Silicon nitride substrate, method for evaluating silicon nitride substrate, evaluation device, evaluation system, and silicon nitride circuit substrate

Abstract

The present application provides a silicon nitride substrate which is less likely to generate color unevenness, an evaluation method for a silicon nitride substrate, an evaluation device, an evaluation system, and a silicon nitride circuit substrate. A silicon nitride substrate has a first surface and a second surface opposite to the first surface. In a measurement surface which is one of the first surface and the second surface, an average half-value width C ave of a value greater than 0 cm ‑1 and less than 5.32 cm ‑1 . The measurement method of the average half-value width C ave : one point in a central portion and four points in edge portions of the measurement surface are taken as measurement points, Raman spectra are measured at the measurement points, respectively, in each of the measured Raman spectra, a half-value width C of a spectral peak which takes a maximum intensity in a range of 850 cm ‑1 or more and 875 cm ‑1 or less is calculated, and an average value of the calculated half-value widths C is taken as the average half-value width C ave .

Description

Technical Field

[0001] This disclosure relates to silicon nitride substrates, evaluation methods, evaluation apparatus, evaluation systems, and silicon nitride circuit boards. Background Technology

[0002] Japanese Patent Application Publication No. 2003-267786 (Patent Document 1) discloses a technique for providing silicon nitride-based ceramic sintered bodies that, compared with existing materials, blacken the color of the sintered body, reduce color unevenness, and have sufficient strength.

[0003] Japanese Patent Application Publication No. 2005-214659 (Patent Document 2) discloses a technology involving a foreign object inspection device that can also identify foreign objects with colors that have low color contrast relative to the wiring substrate.

[0004] Japanese Patent Application Publication Nos. 2016-204206 (Patent Document 3), 2016-204207 (Patent Document 4), 2016-204209 (Patent Document 5), and 2016-204210 (Patent Document 6) disclose a technology for providing a silicon nitride-based ceramic component that is lightweight, has high hardness, excellent resistance to grinding and other processing, and thus has excellent appearance quality.

[0005] Japanese Patent Application Publication No. 9-227240 (Patent Document 7) discloses a technique that can reduce the thickness of the surface color layer in a silicon nitride ceramic sintered body, thereby making the destructive strength characteristics of the surface layer and the inner layer uniform.

[0006] Existing technical documents

[0007] Patent documents

[0008] Patent Document 1: Japanese Patent Application Publication No. 2003-267786

[0009] Patent Document 2: Japanese Patent Application Publication No. 2005-214659

[0010] Patent Document 3: Japanese Patent Application Publication No. 2016-204206

[0011] Patent Document 4: Japanese Patent Application Publication No. 2016-204207

[0012] Patent Document 5: Japanese Patent Application Publication No. 2016-204209

[0013] Patent Document 6: Japanese Patent Application Publication No. 2016-204210

[0014] Patent Document 7: Japanese Patent Application Publication No. 9-227240 Summary of the Invention

[0015] The problem that the invention aims to solve

[0016] Foreign matter and dirt adhering to the surface of a silicon nitride substrate can cause poor contact between the silicon nitride substrate and the solder, as well as poor insulation of the silicon nitride substrate itself. To detect foreign matter and dirt adhering to the surface of the silicon nitride substrate, a visual inspection was performed. One method for visual inspection is to photograph the surface of the silicon nitride substrate using an imaging device such as a CCD camera, and then compare the data from the captured image with pre-registered reference data to detect foreign matter and dirt.

[0017] Sometimes, uneven coloring can occur on the surface of a silicon nitride substrate. This uneven coloring can manifest as a difference in color between the center and the edges of the substrate. If uneven coloring exists on the surface of the silicon nitride substrate, it may be mistakenly detected as foreign matter or dirt during visual inspection.

[0018] In addition, if the warping of the silicon nitride substrate increases, the adhesion between the metal circuit board and the metal heat sink bonded to the silicon nitride substrate by solder will decrease. Due to the cooling process during the bonding process and the thermal stress generated during the thermal cycle of the power module operation, there are problems such as the metal circuit board and the metal heat sink being easily peeled off from the silicon nitride substrate.

[0019] In one aspect of this disclosure, a silicon nitride substrate that is less prone to color unevenness, a method for evaluating the silicon nitride substrate, an evaluation apparatus, and an evaluation system are preferably provided.

[0020] In addition, in one aspect of this disclosure, a silicon nitride substrate with suppressed warpage and less prone to color unevenness, a method for evaluating the silicon nitride substrate, an evaluation apparatus, and an evaluation system are preferably provided.

[0021] Methods for solving problems

[0022] (1) One aspect of this disclosure is a silicon nitride substrate having a first surface and a second surface opposite to the first surface. The average half-width C of the measurement surface, which is one of the first and second surfaces, is measured by the following measurement method. ave The value is greater than 0cm -1 And less than 5.32cm -1 .

[0023] Average half-width C ave Measurement method: One point in the center and four points on the edge of the above-mentioned measurement surface are used as measurement points. Raman spectra are measured at these measurement points respectively. The value at 850 cm⁻¹ is calculated from each of the measured Raman spectra.-1 Above and 875cm -1 The half-width C of the spectral peak with the maximum intensity is obtained within the following range. The average of the calculated half-widths C is taken as the average half-width C. ave .

[0024] The silicon nitride substrate disclosed herein is less prone to color unevenness. Furthermore, warpage is suppressed and color unevenness is less likely to occur.

[0025] (2) Another aspect of this disclosure is an evaluation method for a silicon nitride substrate, which is an evaluation method for evaluating the color unevenness of a silicon nitride substrate. Raman spectra are measured at measurement points on the silicon nitride substrate, and the half-width of the spectral peaks attributable to the lattice vibrations of silicon nitride contained in the Raman spectra is measured. The color unevenness of the silicon nitride substrate is evaluated based on the half-width.

[0026] As another aspect of the evaluation method for silicon nitride substrates disclosed herein, color unevenness can be evaluated even when the area of ​​the measurement point is small. Furthermore, warpage and color unevenness can be evaluated simultaneously.

[0027] (3) Another aspect of this disclosure is an evaluation apparatus for evaluating color unevenness of a silicon nitride substrate, comprising: a data acquisition unit configured to acquire Raman spectra measured at measurement points on the silicon nitride substrate; and a half-width measurement unit configured to measure the half-width of spectral peaks attributable to lattice vibrations of silicon nitride contained in the Raman spectra acquired by the data acquisition unit.

[0028] As another aspect of this disclosure, the evaluation device can evaluate color unevenness even when the area of ​​the measurement point is small. Furthermore, it can simultaneously evaluate warpage and color unevenness.

[0029] (4) Another aspect of this disclosure is an evaluation system for a silicon nitride substrate, which is an evaluation system for evaluating the color unevenness of a silicon nitride substrate, comprising: a Raman measuring device for measuring Raman spectra at measuring points on the silicon nitride substrate, and the evaluation device described in (3) above.

[0030] (5) Another aspect of this disclosure is a silicon nitride circuit board comprising: the silicon nitride substrate, a metal circuit disposed on one side of the silicon nitride substrate, and a metal heat sink disposed on the opposite side of the silicon nitride substrate. Attached Figure Description

[0031] Figure 1 This is a side view showing the structure of the power module and the silicon nitride circuit board.

[0032] Figure 2 This is an explanatory diagram illustrating the manufacturing method of a silicon nitride substrate.

[0033] Figure 3 This is an explanatory diagram showing the state of a stacked configuration of a molded body.

[0034] Figure 4 This is a photograph showing the surface of a silicon nitride substrate with uneven color, which is found in the same batch as the silicon nitride substrate 3Y that has just undergone the nitriding process.

[0035] Figure 5 This is an explanatory diagram illustrating the manufacturing method of a silicon nitride circuit board.

[0036] Figure 6 This is an explanatory diagram showing the configuration of an evaluation system for silicon nitride substrates.

[0037] Figure 7 This is an explanatory diagram showing the arrangement of measurement points on the measurement surface.

[0038] Figure 8 This is an explanatory graph representing the half-width of the fitted spectral peaks.

[0039] Figure 9 This is a table representing the wavenumber, height, half-width, and area of ​​the spectral peaks of the 3X and 3Y spectral peaks of a silicon nitride substrate.

[0040] Figure 10 This is a table representing the luminance L*, chromaticity a* from green to red, chromaticity b* from yellow to blue, chromaticity C*, and half-width at each measurement point during the first evaluation process of silicon nitride substrates 3X and 3Y.

[0041] Figure 11 This is a table representing the luminance L*, chromaticity a* from green to red, chromaticity b* from yellow to blue, chromaticity C*, and half-width at each measurement point during the second evaluation process of silicon nitride substrates 3X and 3Y.

[0042] Figure 12 This represents the average half-width and brightness L of the silicon nitride substrate during the first and second evaluation processes for 3X and 3Y. * The chart.

[0043] Figure 13 This represents the average half-width and chroma C of the silicon nitride substrate 3X and 3Y during the first and second evaluation processes. * The chart.

[0044] Figure 14 It is a graph showing the measured temperature of the molded body in the specified temperature range of the embodiment, the measured temperature of the furnace, and the temperature difference between the molded body and the furnace.

[0045] Figure 15It is a graph showing the measured temperature of the molded body in the specified temperature range of the comparative example, the measured temperature of the furnace, and the temperature difference between the molded body and the furnace.

[0046] Figure 16 This is a table representing the wavenumber, height, half-width, and area of ​​the spectral peaks of silicon nitride substrates 3A and 3B.

[0047] Figure 17 This table represents the luminance L*, chromaticity a* from green to red, chromaticity b* from yellow to blue, chromaticity C*, and half-width at each measurement point during the evaluation process of silicon nitride substrates 3A and 3B.

[0048] Symbol Explanation

[0049] 1…Power module, 2…Silicon nitride circuit board, 3, 3X, 3Y, 3A, 3B…Silicon nitride board, 5…Metal circuit, 7…Metal heat sink, 9, 11…Solder layer, 13…Semiconductor chip, 15…Heat sink, 100A…Molded body, 105…Metal plate, 107…Metal heat sink, 200…Setter, 201…Evaluation system, 203…Raman measurement device, 205…Evaluation device, 206…Laser, 207…Data acquisition unit, 208…Raman scattering light, 209…Data processing unit, 300…Pressure plate, 301…Measurement surface, 305, 307…Spectral peaks, 401, 404…Long side, 402, 403…Short side. Detailed Implementation

[0050] The exemplary embodiments of this disclosure will be described with reference to the accompanying drawings.

[0051] 1. Composition of power module 1 and silicon nitride circuit board 2

[0052] based on Figure 1 The structure of the power module 1 and the silicon nitride circuit board 2 will be described. The silicon nitride circuit board 2 includes a silicon nitride substrate 3, a metal circuit 5, a metal heat sink 7, and solder layers 9 and 11. The power module 1 includes the silicon nitride circuit board 2, a semiconductor chip 13, and a heat sink 15.

[0053] The silicon nitride substrate 3, for example, has a first surface and a second surface. The first surface is the upper surface in the molding process S2 and the sintering process S3, which will be described later. The second surface is the surface opposite to the first surface. The planar shape of the silicon nitride substrate 3 is, for example, a rectangular shape. Planar shape refers to the shape when viewed from the thickness direction of the silicon nitride substrate 3. For example, the length of each side of the silicon nitride substrate 3 is 100 mm or more.

[0054] Metal circuit 5 is made of copper plate. Metal circuit 5 is mounted on the first side of silicon nitride substrate 3 via solder layer 9. Metal heat sink 7 is made of copper plate. Metal heat sink 7 is mounted on the second side of silicon nitride substrate 3 via solder layer 11. Semiconductor chip 13 is mounted on metal circuit 5. Heat sink 15 is mounted on metal heat sink 7.

[0055] 2. Method for manufacturing silicon nitride substrate 3

[0056] For example, it can be done through Figure 2 The method shown manufactures a silicon nitride substrate 3. The manufacturing method includes a slurry preparation step S1, a molded body preparation step S2, a sintering step S3, and a nitriding step S4.

[0057] (2-1) Slurry preparation process S1

[0058] For example, a raw material powder is obtained by adding a sintering aid to silicon powder. Examples of sintering aids include rare earth oxides and magnesium compounds. The raw material powder is then used to prepare a slurry.

[0059] Examples of silicon powders that are industrially available include industrially available grades. The median particle size D50 of the silicon powder before pulverization is preferably 6 μm or more, and more preferably 7 μm or more.

[0060] The preferred BET specific surface area of ​​the silicon powder before pulverization is 3m². 2 / g or less, more preferably 2.5m 2 / g or less. The oxygen content of the silicon powder before pulverization is preferably 1.0% by mass or less, more preferably 0.5% by mass or less. The carbon content of the impurities in the silicon powder before pulverization is preferably 0.15% by mass or less, more preferably 0.10% by mass or less.

[0061] The purity of the silicon powder is preferably 99% or higher, more preferably 99.5% or higher. Oxygen impurities in the silicon powder are one of the main reasons hindering the thermal conductivity of the silicon nitride substrate obtained by reaction sintering. The higher the purity of the silicon powder, the higher the thermal conductivity of the silicon nitride substrate.

[0062] Preferably, the amount of oxygen from magnesium compounds is limited, so that the raw material powder is prepared in a manner in which the amount of impurity oxygen contained in the silicon powder and the total amount of oxygen from magnesium compounds are in the range of 0.1% by mass or more and 1.1% by mass or less relative to the amount of silicon converted to silicon nitride.

[0063] Impurity carbon contained in silicon powder can hinder the growth of silicon nitride particles in silicon nitride substrates obtained through reaction sintering. If the growth of silicon nitride particles is hindered, it is difficult to densify the silicon nitride substrate. If the silicon nitride substrate is difficult to densify, its thermal conductivity and insulation properties will decrease. Therefore, it is preferable to contain as little impurity carbon as possible in the silicon powder.

[0064] It should be noted that in this specification, the BET specific surface area (m²) 2 / g) refers to the value obtained using a BET surface area meter via the BET one-point method (JIS R1626:1996 "Determination of specific surface area of ​​fine ceramic powders based on gas adsorption BET method"). Additionally, the median particle size D50 (μm) refers to the particle size at which the cumulative frequency in the particle size distribution determined by laser diffraction scattering is 50%.

[0065] While not essential in the manufacturing method of the present invention, silicon nitride powder may be included in the raw material powder. However, compared to silicon, using silicon nitride is more costly, so the amount of silicon nitride used is preferably as small as possible. The amount of silicon nitride used is preferably 50 mol% or less of silicon (converted to silicon nitride), more preferably 10 mol% or less, and even more preferably 5 mol% or less. It should be noted that the silicon nitride substrate in this case is a silicon nitride substrate formed by nitriding a sheet-shaped molded body containing the above-mentioned silicon with silicon nitride at a silicon nitride content of 50 mol% or less.

[0066] Examples of rare earth element oxides include oxides of yttrium (Y), ytterbium (Yb), gadolinium (Gd), erbium (Er), and lutetium (Lu). These rare earth element oxides are readily available and stable as oxides. Specific examples of rare earth element oxides include yttrium oxide (Y₂O₃), ytterbium oxide (Yb₂O₃), gadolinium oxide (Gd₂O₃), erbium oxide (Er₂O₃), and lutetium oxide (Lu₂O₃).

[0067] In the silicon nitride substrate, the molar number of rare earth element oxides converted to trivalent oxides (RE₂O₃: RE represents rare earth elements) is designated as M1. In the silicon nitride substrate, the molar number of silicon converted to silicon nitride (Si₃N₄) is designated as M2. M2 is the molar number of silicon nitride obtained when silicon is completely nitrided. In the silicon nitride substrate, the molar number of magnesium compounds converted to MgO is designated as M3.

[0068] The molar ratio of M1 to the total number of moles of M1, M2, and M3 (hereinafter referred to as the molar ratio of rare earth oxides) is, for example, 0.5 mol% or more and less than 2 mol%. When the molar ratio of rare earth oxides is 0.5 mol% or more, the effect of the sintering aid is improved, and the density of the silicon nitride substrate is sufficiently increased. When the molar ratio of rare earth oxides is less than 2.0 mol%, the low thermal conductivity grain boundary phase is less likely to increase, the thermal conductivity of the sintered body is improved, and the amount of expensive rare earth element oxides used is reduced. In particular, the molar ratio of rare earth oxides is preferably 0.6 mol% or more and less than 2 mol%, more preferably 1 mol% or more and less than 1.8 mol%.

[0069] As magnesium compounds, one or more magnesium compounds containing "Si", "N" or "O" can be used. Magnesium oxide (MgO), magnesium silicon nitride (MgSiN2), magnesium silicide (Mg2Si), magnesium nitride (Mg3N2), etc. are preferred as magnesium compounds.

[0070] The ratio of the mass of silicon nitride magnesium to the total mass of magnesium compounds (hereinafter referred to as the silicon nitride magnesium mass ratio) is preferably 87% by mass or more. When the silicon nitride magnesium mass ratio is 87% by mass or more, the oxygen concentration in the resulting silicon nitride substrate can be reduced. When the silicon nitride magnesium mass ratio is less than 87% by mass, the oxygen content in the sintered silicon nitride particles increases, and therefore the thermal conductivity of the sintered silicon nitride substrate decreases. Therefore, a high silicon nitride magnesium mass ratio is preferred to improve the thermal conductivity of the silicon nitride substrate. The silicon nitride magnesium mass ratio is further preferably 90% by mass or more.

[0071] The molar ratio of M3 to the total number of moles of M1, M2, and M3 (hereinafter referred to as the molar ratio of the magnesium compound) is, for example, 8 mol% or more and less than 15 mol%. When the molar ratio of the magnesium compound is 8 mol% or more, the effect of the sintering aid is improved, and the density of the silicon nitride substrate is sufficiently increased. When the molar ratio of the magnesium compound is less than 15 mol%, the low thermal conductivity grain boundary phase is less likely to increase, and the thermal conductivity of the sintered body becomes higher. In particular, the molar ratio of the magnesium compound is preferably 8 mol% or more and less than 14 mol%, and more preferably 9 mol% or more and less than 11 mol%.

[0072] One method for preparing the slurry is as follows: Rare earth oxides and magnesium compounds are added to silicon powder in a prescribed ratio. Next, a dispersion medium is added. The dispersion medium is, for example, an organic solvent. A dispersant is also added if necessary.

[0073] Next, the material is pulverized using a ball mill to produce a slurry containing the raw material powder. The dispersion medium is, for example, an organic solvent. Examples of dispersion media include ethanol, n-butanol, and toluene. Examples of dispersants include sorbitan ester dispersants and polyoxyethylene dispersants.

[0074] The amount of dispersion medium added is preferably 40% by mass or more and 70% by mass or less relative to the total amount of raw material powder. The amount of dispersant added is preferably 0.3% by mass or more and 2% by mass or less relative to the total amount of raw material powder. It should be noted that after dispersion, the dispersion medium can be removed or replaced with other dispersion media as needed.

[0075] (2-2) Molded body manufacturing process S2

[0076] The slurry obtained as described above is then mixed with, for example, a dispersion medium, an organic binder, or a dispersant. Next, vacuum degassing is performed as needed. Then, the viscosity of the slurry is adjusted to a specified range. The result is a coating slurry.

[0077] Next, the obtained coating slurry is formed into a sheet using a sheet forming machine. Then, it is cut to the specified size and dried. The result is a sheet-shaped molded body.

[0078] There are no particular limitations on the organic binders used in the preparation of coating slurries. Examples of organic binders used in the preparation of coating slurries include PVB-based resins, ethyl cellulose-based resins, and acrylic resins. Examples of PVB-based resins include polyvinyl butyral resin. The amount of dispersion medium, organic binder, dispersant, etc., added can be adjusted appropriately according to the coating conditions.

[0079] There is no particular limitation on the method for forming coating slurry into sheets. Examples of methods for forming coating slurry into sheets include doctor blade molding and extrusion molding.

[0080] The sheet-shaped molded body produced in the molding process S2 has a thickness of, for example, 0.15 mm or more and 0.8 mm or less. The produced sheet-shaped molded body is cut to the specified size as needed, for example, using a punching machine.

[0081] (2-3) Sintering process S3, nitriding process S4

[0082] The sintering process S3 includes: a degreasing process to remove the organic binder contained in the molded body, a nitriding process S4 to react the silicon contained in the molded body with nitrogen to form silicon nitride, and a densification sintering process performed after the nitriding process S4.

[0083] The degreasing, nitriding (S4), and densification sintering processes can be performed sequentially in different furnaces or continuously in the same furnace. Furthermore, more than 1600 molded bodies can be heat-treated together in one furnace. The batches of molded bodies heat-treated in the same furnace are identical.

[0084] In sintering process S3, for example, as Figure 3 As shown, multiple molded bodies 100A are stacked on the setter 200. The setter 200 is made of boron nitride (BN). A separating material (not shown) is sandwiched between the molded bodies 100A. A pressure plate 300 is disposed on the multiple molded bodies 100A.

[0085] In this state, multiple molded bodies 100A are placed inside an electric furnace. Next, a degreasing process is performed. Then, in a nitriding apparatus, decarburization is performed at a temperature of 900°C to 1300°C. Next, a nitriding process S4 is performed. In nitriding process S4, the temperature is raised to a specified temperature under a nitrogen atmosphere. Temperature control in nitriding process S4 will be described later.

[0086] Next, a densification sintering process is performed in a sintering apparatus. The densification sintering process is performed, for example, by applying a load of 10 Pa to 1000 Pa to the molded body 100A through a pressure plate 300.

[0087] It should be noted that, as the aforementioned separation material, a boron nitride (BN) powder layer with a thickness of approximately 3 μm or more and 20 μm or less can be cited as an example. The boron nitride powder layer facilitates the separation of the silicon nitride substrate, which has become a sintered body, after the densification sintering process. The boron nitride powder layer is formed, for example, by coating one side of each molded body 100A with boron nitride powder in a slurry state. Methods for coating the boron nitride powder in a slurry state include, for example, spraying, brushing, screen printing, etc. The boron nitride powder preferably has a purity of 95% or more and a median particle size D50 of 1 μm or more and 20 μm or less.

[0088] The silicon nitride substrate is completed through the above processes. In the above manufacturing method, by using high-purity silicon powder, a silicon nitride substrate with a thermal conductivity of 110 W / (m·K) or higher can be manufactured. The silicon nitride substrate in this embodiment is a silicon nitride substrate formed by nitriding silicon contained in a sheet-shaped molded body. The thickness of the silicon nitride substrate is, for example, 0.15 mm or more and 0.8 mm or less.

[0089] (2-4) The relationship between uneven color and temperature control in nitriding process S4

[0090] In the above-described method for manufacturing a silicon nitride substrate, silicon powder is used instead of silicon nitride powder, thus requiring a nitriding step S4. The inventors have recently discovered that, depending on the heating conditions in the nitriding step S4, uneven coloring can occur on the surface of the manufactured silicon nitride substrate.

[0091] In this specification, "the surface of the silicon nitride substrate" can be either the first side or the second side of the silicon nitride substrate. The second side is the side opposite to the first side.

[0092] Additionally, in this specification, "uneven color" refers to, for example, a rectangular silicon nitride substrate where the hue of the central portion differs from that of the edge portion. Figure 4 This is a photograph showing the surface of a silicon nitride substrate from the same batch as the 3Y silicon nitride substrate, immediately after the nitriding process, exhibiting uneven coloring. Figure 4In the image, the central portion of the silicon nitride substrate is white, while the edges are black. That is, the color of the central portion differs from that of the edges.

[0093] The inventors conducted in-depth research into the mechanism of color unevenness on the surface of silicon nitride substrates. They hypothesize that the color unevenness is caused by the following mechanism.

[0094] Nitriding process S4 is performed in a nitrogen atmosphere, for example, as... Figure 3 As shown, multiple sheet-like molding bodies 100A are arranged on the setter 200, and a pressure plate 300 is arranged on the stacked molding bodies 100A for heat treatment. At this time, the molding body 100A sandwiched between the upper and lower molding bodies 100A in the stacked molding bodies 100A is designated as a specific molding body.

[0095] Heat tends to accumulate in the center of a molded object. Therefore, the temperature in the center of the molded object is higher, while the temperature at the edges is lower. Nitriding reactions are more likely to occur in the center of the molded object, but less likely to occur at the edges. Nitriding is an exothermic reaction; therefore, a positive feedback loop of heat generation occurs in the center where nitriding takes place, resulting in a rapid temperature rise. If the temperature rise in the center is large, the temperature may exceed the melting point of silicon, causing the silicon to melt (hereinafter referred to as thermal runaway). Furthermore, due to the increased temperature difference between the center and the edges, insufficient nitriding occurs at the edges. If sintering is performed while the edges are under-nitrided, the incompletely nitrided silicon will remain at the edges. These results are presumably responsible for uneven color distribution.

[0096] Based on the above mechanism, it is believed that the color unevenness on the surface of the silicon nitride substrate is caused by the heating process in the S4 nitriding process. It is argued that if the heating process can be performed simultaneously to achieve a temperature distribution that minimizes the temperature difference between the central and edge areas, thermal runaway and color unevenness can be suppressed.

[0097] Therefore, the heating process in the nitriding process S4 is preferably a process in which the temperature is gradually increased while reducing the temperature difference between the central part and the edge part.

[0098] In the heating step of nitriding process S4, the temperature is raised to the maximum heating temperature. The maximum heating temperature is preferably 1390°C or higher and 1500°C or lower. In the heating step of nitriding process S4, for example, the temperature is raised in a stepped manner. In the heating step, within the range of 1270°C to 1340°C, the average value of the temperature rise per unit time (hereinafter referred to as the slope of the heating temperature) is preferably 3.1°C / h or lower. When the average slope of the heating temperature is 3.1°C / h or lower, uneven color distribution can be suppressed.

[0099] To suppress warpage of the silicon nitride substrate, the cooling conditions for the nitriding and densification sintering processes are preferably controlled as follows: In the nitriding process, it is preferable to perform the process at a cooling rate of 262.2 °C / h or less within the cooling range from the highest heating temperature to 1100 °C. In the densification sintering process, it is preferable to perform the process at a cooling rate of 280.7 °C / h or less within the cooling range from the highest heating temperature to 800 °C. This allows the warpage of the silicon nitride substrate to be suppressed to within the range of 0.1 mm to 1 mm. The warpage amount is a value obtained by measuring the warpage amount according to the SORI standard. Specifically, the least squares plane of the upper surface of the sample is calculated, and the absolute value of the distance from the calculated least squares plane to the highest point of the upper surface of the sample and the absolute value of the distance to the lowest point of the upper surface of the sample are calculated as the warpage amount.

[0100] 3. Manufacturing method of power module 1 and silicon nitride circuit board 2

[0101] based on Figure 5 The manufacturing method of the power module 1 and the silicon nitride circuit board 2 will be described. In step S11, a metal plate 105 and a metal heat sink 107 are soldered onto the silicon nitride substrate 3. Next, in step S12, a portion of the metal plate 105 is removed to form a metal circuit 5. Then, in step S13, multiple silicon nitride circuit boards 2 are obtained by dicing.

[0102] Then, the semiconductor chip 13 and the heat sink 15 are mounted on the silicon nitride circuit board 2. The silicon nitride substrate 3 of the silicon nitride circuit board 2 is manufactured by the above-described "2. Manufacturing method of silicon nitride substrate 3", so it is not easy to produce uneven color and has high thermal conductivity.

[0103] 4. Composition of Evaluation System 201

[0104] based on Figure 6 The configuration of the evaluation system 201 will be described below. The evaluation system 201 is used to evaluate the silicon nitride substrate 3. The evaluation system 201 includes a Raman measurement device 203 and an evaluation device 205.

[0105] The Raman measuring device 203 irradiates a measuring point P, which is part of the measuring surface 301, with a laser 206, and detects the Raman scattered light 208 generated at the measuring point P. Therefore, the Raman measuring device 203 can measure the Raman spectrum of the measuring point P. The irradiation diameter of the laser 206 is approximately 1 μm. Therefore, the Raman measuring device 203 can measure the Raman spectrum of a narrow region.

[0106] The Raman measuring device 203 allows the irradiation position of the laser 206 to be moved independently in the x and y directions. The x and y directions are parallel to the measuring surface 301 and are orthogonal to it. Therefore, the Raman measuring device 203 can measure Raman spectra at multiple measuring points P on the measuring surface 301.

[0107] The evaluation device 205 includes a microcomputer with a CPU and semiconductor memory such as RAM or ROM. The functions of the evaluation device 205 are implemented by the CPU executing a program stored in a non-transitional physical recording medium. Furthermore, the corresponding method is executed by executing this program.

[0108] The evaluation apparatus 205 includes a data acquisition unit 207 and a data processing unit 209. The data acquisition unit 207 acquires Raman spectra from the Raman measuring apparatus 203. The data processing unit 209 performs processing based on the Raman spectra acquired by the data acquisition unit 207 to evaluate the color uniformity of the silicon nitride substrate 3. This processing will be described later. The data processing unit 209 corresponds to the half-width measurement unit.

[0109] 5. Evaluation methods for silicon nitride substrates

[0110] The color unevenness of the silicon nitride substrate 3 can be evaluated using the following methods. For example, evaluation system 201 can be used to evaluate the color unevenness.

[0111] First, a silicon nitride substrate 3, which is to be evaluated, is prepared. The silicon nitride substrate 3 has, for example, a first surface and a second surface. The second surface is the surface opposite to the first surface. The planar shape of the silicon nitride substrate 3 is, for example, a rectangular shape. For example, the length of each side of the silicon nitride substrate 3 is 100 mm or more.

[0112] Next, one of the first and second surfaces is designated as the measurement surface 301. Then, a measurement point P is set on the measurement surface 301.

[0113] The number of measurement points P can be single or multiple. For example, such as Figure 7As shown, five measurement points P1 to P5 can be set. Measurement point P1 is located in the center of measurement surface 301. Measurement points P2 to P5 are located at the edges of measurement surface 301. Measurement points P2 to P5 are located at one of the four corners of measurement surface 301. The distance from measurement point P2 to the long side 401 is 10 mm. The distance from measurement point P2 to the short side 402 is 15 mm. The distance from measurement point P3 to the long side 401 is 10 mm. The distance from measurement point P3 to the short side 403 is 15 mm. The distance from measurement point P4 to the long side 404 is 10 mm. The distance from measurement point P4 to the short side 402 is 15 mm. The distance from measurement point P5 to the long side 404 is 10 mm. The distance from measurement point P5 to the short side 403 is 15 mm.

[0114] Next, Raman spectra are measured at measurement point P using Raman measuring apparatus 203. If multiple measurement points P are set, Raman spectra are measured at each of the multiple measurement points P.

[0115] Next, the data acquisition unit 207 acquires the Raman spectrum measured at measurement point P from the Raman measuring device 203. When the Raman measuring device 203 measures Raman spectra at multiple measurement points P, the data acquisition unit 207 acquires the Raman spectrum at each measurement point P.

[0116] Next, the data processing unit 209 measures the half-width of the spectral peaks attributable to the lattice vibrations of silicon nitride contained in the Raman spectrum acquired by the data acquisition unit 207. When the Raman measuring apparatus 203 measures Raman spectra at multiple measurement points P, the data processing unit 209 measures the half-width of the spectral peaks in the multiple Raman spectra measured at the multiple measurement points P.

[0117] For example, when Raman spectra are measured at measurement points P1 to P5 respectively, the data processing unit 209 measures the half-width C1 of the spectral peaks contained in the Raman spectrum measured at measurement point P1, the half-width C2 of the spectral peaks contained in the Raman spectrum measured at measurement point P2, the half-width C3 of the spectral peaks contained in the Raman spectrum measured at measurement point P3, the half-width C4 of the spectral peaks contained in the Raman spectrum measured at measurement point P4, and the half-width C5 of the spectral peaks contained in the Raman spectrum measured at measurement point P5.

[0118] like Figure 8 As shown, the spectral peak 305 attributable to the lattice vibrations of silicon nitride is, for example, at 850 cm⁻¹. -1 Above and 875cm -1 The peak with the greatest intensity is obtained within the following range.

[0119] The method for measuring the half-width of the data processing unit 209 is as follows. Figure 8As shown, the data processing unit 209 uses a statistical distribution function to fit the spectral peak 305, obtaining the fitted spectral peak 307. Examples of statistical distribution functions include the Lorentz function. The data processing unit 209 measures the half-width C at the spectral peak 307. It should be noted that... Figure 8 In the diagram, A is the wavenumber of spectral peak 307. B is the height of spectral peak 307. D is the area of ​​spectral peak 307.

[0120] When the Raman measuring apparatus 203 measures Raman spectra at multiple measurement points P, the data processing unit 209 calculates the average half-width C. ave Average half-width C ave It is the average half-width C calculated for multiple Raman spectra measured at multiple measurement points P. For example, when Raman spectra are measured at measurement points P1 to P5, the average half-width C is... ave It is the average value of the half-value widths C1 to C5.

[0121] The data processing unit 209 evaluates the degree of color unevenness, for example, based on the half-width C. For example, when the half-width C is greater than 0 cm... -1 Furthermore, if the width at half maximum (C) is less than the threshold, it is considered that color unevenness has been suppressed; if the width at half maximum (C) is greater than the threshold, it is considered that color unevenness is significant. The threshold is, for example, 5.32 cm. -1 It should be noted that, based on the wavenumber resolution of commonly sold Raman spectroscopy devices, the lower limit of the half-value width C can also be set to 0.5 cm. -1 .

[0122] Data processing unit 209, for example, based on average half-value width C ave This is used to evaluate the degree of color unevenness. For example, in the average half-width C... ave Greater than 0cm -1 Furthermore, if the value is less than the threshold, it is determined that color unevenness has been suppressed, and the average half-width C is within this range. ave If the value is above a certain threshold, it is considered a significant color unevenness. For example, the threshold is...

[0123] 5.32cm -1 It should be noted that, based on the wavenumber resolution of commonly sold Raman spectroscopy devices, the average half-width C... ave The lower limit can also be set to 0.5cm. -1 .

[0124] If the interlattice strain of silicon nitride is large, then the half-width value will be large. It is assumed that the interlattice strain of silicon nitride changes due to deviations in the nitriding process of the silicon contained in the wafer. That is, if the nitriding deviation in the silicon nitride substrate 3 is large, then the interlattice strain of silicon nitride will increase, and the half-width value will increase.

[0125] 6. The effect of silicon nitride substrate

[0126] (6-1) The silicon nitride substrate disclosed herein is less prone to color unevenness. In addition, warpage is suppressed and color unevenness is less likely to occur.

[0127] (6-2) The silicon nitride substrate disclosed herein has high thermal conductivity.

[0128] 7. The effectiveness of evaluation methods for silicon nitride substrates

[0129] (7-1) The evaluation method for silicon nitride substrates disclosed herein can evaluate color unevenness even when the area of ​​the measurement point P is small. Therefore, it is possible to evaluate the color unevenness of small silicon nitride substrates used in, for example, micro-devices.

[0130] (7-2) In the evaluation method for silicon nitride substrates disclosed herein, for example, a statistical distribution function can be used to fit the spectral peaks, and the half-width of the fitted spectral peaks can be calculated. In this case, the half-width can be calculated more accurately.

[0131] 8. Examples

[0132] (8-1) Manufacturing of silicon nitride substrates 3X, 3Y, 3A, and 3B

[0133] Silicon nitride substrates 3X, 3Y, 3A, and 3B are manufactured using the method described in "2. Manufacturing Method of Silicon Nitride Substrate 3". Silicon nitride substrates 3X, 3Y, 3A, and 3B have a first surface and a second surface. The planar shape of silicon nitride substrates 3X, 3Y, 3A, and 3B is rectangular. The length of the long side of silicon nitride substrates 3X, 3Y, 3A, and 3B is 200 mm, and the length of the short side is 140 mm. The thickness of silicon nitride substrates 3X, 3Y, 3A, and 3B is 0.32 mm.

[0134] In manufacturing silicon nitride substrates 3X, 3Y, 3A, and 3B, the molar ratio of rare earth oxides is 1.2 mol%, and the molar ratio of magnesium compounds is 9.8 mol%. As mentioned above, the molar ratio of rare earth oxides is the molar ratio of M1 relative to the total number of moles of M1, M2, and M3. As mentioned above, the molar ratio of magnesium compounds is the molar ratio of M3 relative to the total number of moles of M1, M2, and M3.

[0135] In the manufacturing of silicon nitride substrates 3X, 3A, and 3B, during the nitriding process S4, the heating temperature is increased to the maximum heating temperature in a stepwise manner as the heating time progresses. The maximum heating temperature is 1400°C. The average slope of the heating temperature within the temperature rise range from 1270°C to 1340°C is 2.99°C / h. Figure 14 This indicates the measured temperature of the molded body within the specified temperature range, the measured temperature of the furnace, and the temperature difference between the molded body and the furnace, all within the same batch as the silicon nitride substrate 3X. Figure 14 The horizontal axis represents the elapsed time from when the measured temperature of the furnace, which serves as the heating temperature, reaches a reference point of approximately 1300°C. When the measured furnace temperature is approximately 1300°C, the temperature difference between the measured furnace temperature and the temperature of the molded body inside the furnace is less than 20°C. Even in the same batch, the silicon nitride substrate 3X, which underwent heat treatment, did not exhibit a rapid temperature rise in the molded body during the nitriding process, suggesting that "thermal runaway" did not occur. Furthermore, silicon nitride substrates 3A and 3B were heat-treated under the same heating conditions as silicon nitride substrate 3X, therefore, it is also speculated that "thermal runaway" did not occur. It should be noted that silicon nitride substrates 3A and 3B were samples heat-treated in the same batch, while silicon nitride substrate 3X and silicon nitride substrates 3A and 3B were samples heat-treated in different batches.

[0136] In the manufacturing of silicon nitride substrate 3Y, during the nitriding process S4, the heating temperature is increased to the maximum heating temperature in a stepwise manner as the heating time progresses. The maximum heating temperature is 1400°C. The average slope of the heating temperature within the temperature rise range from 1270°C to 1340°C is 4.67°C / h. Figure 15 This indicates the measured temperature of the molded body within the specified temperature range, the measured temperature of the furnace, and the temperature difference between the molded body and the furnace, all within the same batch as the silicon nitride substrate 3Y. Figure 15 The horizontal axis represents the elapsed time from when the measured temperature of the furnace, which serves as the heating temperature, reaches a reference point of approximately 1300°C. When the measured furnace temperature is approximately 1300°C, the temperature difference between the measured furnace temperature and the temperature of the molded body inside the furnace exceeds 20°C, with a maximum of 44.9°C. That is, even for silicon nitride substrates 3Y that have undergone heat treatment in the same batch, a rapid temperature rise in the molded body during the nitriding process occurs, suggesting the possibility of "thermal runaway."

[0137] The thermal conductivity of silicon nitride substrate 3X is 129 W / (m·K). The thermal conductivity of silicon nitride substrate 3Y is 120 W / (m·K). The thermal conductivity of silicon nitride substrates 3A and 3B is 124.4 W / (m·K).

[0138] (8-2) Evaluation of uneven color

[0139] Using the method described in "5. Evaluation Method for Silicon Nitride Substrates" above, the half-width C of silicon nitride substrates 3X, 3Y, 3A, and 3B was calculated respectively. ave The measurement surface 301 is the first surface of silicon nitride substrates 3X, 3A, and 3B, the first surface of silicon nitride substrate 3Y, and the second surface of silicon nitride substrate 3Y. In any measurement surface 301, the measurement point P is set as... Figure 7 The measurement points are shown as P1 to P5.

[0140] The conditions for Raman spectroscopy determination are as follows.

[0141] Raman measurement apparatus: Nanophoton RAMAN force Standard VIS-NIR-HS

[0142] Excitation wavelength: 532.06nm

[0143] Excitation output density: 1.76 × 10⁻⁶ 6 W / cm 2

[0144] ND filter: 99.23% (240 / 255)

[0145] Center wavelength of the beam splitter: 520.00cm -1

[0146] Grille: 1200gr / mm

[0147] Slit width: 50μm

[0148] Exposure time: 1 second

[0149] Average number of times: 20

[0150] Objective lens: TU Plan Fluor 5x / NA 0.15

[0151] The spectral peak attributable to the lattice vibrations of silicon nitride is 305 at 850 cm⁻¹. -1 Above and 875cm -1 The peak with the highest intensity is found within the following range. The Lorentz function is used as the statistical distribution function when fitting the spectral peak 305.

[0152] In addition, luminance L*, chromaticity a*, chromaticity b*, and chroma C* were measured at measurement points P1–P5. Furthermore, for silicon nitride substrates 3X, 3Y, 3A, and 3B, the presence of color unevenness was determined by visual inspection. (The last sentence appears to be incomplete and possibly refers to a measurement of luminance L*.) * chromaticity a * chromaticity b * Chroma C *In the measurements, a colorimeter (manufactured by Konica Minolta, trade name: CR-400) was used. The light source of the colorimeter was a xenon lamp. The measuring instrument was 8 mm in diameter. The illumination diameter was 11 mm. The methods for measuring luminance L*, chromaticity a*, chromaticity b*, and chromaticity C* were in accordance with JIS Z8722. The measurements of luminance L*, chromaticity a*, chromaticity b*, and chromaticity C* were performed under conditions including positively reflected light.

[0153] From measuring Raman spectra to calculating the half-width C ave The processing and measurement of brightness L up to date * chromaticity a * chromaticity b * Chroma C * The processing was also included in the evaluation process. Silicon nitride substrates 3X and 3Y underwent two evaluation processes. Silicon nitride substrates 3A and 3B underwent one evaluation process.

[0154] The wavenumber A, height B, half-width C, and area D of spectral peak 307 are shown in the figure. Figure 9 and Figure 16 It should be noted that "3X-1" refers to the measurement surface 301 being the first surface of the silicon nitride substrate 3X. "3Y-1" refers to the measurement surface 301 being the first surface of the silicon nitride substrate 3Y. "3Y-2" refers to the measurement surface 301 being the second surface of the silicon nitride substrate 3Y. "3A-1" refers to the measurement surface 301 being the first surface of the silicon nitride substrate 3A. "3B-1" refers to the measurement surface 301 being the first surface of the silicon nitride substrate 3B.

[0155] The luminance L*, chromaticity a*, chromaticity b*, chromaticity C*, and half-width C at each measurement point during the first evaluation of silicon nitride substrates 3X and 3Y are shown in the figure. Figure 10 . Figure 11 The luminance L*, chromaticity a*, chromaticity b*, chromaticity C*, and half-width C at each measurement point during the second evaluation processing of silicon nitride substrates 3X and 3Y are shown. The luminance L*, chromaticity a*, chromaticity b*, chromaticity C*, and half-width C at each measurement point during the evaluation processing of silicon nitride substrates 3A and 3B are also shown. Figure 17 Additionally, in Figure 10 , Figure 11 , Figure 17 The diagram also shows the presence or absence of color unevenness and the average value of each measurement calculated for each measurement surface.

[0156] Figure 12 The average half-width C of silicon nitride substrates 3X and 3Y during the first and second evaluation processes is shown. ave and brightness L * . Figure 13The average half-width C of silicon nitride substrates 3X and 3Y during the first and second evaluation processes is shown. ave And chroma C * .exist Figure 12 , Figure 13 The figure also shows the average half-value width C. ave The standard deviation range of the half-value width C centered on the center.

[0157] Average half-width C of silicon nitride substrates 3X, 3A, and 3B without color unevenness ave Less than 5.32cm -1 The average half-width C of the 3Y region of the silicon nitride substrate with uneven color is shown. ave Greater than 5.32cm -1 Therefore, the presence or absence of color unevenness is related to the average half-width C. ave Relevant. Based on this evaluation result, it can be confirmed that the average half-width C can be used as a basis. ave To evaluate color unevenness. Additionally, it can be confirmed that the average half-width C... ave Color unevenness is less likely to occur in small silicon nitride substrates. It should be noted that color unevenness in silicon nitride substrates 3Y is measured by visually observing the presence of minute color unevenness on one of the four sides of the substrate.

[0158] (8-3) Evaluation of warping

[0159] For silicon nitride substrates 3X, 3Y, 3A, and 3B, warpage was measured from the first and second surfaces, respectively. Warpage was measured using a warpage measuring apparatus manufactured by SoftWorks Co., Ltd. This apparatus measures the amount of warpage by placing three line lasers on a plate-shaped test object and taking images using a high-resolution camera. In this warpage measuring apparatus, the least-squares plane of the upper surface of the test object is calculated (defined). Then, the warpage is calculated as the sum of the absolute value of the distance from the calculated least-squares plane to the highest point of the upper surface of the test object and the absolute value of the distance to the lowest point of the upper surface of the test object. This calculation method is based on the SORI (SEMIM1, ASTM F 1451) standard.

[0160] (1) Warp measurement results from the first side

[0161] The warpage of silicon nitride substrate 3X is 0.779 mm, the warpage of silicon nitride substrate 3Y is 0.999 mm, the warpage of silicon nitride substrate 3A is 0.840 mm, and the warpage of silicon nitride substrate 3B is 0.751 mm.

[0162] (2) Warp measurement results from the second side

[0163] The warpage of silicon nitride substrate 3X is 0.653 mm, the warpage of silicon nitride substrate 3Y is 0.879 mm, the warpage of silicon nitride substrate 3A is 0.679 mm, and the warpage of silicon nitride substrate 3B is 0.581 mm.

[0164] (3) Investigation

[0165] It can be seen that the warpage of each silicon nitride substrate 3X, 3Y, 3A, and 3B all meets the acceptable standard of "less than 1mm", but the warpage of the average half-width C is slightly higher. ave Less than 5.32cm -1 In the silicon nitride substrates 3X, 3A, and 3B, the warpage is further suppressed to a low level of less than 0.840 mm.

[0166] 9. Other implementation methods

[0167] The embodiments of this disclosure have been described above, but this disclosure is not limited to the above embodiments and can be implemented in various ways.

[0168] (1) In the evaluation method for silicon nitride substrates disclosed herein, for example, it may be possible not to fit the spectral peaks using a statistical distribution function. In this case, the half-width C of the unfitted spectral peaks can be measured.

[0169] (2) Alternatively, a silicon nitride substrate 3 can be used to manufacture power modules other than power module 1.

[0170] (3) Multiple functions of one component in the above embodiments can be achieved by multiple components, or one function of one component can be achieved by multiple components. Alternatively, multiple functions of multiple components can be achieved by one component, or one function achieved by multiple components can be achieved by one component. Furthermore, a portion of the configuration of the above embodiments can be omitted. Additionally, at least a portion of the configuration of the above embodiments can be added to or replaced in other embodiments.

[0171] (4) In addition to the evaluation system 201 described above, this disclosure can also be implemented in various forms, such as a higher-level system that includes the evaluation system 201 as a component, a program for enabling the computer to function as the evaluation device 205, a recording medium of a non-transitional entity such as a semiconductor memory that records the program, and a quality management method for silicon nitride substrates.

Claims

1. A silicon nitride substrate, comprising a first surface and a second surface opposite to the first surface, formed by nitriding a silicon-containing sheet-like molded body. The silicon nitride substrate has a rectangular planar shape. The length of each side of the silicon nitride substrate is 100 mm or more. In the measurement surface that serves as one of the first and second surfaces, the average half-width C is measured using the following measurement method. ave The value is greater than 0cm -1 And less than 5.32cm -1 , Average half-width C ave The measurement method is as follows: One point in the center and four points on the edge of the measurement surface are used as measurement points. Raman spectra are measured at each measurement point. From each measured Raman spectrum, the value at 850 cm⁻¹ is calculated. -1 Above and 875cm -1 The half-width C of the spectral peak with the maximum intensity is obtained within the following range, and the average value of the calculated half-width C is taken as the average half-width C. ave .

2. The silicon nitride substrate according to claim 1 has a thermal conductivity of 110 W / (m·K) or higher.

3. A method for evaluating a silicon nitride substrate as described in claim 1 or 2, which is a method for evaluating silicon nitride substrates with uneven color. Raman spectra were measured at measurement points on the silicon nitride substrate. The half-width of the spectral peaks attributable to lattice vibrations of silicon nitride contained in the Raman spectrum was determined. The color unevenness of the silicon nitride substrate is evaluated based on the half-width.

4. The evaluation method for silicon nitride substrate according to claim 3, wherein, The spectral peak is at 850 cm⁻¹ -1 Above and 875cm -1 The peak with the greatest intensity is obtained within the following range.

5. The evaluation method for silicon nitride substrate according to claim 3 or 4, wherein, The spectral peaks were fitted using a statistical distribution function. The half-width is calculated from the spectral peaks after the fitting.

6. An evaluation apparatus for evaluating color unevenness of a silicon nitride substrate as described in claim 1 or 2, comprising: The data acquisition unit is configured to acquire Raman spectra measured at measurement points on the silicon nitride substrate; and The half-width measurement unit is configured to measure the half-width of the spectral peaks attributable to the lattice vibrations of silicon nitride contained in the Raman spectrum acquired by the data acquisition unit.

7. An evaluation system for silicon nitride substrates, which is an evaluation system for evaluating color unevenness of silicon nitride substrates. The device comprises: a Raman measuring apparatus for measuring Raman spectra at measuring points on the silicon nitride substrate, and the evaluation apparatus as described in claim 6.

8. A silicon nitride circuit board, comprising: The silicon nitride substrate as described in claim 1 Metal circuitry disposed on one side of the silicon nitride substrate, and A metal heat sink is disposed on the opposite side of the silicon nitride substrate.

Images