Advanced Oxidation of Dexamethasone by Activated Peroxo Compounds in Water Matrices: A Comparative Study

Abstract

The continuous occurrence of steroidal pharmaceutical dexamethasone (DXM) in aqueous environments indicates the need for an efficient removal technology. The frequent detection of DXM in surface water could be substantially reduced by the application of photo-induced advanced oxidation technology. In the present study, Fe²⁺ and UVA-light activated peroxo compounds were applied for the degradation and mineralization of a glucocorticoid, 25.5 µM DXM, in ultrapure water (UPW). The treatment efficacies were validated in real spring water (SW). A 120 min target pollutant degradation followed pseudo first-order reaction kinetics when an oxidant/Fe²⁺ dose 10/1 or/and UVA irradiation were applied. Acidic conditions (a pH of 3) were found to be more favorable for DXM oxidation (≥99%) regardless of the activated peroxo compound. Full conversion of DXM was not achieved, as the maximum TOC removal reached 70% in UPW by the UVA/H₂O₂/Fe²⁺ system (molar ratio of 10/1) at a pH of 3. The higher efficacy of peroxymonosulfate-based oxidation in SW could be induced by chlorine, bicarbonate, and carbonate ions; however, it is not applicable for peroxydisulfate and hydrogen peroxide. Overall, consistently higher efficacies for HO^•-dominated oxidation systems were observed. The findings from the current paper could complement the knowledge of oxidative removal of low-level DXM in real water matrices.

1. Introduction

Human activities, such as the rapid development of industrial processes, ongoing urbanization and growth of population, have an exhausting impact on the environment. One directly affected environmental compartment is water, where the presence of anthropogenic contaminants, primarily in concentrations of ng L⁻¹ to mg L⁻¹, is being reported [1]. Among these substances, pharmaceuticals are a major concern due to their bioactive and often pseudo-persistent properties in environmental compartments, induced by their regular use [2]. Corticosteroids, also called glucocorticoids, are a group of steroidal anti-inflammatory drugs. Dexamethasone (DXM), representative of this group, has been generally used for inflammation, disorders associated with adrenocortical insufficiency, and allergies, and is included in the World Health Organization’s list of essential medicines [3]. Furthermore, it was successfully used in severe cases of respiratory failure during the COVID-19 pandemic due to its ability to minimize the effects of pro-inflammatory cytokines, thereby reducing disease-related mortality [4,5]. Subsequently, DXM-related issues in water bodies have continued to escalate rather than reaching a plateau or debilitation. In addition to human medicine, it is used in veterinary medicine. As 50–90% of DXM is excreted into sewage, aquatic organisms, such as algae, bacteria, and fish, later face possible direct ecotoxicological issues affecting their reproduction [5].

Traditional treatment methods in wastewater treatment plants involve activated sludge processes, which often have a limited capacity for the removal of pharmaceuticals. Secondary treated effluents are one of the most important transmission pathways for DXM into the aquatic environment, e.g., via drinking and surface water, with DXM detected worldwide in surface waters from 0–40 ng L⁻¹, and up to 350 ng L⁻¹ in wastewater treatment plant effluents [4,6]. 1.The above-mentioned compounds have relatively stable properties and complex molecular structures, are challenging to degrade in the natural environment, and can also possibly interact with other organic pollutants, resulting in more complex molecules [7]. Hence, the chemical and biological properties of pharmaceuticals or organic micropollutants, in general, are often limiting factors in the applications of primary water treatment stages, like coagulation–flocculation, filtration, and adsorption, and even at tertiary stages such as disinfection [8,9]. For instance, the high solubility and non-particulate nature of pharmaceuticals causes them to bypass the flocculation step [10]. Unlike inorganic pollutants, the ionized form of these molecules makes them less interactive with classical adsorbent materials [11,12]; very small molecules can pass through filtration units [13].

To relieve this situation, sustainable and environmentally safe remediation processes should be taken into consideration. One approach to consider involves the use of advanced oxidation technologies (AOTs), which are known for their ability to decompose and mineralize organic pollutants to carbon dioxide and inorganic ions [14]. AOTs have proven to be applicable towards species that are toxic and recalcitrant/resistant to biological or physicochemical treatments [15,16].

In general, AOTs are based on a variety of oxidative agents, such as hydrogen peroxide (H₂O₂), persulfates, ozone, ultraviolet (UV) light, that are used to generate reactive species to degrade a wide range of recalcitrant pollutants [17]. The classic use of AOTs, usually referred to as advanced oxidation processes (AOPs), involves hydroxyl radical (HO^•) generation via H₂O₂ reaction with a source of Fe²⁺-ions in an acidic medium—the so-called Fenton reaction (Equation (1)) [17]:

Fe²⁺ + H₂O₂ → Fe³⁺ + HO^• + HO⁻

(1)

The last decade has seen an increase in studies of sulfate radical (SO₄^•–)-based AOTs, which are known to react more selectively with organic pollutants than HO^• [18,19]. SO₄^•− has a higher oxidation potential than HO^• and is present in aqueous solutions for a longer period without requiring a strictly acidic environment, from a pH of 2 to 10 [19]. Moreover, in complex environments containing large amounts of dissolved organic matter and anions, which act as the main HO^• scavengers, the efficiency of HO^•-based AOPs is limited [14].

The source of SO₄^•– is peroxydisulfate (PDS, S₂O₈²⁻) or peroxymonosulfate (PMS, HSO₅⁻), activated with transition metal catalysts, strong oxidants, heat, or UV light, with the indirect formation of HO^• (Equations (2) and (3)) [7,20]:

S₂O₈²⁻ + activator → SO₄^•− + (SO₄^•– or SO₄²⁻)

(2)

HSO₅⁻ + activator → SO₄^•− + (HO^• or HO⁻)

(3)

The activation by transition metals generates single SO₄^•−; however, energy-related activation induces cleavage of the peroxide bond in PDS, directly generating a pair of SO₄^•−. PMS, on the other hand, can produce SO₄^•− along with HO^• [20].

Regarding the wide detection of DXM in surface waters, a promising solution to consider could be photo-induced AOTs. The source of light in photo-induced degradation experiments can be either natural or artificial, the last being preferred thanks to its controllability [21,22]. Natural sunlight, however, is not preferable for fundamental studies due to uncontrollable factors such as intensity variations and the spectral distribution of solar radiation. These can vary based on multiple aspects, e.g., geographical location, day of the year, climate conditions, variations in altitude, and local weather conditions [21,22]. The use of lamps enables researchers to conduct laboratory-scale photodegradation studies in controlled settings, thereby providing reproducible and reliable results [21]. However, for a realistic outcome, the natural characteristics of sunlight, namely the spectral distribution and irradiance level, should be simulated. Thus, UVA light has been selected as a supplementary activator in DXM degradation, as it is a major part of the solar UV radiation component that reaches the earth’s surface.

DXM removal in water matrices has been previously studied by applying chlorination [23], adsorption [24], electrochemical processes [25,26], sonochemical oxidation [27,28], photocatalysis [5,29,30,31,32,33], and UVC-activated H₂O₂ [33]. In addition, thermally activated [6], UVC-activated [34], UV/Al₂O₃-activated [35] and molybdenum carbide-activated [36] PDS have been applied, with a significant focus on advanced photo-induced processes due to the widespread occurrence of DXM in surface water. This study aims to provide complementary knowledge about applications in real water matrices with long-wave irradiation. Findings from this paper could fill a gap in the research concerning the peroxo-compound oxidative removal of low-level concentrations of glucocorticoids.

Therefore, the aim of this study was to investigate and compare the degradation and mineralization efficiency of DXM in ultrapure water (UPW) by activated H₂O₂, PMS and PDS oxidation. The type of oxidant, the effect of initial pH, the addition of ferrous iron as an activator, the use of UVA radiation, and the composition of the spring water (SW) matrix on the efficiency of the target compound decomposition were assessed.

2. Materials and Methods

2.1. Chemicals and Materials

DXM (C₂₂H₂₉FO₅, ≥98%, molecular weight 392.5 g mol⁻¹, Figure 1) was purchased from Alfa Aesar (Heysham, UK). Sodium sulfite (Na₂SO₃, 99%), ethanol (C₂H₆O, 99%), acetonitrile (CH₃CN, LiChrosolv^®, ≥99.9%) and acetic acid (CH₃CO₂H, 100%) were provided by Merck KGaA (Darmstadt, Germany). Ferrous sulfate heptahydrate (FeSO₄·7H₂O, ≥99%), sodium PDS (Na₂S₂O₈, ≥99%), PMS (Oxone^®, KHSO₅·0.5KHSO₄·0.5K₂SO₄), H₂O₂ (PERDROGEN™, ≥30%), sulfuric acid (H₂SO₄, 95–98%) and sodium hydroxide (NaOH, ≥98%) were supplied by Sigma-Aldrich (St. Louis, MO, USA). All the chemicals were of analytical grade and used without further purification. All stock solutions were prepared in UPW (Millipore Simplicity^®UV System, Merck, Kenilworth, NJ, USA) or in bi-distilled water (>18.2 MΩ cm).

SW, without the preceding purification (Valgamaa, Estonia), was used as a real aqueous matrix for the DXM oxidation trials. The SW is collected into a well reservoir and further used as a direct drinking water source. The sample was collected during springtime in April/May with the main parameters given in

Table 1. Chemical composition and main parameters of the SW sample.

Parameter	Unit	Value
pH	-	7.3
Alkalinity	mg CaCO3 L−1	308
Conductivity	µS cm−1	621
Total organic carbon (TOC)	mg L−1	31.3
Fe2+		0.08
Total Fe		0.11
Cl−		15.0
NO3−		8.30
SO42−		9.81
Na+		8.18
K+		8.06
Ca2+		77.1
Mg2+		14.0

. One can see the buffering capacity of the groundwater provided by its alkalinity.

2.2. Experimental Procedure

The experiments were conducted in batch mode at ambient room temperature (22 ± 1 °C) with an initial DXM concentration of 25.5 µM. The concentration employed in this study exceeds those typically reported in contaminated natural and wastewater matrices to ensure the reliability and reproducibility of visible results. DXM solutions (820 mL) in UPW or in SW were treated in a 1.0-L cylindrical glass reactor for a period of 120 min with a permanent agitation speed (400 rpm). The pH of the solution was either unadjusted at 5.6 ± 0.1, adjusted to 3 by using diluted H₂SO₄, or to 11 by using diluted NaOH. Further procedures were carried out as described in our earlier works [37,38]. The Fe²⁺ source (FeSO₄·7H₂O) was added and after its complete dissolution, the reaction was initiated by adding PDS, PMS, or H₂O₂ and simultaneous exposure to the UVA lamp. A low-pressure mercury lamp (11 W OSRAM Dulux S BLUE, Berlin, Germany) located in a quartz tube inside the reactor was used as the UVA source. The lamp was turned on 5 min prior to setting into the reactor for constant output. A water-cooling jacket was used to keep the temperature stable in the reactor. Sample aliquots were taken at pre-determined time intervals from 0 to 120 min and filtered through 0.45 μm pore-size syringe filters (CA, Millipore™, Cork, Ireland). The oxidation quenching was conducted by the addition of EtOH (sample/EtOH volume ratio of 10/1) or Na₂SO₃ at a [oxidant]₀/SO₃²⁻ molar ratio of 1/10 for HPLC–MS or TOC analysis, respectively. All experiments were duplicated; the final data show the results of at least three parallel replicates with a standard error within <5% and are presented as mean.

2.3. Analytical Methods

The DXM concentration was quantified using high-performance liquid chromatography combined with diode array detector and mass spectrometer (Shimadzu HPLC–MS 2020, Shimadzu, Japan) equipped with a Phenomenex Gemini (150 × 2.0 mm, 1.7 mm) NX-C18 (110 Å, 5 µm) column. The analysis was performed using an isocratic method with a mobile phase mixture of 40% acetonitrile (with 0.3% formic acid) and 60% formic acid (0.3%) aqueous solution. Samples (75 µL) were analyzed at a flow rate of 200 µL min⁻¹. Mass spectrometry was applied in full-scan and in positive-ionization mode. Samples were scanned at 50–500 m/z. The concentration of DXM was determined by using the standard multipoint calibration. The results obtained with the MS detector were processed using Shimadzu Lab Solutions software (version number 5.81.1).

Total organic carbon (TOC) was measured by a TOC analyzer multi N/C^® 3100 (Analytik Jena, Jena, Germany) in 20 mL samples with an injection volume of 500 µL for each replicate. TOC was calculated as total carbon (TC) minus total inorganic carbon (TIC). The standard deviation for TIC and TC analysis was less than 1.5% and 1.2%, respectively. Ion chromatography with chemical suppression of the eluent conductivity was used to measure the concentrations of anions in SW (761 Compact IC, Metrohm Ltd., Herisau, Switzerland), whereas the concentrations of cations were measured without chemical suppression (ECO IC, Metrohm Ltd., Herisau, Switzerland). The pH was measured using a digital pH/ion meter (Mettler Toledo S220, Zurich, Switzerland). The electrical conductivity was measured using a digital EC meter (HQ 430d flexi, HACH Company, Loveland, CO, USA). The initial H₂O₂ concentration in the stock solutions was measured spectrophotometrically at λ = 254 nm (ε = 19.6 L mol⁻¹ cm⁻¹). The residual H₂O₂ concentrations in the treated samples were measured by a spectrophotometric method at λ = 4 10 nm with titanium sulfate by a H₂O₂-Ti⁴⁺ complex formation [39] using a Genesys 10S UV–Vis spectrophotometer (Thermo Scientific, Waltham, MA, USA). The residual persulfate concentration in the treated samples was measured spectrophotometrically at λ = 352 nm by a sodium iodide reaction with persulfate towards the formation of I₂ [40]. The alkalinity of the SW was measured by titration with hydrochloric acid (0.1 M) in the presence of methyl orange.

3. Results and Discussion

3.1. Direct Oxidation of DXM

The non-activated uses of three peroxo compounds, PDS, PMS and H₂O₂, in the target compound degradation with a DXM/oxidant molar ratio of 1/1 at an unadjusted pH were compared. Figure S1 shows a minor, <15%, removal of DXM after a 120 min experiment by only adding the oxidant to the studied system.

Two non-activated persulfate compounds showed distinctive efficacies in the direct oxidation of DXM. While no significant decrease (~1%) in DXM concentration was achieved using PDS, the studied pollutant concentration decreased to 12.5% with the PMS addition. Ding et al. [41] determined the structural differences of the oxidants that contribute to different pollutant reduction efficacies. PMS, with an asymmetric peroxide bond (⁻O₃SO-OH), having sulphite on one side and hydrogen on the other, was compared to symmetrical PDS (⁻O₃SO-OSO₃⁻). It was shown that PDS is less active than PMS in the nucleophilic attack of organic pollutants with electron-rich moieties, the latter being more easily oxidized by non-activated/non-catalyzed PMS via electron transfer.

Berruti et al. [42] conducted several studies on the direct PMS oxidation of organic pollutants, such as trimethoprim, sulfonamides and carbamazepine, where the presence of SO₄^•− and HO^• was experimentally excluded. It was proposed that the target pollutant degradation was initiated by singlet oxygen (¹O₂) from the decomposition of PMS or direct PMS oxidation. However, the ¹O₂ involvement in organic molecule degradation was considered marginal. Again, this study led to the conclusion that the main degradation pathway between PMS and pollutants is most likely a direct reaction, where the oxidant is involved in the nucleophilic attack of various organic contaminants with electron-rich moieties. The modest efficacies of both persulfates endorse findings in the literature that despite their strong oxidative properties, the direct reaction rates with most organic pollutants are low [7]. It is necessary to activate them through appropriate ways to destroy the O-O bond and generate strong oxidizing free radicals, ¹O₂, and other reactive oxygen species to degrade organic pollutants in the shortest time and with the highest efficacy [7]. Additionally, PMS has dissociation constants of pK_a1 < 0 and pK_a2 = 9.4; therefore, in a pH range of 4 to 8.5, it is present as an HSO₅⁻ ion [43]. The inhibitory effect of H⁺ could be described by the effect of stabilizing H⁺ on HSO₅⁻.

DXM oxidation by H₂O₂ alone was also negligible (8%), thereby reemphasizing the need for an activator to be effectively used for organic pollutants degradation. This is consistent with the findings of Targhan et al. [44], who reported that despite being a strong oxidant, the full potential of H₂O₂-based systems to degrade organic compounds emerges only when combined with catalysts or additional reagents.

Measured pH values before and after the applied peroxo compounds support the data in Figure S1; the solutions treated with PMS indicated a minor decline from the initial pH, whereas in the case of PDS and H₂O₂, it remained close to the initial unadjusted value.

The mineralization of the target compound showed no changes in the initial concentration of TOC, further confirming that DXM oxidation trials should be performed with the addition of an activator in the treatment process.

3.2. Iron-Activated Oxidation of DXM

3.2.1. The Effect of pH

Further experiments were carried out with the addition of an activator, ferrous iron, at different pH values; the results of the studied combinations are presented in Figure 2. The initial pH was set to strongly acidic at 3, unadjusted at the near circumneutral region, or strongly alkaline at 11. It must be noted that no buffer solution was used to limit possible interference of the formed radicals with the buffer chemicals.

The most efficient DXM degradation was achieved at an acidic pH (Figure 2a), with the highest DXM removal rate (nearly 60%) after the 120 min activated-H₂O₂ treatment. This result is consistent with the classical Fenton process concept, where the acidic medium enhances the generation of HO^•, as a pH between 2.8–3.0 provides maximum catalytic activity for the reaction [45]. In addition, one of the DXM pKa values is 1.8, which suggests that at a low pH range, the compound tends to remain on a molecular structure [27,29]. That could contribute to the high pollutant removal rates at a pH of 3. A similar observation was made by Ghenaatgar et al. [30] during a study on photocatalytic degradation of DXM using zirconium dioxide and tungsten trioxide. Accordingly, the initial pH was set to values from 3 to 10, with the highest and most complete target compound removal achieved at a pH of 3 and 4, respectively.

The activated H₂O₂ treatment also remained effective in the target pollutant removal at an unadjusted pH (Figure 2b), with 39%, while the applied PDS and PMS combinations indicated similar efficacies of 24% and 20%, respectively. Regardless of the peroxo compound studied, the final pH value was close to 4. The decrease in pH could indicate a moderate formation of acidic intermediate degradation products; thus, the decomposition of DXM took place to some extent.

Arvaniti et al. [6] conducted DXM degradation experiments with thermally activated PDS with initial pH values from 3 to 9. Their findings confirmed our observation that the oxidation efficiency of the target pollutant gradually decreased as the pH increased from 3 to 9. Acidic mediums may have an advantage over activated PDS oxidation; these could react with H⁺ and thereby increase the concentration of SO₄^•–, according to Equations (4) and (5). One can conclude that low pH values possibly favor the removal of organic contaminants:

S₂O₈^2– + H⁺ → HS₂O₈⁻

(4)

HS₂O₈⁻ + e⁻ → SO₄^2– + SO₄^•– + H⁺

(5)

According to Gao et al. [46], when Fe²⁺ is used for PDS activation at a pH higher than 4.0, the available Fe²⁺ is transformed into ferrous complexes that inhibit the further reaction between Fe²⁺ and PDS to form SO_4•⁻. Moreover, when the pH is increased to above 4.0, the ferric hydroxide complexes, such as FeOH²⁺, Fe₂(OH)₂⁴⁺, Fe(OH)²⁺, Fe(OH)₃, and Fe(OH)₄⁻, formed in the solution have a low activation capacity to generate SO₄^•− from PDS; therefore, the organic pollutant degradation is hindered [47].

The alkaline conditions seem to have obstructed the DXM degradation noticeably, as nearly 3% by PMS/Fe²⁺, 7% by PDS/Fe²⁺ and 8.5% by H₂O₂/Fe²⁺ of the initial target compound concentration was removed after 120 min of treatment time. No changes in the pH values (11) after the oxidation trials were noted. In the case of activated H₂O₂, the alkaline conditions were most likely unsuitable for oxidant and activator productive interaction, as mentioned earlier. Additionally, a study on DXM degradation via a photocatalytic UV/H₂O₂/MgO process also reported a sudden decline in efficacy when the pH was adjusted to 9 or 11 [32]. The authors indicated that the solution’s alkalinity and H₂O₂ could undergo a reaction that causes the HO^• to decrease. In addition to the above, the impact of the alkaline pH on DXM degradation could also be associated with the assumed ionic form of DXM (pKa 12.42) [6].

Regarding PDS and PMS, base activation is one possible method, especially for in situ applications of chemical oxidation. This allowed the authors to assume that peroxo compounds could have an advantage over H₂O₂; however, the superiority, although negligible, of H₂O₂ was observed. Wacławek et al. [48] also revisited aspects related to both PDS and PMS and concluded that the alkaline activation of these compounds is very strongly dependent on pH. Superoxide radical (O₂^•−) is a common predominant radical for both compounds at these activation parameters, with SO₄^•− and HO^• only characteristic with respect to PDS [48].

Furthermore, the pH value determines the dominant radical species of the process in activated persulfate treatment systems. PDS is known for its stability over a wider range of pH values, while PMS can be sensitive to pH variations. PMS is stable at a pH < 6 or a pH of 12, while at pH = 9, its stability is lowest when half of it decomposes to SO₅²⁻ [49]. It is known that under an alkaline pH, PDS can produce SO₄^•−, which can further transform into HO^•. Thus, the latter could be a lead radical at an alkaline pH, while under neutral pH conditions, both radical species contribute to the organic pollutants’ degradation.

Most likely, a joint inhibiting effect of the pH to the target compound degradation occurs with the use of an iron activator. When the pH is set above 9, species of oxyhydroxylperic acid, Fe(OH)₃^•, Fe(OH)₄⁻, FeOH³⁺, and Fe₂(OH)₃⁴⁺, are produced that have a very modest ability to activate PDS [49]. On the other hand, the activator amount in the used oxidant/activator molar ratio of 1/1 could be unsuitable and limit the process efficacy with both the oxidant and activator doses subjected to upsurge to provide an elevated quantity of liberated radicals. Moreover, the residual oxidant concentration after the applied processes increased with the pH increase, with one exception. A full conversion of H₂O₂ at an unadjusted pH was reached, followed by 80% use at a pH of 3, supporting the high efficacy of the activated oxidant.

Similar to the direct oxidation experiments, no conversion of the target compound was observed with a maximum TOC removal of <1%. This indicated that the studied treatment systems were ineffective; thus, an improvement was needed. The solution was to increase the oxidant dose at acidic pH conditions based on the obtained results so far.

3.2.2. The Effect of Oxidant Dose

Initial experiments with ferrous iron-activated peroxo compounds were carried out at a DXM/oxidant/Fe²⁺ molar ratio of 1/1/1. However, increasing the oxidant dose often showed progress in pharmaceuticals removal and mineralization [37,49]. The pH value adjustment to 3 provided the presence of an activator in its ferrous form, which is known to provide the effective activation potential of H₂O₂ and PDS [37,45]. The results obtained at an oxidant/Fe²⁺ molar ratio of 10/1 are presented in Figure 3.

The results from of our previous studies [37,50] had indicated the high efficacy of H₂O₂-based oxidation, and as anticipated, a 90% removal of the target compound after 120 min was observed. Both of the applied persulfates, on the contrary, had a lower impact on the organic contaminant degradation at an acidic pH. Accordingly, activated PDS and PMS oxidation resulted in 52% and 74% degradation of DXM, respectively. However, compared to 10-fold lower oxidant doses at a pH of 3 (Figure 2), the following significant improvements were observed: 3 times with PDS/Fe²⁺, nearly 2 times with PMS/Fe²⁺ and 1.5 times with the H₂O₂/Fe²⁺ system, respectively.

As previously noted, a similar order in the DXM degradation efficacies were observed, where the peroxide-based process resulted in the highest efficacy, followed by PMS, and PDS, respectively. HO^• formed in the Fe²⁺-activated H₂O₂ oxidative system, interacting quickly with the organic pollutants. The DXM structure, with many OH groups, is prone to responding well to HO^•-initiated oxidation. Therefore, the application of persulfates could result in slightly lower treatment efficacy than Fenton oxidation.

Under the studied strongly acidic condition, DXM degradation trials with all the studied combinations followed pseudo first-order reaction kinetics (R² ≥ 0.91) and can be described with regard to the DXM concentration by Equation (6):

$${\displaystyle \frac{{dC}_{DXM}}{dt}} =-k_{obs}\times C_{DXM}$$

(6)

where k_obs is the observed pseudo first-order rate constant and C_DXM is the DXM concentration. The k_obs constants were calculated from the slopes of the straight lines by plotting ln (C_t/C₀) as a function of time t through linear regression. The k_obs values were as follows: 0.0055 min⁻¹ for PDS, 0.0099 min⁻¹ for PMS, and 0.0155 min⁻¹ for H₂O₂ in the oxidant/Fe²⁺ system, respectively. It should be noted that pseudo first-order reaction conditions were met for the oxidation processes under which the reported k_obs values were obtained.

TOC measurements in the ferrous iron-activated peroxo compound systems at a DXM/oxidant/Fe²⁺ molar ratio 1/10/1 showed an overall low efficiency in terms of converting the target compound to CO₂ and H₂O, with a maximum DXM mineralization of 4%. The residual concentrations of persulfates in the oxidative systems indicated that the molar ratio of oxidant/Fe²⁺ could be further adjusted, either for the lack of an activator or for an excess of oxidant; stepwise addition of the activator could also be considered. However, in the current paper, the actuation to add a stronger activation aid—UVA radiation—to the system was preferred based on the available data of environmental occurrences of the studied compound.

3.3. Photochemical Oxidation of DXM

The Effect of pH

Glucocorticoids are light-sensitive compounds extensively found to be largely unmetabolized in drinking and surface waters, which raises concern for the possibility of toxic photochemical derivatives in the environment [51,52]. Thus, the use of UVA light in close-to-environmental-conditions in the treatment system was more than justified. A set of experiments was carried out at an oxidant/Fe²⁺ molar ratio of 1/1 at a pH of 3, unadjusted, and at a pH of 11; the results are presented in Figure 4 and Figure S2.

Overall, the use of a transition metal activator in the studied combinations proved the importance of the pH value beforehand. Thus, at first, strongly acidic conditions were set (Figure 4a) to provide the presence of the activator in its ferrous form, which was proven to have higher potential in the activation of persulfates and H₂O₂. The symbiosis with UVA light resulted in ≥99% removal of DXM with all the studied oxidants in 30–60 min treatment time. In the PMS system, a rapid 25% decomposition of DXM was noticed in the first 30 s. In the case of H₂O₂, similar removal was achieved at 3 min; in the case of PDS, removal was achieved between 3 and 5 min. Therefore, PMS exposure with UVA light indicated significant oxidative potential. The increased efficiency is most likely promoted by SO₄^•− and HO^•, which are PMS decomposition products under UV radiation by cleavage of the peroxide bond. In the case of PDS, it yields in a pair of SO₄^•−.

The effect of the iron activator addition to the UVA-induced oxidative systems should not be underestimated. In the photo-induced H₂O₂ oxidative system, which under an acidic pH is better known as the photo-Fenton process, the exposure to UV radiation induces more intensive generation of HO^• through the photolysis of the formed Fe³⁺ hydroxo complexes [53]. This could promote the constant enhanced removal of the target compound. In the classical Fenton process, it has been previously well proven that acidic conditions deliver the highest effect.

As the pH increased, the efficiency of DXM removal decreased in all the photo-oxidation systems studied. In the case of the PDS-based process, a DXM reduction of 94% was achieved at an unadjusted pH (Figure 4b), while at a pH of 11 (Figure S2), the efficacy remained at 30%. The use of PMS- and H₂O₂-based processes at the unadjusted pH resulted in comparable efficacies of 85% and 89%, respectively. Moreover, identical target compound removal of 32% was achieved under alkaline conditions. The analogous efficacies of these two peroxo compounds could be due to the presence of HO^• in both activated systems. A pH value in the range of 5–7 is considered most suitable for PDS and PMS activation via UV light [49]. The superiority of SO₄^•− generation occurs at between a pH of 3–5. At higher pH values, the process efficiency could be notably affected, since SO₄^•− and HO^• react with each other (Equation (7)) and a lack of radicals occurs in the oxidation system [49]:

SO₄^•− + HO^• → H⁺ + SO₄²⁻ + 1/2O₂

(7)

For instance, Arman et al. [54] also found that the efficacy of the UVC-induced PDS process applied for DXM degradation decreased in efficacy when adjusting the pH from 3 to 12.

Photo-induced oxidative systems at a pH of 3 showed not only a substantial increase in the rate of DXM degradation but also in the extent of its mineralization (

Table 2. TOC removal and pseudo first-order reaction constants of DXM degradation with the UVA/oxidant/Fe²⁺ system at a DXM/oxidant/Fe²⁺ molar ratio of 1/1/1 in UPW ([DXM]₀ = 25.5 µM).

	PDS		PMS		H2O2
pH	TOC Removal, %	kobs, min−1	TOC Removal, %	kobs, min−1	TOC Removal, %	kobs, min−1
3	38.2	0.137	27.1	0.121	35.1	0.154
unadjusted	6.3	0.023	3.7	0.014	3.1	0.018

). Thus, the addition of UVA radiation to the oxidant/Fe²⁺ systems significantly improved the TOC removal. For example, more than 38% of TOC removal was observed in the PDS-based combination, indicating a 10-fold increase in TOC removal efficiency compared to the PDS/Fe²⁺ system (4%) at the same oxidant dose. In turn, at the unadjusted pH, only the PDS-based system indicated moderate additional TOC removal (6.3%), with a noticeably lower k_obs (R² ≥ 0.97) value compared with that at the pH of 3, but an order of magnitude higher than that without UVA radiation.

In a study by Quaresma et al. [23], where the effect of photodegradation of DXM was studied, it was found that the formation of intermediate degradation products hindered the treatment efficacy due to competitive reactions. Namely, these species competed with the target compound by partially absorbing UV light in photo-oxidation or by reacting with the radical species produced by activated oxidants. The TOC content could indicate the presence of more recalcitrant intermediate degradation products formed during the photo-oxidation and remaining in the solution during the treatment.

The UVA source used in the current study emits its highest intensity at 365 nm, while the UV absorption spectrum of DXM is between 230 and 260 nm [23]. Since there was no overlapping in the wavelengths, the direct UVA-induced degradation of DXM could occur in an extended amount of time compared to UVC. However, more than 50% of the target compound was degraded (k_obs = 0.0065 min⁻¹, R² = 0.99) under UVA direct photolysis at the unadjusted pH, accompanied by negligible mineralization. Under acidic conditions, the direct UVA photolysis indicated a 40% removal of DXM (k_obs = 0.0045 min⁻¹, R² = 0.99) with invariable TOC content. Nevertheless, the photo-effect in the studied oxidant systems is important for providing an effective application.

The results obtained at a pH of 3 suggested that the DXM mineralization could be improved; thus, the oxidant molar concentration was increased by 10-fold corresponding to a DXM/oxidant/Fe²⁺ molar ratio of 1/10/1 (Figure S3).

The use of the PMS-based system indicated a superior efficacy in DXM removal; the target compound was completely decomposed within 5 min of treatment time. With the other studied combinations, DXM was effectively degraded in 15 min. In the PDS-based system, 94% degradation of DXM was observed with residual oxidants present at <93% after 120 min of treatment. The latter indicated that the used molar ratio of oxidant/Fe²⁺ could be modified, or that the stepwise addition of iron activator could be considered to enhance continuous radical generation through the treatment process.

These findings are supported by a slight reduction in TOC removal (34.5%) (Figure 5) compared to the lower PDS dose (38%). On the other hand, the increased pseudo first-order k_obs (R² ≥ 0.97), resulting in between 0.178 to 0.345 min⁻¹, respectively, indicated faster DXM removal. The synergistic effect of directly formed SO₄^•− and HO^• radicals in the PMS oxidative system surpassed the mainly monoradical nature of the UVA-induced PDS-based process in terms of k_obs, while the single HO^•-based activated H₂O₂ system promoted more complete mineralization of the target compound.

3.4. Photochemical Oxidation of DXM in SW

The feasibility of UVA-induced oxidant/Fe²⁺ processes at the most promising DXM/oxidant molar ratio of 1/10 and at a pH of 3 in terms of DXM degradation and mineralization was tested on a real SW sample (Figure 6). The chemical composition of the matrix can significantly influence the oxidation efficacy of the organic contaminants, namely, the most common interfering compounds in water such as carbonates, sulfates, chlorides, bromides and heavy metals. The sample used in the experiments had evident alkalinity followed by a higher extent of calcium, chloride, and magnesium ions than the other constituents (Table 1). Therefore, some interference from the water composition was expected.

The results indicate the inhibiting effect of the SW matrix with the prolonged treatment time for DXM degradation, with a similar degradation profile to that obtained in UPW. The most promising results were achieved by applying PMS-based photo-oxidation. Namely, 99% of the studied corticosteroid was removed by the photo-activated PMS process within 15 min. In turn, the same degradation extent of the target compound by the PDS-based combination was achieved between 60 and 90 min and with H₂O₂ in 30 min. It should be mentioned that the direct UVA photolysis of DXM in SW at a pH of 3 resulted in 39% removal (k_obs = 0.0041 min⁻¹, R² = 1.00) with the TOC value virtually unchanged over 120 min.

Therefore, the application of PMS shows a clear advantage over other tested oxidants in real matrices that simulate environmental conditions. Overall, the efficacy of UVA/oxidant/Fe²⁺ system was validated in the real water sample. In a study on DXM degradation in well water by a sunlight-driven PMS process [42], credible arguments of its beneficial use in surface water and groundwater over PDS and H₂O₂ were described. Amongst these were the anion activation and simultaneous use as a water disinfectant. The presence of halides, e.g., chloride ions, in the studied SW could contribute to the improved removal of organic contaminants through the reaction with HClO and Cl₂, formed via direct oxidation with PMS [41,42].

Bicarbonate and carbonate ions, however, have a much more noticeable impact—they, incidentally, do produce new radicals or act as scavengers via reactions with SO₄^•− and HO^•; however, the newly produced radicals are less active, e.g., CO₃^•− [6,14]. Moreover, the aforementioned ions, together with hydrogen phosphate, have been reported to activate PMS for enhanced oxidative capability, while this is not applicable with H₂O₂ or PDS [49,55]. The alkalinity of the SW matrix used in this research most likely scavenges principal radicals, SO₄^•− and HO^•, to form less reactive species like CO₃²⁻, HCO₃^•, and CO₃^•−. This could explain the degradation profile of DXM in the photo-induced PDS-based system.

In the case of carbonates, the concentration should be carefully monitored to provide increased efficacy, since an elevated amount could have detrimental effects [5,55]. For example, a thermally activated PDS process was applied to degrade 500 μg L⁻¹ of DXM in water [6]. Notably lower degradation efficiency was demonstrated in the case of bicarbonate and chloride ions at 250 mg L ⁻¹ spiked into UPW to degrade DXM. This is in accordance with typically detected concentrations in environmental water matrices. Although the target compound degradation was hindered with both separately studied ions, the addition of bicarbonates decelerated the process rate by nearly 6 times compared to the deceleration of nearly 2 times with chlorides, respectively. This supports the theory presented above; however, in the current research, the SW was a multicomponent mixture with relatively low measured quantities. Therefore, the calculated k_obs values were approximately 2–3 times lower than those obtained for the studied photo-oxidation systems in UPW (Figure 7).

The reduced efficiency of DXM degradation by the studied UVA/oxidant/Fe²⁺ combinations in the real water matrix may also be due to the natural organic matter (NOM) content in SW, since the reactions between radical species and the target compound are hampered by the concurrent oxidation processes towards NOM. The highest TOC removal, 12%, was achieved with the photo-Fenton process, followed by a noticeable decrease in efficiency with PMS-based (7%) and PDS-based (6%) photo-oxidation. This suggests that either the treatment duration or reagent doses should be optimized prior to real water treatment application.

The degradation of DXM in SW also followed a pseudo first-order reaction (R² ≥ 0.99). The calculated k_obs values for photo-oxidation with the studied peroxo compounds can be presented in the following order: PMS (0.315 min⁻¹) > H₂O₂ (0.162 min⁻¹) > PDS (0.06 min⁻¹). These observations clearly indicate that for a steroidal compound, like DXM, the HO^•-dominated oxidation resulted in higher overall treatment efficacy.

3.5. Identification of DXM Transformation Products

A schematic summary of the experimental findings and discussion presented above is given in Figure 8.

The formation of several transformation products (TPs) of DXM was verified by LC–MS analysis. Due to the higher response observed in positive mode for both DXM and TPs, the positive ESI mode was selected for the MS analysis. The measured mass (m/z) for the protonated DXM was 393 or 394. Notably, the transformation products of the target compound were detected only in the UVA-induced combinations. In UPW, the main detected peaks in the activated PDS and PMS systems were at m/z values of 158 (TP1) and 217 (TP2). In turn, m/z values of 287 (TP3) and 325 (TP4) were predominant only in the H₂O₂-based systems, while values of 426 (TP5) were predominant in the H₂O₂- and PMS-based systems. In SW, a small peak at m/z 205 (TP6) was identified in the UVA/Fe²⁺ and UVA/PDS/Fe²⁺ combinations, as well as at 265 (TP7) and 380 (TP8) in the UVA/H₂O₂/Fe²⁺ and UVA/PMS/Fe²⁺ systems, respectively. TP2, TP3 and TP6 were identified by Arman et al. [54] with corresponding m/z values of 221.2, 286.3 and 202.3 after the hydroxylation, de-fluorination, and attack on the five atomic rings of DXM by HO^• or UVC. TP5 could be a carboxylic acid derivative formed after oxidation of the hydroxyacetyl group preceded by hydroxylation of the tetrahydronaphthalenone rings of DXM [26]. TPs with m/z values similar to that of TP5, along with possible degradation pathways, were also proposed in our previous research on DXM degradation using pulsed corona discharge [56]. The decarboxylated and bi-dehydrated derivative, TP4, could result in the decarboxylation/oxidation of TP5 [26]. The oxidative degradation of the methylcyclopentanone ring was reported to lead to the formation of TP7 [6,26,28], while TP8 was most likely another carboxylic acid derivative [6,28]. The previously undetected TP1 could also be proposed as a carboxylic acid derivative, potentially formed via CO₂ loss from TP6. The formation pathways of the identified TPs suggest the presence of several sequential reactions composed of oxidation and dehydration [23].

3.6. Operational Costs of the Applied Processes

The target compound removal efficacy, in terms of its degradation and mineralization merely, might not give a comprehensive overview as to whether the processes are rational to apply. Thus, from an economic point of view, the operational costs could add greater significance. The calculated operational costs are comprised of the oxidant and ferrous activator price and the energy expense (Table S1). The average wholesale market prices in 2024 for H₂O₂, PDS and PMS were considered as EUR 0.55 per kg⁻¹, EUR 1.5 per kg⁻¹ and EUR per 2.5 kg⁻¹, respectively. The ferrous catalyst, FeSO₄·7H₂O, is available for up to EUR 3.60 per kg⁻¹. The energy use of the UVA lamp for a 120 min treatment was then converted into the cost using the average European non-household electric energy price of EUR 0.22 per kWh⁻¹ [57]. The cost calculations were made only for the UVA irradiated oxidative systems, since these indicated higher treatment efficacies. However, when sunlight is used instead of artificial UVA irradiation, there are no energy costs, contributing to a more sustainable approach.

Similar expenses of EUR 5.92–6.12 per m³ in UPW were calculated, regardless of the DXM/oxidant molar ratios of 1/1 and 1/10, as the energy cost is the main contributing factor. On the other hand, the highest target compound mineralization (70%) was achieved in the UVA-induced activated H₂O₂ system (1/10) with the cost of EUR 5.93 per m³. In SW, on the contrary, applying the same system at the same cost results in only 12% mineralization of DXM. This clearly suggests that using natural sunlight instead of artificial irradiation can reduce energy costs. At the same time, it may prolong the treatment duration but could also contribute to water disinfection. This is especially important if potable water is prepared. However, further pilot-scale studies are needed to evaluate possible scenarios, e.g., combinations with other processes, in different types of water matrices. Scaling up the process setup will increase the initial costs and maintenance costs will accrue over time. Compared to other technologies applied for DXM removal, the one initiated with electricity may still have higher costs than alternatives like chlorination or adsorption. However, the feasibility of the process depends on a combined evaluation of overall expenses and TOC removal efficiency, indicating the cost-effectiveness of electrochemical oxidation by higher organic load reduction [58].

4. Conclusions

Ferrous iron and UVA-light activated peroxo compounds were applied for the degradation and mineralization of glucocorticoid DXM in UPW. The feasibility of the most promising combinations for the photo-oxidation of DXM was also controlled in SW. The synergistic presence of activators in the oxidative process at acidic conditions resulted in ≥99% removal of the target compound accompanied by TOC removal of 27–38% by PDS-, PMS- or H₂O₂-based systems, respectively. The increase in pH reduced the treatment efficacies, while 85–94% removal of DXM was reached at the unadjusted pH, followed by ~30% efficacy at a pH of 11. The 10-fold increase in the oxidant dose at a pH of 3 significantly improved the mineralization of DXM by H₂O₂-based oxidation reaching 70% of TOC removal. The target compound degradation followed pseudo first-order reaction kinetics when an increased oxidant dose or/and when UVA irradiation were applied. Accordingly, the highest k_obs values were 0.773 min⁻¹ for the PMS-based and 0.345 min⁻¹ for the H₂O₂-based UVA/oxidant/Fe²⁺ system in UPW. The application of the UVA/oxidant/Fe²⁺ combination at an oxidant/Fe²⁺ molar ratio of 10/1 and at a pH of 3 in SW was effective in degrading 99% of DXM in 15 to 60 min, respectively. The obtained results indicated 2–3 times lower k_obs values and a reduction in TOC removal (6–12%) as compared to 0.178–0.773 min⁻¹ and 34.5–70%, respectively, in UPW, due to the interfering properties of the real matrix composition. Overall, general observations about target compound degradation in the current research indicated consistently higher efficacies for HO^•-dominated oxidation processes. The calculated average operational costs of different oxidant systems were EUR 5.93 per m³ at a DXM/oxidant molar ratio of 1/1 and EUR 6.02 per m³ of 1/10, respectively. A more sustainable approach of using sunlight instead of an artificial UVA source could reduce treatment costs while also providing a disinfection step. The results obtained from this paper could be used in the application of UVA-induced peroxo compounds for the purification of drinking or surface water contaminated with glucocorticoids.