WORKSHOP "ADVANCES IN FIRST-PRINCIPLES COMPUTATIONAL CONDENSED MATTER PHYSICS"
Sponsored by
The European Science Foundation Programme STRUC-Psik
Universidad Autonoma de Madrid
Universidad del Pais Vasco
Spanish Consejo Superior de Investigaciones Cientificas
Motorola PSRL and FPD
Organised by
Emilio Artacho, Universidad Autonoma de Madrid (Spain)
Alberto Garcia, Universdad del Pais Vasco (Spain)
Pablo Ordejon, ICMAB-CSIC (Spain)
Scientific Committee
A. Baldereschi, EPFL, Laussanne (Switzerland)
S. Baroni, SISSA, Trieste (Italy)
M. Gillan, University College of London, London (UK)
K. Kunc, CNRS, University P. and M. Curie, Paris (France)
S. Louie, University of California - Berkeley (USA)
R. M. Martin, University of Illinois at Urbana- Champaign (USA)
M. Methfessel, Inst. fur Semicond. Physics, Frankfurt-Oder (Germany)
R. Needs, Cambridge University, Cambridge (UK)
M. Parrinello, MPI-FKF, Stuttgart (Germany)
K. Terakura, NAIR, Tsukuba (Japan)
D. Vanderbilt, Rutgers University, Rutgers (USA)
E. Wimmer, Molecular Design s.a.r.l., Le Mans (France)
This meeting is related to the series held regularly (in January in odd-numbered years) at the Abdus Salam International Center for Theoretical Physics in Trieste (Italy). These rather large ("MAXI") workshops were originally labeled as dealing with "Total Energies and Forces Methods", and currently with "Electronic Structure Theory and Simulations". In even-numbered years, a smaller ("MINI") workshop is held at different venues. The last MINI was held in Tsukuba, in January 1998. Following the tradition of the previous MINI's, the Madrid workshop was devoted to recent advances in computational condensed matter physics, based on realistic calculations of the electronic structure of polyatomic systems.
The workshop took place in "La Cristalera", a conference center owned and maintained by the Universidad Autonoma de Madrid, and located in Mirafiores, about 40 km north of the city. Transportation to and from the conference center was arranged by the organization. The center provided a quiet environment, with meals and accommodation at the same building, which facilitated discussion and contact.
The workshop was an eventful one. The heaviest snowfall in fifteen years caused a power outage in the Northwest of Madrid province that affected the workshop venue. This forced the relocation of the attendees and a delay of several hours of the start of the Friday session. The program was completed (except for two talks) by tightening the schedule.
The format consisted in 19 invited talks of 40 minutes plus 10 minutes of discussion (after the relocation, the time allocated for talks was reduced to 25-I-5 minutes). Besides the invited talks, there was a display of contributed posters. (There was no specific poster session, but there was time for discussion in between and after the talks.) While waiting (vainly) for the power to be restored, a round table was organized. Among the issues raised were the pros and cons of the use of ab-initio program packages and the possibility of building a public-access repository of routines. It was agreed to work on the idea and make it concrete at the next Trieste meeting, in January 2001.
The workshop was organized around four main topics:
1. Density-functional Molecular Dynamics and Monte Carlo
Mike Gillan gave a review of the recent work on the calculation of free energies of solids and liquids, aimed at the determination of the melting curves of solids, and the new developments to calculate chemical potentials in solutions. He also described the application to the physics of the inner Earth's core. Rafael Ramirez was invited to talk about the recent developments in using Path Integral Monte Carlo simulations to obtain dynamical properties and excitation energies of quantum systems. However, his talk was one of the two that were canceled due to the power problems. Richard Needs described methods to calculate excited electronic states by means of fixed-node diffusion quantum Monte Carlo. He discussed problems associated with these calculations, in particular the lack of a variational principle of general validity for the excitation energies.
2. Theory of magnetic and electronic polarization
No Souza gave a talk about the relation between electronic localization and polarization in insulators. Using a generating function approach he showed how these quantities are intimately related, and gave nice connections to the well known Kohn's theory of the insulating state, in terms of the decomposition of the many-body wave function in localized (disconnected) wave functions. Stefano Baroni described a method to study the spin dynamics in ferromagnetic systems, by means of an adiabatic decoupling of the spin and charge excitations. The method provides an excellent description of the magnon spectrum of iron, and should be much more efficient than other current linear response methods (although it cannot describe Stoner-type excitations). Warren Pickett talked about the recently discovered coexistence of ferromagnetism with singlet superconductivity in ruthenocuprates. He argued about the origin of this coexistence, in terms of the spatial separation of the orbitals responsible for the magnetism (Ru t2g) and for the superconductivity (Cu dx2-y2).
3. Time Dependent DFT
Xavier Gonze discussed on the conceptual aspects of TD-DFT, concerning the formalism (and its connection to many-body perturbation theory), and the use of different potentials and kernels. He showed results for the excited states potential energy surface of (HeH)+, and showed that, comparing to CI results, the exchange-only Optimized Effective Potential performs considerably better then the LDA one. Kazuhiro Yabana presented a method to calculate the optical response of finite systems, by integration of the the time-dependent Kohn-Sham equations in real-time and space. The quadratic scaling of the method with the size of the system allows its use in considerably large systems, with a few hundreds of electrons. Kieron Burke also discussed about the basic approximations of TD-DFT and their accuracy, and on the performance of different XC kernels.
4. Applications
Several applications of first principles calculations were presented in the workshop. An important number of them corresponded to materials related problems, but other fields were also covered.
With regards to materials, several talks were delivered. Erich Wimmer gave a presentation on the combination of electronic structure methods with large-scale database analysis to understand and predict the properties of materials, and presented some commercial tools in this direction. Chris van de Walle presented calculations of defect and impurity states in III-nitride semiconductors. He addressed the issue of the limitations of DFT for these calculations, and showed how these can be overcome with the use of self-interaction and relaxation-corrected pseudopotentials (although the ability of computing the total energy is lost). Magdalena Siodmiak discussed schemes to tackle the complex problem of CVD growth, with the aid of first principles total energy calculations. Nathalie Vast showed calculations of the structure and vibrational properties of icosahedral B4C, and detailed comparison with experimental data. David Singh presented a talk on the use of DFT calculations in the search for new thermoelectric materials. He described the properties which a good thermoelectric material should have, and ideas on how to engineer materials with these properties. The main problem is that these properties (mainly, low thermal conductivity and high electrical conductivity) are to a large extent contradictory. He showed how first principles theory can help in optimize the performance of designed materials. Ruben Perez was invited to talk about tip-surface interactions on semiconductor surfaces, but his talk was cancelled.
There were also several talks on applications to problems other than those of traditional materials science. Jose M. Soler and Daniel Sanchez-Portal talked about the recent applications of order-N LCAO DFT calculations to several complex systems: DNA, liquid silicon surfaces, zintl manganites and monoatomic gold wires. Carme Rovira presented her work on the modelization of biological molecules, and in particular, in HEMS systems. She showed results on the ligand binding properties of myoglobin, with special emphasis on the different behavior of several ligands (O2, CO and NO). Bjorn Winkler discussed the usefulness of first principles DFT calculations in Mineralogy and Crystallography, in particular in the prediction of pressure and temperature-induced structural changes of minerals.
More details may be found in newsletter 37 from page 14