Marianetti Wins DOE Grant on Phonon Interactions
A new approach to the interacting phonon problem
Performing high-risk-high-reward research at universities can be inherently difficult due the nature of funding for scientific research. Even when stated otherwise, government funding agencies cling to research that has basically already succeeded and is extending into a more linear regime of exploration; which is easy to understand given the tight constraints on federal funding. Fortunately, Columbia has a number of seed funding programs in place to act as an incubator for high-risk-high-reward research; one such program being the Columbia Research Initiatives for Science and Engineering (RISE) grant. In 2013, Chris Marianetti was awarded a RISE grant entitled “A new approach to the interacting phonon problem”, which allowed for the support of a postdoctoral researcher and resulted in, among other things, a publication in a noted physics journal, Physical Review Letters. This work was substantial enough to form a foundation for a proposal to the Department of Energy (DOE) Basic Energy Sciences (BES) division, and this was recently (September 2016) funded at a level of $415,000 over a period of three years.
While the topic of “phonon interactions” may sound somewhat obscure, it is actually extremely general. Phonons and their interactions influence, or sometime entirely dictate, many materials properties at non-zero temperatures, including phase stability, thermal transport, and mechanical behavior. Furthermore, there are various experimental phenomena related to phonon interactions which lack a firm theoretical understanding and could have technological relevance, such as intrinsically localized modes. Understanding technologically relevant materials at high temperatures necessitates a strong handle on the interacting phonon problem, and applications run the gamut from turbines in jet engines to nuclear fuel materials to thermoelectrics. In order to be able to successfully predict materials properties at a wide range of temperatures and to understand experimental anomalies, a new approach which can reliably treat the interacting phonon problem is needed. In line with the other research thrusts in Marianetti’s group, the goal is to start with nothing more than the Schrödinger equation (i.e. the laws of quantum physics for interacting nuclei and electrons) and predict materials behavior; which in this case are phonon interactions at a broad range of temperatures.
The proposed research involves two different thrusts. The first thrust is to characterize the irreducible interactions between phonons. Computing phonon’s themselves is a somewhat established technique, and phonons can be precisely computed for a wide range of materials. The same cannot be said for phonon interactions. Therefore, the first step is to compute phonon interactions over a broad range of materials while exploiting the full power of group theory. The longer term goal will be to execute this in the spirit of the Materials Genome Initiative, addressing a broad range of materials, storing the results in a database, and mining information using machine learning techniques. These phonon interactions could immediately be used to perform traditional classical simulations such as molecular dynamics. The second thrust addresses the inclusion of quantum fluctuations when solving this interacting phonon problem. While traditional perturbative and quantum Monte-Carlo techniques have made substantial progress in this space, there is still a need for an accurate, efficient approach which could be broadly applied to real materials.
 Y. Chen, X. Ai, and C. A. Marianetti. First-principles approach to nonlinear lattice dynamics: Anomalous spectra in pbte. Phys. Rev. Lett., 113:105501, Sept. 2014.