ARTICLE pubs.acs.org/Langmuir
Adsorption of Oxygenates on Alkanethiol-Functionalized Pd(111) Surfaces: Mechanistic Insights into the Role of Self-Assembled Monolayers on Catalysis Stephen T. Marshall, Daniel K. Schwartz, and J. William Medlin* Department of Chemical and Biological Engineering, University of Colorado at Boulder, UCB 424, Boulder, Colorado 80309, United States
bS Supporting Information ABSTRACT: Recent work shows that coating a supported palladium catalyst with a self-assembled monolayer (SAM) of alkanethiols can dramatically improve selectivity in the hydrogenation of 1-epoxy-3-butene (EpB) to 1-epoxybutane. Here, we present the results of surface-level investigations of the adsorption of EpB and related molecules on SAM-coated Pd(111), with an aim of identifying mechanistic explanations for the observed catalytic behavior. Alkanethiol SAM-covered Pd(111) surfaces were prepared by conventional techniques and transferred to ultrahigh vacuum, where they were characterized using Auger electron spectroscopy (AES) and temperature-programmed desorption (TPD) of EpB and other probe molecules. Whereas previous studies have shown that EpB undergoes rapid decomposition via epoxide ring opening on uncoated Pd(111), TPD studies show that EpB does not undergo substantial ring opening on SAM-covered surfaces but rather desorbs intact at temperatures less than 300 K. Systematic comparisons of EpB desorption spectra to spectra for other C4 oxygenates suggest that the SAM creates a kinetic barrier to epoxide ring-opening reactions that does not exist on the uncoated surface. The EpB desorption spectra as a function of exposure show behavior similar to the desorption of olefins from Pd(111), indicating that the binding of the olefin functionality, in contrast to that of the epoxide ring, is not significantly perturbed. EpB desorption spectra from surfaces with less well-ordered SAMs show the presence of weakly bound states not observed on well-ordered SAM surfaces. The lower activity observed on catalysts covered with less well-ordered SAMs is hypothesized to occur due to partial confinement of adsorbates into these weakly bound, less active states.
’ INTRODUCTION Homogeneous catalysts are notable for the high selectivity that can be conferred by organic ligands, a feature largely lacking in their heterogeneous counterparts. It is believed that these organic ligands enable precise control over adsorbate interactions. For example, binuclear palladium transition metal complexes with (tert-butyl)2 phosphine ligands can hydrogenate a variety of unsaturated epoxides to saturated epoxides with selectivities as high as 95%.1 However, catalysts similar to this transition metal complex are employed only in a small number of industrial chemical processes due to unfavorable economics resulting from high separation costs or an inability to recover used catalyst.2 Heterogeneous catalysts that incorporate the beneficial properties of their homogeneous counterparts would prove valuable. The chemoselective conversion of unsaturated oxygenates is a daunting challenge on heterogeneous catalysts. Recently, we reported the ability of palladium catalysts coated with self-assembled monolayers (SAMs) of alkanethiols to selectively convert 1-epoxy3-butene (EpB) to 1-epoxybutane with selectivities up to 94%, r 2011 American Chemical Society
whereas uncoated catalysts exhibited a selectivity of 11% under the same conditions.3 Achieving high selectivity to this reaction is difficult due to the inherent strain in the epoxide ring which results in facile ring opening on platinum group metals.4,5 SAMs with different alkyl tail lengths and added functionalities were tested, and three interesting results were observed. First, selectivity for epoxybutane was high regardless of alkyl tail length. Second, activity increased with increasing alkyl tail length from 3 to 18 carbons and with increasing organizational order within the tail phase of the SAM. Last, SAMs formed from the three carbon thioglycerol (1-mercapto-2,3-propanediol), unlike SAMs formed from propanethiol, showed high activity for this reaction. Studies of EpB adsorption on model surfaces have proved invaluable for the design of effective catalysts. Recently, Monnier and colleagues reported 55% selectivity for epoxybutane using a platinumsilver bimetallic catalyst under mild conditions.6 Received: December 10, 2010 Revised: March 19, 2011 Published: April 28, 2011 6731
dx.doi.org/10.1021/la104921p | Langmuir 2011, 27, 6731–6737
Langmuir Studies on Pt and Ag single crystals suggest that this bimetallic catalyst achieves high selectivity through a bifunctional effect by forming a reversibly adsorbed oxametallacycle intermediate incorporating silver atoms, which protects the epoxide ring from further reaction.79 Surface-level investigations of the interaction of epoxides and other reagents with the catalytic metals have aided the design of these systems and predictions for future areas of research.911 Here, we glean similar mechanistic information for the interaction of EpB with SAM-coated surfaces using fundamental surface science techniques in order to explain the three trends described above. Applying a thiol-based SAM is an unusual approach for improving the behavior of catalysts since sulfur is widely regarded as a potent catalyst poison.12,13 However, SAMs confer a number of desirable traits for catalyst modification. SAMs confine thiol adsorbates into a highly regular On many fcc(111) √ structure. √ surfaces, thiols adsorb in a ( 3 3)R30° unit cell with the sulfur “head group” attached to the surface and the hydrocarbon “tail” oriented at a well-defined angle with respect to the surface normal.14,15 Unlike other forms of sulfiding which can result in highly poisonous subsurface sulfur,16 sulfur adsorbed on the metal surface from thiols does not measurably penetrate into the bulk.17 In addition, preparation methods for alkanethiol SAMs are facile, self-limiting, and scalable.17,18 Last, although the current work focuses only on thiol-based SAMs, SAMs may also be comprised of different head groups containing other species including phosphorus,19 selenium,20 and silicon.21 Since our previous investigations show this head group is largely responsible for the observed enhancement in catalysts,3 SAMs have a wide range of potential for improving catalytic systems. An apparent potential problem with using SAM coatings is the seeming lack of available active sites on the surface for adsorption and conversion of reactants. Although SAMs do cover a large fraction of surface sites, the concept of space with regard to binding of olefins is often deceptive. It is well known that ethylene can chemisorb onto surfaces precovered with more than a monolayer of carbonaceous deposits.22 Such deposits are ubiquitous in catalysts used for hydrogenation of olefins,23 so a SAM-coated surface may, in fact, be less covered than a comparable “clean” surface under catalytic conditions. For example, coverage by alkylidynes and other carbonaceous species can exceed one adsorbate per Pt surface atom during ethylene hydrogenation on Pt(110),24 whereas the maximum coverage from alkanethiol SAMs is 1/3 ML on transition metal (111) surfaces.15 The adsorption of molecules on SAM-coated surfaces has been studied in detail.2527 However, these studies generally focus on species such as biomolecules that adsorb onto the tail moieties of SAMs, and interactions of compounds with the underlying metal surface have not been as thoroughly examined. Such investigations may yield insights into fundamental interaction mechanisms between small organic species and SAM-coated surfaces. For example, the modification of gold surfaces with monolayers of chiral thiols including cysteine and penicillamine has been shown to impart enantioselectivity for the adsorption and electrochemical reaction of 3,4-dihydroxyphenylalanine28,29 and the adsorption and separation of propylene oxide enantiomers.30 Fundamental studies of these systems show that thiol modifiers may impart chirality to the inherently achiral underlying metal surface31 and that modification may result in direct interactions between chiral sites on thiols and adsorbates.29 The structure and impact of thiol modification to Pt nanoparticles has also been
ARTICLE
studied for the adsorption and reaction of allyl alcohol.32 These examples illustrate the growing potential and interest in SAMmodified systems for controlling behavior at the site level and the diverse range of fields in which such control is desirable.
’ MATERIALS AND METHODS Experiments were conducted in a stainless steel ultrahigh vacuum (UHV) system described in detail elsewhere.8 The chamber had a base pressure of ∼1010 Torr. The same Pd(111) crystal (Princeton Scientific) was used for all experiments. The crystal was spot welded to a 1.5 mm Ta disk that was held in the chamber on a copper stage using two clips. This sample was cooled using liquid nitrogen in thermal contact with the sample through the copper stage and heated using a resistive heater. The crystal was cleaned using cycles of heating in 5 108 Torr O2 (research grade, Airgas) between 700 and 1000 K, sputtering with Arþ ions (UHP grade, Airgas) at ambient temperature (13 keV, NGI3000-SE sputter gun by LK technologies), and annealing between 700 and 1000 K. Cleanliness was verified using oxygen TPD and Auger electron spectroscopy. The Pd(111) sample was removed from the chamber for SAM deposition and characterization using a wobble stick and a separate load-lock chamber. SAMs were deposited and characterized using techniques described in detail elsewhere.17,33 In short, the Pd(111) sample was removed from vacuum through a load-lock system and immediately inserted into a 1 mM solution of propanethiol, hexanethiol, dodecanethiol, octadecanethiol, or thioglycerol in ethanol (Sigma Aldrich). All thiol purities were >97%, and ethanol was 200 proof HPLC grade. SAMs were deposited over a period greater than 12 h, after which the sample was removed from solution and rinsed with ethanol. The quality of the SAM was characterized by measuring the advancing water contact angle using a custom-built goniometer. The SAM-coated crystal was then reinserted into the UHV chamber through the attached load-lock chamber separated by a gate valve. SAMs on Pd are known to degrade in air, so the sample was exposed to room air for a maximum of 30 min, far less than the 25 days required for substantial oxidation.17 The load-lock chamber was pumped by a turbomolecular pump (Pfeiffer Vacuum) to a pressure less than 106 Torr. The sample was moved into the main chamber and inserted into the sample stage with a magnetic sample transfer arm and a wobble stick. This procedure eliminated the need for baking the chamber after inserting the sample, which would result in decomposition of the SAM coating. 1-Epoxy-3-butene (Alfa Aesar, >98%), 1-epoxybutane (Sigma Aldrich, >99%), crotonaldehyde (2-butenal, Sigma Aldrich, >99.5%), and butyraldehyde (butanal, Sigma Aldrich, >99%) were dosed directly to the sample by sending a pulse of vapor through a 1/16 in. OD tube approximately 2 in. from the sample. Although this method greatly reduces background adsorption and aids in maintaining low system pressure, exposures are not reportable in Langmuirs. The exposure was controlled by varying the pressure of the source for the vapor pulse. For the range of exposures (small to large) reported below for SAM-coated Pd(111), the source pressures were 0.01, 0.03, 0.05, 0.07, and 0.1 Torr. When spectra for only a single coverage are reported in the figures below, it is for a source pressure of 0.1 Torr. Source EpB pressures of 0.5 Torr on SAM-coated surfaces were observed to result in significant multilayer desorption features (see Supporting Information). The same 0.5 Torr exposure of EpB on uncoated Pd(111) followed by TPD results in complete EpB decomposition to yield approximately 0.3 ML CO. Assuming the sticking probability on the SAM-coated surface is not higher than that for the uncoated surface, this coverage places an upper bound on the saturation coverage in the presence of the SAM. However, the true saturation coverage in the presence of the SAM may be significantly lower, particularly since multilayer desorption features are already evident for source pressures of 0.3 Torr. Hexanethiol was 6732
dx.doi.org/10.1021/la104921p |Langmuir 2011, 27, 6731–6737
Langmuir
ARTICLE
exposed to the crystal using a source pressure of approximately 0.1 Torr, with repeated doses used to increase the exposure. Temperature-programmed desorption (TPD) experiments were conducted using a Smart-IQþ quadrupole mass spectrometer (VG Scienta) equipped with a stainless steel shroud. Auger electron spectroscopy (AES) experiments were conducted using a RVL2000 LEED/ AES apparatus including an electron gun (LK Technologies) and a model CMA 2000 (LK Technologies) mounted at a right angle to the electron gun.
’ RESULTS AND DISCUSSION Studying the chemistry of a surface covered with organic ligands is more complicated than studying a clean surface, as SAMs can decompose at elevated temperatures. Supporting Information Figure S1 shows a TPD spectrum for the m/z = 27, 29, and 34 traces (corresponding to ionized hydrocarbon fragments for m/z = 27 and 29 and to H2S for m/z = 34) for a hexanethiol SAM-coated surface. Other masses corresponding to hydrocarbon fragments were observed in the desorption spectra for this SAM; however, they typically mirrored these traces or showed desorption peaks at higher temperatures. While detailed analysis of the decomposition products is not of interest for this work, the data in Figure S1, Supporting Information, show that SAM decomposition occurs on the Pd(111) surface at temperatures above ∼325350 K. In the experiments described below, the temperature of the sample did not exceed 323 K to prevent SAM decomposition. In addition, experiments were repeated for each SAM preparation to ensure there were no changes to the behavior of the SAM coating from thermal decomposition or exposure to adsorbates. The behavior observed for SAMs on Pd(111) is similar to that previously observed for SAMs on polycrystalline Au surfaces.34 Alkanethiol SAMs on Au were observed to change slightly in structure and order with heating up to 350 K. These changes were reversible, and cooling resulted in a structure identical to the original. However, when the SAM-coated surface was heated above 350 K, defects formed in the structure of the SAM that did not disappear on cooling.34 The onset of decomposition appears slightly lower for Pd(111). This effect may be due to the higher activity of Pd for bond-breaking reactions. Prior to TPD experiments, SAMs were characterized by measuring an advancing water contact angle under ambient conditions and using AES in vacuum. Because methyl moieties are more hydrophobic than methylene moieties, increasing order (i.e., crystallinity) in alkane chains presents a more hydrophobic surface resulting in a larger contact angle.17,33 Conversely, for hydroxylated thiols like thioglycerol, increasing order presents a more hydrophilic surface resulting in a smaller contact angle. The contact angles observed for the fully formed propanethiol, hexanethiol, dodecanethiol, and octadecanethiol SAMs were 95°, 115°, 114°, and 117°, respectively. On the basis of previously reported values,17,35 these results indicate that hexanethiol, dodecanethiol, and octadecanethiol form well-ordered SAMs on the Pd surface, whereas propanethiol formed a less wellordered SAM. The advancing contact angle for the thioglycerol SAM was 14°, indicating that a significant amount of thioglycerol coated the surface. However, determining the relative order of the thioglycerol SAM relative to the alkanethiols is not possible with this method. The quality of the SAM coatings was also evaluated using AES. In addition to providing surface composition information, AES is sensitive in some cases to changes in electronic structure.
Figure 1. AES spectra of SAM-coated Pd(111) surfaces. The negative lobes of the S, C, and Pd features are noted by vertical lines (C3, propanethiol; C6, hexanethiol; C12, dodecanethiol; C18, octadecanethiol; TG, thioglycerol).
Table 1. Peak to Peak Intensities and Ratios of Intensities for AES Spectra Shown in Figure 1a peak to peak intensity (au) coating
S
Pd
C
S:Pd ratio
C:Pd ratio
C3
13.7
10.6
3.9
1.3
0.4
C6
14.0
9.4
5.4
1.5
0.6
C12 C18
10.3 4.9
14.1 8.9
9.0 13.8
0.7 0.6
0.6 1.6
TG
17.1
6.8
4.3
2.5
0.6
C6 exposure
10.4
14.4
6.2
0.7
0.4
Error in values of peak to peak intensity and ratios is (0.5 and (0.7, respectively, based on repeated experiments.
a
Significant changes in the local electronic structure of an atom, such as a change in oxidation state, can lead to a shift in the AES feature for that species.36 AES spectra for the coatings employed are shown in Figure 1. These AES spectra provide some details regarding the relative state and order of these surfaces. First, the negative lobe of the palladium feature is present at 330 eV for every coating, which is consistent with the clean Pd surface.37 This result indicates that Pd retains its typical electronic character and is not oxidized like the metal atoms in some organometallic homogeneous catalysts.2 This result does not necessarily indicate that the electronic character of the underlying Pd is the same as for the clean surface, since subtle changes cannot be resolved with AES, but does imply that the surface has not undergone a significant change such as a change in oxidation state. Further investigation of changes in oxidation state using X-ray photoelectron spectroscopy will help in obtaining a detailed picture of more subtle effects on the surface electronic state. The position of the S and C AES features for each of the SAM coatings appears similar except for the S feature for the thioglycerol SAM. While the S features for the alkanethiol SAMs correspond to the typical position of unoxidized sulfur,37 in the thioglycerol spectrum this feature is shifted to lower energy by 2.5 eV. This shift is not as large as the ∼10 eV shift observed in sulfates37 or the oxidation of species such as silicon;36,39 however, it does indicate some change in the electronic state of adsorbed sulfur. The size of the S feature compared to the Pd feature, as measured by the peak-to-peak amplitude (Table 1), is also much larger for thioglycerol than for the alkanethiols, indicating a larger fraction of the surface is covered by sulfur. Although S and Pd 6733
dx.doi.org/10.1021/la104921p |Langmuir 2011, 27, 6731–6737
Langmuir
Figure 2. TPD spectra of EpB and similar molecules on hexanethiol SAMcoated Pd(111). Epoxybutane and butyraldehyde spectra magnified 5 due to low yield: (a) EpB (m/z = 39); (b) crotonaldehyde (m/z = 41); (c) epoxybutane (m/z = 41) 5; (d) butyraldehyde (m/z = 27) 5. See Materials and Methods section for an explanation of the exposures.
Scheme 1. Reaction Pathways of EpB Observed on Clean and SAM-Coated Pd(111)a
a
Pathways for the uncoated surface are taken from ref 40.
features are smaller in the longer alkanethiol coatings presumably due to the shielding effect of thiol tails, comparing the S:Pd ratio of thioglycerol and propanethiol (two thiols of the same length) shows significant surface sulfur enrichment for the former. The larger amount of surface sulfur on the thioglycerol SAM-coated surface may result from an increased packing density in thioglycerol SAMs or from cracking of SC bonds to form surface sulfur, which has been observed on Pd using X-ray photoelectron spectroscopy (XPS).17,18 These results indicate adding functional groups to the SAM tail may result in changes to the surface structure that may affect catalyst behavior compared to similar size tails. Before describing TPD investigations of EpB on SAM-coated Pd(111), it is useful to briefly review previously reported investigations of EpB thermal chemistry on clean Pd(111). Previous studies of EpB adsorption on uncoated Pd(111) surfaces reflect the facile ring opening observed on palladium catalysts. At low temperatures, EpB adsorbs through its olefin functionality with its epoxide ring intact. This adsorption structure was confirmed by high-resolution electron energy loss (HREELS) experiments which show a lack of a CdC stretching mode and the presence of an epoxide ring deformation mode. As the surface is heated to 190 K, the EpB ring opens to form an aldehyde-like intermediate, indicated by a strong CdO stretching feature in HREEL spectra. Above 250 K, this species decarbonylates to form CO and propylene, as observed in
ARTICLE
temperature-programmed desorption (TPD) experiments.40 Significant hydrogen desorption resulting from EpB decomposition is also detected in TPD beginning around 280 K. A representation of this process is shown in Scheme 1. Importantly, no desorption of intact EpB from a chemisorbed state is observed on uncoated Pd(111), because EpB undergoes complete decomposition on the surface by ∼200 K. To determine the nature of EpB binding on the SAM-coated surface, experiments were performed using a variety of molecules that contained similar functional groups. Crotonaldehyde has a structure similar to that of EpB, except that the oxygenate functionality is an aldehyde instead of an epoxide. Epoxybutane is also similar to EpB but lacks the olefin functionality. Finally, butyraldehyde contains an aldehyde functionality and lacks the olefin functionality. TPD spectra for equivalent exposures of EpB and similar oxygenates on a hexanethiol SAM-coated surface are shown in Figure 2. Unlike on clean Pd(111) where the EpB ring opens to form an aldehyde-like intermediate around 190 K and decomposes to form hydrogen below 325 K,40 on SAM-coated surfaces EpB desorbed intact and no H2 desorption was observed until the onset of SAM decomposition above 350 K. EpB behaved similarly to crotonaldehyde, and both molecules desorbed at significantly higher temperatures than their saturated counterparts. This result suggests that both EpB and crotonaldehyde bind through their common olefin functionality. Epoxybutane and butyraldehyde desorbed at lower temperatures than EpB or crotonaldehyde, with no evidence of volatile decomposition products. The lower desorption temperatures indicate that interactions between monofunctional oxygenates and the surface were weaker. Similar to its unsaturated counterpart, epoxybutane did not ring open on the SAM-coated surface. In addition, butyraldehyde desorbed at around 242 K, indicating weaker adsorption than has been previously observed for propionaldehyde on clean Pd(111), which desorbs in multiple peaks up to nearly 350 K.41 More noticeably, aldehydes such as propionaldehyde undergo extensive decomposition on Pd(111), producing H2 and volatile hydrocarbons beginning at temperatures below 300 K during TPD. Such decomposition products are not detected from butyraldehyde on the SAM-coated surfaces. The peak intensities for butyraldehyde and epoxybutane were also much smaller from the same exposure, which could be due to either a reduced density of binding sites available for these molecules or a reduced sticking probability; future experiments will explore this issue. These results indicate the presence of the SAM may weaken the adsorption energy of oxygenates with the surface in addition to poisoning of epoxide ring opening. Additional experiments were performed on the other SAM coatings used in this work, and the same trends were observed. As a whole, these results indicate that the presence of the SAM coating inhibits ringopening reactions of the epoxide functionality and that EpB binds through its olefin functionality. Since the behavior of olefin hydrogenation reactions on metal surfaces can vary with coverage,2442 it is useful to examine the effect of increasing exposure on EpB desorption. Experiments were performed using a range of exposures varying from those that produced very little adsorbed EpB to those that produced multilayer adsorption (physisorption) of EpB. Spectra for submultilayer exposures of EpB on each SAM-coated surface are shown in Figure 3. Note that significant signals for EpB are recorded even at the highest temperatures of the TPD, so that 6734
dx.doi.org/10.1021/la104921p |Langmuir 2011, 27, 6731–6737
Langmuir
ARTICLE
Figure 3. TPD spectra of increasing exposures of EpB (tracked using m/z = 39) on the SAM-coated surfaces. All exposures are submultilayer, indicated by the lack of a physisorption peak ∼150 K (C3, propanethiol; C6, hexanethiol; C12, dodecanethiol; C18, octadecanethiol; TG, thioglycerol). See Materials and Methods section for a description of the exposures.
some EpB in relatively tightly bound states may have remained in subsequent experiments. This may also be due to a slower pumping speed (higher base pressure) of the chamber at the time experiments were conducted; note that the base pressure was lower during the experiments reported in Figure 2, perhaps accounting for the reduced desorption intensity at high temperature. Furthermore, although TPD experiments could be conducted repeatedly on the same SAM with no major change in results, run-to-run variations in the higher temperature desorption intensity (possibly due to small differences in chamber pressure) were also observed. However, the peak temperature and intensity were observed to be highly reproducible (see Figure S2, Supporting Information). These spectra reveal interesting details of EpB chemistry on the SAM-coated surfaces. First, on all surfaces EpB desorption temperature decreases with increasing coverage. A similar coverage-dependent desorption temperature has been observed for olefins desorbing from clean Pd and Pt surfaces and has been attributed to surface crowding that may result in destabilizing interactions between adsorbates, changes to underlying metal electronic structure (through-surface effects), or changes in available ensemble size.22,43,44 For the adsorption of olefins such as ethylene and propylene on clean Pd(111), the maximum desorption temperature for the intact species occurs above 300 K.43,45 The maximum desorption temperature of EpB on the alkanethiol-coated surfaces was approximately 270 K, suggesting a weaker adsorption state on the thiol-covered surface compared to low coverages of olefins on the clean surface. Although the complete decomposition of EpB on uncoated Pd(111) prevented experimental determination of its adsorption strength on the clean (uncoated) surface relative to ethylene, previous theoretical investigations have shown that the adsorption configuration and energy of EpB are similar to those of ethylene on Pt(111).7 Interestingly, other species on the Pd(111) surface can induce a decrease in desorption temperature similar to effects of increasing coverage. On Pd(111) covered with adsorbates such as H or O, the desorption temperature of olefins such as ethylene or propylene shifts from over 300 K to a desorption temperature below 250 K with increasing H or O coverage.43,45 Vibrational spectroscopy studies show that this shift is consistent
with a change in binding state from a di-σ-bound species to the π-bound species that is more reactive toward hydrogenation.46 The maximum desorption temperature of EpB is remarkably similar to the desorption temperature of small exposures of ethylene from surfaces precovered with an amount of hydrogen similar to the full SAM coverage of 0.33 monolayers.43 Although one might expect sulfur to have a significantly different impact on surface adsorption than species such as hydrogen and oxygen, similar destabilization of CdC adsorption has been reported for benzene adsorption on sulfur-coated Pt(111)47 and attributed to steric effects.48 As a whole, these results suggest the mechanism of binding for the olefin functionality of EpB is comparable to the binding of species such as ethylene or propylene on surfaces covered with ubiquitous species such as hydrogen, oxygen, and carbon. The effect of steric repulsions between the alkyl tails and EpB adsorbates on EpB adsorption appears to not be large, but further study is needed to determine possible effects of alkyl tails on EpB sticking probability. The shape of the peaks in the TPD spectra suggest differences between the thioglycerol SAM and the alkanethiol SAMs. The thioglycerol spectra show nearly symmetric peaks; however, the alkanethiol SAM spectra show asymmetric peaks with significant intensity at temperatures higher than the peak temperature. It is possible that such effects are related to the stronger H-bonding interactions between tails, which can increase the temperature at which SAMs decompose in TPD experiments by nearly 100 K49 and are associated with a higher degree of crystallinity, but more study is needed to investigate this hypothesis.50 One prominent feature in the spectra is the difference between the longer, well-ordered alkanethiol SAMs and the less wellordered propanethiol SAM. An additional desorption peak is observed for the propanethiol case around 200 K. This peak temperature is less than that associated with π-bound olefins43,45 yet greater than that for physisorbed EpB observed in multilayer spectra (Figure S3, Supporting Information). This result indicates that a fraction of adsorbates on less well-ordered SAMs is incapable of accessing the more tightly bound states universally occupied on well-ordered SAMs. These species may be too weakly adsorbed to have a high probability of reaction, so less well-ordered SAMs may reduce activity by a reduction in the number of active sites. 6735
dx.doi.org/10.1021/la104921p |Langmuir 2011, 27, 6731–6737
Langmuir
Figure 4. TPD spectrum of the m/z = 39 trace for EpB adsorbed on to a Pd(111) surface precovered with disordered hexanethiol. See Materials and Methods section for a description of the exposures.
In previous studies of alkanethiol SAM-coated catalysts, both the length of the tail and the degree of organizational order within the tail phase were correlated with the catalyst activity; however, which of these two traits caused the observed trend in activity was not clear.3 In order to determine the role of tail length and organizational order, we separated these effects by adsorbing hexanethiol on Pd(111) from the vapor phase in the ultrahigh vacuum system. When exposed in quantities between 10 and 10 000 L, thiols such as hexanethiol form intermediate structures, such as a striped phase and liquid-like phases, that do not have the same organizational order as the complete SAM.15 However, phases like the striped phase tend to adopt a similar unit cell geometry with respect to the position of sulfur atoms on the (111) surface along with a number of void spaces comprised of sites covered by flat lying thiol tails or defects,18,51 so these phases can approximate less well-ordered SAMs with thiols longer than propanethiol. An AES spectrum taken of the hexanethiol-dosed surface before EpB adsorption showed a slightly smaller ratio in C to Pd peak to peak intensity than in the hexanethiol SAM spectrum (Table 1). This result may occur from a smaller thiol coverage on the hexanethiol-exposed surface or from a reduced shielding effect from tails lying parallel to the surface in phases such as the striped phase.18 Figure 4 shows a m/z = 39 TPD spectrum for EpB adsorbed on a Pd(111) surface covered with a moderate exposure of hexanethiol. Changing the thiol exposure over an order of magnitude produced qualitatively similar results to the data shown in Figure 4, likely due to the wide range of exposures that can produce intermediate phases. Multiple desorption states between 200 and 250 K were observed in the spectrum in Figure 4, similar to the desorption of EpB from propanethiol SAM-coated Pd(111). This result suggests that the degree of organizational order, rather than the length of alkyl tails, correlates with the activity of the surface. The limitation of this model lies in the presence of the void spaces that are not present in the complete SAM leading to lower thiol coverage. However, the impact of such voids appears to be minimal, since pathways observed on the clean surface, including decarbonylation to produce volatile CO and propylene, were not observed on the hexanethiol-exposed surface when TPD experiments were run to temperatures above the normal CO desorption temperature of 485 K. Figure 4 also shows an increase in intensity in the m/z = 39 trace at temperatures above room temperature resulting from the decomposition of hexanethiol, which shares this fragment EpB. However, other masses tracked during TPD show that the two peaks between 200 and 250 K can be uniquely identified as intact EpB.
ARTICLE
’ CONCLUSIONS SAM coating offers considerable potential for tailoring the behavior of supported metal catalysts. Previous results of studies regarding SAM-coated catalysts posited two hypotheses for the observed EpB activity and selectivity of these catalysts.3 They were as follows. • High selectivity results from the presence of surface sulfur. Thiol tails do not interact with adsorbates in a manner that confers selectivity. • Activity decreases on shorter and less well-ordered SAMs because disordered tails either physically block surface sites or create entropic barriers to adsorption onto active sites. The results presented here provide additional evidence for the refinement of these hypotheses. First, TPD spectra for saturated and unsaturated analogues of EpB indicate poisoning of ringopening reactions leads to high selectivity for epoxybutane on SAM-coated catalysts. Unlike clean Pd(111), where ring-opening reactions occur at temperatures below 200 K,39 both EpB and its saturated counterpart, epoxybutane, do not ring open on SAM-coated surfaces. The nearly identical desorption peak temperature for the same exposure of EpB and crotonaldehyde and the higher desorption temperature of these molecules compared to their saturated counterparts strongly suggests that surface sulfur causes preferential binding through the olefin functionality. Similarities between desorption spectra with increasing exposure for ethylene on clean and hydrogen-covered Pd(111) and the EpB desorption spectrum from well-ordered SAM-coated surfaces suggest that unlike the oxygenate functionality the binding of EpB’s olefin functionality is similar to that on the clean surface at high coverage. Second, results of EpB desorption spectra as a function of exposure support the hypothesis for catalyst activity. The observation of weaker adsorption states on the propanethiol SAMcoated surface indicates that reduced activity results from a fraction of adsorbates being unable to bind in active sites. Since the techniques used in this article are temperature programmed, they probe only enthalpic contributions to binding, suggesting that less well-ordered coatings must produce additional enthalpic barriers to binding. Such effects could result from either electronic or geometric changes to the surface structure. In either case, these results suggest that the exact reason for changes in catalyst activity with order may be a complex combination of factors. ’ ASSOCIATED CONTENT
bS
Supporting Information. TPD spectra for the decomposition of a hexanethiol SAM, for repeated EpB exposures of the same size, and for multilayer adsorption of EpB on a C18 SAMcoated surface. This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected].
’ ACKNOWLEDGMENT We acknowledge the US Department of Energy, Office of Basic Energy Sciences (Award number DE-SC0005239) and the Renewable and Sustainable Energy Institute (RASEI) for funding this work. S.T.M. acknowledges support from the Peter and Vivian Teets Nanotechnology fellowship and the Department of 6736
dx.doi.org/10.1021/la104921p |Langmuir 2011, 27, 6731–6737
Langmuir Education Graduate Assistantships in Areas of National Need (GAANN).
’ REFERENCES (1) Cho, I. S.; Lee, B.; Alper, H. Tetrahedron Lett. 1995, 36, 6009– 6012. (2) Leeuwen, P. W. N. M. v. Homogeneous catalysis: understanding the art; Kluwer Academic Publishers: Boston, 2004. (3) Marshall, S. T.; O’Brien, M.; Oetter, B.; Corpuz, A.; Richards, R. M.; Schwartz, D. K.; Medlin, J. W. Nat. Mater. 2010, 9, 853–858. (4) Bartok, M.; Fasi, A.; Notheisz, F. J. Catal. 1998, 175, 40–47. (5) Bartok, M.; Fasi, A.; Notheisz, F. J. Mol. Catal. A: Chem. 1998, 135, 307–316. (6) Schaal, M. T.; Pickerel, A. C.; Williams, C. T.; Monnier, J. R. J. Catal. 2008, 254, 131–143. (7) Schaal, M. T.; Hyman, M. P.; Rangan, M.; Ma, S.; Williams, C. T.; Monnier, J. R.; Medlin, J. W. Surf. Sci. 2009, 603, 690–696. (8) Loh, A. S.; Davis, S. W.; Medlin, J. W. J. Am. Chem. Soc. 2008, 130, 5507–5514. (9) Medlin, J. W.; Barteau, M. A.; Vohs, J. M. J. Mol. Catal. A: Chem. 2000, 163, 129–145. (10) Linic, S.; Barteau, M. A. J. Am. Chem. Soc. 2004, 126, 8086– 8087. (11) Monnier, J. R.; Stavinoha, J. L.; Hartley, G. W. J. Catal. 2004, 226, 321–333. (12) Somorjai, G. A. J. Catal. 1972, 27, 453–456. (13) Somorjai, G. A. Introduction to surface chemistry and catalysis; John Wiley & Sons, Inc.: New York, 1994. (14) Poirier, G. E.; Pylant, E. D. Science 1996, 272, 1145–1148. (15) Kondoh, H.; Kodama, C.; Sumida, H.; Nozoye, H. J. Chem. Phys. 1999, 111, 1175–1184. (16) Miller, J. B.; Gellman, A. J. Surf. Sci. 2009, 603, L82–L85. (17) Love, J. C.; Wolfe, D. B.; Haasch, R.; Chabinyc, M. L.; Paul, K. E.; Whitesides, G. M.; Nuzzo, R. G. J. Am. Chem. Soc. 2003, 125, 2597–2609. (18) Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. M. Chem. Rev. 2005, 105, 1103–1169. (19) Gouzman, I.; Dubey, M.; Carolus, M. D.; Schwartz, J.; Bernasek, S. L. Surf. Sci. 2006, 600, 773–781. (20) Han, S. W.; Kim, K. J. Colloid Interface Sci. 2001, 240, 492–497. (21) Sugimura, H.; Hozumi, A.; Kameyama, T.; Takai, O. Surf. Interface Anal. 2002, 34, 550–554. (22) Stuck, A.; Wartnaby, C. E.; Yeo, Y. Y.; King, D. A. Phys. Rev. Lett. 1995, 74, 578–581. (23) Bond, G. C.; Wells, P. B. Adv. Catal. 1964, 15, 91–226. (24) Cremer, P. S.; Su, X. C.; Shen, Y. R.; Somorjai, G. A. J. Am. Chem. Soc. 1996, 118, 2942–2949. (25) Hodneland, C. D.; Lee, Y. S.; Min, D. H.; Mrksich, M. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5048–5052. (26) Jordan, C. E.; Frey, B. L.; Kornguth, S.; Corn, R. M. Langmuir 1994, 10, 3642–3648. (27) Senaratne, W.; Andruzzi, L.; Ober, C. K. Biomacromolecules 2005, 6, 2427–2448. (28) Kang, Y. J.; Oh, J. W.; Kim, Y. R.; Kim, J. S.; Kim, H. Chem. Commun. 2010, 46, 5665–5667. (29) Matsunaga, M.; Nakanishi, T.; Asahi, T.; Osaka, T. Chirality 2007, 19, 295–299. (30) Shukla, N.; Bartel, M. A.; Gellman, A. J. J. Am. Chem. Soc. 2010, 132, 8575–8580. (31) Jadzinsky, P. D.; Calero, G.; Ackerson, C. J.; Bushnell, D. A.; Kornberg, R. D. Science 2007, 318, 430–433. (32) Eklund, S. E.; Cliffel, D. E. Langmuir 2004, 20, 6012–6018. (33) Bain, C. D.; Whitesides, G. M. J. Am. Chem. Soc. 1988, 110, 3665–3666. (34) Bensebaa, F.; Ellis, T. H.; Badia, A.; Lennox, R. B. Langmuir 1998, 14, 2361–2367.
ARTICLE
(35) Marshall, S. T.; Schwartz, D. K.; Medlin, J. W. Sens. Actuators, B: Chem. 2009, 136, 315–319. (36) Kershner, D. C.; Medlin, J. W. Surf. Sci. 2008, 602, 786–794. (37) Davis, L. E. Handbook of Auger Electron Spectroscopy; Physical Electronics Industries: Ismaning, 1976. (38) Sung, Y. E.; Thomas, S.; Wieckowski, A. J. Phys. Chem. 1995, 99, 13513–13521. (39) Kershner, D. C.; Zhang, W.; Medlin, J. W. Surf. Sci. 2008, 602, 3225–3231. (40) Marshall, S. T.; Horiuchi, C. M.; Zhang, W. Y.; Medlin, J. W. J. Phys. Chem. C 2008, 112, 20406–20412. (41) Davis, J. L.; Barteau, M. A. J. Am. Chem. Soc. 1989, 111, 1782– 1792. (42) Thomson, S. J.; Webb, G. J. Chem. Soc., Chem. Commun. 1976, 13, 526–527. (43) Burkholder, L.; Stacchiola, D.; Tysoe, W. T. Surf. Rev. Lett. 2003, 10, 909–916. (44) Calaza, F.; Stacchiola, D.; Neurock, M.; Tysoe, W. T. J. Am. Chem. Soc. 2010, 132, 2202–2207. (45) Stacchiola, D.; Burkholder, L.; Tysoe, W. T. Surf. Sci. 2003, 542, 129–141. (46) Stacchiola, D.; Burkholder, L.; Tysoe, W. T. Surf. Sci. 2002, 511, 215–228. (47) Garfunkel, E. L.; Farias, M. H.; Somorjai, G. A. J. Am. Chem. Soc. 1985, 107, 349–353. (48) Rousseau, G. B. D.; Bovet, N.; Kadodwala, M. J. Phys. Chem. B 2006, 110, 21857–21864. (49) Valiokas, R.; Ostblom, M.; Svedhem, S.; Svensson, S. C. T.; Liedberg, B. J. Phys. Chem. B 2002, 106, 10401–10409. (50) Badia, A.; Lennox, R. B.; Reven, L. Acc. Chem. Res. 2000, 33, 475–481. (51) Sharma, M.; Komiyama, M.; Engstrom, J. R. Langmuir 2008, 24, 9937–9940.
6737
dx.doi.org/10.1021/la104921p |Langmuir 2011, 27, 6731–6737
