Title: Permutation invariant polynomial neural network approach to fitting potential energy surfaces. IV. Coupled diabatic potential energy matrices

Journal Article · · Journal of Chemical Physics
DOI: https://doi.org/10.1063/1.5054310 · OSTI ID:1612363
 [1]; ORCiD logo [2]; ORCiD logo [3]; ORCiD logo [4]
  1. Univ. of New Mexico, Albuquerque, NM (United States); DOE/OSTI
  2. Johns Hopkins Univ., Baltimore, MD (United States); Stanford Univ., CA (United States)
  3. Johns Hopkins Univ., Baltimore, MD (United States)
  4. Univ. of New Mexico, Albuquerque, NM (United States)
+ Show Author Affiliations

A machine learning method is proposed for representing the elements of diabatic potential energy matrices (PEMs) with high fidelity. This is an extension of the so-called permutation invariant polynomial-neural network (PIP-NN) method for representing adiabatic potential energy surfaces. While for one-dimensional irreducible representations the diagonal elements of a diabatic PEM are invariant under exchange of identical nuclei in a molecular system, the off-diagonal elements require special symmetry consideration, particularly in the presence of a conical intersection. A multiplicative factor is introduced to take into consideration the particular symmetry properties while maintaining the PIP-NN framework. Here, we demonstrate here that the extended PIP-NN approach is accurate in representing diabatic PEMs, as evidenced by small fitting errors and by the reproduction of absorption spectra and product branching ratios in both H2O$$(\tilde{X} / \tilde{B})$$ and NH3$$(\tilde{X} $$/Ã) non-adiabatic photodissociation.

Research Organization:
Univ. of New Mexico, Albuquerque, NM (United States)
Sponsoring Organization:
USDOE; USDOE Office of Science (SC)
Grant/Contract Number:
SC0015997
OSTI ID:
1612363
Journal Information:
Journal of Chemical Physics, Journal Name: Journal of Chemical Physics Journal Issue: 14 Vol. 149; ISSN 0021-9606
Publisher:
American Institute of Physics (AIP)Copyright Statement
Country of Publication:
United States
Language:
English

References (83)

Beyond Born-Oppenheimer book January 2006
Multimode Molecular Dynamics Beyond the Born-Oppenheimer Approximation book January 2007
Neural network-based approaches for building high dimensional and quantum dynamics-friendly potential energy surfaces journal October 2014
Molecular potential-energy surfaces for chemical reaction dynamics journal December 2002
Improved direct diabatization and coupled potential energy surfaces for the photodissociation of ammonia journal March 2007
Adiabatic and diabatic representations for atom-diatom collisions: Treatment of the three-dimensional case journal June 1976
The Requisite Electronic Structure Theory To Describe Photoexcited Nonadiabatic Dynamics: Nonadiabatic Derivative Couplings and Diabatic Electronic Couplings journal May 2015
Geometric Phase Effects in Nonadiabatic Dynamics near Conical Intersections journal June 2017
Ab Initio Nonadiabatic Quantum Molecular Dynamics journal February 2018
Ab Initio Potential Energy Surfaces and Quantum Dynamics for Polyatomic Bimolecular Reactions journal March 2018
From ab Initio Potential Energy Surfaces to State-Resolved Reactivities: X + H 2 O ↔ HX + OH [X = F, Cl, and O( 3 P)] Reactions journal May 2015
Recent Progress in Surface Hopping: 2011–2015 journal May 2016
Nonadiabatic Quantum Chemistry—Past, Present, and Future journal November 2011
Permutationally Invariant Polynomial Basis for Molecular Energy Surface Fitting via Monomial Symmetrization journal November 2009
Nonadiabatic Tunneling in Photodissociation of Phenol journal June 2016
Signatures of a Conical Intersection in Adiabatic Dissociation on the Ground Electronic State journal February 2018
Geometric Phase Effects in Chemical Reaction Dynamics and Molecular Spectra journal September 2003
Construction of Diabatic Hamiltonian Matrix from ab Initio Calculated Molecular Symmetry Adapted Nonadiabatic Coupling Terms and Nuclear Dynamics for the Excited States of Na 3 Cluster journal April 2013
Full-Dimensional Quantum Dynamics of Vibrationally Mediated Photodissociation of NH 3 and ND 3 on Coupled Ab Initio Potential Energy Surfaces: Absorption Spectra and NH 2 (Ã 2 A 1 )/NH 2 (X̃ 2 B 1 ) Branching Ratios journal July 2014
HN 2 ( 2 A ‘) Electronic Manifold. II. Ab Initio Based Double-Sheeted DMBE Potential Energy Surface via a Global Diabatization Angle journal April 2008
Energies and Derivative Couplings in the Vicinity of a Conical Intersection Using Degenerate Perturbation Theory and Analytic Gradient Techniques. 1 journal June 1997
Full-Dimensional Quantum State-to-State Nonadiabatic Dynamics for Photodissociation of Ammonia in its A -Band journal March 2014
Feshbach resonances in the exit channel of the F + CH3OH → HF + CH3O reaction observed using transition-state spectroscopy journal July 2017
Direct observation of forward-scattering oscillations in the H+HD→H2+D reaction journal April 2018
Full-dimensional potentials and state couplings and multidimensional tunneling calculations for the photodissociation of phenol journal January 2014
Accurate nonadiabatic dynamics journal January 2016
Recent advances in quantum scattering calculations on polyatomic bimolecular reactions journal January 2017
The direct calculation of diabatic states based on configurational uniformity journal January 2001
An optimal adiabatic-to-diabatic transformation of the 1 2A′ and 2 2A′ states of H3 journal January 2002
Effects of higher order Jahn-Teller coupling on the nuclear dynamics journal March 2004
Interpolation of diabatic potential energy surfaces journal January 2004
Assigning signs to the electronic nonadiabatic coupling terms:  The {H2,O} system as a case study journal April 2005
Interpolation of diabatic potential-energy surfaces: Quantum dynamics on ab initio surfaces journal October 2005
Interpolation of multidimensional diabatic potential energy matrices journal September 2006
Constructing diabatic states from adiabatic states: Extending generalized Mulliken–Hush to multiple charge centers with Boys localization journal December 2008
Toward eliminating the electronic structure bottleneck in nonadiabatic dynamics on the fly: An algorithm to fit nonlocal, quasidiabatic, coupled electronic state Hamiltonians based on ab initio electronic structure data journal March 2010
Atom-centered symmetry functions for constructing high-dimensional neural network potentials journal February 2011
Communication: State-to-state differential cross sections for H 2 O(B̃) photodissociation journal June 2011
A new method to generate spin-orbit coupled potential energy surfaces: Effective relativistic coupling by asymptotic representation journal January 2012
State-to-state photodissociation dynamics of triatomic molecules: H 2 O in the B band journal January 2012
MCSCF study of the avoided curve crossing of the two lowest 1 Σ + states of LiF journal May 1981
Conditions for the definition of a strictly diabatic electronic basis for molecular systems journal December 1982
Dissociation of NH 3 to NH 2 +H journal June 1987
Potential energy surfaces near intersections journal August 1991
Implications of rotation–inversion–permutation invariance for analytic molecular potential energy surfaces journal November 1993
On the representation of coupled adiabatic potential energy surfaces using quasi-diabatic Hamiltonians: A distributed origins expansion approach journal May 2012
Computational determination of the à state absorption spectrum of NH 3 and of ND 3 using a new quasi-diabatic representation of the X̃ and à states and full six-dimensional quantum dynamics journal June 2012
Quasi-diabatic representations of adiabatic potential energy surfaces coupled by conical intersections including bond breaking: A more general construction procedure and an analysis of the diabatic representation journal December 2012
First principles determination of the NH 2 /ND 2 (Ã,X̃) branching ratios for photodissociation of NH 3 /ND 3 via full-dimensional quantum dynamics based on a new quasi-diabatic representation of coupled ab initio potential energy surfaces journal December 2012
Interpolation of multi-sheeted multi-dimensional potential-energy surfaces via a linear optimization procedure journal June 2013
Permutation invariant polynomial neural network approach to fitting potential energy surfaces journal August 2013
Permutation invariant polynomial neural network approach to fitting potential energy surfaces. II. Four-atom systems journal November 2013
Permutation invariant polynomial neural network approach to fitting potential energy surfaces. III. Molecule-surface interactions journal July 2014
Diabatization based on the dipole and quadrupole: The DQ method journal September 2014
On the elimination of the electronic structure bottleneck in on the fly nonadiabatic dynamics for small to moderate sized (10-15 atom) molecules using fit diabatic representations based solely on ab initio electronic structure data: The photodissociation of phenol journal January 2016
Communication: Fitting potential energy surfaces with fundamental invariant neural network journal August 2016
Perspective: Machine learning potentials for atomistic simulations journal November 2016
A new approach for the development of diabatic potential energy surfaces: Hybrid block-diagonalization and diabatization by ansatz journal November 2016
On the incorporation of the geometric phase in general single potential energy surface dynamics: A removable approximation to ab initio data journal December 2016
Dynamic mapping of conical intersection seams: A general method for incorporating the geometric phase in adiabatic dynamics in polyatomic systems journal July 2017
Neural network based coupled diabatic potential energy surfaces for reactive scattering journal August 2017
Construction of diabatic energy surfaces for LiFH with artificial neural networks journal December 2017
Automated construction of potential energy surfaces journal September 2015
Non-adiabaticity: the importance of conical intersections journal May 2016
A comparative study of methods for describing non-adiabatic coupling: diabatic representation of the 1Sigma +/1Pi HOH and HHO conical intersections journal August 1997
On the description of potential energy surfaces exhibiting conical intersections: a compact representation of the energies and derivative couplings and locally diabatic bases for the HOH and OHH portions of the 11A'-21A' seam of conical intersection in water journal April 1998
Permutationally invariant potential energy surfaces in high dimensionality journal October 2009
Single- and multireference electronic structure calculations for constructing potential energy surfaces journal June 2016
Potential energy surfaces from high fidelity fitting of ab initio points: the permutation invariant polynomial - neural network approach journal July 2016
Neural Networks in Chemical Reaction Dynamics January 2012
Studies of the Jahn-Teller effect .II. The dynamical problem
  • Longuet-Higgins, Hugh Christopher; Öpik, U.; Lecorney Pryce, Maurice Henry
  • Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences, Vol. 244, Issue 1236, p. 1-16 https://doi.org/10.1098/rspa.1958.0022
journal February 1958
Geometric Phase Effects in Dynamics Near Conical Intersections: Symmetry Breaking and Spatial Localization journal November 2013
Generalized Neural-Network Representation of High-Dimensional Potential-Energy Surfaces journal April 2007
The geometric phase in molecular systems journal January 1992
Diabolical conical intersections journal October 1996
Training feedforward networks with the Marquardt algorithm journal January 1994
Beyond Born-Oppenheimer journal January 2008
Extremely short-lived reaction resonances in Cl + HD ( v = 1) → DCl + H due to chemical bond softening journal January 2015
Encoding of vinylidene isomerization in its anion photoelectron spectrum journal October 2017
Conical Intersections: Electronic Structure, Dynamics and Spectroscopy book July 2004
Recent Advances in Quantum Dynamics of Bimolecular Reactions journal May 2016
Permutationally Invariant Potential Energy Surfaces journal April 2018
C ONSTRUCTING M ULTIDIMENSIONAL M OLECULAR P OTENTIAL E NERGY S URFACES FROM A B I NITIO D ATA journal October 1999

Cited By (11)

Machine learning enables long time scale molecular photodynamics simulations journal January 2019
Int-Deep: A Deep Learning Initialized Iterative Method for Nonlinear Problems text January 2019
Representation of coupled adiabatic potential energy surfaces using neural network based quasi-diabatic Hamiltonians: 1,2 2 A′ states of LiFH journal January 2019
Bayesian machine learning for quantum molecular dynamics journal January 2019
Two-state diabatic potential energy surfaces of ClH 2 based on nonadiabatic couplings with neural networks journal January 2019
Machine learning enables long time scale molecular photodynamics simulations journal January 2019
Neural network based quasi-diabatic Hamiltonians with symmetry adaptation and a correct description of conical intersections journal June 2019
Vibronically and spin-orbit coupled diabatic potentials for X( 2 P) + CH 4 → HX + CH 3 reactions: Neural network potentials for X = Cl journal June 2019
Diabatic neural network potentials for accurate vibronic quantum dynamics—The test case of planar NO 3 journal October 2019
Machine learning enables long time scale molecular photodynamics simulations text January 2018
Bayesian machine learning for quantum molecular dynamics text January 2019

Similar Records

Related Subjects

71 CLASSICAL AND QUANTUM MECHANICS
GENERAL PHYSICS
Absorption spectroscopy
Artificial neural networks
Chemistry
Diabatization
Machine learning
Photodissociation
Physics
Potential energy surfaces
Quantum chemical dynamics