Biography
Affiliated researcher in the Multiscale Dynamics group (December 2016 - present). Previously employed at International Atomic Energy Agency (IAEA), Vienna, Austria (2009-2016), Emory University, Atlanta, GA (2003-2009), New York University, New York City, NY (1989-2003), Princeton Plasma Physics Laboratory, Princeton, NJ (1986-1989) and on PhD research at FOM Institute for Plasma Physics (now DIFFER), Nieuwegein, Netherlands. The PhD work was done at Max Planck Institute for Plasma Physics, Garching, Germany and at UKAEA Culham Laboratory. Basic university education in Utrecht (theoretical physics) and Eindhoven (mathematics and computing science). Interested in scientific computing and data analysis with applications in plasma physics, molecular modelling, and atomic physics.
Research
Please see also my Google Scholar page.
Fusion energy research
For most of my scientific career I have been active in the international effort towards controlled nuclear fusion. This is the field in which I did my thesis research at the UKAEA Culham Laboratory and at IPP Garching, followed by a post-doc position at Princeton Plasma Physics Laboratory and then a research position at New York University. My research in fusion includes numerical modelling of transport processes in near-wall plasma, magnetohydrodynamic equilibrium computations, development of tools for the interpretation of magnetic measurements on tokamaks, relativistic kinetic theory, and numerical Fokker-Planck studies. My best-known work in the field is the development of the B2 or B2.5 code for fluid simulation of the edge plasma, which is a part of the B2-EIRENE and SOLPS-ITER comprehensive near-wall plasma code systems, but my development of function parameterization for the interpretation of magnetic diagnostics on tokamaks has also had a major impact -- up to this date on the Asdex-Upgrade experiment in Garching.
Quantum chemistry and molecular modelling
In the later years of my time at the Courant Institute my interests moved away from nuclear fusion towards quantum chemistry and molecular modelling. My first major effort in this area concerned the development of representability conditions for the two-body reduced density matrix approach to quantum chemistry. My paper with Ph.D.~student Z.~Zhao, postdoc M.~Fukuda, and NYU colleagues M.~Overton and J.~K.~Percus (2004) offered a very important advance as we recognized and implemented an important family of representability conditions beyond the ones that had been used in previous explorations. We computed ground state energies and also dipole moments of a variety of small molecules, and obtain in all cases an accuracy that is similar to or better than that of the well established high-accuracy CISD and CCSD(T) methods on the same model space using full matrix diagonalization as the benchmark. This work reinvigorated the interest in the reduced density matrix approach to quantum chemistry although the interest remains largely theoretical; the computational cost of the approach is too high relative to that of traditional methods.
Following this work on fundamental quantum chemistry, in about 2002 I became interested in more applied molecular modelling and in particular in the development of fitted potential energy surfaces. In 2003 I moved to Emory University to pursue this interest further in joint work with Joel Bowman. Before my contributions to the field it was common to construct high accuracy global fitted potential energy surfaces only for systems of four nuclei or fewer. For systems of five or more nuclei it was common to use reduced coordinates, freezing certain bond lengths or angles. In addition, certainly for systems of five or more nuclei, the variables in terms of which surfaces were described were normally bond lengths, bond angles, and dihedral angles, making it difficult to describe reactions or rearrangement processes.
My work dramatically changed the scene for global fitted potential energy surfaces. It became almost routine to construct global potential surfaces for systems of 5 to 7 or 8 nuclei and occasionally we dealt with larger systems (water clusters, specifically) by way of a many-body expansion. Key to the success for constructing these global surfaces is to use a functional form that is invariant under the complete symmetry group of permutations of like nuclei. This is technically quite difficult if one goes beyond the first few orders in an expansion, and my work relies on the mathematical theory of invariants of finite groups and on the Magma computational algebra system to help generate the codes. In the end this led to a library of largely computer-generated code of close to 200K dense lines to handle all possible permutational symmetry groups of systems involving up to ten nuclei, anywhere from all equal to all distinct.
During the years 2005 to 2007 I pursued a side effort (next to my main molecular modelling interest) on quantum mechanical correlation functions and the Ring Polymer Molecular Dynamics (RPMD) approach to the treatment of quantum effects on nuclear motion. This was joint work with professor David Manolopoulos of the Chemistry department at Oxford University, who developed the RPMD approach. We had three joint publications on different aspects of RPMD, of which the most important one (Braams and Manolopoulos, 2006) concerns the foundations of RPMD. All effective methods to approximate time correlation functions are to some extent heuristic, but one compares them by analysis of symmetry properties, positivity properties, short-time limits, behavior for model systems, and so on, and of course computational experiments too. In our 2006 work we showed that the original RPMD method is unique in the class in providing the best short-time accuracy.
Peripherally I mention one other piece of work inspired by molecular modelling. In 2001 and 2002 I developed a package of codes, MELKES, to perform reduced Hermite interpolation for data on the vertices of a box in arbitrarily many dimensions. The interpolation is reduced in the sense that it uses fewer degrees of freedom than a tensor product form. The interesting cases are those where the data are function values and gradients (but not mixed first derivatives), or function values, gradients, and hessians. The interpolation is purely local, not involving on any box data except at the vertices of that box. The codes are extremely efficient and are publicly available, but in the application for which they were developed, the construction of potential energy surfaces, they have been superseded by my subsequent approach of global fitting using permutationally invariant basis functions.
Research Management
In 2009 I moved to the International Atomic Energy Agency (IAEA) in Vienna as Head of the small Atomic and Molecular Data Unit, part of the Nuclear Data Section, Division of Physical and Chemical Sciences. In that position, which I held 2009 to 2016, science management and science policy were my main responsibilities. The Atomic and Molecular Data Unit is responsible for the Agency's programme on data for atomic, molecular and plasma-material interaction (A+M+PMI) processes and properties relevant to fusion and other plasma applications. The Unit Head position involves organization of complex Technical Meetings, Coordinated Research Projects (CRPs) and data development and evaluation activities. The Unit Head is a programme developer as well as a scientific and technical expert, and must be capable of initiating new research paths needed for the provision of A+M+PMI data for fusion energy and other plasma applications.
The titles of the CRPs for which I had primary responsibility give a good overview of the topical interests of the Atomic and Molecular Data Unit (see http://www-amdis.iaea.org/CRP/): Plasma-wall Interaction with Reduced-activation Steel Surfaces in Fusion Devices. Plasma-wall Interaction with Irradiated Tungsten and Tungsten Alloys in Fusion Devices. Data for Erosion and Tritium Retention in Beryllium Plasma-facing Materials. Atomic and Molecular Data for State-resolved Modelling of Hydrogen and Helium and their Isotopes in Fusion Plasma. Spectroscopic and Collisional Data for Tungsten from 1 eV to 20 keV. Light Element Atom, Molecule and Radical Behaviour in the Divertor and Edge Plasma Regions. Characterization of Size, Composition and Origins of Dust in Fusion Devices.
During my tenure as Head of the Atomic and Molecular Data Unit and working closely with atomic physicist member of staff Hyun-Kyung Chung we gave special emphasis in our meeting programme to the development of methods and a culture of uncertainty assessment and Uncertainty Quantification (UQ) for calculated atomic and molecular data. We organized a sequence of meetings in this area of which I highlight the IAEA Technical Meeting on Uncertainty Assessment and Benchmark Experiments for Atomic and Molecular Data for Fusion Applications (Vienna, 19-21 December 2016). The IAEA has essentially zero funds for direct research support in the A+M+PMMI area, but through meetings such as these we leverage the IAEA good name and reputation to influence the nature of data-oriented research activities among our data producers and in the atomic and molecular date community.
