Silicon nitride substrate evaluation method, silicon nitride substrate evaluation device, silicon nitride substrate evaluation system, silicon nitride substrate manufacturing method, and power module manufacturing method

Abstract

To provide a silicon nitride substrate evaluation method which allows for evaluating dispersibility of rare earth oxide particles in a pre-sintered silicon nitride substrate.SOLUTION: A silicon nitride substrate evaluation method is provided, comprising: acquiring a Raman spectrum at each of multiple locations in an evaluation region occupying at least a portion of a pre-sintered silicon nitride substrate containing a rare earth oxide; acquiring a distribution of peak intensities caused by lattice vibration of the rare earth oxide in the evaluation region from the Raman spectra acquired at the multiple locations; and evaluating dispersibility of the rare earth oxide particles in the evaluation region based on the distribution.SELECTED DRAWING: Figure 6

Description

[Technical Field]

[0001] The present disclosure relates to a method for evaluating a silicon nitride substrate, an apparatus for evaluating a silicon nitride substrate, a system for evaluating a silicon nitride substrate, a method for manufacturing a silicon nitride substrate, and a method for manufacturing a power module. [Background technology]

[0002] Silicon nitride substrates are used as insulating substrates for power modules, etc. Examples of power modules include control members for motors in electric vehicles, hybrid electric vehicles, railway vehicles, industrial equipment, etc. Silicon nitride substrates are disclosed in Patent Document 1. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 6766509 Summary of the Invention [Problem to be solved by the invention]

[0004] To produce high-quality silicon nitride substrates, the silicon nitride substrates before sintering contain rare earth oxide particles such as Y2O3 particles as a sintering aid. If the dispersion of rare earth oxide particles in the silicon nitride substrate before sintering is poor, the silicon nitride substrate may warp after sintering. Therefore, it is necessary to evaluate the dispersion of rare earth oxide particles in the silicon nitride substrate before sintering. However, it has traditionally been difficult to evaluate the dispersion of rare earth oxide particles in the silicon nitride substrate before sintering.

[0005] In one aspect of the present disclosure, it is preferable to provide a method for evaluating a silicon nitride substrate that can evaluate the dispersibility of rare earth oxide particles in a silicon nitride substrate before sintering. [Means for solving the problem]

[0006] One aspect of the present disclosure is a method for evaluating a silicon nitride substrate, comprising: acquiring Raman spectra at multiple locations in an evaluation region that is at least a portion of an unsintered silicon nitride substrate containing a rare earth oxide; obtaining a distribution of peak intensities in the evaluation region that are due to lattice vibrations of the rare earth oxide based on the Raman spectra acquired at the multiple locations; and evaluating the dispersibility of rare earth oxide particles in the evaluation region based on the distribution.

[0007] According to a method for evaluating a silicon nitride substrate that is one aspect of the present disclosure, the dispersibility of rare earth oxide particles in a silicon nitride substrate before sintering can be evaluated. Another aspect of the present disclosure is an evaluation device for a silicon nitride substrate, comprising: a data acquisition unit configured to acquire, from a Raman measurement device, Raman spectra measured at multiple locations in an evaluation region, which is at least a portion of an unsintered silicon nitride substrate containing a rare earth oxide; a distribution acquisition unit configured to acquire, based on the Raman spectra acquired by the data acquisition unit, a distribution in the evaluation region of peak intensities due to lattice vibrations of the rare earth oxide; and an evaluation unit configured to evaluate the dispersibility of rare earth oxide particles in the evaluation region based on the distribution.

[0008] The silicon nitride substrate evaluation device according to another aspect of the present disclosure can evaluate the dispersibility of rare earth oxide particles in a silicon nitride substrate before sintering. Another aspect of the present disclosure is a method for manufacturing a silicon nitride substrate, which comprises stacking a plurality of pre-sintered silicon nitride substrates containing a rare earth oxide and sintering the stacked silicon nitride substrates, wherein 90% or more of the stacked pre-sintered silicon nitride substrates have a dispersion value Y of 0.5 or less as measured by the following measurement method.

[0009] Measurement method: Raman spectra are obtained at multiple locations in an evaluation area, which is at least a portion of an unsintered silicon nitride substrate containing rare earth oxide. Based on the Raman spectra obtained at the multiple locations, a distribution of peak intensities in the evaluation area due to lattice vibrations of the rare earth oxide is obtained. Rare earth oxide particles are detected in the evaluation area based on the distribution. Voronoi polygons are created for each of the detected rare earth oxide particles. The average value c of the areas of the Voronoi polygons is calculated. The standard deviation d of the areas of the Voronoi polygons is calculated. d / c is taken as the variance value Y.

[0010] According to a method for manufacturing a silicon nitride substrate that is another aspect of the present disclosure, a silicon nitride substrate with reduced warpage can be manufactured. [Brief explanation of the drawings]

[0011] [Figure 1] FIG. 2 is a side view showing the configuration of a power module and a silicon nitride circuit substrate. [Figure 2] 1A to 1C are explanatory diagrams showing a method for manufacturing a silicon nitride substrate. [Figure 3] 1A to 1C are explanatory diagrams showing a method for manufacturing a silicon nitride circuit substrate. [Figure 4] FIG. 1 is an explanatory diagram illustrating the configuration of an evaluation system for a silicon nitride substrate. [Figure 5] FIG. 5A is an explanatory diagram showing the binarized intensity distribution obtained from the unsintered silicon nitride substrate belonging to the first lot, and FIG. 5B is an explanatory diagram showing the binarized intensity distribution obtained from the unsintered silicon nitride substrate belonging to the second lot. [Figure 6] FIG. 6A is an explanatory diagram showing the section division method performed on unsintered silicon nitride substrates belonging to the first lot, and FIG. 6B is an explanatory diagram showing the section division method performed on unsintered silicon nitride substrates belonging to the second lot. [Figure 7] FIG. 7A is an explanatory diagram showing the Voronoi polygon method performed on an unsintered silicon nitride substrate belonging to a first lot, and FIG. 7B is an explanatory diagram showing the Voronoi polygon method performed on an unsintered silicon nitride substrate belonging to a second lot. [Figure 8]FIG. 8A is an explanatory diagram showing a method for forming a Delaunay triangulation network with Y2O3 particles as vertices in a binarized intensity distribution, FIG. 8B is an explanatory diagram showing a method for creating Voronoi polygons based on the Delaunay triangulation network, and FIG. 8C is an explanatory diagram showing the created Voronoi polygons. [Figure 9] 1 is a graph showing the measurement results of the aggregation value X and the dispersion value Y. DETAILED DESCRIPTION OF THE INVENTION

[0012] Exemplary embodiments of the present disclosure will now be described with reference to the drawings. 1. Configuration of power module 1 and silicon nitride circuit board 2 The configurations of a power module 1 and a silicon nitride circuit board 2 will be described with reference to Figure 1. The silicon nitride circuit board 2 includes a silicon nitride substrate 3, a metal circuit 5, a metal heat sink 7, and brazing layers 9 and 11. The power module 1 includes the silicon nitride circuit board 2, a semiconductor chip 13, and a heat sink 15.

[0013] The metal circuit 5 is made of a copper plate. The metal circuit 5 is attached to one side of the silicon nitride substrate 3 by a brazing material layer 9. The metal heat sink 7 is made of a copper plate. The metal heat sink 7 is attached to the other side of the silicon nitride substrate 3 by a brazing material layer 11. The semiconductor chip 13 is attached to the metal circuit 5. The heat sink 15 is attached to the metal heat sink 7.

[0014] 2. Method for manufacturing silicon nitride substrate 3 A method for manufacturing a silicon nitride substrate 3 will be described with reference to FIG. 2. In step (a), raw materials for the silicon nitride substrate 3 are prepared and mixed. The raw materials include silicon nitride powder, a sintering aid, and an organic binder. The sintering aid includes Mg compound powder and Y2O3 powder. The compounding ratio of the silicon nitride powder in the raw materials is, for example, 80 to 98 mass %. The compounding ratio of the Y2O3 powder in the raw materials is, for example, 0.2 to 2.0 mass %. The raw materials are in the form of a slurry. The solid concentration of the slurry is, for example, 30 to 70 mass %. Y2O3 corresponds to a rare earth oxide.

[0015] Next, in step (b), the raw material is formed into sheets, for example, by a doctor blade method, and cut. As a result, sheets 101 are obtained. Next, in step (c), a release agent is printed on the sheets 101. Next, in step (d), multiple sheets 101 are stacked and degreased in the atmosphere. The temperature during degreasing is, for example, 900°C or less, preferably 400 to 800°C. Next, in step (e), the stacked multiple sheets 101 are separated into individual sheets. The sheets 101 separated in step (e) will be referred to as pre-sintered silicon nitride substrates 103 hereinafter.

[0016] Each of the silicon nitride substrates 103 before sintering is subjected to the evaluation method for silicon nitride substrates to calculate the cohesion value X and dispersion value Y. The evaluation method for silicon nitride substrates will be described later. Next, in step (f), a plurality of pre-sintered silicon nitride substrates 103 are stacked. At this time, the pre-sintered silicon nitride substrates 103 to be stacked are selected so that 90% or more of the pre-sintered silicon nitride substrates 103 to be stacked have an agglomeration value X of 0.56 or less and a variance value Y of 0.5 or less. It is preferable to select the pre-sintered silicon nitride substrates 103 to be stacked so that all of the pre-sintered silicon nitride substrates 103 to be stacked have an agglomeration value X of 0.56 or less and a variance value Y of 0.5 or less. Furthermore, the stacked pre-sintered silicon nitride substrates 103 are sintered. By sintering, the pre-sintered silicon nitride substrates 103 become silicon nitride substrates 3. The sintering temperature is, for example, 1600 to 2000°C. Next, in step (g), the stacked plurality of silicon nitride substrates 3 are separated into individual substrates.

[0017] When the agglomeration value X of the pre-sintered silicon nitride substrate 103 is 0.56 or less and the dispersion value Y is 0.5 or less, the silicon nitride substrate 3 obtained by sintering the pre-sintered silicon nitride substrate 103 is resistant to warping. In the above manufacturing method, 90% or more of the pre-sintered silicon nitride substrates 103 to be stacked have an agglomeration value X of 0.56 or less and a dispersion value Y of 0.5 or less, and therefore the silicon nitride substrate 3 manufactured is resistant to warping.

[0018] In step (f), the pre-sintered silicon nitride substrates 103 to be stacked may be selected so that 90% or more of the pre-sintered silicon nitride substrates 103 to be stacked have an agglomeration value X of 0.56 or less. In this case, the silicon nitride substrate 3 to be manufactured is also less likely to warp.

[0019] In step (f), the pre-sintered silicon nitride substrates 103 to be stacked may be selected so that 90% or more of the pre-sintered silicon nitride substrates 103 to be stacked have a dispersion value Y of 0.5 or less. In this case, the silicon nitride substrate 3 to be manufactured is also less likely to warp.

[0020] In step (f), the pre-sintered silicon nitride substrates 103 to be stacked may be selected so that 90% or more of the pre-sintered silicon nitride substrates 103 to be stacked have an aggregation value X of 0.56 or less or a dispersion value Y of 0.5 or less. In this case, the silicon nitride substrate 3 to be manufactured is also less likely to warp. Furthermore, for example, the conditions of one or more of steps (a) to (d) may be changed based on the value of the agglutination value X or the variance value Y, thereby decreasing the value of the agglutination value X or the variance value Y.

[0021] 3. Manufacturing method of power module 1 and silicon nitride circuit board 2 The manufacturing method of the power module 1 and the silicon nitride circuit board 2 will be described with reference to Figure 3. In step (h), a metal plate 105 and a metal heat sink 107 are attached to the silicon nitride substrate 3 by brazing. Next, in step (i), a part of the metal plate 105 is removed to form a metal circuit 5. Next, in step (j), a plurality of silicon nitride circuit boards 2 are obtained by dividing. Furthermore, a semiconductor chip 13 and a heat sink 15 are attached to the silicon nitride circuit board 2. The silicon nitride substrate 3 included in the silicon nitride circuit board 2 is manufactured by the above-mentioned "2. Manufacturing method of silicon nitride substrate 3", and is therefore resistant to warping.

[0022] 4. Configuration of Evaluation System 201 The configuration of the evaluation system 201 will be described with reference to Fig. 4. The evaluation system 201 is used to evaluate the silicon nitride substrate 103 before sintering. The evaluation system 201 includes a Raman measurement device 203 and an evaluation device 205.

[0023] The Raman measurement device 203 irradiates a measurement object with laser light 206 and detects Raman scattered light 208 generated by the measurement object. Thus, the Raman measurement device 203 can measure the Raman spectrum of the measurement object. The irradiation diameter of the laser light 206 is approximately 1 μm. The irradiation diameter is a diameter. Therefore, the Raman measurement device 203 can measure the Raman spectrum of a narrow region.

[0024] The Raman measurement device 203 can move the irradiation position of the laser light 206 independently in the x direction and the y direction. The x direction and the y direction are both parallel to the measurement surface. The x direction is perpendicular to the y direction. Therefore, the Raman measurement device 203 can measure Raman spectra at multiple locations on the measurement object.

[0025] The evaluation device 205 includes a microcomputer having a CPU and a semiconductor memory such as a RAM or a ROM. Each function of the evaluation device 205 is realized by the CPU executing a program stored in a non-transitory tangible recording medium. Furthermore, by executing this program, a method corresponding to the program is performed.

[0026] The evaluation device 205 includes a data acquisition unit 207, a distribution acquisition unit 209, and an evaluation unit 211. The data acquisition unit 207 acquires a Raman spectrum from the Raman measurement device 203. The distribution acquisition unit 209 acquires an intensity distribution (described later) based on the Raman spectrum acquired by the data acquisition unit 207. The evaluation unit 211 evaluates the aggregation and dispersibility of Y2O3 particles in an evaluation region 301 (described later) based on the intensity distribution. Aggregation refers to a dense collection of multiple Y2O3 particles. Aggregation is a characteristic that indicates the degree of aggregation. Dispersibility is a characteristic that indicates how uniformly Y2O3 particles are dispersed in the evaluation region 301.

[0027] 5. Evaluation method for silicon nitride substrates A method for evaluating a silicon nitride substrate will be described with reference to FIGS. 4 to 8. The method for evaluating a silicon nitride substrate is carried out, for example, using an evaluation system 201. As shown in FIG. 4, a Raman measurement device 203 is used to acquire Raman spectra at each of a plurality of locations 303 in an evaluation region 301. The evaluation region 301 is at least a portion of the silicon nitride substrate 103 before sintering. The evaluation region 301 may be a portion of the silicon nitride substrate 103 before sintering, or the entire silicon nitride substrate 103 before sintering. The plurality of locations 303 are arranged at intervals of 0.4 to 1 μm, for example. The plurality of locations 303 are arranged, for example, at equal intervals across the entire evaluation region 301.

[0028] Next, the data acquisition unit 207 of the evaluation device 205 acquires a Raman spectrum from the Raman measurement device 203. Next, the distribution acquisition unit 209 acquires an intensity distribution based on the Raman spectrum acquired by the data acquisition unit 207. The intensity distribution is a distribution of peak intensities caused by lattice vibration of Y2O3 in the evaluation region 301. For example, the intensity distribution is data in which the intensity of the peak caused by lattice vibration of Y2O3 is associated with each of a plurality of locations 303.

[0029] In the Raman spectrum, the peak due to the lattice vibration of rare earth oxides is, for example, 355 cm -1 More than 390cm -1In the Raman spectrum, the peak due to the lattice vibration of Y2O3 is at about 374 cm -1 In the Raman spectrum, the peak due to the lattice vibration of Gd2O3 is at approximately 360 cm -1 In the Raman spectrum, the peak due to the lattice vibration of Er2O3 is at about 380 cm -1 is located. The peak intensity is, for example, the peak height or the peak area. In this embodiment, the peak intensity is determined from the peak area. The peak area is the integral value of the peak height at the full width at half maximum of the peak.

[0030] Next, based on the intensity distribution, the evaluation unit 211 detects Y2O3 particles in the evaluation region 301. The method for detecting Y2O3 particles is, for example, as follows. The evaluation unit 211 collects the intensities of peaks caused by lattice vibrations of Y2O3 from each of the multiple locations 303 and creates a collection of peak intensities. Based on this collection, the evaluation unit 211 creates a graph in which the horizontal axis represents the peak intensity and the vertical axis represents the frequency of elements in the collection. The evaluation unit 211 sets the intensity of the peak with the highest frequency as a threshold. The evaluation unit 211 binarizes the intensity distribution using this threshold. The binarized intensity distribution (hereinafter referred to as binarized intensity distribution 305) is shown in Figures 5A and 5B. The threshold may also be determined by other methods. For example, the threshold may be a preset value.

[0031] 5A and 5B, the white portions are portions of the plurality of locations 303 where the intensity of the peak due to the lattice vibration of Y2O3 is higher than the threshold. The black portions are portions of the plurality of locations 303 where the intensity of the peak due to the lattice vibration of Y2O3 is equal to or lower than the threshold. The evaluation unit 211 determines that Y2O3 particles are present in portions of the plurality of locations 303 where the intensity of the peak due to the lattice vibration of Y2O3 is higher than the threshold.

[0032] FIG. 5A shows a binarized intensity distribution 305 obtained from the silicon nitride substrate 103 before sintering that belongs to the first lot. The first lot and a second lot, which will be described later, are production lots of the silicon nitride substrate 103 before sintering. The first lot is a lot in which the silicon nitride substrate 3 warped little. FIG. 5B shows a binarized intensity distribution 305 obtained from the silicon nitride substrate 103 before sintering that belongs to the second lot. The second lot is a lot in which the silicon nitride substrate 3 warped often.

[0033] The silicon nitride substrates 3 were evaluated for warpage, and those meeting a certain standard were deemed to be acceptable products. The percentage of acceptable products was 96% for the first lot and 87% for the second lot. Next, the evaluation unit 211 calculates the aggregation value X by the interval division method. The calculation method of the aggregation value X is as follows. As shown in Figures 6A and 6B, the evaluation unit 211 divides the binarized intensity distribution 305 into a plurality of intervals of equal area. Note that since the area of ​​the binarized intensity distribution 305 is equal to the evaluation area 301, the evaluation area 301 is divided into a plurality of intervals of equal area. The shape of each interval was rectangular. There were 49 intervals. The area of ​​one interval was 84.6 μm 2 Fig. 6A corresponds to the silicon nitride substrate 103 before sintering that belongs to the first lot, and Fig. 6B corresponds to the silicon nitride substrate 103 before sintering that belongs to the second lot.

[0034] The evaluation unit 211 calculates the number of Y2O3 particles present in each section. The method for calculating the number of Y2O3 particles is as follows. The binarized intensity distribution 305 is a collection of pixels measuring 0.9 μm in height and 0.9 μm in width. The evaluation unit 211 divides the white portion of the binarized intensity distribution 305 into multiple points. Each divided point has a size of 0.9 μm in height and 0.9 μm in width. The evaluation unit 211 recognizes each divided point as a single Y2O3 particle. Next, the evaluation unit 211 calculates the number of recognized Y2O3 particles for each section. The particle size of the Y2O3 powder contained in the raw material of the silicon nitride substrate 3 is approximately 1 μm. Therefore, the size of the Y2O3 particles recognized as above is approximately the same as the particle size of the Y2O3 powder. Furthermore, the evaluation unit 211 calculates the average value a of the number of Y2O3 particles present in one section and the standard deviation b of the number of Y2O3 particles present in one section. For example, if the number of divided sections is m and the numbers of Y2O3 particles present in each section are n1, n2, ... n m a is n1, n2,...n m b is the average value of n1, n2, n m Furthermore, the evaluation unit 211 calculates the aggregation value X, which is b / a.

[0035] The agglomeration value X is a numerical value that represents the degree of variation in the number of Y2O3 particles belonging to each section. The larger the agglomeration value X, the greater the degree of variation in the number of Y2O3 particles belonging to each section. The agglomeration value X is a numerical value that represents the agglomeration of Y2O3 particles. The larger the agglomeration value X, the more agglomerated the Y2O3 particles are. Calculating the agglomeration value X corresponds to evaluating the agglomeration of Y2O3 particles in the evaluation area 301 based on the intensity distribution.

[0036] The measurement results of the cohesion value X are shown in Figure 9. For the unsintered silicon nitride substrate 103 belonging to the first lot, the cohesion value X was 0.457. For the unsintered silicon nitride substrate 103 belonging to the second lot, the cohesion value X was 0.649. This measurement result indicates that when the cohesion value X is large, the silicon nitride substrate 3 is more likely to warp.

[0037] Next, the evaluation unit 211 calculates the variance value Y by the Voronoi polygon method. The method for calculating the variance value Y is as follows. The evaluation unit 211 creates a Voronoi polygon for each of the detected Y2O3 particles, as shown in Figures 7A and 7B. Figure 7A corresponds to the unsintered silicon nitride substrate 103 belonging to the first lot, and Figure 7B corresponds to the unsintered silicon nitride substrate 103 belonging to the second lot.

[0038] Voronoi polygon π i is a set of points x expressed by the following formula (1). Formula (1) π i ={x;lx-x i l≦lx-xj l, i ≠ j, j = 1, …, N} Here, N is the number of Y2O3 particles present in the evaluation area 301. i is a number assigned to the Y2O3 particles. i is any natural number between 1 and N. x i is the coordinate of the Y2O3 particle numbered i. j are the coordinates of the Y2O3 particle numbered j. i is a convex polygon. Each Voronoi polygon π i There is one Y2O3 particle in the evaluation area 301. The evaluation area 301 is divided into N Voronoi polygons. Any point in the evaluation area 301 belongs to one of the Voronoi polygons.

[0039] Voronoi polygons can be created as follows. As shown in Figure 8A, a Delaunay triangular network is formed in the binarized intensity distribution 305, with Y2O3 particles 307 as vertices. Next, as shown in Figure 8B, perpendicular bisectors are created for each side of the triangle. The polygons created by the perpendicular bisectors are the Voronoi polygons shown in Figure 8C.

[0040] The evaluation unit 211 calculates the area of ​​each Voronoi polygon. The evaluation unit 211 calculates the average value c of the areas of the Voronoi polygons. The evaluation unit 211 calculates the standard deviation d of the areas of the Voronoi polygons. For example, if there are N Voronoi polygons and their areas are S1, S2, ...S N Let c be S1, S2, S N d is the average value of S1, S2, S N The evaluation unit 211 sets d / c as the variance Y.

[0041] The variance value Y is a numerical value that represents the degree of variation in the areas of the Voronoi polygons. The larger the variance value Y, the greater the degree of variation in the areas of the Voronoi polygons. The variance value Y is a numerical value that represents the dispersibility of the Y2O3 particles. The smaller the variance value Y, the more uniformly the Y2O3 particles are dispersed. Calculating the variance value Y corresponds to evaluating the dispersibility of the Y2O3 particles in the evaluation area 301 based on the intensity distribution.

[0042] The measurement results of the variance value Y are shown in Figure 9. For the unsintered silicon nitride substrates 103 belonging to the first lot, the variance value Y was 0.375. For the unsintered silicon nitride substrates 103 belonging to the second lot, the variance value Y was 0.619. This measurement result indicates that when the variance value Y is large, the silicon nitride substrate 3 is more likely to warp.

[0043] 6. Benefits of evaluation methods for silicon nitride substrates (1) According to the evaluation method of the silicon nitride substrate of the present disclosure, it is possible to evaluate the aggregation and dispersibility of Y2O3 particles in the silicon nitride substrate 103 before sintering. For example, the manufacturing conditions of the silicon nitride substrate can be changed so that the evaluation results of the aggregation and dispersibility are improved.

[0044] (2) The evaluation method for silicon nitride substrates of the present disclosure uses the intensity of the peak due to the lattice vibration of Y2O3, which makes it possible to suppress the influence of Y compounds and oxides other than Y2O3. (3) In the evaluation method for silicon nitride substrates of the present disclosure, the cohesion value X is calculated by the interval division method, which allows for more accurate evaluation of the cohesion of Y2O3 particles.

[0045] (4) In the evaluation method for silicon nitride substrates of the present disclosure, the variance value Y is calculated by the Voronoi polygon method, which allows for more accurate evaluation of the dispersibility of Y2O3 particles. 7. Other Embodiments Although the embodiments of the present disclosure have been described above, the present disclosure is not limited to the above-described embodiments and can be implemented in various modified forms.

[0046] (1) In the evaluation method for silicon nitride substrates of the present disclosure, only one of the cohesion value X and the dispersion value Y may be calculated. In the evaluation method for silicon nitride substrates of the present disclosure, only one of the cohesion property and the dispersion property may be evaluated. (2) The numerical value representing the cohesion may be a numerical value other than the cohesion value X. The numerical value representing the dispersibility may be a numerical value other than the dispersion value Y.

[0047] (3) The silicon nitride substrate 3 may be used to manufacture a power module other than the power module 1. (4) Instead of or in addition to Y2O3, other rare earth oxides may be used. Examples of other rare earth oxides include erbium oxide and gadolinium oxide. The above-described effects can also be achieved when other rare earth oxides are used. (5) Multiple functions of one component in the above embodiments may be realized by multiple components, or one function of one component may be realized by multiple components. Also, multiple functions of multiple components may be realized by one component, or one function realized by multiple components may be realized by one component. Also, part of the configuration of the above embodiments may be omitted. Also, at least part of the configuration of the above embodiments may be added to or substituted for the configuration of another of the above embodiments.

[0048] (6) In addition to the above-described evaluation system 201, the present disclosure can also be realized in various forms, such as a higher-level system that includes the evaluation system 201 as a component, a program for causing a computer to function as the evaluation device 205, a non-transitory tangible recording medium such as a semiconductor memory on which the program is recorded, and a quality control method for silicon nitride substrates. [Explanation of symbols]

[0049] 1...power module, 2...silicon nitride circuit board, 3...silicon nitride substrate, 5...metal circuit, 7...metal heat sink, 9, 11...brazing material layer, 13...semiconductor chip, 15...heat sink, 101...sheet, 103...silicon nitride substrate before sintering, 105...metal plate, 201...evaluation system, 203...Raman measurement device, 205...evaluation device, 206...laser light, 207...data acquisition unit, 208...Raman scattered light, 209...distribution acquisition unit, 211...evaluation unit, 301...evaluation area, 303...location, 305...binarized intensity distribution, 307...Y2O3 particles

Claims

1. Raman spectra are obtained at a plurality of locations in an evaluation region that is at least a portion of an unsintered silicon nitride substrate containing a rare earth oxide; Based on the Raman spectra acquired at the plurality of locations, a distribution in the evaluation region is acquired in which the intensity of a peak due to lattice vibration of a rare earth oxide is associated with each of the plurality of locations; evaluating the dispersibility of the rare earth oxide particles in the evaluation region based on the distribution; Evaluation method for silicon nitride substrates.

2. 2. A method for evaluating a silicon nitride substrate according to claim 1, comprising: The peak due to the lattice vibration of the rare earth oxide is 355 cm in the Raman spectrum. －１ More than 390cm －１ Within the following range: Evaluation method for silicon nitride substrates.

3. 3. A method for evaluating a silicon nitride substrate according to claim 1 or 2, comprising: The intensity of the peak is the height of the peak or the area of ​​the peak. Evaluation method for silicon nitride substrates.

4. A method for evaluating a silicon nitride substrate according to any one of claims 1 to 3, comprising: Detecting rare earth oxide particles in the evaluation area based on the distribution; creating a Voronoi polygon for each detected rare earth oxide particle; The dispersibility of the rare earth oxide particles is evaluated based on the degree of variation in the areas of the Voronoi polygons created. Evaluation method for silicon nitride substrates.

5. a data acquisition unit configured to acquire, from a Raman measurement device, Raman spectra measured at a plurality of locations in an evaluation region that is at least a portion of an unsintered silicon nitride substrate containing a rare earth oxide; a distribution acquisition unit configured to acquire a distribution in the evaluation region in which the intensity of a peak caused by lattice vibration of a rare earth oxide is associated with each of the plurality of locations based on the Raman spectrum acquired by the data acquisition unit; and an evaluation unit configured to evaluate the dispersibility of rare earth oxide particles in the evaluation area based on the distribution; An evaluation device for a silicon nitride substrate comprising:

6. A silicon nitride substrate evaluation system comprising the Raman measurement device and the silicon nitride substrate evaluation device according to claim 5 .

7. A method for manufacturing a silicon nitride substrate, comprising stacking a plurality of pre-sintered silicon nitride substrates containing rare earth oxides and sintering the stacked silicon nitride substrates, The silicon nitride substrates before sintering are selected so that 90% or more of the silicon nitride substrates to be stacked before sintering have a dispersion value Y of 0.5 or less, as measured by the following measurement method: A method for manufacturing a silicon nitride substrate. Measurement method: Raman spectra are obtained at a plurality of locations in an evaluation region, which is at least a portion of an unsintered silicon nitride substrate containing rare earth oxides. Based on the Raman spectra acquired at the plurality of locations, a distribution in the evaluation area is obtained in which the intensity of peaks resulting from lattice vibration of rare earth oxides is associated with each of the plurality of locations. Rare earth oxide particles are detected in the evaluation area based on the distribution. A Voronoi polygon is created for each detected rare earth oxide particle. The average value c of the areas of the Voronoi polygons is calculated. The standard deviation d of the areas of the Voronoi polygons is calculated. d / c is defined as the variance value Y.

8. A silicon nitride substrate is manufactured by the method for manufacturing a silicon nitride substrate according to claim 7, A power module is manufactured using the silicon nitride substrate. A method for manufacturing a power module.

Images