Concentration polarization

The phenomenon of concentration polarization is one of the most challenging problems within all membrane processes, whether they are driven by osmotic or external pressure.

Concentration polarization denotes the concentration gradients taking place at the mem-brane-solution interface because of selective transfer of species through the semi-permea-ble membrane. (Akther et al., 2015, p. 507.) Due to these gradients, the osmotic pressure difference between feed and draw solutions is much smaller at the membrane active layer than in the bulk (Cath et al., 2006, p. 73). Therefore, the actual water flux across the mem-brane falls, being significantly lower than the theoretical values (Qasim et al., 2015, p. 51).

The membranes used in osmotically driven membrane processes are typically asymmetric:

they consist of a porous support layer and a dense active layer. The concentration polari-zation phenomenon occurring within the support layer is referred to as internal concentra-tion polarizaconcentra-tion (ICP) and on the surface of the membrane active layer as external concen-tration polarization (ECP). It has been shown that the effect of ECP on decreased water flux is negligible compared to that of ICP. (Cath et al., 2006, p. 73.)

In FO applications, the active layer of the membrane faces the feed solution and the porous support layer faces the draw solution. In this orientation, concentrative ECP occurs as the retained solutes of the feed build up on the active layer. Because FO operates under no or low hydraulic pressure, the solutes do not tend to build up on the active layer, and ECP is very low. It can still be further reduced by increasing the flow velocity and turbulence at the membrane surface or by optimizing the water permeation rate. (Qasim et al., 2015, p. 51.)

Dilutive ICP takes place within the porous support layer as the draw solution is diluted by the permeating water (Cath et al., 2006, p. 73). As illustrated in Figure 13, ICP can signifi-cantly decrease the effective osmotic pressure difference across the membrane and result in up to 80 % decline in water permeation rate compared to the theoretical values (Akther et al., 2015, p. 508). Unlike ECP, ICP is more difficult to mitigate because it occurs within the support layer, and alteration of hydrodynamic conditions, such as flowrate of the draw solution, does not influence it (Zhao et al., 2012, p. 9).

Figure 13. Concentration profile across an asymmetric membrane with the active layer facing the feed solution in FO. C₁ and C₅ are the concentrations of the bulk feed and bulk draw solutions, respectively. C₂ and C₄ are the concentrations of the feed–active layer and draw solution–support layer interfaces, respectively. C₃ is the concentration at the active layer–support layer interface. Due to ICP, the effective osmotic pressure across the membrane (Δπ_eff) is much lower than the osmotic pressure difference between bulk feed and bulk draw solutions (Δπ_bulk). (Adapted from Cath et al., 2006, p. 74.)

Loeb, Titelman, Korngold, and Freiman (1997, p. 249) have estimated the water flux behav-ior (J_W) in the presence of dilutive ICP:

D W

F W

1ln Aπ B ,

J K Aπ B J

 + 

=  

 + +  (14)

where K is the solute resistivity to diffusion, A and B are the water and solute permeability coefficients, respectively, and π_D and π_F are the osmotic pressures of the draw and feed solutions, respectively. The solute permeability coefficient, the so-called B-value, is a measure of a membrane’s active layer. A low B-value is desirable as it indicates low solute flux across the membrane. It is determined by measuring the water flux and salt rejection (R) under various hydraulic pressures in RO mode and calculated as follows:



where C_P is the salt concentration in the permeate and C_F is the mean salt concentration in the feed. (Cath et al., 2006, p. 75; Phillip et al., 2010, p. 5172.)

The solute resistivity to diffusion K in Eq. (14) is related to the support layer properties and solute diffusivity by equation

S,

K =D (17)

where S is the structural parameter of the support layer and D is the solute diffusion coeffi-cient (McCutcheon & Elimelech, 2006, p. 240). The so-called S-value describes the struc-tural characteristics of a membrane and is expressed as

s , t τ

S= ε (18)

where t_s is the thickness, τ is the tortuosity, and ε is the porosity of the membrane support layer (McCutcheon & Elimelech, 2006, p. 240). The S-value is a widely used measure for evaluation of a membrane’s propensity to cause ICP (Manickam & McCutcheon, 2015, p.

70).

Equation (14) indicates that a smaller value of K leads to enhancement in water flux and reduced ICP. To attain that, the solute diffusivity D in Eq. (17) should be as high as possible, which can be achieved by increasing the filtration temperature or changing the draw solute (McCutcheon & Elimelech, 2006, p. 246). A small S-value is also preferred as it leads to reduced ICP effects. According to Eq. (18), the support layer should, therefore, be as thin and porous as possible to allow the draw solutes to diffuse more easily inside it. (Akther et al., 2015, pp. 509, 516.)

5.1.5 Fouling

Fouling means the accumulation of material on the surface or in the pores of a membrane weakening the performance of the membrane. It can occur by several mechanisms: pore blockage, deposition of particles on the membrane surface, and adsorption caused by in-teractions between the membrane and solutes or particles. For some molecules, a high level of concentration polarization may also result in gel formation on the membrane sur-face. (Field, 2010, pp. 1–2.) Fouling is a problem that concerns all membrane processes, but is less significant in osmotically driven processes, like FO, that operate under no or low hydraulic pressure. Membranes in such processes require less cleaning and maintenance, have a longer lifetime, and can be more productive over time. Additionally, fouling in FO membranes can be easily removed by backwashing, so there is no or less need for chemical cleaning. (Akther et al., 2015, p. 509.)