Predictive models for adsorption of organic compounds by Graphene nanosheets: comparison with carbon nanotubes

The Linear Solvation Energy Relationships (LSER) technique was applied in the present study for predicting models of organic compounds (OCs) adsorption by Graphene and Graphene oxide (GO), and the results were compared with those of multi-walled carbon nanotube (MWCNT) and single-walled carbon nanotube (SWCNT). Adsorption database of 38 OCs (28 aromatic and 10 aliphatic) for Graphene and 69 OCs (59 aromatic and 10 aliphatic) for GO were collected from the literature and our laboratory. The r² of the LSER models on the adsorption of aromatic OCs by Graphene and GO at three different equilibrium concentrations gradually increased up to OC molecular weight of 400 g/mol, after which a declining trend was observed for GO, while there was no visible change for Graphene. Among descriptors for all LSER models, V (molecular volume) and B (hydrogen bond accepting) for Graphene nanosheets (GNS) and carbon nanotubes (CNT) were the most significant descriptors (p values ≤ 0.05). B term had high value and was negatively correlated with adsorption of all OCs by Graphene (-1.24 to -9.45), GO (-0.55 to -9.31), SWCNT (-0.10 to -5.38) and MWCNT (-1.24 to -1.85). LSER successfully trained models for adsorption of OCs by GNS, and model coefficients were dependent on adsorbent type and OC properties.