Machine-Learning-Accelerated DFT Conformal Sampling of Catalytic Processes.

Computational modeling of catalytic processes at gas/solid interfaces plays an increasingly important role in chemistry, enabling accelerated materials and process optimization and rational design. However, efficiency, accuracy, thoroughness, and throughput must be enhanced to maximize its practical impact. By combining interpolation of DFT energetics via highly accurate Machine-Learning Potentials with conformal techniques for building the training database, we present here an original approach (that we name Conformal Sampling of Catalytic Processes, CSCP), to accelerate and achieve an accurate and thorough sampling of novel systems by exporting existing information on a worked-out case. We use methanol decomposition (of interest in the field of hydrogen production and storage) as a test catalytic reaction. Starting from worked-out Pt-based systems, we show that after only two iterations of active-learning CSCP is able to provide reaction energy diagrams for a set of 7 diverse systems (Pd, Ni, Au, Ag, Cu, Co, Fe) leading to DFT-accuracy-level predictions. Cases exhibiting a change in adsorption sites and mechanisms are also successfully reproduced as tests of catalytic path modification. The CSCP approach thus offers itself as an operative tool to fully take advantage of accumulated information to achieve high-throughput sampling of catalytic processes.

Functionalized Amorphous Carbon Materials via Reactive Molecular Dynamics Simulations.

We derive a database of atomistic structural models of amorphous carbon materials endowed with exohedral functional groups. We start from phases previously derived using the DynReaxMas method for reactive molecular dynamics simulations, which exhibit atomistic and medium-length-scale features in excellent agreement with available experimental data. Given a generic input structure/phase, we develop postprocessing simulation algorithms mimicking experimental preparation protocols aimed at: (1) "curing" the phase to decrease the defect concentration; (2) automatically selecting the most reactive carbon atoms via interaction with a probe molecular species, and (3) stabilizing the phase by saturating the valence of carbon atoms with single-bond functional groups. Although we focus on oxygen-bearing functionalities, they can be replaced with other monovalent groups, such as -H, -COOH, -CHO, so that the protocol is quite general. We finally classify reactive sites in terms of their location within the structural framework and coordination environment (edges, tunnels, rings, aromatic carbons becoming aliphatic) and try to single out descriptors that correlate with tendency to functionalization.

Atomistic mechanisms underlying plastic flow at ultralow yield stress in ductile carbon aerogels.

We investigated carbon aerogel samples with super low densities of 0.013 g cm-3 (graphite is 2.5) and conducted compression experiments showing a very low yield stress of 5-8 kPa. To understand the atomistic mechanisms operating in these super low density aerogels, we present a computational study of the mechanical response of very low-density amorphous carbonaceous materials. We start from our previously derived atomistic models (based on the DynReaxMas method) with a density of 0.16 g cm-3 representing the core regions of carbon aerogels. We considered three different phases exhibiting either a fiber-like clump morphology interconnected with string-like units or a more reticulated framework. We subjected these phases to compression and shear deformations and analyzed the resulting plastic response via an inherent-structure protocol. Strikingly, we find that these materials possess shear plastic relaxation modes with extremely low values of yield stress, negligible with respect to the finite values predicted outside this "zero-stress" region. This is followed by a succession of two additional regimes with increasing yield stress values. Our analysis of the atomistic relaxation mechanisms finds that these modes have a collective and cooperative character, taking the form of nanoscopic shear bands within the clumps. These findings rationalize our experimental observations of very low-stress plastic deformation modes in carbon aerogels, providing the first steps for developing a predictive multi-scale modeling of the mechanical properties of aerogel materials.