The role of long-range forces in the phase behavior of colloids and proteins

The role of long-range forces in the phase behavior of colloids and proteins. Abstract. – The phase behavior of colloid-polymer mixtures, and of solutions of globular proteins, is often interpreted in terms of a simple model of hard spheres with short-ranged attraction. While such a model yields a qualitative understanding of the generic phase diagrams of both colloids and proteins, it fails to capture one important difference: the model predicts fluid-fluid phase separation in the metastable regime below the freezing curve. Such demixing has been observed for globular proteins, but for colloids it appears to be preempted by the appearance of a gel. In this paper, we study the effect of additional long-range attractions on the phase behavior of spheres with short-ranged attraction. We find that such attractions can shift the (metastable) fluid-fluid critical point out of the gel region. As this metastable critical point may be important for crystal nucleation, our results suggest that long-ranged attractive forces may play an important role in the crystallization of globular proteins. However, in colloids, where refractive index matching is often used to switch off long-ranged dispersion forces, gelation is likely to inhibit phase separation. Introduction. – X-ray crystallography is still the standard technique to resolve the three-dimensional structure of globular proteins. But crystallography requires crystals, and protein solutions are notoriously difficult to crystallize. In order to understand the factors that favor crystallization, it is useful to gain insight into the phase behavior of the protein solution. As a first approximation, it is often sufficient to consider proteins as hard spherical bodies, interacting through a short-ranged attractive potential. In fact, Rosenbaum et al. [1] have shown that the crystallization curves for a number of globular protein solutions appear to coincide with those of a system of hard spheres with a rather short-ranged attractive Yukawa interaction, as studied in the simulations of Hagen and Frenkel [2]. These simulations were primarily aimed at modeling the phase behavior of polymer-colloid mixtures. As the attractive Yukawa model is used to model both colloids and globular proteins, one should expect that the conclusions that hold for one system should be transferable to the other. This does indeed appear to be correct, as far as the equilibrium phase behavior is concerned: experimental studies of colloid-polymer mixtures [3] show that, as the range of the attractive interaction between the colloids is shortened, the phase diagram changes in the way predicted …