Chemosphere 40 (2000) 403±409
Solar photocatalytic mineralization of commercial pesticides: acrinathrin S. Malatoa,*, J. Blancoa, A.R. Fern andez-Albab, A. Ag uerab a
b
PSA (Plataforma Solar de Almerõa). Crta Sen es s/n, Tabernas, Almerõa 04200, Spain Pesticide Residue Research Group. Faculty of Sciences. University of Almeria, Almeria 04071, Spain Received 26 March 1999; accepted 8 June 1999
Abstract A comparative study of the degradation of commerical acrinathrin spiked in water using TiO2 photocatalysis and photolysis under sunlight was performed. Samples were analysed by liquid chromatography-diode array detector (HPLC±DAD) and gas chromatography-ion trap-mass spectrometric detector (GC±ITMS). Additional total organic carbon (TOC) analyses were carried out to evaluate the mineralisation rates. One photoproduct, 2-phenoxy benzaldehyde, was unequivocally identi®ed and evaluated by GC±ITMS during the processes. Although acrinathrin is almost destroyed when exposed to irradiation for more than 400 h, photocatalysis with TiO2 noticeably reduced degradation to a few hours. In this case, with the additional presence of peroxydisulphate, in less than 2 h acrinathrin is completely destroyed. Mineralisation of acrinathrin, without catalyst, was only around 50% after 400 h of irradiation. Ó 1999 Elsevier Science Ltd. All rights reserved. Keywords: Acrinathrin; Photocatalysis; Solar energy; Photolysis
1. Introduction The large amount of pesticides necessary in intensive agriculture, such as in the Mediterranean area, has important eects on the quality of environmental water. As a consequence, the presence of herbicides, insecticides or fungicides in super®cial or ground water is well known and has been evaluated in many of the water monitoring programs that are now being carried out in the US and EU (Chiron et al., 1995; Meyer and Thurman, 1996). The removal of these organic chemicals from water is a pressing ecological problem. It is thus important to evaluate the main ways in which pesticides enter runo water in these areas so that action to reduce risk to the
*
Corresponding author. Tel.: +34-9-50-387-900; fax: +34-950-365-015. E-mail address:
[email protected] (S. Malato).
environment may be developed. These are: (i) agricultural pesticide treatment practices; (ii) contaminated rinse-water from washing pesticide applicators and containers; (iii) disposal of unused portions of pesticides; (iv) water baths for industrial rinsing of crops before packaging and (v) from evaporation of water from plants inside greenhouses. It must be noted that the ®rst point leads to widely-dispersed pollution, usually at sub lg/l levels, while the other four points lead to well-de®ned sites at typical concentrations of around 10±100 mg/l of pesticides. Two strategies, not including reduction in the emissions themselves, could be employed to remove these substances from the environment: (i) chemical treatment of drinking water and contaminated surface and groundwater and (ii) chemical treatment of wastewater containing these non-biodegradable compounds. The treatment of the source of contamination, when possible, would always be easier because the ¯ow is smaller, it
0045-6535/00/$ - see front matter Ó 1999 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 5 - 6 5 3 5 ( 9 9 ) 0 0 2 6 7 - 2
404
S. Malato et al. / Chemosphere 40 (2000) 403±409
Almerõa Desalination Plant (conductivity 80% between 300 and 400 nm, volume 2 l, internal diameter: 15 cm) and covered
2.1. Materials and reagents Acrinathrin analytical standard was obtained from Dr. Ehrenstorfer (Augsburg, Germany). The commercial formulation used was RufastÒ LE with 15% w/v of acrinathrin (Rhone Poulenc, Paris, France). 2-Phenoxy benzaldehyde was purchased from Merck (Darmstadt, Germany). Pesticide-grade dichloromethane, ethyl acetate, cyclohexane and anhydrous sodium sulphate were supplied by Panreac (Barcelona, Spain). Titanium dioxide P25 was purchased from Degussa (Frankfurt, Germany) and sodium peroxydisulphate from Probus (Barcelona, Spain). The water used in the degradation assays was obtained from the Plataforma Solar de
Fig. 2. Pilot plant used for the acrinathrin photocatalytic degradation experiments.
S. Malato et al. / Chemosphere 40 (2000) 403±409
with a Pyrex top in order to avoid contamination and evaporation of samples. After that, two of them were exposed to sunlight from 9 a.m. to 4 p.m. daily during March, April and May. The third beaker was kept in the dark. 2.3. Evaluation of solar UV radiation
ACPC VTOT
studies were performed in water at 0.5 and 5.0 mg/l with the extraction method proposed. For this purpose, 1 l volumes of water were spiked with standard solution described above, and extracted at least three times at each concentration. 97.3% mean recovery of 2-phenoxy benzaldehyde was obtained. 2.5. Analytical determinations
A global UV radiometer (KIPP&ZONEN, model CUV3) with typical sensitivity of 264 lVolts Wÿ1 m2 , mounted on a platform tilted 37° (the same angle as the CPCs), provides data in terms of incident WUV mÿ2 . 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 daysÕ experiments and their comparison with other experiments is possible. QUV; n QUV; nÿ1 Dtn UVG;n
405
1
where tn is the experimental time for each sample, UVG;n the average UVG during Dtn , ACPC the collector surface, VTOT the total plant volume and QUV; n is the accumulated energy (per unit of volume, kJ/l) incident on the reactor for each sample taken during the experiment. Sometimes it is necessary to explain the results in terms of illumination time instead of QUV . For this purpose it may be assumed that the average Solar UV on a sunny day for 2 h around noon is about 30 WUV mÿ2 . Under these conditions and in the reactor used in this work, 1 kJ/l is equivalent to 6.5 min illumination time. When this simpli®cation is used, the illumination time calculated by this procedure is called t30 W . A similar procedure has been used for the photolysis experiments. The total area of the transparent beakers (2 l) exposed to sunlight was 0.0182 m2 (diameter height). In this case, 1 kJ/l is equivalent to 61 min illumination time. 2.4. Sample handling During all experiments, the irradiated solutions were made perfectly homogeneous by agitation. Samples were diluted 1:1 with acetonitrile (to avoid product loss in the ®lter due to adsorption and/or retention of TiO2 particles) before ®ltration (0.45 lm) an injection in HPLC± DAD. TOC was measured directly without ®ltration. For GC±ITMS analysis, samples were extracted (l±l) and preconcentrated to detect possible degradation products. A 200 ml sample with 1.0 g of sodium chloride was extracted twice with 75 and 50 ml, respectively, of dichloromethane. The combined organic extracts were ®ltered through anhydrous sodium sulphate and concentrated in a rotary evaporator down to 2±3 ml. This extract was again dried by evaporation with a gentle N2 stream and dissolved in 1 ml of ciclohexane:ethyl acetate (9:1) before injection. 2-phenoxy benzaldehyde recovery
HPLC±DAD analyses were performed using a Hewlett±Packard system HP-1040M Series II/HP-1050 pump (Hewlett-Packard, Palo Alto, CA). The system was equipped with a LiChrospher ODS column, 5 lm particle size, 4.6 mm i:d: 125 mm (Merck, Darmstadt, Germany). The following acetonitrile/water mobilephase program was used: 25/75 (0±2 min), 90/10 (15 min) and 100/0 (17 min). The ¯ow of the mobile phase was 1 ml/min and injection volume was 20 ll. Diluted solutions containing acrinathrin 0.1, 0.5, 1, 5, 10, 20, 25 and 50 mg/l were prepared in acetonitrile:water (1:1) and analysed in duplicate by HPLC±DAD to calculate the linear range of the detector. The calibration curve was constructed at 230 nm and the correlation coecient was found to be 0.98. This curve was used for quantitative analysis. The detection limit (LOD) for acrinathrin under these conditions was 0.08 mg/l. A GC±ITMS Saturn-3 system (Varian, Harbor City, CA) consisting of Varian 3400 gas chromatograph, a Model 1093 septum-programmable injector (SPI) and a 8200 autosampler, was used to identify reaction intermediates. Data acquisition, processing and instrument control were performed by Saturn GC±ITMS workstation version 5.2 software. A DB-5MS (J&W Scienti®c, Folsom, USA) capillary column, 30 m 0:25 lm i.d., 0.25-mm ®lm thickness was connected to the system. GC operating conditions were: 1.0 ll injection volume; solvent plug 0.1 ll; 0.1 s needle hold time in port before injection; 60°C injection port for 0.5 s followed by ramping to 280°C at 150°C/min; 9 psi He column head pressure; oven temperature program: 1.0 min at 60°C, 25°C/min to 180°C, 5°C/min to 280°C (4 min). Transfer line temperature, 280°C; detector manifold temperature, 230°C. ITMS±EI mode operating conditions were as follows: 35 lA ®lament current; 1350 V electron multiplier tube and automatic gain control at 40.000. The mass spectra were monitoring from 50 to 550 m/z and data acquisition was from 4 to 25 min of the chromatogram. For the ITMS±CI mode, the same conditions already described for the EI mode were used. Acetonitrile was selected as the reagent gas. The CI parameters were optimised as follows; maximum ionisation time 2.5 ms, maximum reaction time 50 ms, ionisation storage level 12.5 m/z, reagent ion eject amplitude 8 V and reaction storage level m/z 20. ITMS±MS analyses were performed for a best identi®cation of the metabolites. The
406
S. Malato et al. / Chemosphere 40 (2000) 403±409
operating parameters were optimised by a Varian Toolkit program. Due to the possibility of decomposition processes in the GC system, the response of extracts of pure standards of 2-phenoxy-benzaldehyde was checked and no degradation products were detected. 100 mg portions of 2-phenoxy benzaldehyde analytical standard were dissolved in 100 ml of acetone to give 1 mg/ml stock solution. This solution was used as a spiking solution for the recovery assays. Diluted solutions containing 0.1, 0.5, 1, 5, 10, 20 and 25 mg/l were also prepared in cyclohexane:ethyl acetate (9:1) and analysed by GC±IT±MS to calculate the linear range of the detector. The correlation coecient of the calibration curves was 0.97. The LOD for 2-phenoxy benzaldehyde by GC±ITMS in the EI mode (ion 198) after preconcentration was 0.6 lg/l. TOC determination was done using direct injection of the samples into a Heraeus-Foss Electric TOC-2001 (UV-Peroxydisulphate method, EPA 415.1) in order to monitor the mineralisation rate in the contaminated water.
3. Results and discussion 3.1. Quantitative information Acrinathrin and solar UV (latitude 37°N, longitude 2.4°W) spectra overlap slightly in the 300±330 nm region (Fig. 3) showing that capacity for absorption of solar photons can produce photoalteration processes after being exposed to the environment. But such natural photodegradation (Fig. 4) is very slow: Q 57 kJ/l (t30 W 58 h) to reach half the initial concentration under aerobic well-illuminated conditions (transparent glass, 15 cm ID). This means at least 10 days with perfectly sunny conditions for reduction to one half. 50 mg/l of acrinathrin in the dark are not degraded after 77 days (Fig. 4) at pH 7±8. These eects should be very similar
Fig. 3. UV spectra between 200 and 400 nm of acrinathrin and Almerõa sunlight.
Fig. 4. Kinetics of the disappearance of acrinathrin by photolysis (d), photocatalysis (m), photocatalysis with peroxydisulphate (n) and in the dark (r). Dark experiment is included to clarify the results but X axis can not be used in this case. Time evolution of 2-phenoxy benzaldehyde (right axis) during the photolytic (s) and photocatalytic (n) processes.
when acrinathrin is present in natural waters. In surface water, degradation is extremely slow and in ground water almost negligible. So disposal of acrinathrin into the environment could be very risky. When TiO2 is used (Fig. 4), the degradation rate is increased about 22 times (t30 W 2:7 h, to reach half of the initial concentration) and acrinathrin degradation is practically complete in t30 W 8:5 h. This is around two days in the photoreactor studied. The illumination of TiO2 (band-gap energy, EG 3.2 eV) in water with less than 387 nm wavelength light generates excess electrons in the conduction band (eÿ CB ) and positive ``holes'' (hVB ) in the valence band. TiO2 hv
P EG ! eÿ CB hVB
2
The OH radicals are extremely reactive and readily attack organic molecules degrading them to CO2 and H2 O (and inorganic ions when the organic molecule contains other atoms). One practical problem in using TiO2 as a photocatalyst is electron/hole recombination (reverse of Eq. (2)), which, in the absence of proper electron acceptors, is extremely ecient and thus represents a major energy-wasting step. Oxygen has been chosen for most of the applications for this purpose. In the present case, oxygen is introduced in the pilot plant because the tank (see Fig. 2) is open to the atmosphere permitting continuous contact between the water and air and maintaining the O2 concentration in the water at around 6±8 mg/l. One strategy for inhibiting eÿ /h recombination is to add other (irreversible) electron acceptors to the reaction. The use of inorganic peroxides has been demonstrated to notably enhance the rate of degradation of several organic contaminants (Malato et al., 1996; Martin et al., 1995; Minero et al., 1996; Gr atzel et al.,
S. Malato et al. / Chemosphere 40 (2000) 403±409
1990; Malato et al., 1999b) because they trap photogenerated electrons more eciently than O2 (Pelizzetti et al., 1991). It must be mentioned here that in much of the highly toxic wastewater in which degradation of organic pollutants is the major concern, the addition of an inorganic anion may be justi®ed to enhance the degradation rate of organics. For better results, these additives should ful®l the following criteria: dissociate into harmless by-products and lead to the formation of OH or other oxidising agents. In the case of acrinathrin, the presence of peroxydisulphate in combination with TiO2 (See Figs. 4 and 5) strongly accelerates the degradation rate to more than 200 times compared with photolysis, achieving nearly complete disappearance of acrinathrin in less than two hours. TPS are formed only in photolysis and photocatalysis without peroxydisulphate (2-phenoxy benzaldehyde, see Fig. 4). This TP remains in solution for t30 W 6:5 h in the photocatalysis experiments and for the ®rst 175 h in the case of photolysis. The highest concentration detected was 0.29 mg/l (photolysis). The degradation rate is considerably higher at the beginning of the process than toward the end, which is typical in heterogeneous photocatalytic processes (Minero et al., 1996). Other similar studies of photolysis (Lacorte and Barcel o, 1994; Lacorte et al., 1995; Bertrand and Barcel o, 1991) and photocatalysis (Bertrand and Barcel o, 1991; Pe~ nuela and Barcelo, 1996; Muszkat et al., 1995) in pesticides show that acrinathrin is more resistant to both processes studied. The initial TOC obtained from adding 82 ml of Rufast (15% w/v of acrinathrin) to 247 l of water was almost 50 mg/l (Fig. 5), of which only 29 mg/l are from acrinathrin, and the rest come from the formulation. As the manufacturers refuse to reveal the rest of the compounds in the formulation, these are unknown and must be considered potential contaminants to be eliminated
Fig. 5. Time evolution of the TOC values during the photolysis (n), photocatalysis (.) and photocatalysis with 10 mm peroxydisulphate (d) assays. Dark experiment is also shown (s), but X-axis cannot be used in this case.
407
by the treatment. TOC obtained in the three photodegradation processes studied provided data on the amount of global mineralisation. In Fig. 5, TOC is close to 0 mg/l of TOC in photocatalysis at the end of the experiments and around 30 mg/l, 50% of the initial amount, in the case of photolysis. At ®rst sight, it is evident that the addition of peroxydisulphate substantially bene®ts photocatalytic treatment and leads to very signi®cant reduction of at least 5 times the energy necessary for total mineralisation. This reduction would also reduce the solar collector surface necessary to degrade the organics by the same factor. 3.2. Qualitative information The use of GC±ITMS and HPLC±DAD techniques allowed the identi®cation of both the parent compound and the potential TPs by means of standard and spectral comparison. GC±ITMS was able to identify only one TP. HPLC allowed only acrinathrin to be identi®ed, probably due to the low concentration of TPs present in samples. Photolysis leads to formation of TPs rather than to complete mineralisation as revealed by the high TOC measured after 406 hours of irradiation (see Fig. 5). So for this process, other TPs in addition to 2-phenoxy benzaldehyde may be expected, although they have not been identi®ed, probably because of their strongly polarised nature (e.g., organic acids) as is the case with other pesticides previously studied (Minero et al., 1996; Pelizzetti et al., 1990), or very low response to detectors used in the chromatographic analysis.
Fig. 6. Total ion and selected ion (m/z 181 and 198) GC±ITMS chromatograms of a photocatalysis sample extract after 4 hours of irradiation. Acrinathrin (A) and 2-phenoxy benzaldehyde (B) were detected at retention times of 23.69 and 9.65 min, respectively.
408
S. Malato et al. / Chemosphere 40 (2000) 403±409
Table 1 Main fragment ions and relative abundance (%) of acrinathrin and 2-phenoxy benzaldehyde mass spectrum in electronic ionisation (EI), chemical ionisation (CI) and MS±MS modea m/z (RA %) Acrinathrin
EI
CI
181(100) 93(37) 289(24) 77(21) 208(15)
181(6)
541(2) 2-Phenoxy benzaldehyde
198(100) 169(92) 141(88) 197(85) 77(10) 115(60) 181(51)
a
Tentative asignation MS±MS [C6 H5 OC6 H5 C] [C6 H5 O] [M(CF3 )CH2 ] [C6 H5 ] [C6 H5 OC6 H5 ± CHCN] M
208(100) 125(18) 198(24) 199(100)
198(68) 169(20) 141(20) 197(100) 181(4) 153(30)
M [M+1] [M±CO] [M±1] [C6 H5 ] [M±OH]
Tentative assignation of some fragments has been included.
Fig. 6 shows a total ion, selected ion (181 and 198 ions) and chemical ionisation chromatograms for a sample extract after 4 h of irradiation in photocatalysis. Acrinathrin and one TP could be clearly distinguished at retention times of 9.7 and 23.7 min. The typical ion fragments of these compounds are reported in Table 1. Each molecule was identi®ed by at least three dierent ions. Acrinathrin and 2-phenoxy benzaldehyde were unequivocally identi®ed by comparing the spectra with those obtained by injection of pure standards. In the case of 2-phenoxy benzaldehyde, additional con®rmation resulted from chemical ionisation and MS±MS studies.
4. Conclusions The detoxi®cation of water contaminated by pesticides (usually at low-to-medium concentrations, but highly toxic) is a ®eld where photocatalysis could be applied in the future. But euents are never contaminated by pure pesticides, because they are not marketed that way. So tests are more realistic when they are carried out with commercial products than with pure compounds. More than 400 h (t30 W ) of irradiation are necessary for 90% of acrinathrin to disappear in water, although TOC remains at around 50% of the initial amount. The addition of TiO2 is a very eective method of achieving the destruction of acrinathrin and diminishing TOC until complete mineralisation. Peroxydisulphate has a very important synergetic eect on degradation rates. Almost complete mineralisation of
acrinathrin can be achieved using a CPC system under solar light in 2 and 8 h (t30 W ) for photocatalysis with and without peroxydisulphate, respectively. This enables signi®cant reduction in photoreactor dimensions. GC± ITMS has been shown to be a useful technique for monitoring pesticide phototransformation products in water, but more eort is required in prior extraction and pre-concentration procedures. 2-phenoxy benzaldehyde is the main TP of acrinathrin in photolysis and photocatalysis. But, more TPs can be expected. Acknowledgements This study has been supported by CICYT Project CICYT AMB95-0075-003-002 and PETRI Project PEN96-459 (Pesticide Residue Group of Univ. of Almerõa) and by the Consortium Agreement between CIEMAT, AEPLA; IBACPLAST, ECOSYSTEM, Diputaci on de Almerõa and 9 Councils of the Province of Almerõa-Spain. The authors thank Rhone-Poulenc Agro S.A. (Mr. J. G omez-Arnau) for the pure acrinathrin and the technical information supplied. The authors also thank Mrs. Deborah Fuldauer for correcting the English. References Chiron, S., Valverde, A., Fernandez-Alba, A.R., Barcel o, D., 1995. Automated sample preparation for monitoring ground water pollution by carbamate insecticides and their transformation products. J. AOAC Int. 78, 1346±1356.
S. Malato et al. / Chemosphere 40 (2000) 403±409 Bickley, R.I., 1996. Technologies for environmental protection. Euro Abstracts Report 11, European Commission, Directorate-General XII. Betrand, N., Barcel o, D., 1991. Photodegradation of carbamate pesticides aldicarb carbaryl and carbofuran in water. Anal. Chim Acta 254, 235±244. Gratzel, C.K., Jirousek, M., Gr atzel, M., 1990. Decomposition of organophosphorus compounds on photoactivated TiO2 surfaces. J. Mol. Cat. 60, 375±387. Lacorte, S., Barcel o, D., 1994. Rapid degradation of fenitrotion in estuarine waters. Environ. Sci. Technol. 28, 1159±1163. Lacorte, S., Lartiges, S., Garrigues, P., Barcel o, D., 1995. Degradation of OPs pesticides and their transformation products in estuarine waters. Environ. Sci. Technol. 29, 431±438. Legrini, O., Oliveros, E., Braun, A.M., 1993. Photochemical processes for water treatment. Chem. Rev. 93, 671±698. Malato, S., Richter, C., Blanco, J., Vincent, M., 1996. Photocatalytic degradation of industrial residual waters. Solar Energy 56, 401±410. Malato, S., Blanco, J., Herrmann, J.M. (Eds.) 1999a. Solar Catalysis for Water Decontamination (special issue) Catalysis Today (1 manuscript in press). Malato, S., Blanco, J., Richter, C., Miloe, B., Maldonado, M.I., 1999b. Solar photocatalytic mineralization of commercial pesticides methamidophos. Chemosphere 38, 1145±1156. Martin, S.T., Lee, A.T., Homann, M.R., 1995. Chemical mechanism of inorganic oxidants in the TiO2 /UV process:
409
increased rates of degradation of chlorinated hydrocarbons. Environ. Sci. Technol. 29, 2567±2573. Meyer, M.T., Thurman, E.M., 1996. Herbicide Metabolites in Surface Water and Groundwater. ACS Symposium Series Vol. 630, American Chemical Society, Washington, DC. Minero, C., Pelizzetti, E., Malato, S., Blanco, J., 1996. Large solar plant photocatalytic water decontamination degradation of atrazine. Solar Energy 56, 411±420. Muszkat, L., Bir, L., Fiegelson, L., 1995. Solar photodegradation of xenobiotic contaminants in polluted well water. J. Photochem. Photobiol. A: Chem. 87, 85±88. Pelizzetti, E., Maurino, V., Minero, C., Carlin, V., Pramauro, E., Zerbinati, O., 1990. Photocatalytic degradation of atrazine and others-Triazine herbicides. Environ. Sci. Technol. 24, 1559±1565. Pelizzetti, E., Carlin, V., Minero, C., Gr atzel, M., 1991. Enhancement of the rate of photocatalytic degradation on TiO2 of 2-Chlorophenol 2 7-Dichlorodibenzodioxin and atrazine by inorganic oxidizing species. New J. Chem. 15, 351±359. Pe~ nuela, G.A., Barcel o, D., 1996. Comparative degradation kinetics of alachlor in water by photocatalysis with FeCl3 , TiO2 and photlysis studied by SP-Disk extraction followed by GC techniques. J. Chromatogr. 754, 187±195. Tomlin, C., 1994. The pesticide manual, 10th edition. Crop Protection Publications, British Crop Protection Council and Royal Society of Chemistry, Cambridge UK.
