Lattice thermal conductivity is important for many applications,but experimental measurements or first principles calculations including three-phonon and four-phonon scattering are expensive or even unaffordable.Machi...Lattice thermal conductivity is important for many applications,but experimental measurements or first principles calculations including three-phonon and four-phonon scattering are expensive or even unaffordable.Machine learning approaches that can achieve similar accuracy have been a long-standing open question.Despite recent progress,machine learning models using structural information as descriptors fall short of experimental or first principles accuracy.This study presents a machine learning approach that predicts phonon scattering rates and thermal conductivity with experimental and first principles accuracy.The success of our approach is enabled by mitigating computational challenges associated with the high skewness of phonon scattering rates and their complex contributions to the total thermal resistance.Transfer learning between different orders of phonon scattering can further improve the model performance.Our surrogates offer up to two orders of magnitude acceleration compared to first principles calculations and would enable large-scale thermal transport informatics.展开更多
The prediction of thermal conductivity and radiative properties is crucial.However,computing phonon scattering,especially for four-phonon scattering,could be prohibitively expensive,and the thermal conductivity for si...The prediction of thermal conductivity and radiative properties is crucial.However,computing phonon scattering,especially for four-phonon scattering,could be prohibitively expensive,and the thermal conductivity for silicon after considering four-phonon scattering is significantly under-predicted and not converged in the literature.Here we propose a method to estimate scattering rates from a small sample of scattering processes using maximum likelihood estimation.The calculation of scattering rates and associated thermal conductivity and radiative properties are dramatically accelerated by three to four orders of magnitude.This allows us to use an unprecedented q-mesh(discretized grid in the reciprocal space)of 32×32×32 for calculating four-phonon scattering of silicon and achieve a converged thermal conductivity value that agrees much better with experiments.The accuracy and efficiency of our approach make it ideal for the high-throughput screening of materials for thermal and optical applications.展开更多
基金Z.G.,D.F.,and X.R.acknowledge partial support from the US National Science Foundation through award 2102645Z.H.and X.R.acknowledge partial support from the US National Science Foundation through award 2015946+1 种基金Z.G.acknowledges the Ross Fellowship from Purdue University.L.G.acknowledges the support from the National Science Foundation(DMS-1555072,DMS-2053746,and DMS-2134209)U.S.Department of Energy(DOE)Office of Science Advanced Scientific Computing Research program DE-SC0021142 and DE-SC0023161.
文摘Lattice thermal conductivity is important for many applications,but experimental measurements or first principles calculations including three-phonon and four-phonon scattering are expensive or even unaffordable.Machine learning approaches that can achieve similar accuracy have been a long-standing open question.Despite recent progress,machine learning models using structural information as descriptors fall short of experimental or first principles accuracy.This study presents a machine learning approach that predicts phonon scattering rates and thermal conductivity with experimental and first principles accuracy.The success of our approach is enabled by mitigating computational challenges associated with the high skewness of phonon scattering rates and their complex contributions to the total thermal resistance.Transfer learning between different orders of phonon scattering can further improve the model performance.Our surrogates offer up to two orders of magnitude acceleration compared to first principles calculations and would enable large-scale thermal transport informatics.
基金supported by National Science Foundation(Award No.2102645)supported by National Science Foundation(Award No.2321301)supported by National Science Foundation(Award No.2311848).
文摘The prediction of thermal conductivity and radiative properties is crucial.However,computing phonon scattering,especially for four-phonon scattering,could be prohibitively expensive,and the thermal conductivity for silicon after considering four-phonon scattering is significantly under-predicted and not converged in the literature.Here we propose a method to estimate scattering rates from a small sample of scattering processes using maximum likelihood estimation.The calculation of scattering rates and associated thermal conductivity and radiative properties are dramatically accelerated by three to four orders of magnitude.This allows us to use an unprecedented q-mesh(discretized grid in the reciprocal space)of 32×32×32 for calculating four-phonon scattering of silicon and achieve a converged thermal conductivity value that agrees much better with experiments.The accuracy and efficiency of our approach make it ideal for the high-throughput screening of materials for thermal and optical applications.
