Formation of Refractory Ring-Expanded Triazine Intermediates during

Formation of Refractory. Ring-Expanded Triazine. Intermediates during the. Photocatalyzed Mineralization of the. Endocrine Disruptor Amitrole and. Rel...
0 downloads 0 Views 215KB Size
Download PDF
Environ. Sci. Technol. 2005, 39, 2320-2326

Formation of Refractory Ring-Expanded Triazine Intermediates during the Photocatalyzed Mineralization of the Endocrine Disruptor Amitrole and Related Triazole Derivatives at UV-Irradiated TiO2/H2O Interfaces NATSUKO WATANABE, SATOSHI HORIKOSHI, ATSUSHI KAWASAKI, AND HISAO HIDAKA* Frontier Research Center for the Global Environment Science, Meisei University, 2-1-1 Hodokubo, Hino, Tokyo 191-8506, Japan NICK SERPONE* Dipartimento di Chimica Organica, Universita di Pavia, Via Taramelli 10, Pavia 27100, Italia

Amitrole (ATz, 3-amino-1H-1,2,4-triazole) is a widely employed herbicide with strong estrogenic activity that can lead to abnormalities of the thyroid gland and can cause mutations. The photocatalytic transformation of ATz was carried out at the UV-irradiated TiO2/H2O interface, along with the triazole derivatives Tz (1H-1,2,4-triazole) and DaTz (3,5-diamino-1H-1,2,4-triazole) to assess the decomposition of these herbicides, to identify intermediates, and to elucidate some mechanistic details of the ATz degradation. Conversion of the nitrogens of these triazoles to NH4+ and/ or NO3- ions occurs competitively and depends on the number of amine functions on the five-membered triazole rings. Photomineralization of the substrates in terms of loss of nitrogen to NH4+/NO3- was rather low (ca. 2540%) for each of the triazoles, whereas evolution of CO2 (loss of TOC) was more significant (60-70%), indicating considerable retention of nitrogen in the intermediate products. UV-Vis spectroscopy, TOC assays, FT-IR spectroscopy, proton NMR spectrometry, electrospray LCMS, and molecular orbital calculations were brought to bear in assessing the temporal course of the photocatalyzed process(es). Results show that after cleavage of the triazole ring, the various intermediate fragments recombine to yield ring-expanded six-membered triazine intermediates, which slowly degrade to give the refractory cyanuric acid under the conditions used.

Introduction Amitrole (3-amino-1H-1,2,4-triazole) is commonly used as a herbicide, as a chemical reagent in the photographic industry, and as a hardening agent in chemical resins. This * Address correspondence to either author. E-mail: [email protected] (H.H.); [email protected] (N.S.). 2320

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 7, 2005

endocrine disruptor is highly toxic to animals because of its estrogenic activity (1), the damage it causes to the thyroid gland (2, 3), and its ability to cause mutations (4). Studies on Advanced Oxidation Technologies (AOT) with TiO2 as the photocatalyst have been reported extensively toward making the environment eco-friendly (5). AOTs have proven to be attractive degradation techniques for several organic pollutants leading to facile mineralization to CO2 gas in most cases, except for s-triazines such as atrazine (6). Organic polluting products containing nitrogen atoms are typically converted to NH4+ and NO3- ions by the TiO2 photocatalytic degradation methods (7-14). A study published recently by Catastini et al. (15) showed that amitrole is degraded following photoexcitation of iron(III) aquacomplexes to yield two principal photoproducts: viz., 5-hydroxyamitrole and 2-aminothiazole (urazole) with the former disappearing rapidly with irradiation time. Nearquantitative mineralization of amitrole (0.2 mM) required 160 h (i.e., nearly 1 week) of exposure to artificial UV irradiation, in which all nitrogen atoms were converted to NH4+ ions. We herein report the photodecomposition of the endocrine disruptor amitrole (ATz) in aqueous TiO2 dispersions, together with the related triazole compounds 1H-1,2,4triazole (Tz; no amino group) and 3,5-diamino-1H-1,2,4triazole (DaTz; two amino groups) to assess the extent of mineralization and to aid in delineating possible mechanistic details.

Ring-expanded six-membered triazine substrates formed that proved refractory to further photodegradation under the prevailing conditions used, but nonetheless interesting as a first example of ring expansion in photocatalytic processes. One class of herbicides, the triazoles, was converted to another class of herbicides, the s-triazines. This latter aspect emphasizes the need for a proper choice of AOT methods when wishing to degrade certain pollutant feeds and the need to identify the nature of the intermediate species during the photomineralization process in every instance.

Experimental Section Materials. Amitrole (3-amino-1H-1,2,4-triazole: ATz), 1H1,2,4-triazole (Tz), and 3,5-diamino-1H-1,2,4- triazole (DaTz) were supplied by the Wako Pure Chem. Co. Ltd as highly pure-grade reagents; they were used as received. Titanium dioxide was Degussa P-25 (particle size 20-30 nm, transmission electron microscopy; surface area 53 m2 g-1, BET methods; crystal structure 83% anatase and 17% rutile, X-ray diffraction analysis). Aqueous triazole solutions (0.10 mM, 50 mL) were placed in a 127-mL Pyrex vessel containing TiO2 particles (100 mg; loading, 2.0 g L-1) followed by supersonication to obtain a uniform dispersion. The latter was subsequently purged with oxygen gas prior to UV irradiation with a 75-W Hg lamp that emitted ca. 2.0 mW cm-2 at wavelengths 310-400 nm (maximal emission, 360 nm). Analytical Procedures. Ring opening and decrease of total organic carbon (TOC) during the photodecomposition of ATz, DaTz, and Tz were monitored with a JASCO UV-Visible 10.1021/es049791l CCC: $30.25

 2005 American Chemical Society Published on Web 02/09/2005

FIGURE 1. Variations in the temporal UV spectral patterns during the photodegradation of Tz, ATz, and DaTz (0.1 mM) triazoles in aqueous TiO2 dispersions. spectrophotometer (V-570) and a Shimadzu TOC-5000A analyzer, respectively. Formation of NH4+ and NO3- ions was assayed on a JASCO high-pressure liquid chromatographic system equipped with a CD-5 conductivity detector and either a Y-521 cationic column or an I-524 anionic column. Intermediate products were detected and identified by mass spectrometric methods using electrospray ionization (Agilent Technologies HP1100 series ESI-LC/MSD). Fourier transform infrared spectroscopy (FT-IR) (KBr method; JASCO 620 spectrophotometer) and 1H nuclear magnetic resonance methods (JEOL JNM-AL 300-MHz spectrometer) were used to confirm the nature of the intermediates. Molecular orbital calculations of frontier electron densities and partial charges on all the atoms of the herbicide structures above were carried out at the single determinant (HartreeFock) level for optimization with the minimum energy obtained at the AM1 level. All semiempirical calculations were performed using the MOPAC version 6 in the CAChe package (Fujitsu Co. Ltd) (16). Initial positions of •OH radical attack were inferred from the electron densities, whereas possible modes of closest approach of ATz, DaTz, and Tz molecules on the TiO2 particle surface can be inferred from the partial charges.

Results and Discussion Photodegradations. When irradiated with UV light at energies greater than 3.2 eV (i.e., at wavelengths shorter than 387 nm), the TiO2 particles generate electron (e-)/hole (h+) pairs, after which trapping of holes by surface OH- ions or reactions with H2Oads on the TiO2 particle surface leads to formation of •OH radicals (and H+). Concomitantly, pre-adsorbed molecular oxygen can scavenge conduction band electrons (e-CB) to yield superoxide radical anions (O2•-), which in acidic media convert to hydroperoxy (•OOH) radicals. Although possible under certain circumstances, direct oxidation by valence band holes is not considered likely under our conditions. Temporal changes in the UV spectral patterns during the photocatalyzed degradation of Tz, ATz, and DaTz are portrayed in Figure 1. Maximal absorptions by the initial solutions were observed at 193 nm for Tz, at 196 nm for ATz, and at 192 and 207 nm for DaTz. Following UV irradiation of the appropriate dispersions, the UV absorption spectral features showed some unusual behavior for ATz and DaTz

FIGURE 2. Plots illustrating the temporal course of the formation of (a) NH4+ ions and (b) NO3- ions, (c) the temporal decrease of TOC, and (d) the temporal changes in pH during the photooxidation of Tz, ATz, and DaTz triazoles at TiO2/H2O interfaces. during the first 3 h of illumination. For the Tz triazole the absorption increased in intensity throughout the 24-h irradiation period. After this time, all three triazole derivatives displayed nearly identical spectra, inferring the formation of very similar if not identical intermediate species. We hasten to point out that such spectral behavior is not the result of solubility effects on the triazole derivatives during their photodegradation. Figure 2, panels a and b, depict respectively the temporal evolution of NH4+ and NO3- ions during the photocatalyzed conversion of the Tz, ATz, and DaTz triazoles. Photogeneration of NH4+ ions predominated during the initial stages of the photocatalytic degradation of ATz and DaTz. For the latter two triazoles, formation of NO3- ions necessitated a 2and 3-h induction period, respectively, before these ions could be detected and identified. Subsequently, formation of nitrate ions proceeded somewhat rapidly reaching a plateau beyond the 24-h illumination period. Table 1 reports the relevant first-order kinetics of formation of NH4+ and NO3- ions as well as the kinetics of loss of total organic carbon (TOC) (i.e., evolution of CO2) from the dispersion. Conversion of the ring nitrogens in the Tz triazole ring to ammonium and nitrate ions occurred with NH4+ ions being formed nearly 2-fold faster than NO3- ions. Differences in reactive behavior between Tz and those of ATz and DaTz may be deduced from differences in the chemical structure of the triazoles (see above). Tz has no amine function, so that formation of NH4+ and NO3- ions must of necessity follow from cleavage of the triazole ring. By contrast, both ATz and the DaTz structures possess amine functions. Hence, their conversion to NH4+ ions does not necessarily implicate cleavage of the five-membered triazole rings. However, once VOL. 39, NO. 7, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2321

TABLE 1. First Order Kinetics of Formation of NH4+ and NO3Ions and of Decay of Total Organic Carbon (TOC initial, 2.40 ppm) during the Photocatalyzed Degradation of Tz, ATz, and DaTz Triazoles (initial pH ) 5) parameters

Tz

ATz

DaTz

kNH4+ (h-1) induction period kNO3- (h-1) induction period kTOC (h-1) induction period [NH4+],a mM [NO3-],a mM TOC,a ppm

0.60 ( 0.09 none 0.31 ( 0.04 ∼0 h 0.77 ( 0.12 none 0.022 0.057 1.68 (70%)b

0.43 ( 0.04 none 0.18 ( 0.01 ∼2 h 0.22 ( 0.07 none 0.035 0.094 1.45 (60%)b

0.59 ( 0.04 none 0.13 ( 0.01 ∼3 h 0.17 ( 0.04 none 0.079 0.12 1.59 (65%)b

Quantities of NH4+ and NO3- ions formed and TOC decayed after 24 h of UV irradiation. b Extent of mineralization. a

these functions are converted to ammonium ions, cleavage of the rings must then occur followed by competitive conversion of the ring nitrogens to ammonium and nitrate ions. After an irradiation period of 24 h, the quantity of NH4+ ions produced for each substance followed the order DaTz (0.079 mM) . ATz (0.035 mM) > Tz (0.022 mM); a similar order is obtained for the NO3- ions: DaTz (0.12 mM) > ATz (0.094 mM) > Tz (0.057 mM). Clearly formation of nitrate ions is overall favored in the triazole degradations, aided by oxidation of NH4+ to NO3- at the long irradiation times. The ratio of the quantities of NH4+ ions produced from the conversion of DaTz and ATz, respectively, is nearly 2:1, consistent with the number of amine substituents on the DaTz ring (see structures above) relative to that of ATz. The extent of photoconversion of the triazole nitrogen atoms into NH4+ and NO3- ions for DaTz, ATz, and Tz amounts to 40%, 32%, and 26%, respectively. We deduce that the remaining nitrogen load remains in the nondegradable organic intermediates produced under our experimental conditions (see below). Note that no formation of N2 gas was detected by gas chromatographic techniques during the whole course of the degradation of all three triazoles. The disappearance of TOC during the photooxidative degradation of DaTz, ATz, and Tz is illustrated in Figure 2c. The quantity of initial total organic carbon (2.40 ppm; 0.10 mM) decreased to about 0.81, 0.95, and 0.72 ppm, respectively, for DaTz, ATz, and Tz, corresponding to a photomineralization yield in the ring carbon atoms of 65%, 60%, and 70% after 24 h of UV irradiation (Table 1). The rates of the temporal decrease of TOC and the rates of conversion of the nitrogens into NH4+ and NO3- ions reported in Table 1 show that the order of mineralization is approximately Tz > ATz ≈ DaTz. The photooxidative pathway to achieve this level of carbon dioxide evolved and ring nitrogen atoms converted requires cleavage of the ring and photooxidation of the intermediate species formed. MOPAC simulations have been used in the past as useful tools to infer the position of •OH radical attack and the mode of adsorption (if any) of substrates on the TiO2 surface in elucidating some possible initial stage of the degradation mechanisms. The TiO2 particle surface is positively charged owing to an excess of protons from the photooxidation of H2O and/or to the acidic nature of the initial solution/ dispersion. Accordingly, the most negatively charged atom(s) in the structures of Tz, ATz, and DaTz is expected to interact with TiO2 surface through Coulombic forces, however weak. Enthalpies of physisorption and chemisorption of some substrates on metal oxides have been reported by Emeline and co-workers (17); they range around 30-50 kJ mol-1. In this regard, a molecule will face the positively charged metal oxide surface (at pH < 6) with the closest approach favoring 2322

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 7, 2005

TABLE 2. Calculated Frontier Electron Densities and Partial Charges on All Atoms in the Structures of Tz, ATz, and DaTza radical frontier electron densities

partial charges

atoms

Tz

ATz

DaTz

Tz

ATz

DaTz

N1 N2 C3 N4 C5 N6 N7

0.515 0.390 0.428 0.193 0.470

0.416 0.305 0.159 0.214 0.400 0.494

0.436 0.345 0.201 0.055 0.319 0.270 0.332

0.517 -0.483 0.004 -0.408 -0.167

0.494 -0.636 0.024 -0.525 -0.125 0.276

0.252 -0.648 0.050 -0.566 -0.070 0.264 0.240

a Although calculated, those of hydrogen atoms are not reported. The largest values are in boldface type.

those atoms of the molecule bearing the greater partial negative charge. Calculated frontier electron densities and partial charges for all atoms in the three triazole structures are summarized in Table 2 (those for H atoms are not listed). The most negatively charged atoms are the N2 nitrogens, followed by the N4 atoms. Accordingly, we deduce that the N2 atoms will interact preferentially with the TiO2 particle in all three structures. There exists a large body of evidence that the •OH radicals photogenerated on the UV-irradiated TiO2 surface are the major oxidative agents that oxidize adsorbed substrates. Consequently, the position(s) of •OH radical attack in the triazole structures will be the atom bearing the higher frontier electron density in the DaTz, ATz, and Tz structures. Table 2 shows that the atom with the greater frontier electron density for Tz is the N1 (and C5) atom, for the ATz triazole it is the N1 and N6 atoms, whereas for DaTz it is the N1 atom. Thus, we deduce that the initial •OH radical attack for the Tz substrate should occur preferentially at the N1 (and possibly also at C5) position; for the ATz triazole the attack should occur at the N1 and N6 positions, and for DaTz the prevalent •OH radical attack should also take place at the N1 position. In addition, closer approach of triazoles through the N2 atoms should favor attack at the N1 nitrogens. Figure 2d illustrates the temporal pH changes during the photodegradation of DaTz, ATz, and Tz under UV illumination. The pH of all the initial solutions was ca. pH 5. The isoelectric point of a TiO2 (P-25) aqueous dispersion is at pH 6.3 determined by a ζ-potential analyzer (18). In the first 2 h of irradiation of the dispersion containing the triazoles ATz and DaTz, the positive charge on the TiO2 surface (TiOH2+) decreased toward a more neutral TiO2 surface (TiOH) with increase in pH. For the Tz triazole, however, the TiO2 surface remained positively charged throughout the 24-h period, whereas for the ATz and DaTz systems the surface contained both Ti-OH2+ and Ti-O- terminal groups within 3-8 h of irradiation. Accordingly, we deduce that any adsorption of ATz and DaTz became more significant after irradiation was initiated. pH Dependence of the Photoconversion of Nitrogen Atoms. Table 3 reports the pH dependence of formation of ammonium and nitrate ions during the photocatalyzed degradation of ATz at three different pH values from an initial acidic medium (pH 3) to an alkaline medium (pH 11). The quantity of NH4+ ions formed increased 2-fold from pH 3 to pH 5 for all three triazoles but dropped significantly at pH 11 at 8- and 24-h irradiation times. It is also evident that although formation of NH4+ ions predominates initially, the process favors conversion of nitrogens into NO3- ions at longer irradiation times, the process being more efficient in alkaline media (pH 11). The 24-h data set in Table 3 calls attention to oxidation of ammonia to nitrate ions at the longer

TABLE 3. pH Dependence of the Quantity of NH4+ and NO3Ions Formed in the Photodecomposition of ATz at Three Different pH Values pH 3

pH 5

pH 11

time NH4+ NO3- [NO3-/ NH4+ NO3- [NO3-/ NH4+a (h) (µM) (µM) NH4+] (µM) (µM) NH4+] (µM) 3 8 24 a

14.7 11.2 18.8 29.4 19.6 48.5

0.76 1.56 2.47

27.4 8.9 31.8 64.6 38.3 89.6

0.32 2.03 2.34

NO3- [NO3-/ (µM) NH4+]

52.4 52.1 7.89 14.6 0 175

0.99 1.85

After acidification prior to analysis.

times of irradiation. The increase in the ratio [NO3-/NH4+] from pH 3 to pH 11 is consistent with predictions from a recent theoretical modeling study by Emeline et al. (19) on the activity and selectivity in heterogeneous photocatalysis. The model, which included electric field effects (i.e., surface potentials/charges), predicted that when the surface of the photocatalyst specimen is negative, as it would be herein at pH 11 for TiO2 particles, the valence band holes photogenerated in the bulk lattice of the metal oxide will drift toward the surface, whereas the electrons photogenerated at the surface and in the bulk will drift away from the surface into the bulk lattice. Accordingly, surface oxidation reactions are expected to predominate. Clearly, greater conversion of the triazole nitrogens to NO3- ions should predominate over formation of NH4+ ions at the more alkaline pH values, as observed in Table 3. Mass Spectral Analyses and Identification of Intermediates. The photomineralization of Tz, ATz, and DaTz triazoles was incomplete after 24 h of irradiation, indicating that within this period nondegradable intermediates were formed. The initial solutions and the intermediates produced were analyzed by electrospray LC-MS techniques both in the positive-ion mode ({species + H+} or {species + Na+}) and the negative-ion mode {species - H+}. Molecular mass numbers of the initial substrates and the intermediates produced from the photooxidation of Tz, ATz, and DaTz are displayed in Table 4. The Tz, ATz, and DaTz substrates appeared in the positive mode at masses m/z ) 70, 85, and

100 and in the negative mode at m/z ) 68, 83, and 98, respectively. Survival of Tz lasted for up to ca. 2 h of irradiation, a period which accorded with the time at which formation of NH4+ and NO3- ions increased significantly for the Tz triazole and nearly so for the ATz and DaTz substrates in the formation of NH4+ ions (see Figure 2a,b), whereas the extent of loss of TOC was most pronounced (Figure 2c). After this time, the rates of evolution of ammonium and nitrate ions slowed considerably. For the ATz and DaTz systems the survival time of ca. 3 h of irradiation was in line with formation of nitrate ions and with the decrease of pH (for the ATz; Figure 2d) after the 3-h induction period. Accordingly, we infer that NO3- ions were formed from the photodegradation of the initial intermediates, at least for the decomposition of ATz. At longer irradiation times (>2-3 h), several intermediate species with large mass numbers were detected and identified in both the positive- and negative-ion modes. Their identities are formulated in Scheme 1a-c. Mass spectral analyses of the intermediates for amitrole ATz during the photodegradation carried out at pH 3 and at pH 11 revealed significant differences in the distribution of intermediates formed. At the alkaline pH of 11, fewer intermediates were detected probably because some of the intermediates produced underwent faster oxidation than what might occur at lower pH values. The mass spectral observations at these two pH values are again consistent with expectations from our earlier theoretical model (19). Identification of Nondegradable Intermediates. The FTIR spectral patterns of Tz, ATz, and DaTz triazoles are compared with those recorded after the 24-h irradiation period for the remaining solutions in Figure 3. After 24 h of illumination the solutions were freeze-dried and their FT-IR spectra recorded by the KBr disk method. Except for the coincident IR peaks at 825, 1384, 2822, and 3135 cm-1, the IR spectral patterns were quite different from those of the original triazoles, suggesting that common intermediates with a ring structure were produced from the triazole rings. The band at 3135 cm-1 is consistent with vibrational modes of the -OH and -CONH moieties. The vibrational mode of the -CdN- function in the triazole ring was observed at 1722 cm-1, whereas the -C-NO2 vibration is tentatively assigned

TABLE 4. Positive- and Negative-Ion ESI Mass Numbers (m/z) in Photodecomposition of Tz, ATz, and DaTz after Indicated Period of UV Illumination of Corresponding TiO2 Dispersions (initial pH 5)a time (h)

Tz (m/z)

ATz (m/z)

0 0.5 1 2

68(n), 70(p) 68(n), 70(p) 61(p), 70(p), 73(p), 74(p) 61(p), 70(p), 73(p), 74(p), 210(n), 244(n) 60(p), 61(p), 73(p), 74(p), 210(n), 244(n) 60(p), 61(p), 73(p), 74(p), 210(n), 244(n)

83(n), 85 (p) 83(n), 85 (p) 83(n), 85(p), 99(n), 128(n) 66(n), 127(n), 128(n), 152(n)

5

60(p), 61(p), 73(p), 210(n), 244(n)

86(n), 127(n), 152(n), 210(n)

6

60(p), 61(p), 73(p), 210(n), 244(n)

86(n), 127(n), 152(n), 210(n)

8

60(p), 61(p), 73(p), 210(n), 244(n)

73(p), 74(p), 86(n), 127(n), 210(n)

10

60(p), 61(p), 73(p), 210(n), 244(n)

73(p), 74(p), 86(n), 127(n), 210(n)

24

60(p), 73(p), 210(n), 244(n)

127(n), 210(n)

3 4

a

66(n), 76(p), 83(n), 86(n), 152(n) 66(n), 86(n), 152(n), 210(n)

DaTz (m/z) 98(n), 100(p) 84(n), 98(n), 100(p) 59(p), 74(p), 84(n), 89(p), 98(n), 100(p), 118(p), 130(p) 59(p), 74(p), 84(n), 100(p), 118(p) 59(p), 60(p), 74(p), 84(n), 89(p), 128(n), 155(n), 210(n) 59(p), 74(p), 84(n), 86(n), 118(p), 128(n), 210(n) 59(p), 60(p), 61(p), 74(p), 85(p), 86(n), 89(p), 128(n), 130 (p), 210(n) 59(p), 61(p), 74(p), 86(n), 89(p), 128(n), 130(p), 155(n), 210(n) 59(p), 60(p), 61(p), 74(p), 86(n), 130(p), 210(n) 59(p), 74(p), 86(n), 89(p), 128(n), 130(p), 210(n)

(p), positive-ion mode; (n), negative-ion mode.

VOL. 39, NO. 7, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2323

SCHEME 1. Proposed Mechanisms Based on Identification of Intermediates Formed Chronologically during the Photodegradation of (a) Tz, (b) ATz, and (c) DaTz under UV Illumination in Aqueous TiO2 Dispersions (initial pH 5) by the Electrospray Ionization LC-MS Mass Spectral Methods in both Negative-Ion and Positive-Ion Modes

to the IR band at 1384 cm-1. The band at 825 cm-1 is that of the -NH2 function. Note that although this 825 cm-1 band is seen for the initial ATz and DaTz substrates, it is absent 2324

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 7, 2005

in the Tz spectrum. Except for the 825 cm-1 band, many of the IR signals between ca. 1200 and 400 cm-1 attributable to C-H vibrations of the triazole rings and seen in the spectra

We presume that it is most likely the proton in the N-CHdN group of the triazine ring (see structure below). On the basis of the proton NMR and FT-IR spectral results and from the LC-MS data of Table 4, the following structure is suggested for one of the intermediates with m/z ) 210 (negative-ion mode; n), which appears to be formed not only in the photodegradation of the amitrole (ATz) substrate but also during the photodegradation of the other two triazole derivatives.

FIGURE 3. FT-IR spectral patterns (KBr method) recorded in the photodecomposition of Tz, ATz, and DaTz after 0 and 24 h (freezedried) under UV illumination.

FIGURE 4. 1H NMR spectral patterns in the photodegradation of ATz after 24 h of irradiation in DMSO and D2O solvents. of the initial substrates were no longer significant after the 24-h irradiation period. The photooxidized product(s) of the ATz solution was analyzed by 1H NMR spectroscopy in DMSO and D2O solvents (Figure 4). The photodegraded ATz solution obtained after 24 h of UV illumination was freeze-dried and then dissolved in DMSO and D2O solvents. Proton NMR resonance signals were evident at 0.84 ppm (A), 1.23 ppm (B), 6.51 ppm (C), and 7.06 ppm (D) in the DMSO solvent relative to the TSP standard at 0 ppm. The broad signals A and D are attributed to the protons attached to the nitrogen atoms (quadrupole broadening) in the ATz structure; that is, signal A is assigned to the -NH2 function, whereas signal D is due to the protons of the triazine ring. The singlet peaks at 1.23 ppm (B) and 6.51 ppm (C) originate from the protons attached to the carbon atoms of the ring; that is the proton belonging to an aldehyde group was seen at 1.23 ppm (B), whereas the signal at 6.51 ppm (C) is that of the proton located between two nitrogen functions. The NMR peak of this proton is observed as a singlet signal at 7.81 ppm (E) in D2O solvent and is the same as the hydrogen atom in either the N-CHdN and/or N-CH2-N groups of the triazine or triazole ring, respectively.

Note that the m/z ) 244(n) intermediate triazine species is formed only in the photomineralization of the Tz triazole. The 210 (m/z) species consists of a triazine ring with three substituent groups about the ring carbons. The building blocks of this species likely originate from the condensation of small nitrogen-containing intermediates involving radical recombination reactions. In this regard, the cyanuric acid composed of a triazine skeleton can generally be produced by condensation of three urea molecules. To the extent that formation of urea was observed in the present study, as evidenced by the LC-MS results of Table 4, we carried out a control experiment to examine the fate of urea (0.1 mM) under otherwise identical conditions to those of the triazoles. No intermediate(s) was detected that contained a triazine ring, confirming our suspicion that the species with m/z ) 210(n) must indeed arise from more complex events involving the many fragments formed in the breakup of the triazole rings (see Scheme 1a-c). Over a decade ago, Pelizzetti et al. (6) and other researchers (20, 21) reported that the triazine skeleton, such as that of cyanuric acid, is refractory to TiO2 photocatalytic degradation conditions. By contrast, refractory triazines can easily be decomposed in supercritical water media (22, 23), and apparently cyanuric acid can be photodegraded in fluorinated-TiO2 aqueous dispersions at low pH values and by the Fenton process according to a recent report by Oh and Jenks (24). The earlier study (6) indicated that the Fenton reaction led only to negligible, if any, degradation of cyanuric acid. Our recent study on the degradation of this refractory substrate, under conditions otherwise similar to those of ref 24, with the Fenton reagent and in fluorinated Degussa P25 TiO2 oxygenated dispersions showed no degradation of cyanuric acid even after 6 days of continuous UV irradiation, as evidenced by HPLC/UV and HPLC/IC methods, and by electrospray mass spectral analyses in the positive- and negative-ion modes through the m/z ) 130(p) and 128(n) mass peaks of cyanuric acid (25). From the discussion above, it is not surprising that the complete photodegradation of Tz, ATz, and DaTz triazoles did not fully materialize owing to formation of refractory triazine intermediates (see below), at least under the present experimental conditions. All the control experiments confirmed this assertion. For instance, irradiation of the triazole/ TiO2 dispersions for 17 h with a 200-W mercury lamp produced an intermediate with mass spectral peaks at m/z ) 128(n) and 130(p). Accordingly, formation of cyanuric acid (molecular mass, 129) originates from the triazine intermediates produced from the degradation of the triazoles (Scheme 1a-c): for example, from the m/z ) 210(n) species and others that seem to degrade slowly owing to other VOL. 39, NO. 7, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2325

competing events under the conditions used earlier for the triazoles.

with nonmodified TiO2 particles to terminate at the refractory cyanuric acid system.

The mechanisms proposed for the photodegradation of Tz, ATz, and DaTz triazoles are illustrated in Scheme 1a-c, respectively. The initial step in the photodegradation of Tz involves attack of •OH radicals on the triazole ring leading to cleavage and formation of intermediate a, followed by addition of an •OH radical at the C5 position to give intermediate b. Photodegradation of this intermediate yields NH4+ ions and/or CO2 gas followed by intermediates c-e. Formation of NO3- and NH4+ ions and CO2 evolution continued from the photodecomposition of intermediates c-e producing intermediates f-h, which also mineralized to NO3- ions, NH4+ ions, and CO2 gas. The triazine intermediates (i) with m/z ) 210(n) and 244(n) originate from recombination of the smaller fragments through a series of complex events.

Acknowledgments

The photodegradation of ATz (Scheme 1b) involved attack of the •OH radical on the ring structure yielding intermediate a followed by formation of species b and c through loss of the -NH2 function and ring opening at positions N1-C5 or N1-N2 on further illumination. Intermediate d was generated from ATz by cleavage of the N1-N2 bond and on further irradiation gave intermediates e-g. Intermediates h containing the triazine structure originate from recombination of various (but not exclusively) intermediates (b, c, f, and g) through a series of yet undefined complex events. The initial degradative steps in DaTz was deamination to give amitrole (ATz; intermediate i, see Scheme 1c) and addition of an •OH radical at the N2 position a with subsequent ring breakup to give intermediate b. Continued irradiation led to further deamination (NH4+ and NO3- ions) and to intermediates c and h followed by intermediates d-g. Intermediate j was formed by the same steps involved in the photodecomposition of ATz (Scheme 1b). Again, the two intermediate byproducts denoted as m formed through recombination of smaller precursor species seem to degrade slowly to the refractory cyanuric acid {m/z ) 130(p) and 128(n)}. Concluding Remarks. The results from this study help to re-emphasize the need to identify intermediates formed in any advanced oxidation technology used to detoxify a given pollutant feed. In attempts to show the photocatalyzed mineralization of the endocrine disruptor amitrole and two other related herbicides in TiO2-based photocatalysis, we succeeded only partially in this task. About 60-70% of the triazoles were mineralized to CO2, and only ca. 25-40% of the nitrogens were converted to NH4+ and NO3- ions. The remaining nitrogens formed new species in which triazoles were converted to the six-membered triazine byproducts. In part, we transformed one class of herbicides, the triazoles, into another class of herbicides, namely, the s-triazines that earlier studies (6) had shown their photocatalytic degradation

2326

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 7, 2005

Financial support from the Frontier Research Promotion Foundation and a Grant-in-Aid for Scientific Research 1450544 of the Japanese Ministry of Education, Culture, Sports, Science and Technology (to H.H.) for work carried out in Japan, and from the Ministero dell’Istruzione, Universitn e della Ricerca (MIUR-Roma, to N.S.) for research carried out in Italy is gratefully acknowledged.

Literature Cited (1) Tjalve, H. Arch. Toxicol. 1974, 33, 41. (2) Tsuda, H.; Hananouchi, M.; Tatematsu, M.; Hirose, M.; Hirano, K.; Takahashi, M.; Ito, N. J. Natl. Cancer Inst. 1976, 57, 861. (3) Steinhoff, D.; Weber, H.; Mohe, U.; Boehme, K. Toxicol. Appl. Pharmacol. 1983, 69, 161. (4) Corker, P.; Bonin, A. M.; Stacey, H. H. Mutat. Res. 1992, 283, 7. (5) Ollis, D. F.; Al-Ekabi, H. Photocatalytic Purification and Treatment of Water and Air; Elsevier: Amsterdam, 1993. (6) Pelizzetti, E.; Maurino, V.; Minero, C.; Carlin, V.; Pramauro, E.; Zerbinati, O.; Tosato, M. L. Environ. Sci. Technol. 1990, 24, 1559. (7) Fox, M. A.; Dulay, M. T. Chem. Revs. 1993, 93, 341. (8) Hoffmann, M. R.; Martin, S. T.; Choi, W.; Bahnemann, D. W. Chem. Rev. 1995, 95, 69. (9) Tokiwa, F.; Tsuji, K. J. Phys. Chem. 1971, 75, 23. (10) Poulios, I.; Kositzi, M.; Kouras, A. J. Photochem. Photobiol A: Chem. 1998, 115, 175. (11) Horikoshi, S.; Hidaka, H. J. Photochem. Photobiol A: Chem. 2001, 141, 201. (12) Horikoshi, S.; Watanabe, N.; Mukae, M.; Hidaka, H.; Serpone, N. New J. Chem. 2001, 25, 999. (13) Guillard, C.; Horikoshi, S.; Watanabe, N.; Hidaka, H.; Pichat, P. J. Photochem. Photobiol. A: Chem. 2002, 149, 155. (14) Maurino, V.; Minero, C.; Pelizzetti, E.; Piccinini, P.; Serpone, N.; Hidaka, H. J. Photochem. Photobiol. A: Chem. 1997, 109, 171. (15) Catastini, C.; Rafqah, S.; Mailhot, G.; Sarakha, M. J. Photochem. Photobiol. A: Chem. 2004, 162, 97. (16) Horikoshi, S.; Serpone, N.; Yoshizawa, S.; Knowland, J.; Hidaka, H. J. Photochem. Photobiol. A: Chem. 1999, 120, 63. (17) Emeline, A. V.; Polikhova, S.; Andreev, N.; Ryabchuk, V.; Serpone, N. J. Phys. Chem. B 2002, 106, 5956. (18) Zhao, J.; Hidaka, H.; Takamura, A.; Pelizzetti, E.; Serpone, N. Langmuir 1993, 9, 1646. (19) Emeline, A. V.; Frolov, A. V.; Ryabchuk, V. K.; Serpone, N. J. Phys. Chem. B 2003, 107, 7109. (20) Arnold, S. M.; Hicky, W. J.; Harris, R. F. Environ. Sci. Technol. 1995, 29, 2083. (21) Konstantinou, I. K.; Sakellarides, T. M.; Sakkas, V. A.; Albanis, T. A. Environ. Sci. Technol. 2001, 35, 398. (22) Horikoshi, S.; Wada, Y.; Watanabe, N.; Hidaka, H.; Serpone, N. New J. Chem. 2003, 8, 1216. (23) Horikoshi, S.; Hidaka, H. Chemosphere 2003, 51, 139. (24) Oh, Y.-C.; Jenks, W. S. J. Photochem. Photobiol. A: Chem. 2004, 162, 323. (25) Watanabe, N.; Horikoshi, S.; Hidaka, H.; Serpone, N. Unpublished results.

Received for review February 10, 2004. Revised manuscript received December 28, 2004. Accepted December 30, 2004. ES049791L