Gibbs ensemble Monte Carlo
Phase separation is one of the topics for which simulation techniques have preferentially been focused in the recent past. Different procedures have been used for this purpose. Thus, for the particular case of chain systems, we can employ simulations in the semi-grand canonical ensemble, histogram re-weighting, or characterization of the spinodal curve from the study of computed collective scattering function. The Gibbs ensemble Monte Carlo method has been specifically designed to characterize phase transitions. It was mainly developed by Panagiotopoulos (Ref. 1) to avoid the problem of finite size interfacial effects. In this method, an NVT (or NpT)ensemble containing two (or more) species is divided into two (or more) boxes. In addition to the usual particle moves in each one of the boxes, the algorithm includes moves steps to change the volume and composition of the boxes at mechanical and chemical equilibrium. Transferring a chain molecule from a box to the other requires the use of an efficient method to insert chains. The configurational bias method, is specially recommended for this purpose.