This paper reports the degradation of a solution of 0.314 mM diclofenac (DCF), while using 5–15 mM Oxone as oxidizing agent with the catalytic action of 0.05–0.2 mM Co2+. The best performance was obtained for 10 mM Oxone and 0.2 mM Co2+, achieving the total DCF abatement and 77% removal of chemical oxygen demand after 30 min. Oxidizing of sulfate () and hydroxyl (•OH) radicals was formed by the Co2+/Oxone system. Oxone was firstly oxidized to persulfate ion that was then quickly converted into the above free radicals. For Oxone contents ≥10 mM, the decay of DCF concentration followed a second-order kinetic reaction, but the apparent rate constant changed with the Co2+ concentration used. High-performance liquid chromatography (HPLC) analysis of treated solutions showed the formation of some intermediates, whereas oxalic acid was identified as the prevalent final short-linear carboxylic acid by ion-exclusion HPLC.
Recently, an increasing concentration of the so-called emerging pollutants has been detected in natural waters (Rodriguez et al. 2017). Drugs are an important group that has been found in the aquatic environment at concentrations from ng L−1 to mg L−1 (Sirés & Brillas 2012; Pal et al. 2014), due to their high recalcitrance and resistance to biodegradation. A ubiquitous drug like diclofenac (DCF) has been detected up to 4.4 μg L−1 in surface waters (Moeller et al. 2012; Bonnefille et al. 2018). DCF ([2-(2,6-dichloroanilino)phenyl]acetic acid, C14H11NCl2O2), commercialized as Na+ salt, is a non-steroidal anti-inflammatory drug that can result in human health problems, because the oral consumption of concentrations ≥100 mg can produce gastrointestinal irritations, and superior doses can lead to gastric ulcers (Lagarto et al. 2008; Gonzalez-Rey & Bebianno 2014). Moreover, it has been documented that the prolonged exposure to relatively high concentrations of DCF in the environment affects the health of fish, including their renal lesions, alterations of the gills, genotoxicity and estrogenic effects (Gonzalez-Rey & Bebianno 2014). To avoid these problems, the DCF degradation has been the subject of research by conventional water treatments, showing efficiencies close to 90%, but with long retention times (Baccar et al. 2012; Melo-Guimarães et al. 2013). Advanced oxidation processes (AOPs) have proven their effectiveness for the removal of pharmaceuticals and personal care products from the aqueous phase (Esplugas et al. 2007; Miranda-García et al. 2010; Bernabeu et al. 2011; Sirés & Brillas 2012; Prieto-Rodríguez et al. 2013). AOPs are characterized by the generation of hydroxyl radical (), which is a powerful non-selective transient oxidant species that gives rise to the degradation of the most organic pollutants (Rodríguez-Narváez et al. 2018). The most ubiquitous AOP is the Fenton's reagent, which involves the catalytic decomposition of hydrogen peroxide using ferrous ion (Fenton reaction) to generate (Yalfani et al. 2009). However, the Fenton's reagent possesses several drawbacks like the need for acid pH values for its optimum performance, or the generation of ferric hydroxide sludge that requires an additional separation process and a disposal (Bokare & Choi 2014), which limit its application.
In recent years, to avoid the disadvantages of the Fenton reaction, other oxidant agents (e.g., persulfate or peroxymonosulfate) and transition metals (e.g., cobalt or copper) have been used (Guan et al. 2013; Nie et al. 2014; Yao et al. 2015). These modifications in the Fenton reaction have been called Fenton-like reactions (Wang et al. 2016).
A Fenton-like reaction, which showed high efficiency in the removal of organic contaminants, has been the catalytic decomposition of peroxymonosulfate (Oxone, ) with different transition metals (e.g., iron or cobalt) (Wang et al. 2016). Oxone decomposition is of great significance mainly because of its capability to be catalytically activated for generating sulfate () and peroxymonosulfate radicals (), along with (Anipsitakis & Dionysiou 2003). For many organic pollutants, the oxidation efficiency of the former radical can be even superior to that of (Rodríguez-Narváez et al. 2018).
The aim of this paper is to study the degradation and mineralization of DCF by the Co2+/Oxone system, and the characterization of the radicals produced during the process by Equation (5) to understand the role of generated free radicals. As well, the decay of DCF concentration was analyzed by kinetic models, related to pseudo-first-order and second-order reactions. Oxidation products were detected by the high-performance liquid chromatography (HPLC).
MATERIALS AND METHODS
Sodium diclofenac (>99% purity) was purchased from Sigma-Aldrich. Cobalt(II) sulfate heptahydrate and Oxone® (Oxone + ½KHSO4 + ½K2SO4) were reagent grade purchased from Baker and Sigma-Aldrich, respectively, and used as received. All the solutions were prepared with distilled water. Other chemicals used for analysis were of HPLC or analytical grade purchased from Karal.
The experiments were performed with 50 mL of a 0.314 mM DCF and 0.1 M phosphate buffer solution in an Erlenmeyer flask at ambient temperature (24 ± 2 °C). An amount of the catalytic Co2+ corresponding to 0.05, 0.1 or 0.2 mM was added to the solution, followed by a quantity of the oxidizing agent Oxone corresponding to 5, 10 or 15 mM with vigorous stirring with a magnetic bar to begin the Co2+/Oxone process. The initial pH was 7. Samples of 0.5 mL were withdrawn at the initial time and reaction times of 2, 5, 8, 10, 15, 20, 25, and 30 min to determine the DCF concentration decay by HPLC. To quench the degradation process, 0.5 mL of methanol was added to the reaction mixture immediately after sampling. Trials were made in triplicate and average values are given, along with error bars corresponding to 95% confidence interval in the figures.
The DCF concentration was monitored by reversed-phase HPLC using an Agilent Technologies 1200 Series, equipped with a Phenomenex Gemini 5u C-18 (5 μm, 4.6 mm × 150 mm) column at 40 °C and connected to a UV detector set at λ = 280 nm. The mobile phase consisted of an acetonitrile/methanol/trimethylamine (3 mM, pH 6.2) mixture, which was added with a concentration gradient (Table 1) at a flow rate of 0.8 mL min−1. The peak for DCF appeared at a retention time of 11.4 min.
Generated carboxylic acids were identified and quantified by ion-exclusion HPLC using the above equipment, fitted with a Bio-Rad Aminex HPX 87H (300 mm × 7.8 mm) column at 35 °C and the UV detector selected at λ = 210 nm. The mobile phase was 4 mM H2SO4 at a flow rate of 0.6 mL min−1.
The chemical oxygen demand (COD) of the samples was directly measured according to the 5220D Standard Method (Clesceri et al. 1967). The Oxone concentration was quantified by adding 0.3 mL of concentrated titanium(IV) sulfate to 3 mL of sample and reading the spectrophotometric absorbance at λ = 408 nm (Eisenberg 1943). The content of Oxone + persulfate was measured following the methodology proposed by Liang et al. (2008). An ‘A’ solution of 0.2 g of NaHCO3 and 4 g KI was prepared in 40 mL of distilled water under stirring, remaining equilibrated after 15 min. A sample of 0.1 mL was added to 40 mL of the ‘A’ solution and stirred for 10 min, and then its spectrophotometric absorbance was measured at λ = 352 nm. For the quantification, the procedure of Comninellis (1994) using N,N-dimethyl p-nitrosoaniline (pNDA) was utilized.
RESULTS AND DISCUSSION
Degradation of DCF by the Co2+/Oxone system
Since DCF is only soluble at pH 7, a phosphate buffer was used to follow its degradation (Llinàs et al. 2007). Figure 1 depicts the normalized concentration abatement with time for 0.314 mM of DCF using different Oxone concentrations and a Co2+ concentration of 0.2 mM. A previous control assay without adding the catalyst can also be observed, showing a 25% DCF decay after 30 min of process. This slow removal is due to the expected oxidation of the amino group of DCF by Oxone (Webb & Seneviratne 1995; Eissen et al. 2011). In contrast, Figure 1 highlights a very rapid removal of the drug in the presence of the catalyst up to its total disappearance at 30 min, starting from 10 mM Oxone. The degradation rate increased gradually up to this oxidant agent concentration, but decreased when passing from 10 to 15 mM, indicating that an excess of Oxone caused an inhibition of the process. This behavior agrees with the results reported for the treatment of atrazine with Co2+/Oxone by Ji et al. (2015), who concluded that high Oxone content acted as free radical scavenger. To clarify this phenomenon, more trials with different initial contents of Co2+ and Oxone were analyzed.
Figure 2 shows the elimination of normalized DCF after the action of 10 and 15 mM Oxone with 0.1 or 0.2 mM Co2+. As can be seen, the increase in the Co2+ content always improved the decomposition of the drug with a higher reduction for 10 mM of oxidizing agent with 0.2 mM of Co2+.
The aforementioned results agree with those reported by other authors who mentioned the importance of Co2+ concentration in the Co2+/Oxone system (Ji et al. 2015, 2016; Qin et al. 2016), but disagree with the assumption of others who presupposed that at great oxidant agent concentration, the highest Co2+ content would lead to the highest percentage of degradation, conversely to the trend shown for the k1-values in Table 2. In this scenario, the kinetic model of Chen et al. (2007) confirmed the need for controlling the Co2+ concentration, when operating at low quantities of oxidizing agent to avoid the trapping of free radicals.
In Table 2, it is also possible to observe that the Co2+/Oxone ratio presented a phenomenon of inhibition of the DCF oxidation process, if this ratio is greater than 1:200; but for values between 1:50 and 1:150, there is no significant difference in DCF removal (p = 0.0001). This is in agreement with the work of Ji et al. (2016), where no important changes in tetrabromobisphenol A degradation were observed for molar ratios between 1:120 and 1:160 of Co2+/Oxone.
Figure 3 shows the COD decay determined during 30 min of treatment of 0.314 mM DCF under the best experimental conditions of 10 mM Oxone and 0.2 mM Co2+, as well as for the highest oxidizing agent content of 15 mM Oxone with 0.2 mM Co2+. In the former case, 77% COD removal was found, a value much higher than 48% obtained under the latter conditions. This corroborates that the excess of oxidizing agent allows the consumption of free radicals, which inhibits the degradation process of DCF.
The percentages of COD decay obtained for the DCF solution by the Co2+/Oxone system are consistent with those of Pérez-Estrada et al. (2005), who reported a 40.4% COD abatement after 30 min of photo-Fenton treatment of a similar drug solution. Note that such a process involves the generation of high amounts of as free radical, suggesting that in the Co2+/Oxone system, plays a significant role in the DCF removal, apart from .
Identification of oxidation products
As shown in Figure 4(a), the HPLC chromatogram of 0.314 mM DCF displayed a well-defined peak at retention time of 11.4 min. Figure 4(b) evidences that after 30 min of reaction with 10 mM Oxone and 0.2 mM Co2+, the DCF peak was significantly reduced (>98%); in agreement with the data of Figure 2. Moreover, other peaks can be observed at shorter time, which can be related to aromatic products formed from the DCF oxidation process, as well as final carboxylic acids (Pérez-Estrada et al. 2005; Michael et al. 2014; Mussa et al. 2017).
The formation of short-linear aliphatic carboxylic acids during the treatment of 0.314 mM of DCF with 10 mM Oxone and 0.2 mM Co2+ was confirmed by ion-exclusion HPLC. From this technique, only one strong adsorption peak at 6.9 min was detected, which was identified as oxalic acid. These results propose that further by-products that are more recalcitrant than carboxylic acids persist in the solution until the end of the electrolysis (Sirés & Brillas 2012). It has also been reported by Pérez-Estrada et al. (2005) how the main carboxylic acid was produced under the photo-Fenton treatment of DCF, i.e., when only acted as free radical for its oxidation, as stated above.
Figure 5 highlights a fast generation of oxalic acid during the first 5 min of treatment, then achieves a quasi-steady concentration near 12 mg L−1 for longer times, between 10 and 30 min.
This suggests that after an initial quick destruction of the drug yielding a high accumulation of the final oxalic acid, this final by-product is further generated and destroyed practically at the same rate for up to 30 min of treatment, thereby showing a quasi-constant content. These findings evidence the formation of highly persistent by-products to the attack of free radicals ( and ), even more recalcitrant than oxalic acid, which avoids total DCF mineralization. A mass balance of the final oxalic acid accumulated at 30 min (5.33 mg L−1) indicates that it only contributes 6.22% to the final solution COD of 85.63 mg L−1 (Figure 3); therefore, the rest of the COD, as already mentioned above, possibly comes from other recalcitrant products.
Oxone decomposition and second-order kinetics for DCF
Since Oxone is the oxidizing agent used to degrade DCF in the Co3+/Co2+ system, blank assays were made to determine the change of Oxone concentration under the above experimental conditions, but without drug addition. As an example, Figure 6(a) shows the normalized Oxone abatement for a 15 mM solution under the catalytic action of 0.2, 0.1 and 0.05 mM Co2+. As expected, Oxone was more rapidly removed with increasing Co2+ concentration due to the acceleration of Equations (1) and (2), and about 95% of it disappeared after 30 min of treatment with 0.2 mM Co2+. Kinetic analysis of the concentration decays showed an excellent agreement with a pseudo-first-order decomposition, since good linear ln([Oxone]0/[Oxone]) vs. time plots with R2 > 0.99 were found, as can be seen in Figure 6(b).
Once confirmed that the Oxone decomposition fits with a pseudo-first-order kinetic order, it was studied under all experimental conditions, and the results obtained are collected in Table 3. For 15 mM of oxidant agent, an increasing k-value of 0.036, 0.052 and 0.106 min−1 was found for 0.05, 0.1 and 0.2 mM Co2+, respectively. However, the rise of the rate constant was not linear, since it only grew 2.94 times for a fourfold increase of Co2+ content. This evidences an increasing loss of the catalytic power of Co2+ to oxidize Oxone, because it is more rapidly consumed to destroy the generated S2O82− ion from Equation (4).
When Oxone concentration decreased to 10 mM, the same trend of k as for 15 mM should be expected, but at 10 mM a Co2+ excess (0.2 mM) diminished the k-value in comparison with 0.1 mM of Co2+ (0.0416 vs. 0.054 min−1). The decrease of k-value with Co2+ excess is even more apparent when 5 mM of Oxone was used, because for 0.2 mM it was reduced more than 5 to 10 times, as compared with 0.05 and 0.1 mM, respectively.
The lack of k2-values in Table 2 for 5 mM of oxidizing agent is also noticeable; it was so rapidly and completely destroyed that it gave rise to inconsistent negative rate constants. This allows the conclusion that the second-order kinetic model seems valid for Oxone concentrations ≥10 mM and a molar ratio of Co2+/Oxone < 1:300. However, the limitations and discrepancies of this model indicate that the k1-values of the pseudo-first-order model describe more appropriately the behavior of DCF decay.
Persulfate and hydroxyl radical evolution
To clarify the oxidation power of the Co2+/Oxone process, the evolution of the generated persulfate ion was assessed following the spectrophotometric method of Liang et al. (2008). It was expected that this oxidant was formed from the combination of two anion radicals by Equation (3), and consequently, its evolution profiles could inform about the relative proportion of Co2+ and Co3+ that affected the rates of Equations (1)–(5), especially for the production of the oxidizing free radicals and , originating from their destruction by Equations (4) and (5), respectively. Tests were made with a 15 mM Oxone solution upon addition of 0.05, 0.1 and 0.2 mM Co2+. Surprisingly, no S2O82− was accumulated in the two former solutions, whereas decreasing concentrations of 1.809 and 0.760 mM of this ion were detected at 1 and 2 min of the latter process, respectively. The presence of accumulated persulfate under these conditions can be associated with a very rapid initial oxidation of Oxone via Equations (1)–(3) by the high 0.2 mM Co2+ content in the solution. Nevertheless, the subsequent quicker removal of persulfate by Equations (4) and (5) could account for its complete disappearance at longer time, as well as when less Oxone concentration is present in the reaction medium.
The above results evidence the quick activation of persulfate by Co2+. Several authors have also described the use of persulfate with transition metals to degrade organics (Liang & Guo 2010; Qi et al. 2014; Wu et al. 2017). In a comparative study of Liang & Guo (2010) on the treatment of naphthalene with persulfate catalyzed with Co2+ and zero-valent iron, a much faster activation of persulfate was found with the former ion.
Finally, several assays were made with pNDA solutions containing 10 and 15 mM Oxone and 0.2 or 0.1 mM Co2+ to determine the ability of Co2+/Oxone to generate . For the low pNDA concentrations (<1 mM), maximum production was found at 1 min of treatment, but its low content limited a detailed assessment of further evolution. Better results were obtained for 15 mM pNDA solution, i.e., the same concentration as for the oxidizing agent. Figure 7(a) highlights the evolution of accumulated found for these trials. For 0.2 mM Co2+, a very rapid accumulation of this radical up to 5.30 mM occurred during the first 2.5 min of treatment, whereupon its production stopped to achieve a content near 7.3 mM at 10 min. In contrast, this phenomenon was not observed for 0.1 mM Co2+, where was continuously produced up to 11.3 mM. However, when Oxone concentration was reduced at 10 mM (Figure 7(b)), a very fast content (almost 11 mM) was produced in a similar way for both Co2+ concentrations during the first 2 min of reaction.
Our results then evidence that the relative proportion of generated and depends on the Co2+ content, tested at the given Oxone concentration, the formation of the former oxidant being more favorable than the latter for 15 mM Oxone by using higher amounts of catalyst. This fact can explain the superior oxidation ability of 0.1 over 0.2 mM Co2+ to degrade DCF under these conditions (see Figure 2) because of the greater production of the more potent oxidant , associated with the higher k1- and k2-values obtained for 0.1 mM Co2+ (see Table 2). However, this phenomenon was not observed for 10 mM Oxone, where both rate constants increased progressively at greater catalytic amount (see Table 2), because was more largely produced (Figure 7(b)) than in all cases.
Note that Nie et al. (2014) suggested that at acid pH, the predominant free radical from the thermal activation of persulfate is , compared to . Our results confirm the same behavior for 15 mM Oxone and 0.2 mM Co2+ at long treatment time, whereas at the beginning of the process, is largely produced. The much smaller production of the latter radical by means of other AOPs is also noticeable (Aurioles-López et al. 2016; Huesca-Espitia et al. 2017; Luis Sánchez-Salas et al. 2017; Ramírez-Sánchez et al. 2017). For instance, Huesca-Espitia et al. (2017) reported a maximum concentration of 0.05 mM after 50 min of Fenton reaction with a Fe2+/H2O2 molar ratio of 1:30. The Co2+/Oxone system tested possessed stronger oxidation ability, since it was able to produce near 220-fold higher concentration using 15 mM Oxone and 0.1 mM Co2+ (see Figure 7).
The Co2+/Oxone system achieved the degradation of a DCF solution with an efficiency >95% in a span of 30 min, with 77% of COD removal. The excess of Oxone caused a scavenging effect over the oxidizing free radicals formed. The proportion of and depended on the Co2+ and Oxone concentrations, being more favorable the formation of the former oxidant for higher contents of both reactants. Persulfate ion, generated from Oxone oxidation, was not accumulated in the reaction medium, because it was rapidly transformed into the above free radicals. From the Oxone decay, it was found that the DCF abatement obeyed the second-order kinetics for Oxone concentrations ≥ 10 mM, but the apparent rate constant varied with the amount of Co2+ used, owing to the different relative proportion of both free radicals formed under each experimental condition. Oxalic acid, originating from the cleavage of the benzene moiety of aromatic intermediates, was identified as the final by-product by ion-exclusion HPLC.
The authors would like to acknowledge the economic support from the Universidad de Guanajuato under the Project No. 18 (Convocatoria Institucional de Apoyo a la Investigación Científica 2018). O. M. Rodriguez-Narvaez also would like to thank CONACyT for the graduate fellowship and la Universidad Autónoma de Nayarit for the scholarship as well. O. M. Rodriguez-Narvaez would also like to thank CONACyT for the graduate fellowship and the provided facilities of Mass Spectrometry Laboratory, under the Project CONACyT grant LN 294024.