Publications

Periodic cove edges alter heat conduction in graphene nanoribbons in ways that are not fully captured by geometric boundary scattering. We investigate this effect using a graphene-nanoribbon-specific neuroevolution potential together with molecular dynamics, phonon Monte Carlo simulations, lattice dynamics, and full linearized phonon Boltzmann transport calculations. Cove-edged graphene nanoribbons exhibit lower thermal conductivity than pristine armchair nanoribbons over the length and width ranges considered. The conductivity reduction obtained from molecular dynamics is larger than that predicted by the semiclassical Monte Carlo model. The residual fraction is greatest at short lengths and small widths, whereas the Monte Carlo boundary contribution becomes relatively more important as either dimension increases. Participation ratios and real-space eigenvectors identify partially confined and strongly edge-confined modes in the low- and intermediate-frequency ranges. The cove-edged ribbons also show a smaller full-LBTE correction to the relaxation-time approximation and a lower transport-weighted Normal-to-Umklapp scattering-rate ratio. These results indicate that cove-edge patterning changes both the vibrational eigenstates and the intrinsic scattering balance, thereby reducing the mode-coupling correction beyond the independent-mode approximation.

Understanding phonon transport in alloys is crucial for the design of high-performance electronic and thermoelectric devices. However, conventional theoretical models fail to provide a clear physical picture of phonon scattering caused by atomic disorder in alloys, and their prediction accuracy is limited. In this work, a new substitutional atomic distance model for alloys is proposed, providing an intuitive physical picture. SiGe and InGaAs alloys are taken as representative systems, and their thermal conductivities are calculated, showing good agreement with previous experimental measurements. The results indicate that alloy scattering plays a dominant role in reducing thermal conductivity. This study provides new insights into phonon transport in alloys and offers guidance for tailoring thermal properties through compositional engineering.

Predictive thermal simulation of nanoscale semiconductor devices requires a Monte Carlo transport framework that can transfer mode-resolved phonon information from materials calculations to device-scale heat transport under complex confinement, localized heating, and spatially nonuniform temperature fields. Here we present PhonoMC, a full-band local-deviational phonon Monte Carlo framework that solves the phonon Boltzmann transport equation using locally updated equilibrium backgrounds and temperature-dependent modal properties. This formulation retains full-Brillouin-zone phonon information and reduces the statistical variance associated with full-population sampling in strongly nonuniform temperature fields. PhonoMC reproduces full-population Monte Carlo results with substantially lower noise, captures size-dependent thermal-conductivity suppression in silicon thin films, and resolves localized self-heating in a three-dimensional FinFET structure. Compared with Fourier-based finite-element calculations, PhonoMC predicts larger hotspot temperatures, revealing the contribution of nonlocal phonon transport to device thermal resistance. A PhonoMC-to-FEM bridging conductivity is further extracted from the single-device temperature field and transferred to a larger FinFET-array continuum model, demonstrating a scale-transfer route from phonon-resolved transport to continuum device-level simulation. These results establish PhonoMC as a computational framework for bridging materials-level phonon physics and larger-scale thermal response in nanoscale semiconductor devices.

Understanding nanoscale hotspot thermal transport is crucial in electronic devices. Contrary to common perception, recent experiments show that closely spaced nanoscale multiple hotspots can enhance heat dissipation. Here, the thermal transport in nanoscale multiple hotspot systems is investigated by solving the phonon Boltzmann transport equation. The local thermal conductivity is proposed to describe the non-uniform spatial distribution of heat transport capability in nanoscale multiple hotspot systems. The maximum value exceeds the uniform heating case by up to 27%, which is attributed to the spatially varying fraction of unscattered phonons emitted from hotspots. Moreover, the effects and mechanisms of hotspot spacing on thermal transport are investigated, showing that reducing the hotspot spacing can enhance the heat flux by up to 40%. This work challenges the conventional view that thermal transport capability is spatially uniform throughout the system and provides fundamental insights for thermal management in high-power-density integrated circuits.

Thermal management in nanostructured devices necessitates the accurate and efficient prediction of phonon transport properties. However, solving the Boltzmann transport equation via first-principles calculations is often computationally prohibitive due to the extensive supercells required to model realistic nanostructures. In this work, we present an accelerated, automated workflow that synergizes neuroevolution potentials with compressed sensing techniques to efficiently extract high-order anharmonic force constants. We validate this approach using silicon thin films as a prototype, explicitly accounting for the complexities of surface reconstruction. Our results demonstrate that this framework achieves accuracy comparable to density functional theory while reducing the computational cost by several orders of magnitude. The workflow successfully reproduces phonon dispersion relations and captures the temperature- and size-dependent trends of lattice thermal conductivity, incorporating the critical contribution of inter-mode coherence. This methodology offers a scalable and robust solution for the high-throughput thermal characterization of low-dimensional materials.

High Bandwidth Memory, as a key development trend in future memory chip technology, significantly enhances computer performance. At the same time, the thermal challenges arising from its stacked architecture have drawn considerable attention. Most existing studies on thermal management of high bandwidth memory are based on Fourier's law, neglecting the non-Fourier effects introduced by the micro/nanoscale structures. In this study, the Monte Carlo method is employed to solve the phonon Boltzmann transport equation and investigate the impact of non-Fourier heat transport on the thermal behavior of high bandwidth memory structures. The results reveal that non-Fourier heat transport leads to a junction temperature that is 59.8 K higher than that predicted by Fourier's law. Furthermore, it is found that the phonon transmittance at the chip interlayers has a severe impact on heat dissipation, with the temperature variation reaching up to 56.6 K. These findings provide more accurate thermal insights, which are critical for the optimized design of high bandwidth memory systems.

In this study, lattice dynamics calculations based on the Neuroevolution Machine-learned Potential (NEP) were performed for three types of silicon nanostructures: thin films, nanowires, and quantum dots. The temperature and size dependence of the specific heat capacity was systematically examined. The results reveal a significant enhancement in the specific heat capacity of nanostructures at low temperatures compared to bulk silicon, primarily due to phonon confinement, discrete energy spectra, and the emergence of low-frequency surface vibrational modes. These findings underscore the dominant role of nonlinear acoustic phonons at low temperatures, with increasing contributions from optical modes as the temperature rises. Notably, this work reports the temperature-dependent evolution of local fitting exponents in the specific heat scaling relation $C_v \sim T^{n(T)}$ for nanostructured systems. The high accuracy and computational efficiency of the NEP model allow for detailed characterization of the complex phonon behaviors that govern thermal properties at the nanoscale.

This paper presents an updated comprehensive model for specular reflection from a periodically rough boundary. The model accounts for both phase loss and angular deviations of the reflected wave, enabling a statistical evaluation of the specular reflection probability as a function of wavelength, roughness amplitude, roughness period, and incident angle. Notably, the model reveals oscillatory behavior in the specular reflection probability within the high-frequency regime. The proposed formulation is further incorporated into the Ziman model and applied in Monte Carlo simulations to evaluate thermal conductivity.

This study investigates the dual nature of phonons—encompassing both particle-like and wave-like behaviors—and their roles in thermal transport within pillared graphene nanoribbons (PGNRs). Monte Carlo simulations are employed to evaluate how the presence of pillars affects the thermal conductivity of graphene nanoribbons (GNRs), revealing that pillars significantly reduce thermal conductivity by enhancing phonon-boundary scattering, thereby emphasizing particle effects. A comparison with molecular dynamics simulations enables quantitative assessment of the respective contributions of particle and wave phenomena to the observed reduction in thermal conductivity. Notably, as the width of PGNRs decreases, the influence of wave effects initially increases and then diminishes, suggesting a saturation behavior. Furthermore, this study introduces and evaluates the concept of phonon resonance hybridization depth in PGNRs.

In this study, the temperature fluctuations in three-dimensional confined nanostructures (quantum dots) of germanium, silicon, and diamond were calculated using the \(T^{3/2}\) model based on the particle-in-a-box (PIAB) approach and compared with the Debye \(T^3\) model. The analysis focused on quantum dots with sizes ranging from 1 to 100 nm. According to the PIAB \(T^{3/2}\) model, temperature fluctuations decrease as the temperature decreases, consistent with the principles of statistical physics. In contrast, the Debye \(T^3\) model predicts an increase in temperature fluctuations with decreasing temperature, contradicting the principles of statistical physics. These results emphasize the significant impact of quantum confinement and highlight the limitations of the Debye \(T^3\) model in describing nanoscale systems. Furthermore, distribution diagrams illustrating temperature fluctuations as functions of size and temperature were established for the first time. Based on these diagrams, clear boundaries were defined for the temperature and thermophysical property ranges where reliable predictions can be made.

This study introduces a novel method for calculating the thermal conductivity of graphene using a Monte Carlo approach to evaluate anisotropic three-phonon interactions. The phonon dispersion relation is derived using a force constant model that incorporates up to fifth-order nearest-neighbor interactions, while the phonon density of states (DOS) is computed via a generalized Gilat-Raubenheimer method. A quantitative relationship for the scaling exponent of the specific heat capacity at low temperatures is established, emphasizing the unique two-dimensional characteristics of graphene. To address anisotropic effects, the Monte Carlo approach efficiently identifies three-phonon combinations that adhere to the conservation laws of energy and momentum. The findings highlight the pivotal role of anisotropic phonon interactions in graphene's thermal conductivity. The thermal conductivity values obtained through the iterative method exhibit strong agreement with previous three-phonon calculations, thereby validating the model. Nevertheless, discrepancies with experimental data suggest that incorporating higher-order phonon processes, such as four-phonon scattering, may further improve predictive accuracy.

We estimate the thermal properties of unsmooth Si nanowires, considering key factors such as size (diameter), surface texture (roughness) and quantum size effects (phonon states) at different temperatures. For nanowires with a diameter of less than 20 nm, we highlight the importance of quantum size effects in heat capacity calculations, using dispersion relations derived from the modified frequency equation for the elasticity of a rod. The thermal conductivities of nanowires with diameters of 37, 56, and 115nm are predicted using the Fuchs– Sondheimer model and Soffer’s specular parameter. Notably, the roughness parameters are chosen to reflect the technological characteristics of the real surfaces. Our findings reveal that surface texture plays a significant role in thermal conductivity, particularly in the realm of ballistic heat transfer within nanowires. This study provides practical recommendations for developing new thermal management materials.

The results of the validation of the STEG code using the experiments on two-phase flow across horizontal tube bundles are presented. The experiments are carried out on two water–air SG models, consisting of a transparent vessels, inside of which there is a tube bundle, a perforated sheet and a bead separating the upstream section of the circulation circuit of the model from the downstream one. The void fractions and water velocities are measured in different locations of the models. The brief description of the STEG code based on the 3D two-fluid model is presented. A set of interfacial drag correlations, which was recently developed by the authors, are used in the validation calculations. The peculiarities of the spatial two-phase flows have been established, and a quantitative comparison with experimental data has been performed. Good agreement between the calculated and experimental data is obtained.