Photocatalytic degradation of the herbicides propanil and molinate over aqueous TiO2 suspensions: identification of intermediates and the reaction pathway

Applied Catalysis B: Environmental 34 (2001) 227–239

Photocatalytic degradation of the herbicides propanil and molinate over aqueous TiO2 suspensions: identification of intermediates and the reaction pathway Ioannis K. Konstantinou, Vasilios A. Sakkas, Triantafyllos A. Albanis∗ Department of Chemistry, University of Ioannina, Ioannina 45110, Greece Received 11 March 2001; received in revised form 30 May 2001; accepted 3 June 2001

Abstract The light-induced degradation of propanil and molinate under simulated solar irradiation has been investigated in aqueous solutions containing TiO2 suspensions as photocatalysts. The study focus on the identification of possible intermediate products and the determination of inorganic ions formed during the process, using several powerful analytical techniques such as gas chromatography–mass spectrometry (GC–MS) and ion chromatography (IC). The primary degradation of propanil and molinate has been a fast process with half-lives varied from 4.3 to 2.9 min, respectively, and followed pseudo-first-order kinetics according to the Langmuir–Hinshelwood model. The stoichiometric transformation of organic chlorine into chloride ion and organic sulfur to sulfate ions was observed for propanil and molinate, respectively, whereas oxidation of nitrogen to nitrate ions took place at delayed irradiation times for both herbicides. The mineralization of the organic carbon to CO2 after 240 min of irradiation was found to be ≥95% for both herbicides. Various organic intermediates detected during the treatment have been identified by GC–MS techniques. Based on this by-product identification, a possible multi-step degradation scheme was proposed for each herbicide including hydroxylation, dechlorination, dealkylation and oxidation steps that lead to the mineralization of the starting molecule. This work points to the necessity of extended knowledge of the successive steps in a solar-assisted detoxification process. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Photocatalytic oxidation; TiO2 ; Propanil; Molinate; Photoproduct analysis

1. Introduction Among the compounds of environmental concern, anilides and thiocarbamates derivatives constitute important classes of herbicides that are widely used for weed control in agricultural crops. Within these families, propanil (3 ,4 -dichloropropioanilide) and ∗

Corresponding author. Tel.: +30-651-98348; fax: +30-651-98795. E-mail address: [email protected] (T.A. Albanis).

molinate (S-ethyl-N-hexamethylene-thiocarbamate) are extensively used in Greece and other European countries for the control of some annual grasses and broadleaf weeds in several different crops and especially in rice. Consequently, their residues have been detected in various natural waters of European countries [1,2]. Therefore, the investigation of remediation treatments of polluted waters containing trace amounts of herbicides is of environmental interest. In the last few years, research on new methods for water purification has moved from processes involving

0926-3373/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 6 - 3 3 7 3 ( 0 1 ) 0 0 2 1 8 - 1

228

I.K. Konstantinou et al. / Applied Catalysis B: Environmental 34 (2001) 227–239

phase transfer of a contaminant (e.g. from liquid to solid, such as activated carbon, or from liquid to gas such as air-stripping of volatile contaminants) towards processes involving chemical or biochemical destruction of the contaminant [3–6]. However, chemical oxidation is unable to mineralize all organic substances and in biological treatment the slow reaction rates and the disposal of activated sludge are the drawbacks to be considered [7,8]. Photocatalytic oxidation is one of the emerging technologies for the elimination of organic micropollutants because of the efficiency in their mineralization [8], ideally producing as end products carbon dioxide, water, and inorganic mineral ions [9–12]. Different types of semiconductors have been used in order to generate highly reactive intermediates, primarily the hydroxyl radical (• OH), that initiate a sequence of reactions leading to the partial or total destruction of organic pollutants as chlorophenols, nitrogen containing pesticides and aromatic compounds [6,13–18]. Among the semiconductors used, titanium dioxide (TiO2 ) is considered a very efficient catalyst that, unlike other semiconductors, is non toxic, stable to photocorrosion, low cost and suitable to work using sunlight as energy source [8,18,19]. The last fact is particularly relevant for Mediterranean agricultural areas, where solar irradiation is highly available making this process quite attractive. The primary step in the process involves the generation of conduction band electrons and valence band holes by the illumination of the TiO2 with light energy greater than the band gap energy. Some of the electrons and holes migrate to the surface of the particle before recombination occurs. Oxidation occurs if the photogenerated holes react with adsorbed organic molecules whose oxidation potentials are less positive than the valence band edge [20–22]. The results of many studies are consistent with the oxidation proceeding via hydroxyl and other oxygen containing radicals formed in reactions between the positive holes and hydroxyl groups or water adsorbed at the titania surface [23,24]. These radicals are powerful oxidizing agents capable of attacking a wide variety of organic molecules. The titanium dioxide photocatalyzed degradation of some anilides and thiocarbamates including propanil and molinate has been discussed previously, mostly in terms of degradation kinetics and identification

of by-products [25–27]. The present photocatalytic degradation study of propanil and molinate reveals new transformation products for both herbicides and tries not only to identify these intermediates, but also to incorporate them in an integrated degradation pathway. Moreover, a complete description of the photocatalytic degradation of each herbicide is given, by monitoring the mineralization products such as carbon dioxide and inorganic ions. Thus, the objectives of the study were (i) to determine the main transformation products (TPs) by using gas chromatography–mass spectrometry (GC–MS) techniques and to propose a degradation pathway, (ii) to evaluate the kinetics of disappearance of the tested herbicides and (iii) to monitor the evolution of CO2 and of the main inorganic ions produced during the process.

2. Experimental 2.1. Chemicals Propanil and molinate analytical grade standards (99.5% purity) were purchased from Riedel-de Häen (Germany). Pesticide grade n-hexane, dichloromethane, ethyl acetate and acetone were purchased from Labscan Ltd. (Dublin, Ireland). Barium hydroxide and sodium sulfate (pro-analysis) was from Merck (Darmstadt, Germany). Titanium dioxide (Degussa P25), a known mixture of 80% anatase and 20% rutile form with an average particle size of 30 nm, nonporous with a reactive surface area of 50 ± 10 m2 /g was used as received for all degradation experiments. Phthalic acid and tris(hydroxymethyl)aminomethane (Riedel-de Haën, Seelze, Germany) were used to prepare the ion chromatography eluent. Doubly distilled water, filtered through 0.45 ␮m HA cellulose acetate membranes (Millipore), was used throughout the work. 2.2. Irradiation procedure Irradiation experiments were carried out on stirred aqueous solutions using a 50 ml cylindrical pyrex glass UV-reactor under continuous flushing of O2 at a rate of 15 ml/min. Degradations were performed on 15 ml of aqueous solutions containing the desired concentration of herbicides (10 mg/l) and two different amounts

I.K. Konstantinou et al. / Applied Catalysis B: Environmental 34 (2001) 227–239

of TiO2 (100 and 500 mg/l) at natural pH. The suspensions were allowed to stay in the dark for 60 min under stirring to reach adsorption equilibrium on to semiconductor surface. For the identification of photoproducts 50 ml of pesticides solution was irradiated separately in a 100 ml UV-reactor under the same experimental conditions in order to provide an adequate concentration factor for our applications. The irradiation was carried out using a Suntest CPS + apparatus from Heraeus (Hanau, Germany) equipped with a xenon arc lamp (1500 W) and special glass filters restricting the transmission of wavelengths below 290 nm. Incident photon flow was determined by actinometric method using uranyl oxalate. The total incident light between 270 and 500 nm was 1.12 × 10−7 and 2.97 × 10−7 einstein/s for 15 and 50 ml of irradiated solution, respectively. Chamber and black panel temperature were regulated by pressurized air cooling circuit and monitored using thermocouples supplied by the manufacturer. The temperature of samples did not exceed 35◦ C using tap water cooling circuit for the UV-reactor. 2.3. Analytical procedures 2.3.1. Kinetic studies After irradiation for specific time intervals samples of 5 ml were withdrawn from the reactor. To remove the TiO2 particles the solution samples were filtered through HA 0.45 ␮m filter. After that, the samples were extracted twice with 2.5 ml n-hexane for 1 min using a vortex, dried with a small amount of Na2 SO4 and finally analyzed by GC, quantified by internal standard. The analysis of molinate and propanil was performed using Shimadzu 14A gas chromatographs equipped with flame thermionic detector (FTD) and electron capture detector (ECD 63 Ni), respectively. The DB-1 capillary column, 30 m×0.32 mm i.d. used, contained methylsilicone (J&W Scientific, Folsom, CA) followed a temperature program: 150◦ C (2 min), 150–210◦ C (5◦ C/min), 210◦ C (10 min), 210–270◦ C (10◦ C/min) and 270◦ C (3 min). The detector gases were hydrogen and air, and their flow rates were regulated at 150 and 3.5 ml/min, respectively. The ion source was an alkali metallic salt (Rb2 SO4 ) bonded to a 0.2 mm spiral of platinum wire. The temperatures were set at 220◦ C for the injector and 250◦ C for the detector. For both detectors helium was used as the

229

carrier at 25 cm/s and nitrogen was used as make-up gas at 25 ml/min. For ECD the temperatures were set at 250◦ C for the injector and 300◦ C for the detector. 2.3.2. Mineralization studies Complete mineralization of organic samples to CO2 is very often obtained by using the photocatalytic method. The kinetics of the evolution of CO2 formed was followed by using the method of Chemseddine and Boehm [28]. The evolution of inorganic ions (SO4 2− , NO3 − , Cl− ) was followed by suppressed ion chromatography, using a Shimadzu 500 apparatus equipped with a 100 mm long × 4.6 mm i.d. Shimpack IC-A1 column (Shimadzu) and a conductometric detector CDD-6A (Shimadzu). The eluent was a mixture of phthalic acid (2.5 mM) and tris(hydroxymethyl) aminomethane (2.5 mM) at a flow rate of 1.5 ml/ min. 2.3.3. By-products evaluation A relatively low amount of TiO2 (100 mg/l) was used in the experiments for the identification of organic intermediates, which enables to obtain slower kinetics and provide favorable conditions for the determination of by-products. SPE has been demonstrated to be more efficient than traditional liquid–liquid extraction (LLE) in the analysis of water samples containing very polar intermediates resulting from the photocatalytic degradation process [29,30]. The analytical procedure for SPE extraction technique followed GC–MS–EI mode has been described by our group in previous work [31]. Blank analysis helped us to discard those peaks coming from the sample handling procedure and chromatographic system. The GC was equipped with a DB-5-MS capillary column, 30 m ×0.25 mm i.d. coated with methylphenylsiloxane (5% phenyl) (J&W Scientific, Folsom, CA) used at the following chromatographic conditions: injector temperature 220◦ C, column program of temperatures 55◦ C, 55–200◦ C (5◦ C/min), 200–210◦ C (1◦ C/min), 210◦ C (2 min), 210–270◦ C (20◦ C/min) and 270◦ C (3 min). Helium was used as the carrier gas at 14 psi. The interface was kept at 270◦ C. Qualitative analyses were performed in the electron impact (EI) mode, at 70 eV potential using full scan mode.

230

I.K. Konstantinou et al. / Applied Catalysis B: Environmental 34 (2001) 227–239

Fig. 1. Disappearance of propanil. Curve A: titania in the dark (500 mg/l); curve B: UV-irradiation (λ ≥ 290 nm); curve C: UV-light with titania (100 mg/l); curve D: UV-light with titania (500 mg/l). The inset represents the semi-logarithm plots of the curves C and D.

3. Results and discussion 3.1. Photocatalytic degradation of propanil 3.1.1. Adsorption in the dark and direct photolysis It is generally expected that the observed photocatalytic degradation is governed by the adsorption kinetics. In the present case the adsorption kinetics of propanil onto TiO2 surface was followed for a period of time of 180 min. The adsorption was increased for the first 90 min reaching a plateau after that time (Fig. 1, curve A). The initial concentration of propanil decreased from 13.2 to 10.21 mg/l (i.e. from 6.08 × 10−5 to 4.71 × 10−5 mol/l) under vigorous stirring for 3 h in the dark. This corresponds to a coverage of 0.33 molecule/nm2 which represent a moderately adsorbed organic compound. Irradiation of propanil solution in the absence of TiO2 shows that photolytic decomposition occurs at a low rate

(Fig. 1, curve B) due to the low absorption of the pesticide between the 280–310 nm region. Thus, photochemical processes are scarcely responsible for the observed fast transformations when the solution was irradiated in the presence of titania. The photolysis kinetic parameters of propanil are shown in Table 1. 3.1.2. Kinetics of propanil disappearance Several experimental results indicated that the destruction rates of photocatalytic oxidation of various organic contaminants over illuminated TiO2 fitted the Langmuir–Hinshelwood kinetics [20,32–36]. The Langmuir–Hinshelwood rate form is ν=

dC kKC = dt 1 + KC

(1)

where ν is the oxidation rate of the reactant (mg/l min), C the concentration of the reactant (mg/l), t the illumination time, k the reaction rate constant (mg/l min),

Table 1 Photocatalytic degradation kinetic parameters (rate constants, correlation coefficients, half-lives) of the selected herbicides in distilled water, and in TiO2 suspensions (100 and 500 mg/l) under simulated solar light Herbicides

Photolysis kobs

Propanil Molinate

(min−1 )

0.0020 0.0032

TiO2 (100 mg/l) (min−1 )

R2

t1/2 (min)

kobs

0.951 0.951

346.6 216.6

0.0416 0.0512

TiO2 (500 mg/l) R2

t1/2 (min)

kobs (min−1 )

R2

t1/2 (min)

0.995 0.997

16.7 13.5

0.1623 0.237

0.990 0.969

4.3 2.9

I.K. Konstantinou et al. / Applied Catalysis B: Environmental 34 (2001) 227–239

and K is the adsorption coefficient of the reactant (l/g). The equation can be simplified to
 C0 (2) = kKt = kobs t or Ct = C0 e−kobs t ln C which express a pseudo-first-order reaction [6,32,37], where kobs is the apparent photodegradation rate. The kinetics of propanil disappearance in the presence of different concentrations of TiO2 (100 and 500 mg/l) followed an apparent first-order degradation curves which is consistent to the Langmuir– Hinshelwood model resulting from the low coverage in the experimental concentration range (10 mg/l) (Fig. 1, curves C and D). According to Eq. (2), linear plots of ln(C0 /C) versus time are expected (Fig. 1, inset) allowing for the calculation of rate constants kobs . Table 1 lists the values of kobs and the linear regression coefficients for pseudo-first-order kinetics of propanil photocatalytic degradation. According to these values the appropriate first-order relationship appears to fit well. The effect of the semiconductor concentration on the primary degradation rate is significant, as expected. The kobs values in the presence of 100 and 500 mg/l of TiO2 were 0.0416 and 0.1623 min−1 , respectively, confirming the positive influence of the increased number of TiO2 active sites on the process kinetics. However, this beneficial effect tends to level off and then to decrease for TiO2 concentrations

231

higher than 1–2 g/l depending on the reactor, due to relevant scattering of the incident light [9,38]. 3.1.3. Evolution of the mineralization The evolution of CO2 as a function of the irradiation time in the photocatalytic degradation of propanil is shown in Fig. 2. Taking into account the fact that complete disappearance of propanil in the irradiated reactor occurs after 40 min, whereas the stoichiometric formation of CO2 is observed after 240 min under the same working conditions, it is deduced that transient organic intermediates are likely to be present in the reaction system after the complete destruction of the tested herbicide. The evolution of chloride ions becomes quantitative at 180 min indicating that dechlorination of propanil is faster than mineralization to CO2 . Thus, only dechlorinated transient organics are present in the solution after that time. The presence of several identified chlorinated transformation products discussed below, supports the above observation. However, other studies have reported that complete dechlorination was found contemporarily with the complete disappearance of propanil [27]. Both ammonia and nitrate have been detected generally in different relative concentrations starting from compounds having nitrogen-containing groups, such as amines or nitro derivatives [22]. At long exposure times, conversion of ammonium to nitrate ions has been observed. Under the present

Fig. 2. Evolution of carbon dioxide, chloride and nitrate ions originating from propanil photocatalytic degradation under simulated solar light (TiO2 = 500 mg/l). Curves A, B and C represent the stoichiometric concentrations of CO2 , Cl− , and NO3 − , respectively.

232

I.K. Konstantinou et al. / Applied Catalysis B: Environmental 34 (2001) 227–239

experimental conditions, the quantitative recovery of nitrogen as nitrate would be achieved after long irradiation periods (more than 4 h). In 4 h time period only 32% of the stoichiometric nitrate was formed for propanil. The delayed oxidation of NH4 + to NO3 − is in agreement with previous findings concerning the photocatalytic degradation of other nitrogen containing aromatics [9,22,39]. Moreover, this conversion was found to proceed suddenly probably attributed to the occurrence of autocatalytic reactions by the nitrate ion. 3.1.4. Propanil photocatalytic degradation pathway It is well established that conduction band electrons (e− ) and valence band holes (h+ ) are generated when

aqueous TiO2 suspension is irradiated with light energy greater than its band gap energy (e.g. 3.2 eV). The photogenerated electrons could reduce the organic substrate or react with the adsorbed molecular O2 on the Ti(III)-surface, reducing it to superoxide radical anion O2 •− . The photogenerated holes can also oxidize either the organic molecule directly, or the OH− ions and the H2 O molecules adsorbed at the TiO2 surface, to • OH radicals. Together with other highly oxidant species (peroxide radicals) they are reported to be responsible for the heterogeneous TiO2 photodecomposition of organic substrates. It is known from previous photocatalytic studies that, after the formation of various aromatic derivatives, cleavage of the benzene or other organic rings

Fig. 3. GC–MS–EI total ion chromatogram obtained for a SPE extract of (a) propanil (Prl) solution and (b) molinate (Mol) solution, after 90 min of irradiation with simulated solar light in the presence of TiO2 (100 mg/l). Peaks labeled with numbers correspond to the identified degradation intermediates (see Tables 2 and 3 and Figs. 4 and 7).

I.K. Konstantinou et al. / Applied Catalysis B: Environmental 34 (2001) 227–239

takes place, and different aliphatic products are subsequently formed before complete mineralization [9,32]. The formation of such intermediates was not investigated in the present study that has focused more on the major by-products. Moreover, in other studies the maximum concentration of the main by-products appeared just after the disappearance of the parent compound [18]. As a result, the analysis of sample extracts was obtained after 90 min of irradiation in the presence of 100 mg/l TiO2 and allowed the determination of several transformation products. The analysis of extracts taken from solutions subjected to longer running photocatalytic treatment showed a reduction of the intermediate concentrations and occasionally a small increase for the by-products related to other transformation products. Fig. 3a shows the total ion chromatogram obtained for a SPE extract of propanil solution after 90 min of irradiation. Up to eight compounds could be detected as possible degradation intermediates. Six of them were unequivocally identified by (a) using an identification program of NIST library (compounds 2–5 and 8) with a fit value higher than 70% in all cases (b) by comparing the spectra with previous reported spectra (compounds 2 and 6 [40]). Two compounds (1 and 7) not included in the library were identified by interpretation of the mass spectra. Compound 1 exhibited the molecular ion peak at m/z = 233 and two peaks at m/z = 161 (base peak) and m/z = 163 that correspond to the hydroxylated derivative of propanil (addition of OH group, m/z = [M + 16]+ ) and the characteristic peaks of dichloroaniline, respectively.

233

This means that the hydroxylation takes place in the alkyl chain because the mass spectra of hydroxylated dichloroaniline provided by NIST library shows the lack of peaks at m/z = 161 and 163. Compound 7 was identified as the ring hydroxylated derivative of propanil, exhibited the peaks [M + ] = 233 and m/z = 177 (base peak), m/z = 179 that correspond to the addition of hydroxy group to propanil and the characteristic ions of the hydroxylated dichloroaniline, respectively. Compounds found at R t = 28.22 and 36.75 indicated as unknowns 1 and 2 (U1 and U2), respectively, were not identified but they could be regarded as TPs because their concentration increased and decreased as a function of the reaction time. Table 2 lists the main ions and the retention times obtained for propanil and TPs by GC–MS–EI mode. Based on the previous structure identification of the organic intermediates, a possible photocatalytic degradation pathway of propanil consisting of several steps was proposed in Fig. 4. The relatively low potential of anilides makes them susceptible to photoxidation on the TiO2 surface (2.7 V versus SCE). The reaction may involve either addition of hydroxy radicals or direct hole transfer to the organic substrate. Hydroxyl radicals attack preferentially the aromatic moiety due to their electrophilic character to form the hydroxylated products 7 and 8 but also attack the alkyl chain to form the hydroxylated product 1. The scission of the amide group via positive holes leads to the formation of 3,4-dichloroaniline (2). The dechlorinated products 3 and 5 were formed by the interaction of the trapped electrons on the semiconductor surface with

Table 2 GC–MS–EI retention times (Rt ) and spectral characteristics of propanil identified photoproducts Herbicide/photoproducts

Rt (min)

Characteristic ions (abundance, %)

Propanil Hydroxy-propanil (1a )

31.30 27.97

3,4-Dichloroaniline (2) p-Chloroaniline (3) 3-Hydroxy-4-chloroaniline (4) N-(4-chlorophenyl)-propanamide (5) 3-Hydroxy-4-chloro-propanamide (6) N-(3,4-dichloro-hydroxy-phenyl)-propanamide (7) 3,4-Dichloro-hydroxyaniline (8) U1 U2

19.87 12.93 19.25 25.69 29.93 35.24 32.63 28.22 36.75

217 (11), 163 (43), 161 (71), 57 (100) 233 (64), 231 (90), 188 (42), 161 (100), 163 (60), 109 (24), 63 (60) 163 (69), 161 (100), 126 (15), 90 (21) 129 (31), 127 (100), 92 (16) 143 (100), 145 (30), 113 (27), 78 (59) 183 (13), 129 (31), 127 (100), 57 (60) 199 (18), 201 (5), 143 (10), 145 (29), 79 (27) 233 (14), 235 (6), 177 (68), 179 (42), 57 (100) 177 (63), 179 (41), 113 (9), 57 (100) 197 (25), 141 (100), 106 (34) 200 (66), 107 (100), 94 (23)

a

Corresponding number in the proposed pathway of Fig. 4.

234

I.K. Konstantinou et al. / Applied Catalysis B: Environmental 34 (2001) 227–239

Fig. 4. Proposed photocatalytic degradation pathway of propanil in aqueous solutions by TiO2 and simulated solar light.

propanil and dichloroaniline, as reported for other pesticides [9], releasing chloride ions. The compounds 4 and 6 were formed through hydroxylation reactions of the intermediates 3 and 5. Further attack of hydroxyl radicals and positive holes lead to the formation of di-hydroxy derivatives and finally to the mineralization of propanil and its intermediates. 3.2. Photocatalytic degradation of molinate 3.2.1. Adsorption in the dark and direct photolysis The adsorption kinetics of molinate onto TiO2 surface was also followed for a period of time of 180 min. The adsorption was increased for the first 60 min reaching a plateau after that time (Fig. 5, curve A). The initial concentration of molinate was 9.79 mg/l (5.24×

10−5 ). After addition of titania (0.5 g/l) and stirring in the dark for 180 min, it decreased to 7.88 mg/l (i.e. from 5.24 × 10−5 to 4.21 × 10−5 mol/l). This corresponds to coverage of 0.25 molecule/nm2 , which is of the same order of magnitude as for propanil. Irradiation of molinate solution in the absence of TiO2 shows that photolytic decomposition occurs at a slow rate as well (Fig. 5, curve B) but it is higher than propanil. After 60 min of irradiation, direct photolysis contributed less than 23% to the degradation of molinate. The photolysis kinetic parameters of molinate are shown in Table 1. 3.2.2. Kinetics of molinate disappearance The kinetics of molinate disappearance follows also an apparent first-order degradation curve according to

I.K. Konstantinou et al. / Applied Catalysis B: Environmental 34 (2001) 227–239

235

Fig. 5. Disappearance of molinate. Curve A: titania in the dark (500 mg/l); curve B: UV-irradiation (λ ≥ 290 nm); curve C: UV-light with titania (100 mg/l); curve D: UV-light with titania (500 mg/l). The inset represents the semi-logarithm plots of the curves C and D.

the Langmuir–Hinshelwood model (Fig. 5, curves C and D). Table 1 lists the rate constants and the linear regression coefficients arising from the integral linear transformation curves (Fig. 5, inset). According to these values the appropriate first-order relationship appears to fit well. Photocatalytic experiments in the presence of 100 and 500 mg/l of TiO2 gave kobs values of 0.0512 and 0.237 min−1 , respectively, confirming

the positive influence of the increased TiO2 concentration on the process kinetics. 3.2.3. Evolution of the mineralization The kinetic variation of CO2 and the related inorganic ions is given in Fig. 6. The complete disappearance of molinate occurs after 30 min, whereas the stoichiometric formation of CO2 demands irradiation

Fig. 6. Evolution of carbon dioxide, sulfate and nitrate ions originating from molinate photocatalytic degradation under simulated solar light (TiO2 = 500 mg/l). Curves A, B and C represent the stoichiometric concentrations of CO2 , SO4 2− , and NO3 − , respectively.

236

I.K. Konstantinou et al. / Applied Catalysis B: Environmental 34 (2001) 227–239

ine isocyanate ion (m/z = 126), the base peak of molinate, and gave oxygenated hexahydroazipine isocyanate ions (m/z = 140) and weak molecular ions (m/z = 201). The photoproducts 3 and 6 were identified using the identification program of NIST library with a fit value higher than 70% in both cases. Three compounds (2, 4 and 5) not included in the library were identified by interpretation of the mass spectra. Compound 2 was identified as the de-alkylated derivative of keto-molinate and exhibited a peak at m/z = 171 by the loss of the ethyl group (M −29), corresponding to the [M − 1]+ ion that is explained by the characteristic cleavage of the S–H bond. Compound 4 with molecular ion [M]+ = 203 found at R t = 28.44 min, was associated as the sulfoxide derivative of molinate. The oxidation of S-alkyl groups to sulfoxide derivatives was also reported for other pesticides and for thioethers [31,41]. Compound 5 was identified as the de-alkylated derivative of molinate and exhibited a peak at m/z = 157 due to the loss of alkyl group, corresponding to the [M − 1]+ ion as it is explained previously. Moreover, it exhibits the same ions as molinate. Table 3 lists the main ions and the retention times obtained for molinate and TPs by GC–MS–EI mode. Compound found at R t = 23.52, indicated as U1, was not identified clearly but was regarded as TP, as the mass spectra showed common ions related with molinate and was associated as the S-isomer of molinate. The pathway of such exchange is not fully understood and explained at present thus, this compound was not included in molinate degradation pathway.

period longer than 240 min under the same working conditions. Thus, transient organic intermediates are likely to be present in the reaction system after the complete destruction of molinate. The stoichiometric concentration of SO4 2− was reached just after the complete disappearance of the parent compound in a time period of 40 min. This supports that oxidation of thio-alkyl group is a fast process and only nitrogen containing derivatives are present after that irradiation time. Under the present experimental conditions, the quantitative recovery of nitrogen as nitrate would be achieved after long irradiation periods (more than 4 h). In 4 h time period 66% of the stoichiometric nitrate was formed, i.e. twice the percent formed in the photodegradation of propanil. This is consistent to the fact that higher nitrate to ammonium concentration ratio is produced for compounds containing ring nitrogen than for compounds with the nitrogen atom bonded to the ring or in lateral chains [22]. 3.2.4. Molinate photocatalytic degradation pathway Fig. 3b shows the total ion chromatogram obtained for a SPE extract of molinate solution after 90 min of irradiation. Up to nine compounds could be detected as possible degradation intermediates. Three of them were unequivocally identified as the possible ketomolinate isomers (compounds 1(a)–(c)) by comparing the spectra with previous reported spectra for keto derivatives of molinate [26]. The ␥-keto-derivative was reported in this study for the first time (compound 1(c)). Their mass spectra lacked the hexahydroazip-

Table 3 GC–MS–EI retention times (Rt ) and spectral characteristics of molinate identified photoproducts Herbicide/photoproducts Molinate S-ethyl-pentahydro-1-keto-1H-azepine-1-carbothioate (a) ␣-Keto-molinate (b) ␤-Keto-molinate (c) ␥-Keto-molinate

Characteristic ions (abundance, %)

22.95

187 (7), 126 (55), 98 (9), 83 (16), 55 (100)

25.80 26.42 27.32

201 (18), 140 (26), 112 (100), 69 (95), 56 (78) 201 (14), 140 (25), 112 (100), 69 (31), 56 (33) 201 (15), 140 (58), 112 (100), 69 (79), 56 (72)

20.42 12.86 28.44 15.01 9.80 23.54

171 (40), 143 (7), 110 (89), 82 (100), 55 (74) 113 (5), 112 (12), 85 (55), 56 (100), 55 (90) 203 (2), 142 (99), 98 (34), 88 (100), 57 (58) 157 (14), 142 (65), 98 (16), 55 (100) 99 (17), 98 (21), 84 (12), 71 (9), 56 (100) 187 (2), 126 (81), 98 (12), 83 (27), 55 (100)

(1a )

Pentahydro-1-keto-1H-azepine-1-carbothioic acid (2) Hexahydro-2H-azepine-2-one (3) Molinate sulfoxide (4) Hexahydro-1H-azepine-1-carbothioic acid (5) Hexahydro-1H-azepine (6) U1 a

Rt (min)

Corresponding number in the proposed pathway of Fig. 7.

I.K. Konstantinou et al. / Applied Catalysis B: Environmental 34 (2001) 227–239

237

Fig. 7. Proposed photocatalytic degradation pathway of molinate in aqueous solutions by TiO2 and simulated solar light.

Fig. 7 describes some of the possible reaction pathways involving the intermediates identified by GC–MS–EI (Table 3). In reaction pathway A, the scission of the amino group via positive holes is proposed to involve the formation of an immonium radical cation followed by hydrolysis to give amine and carboxy derivatives [25]. Path B is a hydroxyl radical path. The detected intermediates are all rationalized as being formed by hydrogen abstraction, followed by the addition of oxygen to the alkyl radical and decomposition of the peroxyl radical formed

by C–O or C–C bond cleavage [26]. Keto-molinate (1) was the formed photoproduct and a higher preference for abstraction from the weaker ␣-amino C–H bond versus the inactivated ␤- and ␥-C–H bond will be expected. However, positions ␤ and ␥ also react significantly. The possible complexation of molinate to the semiconductor surface through the nitrogen atom expose all the hydrogens of the N-alkyl chain to the rather unselective attack by surface-bound hydroxyl groups [26]. The oxidized derivative at the position ␥ (4-keto-molinate) was found also as the

238

I.K. Konstantinou et al. / Applied Catalysis B: Environmental 34 (2001) 227–239

oxidation metabolite of molinate in runoff water [42]. The continuous attack of positive holes leads to the formation of dealkylated product (2) and keto-azepine (3). In path C finally, oxidation of S-alkyl group takes place by positive holes attack leading to the sulfoxide derivative (4) and the dealkylated product (5). The continuous attack by positive holes leads to the formation of azepine molecule (6) and finally to the complete mineralization. The great number of compounds detected during the degradation of the above compounds shows the complexity of the photocatalytic process and suggests the existence of various degradation routes resulting in multi-step and interconnected pathways. Costeffective treatments to complete compound mineralization are usually not practicable, and the presence of by-products during and at the end of the water treatment appears to be unavoidable.

4. Conclusions This work shows that heterogeneous titanium dioxide photocatalyzed decomposition is a feasible and fast treatment for the removal of the selected herbicides present at milligram per liter level from agricultural effluents or polluted surface waters. The degradation of the herbicides follows first-order kinetics and complete disappearance is observed in less than 1 h, working with moderate amounts of TiO2 , under simulated solar light. A careful study of the parameters, such as CO2 , inorganic ions formation and identification of by-products, obtained in combination with analytical procedures based on mass spectroscopy techniques provides a thorough knowledge of the degradation mechanism. The analysis of the transformation products reveals the presence of hydroxylation, dechlorination, dealkylation and oxidation steps leading to the mineralization of the starting molecule. References [1] T.A. Albanis, D.G. Hela, T.M. Sakellarides, I.K. Konstantinou, J. Chromatogr. A 823 (1998) 59. [2] T.A. Albanis, S. Galassi, Assessment of the state of pollution of the Mediterranean Sea by herbicides and proposed measures, UNEP Report, 1995.

[3] W. Huang, C.S. Hong, B. Bush, Chemosphere 32 (1996) 1869. [4] K. Banerjee, P.N. Cheremisinoff, S.L. Cheng, Environ. Sci. Technol. 29 (1995) 2243. [5] G. Mascolo, A. Lopez, R. Foldenyi, R. Passino, G. Tiravanti, Environ. Sci. Technol. 29 (1995) 2987. [6] K.H. Wang, Y.H. Hsieh, M.Y. Chou, C.Y. Chang, Appl. Catal. B: Environ. 21 (1999) 1. [7] C. Zwiener, L. Weil, R. Niessner, Int. J. Anal. Chem. 58 (1995) 247. [8] E. Pelizzetti, Solar Energy Mater. Solar Cells 38 (1995) 453. [9] A. Bianco Prevot, M. Vincenti, A. Banciotto, E. Pramauro, Appl. Catal. B: Environ. 22 (1999) 149. [10] D. Bahnemann, J. Cunningham, M.A. Fox, E. Pelizzetti, P. Pichat, N. Serpone, in: G.R. Heltz, R.G. Zepp, D.G. Grosby (Eds.), Aquatic and Surface Photochemistry, Lewis Publishers, Boca Raton, FL, 1994, p. 261. [11] J.M. Hermann, Catal. Today 53 (1999) 115. [12] E. Pelizzetti, E. Pramauro, C. Minero, N. Serpone, E. Borgarello, in: M. Schiavello (Ed.), Photocatalysis and Environment — Trends and Applications, NATO ASI Series, Kluwer Academic Publishers, Dordrecht, The Netherlands, 1988, p. 469. [13] C. Richard, S. Bengana, Chemosphere 33 (1996) 635. [14] W.Z. Tang, C.P. Huang, Water Res. 29 (1995) 745. [15] J.P. Percherancier, R. Chapelon, B. Poyet, J. Photochem. Photobiol. A: Chem. 72 (1995) 261. [16] A. Zaleska, J. Hupka, M. Wiergowski, M. Biziuk, J. Photochem. Photobiol. A: Chem. 135 (2000) 213. [17] J.M. Hermann, J. Disdier, P. Pichat, S. Malato, J. Blanco, Appl. Catal. B: Environ. 17 (1998) 15. [18] J.M. Hermann, C. Guillard, M. Arguello, A. Aguera, A. Tejedor, L. Piedra, A. Fernadez-Alba, Catal. Today 54 (1999) 353. [19] I. Poulios, M. Kositzi, A. Kouras, J. Photochem. Photobiol. A: Chem. 115 (1998) 175. [20] C.S. Turchi, D.F. Ollis, J. Catal. 119 (1989) 483. [21] L.P. Childs, D.F. Ollis, J. Catal. 66 (1980) 383. [22] G.K.C. Low, S.R. Matthews, R.W. McEvoy, Environ. Sci. Technol. 25 (1991) 460. [23] E. Pelizzetti, V. Carlin, C. Minero, E. Pramauro, M. Vinceti, Sci. Total Environ. 123/124 (1992) 161. [24] T.F. Chen, R. Doong, W.G. Lei, Water Sci. Technol. 37 (1998) 187. [25] A. Vidal, Z. Dinya Jr., F. Mogyorodi, Appl. Catal. B: Environ. 21 (1999) 259. [26] M. Sturini, E. Fasani, C. Prandi, A. Casaschi, A. Albini, J. Photochem. Photobiol. A: Chem. 101 (1996) 251. [27] M. Sturini, E. Fasani, C. Prandi, A. Albini, Chemosphere 35 (1997) 931. [28] A. Chemseddine, H.P. Boehm, J. Mol. Catal. 60 (1990) 295. [29] A. Aguera, E. Almansa, A. Tejedor, A.R. Fernadez-Alba, S. Malato, M.I. Maldonado, Environ. Sci. Technol. 34 (2000) 1563. [30] A.R. Fernadez-Alba, A. Aguera, M. Contreras, G. Penuela, I. Ferrer, D. Barcelo, J. Chromatogr. A 823 (1998) 35. [31] I.K. Konstantinou, T.M. Sakellarides, V.A. Sakkas, T.A. Albanis, Environ. Sci. Technol. 35 (2001) 398. [32] J.C. D’Oliveira, G. Al-Sayyed, P. Pichat, Environ. Sci. Technol. 4 (1990) 990.

I.K. Konstantinou et al. / Applied Catalysis B: Environmental 34 (2001) 227–239 [33] D.F. Ollis, C.Y. Hsiao, L. Budiman, C.L. Lee, J. Catal. 88 (1984) 89. [34] K.I. Okamoto, Y. Yamamoto, H. Tanaka, A. Itaya, Bull. Chem. Soc. Jpn. 58 (1985) 2023. [35] H. Al-Ekabi, N. Serpone, J. Phys. Chem. 92 (1988) 5726. [36] A. Topalov, D. Molnar-Gabor, J. Csanadi, Water Res. 33 (1999) 1371. [37] P. Fernadez-Ibanez, S. Malato, F.J. de las Nieves, Catal. Today 54 (1999) 195.

239

[38] A. Assabane, Y.A. Ichcu, H. Tahiri, C. Guillard, J.M. Hermann, Appl. Catal. B: Environ. 24 (2000) 71–87. [39] E. Pramauro, M. Vincenti, V. Augugliaro, L. Palmisano, Environ. Sci. Technol. 27 (1993) 1790. [40] K.W. Moilanen, D.G. Crosby, J. Agric. Food Chem. 20 (1972) 950. [41] M.A. Fox, A.A. Abdel-Wahad, Tetrahedron Lett. 31 (1990) 4533. [42] J. Domagalski, Water Res. Bull. 32 (1996) 953.