Comparison of Machine Learning Models on Performance of Single- and Dual-Type Electrochromic Devices

Toward the Prediction of Electrochromic Properties of WO3 Films: Combination of Experimental and Machine Learning Approaches

WO₃ is the state of the art of electrochromic oxide materials finding technological application in smart windows. In this work, a set of WO₃ thin films were deposited by magnetron sputtering by varying total pressure, oxygen partial pressure, and power. On each film two properties were measured, the electrochemical reversibility and the blue color persistence of Li_xWO₃ films in simulated ambient conditions. With the help of machine learning, prediction maps for such electrochromic properties, namely, color persistence and reversibility, were designed. High-performance WO₃ films were targeted by a global score which is the product of these two properties. The combined approach of experimental measurements and machine learning led to a complete picture of electrochromic properties depending of sputtering parameters providing an efficient tool in regards to time saving.

Evaluation of Transparent ITO/Nano-Ag/ITO Electrode Grown on Flexible Electrochromic Devices by Roll-to-Roll Sputtering Technology

This paper explores the flexible ITO/nano-Ag/ITO multilayer electrodes grown on polyethylene terephthalate (PET) substrates and processed by a continuous roll-to-roll (R2R) sputtering system at room temperature used for flexible electrochromic device (ECD) applications. The effect of the nano-Ag interlayer thickness on the electrical and optical properties of the flexible ITO/nano-Ag/ITO multilayer was thoroughly investigated. By using R2R-sputtered at an Ag DC power of 0.2 kW, we were able to achieve optimal ITO/nano-Ag/ITO multilayer that exhibits a high optical transmittance of 87.19% and the best figure of merit value (30.93 × 10−3 Ω−1). The EC performance and stability of the flexible devices were tested by a cathodic WO3 coloration. Coloring and bleaching tests show that ITO/nano-Ag/ITO multilayers are highly effective conductors, indicating that the R2R sputtering technique is a promising continuous sputtering process in preparing for the fabrication of optical devices and flexible electronics industries.

A Machine Learning Approach for Metal Oxide Based Polymer Composites as Charge Selective Layers in Perovskite Solar Cells

A library of metal oxide-conjugated polymer composites was prepared, encompassing WO₃ -polyaniline (PANI), WO₃ -poly(N-methylaniline) (PMANI), WO₃ -poly(2-fluoroaniline) (PFANI), WO₃ -polythiophene (PTh), WO₃ -polyfuran (PFu) and WO₃ -poly(3,4-ethylenedioxythiophene) (PEDOT) which were used as hole selective layers for perovskite solar cells (PSCs) fabrication. We adopted machine learning approaches to predict and compare PSCs performances with the developed WO₃ and its composites. For the evaluation of PSCs performance, a decision tree model that returns 0.9656 R² score is ideal for the WO₃ -PEDOT composite, while a random forest model was found to be suitable for WO₃ -PMANI, WO₃ -PFANI, and WO₃ -PFu with R² scores of 0.9976, 0.9968, and 0.9772 respectively. In the case of WO₃ , WO₃ -PANI, and WO₃ -PTh, a K-Nearest Neighbors model was found suitable with R² scores of 0.9975, 0.9916, and 0.9969 respectively. Machine learning can be a pioneering prediction model for the PSCs performance and its validation.

Physical Simulation Model of WO3 Electrochromic Films Based on Continuous Electron-Transfer Kinetics and Experimental Verification

Tungsten oxide (WO₃) electrochromic devices have attracted a lot of interest in the energy conservation field and have shown a preliminary application potential in the market. However, it is difficult to quantitatively direct experiments with the existing electrochromic theoretical models, which can restrict the further development of electrochromism. Here, an electrochromic physical simulation model of WO₃ films was built to solve the above problem. Experimentally, the actual electrochromic kinetics of WO₃ in the LiClO₄/propylene carbonate electrolyte was determined as a continuous electron-transfer process by cyclic voltammetry measurement and X-ray photoelectron spectroscopy analysis. Theoretically, the continuous electron-transfer process, Li⁺-ion diffusion process, and the transmittance change process were described by a modified Butler-Volmer equation, Fick's law, and charge versus coloration efficiency/bleaching efficiency coupling equation, respectively. The comparisons between theoretical and experimental data were conducted to verify this model. The shape of the simulated current curves was basically consistent with that of experiments. Besides, the difference of transmittance between the simulation and experiments was less than 8%. The difference between theory and experiment was attributed to the influence of the electric double layer and the actual reaction interface. The success of the simulation was attributed to the accurate description of the electrochromic process by continuous electron-transfer kinetics. This model can be applied in the research of electrochromic mechanisms, experimental result prediction, and novel device development due to its clear physical nature.