Page 53 - Read Online

Allen et al. J Mater Inf 2024;4:35 https://dx.doi.org/10.20517/jmi.2024.72 Page 13 of 15

[Supplementary Figure 8] show that PSC fabricated on NiO + MeO-2PACz using PC 25 performed better
x
than those made using other PC conditions. The results for TA devices fabricated on NiO HTL + MeO-
x
2PACz are also included.

CONCLUSIONS
In this work, we demonstrate that UV-vis absorbance can function as an effective metric for training a
BO-GP model to predict optimal PC conditions for making high-quality MAPbI films. We identified an
3
optimal MAPbI PC condition (PC 25), which produces a UV-vis spectrum closely matching that of the TA
3
MAPbI , achieving similarity metric values significantly better than other PC conditions. Material
3
characterization shows that PC 25 produces smooth MAPbI films with large grains and high crystallinity.
3
Additionally, we used an AI-based segmentation model to determine grain size from SEM images, offering a
quick and more effective analysis alternative to the standard ASTM E112-13 line intercept method. As a
final test, we present p-i-n PSC results using NiO as the HTL. Despite the high quality of MAPbI films
x
3
made using PC 25, the PSC performance is degraded due to MAPbI /NiO interfacial reaction from PC
x
3
processing. Inserting a buffer layer of MeO-2PACz brings average PCEs comparable to that of PSCs made
with TA MAPbI . These results indicate the importance of a high-quality perovskite active layer and
3
possible interactions between the active layer and charge transport layers that cannot be predicted by the
quality of perovskite films alone. When developing a ML framework for process optimization, careful
consideration of device architecture and adjacent materials is needed to define the objective for the BO-GP
models.

DECLARATIONS
Acknowledgments
We thank J. V. Le for developing the script for grain size analysis and M. Valdez for her help and expertise
in scanning electron microscopy.

Authors’ contributions
Developed the project idea: Allen CR, Lee M, Hsu JWP
Performed most experimental work including ML modeling, UV-vis testing/measurements, grain size
analysis, and PSC fabrication: Allen CR
Performed most advanced material characterization work including AFM, XRD, and SEM, and assisted
with PSC fabrication: Bhandari B
Oversaw experimental work and ML modeling: Xu W, Lee M, Hsu JWP
Wrote the first draft of the manuscript: Allen CR, Xu W
Revised and finalized the manuscript: Allen CR, Bhandari B, Xu W, Lee M, Hsu JWP

Availability of data and materials
Data and codes for this paper are available on GitHub (https://github.com/UTD-Hsu-Lab/MAPI-JMI).

Financial support and sponsorship
This work is supported by the National Science Foundation CMMI-2109554. Bhandari B acknowledges the
support of the U.S. Department of Energy’s Office of Energy Efficiency and Renewable Energy under the
Solar Energy Technologies Office Award Number DE-EE0009518. Hsu JWP acknowledges the support of
the Texas Instruments Distinguished Chair in Nanoelectronics.

Conflicts of interest
All authors declared that there are no conflicts of interest.