You are on page 1of 25

Journal of Membrane Science 303 (2007) 428

Review article

Fouling strategies and the cleaning system of NF membranes and factors affecting cleaning efciency
Ahmed Al-Amoudi a,b , Robert W. Lovitt a,
Centre for Complex Fluids Processing, Multidisciplinary Nanotechnology Centre, School of Engineering, University of Wales, Swansea SA2 8PP, UK b Saline Water Conversion Corporation (SWCC), Saline Water Desalination Research Institute, Saudi Arabia Received 17 January 2007; received in revised form 25 May 2007; accepted 6 June 2007 Available online 14 June 2007
a

Abstract Nanoltration membranes play an important role in the desalination of brackish and seawater as well as membrane mediated waste water reclamation and other industrial separations. Fouling of nanoltration (NF) membranes is typically caused by inorganic and organic materials present in water that adhere to the surface and pores of the membrane and results in deterioration of performance (reduced membrane ux) with a consequent increase in costs of energy and membrane replacement. Natural organic matter (NOM) fouling of NF membranes involves interrelationship between physical and chemical interactions and is described in this review. Inorganic fouling due to scale formation of sparingly soluble inorganic salts occurs whenever the ionic salt concentration stream exceeds the equilibrium solubility. Scale formation takes place by homogenous or heterogeneous crystallization mechanisms. Biolm formation also becomes an issue when its thickness and surface coverage reduces permeability. There are two strategies that are usually employed to minimize the effect of fouling. The rst group includes minimizing of fouling by using adequate feed pretreatment, membrane treatment and membrane modication. The second group involves membrane remediation by chemical cleaning which is carried out to restore membrane uxes. A large number of chemical cleaning agents are commercially available, and the commonly used ones fall into six categories: alkalis, acids, metal chelating agents, surfactants, oxidation agents and enzymes. In general, these cleaning agents do improve the membrane ux to certain extent. Combination of these chemical agents has also been tried in order to improve the ux restoration. Even though, many of these cleaning agents can restore the ux over 100% (enhanced ux), they can also impair the selectivity of the membrane reducing of the product water quality. There are many traditional assessment methods for cleaning and at present these are being supplemented by methods using modern surface analysis techniques. These are being now rapidly developed to give a more precise assessment and a better understanding of cleaning processes. Generally, cleaning is assessed by ux, zeta potential measurement, atomic force microscope (AFM) and Fourier transforms infrared technique (FTIR). Atomic force microscope and related techniques are particularly employed in order to evaluate the cleaning efciency and other surface phenomena. There are several factors that can affect the chemical cleaning process which include temperature, pH, concentration of the cleaning chemicals, contact time between the chemical solution and the membrane and the operation conditions such as cross-ow velocity and pressure. The role of temperature and pH in cleaning are membrane dependent. These factors play very important role in ux recovery. A critical review of these factors is also presented. It appears from the literature that only very few papers on cleaning of NF membrane to regenerate membrane performance have been published up to date, and there is an urgent need for extensive research work to investigate fouling mechanisms in order to obtain fundamental understanding of fouling to provide more feasible, cost-effective cleaning and performance restoration procedures. This also provides further strategies for the avoidance of fouling through better pretreatment and more appropriate membrane fabrication and modication. 2007 Elsevier B.V. All rights reserved.
Keywords: Cleaning agents; Nanoltration membrane; Cleaning efciency; Fouling

Corresponding author. Tel.: +44 1792 295709. E-mail addresses: Aalmoudi@swcc.gov.sa, 310981@swan.ac.uk (A. Al-Amoudi), r.w.lovitt@swansea.ac.uk (R.W. Lovitt). 0376-7388/$ see front matter 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.memsci.2007.06.002

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

Contents
1. 2. 3. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fundamentals of NF separation and selectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fouling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.1. Inorganic fouling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.2. Organic fouling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.3. Biofouling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Operational aspect of NF and fouling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Primary location for specic types of fouling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Fouling minimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.1. Coagulation followed by ltrationsedimentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.2. Scale inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.3. Membrane preltration and membrane modication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.4. Sonication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Membrane cleaning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Remediation of the membrane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. General considerations and costs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. Assessment of cleaning agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.1. Type of cleaning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.2. Cleaning mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.3. The impact of cleaning on NF permeate quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Methods of assessing the cleaning effectiveness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.1. Flux measurement (non-destructive) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.2. Atomic force microscopy (AFM) (destructive method) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.3. Fourier transform infrared technique (FTIR) (destructive method) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.4. Zeta potential measurements (non-destructive) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5. Factors affecting chemical cleaning efciency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.1. Effect of cleaning solution pH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.2. Effect of ionic strength of the cleaning solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.3. Effect of cleaning solution concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.4. Effect of cross-ow velocity (hydrodynamic shear) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.5. Cleaning duration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.6. Cleaning frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.7. Effect of temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.8. Effect of pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion and conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 6 6 6 8 10 13 13 13 14 14 14 14 14 15 15 15 16 16 18 18 19 20 20 21 21 22 22 22 22 22 23 24 24 24 24 25

4.

5.

1. Introduction Water scarcity is a major political and economic problem in the many parts of the world especially in the arid regions such as the Middle East, Southern Europe, North and mid Africa, Australia and many states of America such as California, Florida, New Mexico, etc. The shortage in natural fresh water supply for domestic purposes is most acute for the Arabian Peninsula countries Saudi Arabia, Kuwait, Bahrain, Qatar, United Arab Emirates, Oman and Yemen where demand for water increases annually at a rate of 3% or more [1]. In addition, the rapid further reduction of subterranean aquifers, and the increasing salinity of these non-renewable sources will continue to exacerbate the water shortage problems in many areas of the world. Desalination techniques are capable of providing the solution [2]. Desalination or Desalinization refers to water treatment processes that remove salts from saline water. Desalination has already become an acceptable solu-

tion for shortages in conventional water resources and has acknowledged as sustainable and effective process by reputable institutions such as the World Bank [3]. This can be achieved either by thermal processes involving evaporation or by membrane ltration involving separation of ions from water. As water demand increases, environment and safety regulations are becoming more stringent, greater research efforts have been put into the improvement of membrane processes. During the past decade a variety of water treatment membranes has been developed [4]. These membranes have been vastly improved in the area of water ux, salt rejection, and especially in their ability to maintain high performance levels at substantially lower operating pressures than their predecessors [5]. Despite these improvements, a decline in membrane performance over a period resulting from membrane fouling that leads to a decrease in water ux across the membrane and increased salt passage through the membrane [6,7].

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

The membrane can be considered the heart of a desalination plant where the cost of membrane unit is about 2025% of the total capital cost [8]. Consequently, it is very important to be familiar with factors involved in reduction of membrane performance and longevity, in particular membrane fouling. The factors affecting NF separation that can also play an important role in membrane fouling and cleaning are as follows: Membrane properties such as surface roughness pore size distribution, membrane thickness, membrane charge type and charge density. The chemistry of the treated solution such as solute composition, the size, geometry and the charge of the components, the concentration of ions, the pH and the fouling potential of the solution and its interaction with membranes. The operation design of the NF systems, theircapacity, dimensions and ow. The processes environmenttemperature and pressure. This review focuses on the type of foulants that leads to ux reduction of fresh water obtained from desalination plant: explanations of desalination processed and their mechanism of fouling are also reviewed to give clear understanding of the cleaning processes. The main objective of this review is to address the key factors in maintaining and restoring the plant performance and the factors that affect the cleaning efciency. 2. Fundamentals of NF separation and selectivity Nanoltration (NF) membranes are mainly utilized for softening brackish waters. The separation characteristics of nanoltration (NF) stand between ultraltration (UF) and reverse osmosis (RO) and the membrane selectivity has often been attributed to the interchange of both molecular sieving mechanisms characteristic of ultraltration and diffusion mechanisms characteristics of RO. NF membranes are usually made of polyamide based Thin Film Composites (TFC), which are relatively close to RO membranes in chemical structure. However, a key distinguishing feature of RO membranes is their higher rejection of both monovalent and divalent ions, the NF membranes are typically characterized by lower rejection of monovalent ions, but maintaining higher rejection of divalent ions and higher ux than that of RO membranes. In general NF membranes have relatively high charge and also pores in the order of about 1 nm [9]. Consequently both, charge effects and sieving mechanisms inuence the rejection behavior of solutes in NF membranes [10]. Generally, the basic chemical structure of the synthetic polymers used in the preparation of RO, UF and NF membranes are almost same apart from the pore size of the membranes. Therefore overlap of properties of RO and UF with NF in terms of both transport phenomena and consequent fouling is common in the area of water treatment. Consequently the cleaning processes for the RO and UF are also similar for NF membrane. Hence, this literature review sometimes considers UF and RO

membranes and their fouling or/and cleaning system in order to elaborate the points where lack of information on NF fouling and cleaning exists. The forces of the interaction between the membrane surface and particles in solution are important in understanding the fouling phenomena. The normal basis for quantifying particlesurface interaction is DLVO theory where the particlesurface interactions in aqueous environments could be predicted by the summation of van der Waals and electrostatic double layer forces. The Fig. 1 is a schematic description of the DLVO interaction proles and the summation of these two forces. There are several important features about this diagram: Unlike the double layer interaction, the van der Waals interaction potential is largely insensitive to variations in pH and electrolyte concentration. Reducing the interaction between the particles and the membrane as much possible can reduce the fouling phenomena. This can be achieved when the critical values (ux and pressure) arise as a balance between the hydrodynamic force driving solute towards the pore and the electrostatic forces opposing this motion. Critical ux stems from the concept that the higher the ux the stronger is the drag force towards the membrane, the stronger concentration polarization and the higher the compaction of particles. Critical ux is dened as the limiting ux value below which a ux decline over time does not occur [11]. A number of parameters inuenced this critical ux have been discussed in detail and can be found elsewhere [12]. It is maintained that if one operates below the critical ux the fouling can be avoided or minimized. Fouling is common to all types of membrane separation. The type of fouling various from Microltration (MF) membrane processes where hydrodynamic force can predominate to RO membrane processes where hydrodynamic forces have minor effects compared to the forces associated with particles and their interaction with the membrane surface. 3. Fouling 3.1. Background To devise effective cleaning strategies a thorough understanding of membrane fouling and it causes is required and rst part of the review is therefore dedicated to the nature of fouling and fouling processes membrane fouling is an extremely complex phenomenon that has not been dened precisely. In general the term is used to describe the undesirable formation of deposits on membrane surfaces. This occurs when rejected particles are not transported from the surface of the membrane back to the bulk stream. The foulants are typically colloidal materials of one sort or another and these properties and interaction with the membrane dominate fouling/cleaning processes. Colloids are dened as ne suspended particles in the size range of a few nanometres to a few micrometers. Examples of common colloids sized foulant include inorganic (clays, silica salt, and metal oxides), organic (aggregated natural and synthetic organic), biological (bacteria and other micro-organism) [7,1318]. Champlin [19] reported

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

Fig. 1. Schematic energy versus distance proles of DLVO interaction proles (a) Surfaces repel strongly; small colloidal particles remain stable. (b) Surface come into stable equilibrium at secondary minimum if it is deep enough; colloids remain kinetically stable. (c) Surfaces come into secondary minimum; colloids coagulate slowly. (d) The critical coagulation concentration. Surfaces may remain in secondary minimum or adhere; colloids coagulate rapidly. (e) Surfaces and colloids coalesce rapidly. [10].

that removing of the particles size of down to 1 m may not be sufcient to avoid fouling in many cases. Not only do MF and UF process sometimes fail to remove all colloids below a few hundred nm in diameter but also conventional processes used to pre-treat NF feed water fail to remove sub-micron colloids [10]. The high concentration of the rejected ions in the membrane surface could encourage aggregation of dissolved matter in to colloidal sized particles. More to the point, the inuenced of salt retention and concentration polarization in the vicinity of the membrane surface screens electrostatic particlemembrane and particleparticle interactions allowing colloids to foul the membrane. The sites fouling of membrane can be divided into external surface fouling (build-up of a cake/gel-like layer on the upstream face of a membrane) and pore blocking fouling [20]. In a dead-end ltration system, the latter is divided into three types: complete pore blocking (blocking a pore by a particle with approximately the same as the pore size), incomplete pore blocking (intermediate fouling) and standard pore blocking (gradual pore narrowing and constriction by particle that is much smaller than the pore size) [21]. The ux decline of NF membrane is mainly attributed to the pore blocking and it is observed in dead-end and the cross-ow ltration systems.

Vrouwenvelder and Kooij [22] showed that diagnosis of the type/cause of fouling is an essential rst step aiming at controlling fouling. Autopsy gives conclusive information and further understanding about the types and extent of fouling in the membrane ltration plant and provides specic ways for reduction and control of fouling. The tools which have been developed for diagnosis, prediction, reduction, and control of fouling have proven their value in controlling fouling in practice. An overview of the tools is shown in Table 1 [23]. A set of coherent tools has been developed for (i) Determining the fouling potential of the feed water. (ii) Analyzing the fouling of NF and RO membranes. The tools presented can be used to (a) assess the cause of fouling, (b) further dene criteria for feed water to predict and minimize the risk of fouling and (c) evaluate cleaning strategies. Appropriate use of these tools can provide strategies for cleaning to reduce the operational costs of membrane plants (Table 1) [23]. When fouling takes place on the membrane surfaces its causes ux decline leading to an increase in production cost due to increased energy demand, chemical cleaning, reduction in membrane life expectancy and additional labor for maintenance. The types of NF Fouling can be classied on the basis of fouling material into three types [2225]:

8 Continuous on-line monitorBrine stage module with a single spiral wound membrane element

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

Optimizing recovery, acid dose and anti-scalant dose

ScaleGuard

Scaling

1. inorganic fouling due to deposition on membrane surface of inorganic scales (mainly BaSO4 , CaSO4 CaCO3 ), 2. organic fouling due to natural organic material (NOM) found in the process stream (humic acids, protein and carbohydrate), and 3. biofouling due to microbial attachment to membrane surface followed thereafter by their growth and multiplication in presence of adequate supply of nutrients in the pretreated feed or nutrients that deposited on membrane surfaces. 3.1.1. Inorganic fouling Scale formation at the membrane surface is serious problem and resulting from the increased concentration of one or more species beyond their solubility limits and their ultimate precipitation onto the membranes [26]. In order to avoid scaling, it is very important to operate NF systems at conditions lower than the critical solubility limits, unless the water chemistry and physical conditions are adjusted to prevent the type of precipitation. Currently, due to the complexity of the problem, there is no reliable way to predict the limiting concentration level at which there is no a risk of scale formation with a given membrane system and treated water. Similarly, specic antiscalant treatments are hard to dene with condence [27]. Schafer et al. [10], have reported that scaling (scale formation) or precipitation fouling, occurs in a membrane process whenever the ionic product of a sparingly soluble salt in the concentration stream exceeds its equilibrium solubility product. The term membrane scaling is commonly used when the precipitate formed is a hard scale on the surface of the membrane. Scaling usually refers to the formation of deposits of inverse-solubility salts such as CaCO3 , CaSO4 xH2 O, silica, and calcium phosphate. Inorganic scale formation can even lead to physical damage of the NF membrane, and it is difcult to restore NF membrane performance due to the difculties of scale removal and irreversible membrane pore plugging [10]. The greatest scaling potential species in NF membrane are CaCO3 , CaSO4 2H2 O and silica, while the other potential scaling species are BaSO4 , SrSO4 , Ca(PO4 )2 , ferric and aluminium hydroxides [26,28]. Calcium sulphate precipitates in six different phases, dihydrate (so-called gypsum), two hemihydrates and three anhydrites, although at ambient temperatures (about 20 C), gypsum is the most common. The other phases are the product of gypsum dehydration at relatively higher temperature, whereas the calcium carbonate precipitates in three phases: calcite, valerite and aragonite. The most common crystal of calcium carbonate is calcite. It is widely accepted that the crystallization (precipitation) of salts that takes place on the NF membrane surface requires at least two stages, a nucleation stage and a crystal growth from supersaturated solution. It is important to be familiar with the mechanism of scale formation in order to avoid ux reduction through membrane. Gilron and Hasson [27,2931] considered that the ux decline was due to the blockage of the membrane surface by lateral growth of the deposits on the membrane (heterogeneous crystallization (two phase)) whereas Pervov [2931] reported that the ux decline was due to the crystal formation that took place in the bulk solution followed by crystal deposition

Dead end equipment

Table 1 overview of tools available for determining the fouling potential of feed water and fouling diagnosis of NF and or membranes used in water treatment adopted [90]

Specic Oxygen Consumption Rate (SOCR)

Assimilable organic carbon (AOC), cylinder (glass) surface

Fouling diagnosis

Comment

Method

Tools

Biofouling, inorganic, compounds and particles SEM, EDX, XRD, FTIR, AFM XPS, Zeta potential, Contact angle, NMR and Chemical analysis of the foulant by TOC, ICP,HPLC etc. Comparison of fouled and un fouled system isolate the differences both chemical and physical properties of the membrane. This conrmed with water treatment properties provide likely causes of fouling

Integrated diagnosis (autopsy)

Predictive and prevention of Biofouling by determining the (growth) potential of water

Biolm monitor of biofouling formation rate

Biofouling

Non-destructive method for determining active biomass in membrane systems

Biofouling formation rate

Biofouling

Particulate fouling potential of water

Particulate

MFI-UF

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

Fig. 2. Scale formation mechanisms in NF membrane (a) show homogenous precipitation in the liquid phase while (b) show heterogeneous precipitation between liquid phase and solid surface phase and the factor effect the crystallizations [28].

on the surface of the membrane (homogenous crystallization). Clearly, this process will be a mixture of these two extremes and will be affected by membrane morphology and process conditions. Fig. 2 represents homogeneous and heterogeneous modes of crystallization [30]. Her et al. [8,32], reported that if the surface of the solid substrate matches well with the crystal and the interfacial energy between the two solids is smaller than the interfacial energy between the crystal and the solution, then nucleation may take place at a lower saturation ratio on a solid substrate surface (heterogeneous crystallization) rather than in the solution (homogenous crystallization) [8,32]. When the bulk phase becomes supersaturated due to the increasing of concentration polarisation layer, it is possible that both mechanisms of crystallization simultaneously occur in NF system [30]. Aluminium oxide, inorganic salts, clays, sand and biological surfaces can also act as suitable substrates for crystallization [33]. Dydo et al. [8] have reported that most researchers indicate that the gypsum scale precipitates as a bulk phase precipitation process (homogenous crystallization) rather than on membrane surface (heterogeneous crystallization). Lee et al. [30] has also demonstrated that the homogeneous crystallization in the

retentate is a more important mechanism than heterogeneous crystallization of the membrane fouling and ux decline [30]. Hasson et al. [27] have reported that the effect of CaSO4 scaling in RO and NF membranes on the ux decline was a function of the super-saturation level on membrane surface and in the bulk. Various physical and chemical parameters that affect the crystallization process within a membrane system and include temperature [34,35], pH [36], ow velocity, permeation rate [37], types of pretreatment [38], salt concentration and concentration polarization [30,3942], membrane type, materials [37] and metal ions [43]. In addition to these parameters, NOM has also been considered to affect various forms of scaling [44]. These factors also have been summarized as to whether they increase or decrease the scaling, in Table 2 and more information can be found elsewhere [45]. It can be drawn from this table that there are several factors either alone or combined with each other play an important role in crystallization and subsequent cake formation. However, this is further complicated by the nature of mixed solutions of how these alter substantially the solubility product, strength and morphology of precipitates from those in pure state. When the structure of the precipitate is

10 Table 2 scaling factors [38]

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

Value Ionic strength CP Co-precipitation pH Pressure Velocity (ow rate) Temperature Surface morphology High High Presence Higher Higher Higher Higher Higher

Crystallization Increased Increased Increased Increased Increased Decreased Increased Increased

Cause Solubility and supersaturating Solubility and supersaturating Changing structure of the precipitate Solubility decreased Increasing CP and Osmotic pressure at membrane surface (solubility and supersaturating) Higher wall shear rate Solubility decreased Valley blocking

altered so are the intramolecular forces holding the precipitate together [40]. 3.1.2. Organic fouling In general, NF membrane are used in water treatment as alternative processes for the removal of natural organic matter NOM that cause contamination, taints and color and are vehicles for other materials that bind to these substances [46]. Organic fouling could cause either reversible or irreversible ux decline. The reversible ux decline, due to NOM fouling, can be restored partially or fully by chemical cleaning [24]. Whereas the irreversible ux decline can not be restored at all even by rigorous chemical cleaning is applied to remove NOM [47]. Membrane fouling in the presence of NOM can be inuenced by: membrane characteristics [25,4852], including surface structure as well as surface chemical properties, chemistry of feed solution including ionic strength [51,53], pH [48,50,51,5458]; the concentration of monovalent ions and divalent ions [50,51,54,59,60]; the properties of NOM, including molecular weight and polarity [25,49,52,61,62]; the hydrodynamics and the operating conditions at the membrane surface including permeate ux [25,51,6365], pressure [47,50,66], concentration polarization [50], and the mass transfer properties of the uid boundary layer. These factors either increased or decrease the fouling rate have been summarized in Table 3 and more information can be found elsewhere [45]. As it can be seen the Table 3 that the chemical (Ionic strength, NOM fraction, etc.) and physical parameters, such as pressure, velocity,
Table 3 Natural organic matter fouling factors [38] Value Ionic strength concentration pH Divalent cations NOM fraction Molecule or membrane Charge CP Surface morphology Permeate ux (High recovery) Pressure Increased High pH Low pH Presence Hydrophobic Hydrophilic High charge High Higher Higher Higher

and permeate ux, play a major role in NOM fouling at NF membrane surface. Humic substances in aquatic environments are considered to be the major fraction of NOM, are refractory anionic macromolecules of low to moderate molecular weight. Humic substance contains both aromatic as well as aliphatic components with primarily carboxylic (carboxylic functional groups account for 6090% of all functional groups) and phenolic functional groups [67]. As a result, humic substances generally are negatively charged in the pH range of natural waters [18]. Nilson and DiGiano found that only the large molecular weight fraction of NOM contributed to the layer formation. In addition, while studying the effect of NOM properties on fouling of NF membranes, they fractionated NOM into hydrophilic and hydrophobic components. They found that the hydrophobic fraction was the major factor causing permeate ux decline while the hydrophilic fraction had relatively small effect [68]. The hydrophobic fraction of NOM tends to adsorb more than hydrophilic fraction of NOM to the membrane surface. The hydrophobicity of the NOM increases with increasing molecular weight [18,68]. Jucker and Clark have also observed the same trend [69]. The fouling effect of divalent ions on high molecular weight of NOM was more pronounced than with low molecular weight of NOM. Braeken [70] et al. have reported that hydrophobicity and molecular size play an important role in retention of dissolved organic compounds. Hydrophobicity is the most important parameter determining the retention of molecules with a molecular

NOM fouling Rate Increased Increased Increased Increased Increased Decreased Increase Increased Increased Increased Increased

Cause Electrostatic repulsion Hydrophobic forces Electrostatic repulsion Electrostatic repulsion and bridging between NOM and membrane surface Hydrophobicity Electrostatic repulsion Valley blocking Hydrophobicity Compaction

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

11

Fig. 3. Conceptual sketch of the swollen membrane matrix for different ionic environments (a) thick electrical double layer at high pH and low ionic strength and (b) thin electrical double layer at high ionic strength and low pH) [51].

Fig. 4. Conceptual sketch of hypothetical macromolecular endpoints from polyelectrolyte solution theory: (A) high pH, low ionic strength, low solute concentration; (B) low pH, high ionic strength, high solute concentration [51].

weight below the molecular weight cut off (MWCO) of the membrane, presumably binding on or in the membrane surface. The apparent membrane surface structured in the solution is a function of pH and ionic strength. Fig. 3 shows the potential impact of high and low ionic strength on membrane structure. At high ionic strength, the membrane pore size was found to exhibit larger pore size compared at low ionic strength. The forces that control secondary and tertiary structure of NOM are also altered with increasing salt concentration, and results the slow restructuring or transition of the NOM particles [57,71]. For example, the NOM particles can stretch to more linear chains at low concentrations, low ionic strengths and at neutral pH because of higher intermolecular repulsion. Whereas a rigid, compact, spherocolloidal macromolecule is found at high ionic strength, low pH and high solution concentration when intramolecular charge shielding acts to neutralize functional groups (Fig. 4) [53]. The pH also has a major effect on the fouling behavior of humic acids. Typically humic acid contains carboxylic acid groups that lose their charge at acidic pH. At low pH (<4) macromolecules of humic acid have a smaller macromolecular conguration due to increased hydrophobicity and reduced interchain electrostatic repulsion and hence these macromolecules pass more easily through the membrane pores (size exclusion) [51,53]. Mika et al. showed that at pH 45 the membrane and humic acid are almost uncharged and that these conditions promote fouling [48]. Consequently, NOM fouling at high pH is possible when the calcium concentration is high can precipitates and the NOM adsorbs on the calcite surface [58,64]. Schafer has also indicated that the precipitation of a calcium-organic complex is co-precipitated or absorbed occurs at pH 10. He concluded that deposition of calcium increases with pH so more organics are absorbed on the calcite surface [50]. As a results, the Ca2+ and Na+ rejection of the membrane were decreased by foul-

ing [5457]. The same result and trend have found by Manttari and his group [48]. This phenomenon has also elegantly demonstrated by using non-contact AFM imaging in high and low salt environment. Bowen et al. [72] concluded that High electrolyte concentrations give the sharpest images with the best denitions of the pores in case of Ultraltration membrane (ES625). At the high ionic strength (101 M) the isopotential lines follow the pore entrance more closely than at the low ionic strength, not only was the image obtained in 103 M solution less sharp, but also the mean pore size determined was greater than at the higher ionic strengths and also greater than the value obtained in air (Fig. 5). This behaviour will also occur with NF membrane system. Thorsen showed that the most critical particle sizes for NOM fouling fall in the range 0.11.5 m [73]. In a more detailed study, Hong and Elimelech investigated the role of chemical and physical interactions in natural organic matter (NOM) fouling of NF membranes. The role of solution ionic strength, pH, and divalent cations in NOM membrane fouling, as well as the fouling mechanisms involved, are reproduced in Fig. 6. of these chemical factors, the presence of divalent cations, such as calcium and magnesium, has a marked effect on NOM fouling by allowing ionic bridging between NOM molecules. In addition, it was shown that rapid membrane fouling occurs at high permeation rates, even under chemical conditions not favorable for fouling, such as low ionic strength, low levels of divalent cations, and high pH. Therefore the rate of fouling is controlled by the interplay forces in the solute particles between permeation drag and electrostatic double layer repulsion. NOM fouling of NF membranes involves both physical and chemical interactions. The addition of a strong chelating agent (EDTA) to feed water reduces NOM fouling signicantly by removing free and NOM-complexed calcium ions [51]. A more complete understanding of the fouling factors of NOM and membrane characteristics could build the foundation

12

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

Fig. 5. numerically calculated isopotential lines at the entrance to a membrane pore of diameter 0.1 I.tm. (a) In 101 M solution, (b) in104 M solution. The pore size distribution obtained from AFM images at various concentrations. The line BC is the front of the membrane, CD is the internal pore wall, FE is the axis of symmetry along the centre of the pore (due to symmetry, only a half section is shown), AB and AF are the natural boundaries in the solution and DE is a natural boundary in the pore. The front surface and pore wall of the membrane have a normalised potential of 1.0. [60].

required to gure out the mechanisms of NOM fouling during driven-pressure membrane applications. A systematic and comprehensive study is still needed in order to identify key parameters that could be used effectively for the prediction of

NOM fouling in order to maintain the ux. In addition to solute characteristics, a comprehensive understanding of membrane properties is also needed to predict solutemembrane interactions and eventually NOM fouling. Clearly there is scope to

Fig. 6. Schematic description of the effect of solution chemistry on the conformation of NOM macromolecules in the solution and on the membrane surface and the resulting effect on membrane permeate ux. The NOM fouling described in the diagram is applicable for permeation rates above the critical ux. The difference between the two chemical conditions shown becomes less clear at very high permeate ux. At low permeate ux (below the critical ux), no signicant fouling is observed for both conditions [22].

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

13

improve processes and membrane by understanding the nature NOMmembrane interaction. 3.1.3. Biofouling Biofouling is a term used to describe all instances of fouling where biologically active organisms are involved [74]. This is distinct from NOM fouling caused by contaminated organic matter that may be derived from biological systems. Membrane biofouling is caused by bacteria and to a lesser degree, fungi and other eukaryote microorganisms [75]. Biofouling is a dynamic process of microbial colonization and growth, which result in the formation of microbial biolms. Biolm formation invariably precedes biofouling, which becomes an issue only when biolms reach thickness and surface coverage that may cause problems such as declined normalized ux and/or increase in normalized pressure drops during NF or RO operation [22,76]. Many products from biolms have been shown to enhance inorganic precipitation through enhanced nucleation and crystallization kinetics, e.g. carbonate and silicates. Biofouling can be controlled by (1) removal of degradable components from the feed water, (2), ensuring the relative purity of the chemicals dosed and (3) performing effective cleaning procedures. Also, it has been reported that cleaning procedures applied when fouling is not a problem might delay biolm formation [77]. The surface of the membrane offers good site for microbial colonization as it concentrates nutrients for growth. 3.2. Operational aspect of NF and fouling From the mechanisms of fouling process above, many operating procedures have a direct impact on fouling of membranes. This section reviews the effect of membrane process design on fouling. In most cases ux rate is considered as a key design parameters for membrane system and reect membrane productivity. The two factors that lead to deterioration the ux rate are fouling and concentration polarization. In order to overcome these shortcomings the membrane array is to be introduced [21,67,7881]. An appropriate membrane array was considered in designing membrane treatment system in order to reduce the effects of both concentration polarization and to minimize the membrane fouling. Typically, membrane systems use multiple parallel modules so that the plant performance in terms of the product quality and recovery remain identical for a single module (Straight brine stages; typically a single module contains six elements). In the tapered systems (Tappered brine stages; membrane array design 2:1, 3:2:1, 4:2:1), the feed stream is passed through the rst module (or parallel set of modules) and is divided two streams. These streams are the product and the reject stream, the reject stream from rst module (or parallel set of modules) is passed through as feed to the second module (or set of modules). Here the velocities are boosted at each stage by decreasing the number of modules in parallel. Thus it is possible to obtain a high recovery while still avoiding the worst effects of fouling and concentration polarization (Fig. 7). The membrane arrangements are designed with the aim of minimizing fouling and reducing concentration polarization by increasing number of stages and reducing the number of ele-

Fig. 7. Membrane arraysstraight and tapered brine stage (a) straight brine stage one pressure vessel contained six element in series), (b) tapered brine staging in the ratio of 2:1 each presser vessel have six membrane elements), (c) tapered brine staging in the ratio of 3:2:1 each presser vessel contained four membrane elements) and (d) tapered brine staging in the ratio of 4:2:1 each pressure vessel contained four elements).

ments per stage in order to maintain the same or high recovery will involve the following constraints [10,82]: 1. The ow rate should not exceed the maximum ow rate per element, qmax , to avoid large axial pressure drops which could cause membrane element damage such as telescoping. 2. There is a lower limit on the ow rate per element, qmin , in order to control concentration polarization and scaling. 3. There is a maximum recovery for each stage as well as overall maximum recovery in order to minimize the fouling. 3.3. Primary location for specic types of fouling When reviewing the major causes of NF membrane fouling and associated mechanisms, it is very important to understand where the fouling takes place in membrane system in order to arrange the module and optimize uid handling (see section above). Typically the fouling typically takes place either in the lead element (rst element in the pressure vessel) where particles became entrapped on the surface or in the end element (last element in the same pressure vessel) where salts are highly concentrated). Usually organic and metal oxide fouling take place in the rst stage of lead element, metal oxide and colloids deposit early in the process as drag forces are relatively high. However, organic fouling usually occurs heavily in the feed side of the

14

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

Table 4 where fouling occurs rst, adapted from hydranautics technical service bulletin tsp [82] Type of foulant Scaling/silica Metal oxides Colloids Organic Biofouling (rapid) Biofouling (slow) Most susceptible stage of NF/RO Last membrane in last stage First membranes of rst stage First membranes of rst stage First membranes of rst stage First membranes of rst stage Throughout the whole installation

pretreatment reducing Ca2+ and Mg2+ to very low level was achieved using a complex multi-stage process of coagulation and occulation using lime; mechanical bed ltration; weak cation ion exchange and deep-cartridge ltration, were successful in obtaining pretreated water with Ca2+ and Mg2+ under the detection limit. 3.4.2. Scale inhibitors Another approach is to avoid scale formation the addition of scale inhibitors. It is obvious that scaling intensity depends upon the chemical composition of feed water; therefore water with a high scaling potential requires treatment using scale [30]. The chemical species, such as lime and soda or caustic soda are added to hard water in order to remove or reduce the hardness ions. Alkaline chemical additives are added to hard water to raise the pH in order to convert bicarbonates to carbonates and then calcium and magnesium are removed from water as CaCO3 and Mg(OH)2 prior to ltration. Zero hardness water can not be achieved due to the limited solubility of CaCO3 and Mg(OH)2 . 3.4.3. Membrane preltration and membrane modication The application of microltration (MF) as well as ultraltration (UF) as NF prelters has emerged in the last decade as an efcient method in pretreating surface water [88]. Both UF as well as MF membranes offer good physical barrier to colloids, suspended particles as well as microbes. Both MF and UF membrane can be used ahead of desalination units and have capability of ltering out particles in the ranges between 0.005 to 0.1 (UF) whereas 0.1 to 3 (MF) [63]. Attempts have been made to modify membrane surfaces in order to make them less vulnerable to fouling. In some cases, the surface roughness increases membrane fouling by increasing the rate of attachment onto the membrane surface and hence the membranes with a rough surface is more prone to fouling than membrane with a smoother surfaces [50,51]. Colloidal interactions are also important in fouling and charged components tend to cause fouling because of electrostatic attractions between charged components and the membrane (see above). Development of membranes with lower surface charge or surface charge similar to that of the foulant, with hydrophilic character may help solve these specic problems [89]. 3.4.4. Sonication The effect of the particle concentration on the ultrasonic control of the membrane fouling was investigated by Chen et al. [90]. The basic principle of operation is that ultrasound removes particles from the surface by causing particle movement in or near membrane. In this experimental work it was concluded that the ultrasound reduced ceramic membrane fouling by silica particles during cross-ow ltration. At low particle concentrations, there was a little membrane fouling in the presence of ultrasound. However, the permeate recovery of the ultrasound treated membrane decreased with an increased in particle concentration. At low particle concentrations (lower than 0.8 g/L) the particle concentration effect was more apparent when the membrane was far away from the cavitation region. However, at higher

module. While biofouling can be found throughout the ltration stages, however, rapid biofouling was found mostly in the feed side as a result of particle and nutrient attachment [83]. In general, scaling and silica fouling take place in the brine side membrane elements when the concentration of inorganic salts exceeds the solubility limit. The types of the foulants and where they usually cause fouling in typical NF/RO systems are summarized in Table 4 [83]. There are two strategies in order to minimize the effect of fouling, and these can be classied into two major groups minimization and remediation. Both of the strategies are practiced in membrane process industries. 3.4. Fouling minimization It is possible to avoid or control fouling to certain extent by using adequate pretreatment such as coagulation/precipitation, or slow sand ltration and membrane surface modication. 3.4.1. Coagulation followed by ltrationsedimentation Conventional coagulation ltration pretreatment was designed to remove most of the potential foulant materials from the pretreated feed by preltration and more rarely by sedimentation. The degree of the pretreatment, however, is dependent on the raw water quality, particularly its content of organic (including biological) and inorganic suspended matter. Coagulant and coagulant aids can be added in a pretreatment to increase separation efciency. Several studies were have been carried out on the pretreatment side in order to remove the foulant materials by optimizing the operational conditions of the pretreatment process such as ow rate, backwash frequency, pH, etc. [84,85]. Howe and Clark [86], reported studies that were focused on the effect of coagulation on the fouling by dissolved and particulate colloidal matter. Tests with and without preltration were able to provide a comparison between the effect of particulate versus dissolved and colloidal matter. Usually, less than 20% of the fouling in their experiments could be attributed to particulate matter. They concluded that when the water was treated with coagulant, the fouling usually decrease after preltration and suggested that the coagulated particulate matter was able to form a dynamic layer material on the membrane surface. Thus, the ne particles could remove materials that would otherwise foul the membrane. When the dynamic layer was eliminated by preltration the fouling actually worsened [86]. Earlier work [87] on surface water

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

15

particle concentrations of greater than 0.8 g/L the effect of particle concentration was more pronounced when the membrane sonifacation power was close to the cavitation region. Standard water treatments, carried out in order to solve specic pretreatment problems, could lead to further fouling problems. Walton [91], outlined some of these problems as follows: The use of certain phosphate anti-scalants stimulates biological activity in both the internal and external environments. The use of organic biocides to control biological growth often results in organic slime formation and subsequent after growth activity and colloidal entrapment. The use of occulants to control particulate matter results in colloidal iron or aluminium oc fouling, especially in association with organic slimes. The introduction of oxygen or oxidants into anaerobic systems results in iron and sulphur precipitation and potential stimulation of iron hydroxide slime-production. Addition of minerals acid anti-scalant can produce available CO2 allowing biological growth, especially algae and autotrophic bacteria. The use of activated carbon for dechlorination and/or organic removal results in an excellent substrate for bacterial growth producing fouling byproduct; it can also absorb polyelectrolyte and organic antiscalants. 4. Membrane cleaning 4.1. Remediation of the membrane Remediation is usually conducted by chemical cleaning for nearly all membrane processes and application. However, the frequency of the chemical cleaning could range from a routine daily process such as in whey processing to long term annual processes such as in desalination plant according to occurrence of fouling [89]. In general, much of the decline in membrane performance can be corrected by cleaning the membrane. Cleaning can be dened as a process where material is relieved of a substance, which is not an integral part of the material, [92]. Physical cleaning methods include for example: hydrodynamic forward or reverse ushing, permeate back pressure, air spurge and automatic sponge ball cleaning. These methods depend on a mechanical treatment to dislodge and remove foulants from the membrane surface. Application of these methods usually results in a more complex control and design of equipment. The physio-chemical cleaning methods use mechanical cleaning methods with the addition of chemical agents to enhance cleaning effectiveness [93]. Adequate pretreatment and appropriate membrane selection as mentioned above can slow the fouling rate, but the membrane cleaning is an essential step in maintaining the performance of the membrane process. The ideal cleaning processes should not only be effective against several foulants, but gentle to the membranes so as to maintain and restore their characteristics. The optimal (the least membrane damage and maximal effectiveness of cleaning) choice of the cleaning agent is a function of

membrane material as well as foulants. Fu et al. [94] noticed that two NF membranes with different properties (TS 80 and NT47450), fed by the same feed water, required different cleaning processes. The results of cleaning procedures are sometimes very difcult to determine using only ux recovery data. It is of interest to know in what way the cleaning agent interacts with the membrane and whether it actually modies the membrane surface structure and chemistry in such a way that fouling is prevented. It has been noticed that cleaning often increases the ux of the virgin membrane [95]. The chemical reactions between the chemical agents and the foulant takes place either by changing the morphology of the foulant or by altering the surface chemistry of fouling layer in order to remove the foulants from the membrane surfaces [96]. Kosutic and Kunst [97], concluded that an irreversible change in the porous structure of NF membrane was observed as a result of the chemical cleaning. Cleaning may make the pore surfaces more hydrophilic and charged by the adsorption of the chemical agent [98]. Chemical cleaning procedures and commercial membrane cleaning products are almost specied by membrane manufacturers [96,99]. 4.2. General considerations and costs In any membrane processes, the need for proper and periodical cleaning is essential regardless of the type feed be; seawater, brackish water, wastewater or industrial water. The objective of the cleaning processes is to restore membrane performance when it falls below the expected permeate yield typically by about 10%, or feed pressure increase by about 10% and/or differential pressure increase by 1550% [24]. Membrane replacement is a necessary part of the plant operation that is needed to maintain the quality of the product water to the protocol agreed with membrane manufacturers as well as to meet the design productivity when the cleaning processes fail to restore the declined ux [100]. Usually about 10% of the membrane is annually replaced in order to maintain the targeted product quality as well as quantity. It has been reported that the cost of the membrane replacement is about 23% of product water cost from Jeddah SWRO Plant at power cost $ 0.1 kWh1 (water cost = 1.473 $/m3 ) [101]. Although, there are a number of cleaning techniques such as physical or chemical or combination of both, only the chemical cleaning methods are widely used by NF and RO industries for membrane cleaning and regeneration. The complexity and detailed understanding of cleaning processes has not yet been addressed by many researchers and is needed for a clear knowledge of these processes. Although cleaning is intended to restore the ux, it often deteriorates product quality and increases the cleaning frequency affecting plant availability. For example, cleaning processes sometimes takes 12 days to complete in large plants [24]. Desalination plant availability is usually designed to be in the range of 9097% and varies according to the type of water being treated. However, this percentage can be reduced if the cleaning frequency is increased, but the costs routine of plant maintenance, the additional manpower utilization and energy consumed during cleaning processes can increase the overall cost of water production. In general, the chemical consumption of the plant per

16

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

year is about 0.31% of total water treatment cost, nevertheless, the chemical consumption of the cleaning process per year is much higher than the annual chemical consumption for the overall RO process (conditioning etc.,) [102]. These general cost gures exclude the additional facilities, manpower and energy consumed for cleaning. It is well recognized that the energy cost of the plant is about 5060% from the total water cost [101103]. Moch [102] stated that power, itself, can be a half to three quarters of the operational and maintenance costs. In general the cleaning process increases the overall system energy efciency, regardless of the energy consumed during the cleaning. For example by reducing the net driving pressure will be reduced after cleaning by about 1030% which is quiet considerable energy saving, especially during the plant operation [24]. 4.3. Assessment of cleaning agents 4.3.1. Type of cleaning A large number of chemical cleaning agents are commercially available, and commonly used ones fall into six categories: alkalis, acids, metal chelating agents, surfactants, oxidizing agents and enzymes [104,105]. Commercial cleaning products are usually mixture of these chemicals but the actual composition is often not clearly specied. Table 5 [106] shows the chemical cleaning agent recommended by various membrane manufacturers. The table gives the details of chemicals and its concentration to be used for different type foulants. The choice of the preferred cleaning product depends on feed characteristics. For example, acid cleaning is suitable for the removal of precipitated salts, such as CaCO3 , while alkaline cleaning is used to remove adsorbed organics [89]. When surfactant is introduced on the membrane surface in order to restore the membrane ux, surfactant adsorption is possible from hydrophobic interactions between the hydrophobic portion of the membrane surface and hydrophobic tails of the surfactant. In spite of increased electrostatic repulsion between the negatively charged membrane and anionic surfactant, even at a low surfactant concentration may adsorb on to a negative-charged surface due to an ion exchange mechanism leading to a higher concentration of surfactant near to membrane surface compared to bulk solution, which may induce a micellization process at the membrane solution interface [107]. On the other hand, cationic surfactant could lead to a reduction in membrane permeability owing to membrane modication with a cationic surfactant [108,92]. As mentioned earlier, NF membranes are extremely vulnerable to natural organic matter (NOM) fouling, especially in the presence of divalent cations [109,50,51]. Characterization of NOM-fouled membranes by contact angle, zeta potential, and attenuated total reection-fourier transform infrared (ATRFTIR) spectroscopy as well as molecular weight distribution measurements demonstrated that colloidal material with heterogeneous characteristics with variable area of hydrophobicity and charge membrane [104]. These materials are typied by humic acids, fulvic acids, proteins and peptides. Typically, NOM tend to have higher hydrophobic fraction of about 75% compared to about 20% of hydrophilic fraction in water such as Orange

county ground water. While Horsetooth reservoir surface water (HT-SW) found to have high fraction of hydrophilic NOM of about 65% compared to about 27% of hydrophobic NOM fraction. Lee reported that a caustic solution was more effective than citric acid for fouled membrane with the hydrophobic fraction of NOM. On the other hand, chemical cleaning agents were not able to clean fouled membrane by hydrophilic fraction of NOM, because of lack of electrostatic repulsion between NOM acids and the negatively charged membrane surface [104]. This was due to high ionic strength of the feed solution masking the membrane surface charge. In general, alkaline cleaning recovers the ux, while the introduction of alkaline chelating agent further increases the ux. Liikanen et al. [109] reported that alkaline chelatant such as EDTA increased the ux more than plain alkaline cleaning (NaOH) due to membrane charge increase in EDTA alkaline environment, which makes the membrane more open. Liikanen et al. concluded that alkaline and chelating cleaning agents increased membrane ux, but they reduced the ion retention, whereas acidic cleaning could be used in order to recover membrane ion retention. In a recent study, Li et al. [110] noticed that combined simultaneous process of NaOH with sodium dodecyl sulfate (SDS) demonstrated greater cleaning power and cleaning efciency by about more than 100% compared to that of single cleaning with each of NaOH or SDS alone. This is also true when two step method in which SDS cleaning step was performed after caustic treatment [111]. Jacques et al. [112] reported that hydrochloric acid cleaning showed better results than citric acid in removal of the iron deposition on the membrane surface [32]. Song [51,113] reported that sequential use of both caustic and acid cleaning was more effective, in terms of high ux recovery, than caustic or acid alone in removing both acidic and basic fractions of NOM. Also he reported that the caustic cleaning was found to more effective than acid cleaning in removal of the NOM foulants. This is a result of the presence of hydroxyl ions in caustic solutions, which could promote disruption of the foulant layer by these mechanisms: (i) increasing ionic strength, (ii) increasing solubility of NOM foulants, (iii) increasing pH. Increasing the pH should result in an increased negative charge of NOM, because of deprotonation of the carboxyl and phenolic groups. Conversely, decreasing negative charge of NOM has been observed as a result of adsorption of sodium ions to NOM during cleaning with [81,113]. EDTA and SDS were also used as effective cleaning agents in order to remove virtually or all of the NOM foulant material [51,114]. The acid cleaning is effective in removal of precipitated salts (scaling) from the surface of the membrane and from the pore [10]. The polyamide thin lm membrane (TFM) is very sensitive to disruption by the oxidising agent. Powerful oxidation agents have not been used in order to regenerate membrane performance because of oxidation agent typically causes irreversible damage to these membranes. However, there is a procedure based on a patented chemical cleaning using NaOCl where a known concentration is prepared and recirculated through the membrane cells for 20 min at pH > 10, while chlorine oxidation effects were almost negligible [115]. This cleaning procedure at high pH was effective to remove the organic foulant materials

Table 5 chemical cleaning agents recommend by different manufacturers Type of foulant Type of membranes DuPont B-10 CaCO3 HCl at pH 4, citric acid (2%w) pH 4 (NH4 OH), Nutek-NT 600 (5%w), citric acid (2%w) + Na2 EDTA (2%w), pH 4 (NH4 OH) Citric acid (2%w) pH 8 (NH4 OH), EDTA (1.5%w), pH 78 (NaOH/HCl), sodium hydrosulte, Na2 S2 O4 (1%w) NaOH, pH 11, Biz (0.5%w), pH 11 (NaOH) Citric acid (2%w) pH 4 (NH4 OH), sodium hydrosulte, Na2 S2 O4 (1%w), citric acid (2%w) + EDTA (2%w) pH 4 (NH4 OH), v HCl at pH 4, citric acid (2%w) pH 4 (NH4 OH), NaOH, pH 11, Biz (0.5%w), pH 11 (NaOH), Drewperse 738 (1%w), SHMP (1%w) Formalin (0.252%w) followed by Biz (0.25%w) FilmTec FT-30 HCl (0.2%w), phosphoric acid, H3 PO4 (0.5%w), citric acid (2%w), pH 4, sulfamic acid, NH2 SO3 H (0.2%w) Fluid System Citric acid (1%w), pH 2.5 Nitto Denko Citric acid (2%w) pH 4 (NaOH) Toyobo Toray Citric acid (12%), pH 2.5-4 (NH4 or NaOH), ultrasil 70 0.5% pH 2-2.5, peracetic acid 100200 ppm

CaSO4 /BaSO4 /SrSO4 /CaF2

HCl (0.2%w), phosphoric acid, H3 PO4 (0.5%w), Citric acid (2%w), pH 4, sulfamic acid, NH2 SO3 H (0.2%w)

Sodium tripolyphosphate,STP (2%w) + Na4 EDTA (0.85%w), pH 10 (H2 SO4

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

SiO2

NaOH (0.1%w) + Na2 EDTA (0.1%w), pH 12, max 30 C Phosphoric acid, H3 PO4 (0.5%w), sodium hydrosulte, Na2 S2 O4 (1%w), sulfamic acid, NH2 SO3 H (0.2%w)

Citric acid (2%w) pH 4 (NH4 OH) Citric acid (2%w) pH 4 (NH4 OH) -

Metal oxides

Citric acid (1%w), pH 2.5

Citric acid (2%w) pH 4 (NaOH)

Inorganic colloids

NaOH (0.1%w) + sodium dodecylsulfate Na-DSS (0.05%w), pH 12, max 30 C

Sodium tripolyphosphate,STP (2%w) + Na4 EDTA (0.85%w), pH 10 (H2 SO4

Citric acid (2%w) pH 4 (NH4 OH)

Biological matter

NaOH (0.1%w) + Na2 EDTA (0.1%w), pH 12, max 30 C, NaOH (0.1%w) + sodium dodecylsulfate Na-DSS (0.05%w), pH 12, max 30 C, sodium tripolyphosphate, STP (1%w) + trisodium phosphate, TSP (1%w) + EDTA (1%w) NaOH (0.1%w) + Na2 EDTA (0.1%w), pH 12, max 30 C, NaOH (0.1%w) + sodium dodecylsulfate Na-DSS (0.05%w), pH 12, max 30 C, Sodium tripolyphosphate, STP (1%w) + trisodium phosphate, TSP (1%w) + EDTA (1%w)

Sodium tripolyphosphate, STP (1%w) + trisodium phosphate, TSP (1%w) + EDTA (1%w) pH 1011 (HCl)

Sodium Tripolyphosphate, STP (2%w) + Na4 EDTA (0.85%w), pH 10 (H2 SO4 ), Sodium Tripolyphosphate, STP (2%w) + sodium dodecyl benzene sulfonate (0.25%w), pH 10 (H2 SO4

15 ppm chlorine, pH 6.57.5, Formalin (0.52%w)

Sodium lauryl sulfate 0.2% pH 10-11 by NaOH, ultrasil 10 0.7% pH 10-11

Organics

NaOH, pH 11, Biz (0.5%w), pH 11 (NaOH), SHMP (1%w)

Sodium tripolyphosphate, STP (1%w) + trisodium phosphate, TSP (1%w) + EDTA (1%w) pH 1011 (HCl)

Sodium tripolyphosphate,STP (2%w) + Na4 EDTA (0.85%w), pH 10 (H2 SO4 ), sodium tripolyphosphate,STP (2%w) + sodium dodecyl benzene sulfonate (0.25%w), pH 10 (H2 SO4 )

15 ppm chlorine, pH 6.57.5, formalin (0.52%w)

Flow rate velocity as high as possible, pressure as lowest as possible, temperature does not exceed manufacturer recommendation (<45 C).

17

18

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

Fig. 8. Schematic illustration of the change in the organic fouling layer structure by EDTA (a) compact fouling layer formed in the presence of Ca2+ . (b) Loose structure of the fouling layer after EDTA addition [117].

from membrane surface [78]. Hydrogen peroxide also used as oxidizing cleaning agent at high pH in order to clean the membrane from NOM. A combination of both Cl2 and H2 O2 at high pH were noticed that had remarkable increased in the product ux [78]. 4.3.2. Cleaning mechanisms It is reported that the presence of Ca2+ with humic acid increased the fouling rate of humic acid on the membrane surfaces. Li and Elimelech [116], reported that the proposed mechanisms of chemical cleaning with EDTA in order to clean membrane from NOM (humic acid). Since EDTA forms a strong complex with Ca2+ , humic acid molecules associated with Ca2+ ions are replaced by EDTA via a ligand exchange reaction. EDTA cleaning agent does reduce the intermolecular Ca2+ humic acid complexes and humic acid molecules can be more easily rinsed off the membrane surface as illustrated in Fig. 8. The proposed mechanisms of SDS solubilization of Ca2+ humic acid fouled surface at low, moderate and high concentration are illustrated in Fig. 9. Low concentration of SDS is not sufcient to break the intermolecular bridging formed between humic acid and Ca2+ . When moderate concentration of SDS is used, more SDS molecules partition into the foulant layer results in breakup of some Ca2+ binding. Once the SDS concentration exceeds the critical mycella concentration CMC, it is then strong enough to break up all the Ca2+ -induced bridges, resulting in the dissociation of humic acid to the aqueous phase (as indicated by the zero adhesion with SDS shown in Fig. 10). Childress and Elimelech [117], explored the mechanisms of chemical cleaning with SDS and dodecyl trimethylammonium bromide (DTAB, cationic surfactant) at high pH and low pH. Fig. 11 shows the differences in SDS adsorption at low pH and high pH.

At low pH of 3, the membrane initially has a slight positive and adsorption occurs as a result of electrostatic attraction between the positively charge membrane surface and the negative charged polar head of surfactant ions. The surfactant ions start associating with each other and form surfactant aggregates when the concentration of SDS increases causing dramatic change in the surface charge potential. At pH 8 the membrane has negatively charge and the adsorption will be a result of hydrophobic interaction between membrane surface charge and surfactant tail. When the SDS concentration increases, the membrane surface becomes slightly more negative due to a larger number of adsorbed surfactant molecules. Schematics of adsorption mechanisms of DTAB molecules onto membrane surface are shown in Fig. 12. As the concentration increases the membrane surface charge become more positive due to hydrophobic interactions at low pH. At high pH of 8, the adsorption occurs between the membrane surface and charge polar head of the surfactant will be due to electrostatic attraction. Hemi-micelle formation may take place at the very high concentrations [117]. In all cases the cleaning process depends on the type of foulant deposited on the membrane surface, and for a successful cleaning of fouled membranes, identication of the type of foulant is essential which is done by extensive analysis of the foulants. A destructive autopsy, which can provide a scientic foundation on which to optimize the cleaning procedure, is done as a last resort, when cleaning fails to restore membrane performance. 4.3.3. The impact of cleaning on NF permeate quality An impact of cleaning on NF permeate quality has also been observed. According to Liikanen et al. who performed analysis for alkalinity, hardness and conductivity found that the permeate conductivity generally increased after cleaning [118].

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

19

Fig. 10. Interaction forces between the CML colloid probe and the SRHA fouled membrane surface in the presence of various chemical cleaning agents. The test solution contained 20 mg/l SRHA, cleaning chemical as indicated 1 mM NaHCO3 , 1 mM CaCl2 and NaCl to adjust the total ionic strength to 10 mM. the solution pH during the measurements was xed at 8.1 [117].

Fig. 11. Schematic of adsorption of sodium dodecyl sulfate (SDS) molecules into the membrane surface [118]. Fig. 9. Mechanism of humic acid solubilization by SDS (a) low SDS concentration allows association of humic acid (b) moderate SDS concentration allow partial breaking, and (c) SDS concentration exceeding the CMC allowing solubilization of humic acid. The binding sits shows are solely for illustration purposes [117].

of surface analysis such as atomic force microscopy (surface characterization by visualization and measurement of the surface characteristics), FTIR and zeta potential. The three types of measurement are complementary.

Al-Amoudi et al. [24] also recognized that the increase in permeate conductivity after each chemical cleaning specically after high pH cleaning when it was carried out in commercial NF plant at UmmLujj. However, the acid cleaning following high pH cleaning assisted in partial restoration of the ions retention property of membrane [24,118]. This suggests that acidic cleaning had a role in preserving the membrane ion retention capability, probably by making the membranes tighter by charge neutralisation. 4.4. Methods of assessing the cleaning effectiveness There is several assessment methods of cleaning have been used and well established in order to evaluate cleaning efciency. The most common methods are ux measurements, or forms

Fig. 12. Schematic of adsorption of sodium dodecyl trimethyl ammonium bromide (DTAB) molecules into the membrane surface [118].

20

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

4.4.1. Flux measurement (non-destructive) Flux measurement is a direct assessment of fouling and cleaning process and can be made in the applied situation. There is typically a linear relationship between the ux decline and the deposited mass indicating that the ux decline is due to NF fouling resulting from deposition heterogeneous crystallization [30]. Typically the product water ux declines drastically at higher permeation rates in the presence of NOM refers and relates to the transport processes driven by the hydrodynamics force that acts perpendicular to the membrane surface [18,64]. It is important to establish effectiveness of a particular cleaning protocol. The clean water ux can be measured and compared to the ux of the original steady state process. The water ux recovery (WFR) can be calculated [118] as: WFR = Jc J0 (1)
Fig. 13. Graphical depiction of resistance in ltration, rinsing and cleaning [121].

where Jc is the ux after cleaning and J0 is the ux of the virgin membrane. The measurement of initial ux and the ux after cleaning has to be carried out at the pressure and temperature. Several authors [119,4,17] have proposed the comparison of the hydraulic resistance of the cleaned membrane, Rcw , and the intrinsic hydraulic resistance of the membrane to evaluate the cleaning efciency. Permeate ux data was used to evaluate the hydraulic resistance of the membrane (R), according to Darcys law: P R= (2) J and Ruf = Rm + Rf = Rm + Rif + Rrf (3)

concluded that most of cleaning agents used improved the membrane ux after fouling and some of them even restored the ux up to about 95%. Recently, Al-Amoudi and Lovitt et al. [95] from the results of the permeability of the fouled NF-DK membrane before and after cleaning showed that the cleaning process restored the declined ux close to its original value. Moreover, it was also found that the SDS cleaning agents triple the permeability of the virgin membrane. These results suggest that the chemical cleaning does have a major effect on the ux of NF membrane as well on its surface properties. It has been noticed that cleaning often increases the ux and even the permeability of the virgin membrane. 4.4.2. Atomic force microscopy (AFM) (destructive method) There are now many new surface analysis techniques available for assessing membrane fouling and cleaning processes that based on the visualization of the surface of membrane down to the nanometer scale [122]. These images allow the assessment of surfaces and pore by direct measurement of surface morphology in air and in liquid (process relevant) environment. Analysis of these images can be carried out in a number of ways, the most useful being various measurement and/or dimension of pores [122124]. Atomic force microscopy can be used as good tool to evaluate chemical cleaning procedures performance. Using AFM it was showed that there was an accumulation of the particles in the valleys of rough membranes causing more severe ux decline than smooth membranes [125]. Song et al. [113] reported that signicant difference between the surface morphologies of the virgin and fouled membrane is recognized by AFM. Here the root mean square (RMS) surface roughness of a virgin membrane was about 48 nm whereas; RMS of the fouled membrane was about 124 nm. This increase in surface roughness was observed as a result of the presence of humic acid with calcium on the negatively charged membrane surface. Also, workers reported that, RMS of cleaned fouled membrane was about 40 nm compared the virgin of about 48 nm. This decrease in surface roughness was possibly due to the presence of resid-

where P is the transmembrane pressure; J the permeate ux; and Ruf , Rm and Rf , respectively, Ruf is the total resistance of the intrinsic hydraulic resistance of the membrane plus the total resistance of the total fouling membrane (Rf ), the intrinsic hydraulic resistance of the membrane (Rm ), the fRres is residual resistance after cleaning, and the resistance due to membrane fouling, which combines reversible (Rrf ) and irreversible (Rif ) phenomena. The variation of membrane resistance is depicted in Fig. 13. Cleaning can be assumed to be complete when Rcw Rm allowing for experimental error (Fig. 13) [120]. Cleaning efciency (ERW ) can be determined as ERW = Rif Rres 100 Rif (4)

Both WFR and ERW have been used as a measure of cleaning efciency. There is no difference between the above two methods. However, the hydraulic resistances give a more details to understandings fundamental to the ux. By knowing the membrane hydraulic resistance and other fouling resistance, an understanding of some fouling properties could be obtained. The efciency of membrane cleaning is mostly evaluated by ux measurements [121]. Song et al. reported that the chemical cleaning agents tested could not achieve complete ux recovery as a result of residual foulants were strongly embedded in the concavities of membrane surface [113]. However, Zhu et al.

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

21

sive force was reduced signicantly by NaOH compared to that without chemical cleaning addition. 4.4.3. Fourier transform infrared technique (FTIR) (destructive method) FTIR technique is used to investigate the membrane surface properties and the cleaning efciency. Her et al. used FTIR techniques in combination with other techniques to study precipitation scaling attributed to inorganic scales such as CaCO3 and CaSO4 [32]. Zhu and Nystrom [21] have used the FTIR technique to characterize the chemical cleaning efciency. They concluded from the FTIR results that the fouling notably changed the FTIR spectrum. New peaks appeared and fouling obscured some of the peaks of the polysulfone membrane. The results clearly showed that not all of the fouling had been removed by cleaning. Song et al. [113] reported using FilmTec NF-70 membrane that from the FTIR results of virgin, fouled and cleaned membrane that the peak intensity of the virgin were eliminated or severely attenuated due to coating by NOM foulant, whereas the peaks intensity of caustic cleaned membrane tend to be slightly close to the peak intensity of the virgin membrane. The cleaning agent may not remove the NOM completely from the membrane surface by using EDTA. This is supported by FTIR spectra measurements for virgin membranes, fouled membrane and cleaned membrane with and without preoxidizing water where fouled membrane, exposed to preoxidized water and cleaned with caustic solution, had better peak recovery compared to the fouled membrane exposed to raw water and cleaned with caustic solution [113]. 4.4.4. Zeta potential measurements (non-destructive) The electrokinetic properties of a membrane reveal the electrical characteristics of the membrane surface. Membrane electrokinetic properties are determined by streaming potential or electro-osmosis that involves the relative motion of charged surface and an electrolyte solution, with the results often expressed in term of membrane Zeta potential [129], which is dened as the potential at the surface of shear. By measuring streaming potential with solutions of different pH, the membrane surface isoelectric point can be measured. Streaming potential is determined by measuring the electrical potential difference when there is relative motion between a uid containing charged species and charged surface caused by hydrostatic pressure gradient. The interaction of organic and inorganic colloidal substances with membrane surfaces in aqueous media is dependent on the electrostatic properties of both membrane surface and foulant material. Therefore, study of membrane surface characteristics and its interaction with fouling materials is critical area of membrane fouling research. The effect of surface characteristics has generally been overlooked in efforts to select optimal (the best pretreated quality) pretreatment schemes and operating conditions for various membrane separation processes. The zeta potential measurements method can be used not only to evaluate the surface charge of the NF membrane but also to play an important role in the selection of the NF membrane to be used in the treatment of water [51,130].

Fig. 14. Scanning electron microscope image of a silicon collide probe [128].

ual foulants within the concavities of the membrane surface after cleaning. Warczok et al. [111] recently reported that it is possible to determine from AFM images the mean pore distribution and roughness so indicating whether the cleaning procedure was correctly designed or not. The AFM colloid probe which is powerful technique, has been used to measure the force of interaction between colloid particles and the surface of the membrane. Using AFM it is possible to directly measure the force of interaction in process relevant environments where the cantilever tip of AFM is brought in a contact with membrane surface. The most commonly used AFM tips for force measurement are the sharp silicon tips which provide high resolution when measuring surface topography and force in a process relevant environment, However, attaching a sphere to a tipless AFM cantilever has been used to quantify the surface interactions between a sphere and a at surface as well as different material (Fig. 14). These so called colloid probes give a known geometry when approaching or leaving the surface. These probes typically are about 15 m diameter whose surfaces can be treated with many materials including foulants. With different foulants as probe coatings in liquid medium of different salt solutions or cleaning solution, this technique allows an assessment of the foulant membrane interactions and the chemical cleaning processes [126,127]. The study of the electrical double layer interaction between a particle and membrane by AFM also allows assessment of the propensity of the surface to fouling when in use. Force measurement in conjunction with colloid probe technique also allow a direct quantication of membrane fouling through the measurement adhesive force when the probe is retracted from the surface after contact has been made [128]. Adhesive force measurements were performed utilizing carboxylate modied latex (CML) Colloid probe in the presence of various chemical cleaning agents in order to look into the effect of chemical cleaning on foulantfoulant interactions in the fouling layer. Fig. 10 shows that the adhesive forces were measured with and without chemical cleaning agent addition. The eliminated adhesive force was in the presence of EDTA and SDS gave rise to a complete ux recovery. While the remaining adhesive force with NaOH addition indicates a poor cleaning efciency, although, the adhe-

22

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

Recently, workers [95] concluded that using different types of NF membranes (virgin and fouled) were characterized with zeta potential before and after chemical cleaning agents. Results suggest that the cleaning does have a major effect on surface state properties. However, the mechanisms vary from one agent to another; chemical reactions modify membrane surface properties (surface charge) not only between cleaning agent and virgin membrane but also between cleaning agents and fouled membrane. The study also revealed that SDS cleaning agent has the highest effect on surface charge for both virgin and fouled NF membranes. 4.5. Factors affecting chemical cleaning efciency Cleaning mainly involves the dissolution of the material from the membrane surface and several factors could affect the chemical cleaning process. These are: temperature, pH, concentration of the cleaning chemicals, contact time between the chemical solution and the membrane and operation conditions such as cross-ow velocity and pressure [131]. 4.5.1. Effect of cleaning solution pH The pH of cleaning solution affects the recovered permeate ux of fouled membrane with humic acid. Li et al. reported that cleaning efciency of EDTA depend critically on the pH of the solution as a result of deprotonation of functional groups. For example, at high pH 11 all carboxylic groups are deprotonated while at low pH only two or four carboxylic groups molecule are deprotonated. Ang et al. [132] reported that the inuence of solution pH on EDTA cleaning efciency had remarkable effect. They have shown that the cleaning efciency increased noticeably (from 25% to 44%) with increasing pH from 4.9 to 11.0 in the case of EDTA and all carboxylic functional groups of EDTA are deprotonated. The gel layer was broken down more easily at pH 11 compared to lower pH. On the other hand, solution pH has very little effect on SDS cleaning because of the pKa of the sulfate functional group of SDS is at 2.12, implying that SDS is already in its ionic form at ambient and elevated pH values [132]. It was concluded that the cleaning solution pH is a governing factor affecting chemical reaction between deposited foulants and EDTA, whereas the chemical reaction between deposited foulant and SDS is less inuenced by cleaning solution pH. 4.5.2. Effect of ionic strength of the cleaning solution High ionic strength of the NaCl solution alone or in a combination with chemical agent has been found to have an effect on the chemical cleaning system. Lee reported that the recovering ux of hydrophilic NOM fouled membrane was carried out effectively by high ionic strength of NaCl (0.1 M), whereas the citric acid, caustic, and SDS surfactant were not effective to clean the hydrophilic NOM. This could be due to reduced charges of the membrane surface and NOM acids leading to double layer compression and charge screening. Also, it is possible that Cl ions replaced the foulants of hydrophilic NOM associated with Horsetooth surface water (HT-SW) [104]. A combination of high ionic strength and caustic solution were more effective than the acid and SDS solutions for removal hydrophobic of NOM on

fouled membranes [104]. Lee [81] concluded that additional benets from the higher cross-ow velocity and the longer cleaning duration were negligible for high ionic strength cleaning. However, Tran et al. [81] concluded that the cleaning efciency was found to be superior at higher ionic strengths. 4.5.3. Effect of cleaning solution concentration Li et al, reported that at 1 mM (millimole) concentration of SDS had minor effect on the recovered ux, and a similar observe also was made by Lee [104]. These results indicate poor cleaning efciency by SDS at this concentration, when an ultraltration membrane was fouled by surface or ground water NOM. A SDS concentrations of about 10 mM or higher, cleaning completely recovered the initial cleaned membrane water ux. On the other hand, Bartlertt et al. [86] noticed that for UF ceramic and the sintered stainless steel membrane systems, increasing the concentration of NaOH above the optimum value do not improve the cleaning process but rather reduced the cleaning efciency. Li et al. [110] reported that increasing the concentration of surfactant such as CTAB from 0.1 to 0.5 wt.% had an adverse effect on cleaning, possibly due to adsorption of excess surfactants onto membrane surface during the cleaning. In their recent work, Chen et al. [133] noticed that a signicant factor affecting chemical cleaning of both UF and RO membranes is the high pH and the concentration of cleaning solution. Ang et al. [132], reported that cleaning efciency with EDTA and SDS cleaning increased with increasing cleaning agents concentration. They concluded that cleaning with SDS above CMC is critical factor for efcient SDS cleaning, allowing adequate chemical reaction between SDS and the foulant to break down an alginate gel network. Therefore, to better understand the cleaning mechanisms, the stoichiometry between the cleaning agent concentration and amount of foulant on the membrane surface must be considered. 4.5.4. Effect of cross-ow velocity (hydrodynamic shear) Hydraulic cleaning conditions such as cross-ow velocity play important role in membrane ux recovery. Lee [104] reported that in an experiment carried out using fouled UF membrane cleaned with caustic chemical cleaning at low and high cross-ow velocity, he found that the cleaning efciency increased at high cross-ow velocity compared to low crossow velocity, for 1 h duration time [104]. On the other hand, Bartlett et al. [86] noticed that an increase in cross-ow velocity while cleaning seems to have minimal effect on membrane ux fouled with whey protein, for 30 min. The main cause of minimal effect on ux recovery is attributed to the duration of the cleaning and the effect of duration time will be explained in the next section. The cleaning efciency increased with increasing cross-ow velocity when only the chemical reaction between the foulant and cleaning agents has to be high enough to produce a favorable reaction. Otherwise, an increase in cross-ow velocity which result an increase in the shear rate dose not enhance the mass transfer of foulant in the fouling layer to the bulk solution [132]. It can be concluded that the chemical reaction between of the cleaning agent with deposited foulants and associated mass transfer phenomena are quite important in membrane cleaning [132].

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

23

4.5.5. Cleaning duration Hydraulic cleaning conditions, such as cleaning duration, may play important role for ux recovery at actual plants. Longer cleaning time at lower velocity was found more effective in order to remove the NOM foulants from membrane surfaces [104]. In some applications, a cleaning time for 15 min is enough [134], while in other applications of membranes more time is needed for cleaning about 1 h [135] or even longer to reach their maximal cleaning effect in order to restore the membrane performance [24]. In a recent study, Li et al., [110] noticed that increasing cleaning time from 10 to 20 min for surfactant (SDS, CTAB) had greater power on cleaning efciency. Ang et al., [132], in the case of RO membrane reported that, the

cleaning efciency of EDTA increased dramatically when the cleaning time increased from 15 to 60 min. the cleaning efciency of SDS at low concentration with cleaning time of 15 or 60 min was not incorrigible. In other words, a longer contact time did not contribute to enhancing cleaning efciency. Unless there was a favorable chemical reaction between the foulant and cleaning agents in the fouling layer there will be little to reduce the foulantfoulant interaction [132]. The duration of chemical cleaning at different operation conditions has not been discussed to any depth in the literature. However, the few papers on cleaning of NF membrane have been published up to date suggest that shorter ltration cycles and hence more frequent but longer cleaning procedures are benecial as fouling layers

Fig. 15. Schematic representation for effective cleaning of organic-fouled RO membranes. Across-linked fouling layer is formed on the membrane surface in presence of calcium ions, which bind to organic foulants and form bridges between adjacent foulant molecules. During cleaning, the cleaning agent reacts with the foulants in the fouling layer yielding loosened foulants. These reaction products are removed from the fouling layer to the bulk solution through the hydrodynamics/mass transfer. Thus, Efcient cleaning can be achieved through the coupling between the chemical reaction and mass transfer along with the optimization of cleaning conditions responsible for the favourable chemical reaction and mass transfer [133].

24

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

become compact with time and are more difcult to remove [24,78,133,134]. 4.5.6. Cleaning frequency Typical cleaning frequencies for full-scale nanoltration systems treating ground water are 3 months to 2 years, with typical an average of 6 months [112,116]. For surface water the range may vary from once per week to once per 3 months. Cleaning frequencies for full-scale nanoltration system treating seawater are in the range of 46 months [24]. Cleaning frequency plays an important role in the lifetime of the membrane. Increasing cleaning frequency could lead deterioration of the cross-linking network of the membrane surface properties by making the chemical interaction between the cleaning solution molecules and surface functional groups of the membrane stronger [136]. Moreover, it has been noticed that cleaning processes often increases the ux even of the virgin membrane by making the pore more open or altering the membrane surface properties [96,97,99]. The cleaning interval depends on the foulant amount on the membrane surface which is indicated by the decrease in ux or increase in operating pressure. The amount of cleaning is therefore a function of cleaning duration and the frequency. A combination of these factors can give widely varying method as such need optimization. 4.5.7. Effect of temperature The temperatures of the chemical cleaning solution may also play an important role in chemical cleaning system. Generally increased temperature increases the cleaning efciency presumably by increasing transport processes and solubility of the material. However, the sensitivity of membrane materials usually prohibits the use of very high temperature. Generally, membrane manufacturers recommend chemical cleaning should be carried out at lower than 45 C. Bartlett et al., noticed that for the sintered stainless steel membrane and ceramic membrane, further increases in temperature above the optimum value reduced chemical cleaning efciency [86]. Li et al., [110] reported that preheating the cleaning solution up to about 42 C had a signicant impact on increased water ux recovery compared to 25 C where increased cleaning temperature was favorable for desorption of deposits from the membrane surface. Chen et al., [133] noticed that one signicant factor which can affect chemical cleaning of both UF and RO membranes is the cleaning temperature. Ang et al., [132] shown that, the cleaning efciency increased dramatically with increasing temperature from 20 to 40 C in the case of RO membrane. Both the rate of the chemical reaction of EDTA with deposited foulants and the diffusive transport of foulant from the fouling layer to the bulk solution increased as temperature was increased. 4.5.8. Effect of pressure Operating pressure is often related to the hydrodynamics of the system and should be maintained very carefully. Cleaning is to be carried out a minimum pressure in order not onto force the foulant layer onto the surface making it more adhesive. On the other hand, Bartlett et al., [86] reported that these results

demonstrate that any applied pressure during cleaning will not result in maximum cleaning efciency, while there is surface deposits present. The same trend was observed by Sayed [134]. This topic was not covered well in literature and further experiments are needed to nd out the relationship of the operating pressure on effectiveness of cleaning. When reviewing the above major factors of NF membrane cleaning efciency, it is very important to understand that membrane cleaning involves both chemical and physical interactions. The relationship between the coupled effect of dissolution of material from the membrane surface caused by chemical reactions between the cleaning agents and the foulants and the subsequent mass transfer of the dissolved or loosened foulant materials into the bulk solution (Fig. 15) [132]. The effectiveness of cleaning in terms of chemical reactivity depends on the type of cleaning solution, cleaning solution pH, cleaning chemicals concentration and ionic strength while mass transfer of the foulant from the fouling layer to the bulk solution was mainly controlled by the cross-ow velocity and pressure, whereas cleaning time and temperature affects both the mass transfer and chemical reaction. 5. Discussion and conclusion This review is focused on the types of membrane fouling processes that lead to deterioration of the plant performance and there subsequent recovery using membrane cleaning processes. Different types of inorganic fouling and organic fouling have been addressed in this study. Fouling problems lead to higher operation costs, higher energy demand, reduce life time of membrane and increasing cleaning frequency. Our work also reveals that membrane chemical cleaning methods are widely implemented in order to regenerate the membrane performance, even though some cleaning methods have potential limitations. The success of chemical cleaning methods depends on many factors such as nature of the foulant type of cleaning agents, temperature, pH, concentration of the cleaning chemicals, contact time between the chemical solution and the membrane and operation conditions such as cross-ow velocity and pressure. These factors affect the outcome of the cleaning procedure and therefore need thorough investigation in order to establish the optimum cleaning system. Conventional assessment of cleaning by ux measurement has been used in the last decades in order to optimize and evaluate the cleaning procedures. At present there are several modern surface analysis techniques that can assist precisely and rapidly in optimizing the cleaning processes. Fouling is a complex processes largely governed by empirical rules due to poor understanding of processes involved. The main drawback is the poor prediction of fouling rate in the membrane system as a result of limited understanding of the organic fouling and the lack of knowledge concerning the NF module operating conditions at which the crystallization or the scale precipitation associated with inorganic fouling occur. The use of modern surface analysis is required to improve our understanding on mechanisms of fouling or to closely understand how these foulant materials get deposited on membrane surfaces. Nevertheless, it is fundamental to obtain the molecular basis of these

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428

25

processes. More to the point, the use of AFM and other modern technique can assist in reaching the target of understanding fouling phenomena. The direct force measurement between microscopic surfaces as function of distance in controlled environment became possible with development of the surface force apparatus [137]. The fabrication of AFM colloidal probe will undoubtedly be used as one of the powerful technique for providing a deeper understanding of membranefouling interaction process. It also can be used to directly measure the interaction between surfaces of different materials in various aqueous solutions and that can assist particularly in understanding and addressing the complexity of the fouling materials attachment, removal and movement in and around the membrane surfaces [122]. In general, chemical cleaning is conducted mainly to remove the fraction of NOM foulant or to remove the scale formation from NF membrane surfaces or to clean the NF membrane surfaces from both of them in order to regenerate the membrane performance. There are two types of force that can be applied during cleaning, the adhesive force that keep the foulant on the membrane surface and a hydrodynamic shearing force required in removing the foulant from the membrane surface to the bulk. However, in reality the dynamic process associated with movement of foulant material in and around the membrane surfaces is critical during the cleaning processes. It is difcult to obtain clear strategies in both membrane design and avoidance of fouling. Nevertheless, there are several ways to avoid the fouling of membranes and can be classied into three groups: (i), changing operating conditions, (ii), modication of the membrane and (iii), modication of the feed by adding antifoulants prior to ltration system (pretreatment) [66]. Until now these avoidance strategies do not stop the fouling on surface of the membrane, however, they reduce the fouling rate on membrane surfaces and so reducing the frequency and duration of cleaning, both have substantial effect on membrane productivity and longevity. There are several chemical cleaning agents used directly to remove foulant material and an understanding of the roles in chemical has been developed. The broad mechanism of cleaning agents have been inferred from the coupling effect of these chemical reaction between the cleaning agents and the foulant and the mass transfer of the loosened foulant to the bulk and this has lead to improve formulation. Indeed formulations of cleaning agents have been developed for specic foulants. The adsorptiondesorption phenomena are very important in cleaning processes. The surface charge will not only be dependent on the coverage of foulants but also on adsorption phenomena associated the use of cleaning agents. The adsorption phenomena of the cleaning agents are varying and depend on interaction between the membrane and the cleaning agent concentration. The precise chemistry and morphology of the active membrane layer also signicantly affect the adsorption phenomena. The active layer of the membrane may consist of several sub-layers to give a complex composite structure that can signicantly alter rejection and adsorption properties of the surface. Indeed, the cleaning agent may change the membrane surface functionally by allowing chemical changer or by adsorption onto the membrane surface so cause permanent modication

Fig. 16. Venn diagram of factors inuencing operational membrane processes.

and changed membrane surface properties as well as membrane surface functionality [95]. Generally, modern surface analysis techniques can assist not only in selecting the membrane materials with correct surface properties during the membrane casting and grafting process, the aim being to avoid and/or reduce the foulant materials adhesion on the membrane surface. A better understanding membrane functionality in relation to adsorptiondesorption phenomena on the membrane surface will also be obtained. Modern instrumentation will lead to a systematic more predictable view of the fouling and cleaning processes. This understanding is fundamental to providing novel method and materials for membrane manufacture, feed water conditioning and optimal operational systems. Manipulating membrane surface characteristics with chemical and physical aspect of feed water and operation conditions can directly lead to optimized systems with minimized fouling and (Fig. 16) effective cleaning systems. Extensive research work is certainly needed to investigate and explore new ideas and techniques in the eld of membrane cleaning and restoration. It is in the interest of the NF membrane industry that the importance of membrane cleaning be recognized and be addressed with more time and effort being devoted to understand fouling mechanisms and develop more feasible cost-effective, cleaning and restoration procedures for each type of fouling. Apart from this, new membranes which avoid fouling have to be developed in order to reduce cleaning processes and their costs. Fundamental parameters that affect the condition of feed water treatment and operation of the membrane system are very important and have to be given more time subsequently to prevent or reduce fouling and cleaning processes. Physical and physio-chemical cleaning methods have not been given much needed emphasis in order to evaluate the effectiveness of these methods. These methods could be economically attractive but need more investigation. References
[1] A. Al-Rewiali, et al., Performance Evaluation of SWCC SWRO Plants, Research centre and development, Jubail, 2000, pp. 130. [2] T.G. Temperely, The Coming of Age of Desalination. in Proceeding of the IDA World Congress on Desalination and Water Sciences. 1995. United Arab Emirates, Abu Dhabi: Abu Dhabi Publishing Co.

26

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428 [30] S. Lee, J. Kim, C.-H. Lee, Analysis of CaSO4 scale formation mechanism in various nanoltration modules, J. Membr. Sci. 163 (1) (1999) 6374. [31] J. Gilron, D. Hasson, Calcium sulphate fouling of reverse osmosis membranes: ux decline mechanism, Chem. Eng. Sci. 42 (10) (1987) 23512360. [32] M. Alborzfar, G. Jonsson, C. GrOn, Removal of natural organic matter from two types of humic ground waters by nanoltration, Water Res. 32 (10) (1998) 29832994. [33] S. Lee, C.-H. Lee, Effect of operating conditions on CaSO4 scale formation mechanism in nanoltration for water softening, Water Res. 34 (15) (2000) 38543866. [34] Z. Amjad, J. Hooley, Inuence of polyelectrolytes on the crystal growth of calcium sulfate dihydrate, J. Colloid Interface Sci. 111 (2) (1986) 496503. [35] D.G. Troup, J.A. Richardson, Scale nucleation on a heat transfer surface and its prevention, Chem. Eng. Commun. 2 (1978) 167180. [36] P.G. Klepetsanis, P.G. Koutsoukos, Precipitation of calcium sulfate dihydrate at constant calcium activity, J. Cryst. Growth 98 (3) (1989) 480 486. [37] X.-L. Wang, C. Zhang, P. Ouyang, The possibility of separating saccharides from a NaCl solution by using nanoltration in dialtration mode, J. Membr. Sci. 204 (1/2) (2002) 271281. [38] D.R. Kasper, Pre-and posttreatment processes for membrane water treatment system membrane technology conference proceeding, Baltimor, 1993. [39] T.H. Chong, S. Yuen, R. Sheikholeslami, Composite Fouling in Seawater Desalination Precipitation of Calcium Sulfate and Calcium Carbonate, School of chemical engineering and industrial chemistry, Sydney, 2004, pp. 110. [40] R. Sheikholeslami, Mixed saltsscaling limits and propensity, Desalination 154 (2) (2003) 117127. [41] P. Dydo, M. Turek, J. Ciba, Scaling analysis of nanoltration systems fed with saturated calcium sulfate solutions in the presence of carbonate ions, Desalination 159 (3) (2003) 245251. [42] T.H. Chong, R. Sheikholeslami, Thermodynamics and kinetics for mixed calcium carbonate and calcium sulfate precipitation, Chem. Eng. Sci. 56 (18) (2001) 53915400. [43] S.K. Hamdona, R.B. Nessim, S.M. Hamza, Spontaneous precipitation of calcium sulphate dihydrate in the presence of some metal ions, Desalination 94 (1) (1993) 6980. [44] D.L. Kronmiller, What every reverse osmosis water system manager should know, Desalination 98 (13) (1994) 401411. [45] A. Al-Amoudi, R.W. Lovitt, Natural organic matter (NOM) and scaling fouling in NF membranes and its factors: a review, Desalination, in press. [46] P. Fu, H. Ruize, Pilot plant study on ground water natural organic removved by low pressure, Desalination 102 (1995) 4756. [47] A.R. Roudman, F.A. DiGiano, Surface energy of experimental and commercial nanoltration membranes: effects of wetting and natural organic matter fouling, J. Membr. Sci. 175 (1) (2000) 6173. [48] M. Manttari, et al., Fouling effects of polysaccharides and humic acid in nanoltration, J. Membr. Sci. 165 (1) (2000) 117. [49] B. Van der Bruggen, et al., Inuence of molecular size, polarity and charge on the retention of organic molecules by nanoltration, J. Membr. Sci. 156 (1) (1999) 2941. [50] A.I. Schafer, A.G. Fane, T. Waite, Nanoltration of natural organic matter: Removal, fouling and the inuence of multivalent ions, Desalination 118 (1998) 109122. [51] M. Elimelech, et al., Role of membrane surface morphology in colloidal fouling of cellulose acetate and composite aromatic polyamide reverse osmosis membranes, J. Membr. Sci. 127 (1997) 101109. [52] C.R. Bouchard, et al., Study of humic acid adsorption on nanoltration membranes by contact angle measurements, Can. J. Chem. Eng. 75 (2) (1997) 339345. [53] K. Ghosh, M. Schnitzer, Macromolecular structures of humic substances, Soil Sci. 129 (5) (1980) 266276. [54] A.I. Schafer, et al., Natural organic matter removal by nanoltration: effects of solution chemistry on retention of low molar mass acids versus bulk organic matter, J. Membr. Sci. 242 (1/2) (2004) 7385.

[3] T.A. Dabbagh, The Role of Desalination in Sustaining Economic Growth in the Gulf, Proceeding of the IDA World Congress on Desalination and Water Sciences,. 1995. United Arab Emirates, Abu Dhabi: Abu Dhabi Publishing Co. [4] R.J. Petersen, Composite reverse osmosis and nanoltration membranes, J. Membr. Sci. 83 (1) (1993) 81150. [5] Manual, M.T., Manufacturers Technical Manual. [6] M. Barger, R.P. Carnahan, Fouling prediction in reverse osmosis processes, Desalination 83 (13) (1991) 333. [7] D.E. Potts, R.C. Ahlert, S.S. Wang, A critical review of fouling of reverse osmosis membranes, Desalination 36 (3) (1981) 235264. [8] P. Dydo, et al., The nucleation kinetic aspects of gypsum nanoltration membrane scaling, Desalination 164 (1) (2004) 4152. [9] W.R. Bowen, A.W. Mohammad, N. Hilal, Characterisation of nanoltration membranes for predictive purposesuse of salts, uncharged solutes and atomic force microscopy, J. Membr. Sci. 126 (1) (1997) 91105. [10] A.I. Schafer, A.G. Fane, T.D. Waite, NanoltrationPrinciple and Applications, Elsevier Advanced Technology, 2005, pp. 1543. [11] R.W. Field, et al., Critical ux concept for microltration fouling, J. Membr. Sci. 100 (1955) 259272. [12] P. Bacchin, P. Aimar, R.W. Field, Critical and sustainable uxes: theory, experiments and applications, J. Membr. Sci. 281 (2006) 4269. [13] E. Van Houtte, et al., Treating different types of raw water with micro- and ultraltration for further desalination using reverse osmosis, Desalination 117 (13) (1998) 4960. [14] G. Tchobanoglous, et al., Ultraltration as an advanced tertiary treatment process for municipal wastewater, Desalination 119 (13) (1998) 315321. [15] J.J. Porter, Recovery of polyvinyl alcohol and hot water from the textile wastewater using thermally stable membranes, J. Membr. Sci. 151 (1) (1998) 4553. [16] K. Hagen, Removal of particles, bacteria and parasites with ultraltration for drinking water treatment, Desalination 119 (13) (1998) 8591. [17] L. Vera, et al., Cross-ow microltration of biologically treated wastewater, Desalination 114 (1) (1997) 6575. [18] S. Hong, M. Elimelech, Chemical and physical aspects of natural organic matter (NOM) fouling of nanoltration membranes, J. Membr. Sci. 132 (1997) 159181. [19] T.L. Champlin, Using circulation tests to model natural organic matter adsorption and particle deposition by spiral-wound nanoltration membrane elements, Desalination 131 (13) (2000) 105115. [20] T.V. Knyazkova, A.A. Maynarovich, Recognition of membrane fouling: testing of theoretical approaches with data on NF of salt solutions containing a low molecular weight surfactant as a foulant, Desalination 126 (13) (1999) 163169. [21] H. Zhu, M. Nystrom, Cleaning results characterized by ux, streaming potential and FTIR measurements, Colloids Surf. A: Physicochem. Eng. Aspects 138 (2/3) (1998) 309321. [22] H. Vrouwenvelder, et al., Biofouling of membranes for drinking water production, Desalination 118 (1998) 157166. [23] J. Vrouwenvelder, et al., Tools for fouling diagnosis of NF and RO membranes and assessment of the fouling potential of feed water, Desalination 157 (2003) 361365. [24] A.S. Al-Amoudi, A.M. Farooque, Performance, restoration and autopsy of NF membranes used in seawater pretreatment, Desalination 178 (2005) 261271. [25] B. Van der Bruggen, L. Braeken, C. Vandecasteele, Flux decline in nanoltration due to adsorption of organic compounds, Sep. Purif. Technol. 29 (1) (2002) 2331. [26] M.R. Wiesner, et al., Committee Report: membrane processes in potable water, J. AWWA (1992) 5964. [27] D. Hasson, A. Drak, R. Semiat, Inception of CaSO4 scaling on RO membranes at various water recovery levels, Desalination 139 (13) (2001) 7381. [28] K.A. Faller, Reverse osmosis and nanoltration. in AWWA. 1999. Denver. [29] A.G. Pervov, Scale formation prognosis and cleaning procedure schedules in reverse osmosis systems operation, Desalination 83 (13) (1991) 77118.

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428 [55] A.E. Childress, S.S. Deshmukh, Effect of humic substances and anionic surfactants on the surface charge and performance of reverse osmosis membranes, Desalination 118 (13) (1998) 167174. [56] A.E. Childress, M. Elimelech, Effect of solution chemistry on the surface charge of polymeric reverse osmosis and nanoltration membranes, J. Membr. Sci. 119 (2) (1996) 253268. [57] M. Elimelech, W.H. Chen, J.J. Waypa, Measuring the zeta (electrokinetic) potential of reverse osmosis membranes by a streaming potential analyzer, Desalination 95 (3) (1994) 269286. [58] E. Suess, Interaction of organic compounds with calcium-carbonate. 2. Organo-carbonate association in recent sediments, Geochimica Et Cosmochimica Acta 37 (11) (1973) 24352447. [59] W. Yuan, A.L. Zydney, Humic acid fouling during microltration, J. Membr. Sci. 157 (1) (1999) 112. [60] W. Agui, et al., Removal of dissolved humic substances from water with a reverse-osmosis membrane, Sci. Total Environ. 118 (1992) 543550. [61] C. Bellona, et al., Factors affecting the rejection of organic solutes during NF/RO treatmenta literature review, Water Res. 38 (2004) 27952809. [62] B. Van der Bruggen, et al., The use of nanoltration for the removal of pesticides from ground water: an evaluation, Water Sci. Technol. 1 (2001) 99106. [63] M.R. Wiesner, S. Chellam, The promise of membrane technologies, Environ. Sci. Technol. 33 (17) (1999) 360A. [64] L.F. Song, M. Elimelech, Particle deposition onto a permeable surface in laminar-ow, J. Colloid Interface Sci. 173 (1) (1995) 165180. [65] AWWA membrane technology research committee, C.r., Membrane processes in potable water treatment. J. AWWA, 1992. 84 (1): p. 5967. [66] I.l. Roux, et al., Use of chitosan as an antifouling agent in a membrane bioreactor, J. Membr. Sci. 248 (1/2) (2005) 127136. [67] P.H. Given, Humic Substances in Soil, Sediment and Water, WileyInterscience, New York, 1985, p. 692. [68] J.A. Nilson, F.A. DiGiano, Inuence of NOM composition on nanoltration, J. Am. Water Works Assoc. 88 (5) (1996) 5366. [69] C. Jucker, M.M. Clark, Adsorption of aquatic humic substances on hydrophobic ultraltration membranes, J. Membr. Sci. 97 (1994) 3752. [70] L. Braeken, et al., Inuence of hydrophobicity on retention in nanoltration of aqueous solutions containing organic compounds, J. Membr. Sci. 252 (12) (2005) 195203. [71] M.S. Caceci, A. Billon, Evidence for large organic scatterers (50200 nm diameter) in humic-acid samples, Organic Geochem. 15 (3) (1990) 335. [72] W. Richard Bowen, et al., Atomic force microscope studies of membranes: force measurement and imaging in electrolyte solutions, J. Membr. Sci. 126 (1) (1997) 7789. [73] T. Thorsen, Concentration polarisation by natural organic matter (NOM) in NF and UF, J. Membr. Sci. 233 (1/2) (2004) 7991. [74] H.M. lappin-Scott, J.W. Costerton, Bactierial biolms and surface fouling, Biofouling 1 (1989) 323342. [75] H.C. Flemming, et al., Biofoulingthe Achilles heel of membrane processes, Desalination 113 (2/3) (1997) 215225. [76] H.F. Ridgway, H.C. Flemming, Membrane Biofouling in Water Treatment Membrane Processes, McGraw Hill, New York, 1996. [77] N. Hilal, et al., A comprehensive review of nanoltration membranes: treatment, pretreatment, modelling, and atomic force microscopy, Desalination 170 (3) (2004) 281308. [78] A. Farooque, A. al-Amoudi, A.M. Hassan, Chemical cleaning experiments for performance restoration of NF membranes operated on seawater feed, in: IDA Conference, Bahrain, 2002. [79] C. Shankararaman, et al., Effect of pretreatment on surface water nanoltration, J. AWWA 10 (1997) 7789. [80] S. Bertrand, I. Lemaitre, E. Wittman, Performance of a nanoltration plant on hard and highly sulphated water during two years of operation, Desalination 113 (1997) 277281. [81] M.H. Tran-Ha, D.E. Wiley, The relationship between membrane cleaning efciency and water quality, J. Membr. Sci. 145 (1998) 99110. [82] J.S. Taylor, E.P. Jacobs, Reverse osmosis and nanoltration, chapter 9 in water treatment membrane processes, ed. P.E.O.a.M.R.W.e. J. mallevialle, 1996, McGraw Hill.

27

[83] H. Huiting, J. Kappelhof, T. Bosklopper, Operartion of NF/RO plants from reactive to proactive, Desalination 139 (2001) 183189. [84] A. Maartens, P. Swart, E.P. Jacobs, Feed-water pretreatment: methods to reduce membrane fouling by natural organic matter, J. Membr. Sci. 163 (1) (1999) 5162. [85] A.H. Hassan, et al., Optimization of SWRO pre-treatmentPart I-SDI measurement. in IDA World congress, 1995, Abu Dhabi. [86] M. Bartlett, M.R. Birdl, J.A. howell, An experimental study for the development of qualitative membrane cleaning model, J. Membr. Sci. 105 (1995) 147157. [87] V. Mavrov, et al., Desalination of surface water to industrial water with lower impact on the environment. Part 2. Improved feed water pretreatment, Desalination 110 (1/2) (1997) 6573. [88] A.M. Hassan, et al., A new approach to membrane and thermal seawater desalination processes using nanoltration membranes (Part 1), Desalination 118 (13) (1998) 3551. [89] B. Van der Bruggen, et al., Direct nanoltration of surface water using capillary membranes: comparison with at sheet membranes, Sep. Purif. Technol. 31 (2) (2003) 193201. [90] D. Chen, L.K. Weavers, H.W. Walker, Ultrasonic control of ceramic membrane fouling: effect of particle characteristics, Water Res. 40 (4) (2006) 840850. [91] N.R.G. Walton, Reverse osmosis pretreatment-injecting a little chemical control and management, Desalination 82 (1991) 281300. [92] G. Tragardh, 71 (1989) 325, Membrane cleaning Desalination, 1989. 71: p. 325334. [93] S. Ebrahim, Cleaning and regeneration of membranes in desalination and waste water applications, Desalination 96 (1994) 225238. [94] R.A. Bergman, Membrane softening versus lime softening in Florida: a cost comparison update, Desalination 102 (13) (1995) 1124. [95] A. Al-Amoudi, et al., Cleaning results of new and fouled nanoltration membrane characterized by zeta potential and permeability, Sep. Purif. Technol. 54 (2) (2007) 234240. [96] A. Wies, M.R. Bird, M. Nystrom, The chemical cleaning of polymeric UF membranes fouled with spent sulphite liquor over multiple operational cycles, J. Membr. Sci. 216 (2003) 6779. [97] K. KoSutic, B. Kunst, RO and NF membrane fouling and cleaning and pore size distribution variations, Desalination 150 (2002) 113120. [98] M. Nystrom, H. Zhu, Characterization of cleaning results using combined ux and streaming potential methods, J. Membr. Sci. 131 (1/2) (1997) 195205. [99] S.I. Graham, R.L. Reitz, C.E. Hickman, Improving reverse osmosis performance through periodic cleaning, Desalination 74 (1989) 113 124. [100] M. Abdul-Kareem Al-So, Seawater desalinationSWCC experience and vision, Desalination 135 (13) (2001) 121139. [101] R. Osterberg, L. Szajdak, K. Mortensen, Temperature-dependent restructuring of fractal humic acidsa proton-dependent process, Environ. Int. 20 (1) (1994) 7780. [102] I. Mooch, A 21st Century study of global seawater reverse osmosis operating and capital costs, in: IDA Conference, Singapore, 2003. [103] A.M. Hassan, et al., Performance Evaluation of SWCC-SWRO Plants, Research Center, Jubail, 2003, pp. 159. [104] H. Lee, et al., Cleaning strategies for ux recovery of an ultraltration membrane fouled by natural organic matter, Pergamon 35 (2001) 33013308. [105] G. Morel, et al., Surfactant modied ultraltration of nitrate ions removal, J. Membr. Sci. 134 (1997) 4757. [106] Du-Pont Manual Hand Book. [107] H. Seungkwan, M. Elimelech, Chemical and physical aspects of natural organic matter (NOM) fouling of nanoltration membranes, J. Membr. Sci. 132 (1997) 159181. [108] J. Cho, et al., Characterization of clean and natural organic matter (NOM) fouled NF and UF membranes and foulants characterization, Desalination 188 (1998) 101108. [109] R. Liikanan, J. Yli-Kuivila, R. Laukkanen, Efciency of various chemical cleanings for nanoltration membrane fouled by conventionally-treated surface water, J. Membr. Sci. 195 (1995) 265276.

28

A. Al-Amoudi, R.W. Lovitt / Journal of Membrane Science 303 (2007) 428 [125] E.M. Vrijenhoek, S. Hong, M. Elimelech, Inuence of membrane surface properties on initial rate of colloidal fouling of reverse osmosis and nanoltration membranes, J. Membr. Sci. 188 (1) (2001) 115 128. [126] I. Larson, et al., Direct force measurements between TiO2 Surfaces, J. Am. Chem. Soc. 115 (1993) 1188511890. [127] W.R. Bowen, et al., Atomic force microscope studies of membranes, in: T.S. Sorensen (Ed.), Surface Chemistry and Electochemistry of Membrane Surfaces, Surfactant Science Series, vol. Chapter 1, Marcel Dekker Inc., 1999, pp. 137. [128] W. Richard Bowen, T.A. Doneva, Atomic force microscopy studies of nanoltration membranes: surface morphology, pore size distribution and adhesion, Desalination 129 (2) (2000) 163172. [129] K. Kontturi, M. Vouristo, Adsorption of globular proteins on polymeric microltration membranes, Desalination 104 (1996) 99105. [130] M.D. Afonso, G. Hagmeyer, R. Gimbel, Streaming potential measurements to assess the variation of nanoltration membranes surface charge with the concentration of salt solutions, Sep. Purif. Technol. 22/23 (2001) 529541. [131] T. Mohammadi, S.S. Madaeni, M.K. Moghadam, Investigation of membrane fouling, Desalination 153160 (2003) 155. [132] W.S. Ang, S. Lee, M. Elimelech, Chemical hysical aspects of cleaning of organic-fouled reverse osmosis membranes, J. Membr. Sci. 272 (1/2) (2006) 198210. [133] J.P. Chen, S.L. Kim, Y.P. Ting, Optimization of membrane physical and chemical cleaning by a statistically designed approach, J. Membr. Sci. 219 (12) (2003) 2745. [134] S.K. Sayed Razavi, J.L. Harries, F. Sherkate, Fouling and cleaning of membranes in the ultraltration of the aqueous extract of soy our, J. Membr. Sci. 114 (1) (1996) 93104. [135] M.J. Munoz-Aguado, D.E. Wiley, A.G. Fane, Enzymatic and detergent cleaning of a polysulfone ultra-ltration membrane fouled with BSA and Whey, J. Membr. Sci. 117 (1996) 175187. [136] M. Chpman, Bench-scale testing of surfactant modied reverse osmosis/nanoltration membranes, J. Pellegrino A Zydney 115 (1998) 15 32. [137] J.N. Israelachvili, R.K. Tandon, L.R. White, Measurement of forces between two mica surfaces in aqueous polyethylene oxide solution, J. Colloid Interface Sci. 78 (2) (1980) 430443.

[110] L. Xiufen, et al., Chemical cleaning of PS ultralters fouled by the fermentation broth of glutamic acid, Sep. Purif. Technol. 42 (2005) 181187. [111] J. Warczok, et al., Concentaion of apple and paer juices by nanoltration at low pressures, J. Food Eng. 63 (2004) 6370. [112] A.M. Jacques, et al., Integrated multi-objective membrane systems for surface water treatment: pre-treatment of nanoltration by riverbank ltration and conventional ground water treatment, Desalination 118 (1998) 239248. [113] W. Song, et al., Nanoltration of natural organic matter with H2 O2 /UV pretreatment: fouling mitigation and membrane surface characterization, J. Membr. Sci. 241 (1) (2004) 143160. [114] G.S. Murthy, et al., Concentration of xylose reaction liquor by nanoltration for the production of xylitol sugar alcohol, Sep. Purif. Technol. 44 (2005) 205211. [115] H.A. Fremont, et al., Membrane cleaning process and U.S.P., 1988. [116] Q. Li, M. Elimelech, Organic fouling chemical cleaning of nanoltration membranes: measurements and mechanisms, Environ. Sci. Technol. 38 (17) (2004) 46834693. [117] A.E. Childress, M. Elimelech, Relating nanoltration membrane performance to membrane charge (electrokinetic) characteristics, Environ. Sci. Technol. 34 (17) (2000) 37103716. [118] R. Liikanen, J. Yli-Kuivila, R. Laukkanen, Efciency of various chemical cleanings for nanoltration membrane fouled by conventionally-treated surface water, J. Membr. Sci. 195 (2) (2002) 265276. [119] G. Daun, et al., Efciency of cleaning agents for an inorganic membrane after milk ultraltration, J. Dairy Res. 59 (1) (1992) 2938. [120] M.A. Arguello, et al., Enzymatic cleaning of inorganic ultraltration membranes used for whey protein fractionation, J. Membr. Sci. 216 (1/2) (2003) 121134. [121] K.-J. Kim, A.G. Fane, Performance evaluation of surface hydrophilized novel ultraltration membranes using aqueous proteins, J. Membr. Sci. 99 (2) (1995) 149162. [122] W.R. Bowen, A. Teodora, G. Austin, The use of atomic force microscopy to quantify membrane surface electrical properties, Colloids Surf. 201 (2002) 7383. [123] N. Hilal, W.R. Bowen, Atomic force microscope study of the rejection of colloids by membrane pores, Desalination 150 (3) (2002) 289295. [124] W.R. Bowen, T.A. Doneva, Atomic Force Microscopy of Membrane, Madison Avenue, New York, 2002.