THERMALLY CONDUCTIVE GREASE

CROSS REFERENCE TO RELATED APPLICATION

This application claims the benefit of U.S. Provisional Patent Application No. 60/824,599, filed September 5, 2006, the disclosure of which is incorporated by reference herein in its entirety.

BACKGROUND

The invention relates to thermal interface materials and their use. In the computer industry, there is a continual movement to higher computing power and speed. Microprocessors are being made with smaller and smaller feature sizes to increase calculation speeds. Consequently, power flux is increased and more heat is generated per unit area of the microprocessor. As the heat output of the microprocessors increases, heat or "thermal management" becomes more of a challenge. One aspect of thermal management is known in the industry as a "thermal interface material" or "TIM" whereby such a material is placed between a heat source, such as a microprocessor, and a heat dissipation device to facilitate the heat transfer. Such TIMs may be in the form of a grease or a sheet- like material. These thermal interface materials also are used to eliminate any insulating air between the microprocessor and heat dissipation device. TIMs typically are used to thermally connect a heat source to a heat spreader, that is, a thermally conductive plate larger than the heat source, in which case they are referred to as TIM Is. TIMs may also be employed between a heat spreader and a thermal dissipation device such as a cooling device or a finned heat sink in which case such TIMs are referred to as TIM Us. TIMs may be present in one or both locations in a particular installation.

SUMMARY

In one embodiment, the invention provides a thermally conductive grease that comprises 0 to about 49.5 weight percent of carrier oil, about 0.5 to about 25 weight percent of at least one dispersant, and at least about 50 weight percent of thermally conductive particles. The thermally conductive particles comprise a mixture of at least three distributions of thermally conductive particles, each of the at least three distributions of thermally conductive particles having an average (D ₅₀ ) particle size which differs from the other distributions by at least a factor of 5.

In another embodiment, the invention provides a method of making a thermally conductive grease of the invention that comprises the steps of providing carrier oil, dispersant, and thermally conductive particles, and then mixing the carrier oil (if present), dispersant, and thermally conductive particles together. In one aspect, the carrier oil (if present) and dispersant are mixed together, and the thermally conductive particles are mixed sequentially, finest to largest average particle size into the carrier oil and dispersant mixture. In another aspect, the thermally conductive particles are mixed together, and then mixed into the carrier oil (if present) and dispersant mixture. In another aspect, a portion or all of the thermally conductive particles are pre-dispersed with a dispersant prior to mixing the thermally conductive particles into the carrier oil (if present) and dispersant mixture.

In another embodiment, the invention provides a microelectronic package comprising a substrate, at least one microelectronic heat source attached to the substrate, and a thermally conductive grease disclosed in this application on the at least one microelectronic heat source.

In one aspect, the invention provides the above microelectronic package further comprising a heat spreader and thermally conductive grease disclosed in this application between the microelectronic heat source and the heat spreader.

In another aspect, the invention provides a microelectronic package comprising a substrate, at least one microelectronic heat source attached to the substrate, a heat spreader, and a heat dissipation device attached to the heat spreader wherein a thermally conductive grease disclosed in this application is between the heat spreader and the heat dissipation device.

In another aspect, the invention provides a microelectronic package comprising a substrate, at least one microelectronic heat source attached to the substrate, a heat spreader, a thermally conductive grease disclosed in this application between the microelectronic heat source and the heat spreader and a heat dissipation device wherein thermally conductive grease is between the heat spreader and the heat dissipation device.

DETAILED DESCRIPTION

As used herein:

"Grease" means a material having a viscosity of greater than I X lO ⁴ cps (10 Pa.s) at 1/s shear rate and 20 ⁰ C and a viscosity of less than 10 ⁸ cps at 1/sec shear rate and 125°C.

"Thermally conductive grease" means grease having a bulk conductivity of greater than 0.05 W/m-K as measured by the test method Bulk Thermal Conductivity described below.

All numbers are herein assumed to be modified by the term "about," unless stated otherwise. The recitation of numerical ranges by endpoints includes all numbers subsumed within that range (e.g., 1 to 5 includes 1, 1.5, 2, 2.75, 3, 3.80, 4, and 5).

Thermally conductive greases (TCGs) of the invention may contain one or more carrier oils. Carrier oil provides the base or matrix for the TCGs of the invention. Useful carrier oils may comprise synthetic oils or mineral oils, or a combination thereof and are typically flowable at ambient temperature. Specific examples of useful carrier oils include polyol esters, epoxides, silicone oils, and polyolefms or a combination thereof.

Commercially available carrier oils include HATCOL 1106, a polyol ester of dipentaerythritol and short chain fatty acids, and HATCOL 3371, a complexed polyol ester of trimethylol propane, adipic acid, caprylic acid, and capric acid (both available form Hatco Corporation, Fords, NJ); and HELOXY 71 an aliphatic epoxy ester resin, available from Hexion Specialty Chemicals, Inc., Houston TX.

Carrier oil may be present in the TCGs of the invention in an amount of from 0 to about 49.5 weight percent, and in other embodiments, from 0 to not more than about 20 or about 12 weight percent of the total composition. In other embodiments, carrier oil may be present in an amount of at least 2, 1, or 0.5 weight percent of the composition. Carrier oil may also be present in the TCGs of the invention in ranges including from about 0.5, 1, or 2 to about 12, 15, or 20 weight percent.

TCGs of the invention contain one or more dispersants. The dispersant(s) may be present in combination with carrier oil, or may be present in the absence of carrier oil. The dispersants improve the dispersion of the thermally conductive particles (described below) in the carrier oil if present. Useful dispersants may be characterized as polymeric or ionic in nature. Ionic dispersants may be anionic or cationic. In some embodiments, the dispersant may be nonionic. Combinations of dispersants may be used, such as, the combination of an ionic and a polymeric dispersant. In some embodiments, a single dispersant is used. Examples of useful dispersants include, but not limited to, polyamines, sulfonates, modified polycaprolactones, organic phosphate esters, fatty acids, salts of fatty acids,

polyethers, polyesters, and polyols, and inorganic dispersants such as surface-modified inorganic nanoparticles, or any combination thereof.

Commercially available dispersants include those having the tradenames SOLSPERSE 24000 and SOLSPERSE 39000 hyperdispersants, available from Noveon, Inc., a subsidiary of Lubrizol Corporation, Cleveland, OH; EFKA 4046, a modified polyurethane dispersant, available from Efka Additives BV, Heerenveen, The Netherlands; and RHODAFAC RE-610, an organic phosphate ester, available from Rhone-Poulenc, Plains Road, Granbury, NJ.

Dispersant is present in the TCGs of the invention in an amount of at least 0.5 and not more than 50 weight percent, and in other embodiments, not more than 25, 10, or 5 weight percent of the total composition. In another embodiment, dispersant may be present in an amount of at least 1 weight percent. Dispersant may also be present in the TCGs of the invention in ranges including from about 1 to about 5 weight percent.

TCGs of the invention contain thermally conductive particles. Useful thermally conductive particles include those made from or that comprise diamond, polycrystalline diamond, silicon carbide, alumina, boron nitride (hexagonal or cubic), boron carbide, silica, graphite, amorphous carbon, aluminum nitride, aluminum, zinc oxide, nickel, tungsten, silver, and combinations of any of them. Each of these particles is of a different type.

The thermally conductive particles used in the TCGs of the invention are a mixture of at least three distributions of thermally conductive particles. Each of the at least three distributions of thermally conductive particles have an average particle size which differs from the average particle size of the distribution above and/or below it by at least a factor of 5, and in other embodiments, at least a factor of 7.5, or at least a factor of 10, or greater than 10. For example, a mixture of thermally conductive particles may consist essentially of: a smallest particle distribution having an average particle diameter (D ₅₀ ) of 0.3 micrometers; a middle distribution having an average particle diameter (D ₅₀ ) of 3.0 micrometers; and a largest distribution having an average particle diameter (D ₅₀ ) of 30 micrometers. Another example may have average diameter particle distributions having average particle diameter (D ₅₀ ) values of 0.03 micrometers, 0.3 micrometers, and 3 micrometers. The thermally conductive particles used in the TCGs of the invention are a mixture of at least three distributions of thermally conductive particles resulting in at least a trimodal distribution. In such a trimodal distribution, the minima between the peaks (distance

between the baseline of the peaks and the lowest point of the valley between distribution peaks) may be no more than 75, 50, 20, 10 or 5 percent of the interpolated value (height) between adjacent peaks. In some embodiments, the three size distributions are essentially non-overlapping. "Essentially non-overlapping" means that the lowest point of the valley is no more than 5% of the interpolated value between adjacent peaks. In other embodiments, the three distributions have only a minimal overlap. "Minimal overlap" means that the lowest point of the valley is no more than 20% of the interpolated value between adjacent peaks.

Typically, for a trimodal TCG, the average particle size for the third smallest (or smaller) average diameter may range from about 0.02 to about 5.0 micrometers (μm).

Typically, the average particle size for the middle average diameter may range from about 0.10 to about 50.0 μm. Typically, the average particle size for the largest average diameter may range from about 0.5 to about 500 μm.

In some embodiments, it is desirable to provide a TCG having the maximum possible volume fraction thermally conductive particles that is consistent with the desirable physical properties of the resulting TCG, for example, that the TCG conform to the surfaces with which it is in contact and that the TCG be sufficiently flowable to allow easy application.

With this in mind, the conductive particle distributions may be selected in accordance with the following general principles. The distribution of largest diameter particles should have diameters that are smaller than, or nearly bridge, the expected gap between the two substrates to be thermally connected. Indeed, the largest particles may bridge the smallest gap between substrates. When the particles of the largest diameter distribution are in contact with each other, a gap or void volume between the particles will remain. The mean diameter of the middle diameter distribution may be advantageously selected to just fit within the gap or void between the larger particles. The insertion of the middle diameter distribution will create a population of smaller gaps or voids between the particles of the largest diameter distribution and the particles of the middle diameter distribution the dimensions of which may be used to select the mean diameter of the smallest distribution. In a similar fashion, desirable mean particle dimensions may be selected for fourth, fifth, or higher order populations of particles if desired.

Each distribution of thermally conductive particles may comprise the same or different thermally conductive particles in each or any of the at least three distributions. Additionally, each distribution of thermally conductive particles may contain a mixture of different types of thermally conductive particles. The remaining voids may be thought of as being filled with carrier, dispersant(s) and other components with a slight excess to provide flowability. Further guidance in the selection of suitable particle distributions may be found in "Recursive Packing of Dense Particle Mixtures", Journal of Materials Science Letters, 21, (2002), pages 1249-1251. From the foregoing discussion, it will be seen that the mean diameters of the successive particle size distributions will preferably be quite distinct and well separated to ensure that they will fit within the interstices left by the previously packed particles without significantly disturbing the packing of the previously packed particles.

The thermally conductive particles may be present in the TCGs of the invention in an amount of at least 50 percent by weight. In other embodiments, thermally conductive particles may be present in amounts of at least 70, 75, 80, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, or 98 weight percent. In other embodiments, thermally conductive particles may be present in the TCGs of the invention in an amount of not more than 99, 98, 97, 96, 95, 94, 93, 92, 91, 90, 89, 88, 87, 86, or 85 weight percent.

The TCGs and TCG compositions of the invention may also optionally include additives such as antiloading agents, antioxidants, leveling agents and solvents (to reduce application viscosity), for example, methylethyl ketone (MEK), methylisobutyl ketone, and esters such as butyl acetate.

In some embodiments, ZnO is selected for the smallest particles (or third from largest with middle-size or medium-size intervening), diamond or silicon carbide is selected for the medium-size particles, and metal particles are selected for the largest particles.

The TCGs of the invention are generally made by blending dispersant and carrier oil together, and then blending the thermally conductive particles sequentially, finest to largest average particle size into the dispersant/carrier oil mixture. The thermally conductive particles may also be premixed with one another, and then added to the liquid components. Heat may be added to the mixture in order to reduce the overall viscosity and aid in reaching a uniformly dispersed mixture. In some embodiments, it may be desirable to first pretreat or

pre-disperse a portion or all of the thermally conductive particles with dispersant prior to mixing the particles into the dispersant/carrier mixture.

In other embodiments, the TCGs of the invention can be made by solvent casting the blended components, then drying to remove the solvent. For example, the TCG component blend can be provided on a suitable release surface, e.g., a release liner or carrier.

In other embodiments, the TCGs of the invention can be applied to a carrier, or to the device in the intended use, with the aid of an energy source, e.g., heat, light, sound, or other known energy source.

In some embodiments, the invention is further illustrated by observing the force as a function of thickness (e.g., gap in a test device such as described below) for normal and extended tests, as described below with several working examples. Control materials exhibit very similar forces in either test, whereas the inventive materials exhibit much higher resistance to forces opposing reducing the gap after the extended time interval, so that the gap either cannot be closed further, or such that it can be closed more only with great difficulty. In the normal test, the test gap may be reduced by removing or repositioning the mechanical stop which had limited gap closure and applying a nominal load of about ten pounds (about 4.5 kg) to the test fixture. In the case of particularly high viscosity compositions, the settling of the test fixture to the new gap setting may be accelerated by mechanically vibrating the fixture to take advantage of the shear-thinning characteristics which are typical of these materials. One or both of these measures will normally suffice to reduce the gap and to settle onto the new mechanical stopped position until the gap approaches the diameter of the largest particles in the composition, typically between 30 and 60 micrometers. In some embodiments, the thermally conductive grease substantially resists flow after aging for several hours at a temperature above about 50 ⁰ C. In one exemplary embodiment, the material under test was initially equilibrated at a 375 μm gap. Following the collection of thermal data, the gap was reduced to 289 μm with the normal ten pound down force. Again following the collection of thermal data, the gap was reduced to 211 μm, , again with only the normal 10 pound down force. The sample was then thermally characterized and allowed to stand undisturbed for about 15 hours at which time the gap could not be reduced below 200 μm with the normal down force. In a second exemplary embodiment, the material under test was initially equilibrated at a 411 μm gap. The sample was then thermally characterized and allowed to stand undisturbed for about 15 hours at

which time the gap was reduced to 295 μm by the use of 30 pound down force and vibration Again following the collection of thermal data, the gap was reduced to 245 μm, again with a >30 pound down force and substantial vibration. Once more the thermal data was collected and the gap was reduced to 205 μm with a >30 pound down force and even more substantial vibration. The thermal data was again collected, but all attempts to further reduce the gap below 205 μm with a combination of weight and vibration were fruitless. Testing of related compositions did not exhibit a resistance to gap closure at shorter standing intervals of up to and including four hours. Related compositions having similar sized largest particle distributions, but not showing the effect, could be reduced to a 30 to 55 micron gap after standing in the test fixture for similar periods of time. Temperatures encountered during the test depend upon the gap and the thermal characteristics of the composition under test, however they typically are between about 45°C and 85°C with a gradient between the hot and cold surfaces. The resistance to flow described above typically does not build in the manner described at or about room temperature. In addition, simple heating to 75°C or even about 100 ⁰ C followed by cooling to room temperature does not induce the resistance to flow which is seen in the extended test.

A sample can be deemed substantially rigid, or substantially resistant to flow, if the sample cannot be compressed by at least 50 μm in the Extended Test (described below) starting with a gap of at least 150 μm. In some embodiments, the suitable exposure temperatures used when these exemplary materials resist flow or become substantially rigid is above about 70 ⁰ C, above about 100 ⁰ C, above about 110 or 120 ⁰ C, or even higher.

In some embodiments, the suitable exposure times used when these exemplary materials resist flow or become substantially rigid is generally at least a few hours. In other embodiments, this time (hours) is at least about 2, at least about 4, at least about 6, at least about 8, at least about 12, or even longer.

In some embodiments, the thermally conductive grease described herein substantially resists flow after about 12 hours aging at a temperature above about 50 ⁰ C. In other embodiments, the thermally conductive grease described herein becomes substantially rigid after about 12 hours aging at a temperature above about 50 ⁰ C.

In some embodiments, preferred combinations of materials of the present invention incorporate Hatcol 1106 as the carrier, Solsperse 16000 as a sole dispersant, and a blend of

zinc oxide (small particle size distribution), spherical aluminum (large particle size distribution) and either diamond or silicon carbide particles (medium particle size distribution).

The TCGs of the invention may be used in microelectronic packages and may be used to assist in the dissipation of heat from a heat source, for example, a microelectronic die or chip to a thermal dissipation device. Microelectronic packages may comprise at least one heat source, for example, a die mounted on a substrate or stacked die on a substrate, a thermally conductive grease of the invention on the heat source, and may include an additional thermal dissipation device in thermal and physical contact with the die, such as, for example, a thermal spreader. A thermal spreader may also be a heat source for any subsequent thermal dissipation device. The thermally conductive greases of the invention are useful to provide thermal contact between said die and thermal dissipation device. Additionally, TCGs of the invention may also be used in thermal and physical contact between a thermal dissipation device and a cooling device. In another embodiment, the TCGs of the invention may be used between a heat generating device and a cooling device, that is, without using a heat or thermal spreader in between. TCGs of the invention are useful in TIM I and TIM II applications.

EXAMPLES Bulk Thermal Conductivity

Bulk thermal conductivity was measured generally in accord with ASTM D-5470-01 on the TCG samples using a Heat Transfer Tester, available from Custom Automation, Inc., Blaine, MN. The Heat Transfer Tester was built according to Proposal Number 3M- 102204-01 and included such features as: a vision system capable of measuring parallelism and gap between copper meter bars for up to 0.010 inch (0.254 mm) gaps, copper meter bars with 5 resistance temperature detector (RTD) sensors on each meter bar, a cooler to cool the cooled clamping block (to hold the cooled meter bar) having an operating range of from -20 to 100 ⁰ C and can hold the coolant temperature to +/- 0.02 ⁰ C, a 25 IbF load cell mounted on a X-Y micrometer adjust positioning stage, a cooled clamping block (to hold the cooled meter bar) mounted on the load cell, a heated clamping block (to hold the heated meter bar) using resistive heating and has its temperature controlled by a controller and thermocouple, the ability to add weights above the heated clamping block to

adjust the contact force on the meter bars from 5 to 50 N, and means to measure and record temperature, meter bar gap, and contact force at time intervals to a spreadsheet.

The vision system used to measure meter bar gap was calibrated as outlined in the operating procedures provided. The cooler was charged with a 50/50 blend of water and ethylene glycol. The gap between the copper meter bars was set at about 550 micrometers at room temperature. The heater set point was put at 120 ⁰ C and the cooler set point at -5°C, and the unit was allowed to equilibrate. The meter bar gap after equilibration was about 400 micrometers. The surfaces of the hot and cold meter bars were planarized using the individual meter bar turnbuckles until the gap between the meter bars read by each of the three individual cameras fell within a +/- 3 μm range.

Normal Test (Examples 75-85)

The heater set point was put at 120 ⁰ C and the cooler set point at -5°C, and the unit was allowed to equilibrate. The meter bar gap after thermal equilibration was mechanically adjusted to be about 400 micrometers. The surfaces of the hot and cold meter bars were planarized using the individual meter bar turnbuckles until the gap between the meter bars read by each of the three individual cameras fell within a +/- 3 μm range.

An excess of each TCG sample tested was placed on the hot meter bar surface and smoothed across the entire face. The head was then closed and clamped into place with about 10 pounds force (4.5 kg force) pressing down on the head, causing excess TCG sample to ooze out of the meter bar gap until a mechanical stop was reached. This excess TCG was removed with a paper towel or a fine cloth and the pins of the meter bars were cleaned to facilitate accurate measurement of the gap by the three vision cameras. The instrument was allowed to equilibrate for about 10 minutes as data was continuously recorded. Then the mechanical stop was adjusted to allow the meter bar gap to be lowered about 100 μm and the excess TCG sample again oozed out of the gap and was cleaned. The instrument was again allowed to equilibrate for about 10 minutes as data was continuously recorded. This sequence of lowering the meter bar gap in about 100 μm increments, cleaning, and recording data was repeated until a final reading was taken, typically at a meter bar gap of <100 μm. The meter bars were opened back up to about a 400 μm gap, cleaned, and the procedure was repeated for the next sample.

Extended Test (Examples 75-85)

An "extended test" was run identically to the "normal test" except that the sample was allowed to remain in the tester without changing the gap for a minimum of 12 hours. The gap setting chosen is optional, but should be greater than 200 μm to see the effects of the invention more readily. An excess of each TCG sample was placed on the hot meter bar surface and smoothed across the entire face. The head was then closed and clamped into place, causing excess TCG sample to ooze out of the meter bar gap. This excess was removed with a paper towel or a fine cloth and the pins of the meter bars were cleaned to facilitate accurate measurement of the gap by the three vision cameras. The procedure of allowing 10 minute equilibration times, collecting data, lowering the gap, cleaning the vision pins, and repeating the cycle was continued until the gap chosen for an extended time interval was reached. The material was left in the tester, and data continued to be collected, throughout the extended time. The procedure of lowering the head, cleaning the pins, and allowing 10 minute equilibration periods was resumed. In cases of the inventive material with this Extended Test, the gap either could not be closed further, or could be closed only with additional weights and manual vibration of the hot bar. In cases of control material, the gap could continue to be reduced for successive measurements down to the final gap of < 100 μm with no more than the 10 pound weight (4.5 kg force) used on top of the hot bar initially. The data were recorded every 7-8 seconds by the instrument and contained a time/date stamp, the sample name, the force exerted on the TCG in the meter bar gap, each of the individual meter bar gap readings, and each of the 10 RTD sensor temperature readings. The file was downloaded into a spreadsheet for analysis. In the analysis, the last 10 data points recorded at the given gap were averaged, and these averages were used for the calculations.

The power flowing through the TCG sample was calculated using the known bulk conductivity of copper, the dimensions of the copper bars, and the locations of the RTD temperature sensors. Typically, the calculations indicated slightly different wattage flowing down the hot meter bar than down the cold meter bar; these two values were averaged for calculations extending to the TCG sample. The temperature at the surface of each of the meter bars was also extrapolated from a plot of the temperatures and the RTD sensor locations.

The power, the average of the three individual meter bar gaps, the temperature drop across the meter bar gap, and the cross sectional area of the hot/cold meter bars were then used to calculate the temperature gradient, the power flux, and then the thermal impedance for the TCG sample under those conditions. These calculations were completed for each of the meter bar gaps at which the TCG sample had been tested, and the resulting thermal impedance and average gap data was plotted. A line was fitted to the data using spreadsheet software, and the bulk conductivity was calculated as the inverse of the lines' slope. The y-axis intercept and the slope were then used to calculate the thermal impedance at a 100 μm meter bar gap.

Viscosity

The viscosity data on selected samples was generated on a Rheometrics RD A3 viscometer (TA Instruments, Newcastle, DE). The viscometer was run with disposable 1 inch (25.4 mm) diameter parallel plates in the log sweep mode starting at 0.5/sec initial shear rate, taking 5 points/decade up to 1000/sec shear rate. The gap was set at 0.5 mm for a run, and then lowered to 0.25 mm for a second run on some samples; on other samples the gap was set and run only at 0.25 mm. Temperatures of the runs were controlled to either 125°C or 25°C as indicated in the table below. Viscosities were recorded in mPa.s at a 1.25/sec shear rate.

Milling Procedure

Roughly 40 cc of 0.5 mm diameter yttria-stabilized zirconia beads (available from Tosoh, Hudson, OH or from Toray Ceramics, George Missbach & Co., Atlanta, GA) were put into the basket of a Hockmeyer HM-1/16 Micro Mill ("Hockmeyer mill") (Hockmeyer Equipment Corp., Harrison, NJ). The desired MEK and dispersant (SOLSPERSE) were added to the mill chamber and stirred with an air mixer for at least 4 minutes so as to dissolve the dispersant in the solvent. The diamond particles were weighed into the chamber and the contents were stirred for an additional minute to wet out the diamond particles. The resulting mixture was then milled at the maximum speed of the Hockmeyer that avoided splashing. The resulting slurry was poured into a polyethylene container and the solvent was allowed to evaporate until it could not be detected by odor. Details of the compositions milled are shown below.

Glossary

"Sulfonated Bis(pentane dicaprolactone)", an ionic dispersant, was prepared in a reactor equipped with a mechanical stirrer under vacuum to which was added 25 g (0.476 equivalents) 1,5-pentane diol from Aldrich Chemical Co., Milwaukee, WI, 54.3 g (.476 equivalents) caprolactone from Aldrich Chemical Co., and 8.0 g (0.054 equivalents) dimethyl-5-sodiosulfoisophthalate available from DuPont Chemicals, Wilmington, DE. The reactor contents were stirred and heated to 170 ⁰ C with a vacuum at 115 mm mercury.

The reaction was complete after 4 h and a sample was analyzed by infrared spectroscopy. The final product was a clear, low viscosity liquid with a theoretical sulfonate equivalent weight of 1342.

"iC8 Modified silica nanoparticles", a nonionic, inorganic dispersant, was prepared by combining 61.42 g BS1316 isooctyltrimethoxysilane (Wacker Silicones Corp., Adrian, MI), 1940 g l-methoxy-2-propanol and 1000 g NALCO 2326 colloidal silica in a 1 gallon glass jar. The mixture was shaken to ensure mixing and then placed in an oven at 80 ⁰ C overnight. The mixture was then dried in a flow through oven at 150 ⁰ C to produce a white particulate solid. "HIMOD", a sulfonated polyol ionic dispersant, was prepared in a reactor equipped with a mechanical stirrer, nitrogen purge, and distillation apparatus which was charged with dimethyl-5-sodiosulfoisophthalate (42.6 g, 0.144 moles, available from DuPont Chemicals, Wilmington, DE), polyethylene glycol having a molecular weight of 400 (115.1 g, 0.288 moles, from Dow Chemical Co., Midland, MI), and polypropylene glycol having a molecular weight of 425 (122.3 g, 0.288 moles, available from Aldrich Chemical Co., Milwaukee, WI), and xylene (75 g). The reactor was slowly heated to 220 ⁰ C for about 1 hour to remove the xylene. Zinc acetate (0.2 g) was then added to the reactor and the temperature was held at 220 ⁰ C for 4 h with concomitant distillation of methanol from the reaction. The temperature was reduced to about 160 ⁰ C and 0.2 Torr (SI) vacuum was applied to the resulting mixture for 30 minutes. The contents were cooled to 120 ⁰ C under nitrogen to yield a clear, colorless polyol. The OH equivalent was determined to be 310g/mole OH and the theoretical sulfonated equivalent weight was found to be 1882 g polymer/mole sulfonated.

"TCPA HATCOL 3371", an ionic dispersant was prepared in a reactor equipped with a mechanical stirrer, and nitrogen purge into which was added 45 g (0.0241 equivalents) HATCOL 3371 and 3.4 g (0.0121 equivalents) tetrachlorophthalic anhydride. The reactor contents were stirred and heated to 150 ⁰ C with a constant nitrogen purge. The reaction was complete after 4 h and a sample was analyzed by infrared spectroscopy. The final product was a brown, low viscosity liquid with a theoretical acid equivalent weight of 18,127.

"TONE 305 TCPA", an ionic dispersant, was prepared in a reactor equipped with a mechanical stirrer, and nitrogen purge into which was added 10 g (0.1 equivalents) Tone

305 from Dow Chemical Co., and 1.0 g (0.00355 equivalents) tetrachlorophthalic anhydride from Aldrich Chemical. The reactor contents were stirred and heated to 105 ⁰ C with a constant nitrogen purge. The reaction was complete after 4 h and a sample was analyzed by infrared spectroscopy. The final product was a clear, low viscosity liquid with a theoretical acid equivalent weight of 3 , 100.

Sample Preparation

Except as noted in specific Examples, dispersant or mixture of dispersants was weighed into a watch glass. Any other surface active ingredients, if present, were also weighed onto the watch glass. Carrier oil(s), if present, was added to the dispersant(s) and the mixture was stirred with a metal spatula until the dispersant(s) was fully mixed into the carrier oil. Thermally conductive particles were added to the dispersant(s)/carrier oil mixture sequentially, starting with the smallest particle size distribution. Each of the thermally conductive particle distributions was dispersed into the dispersant(s)/carrier oil mixture with a metal spatula before adding the next distribution of thermally conductive particles. If necessary, the thermally conductive grease composition was heated in an oven (110 ⁰ C) to reduce the viscosity of the composition to facilitate mixing of the thermally conductive particles and/or subsequent additions of thermally conductive particles. The resultant thermally conductive greases were transferred into and stored in capped glass vials.

In cases where the thermally conductive particles were pre-dispersed, the amount of dispersant to be carried on the fine thermally conductive particle distribution was calculated. The amount of remaining dispersant necessary for the formulation was then determined and was weighed on to a watch glass. The remaining steps are identical to those described above.

Mixing Procedure I

The antioxidant and silica were weighed into a 115 mm diameter watch glass. The dispersant(s) and the carrier oil were then added, followed by the fine and the medium thermally conductive particle distributions. The mixture was stirred with a metal spatula until the combination of ingredients was a smooth and uniform blend. The coarse particles were then added and the contents of the watch glass were again stirred/ kneaded with the metal spatula until the composite was a smooth and uniform blend. (If necessary, the

mixture was heated in a hot air recirculating oven set at about 100-11O ⁰ C to reduce the sample viscosity and allow easier and more complete mixing and dispersion.) Once the last mineral distribution had been added and fully dispersed, the resultant TIM was then transferred to a glass vial, capped, and held for thermal testing.

Mixing Procedure II

The antioxidant, silica or carbon black, dispersant(s) package, and carrier fluid were all weighed into a polypropylene jar ("Max 100 g White Cup", from Flacktek, Inc., Landrum, SC). The finest of the mineral distributions was then weighed into the cup, and the cup was capped with a corresponding screw-top lid and inserted into a Speedmixer DAC FV (from Flacktek, Inc.). The Speedmixer was run at 3000 rpm for 30 seconds. The unit was opened, the cup removed and opened, and the next coarser particle size was weighed into the cup. The cup was again closed, inserted into the Speedmixer, and run at 3000 rpm for 30 seconds. The unit was again opened, the cup removed and opened, and the coarsest particle size was weighed into the cup. The cup was closed, inserted into the Speedmixer, and run at 3000 rpm for 30 seconds. The Speedmixer was run another cycle at 3300 rpm for one minute. Mixtures containing aluminum powder were optionally heated to about 100 ⁰ C and run for another minute in the Speed mixer at 3300 rpm to assure that the combination of ingredients was a smooth and uniform blend. The resulting TIM material was stored in the mixing cup.

Examples 1-64

The compositions of Examples 1-64 are shown in TABLE 1. The compositions of Examples A-N and 65-74 are shown in TABLE 2. TABLE 3 shows data resulting from the measurement of bulk conductivity and thermal impedance for selected Examples. TABLE 4 shows viscosity data for selected Examples.

TABLE 1

Example 44 contained a 4 ^th thermally conductive particle: DP 2, (4.41 grams), (60 μm).

Examples 46-48 and 50-54 used 0.25, 0.50, or 1.0 μm pre-dispersed diamond particles prepared according to the Milling Procedure and Sample Preparation described above.

Examples A-N and 65-74

Except as noted below, the components were individually weighed into a watch glass and mixed as follows. The silica, antioxidant, dispersants, and the carrier oil were initially combined with both the fine and the medium thermally conductive particles by stirring with a metal spatula until the combination of ingredients was a smooth and uniform blend. The largest particles were then added and the contents of the watch glass were again stirred/kneaded with the metal spatula until the composite was a smooth and uniform blend. If necessary, the thermally conductive grease composition was heated in an oven (110 ⁰ C) to reduce the viscosity of the composition to facilitate mixing of the thermally conductive particles and/or .subsequent additions of thermally conductive particles. The resultant thermally conductive greases were transferred into and stored in capped glass vials. The preparation of certain samples was the same as above except that about 16.5 grams of a

pre-blend of antioxidant, silica, dispersants, and carrier fluid was prepared. The mixture was stirred with a metal spatula until the combination of ingredients was a smooth and uniform blend. Then on a clean watch glass about 0.824 gram of the pre-blend and both the fine and the medium thermally conductive particles were combined with stirring, followed by the largest particles. The certain samples and the pre-blend compositions are described below.

Examples J, K, L, and I were prepared using Pre-blend A. Examples 65, 67, and 71 and Examples M and N were prepared using Pre-blend B.

TABLE 2

TABLE 3

CE 1 = ShinEtsu G751, Sample 1; CE 2 = ShinEtsu G751, Sample 2; CE 3 = Dow Corning TC5022; CE 4 = ShinEtsu G751, Sample 3

TABLE 4

The compositions reported in Table 5, using the amounts reported in Table 6, were prepared by mixing methods described above.

TABLE 5

TABLE 6

Amounts are in grams.

* Example 76 also contained 0.9111 g of dispersant Rhodafac RE-610.

TABLE 7

Example 78-B was a repeat of Example 78-A. Example 78-C exposed the same composition (as in Examples 78-A and 78-B) in an oven set at 80 ⁰ C for about 16 hours, before the sample was cooled to room temperature, before performing the Extended Test.

TABLE 8: Observations from "Extended Tests"

Foreseeable modifications and alterations of this invention will be apparent to those skilled in the art without departing from the scope and spirit of this invention. This invention should not be restricted to the embodiments that are set forth in this application for illustrative purposes.