Skip to main content

Albert F. Wagner

Argonne Distinguished Fellow

Laboratory Emeritus

Biography

Experience

  • Since November 2018, laboratory emeritus
  • B.S. (1966) in Chemistry from Boston College
  • Ph.D. (1972) in Chemical Physics from California Institute of Technology
  • Postdoctoral research (1972–1974) with Dr. Chris Wahl at Argonne National Laboratory
  • Since 1974, on the staff in both research and management capacities at Argonne National Laboratory

Research interests

I am interested in theoretical calculations and simulations of the kinetics and dynamics of molecules and radicals in the gas phase. What dynamical processes control reactive and inelastic encounters between molecules and radicals?  How do these processes change with pressure and temperature?  How do uncertainties in potential energy surfaces influence our calculated kinetics and dynamics?  How can we exploit computer parallelism to accelerate the quantity and quality of our calculations?  In collaboration with others in the group, we attempt to answer these questions with a range of classical to quantum methods using laptops to massively parallel computers. In particular, we are investigating molecular dynamics in bath gases at high pressures, bimolecular reactive tunneling with semi-classical methods, anharmonic partition functions with semi-classical techniques, uncertainty propagation in gas-phase combustion reactions, and highly parallelized localized-orbital electronic structure methods for large systems. 

Gas Phase Molecular Dynamics at High Pressures

The pressure dependence of reaction kinetics is typically modeled by assuming the isolated binary collision (IBC) approximation that states that inelastic collisions are binary events uninterrupted by collisions with other species.  In this research effort, we use molecular dynamics to follow the relaxation of highly rovibrationally excited diatomic, tri-atomic, and polyatomic molecules in bath gases at pressures from tens to hundreds of atmospheres where the IBC approximation can breakdown.  Analysis of the trajectories leads to relaxation mechanisms and mode-specific relaxation pathways.

Semi-classical Tunneling in Bimolecular Reactions

We recently developed an improved analytic semi-classical tunneling theory that is based on readily parallelized, second order vibrational perturbation theory (VPT2) at the saddle point.  For many reactions this improvement extends the reliability of semi-classical theory to energies well below the barrier.  Our approach requires the construction of a composite tunneling potential that incorporates VPT2 saddle point information and the calculated barrier heights in both directions.  We are pursuing a variation of this approach that does not require construction of a potential and is generalizable to higher order vibrational perturbation theory.

Rapid Computation of Gas Phase Chemical Mechanisms

Our group is developing methods and software to rapidly determine gas phase chemical mechanisms using up to exascale computer resources to calculate simultaneously and automatically the thermodynamics and kinetics of a mechanism’s reactions.   Within this large project, I am interested in two efforts.  The first is the propagation of uncertainties in calculated geometries, frequencies, and energetics into the thermodynamics and kinetics and on to the reaction system characteristics (e.g., species profiles, ignition delay, and flame speed). The second effort is the calculation of partition functions at high temperatures where anharmonicity can be important.  Our group has shown that semi-classical corrections to harmonic quantum partition functions offer a feasible approach to high temperature anharmonic partition functions.  I am exploring modifications to these corrections that extend their usefulness to lower temperatures. 

Localized-Orbital Electronic Structure Methods

On massively parallel computers the computational bottleneck for many electronic structure methods is the eigensolve.  For localized-orbital methods (e.g., tight-binding DFT, SIESTA), the eigensolver operates on sparse matrices.  With others we have a developed a highly parallelized eigensolver that exploits that sparsity.  The method slices the eigenvalue spectrum into independent segments for one level of parallelization and uses a segment-specific eigenvalue filter in a Lanczos solution technique for a second level of parallelization.  Effective parallelization has been achieved on up to ~105 cores for problems with up to ~105 atoms.

Recent publications

  • David H.Bross, Ahren W.Jasper, Branko Ruscic, and Albert F.Wagner, Toward accurate high temperature anharmonic partition functions”, Proc. Combust. Inst., doi​.org/​1​0​.​1​0​1​6​/​j​.​p​r​o​c​i​.​2​0​1​8​.​0​5.028
  • Ahren. W. Jasper, Zackary. B. Gruey, Lawrence B. Harding, Yuri Georgievskii, Stephen J. Klippenstein, Albert F. Wagner, Anharmonic rovibrational partition functions for fluxional species at high temperatures via monte carlo phase space integrals”, J. Phys. Chem. 122, 1727-1740 (2018).
  • Murat Keçeli, Fabiano Corsetti, Carmen Campos, Jose E. Roman, Hong Zhang, Àlvaro Vàzquez-Mayagoitta, Peter Zapol, Albert F. Wagner, SIESTA-SIPs: Massively parallel spectrum-slicing eigensolver for an ab initio molecular dynamics package”, J. Comp. Chem. 39, 1806–1814 (2018)
  • Jamin W. Perry and Albert F. Wagner, Pressure effects on the relaxation of an excited hydroperoxyl radical in an argon bath”, Proc. Comb. Inst. 36, 229 (2017)
  • Murat Keceli, Hong Zhang, Peter Zapol, David A. Dixon, and Albert F. Wagner, Shift-and-invert parallel spectral transformation eigensolver: massively parallel performance for density-functional based tight-binding”, J. Comp. Chem. 37, 448–459 (2016)
  • Luis A. Rivera-Rivera, Albert F. Wagner, Tommy D. Sewell, and Don L. Thompson, Pressure effects on the relaxation of an excited nitromethane molecule in an argon bath”, J. Chem. Phys. 142, 014303 (2015)
  • Albert F. Wagner, Richard Dawes, Robert E. Continetti, and Hua Guo, Theoretical/experimental comparison of deep tunneling decay of quasi-bound H(D)OCO to H(D)+CO2”, J. Chem. Phys.141, 054304 (2014)
  • Albert F. Wagner, Improved multidimensional semiclassical tunneling theory”, J. Phys. Chem. A 117, 13089-13100 (2013)