Update 'paper_UAR.tex'

This commit is contained in:
linjie 2020-04-07 19:36:31 +02:00
parent b0ea38bae4
commit 87a601a6de
1 changed files with 1 additions and 1 deletions

View File

@ -154,7 +154,7 @@ Some studies have already been done about the dynamics collision simulation.\cit
\section{Computational Methods} \label{Comput_meth}
\textbf{Exploration of the PES}
{\bf Rque general : their was plenty of ab initio calculation to benchmark the proton transfer etc .. it would be good to mention them and may be to put some data, may be some of them in the manuscirpt and others in the supplementary files}
DFTB is approximated from DFT scheme whose efficiency relies on the use of parameterized integrals with a much lower computational cost. {\bf \cite{Elstner2014,Elstner1998,dftb1,dftb2}} The DFTB approach has been particularly well studied and it has already proven its efficiency to describe chemical processes. \cite{Kruger2005} In this work, we used the second-order version of DFTB, Self Consistent Charge DFTB, with the mio-set for the Slater-Koster tables of integrals. \cite{Elstner1998} To improve the intermolecular interaction, the class IV/charge model 3 (CM3) charges instead of the original Mulliken charges as well as the empirical terms were used to describe dispersion interactions. \cite{Rapacioli2009} For the parameterization of CM3 charges, the bond parameter D$_{OH}$ = 0.129 proposed by Simon and co-workers was applied, \blue{DNH = 0.120 tested by ourselves, (part of this work has been published[]){\bf mettre une ref}} while all other bond parameter values were set to be 0.000, which corresponds to a Mulliken evaluation of the charges.\cite{Simon2012, Simon2013} {\bf In a QM/MM scheme, the Argon atom is treated as a polarizable MM particule interacting with the Uracil-water cluster treated at the DFTB level. Details about this model can be found in the original paper \cite{bzar}. } All the SCC-DFTB calculations in the present work were carried out with the deMonNano code. \cite{demonnanoCode}
DFTB is approximated from DFT scheme whose efficiency relies on the use of parameterized integrals with a much lower computational cost. {\bf \cite{Elstner2014,Elstner1998,dftb1,dftb2}} The DFTB approach has been particularly well studied and it has already proven its efficiency to describe chemical processes. \cite{Kruger2005} In this work, we used the second-order version of DFTB, Self Consistent Charge DFTB, with the mio-set for the Slater-Koster tables of integrals. \cite{Elstner1998} To improve the intermolecular interaction, the class IV/charge model 3 (CM3) charges instead of the original Mulliken charges as well as the empirical terms were used to describe dispersion interactions. \cite{Rapacioli2009} For the parameterization of CM3 charges, the bond parameter D$_{OH}$ = 0.129 proposed by Simon and co-workers was applied, \blue{DNH = 0.140 tested by ourselves, (part of this work has been published[]){\bf mettre une ref}} while all other bond parameter values were set to be 0.000, which corresponds to a Mulliken evaluation of the charges.\cite{Simon2012, Simon2013} {\bf In a QM/MM scheme, the Argon atom is treated as a polarizable MM particule interacting with the Uracil-water cluster treated at the DFTB level. Details about this model can be found in the original paper \cite{bzar}. } All the SCC-DFTB calculations in the present work were carried out with the deMonNano code. \cite{demonnanoCode}
All the energy minima for (H$_2$O)$_{n=3-7}$UH$^+$, have already been obtained in the previous study.\cite{Braud2019} In this present work, we calculated the lowest-energy isomers of (H$_2$O)$_{12}$UH$^+$ cluster. To obtain them, the same two-step theoretical method with the one used for the calculation of lowest energy isomers of clusters (H$_2$O)$_{n=1-7}$UH$^+$ was applied.\cite{Braud2019} Firstly, the potential energy surface (PES) of (H$_2$O)$_{12}$UH$^+$ was roughly explored using the parallel temperature molecular dynamics{\bf move here refs to PTMD} (PTMD) simulations in combination with SCC-DFTB description of the energies and gradients.\cite{Sugita1999, Earl2005} In the PTMD algorithm, 40 replicas with temperatures going linearly from 50 to 350 K were carried out. All the trajectories were 4 ns long, and the integration time step was 0.5 fs. A Nos{\bf \'e} é-Hoover chain of five thermostats with frequencies of 800 cm$^{-1}$ was used to obtain an exploration in the canonical emsemble. \cite{Nose1984, Hoover1985} To avoid any spurious influence of the initial geometry on the PES exploration, three distinct PTMD simulations were carried out. In the three series, a distinct initial proton location was set: on the uracil in two cases and on the water cluster in another one. In the former cases, the u178 and u138 UH$^+$ isomers were used as initial geometries which was named by Pedersen \cite{Pedersen2014} {\bf -> we also used two isomers from Pedersen, reported in this work as u178 and u138 UH$^+$}. 600 geometries per temperature were linearly selected along each PTMD simulation for subsequent geometry optimization leading to 72000 structures optimized at SCC-DFTB level. These structures were sorted in ascending energy order. Secondly, 29 isomers were selected from the 72000 optimized structures at SCC-DFTB level and were optimized at a high accurate MP2/Def2TZVP level, which is a tight criteria for geometry convergence {\bf I don't understand what is the tight criteria ? the convergencey threshold ? } and an ultrafine grid for the numerical integration. \cite{Weigend2005, Weigend2006} From the MP2/Def2TZVP calculation, the lowest energy isomers of cluster (H$_2$O)$_{12}$UH$^+$ were obtained. All MP2 calculations were carried out with the Gaussian 09 package.\cite{GaussianCode}