Simulated Sunlight Photodegradation over Titania in. Aqueous Media: A First Case of Fluorinated Aromatics and. Identification of Intermediates. Claudio Minero ...
Langmuir 1991, 7, 928-936
928
Kinetic Studies in Heterogeneous Photocatalysis, 6, AM1 Simulated Sunlight Photodegradation over Titania in Aqueous Media: A First Case of Fluorinated Aromatics and Identification of Intermediates Claudio Minero and Carlo Aliberti Dipartimento di Chimica Analitica, Universith di Torino, 10125 Torino, Italy
Ezio Pelizzetti' Dipartimento di Chimica Fisica Applicata, Universith di Parma, Parma, Italy
Rita Terzian and Nick Serpone' Department of Chemistry and Biochemistry, Concordia University, Montrkal, Quhbec, Canada H3G 1M8 Received May 25,1990. I n Final Form: July 17, 1990 Fluorinated aromatic compounds have been mineralized to COSand fluoride by irradiation of aqueous, air-equilibrated suspensions (pH 3, H2SO4) containing Ti02 (Degussa P-25anatase) and a fluorinated phenol (2-, 3-, or 4-fluorophenol and 2,4- and 3,4-difluorophenol) with AM1 simulated sunlight (A 2 310 nm). Decomposition of the fluorophenolsis immediately followed by defluorination (530min) as evidenced by the rate of formation of fluoride; only -7540% of fluoride is recovered in solution, the remaining quantity being adsorbed on the Ti02particles. This may have certain consequences on the time needed for COz evolution as fluoride may block some of the Ti02 catalytic sites. Evolution of C02 is 3 times slower (- 90 to 240 min) than decomposition/defluorinationof the starting fluorinated phenols. Apparent kinetic data have been determined by curve-fitting methods on the basis of a simple phenomenological model which implicates consecutive and parallel reactions: A -.+ Bi C. The major oxidizing species is the *OH radical formed by oxidation of surface hydroxyl groups and/or surface bound water molecules with the photocatalyst's valence band holes. It is argued that the distinction between a surface bound and a free (in bulk solution) 'OH radical is in fact a moot point which kinetic considerations alone cannot delineate. A mechanistic route is summarized to rationalize the various intermediate products formed along the temporal course of the photomineralization process: fluorohydroquinone,fluorocatechol, trihydroxybenzene, hydroquinone,catechol, and benzoquinone. The overall order of the total destruction (to C02) of the fluorophenols examined is 2,4-difluorophenol>3-fluorophenol - 4-fluorophenol> 2-fluorophenol 2 3,4-difluorophenol.
-
Introduction The photochemical behavior and fate of aromatic and aliphatic hydrocarbons in the environment are of primordial importance in processes to detoxify waste and groundwaters. Indeed, the photochemical transformation of these environmental substances, whether or not mediated by suitable catalysts, is an attractive alternative nonbiological method to natural biodegradation phenomena.' An inconvenience and a detrimental factor in direct solar photolysis is the lack of sunlight absorption by these substrates, attenuation of the sunlight, and the relatively shallow penetration depth of sunlight in natural aquatic bodies (several contaminants are adsorbed on sediments in lakes and rivers; e.g., polychlorinated biphenyls among others). Thus, a mediator catalyst (photosensitizer)which can absorb the solar photons and mediate the destruction (mineralization) of the substrates, or at least their conversion to environmentally less harmful substances, is desirable. In the last few years, we and others2have demonstrated that total mineralization of representative species, from (1) Leifer, Asa The Kinetics of Environmental Aquatic Photochemistry-Theory and Practice; American Chemical Society: Washington, DC, 1988, Chapters 3 and 4. (2) For references and details, see for example: (a) Serpone, N., Pelizzetti,E., Eds. Photocatalysia-FLndomentalaand Applications; WileyInterscience: New York, 1989. (b) Ollie, D. F.; Pelizzetti, E.; Serpone, N. Feature article submitted for publication in Enuiron. Sci. Technol.
0743-7463/91/2401-0928$02.50/0
the various classes on the list of top priority pollutants of the United States Environmental Protection Agency? to . COz and HC1 (or H2O) is feasible. In several studies,' we examined a variety of materials (metal oxides and others) to act as photocatalysts; Ti02 anatase Degussa P 25 has so far met the criteria imposed on these materials (light absorption, photochemical and chemical stability, cost, and availability). In the present study we examined the photocatalyzed oxidation of mono- and difluorinated phenols. Although these species have yet to be identified in wastewaters and thus identified as having a harmful impact on the environment, some of the fluorinated phenolics may, in the near future, be found in ecosystems because of the biological and/or nonbiological degradation of such industrial products as pharmaceuticals, preservatives of foodstuffs, and high-quality polymer^.^ In particular, the study was undertaken to examine the kinetics of the defluorination of these phenolic substrates and to explore the possible mechanisms of the photooxidative mineral(3) Callahan, M. A.; Slimak, M.; Gbel, N.; May, I.; Fowler, C.; Freed, R.; Jennings, P.; Dupree, P.; Whitmore, F.;Maestri, B.; Holt, B.; Could, C. Water Related Environmental Fate of 129 Priority Pollutants. Report EPA-44014-79-029a,b; US.Government Printing Office: Washington, DC, 1979 (available from NTIS). (4) (a) Barbeni, M.; Pelizzetti, E.; Borgarello, E.; Serpone, N. Chemosphere 1986, 14, 195. (b) Borello, R.; Minero, C.; Pramauro, E.; Pelizzetti, E.; Serpone, N.; Hidaka, H. Enuiron. Toricol. Chem. 1989,8,997. ( c ) Pelizzetti, E.; Borgarello, E.; Serpone, N.; et al. To be submitted. ( 5 ) Seidel, H.Chem. Ind. 1988, If, 62.
0 1991 American Chemical Society
Langmuir, Vol. 7,No. 5, 1991 929
Kinetic Studies in Heterogeneow Photocatalysis Table I. Retention Times (min) of Fluomphenols with CHaCN/HaO Mobile Phase retention timee 40 % 3076 20% compound CH&N CH&N CHsCN 2-fluorophenol (2-FPh) 2.9 4.4 8.2 3-fluorophenol(3-FPh) 3.2 5.2 10.5 4-fluorophenol (4-FPh) 4.6 8.8 2,4-difluorophenol (2,4-DFPh) 3.4 5.7 11.8 3,4-difluorophenol (3,4-DFPh) 3.8 6.8 16.8 ization, which in all cases led to the stoichiometricevolution of COz and fluoride. We also wished to ascertain the role of the position and the number of fluoro substituents on the kinetics of degradation. We have identified several intermediate products and to confirm the proposed mechanism we have examinedtheir photocatalyzed degradation under otherwise identical conditions.
Experimental Section Materials. The fluorinated species 2-, 3-, and 4-fluorophenols, and the 2,4- and 3,4-difluorophenolswere purchased from Aldrich (>99 % purity) and were used without further treatment. All other products were of reagent grade quality. Water was bidistilled and acetonitrile was HPLC grade quality (Lichrosov, Merck). The mobile phase for the liquid chromatographic (HPLC) determinations consisted of a CH&N/H20 mixture which was filtered through a Millipore filter (HA 0.45 pm) that had previously been washed with 100 mL of bidistilled water. The catalyst was Ti02 Degussa P 25 (mostly in the anatase form; BET ca. 55 m2/g; contains such impurities as A120s (
7
-.l
0.2
0
z
-0 10 20 30 40 50 Irradiation Time (min) Figure 5. Plots showing the photodecompositionof 2-fluoropheno1and the nearly concomitant formation of fluorideas a function of irradiation time (min)at constant concentration of Ti02 at 50 mg/L. The relative peak areas for three of the identified intermediates (fluorohydroquinone,fluorocatechol,and catechol; see Table 111) as a function of irradiation time are also indicated. See text for details of the experimental conditions. v .v
0
of fluoride reaches a maximum at approximately 70-80 % , which corresponds to the fractional quantity of fluoride in solution. The remaining 20-30% is adsorbed/photoadsorbed on the catalyst's particle surface. This was confirmed by making up a suspension containing 2 g/L TiO2,4 mg/L F,and 200 pM C2H50H. The suspension was subsequently treated under conditions identical with those used in the photodegradation of the fluorinated organics: in the dark, 15% of the fluoride was adsorbed immediately upon addition of Ti02 to the ethanolic fluoride solution; after 4 h of irradiation, 75% of the fluoride remained in solution. The formation and subsequent decomposition of the intermediate products, fluorohydroquinone (F-HQ),3-fluorocatechol (F-CC), and catechol (CC) obtained from the decomposition of 2-fluorophenol, along with the formation of fluoride are shown in Figure 5 (note that Ti02 is 50 mg/L). After 25 min into the process, 2-FPh has degraded while both F-CC and CC have formed and decomposed (they reach a maximum after -10 min of irradiation); F-HQ also reaches maximum concentration after ca. 10 min and is totally decomposed after 30 min of irradiation. By this time, approximately 20% of the phenol has been converted to carbon dioxide. The other 80 % has probably been converted to aliphatic intermediates that ultimately also degrade, but more slowly, to yield the final mineralization products. Formation of COz is faster for 3-FPh
0- ...o
10 20 30 Irradiation Time (mid Figure 7. First-order plots of the photodecomposition of 2,4and 3,4-difluorophenolas a function of irradiation time (min): concentrationof the difluorophenols,200 pM; temperature, -60 O C ; pH 3 with sulfuric acid; concentration of photocatalyst, 50 mg/L; irradiation source, AM1 simulated sunlight. For other conditions, see text.
0
and 4-FPh (compare t1p = 42 min versus 25 min; last column in Table 11). The rate of decomposition of 4-fluorophenol in a 100 mg/L Ti02 aqueous suspension is 0.16 f 0.01 min-l. Two of the intermediates identified were hydroquinone (HQ) and 4-fluorocatechol. From the curve fitting of the data (Figure 5,for example), the rate of decomposition of the HQ intermediate is 0.16 min-l, identical with the value of 0.16 f 0.01 min-l determined for the degradation of a pure sample of HQ under identical conditions. Photomineralizationof Difluorophenols. Figure 6 shows the decompositionof 2,4-difluorophenol(2,4-DFPh) in suspensions containing 50 or 100 mg/L of TiO2; the process follows good first-order kinetics (0.125 0.006 and 0.238 f 0.011 min-l, respectively) for several halflives. It is complete by 25 min (50mg/L of catalyst) and by 12 min with 100 mg/L of catalyst. Under identical conditions of concentration of catalyst, the decompositions of 2,4-DFPh and 3,4-DFPh are nearly identical (Figure 7); however,when the concentration of Ti02 is 100mg/L, the former degrades more rapidly by a factor of 1.5. The formation of fluoride and carbon dioxide for the mineralization of 2,4-DFPh in shown in Figure 8. Formation of F follows the degradation of the parent substrate 2,4-DFPh kdee = 0.24 0.01 min-' and kF = 0.14 f 0.02 min-'. Similar to the observations noted above, the concentration of fluoride in solution also reaches a maximum value of ca. 75% of the stoichiometric value. The remaining quantity of fluoride is adsorbed on the
*
*
932 Langmuir, Vol. 7, No. 5, 1991
Minero et al.
A 2,4-dif luorophenol
Ti&, 100 mg/L; pHi 3 Simu 1 at ed Sun 1 ig h t
0
60
30
90
120
150
Irradiation Time (min) Figure 8. Plots showing the photodecomposition of 2,4-diflu: orophenol, the concomitantformation of fluoride, and formation of the final product COz in the photomineralization of the fluorophenol;concentrationof the photocatalyst, 100 mg/L. Other conditions are given in Figure 7. c
1.O
. 0 e U
2 0.8
44
c
QJ
2
0.6
0
u 0 QJ
0.4
by comparing the retention times of these products with those from pure samples that are inferred to have been formed. Species so confirmed are summarized in Table 111. Thus, 2-FPh yields fluorohydroquinone (F-HQ), 3-fluorocatechol (3-FCC), and catechol, while the 4-flUOrO analogue gives 4-fluorocatechol (4-FCC) and hydroquinone. In the case of 3-fluorophenol, the aromatic species identified are fluorohydroquinone, 3-fluorocatechol, and 4-fluorocatechol. For the difluorophenols, only one aromatic intermediate was identified: fluorohydroquinone. The nature of other intermediates (see Table IV) was inferred from the similar retention times of the species produced from different starting organic substrates. The symbolsC, F, D, E, and G refer to species in the mechanistic scheme treated later (see Discussion).
Discussion Irradiation of air-equilibrated aqueous suspensions containing the photocatalyst Ti02 (anatase) and a fluorophenol (2-, 3-, or 4-fluorophenol and 2,4- or 3,4-difluorophenol) with AM1 simulated sunlight (A 1 310 nm) leads to the rapid disappearance of the fluorinated hydrocarbons in short time, even for concentrations of catalyst as low as 0.05 or 0.1 g/L. In every case, the photocatalyzed mineralization of the phenolic substrates has been confirmed by product (COz and fluoride) analysis and is consistent with the stoichiometry (reactions 3 and 4)
N
-
H
P 0.2 L 0
z
0
catechol hydroquinone
0.0 I-
0
10
20 30 Irradiation Time (mini
hu
FC,H,OH I
40
Figure 9. First-order plots of the photodecomposition of catechol and hydroquinone as a function of irradiation time (min); concentration of concentrationof the organic substrate, 200 MM; TiOz, 50 mg/L; pH 3 with sulfuric acid. The inset shows the corresponding results in the photodecomposition of 3-flUOrOcatechol, which degrades via a reasonable first-order process. For other conditions, see text.
catalyst's particle surface. Formation of carbon dioxide, which confirms total destruction of the difluorinated phenol, is nearly complete after 90 min of irradiation (kco2 = 0.036 0.002 min-l and t t p = 19 min). For the 3,4difluoro analogue, disappearance of the parent compound is 2 times faster than formation of fluoride in 50 mg/L Ti02 suspensions; in 100 mg/L Ti02 dispersions, the rate of degradation of 3,4-DFPh is somewhat greater than formation of F- within experimental error (see Table 11). Formation of carbon dioxide from the mineralization of 3,4-difluorophenol is 3 times slower than that from the 2,4-difluoro analogue. As a test that the intermediates formed in the degradation of the monofluoro-and difluorophenolsalso degrade in short time under identical conditions, we examined 3-fluorocatechol, hydroquinone, and catechol by irradiating 50 mg/L Ti02 suspensions with AM1 simulated sunlight (Figure 9). The apparent kinetics are summarized in Table 11. Hydroquinone was also examined in a 100 mg/L catalyst suspension. Both HQ and CC degrade via firstorder kinetics; 3-FCC also degrades via reasonably good first-order kinetics (k 0.059 f 0.007 min-1; see inset in Figure 9) although zero-order kinetics cannot be ruled out within the inevitable experimental scatter of the data. Identification of Intermediates. Liquid chromatographic techniques permit identification of intermediate products from the photodegradation of organic substrates
*
-
+ 6.50, Ti02 6CO, + 2H,O + H F
-
(3)
hu
F,C,H,OH
+ 60, Ti02 6CO, + H,O + 2HF
(4)
where HF represents fluoride irrespective of its form. The onset of absorption of light by anatase Ti02 (bandgap3.2 eV) occursat -390 nm. With AM1 simulatad sunlight as the photon source, the photocatalyst absorbs approximately 3-5 76 of the radiation. This leads to the generation of conduction band electrons (e-) and valence band holes (h+),reaction 5. Subsequently, these charge carriers either can recombine within the bulk of the material or can migrate to the particle surface (ca. 30 nm diameter) where they can be trapped, recombine, or be implicated in some redox processes (reduction by e- and oxidation by the h+). The more relevant events taking place on the particle surface are adsorption of molecular oxygen on Ti"* sites which reduce oxygen to the superoxide radical anion, 02.- (reaction 6), while the positive charge carrier can oxidize the surface hydroxyl groups or the surface bound water (reactions 7) to surface hydroxyl
radical^.^
-
412
TiO,
+ hv (lEbg)
e-
+ h+
(5)
At low light fluxes, n = 1.
Ti"-OH-
+ h+ $ Ti'VI'OH
(74
2
Ti"-OH, + h+ TiIvl'OH + H+ (7b) We have already reported10 on the necessity of having both molecular oxygen and water present during the pho(9) Matthews, R. W. J. Chem. Sac., Faraday Tram. 1 1984,80,457. (10) Barbeni, M.; Prameuro, E.; Pelizzetti, E.; Borgarello, E.; Griitzel, M.; Serpone, N. Nouu.J. Chim. 1984,8, 546.
Kinetic Studies in Heterogeneous Photocatalysis
Langmuir, Vol. 7, No. 5, 1991 933
Table 111. Intermediates Identified by HPLC Analyses during the Photocatalyzed Degradation of Fluomphenols intermediates identified compound examined
fluorohydroquinone
3-fluorocatechol
catechol
4-fluorocatechol
hydroquinone
6H
eF
6.
OH
OH
&
OH
&
F
@
OH
Q
F
@ OH
fF
6.
&F
F
OH
Table IV. Summary of Intermediate Products Identified in the Photocatalyzed Mineralization of Fluomphenols identified intermediates
C
D
F
compounds examined catechol
&
3-fluorocatechol fluorohydroquinone
1,2,4-trihydroxybenzene
3-fluorocatechol fluorohydroquinone 4-fluorocatechol
1,2,4-trihydroxybenzene
0
0
E
0
F
hydroquinone
4-fluorocatechol
fluorohydroquinone
4-fluorocatechol
p-benzoquinone
F
6. F
fluorohydroquinone F
tomineralization process. Trapping of the charge carriers, as exemplified by reactions 6 and 7, plays the further role of suppressing somewhat electron/ hole recombination, thereby increasing the competitiveness of the photooxidative process. Reactive Species. All the evidence available to date suggests that the hydroxyl radical may be the major oxidizing species involved in the photomineralization of a wide class of organic s u b s t a n c e ~ . ~ JSuch ~ - ~ ~evidence has involved the observation that a number of hydroxy-
lated species are formed from the degradation of aromatic substrates:l6J6hydroquinones, catechols, 1,2,44rihydroxybenzene, as also witnessed in the present study. As well, (11)Izumi, I.; Dunn, W. W.; Willbourn, K. 0.;Fan, F. F.; Bard, A. J. J. Phys. Chem. 1980,84,3207. (12) Izumi, I.; Fan, F. F.; Bard, A. J. J. Phys. Chem. 1981,85, 218. (13) Fijihira, M.; Satoh, Y.;Om,T. Bull. Chem. SOC.Jpn. 1982,65, 666.
(14) (a) Matthew, R. W. J. Catal. 1986,97,565. (b) Matthew, R. W. Water Res. 1986,20,569. (c) Matthew, R.W. J.Phys. Chem. 1987,91, 3328. (d) Matthem, R.W. Sol. Energy 1987,38,405. (e) Matthew, R. W.A u t . J. Chem. 1987,40,667.
934 Langmuir, Vol. 7, No. 5, 1991
a kinetic isotope effect (H2O/D20) was noted recently in the photooxidation of 2-propanol in Ti02 and ZnO suspensions where formation of 'OH or 'OD appears to be rate determining." Electron spin resonance examinations of irradiated Ti02 have identified the 'OH radical (and no others) under ambient conditions in aqueous dispersions's and in gas-solid systems.lg In aqueous acidic media (pH 31, as used in this study, 02'- protonates to form the hydroperoxide radical HO2' which is subsequently converted to H202.9J2J4vm22 Absorption of UV light by hydrogen peroxide or its reaction with e- can breakdown H202 to produce the 'OH radiIt would appear, however, that the major cal.13J6*x)~21 source of *OHradicals is reactions 7 or their eq~ivalent.99~3 Direct oxidation of the organic substrates by the valence band holes of the photocatalyst does not appear to play a significant role.24 The present study is silent on the possible direct involvement of 0 2 ' - anions in the photodegradation process via either oxidation or reduction of the fluorinated phenols, and organics in but it is worth noting that 02'- can oxidize aromatics (e.g., methylhydroquinone is oxidized by the superoxide radical 1.7 X lo7 M-' s-' 25b). anion with k Although 'OH radicals are produced on the catalyst's particle surface, a suggestion has recently been made24 that these radicals may diffuse away from the surface and react with a substrate in the solution bulk. Employing a reaction diffusion modulus for 'OH radicals, defined as V = k o ~ [ s ] / ( D / L where ~) V is 1 for equal reaction and diffusion rates, k o is ~ 1.4 X 1O'O M-' s-',~~ [SI is the concentration of the substrate (here 200 pM),D is the liquid diffusion coefficient taken as cm2/s, and L is the diffusion length of the 'OH radical from the particle surface (in the present case about 200 A), Turchi and 01lis24have inferred that the photooxidative process need not take place at the catalyst's surface, as the reactive species can diffuse several hundred angstroms into the solution. A recent spin trapping/EPR study, carried out at ambient temperature, has noted that the reaction between 'OH radicals, the lone radicals observed under these conditions, and phenol takes place on the particle surface and that the diffusion length of 'OH is probably not more than a few atomic distances.'& The contention between a 'surface-bound 'OH radical" and a "free solution OH radical" may in fact be a moot point. A semiconductor particle in an aqueous electrolytic system is known to be surrounded by both a compact Stern layer and a Gouy-Chapman diffuse layer,27with the latter
-
(15)(a) Okamoto, K.;Yamamoto,Y.; Tanaka, H.; Tanaka, M.; Itaya, A. Bull. Chem. SOC. Jpn. 1985,58,2015. (b) Okamoto, K.; Yamamoto, Y.; Tanaka, H.; Itaya, A. Bull. Chem. SOC.Jpn. 1985,58,2023. (16)Al-Ekabi, H.;Serpone, N. J. Phys. Chem. 1988,92,5726. (17)Cunningham, J.; Srijaran, S. J.Photochem. Photobiol., A 1988, 43, 329. (18) (a) Ceresa, E. M.; Burlamacehi, L.; Visca, M. J. Mater. Sei. 1983, IS,289. (b) Jaeger, C. D.; Bard, A. J. J. Phys. Chem. 1979,83,3146. (e) Bolton, J. R. Personal communication, 1989. (19)Anpo, M.; Shima, T.; Kubokawa, Y. Chem. Lett. 1985,1799. (20)Cundall, R. B.; Rudham, R.; Salim, M. S.J . Chem. Soc., Faraday Trans. I 1976,72,1642. (21)Harvey, P. R.; Rudham, R.; Ward, S. J. Chem. Soc., Faraday Trans. 1 1983,79,1381. (22)Herrmann, J.-M.; Pichat, P. J. Chem. Soc., Faraday Trans. I 1980,76, 1138. (23)Kormann, C.; Bahnemann, D. W.; Hoffmann, M. R. Enuiron. Sci. Technol. 1988,22,798. (24)Turchi, C.;Ollis, D. F. J. Catal. 1990,122,178. (25)(a) Terzian, R.; Serpone, N.; Meisel, D.; Pelizzetti, E. Work in progress. (b) Rao, P. S.; Hayon, E. J . Phys. Chem. 1975,79,397. (26)(a) Adams, C.E.; Michael, B. D. Trans. Faraday SOC.1967,63, 1171. (b) Land, E. J.; Ebert, M. Trans. Faraday SOC.1967,63,1181. (27)Bard, A. J.; Faulkner, L. R. Electrochemical MethodsFundamentals and Applications; John Wiley: New York,1980; p 500 ff.
Minero et al. potentially attaining severalthousand angstroms.% If the particle surface properties can still influence the 'OH radical species within the diffuse layer, then the above variants are of no consequence. Recent studiesm show that in fact the 'OH radical does not leave the catalyst's particle surface. Apparent Kinetics. Although the fluorophenols examined here do not appear to be adsorbed to the particle surface in the dark, the extent of photoadsorption is often an inaccessible parameter depending, as it were, on adsorption/desorption rates and on whatever other process(es) that takes place subsequent to adsorption. In describing a rate expression for the photomineralization process, we make the implicit assumption that there is a constant fraction, however small, of the organic substrate on the catalyst's oxidatively active surface sites, TiIvl*OH.24 Following 'OH attack on the aromatic ring of the substrate Sade,one or more intermediates (In& and/or I,1) form which subsequently or simultaneously undergo defluorination and fragmentation to aliphatic species that ultimately degrade to produce stoichiometric quantities of CO2
where kb denotes a sum of rate constants for the formation of various intermediate species and ksb represents a sum of rate constants in the fragmentation of these intermediates. Employing the methods from enzyme catalysis as we did earlier,*32 the rate of formation of product will be given by33
where a = (kh+ ksb)/k&, Ks {=ka/(kd + kb))is taken as the photoadsorption coefficient for the various substrates, and N, is the number of oxidative active sites. Expression 9 is reminiscent of a Langmuir-Hinshelwood rate equat i ~ n Further . ~ ~ modification of 9, to take account of the formation of the reactive 'OH species described by reactions 7, leads to rate =
(
4"gApk7roH)k,KsNn[S] krec
{I + aKs[S] + K,[H,O]
+ ZK,[Il + ZK,,[ions]}
X
Here, PA, is the fraction of the particle surface that is irradiated, A, is the particle surface area, TOH is the lifetime of the 'OH radicals, and k,,, is the rate constant for electron/hole recombination events (radiative and nonradiative). The additional terms in the denominator ~~
(28)Gratzel, M. Heterogeneous Photochemical Electron Transfer; CRC Press: Boca Raton, FL, 1989; p 60. (29)Lawless, D.;Serpone, N.; Meisel, D. J.Phys. Chem., in prees. (30)Terzian, R.; Serpone, N.; Minero, C.; Pelizzetti, E. J. Catal., in press. (31)Terzian, R.; Serpone, N.; Minero, C.; Pelizzetti, E.; Hidaka, H. J. Photochem. Photobiol., A 1990,55,243. (32! Minero, C.;Maurino, V.; Pelizzetti, E.; Terzian, R.; Serpone, N. Submitted for publication in New J. Chem. (33)Reference 8,pp 4W406. (34)Reference 8,pp 243-251.
Langmuir, Vol. 7, No. 5, 1991 935
Kinetic Studies in Heterogeneous Photocatalysis indicate the influence of the solvent water, the intermediates, and various ions present in the system; through adsorption, these can act as inhibitors by blocking some of the active surface sites on the photocatalyst. The adsorption isotherm expression for oxygen is included in eq 10 to consider the effect of molecular oxygen on the rate of product formation.30 Expression 10 has the same analytical form as the equations reported by Okamoto and co-workers16and more recently by Turchi and Ollis.% The latter authors noted that with minor variations, the expression for the rate of photomineralization of organic substrates with irradiated Ti02 presents the same saturation type kinetic behavior as protrayed by the Langmuir-Hinshelwood rate law,34 and this whether or not (i) reaction occurs while both reacting species are adsorbed, (ii) a "freen radical species reacts with the adsorbed substrate, (iii) a surface-bound radical reacts with a free substrate in solution, and (iv) the reaction takes place between the free reactants in solution. Under the present circumstances, assigning an operational mechanism for reactions taking place in heterogeneous media to a Langmuir type process, to an Eley-Rideal pathway, or to an equivalent type process is kinetically not possible.s*N Effect of Position a n d Number of Fluoro Substituents. Perusal of the results in Table I1 reveals some interesting trends with respect to the positions of the F substituent(s)on the aromatic ring; however,there appears to be little trends in the rates between monofluorinated and difluorinated substrates. The order of disappearance of the monofluorophenols is 2-FPh 4-FPh > 3-FPh; for the difluoro analogues, the order is 2,4-DFPh 1 3,4-DFPh. Defluorination follows similar trends, namely 2FPh 1 4-FPh 1 3-FPh and 2,4-DFPh > 3,4-DFPh. These are understandable in terms of preferential attack by the 'OH radical on positions ortho and para to the OH ring substituent of the phenols. As well, although catechol and hydroquinone degrade a t the same rate, both decompose faster than 3-fluorocatechol. Evolution of the final product (C02) follows a rather different order in the monofluo4-FPh > 2-FPh by nearly a rophenols: thus, 3-FPh factor of 2, while 2,4-DFPh > 3,4-DFPh by a factor of 3. Clearly, once defluorination has occurred, the final mineralization follows different routes no doubt dictated by the nature of the intermediate species. The overall order of the photomineralization process (COz formation) is 2,4DFPh > 3-FPh 4-FPh > 2-FPh 1 3,4-DFPh. In our earlier studies on the photodegradation of cresols,3O and two depigmenting substances?l the data from the evolution of carbon dioxide was curvefitted to the phenomenological equation that results from reaction 1,s namely
Scheme I 0
6
Scheme I1
r w \
/ 6 \
OH
A
-
and Applications; Serpone, N., Pelizzetti, E.; Eds.;Wiley-Interscience:
where C02 evolution is about an order of magnitude slower than the decomposition of these species; only -20% of the fluorophenols have been converted to carbon dioxide when the quantity of phenols left in solution is zero. We are tempted to attribute these variations to the presence of fluorine in the present systems. We have noted above (see Results) that not all the stoichiometric quantity of fluoride is recovered experimentally; some probably adsorbed on the catalyst's particle surface on which it may compete with the intermediates for the Ti"l'0H sites, thereby acting as an inhibitor and reducing N , (eq 10). Photodegradation Mechanisms. The intermediates that have been identified (Table IV) by HPLC methods permit the inference of a mechanistic pathway commencing with attack of the 'OH radical on the phenyl ring, irrespective of whether the 'OH is free or surface found. Schemes I and I1 summarize the events that lead to formation of the aromatic intermediates. The first step (a) in Scheme I produces a fluorodihydroxycyclohexadienylradical37 which through loss of H2O (step 1)forms a phenoxy radical intermediate; this can interact with the HOz' radical (protonated 02'-, see above) to give back the original substrate A. Further attack of the 'OH radical on the phenoxy radical species produces G and/or F. Loss of fluoride occurs in the formation of G to give benzoquinone for the monofluorinated phenols (observed for 4-fluorophenol; Table IV). The expected products from F are fluorocatechol and fluorohydroquinone as seen for the monofluorophenols but not for the difluorinated analogues. Loss of fluoride by the cyclohexadienyl radical (step 4) yields a fluorohydroxyphenoxy radical which interacts further with .OH radicals to give E (step 5) and D (step 6), while reaction with HOz' gives
(36) Al-Ekabi, H.; Serpone,N. In Photocatalysie-Frtndamentals and Applications; Serpone, N., Pelizzetti, E., Ede.; Wiley-Interscience: New York, 1989; Chapter 14.
Chem. 1989,93, 1938.
-
-
[c]= I[AIo/(~B- kA)}(kB(l- exP(-kAt)) k A ( 1 - exp(-kBt)))
(11)
In all three cases, curve-fits of the data of carbon dioxide evolution gave rates that were just slightly smaller than those of the disappearance of the originalsubstrate. That is, C02 evolution followed closely the disappearance of the original substrate. This is presently not the case for the florinated phenols reported here (see Figures 4 and 8), (35) Pichat, P.; Herrmann, J.-M. In Photocatalysis-Fundamenta~
New York, 1989; Chapter 8.
(37) Draper, R. B.;Fox, M. A.; Pelizzetti, E.; Serpone, N. J. Phys.
936 Langmuir, Vol. 7, No. 5, 1991 C (step 7). SpeciesE is not a possible intermediate for the monofluorophenols, but difluorophenols would produce hydroxybenzoquinone (not observed; Table IV). For the monofluoro analogues, D gives trihydroxybenzene, observed for both 2-FPh and 4-FPh. No fluorinated trihydroxybenzenes have been identified; these species must be thermally unstable as are the trihydroxybenzeneswhich rapidly oxidize and fragment to aliphatic species. In the case where C forms, the intermediate products are hydroquinone for 4-FPh and catechol for 2-FPh. For the difluorinated phenols, both fluorohydroquinone and fluorocatechol are possible and were identified. The pathway for 3-fluorophenolis described separately in Scheme I1 in which intermediates F and D are relevant (see Table IV). Reaction of this phenol with the 'OH radical gives the cyclohexadienyl radical which through loss of water gives the fluorophenoxy species. Further 'OH attack yields F (3-FCCJ4-FCC, and FHQ; Table IV) and subsequently another cyclohexadienylradical which through loss of fluoride and reaction with HO$ ultimately yields the trihydroxybenzene (species D).
Minero et al.
anatase Ti02 suspensions. While decomposition and defluorination of the phenols occur almost concomitantly and in rrlatively short time (130min), complete evolution of COz is rather slow by comparison (90-240 min). This is caused either by slow degradation of the aliphatic intermediates which are the immediate precursors to COz or by a blocking of the oxidative catalytic sites by adsorbed fluoride on the photocatalyst's (TiOz) particle surface. Various aromatic intermediates have been identified which have permitted a tentative mechanistic route to be described. It is relevant that even fluorinated hydrocarbons can be totally destroyed (as demonstrated earlier for chlorinated analogues9 which will no doubt have an important impact on efforts to detoxify wastewaters and on naturally occuring environmental processes.
Conclusions Fluorinated hydrocarbons, as exemplified by the five
Acknowledgment. Our work in Torino and Parma is supported by EN1 Ricerche and the European Economic Community, while our work in Montreal is supported by the Natural Sciences and Engineering Research Council of Canada. We are grateful to these agencies for their financial backing, and we are particularly grateful to the North Atlantic Treaty Organizationfor an exchangegrant (NATO Grant No. CRG 890746).
fluorophenols examined in this work, can also be totally photomineralized to carbon dioxide and fluoride with simulated sunlight irradiation of aqueous, air-equilibrated
(38) Ollis, D. F.;Pelizzetti, E.; Serpone, N. In Photocataly8bFundamentalsand Applications;Serpone,N., Pelizzetti,E., Me.;WileyInterscience: New York, 1989; Chapter 18.
