JP7655035B2 - Thermally conductive resin composition, thermally conductive resin sheet, laminated heat dissipation sheet, heat dissipation circuit board, and power semiconductor device - Google Patents

Description

The present invention relates to a thermally conductive resin composition, a thermally conductive resin sheet, a laminated heat dissipation sheet having a configuration in which a heat dissipation metal layer is laminated on the surface of the thermally conductive resin sheet, and a heat dissipation circuit board having a configuration in which a conductive circuit is further formed.

In recent years, power semiconductor devices used in various fields such as railways, automobiles, industry, and general home appliances are being shifted from conventional Si power semiconductors to power semiconductors using SiC, AlN, GaN, etc. in order to achieve further miniaturization, cost reduction, and efficiency improvement.
Power semiconductor devices are generally used as power semiconductor modules in which a plurality of semiconductor devices are arranged on a common heat sink and packaged.

Various issues have been raised in the practical application of such power semiconductor devices. One of these issues is the problem of heat generation from the devices. Power semiconductor devices can achieve high output and high density when operated at high temperatures. However, there are concerns that the heat generated by device switching may reduce the reliability of power semiconductor devices.

In recent years, particularly in the electrical and electronics fields, heat generation due to the increasing density of integrated circuits has become a major problem, and how to dissipate the heat has become an urgent issue. One method of solving this problem is to use highly thermally conductive ceramic substrates such as alumina substrates and aluminum nitride substrates as heat dissipation substrates on which power semiconductor devices are mounted. However, ceramic substrates have problems such as being easily cracked by impact, and being difficult to make into thin films and compact.

Therefore, thermally conductive resin sheets containing resin and inorganic filler are being considered as alternative products to the ceramic substrates.
Among these, as an inorganic filler, hexagonal boron nitride has attracted attention from the viewpoint of thermal conductivity and the like.
However, since hexagonal boron nitride particles are plate-shaped, they have high thermal conductivity in the plane direction (the a-b axis direction) but low thermal conductivity in the thickness direction (the c axis direction). When this hexagonal boron nitride is mixed with a resin and molded into a sheet, the hexagonal boron nitride tends to be oriented in the flow direction of the resin composition, that is, in the plane direction of the sheet, so the resulting thermally conductive resin sheet has high thermal conductivity in the plane direction but low thermal conductivity in the thickness direction.

In order to improve the anisotropy of the thermal conductivity of a thermally conductive resin sheet, studies have been conducted on reducing particle orientation by using agglomerated boron nitride particles obtained by agglomerating boron nitride particles.
For example, Patent Documents 1 and 2 propose spherical agglomerates obtained by binding boron nitride particles with a binder and then spray-drying the particles.
Moreover, Patent Document 3 proposes hexagonal boron nitride particles in which primary particles of hexagonal boron nitride are aggregated together in a pinecone shape.
Furthermore, Patent Document 4 proposes a granulated powder that contains spherical secondary particles formed by agglomerating scaly boron nitride, in which the density of the primary particles in the core portion of the secondary particles is lower than the density of the primary particles in the shell portion.
Patent Document 5 proposes boron nitride agglomerated particles in which primary particles of hexagonal boron nitride are agglomerated, and the primary particles in the boron nitride agglomerated particles have a house-of-cards structure.

JP 2006-257392 A Special Publication No. 2008-510878 Japanese Patent Application Publication No. 09-202663 JP 2016-044098 A JP 2015-006985 A

Conventional thermally conductive resin sheets sometimes lacked sufficient voltage resistance and thermal conductivity.

Moreover, thermally conductive resin sheets for use in power semiconductor devices are required to have resistance to a reflow process.
The reflow process is one of the processes for assembling a power semiconductor module. In this reflow process, the temperature of the components is rapidly raised to melt the solder and join the metal components together. In recent years, the operating temperature of power semiconductor devices has increased with the increase in output and density, so the solder used in the reflow process is also required to have heat resistance, and it has become common to use high-temperature solder that requires a reflow temperature of 290°C. Therefore, in the reflow process, a process of raising the temperature to around 290°C at which the high-temperature solder flows and then cooling it is repeated.
Furthermore, if the components absorb moisture before the reflow process is performed, deterioration of the components during the reflow process is greatly accelerated, and the voltage resistance performance may be greatly reduced.

Therefore, an object of the present invention is to provide a thermally conductive resin sheet that has good voltage resistance performance and thermal conductivity, and is excellent in moisture absorption and reflow resistance.
In this specification, "moisture absorption reflow resistance" means that even after a moisture absorption reflow test in which a thermally conductive resin sheet is laminated with a metal plate and stored under high temperature and high humidity conditions (e.g., an environment of 85°C and 85% RH for 3 days) and then a reflow test (e.g., 290°C) is performed, the thermally conductive resin sheet has high voltage resistance and does not delaminate at the interface with the metal plate or deform due to foaming of the thermally conductive resin sheet.

A thermally conductive resin composition according to one embodiment of the present invention is a resin composition comprising a thermoplastic resin and boron nitride agglomerated particles, wherein, when the intraparticle pore volume of the boron nitride agglomerated particles, as measured by mercury intrusion porosimetry, is _A1 and the interparticle space volume of the boron nitride agglomerated particles is _B1 , _B1 /( _A1 + _B1 ) is 0.60 or more.

The thermally conductive resin sheet obtained from the thermally conductive resin composition of the present invention has good voltage resistance and thermal conductivity, and is excellent in moisture absorption and reflow resistance.

FIG. 2 is a diagram showing the principle of measuring the intraparticle pore volume and interparticle space volume of boron nitride agglomerated particles by mercury intrusion porosimetry. FIG. 2 is a conceptual diagram of intra-particle pore volume. FIG. 1 is a conceptual diagram of interparticle volume. FIG. 2 is a schematic diagram of a particle cross-section of an example of a boron nitride agglomerated particle having a small interparticle volume. 1 is a SEM photograph (a photograph substituted for a drawing) of a particle surface according to an example of boron nitride agglomerated particles having a small interparticle void volume. FIG. 2 is a schematic diagram of a particle cross-section of an example of a boron nitride agglomerated particle having a large interparticle void volume. 1 is a SEM photograph (a photograph substituted for a drawing) of a particle surface according to an example of a boron nitride agglomerated particle having a large interparticle void volume. FIG. 1 is a schematic diagram of a house of cards structure. 13 shows an example of an SEM photograph of boron nitride agglomerated particles contained in residual ash when the thermally conductive resin sheet of Example 2 is heated to 700° C.

An example of an embodiment of the present invention is described in detail below. However, the present invention is not limited to the following embodiment, and can be modified in various ways within the scope of the gist of the invention.

1. Resin Composition The thermally conductive resin sheet of the present invention is a resin composition containing a thermoplastic resin and boron nitride agglomerated particles, in which, when the intraparticle pore volume of the boron nitride agglomerated particles is defined as _A1 and the interparticle volume is defined as _B1 , the ratio _B1 /( _A1 + _B1 ) is 0.60 or more, as measured by mercury intrusion porosimetry.
Each component will be described in detail below.

(1) Thermoplastic Resin The thermoplastic resin used in the thermally conductive resin composition of the present invention is preferably resistant to flow of the resin due to a decrease in elastic modulus, plastic deformation, and recovery of thermal restoring strain even under reflow conditions for a heat dissipating circuit board, for example, under conditions of 290° C. and 5 minutes. To satisfy this requirement, it is preferable that the glass transition temperature (Tg) of the thermoplastic resin is 300° C. or higher, or the melting point (Tm) is 300° C. or higher.

When the thermoplastic resin is a resin composition consisting of a combination of two or more types of thermoplastic resins that are mutually compatible, it is preferable that the glass transition temperature (Tg) of the resin composition is 300°C or higher, or the melting point (Tm) of the resin composition is 300°C or higher.
On the other hand, when the thermoplastic resin is a resin composition consisting of a combination of two or more types of thermoplastic resins that are incompatible with each other, it is preferable that the glass transition temperature (Tg) of the thermoplastic resin that is the main component of the resin composition is 300°C or higher, or the melting point (Tm) of the thermoplastic resin that is the main component of the resin composition is 300°C or higher.
Here, "main component" means the resin that has the highest mass content in the resin composition, and refers to a resin that accounts for 50 mass% or more of the resin composition, preferably 60 mass% or more, preferably 70 mass% or more, preferably 80 mass% or more, preferably 90 mass% or more, and preferably 95 mass% or more (including 100 mass%).

From the above viewpoint, the thermoplastic resin preferably contains a non-crystalline thermoplastic resin having a glass transition temperature (Tg) of 300° C. or higher and/or a crystalline thermoplastic resin having a melting point (Tm) of 300° C. or higher.
Among these, it is preferable that a non-crystalline thermoplastic resin having a glass transition temperature (Tg) of 300° C. or higher and/or a crystalline thermoplastic resin having a melting point (Tm) of 300° C. or higher is the main component of the thermoplastic resin, i.e., that it accounts for 50% by mass or more of the entire thermoplastic resin, and more preferably 60% by mass or more, more preferably 70% by mass or more, more preferably 80% by mass or more, and even more preferably 90% by mass or more (including 100% by mass).

In the present invention, "non-crystalline thermoplastic resin" refers to a thermoplastic resin that does not have a melting point. On the other hand, "crystalline thermoplastic resin" refers to a thermoplastic resin that has a melting point.

Examples of heat-resistant amorphous thermoplastic resins available as commercially available raw materials include polycarbonate resin (Tg: 152°C), modified polyphenylene ether resin (Tg: 211°C), polysulfone resin (Tg: 190°C), polyphenylsulfone resin (Tg: 220°C), polyethersulfone resin (Tg: 225°C), polyetherimide resin (Tg: 217°C), and the like.
However, none of these commercially available heat-resistant amorphous thermoplastic resins has a glass transition temperature far below 300° C. or higher.
Thermoplastic polyimide resins having a glass transition temperature of 300° C. or higher are commercially available. However, the moldable temperature of the thermoplastic polyimide resin is very high, and the melt viscosity at the moldable temperature is very high, making it difficult to fill a large amount of boron nitride agglomerated particles. Furthermore, since the molecular structure contains imide groups, it is highly hygroscopic, and therefore is not suitable for use as the main component of the thermoplastic resin of the present invention.

For the above reasons, the thermoplastic resin used in the thermally conductive resin composition of the present invention is preferably a resin mainly composed of a crystalline thermoplastic resin having a melting point of 300°C or higher. By using a crystalline thermoplastic resin having a melting point of 300°C or higher as the main component, sufficient heat resistance and durability as a substrate for a power semiconductor device can be obtained. In addition, even if a large amount of boron nitride agglomerated particles are filled, the moldability is good, and voids inside the sheet caused by the shape of the agglomerated particles or internal voids, and voids inside the sheet caused by moisture absorption of the resin components or agglomerated particles can be sufficiently reduced, resulting in good insulation. In addition, a crystalline thermoplastic resin having a melting point of 300°C or higher is less likely to flow, plastically deform, or recover from thermal restoring strain due to a decrease in elastic modulus under reflow conditions, so that the moisture absorption reflow resistance is also good.

The melting point of the thermoplastic resin can be measured by the method specified in JIS K-7121 "Method of measuring transition temperature of plastics - Determination of melting temperature". Specifically, the thermoplastic resin composition is used as a specimen, and 10 mg of the sample is heated from -40°C to 380°C at a heating rate of 10°C/min, held at 380°C for 1 minute, cooled to -40°C at a cooling rate of 10°C/min, held at the same temperature for 1 minute, and then heated again at 10°C/min. The temperature (Tpm) at the apex of the melting peak is read as the melting point (Tm).
The crystalline thermoplastic resin used in the present invention can be one in which at least endothermic peaks due to crystalline melting can be clearly confirmed and the temperature of the apex of the main peak among the peaks is 300° C. or higher.
When a resin composition containing added boron nitride agglomerated particles is measured by a differential scanning calorimeter (DSC), there is no significant difference in melting point, except in cases where the addition of the boron nitride agglomerated particles accelerates deterioration of the matrix resin during heat molding.

From the viewpoint of moisture absorption reflow resistance at 290°C, the melting point of the crystalline thermoplastic resin is more preferably 310°C or higher, of which 320°C or higher, and even more preferably 330°C or higher. On the other hand, there is no particular upper limit to the melting point. From the viewpoint of moldability and productivity, it is more preferably 380°C or lower, of which 370°C or lower, and even more preferably 360°C or lower.

By having the melting point of the thermoplastic resin be 300° C. or higher, the elastic modulus of the resin raw material is unlikely to decrease even under the reflow condition of 290° C. Therefore, the flow deformation of the resin can be suppressed even during the reflow process, and the elastic strain of the resin layer is unlikely to be restored, so that the deterioration of the surface appearance of the thermally conductive resin sheet and the unevenness of the thickness of the thermally conductive resin sheet can be suppressed, and a thermally conductive resin sheet with sufficient strength can be obtained.
In addition, since the elastic modulus of the resin raw material is unlikely to decrease at 290°C, even if the resin composition absorbs moisture in a humid and hot environment and moisture is present in the resin composition, the moisture in the resin composition is unlikely to expand during the reflow process or the process of mounting on a module, and foaming is unlikely to occur in the thermally conductive resin sheet, resulting in good voltage resistance performance and thermal conductivity.

It is believed that there is a relationship between the foaming in the thermally conductive resin sheet during the reflow process and the voltage resistance performance and thermal conductivity, for example, as follows.
When moisture in the resin composition expands during the reflow process or the process of mounting on a module, foaming may occur inside the thermally conductive resin sheet, near the interface between the heat dissipation metal layer and the thermally conductive resin sheet, or near the interface between the conductive circuit pattern and the thermally conductive resin sheet.
For example, when bubbles are generated inside the thermally conductive resin sheet, there is a risk that the voltage resistance performance will be significantly reduced.
When a heat dissipating metal material is laminated on one surface of a thermally conductive resin sheet, if foaming occurs near the interface between the heat dissipating metal material and the thermally conductive resin sheet, the heat dissipating metal material may peel off, resulting in a significant decrease in thermal conductivity.
Furthermore, if a conductive circuit pattern is formed on the other surface of the thermally conductive resin sheet, if foaming occurs near the interface between the conductive circuit pattern and the thermally conductive resin sheet, the conductive circuit pattern may peel off or fall off, causing abnormalities in the electrical circuit or a significant decrease in the thermal conductivity of the thermally conductive resin sheet.

Specific examples of heat-resistant crystalline thermoplastic resins include polybutylene terephthalate resin (PBT, melting point: 224°C), polyamide 6 (nylon 6, melting point: 225°C), polyamide 66 (nylon 66, melting point: 265°C), liquid crystal polymer (LCP, melting point: 320°C to 344°C), polyether ketone resin (melting point: 303°C to 400°C), polytetrafluoroethylene resin (PTFE, melting point: 327°C), tetrafluoroethylene-perfluoroalkoxyethylene copolymer resin (PFA, melting point: 302°C to 310°C), and tetrafluoroethylene-hexafluoropropylene copolymer resin (FEP, melting point: 250°C to 290°C).

As the crystalline thermoplastic resin used in the present invention, liquid crystal polymers, polyether ketone resins, PTFE, and PFA are preferred because they have a melting point of 300° C. or higher.
Among these crystalline thermoplastic resins having a melting point of 300° C. or more, liquid crystal polymers and/or polyether ketone resins are particularly preferred from the viewpoint of moldability, etc. Furthermore, among these, polyether ketone resins are preferred from the viewpoint of adhesion to heat dissipating metal layers such as copper plates.

The polyether ketone resin that can be used in the present invention is a general term for thermoplastic resins having repeating units represented by the following formula (1) (where m and n are 1 or 2):

Examples of polyetherketone resins include polyetherketone (PEK; m=1, n=1, melting point 373°C), polyetheretherketone (PEEK; m=2, n=1, melting point 343°C), polyetherketoneketone (PEKK; m=1, n=2, melting point 303°C to 400°C), polyetheretherketoneketone (PEEKK; m=2, n=2, melting point 358°C), and polyetherketoneetherketoneketone (PEKEKK; copolymer containing both m=1, n=1 structural units and m=1, n=2 structural units, melting point 387°C). All of these are available as commercially available raw materials.

Among polyether ketone resins, polyether ether ketone (PEEK) is particularly preferred, taking into consideration a number of factors, including its sufficiently high melting point and relatively low molding processing temperature, which allows for a shortened molding cycle; its continuous heat resistance of 200°C or higher, which has been proven to allow for use in heat-resistant applications without problems; various grades available in terms of melt viscosity, which is related to molding processability; its structure being the most chemically stable among polyether ketone resins, its excellent resistance to hot water and chemicals; and its price, which has come down due to its use in a wide range of applications.

When polyether ether ketone (PEEK) is used as the crystalline thermoplastic resin in the present invention, it may be blended with other thermoplastic resins.
The type of the other thermoplastic resin is not particularly limited. Among them, the other thermoplastic resin is preferably a compatible resin that has the effect of compensating for the performance that is insufficient when only polyether ether ketone is used for the application of the present invention.

As the compatible resin, polyetherimide (PEI) is more preferable.
Combining PEEK with PEI not only makes it possible to adjust the crystallinity (heat of crystalline fusion) of PEEK, but also makes it possible to adjust the crystallization rate of PEEK and increase the glass transition temperature when the resin composition is in an amorphous state (Tg of PEEK is 143° C., while Tg of PEI is 217° C.).
Since PEI is an amorphous resin, the melting point of the resin itself does not change even when PEEK and PEI are combined.
In addition, in the present invention, by using PEI which has an imide group and is amorphous, it is possible to improve the adhesion to heat dissipating metal materials such as copper plates.

The amount of PEI added is preferably 50% by mass or less, with the total amount of the resin composition being 100% by mass. By adding 50% by mass or less of PEI, the heat resistance in the moisture absorption reflow test can be maintained by the crystallinity of PEEK, while the adhesion to the heat dissipating metal material can be improved.

Various commercially available PEEK products can be used, such as "KetaSpire (registered trademark)" manufactured by Solvay, "Vestakeep (registered trademark)" manufactured by Daicel-Evonik Ltd., and "Victrex PEEK" manufactured by Victrex, which are available from various companies with various melt viscosities.
These PEEK raw materials may be of a single grade, or may be a blend of multiple grades with different melt viscosities, etc.

The melt viscosity of the crystalline thermoplastic resin in the present invention is not particularly limited. In particular, since a relatively large amount of boron nitride agglomerated particles is blended, in order to facilitate hot molding, the melt viscosity of the crystalline thermoplastic resin is preferably 0.60 kPa·s or less, more preferably 0.30 kPa·s or less. By having the melt viscosity within the above range, it is not necessary to set the temperature of the molding machine excessively high, and deterioration of the raw material can be suppressed. On the other hand, the lower limit of the melt viscosity is not particularly limited. In particular, 0.01 kPa·s or more is preferable.
The melt viscosity is a value measured in accordance with ASTM D3835 under conditions of a shear rate of 1000 s ^-1 and a temperature of 400°C.

From the viewpoint of long-term durability in a heated environment, the mass average molecular weight (Mw) of the crystalline thermoplastic resin is preferably 48,000 or more, more preferably 49,000 or more, and even more preferably 50,000 or more. On the other hand, from the viewpoint of moldability, it is preferably 120,000 or less, more preferably 110,000 or less, and even more preferably 100,000 or less.

The MFR of the crystalline thermoplastic resin is preferably 8 g/10 min or more, more preferably 9 g/10 min or more, and even more preferably 10 g/10 min or more, from the viewpoint of moldability and the difficulty of forming voids between the resin and the added boron nitride agglomerated particles. On the other hand, from the viewpoint of long-term durability under a heated environment, the MFR is preferably 180 g/10 min or less, more preferably 170 g/10 min or less, and even more preferably 160 g/10 min or less.
The MFR is a value measured at 380° C. and 5 kgf in accordance with JIS K7210:2014.

(2) Boron Nitride Agglomerated Particles In the boron nitride agglomerated particles of the present invention, when the intraparticle pore volume of the boron nitride agglomerated particles is _A1 and the interparticle space volume of the boron nitride agglomerated particles is _B1 , the ratio _B1 /( _A1 + _B1 ) is preferably 0.60 or more, as measured by mercury intrusion porosimetry.
In the present invention, the intraparticle pore volume and interparticle space volume of the boron nitride agglomerated particles are each determined in accordance with JIS R1655:2003 by the method described in the Examples.
Specifically, first, a mercury intrusion/extrusion curve is measured by mercury intrusion porosimetry. Next, as illustrated in Figure 1, a pore size distribution curve is created with pore size on the horizontal axis and logarithmic differential pore volume on the vertical axis. Between peak a representing intraparticle pores and peak b representing interparticle pores, a diameter (division diameter, X in Figure 1) at which the logarithmic differential pore volume is a minimum value relative to the pore size is read. The integral value of the mercury intrusion/extrusion curve in the region (dashed arrow in Figure 1) where the pore size is larger than the division diameter is the interparticle void volume.
The integral value of the mercury intrusion-extrusion curve in the entire measurement region is the total pore volume, and the intra-particle pore volume is the total pore volume minus the inter-particle volume.

The intraparticle pore volume is the pores a present inside the agglomerated particles in Fig. 2 excluding closed pores into which mercury is not pressed in. This intraparticle pore volume is a value determined by the density, average aspect ratio, particle thickness, etc. of the boron nitride primary particles constituting the agglomerated particles, and is involved in the mechanical properties of the boron nitride agglomerated particles and the thermal conductivity inside the agglomerated particles.
By appropriately reducing the intra-particle pore volume, the strength of the agglomerated particles can be increased. For example, even when a sheet is produced by press molding, the agglomerated particles themselves will not be crushed or excessively deformed, and the isotropy of the agglomerated particles will be maintained, effectively preventing a decrease in thermal conductivity in the thickness direction of the thermal conductive resin sheet.
On the other hand, if the intra-particle pore volume is appropriately increased, sufficient penetration of the resin into the internal pores of the aggregated particles can be ensured, and the generation of voids in the thermally conductive resin sheet can be suppressed, resulting in good voltage resistance performance.

From the above viewpoints, the intraparticle pore volume _A1 is preferably 0.30 mL/g or more, more preferably 0.35 mL/g or more, even more preferably 0.40 mL/g or more, and even more preferably 0.42 mL/g or more. The intraparticle pore volume _A1 is preferably 0.80 mL/g or less, more preferably 0.70 mL/g or less, even more preferably 0.60 mL/g or less, even more preferably 0.55 mL/g or less, and particularly preferably 0.50 mL/g or less.

The interparticle volume is the pore β present between multiple agglomerated particles in Figure 3, and generally represents the degree of packing due to the shape, particle size, and distribution of the agglomerated particles. However, for boron nitride agglomerated particles, in addition to the above, it is a numerical value that also takes into account the orientation state of the primary particles on the agglomerated particle surface (whether the primary particle surface is oriented in the radial direction of the agglomerated particle, or whether the primary particle surface is oriented in the circumferential direction of the agglomerated particle, etc.), the density of the surface primary particles, the average aspect ratio, the particle thickness, etc., and it is also related to the density, average aspect ratio, particle thickness, etc., of the primary particles that make up the entire agglomerated particle. This is because, when the size of the primary particles is excessively large and lacks density, the agglomerated particles themselves are insufficiently strong, and during various operations before blending with the resin, the particles are destroyed or the surface primary particles fall off, and the primary particles isolated by these crushings fill the gaps between the agglomerated particles during mercury intrusion measurement.

Here, the interparticle volume will be described in more detail with reference to FIGS.
For example, when the thickness direction (c-axis direction) of the primary particles near the outermost surface of the agglomerated particles coincides with the radial direction of the agglomerated particles, the interparticle volume is small. Such agglomerated particles have the advantage that they have good coatability when a thermally conductive resin sheet is produced by a wet coating method and that it is easy to obtain a sheet with few residual voids. However, because the primary particles near the outermost surface lie flat so as to cover the agglomerated particles, the thermal resistance at the interface between the agglomerated particle surface and the resin or at the contact interface between multiple agglomerated particles may increase.
In addition, when a plurality of aggregated particles are packed so as to come into contact with each other in a thermally conductive resin sheet, the aggregated particles shown in Fig. 4 and Fig. 5 conduct heat to each other at the surface portions of the aggregated particles through the thickness direction (c-axis direction) of the primary particles. Since the thermal conductivity of the primary particles is low in the thickness direction, there is a limit to the thermal conductivity of the thermally conductive resin sheet in the thickness direction.
In addition, for example, when the primary particles constituting the aggregated particles are excessively large and lack denseness, the interparticle volume is small. Since such aggregated particles lack strength, for example, when a sheet is produced by press molding, the aggregated particles may be broken or the surface primary particles may fall off. Since the isolated primary particles are flat or scaly, the thickness direction (c-axis direction) of the particles may be oriented in the plane direction of the sheet, and the thermal conductivity in the thickness direction of the thermal conductive resin sheet may decrease.

On the other hand, for example, as shown in Figures 6 and 7, when the surface direction (ab axis direction) of the primary particles near the outermost surface of the agglomerated particles is aligned with the radial direction of the agglomerated particles, and these oriented surface primary particles are densely present, the interparticle volume becomes large.
When such aggregated particles are filled in a thermal conductive resin sheet so that multiple aggregated particles contact each other, the contact area between the particles is increased compared to aggregated particles in which the primary particles near the outermost surface lie down to cover the aggregated particles. Therefore, the thermal resistance between the particles is reduced, and the thermal conductivity in the thickness direction of the thermal conductive resin sheet is increased. In addition, if the particle gap volume is large, assuming that the intraparticle pore volume, average particle diameter, and particle size distribution are the same, the contact surface of multiple aggregated particles is deformed by the pressurizing process such as sheet molding as shown in Figure 9, and the contact area increases so that polyhedrons contact each other, forming a state similar to surface contact, thereby further reducing the thermal resistance between the particles.
In addition, for example, when the primary particles constituting the aggregated particles are appropriately small and dense, the interparticle volume is also large. Such aggregated particles have good strength, so that the aggregated particles can be prevented from being destroyed during molding, and the surface primary particles can be prevented from falling off. Therefore, the isolation of the primary particles can be suppressed, and the thermal conductivity in the thickness direction of the thermal conductive resin sheet can be improved.

From the above viewpoints, the particle gap volume _B1 is preferably 0.50 mL/g or more, more preferably 0.55 mL/g or more, even more preferably 0.60 mL/g or more, and even more preferably 0.65 mL/g or more. Furthermore, it is preferably 0.70 mL/g or more, more preferably 0.75 mL/g or more, even more preferably 0.80 mL/g or more, and even more preferably 0.85 mL/g or more. In addition, the particle gap volume _B1 is preferably 1.0 mL/g or less, more preferably 0.95 mL/g or less, and even more preferably 0.90 mL/g or less.

In the present invention, _B1 /( _A1 + _B1 ) indicates the ratio of the interparticle volume to the total pore volume. _B1 /( _A1 + _B1 ) is preferably 0.60 or more, more preferably 0.62 or more, even more preferably 0.64 or more, and even more preferably 0.66 or more. _B1 /( _A1 + _B1 ) is preferably less than 1.00, more preferably 0.90 or less, even more preferably 0.80 or less, even more preferably 0.75 or less, and especially preferably 0.70 or less.
As described above, the strength of agglomerated particles, which is represented by the intraparticle pore volume and interparticle gap volume, the orientation state of primary particles, etc., affect the voltage resistance performance and thermal conductivity. The inventors have found that the balance between the intraparticle pore volume and the interparticle gap volume particularly affects these.
More specifically, when B ₁ /(A ₁ +B ₁ ) is in the above range, the aggregate particles have sufficient strength, so that the collapse of the aggregate particles can be prevented during molding, etc., and the primary particles on the surface of the aggregate particles are radially oriented, so that a heat conduction path between the aggregate particles can be sufficiently formed. In addition, when B ₁ /(A ₁ +B ₁ ) is in the above range, the inside of the aggregate particles maintains a dense structure with isotropic thermal conductivity, while the area where the primary particles on the surface of the aggregate particles are radially oriented deforms appropriately during molding, so that it has a so-called "hard inside, soft outside" structure, so that the contact area between adjacent aggregate particles increases, and the thermal conductivity in the thickness direction of the thermal conductive resin sheet can be further increased.
Furthermore, when B ₁ /(A ₁ +B ₁ ) is in the above range, the inside of the agglomerated particles is dense while the resin sufficiently penetrates into the particles, so that the generation of voids during molding can be suppressed and the voltage resistance performance can be further improved.

Examples of boron nitride agglomerated particles having the above-mentioned _B1 /( _A1 + _B1 ) of 0.60 or more include agglomerated particles having a house of cards structure. After obtaining boron nitride agglomerated particles by a known method, the particles may be subjected to a physical or chemical surface roughening treatment to adjust the _B1 /( _A1 + _B1 ) to 0.60 or more.
Boron nitride agglomerated particles commercially available from various companies use boron nitride primary particles that are relatively highly crystallized during granulation, and as a result, stable primary particle faces (ab faces) tend to stack with each other, resulting in a "cabbage structure" in which the thickness direction of the primary particles near the outermost surface of the agglomerated particles coincides with the radial direction of the agglomerated particles. In agglomerated particles with a "cabbage structure," the primary particles near the outermost surface lie flat so as to cover the agglomerated particles, making it difficult for _B1 /( _A1 + _B1 ) to be 0.60 or more. Therefore, when using agglomerated particles with a "cabbage structure," it is preferable to adjust the surface roughening treatment as described above.

The boron nitride agglomerated particles of the present invention preferably have a spherical shape.
The term "spherical" generally means that the aspect ratio (ratio of major axis to minor axis) is from 1 to 2, preferably from 1 to 1.75, more preferably from 1 to 1.5, and even more preferably from 1 to 1.4.
The aspect ratio can be determined by randomly selecting 200 or more particles from an image of the cross section of a thermally conductive resin sheet taken with a scanning electron microscope (SEM), determining the ratio of the long axis to the short axis of each particle, and calculating the average value.

Alternatively, the "circularity" measured using a particle image analyzer (Malvern Instruments, Morphologi G3S) may be used as an index of "sphericity". This measurement involves observing a projected planar image (two-dimensional image) of the particle, but the "sphericity" can be evaluated by increasing the number of measurements and averaging them.
The circularity is preferably 0.90 or more, more preferably 0.92 or more, even more preferably 0.94 or more, and even more preferably 0.96 or more, with 1 being the upper limit.

From the viewpoint of improving thermal conductivity, the aggregated structure of the boron nitride aggregated particles is preferably a card house structure.
The aggregate structure of the boron nitride aggregated particles can be confirmed by a scanning electron microscope (SEM).

The house of cards structure is a structure in which plate-like particles are stacked in a complex manner without being oriented, as described in "Ceramics 43 No. 2" (published by the Ceramic Society of Japan in 2008). More specifically, it is a structure in which the flat surface of a primary particle forming an aggregated particle is in contact with the end surface of another primary particle present within the aggregated particle. A schematic diagram of the house of cards structure is shown in Figure 8.
The card-house structured aggregated particles have a very high fracture strength due to their structure, and do not collapse even during the pressurizing process performed during the molding of the thermally conductive resin sheet. Therefore, the primary particles that are usually oriented in the longitudinal direction of the thermally conductive resin sheet can be made to exist in a random direction. Therefore, by using the card-house structured aggregated particles, the proportion of the primary particles that have the ab plane oriented in the thickness direction of the thermally conductive resin sheet can be increased, so that the heat can be effectively conducted in the thickness direction of the sheet, and the thermal conductivity in the thickness direction can be further increased.

The boron nitride agglomerated particles having a house-of-cards structure can be produced, for example, by the method described in International Publication No. 2015/119198. It is preferable that the boron nitride agglomerated particles of the present invention are not subjected to a treatment that applies force to the surface of a ball mill or the like.

When using boron nitride agglomerated particles having a house-of-card structure, the particles may be surface-treated with a surface treatment agent.
The surface treatment agent may be a known surface treatment agent such as a silane coupling treatment. In general, direct affinity or adhesion between the boron nitride agglomerated particles and the thermoplastic resin is not observed in many cases, and this is also the case when the boron nitride agglomerated particles having a card house structure are used as the boron nitride agglomerated particles. It is believed that the thermal conductivity attenuation at the interface can be further reduced by increasing the adhesion at the interface between the boron nitride agglomerated particles and the matrix resin by chemical treatment.

The boron nitride agglomerated particles of the present invention can have a larger particle size than when primary particles are used as they are.
By increasing the particle size of the boron nitride agglomerated particles, the number of heat transfer paths between the boron nitride agglomerated particles via the thermoplastic resin with low thermal conductivity can be reduced, and therefore the increase in thermal resistance in the heat transfer paths in the thickness direction can be reduced.

From the above viewpoints, the lower limit of the volume-based maximum particle diameter Dmax (hereinafter also referred to as "maximum particle diameter") of the boron nitride agglomerated particles is preferably 20 μm or more, more preferably 30 μm or more, and even more preferably 50 μm or more. On the other hand, the upper limit of the maximum particle diameter Dmax is preferably 300 μm or less, more preferably 200 μm or less, even more preferably 100 μm or less, and even more preferably 90 μm or less.

The lower limit of the volume-based average particle diameter D50 of the boron nitride agglomerated particles is preferably 10 μm or more, more preferably 20 μm or more, and even more preferably 30 μm or more. On the other hand, the upper limit of the average particle diameter D50 is preferably 200 μm or less, more preferably 100 μm or less, even more preferably 80 μm or less, and even more preferably 60 μm or less.

By having the maximum particle size of the boron nitride agglomerated particles be equal to or less than the above upper limit, when the boron nitride agglomerated particles are contained in a matrix resin, the interface between the matrix resin and the boron nitride agglomerated particles is reduced, resulting in lower thermal resistance, high thermal conductivity, and the formation of a high-quality film without surface roughness. By having the maximum particle size be equal to or greater than the above lower limit, it is possible to obtain a sufficient thermal conductivity improvement effect of the boron nitride agglomerated particles required for power semiconductor devices.

It is also considered that the effect of the thermal resistance at the interface between the matrix resin and the boron nitride agglomerated particles on the thickness of the thermal conductive resin sheet becomes significant when the size of the boron nitride agglomerated particles is 1/10 or less of the thickness of the thermal conductive resin sheet. In particular, in the case of power semiconductor devices, thermal conductive resin sheets with thicknesses of 100 μm to 300 μm are often used, so from the viewpoint of thermal conductivity as well, it is preferable that the volume-based maximum particle size Dmax of the boron nitride agglomerated particles be larger than the above lower limit.
Furthermore, by having the maximum particle diameter Dmax of the boron nitride agglomerated particles be equal to or greater than the above lower limit, not only is the increase in thermal resistance caused by the interface between the boron nitride agglomerated particles and the matrix resin suppressed, but the number of required thermal conduction paths between particles is reduced, increasing the probability that they will be connected from one side to the other side in the thickness direction of the thermal conductive resin sheet.
Furthermore, by having the maximum particle diameter Dmax of the boron nitride agglomerated particles be equal to or greater than the above lower limit, the interfacial area between the matrix resin and the boron nitride agglomerated particles is smaller than when the same mass of particles having a Dmax smaller than the above lower limit is used. This makes it possible to reduce the occurrence of voids that tend to occur at the interface between the matrix resin and the boron nitride agglomerated particles in the thermally conductive resin sheet, making it easier to obtain excellent voltage resistance characteristics.
On the other hand, by having the volume-based maximum particle diameter Dmax of the boron nitride agglomerated particles be equal to or less than the above upper limit, protrusion of the boron nitride agglomerated particles onto the surface of the thermal conductive resin sheet is suppressed, and a good surface shape without surface roughness can be obtained. Therefore, when a sheet is produced by bonding it to a copper substrate, sufficient adhesion can be obtained and excellent voltage resistance characteristics can be obtained.

The ratio (Dmax/thickness) of the size of the boron nitride agglomerated particles to the thickness of the thermally conductive resin sheet is preferably 0.3 or more and 1.0 or less, more preferably 0.35 or more or 0.95 or less, and even more preferably 0.4 or more or 0.9 or less.

The maximum particle size Dmax and the average particle size D50 of the boron nitride agglomerated particles can be measured, for example, by the following method.
A sample in which boron nitride agglomerated particles are dispersed in a solvent, specifically, a sample in which boron nitride agglomerated particles are dispersed in a pure water medium containing sodium hexametaphosphate as a dispersion stabilizer, is subjected to particle size distribution measurement using a laser diffraction/scattering particle size distribution measuring device LA-920 (manufactured by HORIBA, Ltd.), and the maximum particle diameter Dmax and average particle diameter D50 of the boron nitride agglomerated particles can be obtained from the particle size distribution obtained.
The maximum particle size and the average particle size can also be determined using a dry particle size distribution measuring device such as Morphologi G3S (manufactured by Malvern Instruments).
The maximum particle diameter Dmax and average particle diameter D50 of the boron nitride agglomerated particles added to the thermoplastic resin can also be measured in a similar manner to that described above, by dissolving and removing the thermoplastic resin in a solvent (including a heated solvent), or by swelling the thermoplastic resin to reduce the adhesive strength with the boron nitride agglomerated particles and then physically removing it, and further by heating the resin component in the atmosphere to incinerate it and remove it.

(3) Content of each component The lower limit of the content of the thermoplastic resin in 100% by mass of the resin composition of the present invention is preferably 15% by mass or more, more preferably 20% by mass or more. On the other hand, the upper limit of the content of the thermoplastic resin is preferably 40% by mass or less, more preferably 35% by mass or less.
The lower limit of the content of the boron nitride agglomerated particles in 100% by mass of the resin composition of the present invention is preferably 60% by mass or more, more preferably 65% by mass or more, while the upper limit of the content of the boron nitride agglomerated particles is preferably 85% by mass or less, more preferably 80% by mass or less.
When the content of the boron nitride agglomerated particles is equal to or greater than the lower limit, the effect of improving thermal conductivity and the effect of controlling the linear expansion coefficient due to the boron nitride agglomerated particles are satisfactorily exhibited, whereas when the content of the boron nitride agglomerated particles is equal to or less than the upper limit, the moldability of the resin composition and the interfacial adhesion with different materials are improved.

In general, the compounding ratio of a thermally conductive resin composition is often specified by the volume fraction of the matrix resin and the boron nitride agglomerated particles (hence, the area ratio in the cross section of the thermally conductive resin sheet). The thermal conductivity in the thickness direction of the thermally conductive resin sheet is not determined only by the volume fraction, but is affected by various factors such as the above-mentioned preferred particle size, particle orientation state, and particle shape, so in the present invention, the mass fraction is used for convenience in practical compounding.
In particular, when using boron nitride agglomerated particles with a house-of-cards structure, the internal structure of the particle exhibits a house-of-cards structure, and a large number of flat boron nitride primary particles oriented in the radial direction are formed on the particle surface in a protruding state, which is called a burr shape or a confetti shape, and the protrusions of adjacent house-of-cards structure particles come into physical contact with each other, forming a heat transfer path with low thermal resistance in the thickness direction. Therefore, it may not be easy to determine the volume fraction of the matrix resin and the boron nitride agglomerated particles by observation with a normal scanning electron microscope (SEM). Furthermore, when the amount of the boron nitride agglomerated particles with a house-of-cards structure having the characteristics of the present invention is increased, the pressure applied during the hot press molding of the resin composition causes the particles to deform at the contact points, rather than being in point contact as spherical particles, and the contact points are observed to be in linear, i.e., planar contact. When such a contact state is reached, the addition of boron nitride with a house-of-cards structure enables the formation of an efficient heat transfer path. However, even in such a case, it is not easy to determine the respective volume fractions of the matrix resin and the boron nitride agglomerated particles by SEM observation.

The resin composition of the present invention may contain other components in addition to the thermoplastic resin and the boron nitride agglomerated particles, although from the viewpoint of increasing thermal conductivity, it is preferable that the resin composition does not contain other components.
Other components include various antioxidants such as phosphorus-based and phenol-based, phenol acrylate-based and other process stabilizers, heat stabilizers, hindered amine radical scavengers (HAAS), impact modifiers, processing aids, metal deactivators, copper inhibitors, antistatic agents, flame retardants, additives such as silane coupling agents that improve the affinity at the interface between the boron nitride agglomerated particles and thermoplastic resins, as well as additives such as silane coupling agents that can be expected to increase the adhesive strength between the resin sheet and the metal plate material, extenders, etc. When these additives are used, the amount added may be within the range of the amount usually used for these purposes.

2. Thermally conductive resin sheet The thermally conductive resin sheet of the present invention is made of the above-mentioned resin composition, and has the characteristics of excellent moisture absorption reflow resistance and being less susceptible to interfacial peeling due to thermal expansion and thermal contraction when laminated with a metal plate.

The thermal conductivity of the thermal conductive resin sheet in the thickness direction at 25° C. is preferably 16 W/m·K or more, more preferably 18 W/m·K or more, even more preferably 19 W/m·K or more, and even more preferably 20 W/m·K or more. By having a thermal conductivity in the thickness direction equal to or more than the above lower limit, the sheet can be suitably used in power semiconductor devices that operate at high temperatures.
The thermal conductivity can be adjusted by the type of thermoplastic resin, physical properties such as melt viscosity, the value of _B1 /( _A1 + _B1 ) of the boron nitride agglomerated particles, the structure and content, the method of mixing the thermoplastic resin and the boron nitride agglomerated particles, and the conditions in the heating and kneading process described below.

The thermal conductivity can be measured by the following method.
First, the thermal diffusivity a ( ^mm2 /sec) in the thickness direction of the resin sheet is measured by the laser flash method at a measurement temperature of 25° C. Since there are no JIS standards for thermal diffusivity and thermal conductivity of resin-based materials, the measurement is performed in accordance with JIS R1611:2010 (Method for measuring thermal diffusivity, specific heat capacity, and thermal conductivity of fine ceramics by the flash method).
In addition, since JIS R1611:2010 specifies that "the thickness of the sample is 0.5 mm or more and 5 mm or less," the thickness of the resin sheet is adjusted to 0.5 mm or more for measurement. If the thickness of the resin sheet is less than 0.5 mm, multiple sheets may be stacked to adjust the total thickness to 0.5 mm or more for measurement.
Next, the density ρ (g/m ³ ) of the resin sheet is determined by Archimedes' method in accordance with JIS K6268.
Furthermore, in accordance with JIS K7123, the specific heat capacity c (J/(g·K)) at 25° C. is measured using a DSC measuring device.
From these measured values, the thermal conductivity in the sheet thickness direction at 25° C. can be calculated as "H=a×ρ×c".

The lower limit of the thickness of the thermally conductive resin sheet is preferably 50 μm or more, more preferably 60 μm or more, and even more preferably 70 μm or more, while the upper limit of the thickness is preferably 300 μm or less, more preferably 200 μm or less, and even more preferably 160 μm or less.
By making the thickness of the thermally conductive resin sheet 50 μm or more, sufficient voltage resistance characteristics can be ensured. On the other hand, by making the thickness 300 μm or less, it is possible to achieve miniaturization and thinning, particularly when the thermally conductive resin sheet is used in a power semiconductor device, and further, compared to an insulating thermally conductive layer made of a ceramic material, it is possible to obtain the effect of reducing the thermal resistance in the thickness direction by making the thickness thinner.

In the thermally conductive resin sheet of the present invention, in order to reduce the anisotropy of thermal conductivity and increase the thermal conductivity in the thickness direction, it is preferable to reduce the orientation of the primary particles of the boron nitride agglomerated particles.
The orientation of the primary particles can be evaluated by determining the ratio "I(002)/I(100)" of the diffraction peak intensity of the (002) plane to the diffraction peak intensity of the (100) plane when the thermally conductive resin sheet is measured by X-ray diffraction.
The ratio "I(002)/I(100)" is preferably 20 or less, more preferably 17 or less, and even more preferably 15 or less. If the ratio exceeds 20, the boron nitride agglomerated particles collapse in the melt-kneading step or are excessively crushed in the pressurizing step, resulting in an increase in primary particles that are parallel to the surface direction of the thermal conductive resin sheet or form a small angle with the surface direction. In this case, even if the content of the boron nitride agglomerated particles is increased, it is difficult to increase the thermal conductivity in the thickness direction.
The lower limit of the ratio "I(002)/I(100)" is not particularly limited. For example, when using boron nitride agglomerated particles having a house-of-cards structure with high isotropy and therefore very little orientation of the boron nitride primary particles in a specific direction, the ratio "I(002)/I(100)" of the particles alone is about 4.5 to 6.6, so that 4.5 is considered to be the lower limit, except when intentional orientation manipulation is performed.

In the thermally conductive resin sheet of the present invention, when the boron nitride agglomerated particles contained in the residual ash when heated to 700°C have an intraparticle pore volume _A2 and an interparticle volume _B2 measured by mercury intrusion porosimetry, it is preferable that _A2 / _A1 is 0.70 or more and _B2 / _B1 is 0.85 or less.
_A2 / _A1 being 0.70 or more indicates that the state of the primary particles inside the agglomerated particles does not change significantly before and after sheet formation, and therefore the strength of the agglomerated particles is relatively high. Therefore, _A2 / _A1 being 0.70 or more indicates that the structure forming the thermal conductive path inside the original boron nitride agglomerated particles is sufficiently maintained even after sheet formation. It can also be said that there is little generation of isolated primary particles due to the crushing of the agglomerated particles, even though the particles are deformed by the pressure during molding. The little generation of isolated primary particles has the effect of reducing the thermal resistance in the thickness direction of the sheet. The reason why isolated primary particles hinder efficient thermal conduction in the thickness direction is that the primary particles are flat or scaly particles, and therefore the thickness direction (c-axis direction) of the particles is oriented in the plane direction of the sheet in the molded sheet.
On the other hand, _B2 / _B1 being 0.85 or less indicates that the surface of the agglomerated particles is deformed relatively significantly before and after sheet forming. This deformation occurs because the contact area of the contact surfaces of multiple agglomerated particles increases so that polyhedrons come into contact with each other, forming a surface-like contact state, as shown in Fig. 9.
Therefore, when boron nitride agglomerated particles are present in a sheet with both A ₂ /A ₁ and B ₂ /B ₁ satisfying the above numerical ranges, high thermal conductivity can be obtained.

From the above viewpoints, _A2 / _A1 is preferably 0.70 or more, more preferably 0.72 or more, even more preferably 0.75 or more, and even more preferably 0.80 or more. Also, _A2 / _A1 is preferably 1.0 or less, more preferably 0.90 or less, and even more preferably 0.88 or less.
On the other hand, _B2 / _B1 is preferably 0.85 or less, more preferably 0.80 or less, even more preferably 0.75 or less, even more preferably 0.70 or less, still more preferably 0.65 or less, more preferably 0.60 or less, even more preferably _0.50 or _less , and preferably 0.40 or more, more preferably 0.45 or more.

The intraparticle pore volume _A2 of the boron nitride agglomerated particles contained in the residual ash is preferably 0.20 mL/g or more, more preferably 0.25 mL/g or more, even more preferably 0.30 mL/g or more, even more preferably 0.32 mL/g or more, and particularly preferably 0.34 mL/g or more. Also, the intraparticle pore volume _A2 is preferably 0.60 mL/g or less, more preferably 0.50 mL/g or less, even more preferably 0.45 mL/g or less, and even more preferably 0.40 mL/g or less.

The interparticle volume _B2 of the boron nitride agglomerated particles contained in the residual ash is preferably 0.35 mL/g or more, more preferably 0.40 mL/g or more, even more preferably 0.45 mL/g or more, and even more preferably 0.50 mL/g or more. The interparticle volume _B2 is preferably 0.80 mL/g or less, more preferably 0.70 mL/g or less, and even more preferably 0.65 mL/g or less.

The circularity of the boron nitride agglomerated particles contained in the residual ash is preferably 0.85 or more, more preferably 0.90 or more, even more preferably 0.92 or more, and even more preferably 0.94 or more.

3. Method for Producing Thermally Conductive Resin Sheet Hereinafter, an example of the method for producing a thermally conductive resin sheet of the present invention will be described.
An example of the method for producing the thermally conductive resin sheet of the present invention includes a method including a mixing step and a press molding step.

As a conventional manufacturing method, there is a method (wet coating method) in which a matrix resin and a slurry of boron nitride agglomerated particles dispersed in a solvent are applied to a substrate to obtain a thermally conductive resin film. However, when the substrate is coated with the slurry, or when the solvent is not dried sufficiently, bubbles are generated due to the residual solvent in the coating film, which may result in a decrease in voltage resistance. In particular, when boron nitride agglomerated particles are used in which the intraparticle pore volume of _the boron nitride agglomerated particles is _{A1 and the} interparticle volume is _B1 , the primary particles on the agglomerated particle surface are radial, so that the agglomerated particle surface has many irregularities, the slurry is easily thickened, and the slurry is more likely to contain bubbles when kneaded with the resin or when coated on the substrate, so that bubbles are more likely to remain in the resin film, which is likely to cause a decrease in voltage resistance.
Furthermore, in the above-mentioned wet coating method, when the amount of boron nitride agglomerated particles added is increased in order to increase thermal conductivity, streaks may occur during coating, resulting in poor productivity. These streaks are particularly likely to occur when using boron nitride agglomerated particles having the above-mentioned _B1 /( _A1 + _B1 ) of 0.60 or more.

In the manufacturing method of the present invention, the thermoplastic resin is pressed into the voids in the boron nitride agglomerated particles by forming the sheet at a temperature at which the thermoplastic resin exhibits fluidity, so that the sheet is less likely to contain air bubbles. In addition, since the sheet can be obtained without using a solvent in the present invention, there is no foaming caused by the residual solvent. Therefore, the manufacturing method of the present invention can improve the voltage resistance performance of the thermal conductive resin sheet.
In addition, since the manufacturing method of the present invention does not involve a coating step, even if the above-mentioned boron nitride agglomerated particles having _B1 /( _A1 + _B1 ) of 0.60 or more are used, problems due to coating, such as the occurrence of streaks, do not occur, and productivity is good.

As described in the examples of Patent Documents 4 and 5, it was common to use a thermosetting resin as a matrix resin in the past, and therefore wet coating methods were often used to manufacture thermally conductive resin sheets. In this wet coating method, if the aggregated particle surface is uneven, the above-mentioned problems often occur, so it is considered that aggregated particles with as little unevenness on the aggregated particle surface as possible (i.e., the above-mentioned ratio _B1 /( _A1 + _B1 ) is smaller than 0.60) are preferable.
On the other hand, from the viewpoint of thermal conductivity, it is preferable that the primary particles on the surface of the aggregated particles are radial, and such particles often have a ratio of B ₁ /(A ₁ +B ₁ ) greater than 0.60.
Thus, when the conventional manufacturing method is used, there is a trade-off between the improvement in voltage resistance performance and productivity and the improvement in thermal conductivity in terms of the relationship between the intraparticle pore volume _A1 and the interparticle space volume _B1 of the boron nitride agglomerated particles, i.e., _B1 /( _A1 + _B1 ).
According to the manufacturing method of the present invention, agglomerated particles having a _B1 /( _A1 + _B1 ) ratio of 0.60 or more can be used without having to consider the problems that may occur in wet coating methods, and therefore it is possible to improve the voltage resistance and productivity while also increasing the thermal conductivity.

(1) Mixing Step In the mixing step, a powder made of a thermoplastic resin and boron nitride agglomerated particles in which the intraparticle pore volume of the boron nitride agglomerated particles is _A1 and the interparticle volume is _B1 , and _B1 /( _A1 + _B1 ) is 0.60 or more, are stirred and mixed at room temperature.
A conventional manufacturing method involves heating, melting and kneading a matrix resin and boron nitride agglomerated particles. However, this melting and kneading can cause shear fracture of the boron nitride agglomerated particles. In particular, when using boron nitride agglomerated particles with a _B1 /( _A1 + _B1 ) ratio of 0.60 or more, the surface is highly uneven, and shear fracture is likely to occur.
In the present invention, instead of performing the heating and melting kneading, a powder made of a thermoplastic resin and boron nitride agglomerated particles having a _B1 /( _A1 + _B1 ) ratio of 0.60 or more are stirred and mixed at room temperature, which makes it difficult for the agglomerated particles to be subjected to shear fracture and increases the thermal conductivity of the resulting sheet.

(2) Press Molding Step In the press molding step, the mixture obtained in the mixing step is pressed to be molded into a sheet.

As the press molding method, various known press machines for molding thermoplastic resins can be used.
From the viewpoint of preventing deterioration of the resin during the heat pressing, it is particularly preferable to use a vacuum press capable of reducing the amount of oxygen in the press during heating, or a press equipped with a nitrogen replacement device.

In the press molding process, it is preferable to set the pressurizing pressure in order to form the molten kneaded material into a sheet body of uniform thickness, as well as to bond the added boron nitride agglomerated particles together and further to form heat paths by deforming the particle surfaces at the joints, and to eliminate voids and spaces within the sheet.
From this viewpoint, the pressure in the press molding step, as the actual pressure applied to the sample, is usually 8 MPa or more, preferably 9 MPa or more, more preferably 10 MPa or more, and is preferably 50 MPa or less, more preferably 40 MPa or less, and further preferably 30 MPa or less.
By setting the pressure during pressing to the upper limit or less, it is possible to prevent the boron nitride agglomerated particles from being crushed, and a thermally conductive resin sheet having high thermal conductivity can be obtained. Also, by setting the pressing pressure to the lower limit or more, contact between the boron nitride agglomerated particles is improved, making it easier to form a thermally conductive path, and a sheet having high thermal conductivity can be obtained. In addition, since the voids in the resin sheet can be reduced, a thermally conductive resin sheet having a high breakdown voltage can be obtained even after a moisture absorption reflow test.

The set temperature of the press machine in the press molding step is preferably a temperature at which the thermoplastic resin exhibits fluidity, for example, the melting point of the thermoplastic resin as the main component + 30° C. For example, when a polyether ketone resin is used as the main component, the set temperature of the press machine is preferably 370° C. to 440° C., and more preferably 380° C. or higher or 420° C. or lower.
By carrying out press molding at a temperature in this range, the obtained thermally conductive resin sheet can be given good thickness uniformity and high thermal conductivity due to good contact between the added boron nitride agglomerated particles. If the molding temperature is 370°C or higher, the resin viscosity is reduced to a level sufficient for shaping processing, and the molded thermally conductive resin sheet can be given sufficient thickness uniformity. On the other hand, if the set temperature of the press machine is 440°C or lower, deterioration of the resin itself and deterioration of the physical properties of the molded thermally conductive resin sheet can be suppressed.

The pressurization time is usually 30 seconds or more, preferably 1 minute or more, more preferably 3 minutes or more, and even more preferably 5 minutes or more, and is preferably 1 hour or less, more preferably 30 minutes or less, and even more preferably 15 minutes or less.
By being equal to or less than the upper limit, the manufacturing process time of the thermally conductive resin sheet can be reduced, and the cycle time can be shortened compared to a thermally conductive resin sheet using a heat-resistant thermosetting resin, which tends to reduce production costs. Also, by being equal to or more than the lower limit, the thickness of the thermally conductive resin sheet can be sufficiently uniform, internal gaps and voids can be sufficiently removed, and unevenness in thermal conductivity performance and voltage resistance characteristics can be prevented.

4. Laminated Heat Dissipation Sheet The laminated heat dissipation sheet of the present invention is obtained by laminating a heat dissipation metal layer containing a heat dissipation material on one surface of the above-mentioned thermally conductive resin sheet of the present invention.
The heat dissipating material is not particularly limited as long as it is made of a material having good thermal conductivity. In particular, in order to increase the thermal conductivity in the laminated structure, it is preferable to use a heat dissipating metal material, and more preferably to use a flat metal material.
The type of metal material is not particularly limited, but among them, a copper plate, an aluminum plate, an aluminum alloy plate, etc. are preferable because they have good thermal conductivity and are relatively inexpensive.

When a flat metal material is used as the heat dissipating metal material in the laminated heat dissipation sheet, the thickness of the metal material is preferably 0.03 to 6 mm, and more preferably 0.1 mm or more or 5 mm or less, in order to ensure sufficient heat dissipation.

Regarding adhesion to heat-dissipating metal materials, the surface of the metal material on the side that is laminated with the thermally conductive resin sheet may be subjected to surface treatments such as soft etching, burnt plating, and oxidation-reduction treatment to roughen the surface; plating with various metals and metal alloys to ensure adhesion durability; organic surface treatments including amino- and mercapto-based silane coupling treatments; and surface treatments with organic/inorganic composite materials. By carrying out these surface treatments, it is possible to further improve the initial adhesive strength, the durability of the adhesive strength, and the effect of suppressing interfacial peeling after a moisture absorption reflow test.

On the other hand, the side of the heat dissipating metal material opposite to the side laminated with the thermally conductive resin sheet does not have to be a simple flat plate, and may be subjected to processing such as increasing the surface area in order to ensure a contact area with the cooling medium, which is a gas or liquid.
Examples of processing for increasing the surface area include increasing the surface area by roughening the surface with blasting processing, etc., directly forming V-shaped or rectangular grooves or various shapes of unevenness on the heat dissipating metal material by cutting or pressing, joining another metal material that has been further processed to increase the surface area by casting, diffusion bonding, bolt fastening, soldering, brazing, etc. to the heat dissipating metal layer made of a flat metal material, or embedding a metal pin, etc. Furthermore, it is also possible to directly press-laminated a thermally conductive resin sheet onto a heat dissipating metal layer having a cavity for passing a cooling medium. However, since the pressing pressure between the thermally conductive resin sheet and the heat dissipating metal plate is relatively high, in these cases, it is preferable to integrate the laminated and integrated flat metal material and the thermally conductive resin sheet with a grooved metal plate or a metal layer having a cavity for flowing a refrigerant in a later process by soldering, brazing, bolt joining, etc.

When laminating and integrating the heat-dissipating metal material and the thermally conductive resin sheet in the laminated heat dissipation sheet, press molding, which is a batch process, can be preferably used. In this case, the press equipment and press conditions are the same as the range of press molding conditions for obtaining the thermally conductive resin sheet described above.

The heat dissipating circuit board of the present invention has the above-mentioned laminated heat dissipation sheet, that is, the heat dissipating metal layer is laminated on one surface of the thermally conductive resin sheet of the present invention, and a circuit board is formed on the surface of the thermally conductive resin sheet other than the heat dissipating metal layer by, for example, etching or the like.

The heat dissipating circuit board is preferably configured as an integrated structure of "heat dissipating metal layer/thermally conductive resin sheet/conductive circuit". The state before the circuit etching is, for example, an integrated structure of "heat dissipating metal layer/thermally conductive resin sheet/conductive circuit forming metal layer", in which the conductive circuit forming metal layer is flat and formed on the entire surface of one side of the thermally conductive resin sheet, or formed on a partial area.

The material of the metal layer for forming the conductive circuit is not particularly limited. In particular, it is generally preferable to form it from a copper thin plate having a thickness of 0.05 mm to 1.2 mm from the viewpoints of electrical conductivity, etching properties, cost, etc.

The breakdown voltage of the heat dissipating circuit board is preferably 40 kV/mm or more, more preferably 50 kV/mm or more, even more preferably 60 kV/mm or more, and even more preferably 80 kV/mm or more. With a breakdown voltage of 40 kV/mm or more, even a thermally conductive resin sheet with a thickness of, for example, 100 μm can have a breakdown voltage of 4 kV or more, and with a breakdown voltage of 80 kV/mm or more, even a thickness of 50 μm can have a breakdown voltage of 4 kV or more. Therefore, while using a thin thermally conductive resin layer that is advantageous in terms of thermal resistance, it has sufficient voltage resistance performance and can suppress the occurrence of breakdown when a high voltage is applied.

6. Power Semiconductor Device The thermally conductive resin sheet of the present invention and the laminated heat dissipation sheet of the present invention can be suitably used as a heat dissipation sheet for power semiconductor devices, and a highly reliable power semiconductor module can be realized.
The power semiconductor device is a power semiconductor device that uses the above-mentioned thermally conductive resin sheet or the above-mentioned laminated heat dissipation sheet, and the above-mentioned thermally conductive resin sheet or the above-mentioned laminated heat dissipation sheet is mounted on a power semiconductor device apparatus as a heat dissipation circuit board.
The power semiconductor device has a high thermal conductivity and hence a heat dissipation effect, which allows it to achieve high output and high density with high reliability.
In a power semiconductor device, conventionally known materials can be appropriately used for the materials other than the thermally conductive resin sheet or laminated heat dissipation sheet, such as aluminum wiring, sealing material, packaging material, heat sink, thermal paste, and solder.

<Explanation of terms>
In the present invention, when expressed as "X to Y" (X and Y are arbitrary numbers), unless otherwise specified, it includes the meaning of "X or more and Y or less", as well as "preferably larger than X" or "preferably smaller than Y".
Furthermore, when it is expressed as "X or more" (X is any number) or "Y or less" (Y is any number), it also includes the intention that "it is preferably greater than X" or "it is preferably less than Y".
In the present invention, the term "sheet" conceptually includes a sheet, a film, and a tape.

The present invention will be explained in more detail below with reference to examples. However, the present invention is not limited to the following examples as long as it does not depart from the gist of the invention.

<Examples 1 to 3 and Comparative Examples 1 to 4>
The materials used, the manufacturing method, and the measurement conditions and evaluation methods of the thermally conductive resin sheets in Examples 1 to 3 and Comparative Examples 1 to 4 are as follows.

[Materials used]
(Thermoplastic resin)
Thermoplastic resin 1: Polyether ether ketone "KetaSpire KT-880FP" (manufactured by Solvay, melting point: 343°C, melt viscosity: 0.15 kPa·s (400°C), average particle size (D50): 30.0 to 45.0 μm, MFR: 86 g/10 min, mass average molecular weight (Mw): 58,000 was used.

(Thermosetting resin)
Thermosetting resin 1: Epoxy resin composition (bisphenol F type epoxy resin (manufactured by Mitsubishi Chemical Corporation, weight average molecular weight in terms of polystyrene: 60,000) 8.74 parts by mass, hydrogenated bisphenol A type liquid epoxy resin (manufactured by Mitsubishi Chemical Corporation) 10.93 parts by mass, p-aminophenol type liquid epoxy resin (manufactured by Mitsubishi Chemical Corporation) 2.62 parts by mass, phenol resin curing agent "MEH-8000H" (manufactured by Meiwa Chemical Industry Co., Ltd.) 5.73 parts by mass, 1-cyanoethyl-2-undecylimidazole "C11Z-CN" (manufactured by Shikoku Chemical Industry Co., Ltd., molecular weight 275) as a curing catalyst 0.48 parts by mass) was used.

Boron nitride agglomerated particles (thermally conductive filler)
The thermally conductive filler used as the boron nitride agglomerated particles is as follows:
Thermally conductive filler 1: boron nitride agglomerated particles having a house-of-cards structure (B ₁ /(A ₁ +B ₁ )=0.664, average particle size (D50): 35 μm, maximum particle size Dmax: 90 μm)
Thermally conductive filler 2: boron nitride agglomerated particles having a house-of-cards structure (B ₁ /(A ₁ +B ₁ )=0.662, average particle size (D50): 35 μm, maximum particle size Dmax: 90 μm)
Thermally conductive filler 3: boron nitride agglomerated particles having a house-of-cards structure (B ₁ /(A ₁ +B ₁ )=0.648, average particle size (D50): 35 μm, maximum particle size Dmax: 90 μm)
Thermally conductive filler 4: boron nitride agglomerated particles having a house-of-cards structure (B ₁ /(A ₁ +B ₁ )=0.516, average particle size (D50): 34 μm, maximum particle size Dmax: 90 μm)
Thermally conductive filler 5: boron nitride agglomerated particles having a house-of-cards structure (B ₁ /(A ₁ +B ₁ )=0.556, average particle size (D50): 40 μm, maximum particle size Dmax: 90 μm)
Thermally conductive filler 6: boron nitride agglomerated particles having a cabbage structure (the thickness direction of the primary particles near the outermost surface of the agglomerated particles coincides with the radial direction of the agglomerated particles) ("PTX25" (manufactured by Momentive Corp., _B1 /( _A1 + _B1 )=0.441, average particle diameter (D50) 25 μm, specific surface area 7 ^m2 /g)

Thermally conductive fillers 1 and 2 were prepared by the method described below.
For the thermally conductive fillers 1 and 2, the processes up to (raw material), (preparation of slurry), (granulation), and (pyrolysis) described below were carried out as one batch, and only the furnace treatment at 2000° C. in the section (production of boron nitride agglomerated particles) was carried out sequentially as separate batches.

(Raw materials)
The raw materials used were 10,000 g of hexagonal boron nitride (hereinafter referred to as "raw h-BN powder") having a half-width of the (002) plane peak of 2θ = 0.67° and an oxygen concentration of 6.0 mass% as measured by powder X-ray diffraction measurement (Cu-Kα), 11,496 g of a binder (Taki Chemical Industry Co., Ltd.'s "Taxeram M160L", solids concentration 21 mass%), and 250 g of a surfactant (Kao Corporation's surfactant "ammonium lauryl sulfate", solids concentration 14 mass%).

(Preparation of Slurry)
The raw material h-BN powder was weighed out in the above amount into a resin bottle, and then the binder was added in the above amount. Furthermore, the surfactant was added in the above amount, and then zirconia ceramic balls were added, and the mixture was stirred for 1 hour on a pot mill rotating table to obtain a BN slurry. The viscosity of this slurry was 810 mPa·s.

(Granulation)
The BN slurry was spray-dried using a spray dryer (Okawara Kakoki Co., Ltd., FOC-20) at a disk rotation speed of 20,000 to 23,000 rpm and a drying temperature of 80° C. to obtain spherical BN granulated particles.

(Thermal decomposition)
The BN granulated particles were heat-treated in an air atmosphere at 700° C. for 5 hours to obtain precursor particles.

(Preparation of boron nitride agglomerated particles)
The precursor particles were filled in a circular graphite crucible with a lid in the shape of a disk, the atmosphere in the furnace was replaced with nitrogen gas at normal pressure at room temperature, the temperature was raised to 2000° C. at a rate of 83° C./hour while passing nitrogen gas, and after reaching 2000° C., the temperature was maintained for 5 hours while passing nitrogen gas, and then cooled to room temperature. The part that had been in contact with the graphite was removed from the fired material taken out of the crucible, and the material was manually crushed using a mortar and pestle, and then treated with a roll mill to obtain spherical agglomerated particles of boron nitride having a card house structure.
The obtained boron nitride agglomerated particles were sieved using a sieve with 90 μm openings, and only the particles that passed through the sieve were used as the thermally conductive filler.
In addition, after the above-mentioned furnace treatment at 2000°C, thermally conductive fillers 1 and 2 were also subjected to manual crushing and roll mill treatment separately. However, the fact that the 2000°C furnace treatment was performed sequentially as separate batches did not result in any difference in the measurement results by the mercury intrusion method.

Thermally conductive filler 3 was produced in the same manner as thermally conductive filler 1, except that 5,000 g each of hexagonal boron nitride having an oxygen concentration of 5 mass % and hexagonal boron nitride having an oxygen concentration of 7.5 mass % was mixed as the raw materials.
Thermally conductive filler 4 was made by dry-treating thermally conductive filler 2 for 30 minutes in a bead mill using nylon balls with a diameter of 10 mm, resulting in spherical aggregated particles in which the primary particles on the surface of the aggregated particles were bent to cover the aggregated particles. As shown in Fig. 5, from the SEM observation image of the aggregated particles alone, it is determined that the aggregated particles themselves are not broken or damaged, and the internal house-of-cards structure is maintained.
Thermally conductive filler 5 was prepared in the same manner as thermally conductive filler 1, except that 10,000 g of hexagonal boron nitride with an oxygen concentration of 7.5% by mass was used as a raw material, and then dry-treated for 30 minutes in a bead mill using nylon balls with a diameter of 10 mm to obtain spherical agglomerated particles in which the primary particles on the surface of the agglomerated particles were bent to cover the agglomerated particles. From the SEM observation image of the agglomerated particles alone, it was determined that the agglomerated particles themselves of filler 5 were not broken or damaged, and the internal house-of-cards structure was also maintained.

[Preparation of thermally conductive resin sheets of Examples 1 to 3 and Comparative Examples 1 to 3]
The matrix resin powder and the thermally conductive filler powder were mixed at room temperature, and the obtained powder mixture was pressed for 10 minutes in a high-temperature vacuum press device (manufactured by Kitagawa Seiki Co., Ltd.) at a press temperature of 395°C and a press surface pressure of 10 MPa to obtain a thermally conductive resin sheet having a thickness of 150 μm and a size of 15 cm square, and a thermally conductive resin sheet having a thickness of 500 μm and a size of 15 cm square. At this time, the thickness of the sheet was adjusted by adjusting the amount of powder. However, only in Example 2, pressing was performed for 10 minutes at a press temperature of 395°C and a press surface pressure of 20 MPa.
The 150 μm thick thermally conductive resin sheet was used as a specimen for a moisture absorption reflow test described later, and the 500 μm thick thermally conductive resin sheet was used as a specimen for measuring thermal conductivity described later.

Here, the 10 minutes mentioned above means that the inside of the vacuum press was preheated to 150°C, the powder mixture was added there as the press feed composition, a light pressure of several MPa was applied to the powder mixture while operating the vacuum pump, the temperature inside the press was set to 395°C, and after 40 minutes of heating, the press surface pressure was set to 10 MPa for 10 minutes. After the 10 minutes had passed, the temperature inside the press was again set to 150°C, and when the internal temperature approached 150°C, the vacuum was released and the thermally conductive resin sheet was removed.

The press-prepared structure is a structure required for one press in a batch, in which a frame-shaped spacer 6 mm thick, 20 cm long and wide on the outside, and an opening of 15 cm x 15 cm long and wide on the inside is placed on the lower plated plate, the powder mixture is scattered in the spacer in the amount of mass required to obtain a 150 μm-thick pressed sheet or a 500 μm-thick pressed sheet, a 5.85 mm-thick drop lid (when taking a sample of 150 μm thickness) or a 5.50 mm-thick drop lid (when taking a sample of 500 μm thickness) and a 14.6 cm x 14.6 cm long and wide drop lid is fitted into the 15 cm x 15 cm opening, and the upper plated plate is placed on top.

[Preparation of thermally conductive resin sheet of Comparative Example 4]
A thermally conductive filler was added so that the total amount of the epoxy resin composition and the thermally conductive filler was 100% by mass, and then 37.2% by mass of a mixed solution of methyl ethyl ketone and cyclohexanone (mixing ratio (volume ratio) 1:1) was added and mixed so that the total solid content concentration of the epoxy resin composition and the thermally conductive filler was 62.8% by mass, to obtain a coating slurry (coating liquid for sheets). When mixing these, after manual stirring, stirring was performed for 2 minutes using a planetary stirrer "Awatori Rentaro AR-250".
The coating slurry obtained above was applied onto a polyethylene terephthalate film (hereinafter also referred to as "PET film") having a thickness of 38 μm by a doctor blade method, and then heated and dried at 60° C. for 120 minutes, and then pressed at a press temperature of 42° C. and a press surface pressure of 15 MPa for 10 minutes to obtain an uncured epoxy resin sheet-shaped molding. This epoxy resin sheet-shaped molding was used as a specimen for a moisture absorption reflow test described later.

A thermally conductive resin sheet with a thickness of 500 μm used for the thermal conductivity measurement described later was prepared by the following method. The uncured epoxy resin sheet-like molded body was cut into a size of 10 cm x 10 cm, and four of the molded bodies were stacked to prepare a press-prepared structure. Using the same spacers and drop lids as those used to obtain 500 μm-thick sheets in Examples 1 to 3 and Comparative Examples 1 to 3, the press was performed at a press temperature of 175°C and a press surface pressure of 10 MPa for one hour to obtain a thermally conductive resin sheet with a thickness of 500 μm. The excess thickness of the laminated uncured epoxy resin sheet-like molded body was absorbed by the empty volume between the sheet dimensions of 10 cm x 10 cm and the spacer dimensions of 15 cm x 15 cm.

<Measurement conditions and evaluation method>
The thermally conductive fillers 1 to 6 and the thermally conductive resin sheets of Examples 1 to 3 and Comparative Examples 1 to 4 were measured and evaluated by the following methods.

[Physical properties of agglomerated particles (raw powder)]
(Dmax and D50)
The maximum particle diameter Dmax and average particle diameter D50 of the boron nitride agglomerated particles were measured by the following method.
A sample was prepared by ultrasonically dispersing 20 mg of boron nitride agglomerated particles in 10 mL of a pure water medium containing sodium hexametaphosphate, and the particle size distribution of the sample was measured using a laser diffraction/scattering particle size distribution analyzer LA-920 (manufactured by HORIBA, Ltd.). From the particle size distribution thus obtained, the maximum particle diameter Dmax and the average particle diameter D50 of the boron nitride agglomerated particles were determined.

(Intra-particle pore volume _A1 , inter-particle void volume _B1 and _B1 /( _A1 + _B1 ))
The intraparticle pore volume A ₁ (mL/g) and interparticle volume B ₁ (mL/g) of the boron nitride agglomerated particles were measured according to JIS R1655:2003 "Test method for pore size distribution of fine ceramics molded bodies by mercury intrusion porosity method".
Specifically, a Micromeritics Autopore IV mercury porosimeter was used, 200 mg of sample was filled into a measurement cell, and the cell was subjected to a reduced pressure treatment (50 μmHg or less) for 10 minutes, after which the total pore volume was determined, and the mercury intrusion/extrusion curve was measured. A pore size distribution curve was created with the pore size on the horizontal axis and the logarithmic differential pore volume on the vertical axis, and the diameter (division diameter) at which the logarithmic differential pore volume was a minimum value with respect to the pore size was read between peak a representing the intraparticle pores and peak b representing the interparticle pores, assuming that the pores are cylindrical. The interparticle volume B ₁ (mL/g) was calculated from the integral value of the mercury intrusion/extrusion curve in the region where the pore size is larger than the division diameter.
Furthermore, the inter-particle pore volume _B1 was subtracted from the total pore volume to determine the intra-particle pore volume _A1 (mL/g). From the measurement results, " _B1 /( _A1 + _B1 )" was calculated.
When filling the cell of the mercury porosimeter with the sample, no special operations such as tapping are performed. However, since the measurement itself involves forcing mercury, which has a high specific gravity, the measurement results are not dependent on the initial filling condition of the sample and depend only on the properties of the boron nitride agglomerated particles.
Although there may be boron nitride agglomerated particles that are destroyed by the above-mentioned mercury intrusion, such particles do not satisfy the conditions of the present invention.

(Circularity)
The circularity of the thermally conductive filler 2 and the thermally conductive filler 4 was measured using a particle image analyzer (Malvern, Morphologi G3S). For all of the boron nitride agglomerated particles, there may be boron nitride particles that do not form agglomerates and remain as primary particles, or particles that once formed agglomerates but then dropped off from the agglomerates during subsequent handling to become primary boron nitride particles. Therefore, classification was performed using air flow dispersion at 1 Bar, and then image analysis was performed to measure the circularity. The circularity was measured using Morphologi by measuring and calculating the particle perimeter and the perimeter of a circle with an area equal to the particle area, with the former as the denominator and the latter as the numerator. Measurements were performed on 10,000 particles, and the average value was taken as the circularity.

[Physical properties of agglomerated particles (sheet ash)]
(Intra-particle pore volume _A2 , inter-particle void volume _B2 , _A2 / _A1 and _B2 / _B1 )
The intraparticle pore volume A ₂ (mL/g) and interparticle volume B ₂ (mL/g) of the boron nitride agglomerated particles contained in the residual ash when the thermally conductive resin sheet was heated to 700° C. were measured by mercury intrusion porosimetry as follows.
First, a laminated heat dissipation sheet consisting of a heat conductive resin layer of 150 μm thickness, laminated with a copper plate, was prepared for the moisture absorption reflow test described later, and was made of "heat dissipation metal plate material (copper plate, thickness 2 mm) / heat conductive resin sheet (thickness 0.15 mm) / copper plate for forming conductive circuit (thickness 0.5 mm)". The copper plate was removed by etching, and only the resin sheet was isolated. 400 mg of the sample (resin sheet area is about 3.6 cm x 3.6 cm) was taken and heated in a heating furnace in the air at 700 ° C for 5 hours to decompose and remove the resin content, and the heated ash was obtained.
The sheet before heating and the ash content after heating were each weighed, and it was confirmed that the mass of the ash content approximately coincided with the calculated mass of the boron nitride agglomerated particles blended.
Next, a Micromeritics Autopore IV mercury porosimeter was used, and 200 mg of sample was loaded into the measurement cell. The interparticle void volume _B2 (ml/g) and intraparticle pore volume _A2 (ml/g) of the boron nitride agglomerated particles in the sheet ash were determined in the same manner as in the measurements of _A1 and _B1 described above.
The residual ratios A2 _/A1 and _B2/B1 were calculated from the intraparticle pore volume A1 and interparticle space volume B1 of the fillers used in Examples 1 to 3 and Comparative Examples ₁ to ₄ , respectively, and the intraparticle pore volume _A2 and interparticle space _{volume B2 of} the ash of the thermally conductive resin sheet.

(Circularity)
For Example 2 and Comparative Example 1, the circularity of the ash of the thermally conductive resin sheet was measured using a particle image analyzer (Malvern, Morphologi G3S). In addition, since it is possible that the boron nitride particles in each sheet ash are primary particles without forming aggregates, or particles that once formed aggregate particles but fell off from the aggregate particles during subsequent handling or the pressurizing process for sheet molding to become primary boron nitride particles, etc., classification was performed by air flow dispersion at 1 Bar, and then image analysis was performed to measure the circularity. The circularity measurement using Morphologi was performed by measuring and calculating the particle perimeter and the perimeter of a circle with an area equal to the particle area, with the former as the denominator and the latter as the numerator. Measurements were performed for 10,000 particles, and the average value was taken as the circularity.

[Thermal conductivity at 25° C.]
Measurement samples were cut into 10 mm squares from the 500 μm thick thermally conductive resin sheets (specimens) obtained in Examples 1 to 3 and Comparative Examples 1 to 4, and after a thin layer of laser light absorbing spray (Fine Chemical Japan's Blackguard Spray FC-153) was applied to both sides and dried, the thermal diffusivity a ( ^mm2 /sec) in the thickness direction of the resin sheet at a measurement temperature of 25° C. was measured by the laser flash method using a xenon flash analyzer (NETZSCH's LFA447 NanoFlash300). The measurement was performed on five points cut out from the same sheet, and the arithmetic average value was calculated.

Next, the density ρ (g/m ³ ) of the resin sheet was determined by the Archimedes method using a specific gravity measuring device (manufactured by A&D Co., Ltd.) in accordance with JIS K 6268. In addition, the specific heat capacity c (J/(g·K)) at 25° C. was measured using a DSC measuring device (ThermoPlusEvo DSC8230, manufactured by Rigaku Corporation) in accordance with JIS K 7123.
From these measured values, the thermal conductivity in the sheet thickness direction at 25° C. was calculated as "H=a×ρ×c".

In addition, since there is no JIS standard regarding the thermal diffusivity and thermal conductivity of resin-based materials, the thermal diffusivity a ( ^mm2 /sec) was determined based on JIS R1611:2010 (Method for measuring thermal diffusivity, specific heat capacity, and thermal conductivity of fine ceramics by the flash method). Since the standard stipulates that "the thickness of the sample shall be 0.5 mm or more and 5 mm or less," only the thickness of the sample used for thermal conductivity measurement was adjusted to 0.5 mm and then measured.

[Breakdown voltage before moisture absorption and reflow test (BDV)]
(Preparation of heat dissipation circuit board)
The thermally conductive resin sheets having a thickness of 150 μm produced in Examples 1 to 3 and Comparative Examples 1 to 4 were cut into a size of 40 mm×80 mm to prepare thermally conductive resin sheets for circuit boards.
On the other hand, a copper plate having a thickness of 2000 μm and a size of 40 mm x 80 mm was prepared as a metal plate material for heat dissipation, and a copper plate having a thickness of 500 μm and a size of 40 mm x 80 mm was prepared as a copper plate for forming a conductive circuit, one of each being prepared for each thermally conductive resin sheet for the circuit board.
One side of a 2000 μm thick copper plate and one side of a 500 μm thick copper plate were each polished in advance with #100 sandpaper to roughen the surface, and the thermally conductive resin sheet for circuit boards was sandwiched between the copper plates of different thicknesses so that the roughened surface faced the thermally conductive resin sheet for circuit boards. The copper plates were pressed at a press temperature of 390° C. and a press surface pressure of 13 MPa for 10 minutes to obtain a laminated heat dissipation sheet consisting of "heat dissipating metal plate material (copper plate)/thermal conductive resin sheet/copper plate for forming conductive circuits."

On the other hand, for Comparative Example 4, which is a thermally conductive resin sheet made of a thermosetting resin, a 150 μm thick epoxy resin sheet molded body that was uncured or in which the curing reaction had only progressed very slightly was sandwiched between the above-mentioned roughened metal plate material for heat dissipation (copper plate) and a copper plate for forming a conductive circuit, and vacuum pressed for 30 minutes at a press temperature of 175°C and a press surface pressure of 10 MPa to complete the curing reaction of the thermosetting resin, thereby obtaining a laminated heat dissipation sheet consisting of "metal plate material for heat dissipation (copper plate)/thermally conductive resin sheet/copper plate for forming a conductive circuit."

The copper plate for forming the conductive circuit of each laminated heat dissipation sheet was then etched and patterned to obtain a heat dissipation circuit board. The pattern was such that two copper plates for the conductive circuit with a circular pattern of φ25 mm remained on a 40 mm x 80 mm thermally conductive resin sheet.

(Measurement of dielectric breakdown voltage (BDV))
The heat dissipating circuit boards prepared using the thermally conductive resin sheets obtained in Examples 1 to 3 and Comparative Examples 1 to 4 by the above method were immersed in Fluorinert FC-40 (manufactured by 3M), and a φ25 mm electrode was placed on a φ25 mm copper plate patterned by etching on the heat dissipating circuit board using an ultra-high voltage withstand voltage tester 7470 (manufactured by Keisoku Gijutsu Kenkyusho Co., Ltd.), a voltage of 0.5 kV was applied, and the voltage was increased by 0.5 kV every 60 seconds until dielectric breakdown occurred. The measurement was performed at a frequency of 60 Hz and a voltage increase rate of 1000 V/sec.

When the breakdown voltage per unit thickness (converted value for a thickness of 1 mm) was 60 kV/mm or more, it was indicated as "◯ (good)", when it was 40 kV/mm or more and less than 60 kV/mm, it was indicated as "△ (not good)", and when it was less than 40 kV/mm, it was indicated as "× (poor)". The measurement results are shown in Table 1.
The evaluation of the dielectric breakdown voltage (BDV) before the moisture absorption reflow test can be used as an evaluation of the voltage resistance performance.

[Breakdown voltage after moisture absorption reflow test (BDV)]
(Measurement of dielectric breakdown voltage (BDV))
Using the thermally conductive resin sheets obtained in Examples 1 to 3 and Comparative Examples 1 to 4, a heat dissipating circuit board was prepared in the same manner as in the above [Breakdown voltage (BDV) before moisture absorption reflow test], and stored in an environment of 85 ° C. and 85% RH for 3 days using a thermohygrostat SH-221 (manufactured by Espec Corporation). After that, the temperature was raised from room temperature to 290 ° C. in 12 minutes in a nitrogen atmosphere within 30 minutes, and the temperature was kept at 290 ° C. for 10 minutes, and then cooled to room temperature (moisture absorption reflow test). Thereafter, the heat dissipating circuit board was immersed in Fluorinert FC-40 (manufactured by 3M Corporation), and a φ25 mm electrode was placed on a φ25 mm copper plate patterned by etching on the heat dissipating circuit board using an ultra-high voltage withstand voltage tester 7470 (manufactured by Keisoku Gijutsu Kenkyusho Co., Ltd.), and a voltage of 0.5 kV was applied, and the voltage was increased by 0.5 kV every 60 seconds, and measurements were performed until the insulation breakdown occurred. The measurement was carried out at a frequency of 60 Hz and a voltage rise rate of 1000 V/sec.

When the breakdown voltage per unit thickness (converted value for a thickness of 1 mm) was 60 kV/mm or more, it was indicated as "◯ (good)", when it was 40 kV/mm or more but less than 60 kV/mm, it was indicated as "△ (not good)", and when it was less than 40 kV/mm, it was indicated as "× (poor)". The measurement results are shown in Table 1.
The evaluation of the dielectric breakdown voltage (BDV) after the moisture absorption reflow test can be used as an evaluation of the moisture absorption reflow resistance.

[Observation of sample condition after moisture absorption reflow test]
The heat dissipating circuit boards prepared in Examples 1 to 3 and Comparative Examples 1 to 4 were subjected to a moisture absorption reflow test in the same manner as described above, and then the interface between the copper electrode with a diameter of 25 mm patterned by the above etching and the thermally conductive resin sheet was observed using an ultrasonic imaging device FinSAT (FS300III) (manufactured by Hitachi Power Solutions). The measurement was performed using a probe with a frequency of 50 MHz, a gain of 30 dB, a pitch of 0.2 mm, and by placing the sample in water. Those that did not show peeling, lifting, or voids at the interface were marked as "○ (good)", and those that showed peeling, lifting, or voids at the interface were marked as "× (poor)". The evaluation results are also shown in Table 1.
The evaluation of interfacial peeling after the moisture absorption reflow test can be made as an evaluation of whether or not interfacial peeling due to thermal expansion and thermal contraction and deformation due to foaming of the thermal conductive resin sheet are likely to occur when the thermal conductive resin sheet is laminated with a metal plate and subjected to a reflow process.

Comparing Examples 1 to 3 with Comparative Examples 1 to 3, it was found that by using boron nitride agglomerated particles in which B ₁ /(A ₁ +B ₁ ) is 0.60 or more, a heat dissipation sheet with significantly high thermal conductivity can be obtained.
In addition, in Examples 1 to 3, _A2 / _A1 before and after sheet forming was 0.70 or more, and _B2 / _B1 was 0.85 or less, so it is believed that the boron nitride agglomerated particles used in these examples maintain the internal structure of the agglomerated particles themselves even after sheet forming, thereby maintaining high thermal conductivity inside the agglomerated particles, and at the same time, the surfaces of the contact parts between adjacent particles are significantly deformed, thereby suppressing the thermal resistance between particles to a low level. From the SEM photograph of the boron nitride agglomerated particles contained in the ash of the thermal conductive resin sheet of Example 2 shown in Figure 9, it can be confirmed that while the internal structure of the agglomerated particles is maintained, the surfaces of the contact parts between adjacent particles are significantly deformed.
In addition, when comparing the values of the breakdown voltage before and after the moisture absorption reflow test between Examples 1 to 3 and Comparative Example 4, it was found that the use of a thermoplastic resin having a melting point of 300°C or higher as the matrix resin improved the voltage resistance performance and moisture absorption reflow resistance.
Furthermore, in Examples 1 to 3, the manufacturing method includes a mixing step of mixing a powder made of thermoplastic resin with boron nitride agglomerated particles, and a press molding step of pressing the mixture to form a sheet. This makes it possible to achieve both high voltage resistance and high thermal conductivity without impairing the inherent properties of the boron nitride agglomerated particles, which have high thermal conductivity, and also suppresses the formation of voids in the sheet during molding.
In Comparative Example 4, an epoxy resin and boron nitride agglomerated particles having a _B1 /( _A1 + _B1 ) of 0.60 or more were used to form a sheet by a conventional wet coating method. It is estimated that the interparticle volume of the boron nitride agglomerated particles is too large for use in a wet coating method, and streaks were generated when the film was coated and formed, and the sheet also contained many air bubbles, which is thought to have reduced the voltage resistance performance. Due to the influence of the air bubbles in the sheet, the thermal conductivity was also inferior to that of Example 1, which contained the same amount of the same boron nitride agglomerated particles.

Claims (17)

A resin composition comprising a thermoplastic resin and boron nitride agglomerated particles,
the resin having the highest content by mass ratio among the resins contained in the resin composition is a crystalline thermoplastic resin having a melting point of 300° C. or more;
The resin composition contains 15% by mass or more and 40% by mass or less of the thermoplastic resin and 60% by mass or more and 85% by mass or less of the boron nitride agglomerated particles, based on 100% by mass of the resin composition;
a pore volume within the boron nitride agglomerated particles measured by mercury intrusion porosimetry of A1 and a volume of interparticle spaces of the boron nitride agglomerated particles measured by mercury intrusion porosimetry of B1, the ratio B1/(A1+B1) being 0.60 or more. The resin composition according to claim 1, wherein the interparticle volume B1 is 0.85 mL/g or more. The resin composition according to claim 1 or 2, wherein the crystalline thermoplastic resin having a melting point of 300°C or higher is a polyether ketone resin. The resin composition according to claim 3, wherein the polyether ketone resin is polyether ether ketone. The resin composition according to any one of claims 1 to 4, wherein the boron nitride agglomerated particles have a house of cards structure. The resin composition according to any one of claims 1 to 5, wherein the volume-based average particle diameter D50 of the boron nitride agglomerated particles is 10 μm or more and 200 μm or less. A thermally conductive resin sheet comprising the resin composition according to any one of claims 1 to 6. 8. The thermally conductive resin sheet according to claim 7, wherein, for boron nitride agglomerated particles contained in residual ash when the thermally conductive resin sheet is heated to 700°C, the intraparticle pore volume measured by mercury intrusion porosimetry is _A2 and the interparticle volume is _B2 , _A2 / _A1 is 0.70 or more and _B2 / _B1 is 0.85 or less. The thermally conductive resin sheet according to claim 7 or 8, wherein the circularity of the boron nitride agglomerated particles contained in the residual ash when the thermally conductive resin sheet is heated to 700°C is 0.85 or more. The thermally conductive resin sheet according to any one of claims 7 to 9, having a thickness of 50 μm or more and 300 μm or less. The thermally conductive resin sheet according to any one of claims 7 to 10, having a thermal conductivity in the thickness direction at 25°C of 16 W/m·K or more. A laminated heat dissipation sheet having a configuration in which a heat dissipation metal layer is laminated on one surface of the thermally conductive resin sheet according to any one of claims 7 to 11. A heat dissipating circuit board having the laminated heat dissipation sheet according to claim 12. The heat dissipation circuit board according to claim 13, comprising a conductive circuit formed on the other surface of the thermally conductive resin sheet. A power semiconductor device having a heat dissipation circuit board according to claim 13 or 14. A mixing step of obtaining a mixture of powder made of a thermoplastic resin and agglomerated particles of boron nitride to prepare a resin composition;
A press molding step of pressing the resin composition into a sheet,
the resin having the highest content by mass ratio among the resins contained in the resin composition is a crystalline thermoplastic resin having a melting point of 300° C. or more;
The resin composition contains 15% by mass or more and 40% by mass or less of the thermoplastic resin and 60% by mass or more and 85% by mass or less of the boron nitride agglomerated particles, based on 100% by mass of the resin composition;
a pore volume within a particle of the boron nitride agglomerated particle raw material measured by mercury intrusion porosimetry is defined as A1, and an interparticle volume therebetween is defined as B1, where B1/(A1+B1) is 0.60 or more. The method for producing a thermally conductive resin sheet according to claim 16, wherein in the mixing step, the powder made of the thermoplastic resin and the boron nitride agglomerated particles are stirred and mixed at room temperature.

Images