Comparison of the degradation of molecular and ionic ibuprofen in a UV/H₂O₂ system

Pharmaceutical and personal care products (PPCPs) are commonly used chemicals, which can be easily found in various water matrices, such as wastewater effluent, seawater and surface water (Kim et al. 2007; Yu et al. 2009; Yuan et al. 2009). PPCPs are usually present in the environment at the order of ng L⁻¹ and μg L⁻¹ (Kim et al. 2007; Yu et al. 2009). The impact on both the human body and ecological environments is unclear under the low residual PPCP concentrations in surface water. Many toxicological studies have shown that fish and other aquatic organisms show reproductive disorders and behavioral changes when exposed to PPCPs (Christensen et al. 2006; Flippin et al. 2007). Ibuprofen (IBU), a common anti-inflammatory analgesic, is chosen as the target recalcitrant PPCP due to its frequent occurrence in surface water and potential long-term adverse effects. The detection of IBU in the surface water has been reported in the concentration range of 0.087 to 5 μg L⁻¹ (Dębska et al. 2004; Kim et al. 2007). Flippin et al. (2007) found that there was a delay in the spawning time and breeding period of medaka (Oryzias latipes) that were exposed to aqueous solutions containing 0–100 μg L⁻¹ IBU for 6 weeks.

However, current wastewater treatment plants are not specially designed to remove PPCPs (Kim et al. 2007; Vieno et al. 2007). Kim et al. (2007) investigated the removal of IBU and other drugs at a pilot scale. The results showed that the traditional water treatment technologies, such as coagulation, precipitation and filtration, cannot effectively remove these compounds. Vieno et al. (2007) reported that rapid sand filtration following coagulation and sedimentation only eliminated an additional 10% of the studied PPCPs, including ibuprofen. More than 95% of ibuprofen was found to be removed in the aeration tank, with aerobic biodegradation being the dominant mechanism (Smook et al. 2008). But the biological removal of PPCPs by secondary treatment process is unreliable, ranging from no removal to complete removal (Yu et al. 2009). Therefore, it is necessary to develop more efficient and stable water treatment methods to remove such compounds.

The ^•OH (E₀ = 1.89–2.72 V vs. NHE) is a strong oxidizing agent, which has a high oxidation-reduction potential (Vogna et al. 2004). ^•OH based advanced oxidation technologies (AOTs), including UV/H₂O₂, Fenton, ozonation, and sonication, can oxidize target contaminants by attacking electron-rich sites on molecules (Crittenden et al. 2005). They can effectively degrade PPCPs under operating conditions of normal temperature and pressure (Vogna et al. 2004). Laboratory studies have demonstrated that ^•OH based AOTs have a high ability to degrade PPCPs in deionized water, surface water, saltwater and wastewater (Grebel et al. 2010; Guo et al. 2013; Kwon et al. 2015; Yang et al. 2016). However, there are still some problems that need to be studied before large-scale application, such as the effect of target compound species and matrix components (i.e. natural organic matter (NOM), inorganic anions).

In recent years, studies have reported the degradation of IBU in UV and UV/H₂O₂ systems (Szabó 2010; Chianese et al. 2016; Peng et al. 2017). The IBU degradation rate is related to initial IBU concentration, initial H₂O₂ concentration, light intensity, effective optical path, solution pH, IBU properties (i.e. IBU species, absorption characteristics, secondary reaction rate constant with ^•OH), and matrix components (i.e. NOM, inorganic anions) in the UV/H₂O₂ system (Crittenden et al. 2005). Due to the different structures and properties of IBU species (molecular and ionic forms), Chianese et al. (2016) found that ionic IBU was degraded more rapidly than the molecular form under light conditions. However, the absorption characteristics of IBU species were not determined. Kwon et al. (2015) found that the degradation of IBU was accelerated in UV/H₂O₂ system in neutral conditions, but the comparison of IBU degradation between the ionic form and the molecular form was not studied. Moreover, previous studies have also identified that the H₂O₂ dosage and matrix components (i.e. NOM, inorganic anions) have impacts on PPCPs degradation (Guo et al. 2013; Xiao et al. 2015). However, studies of the correlation impacts between IBU and H₂O₂ dosage and matrix components are studied.

Hence, the degradation kinetics of ionic and molecular IBU were thoroughly investigated under two different treatment methods (i.e. UV and UV/H₂O₂). The molar absorptivity and quantum yield of IBU species at the wavelength of 254 nm were measured. The second-order reaction rate constants of IBU species with ^•OH were determined by using competitive kinetics. The mechanism of IBU degradation in the UV/H₂O₂ system was studied by establishing a pseudo first-order reaction kinetics model based on a steady-state approach. Then the effects of H₂O₂ dosage and matrix components (such as NOM, inorganic anions) on the photodegradation kinetics of IBU were further evaluated.

Ibuprofen (IBU, 99%), 4-chlorobenzoic acid (pCBA, 99%), disodium phosphate (99%), sodium dihydrogen phosphate (99%), and humic acid (technical) were purchased from Sigma Aldrich. Sodium chloride (guaranteed reagent), sodium sulfate (guaranteed reagent), sodium nitrate (guaranteed reagent), hydrogen peroxide (30% by weight), sulfuric acid (guaranteed reagent), potassium permanganate (analytical grade), sodium sulfate (analytical grade) and sodium oxalate (analytical grade) were purchased from Sinopharm Chemical Reagent, China.

IBU and pCBA stock solutions were prepared in deionized water and stored at 4 °C in the dark. For the kinetic studies, the initial concentrations of IBU and pCBA in the working solutions were 10 μM (2.06 mg L⁻¹ and 1.57 mg L⁻¹, respectively). The solution pH was adjusted to pH = 7.55 or pH = 3.00 with 10 mM phosphate buffer solution. The selection of pH = 7.55 is based on two factors: (1) pH = 7.55 is environmentally relevant and (2) IBU is a weak acid with pK_a = 4.9, thus the majority (i.e. 99.8%) of IBU is in its ion form at pH = 7.55.

The photochemical reactor is a cylindrical glass tube with an effective volume of 450 mL, with two casings. The schematic diagram of the photochemical reactor was shown in our previous study (Yang et al. 2017). The outer casing is connected to the water bath circulator (SC150–A25B, Thermo Fisher Scientific) and maintains a constant system temperature (20 ± 1 °C). A low-pressure UV lamp (GPH212T5 L/4, 10 W, Heraeus) is placed in the quartz cold trap in the center of the inner casing, and a small stirring bar was placed at the bottom of the reactor to ensure homogeneous UV exposure. The water bath circulator and the ultraviolet lamp were pre-opened to stabilize the system temperature and light intensity. One millilitre of solution was taken from the reactor at scheduled times using a 2.5 mL glass syringe (Gastight 1001, Hamilton) and chemical analysis in the solution was immediately carried out by ultra-performance liquid chromatography (Waters ACQUITY H–Class). All kinetic experiments were replicated independently at least three times.

The average light intensity per volume (I₀) in the UV reactor was estimated to be 6.16 × 10⁻⁶ Einstein L⁻¹ s⁻¹ (which corresponds to a power output of 3.83 mW cm⁻²) with potassium ferrioxalate chemical actinometry (Parker 1953) and the effective optical path length (b) was 1.32 cm measured by H₂O₂ chemical actinometry (Beltrán et al. 1995) in this study.

The emission wavelength and intensity of the UV lamp were determined using fiber optic spectrometers (USB 2000+, Ocean Optics) and the absorption spectra of IBU and pCBA in solution were measured by a UV–1800 spectrometer (Shimadzu, Japan). The pH of the solutions was measured by a S220 pH meter (Mettler Toledo). The concentration of H₂O₂ was measured by the KMnO₄ titration method (Klassen et al. 1994).

Analysis of IBU and pCBA was performed using ultra-performance liquid chromatography (Waters ACQUITY H–Class) with a BEH C18 column (1.7 μm, 2.1 mm × 50 mm, Waters). The column temperature was set at 35 °C. An isocratic mobile phase of 50% acetonitrile and 50% phosphate buffer (20 mM, pH = 3.00) was used for the quantification of IBU and 30% acetonitrile and 70% phosphate buffer (20 mM, pH = 3.00) was used for the quantification of pCBA. The sample injection volume was 5 μL and a flow rate was 0.3 mL min⁻¹. The UV detection wavelengths of IBU and pCBA were set at 220 nm and 238 nm, respectively.

The degradation of IBU in the UV/H₂O₂ system at pH = 7.55 fitted the first-order kinetic models with a higher R² (0.994) than zero-order (R² = 0.826) and second-order kinetics (R² = 0.708), indicating that IBU degradation follows first-order kinetics. The same result was obtained under direct UVs and at pH = 3.00 in the UV/H₂O₂ system. Under UV intensity of 3.83 mW cm⁻² and initial IBU concentration of 10 μM (2.06 mg L⁻¹), the initial direct photolysis degradation rates of IBU were 4.25 × 10⁻³ μM min⁻¹ (pH = 3.00) and 28.80 × 10⁻³ μM min⁻¹ (pH = 7.55), respectively. As shown in Figure 1, IBU absorbed light with a relatively low molar absorption coefficient at the wavelengths of 200–230 nm, thus it is expected that direct photolysis of IBU by UV light at 254 nm is low.

Figure 1 illustrates the decadic molar extinction coefficient for IBU with reference to Hg lamp emission spectra. At the wavelength of 254 nm, the ɛ values of IBU were 248.41 M⁻¹ cm⁻¹ (pH = 3.00) and 283.64 M⁻¹ cm⁻¹ (pH = 7.55), which were relatively low compared to the values of other PPCPs (Yang et al. 2016). The value of at 254 nm determined by Kwon et al. (2015) was 256 M⁻¹ cm⁻¹ (pH = 7.00), which is between the two values we measured. The molar extinction coefficient of IBU was positively correlated with the pH value, indicating that the was dependent on the species of IBU.

Due to both and at pH = 3.00 being lower than at pH = 7.55, the direct photolysis degradation rates of molecular IBU is slower than ionic IBU. As shown in Figure 2, the degradation kinetics of IBU was significantly enhanced by adding 625 μM (21.25 mg L⁻¹) H₂O₂ compared to direct photolysis. In the UV/H₂O₂ system, the initial degradation rate of ionic IBU at pH = 3.00 (10.62 μM min⁻¹) was also slightly faster than that of the molecular form at pH = 7.55 (8.61 μM min⁻¹). The dark reaction experiments showed that IBU did not degrade without UV irradiation, and the same results were obtained by adding 625 μM (21.25 mg L⁻¹) H₂O₂ to the dark reaction experiments. These results indicate that degradation of IBU in UV/H₂O₂ systems includes direct photolysis and radical degradation, but radical degradation plays a major role (more than 95%). The enhanced degradation of IBU with the addition of H₂O₂ is due mainly to ^•OH radical mediated oxidation. The results are consistent with the results reported by Xiao et al. (2015), and they also demonstrated that ^•OH plays a major role in the degradation of iodinated trihalomethanes by UV/H₂O₂. Therefore, the degradation rate constants of IBU largely depend on the formation of ^•OH in the UV system.

The degradation of target compounds in UV-based processes can be predicted and explained by the method of the steady-state approximation for the kinetic description of radicals (Yuan et al. 2009). The method is established based on the assumption that the radicals (i.e. ^•OH) produced by UV photolysis of H₂O₂ play a major role on the degradation of target compounds (Yuan et al. 2009). The reactions in the UV/H₂O₂ system and their rate constants are presented in Table 1.

The averages of and were 2.24 × 10⁻⁷ M s⁻¹ and 1.42 × 10⁻² s⁻¹ at pH = 3.00, then 2.24 × 10⁻⁷ M s⁻¹ and 1.77 × 10⁻² s⁻¹ at pH = 7.55, respectively.

The of molecular and ionic IBU with ^•OH were calculated to be 3.47(±0.11) × 10⁹ M⁻¹ s⁻¹ (pH = 3.00) and 5.89(±0.19) × 10⁹ M⁻¹ s⁻¹ (pH = 7.55), which were consistent with the values determined by competitive kinetics. The average of were calculated to be 4.06 × 10⁻¹² M and 2.93 × 10⁻¹² M at pH = 3.00 and pH = 7.55, respectively. Kwon et al. (2015) has determined the to be 0.27 × 10⁻¹² M in the UV/H₂O₂ system ([H₂O₂]₀ = 0.5 mM, I₀ = 0.5 mW cm⁻², b = 0.79 cm, [IBU]₀ = 10 μM, pH = 7.00), which was lower than the value we have measured (4.06 × 10⁻¹² M at pH = 3.00 and 2.93 × 10⁻¹² M at pH = 7.55). This is mainly due to both I₀ and b in the research of Kwon et al. (2015) are lower than that in this study (I₀ = 3.83 mW cm⁻², b = 1.32 cm).

Thus, under the same initial conditions, the secondary reaction rate constants of molecular and ionic IBU with ^•OH and the steady-state concentration of ^•OH determine the IBU degradation rate in the UV/H₂O₂ system. These also explain why the IBU degradation rate at pH = 7.55 is faster than at pH = 3.00 in UV/H₂O₂ system.

The concentration of H₂O₂ in the UV/H₂O₂ system affected the degradation rates of IBU. As shown in Figure 4, the experimental results of the pseudo first-order rate constants for IBU degradation increased from 2.25 × 10⁻³ s⁻¹ to 10.24 × 10⁻³ s⁻¹ when the H₂O₂ concentration increased from 45 to 375 μM (1.53 mg L⁻¹–12.75 mg L⁻¹) at pH = 7.55. This result could be well predicted by the kinetic model. In Figure 4, the was consistent with the modeling results ⁠. The calculated contributions of direct photolysis and ^•OH to IBU degradation are also shown in Figure 4. The ^•OH was the main reactive species in the UV/H₂O₂ system, whose contribution to IBU degradation was always greater than 97% with the increase of H₂O₂ concentration from 45 μM to 375 μM (from 1.53 mg L⁻¹ to 12.75 mg L⁻¹) (i.e. the contribution of direct photolysis was less than 3%). The calculated contributions of direct photolysis and ^•OH were 2.67% and 97.33% at the initial concentration of H₂O₂ (625 μM) at pH = 7.55, which were consistent with experimental data (Equation (17)). Guo et al. (2013) found that the apparent rate constant (k_app) of ciprofloxacin (CIP) increased from 0.39 × 10⁻³ s⁻¹ to 3.72 × 10⁻³ s⁻¹ when H₂O₂ concentrations were in the range of 0–5 mM. The modeling result also showed that a negative effect would be found as the concentration of H₂O₂ further increased, but only when the concentration of H₂O₂ was above 10 mM, when the decreasing trend of occurred (the inset in Figure 4). The possible explanation is that ^•OH could be scavenged by the excess H₂O₂ with the second-order rate constant of 1.4 × 10⁹ M⁻¹ s⁻¹ (Buxton et al. 1988).

Since humic acid is the main component of organic matter, the effected of organic matter on the degradation of IBU in the UV/H₂O₂ system at pH = 7.55 was studied by adding different concentrations of humic acid (0–2.87 mgC L⁻¹). Figure 5 shows that decreased from 3.55 × 10⁻³ s⁻¹ to 1.79 × 10⁻³ s⁻¹ with humic acid concentration increasing from 0 mgC L⁻¹ to 2.87 mgC L⁻¹. Moreover, of the model calculation slowed down slightly faster than that of the experimental data with the increasing humic acid concentration. These indicated that humic acid not only acted as a radical scavenger and the inner filter in the UV/H₂O₂ system, but also reacted with hydroxyl radicals to form secondary radicals that can continue to react with IBU. Then, the kinetic model was used to estimate the relative contributions of the inner filter effect (Equation (8) can be modified to A = b(ɛ_IBU[IBU] + [H₂O₂] + ɛ_NOM [humic acid]), the ɛ_NOM of humic acid being 0.10 L mgC⁻¹ cm⁻¹ as measured in this work) and radical scavenger to the decrease of k_obs. As shown in Figure 5, k_cal,obs changed slightly as compared to the experimental result (magenta line), if the inner filter effect of humic acid was ignored (i.e. assuming that ɛ of humic acid was zero). While the calculated values of k_cal,obs greatly deviated from the experimental result (green line), if the radical scavenging effect of humic acid was left out (i.e. assuming that the second-order rate constant of reactions between humic acid and ^•OH (⁠ in Table 1) in Equation (12) were zero). These contrasting results indicate that the radical scavenger effect of humic acids has a greater significant role than its inner filter effect in decreasing IBU degradation rate in the UV/H₂O₂ process. Even in the presence of humic acid, the calculated contributions of ^•OH was still clearly higher than the direct photolysis (Figure 5). This phenomenon suggests that ^•OH is still the most important candidate regarding IBU degradation.

As reported in a previous study, is the predominant carbonate species in neutral conditions (Kwon et al. 2015). As shown in Figure 6, the presence of has an inhibitory effect on IBU degradation. The k_obs of IBU decreased significantly when was present. When the initial concentration of was increased to 5 mM, the k_obs of IBU was further decreased by 25%. The inhibitory effect of on other PPCPs degradation in the UV/H₂O₂ system has also been reported in other literature (Grebel et al. 2010; Yang et al. 2014; Kwon et al. 2015). This could be mainly attributed to the scavenging effects of (⁠ = 8.5 × 10⁶ M⁻¹ s⁻¹) (Buxton et al. 1988; Grebel et al. 2010).

In the UV/H₂O₂ process, although the PPCPs can easily be degraded in most cases, the complete mineralization of PPCPs may be an uneconomical goal for treatment (Yuan et al. 2009). The cost limits the wide application of the process in practical engineering at full scale (Yu et al. 2009). Hence, two aspects should be further addressed in future studies, including the full-scale operation cost and the toxicity of byproducts.

The degradation rate of IBU was significantly increased after the addition of H₂O₂ in the UV reactor, indicating the major primary role of ^•OH. Due to the higher molar absorption coefficient and quantum yield, direct UV photolysis of ionic IBU is faster than of the molecular form. In the UV/H₂O₂ system, the degradation rate of ionic IBU is also slightly faster than that of molecular IBU, and the main determinants are and ⁠. The effect of matrix components on IBU degradation in the UV/H₂O₂ system was predicted and simulated by developing a steady-state approximation and kinetic parameters method. The k values and effect of matrix components are helpful for predicting and explaining IBU degradation mechanisms and selecting better performing treatment processes for the removal of PPCPs in wastewater.

Funding from the National Nature Science Foundation of China (No. 51674305), Hunan Graduate Student Innovation program (No. CX2017B465), and the Innovation-Driven Project (2017CX010) of Central South University are gratefully acknowledged.