Low-pressure gas adsorption analysis based on density functional theory (DFT) or non-local DFT (NLDFT) has become an increasingly reliable method for the characterization of nanopore structures in porous materials. The accuracy of the characterization of nanoporous structures in organic-rich shales can be improved by using a variety of probe molecules and models: carbon dioxide, nitrogen, and argon are considered using the CO2-DFT, N-2/Ar-DFT, N-2/Ar-NLDFT models and so on. Three types of shale sample with different maturity levels (i.e., mudstone, oil shale, and gas shale) are studied to investigate the effect of soluble components and maturity on the nanopore structure. The results show that the CO2-DFT and N-2/Ar-DFT models are more suitable than the other models based on DFT or NLDFT and that composite CO2-Ar adsorption analysis is the most accurate for assessing micro-, meso-, and macropores in the shales over the complete nanopore range (similar to 0.33-100 nm). The best method (i.e., DFT-based CO2-Ar analysis) is applied to characterize the pore structures of original and extracted shales, revealing that solvent extraction can increase the total pore volume (i.e., micro-, meso-, and macropore). However, the effect is not obvious in overmature gas shale. Micropores in the three shales display favorable pore size distributions of 0.4-0.7, 0.7-0.9, and 1.0-2.0 nm, respectively. (C) 2017 Elsevier Ltd. All rights reserved.