Thermal conductance at the interface between crystals using equilibrium and nonequilibrium molecular dynamics

Abstract

In this article, we compare the results of non-equilibrium (NEMD) and equilibrium (EMD) molecular dynamics methods to compute the thermal conductance at the interface between solids. We propose to probe the thermal conductance using equilibrium simulations measuring the decay of the thermally induced energy fluctuations of each solid. We also show that NEMD and EMD give generally speaking inconsistent results for the thermal conductance: Green Kubo simulations probe the Landauer conductance between two solids which assumes phonons on both sides of the interface to be at equilibrium. On the other hand, we show that NEMD give access to the out-ofequilibrium interfacial conductance consistent with the interfacial flux describing phonon transport in each solid. The difference may be large and reaches typically a factor 5 for interfaces between usual semi-conductors. We analyze finite size effects for the two determinations of the interfacial thermal conductance, and show that the equilibrium simulations suffer from severe size effects as compared to NEMD. We also compare the predictions of the two above mentioned methods -EMD and NEMD-regarding the interfacial conductance of a series of mass mismatched Lennard-Jones solids. We show that the Kapitza conductance obtained with EMD can be well described using the classical diffuse mismatch model (DMM). On the other hand, NEMD simulations results are consistent with a out-of-equilibrium generalisation of the acoustic mismatch model (AMM). These considerations are important in rationalizing previous results obtained using molecular dynamics, and help in pinpointing the physical scattering mechanisms taking place at atomically perfect interfaces between solids, which is a prerequesite to understand interfacial heat transfer across real interfaces.

Atomic density effects on temperature characteristics and thermal transport at grain boundaries through a proper bin size selection

“…Numerous MD studies showed that the temperature drops abruptly at the interface due to the phonon scattering and ITR. 7,[13][14][15] Meanwhile, it was additionally found that there are other temperature gradients at solid-solid interfaces. [10][11][12] These gradients are much steeper than those in the bulk solid.…”

“…MD simulation has been widely applied to investigate the thermal transport at GBs. [7][8][9][10][11][12][13][14][15][16][17][18][19][20] By using MD simulation, it is possible to directly observe the independent molecular motion, which has so far been very difficult to obtain through experiments. In addition, the ITR can be interpreted by phonons, which are thermal energy carriers which play an important role when studying nanoscale heat transfer in a superlattice.…”

Strong Coupling between Nanofluidic Transport and Interfacial Chemistry: How Defect Reactivity Controls Liquid–Solid Friction through Hydrogen Bonding

“…In order to confirm the results of the friction calculations with Eq. (1), and following a similar approach applied to the measure of thermal conductance, 55 we also estimated the friction coefficient using an alternative Green-Kubo formula involving the initial slope of the autocorrelation function of the slip velocity fluctuations at equilibrium, C vv = v slip (t)v slip (0) :…”

Relationship of thermal boundary conductance to structure from an analytical model plus molecular dynamics simulations

“…On the other hand, MD simulations have been successfully used to directly calculate thermal boundary resistance [21][22][23][24][25][26][27] . These studies established the effects of sizes [28][29][30][31] , temperature 21,26,32 , mass differential of the two layers 21,26,27 , lattice mismatch between layers 26,27 , interfacial defects 26,27 , stiffness of the materials, and bond strength at the interface 22,24,32,33 . While the MD data provided knowledge significantly beyond that achieved from analytical models (e.g., the acoustic mismatch model and the diffuse mismatch model 2,9 ), our current understanding of the thermal boundary conduc-tance is far from the material/structure design requirement.…”