Capturing Anharmonic Vibrational Motion in First-Principles Simulations
Location: On-line, hosted by CECAM-HQ
In traditional electronic structure studies, the nuclear motion is typically neglected, albeit having a profound impact on essentially all properties of matter. Even when assuming that the dynamics of electrons and nuclei follows the Born-Oppenheimer approximation, nuclear fluctuations manifest themselves either directly through atomic structural changes in molecules and solids, or indirectly through changes in the electronic structure, due to vibronic coupling. Moreover, the nuclear dynamics defines different response properties of matter, like vibrational excitations, heat transport, diffusion properties, and many more.
Pioneering first-principles studies capturing these effects have relied on the harmonic approximation for the nuclear degrees of freedom or on perturbative expansions thereof, thus accurately capturing "weak" anharmonicity at best. These studies were nevertheless successful, being able to shed light on the temperature dependence of electronic band-gaps , on intricate vibrational energy reorganization , and on the character of vibrational fingerprints of molecules . However, over the last decade, with the increase in accuracy of electronic structure methods and the emergence of materials that combine soft and hard matter, it has become evident that a naïve modeling of the nuclear motion in the harmonic approximation is not sufficiently accurate or even qualitatively incorrect in many cases [4, 5]. This realization has fueled an array of recent methodological developments, ranging from higher-order perturbative methods to novel ab initio molecular dynamics approaches – sometimes aided and improved by machine-learning algorithms [6,7]. To cite a few developments where the organizers have been directly involved, one could name novel, fully anharmonic methods for the computational prediction of thermal conductivities  and of temperature-dependent electronic properties  in solids and the inclusion of quantum anharmonicity in vibrational properties [10,11], both applicable to realistic molecules and solids.
Clearly, the challenges one faces when investigating the vibrational motion vary with the actual material of interest. Within ab initio electronic structure simulations, achieving convergence with respect to system size (viz. reciprocal q-points ) has been a massive hurdle for inorganic solids, whereas achieving convergence with respect to the configurational entropy of the nuclear degrees of freedom has been equally challenging in organic molecular matter. Ongoing efforts in the scientific community focus on developing more accurate methodologies and on reducing the cost of these simulations. However, the molecular and solid state communities have few opportunities to meet each other, even if the challenges faced by both when addressing anharmonic nuclear motion are similar. We thus believe it is timely to organize a workshop where expert and prominent code and method developers from these two communities are brought together. We plan a workshop program that will not be aimed at showcasing the participants' achievements, but rather at discussing the nuts and bolts of the methods and implementations – an aspect that often falls short of other presentation formats. Along these lines, we will strongly encourage the participation and involvement of young researchers that are involved in the code development on a day-to-day basis.
Christian Carbogno (Fritz-Haber-Institut der Max-Planck-Gesellschaft) - Organiser & speaker
Mariana Rossi (Max Planck Institute for the Structure and Dynamics of Matter) - Organiser & speaker