Thermal transport at a nanoparticle-water interface: A molecular dynamics and continuum modeling study

Abstract

Heat transfer between a silver nanoparticle and surrounding water has been studied using molecular dynamics (MD) simulations. The thermal conductance (Kapitza conductance) at the interface between a nanoparticle and surrounding water has been calculated using four different approaches: transient with/without temperature gradient (internal thermal resistance) in the nanoparticle, steady-state non-equilibrium and finally equilibrium simulations. The results of steady-state non-equilibrium and equilibrium are in agreement but differ from the transient approach results. MD simulations results also reveal that in the quenching process of a hot silver nanoparticle, heat dissipates into the solvent over a length-scale of ~ 2nm and over a timescale of less than 5ps. By introducing a continuum solid-like model and considering a heat conduction mechanism in water, it is observed that the results of the temperature distribution for water shells around the nanoparticle agree well with MD results. It is also found that the local water thermal conductivity around the nanoparticle is greater by about 50 percent than that of bulk water. These results have important implications for understanding heat transfer mechanisms in nanofluids systems and also for cancer photothermal therapy, wherein an accurate local description of heat transfer in an aqueous environment is crucial.

Enhanced Heat Transfer with Metal-Dielectric Core-Shell Nanoparticles

“…Similarly, heat transport in liquid water can also be modelled as diffusive. Recent MD simulations [79] have indeed showed that, in the vicinity of a metallic nanoparticle, the local thermal conductivity of water may be described by bulk water conductivity, except in the immediate vicinity (2 nm) of the nanoparticle and at very short time scales (t < 5 ps).…”

Optomechanical Processing of Silver Colloids: New Generation of Nanoparticle–Polymer Composites with Bactericidal Effect

“…Unfortunately, it is impossible to measure temperature directly at the nanoscale, thus molecular dynamics (MD) simulations are usually used. The problem of heat transfer between AgNPs and water was intensively studied by Rajabpour et al [54], who calculated values of interfacial thermal conductance for that system. Other important data, temperature profiles for nanoparticles submerged in different media, were determined by MD simulations in [55].…”

“…To estimate the meaningfulness of the continuum model, used for all further calculations, temperature profiles for the gold-octane system were calculated using finite element method (FEM) and compared with the data provided in [55] (Figure 12). Macroscopic constants of gold and octane were used and only Kapitza conductance values (G) were taken from the publication [54]. The comparison of these two methods suggests that although the continuum model is unable to capture small local increases in temperature in the vicinity of the NP (arising from enhanced density of the adsorbed liquid layer) it is precise enough for further calculations, providing an accurate description for the temperature of the NP core and surrounding liquid for distances above 2 nm.…”

“…The temperature on the outside of the computational cell was fixed at 0 and corresponding interfacial conditions were applied. Kapitza conductance from [54] was used to define contact conductance of nanoparticle-water interface. Since the literature data on the conductance at the nanoparticle-polymer interface vary significantly, the calculations were simulated for values of 10, 25, 50, 75, 100 MW m −2 •K −1 (the same value is assumed for all simulated polymers).…”

Cellular Thermometry Considerations for Probing Biochemical Pathways

“…On the other hand, continuum models such as finite element methods (FEM) approximate the temperature discontinuity across dissimilar interfaces using the interfacial resistance as an input parameter [48], [54]. Typically, the thermal interfacial resistances are either measured experimentally [33], [54] or calculated from MD simulations [48], [53]. MD based simulations are usually limited to confined domains of few nm 2 -μm 2 and are also limited by the computational capacity to be able to model an entire cell.…”