The quantum chemical treatment of molecular liquids is a challenging task, even by today's computer standards. In order to effectively observe bulk-like properties, a large number of molecules must be included in the calculation. However, conventional quantum chemical methods have a computational cost which scales at least with the fourth power of the system size. This limits the range of research to small clusters and/or low levels of accuracy.

In recent years, many-body approaches have been proposed, in which the energy of the system is decomposed into 1-,2-,..., N-body terms. Due to the weak interactions between the monomers, the series is known to converge rather quickly, with the 3rd and 4th order expansions already showing close to converged results. The topic of this talk is the attempt at further truncating the N-body expansion while retaining the level of accuracy, and designing a general linear scaling algorithm for use in the study of molecular clusters.