Unbiased crystal structure prediction
given by Stefano de Gironcoli
Start: 10-01-2021 End: 28-02-2021 Room: 131
The stability of a material polymorph with interesting (optoelectronic, thermal, electrical, ....) properties crytically depends on the competing phases, either with the same stoichiometry or accessible via chemical decomposition. Predicting the crystal structure of a material from scratch is a challenging problem that depends on many kinetic factors but is primarily dominated by thermodynamics and ultimately by the energetic stability of the competing candidate structures.
This kind of problems can be addressed combining i) an efficient strategy to explore the crystal stucture phase space and ii) an accurate and efficient way to evaluate the needed energetics. Among the openly available tools for crystal structure exploration our group has some experience with the USPEX (Universal Structure Predictor: Evolutionary Xstallography) code, as for the accurate DFT calculations the Quantum ESPRESSO distribution will be employed.
CsPbI3 in the perovskite structure and its hybrid organic/inorganic relatives ( for instance the Methylammonium Lead Iodide (CH3NH3)PbI3 or MAPI ) form an interesting family of materials for photovoltaic applications with energy conversion efficiency in the 15-20% range. The main problem associated to the Lead Iodide family is the toxicity of its Lead component.
A possible alternative involving Tin (completely or at least partially replacing Lead) is much less studied and an assessment of its stability in the perovskite structure is highly desirable. In particular it is interesting to establish the relative stability of competing polymorphs with CsSnI3 stoichiometry as well as the energy balance of possible decomposition products such as CsI and SnI2, not forgetting Tin higher oxydation state Stannic Iodide, SnI4.
Interested Students should contact Stefano de Gironcoli for an introductory discussion and keep in touch regularly to discuss progresses and address problems of this research project.