AmericanChemicalSociety.com

Our research project focuses on accurate theoretical prediction of the structural, electronic, and optical properties of core-shell semiconductor nanocrystals, such as CdSe/CdS, CdSe/ZnSe, and CdSe/ZnS systems. The study of core-shell nanostructures is carried out using first principles density functional and time-dependent density functional methods implemented on a parallel computational platform. The goal of this project is to expand the knowledge of chemistry and physics of semiconductor heterostructures and accelerate the development of composite materials for efficient conversion of solar energy.

Major research activities during the first year of the project included testing the accuracy and efficiency of the PARSEC (Pseudopotential Algorithm for Real-Space Electronic Calculations) and SIESTA (Spanish Initiative for Electronic Simulations with Thousands of Atoms) density functional electronic structure codes in application to core-shell semiconductor particles, structural optimization of traditional (CdSe core and CdS, ZnSe, or ZnS shell) and inverted (CdS, ZnSe, or ZnS core and CdSe shell) heterocrystals, and the development of an efficient massively-parallel time-dependent density functional computational algorithm in the framework of the PARSEC code for computing the optical properties of core-shell nanocrystals.

The optimization of equilibrium geometries of core-shell nanoparticles presents considerable challenges for density functional computational methods. These challenges are mainly related to the lack of a simple surface passivation scheme for the group II-IV semiconductor nanocrystals and to the differences in the lowest-energy crystal structures of different group II-IV semiconductors. For the outer surfaces of the CdS, ZnSe, ZnS, and CdSe shells, we used a simple but efficient surface passivation technique based on fictitious partially charged hydrogen atoms. The value of fractional charge was selected according to the type of covalent bond. For the group II-IV semiconductors, we used fictitious hydrogen atoms with a nuclear charge of 0.5 e and 1.5 e. We found that this approach completely removes the electronic states associated with the surface hydrogen atoms from the gap of semiconductor heterocrystals. The optimized geometries of core-shell nanocrystals were obtained by considering different types of crystal lattice for the core and shell semiconductors and selecting heterocrystals with the lowest total energies after structural relaxation. We found that the lattice symmetry of the group II-IV semiconductors in a core-shell nanoparticle did not always coincide with the most stable form of the bulk crystalline lattice. For example, while the sphalerite CdSe structure is unstable and converts to the wurtzite form in the bulk, CdSe/ZnSe core-shell nanocrystals based on the sphalerite form of CdSe were found to be lower in energy than those based on the wurtzite form of CdSe.

The accuracy of the PARSEC and SIESTA electronic structure codes in application to semiconductor heterocrystals was tested by comparing the structures, total energies, and eigenvalue spectra of small optimized sore-shell nanoparticles obtained with PARSEC and SIESTA to those computed using common quantum chemistry program packages, such as Turbomole and Gaussian 2003. In most cases, the energies and structures of semiconductor nanoparticles calculated using the PARSEC and SIESTA codes were in good agreement with those obtained using the Turbomole and Gaussian 2003 packages. These calculations were done in collaboration with the Theoretical Chemistry Group at Los Alamos National Laboratory (LANL). The research collaboration between the PI and LANL was supported by the Center for Integrated Nanotechnologies (CINT).

Due to the complex structure and large size of core-shell nanocrystals, time-dependent density functional calculations for the electronic excitation energies and optical spectra of these systems require the use of an efficient parallel computational algorithm. To address this problem, we developed a massively-parallel version of the time-dependent density functional code and implemented it in the framework of the PARSEC electronic structure program. The developed code uses symmetry operations to reduce the size of the Hamiltonian matrix and the number of matrix elements calculated within the time-dependent density functional algorithm. The accuracy of time-dependent density functional calculations was improved by using asymptotically correct exchange-correlation functionals within the PARSEC code. This work was done in collaboration with the Center for Computational Materials at the Institute for Computational Engineering and Sciences and the University of Texas at Austin (UTA) during the PI's sabbatical leave from September 2008 through May 2009.

The collaborative partnership with the Center for Computational Materials at UTA and the Theoretical Chemistry Group at LANL created through this project helped advance the PI's research program at New Mexico State University (NMSU). The PARSEC program package represents one of the primary computational tools used by the PI's research group at NMSU. The new computational algorithms developed in collaboration with the Center for Computational Materials at UTA will help the PI maintain a competitive advantage over other research teams working in this field. The research conducted in connection with this project has generated several peer-reviewed publications in high-ranked U.S. and international physical and chemical journals. The results of this study will be used to attract additional external funding and further expand the PI's research program at NMSU.

This research project has also benefited graduate students at NMSU. The major part of the project budget is used to support NMSU graduate students through stipends and benefits. During the first year of the project, two graduate students were involved in research related to this proposal. The research carried out under this project will serve as a basis of Ph.D. dissertations for these two students.