Upgrading the eastern wastewater treatment plant effluent quality in Egypt for reuse by means of in-line coagulation followed by ultrafiltration

BOD

biological oxygen demand

COD

chemical oxygen demand

DOC

dissolved organic carbon

IRF

irreversible fouling coefficient

MWCO

molecular weight cut-off

PVDF

polyvinyl-dene fluoride

PES

polyethersulfone

RF

reversible fouling coefficient

R_t

total fouling coefficient

TMP

transmembrane pressure

TSS

total suspended solids

UF

ultrafiltration

UV₂₅₄

ultraviolet absorbance at 254 nm

The reclamation of wastewater has become one of the most reliable strategies to tackle the persisting problem of water scarcity worldwide (Lee et al. 2019). The Eastern Wastewater Treatment Plant (EWWTP) is the second largest treatment plant in Egypt with a capacity of 800,000 m³/day and consists of preliminary, primary, and secondary treatments. To make the effluent reusable, advanced treatment is required. Many studies have been conducted to investigate the appropriate technology to improve the quality of the secondary effluent obtained from the EWWTP (Moursy et al. 2018; AbdelMoula et al. 2021) or even other Wastewater Treatment Plant (WWTP) in the Middle East (Arif et al. 2018; Moursy et al. 2018; Arif et al. 2020; AbdelMoula et al. 2021; Juneidi et al. 2022). However, all of these studies used mathematical modeling, by means of the GPS-X (v.8) software. Membrane technology, applied as a membrane bioreactor, has been demonstrated as an effective technology in providing the highest quality with approximately minimal health risks. Owing to the inherent advantages of ultrafiltration (UF), namely, the removal of colloids, suspended solids, and microorganisms with great efficiency in addition to its low-pressure requirements, it has been a popular option in wastewater treatment (Liu et al. 2018). UF could also remove the nutrients, such as phosphorous and nitrogen, which exist in the wastewater and lead to a very serious problem when being discharged, eutrophication. For an illustration, Zheng et al. (2012) recorded an efficient removal of phosphorous found in secondary effluent using UF. However, the applicability of membranes is still limited due to fouling, the major curb in membranes (Yu et al. 2016b; Liu et al. 2019a; Jun et al. 2020).

Concerning the secondary effluent, the main contributor to UF fouling is the effluent organic matter. In general, it is composed of natural organic matter, which involves biopolymers (such as polysaccharides and proteins), humics, and soluble microbial products, such as extracellular polymeric substances (Tian et al. 2018). Their accumulation can cause severe fouling (Huang et al. 2007). Moreover, many researchers have noticed a lot of parameters that impacted fouling considerably, such as membrane properties like morphology, charge, and hydrophobicity. Also, the solution properties include calcium concentration, ionic strength, and pH, in addition to the characteristics of the organic matter such as molecular weight, hydrophobicity, charge density, impact fouling rate (Aly 2015; Schäfer et al. 2000; Fan et al. 2001; Howe & Clark 2002; Ladouceur & Narbaitz 2022). Thus, many pretreatment methods have been implemented to control fouling, resulting in an improvement in the effluent of the membrane.

Among different types of pretreatments, the effectiveness and reliability of chemical coagulation have been demonstrated, not only in improving the water quality but also in alleviating membrane fouling (Liu et al. 2018; Li et al. 2020; Xu et al. 2022). Coagulants can convert the hydrophobic charged substances into large flocs that could be removed by either sedimentation or by being retained on the membrane's surface (Zhang et al. 2015). As compared to traditional coagulation, in-line coagulation, which does not include the sedimentation phase, showed better results and effectiveness (Guigui et al. 2002), and it could slow down the deposition of the contaminants, which makes the formed cake layer loose (Chen et al. 2007). Xu et al. (2021) demonstrated the efficiency of combining in-line coagulation and UF using titanium xerogel to treat synthetic algae-loaded water, such that the coagulated flocs improved the UF efficiency by reducing the fouling resistance. It has been reported that alum (5.0 mg/L) and ferric chloride (FeCl₃) (5.0 mg/L), which were injected in-line before UF during the treatment of secondary effluent (Aly et al. 2021), removed 26.0 and 35.0% of dissolved organic carbon, in addition to 95.0 and 98.0% of turbidity, respectively. Besides, alum was found to be efficient in mitigating the flux decline in a system rich with humics (Listiarini et al. 2009). However, the coagulants could not remove all the pollutants, especially those organics of low and medium molecular weights, which cause severe fouling (Yu et al. 2016a; Chen et al. 2017). Owing to the beneficial effect of oxidation (e.g. preozonation) in alleviating fouling, removing low molecular weight organics, and reducing microorganisms (Liu et al. 2011; Yu & Graham 2015), there has been a great interest in combining it with coagulation as a pretreatment to the membrane. This combination led to great results in controlling fouling and improving water characteristics (Liu et al. 2011; Van Geluwe et al. 2011; Lin et al. 2013).

Different oxidants have been used, e.g. ozone, chlorine, and permanganate, combined with coagulation. Despite being effective in the removal of organics and control of fouling (Tian et al. 2013; Wang et al. 2016; Yu et al. 2016a), almost all of them have reverse effects. For example, ozone and chlorine may produce disinfection byproducts. Permanganate may increase magnesium concentration in the effluent. On the contrary, ferrate (Fe^VI) has no chemical residual; thus, it is considered an environmentally friendly chemical (Yu et al. 2016b; Liu et al. 2018). Its oxidation potential is very high (E^o = 2.2 V), greater than that of ozone (2.07 V). It is not only an effective oxidant but also a reliable coagulant, as it can be reduced to Fe(III) types (Liu et al. 2019a, b) in addition to being an effectual disinfectant (He et al. 2021). As a result, it has been used with other coagulants in the treatment of water (Amano et al. 2018), in which integration between oxidation (ferrate) and coagulation (poly-aluminum chloride) was able to improve water quality. This was proved by another study, by Chen et al. (2020), which confirmed that this integration mitigated fouling considerably during secondary effluent treatment. Al Umairi et al. (2021) reported that the usage of ferrate (10.0 mg/L) and alum (6.0 mg/L) showed effective results in minimizing the microbial level, organics, and inorganics when treating wastewater.

To date, no study, as far as we know, has investigated the advanced treatment of the EWWTP practically by a polyethersulfone (PES) UF for reuse purposes; only models were used. Unlike ferric chloride and alum, for which several studies have been conducted to analyze their performance as in-line coagulants for attenuating fouling (Tang et al. 2019; Aly et al. 2021; Brover et al. 2022; Ladouceur & Narbaitz 2022), the performance of either sodium ferrate (Na₂FeO₄) solely or in combination with ferric chloride, as an in-line coagulant, was also monitored and compared to them. Thus, the main aims of this study were as follows: (1) investigate the performance of in-line coagulation with UF in improving the secondary effluent of EWWTP for reuse purposes and the applicability of the permeate in agriculture irrigation; (2) assign the most optimum dosages of each investigated coagulant; (3) compare the effect of these dosages on enhancing UF performance; (4) study the impact of applying coagulation and oxidation together; and (5) investigate the correlation between influent characteristics and membrane fouling.

The secondary effluent investigated in this study was obtained from the EWWTP, in Alexandria, Egypt, which was dug up twice a week. The treatment type was activated sludge consisting of 12 aeration tanks and 12 sedimentation tanks. Table 1 displays its characteristics as collected from the plant. The studied water was moved, then stored in polyethylene tanks (22.0 L), held at 4.0 °C in the refrigerator, and then pumped to the membrane modules when the temperature had been brought back to room temperature, 20 °C.

Before doing UF tests, separate jar test experiments were conducted utilizing a programmable six-paddle gang stirrer (217011163, Phipps and Bird, USA) to determine the ideal pretreatment conditions. The jar test was conducted in the Faculty of Engineering's Sanitary Engineering Laboratory, Alexandria University, at room temperature using the coagulation and flocculation jar test of water standard procedure. The dosages of each coagulant had been added after the secondary effluent had first been mixed for 60 s. Based on the literature (Yu et al. 2016b; Aly et al. 2021; Ding et al. 2022; Yang et al. 2022), different dosages of each coagulant were selected. Sodium ferrate (Na₂FeO₄) was tested at seven different concentrations: 0.5, 2.5, 5.0, 7.5, 10.0, 12.5, and 15.0 mg/L. While for ferric chloride (FeCl₃), six dosages – 5.0, 7.5, 10.0, 12.5, 15.0, and 20.0 mg/L – were examined. In the same way, six alum dosages – 15.0, 20.0, 25.0, 30.0, and 50.0 mg/L – were looked at. Three mixing ratios, 2:1 (5.0:2.5 mg/L), 4:1 (5.0:1.25 mg/L), and 10:1 (5.0:0.5 mg/L), were tested for the integration of ferric chloride and sodium ferrate. Turbidity, COD, and bacteriology have been monitored for jar test analysis, but turbidity was the parameter that could be controlled, which made it easier to choose the appropriate dosages (Aly et al. 2021; Yang et al. 2022).

The UF experiments were run in a dead-end mode using two different dosages of each coagulant as in-line coagulants. Each experiment involved injecting the coagulant into the secondary effluent feed tank and mixing it quickly for 1 min at 296 s⁻¹ (Ladouceur & Narbaitz 2022). Then, the feed water was pumped through the membrane module at a constant rate (11.4 L/h). Each experiment was set to end after 120 min of permeation, and it was carried out in the same order: the tank was first emptied, then permeated for 30 min, followed by 5 min of backwashing with ultra-pure water at 1.0 bar. To calculate the transmembrane pressure (TMP), two pressure gauges were connected at the influent and the effluent feed of the UF to record the pressure readings every 2 min. Additionally, a graduated cylinder and a stopwatch were used to keep track of the permeating flow rate. It is important to note that the TMP and flux values were both set to 20.0 °C before any calculations, as shown by Aly (2015). The pH meter and temperature sensor were also tracked and recorded.

Following each experiment, the module was chemically backwashed with 5.0 g/L of citric acid solution and 200 mg/L of sodium hypochlorite. To allow for chemical reactions, the chemical backwash period of each chemical was divided into two stages: first, the module was backwashed with the chemical for 5.0 min, and then the chemical solution was left inside the module for 20 min. To maintain and restore its reliable operation, the UF membrane requires a typical amount of time, as mentioned by Ladouceur & Narbaitz (2022). Following the application of the two chemicals, the module was backwashed for 20.0 min with distilled water to remove any remaining chemicals or foulants.

Every 30 min, samples were taken at the UF's inlet and outlet and placed in clean glass bottles. A turbidity bench meter (Model HACK/2100N, China) was used to measure turbidity, and a pH analyzer (Model Thermo Scientific Orion StarA111, Singapore) was used to measure pH values. The open reflux method (5220 B) was used to measure COD values. Orthophosphate was measured using the ascorbic acid method by a spectrophotometer (Hach DR 2010), bought from Hach Company at wavelength 890 nm. Nitrate was determined using the spectrophotometer according to the cadmium reduction method 4500-NO₃ E. The 1,10-phenanthroline indicator was used to analyze the iron using the standard method (3500-Fe B) (phenanthroline method 8008) and the spectrophotometer was used to measure the results. Finally, bacterial concentrations were determined using the plate count method. All analytical measurements have been conducted according to Standard Methods (2017).

Using IMB SPSS Statistics version 26.0, a statistical analysis was carried out to examine the quantitative relationship between the influent characteristics (COD and turbidity) and the fouling resistances of the UF.

Eight experiments have been carried out to assess the effects of the different coagulant types and dosages, including one to test the secondary effluent (control one), two experiments for each coagulant (alum, ferric chloride, and sodium ferrate) at two different dosages, and the final experiment involving the mixture of FeCl₃ and Na₂FeO₄ at a ratio of 4:1, respectively. First, tests have been done using 15.0 mg/L of alum, 5.0 mg/L of ferric chloride, and 0.5 mg/L of Na₂FeO₄. Second, the tested dosage of each coagulant was lower: 10.0 mg/L of alum, 2.5 mg/L of FeCl₃, and 0.2 mg/L of Na₂FeO₄. Additionally, the combination approach added much smaller amounts of ferric and ferrate (0.5 mg/L of ferric and 0.125 mg/L of ferrate).

Both the UF permeates and the influent feed immediately before entering the UF have been collected and analyzed, and the results are presented in Table 2. The impact of the in-line coagulation pretreatment alone varied depending on the coagulant type and dosage. Most dosages used had minimal effects on turbidity; the ferrate concentration of 0.2 mg/L had the highest removal efficiency of 31.4%. Contrarily, according to Yang et al. (2022), it increased turbidity at 0.5 mg/L, which may be related to the production of iron (hydro) oxide molecules due to the self-decomposition of ferrate. In a manner like He et al. (2023), 0.2 mg/L of ferrate followed by 2.5 mg/L of ferric chloride both contributed to the greatest COD reduction, followed by 2.5 mg/L of FeCl₃ and a higher dosage of alum consistent with Aly et al. (2021).

Regarding the removal of inorganic matter, except for alum dosages, the iron content of the influent typically increased after coagulation. Due to the iron's adsorption into the formed flocs, 5.0 mg/L of FeCl₃ had the best removal efficiency (89.76%), followed by 15.0 and 10.0 mg/L of alum. Like the findings of Li et al. (2019) and Khouni et al. (2020), alum produced the best results for phosphorus reduction, followed by ferrate. The best removal of nitrate was achieved by the two dosages of ferrate, followed by 2.5 mg/L of FeCl₃ and 10.0 mg/L of alum.

The effects of the hybrid process (in-line coagulation followed by UF) on the permeate quality are also summarized in Table 2. The UF caused a turbidity reduction of 85.76% and a COD reduction of about 18.2%. The same outcomes were demonstrated by Muthukumaran et al. (2011), who concluded that the used PES membrane was susceptible to minor fouling. Most of the experiments showed that in-line coagulation increased the percentage of turbidity reduction, except for those involving 15.0 mg/L of alum, which is consistent with Ladouceur & Narbaitz (2022), and 5.0 mg/L of FeCl₃. The highest reduction in turbidity was observed by applying 0.2 mg/L of Na₂FeO₄, followed by FeCl₃/Na₂FeO₄, and then 10 mg/L of alum consistence with Tang et al. (2022) and Yu et al. (2016b). In-line coagulation significantly increased the removal efficiency of COD owing to its capacity to aggregate small organic compounds and particles into larger ones that could settle on the membrane surface at low coagulant doses and neutral pH values using a charge neutralization mechanism (Aly et al. 2021). The highest removal efficiency of COD was observed by FeCl₃/Na₂FeO₄ (73.84%), ferrate at a higher dosage (54.33%), ferrate at a lower dosage (41.30%), and ferric at 2.5 mg/L (39.27%). Similar findings regarding the superior effect of the ferrate, which outperformed FeCl₃ were made by Ding et al. (2022) and Yang et al. (2022), and (2023). This may be attributed to the larger specific area of the formed flocs, which can adsorb more substances and reduce more organics (Ding et al. 2022). Similar results were found by Aly et al. (2021) and Ladouceur & Narbaitz (2022) who found that increasing the dosage of alum and ferric decreased the removal efficiency of COD in contrast to ferrate.

A good removal efficiency of nitrate, iron, and phosphate has been reported when using coagulants with membrane filtration (Zhao et al. 2022). It can be noticed from the results provided in Table 3 that it is obvious that the process used in the current study (in-line coagulation followed by UF) had great efficiency in governing the environmental risks related to nitrate, iron, and phosphate in wastewater reuse.

The effluent quality was considerably improved, making the permeate suitable for irrigation of Grade A food (food eaten without being peeled, like apples and grapes). The turbidity value in all the treatment cases was reduced to less than 5.0 Nephelometric Turbidity unit (NTU), the maximum limit set by the Egyptian regulations, regarding Grade A food (ECP 2015). Except when using UF alone or using 15.0 mg/L of alum, the highest recorded value of COD was 54.73 mg/L, and in most cases, the value was less than or equal to 40.0 mg/L, the maximum accepted value per Egyptian Governmental Law (1982). The application of FeCl₃ and Na₂FeO₄ together produced the best results, followed by 0.5 mg/L of ferrate, which produced COD concentrations of 14.32 and 25.0 mg/L, respectively. According to FAO (2021), applying the UF alone will not be effective for reducing nitrate because it would result in 6.22 mg/L of nitrate, which is higher than the allowed limit (5.0 mg/L). However, the in-line coagulation and UF combined process produced an effective low nitrate concentration. Iron and phosphate concentrations in the raw water were lower than those allowed by the Egypt Decree (2013); 3 and 30 mg/L, respectively. These findings showed that the investigated treatment processes turned the effluent of the EWWTP from being a completely unsuitable option for irrigation to convenient, safe water for agricultural purposes, even in the case of food eaten without being peeled.

A permeability test that was carried out before each experiment showed that the membrane restored almost its entire integrity, which is noteworthy. Additionally, in the permeability test that was completed right before each experiment (J₀), the flux of each experiment was normalized to the flux of pure water. Figure 5 displays the performance of the chosen coagulant dosages. Throughout the experiments, the permeate flux decreased and the TMP values rose in every experiment that was carried out. The control experiment revealed a discernible rise in the TMP as well as a fall in the normalized flux. The normalized flux values significantly decreased and reached 0.788 at the end of the first cycle, while the TMP values significantly increased during the first cycle from 0.16 to 0.198 bar. The same results are observed by Yang et al. (2022) and Ding et al. (2022) when treating secondary effluent using UF and by Cheng et al. (2016) and (2017) when treating surface water. However, minimal fouling has been observed in the current study. This may be attributed to the use of UF membranes with different materials. For instance, polyvinyl-dene fluoride UF studied by Cheng et al. (2016) experienced a 47% reduction in flux (J₀) during the first cycle. Muthukumaran et al. (2011) investigated PES UF membranes (the same one used in the present study) to treat synthetic secondary effluent, and the membrane's flux remained nearly constant throughout the entire run (90 min). The PES membrane was, therefore, thought to have passed more organics and foulants through it. Howe & Clark (2002), while studying different types of surface waters, concluded that the main fouling mechanism of 0.2 μm PES is adsorption and that it does not retain clogging agents, which differs from other polypropylene membranes. In addition, PES is highly hydrophilic and has a higher ability to adsorb biopolymers and proteins on hydrophobic surfaces (Cheryan 1998; ABDEL-KARIM et al. 2022; Salimi et al. 2022).

The reduction percentage of the IRF was greatly improved by the greater FeCl₃ dosage than by the lower, with identical results achieved by Aly et al. (2021). Because of its capacity to reduce pore blockage, particularly at higher dosages where bigger particles have been produced (Brover et al. 2022). Conversely, a low FeCl₃ dosage (2.5 mg/L) gave better results in lowering the reversible one.

In terms of ferrate (Na₂FeO₄), 0.5 mg/L (0.003 mM/L) reduced the IRF by 32.46%, which was slightly less than the findings obtained by 5.0 mg/L of FeCl₃ (0.03 mM/L) (Ding et al. 2022; Yang et al. 2023). The irreversible resistance was conversely increased by the decreased ferrate dosage, which may be related to the oxidation of low molecular weight particles that might easily obstruct membrane pores (Yang et al. 2022). Nevertheless, the small ferrate dosage decreased the RF by 21.6%, demonstrating that it can remove some of the organics and high molecular weight particles.

The mixture of FeCl₃ and Na₂FeO₄ significantly reduced the total and the IRF due to their efficient oxidation and coagulation abilities explained in the transfer process of 1e- and 2e- resulting in the production of Fe(V) and Fe(IV), which had a great capability of oxidation, in addition to the self-decomposition of ferrate-generating OH oxygen species. In addition, iron (hydro) oxide was formed (Chen et al. 2021; Yu et al. 2016b), which was able to adsorb some of the smaller compounds. However, the extra iron hydroxide that was deposited on the membrane's surface caused the RF coefficient to rise. According to He et al. (2023), ferrate and NaClO have a beneficial effect in lowering membrane resistance. It is vital to note that the elimination of IRF is crucial to the overall efficiency of UF because it reduces the demand for chemicals and lengthens the lifetime of the membrane. In conclusion, the two doses of FeCl₃, followed by 0.5 mg/L of Na₂FeO₄ were the second most effective dosages for reducing total and IRF after FeCl₃/Na₂FeO₄. Additionally, Na₂FeO₄ at 0.2 mg/L demonstrated the best management of reversible fouling. According to Aly et al. (2021), who found that ferric salts were more successful as a fouling management coagulant than alum, the applied dosages of alum in the current study were the least effective.

The sustainability of the in-line coagulation, used as a pretreatment for the PES UF and treating the real secondary effluent, was demonstrated. It proved its effectiveness in turning a very large amount of water, EWWTP effluent, into hygienic water of high quality, suitable for irrigation (Grade A), as aimed by the 2030 Sustainable Development Goals, without having any negative impact on public health. Under the investigated conditions and the composition of the investigated secondary effluent, the type and dose of the coagulant had a substantial impact on the process's effectiveness; for example, raising the alum dosage from 10.0 to 15.0 mg/L resulted in an increase in total fouling and IRF that was greater than in the case of no coagulant. Additionally, it was determined that the two ferric-based coagulants (FeCl₃ and Na₂FeO₄) performed significantly better than the aluminum-based coagulant (alum). By lowering both the organic matter (COD) and the inorganic one, the green coagulant treatment at 0.5 mg/L significantly improved the effluent quality. FeCl₃ (2.5 mg/L) performed almost as well in decreasing IRF as Na₂FeO₄ (0.5 mg/L) at a fifth of the molar concentration of Fe, but significantly outperforming it in terms of total fouling control. The combination of FeCl₃ and Na₂FeO₄ produced the best results in terms of both permeate quality enhancement and fouling mitigation. During this efficient combination, the COD concentration and IRF coefficient showed the least increase in TMP and the maximum decrease percentage, respectively. Since IRF is the most important due to its impact on membranes' lifetime, studying the appropriate pretreatment to reduce foulants correlated to it is essential.

Overall, ferrate as in-line pretreatment coagulation before UF is a promising technology to efficiently reduce both organic and inorganics from EWWTP effluent and improve UF membrane fouling in wastewater reclamation. In future studies, this technology may be investigated using different WWTP effluents, and different materials of UF should be investigated.

This paper is based upon work supported by the Science, Technology & Innovation Funding Authority (STDF) under the grant (Young Researcher).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.