Recording of my presentation at the Molecular Simulation with Machine Learning On-line workshop this summer. It presents the hybrid perovskites as a prime example of "Dynamic Solids" and gives overview of the research and developments we have made in the last 3 years.
At record pace, the solar cell efficiency of MAPbI3 based solar cells has increased from a few percent to over 20%. A perovskite crystal film can be grown by simple wet chemistry and performs well even for much thinner films than silicon. I focus on the physical mechanisms determining this success in the larger class of ABX3 hybrid perovskite semiconductors. At room temperature, a cubic framework is formed by PbI6 octahedra enclosing Methylammonium (MA) molecules. The MA+ Pb2+ I-3 system is held together by a mixuture of covalent and ionic bonds, whereby the MA molecule is chemically decoupled from the PbI framework. Weak (van der Waals, hydrogen) bonds between the MA molecule and the cage result in an effective rotational barrier of ~30 meV, thereby allowing the molecule to spin around. In my view, the ionic nature of the crystal, the high degree of dynamics in combination with the 1.6 eV bandgap at this particular chemical composition are the dominant factors for the remarkable solar cell efficiency. Therefore, I use this appealing dynamical system to study the following interconnected topics:
Finite temperature atomic structure prediction: Ab-initio molecular dynamics calculations of large supercells at finite temperature. Point of care: different density functional approximations describe the weak interactions between the molecules and the cage differently.
The finite temperature dielectric function: Ionic contributions have a large effect on the dielectric function of the perovskite crystal. The static dielectric constant increases from ~7 (electronic contributions only) to ~30 (elec. + PbI motion) to values as large as ~100 (elec. + PbI motion + MA rotation). In the last step the MA molecules align to an external electric field because they carry an intrinsic dipole moment. The question is, at what frequencies do these effects set in, and how applicable is the harmonic approximation here? We are devolping a new method to study these effects. Currently we are testing it on SrTiO3 and BaTiO3.
First-principles absorption spectra: A combination of GW -BSE approximations can be applied to predict the correct onset of absorption (the optical band gap) and the dielectric function of new (hybrid) perovskite materials.
"Ionic dressing of the excited state." How, and to what extent, do excited quasiparticles couple to the ionic lattice?