Atomistic simulation studies of complex carbon and silicon systems using environment-dependent tight-binding potentials

Fast and Accurate Prediction of Material Properties with Three-Body Tight-Binding Model for the Periodic Table

“…While there are various methods to alleviate this problem by including neighborhood-dependent hoppings [42][43][44][45], here, we directly include three-body terms in our fitting [41]. For example, consider H pzA,sB , the interaction between the p z -orbital on atom A and the s-orbital on atom B in Fig.…”

“…First, in addition to the typical two-body (two-center) atom-atom interactions, we use three-body (three-center) terms [41] to predict the tightbinding Hamiltonian from atomic positions. Including explicit three-body terms allows the Hamiltonian matrix elements between a pair of atoms to be environment dependent [42][43][44][45]. This creates a more transferable model that can be applied with equal accuracy to many crystal structures and that better takes advantage of the abundance of DFT data available from modern computational resources.…”

TeaNet: universal neural network interatomic potential inspired by iterative electronic relaxations

“…E /atom 5 6 7 10 8 12 14 2 9 10 5 6 3 8 6 9 3 5 −427.13 −3.34 0 0 0 0 0 0 0 0 1 0 0 6 0 0 0 0 1 0 −15.13 −1.89 0 0 0 0 16 0 0 0 0 0 0 0 0 0 0 0 0 0 −88.14 −5.51 2 1 1 1 0 1 1 0 2 1 3 0 0 1 0 1 0 1 −39.07 −2.44 0 0 0 1 1 0 0 2 2 0 1 1 1 0 3 0 0 0 −42.05 −3.50 1 1 2 2 3 0 0 0 0 0 1 1 0 0 0 2 2 1 −39.07 −2.44 0 1 1 1 1 0 0 3 0 1 0 0 1 1 0 0 0 0 −32.15 −3.22 8 0 0 0 0 8 0 0 0 0 0 0 0 0 0 0 0 0 −71.87 −4.49 1 0 1 1 0 3 5 0 1 0 1 1 1 1 1 1 1 1 −66.54 −3.33 1 1 3 1 0 0 1 2 0 1 1 0 0 2 0 1 1 1 −46.73 −2.92 0 1 0 1 0 2 1 1 0 0 1 1 0 0 0 2 1 1 −38. 19 −3.18 0 2 1 1 1 1 1 0 2 0 0 1 1 1 0 1 0 3 −46.10 −2.88 0 0 8 0 16 0 8 16 8 8 0 0 16 0 32 8 8 0 −461.06 −3.60 0 0 0 0 0 0 0 0 0 0 0 0 15 0 0 1 0 0 −44.16 −2.76 11 8 8 6 8 11 6 3 5 12 9 6 6 6 6 4 6 7 −374.90 −2.93 0 0 0 0 0 0 0 0 0 0 0 0 16 0 0 0 0 0 −39.70 −2.48 6 0 0 0 0 8 2 0 0 0 0 0 0 0 0 0 0 0 −75.53 −4.72 1 0 0 1 0 1 1 0 2 0 0 0 2 3 1 2 0 2 −56.83 −3.55 1 1 0 0 0 0 1 2 2 1 1 0 0 0 0 2 5 0 −32.44 −2.03 0 0 1 1 1 0 1 0 0 1 2 0 0 0 0 0 0 1 −7.19 −0.90 0 0 0 0 0 0 8 0 0 0 0 0 0 0 0 0 0 0 −40.96 −5.12 0 2 0 1 0 1 1 1 0 1 0 1 2 1 1 0 4 0 −37.51 −2.34 0 0 0 10 0 10 0 0 0 0 0 0 0 0 0 0 0 0 −93.71 −4.69 1 2 3 1 0 1 1 0 0 2 1 1 0 2 0 2 2 1 −44.09 −2.20 4 0 0 0 0 4 2 2 0 0 0 0 0 0 0 0 0 0 −51.94 −4.33 0 0 0 1 2 0 3 0 1 0 0 1 0 1 2 1 0 0 −53.71 −4.48 0 0 0 0 0 0 0 0 0 4 0 0 0 4 0 0 0 0 −11.15 −1.39 0 1 0 0 0 2 1 3 1 2 0 3 0 1 0 0 0 2 −50.47 −3.15 0 0 8 0 0 0 0 8 0 0 0 0 0 0 0 0 0 0 −57.67 −3.60 2 2 0 1 2 1 0 0 0 1 1 0 1 2 0 0 1 2 −29.53 −1.85 14 7 4 0 2 5 13 8 8 7 5 10 5 9 11 7 7 6 −378.64 −2.96 0 0 0 0 0 9 0 0 0 0 0 0 0 0 0 9 0 0 −74.44 −4.14 0 1 0 0 0 0 0 0 7 0 0 0 0 0 0 0 0 0 −6.89 −0.86 0 0 0 0 0 0 9 0 0 0 9 0 0 0 0 0 0 0 −46.17 0 0 2 1 3 1 0 2 1 1 0 0 0 1 1 1 0 2 −59.94 −3.75 0 0 0 0 7 0 1 0 0 0 0 0 0 0 0 0 0 0 −33.89 −4.24 1 0 3 2 1 1 1 0 1 1 0 0 0 0 0 0 1 0 −29.74 −2.48 0 1 0 0 0 0 0 0 0 6 0 0 0 0 1 0 0 0 −0.06 −0.01 7 0 0 0 0 7 0 1 0 0 0 0 0 0 0 1 3 0 −74.36 −3.91 1 0 1 2 2 1 1 0 1 0 1 1 2 2 1 1 1 2 −54.96 −2.75 0 8 0 0 0 0 0 0 0 0 0 0 0 0 0 8 0 0 −22.66 −1.42 0 0 0 0 2 1 1 1 0 1 0 2 1 1 0 0 1 1 −31.12 −2.59 0 0 1 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 −5.95 −1.98 0 0 0 0 0 0 0 0 0 0 9 0 0 0 0 0 9 0 −47.85 −2.66 0 0 2 0 1 2 0 0 0 1 2 1 2 1 1 2 1 0 −52.83 −3.30 7 0 0 0 0 0 9 4 0 0 0 0 0 0 0 0 0 0 −60.71 −3.04 0 2 2 0 2 0 1 2 5 2 0 0 0 0 1 0 0 1 −49.48 −2.75 1 2 1 0 1 0 2 1 0 0 0 0 2 1 0 0 1 0 −40.21 −3.35 0 0 16 0 0 0 8 0 8 0 0 8 0 8 0 0 16 0 −185.75 −2.90 1 2 0 2 1 1 1 3 0 2 0 1 2 1 2 0 0 1 −61.69 −3.08 0 0 0 0 0 8 0 0 0 0 0 0 0 0 0 0 8 0 −59.34 −3.71 0 2 0 0 1 0 0 1 2 1 1 0 2 0 1 1 3 1 −42.74 −2.67 1 0 0 1 0 0 1 3 1 1 2 2 0 0 1 0 1 2 −37.49 −2.34 3 0 1 0 0 1 0 0 0 1 0 1 1 0 2 2 4 0 −39.57 −2.47 0 2 0 0 1 2 2 0 0 1 0 2 0 1 0 0 0 1 −33.80 −2.82 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 15 0 −29.37 −1.84 0 0 0 0 8 0 0 0 0 8 0 0 0 0 0 0 0 0 −39.13 −2.45 1 2 0 2 0 0 1 1 1 0 0 2 0 1 0 2 1 2 −45.19 −2.82 0 0 16 0 16 0 0 0 0 0 0 0 0 16 0 0 0 16 −170.52 −2.66 0 8 8 8 16 8 24 8 0 8 16 8 0 8 8 0 0 0 −483.13 −3.77 13 0 0 0 0 7 1 1 0 0 0 0 0 0 0 0 0 0 −88.00 −4.00 0 1 0 0 2 4 0 0 0 1 2 1 1 3 0 1 0 0 −56.86 −3.55 9 0 0 0 0 0 0 0 0 0 0 0 0 0 9 0 0 0 −38.97 −2.17 0 0 0 2 1 0 1 0 1 1 2 0 1 0 1 1 0 1 −33.08 −2.76 0 1 0 1 0 1 1 0 0 0 0 1 0 0 1 0 0 2 −17.18 −2.15 0 0 1 1 3 1 0 0 2 0 3 1 1 0 1 1 1 2 −45.47 −2.53 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 9 0 0 −45.55 −2.53 0 0 0 0 1 1 0 0 0 0 0 0 0 2 2 0 0 2 −12.91 −1.61 0 2 0 0 1 1 1 1 1 1 1 3 0 0 1 1 1 1 −44.00 −2.75 9 8 9 8 7 10 8 8 4 6 2 7 8 6 5 10 8 5 −370.09 −2.89 0 0 9 0 0 0 0 0 0 0 0 9 0 0 0 0 0 0 −11.95 −0.66 1 0 1 1 2 0 0 0 1 0 0 0 0 0 0 1 1 0 −24.14 −3.02 0 0 0 0 18 0 0 0 0 0 0 0 0 0 0 0 0 2 −83.57 −4.18…”

“…In other words, GNN with certain architecture can be interpreted as the physics-based model of the corresponding IPs. In this paper, we propose a NNIP form that can be considered a superset of MEAM/Tersoff potentials while mimicking electronic total-energy relaxation [16] in a local orbital (tight-binding) basis [18][19][20], named the tensor embedded atom network (TeaNet). In section II, we modify the architecture of GNN with new components (edge-associated in addition to node-associated variables) that fully represent the corresponding physics-based IP.…”

“…In section II, we modify the architecture of GNN with new components (edge-associated in addition to node-associated variables) that fully represent the corresponding physics-based IP. Rank-2 tensors as well as vector variables are introduced in the network and the model can naturally represent propagation of orientation-dependent Hamiltonian information, that is natural in local orbital basis calculations [18][19][20]. We have also adopted ResNet architecture and recurrent GNN initialization to accelerate computations.…”