Crystallization has exhibited great potential in downstream processing of the proteins. In order to provide the required amounts of crystalline proteins, efficient and reproducible methods for the crystallization of various kinds of proteins on a large scale are required. However, due to the complexity of the proteins, such methods are not found yet. Hence, new strategies have been sought in order to better understand and improve the crystallization processes in order to avoid large amounts of salts as crystallizing agent. For those purposes the "solvent freeze out technology" for crystallizing proteins is introduced here as an advanced crystallization technology which is efficient and reproducible. This new technique in protein crystallization involves the freeze out (a layer crystallization for the solvent, known from melt crystallization) of the solvent in order to enrich and therefore supersaturate the protein solution. The process is thus primarily a solid-layer melt crystallization technique with a secondary solution crystallization step generating the protein crystals. Problems and advantages of the technology are discussed on the basis of experimental results. As such it suffers from a problem common to solid-layer melt crystallization, namely inclusions of impurities in the crystalline solid layer. In the specific case presented here, the major impurity found in the ice layer is the protein. The dependence of the concentration of the protein found in the ice upon the final temperature of the cold finger as well as on the cooling rate applied has been investigated and is presented here. The question of the maximum allowable layer growth rate is discussed on the basis of the known criteria. The major parameters concerning the inclusion effect of solute in the ice are the ice growth rate (vice), mass transfer coefficient of the solute (k), the interface temperature (ti) and the interface concentration (ci) of the solute between the ice and the solution with the boundary layer conditions. In order to avoid inclusions in the ice, the allowable maximum ice growth rate (v max) for the process should be defined depending on the optimum criterion of v ice/k. The empirical evaluation of the estimated criterion of vice/k to the freeze out protein crystallization process is assessed on the basis of separation efficiency which is represented as effective distribution coefficient (k eff).