Metals, heavy metals and microorganism removal from spent filter backwash water by hybrid coagulation-UF processes

Recently, water reuse of spent filter backwash water (SFBW) has attracted particular attention in most countries because of water scarcity. The main point is that the SFBW acts as a permanent source until the water treatment plant is working. During most water treatment processes, SFBW is generated from 2 to 10% of the total plant production (Raj et al. 2008). Filter backwashing is conducted to remove all captured material through the filter bed during filtration. Thus, there are some concerns regarding its reuse because of high concentration of metals, heavy metals, natural organic matters, microorganisms, and colloidal materials (Walsh et al. 2008).

Heavy metals’ exposure routes include absorption, inhalation, and ingestion. Ingestion through drinking water is the major source of heavy metals’ exposure in some areas. Heavy metal contamination in drinking water causes health problems such as shortness of breath, neurotoxic, mutagenic and teratogenic effects with various types of cancers depending on the heavy metal type (Chowdhury et al. 2016). It is obvious that direct recycling of SFBW can jeopardize the quality of the treated water because of the high concentration of contaminants. SFBW treatment can be conducted by different methods. Previous studies have shown that membrane separation is an effective way. Membrane filtration such as microfiltration and ultrafiltration (UF) remove particulates, colloids, and pathogens effectively and has been believed to be the water treatment technique of the 21st century. They need a smaller footprint, have low energy consumption and produce clean water (Ang et al. 2015); however, fouling is a big problem in membrane processes, especially for SFBW treatment.

To overcome this problem, a combination of membrane system with chemical or physical processes, such as coagulation and sedimentation, has been introduced (Wang et al. 2014; Lai et al. 2015). UF acts as an effective barrier to microorganisms, suspended particles, and colloids (Peter-Varbanets et al. 2011; Chen et al. 2014). Coagulation is an effective process to remove the majority of the particulate matter, suspended solids, colloids, organic matter and some dissolved matter (Gao & Yue 2005; Zhao et al. 2008; Yu et al. 2013) before the membrane process.

Most previous articles about backwash water treatment have addressed the fouling issue (Huang et al. 2011; Li et al. 2012; Yu et al. 2013; Wang et al. 2016). Only a few have dealt with the quality of treated water for drinking or other application (Raj et al. 2008; Zhang et al. 2008).

The purpose of this study was to understand the level of contaminants and health risk related to metals (Fe and Al), heavy metals (Pb, As, and Cd), and microorganisms (total and fecal coliforms (TC/FC)) in the SFBW of Isfahan Water Treatment Plant before and after treatment with pilot-scale hybrid coagulation-UF processes, and comparison between the two coagulants: poly-aluminum ferric chloride (PAFCl) and FeCl₃, as pretreatment.

Samples of SFBW were collected from Babashaikhali water treatment plant in Isfahan, Iran, during the winter season. This plant treats 12 m³/s of water using coagulation, flocculation, clarification, and sand filtration processes. In this plant, PACl is used as a coagulant and 48 filters purify the water. Each filter backwash needs 500 m³ of water. Considering 48 filters with a 24 h cleaning interval, it accounts for over 2.25% of the raw water entering the plant. Hence, during the water treatment process, approximately 24,000 m³/d of SFBW is generated. The characteristics of raw SFBW are listed in Table 1. In this study, about 3,000 L of raw SFBW was collected from the water treatment plant. It was transferred to the laboratory and used for examination.

In this study, continuous processes including primary sedimentation, coagulation, flocculation, clarification, and UF were used for SFBW treatment. For all the sections of the pilot unit, except the UF membrane, the flow rate was 10 L/h. Hydraulic retention time (HRT) for this section, except the UF membrane, was 60, 6, 48, and 192 min. Figure 1 shows the experimental set-up of the pilot unit used in the experiments. According to the authors’ previous study (Mahdavi et al. 2016), coagulation with PAFCl and FeCl₃ was conducted at the natural pH of SFBW (8.3), and pre-determined dosage of PAFCl (15 mg/L) and FeCl₃ (40 mg/L) was continuously added into the rapid mixing unit (speed was 80 rpm and HRT was 6 min). The coagulated water then passed though the two flocculation tanks, with a 40 rpm mixing intensity. After this, water was introduced for clarification, and then the treated water was fed to the UF membrane module. Investigation for the optimum dose selection was carried out for both coagulants in a continual manner for 4 days separately (Figure 1, section 4). The optimum dose was selected to produce the best water quality based on turbidity and color results. These parameters were analyzed after two HRT of the second clarification (inflow was 10 L/h). After that about 1,000 L of raw SFBW was treated with PAFCl and FeCl₃ separately. Then, 800 L of treated water with PAFCl and 800 L of treated water with FeCl₃ entered the UF membrane process separately (inflow was 8 L m⁻² h⁻¹). Some parameters, such as turbidity, color, pH, TC, and FC were detected about ten times during the pilot operation, but metals and heavy metals were detected three times in the optimum dose and quality of treated water.

The UF membrane was made of hollow-fiber polypropylene, with a nominal pore size of 0.01–0.2 μm. The total membrane area of UF was 0.1 m²/module. The UF module was operated in dead-end mode with constant filtration of about 8 L m⁻² h⁻¹ at a trans-membrane pressure of 300 Pa. It was operated in a cycle of 60 min filtration and 1 min backwashing with permeate in the reverse direction.

Samples were analyzed for residual aluminum (Al), iron (Fe), lead (Pb), arsenic (As) cadmium (Cd), TC, and FC. All the experiments were conducted according to the Standard Methods for the Examination of Water and Wastewater.

Turbidity, color, and pH of the samples were measured by TN-100 (EUTECH) turbid meter, DR 5000-HACH LANGE and pH-meter model CG 824, respectively. Al was analyzed by DR 5000-HACH LANGE, method 8326. Fe, Pb, As, and Cd were analyzed by inductively coupled plasma (ICP).

From Table 1 it can be seen that the SFBW sample had high turbidity, color, Fe, Al, and heavy metals’ concentration in comparison with raw water. Of course, the quality of raw water was very high because of very low turbidity, metals and heavy metals’ concentration. From Table 1 it can be concluded that the concentrations of metals and heavy metals in raw water were lower than the EPA guideline. On the other hand, raw SFBW has a very high concentration of turbidity, color, Fe, Al, Pb, TC, FC and, to some extent, Cd. Results showed that during the water treatment process, water contaminants were removed or accumulated on filter beds. Subsequently, filter backwash removed this material from the filters. Low concentrations of TC and FC in SFBW may be related to pre-ozonation and pre-chlorination in the water treatment plant. Metals and heavy metal concentrations in SFBW samples were found to be in the order: Fe > Al > Pb > Cd > As.

The effect of various doses (5 to 60 mg/L) of PAFCl and FeCl₃ is presented in Figure 2, to determine the optimum dose. Results showed that the optimum dose for PAFCl and FeCl₃ was 15 and 40 mg/L, respectively. At these doses, turbidity removal was 99.6 and 99.4%, respectively (Figure 2). For each dose, turbidity was measured after two HRT. The removal efficiency of turbidity for PAFCl was increased to 15 mg/L. Above this point, re-stabilizing of colloidal suspension and, subsequently, an increase in turbidity occurred. While for FeCl₃, the removal efficiency of turbidity increased until 40 mg/L and then it worsened above 40 mg/L due to re-stabilization of colloidal suspension. Therefore, coagulation was conducted with two doses and treated water was analyzed for predetermined parameters.

For the membrane process, the quality of input water is very important because of fouling problems. In this study, by applying optimum doses of PAFCl and FeCl₃, treated water turbidity reached 2.4 and 3.9 NTU, respectively. After this, treated SFBW was used in the UF membrane process and microbial quality of treated water was investigated.

With regards to coagulation and flocculation, most bacteria and protozoa are considered as particles, and most viruses as colloidal organic particles. Thus, removal of turbidity has an indirect relationship with microbial reduction. In this result, the two coagulants showed a different influence on microbial reduction. From Figure 3 it can be seen that PAFCl showed good removal efficiency in comparison with FeCl₃.

Under optimal coagulation conditions with PAFCl (15 mg/L), the average removal of TC and FC was 88 and 79%, and by PAFCl-UF process, it reached 100 and 100%, respectively. Also for FeCl₃ (40 mg/L), removal efficiency for TC and FC was 81 and 72%, respectively. This value in the FeCl₃-UF process reached 100 and 100%, respectively. During coagulation processes, turbidity, colloids and microorganisms were removed by various mechanisms such as adsorption, precipitation, charge neutralization, and enmeshment (Ebrahimi et al. 2015).

Low coagulant consumption and good removal efficiency is related to the charge neutralization mechanism. In this study, PAFCl with a lower dose of 15 mg/L gave better efficiency for turbidity, TC and FC removal and is related to positively charged PAFCl flocs that can adsorb negatively charged colloids and microorganisms. While for FeCl₃, it is necessary to add a higher dose to get enough positive charge. Therefore, in this situation, the mechanism changed to sweep-flocculation. PAFCl is a pre-hydrolyzed coagulant that had more positive charge in comparison with the traditional coagulants. This coagulant has Al and Fe ions and can produce Fe(OH)₃ and Al(OH)₃ complex (Edzwald 1993). In UF membrane process, physical sieving and straining was the main mechanism for the removal of particles and microorganisms. The size of most bacteria is about 0.1 to more than 10 μm, while the pore sizes of UF membranes are smaller than bacteria, so in this study, the very low pore size of UF removed all TC, FC, and particulate matter; this result is in accordance with the study of Xiao et al. (2012). Also, Reissmann & Uhl's (2006) study on reuse of SFBW from drinking water treatment showed that during the whole experiment, coliforms, Escherichia coli and Clostridium were not detected in the permeate. Thus, these results prove that UF is a very good barrier for TC/FC.

The quality of SFBW treated with coagulation and the UF membrane process is presented in Table 2. Coagulation was conducted with PAFCl and FeCl₃ at optimum doses. It can be seen that the two coagulants showed a different influence on SFBW quality. From Table 2, it can be concluded that PAFCl showed good removal efficiency for all the experimental parameters, except residual Al concentration.

As indicated in Table 2, residual turbidity and true color after coagulation with PAFCl and FeCl₃ reached 2.4 ± 0.14, 3.9 ± 0.14 NTU and 2 ± 0, 5 ± 1.4 Pt. Co. units, respectively, and in PAFCl-UF and FeCl₃-UF output, they reached 0.1 NTU and 0 Pt. Co. units, respectively. The level of these parameters in the UF filtrate is very low and meets the US-EPA drinking water standard. The most important parameters in SFBW reuse are metals and heavy metals that have very important effects on human health (Berthon 2002). They are removed by different mechanisms such as adsorption, precipitation, and enmeshment during water treatment processes. Also, under different conditions, diverse mechanisms such as charge neutralization, precipitation, bridge-aggregation, adsorption, and sweep-flocculation help to remove particles, organic substances, and other pollutants during coagulation (Ebrahimi et al. 2015). The results of this study showed that the concentration of iron and aluminum in raw SFBW was 4 ± 0.14 and 0.4 ± 0.028 mg/L, respectively. After coagulation with PAFCl and FeCl₃ at optimum doses, Fe concentration reached 0.02 ± 0.01 and 0.12 ± 0.01 mg/L, respectively and after PAFCl-UF and FeCl₃-UF processes, it reached 0 and 0.08 ± 0.004 mg/L, respectively. Al concentration after coagulation with PAFCl and FeCl₃ at optimum doses was 0.046 ± 0.02 and 0.031 ± 0.01 mg/L, respectively, and after PAFCl-UF and FeCl₃-UF processes, it reached 0.031 ± 0.002 and 0.02 ± 0.001 mg/L, respectively. SFBW treated with PAFCl has, to some extent, more Al concentration than SFBW treated with FeCl₃, and SFBW treated with FeCl₃ has, to some extent, more Fe concentration than SFBW treated with PAFCl. Maybe it is related to composition property of the coagulant that repels some Fe or Al in treated water. However, the concentrations of Fe and Al in treated SFBW were low and met the drinking water standards.

The concentration of Pb, As, and Cd in raw SFBW was 217 ± 9.9, 2.36 ± 0.5 and 4 ± 0.7 μg/L, respectively. After coagulation with PAFCl (at optimum doses), it reached 0, 0, and 0.19 ± 0.012 μg/L, respectively. Also, for FeCl₃, it reached 0, 0. and 0.35 ± 0.08 μg/L, respectively. The concentration of Pb, As, and Cd in SFBW treated with PAFCl-UF process was 0, 0, and 0.15 ± 0.01 μg/L, respectively, and for SFBW treated with FeCl₃-UF process, it was 0, 0, and 0.3 ± 0.07 μg/L, respectively. It can be seen that both coagulants had good efficiency for heavy metals’ removal. However, PAFCl removed heavy metals better than FeCl₃. Metals and heavy metals’ concentrations in SFBW treated with PAFCl were found to be in the order: Al > Fe > Cd > As and Pb. For coagulation with FeCl₃, this order was Fe > Al > Cd > As and Pb. After the PAFCl-UF process, the order of metals and heavy metals was Al > Cd > Fe, As and Pb. Eventually, this order for the FeCl₃-UF process was Fe > Al > Cd > , As and Pb.

It can be seen that coagulation reduced most of the metals and heavy metals in treated SFBW, therefore, it can be concluded that the highest amount of metals and heavy metals are related to constituents that are attached to particles, organic matter, clay or silt during the water treatment process. Thus, with the removal of particles, colloids, and suspended solids during coagulation, most of the metals and heavy metals were removed effectively. Previous studies have shown that clay and soil have very good potential for heavy metal sorption and heavy metals like Pb were primarily located in the silt fraction (Appel & Ma 2002; Sutherland 2003). Also, other studies have proposed that Fe-oxide fraction has the ability to capture heavy metals (Orrono & Lavado 2009).

In this study, coagulation with FeCl₃ that occurred under relatively higher coagulant concentration (40 mg/L), could have resulted from the sweep-flocculation process; so, by enmeshment of particles and organic matter in the formed flocs, all attached metals and heavy metals were consequently removed. PAFCl is a pre-hydrolyzed coagulant that has more positive charges in comparison with the traditional coagulants. It has both Al and Fe ions; so, it uses both properties of Fe(OH)₃ and Al(OH)₃ complex. Thus, positively charged PAFCl flocs can adsorb negatively charged organic matter, colloids, and particles that attracted metals and heavy metals (Ebrahimi et al. 2015). On the other hand, coagulation with PAFCl occurs under relatively lower coagulant concentration which could be a sign of the adsorption mechanism. Thus, it is predicted that the dominant mechanism for metals and heavy metals’ removal by PAFCl is adsorption.

The CDI and HRI values of the selected metals and heavy metals, before and after treatment, are summarized in Tables 4 and 5. HRI value less than one is considered to be safe for consumers (Khan et al. 2008).

The results showed that CDI values of Fe, Al, Pb, As, and Cd in raw SFBW were 111.1, 11.1, 6, 0.1, and 0.11 μg/kg-day, respectively, for adults and they were 122.3, 12.2, 6.6, 0.07, and 0.12 μg/kg-day, respectively, for children. For SFBW treated with PAFCl, the CDI values of Fe, Al, Pb, As, and Cd for adults reached 0.556, 1.278, 0, 0, and 0.005 μg/kg-day, respectively. Also for children, these amounts reached 0.612, 1.4, 0, 0, and 0.006 μg/kg-day, respectively. For SFBW treated with FeCl₃, the CDI values of the mentioned parameters for adults reached 3.33, 0.861, 0, 0, and 0.01 μg/kg-day, respectively, and for children they reached 3.67, 0.948, 0, 0, and 0.011 μg/kg-day, respectively.

The amounts of CDI for Fe, Al, Pb, As, and Cd in SFBW treated with PAFCl-UF process were 0, 0.861, 0, 0, and 0.004 μg/kg-day, respectively, for adults and 0, 0.948, 0, 0, and 0.005 μg/kg-day, respectively, for children. For SFBW treated with FeCl₃-UF process, the CDI values of Fe, Al, Pb, As, and Cd for adults were 2.22, 0.556, 0, 0, and 0.008 μg/kg-day, respectively, while for children they were 2.44, 0.612, 0, 0, and 0.009 μg/kg-day, respectively. It can be seen that CDI values in raw SFBW were greater than that of the treated SFBW. Also, for SFBW treated with PAFCl and PAFCl-UF processes, these values were lower in comparison with that of FeCl₃ and FeCl₃-UF processes, except for aluminum (Table 3).

With regards to the raw and treated SFBW quality in this study, the CDIs of metals and heavy metals for raw SFBW were found to be in the order: Fe > Al > Pb > Cd > As. For SFBW treated with PAFCl-UF process, the order was Al > Cd and the others were zero, while for FeCl₃-UF process, it was Fe > Al > Cd and the others were zero.

The health risk indices’ values of Fe, Al, Pb, As, and Cd in raw SFBW for adult consumption were 0.16, 0.002, 1.72, 0.22, and 0.22, respectively (Table 5). It can be seen that raw SFBW has a very low concentration of metals and heavy metals. Also, HRIs calculation showed that the values were less than 1 except for Pb, which indicates no risk of exposure to the use of this water. The same results were observed for child consumption.

The HRIs indices’ values of Fe, Al, Pb, As, and Cd in SFBW treated with PAFCl-UF process for adult consumption were 0, 0.00012, 0, 0, and 0.0083, respectively, and for children they were 0, 0.00013, 0, 0, and 0.0091, respectively. These values for SFBW treated with the FeCl₃-UF process and adult consumption were 0.0031, 0.00008, 0, 0, and 0.0167, respectively, and for children they were 0.0035, 0.00009, 0, 0, and 0.0183, respectively.

It can be seen that HRIs indices’ values of Fe, Al, Pb, As, and Cd in SFBW treated with PAFCl-UF and FeCl3-UF processes were less than 1 which indicates no risk of exposure to the use of this water. By comparing between the quality of water produced with PAFCl-UF and FeCl3-UF processes, it can be seen that HRIs indices’ values for Fe and As are lower in the PAFCl-UF process while HRIs’ value of Al is lower in the FeCl₃-UF process. Maybe, this is because PAFCl contains the Al component. In this study, the concentration of Fe and Cd in water treated with PAFCl-UF was lower than that of the FeCl₃-UF process, and Pb and As concentration were zero for both processes. All the concentrations of metals and heavy metals in the SFBW treated with both processes met the drinking water standard according to EPA guidelines. Based on the quality of SFBW treated with the PAFCl-UF process, the HRIs of selected metals and heavy metals were found to be in the order: Cd > Al > Fe = Pb = As (HRIs values for Fe, Pb, and As were zero). For the FeCl₃-UF process, this order was Cd > Fe > Al > Pb = As (HRIs values for Pb and As were zero).

• SFBW treated with the coagulation-UF membrane process was colorless and had a turbidity of 0.1 NTU, and TC/FC were undetectable.

• All concentrations of metals and heavy metals in SFBW treated with both processes met the drinking water standard according to EPA guidelines.

• PAFCl showed good removal efficiency in comparison with FeCl₃.

• Based on the quality of SFBW treated with the PAFCl-UF process, the HRIs of selected metals and heavy metals were found to be in the order: Cd > Al > Fe = Pb = As (HRIs values for Fe, Pb, and As were zero). For the FeCl₃-UF process, this order was Cd > Fe >Al > Pb = As (HRIs values for Pb and As were zero).

• HRIs indices’ values of Fe, Al, Pb, As, and Cd in SFBW treated with PAFCl-UF and FeCl₃-UF processes were less than 1, which indicates no risk of exposure to the use of this water.

The authors acknowledge the fund provided by the Isfahan University of Medical Science, School of Health, Environmental Health Engineering Department, Isfahan, Iran for this PhD thesis with 394281 code number. This article is dedicated to Shawkat Shaikhmoradi.