An adsorption isotherm identification method based on CNN-LSTM neural network.

Area of Science:

• Materials Science
• Chemical Engineering
• Computational Chemistry

Background:

• Adsorption isotherm morphology provides critical insights into adsorption mechanisms.
• Traditional methods for identifying adsorption isotherms are often inefficient and error-prone.
• Neural network approaches offer a promising solution for rapid and accurate identification.

Purpose of the Study:

• To develop and evaluate a hybrid convolutional neural network (CNN) and long short-term memory (LSTM) model for classifying adsorption isotherms.
• To create a comprehensive training database using theoretical adsorption equations, reducing experimental costs and time.
• To assess the model's performance using F1-score, ROC curves, and AUC metrics.

Main Methods:

• A hybrid CNN-LSTM model was implemented using Python 3.9, TensorFlow 2.11.0, and Keras 2.10.0.
• Theoretical adsorption isotherms were generated via adsorption equations to build training and validation datasets.
• Model performance was evaluated using F1-score, receiver operating characteristic (ROC) curves, and area under the ROC curve (AUC).

Main Results:

• The CNN-LSTM model achieved 100% accuracy on both training and validation sets.
• The model demonstrated a mean F1-score of 0.8885 on the testing set.
• Both macro-average and micro-average AUC values exceeded 0.95, indicating strong generalization ability.

Conclusions:

• The hybrid CNN-LSTM model provides a highly accurate and efficient method for adsorption isotherm identification.
• This AI-driven approach overcomes the limitations of traditional methods, enabling faster and more precise analysis.
• The developed methodology has significant potential for advancing research in adsorption science and related fields.