Abstract

To reveal the mechanisms of the influence of Ca2+ on membrane fouling with humic acid (HA), the adhesion forces of HA with both other HA molecules and the membrane, the HA fouling layer structure, HA fouling experiments, and the HA rejections at a wide range of Ca2+ dosages were investigated. The results indicated that the effect of Ca2+ on HA fouling can be divided into three stages. At lower ionic strength (IS) of CaCl2, the change in electrostatic forces is the main factor in controlling HA fouling behavior; i.e., increasing Ca2+ dosages resulted in more serious membrane fouling. When the IS of CaCl2 reached 10 mM, HA aggregates became the dominant factor in the fouling process, which could result in a porous fouling layer accompanied by less membrane fouling. Interestingly, much weaker membrane fouling was observed when the IS increased to 100 mM and the HA rejection began to decline. This was because a stronger hydration repulsion force was generated, which could weaken the compactness of the fouling layer and the adhesion forces of HA with both the membrane and HA, while enabling smaller-sized HA to pass more easily into the permeate, which led to less membrane fouling and a lower HA rejection.

INTRODUCTION

Humic acid (HA) is the main compound of natural organic matter, which exists widely in surface water, municipal wastewater, seawater, and other aquatic environments. HA is also a widely known precursor to potentially carcinogenic disinfection byproducts (Kyzas et al. 2017); therefore, removing HA from water or wastewater is a widespread concern. Ultrafiltration (UF) technology is one of the effective technologies for separating HA foulants from water sources (Szymański et al. 2016). However, HA fouling of membranes causes a rapid and irreversible loss of flux through the membrane, which is one of the major limiting factors for the successful application of UF technology in water or wastewater treatment (Shankar et al. 2017).

Many factors influence HA fouling of membranes, such as feedwater solution chemistry, membrane properties, and hydrodynamic and operating conditions (Lin et al. 2015). Among them, the most typical and widespread divalent ion in waters and wastewaters (such as natural waters, municipal wastewater, and seawater), i.e., the Ca2+ ion, has been widely reported to be a crucial factor that could seriously affect the HA fouling behavior in membranes (Wang et al. 2015). The functional groups in HA molecules easily interact with Ca2+ via electrical double layers, neutralization, or complexation, which can significantly influence the HA fouling layer structure and the deposition behavior of HA onto the membrane surface, which then will affect the fouling potential (Sutzkover-Gutman et al. 2010). Therefore, unraveling the mechanisms of the effect of Ca2+ on HA fouling behavior in membranes has continued to be of interest.

Some previous studies have reported the effects of Ca2+ on membrane fouling with HA; however, the influence of Ca2+ on membrane fouling with HA is still unclear and some contradictory findings were obtained.

For the most part, the HA fouling rate and extent have been exacerbated in the presence of divalent Ca2+ (Tang et al. 2007; Tian et al. 2013; Lin et al. 2015). Most researchers believe that with the increase in Ca2+ concentrations, more Ca2+ crosslinking or bridging interactions occur between HA molecules or between the membrane surface and HA, which could cause a more compacted gel network in the fouling layer and a rapid deposition of HA onto the membrane surface. This eventually leads to the phenomenon of more serious membrane fouling in the presence of Ca2+ (Sutzkover-Gutman et al. 2010; Tian et al. 2013).

However, several other research groups investigating the effects of Ca2+ on membrane fouling with HA found that HA fouling rate/extent decreased visibly at some Ca2+ concentrations, which is contrary to the findings mentioned above (Katsoufidou et al. 2005; Nghiem et al. 2006; Katsoufidou et al. 2008; Nghiem et al. 2008; Shao et al. 2011). Until now, however, the reason for reduced membrane fouling with Ca2+ has still been unclear and only tentative explanations have been given. For example, Shao et al. speculated that the complexation/bridging interactions between HA molecules are enhanced with the increase in Ca2+ dosage, which led to the formation of large aggregates. These large HA aggregates are retained on the membrane surface, which results in a more porous fouling layer with a comparatively higher permeability. Consequently, the membrane fouling rate/extent began to decrease (Shao et al. 2011). However, Nghiem et al. found that although membrane fouling decreased at higher Ca2+ dosages, the foulant rejection also decreased. This result appears not to support the speculation that larger HA aggregates are formed at higher Ca2+ dosages. Nghiem et al. did not present any further discussion or pursue any further studies to find the cause of this phenomenon (Nghiem et al. 2006, 2008).

The studies mentioned above indicate that two kinds of opposing results about the effects of Ca2+ on HA fouling behavior have been demonstrated to date. However, it appears that no study has focused on identifying the reason for the different characteristics with Ca2+ shown in membrane fouling with HA, and particularly the reason for the enhancement or mitigation mechanisms of Ca2+ on membrane fouling are still unclear or debatable. Most existing explanations appear to be simple descriptions or speculation based on experimental observations (Singh & Song 2005). Therefore, more research is needed into the effect of different Ca2+ dosages on membrane fouling with HA, which may be helpful in resolving the above issues and also effectively explain the different characteristics observed while using Ca2+ in membrane fouling with HA.

In the present work, polyvinylidene fluoride (PVDF) UF membranes were used to filter HA solutions for a wide range of Ca2+ dosages. Atomic force microscopy (AFM) associated with specific colloidal probes was used to determine the adhesion forces of HA with both HA and PVDF membrane at corresponding Ca2+ dosages. A quartz crystal microbalance with dissipation monitoring (QCM-D) was used to investigated the influences of different Ca2+ dosages on the HA fouling layer structure. These results were combined with the HA rejection and the hydrodynamic diameter of HA at corresponding Ca2+ dosages to give a better explanation for the HA fouling mechanism at different Ca2+ concentrations.

MATERIALS AND METHODS

Humic acids and membrane preparation

HA was purchased from Sigma-Aldrich (St Louis, MO, USA), and stock HA (1 g/L) solution was prepared. The fresh HA working solution was prepared before each HA filtration experiment and maintained at a pH of 7.0 and a dissolved organic carbon (DOC) concentration of 10 mg/L. Throughout this work, 0.01 mol/L NaOH or 0.01 HCl mol/L solutions were used to adjust the pH values of the HA working solutions as needed, and 0.5 mol/L CaCl2 stock solution was used to adjust the ionic strength (IS) of the HA working solutions. The IS of CaCl2 of the HA working solution was set at 0, 1, 10, and 100 mM. These values were selected because the range of 0–1 mM represents very soft water or wastewater, while 10 mM corresponds to ‘hard’ water and 100 mM would be encountered in special water or wastewater feed, such as seawater and brine or in chemical cleaning. Moreover, unless otherwise stated, all reagents and chemicals were analytical reagents with a purity exceeding 99% and ultrapure water was used to prepare all working solutions.

The present work used flat-sheet PVDF UF membranes that were synthesized in the laboratory by a phase inversion method. PVDF and N, N-dimethylacetamide were the predominant materials for the preparation of membranes, which were purchased from Solvay Advanced Polymers Co. (Solef 1015; USA) and Tianjin Fucheng Chemical Reagent Co. (Tianjin, China), respectively. The preparation procedure and characteristics of PVDF UF membranes were described in detail previously (Miao et al. 2017); the pure water flux, surface roughness, and contact angle of the membranes were 300 ± 20 L/(m2·h), 17.5 nm, and 75 ± 2°, respectively.

HA filtration and flux recovery tests

As previously reported (Miao et al. 2017), a stirred dead-end filtration system was used in all HA filtration experiments with PVDF UF membranes. The effective filtration area of membranes was 3.32 × 10−3m2 and the constant pressure of the filtration system was provided by a compressed nitrogen gas cylinder. During the filtration experiments, a computer connected to an electronic balance recorded the permeate flux (J) data of the corresponding tested membrane continuously. Unless otherwise noted, a new PVDF membrane was used in each filtration experiment. The HA filtration experiment procedure was as described in the following paragraphs.

First, a clean PVDF membrane was compacted at pressure of 0.15 MPa until the permeate flux stabilized. The pure water flux (J0) of the membrane was then measured at pressure of 0.10 MPa. And finally, at pressure of 0.10 MPa, the HA feed solution was filtered with the desired IS of CaCl2 for 120 min. The permeate flux of the membrane at this stage was recorded continuously and J/J0 was used to evaluate the flux decline behavior in PVDF membranes for convenient comparison. The fouling experiments were repeated at least six times for each IS condition of CaCl2.

After the corresponding filtration test, a physical cleaning process was carried out on the HA-fouled membrane: the HA-fouled membrane was put into a 400 mL deionized water in a beaker, which was then placed in a multivibrator and vibrated for 2 min. Finally, the permeate flux (Jr) of the cleaned membrane was determined at pressures of 0.10 MPa and cleaning efficiency was evaluated by the flux recovery rate (Jr/J0).

The HA rejection (R) of the PVDF membranes for different IS of CaCl2 was measured as:  
formula
where Cp is the DOC concentration of the HA permeate and Cf is the DOC concentration of the corresponding HA feed solution. The filtration times at which the HA rejections were measured were identical.

Adhesion force measurements

First, carboxyl colloidal probes were developed by our research group because HA is rich in carboxyl groups. The preparation method of the carboxyl-group probe was followed as detailed previously (Miao et al. 2015). The carboxyl colloidal probes were used as HA surrogates and associated with a MultiMode 8.0 AFM (Bruker, Germany) to determine the adhesion forces of HA with both HA and PVDF membranes at different IS of CaCl2.

All measurements of adhesion forces were performed in a fluid cell under contact mode. In measuring the adhesion force of HA with both HA and PVDF membranes, the test samples were an HA-fouled membrane and a clean PVDF membrane, respectively. The HA solution that was used in the corresponding filtration experiments was used as a test solution for HA–HA adhesion forces and an aqueous solution with desired IS of CaCl2 was used as a test solution for PVDF membrane–HA adhesion forces. To minimize the intrinsic error resulting in part from membrane surface heterogeneity, the adhesion force under the desired IS of CaCl2 was determined at six different locations. At least 15 force curves were achieved for each site. In addition, the integrity of all used carboxyl probes was checked by a scanning electron microscope to ensure the precision of the adhesion force measurements.

QCM-D experiments

A QCM-D system (E1; Qsence, Sweden) was used to investigate the structure of the HA fouling layers at different Ca2+ dosages. Before each QCM-D experiment, a PVDF-coated sensor crystal was developed and thoroughly cleaned using the preparation method and process described previously (Miao et al. 2017).

In the QCM-D experiments, the flow module of the system was equipped with a clean PVDF-coated sensor crystal. This system was then pre-equilibrated with ultrapure water to achieve a stable baseline. Finally, the HA feed solution with the targeted Ca2+ dosage was pumped into the system for 25 min. During this stage, increase in dissipation (ΔD) and decrease in frequency (Δf) from the baseline were monitored continuously.

The │ΔD/Δf│ ratio, i.e., the energy dissipation value per unit adsorbed mass change, provides information about the HA fouling layer structure at the corresponding Ca2+ concentration. That is, a higher │ΔD/Δf│ ratio suggests the formation of a relatively nonrigid and looser HA fouling layer structure, whereas a lower │ΔD/Δf│ ratio indicates a denser HA fouling layer (Chang & Bouchard 2013).

All QCM-D experiments were conducted at 23 °C. To maintain laminar flow in the module, the flow rate through the measurement chamber was kept at 0.1 mL/min. Both Δf and ΔD at the third overtone were used. Each QCM-D measurement was carried out at least three times.

Analyses

At various ISs of CaCl2, the zeta potentials and hydrodynamic diameters of HA were measured with a Zetasizer Nano instrument (ZS90; Malvern Instruments, Malvern, UK). The zeta potentials of PVDF membranes were quantified with an electrokinetic analyzer for solid surface analysis (SurPASS; Anton Paar GmbH, Graz, Austria). A total organic carbon analyzer (CPN; Shimadzu, Kyoto, Japan) was employed to measure the DOC concentrations of HA feed solutions or permeate.

RESULTS AND DISCUSSION

Charge characteristics of PVDF membranes and HA at different Ca2+ concentrations

The charge characteristics of the PVDF membrane and HA molecules are closely correlated with the electrostatic interaction forces of HA with both HA and membrane, which may be easily affected by the Ca2+ concentration (Yamamura et al. 2008). Therefore, the zeta potentials of PVDF membranes and HA at IS of 0, 1, 10, and 100 mM CaCl2 were determined systematically and are summarized in Table 1.

Table 1

Zeta potentials of PVDF membranes and HA surface at various Ca2+ dosages

IS (mM) 10 100 
Zeta potential (mV) PVDF membrane −29.7 −11.9 −5.5 −4.3 
HA −46.7 −19.7 −13.6 −8.3 
IS (mM) 10 100 
Zeta potential (mV) PVDF membrane −29.7 −11.9 −5.5 −4.3 
HA −46.7 −19.7 −13.6 −8.3 

As shown in Table 1, the HA molecules and PVDF membrane were negatively charged at all IS of CaCl2 investigated, which indicates that the electrostatic force of HA with both HA and PVDF membrane were repulsion forces. Furthermore, with increasing IS of CaCl2, the zeta potentials of both the HA molecules and the membrane decreased significantly, which was mainly because the charge neutralization, compression of the electrical double layer, and bridging action or complexation of Ca2+ were strengthened (Shen et al. 2011). This should also indicate that the electrostatic repulsion forces of HA with both PVDF and HA decreased with increasing Ca2+ dosage.

Influences of Ca2+ concentrations on the macro membrane fouling with HA

HA filtration tests were carried out at IS of 0, 1, 10, and 100 mM CaCl2, respectively. The normalized flux decline curves and the flux recovery efficiency of HA-fouled membranes at corresponding Ca2+ dosages are presented in Figures 1 and 2, respectively. As is clearly shown in the figures, with the increase of Ca2+ dosages, the variation trend in the flux decline rate and the flux recovery rate can be divided into two stages.

Figure 1

Normalized flux with filtration time of HA-fouled membranes at different Ca2+ dosages.

Figure 1

Normalized flux with filtration time of HA-fouled membranes at different Ca2+ dosages.

Figure 2

The flux recovery rate of HA-fouled membranes at different Ca2+ dosages.

Figure 2

The flux recovery rate of HA-fouled membranes at different Ca2+ dosages.

First, when increasing the IS from 0 to 1 mM, the extent of flux decline increased sharply from 33% to 77%, and the flux recovery rate decreased from 98% to 27%. Clearly, the membrane fouling rate and irreversible fouling were exacerbated significantly with the addition of a Ca2+ ion. This is in accordance with the results of most previous investigations (Seidel & Elimelech 2002; Tang et al. 2007; Tian et al. 2013). These studies suggested that the electrostatic repulsive force of HA with both HA and PVDF membrane would decrease with the addition of a Ca2+ ion, which could result in a denser HA fouling layer and much more serious membrane fouling.

In contrast, when the IS of CaCl2 exceeded 1 mM, the flux decline rate and irreversible fouling extent began to decrease with increasing Ca2+ dosage. In particular, when the IS reached 100 mM, the flux of HA-fouled membranes only declined by 24%; meanwhile the flux was almost fully recovered after cleaning. Similar phenomena have also been reported by Shao et al. (2011) and Katsoufidou et al. (2008), who investigated the HA fouling behavior of UF membranes at different Ca2+ dosages. They suggested that, with increasing Ca2+ concentrations, the complexation/bridging interactions of HA with both membrane and HA were gradually enhanced, which thus promoted the formation of larger HA aggregates. Such larger HA aggregates apparently tended to form looser and more porous HA fouling with relatively high permeability, which then resulted in less membrane fouling. However, as shown in Figure 3, the HA rejections of the PVDF membrane decreased significantly at IS of 100 mM CaCl2. This observation appeared to contrast somewhat with the above speculation that larger HA aggregates form at higher Ca2+ dosages because, compared with the relatively smaller HA molecules, the larger HA aggregates are more easily retained on the membrane surface (Yuan & Zydney 1999).

Figure 3

The HA rejections of a PVDF membrane at different IS of CaCl2.

Figure 3

The HA rejections of a PVDF membrane at different IS of CaCl2.

Based on the above discussion, it is clear that membrane fouling was enhanced with Ca2+ at relatively lower IS. In contrast, when the Ca2+ dosage exceeded a critical value, membrane fouling began to decline with increasing Ca2+ dosages. This phenomenon appears to suggest that the influence of Ca2+ on the HA fouling of membranes is different at different Ca2+ dosages. However, few researchers seem to have conducted insightful and systematic studies to unravel the reasons for the ‘different influence behavior’. Therefore, further studies are required.

Influences of Ca2+ concentrations on the adhesion forces of HA with both HA and PVDF

At IS of 0, 1, 10, and 100 mM CaCl2, the adhesion forces of HA with both HA and PVDF membrane were measured systematically. The corresponding representative adhesion force curves are presented in Figure 4.

Figure 4

The typical adhesion force curves of HA with both HA and PVDF membrane at different IS of CaCl2.

Figure 4

The typical adhesion force curves of HA with both HA and PVDF membrane at different IS of CaCl2.

It is clear that, with increasing Ca2+ dosages, the variation trend of the adhesion forces of HA with both HA and PVDF membrane was divided into two stages, which is similar to the variation in the characteristics of the flux decline and recovery rate/extent of corresponding HA-fouled membranes (as shown in Figures 1 and 2).

At IS of 0, 1, and 10 mM CaCl2, the average adhesion forces of HA with PVDF membrane were 6.6 ± 0.38, 7.5 ± 0.57, and 9.6 ± 0.76 mN/m, respectively, and the average adhesion forces of HA with HA were 0.50 ± 0.03, 1.71 ± 0.11, and 2.15 ± 0.18 mN/m, respectively. Clearly, in the IS of 0–10 mM CaCl2, the adhesion forces of HA with both HA and PVDF membrane increased gradually with increasing Ca2+ dosages. This was mainly because the complexation/bridging actions, neutralization, or charge screening were strengthened by increasing the Ca2+ dosage, which reduced the net charges of the HA molecules and the membrane. This reduction of net charges led to a decrease in the electrostatic repulsion forces of HA with both HA and membrane, eventually leading to the corresponding adhesion forces increasing with increasing Ca2+ dosages (Mo et al. 2008).

While increasing the IS from 10 to 100 mM, the average PVDF–HA adhesion force decreased sharply from 9.6 ± 0.76 to 2 ± 0.12 mN/m and the HA–HA adhesion force decreased from 2.15 ± 0.18 to 0.52 ± 0.04 mN/m. It is interesting to note that in this range of IS, the adhesion forces of HA with both HA and PVDF membrane began to decline with increasing Ca2+ dosages, although the electrostatic repulsion forces of HA with both HA and PVDF membrane still decreased with increasing Ca2+ dosages (as discussed in section ‘Charge characteristics of PVDF membranes and HA at different Ca2+ concentrations’). This phenomenon is contrary to the observations at IS of 0–10 mM CaCl2, which may be explained effectively by the balance of electrostatic interaction forces and hydration repulsion forces.

As mentioned above, the electrostatic repulsion forces of HA with both HA and PVDF membrane decreased with an increase in Ca2+ dosage. Simultaneously, because Ca2+ is one of the most typical hydrated ions, the hydration repulsion forces of HA with both HA and PVDF membrane would be provoked and increase with an increase in Ca2+ dosage, which is a relatively short-range repulsive force and difficult to distinguish from the van der Waals and electrostatic forces at lower Ca2+ concentrations (Butt et al. 2005). That is, with the increase in Ca2+ dosage, the electrostatic repulsion forces of HA with both HA and PVDF membrane decreased, whereas their hydration repulsion forces increased (Tansel et al. 2006). Therefore, the balance of electrostatic interaction and hydration repulsion forces might be able to explain the variation characteristics of their corresponding adhesion force.

That is, when increasing IS from 0 to 10 mM, the increase in the adhesion forces of HA with both HA and PVDF membrane was mainly attributed to the decrease of electrostatic repulsion forces because the increase in hydration repulsion forces was weak and masked by the changes in the electrostatic forces (Israelachvili 2012). Conversely, when the IS increased to 100 mM, the amount of hydrated Ca2+ in the vicinity of the HA and membrane surfaces increased markedly, which resulted in a much stronger hydration repulsion force of HA with both HA and membrane (Valle-Delgado et al. 2011). This stronger hydration repulsion force masked the corresponding changes in the electrostatic repulsion force, which eventually led to a decline in the adhesion forces of HA–HA and HA–PVDF.

This is also the reason why the HA rejection at IS 100 mM visibly declined. The stronger hydration repulsion force could result in a higher diffusion coefficient of HA molecules; therefore, the HA with a relatively smaller size could pass more easily through the membrane and into the permeate, thus weakening the HA rejection of PVDF membranes at higher Ca2+ dosages (Wang et al. 2013a, 2013b).

Combining the above discussion about the micro-adhesion forces of HA–HA and PVDF–HA with those of macro fouling behaviors at corresponding Ca2+ dosages, it is easy to find that, at some range of IS of CaCl2, the variation characteristics of macro membrane fouling appear to be well explained by the corresponding changes in the adhesion forces of HA–HA and PVDF–HA.

For example, when increasing the IS of CaCl2 from 0 to 1 mM, the adhesion forces of HA–HA and PVDF–HA increased, which could accelerate the deposition of HA onto the membrane surface and result in a much denser fouling layer (Miao et al. 2017). Thus, the corresponding membrane fouling rate/extent and irreversible fouling were enhanced. In addition, when increasing the IS of CaCl2 from 10 to 100 mM, the adhesion forces of HA with both HA and PVDF membrane began to decline because of the stronger hydration repulsion forces, which resulted in a great decrease in the membrane fouling rate/extent and irreversible fouling. Those observations are in line with previous viewpoints, namely, the stronger or weaker the adhesion force of the membrane–foulant and foulant–foulant, the more or less serious the irreversible fouling and flux decline rate/extent (Wang et al. 2013a, 2013b).

However, it is noteworthy that, when increasing the IS of CaCl2 from 1 to 10 mM, the adhesion forces of HA with both HA and membrane increased with increasing Ca2+ dosage, whereas the irreversible fouling and flux decline rate/extent of the HA-fouled membrane declined significantly. That is, the increase in adhesion forces of HA with both HA and membrane was accompanied by less membrane fouling. This seems to suggest that, at IS of 10 mM CaCl2, the changes in macro membrane fouling behavior do not appear to be fully explained by the changes of corresponding adhesion forces. In addition, the HA fouling mechanisms of membranes at different Ca2+ dosages were different; therefore, further studies need to analyze and resolve these issues.

Effects of Ca2+ concentrations on HA fouling layer structure

Based on the macro fouling experiments and the micro PVDF–HA/HA–HA adhesion forces described above, the HA fouling layer structures on membrane surfaces at corresponding Ca2+ concentrations were investigated systematically by QCM-D in conjunction with self-made PVDF-coated sensor crystals. The results are presented in Figure 5. It is clear that the variation trend of the │ΔD/Δf│ value with increasing Ca2+ dosage can be divided into two stages.

Figure 5

The │ΔD/Δf│ values for the HA fouling layer at various IS of CaCl2.

Figure 5

The │ΔD/Δf│ values for the HA fouling layer at various IS of CaCl2.

When increasing the IS of CaCl2 from 0 to 1 mM, the │ΔD/Δf│ value decreased rapidly from 10.6 × 10−8 to 1.4 × 10−8, which suggests that the HA fouling layer became much denser with addition of Ca2+. Combined with the results of macro membrane fouling and microadhesion forces at corresponding Ca2+ dosages, it is easy to find that the effects of Ca2+ on HA fouling behavior completely followed the Derjaguin–Landau–Verwey–Overbeek theory in this range of Ca2+ dosages (Israelachvili 2012). That is, with the increase in Ca2+ dosage, the electrostatic repulsion forces of HA–HA and PVDF–HA decreased, which accelerated the deposition of HA onto the membrane surface, and formation of a more compacted HA fouling layer finally led to a more serious membrane fouling rate and irreversible fouling.

In contrast to the observations at IS of 0–1 mM, the │ΔD/Δf│ value increased from 1.4 × 10−8 to 5.2 × 10−8 when increasing the IS from 1 to 10 mM, which indicated that the HA fouling layer became loose and porous at an IS of 10 mM, compared with that at an IS of 1 mM. It is generally accepted that the stronger the adhesion force of HA–HA, the more compact the HA fouling layer (Wang et al. 2013a, 2013b). However, at an IS of 10 mM, the adhesion force of HA–HA was the strongest, whereas the HA fouling layer became looser. This phenomenon could well be explained by the aggregation of HA molecules.

As shown in Table 2, the hydrodynamic diameter of HA increased sharply from 222 nm to 768 nm when increasing the IS from 1 to 10 mM. This is mainly because HA is rich in carboxyl functional groups and Ca2+ promotes bridging between the carboxyl functional groups; i.e., two adjacent HA molecules aggregate together through their bridging actions. The aggregation phenomena of HA intensify as the Ca2+ dosage continues to increase, which could result in the formation of larger HA aggregates (Katsoufidou et al. 2005). These larger aggregates were much more difficult to pass through the membrane. As a result, it is much easier for these large HA aggregates to be retained on a membrane surface with a more porous fouling layer (Yuan & Zydney 2000). This phenomenon suggests that a pretreatment choice, such as coagulation/flocculation, which could result in the formation of larger foulant aggregates, may be an effective measure to mitigate membrane fouling.

Table 2

Hydrodynamic diameter of HA at different Ca2+ dosages

Ionic strengths 0 mM 1 mM 10 mM 100 mM 
Hydrodynamic diameter (nm) 212 ± 14 222 ± 23 768 ± 31 1,379 ± 108 
Ionic strengths 0 mM 1 mM 10 mM 100 mM 
Hydrodynamic diameter (nm) 212 ± 14 222 ± 23 768 ± 31 1,379 ± 108 

The │ΔD/Δf│ increased rapidly to 19.6 × 10−8 when the IS of CaCl2 further increased to 100 mM, which indicated that the HA fouling layer became increasingly porous. As shown in Table 2, the hydrodynamic diameter of HA at an IS of 100 mM increased to 1,379 nm. This observation is similar to that at an IS of 10 mM, which could be explained by the aggregation phenomena of HA molecules. However, it is worth noting that the HA rejection began to decrease at an IS of 100 mM, which is different from that at an IS of 10 mM. This may be because a stronger hydration repulsion force is generated at an IS of 100 mM, which becomes another key factor in controlling the HA fouling of membrane other than the aggregates of HA molecules. The stronger hydration repulsion could effectively weaken the compactness of the HA fouling layer and the adhesion forces of HA with both HA and PVDF membrane. Meanwhile, it results in a higher diffusion coefficient of HA molecules, which causes the HA with relatively lower size to pass more easily through the membrane and into the permeate. Therefore, the HA rejection of PVDF membranes is weakened (Wang et al. 2013a, 2013b), which eventually leads to the fouling rate/extent and irreversible fouling of HA-fouled membrane steadily becoming weaker. It is worth noting that the decrease in membrane fouling at an IS value of 100 mM CaCl2 appears to be at the cost of sacrificing HA removal efficiency. Therefore, when the hydration repulsion force is generated, the mitigation extent of membrane fouling and the foulant rejections should be considered comprehensively in actual operation processes.

CONCLUSION

Based on the above analysis, it is easy to find that the effects of Ca2+ on HA fouling behavior were different for different Ca2+ dosages and could be divided into three stages.

When increasing the IS of CaCl2 from 0 to 1 mM, membrane fouling with HA was mainly controlled by the changes of electrostatic repulsive forces, and membrane fouling rate/extent and irreversible fouling were increased with increasing Ca2+ dosage. This is mainly because the electrostatic repulsive forces of HA with both HA and PVDF membranes decreased with the addition of Ca2+ ions, which could promote the deposition of HA onto the membrane surface, and the formation of a more compacted fouling layer, eventually leading to more serious membrane fouling.

When the IS of CaCl2 increased to 10 mM, the aggregates of HA molecules appeared to be the primary factor controlling membrane fouling behavior, which could cause less membrane fouling rate/extent and irreversible fouling. To be more precise, the increase in the Ca2+ dosage could promote the formation of large HA aggregates, which could not penetrate the membrane and will be retained on the membrane surface. This not only improved the HA rejection, but also resulted in a more porous HA fouling layer, accompanied by less membrane fouling.

When the IS of CaCl2 reached 100 mM, the membrane fouling rate/extent and irreversible fouling decreased even more sharply and the HA rejection also began to decline. At this higher Ca2+ dosage, in addition to the aggregates of HA molecules, the hydration repulsion forces became another dominant factor controlling membrane fouling, which not only could weaken the adhesion forces of HA with both HA and PVDF membrane, but also caused the HA with relatively smaller size to pass more easily through the membrane and into the permeate, thus leading to less membrane fouling and a lower HA rejection.

ACKNOWLEDGEMENTS

We thank the Innovative Research Team of Shaanxi Province (2017KCT-19-01), the Hong Kong Scholars Program (XJ2017037), the National Natural Science Foundation of China (51608429), the Natural Science Foundation of Shaanxi Province (2016JQ5067), the Educational Commission of Shaanxi Province of China (16JS062), the key industrial chain (group) project of Shaanxi Province (2017ZDCXL-GY-07-01), and the Postdoctoral Science Foundation of Shaanxi Province for the support of this research.

REFERENCES

REFERENCES
Butt
,
H. J.
,
Cappella
,
B.
&
Kappl
,
M.
2005
Force measurements with the atomic force microscope: technique, interpretation and applications
.
Surface Science Reports
59
(
1
),
1
152
.
Chang
,
X.
&
Bouchard
,
D. C.
2013
Multiwalled carbon nanotube deposition on model environmental surfaces
.
Environmental Science & Technology
47
(
18
),
10372
10380
.
Israelachvili
,
J. N.
2012
Intermolecular and Surface Forces, revised
3rd edn
.
Academic Press
,
London
.
Katsoufidou
,
K.
,
Yiantsios
,
S. G.
&
Karabelas
,
A. J.
2005
A study of ultrafiltration membrane fouling by humic acids and flux recovery by backwashing: experiments and modeling
.
Journal of Membrane Science
266
(
1
),
40
50
.
Kyzas
,
G. Z.
,
Bikiaris
,
D. N.
&
Lambropoulou
,
D. A.
2017
Effect of humic acid on pharmaceuticals adsorption using sulfonic acid grafted chitosan
.
Journal of Molecular Liquids
230
,
1
5
.
Miao
,
R.
,
Wang
,
L.
,
Deng
,
D.
,
Li
,
S.
,
Wang
,
J.
,
Liu
,
T.
,
Zhu
,
M.
&
Lv
,
Y.
2017
Evaluating the effects of sodium and magnesium on the interaction processes of humic acid and ultrafiltration membrane surfaces
.
Journal of Membrane Science
526
,
131
137
.
Nghiem
,
L. D.
,
Oschmann
,
N.
&
Schäfer
,
A. I.
2006
Fouling in greywater recycling by direct ultrafiltration
.
Desalination
187
(
1
3)
,
283
290
.
Shankar
,
V.
,
Heo
,
J.
,
Al-Hamadani
,
Y. A.
,
Park
,
C. M.
,
Chu
,
K. H.
&
Yoon
,
Y.
2017
Evaluation of biochar-ultrafiltration membrane processes for humic acid removal under various hydrodynamic, pH, ionic strength, and pressure conditions
.
Journal of Environmental Management
197
,
610
618
.
Shen
,
Y.
,
Kim
,
H.
,
Tong
,
M.
&
Li
,
Q.
2011
Influence of solution chemistry on the deposition and detachment kinetics of RNA on silica surfaces
.
Colloids and Surfaces B: Biointerfaces
82
(
2
),
443
449
.
Singh
,
G.
&
Song
,
L.
2005
Quantifying the effect of ionic strength on colloidal fouling potential in membrane filtration
.
Journal of Colloid and Interface Science
284
(
2
),
630
638
.
Sutzkover-Gutman
,
I.
,
Hasson
,
D.
&
Semiat
,
R.
2010
Humic substances fouling in ultrafiltration processes
.
Desalination
261
(
3
),
218
231
.
Szymański
,
K.
,
Morawski
,
A. W.
&
Mozia
,
S.
2016
Humic acids removal in a photocatalytic membrane reactor with a ceramic UF membrane
.
Chemical Engineering Journal
305
,
19
27
.
Tansel
,
B.
,
Sager
,
J.
,
Rector
,
T.
,
Garland
,
J.
,
Strayer
,
R. F.
,
Levine
,
L.
&
Bauer
,
J.
2006
Significance of hydrated radius and hydration shells on ionic permeability during nanofiltration in dead end and cross flow modes
.
Separation and Purification Technology
51
(
1
),
40
47
.
Tian
,
J. Y.
,
Ernst
,
M.
,
Cui
,
F.
&
Jekel
,
M.
2013
Effect of different cations on UF membrane fouling by NOM fractions
.
Chemical Engineering Journal
223
,
547
555
.
Valle-Delgado
,
J. J.
,
Molina-Bolívar
,
J. A.
,
Galisteo-González
,
F.
&
Gálvez-Ruiz
,
M. J.
2011
Evidence of hydration forces between proteins
.
Current Opinion in Colloid & Interface Science
16
(
6
),
572
578
.
Wang
,
L.
,
Miao
,
R.
,
Wang
,
X.
,
Lv
,
Y.
,
Meng
,
X.
,
Yang
,
Y.
&
Ju
,
K.
2013a
Fouling behavior of typical organic foulants in polyvinylidene fluoride ultrafiltration membranes: characterization from microforces
.
Environmental Science & Technology
47
(
8
),
3708
3714
.
Wang
,
L. L.
,
Wang
,
L. F.
,
Ye
,
X. D.
&
Yu
,
H. Q.
2013b
Hydration interactions and stability of soluble microbial products in aqueous solutions
.
Water Research
47
(
15
),
5921
5929
.
Yamamura
,
H.
,
Kimura
,
K.
,
Okajima
,
T.
,
Tokumoto
,
H.
&
Watanabe
,
Y.
2008
Affinity of functional groups for membrane surfaces: implications for physically irreversible fouling
.
Environmental Science & Technology
42
(
14
),
5310
5315
.
Yuan
,
W.
&
Zydney
,
A. L.
2000
Humic acid fouling during ultrafiltration
.
Environmental Science & Technology
34
(
23
),
5043
5050
.