Thomas

Population dynamics for a generalized three‐state Jahn‐Teller system: Quantum‐mechanical (full curves) versus quantum‐classical propagations by means of fewest switches surface hopping (dashed curves) and single switch surface hopping (dash‐dotted curves).

WavePacket is an open‐source program package for numerical simulations in quantum dynamics. Building on the previous Part I (Schmidt and Lorenz, Comput. Phys. Commun. 2017, 213, 223] and Part II (Schmidt and Hartmann, Comput. Phys. Commun. 2018, 228, 229] which dealt with quantum dynamics of closed and open systems, respectively, the present Part III adds fully classical and mixed quantum‐classical propagation techniques to WavePacket. There classical phase‐space densities are sampled by trajectories which follow (diabatic or adiabatic) potential energy surfaces. In the vicinity of (genuine or avoided) intersections of those surfaces, trajectories may switch between them. To model these transitions, two classes of stochastic algorithms have been implemented: (1) Tully's fewest switches surface hopping and (2) Landau–Zener‐based single switch surface hopping. The latter one offers the advantage of being based on adiabatic energy gaps only, thus not requiring nonadiabatic coupling information any more. The present work describes the MATLAB version of WavePacket 6.1.0, which is essentially an object‐oriented rewrite of previous versions, allowing to perform fully classical, quantum‐classical and quantum‐mechanical simulations on an equal footing, that is, for the same physical system described by the same WavePacket input. The software package is hosted and further developed at the Sourceforge platform, where also extensive Wiki‐documentation as well as numerous worked‐out demonstration examples with animated graphics are available. © 2019 Wiley Periodicals, Inc.

The Journal of Chemical Physics, Volume 151, Issue 6, August 2019.
The absorption spectrum of the vibronically allowed S1(1A2) ← S0(1A1) transition of formaldehyde is computed by combining multiplicative neural network (NN) potential surface fits, based on multireference electronic structure data, with the two-layer Gaussian-based multiconfiguration time-dependent Hartree (2L-GMCTDH) method. The NN potential surface fit avoids the local harmonic approximation for the evaluation of the potential energy matrix elements. Importantly, the NN surface can be constructed so as to be physically well-behaved outside the domain spanned by the ab initio data points. A comparison with experimental results shows spectroscopic accuracy of the converged surface and 2L-GMCTDH quantum dynamics.

The Journal of Chemical Physics, Volume 150, Issue 24, June 2019.
On the basis of a new extensive database constructed for the purpose, we assess various Machine Learning (ML) algorithms to predict energies in the framework of potential energy surface (PES) construction and discuss black box character, robustness, and efficiency. The database for training ML algorithms in energy predictions based on the molecular structure contains SCF, RI-MP2, RI-MP2-F12, and CCSD(F12*)(T) data for around 10.5 × 106 configurations of 15 small molecules. The electronic energies as function of molecular structure are computed from both static and iteratively refined grids in the context of automized PES construction for anharmonic vibrational computations within the n-mode expansion. We explore the performance of a range of algorithms including Gaussian Process Regression (GPR), Kernel Ridge Regression, Support Vector Regression, and Neural Networks (NNs). We also explore methods related to GPR such as sparse Gaussian Process Regression, Gaussian process Markov Chains, and Sparse Gaussian Process Markov Chains. For NNs, we report some explorations of architecture, activation functions, and numerical settings. Different delta-learning strategies are considered, and the use of delta learning targeting CCSD(F12*)(T) predictions using, for example, RI-MP2 combined with machine learned CCSD(F12*)(T)-RI-MP2 differences is found to be an attractive option.

Revised versions of our pob‐TZVP and pob‐DZVP basis sets have been derived for the elements HBr for periodic quantum‐chemical solid state calculations, denoted as pob‐TZVP‐rev2 and pob‐DZVP‐rev2. To reduce the basis set superposition error, the authors took into account the counterpoise energy of hydride dimers as an additional parameter in the basis set optimization. The overall performance of the rev2 basis sets is significantly improved compared to the original pob basis sets.

Revised versions of our published pob‐TZVP [Peintinger, M. F.; Oliveira, D. V. and Bredow, T., J. Comput. Chem., 2013, 34 (6), 451–459.] and unpublished pob‐DZVP basis sets, denoted as pob‐TZVP‐rev2 and pob‐DZVP‐rev2, have been derived for the elements HBr. It was observed that the pob basis sets suffer from the basis set superposition error (BSSE). In order to reduce this effect, we took into account the counterpoise energy of hydride dimers as an additional parameter in the basis set optimization. The overall performance, portability, and SCF stability of the resulting rev2 basis sets are significantly improved compared to the original pob basis sets. © 2019 Wiley Periodicals, Inc.

Density functional theory and flow‐reactor experiments are used to study the adsorption of SO₂ on gas‐phase small gold cluster anions and their hydroxide derivatives. All clusters adsorb up to four SO₂ molecules, and the gold‐hydroxide clusters are more active for SO₂ adsorption.

Abstract

The binding of SO₂ on gas‐phase gold cluster anions, Au_N ⁻, and their hydroxide counterparts, Au_NOH⁻, have been studied using density functional theory combined with flow reactor/time‐of‐flight mass spectrometry techniques. SO₂ is adsorbed on all of the Au_N ⁻ and Au_NOH⁻ clusters (N = 1‐8) and the hydroxide clusters are more active than the bare anionic clusters. Successive additions of SO₂ molecules (up to four) have been analyzed. In all cases, anionic clusters are shown to bind multiple SO₂ molecules. Theoretical analyses are in agreement with the experimental results, showing that the addition of more than one molecule is thermodynamically favorable. Larger clusters do not necessarily absorb more molecules, as different SO₂ binding motifs on these clusters are present. These results provide important insight for the potential use of these anionic clusters as SO₂ hunters.

autoCAS program for fully automated multiconfigurational calculations is presented. The program requires a minimal input, comparable to that of a density‐functional theory calculation and automatically selects an appropriate active space based on orbital entanglement entropies calculated from a partially converged density‐matrix renormalization group calculation. Hence, autoCAS turns multiconfigurational calculations into a black‐box method. It is implemented as a graphical user interface and is part of the SCINE project for chemical interaction networks.

We present our implementation autoCAS for fully automated multiconfigurational calculations, which we also make available free of charge on our webpages. The graphical user interface of autoCAS connects a general electronic structure program with a density‐matrix renormalization group program to carry out our recently introduced automated active space selection protocol for multiconfigurational calculations (Stein and Reiher, J. Chem. Theory Comput., 2016, 12, 1760). Next to this active space selection, autoCAS carries out several steps of multiconfigurational calculations so that only a minimal input is required to start them, comparable to that of a standard Kohn–Sham density‐functional theory calculation, so that black‐box multiconfigurational calculations become feasible. Furthermore, we introduce a new extension to the selection algorithm that facilitates automated selections for molecules with large valence orbital spaces consisting of several hundred orbitals. © 2019 Wiley Periodicals, Inc.

CRYSPLOT is a publicly accessible web‐based tool (http://crysplot.crystalsolutions.eu) to visualize physico‐chemical properties of periodic systems (i.e., crystals, surfaces and polymers) as computed with the CRYSTAL code. It has been designed as a very intuitive graphical tool, a low entry‐level interface, to all types of users, from beginners to experts, and purposes, from education to research, as shown with selected examples.

CRYSPLOT is a web‐oriented tool (http://crysplot.crystalsolutions.eu) to visualize computed properties of periodic systems, in particular, as computed with the CRYSTAL code. Along with plotting, CRYSPLOT also permits the modification and customization of plots to meet the standards required for scientific graphics. CRYSPLOT has been designed with advanced and freely available graphical Javascript libraries as Plotly. The programming language used is Javascript. The code parses the input files, reads the data, and organizes them into objects ready to be plotted with the plotly.js library. It is modular and flexible so that it is very simple to add other input data formats. The new graphical tool is presented in details along with selected applications on metal–organic frameworks to show some of its capabilities. © 2019 Wiley Periodicals, Inc.

The Journal of Chemical Physics, Volume 150, Issue 14, April 2019.
The near-equilibrium potential energy surface (PES) of the [math] state of SO2 is developed from explicitly correlated spin-unrestricted coupled cluster calculations with single, double, and perturbative triple excitations with an augmented triple-zeta correlation-consistent basis set. The lowest-lying ro-vibrational energy levels of several sulfur isotopologues have been determined using this PES. It is shown that the new ab initio PES provides a much better description of the low-lying vibrational states than a previous PES determined at the multi-reference configuration interaction level. In particular, the theory-experiment agreement for the three lowest-lying vibrational transitions is within 1–3 cm−1.

Zwei Mathematiker finden den schnellsten Multiplikations-Algorithmus - und beweisen, dass es keinen schnelleren geben kann.

A perturbation theory‐based algorithm for the iterative orbital update in complete active space self‐consistent‐field (CASSCF) calculations is presented. It is based on single‐excitation amplitudes obtained from a perturbative configuration interaction calculation using the Dyall Hamiltonian as a zeroth‐order Hamiltonian. These amplitudes are used for the construction of the exponential of an anti‐Hermitian matrix which can be used for the orbital update. Combined with DIIS, this approach leads to rapid convergence of the CASSCF iteration procedure.

A perturbation theory‐based algorithm for the iterative orbital update in complete active space self‐consistent‐field (CASSCF) calculations is presented. Following Angeli et al. (J. Chem. Phys. 2002, 117, 10525), the first‐order contribution of singly excited configurations to the CASSCF wave function is evaluated using the Dyall Hamiltonian for the determination of a zeroth‐order Hamiltonian. These authors employ an iterative diagonalization of the first‐order density matrix including the first‐order correction arising from single excitations, whereas the present approach uses the single‐excitation amplitudes directly for the construction of the exponential of an anti‐Hermitian matrix resulting in a unitary matrix which can be used for the orbital update. At convergence, the single‐excitation amplitudes vanish as a consequence of the generalized Brillouin's theorem. It is shown that this approach in combination with direct inversion of the iterative subspace (DIIS) leads to very rapid convergence of the CASSCF iteration procedure. © 2019 Wiley Periodicals, Inc.