| Popis: |
Membrane microfiltration processes are used in for example the food, biotechnology, chemical and pharmaceutical industry, and more generally in e.g. wastewater treatment. Microfiltration is mostly used to separate components that are greatly different in size, e.g. micro-organisms from water, but rarely to fractionate components that are of similar size. This latter option would be interesting for many applications, since it would lead to enriched starting materials and possibly new products, but is hampered by accumulation of components in and on the membrane due to size exclusion by the pores. This leads to flux reduction and increased retention of components in time, basically the accumulated layer determines which components can pass the membrane (see Figure 1). Figure 1. Schematic representation of a cross-flow microfiltration process with a decrease in permeate flux over the length of the membrane due to pore blocking and particle adsorption in the pores and on the membrane walls. Most research focusses on accumulation mechanisms (concentration polarization, cake formation and adsorption) and concepts targeted at controlling particle accumulation. One example is back-pulsing, but this only gives a short term solution leading to extensive cleaning procedures given the way membranes are currently operated in practice. Clearly it would be beneficial if accumulation could be prevented, and through that, more stable operation could be achieved. This thesis presents how flow-induced particle migration can be used for stable membrane flux and retention of components in time. The particle migration mechanisms that are considered in this thesis, shear-induced diffusion, inertial lift, and fluid skimming, act on particles that are typically between 0.1 and 10 micron. They induce separation of components in the fluid moving (larger) particles away from the membrane, therewith facilitating separation; basically pore size no longer determines particle permeation. In the thesis it will be shown that these effects improve processing of dilute suspensions and make processing of highly concentrated systems possible, which is beyond the scope of current microfiltration processes. Before the design of these processes, methods to measure velocity and concentration profiles in microfluidic devices are described, compared and evaluated. The small dimensions of these devices will cause particles to migrate; as is used later in the thesis to facilitate segregation and separation. A drawback of the small dimensions is that they make measurement of velocity and concentration gradients difficult. Based on our evaluation, Nuclear Magnetic Resonance (NMR) and Confocal Scanning Laser Microscopy (CSLM), although expensive, are the most promising techniques to investigate flowing suspensions in microfluidic devices, where one may be preferred over the other depending on the size, concentration and nature of the suspension, the dimensions of the channel, and the information that has to be obtained. CSLM is used to study the behaviour of suspensions, between 9 and 38 volume%, at the particle level. Under Poiseuille flow in a closed microchannel, shear-induced diffusion causes migration in these suspensions. Under all measured process conditions, particles segregate on size within an entrance length of around 1000 times the channel height. Mostly, the larger particles migrate to the middle of the channel, while the small particles have high concentrations near the walls. This indicates that the small particles could be collected from their position close to the wall and that this principle can be applied to microfiltration (see Figure 2). Figure 2. Schematic representation of a cross-flow microfiltration process with a constant permeate flux over the length of the membrane due to shear-induced particle migration in combination with the use of a closed entrance length and large pores. Microfiltration of emulsions proves that the small particles can be removed without accumulation of particles in and on the membrane, as long as the process conditions are chosen appropriate. The membrane cross-flow module consists of a closed channel to allow particles to migrate due to shear-induced diffusion followed by a membrane with 20 micron pores, being much larger than the particles, where fractions of these emulsions can be removed. The emulsions consist of small droplets (~2.0 micron) and large droplets (~5.5 micron), with total concentrations between 10 and 47% and different ratios between small and large particles. As expected, the size of the emulsion droplets in the permeate is a function of trans-membrane pressure, membrane design and oil volume fraction (i.e., of the total, the small and the large particles). The guidelines for appropriate process conditions are described and application of the right process conditions leads to very high selectivity. This means that the permeate only consists of small droplets, and on top of that their concentration is higher than in the original emulsion. Especially at high droplet concentration (which is known to cause severe fouling in regular membrane filtration), these effects are occurring as a result of shear-induced diffusion. If only small particles are targeted in the permeate, the module can be operated at fluxes of 40 L/(h•m²); if fractionation is targeted the fluxes can be maintained considerably higher (2-10 fold higher). Separation of concentrated suspensions is currently done by dilution and since the process based on shear-induced diffusion works well at low velocities and high concentrations, industrial application could have major benefits in terms of energy and water use. An outlook is given on how current industrial processes can be designed and improved in terms of energy consumption by making use of particle migration. It is shown that return of investment of installation of these new membrane modules is short compared to the membrane life time, due to high energy savings. In order to reach this, it will be necessary to take unconventional process conditions that target particle migration and membrane designs as a starting point. Besides concentrated suspensions, also dilute suspensions benefit from particle migration. Migration phenomena can induce fractionation of yeast cells from water in dilute suspensions, using micro-engineered membranes having pores that are typically five times larger than the cells. The observed effects are similar to fluid skimming (in combination with inertial lift), and the separation performance can be linked to the ratio between cross-flow and trans-membrane flux, which is captured in a dimensionless number that can predict size of transmitted cells. For sufficiently high cross-flow velocity, the particles pass the pore and become part of the retentate; the separation factor can simply be changed by changing the ratio between cross-flow velocity and trans-membrane flux. Since the membranes have very large pores, fouling does not play a role and constant high trans-membrane flux values of 200–2200 L/(h•m2) are reached for trans-membrane pressures ranging from 0.02 to 0.4 bar. In conclusion, particle migration can improve (membrane) separation processes and even has the potential to lead to totally new separation processes. Particle migration can be advantageous in both dilute as well as concentrated systems, leading to reduced fouling, reduced energy and water consumption and a reduction in waste. This can all be achieved at production capacity similar or better than currently available in microfiltration processes. |
