Figure 3 - uploaded by Gongqiang Li
Content may be subject to copyright.
Predicted PCEs of model B trained by a) RF, b) GBRT, c) SVM, d) ANN (one hidden layer), and e) GBRT-SVM-ANN learners versus experimental PCEs for the training set and test set.
Source publication
The development of highly efficient dye‐sensitized solar cells (DSSCs) is greatly hindered by the lack of a reliable and understandable quantitative structure‐property relationship (QSPR) model. Herein, an accurate, robust and interpretable QSPR model is established by combining the machine learning technique and computational quantum chemistry, an...
Contexts in source publication
Context 1
... predictive performance of model A trained by different ML learners is shown in Table 1 and Figure S2, Supporting Information. The accuracy through cross-validation of both tree-based RF (r val ¼ 0.57) and GBRT (r val ¼ 0.57) models is A set of extended descriptors obtained by the frequency and excited-state calculations are further incorporated to train model B. As shown in Figure 3 and Table 1, the values of r val and r test of all learners are both greater than 0.70, indicating that the performance of model B is significantly improved as compared with model A, especially for RF (r val ¼ 0.75, r test ¼ 0.70, MAE ¼ 0.86, and RMSE ¼ 1.08) and GBRT (r val ¼ 0.76, r test ¼ 0.76, MAE ¼ 0.78, and RMSE ¼ 0.97) learners. We suspect that the tree-based learners need more complex descriptors to capture the subtle features of dyes in a small data. ...
Context 2
... on the predicted PCEs and chemical stability, 500 high-performance molecules are identified for more accurate screening. The distribution of predicted PCEs of the preselected 500 molecules using model B trained by different learners is shown in Figure S3, Supporting Information (see Supporting Data 2 for the predicted PCEs). It is important to note that extracting useful information for dye design from the virtual screening is more meaningful than the specific values of predicted PCEs. ...
Citations
... In addition, ML-based approaches can predict several photophysical and photochemical properties, such as λ abs , λ em , Stokes shifts, molar absorption, UV-visible-NIR spectra, excitation energy, optical properties, etcetera [6][7][8][9][10][11][12][13][14][15][16]. Aside from these characteristics, data-driven ML models have predicted the power conversion efficiency of dye-sensitized solar cells [17][18][19][20]. Furthermore, utilizing ML screening models, comprehend the aggregation-induced emission effect of fluorescent materials [21][22][23]. ...
... However, such approaches are timeconsuming and complex in nature. In recent years, machine learning (ML) techniques have been used as a very effective tool for studying and estimating material properties, including photophysical properties, quite successfully [6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23]. These approaches are data-driven, statistical in nature, and quite easy to operate. ...
... The aforesaid data-based techniques affording essential understanding of the intrinsic properties of the photocatalytic system could be a prospective tool for discovery of the highly efficient dye-sensitized TiO 2 photocatalytic system for water splitting and H 2 evolution reaction [131]. The below Table 4 describes the summary of ML algorithms applied to data bases of materials and dyes used in various applications such as in Agro waste [132], fabricating dye sensitized solar cells (DSSC) and optical sensors [133][134][135][136][137][138][139], and photocatalytic splitting of water into hydrogen production [140][141][142]. The most commonly used ML algorithms are ANN, RF, GBRT, and DT. ...
... In 2020, Wen et al. [86] utilized a combination of quantitative structure-property relationships and ML to identify novel dyes for DSSCs. They explored a database consisting of nearly 10,000 dyes and employed various ML algorithms, including RF, GBR Tree, SVM, and ANN, to mine the database. ...
... Structure of two new molecules with the highest efficiency[86]. ...
This study explores the potential use of Machine Learning (ML) techniques to enhance three types of nano-based solar cells. Perovskites of methylammonium-free formamidinium (FA) and mixed cation-based cells exhibit a boosted efficiency when employing ML techniques. Moreover, ML methods are utilized to identify optimal donor complexes, high blind temperature materials, and to advance the thermodynamic stability of perovskites. Another significant application of ML in dye-sensitized solar cells (DSSCs) is the detection of novel dyes, solvents, and molecules for improving the efficiency and performance of solar cells. Some of these materials have increased cell efficiency, short-circuit current, and light absorption by more than 20%. ML algorithms to fine-tune network and plasmonic field bandwidths improve the efficiency and light absorption of surface plasmonic resonance (SPR) solar cells. This study outlines the potential of ML techniques to optimize and improve the development of nano-based solar cells, leading to promising results for the field of solar energy generation and supporting the demand for sustainable and dependable energy.
... By understanding the characteristics of natural sensitized Maddah investigated to enhance the performance of DSSCs through ML [260]. Wen et al illustrated a quantitative structure-property relationship model by combining MLand computational quantum chemistry for exploring various organic dyes capable of being integrated in organic DSSCs [261]. Al-Sabana and Abdellatif established an attempt to used ML algorithm, random-forest in optimization process for the DSSCs in tern of efficiency. ...
Flexible dye-sensitized solar cells (FDSSCs) show a huge potential for stretchable electronics and portable power sources due to their lightweight, handy, flexibility, cost-effective, and easy processing. This paper introduces basic operating principles and design opportunities for maximum efficiencies for FDSSCs. Flexible polymers or metal substrates, enabling cost reduction due to large volume production with roll to roll manufacturing technique. DSSCs achieved a PCE of 14.30 % on rigid conductive substrates, 10.28 % on flexible metal substrates, and 8 % on plastic substrates. A brief distinction has been made on different substrates, preparation of charge transfers materials, coating and printing techniques and processing methods for enhancing the performance of FDSSCs. We also highlight issues pertaining to progress in the stability of devices and the commercialisation of FDSCs technologies will be explained.
... Yaping et al. [87] combined ML with computational quantum chemistry results in the establishment of an accurate, reliable, and interpretable QSPR model. Using this model, virtual screening as well as the evaluation of synthetic accessibility are carried out to find new effective and synthetically accessible organic dyes for DSSCs. ...
Solar photovoltaic (PV) technology has merged as an efficient and versatile method for converting the Sun's vast energy into electricity. Innovation in developing new materials and solar cell architectures is required to ensure lightweight, portable, and flexible miniaturized electronic devices operate for long periods with reduced battery demand. Recent advances in biomedical implantable and wearable devices have coincided with a growing interest in efficient energy-harvesting solutions. Such devices primarily rely on rechargeable batteries to satisfy their energy needs. Moreover, Artificial Intelligence (AI) and Machine Learning (ML) techniques are touted as game changers in energy harvesting, especially in solar energy materials. In this article, we systematically review a range of ML techniques for optimizing the performance of low-cost solar cells for miniaturized electronic devices. Our systematic review reveals that these ML techniques can expedite the discovery of new solar cell materials and architectures. In particular, this review covers a broad range of ML techniques targeted at producing low-cost solar cells. Moreover, we present a new method of classifying the literature according to data synthesis, ML algorithms, optimization, and fabrication process. In addition, our review reveals that the Gaussian Process Regression (GPR) ML technique with Bayesian Optimization (BO) enables the design of the most promising low-solar cell architecture. Therefore, our review is a critical evaluation of existing ML techniques and is presented to guide researchers in discovering the next generation of low-cost solar cells using ML techniques.
... For example, using DFT based calculations one can create databases for electronic and structural properties for a large number of molecules. These databases can be further evaluated to select effective systems for photo-induced charge transfer processes [19]. These databases can also be used for machine learning and for further molecular modelling [3,20,21]. ...
... Moreover, the efficiency and its research challenges towards DSSCs have been clearly reviewed [136]. To solve such problems, recently, the discovery of new materials, such as 2D and high selectivity catalysts, have been emerged as promising materials, and their identifications have been identified by machine learning data-driven approach [137]. Especially, indoor solar cell is a strong positive influence on the ecology of the Internet of Things (IoTs). ...
Production of green energy by using environment friendly and cost-effective components is attracting the attention of the research world and is found to be a promising approach to replace nonrenewable energy sources. Among the green energy sources, dye-sensitized solar cells (DSSCs) are found to be the most alternative way to reduce the energy demand crises in current situation. The efficiency of DSSCs is dependent on numerous factors such as the solvent used for dye extraction, anode and cathode electrodes, and the thickness of the film, electrolyte, dye, and nature of FTO/ITO glasses. The efficiency of synthetic dye-based DSSCs is enhanced as compared to their counterparts. However, it has been found that many of the synthetic sensitizers used in DSSCs are toxic, and some of them are found to cause carcinogenicity in nature by forming a complex agent. Instead, using various parts of green plants such as leaves, roots, steam, peel waste, flowers, various spices, and mixtures of them would be a highly environmentally friendly and good efficient. The present review focuses on and summarizes the efficiency affecting factors, the various categories of natural sensitizers, and solvent effects. Furthermore, the review work assesses the experimentally and computationally obtained values and their progress in development.
... Various machine learning (ML) models could be applied to efficiently screen out potential candidates with higher power conversion efficiency (PCE) [11][12][13][14][15][16][17][18][19][20][21], open-circuit voltage (V OC ) [20][21][22], short-circuit current (J SC ) [21][22][23], and fill factor (FF) [21,22] of OSCs from the experimental databases and the high throughput density functional theory (DFT) calculations. As a key factor to achieve high V OC and photocurrent simultaneously, the CT state is believed to be the intermediate between exciton dissociation and charge extraction across the D/A interface [24][25][26][27]. ...
Charge transfer and transport properties are crucial in the photophysical process of exciton dissociation and recombination at the donor/acceptor (D/A) interface. Herein, machine learning (ML) is applied to predict the charge transfer state energy (ECT) and identify the relationship between ECT and intermolecular packing structures sampled from molecular dynamics (MD) simulations on fullerene- and non-fullerene-based systems with different D/A ratios (RDA), oligomer sizes, and D/A pairs. The gradient boosting regression (GBR) exhibits satisfactory performance (r = 0.96) in predicting ECT with π-packing related features, aggregation extent, backbone of donor, and energy levels of frontier molecular orbitals. The charge transport property affected by π-packing with different RDA has also been investigated by space-charge-limited current (SCLC) measurement and MD simulations. The SCLC results indicate an improved hole transport of non-fullerene system PM6/Y6 with RDA of 1.2:1 in comparison with the 1:1 counterpart, which is mainly attributed to the bridge role of donor unit in Y6. The reduced energetic disorder is correlated with the improved miscibility of polymer with RDA increased from 1:1 to 1.2:1. The morphology-related features are also applicable to other complicated systems, such as perovskite solar cells, to bridge the gap between device performance and microscopic packing structures.
... This study helps to synthesize new sensitizers with desired properties and UV-vis spectra. Similarly, Yaping Wen et al 43 In this review, the authors present an extensive analysis of ML techniques used in predicting the design and efficiency of DSSCs. The review gives an analysis of the existing techniques employed for finding the best suitable design, optimum efficiency, and best suitable material for the fabrication of DSSC. ...
... Therefore, many teams of researchers applied ML approaches to determine the nature and optimum values of these attributes for achieving the highest performance of DSSC. 43,117,118 The Dye is responsible for harvesting the solar energy, generating electrons, and transferring the electrons to the conduction band of semiconductor oxide in DSSC. Different types of dye sensitizers such as organic, 119 inorganic, 120 and natural 121 have been used to fabricate DSSC. ...
... They identified the efficient and synthetically accessible organic dyes for DSSC using the QSPR model. 43 Hosseinnezhad et al employed the ANN and Genetic Algorithm (GA) to determine the device's efficiency using dye aggregation as an input. 100 They explained that the hybrid model of ANN and GA could predict the optimum efficiency of DSSC based on the parameters such as the volume ratio of organic dyes, the concentration of antiaggregation, and temperature. ...
Photovoltaic technology attracts researchers from industry and academia due to its potential in producing electricity directly from the sunlight. Among all the photovoltaic devices, the dye‐sensitized solar cell has gained preference due to its low‐cost fabrication and versatility in electrolytes, dye, substrate, and catalyst. The optical, electrical, and structural properties of the materials determine the power conversion efficiency of a solar cell. But, conducting experiments in the laboratory to identify the suitable materials for the fabrication of an efficient solar cell requires much time, cost, and human effort. The proven potential of machine learning techniques in pattern matching and computer vision motivated the researchers to employ these techniques for predicting the efficiency of solar cells. The research works conducted so far show the applications of these techniques in predicting the optimum efficiency, best suitable design, and material for the fabrication of Dye‐Sensitized Solar Cells (DSSCs). In this paper, the authors present a comprehensive review of the machine learning techniques employed and the types of input data used for predicting the design and efficiency of solar cells. They also give essential insights into the selection of optimum parameters for selecting the materials for fabricating a substrate, dye sensitizer, semiconductor, electrolyte, and catalyst for designing the most efficient dye‐sensitized solar cell without conducting experiments in the laboratory. This paper may prove a time and cost‐saving assistant for developing a customized neural network model for predicting the efficiency of a DSSC from the dataset available in the literature. Artificial Neural Network (ANN) model is useful to identify the suitable materials for efficient DSSC assembly. The tailored neural network models minimize the need for hit and trial experiments. The hybrid of ANN and Genetic Algorithm (GA) offers a low‐cost technological solution for DSSC assembly. The experimental data are vital for extracting useful insights for DSSC fabrication. ANN for predicting properties of an efficient DSSC
... Specifically, the works in Refs. [13][14][15][16] were introduced for machine learning algorithms related to DSSC optimization. In Ref. [14], a valuable attempt was demonstrated to optimize the sensitizers used in DSSC fabrication. ...
... In Ref. [14], a valuable attempt was demonstrated to optimize the sensitizers used in DSSC fabrication. Alternatively, Ref. [15] illustrated an artificial integument model for exploring various organic dyes capable of being integrated in organic DSSCs. However, by screening the literature, no previous investigation on the optimization of the active mesoporous layer for maximizing DSSC efficiency was addressed. ...
This paper provides an attempt to utilize machine learning algorithm, explicitly random-forest algorithm, to optimize the performance of dye sensitized solar cells (DSSCs) in terms of conversion efficiency. The optimization is implemented with respect to both the mesoporous TiO2 active layer thickness and porosity. Herein, the porosity impact is reflected to the model as a variation in the effective refractive index and dye absorption. Database set has been established using our data in the literature as well as numerical data extracted from our numerical model. The random-forest model is used for model regression, prediction, and optimization, reaching 99.87% accuracy. Perfect agreement with experimental data was observed, with 4.17% conversion efficiency.
