Merge branch 'master' of git.irsamc.ups-tlse.fr:loos/Ec
This commit is contained in:
commit
966e02d4a6
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
BIN
Manuscript/Cover_Letter/CNRS_logo.pdf
Normal file
BIN
Manuscript/Cover_Letter/CNRS_logo.pdf
Normal file
Binary file not shown.
40
Manuscript/Cover_Letter/CoverLetter.tex
Normal file
40
Manuscript/Cover_Letter/CoverLetter.tex
Normal file
@ -0,0 +1,40 @@
|
||||
\documentclass[10pt]{letter}
|
||||
\usepackage{UPS_letterhead,color,mhchem,mathpazo,ragged2e}
|
||||
\newcommand{\alert}[1]{\textcolor{red}{#1}}
|
||||
|
||||
\begin{document}
|
||||
|
||||
\begin{letter}%
|
||||
{To the Editors of the Journal of Chemical Physics}
|
||||
|
||||
\opening{Dear Editors,}
|
||||
|
||||
\justifying
|
||||
|
||||
Please find enclosed our manuscript entitled \textit{``Accurate full configuration interaction correlation energy estimates for five- and six-membered ring molecules''},
|
||||
which we would like you to consider as a Regular Article in the \textit{Journal of Chemical Theory and Computation}.
|
||||
|
||||
Thanks to the selected configuration interaction (SCI) algorithm named \textit{Configuration Interaction using a Perturbative Selection made Iteratively} (CIPSI), in the present manuscript, we have been able to produce FCI-quality frozen-core correlation energies for twelve cyclic molecules in the correlation-consistent double-$\zeta$ Dunning basis set (cc-pVDZ).
|
||||
These estimates, which are likely accurate to a few tenths of a millihartree, have been obtained by extrapolating \textit{orbital-optimized} CIPSI energies to the FCI limit.
|
||||
Compared to natural orbitals, we have shown that, by using energetically optimized orbitals, one can reduce the size of the variational space by one order of magnitude for the same variational energy.
|
||||
|
||||
Having at hand these reference FCI energies, we have been able to faithfully benchmark and study the convergence properties of three families of popular electronic structure methods: i) MP2, MP3, MP4, and MP5, ii) CC2, CC3, and CC4, and iii) CCSD, CCSDT, and CCSDTQ.
|
||||
The performance of the ground-state gold standard CCSD(T) as well as the completely renormalized CC model, CR-CC(2,3), are also investigated.
|
||||
|
||||
Because of the novel theoretical developments described in this manuscript and the usefulness of benchmark sets and their corresponding reference data for the electronic structure community, we expect this work to be of interest to a wide audience within the chemistry and physics communities.
|
||||
|
||||
We suggest Cyrus Umrigar, Sandeep Sharma, Nicholas Mayhall, Janus Eriksen, Seiichiro Ten-no, and Piotr Piecuch as potential referees.
|
||||
|
||||
We look forward to hearing from you soon.
|
||||
|
||||
\closing{Sincerely,}
|
||||
|
||||
|
||||
\end{letter}
|
||||
\end{document}
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
70
Manuscript/Cover_Letter/UPS_letterhead.sty
Normal file
70
Manuscript/Cover_Letter/UPS_letterhead.sty
Normal file
@ -0,0 +1,70 @@
|
||||
%ANU etterhead Yves
|
||||
%version 1.0 12/06/08
|
||||
%need to be improved
|
||||
|
||||
|
||||
\RequirePackage{graphicx}
|
||||
|
||||
%%%%%%%%%%%%%%%%%%%%% DEFINE USER-SPECIFIC MACROS BELOW %%%%%%%%%%%%%%%%%%%%%
|
||||
\def\Who {Pierre-Fran\c{c}ois Loos}
|
||||
\def\What {Dr}
|
||||
\def\Where {Universit\'e Paul Sabatier}
|
||||
\def\Address {Laboratoire de Chimie et Physique Quantiques}
|
||||
\def\CityZip {Toulouse, France}
|
||||
\def\Email {loos@irsamc.ups-tlse.fr}
|
||||
\def\TEL {+33 5 61 55 73 39}
|
||||
\def\URL {} % NOTE: use $\sim$ for tilde
|
||||
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% MARGINS %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
\textwidth 6in
|
||||
\textheight 9.25in
|
||||
\oddsidemargin 0.25in
|
||||
\evensidemargin 0.25in
|
||||
\topmargin -1.50in
|
||||
\longindentation 0.50\textwidth
|
||||
\parindent 5ex
|
||||
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%% ADDRESS MACRO BELOW %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
|
||||
\address{
|
||||
\includegraphics[height=0.7in]{CNRS_logo.pdf} \hspace*{\fill}\includegraphics[height=0.7in]{UPS_logo.pdf}
|
||||
\\
|
||||
\hrulefill
|
||||
\\
|
||||
{\small \What~\Who\hspace*{\fill} Telephone:\ \TEL
|
||||
\\
|
||||
\Where\hspace*{\fill} Email:\ \Email
|
||||
\\
|
||||
\Address\hspace*{\fill}
|
||||
\\
|
||||
\CityZip\hspace*{\fill} \URL}
|
||||
}
|
||||
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%% OTHER MACROS BELOW %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
\signature{\What~\Who}
|
||||
|
||||
\def\opening#1{\ifx\@empty\fromaddress
|
||||
\thispagestyle{firstpage}
|
||||
\hspace*{\longindendation}\today\par
|
||||
\else \thispagestyle{empty}
|
||||
{\centering\fromaddress \vspace{5\parskip} \\
|
||||
\today\hspace*{\fill}\par}
|
||||
\fi
|
||||
\vspace{3\parskip}
|
||||
{\raggedright \toname \\ \toaddress \par}\vspace{3\parskip}
|
||||
\noindent #1\par\raggedright\parindent 5ex\par
|
||||
}
|
||||
|
||||
%I do not know what does the macro below
|
||||
|
||||
%\long\def\closing#1{\par\nobreak\vspace{\parskip}
|
||||
%\stopbreaks
|
||||
%\noindent
|
||||
%\ifx\@empty\fromaddress\else
|
||||
%\hspace*{\longindentation}\fi
|
||||
%\parbox{\indentedwidth}{\raggedright
|
||||
%\ignorespaces #1\vskip .65in
|
||||
%\ifx\@empty\fromsig
|
||||
%\else \fromsig \fi\strut}
|
||||
%\vspace*{\fill}
|
||||
% \par}
|
||||
BIN
Manuscript/Cover_Letter/UPS_logo.pdf
Normal file
BIN
Manuscript/Cover_Letter/UPS_logo.pdf
Normal file
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
@ -1,13 +1,36 @@
|
||||
%% This BibTeX bibliography file was created using BibDesk.
|
||||
%% http://bibdesk.sourceforge.net/
|
||||
|
||||
%% Created for Pierre-Francois Loos at 2021-07-26 11:11:15 +0200
|
||||
%% Created for Pierre-Francois Loos at 2021-07-28 21:47:27 +0200
|
||||
|
||||
|
||||
%% Saved with string encoding Unicode (UTF-8)
|
||||
|
||||
|
||||
|
||||
@misc{Marie_2021b,
|
||||
archiveprefix = {arXiv},
|
||||
author = {Antoine Marie and F{\'a}bris Kossoski and Pierre-Fran{\c c}ois Loos},
|
||||
date-added = {2021-07-28 21:46:38 +0200},
|
||||
date-modified = {2021-07-28 21:46:50 +0200},
|
||||
eprint = {2106.11305},
|
||||
primaryclass = {physics.chem-ph},
|
||||
title = {Variational coupled cluster for ground and excited states},
|
||||
year = {2021}}
|
||||
|
||||
@article{Kossoski_2021,
|
||||
author = {Kossoski, F{\'a}bris and Marie, Antoine and Scemama, Anthony and Caffarel, Michel and Loos, Pierre-Fran{\c c}ois},
|
||||
date-added = {2021-07-28 21:45:35 +0200},
|
||||
date-modified = {2021-07-28 21:45:52 +0200},
|
||||
doi = {10.1021/acs.jctc.1c00348},
|
||||
journal = {J. Chem. Theory Comput.},
|
||||
number = {0},
|
||||
pages = {null},
|
||||
title = {Excited States from State-Specific Orbital-Optimized Pair Coupled Cluster},
|
||||
volume = {0},
|
||||
year = {0},
|
||||
Bdsk-Url-1 = {https://doi.org/10.1021/acs.jctc.1c00348}}
|
||||
|
||||
@article{Chilkuri_2021,
|
||||
abstract = {Selected configuration interaction (SCI) methods, when complemented with a second-order perturbative correction, provide near full configuration interaction (FCI) quality energies with only a small fraction of the Slater determinants of the FCI space. However, a selection criterion based on determinants alone does not ensure a spin-pure wave function. In other words, such SCI wave functions are not eigenfunctions of the {\^S}2 operator. In many situations (bond breaking, magnetic system, excited state, etc.), having a spin-adapted wave function is essential for a quantitatively correct description of the system. Here, we propose an efficient algorithm which, given an arbitrary determinant space, generates all the missing Slater determinants allowing one to obtain spin-adapted wave functions while avoiding manipulations involving configuration state functions. For example, generating all the possible determinants with 6 spin-up and 6 spin-down electrons in 12 open shells takes 21 CPU cycles per generated Slater determinant. The selection is still done with individual determinants, and one can take advantage of the basis of configuration state functions in the diagonalization of the Hamiltonian to reduce the memory footprint significantly.},
|
||||
author = {Vijay Gopal Chilkuri and Thomas Applencourt and Kevin Gasperich and Pierre-Fran{\c c}ois Loos and Anthony Scemama},
|
||||
@ -432,10 +455,10 @@
|
||||
year = {1985},
|
||||
Bdsk-Url-1 = {https://doi.org/10.1016/0009-2614(85)80934-9}}
|
||||
|
||||
@article{Marie_2021,
|
||||
@article{Marie_2021a,
|
||||
author = {Antoine Marie and Hugh G. A. Burton and Pierre-Fran{\c c}ois Loos},
|
||||
date-added = {2021-06-18 08:52:50 +0200},
|
||||
date-modified = {2021-06-18 08:53:45 +0200},
|
||||
date-modified = {2021-07-28 21:46:52 +0200},
|
||||
doi = {10.1088/1361-648X/abe795},
|
||||
journal = {J. Phys.: Condens. Matter},
|
||||
pages = {283001},
|
||||
@ -1430,10 +1453,10 @@
|
||||
Bdsk-Url-1 = {https://link.aps.org/doi/10.1103/PhysRevA.5.50},
|
||||
Bdsk-Url-2 = {https://doi.org/10.1103/PhysRevA.5.50}}
|
||||
|
||||
@article{Piecuch_2002,
|
||||
@article{Piecuch_2002b,
|
||||
author = {Piotr Piecuch and Karol Kowalski and Ian S. O. Pimienta and Michael J. Mcguire},
|
||||
date-added = {2021-05-06 15:31:25 +0200},
|
||||
date-modified = {2021-05-06 15:31:25 +0200},
|
||||
date-modified = {2021-07-26 18:04:21 +0200},
|
||||
doi = {10.1080/0144235021000053811},
|
||||
journal = {Int. Rev. Phys. Chem.},
|
||||
pages = {527-655},
|
||||
@ -2446,71 +2469,78 @@
|
||||
Bdsk-Url-1 = {https://doi.org/10.13140/RG.2.1.3187.9766}}
|
||||
|
||||
@article{gamess,
|
||||
title = {Recent developments in the general atomic and molecular electronic structure system},
|
||||
volume = {152},
|
||||
issn = {0021-9606, 1089-7690},
|
||||
url = {http://aip.scitation.org/doi/10.1063/5.0005188},
|
||||
doi = {10.1063/5.0005188},
|
||||
language = {en},
|
||||
number = {15},
|
||||
urldate = {2020-06-18},
|
||||
journal = {The Journal of Chemical Physics},
|
||||
author = {Barca, Giuseppe M. J. and Bertoni, Colleen and Carrington, Laura and Datta, Dipayan and De Silva, Nuwan and Deustua, J. Emiliano and Fedorov, Dmitri G. and Gour, Jeffrey R. and Gunina, Anastasia O. and Guidez, Emilie and Harville, Taylor and Irle, Stephan and Ivanic, Joe and Kowalski, Karol and Leang, Sarom S. and Li, Hui and Li, Wei and Lutz, Jesse J. and Magoulas, Ilias and Mato, Joani and Mironov, Vladimir and Nakata, Hiroya and Pham, Buu Q. and Piecuch, Piotr and Poole, David and Pruitt, Spencer R. and Rendell, Alistair P. and Roskop, Luke B. and Ruedenberg, Klaus and Sattasathuchana, Tosaporn and Schmidt, Michael W. and Shen, Jun and Slipchenko, Lyudmila and Sosonkina, Masha and Sundriyal, Vaibhav and Tiwari, Ananta and Galvez Vallejo, Jorge L. and Westheimer, Bryce and Wloch, Marta and Xu, Peng and Zahariev, Federico and Gordon, Mark S.},
|
||||
month = apr,
|
||||
year = {2020},
|
||||
pages = {154102}
|
||||
}
|
||||
author = {Barca, Giuseppe M. J. and Bertoni, Colleen and Carrington, Laura and Datta, Dipayan and De Silva, Nuwan and Deustua, J. Emiliano and Fedorov, Dmitri G. and Gour, Jeffrey R. and Gunina, Anastasia O. and Guidez, Emilie and Harville, Taylor and Irle, Stephan and Ivanic, Joe and Kowalski, Karol and Leang, Sarom S. and Li, Hui and Li, Wei and Lutz, Jesse J. and Magoulas, Ilias and Mato, Joani and Mironov, Vladimir and Nakata, Hiroya and Pham, Buu Q. and Piecuch, Piotr and Poole, David and Pruitt, Spencer R. and Rendell, Alistair P. and Roskop, Luke B. and Ruedenberg, Klaus and Sattasathuchana, Tosaporn and Schmidt, Michael W. and Shen, Jun and Slipchenko, Lyudmila and Sosonkina, Masha and Sundriyal, Vaibhav and Tiwari, Ananta and Galvez Vallejo, Jorge L. and Westheimer, Bryce and Wloch, Marta and Xu, Peng and Zahariev, Federico and Gordon, Mark S.},
|
||||
doi = {10.1063/5.0005188},
|
||||
issn = {0021-9606, 1089-7690},
|
||||
journal = {The Journal of Chemical Physics},
|
||||
language = {en},
|
||||
month = apr,
|
||||
number = {15},
|
||||
pages = {154102},
|
||||
title = {Recent developments in the general atomic and molecular electronic structure system},
|
||||
url = {http://aip.scitation.org/doi/10.1063/5.0005188},
|
||||
urldate = {2020-06-18},
|
||||
volume = {152},
|
||||
year = {2020},
|
||||
Bdsk-Url-1 = {http://aip.scitation.org/doi/10.1063/5.0005188},
|
||||
Bdsk-Url-2 = {https://doi.org/10.1063/5.0005188}}
|
||||
|
||||
@article{kowalski_2000a,
|
||||
@article{Kowalski_2000a,
|
||||
author = {Kowalski, Karol and Piecuch, Piotr},
|
||||
title = {{The method of moments of coupled-cluster equations and the renormalized CCSD[T], CCSD(T), CCSD(TQ), and CCSDT(Q) approaches}},
|
||||
date-modified = {2021-07-26 18:04:54 +0200},
|
||||
doi = {10.1063/1.481769},
|
||||
issn = {0021-9606},
|
||||
journal = {J. Chem. Phys.},
|
||||
volume = {113},
|
||||
month = {Jul},
|
||||
number = {1},
|
||||
pages = {18--35},
|
||||
year = {2000},
|
||||
month = {Jul},
|
||||
issn = {0021-9606},
|
||||
publisher = {American Institute of Physics},
|
||||
doi = {10.1063/1.481769}
|
||||
}
|
||||
|
||||
@article{kowalski_2000b,
|
||||
author = {Kowalski, Karol and Piecuch, Piotr},
|
||||
title = {{Renormalized CCSD(T) and CCSD(TQ) approaches: Dissociation of the N2 triple bond}},
|
||||
journal = {J. Chem. Phys.},
|
||||
title = {{The method of moments of coupled-cluster equations and the renormalized CCSD[T], CCSD(T), CCSD(TQ), and CCSDT(Q) approaches}},
|
||||
volume = {113},
|
||||
year = {2000},
|
||||
Bdsk-Url-1 = {https://doi.org/10.1063/1.481769}}
|
||||
|
||||
@article{Kowalski_2000b,
|
||||
author = {Kowalski, Karol and Piecuch, Piotr},
|
||||
date-modified = {2021-07-26 18:04:57 +0200},
|
||||
doi = {10.1063/1.1290609},
|
||||
issn = {0021-9606},
|
||||
journal = {J. Chem. Phys.},
|
||||
month = {Oct},
|
||||
number = {14},
|
||||
pages = {5644--5652},
|
||||
year = {2000},
|
||||
month = {Oct},
|
||||
issn = {0021-9606},
|
||||
publisher = {American Institute of Physics},
|
||||
doi = {10.1063/1.1290609}
|
||||
}
|
||||
@article{piecuch_2002,
|
||||
title = {{Renormalized CCSD(T) and CCSD(TQ) approaches: Dissociation of the N2 triple bond}},
|
||||
volume = {113},
|
||||
year = {2000},
|
||||
Bdsk-Url-1 = {https://doi.org/10.1063/1.1290609}}
|
||||
|
||||
@article{Piecuch_2002a,
|
||||
author = {Piecuch, Piotr and Kucharski, Stanis{\l}aw A. and Kowalski, Karol and Musia{\l}, Monika},
|
||||
title = {{Efficient computer implementation of the renormalized coupled-cluster methods: The R-CCSD[T], R-CCSD(T), CR-CCSD[T], and CR-CCSD(T) approaches}},
|
||||
date-modified = {2021-07-26 18:04:16 +0200},
|
||||
doi = {10.1016/S0010-4655(02)00598-2},
|
||||
issn = {0010-4655},
|
||||
journal = {Comput. Phys. Commun.},
|
||||
volume = {149},
|
||||
month = {Dec},
|
||||
number = {2},
|
||||
pages = {71--96},
|
||||
year = {2002},
|
||||
month = {Dec},
|
||||
issn = {0010-4655},
|
||||
publisher = {North-Holland},
|
||||
doi = {10.1016/S0010-4655(02)00598-2}
|
||||
}
|
||||
@article{piecuch_2005,
|
||||
title = {{Efficient computer implementation of the renormalized coupled-cluster methods: The R-CCSD[T], R-CCSD(T), CR-CCSD[T], and CR-CCSD(T) approaches}},
|
||||
volume = {149},
|
||||
year = {2002},
|
||||
Bdsk-Url-1 = {https://doi.org/10.1016/S0010-4655(02)00598-2}}
|
||||
|
||||
@article{Piecuch_2005,
|
||||
author = {Piecuch, Piotr and W{\l}och, Marta},
|
||||
title = {{Renormalized coupled-cluster methods exploiting left eigenstates of the similarity-transformed Hamiltonian}},
|
||||
date-modified = {2021-07-26 18:04:49 +0200},
|
||||
doi = {10.1063/1.2137318},
|
||||
issn = {0021-9606},
|
||||
journal = {J. Chem. Phys.},
|
||||
volume = {123},
|
||||
month = {Dec},
|
||||
number = {22},
|
||||
pages = {224105},
|
||||
year = {2005},
|
||||
month = {Dec},
|
||||
issn = {0021-9606},
|
||||
publisher = {American Institute of Physics},
|
||||
doi = {10.1063/1.2137318}
|
||||
}
|
||||
title = {{Renormalized coupled-cluster methods exploiting left eigenstates of the similarity-transformed Hamiltonian}},
|
||||
volume = {123},
|
||||
year = {2005},
|
||||
Bdsk-Url-1 = {https://doi.org/10.1063/1.2137318}}
|
||||
|
||||
@ -73,31 +73,32 @@
|
||||
\newcommand{\LCPQ}{Laboratoire de Chimie et Physique Quantiques (UMR 5626), Universit\'e de Toulouse, CNRS, UPS, France}
|
||||
\newcommand{\CEISAM}{Universit\'e de Nantes, CNRS, CEISAM UMR 6230, F-44000 Nantes, France}
|
||||
|
||||
\title{Reference correlation energies in finite Hilbert spaces: five- and six-membered rings}
|
||||
\title{Accurate full configuration interaction correlation energy estimates for five- and six-membered ring molecules}
|
||||
|
||||
\author{Yann Damour}
|
||||
\author{Yann \surname{Damour}}
|
||||
\affiliation{\LCPQ}
|
||||
\author{Micka\"el V\'eril}
|
||||
\author{Micka\"el \surname{V\'eril}}
|
||||
\affiliation{\LCPQ}
|
||||
\author{Michel Caffarel}
|
||||
\author{F\'abris \surname{Kossoski}}
|
||||
\affiliation{\LCPQ}
|
||||
\author{Denis Jacquemin}
|
||||
\author{Michel \surname{Caffarel}}
|
||||
\affiliation{\LCPQ}
|
||||
\author{Denis \surname{Jacquemin}}
|
||||
\affiliation{\CEISAM}
|
||||
\author{Anthony Scemama}
|
||||
\author{Anthony \surname{Scemama}}
|
||||
\email{scemama@irsamc.ups-tlse.fr}
|
||||
\affiliation{\LCPQ}
|
||||
\author{Pierre-Fran\c{c}ois Loos}
|
||||
\author{Pierre-Fran\c{c}ois \surname{Loos}}
|
||||
\email{loos@irsamc.ups-tlse.fr}
|
||||
\affiliation{\LCPQ}
|
||||
|
||||
% Abstract
|
||||
\begin{abstract}
|
||||
In the continuity of our recent work on the benzene molecule [\href{https://doi.org/10.1063/5.0027617}{J.~Chem.~Phys.~\textbf{153}, 176101 (2020)}], itself motivated by the blind challenge of Eriksen \textit{et al.} [\href{https://doi.org/10.1021/acs.jpclett.0c02621}{J.~Phys.~Chem.~Lett.~\textbf{11}, 8922 (2020)}] on the same system, we report reference frozen-core correlation energies for twelve five- and six-membered ring molecules (cyclopentadiene, furan, imidazole, pyrrole, thiophene, benzene, pyrazine, pyridazine, pyridine, pyrimidine, tetrazine, and triazine) in the standard correlation-consistent double-$\zeta$ Dunning basis set (cc-pVDZ).
|
||||
This corresponds to Hilbert spaces with sizes ranging from $10^{28}$ (for thiophene) to $10^{36}$ (for benzene).
|
||||
Our estimates are based on energetically optimized-orbital selected configuration interaction (SCI) calculations performed with the \textit{Configuration Interaction using a Perturbative Selection made Iteratively} (CIPSI) algorithm.
|
||||
The performance and convergence properties of several series of methods are investigated.
|
||||
In the continuity of our recent work on the benzene molecule [\href{https://doi.org/10.1063/5.0027617}{J.~Chem.~Phys.~\textbf{153}, 176101 (2020)}], itself motivated by the blind challenge of Eriksen \textit{et al.} [\href{https://doi.org/10.1021/acs.jpclett.0c02621}{J.~Phys.~Chem.~Lett.~\textbf{11}, 8922 (2020)}] on the same system, we report accurate full configuration interaction (FCI) frozen-core correlation energy estimates for twelve five- and six-membered ring molecules (cyclopentadiene, furan, imidazole, pyrrole, thiophene, benzene, pyrazine, pyridazine, pyridine, pyrimidine, s-tetrazine, and s-triazine) in the standard correlation-consistent double-$\zeta$ Dunning basis set (cc-pVDZ).
|
||||
Our FCI correlation energy estimates, with estimated error below 1 millihartree, are based on energetically optimized-orbital selected configuration interaction (SCI) calculations performed with the \textit{Configuration Interaction using a Perturbative Selection made Iteratively} (CIPSI) algorithm.
|
||||
Having at our disposal these accurate reference energies, the respective performance and convergence properties of several popular and widely-used families of quantum chemistry methods are investigated.
|
||||
In particular, we study the convergence properties of i) the M{\o}ller-Plesset perturbation series up to fifth-order (MP2, MP3, MP4, and MP5), ii) the iterative approximate single-reference coupled-cluster series CC2, CC3, and CC4, and iii) the single-reference coupled-cluster series CCSD, CCSDT, and CCSDTQ.
|
||||
The performance of the ground-state gold standard CCSD(T) is also investigated.
|
||||
The performance of the ground-state gold standard CCSD(T) as well as the completely renormalized CC model, CR-CC(2,3), are also investigated.
|
||||
\end{abstract}
|
||||
|
||||
% Title
|
||||
@ -107,18 +108,18 @@ The performance of the ground-state gold standard CCSD(T) is also investigated.
|
||||
\section{Introduction}
|
||||
\label{sec:intro}
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
Electronic structure theory relies heavily on approximations. \cite{Szabo_1996,Helgaker_2013,Jensen_2017}
|
||||
Non-relativistic electronic structure theory relies heavily on approximations. \cite{Szabo_1996,Helgaker_2013,Jensen_2017}
|
||||
Loosely speaking, to make any method practical, three main approximations must be enforced.
|
||||
The first fundamental approximation, known as the Born-Oppenheimer approximation, usually consists in assuming that the motion of nuclei and electrons are decoupled. \cite{Born_1927}
|
||||
The first fundamental approximation, known as the Born-Oppenheimer (or clamped-nuclei) approximation, consists in assuming that the motion of nuclei and electrons are decoupled. \cite{Born_1927}
|
||||
The nuclei coordinates can then be treated as parameters in the electronic Hamiltonian.
|
||||
The second central approximation which makes calculations feasible by a computer is the basis set approximation where one introduces a set of pre-defined basis functions to represent the many-electron wave function of the system.
|
||||
In most molecular calculations, a set of one-electron, atom-centered gaussian basis functions are introduced to expand the so-called one-electron molecular orbitals which are then used to build the many-electron Slater determinants.
|
||||
In most molecular calculations, a set of one-electron, atom-centered Gaussian basis functions are introduced to expand the so-called one-electron molecular orbitals which are then used to build the many-electron Slater determinants.
|
||||
The third and most relevant approximation in the present context is the ansatz (or form) of the electronic wave function $\Psi$.
|
||||
For example, in configuration interaction (CI) methods, the wave function is expanded as a linear combination of Slater determinants, while in (single-reference) coupled-cluster (CC) theory, \cite{Cizek_1966,Paldus_1972,Crawford_2000,Piecuch_2002,Bartlett_2007,Shavitt_2009} a reference Slater determinant $\PsiO$ [usually taken as the Hartree-Fock (HF) wave function] is multiplied by a wave operator defined as the exponentiated excitation operator $\hT = \sum_{k=1}^\Nel \hT_k$ (where $\Nel$ is the number of electrons).
|
||||
For example, in configuration interaction (CI) methods, the wave function is expanded as a linear combination of Slater determinants, while in (single-reference) coupled-cluster (CC) theory, \cite{Cizek_1966,Paldus_1972,Crawford_2000,Piecuch_2002b,Bartlett_2007,Shavitt_2009} a reference Slater determinant $\PsiO$ [usually taken as the Hartree-Fock (HF) wave function] is multiplied by a wave operator defined as the exponentiated excitation operator $\hT = \sum_{k=1}^\Nel \hT_k$ (where $\Nel$ is the number of electrons).
|
||||
|
||||
The truncation of $\hT$ allows to define a hierarchy of non-variational and size-extensive methods with improved accuracy:
|
||||
The truncation of $\hT$ allows to define a hierarchy of non-variational and size-extensive methods with increasing levels of accuracy:
|
||||
CC with singles and doubles (CCSD), \cite{Cizek_1966,Purvis_1982} CC with singles, doubles, and triples (CCSDT), \cite{Noga_1987a,Scuseria_1988} CC with singles, doubles, triples, and quadruples (CCSDTQ), \cite{Oliphant_1991,Kucharski_1992} with corresponding computational scalings of $\order*{\Norb^{6}}$, $\order*{\Norb^{8}}$, and $\order*{\Norb^{10}}$, respectively (where $\Norb$ denotes the number of orbitals).
|
||||
Parallel to the ``complete'' CC series presented above, an alternative family of approximate iterative CC models have been developed by the Aarhus group in the context of CC response theory \cite{Christiansen_1998} where one skips the most expensive terms and avoids the storage of the higher-excitation amplitudes: CC2, \cite{Christiansen_1995a} CC3, \cite{Christiansen_1995b,Koch_1997} and CC4. \cite{Kallay_2005,Matthews_2021}
|
||||
Parallel to the ``complete'' CC series presented above, an alternative family of approximate iterative CC models has been developed by the Aarhus group in the context of CC response theory \cite{Christiansen_1998} where one skips the most expensive terms and avoids the storage of the higher-excitation amplitudes: CC2, \cite{Christiansen_1995a} CC3, \cite{Christiansen_1995b,Koch_1997} and CC4. \cite{Kallay_2005,Matthews_2021}
|
||||
These iterative methods scale as $\order*{\Norb^{5}}$, $\order*{\Norb^{7}}$, and $\order*{\Norb^{9}}$, respectively, and can be seen as cheaper approximations of CCSD, CCSDT, and CCSDTQ.
|
||||
Coupled-cluster methods have been particularly successful at computing accurately various properties for small- and medium-sized molecules.
|
||||
\cite{Kallay_2003,Kallay_2004a,Gauss_2006,Kallay_2006,Gauss_2009,Chrayteh_2021,Sarkar_2021}
|
||||
@ -129,32 +130,33 @@ The non-variationality of truncated CC methods being less of an issue than the s
|
||||
However, a different strategy has recently made a come back in the context of CI methods. \cite{Bender_1969,Whitten_1969,Huron_1973}
|
||||
Indeed, selected CI (SCI) methods, \cite{Booth_2009,Giner_2013,Evangelista_2014,Giner_2015,Caffarel_2016b,Holmes_2016,Tubman_2016,Liu_2016,Ohtsuka_2017,Zimmerman_2017,Coe_2018,Garniron_2018} where one iteratively selects the important determinants from the FCI space (usually) based on a perturbative criterion, has been recently shown to be highly successful in order to produce reference energies for ground and excited states in small- and medium-sized molecules \cite{Caffarel_2014,Caffarel_2016a,Scemama_2016,Holmes_2017,Li_2018,Scemama_2018,Scemama_2018b,Li_2020,Loos_2018a,Chien_2018,Loos_2019,Loos_2020b,Loos_2020c,Loos_2020e,Garniron_2019,Eriksen_2020,Yao_2020,Williams_2020,Veril_2021,Loos_2021} thanks to efficient deterministic, stochastic or hybrid algorithms well suited for massive parallelization.
|
||||
We refer the interested reader to Refs.~\onlinecite{Loos_2020a,Eriksen_2021} for recent reviews.
|
||||
SCI methods are based on a well-known fact: amongst the very large number of determinants contained in the FCI space, only a relative small fraction of them significantly contributes to the energy.
|
||||
SCI methods are based on a well-known fact: amongst the very large number of determinants contained in the FCI space, only a tiny fraction of them significantly contributes to the energy.
|
||||
Accordingly, the SCI+PT2 family of methods performs a sparse exploration of the FCI space by selecting iteratively only the most energetically relevant determinants of the variational space and supplementing it with a second-order perturbative correction (PT2). \cite{Huron_1973,Garniron_2017,Sharma_2017,Garniron_2018,Garniron_2019}
|
||||
Although the formal scaling of such algorithms remains exponential, the prefactor is greatly reduced which explains their current attractiveness in the electronic structure community and much wider applicability than their standard FCI parent.
|
||||
Note that, very recently, several groups \cite{Aroeira_2021,Lee_2021,Magoulas_2021} have coupled CC and SCI methods via the externally-corrected CC methodology, \cite{Paldus_2017} showing promising performances for weakly and strongly correlated systems.
|
||||
|
||||
A rather different strategy in order to reach the holy grail FCI limit is to resort to M{\o}ller-Plesset (MP) perturbation theory, \cite{Moller_1934}
|
||||
which popularity originates from its black-box nature, size-extensivity, and relatively low computational scaling, making it easily applied to a broad range of molecular systems.
|
||||
Again, at least in theory, one can obtain the exact energy of the system by ramping up the degree of the perturbative series. \cite{Marie_2021}
|
||||
whose popularity originates from its black-box nature, size-extensivity, and relatively low computational scaling, making it easily applied to a broad range of molecular systems.
|
||||
Again, at least in theory, one can obtain the exact energy of the system by ramping up the degree of the perturbative series. \cite{Marie_2021a}
|
||||
The second-order M{\o}ller-Plesset (MP2) method \cite{Moller_1934} [which scales as $\order*{\Norb^{5}}$] has been broadly adopted in quantum chemistry for several decades, and is now included in the increasingly popular double-hybrid functionals \cite{Grimme_2006} alongside exact HF exchange.
|
||||
Its higher-order variants [MP3, \cite{Pople_1976}
|
||||
MP4, \cite{Krishnan_1980} MP5, \cite{Kucharski_1989} and MP6 \cite{He_1996a,He_1996b} which scale as $\order*{\Norb^{6}}$, $\order*{\Norb^{7}}$, $\order*{\Norb^{8}}$, and $\order*{\Norb^{9}}$ respectively] have been investigated much more scarcely.
|
||||
However, it is now widely recognised that the series of MP approximations might show erratic, slowly convergent, or divergent behavior that limits its applicability and systematic improvability. \cite{Laidig_1985,Knowles_1985,Handy_1985,Gill_1986,Laidig_1987,Nobes_1987,Gill_1988,Gill_1988a,Lepetit_1988,Malrieu_2003,Marie_2021}
|
||||
However, it is now widely recognized that the series of MP approximations might show erratic, slowly convergent, or divergent behavior that limits its applicability and systematic improvability. \cite{Laidig_1985,Knowles_1985,Handy_1985,Gill_1986,Laidig_1987,Nobes_1987,Gill_1988,Gill_1988a,Lepetit_1988,Malrieu_2003,Marie_2021a}
|
||||
Again, MP perturbation theory and CC methods can be coupled.
|
||||
The CCSD(T) method, \cite{Raghavachari_1989} known as the gold-standard of quantum chemistry for weakly correlated systems, is probably the most iconic example of such coupling.
|
||||
The CCSD(T) method, \cite{Raghavachari_1989} where one includes iteratively the single and double excitations and perturbatively (from MP4 and partially MP5) the triple excitations, known as the ``gold-standard'' of quantum chemistry for weakly correlated systems thanks to its excellent accuracy/cost ratio, is probably the most iconic example of such coupling.
|
||||
|
||||
Motivated by the recent blind test of Eriksen \textit{et al.}\cite{Eriksen_2020}~reporting the performance of a large panel of emerging electronic structure methods [the many-body expansion FCI (MBE-FCI), \cite{Eriksen_2017,Eriksen_2018,Eriksen_2019a,Eriksen_2019b} adaptive sampling CI (ASCI), \cite{Tubman_2016,Tubman_2018,Tubman_2020} iterative CI (iCI), \cite{Liu_2014,Liu_2016,Lei_2017,Zhang_2020} semistochastic heat-bath CI (SHCI), \cite{Holmes_2016,Holmes_2017,Sharma_2017} the full coupled-cluster reduction (FCCR), \cite{Xu_2018,Xu_2020} density-matrix renormalization group (DMRG), \cite{White_1992,White_1993,Chan_2011} adaptive-shift FCI quantum Monte Carlo (AS-FCIQMC), \cite{Booth_2009,Cleland_2010,Ghanem_2019} and cluster-analysis-driven FCIQMC (CAD-FCIQMC) \cite{Deustua_2017,Deustua_2018}] on the non-relativistic frozen-core correlation energy of the benzene molecule in the standard correlation-consistent double-$\zeta$ Dunning basis set (cc-pVDZ), some of us have recently investigated the performance of the SCI method known as \textit{Configuration Interaction using a Perturbative Selection made Iteratively} (CIPSI). \cite{Huron_1973,Giner_2013,Giner_2015,Garniron_2018,Garniron_2019} on the very same system \cite{Loos_2020e} [see also Ref.~\onlinecite{Lee_2020} for a study of the performance of phaseless auxiliary-field quantum Monte Carlo (ph-AFQMC) \cite{Motta_2018}].
|
||||
In the continuity of this recent work, we report here a significant extension by estimating the (frozen-core) FCI/cc-pVDZ correlation energy of twelve cyclic molecules (cyclopentadiene, furan, imidazole, pyrrole, thiophene, benzene, pyrazine, pyridazine, pyridine, pyrimidine, tetrazine, and triazine) with the help of CIPSI employing energetically-optimized orbitals at the same level of theory. \cite{Yao_2020,Yao_2021}
|
||||
In the continuity of this recent work, we report here a significant extension by estimating the (frozen-core) FCI/cc-pVDZ correlation energy of twelve cyclic molecules (cyclopentadiene, furan, imidazole, pyrrole, thiophene, benzene, pyrazine, pyridazine, pyridine, pyrimidine, s-tetrazine, and s-triazine) with the help of CIPSI employing energetically-optimized orbitals at the same level of theory. \cite{Yao_2020,Yao_2021}
|
||||
These systems are depicted in Fig.~\ref{fig:mol}.
|
||||
This set of molecular systems corresponds to Hilbert spaces with sizes ranging from $10^{28}$ (for thiophene) to $10^{36}$ (for benzene).
|
||||
This set of molecular systems corresponds to Hilbert spaces with sizes ranging from $10^{29}$ to $10^{36}$.
|
||||
In addition to CIPSI, the performance and convergence properties of several series of methods are investigated.
|
||||
In particular, we study i) the MP perturbation series up to fifth-order (MP2, MP3, MP4, and MP5), ii) the CC2, CC3, and CC4 approximate series, and ii) the ``complete'' CC series up to quadruples (\ie, CCSD, CCSDT, and CCSDTQ).
|
||||
The performance of the ground-state gold standard CCSD(T) is also investigated.
|
||||
The performance of the ground-state gold standard CCSD(T) as well as the completely renormalized (CR) CC model, CR-CC(2,3), \cite{,Kowalski_2000a,Kowalski_2000b,Piecuch_2002a,Piecuch_2002b,Piecuch_2005} are also investigated.
|
||||
|
||||
%%% FIG 1 %%%
|
||||
\begin{figure*}
|
||||
\includegraphics[width=\linewidth]{mol}
|
||||
\includegraphics[width=0.8\linewidth]{ring5}
|
||||
\includegraphics[width=\linewidth]{ring6}
|
||||
\caption{
|
||||
Five-membered rings (top) and six-membered rings (bottom) considered in this study.
|
||||
\label{fig:mol}}
|
||||
@ -168,6 +170,7 @@ In Sec.~\ref{sec:res}, we report our reference FCI correlation energies for the
|
||||
These reference correlation energies are then used to benchmark and study the convergence properties of various perturbative and CC methods (Sec.~\ref{sec:mpcc_res}).
|
||||
Finally, we draw our conclusions in Sec.~\ref{sec:ccl}.
|
||||
|
||||
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
\section{CIPSI with optimized orbitals}
|
||||
\label{sec:OO-CIPSI}
|
||||
@ -202,8 +205,7 @@ We then linearly extrapolate, using large variational wave functions, the CIPSI
|
||||
Further details concerning the extrapolation procedure are provided below (see Sec.~\ref{sec:res}).
|
||||
|
||||
Orbital optimization techniques at the SCI level are theoretically straightforward, but practically challenging.
|
||||
\toto{
|
||||
Most of the technology presented here has been borrowed from complete-active-space self-consistent-field (CASSCF) methods \cite{Werner_1980,Werner_1985,Sun_2017,Kreplin_2019,Kreplin_2020} but one of the strength of SCI methods is that one does not need to select an active space and to classify orbitals as active, inactive, and virtual orbitals.}
|
||||
Some of the technology presented here has been borrowed from complete-active-space self-consistent-field (CASSCF) methods \cite{Werner_1980,Werner_1985,Sun_2017,Kreplin_2019,Kreplin_2020} but one of the strength of SCI methods is that one does not need to select an active space and to classify orbitals as active, inactive, and virtual orbitals.
|
||||
Here, we detail our orbital optimization procedure within the CIPSI algorithm and we assume that the variational wave function is normalized, \ie, $\braket*{\Psivar}{\Psivar} = 1$.
|
||||
|
||||
As stated in Sec.~\ref{sec:intro}, $\Evar$ depends on both the CI coefficients $\{ c_I \}_{1 \le I \le \Ndet}$ [see Eq.~\eqref{eq:Psivar}] but also on the orbital rotation parameters $\{\kappa_{pq}\}_{1 \le p,q \le \Norb}$.
|
||||
@ -294,62 +296,18 @@ are the one- and two-electron integrals, respectively.
|
||||
Because the size of the CI space is much larger than the orbital space, for each macroiteration, we perform multiple \textit{microiterations} which consist in iteratively minimizing the variational energy \eqref{eq:Evar_c_k} with respect to the $\Norb(\Norb-1)/2$ independent orbital rotation parameters for a fixed set of determinants.
|
||||
After each microiteration (\ie, orbital rotation), the one- and two-electron integrals [see Eqs.~\eqref{eq:one} and \eqref{eq:two}] have to be updated.
|
||||
Moreover, the CI matrix must be re-diagonalized and new one- and two-electron density matrices [see Eqs.~\eqref{eq:one_dm} and \eqref{eq:two_dm}] are computed.
|
||||
Microiterations are stopped when a stationary point is found, \ie, $\norm{\bg}_\infty < \tau$, where $\tau$ is a user-defined threshold which has been set to $10^{-3}$ a.u.~in the present study, and a new CIPSI selection step is performed.
|
||||
Microiterations are stopped when a stationary point is found, \ie, $\norm{\bg}_\infty < \tau$, where $\tau$ is a user-defined threshold which has been set to $10^{-4}$ a.u.~in the present study, and a new CIPSI selection step is performed.
|
||||
Note that a tight convergence is not critical here as a new set of microiterations is performed at each macroiteration and a new production CIPSI run is performed from scratch using the final set of orbitals (see Sec.~\ref{sec:compdet}).
|
||||
This procedure might sound computationally expensive but one has to realize that the microiterations are usually performed only for relatively compact variational spaces.
|
||||
Therefore, the computational bottleneck remains the diagonalization of the CI matrix for very large variational spaces.
|
||||
|
||||
%\begin{equation}
|
||||
% \Evar = \sum_{pq} h_p^q \gamma_p^q + \frac{1}{2} \sum_{pqrs} v_{pq}^{rs} \Gamma_{pq}^{rs},
|
||||
%\end{equation}
|
||||
|
||||
To enhance the convergence of the microiteration process, we employ an adaptation of the Newton-Raphson method known as ``trust region''. \cite{Nocedal_1999}
|
||||
This popular variant defines a region where the quadratic approximation \eqref{eq:EvarTaylor} is an adequate representation of the objective energy function \eqref{eq:Evar_c_k} and it evolves during the optimization process in order to preserve the adequacy via a constraint on the step size preventing it from overstepping, \ie, $\norm{\bk} \leq \Delta$, where $\Delta$ is the trust radius.
|
||||
By introducing a Lagrange multiplier $\lambda$ to control the trust-region size, one replaces Eq.~\eqref{eq:kappa_newton} by $\bk = - (\bH + \lambda \bI)^{-1} \cdot \bg$.
|
||||
The addition of the level shift $\lambda \geq 0$ removes the negative eigenvalues and ensures the positive definiteness of the Hessian matrix by reducing the step size.
|
||||
By choosing the right value of $\lambda$, $\norm{\bk}$ is constrained into a hypersphere of radius $\Delta$ and is able to evolve from the Newton direction at $\lambda = 0$ to the steepest descent direction as $\lambda$ grows.
|
||||
The evolution of the trust radius during the optimization and the use of a condition to reject the step when the energy rises ensure the convergence of the algorithm.
|
||||
More detail can be found in Ref.~\onlinecite{Nocedal_1999}.
|
||||
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
\section{Computational details}
|
||||
\label{sec:compdet}
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
The geometries of the twelve systems considered in the present study were all obtained at the \toto{frozen-core?} CC3/aug-cc-pVTZ level of theory and were extracted from a previous study. \cite{Loos_2020a}
|
||||
Note that, for the sake of consistency, the geometry of benzene considered here is different from one of Ref.~\onlinecite{Loos_2020e} which was obtained at a lower level of theory [MP2/6-31G(d)]. \cite{Schreiber_2008}
|
||||
The MP2, MP3, MP4, CC2, CC3, CC4, CCSD, CCSDT, and CCSDTQ calculations were performed with CFOUR, \cite{Matthews_2020} the CCSD(T), CCSD(TQ), CCSD[T], R-CCSD[T] and R-CCSD(T) calculations were made with GAMESS 2014R1, \cite{gamess,kowalski_2000a,kowalski_2000b,piecuch_2002,piecuch_2005} and MP5 calculations were computed with GAUSSIAN 09. \cite{g09}
|
||||
The CIPSI calculations were performed with QUANTUM PACKAGE. \cite{Garniron_2019}
|
||||
In the current implementation, the selection step and the PT2 correction are computed simultaneously via a hybrid semistochastic algorithm.\cite{Garniron_2017,Garniron_2019} %(which explains the statistical error associated with the PT2 correction in the following).
|
||||
Here, we employ the renormalized version of the PT2 correction which was recently implemented and tested for a more efficient extrapolation to the FCI limit thanks to a partial resummation of the higher orders of perturbation. \cite{Garniron_2019}
|
||||
We refer the interested reader to Ref.~\onlinecite{Garniron_2019} where one can find all the details regarding the implementation of the PT2 correction and the CIPSI algorithm.
|
||||
For all these calculations, we consider Dunning's correlation-consistent double-$\zeta$ basis (cc-pVDZ).
|
||||
|
||||
Although the FCI energy has the enjoyable property of being independent of the set of one-electron orbitals used to construct the many-electron Slater determinants, as a truncated CI method, the convergence properties of CIPSI strongly dependent on this orbital choice.
|
||||
In the present study, we investigate, in particular, the convergence behavior of the CIPSI energy for two sets of orbitals: natural orbitals (NOs) and optimized orbitals (OOs).
|
||||
Following our usual procedure, \cite{Scemama_2018,Scemama_2018b,Scemama_2019,Loos_2018a,Loos_2019,Loos_2020a,Loos_2020b,Loos_2020c,Loos_2020e} we perform first a preliminary SCI calculation using HF orbitals in order to generate a SCI wave function with at least $10^7$ determinants.
|
||||
Natural orbitals are computed based on this wave function and they are used to perform a new CIPSI run.
|
||||
Successive orbital optimizations are then performed, which consist in minimizing the variational CIPSI energy at each iteration up to approximately $2 \times 10^5$ determinants.
|
||||
When convergence is achieved in terms of orbital optimization, as our ``production'' run, we perform a new CIPSI calculation from scratch using this set of optimized orbitals.
|
||||
Using optimized orbitals has the undeniable advantage to produce, for a given variational energy, more compact CI expansions (see Sec.~\ref{sec:res}).
|
||||
For the benzene molecule, we also explore the use of localized orbitals (LOs) which are produced with the Boys-Foster localization procedure \cite{Boys_1960} that we apply to the natural orbitals in several orbital windows in order to preserve a strict $\sigma$-$\pi$ separation in the planar systems considered here. \cite{Loos_2020e}
|
||||
Because they take advantage of the local character of electron correlation, localized orbitals have been shown to provide faster convergence towards the FCI limit compared to natural orbitals. \cite{Angeli_2003,Angeli_2009,BenAmor_2011,Suaud_2017,Chien_2018,Eriksen_2020,Loos_2020e}
|
||||
As we shall see below, employing optimized orbitals has the advantage to produce an even smoother and faster convergence of the SCI energy toward the FCI limit.
|
||||
\toto{Note both localized and optimized orbitals do break the spatial symmetry.}
|
||||
Unlike excited-state calculations where it is important to enforce that the wave functions are eigenfunctions of the $\Hat{S}^2$ spin operator, \cite{Chilkuri_2021} the present wave functions do not fulfill this property as we aim for the lowest possible energy of a closed-shell singlet state.
|
||||
We have found that $\expval*{\Hat{S}^2}$ is, nonetheless, very close to zero for each system.
|
||||
|
||||
The present calculations have been performed on the AMD partition of GENCI's Irene supercomputer.
|
||||
Each Irene's AMD node is a dual-socket AMD Rome (EPYC) CPU at 2.60 GHz with 256GiB of RAM, with a total of 64 physical cores per socket.
|
||||
These nodes are connected via Infiniband HDR100.
|
||||
In total, the present calculations have required around 3~million core hours.
|
||||
|
||||
All the data (geometries, energies, etc) and supplementary material associated with the present manuscript are openly available in Zenodo at \href{http://doi.org/XX.XXXX/zenodo.XXXXXXX}{http://doi.org/XX.XXXX/zenodo.XXXXXXX}.
|
||||
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
\section{Results and discussion}
|
||||
\label{sec:res}
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
|
||||
More details can be found in Ref.~\onlinecite{Nocedal_1999}.
|
||||
%%% FIG 2 %%%
|
||||
\begin{figure*}
|
||||
\includegraphics[width=0.24\textwidth]{Cyclopentadiene_EvsNdet}
|
||||
@ -368,7 +326,7 @@ All the data (geometries, energies, etc) and supplementary material associated w
|
||||
\includegraphics[width=0.24\textwidth]{Triazine_EvsNdet}
|
||||
\caption{$\Delta \Evar$ (solid) and $\Delta \Evar + \EPT$ (dashed) as functions of the number of determinants $\Ndet$ in the variational space for the twelve cyclic molecules represented in Fig.~\ref{fig:mol}.
|
||||
Two sets of orbitals are considered: natural orbitals (NOs, in red) and optimized orbitals (OOs, in blue).
|
||||
The CCSDTQ correlation energy is represented as a thick black line.
|
||||
The FCI estimate of the correlation energy is represented as a thick black line.
|
||||
\label{fig:vsNdet}}
|
||||
\end{figure*}
|
||||
%%% %%% %%%
|
||||
@ -393,7 +351,7 @@ All the data (geometries, energies, etc) and supplementary material associated w
|
||||
Two sets of orbitals are considered: natural orbitals (NOs, in red) and optimized orbitals (OOs, in blue).
|
||||
The five-point weighted linear fit using the five largest variational wave functions for each set is depicted as a dashed black line.
|
||||
The weights are taken as the inverse square of the perturbative corrections.
|
||||
The CCSDTQ correlation energy is also represented as a thick black line.
|
||||
The FCI estimate of the correlation energy is represented as a thick black line.
|
||||
\label{fig:vsEPT2}}
|
||||
\end{figure*}
|
||||
%%% %%% %%%
|
||||
@ -426,9 +384,11 @@ All the data (geometries, energies, etc) and supplementary material associated w
|
||||
CCSDTQ & $-193.5465$ & $-738.2$ & $-229.4100$ & $-766.7$ & $-225.6123$ & $-776.9$ & $-209.5860$ & $-757.4$ & $-552.0485$ & $-727.5$ \\
|
||||
\hline
|
||||
CCSD(T) & $-193.5439$ & $-735.6$ & $-229.4073$ & $-764.0$ & $-225.6099$ & $-774.5$ & $-209.5836$ & $-754.9$ & $-552.0458$ & $-724.8$
|
||||
\\
|
||||
\\
|
||||
CR-CC(2,3)& $-193.5439$ & $-735.6$ & $-229.4075$ & $-764.2$ & $-225.6098$ & $-774.5$ & $-209.5835$ & $-754.9$ & $-552.0459$ & $-724.9$
|
||||
\\
|
||||
\hline
|
||||
FCI & & $-739.2(1)$ & & $-768.2(1)$ & & $-778.2(1)$ & & $-758.5(1)$ & & $-728.9(3)$\\
|
||||
FCI & & $-739.2(1)$ & & $-768.2(1)$ & & $-778.2(1)$ & & $-758.5(1)$ & & $-728.9(3)$\\
|
||||
\end{tabular}
|
||||
\end{ruledtabular}
|
||||
\end{table*}
|
||||
@ -443,7 +403,7 @@ All the data (geometries, energies, etc) and supplementary material associated w
|
||||
\label{tab:Tab6-VDZ}}
|
||||
\begin{ruledtabular}
|
||||
\begin{tabular}{lcccccccccccccc}
|
||||
& \mc{2}{c}{Benzene} & \mc{2}{c}{Pyrazine} & \mc{2}{c}{Pyridazine} & \mc{2}{c}{Pyridine} & \mc{2}{c}{Pyrimidine} & \mc{2}{c}{Tetrazine} & \mc{2}{c}{Triazine} \\
|
||||
& \mc{2}{c}{Benzene} & \mc{2}{c}{Pyrazine} & \mc{2}{c}{Pyridazine} & \mc{2}{c}{Pyridine} & \mc{2}{c}{Pyrimidine} & \mc{2}{c}{s-Tetrazine} & \mc{2}{c}{s-Triazine} \\
|
||||
\cline{2-3} \cline{4-5} \cline{6-7} \cline{8-9} \cline{10-11} \cline{12-13} \cline{14-15}
|
||||
Method & $E$ & $\Delta E$ & $E$ & $\Delta E$ & $E$ & $\Delta E$ & $E$ & $\Delta E$
|
||||
& $E$ & $\Delta E$ & $E$ & $\Delta E$ & $E$ & $\Delta E$ \\
|
||||
@ -464,6 +424,7 @@ All the data (geometries, energies, etc) and supplementary material associated w
|
||||
CCSDTQ & $-231.5826$ & $-860.4$ & $-263.6053$ & $-902.3$ & $-263.5770$ & $-907.1$ & $-247.5960$ & $-880.8$ & $-263.6126$ & $-899.0$ & $-295.5712$ & $-955.4$ & $-279.6331$ & $-915.8$ \\
|
||||
\hline
|
||||
CCSD(T) & $-231.5798$ & $-857.5$ & $-263.6024$ & $-899.4$ & $-263.5740$ & $-904.1$ & $-247.5929$ & $-877.7$ & $-263.6099$ & $-896.2$ & $-295.5680$ & $-952.2$ & $-279.6305$ & $-913.1$ \\
|
||||
CR-CC(2,3) & $-231.5792$ & $-856.9$ & $-263.6017$ & $-898.7$ & $-263.5732$ & $-903.3$ & $-247.5922$ & $-877.1$ & $-263.6091$ & $-895.5$ & $-295.5670$ & $-951.3$ & $-279.6298$ & $-912.5$ \\
|
||||
\hline
|
||||
FCI & & $-862.9(3)$ & & $-904.6(4)$ & & $-908.8(1)$ & & $-882.7(4)$ & & $-900.5(2)$ & & $-957.9(4)$ & & $-918.4(3)$\\
|
||||
\end{tabular}
|
||||
@ -472,155 +433,108 @@ All the data (geometries, energies, etc) and supplementary material associated w
|
||||
\end{squeezetable}
|
||||
%%% %%% %%%
|
||||
|
||||
|
||||
%%% TABLE III %%%
|
||||
\begin{squeezetable}
|
||||
\begin{table}
|
||||
\caption{
|
||||
Extrapolated correlation energies $\Delta \Eextrap$ (in \SI{}{\milli\hartree}) for the twelve cyclic molecules represented in Fig.~\ref{fig:mol} and their associated fitting errors (in \SI{}{\milli\hartree}) obtained via weighted linear fits with a varying number of points.
|
||||
Two sets of orbitals are considered: natural orbitals and optimized orbitals.
|
||||
The weights are taken as the inverse square of the perturbative corrections.
|
||||
For a $m$-point fit, the $m$ largest variational wave functions are used.
|
||||
\label{tab:fit}}
|
||||
\begin{ruledtabular}
|
||||
\begin{tabular}{lccc}
|
||||
Molecule & Number of & \mc{2}{c}{Fitting parameters} \\
|
||||
\cline{3-4}
|
||||
& fitting points & $\Delta \Eextrap$ & Fitting error \\
|
||||
\begin{tabular}{lccccc}
|
||||
Molecule & Number of & \mc{2}{c}{Natural orbitals} & \mc{2}{c}{Optimized orbitals} \\
|
||||
\cline{3-4}\cline{5-6}
|
||||
& fitting points & $\Delta \Eextrap$ & Fitting error & $\Delta \Eextrap$ & Fitting error \\
|
||||
\hline
|
||||
Cyclopentadiene & 3 & $-739.295$ & $0.199$ \\
|
||||
& 4 & $-739.309$ & $0.088$ \\
|
||||
& \bf5 & $\bf-739.230$ & $\bf0.074$ \\
|
||||
& 6 & $-739.304$ & $0.072$ \\
|
||||
& 7 & $-739.292$ & $0.055$ \\
|
||||
% & 8 & $-739.119$ & $0.131$ \\
|
||||
Cyclopentadiene & 3 & $-740.639$ & $0.273$ & $-739.295$ & $0.199$ \\
|
||||
& 4 & $-740.243$ & $0.306$ & $-739.309$ & $0.088$ \\
|
||||
&\bf5 & $-740.047$ & $0.242$ & $\bf-739.230$& $\bf0.074$ \\
|
||||
& 6 & $-739.952$ & $0.187$ & $-739.304$ & $0.072$ \\
|
||||
& 7 & $-739.761$ & $0.204$ & $-739.292$ & $0.055$ \\
|
||||
\hline
|
||||
Furan & 3 & $-767.790$ & $0.064$ \\
|
||||
& 4 & $-768.104$ & $0.196$ \\
|
||||
& \bf5 & $\bf-768.194$ & $\bf0.135$ \\
|
||||
& 6 & $-768.060$ & $0.131$ \\
|
||||
& 7 & $-768.086$ & $0.101$ \\
|
||||
% & 8 & $-767.904$ & $0.154$ \\
|
||||
Furan & 3 & $-766.090$ & $0.729$ & $-767.790$ & $0.064$ \\
|
||||
& 4 & $-766.445$ & $0.459$ & $-768.104$ & $0.196$ \\
|
||||
&\bf5 & $-766.582$ & $0.318$ & $\bf-768.194$ &$\bf0.135$ \\
|
||||
& 6 & $-766.366$ & $0.288$ & $-768.060$ & $0.131$ \\
|
||||
& 7 & $-766.507$ & $0.254$ & $-768.086$ & $0.101$ \\
|
||||
\hline
|
||||
Imidazole & 3 & $-778.295$ & $0.356$ \\
|
||||
& 4 & $-778.270$ & $0.150$ \\
|
||||
& \bf5 & $\bf-778.178$ & $\bf0.105$ \\
|
||||
& 6 & $-778.174$ & $0.072$ \\
|
||||
& 7 & $-778.051$ & $0.099$ \\
|
||||
% & 8 & $-777.992$ & $0.089$ \\
|
||||
Imidazole & 3 & $-778.148$ & $2.197$ & $-778.295$ & $0.356$ \\
|
||||
& 4 & $-777.436$ & $1.107$ & $-778.270$ & $0.150$ \\
|
||||
&\bf5 & $-776.300$ & $0.996$ & $\bf-778.178$ &$\bf0.105$ \\
|
||||
& 6 & $-776.104$ & $0.712$ & $-778.174$ & $0.072$ \\
|
||||
& 7 & $-776.098$ & $0.541$ & $-778.051$ & $0.099$ \\
|
||||
\hline
|
||||
Pyrrole & 3 & $-758.650$ & $0.321$ \\
|
||||
& 4 & $-758.389$ & $0.174$ \\
|
||||
& \bf5 & $\bf-758.460$ & $\bf0.110$ \\
|
||||
& 6 & $-758.352$ & $0.100$ \\
|
||||
& 7 & $-758.347$ & $0.075$ \\
|
||||
% & 8 & $-758.357$ & $0.059$ \\
|
||||
Pyrrole & 3 & $-758.309$ & $0.447$ & $-758.650$ & $0.321$ \\
|
||||
& 4 & $-758.749$ & $0.393$ & $-758.389$ & $0.174$ \\
|
||||
&\bf5 & $-758.405$ & $0.359$ & $\bf-758.460$ &$\bf0.110$ \\
|
||||
& 6 & $-758.136$ & $0.334$ & $-758.352$ & $0.100$ \\
|
||||
& 7 & $-757.990$ & $0.283$ & $-758.347$ & $0.075$ \\
|
||||
\hline
|
||||
Thiophene & 3 & $-728.744$ & $0.691$ \\
|
||||
& 4 & $-729.052$ & $0.331$ \\
|
||||
& \bf5 & $\bf-728.948$ & $\bf0.203$ \\
|
||||
& 6 & $-728.987$ & $0.140$ \\
|
||||
& 7 & $-729.067$ & $0.117$ \\
|
||||
% & 8 & $-728.876$ & $0.162$ \\
|
||||
Thiophene & 3 & $-728.054$ & $0.134$ & $-728.744$ & $0.691$ \\
|
||||
& 4 & $-728.240$ & $0.139$ & $-729.052$ & $0.331$ \\
|
||||
&\bf5 & $-728.243$ & $0.087$ & $\bf-728.948$ &$\bf0.203$ \\
|
||||
& 6 & $-728.242$ & $0.062$ & $-728.987$ & $0.140$ \\
|
||||
& 7 & $-728.420$ & $0.144$ & $-729.067$ & $0.117$ \\
|
||||
\hline
|
||||
Benzene & 3 & $-862.325$ & $0.279$ \\
|
||||
& 4 & $-863.024$ & $0.424$ \\
|
||||
& \bf5 & $\bf-862.890$ & $\bf0.266$ \\
|
||||
& 6 & $-862.360$ & $0.383$ \\
|
||||
& 7 & $-862.083$ & $0.339$ \\
|
||||
% & 8 & $-861.711$ & $0.370$ \\
|
||||
Benzene & 3 & $-860.350$ & $0.496$ & $-862.325$ & $0.279$ \\
|
||||
& 4 & $-861.949$ & $0.811$ & $-863.024$ & $0.424$ \\
|
||||
&\bf5 & $-861.807$ & $0.474$ & $\bf-862.890$ &$\bf0.266$ \\
|
||||
& 6 & $-861.110$ & $0.539$ & $-862.360$ & $0.383$ \\
|
||||
& 7 & $-861.410$ & $0.444$ & $-862.083$ & $0.339$ \\
|
||||
\hline
|
||||
Pyrazine & 3 & $-904.867$ & $1.420$ \\
|
||||
& 4 & $-904.588$ & $0.650$ \\
|
||||
& \bf5 & $\bf-904.550$ & $\bf0.385$ \\
|
||||
& 6 & $-903.982$ & $0.439$ \\
|
||||
& 7 & $-903.746$ & $0.359$ \\
|
||||
% & 8 & $-903.549$ & $0.311$ \\
|
||||
Pyrazine & 3 & $-904.148$ & $0.035$ & $-904.867$ & $1.420$ \\
|
||||
& 4 & $-904.726$ & $0.377$ & $-904.588$ & $0.650$ \\
|
||||
&\bf5 & $-904.274$ & $0.383$ & $\bf-904.550$ &$\bf0.385$ \\
|
||||
& 6 & $-903.980$ & $0.341$ & $-903.982$ & $0.439$ \\
|
||||
& 7 & $-903.621$ & $0.370$ & $-903.746$ & $0.359$ \\
|
||||
\hline
|
||||
Pyridazine & 3 & $-909.292$ & $0.024$ \\
|
||||
& 4 & $-908.808$ & $0.230$ \\
|
||||
& \bf5 & $\bf-908.820$ & $\bf0.133$ \\
|
||||
& 6 & $-908.342$ & $0.303$ \\
|
||||
& 7 & $-908.368$ & $0.224$ \\
|
||||
% & 8 & $-908.229$ & $0.198$ \\
|
||||
Pyridazine & 3 & $-910.856$ & $3.053$ & $-909.292$ & $0.024$ \\
|
||||
& 4 & $-908.222$ & $1.834$ & $-908.808$ & $0.230$ \\
|
||||
&\bf5 & $-909.282$ & $1.191$ & $\bf-908.820$ &$\bf0.133$ \\
|
||||
& 6 & $-912.566$ & $1.727$ & $-908.342$ & $0.303$ \\
|
||||
& 7 & $-910.694$ & $2.210$ & $-908.368$ & $0.224$ \\
|
||||
\hline
|
||||
Pyridine & 3 & $-883.363$ & $0.047$ \\
|
||||
& 4 & $-883.413$ & $0.029$ \\
|
||||
& \bf5 & $\bf-882.700$ & $\bf0.405$ \\
|
||||
& 6 & $-882.361$ & $0.341$ \\
|
||||
& 7 & $-882.023$ & $0.330$ \\
|
||||
% & 8 & $-881.732$ & $0.322$ \\
|
||||
Pyridine & 3 & $-883.025$ & $3.919$ & $-883.363$ & $0.047$ \\
|
||||
& 4 & $-883.862$ & $1.869$ & $-883.413$ & $0.029$ \\
|
||||
&\bf5 & $-881.664$ & $1.760$ & $\bf-882.700$ &$\bf0.405$ \\
|
||||
& 6 & $-880.422$ & $1.456$ & $-882.361$ & $0.341$ \\
|
||||
& 7 & $-880.191$ & $1.084$ & $-882.023$ & $0.330$ \\
|
||||
\hline
|
||||
Pyrimidine & 3 & $-900.817$ & $0.726$ \\
|
||||
& 4 & $-900.383$ & $0.356$ \\
|
||||
& \bf5 & $\bf-900.496$ & $\bf0.214$ \\
|
||||
& 6 & $-900.698$ & $0.190$ \\
|
||||
& 7 & $-900.464$ & $0.206$ \\
|
||||
% & 8 & $-900.226$ & $0.227$ \\
|
||||
Pyrimidine & 3 & $-900.386$ & $1.884$ & $-900.817$ & $0.726$ \\
|
||||
& 4 & $-901.441$ & $0.991$ & $-900.383$ & $0.356$ \\
|
||||
&\bf5 & $-900.354$ & $0.865$ & $\bf-900.496$ &$\bf0.214$ \\
|
||||
& 6 & $-900.240$ & $0.594$ & $-900.698$ & $0.190$ \\
|
||||
& 7 & $-899.689$ & $0.565$ & $-900.464$ & $0.206$ \\
|
||||
\hline
|
||||
Tetrazine & 3 & $-957.559$ & $0.246$ \\
|
||||
& 4 & $-957.299$ & $0.160$ \\
|
||||
& \bf5 & $\bf-957.869$ & $\bf0.349$ \\
|
||||
& 6 & $-957.744$ & $0.247$ \\
|
||||
& 7 & $-957.709$ & $0.183$ \\
|
||||
% & 8 & $-957.558$ & $0.176$ \\
|
||||
s-Tetrazine & 3 & & & $-957.559$ & $0.246$ \\
|
||||
& 4 & & & $-957.299$ & $0.160$ \\
|
||||
&\bf5 & & & $\bf-957.869$ &$\bf0.349$ \\
|
||||
& 6 & & & $-957.744$ & $0.247$ \\
|
||||
& 7 & & & $-957.709$ & $0.183$ \\
|
||||
\hline
|
||||
Triazine & 3 & $-919.596$ & $0.105$ \\
|
||||
& 4 & $-918.457$ & $0.538$ \\
|
||||
& \bf5 & $\bf-918.355$ & $\bf0.312$ \\
|
||||
& 6 & $-918.206$ & $0.226$ \\
|
||||
& 7 & $-917.876$ & $0.267$ \\
|
||||
% & 8 & $-917.533$ & $0.308$ \\
|
||||
s-Triazine & 3 & $-917.221$ & $0.693$ & $-919.596$ & $0.105$ \\
|
||||
& 4 & $-918.723$ & $0.913$ & $-918.457$ & $0.538$ \\
|
||||
&\bf5 & $-917.402$ & $0.956$ & $\bf-918.355$ &$\bf0.312$ \\
|
||||
& 6 & $-916.517$ & $0.862$ & $-918.206$ & $0.226$ \\
|
||||
& 7 & $-916.544$ & $0.643$ & $-917.876$ & $0.267$ \\
|
||||
\end{tabular}
|
||||
\end{ruledtabular}
|
||||
\end{table}
|
||||
\end{squeezetable}
|
||||
%%% %%% %%%
|
||||
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
\subsection{CIPSI estimates}
|
||||
\label{sec:cipsi_res}
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
|
||||
We first study the convergence of the CIPSI energy as a function of the number of determinants.
|
||||
Our motivation here is to generate FCI-quality reference correlation energies for the twelve cyclic molecules represented in Fig.~\ref{fig:mol} in order to benchmark, in a second time, the performance and convergence properties of various mainstream MP and CC methods (see Sec.~\ref{sec:mpcc_res}).
|
||||
For the natural and optimized orbital sets, we report, in Fig.~\ref{fig:vsNdet}, the evolution of the variational correlation energy $\Delta \Evar = \Evar - \EHF$ (where $\EHF$ is the HF energy) and its perturbatively corrected value $\Delta \Evar + \EPT$ with respect to the number of determinants $\Ndet$ for each cyclic molecule.
|
||||
As compared to natural orbitals (solid red lines), one can see that, for a given number of determinants, the use of optimized orbitals greatly lowers $\Delta \Evar$ (solid blue lines).
|
||||
Adding the perturbative correction $\EPT$ yields similar curves for both sets of orbitals (dashed lines).
|
||||
This indicates that, for a given number of determinants, $\EPT$ (which, we recall, provides a qualitative idea to the distance to the FCI limit) is much smaller for optimized orbitals than for natural orbitals.
|
||||
This is further evidenced in Fig.~\ref{fig:vsEPT2} where we show the behavior of $\Delta \Evar$ as a function of $\EPT$ for both sets of orbitals.
|
||||
From Fig.~\ref{fig:vsEPT2}, it is clear that, using optimized orbitals, the behavior of $\Delta \Evar$ is much more linear and produces smaller $\EPT$ values, hence facilitating the extrapolation procedure to the FCI limit (see below).
|
||||
The five-point weighted linear fit using the five largest variational wave functions are also represented (dashed black lines), while the CCSDTQ correlation energy (solid black line) is reported for comparison purposes in Figs.~\ref{fig:vsNdet} and \ref{fig:vsEPT2}.
|
||||
|
||||
%%% FIG 4 %%%
|
||||
\begin{figure}
|
||||
\includegraphics[width=\linewidth]{Benzene_EvsNdetLO}
|
||||
\caption{$\Delta \Evar$ (solid) and $\Delta \Evar + \EPT$ (dashed) as functions of the number of determinants $\Ndet$ in the variational space for the benzene molecule.
|
||||
Three sets of orbitals are considered: natural orbitals (NOs, in red), localized orbitals (LOs, in green), and optimized orbitals (OOs, in blue).
|
||||
The CCSDTQ correlation energy is represented as a thick black line.
|
||||
The FCI estimate of the correlation energy is represented as a thick black line.
|
||||
\label{fig:BenzenevsNdet}}
|
||||
\end{figure}
|
||||
%%% %%% %%%
|
||||
|
||||
Figure \ref{fig:BenzenevsNdet} compares the convergence of $\Delta \Evar$ for the natural, localized, and optimized sets of orbitals in the particular case of benzene.
|
||||
As mentioned in Sec.~\ref{sec:compdet}, although both the localized and optimized orbitals break the spatial symmetry to take advantage of the local nature of electron correlation, the latter set further improve on the use of former set.
|
||||
More quantitatively, optimized orbitals produce the same variational energy as localized orbitals with, roughly, a ten-fold reduction in the number of determinants.
|
||||
A similar improvement is observed going from natural to localized orbitals.
|
||||
\titou{Comment on PT2 for localized orbitals.}
|
||||
Accordingly to these observations, all our FCI correlation energy estimates have been produced with the set of optimized orbitals.
|
||||
|
||||
To do so, we have then extrapolated the orbital-optimized variational CIPSI correlation energies to $\EPT = 0$ via a weighted five-point linear fit using the five largest variational wave functions (see Fig.~\ref{fig:vsEPT2}).
|
||||
The fitting weights have been taken as the inverse square of the perturbative corrections.
|
||||
Our final FCI correlation energy estimates are reported in Tables \ref{tab:Tab5-VDZ} and \ref{tab:Tab6-VDZ} for the five- and six-membered rings, respectively, alongside their corresponding fitting error.
|
||||
The stability of these estimates are illustrated by the results gathered in Table \ref{tab:fit} where we report, for each system, the extrapolated correlation energies $\Delta \Eextrap$ and their associated fitting errors obtained via weighted linear fits varying the number of fitting points from $3$ to $7$.
|
||||
Although we cannot provide a mathematically rigorous error bar, the data provided by Table \ref{tab:fit} show that the extrapolation procedure is robust and that our FCI estimates are very likely accurate to a few tenths of a millihartree.
|
||||
Logically, the FCI estimates for the five-membered rings seem slightly more accurate than for the (larger) six-membered rings.
|
||||
Note that it is pleasing to see that, although different geometries are considered, our present estimate of the frozen-core correlation energy of the benzene molecule in the cc-pVDZ basis is very close to the one reported in Refs.~\onlinecite{Eriksen_2020,Loos_2020e}.
|
||||
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
\subsection{Benchmark of CC and MP methods}
|
||||
\label{sec:mpcc_res}
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
|
||||
%%% FIG 5 %%%
|
||||
\begin{figure*}
|
||||
\includegraphics[width=0.32\textwidth]{Cyclopentadiene_MPCC}
|
||||
@ -640,12 +554,12 @@ Note that it is pleasing to see that, although different geometries are consider
|
||||
\includegraphics[width=0.32\textwidth]{Triazine_MPCC}
|
||||
\caption{Convergence of the correlation energy (in \SI{}{\milli\hartree}) as a function of the computational cost for the twelve cyclic molecules represented in Fig.~\ref{fig:mol}.
|
||||
Three series of methods are considered: i) MP2, MP3, MP4, and MP5 (blue), ii) CC2, CC3, and CC4 (green), and iii) CCSD, CCSDT, CCSDTQ (red).
|
||||
The CIPSI estimate of the correlation energy is represented as a black line for reference.
|
||||
The FCI estimate of the correlation energy is represented as a black line.
|
||||
\label{fig:MPCC}}
|
||||
\end{figure*}
|
||||
%%% %%% %%%
|
||||
|
||||
%%% TABLE III %%%
|
||||
%%% TABLE IV %%%
|
||||
\begin{squeezetable}
|
||||
\begin{table}
|
||||
\caption{
|
||||
@ -654,7 +568,7 @@ Note that it is pleasing to see that, although different geometries are consider
|
||||
\label{tab:stats}}
|
||||
\begin{ruledtabular}
|
||||
\begin{tabular}{lcdddd}
|
||||
Method & Scaling & \tabc{MAE} & \tabc{MSE} & \tabc{Min} & \tabc{Max} \\
|
||||
Method & Scaling & \tabc{MAE} & \tabc{MSE} & \tabc{Max} & \tabc{Min} \\
|
||||
\hline
|
||||
MP2 & $\order{N^5}$ & 68.4 & 68.4 & 80.6 & 57.8 \\
|
||||
MP3 & $\order{N^6}$ & 46.5 & 46.5 & 58.4 & 37.9 \\
|
||||
@ -669,29 +583,110 @@ Note that it is pleasing to see that, although different geometries are consider
|
||||
CCSDT & $\order{N^8}$ & 4.5 & 4.5 & 6.3 & 3.0 \\
|
||||
CCSDTQ & $\order{N^{10}}$& 1.8 & 1.8 & 2.6 & 1.0 \\
|
||||
\hline
|
||||
CCSD(T) & $\order{N^7}$ & 4.5 & 4.5 & 5.7 & 3.6 \\
|
||||
CCSD(T) & $\order{N^7}$ & 4.5 & 4.5 & 5.7 & 3.6 \\
|
||||
CR-CC(2,3) & $\order{N^7}$ & 5.0 & 5.0 & 6.6 & 3.6 \\
|
||||
\end{tabular}
|
||||
\end{ruledtabular}
|
||||
\end{table}
|
||||
\end{squeezetable}
|
||||
%%% %%% %%%
|
||||
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
\section{Computational details}
|
||||
\label{sec:compdet}
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
The geometries of the twelve systems considered in the present study were all obtained at the CC3/aug-cc-pVTZ level of theory and were extracted from a previous study. \cite{Loos_2020a}
|
||||
Note that, for the sake of consistency, the geometry of benzene considered here is different from the one of Ref.~\onlinecite{Loos_2020e} which was obtained at a lower level of theory [MP2/6-31G(d)]. \cite{Schreiber_2008}
|
||||
The MP2, MP3, MP4, CC2, CC3, CC4, CCSD, CCSDT, and CCSDTQ calculations were performed with CFOUR, \cite{Matthews_2020} the CCSD(T) and CR-CC(2,3) calculations were made with GAMESS 2014R1, \cite{gamess} and MP5 calculations were computed with GAUSSIAN 09. \cite{g09}
|
||||
The CIPSI calculations were performed with QUANTUM PACKAGE. \cite{Garniron_2019}
|
||||
In the current implementation, the selection step and the PT2 correction are computed simultaneously via a hybrid semistochastic algorithm.\cite{Garniron_2017,Garniron_2019} %(which explains the statistical error associated with the PT2 correction in the following).
|
||||
Here, we employ the renormalized version of the PT2 correction which was recently implemented and tested for a more efficient extrapolation to the FCI limit thanks to a partial resummation of the higher orders of perturbation. \cite{Garniron_2019}
|
||||
We refer the interested reader to Ref.~\onlinecite{Garniron_2019} where one can find all the details regarding the implementation of the PT2 correction and the CIPSI algorithm.
|
||||
For all these calculations, we consider Dunning's correlation-consistent double-$\zeta$ basis (cc-pVDZ).
|
||||
|
||||
Although the FCI energy has the enjoyable property of being independent of the set of one-electron orbitals used to construct the many-electron Slater determinants, as a truncated CI method, the convergence properties of CIPSI strongly dependent on this orbital choice.
|
||||
In the present study, we investigate, in particular, the convergence behavior of the CIPSI energy for two sets of orbitals: natural orbitals (NOs) and optimized orbitals (OOs).
|
||||
Following our usual procedure, \cite{Scemama_2018,Scemama_2018b,Scemama_2019,Loos_2018a,Loos_2019,Loos_2020a,Loos_2020b,Loos_2020c,Loos_2020e} we perform first a preliminary SCI calculation using HF orbitals in order to generate a SCI wave function with at least $10^7$ determinants.
|
||||
Natural orbitals are computed based on this wave function and they are used to perform a new CIPSI run up to $8 \times 10^7$ determinants.
|
||||
Successive orbital optimizations are then performed, which consist in minimizing the variational CIPSI energy at each macroiteration up to approximately $2 \times 10^5$ determinants.
|
||||
When convergence is achieved in terms of orbital optimization, as our production run, we perform a new CIPSI calculation from scratch using this set of optimized orbitals to $8 \times 10^7$ determinants.
|
||||
Using optimized orbitals has the undeniable advantage to produce, for a given variational energy, more compact CI expansions (see Sec.~\ref{sec:res}).
|
||||
For the benzene molecule, we also explore the use of localized orbitals (LOs) which are produced with the Boys-Foster localization procedure \cite{Boys_1960} that we apply to the natural orbitals in several orbital windows in order to preserve a strict $\sigma$-$\pi$ separation in the planar systems considered here. \cite{Loos_2020e}
|
||||
Because they take advantage of the local character of electron correlation, localized orbitals have been shown to provide faster convergence towards the FCI limit compared to natural orbitals. \cite{Angeli_2003,Angeli_2009,BenAmor_2011,Suaud_2017,Chien_2018,Eriksen_2020,Loos_2020e}
|
||||
As we shall see below, employing optimized orbitals has the advantage to produce an even smoother and faster convergence of the SCI energy toward the FCI limit.
|
||||
Note that both localized and optimized orbitals do break the spatial symmetry.
|
||||
Unlike excited-state calculations where it is important to enforce that the wave functions are eigenfunctions of the $\Hat{S}^2$ spin operator, \cite{Chilkuri_2021} the present wave functions do not fulfill this property as we aim for the lowest possible energy of a closed-shell singlet state.
|
||||
We have found that $\expval*{\Hat{S}^2}$ is, nonetheless, very close to zero for each system.
|
||||
|
||||
The present calculations have been performed on the AMD partition of GENCI's Irene supercomputer.
|
||||
Each Irene's AMD node is a dual-socket AMD Rome (EPYC) CPU at 2.60 GHz with 256GiB of RAM, with a total of 64 physical cores per socket.
|
||||
These nodes are connected via Infiniband HDR100.
|
||||
In total, the present calculations have required around 3~million core hours.
|
||||
|
||||
All the data (geometries, energies, etc) and supplementary material associated with the present manuscript are openly available in Zenodo at \href{http://doi.org/XX.XXXX/zenodo.XXXXXXX}{http://doi.org/XX.XXXX/zenodo.XXXXXXX}.
|
||||
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
\section{Results and discussion}
|
||||
\label{sec:res}
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
\subsection{CIPSI estimates}
|
||||
\label{sec:cipsi_res}
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
|
||||
We first study the convergence of the CIPSI energy as a function of the number of determinants.
|
||||
Our motivation here is to generate FCI-quality reference correlation energies for the twelve cyclic molecules represented in Fig.~\ref{fig:mol} in order to benchmark, in a second time, the performance and convergence properties of various mainstream MP and CC methods (see Sec.~\ref{sec:mpcc_res}).
|
||||
For the natural and optimized orbital sets, we report, in Fig.~\ref{fig:vsNdet}, the evolution of the variational correlation energy $\Delta \Evar = \Evar - \EHF$ (where $\EHF$ is the HF energy) and its perturbatively corrected value $\Delta \Evar + \EPT$ with respect to the number of determinants $\Ndet$ for each cyclic molecule.
|
||||
As compared to natural orbitals (solid red lines), one can see that, for a given number of determinants, the use of optimized orbitals greatly lowers $\Delta \Evar$ (solid blue lines).
|
||||
Adding the perturbative correction $\EPT$ yields similar curves for both sets of orbitals (dashed lines).
|
||||
This indicates that, for a given number of determinants, $\EPT$ (which, we recall, provides a qualitative idea to the distance to the FCI limit) is much smaller for optimized orbitals than for natural orbitals.
|
||||
This is further evidenced in Fig.~\ref{fig:vsEPT2} where we show the behavior of $\Delta \Evar$ as a function of $\EPT$ for both sets of orbitals.
|
||||
From Fig.~\ref{fig:vsEPT2}, it is clear that, using optimized orbitals, the behavior of $\Delta \Evar$ is much more linear and produces smaller $\EPT$ values, hence facilitating the extrapolation procedure to the FCI limit (see below).
|
||||
The five-point weighted linear fit using the five largest variational wave functions are also represented (dashed black lines), while the FCI estimate of the correlation energy (solid black line) is reported for reference in Figs.~\ref{fig:vsNdet} and \ref{fig:vsEPT2}.
|
||||
|
||||
Figure \ref{fig:BenzenevsNdet} compares the convergence of $\Delta \Evar$ for the natural, localized, and optimized sets of orbitals in the particular case of benzene.
|
||||
As mentioned in Sec.~\ref{sec:compdet}, although both the localized and optimized orbitals break the spatial symmetry to take advantage of the local nature of electron correlation, the latter set further improve on the use of former set.
|
||||
More quantitatively, optimized orbitals produce the same variational energy as localized orbitals with, roughly, a ten-fold reduction in the number of determinants.
|
||||
A similar improvement is observed going from natural to localized orbitals.
|
||||
According to these observations, all our FCI correlation energy estimates have been produced with the set of optimized orbitals.
|
||||
|
||||
To do so, we have then extrapolated the orbital-optimized variational CIPSI correlation energies to $\EPT = 0$ via a weighted five-point linear fit using the five largest variational wave functions (see Fig.~\ref{fig:vsEPT2}).
|
||||
The fitting weights have been taken as the inverse square of the perturbative corrections.
|
||||
Our final FCI correlation energy estimates are reported in Tables \ref{tab:Tab5-VDZ} and \ref{tab:Tab6-VDZ} for the five- and six-membered rings, respectively, alongside their corresponding fitting error.
|
||||
The stability of these estimates are illustrated by the results gathered in Table \ref{tab:fit} where we report, for each system, the extrapolated correlation energies $\Delta \Eextrap$ and their associated fitting errors obtained via weighted linear fits varying the number of fitting points from $3$ to $7$.
|
||||
Although we cannot provide a mathematically rigorous error bar, the data provided by Table \ref{tab:fit} show that the extrapolation procedure is robust and that our FCI estimates are very likely accurate to a few tenths of a millihartree.
|
||||
Logically, the FCI estimates for the five-membered rings seem slightly more accurate than for the (larger) six-membered rings.
|
||||
Note that it is pleasing to see that, although different geometries are considered, our present estimate of the frozen-core correlation energy of the benzene molecule in the cc-pVDZ basis is very close to the one reported in Refs.~\onlinecite{Eriksen_2020,Loos_2020e}.
|
||||
|
||||
Table \ref{tab:fit} does report extrapolated correlation energies and fitting errors for both natural and optimized orbitals.
|
||||
Again, the superiority of the latter set is clear as the variation in extrapolated values and fitting error is much larger with the natural set.
|
||||
Taking cyclopentadiene as an example, the extrapolated values vary by almost \SI{1}{\milli\hartree} with natural orbitals and less than \SI{0.1}{\milli\hartree} with the optimized set.
|
||||
The fitting errors follow the same trend.
|
||||
|
||||
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
\subsection{Benchmark of CC and MP methods}
|
||||
\label{sec:mpcc_res}
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
|
||||
Using the CIPSI estimates of the FCI correlation energy produced in Sec.~\ref{sec:cipsi_res}, we now study the performance and convergence properties of three series of methods: i) MP2, MP3, MP4, and MP5, ii) CC2, CC3, and CC4, and iii) CCSD, CCSDT, and CCSDTQ.
|
||||
Additionally, we also report CCSD(T) correlation energies.
|
||||
Additionally, we also report CCSD(T) and CR-CC(2,3) correlation energies.
|
||||
All these data are reported in Tables \ref{tab:Tab5-VDZ} and \ref{tab:Tab6-VDZ} for the five- and six-membered rings, respectively.
|
||||
In Fig.~\ref{fig:MPCC}, we show, for each molecule, the convergence of the correlation energy for each series of methods as a function of the computational cost of the corresponding method.
|
||||
The FCI correlation energy estimate is represented as a black line for reference.
|
||||
Key statistical quantities [mean absolute error (MAE), mean signed error (MSE), and minimum (Min) and maximum (Max) absolute errors with respect to the FCI reference values] are also reported in Table \ref{tab:stats} for each method as well as their formal computational scaling.
|
||||
|
||||
First, we investigate the ``complete'' and well-established series of methods CCSD, CCSDT, and CCSDTQ.
|
||||
Unfortunately, CC with singles, doubles, triples, quadruples and pentuples (CCSDTQP) calculations are out of reach here. \cite{Hirata_2000,Kallay_2001}
|
||||
Unfortunately, CC with singles, doubles, triples, quadruples, and pentuples (CCSDTQP) calculations are out of reach here. \cite{Hirata_2000,Kallay_2001}
|
||||
As expected for the present set of weakly correlated systems, going from CCSD to CCSDTQ, one improves systematically and quickly the correlation energies with MAEs of $39.4$, $4.5$, \SI{1.8}{\milli\hartree} for CCSD, CCSDT, and CCSDTQ, respectively.
|
||||
As usually observed, CCSD(T) (MAE of \SI{4.5}{\milli\hartree}) provides similar correlation energies than the more expensive CCSDT method by computing perturbatively (instead of iteratively) the triple excitations.
|
||||
As usually observed, CCSD(T) (MAE of \SI{4.5}{\milli\hartree}) provides similar correlation energies than the more expensive CCSDT method by computing perturbatively (instead of iteratively) the triple excitations, while CCSD(T) and CR-CC(2,3) perform equally well.
|
||||
|
||||
|
||||
Second, let us look into the series of MP approximations which is known, as mentioned in Sec.~\ref{sec:intro}, to potentially exhibit ``surprising'' behavior depending on the type of correlation at play.\cite{Laidig_1985,Knowles_1985,Handy_1985,Gill_1986,Laidig_1987,Nobes_1987,Gill_1988,Gill_1988a,Lepetit_1988,Malrieu_2003}
|
||||
(See Ref.~\onlinecite{Marie_2021} for a detailed discussion).
|
||||
Second, let us look into the series of MP approximations which is known, as mentioned in Sec.~\ref{sec:intro}, to potentially exhibit ``surprising'' behaviors depending on the type of correlation at play.\cite{Laidig_1985,Knowles_1985,Handy_1985,Gill_1986,Laidig_1987,Nobes_1987,Gill_1988,Gill_1988a,Lepetit_1988,Malrieu_2003}
|
||||
(See Ref.~\onlinecite{Marie_2021a} for a detailed discussion).
|
||||
For each system, the MP series decreases monotonically up to MP4 but raises quite significantly when one takes into account the fifth-order correction.
|
||||
We note that the MP4 correlation energy is always quite accurate (MAE of \SI{2.1}{\milli\hartree}) and is only a few millihartree higher than the FCI value (except in the case of tetrazine where the MP4 number is very slightly below the reference value): MP5 (MAE of \SI{9.4}{\milli\hartree}) is thus systematically worse than MP4 for these weakly-correlated systems.
|
||||
We note that the MP4 correlation energy is always quite accurate (MAE of \SI{2.1}{\milli\hartree}) and is only a few millihartree higher than the FCI value (except in the case of s-tetrazine where the MP4 number is very slightly below the reference value): MP5 (MAE of \SI{9.4}{\milli\hartree}) is thus systematically worse than MP4 for these weakly-correlated systems.
|
||||
Importantly here, one notices that MP4 [which scales as $\order*{N^7}$] is systematically on par with the more expensive $\order*{N^{10}}$ CCSDTQ method which exhibits a slightly smaller MAE of \SI{1.8}{\milli\hartree}.
|
||||
|
||||
Third, we investigate the approximate CC series of methods CC2, CC3, and CC4.
|
||||
@ -709,8 +704,10 @@ Compared to natural orbitals, we have shown that, by using energetically optimiz
|
||||
|
||||
Thanks to these reference FCI energies, we have then benchmarked three families of popular electronic structure methods: i) the MP perturbation series up to fifth-order (MP2, MP3, MP4, and MP5), ii) the approximate CC series CC2, CC3, and CC4, and iii) the ``complete'' CC series CCSD, CCSDT, and CCSDTQ.
|
||||
Our results have shown that, with a $\order*{N^7}$ scaling, MP4 provides an interesting accuracy/cost ratio for this particular set of weakly correlated systems, while MP5 systematically worsen the perturbative estimates of the correlation energy.
|
||||
We have evidenced that CC3 (where the triples are computed iteratively) also outperforms the perturbative-triples CCSD(T) method with the same $\order*{N^7}$ scaling but also its more expensive parent, CCSDT.
|
||||
We have evidenced that CC3 (where the triples are computed iteratively) also outperforms the perturbative-triples CCSD(T) method with the same $\order*{N^7}$ scaling, its completely renormalized version CR-CC(2,3), but also its more expensive parent, CCSDT.
|
||||
A similar trend is observed for the methods including quadruple excitations, where the $\order*{N^9}$ CC4 model has been shown to be more accurate than CCSDTQ [which scales as $\order*{N^{10}}$].
|
||||
Of course, the present trends are only valid for this particular class of (weakly-correlated) systems and it would be desirable to provide more variety in terms of systems in the future by including more challenging systems such as, for example, transition metal compounds.
|
||||
Some work along this line is currently being done.
|
||||
|
||||
As perspectives, we are currently investigating the performance of the present approach for excited states in order to expand the QUEST database of vertical excitation energies. \cite{Veril_2021}
|
||||
We hope to report on this in the near future.
|
||||
@ -718,8 +715,8 @@ The compression of the variational space brought by optimized orbitals could be
|
||||
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
\begin{acknowledgements}
|
||||
This work was performed using HPC resources from GENCI-TGCC (2021-gen1738) and from CALMIP (Toulouse) under allocation 2021-18005, and was also supported by the European Centre of Excellence in Exascale Computing TREX - Targeting Real Chemical Accuracy at the Exascale. This project has received funding from the European Union's Horizon 2020 - Research and Innovation program - under grant agreement no. 95216.
|
||||
This project has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (Grant agreement No.~863481 and 952165).
|
||||
This work was performed using HPC resources from GENCI-TGCC (2021-gen1738) and from CALMIP (Toulouse) under allocation 2021-18005, and was also supported by the European Centre of Excellence in Exascale Computing TREX --- Targeting Real Chemical Accuracy at the Exascale. This project has received funding from the European Union's Horizon 2020 --- Research and Innovation program --- under grant agreement no.~952165.
|
||||
This project has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (Grant agreement No.~863481).
|
||||
\end{acknowledgements}
|
||||
%%%%%%%%%%%%%%%%%%%%%%%%%
|
||||
|
||||
|
||||
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
@ -23,3 +23,4 @@
|
||||
10485760 -231.4529267946 -0.1034069718 0.0002056718 -0.1015824747 0.0002020429
|
||||
20971520 -231.4645290852 -0.0942474663 0.0001880281 -0.0927652325 0.0001850710
|
||||
41943040 -231.4768202820 -0.0842614465 0.0001668731 -0.0831039064 0.0001645807
|
||||
83886080 -231.4895438170 -0.0746164115 0.0003622034 -0.0737313072 0.0035790692
|
||||
|
||||
@ -23,3 +23,4 @@
|
||||
10485760 -247.4637428655 -0.1057773192 0.0002104625 -0.1039205954 0.0002067683
|
||||
20971520 -247.4753699775 -0.0967407932 0.0001933432 -0.0952218594 0.0001903075
|
||||
41943040 -247.4876731985 -0.0865878863 0.0001719015 -0.0853982537 0.0001695398
|
||||
83886080 -247.5004308830 -0.0770957245 0.0003815831 -0.0761773500 0.0003770376
|
||||
|
||||
Loading…
Reference in New Issue
Block a user