Optimizing methanol synthesis from CO₂ using graphene-based heterogeneous photocatalyst under RSM and ANN-driven...

Area of Science:

• Catalysis
• Materials Science
• Chemical Engineering

Background:

• CO2 conversion to methanol is crucial for sustainable energy.
• Graphene-based catalysts enhance catalytic efficiency.
• Optimizing catalytic processes requires robust predictive models.

Purpose of the Study:

• To assess linear (RSM) and nonlinear (ANN) regression models for estimating in situ catalytic CO2 transformations.
• To optimize methanol yield using TiO2/Cu coupled with hydrogen exfoliation graphene (HEG).
• To compare the predictive performance of RSM and ANN models.

Main Methods:

• Experimentation and optimization of methanol yield using response surface methodology (RSM) and artificial neural network (ANN).
• Investigation of HEG loading (10-40 wt%) on TiO2/Cu catalyst performance.
• Analysis of influencing parameters like HEG dosing and CO2 inflow rate at pH 3.

Main Results:

• 30 wt% HEG loading on the TiO2/Cu catalyst yielded the highest methanol conversion efficiency.
• Optimal methanol yields predicted by RSM and ANN were 36.3 mg/g and 37.3 mg/g, respectively.
• The nonlinear regression-based ANN model demonstrated a superior determination coefficient (R² > 0.985) compared to RSM (R² ~ 0.97).

Conclusions:

• The ANN model, with 9 input neurons and 1 hidden layer, provided predictions closer to experimental outcomes.
• Both RSM and ANN models performed well, but ANN showed better accuracy in predicting optimal methanol yields.
• The study highlights the effectiveness of HEG-TiO2/Cu catalysts and the superiority of ANN for process optimization in CO2 transformations.