Direct Addition Mechanism during the Catalytic Hydrogenation of

Jun 16, 2016 - Direct Addition Mechanism during the Catalytic Hydrogenation of Olefins over Platinum Surfaces ... *E-mail: [email protected]. .... Figure ...
2 downloads 6 Views 1MB Size
Download PDF
Letter pubs.acs.org/JPCL

Direct Addition Mechanism during the Catalytic Hydrogenation of Olefins over Platinum Surfaces Yujung Dong, Maryam Ebrahimi,‡ Aashani Tillekaratne,# and Francisco Zaera* Department of Chemistry and UCR Center for Catalysis, University of California, Riverside, California 92521, United States S Supporting Information *

ABSTRACT: The mechanism of the hydrogenation of olefins catalyzed by metal surfaces was probed by using isotope labeling in conjunction with a high-flux effusive molecular beam setup capable of sustaining steady-state conversion under wellcontrolled ultrahigh vacuum (UHV). The unique conditions afforded by this instrument, namely, a single collision regime and impinging frequencies equivalent to pressures in the mTorr range, led to the clear identification of two competing pathways: a multiple H−D isotope exchange channel explained by the well-known Horiuti−Polanyi mechanism but with an unusually high probability for β-hydride elimination from the alkyl surface intermediate (versus its reductive elimination to the alkane), and a direct addition route that produces dideuterated alkanes selectively. The latter may follow an Eley−Rideal mechanism involving an adsorbate (either the olefin or the hydrogen/deuterium atoms resulting from dissociative adsorption of H2/D2) and a gas-phase molecule (the other reactant), or, alternatively, it could reflect the limited diffusion of the hydrogen atoms on the surface under catalytic conditions because of site blocking by the islands of strongly bonded carbonaceous (alkylidyne) layers present during catalysis. Regardless, our data clearly show that the distribution of alkane isotopologues obtained from the conversion of olefins with deuterium can deviate significantly from statistical expectations.

A

hydrogen in the reaction mixture, the resulting alkanes display a wide range of isotope substitutions.12−17 These distributions of alkane isotopologues have in general been accounted for by multiple olefin−alkyl interconversions within the Horiuti− Polanyi mechanism, but deviations from what can be justified that way have been reported and explained via the recollection of additional mechanistic steps, which have included the disproportionation of adsorbed olefins and/or alkyl moieties, the cross reaction between adsorbed olefins and alkyls or between either of those species and molecular deuterium, the initial dehydrogenation of adsorbed olefins to form vinyl surface intermediates, and the adsorption and reaction of hydrogen in different types of competitive and noncompetitive adsorption sites.4,15,18,19 All of these mechanistic possibilities were debated in the early literature but have been almost forgotten because of the lack of compelling evidence to make a clear case for their need to explain the experimental isotopologue distributions. Particularly relevant to our discussion is the proposal by Bond of a “direct addition” mechanism (graphical abstract, bottom pathway), which he introduced to justify the excess in CnH2nD2 production sometimes seen in the CnH2n + D2 conversions.4,13,20 To specifically probe this step, we have designed an experiment based on the use of high-flux effusive molecular beams where olefin hydrogenation and H−D isotope exchange reactions are catalyzed by a platinum surface under

lthough the catalytic hydrogenation of olefins promoted by transition metals is one of the oldest and most studied processes used in industry,1−3 some details of its basic mechanism remain unresolved to date.4,5 In this study, we use a newly developed high-flux molecular beam instrument to explore a previously suggested but not proven pathway involving the direct addition of hydrogen molecules to the carbon−carbon double bond. We provide kinetic evidence for such a pathway based on results from experiments using isotope labeling and argue that the addition may not be direct from the gas phase but rather related to the limited mobility of the hydrogen atoms on the surface resulting from the H2 (or D2) dissociative adsorption. The general framework to explain the surface chemical steps involved in olefin hydrogenation on metal surfaces was provided almost a century ago by Horiuti and Polanyi.6 Their key proposal is that hydrogenation takes place in a stepwise fashion, via the sequential and reversible incorporation of single surface hydrogen atoms to first form an alkyl intermediate and then the final alkane product (graphical abstract, top pathway). Double-bond migration and isomerization reactions, as well as H−D isotope scrambling, compete with this hydrogenation via the rapid interconversion between the adsorbed olefin and alkyl species. Selectivity among those reactions is defined by the relative rates of the steps that the half-hydrogenated surface alkyl moiety can undergo: reductive elimination with adsorbed atomic hydrogen to form the alkane7,8 or β-hydride elimination to regenerate an adsorbed ethylene molecule.9−11 Much of the knowledge available on this mechanism has been generated by experiments using isotope labeling. It has been well established that when deuterium is used instead of © XXXX American Chemical Society

Received: May 20, 2016 Accepted: June 14, 2016

2439

DOI: 10.1021/acs.jpclett.6b01103 J. Phys. Chem. Lett. 2016, 7, 2439−2443

Letter

The Journal of Physical Chemistry Letters single-collision conditions.21,22 Our experiments are carried out under the controlled ultrahigh vacuum (UHV) environment typically used in modern surface-science experiments but with collimated gas beams that afford the emulation of the highimpinging-frequencies conditions encountered in catalysis locally on a small (∼2 mm) spot on the surface of the catalyst. Several characteristics of our experimental setup make it unique and ideal for the study of the direct addition mechanism, namely: (1) because the reactants impinge on the surface only once and because the products are immediately pumped away upon their desorption from the surface, only primary reactions are probed; there is no interference from possible readsorption of isotope-exchanged olefins, for instance; (2) the beam fluxes used are equivalent to an intermediate pressure range, between UHV and the atmospheric values typically used in catalysis, where the transition to catalytic kinetics occurs;5 and (3) the high hydrogen-to-olefin mixture ratios used are expected to enhance the hydrogenation reaction and to minimize interference from the strongly adsorbed hydrocarbon adsorbates known to form on the surface of the hydrogenation catalysts.5,23,24 Figure 1 displays a typical example of the data obtained with our molecular beam arrangement, in this case for the

isotopologues, indicating the occurrence of multiple H−D exchange steps during a single adsorbed ethylene residence time, and also that the reaction can be sustained catalytically, in this case at a total TOF of approximately 2 ML/s for a minimum of 300 s (resulting in a total turnover number, TON, of over 600 molecules per platinum surface atom). An alternative way to analyze the composition of the reaction mixtures is by taking a full mass spectrum of the gas while under steady-state catalytic conditions. Typical data from such experiments are reported in the left panel of Figure 2, in this

Figure 2. Contrast of mass spectra data obtained from experiments carried out with our high-flux molecular beam (MB, top, red trace) versus in a high-pressure cell (HP, batch reactor; bottom, blue trace). The left panel shows the raw traces, and the right panel the product distributions obtained from their deconvolution. Particularly noteworthy is the dual maxima seen in the product distribution from the molecular beam experiment. Also provided is the molecular beam deuterium distribution calculated by using a model based on the Horiuti−Polanyi mechanism (filled red circles), as discussed in the text.

case for a D2:C2D4 40:1 ratio and a total flux of 8000 ML/s (top, red trace). A second mass spectrum (bottom, blue trace) is reported for comparison, from a more conventional experiment carried out by using a so-called high-pressure cell, which is used to isolate the Pt surface from the UHV environment and enclose it in a small volume that can be pressurized and used as a batch reactor.21,26 The fact that both deuteration and H−D isotope scrambling take place in both of these experiments is immediately apparent by the detection of significant signal intensities all the way to 36 amu. On the other hand, the final product distributions are clearly different in each case. This is better evaluated after deconvolution of the data by comparing the resulting ethane product distributions (Figure 2, right panel). The distribution obtained with the high-pressure cell peaks at ethane-d1 and decreases in an approximately exponential way for the ethanes with higher deuterium substitutions, a result quite typical of these systems.13,16,17,27 The distribution seen in the molecular beam experiments, on the other hand, is unique. For one, it shows two (instead of one) local maxima, at ethane-d2 and ethane-d6. Moreover, above ethane-d3, the yield increases with increasing number of deuterium substitutions. As far as we know, this type of alkane isotopologue distribution is unprecedented and has not been reported before with any olefin or transition-metal catalyst. The high-deuterium end of these distributions can be easily

Figure 1. (Left) Kinetic data from a high-flux molecular beam experiment in the form of mass spectrometer signal versus time for the masses in the 27−36 amu range. (Right) Corresponding data for the evolution of the reaction rates of the different ethylene and ethane isotopologues, extracted via deconvolution of the raw data and expressed in terms of TOFs (in molecules per surface Pt atom per second). High conversion rates were sustained in these experiments in a steady-state regime typical of catalytic processes; a total TON (in molecules per Pt surface atom) of about 600 was measured in this case.

deuteration of ethylene using a 250:1 D2 + C2H4 reaction mixture and a total beam flow of 1750 ML/s (monolayers, in molecules per platinum surface atom, per second). The left panel displays the evolution of the raw mass spectrometer signal intensities for the key masses as a function of time, and the right panel shows the result from deconvolution of those data using an established matrix-based procedure25 to extract information on the time evolution of the turnover frequencies (TOFs, in ML/s) of the products. It can be seen from the results that the product mixture contains all ethane 2440

DOI: 10.1021/acs.jpclett.6b01103 J. Phys. Chem. Lett. 2016, 7, 2439−2443

Letter

The Journal of Physical Chemistry Letters

justified by the logarithmic distribution implied by the use of σ is negligible; most of the C2H4D2 detected ( f DA = 0.96) must come from a different source, the direct addition pathway in the terminology introduced by Bond. This is also unique because in the past deviations in the ethane-d2 yield from those predicted by the Horiuti−Polanyi mechanism have amounted to about half of the total at most regardless of the nature of the catalyst or the support, the metal loading, the reaction temperature, the extent of the reaction, or the gas composition (Table 1).13,16,17,27 We report values close to 100% instead. The high deuterium content of our beams may explain some of this difference, but that cannot be the whole story because past experiments with deuterium-rich mixtures (sixth entry in Table 1, for instance)13 did not result in the same high values for f DA seen here. We think that the deviations from the past results in our data are due to the different pressure regime explored in our work; on the basis of the kinetic theory of gases, the 1 × 103−1 × 104 ML/s total flux range of our molecular beam corresponds to impinging frequencies similar to those in a gas at pressures between 1 and 10 mTorr, several orders of magnitude lower than those typically used in catalysis. Regarding the differences in σ, a more complex explanation is in order. The kinetic behavior of our D2 + C2H4/Pt system was followed systematically as a function of the different controllable parameters of the experiment. The most interesting trends were seen as a function of ethylene flux, F(C2H4), as shown in Figure 3. Because the total flux was held constant in the

understood in terms of the relative rates of the steps in the Horiuti−Polanyi mechanism (graphical abstract, top line) by using a formalism developed by Kemball,14 or, alternatively, by employing the simplified scheme advanced by Bond.13 The latter requires the fit of only two parameters to the data; the first is the ratio σ of the yield of alkane-dn to that of alkane-dn−1 (for n ≥ 3 in the case of ethane), which under most conditions reflects the ratio of the rates of the β-hydride versus reductive elimination steps from the adsorbed alkyl intermediates, and the second is the fraction f DA of the ethane-d2 produced that cannot be accounted for by extrapolation of the high-deuterium substitution side of the distribution using this σ parameter (the fraction that Bond associated with the direct addition mechanism). The best fit of our experimental molecular beam distribution to a single parameter is reported as filled red circles in the right panel of Figure 2. The fact that our σ parameter is higher than unity (σ = 1.97) reflects a large preference for β-hydride over reductive elimination steps. It should be noted that this is usually not the case; the values of σ calculated from data reported in the past are all around 0.5 (selected numbers are provided in Table 1). Also, the high σ value is at first sight Table 1. Kinetic Parameters from Fits of the Deuterium Distributions Reported in Figures 1−3 to a Model Based on the Horiuti−Polanyi Mechanisma experiment Pt/SiO2f

% conversion

P/Torr or F/ML s−1 b

D2:C2H4

σd

f DAe

1% 0.1% Pt/SiO2g 0.04% Pt/ SiO2h 1% Pt/SiO2f 1% Pt/Al2O3f 1% Pt/Al2O3f HP cell, Pt(111)i

48 15 1−5

118 200b 175b

1.3 1 6

0.49 0.52 0.60

0.40 0.49 0.55

100 47 39 100

115 (273 K)b 111.5b 255 (273 K)b 30b

8.3 1.2 125 2

0.48 0.42 0.21 0.44

0.59 0 0.19 0.26

HP cell, 1 min HP cell, 10 min

70 100

52b 52b

25 25

0.75 0.60

0.55 0.50

MBj MBk MBl MBl MBl

30 80 75 65 15

1750c 8000c 7000c 7000c 7000c

250 40 10000 5000 100

1.20 1.97 11.9 2.65 0.92

0.94 0.96 1.00 0.97 0.92

a

Details of this model are provided in the text. Selected data from the literature are added for reference. T ≈ 300 K. bP/Torr. cF/ML s−1 dσ = C2H5−xDx+1/C2H6−xDx (x ≥ 3). ef DA = fraction of C2H4D2 made via the direct addition of D2 to C2H4. fReference 13. gReference 16. h Reference 27. iReference 17. jFigure 1. kFigure 2. lFigure 3.

Figure 3. (Left) Average number of deuterium substitutions per molecule of ethane produced ⟨D⟩ (top, red trace) and conversion (as a molar fraction; bottom, blue trace) as a function of ethylene flux for experiments using a total beam flux of 7000 ML/s. (Right) Corresponding ethane isotopologue distributions.

somewhat counterintuitive because high D2:C2H4 ratios, as used here, should favor the bimolecular reductive elimination reaction: the rate of that step depends on the surface coverage of hydrogen/deuterium atoms. It would appear that overall accessibility to surface open sites, which is inhibited by the high hydrocarbon coverages produced at low hydrogen/deuteriumto-ethylene ratios, is what controls the ability of the ethylene molecules to undergo H−D scrambling. Further discussion of this point is provided below. Another observation that derives from the molecular beam data in Figure 2 is that the amount of ethane-d2 that can be

reported sequence, higher olefin fluxes correspond to lower D2:C2 H4 ratios, which in Figure 3 were varied from approximately 10000 to 100. Several clear trends can be extracted from these data. First, the conversion decreases with increasing ethylene flux because the reaction rate is known to display a negative reaction rate order with respect to ethylene pressure.4 Second, the average number of deuterium substitution per ethane molecule also decreases with increasing ethylene flux. This is because less multideuterium-substituted ethane molecules are produced, as also indicated by a decrease in the value of σ (Table 1); the effect is due to the decrease in 2441

DOI: 10.1021/acs.jpclett.6b01103 J. Phys. Chem. Lett. 2016, 7, 2439−2443

Letter

The Journal of Physical Chemistry Letters the relative D2:C2H4 ratio. Finally, the value of f DA also decreases with F(C2H4) (Table 1). It is worth noting, however, that the contribution from the direct addition mechanism is quite high in all of the experiments reported here; the lowest value measured was f DA = 0.92, for the D2:C2H4 = 100 ratio. In summary, the combination of the use of high D2:C2H4 ratios and an experimental setup to study catalytic reactions under single-collision conditions and in an impinging rate regime equivalent to mTorr pressures has afforded the identification of a unique kinetic regime for the catalytic deuteration of ethylene where (1) the probability for isotope scrambling on the surface is much higher than that for the hydrogenation of the ethyl surface intermediate and (2) the fraction of ethane-d2 produced is unusually high, with more than 90% originating from direct incorporation of two deuterium atoms on ethylene that has not gone through any H−D scrambling. It is tempting to interpret the latter result as an indication of the direct incorporation of D2 molecules from the gas phase into adsorbed ethylene (as the term “direct addition” suggests), even if this type of so-called Eley−Rideal20 mechanism is quite rare in heterogeneous catalysis28,29 and difficult to justify here given the steric and energetic constrains implied (the D−D bond is quite stable and difficult to break). Intriguingly, recent DFT calculations have identified a lowenergy “reverse” Eley−Rideal mechanism whereby olefin molecules arriving from the gas phase react directly with hydrogen/deuterium atoms on the surface.30 This mechanism was proposed for the formation of alkyl intermediates but could be conceived for the direct incorporation of two deuterium atoms in one single step as well. At first sight, such concerted addition of two deuterium atoms at once to an incoming olefin molecule may seem implausible, but the limited mobility that deuterium atoms have on the surface of the working catalyst surface (discussed in more detail below and elsewhere)31 provides for the possibility of the placement of the two D atoms in the appropriate geometry for the three-body transition state required in this case. We would like to offer yet another alternative explanation for the direct addition mechanism, one that still involves adsorbed hydrogen/deuterium atoms, not gas-phase molecules, but acknowledges the possibility of their limited mobility on the surface. This idea has been recently advanced by us based on independent studies on the H2 + D2 isotope scrambling during olefin hydrogenation catalysis.31 It was determined that the initial rate of HD production from H2 + D2 + C2H4 mixtures can be quite low but can switch in a highly nonlinear way halfway during the course of the olefin conversion. We hypothesize that the transition occurs when the size of the islands of the strongly bound hydrocarbon deposits (alkylidynes) that are known to be present on the surface of the hydrogenation catalysts17,21,32,33 is reduced below a critical size, at which point channels of clean Pt sites are opened to allow for the adsorbates to move more freely. Before that point is reached, we suggest that the adsorption sites available for hydrogen molecules are isolated, so that the H and D atoms deposited from H2 and D2 adsorption, respectively, cannot scramble and recombine as HD. If this is the case, it could be conceived that the same nonstatistical distribution of hydrogen isotopes on the surface could force many individual ethylene molecules to adsorb next to and incorporate two adjacent deuterium atoms (produced by dissociation of one single D2 molecule) and that multiple ethylene−ethyl interconversions would be confined to the involvement of those atoms only and

to not lead to any change in the overall isotopologue composition of the final ethane product. Unfortunately, it is difficult to test this model experimentally or to evaluate the relative strengths and weaknesses of this versus the other explanations. Experimentally, the kinetics of these systems are too complex to use for the direct differentiation between the typical Langmuir−Hinshelwood (with both reactants adsorbed on the surface) and the rarer Eley−Rideal (which involves a species from the gas phase) mechanisms; the competitive adsorption of the reactants for different surface sites, the presence of carbonaceous deposits on the surface, and the limited mobility of the adsorbed species all interfere with a straight kinetic analysis. It may be possible to reproduce the experimental deuterium distributions with Monte Carlo calculations that incorporate the spatially heterogeneous distribution of adsorbates on the surface (mean-field deterministic kinetics would not be able to capture this effect), but that is not a trivial task, and it is beyond the scope of our current report. What is clear is that we have shown experimentally that it is possible to set up olefin catalytic hydrogenation reactions under conditions where the isotopologue product distributions (when deuterium is used) deviate significantly from that expected based on statistical arguments using the well-established Horiuti−Polanyi mechanism. We believe that the reported new distributions provide compelling, albeit indirect, evidence for the coexistence of two competing hydrogenation mechanisms, the stepwise Horiuti−Polanyi pathway and a second “direct addition” route. This is a central mechanistic idea widely discussed in the past but not, to date, supported by the available data. In that context, we suggest that our results may open new avenues to revive this discussion.

■

EXPERIMENTAL SECTION The molecular beam kinetic measurements were carried out in an ultrahigh vacuum (UHV) apparatus described in previous publications.21,22 A polished 1 cm in diameter polycrystalline Pt disk approximately 1 mm in thickness was used as the catalyst. The central element of our instrument is the gas doser made out of a single capillary, 150 μm in diameter and 1.2 cm in length, used to generate the collimated high-flux effusive molecular beams required to sustain the catalytic olefin conversion. The reaction mixtures were premixed in a calibrated volume placed behind the leak valve used to set the beam flux, which was quantified by following the drop in the pressure in the backing volume versus time using a MKS Baratron capacitance manometer and is reported in units of ML/s (monolayers per second, assuming a Pt surface density of 1.5 × 1015 Pt sites/cm2). A UTI 100C quadrupole mass spectrometer interfaced to a personal computer was used to follow the partial-pressure data during the kinetic runs. More details on this instrument and the procedures used are provided in the Supporting Information.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpclett.6b01103. Experimental details (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. 2442

DOI: 10.1021/acs.jpclett.6b01103 J. Phys. Chem. Lett. 2016, 7, 2439−2443

Letter

The Journal of Physical Chemistry Letters Present Addresses

genation of Ethylene on Pt(111) Single Crystal Surfaces. ACS Catal. 2012, 2, 2259−2268. (22) Ebrahimi, M.; Simonovis, J. P.; Zaera, F. Near-Unity Reaction Probability in Olefin Hydrogenation Promoted by Heterogeneous Metal Catalysts. J. Phys. Chem. Lett. 2014, 5, 2121−2125. (23) Thomson, S. J.; Webb, G. Catalytic Hydrogenation of Olefins on Metals: A New Interpretation. J. Chem. Soc., Chem. Commun. 1976, 526−527. (24) Davis, S. M.; Zaera, F.; Somorjai, G. A. The Reactivity and Composition of Strongly Adsorbed Carbonaceous Deposits on Platinum. Model of the Working Hydrocarbon Conversion Catalyst. J. Catal. 1982, 77, 439. (25) Wilson, J.; Guo, H.; Morales, R.; Podgornov, E.; Lee, I.; Zaera, F. Kinetic Measurements of Hydrocarbon Conversion Reactions on Model Metal Surfaces. Phys. Chem. Chem. Phys. 2007, 9, 3830−3852. (26) Tillekaratne, A.; Simonovis, J. P.; Zaera, F. Ethylene Hydrogenation Catalysis on Pt(111) Single-Crystal Surfaces Studied by Using Mass Spectrometry and in Situ Infrared Absorption Spectroscopy. Surf. Sci. 2015, DOI: 10.1016/j.susc.2015.11.005. (27) Rekoske, J. E.; Cortright, R. D.; Goddard, S. A.; Sharma, S. B.; Dumesic, J. A. Microkinetic Analysis of Diverse Experimental Data for Ethylene Hydrogenation on Platinum. J. Phys. Chem. 1992, 96, 1880− 1888. (28) Harris, J.; Kasemo, B. On Precursor Mechanisms for Surface Reactions. Surf. Sci. 1981, 105, L281−L287. (29) Rettner, C. T.; Auerbach, D. J.; Tully, J. C.; Kleyn, A. W. Chemical Dynamics at the Gas−Surface Interface. J. Phys. Chem. 1996, 100, 13021−13033. (30) Li, J.; Fleurat-Lessard, P.; Zaera, F.; Delbecq, F. Mechanistic Investigation of the Cis/Trans Isomerization of 2-Butene on Pt(111): DFT Study of the Influence of the Hydrogen Coverage. J. Catal. 2014, 311, 190−198. (31) Simonovis, J.; Zaera, F. Abrupt Increase in Hydrogen Diffusion on Transition-Metal Surfaces During Hydrogenation Catalysis. Chem. Sci. 2016, DOI: 10.1039/C6SC01249C. (32) Cremer, P. S.; Su, X.; Shen, Y. R.; Somorjai, G. A. Ethylene Hydrogenation on Pt(111) Monitored in Situ at High Pressures Using Sum Frequency Generation. J. Am. Chem. Soc. 1996, 118, 2942−2949. (33) Ohtani, T.; Kubota, J.; Kondo, J. N.; Hirose, C.; Domen, K. InSitu Observation of Hydrogenation of Ethylene on a Pt(111) Surface under Atmospheric Pressure by Infrared Reflection Absorption Spectroscopy. J. Phys. Chem. B 1999, 103, 4562−4565.

M.E.: Université du Québec, INRS-ÉMT, 1650 Boul. Lionel Boulet, Varennes, QC Canada J3X 1S2. # A.T.: Department of Chemistry, Faculty of Science, University of Colombo, Colombo 03, Sri Lanka. ‡

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Funding for this project was provided by a grant from the U.S. National Science Foundation (CHE-1359668).

■

REFERENCES

(1) Beeck, O. Catalysis − A Challenge to the Physicist (as Exemplified by the Hydrogenation of Ethylene over Metal Catalysts). Rev. Mod. Phys. 1945, 17, 61−71. (2) Satterfield, C. N. Heterogeneous Catalysis in the Industrial Practice, 2nd sub ed.; Krieger Pub Co: Malabar, FL, 1996. (3) Armor, J. N. A History of Industrial Catalysis. Catal. Today 2011, 163, 3−9. (4) Bond, G. C. Metal-Catalysed Reactions of Hydrocarbons; Springer: New York, 2005. (5) Zaera, F. Key Unanswered Questions About the Mechanism of Olefin Hydrogenation Catalysis by Transition-Metal Surfaces: A Surface-Science Perspective. Phys. Chem. Chem. Phys. 2013, 15, 11988−12003. (6) Horiuti, I.; Polanyi, M. Exchange Reactions of Hydrogen on Metallic Catalysts. Trans. Faraday Soc. 1934, 30, 1164−1172. (7) Cremer, P. S.; Su, X.; Somorjai, G. A.; Shen, Y. R. High Pressure Catalytic Processes Studied by Infrared-Visible Sum Frequency Generation. J. Mol. Catal. A: Chem. 1998, 131, 225−241. (8) Chrysostomou, D.; French, C.; Zaera, F. The Thermal Chemistry of Propyl Groups on Pt(111). Catal. Lett. 2000, 69, 117−128. (9) Zaera, F. Direct Observation of β-Hydride Elimination Reactions on Metal Surfaces. J. Am. Chem. Soc. 1989, 111, 8744−8745. (10) Ye, P.; Gellman, A. J. Transition State for β-Hydride Elimination in Alkyl Groups on Pt(111). J. Phys. Chem. B 2006, 110, 9660−9666. (11) Lee, I.; Hong, J.; Zaera, F. The Stereoselectivity of the Dehydrogenation of Alkyl Groups on Pt(111) Single-Crystal Surfaces. J. Phys. Chem. C 2011, 115, 982−989. (12) Twigg, G. H. The Mechanism of the Catalytic Hydrogenation of Ethylene. Discuss. Faraday Soc. 1950, 8, 152−159. (13) Bond, G. C. The Reaction of Ethylene with Deuterium over Various Types of Platinum Catalyst. Trans. Faraday Soc. 1956, 52, 1235−1244. (14) Kemball, C. 146. The Deuteration and Exchange of Ethylene on Evaporated Metal Catalysts at Low Temperatures. J. Chem. Soc. 1956, 735−743. (15) Horiuti, J.; Miyahara, K. Hydrogenation of Ethylene on Metallic Catalysts, NSRDS-NBC No. 13; National Bureau of Standards, 1968. (16) Briggs, D.; Dewing, J.; Burden, A. G.; Moyes, R. B.; Wells, P. B. Support Effects in Ethene Hydrogenation Catalyzed by Platinum. J. Catal. 1980, 65, 31−35. (17) Zaera, F.; Somorjai, G. A. Hydrogenation of Ethylene over Platinum (111) Single-Crystal Surfaces. J. Am. Chem. Soc. 1984, 106, 2288−2293. (18) Farkas, A.; Farkas, L. The Catalytic Interaction of Ethylene and Heavy Hydrogen on Platinum. J. Am. Chem. Soc. 1938, 60, 22−28. (19) Goddard, S. A.; Cortright, R. D.; Dumesic, J. A. Deuterium Tracing Studies and Microkinetic Analysis of Ethylene Hydrogenation over Platinum. J. Catal. 1992, 137, 186−198. (20) Farkas, A.; Farkas, L.; Rideal, E. K. Experiments on Heavy Hydrogen. IV. Hydrogenation and Exchange Reaction of Ethylene with Heavy Hydrogen. Proc. R. Soc. London, Ser. A 1934, 146, 630− 639. (21) Tilekaratne, A.; Simonovis, J. P.; López Fagúndez, M. F.; Ebrahimi, M.; Zaera, F. Operando Studies of the Catalytic Hydro2443

DOI: 10.1021/acs.jpclett.6b01103 J. Phys. Chem. Lett. 2016, 7, 2439−2443