|Posted on May 21, 2015 at 9:50 AM|
In the last years we have been working together with Josep Maria Luis on the ability of density functional theory (DFT) to reproduce nonlinear optical properties (NLOP). We started by studing the so-called catastrophe of DFT to reproduce polarizabilities and second-order polarizability on long conjugated chains as the size of chain increases. We supervised together the Bachelor Thesis of Natalia Abulí and recently the Erasmus stay of Sebastian Sitkiewicz, both devoted to this research topic.
DFT has found applications on a wide variety of scientific areas due to its remarkable combination of efficiency and accuracy. It is being applied with success to the fields of bioinorganic chemistry, material science, drug design, biochemistry and nano-technology, among many others. The exact density functional expressions of many energy components have not been found, and as a result, the construction of new functionals in DFT has become a complicated task, often untangled by the recourse of fitting parameters with the aid of experimental results. In this regard, the errors in DFT calculations are hardly predictable and for each new scientific challenge functionals must be calibrated against the expensive standard ab initio methods to assess their performance. DFT has reached a state of saturation, and the design of new strategies for constructing DFT functionals is now of utmost importance.
Last year we put together an ambitious research project to construct new DFT functionals that do not suffer from the above-mentioned drawback. We based the strategy on the same idea I suggested for DFTCorr, which combined with Josep Maria outstanding expertise on NLOP, resulted in the research projecte entitled: "DFT functionals for the calculation of nonlinear optical properties" (acronym: NLOPDFT). This project was submitted to 2014's call "Generación de Conocimiento" of the Spanish Ministry (MINECO) with the following team: Sebastian Sitkiewicz, Mauricio Rodríguez-Mayorga, Eloy Ramos-Córdoba, Marc Garcia-Borràs, Josep Maria Luis and myself.
The goal of NLOPDFT is to use a genuinely new strategy to design density functionals for the calculation of NLOPs, which will lead ultimately to an all-purpose functional that yields reasonably accurate results in most applications. This strategy is physically motivated and consists in using variable amounts of the components of the exchange-correlation functionals regulated by electron correlation measures that enter into the functional expressions. Overcoming one of the most important DFT pitfalls —the description of NLOPs—, promises to furnish new functionals with the flexibility to accurately describe a wider range of properties, paving the way towards the development of all-purpose functionals. This project gives the recipe to construct an unprecedented class of local hybrid functionals and range-separate hybrids with more flexibility than its predecessors. Our preliminary results following this strategy show a drastic improvement of hybrid functionals. This project shall leap forward DFT development and make an impact in a number of fields where DFT has found applications.
A few weeks ago, the MINECO decided to grant our project with 60,000€ and a PhD fellowship. Therefore, there is an opening for a PhD position to work in this project. The PhD fellowship is a four-year grant including a free one-year master in the MaCMOM master organized by the IQCC at the University of Girona. Interested candidates should have an excellent academic track record, preferably with a good background on computational chemistry. Prospective candidates should send their CV by e-mail to either JosepM ([email protected]) or myself ([email protected]).
[Polarizability and second-order polarizabilities in polyacetylene chains. From: Limacher et al. JCP 130, 194114 (2009)]