Efficient generation of sum-of-products representations of high-dimensional potential energy surfaces based on multimode expansions.

The transformation of multi-dimensional potential energy surfaces (PESs) from a grid-based multimode representation to an analytical one is a standard procedure in quantum chemical programs. Within the framework of linear least squares fitting, a simple and highly efficient algorithm is presented, which relies on a direct product representation of the PES and a repeated use of Kronecker products. It shows the same scalings in computational cost and memory requirements as the potfit approach. In comparison to customary linear least squares fitting algorithms, this corresponds to a speed-up and memory saving by several orders of magnitude. Different fitting bases are tested, namely, polynomials, B-splines, and distributed Gaussians. Benchmark calculations are provided for the PESs of a set of small molecules.

[1]  Joel M. Bowman,et al.  Vibrational self-consistent field method for many-mode systems: A new approach and application to the vibrations of CO adsorbed on Cu(100) , 1997 .

[2]  G. Rauhut.  Efficient calculation of potential energy surfaces for the generation of vibrational wave functions. , 2004, The Journal of chemical physics.

[3]  Weitao Yang,et al.  The collocation method for bound solutions of the Schrödinger equation , 1988 .

[4]  V. Szalay.  The generalized discrete variable representation. An optimal design , 1996 .

[5]  T. Carrington,et al.  A multi-dimensional Smolyak collocation method in curvilinear coordinates for computing vibrational spectra. , 2015, The Journal of chemical physics.

[6]  J. Light,et al.  On distributed Gaussian bases for simple model multidimensional vibrational problems , 1986 .

[7]  T. Carrington,et al.  Solving the Schroedinger equation using Smolyak interpolants. , 2013, The Journal of chemical physics.

[8]  S. Carter,et al.  Pruning the Hamiltonian Matrix in MULTIMODE: Test for C2H4 and Application to CH3NO2 Using a New Ab Initio Potential Energy Surface. , 2015, The journal of physical chemistry. A.

[9]  C. Loan.  The ubiquitous Kronecker product , 2000 .

[10]  Eduard Matito,et al.  Using Electronic Energy Derivative Information in Automated Potential Energy Surface Construction for Vibrational Calculations. , 2010, Journal of chemical theory and computation.

[11]  Joel M Bowman,et al.  Variational calculations of vibrational energies and IR spectra of trans- and cis-HOCO using new ab initio potential energy and dipole moment surfaces. , 2013, The journal of physical chemistry. A.

[12]  B. Braams,et al.  Ab-Initio-Based Potential Energy Surfaces for Complex Molecules and Molecular Complexes , 2010 .

[13]  V. Szalay.  Discrete variable representations of differential operators , 1993 .

[14]  Hans-Dieter Meyer,et al.  Product representation of potential energy surfaces. II , 1996 .

[15]  D. Truhlar,et al.  Gradient-based multiconfiguration Shepard interpolation for generating potential energy surfaces for polyatomic reactions. , 2010, The Journal of chemical physics.

[16]  Yuan Lei,et al.  Iterative method for the least squares problem of a matrix equation with tridiagonal matrix constraint , 2012 .

[17]  D. Fausett,et al.  Large Least Squares Problems Involving Kronecker Products , 1994, SIAM J. Matrix Anal. Appl..

[18]  Nicholas C. Handy,et al.  On the representation of potential energy surfaces of polyatomic molecules in normal coordinates , 2002 .

[19]  S. Carter,et al.  MULTIMODE: A code to calculate rovibrational energies of polyatomic molecules , 2003 .

[20]  Zhen Xie,et al.  Permutationally Invariant Polynomial Basis for Molecular Energy Surface Fitting via Monomial Symmetrization. , 2010, Journal of chemical theory and computation.

[21]  H. Meyer,et al.  The multigrid POTFIT (MGPF) method: grid representations of potentials for quantum dynamics of large systems. , 2013, The Journal of chemical physics.

[22]  Jacob Kongsted,et al.  Automatic generation of force fields and property surfaces for use in variational vibrational calculations of anharmonic vibrational energies and zero-point vibrational averaged properties. , 2006, The Journal of chemical physics.

[23]  Joel M. Bowman,et al.  Permutationally invariant potential energy surfaces in high dimensionality , 2009 .

[24]  Donald L. Thompson,et al.  Ab initio dynamics: HeH+ + H2 → He + H3+ (C2ν) classical trajectories using a quantum mechanical potential‐energy surface , 1973 .

[25]  Lawrence B. Harding,et al.  Ab initio calculations of electronic and vibrational energies of HCO and HOC , 1986 .

[26]  Sergei Manzhos,et al.  A random-sampling high dimensional model representation neural network for building potential energy surfaces. , 2006, The Journal of chemical physics.

[27]  Frank Otto,et al.  Multi-layer Potfit: an accurate potential representation for efficient high-dimensional quantum dynamics. , 2013, The Journal of chemical physics.

[28]  Willi-Hans Steeb,et al.  Matrix Calculus and the Kronecker Product with Applications and C++ Programs , 1997 .

[29]  A. Varandas,et al.  Roadmap to spline-fitting potentials in high dimensions , 2013, Journal of Mathematical Chemistry.

[30]  Guntram Rauhut,et al.  Modeling of high-order many-mode terms in the expansion of multidimensional potential energy surfaces: application to vibrational spectra. , 2009, The Journal of chemical physics.

[31]  J. Bowman,et al.  Infrared identification of the Criegee intermediates syn- and anti-CH3CHOO, and their distinct conformation-dependent reactivity , 2015, Nature Communications.

[32]  M. Minkoff,et al.  Interpolating moving least-squares methods for fitting potential energy surfaces: using classical trajectories to explore configuration space. , 2009, The Journal of chemical physics.

[33]  L. Radom,et al.  Proton-transport catalysis and proton-abstraction reactions: An ab initio dynamical study of X+HOC+ and XH++CO (X=Ne, Ar, and Kr) , 2000 .

[34]  Sergei Manzhos,et al.  Using neural networks to represent potential surfaces as sums of products. , 2006, The Journal of chemical physics.

[35]  T. Carrington,et al.  Using multi-dimensional Smolyak interpolation to make a sum-of-products potential. , 2015, The Journal of chemical physics.

[36]  B. Braams,et al.  On using low-order Hermite interpolation in `direct dynamics' calculations of vibrational energies using the code `MULTIMODE' , 2001 .

[37]  H. Rabitz,et al.  Reproducing kernel Hilbert space interpolation methods as a paradigm of high dimensional model representations: Application to multidimensional potential energy surface construction , 2003 .

[38]  T. Crawford,et al.  Communication: Spectroscopic consequences of proton delocalization in OCHCO⁺. , 2015, The Journal of chemical physics.

[39]  Jun Li,et al.  High-Level, First-Principles, Full-Dimensional Quantum Calculation of the Ro-vibrational Spectrum of the Simplest Criegee Intermediate (CH2OO). , 2014, The journal of physical chemistry letters.

[40]  T. Carrington,et al.  A general discrete variable method to calculate vibrational energy levels of three‐ and four‐atom molecules , 1993 .

[41]  Joel M. Bowman,et al.  High-dimensional ab initio potential energy surfaces for reaction dynamics calculations. , 2011, Physical chemistry chemical physics : PCCP.

[42]  J. Bowman,et al.  Full-dimensional (15-dimensional) ab initio analytical potential energy surface for the H7+ cluster. , 2012, The Journal of chemical physics.

[43]  Gabriel G. Balint-Kurti,et al.  The Fourier grid Hamiltonian method for bound state eigenvalues and eigenfunctions , 1989 .

[44]  J. Light,et al.  Generalized discrete variable approximation in quantum mechanics , 1985 .

[45]  A. Barrlund.  Efficient Solution of Constrained Least Squares Problems with Kronecker Product Structure , 1998 .

[46]  So Hirata,et al.  Multiresolution potential energy surfaces for vibrational state calculations , 2007 .

[47]  D. Fausett,et al.  Improved parallel QR method for large least squares problems involving Kronecker products , 1997 .

[48]  David R. Forsey,et al.  Surface fitting with hierarchical splines , 1995, TOGS.

[49]  Michael A. Collins,et al.  Molecular potential-energy surfaces for chemical reaction dynamics , 2002 .