Accurate vibrational spectra via molecular tailoring approach: A case study of water clusters at MP2 level

Abstract

In spite of the recent advents in parallel algorithms and computer hardware, high-level calculation of vibrational spectra of large molecules is still an uphill task. To overcome this, significant effort has been devoted to the development of new algorithms based on fragmentation methods. The present work provides the details of an efficient and accurate procedure for computing the vibrational spectra of large clusters employing molecular tailoring approach (MTA). The errors in the Hessian matrix elements and dipole derivatives arising due to the approximation nature of MTA are reduced by grafting the corrections from a smaller basis set. The algorithm has been tested out for obtaining vibrational spectra of neutral and charged water clusters at Møller-Plesset second order level of theory, and benchmarking them against the respective full calculation (FC) and/or experimental results. For (H2O)16 clusters, the estimated vibrational frequencies are found to differ by a maximum of 2 cm(-1) with reference to the corresponding FC values. Unlike the FC, the MTA-based calculations including grafting procedure can be performed on a limited hardware, yet take a fraction of the FC time. The present methodology, thus, opens a possibility of the accurate estimation of the vibrational spectra of large molecular systems, which is otherwise impossible or formidable.

Molecular Tailoring Approach for the Estimation of Intramolecular Hydrogen Bond Energy

“…The molecular tailoring approach (MTA) is a fragmentation-based technique developed by Gadre et al [ 68 , 69 , 70 , 71 , 72 , 73 , 74 , 75 , 76 , 77 , 78 , 79 , 80 , 81 , 82 ]. Within MTA, a spatially extended molecular system under consideration is partitioned into a set of overlapping fragments (called the “main fragments”) on which ab initio calculations for one-electron property or the energy are carried out.…”

“…Recently, we have widened the application of the MTA-based method for the estimation of individual O-H•••O HB energy and their cooperativity contribution in water clusters [88]. Many works reported in the literature focus on the global minimum structure of water clusters of variable sizes [73,75,76,[100][101][102][103]. In spite of a large number of studies, the nature of water clusters at the molecular level is not fully understood.…”

“…Thus, it was felt necessary to come up with a direct theoretical method for a reliable estimation of IHB energy. Deshmukh and Gadre [ 66 , 67 ] proposed such a procedure, based on the in-house developed molecular tailoring approach (MTA) for the ab initio treatment of large molecular systems [ 68 , 69 , 70 , 71 , 72 , 73 , 74 , 75 , 76 , 77 , 78 , 79 , 80 , 81 , 82 ]. The MTA currently enables the calculation of one-electron properties, geometry optimization, and the calculation of vibrational infrared and Raman spectra of large molecules/clusters using DFT or correlated methods.…”

“…This error can be further reduced using the so-called grafting procedure embedded in the current version of MTA [77,78].Gadre et al first proposed the MTA methodology for estimating the electrostatic properties of large, closed-shell molecules[68]. However, in the last 24 years, the scope of MTA was extended for the estimation of molecular energy[69], geometry optimization[70,71], the estimation of the Hessian matrix[72], the computation of vibrational infrared[73] and Raman spectra[74], and binding energies of large molecular clusters and complexes[75][76][77][78][79][80][81][82]. Since the fragment computations are independent of each other, MTA has the advantage that the energy computation of the parent molecule is intrinsically parallel.…”

Simulations of Chemical Reactions with the Frozen Domain Formulation of the Fragment Molecular Orbital Method

“…(a) linear scaling methods, 7-10 (b) various partitioning schemes, involving subsystems or fragments [11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29] and (c) model approaches, often of semiempirical nature. 30,31 Despite very significant progress in both method development and computation hardware efficiency, an application of fully ab initio methods to the study of chemical reactions of large molecular systems is still difficult, as it requires large computational resources as well as necessary methodological and software development to evaluate the matrix of second derivatives (the Hessian), do transition state search, map a reaction path, or perform molecular dynamics (MD) simulations.…”