• 
  Gries, Thomas William: Predicting Performance and Stability in Perovskite Photovoltaics by Quantifying Extraction and Recombination at Semiconductor/Selective-Contact Interfaces. , Dissertation, Universität Bielefeld, 2025

Abstract:
The commercialization of perovskite photovoltaics relies on controlling the interfaces, where efficiency and stability are critically determined. These heterojunctions between the perovskite absorber and charge-selective contacts represent non-radiative recombination channels, which limit charge extraction and initiate long-term degradation. However, their microscopic dynamics remain difficult to access, as most experimental probes average over the bulk or provide only indirect signals. This thesis develops a strategy to open this “black box” by uniting surface-sensitive and optoelectronic characterization with drift-diffusion (DD) modeling. The first study investigated TiO2 co-doped with Nb(V) and Sn(IV) as an electron-selective contact for CsPbI3. Kelvin-probe force microscopy and time-of-flight secondary ion mass spectrometry showed that co-doping increased bulk conductivity and suppressed interface trap states. A DD model was parameterized and quantitatively fitted to transient surface photovoltage (trSPV) data. The simulations revealed that co-doping reduced the interface hole recombination velocity and increased the number of extracted electrons. These microscopic improvements translated into macroscopic improvements in open-circuit voltage and fill factor, and extended device stability. The second study moved from materials optimization to a conceptual generalization. It addressed the apparent paradox that steady-state photoluminescence (PL) is often quenched when device performance improves. Applying PL, trSPV, and DD modeling across three classes of semiconductor/CSC heterojunctions showed that PL quenching can arise not from increased non-radiative losses but from efficient charge extraction. The decisive parameter is the ratio of interface to bulk lifetimes (τinterf/τbulk), which determines whether PL quenching or enhancement is observed. From this insight, a decision tree was derived that classifies interface quality from PL and trSPV alone, validated on more than 80 heterojunctions. These results demonstrate that buried interfaces can be understood and optimized through the combination of chemical modifications with predictive modeling. This thesis contributes both a concrete route to more efficient and stable perovskite photovoltaics and a generalizable methodology for distinguishing recombination from extraction at semiconductor interfaces. By establishing this link between microscopic interface physics and macroscopic device function, it points toward accelerated, rational design of photovoltaic technologies.