Characterization of Fouling in Used Reverse Osmosis Modules with Methods to Recycle and Repurpose Via Transformation

Abstract

Used reverse osmosis (RO) membranes have become a concerning issue, as there is significant difficulty in properly disposing of them without damaging the environment, which is exacerbated by the rapidly growing industry of RO membranes. More investigation needs to be done on RO membrane fouling and potential solutions to such issues. Therefore, this paper aims to characterize the types of fouling on used reverse osmosis modules and their preventative methods. In addition, the testing conditions, oxidative chemicals, results of the transformation process, and specific, potential applications of recycled membranes will be investigated. There are 3 primary types of fouling: organic, inorganic, and biofouling. The pH of inorganic fouling was investigated, and it was found that a pH greater than 10.5 was conducive to less fouling, and the fouling of hardness and silica could be reversed by antiscalants and coagulants. Biofouling could be minimized using a variety of chemical oxidative methods with their advantages and disadvantages, but organic fouling was more complex. The goal of the transformation processes was to turn discarded RO membranes into nanofiltration or microfiltration membranes. This transformation process is done by peeling the active polyamide layer underneath to reveal the polysulfone layer, which possesses similar properties to the aforementioned membrane types. The most effective transforming agent was found to be potassium permanganate, rejecting up to 96% of suspended solids, but sodium hypochlorite was also found to be as effective or better. Sodium hydroxide was also investigated as an oxidative chemical. For the method of transformation, active recirculation was more effective than passive immersion. Specific applications of the transformed membranes include pathogen removal, water treatment systems, irrigation systems, and especially anion exchange membranes in electrodialysis. 

1. Introduction

Water is a central resource necessary for survival, which is why demand will continue to be high as time passes, especially with the rapid advancement of technologies. In the modern world, there already exists a variety of strategies to manage, recycle, and process water, whether it be for consumption, sanitation, or the processing of other materials. A large industry in water purification already exists, with the number of desalination plants increasing by 70% over the last 2 decades, and provides usable water on a large scale to sustain communities, demonstrating the need and importance of water. The size and scale of the plants have also increased, going up to 500,000 cubic meters of water filtered per day globally over the last 2 decades¹.

One of the most prominent strategies to provide clean water is water filtration, purifying the water in the surroundings to fulfill demand in an effective way. However, there are many problems and risks associated with membrane technology. Water filtration may require an inefficient amount of energy or extreme conditions, which may damage the membrane and have negative impacts on the climate. Replacing membranes is also resource-intensive and sometimes requires consistent maintenance to keep them functioning. Their production may take a long time, require lots of resources, or be detrimental to the environment. As a result, membranes are often discarded into landfills, and factoring in the large industry of membrane filtration, the waste will continually add up at an alarming rate. There already exists a large problem with membrane disposal, and it will continue to worsen if solutions are not implemented to recycle these membranes, specifically reverse osmosis membranes. Therefore, there is a need to find alternative disposal methods for these membranes to minimize environmental impact. 

Reverse osmosis (RO) membranes are a key type of membrane in water filtration. They have extremely small pore sizes compared to ultrafiltration or nanofiltration membranes, thus requiring a high-pressure gradient across the membrane. This high pressure leads to fast fouling, which leads to low water permeability, recovery, and quality. Another reason why fouling occurs is because of physical and chemical interactions causing gel layers, inorganic particles, filter cake, etc., to form rapidly². The flux is usually not very high due to reverse osmosis membranes needing to operate through diffusion, which heavily increases the likelihood of fouling very quickly. These problems ultimately result in large amounts of landfill waste that quickly accumulate due to a lack of efficient and convenient options for disposal. For example, more than 840,000 end-of-life RO membrane modules (>14,000 tons of plastic waste) are disposed of in landfills annually³. RO users also do not have many alternatives for disposal and therefore must resort to options with negative consequences⁴.

However, reverse osmosis is still effective in desalinating water and providing drinking water for the population, which is why there has been such an emphasis on it. To continue utilizing these plants, the problems must be addressed. The potential for reuse has been considered by cleaning the membranes or positioning the used membrane at the front of the filtration system⁵. Many attempts have been made to transform end-of-life reverse osmosis membranes into filtration systems with larger pore sizes by removing the active polyamide layer using a variety of strong oxidizing agents such as potassium permanganate or sodium hydroxide¹. Their results were largely varied and depended on many factors, ranging from concentration to the method of applying the oxidizing agent. Innovative strategies have been proposed to repurpose these membranes in other filtration processes, conserving materials and reducing waste byproducts. However, these studies on potentially reducing costs or conserving membranes have not been consolidated and have varying conditions and degrees of success. These solutions have been mostly independent, with no main solution to address RO membrane waste, which this review aims to address. Lastly, investigations into preventative measures are also crucial because the energy and resources needed to recycle the membranes can be avoided if membrane fouling or damage can be kept to a minimum, which is why pretreating the feed for serious RO fouling agents can help stop landfill waste^{2 ,6 ,7}. 

Figure 1 demonstrates the priority of effective disposal for RO membranes. This study will go into more detail about reuse, recycling, and treatment. It is preferred to reduce membrane fouling and prolong the use before needing to reuse the membrane in some other function. If reuse is not possible, recycling is considered, as well as recovering the energy that can be restored from the membrane. Disposal is last, as that is the problem that should be avoided at all costs to reduce environmental impact and waste. 

This paper intends to study the effect of fouling on membranes, classify the types of fouling, and consider prevention methods for the fouling, as well as go in-depth about the variety of ways to reduce reverse osmosis waste, including reusing, repurposing, recycling, and transforming end-of-life membranes. Transformed and repurposed membranes will be assessed compared to commercial, new membranes, while the structure of used and fouled membranes will be analyzed. 

2. Effects of fouling

The effect of fouling on reverse osmosis (RO) membranes needs to be closely evaluated to develop a solution. In this section, the variety of foulants, their effects on the membrane, and their category will be discussed, as well as potential preventative measures that can be taken. 

In a study conducted by Minier-Matar et al., the process water from an industrial RO facility is characterized and studied to analyze the fouling types⁶. The study characterized RO membrane fouling into 3 groups: inorganic, biofouling, and organic fouling. 

Inorganic fouling consists of hardness ions and silica-type materials that scale on the membrane and reduce the flux. Examples include calcium sulfate, calcium carbonate, and barium sulfate. This kind of fouling, however, can be easily prevented by using scale inhibitors⁶. Xiaochen Sun et al. set up 5 experiments using RO units with differing silica compositions along with the same concentration of transparent exopolymer particles(TEP), a kind of translucent organic material that is present in aquatic habitats and has been linked to organic and biofouling and is usually more difficult to handle than extracellular polymeric substances(EPS)². Predominantly found in marine habitats, TEP has unique properties like high viscosity, strengthened adhesion, and high strain performance, along with the ability to transform its shape to pass larger pore sizes, to then foul RO membranes. TEP has also been shown to easily adhere to RO membranes and to cause severe fouling quickly. The results of Sun et al.’s study showed that as silica concentration increased, the more severe the membrane fouling became due to the polymerization of silica, with the decrease in flux becoming worse in the presence of TEP. The membrane fouling plateaued and remained around the same level at 6 millimolar (mM), which is a measurement of how concentrated the silica is, indicating that membrane fouling reached its most severe level at 6 mM of silica, as shown in Figure 2. The reversibility of fouling via hydraulic cleaning correlated negatively with silica concentration, with concentrations of silica 6 mM and above very difficult to reverse. In general, however, inorganic fouling is not a significant issue as there are adequate countermeasures such as antiscalants and hardness removers.

Similar fouling mechanisms have also been observed in biofouling, where biofilm formation can lead to irreversible damage. Biofouling is made of biological matter such as polysaccharides and proteins that can be prevented using biocides⁶. As a study by Matin et al. puts it, biofouling is composed of biofilms, which are microbial cells on a surface that are enclosed by EPS⁸. These cells are not easily removed and are present in almost all bodies of water. Attempts to remove this fouling through stronger rinsing or cleaning may risk damaging the membrane. Biofilms can form on many different surfaces by adhering to those surfaces, and the microorganisms in these biofilms feed off of nutrients accumulating in the water phase. The microorganisms excrete EPS and form biofilms in which they are embedded. Therefore, dissolved nutrients are converted from the feed solution to an immobilized semisolid state on the surface of the RO membrane. EPS accounts for the majority of the total organic carbon of biofilms and is the primary material for the structure of biofilms. Consisting of polysaccharides, different proteins, and other macromolecules, the biofilm’s properties are derived from EPS and bacterial cells. EPS is also responsible for adhesion by a slime matrix that holds the biofilm together and for antimicrobial resistance. Modifying the surface of the membrane rather than pretreatment seems to be more effective, as pretreatment is supposed to minimize microbe concentration and not prevent the adhesion of bacteria. 

Organic fouling possesses similar qualities to biofouling, but has a greater variety of potential foulants. Organic fouling comprises many different foulants that can cause irreversible fouling, but not all organic foulants are detrimental to the membrane. Organic foulants are usually measured in total organic carbon (TOC) and dissolved organic carbon (DOC), but neither gives detailed descriptions of the foulants, only their organic content. Identifying specifically which organic particles cause fouling is crucial to obtaining effective pretreatment of the feed. In the same study by Minier-Matar et al., a liquid chromatography system is used with an organic carbon detector (called LC-OCD) to separate soluble organics based on several properties, such as their interactions with water and their molecular weight⁶. This kind of fouling takes the most research and investigation to develop effective preventative measures due to its variety. It is also related to biofouling and inorganic fouling, and such connections must be taken into account.

External types of fouling include the degradation of the active membrane layer in RO modules, which reduces the efficiency of the membrane, and the growth of bacterial cells on the membrane, causing irreversible fouling. These different types of fouling are interconnected with each other, and thus must be discussed together to devise a solution to membrane fouling. Such categorization and description could lead to easier methods of reducing membrane waste. Although there are many antifouling and cleaning agents used in an attempt to reverse these conditions, repetitive exposure can slowly affect their rejection efficiencies, promoting earlier and more frequent disposal. 

3. Reversibility and fouling prevention

Reversibility and fouling are major factors in deciding the usability of RO membranes. Most of the time, through prolonged use, the fouling on RO membranes is irreversible if the foulant concentration in the feed is high. Therefore, most methods of prevention include the transformation of membranes to make backwashing and other cleaning methods more effective, which will be discussed later in the paper, as well as preventative methods such as pretreating the feed for problematic foulants. 

3.1 Effect of pH on inorganic fouling

In a study by Jingjing Sun et al., a single-step and double-step process of removing silica and hardness ions was investigated, using a hardness removal agent and desilication process⁷. It also investigated the effect of pH to relieve fouling and scaling. The single-step process had iron chloride added to the wastewater with sodium hydroxide. The double-step included 3 different methods of producing effluent: adding sodium hydroxide, then magnesium chloride; adding sodium hydroxide, then iron chloride; and adding sodium hydroxide and magnesium together in the first step, then iron chloride in the second. 

The results for both processes suggested that as basicity increased, the less concentrated the calcium and magnesium ions became. For silica, a specific range of pH from 10.5 to 11.0 was conducive to low silica content, where pH is a scale of how acidic a solution is, with higher values being less acidic. In a similar study by Xiaochen Sun et al., one of the goals was to investigate the effect of pH on silica scaling and fouling behavior². Running experiments with only xanthan gum, silica, or both, they found that xanthan gum had a minimal effect on membrane flux reduction, having a J/J₀ value of 0.8 to 1(signaling the ratio of experimental flux to normalized flux), while silica significantly decreased the membrane flux at a pH of 9 due to the release of silica ions in the solution causing greater membrane fouling. In Figure 3, both (a) and (b) show that the pH of 9 has drastically lower flux rates than the other experiments at another pH, regardless of whether or not TEP was involved, as silica was the main factor lowering flux, with a J/J₀ value of around 0.2 at 600 minutes. 

Both of these studies indicate that the pH for treating inorganic fouling should be kept relatively high to have acceptable removal of silica and hardness ions. Although pH alone is not enough to entirely mitigate the effects of scaling, it is beneficial to maintain conditions that contribute to low hardness content.

3.2 Pretreatment 

Some methods of prevention include filtering out active foulants that are the primary sources of RO fouling, like organics. In addition to categorizing the fouling in RO filtration, Minier-Matar et al. also ran small-scale experiments using RO feed, reject, active carbon filter feed(ACF), and RO permeate from an industrial facility to observe the organic fouling from the facility⁶. One type of experiment was run using the ACF feed to investigate the effect of organics on RO membranes, and another type of experiment was run with regular RO feed that was not pretreated, with both having similar normalized flux. The results demonstrated that there was an immediate 10% decrease in flux from 1.8 to 1.61 L/m²h bar for the RO feed experiment because the RO membrane was fouled by organic foulants. On the other hand, the RO feed experiment showed stable performance with no decline in flux or fouling trends after being treated with the ACF, suggesting that ACF removes problematic organic compounds that could deposit on the RO membrane surface. Without ACF pretreatment, there was a 10% decline in RO permeability. ACF also removed 50% of hydrophobic organics that caused membrane fouling, reducing the hydrophobic fraction of TOC from 1.3 to 0.65 mg/L in ACF-treated RO feed, and there was no observed accumulation of organics on the RO membranes after ACF. These results show that pretreating the feed can be beneficial in preserving the lifetime of RO membranes, reducing their fouling, and minimizing disposal costs .

In conjunction, coagulants such as iron chloride and magnesium chloride can be used to mitigate silica and hardness fouling, as referenced by Jingjing Sun et al.⁷. In the single-step process, iron chloride was overall shown to be effective in lowering calcium and magnesium ions due to the adsorption capacity of the large colloidal ions. When iron chloride concentrations were greater than 500 mg/L, silica was below 5 mg/L, and total hardness concentration was lower than 45 mg/L. In the double-step, adding both magnesium chloride and iron chloride together was much more effective than by themselves, with experiments varying one anticoagulant causing hardness, silica, and ion concentration to decrease. Optimal concentrations of magnesium chloride and iron chloride were 600 and 100 milligrams per liter, respectively. It was also found that utilizing a double-step process or treating hardness and silica separately was more cost-effective and more efficient as well, with the smallest total unit price and the cheapest disposal cost.

These 2 papers indicate the potential for effective pretreatment for RO users to minimize the waste produced. Although not directly recycling or reusing, these methods still benefit RO filtration systems by inducing longer usage and lowering the frequency of disposal. By using these methods, the environmental impact can be minimized and reduced, and RO membranes will not have to be backwashed or replaced as often. 

Biofouling impacts RO processes through phases, which include rapid fouling at first due to early attachment of microorganisms and a slow phase, which results from equilibrium between biofilm growth and biofilm loss(cell detachment)⁸. The primary methods of minimizing fouling are using feed pretreatment methods to filter out the biofoulants and to restore the membrane by using chemicals to clean it. 

One option for pretreatment to remedy biofouling is to use microfiltration(MF) or ultrafiltration(UF) membranes that are more efficient than conventional pretreatment processes due to spatial and chemical factors. Matin et al. looked at other studies investigating both pretreatment methods, and through various studies, they found that conventional pretreatment either failed or was inconsistent in preventing RO fouling⁸. They also discussed the uncertainty of using traditional methods on raw water because of the water’s fluctuating composition, and therefore, a specific pretreatment technology would be inefficient. UF systems seem to be better than MF systems when filtering suspended solids and microorganisms, making the systems an ideal alternative to conventional methods. 

Another option is to apply biocides like chlorine, ozone, or ultraviolet(UV) light either continuously or periodically. This application can also be used to preserve membrane-related components. Chlorine has been reported to limit microbial growth and prevent biofilm formation in water containing 0.04 to 0.05 mg/L of free chlorine. However, using chlorine could potentially peel the polyamide layer of the RO membrane and contribute to the creation of harmful byproducts, as well as being ineffective in targeting certain pathogens that display resistance to high pH and chlorination. 

Ozone has also been widely used due to its strong oxidative power and few byproducts. It is effective in cleaning biofilms and biomass as it weakens the biofilm matrix and allows it to be removed through shear forces, but it suffers from the disadvantage of cost, being four times more expensive than chlorination, and undesired degradation of the membrane surface. 

UV light has also been tested due to its ability to inactivate microorganisms. This method does not produce byproducts, is independent of pH, and effectively targets organic compounds serving as nutrients for microbial growth. In testing, UV light demonstrated effective treatment of microorganism control, completely removing a test organic compound and preventing flux decline, as well as maintaining salt rejection^{8 ,9}. The results of disinfection and chemical cleaning are presented in Figure 4, where the strongly oxidative chemicals either produce harmful side products or contribute to the corrosion of the RO membrane. Physical disinfection methods are also considered, such as sand filtration, pretreatment, and UV irradiation.

Membrane cleaning involves weakening the biofilm matrix by using chemicals like bactericidal agents and chlorinated water to interfere with the bonding and removal of the biofilm using shear forces. Multiple combinations of chemicals are available, such as an enzyme-based dispersant with a bactericidal agent; however, the ideal combination as well as the process of shear forces need to be investigated⁸.

Lastly, there is a possibility of using kinetics, specifically sonication technology, to dislodge and break up filter cake formation, reducing fouling. Concentration polarization and foulant deposition can be reduced by achieving chaotic vortices near the membrane surface. In a study by Wang et al., 2 types of filtration systems were investigated in cross-flow filtration experiments, static and dynamic filtration¹⁰. Static filtration uses no moving parts but has turbulence and curved tubes to reduce particle clumping. Dynamic filtration vibrates the filter or moves a neighboring part, specifically a pulsatile or oscillating flow. Figure 5 illustrates the setup for the cross-flow filtration system as well as the membrane section of the system.

By constantly altering the shear rate, cake deposition is reduced. A negative-differential-resistance oscillator was used under constant pressure to automatically produce a pulsating flow without requiring external controllers. The results of the study showed that in one pulsating period, the forward rate lasted for 49% of the period, backward flow made up 24% of the flow, and the rest of the time was chaotic flow. Ultimately, this alternation leads to a scouring effect by disrupting the polarized layer and removing foulant from the membrane surface. Negative pressure present causes backflushing as well. In testing the oscillator with aluminum hydroxide suspensions and pneumococcus fermentation broth, the flux increased by 5.0 times that of static filtration for aluminum hydroxide, and 1.7 times that of static filtration for the broth. The specific oscillator used in the study also does not need complex mechanisms to function. 

This study presents an interesting solution to membrane fouling, as most other methods involve treatment of the feed before filtration to limit fouling and preserve the RO membranes. Although a unique idea, it presents promising results that could be used to address the problem of membrane fouling, thus increasing membrane lifetime. However, one drawback of the study is whether or not the oscillator could be specifically applied to reverse osmosis membranes, as the paper discusses fouling in general. The oscillator could potentially damage the filter, requiring disposal even before the membrane has fouled. Another drawback is that the oscillator was tested with known suspensions and potential foulants. This knowledge may not apply to reverse osmosis foulants because the oscillator working for the aluminum hydroxide suspension does not mean the oscillator must work for reverse osmosis foulants. Durable foulants like silica and TEP may not break down to prevent fouling. Nonetheless, the paper introduces a solution to fouling that holds promise.

3. Transformation and repurposing membranes

Another avenue for reusing reverse osmosis (RO) membranes is by transforming them. Due to the difficulty of backwashing and directly reusing fouled RO membranes, investigations have turned to transforming these membranes to membranes with larger pore sizes, namely to membranes used in microfiltration (MF), ultrafiltration (UF), and nanofiltration (NF). 

To understand how RO membranes can be transformed into those with smaller pore sizes, the structure of these membranes must be understood. A composite RO membrane usually consists of a thin aromatic polyamide layer(usually the active layer), with a porous polysulfone inner layer, and a polyester webbing. Plastics are also used in RO membranes such as polypropylene, polyester, acrylonitrile butadiene styrene, and fiberglass. These plastics’ uses include spacers, end-caps, and outer casing. Lastly, the glued parts (consisting of epoxy-related compounds) took up the smallest percentage of the weight of the membrane¹.

3.1 Transforming agents and testing conditions

Membrane conversion to those that are comparable to ultrafiltration(UF) products requires strongly oxidative chemicals to purposely degrade the polyamide active layer. This reveals the polyester and polysulfone support layers that are similar to UF materials. Therefore, the most effective oxidation chemical is investigated¹.

Sodium hypochlorite, hydrogen peroxide, and potassium permanganate were tested in 2 different conditions in a study investigated by Lawler et al.¹. Potassium permanganate, at first, was found to be the most efficient agent to transform old membranes. Studies from Rodriguez et al. and Veza et al. also corroborate this finding^{11 ,12}. Lawler et al. also investigated the testing of these transformed membranes in the tertiary treatment of wastewater. The membranes were able to remove up to 96% of suspended solids before RO treatment, albeit with high levels of fouling. The optimal dose of potassium permanganate was recirculating approximately 1000 mg/L for 1 or 2 hours.

However, there have been signs that sodium hypochlorite may be more effective than potassium permanganate. Lawler et al.¹ found that sodium hypochlorite had greater, more extensive degradation of the active layer when assessing the converted membranes’ permeability and salt rejection rather than the percentage of solids removed. The study suggests that using permanganate only allowed for a two times increase in permeability. In addition, sodium hypochlorite is more controllable than permanganate, as the permanganate ion is a very strong oxidizer. After transforming using this chemical, the salt rejection had dropped to 5%, compared to the 35-50% of directly reusing RO membranes, showing that NaClO can almost completely remove the polyamide layer, leading to UF-like filters. The converted membrane and commercially available membranes also demonstrated similar rejection performance. 

Many other studies also mention using sodium hypochlorite as an oxidative agent, such as in a study by Lejarazu-Larranaga et al. and Garcia-Pacheco et al.^{3 ,5}. In the latter, it was found that the longer the membranes were immersed in the solution of the transformation agent, which in addition to sodium hypochlorite included N-methyl-2-pyrrolidone(NMP) and acetone, the better the permeability of the resulting membranes. Acetone and NMP did not have substantial changes(more than 30% or more increase in permeability), so only sodium hypochlorite led to a noticeable increase in permeability: around 385% in 20 hours of immersion. The study by Veza et al. compiled data that suggested sodium hypochlorite was one of the better agents to peel the polyamide layer of the membrane¹². Interestingly, this study places sodium hydroxide as a better peeling/oxidative agent than sodium hypochlorite, with a simplified peeling effectiveness of 19.2% compared to the 17.3% of sodium hypochlorite. Figure 3 demonstrates that sodium hypochlorite had both lower peeling effectiveness and simplified peeling effectiveness. Peeling effectiveness is the fraction of feed water that permeates through, multiplied by the fraction of salt that passes through, divided by the pressure during the peeling. The higher the peeling effectiveness, the more effective the filter is. The simplified peeling effectiveness applies to membranes that were immersed in solution without pressure and is the peeling effectiveness without dividing by the pressure during the peeling.

One possible explanation of this finding could be that the sodium hypochlorite was recirculated and immersed, while the sodium hydroxide was completely recirculated, leading to more effective peeling. Another factor is that the dosage of sodium hydroxide was slightly higher at 7680 mg/L compared to the 6000 mg/L of sodium hypochlorite. A potential fault of the study could be that they investigated potassium permanganate in much further depth than either of the 2 aforementioned chemicals. 

In fact, there may be some problems with using potassium permanganate as an oxidative agent, even though it facilitated the most thorough removal of the polyamide active layer and had an adequate performance. In another study by Lawler et al. investigating the oxidative treatment of membranes, using potassium permanganate led to stable, high flux but lower salt rejection¹³. The most effective potassium permanganate treatment only led to a salt rejection reduction down to 85%. Instead, sodium hypochlorite in a particular study investigated had a much higher flux due to larger doses of the oxidative chemical. Sodium hydroxide and potassium permanganate had successful degradation of the membrane surface, but the flux was lower, possibly due to a manganese oxide layer forming. These findings regarding potassium permanganate are consistent with previous studies, which explains why several studies chose to use sodium hypochlorite instead of potassium permanganate. 

Storage of the membranes is also crucial to performance and flux. In the study by Lawler et al., it was found that permeability loss and flux decrease due to drying out were irrecoverable, although the salt rejection of the dried membranes was the same as if it had been stored wet¹³. This finding is shown in Figure 7, where it is clear that prewetting is required for high permeability. Some possibilities suggested include the collapse of pores through capillary forces, which causes rehydration to be very difficult, lowering hydraulic performance. Another possibility is the reduction of water-to-polymer interactions. The findings from Garcia-Pacheco et al. agree with the study that dry membranes result in lower average permeability and that keeping end-of-life membranes wet is crucial⁵. 

Referring back to the study of Lawler et al.¹, the method of transformation was also investigated. Active recirculation involves circulating the oxidative chemical around the used RO membranes, while passive immersion simply places the module in the oxidative solution. A study by Rodriguez et al. also looked at similar methods to clean and wash the membrane before transformation, by recirculation and immersion¹¹. 

3.2 Testing results

The results of many of these studies were successful and demonstrated the promise of using transformed membranes effectively. A parameter used to measure the effectiveness of the membranes is to measure turbidity, a measure of how clear water is based on scattered particles in the water. The units of turbidity are nephelometric turbidity units(NTU). The more particles in the water, the cloudier it appears and the higher the turbidity. Because turbidity is usually caused by suspended solids in water, it is an accurate method to assess the quality of transformed membranes. In the study by Veza et al., their feed water flow to the processed filters had turbidity values of 8 to 10 NTU reduced down below 0.6 NTU after the filtration. In the second filtration stage, turbidity was further reduced to 0.1 NTU¹². The suspended solids were 98% filtered, too. 

This high level of success was also present in other studies. In the study by Garcia-Pacheco et al., a common trend of longer transformation durations led to higher permeability rates, with nanofiltration properties being obtained after 36-hour to 122-hour exposure, and ultrafiltration properties being obtained after 242 hours of exposure⁵. In addition, the effect of pH was thoroughly investigated on the transformation process, and it was found that basic conditions were more efficient in obtaining better permeability for the membranes. Intriguingly, the study mentions sodium hypochlorite as a good cleaning agent for organic fouling, indicating another solution to minimize fouling. This study also studied the rejection of monovalent and divalent ions. It was found that a higher pH led to lower rejection for ions in general, and there was an inverse correlation between pH and ions rejected. Monovalent ions were rejected much less than divalent ions due to a difference in the magnitude of charge. At 122 hours of transformation duration and a pH of 10.5, all of the ions had less than 20% rejection. 

Overall, recirculation was a more effective method than passive immersion. This is supported by findings from multiple studies as well. For example, from Veza et al., the transformation of the membranes that only utilized passive immersion did not have a reportable peeling effectiveness, and the simplified peeling effectiveness was below 1%¹².

In the study by Rodriguez et al., there was another washing method used before the transformation process to remove fouling, which ultimately affected peeling effectiveness (PE)¹¹. Recirculation in a washing context meant having the washing solution recycled through the membrane for a duration of time(in this study, 18 hours) while membrane immersion soaked the membrane in the washing solution (soaked for 24 hours). Similarly to the transforming method of degrading the active layer on membranes, by recirculating the washing solution, which comprised sodium bisulfite, the membranes had higher PE than simply being immersed. Figure 8 supports this conclusion as the membranes that solely used the immersion method did not have comparable peeling effectiveness.

The transformed membranes needed to be compared to commercial membranes of that type, so for instance, comparing commercial UF membranes to transformed RO membranes. Lawler et al. found that the rejection performance and the molecular weight cut-off (MWCO) ranged from 10 to 100 kilodaltons, where daltons are a unit of measuring atomic mass and are defined as one-twelfth the mass of an atom of carbon-12 isotope, categorizing them as UF membranes¹. Veza et al. explain that the near absence of rejection of dextrose at 122-hour exposure time and a pH of 10.5 indicates that the transformed membrane behaves like ultrafiltration¹². Mohamedou et al. also found that RO membranes act like NF membranes with similar uses to NF applications¹⁴. Old membranes exhibited 2 times more hydraulic permeability than commercial membranes but only 35-50% salt rejection for 10 to 30 bar, where bar is a unit of pressure.

The most significant drawback was fouling. For Veza et al., the used membranes demonstrated rapid fouling, requiring frequent cleanings. They were cleaned using 2 methods: a short flushing with reject water and with chemical cleanings using chemical detergent every few days. Similarly, the study by Garcia-Pacheco et al. had one membrane that, despite showing nanofiltration membrane properties, ultrafiltration membrane properties were not achieved, regardless of how long the transformation process took⁵. The study attributes this to severe surface fouling of the membrane that could not be reversed. This severe surface fouling prevents the attack of sodium hypochlorite and thus prevents degradation of the polyamide and polysulfone layer. 

Another factor was the structure of the polyamide surface. The paper proposes doing initial chemical cleaning to remove the membrane fouling and allow for better interaction, but this has not been investigated. The study by Lawler et al. noted that fouling was not a major issue; however, the transformed membranes are still vulnerable to resistance increase¹³. Compared with a commercial UF membrane, attempting to reverse the fouling in the transformed membranes was not as effective.

3.3 Anion exchange membrane innovation

One article investigates the use of end-of-life RO membranes as a support for anion exchange membranes. Lejarazu-Larranaga et al.³ investigate using discarded RO membranes to support such exchange membranes in electrodialysis. By doing so, much of the waste in RO can be reused. The study follows the methods above to degrade the polyamide layer of the old membranes and obtain ultrafiltration and nanofiltration membrane properties. Still, it suggests that extremely damaged membranes, either chemically or mechanically, should be indirectly recycled, as it is more efficient and would be more useful in opening up applications of the old membranes. 

Electrodialysis (ED) is a separation technology driven by electricity with an anion and cation exchange stack that is charged, thus being able to remove ions from feed water. By processing the anion exchange resin first, casting it onto the recycled membrane, and then evaporating the solvent, the recycled anion exchange membrane (AEM) is set up optimally. Figure 9 shows this process, with the AEM being prepared using a doctor blade to cast the solution with different thicknesses, as well as the steps for solvent evaporation and phase inversion. The wet phase inversion is required to keep the membranes in a wet condition to prevent pore collapse. 

The study indicates that the permselectivity of the membranes increased as the evaporation time increased, with the permselectivity being on the same level as commercial AEM. However, when permselectivity was enhanced, electrical resistance also increased, possibly due to the formation of a denser membrane surface. Similar to the previous papers, the recycled RO membrane did not exhibit permselectivity, as ion rejection should decrease after stripping the active polyamide layer. Ultimately, the anion exchange resin cast onto the support gives the support new electrochemical and salt rejection properties. Later in the paper, when discussing diffusion coefficients, the recycled AEM also had high permselectivity and electrical resistance.

Lejarazu Larranaga et al. also critically investigate how useful a support structure is to the AEM and how it can improve mechanical properties. Therefore, the study also prepares a membrane without mechanical support to compare. Although the tensile strength of the recycled membrane support alone is much higher than the transformed recycled membrane(71 MPa to 17 MPa, where Pa represents Pascals, a unit of pressure), the transformed recycled membrane still had a higher tensile strength than the commercial AEM(with a tensile strength of 14 MPa) and the AEM without RO support membranes(9 MPa). This is attributed to enhanced structural stability of the membrane support, but could also be the membrane thickness and the polymer binder. These factors lead to a more stable AEM than commercial ones. 

Lastly, the recycled AEM was tested in desalination experiments in this study. The 2 experiments consist of the recycled AEM with commercial cation exchange membranes and the commercial cation and anion exchange membranes. The study utilizes feed spacers to promote turbulence in both types of experiments. As salt concentration decreased, the flux of ions decreased due to the absence of ions to carry electric charge. In addition, the percentage of salt removal for the recycled AEM was 84.51%, very close to the 85.08% of commercial AEM, meaning that utilizing recycled RO membranes as a support is plausible. However, recycled AEM took more time to produce the same volume of fresh water as the commercial counterpart due to high electrical resistance, taking 14.5 hours compared to the commercial’s 6.5 hours. 

Overall, the possibility of using recycled AEM for electrodialysis of brackish water desalination was confirmed and successful. The desalination rate is slightly higher than commercial membranes under the same conditions, but further work can be conducted to reduce the electrical resistance and to see if the recycled AEM has application in real scenarios. The paper presents an interesting idea to repurpose membranes that have proven to be effective. Instead of turning RO membranes into ultrafiltration or nanofiltration membranes, the idea of using the membrane as support was investigated and presents another avenue to reduce waste and recycle RO membranes. By transforming discarded RO membranes, not only have the waste issues been addressed, but they can also assist desalination technology. 

3.4 Applications of transformed membranes

Applications of transformed RO membranes include pretreatment of desalination technology or treatment of wastewater¹. These membranes would be able to filter suspended solids, large organics, and even some pathogens from feed water. Rural areas could utilize this water purification strategy, and gravity-driven filtration processes have also been considered using these recycled membranes. Thus, there is potential for these recycled membranes to be used in developing societies where large-scale industrial water treatment plants are not feasible. 

Another treatment application of transformed membranes is pathogen removal. Particles that had similarities to virus particles were used to test based on the convenience of manufacturing. The results and analysis of the permeate across cross-flow filtration showed that a converted RO membrane had rejections of citrate stabilized silver nanoparticles particles 99.5% and above, being comparable to a 10 kDa UF membrane. The log removal values of the converted membrane and the commercial membrane were 2.42 and 2.57, respectively. These values are a measure of how much of the particles were removed, with a value of 2 removing 99% of the particles and a value of 3 removing 99.9% of the particles. These particles were selected to represent viruses being filtered out. The particles were also similar to the physical traits of select waterborne pathogens, such as having similar size, shape, and charge. Therefore, pathogen removal could be a possible application of reused RO membranes, meaning transformed membranes can also fulfill the role of commercial membrane filtration of pathogens^{13 ,15}. More investigation could be done, however, to investigate the effectiveness of the membranes using genuine pathogens and viruses that are present in UF feed, as operating conditions can affect the log removal values.

Because these membranes have been shown to be promising and demonstrate effective filtration properties, they have several suitable applications. For example, water disinfection, industrial wastewater treatment, and residential wastewater treatment all utilize membranes with ultrafiltration or nanofiltration-like properties; thus, treating the wastewater with these membranes would reduce waste for RO processes while also conserving resources for the production of new commercial membranes¹³. NF-like membranes would be able to remove a large variety of pollutants in groundwater or surface water in water treatment, such as removing hard ions, nitrates, partial desalination, etc¹⁵. 

In addition, application in the food and pharmaceutical industry would not be sanitary and thus inapplicable. However, these membranes can be used in agriculture, especially for crops. Even after conventional treatment, high salinity makes wastewater unable to be used for reuse. Desalination of wastewater using the transformed membranes is then a viable strategy considering the amount of discarded RO membranes. Therefore, wastewater could be conditioned to be reused in irrigation systems as well as minimize the waste produced by RO facilities, effectively repurposing the membranes. Due to the variation in the transformed membranes, more testing will need to be done for specific membranes to determine whether or not they can successfully be utilized in a certain process. In addition, Mohamedou et al. propose that old RO membranes that act like NF membranes can be used for desalination pretreatment, selective demineralization of brackish water, or further uses in coral growth studies¹⁴. 

4. Future work

Transformed membranes hold promise for water filtration and demonstrate capabilities similar to those of commercial membranes. However, further study is required to implement the findings discussed. The next steps would be testing these membranes under real wastewater treatment applications and seeing if the large-scale transformation of membranes is viable. The experiments were done in a lab setting under controlled conditions; therefore, if the conditions of industrial treatment are applied, then the true effectiveness of the membrane can be evaluated. For example, in the experiment by Minier-Matar et al., the experiment performed used real process water from an RO facility to test in a lab setting, effectively mimicking the industrial scale of the experiment¹⁶. However, the studies by Garcia-Pacheco et al. and Veza et al. are pilot plants and only validate testing in a controlled setting^{5 ,12}. Thus, work could be done to establish the effectiveness of peeled membranes at a commercial level to effectively reduce waste.

More experiments could also be performed with kinetics because many of the studies investigated in this review used feed spacers to promote turbulence, but did not investigate using sonication as an effective antifouling method. In addition, the study by Wang et al. did not evaluate the performance of RO membranes¹⁰. A new experiment could be designed to assess the flux and fouling rates of RO membranes, as well as address some of the concerns associated with using sonication technology, such as the study using known foulants. The study could use silica and transparent exopolymer particles as test foulants, as these encompass a range of different foulants. Experiments regarding kinetics could also be used to address the lack of explanation for shear rates in membrane cleaning in the study by Matin et al.⁸. The kinetics experiment was shown to have significant shear rates due to the alternating flow(49% forward flow in a pulsating period and 24% backward). Therefore, an experiment could be conducted to investigate the effectiveness of biofouling mitigation through the use of sonication technology.

Additionally, modified experiments could be conducted, for instance, finding ways to reduce the electrical resistance of anion exchange membranes that use transformed reverse osmosis membranes³. Further experiments vary the evaporation time, the membrane’s thickness, considering the membrane’s activation, or using smaller particles when creating the polymeric solution. Lastly, there was not much attention paid to the fouling of transformed membranes. Although the fouling of the membranes was easily cleared, they still posed a problem. In the study by Lawler et al., membranes are still vulnerable to fouling and resistance increases compared to commercial ultrafiltration membranes and are not as responsive to cleaning solutions¹³. Further investigation could be done to analyze how chemical cleaning on these membranes is different from commercial membranes and how fouling occurs with transformed membranes.

Conclusion

In this study, a multitude of ways to repurpose and reuse reverse osmosis(RO) modules were considered. Many methods already exist to recycle these membranes, and the study characterized and investigated the different types of fouling that occur with RO modules and their prevention, as well as the transformation of these membranes to use them in another application, thus minimizing waste both in the creation and disposal of membranes. 

The types of fouling, such as inorganic fouling, biofouling, and organic fouling, were discussed in detail, mentioning the causes, types, structure, and prevention, as well as how influential that type of fouling was in preserving RO membranes. A variety of preventative methods were discussed, primarily pretreatment to avoid damaging or significantly altering the RO membranes, but including investigations into other aspects such as pH and kinetics. These investigations help to diagnose the properties of fouling to better acknowledge prevention methods. The methods of prevention differed for each type of fouling and, therefore, must be considered in pretreatment for RO filtration systems. 

In addition, the process of transformation was discussed in great detail. Many of the studies investigated promoted potassium permanganate as the most effective transforming agent, but sodium hydroxide and sodium hypochlorite were also effective in transformation. The method of transforming, as well as its conditions, were evaluated to conclude the optimal transformation methods. The results of the experiments for the transformed membranes were also promising, showing that transformed RO membranes can function with similar effectiveness as commercial nanofiltration membranes or ultrafiltration membranes, depending on the testing conditions. The fouling present in the membranes was not significant, but the effects were more noticeable in the cleaning processes than in commercial membranes. Discarded RO membranes were also considered for anion exchange membrane creation and were deemed successful, although with the drawback of higher electrical resistance compared with commercial ones. 

References

1. Lawler, W., Bradford-Hartke, Z., Cran, M. J., Duke, M., Leslie, G., Ladewig, B. P., & Le-Clech, P.  (2012). Towards new opportunities for reuse, recycling and disposal of used reverse osmosis membranes. Desalination, 299, 103–112. https://doi.org/10.1016/j.desal.2012.05.030 [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩]
2. Sun, X., Duan, L., Liu, Z., Gao, Q., Liu, J., & Zhang, D. (2024). The mechanism of silica and transparent exopolymer particles (TEP) on reverse osmosis membranes fouling. Journal of Environmental Management, 349. https://doi.org/10.1016/j.jenvman.2023.119634 [↩] [↩] [↩] [↩] [↩] [↩]
3. Lejarazu-Larrañaga, A., Molina, S., Ortiz, J. M., Navarro, R., & García-Calvo, E. (2020). Circular economy in membrane technology: Using end-of-life reverse osmosis modules for preparation of recycled anion exchange membranes and validation in electrodialysis. Journal of Membrane Science, 593. https://doi.org/10.1016/j.memsci.2019.117423 [↩] [↩] [↩] [↩] [↩]
4. Lawler, W., Bradford-Hartke, Z., Cran, M. J., Duke, M., Leslie, G., Ladewig, B. P., & Le-Clech, P.  (2012). Towards new opportunities for reuse, recycling and disposal of used reverse osmosis membranes. Desalination, 299, 103–112. https://doi.org/10.1016/j.desal.2012.05.030 [↩] [↩]
5. García-Pacheco, R., Landaburu-Aguirre, J., Molina, S., Rodríguez-Sáez, L., Teli, S. B., & García-Calvo, E. (2015). Transformation of end-of-life RO membranes into NF and UF membranes: Evaluation of membrane performance. Journal of Membrane Science, 495, 305–315. https://doi.org/10.1016/j.memsci.2015.08.025 [↩] [↩] [↩] [↩] [↩] [↩]
6. Minier-Matar, J., AlShamari, E., Raja, M., Khan, F., Al-Maas, M., Hussain, A., & Adham, S. (2024). Detailed organic characterization of process water to evaluate reverse osmosis membrane fouling in industrial wastewater treatment. Desalination, 572. https://doi.org/10.1016/j.desal.2023.117128 [↩] [↩] [↩] [↩] [↩] [↩]
7. Sun, J., Chen, S., Dong, Y., Wang, J., & Nie, Y. (2023). Removal of hardness and silicon by precipitation from reverse osmosis influent: Process optimization and economic analysis. Journal of Water Process Engineering, 55. https://doi.org/10.1016/j.jwpe.2023.104147 [↩] [↩] [↩]
8. Matin, A., Khan, Z., Zaidi, S. M. J., & Boyce, M. C. (2011). Biofouling in reverse osmosis membranes for seawater desalination: Phenomena and prevention. In Desalination (Vol. 281, Issue 1, pp. 1–16). https://doi.org/10.1016/j.desal.2011.06.063 [↩] [↩] [↩] [↩] [↩] [↩] [↩]
9. Rho, H., Yu, P., Zhao, Z., Lee, C. S., Chon, K., Perreault, F., Alvarez, P. J. J., Amy, G., & Westerhoff, P. (2022). Inhibition of biofouling on reverse osmosis membrane surfaces by germicidal ultraviolet light side-emitting optical fibers. Water Research, 224. https://doi.org/10.1016/j.watres.2022.119094 [↩]
10. Wang, J. M., Jin, Q. Q., Zhang, Y. Y., Fang, H. C., & Xia, H. M. (2021). Reducing the membrane fouling in cross-flow filtration using a facile fluidic oscillator. Separation and Purification Technology, 272. https://doi.org/10.1016/j.seppur.2021.118854 [↩] [↩] [↩]
11. Rodriguez, J. J., Jimhez, V., Trujillo, O., & Vezab, J. M. (2002). Reuse of reverse osmosis membranes wastewater treatment in advanced. In DESALINATION (Vol. 150). www.elsevier.com/locate/desal [↩] [↩] [↩] [↩]
12. Veza, J. M., & Rodriguez-Gonzalez, J. J. (2003). European Desalination Society, International Water Association. In Desalination (Vol. 15). www.elsevier.com/locate/desal [↩] [↩] [↩] [↩] [↩] [↩] [↩]
13. Lawler, W., Antony, A., Cran, M., Duke, M., Leslie, G., & Le-Clech, P. (2013). Production and characterisation of UF membranes by chemical conversion of used RO membranes. Journal of Membrane Science, 447, 203–211. https://doi.org/10.1016/j.memsci.2013.07.015 [↩] [↩] [↩] [↩] [↩] [↩] [↩]
14. Ould Mohamedou, E., Penate Suarez, D. B., Vince, F., Jaouen, P., & Pontie, M. (2010). New lives for old reverse osmosis (RO) membranes. Desalination, 253(1–3), 62–70. https://doi.org/10.1016/j.desal.2009.11.032 [↩] [↩]
15. van der Bruggen, B., & Vandecasteele, C. (n.d.). Removal of pollutants from surface water and groundwater by nanofiltration: overview of possible applications in the drinking water industry. www.elsevier.com/locate/envpol [↩] [↩]
16. Minier-Matar, J., AlShamari, E., Raja, M., Khan, F., Al-Maas, M., Hussain, A., & Adham, S. (2024). Detailed organic characterization of process water to evaluate reverse osmosis membrane fouling in industrial wastewater treatment. Desalination, 572. https://doi.org/10.1016/j.desal.2023.117128 [↩]