 |
|
 |
||||||
|
||||||
 |
Home
| Mission
|
about SciDAC
|
Contact Us |
|||||
 |
 |
 |
||||
|
Alumni ProjectRelativistic Quantum Chemistry SoftwarePI:Russell M. PitzerBruce E. Bursten Isaiah Shavitt Department of Chemistry Ohio State University SummaryRelativistic quantum chemistry software is very helpful for understanding spectroscopic data, energetic data, and reaction rate processes of chemical species containing heavy atoms. The development of significantly more capable software together with next-generation hardware should make theoretical calculations much more useful in studies of heavy-element species. The software must have effective scaling in its ability to (1) handle larger chemical species, and (2) make effective use of many computer processors. Background InformationAffiliated SciDAC researchers are Walter C. Ermler and Maria Marino, University of Memphis, working on developing relativistic effective atomic core potentionals, which provide a more effective and flexible way to treat the electronic cores of heavy atoms.Collaborators on our basic quantum chemistry codes (Columbus suite) are Hans Lischka (University of Vienna) and Ron Shepard (Argonne National Laboratory). A server with software for coordinating code development is being established at the Ohio Supercomputer Center. During 2002, the Vienna group contributed a parallel version of the Columbus programs which included relativistic effects, but the latter were only partly tested. Porting capabilities to other machines were also unfinished. Thus we are working on the porting to other types of computers and will then continue and expand the testing of relativistic calculations. The plan to make the programs scalable to larger molecules or ions is to convert current memory allocation methods to Fortran 90 form, updating a number of other features as well. Example of current application (supported by DOE): The UO2 molecule (radical) has two unpaired electrons, is quite reactive, and therefore is difficult to observe. It has been studied in an inert matrix by the L. Andrews group at the University of Virginia and in the gas phase (laser ablation) by the M. Heaven group at Emory University. It is linear in the ground electronic state and presumably also in its excited electronic states because the two unpaired electrons are the ones excited and they play little role in the bonding. Earlier calculations at the University of Toulouse showed that the ground electronic state is (5f)1(7s)1. Some time ago, the Heaven group found a transition a t 29,500 cm-1. Our calculations, with a present version of the Columbus programs, were limited to wave function expansions of 1.0 to 1.5 million terms due to our spectroscopic analysis programs. Our results so far show that the excited state mentioned is due to a 7s ® 7ps transition and is the 31st excited state. We looked for lower-energy transitions of 7s ® 7pp character because they should have appreciable intensity also. We found them in the 17,000 cm-1 to 20,000 cm-1 range. A minimum of two such states were expected, but the intensity was spread to additional states due to spin-orbit mixing. Some states in this region have now been observed by the Heaven group. We calculated lower intensities to a few other states among the 32 we studied, but these involved transitions where the major component of one of the wave functions did not contribute to the transition moment. The excited states in this region are of the (5f)1(7p)1 and (5f)2 types, with considerable mixing between them. The 5f orbitals in these two types of wave functions are not the same because the there is more shielding of the U nucleus in the second type than in the first type. Thus most of the wave functions have considerable multireference character, making the need for flexible wave function methods quite apparent. The version of Columbus software under development should add considerable accuracy to this type of work and extend the size of molecules or ions to which it is applicable. Other examples being developed, chosen for a range of types of molecules or complexes are (1) the interaction of B10H14 with I- (known complex) and with Xe (unknown complex), and (2) the intensities of lanthanide transitions in crystals, such as Er3+ doped into GaN. For further information on this subject contact:Russell M. Pitzer
|
||||
Home | ASCR | Contact Us | DOE disclaimer |
|
|
