Quantitative and Atomic-Scale View of CO-Induced ... - ACS Publications

Subscriber access provided by University of Newcastle, Australia

Article

Quantitative and Atomic Scale View of CO-Induced Pt Nanoparticle Surface Reconstruction at Saturation Coverage via DFT Calculations Coupled with in-situ TEM and IR Talin Avanesian, Sheng Dai, Matthew James Kale, George W. Graham, Xiaoqing Pan, and Phillip Christopher J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b01081 • Publication Date (Web): 06 Mar 2017 Downloaded from http://pubs.acs.org on March 7, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of the American Chemical Society is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

Quantitative and Atomic Scale View of CO-Induced Pt Nanoparticle Surface Reconstruction at Saturation Coverage via DFT Calculations Coupled with in-situ TEM and IR Talin Avanesian†,‡, Sheng Dai§,‡, Matthew J. Kale†,‡, George W. Graham§, Xiaoqing Pan§,|| *, Phillip Christopher†,⊥,# * †

Department of Chemical & Environmental Engineering, University of California, Riverside, Riverside, California 92521, United States. §

Department of Chemical Engineering and Materials Science, University of California Irvine, Irvine, CA, 92697, United States.

||

Department of Physics and Astronomy, University of California Irvine, Irvine, CA, 92697, United States.

⊥Program in Materials Science and Engineering, University of California, Riverside, Riverside, California 92521, United States. #

UCR Center for Catalysis, University of California, Riverside, Riverside, California 92521, United States.

ABSTRACT: Atomic scale insights into how supported metal nanoparticles catalyze chemical reactions are critical for the optimization of chemical conversion processes. It is well known that different geometric configurations of surface atoms on supported metal nanoparticles have different catalytic reactivity and that the adsorption of reactive species can cause reconstruction of metal surfaces. Thus, characterizing metallic surface structures under reaction conditions at atomic scale is critical for understanding reactivity. Elucidation of such insights on high surface area oxide supported metal nanoparticles has been limited by less than atomic resolution typically achieved by environmental transmission electron microscopy (TEM) when operated under realistic conditions and a lack of correlated experimental measurements providing quantitative information about the distribution of exposed surface atoms under relevant reaction conditions. We overcome these limitations by correlating Density Functional Theory (DFT) predictions of adsorbate-induced surface reconstruction visually with atom-resolved imaging by in-situ TEM and quantitatively with sample averaged measurements of surface atom configurations by in-situ infrared (IR) spectroscopy all at identical saturation adsorbate coverage. This is demonstrated for Platinum (Pt) nanoparticle surface reconstruction induced by CO adsorption at saturation coverage and elevated (> 400K) temperature, which is relevant for the CO oxidation reaction under cold-start conditions in the catalytic convertor. Through our correlated approach, it is observed that the truncated octahedron shape adopted by bare Pt nanoparticles undergoes a reversible, facet selective reconstruction due to saturation CO coverage, where {100} facets roughen into vicinal stepped high miller index facets, while {111} facets remain intact.

1.

Introduction

Supported metal catalysts contain metal nanoparticles with exposed surface structures that can physically reconstruct under reaction conditions due to the adsorption of reactive molecules.1–3 It is well established that different geometric configurations of metal surface atoms have different reactivity, and that characterizing the metal surface structure under industrially-relevant reaction conditions at an atomic scale is critical for understanding catalytic performance.4–6 Reconstruction of Pt surfaces due to the adsorption of CO is one of the most studied examples of adsorbate-mediated catalytic metal surface reconstruction, and is the first step in the CO oxidation reaction that occurs in automotive catalytic converters.7,8

Atomic scale descriptions of Pt surface reconstruction by adsorbed CO have been effectively achieved on model Pt single crystals using scanning tunneling microscopy (STM), where the CO adsorption induced formation of increased concentrations of under-coordinated (UC, atoms with a metal-metal coordination number of 6 or 7) Pt surface atoms has been observed for all facets, except {111}.9–13 However, detailed descriptions of CO-induced reconstruction of model Pt single crystals have not yet translated into a complete picture of the same process on high surface area oxide-supported Pt nanoparticles.5,7,14,15 This is due to the higher degree of structural complexity associated with supported metal nanoparticles, compared to model single crystals, where multiple surface facets with varying levels of coordination are simultaneously exposed.16,17

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 11

In-situ TEM analyses of CO induced Pt nanoparticle surface reconstruction have shown direct evidence of increased concentrations of UC surface sites on high surface area supported catalysts. However, coverage of CO on Pt surfaces varies with CO pressure up to ~10 torr,18 and insitu TEM reports operating below saturation coverage make it difficult to correlate changes in atomic surface structure to adsorbate coverage.15 In-situ TEM reports operating at saturation CO coverage on Pt nanoparticle surfaces have also shown evidence of CO induced reconstructions, but have been limited by less than atomic resolution,14 and in all cases TEM provides only extremely small sample sizes.19,20 In-situ spectroscopic interrogations have also shown signatures of CO-induced reconstruction of Pt nanoparticle surfaces that are consistent with increased concentrations of UC Pt atoms.7,21–24 Spectroscopic analysis is typically a sample averaged measurement that probe large numbers of nanoparticles simultaneously, suggesting that CO induced reconstruction of Pt nanoparticle surfaces is pervasive throughout a majority of Pt nanoparticles in a given sample. However, no previous analyses have provided a substantiated view of how saturated layers of CO on Pt nanoparticles induce changes in the surface structure at the atomic scale, both from the view point of how Pt surface structures geometrically change and how this quantitatively changes the distribution of active site geometries.

used to model low Miller-index Pt surface facets {100}, {110}, and {111} and higher index facets, {210}, {211}, {310}, and {311} which have been shown to well represent the behavior of catalytic particles with diameters greater than ~1.6 nanometers.28 The consecutive metal slabs were separated by 15 Å vacuum space in z direction in periodically repeated unit cells in x and y directions. A grid space of h = 0.2 was used. The Brillouin zone sampling for all the unit cells was performed with a 8×8×1 Monkhorst–Pack kpoint set, which was assured to be sufficient for the convergence of the smallest unit cells.29

Here, we characterize atomic scale details associated with the structural rearrangement of supported Pt nanoparticle surfaces induced by the adsorption of CO at saturation coverage and elevated temperature. This is achieved by using density functional theory (DFT) based Wulff construction models to predict reconstruction of Pt particles induced by CO saturation coverage, followed by visually and quantitatively correlating these results with measurements made by in-situ scanning transmission electron microscopy (STEM) and infrared (IR) spectroscopy, respectively. It was observed that the truncated octahedron shape adopted by bare Pt nanoparticles undergoes a reversible, facet specific reconstruction due to CO adsorption, where flat {100} facets roughen into vicinal stepped high miller index facets, while flat {111} facets remain intact. These findings have important ramifications for understanding reactivity of Pt catalysts at high CO coverage, for example at cold start conditions in automotive catalytic converters,7,8,25 and the general approach is expected to be useful for elucidation of surface structures on other catalytic materials that operate at high adsorbate coverage under reaction conditions.

hkl ads lated as hkl where hkl is the interfacial energy of the adsorbate-covered facet {hkl}, θ is the fractional adsorbate coverage, Eads is the coverage dependent molecular adsorption energy, and A is the surface area per adsorbate at 1 monolayer coverage. Further details are provided in the SI. Based on surface free energies of bare and CO saturated Pt surfaces, the Wulff construction method was used to predict nanoparticle shapes.30,31

2.

Methods 2.1.

Calculation Details

Pt surface energies as a function of facet and CO coverage were calculated within the DFT framework using the real space grid-based projector-augmented wave method (GPAW) open source code.26 The revised Perdew−Burke−Ernzerhof (RPBE) form of the generalized gradient approximation (GGA) was used to approximate exchange and correlation effects.27 Periodic slabs were

The surface energies of bare Pt facets,

γ

vac hkl

= (E

− NE

)

vac γ hkl

, were calcu-

2A

slab bulk cell lated as where Eslab is the total energy of the slab, N is the number of atoms in the slab, Ebulk is the energy of each atom in bulk and Acell is the surface area of unit cell. Relative surface free energies for the clean facets with respect to the {111} facet were converged as a function of slab thickness and number of relaxed layers (see details in Supporting information (SI) and Table S1-S2). The calculated surface energies were then normalized to the experimental surface energy for polycrystalline Pt (2.49 J/m2). The results are shown in Table S2.

The surface energies of CO covered facets were calcu-

γ ads = γ vac + θ E

A

γ ads

2.2. In-situ Scanning Transmission Electron Microscopy A carbon-supported Pt catalyst (Pt/C) was used for insitu STEM imaging to maximize contrast between the metal and support, and minimize metal-support interactions that may influence metal facet interfacial energies and the particle shape.32 The catalyst contained 4.8 ± 0.6 nm Pt particles, which adopted face centered cubic (fcc) crystal structures (Figure S1). In-situ electron microscopy was performed on a JEOL JEM-3100-R05 transmission electron microscope equipped with two spherical aberration correctors and a 300 kV cold field emission gun. high-angle annular dark field (HAADF)-STEM images were recorded using a convergence semi angle of 22 mrad and inner and outer collection angles of 83 and 165 mrad, respectively. To minimize beam irradiation, a relatively small beam current of 20 pA was used for imaging33 and the electron beam was turned off except for during image collection. For in-situ analysis Pt/C samples were first dispersed in a solvent, and the suspension was deposited directly onto a thermal E-chip, which is equipped with a thin ceramic


Page 3 of 11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


heating membrane controlled by the Protochips AtmosphereTM system. A second E-chip window was then placed on top of the thermal chip in the holder, creating a thin gas cavity sealed from the high vacuum of the TEM column. The Pt nanoparticles were situated between two SiN

membranes, each 30-50 nm in thickness, with a 5 micron gap in between. Importantly, the control system of the gas cell has a built-in leak-check function where after initial pump-

Figure 1. Calculation of bare and CO-saturated surface free energies for various Pt facets. a, Schematic diagram depicting the driving force for adsorbate induced reconstruction of surfaces through modification of competing facet surface energies. b, DFT-calculated CO coverage dependent surface energy of various Pt surface facets, see methods. We note that coverage increased above 1 for higher index facets, because a coverage of 1 correlates to CO saturation of all the lowest coordination number exposed adsorption sites for each facet. c, Models of various Miller indices of Pt slabs at the identified saturation CO coverage, identified from the DFT calculations shown in b.

down, the system will not allow progression to the next step if there is a significant leak in the gas cell system. In addition, since the static mode was used in our in-situ experiments (ie. gas was pumped into the cell, but was not continuously flowing), any gas leaks can be directly detected by monitoring cell pressure as a function of time during the experiment. Based on these approaches we can be sure that the relative partial pressure of CO in the insitu cell was maintained at ~25 Torr of CO, and this saturation CO coverage on Pt. HAADF-STEM image simulation was performed using the QSTEM simulation package.34 The simulations were carried out using a 512×512 pixel area and a single slice thickness of 1.96 Å. The microscopy parameters used for the simulations were the same as those for imaging. Further details are provided in the SI. 2.3. Quantitative in-situ Infrared Spectroscopy Pt/Al2O3 catalysts of various sizes were synthesized using incipient wetness impregnation and controlled sintering.7,35 The α- Al2O3 support was chosen to mimic the weakly interacting C support used in the in-situ STEM experiments, and maximize signal quality in the IR measurements. IR measurements were performed using a Praying Mantis diffuse reflection accessory (Harrick Scientific) and a Nicolet iS10 FTIR (Thermo Scientific) with 4 cm-1 resolution. A reduced catalyst at 298 K and 363 K were used as a background for room temperature and elevated temperature spectral measurements, respectively, and all spectra were obtained in Kubelka-Munk units (KM) units. Catalysts were loaded into a Harrick High temperature reaction chamber and reduced in-situ at 500 K in UHP H2

(Airgas) for 1 hour, followed by flushing the reactor for 3060 minutes with He (99.999%, Airgas) at 417K. After obtaining background spectra at 363K and 298K, each catalyst was exposed to a gas stream of 1% CO/He at 298K until spectra became stable. After 10 minutes of exposure to CO at 298K, the reactor was heated to 363 K, and spectra were taken for 30 minutes. Further details are provided in the SI regarding the experiments and the spectra deconvolution and quantitification. 3.

Results and Discussion

Equilibrium shapes of adsorbate-free, bare metal nanoparticles are dictated by free energies of competing surface facets and can be predicted using the Wulff construction method.30,31 Well-coordinated (WC, atoms with a metal-metal coordination number greater than 7), closedpacked metal surface facets generally have the lowest free energies and thus, most metal nanoparticles primarily expose {111} and {100} surface facets, adopting truncated octahedron shapes. Adsorbate-induced metal surface reconstructions are driven by changes in relative interfacial surface energies of competing surface facets upon the adsorption of molecular or atomic species. For example, Figure 1a shows an energy diagram of arbitrary surface facets {abc} and {xyz}. While {abc} is more stable than {xyz} in vacuum environment, the structure sensitivity of molecular adsorption can stabilize the surface energy of {xyz} to become the lowest energy facet, thereby driving surface reconstruction. To analyze thermodynamic driving forces for COinduced reconstruction of Pt nanoparticle surfaces, surface energies of exposed facets were calculated with DFT



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

and used as inputs to Wulff construction models (see Methods and SI for details). The low index facets used in this study ({111}, {100} and {110}) were chosen because they represent the most common facets found on bare Pt particles and the higher index facets ({211}, {311}, {210}, {310}) are the vicinal, stepped surfaces that are most likely to

Page 4 of 11

form upon “roughening” of the low index facets. We neglected the potential influences of support interactions on the shapes and reconstruction of Pt particles due to the minimal expected influence of weakly interacting α-Al2O3 and C supports

Figure 2. Visual comparison of DFT-calculated and in-situ STEM measurements of facet-specific CO induced Pt nanoparticle surface reconstruction. Wulff constructions of a 9.2 nm Pt particle based on DFT calculated surface free energies for a, bare surfaces, and b, CO saturated surfaces. Both images are tilted slightly off the zone axis (ZA). The green atoms represent WC Pt atoms, and the blue atoms represent UC Pt edge and corner atoms. The layers to the right are the top 3-4 {100} layers of the particle model, with the zone axis going into the page. Aberration-corrected STEM images of a 9 nm Pt particle taken along the zone axis c, at 423 K in 500 Torr N2, and d, at 423 K in 500 Torr of 5% CO in Ar (25 Torr CO). Layers 0 (for d only), 1, and 3 are labeled in each image for comparison. Below each image is the intensity analysis for layer 1 of each corresponding image. Each peak corresponds to an atomic column along the zone axis. Simulated STEM-HAADF image based on layers 0-6 of the {100} facets of the e, clean and f, CO saturated 9.2 nm Wulff constructions, along with the corresponding intensity analysis of layer 1 for each particle model.

(those used in our experiments) on particle shape. This is justified further below by the good agreement in the shapes of Pt particles observed by TEM in the considered catalysts and the predicted shapes by Wulff construction.Coverage dependent CO adsorption energies were calculated and used to obtain coverage dependent interfacial free energies for each facet as shown in Figure 1b and Tables S3-S4. CO coverage was gradually increased (to above 1 monolayer (ML) for some facets) to find the expected CO coverage at saturation on each facet. 1 ML coverage was defined as the state where all atoms with the lowest coordination number (most stable adsorption sites) on a given facet are covered by CO molecules. The configuration of surface atoms of higher index facets allows adsorption of additional CO molecules on higher coordinated sites, thus resulting in coverage greater than 1 ML in the context of our definition. At low coverage, exothermic CO adsorption stabilizes the interfacial energy of all Pt facets. With increasing coverage, CO adsorption energies decrease, and ultimately increasing coverage causes increased interfacial free energy. The minimum

interfacial free energy for each surface facet was considered representative of the interfacial free energy at saturation CO coverage, as is typically found at partial pressures ~10 Torr CO, where our experiments are operated.18,36 Fig 1c shows the configuration of CO at the saturation coverage for all facets investigated here. The calculated relative surface energies, and coverage dependent adsorption energies are consistent with both experiments and theory (see Tables S1-S4).8,37–39 Pt {111} was calculated to be the most stable facet in the bare and saturation CO coverage cases. However, the surface free energy of the second most stable bare facet, {100}, becomes comparable to the surface free energy of stepped vicinal surfaces, {210} and {310}, under saturation CO coverage, suggesting that there is a thermodynamic driving force for reconstruction of {100} facets at saturation CO coverage to induce the formation of more UC Pt surface atoms.18,36 The reconstruction of {100} facets into {211} and {311} facets (which also show comparable surface energy to {100} at saturation coverage) is not expected on well-defined particles because of the large difference in


Page 5 of 11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the orientation of these facets in the bulk atomic structure. Figure 2a shows the predicted Wulff construction of an adsorbate-free 9.2 nm Pt particle tilted slightly off the zone axis, where it is seen that atomically flat {111} and {100} facets terminate the particle, forming a truncated octahedron shape. At saturation CO coverage, an additional atomic layer on top of the {100} facet is predicted to form (labeled layer 0), as shown in Figure 2b. From direct comparison of the two equilibrium particle models of similar size, it is apparent that the newly formed layer is created through the migration of Pt atoms from lower {100} layers, although the dynamics of this migration is not directly observed with the current theoretical techniques. The DFT calculations predict that the terminating {100} facets on bare Pt particles will reconstruct to vicinal stepped {210} and {310} surfaces at saturation CO coverage, whereas {111} facets remain intact and the overall particle shape remains largely unchanged, albeit with slightly “rounded” edges.14,15 To directly image the DFT predicted CO induced facet selective Pt surface reconstruction, in-situ STEM experiments were performed using a gas cell system which allows for dynamic observation of nanomaterials heated in reactive gases at atmospheric pressure and atomic resolution.40–42 Figure S1 shows a typical (HAADF) STEM image of a ~3 nm diameter Pt nanoparticle at 423 K in an inert environment consisting of 500 Torr N2. STEM characterization shows a truncated octahedral morphology of the Pt nanoparticle, in excellent agreement with the DFT-based Wulff construction predictions of an adsorbate free Pt particle of the same size (Fig. S1). Similar geometries were observed generally for other particles in this sample. This justifies the assumption of neglecting the weakly interacting supports in the Wulff construction models. Figure 2c shows a HAADF-STEM image of a ~9 nm diameter Pt nanoparticle at 423 K in 500 Torr of N2 with atomically smooth {100} and {111} surfaces, viewed along the zone axis. The flat {100} and {111} surfaces were stable under N2 atmosphere at 423 K, with no morphological changes identified from images during an elapsed time of 30 minutes. See SI, Methods and Figure S2 for a discussion on how beam effects were minimized during imaging. N2 was pumped out of the cell and 500 Torr of a 5% CO environment was introduced into the gas cell for ~10 minutes while the temperature was maintained at 423 K, where the Pt nanoparticle surface is saturated with CO.36 We note that no leaks were detected during the duration of the experiment, such that at 25 Torr CO used in the experiment, saturation CO coverage on the Pt nanoparticle surfaces was ensured. It has been shown that the surface reconstruction process is fast at this temperature in analogous conditions.7,14 Figure 2d shows a HAADF-STEM image of the same nanoparticle as in Figure 2c after exposure to the CO environment for 10 minutes. For a clear comparison of structural differences induced by CO exposure, layers 1 and 3 were marked as a reference in both images. Upon CO adsorption, a new atomic plane of Pt atoms (marked as layer 0) formed on top of the original {100} surface, while the {111} facets re-

mained unchanged, which is in good agreement with DFT predicted Wulff constructions shown in Figures 2a and b and the simulated HAADF-STEM images derived from the Wulff constructions, Figures 2e and f respectively. Larger view images of this transformation are shown in Figure S4. Analysis of the relative intensity of each atomic Pt column (proportional to the number of Pt atoms in each column) in layer 1 is shown below each corresponding image (Figs. 2c and d). The relative intensity of each atomic column when moving away from the particle center decreases upon CO adsorption, suggesting that layer 0 was formed from migration of Pt atoms that were near the edges of layer 1 in the adsorbate-free conditions. The simulated HAADF-STEM images in Figures 2e and f were used to perform similar spectral intensity analysis. The analysis of layer 1 in the simulated images of the bare and CO saturated Pt particle, shown below Figures 2e and f, exhibit the same loss of intensity near the edges of the layer upon CO adsorption as observed in the experimental images. Since the intensity of HAADF-STEM images is proportional to the number of atoms in each atomic column, the loss of the atoms in Layer 1 following CO adsorption can be quantified through the change in integrated intensity of the atomic columns. Layer 1 in the experimental images lost 26.9% of total intensity, corresponding to an expected loss in the same percentage of atoms in this layer. The DFT predicted Wulff construction, and thereby the STEM image simulation, showed a 30% loss in atoms located in Layer 1 following CO adsorption in almost quantitative agreement with the experimental results. A similar comparative analysis was performed on layer 0, although the results suggest that layer 0 in the experimental measurements is not structured identically to the Wulff construction model (Fig. S3). This can be accounted for based on the expected fluxional geometric structure of this outermost layer. Based on previous IR and TEM analysis it was inferred that the Pt surface structure would reversibly reconstruct to the original state when CO desorbed from the Pt surface in response to increased temperature, above ~500 K.7,14 To observe the reversibility of the CO induced Pt surface reconstruction, the cell temperature was increased to 573 K in the CO environment, inducing CO desorption. After 10 minutes at 573 K, the loss of layer 0 was observed, and essentially the Pt particle reverted to the original shape shown in Figure 2c, (Fig. S4). This is in good agreement with previous reports of reversible CO induced reconstruction of Pt surfaces. Despite having direct observation at the atomic scale of the {100} facet selective CO induced Pt nanoparticle surface reconstruction and excellent agreement with DFT predictions, STEM analysis provides an extremely small sample size (only 1 in this case), and the Wulff construction only represents the thermodynamically most stable particle structure. This raises questions about the universality of the proposed reconstruction across an entire catalyst sample. Recently, we showed that quantitative IR spectroscopy could provide sample-averaged measure-



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ments of the fraction of total exposed Pt atoms existing as either WC or UC Pt sites. This is achieved by measuring IR spectra of CO saturated Pt nanoparticle surfaces, integrating relative peak areas associated with site-specific CO vibrational stretching modes, and normalizing each area by mode-specific attenuation coefficients (Fig. S5).7,43 To quantitatively measure CO induced reconstruction of Pt nanoparticles via IR spectroscopy, Pt/α-Al2O3 cata-

Page 6 of 11

lysts with varied average Pt particle diameter (1.8 ± 0.3 nm, 8.1 ± 6.2 nm and 17 ± 9 nm, Fig. S6) were synthesized (see SI).7,35 Pt catalysts were reduced in an in-situ IR cell and exposed to 1% CO/He (7.5 Torr CO) at 298 K and atmospheric pressure, resulting in saturation CO coverage (Fig. 3). The observed stretching mode at 2085-2098 cm-1 was assigned to the collective vibration of linearly bound CO adsorbed on WC Pt

Figure 3. Quantitative correlation between IR measurements and DFT predictions of CO-induced Pt nanoparticle surface reconstruction. In-situ IR spectra associated with a time evolution of a, a pre-reduced 17 ± 9 nm Pt/Al2O3 catalyst b, a prereduced 8.1 ± 6.2 nm Pt/Al2O3 catalyst, and c, a pre-reduced 1.8 ± 0.3 nm Pt/Al2O3 catalyst in a stream of 1% CO/He at room temperature (black) followed by a fast temperature ramp (1-2 minutes) to 363K (colors) which was maintained for 30 minutes. All spectra are non-normalized and are presented in KM units. The inset in a shows example linear adsorption geometries of CO on a Pt nanoparticle on WC (green) and UC (blue) atoms, and the assigned vibrational frequencies (vCO). Additional CO molecules are excluded in this illustration for clarity. d, The change (Δ) in fraction of Pt surface atoms existing as UC Pt atoms, caused by CO induced reconstruction, as a function of particle size calculated from the DFT based Wulff construction models, and measured by IR for the three considered catalysts.

sites, and the stretching mode at 2060-2078 cm-1 was assigned to the collective vibration of linearly bound CO adsorbed on UC Pt sites (Fig. 3a inset).44,45 It has been shown that the vibrational frequency of various stretching modes is related to the charge transfer between the adsorbate and the metal, where CO adsorbed on UC Pt sites has a larger amount of charge transfer and thus a more red-shifted vibrational frequency.44,46 In addition, we cannot distinguish between CO stretching modes on Pt {111} and Pt {100} due to their very similar frequencies. As a

result we describe both stretching modes as CO bound to WC Pt sites, and cannot directly differentiate changes in CO bound to (100) versus (111) sites. 44,47 The fraction of exposed Pt atoms existing as WC and UC sites at room temperature was quantified and compared to the expected fractions from DFT based Wulff construction models of bare Pt particles of the same size, in addition to expected fractions from TEM inferred particle sizes coupled with geometric models (Fig. S7 and Table S7).


Page 7 of 11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


It is important to note that fraction of surface Pt atoms existing as UC and WC sites measured from quantitative IR, as well as UC and WC site fractions based on model predictions using an average TEM particle size may be impacted by broad particle size distributions, because of the nonlinear dependence of site fraction on particle size (Fig. S7C). However, quantitative IR allows for a sampleaveraged measurement of WC and UC site fractions in the reactor and provides a more accurate description of the site distribution within the material than an averaged TEM particle size. The excellent agreement between WC and UC site fractions inferred from CO probe molecule IR measurements at 298 K and model predictions of the bare Pt particles from the Wulff construction justifies that the approach provides a quantitative measure of the surface structure of bare Pt particles, and more generally justifies the quantitative nature of the approach.7 The quantitative agreement of CO probe molecule IR measurements at 298 K with predictions of bare Pt surface structure is consistent with the 0.4-0.5 eV kinetic barrier previously observed for CO induced reconstruction of a Pt{100} single crystal.12 Thus, quantitative measurements of the WC and UC site concentrations on bare Pt nanoparticle surfaces can be inferred from IR measurements made at saturation CO coverage and room temperature, where the surfaces are kinetically trapped from reconstruction. Upon heating catalysts to 363 K in the CO environment (while maintaining saturation CO coverage), the stretching mode corresponding with CO adsorbed on UC Pt sites becomes more defined and grows in intensity, ultimately reaching steady state after 10-20 minutes (Figs. 3a, 3b, 3c). The increased intensity at elevated temperature of the CO stretching mode associated with adsorption on UC sites is direct evidence of CO induced Pt surface reconstruction. The difference in fraction of WC and UC Pt sites from quantitative IR measurements at 298 and 363 K provides a measure of the amount of reconstruction induced by saturation CO coverage. To substantiate the {100} facet selective Pt surface reconstruction inferred from DFT calculations and in-situ STEM measurements, IR measurements of the amount of Pt surface reconstruction (change in fraction of total Pt exposed existing as UC Pt atoms) were compared to predictions from the adsorbate-free and CO covered DFT-based Wulff construction models as a function of particle size (Figs. 3d and S7). The IR measurements and predictions based on Wulff construction models show quantitative agreement with an increasing amount of reconstruction observed as a function of particle size, from ~2% for 2 nm particles, to ~10% for ~15 nm Pt particles. The variation in the calculated change in UC Pt atom fraction from DFT based Wulff constructions as a function of size is expected, as the calculation is strongly dependent of the number of atoms in the outer shell of the particle, and is not an analytical function. This is due to the lack of a complete outer shell expected for all sizes that are not closed shell structures. The excellent agreement between sample averaged quantitative IR measurements, predictions from DFT-based Wulff constructions, and in-situ STEM measurements, is strong evidence that

adsorption of CO at saturation coverage on Pt nanoparticle catalysts induces the facet selective reconstruction of {100} facets into higher index stepped facets and that this behavior is representative of the whole catalyst sample. It is important to point out what conclusions can be drawn from the correlated analysis executed here. The quantitative IR analysis provides evidence that at saturation CO coverage, Pt nanoparticle surfaces undergo a facet selective reconstruction where {100} facets significantly roughen thereby increasing the fraction of total exposed Pt sites that are UC. If {111} facets were reconstructing as well, then the changes in IR spectra at elevated temperature would have been more significant (should be much greater than the ~10% change we observed) due to the very large fraction of total Pt atoms exposed in {111} facets. However, it could be possible that the CO-induced {100} facet selective reconstruction take other geometric forms than the one shown in Figure 2, as it is possible other reconstructed states of the {100} facet with increased UC site concentrations could be close to energetically degenerate. It is interesting to consider that previous studies of CO induced reconstruction of polycrystalline Pt surfaces show that at higher temperatures and different environmental conditions, as explored here, it was observed that the (111) surface reconstructs.14,48,49 This is suggestive that the specific nature of adsorbate induced reconstruction on metal nanoparticle surfaces is directly related to the coverage of different adsorbates, which control relative surface energies. Future studies will look to directly relate the composition and coverage of various adsorbates as a function of environmental conditions and temperature to the induced reconstruction. Based on our results, we argue that combining DFTbased predictions of adsorbate induced surface reconstruction with atom resolved imaging via in-situ TEM at saturation coverage and quantitative, sample averaged insitu IR spectroscopy measurements provides a unique multifaceted approach for elucidating catalytic structures under reaction conditions. When combined, these techniques allow for a sample-averaged method of observing technical catalysts under realistic operating conditions at the atomic level, while characterizing a 3-dimensional geometric structure of catalytic materials under various conditions. This approach is expected to be useful for identifying the exposed surface structure on various supported metal catalysts under reactions conditions, and have impacts on active site elucidation for structurally dynamic catalytic systems, as we have shown previously that CO induced reconstruction of Pt surfaces masks the inherent and expected structure sensitivity of the CO oxidation reaction.7 We note finally that the theoretical approaches used here to predicting the adsorbate induced reconstruction provide only thermodynamic energy minimums of the reconstruction and don’t provide information about the time resolved process of reconstruction or the potential existence of other reconstructions with degenerate energies. Recent advances in ab initio molecular dynamics calculations are likely to provide important



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

and interesting details about how the reconstruction proceeds, linking to the two states (initial and reconstructed) that we have observed.50–52 4.

Conclusions

In conclusion, we demonstrated a correlated approach towards elucidating the process of adsorbate-induced reconstruction of metal particle surfaces in high surface area catalysts. The approach relies on DFT based Wulff constructions to correlate atom resolved measurements made by in-situ STEM and sample averaged, quantitative measurements made by in-situ IR. It was demonstrated that at saturation CO coverage, Pt nanoparticle surfaces undergo a facet selective reconstruction of {100} facets into vicinal stepped surfaces with high concentrations of UC Pt atoms. Based on the level of detail obtained in these studies, and their relationship to understanding catalytic properties, it is expected that this correlated approach will be useful for elucidating the operating structures of other heterogeneous catalytic materials.


Page 8 of 11

Page 9 of 11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


ASSOCIATED CONTENT Supporting Information. Details of DFT calculations, Wulff constructions, in addition to experimental details for STEM and IR experiments. This material is available free of charge via the Internet at http://pubs.acs.org.

(15)

(16) (17)

AUTHOR INFORMATION Corresponding Author

(18)

* Correspondence to: [email protected], [email protected]

Author Contributions ‡

These authors contributed equally.

Notes

(19) (20) (21)

The authors declare no competing financial interests.

(22) (23)

ACKNOWLEDGMENT

(24)

P.C. acknowledges funding from University of California, Riverside, and U.S. Army Research Office through the YIP program, grant no. W911NF-14-1-0347. G.W.G. and X.P. acknowledge funding from the National Science Foundation (NSF) Grants No. CBET-1159240 and No. DMR-0723032. DFT calculations were performed using the Extreme Science and Engineering Discovery Environment (XSEDE), which is supported by NSF grant number OCI-1053575. Particle size measurements for the Pt/Al2O3 catalysts were performed at the UCR Central Facility for Advanced Microscopy and Microanalysis (CFAMM) and Irvine Materials Research Institute (IMRI).

REFERENCES (1) (2) (3)

(4)

(5) (6)

(7) (8)

(9)

(10)

(11) (12) (13) (14)

(25) (26)

(27) (28)

Hansen, P. L.; Wagner, J. B.; Helveg, S.; Rostrup-Nielsen, J. R.; Clausen, B. S.; Topsøe, H. Science 2002, 295, 2053. Nolte, P.; Stierle, A.; Jin-Phillipp, N. Y.; Kasper, N.; Schulli, T. U.; Dosch, H. Science 2008, 321 (September), 1654. Tao, F.; Nguyen, L.; Zhang, S.; Li, Y.; Tang, Y.; Zhang, L.; Frenkel, A. I.; Xia, Y.; Salmeron, M. B. Nano Lett. 2016, 16, 5001. Honkala, K.; Hellman, A.; Remediakis, I. N.; Logadottir, A.; Carlsson, A.; Dahl, S.; Christensen, C. H.; Nørskov, J. K. Science 2005, 307, 555. Rupprechter, G.; Weilach, C. J. Phys. Condens. Matter 2008, 20 (18), 184019. Yoshida, H.; Kuwauchi, Y.; Jinschek, J. R.; Sun, K.; Tanaka, S.; Kohyama, M.; Shimada, S.; Haruta, M.; Takeda, S. Science 2012, 335, 317. Kale, M. J.; Christopher, P. ACS Catal. 2016, 6, 5599. Allian, A. D.; Takanabe, K.; Fujdala, K. L.; Hao, X.; Truex, T. J.; Cai, J.; Buda, C.; Neurock, M.; Iglesia, E. J. Am. Chem. Soc. 2011, 133, 4498. Tao, F.; Dag, S.; Wang, L.-W.; Liu, Z.; Butcher, D. R.; Bluhm, H.; Salmeron, M.; Somorjai, G. A. Science 2010, 327, 850. Thostrup, P.; Christoffersen, E.; Lorensen, H. T.; Jacobsen, K. W.; Besenbacher, F.; Nørskov, J. K. Phys. Rev. Lett. 2001, 87 (12), 126102. Tao, F.; Dag, S.; Wang, L.-W.; Liu, Z.; Butcher, D. R.; Salmeron, M.; Somorjai, G. A. Nano Lett. 2009, 9 (5), 2167. van Beurden, P.; Bunnik, B. S.; Kramer, G. J.; Borg, A. Phys. Rev. Lett. 2003, 90 (6), 66106. Kim, J.; Noh, M. C.; Doh, W. H.; Park, J. Y. J. Am. Chem. Soc. 2016, 138, 1110. Vendelbo, S. B.; Elkjær, C. F.; Falsig, H.; Puspitasari, I.; Dona, P.; Mele, L.; Morana, B.; Nelissen, B. J.; van Rijn, R.;

(29) (30) (31) (32)

(33)

(34)

(35) (36) (37) (38) (39) (40)

(41)

(42) (43)

Creemer, J. F.; Kooyman, P. J.; Helveg, S. Nat. Mater. 2014, 13, 884. Yoshida, H.; Matsuura, K.; Kuwauchi, Y.; Kohno, H.; Shimada, S.; Haruta, M.; Takeda, S. Appl. Phys. Express 2011, 4, 65001. Harris, P. J. F. Nature 1986, 323 (6091), 792. Gontard, L. C.; Chang, L. Y.; Hetherington, C. J. D.; Kirkland, A. I.; Ozkaya, D.; Dunin-Borkowski, R. E. Angew. Chem. Int. Ed. 2007, 46 (20), 3683. Longwitz, S. R.; Schnadt, J.; Vestergaard, E. K.; Vang, R. T.; Lægsgaard, E.; Stensgaard, I.; Brune, H.; Besenbacher, F. J. Phys. Chem. B 2004, 108 (38), 14497. Tao, F.; Crozier, P. A. Chem. Rev. 2016, 116, 3487. Jinschek, J. R. Chem. Commun. 2014, 50, 2696. Elsen, A.; Jung, U.; Vila, F.; Li, Y.; Safonova, O. V.; Thomas, R.; Tromp, M.; Rehr, J. J.; Nuzzo, R. G.; Frenkel, A. I. J. Phys. Chem. C 2015, 119, 25615. Singh, J.; van Bokhoven, J. A. Catal. Today 2010, 155, 199. Boubnov, A.; Gänzler, A.; Conrad, S.; Casapu, M.; Grunwaldt, J.-D. Top. Catal. 2013, 56, 333. Gänzler, A. M.; Casapu, M.; Boubnov, A.; Müller, O.; Conrad, S.; Lichtenberg, H.; Frahm, R.; Grunwaldt, J.-D. J. Catal. 2015, 328, 216. Kale, M. J.; Gidcumb, D.; Gulian, F. J.; Miller, S. P.; Clark, C. H.; Christopher, P. Appl. Catal., B 2017, 203, 533. Enkovaara, J.; Rostgaard, C.; Mortensen, J. J.; Chen, J.; Dułak, M.; Ferrighi, L.; Gavnholt, J.; Glinsvad, C.; Haikola, V.; Hansen, H. A.; Kristoffersen, H. H.; Kuisma, M.; Larsen, A. H.; Lehtovaara, L.; Ljungberg, M.; Lopez-Acevedo, O.; Moses, P. G.; Ojanen, J.; Olsen, T.; Petzold, V.; Romero, N. A.; Stausholm-Møller, J.; Strange, M.; Tritsaris, G. A.; Vanin, M.; Walter, M.; Hammer, B.; Häkkinen, H.; Madsen, G. K. H.; Nieminen, R. M.; Nørskov, J. K.; Puska, M.; Rantala, T. T.; Schiøtz, J.; Thygesen, K. S.; Jacobsen, K. W. J. Phys. Condens. Matter 2010, 22 (25), 253202. Hammer, B.; Hansen, L.; Nørskov, J. K. Phys. Rev. B 1999, 59 (11), 7413. Li, L.; Larsen, A. H.; Romero, N. A.; Morozov, V. A.; Glinsvad, C.; Abild-Pedersen, F.; Greeley, J.; Jacobsen, K. W.; Nørskov, J. K. J. Phys. Chem. Lett. 2013, 4 (1), 222. Monkhorst, H. J.; Pack, J. D. Phys. Rev. B 1976, 13 (12), 5188. Wulff, G. Z. Kryst. 1901, 34, 449. Barmparis, G. D.; Remediakis, I. N. Phys. Rev. B 2012, 86 (8), 1. Hansen, K.; Worren, T.; Stempel, S.; Lægsgaard, E.; Bäumer, M.; Freund, H.-J.; Besenbacher, F.; Stensgaard, I. Phys. Rev. Lett. 1999, 83, 4120. Chi, M.; Wang, C.; Lei, Y.; Wang, G.; Li, D.; More, K. L.; Lupini, A.; Allard, L. F.; Markovic, N. M.; Stamenkovic, V. R. Nat. Commun. 2015, 6, 8925. Koch, C. T. Determination of core structure periodicity and point defect density along dislocations, Arizona State University, 2002. Kale, M. J.; Avanesian, T.; Xin, H.; Yan, J.; Christopher, P. Nano Lett. 2014, 14, 5405. Bourane, A.; Bianchi, D. J. Catal. 2003, 218, 447. Vitos, L.; Ruban, A. V.; Skriver, H. L.; Kollár, J. Surf. Sci. 1998, 411, 186. Wen, Y.-N.; Zhang, J.-M. Solid State Commun. 2007, 144, 163. Yeo, Y. Y.; Vattuone, L.; King, D. A. J. Chem. Phys. 1997, 106 (1), 392. Zhang, S.; Plessow, P. N.; Willis, J. J.; Dai, S.; Xu, M.; Graham, G. W.; Cargnello, M.; Abild-Pedersen, F.; Pan, X. Nano Lett. 2016, 16 (7), 4528. Matsubu, J. C.; Zhang, S.; DeRita, L.; Marinkovic, N. S.; Chen, J. G.; Graham, G. W.; Pan, X.; Christopher, P. Nat. Chem. 2017, 9, 120. Dai, S.; Zhang, S.; Katz, M. B.; Graham, G. W.; Pan, X. ACS Catal. 2017, 7, 1579. Matsubu, J. C.; Yang, V. N.; Christopher, P. J. Am. Chem.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(44) (45) (46) (47) (48) (49) (50)

Soc. 2015, 137, 3076. Greenler, R. G.; Brandt, R. K. Colloids Surf., A 1995, 105, 19. Xu, J.; Yates, J. T. Surf. Sci. 1995, 327, 193. Blyholder, G. J. Phys. Chem. 1964, 68 (10), 2772. Crossley, A.; A. King, D. Surf. Sci. 1980, 95 (1), 131. Flytzani-Stephanopoulos, M.; Schmidt, L. D. Prog. Surf. Sci. 1979, 9 (3), 83. García-Diéguez, M.; Iglesia, E. J. Catal. 2013, 301, 198. Wang, Y. G.; Yoon, Y.; Glezakou, V. A.; Li, J.; Rousseau, R. J.

(51) (52)

Page 10 of 11

Am. Chem. Soc. 2013, 135 (29), 10673. Wang, Y.-G.; Mei, D.; Glezakou, V.-A.; Li, J.; Rousseau, R. Nat. Commun. 2015, 6, 6511. Xu, C.-Q.; Lee, M.-S.; Wang, Y.-G.; Cantu, D. C.; Li, J.; Glezakou, V.-A.; Rousseau, R. ACS Nano 2017, 11, 1649.


Page 11 of 11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


TOC Image:


11

Quantitative and Atomic-Scale View of CO-Induced ... - ACS Publications

Recommend Documents