Amorphous materials are being described by increasingly powerful computer simulations, but new approaches are still needed to fully understand their intricate atomic structures. Here, we show how machine‐learning‐based techniques can give new, quantitative chemical insight into the atomic‐scale structure of amorphous silicon ( a ‐Si). We combine a quantitative description of the nearest‐ and next‐nearest‐neighbor structure with a quantitative description of local stability. The analysis is applied to an ensemble of a ‐Si networks in which we tailor the degree of ordering by varying the quench rates down to 10 10 K s −1 . Our approach associates coordination defects in a ‐Si with distinct stability regions and it has also been applied to liquid Si, where it traces a clear‐cut transition in local energies during vitrification. The method is straightforward and inexpensive to apply, and therefore expected to have more general significance for developing a quantitative understanding of liquid and amorphous states of matter.

In this paper, we offer large and realistic models of amorphous carbon spanning densities from 0.95 g/cm 3 to 3.5 g/cm 3 . The models are designed to agree as closely as possible with experimental diffraction data while simultaneously attaining a local minimum of a density functional Hamiltonian. The structure varies dramatically from interconnected wrapped and defective sp 2 sheets at 0.95 g/cm 3 to a nearly perfect tetrahedral topology at 3.5 g/cm 3 . Force Enhanced Atomic Refinement (FEAR) was used and is shown here to be computationally superior and more experimentally realistic than conventional ab initio melt quench methods. We thoroughly characterize our models by computing structural, electronic and vibrational spectra. The vibrational density of states of the 0.95 g/cm 3 model is strikingly similar to monolayer amorphous graphene. Our sp 2 /sp 3 ratios are close to experimental predictions where available, a consequence of compelling a satisfactory fit for pair correlation function. arXiv:1712.01437v1 [cond-mat.dis-nn]

In this paper, we provide evidence that low density nano-porous amorphous carbon (a-C) consists of interconnected regions of amorphous graphene (a-G). We include experimental information in producing models, while retaining the power and accuracy of ab initio methods with no biasing assumptions. Our models are highly disordered with predominant sp2 bonding and ring connectivity mainly of sizes 5-8. The structural, dynamical and electronic signatures of our 3-D amorphous graphene are similar to those of monolayer amorphous graphene. We predict an extended x-ray absorption fine structure (EXAFS) signature of amorphous graphene. Electronic density of states calculations for 3-D amorphous graphene reveal similarity to monolayer amorphous graphene and the system is non conducting.

Amorphous materials are being described by increasingly powerful computer simulations,b ut new approaches are still needed to fully understand their intricate atomic structures.Here,weshowhow machine-learning-based techniques can give new,quantitative chemical insight into the atomic-scale structure of amorphous silicon (a-Si). We combine aquantitative description of the nearest-and next-nearestneighbor structure with aq uantitative description of local stability.T he analysis is applied to an ensemble of a-Si networks in whichwetailor the degree of orderingbyvarying the quenchr ates down to 10 10 Ks À1 .O ur approach associates coordination defects in a-Si with distinct stability regions and it has also been applied to liquid Si, where it traces ac lear-cut transition in local energies during vitrification. The method is straightforwarda nd inexpensive to apply,a nd therefore expected to have more general significance for developing aquantitative understanding of liquid and amorphous states of matter.Supportinginformation and the ORCID identification number(s) for the author(s) of this article can be found under: https://doi.org/10.