Automated Multimodal Mapping of Optical and Relative Surface-Potential Responses in a FAPbI3–BDAPbI4–PEA2PbI4 Halide Perovskite Library
Abstract
High-throughput (HT) characterization has accelerated materials screening, however most workflows for functional characterization rely on a single measurement modality. This approach leaves the relationship between different aspects of materials response, such as optical emission and local electronic behavior, largely unassessed within the same campaign. Multimodal characterization, particularly pairing probe-based electronic measurements with optical screening, remained comparatively underdeveloped. Here we address this gap by combining automated Kelvin probe force microscopy under illumination (photoKPFM) with HT photoluminescence (PL) peak-intensity mapping to screen a formamidinium lead iodide (FAPbI3), 1,4-butanediammonium lead iodide (BDAPbI4), and phenethylammonium lead iodide (PEA2PbI4) combinatorial halide perovskite library. A Gaussian-process Bayesian optimization with pure exploration workflow selects photoKPFM measurements to reconstruct the composition-dependent surface potential response, benchmarked on a binary FAPbI3–PEA2PbI4 line before extension to the ternary system. Comparing photoKPFM against PL peak intensity across 13 unique nominal ternary compositions, measured on independently prepared replicate libraries, shows that both descriptors broadly prioritize PEA2PbI4-rich compositions, but individual composition rankings frequently disagree, and exploratory correlation analysis finds no statistically significant global association between the two measurements after multiple-comparison correction. Rather than indicating redundancy, this divergence shows that photoKPFM and PL provide complementary, non-interchangeable candidate-selection information, jointly identifying compositions that rank consistently strong by both descriptors and compositions whose priority depends on which functional descriptor is used. This work demonstrates how pairing automated probe-based and optical HT methods surfaces candidates that single-modality screening would rank differently, motivating multimodal characterization as a general, extensible strategy for combinatorial materials discovery beyond this specific material system.
Related
- Multimodal Bioinspired Self‐Healing Composites Enabling Durable and High‐Efficiency Wearable Perovskite Light‐Emitting Diodes
- Chemistry-Informed Multimodal Model for Structure Elucidation in Automated Reaction Discovery
- First-Principles Investigation of Multimodal Toxic Gas Sensing in Carbon-Tuned hBN-Graphene Alloys: Chemiresistive, Work-Function, and Optical Responses
- Degradation and Stabilization of Inverted Perovskite Solar Cells
- LiteratureAgent: Linking Perovskite Literature to Database Expansion, Knowledge Graphs, and Predictive Modeling
BibTeX
@misc{marshall2026automated,
title = {{Automated Multimodal Mapping of Optical and Relative Surface-Potential Responses in a FAPbI3–BDAPbI4–PEA2PbI4 Halide Perovskite Library}},
author = {Jordan Marshall and Boris Slautin and Hochan Song and Sergei Kalinin and Mahshid Ahmadi},
howpublished = {ChemRxiv},
year = {2026},
month = aug,
doi = {10.26434/chemrxiv.15008093/v1},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008093/v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Automated Multimodal Mapping of Optical and Relative Surface-Potential Responses in a FAPbI3–BDAPbI4–PEA2PbI4 Halide Perovskite Library
AU - Jordan Marshall
AU - Boris Slautin
AU - Hochan Song
AU - Sergei Kalinin
AU - Mahshid Ahmadi
PY - 2026
DA - 2026/08/31
PB - ChemRxiv
DO - 10.26434/chemrxiv.15008093/v1
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008093/v1
AB - High-throughput (HT) characterization has accelerated materials screening, however most workflows for functional characterization rely on a single measurement modality. This approach leaves the relationship between different aspects of materials response, such as optical emission and local electronic behavior, largely unassessed within the same campaign. Multimodal characterization, particularly pairing probe-based electronic measurements with optical screening, remained comparatively underdeveloped. Here we address this gap by combining automated Kelvin probe force microscopy under illumination (photoKPFM) with HT photoluminescence (PL) peak-intensity mapping to screen a formamidinium lead iodide (FAPbI3), 1,4-butanediammonium lead iodide (BDAPbI4), and phenethylammonium lead iodide (PEA2PbI4) combinatorial halide perovskite library. A Gaussian-process Bayesian optimization with pure exploration workflow selects photoKPFM measurements to reconstruct the composition-dependent surface potential response, benchmarked on a binary FAPbI3–PEA2PbI4 line before extension to the ternary system. Comparing photoKPFM against PL peak intensity across 13 unique nominal ternary compositions, measured on independently prepared replicate libraries, shows that both descriptors broadly prioritize PEA2PbI4-rich compositions, but individual composition rankings frequently disagree, and exploratory correlation analysis finds no statistically significant global association between the two measurements after multiple-comparison correction. Rather than indicating redundancy, this divergence shows that photoKPFM and PL provide complementary, non-interchangeable candidate-selection information, jointly identifying compositions that rank consistently strong by both descriptors and compositions whose priority depends on which functional descriptor is used. This work demonstrates how pairing automated probe-based and optical HT methods surfaces candidates that single-modality screening would rank differently, motivating multimodal characterization as a general, extensible strategy for combinatorial materials discovery beyond this specific material system.
ER -