Page 42 - Read Online

Page 2 of 15 Allen et al. J Mater Inf 2024;4:35 https://dx.doi.org/10.20517/jmi.2024.72

INTRODUCTION
In recent years, the efficiency of perovskite solar cells (PSCs) has improved remarkably, with power
[1]
conversion efficiencies (PCEs) reaching up to 26.7% . Of the primary benefits for perovskites are the low
cost and outstanding optoelectronic properties of solution-processed films, making PSCs an attractive
addition to next-generation photovoltaic technologies . Perovskite active layers often require a thermal
[2]
annealing (TA) step to convert a deposited precursor film into a fully crystalline film that has high
absorbance and large grains, and thus long diffusion length and high mobility . However, this process
[3,4]
typically takes tens of minutes of annealing at 100 to 150 °C and is a bottleneck for large-scale PSC
manufacturing . Previous works on photonic curing (PC), utilizing a xenon flash lamp to deliver intense
[5]
[6-8]
broadband light to the film, reduced the annealing time of perovskite films to ~20 ms . Therefore, this
technique can be a candidate to replace TA in the scale-up manufacturing of PSCs.

All previous works on using PC for crystallizing perovskite films vary only by the amount of time over
which light illuminates the sample, the pulse length (ms), and the energy the light delivers to the sample in
one pulse, the radiant energy (J/cm ) . In this work, we use a more sophisticated pulse that includes
2 [6,7,9-12]
micro-pulses (µpulse), which split a single pulse into several smaller sub-pulses with a specified duty cycle.
Using these additional features allows us to shape the temperature profile of the thin film, ultimately gaining
more control over how the film crystallizes. The addition of these two variables requires optimization of a
four-dimensional input space. In the case of problems with only two variables, a typical varying
one-variable-at-a-time approach is often sufficient to properly parameterize the space. However, this
method often fails in higher dimensional input spaces, where the interdependence of the variables requires
an impractically large number of test conditions to confidently reach any conclusion. Xu et al. showed
success in using Bayesian optimization coupled with Gaussian process regression (BO-GP) as an effective
tool to optimize the PC of a different methylammonium lead iodide (MAPbI ) recipe using the device PCE
3
as the objective for optimization . While PCE is the ultimate goal of the MAPbI PC optimization process,
[12]
3
making and testing a set of PSCs can take as long as two days to complete. When coupled with the fact that
each sample needs to be produced numerous times to check for reproducibility, relying on PCE as the
objective function in BO is labor-intensive and time-consuming, presenting a bottleneck in processing
optimization.

Various studies have applied machine learning (ML) modeling along with high-throughput material,
optical, and electronic characteristics to optimize perovskite materials [13-15] , e.g., finding optimal triple-cation
[16]
perovskite composition using photoluminescence and employing machine vision and optical imaging of
perovskite films to predict film quality and estimate short-circuit current density . In this work, we
[17]
perform BO to optimize PC conditions to crystallize MAPbI by measuring their ultraviolet-visible (UV-vis)
3
absorbance spectra, which are used as a proxy for good PSCs. We quantitatively compare the UV-vis
absorbance around the bandgap (600-850 nm) for TA and PC MAPbI using mathematical similarity
3
metrics. UV-vis absorbance is chosen because it is a fast material characterization method, in addition to
providing crucial information about MAPbI thin film properties. Beyond light absorption, shifts and
3
changes in the shape of the absorbance curve can indicate grain size and uniformity (including the presence
of pinholes, defects, and intermediaries) , film thickness , and crystallinity , which all play key roles in
[20]
[18]
[19]
determining the PCEs of MAPbI PSCs.
3
Furthermore, we employ a ML method to improve grain size determination. The usual ASTM E112-13 line
intercept method of grain size determination is time-consuming and based on a limited amount of data
[21]
from the image. By implementing an AI image segmentation model, we use data from the entire image and