Evaluation of site-site bridge diagrams for molecular fluids

The presence of bridge functions in formally exact integral equation theories is the primary obstacle preventing the extraction of exact fluid structure from these theories. The bridge functions are typically neglected but in many fluids their impact may be significant. Each bridge function can be subdivided into bridge diagrams, which are well defined but difficult to evaluate. The calculation of bridge diagrams for the Chandler–Silbey–Ladanyi (CSL) integral equation theory [D. Chandler, R. Silbey, and B. Ladanyi, Mol. Phys. 46, 1335 (1982)] is the subject of this paper. In particular, we evaluate the diagrams required to yield an exact theory up to the first power in density [O(ρ1)] and provide algorithms that remain feasible for any molecule. Further, the bridge diagrams are evaluated and compared with the f-bond and h-bond formulations. Exact bridge diagrams are numerically evaluated for several chiral molecules, for two polar dimers, and for SPC/E water [W. L. Jorgensen, J. Chandrasekhar, J. D. Madura, R. W. Impey, and M. L. Klein, J. Chem. Phys. 79, 926 (1983)]. The quality of the diagrams is assessed in two ways: First, the predicted interatomic distributions are compared with those obtained from Monte Carlo simulations. Second, the connectivity constraints are evaluated and the errors in satisfying these exact relationships are compared for the f-bond and h-bond formulations. For apolar fluids, a clear improvement in CSL theory is evident with the inclusion of O(ρ0) and O(ρ1) diagrams. In contrast, for polar fluids, the inclusion of bridge diagrams does not lead to improvement in the structural predictions.