TY - JOUR
T1 - Machine learning-aided prediction and optimization of specific capacitance in functionalized GO-based Mn/Fe co-doped Bi2O3 nanocomposites
AU - Mane, Vijay A.
AU - Chavan, Kartik M.
AU - Munde, Sushant S.
AU - Dake, Dnyaneshwar V.
AU - Raskar, Nita D.
AU - Sonpir, Ramprasad B.
AU - Dhole, Pravin V.
AU - Gattu, Ketan P.
AU - Somvanshi, Sandeep B.
AU - Kayande, Pavan R.
AU - Pawar, Jagruti S.
AU - Dole, Babasaheb N.
N1 - Publisher Copyright:
© 2025 Elsevier Ltd
PY - 2026/2/10
Y1 - 2026/2/10
N2 - In this research, tungsten-decorated graphene oxide (GO)-based 5 % Mn/Fe co-doped Bi2O3 (MFBGOT) was successfully synthesized using a hydrothermal technique. These nanocomposites were thoroughly examined for their structural, morphological, surface, and electrochemical characteristics. X-ray diffraction (XRD) analysis confirmed a cubic crystalline structure with a significantly reduced crystallite size (∼28.63 nm), suggesting efficient integration of dopant elements. Field emission scanning electron microscopy (FESEM) displayed a nanoflower-sheet-like morphology, which increased surface roughness and active site exposure. According to Brunauer–Emmett–Teller (BET) analysis, the sample exhibited a high specific surface area of 85.23 m2/g and mesoporous nature, with an average pore size of 15.78 Å, evidenced by Type IV isotherms and BJH pore distribution. Electrochemical testing revealed an impressive specific capacitance of 1037 F g−1 at a current density of 0.5 A/g of MFBGOT sample, over twice that of the PBGOT sample (420 F g−1), along with excellent cycling stability, retaining 77 % of its capacitance after 2000 cycles. This outstanding performance is attributed to the combined effects of dual-metal doping, improved wettability, and efficient ion transport. X-ray photoelectron spectroscopy (XPS) validated the successful co-doping and the presence of mixed oxidation states (Mn2+/3+ and Fe3+), which enhance redox reactions. Furthermore, machine learning (ML) algorithms, including Random Forest, Gradient Boosting, Support Vector Regression (SVR), ElasticNet, and Ridge Regression, were applied to predict specific capacitance based on a dataset comprising 141 different materials. Among them, the Gradient Boosting model showed the best performance with a high test R2 value of 0.9202, highlighting the significant influence of the electrode material and electrolyte type. This research presents a promising approach that combines advanced material engineering with machine learning to optimize energy storage solutions.
AB - In this research, tungsten-decorated graphene oxide (GO)-based 5 % Mn/Fe co-doped Bi2O3 (MFBGOT) was successfully synthesized using a hydrothermal technique. These nanocomposites were thoroughly examined for their structural, morphological, surface, and electrochemical characteristics. X-ray diffraction (XRD) analysis confirmed a cubic crystalline structure with a significantly reduced crystallite size (∼28.63 nm), suggesting efficient integration of dopant elements. Field emission scanning electron microscopy (FESEM) displayed a nanoflower-sheet-like morphology, which increased surface roughness and active site exposure. According to Brunauer–Emmett–Teller (BET) analysis, the sample exhibited a high specific surface area of 85.23 m2/g and mesoporous nature, with an average pore size of 15.78 Å, evidenced by Type IV isotherms and BJH pore distribution. Electrochemical testing revealed an impressive specific capacitance of 1037 F g−1 at a current density of 0.5 A/g of MFBGOT sample, over twice that of the PBGOT sample (420 F g−1), along with excellent cycling stability, retaining 77 % of its capacitance after 2000 cycles. This outstanding performance is attributed to the combined effects of dual-metal doping, improved wettability, and efficient ion transport. X-ray photoelectron spectroscopy (XPS) validated the successful co-doping and the presence of mixed oxidation states (Mn2+/3+ and Fe3+), which enhance redox reactions. Furthermore, machine learning (ML) algorithms, including Random Forest, Gradient Boosting, Support Vector Regression (SVR), ElasticNet, and Ridge Regression, were applied to predict specific capacitance based on a dataset comprising 141 different materials. Among them, the Gradient Boosting model showed the best performance with a high test R2 value of 0.9202, highlighting the significant influence of the electrode material and electrolyte type. This research presents a promising approach that combines advanced material engineering with machine learning to optimize energy storage solutions.
KW - Bismuth oxide
KW - Electrochemical performance
KW - Graphene oxide
KW - Machine learning prediction
UR - https://www.scopus.com/pages/publications/105027441307
U2 - 10.1016/j.est.2025.120114
DO - 10.1016/j.est.2025.120114
M3 - Article
AN - SCOPUS:105027441307
SN - 2352-152X
VL - 146
JO - Journal of Energy Storage
JF - Journal of Energy Storage
M1 - 120114
ER -