Crystallization of membrane proteins continues to be a field of study that proves to be more of an art than science. Membrane proteins are a fundamental component of many biological processes such as transport and ion regulation through a cell membrane. Membrane proteins also serve as a channel for signal transduction from one side of the cell membrane to the other and establish contact with the proteins of neighboring cells to form tissues such as muscles and organs. Proteomics has shown how the functionality of membrane proteins can be determined by their structure. Structural determinations are performed using various methods such as X-ray diffraction, electron microscopy, and solid state NMR. All methods are well developed and understood however; they all require a sample with some degree of crystallinity. Crystallization is the limiting step behind the proteomics of membrane proteins. Factors such as quantity, amphiphilicity, and non-rigid conformations increase the difficulty of determining crystallization conditions. The 1988 Nobel Prize in Chemistry was given in recognition for the 3D structural determination of the membrane protein bacteriorhodopsin from a crystal grown using detergent solublized protein. This is now referred to as the in-surfo crystallization method. However, the in-surfo method has serious limitations such as the choice of detergent for solubilization, denaturation due to the loss of the lateral membrane pressure, and applicability to various proteins. The in-meso crystallization method alleviates these problems by allowing the proteins to be crystallized from an artificial membrane-like environment. In this method it has been proposed that the bicontinuous region supplies protein for crystallization as nucleation and growth occur in a locally lamellar region formed from a precipitant-induced concentration change to form stacked 2D sheets. The rate of protein diffusion has been shown to be a function of the curvature of the lipid membrane. The formation of this artificial mesophase involves the mixing of non-Newtonian fluids having viscosities differing by ~30x. Traditionally this is done using a manual syringe or centrifuge-style mixer, requiring volumes of 20 µL of protein or more. Once the mesophase is formed high throughput matrix screening can be used to determine crystallization conditions in well-established well plate technology. These methods involve the use of large expensive fluid handling equipment, which can be on a scale larger than the amount of protein available. We report on the design and testing of a microfluidic platform designed for mixing viscous liquid to achieve in-meso membrane protein crystallization on-chip at volumes of >20 nL per sample. This platform allows for the preparation of multiple mesophase samples at once, while also facilitating high throughput screening of multiple precipitant solutions on a single chip. In addition to screening for appropriate crystallization conditions, independent control over precipitant injection has the potential to allow for kinetic control of crystal growth done in-meso through the use of additives to control the rate of protein diffusion by affecting the curvature of the protein-containing mesophases. The membrane-bound proton pump bacteriorhodopsin is used as a model system.