Subscriber access provided by GRIFFITH UNIVERSITY
Article
Towards rational design of Cu/SSZ-13 selective catalytic reduction catalysts: Implications from atomic-level understanding of hydrothermal stability James Song, Yilin Wang, Eric D Walter, Nancy M Washton, Donghai Mei, Libor Kovarik, Mark H Engelhard, Sebastian Prodinger, Yong Wang, Charles H. F. Peden, and Feng Gao ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.7b03020 • Publication Date (Web): 24 Oct 2017 Downloaded from http://pubs.acs.org on October 24, 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.
ACS Catalysis 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 16
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
ACS Catalysis
Towards rational design of Cu/SSZ-13 selective catalytic reduction catalysts: Implications from atomic-level understanding of hydrothermal stability James Song,a,c Yilin Wang,a Eric D. Walter,b Nancy M. Washton,b Donghai Mei,a,* Libor Kovarik,b Mark H. Engelhard, b Sebastian Prodinger, a Yong Wang,a,c Charles H.F. Peden,a Feng Gaoa,* a
Institute for Integrated Catalysis, Pacific Northwest National Laboratory, P.O. Box 999, Richland, WA 99354, United States b Environmental Molecular Sciences Laboratory, Pacific Northwest National Laboratory, P.O. Box 999, Richland, WA 99354, United States c The
Gene & Linda Voiland School of Chemical Engineering and Bioengineering, Washington State University, P.O. Box 646515, Pullman, WA 99164, United States
ABSTRACT: The hydrothermal stability of Cu/SSZ-13 SCR catalysts has been extensively studied, yet atomic level understanding of changes to the zeolite support and the Cu active sites during hydrothermal aging are still lacking. In this work, via the utilization of spectroscopic methods including solid-state 27Al and 29Si NMR, EPR, DRIFTS, and XPS, together with imaging and elemental mapping using STEM, detailed kinetic analyses, and theoretical calculations with DFT, various Cu species, including two types of isolated active sites and CuOx clusters, were precisely quantified for samples hydrothermally aged under varying conditions. This quantification convincingly confirms the exceptional hydrothermal stability of isolated Cu2+-2Z sites, and the gradual conversion of [Cu(OH)]+-Z to CuOx clusters with increasing aging severity. This stability difference is rationalized from the hydrolysis activation barrier difference between the two isolated sites via DFT. Discussions are provided on the nature of the CuOx clusters, and their possible detrimental roles on catalyst stability. Finally, a few rational design principles for Cu/SSZ-13 are derived rigorously from the atomic-level understanding of this catalyst obtained here.
KEYWORDS: selective catalytic reduction, Cu/SSZ-13, hydrothermal aging, EPR, NMR, TEM, DRIFTS. 1. INTRODUCTION The launching of commercial NOx storage-reduction (NSR) and ammonia selective catalytic reduction (NH3SCR) catalysts in the past decade or so heralded the arrival of a new era in catalytic exhaust abatement for the transportation industry; an era driven by a need for increasing fuel efficiency and decreasing carbon footprint for sustainable development.1-3 Lean fuel combustion at low temperatures, an essential option for maximizing fuel efficiency, poses two grand challenges for exhaust aftertreatment: (1) NOx removal under highly oxidizing conditions; and (2) removal of other harmful components in the exhausts at ever decreasing temperatures. Even though the latter can be partially overcome by vehicle design to efficiently utilize exhaust thermal energy, e.g., via the use of close-coupled catalytic convertors, practical solutions must ultimately rely on the development of active (low-temperature), selective, and robust catalysts.
Cu/SSZ-13 SCR materials have been generally recognized as the best lean NOx abatement catalyst on the market, with significant advantages in its superb activity and robustness over NSR and other SCR catalysts.2, 4, 5 The development and refinement of Cu/SSZ-13 SCR catalysts has witnessed rapid progress in recent years. Efforts have been made in varying Cu loading,6-9 Si/Al ratio,10 zeolite particle size/shape,11 and the addition of cocatalysts12 to promote its catalytic performance and robustness, as well as fundamental investigations regarding the nature of active centers and detailed SCR mechanisms.2, 4, 8, 9, 13-29 Novel routes for SSZ-13 and Cu/SSZ-13 synthesis, including one-pot methods, have also been extensively investigated.30-34 During the initial decade of research and development, two major degradation mechanisms have been identified for this catalyst: sulfur poisoning and hydrothermal aging.27, 35 Sulfur poisoning is typically reversible; adsorbed sulfur (mainly sulfate) can be removed during active hightemperature regeneration treatments. In contrast, hy1
ACS Paragon Plus Environment
ACS Catalysis
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
drothermal degradation is typically irreversible due to permanent structural changes such as the loss of Brønsted acid sites, and changes to the chemical or physical structure of Cu active centers. Therefore, improving hydrothermal stability is of vital importance for designing and manufacturing newer generation Cu/SSZ13 catalysts. As such, hydrothermal stability of Cu/SSZ13 catalysts, unsurprisingly, has been extensively studied over the past years.5, 36-40 However, an atomic level understanding of Cu/SSZ-13 hydrothermal degradation mechanisms are still lacking; this will be the aim of the current investigation. Coincident with fundamental research undertaken over the last decade, commercial catalysts also evolved as new information became available. The first commercial Cu/SSZ-13 catalyst, developed by the BASF Corporation, had a Si/Al ratio of 17.5 and a Cu loading of 2.8%, corresponding to a 100% ion exchange level, where ion exchange level is defined as × 200%.36 More recent state-of-the-art commercial catalysts, on the other hand, contain a lower Si/Al ratio of ~10 and an ion exchange level of ~60%.27, 41 Such changes in catalyst formulation may be a result of performance optimization based on trial-and-error; in the present study we show that a number of rational design principles for the Cu/SSZ-13 SCR catalyst can now be derived from a detailed atomic-level understanding of the nature of the active centers, and the aforementioned formulation changes can be understood within the framework of these rational design principles. 2. EXPERIMENTAL AND THEORETICAL METHODS 2.1. Catalyst Preparation SSZ-13 materials (Si/Al = 6, 12, and 35) were synthesized hydrothermally using a method described in detail elsewhere.10 To prepare Cu/SSZ-13 catalysts, the SSZ-13 materials were first changed into NH4+ form, and then to Cu2+ form via aqueous solution ion-exchange methods. Cu loading was readily controlled by varying concentrations of the CuSO4 solutions used for exchange.9 To determine Cu, Al, and Si contents of the catalysts, elemental analysis with inductively coupled plasma atomic emission spectroscopy (ICP-AES) was performed at Galbraith Laboratories (Knoxville, TN, USA). The majority of the present study was conducted on a catalyst with Si/Al = 12 and a Cu loading of 2.1% (ion exchange level ~55%), with a composition close to the current industrial state-of-the-art catalyst. Portions of the fresh catalyst were hydrothermally aged in air containing 10% water vapor at 550, 600, 650, 700, 750, and 800 °C for 16 h, and at 900 °C for 2 h. These are denoted as HTA samples (for example, HTA-600 represents the sample aged at 600 °C). A few other samples used in our previous studies, i.e., Cu loaded Si/Al = 6
Page 2 of 16
and 35 samples, were also used here.9, 10 These will be described in more detail below. 2.2. Catalyst Characterization BET surface areas and micropore volumes of the fresh and aged samples were measured with a Quantachrome Autosorb-6 analyzer, with micropore volumes determined using the t-plot method. Prior to analysis, the samples were dehydrated under vacuum overnight at 250 °C. For selected samples, more accurate analyses were conducted on a Micromeritics ASAP 2000 instrument with Ar as the probe molecule. Powder X-ray diffraction (XRD) measurements were performed on a Philips PW3040/00 X’Pert powder X-ray diffractometer with Cu Kα radiation (λ = 1.5406 Å). Data were collected with 2θ ranging from 5º to 50º using a step size of 0.02º. Temperature-programmed reduction (TPR) experiments were performed on a Micromeritics AutoChem II analyzer. TPR was carried out on ambient samples (i.e., samples stored in typical open laboratory environment and saturated with moisture) and, typically, ~50 mg of sample was used for each measurement. TPR was carried out in 5% H2/Ar at a flow rate of 30 cm3/min. The temperature was ramped linearly from ambient to 700 °C at 10 °C/min and H2 consumption was monitored with a TCD detector. 27Al
solid-state MAS-NMR analyses of the samples were conducted on a 20 Tesla Agilent VNMRS system utilizing a 1.6 mm HXY probe in direct polarization (DR) mode tuned to 221.4119 MHz. A calibrated π/20 pulse of 0.30 µs with a recycle delay of 0.5 s was used to collect 32,000 transients. Samples were spun at 34 kHz at 20 °C. Time domain free induction decays were apodized with 150 Hz of Lorentzian broadening after zero filling twice. 27Al chemical shifts are reported relative to a 0.1 M aqueous AlCl3 solution. 29Si experiments were conducted on a 14.0 Tesla Agilent VNMRs system utilizing a 5.0 mm HXY prove in DR mode tuned to 119.1352 MHz. A calibrated π/4 pulse of 4.75 µs with a 90 s recycle delay was used to collect 160 transients. Samples were spun at 12 kHz at 20 °C. Time domain free induction decays were apodized with 250 Hz of Lorentzian broadening after zero filling twice. 29Si chemical shifts are reported relative to a Si(CH3)4 standard. Electron paramagnetic resonance (EPR) experiments were conducted on a Bruker E580 X-band spectrometer equipped with a SHQE resonator and a continuous flow cryostat.8 Powder samples (∼10 mg) were contained in 4 mm OD quartz tubes. Microwave power was 20 milliwatts, and the frequency was 9.34 GHz. The field was swept by 6000 G in 84 s, and modulated at 100 kHz with 10 G amplitude. A time constant of 41 ms was used. For hydrated samples, measurements were performed at -150 °C to “freeze” Cu(II) ions. This is neces2
ACS Paragon Plus Environment
Page 3 of 16
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
ACS Catalysis
sary since antiferromagnetic interactions between mobile Cu(II) ions can cause signal loss. For samples dehydrated by flowing dry He through the samples at 250 °C for 60 min, measurements were done after cooling the samples back to -150 °C. Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS) measurements were acquired on a Nicolet 6700 FT-IR spectrometer with an integrated Praying Mantis cell connected to a gas manifold system that was assembled in-house. To study Brønsted acidity of the catalysts, the OH stretching vibrational region (νOH) was examined on samples dehydrated at 550 °C for 1 h in flowing O2 and cooled back to 150 °C. Spectra were then collected against a KBr background spectrum. The nature of the Cu ions was studied by examining the TOT vibrational region of the samples using a method developed by Luo et al.27, 41 In this case, a spectrum of a dehydrated sample obtained at 100 °C was used as the background. Following which, a continuous flow of 0.5% NH3/He was introduced to the sample at this temperature, and spectra were continuously collected until they no longer changed with time. As will be shown below, the negative TOT vibrational features are representative of the nature of cationic Cu species. 2.3. SCR and NH3 Oxidation Reaction Testing Measurements for standard NH3-SCR (4NO + 4NH3 + O2 = 4N2 + 6H2O) and the major side reaction, nonselective NH3 oxidation (4NH3 + 3O2 = 2N2 + 6H2O), were carried out in a plug-flow reactor system equipped with a MKS MultiGas 2030 FTIR gas analyzer. In brief, for standard NH3-SCR activity testing, the feed gas contained 360 ppm NO, 360 ppm NH3, 14% O2, ~2.5% H2O, and balanced with N2 for a total flow rate of 1 L/min. For NH3 oxidation, NO was not included in the feed while concentrations of other gases and the total flow rate maintained. For light-off measurements, 200 mg catalyst (60-80 mesh) was used resulting in a gas hourly space velocity (GHSV) of ~200,000 h-1. To collect data for kinetics analyses, care was taken to ensure the absence of mass and heat transfer limitations by applying standard methods including varying particle size, space velocity, and Cu loading. Typically, 60 mg catalyst was used (GHSV ≈ 650,000 h-1) and conversions were kept low (< 20%). Under similar conditions, our previous study has demonstrated that the Koros-Nowak criterion is obeyed for moderate Cu-loaded catalysts, indicating the absence of mass and heat transfer limitations.9 Before reaction measurements, the samples were exposed to flowing 14% O2 + 2.5% H2O in N2 at 550 °C for 2 h. This pretreatment was necessary to stabilize the catalysts in order to obtain highly reproducible results. More details on the experimental setup and mathematical equations used for data analyses can be found elsewhere.8, 9
2.4. DFT Calculations Periodic DFT calculations were performed using the CP2K code.42 All DFT calculations employed a mixed Gaussian and planewave basis sets. Core electrons were represented with norm-conserving Goedecker-TeterHutter pseudopotentials,43-45 and the valence electron wavefunctions were expanded in a double-zeta basis set with polarization functions,46 along with an auxiliary plane wave basis set with an energy cutoff of 360 eV. The generalized gradient approximation exchangecorrelation functional of Perdew, Burke, and Enzerhof (PBE)47 was used. The SSZ-13 zeolite structure was modelled using a hexagonal unit cell (36 T atoms total) with the size parameters of 13.6750×23.6858×14.7670 Å3. To mimic Si/Al ratios of technically relevant Cu/SSZ13 catalysts, 6 Si atoms within the SSZ-13 model structure were replaced by 6 Al atoms, obtaining a model structure with the Si/Al ratio of 11. Six H atoms were also introduced on the O1 position of four O atoms connected with the Al atom to keep the structure charge neutral.48 On the basis of previous theoretical studies,22, 24, 25, 28 “naked” Cu2+ ions (i.e., ions with no extra lattice ligands) located in the window of 6-membered rings (6MR) and balanced with 2 framework negative charges (denoted as Cu2+-2Z below, with Z representing a framework negative charge), and [Cu(OH)]+ ions located in the window of 8MRs balanced with 1 framework negative charge (denoted as [Cu(OH)]+-Z) were optimized as the initial configurations of the active centers. The activation barrier (∆H≠) and Gibbs free energy of activation (∆G≠) were calculated as follows: "# %# ∆H = ∆ + ∆ + ∆ + ∆ + ! − !
(1)
∆& = ∆' − (∆)
(2)
"# %# where ( ! - ! ) represents electronic energy differences between the transition state (TS) and the initial state (IS) of elementary dealumination (i.e., framework Al removal) reaction steps. Transition states of the elementary steps were located using the climbing image nudged elastic band (CI-NEB) method,49, 50 using nine intermediate images along the reaction pathway between initial and final states. The identified transition states were then confirmed by vibrational analysis. More calculation details can be found in our previous work.29
3. RESULTS 3.1. Hydrothermal Aging Effects on Catalytic Performance Properly ion-exchanged Cu/SSZ-13 in the asprepared (“fresh”) form should only contain isolated Cu ions. In fully hydrated samples, these are present as [Cu(H2O)6]2+ and [Cu(OH)(H2O)5]+ ions located in the 3
ACS Paragon Plus Environment
ACS Catalysis Chabazite cages.28 In the dehydrated form, they convert to “naked” Cu2+ and [Cu(OH)]+ ions coordinated with zeolite lattice oxygens in 6MRs and 8MRs, respectively. More specifically, the thermodynamically most stable Cu2+ ions locate in windows of 6MRs, adopt 3 or 4 Cu-O coordination, and are balanced by paired framework charges (i.e., as Cu2+-2Z). [Cu(OH)]+ ions locate in 8MRs, adopt 3 Cu-O coordination (two framework O together with –OH), and are balanced by a single framework charge (as [Cu(OH)]+-Z).16 Remaining Brønsted acid sites are also critical for SCR because these are largely responsible for ammonia storage.51 These three important sites are schematically illustrated in Scheme 1.
tivities. The higher activity for [Cu(OH)]+-Z in catalyzing NH3 oxidation is not fully understood at this time, but seems likely to be related to either its lower redox barrier than Cu2+-2Z since it binds the zeolite framework more weakly, or to its higher tendency to dimerize, thereby forming Cu-ion dimers that are more active in O2 activation.23 Collectively, Figures 1 (a) and (b) indicate that, for catalysts containing only these two types of isolated Cu-ions, SCR selectivities at elevated temperatures (e.g., 500 °C) should be no lower than ~90%. Thus, SCR selectivity at relatively high temperatures can be used as a criterion to judge whether or not a Cu/CHA catalyst contains predominantly isolated Cu ions. Reaction Temperature (°C)
(a)
550
Scheme 1: structures of Cu2+, [Cu(OH)]+, and Brønsted acid sites in Cu/SSZ-13.
We have shown previously that, by controlling Si/Al ratios and Cu loadings, model catalysts that only contain one of the two Cu-ion active sites can be readily synthesized.10 In catalysts with low Si/Al ratios and low Cu loadings, Cu2+-2Z species predominate, while in high Si/Al samples, [Cu(OH)]+-Z ions are the primary species because paired Al sites (Al-Si-Al or Al-Si-Si-Al) are rare in this latter case.28 The presence of either of these two Cu moieties is readily identified by hydrogen temperature-programmed reduction (H2-TPR); this is based on the fact that Cu2+-2Z reduces at higher temperatures than [Cu(OH)]+-Z (Figure S1).10 Next, using model catalysts with predominantly Cu2+2Z sites (Si/Al = 6, Cu/Al = 0.016 and 0.044) or [Cu(OH)]+-Z sites (Si/Al = 35, Cu/Al = 0.1), hightemperature SCR performance is examined by comparing normalized rates (mole NH3 mole Cu-1 s-1) between standard SCR and the major side reaction, nonselective NH3 oxidation. Note that NH3 oxidation maintains low until ~350 °C for Cu/SSZ-13 catalysts; therefore, only at temperatures above ~350 °C could SCR selectivities for Cu/SSZ-13 be properly evaluated. As shown in Figure 1(a), standard SCR rates level off at temperatures above ~400 °C, indicating a transition from kinetic control to mass transfer control. Even so, SCR rates are two orders of magnitude higher than NH3 oxidation rates under our reaction conditions, indicating excellent hightemperature SCR selectivities for Cu2+-2Z sites. The same measurements performed for [Cu(OH)]+-Z sites (Figure 1(b)) reveal that SCR rates again level off above ~450 °C, and NH3 oxidation displays two distinct kinetic regimes likely due to a transition in reaction mechanism.8 For this latter catalyst, SCR rates are about one order of magnitude higher than NH3 oxidation rates at elevated temperatures, indicating ~90% SCR selec-
Normalized Rate (mole NH3 per mole Cu per s)
100
500
400
450
350
standard SCR
10-1 Si/Al = 6 Cu/Al = 0.016
10-2 NH3 oxidation
10
-3
Si/Al = 6 Cu/Al = 0.044
10-4 1.2
1.3
1.4
1.5
1.6
−1
1000/T (K ) Reaction Temperature (°C)
(b) 10
Normalized Rate (mole NH3 per mole Cu per s)
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 4 of 16
-1
550
500
450
400
350
standard SCR
10-2
Si/Al = 35 Cu/Al = 0.1
10-3
NH3 oxidation
10-4 1.2
1.3
1.4
1.5
1.6
−1
1000/T (K ) Figure 1. Comparison between normalized standard SCR and NH3 oxidation rates (mole NH3 per mole Cu per s) for model Cu/SSZ-13 catalysts that contain predominately Cu2+-2Z sites (a) and [Cu(OH)]+-Z sites (b). Reactant feed contains 350 ppm NO (for standard SCR only), 350 ppm NH3, 14% O2, 2.5% H2O balanced with N2 at a GHSV of ~400,000 h-1.
For our Si/Al = 12 with Cu loading = 2.1 wt% (from ICP) catalyst in its fresh form, Cu(II) quantification with EPR (vide infra) confirms a monomeric Cu(II) loading of 2.08 wt%, almost identical to the total Cu loading de4
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
ACS Catalysis
rived from ICP. This strongly suggests that this sample also contains predominantly isolated Cu-ions. NOx and NH3 light-off curves for standard SCR on these Cu/SSZ13 catalysts, both fresh and HTA-800, are displayed in Figure S2(a). Below ~350 °C, NOx and NH3 conversions are essentially identical demonstrating excellent SCR selectivities. At higher temperatures, NH3 conversions become higher, with the extra NH3 consumption attributable to the nonselective NH3 combustion side reaction as described above. At 500 °C, NH3 conversion is ~10% higher than that of NOx for the fresh catalyst, consistent with the dominance of isolated Cu-ions in this sample. However for the HTA-800 sample, NH3 conversions become ~20% higher than that of NOx at 500 °C. This result indicates that some new Cucontaining moieties form during hydrothermal aging, and that these new Cu species are more active than isolated Cu ions in catalyzing NH3 combustion at elevated temperatures. It is now generally agreed that the new Cu-containing moieties that catalyze NH3 combustion are CuOx clusters that form during hydrothermal aging.5, 36, 39 Figure S2(b) presents NO conversions as a funcx tion of temperature on the first-generation BASF commercial Cu/SSZ-13 SCR catalyst hydrothermally aged at 800 °C for 16 h.36 This catalyst is even less SCR selective at elevated temperatures, with NOx conversions of only ~65% at 500 °C. This indicates that the high ionexchange level adopted for this catalyst has the disadvantage of generating even more unwanted CuOx species during hydrothermal aging. This topic will be discussed further below. To gain a more detailed understanding of the hydrothermal aging effects on SCR and NH3 oxidation, these reactions were also studied under kinetic-controlled differential conditions. For standard SCR, reaction rates were normalized using the first-order kinetic equation, + i.e., * = −ln 1 − 0 , where F is the NO flow rate , (moles of NO per s), w is the mass of the catalyst (g), and x is the NO conversion.9, 52 The Arrhenius equation, 2=
[45]7
:;<
= 89 => , was then used for estimating pre-
exponential factors (A) and apparent activation energies (Ea), where k is the rate constant and [NO]0 the molar concentration of NO in the feed.52 Figure 2 presents Arrhenius plots (k vs. 1/T) for fresh and HTA Cu/SSZ13 samples (Si/Al = 12; Cu 2.1%). An expected trend exists where k decreases with increasing aging temperature (the HTA-600 sample is an outlier for unknown reasons). Using data shown in Figure 2, A and Ea were calculated and the results are given in Table 1. The apparent activation energies are typical for standard NH3SCR over Cu/SSZ-13 under kinetic control,10, 17 and the similar Ea values among the catalysts indicate similar active centers and reaction pathways. In other words, hydrothermal aging does not alter the nature of the active sites in this catalyst. In contrast, pre-exponential
factors change dramatically with hydrothermal aging. The moderately aged HTA-600/650/700 samples display pre-exponential factors two orders of magnitude lower than the fresh catalyst, while the heavily aged HTA-750/800/900 ones display prefactors three orders of magnitude lower. This indicates that effective collisions between reactants and Cu active sites decrease markedly with increasing aging temperature, perhaps as a result of the chabazite framework exhibiting stronger confinement of the Cu active sites with increasing aging temperature. This suggested explanation will be discussed further below.
Temperature (°C) 140
150
130
10-2
Rate Constant k (s-1)
Page 5 of 16
Fresh HTA-550 HTA-600 HTA-650 HTA-700 HTA-750 HTA-800 HTA-900
10-3
2.34
2.36
2.38
2.40
2.42
2.44 -1
2.46
2.48
2.50
1000/T (K ) Figure 2. Arrhenius plots for standard SCR on fresh and HTA Cu/SSZ-13 samples (Si/Al = 12, Cu 2.1%), where the rate constant, k, was calculated using NOx conversion data collected between 130 and 150 °C under differential conditions. The feed gas contained 360 ppm NO, 360 ppm NH3, 14% O2, ~2.5% H2O, and balanced with N2 at a total flow rate of 1 L/min. The GHSV was estimated to be ~650,000 h-1.
NH3 oxidation results for the fresh and HTA samples are displayed in Figure S3. Under low temperature (≤ 325 °C) and high space velocity (i.e., “differential”) conditions (Figure S3(a)), a clear trend, that activity decreases with increasing aging temperature, is observed. It is important to note that even though CuOx clusters form during hydrothermal aging, these moieties do not appear to catalyze NH3 oxidation at such low temperatures; otherwise, aged catalysts would display activities higher than the fresh catalyst. In “light-off” measurements at a lower space velocity, reaction results displayed in Figure S3(b) reveal a rather complex temperature dependence. While at reaction temperatures ≤ 300 °C the trends found in Figure S3(a) are essentially preserved; notably, HTA-800/900 samples become more active than HTA-700/750 at 400 °C and above. Because the reaction is not under kinetic control for these latter measurements, rigorous and detailed kinetics analysis is not possible. However, these results do clearly indi5
ACS Paragon Plus Environment
ACS Catalysis cate that the nature of the CuOx clusters changes dramatically with aging temperature. Possible explanations are given in Supporting Information S3.
Table 2: Micro- and meso-pore volume analysis of the fresh and HTA-700/800 samples measured with Micromeritics ASAP 2000 via Ar absorption. Sample
Table 1: Preexponential factors (A) and apparent activation energies (Ea) derived from Arrhenius plot analysis of lowtemperature standard SCR.
Total Pore Volume (cm3/g)
0.23
0.047
0.277
0.23
0.159
0.389
HTA-800
0.18
0.202
0.382
3
Pore volume (cm /g)
Pore size distribution
0.35
(Horvath Kawazoe + Saito Foley)
0.30
Parent
0.25
700 ºC HTA
0.20
800 ºC HTA
0.15 0.10 0.05 0 0
5
(b)
10 Pore width (Å)
15
20
Mesopore size distribution BJH
0.10
Parent
3
Textural properties of the fresh and HTA samples were probed with surface area/pore volume analysis (N2 adsorption on a Quantachrome AutoSorb-6) and XRD measurements. The results are displayed in S4. As evidenced from Figure S4, XRD is largely insensitive for revealing changes caused by hydrothermal aging; that is, XRD patterns for the mildly and heavily aged materials are strikingly similar. As shown in Table S1, surface areas of the catalysts also experienced only minor changes; e.g., aging up to 750 °C caused less than a 5% decrease in surface area from the fresh catalyst. However, the adsorption of N2 in the pore constraints of zeolites at -196 °C is hindered to some degree by its quadrupolar moment.53-55 For selected samples, i.e., the fresh and HTA-700/800 ones, porosity was more accurately analyzed using Ar adsorption at -186 °C on a Micromeritics ASAP 2000 to avoid the quadrupolar moment, where adsorption/desorption isotherms are shown in Figure S5. The higher partial pressure required for filling of micropores also allowed us to more accurately probe the porous environment.55 Based on these experiments, micro- and mesopore distributions are plotted in Figure 3 (a) and (b), respectively, and the corresponding pore volumes are presented in Table 2. These results demonstrate that hydrothermal aging induces no change in micropore size (Figure 3(a)); however, it does cause mesopore formation. For the HTA-800 sample in particular, the amount of mesopores that are ≤ 4nm increases significantly (Figure 3(b)). Importantly, quantification results in Table 2 clearly demonstrate that mesopore formation in the HTA-700 sample does not appear to affect micropore volume; however in the HTA-800 sample, the extra mesopore formation is clearly at the expense of micropores. Note that for zeolite materials, mesopore formation during hydrothermal aging is a common feature.56 However, the influence of Cu on mesopore formation in Cu/zeolites during hydrothermal aging, has not been well documented. Next, in order for zeolite structural changes to be more precisely determined, solid-state NMR was used to acquire 27Al and 29Si spectra of the fresh and HTA samples.
Meso Pore Volume (cm3/g)
Fresh
(a) 3.2. Textural Properties, Spectroscopy and Microscopy Investigations
Micro Pore Volume (cm3/g)
HTA-700
Pore volume (cm /g)
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 6 of 16
700 ºC HTA 0.08
800 ºC HTA
0.06
0.04
50
100 Pore width (Å)
150
200
Figure 3: (a) Micropore size distributions of the fresh, HTA700 and HTA-800 samples determined with the HorvathKawazoe and Saito-Foley methods. (b) Mesopore size distributions of the fresh, HTA-700 and HTA-800 samples determined with the BJH method.
Figure 4(a) presents 27Al spectra acquired for hydrated, ambient samples. In these spectra, tetrahedrallycoordinated framework Al (AlF) features appear at ~57.5 ppm and much smaller octahedrally-coordinated extraframework Al (EFAl) signals appear at ~0 ppm.57 Taking the AlF signal area of the fresh sample as unity, the amounts of AlF in HTA samples were normalized, and the results are displayed adjacent to each spectrum. The results demonstrate that this technique is much more sensitive than the textural characterization methods described above in revealing degradation caused by hydrothermal aging. It is worth noting that hydrothermal aging at 550 and 600 °C for 16 h results in almost no dealumination while at higher aging temperatures, 6
ACS Paragon Plus Environment
Page 7 of 16
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
ACS Catalysis
the extent of dealumination increases progressively with increasing temperature. Note also that the portion of Al detached from the zeolite framework upon dealumination does not appear at ~0 ppm but rather remains NMR silent. This phenomenon is typical for hydrothermally aged Cu/zeolites. Possible explanations are that (1) this portion of Al stays adjacent to paramagnetic Cu sites and is thus invisible to NMR, or (2) the detached Al is in highly distorted and heterogeneous sites, which broadens their lines to the point of undetectability.5, 11
aging temperatures of 650 °C and higher, the -105 ppm feature progressively declines, consistent with catalyst dealumination shown in Figure 4(a).
(a)
Hydrated Cu/SSZ-13, Si/Al = 12, Cu loading 2.1%
Fresh HTA-550 HTA-600 HTA-650 HTA-700 HTA-750 HTA-800 HTA-900
57.5
(a)
AlIV Fresh
1
HTA-550
1
HTA-600
0.97
HTA-650
0.92
HTA-700
0.87
HTA-750
0.79
HTA-800
0.70
HTA-900
0.61
100
80
60
40
20
0
-20
2400
2600
2800
3000
3200
3400
Field (G) (b) Dehydrated Cu/SSZ-13, Si/Al = 12, Cu loading 2.1%
Fresh HTA-550 HTA-600 HTA-650 HTA-700 HTA-750 HTA-800 HTA-900
-40
27
Al Chemical Shift (ppm from AlCl3 (aq)) -111
(b) -105
2850 2900 2950 3000 3050 3100
Fresh
3600
2400
HTA-550
2600
2800
3000
3200
3400
3600
Field (G)
HTA-600
Figure 5. EPR spectra of (a) hydrated, and (b) dehydrated fresh and HTA Cu/SSZ-13 catalysts. Spectra were acquired at 125 K. Sample dehydration was achieved by flowing dry He through the samples at 250 °C for 1 h. Line-shape changes for the dehydrated samples in the hyperfine region are displayed in the inset to (b).
HTA-650 HTA-700 HTA-750 HTA-800 HTA-900 -80
-90
-100
-110
-120
-130
-140
29
Si Chemical Shift (ppm from TMS)
Figure 4. (a) 27Al and (b) 29Si magic angle solid-state NMR of the fresh and HTA Cu/SSZ-13 catalysts studied here. Spectra were acquired on ambient samples.
The corresponding 29Si NMR spectra are displayed in Figure 4(b). The fresh catalyst contains two framework tetrahedral Si (Q4) features: one at -105 ppm attributed to tetrahedral Si with three Si and one Al neighbors (i.e., Si(OSi)3(OAl)) and one at -111 ppm to tetrahedral Si with four Si neighbors (i.e., Si(OSi)4).57 At hydrothermal
Next, hydrothermal aging effects to the active sites, i.e., isolated Cu ions, are described and quantified through the use of electron paramagnetic resonance (EPR). Under fully hydrated ambient conditions isolated Cu ions stay exclusively at a +2 oxidation state under this condition (i.e., as [Cu(H2O)6]2+ and [Cu(OH)(H2O)5]+ ions), and are therefore amenable to quantitative interrogations via ERP. However, peak broadening resulting from Cu mobility and antiferromagnetic interactions must be ameliorated; and this is done by conducting the experiments at cryogenic temperatures to freeze Cu mobility.8 Figure 5(a) presents EPR spectra for hydrated samples collected at -150 °C. Signal lineshapes for the samples are very similar, indicating that [Cu(H2O)6]2+ and [Cu(OH)(H2O)5]+ ions are indistin7
ACS Paragon Plus Environment
ACS Catalysis guishable under the measurement conditions. Using Cuimide chelated standard solutions as reference, these Cu ions are readily quantified, and the results are shown in Table S2. For selected samples, such quantifications were conducted more than one time. The repeated measurements suggest errors substantially lower than 5%. Note again that in the fresh sample, this quantification (2.08 wt%) matches total Cu loading quantification via ICP (2.1 wt%) very well. In the aged samples, signal intensities decrease due to conversion of EPR active Cuions to EPR silent CuOx clusters. 1.6
Cu Quantity (wt.%)
Cu2+ 1.2
CuOx
Based on their different EPR responses (i.e., Cu2+-2Z is EPR active in both hydrated and dehydrated forms, [Cu(OH)]+-Z is only EPR active in the hydrated form, and CuOx is always EPR silent), the amount of the three species in the sample series were estimated and the results are displayed in Table S2 and also plotted in Figure 6. With increasing aging temperature, [Cu(OH)]+Z content monotonically decreases while CuOx content monotonically increases. Cu2+ content first increases, reaching a highest value of ~1.5 wt.% in HTA-700, and then declines slightly before stabilizing at ~1.2 wt.% at higher aging temperatures. Several key points are worth mentioning. (1) In the fresh sample, Cu2+-2Z is not saturated even though this species is thermodynamically more stable than [Cu(OH)]+-Z. (2) At aging temperatures ≤ 700 °C, a portion of [Cu(OH)]+-Z gradually converts to CuOx clusters and a portion to Cu2+-2Z. The latter reaction can be described as follows:
0.8
[Cu(OH)]+ + H+ = Cu2+ + H2O 0.4 +
Cu(OH)
Fr es h TA -5 50 H TA -6 00 H TA -6 50 H TA -7 00 H TA -7 50 H TA -8 00 H TA -9 00
0.0
H
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 8 of 16
Figure 6. Estimation of Cu2+, Cu(OH)+ and CuOx in fresh and HTA samples.
Upon full dehydration, [Cu(H2O)6]2+ converts to “naked” Cu2+, which remains EPR active. However, [Cu(OH)(H2O)5]+ converts to “naked” [Cu(OH)]+ or even undergoes further chemical reactions; i.e., “autoreduction” to Cu+ or condensation to dimers ([Cu-OCu]2+) via dehydration.19 Significantly, all of these latter species are EPR silent due to the lack of paramagnetic electrons (for Cu+), or a pseudo Jahn-Teller effect (for [Cu(OH)]+),19 or signal broadening from antiferromagnetic interactions (for [Cu-O-Cu]2+).58 The antiferromagnetic effect also applies to CuOx clusters rendering them EPR invisible. Figure 5(b) presents EPR spectra for the dehydrated samples. In this case, signal lineshapes do change considerably upon hydrothermal aging. For example, aged samples display new features in the hyperfine region (magnified in the figure insert), likely due to strong coupling between Cu atoms in the CuOx clusters and the remaining isolated Cu ions. Detailed spectra analyses warrant an independent study and will not be addressed further here. In the present study, spectra shown in Figure 5(b) are primarily used for quantification of EPR active species, again referenced to Cu-imide standard solutions, with the results also presented in Table S2.
(3)
The former chemistry, leading to CuOx clusters, is more complex; in particular, the nuclearity and charge state for various CuOx clusters can be quite different and are still largely unknown. (3) At aging temperatures of 750 °C and above, Cu2+-2Z stabilizes at ~1.2 wt.% indicating that this portion of Cu2+ must be located in windows of 6MR with paired Al sites; i.e., at the energetically most stable locations for Cu2+.16 This indicates then that a small portion of Cu2+ in the HTA-700 sample (~0.3 wt.%) is stabilized by paired framework charges not located in the same 6MR. This amount of Cu2+ is less stable, ultimately converting to CuOx at higher aging temperatures. [Cu(OH)]+-Z conversion to Cu2+-2Z during hydrothermal aging demonstrates that [Cu(OH)]+-Z formation commences prior to Cu2+-2Z saturation for the fresh catalyst. This is somewhat unexpected since theoretical studies have concluded that Cu2+-2Z is thermodynamically more stable than [Cu(OH)]+-Z.24, 28 One likely explanation is that [Cu(OH)]+-Z is readily kinetically stabilized such that it can form prior to Cu2+-2Z saturation. During hydrothermal aging, the catalyst gains sufficient thermal energy and time to relax, leading to gradual [Cu(OH)]+-Z conversion to Cu2+-2Z. Another possible explanation for [Cu(OH)]+-Z conversion to Cu2+-2Z is that framework Al redistributes to generate extra “paired” Al sites to accommodate more Cu2+-2Z configurations during hydrothermal aging. To test this hypothesis, Co2+ titration experiments were performed; the procedures and results are collectively displayed in S7. Note that Co2+ ions selectively titrate paired Al sites.28 These titration results (Table S3) demonstrate that framework Al redistribution does not occur during hydrothermal aging, corroborating the argument that [Cu(OH)]+-Z population prior to Cu2+-2Z saturation in the fresh catalyst is due to kinetic reasons. During hy8
ACS Paragon Plus Environment
Page 9 of 16
drothermal aging, [Cu(OH)]+-Z gradually converts to the thermodynamically more stable Cu2+-2Z until the latter saturates. Kinetic stabilization of [Cu(OH)]+-Z has also been found in the current state-of-the-art SCR catalyst, suggesting that this is a general phenomenon.41
3573
3600
3650
3730
Absorbance (a.u.)
(a)
HTA-900 HTA-800 HTA-750 HTA-700 HTA-650 HTA-600 Fresh 4000
3900
3800
3700
3600
3500
3400
-1
Wavenumber (cm ) (b)
950
900
Fresh HTA-550
Absorbance (a.u.)
HTA-600
ing hydrothermal aging temperatures, intensities for both Si-O(H)-Al and Cu-OH decrease, fully consistent with 27Al NMR spectra shown in Figure 4(a) and [Cu(OH)]+-Z estimation displayed in Figure 6. It has been demonstrated recently that perturbed framework vibrations (i.e., TOT vibrations) for Cu/SSZ13 arise when Cu2+ and [Cu(OH)]+ move close to the SSZ-13 framework charged sites, for example during dehydration.7, 20 It is also known that NH3, as a strong ligand to Cu ions, can solvate and, thus, move these ions away from the framework exchange sites. This results in the elimination of the perturbed framework vibrations. Taking advantage of these phenomena, Luo et al. demonstrated that, by using a dehydrated Cu/SSZ-13 as the background for IR analysis, NH3-saturated Cu/SSZ13 displays well-revolved negative TOT bands that nicely represent the formation of solvated Cu2+ and [Cu(OH)]+ species.27 This same method was adopted here and the spectra are shown in Figure 7(b). For the fresh and HTA-550/600 samples, well-resolved negative T-O-T bands are found at ~900 and ~950 cm-1, representing the solvation of both Cu2+-2Z and [Cu(OH)]+-Z, respectively. At higher aging temperatures, the ~950 cm-1 band diminishes and becomes unresolved, whereas the ~900 cm-1 band shifts to ~920 cm-1 but the intensity maintains. These results are consistent with EPR estimations presented in Figure 6, demonstrating the remarkable hydrothermal stability of Cu2+-2Z.
HTA-650 920
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
ACS Catalysis
HTA-700 HTA-750 HTA-800 HTA-900
1200
1100
1000
900
800 -1
700
Wavenumber (cm )
Figure 7. DRIFTS spectra of the fresh and HTA Cu/SSZ-13 catalysts (Si/Al = 12 and a Cu loading of 2.1%): (a) OH vibrational region; and (b) TOT vibrational region perturbed by NH3 adsorption.
DRIFTS was used to gain further insights into hydrothermal aging effects. It was first applied to examine the OH vibrational (νOH) region of the catalysts. As shown in Figure 7(a), the fresh Cu/SSZ-13 catalyst has 4 prominent νOH features at 3730, 3650, 3600, and 3573 cm-1. The 3730 cm-1 band is attributed to Si-OH and the 3600/3573 cm-1 signals are attributed to Brønsted acid sites (i.e., Si-O(H)-Al). The 3650 cm-1 feature is largely attributable to Cu-OH, but contributions from terminal Al-OH species cannot be ruled out.13, 18, 20 With increas-
Figure 8. HAADF STEM images of a fresh Cu/SSZ-13 catalyst particle, and the corresponding elemental mapping for Cu, O, Al, Si and Al+Si. Note that all elements are uniformly distributed. More images are shown in Figure S6.
Taken together, these results convincingly demonstrate that [Cu(OH)]+-Z converts to CuOx clusters during hydrothermal aging. In contrast to the two distinct isolated Cu ions, the physical and chemical properties of CuOx clusters, e.g., their nuclearity, location, charge state, etc., are largely unknown. To gain a more thorough understanding of these latter species, scanning 9
ACS Paragon Plus Environment
ACS Catalysis
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
TEM (STEM) imaging and XPS analysis were utilized, with results discussed below. We have recently shown that FeOx clusters in hydrothermally aged Fe/SSZ-13 catalysts, measuring ~1nm in size, are readily detectable with STEM.59 In the present study, STEM was initially used to attempt imaging of CuOx clusters in the HTA samples to elucidate their location and nuclearity. However, no CuOx aggregates of any size were observed in any of the HTA samples. Instead, STEM was used here mainly for elemental mapping purposes. Figure 8 depicts HAADF images of a representative fresh Cu/SSZ-13 particle, with elemental mapping for Cu, O, Al, Si, and Al+Si. All images display a high degree of homogeneity, consistent with an expected random T site distribution and atomic dispersion of Cu ions. More images of the fresh catalyst are shown in S8 (Figure S6(a-d)), further corroborating the results shown in Figure 8. Images of representative HTA-800 particles and the corresponding elemental mappings are displayed in Figure 9. Again, these images are highly homogeneous. Importantly, no enrichment of Cu and Al is found in the surface layers suggesting the absence of Cu and Al segregation to the surface. More images of the HTA-800 catalyst are shown in Figure S7(a-c).
Page 10 of 16
material. Cu quantification via XPS is known to underestimate the bulk Cu content for highly dispersed Cu in zeolites, which is consistent with our data indicating that the Cu contents are under-measured by at least 50% for both samples.60, 61 A comparison between the fresh and HTA-800 samples demonstrates that EFAl from dealumination, and CuOx formed via [Cu(OH)]+-Z agglomeration do not segregate to the zeolite surface during hydrothermal aging, consistent with STEM imaging. Therefore, it can be posited that CuOx clusters formed during hydrothermal aging (with Al incorporation) have sizes in the sub-nanometer range, and that they reside inside CHA pores and channels. Table 3: Surface compositions of Fresh and HTA-800 Cu/SSZ13 samples determined with XPS.
3.3. Computational Studies The experiential investigations described above clearly demonstrate that dealumination and detachment of [Cu(OH)]+-Z sites occurs during hydrothermal aging, while Cu2+-2Z survives even after very harsh aging. DFT was applied next to provide further insights into dealumination and Cu-ion detachment mechanisms.
Figure 9. HAADF STEM images of a group of HTA-800 Cu/SSZ13 catalyst particles, and the corresponding elemental mapping for Cu, O, Al, Si and Al+Si. Note that all elements are uniformly distributed. More images are shown in Figure S7.
To further confirm the absence of Cu and Al surface segregation, the concentrations of Cu, O, Al, and Si in the near-surface region of the fresh and HTA-800 samples were probed with XPS. Samples were evaluated in vacuum at ambient temperature, without any pretreatment prior the measurements. For each sample, measurements were conducted at 3 different locations to confirm composition uniformity. Narrow scan Cu 2p, Al 2p, Si 2p and O 1s spectra are displayed in Figure S8, and averaged surface region elemental weight percentages are presented in Table 3. For both the fresh and aged samples, the Si/Al ratios are close to the parent SSZ-13
Figure 10. DFT calculated reaction energetics for SSZ-13 dealumination. Activation barriers for the 4 Al-O-Si bond cleavage steps are marked in the figure. The optimized intermediate structures, denoted with parenthesized letters A to K, are displayed in Figure S9.
Zeolite dealumination is known to occur via sequential Al-O(H)-Si bond hydrolysis. SSZ-13 dealumination has been studied in recent years.62-65 As pointed out by Nielsen et al, even for this simple zeolite (one T site), there are still a few hundred possibilities of hydrolysis order and mechanisms.64 Clearly elucidating the ener10
ACS Paragon Plus Environment
Page 11 of 16
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
ACS Catalysis
getically most favorable pathway is a formidable task. In reality, it is conceivable that multiple dealumination pathways energetically similar coexist. Herein, we only simulate one of the possible pathways based on a generally accepted reaction mechanism that hydrolysis is initiated by water adsorption on Al, and followed bond hydrolysis of each Al-O-Si, where the proton moves to one of the remaining Al-O-Si bond until detached Al(OH)3 is formed. Figure 10 presents the calculated Gibbs reaction free energy profiles along the reaction coordinate at 600 °C and ambient pressure, and the relaxed intermediate structures are displayed in S11(Figure S9(A-K)). The highest Gibbs activation free energy of 143 kJ/mol occurs for the hydrolysis of the third Al-O-Si bond. This value appears to be lower than values simulated by Swang and coworkers (190 kJ/mol),62-64 but higher than numbers obtained by Silaghi et al. (114 kJ/mol at RT).65 Again, our aim here is not to find the lowest energy pathway, but rather demonstrate that SSZ-13 dealumination is energetically demanding, and this is why it only minimally occurs at aging temperatures ≤ 600 °C (Figure 4(a)). Note that the study by Silaghi et al.65 demonstrated a decrease in the dealumination energy barrier with increasing temperature, which is in line with our experimental results.
CuOx clusters. These processes are not simulated here. Based on the generally accepted consensus that a charged [Cu(NH3)2]+ migrates readily between Chabazite pores,28, 66 it follows that migration of the chargeneutral Cu(OH)2 in SSZ-13 should not be too demanding, both sterically and energetically. Finally, for Cu2+-2Z sites located in 6MRs, hydrolysis Gibbs free energies were simulated and the results are depicted in Figure 12, with optimized structures for intermediates shown in Figure S11. In this case, formation of the first Cu-OH bond requires overcoming an energy barrier of 92 kJ/mol. However, formation of the second Cu-OH bond, which leads to Cu detaching from the zeolite, is highly demanding with a formidable energy barrier of 233 kJ/mol. Note that this is fully consistent with the EPR quantification results shown in Figure 6 demonstrating that Cu2+-2Z possesses much higher hydrothermal stability than [Cu(OH)]+-Z.
Figure 12. DFT calculated reaction energetics for Cu2+-2Z hydrolysis. The optimized intermediate structures, denoted with parenthesized letters A to H, are displayed in Figure S11.
4. DISCUSSION
Figure 11. DFT calculated reaction energetics for [Cu(OH)]+-Z hydrolysis. The optimized intermediate structures, denoted with parenthesized letters A to D, are displayed in Figure S10.
For [Cu(OH)]+-Z species located at charged sites within 8MRs, a conceivable pathway for their removal from internal zeolite surfaces is via hydrolysis as follows: [Cu(OH)]+ + H2O = Cu(OH)2 + H+
(4)
Energy profiles for this process are shown in Figure 11, and the optimized structures for key intermediates and the final state are displayed in Figure S10. The highest Gibbs free energy barrier for the process is 155 kJ/mol, slightly higher than the simulated highest energy barrier for dealumination. It is anticipated that, upon formation, Cu(OH)2 migrates and agglomerates to yield
In heterogeneous catalysis, development of structurefunction correlations at an atomic level is highly important but very challenging. Simplified models, notably two dimensional model catalysts (e.g., single crystals), where active sites can be analyzed and described at an atomic level, have been studied for decades to deduce some firm structure-function relationships for certain catalytic reactions.67-69 The concept of utilizing simplified model catalysts in technically relevant forms (e.g., powdered catalysts) has also witnessed considerable developments in recent years. For example, single site heterogeneous catalysts have been developed on welldefined supports (mesoporous silica, microporous zeolites, etc.) to derive principles for the design of new catalysts and the improvement of existing ones.70 Recently, Corma and coworkers demonstrated that Pt supported on SSZ-13 can be prepared with tunable nuclearity from single atoms to encapsulated nanoparticles as single site catalysts.71 Cu/SSZ-13 can be categorized as a single 11
ACS Paragon Plus Environment
ACS Catalysis
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
site heterogeneous catalyst for two reasons: (1) SSZ-13 has a well-defined and relatively simple Chabazite structure constructed by stacking double six-membered ring prisms with only one type of T site; and (2) when ion-exchanged properly, this catalyst contains only isolated, monomeric Cu ions. Under dehydrated conditions, these ions are located at 6MR and 8MR windows while, under wet (including low-temperature SCR) conditions, they are solvated and positioned within Chabazite cages.2 To our knowledge, this catalyst is one of the few that are both technically highly important and structurally very well defined. As such, rational design principles for this catalyst can and should be derived from detailed atomic level understandings of the nature of the active centers, especially with respect to their durability, a property that is essential for practical vehicle applications. To our knowledge, such design principles have not yet been established, despite the rather extensive recent work on understanding hydrothermal stability of Cu/Chabazite catalysts.5, 36, 38, 39, 72 4.1. The Fate of Cu Species during Hydrothermal Aging It is well-known that Cu-O-Cu dimers, CuOx clusters, and bulk CuO are EPR silent even though the Cu(II) atoms are paramagnetic. Signal smearing is due to antiferromagnetic interactions: adjacent Cu2+ ions tend to have opposed spins as a result of antiferromagnetic super exchange.58 For isolated Cu(II) ions in zeolites, we have shown previously that they are EPR active and 100% detectable in a fully hydrated form.8 This notion is again confirmed here by comparing quantification results from EPR and ICP of the fresh Cu/SSZ-13 catalyst used in this study. Upon full dehydration, [Cu(H2O)6]2+ converts to naked Cu2+, which remains EPR active. However, [Cu(OH)(H2O)5]+ can undergo a number of chemical transformations, including (1) full dehydration to “naked” [Cu(OH)]+, (2) autoreduction to Cu+, and (3) formation of Cu-dimers via dehydration condensation. Importantly, all of these latter species are EPR silent. In particular, Godiksen et al. recently suggested that [Cu(OH)]+ is EPR silent due to a pseudo Jahn-Teller effect,19 with absorbed microwave energy by [Cu(OH)]+ dissipating too rapidly for the time scale of EPR; that is, the excited state relaxes before it is detected. For Cu+, EPR ‘silence’ is due to the lack of paramagnetic electron and, for Cu-dimers, to antiferromagnetic interactions. Overall, based on the varying EPR characteristics for Cu2+, [Cu(OH)]+ and CuOx, their populations can be readily estimated when (1) the total Cu loading is known (e.g., via ICP), and (2) EPR active Cu contents are known in both hydrated and dehydrated samples. 4.1.1. Changes in Cu Distributions as Reflected in Performances of SCR and NH3 oxidation
Page 12 of 16
The Cu quantification results (Table S2 and Figure 6) demonstrate exceptional hydrothermal stability for Cu2+-2Z. This notion is further corroborated by NH3 titration DRIFTS data displayed in Figure 7(b), and by DFT simulations shown in Figure 12. In contrast, [Cu(OH)]+-Z agglomerates during hydrothermal aging to an extent that increases with increasing aging temperature. By comparing SCR kinetics of the fresh and HTA catalysts, the important question of which of the two sites is more active in catalyzing standard SCR can now be addressed for the first time. The kinetics parameters shown in Table 1 indicate that both sites have similar apparent activation energies for SCR consistent with theoretical predictions by Paolucci et al.28 However, [Cu(OH)]+-Z appears to be more accessible to the reactants, as evidenced by the 3 orders of magnitude decrease in preexponential factors when the fresh sample (heavily populated with [Cu(OH)]+-Z) and the HTA800/900 samples (predominately Cu2+-2Z) are compared. At an SCR reaction temperature of 150 °C, reaction rates were normalized on a per isolated Cu-ion basis (i.e., TOF), and the results are shown in Figure S12. Without considering other factors that may influence SCR rate including, for example, Brønsted acidity and NH3 storage contributions,10 a first approximation indicates that [Cu(OH)]+-Z sites are ~1.5 times more active than Cu2+-2Z. Note that at a reaction temperature of 150 °C, this small difference in reactivity can be readily realized when the reaction activation barriers for the [Cu(OH)]+-Z sites are a few kJ/mol lower than those for the Cu2+-2Z sites. This is very likely since the latter sites are charge-balanced by stronger, paired framework charges. An alternative explanation could be that the NH3 solvated Cu2+-2Z sites are sterically slightly less favorable than NH3 solvated [Cu(OH)]+-Z sites for NO adsorption, a notion in line with the prefactor variation trend shown in Table 1. Based on this simple activity comparison, increasing the populations of [Cu(OH)]+-Z sites appears to benefit low-temperature SCR activity. However, the limited hydrothermal stability of [Cu(OH)]+-Z complicates this Cu loading optimization strategy, as further discussed immediately below. The CuOx cluster formation mechanism is not entirely clear, but seems likely to involve [Cu(OH)]+-Z hydrolysis, Cu(OH)2 migration, and agglomeration via a dihydroxylation-condensation sequence. The high activation barrier (155 kJ/mol) for the hydrolysis step, as derived from DFT (Figure 11), is consistent with the fact that [Cu(OH)]+-Z conversion to CuOx requires rather high temperatures. Even though there are still uncertainties regarding the nature of CuOx clusters, e.g., their nuclearity, location and charge state, from XPS (Table 3) and STEM mapping (Figure 9) analyses, it can readily be concluded that these are not located on the zeolite external surfaces and their sizes are quite small. Thus, we propose that these clusters reside in pores of SSZ-13, 12
ACS Paragon Plus Environment
Page 13 of 16
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
ACS Catalysis
including the primary micropores and secondary mesopores that are generated during hydrothermal aging (Figure 3 and Table 2). Another long standing question regarding the nature of CuOx is whether CuOx species and EFAl (from dealumination, presumably as Al(OH)3) interact. From Figure 4(a), the largely NMR invisibility for EFAl suggests that they likely reside in the vicinity of paramagnetic Cu sites. Recently, Vennestrøm et al. proposed that EFAl interacts with Cu during hydrothermal aging to form CuAl2O4-like species that contains a single Cu(II) atom.73 For such a species, the Cu(II) center should be EPR active, which is inconsistent with our EPR quantification results demonstrating that the CuOx sites must be multinuclear (i.e., EPR silent). To corroborate this argument, a 1% CuO and γ-Al2O3 mixture was calcined to 1000 °C to form CuAl2O4. The majority of Cu in this sample was indeed EPR active (data not shown). NH3 oxidation results shown in Figure S3 provide important hints about CuOx-EFAl interactions. Particularly, Figure S3(b) shows that at 400 °C and above, NH3 oxidation activity for HTA-800 becomes much higher than HTA-750. Note that the quantities of CuOx in these two samples are very similar (Figure 6). This indicates that the nature of CuOx changes significantly when the aging temperature increases from 750 to 800 °C. In Supporting Information S3, some speculations are given. Without further (spectroscopic) evidence, however, no more details on the (Al containing) CuOx structures can be given at this time. From the discussions provided above, CuOx formation becomes unavoidable during hydrothermal aging for Cu/SSZ-13 catalysts with Cu loadings beyond Cu2+-2Z saturation. The clear detrimental effects of these clusters, especially in catalysts aged at 800 °C and above, is that they become highly active for NH3 oxidation above ~350 °C, thus lowering high-temperature SCR selectivity (Figure S2). Under normal on-road operation conditions, this is not a major concern since the SCR catalyst bed typically operates between 200-400 °C. However, higher catalyst bed temperatures are realized during high-temperature maintenance events including, e.g., soot combustion and SCR desulfurization.27 For Cu/zeolite SCR catalysts with lower stability (e.g., Cu/ZSM-5), it is generally believed that structural degradation during hydrothermal aging is due to dealumination.74 For the structurally more rigid SSZ-13, dealumination should be less of a problem for the structural integrity of SSZ-13 because, for example, it can be formed over a wide range of Si/Al ratios. Indeed, we have shown recently that crystallinity maintains for low loaded Fe/SSZ-13 even with ~80% dealumination (determined with 27Al NMR).59
4.1.2. Effect of CuOx on Zeolite Structural Integrity If dealumination is the primary cause of structural degradation for Cu/SSZ-13, then it follows that increasing Cu loading will increase Cu/SSZ-13 hydrothermal stability since Cu protects framework Al. However, this is quite contrary to recent discoveries that Cu/zeolites with high Cu loadings degrade more readily during hydrothermal aging. 36, 38, 39 Instead, we next show that structural degradation of Cu/SSZ-13 during hydrothermal aging is due to the presence of CuOx clusters rather than dealumination, with the clusters aggressively destroying the SSZ-13 structure during high-temperature hydrothermal aging. For example, Nam and coworkers showed that high exchange levels in Cu/SSZ-13 catalysts with varying Si/Al ratios lead to much lowered hydrothermal stability at 850 °C.39 Still, a satisfactory explanation for this important phenomenon is lacking. In the present study, via detailed porosity analysis of the fresh, HTA-700, and HTA-800 samples (Figure 3), it was found that secondary mesoporosity is generated during hydrothermal aging at 700 °C, but the primary micropores are essentially unaffected. In contrast for an aging temperature of 800 °C, the newly generated mesopores largely result from micropore damage. These findings clearly demonstrate the destructive nature of CuOx for the first time. It is conceivable that the CuOx clusters are larger (due to more extensive agglomeration and Al incorporation) and more mobile at the higher aging temperature of 800 °C relative to 700 °C. In this respect, the structural damaging effect, reflected in the mesopore generation at the expense of micropore damage, can be considered to be caused from transport of CuOx clusters with sizes larger than the SSZ-13 pore openings (~3.8 Å). Accumulation of such secondary pores and channels will eventually lead to permanent structure damage of Cu/SSZ-13; note that such processes will depend on temperature, aging time and Cu loading, where higher aging temperatures, longer aging durations and higher Cu loadings promote more damage. Prior literature is fully in line with this structure damage mechanism.5, 36, 38, 39, 72 On the other hand, dealumination alone is much less destructive for this catalyst. 4.2. Practical Implications Finally, some rational design principles can be suggested based on the atomic-level understanding gained during the course of this study. (1) For the enhancement of low-temperature NOx conversion, increasing ion-exchange levels that populate [Cu(OH)]+-Z species to near saturation will be beneficial. However, catalysts thus formulated will not be as hydrothermally stable. (2) For the enhancement of high-temperature selectivity, maximizing Cu2+-2Z and minimizing [Cu(OH)]+-Z is beneficial. Such a formulation may render relatively high residual Brønsted acid site densities that are vul13
ACS Paragon Plus Environment
ACS Catalysis
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
nerable to dealumination. In this case, introduction of a cocation that neutralizes some Brønsted acid sites can pose benefits.12 (3) For optimum hydrothermal stability of Cu/SSZ-13 catalysts, preventing the formation of large CuOx clusters is critical. In this case, total Cu loading must be based on Si/Al ratios of the zeolite. Nam and coworkers suggested an optimum ion-exchange level of ~70%.39 This number should be further lowered since CuOx formation is clearly the most detrimental factor, much more than dealumination, in influencing long term stability of this catalyst. (4) Finally, Si/Al ratios of the zeolite substrate should also be optimized in order to incorporate enough active Cu ions for low-temperature activities. The change from Si/Al = 17.5 for the first generation catalyst to Si/Al = 10 for the current state-of-the-art catalyst reflects efforts for this optimization. 5. CONCLUSION Using model Cu/SSZ-13 catalysts that contain either Cu2+-2Z or [Cu(OH)]+-Z active centers, their individual SCR characteristics can be isolated and defined. Cu2+-2Z is highly selective under all reaction temperatures relevant to application, while [Cu(OH)]+-Z is slightly less selective at elevated reaction temperatures. The Cu species that primarily result in a deterioration of hightemperature SCR selectivity are CuOx clusters. From detailed EPR analyses, the amounts of Cu2+-2Z, [Cu(OH)]+-Z, and CuOx clusters can be accurately estimated. This quantification convincingly highlights the remarkable hydrothermal stability of Cu2+-2Z, along with the gradual conversion of [Cu(OH)]+-Z to CuOx clusters. SCR and NH3 oxidation kinetics are nicely correlated to these structural changes. Diffusion of CuOx clusters larger than the primary pore openings of SSZ13 can be regarded as a primary cause for structure destabilization of this catalyst during hydrothermal aging. To conclude, rational design principles can now be derived based on such detailed atomic-level and nearatomic-level understanding of the hydrothermal stability of these catalysts. These principles follow from this work’s elucidation of previously unexplained structurefunction phenomena observed for Cu/SSZ-13, and can guide the future development of optimum SCR catalysts.
AUTHOR INFORMATION Corresponding Author * F. Gao (
[email protected]) and D. Mei (
[email protected]) are the authors to whom correspondence should be addressed.
ASSOCIATED CONTENT
Page 14 of 16
Supporting Information. S1: H2-TPR results for model Cu/SSZ-13 samples. S2: light-off (NOx and NH3 conversion vs. temperature) curves for standard SCR over fresh and HTA800 samples, and a NOx light-off curve for a hydrothermally aged BASF commercial catalyst. S3: NH3 oxidation conversions vs. temperature for fresh and HTA Cu/SSZ-13 samples. S4: XRD patterns, and BET surface area and porosity analysis results obtained with N2 adsorption using a Quantachrome Autosorb-6 analyzer. S5: Adsorption isotherms for selected samples (fresh, HTA-700, and HTA-800) obtained using Ar with a Micromeritics ASAP 2000 analyzer. S6: detailed Cu 2+ species estimations with EPR. S7: Co titration results. S8: additional STEM and elemental mapping images for a fresh sample. S9: additional STEM and elemental mapping images for the HTA-800 sample. S10: Cu 2p, Al 2p, Si 2P, and O 1s region XPS spectra for fresh and HTA-800 samples. S11: optimized relaxed intermediate structures present during SSZ13 dealumination. S12: optimized intermediate structures present along the hydrolysis pathway for removal of + [Cu(OH)] -Z from cationic positions within the 8MRs. S13: optimized intermediate structures present along the hydroly2+ sis pathway for the removal of Cu -2Z from cationic positions within the 6MRs. This material is available free of charge via the Internet at http://pubs.acs.org.
ACKNOWLEDGMENT The authors gratefully acknowledge the US Department of Energy (DOE), Energy Efficiency and Renewable Energy, Vehicle Technologies Office for the support of this work. Computing time was granted by a user proposal at the William R. Wiley Environmental Molecular Sciences Laboratory (EMSL) and by the National Energy Research Scientific Computing Center (NERSC). The experimental studies described in this paper were performed in the EMSL, a national scientific user facility sponsored by the DOE’s Office of Biological and Environmental Research and located at Pacific Northwest National Laboratory (PNNL). PNNL is operated for the US DOE by Battelle.
REFERENCES 1. Epling, W. S.; Campbell, L. E.; Yezerets, A.; Currier, N. W.; Parks, J. E. Catal. Rev. 2004, 46, 163-245. 2. Beale, A. M.; Gao, F.; Lezcano-Gonzalez, I.; Peden, C. H. F.; Szanyi, J. Chem. Soc. Rev. 2015, 44, 7371-7405. 3. Nova, I.; Tronconi, E., Urea-SCR Technology for deNOx After Treatment of Diesel Exhausts. Springer Science+Business Media: New York, 2014. 4. Gao, F.; Kwak, J. H.; Szanyi, J.; Peden, C. H. F. Top. Catal. 2013, 56, 1441-1459. 5. Kwak, J. H.; Tran, D.; Burton, S. D.; Szanyi, J.; Lee, J. H.; Peden, C. H. F. J. Catal. 2012, 287, 203-209. 6. Kwak, J. H.; Tran, D.; Szanyi, J.; Peden, C. H. F.; Lee, J. H. Catal. Lett. 2012, 142, 295-301. 7. Kwak, J. H.; Zhu, H. Y.; Lee, J. H.; Peden, C. H. F.; Szanyi, J. Chem. Commun. 2012, 48, 4758-4760. 8. Gao, F.; Walter, E. D.; Karp, E. M.; Luo, J. Y.; Tonkyn, R. G.; Kwak, J. H.; Szanyi, J.; Peden, C. H. F. J. Catal. 2013, 300, 2029. 14
ACS Paragon Plus Environment
Page 15 of 16
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
ACS Catalysis
9. Gao, F.; Walter, E. D.; Kollar, M.; Wang, Y. L.; Szanyi, J.; Peden, C. H. F. J. Catal. 2014, 319, 1-14. 10. Gao, F.; Washton, N. M.; Wang, Y. L.; Kollár, M.; Szanyi, J.; Peden, C. H. F. J. Catal. 2015, 331, 25-38. 11. Prodinger, S.; Derewinski, M. A.; Wang, Y. L.; Washton, N. M.; Walter, E. D.; Szanyi, J.; Gao, F.; Wang, Y.; Peden, C. H. F. Appl. Catal. B-Environ. 2017, 201, 461-469. 12. Gao, F.; Wang, Y. L.; Washton, N. M.; Kollar, M.; Szanyi, J.; Peden, C. H. F. ACS Catal. 2015, 5, 6780-6791. 13. Giordanino, F.; Vennestrom, P. N. R.; Lundegaard, L. F.; Stappen, F. N.; Mossin, S.; Beato, P.; Bordiga, S.; Lamberti, C. Dalton Trans. 2013, 42, 12741-12761. 14. Kwak, J. H.; Lee, J. H.; Burton, S. D.; Lipton, A. S.; Peden, C. H. F.; Szanyi, J. Angew. Chem. Int. Edit. 2013, 52, 9985-9989. 15. Szanyi, J.; Kwak, J. H.; Zhu, H. Y.; Peden, C. H. F. Phys. Chem. Chem. Phys. 2013, 15, 2368-2380. 16. Andersen, C. W.; Bremholm, M.; Vennestrøm, P. N. R.; Blichfeld, A. B.; Lundegaard, L. F.; Iversen, B. B. IUCrJ 2014, 1, 382-386. 17. Bates, S. A.; Verma, A. A.; Paolucci, C.; Parekh, A. A.; Anggara, T.; Yezerets, A.; Schneider, W. F.; Miller, J. T.; Delgass, W. N.; Ribeiro, F. H. J. Catal. 2014, 312, 87-97. 18. Giordanino, F.; Borfecchia, E.; Lomachenko, K. A.; Lazzarini, A.; Agostini, G.; Gallo, E.; Soldatov, A. V.; Beato, P.; Bordiga, S.; Lamberti, C. J. Phys. Chem. Lett. 2014, 5, 15521559. 19. Godiksen, A.; Stappen, F. N.; Vennestrom, P. N. R.; Giordanino, F.; Rasmussen, S. B.; Lundegaard, L. F.; Mossin, S. J. Phys. Chem. C 2014, 118, 23126-23138. 20. Kwak, J. H.; Varga, T.; Peden, C. H. F.; Gao, F.; Hanson, J. C.; Szanyi, J. J. Catal. 2014, 314, 83-93. 21. Lezcano-Gonzalez, I.; Deka, U.; Arstad, B.; Van Yperen-De Deyne, A.; Hemelsoet, K.; Waroquier, M.; Van Speybroeck, V.; Weckhuysen, B. M.; Beale, A. M. Phys. Chem. Chem. Phys. 2014, 16, 1639-1650. 22. Paolucci, C.; Verma, A. A.; Bates, S. A.; Kispersky, V. F.; Miller, J. T.; Gounder, R.; Delgass, W. N.; Ribeiro, F. H.; Schneider, W. F. Angew. Chem. Int. Edit. 2014, 53, 1182811833. 23. Verma, A. A.; Bates, S. A.; Anggara, T.; Paolucci, C.; Parekh, A. A.; Kamasamudram, K.; Yezerets, A.; Miller, J. T.; Delgass, W. N.; Schneider, W. F.; Ribeiro, F. H. J. Catal. 2014, 312, 179-190. 24. Zhang, R. Q.; McEwen, J. S.; Kollar, M.; Gao, F.; Wang, Y. L.; Szanyi, J.; Peden, C. H. F. ACS Catal. 2014, 4, 4093-4105. 25. Janssens, T. V. W.; Falsig, H.; Lundegaard, L. F.; Vennestrom, P. N. R.; Rasmussen, S. B.; Moses, P. G.; Giordanino, F.; Borfecchia, E.; Lomachenko, K. A.; Lamberti, C.; Bordiga, S.; Godiksen, A.; Mossin, S.; Beato, P. ACS Catal. 2015, 5, 2832-2845. 26. Moreno-Gonzalez, M.; Hueso, B.; Boronat, M.; Blasco, T.; Corma, A. J. Phys. Chem. Lett. 2015, 6, 1011-1017. 27. Luo, J. Y.; Wang, D.; Kumar, A.; Li, J. H.; Kamasamudram, K.; Currier, N. W.; Yezerets, A. Catal. Today 2016, 267, 3-9. 28. Paolucci, C.; Parekh, A. A.; Khurana, I.; Di Iorio, J. R.; Li, H.; Caballero, J. D. A.; Shih, A. J.; Anggara, T.; Delgass, W. N.; Miller, J. T.; Ribeiro, F. H.; Gounder, R.; Schneider, W. F. J. Am. Chem. Soc. 2016, 138, 6028-6048. 29. Gao, F.; Mei, D. H.; Wang, Y. L.; Szanyi, J.; Peden, C. H. F. J. Am. Chem. Soc. 2017, 139, 4935-4942.
30. Ren, L. M.; Zhu, L. F.; Yang, C. G.; Chen, Y. M.; Sun, Q.; Zhang, H. Y.; Li, C. J.; Nawaz, F.; Meng, X. J.; Xiao, F. S. Chem. Commun. 2011, 47, 9789-9791. 31. Martinez-Franco, R.; Moliner, M.; Thogersen, J. R.; Corma, A. ChemCatChem 2013, 5, 3316-3323. 32. Martin, N.; Moliner, M.; Corma, A. Chem. Commun. 2015, 51, 9965-9968. 33. Wang, X.; Wu, Q. M.; Chen, C. Y.; Pan, S. X.; Zhang, W. P.; Meng, X. J.; Maurer, S.; Feyen, M.; Muller, U.; Xiao, F. S. Chem. Commun. 2015, 51, 16920-16923. 34. Zhao, Z. C.; Yu, R.; Zhao, R. R.; Shi, C.; Gies, H.; Xiao, F. S.; De Vos, D.; Yokoi, T.; Bao, X. H.; Kolb, U.; Feyen, M.; McGuire, R.; Maurer, S.; Moini, A.; Muller, U.; Zhang, W. P. Appl. Catal. BEnviron. 2017, 217, 421-428. 35. Zhang, L.; Wang, D.; Liu, Y.; Kamasamudram, K.; Li, J. H.; Epling, W. Appl. Catal. B-Environ. 2014, 156, 371-377. 36. Schmieg, S. J.; Oh, S. H.; Kim, C. H.; Brown, D. B.; Lee, J. H.; Peden, C. H. F.; Kim, D. H. Catal. Today 2012, 184, 252-261. 37. Ye, Q.; Wang, L. F.; Yang, R. T. Appl. Catal. A-Gen. 2012, 427, 24-34. 38. Blakeman, P. G.; Burkholder, E. M.; Chen, H. Y.; Collier, J. E.; Fedeyko, J. M.; Jobson, H.; Rajaram, R. R. Catal. Today 2014, 231, 56-63. 39. Kim, Y. J.; Lee, J. K.; Min, K. M.; Hong, S. B.; Nam, I. S.; Cho, B. K. J. Catal. 2014, 311, 447-457. 40. Lezcano-Gonzalez, I.; Deka, U.; van der Bij, H. E.; Paalanen, P.; Arstad, B.; Weckhuysen, B. M.; Beale, A. M. Appl. Catal. B-Environ. 2014, 154, 339-349. 41. Luo, J. Y.; Gao, F.; Kamasamudram, K.; Currier, N. W.; Peden, C. H. F.; Yezerets, A. J. Catal. 2017, 348, 291–299. 42. VandeVondele, J.; Krack, M.; Mohamed, F.; Parrinello, M.; Chassaing, T.; Hutter, J. Comput. Phys. Commun. 2005, 167, 103-128. 43. Goedecker, S.; Teter, M.; Hutter, J. Phys. Rev. B 1996, 54, 1703-1710. 44. Hartwigsen, C.; Goedecker, S.; Hutter, J. Phys. Rev. B 1998, 58, 3641-3662. 45. Krack, M.; Parrinello, M. Phys. Chem. Chem. Phys. 2000, 2, 2105-2112. 46. VandeVondele, J.; Hutter, J. J. Chem. Phys. 2007, 127, 114105. 47. Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865. 48. Nielsen, M.; Brogaard, R. Y.; Falsig, H.; Beato, P.; Swang, O.; Svelle, S. ACS Catal. 2015, 5, 7131-7139. 49. Henkelman, G.; Uberuaga, B. P.; Jonsson, H. J. Chem. Phys. 2000, 113, 9901-9904. 50. Mills, G.; Jonsson, H.; Schenter, G. K. Surf. Sci. 1995, 324, 305-337. 51. Kamasamudram, K.; Currier, N. W.; Chen, X.; Yezerets, A. Catal. Today 2010, 151, 212-222. 52. Long, R. Q.; Yang, R. T. J. Catal. 2002, 207, 274-285. 53. Groen, J. C.; Peffer, L. A. A.; Pérez-Ramı́rez, J. Microporous Mesoporous Mater. 2003, 60, 1-17. 54. Venero, A. F.; Chiou, J. N. MRS Proceedings 1987, 111. 55. Ravikovitch, P. I.; Vishnyakov, A.; Russo, R.; Neimark, A. V. Langmuir 2000, 16, 2311-2320. 56. van Donk, S.; Janssen, A. H.; Bitter, J. H.; de Jong, K. P. Catal. Rev. 2003, 45, 297-319. 57. Klinowski, J. Chem. Rev. 1991, 91, 1459-1479. 58. Morton, J. R. J. Chem. Soc. Farad. Trans. 1989, 85, 4091-4098.
15
ACS Paragon Plus Environment
ACS Catalysis
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 16 of 16
59. Kovarik, L.; Washton, N. M.; Kukadapu, R.; Devaraj, A.; Wang, A. Y.; Wang, Y. L.; Szanyi, J.; Peden, C. H. F.; Gao, F. ACS Catal 2017, 7, 2458-2470. 60. Corma, A.; Palomares, A.; Marquez, F. J. Catal. 1997, 170, 132-139. 61. Pereda-Ayo, B.; De La Torre, U.; Illan-Gomez, M. J.; Bueno-Lopez, A.; Gonzalez-Velasco, J. R. Appl. Catal. B-Environ. 2014, 147, 420-428. 62. Malola, S.; Svelle, S.; Bleken, F. L.; Swang, O. Angew. Chem. Int. Edit. 2012, 51, 652-655. 63. Fjermestad, T.; Svelle, S.; Swang, O. J. Phys. Chem. C 2013, 117, 13442-13451. 64. Nielsen, M.; Brogaard, R. Y.; Falsig, H.; Beato, P.; Swang, O.; Svelle, S. ACS Catal 2015, 5, 7131-7139. 65. Silaghi, M. C.; Chizallet, C.; Sauer, J.; Raybaud, P. J. Catal. 2016, 339, 242-255. 66. Clemens, A. K. S.; Shishkin, A.; Carlsson, P. A.; Skoglundh, M.; Martinez-Casado, F. J.; Matej, Z.; Balmes, O.; Harelind, H. ACS Catal 2015, 5, 6209-6218. 67. Gao, F.; Goodman, D. W. Annu. Rev. Phys. Chem. 2012, 63, 265-286. 68. Goodman, D. W. J. Catal. 2003, 216, 213-222. 69. Baumer, M.; Freund, H. J. Prog. Surf. Sci. 1999, 61, 127-198. 70. Thomas, J. M.; Raja, R.; Lewis, D. W. Angew. Chem. Int. Edit. 2005, 44, 6456-6482. 71. Moliner, M.; Gabay, J. E.; Kliewer, C. E.; Carr, R. T.; Guzman, J.; Casty, G. L.; Serna, P.; Corma, A. J. Am. Chem. Soc. 2016, 138, 15743-15750. 72. Wang, D.; Jangjou, Y.; Liu, Y.; Sharma, M. K.; Luo, J. Y.; Li, J. H.; Kamasamudram, K.; Epling, W. S. Appl. Catal. BEnviron. 2015, 165, 438-445. 73. Vennestrom, P. N. R.; Janssens, T. V. W.; Kustov, A.; Grill, M.; Puig-Molina, A.; Lundegaard, L. F.; Tiruvalam, R. R.; Concepcion, P.; Corma, A. J. Catal. 2014, 309, 477-490. 74. Fickel, D. W.; D'Addio, E.; Lauterbach, J. A.; Lobo, R. F. Appl. Catal. B-Environ. 2011, 102, 441-448.
TOC
16
ACS Paragon Plus Environment
