Why increasing flux is much more challenging than reducing membrane fouling?

Sustainable flux has hardly increased in last decades

   Membrane technology has been rapidly improved in recent decades. But, the improvements are mainly about lower fouling, lower energy consumption, easier installation and operation, more reliable membranes and modules, etc. rather than higher flux. 

  • In RO membranes for surface water treatment, the typical design pressure of 15 bar (or ~225 psi) has been reduced to 10 bar (or ~150 psi) with low pressure RO and again to 7 bar (or ~100 psi) with ultra-low pressure RO. These low operating pressure was achieved by improving membrane permeability without loosing too much of rejection efficiency. But, operating flux remains nearly identical at 15-25 LMH (or 9-15 gfd) in typical applications.
  • Immersed membranes have greatly evolved during the last two decades. The sustainable flux of the hollow fiber membranes with small inner diameter (<0.3 mm) was 10-15LMH in early 1990s. But, the sustainable flux has increased to 20-30 LMH after dual layer hollow fibers with bigger inner diameter (~0.9 mm) were developed in late 1990s. It is believed that the lower internal pressure drop is the major contributor to the increased flux as discussed here. With little changes in hollow fiber dimensions and operating methods since  then, little flux increase has been realized despite of the large improvement in membrane permeability. Meanwhile, reliability of immersed membranes has been improved dramatically due to the improvement in module/cassette structures, aeration methods, optimized biological process design that allows less membrane fouling, etc
  • Tubular membranes has been used for decades with some improvement in membrane materials, dimensions, design strategies, etc. However, operating pressure and flux have hardly improved.

Flux is self-limiting and dominated by hydrodynamics on membrane surface

   As discussed here, membrane resistance is a minor portion of total resistance in a typical MF/UF processes regardless of the membrane configuration. Therefore, it is obvious that reducing membrane resistance (or improving membrane permeability) does not give a significant edge to the filtration process in general when it comes to MF/UF.

   Fig. 1 illustrates the membrane filtration process, where the cake layer formed on membrane surface acts as a filtration barrier while concentration polarization layer is in dynamic equilibrium with the cake layer. If liquid velocity increases, shear rate increases on the cake surface and thinner cake layer can be formed. The high liquid velocity can also reduce the thickness of the concentration polarization layer. It is noteworthy that the membrane property is not a factor affecting the cake layer growth.

   If flux increases whatever the reasons are, cake layer can build up quicker under a given condition due to the increase of drag force toward the membrane surface. Moreover, as discussed here, cake layer compaction expedites especially in the bottom the layer contacting the membrane surface. As a result of the faster cake layer growth and the subsequent cake layer compaction, flux decreases quickly near to the original level.

   The extreme case of the self-limiting nature of flux in MF/UF process is observed in the mass transfer controlled region as discussed with Fig. 1 here, which is commonly observed in tubular membrane process. Tubular membranes typically run under mass-transfer controlled region and increasing TMP results higher flux only for short term at best.

FluxCo1

Fig. 1. Contributing factors for membrane fouling.

 

Hydrodynamics on membrane surface is largely unchangeable for a given membrane configuration

 

   The hydrodynamic conditions on membrane surface are mainly affected by membrane module configuration and operating condition. For a given membrane module configuration, operating conditions can be modified by changing crossflow velocity, aeration rate, suspended solids in water, etc., but there are not much flexibility in changing those due to the following reasons.

  • Higher air flow rate for immersed membranes certainly increases sustainable flux to some degree, but it comes with higher operating costs. Air flow rate must be optimized considering the flux, cost of aeration, cost of membranes, etc. The current design fluxes and aeration rates are the outcome of such optimization.
  • Increasing crossflow velocity in tubular membrane not only causes higher operating costs, but also causes higher pressure drop along the tubes. As discussed here, high crossflow velocity causes high pressure loss along the tube and this directly means high pressure in tube inlet. Due to the pressure limit of the tube, either tube length or crossflow velocity must be compromised.
  • Membrane flux may be able to be raised by reducing MLSS to some extent at least in theory, but larger aeration tank is required to keep the F/M ratio in an optimum range. However, the relation between MLSS and sustainable flux is very fuzzy and there is no simple correlation exists.

   Therefore, obtaining higher flux for a given membrane configuration is not readily possible due to the process optimization issues along with cost issues. If higher sustainable fluxes are desired, fundamentally different membrane module configurations should be devised, which allows better hydrodynamics on membrane surface.

Surface modifications of membrane may result in prolonged filtration cycle, but no high flux

   Membrane surface properties are important factors affecting initial membrane fouling as discussed here. By making membrane surface more hydrophilic and smooth, initial membrane fouling can be slowed down to some extent. As a result, membrane cleaning interval can be prolonged at a same flux. However, due to the following reasons, the improved membrane surface properties do not necessarily mean higher operating flux.

  • Once initial cake layers are formed, membrane surface properties no longer are the major factors affecting subsequent cake layer formation. Therefore, membrane surface properties affect only the initial membrane fouling.
  • If flux is increased, more foulants deposit on membrane surface and the initial membrane fouling completes quicker. The subsequent fouling process is dominated by the hydrodynamic conditions under the filtration condition.

   As a result, little flux increase has been realized in last decades for a given membrane module configuration in spite of the enormous efforts put on membrane surface modifications.

Analogy with the speed of passenger jet

   It appears that the time required to fly from one continent to other has been marginally reduced over the last 3-4 decades. Meanwhile the only supersonic passenger jet, Concorde, stopped its service in 2003. Why there has been only little improvement in the passenger jet speed despite of remarkable progress in turbine engine? Unlike the cars on the ground of which speed may be limited by how quickly we can respond to the sudden accident, there seems no barriers that limit plane speed. The simple answer might be “it is not economical to raise the speed due to the aerodynamic limitation around the wings”. Since drag force is proportionally to the square of plane speed, less fuel efficiency is inevitable at high speed. Under the current economic situation, aircraft industry is focused on improving fuel economy rather than improving speed as far as passenger jets are concerned.

   Likewise, MBR technology has been evolved to obtain a high energy efficiency by modifying module design, aeration method, suction method, etc. except in the relatively short period of initial technology development period. It is quite obvious that fundamentally different hydrodynamic conditions are required to increase sustainable flux beyond 20-30 LMH, which can hardly be achieved with the currently available immersed membranes. Unless the entire module structure and the way upflow is generated are changed, it is difficult to exceed the current flux ranges.

   Interestingly there is one membrane called AirliftTM (Norit division of Pentair), where circulation pumps are used to recycle mixed liquor through vertically mounted tubular membranes while bubbling through the tubes. Due to the higher upflow velocity and better defined hydrodynamics, steadily achievable flux is known to be > 40 LMH, which is much higher than that of immersed hollow fiber membranes. Norit claims the overall energy efficiencies are as good as the state of the art immersed hollow fiber membranes despite of the different module configuration and the high flux. It will be interesting to see how these claims turn out while the similar was not the case in passenger jet industry.

 

© Seong Hoon Yoon