D Exchange - American Chemical Society

Mar 4, 2005 - Interfacial Chemistry and Engineering Group, Pacific Northwest ... Isotopic H/D exchange between coadsorbed acetone and water on the ...
0 downloads 0 Views 531KB Size
Download PDF
Langmuir 2005, 21, 3451-3458

3451

Acetone and Water on TiO2(110): H/D Exchange Michael A. Henderson* Interfacial Chemistry and Engineering Group, Pacific Northwest National Laboratory, P.O. Box 999, MS K8-93, Richland, Washington 99352 Received September 20, 2004. In Final Form: January 24, 2005 Isotopic H/D exchange between coadsorbed acetone and water on the TiO2(110) surface was examined using temperature programmed desorption (TPD) as a function of coverage and two surface pretreatments (O2 oxidation and mild vacuum reduction). Coadsorbed acetone and water interact repulsively on reduced TiO2(110) on the basis of results from the companion paper to this study, with water exerting a greater influence in destabilizing acetone and acetone having only a nominal influence on water. Despite the repulsive interaction between these coadsorbates, about 0.02 monolayers (ML) of a 1 ML d6-acetone on the reduced surface (vacuum annealed at 850 K to a surface oxygen vacancy population of 7%) exhibits H/D exchange with coadsorbed water, with the exchange occurring exclusively in the high-temperature region of the d6-acetone TPD spectrum at ∼340 K. The effect was confirmed with combinations of d0acetone and D2O. The extent of exchange decreased on the reduced surface for water coverages above ∼0.3 ML due to the ability of water to displace coadsorbed acetone from first layer sites to the multilayer. In contrast, the extent of exchange increased by a factor of 3 when surface oxygen vacancies were pre-oxidized with O2 prior to coadsorption. In this case, there was no evidence for the negative influence of high water coverages on the extent of H/D exchange. Comparison of the TPD spectra from the exchange products (either d1- or d5-acetone depending on the coadsorption pairing) suggests that, in addition to the 340 K exchange process seen on the reduced surface, a second exchange process was observed on the oxidized surface at ∼390 K. In both cases (oxidized and reduced), desorption of the H/D exchange products appeared to be reaction limited and to involve the influence of OH/OD groups (or water formed during recombinative desorption of OH/OD groups) instead of molecularly adsorbed water. The 340 K exchange process is assigned to reaction at step sites, and the 390 K exchange process is attributed to the influence of oxygen adatoms deposited during surface oxidation. The H/D exchange mechanism likely involves an enolate or propenol surface intermediate formed transiently during the desorption of oxygen-stabilized acetone molecules.

1. Introduction Water and hydroxyl groups are common adsorbed species on oxide surfaces, and the influence of these species on surface phenomena are the subject of much interest,1,2 particularly in the area of heterogeneous photocatalysis.3-6 Many examples exist in the TiO2 photocatalysis literature in which water is seen to either promote, poison, or have little effect on photocatalytic rates, depending on the nature of the species being photocatalyzed. Photo-oxidation of acetone on TiO2 surfaces in an example of a process in which trace amounts of water appear to promote the reaction rate,7-10 but moderate to high relative humidity levels inhibit conversion.8-15 In the companion paper to this one,16 it is shown that low coverages of water exert * E-mail: [email protected] (1) Thiel, P. A.; Madey, T. E. Surf. Sci. Rep. 1987, 7, 211. (2) Henderson, M. A. Surf. Sci. Rep. 2002, 46, 1. (3) Fu, X.; Zeltner, W. A.; Anderson, M. A. Stud. Surf. Sci. Catal. 1996, 103, 445. (4) Linsebigler, A. L.; Lu, G.; Yates, J. T., Jr. Chem. Rev. 1995, 95, 735. (5) Fox, M. A.; Dulay, M. T. Chem. Rev. 1993, 93, 341. (6) Serpone, N.; Khairutdinov, R. F. Stud. Surf. Sci. Catal. 1996, 103, 417. (7) Coronado, J. M.; Zorn, M. E.; Tejedor-Tejedor, I.; Anderson, M. A. Appl. Catal. B 2003, 43, 329. (8) Chang, C. P.; Chen, J. N.; Lu, M. C. J. Environ. Sci. Health A 2003, 38, 1131. (9) Chen, S.; Cheng, X.; Tao, Y.; Zhao, M. J. Chem. Technol. Biotechnol. 1998, 73, 264. (10) El-Maazawi, M.; Finken, A. N.; Nair, A. B.; Grassian, V. H. J. Catal. 2000, 191, 138. (11) Pearl, J.; Ollis, D. F. J. Catal. 1992, 136, 554. (12) Raillard, C.; Hequet, V.; Le Cloirec, P.; Legrand, J. J. Photochem. Photobiol. A 2004, 163, 425. (13) Vorontsov, A. V.; Kurkin, E. N.; Savinov, E. N. J. Catal. 1999, 186, 318. (14) Kim, S. B.; Hong, S. C. Appl. Catal. B 2002, 35, 305.

a destabilizing influence on η1-bound acetone molecules (η1-acetone is the major form of adsorbed acetone on the TiO2(110) surface), but surfaces populated with reactive oxygen species (from interaction of O2 with surface oxygen vacancies) result in formation of an acetone-oxygen complex17 that shows little or no susceptibility to the destabilizing influence of coadsorbed water. Observations of H/D exchange between acetone and OH groups on Al2O3,18 Fe2O3,19 MgO,20 NiO,20 TiO2,21,22 SnO2,23 ZrO2,24 Ni/SiO2,25 reduced Ni/CuO,26 Cu/MgO,27 and several zeolites28-30 reveal the presence of complex H/D interactions between adsorbed organics and surface hydroxyl groups. The key surface intermediate that (15) Choi, W.; Ko, J. Y.; Park, H.; Chung, J. S. Appl. Catal. B 2001, 31, 209. (16) Henderson, M. A. Langmuir 2005, 21, xxxx. (17) Henderson, M. A. J. Phys. Chem. B 2004, 108, 18932. (18) Hanson, B. E.; Wieserman, L. F.; Wagner, G. W.; Kaufman, R. A. Langmuir 1987, 3, 549. (19) Busca, G.; Lorenzelli, V. J. Chem. Soc., Faraday Trans. 1 1982, 78, 2911. (20) Miyata, H.; Toda, Y.; Kabokawa, Y. J. Catal. 1974, 32, 155. (21) Shannon, I. R.; Lake, I. J. S.; Kemball, C. Trans. Faraday Soc. 1971, 67, 2760. (22) Griffiths, D. M.; Rochester, C. H. J. Chem. Soc., Faraday Trans. 1 1978, 74, 403. (23) Thornton, E. W.; Harrison, P. G. J. Chem. Soc., Faraday Trans. 1 1975, 71, 2468. (24) Yamaguchi, T.; Nakano, Y.; Tanabe, K. Bull. Chem. Soc. Jpn. 1978, 51, 2482. (25) Young, R. P.; Sheppard, N. J. Catal. 1971, 20, 333. (26) Burwell, R. L., Jr.; Patterson, W. R.; Roth, J. A. J. Am. Chem. Soc. 1971, 93, 839. (27) Chikan, V.; Molnar, A.; Balazsik, K. J. Catal. 1999, 184, 134. (28) Xu, M.; Wang, W.; Hunger, M. Chem. Commun. 2003, 722. (29) Senkyr, G.; Noller, H. J. Chem. Soc., Faraday Trans. 1 1975, 71, 997. (30) Florian, J.; Kubelkova, L. J. Phys. Chem. 1994, 98, 8734.

10.1021/la0476581 CCC: $30.25 © 2005 American Chemical Society Published on Web 03/04/2005

3452

Langmuir, Vol. 21, No. 8, 2005

Henderson

facilitates the H/D exchange on these oxides appears to be an acetone enolate species. The enolate is formed by hydrogen atom transfer from one of the acetone molecule’s methyl groups to a surface anion site. This process is initiated on oxides by either acidic or basic sites, although on TiO2 it appears to be only basic sites that come into play because surface OH groups are not deemed acidic enough.22 The exchange process presumably involves H/D scrambling on the surface and recombination of a different OD/OH group with the enolate to reform the isotopically modified acetone. The role of acetone conversion to enolate has not been explored on single-crystal TiO2 surfaces, although enolate formation has been observed on the Znterminated surface of ZnO(0001),31 and was found to be favorable on MgO(100) on the basis of ab initio quantum mechanical calculations.32 These studies provide motivation for examining how ketones interact with hydroxyl groups and water on well-defined TiO2 single-crystal surfaces. In this paper, it is shown that the acetoneoxygen complex, as well as acetone molecules bound at step sites, exchange H/D atoms with coadsorbed OH/OD groups TiO2(110) commensurate with acetone (and recombinative water) desorption. 2. Experimental Section The base pressure of the UHV system used in this study was 2 × 10-10 Torr. The TiO2(110) crystal used in this study was obtained from First Reaction with dimensions of 10 mm × 10 mm × 1.5 mm. The crystal was epi-polished on both sides to provide maximum thermal contact with the Au-foil-covered Ta heating plate, which was resistively heated via 0.8 mm Ta wires spot-welded to the back of the plate. The crystal was cleaned by cycles of Ar+ sputtering and annealing until Auger electron spectroscopy (AES) and secondary ion mass spectrometry (SIMS) analyses deemed the surface free of detectable levels of impurities. After repeated sputter/anneal cycles, the crystal became blue and its surface showed evidence of oxygen vacancies on the basis of H2O temperature programmed desorption (TPD).33 Two surface pretreatments were used in this study. The pretreatment referred to as ‘the reduced TiO2(110) surface’ was prepared by annealing the crystal in UHV at 850 K for 10 min prior to each experiment. This surface possessed an oxygen vacancy population of about 7%. The pretreatment referred to as ‘the oxidized surface’ was obtained by exposing the reduced surface to g20 L O2 at 95 K followed by flashing to room temperature (RT) before recooling for gas exposures. Additional details on the use of these surface pretreatments in acetone surface chemistry on TiO2(110) are found elsewhere.17 Research-grade d6-acetone, d0-acetone, D2O, and H2O were obtained from Aldrich and were further purified using LN2 freeze-pump-thaw cycles. With the exception of D2O, each source was deemed to be isotopically pure on the basis of analysis of TPD spectra from multilayer exposures. The D2O source possessed a significant amount of HDO due to inadvertent isotope contamination. TPD spectra (see below) indicate that approximately 50% of the D2O source was HDO. Some of the H contamination may have arisen from the gas handling system despite efforts to acclimate its surfaces to D2O by repeated exposures to Torr pressures of the vapor. All gas exposures were performed with a calibrated directional doser positioned ∼1 mm from the face of the crystal and at a crystal temperature of 95 K. Coverages were estimated on the basis of the total gas exposure emitted from the calibrated doser during the dose assuming that both acetone and water adsorbed on TiO2(110) at 95 K with near unity sticking probabilities. Coverages are expressed in units of ML (monolayers) where ‘1 ML’ is equivalent to the areal density five-coordinated Ti4+ cation sites on the ideal TiO2(110) surface (5.2 × 1014 molecules/cm2). The heating rate for all TPD measurements was 2 K/s. (31) Vohs, J. M.; Barteau, M. A. J. Phys. Chem. 1991, 95, 297. (32) Oviedo, J.; Sanz, J. F. Surf. Sci. 1998, 397, 23. (33) Henderson, M. A. Langmuir 1996, 12, 5093.

Figure 1. TPD spectra of various masses from ∼1 ML d6acetone adsorbed on the reduced surface of TiO2(110). Several spectra are displaced vertically and are paired for clarity.

3. Results and Discussion Evidence for H/D exchange between coadsorbed d6acetone and H2O on oxidized TiO2(110) was first observed using high-resolution electron energy loss spectroscopy (HREELS).17 Although HREELS spectra of multilayer d6acetone showed no evidence for the presence of C-H stretches, intensity due to a C-H stretching mode appeared in HREELS spectra after desorption of the multilayer by heating to 135 K. The source of H was presumably due to background H2O adsorption prior to d6-acetone exposure. In this study, TPD was used to better characterize the isotopic exchange between acetone and water as a function of coverage and surface pretreatment. 3.1. H/D Isotope Exchange on Reduced TiO2(110). In this section, the exchange of H/D between coadsorbed acetone and water is explored on the reduced TiO2(110) surface, which possesses 7% surface oxygen vacancy sites. The major QMS cracking fragments of d6-acetone are CD3 (mass 15), CD3CO (mass 46), and the parent (mass 64). Figure 1 shows comparisons of the mass 46 versus mass 45 ((CD3)CO versus (CHD2)CO), mass 64 versus mass 63 ((CD3)2CO versus (CHD2)(CD3)CO), and mass 18 versus mass 17 (CD3 versus CHD2) QMS cracking fragment pairs from ∼1 ML d6-acetone adsorbed on reduced TiO2(110). H2O adsorption from the background occurred prior to d6-acetone adsorption. Depending on the day-to-day conditions, it was not uncommon to encounter about 0.050.1 ML of background water adsorption during the course of cooling the TiO2(110) crystal from 700 to 90 K (especially if the chamber had been previously exposed to high pressures of O2). The mass 17, 45, and 63 TPD traces show preferential desorption at 345 K relative to those of mass 18, 46, and 64, respectively. The mass 17, 45, and 63 desorption signals at 345 K did not originate from nonTiO2(110) surface (e.g., the sample holder or thermocouple assembly) because d6-acetone was only dosed on the crystal

Acetone and Water on TiO2(110): H/D Exchange

Langmuir, Vol. 21, No. 8, 2005 3453

Table 1. Ratios of QMS Fragment Intensities from d6-, d5-, d1-, and d0-Acetone this studya d6-acetone d5-acetone d1-acetone d0-acetone

NIST Chemical WebBook 36

methyl/parent

acetyl/parent

methyl/parent

acetyl/parent

mass 18/mass 64 4:1 mass 17/mass 63 3.5:1b.c mass 16/mass 59 ∼1.5:1b,c mass 15/mass 58 3:1

mass 46/mass 64 8.5:1 mass 45/mass 63 4:1b mass 44/mass 59 ∼7:1b,c mass 43/mass 58 7:1

mass 18/mass 64 0.5:1

mass 46/mass 64 2.7:1

mass 15/mass 58 0.4:1

mass 43/mass 58 1.6:1

a Ratios rounded to nearest half integer. b Ratios should be doubled for comparison with d - and d -acetone since only one of the two 6 0 methyl groups in the molecule is affected. c Accurate determination of ratios affected by signal contributions and/or high background from water and CO2.

(Hereafter, this molecule is referred to as 2-propenol.) In this case, H/D exchange between the adsorbate’s -OD group H2O/OH likely would be rapid and statistical on the basis of the amounts of d6-2-hydroxypropene formed and H2O/OH available. However, Xu and co-workers34 have calculated that formation of 2-propenol from acetone is endothermic by about 70 kJ/mol in the gas phase. Additional evidence against this species being responsible for the 345 K features comes from mass spectral data by Turecek and Hanus.35 These authors compared the mass spectra of acetone and 2-propenol and concluded that several cracking fragments, such as mass 29 (CHO and/ or C2H5), 31 (CH3O), 39 (C3H3), and 57 (C3H5O), could be used to differentiate between these two molecules. In the fully deuterated analogue, these masses would be at 30 (CDO), 34 (C2D5 and/or CD3O), 42 (C3D3), and 62 (C3D5O). Although trace mass 62 signal was often found in conjunction with the 345 K peak (see below), examination of the other mass signals from many different TPD experiments such as those shown in Figure 1 did not

provide evidence for the 345 K feature being associated with desorption of d5-2-propenol (CD3-C(OH)dCD2). On the basis of these arguments, the best model for explanation of the 345 K TPD features in Figure 1 is due to H/D exchange into d6-acetone resulting in desorption of d5acetone. As will be shown below, the same process occurs for coadsorption of d0-acetone and D2O. Published mass spectra of d1- or d5-acetone were not found in the literature, although those of d0- and d6-acetone are available on the NIST Chemistry WebBook.36 Table 1 compares the mass spectral cracking ratios for methyl/ parent and acetyl/parent obtained from the QMS used in this study with those listed in the NIST Chemistry WebBook. The mass 46-to-mass 64 and mass 18-to-mass 64 ratios for d6-acetone in this study were about 8.5:1 and 4:1, respectively. The QMS used in this study (Extrel C-50) has a transmission function that is different from those of conventional spectrometers, owing to the presence of a Bessel box used for SIMS analysis. Although not equivalent, the relative changes in those ratios on going from d0-acetone to d6-acetone are consistent with those on the NIST website. The QMS cracking yields of both methyl- and acetyl-derived masses increased relative to that of the parent on deuteration of the molecule. Table 1 also shows results from an effort to obtain methyl/parent and acetyl/parent ratios for d1- and d5-acetone obtained in TPD from H/D exchange into d0- and d6-acetone, respectively. The analyses for those ratios was complicated by the necessity of subtracting high background signals in TPD due to QMS cracking contributions from O at mass 16, OH from mass 17, and CO2 from mass 44. The resulting signals for CH2D (mass 16), CHD2 (mass 17), and CH2DCO (mass 44) were considerably ‘noisier’ than the acetone-related mass signals with no contributions from background gases. An additional complication is that H/D exchange in these TPD studies occurred predominately into one of the acetone molecule’s methyl groups, with the unaffected methyl group still contributing signal in the QMS spectrum. For example, QMS cracking of d1-acetone should contribute equivalent signals at mass 15 and 16 or at mass 43 and 44 as a result of a 50:50 chance of the d1-methyl being in the detected species. To obtain an accurate reflection of the extent of H/D exchange, the mass 16 signal from CHD2 contribution to cracking of d1-acetone should be doubled in a comparison with the methyl signals of d0- or d6-acetone. This method of comparison may not be valid since comparisons the d0- and d6-acetone signals indicate that deuteration does affect the QMS cracking ratios. Nevertheless, doubling the mass 45-to-mass 63 ratio from d5-acetone (which is the most reliable signal ratio because of background issues associated with the

(34) Xu, X. S.; Hu, Z.; Jin, M. X.; Liu, H.; Ding, D. J. Mol. Struct. (THEOCHEM) 2003, 638, 215.

(35) Turecek, F.; Hanus, V. Org. Mass Spectrom. 1984, 18, 631. (36) NIST Chemistry WebBook; http://webbook.nist.gov/chemistry/.

face. Intentional exposure of d6-acetone to non-TiO2(110) surfaces (i.e., the sample holder) did not result in these desorption features. The possibility that an impurity, dosed along with d6-acetone, either desorbing at 345 K or generating decomposition products that desorbed at 345 K was excluded because there was no evidence for such impurities in TPD of a thick d6-acetone ice layers. The possibility that the mass 17, 45, and 63 signals at 345 K were due to a trace amount of d5-acetone in the d6-acetone source was ruled out because these masses exhibited a different desorption profile than those masses associated with d6-acetone. The mass 17, 45, and 63 traces tracked each other during the 345 K desorption event, and the relative intensities of the masses among themselves (e.g., mass 17 versus mass 45) at 345 K resembled those for mass 18, 46, and 64. These observations lead to the conclusion that the 345 K signals in the mass 17, 45, and 63 set are from d5-acetone formed from H/D exchange on the surface of TiO2(110) between d6-acetone and H2O adsorbed from the background. There is also a possibility that instead of d5-acetone desorption, the 345 K features could have arisen from desorption of an isomer of d5-acetone desorbed at 345 K. The formation of acetone enolate on oxide surfaces is wellknown in the catalytic literature (discussed in more detail below). It is conceivable that enolation of d6-acetone could result in formation and desorption of d6-2-hydroxypropene, an isomer of d6-acetone, according to the following reaction:

CD3-C(O)-CD3 f CD3-C(OD)dCD2

3454

Langmuir, Vol. 21, No. 8, 2005

Figure 2. TPD spectra of various masses from adsorption of ∼1 ML d6-acetone adsorbed on the reduced surface of TiO2(110) previously exposed to sufficient water to fill all oxygen vacancy sites with OHbr groups. (See text for further details.) The lower two traces (masses 18 and 19) correspond to TPD of OHbr groups in the absence of coadsorbed d6-acetone. Several spectra are displaced vertically and are paired for clarity.

others) fits well with the trend from d0- and d6-acetone. The other ratios listed from d1- and d5-acetone were inconsistent with their respective trends. Evidence for desorption of HDO from H/D exchange between coadsorbed d6-acetone and H2O is shown in Figure 2. In this figure, approximately 1 ML of d6-acetone was adsorbed on a reduced TiO2(110) surface in which the vacancy sites were previously exposed to H2O. The latter treatment was accomplished by preheating a multilayer H2O exposure to 400 K followed by cooling to 95 K for subsequent d6-acetone exposure. TPD spectra (masses 18 and 19) obtained after this treatment (without d6-acetone) are shown at the bottom of Figure 2. Previous work by several groups has shown that water dissociates at oxygen vacancies on TiO2(110), generating double the vacancy concentration of bridging OH groups (OHbr) (see references in ref 2). The amount of recombinative water desorption at 500 K is equivalent to the vacancy population (about 0.07 ML in this case). Although a small amount of molecular water desorption was also detected at ∼310 K, there was no mass 19 desorption signal indicating the absence of D on the surface. The upper portion of Figure 2 shows TPD of ∼1 ML d6-acetone adsorbed on TiO2(110) after filling vacancies with OHbr groups. The majority of the mass 18 signal in the ‘raw’ data was from the CD3 QMS cracking fragment of d6-acetone, but this could be removed by subtraction of an appropriately scaled mass 46 (CD3CO) trace. (Note that small differences in the ‘sharp’ desorption features at 125 and 175 K for the two spectra resulted in incomplete subtraction of these features from the mass 18 spectrum.) The resulting spectrum containing the H2O contribution (and a small amount from OD) shows features at ∼280 K due to molecular water

Henderson

desorption and at ∼500 K due to OHbr recombinative desorption, as well as a weak feature at ∼350 K not typically observed in the TPD of water on clean TiO2(110). The corresponding mass 19 trace also showed evidence for HDO desorption at these three temperatures. (A small contribution at mass 19 from d6-acetone, presumably resulted from ion-molecule reactions in the QMS, yielding CHD3. This contribution was also removed using a scaled version of the mass 46 trace.) In comparison with the H2O desorption trace (corrected mass 18), proportionally more HDO signal was detected in the region between 350 and 450 K. The total amount of H2O + HDO desorbing at 500 K (OHbr/ODbr recombination) was approximately the same as that from H2O on the clean surface (lower traces), although more total water was present in the coadsorbed case, presumably due to additional background adsorption during d6-acetone exposure. Nevertheless, these data indicate that OHbr groups were not consumed or displaced by coadsorbed d6-acetone during the exchange process. This does not preclude involvement of OHbr groups in H/D exchange with d6-acetone, but these results, as well as those discussed below, suggest that the OHbr groups are not directly involved in the exchange process. Also, comparison of the size of the H2O and HDO features evolving below 300 K suggest that molecular water may not be directly involved and that exchange occurs predominately after desorption of molecularly adsorbed water (see below). The total HDO TPD peak area (200-600 K) was about 50% greater in the coadsorption case than the H2O TPD peak area from OHbr recombination on the clean surface (lower traces in Figure 2). The latter was contributed to by about 0.14 ML of H atoms (two OHbr groups per vacancy for the 0.07 ML coverage of vacancies), whereas desorption of HDO possessed one D atom per detected molecule. Therefore, the amount of H/D exchange based on water TPD was about 0.1 ML (1.5 × 0.07 ML) for a ∼1 ML d6acetone starting coverage. This value is about twice that estimated on the basis of calibration of the d5-acetone yield (see below). Additional evidence for H/D exchange between coadsorbed acetone and water was obtained from either combinations of d6-acetone with H2O or d0-acetone with D2O. Figure 3 shows four sets of TPD traces resulting from coadsorption of an acetone isotope with a water isotope on the reduced TiO2(110) surface. The parent signals (in the range of masses 58-64) were used for these experiments even though their signal intensities were roughly half of those from the acetyl QMS cracking fragments (masses 43-46) because the signal from the d1-acetyl cracking fragment of d1-acetone was only a small contribution to the mass 44 signal dominated by background CO2. In each case in Figure 3, the presence of water on the surface ‘split’ the clean surface acetone desorption feature into two peaks. As discussed in the companion paper to this one,16 repulsions between water and acetone destabilize some acetone molecules, resulting in desorption at lower temperature. The peak water desorption occurred in the minimum between the two acetone peaks. The high-temperature acetone desorption feature occurred after water desorption and resembled that seen on the water-free surface. Traces labeled in Figure 3A show the coadsorption case in which 0.55 ML H2O was adsorbed on 0.25 ML d0-acetone. The significance of these data in the context of H/D exchange is to illustrate that only a very weak mass 59 signal was observed in the absence of a deuterated species. This small signal came from H-attachment to the parent d0-acetone molecule within QMS ionizer. (Similar

Acetone and Water on TiO2(110): H/D Exchange

Langmuir, Vol. 21, No. 8, 2005 3455

Figure 3. TPD spectra of various masses from coadsorption of isotopically labeled water and acetone on the vacuum reduced surface of TiO2(110): (A) 0.55 ML H2O on 0.25 ML d0-acetone; (B) 0.5 ML D2O on 0.25 ML d0-acetone; (C) 0.5 ML H2O on 0.18 ML d6-acetone; and (D) 0.5 ML D2O on 0.18 ML d6-acetone. The latter three sets are displaced vertically for clarity.

Figure 4. TPD spectra of various masses from adsorption of ∼1 ML d6-acetone adsorbed on the oxidized surface of TiO2(110) previously exposed to sufficient water to fill all oxygen vacancy sites with OHbr groups. (See text for further details.) Several spectra are displaced vertically and are paired for clarity.

signals are seen in documented mass spectra of d0- and d6-acetone.36) The same mass signals are shown in Figure 3B from coadsorption of 0.5 ML D2O on 0.25 ML d0-acetone. In contrast to the weak mass 59 signal from coadsorption of H2O and d0-acetone in Figure 3A, the mass 59 signal was more intense in Figure 3B for coadsorption of D2O and d0-acetone. This is interpreted to D exchange into d0-acetone to form d1-acetone. As was the case for data in Figure 1, this exchange occurred preferentially in the acetone desorption feature above RT. Complementary data for the coadsorption of 0.5 ML H2O with 0.25 ML d6-acetone is shown in Figure 3C. Comparison of the mass 64 (d6acetone) and 63 (d5-acetone) traces indicates that H exchanged into d6-acetone to generate d5-acetone. As a further control test, the mass 64 and 63 traces labeled in Figure 3D show that coadsorption of 0.5 ML D2O with 0.25 ML d6-acetone resulted in a diminished mass 63 signal at ∼340 K. (As mentioned in the Experimental Section, the D2O source was not isotopically pure, which likely accounts for persistence of a mass 63 signal in the d6acetone and D2O coadsorption experiment.) 3.2. H/D Isotope Exchange on Oxidized TiO2(110). The extent of H/D exchange between acetone and water increased if oxygen vacancy sites on the TiO2(110) surface were exposed to O2 prior to coadsorption. Figure 4 shows TPD data from ∼1 ML d6-acetone adsorbed on oxidized TiO2(110) that compliments that shown in Figure 1. In this case, coadsorbed H2O came from background adsorption (∼0.1 ML). The main QMS cracking fragments from d6-acetone (masses 18, 46, and 64) are paired with the corresponding deuterium-containing fragments from d5acetone (masses 17, 45, and 63) with the same multiplication factors used in Figure 1. The effect of surface oxidation on the chemistry of acetone on TiO2(110) has

been discussed previously.17 Oxygen species (molecular and/or atomic) deposited on the surface during vacancy oxidation by O2 react with adsorbed acetone molecules to generate an acetone-oxygen complex that possesses a C-O stretching frequency midway between that expected from a single and double C-O bond based on HREELS analysis. This acetone-oxygen complex decomposes in TPD at 375 K, liberating predominately acetone in the gas phase but also producing small amounts of other decomposition products (e.g., acetate). Data in Figure 4 show that the amount of exchange significantly increased relative to that observed on the reduced surface (Figure 1). In contrast to the reduced surface case, two temperature regions for H/D exchange were observed on the oxidized surface. A low-temperature exchange process at ∼340 K was similar to that seen on the reduced surface, but an additional exchange process at higher temperature (395 K) was also seen on the oxidized surface. The latter appears related to decomposition of the acetone-oxygen complex that peaks in TPD at slightly lower temperature (375 K). Results in Figures 1-4 indicate that acetone molecules exchange H/D atoms with adsorbed water and/or hydroxyls on TiO2(110), with the exchanged product evolving in TPD above RT. Although not shown in these figures, TPD also detected trace amounts of products resulting from multiple H/D exchange events. For example, a very weak mass 62 signal suggestive of d4-acetone evolved at a slightly higher temperature than the mass 63 signal in the TPD of H2O coadsorbed with d6-acetone. The extent of multiple exchanges is likely limited in these experiments by the constraints of UHV. Other acetone isotopes (d3-, d2-, d1-, and d0-) may form if high coverages of water and acetone were sustained at sufficient temperature or if desorbed molecules were permitted to revisit the surface.

3456

Langmuir, Vol. 21, No. 8, 2005

Figure 5. Fractional yields of d5-acetone resulting from H/D exchange between coadsorbed H2O and d6-acetone as a function of H2O coverage on the oxidized (filled symbols) and reduced (open symbols) surfaces of TiO2(110). Starting coverages of d6acetone were ∼1 ML (triangles) and ∼0.2 ML (circles).

Figure 5 shows data for the extent of single exchange events between d6-acetone and H2O for two coverages of d6-acetone (∼0.2 ML, circles; ∼1 ML, triangles) on the oxidized (filled symbols) and reduced (open symbols) surfaces of TiO2(110). Yields of d5-acetone were determined in the following manner. Signals from acetyl cracking fragments (masses 45 and 46) of d5- and d6-acetone were used instead of the parent signals (masses 63 and 64) because the former were more intense and possessed better signal-to-noise. Contributions to mass 45 from H/D scrambling of d6-acetone in the QMS were removed as discussed above. Assuming that the probabilities for generating d2-acetyl (CHD2CO, mass 45) or d3-acetyl (CD3CO, mass 46) from QMS cracking of d5-acetone (CHD2C(O)-CD3) are equivalent, then the amount of d5-acetone produced from H/D exchange between d6-acetone and H2O should be twice that of the mass 45 signal. The yield of d5-acetone is then equivalent to twice the mass 45 signal divided by the total acetone coverage, which is the sum of twice the mass 45 signal plus the mass 46 signal minus its d3-acetyl contribution from cracking of d5-acetone. This works out to be twice the mass 45 signal divided by the sum of the mass 45 and mass 46 signals. These data are displayed in Figure 5 as a function of the H2O coadsorption coverage. Assuming no exchange in the absence of coadsorbed water, the amount of H/D exchange increased abruptly in each case with RT), the role of water transiently present on the surface during recombinative desorption of OHt groups cannot be ruled out. The activated form of acetone is likely an acetone enolate or propenol species. These results illustrate both the complexity of simple organic chemistry on TiO2 surfaces and the utility of the UHV approach in studying this chemistry. Acknowledgment. This work was funded by the Office of Basic Energy Sciences, Division of Materials Sciences. Pacific Northwest National Laboratory is a multiprogram national laboratory operated for the U.S. Department of Energy by the Battelle Memorial Institute under Contract No. DE-AC06-76RLO 1830. The research reported here was performed in the William R. Wiley Environmental Molecular Science Laboratory, a Department of Energy user facility funded by the Office of Biological and Environmental Research. LA0476581