法国国家科研中心(CNRS)博士后研究项目

Institution

Le Centre National de la Recherche Scientifique (CNRS), France 

Position

Post-doc in TDDFT theory, south-west France

Location

Bordeaux and Pau, south-west France

Date Posted

Sep 28 2007

STARTING DATE

January 2008

DURATION

1 year, renewable twice

Salary

NET PAY:ca. 2100euros

Description

Two post-doc and one PhD positions are open as part of an ANR-funded project in south-west France: NOSSI, ‘Nouveaux outils pour la simulation des solides et interfaces’ (‘New tools for the simulation of solids and interfaces’).

This advertisement is for POST-DOC 1:

CALCULATION OF OPTICAL SPECTRA OF LARGE MOLECULES FROM TDDFT LINEAR RESPONSE, USING A NEW ALGORITHM

HOST UNIVSERSITIES: University of Bordeaux 1, University of Pau, France

HOST LABORATORIES: Centre de Physique Moléculaire Optique et Hertzienne (CPMOH-UMR5798 CNRS and UB1) and
Institut pluridisciplinaire de recherche sur l’environnement et les matériaux (IPREM-UMR5254 CNRS and UPPA)

RESEARCH PROJECT:

This position is to finalize deployment of a novel O(N3) TDDFT code for the calculation of molecular optical spectra using the linear response equations of Gross, Petersilka, Gossmann (1) rather than Casida¿s equations (2) that are used and implemented in most available DFT codes. Part of our motivation is that these equations are more convenient for systems with boundary conditions, such as molecules adsorbed on surfaces.

A new algorithm developed by D. Foerster eliminates the need of a severe approximation made in an earlier attempt to solve these equations by Blase and Ordejon (3) while avoiding combinatorial explosion by a mathematical technique that is new in this context (4). Acceleration compared to existing O(N3) codes will be explored in calculations starting with the reference compound benzene and continuing with progressively larger molecules such as cyano-anthracenes and dyes like indigo and naphthol yellow, comparing throughout with experimental data and checking scaling with system size.

The candidate will participate in the last stages of code development and optimization which may involve recoding J.Talman’s fast Hankel transform (5) that plays an essential role in this project. The code will be interfaced with the SIESTA package, implying optimisation and parallel programming. These tasks will be carried out in contact with, respectively, one of the originators of the siesta code (D. Sanchez-Portal at DIPC San Sebastian, Spain) and with computer scientists and applied mathematicians at Bordeaux (O. Coulaud) while making use of the ~ 170 parallel processors that are available at the local computer center.

Because of the rather general caracter of the concepts and mathematical techniques to be used, experience gained by the candidate in the context of this project should be transposable to, and useful in, other fields of computational materials science.

PREVIOUS EXPERIENCE: Candidates should have a strong background in scientific programming and density functional theory. Experience with parallel programming is not necessary, but will be greatly appreciated.

CONTEXT:

The overall project NOSSI project aims to:

-Couple dl_poly and Siesta to provide a QM/MM package for solids and interfaces.
-Introduce a new TDDFT algorithm for electronic spectra.
-Provide tools for visualisation and steering of running, coupled, parallel dl_poly+Siesta jobs
-Provide new molecular mechanics force-fields for simulation of adsorption in zeolites or on surfaces.

ENVIRONMENT:

The university of Bordeaux is recognized for its excellence in several fields relevant to this project: optical spectroscopy, molecular and solid state chemistry and meso- and nanostuctured materials. In the domain of applied mathematics, it is renowned for its development of computational algorithms. The university of Pau has proven expertise in the fields of materials and surface characterization and solid state quantum chemistry. The position offers opportunities for cross-disciplinary research in a well-equipped and pleasant working environment. Candidates will be expected to interact closely with colleagues providing experimental data for the validation of the calculations. Travel between the participating centres (Pau, Bordeaux) will be required.

REFERENCES:

1 M. Petersilka, U. J. Gossmann, and E. K. Gross, Phys. Rev. Lett. 76, 1212 (1996).

2 M.E. Casida, in Recent Advances in Density Functional Methods, Part I, edited by D.P. Chong (Singapore, World Scientific, 1995), p. 155; Christine Jamorski, Mark E. Casida, and Dennis R. Salahub, J. Chem. Phys. 104, 5134 (1996).

3 X. Blase and P. Ordejon Phys. Rev. B 69 (2004) 085111.

4 D. Foerster, to be published in Journal of Chemical Physics.

5 J. Talman, J. Comp. Phys. 29, (1978) 35; Comput. Phys. Comm. 30 (1983), 93.

Contact

I. Baraille, isabelle.baraille@univ-pau.fr, Tel : +33 (0) 55940 7845
D. Foerster, d.foerster@cpmoh.u-bordeaux1.fr, Tel: +33 (0) 54000 2507 (after October 1st).