- 1 August 2007 -

Introduction to membranes: fouling control

As part of a series of articles looking at membrane filtration technology in water and wastewater treatment, Graeme Pearce looks at the key issue of fouling.

Background

This article discusses the key operational issue of fouling, and how to control it. Subsequent articles will examine the benefits of membrane filtration as pre-treatment to reverse osmosis, and the series will be concluded with a review of commercial products. The review will categorise products, and indicate the types of application for which different products are best employed.

The nature of fouling

Fouling in UF/MF is complex, with multiple interactions to consider between the various fouling constituents in the feed, and between these constituents and the membrane surface. Fouling can be characterised according to the nature of the constituent responsible, the mechanism by which it operates, or by the strategy adopted to control it.

Fouling constituents can be categorised as follows:

. Particulates
. Organic
. Inorganic
. Micro-biological organisms.

Particulates could be inorganic or organic and act as foulants due to their ability to blind or block the surface. The organic category covers dissolved components and colloids which would attach to the surface by adsorption. The inorganic category includes dissolved components which tend to precipitate onto the surface due to a pH change (scaling), or due to oxidation (e.g. iron or manganese oxides). Inorganics may also be present as coagulant residuals. The microbiological category covers vegetative matter such as algae, and organisms such as bacteria, which can form colonies and cause bio-fouling.

Fouling occurs due to a combination of chemical and physical interactions. Constituents in the feed can become attached to the membrane surface due to chemical binding and/or the interaction of surface properties, such as the degree of hydrophilicity, or charge effects. In addition, the fouling constituents will tend to physically blind the surface and block the pores, or hinder transport to the surface by the development of a cake layer. The combination of chemical and physical effects will control the degree of attachment. This will determine how severe the fouling is, and what strategies will be effective in controlling it.

Fouling mechanisms

UF and MF membranes used for water and wastewater applications tend to be rated at 0.01 to 0.1 µm, since at this rating, the majority of particles commonly found in the feed will be removed. Particles significantly above this size will be removed by pre-treatment, which might include clarification or flotation, and in some cases a media filter as well. If this type of pre-treatment is not used, a screen will be used to protect the membrane surface from larger particles, to avoid impact damage.

Particles smaller than 0.01 µm will be controlled by surface charge, and will tend to agglomerate, or bind to the membrane surface. This type of particle is often described as a colloid, and may be organic or inorganic in nature. Coagulation is a useful pre-treatment for this type of constituent since it will destabilise the colloids, and reduce their tendency to bind to the membrane. The most commonly used coagulants are inorganic salts with trivalent cations such as alum or ferric salts. However, it is important to note that divalent cations, particularly calcium, are also effective at destabilizing colloids to some degree, with the consequence that membranes are much less prone to organic fouling from colloids in high hardness/high alkalinity sources.

Particles foul the membrane by the following mechanisms:

. Pore blocking
. Cake formation
. Concentration polarisation.

The mechanisms of particulate fouling can be seen as a progressive development depending on the concentration of particles present in the feed, and the length of time before action is taken to mitigate their effect. Initially, particles will begin to deposit on the membrane surface, restricting the pore openings to varying degrees. The initial phase of particulate fouling in known as pore blocking, which can entail the plugging of a pore (complete), the constriction of the pore opening due to deposition of particles around the pore entry (standard), or a combination of the two as the build up of deposited particles begins to bridge the pore openings (intermediate).

The next stage of particulate fouling involves the development of a cake layer on the membrane surface, as additional particles continue to be deposited on the initial layer. As soon as the cake starts to form, the cake layer will control transport and removal, and effectively takes over the role of the membrane. The cake can have a beneficial effect in improving removal efficiency, and protecting the surface from adsorptive fouling, but often at the expense of permeability. The phenomenon therefore needs to be controlled, particularly in water applications, where solids concentrations are low, and the fine particulates can form a dense cake. Cake permeability can be affected by particle shape and size, particle deformability, and operating parameters such as transmembrane pressure.

The final stage of particulate fouling is that of concentration polarization, in which the particle concentration is allowed to build up in the feed channel to such an extent that transport to the membrane surface becomes limited. However, feed channel hydrodynamics are normally designed to minimize this effect.

Fouling control

Membrane filtration processes for water and wastewater treatment normally use dead end or directflow designs with intermittent backwash, sometimes combined with air scour either during the filtration and/or backwash cycle. The backwash controls the build up of fouling constituents by expelling particles from the membrane surface on a regular basis. This type of operation is designed to remove loosely attached foulants in a simple inexpensive physical process. Foulants which are not removed by backwash may require the addition of chemicals to improve the efficiency of removal. The various processes used are described below, and fall into three main categories:

. Prevention - Physical, e.g. backwash, air scour, or forward flush
. Maintenance - Chemical Wash (CW) or Chemical Enhanced Backwash (CEB)
. Recovery - Clean-In-Place (CIP)

The characteristics of the various processes are described below.

Backwash

. Regular intermittent process to address particle fouling, normally undertaken 1-4 times/hour
. Reverses the effect of pore plugging due to high velocities
. Controls the build up of particles at the membrane surface
. Reduces the particle concentration in the feed channel
. Reduces the effect of concentration polarisation
. Creates surface shear to dislodge surface attachment.

Air scour

. Used as part of a maintenance strategy between once/cycle to once/day (n.b. can be mechanically aggressive to the membrane fibre)
. Improves mass transfer and displacement action
. Effective for reversing pore plugging, particularly as TMP rises.

Forward flush

. Can be undertaken during the filtration cycle, or as part of the backwash routine (can be expensive in terms of reduced recovery)
. Improves shear
. Particle concentration build up effectively removed.

Chemical wash

. Used as part of a maintenance strategy on a periodic basis of between several times per day to once per week
. Alkali or chlorine soak to combat organic fouling
. Acid soak to combat inorganic fouling
. Biocide soak to combat bio-fouling.

Clean-in-place

. Used as part of a restoration strategy with heavy or tenacious fouling, normally undertaken between once per week to once in several months, often using the same chemicals as for chemical wash
. Extended soak and preferably recirculation, also sometimes with heating, to enhance effectiveness of chemicals.

Sustainable flux

The concept of critical flux has been used to describe the relationship between flux and fouling rate in controlled steady state environments The idea is that there is a critical flux below which no fouling occurs. Above this level, fouling occurs, the extent of which is a function of flux. However, in the application of UF/MF membranes in the water industry, directflow designs are used, and this is a pseudo steady state operation with different fouling characteristics to crossflow. In directflow, a degree of fouling occurs in the filtration cycle even at low fluxes, and this fouling may not be fully removed during the backwash cycle. Accordingly, it is necessary to develop different tools to understand, predict, and control membrane fouling. A practical tool for providing design guidelines for commercial plants is the concept of sustainable flux.

The sustainable flux is the flux at which a modest degree of fouling occurs, providing an acceptable compromise between capex (by using a high flux) and opex (by restricting the fouling rate). The value is dependent on feed characteristics, membrane characteristics, process design, and operational requirements (e.g. by use of an acceptable cleaning frequency). Pilot trials can be used to establish the relationship between flux and fouling rate for a particular set of circumstances, and evaluate a sustainable flux for a commercially competitive design and operation.

Fouling rates increase exponentially with flux, so the optimum flux is quite sharply defined for a given membrane and process design. The designer and end user need to define an acceptable cleaning frequency, and calculate the design flux from the fouling rate curve. Typical cleaning frequencies for pressure driven formats can be quite frequent, e.g. of the order of once/week to once /month, since the chemicals are contained within the module housings and downtimes are relatively low. Submerged formats have up to four times the chemical use per clean together with longer downtimes due to soak requirements and transfer time, so that optimum cleaning frequencies are typically once/month to once/quarter. Taking account of the fouling rate behaviour and the acceptable cleaning frequency, membrane permeability guidelines can be produced for the designer and operator for any system to provide a reliable control algorithm for stable long term performance.

Conclusions

Fouling is caused by a complex interaction between various constituents in the feed stream. It can be controlled by a combination of physical processes, such as backwash and air scour, and chemical processes such as chemical washes and clean in place.

The rate of fouling increases exponentially with flux. Commercial plants provide an optimal compromise between flux and fouling rate through the identification of the sustainable flux which provides a trade off between reduced capex at high flux, and reduced opex at low fouling rate. Membrane permeability can then be monitored to ensure stable long term performance. .

Contact:
Graeme Pearce at graemekpearce@btinternet.com
For references, contact l.nickels@elsevier.com