Understanding perovskite solar cell physics through combining modelling and machine learning
Alison Walker a, Samuel McCallum b, Jamie Lerpiniere a
a Department of Physics University of Bath, BA2 7AY, UK
b Department of Mathematical Sciences, University of Bath, BA2 7AY, UK
International Conference on Hybrid and Organic Photovoltaics
Proceedings of International Conference on Hybrid and Organic Photovoltaics (HOPV24)
València, Spain, 2024 May 12th - 15th
Organizer: Bruno Ehrler
Invited Speaker Session, Alison Walker, presentation 100
DOI: https://doi.org/10.29363/nanoge.hopv.2024.100
Publication date: 6th February 2024

This talk will cover our recent work combining machine learning and simulation methodologies to produce much faster and more direct characterisation of materials and devices. We have created a virtual model through a combination of device transport models and machine learning. This combination can be used to test hypotheses about the physical processes within these devices. These processes include the role of interfaces where charge accumulation/depletion can occur and traps where nonradiative recombination takes place are often located. I will show how we address outstanding questions on charge transport in lead halide perovskites.


My main topic will concern how machine learning can solve the inverse parameter problem, where a device model (here the drift-diffusion code IonMonger) is combined with Bayesian parameter estimation to deduce the input parameters for the device model from experimental measurements of the device characteristics [1]. This approach allows us to pinpoint causes of features seen in the measurements using only a few hours of computation. The virtual model can be continuously updated to reflect the current output of a fabricated laboratory device. Through these updates the materials processes underlying changes in the devices’ outputs can be identified. Accurately and rapidly simulating the function and performance of the lab device opens up the possibility of pinpointing the origins of degradation and allows improvements to be made much more quickly in future device iterations.

The talk will also describe how we efficiently search the input space using Bayesian optimization to minimize the difference between the simulation output from our mesoscale simulations with the code BoltMC, and a set of experimental results. BoltMC uses Boltzmann transport theory implemented via ensemble Monte Carlo to provide insight into mobility-limiting mechanisms [2]. From our analysis, we conclude that the formation of large polarons, quasiparticles created by the coupling of excess electrons or holes with ionic vibrations, cannot explain the experimentally observed temperature dependence of electron mobility.

We thank the UK Engineering and Physical Sciences Research Council (EPSRC) for a doctoral training partnership studentship (JEC) and for a Centre for Doctoral Training in Statistical Applied Mathematics (SAMBa) studentship (SGM).

© FUNDACIO DE LA COMUNITAT VALENCIANA SCITO
We use our own and third party cookies for analysing and measuring usage of our website to improve our services. If you continue browsing, we consider accepting its use. You can check our Cookies Policy in which you will also find how to configure your web browser for the use of cookies. More info