This EArly-concept Grants for Exploratory Research (EAGER) grant provides funding for the processing and characterization of novel indium-graphene and copper-graphene composites for heat spreader applications. The processing steps will consist of rolling indium and copper foils with dispersed graphene oxide followed by intermediate heating so that agglomeration of graphene in the composites is prevented. The metal oxide formed by reduction of graphene oxide to graphene will be removed by indium flux or by subjecting the copper-graphene composite to hydrogen at higher temperature. The size and uniform distribution of graphene in the composites will be characterized by advanced methods that include scanning and transmission electron microscopy. The thermal conductivity will be determined experimentally as a function of temperature and volume fraction of graphene in the composites. Measurement of interfacial thermal conductance between graphene and the matrix metal will be performed to verify that it is not a limiting factor to achieve high thermal conductivity of the composites. Modeling of the thermal transport properties of the composites will be carried out to evaluate the importance of microstructural features, interfacial thermal conductance, and thermal conductivity of graphene and metal matrix. Semiconductor wafers will be bonded to the copper-graphene composites using the indium-graphene composite solder. The bonded structures will be subjected to thermal cycling to determine the stability and resistance to failure by debonding.
If successful, the results of this research will lead to significant improvements in the thermal conductivity of the thermal interface material and heat spreader. The bottleneck to thermal energy dissipation, which is very critical to the reliability and operation of high frequency and high power electronic devices and lasers, will be eliminated. The primary goal of the present research is to process graphene composites with uniform dispersion of graphene and achieve high heat spreader capability. Determination of thermal transport properties in terms of the properties of the individual components will be helpful to determine the contribution from graphene which is considered the future novel electronic material. Low cost manufacturing methods developed in this project will enable semiconductor and laser packaging technologies to overcome the thermal management problems.
Low cost processing of high thermal conductivity indium-graphene (In-gr), indium-gallium-graphene (In-Ga-gr), and copper-graphene (Cu-gr) composites was investigated and thermal conductivity was characterized. We have used two methods to prepare In-gr, In-Ga-gr and one method for the Cu-gr composites. Graphene oxide (GO) was synthesized by oxidation of microcrystalline graphite and chemical exfoliation. In the first method, the composites were prepared with the GO. The suspension of GO in isopropyl alcohol was deposited on In or In-Ga foils and repeated folding, rolling, and annealing was used to achieve uniform distribution of GO. The Cu-gr composites were prepared by electrochemical co-deposition of Cu and GO at low current density using slightly acidic copper sulfate bath containing suspension of GO. The distribution of GO was very uniform. The composites were heated to reduce the GO to graphene. Indium oxide and indium gallium oxide formed due to reduction of GO were removed using an indium flux. The composites with Cu were heated in hydrogen near 400oC so that copper oxide formed by reduction of GO to graphene was converted back to pure Cu. Further annealing under vacuum at 400oC was used to remove residual oxygen. Also, the volume fraction of graphene is easily controlled by increasing the suspension of GO in the electrochemical bath. The GPs in Cu were uniformly distributed and randomly oriented to enable isotropic thermal conductivity. The preparation of In-gr composites by reduction of GO and use of flux to remove the indium oxide was found to give rise to loss of some indium in the flux. Price of indium is increasing and therefore we have used an alternate method wherein indium oxide formation is avoided. In this second method, GO was reduced thermally at 400oC in vacuum. Also, Nd-YAG laser beam (wave length =266 nm) incidence was used to reduce GO to graphene. The resulting graphene was suspended in isopropyl alcohol, sonicated for three hours, and was used to form the composites. We find that thermal reduction of GO to prepare composites with In is a more cost effective method. These methods of preparing graphene composites with In, In-Ga and Cu are very low cost and industrially suitable. Microstructural characterization showed a uniform distribution of graphene with random orientation. The thermal conductivity of indium improved from 0.7 to 1.6 Wcm-1K-1 upon introducing graphene to volume fraction of 0.18. The factor of improvement was higher in In-Ga-gr composites. The thermal conductivity of Cu was improved from 3.8 to 5.0 Wcm-1K-1 upon introducing graphene to volume fraction between 0.2 and 0.25. The interface thermal conductance between graphene and indium or copper in the a-b plane was found to be high so that it is not a limiting factor in the improvement of thermal conductivity of the composites. The two factors that were found to be important for improvement in the thermal conductivity are the size and thickness of graphene. Exfoliation of GO by sonication for extended period of time is responsible for smaller average size of graphene. Also, graphene platelets are better because phonons in thin graphene film are quenched by the matrix. Determination of volume fraction of graphene from backscattering SEM images is suitable to determine thicker graphene because contrast is strong. Thinner graphene with single or two layers does not provide a strong contrast. Electrical conductivity of thicker graphene is low and that of thinner graphene is high. Therefore, we have used combination of electrical and thermal conductivity measurements and image analysis to determine the volume fraction of graphene. It is important to eliminate all the point defects and residual impurities so that the improvement in electrical conductivity is only a measure of the graphene present. Reliability of the bond strength using In-gr as interface material and Cu-gr as heat spreader was tested by bonding a thin silicon wafer (180 mm). The bonded sample is cycled through heating and cooling between 270 and 350 K. The thermal conductivity of the bonded structure was measured after two hundred cycles. It was found that the wafer bond was stable. Two graduate students completed MS thesis and graduated. An undergraduate student has worked as REU participant with funding provided by the North Carolina State University undergraduate scholarship program. Sixteen publications were prepared. Results were published on the internet. Many industries showed interest in the technology. Data on electrical resistivity; three-omega signal data on thermal conductivity of In-gr, In-Ga-gr, and Cu-gr; thermo reflectance data on the Cu-gr and In-gr; and Raman data in excel format, Optical, SEM and TEM micrographs in TIFF format, and X-ray data in excel format are produced. Samples, physical collections and other products include In-gr, In-Ga-gr, and Cu-gr. samples, graphene, thermo reflectance set up, 3w and electrical resistivity set up in temperature controlled chamber. Technology transfer is in progress.
