Machine learning-assisted design of thermosensitive highly efficient adsorption agents modified with cucurbituril for lithium extraction from Salt Lakes

Escalating worldwide lithium shortages have greatly promoted the development of lithium recovery techniques from salt lake brines. Given the prominent technical drawbacks of traditional adsorbents during brine lithium enrichment, machine learning was adopted to rapidly synthesize a magnetically separable high-efficiency adsorbent functionalized with cucurbit[7]uril (CB[7]) Cucurbit[7]uril-based magnetic biochar (CBMBC) has been successfully prepared. Under optimal adsorption conditions, CBMBC exhibited a lithium adsorption capacity as high as 48.87 mg/g, and in Lagute Lake brine applications it showed an adsorption performance 11.3 times higher than that of commercially available carbon-based lithium adsorbents. In addition, CBMBC possesses excellent photothermal performance; COMSOL simulation confirmed that CBMBC forms pronounced temperature and humidity gradients under solar irradiation and enhances air flow, thereby increasing the diffusion rate of adsorbate molecules migrating from the aqueous solution to the adsorbent surface, and the lithium adsorption capacity of CBMBC under one-sun irradiation increased by 36.7 %. Significantly, integrated analyses combining in-situ Raman physical characterization and theoretical simulations verify that the CBMBC material delivers outstanding lithium uptake capacity via diverse intermolecular forces. These interactions stem from nitrogen/oxygen sites at CB[7] cavities, surface amino moieties of carbon sheets, as well as carbon framework structures. Collectively, this work establishes a promising route toward industrially feasible lithium recovery from salt lake brines.