Table 3 A brief explanation of the previous works for the application of AI methods in the discovery of Li-ISMs.
From: Materials discovery of ion-selective membranes using artificial intelligence
Heading | Type and model of computational chemistry calculations and AI | summery | The most important result | Ref |
|---|---|---|---|---|
Improving membranes design using advanced computational methods (MD, DFT, etc.) | MD and DFT | ∙ water absorption on AEM | ✓ structure of cationic groups has a significant effect on water absorption on AEM | |
DFT | ∙ modification of a layer of ion-imprinted polymer to the PVDF (Poly vinylidene fluoride) membrane with a molecular-scale design | ✓ molecular-level design with DFT can increase ion-ion selectivity in membrane construction | ||
DFT | ∙ diffusion mechanism of hydroxide ions and protons along the water wires | ✓ electronic structure has an important effect on the water wire conductivity in the classical nuclei simulations | ||
Improvement of DFT performance by utilizing ML | DL and ML | ∙ simulate large systems using data from DFT on small systems | ✓ Forces predicted by ML in molecular simulation can be calculated accurately by qualitative dynamic properties of materials | |
MD and AI | ∙ Examine diffusion mechanisms using common computational methods such as DFT calculation and MD and Eliminate their limitations by AI and ML. | ✓ AI can predict diffusion mechanisms completely automatically using existing datasets in these fields |
