We present an empirical potential developed for silicon under conditions of strong electronic excitation. We show the essentially athermal nature of the melting transition when the electronic temperature is extremely high. The resulting liquid is shown to be distinct from ordinary liquid silicon. For less intense excitations, we determine the thermal melting temperature and demonstrate the possible existence of a regime where ordinary thermodynamic melting can occur but at a reduced temperature Tm. We show laser-induced softening of the lattice can lead to lattice cooling for very short time scales (∼100 fs), an effect never before recognized.

We present a detailed analysis of a recently-developed empirical potential to describe silicon under conditions of strong electronic excitation. The parameters of the potential are given as smooth functions of the electronic temperature Te, with the dependence determined by fitting to finite-temperature density-functional theory calculations. We analyze the thermodynamics of this potential as a function of the electronic temperature Te and lattice temperature Tion. The potential predicts phonon spectra in good agreement with finite-temperature density-functional theory, including the previously predicted lattice instability. We predict that the melting temperature Tm decreases strongly as a function of Te. Electronic excitation has a strong effect on the rate of crystallization from the melt. In particular, high Te results in very slow kinetics for growing crystal from the melt, due mainly to the fact that diamond becomes much less stable as Te increases. Finally, we explore annealing amorphous Si (a-Si) below Tm, and find that we cannot observe annealing of a-Si directly at high Te. We hypothesize that this is also due to the decreased stability of the diamond structure at high Te.

We report molecular-dynamics simulations of Pd:H to elucidate transport properties, with special focus placed on determining the temperature dependence of the heat of transport Q Ã. Simulation results are analyzed using the Green-Kubo approach. It is found that Q Ã describing the thermodiffusion of hydrogen increases linearly with temperature. By contrast, the reduced heat of transport Q Ã0 ¼ Q Ã À h 2 , with h 2 the partial enthalpy of hydrogen, is approximately temperature independent. By computing separately the potential, kinetic, and virial contributions to Q Ã , it is possible to understand key features of the thermodiffusion process. In particular, the sum of the kinetic and potential energy of hydrogen atoms is increased above that of an average hydrogen atom by an amount comparable to the migration energy during a successful hop. However, the virial term in the energy flux is less than what would be expected based on the average local stress contribution due to the hydrogen atoms. Detailed calculations show that the relevant component of the stress tensor due to a hopping hydrogen atom exhibits a minimum at the transition state. Hence, while Q Ã has significant positive contributions due to the excited nature of the hopping hydrogen atom, the reduced heat of transport Q Ã0 can still be negative. The results here present important insight into the failure of simple kinetic theories of thermodiffusion, and provide a new perspective that can be tested on other systems. V

No abstract

No abstract