JP2009102217A - MANUFACTURING METHOD OF GaN SUBSTRATE, MANUFACTURING METHOD OF EPITAXIAL WAFER, MANUFACTURING METHOD OF SEMICONDUCTOR ELEMENT AND EPITAXIAL WAFER - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a manufacturing method of a GaN substrate capable of obtaining an epitaxial wafer having less deflection than needed when preparing the epitaxial wafer and also capable of reducing a cost, to provide a manufacturing method of the epitaxial wafer, to provide a manufacturing method of a semiconductor element and to provide the epitaxial wafer. <P>SOLUTION: When the radius of the GaN substrate 10 is r (m), the thickness is t1 (m), the deflection of the GaN substrate before the epitaxial wafer 20 is formed is h1 (m), the thickness of an Al<SB>x</SB>Ga(1-x)N layer 21 is t2 (m), a deflection of the epitaxial wafer 20 is h2 (m), the lattice constant of the GaN is a1 and the lattice constant of AlN is a2, the value of the t1 obtained by (1.5×10<SP>11</SP>×t1<SP>3</SP>+1.2×10<SP>11</SP>×t2<SP>3</SP>)×(1/(1.5×10<SP>11</SP>×t1)+1/(1.2×10<SP>11</SP>×t2))/(15.96×x×(1-a2/a1))×(t1+t2)+(t1×t2)/(5.32×x×(1-a2/a1))-(r<SP>2</SP>+h<SP>2</SP>)/2h=0 is the minimum thickness of the GaN substrate and the GaN substrate having the thickness of more than the minimum thickness and less than 400 μm is formed. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

The present invention relates to a GaN substrate manufacturing method, an epi-wafer manufacturing method, a semiconductor device manufacturing method, and an epi-wafer, and in particular, has a c-plane, and an Al _x Ga _(1-x) N layer and a GaN layer on the c-plane. The present invention relates to a GaN substrate manufacturing method, an epiwafer manufacturing method, a semiconductor device manufacturing method, and an epiwafer for manufacturing an epiwafer by sequentially stacking at least two layers.

  Conventionally, a GaN (gallium nitride) substrate has been used as a substrate for a semiconductor element such as a light emitting diode (LED) or a laser diode (LD). Since GaN has an energy band gap of 3.4 eV and high thermal conductivity, when used as a substrate, an electrode can be provided on the back surface thereof, and the driving (operating) voltage of the semiconductor element can be reduced. .

Such a GaN substrate is manufactured as described in Non-Patent Document 1, for example. Non-Patent Document 1 discloses that a GaN substrate is manufactured by performing the following steps. That is, a buffer layer made of GaN with a thickness of 60 μm is formed on a GaAs (gallium arsenide) substrate by HVPE (Hydride Vapor Phase Epitaxy) method. Thereafter, a GaN layer having a thickness of 500 μm is formed on the buffer layer by HVPE. Thereafter, the GaAs substrate is removed, and a GaN substrate having a thickness of 495 ± 10 μm is obtained by polishing.
K. Motoki et al., "Preparation of large GaN substrates", Materials Science and Engineering B93 (2002), pp.123-125

  In Non-Patent Document 1, since a plurality of GaN substrates are not cut out in the thickness direction from one GaN bulk single crystal, there is a problem in that it takes cost to obtain one GaN substrate.

  In order to reduce the cost of a single GaN substrate, a technology for manufacturing a GaN substrate by manufacturing an ingot made of GaN having an increased thickness and cutting out a plurality of GaN substrates in the thickness direction from the ingot. Conceivable. When the thickness of the GaN substrate to be cut out is 495 ± 10 μm described in Non-Patent Document 1, the thickness is more than necessary depending on the warping performance required for a semiconductor device manufactured using this GaN substrate. There is. In this case, cost reduction is not sufficient.

  In order to further reduce the cost, it is desirable to cut out the GaN substrate with a thin thickness from the ingot, but the GaN substrate cut out with a thin thickness may be cracked by processing such as polishing. Even when cracking does not occur, if an epitaxial layer is formed on the GaN substrate, the warpage of the epi wafer provided with the GaN substrate and the epitaxial layer may increase. In this case, photolithography or the like cannot be performed when forming electrodes on the epiwafer, and the epiwafer cannot be used as a semiconductor element.

  Accordingly, the present invention is to provide a GaN substrate manufacturing method, an epiwafer manufacturing method, a semiconductor device manufacturing method, and an epiwafer which can be reduced in warpage or less when an epiwafer is formed.

The method for producing a GaN substrate of the present invention includes an Al _x Ga having a c-plane and having an Al composition x exceeding 0 and not more than 0.3 and having a thickness exceeding 0 and not more than 30 nm on the c-plane. _{(1-x) A} method of manufacturing a GaN substrate for manufacturing an epi-wafer by sequentially stacking at least two layers of an N layer and a GaN layer, and the following steps are performed. First, the thickness of the GaN substrate t1 (m), the radius of GaN substrate _{r (m), Al x Ga} (1-x) the thickness of the N layer _{t2, Al x Ga (1-} x) of Al in N layer When the composition is x, the warp of the epi-wafer is h (m), the lattice constant of GaN is a1, and the lattice constant of AlN is a2.
(1.5 × 10 ¹¹ × t1 ³ + 1.2 × 10 ¹¹ × t2 ³ ) × {1 / (1.5 × 10 ¹¹ × t1) + 1 / (1.2 × 10 ¹¹ × t2)} / {15 .96 × x × (1−a2 / a1)} × (t1 + t2) + (t1 × t2) / {5.32 × xx (1−a2 / a1)} − (r ² + h ² ) / 2h = 0 ... (Formula 1)
The value of t1 obtained by the above is determined as the minimum thickness of the GaN substrate. Then, a GaN substrate having a thickness not smaller than 400 μm is cut out from the ingot made of GaN.

  As a result of intensive studies on a method for determining the thickness of a GaN substrate used for an epi-wafer having the above structure, the present inventor has found the above formula 1. In other words, since the minimum thickness for satisfying the set epi-wafer warp h can be determined by the above formula 1, a GaN substrate obtained in the thickness direction by cutting out a GaN substrate from one ingot with a thickness greater than the minimum thickness and less than 400 μm. Can be increased. Therefore, it is possible to manufacture a GaN substrate that can be warped less than that required when an epi-wafer is formed, and that can reduce costs.

In the case of manufacturing a GaN substrate used in the epitaxial wafer comprising a plurality of _{Al x Ga (1-x)} N layer, the total thickness of the plurality of _{Al x Ga (1-x)} N layer and t2 . Further, in the plurality of Al _x Ga _(1-x) N layers, the composition ratio of Al in the Al _x Ga _(1-x) N layer that occupies the largest total thickness t2 is defined as a composition x. The composition x of Al is a molar ratio.

  Preferably, in the method of manufacturing the GaN substrate, in the cutting step, a GaN substrate having a thickness not less than the minimum thickness and not less than 100 μm and less than 250 μm is formed.

  When the minimum thickness is less than 100 μm, by making the thickness of the GaN substrate 100 μm or more, handling is easy and the number of GaN substrates in the thickness direction that can be manufactured from one ingot can be increased. When the minimum thickness is less than 100 μm, by making the thickness of the GaN substrate less than 250 μm, a GaN substrate that is easier to handle can be manufactured.

In the epiwafer manufacturing method in one aspect of the present invention, the following steps are performed. First, the GaN substrate is manufactured by the GaN substrate manufacturing method. An Al _x Ga _(1-x) N layer is formed on the c-plane of the GaN substrate. A GaN layer is formed on the Al _x Ga _(1-x) N layer.

According to the method for manufacturing an epi-wafer in one aspect of the present invention, even if an Al _x Ga _(1-x) N layer and a GaN layer are formed on a GaN substrate, the warp of the epi-wafer is less than h and the GaN is thin. An epi wafer having a substrate can be manufactured. Therefore, the manufactured epi-wafer can be used for a semiconductor element, and an epi-wafer with reduced cost can be manufactured by reducing the manufacturing cost of the GaN substrate.

In the epiwafer manufacturing method according to another aspect of the present invention, the following steps are performed. First, the GaN substrate is manufactured by the GaN substrate manufacturing method. A GaN layer is formed on the c-plane of the GaN substrate. An Al _x Ga _(1-x) N layer is formed on the GaN layer. A GaN layer is formed on the Al _x Ga _(1-x) N layer.

According to the method for manufacturing an epi-wafer in another aspect of the present invention, even if a GaN layer, an Al _x Ga _(1-x) N layer, and a GaN layer are formed in this order on a GaN substrate, the warp of the epi-wafer is h or less. Thus, an epi wafer having a thin GaN substrate can be manufactured. Therefore, the manufactured epi-wafer can be used for a semiconductor element, and an epi-wafer with reduced cost can be manufactured by reducing the manufacturing cost of the GaN substrate.

  In the method for manufacturing a semiconductor device of the present invention, the following steps are performed. First, an epi-wafer is manufactured by the epi-wafer manufacturing method. Then, an electrode is formed on the epi-wafer.

  According to the method for manufacturing a semiconductor device of the present invention, even when an electrode is formed, an epiwafer having a thin GaN substrate with an epiwafer warpage of h or less can be manufactured. Therefore, it is possible to manufacture a semiconductor element whose cost is reduced by reducing the manufacturing cost of the GaN substrate.

The epi-wafer of the present invention includes a GaN substrate, an Al _x Ga _(1-x) N layer formed on the c-plane of the GaN substrate, and a GaN layer formed on the Al _x Ga _(1-x) N layer. Is an epi-wafer. The thickness of the GaN substrate is less than 250 μm. The warp of the epi wafer is 100 μm or less.

  According to the epi-wafer of the present invention, the thickness of the GaN substrate that could not be achieved conventionally can be reduced to less than 250 μm, and an epi-wafer with a warp of 100 μm or less can be realized by the above formula 1.

  According to the GaN substrate manufacturing method, epiwafer manufacturing method, semiconductor element manufacturing method, and epiwafer of the present invention, it is possible to manufacture a GaN substrate that is less than the warp required when an epiwafer is formed and that can reduce costs.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following drawings, the same or corresponding parts are denoted by the same reference numerals, and description thereof will not be repeated.

(Embodiment 1)
FIG. 1 is a schematic sectional view showing an epi-wafer in the first embodiment of the present invention. As shown in FIG. 1, the epi-wafer 20 includes a GaN substrate 10, an Al _x Ga _(1-x) N layer 21, and a GaN layer 22. The GaN substrate 10 has a c-plane. The Al _x Ga _(1-x) N layer 21 is formed on the c-plane of the GaN substrate 10. The GaN layer 22 is formed on the Al _x Ga _(1-x) N layer 21.

  The thickness t1 of the GaN substrate 10 is less than 250 μm, and the warp of the epi-wafer 20 is 100 μm or less. More preferably, the thickness t1 of the GaN substrate 10 is not less than 100 μm and less than 250 μm, and the warp of the epi-wafer 20 is not less than 2 μm and not more than 85 μm. In particular, it is more preferable that the thickness t1 of the GaN substrate 10 is 120 μm or more and 240 μm or less, and the warp of the epi-wafer 20 is 2 μm or more and 50 μm or less. Within this range, the warpage performance required for the LED is satisfied, and the thickness tl of the GaN substrate is small. The thickness t1 of the thickest GaN substrate 10 calculated from the above equation 1 when the warp of the epi-wafer 20 is 12 μm is 250 μm. Similarly, the thickness t1 of the thickest GaN substrate 10 calculated from the above formula 1 when the warp of the epi-wafer 20 is 50 μm is 120 μm. That is, in the epi-wafer 20 in the present embodiment, the relationship between the thickness t1 of the GaN substrate 10 and the warp h of the epi-wafer 20 is equal to or less than the warp h of the epi-wafer 20 set by the above equation 1.

  When the GaN substrate 10 is warped, the thickness t1 of the GaN substrate 10 is the thickness of the GaN substrate 10 at a virtual line extending a line along the side surface including both ends of an arbitrary radius of the GaN substrate 10. .

The Al composition x of the Al _x Ga _(1-x) N layer 21 is not particularly limited as long as it exceeds 0 and is 0.30 or less, but is preferably 0.05 or more and 0.30 or less. More preferably, it is 18 or more and 0.30 or less. In this case, an epi-wafer 20 that is suitably used for a semiconductor element is obtained.

The thickness t2 of the Al _x Ga _(1-x) N layer 21 is not particularly limited as long as it exceeds 0 and is 30 nm, but is preferably 10 nm or more and 30 nm or less, and more preferably 20 nm or more and 30 nm or less. In this case, an epi-wafer 20 that is suitably used for a semiconductor element is obtained.

When the epi wafer 20 is warped, the thickness t2 of the Al _x Ga _(1-x) N layer 21 is an epi wafer at an imaginary line extending a line along a side surface including both ends of an arbitrary radius of the epi wafer 20. The thickness is 20.

The Al _x Ga _(1-x) N layer 21 may be a plurality of layers, and a GaN layer may be provided between the plurality of Al _x Ga _(1-x) N layers 21. When the Al _x Ga _(1-x) N layer 21 is composed of a plurality of layers, the total thickness is defined as the thickness t 2 of the Al _x Ga _(1-x) N layer 21. When the Al _x Ga _(1-x) N layer 21 is composed of a plurality of layers, the composition of Al constituting the Al _x Ga _(1-x) N layer that occupies the most thickness t2 is defined as the composition x. .

The thickness t3 of the GaN layer 22 is smaller than the thickness t1 of the GaN substrate 10. The GaN layer 22 may be a plurality of layers, and an Al _x Ga _(1-x) N layer may be further provided between the plurality of GaN layers.

  The radius r (see FIG. 7) of the GaN substrate 10 is preferably 25 mm or more. If it is 25 mm or more, it can be carried out at a cost commensurate with the mass production process.

  Next, with reference to FIG. 1 and FIG. 2, the manufacturing method of the epi-wafer in the present embodiment will be described. FIG. 2 is a flowchart showing an epi-wafer manufacturing method in the present embodiment.

As shown in FIG. 2, first, the thickness of the GaN substrate 10 is t1 (m), the radius of the GaN substrate 10 is r (m), the thickness of the Al _x Ga _(1-x) N layer 21 is t2 (m), When the warp of the epi-wafer 20 is h (m), the lattice constant of GaN is a1 (m), and the lattice constant of AlN is a2 (m),
1.5 × 10 ¹¹ × t1 ³ + 1.2 × 10 ¹¹ × t2 ³ ) × {1 / (1.5 × 10 ¹¹ × t1) + 1 / (1.2 × 10 ¹¹ × t2)} / {15. 96 × xx × (1−a2 / a1)} × (t1 + t2) + (t1 × t2) / {5.32 × xx (1−a2 / a1)} − (r ² + h ² ) / 2h = 0 ·・ ・ (Formula 1)
The value of t1 obtained by the above is determined as the minimum thickness of the GaN substrate (step S1).

The above-mentioned “GaN lattice constant a1” and “AlN lattice constant a2” mean lattice constants in a direction perpendicular to the c-axis. GaN and AlN are hexagonal. The hexagonal lattice constant includes lattice constants a1 and a2 in the direction perpendicular to the c-axis and lattice constants c1 and c2 in the c-axis direction. For GaN, the lattice constant in the direction perpendicular to the c-axis is a1 = 3.19 × 10 ⁻¹⁰ m (3.19 cm), and the lattice constant in the c-axis direction is c1 = 5.19 × 10 ⁻¹⁰ m (5. 19). For AlN, the lattice constant in the direction perpendicular to the c-axis is a2 = 3.11 × 10 ⁻¹⁰ m (3.11 cm), and the lattice constant in the c-axis direction is c2 = 4.98 × 10 ⁻¹⁰ m (4. 98Å).

  Specifically, first, the radius r of the GaN substrate 10 is obtained. For the radius r, an ingot made of GaN may be prepared, and the length to be the radius r of the GaN substrate 10 in this ingot may be measured. The radius r may be, for example, the radius of the base substrate when the ingot includes a base substrate and a layer made of GaN formed on the base substrate.

Further, the thickness t2 of the epi-wafer 20 to be manufactured, the warp h of the epi-wafer 20 (see FIG. 7), and the Al composition x constituting the Al _x Ga _(1-x) N layer 21 are set. The warp h of the epi-wafer 20 is set to a value that allows photolithography using a stepper or aligner, for example. As described above, a1 that is the lattice constant of GaN and a2 that is the lattice constant of AlN are known values. Then, by substituting r, t2, h, x, a1, and a2 into Equation 1, the thickness t1 of the GaN substrate 10 is obtained. This thickness t1 is the minimum thickness.

Here, when the Al _x Ga _(1-x) N layer 21 is composed of a plurality of layers, the total thickness is set as the thickness t 2 of the Al _x Ga _(1-x) N layer 21. When the Al _x Ga _(1-x) N layer 21 is composed of a plurality of layers, the composition of Al constituting the Al _x Ga _(1-x) N layer that occupies the most thickness t2 is set as the composition x. To do.

  Next, a GaN substrate having a thickness not less than the minimum thickness and less than 400 μm is cut out from the ingot made of GaN (step S2).

  Specifically, the thickness of the GaN substrate cut out from the prepared ingot is determined. Thereafter, the GaN substrate is cut out from the ingot by slicing or the like so that each GaN substrate has this thickness. The method for cutting out the GaN substrate is not particularly limited, and a generally known method can be adopted.

  The thickness of the GaN substrate cut out from the ingot is not less than the minimum thickness determined by the above formula 1 and not more than 400 μm, preferably not less than the minimum thickness and not less than 100 μm and less than 250 μm. When cutting out from the ingot with the minimum thickness determined by the above equation 1, a GaN substrate that can be used for an epi-wafer with a set warp h is obtained. When handling is difficult with this minimum thickness GaN substrate, it is preferable to cut out with a thickness exceeding the minimum thickness and capable of handling. In particular, handling is facilitated when the thickness of the GaN substrate to be cut out is 100 μm or more. If the thickness of the GaN substrate to be cut out is less than 400 μm, handling is very easy. When the thickness of the GaN substrate to be cut out is less than 250 μm, the number of pieces to be cut out can be increased and handling becomes easy.

  By performing the above steps, the GaN substrate 10 can be manufactured. The GaN substrate 10 manufactured by this GaN substrate manufacturing method may be further polished. When manufacturing the epi-wafer 20 shown in FIG. 1, the following steps are further performed.

Next, the Al _x Ga _(1-x) N layer 21 is formed on the c-plane of the GaN substrate 10, and the GaN layer 22 is formed on the Al _x Ga _(1-x) N layer 21 (step S3). . In this step S3, the Al _x Ga _(1-x) N layer 21 and the GaN layer 22 are formed by the epitaxial growth method. This growth method is not particularly limited. For example, a vapor phase method such as a sublimation method, HVPE method, MBE (Molecular Beam Epitaxy) method, MOCVD (Metal Organic Chemical Vapor Deposition) method, Examples include liquid phase method.

In this step S3, an Al _x Ga _(1-x) N layer 21 having an Al composition x exceeding 0 and 0.3 or less and a thickness exceeding 0 and 30 nm or less is formed. In this range, the above formula 1 is obtained on the assumption that the warp h is the largest in the epi-wafer 20 within this range, so that the set warp h of the epi-wafer 20 and the warp h of the epi-wafer 20 are satisfied. Therefore, the thickness of the GaN substrate 10 does not become too large. Therefore, the above formula 1 can be applied to the epi-wafer 20 having the structure in the above range.

When the Al _x Ga _(1-x) N layer 21 is composed of a plurality of layers, the total thickness thereof is set so that the thickness t2 of the Al _x Ga _(1-x) N layer 21 exceeds 0 and is 30 nm or less. In addition, the Al _x Ga _(1-x) N layer 21 is formed so that the composition x of the Al _x Ga _(1-x) N layer occupying the largest thickness t2 exceeds 0 and becomes 0.3 or less. The

By performing the above steps, the epi-wafer 20 shown in FIG. 1 is manufactured. After the epi wafer 20 is manufactured, the surface of the GaN substrate 10 opposite to the surface on which the Al _x Ga _(1-x) N layer 21 is formed may be further polished. The epi-wafer 20 manufactured in this way satisfies the warp h set before manufacturing, and is the thinnest in the range in which the thickness of the substrate 10 can be handled. Therefore, since the substrate 10 with a reduced manufacturing cost is used, the manufacturing cost of the epi wafer 20 can also be reduced.

(Embodiment 2)
FIG. 3 is a schematic cross-sectional view showing an epi-wafer in the second embodiment of the present invention. As shown in FIG. 3, the epi-wafer 30 in this embodiment includes a GaN substrate 31, a buffer layer 32, an active layer 33, an electron block layer 34, and a contact layer 35.

  FIG. 4 is a schematic cross-sectional view showing a semiconductor element according to the second embodiment of the present invention. The LED 40 as a semiconductor element in the second embodiment includes the epi-wafer 30 and electrodes 41 and 42 shown in FIG.

In the epi-wafer 30 and the LED 40, the thickness of the GaN substrate 31 is t31. A buffer layer 32 is formed on the c-plane of the GaN substrate 31. The buffer layer 32 is made of n-type GaN, for example, and has a thickness of t32. An active layer 33 is formed on the buffer layer 32. The thickness of the active layer 33 is t33. The active layer 33 has a multiple quantum well structure made of, for example, InGaN and GaN. The active layer 33 may be made of a single semiconductor material. An electron blocking layer 34 is formed on the active layer 33. The electron block layer 34 is made of, for example, p-type Al _x Ga _(1-x) N and has a thickness of t34. A contact layer 35 is formed on the electron blocking layer 34. The contact layer 35 is made of p-type GaN, for example, and has a thickness of t35.

  In the LED 40, an electrode 41 is formed on the surface of the GaN substrate 31 opposite to the surface on which the buffer layer 32 is formed. The electrode 41 is made of, for example, titanium and aluminum. An electrode 42 is formed on the contact layer 35. The electrode 42 is made of, for example, nickel and gold.

  Next, with reference to FIGS. 3 to 5, a method for manufacturing the GaN substrate 31, the epiwafer 30 and the LED 40 in the present embodiment will be described. FIG. 5 is a flowchart showing a method for manufacturing a semiconductor element in the present embodiment.

First, the minimum thickness of the GaN substrate 10 is determined using the above formula 1 (step S1). To determine the minimum thickness, to set the Al _x Ga _(1-x) the thickness of the N layer t2, Al _x Ga _(1-x) N layer of Al composition x, and epi-wafer warpage h in Formula 1.

In the present embodiment, since the GaN substrate 31 used for the epi-wafer 30 of FIG. 3 is manufactured, the thickness t2 of the Al _x Ga _(1-x) N layer in Equation 1 is t34. Since the active layer 33 if it contains _{Al x Ga (1-x)} N layer, _{Al x Ga (1-x)} N layer included in active layer 33 is very thin, active layer 33 The sum of the thickness of the Al _x Ga.sub. _(1-x) N layer and the thickness t34 of the electron block layer 34 and the thickness t34 of the electron block layer 34 are substantially the same value.

In Formula 1, the composition x of the Al _x Ga _(1-x) N layer is the composition x of the electron block layer 34. Here, even when the active layer 33 contains an Al _x Ga _(1-x) N layer, of the Al _x Ga _(1-x) overall thickness constituting the N layer t2, the most Since the occupying layer is the electron block layer 34, the Al composition x of the Al _x Ga _(1-x) N layer in Formula 1 is the Al composition x of the electron block layer 34.

  Further, the warp h required for the manufactured epi-wafer 30 is set. Since the epiwafer 30 is an epiwafer having an LED structure, it is, for example, 100 μm or less.

  Further, the radius r of the GaN substrate 31 is obtained. For example, as described above, the radius r can be obtained by measuring the radius of the base substrate used in the ingot.

The set Al _x Ga _(1-x) N layer thickness t 2, Al _x Ga _(1-x) N layer Al composition x, epi-wafer warpage h, and radius r are substituted into Equation 1 to obtain the GaN substrate 31. The minimum thickness t1 is obtained.

  Next, a GaN substrate 31 having a thickness not less than the minimum thickness and less than 400 μm is cut out from the ingot made of GaN (step S2). When the handling of the GaN substrate having the minimum thickness t1 obtained by Equation 1 is difficult, the thickness of the GaN substrate is changed to a thickness that can be handled with a thickness of 400 μm or less. When used for the epitaxial wafer 30 having an LED structure, the thickness of the GaN substrate is preferably determined to be not less than the minimum thickness and less than 250 μm. The GaN substrate 31 is manufactured by cutting out from the ingot with the determined thickness.

  Next, the buffer layer 32 is formed on the c-plane of the GaN substrate 31. Next, the active layer 33 is formed on the buffer layer 32. Next, the electron block layer 34 is formed on the active layer 33. Next, a contact layer 35 is formed on the electron block layer 34. The method for forming these epitaxial layers is not particularly limited, and is formed by, for example, the HVPE method or the MOCVD method.

  The epi-wafer shown in FIG. 3 is manufactured by performing the above steps. When manufacturing LED40, the following processes (step S4) are further implemented.

  Next, an electrode 41 is formed on the surface of the GaN substrate 31 opposite to the surface on which the buffer layer 32 is formed. An electrode 42 is formed on the contact layer 35. The method for forming the electrodes 41 and 42 is not particularly limited. For example, a mask is formed on the front and back surfaces of the epi-wafer 30 by photolithography, for example, a metal layer to be an electrode is formed by vapor deposition, and the electrodes 41 and 42 are formed by lift-off, for example.

The LED 40 shown in FIG. 4 is manufactured by performing the above steps.
Since other configurations are the same as those in the first embodiment, description thereof will not be repeated.

[Example]
Examples of the present invention will be described below. In the present embodiment, as shown in FIG. 1, a GaN substrate 10, an Al _x Ga _(1-x) N layer 21 formed on the c-plane of the GaN substrate 10, and an Al _x Ga _(1-x) N In order to find Equation 1 for determining the minimum thickness of the GaN substrate used in the epi-wafer 20 having three or more layers including the GaN layer 22 formed on the layer 21, the thickness t1 of the GaN substrate and the warp h of the epi-wafer I investigated the relationship with.

FIG. 6 is a schematic cross-sectional view showing an epi _- wafer composed of a GaN substrate and an Al _x Ga _(1-x) N layer formed on the GaN substrate. As shown in FIG. 6, in an epi-wafer 100 having a two-layer structure including a GaN substrate 10 and an Al _x Ga _(1-x) N layer 21 formed on the c-plane of the GaN substrate 10, the lattice constant of GaN Since this is different from the lattice constant of Al _x Ga _(1-x) N, it is known that the epitaxial wafer 100 having this two-layer structure is distorted. Specifically, since the lattice constant of Al _x Ga _(1-x) N is smaller than the lattice constant of GaN, tensile stress is applied to the Al _x Ga _(1-x) N layer 21 side, and the GaN substrate 10 side is subjected to tensile stress. Compressive stress is applied. As a result, as shown in FIG. 6, the epi-wafer 100 is warped to have a downward convex shape. With reference to FIG. 6, the relationship between the thickness of the GaN substrate 10 and the warp of the epi-wafer 20 for the epi-wafer 100 having the two-layer structure will be described. It is assumed that the Al _x Ga _(1-x) N layer 21 is coherently grown on the GaN substrate 10 with a thickness that does not cause lattice relaxation.

For the two-layer epi-wafer 100, the stress applied to the GaN substrate 10 and the Al _x Ga _(1-x) N layer 21 is P, the depth of the GaN crystal and the Al _x Ga _(1-x) N layer 21 is b, and the GaN substrate. the thickness of 10 to t1, the bending moment M1 of the GaN substrate 10, the Young's modulus of GaN substrate 10 E1, the radius of GaN substrate 10 r, the second moment of GaN substrate ^{10 I1 (= b × t1 3} /12 ), the Al _x Ga _(1-x) the thickness of the N layer 21 t2, Al _x Ga _(1-x) the bending moment of the N layer 21 M2, Al _x Ga _(1-x) Young's modulus of the N layer 21 _{E2, Al x Ga (1-} x) the bending moment of the N layer ^{21 I2 (= b × t2 3} /12), the bending moment of epitaxialwafer M, when the warp of epitaxialwafer 20 is h, these relationships, the following These are represented by formulas 2 to 4.

M = M1 + M2 = P × (t1 + t2) / 2 (Expression 2)
M1 = t1 × I1 / ρ = E1 × b × t1 ³ / (12 × ρ) (Expression 3)
M2 = t2 × I2 / ρ = E2 × b × t2 ³ / (12 × ρ) (Expression 4)

From the above formulas 2 to 4,
b × (E1 × t1 ³ + E2 × t2 ³ ) / (12 × ρ) = P × (t1 + t2) / 2 (Formula 5)
Is obtained.

From this equation 5,
P = b × (E1 × t1 ³ + E2 × t2 ³ ) / {6 × ρ × (t1 + t2)} (Expression 6)
Is obtained.

Next, the lattice mismatch strains of the GaN substrate 10 and the Al _x Ga _(1-x) N layer 21 are k × Δ and −k × Δ, respectively, where Δ is the lattice mismatch degree and k is the proportionality constant. is there. The strain due to the axial force of the GaN substrate 10 and the Al _x Ga _(1-x) N layer 21 is P / (E2 × b × t2) and −P / (E1 × b × t1), respectively. Further, the strain due to bending of the GaN substrate 10 and the Al _x Ga _(1-x) N layer 21 is t2 / (2 × ρ) and −t1 / (2 × ρ), respectively. The degree of lattice mismatch (Δ = −Δ) is x × (a2 / a1-1) where the lattice constant of GaN is a1 and the lattice constant of AlN is a2.

Since the strain at the interface between the GaN substrate 10 and the Al _x Ga _(1-x) N layer 21 needs to match, from the above relationship and the above equation 6,
k × Δ + P / (E2 × b × t2) + t2 / (2 × ρ) = − k × Δ−P / (E1 × b × t1) −t1 / (2 × ρ) (Expression 7)
Is obtained.

From this equation 7,
ρ = − (E1 × t1 ³ + E2 × t2 ³ ) × {1 / (E1 × t1) + 1 / (E2 × t2)} / {12k × Δ × (t1 + t2)} − (t1 + t2) / (4k × Δ) .... (Formula 8)
Is obtained.

  In addition, as shown in FIG. 6, θ is a half of an angle at which a virtual line obtained by extending a line along each side surface including an end of a certain diameter on the c-plane of the epi-wafer 100 intersects, and the length of this virtual line Where ρ is the epi-wafer warp and h is the following relationship.

h = ρ−ρ × cos θ (Equation 9)
r = ρ × sin θ (Expression 10)
Substituting Equation 9 and Equation 10 into sin ² θ + cos ² θ = 1,
(R / ρ) ² + (1−h / ρ) ² = 1 (Expression 11)
Is established.

  From the above, with respect to the two-layer epiwafer 100, the expressions showing the relationship between the thickness t1 of the GaN substrate 10 and the warp h of the epiwafer 100 are Expression 8 and Expression 11.

Next, as shown in FIGS. 1 and 7, the inventor formed a GaN substrate 10 having a c-plane and a c-plane, and the Al composition x exceeds 0 and is 0.3 or less. Three or more epitaxial wafers each including an Al _x Ga _(1-x) N layer 21 having a thickness of more than 0 and 30 nm or less and a GaN layer 22 formed on the Al _x Ga _(1-x) N layer 21 The relationship between the thickness of the GaN substrate 10 used for 20 and the epiwafer 20 was examined as follows. FIG. 7 is a schematic cross-sectional view showing a state in which the warp of the three-layer epi-wafer has occurred. That is, FIG. 7 shows a state in which warpage has occurred in the epi-wafer 20 of FIG.

First, the inventor of the present invention uses three or more layers of the epi-wafer 20 to form a layer formed on the Al _x Ga _(1-x) N layer 21 (the GaN layer 22 in FIGS. 1 and 7) on the GaN substrate 10. Since the thickness is very small compared to the above, it was considered that the equation 8 applied to the two-layer epiwafer was corrected and applied to the epi-wafer 20 having three or more layers.

Second, in 3 or more layers of epitaxialwafer 20, since the GaN layer 22 is formed on the _{Al x Ga (1-x)} N layer 21, the _{Al x Ga (1-x)} N layer 21 Then, strain caused by the GaN substrate 10 and strain caused by the GaN layer 22 are generated. On the other hand, in the two-layer epi-wafer, distortion occurs only on one surface. For this reason, the present inventor considered that the Young's modulus E2 of the Al _x Ga _(1-x) N layer 21 in Equation 8 applied to the two-layer epiwafer cannot be applied to the three or more epiwafers 20. Therefore, it was considered to replace the Young's modulus E2 in Formula 8 with another constant E.

Third, the proportionality constant k in Equation 8 is proportional to the difference in lattice constant between GaN and Al _x Ga _(1-x) N at the interface between the GaN substrate 10 and the Al _x Ga _(1-x) N layer 21. The value to be Therefore, when the GaN layer 22 is formed on the Al _x Ga _(1-x) N layer 21 in the epi-wafer 20 having three or more layers, the proportionality constant k considering only the difference in lattice constant from the GaN substrate 10 is It was considered that the method could not be applied to the epitaxial wafer 20 having three or more layers. Therefore, it was considered to replace the proportionality constant k in Equation 8 with another constant K.

  As described above, the following equation 12 was assumed by substituting E2 = E and k = K in the above equation 8.

ρ = − (E1 × t1 ³ + E × t2 ³ ) × {1 / (E1 × t1) + 1 / (E × t2)} / {12K × Δ × (t1 + t2)} − (t1 + t2) / (4K × Δ) ... (Formula 12)
In this example, the Al _x Ga _(1-x) N layer has an Al composition x exceeding 0 and not more than 0.3 and a thickness exceeding 0 and not more than 30 nm. In the two-layer epi-wafer 100, E2 and k are constants determined by the Al composition x of the Al _x Ga _(1-x) N layer. Further, in the Al _x Ga _(1-x) N layer of the epi-wafer, when the Al composition x exceeds 0 and is 1 or less and the thickness is not limited, the fluctuation range between E and K becomes large. In this embodiment, since the fluctuation range of E and K is considered to be substantially the same, in the Al _x Ga _(1-x) N layer of the epi-wafer 20, the Al composition x exceeds 0 and is 0.3 or less. In addition, the thickness exceeds 0 and is 30 nm or less.

  Next, in order to determine the values of E and K in Expression 12, the present inventor manufactured seven epiwafers 30 (samples 1 to 7) having the structure of the LED 40 shown in FIG. And the warp of the epi-wafer 30 were measured.

Specifically, for the LED wafer epi wafer 30 in the sample 1 shown in FIG. 3, first, a GaN substrate 31 having a thickness t31 of 350 μm and a radius r of 25 mm was prepared. Thereafter, a buffer layer 32 made of n-type GaN having a thickness t32 of 2.0 μm was formed on the GaN substrate 31 by MOCVD. Next, an active layer 33 made of InGaN / GaN 3QW (quantum well) having a thickness t33 of 0.05 μm was formed on the buffer layer 32 by MOCVD. Next, an electron blocking layer 34 having a thickness t34 of 0.02 μm and made of Al _0.18 Ga _0.82 N was formed on the active layer 33 by MOCVD. Next, a contact layer 35 made of p-type GaN having a thickness t35 of 0.2 μm was formed on the electron block layer 34 by MOCVD. Thereby, the epiwafer in the sample 1 was manufactured.

  The epi-wafer in the sample 2 has basically the same configuration as the epi-wafer in the sample 1, but differs only in that the thickness of the GaN substrate is 250 μm.

  The epi-wafer in the sample 3 has basically the same configuration as the epi-wafer in the sample 1, but differs only in that the thickness of the GaN substrate is 200 μm.

The epi-wafer in the sample 4 has basically the same configuration as the epi-wafer in the sample 1, except that the Al composition x in the Al _x Ga _(1-x) N layer is 0.30 and the Al _x Ga. _{(1-x) The} only difference is that the thickness of the N layer is 0.03 μm.

The epi-wafer in the sample 5 has basically the same configuration as the epi-wafer in the sample 1, but the Al composition in the Al _x Ga _(1-x) N layer is 0.30, and the Al _x Ga _{( 1-x) It} differs only in that the thickness of the N layer is 0.03 μm and the thickness of the GaN substrate is 250 μm.

The epi-wafer in the sample 6 has basically the same configuration as the epi-wafer in the sample 1, but the Al composition in the Al _x Ga _(1-x) N layer is 0.30, and the Al _x Ga _{( 1-x) It} differs only in that the thickness of the N layer is 0.03 μm and the thickness of the GaN substrate is 200 μm.

The epi-wafer in the sample 7 has basically the same configuration as the epi-wafer in the sample 1, but the Al composition in the Al _x Ga _(1-x) N layer is 0.30, and the Al _x Ga _{( 1-x) It} differs only in that the thickness of the N layer is 0.03 μm and the thickness of the GaN substrate is 150 μm.

The epiwafers of Sample 8 and Sample 9 were epiwafers not shown in FIG. FIG. 8 is a schematic cross-sectional view showing the epi-wafers of Sample 8 and Sample 9. Specifically, a GaN substrate 51 having a thickness t51 of 400 μm and a radius of 25 mm was prepared. Thereafter, an n-type Al _0.05 Ga _0.95 N layer 52 with a thickness t52 of 2 μm, a GaN layer 53 with a thickness t53 of 0.2 μm, and an InGaN / GaN 3QW layer with a thickness t54 of 0.05 μm on the GaN substrate 51 by MOCVD. 54, a p-type Al _0.18 Ga _0.82 N layer 55 having a thickness t55 of 0.02 μm, a GaN layer 56 having a thickness t56 of 0.2 μm, an Al _0.05 Ga _0.95 N layer 57 having a thickness t57 of 0.4 μm, and a thickness t58 of 0 A GaN layer 58 having a thickness of 2 μm was sequentially formed.

  And about the obtained epiwafer of samples 1-9, curvature was measured about five diameters and the average value of five curvatures was made into the curvature of the epiwafer of samples 1-9. The results are shown in Table 1.

For samples 1-7, the thickness t1 of the GaN substrate 10 was determined so that the warp h of the epiwafer was satisfied, and E and K were determined for these seven samples by the least square approximation method. As a result, it was found that E = 1.2 × 10 ¹¹ and K = 1.33.

In Equations 8 and 12, E1 is the Young's modulus of GaN, that is, E1 = 1.5 × 10 ¹¹ (GPa).

  Therefore, by substituting E and K into the above equation 12, the following equation 13 is obtained.

ρ = − (1.5 × 10 ¹¹ × t1 ³ + 1.2 × 10 ¹¹ × t2 ³ ) × {1 / (1.5 × 10 ¹¹ × t1) + 1 / (1.2 × 10 ¹¹ × t2)} /{12×1.33×Δ×(t1+t2)}−(t1+t2)/(4×1.33×Δ) (Equation 13)

Here, by substituting the equation of ρ = (x ² + h ² ) / 2h derived from Equation 11 and the degree of lattice mismatch Δ = −Δ = x × (a2 / a1-1) into Equation 13,
(1.5 × 10 ¹¹ × t1 ³ + 1.2 × 10 ¹¹ × t2 ³ ) × {1 / (1.5 × 10 ¹¹ × t1) + 1 / (1.2 × 10 ¹¹ × t2)} / {15 .96 × x × (1−a2 / a1)} × (t1 + t2) + (t1 × t2) / {5.32 × xx (1−a2 / a1)} − (r ² + h ² ) / 2h = 0 ... (Formula 1)
Is derived.

  Next, with reference to FIG. 9, the formula 1 and the formula of the two-layered epi-wafer relating to the epitaxial wafer having three or more layers of the present invention will be compared.

FIG. 9 is a diagram showing the relationship between the thickness t1 of the GaN substrate and the warp h of the epiwafer for the samples 1 to 9 derived from the equations 8 and 11 for the equations 1 and 2 of the epitaxial wafer of the present invention. . “Equation 1 of the present invention” in FIG. 9 indicates the warp h (m) of the epi-wafer when x = 0.18 in Equation 1 above. “Conventional Formula 1” is the proportional constant of the two-layered epi-wafer, E2 = 1.78 × 10 ¹¹ (GPa), which is the Young's modulus of x = 0.18 and Al _0.18 Ga _0.82 N in Formula 8 and Formula 11 above. The warp h (m) of the epi wafer when the proportionality constant k = 1.00 is shown. “Conventional equation 2” is the proportionality of E2 = 1.97 × 10 ¹¹ (GPa), Al _0.30 Ga _0.70 N which is the Young's modulus of x = 0.30 and Al _0.30 Ga _0.70 N in the above equations 8 and 11. The warp h (m) of the epi-wafer when the constant of proportionality k = 1.00 is shown. E1 in the formula 1 of the present invention and the conventional formulas 1 and 2 is the Young's modulus of GaN, that is, E1 = 1.5 × 10 ¹¹ .

As shown in Formula 1 of the present invention and Samples 1 to 7 in FIG. 9, the warp h of the epi wafer produced in Formula 1 of the present invention, the thickness t 2 of the epi wafer, and the Al composition x of the epi wafer are set. If the thickness is equal to or greater than the thickness of the GaN substrate 10 calculated by obtaining the radius r, the epi wafer on which the Al _x Ga _(1-x) N layer and the GaN layer are formed can always satisfy the set warp h. I understood. Further, it has been found that the thickness of the GaN substrate 10 for satisfying the set warp h of the epi wafer 20 and the warp h of the epi wafer 20 does not become too large.

Further, as shown in Formula 1 of the present invention and Samples 8 and 9 in FIG. 9, Formula 1 is applied to an epi-wafer having an Al _x Ga _(1-x) N layer that has a thickness of more than 0 and not less than 30 nm. It turns out that it is not appropriate to do. An epiwafer having an Al _x Ga _(1-x) N layer having an Al composition x exceeding 0 and not more than 0.3 and having a thickness exceeding 0 and not more than 30 nm is the epiwafer 20 in this range. Since it has been found that the above equation 1 is obtained so that the warp h becomes the largest, that is, x = 0.30 and t2 = 30 nm, the above equation 1 can be applied to the epi-wafer 20 having the structure in the above range. I understood.

  Therefore, it was found that Equation 1 found by examining E and K assuming Equation 12 is applicable to an epi-wafer having three or more layers having the above structure.

On the other hand, the thickness t1 (for example, 200 μm) of the GaN substrate when the epiwafer warpage h (for example, 6 μm) is set in the equation for the two-layered epiwafer at x = 0.18 (conventional expression 1 in FIG. 9) is the same epiwafer. The thickness was smaller than the thickness (for example, 250 μm) of the GaN substrates of Samples 1 and 2 where x = 0.18 satisfying the warp h (for example, 6 μm). That is, in the equation of the two-layer epiwafer at x = 0.18, the warp h of the epiwafer to be produced, the thickness t2 of the epiwafer, and the Al composition x of the epiwafer are set, and the Al _x Ga _(1-x) N to be produced is set. An epitaxial wafer in which a GaN substrate is manufactured with the thickness of the GaN substrate determined by substituting the Young's modulus, and an Al _x Ga _(1-x) N layer and a GaN layer are formed thereon does not satisfy the set warp h of the epi wafer. It was. Therefore, even when producing a GaN substrate 10 having a thickness determined by the expression for epitaxialwafer two layers, when manufacturing the epitaxial wafer using the GaN substrate, because they may not meet the warping performance, GaN substrate 10 and Al _x It has been found that the present invention is not applicable to an epitaxial wafer having three or more layers including the Ga _(1-x) N layer 21 and the GaN layer 22.

Further, in the equation for the two-layer epiwafer at x = 0.30 (conventional equation 2 in FIG. 9), the thickness t1 of the GaN substrate when the epiwafer warp h is set in the equation for the two-layer epiwafer is the same epiwafer. It was larger than the thickness of the GaN substrate of Samples 4 to 7 where x = 0.30 satisfying the warp h. That is, in the equation of the two-layer epiwafer at x = 0.30, the warp h of the epiwafer to be produced, the thickness t2 of the epiwafer, and the Al composition x of the epiwafer are set, and the Al _x Ga _(1-x) N to be produced is set. A GaN substrate having a thickness obtained by substituting the Young's modulus is manufactured, and an epi wafer on which an Al _x Ga _(1-x) N layer and a GaN layer are formed has a thickness tl of the GaN substrate obtained by Equation 1 of the present invention. It was bigger than. That is, it was found that the thickness of the GaN substrate obtained by the equation of the two-layer epiwafer at x = 0.30 is not an optimum equation for satisfying the set epiwafer warp h.

  As described above, according to the present embodiment, an equation for finding the minimum thickness of the GaN substrate for satisfying the warpage of the epiwafer set in the epiwafer having three or more layers was found. Further, it was confirmed that the minimum thickness satisfying the set epi-wafer warp can be determined according to Equation 1.

  The embodiments and examples disclosed above are to be considered in all respects as illustrative and not restrictive. The scope of the present invention is shown not by the above embodiments and examples but by the scope of claims, and is intended to include all modifications and variations within the meaning and scope equivalent to the scope of claims. .

The present invention is particularly advantageous for a method of manufacturing a GaN substrate having an c-plane and manufacturing an epiwafer by sequentially stacking an Al _x Ga _(1-x) N layer and a GaN layer on the c-plane. Can be applied to.

It is a schematic sectional drawing which shows the epi wafer in Embodiment 1 of this invention. It is a flowchart which shows the manufacturing method of the epi wafer in Embodiment 1 of this invention. It is a schematic sectional drawing which shows the epi wafer in Embodiment 2 of this invention. It is a schematic sectional drawing which shows the semiconductor element in Embodiment 2 of this invention. It is a flowchart which shows the manufacturing method of the semiconductor element in this Embodiment. Is a schematic sectional view showing an epitaxialwafer consisting of _{Al x Ga (1-x)} N layer formed on the GaN substrate and a GaN substrate. It is a schematic sectional drawing of the state which the curvature of the 3-layer epiwafer generate | occur | produced. 5 is a schematic cross-sectional view showing an epi-wafer of samples 8 and 9. FIG. It is a figure which shows the relationship between the thickness t1 of a GaN substrate, and the curvature h of an epiwafer about the formula about the formula 1 and 2 layer epiwafer of this invention, and the samples 1-9.

Explanation of symbols

10, 31 GaN substrate, 20, 30 epi-wafer, 21 Al _x Ga _(1-x) N layer, 22 GaN layer, 32 buffer layer, 33 active layer, 34 electron blocking layer, 35 contact layer, 40 LED, 41, 42 electrode.

Claims (5)

an Al _x Ga _(1-x) N layer having a c-plane and having an Al composition x of more than 0 and not more than 0.3 and a thickness of more than 0 and not more than 30 nm on the c-plane and GaN A method for producing a GaN substrate for producing an epi-wafer by sequentially laminating at least two layers with a layer,
The thickness of the GaN substrate is t1 (m), the radius of the GaN substrate is r (m), the thickness of the Al _x Ga _(1-x) N layer is t2 (m), and the Al _x Ga _(1-x) When the composition of Al in the N layer is x, the warp of the epi-wafer is h (m), the lattice constant of GaN is a1, and the lattice constant of AlN is a2.
(1.5 × 10 ¹¹ × t1 ³ + 1.2 × 10 ¹¹ × t2 ³ ) × {1 / (1.5 × 10 ¹¹ × t1) + 1 / (1.2 × 10 ¹¹ × t2)} / {15 .96 × xx (1-a2 / a1)} × (t1 + t2) + (t1 × t2) / {5.32 × xx (1-a2 / a1)} − (r ² + h ² ) / 2h = 0 ... (Formula 1)
Determining the value of t1 determined by the minimum thickness of the GaN substrate;
And a step of cutting out a GaN substrate having a thickness not less than 400 μm and not less than the minimum thickness from an ingot made of GaN.   2. The method of manufacturing a GaN substrate according to claim 1, wherein in the cutting step, the GaN substrate having a thickness equal to or greater than the minimum thickness and equal to or greater than 100 μm and less than 250 μm is formed. A step of producing a GaN substrate by the method of producing a GaN substrate according to claim 1 or 2,
Forming an Al _x Ga _(1-x) N layer on the c-plane of the GaN substrate;
And a step of forming a GaN layer on the Al _x Ga _(1-x) N layer. A step of manufacturing an epi-wafer by the method of manufacturing an epi-wafer according to claim 3;
And a step of forming an electrode on the epi-wafer. Epi wafer comprising a GaN substrate, an Al _x Ga _(1-x) N layer formed on the c-plane of the GaN substrate, and a GaN layer formed on the Al _x Ga _(1-x) N layer Because
The GaN substrate has a thickness of less than 250 μm;
An epi-wafer in which the warp of the epi-wafer is 100 μm or less.

Images