The dye-sensitized solar cell database

Automatic Prediction of Peak Optical Absorption Wavelengths in Molecules Using Convolutional Neural Networks

Molecular design depends heavily on optical properties for applications such as solar cells and polymer-based batteries. Accurate prediction of these properties is essential, and multiple predictive methods exist, from ab initio to data-driven techniques. Although theoretical methods, such as time-dependent density functional theory (TD-DFT) calculations, have well-established physical relevance and are among the most popular methods in computational physics and chemistry, they exhibit errors that are inherent in their approximate nature. These high-throughput electronic structure calculations also incur a substantial computational cost. With the emergence of big-data initiatives, cost-effective, data-driven methods have gained traction, although their usability is highly contingent on the degree of data quality and sparsity. In this study, we present a workflow that employs deep residual convolutional neural networks (DR-CNN) and gradient boosting feature selection to predict peak optical absorption wavelengths (λmax) exclusively from SMILES representations of dye molecules and solvents; one would normally measure λmax using UV-vis absorption spectroscopy. We use a multifidelity modeling approach, integrating 34,893 DFT calculations and 26,395 experimentally derived λmax data, to deliver more accurate predictions via a Bayesian-optimized gradient boosting machine. Our approach is benchmarked against the state of the art that is reported in the scientific literature; results demonstrate that learnt representations via a DR-CNN workflow that is integrated with other machine learning methods can accelerate the design of molecules for specific optical characteristics.

Autonomous, multiproperty-driven molecular discovery: From predictions to measurements and back

A closed-loop, autonomous molecular discovery platform driven by integrated machine learning tools was developed to accelerate the design of molecules with desired properties. We demonstrated two case studies on dye-like molecules, targeting absorption wavelength, lipophilicity, and photooxidative stability. In the first study, the platform experimentally realized 294 unreported molecules across three automatic iterations of molecular design-make-test-analyze cycles while exploring the structure-function space of four rarely reported scaffolds. In each iteration, the property prediction models that guided exploration learned the structure-property space of diverse scaffold derivatives, which were realized with multistep syntheses and a variety of reactions. The second study exploited property models trained on the explored chemical space and previously reported molecules to discover nine top-performing molecules within a lightly explored structure-property space.

Machine Learning Models for Predicting Molecular UV-Vis Spectra with Quantum Mechanical Properties

Accurate understanding of ultraviolet-visible (UV-vis) spectra is critical for the high-throughput synthesis of compounds for drug discovery. Experimentally determining UV-vis spectra can become expensive when dealing with a large quantity of novel compounds. This provides us an opportunity to drive computational advances in molecular property predictions using quantum mechanics and machine learning methods. In this work, we use both quantum mechanically (QM) predicted and experimentally measured UV-vis spectra as input to devise four different machine learning architectures, UVvis-SchNet, UVvis-DTNN, UVvis-Transformer, and UVvis-MPNN, and assess the performance of each method. We find that the UVvis-MPNN model outperforms the other models when using optimized 3D coordinates and QM predicted spectra as input features. This model has the highest performance for predicting UV-vis spectra with a training RMSE of 0.06 and validation RMSE of 0.08. Most importantly, our model can be used for the challenging task of predicting differences in the UV-vis spectral signatures of regioisomers.

Can range-separated functionals be optimally tuned to predict spectra and excited state dynamics in photoactive iron complexes?

Density functional theory is an efficient computational tool to investigate photophysical and photochemical processes in transition metal complexes, giving invaluable assistance in interpreting spectroscopic and catalytic experiments. Optimally tuned range-separated functionals are particularly promising, as they were created to address some of the fundamental deficiencies present in approximate exchange-correlation functionals. In this paper, we scrutinize the selection of optimally tuned parameters and its influence on the excited state dynamics, using the example of the iron complex [Fe(cpmp)₂]²⁺ with push-pull ligands. Various tuning strategies are contemplated based on pure self-consistent DFT protocols, as well as on the comparison with experimental spectra and multireference CASPT2 results. The two most promising sets of optimal parameters are then employed to carry out nonadiabatic surface-hopping dynamics simulations. Intriguingly, we find that the two sets lead to very different relaxation pathways and timescales. While the set of optimal parameters from one of the self-consistent DFT protocols predicts the formation of long-lived metal-to-ligand charge transfer triplet states, the set in better agreement with CASPT2 calculations leads to deactivation in the manifold of metal-centered states, in better agreement with the experimental reference data. These results showcase the complexity of iron-complex excited state landscapes and the difficulty of obtaining an unambiguous parametrization of long-range corrected functionals without experimental input.

Technoeconomic Analysis of Dye Sensitized Solar Cells (DSSCs) with WS2/Carbon Composite as Counter Electrode Material

Exploration of clean and renewable energy materials is necessary due to the coming energy crisis and environmental problems. Solar energy is one of the favorable energy sources because of the continuous energy reservoir and its affluence. Silicon-based solar devices are expensive due to their complicated production process, which limits this technology for urban and other commercial applications. Among the third generation of solar cells, Dye-Sensitized Solar Cells (DSSCs) have attracted widespread attention as potential cost-effective alternatives to silicon-based solar cells. In this paper, the commercializing potential of the DSSCs is investigated. A module is introduced where the materials, equipment, and distribution of direct manufacturing costs are calculated. The manufacturing costs and the Levelized Cost of Energy (LCOE) of these DSSCs for a system lifetime of 25 years were determined to be USD 22.40 per m2 and USD 0.0438 per kWh and the module price of this technology is USD 0.18 per W and the total installed system cost is USD 0.88 per W in Kansas which suggest that this PV technology could challenge other leading PV technologies.

Perovskite- and Dye-Sensitized Solar-Cell Device Databases Auto-generated Using ChemDataExtractor

The number of scientific publications reporting cutting-edge third-generation photovoltaic devices is increasing rapidly, owing to the pressing need to develop renewable-energy technologies that address the climate-change crisis. Consequently, the field could benefit from a central repository where photovoltaic-performance metrics, such as the power-conversion efficiency (η) are recorded. We present two automatically generated databases that contain photovoltaic properties and device material data for dye-sensitized solar cells (DSCs) and perovskite solar cells (PSCs), totalling 660,881 data entries representing 57,678 photovoltaic devices. The databases were generated by applying the text-mining toolkit ChemDataExtractor on a corpus of 25,720 articles. A multi-faceted evaluation, incorporating manual and automatic methods, was applied to ensure that the data contained therein were of the highest quality, with precision metrics ranging from 73.1% to 95.8%. The DSC database contains 475,045 entries representing 41,680 devices, and the PSC database contains 185,836 entries representing 15,818 devices. The databases are available in MongoDB and JSON formats, which can be queried in Python, R, Java and MATLAB for data-driven photovoltaic materials discovery.

Multi-fidelity prediction of molecular optical peaks with deep learning

Optical properties are central to molecular design for many applications, including solar cells and biomedical imaging. A variety of ab initio and statistical methods have been developed for their prediction, each with a trade-off between accuracy, generality, and cost. Existing theoretical methods such as time-dependent density functional theory (TD-DFT) are generalizable across chemical space because of their robust physics-based foundations but still exhibit random and systematic errors with respect to experiment despite their high computational cost. Statistical methods can achieve high accuracy at a lower cost, but data sparsity and unoptimized molecule and solvent representations often limit their ability to generalize. Here, we utilize directed message passing neural networks (D-MPNNs) to represent both dye molecules and solvents for predictions of molecular absorption peaks in solution. Additionally, we demonstrate a multi-fidelity approach based on an auxiliary model trained on over 28 000 TD-DFT calculations that further improves accuracy and generalizability, as shown through rigorous splitting strategies. Combining several openly-available experimental datasets, we benchmark these methods against a state-of-the-art regression tree algorithm and compare the D-MPNN solvent representation to several alternatives. Finally, we explore the interpretability of the learned representations using dimensionality reduction and evaluate the use of ensemble variance as an estimator of the epistemic uncertainty in our predictions of molecular peak absorption in solution. The prediction methods proposed herein can be integrated with active learning, generative modeling, and experimental workflows to enable the more rapid design of molecules with targeted optical properties.

QSPR modeling of absorption maxima of dyes used in dye sensitized solar cells (DSSCs)

Dye-sensitized solar cells (DSSCs) have recently received a significant attention as possible sources of renewable energy. As a result, a significant effort is being made to develop organic dyes for highly power conversion efficient DSSCs, in order to overcome the disadvantages of previous solar cell systems, such as cost reduction, weight reduction, and production methods that minimize environmental pollution. As shown by multiple recent research publications, computational techniques such as quantitative structure-property relationship (QSPR) modeling may aid in the development of suitable dyes for DSSCs satisfying many fundamental desired characteristics. The current report provides robust, externally verified QSPR models for five chemical classes of organic dyes (Triphenylamines, Phenothiazines, Indolines, Porphyrins and Coumarins) based on experimentally determined absorption maxima values. The size of the dye data points utilized to develop the models is the largest known to date. The QSPR models were constructed using only two-dimensional descriptors with clear physicochemical meaning. Using the best subset selection approach, we built 5, 3, 4, 3 and 2 descriptor models for the Triphenylamine, Phenothiazine, Indoline, Porphyrin and Coumarin classes, respectively. The models were validated both internally and externally, and then consensus predictions were made for specific categories of dyes using the developed partial least squares (PLS) models, and the "Intelligent consensus predictor" tool (http://teqip.jdvu.ac.in/QSAR_Tools/) was used to determine whether the quality of test set compound predictions can be improved through the "intelligent" selection of multiple PLS models. We identified from the insights gained from the developed models several chemical attributes that are important in enhancing the absorption maxima. Thus, our study may be utilized to predict the λ_max values of novel or untested organic dyes and to give insights that will aid in the development of new dyes for use in solar cells with increased λ_max values and enhanced power conversion efficiency.

Data-driven approach towards identifying dyesensitizer molecules for higher power conversion efficiency in solar cells

Machine learning (ML) research based on the quantitative structure–property relationship (QSPR) has been applied for the development of highly efficient dye-sensitized solar cells (DSSCs).

UV-adVISor: Attention-Based Recurrent Neural Networks to Predict UV-Vis Spectra

Ultraviolet-visible (UV-Vis) absorption spectra are routinely collected as part of high-performance liquid chromatography (HPLC) analysis systems and can be used to identify chemical reaction products by comparison to the reference spectra. Here, we present UV-adVISor as a new computational tool for predicting the UV-Vis spectra from a molecule's structure alone. UV-Vis prediction was approached as a sequence-to-sequence problem. We utilized Long-Short Term Memory and attention-based neural networks with Extended Connectivity Fingerprint Diameter 6 or molecule SMILES to generate predictive models for the UV spectra. We have produced two spectrum datasets (dataset I, N = 949, and dataset II, N = 2222) using different compound collections and spectrum acquisition methods to train, validate, and test our models. We evaluated the prediction accuracy of the complete spectra by the correspondence of wavelengths of absorbance maxima and with a series of statistical measures (the best test set median model parameters are in parentheses for model II), including RMSE (0.064), R² (0.71), and dynamic time warping (DTW, 0.194) of the entire spectrum curve. Scrambling molecule structures with the experimental spectra during training resulted in a degraded R², confirming the utility of the approaches for prediction. UV-adVISor is able to provide fast and accurate predictions for libraries of compounds.

Dynamics of photoconversion processes: the energetic cost of lifetime gain in photosynthetic and photovoltaic systems

The continued development of solar energy conversion technologies relies on an improved understanding of their limitations. In this review, we focus on a comparison of the charge carrier dynamics underlying the function of photovoltaic devices with those of both natural and artificial photosynthetic systems. The solar energy conversion efficiency is determined by the product of the rate of generation of high energy species (charges for solar cells, chemical fuels for photosynthesis) and the energy contained in these species. It is known that the underlying kinetics of the photophysical and charge transfer processes affect the production yield of high energy species. Comparatively little attention has been paid to how these kinetics are linked to the energy contained in the high energy species or the energy lost in driving the forward reactions. Here we review the operational parameters of both photovoltaic and photosynthetic systems to highlight the energy cost of extending the lifetime of charge carriers to levels that enable function. We show a strong correlation between the energy lost within the device and the necessary lifetime gain, even when considering natural photosynthesis alongside artificial systems. From consideration of experimental data across all these systems, the emprical energetic cost of each 10-fold increase in lifetime is 87 meV. This energetic cost of lifetime gain is approx. 50% greater than the 59 meV predicted from a simple kinetic model. Broadly speaking, photovoltaic devices show smaller energy losses compared to photosynthetic devices due to the smaller lifetime gains needed. This is because of faster charge extraction processes in photovoltaic devices compared to the complex multi-electron, multi-proton redox reactions that produce fuels in photosynthetic devices. The result is that in photosynthetic systems, larger energetic costs are paid to overcome unfavorable kinetic competition between the excited state lifetime and the rate of interfacial reactions. We apply this framework to leading examples of photovoltaic and photosynthetic devices to identify kinetic sources of energy loss and identify possible strategies to reduce this energy loss. The kinetic and energetic analyses undertaken are applicable to both photovoltaic and photosynthetic systems allowing for a holistic comparison of both types of solar energy conversion approaches.

Dye-sensitized solar cells strike back

Dye-sensitized solar cells (DSCs) are celebrating their 30th birthday and they are attracting a wealth of research efforts aimed at unleashing their full potential. In recent years, DSCs and dye-sensitized photoelectrochemical cells (DSPECs) have experienced a renaissance as the best technology for several niche applications that take advantage of DSCs' unique combination of properties: at low cost, they are composed of non-toxic materials, are colorful, transparent, and very efficient in low light conditions. This review summarizes the advancements in the field over the last decade, encompassing all aspects of the DSC technology: theoretical studies, characterization techniques, materials, applications as solar cells and as drivers for the synthesis of solar fuels, and commercialization efforts from various companies.

Catalytic Effect of 1,4-Dioxane on the Kinetics of the Oxidation of Iodide by Dicyanobis(bipyridine)iron(III) in Water

Dye-sensitized solar cells (DSSCs) are a technically and financially viable alternative to today’s photovoltaic systems using p-n junctions. The two functions are isolated here, which are unlike traditional systems where the semiconductor is thought to perform both light absorption and charge carrier transport. This article discusses the potential use of dicyanobis(bipyridine)iron(III) to oxidize iodide as a sensitizer in DSSCs. However, it is critical to understand the kinetics of this essential process in order to understand the mechanism of electron transport. The oxidation of iodide by dicyanobis(bipyridine)iron(III) in three reaction media was studied: water, 10% v/v 1,4-dioxane-water, and 20% v/v 1,4-dioxane-water. The reaction was carried out in a regular laboratory setting, with no special sensitive conditions or the use of expensive materials, making it a cost-effective and practical method. Dicyanobis(bipyridine)iron(III) oxidized iodide in selected media at 0.06 M ionic strength and constant temperature. The reaction was subjected to a spectrophotometric analysis. The data were acquired by measuring the rise in visible absorbance as a function of time after the formation of dicyanobis(bipyridine)iron(II). The reaction proceeded with an overall fractional (0.5), first order, and third order in water, 10% media, and 20% media, respectively. The presence of dicyanobis(bipyridine)iron(III) in either of the reaction media had no effect on the rate. The effect of protons (H+) on the rate constant indicated resistance in water and catalysis in dioxane-water media containing 10–20% dioxane. When the ionic strength was raised, there was no change in the rate constant in water, but there was a deceleration in both binary solvent media. In an aqueous medium, the thermodynamic parameters of activation were computed as Ea 46.23 kJ mol−1, 24.62 M s−1, ΔH# 43.76 kJ mol−1, ΔS# −226.5 J mol−1 K−1, and ΔG# 111.26 kJ mol−1 (25 °C). By increasing the rate of the reaction to its maximum, this study discovered the binary solvent media with the highest catalytic efficiency, i.e., 20% v/v 1,4-dioxane-water, which may increase the efficiency of DSSCs without using any expensive material or unusual experimental conditions.

Evaluation of Molecular Fingerprints for Determining Dye Aggregation on Semiconductor Surfaces

Dye aggregation plays an important role in determining the photovoltaic performance of dye sensitized solar cells. Compared with the spectra observed in solution, it is, apriori, difficult to ascertain whether a dye is likely to show hypsochromic (H) or bathochromic (J) aggregation, until after adsorption onto the semiconductor electrode. Herein, we show that molecular fingerprint-based methods provide a fast and efficient way to discriminate between H- and J-aggregating dyes. The efficacy of the fingerprint-based classification models is demonstrated with a diverse set of over 3000 organic dyes dissolved in different solvents. Requiring only the structure of the dye and the polarity of the solvent used, the machine learning model achieves close to 80 % classification accuracies that are comparable with models based on a combination of fragment counts and topological indices. For interested researchers, we have bundled the prediction tools as an R package.

An Open Access Data Set Highlighting Aggregation of Dyes on Metal Oxides

The adsorption of a dye to a metal oxide surface such as TiO2, NiO and ZnO leads to deprotonation and often undesirable aggregation of dye molecules, which in turn impacts the photophysical properties of the dye. While controlled aggregation is useful for some applications, it can result in lower performance for dye-sensitized solar cells. To understand this phenomenon better, we have conducted an extensive search of the literature and identified over 4000 records of absorption spectra in solution and after adsorption onto metal oxide. The total data set comprises over 3500 unique compounds, with observed absorption maxima in solution and after adsorption on the semiconductor electrode. This data may serve to provide further insight into the structure-property relationships governing dye-aggregation behaviour.

Comparative dataset of experimental and computational attributes of UV/vis absorption spectra

The ability to auto-generate databases of optical properties holds great prospects in data-driven materials discovery for optoelectronic applications. We present a cognate set of experimental and computational data that describes key features of optical absorption spectra. This includes an auto-generated database of 18,309 records of experimentally determined UV/vis absorption maxima, λ_max, and associated extinction coefficients, ϵ, where present. This database was produced using the text-mining toolkit, ChemDataExtractor, on 402,034 scientific documents. High-throughput electronic-structure calculations using fast (simplified Tamm-Dancoff approach) and traditional (time-dependent) density functional theory were executed to predict λ_max and oscillation strengths, f (related to ϵ) for a subset of validated compounds. Paired quantities of these computational and experimental data show strong correlations in λ_max, f and ϵ, laying the path for reliable in silico calculations of additional optical properties. The total dataset of 8,488 unique compounds and a subset of 5,380 compounds with experimental and computational data, are available in MongoDB, CSV and JSON formats. These can be queried using Python, R, Java, and MATLAB, for data-driven optoelectronic materials discovery.

Measurement(s)	ultraviolet–visible spectrum • absorption wavelength • extinction coefficient
Technology Type(s)	digital curation

Machine-accessible metadata file describing the reported data: 10.6084/m9.figshare.10304897

Prediction of Absorption Spectrum Shifts in Dyes Adsorbed on Titania

Dye adsorption on metal-oxide films often results in small to substantial absorption shifts relative to the solution phase, with undesirable consequences for the performance of dye-sensitized solar cells and optical sensors. While density functional theory is frequently used to model such behaviour, it is too time-consuming for rapid assessment. In this paper, we explore the use of supervised machine learning to predict whether dye adsorption on titania is likely to induce a change in its absorption characteristics. The physicochemical features of each dye were encoded as a numeric vector whose elements are the counts of molecular fragments and topological indices. Various classification models were subsequently trained to predict the type of absorption shift i.e. blue, red or unchanged (|Δλ| ≤ 10 nm). The models were able to predict the nature of the shift with a good likelihood (~80%) of success when applied to unseen data.

Aggregation-Free Organic Dyes Featuring Spiro[dibenzo[3,4:6,7]cyclohepta[1,2-b]quinoxaline-10,9'-fluorene] (SDBQX) for Dye-Sensitized Solar Cells

Three novel organic dyes coded as FHD4-1, FHD4-2, and FHD4-3 featuring spiro[dibenzo[3,4:6,7]cyclohepta[1,2-b]quinoxaline-10,9'-fluorene] (SDBQX) moieties are designed to inhibit dye aggregation to improve the performance of dye-sensitized solar cells (DSSCs). The consistent absorption onsets of FHD4-1, FHD4-2, and FHD4-3 in solutions and adsorbed on TiO2 films indicate that these dyes are aggregation-free dyes. Therefore, coadsorption with chenodeoxycholic acid (CDCA) of these three dyes reduces the performance of DSSCs because no inhibition effect for dye aggregation is needed, but, on the contrary, the dye loading amount is reduced after addition of CDCA.

Dye-Sensitized Solar Cells: Fundamentals and Current Status

Dye-sensitized solar cells (DSSCs) belong to the group of thin-film solar cells which have been under extensive research for more than two decades due to their low cost, simple preparation methodology, low toxicity and ease of production. Still, there is lot of scope for the replacement of current DSSC materials due to their high cost, less abundance, and long-term stability. The efficiency of existing DSSCs reaches up to 12%, using Ru(II) dyes by optimizing material and structural properties which is still less than the efficiency offered by first- and second-generation solar cells, i.e., other thin-film solar cells and Si-based solar cells which offer ~ 20-30% efficiency. This article provides an in-depth review on DSSC construction, operating principle, key problems (low efficiency, low scalability, and low stability), prospective efficient materials, and finally a brief insight to commercialization.