Photocatalytic Treatment of Diuron by Solar Photocatalysis: Evaluation

Rafael R. Solís, F. Javier Rivas, Ana Martínez-Piernas, Ana Agüera. .... Maria José Farré, Manuel Ignacio Maldonado, Wolfgang Gernjak, Isabel Oller, S...
0 downloads 0 Views 242KB Size
Environ. Sci. Technol. 2003, 37, 2516-2524

Photocatalytic Treatment of Diuron by Solar Photocatalysis: Evaluation of Main Intermediates and Toxicity S . M A L A T O , * , † J . C AÄ C E R E S , † A . R . F E R N AÄ N D E Z - A L B A , ‡ L . P I E D R A , ‡ M. D. HERNANDO,‡ A. AGU ¨ ERA,‡ AND J. VIAL‡ Plataforma Solar de Almerı´a-CIEMAT, Ctra. Sene´s Km. 4, 04200-Tabernas, Almerı´a, Spain, and Pesticide Residue Research Group, Faculty of Sciences, University of Almerı´a, 04071-Almerı´a, Spain

The technical feasibility, mechanisms, and performance of degradation of aqueous diuron (22 mg/L) have been studied at pilot scale in two well-defined photocatalytic systems of special interest because natural UV light can be used: heterogeneous photocatalysis with titanium dioxide and homogeneous photocatalysis by photoFenton. Equivalent pilot-scale (made up of Compound Parabolic Collectors (CPCs) specially designed for solar photocatalytic applications) and field conditions used for both allowed adequate comparison of the degree of mineralization and toxicity achieved as well as the transformation products generated en route to mineralization by both systems. Total disappearance of diuron is attained by both phototreatments in 45 min. 100% of chlorine was recovered as chloride, but total recovery of nitrogen as inorganic ions was not attained. 90% of mineralization was reached after 200 min of photocatalytic treatment, but toxicity measured by two different bioassays (Daphnia magna and a Microalga) was reduced to below the threshold (EC50%) in a shorter time. Transformation products evaluated by LC-IT-MS by direct injection of the samples were the same in both cases. The main differences between the two processes were in the amount of transformation products (DPs) generated, not in the DPs detected, which were always the same.

Introduction One unresolved environmental problem is the pollution of soils and aquatic systems by chemicals used in agriculture. One of the most commonly used herbicides is diuron [3-(3,4dichlorophenyl)-1,1-dimethylurea], a phenylurea, because it inhibits photosynthesis. Its action as a herbicide was first described in 1951 (1) and was marketed by E. I. du Pont de Nemours and Co (now DuPont) in 1954. The world production of diuron is around 14 000 to 16 000 tons per year (2). Diuron is manufactured by reacting 3,4-dichlorophenyl isocyanate with dimethylamine, which yields a colorless solid. It is stable at neutral pH at room temperature and is hydrolyzed by acids and alkalis. The main byproduct formed is 3,4-dichloroaniline (3), a very toxic intermediate (4). Diuron * Corresponding author phone: 34-950387940; fax: 34-950365015; e-mail: [email protected]. † Plataforma Solar de Almerı ´a-CIEMAT. ‡ University of Almerı ´a. 2516

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 11, 2003

is soluble in water (36.4 mg/L, 25 °C) and is highly persistent, with a half-life in soil of over 300 days. The strong toxicity of other main byproducts of its degradation [1-(3,4-dichlorophenyl)urea and methylurea] has also been described (5, 6). Therefore, because of its widespread use all over the world, its solubility in water, its strong persistence, and toxic intermediates, it is clearly bound to seriously contaminate soil (7) and aquatic environments at a mg/L level. Diuron can also be degraded by photochemical pathways (8, 9) and ozonation (10), but a wide range of photoproducts are formed with unknown ecotoxicological effects. Biodegradation of diuron by sludge of a sewage plant is almost negligible, and only a small quantity of its transformation product 1-(3,4dichlorophenyl)methylurea was detected with a BOD ) 0 (11). Biodegradation of diuron by fungi has exhibited higher toxicity of the metabolites than of the diuron (5). Due to all these reasons, diuron is included in the list of priority hazardous substances of the European Union (12) as a substance subject to emission controls and quality standards at the Community level in order to achieve “progressive reduction of discharges, emissions and losses”. In this context, it is important to develop suitable and efficient water treatment methods for removing diuron. Many technologies for pesticide removal including adsorption, filters, biological treatment, and advanced oxidation processes (AOP) have been reported. Among the AOPs, photocatalysis has been studied as an emerging successful technology for the decontamination of biorecalcitrant effluents (13). Unlike traditional nondestructive methods such as volatilization or adsorption onto a solid phase, this process has the advantage of destroying the organic compounds by oxireduction reactions. The mechanisms of this process have been described in detail elsewhere (14). In the past few years, solar photocatalysis has been shown in extensive research at pilot-plant scale (15, 16) to be an important alternative for pesticide degradation (17). This work focuses on one aspect of AOP research not very often addressed, as is the application of a complete analytical procedure for suitable evaluation of the AOP from the point of view of the environmental chemistry. The scope includes not only following of the process kinetics by determination of parent compound disappearance, complete mineralization, and release of anions but also evaluation of the main intermediates generated (DPs) by LC-MS and biotoxicity of the treated water. With this procedure, complete control of the degradation process could be achieved, and the environmental “suitability” of the overall treatment can be guaranteed. Moreover, this procedure has been used for comparing two alternative solar photocatalytic treatments: TiO2 heterogeneous-phase photocatalysis and homogeneousphase photo-Fenton.

Experimental Section Photoreactor. All the experiments were carried out under sunlight in compound parabolic collectors (CPC) at the Plataforma Solar de Almerı´a (PSA, latitude 37° N, longitude 2.4° W). The pilot plant (18) is made up of twin systems, each having three collectors, one tank, and one pump. Each collector (1.03 m2 each) consists of eight Pyrex tubes connected in series and mounted on a fixed platform tilted 37° (local latitude). The water flows at 20 L/min directly from one module to another and finally into a tank. The total volume (VT) of the reactor (40 L) is separated into two parts: 22 L total irradiated volume (in Pyrex tubes) (Vi) and the dead reactor volume (tank + connecting tubes). At the beginning of the experiments, with collectors covered, all 10.1021/es0261170 CCC: $25.00

 2003 American Chemical Society Published on Web 04/15/2003

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). 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 WUV m-2. This gives an idea of the energy reaching any surface in the same position with regard to the sun. With eq 1, combination of the data from several days’ experiments and their comparison with other photocatalytic experiments is possible

t30W,n ) t30W,n-1 + ∆tn

UV Vi ; ∆tn ) tn - tn-1 30 VT

(1)

where tn is the experimental time for each sample, UV is the average solar ultraviolet radiation measured during ∆tn, 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). As the CPCs do not concentrate light inside the photoreactor, the system is outdoors and is not thermally insulated, the maximum temperature achieved inside the reactor during the experiments is 25 °C. All test were performed in November. Chemicals. Technical-grade Diuron (98.5%) was supplied by Aragonesas Agro S. A. (Madrid, Spain). Analytical-standard Diuron was purchased from “Dr. Ehrenstorfer GmbH” (Augsburg Germany). Analytical-grade organic solvents were used for HPLC-UV (High-Pressure Liquid Chromatography with Ultraviolet Detector). The heterogeneous photocatalytic degradation tests were carried out using a slurry solution (200 mg/L of TiO2) of Degussa (Frankfurt, Germany) P-25 titanium dioxide (surface area 51-55 m2 g-1). For the photoFenton experiments (0.05 mM iron), the following chemicals were used: iron sulfate (FeSO4-7H2O), hydrogen peroxide reagent grade (30%), and sulfuric acid for pH adjustment (around 2.7-2.8). The concentration of peroxide in the reactor was determined by frequent analyses (iodometric titration) and maintained constant (around 15 mM) by adding small amounts as consumed. 2500 U/mg bovine liver catalase acquired from Fluka Chemie AG (Buchs, Switzerland) was used to eliminate the remaining H2O2 after sampling in photoFenton experiments, for which 0.5 mL of catalase solution (0.1 g/L) were added to 25 mL of sample (6 < pH < 7) and after 10 min hydrogen peroxide was completely decomposed. Using this procedure, up to 20 mM of hydrogen peroxide can be decomposed. The water used in the experiments was obtained from the PSA Distillation Plant (conductivity < 10 microS cm-1, Cl- ) 0.2 mg/L, NO3- ) 0.5 mg/L, organic carbon < 0.5 mg L-1). Analytical Determinations. Diuron was analyzed using reverse-phase liquid chromatography (at 0.5 mL/min) with UV detection in an HPLC (Hewlett-Packard, series 1100) equipped with a diode array detector (DAD) with C-18 column (LUNA 5micron-C18, 3 × 150 mm from Phenomenex). The mobile-phase composition and wavelength were H2O/ methanol at a ratio of 40/60 at 254 nm. Under these conditions, the detection limit of diuron is 0.02 mg/L, and the quantification limit is 0.1 mg/L TOC. Total Organic Carbon (TOC) was analyzed by direct injection of the filtered samples into a Shimadzu-5050A TOC analyzer calibrated with standard solutions of hydrogen potassium phthalate. Formation of inorganic anions was followed by LC-IC (Dionex120, anions column IonPAc AS14, 250 mm long). The eluent for inorganic anions was Na2CO3/NaHCO3 (1 mM/3.5 mM). The eluent for carboxylic acids was Na2B4O7 1 mM. Nessler spectrophotometric methods were used for ammonium. Analyses of DPs by LC-ES-MS (Liquid ChromatographyElectro Spray-Mass Spectrometry) was performed using a

Ion Trap (IT) system: Agilent Series 1100 MSD G2445A (Palo Alto, CA, U.S.A.) equipped with ES and Atmospheric Pressure Chemical Ionization (APCI) interfaces and controlled by Bruker Daltonics software. System calibration was made by MSD Trap control software. ES polarity ionization was set for positive mode under the following conditions: Nebulizer pressure: 50 psig, drying-gas flow: 10 mL‚min-1, drying-gas temperature: 325 °C, capillary voltage: 3500 V. Positive ionization mode was also set for APCI interface. Parameters for APCI positive mode were as follows: vaporizer temperature: 275 °C, Corona discharge intensity: 4 mA, drying gas flow: 10 mL‚min-1, nebulizer pressure: 50 psig, drying gas temperature: 325 °C, capillary voltage: 3500 V. To obtain information about the structure of Diuron and its byproducts, different fragmentor voltages of 60, 90, and 120V were used to produce mass fragmentation of the molecules. MSn was used to get more fragmentation and detailed information about the possible structures. Chromatographic separation was performed by using an Agilent Series 1100 liquid chromatograph (Palo Alto, CA, U.S.A.) equipped with a binary solvent delivery system, autosampler, and column heater. The column used was an XTerra MS C8 column (100 × 2.1 mm - 3.5 µm) purchased from Waters (Milford, MA). A gradient elution was made up using binary gradient of acetonitrile (solvent A) and ammonium formate 50 mM, acetonitrile 5%, acidified with formic acid to pH 3.5 (solvent B) as follows: Linear gradient from 10 to 50% of B in 10 min and then 100% B in 10 min. After 1 min at 100% acetonitrile, the mobile phase was returned to initial conditions. The flow rate was kept at 0.25 mL‚min-1 and elution was monitored by UV at 254 nm. Toxicity Studies. For Daphnia magna and Selenastrum capricornotum bioassays, stock solutions were prepared by dilution in a specific culturing medium. Points representing the evolution of toxicity are the arithmetic mean of three toxicity measurements. Reproducibility studies were performed with replicates of the toxicity bioassays under the same conditions for 3 days. Reproducibility, expressed in terms of coefficient of variation, were 22% and 20% for D. magna and S. capricornutum, respectively. Daphnia magna. The toxicity of diuron and its photoproducts for crustaceans was assessed using commercially available Toxkit Daphtoxkit (Creasel, Belgium). The toxicity studies were performed in accordance with testing conditions prescribed by OECD Guideline 202 (1995) and ISO 6341. Acute toxicity was assessed by noting the effects of the test compounds on the motility of Daphnia magna. This bioassay using dormant eggs (ephippia) was treated as described in the standard operational procedure to induce hatching and the experiments were then carried out on daphnids less than 24 h old. All the tests were performed in the dark at a constant temperature of 20 ( 1 °C. The neonates are considered immobilized after 24 and 48 h of incubation, if they lie on the bottom and do not resume swimming within 15 s of observation. The toxicity end-point (EC50) was determined as the concentration estimated to immobilize 50% of the daphnids after 24 and 48 h exposure. Microalgae. The alga growth-inhibition test was performed using the freshwater green microalga Selenastrum capricornotum according to the prescriptions of the OECD (Guideline 201, 1995) and using the commercially available Toxkit Algaltoxkit (Creasel, Belgium). Algaltoxkit makes use of deimmobilized microalga beads in an inert matrix. After deimmobilization and transfer into an adequate culturing medium, the microalgae resume their growth immediately. The initial number of algal cells was adjusted to 106 cells/ mL, and the test tubes were then kept at 25 °C in an incubator under continuous illumination. Each test tube containing the test compound, and algae in the medium was incubated at 25 °C for 3 days. Inhibition of alga growth relative to control VOL. 37, NO. 11, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2517

FIGURE 1. Disappearance of diuron as a function of t30W (illumination time). Mineralization and chloride, nitrate and ammonia release is also shown. TiO2 (200 mg/L): White symbols. Photo-Fenton (0.05 mM Fe): Black symbols. was determined by measurement of the OD in spectrophotometer with a filter at 670 nm. The 72-h EC50 for Selenastrum capricornotum was calculated as the test concentration resulting in a 50% reduction in growth relative to the control.

Results and Discussion All experiments were performed at the highest diuron concentration (around 22 mg/L) attained dissolving it in water at around 10 °C (ambient temperature at Plataforma Solar de Almerı´a in November). The intention was to attain the highest possible initial concentration for 100% inhibition (in the toxicity tests) and evaluation of intermediates. All tests were repeated three times. Several blank experiments, at the same initial concentrations as the photocatalytic experiments, were performed to guarantee that the results obtained during the photocatalytic tests were consistent and not due to hydrolysis and/or photolysis. Hydrolysis experiments were done at different pH (2.7, 5, 7, and 9). Hydrolysis was performed at pH 2.7 because photo-Fenton tests were done at this pH. No hydrolysis was detected in any case after 72 h. Furthermore, results obtained with solar illumination without catalysts indicate that degradation was always insignificant (44% diuron degradation after 48 h of illumination, 0% TOC disappearance) compared to photocatalytic treatments (see Figure 1). Diuron and solar UV spectra slightly overlap in the 300-to-330-nm region showing that absorption of solar photons can produce photoalteration after exposure to the environment, but mineralization never occurs. As mentioned above, different degradation products may also be formed, so disposal of diuron into the environment could be very risky. The pesticide was successfully degraded by both photocatalytic procedures (see Figure 1). The “dark” Fenton reaction produced very slight mineralization of the pesticide before illumination. Nevertheless, mineralization (i.e., disappearance of TOC) can be attained only after irradiation. This disappearance of parent compounds in the dark by Fenton has been described by a multitude of authors, but in this case, it is very slight because of the small quantity of Fe2+ used (0.05 mM), which is rapidly (less than 10 min) converted to Fe3+ by hydrogen peroxide. The total disappearance of diuron was obtained at 45 min (t30W) by photo-Fenton and approximately at the same time by TiO2. The photocatalytic disappearance of diuron with TiO2 followed apparent first2518

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 11, 2003

order kinetics, as is usual in heterogeneous photocatalysis for other pollutants when initial concentration is low enough and no catalyst saturation occurs. It should be emphasized that DPs formed during diuron decomposition could also be competitive on the surface of the TiO2. Their concentration varies throughout the reaction up to their mineralization and thus, the following equation (based on Langmuir Hiselwwod kinetic model, used commonly for describing photocatalysis by TiO2 kinetics (19)) can describe the kinetics

krKC

r)

(2)

n

1 + KC +

∑ K C (i)1,n) i i

i)1

where kr is the reaction rate constant, K is the reactant (diuron) adsorption constant, C is diuron concentration at any time, Ki is the DP adsorption constant, and Ci is DP concentration at any time. When C0 (22 mg/L of diuron) is low enough, eq 3 can be simplified (1 + KC + Σ ...) 1) to an order one reaction rate equation:

r ) kapC

(3)

This was confirmed by the linear behavior of Ln(C0/C) as a function of t30W. A constant rate of kap ) 0.092 min-1 was observed. Similar behavior was observed with the photoFenton treatment when the C0 used for the calculation corresponds to the beginning of illumination and the dark reaction is not considered. The Fenton reactant consists of an aqueous solution of hydrogen peroxide and ferrous ions providing hydroxyl radicals (eq 4). When the process is complemented with UV/visible radiation (eq 5), it is called photo-Fenton. In this case, the process becomes catalytic.

Fe2+ + H2O2 f Fe3+ + OH- + •OH

(4)

Fe3+ + H2O 9 8 Fe2+ + H+ + •OH hν

(5)



OH + diuroni f DPs

(6)

Assuming that the reaction between •OH radicals and diuron is the rate-determining step, the rate equation is

then written as

reaction is given here only to its most oxidized state.

r ) kOH[•OH]C ) k′apC

(7)

where C is diuron concentration, kOH is the reaction rate constant, and k′ap is a pseudo-first-order constant. This was confirmed by the linear behavior of Ln(C0/C) as a function of t30W. A constant rate of k′ap ) 0.20 min-1 was observed. It should be emphasized that, being the first-order kinetics constant of photo-Fenton higher than that of TiO2, total diuron disappearance (