Title:  Coupling between Membrane Processes and Crystallization Operations

Journal:  Ind. Eng. Chem. Res. 2010, 49, 5489?5495

Authors:  Catherine Charcosset,* Roland Kieffer, Denis Mangin, and Francois Puel
UniVersite de Lyon, F-69622 Lyon, France

This paper presents Reverse Osmosis (RO),  Membrane Distillation (MD), Membrane Contactors (MC), and Membrane Templates (MT) as reported membrane techniques for Membrane Crystallization (MCr).  The main features of MCr are:
- control of the supersaturation due to defined mass transfer across the membrane.
- membrane as heterogeneous nucleation-inducing substrate
- control of crystal size, shape, and purity
- reduced energy consumption compared to conventional technologies

1.) RO:  use of hollow fiber membranes.  Precipitation of calcium oxalate. Osmotic dewatering. crystallization of biomolecules. Solvent recovery from water. 6-fold lower energy consumption than in in evaporative crystallization. A main problem is concentration polarization, which reduces the mass flux across the membrane.
Advantages: better supersaturation control and larger crystal size than in conventional crystalizers. easy scale up and flexible control parameters. 6 to 10 times lower energy conversion than evaporative crystallizers.

2.) MD:  Only very small feed amounts have been investigated, a few tenths of a microliter to a few mililiters. Particle generation occurs mainly due to secondary nucleation, which is the destruction of already formed crystals in the unit due to mixing or in the pump.  Large crystals are necessary in biosciences to determine the crystal structure, which in turn is a way to determine the 3D chemical structure of the substance forming that crystal.  3Dstructures of molecules are vitally important in biotech. Shortened induction periods and higher growth rates compared to conventional technologies.  By controlling the flux, the supersaturation can accurately be adjusted and even crystal polymorphism (needle-like shape, prismatic shape...) can be controlled.
Advantages: MD-MCr produces especially large and well shaped crystals of high purity.  Production of specific polymorphs is possible.  Modularity.
Problem: high feed concentrations reduce the vapor pressure and thus reduce trans-membrane water flux.  At very high supersaturations crystals deposit on the membrane surface and a further but sharp flux decline is the result.

3.) MC: emulsification. mixing and reaction. precipitation. high mixing quality due to large number of pores, thus better than conventional mixer.  Antisolvent precipitation (which means that the membrane separates two liquids, one dissolves the solute and one doesn't. The antisolvent permeates through the membrane and causes the solute to crystallize).  Experiments using hollow fibre membranes showed fouling, however the fouling occured inside the fiber not on the membrane surface, so that use of larger fiber diameters significantly reduced fouling and flux decline.  Low chemical conversion (3.7 to 20.5%). 
Advantages: compared to conventional crystallizers the MC-MCr produced a higher number of nuclei/crystals per unit volume.

4.) MT: here the membrane has a defined shape and is used as a mold for crystallization to occur in.  The reactants are introduced into the mold and large, high quality single crystals are formed.
Advantage: single crystals with unique shapes.
Problem: Only one crystal per mold because the mold (membrane) is destroyed to release the crystal.

The advantages of membrane processes for crystallization operations are:
- heterogeneous nucleation:  The membrane surface fosters crystal nucleation
- supersaturation:  Due to the large number of pores local supersaturation peaks are low.
- reduced energy consumption:  between 6 to 10 times less energy than conventional technologies
- ease of scale-up:  Due to modularity.

Problems of membrane crystallization are:
- fouling and scaling: cleaning and recovering of the membrane necessary.

Contribution to the lab's work:
The paper clearly states that the MCr technology has not yet found industrial application and that it requires further optimization.  It also shows that research focus to date was more on crystal quality for applications such as protein crystallization and not, as would be desirable for the desalination industry, on salt/water procuctivity maximization.  It also lists the membrane processes that have been researched to date for their cpability of crystal formation.