Dr. Darío M. Mitnik
UBA: Departmento de Física, Universidad de Buenos Aires, Argentina   +541 (1) 4-576-3390 (#802)
IAFE: Instituto de Astronomía y Física del Espacio, Buenos Aires, Argentina   +541 (1) 4-788-1916 (#231)
e-mail: dmitnik@df.uba.ar

 
RESEARCH EXPERIENCE AND INTERESTS

I have performed research in several aspects of theoretical atomic physics and plasma physics. Following is a brief description of my research experience.

Plasma modeling: My doctoral dissertation was a study of Excitation-Autoionization processes in the Ni-like to Kr-like isoelectronic sequences of ionized heavy elements and the effect on the fractional ion-abundance. During this work, I worked extensively on the problem of ionization state distributions in plasmas, and on the implementation of a collisional-radiative model for the population of complex ions. This research led to understanding of the very complex problem of irregularities in the plasma spectra emitted in Tokamak experiments. I was also working on the spectral modeling of heliumlike ions, in particular, trying to solve the problem of the anomalous experimental Kbeta/Kalpha line ratios obtained from Tokamak and EBIT devices.

Electron-ion scattering: I am involved in calculations of electron impact excitation, ionization and recombination. The theoretical challenge in the determination of accurate cross sections is multifold, and varies according to the different ions. Therefore, many approaches are needed in order to calculate these atomic processes using the appropriate method for each case. In my research, I am using many of these approaches, like distorted-waves, close-coupling, configuration-averaged and time-dependent methods.

Electron-ion recombination: I made theoretical calculations of very complicated dielectronic recombination spectra. Our theoretical calculations of the O-shell dielectronic recombination of U28+ agrees very well with the experimental results. This is one of the most complicated dielectronic spectra published to date. In many experimental recombination spectra, unusually large rate coefficients have been found at low energies. I am investigating the source of these enhancements. My research includes the investigation of interference effects between the radiative and dielectronic recombination, through the searching of unusual asymmetric resonances. I am also participating in research on field enhanced dielectronic recombinations and on the effect of overlapping resonances. I performed sophisticated calculations of dielectronic recombination of highly-ionized ions. The high resolution achieved in current experiments requires the presentation of the theoretical results detailed at the level of natural linewidths. In particular, for highly-ionized heavy ions, fully relativistic calculations, including Breit and quantum electrodynamic effects, are needed.

ADAS Project: I am involved in data generation for the ADAS (Atomic Data and Analysis Structure) project. This is an interconnected set of computer codes and data collections for modeling the radiating properties of ions and atoms in plasma and for assisting in the analysis and interpretation of spectral measurements. This project includes calculations on electron-ion excitation, ionization, and recombination cross sections and rate coefficients.

Time-dependent methods: I am working with different methods that solve the time-dependent Schrodinger equation on a partitioned lattice. This technique has been employed in the Auburn University atomic physics group for solving many different problems ( e.g. electron impact ionization, photoionization, electron-impact double-ionization, and even to solve the Gross-Pitaevskii equation for the calculation of Bose-Einstein condensates). The time-dependent close-coupling lattice method solves directly the time-dependent Schrodinger equation by the partition of the radial wavefunction over the many processors found on a massively parallel computer. The wavefuncions are evolved in time using explicit second-order differencing. I am currently exploring different algorithms for the time-propagation, the possibility of its implementation on parallel computers, and also the best use of the dual shared and distributed memory environment, typical for the next generation of supercomputers.

Parallelization of the R-matrix package: The R-matrix method is an extremelly successful theoretical approach, that has been used in atomic and molecular physics for almost forty years. However, this method has a severe drawback in what is known as ``stage 3", the matrix diagonalization of the electron-ion Hamiltonian. The size of the matrices involved in these calculations requires extremely time-consuming calculations, beyond the capabilities of regular computers. I finished an adaptation of the STG3 R-matrix program (from Queen's University at Belfast), in order to run it on the distributed-memory parallel Cray T3E-900 supercomputer, and also on the IBM SP supercomputer at NERSC (National Energy Research Supercomputer Center) at Oakland CA, USA. The diagonalization of the total Hamiltonian is carried out using recently developed distributed memory numerical software, as found in the ScaLAPACK library. Now the diagonalization calculation can be distributed on different nodes, producing a substantial decrease in wall-clock turnaround. Recently, I successfully finished an adaptation of the further steps in the R-matrix package for parallel computers. This include the STGF code (for the outer-region), and STGIC code (for generation of collision strengths in intermediate coupling).

R-matrix theory: I am participating in the development of a ``toy" R-matrix code. This program solves the electron-hydrogenic ion scattering under the simplifying assumption (known as the Temkin-Poet model) that the coordinate wavefunction is spherically symmetric with respect to both projectile and target electron positions. Although the model ignores states having non zero values of the angular momentum, the model is still complex enough to include some effects that make the full problem difficult to solve. It includes strong coupling between an infinite number of open channels and requires an adequate representation of the continuum states. This work on the toy program will enable us to study some of the general features of the R-matrix theory.

Computational physics: I developed and participated in the development of many computer codes for atomic physics calculations. I achieved an expertise in the use of the relativistic computer code package HULLAC (Hebrew University - Lawrence Livermore Atomic Code). I am familiar with the use of many other atomic physics codes, including the MCHF atomic structure package by C. Froese Fischer, Cowan's atomic structure program, and a modified version of the QUB R-matrix suite. I have also a broad experience in different computer systems, from desktop workstations to parallel supercomputers.

Experimental work: While finishing my Ph.D, I did experimental research at the 14UD Pelletron accelerator at the Weizmann Institute, Israel. I completed a short Postdoctoral project under the supervision of Prof. M. Paul. I joined his group working on the measurement of electron affinity and absolute photodetachment cross section of negative ions, by using Laser Excitation and Accelerator Mass Spectrometry (LAMS) technique.

Complete list of Publications

 

Talk: Atomic Physics with Supercomputers

Talk: Parallelization of the R-matrix package

The Parallel RMTRX-I Package User's Guide

 
 
VITA PUBLICATIONS RESEARCH CODES LINKS PERSONAL