Ind. Eng. Chem. Res. 2006, 45, 4015-4018
4015
An Examination of the H/D Isotope Substitution Effect on Selectivity and Activity in the Cavitating Ultrasound Hydrogenation of Aqueous 3-Buten-2-ol and 1,4-Pentadien-3-ol on Pd-black K. R. Boyles, S. M. Chajkowski, R. S. Disselkamp,* and C. H. F. Peden Institute for Interfacial Catalysis, Pacific Northwest National Laboratory, 3335 Q AVenue, P.O. Box 999, MS K8-93, Richland, Washington 99352
In this study, cavitating ultrasound (CUS) processing is compared to the traditional (stirred/silent, SS) method in the aqueous phase hydrogenation, via H/D isotope substitution of 3-buten-2-ol (3B2OL) and 1,4-pentadien3-ol (14PD3OL) using a Pd-black catalyst, all at 298 K. The products that are formed include 2-butanol and 2-butanone for 3B2OL, and 3-pentanol and 3-pentanone for 14PD3OL. The H and D isotope-dependent chemistries are accomplished using H2- versus D2-mediated H/D addition, as well as H2O versus D2O solvents for alcohol -OH and -OD isotope substitution. Several conclusions are presented. For example, the H2CUS processing of 3B2OL in water compared to D2O reveals ketone selectivities that are twice as large. This suggests that the alcohol-dependent enol tautomerization to ketone reaction is slower with deuterium, and possibly a rate-controlling event. In addition, for CUS processing of both 3B2OL and 14PD3OL in water, the similarity in ketone selectivities (all ∼17%) for H2 compared to D2 addition suggests that adsorbed H and D isotopes have comparable surface diffusion rates and, hence, result in almost-equal selectivities. 1. Introduction
2. Experimental Section
The effect of ultrasound processing on heterogeneous catalysis has been previously documented,1-4 with most studies focusing on activity enhancement and a smaller subset of studies examining selectivity issues. Our understanding is perhaps most lacking in assigning how ultrasound processing effects processes at the molecular level. An under-utilized technique applied to sonocatalysis is the use of isotope substitution. For hydrogenation reactions, through the use of H2 compared to D2 gas for H/D atom addition, and using H2O compared to D2O solvents for aqueous phase alcohol substitution, a wealth of information can be obtained comparing reaction selectivities and activities. This information, when interpreted, has the potential to yield molecular-level (e.g., mechanistic) information about reacting systems when comparing ultrasound processing to traditional stirred/silent (SS) control experiments.
The experimental approach used here, except for minor variations noted or where discussion is warranted, is the same as that described earlier.5 A Branson Ultrasonics Digital model 450 Sonifier II unit, operating at 20 kHz, was equipped with a jacketed reaction vessel to maintain near-isothermal conditions and operated in batch mode. Reactions used 50 mL of solution (18 MΩ-cm deionized water or 99.9% D2O, Aldrich Chemical Co.) with periodic sampling for gas chromatography/mass spectroscopy (GC/MS) analyses (seven 0.45-µm hydrophilic polytetrafluoroethylene (PTFE)-filtered samples were collected). During either SS or CUS treatments, a total H2 (99.99+% purity, Matheson Tri-gas) or D2 (99.997+% purity, Matheson Tri-gas) pressure of 65 psig (5.4 atm) was used. All substrate concentrations were 100 mM (Aldrich Chemical Co., 97+% purity) and either 1.0, 1.5, or 5.0 mg ((0.2 mg) of Pd-black (Aldrich Chemical Co., 99.9% purity, metals basis, 42 m2/g N2 BET surface area) were utilized. The bulk reaction temperature was maintained at 298 ( 3 K using a water-bath circulation unit (even during ultrasound treatment; see ref 9 for details). For all experiments, the catalyst was pre-reduced using noncavitating ultrasound for 2 min at 90% amplitude (360 W). Next, the substrates and inert dopant were added to the reaction mixture. For the 3B2OL substrate, 100 mM of 1-propanol inert dopant was used. Conversely, for the 14PD3OL substrate, 275 mM of 1-propanol was used. The greater dopant concentration used in the 14PD3OL systems was due to the comparatively greater difficulty in cavitating this (di-olefin) substrate. The concentrations chosen, in part, were based on a cavitation time onset of less than ca. 6 s for each substrate. In our study here for 14PD3OL, CUS processing yielded no difference in selectivity when 1-propanol was used, versus 50 mM 1-pentanol inert dopant; therefore, we will not consider the different amounts of inert dopants that have been used to be significant. Samples collected following SS or CUS processing were run on an Hewlett-Packard model 5890/5972MSD GC/MS system, using a Cyclosil-B 30-m column. The chiral column provided no additional benefit, compared to a nonchiral column in this
In principle, either noncavitating or cavitating conditions can exist during ultrasound processing.5 Consistent with earlier work by our group,5-7 here, we focus on comparing cavitating ultrasound (CUS) to traditional (SS) processing, because this affords the greatest contrast in selectivity and activity. The aqueous hydrogenation investigation here examines the substrates of 3-buten-2-ol (3B2OL) and 1,4-pentadien-3-ol (14PD3OL) on Pd-black. The observed products that are formed for these systems include the saturated alcohol and ketone for each substrate, in particular, 2-butanol and 2-butanone for 3B2OL, and 3-pentanol and 3-pentanone for 14PD3OL. Upon assuming a standard olefin hydrogenation mechanism for each substrate,8 where a H or D atom addition occurs, yielding an adsorbed alkyl radical intermediate prior to full hydrogenation, several tentative conclusions about the effects of CUS at the molecular level can be proposed. * To whom correspondence should be addressed. Tel. 509-376-8209. FAX 509-376-5106. E-mail:
[email protected].
10.1021/ie051347j CCC: $33.50 © 2006 American Chemical Society Published on Web 04/21/2006
4016
Ind. Eng. Chem. Res., Vol. 45, No. 11, 2006
Scheme 1. Reaction Scheme for 3-Buten-2-ol (3B2OL)
study, nor did it hinder or make analyses cumbersome. Experiments were repeated at least twice, and duplicate injections were used. Whenever possible, single-point calibrations of total ion counts to concentrations that use authentic standards (i.e., substrate, saturated alcohols, and final ketones) were used. Finally, we utilized both CUS processing and SS reaction methods. During CUS treatment, the applied power was 270 W at 100% amplitude. 3. Results and Discussion A. Proposed Reaction Mechanisms. Based on the generally accepted model of olefin hydrogenation of Horiuti and Polayni,8 it is straightforward to postulate the reaction mechanism that leads to the formation of ketone and saturated alcohol for our two substrates. A postulated reaction mechanism for 3B2OL is given in Scheme 1, where it can be observed that, initially, a single H-atom addition to the substrate occurs, leading to the C3 alkyl radical intermediate. This intermediate, in turn, serves as a source of the two products (i.e., a branching point) in that either a second H-atom adds to the surface-bound alkyl radical, leading to 2-butanol (via reaction 2), or that a unimolecular C2 H-atom elimination occurs from the substrate to yield the enol (via reaction 3) that then tautomerizes to 2-butanone (via reaction 4). Based in Scheme 1, it is straightforward to predict, at least qualitatively, the effect of H/D isotope substitution on the various reaction steps. For example, using D2O instead of water as a solvent will result in deuteration of the (substrate) alcohol throughout the reaction sequence. Thus, aside from general solvent effects (which are assumed to affect activity more than selectivity) for all reaction steps, this scheme predicts that primarily the enol tautomerization step will be most affected, Scheme 2. Reaction Scheme for 1,4-Pentadien-3-ol (14PD3OL)
because the olefin/O-H versus olefin/O-D bond rearrangement processes can be expected to be the most different. Similarly, when comparing H2 versus D2 hydrogenation, the selectivity is expected to be different, because the H-addition step (via reaction 2) may differ between H-addition and D-addition; however, the unimolecular H-elimination step (via reaction 3) is identical, because it remains unchanged when different hydrogenation gases are used. Therefore, it can be concluded that differences will arise solely from differences in H- versus D-atom addition (via reaction 2). A somewhat more involved reaction scheme is required for 14PD3OL hydrogenation, and that is illustrated in Scheme 2. In this scheme, for 14PD3OL, because this substrate is a diolefin, there is hydrogenation (H-atom addition via reaction 1) to the first C4 alkyl radical intermediate, which serves as the first branching point for either further hydrogenation (reaction 2) or enol formation (reaction 5). The latter enol, again, is expected to undergo tautomerization to the potentially stable intermediate 1-penten-3-one (reaction 7) and, from here, eventually hydrogenate to 3-pentanone (reaction 9). Alternatively, the intermediate 1-penten-3-ol can gain a second H-atom (reaction 3) to a second C2 surface-bound alkyl radical and serve as a second branching point for either additional hydrogenation to the saturated alcohol (3-pentanol, reaction 4), or another unimolecular H-elimination process (reaction 6) to the enol and (via reaction 8) form 3-pentanone. Thus, because there are two surface-bound alkyl radicals (i.e., two branching points), there are two opportunities for 3-pentanone formation, whereas Scheme 1 (presented previously for 3B2OL) only had one opportunity for ketone formation. B. Data Interpretation. Numerous experimental combinations of process conditions (SS or CUS), hydrogenation gas (H2 or D2), and solvent (H2O or D2O) have been explored. A summary of the combinations that we have chosen for study is presented in Table 1. In this table, the experiments for 3B2OL are labeled B1-B7 and those for 14PD3OL are labeled P1P6. The second column lists the experimental conditions, whereas the third column lists the initial substrate concentrations based on 100 mM of substrate and the amount of catalyst used. The penultimate column lists the final (extent of reaction >95%) selectivities to ketone (2-butanone or 3-pentanone) and the final column lists the pseudo-first-order substrate reaction rate
Ind. Eng. Chem. Res., Vol. 45, No. 11, 2006 4017 Table 1. Summary of 3-Buten-2-ol (3B2OL) and 1,4-Pentadien-3-ol (14PD3OL) Experiments substrate
conditionsa
concentration (M/g-cat.)
% selectivity to final ketoneb
k (min-1)
3B2OL - B1 3B2OL - B2 3B2OL - B3 3B2OL - B4 3B2OL - B5 3B2OL - B6 3B2OL - B7 14PD3OL - P1 14PD3OL - P2 14PD3OL - P3 14PD3OL - P4 14PD3OL - P5 14PD3OL - P6
H2/H2O - CUS H2/D2O - CUS H2/H2O - SS H2/D2O - SS D2/H2O - CUS D2/H2O - SS D2/H2O - SS H2/H2O - CUS H2/D2O - CUS H2/H2O - SS D2/H2O - CUS D2/H2O - SS D2/H2O - SS
67. 67. 67. 67. 67. 67. 20. 100. 100. 100. 100. 100. 20.
16. 9. 42. 32. 17. 36. 68. 17. 16. 28. 18. 32. 42.
2.7 4.3 0.0038 0.010 3.3 0.012 0.030 3.5 2.3 0.0068 2.3 0.0070 0.042
a The abbreviations CUS and SS are defined as cavitating ultrasound and stirred/silent processing, respectively. b The percent selectivity to final ketone plus saturated alcohol sum to 100%.
coefficients. The estimated uncertainty (95% confidence interval) of the pseudo-first-order rate coefficients are (20%, whereas the error uncertainty for the ketone selectivities are (3%. The dataset contained in Table 1 enables numerous conclusions to be made regarding the reaction systems. The differences in initial concentrations (e.g., 67 versus 100 M/g-cat.) arise from the chosen convenience of having similar activities and, therefore, comparable reaction times. The conclusions from these datasets, many of which are tentative, are now discussed in detail. First, the activity of CUS compared to SS processing were at least ∼100-fold larger, which is consistent with earlier findings.5 This is readily observed by comparing the reaction rate coefficients of the CUS to SS processes for, e.g., experiments B1 versus B3 for 3B2OL or experiments P1 versus P3 for 14PD3OL. Second, an examination of variable catalyst loading experiments for SS-D2 processing has been performed to test whether mass transfer of D2 gas to the catalyst surface, or its subsequent D-atom surface diffusion, affects the reaction selectivity (compare experiment B6 versus experiment B7, and experiment P5 versus experiment P6). These experiments show that a higher catalyst loading resulted in an increased ketone selectivity. Thus, this suggests that mass transfer of D2 gas to the palladium surface may have had a role, such that a higher catalyst loading (e.g., larger catalyst surface area) could have reduced the surface D-atom concentrations on the catalyst surface and, in turn, reduced the saturated alcohol formation (e.g., via D-atom addition to surface alkyl radicals). According to our postulated schemes, this corresponds to a reduction of reaction 2, relative to reaction 3 of Scheme 1, and reaction 2, compared to reaction 5, and/or reaction 4, compared to reaction 6 of Scheme 2. Third, a comparison of ketone selectivities for water to D2O solvents can be made to test whether the enol rearrangement process has a significant role in determining the product yield. This is accomplished by examining the H2-CUS processing of 3B2OL in water versus D2O (compare selectivities for experiments B1 and B2). The ketone selectivity is observed to be a factor of ∼2 larger for the water solvent. It can be suggested that the larger ketone selectivity for water arises from a larger -OH, compared to -OD enol tautomerization to ketone rearrangement (e.g., reaction 4 of Scheme 1), which may be a H/D-atom tunneling event and is substantially slower with deuterium and possibly a rate-controlling process. This 3B2OL result is perplexing in that only enol rearrangement, as discussed previously, is thought to be affected through alcohol deuteration, and that it is generally understood that enol tautomerization is rapid and not rate-controlling. In contrast to the 3B2OL system,
the comparable 14PD3OL system (compare experiments P1 and P2) suggests, for this more-complex reaction system, a negligible H/D isotope difference. Here, enol tautomerization may be sufficiently rapid for both alcohol -OH and -OD isotopes, so as to not manifest itself in a ketone selectivity difference. Further support for this assignment is derived from the observation of initial pseudo-first-order substrate decay rate coefficient values. Namely, comparison of experiments B1 and P1 (4.0 and 3.5 M/g-cat. min, respectively) shows that their similar activity and ketone yields suggest that, for OH enol rearrangements, they are not rate-controlling factors. Conversely, for the alcohol deuteration experiments B2 and P2, their activities differ by a factor of ∼2.8 (6.4 and 2.3 M/g-cat. min, respectively) and their ketone selectivities also differ by a factor of ∼2. Thus, the slowed -OD enol tautomerization for 3B2OL only has likely become a rate-controling factor. A fourth issue that can be examined by the dataset of Table 1 is the comparison of CUS-water solvent processing, using H2 versus D2 gas in each substrate’s hydrogenation. In principle, this information may yield tentative information about the hydrogenation process of these two isotopes, such as H-atom versus D-atom addition to chemisorbed alkyl radical intermediates (compare experiment B1 to experiment B5 and experiment P1 to experiment P4). In these CUS processing experiments, they all have almost-identical ketone selectivities of ∼17%. Therefore, it can be suggested that differences arising from the use of H2, compared to the use of D2, indicates that both H/D isotopes have facile surface diffusion and, hence, result in almost-equal selectivities. Furthermore, their comparable H/Dsurface diffusion rate processes suggest that their CUS surface diffusion activation energy lies well below the actual energy contained at the catalyst surface, perhaps because of the (extreme) conditions present during CUS processing. A final issue worthy of mention is a discussion of the concentration of intermediates during the reactions as measured by the sampling taken during the course of all reactions investigated. The experiments that use 3B2OL did not yield any intermediates of notable concentration. However, the experiments on 14PD3OL, conversely, formed a substantial 1-penten3-ol (1PE3OL) intermediate concentration in the experiments. For example, experiment P4 of Table 1 yielded an intermediate 1PE3OL concentration of 42% for the D2/H2O CUS system. According to Scheme 2 (for 14PD3OL, presented previously), it can be suggested that the reaction velocity of the combined reactions 1 and 2 are of comparable magnitude to that of reaction 3, yielding substantial 1PE3OL. Furthermore, data for the D2/ H2O-SS system contained in experiment P5 of Table 1 yielded an intermediate 1PE3OL concentration of only 18%. The
4018
Ind. Eng. Chem. Res., Vol. 45, No. 11, 2006
observed differences between the 1PE3OL intermediate concentrations for the CUS system of 42% (experiment P4) compared to the SS result of 18% (experiment P5) is not known, but may be due to the possibility that 1PE3OL for the SS system desorbs less from the palladium surface following reaction 2, compared to CUS processing. The cause of this phenomenon would be the enhanced energy gained by the 1PE3OL intermediate at the palladium surface arising from CUS processing. Acknowledgment K.R.B. was supported from a Department of Energy (DOE) Community College Initiative (CCI) award, whereas S.M.C. was supported from a DOE Pre-service Teachers (PST) award, each for their 10-week internship at Pacific Northwest National Laboratory (PNNL). This project was performed in the W.R. Wiley Environmental Molecular Sciences Laboratory (EMSL) at PNNL, and funded in part from a Laboratory Directed Research and Development (LDRD) grant administered by PNNL. PNNL is operated by Battelle Memorial Institute for the U.S. Department of Energy. The authors would like to thank Dr. James F. White (PNNL) for stimulating discussions during the course of this work. Literature Cited (1) Suslick, K. S. In Handbook of Heterogeneous Catalysis, Vol. 3; Ertl, G., Knozinger, H., Weitkamp, J., Eds.; Wiley-VCH: Weinheim, Germany, 1997.
(2) Mason, T. J. Sonochemistry; Oxford University Press: Oxford, U.K., 2000. (3) Luche, J. L., Ed. Synthetic Organic Sonochemistry; Plenum: New York, 1998. (4) Torok, B.; Balazsik, K.; Torok, M.; Szollosi, Gy.; Bartok, M. Asymmetric sonochemical reactionssEnantioselective hydrogenation of alpha-ketoesters over platinum catalysts. Ultrason. Sonochem. 2000, 7, 151. (5) Disselkamp, R. S.; Ya-Huei Chin, Peden, C. H. F. The effect of cavitating ultrasound on the heterogeneous aqueous hydrogenation of 3-buten-2-ol on Pd-black, J. Catal. 2004, 227, 552. (6) Disselkamp, R. S.; Denslow, K. M.; Hart, T. R.; White, J. F.; Peden, C. H. F. The effect of cavitating ultrasound on the aqueous phase hydrogenation of cis-2-buten-1-ol and cis-2-penten-1-ol on Pd-black. Appl. Catal., A 2005, 288, 62. (7) Disselkamp, R. S.; Peden, C. H. F. The Effect of Ultrasound on the Isomerization versus Reduction Reaction Pathways in the Hydrogenation of 3-Buten-2-ol and 1,4-Pentadien-3-ol on Pd-black. In 2004 Catalysis of Organic Reactions (ORCS) Proceedings; Sowa, J. R., Jr., Ed.; Taylor and Francis: New York, 2005; p 303. (8) Horiuti, J.; Polayni, M. Exchange Reaction of Hydrogen on Metal Catalysts. Trans. Faraday Soc. 1934, 30, 1164. (9) Disselkamp, R. S.; Judd, K. M.; Hart, T. R.; Peden, C. H. F.; Posakony, G. J.; Bond, L. J. A comparison between conventional and ultrasound-mediated heterogeneous catalysis: hydrogenation of 3-buten1-ol aqueous solutions, J. Catal. 2004, 221, 347.
ReceiVed for reView December 2, 2005 ReVised manuscript receiVed March 22, 2006 Accepted March 28, 2006 IE051347J
