Effect of Phosphate on Ultrafiltration Membrane Performance After Predeposition of Fe 3 O 4

Abstract

The effect of predeposition of Fe₃O₄ on ultrafiltration (UF) processes has been widely studied in recent years. However, there have been no reports of effects of phosphate on the predeposition process. In this article, the fluxes of humic acid (HA) solutions filtered by Fe₃O₄+UF and Fe₃O₄+H₂PO₄⁻+UF were studied and filtered water was analyzed. Results indicate that the optimum concentration of Fe₃O₄ for the predeposition process was 0.45 g/L. Filtration fluxes of HA solutions using UF, Fe₃O₄+UF, and Fe₃O₄+H₂PO₄⁻+UF were 0.32, 0.52, and 0.41, respectively, after 10 min. Removal efficiencies of HA filtrated by UF, Fe₃O₄+UF, and Fe₃O₄+H₂PO₄⁻+UF were 32.6%, 41.1%, and 35.8%, respectively, according to dissolved organic matter results. After Fe₃O₄ adsorbed phosphate, the effect of Fe₃O₄ predeposition on removal of high and medium molecular weight HA was weaker. However, Fe₃O₄ predeposition enhanced removal of low molecular weight HA. Scanning electron microscopy and atomic force microscopy demonstrated that phosphate decreased compactness of the Fe₃O₄ predeposition layer. These data provide new insights into use of Fe₃O₄ predeposition in UF applications.

Introduction

Ultrafiltration (UF) is one of the most widely used water treatment technologies in the 21st century and has many advantages, including simplicity, compact size, product quality, and pathogen removal (Bai et al., 2013; Tian et al., 2013). However, the widespread application of UF technology is limited by membrane fouling, which constitutes a major bottleneck (Li et al., 2014; Liu et al., 2014). As the main component of organic matter, humic acid (HA) is considered to be an important contributor to membrane fouling, and its mechanism has been widely studied (Kumar et al., 2016; Xie et al., 2017). To effectively reduce membrane fouling, many methods have been intensively studies, including membrane modification and process improvement. Among these combination processes using UF technology in tandem with other purification methods are promising approaches. In recent years, some scholars have proposed that predepositing adsorbents onto the surface of the UF membrane to form a protective layer could not only ensure the efficient removal of pollutants but could also alleviate the problem of membrane fouling (Zhao and Wang, 2016). One study showed that precoating can effectively filter or retain pollutants; however, the effect of precoating on membrane fouling was limited (Ma et al., 2015). At present, the main adsorbents used in predeposition technology are carbon nanotubes (Ajmani et al., 2012), magnetic ion resin (Yiantsios et al., 2001), iron nanoparticles (Ma et al., 2015), activated carbon (Zhang et al., 2003), quartz sand, aluminum oxide particles (Campinas and Rosa, 2010), and iron oxide particles (Yu and Graham, 2015).

In the past, ferric coagulants were frequently used in predeposition studies due to their low price and high efficiency. Ferric salts have many advantages, such as a high specific surface area, low cost, and the ability to absorb a variety of organic and inorganic substances (Kim et al., 2008). In addition, ferric salts have the advantage of easy separation and recycling (Campinas and Rosa, 2010; Yu and Graham, 2015), especially because magnetic iron adsorbents can be effectively separated by application of an external magnetic force. To our knowledge, research studies on the influence of inorganic ions in raw water on the predeposition process have not been performed in any detail. In particular, phosphate levels are high in secondary effluent from wastewater treatment plants and the effect of phosphate on predeposition has not been well studied. Some scholars believe that there is competition between phosphate and organic molecules for adsorption to iron, aluminum, and other positively charged metals because of their similar negative charge (Antelo et al., 2007; Qin et al., 2012). Studies have also confirmed the competitive adsorption between phosphate and HA for metal ions (Fu et al., 2013).

In this study, we examined the differences in Fe₃O₄ adsorption in the presence and absence of phosphate during the predeposition process and its effects on membrane fouling. HA removal efficiencies under these two conditions were investigated using synthetic wastewaters in the laboratory as an example. Phosphate inhibited Fe₃O₄ absorption under our experimental conditions. The mechanism underlying the influence of phosphate on Fe₃O₄ predeposition process was examined using multiple approaches. The implications of our findings are discussed.

Experimental protocols

Chemical agents and materials

Chemical reagents used in this study were humic acid (HA; Aldrich), Fe₃O₄, KH₂PO₄, KNO₃, NaOH, NaNO₃, and HNO₃. Most of the reagents were pure (Analytical Reagents), except HA (chemical pure). Fe₃O₄ was purchased from Beijing Dk Nano technology Co., Ltd., and its specific physical parameters were as follows: color: black; content: 99.9%; particle morphology: spherical; density: 4.8–5.1 g/cm³.

In experiments, 1 g/L HA solution was prepared by dissolving a certain amount of HA in deionized (DI) water. Then the HA solution was filtered using a 0.45 μm pore size filter membrane and stored at 4°C as a reserve liquid. Polyvinylidene fluoride [PVDF; Ande Membrane Separation Technology & Engineering (Beijing) Co., Ltd, China] flat sheet UF membranes (100 kDa cutoff) were used and their characteristics are shown in Table 1.

Table 1.

Membrane Parameter

Index	Parameter values
Temperature/°C	5–38
pH	2–13
Cutting molecular weight/Da	100,000
Effective filtration area/m²	0.45
Working pressure/MPa	0.35

UF experiments

Each membrane was soaked in DI water for at least 24 h to remove impurities. The filtration process is shown in Fig. 1. The filtration pressure provided by nitrogen gas was maintained at 0.10 Mpa. Raw water was filtered through the membrane fixed at the bottom of a stirred UF cell and the filtered water was collected in a beaker placed on an electronic balance. The weight data were recorded every minute and transferred to a computer.

FIG. 1.

Schematic diagram of ultrafiltration test device.

Before the start of the experiment, 350 mL DI water was filtered through the membrane to test the DI water flux and the stable filtration flux was recorded as J₀. In the experiment, the filtration flux of HA was recorded as J and normalized flux J/J₀ as a function of time was used to reflect the change of filtration flux.

Process 1 (UF)

First, the UF membrane was soaked in DI water for 24 h, and then 350 mL pure water was filtered to measure the membrane flux J₀. After that, 350 mL 20 mg/L HA solution was filtered through the membrane.

Process 2 (Fe₃O₄+UF)

17.5 mg Fe₃O₄ was weighed and put in a beaker containing 350 mL DI water. The beaker was shocked by an ultrasonic washer for 5 min, and then the pH of the solution in beaker was adjusted to 7.0. The solution was poured into the UF stirred cell and was stirred for 1 min using nonmagnetic stirrer. After that, the cell was left standing still for 10 min and then 7 mL 1 g/L HA solution was dripped into the cell. Finally, the cell was placed in a stable place for 30 min and then the solution in the cell was filtered. The other five groups in these experiments were the same as above (using 52.5, 157.5, 472.5, 1417.5, and 4252.5 mg Fe₃O₄).

Process 3 (Fe₃O₄+H₂PO₄⁻+UF)

157.5 mg Fe₃O₄ was weighed and poured into a beaker containing 350 mL phosphate solution (0.05 mmol/L). The pH of the solution was adjusted to 7.0 and the solution was stirred for 4 h using a nonmagnetic stirrer. The solution was then poured into a stirred UF cell and stirred for 1 min using a nonmagnetic stirrer. The cell was then stably placed still for 10 min and 7 mL 1 g/L HA solution was dripped into the cell. Finally, the cell was placed in a stable place for 30 min and the solution in the cell was filtered. All filtration experiments were repeated thrice and the mean values were calculated.

Analytical methods

A pH meter (PB-10; Sartorius, Germany) was used to monitor pH throughout the experiment. A stirred UF cell (8400; Millipore Co.) was used in the UF experiments. Variations in UF membrane flux as a function of time were recorded by a data logger (ME2002E; Mettler Toledo, Switzerland and USA). Samples were stirred using a ZR4-6 coagulation test mixer (Zhongrun Water industry Technology Development Co., China). The molecular weight distribution of HA was determined by gelpermeation chromatography (GPC; Agilent Technologies; Detector:UV₂₅₄; Column: TSK; Temperature: 25°C). The samples were dispersed using an ultrasonic cleaner (YKTD-240W; YDTD, China). Zeta potential was measured using a zeta electric potential analyzer (Zetasizer NanoZS90, England). The microstructure of Fe₃O₄ was measured by a specific surface analyzer (BELSORP-mini II; NIKKISO GROUP, Japan). The particle size distribution was measured using a laser particle size analyzer (MS2000 Hydro 2000 MU, Malvern, England). The dissolved organic content of water samples was measured using a total organic carbon analyzer (TOC-VCPH; Shimadzu Co., Japan). The predeposited membranes were scanned using a scanning electron microscope (SEM) (S-3000N; Hitachi High-Technologies Co., Japan) and a high-resolution atomic force microscope (AFM) (Multimode-8; Bruker Co.).

Results and discussion

Flux decline caused by a critical dose of adsorbents

Figure 2 revealed the effect of Fe₃O₄ levels on the flux of HA solution filtered by Fe₃O₄+UF. When the Fe₃O₄ concentration was 0.05, 0.15, 0.45, 1.35, and 4.05 g/L, the filtration flux of HA solution was 2.9%, 5.4%, 20.1%, 16.3%, and 0.5% higher, respectively, after 10 min compared with that filtered by UF. Thus, there was an optimum concentration (0.45 g/L) for Fe₃O₄ predeposition, which alleviated membrane fouling and enhanced the filtration flux.

FIG. 2.

Normalized flux declines when filtering raw water with predeposited Fe₃O₄ at different concentrations.

Filtration fluxes of HA solution under different conditions (UF, Fe₃O₄+UF, and Fe₃O₄+H₂PO₄⁻+UF) were investigated according to the analysis shown in Fig. 2. The results are shown in Fig. 3.

FIG. 3.

Normalized flux decline when filtering raw water by different processes.

This shows that the filtration fluxes of HA solution were Fe₃O₄+UF>Fe₃O₄+H₂PO₄⁻+UF>UF. The filtration fluxes of HA solution filtrated by UF, Fe₃O₄+UF, and Fe₃O₄+H₂PO₄⁻+UF were 0.32, 0.52, and 0.41, respectively, after 10 min. The filtration of HA by UF decreased by 68.3% after 10 min, which demonstrated that the HA solution had seriously contaminated the filtration membrane. The filtration flux of HA by Fe₃O₄+UF was 20.1% higher than that filtered by UF after 10 min. This demonstrated that the predeposition process effectively improved the filtration flux and alleviated membrane fouling. The effect of Fe₃O₄ predeposition on HA filtration was consistent with previous studies (Schippers et al., 2001; Harman et al., 2010; Malczewska et al., 2015). After Fe₃O₄ adsorption of phosphate for 4 h (0.71 mg/g), the filtration flux of the HA solution by Fe₃O₄+H₂PO₄⁻+UF was higher than that filtered by UF, but it was lower than that filtered by Fe₃O₄+UF. After 10 min, the filtration flux of HA solution by Fe₃O₄+H₂PO₄⁻+UF was 10.8% lower than that filtered by Fe₃O₄+UF. It was speculated that phosphate occupied adsorption sites on the surface of Fe₃O₄ particles and the adsorption capacity of Fe₃O₄ on HA molecules was therefore weakened. In conclusion, the Fe₃O₄ predeposition process alleviated membrane fouling and its effect was weakened after adsorption of phosphate.

Particle size analysis

The microstructure of Fe₃O₄ was then measured with the following results: mean particle size: 201.32 nm; specific surface area: 10.71 m²/g; total pore volume: 0.02 mL/g; and mean pore size: 8.37 nm. The particle size distribution of Fe₃O₄ is presented in Fig. 4.

FIG. 4.

Fe₃O₄ particle size distribution.

The mean Fe₃O₄ particle size was 201.32 nm, which was small enough to make the deposition layer dense and was favorable for the interception of HA molecules. However the Fe₃O₄ particle size distribution was not consistent with this. There were five peaks at 0.112, 0.562, 2.818, 17.783, and 100 μm, respectively. The average diameter of Fe₃O₄ particles was 56.17 μm. It was inferred that Fe₃O₄ particles must have joined with one another, which made the measured particle size larger than the actual size. The specific surface area of Fe₃O₄ and its total pore volume were not large, which indicated that the adsorption ability of Fe₃O₄ was not strong. The mean pore size of Fe₃O₄ was 8.37 nm, which is large enough to retain HA molecules. The efficiency of the predeposition process on HA removal depended on its adsorption function and interception function.

Effect of different processes on removal of organic compounds

Analysis of dissolved organic carbon in the water treated by different processes was carried out to identify the influence of phosphate on the filtration of HA as it relates to the predeposition process. The results are shown in Table 2.

Table 2.

Humic Acid Removal Efficiencies by Different Treatment Processes

Filtration process	DOC/(mg/L)	Removal rate/%
Raw water	10.85	—
UF	7.31	32.6
Fe₃O₄+UF	6.39	41.1
Fe₃O₄+H₂PO₄⁻+UF	6.97	35.8

Fe₃O₄ dosage = 0.45 g/L, H₂PO₄⁻ = 1.55 mg/L, HA = 20 mg/L, pH = 7.0.

DOC, dissolved organic matter; HA, humic acid; UF, ultrafiltration.

Removal efficiencies of HA filtered by UF, Fe₃O₄+UF, and Fe₃O₄+H₂PO₄⁻+UF were 32.6%, 41.1%, and 35.8%, respectively. The predeposition process of Fe₃O₄ effectively improved HA removal efficiency and alleviated membrane fouling. Phosphate weakened the effect of Fe₃O₄ predeposition on the efficiency of HA removal.

To further analyze the effect of the predeposition process on the removal of HA with different molecular weights, water treated by different processes was analyzed by gel chromatography. The results are displayed in Fig. 5.

FIG. 5.

Molecular weight distribution of raw water before and after treatment by different filtration processes.

The molecular weight distribution of HA changed after treatment with different processes. The peak value of HA in raw water was 10,671 Da and was reduced to 10,433 Da after UF, which indicated UF efficiently removed high molecular weight HA. The peak value of HA after Fe₃O₄+UF was 9,432 Da and it was lower than that after UF. This demonstrates that there is a synergistic effect between UF and Fe₃O₄ predeposition that increased the removal efficiency of medium to high molecular weight HA. The peak value of HA after Fe₃O₄+H₂PO₄⁻+UF increased to 10,201 Da and it was much higher than that after Fe₃O₄+UF, which inferred that phosphate weakened the effect of Fe₃O₄ predeposition on medium to high molecular weight HA.

Table 3 shows the effect of different treatment processes on the removal efficiencies of HA with different molecular weights. Changes were calculated by the integration of different response time periods.

Table 3.

The Humic Acid Removal Efficiencies from Raw Water by Different Processes According to Different Molecular Weight Ranges

Molecular weight (Da)	13,000–66,000	7,400–13,000	3,900–7,400
UF	21.0	13.2	11.8
Fe₃O₄+UF	46.4	43.4	28.6
Fe₃O₄+H₂PO₄⁻+UF	37.4	32.0	34.9

Removal efficiencies of high, medium, and low molecular weight HA filtered by Fe₃O₄+UF were 25.3%, 29.1%, and 17.3% higher, respectively, than after filtration with UF alone. This shows that predeposition of Fe₃O₄ effectively enhanced the effect of UF and increased the HA removal efficiency, especially for medium and high molecular weight HA. The removal efficiencies of high and medium molecular weight HA filtered by Fe₃O₄+H₂PO₄⁻+UF were 9.0% and 9.0% lower than those filtered by Fe₃O₄+UF. However, the efficiency of removal of low molecular weight HA by Fe₃O₄+UF+H₂PO₄⁻ increased by 3.6%. This suggests that phosphate weakens the adsorption of high and medium molecular weight HA on the Fe₃O₄ predeposition layer, while the adsorption of low molecular weight HA by the Fe₃O₄ predeposition layer was not affected.

Mechanism of membrane fouling reduction by predeposition of Fe₃O₄

Zeta potential

Zeta potentials of Fe₃O₄ in the presence and absence of phosphate at pH = 3–9 were measured. The result is presented in Fig. 6. The Zeta potential of Fe₃O₄ with phosphate was lower than in the absence of phosphate. The difference in the Zeta potentials of Fe₃O₄ increased at pH = 3–7 and then decreased at pH = 7–9. Under acidic conditions, the surface of Fe₃O₄ was positively charged and it attracted phosphate. The positive charges on the surface of Fe₃O₄ increased along with the degree of acidity. Therefore, the attractive electric force between Fe₃O₄ and phosphate increased with decreasing pH and more phosphate was adsorbed on the surface of Fe₃O₄. However, the increase in the level of adsorbed phosphate was far less than the increase of positive charges with increasing acidity, which contributed to a smaller change in Zeta potential. In contrast, under more alkaline conditions, the surface of Fe₃O₄ was negatively charged and it no longer bound to phosphate. This is because the negative charges on the surface of Fe₃O₄ increased along with alkalinity. Consequently, the repulsive electric force between Fe₃O₄ and phosphate increased and less phosphate was adsorbed on the surface of Fe₃O₄. As a result, the Zeta potential change was reduced as alkalinity increased.

FIG. 6.

Zeta potentials of Fe₃O₄ with/without phosphate.

SEM images of membranes

To observe the variation of morphology of Fe₃O₄ deposition layer preadsorption and postadsorption of phosphate, the surfaces of UF membranes after different treatment processes were tested by SEM. The results are shown in Fig. 7.

FIG. 7.

Scanning electron microscope characterization of membrane by different processes: (a) UF; (b) UF+HA; (c) Fe₃O₄+UF+HA; (d) Fe₃O₄+H₂PO₄⁻+UF+HA.

It can be seen clearly from Fig. 7(a) that the pore size distribution of the UF membrane was mainly in the 30–100 nm range, which is much smaller compared with Fe₃O₄. Accordingly, the UF membrane could support Fe₃O₄ particles without Fe₃O₄ particles traversing through the membrane. As shown in Fig. 7b, there was a layer that was formed by HA caked on the surface of the UF membrane. This was the main reason that led to the decreased flux of HA during UF. As revealed in Fig. 7c, Fe₃O₄ particles exhibited a complete crystal structure and were arranged in an ordered manner, which made the Fe₃O₄ predeposition layer more compact and smooth. In addition, the small diameter made the gap between the particles small, which created a high probability for HA molecules to collide with Fe₃O₄ particles. Therefore, Fe₃O₄ particles adsorbed HA molecules easily and the filtration flux was improved as a result. Figure 7d indicates that Fe₃O₄ particles, which adsorbed phosphate, still have a complete crystal structure. The particles were arranged closely like a cake. Also, there were defects in the integrity of the predeposition layer formed by Fe₃O₄+H₂PO₄⁻+UF. Furthermore, the compactness of the Fe₃O₄ particles decreased. The above results show how a portion of the HA molecules was able to penetrate through the membrane, and hence, the effect of Fe₃O₄ predeposition decreased in the presence of phosphate

AFM images of membranes

To further study their composition, the UF membrane under different treatment conditions were examined by AFM to study the influence of the Fe₃O₄ predeposition process on membrane fouling. The results are presented in Fig. 8 and Table 4.

FIG. 8.

Three-dimensional shapes of membrane after different processes: (a) UF; (b) UF+HA; (c) Fe₃O₄+UF+HA; and (d) Fe₃O₄+H₂PO₄⁻+UF+HA.

Table 4.

Roughness of Membranes Made Using Different Processes

Sample	R_a/nm	R_q/nm	Z/nm
UF	116.5	146.9	902
HA+UF	71.5	88.4	554
HA+Fe₃O₄+UF	147.0	186.6	2,130
HA+Fe₃O₄+H₂PO₄⁻+UF	164.7	194.8	2,165

The three parameters (R_a, R_q, and Z) reflect the roughness of the surface of membranes tested by high-resolution atomic force microscope in different ways. R_a: Average roughness, the arithmetic mean value of the absolute value of the height of each point of the membranes relative to the zero plane. R_q: Root mean square roughness, the root mean square of the height of each point of membranes relative to the zero plane. Z: the maximum height of the profile, which is between the diaphragm surface peak and valley line distance.Fe₃O₄ dosage = 0.45 g/L, H₂PO₄⁻ = 1.55 mg/L, HA = 20 mg/L, pH = 7.0

The results show that the predeposition process increased the roughness of the UF membrane surface, consistent with previous research (Diallo et al., 2005; Xiarchos and Doulia, 2006). It was speculated the predeposition layer could work by intercepting HA molecules; so the number of HA molecules that reached the membrane surface was reduced greatly. In addition, it also cut off the adsorption function between organic matter and the membrane surface (Ma et al., 2017). Hence, membrane fouling was alleviated and the filtration flux increased. The interception effect was mainly dependent on the adsorption ability and compactness of the Fe₃O₄ predeposition layer.

As observed in Table 4, the roughness of membrane surface of Fe₃O₄+UF increased significantly after filtering HA compared with the roughness of direct UF. The values of R_a, R_q, and Z increased by 26.2%, 27.2%, and 136%, respectively. The roughness of the membrane surface of Fe₃O₄+H₂PO₄⁻+UF was largest. The values of R_a, R_q, and Z of it were 12.0%, 4.4%, and 0.2% higher than those of Fe₃O₄+UF, which inferred that phosphate caused the increased roughness. The increased roughness appeared as the decreased compactness of the Fe₃O₄ predeposition layer. Consequently, the ability of predeposited Fe₃O₄ to intercept HA molecules decreased.

Conclusions

Fe₃O₄ predeposition effectively alleviated membrane fouling and increased filtration flux of raw water by UF. However, after adsorption of phosphate, the effect of Fe₃O₄ predeposition on membrane flux and HA removal was reduced. There was an optimum concentration for the Fe₃O₄ predeposition process. The filtration flux of HA and HA removal efficiency were both highest under optimum concentration. The removal efficiency of Fe₃O₄ predeposition on HA with medium and high molecular weight increased after phosphate adsorption, but the removal efficiency of low molecular weight HA changed very little. The adsorption and interception functions of Fe₃O₄ predeposition weakened after adsorption of phosphate. As a result, the study of the effect of phosphate on the Fe₃O₄ predeposition process would be helpful in the application of an effective method for alleviating membrane fouling.

Footnotes

Acknowledgments

This work was supported by the Funds for the National Natural Science Foundation of China (Grant Nos. 51678027 and 51678026).

Author Disclosure Statement

No competing financial interests exist.

References

Ajmani

G.S.

, Goodwin

, Marsh

, Fairbrother

D.H.

, Schwab

K.J.

, Jacangelo

J.G.

, and Huang

(2012). Modification of low pressure membranes with carbon nanotube layers for fouling control. Water Res, 46, 5645.

Antelo

, Arce

, Avena

, Fiol

, López

, and Macías

(2007). Adsorption of a soil humic acid at the surface of goethite and its competitive interaction with phosphate. Geoderma, 138, 12.

Bai

, Qu

, Liang

, Ma

, Chang

, Wang

, and Li

(2013). Membrane fouling during ultrafiltration (UF) of surface water: Effects of sludge discharge interval (SDI). Desalination, 319, 18.

Campinas

, and Rosa

M.J.

(2010). Removal of microcystins by PAC/UF. Sep Purif Technol, 71, 114.

Chowdhury

S.R.

, and Yanful

E.K.

(2010). Arsenic and chromium removal by mixed magnetite–maghemite nanoparticles and the effect of phosphate on removal. J Environ Manag, 91, 2238.

Diallo

M.S.

, Christie

, Swaminathan

, Johnson

J.H.

Jr. and Goddard

W.A.

3rd . (2005). Dendrimer enhanced ultrafiltration. 1. Recovery of Cu(II) from aqueous solutions using PAMAM dendrimers with ethylene diamine core and terminal NH2 groups. Environ Sci Technol, 39, 1366.

Z.Y.

, Fengchang

W.U.

, Song

, Lin

, Bai

, Zhu

, and Geisy

J.P.

(2013). Competitive interaction between soil-derived humic acid and phosphate on goethite. Appl Geochem, 36, 125.

Harman

B.I.

, Koseoglu

, Yigit

N.O.

, Beyhan

, and Kitis

(2010). The use of iron oxide-coated ceramic membranes in removing natural organic matter and phenol from waters. Desalination, 261, 27.

Kim

, Cai

, and Benjamin

M.M.

(2008). Effects of adsorbents on membrane fouling by natural organic matter. J Memb Sci, 310, 356.

10.

Kumar

, Gholamvand

, Morrissey

, Nolan

, Ulbricht

, and Lawler

(2016). Preparation and characterization of low fouling novel hybrid ultrafiltration membranes based on the blends of go−tio 2, nanocomposite and polysulfone for humic acid removal. J Memb Sci, 506, 38.

11.

, Liang

, Qu

, Shao

, Yu

, Han

Z.S.

, Du

, and Li

(2014). Control of natural organic matter fouling of ultrafiltration membrane by adsorption pretreatment: Comparison of mesoporous adsorbent resin and powdered activated carbon. J Memb Sci, 471, 94.

12.

Liu

, Li

, Yang

, Ye

, Ji

, Ren

, and Zhou

(2014). Analysis of the major particle-size based foulants responsible for ultrafiltration membrane fouling in polluted raw water. Desalination, 347, 191.

13.

, Wang

, Hu

, Jefferson

W.A.

, Liu

, and Qu

(2017). Antifouling by pre-deposited al hydrolytic flocs on ultrafiltration membrane in the presence of humic acid and bovine serum albumin. J Memb Sci, 538, 34.

14.

, Yu

, Jefferson

W.A.

, Liu

, and Qu

(2015). Modification of ultrafiltration membrane with nanoscale zerovalent iron layers for humic acid fouling reduction. Water Res, 71, 140.

15.

, Rajabzadeh

, and Matsuyama

(2015). Preparation of antifouling poly(vinylidene fluoride) membranes via different coating methods using a zwitterionic copolymer. Appl Surf Sci, 357, 1388.

16.

Malczewska

, Liu

, and Benjamin

M.M.

(2015). Virtual elimination of mf and uf fouling by adsorptive pre-coat filtration. J Memb Sci, 479, 159.

17.

Qin

, Liu

, and Wang

(2012). Fractionation and kinetic processes of humic acid upon adsorption on colloidal hematite in aqueous solution with phosphate. Chem Eng J, 209, 458.

18.

Schippers

J.C.

, Galjaard

, Buijs

, Beerendonk

, and Schoonenberg

(2001). Pre-coating (EPCE^®) UF membranes for direct treatment of surface water. Desalination, 139, 305.

19.

Tian

J.Y.

, Ernst

, Cui

, and Jekel

(2013). Correlations of relevant membrane foulants with uf membrane fouling in different waters. Water Res, 47, 1218.

20.

Xiarchos

, and Doulia

(2006). Interaction behavior in ultrafiltration of nonionic surfactant micelles by adsorption. J Colloid Interface Sci, 299, 102.

21.

Xie

, Luo

W.H.

, and Gray

(2017). Synchrotron Fourier transform infrared mapping: A novel approach for membrane fouling characterization. Water Res, 111, 375.

22.

Yiantsios

S.G.

, and Karabelas

A.J.

(2001). An experimental study of humic acid and powdered activated carbon deposition on UF membranes and the removal by backwashing. Desalination, 140, 195.

23.

, and Graham

N.J.D.

(2015). Performance of an integrated granular media—Ultrafiltration membrane process for drinking water treatment. J Memb Sci, 492, 164.

24.

Zhang

, Chen

, Wang

, Zhang

, Zhao

, and Ma

(2015). Membrane fouling control in the integrated process of magnetic anion exchange and ultrafiltration. Desalin Water Treat, 57, 1.

25.

Zhao

, and Wang

(2016). A loose nano-filtration membrane prepared by coating hpan uf membrane with modified pei for dye reuse and desalination. J Memb Sci, 524, 214.

Effect of Phosphate on Ultrafiltration Membrane Performance After Predeposition of Fe 3 O 4

Abstract

Abstract

Introduction

Experimental protocols

Chemical agents and materials

UF experiments

Process 1 (UF)

Process 2 (Fe3O4+UF)

Process 3 (Fe3O4+H2PO4−+UF)

Analytical methods

Results and discussion

Flux decline caused by a critical dose of adsorbents

Particle size analysis

Effect of different processes on removal of organic compounds

Mechanism of membrane fouling reduction by predeposition of Fe3O4

Zeta potential

SEM images of membranes

AFM images of membranes

Conclusions

Footnotes

Acknowledgments

Author Disclosure Statement

References

Process 2 (Fe₃O₄+UF)

Process 3 (Fe₃O₄+H₂PO₄⁻+UF)

Mechanism of membrane fouling reduction by predeposition of Fe₃O₄