I want to simulate water diffusion in polymer matrix using Gromacs. I would like to know about system size effects. Then, I chose system with 10, 20 and 30 chains. In all systems, water content is invariant regarding to polymer chain ( 80 water molecule for each chain in all systems). When I did water cluster analysis(gmx clustsize), number of clusters was the same ( 3 for each system) and with increasing system size(number of chain and water with a constant ratio) I just saw an increase in each cluster size ( 3 cluster in each system but with different sizes). Is it a true conclusion that with increase in system size and hence number of molucules we must expect for increase in cluster size not cluster numbers? is Cluster number fixed and could we correlate it to macroscopic sample?