Biofouling phenomena in membrane distillation: mechanisms and mitigation strategies

Membrane distillation (MD) is envisaged as a cost-e ﬀ ective water desalination technology. When operated by low-grade energy, MD surpasses the cost challenges faced by other water desalination technologies. Although MD operates under conditions that minimize the survival of bio ﬁ lm causing microorganisms, their development and succession is being increasingly reported. This is believed to be caused by the presence of halophiles and thermophiles in the feed solution, inducing signi ﬁ cant e ﬃ ciency losses. Therefore, biofouling mitigation remains crucial. This study reports current developments toward MD biofouling and mitigation strategies. Also, e ﬀ ects of membrane biofouling on process performance are brie ﬂ y highlighted. This provides an in-depth understanding of measures required to minimize biofouling of MD systems.


Introduction
The unprecedented rise in climate change and an increase in population growth have resulted in signicant global water shortages. 1,2Over the last century, water consumption has increased by approximately 600%, largely due to urbanization. 3urthermore, based on scholarly reports, nearly 1.8 billion people are projected to live in water stressed areas by 2025. 4 These alarming statistics and the rise in demand for potable water have stimulated the need to explore freshwater supply alternatives.Currently, scientists and engineers are developing sustainable solutions through research to address existing water shortages.
Membrane-based separation technologies have emerged as promising approaches to supply water to coastal areas.These include reverse osmosis (RO), forward osmosis (FO), nano-ltration (NF) and membrane distillation (MD). 5Although RO is the leading water desalination technology, its high energy consumption and extreme operating pressure increases operational costs.These costs are signicantly high compared to the volume of water produced.Therefore, alternative membrane technologies of low-cost should be considered.MD is envisaged as a cost-effective membrane technology, not only because of its ability to exploit solar thermal energy but its capacity to operate under moderate conditions, requiring inexpensive equipment. 6The driving force behind MD is the vapor pressure difference existing across the membrane, induced by differences in temperature (Fig. 1a). 7As illustrated in Fig. 1b, both mass (vapor) and heat are transferred through the membrane.The mechanism of mass transport is well explained by Poisseuille ow, molecular diffusion, and Knudsen diffusion, conditional to the existence of trapped air in the pores and the membrane pore size. 8Mass transport depends on the type of membrane used, its properties, the concentration of the feed solution, and the mass transfer coefficient.Further advantages of MD include high salt rejection ($99%), minimal sensitivity to salinity of the feed solution, low hydrostatic temperature and pressure requirements and the feasible utilization of renewable energy and low-grade/residual energy. 9chwantes et al. (2013) 10 used the waste heat generated from a diesel power station to drive a pilot-scale MD system.The waste heat supplied feed temperatures of 70-80 °C, enabling water production of 4 m 3 per day.Although MD displays attractive features, its industrial application has been hindered by low performance module design, production of low water uxes (compared to current technologies), and membrane fouling. 11,12embrane fouling refers to the deposition and growth of inorganic, organic, colloidal and biological substances on the membrane surface. 14Inorganic fouling involves the deposition of mineral salts on the membrane surface following precipitation whereas organic fouling refers to the deposition of organic matter on the membrane surface. 16Membrane fouling is reported to increase membrane wetting, causing a reduction in distillate quality. 17Moreover, membrane fouling introduces regular membrane cleaning and frequent membrane replacements, further increasing costs of water production. 18Current reported studies mostly focus on inorganic and organic fouling, with limited information on the dynamics and succession of biofouling. 19By denition, biofouling refers to the deposition and accumulation of microorganisms on the surface of the membrane and within its pores, leading to the formation of a biolm. 20The biolm consists of microbial cells and extracellular polymeric substances (EPS), covering the surface of the membrane.Biofouling of MD systems was overlooked with the perceptions that its operating conditions would minimize bio-lm development.However, an in-depth study of fouling layers has revealed the increased presence of biolms. 21][26] Krivorot et al., (2011) 24 studied biofouling occurrence in MD, using coastal seawater as the feed.Reportedly, microbial attachment increased quantitatively as a function of time, thus affecting process performance.A 34% reduction in water ux was recorded, markedly due to membrane pore blockage caused by biolm formation.Furthermore, ux decline was caused by the growth of crystal salts on the membrane surface.Another phenomenon contributing to biofouling in MD is temperature polarization (TP), referring to the difference in temperature between the bulk feed stream and the membrane surface. 6The TP stimulates microbial growth at the cooler membrane interface. 24Furthermore, TP decreases the driving force in MD. 27 A detailed review of TP and its holistic effects on MD is reported elsewhere. 28Membrane wetting and a subsequent reduction in salt rejection induced by microbial accumulates on the membrane surface was reported by Bogler and Bar-Zeev (2018). 22Membrane wetting was clearly observed at feed temperature of 65 °C using optical coherence tomography (OCT) and facilitated penetration of bacteria and endospores through the membrane towards the distillate.Microbial deposition was dependent on the composition of feed solution and nutrient availability. 26To establish high performing MD with excellent resistance to biofouling, elucidation of biofouling mechanism and its effects is essential.A recent review presenting biolm development in MD systems was reported. 6owever, specic illustration and collective reporting of microorganisms causing biofouling in MD was not highlighted.Therefore, the current review provides detailed insights of MD biofouling with particular focus on the survival of microbes under harsh MD operating conditions.Additionally, the mechanism of biofouling, its formation, causes, effects, and prevention for efficient MD operation are discussed.

Biofilm formation
Biolm formation follows a pathway of successive steps including; (1) conditioning lm formation with migration and  [13][14][15] adhesion of bacterial cells to the membrane surface, (2) EPS secretion, growth and maturation of bacterial cells, and (3) proliferation and cell detachment for colonization of new areas. 30For control and efficient mitigation of biofouling, insights into its mechanism must be expounded and understood.

Conditioning lm formation and surface attachment
A conditioning lm composed of macromolecules, organic matter, proteins, amino acids, and nucleic acids initially covers the surface of the membrane.Alongside, dead bacteria and soluble microbial products (SMP's) further adhere on the membrane.The presence of conditioning lm alters the membrane surface, promoting better adhesion of cells.Planktonic (free-oating) cells migrate from the bulk solution to the membrane surface through the Brownian motion, where movement is induced by the collision of particles. 15However, attachment is not permanent, since planktonic cells preferentially select sites for adhesion through cell locomotion using agella and pili. 31Cells exhibiting better motility and microbial activity attach rst on membrane surfaces. 32Following attachment, cells undergo proliferation, where constant multiplication and division take place.Furthermore, transparent exopolymer substances (TEP) and gel-like polysaccharides (aggregates) adsorb on the membrane surface to inuence fouling. 33The TEPs generate a viscous surface, thus providing a transport for microbial cells to the membrane surface.However, due to their complex chemical composition, specic contribution of TEPs to fouling is not known.According to Zhang et al. 2018, 34 interaction between cells and the membrane (cell-to-surface) occur largely through electrostatic and hydrophobic-hydrophobic processes.Although positively charged bacterial cells adhere readily on membrane surfaces, negatively charged bacterial cells tend to adhere less, especially in cases where the membrane surface is predominantly negatively charged and the ionic strength of the feed solution is low. 35olutions of low ionic strength contain less ions to counter-act the electronegativity of bacterial cells, resulting in increased repulsion between the membrane and the cell.However, at high ionic strength, cells adhere irreversibly on membrane surfaces.Abu-Lail and Camesano (2003) 36 evaluated the bio-adhesion of Pseudomonas putida KT2442 and noted low adhesion rates at low ionic strength.This was attributed to the high energy barrier required for cell adhesion to take place. 37dditionally, hydrophobic surfaces tend to adsorb hydrophobic microorganisms.Bacteria contain hydrophobins (known as adhesins), promoting adhesion and formation of conditioning layer.This layer initiates biolm development. 27ori and Matsumoto (2010) 35 studied the binding mechanism of Staphylococcus aureus on hydrophilic and hydrophobic surfaces.Notably, macromolecules tethered more on hydrophobic surfaces than hydrophilic surfaces.This was attributed to cell-surface contact time and adhesion force.According to Fabre et al. (2018), 38 large quantities of protein adhered on hydrophobic surfaces compared to hydrophilic surfaces.This phenomenon was described by reduction in free energy and a rise in entropy of the system. 39However, full assurance of absolute adherence/non-adherence of cells is not well understood since bacteria exhibit endless mechanisms enabling their adaptation under different conditions. 40For instance, hydrophilic bacteria adsorb on hydrophilic surfaces too, further promoting membrane fouling.Theoretically, polysaccharides (the major constituent of microbial cell exopolymer) favourably adhere on hydrophilic surfaces and form a gel-like rm matrix layer.Also, membrane roughness increases microbial attachment on membrane surfaces. 41Bernstein, Belfer, and Freger (2011) 42 reported conicting ndings where fewer cells were deposited at high membrane surface roughness.In contrast, other studies have found no correlation between surface roughness and bacterial deposition.Hence, extensive research is required to ascertain these contradictions.
The presence of high organic material in the feed solution is reported to increase the deposition of microbial contaminants on surfaces (Bogler, Lin, and Bar-zeev 2017) 27 (Fig. 2).According to Chen et al., (2021), 25 deposition of organic matter on membrane surface induced the formation of a conditioning layer, enabling more bacteria to adsorb on the surface.Other forces driving the adhesion of cells on membrane surfaces include thermodynamic and hydrodynamic interactions, polarity, oxygen content, bacteria abundance, process cross ow velocity and permeate ux.Furthermore, process parameters such as feed spacers and surface charge enhance biolm development.Accumulation of microorganisms on feed spacers reduces the ow of the feed stream, thus affecting water ux. 27

Biolm formation and EPS secretion
Following attachment, bacterial cells secrete EPS.The EPS and bacteria form a highly organised 3-dimensional structure known as a biolm (Fig. 3).The biolm forms between the solid and the liquid phase, where ultimate separation takes place.The EPS immobilizes and encapsulates bacterial cells, enhancing the rmness of the biolm.Though biolms vary in composition, they consist of bacteria (dead and alive), fungi, eukaryotic organisms, EPS, microalgae, and archaea. 43,44The biolm structure contains interstitial water channels, facilitating the movement of oxygen, nutrients, and genetic material. 27Bacteria are usually larger than other biolm microorganisms and typically range from 0.5-2 mm. 35Movement of bacteria through the biolm results in the colonization of new areas.Also, EPS contains polysaccharides, lipoproteins, proteins, glycoproteins, and carbohydrates contributing towards bacteria resistance against inhibitors. 45,46Furthermore, EPS bind the cells to the surface and maintain a stable environment.The EPS enhances communication within cells and act as a source of nutrients for the bacteria. 6Bacteria are known to release signicant EPS under thermal stress to overcome destruction by MD operational conditions. 32During membrane colonization, bacteria produce an exopolymer matrix, which grows and multiplies through the help of nutrients to adapt to the new environment (Fig. 3).
Notably, different biolm layers contain different pore-size distributions.Goh et al., (2013) 23 used evapoporometry to study the particle-size distribution of the biolm layer caused by two sludge solutions of different hydrophobicities.Reportedly, the hydrophilic sludge containing smaller pores was responsible for vapor pressure reduction.Moreover, smaller pores depressed the vapor ux, owing to the Kelvin effect rather than the effect of an increase in hydraulic resistance.During Fig. 2 .Mechanistic relationship of the salinity, heat, nutrient, and flow stresses on the distribution of bacteria on membrane surfaces during MD operation. 30,32g. 3 .Biofilm progression on membrane surfaces.Organic matter forms a conditioning layer enabling attachment of bacterial cells.The EPS are subsequently secreted to further promote attachment of bacterial cells and their colonies.During this process, the biofilm forms thus promoting dispersion and continuous growth of bacterial clusters. 15,27iolm development, bacteria senses each other within the vicinity, a process known as quorum sensing (QS).Bacteria secrete autoinducers during QS development, enhancing their communication. 47The QS further promotes microbial social activities and enhance community behaviour through the expression of certain genes. 48The dispersion of the biolm is augmented by QS. 49 Thus, QS further encourages biolm development.According to Zheng et al., (2022), 32 the development and succession of the microbial community is exerted by QS, amongst other factors.The obstinacy of biolm development and EPS secretion have deemed membrane biofouling a key research concern.

Enumeration and identication of microbial cells on the membrane surface
Biolm identication on the membrane surface is evaluated using a variety of techniques including scanning electron microscopy coupled with energy dispersive spectroscopy (SEM-EDS), optical coherence tomography (OCT) and confocal laser scanning microscopy (CLSM). 50,51Zodrow et al., (2014) 52 evaluated membrane biolm formation using CLSM, assessing the architecture, heterogeneity and biovolume of the microbial community.The biolms were heterogeneous and contained several colonies, with a plethora of Burkholderiales, Rhodobacterales, and Flavobacteriales.Bogler & Bar-Zeev, (2018) 22 identied the presence of polysaccharides and detected dead and live bacterial cells using CLSM, with a further evaluation of the average biolm thickness and total biovolume.Overall, feed operating temperatures of 55 °C led to a slightly thicker biovolume compared to 47 °C.This was attributed to provision of the optimum growth temperature of Anoxybacillus sp.Krivorot et al., (2011) 24 evaluated the progression of biolm development on the membrane surface using SEM.The experiment was conducted over a period of 19 days.Reportedly, deposition of microbial cells increased qualitatively as a function of time.Initial bacterial attachment was detected aer 28 h, with the conditioning lm seen only aer 20 h of operation.From 48 h towards 19 days, the biolm formed, thus covering the membrane surface.This later caused a reduction in process performance.
Physiochemical characteristics and viscoelastic properties of the biolm are largely determined via UV/vis spectrometry, ow cytometry, atomic force microscopy (AFM) and Fourier Transform Infrared Spectroscopy (FTIR). 27Phattaranawik et al., (2009) 21 utilized UV/vis spectrometry to quantify the concentration and composition of the deposited biolm on the surface of polyvinylidene uoride (PVDF) and polytetrauoroethylene (PTFE) membranes.Reportedly, the fouling layer was composed of protein, polysaccharides, and EPS.Since the concentration of polysaccharides on PVDF was higher than PTFE, it was envisaged that polysaccharides played a crucial role in ux decay.The ux of PVDF decreased considerably while that of PTFE remained stable.In another study, Krivorot et al., (2011) 24 determined the protein content of the biolm developed on the membrane surface using the Lowry protein assay.Notably, the protein concentration deposited on the membrane gave an estimation of the biolm thickness on the surface, further conrming SEM results.In another study, Goh et al., (2013) 23 employed FTIR to study and conrm the presence of polysaccharides and proteins, indicating the presence of bacteria.Specically, peaks at 1500-1700 cm −1 , 950-1170 cm −1 and 2800-3000 cm −1 were attributed to proteins, polysaccharides, and fatty acids, respectively.The use of highly specialized sequencing techniques for identication of bacterial species within the biolm structure has gained signicant attraction.For instance, Goh et al., (2013), 23 identied bacterial species on fouled membranes through DNA amplication using polymerase chain reaction (PCR).Bacterial organisms were predominantly hydrophilic, thermophilic, and halotolerant species.These included Rubrobacter taiwanensis, Caldalkalibacillus uzonensis, Caldalkalibacillus uzonensis, Tepidimonas sp, and Meiothermus hypogaeus.More recently, specialized techniques such as 16S rDNA and gene sequencing have gained research consideration.With these, bacterial species, growth patterns and behavioural changes of specic microorganisms are identied.Liu et al., (2020a) 55 studied the bacterial composition of the biolm on the membrane surface using 16S rRNA and gene sequencing following DNA extraction.The abundance of live bacteria was higher during initial biolm development stage which sharply declined as a function of time.This decrease was associated with an increase in salt crystal deposition on the membrane.A further increase in organic and inorganic substances contributed to the remarkable succession and evolution of biolm bacterial community.

Microorganisms responsible for membrane biofouling in MD
Biofouling in MD is predominantly exacerbated by thermophilic, mesophilic, and halophilic bacteria present in the feed solution.Since feed solutions sourced from different environmental locations contain microorganisms of different identity, analytical assessment of feed solution composition is imperative.For example, feed solutions sourced from marine environments cause biofouling through deposition of phyla Firmicutes, Bacteroidetes and Proteobacteria, largely at high temperatures and water salinity. 53From a general perspective, bacteria enhancing biolm development in MD include genera Mycobacterium, Bacillus, Lactobacillus, Cytophaga, and Flavobacterium. 45Notably, not all microorganisms present in the feed solution cause biofouling in MD.This uncertainty of events emphasises the importance of extensive biofouling assessment in MD systems. 26,29Gryta (2002) 26 evaluated biofouling occurrence in MD.Prior to MD operation, the feed solution was characterized by genera Pseudomonas, Penicillium bacteria, Aspergillus fungi and species S. faecalis.Evaluation of biolm on membrane surface displayed the presence of S. faecalis and Aspergillus fungi only.Genera Pseudomonas growth was hindered by oxygen deciency and high-water salinity.However, S. faecalis presented resistance to process conditions, thuspromoting membrane fouling.In another study, Zheng et al.
(2022) 32 reported a change of plump sphere or short rod to lankier rhabditiform with a microbial community transformation from Algoriphagus, Marinobater, Sulfurihydrogenibium to Chelativorans, Acinetobater, Idiomarina.Change in microbial community was attributed to elevated temperatures and high saline conditions.Moreover, the latter strains notably survived due to their high motility, good quorum sensing effect and EPS secretion.Microbial succession is detrimental to MD as it leads to production of more EPS, thus decreasing membrane life span and process performance.In certain instances, microbial communities present in the feed solution do not evolve, but deposit on the membrane surface.Zodrow et al., (2014) 52 evaluated the biofouling impact of seawater, predominantly characterized by Octadecabacter, Sediminicola, Loktanella, and Pelagibacteraceae.These strains were detected on the membrane surface, although in varying concentrations.The most abundant strain was mesophilic Octadecabacter, due to its high temperature and salinity resistance.Other strains detected included thermophilic Bacillales and spore self-protective Ralstonia.Additional microbes identied from the biolm are presented in Table 1.
Different modes of MD operations (closed and open loop) affect biolm development differently. 19,32Liu et al., (2020a) 55 used lake water (Xuanwu Lake) characterized by Proteobacteria, Actinobacteria, Bacteroidetes, and Cyanobacteria to assess biolm development in MD.Early biolm development consisted of genera Acidovorax and Acetobacteraceae, which was replaced by thermophilic Methyloversatilis at stage 2 of ux decay.Only Gammaproteobacteria and Deinococcus-Thermus were detected from the biolm under closed loop operation.The viability of these strains was explained by their halotolerant mechanism induced by a change in morphology to withstand heat. 32Under open loop operation, the membrane biolm was dominated by Anoxybacillus, Meiothermus, Schlegelella, Tepidimonas, and Vulcaniibacterium.These examples show the dependence of microorganisms on process parameters.The impact of feed salinity was further evaluated by Chen et al., (2021), 25 where the membrane biolm was composed of Proteobacteria, Bacteroidetes, Deinococcus-Thermus, Tepidimonas, Meiothermus, OLB14_norank, Env.OPS 17_norank and Schlegelella.The increase in feed concentration induced by continuous MD operation inhibited the growth of OLB14_norank, Schlegelella, and Tepidimonas.Evidently, MD conditions affect  microbial diversity, mobility and activity, leading to a biological selection of cells. 54

Effects of biofouling on process performance
Microbial accumulation on membrane surface produces undesirable effects such as permeate ux decays, salt rejection decays and a reduction in membrane lifespan.These effects are largely caused by vapor pressure depression/pore blockage, membrane wetting, and the accumulation of EPS and protein.Microbial accumulates further adsorb on feed spacers, thus exacerbating fouling.However, it is worth noting that decay in MD process performance is complex and may not be attributed to biofouling alone.Table 2 summarizes the impact of biofouling on water ux and salt rejection.

Permeate ux decay
Membrane biolm development notably reduces water ux.The water ux decreases due to: (1) membrane pore-blockage, (2) vapor pressure reduction, and (3) temperature and concentration polarization effects (Fig. 4).Commonly, permeate ux decay is caused by membrane pore blockage induced by the biofouling layer. 17,52However, degree of pore blockage determines the extent at which vapour pressure decays, hence ux decline.Goh et al., (2013) 23 noted a 60% decrease in water ux caused by the accumulation of microbial contaminants on the surface of the membrane.These contaminants essentially caused a 32% decrease in the vapor pressure, thus reducing the rate of water recovery.This was ascribed to an increase in heat and mass transfer resistance across the boundary layer.Resistance to heat transfer minimizes water vaporization, thus reducing process performance. 6Liu et al., (2020a) 55 evaluated the occurrence of biofouling on a closed loop direct contact membrane distillation (DCMD).Reportedly, the ux decreased by 55.79% over time due to the increased deposition of foulants.
Permeate ux decay of MD systems vary depending on feed temperature.This phenomenon was reported by Bogler & Bar-  a Where T F , T P , J I , J F , k I and k F represent the feed and permeate temperature, initial and nal water ux, initial and nal permeate conductivity respectively.
Fig. 4 .Temperature polarization phenomenon in MD.T f,in , T fm , T pm and T pb represent temperatures in the bulk feed solution, near the membrane surface (feed side), near the membrane surface (permeate side) and in the bulk permeate, respectively. 58eev, (2018). 22The feed stream temperatures were controlled at 47 °C, 55 °C and 65 °C with corresponding ux declines of 30%, 78% and 32%, respectively.Flux decay was attributed to; (1) rapid growth of Anoxybacillus sp at the operating temperatures, (2), membrane pore wetting, (3) TP and CP, and (4) vapor pressure depression and hydraulic resistance.Though temperature may be increased beyond bacterial optimum growth temperature, cell adaptability at new operating conditions continue to threaten MD performance.For these reasons, the growth of biolms increase at higher operating temperatures (Elcik et al., 2022). 57

Salt rejection decline
Membrane wetting is a fundamental aspect requiring signicant attention in MD systems.Wetting promotes the passage of feed in liquid form, thus reducing salt rejection. 5,55,59Deposition of hydrophilic biolms causes membrane wetting.Bogler & Bar-Zeev, (2018) 22 reported a reduced salt rejection caused by EPSinduced membrane wetting.A signicant difference in salt rejection was obtained for different feed operating temperatures.30-fold and 90-fold increase in distillate salinity was recorded at 55 °C and 65 °C respectively.90-fold increase was attributed to the improved bacterial growth conditions causing rapid biolm formation.Evidently, salt rejection and water ux declined upon inoculation of feed water with bacteria (Anoxybacillus sp.).However, bacterial lm conditioning occurring on membrane surface may not affect process performance, especially in cases where hydrophobicity of membrane pores is not altered.This phenomenon was reported by Zodrow et al., (2014) 52 .Although the contact angle of the hydrophobic membrane decreased from 134 ± 4°to 32 ± 6°due to deposition of foulant layer, no membrane wetting occurred.The hydrophilic layer grew on the membrane surface without changing pore hydrophobicity.

Frequency of membrane replacement
A decrease in membrane lifespan caused by the accumulation and persistence of biolm development leads to frequent membrane replacements.This causes signicant increases in operational costs. 5The frequency of membrane replacement largely depends on the type of feed characteristics and hydrodynamic conditions. 60In desalination plants, membrane replacement occurs within 4-5 years whereas industries treating dairy products replace membranes within 3 years. 60In general, membrane replacements are not recommended due to interruptions in the process and signicant amount of labour requirements. 61

Mitigation of biofouling in MD
Major strategies used to mitigate biofouling in MD include pretreatment of the feed solution, membrane modication and cleaning (Fig. 5).Detailed impact of these strategies towards water ux and salt rejection stabilities is presented in Table 3.

Pre-treatment of the feed solution
Pre-treatment of the feed solution is a useful tool to minimize biofouling.This involves removal of potential foulants from the feed solution before MD processing.Pre-treated feed improves resistance to ux decay and reduces frequency of membrane cleaning. 36,62Pre-treatment methods are largely dependent on the composition of the feed solution. 64There are three types of pre-treatment methods widely applied in membrane-based technologies.These are physical, chemical and hybrid pretreatment processes.Physicochemical pre-treatment processes.Physical and chemical pre-treatment processes include ltration (ultra/ micro), sonication, coagulation, and chlorination, amongst others.Ultraltration (UF) and microltration (MF) remove foulants from the feed solution using size-exclusion mechanism.The MF and UF minimize nutrients required for microbial growth.The UF was reported to effectively remove algal cells (96%) and non-algal cells (98%) from the feed solution in a seawater reverse osmosis (SWRO) plant in Saudi Arabia. 65It is worth noting that planktonic cells passing through UF reproduce, thus enhancing biolm development. 66For example, Zodrow et al., (2014) 52 evaluated MF pre-treatment of RO and MD processes.Notably, inefficient removal of organic and biological matter from the feed stream caused blockage of membrane pores, thus a decline in water ux within 12 h of operation.Pre-treatment processes are currently evaluated for MD systems with limited literature reported. 56lternatively, sonication (with ultrasound frequencies of $18 kHz) is strategically used to manage bacterial growth. 67ouling is controlled through cavitation, where production of strong convective currents is triggered. 68However, this pretreatment process is rarely reported.Mathieu et al. (2019) 69 applied ultrasound (frequency of 46 kHz) to inhibit bacterial Fig. 5 .Latest developments of biofouling control strategies in MD. 62 growth in bulk drinking water "N" spiked with 100 mg L −1 of Ca(OH) 2 .Evidently, a 7-fold reduction in bacterial growth was reported.Ultraviolet (UV) treatment is another alternative, where hydroxyl radicals are produced to inhibit bacterial growth. 44During UV pre-treatment, bacterial DNA is broken down while proteins are effectively denatured. 44However, UV is costly, thus limiting its application in MD water purication. 70ue to prevention of UV rays through the feed solution by contaminants, UV treatment is limited to feed turbidity.Generally, wavelength of 200-400 nm is required to initiate cell death. 6To minimize reactivation of microbial cells, UV is applied in conjunction with other mitigation strategies.
Coagulation, chlorine and ozone injection are some of the chemical processes used during pre-treatment.Coagulation is a cost-effective and most convenient process, involving stabilization of particulates through the addition of a coagulant. 71norganic coagulants include ferric chloride (FeCl 3 ), aluminium chloride (AlCl 3 ) and polymeric aluminium chloride (PAC) whereas organic coagulants include polyacrylamide (PAM), poly dimethyl diallyl ammonium chloride (PDMDAAC) and microbial occulants. 72,73Since higher doses of inorganic occulants are known to cause secondary pollution due to the presence of residual metal ions, a hybrid system of organic and inorganic coagulants is used to improve process performance.Zhang et al.
(2022) 63 used a hybrid system to evaluate the performance of microbial occulants modied with PAC (MMF/PAC) to minimize membrane fouling.Cake layer formation on the membrane surface was effectively minimized.Management of biolm development has been predominantly realized through chlorine dosing. 6Although chlorination of the feed stream effectively prevents biolm development, it produces harmful mutagenic and carcinogenic by-products. 74Moreover, chlorination damages polymeric membranes due to its high oxidizing ability. 63Spore-forming microbes such as Bacillus are reported to resist chlorination, thus making pre-treatment ineffective. 44lso, ozone (O 3 ) is used to control the growth of bacterial cells in MD systems.Owing to its high oxidizing properties, ozone inactivates viruses, bacteria and organic contaminants. 44Yong Zhang et al. (2016) 75 coupled a DCMD system with ozone injection to treat organic pollutants from wastewater.49% ux recovery was attained.Although ozonation minimizes membrane fouling, its application is limited to process costs. 6imilarly, ozone is insoluble at high temperature, thus making it unsuitable for MD applications.Other oxidative chemicals used to reduce biofouling include peracetic acid and hydrogen peroxide.
Hybrid pre-treatment processes.The complete removal of microorganisms require hybrid rather than a standalone process.Therefore, a combination of these processes can be chosen from a wide range including ozonation, chlorination and UV irradiation.Stand-alone processes portray a certain level of limitation, hence the need for hybridization.Chen et al.
Pre-treatment and membrane cleaning

Magnetic coagulation and HCl cleaning
Wastewater 65 days 97% ux and salt rejection were restored aer cleaning (at high HCl concentration)  (2022) 76

Membrane cleaning
Membrane cleaning involves removal of microbial accumulates from membrane surfaces.The concept of cleaning was introduced to restore the original performance of the fouled membranes. 6Furthermore, membrane cleaning reduces frequency of membrane replacements. 78There are two types of cleaning methods widely applied in membrane-based technologies: (1) physical cleaning and (2) chemical cleaning.Physical cleaning.Physical cleaning involves the use of deionized water during MD operation to facilitate foulant detachment. 15The use of water guarantees minimal damage to the membrane structure.During cleaning, the feed solution is pumped into the module, followed by deionised water through the system at higher owrates.These owrates ensure increased shear forces to enable detachment of foulants from the membrane surface and spacers. 27In other instances, high pressure water is utilized. 67Lyly et al. (2021) 79 used deionised water to restore the performance of PTFE membranes during treatment of synthetic seawater spiked with BSA and microalgae (Cylindrotheca fusiformis).The following steps were implemented for physical cleaning: (1) ow of pure cold-water (at 25 °C) for 10 min, followed by (2) hot water (at 60 °C) for 10 min and nally (3) cold-water for 5 min.Reportedly, 62.69% of water ux was restored, displaying promising results for future applications.
Chemical cleaning.Chemical cleaning involves the use of chemicals to continuously remove foulants from the membrane surface.To minimize membrane damage, frequency and concentration of the cleaning agent should be optimized. 64ominating chemical cleaning involves removal of inorganic and organic foulants using acidic and basic solutions respectively with limited information on biofouling treatment. 80heng et al., (2022) 32 used HCl and water to remove microbial accumulates from the membrane surface.The process ensured complete removal of Euryarchaeota.However, full restoration of the ux was not achieved, largely due to Idiomarina cells' affinity for the membrane surface.Krivorot et al., (2011)  24 investigated the removal of biofoulants from the membrane surface through ushing with a series of steps: (1) NaOH (pH 12 at 40 °C), (2) distilled water, and (3) 70% ethanol.Notably, the original ux of the membrane was fully restored.To minimize cleaning requirements, modication of the membranes is recommended to render them resistant to biofouling (Bogler et al., 2017). 27his approach ensures minimal waste disposal, thus ensuring environmental safety. 81

Membrane modication
Modication involves the alteration of membrane properties to render them resistant to biofouling.Membranes are modied through various approaches; including incorporation of metal nanoparticles (MNPs) (Fig. 6) and alteration of membrane surface properties such as surface roughness, membrane hydrophobicity and surface charge. 59,82ncorporation of MNPs.Attachment and proliferation of microbial cells on membrane surfaces is largely minimized through the incorporation of biocidal MNPs.Upon oxidation, ionic counterparts of biocidal MNPs are released to react with microbial cells.Biocidal MNPs capable of inactivating bacterial cells include Ag, Ti, Zn and Fe. 15,83Inactivation occurs due to the electrostatic interaction between positively charged metal ions and negatively charged thiol groups on bacterial DNA. 84Nthunya et al., (2020) 85 modied polyvinylidene uoride (PVDF) membrane using silver nanoparticles (AgNPs) and functionalized carbon nanotubes (f-MWCNTs), achieving successful biofouling, colloidal and organic fouling control in MD.Although fouling mitigation is evident, biocidal MNPs tend to leach from the membrane surface, thus reducing fouling control. 27In addition, leaching of MNPs reduces the quality of produced water.Hence, ensuring the stability of these MNPs on membrane surfaces is crucial.
Alteration of membrane surface properties.Rendering MD membranes superhydrophobic is a useful approach to mitigate biofouling, where the surface energy of the membrane is reduced. 85Ideally, superhydrophobic membranes are characterized by contact angles $150°.Teoh et al. (2022) 86 enhanced the surface hydrophobicity of PVDF membranes using nylon taffeta substrate for MD treatment of aquaculture wastewater.Self-cleaning microtextured membrane was characterized by water contact angle of 153°and a sliding angle of 8.9°.Notably, the water ux (30 kg m −2 h −1 ) remained stable.Also, hydrophilic coating of hydrophobic membranes is reported to minimize biofouling.This is achieved through fabrication of Janus membranes.During the process, membranes of asymmetric wettabilities, i.e. hydrophobic/hydrophilic combinations are produced. 27Yang et al. (2010) 87 graed zwitterionic polysulfobetaine methacrylate (polySBMA) to polypropylene (PP) membrane.Water contact angle of the active site was reduced from 145°to 15°.Reportedly, graing density of 560 mg cm −2 provided resistance to bacterial adhesion.This was attributed to a decrease in interaction between hydrophilic bacteria (E.coli) and membrane active surface.Furthermore, hydrophobic membranes are coated with a thin layer containing hydrophilic MWCNTs and AgNPs to reduce combined fouling of organic, inorganic and biofoulants.While uncoated membranes face ux decays of approximately 90% and salt rejection decays of 6%, hydrophilic coating of membranes minimized ux and salt rejection decays to 24% and 0.75% respectively.Therefore, hydrophilic modication of membranes maintains process resistance to fouling. 88

Other biofouling mitigation strategies
Other techniques used to control biofouling of MD systems include air-sparging (bubbling), quorum quenching (QQ), nitric oxide (NO x ) injection and alteration of hydrodynamic conditions.The NO x enhances biolm dispersal by decreasing 60% of membrane biolm coverage. 89However, due to NO x instability and insolubility, their application in water treatment is minimal.The QQ involves the use of quorum quenching enzymes to deactivate signalling molecules required for biolm development. 6The quorum sensing system regulates the communication and aggregation of bacterial species.Its suppression assists in managing biofouling in membrane systems.Gram-negative and Gram-positive bacteria use signalling molecules such as N-acyl-homoserine lactones (AHLs) and autoinducing peptides (AIPS) to promote biolm development and successions. 90To minimise these developments, QQ enzymes such as AHLs lactases damage the lactone rings of the cells.This reduces effective communication between cells, thus slowing biolm progression. 47,91,92Air-sparging involves the bubbling of air in the feed stream to increase ow rate, shear stress and turbulence, thus preventing membrane clogging. 27hang et al. (2021) 93 studied air bubbling to minimize biolm development.Reportedly, membrane deposition of inorganic, organic and microalgae was signicantly reduced leading to a recovery of 16.95% water ux.This was attributed to an increase in turbulence ow at the feed-membrane interface.Comparatively, shear turbulence of gas bubbling removes foulants better than shear turbulence of feed owrates. 94However, the increase in air-bubbling beyond a certain threshold reduces the permeate ux. 95Therefore, optimization of air-water ratio is imperative. 27Lastly, hydrodynamic conditions affect MD process performance.Silva et al. (2018) 96 evaluated different operating conditions such as solute concentration, owrates, and different membrane modules on process performance.Reportedly, water ux was higher in a perpendicular ow (Wcell) of the feed to the membrane compared to a parallel ow (H-cell).This was ascribed to a decrease in temperature and concentration polarization.Remarkably, feed owrate should be optimised to maintain high water ux while simultaneously preventing biofouling.According to Zheng et al. (2022) 32 bacterial cells got thinner and longer as feed ow passed through the membrane.The inlet was less fouled by Chelativorans, Acinetobater, and Sphingobium.On one hand, temperature tolerant bacteria such as Idiomarina and Phenylobacterium were detected in abundance at the inlet where process velocity is high.Although a high feed owrate is desired to minimize the feed boundary layer through high shear stress, 96 caution must be exercised to minimize bacterial succession and abundance.

Conclusions and future outlook
MD remains a promising technology to purify high saline wastewater and seawater.However, hydrophobic membranes used in MD are susceptible to biofouling, a process involving the deposition and accumulation of microorganisms.Requiring nutrient availability to mature into a biolm, these microorganisms largely originate from the feed solution.Several factors affecting microbial deposition include membrane properties, hydrodynamic conditions, feed solution properties and membrane module designs.Mitigation of biofouling is essential as it affects membrane lifespan, increases operational costs, and diminishes permeate water quality.Moreover, large scale application of MD has been hindered by fouling, amongst other factors.Various approaches minimizing biofouling are proposed.These include pre-treatment of the feed solution, membrane cleaning, and metal nanoparticle and surface alteration-based modication.Current reported studies largely focus on the characterization of the microbial community within the biolm structure.Although this is essential, research work focusing on different approaches to mitigate biofouling is essential.Moreover, various module designs need to be explored.To establish the sustainable application of MD systems at pilot scale, fouling experiments should be carried over long operating periods (600 h and more).Elucidating cellto-surface interactions using various mathematical models to simulate real operating conditions requires experimental attention.Directing efforts toward lessening attachment will ensure minimal EPS secretion and hinder biolm progression.Optimization of MNPs concentration on membranes to determine a balance between optimum activity over long periods and the rate of leachability is required.Indeed, membrane biofouling in MD requires signicant experimental work to mitigate the existing challenges.

Table 1
Biofouling-causing microbial organisms in MD a

Table 2
A summarized impact of biofouling on MD efficiency a

Table 3
Mitigation strategies of biofouling in MD a evaluated a combination of UV and chlorine to mitigate the growth of Staphylococcus aureus.Notably, the generation of free radicals such as OHc and Clc minimized bacterial growth and inactivated the photo-reactivation of the bacteria.In another study, O 3 , UV/O 3 and UV-assisted peroxidation (UV/ H 2 O 2 ) were used to pre-treat the feed solution (wastewater) in DCMD systems (Kumar et al. 2020).