in two stages, and chamfering polishing is performed between the first stage mechanical polishing and the second stage mechanical polishing. In addition, the half width ratio of the present invention can be reduced by taking a sufficient time for chemical mechanical polishing performed after mechanical polishing.
Further, when the crystallinity of the seed used for growing the as-grown crystal is low, the crystallinity of the as-grown crystal tends to be low. Therefore, the half-width ratio of the present invention can also be reduced by growing an as-grown crystal using a highly crystalline seed as a seed. In addition, the detailed manufacturing method for obtaining the periodic table group 13 metal nitride substrate of this invention is mentioned later.

  Moreover, the periodic table group 13 metal nitride substrate of the present invention is obtained when the half-value width of the X-ray rocking curve is measured by changing the penetration depth from 500 nm to 1000 nm from the substrate main surface to the substrate thickness direction. It is preferable that the average value of the full width at half maximum is 200 arcsec or less. A portion having a penetration depth of 500 nm to 1000 nm in the substrate is a portion showing bulk crystallinity, and that the average value of the half-value width is 200 arcsec or less at this position is close to the value in the complete crystal and high in crystallinity. means. More preferably, it is 150 arcsec or less, and further preferably 100 arcsec or less.

In addition, the periodic table group 13 metal nitride substrate of the present invention performs X-ray rocking curve measurement at a penetration depth of 300 nm in the substrate thickness direction from the substrate main surface, and the obtained dispersion of the inclination Δθ of the crystal lattice plane is The angle is preferably 0.05 ° or less.
As a result of investigations by the present inventors, it has been found that the main cause of the expansion of the half-value width is the inclination Δθ of the crystal lattice plane. The inclination Δθ of the crystal lattice plane means the inclination of the crystal lattice plane with respect to the main surface. The larger the inclination Δθ of the crystal lattice plane, the more the crystal plane is inclined from the main surface. Further, the dispersion of the crystal lattice plane inclination Δθ represents the strain of the crystal, and the smaller the dispersion value, the smaller the strain.

A method for calculating the dispersion of the inclination Δθ of the crystal lattice plane will be described by taking as an example the case where the Group 13 metal nitride of the periodic table is gallium nitride.
First, the rocking curve of the complete gallium nitride crystal is calculated, the ratio of these calculated peaks is changed, the experimental values are fitted, and the abundance ratio of Δθ is calculated. FIG. 4 shows the rocking curve and fitting when the penetration depth corresponds to 330 nm. FIG. 4 is calculated by changing the inclination of the crystal plane by 0.005 ° in a range of ± 0.03 °.
The abundance ratio of Δθ at this time was plotted as shown in FIG. 5, and a value at a portion where the abundance ratio was 0.5 was obtained, and the value divided by 2 was defined as a dispersion of Δθ.

The dispersion of Δθ at the penetration depth of 300 nm from the substrate main surface determined as described above is preferably 0.050 ° or less, more preferably 0.030 ° or less,
The angle is more preferably 0.020 ° or less, and particularly preferably 0.015 ° or less.

In addition, the periodic table group 13 metal nitride substrate of the present invention performs X-ray rocking curve measurement at a penetration depth of 400 nm from the main surface of the substrate in the thickness direction of the substrate. It is preferable that it is 0.040 degrees or less.
The dispersion of Δθ at the penetration depth of 400 nm determined as described above is preferably 0.020 ° or less, and more preferably 0.015 ° or less.

In addition, the periodic table group 13 metal nitride substrate of the present invention performs X-ray rocking curve measurement at a penetration depth of 500 nm from the main surface of the substrate in the thickness direction of the substrate. It is preferable that it is 0.030 degrees or less.
The dispersion of Δθ at the penetration depth of 500 nm determined as described above is preferably 0.020 ° or less, more preferably 0.015 ° or less, and 0.010 ° or less. Is particularly preferred.

  Further, the periodic table group 13 metal nitride substrate of the present invention performs an X-ray rocking curve by changing the penetration depth from 300 nm to 400 nm in the substrate thickness direction from the substrate main surface, and the horizontal axis represents the penetration depth. It is preferable that the slope of the linear line formed when the vertical axis is plotted as the variance of the inclination Δθ of the crystal lattice plane is −0.01 ° / 100 nm or more. The periodic table group 13 metal nitride substrate showing the inclination of such a linear straight line has a high crystallinity and tends to have a small strain change particularly in the depth direction. This linear line can be obtained as a linear approximation line by the method of least squares.

  Hereinafter, the manufacturing method for obtaining the periodic table group 13 metal nitride substrate of the present invention will be described. The periodic table group 13 metal nitride substrate of the present invention has very high crystallinity, and it is important that scratches due to chipping do not occur on the substrate surface in the polishing process. As an example of a manufacturing method that does not cause such chipping, a method of performing mechanical polishing in two stages in polishing and chamfering polishing between them will be described, but the present invention is not limited to the following manufacturing method. . Of the mechanical polishing performed in two stages, the first mechanical polishing is referred to as first mechanical polishing (hereinafter sometimes simply referred to as “first polishing”) and is performed after chamfering. The polishing is referred to as second mechanical polishing (hereinafter, this may simply be abbreviated as “second polishing”).

<First mechanical polishing>
In general, a periodic table group 13 metal nitride substrate is obtained by performing processing such as slicing, grinding, and polishing from an as-grown crystal. Usually, the first mechanical polishing of the present invention is performed by slicing an as-grown crystal, chamfering the side surface, and performing processing such as grinding or etching on the back surface or the surface as necessary, and then group 13 metal nitride in the periodic table. This is performed on a physical crystal (hereinafter sometimes referred to as a nitride crystal). These processes should just employ | adopt a well-known method, and a processing method and an order are not limited. In addition, the above treatment is an example, and the first mechanical polishing of the present invention is not necessarily applied to a crystal that has been subjected to all such treatments.

For the first polishing, a known method of polishing by attaching the back side of the nitride crystal so as to face the plate can be employed. Usually, in the polishing step, mechanical polishing is performed with the abrasive grain size being reduced stepwise. The first polishing is a stage of mechanical polishing with a relatively large abrasive particle size, but the abrasive particle size does not necessarily have to be large. When the first polishing is performed with a small abrasive particle size, the time required for the first polishing becomes longer.
Therefore, the grain size R1 of the abrasive grains used for the first polishing is preferably large, and usually R1 is 0.5 μm or more, preferably 1 μm or more. On the other hand, the upper limit is usually 30 μm or less, preferably 15 μm or less, more preferably 3 μm or less. As the abrasive grains, diamond is usually used, but it is not necessarily limited to this as long as polishing is possible.
Also, the polishing speed of the first polishing is not particularly limited, and in the case of polishing by rotation, the normal rotation speed is usually 10 rpm or more, preferably 20 rpm or more. On the other hand, the upper limit is usually 50 rpm or less, preferably 40 rpm or less. Further, the pressure when the crystal is arranged to face the plate is not particularly limited, and is usually 50 g / cm ² or more and 500 g / cm ² or less. In addition, the time for performing the first polishing may be appropriately adjusted according to the thickness of the machining allowance, and the polishing rate may be appropriately adjusted according to the polishing conditions.

<Second mechanical polishing>
The second mechanical polishing of the present invention can be performed by the same method as the first mechanical polishing. In addition, since the second mechanical polishing is a finish polishing, one having a small abrasive grain size is usually used. The size may be appropriately set by those skilled in the art depending on the degree of finishing, but the grain size R2 of the abrasive grain used in the second mechanical polishing is preferably smaller than the abrasive grain R1 of the abrasive grain used in the first mechanical polishing, Usually, R2 is 0.05 μm or more, preferably 0.125 μm or more. On the other hand, the upper limit is preferably 0.5 μm or less. As the abrasive grains, diamond, SiC powder or the like is usually used.
Also, the polishing speed of the second polishing is not particularly limited, and in the case of polishing by rotation, the normal rotation speed is 10 rpm or more, preferably 20 rpm or more. On the other hand, the upper limit is usually 50 rpm or less, preferably 40 rpm or less. Further, the pressure when the crystal is arranged to face the plate is not particularly limited, and is usually 50 g / cm ² or more and 500 g / cm ² or less. In addition, the implementation time of the second polishing may be appropriately adjusted according to the thickness of the machining allowance, and the polishing rate may be appropriately adjusted according to the polishing conditions.

  Both the first mechanical polishing and the second mechanical polishing are steps for making the nitride crystal to a desired thickness, and the time for performing each step is the desired thickness and the abrasive grains. A person skilled in the art appropriately sets the particle size according to the size.

<Chamfering polishing>
Chipping can be effectively prevented by performing chamfering before the second polishing in which the grain size of the abrasive grains is small. And the crystallinity of the board | substrate in this invention can be raised, and ratio with a perfect crystal can be made small.
By carrying out the chamfering polishing of the present invention, a fringe is formed on the outer peripheral portion of the nitride crystal to round the edge portion of the crystal so that chipping at the edge hardly occurs. It is preferable that the peripheral portion of the outer peripheral portion is in the center of 20 μm or more from the outermost edge of the crystal, more preferably 30 μm or more, further preferably 40 μm or more, preferably 5,000 μm or less, More preferably, it is 3,000 μm or less.

  The outer peripheral portion of the nitride crystal means the outer peripheral portion of the crystal surface where chipping is likely to occur during polishing, and is not strictly specified, but usually means about 500 μm on the outer edge side of the crystal, It is preferably about 300 μm.

The method of chamfering polishing of the present invention is not limited as long as the chamfering can be formed on the outer peripheral portion to round the edge portion of the crystal, but it is preferably performed by chemical mechanical polishing (CMP). The chemical mechanical polishing in this case is referred to as first chemical mechanical polishing in this specification.
As the chamfering polishing, it is preferable to perform the first chemical mechanical polishing under conditions that sink into a polishing pad or the like because the edge portion of the crystal can be more effectively polished.
When the chamfering polishing is performed by chemical mechanical polishing (CMP), it is possible to make the outer peripheral portion of the nitride crystal difficult to chip by performing it for approximately two hours or more. The upper limit of the chemical mechanical polishing time is not particularly limited, but it is preferably 10 hours or less, more preferably 5 hours or less in consideration of the time required for the entire polishing.

The first chemical mechanical polishing can be carried out by employing a known method, for example, using a colloidal silica slurry and a suede pad. A known material may be used for the suede pad, and examples thereof include polyurethane. The particle size of the abrasive grains (colloidal silica) contained in the colloidal silica slurry is usually 10 nm or more, and preferably 30 nm or more. The upper limit is usually 100 nm or less. The pH of the colloidal silica slurry is preferably 0.8 or more and 2.5 or less.
Further, the polishing speed of the first chemical mechanical polishing is not particularly limited, and in the case of polishing by rotation, the normal rotation speed is 50 rpm or more and 200 rpm or less. Further, the pressure at the time of disposing the crystal so as to face the plate is not particularly limited, and is usually 100 g / cm ² or more and 1500 g / cm ² or less. The time for performing the first chemical mechanical polishing is usually 1 hour or longer, preferably 2 hours or longer, and usually 10 hours or shorter, preferably 5 hours or shorter. About a polishing rate, what is necessary is just to adjust suitably according to polishing conditions.
The nitride semiconductor substrate after the completion of the first chemical mechanical polishing may have a surface roughness Rms of 10 nm or more, 50 nm or more, or 100 nm or more.

<Second chemical mechanical polishing>
The crystallinity of the substrate can also be increased and the ratio to the complete crystal can be reduced by further performing chemical mechanical polishing on the surface of the nitride crystal obtained by the second mechanical polishing. In addition, in order to distinguish from the chemical mechanical polishing in chamfering polishing, in this specification, it calls a 2nd chemical mechanical polishing.
The second chemical mechanical polishing is a process for removing the work-affected layer present on the surface of the nitride crystal. Although the work-affected layer on the outer periphery of the crystal has been roughly removed by the chamfering described above, a high-quality nitride substrate is provided by removing the work-affected layer on the other crystal surface in this step. be able to.

Similar to the first chemical mechanical polishing, the second chemical mechanical polishing can be carried out by employing a known method, for example, using a colloidal silica slurry and a suede pad. A known material may be used for the suede pad, and examples thereof include polyurethane. In addition, it is preferable to use a pad that is harder than the pad used in the first chemical mechanical polishing because there is less debris from the end of the crystal.
The particle size of the abrasive grains (colloidal silica) contained in the colloidal silica slurry is usually 10 nm or more, and preferably 30 nm or more. The upper limit is usually 100 nm or less. The pH of the colloidal silica slurry is preferably 0.8 or more and 2.5 or less.
Further, the polishing speed of the second chemical mechanical polishing is not particularly limited, and in the case of polishing by rotation, the normal rotation speed is 50 rpm or more and 200 rpm or less. Further, the pressure at the time of disposing the crystal so as to face the plate is not particularly limited, and is usually 100 g / cm ² or more and 1500 g / cm ² or less. About a polishing rate, what is necessary is just to adjust suitably according to polishing conditions.

The first chemical mechanical polishing is different from the first chemical mechanical polishing in that the first chemical mechanical polishing is intended to remove the work-affected layer on the outer periphery of the crystal, and the second chemical mechanical polish is used to remove the work-affected layer on the entire nitride crystal surface. It is intended to be removed. Therefore, the implementation time of the second chemical mechanical polishing is longer than the implementation time of the first chemical mechanical polishing, and is usually 10 hours or more, preferably 15 hours or more. On the other hand, the upper limit is not particularly limited, but is preferably 20 hours or less in consideration of the time required for the entire polishing process.
The nitride substrate after completion of the second chemical mechanical polishing preferably has a surface roughness Rms of 1 nm or less.

  Next, a method for obtaining an as-grown crystal will be described. The as-grown crystal can be obtained by growing a seed by a known method. Here, the HVPE growth will be described as an example of the GaN crystal.

  FIG. 1 is a diagram for explaining an outline of a configuration example of an apparatus used for GaN growth by the HVPE method. In the figure, reference numeral 100 is a reactor, 101 to 104 are gas introduction pipes, 105 is a reservoir, 106 is a heater, 107 is a substrate holder, and 108 is a gas discharge pipe.

As a material of the reactor 100, quartz, polycrystalline boron nitride (BN) stainless steel or the like is used. A preferred material is quartz. The reactor 100 is filled with atmospheric gas in advance before starting the reaction. Examples of the atmospheric gas include H ₂ gas, N ₂ gas, inert gas such as He, Ne, and Ar. These gases may be mixed and used.

  Carbon is preferable as the material of the substrate holder 107, and a material whose surface is coated with SiC is more preferable. The shape of the substrate holder 107 is not particularly limited as long as the base substrate 10 of the present invention can be installed, but there is no structure upstream of the growing crystal when the crystal is grown. It is preferable. If there is a structure in which a crystal may grow on the upstream side, a polycrystal adheres to the structure, and HCl gas is generated as a product, which adversely affects the crystal to be grown. The size of the base substrate mounting surface of the substrate holder 107 is preferably smaller than the base substrate 10 to be mounted. That is, it is more preferable that the substrate holder 107 is hidden by the size of the base substrate 10 when viewed from the gas upstream side.

When the base substrate 10 is placed on the substrate holder 107, the base substrate 10 is preferably placed so that the growth surface of the base substrate 10 faces the upstream side of the gas flow (above the reactor in FIG. 1). By placing the base substrate 10 in this way, a more uniform GaN crystal having excellent crystallinity can be obtained.
As the base substrate 10, a target periodic table group 13 metal nitride crystal such as GaN, InGaN, or AlGaN is used, a metal oxide such as sapphire, ZnO, or BeO, or a silicon such as SiC or Si. Inclusions, GaAs or the like can be used. Considering consistency with the group 13 metal nitride of the periodic table to be grown, the group 13 metal nitride of the periodic table is most preferable. In addition, when a crystal other than the Group 13 metal nitride crystal of the periodic table is used as the base substrate, a group 13 metal nitride layer of the periodic table is formed on the crystal in order to enhance the consistency (template) Substrate) may be used. The shape of the base substrate is not particularly limited, and may be a flat plate shape or a rod shape.

The reservoir 105 is charged with a raw material to be a Ga source. Examples of other Group 13 metal sources of the periodic table include Al and In. A gas that reacts with the raw material put in the reservoir 105 is supplied from an introduction pipe 103 for introducing the gas into the reservoir 105. For example, when a raw material serving as a Ga source is put in the reservoir 105, HCl gas can be supplied from the introduction pipe 103. At this time, the carrier gas may be supplied from the introduction pipe 103 together with the HCl gas. Examples of the carrier gas include H ₂ gas, N ₂ gas, inert gas such as He, Ne, and Ar. These gases may be mixed and used. The carrier gas may be the same as or different from the atmospheric gas, but is preferably the same.

From the introduction pipe 101, a source gas serving as an N source is supplied. Normally, NH ₃ gas is supplied. A carrier gas is supplied from the introduction pipe 102. As the carrier gas, the same carrier gas supplied from the introduction pipe 103 can be exemplified. The carrier gas supplied from the introduction pipe 102 and the carrier gas supplied from the introduction pipe 103 are preferably the same. A dopant gas can also be supplied from the introduction pipe 102. For example, an n-type dopant gas such as SiH ₄ or SiH ₂ Cl ₂ can be supplied.

  An etching gas can be supplied from the introduction pipe 104. As an etching gas, a chlorine-based gas can be used, and HCl gas is preferably used. Etching can be performed by setting the flow rate of the etching gas to about 0.1% to 3% with respect to the total flow rate. A preferable flow rate is about 1% with respect to the total flow rate. The gas flow rate can be controlled by a mass flow controller (MFC) or the like, and the individual gas flow rates are preferably always monitored by MFC.

  The gases supplied from the introduction pipes 101, 102, and 104 may be exchanged with each other and supplied from different introduction pipes. In addition, the source gas and the carrier gas serving as a nitrogen source may be mixed and supplied from the same introduction pipe. Further, a carrier gas may be mixed from another introduction pipe. These supply modes can be appropriately determined according to the size and shape of the reactor 100, the reactivity of the raw materials, the target crystal growth rate, and the like.

  The gas discharge pipe 108 is generally installed so that it can be discharged from the reactor inner wall on the side opposite to the introduction pipes 101 to 104 for introducing gas. In FIG. 1, a gas discharge pipe 108 is installed on the bottom surface of the reactor located opposite to the top surface of the reactor where the introduction pipes 101 to 104 for gas introduction are installed. When the introduction pipe for introducing gas is installed on the right side of the reactor, the gas discharge pipe is preferably installed on the left side of the reactor. By adopting such an aspect, a gas flow can be stably formed in a certain direction.

  Crystal growth by the HVPE method is usually performed at 800 ° C. to 1200 ° C., preferably 900 ° C. to 1100 ° C., more preferably 925 ° C. to 1070 ° C., and more preferably 950 ° C. to 1050 ° C. preferable. The pressure in the reactor is preferably 10 kPa to 200 kPa, more preferably 30 kPa to 150 kPa, and even more preferably 50 kPa to 120 kPa. The etching temperature and pressure at the time of etching may be the same as or different from the crystal growth temperature and pressure.

  EXAMPLES Hereinafter, although an Example and a comparative example are given and this invention is demonstrated further in detail, it is not limitedly interpreted only to the specific form shown in the following Examples.

<Example>
(Manufacture of substrate A)
A C-plane gallium nitride crystal having an appropriate thickness was prepared from an as-grown gallium nitride crystal grown from a gallium nitride substrate having a C-plane as a main surface by slicing or back surface (N-plane) grinding.
Next, in order to polish the surface (Ga surface) of the crystal, the back surface (N surface) was attached to a polishing plate with wax, and the surface (Ga surface) was polished (first mechanical polishing).
Surface polishing was performed by first mechanical polishing (lapping) using diamond loose abrasive grains having an average diameter of 3 μm, and then lapping using diamond loose abrasive grains having an average diameter of 1 μm.
After lapping using 1 μm diamond abrasive grains, first chemical mechanical polishing (polishing) was performed for 2 hours using colloidal silica and a suede pad (made of urethane).
Mechanical polishing (second mechanical polishing) was further applied to the crystal surface after polishing. First, mechanical polishing (lapping) was performed using diamond loose abrasive grains having an average diameter of 0.5 μm, and then mechanical polishing (wrapping) was performed using diamond loose abrasive grains having an average diameter of 0.25 μm.
After lapping using 0.25 μm diamond abrasive grains, second chemical mechanical polishing (polishing) was performed using colloidal silica and a suede pad (made of urethane) to obtain a GaN substrate A having a thickness of 420 μm. The polishing time for the second chemical mechanical polishing was 25 hours as a result of adjusting according to the polishing rate. Scratch inspection was performed on the crystal surface after the second chemical mechanical polishing using a fluorescent microscope at an observation magnification of 200 times, but no scratch was observed.

(Manufacture of substrate B)
Polishing was performed in accordance with the substrate A, and the substrate was changed from a GaN substrate to a sapphire substrate to produce a 385 μm thick GaN substrate B. The polishing time for the second chemical mechanical polishing was 48 hours as a result of adjusting according to the polishing rate.
Moreover, although the crystal | crystallization surface after performing 2nd chemical mechanical polishing was scratch-inspected with the observation magnification of 200 times using the fluorescence microscope, the scratch was not observed.

(Measurement of grazing incidence X-ray diffraction)
With respect to the obtained substrates A and B, synchrotron radiation was used as an X-ray source, and X-rays were incident on the C-plane GaN crystal surface at a low angle. The wavelength of the X-ray used was measured by changing between 1.51 and 1.66 inches, and the number of measurement points was 23 to 40 points. For monochromatization and collimation of the X-ray beam, a Si (111) two-crystal spectrometer was used as a pre-crystal. The sample was set in a goniometer capable of ω rotation, in-plane rotation, and tilt rotation, and placed on the downstream side of the spectroscopic crystal. The incident angle of X-rays was set to an angle at which GaN (103) reflection was obtained. The rotational accuracy of each axis of the goniometer was ω rotation: 1/100 second, in-plane rotation: 1/100 degree, and tilt rotation: 1/100 degree.

(Measurement data analysis)
The full width at half maximum of the rocking curve for each measured wavelength was calculated. Furthermore, the X-ray penetration depth for each wavelength was calculated, and a depth-half-width curve was obtained. Except for special cases, these curves have a minimum point in the depth-half-width curve when the X-ray incident angle on the sample reaches the total reflection critical angle. The full width at half maximum at this minimum point was evaluated by actual measurement / theoretical value. For each sample, the half-value width at the minimum point and the actually measured value / theoretical value are shown in Table 1. It was found that by optimizing the polishing time, an ideal surface close to no distortion can be obtained.

<Comparative example>
(Manufacture of substrate C, D, GaN template)
A GaN substrate C having a thickness of 300 μm was manufactured by polishing according to the substrate A except that the first chemical mechanical polishing was not performed. The polishing time for the second chemical mechanical polishing was 9 hours as a result of adjusting according to the polishing rate. Moreover, although the crystal | crystallization surface after performing 2nd chemical mechanical polishing was scratch-inspected with the observation magnification of 200 times using the fluorescence microscope, the scratch was not observed.
Next, polishing was performed according to the substrate A except that the second chemical mechanical polishing was not performed, and a GaN substrate D having a thickness of 300 μm was manufactured. Further, the crystal surface after the second mechanical polishing was subjected to scratch inspection at an observation magnification of 200 times using a fluorescence microscope, but many scratches were observed.
Moreover, a GaN template was prepared by growing a GaN crystal having a thickness of several μm on a sapphire substrate by MOCDV method.

(Measurement of small-angle incident X-ray diffraction, analysis of measurement data)
The above measurement and analysis were similarly performed on the substrates C, D, and the GaN template. The number of measurement points on the rocking curve was 9 to 23 points. For each sample, the half-value width at the minimum point and the result of the actual measurement / theoretical value are shown in Table 1. However, since the GaN substrate D has no minimum value, the half-value width at the X-ray penetration depth of about 112 nm from the surface was taken as the measurement value.
These results are plotted on a graph and shown in FIG.

From Table 1, the value of (FWMH) ⁱ _m / (FWMH) ⁱ _t of the GaN substrate D in which a large number of scratches were observed in the fluorescence microscope observation was approximately compared with that of the substrates A to C in which no scratch was observed. It was about 7-8 times larger. That is, it can be seen that the value of (FWMH) ⁱ _m / (FWMH) ⁱ _t reflects the crystallinity near the crystal surface very well.
In particular, the value of the first chemical mechanical polishing of the GaN substrate C was not carried out _{^{(FWMH) i m / (FWMH}} ) i t is compared to that of GaN substrate A and B was carried out first chemical mechanical polishing, It was about 5% to 15% larger. In the GaN substrate C, no scratches were observed in the fluorescence microscope observation, but with the fact that the first chemical mechanical polishing was not performed, chipping occurred at the end of the crystal during the second mechanical polishing, and the chip caused the chip surface Scratches are generated, and the distortion generated in the vicinity of the crystal surface due to the scratches cannot be removed even by the subsequent second mechanical polishing, and remains as (FWMH) ⁱ _m / (FWMH) ⁱ _t The value is considered to be affected. In other words, the value of (FWMH) ⁱ _m / (FWMH) ⁱ _t is considered to be an index that reflects the crystallinity in the vicinity of the crystal surface, which cannot be grasped by observation with a fluorescence microscope, very well.

Next, the rocking curve measured for the GaN substrate A was fitted by the following procedure as a superposition of rocking curves of complete crystals obtained by a plurality of theoretical calculations.
i) First, a rocking curve of a complete crystal (theoretical calculation rocking curve) was obtained by theoretical calculation.
ii) Secondly, the first theoretically calculated rocking curve is placed at the angle (Δθ = 0) at the maximum value of the rocking curve obtained by measurement, and the high angle side (+ Δθ side) and the low angle side (−Δθ side) A total of about 10 to 20 pieces were arranged at regular intervals.
iii) Third, the strength of each theoretically calculated rocking curve was adjusted to best match the actually measured rocking curve. For these fitting calculations, commercially available scientific calculation software or the like was used. FIG. 4 shows an example of fitting a rocking curve at a depth of 330 nm with 13 theoretically calculated rocking curves. The strength of each rocking curve is standardized.
iv) Fourth, plot the intensity obtained by fitting against each Δθ to create a distribution curve graph, calculate the half-value widths of these, and ½ of these values as the variance value of Δθ did. FIG. 5 shows a graph created from the above fitting results.
v) Fifth, the data processing similar to i) to iv) was performed for each depth, and the dispersion value of Δθ against the depth was plotted. Table 2 shows the dispersion value of Δθ of the substrate A, and FIG. 6 shows a graph plotting the dispersion value of Δθ for the substrate A.

(Device evaluation)
An LED structure device was fabricated by MOCVD on a substrate prepared by the same method as that for substrate B and on a GaN template substrate. The PL emission wavelength was about 400 nm. Table 3 shows the emission intensity. The former device showed high emission intensity.

DESCRIPTION OF SYMBOLS 1 Radiation light source 2 Spectroscopic crystal 3 Ion chamber 4 Goniometer 5 Detector 10 Base substrate 100 Reactor 101 Gas introduction pipe 102 Gas introduction pipe 103 Gas introduction pipe 104 Gas introduction pipe 105 Reservoir 106 Heater 107 Substrate holder 108 Gas exhaust pipe

Claims (6)

For the periodic table group 13 metal nitride substrate, the minimum value of the half value width obtained when the half value width of the X-ray rocking curve is measured by changing the penetration depth from the substrate main surface in the substrate thickness direction ( FWMH) ⁱ _m ,
When the minimum value of the half width calculated from the dynamic diffraction theory of the group 13 metal nitride perfect crystal of the periodic table is defined as (FWMH) ⁱ _t ,
(FWMH) ⁱ _m / (FWMH) ⁱ _t ≦ 1.25
The periodic table group 13 metal nitride substrate satisfying   With respect to the periodic table group 13 metal nitride substrate, the half value width obtained when the half value width of the X-ray rocking curve is measured by changing the penetration depth from 500 nm to 1000 nm in the substrate thickness direction from the substrate main surface. The periodic table group 13 metal nitride substrate according to claim 1, wherein the average value is 200 arcsec or less.   An X-ray rocking curve measurement was performed on the periodic table group 13 metal nitride substrate at a penetration depth of 300 nm from the substrate main surface in the substrate thickness direction, and the dispersion of the obtained crystal lattice plane inclination Δθ was 0.050 °. The periodic table group 13 metal nitride substrate according to claim 1 or 2, wherein:   An X-ray rocking curve measurement at a penetration depth of 400 nm from the main surface of the periodic table group 13 metal nitride substrate in the thickness direction of the substrate was performed, and the dispersion of the obtained crystal lattice plane inclination Δθ was 0.040 °. The periodic table group 13 metal nitride substrate according to any one of claims 1 to 3, wherein:   An X-ray rocking curve measurement was performed at a penetration depth of 500 nm from the principal surface of the periodic table group 13 metal nitride substrate in the thickness direction of the substrate, and the dispersion of the obtained crystal lattice plane inclination Δθ was 0.030 °. The periodic table group 13 metal nitride substrate according to any one of claims 1 to 4, wherein:   For the periodic table group 13 metal nitride substrate, X-ray rocking curve measurement was performed by changing the penetration depth from 300 nm to 400 nm from the substrate main surface to the substrate thickness direction, the horizontal axis represents the penetration depth, and the vertical axis represents the penetration depth. 6. The periodic table according to claim 1, wherein the slope of the linear line formed when plotted as the dispersion of the inclination Δθ of the crystal lattice plane is −0.01 ° / 100 nm or more. Group metal nitride substrate.