Environ. Sci. Technol. 2005, 39, 8300-8306
Photo-Fenton Degradation of Diclofenac: Identification of Main Intermediates and Degradation Pathway
reported involving alternative degradation processes (ozonization, UV/H2O2, or photolysis), indicating that diclofenac degradation follows different pathways, depending on the treatment applied.
LEO Ä N I D A S A . P EÄ R E Z - E S T R A D A , SIXTO MALATO,† WOLFGANG GERNJAK,† ANA AGU ¨ E R A , * ,‡ E . M I C H A E L T H U R M A N , ‡ IMMA FERRER,‡ AND A M A D E O R . F E R N AÄ N D E Z - A L B A ‡ Plataforma Solar de Almerı´a, CIEMAT, Carretera Sene´s Km. 4, 04200 Tabernas, Espan ˜ a, and Hydrogeology and Analytical Chemistry Department, University of Almerı´a, Ctra. Sacramento s/n, 04120 La Can ˜ ada, Almerı´a, Spain
The presence of pharmaceutical compounds in natural water has recently been reported (1-4), and sewage treatment plants have been pointed out as the main source of their discharge into the environment, because conventional treatment fails to eliminate them (5-8). This behavior is inherent to their lipophilic and biologically persistent action, which is designed to maintain therapeutic activity until their specific physiological function in humans or animals has been completed. Among the groups of pharmaceutical compounds of greatest environmental interest are the nonsteroidal antiinflammatory drugs (NSAIDs). Concentrations of NSAIDs in the range of several hundred nanograms per liter have been found in European rivers (9-12). Diclofenac is a commonly used analgesic, antiarthritic, and antirheumatic NSAID. Although it has been proven that diclofenac is rapidly degraded by direct photolysis under normal environmental conditions (2, 13, 14), it is still one of the most frequently detected compounds in the water cycle at concentrations up to 1.2 µg L-1 (1, 3, 10, 15). This suggests a continuous significant input of diclofenac into the environment that can be explained in part by the inefficiency of typical biological treatments in removing this compound. Although the ecotoxicity of diclofenac is relatively low and acute effects rather improbable at the concentration levels present in the environment (around 1000 times lower than the effective concentrations), it has been demonstrated that in combination with other pharmaceuticals present in water samples, the toxic effect can be considerably increased, even at concentrations at which the substances alone showed either no effect at all or only a very slight one (16). On the other hand, there is evidence that prolonged exposure to environmentally relevant concentrations of diclofenac leads to impairment of the general health of fish, inducing renal lesions and alteration of the gills, at the lowest observed effect concentration (LOEC) of 5 µg/L (17). Because of the importance of this compound, a crucial need for more enhanced technologies that can reduce its presence in the environment has become evident. In this sense, advanced oxidation processes (AOPs) represent a good choice. AOPs are characterized by the production of hydroxyl radicals (•OH), which are very reactive and produce the pollutant’s mineralization in the final stages. Recent research on the removal of diclofenac from polluted and drinking water by ozonization and AOPs (O3/H2O2, H2O2/UV, γ-radiolysis, and photo-Fenton) (18-21) has reported complete degradation in all cases, although only photo-Fenton treatment achieved total mineralization (21). Photo-Fenton is one of the AOPs for which the solar technologies have been most extensively studied and developed (22-25), because its lowcost, easy-to-handle technology is well adapted to smallto-medium-scale renewable energy facilities. Sagawe et al. (22) report an extensive compilation of the reaction mechanisms involved in the photo-Fenton chemistry. In this in-depth pilot-scale study, diclofenac in aqueous media was degraded by photo-Fenton to identify the intermediate degradation products (DPs) formed and establish the reaction pathway. Gas and liquid chromatography-mass spectrometry (GC/MS and LC/MS) techniques,
†
In recent years, the presence of pharmaceuticals in the aquatic environment has been of growing interest. These new contaminants are important because many of them are not degraded under the typical biological treatments applied in the wastewater treatment plants and represent a continuous input into the environment. Thus, compounds such as diclofenac are present in surface waters in all Europe and a crucial need for more enhanced technologies that can reduce its presence in the environment has become evident. In this sense, advanced oxidation processes (AOPs) represent a good choice for the treatment of hazardous nonbiodegradable pollutants. This work deals with the solar photodegradation of diclofenac, an antiinflammatory drug, in aqueous solutions by photo-Fenton reaction. A pilotscale facility using a compound parabolic collector (CPC) reactor was used for this study. Results obtained show rapid and complete oxidation of diclofenac after 60 min, and total mineralization (disappearance of dissolved organic carbon, DOC) after 100 min of exposure to sunlight. Although diclofenac precipitates during the process at low pH, its degradation takes place in the homogeneous phase governed by a precipitation-redissolutiondegradation process. Establishment of the reaction pathway was made possible by a thorough analysis of the reaction mixture identifying the main intermediate products generated. Gas chromatography-mass spectrometry (GC/ MS) and liquid chromatography coupled with time-offlight mass spectrometry (LC/TOF-MS) were used to identify 18 intermediates, in two tentative degradation routes. The main one was based on the initial hydroxylation of the phenylacetic acid moiety in the C-4 position and subsequent formation of a quinone imine derivative that was the starting point for further multistep degradation involving hydroxylation, decarboxylation, and oxidation reactions. An alternative route was based on the transient preservation of the biphenyl amino moiety that underwent a similar oxidative process of C-N bond cleavage. The proposed degradation route differs from those previously * Corresponding author phone: +34-950015531; fax: +34950015483; e-mail:
[email protected]. † Plataforma Solar de Almerı ´a, CIEMAT. ‡ University of Almerı ´a. 8300
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 21, 2005
Introduction
10.1021/es050794n CCC: $30.25
2005 American Chemical Society Published on Web 09/21/2005
and enables their comparison with other photocatalytic experiments:
UV Vi t30W,n ) t30W,n-1 + ∆tn 30 VT
FIGURE 1. Disappearance of 50 mg/L diclofenac and DOC by photoFenton treatment (0.05 mM Fe) as a function of t30w (illumination time). (Inset) Time evolution of pH during the experiment. as a valid simple approach to exhaustive study and identification of unknowns (26-30), were employed for this. In addition to the superior identification capability of the GC/ EI-MS full scan mass spectra, LC/MS has become the preferred analytical instrument for analyzing relatively polar and nonvolatile organic molecules without derivation. Furthermore, time-of-flight analysis (TOF-MS) with electrospray ionization (ESI) has gained popularity due to its sensitivity, theoretically unlimited mass range, high mass resolution, and capability for highly accurate mass determination. This work is part of a complete study about this contaminant, which has included also evaluation of DPs generated through photolysis under sunlight (14) and photo-Fenton process parameter optimization (31).
Experimental Section Photoreactor. All experiments were performed in a compound parabolic collector (CPC) solar pilot plant with a total volume of 35 L, illuminated volume of 22 L, and irradiated collector surface of 3.08 m2 (23). The plant works in batch mode, continuously recirculating the entire volume through the reactor and maintaining turbulent flow in the absorber tubes. The temperature inside the reactor depends on season and weather but is usually between 30 and 40 °C. At the beginning of the experiments, with collectors covered, all the chemicals are added to the tank and mixed until constant concentration is achieved throughout the system. Then the cover is removed and samples are collected at predetermined times (t). Hydrogen peroxide concentration was maintained at around 200-400 mg‚L-1 by continual addition as consumed. This quantity has been considered optimal to avoid excessive hydrogen peroxide consumption, without delaying diclofenac degradation rate, according to previous results (31). At the beginning of the test, 12 mM was added. From previous results (31) we knew that around 3 mM was consumed and consequently needed to be added every 15 min to recover the initial concentration. So, hydrogen peroxide was added after the sample was taken and the consumption was evaluated by frequent analyses, as depicted in Figure 1. Solar ultraviolet radiation (UV) was measured by a global UV radiometer (Kipp & Zonen, model CUV3), mounted on a platform tilted 37° (the same angle as the CPCs), which provides data in terms of incident wattsUV per square meter and indicates the energy reaching any surface in the same position with regard to the sun. Equation 1 permits data from several days’ experiments to be combined
(1)
where tn is the experimental time for each sample, UV is the average solar ultraviolet radiation measured during ∆tn, ∆tn ) tn - tn-1, and t30W is a “normalized illumination time”. In this case, time refers to a constant solar UV power of 30 W m-2 (typical solar UV power on a perfectly sunny day around noon). Hydrolysis and photolysis experiments were performed in 3-L Pyrex beakers (UV transmissivity >80% between 320 and 400 nm and around 40% at 300 nm; internal diameter 15 cm) and covered with a Pyrex top (not airtight) to avoid sample contamination and evaporation. Beakers were kept in the dark during hydrolysis tests and were exposed to direct sunlight and continuously stirred during the photolysis test. The maximum temperature inside the beakers was 35 °C. Chemicals. Diclofenac sodium salt (C14H10Cl2NNaO2) was from Sigma-Aldrich and was used as obtained. Analyticalgrade organic solvents were used for HPLC-UV/vis. In the photo-Fenton experiments (0.05 mM iron) iron sulfate (FeSO4‚7H2O) and reagent-grade hydrogen peroxide (30%) were used. Demineralized water was used in all the experiments, obtained from the PSA (Plataforma Solar de Almerı´a) distillation plant (conductivity < 10 µS cm-1, Cl- ) 0.7-0.8 mg L-1, NO3- ) 0.5 mg L-1, organic carbon < 0.5 mg L-1). Sample Treatment. Solid-phase extraction was employed for preconcentration of samples prior to analysis. Oasis HLB (hydrophilic/lipophilic balance) was used as the sorbent, which ensures good recovery of compounds in a wide range of polarities. The cartridges were placed in a vacuum cube (provided by Supelco) and conditioned with 2 mL of methanol, 2 mL of deionized water, 2 mL of 0.1 N chlorhydric acid, and 2 mL of water. After the conditioning step, 50 mL aliquots of the water samples were loaded at a flow rate of approximately 10 mL min-1. Elution was performed with 2 × 4 mL of methanol at 1 mL min-1. The eluates obtained were concentrated to dryness by solvent evaporation with a gentle nitrogen stream and recomposed to a final volume of 1 mL in methanol. The extracts were stored in amber vials and refrigerated until chromatographic analysis to prevent further degradation. Analytical Determinations. Dissolved organic carbon (DOC) was measured after filtration by means of a TOC analyzer, model Shimadzu TOC 5050A, equipped with an ASI5000 autosampler. Diclofenac concentration was analyzed by reverse-phase liquid chromatography (flow 0.5 mL‚min-1) with UV detector in a HPLC-UV (Agilent Technologies, series 1100) with a C-18 column (LUNA 5 µm, 3 × 150 mm, from Phenomenex). Diclofenac was determined by isocratic elution with 50% aqueous buffer solution with 10 mM ammonium formiate/50% acetonitrile. Samples were prepared by dilution with acetonitrile and subsequent filtration. Ammonia concentration was determined with a Dionex DX-120 ion chromatograph equipped with a Dionex Ionpac CS12A 4 × 250 mm column. Isocratic elution was done with H2SO4 (10 mM) at a flow rate of 1.2 mL‚min-1. Anion and carboxylic acid concentrations were measured with a Dionex DX-600 ion chromatograph on a Dionex Ionpac AS11-HC 4 × 250 mm column. Flow rate was 1.5 mL‚min-1 and elution was done with NaOH gradient programs. H2O2 concentration was determined by iodometric titration and temperature was measured directly in the pilot plant. Identification of Degradation Products. GC/MS. GC/MS analyses were run on an HP 6890 series gas chromatograph (Agilent Technologies, Palo Alto, CA) interfaced to an HP 5973 selective mass detector. Separation was carried out in VOL. 39, NO. 21, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8301
an HP-5MS capillary column (5%diphenyl/95% dimethylsiloxane), 30 m, 0.25-mm i.d., and 0.25-µm-thick film. A splitsplitless injector was used under the following conditions: injection volume 10 µL, injector temperature 250 °C, initial pulse pressure 30 psi (1.5 min), split flow 50.0 mL/min, and split time 1.5 min. The helium carrier gas flow was 1 mL/ min. The oven temperature program was 1.0 min at 105 °C, 25 °C/min to 180 °C, 5 °C/min to 230 °C (1 min). Typical MSD operating conditions were optimized by the autotuning software. Electron impact (EI) mass spectra were monitored from 50 to 400 m/z. The ion source and quadrupole analyzer temperatures were set at 230 and 106 °C, respectively. By use of methane as the reagent gas, analyses were also performed in positive chemical ionization (PCI) mode. Quadrupole and ion-source temperatures were set at 120 and 250 °C, respectively. LC/TOF-MS. Liquid chromatography/electrospray-timeof-flight mass spectrometry (LC/ESI-TOF-MS), with both positive and negative ionization, were used to detect diclofenac phototransformation products. Separation was by HPLC system (vacuum degasser, autosampler, and binary pump) (Agilent Series 1100) equipped with a reverse-phase 150 mm × 4.6 mm C8 analytical column and 5-µm particle size (Zorbax Eclipse XDB-C8). Column temperature was 25 °C. A and B mobile phases were acetonitrile and water with 0.1% formic acid, respectively. A linear gradient progressed from 15% A to 100% A in 30 min, and it was maintained at 100% A for 5 min. The flow rate was 0.6 mL/min, and injection volume was 50 µL. This HPLC system was connected to an Agilent MSD time-of-flight mass spectrometer with an electrospray interface. LC/TOF-MS accurate mass spectra were recorded from 50 to 1000 m/z. The data recorded was processed with Applied Biosystems/MDS-SCIEX Analyst QS software (Frankfurt, Germany) with application-specific accurate mass additions provided by the Agilent MSD TOF software. The mass axis was calibrated by use of the mixture provided by the manufacturer in the m/z 50-3200 range. A second orthogonal sprayer was used with a reference solution as a continuous calibration using the following reference masses: 121.0509 and 922.0098 m/z (resolution 9500 ( 500 @ 922.0098 m/z).
Results and Discussion Hydrolysis and Photolysis Assays. To ensure that the results obtained during the photo-Fenton tests were consistent and not due to hydrolysis, three hydrolysis experiments were done at different pH (3, 7, and 9). No hydrolysis was detected in any case, but precipitation of diclofenac was observed at pH 3. Diclofenac is a very soluble (water solubility 50 g/L at 25 °C at pH 7), acidic pharmaceutical (pKa ) 4.15) (32) that becomes almost insoluble below pH 4. So under these acidic conditions, diclofenac precipitated and disappeared from the solution. No diclofenac was degraded because redissolution of the samples with acetonitrile yielded the initial diclofenac concentration. It is very important to take this into consideration, because during the photo-Fenton assays, pH can easily drop to