Comparative analysis of performance and fouling characteristics of microfiltration and ultrafiltration polycarbonate membrane

Seawater Reverse Osmosis is the most popular desalination technology for providing clean water. However, several problems in SWRO operations occurs, namely the decrease in membrane performance due to fouling. Fouling on the membrane is generally caused by high salinity and organic content in seawater. Therefore, pre-treatment technology is needed to improve water quality and reduce the workload of SWRO. This study aims to determine the suitable pre-treatment technology, by examining the removal efficiency of parameters in water using Ultrafiltration and microfiltration membranes. In this study, feed water was obtained from treated seawater. The experiment employed an average pore size of 0.01 micron polycarbonate track etched (PCTE) ultrafiltration membrane and 0.2 micron polycarbonate (PC) microfiltration membrane, respectively, with a dead-end filtration method and constant flux values at 60 L/m2.h and 120 L/m2.h. The choice of polycarbonate membrane is based on several advantages, such as high durability and chemical resistance. Water quality parameters such as turbidity, total dissolved solid (TDS), conductivity, dissolved oxygen (DO), organic substances (UV-Vis), and chemical oxygen demand (COD) were observed to determine the performance of each membrane types. The results showed that the operation of ultrafiltration membranes able to remove high amount of turbidity and COD with 88 ± 4 % and 86 ± 12 % removal percentage. Moreover, lesser removal efficiency was found for DO, TDS, UV-Vis and conductivity employing ultrafiltration membrane. In comparison to microfiltration, Ultrafiltration membrane was revealed as promising pretreatment for SWRO with higher retention of measured parameters and better membrane filtration performance.


Introduction
Access to clean water supply become a main concern in a large city in Indonesia due to high population growth.The water demand of DKI Jakarta is revealed to be 547.5 million cubic meters per year (150 liters per day).Nevertheless, only 54% of the population can be fulfilled by the regional water utility (PDAM).This situation led residents to seek alternative sources of clean water, which tend to be unsafe due to poor quality.Additionally, the extensive infrastructure development and waste disposal in DKI Jakarta have caused significant water pollution.About 80% of the groundwater in the Jakarta Aquifer Basin does not meet the standards set by the Minister of Health regarding Drinking Water Quality Requirements.Moreover, using groundwater as a water supply resulted in greater problems such as land subsidence.
The northern part of Jakarta is the most contaminated area in terms of groundwater due to high levels IOP Publishing doi:10.1088/1755-1315/1263/1/012058 2 of iron (Fe), sodium (Na), chloride (Cl), total dissolved solids (TDS), and electrical conductivity influenced by seawater intrusion.Therefore, one possible technology for providing clean water in coastal areas is seawater desalination technology.Seawater desalination using reverse osmosis membrane is a process that involves high-pressure pumping to pass seawater through semi-permeable snot only promises a stable operation but also high separation performances [2].As a result, Seawater reverse osmosis (SWRO) has gained the limelight technology to produce fresh water and holds roughly a 69% market share among all desalination technologies [3].
There are several process treatments in SWRO including Coarse Screen, Dissolved Air Flotation, Automatic Filter Screen and Ultrafiltration (UF), and Seawater Reverse Osmosis.However, several factors affect the energy consumption of the RO unit, such as the salinity of feed water and the recovery rate of the system [4].Continuous operation can also cause problems in the form of membrane fouling, a process by which the particles, colloidal particles, or macromolecules are deposited or adsorbed onto the membrane pores or surface by physical and chemical interactions or mechanical action, which results in smaller or blocked membrane pores [5].This phenomenon can decrease the quality of feed water, or in a larger scale, higher operating cost because of chemical cleaning or membrane replacement.
Successful operation of a SWRO plant depends on the ability of the pretreatment system to maintain the feedwater quality before RO process, thus membrane-based pretreatment processes such as microfiltration and ultrafiltration have found practical application worldwide [6].Compared to microfiltration, ultrafiltration membrane can remove smaller particles because of the smaller pore size and silt density index (SDI) value, which makes it an ideal SWRO pretreatment technology [7].Ultrafiltration membrane also requires comparable operating pressure than microfiltration membrane.
Over 90% of UF and MF membranes are made up of polymeric materials, many of which have successfully been commercialized [8], such as polycarbonate.Polycarbonates (PCs) are polyesters of carbonic acid, specially derived from phosgene or diphenyl carbonate [9].PCs are tough and glassy thermoplastic polymers of particular interest for separation [10].Polycarbonate membrane filters offer several advantages, including high chemical resistance, broad compatibility with various solvents, good thermal stability, and easy handling.To some extent, the smooth and flat surfaces give an advantage for PC membranes as anti-fouling performance, because foulants would not be trapped inside compared to the rough membrane surfaces [11].
This paper is aimed to investigate membrane performances including filtration performance as well as removal efficiency of 0,01-micron average pore size polycarbonate track-etched ultrafiltration membrane and 0,2-micron average pore size polycarbonate microfiltration membrane.A real treated seawater are filtered through the membrane operated in dead-end filtration mode with constant flux operation method at Lab scale experiment.The experiment is performed to compare the performance of both membranes for the suitable pretreatment technology for SWRO.

Feed water characterization
Feedwater was collected using the grab sampling method at the outlet of the dissolved air flotation (DAF) unit from SWRO treatment plant.Feed water characteristics then measured by various parameters.Organic parameters including COD and UV-Vis was determined using a spectrophotometer (DR2000, Hach, USA) and ultraviolet-visible spectrophotometer (DR6000, Hach, USA), respectively.Saline parameters (TDS and Conductivity), temperature, and Dissolved Oxygen (DO) are monitored by multi-parameter meter (HQ40d, Hach, USA).Other sum parameters (Turbidity and pH) were measured by turbidity meter (TU-2016, Lutron, Taiwan) and pH meter (HI 98107, Hanna, USA), respectively, as an initial study of the existing conditions related to fouling issues.Feed waters were collected twice during the periods of March -May 2023.Summary of feed water characteristics are shown in Table 1.

Membrane characterization
The polycarbonate track etched (PCTE) membrane (Sterlitech Corporation) and Polycarbonate (PC) membrane (Isopore) with average pore size of 0.01 micron and 0,2 micron pore size with a diameter of 47 mm were used, respectively.Prior experiment, membrane was soaked in a NaOCl 200 ppm solution for 24 hours to eliminate material which has adsorbed onto the membrane that cannot be removed hydraulically during production process [12].Residual ions could potentially affect the experimental results.Soaking was also conducted to open the membrane pores, preventing spontaneous fouling during the experiment.Furthermore, the membrane was soaked in distilled water to neutralize the membrane's pH and to remove any remaining particles on the membrane surface from the soaking process.

Experiment setup
Reactor for lab-scale filtration experiments (see.figure 1) consists of feed tank with a magnetic stirrer for homogenous solution.BT300-2J peristaltic pump is then used to transfer feed water to a 47 mm Millipore membrane filter, consisting of a membrane holder for placing flat sheet membranes, which uses a dead-end flow system.This setup was used to determine the membrane's efficiency in removing specific parameters from the water sample that will pass through the membrane layer.Peristaltic pump speed is set to 2 rotation per minute equal to the 120 L/m²H operational flux, and the experimental pressure i monitored by a pressure gauge placed between the peristaltic pump and membrane holder.At the permeate side, the volume of permeate water was measured using digital scale with a precision of 0.01 gr.The experiments use room temperature (26ºC ± 2) and run until the permeate water stops flowing as an indication of membrane fouling.4

Filtration experiments
Both microfiltration and ultrafiltration membrane were conducted using the same methods of experiment.A pure water permeability test was conducted before the filtration process with distilled water for 5 minutes to clean the membrane after the soaking process.The feed water was pumped to membrane holder through the membrane.Several data such as volume, time, and pressure are recorded until experiments are finished every 1 minutes at early phase and 5 minutes at later stage.Operational at constant flux determines the amount of permeate that can pass through the membrane per unit area per unit of time.Pressure represents the magnitude of the driving force required to pass the feed water through the membrane.Time indicates the duration of membrane operation until the membrane experiences fouling under specific operating conditions.Subsequently, permeate water characteristics are measured using the same methods as feed water characteristics.Membrane performances are subsequently determined with equations from Table 2.
Table 2. Equations for membrane performance analysis.

Equation Description
Membrane Retention: CP: solute concentration in permeate, and CF: solute concentration in feed solution.
Membrane Permeability: J: water flux, ΔP: transmembrane pressure, where Pfeed is the pressure on the feed side and Ppermeate are pressure on the permeate side.W: permeability Normalized Permeability W': normalized pure water permeability ratio, WVsp: specific filtered volume permeability, W0: virgin membrane permeability.

Membrane retention
Microfiltration and ultrafiltration membrane retention are shown ini Table 3.In general, ultrafiltration membrane can achieve a better retention rate compared to microfiltration membrane.This phenomenon is prevalent in organic parameters, such as turbidity with 87,64% retention rate.Previous studies found that in the removal of turbidity and particulate matter, the UF process has a higher removal rate than the conventional process, with effluent turbidity is stable below 0.1 NTU, and the removal rate of particulate matter is up to 99.9% [13].Higher retention rates can also be described as an effect of the smaller pore size of ultrafiltration membrane.However, the similar retention rates in MF membrane are in-line with previous studies, which found that 95% removal of turbidity from seawater can be achieved by ceramic membranes, with similar properties compared to polycarbonate membranes, mainly higher chemical and mechanical strength [14].Smaller pore size enables UF membranes to remove microscopic compounds and other smaller substances in the feed water, as well as small organics matter which is represented by 91,83% COD removal.A previous study found that UF was more effective than MF in removing COD from water, with a removal rate of 99% compared to 75% for MF [15].In comparison, UV274 have a less number, with 38 % and 41% retention rate for UF and MF respectively.Slight decrease in UV274 is a result of removal of some organic matters, thus reducing organic matter ability to absorb UV light in water [16].
Saline parameters such as TDS are less affected by porous membrane filtration, with 15,59 % and 15,16 % retention rate for UF and MF processes respectively.Because of the small size of salt ions, UF membranes are unable to completely removed TDS components from water.TDS represents substances present in a liquid in molecular, ionized, or micro-granular (colloidal sol) suspended form that is present in water, and while certain organic particles can be removed, certain inorganic substances such as salt, fluoride, and other ions are able to pass through because of the smaller size in comparison to UF membrane pore and can only be completely removed by RO membrane.While previous study found that the TDS removal rate increased significantly as the pore size of the membrane decreased, an insignificant difference in membrane pore size between UF and MF membrane results in a similar retention rate.

Membrane Performance
Investigation the performance of both microfiltration (MF) and ultrafiltration (UF) membrane was performed by correlation of normalized permeability with filtered specific volume.Fluctuations in membrane fouling rate curves are a result of unstable filtration condition measurements, mainly the unstable digital pressure gauge reading.Normalized permeability refers to a dimensionless measure of the permeability of the membrane.Specific volume refers to the volume of permeate produced in certain period.Membrane performance can be observed from the filtration curve (see Figure 2 and 3).It could be seen that different filtration curves declined for both MF and UF.In general, filtration curve of feed water using UF membrane revealed a longer performance with higher filtered volume in comparison to filtration using microfiltration membrane.The severe of MF membrane fouling was caused by pore blocking associated with large (macromolecular) hydrophilic molecules and/or organic colloids [17][18].
Furthermore, to understand fouling behavior.investigation of filtration curves was conducted by separating the curves into stages [19][20][21].The idea of separating the curves into two stages based on previous study.In a filtration cycle, foulant particles had tendency to either deposited on membrane surface or inside the membrane explained as an early phase filtration cycle followed by deposition on top of the foulant at later stages that well-known as Cake layer formation [19].
In this study, as a first stage, a linear permeability decline of 0,55 with the filtered specific volume of 150 L/m² in the case of UF membrane (Figure 2) was seen.On the other hand, a steep decline of permeability In case of UF, the permeability decline of 0,55 with a specific filtered volume of 75 L/m² was observed.Despite similar permeability decline, the ability of membrane filtration of MF was found to be less than almost 2 times compared to UF.Moreover, in the later stage, similar to the initial stage, linear permeability decline of UF membrane with the decline from 0,45 to 0,2 was observed, while steep decline of permeability of MF membrane from 0,45 to 0,05 was identified.The steep decline might indicate a fast cake layer and strong formation that blocked the filtration process and decrease the ability of the membrane to filter the water.In comparison between the first and second stage, it could be seen that earlier stages of fouling tended to be faster and sever fouling in early stages rather than cake layer phase [20] [21].

6
These results are in line with previous studies, which stated that in PCTE membrane, filtration transformed the fouling mechanism from membrane blocking to cake filtration at a lower filtration flux or lower particle accumulation for the membrane with a larger pore size or under a lower filtration pressure, thus MF membrane is less resistant to fouling and particle deposition [22].Results from both graphs can be concluded by higher Intermediate blocking fouling rate than cake filtration fouling rate, which contributes to the higher extent of irreversible fouling [20].
In correlation with membrane productions, better performance of UF membrane can also be attributed to the manufacturing process of both membranes.UF membranes use track-etched method, which create an evenly distributed pore size and shape, which on average are circular shaped [23].That circular shape creates a smoother surface with better antifouling properties, in contrast to MF membrane with larger pore size but with an uneven distribution, thus creating an unstable filtration performance that leads to foulant deposition.

Conclusion
Comparative studies of membrane performance between microfiltration and ultrafiltration membrane performance as a pre-treatment method in SWRO found that both ultrafiltration and microfiltration membrane can achieve a significant organic parameters retention rate with 87% for turbidity and 91 % for COD removal rate with better removal of conductivity and TDS in range between 15-38 %.Furthermore, related to membrane performance, ultrafiltration membrane revealed less membrane fouling and better membrane performance.

Figure 1 .
Figure 1.Schematics of lab-scale constant flux membrane filtration experiments.

Table 3 .
Removal efficiency of microfiltration and ultrafiltration membrane.