Colloid rejection

Colloids are retained effectively by UF due to the small pore sizes of the membranes, compared to MF. However, if colloids are very small, then pore penetration can occur. Kim et al (1993) found a higher colloid rejection in stirred conditions using silver sol. Particle penetration into the membrane was highest at low salt concentrations. In the absence of salt, particle-membrane interactions dominated, whereas at high salt concentrations aggregation enhanced rejection. 

As shown in Table 5.5 (No 1, 2, 3), HA concentration was varied from 5 to 20 mgL" DOC, but the increase in concentration had only a marginal effect on colloid rejection. The aggregates were not redispersed due to the addition of organics, which stabilised the particles in the OPS case. The deposition of DOC (results not shown) on the membrane increases with concentration, but colloid deposition is constant at about 90%. Organic rejection decreases with increased concentration. 

Colloid rejection is almost complete for all sizes as shown in Table 5.3, and deposition is in the order of 90%. Flux decline increases with decrease in primary particle size (see Figure 5.4). This is most likely attributable to the formation of a cake of lower porosity with more interstitial organics. 

Experiments were carried out at 0.5, 2.5, and 4 mM CaCH. Results are shown in Figure 5.9B, and the trend observed is to that of the SPO data. Calcium can destabilise colloids that were stabilised by-organics. This was observed to occur at a concentration between 2.5 and 4 mM CaCl 2, with a resultant increase in colloid rejection from 15 to 95% and a greater flux decline, as shown in Table 5.7 (No 1, 4, 5). Tills corresponds to the effect of calcium on stabilised colloids reported by Amirbahman and Olson (1995), which was described in more detail in Chapter 2. Deposition increased with calcium concentration, which indicates that the destabilisation is always present to some extent. The calcium was added after the colloids were stabilised with HA, which is a different scenario to NOM, where salt (which is in the NOM powder) is added simultaneously. In this case, the calcium provides a full destabilisation of the organic-coated colloids at 4 mM, leading to complete rejection and deposition. 

Ferric chloride pretreatment leads to a higher WQP due to the higher organic and colloid rejection. While the value is stable in NF (although minor changes can occur as shown in Chapter 7), a correction of WQP due to the improved rejection is carried out for MF and UF. 

A total colloid rejection due to coagulation in MF increases the WQP from 57.7 to 111.5. With an average organic rejection of 30% and 63% at ferric chloride dosages of 25 and 100 mgL (calculated from Table 8.7), the WQP further increases to 130 and 163, respectively. 

The rejection of both systems, MF and RO, is summarised in Table A1.3. There is a large increase in DOC rejecbon during the MF operation, from 9.5 to 55.5%, and in colloid rejection reflected in Fe and Al. This increase in colloid rejection is probably due to retention of small colloids (smaller than the membrane pore size) as shown in Figure 5.1 and also characterised in Chapter 2. Feed concentradons during MF operadon are stable and no concentradon occurred as in RO. The RO rejecdon increased with feed concentradon to almost 100% for all parameters. 

Electroultrafiltration (EUF) combines forced-flow electrophoresis (see Electroseparations,electrophoresis) with ultrafiltration to control or eliminate the gel-polarization layer (45—47). Suspended colloidal particles have electrophoretic mobilities measured by a zeta potential (see Colloids Elotation). Most naturally occurring suspensoids (eg, clay, PVC latex, and biological systems), emulsions, and protein solutes are negatively charged. Placing an electric field across an ultrafiltration membrane faciUtates transport of retained species away from the membrane surface. Thus, the retention of partially rejected solutes can be dramatically improved (see Electrodialysis). 

Process Water Purification Boiler feed water is a major process apphcation of RO. Sealants and colloids are particularly well rejected by membranes, and TDS is reduced to a level that makes ion exchange or continuous deionization for the residual ions very economic. Even the extremely high quahty water required for nuclear power plants can be made from seawater. The iiltra-high quahty water required for production of electronic microcircuits is usually processed starting with two RO systems operating in series, followeci by many other steps. 

The RO system removes 90-95 % of the dissolved solids in the raw water, together with suspended matter (including colloidal and organic materials). The exact percent of product purity, product recovery and reject water depends on the amount of dissolved solids in the feedwater and the temperature at which the system operates. 

Major problems inherent in general applications of RO systems have to do with (1) the presence of particulate and colloidal matter in feed water, (2) precipitation of soluble salts, and (3) physical and chemical makeup of the feed water. All RO membranes can become clogged, some more readily than others. This problem is most severe for spiral-wound and hollow-fiber modules, especially when submicron and colloidal particles enter the unit (larger particulate matter can be easily removed by standard filtration methods). A similar problem is the occurrence of concentration-polarization, previously discussed for ED processes. Concentration-polarization is caused by an accumulation of solute on or near the membrane surface and results in lower flux and reduced salt rejection. 

As RO membranes become looser their salt rejection falls (see Section 31.8.1). Eventually a point is reached at which there is no rejection of salts, but the membrane still rejects particulates, colloids and very large molecules. The membrane pore size can be tailored to a nominal molecular weight cut-off. The resulting filtering process is called ultra-filtration. 

Generally, the effectiveness of the separation is determined not by the membrane itself, but rather by the formation of a secondary or dynamic membrane caused by interactions of the solutes and particles with the membrane. The buildup of a gel layer on the surface of an ultrafiltration membrane owing to rejection of macromolecules can provide the primary separation characteristics of the membrane. Similarly, with colloidal suspensions, pore blocking and bridging of... 

This mechanism is based on the known importance of hydroxides in other deposition reactions, such as the anomalous codeposition of ferrous metal alloys [38-39], Salvago and Cavallotti claim an analogy with the mechanism of Ni2 + reduction from colloids in support of their proposed mechanism. There is no direct evidence for the hydrolyzed species, however. Furthermore, the mechanism does not explain two experimentally observed facts Ni deposition will proceed if the Ni2 + and the reducing agent are in separate compartments of a cell [36, 37] and P is not deposited in the absence of Ni2 +. The chemical mechanism does not take adequate account of the role of the surface state in catalysis of the reaction. It has no doubt been the extreme oversimplification, by some, of the electrochemical mechanism that has led other investigators to reject it. 

The thermodynamic approach does not make explicit the effects of concentration at the membrane. A good deal of the analysis of concentration polarisation given for ultrafiltration also applies to reverse osmosis. The control of the boundary layer is just as important. The main effects of concentration polarisation in this case are, however, a reduced value of solvent permeation rate as a result of an increased osmotic pressure at the membrane surface given in equation 8.37, and a decrease in solute rejection given in equation 8.38. In many applications it is usual to pretreat feeds in order to remove colloidal material before reverse osmosis. The components which must then be retained by reverse osmosis have higher diffusion coefficients than those encountered in ultrafiltration. Hence, the polarisation modulus given in equation 8.14 is lower, and the concentration of solutes at the membrane seldom results in the formation of a gel. For the case of turbulent flow the Dittus-Boelter correlation may be used, as was the case for ultrafiltration giving a polarisation modulus of ... 

Reverse osmosis performs a separation without a phase change. Thus, the energy requirements are low. Typical energy consumption is 6 to 7 kWh/m2 of product water in seawater desalination. Reverse osmosis, of course, is not only used in desalination, but also for producing high-pressure boiler feedwater, bacteria-free water, and ultrapure water for rinsing electronic components—because of its properties for rejecting colloidal matter, particle and bacteria. 

As described above, the initial cause of membrane fouling is concentration polarization, which results in deposition of a layer of material on the membrane surface. The phenomenon of concentration polarization is described in detail in Chapter 4. In ultrafiltration, solvent and macromolecular or colloidal solutes are carried towards the membrane surface by the solution permeating the membrane. Solvent molecules permeate the membrane, but the larger solutes accumulate at the membrane surface. Because of their size, the rate at which the rejected solute molecules can diffuse from the membrane surface back to the bulk solution is relatively low. Thus their concentration at the membrane surface is typically 20-50 times higher than the feed solution concentration. These solutes become so concentrated at the membrane surface that a gel layer is formed and becomes a secondary barrier to flow through the membrane. The formation of this gel layer on the membrane surface is illustrated in Figure 6.6. The gel layer model was developed at the Amicon Corporation in the 1960s [8],... 

Because of the effect of the secondary layer on selectivity, ultrafiltration membranes are not commonly used to fractionate macromolecular mixtures. Most commercial ultrafiltration applications involve processes in which the membrane completely rejects all the dissolved macromolecular and colloidal material in the feed solution while completely passing water and dissolved microsolutes. Efficient fractionation by ultrafiltration is only possible if the species differ in molecular weight by a factor of 10 or more. 

As the liquid passes through the membrane in crossflow filtration, the particles, macromolecules, colloids, and so on, rejected by the membrane will accumulate in... 

In micro- and ultrafiltrations, the mode of separation is by sieving through line pores, where microfiltration membranes filter colloidal particles and bacteria from 0.1 to 10 mm, and ultrafiltration membranes filter dissolved macromolecules. Usually, a polymer membrane, for example, cellulose nitrate, polyacrilonytrile, polysulfone, polycarbonate, polyethylene, polypropylene, poly-tretrafhioroethylene, polyamide, and polyvinylchloride, permits the passage of specific constituents of a feed stream as a permeate flow through its pores, while other, usually larger components of the feed stream are rejected by the membrane from the permeate flow and incorporated in the retentate flow [10,148,149],... 

W.R. Bowen and A.O. Sharif, Hydrodynamic and colloidal interaction effects on the rejection of a particle larger than a pore in microfiltration and ultrafiltration membranes, Chem. Eng. Sci. 53 (1998) 879-890. 

Ultrafiltration (UF) refers to the removal of high molecular weight colloids (10,000 MW) up to particles less than 0.05 pm in diameter [10]. Like MF, UF places a mechanical barrier into the flow stream to separate the solid and liquid phases. The most common UF is a cross-flow hollow fiber type whereby a UF module contains hundreds of hollow microfibers. Whether or not the medium to be filtered flows inside or outside, the microfibers depends on the characteristics of the waste stream. UF is different from MF not only because UF can filter very small particles and some colloids, but also because of the cross-flow dynamics inside the UF module that keeps the surface of the hollow libers clean. Because of the cross flow, UF modules require a reject stream as well as a permeate stream. In other words, 100% of the liquid that enters the UF module does not exit as permeate. Figure 19.1 shows the differences between the UF and microlilters with respect to flow path. 

---