Removal of Fluticasone Propionate and Clobetasol Propionate by Calcium Peroxide: Synergistic Effects of Oxidation, Adsorption, and Base Catalysis

Ai Zhang,¹ Xueyan Yin,¹ Xue Shen¹ and Yanan Liu^*,1,2

¹ College of Environmental Science and Engineering, Donghua University, 2999 North Renmin Road, Shanghai 201620, China

² Shanghai institute of pollution control and ecological security, Shanghai 200092, China

*E-mail: liuyanan@dhu.edu.cn

Abstract

Performance of calcium peroxide (CaO₂) in fluticasone propionate (FP) and clobetasol propionate (CP) removal was investigated. The removal mechanisms in CaO₂ treatment were explored. Specifically, degradation intermediates of FP and CP in CaO₂ treatment were firstly identified. CaO₂ treatment can be a practical and effective technology for FP and CP removal because of its synergistic effects in oxidation, adsorption, and base catalysis. Over 99.4% and 84.4% of FP and CP could be removed in 0.5 min at CaO₂ dosage of 0.1 - 4 g L^-1 at initial pH of 7, respectively. HO• was the dominant radicals responsible for FP and CP degradation in CaO₂ oxidation while •O₂^- also accelerated the removal. The suitable increase of solution pH (< 9) by Ca(OH)₂ released from CaO₂ and the adsorption effects of CaO₂ particles also contribute to FP and CP removal. CaO₂ treatment remained effective in complex water matrix with the presence of HA and common anions except for HCO₃^-. Moreover, nine intermediates were revealed for the first time during CaO₂ treatment indicating the decrease of glucocorticoid toxicity by CaO₂ treatment.

Table of Content

It is the first paper systematically clarifying an effective advanced oxidation process removing FP and CP pollutants.

 

 

 

Keywords: Fluticasone propionate (FP); Clobetasol propionate (CP); Advanced oxidation processes (AOPs); Degradation intermediates; Synergistic effects.

---

1. Introduction

The existence of endocrine disrupting compounds (EDCs) in water has caused a widespread of concerns for the past 20 years, due to their potential ecological hazards.1 EDCs are chemicals that have negative effects on the endocrine systems of humans and wildlife. EDCs include natural estrogens, natural androgens, artificial synthetic estrogens, phytoestrogens, and other industrial compounds.2 Research on environmental EDCs mainly focuses on the chemicals that disrupt the sex hormone signaling pathways, especially estrogenic and androgenic compounds.3 In the recent years, the disruption of corticosteroid signaling pathways in wildlife and humans from environmental chemicals, such as glucocorticoids (GCs), has attracted increasing attention.4 Gcs are a kind of steroid hormones which have anti-inflammatory effects.5 They are widely used as medicals for humans and animals.6 The presence of toxicity of GCs in surface water and wastewater was first detected in the Netherlands.7 Then they were discovered frenquently in other countries, such as the United States,8 the Czech Republic,9 Switzerland,9 Australia,10 and Japan.1 Increasing amount of GCs is detected in aquatic environment in different areas as summarized in Table S1. In a study conducted in France, the GC concentration reported reached up to 23 μg L-1 in a sample of an industrial sewage treatment plant (STP).11 Exposure to GCs even at low concentration levels could cause adverse effects on fish reproduction, growth, and development.12 Therefore, occurrence and fate of GCs in the environment have been increasingly concerned.13

Although the occurrence and environmental toxicity of GCs have been confirmed, the research on GC removal is still in its infancy. Moreover, most researches on GCs selected cortisone, prednisolone, cortisol, prednisone, or dexamethasone (Dex) as target compounds because of their high concentrations in environment,1 whereas the other GCs have not been sufficiently investigated. However, Jia et al.5 reported that four synthetic GCs (clobetasol propionate (CP), triamcinolone acetonide (TA), fluocinolone acetonide (FA), and fluticasone propionate (FP)), instead of Dex, prednisolone, prednisone, cortisol, and cortisone, are the major GC activity contributors in environmental samples. Among the four synthetic GCs, FP and CP have much higher bioactivity (70 times and 38 times as high as the activity of dexamethasone, respectively) and large consumption and release (among the top 100 most prescribed drugs in the United States).5 Their molecular structures and atoms label are shown in Fig. 1. These two synthetic glucocorticoids (FP and CP) should be prioritized in future study on GC removal.

Information about the removal of FP and CP is rarely available, most of which only focus on the removal in STPs.13-15 Unfortunately, these researches have shown that traditional treatment processes

Fig. 1 Molecular structure and atoms’ label for FP (A) and CP (C); Calculated frontier electron densities of FP (B) and CP (D) (red for positive and blue for negative).

involved in STPs are not fully effective in eliminating FP and CP.10 In contrast, the concentrations of some GCs in the final effluent of STPs are even larger than those in influent since their conjugates are hydrolyzed during STP treatment.16 The incomplete removal and even increase of GCs in STPs will worsen their pollution to surface water, soil, and groundwater through effluents, irrigation, landfills, and leachate inevitably17 as shown in Fig. S1. Therefore, other advanced or more efficient technologies were required urgently to remove FP and CP pollution.

As an environmentally friendly chemical being able to release oxygen (O ) and hydrogen peroxide (H O ) at controllable rate, 2 2 2 calcium peroxide (CaO ) has been widely applied in environmental 2 remediation as summarized in Table S2.18 Compared with other peroxides which can produce H O (such as magnesium peroxide), 2 2 CaO is more economical, efficient, and safe.19,20 The annual US 2 market for CaO is a few hundred metric tons and the price is about $ 2 2.8/kg.21 The commercially available material is commonly used in pharmaceuticals and dough conditioning formulations to provide better handling dough.21 Moreover, CaO could release H O for a 2 2 2 long time and decrease H O disproportionation,18 leading to higher 2 2 micro-pollutant removal efficiencies than using liquid H O .22 In 2 2 addition, CaO particles could provide active surface sites for 2 oxidation or low steady-state levels of H O to reaction with “built- 2 2 in” activators and offer an alkaline environment conducive to certain oxidation pathways.23 However, little information is available on the application of CaO to remove GCs. Particularly, the removal 2 efficiencies and degradation mechanisms of FP and CP during oxidation have not been elucidated clearly.

Therefore, it is of great interest to examine if the CaO 2 treatment could remove FP and CP pollution effectively. There is also a major data gap regarding the removal mechanisms of FP and CP during CaO treatment. So this study targets on (1) the feasibility 2 of CaO treatment for FP and CP, (2) the removal mechanisms of 2 CaO for remove FP and CP, (3) the influence of coexisting inorganic 2 ions and HA in matrix on FP and CP removal, (4) the degradation intermediates of FP and CP during CaO oxidation. This is the first 2 study on the removal of FP and CP by CaO . The synergistic effects 2 of oxidation, adsorption, and base catalysis of CaO in pollutant 2 removal are also raised for the first time.

2. Materials and methods

2.1 Materials

The FP and CP standards were obtained from Sigma-Aldrich (St. Louis, MO, USA). Stock FP and CP solutions were prepared in acetonitrile at 4 oC in brown bottles. Analytical grade CaO (75% CaO , 2 2 25% Ca(OH) ) and HA (fulvic acid > 90%) were purchased from 2 Aladdin Reagent Co. Ltd. (Shanghai, China). Methanol, acetonitrile, water and acetic acid (high-performance liquid chromatography (HPLC) grade) were purchased from Sigma-Aldrich, USA. Tert-butyl alcohol ((CH ) OH, TBA, 99.0%), trichloromethane (CHCl , 99.0%), 3 3 3 and various inorganic salts were obtained from Shanghai Lingfeng Chemical Reagent Co. Ltd. (Shanghai, China).

2.2 Experimental procedures

2.2.1 CaO₂ performance on FP and CP removal

The ability of CaO in removing FP and CP at different dosages was 2 determined. Experiments were conducted in a 1 L, magnetically stirred, brown flasks at 25 oC. Desired stock FP or CP solution was diluted with water to the desired concentration. Specific amount of CaO was added to start the reaction. At desired time intervals, 2 samples were withdrawn, and excess Na S O was immediately 2 2 3 added to the samples to quench the reaction. Effects of CaO dosage 2 (0.05 - 4 g L-1) was investigated at the initial pH value of 7 and 0.02 mM FP or CP . Controlled experiment without CaO was carried out 2 simultaneously. As there were little hydrolytic stability data for FP and CP, hydrolysis behavior was examined in this study at different pH values (2, 4, 6, 8, 10, and 12) at 25 oC at initial FP or CP concentration of 0.02 mM, just as with the procedures described above without the addition of CaO2.

2.2.2 ROS identification and their effects

TBA (scavenge HO•) and CHCl (scavenge •O - )20 had been applied 3 2 to observe the changes of FP and CP removal efficiencies during CaO treatment at unadjusted pH, CaO dosage of 0.1 g L-1, and FP or 2 2 CP at 0.02 mM. ROS scavenger concentrations were set at 35, 173, and 347 mM. The metal elements in CaO reagent were identified 2 using an inductive coupled plasma emission spectrometer (ICP, Agilent 720ES) as described in Supporting Information (SI).

2.2.3 Effects of solution pH and CaO₂ adsorption on FP and CP removal

Effects of initial pH values from 1.8 to 11.6 adjusted by 1 M of H SO 2 4 or NaOH were investigated at 0.1 g L-1, CaO and 0.02 mM FP or CP. 2 Samples were withdrawn to test solution pH and concentrations of FP and CP as described in section 2.2.1.

The adsorption experiments were carried out as described in section 2.2.1 at CaO of 0.1 g L-1, FP or CP at 0.2 mM, and unadjusted 2 pH . After stirring for 0 and 18 min, the solutions were filtered through 0.22 μm glass fiber filters and the residuals on the filters were freezedried and investigated by scanning electron microscopy (SEM) equipped with energy dispersive spectroscopy (EDS; FEI Quanta 200).

2.2.4 Effects of solution matrix on FP and CP removal

The effects of inorganic ions (Cl-, HCO -, SO 2-, NO -, PO 3-) on FP 3 4 3 4 and CP removal were investigated at ion concentrations of 1 to 100 mM. Effects of precipitation of Ca2+ and CO 2- or SO 2- on CP 3 4 removal by adsorption were also evaluated at various Ca2+ concentrations. Effects of HA in matrix on FP and CP removal by CaO were tested at HA of 0, 0.1, 0.5, 1, and 2 g L-1 as described in 2 section 2.2.1 at unadjusted pH, CaO of 0.1 g L-1, and FP or CP at 0.02 mM. The adsorption effects of HA at 0.1, 0.5, 1, and 2 g L-1 were also evaluated.

2.2.5 Identification of byproducts

For identification of byproduct, the initial FP or CP concentration was set at 0.1 mM. After oxidized by 0.1 g L-1 CaO for 0, 0.5, 2, 4, 2 8, 10, and 15 min, 2 mL water sample was withdrawn, respectively. The samples were filtered and detected immediately.

2.3 Analytical methods

For analysis of FP and CP, 100 mL of water sample was enriched by solid phase extraction (SPE) according to the method described by Jia et al.5 An Agilent 1100 HPLC (USA) was used to analyze the GC concentrations in samples. The column in HPLC was a reversedphase C-18 column (4.6 mm × 250 mm, 5 μm, Agilent, USA). Detailed information such as mobile phase and the procedure are described in SI.

For identification of intermediates of FP and CP, an Agilent 1290 ultra-high-performance liquid chromatography (UHPLC) was employed for separation. Samples were loaded on an Agilent Zorbax Extend - C18 analytical column (2.1 mm × 100 mm, 1.8 μm). A high-resolution Agilent 6540 Quadrupole Time-of-Flight mass spectrometer (QTOF-MS) system was used in negative mode. Detailed sample preparation, instrument analysis, and method performance are provided in detail in SI.

3. Results and Discussion

3.1 Effectiveness of CaO₂ in removing FP and CP

The effects of CaO dosage on FP and CP removal efficiencies were 2 shown in Figs. 2a&b. As shown in Fig. 2a, FP was removed very fast during CaO treatment. Over 99.4 % of FP was removed in 0.5 min at 2 CaO dosage of 0.05 - 4 g L-1. For CP, over 91.3 % could be removed 2 in 18 min at CaO of 0.05 - 4 g L-1 (Fig. 2b). So CaO treatment is 2 2 effective on FP and CP removal.

Particularly, the blank tests for FP and CP removal without CaO addition showed minor loss (< 10 %). Notable hydrolysis of FP 2 and CP were observed at all investigated pH values as presented in Figs. 2c&d. After 20 h of hydrolysis at 25 oC, 22.5, 18.3, 20.8, 16.5, 14.9, and 18.5% of FP and 26.1, 25.9, 20.9, 78.4, 62.0, and 92.8% of

Fig. 2 Removal efficiencies of FP (a) and CP (b) at different CaO dosages (unadjusted pH , initial FP and CP concentration = 0.02 mM) and 2 C/Co (concentration ratio of residual GC to initial GC) of FP (c) and CP (d) during hydrolysis process at different initial pH (initial FP and CP concentration = 0.02 mM).

CP were hydrolyzed at pH 2, 4, 6, 8, 10, and 12, respectively (Figs. 2c&d). The results confirmed that FP hydrolysis was unreliable on pH while CP hydrolysis was closely dependent on pH values. An alkaline pH solution was found to be favourable for CP hydrolysis (Fig. 2d). As the hydrolysis products may convert back to the parent compounds at certain circumstances, it is still necessary to remove them by other methods.24 As the studied reaction time for CaO 2 treatment was only 18 min (0.3 h), the FP and CP removal efficiency by hydrolysis (< 10% in 18 min) could be neglected during CaO 2 treatment. The results in Figs. 2c&d explained the phenomenon of the blank tests in Figs. 2a&b. Their hydrolysis products and mechanisms would be identified and evaluated in our future study.

With raising CaO dosages, the removal efficiencies of FP and 2 CP increased at low CaO dosages and decreased at high CaO 2 2 dosages (shown in Fig. 2a). At CaO of 0.05 to 0.1 g L-1, the FP 2 removal efficiency at 14 min is proportional to CaO dosage (Fig. 2 2a). However, at CaO dosage exceeding 0.1 g L-1, the FP removal 2 efficiency at 14 min decreased (Fig. 2a). For CP, the removal efficiencies promoted significantly from 64.1 to 91.3 % in 2 min with increasing CaO from 0.05 to 1 g L-1. However, increasing CaO 2 2 over 1 g L-1 decreased CP removal efficiency from 91.3 to 88.3 % (Fig. 2b). The reason may be explained by HO• generation. With raising CaO dosage, the amount of H O increased as shown in Eq. 2 2 2 (1) and more HO• were generated to degrade FP and CP:

However, excessive CaO generates an excessive H O which could 2 2 2 inversely scavenge the reactive oxygen species (ROS) as shown in Eqs. (2) and (3):25

These chain reactions involved in CaO treatment were explored and 2 verified as the following section.

3.2 Effect of ROS on FP and CP removal during CaO oxidation

In order to verify the function of ROS for FP and CP removal during CaO treatment, TBA and chloroform were added in CaO system to 2 2 scavenge HO• and •O -, respectively. As shown in Fig. 3a, the FP and 2 CP removal were inhibited significantly by TBA. Increasing TBA concentration from 0 to 347 mM, the FP and CP removal efficiencies decreased for 3% (from 99.6% to 96.5%) and 37% (from 84.9% to 48.0%) in 2 min, respectively. Northup and Cassidy26 also obtained the similar conclusion during the removal of 2-propanol in CaO 2 system. During •O - scavenging tests, the FP and CP removal 2 efficiencies decreased 2% (from 99.6% to 97.5%) and 29% (from 84.9% to 56.0%) in 2 min with the increase of chloroform from 0 to 347 mM, respectively (shown in Fig. 3b). This result showed that both HO• and •O - contributed to FP and CP removal. As TBA has 2 more obvious inhibition effects than chloroform (Fig. 3b), most of the FP and CP were removed by the oxidation of HO• as shown in Fig. 3c.

The raw materials for producing of commercial CaO (purity less 2 than 80%) usually contain marbles. Since various mineral impurities originally are present in the marble, they may inevitably exist in CaO reagent, catalyzing ROS generation in CaO treatment. Various 2 2 metal ions were detected in CaO reagent as shown in Fig. 3c. 2 Among these ions, Mg, Fe, and Al ions were commonly detected at relatively high concentrations (Fig. 3c). Our previous study25 demonstrated that the Mg, Al, Co, Fe, and Mn ions could accelerate the EDC removal during advanced oxidation processes. These metals in CaO reagent may participate in Fenton-type reactions and 2 leading to producing of •OH radicals as shown in Eqs. (4) and (5),27

where M represents the metal ions. Based on the above results, the reacting system of CaO treatment applicable in this study is 2 summarized in Fig. 3c.

For comparison, performance of H O and Fenton oxidation on 2 2 CP removal were studied as shown in Fig. S2. Compared with individual H O treatment, Fenton oxidation achieved higher CP 2 2 removal efficiencies. At H O dosage of 0.47, 0.94, and 1.88 g L-1 2 2 during Fenton oxidation, CP removal efficiencies were 7.6, 24.8, and 34.8% at 18 min, respectively. Notably, CaO treatment achieved 2 much higher CP removal efficiencies (96.3, 95.7, and 95.5% CP removal at CaO dosage of 1, 2, and 4 g L-1, at 18 min, respectively 2 (Fig. 2b) than Fenton oxidation at equivalent H O concentrations 2 2 calculated by Eq. (1), indicating the efficiency of CaO treatment and 2 the existing of other mechanisms besides oxidation during CaO 2 treatment.

Fig. 3 Effects of TBA (a) and chloroform (b) at various concentrations on FP and CP removal during CaO oxidation at 2 2 min and proposed reaction system of CaO oxidation (c).

3.3 Effects of solution pH and CaO₂ adsorption on FP and CP removal

As shown in section 3.2, TBA and chloroform could not thoroughly inhibit FP and CP removal in CaO treatment. So other mechnisms 2 may also be involved in FP and CP removal during CaO treatment. 2 CaO could generate Ca(OH) in water (Eq. (6)), increasing the pH 2 2 solution of the adsorbed FP and CP as a solid form:

So the effects of solution pH and CaO adsorption on FP and CP 2 removal were evaluated in our study.

As shown in Figs. 4a&b, the FP and CP removal efficiencies altered significantly at different solution pH values. Adding CaO for 2 1 g L-1 increased the solution pH from 1.8, 3.7, 6.8, 8.1, 9.7, 11.6 to 1.9, 3.8, 7.9, 9.5, 10.8, and 11.9, respectively. It was clear that the highest removal efficiencies of FP and CP were both obtained at solution pH of 7.9 (Figs. 4a&b). The removal efficiencies of FP and CP were lowest at acidic pH conditions of 1.9 and 3.8 (corresponding to initial pH of 1.8 and 3.7, respectively) (Figs. 4a &b). The solution pH could affect the generation of H O in CaO 2 2 2 treatment as shown in Eq. (1).20 Arienzo21 found that adding phosphoric acid to CaO slurry sharply increased the H O release. At 2 2 2 pH values of 1.9 - 3.8, large amount of H O may release rapidly 2 2 from CaO , causing excessive disproportionation as presented in Eq. 2 (7)25 and lead to low FP and CP removal efficiencies:

However, a high solution pH over 9 is also unfavorable for CaO 2 treatment (Figs. 4a&b). At solution pH over 9, CaO was difficult to 2 dissolve to generate H O ,21 and the main reaction during CaO 2 2 2 treatment was generating O instead of H O (shown in Eq. 6),18 2 2 2 causing a decrease in FP and CP removal efficiencies. In comparison, at solution pH of 7.9, CaO could persistently release 2 H O for a long time and the disproportionation of H O could be 2 2 2 2 reduced sharply.21 So the optimum CaO dosage should be decided 2 considering the improved solution pH by released Ca(OH) and the 2 initial pH values in practical situation. The CaO treatment could be 2 the most effective at near-neutral initial pH values. The suitable increase of solution pH (< 9) by Ca(OH) is favorable for FP and CP 2 removal during CaO treatment.

With the SEM-mapping of CaO particles before and after 2 treatment, the adsorption effect of CaO particles was investigated as 2 shown in Figs. 4(c–h). C and O elements detected on CaO particles

Fig. 4 Effects of initial pH on FP (a) and CP (b) removal efficiencies and changes of solution pH (CaO at 0.1 g L-1); EDS distribution maps for 2 CaO particles (c and d), C (e and f), and O (g and h) before (c, e, and g) and after (d, f, and h) treatment; SEM images of CaO before (i) and 2 2 after (j) treatment; Proposed roles of CaO during FP and CP removal (k) .

after treatment (Figs. 4f&h) were much more than those before treatment (Figs. 4e&g). The surface morphology of CaO particles 2 during FP treatment at 0 min and 18 min were also characterized using SEM (Figs. 4i&j). As shown in Fig. 4i, the SEM images indicate porous and semi spherical – flaky shape CaO particles at 0 2 min, indicating the existence of surface area for adsorption. However, the CaO particles are agglomerated after reaction for 18 2 min as shown in Fig. 4j, which may decrease the surface area for adsorption. Zhang et al.28 have detected the pore volume and the Brunauer-Emmett-Teller (BET) specific surface area of CaO 2 particles during oxidation and found that the CaO particles could 2 adsorb organic pollutants and its adsorption effects decreased during the oxidation. So increasing surface area of CaO particles and 2 improving their dispersion capacity would be beneficial for pollutant removal by adsorption during CaO treatment. Based on the above 2 results, the efficient FP and CP removal could be attributed to the synergistic effects of oxidation, adsorption, and base catalysis as summarized in Fig. 4k.

3.4 Effect of matrix on FP and CP removal during CaO treatment

3.4.1 HA effect

As FP and CP could be adsorbed by CaO , natural organic matter in 2 water matrix may also affect FP and CP removal by adsorption during CaO treatment. HA can effectively absorb FP and CP as shown in Fig. 2 5a. With the increase of HA concentration in the absence of CaO , FP 2 and CP removal efficiencies by HA adsorption increased (Fig. 5a). However, with the addition of CaO , the removal efficiencies of FP and 2 CP increased with the increase of HA concentrations and then decreased (Figs. 5b&c). The presence of HA promoted CP removal at low HA concentrations (≤ 0.5 g L-1) and inhibited FP and CP removal at high HA concentrations (> 0.5 g L-1) during CaO treatment (Figs. 2 5b&c). Basically, HA could interact with target compounds by adsorption, catalytic reactions, and competition.29 The adsorption and competition are two opposite effects to FP and CP removal. HA can improve FP and CP removal efficiencies by adsorption. On the other hand, HA could react with HO• easily with high rate constants30 and compete with FP and CP for the ROS during CaO oxidation, causing 2 the decrease in removal efficiencies (Fig. 5d). Thus, the overall role of HA in CaO treatment depends on the balance between the two 2 opposite effects as shown in Fig. 5d. According to the results (Fig. 5), CaO is favorable for FP and CP removal from waters containing low 2 HA concentrations (≤ 0.5 g L-1).

Fig. 5 Role of HA during CaO oxidation (a: GC adsorption by HA at various HA concentrations without CaO addition; b and c: Effects of HA 2 2 on FP (b) and CP (c) removal at various HA concentrations; d: Coupling induction of HA adsorption and competing effects in CaO system for 2 GC removal.).

3.4.2 Anion effect

Effects of 5 common inorganic ions in matrix on FP and CP removal were evaluated and shown in Fig. 6. For FP, the presence of the 5 anions from 1 to 100 mM does not show any effects on its removal except for the HCO - (Figs. 6a - e). 3 Increasing HCO - concentration to 100 mM resulted in the 3 decrease of FP removal efficiency to 42.6 % within 18 min during CaO oxidation (Fig. 6b). For CP, the presence of the Cl-, 2 HCO -, and SO 2- from 1 to 100 mM significantly inhibits its 3 4 removal while NO - has no obvious effect and PO 3- has 3 4 promotion effect from 1 to 100 mM (Figs. 6f-j). The CP removal by CaCO and CaSO adsorption could be neglected (< 15 %) as 3 4 shown in Fig. S3. The possible mechanisms for the effects of these anions on pollutant removal during HO• or CaO oxidation 2 were summarized in Figs. 6k - o. Briefly, Cl-, HCO -, and SO 2- 3 4 could consume HO• and formed other radicals much less reactive than HO• radicals (as shown in Figs. 6k, 6l, and 6n), causing the decrease in FP and CP removal efficiencies. It is worth noted that PO 3- addition could accelerate CP removal during CaO treatment 4 2 (increased by 15 % at 100 mM for 2 min). Yang et al. 31 found that the existence of PO 3- in CaO system could lead to the generation of 4 2 ROS (HPO 2−), thus promoting pollutant removal. Moreover, PO 3- 5 4 could stabilize H O 31 as shown in Fig. 6o. So CaO is suitable for FP 2 2 2 and CP removal from waters containing NO - and PO 3-.

3.5 Degradation intermediates of FP and CP by CaO oxidation

Analyses of treated solutions at different reaction times by UHPLCQTOF and extraction of chemical features by Agilent Mass Profinder and Mass Profiler Professional (MPP) software based on highresolution TOF scan data revealed 5 distinct FP decomposition products and 4 distinct CP decomposition products as listed in Table 1 which have never been reported before. Their Extracted-ion mass

Fig. 6 Effects of inorganic ions (a and f: Cl-; b and g: HCO -; c and h: NO -; d and i: SO 2-; e and j: PO 3-) on FP (a - e) and CP (f - j) removal at 3 3 4 4 various concentrations and their roles (k: Cl-; l: HCO -; m: NO -; n: SO 2-; o: PO 3-) in HO• or CaO oxidation systems.

chromatographs and the mass spectrometry fragments of the LCQTOF results can be found in Figs. 7a - 7e, 8a-8d and S4.

According to the frontier orbital theory by Fukui,32 the frontier electron densities of C, O, S, F, and Cl atoms in FP and CP are listed in Table S3 and shown more visually in Figs. 1b and 1d. Seen from Table S3 and Figs. 1b and 1d, S19, O21, F34, F30, and C17 atoms in FP and Cl20, O26, O25, F29, and C17 atoms in CP have the largest frontier electron densities, indicating that these atoms are easy to be attacked by HO• radicals.

Combining the molecular formulas (Table 1) with the mass spectrometry fragments of the LC-QTOF results (Fig. S4), the possible byproducts of FP and CP during CaO treatment were 2 proposed and shown in Figs. 7f - 7j and 8e - 8h. The changes of their concentrations are also exhibited in Figs. 7k - o and 8i - 8l. As shown in Figs. 7c, 8a - 8d, concentrations of C H F O S, C H O , C H O , C H O , and C H FO increased with time at first and 23 30 3 23 32 2 25 43 3 then decreased, meaning that they can be degraded further. On the contrary, C H F O S, C H FO , C H F O S, and C H F O (shown 23 33 3 4 21 27 5 22 33 3 3 19 28 2 5 in Figs. 7a, 7b, 7d, and 7e) are stable byproducts.

Sandra et al.33 reported that decreased glucocorticoid activity could be achieved by elimination of the double bond at C1 and C6 and substitution at the C17 position. They also reported that the lipophilicity of the steroid and the metabolic resistance of GCs could be greatly increased by halogenation (fluorination/chlorination) of the B ring at the C10 and/or C7 and by adding acetal or ester groups at C17 and C16. For FP and CP, the addition and epoxidation of the unsaturated bond of C1 and C6 (Figs. 7b - 7e, 8b, and 8d), changes by oxidizing ester group at C17 (Figs. 7a, 7c - 7e, 8a - d), and the dehalogenation (defluorination) at C10 (Figs. 7b and 8a - 8c) represents the decrease of GC activity during CaO treatment.

Fig. 7 The extracted-ion mass chromatographs for the intermediates of FP (a - e) during CaO oxidation. All figures presented the relatively 2 higher intensity of each product at the corresponding reaction time. Chemical structures of byproducts of FP (f – g). The change of peak area of each byproducts of FP (k – o).

Table 1 General information on identified byproducts of FP and CP in CaO₂ treatment.

Product	Specie	m/z	Chemical formula	Measured M	Diff (ppm)	Score
FP-P1	(M-H)-	461.1985	C23H33F3O4S	462.2052	-2.03	91.46
FP-P2	(M-H)-	377.1774	C21H27FO5	378.1843	-0.91	99.19
FP-P3	(M-H)-	419.1877	C21H31F3O3S	420.1946	-1.33	93.29
FP-P4	(M-H)-	433.2032	C22H33F3O3S	434.2103	-1.02	92.80
FP-P5	(M-H)-	373.1822	C19H28F2O5	374.1905	2.44	96.45
CP-P1	(M-H)-	277.1810	C17H26O3	278.1882	-0.30	99.26
CP-P2	(M-H)-	353.2121	C23H30O3	354.2195	0.71	98.97
CP-P3	(M-H)-	339.2328	C23H32O2	340.2402	0.27	98.62
CP-P4	(M-H)-	409.3112	C25H43FO3	410.3196	3.01	85.75

4. Conclusions

This study researched the feasibility of CaO in removing FP and CP. 2 Efficient FP and CP removal were achieved due to the coupled effects of oxidation, adsorption, and alkaline condition in CaO 2 treatment. HO• radicals played the most important role for FP and CP removal by CaO while •O - also contributs. The suitable increase 2 2 of solution pH (< 9) resulting from Ca(OH) releasing and the 2 adsorption effects of CaO particles also contribute to FP and CP 2 removal.

CaO treatment remained effective in complex water matrix 2 with the presence of humic acid (HA) and common anions except for HCO -. The presence of HA could promote FP and CP removal at 3 low HA concentrations (≤ 0.5 g L-1) while PO 3- could also accelerate 4 FP and CP removal.

Nine degradation intermediates of FP and CP were identified for the first time, indicating the decrease of GC toxicity. The results show that CaO treatment could be efficient for FP and CP removal.

Fig. 8 The extracted-ion mass chromatographs for the intermediates of CP (a – d) during CaO oxidation. All figures presented the relatively 2 higher intensity of each product at the corresponding reaction time. Chemical structures of byproducts of CP (e – h). The change of peak area of each byproducts of CP (i – l)

Conflict of interest

The authors declare no competing financial interest and human conflicts.

Acknowledgments

This work was completed by the financial support of the National Natural Science Foundation of China (No. 51708096 and No. 51578122), the Fundamental Research Funds for the Central Universities (Grant no. 2232018D3-31), and the China Postdoctoral Science Foundation (Grant no. 2017M611423). All the financial supports are gratefully acknowledged.

Appendix A. Supplementary data

Supplementary data to this article can be found in the supporting information.

References:

1. G. Suzuki, K. Sato, T. Isobe, H. Takigami, A. Brouwer and K. Nakayama, Sci. Total Environ., 2015, 527, 328-334.
2. C. Huang, Z. B. Zhang, S. M. Wu, Y. B. Zhao and J. Y. Hu, Chemosphere, 2010, 80, 608-612.
3. A. Odermatt and C. Gumy, Biochem. Pharmacol., 2008, 76, 1184-1193.
4. A. Jia, S. M. Wu, K. D. Daniels and S. A. Snyder, Environ. Sci. Technol., 2016, 50, 2870-2880.
5. K. Nakayama, K. Sato, T. Shibano, T. Isobe, G. Suzuki and S. I. Kitamura, Environ. Toxicol. Chem., 2016, 35, 946-952. 
6. N. Creusot, S. Ait-Aissa, N. Tapie, P. Pardon, F. Brion, W. Sanchez, E. Thybaud, J. M. Porcher and H. Budzinski, Environ.Sci. Technol., 2014, 48, 3649-3657.
7. B. M. C. Almroth, L. M. Gunnarsson, F. Cuklev, J. Fick, E. Kristiansson and D. G. J. Larsson, Sci. Total Environ., 2015, 511, 37-46.
8. H. Chang, J. Y. Hu and B. Shao, Environ. Sci. Technol., 2007, 41, 3462-3468.
9. Z. L. Fan, S. M. Wu, H. Chang and J. Y. Hu, Environ. Sci. Technol., 2011, 45, 2725-2733.
10. T. Isobe, K. Sato, K. Joon-Woo, S. Tanabe, G. Suzuki and K. Nakayama, Environ. Sci. Pollut. Res., 2015, 22, 14127-14135.
11. J. Roberts, P. A. Bain and A. Kumar, Environ. Toxicol. Chem.,2016, 34, 2271-2281.
12. A. Piram, A. Salvador, J. Y. Gauvrit and P. Lanteri, R. Faure, Talanta., 2008, 74, 1463-1475.
13. D. Shargil, P. Fine, Z. Gerstl, I. Nitsan and D. Kurtzman, Sci. Total Environ., 2016, 541, 742-749.
14. S. Lu, X. Zhang and Y. Xue, J. Hazard. Mater., 2017, 337, 163-177.
15. Y. Li, J. Wang, A. Zhang and L. Wang, Water Res., 2015, 83, 84-93.
16. X. Zhang, X. Gu, S. Lu, Z. Miao, M. Xu, X. Fu, Z. Qiu and Q. Sui, J. Hazard. Mater., 2015, 284, 253-260.
17. A. C. Ndjou’ou and D. Cassidy, Chemosphere, 2006, 65, 1610- 1615.
18. B. Yang, J. J. Pignatello, D. Qu and B. Xing, Environ. Eng. Sci., 2016, 33, 193-199.
19. R. Prewo and J. J. Stezowski, J. Am. Chem. Soc., 1979, 101, 7657-7660.
20. A. Zhang and Y. Li, Sci. Total Environ., 2014, 493, 307-323.
21. A. Northup and D. Cassidy, J. Hazard. Mater., 2008, 152, 1164-1170.
22. I. L. Marta, Appl. Catal. Environ., 1999, 23, 89-114.
23. M. Arienzo, Chemosphere, 2000, 40, 331-337.
24. C. Q. Tan, N. Y. Gao, Y. Deng, Y. J. Zhang, M. H. Sui, J. Deng and S. Q. Zhou, J. Hazard. Mater., 2013, 260, 1008-1016.
25. Y. Qian, J. Zhou, Y. Zhang, J. Chen and X. Zhou, Sep. Purif. Technol., 2016, 166, 222-229.
26. X. Yang, J. He, Z. Sun, A. Holmgren and D. Wang, J. Environ. Sci. – China, 2016, 39, 69-76.
27. L. He, X. M. Sun, X. F. Sun, J. Gao, C. X. Zhang, H. J. Cao and S. G. Wang, Can. J. Chem. 2014, 92, 831-837.
28. W. Sandra, C. S. Leopold and R. H. Guy, Eur. J. Pharm. Biopharm., 2008, 68, 453-466.