Search

Interfacial thermal resistance (ITR) is a critical property for the performance of nanostructured devices where phonon mean free paths are larger than the characteristic length scales. The affordable, accurate and reliable prediction of ITR is essential for material selection in thermal management. In this work, the state-of-the-art machine learning methods were employed to realize this. Descriptor selection was conducted to build robust models and provide guidelines on determining the most important characteristics for targets. Firstly, decision tree (DT) was adopted to calculate the descriptor importances. And descriptor subsets with topX highest importances were chosen (topX-DT, X = 20, 15, 10, 5) to build models. To verify the transferability of the descriptors picked by decision tree, models based on kernel ridge regression, Gaussian process regression and K-nearest neighbors were also evaluated. Afterwards, univariate selection (UV) was utilized to sort descriptors. Finally, the top5 common descriptors selected by DT and UV were used to build concise models. The performance of these refined models is comparable to models using all descriptors, which indicates the high accuracy and reliability of these selection methods. Our strategy results in concise machine learning models for a fast prediction of ITR for thermal management applications.

Boron nitride nanosheets (BNNS) hold the similar two-dimensional structure as graphene and unique properties complementary to graphene, which makes it attractive in application ranging from electronics to energy storage. The exfoliation of boron nitride (BN) still remains challenge and hinders the applications of BNNS. In this work, the preparation of BNNS has been realized by a shear-assisted supercritical CO2 exfoliation process, during which supercritical CO2 intercalates and diffuses between boron nitride layers, and then the exfoliation of BN layers is obtained in the rapid depressurization process by overcoming the van der Waals forces. Our results indicate that the bulk boron nitride has been successfully exfoliated into thin nanosheets with an average 6 layers. It is found that the produced BNNS is well-dispersed in isopropyl alcohol (IPA) with a higher extinction coefficient compared with the bulk BN. Moreover, the BNNS/epoxy composite used as thermal interface materials has been prepared. The introduction of BNNS results in a 313% enhancement in thermal conductivity. Our results demonstrate that BNNS produced by supercritical CO2 exfoliation show great potential applications for heat dissipation of high efficiency electronics.

Thermal interface materials (TIMs) are crucial components of high density electronics and the high thermal conductivity of graphite makes this material an attractive candidate for such applications. We report an investigation of the in-plane and through-plane electrical and thermal conductivities of thin thermal interface layers of graphite nanoplatelet (GNP) based composites. The in-plane electrical conductivity exceeds its through-plane counterpart by three orders of magnitude, whereas the ratio of the thermal conductivities is about 5. Scanning electron microscopy reveals that the anisotropy in the transport properties is due to the in-plane alignment of the GNPs which occurs during the formation of the thermal interface layer. Because the alignment in the thermal interface layer suppresses the through-plane component of the thermal conductivity, the anisotropy strongly degrades the performance of GNP-based composites in the geometry required for typical thermal management applications and must be taken into account in the development of GNP-based TIMs.

Chemically modified single-walled carbon nanotubes (SWNTs) with varying degrees of functionalization were utilized for the fabrication of SWNT thin film catalyst support layers (CSLs) in polymer electrolyte membrane fuel cells (PEMFCs), which were suitable for benchmarking against the US DOE 2017 targets. Use of the optimum level of SWNT -COOH functionality allowed the construction of a prototype SWNT-based PEMFC with total Pt loading of 0.06 mg(Pt)/cm²--well below the value of 0.125 mg(Pt)/cm² set as the US DOE 2017 technical target for total Pt group metals (PGM) loading. This prototype PEMFC also approaches the technical target for the total Pt content per kW of power (0.125 g(PGM)/kW) at cell potential 0.65 V: a value of 0.15 g(Pt)/kW was achieved at 80°C/22 psig testing conditions, which was further reduced to 0.12 g(Pt)/kW at 35 psig back pressure.