Multivariate Analysis of CO2 Electroreduction for Fuel and Value-Added Chemical Production Using Explainable Machine Learning
International Journal of Energy Research, cilt.2026, sa.1, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 2026 Sayı: 1
- Basım Tarihi: 2026
- Doi Numarası: 10.1155/er/2400202
- Dergi Adı: International Journal of Energy Research
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Aerospace Database, Compendex, Environment Index, INSPEC, Directory of Open Access Journals, Academic Search Ultimate (EBSCO), Natural Science Collection (ProQuest), Earth, Atmospheric, & Aquatic Science Collection (ProQuest), Engineering Source (EBSCO), Materials Science & Engineering Collection (ProQuest), Technology Collection (ProQuest)
- Anahtar Kelimeler: box and whisker plot, classification, CO2 electroreduction, correlation, machine learning, SHAP
- Gazi Üniversitesi Adresli: Evet
Özet
This work presents a data-driven assessment of CO2 electroreduction (CO2R) performance trends toward sustainable carbon utilization to produce chemical feedstocks and renewable fuels, using an expanded dataset of 662 observations from 47 studies published between 2003 and 2025. To enable both performance benchmarking and selectivity mapping, the highest Faradaic efficiency (FE) and the most selective product were chosen as target variables, while 13 variables (e.g., cathode catalyst elements, preparation method, and support type) were used as descriptors. First, exploratory data analysis was conducted to quantify how the reported highest FE values evolved over time. This analysis revealed a clear improvement in performance, with the median of the highest FEs increasing from 32.0% (2003–2013) to 47.8% (2022–2025). Furthermore, comparative analysis revealed shifts in system configurations, including increased use of chemical dealloying and Nafion 212 membranes. Subsequently, SHAP analysis quantified the direction and magnitude of feature effects on FE, while parallel coordinate plots highlighted multivariable patterns governing product selectivity (e.g., CO, HCOOH, and C2H4). Notably, key findings include a strong positive association between Sn atomic percentage and HCOOH FE, as well as the identification of preparation methods and electrolyte conditions linked to high C2H4 and CH4 selectivity. Overall, the proposed framework supports rational CO2 electrolysis design and performance-oriented optimization of integrated CO2R and renewable hydrogen systems.