3674
Langmuir 2004, 20, 3674-3678
Photooxidation of a Mustard Gas Simulant over TiO2-SiO2 Mixed-Oxide Photocatalyst: Site Poisoning by Oxidation Products and Reactivation D. Panayotov, P. Kondratyuk, and J. T. Yates, Jr.* Department of Chemistry, Surface Science Center, University of Pittsburgh, Pittsburgh, Pennsylvania 15260 Received October 10, 2003. In Final Form: February 11, 2004 The photooxidation of 2-chloroethyl ethyl sulfide (2-CEES), a simulant for mustard gas, was studied using transmission IR spectroscopy on a mixed-oxide TiO2-SiO2 photocatalyst. Ultraviolet irradiation in the photon energy range from 2.1 to 5 eV was employed at a catalyst temperature of 200 K. Rapid photooxidation was observed by the loss of infrared intensity in the ν(CHx) stretching region, and concomitant infrared features of adsorbed oxidation products were observed to develop. The oxidation products, captured on the photocatalyst at 200 K, were found to block 2-CEES readsorption. Upon heating the poisoned photocatalyst to about 300 K, infrared measurements indicate that the adsorbed CO2 oxidation product was desorbed. The capability for full readsorption of 2-CEES was achieved upon heating the poisoned photocatalyst to 397 K, and continued rapid photooxidation of the 2-CEES was then possible at about 1/3 the rate found for the fresh catalyst. Thus thermal treatment at 397 K of oxidation-product-poisoned TiO2-SiO2 material is able to partially restore the TiO2-SiO2 photooxidation activity.
I. Introduction The destruction of mustard gas by photooxidation over TiO2-based photocatalysts may form a new decontamination technology, following the demonstrated use of photoactivated TiO2 for many environmental cleanup applications.1,2 Photocatalytic oxidation of nerve agent simulant (organophosphorus compounds)3-5 and mustard gas simulant (organosulfur compounds)6-13 has been studied in the past decade. These compounds can be completely mineralized via multiple steps involving several intermediate products. In most cases, CO2, H2O, and inorganic salts are the final products and no hazardous final byproducts are formed. However, photocatalytic oxidation over TiO2 can be kinetically retarded due to the accumulation of partially oxidized intermediate species on the catalyst surface,14 a poisoning process which can occur rapidly in some cases. Both diethyl sulfide (DES) and 2-chloroethyl ethyl sulfide (2-CEES) have been studied as simulants of bis (2chloroethyl) sulfide, the main component of mustard gas. (1) Mills, A.; Le Hunte, S. J. Photochem. Photobiol., A 1997, 108, 1. (2) Fujishima, A.; Hashimoto, K.; Watanabe, T. TiO2 Photocatalysis: Fundamentals and Applications; BKC: Tokyo, 1999. (3) O’Shea, K. E.; Beightol, S.; Garcia, I.; Aguilar, M.; Kalen, D. V.; Cooper, W. J. J. Photochem. Photobiol., A 1997, 107, 221. (4) Obee, T. N.; Satyapal, S. J. Photochem. Photobiol., A 1998, 118, 45. (5) Rusu, C. N.; Yates, J. T., Jr. J. Phys. Chem. B 2000, 104, 12299. (6) Davidson, R. S.; Pratt, J. E. Tetrahedron Lett. 1983, 24 (52), 5903. (7) Yang, Y.-C. U.S. Army Chemical Research Development and Engineering Center (CRDEC): Aberdeen Proving Ground, MD, May 1989; p 83. (8) Fox, M. A.; Kim, Y.-S.; Abdel-Wahab, A. A.; Dulay, M. Catal. Lett. 1990, 5, 369. (9) Vorontsov, A. V.; Savinov, E. V.; Davydov, L.; Smirniotis, P. G. Appl. Catal., B 2001, 32, 11. (10) Vorontsov, A. V.; Davydov, L.; Reddy, E. P.; Lion, C.; Savinov, E. N.; Smirniotis, P. G. New J. Chem. 2002, 26, 732. (11) Kozlov, D. V.; Vorontsov, A. V.; Smirniotis, P. G.; Savinov, E. N. Appl. Catal., B 2003, 42, 77. (12) Martyanov, I.; Klabunde, K. J. Environ. Sci. Technol. 2003, 37, 3448. (13) Panayotov, D. A.; Paul, D. K.; Yates, J. T., Jr. J. Phys. Chem. B 2003, 107, 10571. (14) Peral, J.; Ollis, D. F. J. Mol. Catal. A: Chem. 1997, 115, 347.
Vorontsov et al.9-11 have studied the photocatalytic gasphase oxidation of DES over several types of TiO 2. Several organic products were detected in the gas phase and on the catalyst surface corresponding to C-S bond cleavage and to oxidation of sulfur and carbon atoms. Deactivation of the TiO2 photocatalyst, as evidenced by the increased time required for complete mineralization of DES,11 has been established. Carbonate and sulfate species, detected on the TiO2 surface after complete DES oxidation (using FT-IR diffuse reflectance spectroscopy), are suggested11 as being responsible for TiO2 deactivation. In their recent study, Martyanov and Klabunde12 using gas chromatography-mass spectrometry (GC-MS) analytical procedures have observed a wide range of oxidation products during photocatalytic oxidation of 2-CEES over TiO2. Accumulation of final surface products, particularly sulfate ions, appeared to be responsible for TiO2 deactivation. Therefore, this technology requires that the reactivation of the photocatalyst can be achieved, and different regeneration techniques have been examined.15-18 Exposure of the used catalyst to flowing air and UV illumination for a certain time is capable of full restoration of activity in the case of o-xylene oxidation15 and partial recovery of activity in the case of benzene and toluene photooxidation.16 It has been shown that in the removal of nitrogencontaining organic compounds, TiO2-based photooxidation catalysts may be reactivated by water washing to remove nitrate products which slowly deactivate the catalyst.17 Severe deactivation of nanoparticle TiO2 catalysts due to the accumulation of partially oxidized intermediates on active sites has been reported for toluene photooxidation.18 Complete recovery of photoactivity is achieved after thermal regeneration of the catalyst at temperatures above (15) Ameen, M. M.; Raupp, G. B. J. Catal. 1999, 184, 112. (16) d’Hennezel, O.; Pichat, P.; Ollis, D. F. J. Photochem. Photobiol., A 1998, 118, 197. (17) Alberici, R. M.; Canela, M. C.; Eberlin, M. N.; Jardim, W. F. Appl. Catal., B 2001, 30, 389. (18) Cao, L.; Gao, Z.; Suib, S. L.; Obee, T. N.; Hay, S. O.; Freihaut, J. D. J. Catal. 2000, 196, 253.
10.1021/la0303815 CCC: $27.50 © 2004 American Chemical Society Published on Web 03/31/2004
Photooxidation of a Mustard Gas Simulant
500 K. Using platinum-loaded TiO2 results in lower photooxidation rates for toluene but facilitates the removal of photocatalyst poisons. Recently, transient kinetic models for TiO2 deactivation during photocatalytic oxidation of aromatics have been reported.19,20 A two-site model20 provides reasonable fits for deactivation and regeneration kinetics. Titania-silica mixed oxides exhibit different surface chemical and photochemical properties as compared to titania itself.21 The TiO2-SiO2 photocatalyst is reported to display enhanced photooxidation activity, compared to pure TiO2 materials.22-24 The mixed oxide was found to deactivate more slowly than TiO2 during photooxidation of toluene.25 It is suggested that the generation of new surface sites may be responsible for the increased activity and stability of TiO2-SiO2 compared to TiO2. We have employed a mixed oxide, TiO2-SiO2, as a photocatalyst for the destruction of 2-chloroethyl ethyl sulfide, which is a simulant for mustard gas.13 Infrared spectroscopy has been used to observe the formation of adsorbed CO2 as the final oxidation product possible, as well as the formation of partially oxidized products containing carbonyl functionalities.13 The goal of the work to be reported here is to determine whether these adsorbed products deactivate the photocatalyst and to study how the photocatalytic activity may be restored. The 2-CEES photooxidation experiments carried out here were done at 200 K. Using this low temperature, we are able to observe the photooxidation chemistry of the molecule in the absence of thermally induced chemistry. Subsequently, heating of the surface reveals the thermal chemistry by means of infrared spectroscopic changes. II. Experimental Section A. Materials. The TiO2-SiO2 sample was prepared by the modified aerogel procedure in the laboratory of Professor K. Klabunde, Kansas State University, as reported previously.26,27 Production of mixed oxides which are intimately intermingled, essentially molecular in nature, is the main advantage of this procedure. The TiO2-SiO2 mixed oxide (50 mol % TiO2 and 50 mol % SiO2) has very high surface area, 680 m2/g, large pore volume, 2.97 cm3/g, and average pore diameter, 175 Å. The powdered TiO2-SiO2 sample is hydraulically pressed into the openings of a flat tungsten grid as a circular spot 7 mm in diameter, typically weighing 1-1.5 mg (1.3-1.9 mg/cm2). The tungsten grid28 (0.0508 mm thick, with 0.22 × 0.22 mm square holes) was obtained from Buckbee-Mears, St. Paul, MN. The TiO2-SiO2 samples on the grid support assembly are then placed into the dual beam IR-UV photoreactor and evacuated. 2-Chloroethyl ethyl sulfide (98%) from Aldrich was used in this work. The liquid was transferred under nitrogen gas to a glass bulb and purified by five freeze-pump-thaw cycles. The bulb is attached to a stainless steel vacuum system, and the vapor is transferred to the dual beam IR-UV photoreactor. Oxygen (99.8%) from VWSCO was used as received. B. IR Measurements. The dual beam IR-UV photoreactor used in this work has been described previously.29,30 The stainless steel photoreactor can be used over wide ranges of temperature (19) Lewandowski, M.; Ollis, D. F. Appl. Catal., B 2003, 43, 309. (20) Lewandowski, M.; Ollis, D. F. Appl. Catal., B 2003, 45, 223. (21) Gao, X.; Wachs, I. E. Catal. Today 1999, 51, 233. (22) Anderson, C.; Bard, A. J. J. Phys. Chem. 1995, 99, 9882. (23) Anderson, C.; Bard, A. J. J. Phys. Chem. B 1997, 101, 2611. (24) Deng, Z.; Wang, J.; Zhang, Y.; Weng, Z.; Zhang, Z.; Zhou, B.; Shen, J.; Cheng, L. Nanostruct. Mater. 1999, 11, 1313. (25) Me´ndez-Roma´n, R.; Cardona-Martı´nez, N. Catal. Today 1998, 40, 353. (26) Panayotov, D. A.; Yates, J. T., Jr. J. Phys. Chem. B 2003, 107, 10560. (27) Carnes, C. L.; Kapoor, P. N.; Klabunde, K. J.; Bonevich, J. Chem. Mater. 2002, 14, 2922. (28) Ballinger, T. H.; Wong, J. C. S.; Yates, J. T., Jr. Langmuir 1992, 8, 1676.
Langmuir, Vol. 20, No. 9, 2004 3675 (from 100 to 1500 K) and pressure (from ∼10-8 to ∼760 Torr). The grid is fixed by a pair of nickel clamps alongside the edges and is held on a power/thermocouple feed-through, acting as a re-entrant Dewar in the dual beam IR-UV photoreactor. A type-K thermocouple is spot-welded to the top-center region of the grid. The sample temperature may be adjusted accurately and programmed by cooling the re-entrant Dewar with liquid N2 and ohmic heating of the tungsten grid via a Honeywell digital temperature controller.31 The grid is oriented at a 45° angle to the IR and the UV beams. The dual beam IR-UV photoreactor involves two orthogonal optical paths using KBr windows mounted on concentric Viton O-rings which are differentially pumped to prevent leaks. The photoreactor is connected to the ultrahigh vacuum system and is mounted on a computer-controlled translation system (Newport Corp.). The translation system allows one to align different positions on the grid with respect to the IR beam with (1 µm accuracy29 in both horizontal and vertical directions. The upper one-third position on the grid is empty and is used for IR measurements of the gas phase and the background absorbance in the same experiment. The middle and the bottom positions on the grid are used for IR measurements of catalyst samples. The IR spectra were measured with a Mattson FTIR spectrometer (Research Series I). The spectrometer is equipped with a liquid N2 cooled HgCdTe detector and is purged by nitrogen gas. The IR spectra within the region 4000-500 cm-1 were collected in the ratio mode at a resolution of 4 cm-1. The number of accumulated scans for each spectrum was set to 200 during the adsorption experiments, but it was decreased to 50 scans for the kinetic experiments in order to reduce the acquisition time. C. Vacuum System. The stainless steel vacuum system is pumped simultaneously with a Pfeiffer Vacuum 60 L/s turbomolecular pump and a Varian 20 L/s ion pump. The base pressure of the system is ∼10-8 Torr, after 24 h baking of the photoreactor and system while pumping with both pumps. A MKS capacitance manometer (Baratron, type 116A, range of 10-3-103 Torr) is used for reactant gas pressure measurements. D. Sample Activation, Adsorption, and Photooxidation Experiments. The fresh TiO2-SiO2 sample was treated in vacuum at 675 K for 4 h, treated with 6.2 Torr O2 for 0.5 h at the same temperature, and then treated in vacuum at 1026 K for 3 h. After that, the sample was cooled to room temperature and the reference spectra were taken. The adsorption of 2-CEES on the surface of fresh and used catalyst samples was carried out at 200 K. Seven consecutive dosages of 2-CEES with increased vapor pressure from 0.100 to 2 Torr were employed. These exposures of 2-CEES ensured the achievement of a saturation level for 2-CEES adsorption coverage on the TiO2-SiO2 surface.26 The photooxidation experiments were carried out at 200 K in the presence of 20.5 Torr O2. The UV light source was a 350 W high-pressure Hg arc (Oriel Corp. 18) lamp. The light was filtered with a water filter to remove the IR radiation. The intensity of the UV radiation on the sample was 240 mW/cm2 in the photon energy range of 2.1-5 eV.
III. Experimental Results and Discussion The photooxidation of 2-CEES is conveniently measured by the observation of the removal of ν(CHx) stretching mode absorbance in the 3000 cm-1 region.13 In addition, the formation of some adsorbed products may be observed in the 2400-1500 cm-1 region due to adsorbed carbon dioxide formation and to the formation of organic products containing carbonyl functionalities. Figure 1A shows the loss of ν(CHx) mode intensity during photooxidation using 20.5 Torr of O2(g). Figure 1B shows the development of infrared modes due to the production of various oxidation products which accompanies the loss of the 2-CEES molecule. The continuous increase in absorbance at 2355 (29) Mawhinney, D. B.; Rossin, J. A.; Gerhart, K.; Yates, J. T., Jr. Langmuir 1999, 15, 4617. (30) Rusu, C. N.; Yates, J. T., Jr. J. Phys. Chem. B 2000, 104, 1729. (31) Muha, R. J.; Gates, S. M.; Basu, P. Rev. Sci. Instrum. 1985, 56, 613.
3676
Langmuir, Vol. 20, No. 9, 2004
Panayotov et al.
Figure 3. Reactivation of TiO2-SiO2 by heating. Figure 1. Photooxidation of adsorbed 2-CEES on TiO2-SiO2. The initial condition was achieved by saturation of the photocatalyst with 2-CEES at 200 K.
Figure 2. 2-CEES readsorption after extensive photooxidation. Readsorption of 2-CEES was carried out up to saturation.
cm-1 reflects the mineralization of 2-CEES, that is, the accumulation of the final fully oxidized species, CO2(ads), on the photocatalyst surface. CO2(ads) must represent the result of multiple elementary photooxidation steps as a molecule as complex as 2-CEES is destroyed by the production of a sequence of intermediate oxidation products which ultimately reach CO2(ads) product.13 The bands at 1709 and 1560 cm-1 were attributed13 to carbonyl stretching modes due to aldehyde and carboxylate species, respectively. These species are the intermediate partially oxidized products formed simultaneously with the loss of alkyl absorbance (ν(CHx) modes, Figure 1A), that is, the loss of 2-CEES molecules. The kinetics of the loss of the ν(CH2)s mode absorbance during photooxidation is plotted in experiment I (starting at point I) in Figure 2A, working at 20.5 Torr O2 pressure and 200 K. A smooth and substantial decrease in intensity
is observed. Following the destruction of about 75% of the adsorbed 2-CEES molecules, readsorption was attempted using an identical exposure to 2-CEES as employed initially at point I. The behavior of the ν(CH2)s mode intensity upon readsorption is shown at point II for experiment II. It is observed that readsorption of 2-CEES does not cause a substantial increase in ν(CH2)s absorbance in experiment II. The spectral changes in the ν(Si-OH) and the ν(CHx) regions after 2-CEES adsorption in experiment I and experiment II are shown in Figure 2B,C. In Figure 2B, two groups of spectral features are observed to develop upon the initial adsorption of 2-CEES: (1) hydrogenbonded Si-OH groups at 3566 and 3330 cm-1; (2) ν(CHx) stretching modes in the 3000 cm-1 region. The ν(Si-OH) mode at 3566 cm-1 is due to hydrogen bonding to the Cl moiety of 2-CEES, as has been reported previously;26 the ν(Si-OH) mode at 3330 cm-1 is due to hydrogen bonding to the S moiety of 2-CEES and has also been reported previously.26 The Si-OH groups act as anchor groups for the 2-CEES molecules for both types of hydrogen bonding to the molecules, holding the 2-CEES in place for photooxidation at the neighboring photoactive TiO2 sites, as reported previously.13 During photooxidation, the low-intensity mode which develops as a shoulder at 3200 cm-1 (Figure 2C) is assigned to Si-OH groups bonded to an oxidation product. It is noted that in experiment II, the pair of 2-CEES hydrogenbonded Si-OH groups are not restored to their initial intensity by readsorption of 2-CEES on the partially deactivated photocatalyst surface, nor are the ν(CHx) modes of the 2-CEES substantially restored. Figure 3 shows the spectral behavior of the TiO2-SiO2 photocatalyst as it is reactivated by brief heating to the desired temperature. All the spectra are taken at 200 K after heating. In the top spectrum, obtained after extensive photooxidation of 2-CEES, we observe the residual 2-CEES by means of the ν(CHx) mode absorbance in the 3000 cm-1 region showing the characteristic three-peak feature. In addition, the strong mode due to adsorbed CO2 product
Photooxidation of a Mustard Gas Simulant
Langmuir, Vol. 20, No. 9, 2004 3677
Figure 4. Thermal reactivation of TiO2-SiO2 following 2-CEES photooxidation. Partial restoration of photooxidation activity.
at 2355 cm-1 is observed. Three hydrogen-bonded Si-OH modes at 3606, 3330, and 3200 cm-1 are also observed. The modes at 3606 and 3330 cm-1 are due to the remaining 2-CEES hydrogen bonded by the Cl and S moieties to Si-OH groups. The mode at 3200 cm-1, which appears upon photooxidation of adsorbed 2-CEES, is assigned to one or more types of hydrogen bond to partially oxidized products. Upon heating in vacuum to temperatures higher than 200 K, the product CO2 (2355 cm-1) is first observed to desorb and this process is completed at about 300 K. During the loss of the adsorbed CO2, little change occurs in the relative intensities of the three hydrogenbonded Si-OH groups, indicating that CO2 is bound directly to the TiO2-SiO2 photocatalyst sites and not to Si-OH sites. The loss of adsorbed CO2 below 300 K is consistent with thermal desorption studies of CO2 adsorbed on the TiO2(110) surface.32,33 In these studies, the CO2 was not hydrogen bonded since -OH groups were not present on the surface. Above 300 K, in Figure 3, the main spectral change involves the loss of absorbance near 3200 cm-1, as may be seen most easily in the spectra obtained by heating to 356 and 397 K. This indicates that partially oxidized organic products, which are hydrogen bonded to Si-OH groups, are removed by this thermal treatment above 300 K. The removal of oxidation products at 397 K causes partial reactivation of the TiO2-SiO2 photooxidation catalyst, although the ability to adsorb 2-CEES is completely restored as seen in Figure 4. Experiment I shows the kinetics of the photooxidation of 2-CEES on the fresh photocatalyst, and this spectrum is presented for comparison with the photooxidation kinetics experiment on reactivated (397 K) TiO2-SiO2 photocatalyst, called experiment III. It may be seen that the capability for 2-CEES adsorption is fully restored for the used catalyst by heating to 397 K in a vacuum, since the C-H absorbance at 2970 cm-1 is completely restored to the value observed in experiment I on the fresh catalyst. The rate of photooxidation is partially restored in experiment III on the reactivated photocatalyst, compared to the rate on the (32) Henderson, M. A.; Epling, W. S.; Perkins, C. L.; Peden, C. H. F. J. Phys. Chem. B 1999, 103, 5328. (33) Thompson, T. L.; Diwald, O.; Yates, J. T., Jr. J. Phys. Chem. B 2003, 107, 11700.
fresh catalyst in experiment I. Thus, in experiment I, a 50% reduction in 2-CEES coverage occurs after about 30 min of photooxidation, whereas the same change requires about 150 min in experiment III. Such partial deactivation effects have been observed for diethyl sulfide photooxidation on TiO2.11 The results from this study clearly show that accumulation of hydrogen-bonded partially oxidized products restrains the readsorption of the reactant molecule, 2-CEES, that is, its hydrogen bonding via Cl and S moieties to the surface Si-OH groups diminishes (Figure 2). This is in accordance with the generally accepted understanding of deactivation phenomena occurring on TiO2-based photocatalysts, that the buildup of partially oxidized species hinders the readsorption of the reactant molecule.15,16,18-20,25 The possibility that the adsorption of the final product, CO2, and the formation of carbonate species are also involved in surface deactivation as observed by others34 cannot be excluded. It was recently observed that CO2 adsorbs on anion vacancy defect sites on TiO233 which are preferential sites for O2 adsorption and thus important for photooxidation chemistry. The less than complete restoration of photooxidation activity of the TiO2-SiO2 catalyst after treatment at 400 K may be caused by an irreversible adsorption of sulfur-containing species on titanium active sites as suggested recently for 2-CEES12 and DES11 photooxidation over TiO2 catalysts. Such an effect would not interfere with 2-CEES adsorption on SiOH groups on the TiO2-SiO2 photocatalyst. IV. Conclusions The photooxidation of 2-CEES on a TiO2-SiO2 photocatalyst has been studied using transmission IR spectroscopy. The following conclusions have been obtained: 1. Photooxidation of 2-CEES at 200 K results in the formation of adsorbed CO2 product (2355 cm-1) as well as other partially oxidized products exhibiting carbonyl stretching modes at 1709 and 1560 cm-1. The partially oxidized products block readsorption of 2-CEES by Si-OH groups on the used photocatalyst. 2. The CO2 product may be desorbed by heating to about 300 K. Adsorbed CO2 product is not bound by hydrogen bonding to Si-OH groups. (34) Larson, S. A.; Falconer, J. L. Appl. Catal., B 1994, 4, 325.
3678
Langmuir, Vol. 20, No. 9, 2004
3. Other carbonyl-containing products may be desorbed by heating in a vacuum to about 400 K. These products are observed to be adsorbed via hydrogen bonding to Si-OH groups causing the formation of a broad hydrogenbonded spectral feature centered at about 3200 cm-1. 4. Reactivation by heating to 400 K restores the ability of TiO2-SiO2 photocatalysts to fully adsorb 2-CEES. The reactivated material exhibits an ability to continue to photooxidize 2-CEES at about 1/5 the rate observed for the fresh photocatalyst.
Panayotov et al.
Acknowledgment. We acknowledge with thanks the support of this work by the DoD Multidisciplinary University Research Initiative (MURI) program administered by the Army Research Office under Grant DAAD 19-01-0-0619. We also thank J. Q. Wang and Professor K. J. Klabunde for supplying the TiO2-SiO2 photocatalyst powder and the characteristic Brunauer-Emmett-Teller data on surface area and pore size distribution. LA0303815
