Surface Structure Dependence of the Dry ... - ACS Publications

May 23, 2017 - Robert A. HoytMatthew M. MontemoreE. Charles H. SykesEfthimios ... Baran Eren , Heath Kersell , Robert S. Weatherup , Christian Heine ...
1 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/JPCC

Surface Structure Dependence of the Dry Dehydrogenation of Alcohols on Cu(111) and Cu(110) Zhi-Tao Wang,†,‡ Yunfei Xu,‡ Mostafa El-Soda,† Felicia R. Lucci,† Robert J. Madix,§ Cynthia M. Friend,‡,§ and E. Charles H. Sykes*,† †

Department of Chemistry, Tufts University, Medford, Massachusetts 02155, United States Department of Chemistry and Chemical Biology and §Harvard John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge, Massachusetts 02138, United States

‡

S Supporting Information *

ABSTRACT: The non-oxidative dehydrogenation of alcohols is considered as an important method to produce aldehydes for the chemical industry and hydrogen gas. However, current industrial processes are oxidative, meaning that the aldehydes are formed along with water, which, in addition to being less energy efficient, poses separation problems. Herein the production of aldehydes from methanol and ethanol on clean and dry Cu(111) and Cu(110) surfaces was investigated in order to understand the catalytic anhydrous dehydrogenation of alcohols. Both ethanol and methanol preferentially react under ultrahigh vacuum conditions at surface defects to yield acetaldehyde and formaldehyde, respectively, in the absence of surface oxygen and water. The amount of alkoxide reaction intermediates measured by scanning tunneling microscopy, and aldehyde and hydrogen products detected by temperature programmed reaction, are increased by inducing more defects in the Cu substrates with Ar ion sputtering. This work also reveals that the Cu model surfaces are not poisoned by the reaction and exhibit 100% selectivity for alcohol dehydrogenation to aldehyde and hydrogen.

■

INTRODUCTION Aldehydes are important precursors or reactants in the production of a large number of industrial chemicals, such as plastics, resins, explosives, and drugs.1,2 The oxidative dehydrogenation of alcohols over noble-metal-based catalysts in the presence of O2 is generally used for commercial catalytic processes,2−4 which creates a significant energy cost because of the need to separate the water byproduct.5 Therefore, there is considerable interest in the development of catalytic nonoxidative alcohol dehydrogenation pathways in which, instead of water, valuable H2 is produced. There are a few studies of bulk and supported transition metal oxides as catalysts for the conversion of alcohols to aldehydes and hydrogen.6−9 These catalysts, however, are slowly reduced to metallic nanoparticles by oxygen consumption during the alcohol dehydrogenation reactions and become inactive for non-oxidative dehydrogenation.4,6−9 Therefore, it is timely to investigate oxygen-free systems that are selective to aldehydes and do not deactivate during the alcohol dehydrogenation reaction. Cu-based catalysts are of interest since they have been shown to have high selectivity for ethanol dehydrogenation to acetaldehyde and H2, although frequent regeneration is required.2 In terms of gaining a fundamental understanding of Cu’s role in the dehydrogenation reaction, ultrahigh vacuum (UHV) studies offer an excellent way to quantitatively probe oxygenfree reactions on well-defined metal single crystals and, being accessible to scanning probe studies, to enable an understanding of the catalytic mechanism at the molecular level. © XXXX American Chemical Society

Alcohol dehydrogenation on single crystal surfaces of transition metals has been studied previously.10−18 However, the majority of these surfaces required surface oxygen to facilitate the dehydrogenation of alcohols to aldehydes because of the low surface activity. For Cu single crystals, alcohol dehydrogenation to aldehyde at low conversions was reported on oxygen-free Cu(110) and (210) surfaces under UHV conditions.11,12,19,20 Photoelectron spectroscopy and temperature programmed experiments showed that the first step in the reaction of an alcohol with a Cu single crystal surface is dissociation of the O−H bond to form the alkoxide. In addition to bond activation by the metal surface, intermolecular interactions between selfassembled molecules on surfaces can also lead to dehydrogenation.21,22 Our recent temperature programmed reaction study revealed that surface-bound water can enhance methanol dehydrogenation on Cu(111) due to the formation of hydrogen-bonded clusters between water and methanol, which increase the conversion to methoxy by stabilizing the removed hydrogen atom.10,23−25 The disadvantage of cocatalysis by water is that it needs to be separated from the aldehyde product. Previous UHV studies of alcohol dehydrogenation on Cu(110) and (210) were all performed by depositing the alcohols at low temperatures at which the Cu surface can trap the small amounts of chamber background Received: March 28, 2017 Revised: May 23, 2017 Published: May 23, 2017 A

DOI: 10.1021/acs.jpcc.7b02957 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

alcohol at 1 × 10−8 mbar to achieve the desired alcohol exposures while the sample was held at 200 K in order to isolate adsorbed alkoxy species and prevent water adsorption on Cu. The temperature programmed experiments were run using a linear heating rate of 5 K/s. Surface oxygen was created by dissociating the background-dosed molecular oxygen (>99% purity) on Cu(110) at 300 K. The oxygen exposure was varied from 0 to 20 langmuirs. The coverages of adsorbed O were calibrated by AES by comparing to the peak-to-peak amplitudes of the oxygen AES signal to saturation coverage, which is known to be 0.50 ML.26 The roughened Cu(110) surfaces were prepared by Ar+ sputtering (1 h sputtering at 1 keV and 3 μA) without the subsequent annealing. Note that the sputtering gun setups are different in the two temperature programmed reaction chambers; therefore, different sputtering parameters were used to achieve similar results.

water, which was not previously considered as co-catalyst.11,12,19,20 Therefore, desorption experiments at temperatures at which water is not adsorbed on the surface are important in order to conclusively rule out the effect of water. In addition, the techniques used in previous reports lack spatial resolution, and as such the active sites for alcohol dehydrogenation on Cu model surfaces have not been studied with atomicscale resolution. Herein, a detailed, quantitative investigation of the reactions of methanol and ethanol on Cu(111) and Cu(110) surfaces is reported and provides further insight into the low conversion rates of ethanol and methanol to their respective aldehydes under UHV conditions. This work demonstrates that alcohol dehydrogenation occurs at the steps and kinks on the Cu single crystals in the absence of water and surface oxygen. The amount of dehydrogenation can be tuned by changing the surface defect density via thermal annealing and Ar ion (Ar+) sputtering of the Cu substrates. Oxygen-free Cu model surfaces exhibit no deactivation and high selectivity for the alcohol dehydrogenation reaction.

■

RESULTS AND DISCUSSION Influence of Defects: Flat and Roughened Cu(111) Surfaces. Ethanol can be converted in small yields to acetaldehyde by reaction with a clean, flat single crystal of Cu(111). A temperature programmed trace taken on Cu(111) exposed to 0.8 langmuir of ethanol at 85 K (Figure 1a) shows that most of the ethanol desorbs reversibly from Cu(111) upon heating. The peaks at 220 and 180 K correspond to ethanol desorption from steps and terraces, respectively (see Support Information Figure S1). Very small acetaldehyde and H2 signals were detected at 325 and 370 K in the experiment (Figure 1a), indicating a small yield for ethanol dehydrogenation. In general, low index surfaces have weaker binding and lower reactivity than high index or rough surfaces. Considering that Cu(111) is the most thermodynamically stable surface of Cu single crystal, the observed low ethanol dehydrogenation activity suggests that the flat Cu(111) terraces have a higher barrier to reaction than for ethanol desorption. The results also suggest that the low concentration of defects in the form of surface step edges on Cu(111) may be responsible for the small amount of dehydrogenation. To test this hypothesis, temperature programmed experiments were performed with ethanol on both flat and roughened Cu surfaces created by Ar+ sputtering. To eliminate the potential influence of different ethanol sticking coefficients for the flat and rough surfaces, we exposed both surfaces to 3 langmuirs of ethanol to saturate both terraces and step sites. Ethanol desorption is observed at 150 and 180 K from both the flat and sputtered Cu(111) surfaces (Figure 1b), which corresponds to the multilayer and monolayer desorption from terraces sites, respectively. There is a shoulder in the desorption trace at 220 K on the flat surface corresponding to desorption from steps which increases in intensity after the surface is roughened by sputtering. This result, combined with the STM data discussed later, indicates that there is indeed a much higher step density on the sputtered surface than on the flat one. In addition, a new peak at 255 K appears in the desorption trace from the sputtered surface, indicating the existence of surface sites that bind ethanol more strongly than at step edges on Cu(111). STM images shown later show that after annealing only kink sites at the step edges of the sputtered Cu(111) show the presence of adsorbates; therefore, the peak at 255 K is assigned to ethanol desorption from kink sites. Quantitative analysis of the temperature programmed traces corrected for ionization cross sections, fragmentation patterns, and quadrupole sensitivity reveals that ∼25% less ethanol desorbs from the

■

EXPERIMENTAL SECTION The STM experiments were conducted in an Omicron UHV low temperature scanning tunneling microscope (LT-STM). Cu(111) substrates were cleaned under UHV by multiple alternating Ar+ sputtering (1 keV/15 μA) and thermal annealing (1000 K) cycles. The roughened surfaces were prepared through short Ar+ sputtering (30 s sputtering at 1 keV and 15 μA) without a subsequent annealing step. Ethanol and methanol were introduced into the STM chamber with background dosing at the substrate temperature of 5 K. Exposures are quoted in langmuirs (1 langmuir = 1 × 10−6 Torr·s). STM images were recorded at 5 K with a background pressure below 1 × 10−11 mbar. The second UHV chamber incorporates a Hiden quadrupole mass spectrometer for performing temperature programmed desorption experiments on Cu(111) with a base pressure below 1 × 10−10 mbar. A clean Cu(111) sample was prepared via Ar+ sputtering and annealing to 750 K. Ethanol and methanol were deposited on the Cu(111) crystal by varying the backfilling time of ethanol and methanol at 1 × 10−8 mbar to achieve the desired alcohol exposures while the sample was held at 85 or 180 K, respectively. Temperature programmed experiments were run using a linear heating rate of 1 K/s. Both molecular ethanol and methanol were tracked by the major contributors to their cracking pattern m/z 31, and the traces for acetaldehyde and formaldehyde were produced by subtracting the ethanol and methanol contributions separately from the trace of m/z 29. H2O and H2 desorptions were detected by tracking their parent ions, m/z 18 and 2, respectively. The roughened Cu surfaces were prepared in this chamber by Ar+ sputtering for 10 min at 1.5 keV and 10 μA without subsequent annealing. The temperature programmed experiments on Cu(110) were performed in the third UHV chamber with a base pressure of ∼1 × 10−9 mbar. The chamber was outfitted with a mass spectrometer for the detection of products and Auger electron spectroscopy (AES) to verify the cleanliness of the surface and to calibrate the surface O coverages on Cu(110). The Cu(110) single crystal was cleaned via 1 h Ar+ sputtering at room temperature followed by thermal annealing at 850 K for 30 min, and its cleanliness was confirmed by AES. Alcohols were deposited onto Cu(110) by varying the backfilling time of B

DOI: 10.1021/acs.jpcc.7b02957 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

roughened Cu(111) due to the uptake of background water during cooling and alcohol dosing (Figure 1b). In order to eliminate the influence of water on the reaction, a second set of experiments were performed in which ethanol was adsorbed above the desorption temperature of molecular water on the Cu(111) surface (180 K) prior to the temperature programmed measurements, and no desorption of water was observed (see Figure S3). Consistent with the result shown in Figure 1, roughened Cu(111) has a significantly higher yield of acetaldehyde (∼10 times) than flat Cu(111) (Figure 2a) for the dry dehydrogenation reaction. As a comparison, these water-free experiments were also performed for methanol on both flat and roughened Cu(111) (see Figure S4). The results reveal that more (∼14 times) formaldehyde was produced on the roughened surface than on the flat one (Figure 2b). Clearly, anhydrous dehydrogenation proceeds readily on the roughened surfaces. Further, under the conditions studied here the dry dehydrogenation of alcohols also occurs at defects on the (110) facet of Cu. Both flat and sputtered Cu(110) surfaces were exposed to saturation coverages of methanol and ethanol (3 langmuirs) at 200 K, at which temperature Cu(110) is free of co-adsorbed water (see Figure S5). The presence of desorption peaks for acetaldehyde and formaldehyde on flat Cu(110) (Figure 2c,d) demonstrates dry dehydrogenation of ethanol and methanol on clean, flat Cu(110). Just as with Cu(111), more acetaldehyde and formaldehyde were produced on the roughened surface created by Ar + sputtering, demonstrating that dry alcohol dehydrogenation also occurs at surface defects of Cu(110). Whether the rate of product formation was limited by desorption or the reaction was easily determined by comparing desorption temperatures of the product species themselves to the temperatures at which they were formed in the alcohol dehydrogenation. Molecularly adsorbed acetaldehyde and formaldehyde desorbed from the Cu single crystals at temperatures below 250 K (see Figure S6 and refs 12, 25, 28, and 29). The reaction products formed well above 250 K (Figure 2); the formation of acetaldehyde and formaldehyde are both reaction rate-limited processes. In contrast, H2 evolves from the surface in desorption rate-limited processes at higher temperature than the aldehydes. For example, while acetaldehyde desorbs at 325 K when ethanol reacts with Cu(111), H2 was detected between 330 and 370 K (Figure 1) due to the high energy barrier for recombinative desorption of H2 from Cu.30 Our results indicate that ethanol has a higher reactivity than methanol on both (111) and (110) Cu surfaces and that Cu(110) steps are more active than Cu(111) steps for ethanol and methanol dehydrogenation. Acetaldehyde desorbs from roughened Cu(111) and Cu(110) at 325 and 305 K, respectively, and formaldehyde desorbs at 370 and 347 K (Figure 2). Since the evolution of both acetaldehydes is reaction-limited, the lower temperature for acetaldehyde formation indicates that the energetic barrier for producing aldehyde is lower for ethoxy than methoxy on both Cu surfaces, in agreement with previous work.11,12 Therefore, the reactivity of alcohols discussed here is dominated by the C−H bond activation, which, estimated from desorption temperatures using the Redhead equation and a pre-factor of 1012,31,32 are 19.3 and 17.6 kcal/mol for ethanol on Cu(111) and Cu(110), respectively, and 22.1 and 19.8 kcal/mol for methanol on Cu(111) and Cu(110).

Figure 1. Temperature programmed traces for the reaction of ethanol on flat and roughened Cu(111) crystals after (a) 0.8 and (b) 3 langmuir ethanol exposures at 85 K. Desorption traces taken on the flat (black) and roughened (red) Cu(111) surfaces for reactants ethanol (m/z 31), and the products acetaldehyde (m/z 29), H2O (m/z 18), and H2 (m/z 2) are plotted.

monolayer (from terraces, steps, and kinks) on the roughened surface than from the flat one. Despite this difference, the yields of acetaldehyde and H2 on the roughened Cu(111) are ∼10 times higher than those on the flat surface. Further analysis of the traces of ethanol and acetaldehyde (see Table S1) reveals that ethanol dehydrogenation contributes ∼63% to the difference in ethanol desorption on the roughened surface vs the flat surface. The remaining difference is attributed to the lower initial ethanol coverage on the roughened surface because that ethanol cannot form the densely packed hydrogen bonded arrays on the rough surfaces created in the Ar+ sputtering.10 Since only surface defects (steps and kinks) on Cu(111) significantly increase after an Ar+ sputtering, the dehydrogenation of ethanol on Cu(111) must be induced by these undercoordinated sites. Dry Alcohol Dehydrogenation on Cu(111) and Cu(110). Water does not play a role in the experiments reported here. Even after a bakeout, trace amounts of water are present in UHV chambers, and water has been shown to facilitate alcohol dehydrogenation on Cu (see Figure S2).25,27 Indeed, small water peaks were observed in temperature programmed experiments around 150 K on both flat and C

DOI: 10.1021/acs.jpcc.7b02957 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

Figure 2. Temperature programmed traces for aldehyde yield for alcohol dehydrogenation on the flat and roughened Cu model surfaces: (a) formaldehyde from Cu(111), (b) acetaldehyde from Cu(111), (c) formaldehyde from Cu(110), and (d) acetaldehyde from Cu(110). Ethanol and methanol were exposed to Cu(111) and Cu(110) at saturation coverage at 180 and 200 K, respectively. Desorption traces taken on the flat (black) and roughened (red) surfaces and aldehyde products (m/z 29) are plotted.

dehydrogenate to the corresponding aldehyde (see Figure S9).12,13 Note that similar sub-monolayer ordered surface oxygen structures cannot be easily formed on Cu(111); hence, temperature programmed spectra of ethanol were used for calibration of defect sites on Cu(111). The concentration of active sites on Cu(110) was determined to be ∼7% for higher alcohols, including ethanol, and between 4% and 7% for methanol as shown in Figure 3. Imaging of Reaction Intermediates by STM: Verification of Step Edge Reactivity. The intermediates of alcohol dry dehydrogenation were directly observed at the step edges on both flat and roughened Cu(111) using LT-STM. As an example, the evolution of surface structures on both flat and roughened Cu(111) after ethanol exposure and thermal annealing is shown in Figure 4. Ethanol molecules are randomly distributed on Cu(111) terraces after 0.5 langmuir ethanol exposure at 5 K because they are immobile at this temperature (Figure 4a). Following annealing at 160 K all the ethanol desorbs from the Cu(111) terraces (Figure 4b). The slight difference of ethanol desorption temperatures measured through the temperature programmed experiment and STM is attributed to the different heating rates applied to the substrates in the two different experimental setups, the temperature measurement in the former experiment being the most accurate (see Supporting Information). Note that the flat Cu(111) surface has large terraces, but the images shown here were deliberately selected to show that there are some step edges present. Sites for ethanol dehydrogenation on the roughened Cu(111) are observable with STM and appear to be exclusively under-coordinated Cu sites. In contrast to the flat

Aldehydes are more easily produced on Cu(110) than on Cu(111), indicating the higher activity of Cu(110) steps than Cu(111) steps. This can be attributed to the existence of more under-coordinated Cu atoms at steps of Cu(110) (6-fold coordinated at (100) steps) than at steps of Cu(111) (7-fold coordinated). Density of Active Sites on As Prepared Cu(111) and (110). The aldehyde yields observed on the Cu(111) surface were more enhanced by sputtering than on Cu(110) due to the intrinsically higher density of defect sites on the Cu(110) crystal. Specifically, while aldehyde yields were at least 10 times higher on the roughened Cu(111) surfaces than flat surface, sputtering only increased the aldehyde yields about ∼2−3 times on Cu(110) (Figure 2). The defect densities were quantified as follows. The density of step sites on flat Cu(111) was determined to be approximately 2% of all the surface adsorption sites for ethanol by analyzing the desorption traces areas in Figure 1b (see Supporting Information). Further analysis reveals that ∼1% of the total surface sites activate ethanol under the conditions used in these experiments (see Supporting Information), which is consistent with the density of step edges on the flat surface. To calibrate the effective active site density on Cu(110) pre-adsorbed O was titrated with the alcohols, and the amount of aldehyde formed was compared to the results from the oxygen-free surface. The known saturation coverage of surface atomic oxygen of 0.5 ML in the (1 × 2) reconstruction on Cu(110) was used to calibrate the O precoverages (see Figures S7 and S8).26 It is known that at low O coverages each oxygen adatom reacts with two alcohol molecules to form two alkoxy species, which subsequently D

DOI: 10.1021/acs.jpcc.7b02957 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

the ethanol species from the terraces (Figure 4d); however, adsorbates remain at steps and kinks (highlighted in the inset). Close inspection reveals that the adsorbates at steps on sputtered Cu(111) also appear at the steps on the flat Cu surface after the 160 K annealing step (Figure 4b, white circles in the inset). These adsorbates are very stable and are still present after further annealing to 290 K (see Figure S10). This temperature is higher than that for the desorption of molecular ethanol from all surface sites (terraces, steps, and kinks) on Cu but lower than the temperature at which acetaldehyde forms. Since hydrogen appears as a depression in STM images,33 these adsorbates must be the intermediates of ethanol dehydrogenation and are assigned to ethoxy species, as discussed next. The significant number of intermediates observed at the defect sites on the roughened surface and their thermal stability further confirms that ethanol dehydrogenation occurs at undercoordinated sites on Cu(111). The impact of water can be excluded in our STM experiments because the extremely cold cryostat (5 K) traps all the chamber background water, and the surface stays completely clean for up to 24 h. Therefore, the ethanol dehydrogenation at Cu(111) steps occurs dry. Mechanism of Alcohol Dehydrogenation on Model Cu Surfaces. Ethanol and methanol dehydrogenation on the Cu model surfaces involves a two-step reaction in which O−H and C−H bonds are activated at two distinct temperatures. In the process of alcohol dehydrogenation to aldehyde, it was previously reported that on Cu(110) the cleavage of the O−H bond is more facile than C−H, leading to the formation of stable alkoxy species at low temperatures.11 This was confirmed here, for example, by Fourier transform infrared reflection− absorption spectroscopy (FT-IRAS) following the adsorption of ethanol on the roughened Cu(111) and annealing to 260 K; the absence of O−H but persistence of C−H vibration modes is consistent with an ethoxy intermediate on Cu(111) (see Figure S11). Therefore, the reaction intermediates at surface steps observed in the STM images are attributed to ethoxy species. Importantly, by combining the STM and temperature programmed experimental results, we can conclude that ethanol undergoes a two-step dehydrogenation at Cu surface defects during the thermal annealing in accordance with previous work on Cu(110);12 the O−H bond is broken below 160 K to form ethoxy, and then the C−H bond is cleaved at 325 K to form acetaldehyde, which desorbs immediately in a reaction rate limited process followed by the recombination of surface-bound hydrogen in a desorption rate limited process. Active Site Stability and Reaction Selectivity. Cu model surfaces exhibit no deactivation after multiple temperature programmed experiments and high selectivity to the desired aldehyde and hydrogen products of alcohol dehydrogenation. An important aspect of any catalytic cycle is the fate of the active site after the reaction has occurred, e.g., poisoning with byproducts. Repeated temperature programmed experiments shows that the acetaldehyde yield stays constant upon repeated reaction cycles on the same flat Cu(111) surface (see Figure S12). This confirms that no deactivation occurs during ethanol dehydrogenation on Cu(111). Furthermore, STM images after the deposition of ethanol and annealing to 450 K revealed a completely clean surface with no residual surfacebound species that could poison the active sites for dry dehydrogenation (see Figure S13). In terms of reaction selectivity, no species except the characteristic m/z fragments for alcohols, H2, and aldehydes in the temperature programmed

Figure 3. Titration of the active site concentration on flat Cu(110) for the formation of methoxy (green), ethoxy (blue), 1-propoxy (red), and 1-butoxy (black) on O-free Cu(110). Alcohols were deposited onto flat Cu(110) at 200 K (green asterisk) and 270 K (green dot) for methanol, 200 K (blue asterisk) and 250 K (blue dot) for ethanol, 270 K for 1-propanol (red square), and 270 K for 1-butanol (black square).

Figure 4. (a, b) STM images (50 nm × 50 nm) of flat Cu(111) after 0.5 langmuir ethanol exposure at 5 K (a) followed by a 160 K annealing step (b; derivative image). The inset of (b) (15 nm × 15 nm) shows ethoxy adsorbed at Cu(111) steps (highlighted with white circles). (c, d) STM images (50 nm × 50 nm) of the roughened Cu after 0.5 langmuir ethanol exposure at 5 K (c) followed by a 160 K annealing step (d; derivative image). The inset (15 nm × 15 nm) in (c) shows ethanol molecules adsorbed on a small terrace of Cu(111). The inset (15 nm × 15 nm) in (d) highlights Cu steps and kinks with adsorbed ethoxy intermediates. STM images were acquired at 5 K using non-perturbative scanning parameters: 10−30 mV, 30 pA.

Cu surface, many more steps and kinks appear on the sputtered Cu(111) surface in the STM image (Figure 4c,d). Following exposure of the sputtered surface to 0.5 langmuir of ethanol, randomly distributed ethanol molecules are observed on the small terraces present (Figure 4c, inset). As expected from the results on flat Cu(111), annealing to 160 K completely removes E

DOI: 10.1021/acs.jpcc.7b02957 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

(5) Su, S.; Zaza, P.; Renken, A. Catalytic Dehydrogenation of Methanol to Water-Free Formaldehyde. Chem. Eng. Technol. 1994, 17, 34−40. (6) Hashimoto, K.; Toukai, N. Dehydrogenation of Alcohols over Potassium Zinc Aluminum Silicate Hydroxide. J. Mol. Catal. A: Chem. 1999, 145, 273−280. (7) Ren, L.-P.; Dai, W.-L.; Cao, Y.; Li, H.; Fan, K. First Observation of Highly Efficient Dehydrogenation of Methanol to Anhydrous Formaldehyde over Novel Ag-SiO2-MgO-Al2O3 Catalyst. Chem. Commun. 2003, 3030−3031. (8) Ren, L.-P.; Dai, W.-L.; Yang, X.-L.; Cao, Y.; Li, H.; Fan, K.-N. Novel Highly Active Ag−SiO2−Al2O3−ZnO Catalyst for the Production of Anhydrous HCHO from Direct Dehydrogenation of CH3OH. Appl. Catal., A 2004, 273, 83−88. (9) Takagi, K.; Morikawa, Y.; Ikawa, T. Catalytic Activities of Coppers in the Various Oxidation States for the Dehydrogenation of Methanol. Chem. Lett. 1985, 14, 527−530. (10) Boucher, M. B.; Marcinkowski, M. D.; Liriano, M. L.; Murphy, C. J.; Lewis, E. A.; Jewell, A. D.; Mattera, M. F. G.; Kyriakou, G.; Flytzani-Stephanopoulos, M.; Sykes, E. C. H. Molecular-Scale Perspective of Water-Catalyzed Methanol Dehydrogenation to Formaldehyde. ACS Nano 2013, 7, 6181−6187. (11) Bowker, M.; Madix, R. J. XPS, UPS and Thermal Desorption Studies of Alcohol Adsorption on Cu(110). Surf. Sci. 1982, 116, 549− 572. (12) Wachs, I. E.; Madix, R. J. The Selective Oxidation of CH3OH to H2CO on a Copper(110) Catalyst. J. Catal. 1978, 53, 208−227. (13) Bowker, M. Active Sites in Methanol Oxidation on Cu(110) Determined by STM and Molecular Beam Measurements. Top. Catal. 1996, 3, 461−468. (14) Gong, J.; Flaherty, D. W.; Ojifinni, R. A.; White, J. M.; Mullins, C. B. Surface Chemistry of Methanol on Clean and Atomic Oxygen Pre-Covered Au(111). J. Phys. Chem. C 2008, 112, 5501−5509. (15) Barros, R. B.; Garcia, A. R.; Ilharco, L. M. Effect of Oxygen Precoverage on the Reactivity of Methanol on Ru(001) Surfaces. J. Phys. Chem. B 2004, 108, 4831−4839. (16) Fukui, K.-i.; Motoda, K.; Iwasawa, Y. Selective Oxidation of Methanol by Extra Oxygen Species on One-Dimensional Mo Rows of a Mo(112)−(1 × 2)-O Surface. J. Phys. Chem. B 1998, 102, 8825− 8833. (17) Felter, T. E.; Weinberg, W. H.; Lastushkina, G. Y.; Zhdan, P. A.; Boreskov, G. K.; Hrbek, J. The Adsorption of Methanol on Ag(111) and Its Reaction with Preadsorbed Oxygen. Appl. Surf. Sci. 1983, 16, 351−364. (18) Parmeter, J. E.; Xudong, J.; Goodman, D. W. The Adsorption and Decomposition of Methanol on the Rh(100) Surface. Surf. Sci. 1990, 240, 85−100. (19) Chen, A. K.; Masel, R. Direct Conversion of Methanol to Formaldehyde in the Absence of Oxygen on Cu(210). Surf. Sci. 1995, 343, 17−23. (20) Sexton, B. A.; Hughes, A. E.; Avery, N. R. Surface Intermediates in the Reaction of Methanol, Formaldehyde and Methyl Formate on Copper (110). Appl. Surf. Sci. 1985, 22, 404−414. (21) Contini, G.; Gori, P.; Di Giovannantonio, M.; Zema, N.; Turchini, S.; Catone, D.; Prosperi, T.; Palma, A. Interplay between Supramolecularity and Substrate Symmetry in the Dehydrogenation of D-Alaninol on Cu(100) and Cu(110) Surfaces. J. Phys. Chem. C 2013, 117, 10545−10551. (22) Supramolecular Chemistry at Surfaces; The Royal Society of Chemistry: 2016; Chapter 5, pp 252−302. (23) Pan, M.; Pozun, Z. D.; Yu, W.-Y.; Henkelman, G.; Mullins, C. B. Structure Revealing H/D Exchange with Co-Adsorbed Hydrogen and Water on Gold. J. Phys. Chem. Lett. 2012, 3, 1894−1899. (24) Brush, A. J.; Pan, M.; Mullins, C. B. Methanol O−H Bond Dissociation on H-Precovered Gold Originating from a Structure with a Wide Range of Surface Stability. J. Phys. Chem. C 2012, 116, 20982− 20989. (25) Shan, J.; Lucci, F. R.; Liu, J.; El-Soda, M.; Marcinkowski, M. D.; Allard, L. F.; Sykes, E. C. H.; Flytzani-Stephanopoulos, M. Water Co-

experiments on Cu model surfaces were detected, indicating 100% selectivity for alcohol dehydrogenation.

■

CONCLUSIONS Cu(111) and Cu(110) model surfaces dehydrogenate alcohols to aldehydes and H2 in the absence of surface oxygen or water. STM reveals that dehydrogenation occurs at surface defects in the form of step edges and kinks. Consistent with this, the degree of dehydrogenation can be increased by increasing the density of defect sites by roughening the surface via Ar+ sputtering. The combination of temperature programmed experiments and STM studies indicates that alcohol molecules undergo a two-step dehydrogenation at Cu surface defects: the O−H bond is broken at low temperature to form alkoxy intermediates, and then the C−H bond is broken at higher temperature to form aldehyde and H2 by H adatom recombination. Importantly, alcohol dehydrogenation to the desired aldehyde and H2 products occurs with 100% selectivity on the Cu surfaces, and the surfaces do not deactivate after repeated reaction cycles. This work demonstrates the effectiveness of surface defects for catalyzing the dry dehydrogenation of alcohols on Cu and suggests that if highly defective Cu architectures could be synthesized and remain stable under reaction conditions they may offer a new approach for the dry dehydrogenation of alcohols.

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcc.7b02957. Figures S1−S13 (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail [email protected]. ORCID

Cynthia M. Friend: 0000-0002-8673-9046 E. Charles H. Sykes: 0000-0002-0224-2084 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We gratefully acknowledge support by Integrated Mesoscale Architectures for Sustainable Catalysis (IMASC), an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Basic Energy Sciences under Award DE-SC0012573.

■

REFERENCES

(1) Reuss, G.; Disteldorf, W.; Grundier, O.; Hilt, A. In Ullmann’s Encyclopedia of Industrial Chemistry, 5th ed.; Ullmann, I. F., Gerhartz, W., Yamamoto, Y. S., Campbell, F. T., Pfefferkorn, R., Rounsaville, J. F., Eds.; Wiley-VCH: Deerfield Beach, FL, 1985; p 619. (2) Eckert, M.; Fleischmann, G.; Jira, R.; Bolt, H. M.; Golka, K. Acetaldehyde. In Ullmann’s Encyclopedia of Industrial Chemistry; WileyVCH Verlag GmbH & Co. KGaA: 2000. (3) Usachev, N. Y.; Krukovskii, I. M.; Kanaev, S. A. The Nonoxidative Methanol Dehydrogenation to Formaldehyde: (A Review). Pet. Chem. 2004, 44, 379−394. (4) Dai, W.; Ren, L. Dehydrogenation of Alcohols: Formaldehyde. Handbook of Heterogeneous Catalysis; Wiley-VCH: Weinheim, 1997; Vol. 101, pp 994−1006. F

DOI: 10.1021/acs.jpcc.7b02957 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C Catalyzed Selective Dehydrogenation of Methanol to Formaldehyde and Hydrogen. Surf. Sci. 2016, 650, 121−129. (26) Hartmann, N.; Madix, R. J. Growth and Ordering of Cu−O Islands During Oxygen Adsorption on Cu(1 1 0) at 470 K. Surf. Sci. 2001, 488, 107−122. (27) Li, H.-J.; Lausche, A. C.; Peterson, A. A.; Hansen, H. A.; Studt, F.; Bligaard, T. Using Microkinetic Analysis to Search for Novel Anhydrous Formaldehyde Production Catalysts. Surf. Sci. 2015, 641, 105−111. (28) Gomes, J. R. B.; Gomes, J. A. N. F.; Illas, F. First-Principles Study of the Adsorption of Formaldehyde on the Clean and Atomic Oxygen Covered Cu(1 1 1) Surface. J. Mol. Catal. A: Chem. 2001, 170, 187−193. (29) Bryden, T. R.; Garrett, S. J. Adsorption and Polymerization of Formaldehyde on Cu(100). J. Phys. Chem. B 1999, 103, 10481−10488. (30) Anger, G.; Winkler, A.; Rendulic, K. D. Adsorption and Desorption Kinetics in the Systems H2/Cu(111), H2/Cu(110) and H2/Cu(100). Surf. Sci. 1989, 220, 1−17. (31) Madix, R. J.; Telford, S. G. Primary and Secondary Kinetic Isotope Effects for Methoxy Dehydration and Cu(110): Absence of Tunnelling Corrections. Surf. Sci. 1995, 328, L576. (32) Redhead, P. A. Thermal Desorption of Gases. Vacuum 1962, 12, 203−211. (33) Jewell, A. D.; Peng, G.; Mattera, M. F. G.; Lewis, E. A.; Murphy, C. J.; Kyriakou, G.; Mavrikakis, M.; Sykes, E. C. H. Quantum Tunneling Enabled Self-Assembly of Hydrogen Atoms on Cu(111). ACS Nano 2012, 6, 10115−10121.

G

DOI: 10.1021/acs.jpcc.7b02957 J. Phys. Chem. C XXXX, XXX, XXX−XXX