There is considerable interest in developing a nonfluorous capping ligand that is effective in dispersing nanoparticles in supercritical carbon dioxide. To augment the experimental search for such a molecule, a simulation protocol is developed, involving atomistic molecular dynamics simulations, which captures the relevant physical phenomena related to the effectiveness of capping ligands in stabilizing nanoparticle dispersions, evidenced by a consistency with empirical observation. The method is used to determine why the cheap and benign alkanethiol ligands are effective in supercritical ethane and several organic solvents but not in supercritical CO 2 at convenient conditions of temperature and pressure; n-dodecanethiol is used as the representative ligand. We conclude that the ineffectiveness of the alkanethiol ligands in CO 2 is primarily because they cannot compensate for the quadrupolar interactions which account for a substantial portion of the cohesive energy of bulk CO 2 , which is lost to the CO 2 molecules when transitioning from bulk solvent to the confined region between approaching nanoparticles, thus reducing the degree of osmotic repulsion. Ethane, which largely lacks significant electrostatic interactions, is adequately compensated for transitioning into the confinement by surface dispersion interactions. Hence, ligands that would be more effective in stabilizing nanoparticle dispersions in supercritical CO 2 should interact with CO 2 not only via dispersion interactions but also via its quadrupole. We find that the solvent does not penetrate the densely packed ligandpassivation layer at the nanoparticle surface, and solvent interacts only with the ligand tips. Hence, experiments intended to screen for effective ligands should probe ligand behavior not in an isotropic bulk system but in heterogeneous conditions similar to the ligand-solvent interface.