water research 43 (2009) 661–668 Available at www.sciencedirect.com journal homepage: www.elsevier.com/locate/watres Decontamination industrial pharmaceutical wastewater by combining solar photo-Fenton and biological treatment C. Sirtoria,b,c, A. Zapataa, I. Ollera, W. Gernjaka,d, A. Agüerab, S. Malatoa,* a Plataforma Solar de Almerı́a (CIEMAT), Carretera Senés, km 4, 04200 Tabernas (Almerı́a), Spain Pesticide Residue Research Group, University of Almerı́a, 04120 Almerı́a, Spain c The Capes Foundation, Ministry of Education of Brazil, PO Box 365, Brası́lia-DF 70359-970, Brazil d The University of Queensland, Advanced Water Management Centre (AWMC), Qld 4072, Australia b article info abstract Article history: Characterization and treatment of a real pharmaceutical wastewater containing 775 mg Received 24 July 2008 dissolved organic carbon per liter by a solar photo-Fenton/biotreatment were studied. Received in revised form There were also many inorganic compounds present in the matrix. The most important 11 November 2008 chemical in this wastewater was nalidixic acid (45 mg/L), an antibiotic pertaining to the Accepted 12 November 2008 quinolone group. A Zahn–Wellens test demonstrated that the real bulk organic content of Published online 27 November 2008 the wastewater was biodegradable, but only after long biomass adaptation; however, the nalidixic acid concentration remained constant, showing that it cannot be biodegraded. An Keywords: alternative is chemical oxidation (photo-Fenton process) first to enhance biodegradability, Immobilized biomass reactor followed by a biological treatment (Immobilized Biomass Reactor – IBR). In this case, two Nalidixic acid studies of photo-Fenton treatment of the real wastewater were performed, one with an Photo-Fenton excess of H2O2 (kinetic study) and another with controlled H2O2 dosing (biodegradability Solar photocatalysis and toxicity studies). In the kinetic study, nalidixic acid completely disappeared after 190 min. In the other experiment with controlled H2O2, nalidixic acid degradation was complete at 66 mM of H2O2 consumed. Biodegradability and toxicity bioassays showed that photo-Fenton should be performed until total degradation of nalidixic acid before coupling a biological treatment. Analysis of the average oxidation state (AOS) demonstrated the formation of more oxidized intermediates. With this information, the photo-Fenton treatment time (190 min) and H2O2 dose (66 mM) necessary for adequate biodegradability of the wastewater could be determined. An IBR operated in batch mode was able to reduce the remaining DOC to less than 35 mg/L. Ammonium consumption and NO 3 generation demonstrated that nitrification was also attained in the IBR. Overall DOC degradation efficiency of the combined photo-Fenton and biological treatment was over 95%, of which 33% correspond to the solar photochemical process and 62% to the biological treatment. ª 2008 Elsevier Ltd. All rights reserved. 1. Introduction Industrial wastewater is often polluted by toxic or nonbiodegradable organic compounds. Special attention currently focuses on pharmaceuticals (Joss et al., 2005, 2006). Their common consumption in human and veterinary medicine generates a diverse range of residual pollutants (pharmaceuticals þ metabolites) that reach the aquatic environment through wastewater (Jones et al., 2001; Heberer, 2002). Antibiotics are of particular concern, as they can induce bacterial * Corresponding author. Tel.: þ34 950387940; fax: þ34 950365015. E-mail address: sixto.malato@psa.es (S. Malato). 0043-1354/$ – see front matter ª 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2008.11.013 662 water research 43 (2009) 661–668 resistance, even at low concentrations (Hernández et al., 2007; Pauwels and Verstraete, 2006; Purdom et al., 1994; Schwartz et al., 2003). Nalidixic acid is a synthetic antibacterial agent frequently used in the treatment of urinary tract infections involving Gram-negative organisms (Othman et al., 1988). Alternatives to the conventional activated sludge treatment are employed for nonbiodegradable or toxic industrial wastewater. Among these, chemical oxidative treatments, and especially, Advanced Oxidation Processes (AOP), are well known for their capacity for oxidizing and mineralizing almost any organic contaminant (Comninellis et al., 2008). Nevertheless, technical applications are still scarce. As the process costs may be considered the main obstacle to their commercial application, several promising cost-cutting approaches have been proposed, such as integration of AOPs as part of a treatment train. In the typical basic process design approach an AOP pretreats nonbiodegradable or toxic wastewater, and once biodegradability has been achieved, the effluent is transferred to a cheaper biological treatment. The key is to minimize residence time and reagent consumption in the more expensive AOP stage by applying an optimized coupling strategy (Scott and Ollis, 1997). Other proposed costcutting measures are the use of renewable energy sources, i.e., sunlight as the irradiation source for running the AOP. PhotoFenton has been successfully demonstrated in real wastewater containing high organic loads in complicated matrixes as a suitable treatment for this purpose (Da Hora Machado et al., 2004; Gernjak et al., 2007; Maciel et al., 2004; Moraes et al., 2004; Rodrigues de Souza et al., 2006). Nevertheless, there are very few studies that combine the information of chemical analysis, toxicity analysis and biodegradability analysis to study the viability of the combination of photo-Fenton and biological treatment on actual industrial wastewater, not only model wastewater. Some of the few available studies were conducted in our group (Malato et al., 2007; Zapata et al., 2008), but these show different results regarding coupling strategy for different wastewaters. Hence, there is still a major need for a scientific rationale on which an ‘‘a priori’’ choice of the most appropriate treatment can be based and additional case-studies like the present one are required to enhance the common knowledge database. The aim of this study is to provide a strategy for determining the best way of combining Advanced Oxidation Processes (in this case photo-Fenton) and biological treatment (immobilized biomass reactor) to achieve the mineralization and detoxification of a real pharmaceutical wastewater containing nalidixic acid. Solar de Almerı́a (PSA) distillation plant (conductivity < 10 mS/ organic carcm, Cl ¼ 0.2–0.3 mg/L, NO 3 < 0.2 mg/L, bon < 0.5 mg/L). The experiments were performed using iron sulfate heptahydrate (FeSO4$7H2O), reagent-grade hydrogen peroxide (30% w/v) and sulfuric acid for pH adjustment, all purchased from Panreac. The photo-treated solutions were neutralized by NaOH (reagent-grade, Panreac) for toxicity and biodegradability analyses, and for discharge of the phototreated sample into the bioreactor. Industrial pharmaceutical wastewater is described in detail in Section 3. 2.2. Solar photochemical treatment All solar photochemical experiments were performed in a pilot plant made up of Compound Parabolic Collectors (CPCs) designed for solar photocatalytic applications. This reactor is composed of two modules with 12 Pyrex glass tubes mounted on a fixed platform tilted 37 (local latitude). The total illuminated area is 3 m2 and the volume is 40 L, 22 L of which are irradiated volume. At the beginning of all the photo-Fenton experiments, the solutions studied were directly added to the photoreactor, and a sample was taken after 15 min of homogenization (initial concentration). Then the pH was adjusted with sulfuric acid and another sample was taken after 15 min to confirm the pH. Afterwards, iron salt was also added (FeSO4$7H2O) and homogenized well for 15 min before a sample was taken. Finally an initial dose of hydrogen peroxide was added and samples were taken to evaluate the degradation process. Photo-Fenton experiments were carried out at a pH adjusted to 2.6–2.8 (H2SO4, 2 N) and Fe2þ concentration of 20 mg/L. In the kinetic study, the initial hydrogen peroxide concentration was around 300 mg/L and was maintained between 200 and 400 mg/L during the experiments. Solar ultraviolet radiation (UV) was measured by a global UV radiometer (KIPP&ZONEN, model CUV 3) mounted on a platform tilted 37 (the same as the CPCs). With Eq. (1), combination of the data from several days’ experiments and their comparison with other photocatalytic experiments is possible (Malato et al., 2003). t30W;n ¼ t30W;n1 þ Dtn UV Vi ; 30 VT Dtn ¼ tn  tn1 (1) 2. Experimental where tn is the experimental time for each sample, UV is the average solar ultraviolet radiation measured during Dtn, and t30W is a ‘‘normalized illumination time’’. In this case, time refers to a constant solar UV power of 30 W m2 (typical solar UV power on a perfectly sunny day around noon). VT is the total volume of the water loaded in the pilot plant (40 L), Vi is the total irradiated volume (22.0 L). 2.1. Chemicals 2.3. The nalidixic acid standard was provided by Fluka (ref. code 70162, 25 g). HPLC-grade methanol was supplied by Merck (Germany). A Milli-Q ultra-pure water system from Millipore (Milford, MA, USA) was used throughout the study to obtain the HPLC-grade water used in the analyses. Formic acid (purity, 98%) was obtained from Fluka (Germany). Distilled water used in the pilot plant was supplied by the Plataforma Analytical determinations The nalidixic acid was analyzed by liquid chromatography (flow rate 0.5 mL/min) in an HPLC–UV (Agilent Technologies, series 1100) with a C-18 column (LUNA 5 mm, 3  150 mm, from Phenomenex). The isocratic method used formic acid 25 mM/ methanol 50/50, l ¼ 254 nm. Ammonium and Naþ concentration were determined with a Dionex DX-120 ion chromatograph equipped with a Dionex Ionpac CS12A 4 mm  250 mm water research 43 (2009) 661–668 column. Isocratic elution was done with H2SO4 (10 mM) at  a flow rate of 1.2 mL/min. Anion concentrations (NO 3 and Cl ) were determined with a Dionex DX-600 ion chromatograph using a Dionex Ionpac AS11-HC 4 mm  250 mm column. The gradient program was pre-run for 5 min with 20 mM NaOH, an 8-min injection of 20 mM of NaOH, and 7 min with 35 mM of NaOH, at a flow rate of 1.5 mL/min. Mineralization was followed by measuring the dissolved organic carbon (DOC) by direct injection of filtered samples into a Shimadzu-5050A TOC analyzer with an NDIR detector and calibrated with standard solutions of potassium phthalate. Chemical oxygen demand (COD) was measured with MerckÒ Spectroquant kits. Total iron concentration was monitored by colorimetric determination with 1,10-phenanthroline, following ISO 6332, using a Unicam-2 spectrophotometer. Hydrogen peroxide was analyzed by a fast, simple spectrophotometric method using ammonium metavanadate, which allows the H2O2 concentration to be determined immediately based on a red-orange peroxovanadium cation formed during the reaction of H2O2 with metavanadate, maximum absorption of which is at 450 nm. The peroxide concentrations are calculated from absorption measurements by a ratio found by Nogueira et al. (2005). 2.4. Toxicity and biodegradability assays Toxicity of the initial wastewater and selected pre-treated samples was evaluated with BiofixÒ Lumi-10, a commercial assay. The test is based on the inhibition of the luminescence emitted by the marine bacteria Vibrio fischeri. The reagent is a freeze-dried preparation of a specially selected strain of the marine bacterium V. fischeri (formerly known as Photobacterium phosphoreum, NRRL number B-11177). The drop in light emission was measured after contact periods of 30 min. Hydrogen peroxide present in the samples from photo-Fenton experiments was removed prior to toxicity analysis using catalase (2500 U/mg bovine liver; 100 mg/L) acquired from Fluka Chemie AG (Buchs, Switzerland) after adjusting the sample pH to 7. Biodegradability of the photo-Fenton pre-treated pharmaceutical wastewater at different stages was evaluated by a Zahn–Wellens test (an adaptation of the EC protocol, Directive 88/303/EEC). Activated sludge from the El Ejido wastewater treatment plant (Almerı́a-Spain), mineral nutrients and the test material as the sole carbon source were placed together in a 0.25-L glass vessel equipped with an agitator. The test was continued for 28 days at 20–25  C and under diffuse illumination (or in a dark room). Degradation was monitored by DOC determination of the filtered solution, daily or at other appropriate regular time intervals. The ratio of DOC eliminated after each interval to initial DOC is expressed as the percentage of biodegradation. Samples analyzed are considered biodegradable when the biodegradation percentage is over 70% (EPA, 1996). 2.5. Biological system The selected biological reactor is an IBR (Immobilized Biomass Reactor). The IBR consists of a 160-L flat-bottom tank filled with 90–95 L of polypropylene 15-mm Pall Ring supports 663 colonized by activated sludge from the conventional aerobic wastewater treatment plant at El Ejido (Almerı́a). The system is also provided with a 100-L conditioner tank with pH control connected to the IBR through a recirculation pump. The operation flux is 500 L/h. Dissolved oxygen, pH and temperature were automatically measured and registered. Total volume of both tanks was 150 L. Startup and adaptation of the biological reactor began with the immobilization of the sludge on the ring supports, which took 2 days. After this, the system was maintained with controlled additions of glucose and ammonium chloride, keeping the carbon/nitrogen ratio at 100/20. The next step was to adapt the sludge to high NaCl content (approximately 5 g/L). During the whole adaptation process the analytical controls used to evaluate the IBR state were the total suspended solid, DOC, pH and Nitrogen concentration (as ammonium and nitrate). 3. Results and discussion 3.1. Matrix characterization Firstly, the main parameters of the industrial pharmaceutical wastewater were evaluated (Table 1). One relevant point was the high conductivity, associated with the presence of large amounts of inorganic ions, such as chloride and sodium, found in grams per liter. The sample further contained a significant concentration of suspended solids, a DOC of around 775 mg/L and COD (chemical oxygen demand) of 3420 mg/L. The most important organic compound studied in the matrix was nalidixic acid (Fig. 1). The concentration of this compound in the wastewater was 45 mg/L, with the main DOC component being acetate. An initial biodegradability test was performed at two different dilutions (1:2 and 1:8) of the original wastewater. The 1:2 dilution was selected according to the initial DOC recommended in Zahn–Wellens standard methodology. The other dilution (1:8) was studied in order to understand the behavior of sludge in contact with small concentrations of nalidixic acid. As shown in Fig. 2, both samples were biodegradable. Table 1 – Main parameters of industrial pharmaceutical wastewater. Parameter pH Conductivity DOC COD Acetate Nalidixic acid TSS Cl PO3 4 SO2 4 Naþ Ca2þ Amount 3.98 7 mS cm1 775 mg L1 3420 mg L1 1.9 g L1 45 mg L1 0.407 g L1 2.8 g L1 0.01 g L1 0.16 g L1 2 g L1 0.02 g L1 664 water research 43 (2009) 661–668    ClOH þ Hþ /Cl þ H2 O  (3)    Cl þ Cl /Cl2 (4) The negative effect of Cl on the photo-Fenton efficiency is also related to the formation of iron(III)-chlorocomplexes, concomitantly with the inhibited formation of iron(III)hydroperoxide complexes, the reactive species in the photoFenton process. Nevertheless, Cl and Cl2 are strong oxidants   (ESHE, Cl /Cl ¼ 2.41 V; ESHE, Cl2/2Cl ¼ 2.09 V) and could also oxidize organic solutes, so that the mineralization rate would not be lowered drastically. Similarly, chloride ions also make higher consumption of hydrogen peroxide necessary for the same mineralization through by Reactions (5) and (6). De Laat and Le (2006) conclude that over 100 mM are needed to inhibit formation of iron(III)-hydroperoxide complexes and reduce the reaction rate, and that Cl and Cl2 are reactive enough to degrade a wide range of organic compounds. Therefore, the mineralization rate of nalidixic acid dissolved in saline water (5 g/L NaCl) could be expected not to be drastically lowered.  Fig. 1 – Chemical structure of nalidixic acid.  However, the concentration of nalidixic acid remained constant in both cases. Adaptation also took much longer (around 15 days) for the 1:2-diluted sample, mainly due to the higher nalidixic acid concentration. This pharmaceutical may therefore be considered recalcitrant and possibly inhibitory to sludge metabolism, though not very acutely toxic. Nalidixic acid inhibition of sludge metabolism was also confirmed during photo-Fenton tests, as discussed below.   3.2.      Cl þ H2 O2 /HO2 þ Cl þ Hþ Solar photo-Fenton treatment    (5)  Cl2 þ H2 O2 /HO2 þ 2Cl þ Hþ The first study was the photo-Fenton degradation of 30 mg/L of nalidixic acid standard in dematerialised saline water (5 g/L of NaCl). These conditions were selected to simulate the ionic strength of the real wastewater. The pH was kept at 2.6–2.8, the temperature 30–40  C and Fe2þ concentration was 20 mg/L that is considered as optimum iron concentration for solar photoreactors selected for this study (Malato et al., 2004). The initial nalidixic acid standard mineralization rate in distilled water and with 5 g/L NaCl was 0.61 mg L1 min1 and 0.38 mg L1 min1, respectively. The consumption of hydrogen peroxide at 60% mineralization was 12 mM and 15 mM, respectively. The presence of inorganic species, like Cl, affects the photo-Fenton process, as Cl acts as a hydroxyl radical scavenger. De Laat and Le (2006) explain this by considering that less reactive species, such as chlorine atoms (Cl ) and dichloride anion radicals (Cl2), are generated, by Reactions (2)–(4).      Cl þ HO /ClOH  (6) Fig. 3 shows photo-Fenton treatment of the real wastewater described in Table 1. The experimental conditions were 20 mg/L of total iron and H2O2 concentration was maintained between 200 and 400 mg/L. 90% of the initial DOC was removed in 400 min of illumination time and the total H2O2 consumed was 180 mM. Nalidixic acid had completely disappeared at 190 min with 72 mM of H2O2 consumed. In view of these considerations, it is not recommended to mineralize 90% of DOC, because of the very long time and huge amount of hydrogen peroxide needed (around 6 g/L) for such a high mineralization level. Moreover, the results shown in Fig. 2 show that the biodegradability of the wastewater was not bad, and therefore, nalidixic acid degradation is the main goal of treatment. So it is important to determine the behavior of biodegradability and toxicity of wastewater during the photoFenton treatment, mainly during nalidixic degradation and (2) 200 800 DOC H2O2 consumed 25 40 % biodegradability % biodeg. (1:2) % biodeg. (1:8) Nalidixic acid (1:2) Nalidixic acid (1:8) 40 15 10 5 20 Nalidixic acid (mg/L) 60 DOC (mg/L) 20 80 400 0 5 10 15 20 25 Time (days) Fig. 2 – Zahn–Wellens biodegradability test of the pharmaceutical wastewater: 1:2 and 1:8 dilutions. 0 30 120 20 10 80 0 0 50 100 150 200 200 Total elimination of nalidixic acid (<1 mg/L) H2O2 consumed: 70 mM 40 0 0 0 0 160 30 H2O2 consumed (mM) 600 Nalidixic acid (mg/L) 100 50 100 150 200 250 300 350 400 t30W (min) Fig. 3 – Mineralization of industrial pharmaceutical wastewater and H2O2 consumed during photo-Fenton. Inset shows degradation of nalidixic acid during the same test. 665 water research 43 (2009) 661–668 shortly afterwards to find the best treatment time with the minimum hydrogen peroxide consumption. Toxicity and biodegradability assays Toxicity and biodegradability of the real wastewater treated by photo-Fenton were evaluated at different stages of the process in order to determine the optimal point for coupling to the biological process. To do this, it was necessary to reproduce the previous experiment maintaining all the parameters except H2O2 dosing the same. H2O2 was added so samples would be representative of the photocatalytic process. H2O2 was added (a small quantity, on the order of a few mM) to the photoreactor, and consumption monitored, and after the peroxide was consumed, a sample was taken for bioassay. H2O2 was again added, and another sample was taken after all the peroxide had been consumed. This procedure of ‘‘additiontotal consumption-addition’’ was repeated until significant mineralization (70%), ensuring that no H2O2 remained (which could affect the bioassays). It also prevents any reaction in the dark after the sample is taken form the photoreactor, since analyses are not performed until the H2O2 has been completely consumed. Moreover, as nalidixic acid and DOC are also determined in these experiments, results can be compared with the kinetic results shown in Fig. 3. Another option could be to work with excess H2O2 (as the experiment shown in Fig. 3), and eliminate the remaining H2O2 before applying the bioassay. But methods for eliminating H2O2 based on catalase, MnO2, etc, could change the matrix composition and/or the added chemicals could change the response of the bioassays. Therefore, the data shown are directly related to H2O2 consumption, instead of illumination time. In this experiment, 12 samples suitable for analyzing toxicity and biodegradability were taken (Fig. 4). Total degradation of nalidixic acid was attained at 66 mM of H2O2 consumed, a consumption very similar to the experiment shown in Fig. 3, confirming that experiments performed S1 800 60 S5 700 DOC Nalidixic acid S10 500 S15 400 300 30 20 200 10 100 0 0 20 40 60 80 0 100 100 DOC Total elimination of nalidixic acid 1:3 dil. 700 80 600 60 500 400 40 300 200 20 H2O2 consumed (mM) Fig. 4 – Mineralization of real wastewater during the photoFenton process as function of hydrogen peroxide dose (addition-total consumption-addition procedure). Sn (from S1 to S17) are selected samples for toxicity and biodegradability studies. 800 100 0 0 S1 S4 S10 S11 S13 S14 S15 S16 Sample Fig. 5 – Toxicity bioassay (Vibrio fischeri) of 1:3 diluted samples and DOC during the photo-Fenton process. DOC (mg/L) DOC (mg/L) 40 Nalidixic acid (mg/L) 600 50 % inhibition Vibrio fischeri 3.3. by the ‘‘addition-total consumption-addition’’ procedure are comparable to others in which excess H2O2 is added (a certain amount of H2O2 produces a specific DOC mineralization and nalidixic acid degradation). A first group of samples was selected for the V. fischeri toxicity bioassay and COD analysis (S1, S5, S6, S7, S8, S9, S11, S12, S13, S14, S15), and a second group for Daphnia magna toxicity analysis and Zahn–Wellens test (S6, S9, S10, S12, S13, S14). V. Fischeri toxicity was evaluated at dilution of 1:3, 1:6 and undiluted. Fig. 5 shows only results with the 1:3 dilution along with DOC. All the toxicity results (diluted and undiluted) were quite similar, and it can therefore be concluded that photoFenton treatment did not decrease V. Fischeri toxicity. D. magna bioassays demonstrated similar behavior. All microcrustaceans in all samples died in 24 h. Therefore, toxicity bioassays show that photo-Fenton was unsuccessful. But both bioassays have being described as very sensitive (Hernando et al., 2007) and this result is not surprising for real wastewater containing hundreds of mg/L of different organic compounds (fro example, carboxylic acids) and their degradation intermediates. Sometimes toxicity tests (usually a quick method) can help select the stage of an AOP treatment at which the water becomes nontoxic and, presumably, biodegradable (Gutiérrez et al., 2002; Hernando et al., 2005; Lapertot and Pulgarin, 2006; Lapertot et al., 2008). In other words, toxicity tests such as V. fischeri can detect toxic response in a short time (5–30 min), should be retested afterwards for biodegradability (usually a more time-consuming method). But in view of the results shown in Fig. 5, no such information was found from the toxicity assays. Variation in the COD during the experiment was also determined along with DOC (Fig. 6). The considerable decrease in this parameter agrees with the strong oxidation of organic matter. The efficiency of the oxidative process is more clearly shown by the AOS parameter (average oxidation state), which can be calculated by Eq. (7), in which DOC and COD are expressed in moles of C/L and of O2/L, respectively, at the sampling time. AOS is between þ4 for CO2, the most oxidized state of C, and 4 for CH4, the most reduced state of C. As observed in Fig. 6, the maximum AOS of 0.3 was reached 666 water research 43 (2009) 661–668 1 1500 -1 1000 % Biodegradability AOS DOC, COD (mg/L) 0 2000 80 80 Biodegradable 60 Initial sample (1:2) Samples containing nalidixic acid S7 S10 S11 40 Samples without nalidixic acid S13 S14 S15 20 500 0 -2 0 0 10 20 30 40 50 60 70 0 80 Fig. 6 – AOS (see Eq. (7)) evolution during the photo-Fenton process. after approximately 25 mM of H2O2 consumed and remained around there until the end of the test. AOS usually increases with treatment time until almost reaching a plateau. These results suggest that more oxidized organic intermediates are formed at the beginning of the photocatalytic process, and after a certain time, the chemical nature of most of them no longer varies substantially (Sarria et al., 2002), even if the photo-Fenton treatment continues. Formation of more oxidized intermediates indirectly demonstrates that the treatment can improve biodegradability. At the moment that AOS stabilizes, the chemical treatment is only mineralizing organic contaminants, but with no partial oxidation. The changes in AOS were taken into account in selecting the treatment stage at which the water might presumably be biodegradable. Zahn–Wellens tests (usually a time-consuming method) were therefore only done on samples >S6 (H2O2 consumption 20 mM). Before this treatment stage the concentration of nalidixic acid and toxicity was rather high and AOS increased. 4ðDOC  CODÞ DOC 10 15 20 25 40 20 0 30 Time (days) H2O2 consumed (mM) AOS ¼ 5 60 % Biodegradability 2500 100 100 COD DOC AOS 3000 (7) Zahn–Wellens tests (Fig. 7) were performed on six undiluted samples at different stages during the photo-Fenton process. All samples were at least 90% biodegradable at the end of the Z–W assay. In three of them (S7, S10 and S11), in which nalidixic acid concentrations were 20.7, 8.5 and 4.6 mg/L, respectively, 70% biodegradability was attained only after 10 days of biotreatment (samples S10 and S11 after 8 days). On the other hand, samples with very low concentrations (<1 mg/L) or without nalidixic acid (S13, S14 and S15) were biodegradable after 3 days. Results show that untreated samples need much longer adaptation periods than treated samples and that nalidixic acid (at concentration as low as 4.6 mg/L) is also detrimental in the sense that they reduce biodegradation efficiency. These results demonstrated that photo-Fenton should be performed until total degradation of nalidixic acid before coupling a biological treatment and that AOS determination is an appropriate technique for selecting those samples to be tested by Z–W assay. Fig. 7 – Zahn–Wellens test for selected samples during the photo-Fenton process (initial sample is also shown). 3.4. Combined solar photo-Fenton and biological system When the best photo-Fenton wastewater pretreatment for biodegradability had been determined, the combined photoFenton/biological treatment was carried out in a pilot bioreactor. Before performing the experiment in the combined system, the IBR was inoculated with 150 L of concentrated activated sludge from the El Ejido wastewater treatment plant (Almerı́a, Spain). Then, recirculation was maintained between the conditioner tank and the IBR in order to ensure optimum fixation of the sludge on the propylene Pall Ring supports. Total suspended solids, DOC and inorganic ion concentration (mainly ammonia and nitrate) were measured daily. The system was maintained with controlled addition of glucose, a pharmaceutical (CAS number: 1953-04-4) less readily biodegradable than glucose, and ammonium chloride, keeping the carbon/nitrogen ratio at 100/20. The next step was adapting the sludge to high salinity, as the wastewater contains large quantities of NaCl (see Table 1). During the whole adaptation process, NaCl concentration was increased gradually in five cycles from 1 to 5 g/L. Startup and adaptation of the biological reactor was done while simultaneously performing several runs with the photoFenton reactor in order to accumulate enough pre-treated wastewater to add to the bioreactor, as the photo-Fenton plant only has a 40 L volume and the bioreactor (IBR þ conditioner tank) has 150 L. The different photo-Fenton runs were accumulated in a large volume tank (neutralization tank) connected to the bioreactor where the pH was adjusted roughly to 7, as automatic pH control is in the bioreactor itself. In the coupled system, wastewater containing nalidixic acid (45 mg/ L) was pre-treated by photo-Fenton until its total elimination during 190 min of illumination time with 20 mg/L of Fe2þ and 66 mM of H2O2 consumed. The chemical characterization of the photo-Fenton effluent was 530 mg/L of DOC and 6.5 g/L of NaCl. It should be remarked that the same dose of H2O2 (66 mM) did not always accomplish exactly the same mineralization, as observed in Fig. 4, where less than 530 mg/L were attained. But the scope of the treatment was the complete (<1 mg/L) degradation of nalidixic acid. The pre-treated water research 43 (2009) 661–668 600 N-NH4+, N-NO3- (mg/L) 100 DOC (mg/L) 500 400 300 N-NH4+ 80 60 Acknowledgments 40 20 0 20 40 60 80 100 120 100 0 0 20 40 60 80 100 120 Time (hours) Fig. 8 – Mineralization of the photo-Fenton pre-treated wastewater in the IBR. effluent was pumped from the neutralization tank to the conditioner tank connected to the IBR. The system was operated in batch mode, with a recirculation flow rate of 500 L/h between the conditioner tank and the IBR until the effluent was bio-mineralized (final DOC 35 mg/L). The DOC and evolution of nitrogen (as NHþ 4 and NO3) during biological treatment are shown in Fig. 8. DOC went down 495 mg/L in 4 days, a result similar to the Z–W test. NHþ 4 (NH4Cl, in a 68 mg/L concentration of N) was added the first day to enable nitrifying bacteria to metabolize the organic  carbon. The consumption of NHþ 4 and the generation of NO3 demonstrated nitrification and N assimilation through biomass grow as overall N content decreased. Overall mineralization efficiency of the combined photoFenton and biological treatment in batch mode for the degradation of the real pharmaceutical wastewater was over 95%, of which 33% corresponds to the solar photochemical process and 62% to the biological treatment. V. Fischeri and D. magna bioassays were also performed on the biotreatment effluent showing below-threshold toxicity. Therefore, the combined treatment was also successful from the viewpoint of biotoxicity. 4. accomplished by the solar photo-Fenton treatment and 62% by the biological treatment. N-NO3- 0 200 667 Conclusions  It has been demonstrated that a toxic industrial wastewater containing a biorecalcitrant compound (nalidixic acid) can be successfully treated by photo-Fenton after long treatment with heavy consumption of hydrogen peroxide, but without decreasing toxicity.  Photo-Fenton successfully enhanced the wastewater biodegradability.  Suitable selection of the photo-Fenton treatment time and hydrogen peroxide dose necessary to reach the biodegradability threshold made it possible to degrade the remaining DOC in a pilot aerobic bioreactor, and detoxify the wastewater.  The global efficiency in the combined solar photo-Fenton þ IBR system operated in batch mode was 95% of DOC elimination (initial DOC of 775 mg/L), of which 33% was The authors wish to thank the European Commission for financial support for the INNOWATECH project under the Sixth Framework Programme, within the ‘‘Global Change and Ecosystems Program’’ (Contract no: 036882) and AUSTEP (Italy) for providing the wastewater. Ana Zapata and Carla Sirtori thank the Spanish Ministry of Education and Science and the Capes Foundation – Brazil Ministry of Education, respectively, for their Ph.D. research grants. references Comninellis, C., Kapalka, A., Malato, S., Parsons, S.A., Poulios, I., Mantzavinos, D., 2008. Advanced oxidation processes for water treatment: advances and trends for R&D. Journal of Chemical Technology and Biotechnology 83, 769–776. Da Hora Machado, A.E., Xavier, T.P., de Souza, D.R., de Miranda, J. A., Duarte, E.T.F.M., Ruggiero, R., de Oliveira, L., Sattler, C., 2004. Solar photo-Fenton treatment of chipboard production wastewater. Solar Energy 77, 583–589. De Laat, J., Le, T.G., 2006. Effects of chloride ions on the iron(III)catalyzed decomposition of hydrogen peroxide and on the efficiency of the Fenton-like oxidation process. Applied Catalysis B: Environmental 66, 137–146. EPA-United States Environmental Protection Agency, April 1996. Prevention, Pesticides and Toxic Substances (7101). Fates, Transport and Transformation Test Guidelines OPPTS 835. 3200 Zahn–Wellens/EMPA Test. EPA 712-C-96–084. EPA. Gernjak, W., Krutzler, T., Malato, S., Bauer, R., 2007. Photo-Fenton treatment of olive mill wastewater applying a combined Fenton/flocculation pretreatment. Journal of Solar Energy Engineering 129, 53–59. Gutiérrez, M., Etxebarrı́a, J., de las Fuentes, L., 2002. Evaluation of wastewater toxicity: comparative study between MicrotoxÒ and activated sludge oxygen uptake inhibition. Water Research 36, 919–924. Heberer, T., 2002. Occurrence, fate, and removal of pharmaceutical residues in the aquatic environment: a review of recent research data. Toxicology Letters 131, 5–17. Hernández, F., Sancho, J.V., Ibáñez, M., Guerrero, C., 2007. Antibiotic residue determination in environmental waters by LC–MS. Trends in Analytical Chemistry 26, 466–485. Hernando, M.D., De Vettori, S., Martı́nez Bueno, M.J., FernÁndezAlba, A.R., 2007. Toxicity evaluation with Vibrio fischeri test of organic chemicals used in aquaculture. Chemosphere 68, 724–730. Hernando, M.D., Fernández-Alba, A.R., Tauler, R., Barceló, D., 2005. Toxicity assays applied to wastewater treatment. Talanta 65, 358–366. Jones, O.A.H., Voulvoulis, N., Lester, J.N., 2001. Human pharmaceuticals in the aquatic environment: a review. Environmental Technology 22, 1383–1394. Joss, A., Keller, E., Alder, A.C., Göbel, A., McArdell, C.S., Ternes, T. A., Siegrist, H., 2005. Removal of pharmaceuticals and fragrances in biological wastewater treatment. Water Research 39, 3139–3152. Joss, A., Zabczynski, S., Göbel, A., Hoffmann, B., Löffler, D., McArdell, C.S., Ternes, T.A., Thomsen, A., Siegrist, H., 2006. Biological degradation of pharmaceuticals in municipal 668 water research 43 (2009) 661–668 wastewater treatment: proposing a classification scheme. Water Research 40, 1686–1696. Lapertot, M., Pulgarin, C., 2006. Biodegradability assessment of several priority hazardous substances: choice, application and relevance regarding toxicity and bacterial activity. Chemosphere 65, 682–690. Lapertot, M., Ebrahimi, S., Oller, I., Maldonado, M.I., Gernjak, W., Malato, S., Pulgarin, C., 2008. Evaluating Microtox as a tool for biodegradability assessment of partially treated solutions of pesticides using Fe3þ and TiO2 solar photo-assisted processes. Ecotoxicology and Environmental Safety 69, 546–555. Maciel, R., Sant’Anna Jr., G.L., Dezotti, M., 2004. Phenol removal from high salinity effluents using Fenton’s reagents and photo-Fenton reactions. Chemosphere 57, 711–719. Malato, S., Blanco, J., Maldonado, M.I., Oller, I., Gernjak, W., PérezEstrada, L., 2007. Coupling solar photo-Fenton and biotreatment at industrial scale: main results of a demonstration plant. Journal of Hazardous Materials 146 (3), 440–446. Malato, S., Blanco, J., Maldonado, M.I., Fernandez-Ibañez, P., Alarcon-Padilla, D., Collares-Pereira, M., Farinha-Mendes, J., Correia de Oliveira, J., 2004. Engineering of solar photocatalytic collectors. Solar Energy 77, 513–524. Malato, S., Blanco, J., Vidal, A., Alarcón, D., Maldonado, M.I., Cáceres, J., Gernjak, W., 2003. Applied studies in solar photocatalytic detoxification: an overview. Solar Energy 75, 329–336. Moraes, J.E.F., Quina, F.H., Nascimento, C.A.O., Silva, D.N., Chiavone-Filho, O., 2004. Treatment of saline wastewater contaminated with hydrocarbons by the photo-Fenton process. Environmental Science and Technology 38, 1183–1187. Nogueira, R.F.P., Mirela, C.O., Paterlini, W.C., 2005. Simple and fast spectrophotometric determination of H2O2 in photo-Fenton reactions using metavanadate. Talanta 66, 86–91. Othman, S., Muti, H., Shaheen, O., Awidi, A., Al-Turk, W.A., 1988. Studies on the adsorption and solubility of nalidixic acid. International Journal of Pharmaceutics 41, 197–203. Pauwels, B., Verstraete, W., 2006. The treatment of hospital wastewater: an appraisal. Water Health 4, 405–416. Purdom, C.E., Hardiman, P.A., Bye, V.J., Eno, N.C., Tyler, C.R., Sumpter, J.P., 1994. Estrogenic effects of effluents from sewage treatment works. Journal of Chemical Ecology 8, 275–285. Rodrigues de Souza, D., Duarte, E.T.F.M., Girardi, G.S., Velani, V., da Hora Machado, A.E., Sattler, C., de Oliveira, L., de Miranda, J. A., 2006. Study of kinetic parameters related to the degradation of an industrial effluent using Fenton-like reactions. Journal of Photochemistry and Photobiology A: Chemistry 179, 269–275. Sarria, V., Parra, S., Adler, N., Peringer, P., Benitez, N., Pulgarin, C., 2002. Recent developments in the coupling of photoassisted and aerobic biological processes for the treatment of biorecalcitrant compounds. Catalysis Today 76, 301–315. Schwartz, T., Kohnen, W., Jansen, B., Obst, U., 2003. Detection of antibiotic-resistant bacteria and their resistance genes in wastewater, surface water, and drinking water biofilms. FEMS Microbiology Ecology 43, 325–335. Scott, J.P., Ollis, D.F., 1997. Integration of chemical and biological oxidation processes for water treatment II: recent illustrations and experiences. Journal of Advanced Oxidation Technology 2, 374–381. Zapata, A., Oller, I., Gallay, R., Pulgarı́n, C., Maldonado, M.I., Malato, S., Gernjak, W., 2008. Comparison of photo-Fenton treatment and coupled photo-Fenton and biological treatment for detoxification of pharmaceutical industry contaminants. Journal of Advanced Oxidation Technologies 11 (2), 261–269.