Properties of the cake layer formed during crossflow microfiltration (original) (raw)
Related papers
Cake growth mechanism in cross-flow microfiltration of mineral suspensions
Journal of Membrane Science, 1995
This paper describes the fouling mechanism in cross-flow filtration of a mineral CaCO3 suspension. Since the minimum particle size ( 1 /zm for a mean of 6/zm) was larger than the pore size (0.2/zm) of the ceramic membrane, fouling is in this case exclusively superficial and Brownian diffusion effects are negligible. The transmembrane pressure was increased and decreased in steps at constant velocity to form successive cycles. The permeate flux variation presented a hysteresis which is interpreted as an irreversible compression of the cake deposited on the membrane. Cake resistance was measured as a function of the pressure in the permeation experiments. The porosity was calculated from Ergun's equation and found to be a function of the pressure gradient in the cake. The mean cake thickness in the microfiltration experiments was calculated as a function of the pressure drop across the cake using the porosity data from the permeation experiments. The thickness was found to be a significant fraction of the membrane radius (up to 38%). This result suggested that the cake growth was limited by the shear stress acting on its surface. The cake grew as long as the top particle layer obeyed the Coulomb friction equilibrium. When due to an increase in shear stress by lumen reduction the Coulomb equilibrium was not satisfied, incident particles were ripped off the surface and the cake ceased to grow. However, subsequent pressure reduction did not decrease the cake thickness because of its irreversible compressibility.
Effect of particle size on the performance of cross-flow microfiltration
Advanced Powder Technology, 2006
The effects of particle size on the cake properties and the performance of cross-flow microfiltration are studied. A particulate sample with a wide size distribution range from submicron to micron is used in experiments. The probabilities of particle deposition are analyzed based on a force analysis. Since the major forces in determining the particle deposition and packing in the filter cake are different for submicron and micron particles, the particle size plays an important role in the filtration performance. Cake properties, such as mass, porosity and average specific filtration resistance of the cake, are calculated theoretically and compared with experimental data. Except for the overestimation of the mean particle size for about 1 μm, the calculated results of the pseudo-steady filtration rate and cake properties under various operating conditions agree fairly well with the experimental data.
Effective hydraulic resistance of the first cake layers at the membrane surface in microfiltration
Desalination, 2002
The usual procedure for predicting the filtration flux in microfiltration processes is to assume that the total hydraulic resistance to filtration flow can be calculated as the mere addition of the membrane resistance and the intrinsic cake layer resistance. It is shown that this assumption is acceptable when the cake layer is sufficiently thick but questionable when the cake layer is thin, i.e. at the beginning of the microfiltration process when the filtration flux decreases drastically. In fact, the simple additive model underestimates the total hydraulic resistance in the early stages of the cake formation. This problem is studied using two dimensional boundary element methods to solve the coupled problem of Darcy flow in the cake layer and Stokes flow in the upper fluid region. For a given cake layer thickness, the numerical results show that the effective resistance of the cake layer increases as the membrane surface porosity decreases.
Journal of Membrane Science, 2000
This paper investigates the effect of cyclic variations of transmembrane pressure (TMP), velocity and concentration on the particle cake formed on an organic membrane during the microfiltration of CaCO 3 suspension. The steady-state permeate flux was measured in a series of tests in which the TMP was varied between 0.25 and 3 bar, the velocity was changed between 0.25 and 2 m/s and the particle concentration was successively raised and diluted between 1 and 700 g/l. In these tests, the permeate flux was observed to exhibit an important hysteresis due to the irreversibility of the cake formed. When the change in operating conditions resulted in cake growth (TMP or concentration increase, velocity decrease), the permeate flux was found to be governed simultaneously by the resistance of the cake layer and concentration polarization with shear-induced diffusion. When the TMP or concentration were reduced, the cake thickness remained constant because of particle cohesion. When the velocity was increased, the cake thickness also remained constant until a critical velocity corresponding to the erosion shear stress for the cake was reached. Above this critical velocity, the cake was progressively eroded, starting from the membrane outlet, and completely removed if the velocity became high enough. This critical velocity decreased when the concentration rose because the suspension viscosity and density increased. The erosion shear stress was found to be a linear function of the normal stress on the cake surface. The main conclusion is that, when a cake has been formed, the permeate flux is governed not only by the present operating conditions but also by the previous ones.
Direct observation of particle deposition on the membrane surface during crossflow microfiltration
Journal of Membrane Science, 1998
In the cross¯ow micro®ltration of particles, a deposit cake layer tends to form on the membrane and this usually controls the performance of the ®ltration process. This paper presents observations of particle deposition on membrane surfaces using a non-invasive, in situ, continuous direct observation through the membrane (DOTM) technique. The particles used in the experiments were typical of micro®ltration processes, yeast (mean diameter 5 mm) and latex beads (3, 6.4 and 12 mm). The ®ltration tests were conducted in the imposed¯ux mode, so that the¯ux could be controlled at, below, or above the``critical ux''. Below the critical¯ux, the particle deposition was negligible; near the critical¯ux the particle deposition was signi®cant; and above the critical¯ux, particle layers were formed on the membrane surface. Rolling of the particles was observed during the ®ltration of 6.4 mm latex near the critical¯ux whereas a¯owing cake layer was observed during the ®ltration of 3 mm latex. The particle size distribution of the deposited particles changed with the cross¯ow velocity, with smaller particles deposited on the membrane at higher cross¯ow velocity. Comparison of the normalised¯ux (J/ÁP) with the membrane area coverage by the particles revealed that for ®ltration of latex particles``¯ux percentage (with respect to the clean membrane)'' was marginally greater than the percentage of uncovered membrane area, whereas for ®ltration of yeast, the``¯ux percentage'' was signi®cantly less than the uncovered area percentage due to the deposition of smaller cell debris species. This paper demonstrates that DOTM is a powerful technique for the study of fundamentals of particle deposition and interactions between the particles and the membrane.
A surface-renewal model for constant flux cross-flow microfiltration
A mathematical model using classical cake-filtration theory and the surface-renewal concept is formulated for describing constant flux, cross-flow microfiltration (CFMF). The model provides explicit analytical expressions for the transmembrane pressure drop (TMP) and cake-mass buildup on the membrane surface as a function of filtration time. The basic parameters of the model are the membrane resistance, specific cake resistance, and rate of surface renewal. The surface-renewal model has two forms: the complete model, which accounts for cake compressibility; and a subsidiary model for incompressible cakes, which can be derived from the complete model. The subsidiary model is correlated against some of the experimental TMP data reported by Miller et al. (J Membrane Sci 2014, 452, 171) for constant flux CFMF of a soybean-oil emulsion in a cross-flow filtration cell having unmodified and surface-modified, fouling-resistant membranes, and has an average root-mean-square (RMS) error of 6.2%. The complete model is fitted to the experimental TMP data reported by Ho and Zydney (J Membrane Sci, 2002, 209, 363) for constant flux microfiltration of a bovine serum albumin solution in a stirred cell using polycarbonate track-etched membranes and has an average RMS error of 11.5%. This model is also correlated against the TMP data of Kovalsky et al.
Journal of Membrane Science, 1997
The transition from concentration polarization to cake formation has been studied for the membrane filtration of colloidal Ž. silica by imposing flux and observing the system response. A critical flux J has been measured, below which crit transmembrane pressure drop, D P, is stable for increasing and decreasing flux. The flux-pressure profiles for operations Ž. Ž. below J show little for MF or negligible for UF hysteresis. Above J the pressure has a period of instability for crit crit increasing and decreasing flux, and there is significant hysteresis. It appears that once J is exceeded, the colloids in the crit polarized layer form a consolidated cake structure that is slow to depolarize and which reduces the flux. Evidence for cake deposition was obtained from electron micrographs. The depolarization can be increased by crossflow, by washing, and increasing pH. It was observed that the slow incrementation of flux to a given high value can result in significantly lower D P than the direct application of that flux. These differences are ascribed to formation of a stagnant, highly concentrated layer near the membrane surface due to consolidation and aggregation of solute resulting from very rapid flux increases.
Journal of Membrane Science, 2014
The concept of a critical permeation flux for the onset of particle deposition in crossflow microfiltration (CFMF) is well-established. However, the critical flux is known to be a function of process parameters such as the particle size, bulk concentration and crossflow velocity. The critical modified Peclet number (Pe crit ) is explored here as a generalized criterion for the onset of particle deposition that incorporates the effects of these process parameters as well as axial position along the membrane. Proper determination of Pe crit requires accurate prediction of the concentration polarization boundary layer thickness δ c and shear-induced diffusion coefficient D s . The classical Lévêque model is adapted to allow for the effect of the permeation flux on the velocity profile. Moreover, the assumptions of a constant concentration at the membrane surface c w and constant D s that have been made in prior studies are relaxed in an improved numerical solution to the convective diffusion equation that is used to predict δ c and D s . The critical permeation flux is determined from particle deposition data for 6 and 10 μm latex spheres taken via Direct Observation Through the Membrane (DOTM) characterization. A constant value of Pe crit = 4.00 0.08 is found to characterize the effects of particle diameter, bulk concentration and crossflow velocity as well as axial position on the onset of particle deposition.
Crossflow ultrafiltration of cake forming solutes: a non-steady state model
Desalination, 2005
The aim of this work was to check the validity of a non-empirical non-steady state model, which was initially developed to predict permeate flux for dead-end ultrafiltration, in the crossflow ultrafiltration of macromolecules. All the model parameters were theoretically calculated. Consequently, the model is capable of making predictions without the need of experimentation. The results estimated by the model were compared with those experimentally obtained in an ultrafiltration pilot plant. Tests were performed with Carbosep M2 ceramic membranes (Orelis, France) at different transmembrane pressures (0.1, 0.2, 0.3 and 0.4 MPa) and crossflow velocities (1, 2 and 3 m/s). Polyethylene glycol of 35,000 g/mol was used as a standard macromolecule in the fouling experiments. The validity of the model is restricted to the case when cake layer formation is the fouling mechanism. Model predictions are in agreement with experimental results for a crossflow velocity of 1 m/s, when the fouling mechanism is more likely to be cake formation. For other fouling mechanisms the predictions are not so good.