Search

This paper presents a comprehensive approach for designing polymer acceptors for organic photovoltaic applications through the generation of an extensive database and the application of machine learning (ML) techniques. Over 40 ML models are trained for the prediction of power conversion efficiency (PCE). Histgradient boosting regressor has appeared as best model. Almost 10 k polymers are generated and their PCE values are predicted. The chemical space of polymers has been visualized and analyzed. Cluster analysis revealed significant differences among the selected polymers. Additionally, an assessment of synthetic accessibility for these polymers indicated that the majority can be synthesized with relative ease. [ABSTRACT FROM AUTHOR]

Machine learning (ML) analysis has gained huge importance among researchers for predicting multiple parameters and designing efficient donor and acceptor materials without experimentation. Data are collected from literature and subsequently used for predicting impactful properties of organic solar cells such as power conversion efficiency (PCE) and energy levels (HOMO/LUMO). Importantly, out of various tested models, hist gradient boosting (HGB) and the light gradient boosting (LGBM) regression models revealed better predictive capabilities. To achieve the prediction effectively, the selected (best) ML regression models are further tuned. For the prediction of PCE (test set), the LGBM shows the coefficient of determination (R2) value of 0.787, which is higher than HGB (R2 = 0.680). For the prediction of HOMO (test set), the LGBM shows R2 value of 0.566, which is higher than HGB (R2 = 0.563). However, for the prediction of LUMO (test set), the LGBM shows R2 value of 0.605, which is lower than HGB (R2 = 0.606). Among the three predicted properties, prediction ability is higher for PCE. These models help to predict the efficient acceptors in a short time and less computational cost. [ABSTRACT FROM AUTHOR]

Context: Selecting high performance polymer materials for organic solar cells (OSCs) remains a compelling goal to improve device morphology, stability, and efficiency. To achieve these goals, machine learning has been reported as a powerful set of algorithms/techniques to solve complex problems and help/guide exploratory researchers to screen, map, and develop high performance materials. In present work, we have applied machine learning tools to screen data from reported studies and designed new polymer acceptor materials, respectively. Quantitative structure-activity relationship (QSAR) models were generated using machine learning-assisted simulation techniques. For this purpose, 3000 molecular descriptors are generated. Consequently, molecular descriptors having key effect on power conversion efficiency (PCE) were identified. Moreover, numerous regression models (e.g., random forest and bagging regressor models) were developed to predict the PCE. In particular, new materials were designed based on the similarity analysis. The GDB17 chemical database consisting of 166 million organic molecules in an ordered form is used for performing similarity analysis. A similarity behavior between GDB17 materials and the materials reported in literature is studied using RDKit (a cheminformatics software). Noteworthily, 100 monomers proved to be unique and effective, and PCEs of these monomers are predicted. Among these monomers, four monomers exhibited PCE higher than 14%, which is better than various reported studies. Our methodology provides a unique, time- and cost-efficient approach to screening and designing new polymers for OSCs using similarity analysis without revisiting the reported studies. Methods: To perform machine learning analysis, data from reported studies and online databases was collected. Different molecular descriptors were generated for polymer materials utilizing Dragon software. 3D structures of studied molecules were applied as input (SDF; structure data file format). Importantly, about 3000 molecular descriptors were generated. Comma-separated value (.csv) file format was used to export these molecular descriptors. To shortlist best descriptors, univariate regression analysis was performed. These descriptors were further utilized for training machine learning models. Moreover, necessary packages of Python for data analysis and visualization were imported such as Matplotlib, Numpy, Pandas, Scikit-learn, Seaborn, and Scipy. Random forest and bagging regressor models were applied for performing machine learning analysis. A cheminformatics software, RDKit, was applied for similarity analysis. [ABSTRACT FROM AUTHOR]

[Display omitted] • Machine learning is used to predict the power conversion efficiency (PCE). • Gradient boosting regressor outperforms other models in PCE prediction. • 30 polymer donors have identified and evaluated for synthesis. • Selected donors have exhibited significant structural similarity. The design of novel polymer donors for organic solar cells has been a major research focus for decades, but discovering unique materials remains challenging due to the high cost of experimentation. In this study, machine learning models are employed to predict power conversion efficiency (PCE), Mordred descriptors are used for model training. Among the four machine learning models evaluated, the gradient boosting regressor emerged as the best-performing model. Additionally, a chemical library of polymer donors was generated and analyzed using various measures. 30 donors with highest PCE are selected and their synthetic accessibility is evaluated. Similarity analysis has indicated much resemblance in selected polymer donors. [ABSTRACT FROM AUTHOR]

[Display omitted] The communalization of all-polymer solar cells depends on the cost of active layer materials. We have introduced a framework to find the easily synthesizable polymers. Breaking Retrosynthetically Interesting Chemical Substructures (BRICS) method is used to generate a large database of polymers and synthetic accessibility of generated polymers is predicted. The generated database is visualized using various methods. Energy levels of polymers are also predicted using pretrained machine learning models. Polymers are screened on the basis of predicted properties. Library of polymers are displayed using the T -distributed Stochastic Neighbor Embedding (t -SNE) visualization. Structure Activity Landscape Index (SALI) visualization is also used. A significant change is observed in synthetic accessibility score on structural changes. The histagradient boosting regressor is used to predict the energy levels of polymers that energy levels play significant role in the selection of materials for organic photovoltaic cells. Synthetic accessibility of polymers is analyzed and a significant number of polymers are easy to synthesize. Thirty polymers are selected through screening process that are potential candidates for organic photovoltaic cells. [ABSTRACT FROM AUTHOR]

Organic solar cells (OSCs) are gaining fame for their cost-effective solution processing. Machine learning is increasingly popular for material design in OSCs. In this study, molecular fingerprints are used to train over 40 machine learning models. The random forest regressor emerges as the most predictive one. 10k new dyes are generated. A pre-trained ML model is used to predict their reorganization energy values. Dyes are selected on the basis of reorganization energy, dyes with lower reorganization energy are retrained. The synthetic accessibility of chosen dyes is then analyzed. Chemical similarity analysis has indicated reasonable resemble among selected dyes. [Display omitted] • Machine learning optimized dye design for lower hole reorganization energy. • Trained over 40 models, Random Forest is identified as top predictor. • Generated 10k new dyes for comprehensive analysis. • Dyes with lower reorganization energy are prioritized and retrained. • Analysis confirms accessibility of chosen dyes for practical application. [ABSTRACT FROM AUTHOR]

[Display omitted] • Machine learning is used to predict power conversion efficiency of organic solar cells. • Gradient boosting regression has identified as best ML model. • Designed and evaluated 10,000 small molecule donors. • Chemical similarity analysis has revealed reasonable structural resemblance. • Synthetic accessibility has indicated easy synthesis for majority of selected donors. Designing of materials for organic solar cells is a tedious process. In present study, machine learning (ML) is used to predict the power conversion efficiency (PCE). Over 40 ML models are tried. Gradient boosting regression is appeared as best model. 10k small molecule donors are designed. Their PCE values are predicted using best model. The library of generated donors is visualized using various tools. Chemical similarity analysis is done to study structural behavior of selected donors. Reasonable resemblance is found. Synthetic accessibility assessment has indicated easy synthesis for majority of selected small molecule donors. The introduced framework has ability to find the efficient materials in short time. [ABSTRACT FROM AUTHOR]

Quantum mechanism calculations were performed to study the relationship between strength of electron‐donating, electron‐withdrawing groups and electronic, photochemical, charge transport properties. Electron‐donating groups blueshifted the ultraviolet (UV)/visible spectra, while electron‐withdrawing groups redshifted the UV/visible spectra. Inverse relationship observes between Hammett parameter and reorganization energy. Small molecules acceptors with electron‐withdrawing substituents showed higher electron mobility. This study can pave way for experimental chemists to synthesize efficient small molecule acceptors. Effect of different electron‐releasing and electron‐withdrawing groups on pi‐spacer of SM acceptor was studied. Strong electron‐releasing and electron‐withdrawing groups significantly reduced the band gap. Electron‐withdrawing groups red‐shifted the absorption spectra. A reverse relationship was observe between reorganization energy and Hammett parameters. SM acceptor with electron‐withdrawing groups showed higher charge mobility. [ABSTRACT FROM AUTHOR]

[Display omitted] • The key parameters of HOMO, LUMO, and λ max values associated with the organic molecules was utilized to perform machine learning (ML) analysis. • The dataset of each parameter was utilized to train the model and the validity of the models was tested against the test dataset. • The Hist gradient boosting (HGB) model exhibited best results for the prediction of the understudy parameters. • The fabrication of high-performance organic solar cells (OSCs) devices was also suggested by using the ML models. Typical research design associated with organic solar cells (OSCs) is conventionally considered time-consuming and laborious because the selection of the materials as the core, pi-acceptor, and terminal groups required for the engineering of these devices is done via hit and trial methodology. The advanced data-driven approaches, particularly machine learning (ML), have materialized as the robust technique for identifying the organic materials for the fabrication of the OSCs devices. The key parameters of highest occupied molecular orbital (HOMO), lowest unoccupied molecular orbital (LUMO), and maximum absorption wavelength (λ max) were selected for developing the ML models. The molecular descriptor associated with each parameter was investigated and the relative contribution of the understudy descriptors in the training of the ML model was studied by using the relative importance test. The Hist gradient boosting (HGB) model exhibited the best results for performing the predictive analysis of all three parameters. Moreover, the chemical database was constructed based on the academic literature to develop the high-performance OSCs devices, and the trained HGB model was applied to predict the HOMO, LUMO, and λ max values for these newly designed OSCs devices. Synthetic accessibility of designed molecules is also predicted which revealed that the suggested new organic molecules can be easily commercialized via experimentation. Highly encouraging results in terms of the understudy key parameters were acquired by this ML approach indicating that the data-driven approaches hold extreme potential for engineering high-performance OSCs devices. [ABSTRACT FROM AUTHOR]

• Machine learning assisted designing of hole-transporting materials for high performance perovskite solar cells. • About 04 machine learning regresssor models are applied for predicting reorganization energy (Rh). • Chemical similarity analysis is used for screening potential candidates for perovskite solar cells. • 30 potential compounds are identified that could be synthesized with ease. In recent years, the advancement of perovskite solar cells has accelerated, leading to continuous performance improvements. Over the past few years, machine learning (ML) has gained popularity among scientists researching perovskite solar cells. In this study, ML is used to screen hole-transporting materials for perovskite solar cells. To construct machine-learning (ML) models, data from prior investigations are collected. Out of four machine learning algorithms trained for predicting reorganization energy (Rh), the gradient boosting regression model stood out as the most effective, attaining an R2 value of 0.89. Data visualization analysis is then utilized to scrutinize the patterns within the dataset. 10,000 new compounds are generated. Chemical space of generated compounds is visualized using various measures. Minor structural modifications resulted in only a slight alteration in reorganization energy (Rh). The newly introduced multidimensional framework has the potential to efficiently screen materials in a short amount of time. [ABSTRACT FROM AUTHOR]

Bandgap is a key parameter for selecting suitable materials for a broad range of applications. Organic solar cells (OSCs) are emerging as powerful devices due to their low-cost solution processing. Developing OSCs necessitates producing effective materials in a computationally cost-effective and rapid manner. Machine learning has become popular and well-recognized among researchers to screen and design high performance materials for OSCs. Machine learning models require data from the literature (reported studies or databases) to effectively predict targeted properties. To unveil the hidden dataset patterns, a thorough data visualization analysis is conducted. Importantly, multiple database mining is performed for designing low band gap organic semiconductors. Molecular descriptors are utilized to train machine learning models. Importantly, about 22 different machine learning models are tested. Among all models, extra trees regressor shows higher predictive capability. Residuals, learning curve and validation curve are also drawn for extra trees regressor. Feature importance analysis determines the significance of the features. Moreover, library enumeration and similarity analysis further facilitate designing of high-performance semiconductor materials. This work may help in screening and designing efficient semiconductors having low band gap for increasing the efficiency of OSCs. [Display omitted] • Machine learning is applied to design new low band gap organic semi-conductors for OSCs. • Data mining and property prediction strategies are applied to screen potential candidates for OSCs. • The chemical similarity analysis and library enumeration techniques are performed. • More than >20 different regression models are developed and used for better prediction. [ABSTRACT FROM AUTHOR]

Designing effective materials for organic solar cells (OSCs) is a challenging and time-consuming process. To achieve high performance OSCs, efficient designing/screening of materials is essential. In recent years, machine learning (ML) has captured the attention of the scientific community working on OSCs. In present study, efficiency of building blocks is predicted by using different ML models. Machine learning analysis is performed for predicting power conversion efficiency (PCE) as a dependent variable and molecular descriptors as independent factors. Moreover, similarity analysis (Tanimoto similarity) is used to screen structures based on the similarity between structures present in the databases and reference (given) structures. RDkit is used to calculate Tanimoto index and compare the fingerprints of molecules present within the database with fingerprint of reference/query structure. The monomer of three famous polymer donors PM6, PBT7-Th and PDPP3T are used as reference molecules for similarity analysis. The best buildings blocks are selected based on the results obtained from similarity analysis. The high efficiency screened building units are connected to design new polymers. PCE values of newly designed monomers are predicted using already trained machine learning models. This proposed framework can screen and design effective polymers for OSCs and predict their PCE without any experimentation in minimum time with marginal computational cost. [Display omitted] • Machine learning based approach is used to design new polymers for organic solar cells. • New insights are provided for virtual screening of building blocks for designing polymer materials. • Similarity analysis is performed to screen easily synthesizable organic building blocks. • Among newly designed polymers, 4 polymers show power conversion efficiency of above 12%. • Regression analysis and shapley additive explanations are performed. [ABSTRACT FROM AUTHOR]

[Display omitted] • Machine learning analysis based approach is used to design new organic semiconductors for organic solar cells. • New insights are provided for high throughput screening and reorganization energy prediction. • Similarity analysis is performed to screen easily synthesizable organic building blocks to design > 1000 new organic semiconductors. • Nearly ∼ 22 regression models were developed and selected based on better predictive capability. Organic solar cells (OSCs) are ecofriendly and an inexpensive source of electricity production. However, high-throughput screening and designing new materials without performing trial-and-error experimental procedures is essential for the future commercialization of OSCs. Herein, a machine learning assisted approach is applied to design efficient organic semiconductors for OSCs in a fast and computationally cost-effective manner. Experimental and theoretical data from previous studies (databases) is collected for training of machine learning models to predict various properties of organic semiconductor materials such as reorganization energy. Moreover, high-throughput screening is performed to screen potential materials for OSCs. To evaluate the database's trends, data visualization analysis is performed. Moreover, Cook's distance is used to detect outliers in the machine learning models. Importantly, out of 22 tested models, only two models i.e., random forest regressor and extra trees regressor have shown better predictive capability. To check the applicability of this innovative approach, >1000 new organic semiconductors are designed by utilizing easily synthesizable organic building blocks. This machine learning approach can be used for high-throughput screening and designing of efficient materials for OSCs. [ABSTRACT FROM AUTHOR]