WO1995011798A1

WO1995011798A1 - Thermally-conductive di-electric composite materials, and methods of forming same

Info

Publication number: WO1995011798A1
Application number: PCT/US1994/012319
Authority: WO
Inventors: Eric A. Groat; Thomas J. Miroz
Original assignee: Advanced Refractory Technologies, Inc.
Priority date: 1993-10-29
Filing date: 1994-10-28
Publication date: 1995-05-04
Also published as: AU8126894A

Abstract

A composite material, occupying a volume, includes a thermally-conductive material (e.g. AIN, BeO, etc.) having a plurality of solid particles (e.g., granules, whiskers, etc.). The particles are bonded together to form at least one thermally-conductive path from one point in the volume to another point in the volume. A thermally-insulative material (e.g., a glass, a polymer, etc.) occupies at least portion of the volume not occupied by the thermally-conductive material. The composite material is formed by a method, which includes the steps of: creating a preform of a thermally-conductive material having a plurality of solid particles bonded together to form at least one thermally-conductive path between two points in the preform, and infiltrating the preform with a thermally-insulative material.

Description

THERMALLY-CONDUCTIVE DI¬ ELECTRIC COMPOSITE MATERIALS, AND METHODS OF FORMING SAME

Technical Field The present invention relates generally to composite materials, and, more particularly, to improved dielectric composite materials having aluminum nitride particles sintered to form a network of thermally-conductive paths, with an electrically- and ther¬ mally-insulative filler (e.g., a glass, a polymer, etc.) occupying at least a portion of the volume of the aluminum nitride preform.

Background Art

There is a clear need for materials which exhibit low dielectric properties while being thermally conductive. For example, these materials can be used in many types of electronic packages to convey heat away from the various parts and components, while acting as an electrical insulator. Aluminum nitride (AIN) and beryllium oxide (BeO) are two materials that are known to have high thermal conductivities and moder¬ ate dielectric constants. Thus, these two materials would appear to be prime candidates for use in such composites. However, AIN is comparatively expensive, and BeO is highly toxic. Also, further reduction in dielectric constant is of interest for advanced electronic applications. Accordingly, because of these factors, the use of such materials has been somewhat limited. Composite systems that utilize the thermal conductivity of AIN or

BeO, and the low dielectric constant of some polymers and glasses, are also of interest.

An example of a glass composite is illustrated in U.S. Patent No. 5,214,005, which appears to disclose a glass-AIN composite material in which a body is formed of

AIN grains and glass powder. This blend or mixture is then molded and sintered at a temperature of about 900 °C. While this sintering temperature is less than the bonding temperature of AIN, the resulting article is reported to have a heat conductivity of 30 W/m^'K or more, and a reduction of dielectric constant by about 45% relative to dense monolithic AIN.

U.S. Patent No. 5,102,749 appears to disclose a composite material formed of AIN and borosilicate glass. Another specific example is shown in U.S. Patent No. 5,017,434. See also, published European Patent Application No. 92 101 704.2, filed February 3, 1992. This European Patent Application claims Convention priority of Japa¬ nese Patent Application No. 3-033345, filed February 4, 1991, and is believed to corre¬ spond to U.S. Patent No. 5,214,005. The need for composite materials having a high thermal conductivity and a low dielectric constant, has lead to much research in the area of filled-polymer and filled- glass systems. These composite systems typically use a chemically-inert solid filler to modify the thermal and dielectric properties of a polymer or glass, without degrading the other structural and electronic properties of interest. Metal particulate and short fibers

(Le., whiskers) have been used successfully to produce thermally- and electrically-conduc¬ tive polymer composites. See, e.g., D. Bigg, "Mechanical, Thermal, and Electrical Proper¬ ties of Metal Fiber-Filled Polymer Composites", Poly. Eng. & ScL, 19 [12] 1188-1192 (1979); and A. Holbrook, "Aluminum Flake Filled Conductive Plastics for EMI Shielding and Thermal Conductivity", Int. J. Pwdr. Metal., 22 [1] 39-45 (1986). However, many of the potential applications require high electrical resistivity, which prohibits the use of metal-filled systems.

Ceramic particles have been used with a degree of success where high ther¬ mal conductivity and high electrical resistivity are required. Because the thermal trans- port mechanism in these materials is believed to be phonon-based, rather than electron- based, the thermal conductivity is regarded as being independent of the electrical con¬ ductivity. Composite materials having highly-conductive fillers, such as alumina, boron nitride (BN), AIN and diamond, have been examined for possible use in this application. Despite the higher thermal conductivities of AIN and diamond over alumina and BN, the thermal conductivity of samples described in the prior art appears to be relatively insen¬ sitive to filler type. In fact, in some cases, BN-filled systems displayed higher thermal conductivities at lower solids contents than AIN-filled systems. Thus, it is expected that some other variable has a much larger effect on the thermal conductivity of the compos¬ ite, than the particular conductivity of the filler. See, e.g., P. Bujard et al, "Thermally Conductive Aluminum Nitride-Filled Epoxy Resin", Fifth Annual IEEE SEMI-THERM™ Symposium, IEEE, Piscataway, N.J., 126-130 (1989); P. Bujard, "Thermal Conductivity of Boron Nitride Filled Epoxy Resins: Temperature Dependence and Influence of Sample Preparation", Intersoc. Conf. Therm. Phenom. Fabr. Ope Electron. Compon., IEEE, New York, N.Y., 41049 (1988); N. Tsutsumi et al, "Measurement of Thermal Diffusivity of Filler-Polyimide Composites by Flash Radiometry", . Appl. Poly. ScL, 30 [5] 2226-2235

(1985); and Y. Agari et al, "Estimation on Thermal Conductivities of Filled Polymers", /. App. Poly. ScL, 32 [7] 5705-5712 (1986).

Upon information and belief, there is a continued need to form a composite material of a thermally-conductive material (e.g., AIN, BeO, or the like), and a thermally- insulative filler. Additionally, improved materials exhibiting both high thermal conduc- tivity and low dielectric constant, are of interest. Moreover, there is also believed to be a continued and further need for an improved method of forming such composite mate¬ rials in order that the intended thermal conductivity properties may be more predictably obtained. There is also believed to be a need for an improved composite material which can afford the desired thermal conductivity while minimizing the quantity of thermally- conductive material required in such a composite.

Disclosure of the Invention The present invention provides an improved composite material which occu¬ pies a volume. The improved material includes a thermally-conductive, yet electrically- resistive, component material (e.g., AIN, BeO, etc.) having a plurality of solid particles (Le., granules, whiskers, grains, platelets, etc.), the individual particles being bonded together to form at least one thermally-conductive path from one point in the volume to another point in the volume; and a thermally-insulative dielectric material (e.g., glass, polymer, etc.) occupying at least a portion of the volume not occupied by the thermally- conductive material. The thermally-conductive material may form a network of thermal¬ ly-conductive paths, and the filler material may form a complementary interpenetrating network.

In another aspect, the invention provides the method of creating such a composite material. The improved method broadly includes the steps of: creating a preform of thermally-conductive material having a plurality of solid particles bonded together to form at least one thermally-conductive path from one point to another point; and infiltrating the preform with a thermally-insulative material. The steps of creating the preform may include the steps of: mixing particles (e.g., a powder) of the thermally- conductive material with an organic binder; forming (Le., shaping under the influence of pressure) a preform; and heating the preform to remove the binder therefrom and to create particle-to-particle bonds between the particles of thermally-conductive material. Thus,- the sintered thermally-conductive material will create a matrix-like network of thermally-conductive paths through the volume. The method may include the further step of allowing the thermally-insulative material to solidify so as to prevent its unintend- ed separation from the thermally-conductive material and/or exposing the thermally- conductive material at separated portions on the surface of the infiltrated preform to create a preferential thermally-conductive path therebetween.

In still another aspect, the invention provides an improved composite mate¬ rial which is made according to the foregoing method. Accordingly, the general object of the invention is to provide an improved composite material having a low dielectric constant and a high thermal conductivity.

Another object is to provide an AIN-glass composite material.

Another object is to provide an AIN-polymer composite material.

Another object is to provide an improved AlN-filler composite material which maximizes thermal conductivity while minimizing AIN content.

Still another object is to provide an improved method of manufacturing a composite material having AIN and a filler (e.g., a glass, a polymer, etc.), in which the actual thermal conductivity of the composite will more predictably meet its intended goal. These and other objects and advantages will become apparent from the foregoing and ongoing written specification, the drawings, and the appended claims.

Brief Description of the Drawings Fig. 1 is a flow chart showing the various operative steps used in forming the improved AIN-polymer composite material. Fig. 2 is a microphotograph, taken at a magnification of 1000 times, of a

"mixed" AIN-polymer composite material having an AIN loading of about 55 volume per¬ cent.

Fig. 3 is a microphotograph, taken at a magnification of 5000 times, of the "mixed" composite material shown in Fig. 2. Fig. 4 is a microphotograph, taken at a magnification of 1000 times, of a sintered AIN preform having an AIN loading of about 77 volume percent.

Fig. 5 is a microphotograph, taken at a magnification of 5000 times, of the sintered AIN preform shown in Fig. 4.

Fig. 6 is a microphotograph, taken at a magnification of 1000 times, of the sintered preform shown in Figs. 4 and 5, after it has been back-filled and infiltrated with a thermally-insulative polymer.

Fig. 7 is a microphotograph, taken at a magnification of 5000 times, of the back-filled sintered preform shown in Fig. 6.

Fig. 8 is a plot of the thermal conductivity (ordinate) of various "mixed" composite samples as a function of AIN volume loading (abscissa).

Fig. 9 is a plot of the thermal conductivity (ordinate) of various "infiltrated" composite samples as a function of AIN volume loading (abscissa).

Description of the Preferred Embodimentfs) At the outset, it should be clearly understood that like reference numerals are intended to identify the same structural elements, portions or surfaces consistently throughout the several drawing figures, as such elements, portions or surfaces may be further described or explained by the entire written specification, of which this detailed description is an integral part. Unless otherwise indicated, the drawings are intended to be read (e.g., cross-hatching, arrangement of parts, proportion, degree, etc.) together with the specification, and are to be considered a portion of the entire written descrip¬ tion of this invention. As used in the following description, the terms "horizontal", "verti¬ cal", "left", "right", "up" and "down", as well as adjectival and adverbial derivatives thereof (e.g., "horizontally", "rightwardly", "upwardly", etc.), simply refer to the orientation of the illustrated structure as the particular drawing figure faces the reader. Similarly, the terms "inwardly" and "outwardly" generally refer to the orientation of a surface relative to its axis of elongation, or axis of rotation, as appropriate.

Theory

Most filled-polymer systems are produced by a simple mixing process. In this process, a powdered filler and the polymer are mixed together, cast, and subsequent¬ ly cured to form the desired article. This mixing process results in particles that appear to be well-dispersed in the polymer matrix, but which are believed to be separated from one another by a polymer film.

Models of the composite structure effect on properties have been construct- ed for two-dimensional macroscopic systems. These models appear to describe some microscopic systems in either two or three dimensions. Because the standard "mixed" filled-polymer system does not have interpenetrating networks (Le., only the filler phase is continuous), the thermal conductivity of the resulting "mixed" composite material may be initially described by the perpendicular plate model:

[1] J_ = k. ₊

-K. A. K.p where K, represents the thermal conductivity of the composite, K, and K_p are the thermal conductivities of the filler and polymer, respectively, and f_β and f are the volume frac¬ tions of the filler and the polymer, respectively. The thermal conductivity of the result¬ ing composite does not increase linearly with increased filler content, and appears to be dominated by the thermal conductivity of the least-conductive component (Le., the poly- mer).

The perpendicular plate model requires adjustment to accurately predict the thermal conductivity of actual filled-polymer composite materials. Several variables, mostly related to three-dimensional influences on particle packing and shape, must be considered in the model to better represent the real system. There are a number of proposed models in the prior art. However, Nielsen's model is believed to best repre¬ sent actual experimental data. T. Lewis and L. Nielsen, "Dynamic Mechanical Properties of Particulate-Filled Polymers", J. App. Polym. ScL, 14:1449-1471 (1970). Bujard has reportedly used Nielsen's model to fit the thermal conductivity results obtained from AIN- and BN-filled polymers. However, in his experiments, Bujard used a non-spherical shape factor to fit the AIN data, despite the use of an approximately spherical powders. Additionally, Bujard reported that the thermal conductivity of the composite material with a given filler concentration is reduced by additional mixing of this system. He at¬ tempted to explain these results in terms of conductive "clusters" of filler particles. These "clusters" are reportedly much more efficient in conducting heat as no polymer interface exits between the individuals particles within the cluster. If mixing efficiency is increased, the clusters break down into individual particles, reducing the thermal con- ductivity of the resulting composite.

Should the two phases or components become interconnected and inter¬ penetrating, the thermal behavior of the system could be best described by the parallel plate model:

In this case, each component affects the thermal conductivity of the composite material as a proportional function of its volume concentration. Thus, the response is substantial¬ ly linear over the entire range of filler concentration.

When mixing thermally-conductive fillers into a fluid polymer, it is expected that the filler particles will be wet by the polymer, and will therefore not able to come into direct contact with other filler particles because of the presence of an intermediate polymer film. This is believed to be particularly true in cases where the raw composite material is mixed using high shearing forces to a constant viscosity. In this case, it is believed likely that each filler particle will be completely wetted by the polymer, and, therefore, will be isolated from all other filler particles. Because the thermal conductivi¬ ty of the polymer is much lower than that of the filler, it is believed to act as a thermal barrier between the filler particles, and appears to dominate the properties of the com¬ posite material. In effect, the individual filler particles may be viewed as islands in a sea of polymer.

The structure-based theories presented for thermal conductivities of com¬ posite mixtures is generally applicable to any system where a thermal-conductor phase and a thermal-insulator phases are both present. Therefore, it can be assumed that the thermal conductivity outcome will not change significantly as different thermal insulators are used for the filler component, as long as the structure of the thermally-conductive phase remains the same. In the present invention, the process does not involve a chemi- cal reaction between the composite phases. Therefore no change will occur in either the thermally-conductive or the thermally-insulative phases.

Dielectric properties of composites are not bound by structural consider¬ ations, but rather are best described by a simple rule of mixtures law. Thus, the mixture of two materials with two distinct dielectric constants results in a composite that exhibits a dielectric constant that is equivalent to the sum of the products of the dielectric con¬ stants of the individual phases with their observed volume fractions. U.S. Patent No. 5,214,005 demonstrates this effect for a particular AlN-glass mixed system. However, the effect is believed to be universal to all forms of materials. Thus, any mixture of AIN and a low dielectric constant polymer or glass, regardless of the microstructure, is believed to result in a composite with a dielectric constant lower than that of monolithic AIN.

Experimental Protocol

It is believed that composite materials having much higher thermal conduc¬ tivities can be achieved when particle-to-particle connections of thermally-conductive filler are maximized. The improved method of obtaining this structure is to first prepare a preform of AIN in which each AIN grain is strongly bonded to its adjacent grains. The porosity between the grains can then be back-filled with the polymer. Such a structure allows particle-to-particle grain contact to remain, and is believed to result in the reten¬ tion of an unimpeded thermal flow path between different points in the composite mate¬ rial. This structure then resembles the parallel plate model, characterized by a linear response of the intended characteristic with increased filler content. Presumably, this type of composite would exhibit much higher thermal conductivity than standard "mixed" systems, due to the intimate touching contact between filler particles. However, because less thermally-conductive material is utilized, the cost is minimized, and the dielectric constant of the composite material can be decreased. In the following examples, commercially-available AIN powders were used in all composite samples. The particular AIN powders used were AINel A-100 and AINel A-100SD, both available from Advanced Refractory Technologies, Inc. of Buffalo, New York. A low-viscosity epoxy system was selected for the polymeric matrix, as it allowed back-filling infiltration of even micro-diameter pour channels. This epoxy is identified as "Low Viscosity Spurr Kit", available from SPI Supplies, a division of Structure Probe, Westchester, Pa.

Test samples were then prepared using two general methods. The first method, termed the "mixed" composite system, entailed mixing the AIN powder directly into the fluid polymer. Nominal AIN concentrations ranging from 25 to 55 volume per- cent were prepared. Samples above 55 volume percent were found to be too viscous to be formed. The mixture was stirred until the powder was completely wetted by the poly¬ mer, and the appearance was uniform. Samples were placed in a vacuum chamber for about two hours in order to degas the polymer. Each composition was then thermally- cured into disk-shaped samples approximately 7 cm in diameter and about 1.5 cm high. The second method, termed the "infiltrated" preform composite system, utilized a partially-sintered AIN preform, that was subsequently back-filled with the ep¬ oxy. Referring now to Fig. 1, the preforms were prepared by first mixing AIN powder with a resinous binder, as indicated in box 10. The mixture was initially shaped, as indi¬ cated in box 11, and was subsequently pressed to the desired shape and density, as indi- cated in box 12. A spray-dried grade of AIN powder was used for this experiment, as it resulted in better pressing uniformity and greater green strength in the pressed part. In order to obtain samples with varying porosities, parts were pressed at pressures ranging from about 35 Mpa to about 210 Mpa. The preforms were then sintered in nitrogen or argon at about 1850 °C for about three hours, as indicated in box 13. This yielded a strong, yet porous, assemblage of bonded particles. After sintering, the preforms were machined (as necessary) to shape, as indicated in box 14, and were placed into contain¬ ers, as indicated in box 15. Epoxy was added to cover the part, and to infiltrate the preform, as indicated in box 16. The containers were placed in a vacuum chamber for several hours. The containers were then allowed to cycle back to atmospheric pressure, and were then heated to cure the epoxy, as indicated in box 17. The vacuum-to-atmo¬ spheric pressure cycle was found to be sufficient to completely infiltrate the preforms with epoxy. Attempts to infiltrate as-pressed unsintered samples with epoxy were not successful. Parts prepared from spray-dried powders could not be completely infiltrated, possibly due to interaction with the pressing binders. Parts pressed from raw powders disintegrated during attempted infiltration, resulting in a non-connected AIN network within the epoxy.

Once the samples were cured, they were removed from their containers, as indicated in box 18, and were cut to size using a diamond blade, and machined to expose separated portions on their outer surfaces, as indicated in box 19. In the case of the "mixed" system samples, each test sample was cut from the center of the disk. For the "infiltrated" samples, each sample was cut entirely from within the boundary of the AIN preform. All samples were then machined to the final test dimensions of about 2.5 cm x 2.5 cm x 0.8 cm. Sample densities were obtained for each composite by geometric measurements on the machined test sample. A ground sample of neat epoxy was also prepared for comparison testing. Actual volume loadings of AIN were calculated from these density results for every sample.

Portions of select composite samples were fractured and mounted for SEM analysis. SEM microphotographs were obtained during a combination signal from sec¬ ondary and back scatter electrons. The partial use of back scatter electrons provides some contrast between the AIN and polymer materials.

Machined samples were tested for thermal conductivity by a comparative technique using a Quickline™ - 10 thermal conductivity analyzer, available from Anter Laboratories, Inc., Pittsburgh, Pennsylvania 15235. Reference samples were selected with thermal conductivities close to the values obtained for the test samples.

Comparative Results/Discussion:

Figs. 2 and 3 show the microstructure of an AIN-polymer composite pre¬ pared by the "mixed" system method. The volume was prepared at an AIN loading of about 55 volume percent. Fig. 2 shows that the composite is well mixed, and that each AIN particle is separated from neighboring particles by the polymer matrix. Fig. 3 shows that the individual AIN particles appear to be discrete, and separated by the polymer, with a consistency somewhat resembling peanut brittle.

Table 1 shows the composite density and actual AIN loading for various experimental sample. The actual AIN loading compared well with the intended solids content.

Table 1: Thermal Conductivity of "Mixed" Composite System

Sample Type Composite Density Actual Volume % Thermal Conduc¬ tivity

100% Polymer 1.10 gm/cc 0.0% AIN 0.22 W/mK

25 vol.% AIN 1.68 gm/cc 27.0% AIN 0.60 W/Mk

35 vol.% AIN 1.87 gm/cc 36.0% AIN 0.86 W/mK

45 vol.% AIN 2.05 gm/cc 44.0% AIN 1.20 W/mK

55 vol.% AIN 2.24 gm/cc 53.0% AIN 1.39 W/mK

Thermal conductivity of each sample is also given in Table 1. The thermal conductivity of the neat polymer was measured at 0.22 W/mK. As expected, the thermal conductivity increases with addition of AIN, to a maximum value of about 1.39 W/mK, corresponding to a 55 volume percent AIN concentration. Thermal conductivity as a function of AIN loading is illustrated in Fig. 8. The increase in thermal conductivity with concentration is shown to be monotonic, with no discontinuities. The data has been fit to Nielsen's model using a packing fraction of 0.61, as determined by the pressed density of the pow¬ der pressed at 10 Mpa, a particle thermal conductivity of 6 W/mK, and a particle geo¬ metric factor (A) of 1. The particle thermal conductivity was estimated by first calcu¬ lating the effect of 0.4 percent internal oxygen within the particle using the relationship previously developed by Slack. See, e.g., G. Slack, "Nonmetallic Crystals With High Thermal Conductivity"' /. Phys. Chem. Solids, 34:321-335 (1973). Then, the remaining oxygen was assumed to be present at each particle surface in the form of a hydrated aluminum oxide with an estimated thermal conductivity of 1 W/mK. The effect was calculated using the perpendicular plate model. The value of 6 W/mK calculated by this method, compares well with the value obtained in Bujard's work. The results were compared to Bujard's earlier work in this system. He found that a particle geometry factor of 8 was required for data fit (data reproduced in Fig. 9). Bujard suggested that this indicated that AIN particles were poorly mixed with the polymer, and thus formed conductive clusters. With the current data, the low geo¬ metric factor suggests little or no cluster formation. Thus, the system was well mixed. Bujard also reported that thermal conductivity decreased with increased mixing efficien¬ cy, due to the thermal conductivity effect of clusters as predicted by the Nielsen model). This effect explains the difference in the thermal conductivities obtained with the current samples, as compared to Bujard's data.

Figs. 4 and 5 show the microstructure of a sintered AIN preform having an AIN loading of about 77%. Note the well-developed bonded regions between the parti¬ cles, as well as the size and distribution or porosity. Transmission electron microscopy has shown that the bonded regions contain no secondary phase. Thus, the flow of ther¬ mal energy between particles is unimpeded by the particle-particle interfaces. Because sintering aids were intentionally left out of the powder, the individual particles necked together and grew via vapor phase transport, as shown in Fig. 5. However, no densifica- tion occurred. The pore structure also appears to be interconnected, and is generally in the same size range as the grain size. The resulting part exhibited substantial strength, but contained on the order of 23% porosity, which was interconnecting. The porosity concentration of the samples in general was found to correlate with the original pressing pressure used to fabricate the sample.

Figs. 6 and 7 show the microstructure of the AIN preform shown in Figs. 4 and 5 after the preform has been back-filled with polymer. The polymer completely infiltrates the pores throughout the part. Inter-granular neck regions have not been affected by the infiltration. The individual grains remain interconnected.

Table 2 shows the composite density and actual AIN volume percent calcu¬ lations for the "infiltrated" preform composites. AIN volume percentages ranged from 58 percent to 75 percent. Thermal conductivities ranged from 22.2 W/mK to 45.1 W/mK. The composite results appear to be insensitive to the sintering condition. The thermal conductivities of these samples are more than an order of magnitude higher than the values previously obtained for the "mixed" polymer system. The continuous nature of the AIN matrix is believed to allow the freer and less-impeded transfer of thermal energy through the composite structure.

Table 2: Thermal Conductivity of "Infiltrated" Composite Systems

Sinter Condition Composite Density Actual Vol.% Thermal Conduc¬ tivity

1 (N₂) 2.36 gm/cc 58.0% AIN 28.6 W/mK

1 (N₂) 2.40 gm/cc 60.0% AIN 23.6 W/mK

1 (N₂) 2.73 gm/cc 75.0% AIN 27.3 W/mK

2 (Ar) 2.42 gm/cc 61.0% AIN 24.0 W/mK

2 (Ar) 2.41 gm/cc 61.0% AIN 22.2 W/mK

2 (AT) 2.50 gm/cc 65.0% AIN 26.1 W/mK

2 (AT) 2.66 gm/cc 72.0% AIN 45.1 W/mK

Fig. 9 shows the thermal conductivity of the "infiltrated" samples as a func¬ tion of AIN volume loading. The data is generally consistent with the linear relationship anticipated by the parallel plate mixing model. The single sample which achieved a thermal conductivity greater than 30 W/mK appears to be inconsistent with the results obtained from the other samples.

Extrapolation of the best-fit line to 100 percent AIN content provides an estimate of the thermal conductivity phase of approximately 40 W/mK. Oxygen analysis on the preforms was measured to be 0.7 percent. Because the surface area of the pre¬ form was below 0.5 m²/g, it is expected that most of this oxygen is contained with the lattice, where it has the greatest effect on thermal conductivity. Thus, a thermal conduc¬ tivity as low as 40 W/mK for this sintered part is not unusual. This value is considerably higher than the value calculated for the powder in the "mixed" system, as previously described. This result is believed to be obtained because the direct contact between the individual grains in the sintered preform eliminates the phonon-scattering effect of oxy¬ gen at the particle surfaces.

Because the oxygen content is so high, it is expected that the thermal con¬ ductivity of the preform (and likewise of the resulting composite) could be substantially increased by utilizing preform sintering techniques that would further reduce the lattice oxygen. In fact, the large effect that lattice oxygen appears to have on the thermal conductivity of the AIN, may be the source of the observed spread in the data. Small changes in oxygen content and the location between samples would be expected to result in noticeable differences in thermal conductivity.

Conclusion

Physical contact between the individual particles in AlN-filled polymer com¬ posite materials has been shown to have a significant effect on the thermal conductivity of such composites. Standard "mixed" samples exhibited thermal conductivity values con¬ sistent with a modified perpendicular plate composite model. The thermally-conductive material may be either a solid, a liquid, a gas, or a combination thereof.

Samples prepared with sintered AIN preforms exhibited thermal conductivi¬ ties an order of magnitude higher than standard "mixed" samples. The enhanced thermal conductivity of this group is believed to be related to the connectivity of the AIN parti¬ cles. The thermal conductivity response to increased filler concentration appears to fit a parallel plate composite model.

By selectively processing an AlN-filled composite system so as to insure inter-particle contact, the thermal conductivity of the resulting composite can be signifi¬ cantly modified.

Therefore, while a preferred form of composition of the improved compos- ite material has been shown and described, and various changes and modifications there¬ of discussed, persons skilled in this art will readily appreciate that various additional changes and modifications may be made without departing from the spirit of the inven¬ tion, as defined and differentiated by the following claims.

Claims

What is claimed is:

1. A composite material occupying a volume, comprising: a thermally-conductive material having a plurality of solid particles, said particles being bonded together to form at least one thermally-conductive path from one point in said volume to another point in said volume; and a thermally-insulative material occupying at least a portion of said volume not occupied by said thermally-conductive material and interconnected throughout said thermally-conductive material.

2. A composite material as set forth in claim 1 wherein said solid particles are granular.

3. A composite material as set forth in claim 1 wherein said thermally-conduc¬ tive material is a dielectric material.

4. A composite material as set forth in claim 3 wherein said thermally-conduc- tive material is aluminum nitride.

5. A composite material as set forth in claim 1 wherein said thermally-insulat¬ ive material is a glass.

6. A composite material as set forth in claim 1 wherein said thermally-insulat¬ ive material is a polymer.

7. A composite material as set forth in claim 1 wherein said thermally-conduc¬ tive material forms a three-dimensional network of thermally-conductive paths. 8. A composite material as set forth in claim 1 wherein said thermally-conduc¬ tive material and said thermally-insulative material form interpenetrating networks.

9. A composite material as set forth in claim 1 wherein said thermally-insulat¬ ive material has a dielectric constant less than or equal to the dielectric constant of said thermally-conductive material.

10. A composite material as set forth in claim 1 wherein no chemical interaction occurs between said thermally-conductive material and said thermally-insulative material.

11. The method of creating a composite material, comprising the steps of: creating a preform of a thermally-conductive material having a plurality of solid particles bonded together to form at least one thermally-conductive path from one point to another point; and infiltrating said preform with a thermally-insulative material; thereby to create a thermally-conductive composite material.

12. The method as set forth in claim 11, and further comprising the additional step of: causing the thermally-insulative material to solidify so as to prevent unin¬ tended separation thereof from said thermally-conductive material.

13. The method as set forth in claim 11, and further comprising the additional step of: exposing said thermally-conductive material at at least one portion of the surface of the infiltrated preform.

14. The method as set forth in claim 13 wherein said thermally-conductive mate¬ rial is exposed at at least two spaced portions of said infiltrated preform outer surface to establish a preferential thermally-conductive path therebetween. 15. The method as set forth in claim 11 wherein the step of creating said pre¬ form includes the steps of: mixing particles of said thermally-conductive material with an organic binder; forming the preform; and heating the preform to create particle-to-particle bonds between the parti¬ cles of thermally-conductive material and to remove the binder from the preform.

16. The method as set forth in claim 15 wherein the heating step causes the particles of thermally-conductive material to form a network of particle-to-particle bonds.

17. The method as set forth in claim 16 wherein the heating step occurs in a non-oxidizing atmosphere.

18. The method as set forth in claim 11 wherein the preform is infiltrated by: placing the preform in an enclosure; and introducing the thermally-insulative material into the enclosure.

19. The method as set forth in claim 11 wherein the infiltration step is provided by: filling at least some of the interstices of the preform with a glass powder; and heating the preform and powder to cause the powder to liquify about the thermally-conductive material.

20. The method as set forth in claim 11 wherein the step of infiltrating the preform with thermally-insulative material is pressure-assisted.

21. A composite material occupying a volume, formed by a process which in¬ cludes the steps of: creating a preform of a thermally-conductive material having a plurality of solid particles bonded together to form at least one thermally-conductive path from one point to another point; and infiltrating said preform with a thermally-insulative material; thereby to create a thermally-conductive composite material.

22. A composite material as set forth in claim 21, wherein the process of its formation further includes the additional step of: exposing said thermally-conductive material at at least one portion of the infiltrated preform.

23. A composite material as set forth in claim 22, wherein the process of its formation further includes the additional step of: causing the thermally-insulative material to solidify so as to prevent unin¬ tended separation thereof from said thermally-conductive material.

27. A composite material as set forth in claim 25, wherein the thermally-conduc¬ tive material is exposed at at least two spaced portions of said infiltrated preform to establish a preferential thermally-conductive path therebetween.