The dynamics of glass forming liquids shows a tremendous slowing down if temperature is decreased: Relaxation times, transport coefficients such as the diffusion constant or the shear viscosity grow by more than 15 orders of magnitude within narrow temperature range. One of the central goals in the field of glass physics is to understand the origin of this dramatic slowing down of the dynamics.There are several scenarios that seem to be able to explain the slowing down of the dynamics and the low temperature state of the glass former. A very popular one is to rationalize the slow dynamics by invoking the existence of a thermodynamic transition, the so-called ideal glass transition from the liquid to an ideal glass states. This ideal glass transition occurs when the configurational entropy, defined as the logarithm of the number of available states, becomes zero. Due to thissingular behavior of the configurational entropy, the relaxation time andtransport coefficients diverge if the transition point is approached [1].However, confirming the existence of the ideal glass transition is a verydifficult task since in practice the system falls out of equilibriumbefore reaching this putative ideal glass transition temperature,thus making the test of this theory most difficult. Furthermore, otherscenarios are also able to explain the slow dynamics without resortingany thermodynamic transition [2].In the present study, we investigate the existence of the idealglass transition by using computer simulation of a canonical binaryLennard-Jones glass former in three dimensions. Massive computationaleffort and efficient sampling algorithm (parallel tempering) allow us toreach equilibrium states that are very deeply supercooled. By measuringthe configurational entropy directly (obtained via thermodynamicintegration), we find that the configuration entropy does not go to zeroat a finite temperature. This implies that the thermodynamic singularityis indeed avoided before reaching a putative ideal glass transitiontemperature predicted by previous studies. Our results indicates thatcontrary to the previous results the system remains in a liquid statedown to zero temperature without showing an ideal glass transition.Furthermore, we analyze the microscopic structure and the potentialenergy landscape of the system to clarify the mechanism that leads tothe avoidance of the ideal glass transition. We find that locallyfavored structures present in the liquid state hardly change in thetemperature range we consider, which rules out another possibility forthe slow dynamics, e.g. the transformation between two distinct liquids,so-called liquid-liquid transition [3]. Instead, we find the potentialenergy landscape contains extensive number of minima even very lowtemperature, which rationalize the avoided singularity.Our results support a theoretical argument that the ideal glasstransition can not exist at bulk system in finite dimensions [5].[1] Adam G and Gibbs JH, J. Chem. Phys., 43, 139 (1965)[2] Chandler D and Garrahan JP, Annu. Rev. Phys. Chem., 61, 191 (2010)[3] Speck T, Royall CP, and Williams SR, arXiv:1409.0751, (2014)[4] Saksaengwijit A and Heuer A, Phys. Rev. Lett., 93, 235701 (2004)[5] Stillinger FH, J. Chem. Phys., 88, 7818 (1988)

We discuss two procedures to obtain empirical potentials from ab initio trajectories. The first methodconsists in adjusting the parameters of an empirical pair potential so that the radial distribution functionsextracted from classical simulations using this potential match the ones extracted from the ab initio sim-ulations. As a case study, we consider the example of amorphous silica, a material that is highly relevantin the field of glass science as well as in geology. With our approach we are able to obtain an empiricalpotential that gives a better description with respect to structural and thermodynamic properties thanthe potential proposed by van Beest, Kramer, and van Santen, and that has been very frequently usedas a model for amorphous silica. The second method is the so-called ‘‘force matching” approach proposedby Ercolessi and Adams to obtain an empirical potential. We demonstrate that for the case of silica thismethod does not yield a reliable potential and discuss the likely origin for this failure.