***************************************************************
**** ****
**** ESF Programme ****
**** ****
**** RELATIVISTIC EFFECTS IN HEAVY ELEMENT CHEMISTRY ****
**** AND PHYSICS ****
**** ****
***************************************************************
Newsletter No. 15 (9. October 1995)
______________________________________________________________
Editor: Bernd Hess, hess@uni-bonn.de
Tel. 49-228-732920
FAX 49-228-732251
______________________________________________________________
The programme 'Relativistic Effects in Heavy-Element Chemistry and Physics'
('REHE') has been initiated by the European Science
Foundation in November 1992 and it is expected to run for 5 years, i.e.
from 1993 through 1997. The programme is intended to strengthen the in-
dicated "field" and to facilitate interactions between European scientists
concerned with related topics.
The 'Steering Committee' of the programme has at present the following
members:
E. J. Baerends (Amsterdam)
J.P. Daudey (Toulouse)
K. Faegri (Oslo)
I.P. Grant (Oxford)
B. Hess (Bonn, Vice-Chairman)
H. U. Karow (ESF)
J. Karwowski (Torun)
P. Pyykko (Helsinki, Chairman)
K. Schwarz (Vienna)
A. Sgamellotti (Perugia).
================================================================================
--- E D I T O R I A L
Please send material for the forthcoming newsletter to my attention,
hess@uni-bonn.de
The newsletter will be sent out every second month around the 10th day
of the month. Contributions should arrive in Bonn until the end of the
preceding month.
| The next newsletter (#16) is scheduled for Beginning of December 1995.
| Please send your contributions until end of November 1995.
Please send material >by e/mail< that enables us to fill the
following topics in forthcoming newsletters
All REHE newsletters are now available on www under URL
http://pcgate.thch.uni-bonn.de/tc/hess/esf/nl.html
================================================================================
--- F E L L O W S H I P S
In the framework of the REHE programme, there is support available
for visits of doctoral students and also for senior scientists at
institutions in a foreign partner country. This support covers visits
lasting 2-4 months ("long-term visits") which will give the holders
time to acclimatize to the methods used in the host laboratory as well as
short visits ("short-term visits") of only a few days.
Please send a short application detailing the project, the names of the
scientists involved and the aproximate date and duration of the visit
to either Pekka Pyykko or Bernd Hess. Please refer to REHE newsletter #7
for details.
Should the planned dates of your stay change for any reason, you are
requested to notify the Chairman as soon as possible with a copy to
the ESF.
================================================================================
--- R E S E A R C H N E W S AND R E L A T E D I N F O R M A T I O N
Summaries of recent research or comments to it (up to 1 page),
which are of general interest to the 'REHE' community, may
be submitted by any colleague preferrably by E-mail to my attention.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
[communicated by Maria Barysz]
Report on results obtained during two-months fellowship
granted to Maria Barysz within the REHE project.
Authors: Maria Barysz, Norbert Flocke, Jacek Karwowski, Geerd H. F. Diercksen
Place: Max-Planck-Institut fur Astrophysik, Garching bei Munchen
Time: May-June 1995
Subject: Developing methods for the simultaneous inclusion of
electron correlation and relativistic effects into the theory of
molecular structure
The project consists of two parts:
A Numerical Comparison Of Several Two-Component Relativistic
Approaches In Gaussian Bases.
Two invariants of the unitary transformation of the one-electron
basis set - the trace and the dispersion of the one-electron
integral matrices - have been calculated using several different
Gaussian basis sets with several
different two-component relativistic approaches to the theory of
molecular structure. The approaches considered include a
four-component kinetically balanced approach, the algebraic
no-pair approximation, the standard Pauli approximation and an
algebraic modification of the Pauli approximation. The results
confirm the high accuracy of the algebraic no-pair approach.
Relativistic one Component Coupled Cluster Method
in the Breit-Pauli and No-pair approximations.
A single reference Coupled Cluster (CCSD) code has been linked
to the OpenMol Program. Within OpenMol the CCSD method has been
implemented in several versions:
1. Non-relativistic CCSD with the Hartre-Fock reference function
2. CCSD with the relativistic no-pair Hartree-Fock orbitals
3. Breit-Pauli relativistic CCSD with the
non-relativistic Hartree-Fock orbitals.
4. No-pair relativistic CCSD with the non-relativistic
Hartree-Fock orbitals.
Only one-electron spin independent relativistic corrections have
been included at this stage.
In order to test this approach,some studies on the Ne and Ar isoelectronic
series are being carried out
This project will result in two research papers.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
[communicated by Jacek Kobus]
Account of a research carried out under REHE grant (No. 16-95)
entitled "Finite difference (Dirac)-Hartree-Fock method for diatomic
molecules" (as reported by J.Kobus and S.Wilson)
The research reported in this letter was carried out during Dr. Jacek
Kobus' stay in the Rutherford Appleton Laboratory, Chilton, UK (17
July - 15 September 1995). The projects were undertaken which may be
summarized as follow:
1. The molecular nuclear volume effect
As molecular structure calculations advance towards systems consisting
of heavy atoms the validity of using point-like nuclei needs
reconsidering. At the nonrelativistic level we were searching for a
molecular nuclear volume effect, i.e. the influence of the finite
nuclei charge distribution on the geometry dependence of the
properties of heavy diatomic systems. The Fermi nucleus model (with
the standard values of the "half-density" and "skin thickness"
parameters) was employed and the calculations carried out both within
the finite basis set and finite difference approaches.
The Sn-Zr^{+89} and Th_{2}^{179+} one-electron systems were studied.
Although the ground state energy of these systems was calculated to
within micro hartree accuracy (10 significant figures) no molecular
nuclear volume effect was detected for the bond distances 1 and 2
bohr. However, the kinetic and potential energy of these systems, when
monitored separately, did exhibit the molecular effect due to the
finite charge distribution. It is worth noting that the changes were
more pronounced for the Sn-Zr^{89+} than Th_2^{+179+}. It is expected
that the molecular properties which depend on the behaviour of
orbitals in the vicinity of nuclei would be influenced by the finite
nuclei charge distribution.
2. Graphical analysis of the finite basis set wavefunction.
Over the last couple of years we have put a lot of effort into the
construction of universal even-tempered Gaussian basis sets yielding
ground state Hartree-Fock energy of the BF, CO, N_{2}, AlF, GaF
molecules to within micro hartree accuracy. In order to assess the
quality of the basis sets used we examined (both graphically and
numerically) the differences between the BF basis set molecular
orbitals and their finite difference counterparts. The behaviour of
the local energy was also studied. It has turned out that the local
energy can be used a as good measure of quality of molecular orbitals
and hence of their finite basis set representation. It is hoped that
this measure may be of great help when assessing basis sets for
systems for which the finite difference HF results are not available
for comparison.
As a useful by-product of this investigation, the results of
GAUSSIAN94 calculations can now be fed into the finite difference HF
program providing a good initial guess of molecular orbitals and
potentials. This is extremely important from the point of view of the
very cpu and memory intensive finite difference HF calculations for
the InF and TlF systems which are planned.
3. Improvements to finite difference diatomic HF code
The finite difference HF code has been reconstructed to reduce
considerably the memory requirements of the program (at the expense of
increased I/O operations) and thus to make calculations of the InF and
TlF systems feasible within available computer resources. The program
can also undergo parallelization since in every self-consistent-field
iteration there are groups of exchange potentials that can be relaxed
simultaneously. The Parallel Virtual Machine software was investigated
for this purpose and a scheme for the parallelization of the code was
worked out.
Once the highly accurate nonrelativistic (finite basis set and finite
difference) HF calculations for heavy molecular systems are mastered,
it will be possible to address with confidence the problem of solving
the Dirac-Hartree-Fock equations for such systems.
One of us (JK) would like to thank the REHE Steering Committee for the
prompt reply to the application in early June which made his summer
visit to RAL possible.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
[communicated by Jaap Snijders]
Short report on the visit of J.G.Snijders to A.J.Sadlej (university of
Lund, Sweden), from august 19 1995 till september 17 1995.
The relation between Direct Perturbation Theory and the CPD/ZORA method
studied by us earlier was further clarified. It was shown that it is
possible to formulate the ZORA and FORA method in a perturbative way with
the non-relativistic Hamiltonian as a zeroth order starting point. However,
test calculations on a hydrogen like ion with high nuclear charge, where
everything can be worked out analytically, show that it is in general not
sufficient to take into account the ZORA Hamiltonian in a perturbative way.
In particular, non-analyticity near the nucleus that the exact ZORA
wavefunction shares with the (large components of) the Dirac solution, and
which are only taken into account in higher order perturbation theory,
turns out to be crucial in obtaining energies that are sufficiently
accurate, in particular for s orbitals. The effect of the behaviour near
the nucleus is even more important if the FORA corrections are calculated.
We have also analyzed in detail to what extent the Pauli (order c-2)
corrections are included in ZORA (apart from the higher order contributions
that are responsible for the regularization). It turns out ZORA contains
the complete Darwin and Spin-orbit corrections, but a small part of the
mass-velocity correction is absent and is only (completely) recovered at
the FORA level. The error is, however, shown to be much smaller in a
neutral atom than in a hydrogen-like ion, explaining the great accuracy of
ZORA for the valence levels of neutral atoms.
The results of the investigation have been written up in a publication that
will shortly be submitted
Jaap G. Snijders and Andrzej J. Sadlej
Perturbation vs. variation treatment of regular relativistic hamiltonians
Chem. Phys. Lett. (to be submitted)
We have also started to formulate a way in which the so called "scaled
ZORA" (or CPD-4) method could be used in a many electron ab-initio setting
with a view towards relativistic calculations at the CI or MCSCF level.
However, the results are still somewhat premature, and work on this problem
will continue in the near future.
J.G.Snijders
A.J.Sadlej
prof. dr. J.G.Snijders
Department of Theoretical Chemistry
Vrije Universiteit
De Boelelaan 1083
1081 HV Amsterdam
Nederland
Tel. direct (31)-20-4447624
Tel. secr. (31)-20-4447644
Fax. (31)-20-4447643
E-mail: snijders@cchem.vu.nl
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
[communicated by Jacek Styszynski]
Report on the research stay in the Laboratory of Chemical Physics
(Prof. W.C. Nieuwpoort) of the University of Groningen
The package MOLFDIR developed by Groningen group is an advanced
suite of programs for studying electronic structure of molecules including
relativistic and correlation effects. In 1993 a relativistic variant of the
restricted active space configuration interaction (RASCI) method and most
recently, relativistic versions of second order Möller-Pleset perturbation
theory (MP2) and coupled clouster with single and double (CCSD) excitations
and CCSD corrected for the effect of triple excitations (CCSD(T)) methods
were added to this package.
Within the frame of this project relativistic and correlation
effects were studied in hydrogen halide molecules (HX, X=F, Cl, Br, I, At).
In our study we employed two types of basis sets: polarised valence
double-zeta (pVDZ) and triple-zeta (pVTZ). For each molecule the
calculations at seven different internuclear distances H - X were performed
using DF, RASCI, MP2, CCSD and CCSD(T) methods. Additionally , the basis
set superposition error (BSSE) was estimated for these calculations using
the counterpoise method. The obtained potential curves were fitted to
fourth order polynomials, giving the H - X equilibrium separation re and
fundamental harmonic frequency we. The dissociation energy De was
calculated as a difference between the sum of atomic energies and the
molecular energy. In order to estimate the effect of relativity on the
spectroscopic constants, corresponding nonrelativistic calculations were
also performed (using the two-component option in MOLFDIR package).
The relativistic effects are seen clearly in the decrease of the
dissociation energy of the molecule (about 1/3 of the valence p spin-orbit
splitting), which in case of HI is about 7.5 kcal/mol. The relativistic
effect on frequency is also important, giving the shift of ~40 cm-1 for HI.
Finally, a slight relativistic contraction of a bond is found (~0.005Å for
HI). Reported calculations were performed on the CRAY J916 at RUG Groningen
and the CRAY YMP-C98 at SARA Amsterdam. A paper discussing results of those
calculations is in preparation.
Acknowledgement.
I would like to thank Prof. W.C. Nieuwpoort for accepting the idea
of my stay in his group, help and great hospitality during my 3-month
stay. I am very grateful to Dr. Luuk Visscher for suggesting the subject of
my research stay in Groningen, many illuminating discussions, help and
hospitality. Thanks are also due to Bert de Jong for many discussions and
help with MOLFDIR package. The computational grant from the Netherlands
National Computing Facilities Foundation (NCF) is kindly acknowledged.
Jacek Styszynski
Lab. of Chemical Physics
University of Groningen
Nijenborgh 4
9747 AG Groningen
The Netherlands
e-mail: jaceks@chem.rug.nl
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
================================================================================
--- P A P E R S F U N D E D B Y R E H E
>>> please send a preprint of papers funded by REHE to Bernd A. He\ss,
>>> Institut f\"ur Physikalische und Theoretische Chemie, Universit\"at Bonn,
>>> 53115 Bonn, Germany
================================================================================
--- C O N F E R E N C E N E W S
'Conference News' (in general they should NOT overrun about 1 page)
may be provided by organizers or their scientific secretaries. --
For meetings and workshops supported by ESF the submission of such
a report is a m u s t . To facilitate my job the reports should
be forwarded to my attention via E-mail.
Also please send information about conferences that might be of interest
for the members of the REHE community.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
******************************************************************
**** ****
**** Winter School in Theoretical Chemistry ****
**** ****
**** CALCULATION OF NMR PARAMETERS ****
**** ****
******************************************************************
TIME: 11-14 December, 1995. (Start at 1030 on Monday)
VENUE: Department of Chemistry, University of Helsinki,
A.I. Virtasen aukio 1, Helsinki, Finland.
LANGUAGE: English.
LECTURERS: U. Fleischer (Bochum): IGLO and other applications
J. Gauss (Karlsruhe): Advanced correlated methods
J. Jokisaari (Oulu): Experimental determination of NMR
parameters (2h)
M. Kaupp (Stuttgart): DFT and NMR parameters
W. Kutzelnigg (Bochum): Magnetic susceptibilities
J. Oddershede (Odense): Basic theories of magnetic properties (6h)
H. Nakatsuji(Kyoto) : Spin-orbit effects
P. Pyykko" (Helsinki): Early treatments of relativistic effects
(1h); Nuclear quadrupole effects (1h)
T.K. Rebane (St.Petersburg) (*)
K. Ruud (Oslo): Variational (MCSCF) methods
(*) Not yet confirmed.
The program will consist of a trunk course by Oddershede and of
particular subjects, treated by the others.
CONTRIBUTED Any member of the School is welcome to put up
PAPERS: a poster on the premises.
PROCEEDINGS: None.
LEVEL: Post-graduate and research, for scientists already involved in
computational quantum chemistry.
The course will give 2 credits.
ADMISSION: The participants are requested to register
by 15 November, 1995 to the address below. The sooner you can
do it, the better, due to room reservation problems.
LOCAL
SECRETARY: Mrs. Susanne Lundberg, Dept. of Chemistry,
P.O.Box 55 (A.I. Virtasen aukio 1),
00014 University of Helsinki, Finland.
Telephone: 358-0-191 40170.
FAX: 358-0-191 40169.
E-mail: Susanne.Lundberg@helsinki.fi
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
================================================================================
--- P O S I T I O N S available
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
[no material for this section in the current newsletter]
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
================================================================================
--- P O S I T I O N S sought
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
[no material for this section in the current newsletter]
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
================================================================================
--- ADDRESS LIST
The REHE address list comprises 170 scientists as of Oct 9, 1995; the
next address list will be provided with newsletter no. 16
In order to join the REHE mailing list, please complete the form below
>>> PLEASE include TEL, FAX, E-MAIL <<<
=================================================================
I am interested in receiving the REHE newsletter
NAME
-----------------------------------------------------------------
ADDRESS
-----------------------------------------------------------------
-----------------------------------------------------------------
TEL
-----------------------------------------------------------------
FAX
-----------------------------------------------------------------
E-MAIL
-----------------------------------------------------------------
MAIN RESEARCH INTERESTS
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
End of REHE Newsletter No. 15