Accelerating the discovery of high-performance donor/acceptor pairs in photovoltaic materials via machine learning and density functional theory

Opportunities and challenges for machine learning to select combination of donor and acceptor materials for efficient organic solar cells

“…Lu et al 91 created a dataset of 717 D : A pairs of OSCs from the literature (379 papers, published during 2015-2020). In their dataset, donors are both small molecules and polymers (# = 192), while all acceptors are NFAs (# = 377).…”

“…Wang et al 90 collected a dataset of 164 NFSMAs from the literature with PBDB-T as the donor. In 2022, Lu et al 91 created a dataset of 717 NFA based organic solar cells for PCE prediction.…”

Active Discovery of Donor:Acceptor Combinations For Efficient Organic Solar Cells

“…13,22 Studies on high-throughput screening by creating a large search test space have been performed to discover potential OSC candidates. 36,39,40 Organic semiconducting materials have tremendous scope for structural flexibility and rich design space, allowing a wide range of optoelectronic characteristics and FMOs to be tuned. 41 Innumerable OSC materials can be synthesized from derivatives of commonly used donor and acceptor materials by altering their donor moiety, acceptor moiety, side chain, core, and end-capping group.…”

“…Interest in ML is increasing in material science-related fields because of the availability of massive data sets, improved algorithms, and exponentially increasing computing power. Various ML models are used for predicting power conversion efficiency (PCE), − short circuit current density ( J SC ), ,,, open-circuit voltage ( V OC ), ,,,, fill factor (FF) , nonradiative voltage loss (Δ V NR ), and frontier molecular orbitals (FMO). , Studies on high-throughput screening by creating a large search test space have been performed to discover potential OSC candidates. ,, …”

Machine Learning Predicts Electrospray Particle Size

“…Thus, methods that can reduce cost, material expenditure and time will accelerate ES product developments. Alternative to this empirical approach is to leverage in silico tools to diminish the number of experiments needed to achieve the desired products [28][29][30][31]. An emerging in silico tool is machine learning (ML), a subfield of artificial intelligence (AI), which uses historical data to predict future outcome.…”