Competitive study of homogeneous and heterogeneous Fenton-like flow-through propoxur oxidation in ROC solution

The ever-increasing application of reverse osmosis (RO) in municipal and industrial wastewater treatment has provided an affordable alternative for the sustainable reclamation of water resources. Nonetheless, this process produces a concentrated waste stream (ROC) that needs to be treated and/or disposed of. Municipal wastewater ROC contains rejected substances at 4–7 times higher concentrations than the feed water (Yaqub et al. 2022) and has a reduced volume of 15–20% (Xiang et al. 2019). The composition of the ROC depends on the properties of the raw wastewater and the specific wastewater treatment process applied. It consists mainly of nutrients, dissolved inorganic salts, and organic substances such as micro-pollutants (MPs). The range of the physicochemical parameters of ROC can be found in Deng (2020) and Valdés et al. (2021).

MPs consist of natural and anthropogenic substances, including pharmaceuticals, personal care products, steroid hormones, industrial chemicals, pesticides, and surfactants. The concentrations of micro-pollutant in ROC range from a few ng/L to hundreds of mg/L. The occurrence of 55 emerging MPs in nanofiltration and RO municipal wastewater concentrates was reviewed by Deng (2020).

Several studies have comprehensively reviewed the physicochemical and biological techniques explored for the removal of MPs, including adsorption, membrane processes, advanced oxidation processes (AOPs), and their integration (Rathi et al. 2021; Sivaranjanee & Kumar 2021; Shahid et al. 2021; Kumar et al. 2022; Morin-Crini et al. 2022).

Management of ROC often poses challenges, as removal of organic pollutants from ROC by conventional methods is difficult due to high salinity and the organic pollutants recalcitrant nature. High salinity inhibits microbial activity mainly by causing water loss from the cell membranes and reducing enzyme activities. This can result in plasmolysis or cell death (Shi et al. 2015). Additionally, high salinity can impact oxidation processes by scavenging bulk and/or surface oxidants, decreasing dissolution, inactivating active sites, and altering the nature of the organics (Yuan et al. 2022).

AOPs are utilized for the partial or complete mineralization of contaminants through in-situ generation of highly reactive and low selectivity radicals, including hydroxyl, hydroperoxyl, and superoxide. Sulfate and chloride radicals are also included in a wider definition of AOPs. AOPs are based on ozone, hydrogen peroxide electrochemical reactions, photolysis, sonolysis, and their combinations. Several recent comprehensive reviews have been published on AOPs in general and specifically for treating emerging trace organic contaminants (Giwa et al. 2021; Krishnan et al. 2021; Tufail et al. 2021; Priyadarshini et al. 2022; Saravanan et al. 2022; Gonzaga et al. 2023; Wang et al. 2023). Moreover, Xiang et al. (2019), Deng (2020), and Arola et al. (2019) have reviewed various AOPs that are applicable to ROC treatment for municipal wastewater reclamation.

The Fenton (H₂O₂/Fe²⁺) and Fenton-like (H₂O₂/Fe³⁺) reaction can be either homogeneous or heterogeneous, with the latter involving the use of solid catalyst such as iron oxyhydroxide, iron oxides (ferrihydrite, goethite, hematite, and magnetite), and zero-valent iron. Iron-based catalysts have several advantages, including their abundance, low cost, high catalytic activity, low toxicity, efficient recovery, chemically stability over a wide pH environmental friendliness, and sustainability (Hussain et al. 2021; Thomas et al. 2021). In both Fenton and Fenton-like systems, the primary reactions generate hydroxyl radicals (HO•), hydroperoxy (HO₂•), and peroxide radicals O₂•⁻, as shown in Equations (1)–(3). The rate constant of reaction 1 (40–80 1/M s) is much higher than that of reaction 2 (0.1–1 × 10⁻² 1/M s) (Thomas et al. 2021), indicating the importance of the Fe³⁺/Fe²⁺ redox cycle for highly efficient oxidation.

The Fenton process is a preferred AOP due to its versatile applications, capacity to withstand interference, simple operation, rapid degradation rate, and improved biodegradability of organic contaminants. However, it exhibits certain limitations such as a restricted operational pH range of 2–4 and the need for neutralization. The formation of iron sludge and consequent discharge of iron into the environment further contribute to its drawbacks (Zhang et al. 2019). The heterogeneous Fenton process offers several advantages over the homogeneous Fenton process. These include reduced iron sludge production, a wider working pH range, catalyst reusability, reduced iron leaching, long-term catalyst stability, and effective Fe³⁺/Fe²⁺ interconversion (Xavier et al. 2013; Sreeja & Sosamony 2016; Zhang et al. 2019). However, it is limited by the diffusion rate of the target compounds onto the surface of the catalyst and/or its boundary layer, resulting in a slower oxidation rate than the homogeneous process at the same concentrations (Litter & Slodowicz 2017). The catalyst's synthetic conditions and costs, as well as the reactor, pose challangers (Zhang et al. 2019).

Propoxur (PR), 2-isopropoxy phenyl N-methyl carbamate, commercially known as Baygon® is a broad-spectrum insecticide that can persist in the environment and pose a risk to human health and the ecosystem. It has been detected in surface water, groundwater, and soil, indicating widespread environmental contamination. Exposure to PR has been linked to various adverse health effects, such as nervous system stimulation, enzyme activation, stunted growth, and cancer (Vandana et al. 2001). AOPs can be effective for the removal of PR from contaminated water hence contribute to reducing its potential adverse effects on human health and the environment. It has been reported that PR was oxidized via direct photolysis (UV) (Benitez et al. 1994; Sanjuán et al. 2000), photo-catalysis (UV/TiO₂) (Sanjuán et al. 2000), UV/TiO₂/GAC (Lu 1999), O₃/UV (Benitez et al. 1994), and electrochemically (Guimarães Selva & Longo Cesar Paixão 2016).

In this study, the removal of PR from synthetic ROC solution by homogeneous and heterogeneous Fenton-like oxidation in a ceramic membrane reactor was compared. PR was chosen as a model MP compound due to its high water solubility and widespread use as a broad-spectrum insecticide.

The chemicals were of analytical grade and were used as-received without further purification. PR was supplied by Supelco (China), NaCl, KCl, NaHCO₃ (Frutarom Industries Ltd, Israel), Na₂SO₄ (Bio-Lab Ltd, Israel), MgCl₂·6H₂O (Merck, Germany), CaCl₂·2H₂O (Spectrum, USA), Na₂HPO₄ (Riedel-de Haën, Germany), and FeCl₃·6H₂O (Alfa Aesar, USA).

Synthetic ROC solution was prepared by dissolving NaCl, NaHCO₃, Na₂SO₄, CaCl₂·2H₂O, KCl, Na₂HPO₄·2H₂O, and MgCl₂·6H₂O in deionized water (DI). The composition of the solution is listed in Table 1. PR oxidation experiments were conducted in both DI and ROC solutions. It is important to note that the addition of both the homogeneous and heterogeneous catalysts introduces counter ions to the aqueous solution, namely sodium and chloride.

A freshly prepared amorphous ferric chloride-based catalyst was synthesized using the precipitation method. The catalyst was prepared by diluting a stock solution of FeCl₃·6H₂O (1.3 M) in DI water to the desired concentration, followed by the addition of a predetermined amount of NaOH solution (1.0 M) to achieve the desired pH.

The particle sizes were measured using a laser diffraction particle size analyzer (Mastersizer 2000, Malvern, UK). The zeta potential was measured using a ZetaSizer Nano-ZS analyzer (Malvern, UK) in a disposable folded capillary cell (DTS1060). The morphology of the synthesized catalyst was studied using Philips CM120 Transmission Electron Microscopy (TEM) and an Olympus BH2 Light Microscope connected to an Optronics CCD camera. Raman spectra were obtained with a Horiba Jobin Yvon LABRAM HR micro-Raman spectrometer equipped with a Nd-YAG laser (100 mW, 532.2 nm) and diffraction gratings of 1,800 grooves 1/mm. The Raman shifts were detected using a Peltier-cooled, slow-scan, CCD matrix detector, and ranged from 180 to 450 1/cm.

A Masterflex peristaltic pump (model 77800-50, Cole-Parmer, USA) was used to feed a solution containing PR and hydrogen peroxide into the reactor. The flow rates were set at 0.13–0.59 L/h, resulting in residence times (RTs) of 20–90 min. A second pump with an identical flow rate was used to pumped the product water (i.e., permeate) at fluxes of 13–59 L/m² h and a pressure of <25 mbar. The submerged membrane operated in a dead-end mode.

At the beginning of the experiment, the heterogeneous catalyst was added to the reactor at concentrations ranging from 252 to 1,495 mg/L. The experiments were carried out either in ROC or DI water solutions with a volume of 192 mL, a PR concentration of approximately 1 mg/L, and hydrogen peroxide concentrations ranging from 140 to 1,096 mg/L at molar ratios of 0.7–1.1 iron to hydrogen peroxide. A third pump was used to continuously dose the homogeneous catalyst. These experiments were carried out in ROC solution, at pH 3.0, 150 mg/L H₂O₂, and 250 mg/L Fe³⁺.

Experiments were conducted to test the effects of sulfate, chloride, and bicarbonate on the PR oxidation in ROC solution, both in the presence and absence of these ions. The experiments were performed at pH 4.0, 150 mg/L H₂O₂, and 250 mg/L catalyst. Notably, these experiments were conducted using a full ROC solution composition, with the omission of one of the aforementioned anions. This is in contrast to the batch to the batch experiments described in the next section, which were carried out in a single salt solution with PR and the tested ion dissolved in DI water.

pH values of 3.0 and 4.0 were maintained constant throughout the experiments using the homogeneous and heterogeneous catalysts, respectively. Experiments were carried out for at least five RTs after the system had reached a steady state. The reported values of C/C₀ represent the normalized concentration at steady-state conditions.

The efficiency of batch PR oxidation in ROC solution was investigated using both homogeneous and heterogeneous catalysts. The homogeneous catalyst was tested at pH values of 2.5, 3.0, and 3.5, while the heterogeneous catalyst was tested at pH values of 4.0, 5.0, 6.0, and 7.8. The lowest pH tested for the heterogeneous catalyst was 4, as this is the threshold for ferric chloride coagulation (Wei et al. 2017). The pH was maintained constant throughout the experiments. The experiments were conducted in 100 mL glass beakers containing 50 mL of ROC solution with 1 mg/L of PR for up to 4 h. The molar ratio of Fe to H₂O₂ was kept constant at 1, with 84 mg/L of the homogeneous catalyst (50 mg/L H₂O₂) and 250 mg/L of the heterogeneous catalyst (150 mg/L H₂O₂). The concentrations of the reactants were adjusted to achieve comparable PR removal rates with both catalysts. The solution was stirred using a magnetic stir bar at room temperature (22 ± 2 °C) and sampled periodically to measure the concentrations of PR, hydrogen peroxide, and iron. Prior to analysis, samples were filtered using a 0.45 μm filter to separate the heterogeneous catalysts from the solution.

Batch Fenton-like oxidation experiments designed to study the effect of the catalyst concentration were carried out in a glass beaker containing 25 mL ROC solution with 1 mg/L PR for up to 3.5 h. These experiments were conducted at pH 3.0 and 50 mg/L H₂O₂ for the homogeneous catalyst, and at pH 4.0 and 150 mg/L H₂O₂ for the heterogeneous catalyst. The different reactant concentrations were used so as to obtain similar PR removal efficiencies using both of the catalysts. The catalyst dosages were varied from 8 to 93 mg/L for the homogeneous catalyst and 71 to 256 mg/L for the heterogeneous catalyst, while the oxidant dosage was kept constant.

To examine the effect of accompanying ions on the removal of PR, batch experiments were conducted in the presence of each of the ROC components dissolved in DI water at the concentrations listed in Table 1. The experiments were conducted in glass beakers containing 50 mL of solution for up to 25 min. The homogeneous catalyst was used at pH 3.0 with 25 mg/L H₂O₂, while the heterogeneous catalyst was used at pH 4.0 with 150 mg/L H₂O₂. The molar ratio of Fe to H₂O₂ was kept constant at 1.

The counter ion for the cations (calcium, potassium, and magnesium) was chloride, while the counter ion for the anions (chloride, bicarbonate, phosphate, and sulfate) was sodium. Initially, the effect of the counter ions was tested by dissolving NaCl in DI water at a concentration of 1,628 mg/L. The results served as a benchmark for determining the effect of the other tested ions.

The concentration of PR was analyzed using an Agilent 1260 Infinity LC (Agilent, US) equipped with a Zorbax Eclipse XDB-C18 column. The eluent consisted of a 70:30 (v/v) mixture of HPLC-grade acetonitrile and HPLC-grade water, at a flow rate of 0.4 mL/min. PR concentration was measured at a wavelength of 271 nm, using an injection volume of 10 μL, and a column oven temperature of 30 °C.

Hydrogen peroxide concentration was determined by the Ghormley method (Klassen et al. 1994). Iron concentration was determined using Hach method 8008 FerroVer^® Iron Reagent Powder Pillows (DR2800 spectrophotometer, Hach, Germany). All experiments carried out in triplicate revealed a standard deviation average of up to 8% in PR analysis and 5% in hydrogen peroxide and iron analysis.

Smaller aggregates were obtained in the ROC solution as compared to DI water, as seen in Figure 2(c). This is because the electrolyte ions hinder condensation and polymerization processes, resulting in smaller particle sizes (Wei & Semiat 2017). Measurement of the zeta potential (Figure 2(d)) revealed a higher point of zero charge (8.0) in DI water than in ROC solution (6.1). The electrolytes decrease the zeta potential, indicating reduced electrostatic repulsion, initially promoting the aggregation of iron nanoparticles before the ion hindering effect becomes predominant (Wei & Semiat 2017). Also seen in Figure 2(c) is that larger aggregates were obtained in both DI and ROC solutions as the pH was increased. It has been reported that the threshold of ferric chloride coagulation is around pH 4. At pH < 4, repletion between positively charged monomers limits the bridging process. As the pH increases, the generation of polycationic structures is promoted, resulting in precipitated particles in the aqueous solution (Wei et al. 2017).

Control experiments were conducted in DI water with PR, PR + H₂O₂ (130 mg/L), and PR + Fh (250 mg/L) at a residence time of 44 min and pH of 4.0. In these experiments, the feed PR concentration was 1 mg/L. The results revealed negligible changes in the PR concentrations, indicating that the ceramic membrane and/or the Fh did not adsorb the PR and the hydrogen peroxide did not oxidize it without a catalyst. An Fh rejection of >99.6% by the ceramic membrane was obtained at pH ≥ 4.0.

The pH can have an impact on heterogeneous catalysts by altering the solubility of iron on the catalyst surface and/or activity of its active sites. The fact that the amorphous Fh catalysis was not efficient at elevated pH levels suggests that the pH did not affect its active sites but rather the solubility of iron. An increase in pH slows the iron leaching from the heterogeneous catalyst, leading to inactivation in the bulk solution via hydrolysis and ferric hydroxide sludge precipitation (Wang et al. 2016). Homogeneous Fenton reactions are feasible at pH < 4 because the interconversion of Fe²⁺ and Fe³⁺ maximizes the process efficiency (Hussain et al. 2021). Acidity aids in protonating hydrogen peroxide, which enhances its reactivity and increases its propensity to produce hydroxyl radicals rather than decomposing into water (Jung et al. 2009).

Both catalysts exhibited similar removal efficiencies at the same molar ratio (Figure 7). However, to achieve equivalent removal percentages as Fe³⁺, the heterogeneous oxidation required three times higher concentrations of H₂O₂ and Fh, as well as prolonged reaction time. As mentioned earlier, heterogeneous catalysis can be limited by factors such as mass transfer, catalyst surface area, number of active sites, and the reduction rate of Fe³⁺.

During the homogeneous Fenton-like oxidation process, Fe³⁺ was continuously dosed into the reactor. Initially, about five times as much Fh was added as compared to Fe³⁺ (1,500 and 250 mg/L, respectively). Within about six RTs, the total iron consumed (253 g/m³ product water) was equalized for both catalysts, after which the Fe³⁺ concentration exceeded that of the Fh. This suggests a preference for heterogeneous catalysis during prolonged operations, as less Fh would be required. Additionally, using Fh as a catalyst has the added advantage that the product water does not contain any iron.

As seen in Figure 9, a linear relationship was observed in DI water solutions, between residence time and PR removal, with higher removal rates as residence time increased. The highest PR removal of 89.7% was obtained at a RT of 87.5 min. In contrast, an asymptotic relationship between residence time and PR removal was observed in the ROC solution. Only a small change in the PR removal performance was found following the RT of 64 min. Longer RTs typically result in higher removal rates due to increased reaction time. Insufficient contact between the target compounds and the heterogeneous catalyst, as well as the limiting mass transfer rates, may lead to a lower oxidation efficiency (An et al. 2022). Nonetheless, PR oxidation catalyzed by Fe³⁺ (150 and 250 mg/L H₂O₂ and Fe³⁺, respectively) was not affected by the RTs of 20–60 min (data not shown). These results suggest that, for the tested concentrations of the hydrogen peroxide and Fe³⁺, RT of 20 min was sufficient to effectively remove the PR to a maximum level of 68 ± 3.4%, and that the reaction rate was shorter than 20 min.

To better understand the effect of the ROC solution on the PR Fenton-like oxidation (both homogenous and heterogeneous), batch experiments were performed in DI water solution with the addition of each of the ROC solution components at the concentrations listed in Table 1. It should be noted that the concentration of the ions may also contribute to their effect on the PR oxidation. The results were compared to those of PR oxidation in NaCl solution (1.6 g/L) in order to eliminate the effect of the counter ions, chloride and sodium, that accompany the tested ions. These experiments were conducted at a molar ratio of 1 Fe to H₂O₂ with six times higher concentrations of H₂O₂ catalyzed by Fh. The higher concentrations were chosen to achieve PR removal within the time frame of Fe³⁺.

Inorganic ions may inhibit AOPs through two main mechanisms: hydroxyl radical scavenging and iron complexation. Table 2 lists the reactions of the inorganic anions present in the ROC solution with hydroxyl radicals. These inorganic radicals exhibit lower reactivity, higher selectivity, and longer half-lives than hydroxyl radicals (De Laat et al. 2004; Patra et al. 2020). Inorganic anions may also interfere with the formation of O₂·^–/HO₂· radicals hindering the regeneration of Fe²⁺ from Fe³⁺ reduction (Equation (2)), resulting in lower hydroxyl radicals generation (Equations (1); De Laat et al. 2004).

Table 2

Primary reactions of hydroxyl radical with inorganic anions

Reaction .	Rate constant (1/M s) .	Ref. .
	4.3 × 109	Buxton et al. (1998)
	1.4 × 107	de Oliveira et al. (2015)
		Kim et al. (2019)
	8.5 × 106	Buxton et al. (1998)
	3.9 × 108	Buxton et al. (1998)
	2.2 × 106	Martire & Gonzalez (2001)
	8.0 × 105	Martire & Gonzalez (2001)
	2.6 × 106	Martire & Gonzalez (2001)

Reaction .	Rate constant (1/M s) .	Ref. .
	4.3 × 109	Buxton et al. (1998)
	1.4 × 107	de Oliveira et al. (2015)
		Kim et al. (2019)
	8.5 × 106	Buxton et al. (1998)
	3.9 × 108	Buxton et al. (1998)
	2.2 × 106	Martire & Gonzalez (2001)
	8.0 × 105	Martire & Gonzalez (2001)
	2.6 × 106	Martire & Gonzalez (2001)

View Large

The complexation of anions with aqueous and/or surface-bound Fe²⁺ and Fe³⁺ affects the distribution and reactivity of the iron species (De Laat et al. 2004) for both homogeneous and heterogeneous catalysts. For heterogeneous catalysis, surface complexations may block active sites on the Fh surface, decreasing the oxidation efficiency (Sheng et al. 2020). The four anions' surface complexation on Fh is considered inner-sphere for phosphate and carbonate, outer-sphere for chloride, and both inner- and outer-sphere for sulfate, depending on the pH, ionic strength, surface coverage, and hydration (Kumar et al. 2014). The molecular structure of sulfate inner-sphere complexes was determined as bidentate − binuclear (Gu et al. 2016) while that of carbonate and phosphate double-corner bidentate complexes (Kumar et al. 2014). In the solution, sulfate, chloride, and phosphate anions form complexes such as FeSO₄⁺, FeSO₄, FeCl²⁺, FeCl⁺ (de Oliveira et al. 2015), FeH₂PO₄²⁺, FeHPO₄⁺, and Fe(H₂PO₄)₂⁺ (Al-Sogair et al. 2002).

The removal of PR from synthetic ROC solution was studied using homogeneous (Fe³⁺) and heterogeneous (Fh) Fenton-like oxidation in a submerged ceramic membrane reactor was studied. The freshly synthesized heterogeneous Fh catalyst had an amorphous, layered, porous structure consisting of nanoparticles (5–16 nm) in aggregates with a pH_PZC of 6.1. The size of the aggregates ranged from 33 to 49 μm, depending on the pH of the solution.

Batch experiments revealed that the most favorable conditions for PR oxidation were achieved at a molar ratio of 0.75 iron to hydrogen peroxide and pH values of 3.0 and 4.0 for homogeneous and heterogeneous catalysis, respectively. The flow-through experiments showed that Fe³⁺ exhibited higher catalytic activity than Fh, with 72 and 13% of PR removed and 39 and 2% of hydrogen peroxide consumed, respectively. By increasing the concentrations of hydrogen peroxide and Fh, the PR removal could be increased to comparable levels as those obtained by Fe³⁺ catalysis. Although Fh required about five times higher reactant concentrations, the total iron consumed was equal to that of the Fe³⁺ within six RTs. After which, the Fe³⁺ concentration exceeded that of the Fh, as ≥99.6% rejection of the Fh was obtained by the ceramic membrane. Hence, for prolonged operations, less Fh would be required. A maximal PR removal of 87% was obtained at a residence time of 88 min.

The ionic composition of the ROC solution was found to inhibit the PR oxidation, requiring higher reactant concentrations to achieve the same removal efficiencies as in DI water solution. PR oxidation in a single salt solution of the ROC composition revealed a negligible effect of the cations, while, the anions hindered PR removal by scavenging hydroxyl radicals and forming iron complexations. Under steady-state conditions, excluding bicarbonate, sulfate, or chloride from the ROC solution did not significantly affect PR oxidation. However, in the absence of sulfate and chloride, the number of RTs required to reach steady state increased, suggesting that outer-sphere complexation is the predominant mechanism of the anion adsorption onto the Fh in the ROC solution.

The authors would like to express their gratitude for the funding provided by the German Ministry of Education and Research (BMBF) and Israeli Ministry of Science & Technology (MOST) through the German-Israeli water technology cooperation (project WT1902).

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

The authors declare there is no conflict.