The impact of secondary operating temperature on in-line coagulation/flocculation and fouling of membranes used in tertiary treatment

Abstract

In-line coagulation/flocculation is a promising approach for fouling mitigation for membranes of tertiary treatment of wastewaters. However, the potential to minimize membrane fouling under low temperatures by optimizing coagulant dose during in-line coagulation/flocculation has not been examined. In the present study, the effect of secondary operating temperature on the performance of in-line coagulation/flocculation, and subsequent membrane fouling due to the differing SBR effluent components, was differentiated over a range of alum dosages (0–1.0 mM). The results demonstrated that in-line coagulation/flocculation achieved similar DOC reduction for effluents from SBR operated at 8 and 20 °C, however, the reduction in high and low MW organics by in-line coagulation/flocculation were higher for the effluent from SBR operated at 8 °C than that of 20 °C. Moreover, the reduction in high MW organics by in-line coagulation/flocculation were greater than those of low MW organics. Filtration tests revealed that in-line coagulation/flocculation reduced the development of cake fouling more than intermediate pore blocking and this was more obvious with the effluent from the SBR operated at 8 °C. The preferred alum dosages to control total membrane resistance accumulation were 0.2 and 0.1 mM for the effluents from the SBR operated at 8 and 20 °C, respectively. However, the temperature corrected membrane resistances were two times higher at low filtration temperature than those of 20 °C regardless of coagulant dosages. It was concluded that the potential of in-line coagulation/flocculation with alum for membrane fouling alleviation was limited in cold regions.

---

Water impact

This study investigates the interactive effect of secondary operating temperature on effluent characteristic, in-line coagulation/flocculation performance and ultrafiltration performance. The preferred dosage of coagulant and fouling mechanisms for fouling mitigation for tertiary membranes filtrations of a secondary effluent at the low temperatures and room temperature were compared. This study provides a design reference when coagulation/flocculation is employed to mitigate fouling of tertiary membranes under low temperature conditions.

1. Introduction

More than 20% of the world's population lives in a cold climate for several months every year.¹ Consequently, the need for effective treatment processes that operate effectively in cold environments is urgent. Ultrafiltration (UF) is gaining increasing acceptance for tertiary treatment of wastewaters as it is able to effectively and efficiently produce high quality treated water.² However, membrane fouling that reduces permeability and leads to increased operating costs and energy consumption is a significant barrier to it's adoption.³ Furthermore, membrane fouling has been demonstrated to be accentuated under cold operating conditions.^4,5 Under these circumstances, it is imperative to develop fouling mitigation strategies that are specific to cold weather operations.

Low temperatures have been found to negatively impact the performance of ultrafiltration (UF) membranes in a complex manner when employed for domestic wastewater treatment. Abu-Obaid et al. reported that in a full-scale tertiary membrane treatment process, seasonally low temperatures (as low as 8 °C) resulted in increased release of soluble microbial products (SMPs) from biomass, that negatively impacted membrane fouling and the dewaterability of the activated sludge.⁶ Polysaccharides and low molecular weight organics (LMW) accounted for most of the increase in SMPs rather than intermediate molecular weight organics.⁷ The higher levels of polysaccharides and LMW organics generated under the low secondary operating temperatures corresponded to increased cake layer formation and pore blocking respectively when membrane filtration was employed for tertiary treatment.⁸ Low temperature operation also led to a decrease in membrane permeability and membrane pore size shrinkage.⁹ Considering that the primary foulants during low temperature conditions are produced through biomass metabolism in secondary treatment, an efficient pre-treatment that can mitigate fouling associated with these SMPs would benefit subsequent tertiary membrane treatment.

In-line coagulation/flocculation has been widely used as a pre-treatment strategy prior to membrane filtration to mitigate fouling and enhance contaminant removal. Coagulation/flocculation with hydrolyzing metallic salts can either agglomerate the SMPs into flocs or modify the SMPs' physical–chemical characteristics,^10,11 thereby mitigating the severity of flux decline when filtering SMP rich waters.^12,13 Furthermore, the use of in-line coagulation/flocculation can avoid the use of a settling stage to reduce the footprint. Previous studies considered the filtration of waters with dissolved organic carbon (DOC) concentrations ranging from 4.5 to 8.0 mg L⁻¹.^14,15 However, recent research has revealed that low temperature operation (8 °C) of secondary treatment generates DOC at higher concentrations (i.e. 15 mg L⁻¹) and with a different composition than previously tested.¹⁶ Given the different properties of the feedwater entering coagulation/flocculation in low temperature conditions, it is anticipated that the preferred dosage of coagulants, and the mechanisms of membrane fouling, will differ for low temperature conditions.

The comprehensive impact of low temperatures on biological treatment, in-line coagulation/flocculation and ultrafiltration has not been studied in detail, and no attention has been paid to differentiate the previously discussed complex interactions. The objectives of the present study were to 1) investigate the effect of secondary operating temperature on the performance of in-line coagulation/flocculation due to changed secondary effluent characteristics, 2) evaluate membrane performance with different coagulant dosages and identify fouling mitigation mechanisms that can be achieved with in-line coagulation/flocculation. In the present study, alum was selected as the coagulant as it has been widely demonstrated to be effective prior to membranes for fouling mitigation.^10,17 Bench-scale sequencing batch reactors (SBRs) treating municipal wastewater were operated at 20 and 8 °C to provide feed water to bench-scale membrane systems used to assess fouling. The SBR effluents were coagulated/flocculated with different dosages of alum prior to direct membrane filtration (i.e. without settling). Coagulation/flocculation and filtration tests were conducted at one temperature to avoid having more than one variable influencing the filtration outcomes. Liquid chromatography-organic carbon detection (LC-OCD) analysis was employed to quantify SMP concentrations in the raw and coagulated/flocculated SBR effluents, providing insight into the potential types of foulants that were impacted by temperature of SBR effluent and alum dosage. Fouling models were employed to facilitate the analysis of the time series data, enabling the results from experiments undertaken with different temperatures and coagulant dosages to be compared. The membrane resistances were corrected by temperature using data from previous filtration tests that were conducted under low temperatures. The outcomes of the present study provide novel insights into how coagulant dosing should be modified to mitigate fouling of tertiary membranes during periods of low temperature.

2. Materials and methods

2.1 Secondary treatment operation

The secondary effluents used in the present study were obtained from SBRs that were treating municipal wastewater at a hydraulic retention time (HRT) of 10 hours and a sludge retention time (SRT) of 25 days at different temperatures. Additional details of the SBR configuration can be found elsewhere.⁷ A control SBR was operated at a temperature of 20 °C, considered to be representative of typical warm weather temperatures for secondary treatment, while a test SBR was operated at 8 °C, considered to be representative of winter operating conditions in Ontario Canada. Both bench-scale SBRs were operated for 3 months which was longer than 3 SRT (SRT = 25 days) to reach pseudo steady state. The effluent quality parameters (DOC, COD, turbidity, etc.) were measured over time and they did not change substantially. A detailed description of SBR operation is provided in the ESI.†

2.2 In-line coagulation/flocculation and filtration tests

The experiments were designed to be representative of in-line coagulation/flocculation immediately prior to UF filtration without sedimentation (i.e. direct filtration). In-line coagulation/flocculation experiments were conducted in 1 L beakers using a programmable jar test apparatus (PB-700, PHIPPS&BIRD, USA). Alum powder (Sigma-Aldrich, USA) was added into the beakers to achieve a range of dosages (0, 0.05, 0.1, 0.2, 0.5, 1.0 mM as Al). Dosages were selected based on previous research on in-line coagulation/flocculation with alum.¹⁸ The pH of the effluents before and after coagulation/flocculation was monitored. The dosed water samples were mixed rapidly at 200 rpm for 1 min and then flocculated at 35 rpm for 20 min to allow floc growth to occur. This mixing condition was selected as suggested by Yu et al.¹⁹ and Liu et al.²⁰ After mixing, coagulated/flocculated water was immediately used for filtration tests. The filtration tests lasted for approximately 16 hours during which air scouring was continuously applied as a fouling mitigation approach.

As one of the goals of the study was to evaluate the interaction between secondary effluent characteristics with effects of secondary operating temperatures and coagulant dose, in-line coagulation/flocculation and filtration tests were conducted at room temperature (∼20 °C). To achieve this, the effluents from the SBR operated at 8 °C were warmed up to 20 °C in a water bath prior to the in-line coagulation/flocculation and filtration tests. Fig. 1 illustrates the schematic diagram of test plan. The terms test SBR and control SBR are subsequently employed to refer to the SBRs that were operated at 8 °C and 20 °C, respectively.

Fig. 1 Schematic diagram of test plan.

The filtration tests were conducted with custom bench scale hollow fibre ultrafiltration membrane modules in dead-end mode. The volume of the filtration tank was 2 L. Each membrane module consisted of three ZeeWeed-1000 hollow-fibres (Veolia, Canada). The nominal pore size of the ZeeWeed-1000 membrane is 0.02 μm. The length of each hollow-fibre was 500 mm. The total membrane surface area of each module was 4475 mm². Each test consisted of 30 permeation cycles that included 30 min of filtration and 2 min of back pulsing and air scouring. Back pulsing and air scouring were employed to remove hydraulically reversible fouling from the membrane. The filtration and back pulsing fluxes were 24 and 48 LMH, respectively. An air diffuser (coupled with an air flow meter), located at the bottom of tank, provided aeration for mixing and membrane scouring. The aeration flow was set to 0.75 L min⁻¹ as optimized by Akhondi et al.²¹ All filtration tests were conducted in triplicate.

2.3 Water quality analysis

Raw municipal wastewater and SBR effluent with and without prior coagulation/flocculation were filtered through 0.45 μm nylon filters prior to analysis of soluble components. Soluble chemical oxygen demand (sCOD) dissolved organic carbon (DOC) and UV₂₅₄ were measured as indicators of dissolved organic matter using Standard Methods 5220 (APHA, 2005), a total organic carbon analyzer (TOC-L, Shimadzu, Japan) and a spectrophotometer (8453, Agilent, USA), respectively. pH was measured using a portable pH meter (SP80PD, VWR, USA). Suspended solids (SS) and turbidity were analyzed as indicators of particulate matter using Standard Method 2540D (APHA, 2005) and a turbidity meter (2100N, Hach, USA), respectively. Total phosphorus (TP) was measured according to Standard Methods (APHA, 2005) to indicate the P removal performance achieved by coagulation.

LC-OCD analysis was used to provide insight into the DOC composition of the SBR effluents, coagulated/flocculated SBR effluents and membrane permeates, as outlined by Huber et al.²³ The biopolymer concentrations reported by the LC-OCD analysis were divided into polysaccharide and protein components assuming all of the dissolved organic nitrogen (DON) in the biopolymers was bound in proteinaceous matter and employing a typical C : N mass ratio of 3 for proteins.²⁴

The removal rate (%) in substance concentrations was used to quantify the treatment performance by coagulation/flocculation, and was calculated from the test data using eqn (1):

where C₁ is the substance concentration in the SBR effluent without coagulation/flocculation, C₂ is the substance concentration in the SBR effluent with coagulation/flocculation.

2.4 Membrane fouling analysis

Three resistance-based fouling indices, based on the resistance-in-series model,²⁵ were used to quantify the evolution of fouling in the filtration tests. Total resistance (R_t,i) quantified the sum of intrinsic membrane (R_m), hydraulically reversible (R_rev,i) and irreversible (R_irr,i) resistances at the end of cycle i, and was calculated from the test data using eqn (2):where TMP was the measured transmembrane pressure (Pa), η was the permeate viscosity (Pa s), and J was the measured flux (m³ m⁻² s⁻¹). Permeate viscosity was calculated according to eqn (3):²⁶

η = 0.497(T + 42.5)^−1.5 (3)

with temperature (T) in degrees Celsius. This equation has been widely used for temperatures ranging 5–35 °C,^3,26,27 that spanned the range of temperatures employed in the present study.

R _rev,i quantified the resistance that accumulated in each cycle that was recovered during physical cleaning between cycles and was calculated for each cycle as per eqn (4).

R_rev,i = Rt^f_i − Rtⁱ_i+1 (4)

where Rt^f_i was the final membrane resistance of cycle i, and Rtⁱ_i+1 was the initial membrane resistance of filtration cycle i + 1.

R _irr,i was defined as the resistance that accumulated during a cycle and that was not removed by physical cleaning and was calculated for each cycle as per eqn (5).

R_irr,i = R_t,i − R_rev,i − R_m (5)

The fouling rate (F_i) in each cycle was examined as it can provide insight into the possible mechanisms responsible for fouling,²⁸ and was calculated as per eqn (6).where Rt^f_i and Rtⁱ_i were the final and initial membrane resistances of cycle i, and Δt was the duration of the cycles.

The combined fouling models, established by Bolton et al.²⁹ were used to extract relevant kinetic constants from the time series TMP data collected during the filtration tests. The five constant flow combined fouling models were summarized in ESI† (Table S2). Cake-complete, cake-intermediate, complete-standard, intermediate-standard and cake-standard models were considered. The normalized TMP values (P/P₀) over time from the triplicate experiments were used in the model fitting. The best fit was determined by minimizing the sum of squared residuals (SSR) where the residual was equal to the difference between measured data and model prediction.

2.5 Statistical analysis

Student t-tests were employed to compare the characteristics of effluents from the SBRs operated at 8 and 20 °C. Where appropriate the mean values observed at each operating condition were compared with Tukey tests. Data sets were considered statistically different at a 95% confidence interval (P < 0.05). All statistical analyses were performed using the Origin Pro 2020 software package.

3. Results and discussion

The effect of temperature of SBR effluent on coagulation/flocculation, as characterized by water quality parameters (i.e. DOC and DOC fractions), was assessed at various coagulant doses. Filtration tests were conducted and the relationship between membrane fouling and DOC fractions in the coagulated/flocculated SBR effluent was quantitatively evaluated. The results from the ultrafiltration tests were used to estimate fouling indices that provided insights into the effect of the temperature of SBR effluent operating temperature on the fouling mitigation performance of alum addition.

3.1 Effect of secondary operating temperature on the performance of in-line coagulation/flocculation

3.1.1 Conventional water quality responses. The water quality of the effluents from the test and control SBRs, as assessed by a range of conventional parameters, was compared using student t-tests. The two SBRs were fed with the same raw wastewater and hence the differences in effluent characteristics were attributed to the effect of SBR operating temperature. To summarize, significant effects of SBR operating temperatures on effluent sCOD, UV₂₅₄ and TP were observed. The increased concentrations in these soluble components were attributed to reduced biological reaction rates at low temperatures.^7,30

The performance of in-line coagulation/flocculation of the effluents from the SBRs that were operated at different temperatures over a range of alum doses (0, 0.05, 0.1, 0.2, 0.5 and 1.0 mM as Al) was investigated. The monitored characteristics of the SBR effluents, with and without coagulation/flocculation before filtration, are illustrated in Fig. 2 for the two temperatures of SBR effluent considered. With in-line coagulation/flocculation, the reduction in sCOD was not significant until the alum dosages increased to 0.5 mM for both temperatures considered and reached 30 ± 2% and 40 ± 3% at the dosage of 1.0 mM for the effluents from the control and test SBR, respectively. This indicates limited sCOD was coagulated and incorporated into flocs at the low alum dosage range (0–0.2 mM), which was confirmed by the limited generation in turbidity (Fig. 2D) and SS (Fig. 2E) at the corresponding alum dosage range. This was likely due to different coagulation mechanisms occurring in the different ranges of doses. Low doses (0–0.2 mM) of coagulant have been found to only modify organic properties, without significant formation of precipitates,¹⁸ and this trend was not impacted by the temperatures of the SBR effluents considered in the present study. However, the low temperature SBR effluent resulted in higher residual sCOD due to greater sCOD generation at that SBR operating temperature.

Fig. 2 Monitored characteristics of SBR effluents, with and without coagulation/flocculation (c/f): (A) pH, (B) sCOD, (C) UV₂₅₄, (D) turbidity, (E) SS and (F) TP. (Error bars represent standard deviations).

The concentration of TP of the SBR effluents ranged from 2–3 mg L⁻¹ with the higher concentrations for the effluents from the test SBR (Fig. 2F). It was observed that alum dosages of 0.2 and 0.5 mM were required for the effluents from the test and control SBR to achieve a coagulated TP value of 0.1 mg L⁻¹, which is a typical design objective in a water resource recovery facility in Ontario, Canada.⁶ Dosages of 0.2 and 0.5 mM were subsequently compared with that required for fouling mitigation.

3.1.2 Removal of DOC and DOC fractions. The concentrations of DOC and DOC fractions (Table 1) in the effluents from control and test SBRs were compared using student t-tests. The concentrations of DOC, polysaccharide, protein and low molecular weight organics were significantly higher in the effluents from the test SBR than those of the control SBR. With in-line coagulation/flocculation, the DOC removal rate did not differ significantly between the two SBR effluents (Fig. 3). However, as illustrated in Fig. 4 the removals of organics with different molecular weights were significantly different for the two temperatures of SBR effluent. It was concluded that in-line coagulation/flocculation selectively removes the different DOC fractions.

Table 1 Effluent concentrations of DOC, DOC fractions, and the ratio of DOC fractions to DOC in the SBRs operated at 8 and 20 °C (± represents standard deviation of mean data)

SBR operating temperature (°C)		DOC	DOC fractions
			High MW		Intermediate MW	Low MW
			Polysaccharide	Protein
8	Concentration (mg L^−1)	14.9 ± 1	2.04 ± 0.15	0.78 ± 0.02	4.0 ± 0.20	4.75 ± 0.13
	Fraction of DOC (%)	—	13.7	5.2	26.9	31.9
20	Concentration (mg L^−1)	10.5 ± 1	0.69 ± 0.08	0.30 ± 0.04	4.1 ± 0.23	3.07 ± 0.12
	Fraction of DOC (%)	—	6.6	2.9	39.0	29.2

---

Fig. 3 Impact of alum dosage on DOC removal rates (error bars represent standard deviations).

Fig. 4 Impact of alum dosage on removal rates of (A) high MW organics, (B) low MW organics, (C) polysaccharide, and (D) protein (error bars represent standard deviations).

In terms of the high MW organics, at low-intermediate doses (0.05–0.20 mM) the removal rates with coagulation/flocculation were higher for the effluent from the test SBR than that of the control SBR. The concentrations of high MW organics are the sum of those of the polysaccharides and proteins. The removal rates for polysaccharides for the two SBR operating temperatures at the corresponding doses (0.05–0.2 mM) were not significantly different (Fig. 4C), however the removal rates for proteins were higher for the effluent from test SBR. Therefore, the higher removal in high MW organics by coagulation/flocculation for the effluent from the test SBR was attributed to the greater reduction in proteins. The significant removal in proteins at low-intermediate coagulant doses was likely due to the reactions between alum and the amino functional groups in the proteins.^14,31 However, when alum dosage increased to 0.5 mM, the removal rates for high MW organics were lower for the effluent from the test SBR than that of control SBR and this was attributed by both lower removal of polysaccharides and proteins. The different removal rates for high MW organics suggests that capture of these organics by precipitation and sweep flocculation was affected by their temperature dependent concentrations.

However, it was observed that the removal rates for low MW organics were relatively low and did not increase as coagulant dosage increased which indicates limited removal efficiency by coagulation/flocculation of low MW organics with alum. The low removal of low MW organics was more obvious for the effluent from the test SBR. The selective treatment efficiency of high MW organics might have been due to their larger variety of functional groups, allowing an enhanced interaction with metal hydroxide precipitates.³² It was previously demonstrated that high MW organics concentration correlated well with cake fouling¹⁶ and hence the cake fouling was expected to be mitigated in the subsequent filtration tests.

3.2 Fouling mitigation by in-line coagulation/flocculation versus secondary operating temperature

The impact of applying in-line coagulation/flocculation on membrane fouling for the two temperatures of SBR effluent considered was assessed at different alum dosages. Membrane fouling was characterized by total, hydraulically reversible and irreversible resistances (R_t,i, R_rev,i and R_irr,i). The slopes of linear regressions of the membrane resistances with respect to cumulative permeation cycle (i.e. cycle number), presented in Fig. 5, were used as estimates of the fouling rates. The fouling rates are listed in Table 2.

Fig. 5 Membrane fouling indices vs. alum dose for effluents from SBRs operated at (A) 8 °C and (B) 20 °C; dashed lines indicate the rates of fouling (error bars represent standard deviations).

Table 2 Fouling rates (mean ± standard deviation) at different dosages for the two temperatures of SBR effluent

Temperature of SBR effluent (°C)	Alum dosage (mM)	Total fouling rates (m^−1 min^−1)	Hydraulically reversible fouling rates (m^−1 min^−1)	Hydraulically irreversible fouling rates (m^−1 min^−1)
8	0	3.2 × 10^10 ± 5.9 × 10^8	2.8 × 10^9 ± 3.4 × 10^7	2.9 × 10^10 ± 7.2 × 10^8
	0.05	2.1 × 10^10 ± 5.1 × 10^8	2.6 × 10^9 ± 3.4 × 10^7	1.8 × 10^10 ± 3.7 × 10^8
	0.10	4.7 × 10^9 ± 2.4 × 10^7	—	6.8 × 10^9 ± 3.1 × 10^7
	0.20	—	—	2.8 × 10^9 ± 3.6 × 10^7
	0.50	—	—	—
	1.00	—	—	—
20	0	1.2 × 10^10 ± 2.5 × 10^8	—	1.1 × 10^10 ± 5.5 × 10^8
	0.05	6.4 × 10^9 ± 3.6 × 10^7	—	6.8 × 10^9 ± 4.4 × 10^7
	0.10	—	—	3.6 × 10^9 ± 3.7 × 10^7
	0.20	—	—	2.1 × 10^9 ± 4.5 × 10^7
	0.50	—	—	—
	1.00	—	—	—

---

Values of R_t,i were initially examined to assess the effect of alum dosage on the trends in the combined resistances over the multiple filtration cycles. The effluent from the control SBR was considered as a reference condition. It was observed that the R_t,i values in the first filtration cycle decreased by 15 ± 2% as dosage increased to 1.0 mM indicating moderate mitigation of total fouling by in-line coagulation/flocculation. The fouling rates were compared between doses to assess the impact of alum addition on longer term development of total resistance. It was observed that the fouling rate decreased by 47 ± 5% (P < 0.05) when the alum dosage was increased from 0 to 0.05 mM. Note that, the fouling rates were not significantly different from 0 for higher alum dosages, indicating no accumulation of R_t,i over the multiple filtration cycles (Table 2). The results demonstrate that in-line coagulation/flocculation for the effluent from the control SBR effectively reduced the development of total resistance at relatively low alum dosages.

Values of R_t,i for the effluent from the SBR operated at 8 °C were compared to those for the effluent from the control SBR. It was noted that without in-line coagulation/flocculation, the values of R_t,i were consistently higher for the effluents from the test SBR than those of the control. The greater R_t,i values associated with the effluent from the SBR operated at a lower temperature were attributed to the higher content of high and low MW organics (Table 2), that have previously been demonstrated to contribute to total membrane resistance.⁸ The R_t,i values in the first filtration cycle decreased modestly as the alum dosage increased from 0 to 1.0 mM which was consistent with the responses observed with the effluent from the control SBR. The fouling rates for the test SBR effluents decreased by 85 ± 6% as dosage increased to 0.10 mM. At higher alum dosages, the fouling rates were not significant, indicating that R_t,i values did not increase with filtration cycle. The reduction in initial R_t,i values and fouling rates by in-line coagulation/flocculation were both greater for the effluent from the test SBR when compared to the corresponding values for the effluent from the control SBR. However, the R_t,i values after in-line coagulation/flocculation were still higher for the effluent from the SBR operated at a lower temperature. The minimum coagulant dosages required to avoid R_t,i accumulation were 0.2 and 0.1 mM for the effluents from the test and control SBR, respectively. Although R_t,i values were constant with cycles at the dosage of 0.2 mM for the effluent from the test SBR. it is expected that a higher dosage is needed to reach equivalent R_t,i values of the effluent from the control SBR at the dosage of 0.1 mM.

The R_rev,i values were analyzed to obtain insights into the impact of alum addition on the development of hydraulically reversible fouling for the two temperatures of SBR effluents considered. For the control SBR, the hydraulically reversible fouling rates were not significant in all cases. This indicates that the conditions considered effectively prevented foulant accumulation. The means of the R_rev,i values were compared to gratify the impact of alum dosage. It was observed that when alum dosage increased from 0 to 0.5 mM, the mean R_rev,i values decreased by 90 ± 6% and no additional benefit was observed at higher alum dosages. The results indicate that of the coagulant dosage considered that 0.5 mM achieved the maximum improvement for the control SBR in terms of the control of hydraulically reversible fouling.

The impact of alum addition on hydraulically reversible fouling mitigation was then quantified for the effluent from the test SBR to determine whether temperature of SBR effluent had an impact on this response. With this effluent the hydraulically reversible fouling rates were significantly different from 0 for the non-coagulated effluent (2.8 × 10⁹ ± 3.4 × 10⁷ m⁻¹ per cycle) and the effluent that was dosed with 0.05 mM of alum (2.6 × 10⁹ ± 3.4 × 10⁷ m⁻¹ per cycle). The results indicate that foulants were accumulating on the membrane surfaces over multiple filtration cycle despite the use of back pulsing and air scouring at the end of each cycle. However, the rates of fouling became insignificant for dosages greater than 0.1 mM which was in alignment with the significant reduction in polysaccharides at this dosage (Fig. 4C). However, at all dosages between 0.1 to 1.0 mM the values of R_rev,i for the effluent from the test SBR were higher than those of the control. The results indicate that greater dosages were required to achieve equivalent reduction in R_rev,i values when the operating temperature of the SBR decreased.

Values of R_irr,i were then analyzed to assess how in-line coagulation/flocculation impacted the development of hydraulically irreversible fouling for the two temperatures of SBR effluent considered (Fig. 5). For the effluent from the control SBR, the hydraulically irreversible fouling rates decreased progressively with dosage until an 81% reduction in fouling rates was attained at 0.2 mM. The decreased rate of development of hydraulically irreversible fouling with increased alum dosage may have been due to the increased formation of flocs at higher doses that led to a thicker cake layer which would prevent the low MW organics from directly blocking membrane pores.³³ It was concluded that for the dosages considered that preferred for hydraulically irreversible fouling control was 0.2 mM for the effluent from the control SBR. Higher doses did not result in lower R_irr,i values.

The R_irr,i responses for the effluent from the test SBR were compared to that of the control. It was observed that the hydraulically irreversible fouling rate progressively decreased with alum dosage. When the dosage increased to 0.2 mM, the hydraulically irreversible fouling rate reduced by 90% relative to the undosed effluent. While the results indicate that R_irr,i values were constant with cycle, for dosages of 0.5 and 1.0 mM, the means of the R_irr,i values for the effluent from the test SBR were 40% higher than those of the control. Consequently, for the conditions investigated, hydraulically irreversible fouling could be controlled at a dosage of 0.2 mM for the effluents from the two temperatures of SBR effluent considered. The R_irr,i values were higher for the effluent from the test SBR. Additional increments in dosage above 0.2 mM did not substantially reduce hydraulically irreversible fouling. Therefore, for the coagulant dosages investigated the dosage of 0.2 mM represented an upper limit on the extent to which this response could be improved with alum for the two temperatures of SBR effluent considered.

3.3 Fouling mitigation mechanisms by in-line coagulation/flocculation

The filtration tests generated time series data on TMP development over multiple filtration cycles. The combined fouling models described by Bolton et al.²⁹ (Table S2†) were employed to extract summative parameters from the time series data. For all conditions investigated, the time series data could best be modelled using the combined cake-intermediate pore blocking model (SSE < 0.7 and R² > 0.8), suggesting that cake and intermediate blocking were the dominant contributors to membrane fouling. The parameters estimated for the combined cake-intermediate pore blocking model for the different dosages and the temperature of SBR effluents considered are presented (Fig. 6).

Fig. 6 Fitted model parameters (A) K_c and (B) K_i under different alum dosages (error bars represent standard deviations).

The impacts of alum dosage on cake fouling through intermediate pore blocking were evaluated by comparing the estimated rate for cake fouling (K_c) and intermediate pore blocking (K_i), respectively, for the effluents from the two temperatures of SBR effluent. As illustrated in Fig. 6A, the values of K_c were consistently higher for the effluent from the test SBR. This was attributed to the greater concentrations of high MW organics after in-line coagulation/flocculation. The K_c values declined asymptotically with dosage to a limiting value for the effluents generated at both temperatures. The asymptotic values were attained at doses greater than 0.5 mM and were 75% and 56% less than the un-dosed values for the effluents from the control and test SBR, respectively. The increased concentration of SS and the effective decrease of high MW organics after in-line coagulation/flocculation have been reported to facilitate the formation of a more porous cake layer which presented less fouling propensity.³⁴ Based on these results, it could be concluded that there is an alum dosage (0.5 mM) above which there is no further improvement in K_c values irrespective of temperature of SBR effluent. For the conditions investigated, the dosage of 0.5 mM was also found to meet common requirements for TP control which was previously discussed in section 3.1. Similarly, K_i values decreased asymptotically with dosage. The K_i values for the effluent from the test SBR were consistently higher than those of the control, and there was no indication that increasing the alum dosage could reduce the K_i values for the 8 °C effluents to match those of the 20 °C effluents. The results indicate that alum addition is only partially effective in reducing intermediate pore blocking and that its effectiveness declines as the temperature of SBR effluent declines.

Correlation between K_c and K_i values and the concentrations of high and low MW organics was examined separately for the two temperatures of SBR effluent considered because the initial concentrations were different for the two cases. The values of K_c were correlated with the concentration of high MW organics for the two temperatures of SBR effluent (Fig. 7); and not correlated with low MW organics (data not shown). The correlation coefficient for the effluent from the test SBR was 0.95, which was greater than that of the effluent from the control SBR (0.85). In addition, the K_c values were observed to be correlated to the concentrations of polysaccharides and proteins with the correlation coefficient with polysaccharides being greater than that with proteins (data not shown). Considering that the fraction of polysaccharides in the effluent from the test SBR was much higher than that of the control, the stronger correlation between K_c values and high MW organics was attributed to the greater production of polysaccharides for the SBR operated at lower temperatures. It was observed from Fig. 7 that the values of K_i were correlated with the concentrations of high MW organics for the two temperatures of SBR effluent; and not correlated with low MW organics. The lack of correlation between K_i values and low MW organic concentrations could likely be due to the relatively narrow range of concentrations that the membrane was exposed to as alum dosing did not substantially remove these materials. The observed decrease in pore blocking by coagulation/flocculation (Fig. 6) may have been due to the development of a cake layer from floc production that then entrapped the low MW organics and prevented them from blocking the membrane pores.³⁵ In summary, the results suggest that fouling mitigation by coagulation/flocculation with alum could be mainly attributed to the reduction of high MW organics.

Fig. 7 Linear correlation between fitted parameters and DOC fraction concentrations for SBRs operated at (A) 8 °C and (B) 20 °C.

3.4 Effect of filtration temperature on membrane performance

In the experiments described in the previous sections, the in-line coagulation/flocculation and filtration temperatures were maintained at 20 °C to avoid having more than one variable influencing the outcomes. However, in cold regions, in-line coagulation/flocculation and filtration are both conducted under low temperatures. Hence, the effect of temperature on the performance of in-line coagulation/flocculation and filtration should also be considered. The temperature of coagulation/flocculation, in the range of 2–8 °C has been found to have little effect on the destabilization of particles after alum addition when compared to a temperature of 22 °C.³⁶ Therefore, the effect of temperature on coagulation/flocculation of secondary effluents was neglected in the present study.

The impact of low filtration temperatures on the development of membrane resistance was previously reported for a range of DOC concentrations.²² In this prior work, the ratio of the total membrane resistances observed at 20 and 8 °C were found to range from 0.49 to 0.55 and there was no correlation between this ratio and the development of resistance. Hence, in the present study the total membrane resistances observed at the end of the filtration tests conducted with effluent from the test SBR were adjusted for filtration at 8 °C using the mean of the previously observed ratios (0.52).

The total membrane resistances when the filtration temperatures were the same as the SBR temperatures and the ratio of the values for 8 °C to the 20 °C systems are plotted versus alum dosage in Fig. 8. It was observed in-line coagulation/flocculation can effectively reduce the total membrane resistances at an alum dosage of 0.2 and 0.1 mM for the effluents from the test and control SBR, respectively. From Fig. 8 it can be seen that the total membrane resistances were more than two times higher for the test SBR effluents filtered at 8 °C than the control SBR effluents filtered at 20 °C for all alum doses. The results suggest that although in-line coagulation/flocculation can impede the accumulation of membrane resistance by reducing the DOC and DOC fraction concentrations that are produced under low SBR operating temperature, the ability to reduce the negative effect of low filtration temperature on membrane performance (i.e. membrane shrinkage) was limited. Hence, other membrane fouling mitigation strategies (i.e. enhanced membrane cleaning) should be considered along with in-line coagulation/flocculation when operating at cold temperatures.

Fig. 8 Total membrane resistance where filtration temperature was equivalent to SBR temperature.

4. Conclusions

The interaction between secondary operating temperature, in-line coagulation/flocculation performance and subsequent membrane fouling, was comprehensively studied for the first time in this novel study. Performance with respect to reduction in DOC fractions and fouling mitigation by in-line coagulation/flocculation was compared at two secondary operating temperatures (8 and 20 °C). In-line coagulation/flocculation was more effective in removing high MW organics than low MW organics. However, at the lower secondary operating temperature, lower treatment efficiencies of both high and low MW organics by in-line coagulation/flocculation were observed due to the greater concentration of these organics in the low temperature SBR effluent. The total, hydraulically reversible and irreversible resistance after coagulation/flocculation were consistently higher for the effluent from the SBR operated at 8 °C; this was attributed to higher residual high and low MW organics in these effluents. The preferred alum dosages to mitigate total membrane fouling were 0.2 and 0.1 mM, respectively, for the effluent from the SBR operated at 8 and 20 °C. When filtration was conducted at the same temperature as SBR operation, the total membrane resistances for the 8 °C system were two times higher than the 20 °C system regardless of alum dose. The result of the preferred coagulant dosage and the predominant mechanisms provide a design reference when in-line coagulation/flocculation is employed to mitigate fouling of tertiary membranes under low temperature conditions.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the Natural Science Research Project of Anhui Educational Committee (No. 2022AH050817), Scientific Research Foundation for High-level Talents of Anhui University of Science and Technology (No. 2022yjrc51), and the Natural Sciences & Engineering Research Council of Canada (NSERC).

References

1. F. Xiao, J. Huang, B. Zhang and C. Cui, Effects of low temperature on coagulation kinetics and floc surface morphology using alum, Desalination, 2009, 237(1–3), 201–213.
2. J. Yang, M. Monnot, T. Eljaddi, L. Ercolei, L. Simonian and P. Moulin, Ultrafiltration as tertiary treatment for municipal wastewater reuse, Sep. Purif. Technol., 2021, 272, 118921.
3. B. Liu, Y. Jun, C. Zhao, C. Zhou, T. Zhu and S. Shao, Using Fe(II)/Fe(VI) activated peracetic acid as pretreatment of ultrafiltration for secondary effluent treatment: Water quality improvement and membrane fouling mitigation, Water Res., 2023, 244, 120533.
4. A. Tikka, W. Gao and B. Liao, Reversibility of membrane performance and structure changes caused by extreme cold water temperature and elevated conditioning water temperature, Water Res., 2019, 151, 260–270.
5. T. Steinhauer, S. Hanély, K. Bogendörfer and U. Kulozik, Temperature dependent membrane fouling during filtration of whey and whey proteins, J. Membr. Sci., 2015, 492, 364–370.
6. S. Abu-Obaid, P. Bérubé and W. J. Parker, Characterization of performance of full-scale tertiary membranes under stressed operating conditions, Water Sci. Technol., 2020, 571–584.
7. C. Tao, W. Parker and P. Bérubé, Characterization and modelling of soluble microbial products in activated sludge systems treating municipal wastewater with special emphasis on temperature effect, Sci. Total Environ., 2021, 779, 146471.
8. C. Tao, W. Parker and B. Pierre, Evaluation of the impact of SBR operating temperature and filtration temperature on fouling of membranes used for tertiary treatment, Sep. Purif. Technol., 2022, 294, 121194.
9. L. Cui, C. Goodwin, W. Gao and B. Liao, Effect of cold water temperature on membrane structure and properties, J. Membr. Sci., 2017, 540, 19–26.
10. W. Wang, R. Li, F. Bu, Y. Gao, B. Gao and Q. Yue, et al., Coagulation and membrane fouling mechanism of Al species in removing humic acid: Effect of pH and a dynamics process analysis, Sep. Purif. Technol., 2023, 309, 123130.
11. K. Kimura and K. Kume, Irreversible fouling in hollow-fiber PVDF MF/UF membranes filtering surface water: Effects of precoagulation and identification of the foulant, J. Membr. Sci., 2020, 602, 117975.
12. K. Guo, Z. Wang, J. Wu, W. Li, B. Liu and Q. Yue, et al., Insights into the Mechanism of Coagulation Pretreatment for Membrane Fouling Control from the Perspective of Organic Removal Based on a Sensitive SEC-DAD-FLD-OCD/OND Method, ACS EST Eng., 2022, 851–861.
13. T. Liu, Y. Lian, N. Graham, W. Yu, D. Rooney and K. Sun, Application of polyacrylamide flocculation with and without alum coagulation for mitigating ultrafiltration membrane fouling: Role of floc structure and bacterial activity, Chem. Eng. J., 2017, 307, 41–48.
14. Z. Chen, B. Yang, Q. Wen and Y. Tang, Application of potassium ferrate combined with poly-aluminum chloride for mitigating ultrafiltration (UF) membrane fouling in secondary effluent: Comparison of oxidant dosing strategies, Chemosphere, 2021, 274, 129862.
15. J. Liu, K. He, J. Zhang, C. Li and Z. Zhang, Coupling ferrate pretreatment and in-situ ozonation/ceramic membrane filtration for wastewater reclamation: Water quality and membrane fouling, J. Membr. Sci., 2019, 590, 117310.
16. C. Tao, W. Parker and P. Bérubé, Interaction of operating HRT and temperature on fouling of tertiary membranes treating municipal wastewater, Sci. Total Environ., 2023, 858, 159812.
17. P. Jutaporn, R. M. Cory, P. C. Singer and O. Coronell, Efficacy of selected pretreatment processes in the mitigation of low-pressure membrane fouling and its correlation to their removal of microbial DOM, Chemosphere, 2021, 277, 130284.
18. N. M. Peleato, R. L. Legge and R. C. Andrews, Characterization of UF foulants and fouling mechanisms when applying low in-line coagulant pre-treatment, Water Res., 2017, 126, 1–11.
19. W. Yu, N. Graham, H. Liu and J. Qu, Comparison of FeCl3 and alum pre-treatment on UF membrane fouling, Chem. Eng. J., 2013, 234, 158–165.
20. T. Liu, Z. Chen, W. Yu, J. Shen and J. Gregory, Effect of two-stage coagulant addition on coagulation-ultrafiltration process for treatment of humic-rich water, Water Res., 2011, 45(14), 4260–4268.
21. E. Akhondi, F. Wicaksana and A. G. Fane, Evaluation of fouling deposition, fouling reversibility and energy consumption of submerged hollow fiber membrane systems with periodic backwash, J. Membr. Sci., 2014, 452, 319–331.
22. C. Tao, W. Parker and B. Pierre, Assessing the role of cold temperatures on irreversible membrane permeability of tertiary ultrafiltration treating municipal wastewater, Sep. Purif. Technol., 2022, 278.
23. S. a. Huber, Getting familiar with LC-OCD Data Processing and Data Interpretation, LC-OCD -OND Tutorial, 2010.
24. S. Huber, A. Balz, M. Abert and W. Pronk, Characterisation of aquatic humic and non-humic matter with size-exclusion chromatography - organic carbon detection - organic nitrogen detection (LC-OCD-OND), Water Res., 2011, 45(2), 879–885.
25. A. Nguyen, J. Tobiason and K. Howe, Fouling indices for low pressure hollow fiber membrane performance assessment, Water Res., 2011, 45(8), 2627–2637.
26. P. van den Brink, O. A. Satpradit, A. van Bentem, A. Zwijnenburg, H. Temmink and M. van Loosdrecht, Effect of temperature shocks on membrane fouling in membrane bioreactors, Water Res., 2011, 45(15), 4491–4500.
27. C. Chen, M. Sun, J. Chang, Z. Liu, X. Zhu and K. Xiao, et al., Unravelling temperature-dependent fouling mechanism in a pilot-scale anaerobic membrane bioreactor via statistical modelling, J. Membr. Sci., 2022, 644, 120145.
28. F. Meng, S. Zhang, Y. Oh, Z. Zhou, H. S. Shin and S. R. Chae, Fouling in membrane bioreactors: An updated review Fangang, Water Res., 2017, 114, 151–180.
29. G. Bolton, D. LaCasse and R. Kuriyel, Combined models of membrane fouling: Development and application to microfiltration and ultrafiltration of biological fluids, J. Membr. Sci., 2006, 277(1–2), 75–84.
30. L. A. Lishman, R. L. Legge and G. J. Farquhar, Temperature effects on wastewater treatment under aerobic and anoxic conditions, Water Res., 2000, 34(8), 2263–2276.
31. P. Arunkumar, V. Sadish Kumar, S. Saran, H. Bindun and S. P. Devipriya, Isolation of active coagulant protein from the seeds of Strychnos potatorum–a potential water treatment agent, Environ. Technol., 2019, 40(12), 1624–1632.
32. J. Haberkamp, A. S. Ruhl, M. Ernst and M. Jekel, Impact of coagulation and adsorption on DOC fractions of secondary effluent and resulting fouling behaviour in ultrafiltration, Water Res., 2007, 41(17), 3794–3802.
33. K. Kimura, T. Kakuda and H. Iwasaki, Membrane fouling caused by lipopolysaccharides: A suggestion for alternative model polysaccharides for MBR fouling research, Sep. Purif. Technol., 2019, 223, 224–233.
34. J. Liu, Z. Zhang, Z. Liu and X. Zhang, Integration of ferrate (VI) pretreatment and ceramic membrane reactor for membrane fouling mitigation in reclaimed water treatment, J. Membr. Sci., 2018, 552, 315–325.
35. R. Naim, R. Epsztein, A. Felder, M. Heyer, M. Heijnen and V. Gitis, Rethinking the role of in-line coagulation in tertiary membrane filtration of municipal effluents, Sep. Purif. Technol., 2014, 125, 11–20.
36. F. Xiao, J. Ma, P. Yi and J. Huang, Effects of low temperature on coagulation of kaolinite suspensions, Water Res., 2008, 42(12), 2983–2992.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ew00239c

---