We introduce a large “synthetic” dataset of atomistic structures and energies, generated using a fast machine-learning model, and we demonstrate its usefulness for supervised and unsupervised ML tasks in chemistry.
28 February 2024 in The Journal of Chemical Physics
Amorphous ice phases are key constituents of water’s complex structural landscape. This study investigates the polyamorphic nature of water, focusing on the complexities within low-density amorphous ice (LDA), high-density amorphous ice, and the recently discovered medium-density amorphous ice (MDA). We use rotationally invariant, high-dimensional order parameters to capture a wide spectrum of local symmetries for the characterization of local oxygen environments. We train a neural network to classify these local environments and investigate the distinctiveness of MDA within the structural landscape of amorphous ice. Our results highlight the difficulty in accurately differentiating MDA from LDA due to structural similarities. Beyond water, our methodology can be applied to investigate the structural properties and phases of disordered materials.
We use atomistic and coarse-grained machine-learning models to address a long-standing question: to what extent are ZIFs analogous to inorganic AB2 phases?
The splitting of the principal peak in the structure factor of liquid water is commonly interpreted as evidence of a competition between two distinct local environments. Here, we show that this peak splitting is captured by medium-range topological features of the hydrogen-bond network. Using atomistic simulations, we systematically decompose the structure factor into contributions from hydrogen-bonded rings of different sizes. We find that 5–8-membered rings, which dominate the network topology of liquid water at low temperatures, can directly explain the experimentally observed bimodal scattering signal. Among these, 5-membered rings are particularly persistent, maintaining distinct structural signatures even above room temperature. Our findings establish a direct link between the network topology of liquid water and experimentally accessible diffraction features and provide a topological perspective on the nature of the split first diffraction peak in tetrahedral materials.
We report the magnetic behaviour of the hybrid perovskites [Gua]Mn1–xFe2x/3(HCOO)3 (0 <x 0.6. The behaviour is rationalised in terms of a simple microscopic model of percolation-induced ferrimagnetism. Monte Carlo simulations driven by this model reproduce the experimental dependence of magnetic susceptibility on x and show that, at intermediate compositions, domains of short-range vacancy order lead to the emergence of local magnetisation. Our results open up a new avenue for the design of multiferroic hybrid perovskites.