Research

Take away computation from the world and everything is wrapped in blind ignorance -- Isidore of Seville quoted from Peter Hunter Blair

Recent work:

Proof that H passivation is the origin of low doping efficiency in the important photovoltaic material a-Si:H.

Realistic computer models of amorphous GaN have completely delocalized conduction tails. One would expect that electrons in the tail would have much higher mobility than in a material like a-Si or chalcogenide glasses. It bolsters our old argument that a-GaN might have unique potential as an electronic material.

We are developing the ab initio theory of transport in disordered materials. The idea is to make use of the basic output from current ab initio codes (Kohn-Sham states and energies) and classical normal modes and frequencies to formulate a predictive theory of transport. The quantization of the lattice vibrations is fully included and the full temperature and frequency dependent conductivity and Hall effect is the outcome. Papers and preprints here and here. See also an interesting and closely related study of “The work done by an external electromagnetic field”

What does diffraction data imply about the structure of an amorphous system? Until recently, it was believed that it was a necessary condition that an atomistic model should reproduce diffraction data. We show for homogeneously disordered materials that the information in the data is far closer to sufficient than previously believed. With M. Cliffe, A. L. Goodwin and M. T. Dove, we make this report here APS Physics Viewpoint here. It is also highlighted in the Oxford Science Blog.

What is the nature of the trapping centers and the ion hopping dynamics by direct calculation in glassy chalcogenide materials? These systems are standard examples of fast ion conductors. These processes are relevant to batteries. These materials are a promising candidate to replace current FLASH memory devices. Paper here.

What is the atomistic origin of the intermediate phase? Raman and calorimetric studies on GeSe glasses have provided evidence for the existence of the intermediate phase (IP) in chalcogenide and other glasses. With Prof. Gang Chen, we present X-Ray Absorption Near Edge Structure (XANES) measurements on germanium selenide glasses in the IP composition range, and detect an electronic signature of the IP. Ab initio molecular dynamics (MD) based models of these glasses are discussed, and an atomistic picture of the IP, based upon the models and available experiments is presented. We show that these models reproduce detailed composition-dependent structure in the XANES measurements, and otherwise appear to properly represent the structure of the GeSe glasses near the IP. For a more detailed analysis of the “self organization” see this paper.

Where do universal Urbach edges “come from”? Disordered systems always show an exponential band tail at both valence and conduction edges as discovered by Urbach in 1953. Here, for the key amorphous material, a-Si, we show exactly which structural elements in the model give rise to these “Urbach tails”. The short answer is that there are filaments of long and short bonds that are highly connected and correlated, and it is exactly these filaments that give rise to the Urbach edges in Si. Short bonds, valence edge; long bonds, conduction edge. We also show how Urbach tails arise from relaxed point defects in c-Si. Additional background here. We have generalized all of this to amorphous silica, and observed closely related phenomena in amorphous graphene and even carotene-beta [here].

Older work:

What causes light-induced photovoltaic degradation in solar cells? Clear evidence that light-soaking a-Si:H creates Si-dihydride here. More complete discussion here (2006 Institute of Physics Highlight) and here.

Universal features of the localized to extended transition for varied disorder, interactions, electrons and phonons here.

A novel scheme to merge ab initio simulation and experimental information to model glasses and amorphous materials. Use all of the information available to make the best models! "Experimentally Constrained Molecular Relaxation": in J. Phys. Cond. Matt . and Phys. Rev. B .

A general result: "The Electron-Phonon Coupling is Large for Localized States". The paper not only makes the observation from simulation, but explains why with a simple model.

An improved "Reverse Monte Carlo" scheme for modeling amorphous materials . With sufficient constraints, electronically realistic models are possible.

We have found that a surprisingly simple method can produce realistic models of binary amorphous materials , by "decorating" the bonds of tetrahedral elemental amorphous semiconductor models (like Ge or Si). Results here: Realistic Models of Binary glasses from Models of Tetrahedral Amorphous Semiconductors

Pressure-induced phase transitions in amorphous materials: We have published a series of papers elucidating amorphous to amorphous transitions in silicon , germanium , glassy GeSe 2 , GaAs and paracrystalline Si : the first accurate calculations of insulator-metal transitions between amorphous states.

The finite-temperature Anderson problem: the existence of diffusion in our favorite random lattice for T>0! Atomistic simulation of the finite-temperature Anderson localization problem: Phys. Stat. Sol. (2002) ; Jun Li and D. A. Drabold, A portrait of hopping between localized states: a density functional simulation of the finite-temperature Anderson problem ,

We (Sergei Taraskin, DD and S. R. Elliott) have found exact asymptotic expansions for the single-particle density matrix in two and three dimensions . PDF here , published paper in Phys. Rev. Lett. here . Extended to a one band model of a metal here.

How does the density matrix really decay in metals and insulators according to a realistic Hamiltonian? This is a fundamental decay or correlation length in a given material, and is a key to efficient approaches to electronic structure calculations. Properties of the density matrix from realistic calculations , Phys Rev B 63 233109 (2001).

The first ab initio calculation of photostructural effects in a binary chalcogenide, and hints of a theory of photomelting, Direct calculation of light-induced structural change and diffusive motion in glassy As2Se3 with Jun Li.

The first ab initio calculation of photostructural effects in a glass, Direct molecular dynamic simulation of light-induced structural change in amorphous selenium, with X. Zhang.

Does anamolous low-T specific heat come from local vibrational modes of defects in glasses? Phys. Rev. B 61 5376 (2000). Links to images of the modes here

How to build Wannier functions order N using ab initio methods: Order-N projection method for first-principles computations of electronic quantities and Wannier functions

Maximum Entropy in Condensed Matter Theory ; A maximum entropy approach to closing hierarchies and summing/analytically continuing perturbation series [examples include the quantum harmonic oscillator with octic perturbation and the virial equation of state for hard spheres] ( Maximum Entropy and Bayesian Methods , Kluwer, 1991. See also, D. A. Drabold and G. L. Jones, Maximum-entropy approach to series extrapolation and analytic continuation