A truly complex fluid: particles with random interactions

We use computer simulations to study multi-component systems in which all the particles are different (APD). The particles are assumed to interact via Lennard-Jones potentials with identical size parameters, but with pair interaction parameters generated at random from some distribution. We analyze these systems at temperatures above and below the freezing transition and find that APD fluids relax into a non-random state characterized by clustering of particles according to the values of their pair interaction parameters (Neighborhood Identity Ordering - NIO). We study the NIO using the random bond lattice model and show that the transition from frozen to annealed disorder depends not only on temperature but also on system size. We use a variant of the APD model to study the competition between specific and non-specific interactions and show that contrary to intuitive expectations the latter can assist in the formation of specific complexes. The relevance of our results to biological systems is discussed.