Impact of Inorganic Anions on the Photodegradation of Herbicide Residues in Water by UV/Persulphate-Based Advanced Oxidation

1. Introducción

Pesticides constitutes an issue of considerable political interest in the fields of environment, agriculture and human health under European Legislation. The Directive on the Sustainable Use of Pesticides (2009/128/EC) [1] aims to reduce the risks and impacts of pesticide use on human health and the environment, and to promote the use of integrated pest management and alternatives such as non-chemical approaches. In this context, water quality legislation is key to protecting public health and the environment in the EU. The Green Deal [2] and its associated strategies aim to reduce pesticide use and risks by 50% by 2030, with a focus on protecting ecosystems and enhancing biodiversity. For this purpose, the Water Framework Directive (WFD) 2000/60/EC [3] and its “daughter directives” such as the Groundwater Directive (GWD) 2006/118/EC [4] and the Environmental Quality Standard Directive (EQSD) 2013/39/EU [5] help to protect European waters from pesticide and chemical pollution. In line with advances in science and technology over the past decades, on 26 October 2022, the EU Commission published its proposal to amend the WFD and EQSD [6].

These directives focus largely on so-called priority substances. The Commission now proposes a simpler and more general definition of priority substances as chemicals that "present a significant risk to or via the aquatic environment in a high proportion of Member States", leaving open how many water bodies and Member States would be considered a "high proportion". Priority hazardous substances are “priority substances that are also identified as hazardous on the basis that they are recognised in scientific reports, relevant Union legislation or relevant international agreements as being toxic, persistent and liable to bio-accumulate, or as giving rise to an equivalent level of concern where this concern is relevant to the aquatic environment". The first list of 33 priority substances in the field of water policy was published in 2001 [7]. Subsequently, Directive 2008/105/EC [8] established Environmental Quality Standards (EQS) for these priority substances and certain other pollutants, which were subsequently updated by the EQSD Directive, covering 46 substances and groups of substances. Article 8 of the EQSD establishes a "watch list" of substances for which EU-wide monitoring data will be collected to provide adequate data for possible future inclusion in the priority list. The watch list was updated by the Commission in a series of implementing decisions, first in 2015 [9], then in 2018 [10], 2020 [11] and finally in 2022 [12]. These monitoring activities have led to the identification of several new priority substances, but not all of the substances originally included in a watch list are of EU-wide concern. Indeed, for the identification of new priority substances, the watch list mechanism seems to act as a filter. The updating of the list of priority substances and compounds on the watch list by delegated acts could speed up the process and make it less cumbersome. This argument, together with the "one substance, one assessment" argument, is used to justify transferring future work on priority substances and the setting of EQS values to the European Chemicals Agency (ECHA). In addition, the new Drinking Water Directive (DWD) 2020/2184 [13] sets quality objectives for pesticides at the tap and includes a “risk-based approach”, provisions for the protection of raw water and the consideration of pesticide metabolites.

On the other hand, since the adoption of the EU Urban Wastewater Treatment Directive (UWWTD) 91/271/EEC [14] in 1991, the quality of Europe’s rivers, lakes and seas has improved dramatically. EU countries have used EU funding to build collection systems and wastewater treatment plants. However, there is still pollution that needs to be tackled and is not covered by current legislation. According to the European Environment Agency (EEA), a report based on data from Member States shows that only 38% of EU surface water bodies are in good chemical status, 46% are not in good chemical status and 16% are in unknown chemical status [15]. Another recent EEA technical report provides an overview of the occurrence of pesticides and their main metabolites (breakdown or reaction products) in surface water (lakes and rivers, 180 pesticides) and groundwater (159 pesticides) focusing on the European level (up to 39 European countries) from 2007 to 2017 [16]. This report shows exceedance rates of 5 to 15 % for herbicides and 3 to 8 % for insecticides in surface water, while in groundwater the percentages were around 7 % for herbicides and less than 1 % for insecticides. Fungicides appeared to be less frequently exceeded for both surface water and groundwater. In order to achieve maximum crop yields, herbicides are widely used. Agricultural activities, cleaning of herbicide containers and equipment, and agricultural and herbicide manufacturing effluents are the main sources of herbicide water pollution. Two herbicides, isoproturon and terbuthylazine, are among the most frequently reported pesticides in surface and groundwater in Europe. More recently, another study proposes a protocol for a systematic review and meta-analysis to know the prevalence of pesticides in European water resources (watersheds, aquifers, rivers, sea and springs), wastewater (influent and effluent) and drinking water [17].

To address water pollution, the Commission has proposed an update of the Directive [18] to enable a significant reduction in the discharge of pollutants into the EU’s lakes, rivers and seas. In this sense, for some of the most persistent pollutants, such as recalcitrant pesticides, the polluted water (urban, agricultural and industrial) treated by conventional wastewater treatment plants (WWTPs) is in some cases insufficient to achieve the legally required level of purity [19]. This issue is of particular concern in areas where low rainfall does not provide sufficient water resources to meet the needs of agriculture, requiring increased reuse of wastewater treatment plants. Water scarcity and the uneven geographical distribution of rainfall are concerns in arid and semi-arid areas where water management strategies advocate the reuse of treated wastewater in agriculture due to climate change. The EU has also addressed this issue by revising the minimum requirements for water re-use in the context of integrated management under Regulation 2020/741/EU on minimum requirements for water re-use [20]. The aim of this Regulation is to ensure the safety of recycled water for agricultural irrigation, promote circularity, support resilience to climate change and contribute to the objectives of the WFD by addressing water scarcity and related pressures on water resources. Growing public concern about the presence of pesticides in both wastewater and drinking water is leading to the development of new regulations that will undoubtedly have an impact on the design and operation of wastewater treatment plants in the coming years.

As stated in the EQSD, the sources of pollution must be identified, and the emissions of pollutants must be treated at source in the most economic, environmental and sustainable way. Concerns have increased, as has the need to identify them and apply new and effective techniques for their reduction and elimination. Conventional water treatment processes and technologies reduce pollution significantly, but not as much as current regulations require. The refractory nature of some pesticides, which can be toxic to the micro-organisms used, often renders biological treatments ineffective [21]. In addition, approaches based on adsorption, flocculation and ion exchange are not fully effective because they do not destroy the contaminant, but rather transfer it to a solid phase that requires subsequent treatment. Membrane technology has been used extensively in recent years to remove micropollutants from wastewater, but during the long-term process, the pollutants gradually accumulate in the adsorption materials until they reach saturation, after which they become inactive and consequently the pollutant removal and membrane filtration efficiency decreases [22], so it is usually combined with other techniques such as ozonation, activated carbon or photooxidation [23].

Advanced Oxidation Processes (AOPs) have been the most studied in recent years for the removal of pesticides and other micropollutants from wastewater [24-26]. Pesticides are oxidised by highly reactive oxygen species (ROS), mainly hydroxyl radicals (HO^•/OH^-, E⁰ = 1.9-2.7 V vs. NHE) and others such as superoxide anion (O₂^•-) and hydridodioxygen (HO₂^•). The main advantage of these technologies is that they achieve the removal, or at least the reduction, of pesticide levels by mineralisation rather than by transfer, as happens in conventional processes [27]. These technologies are particularly interesting in areas characterised by intensive agriculture and specific climatic patterns, where annual solar radiation is very high and water is scarce. More recently, AOPs based on sulphate radical anion (SO₄^•−) have received increasing attention as an alternative to conventional HO^• based AOPs [28-30]. SO₄^•− has similar redox potential (SO₄^•−/SO₄⁼, E⁰ = 2.6-3.1 V vs. NHE) compared to HO^•, higher selectivity and, in some cases, longer half-lives (t_1/2 = 30-40 μs) than the HO^• (t_1/2 ≤ 1μs), allowing more contact and transfer mass between radical and organic compounds [31]. SO₄^•− are typically generated from peroxydisulphate (S₂O₈⁼, PDS), henceforth persulphate (PS) or peroxymonosulphate (HSO₅^-, PMS) that can be activated by energy (photochemical, sonochemical and thermal), carbonaceous materials, electrochemical activation, alkaline conditions, transition metal ions and various other oxidants such as ozone, hydrogen peroxide and calcium peroxide [32,33]. PMS has a smaller bond length (1.460 Å) than PS (1.497 Å), which explains its higher dissociation energy of the O-O bond (377 kJ mol^-1) compared to PS (92 kJ mol^-1) [28]. In most cases, UV/PS showed a better ability to oxidise organic pollutants than UV/PMS. This can be attributed to the quantum yield of radical formation, as the quantum yield of PS₂₅₄ is about 1.8 M Einstein^-1, which is much higher than that of PMS₂₅₄ (0.5 M Einstein^-1) [28]. As a result, more energy is required for PMS to generate a radical during the homolytic cleavage of the peroxide bond [34]. In the thermal/photochemical decomposition pathway of PS, two mols of SO₄^•− are formed per mole of S₂O₈⁼ as a result of the cleavage of the peroxide bond (Eq. 1). SO₄^•− reacts with H₂O at all pHs to form HO^• (Eq. 2) which is the primary reactive species under basic conditions (Eq. 3). At acidic (pH < 7) and basic (pH > 9) pHs, SO₄^•− and HO^• are the dominant reactive species, respectively, while both radicals participate similarly in reactions at circumneutral pHs [28,35]. Finally, SO₄^•- promotes the mineralisation of pesticides to CO₂ and H₂O (Eq. 4) according to the following reactions:

$${S_{2}O}_8^{2-}+ h \left(UV\right){2SO}_4^{•-}\hspace{1em}\hspace{1em}k=5.8\times {10}^{•5} {\mathrm{M}}^{-1}{s}^{\mathrm{-}1}$$

(1)

$${SO}_4^{•-}+H_{2}O {HO}^{•}+{SO}_4^{2-}+H^{+}\hspace{1em}\hspace{1em}k=200 {\mathrm{M}}^{-1}{s}^{\mathrm{-}1}$$

(2)

$${SO}_4^{•-}+{HO}^{-} {HO}^{•}+{SO}_4^{2-}\hspace{1em}\hspace{1em}k=6.5\times {10}^7 {\mathrm{M}}^{-1}{s}^{\mathrm{-}1}$$

(3)

$${SO}_4^{•-}+Pesticides\to Intermediates\to {SO}_4^{2-}{+CO}_2+H_{2}O$$

(4)

However, both (SO₄^•− and HO^•) undergo rapid and unwanted reactions (Eq. 5-7) where they are consumed, which may be limitations to their effectiveness [36]:

$$2{SO}_4^{•-} {S_{2}O}_8^{2-}\hspace{1em}\hspace{1em}k=8.1\times {10}^8 {\mathrm{M}}^{-1}{s}^{\mathrm{-}1}$$

(5)

$${SO}_4^{•-}+{HO}^{•} {HSO}_5^{-}\hspace{1em}\hspace{1em}k=1.0\times {10}^{10} {\mathrm{M}}^{-1}{s}^{\mathrm{-}1}$$

(6)

$${SO}_4^{•-}+{S_{2}O}_8^{2-} {SO}_4^{2-}+{S_{2}O}_8^{•-}\hspace{1em}\hspace{1em}k=6.1\times {10}^{-5} {\mathrm{M}}^9{s}^{\mathrm{-}1}$$

(7)

As general rule, SO₄^•− is more susceptible to electron transfer reactions than HO^•. In contrast to SO₄^•−, HO^• is more prone to hydrogen abstraction or addition [31,37]. Thus, SO₄^•−-based oxidation is presented as an alternative oxidative treatment to AOPs based on HO^• generation. SO₄^•− has some unique properties, such as being a very strong electron acceptor, which allows the degradation of persistent pesticides that are refractory to the HO^•. In addition, S₂O₈⁼, an environmentally friendly nontoxic oxidizer, avoids the problems of transport limitations because it is relatively stable and can be produced in high quantities.

Overall, the elimination of pesticides and other micropollutants by HO^• and SO4^•- based-AOPs is largely influenced by the quality of the water matrix (dissolved constituents), which can have neutral, inhibitory or promoting, depending on the process and the mechanism by which these water constituents react [28,38]. In addition to organic species, inorganic species such as Cl^-, HCO₃^-/CO₃^2-, NO₃^-/NO₂^-, PO₄H^3- and/or SO₄^2-, among others, can also acts as inhibitors either by scavenging and generating radicals such as ClOH^•−, Cl^•, Cl₂^•−, CO₃^•−, HCO₃^•−, Br^•, Br₂^•−, NO^•, and/or H₂PO₄^•, which have a lower E⁰ than HO^• and SO₄^•- or by promoting the formation of ROS, as in the case of NO₃⁻, which is capable of generating HO^• and NO₂^• radical species that promote the photodegradation of pesticides, especially those for which indirect photolysis is the main reactive pathway. For all the above mentioned, the aim of this work has been to evaluate the impact of inorganic ions on the removal of two herbicides (isoproturon and terbuthylazine) commonly found in wastewater using UV/persulphate-based advanced oxidation under laboratory conditions.

2. Materials and Methods

2.1. Herbicides, solvents and Reagents

Analytical standards of the herbicides, isoproturon (ISP), terbuthylazine (TBZ) and their main metabolites were purchased from Dr. Ehrenstorfer GmBh (Augsburg, Germany), all with purity greater than 95%. The main physical-chemical characteristics of the herbicides studied are shown in Table 1. Titanium dioxide (70 anatase/30 rutile, 99.5%, BET 50 m² g^-1, size <21 nm) Aeroxide™ P25 was supplied by Nippon Aerosil Co Ltd. (Osaka, Japan). Solvents (H₂O, CH₃CN, CH₃OH, CH₂O₂), all HPLC-grade and reagents (Na₂S₂O₈, NaCl, Na₂SO₄ and NaHCO₃) with purity > 98.5% were provided by Scharlab (Barcelona, Spain).

Table 1. Main physico-chemical properties of the herbicides studied.

Herbicide1	Structure	Formula	MW2	SH2O3	log Kow4	H5
IsoproturonPU		C₁₂H₁₈N₂O	206.3	70	2.5	1.5×10-5
TerbuthylazineTZ		C₉H₁₆ClN₅	229.7	7	3.4	2.3×10-3

¹ PU: Phenulurea; TZ: s-Triazine; ² Molecular weight; ³ Water solubility (mg L^-1); ⁴ Partition coefficient n-octanol/water; ⁵ Henry Law Constant (Pa m³ mol^-¹).

Table 1. Main physico-chemical properties of the herbicides studied.

Herbicide1	Structure	Formula	MW2	SH2O3	log Kow4	H5
IsoproturonPU		C₁₂H₁₈N₂O	206.3	70	2.5	1.5×10-5
TerbuthylazineTZ		C₉H₁₆ClN₅	229.7	7	3.4	2.3×10-3

¹ PU: Phenulurea; TZ: s-Triazine; ² Molecular weight; ³ Water solubility (mg L^-1); ⁴ Partition coefficient n-octanol/water; ⁵ Henry Law Constant (Pa m³ mol^-¹).

2.2. Experimental Setup

The degradation of the herbicides was carried out by the UV/S₂O₈⁼ or UV/TiO₂ process in a 2 L (10 cm Ø x 25 cm long) Pyrex batch cylindrical photoreactor (SBS, Barcelona, Spain) equipped with a magnetic stirrer, as previously described [39]. The study was carried out in batch circulation mode at a flow rate of 600 mL min^-1. The UV light source was two Philips LIT 8 W low-pressure mercury lamps with peak emission at 254 and 366 nm. Typical irradiance of approximately 10 mW cm^-2 was measured using a Delta Ohm HD 2102.2 radiometer (Caselle di Selvazzano, Italy). Nitrogen was continuously fed into the photoreactor at a flow rate of 0.5 L min^-1 to ensure that oxygen did not interfere with the process. The system was thermostatically controlled by circulating water to maintain a constant temperature of 23±1 ⁰C. The photochemical reactor was filled with 2000 mL of Type II analytical grade water (pH 6.7, ORP 220 mV, conductivity < 20 µS cm^-1, TOC < 30 µg L^-1) and spiked with 0.2 mg L^-1 of each herbicide. After the addition of 250 mg L^-1 Na₂S₂O₈ or 250 mg L^-1 of TiO₂, the two UV lamps (1 x 254 nm + 1 x 366 nm or 2 x 366 nm) were switched on and degradation was studied for 120 min. All experiments were replicated three times (n = 3). Table 2 summarises the test carried out.

2.3. Sample Preparation and Analytical Determinations

Sample (25 mL) was passed through an Oasis^® HLB 60 µm (500 mg) extraction cartridge purchased from Waters (Milford, MA, USA) using a Visiprep™ SPE vacuum manifold supplied by Supelco (Madrid, Spain) at a flow rate of approximately 3 mL min^-1. The extraction cartridge was previously conditioned with 3 mL of CH₃OH and equilibrated with 3 mL of acidified ultrapure water (pH = 3). After passing through the cartridge, the sample was washed with 5 mL of Milli-Q water (18.2 MΩ·cm), the eluate was discarded and the column was dried with air. The analytes were then eluted with 5 mL CH₃OH and the extracts were evaporated to dryness. Finally, the residues were dissolved in 500 μL of CH₃CN and filtered through 0.22 μm polytetrafluoroethylene (PTFE) syringe filters prior to the chromatographic analysis.

Table 2. Summary of the trials carried out.

Trials		pH	aCE	bSO4=	bCl-	bHCO3-
1	UV (1 x 254 nm + 1 x 366 nm)	6.7	< 5	-	-	-
2	UV (2 x 366 nm)	6.7	< 5	-	-	-
3	Na2S2O8	5.1	240	-	-	-
4	Na2S2O8/UV (254/366 nm)	5.1	242	-	-	-
5	Na2S2O8/UV (366 nm)	5.0	246	-	-	-
6	TiO2/UV (366 nm)	5.3	< 5	-	-	-
7	Na2S2O8/UV (366 nm)	4.7	760	250	-	-
8	Na2S2O8/UV (366 nm)	8.3	377	-	-	125
9	Na2S2O8/UV (366 nm)	5.2	610	-	150	-
10	Na2S2O8/UV (366 nm)	8.4	1200	250	150	125

^a µS cm^-1; ^b mg L^-1.

Table 2. Summary of the trials carried out.

Trials		pH	aCE	bSO4=	bCl-	bHCO3-
1	UV (1 x 254 nm + 1 x 366 nm)	6.7	< 5	-	-	-
2	UV (2 x 366 nm)	6.7	< 5	-	-	-
3	Na2S2O8	5.1	240	-	-	-
4	Na2S2O8/UV (254/366 nm)	5.1	242	-	-	-
5	Na2S2O8/UV (366 nm)	5.0	246	-	-	-
6	TiO2/UV (366 nm)	5.3	< 5	-	-	-
7	Na2S2O8/UV (366 nm)	4.7	760	250	-	-
8	Na2S2O8/UV (366 nm)	8.3	377	-	-	125
9	Na2S2O8/UV (366 nm)	5.2	610	-	150	-
10	Na2S2O8/UV (366 nm)	8.4	1200	250	150	125

^a µS cm^-1; ^b mg L^-1.

The determination of herbicide residues was carried out using an HPLC system consisting of a Waters e2695 separation module (Waters, Mildford, USA) equipped with a quaternary pump and an autosampler and coupled to a Waters 2998 photodiode array detector (PDA). Data were processed using Water Empower software (version 3). Separation was performed on a 100 mm × 4.6 mm, 5 μm, Phenomenex Kinetex XB-C₁₈ analytical column (Madrid, Spain) with a mobile phase consisting of CH₃CN (solvent A) and an aqueous solution of HCOOH (0.1%) (Solvent B) under gradient mode, with the temperature of the column oven set to 25 °C. The following gradients were used: 30% of A for 1 min, linearly increased to 90% of A for 9 min, held for 1 min and decreased to 30% of A for 2 min to allow equilibration before the next injection (5 min). The flow rate was maintained (0.5 mL min^-1) and the volume injected was 50 μL. Confirmation criteria were retention times and recovered spectra (190-400 nm). The detection wavelengths were 222 nm and 241 nm for TBZ and ISP, respectively. Standard solutions containing the herbicides were used to construct calibration curves (0.1-1000 μg L^-1). The limits of detection (LOD) and limits of quantitation (LOQ) were obtained by dividing the signal-to-noise ratios of 3 and 10 times, respectively, by the angular coefficients of the calibration curves. In addition, HPLC-MS² analysis of the intermediates generated during the UV/PS process was performed on an Agilent 1200 HPLC system equipped with the same analytical column as previously described and coupled to an Agilent G6410A triple quadrupole (QqQ) mass spectrometer (MS) operating in electrospray (ESI) positive ion mode. The most abundant fragment ion was selected for quantification and the second for identity confirmation. Data acquisition was performed using MassHunter software.

Dissolved organic carbon (DOC) content was determined using a Multi N/C 3100 TOC Analyzer (Analytic Jena AG, Jena, Germany) after passing the samples through a nylon syringe filter (0.45 mm) prior to analysis to remove particulate OC from the sample. A GLP 21 pH-meter and GLP 31 conductivity-meter both of Crison Instruments (Barcelona, Spain) were used for pH and conductivity measurements, respectively. The statistical software SigmaPlot (Systat, Software Inc., San Jose, CA, USA) v.15 was used to fit the experimental data.

3. Results and Discussion

3.1. Photodegradation of herbicides by UV, Na₂S₂O₈ and UV/Na₂S₂O₈

The efficiency of degradation of the herbicides has been defined according to the following equation (Equation 8):

$$R_r\left(\%\right)=\left(\frac{H_{0 }-H_t}{H_0}\right)x 100$$

(8)

where R_r is the removal rate, H₀ is the initial concentration of herbicide and H_t is the concentration of herbicide at time t.

UV radiation has been proposed as an effective method of removing herbicides from water [40]. Most herbicides are photoactive because their structure generally contains aromatic rings, heteroatoms and other functional groups that make them susceptible to absorbing UV-vis radiation (direct photolysis) or to reacting with photosensitive species capable of inducing photodegradation of herbicides in water (indirect photolysis). The results of herbicide degradation using UV (254/366 nm), UV (366 nm), PS alone, UV (254/366 nm)/PS and UV (366 nm)/PS combined (trials 1-5) are shown in Figure 1.

Minimal degradation (R_r about 2%) was observed for both herbicides when PS was used in the dark. When UV irradiation (254 or 366 nm) was used, the R_r of ISP was increased to 16% (254 nm +366 nm) and 10% (366 nm only), while TBZ removal was 15% (254 nm +366 nm) and 7% (366 nm only) after 120 min photoperiod. Some studies showed that TBZ undergoes limited photodegradation and is considered photolytically and hydrolytically stable under at environmentally relevant temperatures and pH conditions [41]. However, photodegradation was drastically increased in both cases when PS was activated with UV, either at a combined wavelength (254 + 366 nm) or at 366 nm only. After 5 min of irradiation (254 + 366 nm), the levels of both herbicides in absence of inorganic anions were below the LOD, whereas after 30 min of irradiation with 366 nm only, both herbicides practically disappeared. Similar results were obtained by Lin and Wu [42], who compared the effectiveness of PS activation under two wavelengths (254 nm and 365 nm) on the photodegradation of polyvinyl alcohol. PS Activation by UV light might not be a promising method in industrial applications because of the high cost and the requirement of complex equipment. Although the activation of PS by UV/254 nm improve the efficiency of the process, it is important to remark that of the total radiant energy emitted by the Sun, only the part with a wavelength between 280-3000 nm, known as short-wave radiation, reaches the surface of our planet. This includes UVB (280-315 nm) and UVA (315-400 nm), but not UVC (100-280 nm). Therefore, and considering that the UVC fraction does not reach the earth’s surface, this fraction cannot be used in solar treatments for water purification because the spectrum of sunlight at earth’s surface begins around 300 nm. The UV-365 nm/PS can therefore be considered as a clearly economical and effective process for the degradation of the herbicides studied using natural sunlight. For this reason, the UVA fraction (366 nm) was used in the subsequent tests, in view of the satisfactory results obtained.

3.2. Comparing UV/TiO₂ and UV/Na₂S₂O₈ Efficiency

To compare the effectiveness on the degradation of the two herbicides, UV/PS (test 5) and heterogeneous photocatalysis using UV/TiO₂ as catalyst (test 6) were compared. Figure 2 shows the evolution of the residues of both herbicides under these conditions. As can be seen, in both cases a higher efficacy was observed for UV/PS than for UV/TiO₂, which is more notorious in the case of TBZ.

According to the law of mass action, the photodegradation of the herbicides could be modelled assuming pseudo-first order kinetics [43]. The kinetic study of the experimental results showed that the degradation process of both herbicides under UV basically conformed to the single first-order (SFO) kinetic model according to the following equations (Equation 9-11):

$$- \frac{d\left[H\right]}{dt}= kt$$

(9)

$${\left[H\right]}_t={\left[H\right]}_0 e^{-kt}$$

(10)

$$\ln {\left[H\right]}_t=\ln {\left[H\right]}_o-kt\frac{\ln {\left[H\right]}_o}{\ln {\left[H\right]}_t}=kt$$

(11)

where [H]_t is the concentration of herbicide in solution at time t, [H]₀ is the initial herbicide concentration in solution, t is the time (min), and k is the apparent reaction rate constant (min⁻¹). The slope of ln(C₀/C) plotted against time gives the value of k. From the above equation, the time required for X% disappearance of herbicides (disappearance time) from the water can be calculated according to Equation 12:

$${DT}_x=ln \left(\frac{100}{100-x}\right)/k$$

(12)

The fit to the selected model of the experimental data obtained is shown in Table 3. Consistent with the data shown in this table, the SFO model fitted the exponential decay curve realistically with R² ≥ 0.97 in all cases and standard error of estimation (S_y/x) < 0.05. The ratio between the apparent rate constants (k_PS/k_TiO2) was > 1 in both cases, 1.1 for ISP and 3.3 for TBZ, indicating a higher reaction rate for both herbicides using UV/PS compared to UV/TiO₂. For this reason, UV/PS was used to study the impact of inorganic anions on the photodegradation on the herbicides studied.

3.3. Effect of Inorganic Anion Content on Herbicide Photodegradation Using Na₂S₂O₈/UV

The composition of the water on the photodegradation process is critical to the assessment of its suitability for real wastewater treatment, and its impact can be complex in some cases. In addition to dissolved organic matter, the unfavorable effects of some inorganic ions (mainly anions) such as Cl^-, HCO₃^-/CO₃^2-, NO₃^-/NO₂^-, PO₄H^3- and/or SO₄^2-, among others can be explained by the fact that they reduce the oxidising power of the solution. The scavenging of HO^• and SO₄^•- by various anions generates the corresponding radicals such as ClOH^•−, Cl^•, Cl₂^•−, CO₃^•−, HCO₃^•, Br^•, Br₂^•−, NO^•, and/or H₂PO₄^•, which have a lower E⁰ than HO^• and SO₄^•- or by promoting the formation of ROS, as in the case of NO₃⁻, which is capable of generating HO^• and NO₂^• radical species, which promote the photodegradation of pesticides, especially those for which indirect photolysis is the main reaction pathway [28,38,44]. Some anions and cations present in water are transparent to solar radiation, while nitrate (λ = 303 nm) and nitrite (λ = 355 nm) show some adsorption. Both absorb light and undergo homolysis to form HO^• and nitrogen reactive species such as NO^•, NO₂^•, N₂O₃, and/or N₂O₄ such as leading to the herbicide degradation, although HO^•can be further scavenged by NO₂^- to form NO₂^• [45]. However, the concentrations of NO₃^-/NO₂ in real wastewater is usually low (< 5 mg L^-1). For this reason, we have assessed the influence of SO₄^2-, Cl^- and HCO₃^-, which are present at much higher concentrations.

As shown in Figure 3, the removal efficiency of the herbicides was not affected by the presence of sulphate at the concentration studied. Even, the presence of SO₄^2- in the solution (test 7) has a positive effect on the photodegradation of both herbicides, especially for ISP, because higher values of k were obtained according to the results shown in Table 4. SO₄⁼ is not a strong scavenger of HO^• such as Cl^-, HCO₃^- and CO₃^2-. On the contrary, it has been shown to promote the oxidative degradation of the antibiotics chloramphenicol [46] and ciprofloxacin [47] or polyvinyl alcohol [42], as occurs in our case. According to Cabrera-Reina et al. [48], the removal efficiency of acetamiprid, carbamazepine and caffeine was also unaffected by sulphate concentration in the range 0-550 mg L^-1. Therefore, sulphate impact on activated PS processes, summarized in Equation 13-14, is very low.

$${{SO}_4^{2-}+{HO}^{\bullet }\to SO}_4^{\bullet -}+{OH}^{-}$$

(13)

$${{{{SO}_4^{2-}+SO}_4^{\bullet -}\to S}_2{O}_8^{2-}+e}^{-}$$

(14)

Chloride ion (Cl⁻) is one of major inorganic anions in natural water. The adverse effect of Cl⁻, especially at high concentration on the degradation of many organic micropollutants has been reported by some authors [28]. In the case of ISP, a minimal influence was observed when chloride ions (150 mg L^-1) are present in the solution because k_PS/K_PS/Cl is 0.98 (Table 4). However, a significant influence was observed for TBZ where the ratio k_PS/K_PS/Cl was 4.1 as can be seen in Figure 3. The adverse effects of high Cl⁻ concentrations could favour the scavenging of HO^• and SO₄^•− according to the following reactions (Equation 15-18):

$${{Cl}^{-}+ { SO}_4^{\bullet -} \to {SO}_4^{2-}+Cl}^{\bullet }$$

(15)

$${Cl}^{-}+{HO}^{\bullet } \to {ClOH}^{\bullet }$$

(16)

$${ClOH}^{\bullet }+H^{+}\to H_{2}O+{Cl}^{\bullet }$$

(17)

$${Cl}^{-}+{Cl}^{\bullet } \to {Cl}_2^{\bullet -}$$

(18)

Compared to HO^•-based AOPs, radical scavenging by Cl^- in SO₄^•--based AOPs is a major challenge. The reaction between Cl^- and HO^• gives ClOH^•- (Equation 16), but this is a reversible reaction that mostly returns to HO^• [49]. In contrast, SO₄^•- produces Cl^• (E⁰ = 2.4 V) by one-electron abstraction from Cl^- (Equation 15), which is often reflected in lower efficiencies in organic pollutant removal as well as higher generation of reactive chlorine species such as Cl₂^•- (E⁰ = 2.1 V), ClO₂^• (E⁰ = 0.9 V), (ClO^• E⁰ = 1.4 V) and ClOH^•- (E⁰ = 1.9) V, all of them with lower E⁰ than HO^• and SO₄^•- [50].

An overall negative effect of HCO₃^- and CO₃^2- anions on the SO₄^•--based AOPs has traditionally been assumed [51]. The alkalinity of a water body is mainly contributed by CO₃^2-and HCO₃^- ions, which are generally present in natural waters ranging from 50-250 mg L^-1. Both CO₃^2-and HCO₃^- are known to be radical scavengers in AOPs [28]. Among them, the radical scavenging ability of CO₃^2- is stronger than that of HCO₃^- due to its higher reaction rate constant with OH^• (3.9 × 10⁸ M^-1 s^-1 vs 8.5 × 10⁶ M^-1 s^-1) [52].

However, depending on the chemical structure of the pesticides, the HCO₃^-/CO₃² effect will be neutral, positive or negative for their degradation rate, which makes it very difficult to predict due to the complex mixtures of pollutants always present in wastewater [53,54]. Although CO₃^•- have low redox potential than HO^•, it could exhibit better removal performing in degrading organic micropollutants, which could be due to its high selectivity and longer survival time in solution [55].

When HCO₃^- was added, the initial pH of the aqueous solution changed to about 8.3, rather than pH 6.7 without the addition of any anions, and the basic conditions reduced the efficiency of degradation of both herbicides, as discussed by some authors [56]. CO₂, CO₃^2- and HCO₃^- are present in aqueous media at pH > 4. CO₃^2- and HCO₃^-, responsible of water alkalinity, can compete with herbicides for HO^• and SO₄^•- to generate other weaker radicals, such as CO₃^•- and/or HCO₃^•. Above pH = 10.3, CO₃^2- is the prevalent specie, but at pH below 8.3 all CO₃^2- has been converted to HCO₃^- [57]. As pH decreases, HCO₃^- also decreases and dissolved CO₂ increases (Figure 4). The relationship between CO₃^2-/HCO₃^- and solution pH is represented as follows (Equation 19):

$$pH={pK}_a-log\frac{\left[{HCO}_3^{-}\right]}{\left[{CO}_3^{2-}\right]} {pK}_a=10.3$$

(19)

The high concentration (125 mg L^-1) of HCO₃^- strongly scavenges HO^• and SO₄^•-, generating electrophilic CO₃^•-, a more selective and less reactive reaction radical with a lower electrode potential (E⁰ = 1.6 V) than HO^• and SO₄ [58]. Detailed reaction mechanisms between CO₃^•- and micropollutants are still largely unclear. In previous studies, the single electron transfer was determined to be the dominant pathway [52]. The nature of CO₃^•- makes it more selective towards electron-rich moieties such as such as -NH₂, -OH and aromatic rings. HCO₃^• has lower reaction rate with organic pollutants than HO^•. For SO₄^•-, it has lower reaction rate with CO₃^2- (6.1 × 10⁶ M^-1 s^-1) and similar reaction rate with CO₃^2- (9.1 × 10⁶ M^-1 s^-1) compared to HO^• [59]. Nevertheless, the presence of CO₃^2-/HCO₃^- can cause the transformation of HO^• and SO₄^•-, which can further affect the performance of AOPs. In radical-based treatment processes, CO₃^•- and HCO₃^• can be formed by oxidation of CO₃^2- and HCO₃^- with highly reactive radical species, such as HO^• and SO₄^•- (Equation 20-23) [60]:

$${CO}_3^{2-}+ {HO}^{\bullet } \to {CO}_3^{\bullet -}+ {OH}^{-}$$

(20)

$${CO}_3^{2-}+{ SO}_4^{\bullet -} \to {CO}_3^{\bullet -}+{SO}_4^{2-}$$

(21)

$${{HCO}_3^{-}+{HO}^{\bullet } \to H}_{2}O+{CO}_3^{\bullet -}$$

(22)

$${{HCO}_3^{-}+{ SO}_4^{\bullet -} \to {SO}_4^{2-}+{{HCO}_3^{\bullet } \to {SO}_4^{2-}+H}^{+}+CO}_3^{\bullet -}.$$

(23)

In addition, the carbonate precipitation can result in fouling on the membrane surface, which further affects the removal efficiency of micropollutants [61].

After studying the individual effects of SO₄⁼, Cl^- and HCO₃^- on the UV/PS process, their combined effect was then assessed (test 10). The R_r of both herbicides is presented in Figure 3. As can be seen, Cl^- (150 mg L^-1) and HCO₃^- (125 mg L^-1) generally had a negative effect on herbicide R_r, although the magnitude depended on the chemical structure of each herbicide. As can be observed in Table 4 the R values ranged from 0.976 to 0.996 for TBZ and 0.988 to 0.998 for ISP, with standard errors of estimation less than 0.08 and 0.05, respectively, indicating good fit. The time to 50% disappearance (DT₅₀) for ISP was slightly higher in the presence of anions (8.2 min) than in the corresponding experiment carried out in the absence of anions (5.5 min). However, a substantial effect was observed for TBZ, as the DT₅₀ was approximately 23 times higher than in the absence of anions, mainly due to the impact of HCO₃^-.

The initial pH of the aqueous solution of herbicides in absence/presence of inorganic anions decreased in all cases during the photoperiod (Table 5). The acidification may be caused by the formation of acidic products as consequence of the herbicide degradation, and acidic photoproducts derived from S₂O₈⁼, such as HSO₄⁻, with release of H⁺ (Equation 24, 25) as reported by Yang et al. [28]:

$${SO}_4^{•-}+ H_{2}O {HO}^{•}+H{SO}_4^{-}$$

(24)

$$H{SO}_4^{-} {SO}_4^{2-}+H^{+}$$

(25)

In addition, the concentration of SO₄^2- increased significantly throughout the process (Equation 25), as can be seen from the EC values shown in Table 6, due to the transformation of the initially added S₂O₈^2- into SO₄². Various methods have been proposed to remove SO₄⁼ from water over the past decades, such as adsorption on activated carbon, neutralisation with calcium carbonate, biological treatment, reverse osmosis and dialysis, and ion exchange. The choice of wastewater treatment method is usually based on the type of wastewater, the removal rate, the waste concentration and the cost of treatment. Of these, adsorption on an ion exchange resin is the most popular method for removing SO₄⁼ from water and wastewater and shows good potential for industrial wastewater treatment [62].

Figure 5 shows the evolution of DOC during the different tests carried out. It is important to monitor the process using this tool because only DOC values close to zero guarantee that recalcitrant pollutants or intermediates with greater persistence and toxicity than the initial ones do not persist. In the absence of any ions (test 5) and in the presence of SO₄⁼ (test 7) a high mineralisation rate was observed because the remaining DOC after 120 min was less than 5%. However, in the presence of HCO₃^- (test 8) and Cl^- (test 9), the residual DOC concentration after 120 min was 20% and 10% respectively, while in the presence of all anions in the reaction solution (test 10), the residual DOC concentration at the end of the experiment was significantly increased (37%). These residual DOC fraction may be due to the partial removal of herbicides in aqueous solution and the formation of non-degradable organic intermediates produced during the irradiation time.

3.4. Degradation Pathway of Herbicides with UV/Na₂S₂O₈

To gain a better understanding of the reaction mechanisms involved in the photodegradation of both herbicides, the evolution of key intermediates during the irradiation experiment was also followed by comparison with authentic analytical standards using quantitative HPLC-MS² analysis. Figure 6 shows the proposed degradation pathways for both herbicides. The metabolic pathway involved in the degradation of ISP mainly involves two successive N-demethylations followed by hydrolysis to the aniline-based metabolite which can then be further degraded. Photooxidation of TBZ leads to dealkylation of the amine groups, dechlorination and subsequent hydroxylation [66].

Three transformation products of ISP, 1-(4-Isopropylphenyl)-3-methylurea (ISP-1), 1-(4-isopropylphenyl)urea (ISP-2) and 4-isopropylaniline (ISP-3) were identified during experiment 10, although their concentrations were below their LOQs (<-2 µg L^-1) at the end of the photoperiod (120 min). The maximum concentration peak was found after 30 min from the start, reaching levels between 4 and 8 µg L^-1. These intermediates were also isolated by other authors in aqueous suspensions of TiO₂ and ZnO under UV light [63,64]. In the case of TBZ, the photooxidation of the parent compound was accompanied by the appearance of desethyl-terbuthylazine (TBZ-1), deisopropyl-atrazine (TBZ-2) and terbuthylazine-2-hydroxy (TBZ-3), whose residual levels after 120 min were in the range of 0.5-1.2 µg L^-1, respectively. Fenoll et al [65] found TBZ-1 and TBZ-3 intermediates using TiO₂ and ZnO as photocatalysts. The presence of these degradants confirms the behaviour observed in the evolution of DOC indicating that there is no total mineralisation of the herbicides in the presence of anions. It should be remembered that for mineralisation of the pollutant to occur, not only must it disappear, but all the organic carbon must be converted into inorganic carbon in the form of CO₂, since the DOC value is independent of the oxidation state of the compounds present in the aqueous solution.

4. Conclusions

The need to increase the supply of water is linked to both the scarcity of the quantity available and the deterioration of its quality. There are many families of chemical compounds, many of which are organic in nature, that can have a negative impact on natural ecosystems and adversely affect water quality, thus posing a risk to human health. Among these, pesticides stand out for their toxicity and persistence, and their control is currently a priority at institutional level. Many of them are carcinogenic, endocrine disrupting or even teratogenic and highly persistent in the environment due to their complex structures and resistance to degradation. As a result of the current water scarcity, more emphasis should be placed on the 3Rs (Recover, Recycle and Reuse) approach to wastewater treatment, both industrial and agricultural.

Each water treatment method has its own drawbacks, including those related to feasibility, availability of equipment or space, disposal efficiency, environmental impact, sludge and by-product generation, operational difficulties and/or pre-treatment requirements. For this reason, it is now essential to develop remediation techniques that favour the total elimination of all traces of pollutants in water, both of natural and anthropogenic origin, in order to promote its subsequent reuse, all with the aim of achieving the zero-pollution goal announced in the European Green Deal in relation to the Chemicals Sustainability Strategy.

The need for treatment technologies that minimise environmental hazards at a reasonable cost has been driven by increasingly stringent environmental regulations regarding the presence of emerging contaminants in wastewater and natural systems. Conventional water treatment processes and technologies reduce the level of water pollution to a significant extent, but not to the extent required by current legislation. In addition, approaches based on adsorption, flocculation and ion exchange are not fully effective. Rather than destroying the contaminant, they transfer it to a solid phase that requires subsequent treatment. Due to the refractory nature of contaminants, biological treatments are often ineffective, which can be toxic to the microorganisms used. The use of artificial UV light in photochemical processes may not be a promising method for environmental applications due to high costs and the need for complex equipment. However, in the case of solar-powered applications, the process is not expensive because of the use of solar radiation.

AOPs, usually applied after biological processes, have recently emerged as effective tertiary treatments for the removal of pesticides and other micropollutants, but the oxidation rates of individual compounds can be strongly influenced by the constituents of the water matrix. Overall, the removal of the herbicides studied (isoproturon and terbuthylazine) by HO^• and SO4^•--based AOPs is largely influenced by the quality of the water matrix (dissolved constituents), which can be neutral, inhibitory or promoting, depending on the process and the mechanism by which these water constituents react. In addition to organic species, inorganic species, mainly chloride and carbonate/bicarbonate ions can also act as inhibitors either by scavenging or by generating new radicals that have a lower E⁰ than HO^• and SO₄^•-. Herbicide photodegradation by UV/PS treatment is faster in the absence of inorganic compounds, largely due to the absence in pure water of organic and inorganic compounds that can absorb UV radiation and generate other ROS with lower oxidation potential.