Physicochemical Stabilization of Pt against Sintering for a

Mar 23, 2016 - The physically stabilized Pt catalyst is denoted as “P-PtGa”, whereas the catalyst that is simultaneously stabilized physically and...
0 downloads 8 Views 9MB Size
Download PDF
This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.

Research Article pubs.acs.org/acscatalysis

Physicochemical Stabilization of Pt against Sintering for a Dehydrogenation Catalyst with High Activity, Selectivity, and Durability Juhwan Im and Minkee Choi* Department of Chemical and Biomolecular Engineering, Korea Advanced Institute of Science and Technology, Daejeon 34141, Republic of Korea S Supporting Information *

ABSTRACT: Suppressing irreversible catalyst deactivation is critical in heterogeneous catalysis. In particular, deactivation via sintering of active sites is a significant issue for reactions involving harsh reaction/ regeneration conditions. In this work, we developed a PtGa/γ-Al2O3 alkane dehydrogenation catalyst with exceptionally high activity, selectivity, and long-term stability by markedly suppressing Pt sintering under harsh conditions (reaction/regeneration at >823 K). To stabilize Pt, physical and chemical stabilization strategies were synergistically combined. For the former, Pt was introduced during the synthesis of γAl2O3 via sol−gel chemistry, which can increase the interfacial contact between Pt and γ-Al2O3 due to the partial entrapment of Pt in γ-Al2O3. For the latter, atomically dispersed Ce was doped on γ-Al2O3, which can stabilize Pt via strong Pt−O−Ce interactions. Because of effective Pt stabilization, the catalyst showed remarkably steady activity and selectivity behaviors over the repeated reaction cycles, although the catalyst is regenerated via simple oxidation rather than industrially used oxychlorination. The Pt stabilization strategies reported in this work can be applied to other metal-catalyzed reactions that involve severe reaction/regeneration conditions. KEYWORDS: heterogeneous catalysis, dehydrogenation, stability, regeneration, sintering

■

INTRODUCTION Light olefins are important raw materials used in petrochemistry for the production of various chemicals such as polymers, oxygenates, and important chemical intermediates.1 Light olefins have been mainly produced in steam cracking and fluid catalytic cracking (FCC) of naphtha and other oil products. However, with a recent increase in shale-gas production, the conventional processes are less cost-effective than ethane cracker, which causes a decrease in the production of propene and C4 olefins. From this background, the onpurpose production of >C3 olefins by alkane dehydrogenation recently has gained increasing scientific attention.1,2 The dehydrogenation of light alkanes is highly endothermic and is thermodynamically limited, which generally requires very high reaction temperatures (>823 K).2 Under such harsh conditions, undesirable side reactions such as cracking, hydrogenolysis, and isomerization can take place. In particular, heavy coke formation causing rapid catalyst deactivation is unavoidable, and thus the catalytic processes generally require frequent catalyst regenerations. This means that the dehydrogenation catalysts should have high thermal stability under both reductive and oxidative atmospheres. Due to their outstanding olefin selectivity and catalyst stability, PtSn3−6 (Oleflex process from UOP) and CrOx7−9 (Catofin process from CB&I Lummus) supported on Al2O3 © XXXX American Chemical Society

have been widely used in commercial dehydrogenation processes. Although substantial improvements in catalysts and process conditions have been made for both technologies, challenges related to catalyst stability and regenerability still should be overcome. For instance, PtSn catalysts require oxychlorination for regeneration in order to maintain their catalytic properties, because coke burning via simple oxidation causes metal sintering and a gradual change of the alloying state of the bimetallic catalysts.3 The oxychlorination requires the use of corrosion-resistant metallurgy for reactor design and an additional process to eliminate corrosive chlorine compounds in the vent gas.1,10 In the case of a CrOx catalyst, the catalyst regeneration process via coke combustion causes the incorporation of CrOx into the Al2O3 framework, leading to a gradual loss of accessible catalytically active sites.8 Furthermore, the high cost of Pt and environmental concerns associated with the use of Cr have spurred a search for alternative catalyst compositions.11 Various transition-metal oxides such as VOx,12 MoOx,13 InOx,14 GaOx,15−20 ZrOx,21 and Pt catalysts supported on various zeolites (e.g., Fe-ZSM-5 and KL)22,23 have been widely studied. Received: February 1, 2016 Revised: March 21, 2016

2819

DOI: 10.1021/acscatal.6b00329 ACS Catal. 2016, 6, 2819−2826

Research Article

ACS Catalysis Recently, Sattler et al.15 reported a Pt-promoted Ga/γ-Al2O3 species (PtGa/γ-Al2O3) as a highly promising catalyst for propane dehydrogenation, which showed remarkably high activity, selectivity, and stability. It was proposed that Ga3+ (3 wt %) substituted within the γ-Al2O3 framework is a main active site for C−H activation, whereas a minute amount of Pt (0.1 wt %) as a promoter facilitates recombination of hydrogen atoms into H2. The catalyst showed a ca. 30% conversion drop at the very early period of reaction/regeneration cycles, but thereafter the catalyst showed stable catalytic performance. Considering that the catalyst was regenerated by simple oxidation in air (not by oxychlorination), this catalyst stability is already remarkable (note that a commercial PtSn/Al2O3 catalyst loses catalytic activity almost completely under similar regeneration conditions).3 The reason for the initial rapid drops in conversion and selectivity was not clearly elucidated in the aforementioned work, but it can be attributed to Pt sintering during catalyst regeneration, because the authors reported significant Pt sintering after oxidation treatment at high temperature.15 In this respect, it is reasonably expected that stabilization of Pt against sintering under such harsh reaction/regeneration conditions can even further improve the catalytic performance of PtGa/γ-Al2O3 catalyst. In the present work, we investigated physical and chemical methods for stabilizing supported Pt catalysts against thermal sintering and synergistically combined them for designing PtGa/γ-Al2O3 catalyst (0.1 wt % Pt and 3 wt % Ga) with remarkably enhanced activity, selectivity, and long-term catalyst stability in alkane dehydrogenations. Catalytic reactions were conducted under industrially relevant conditions (i.e., reactions in the absence of H2 in the reactant feeds and catalyst regeneration via air calcination after 1 h reaction). Rigorous analyses using complementary characterization techniques unequivocally elucidated that the suppression of Pt sintering during repeated reaction/regeneration cycles can indeed lead to the exceptionally stable catalytic performances.

Scheme 1. Schematic Representation of Catalyst Preparation

phase (Figure 1a). According to N2 adsorption−desorption isotherms (Figure 1b), the catalysts prepared by sol−gel methods (P-PtGa and PC-PtGa) showed 15−30% higher Brunauer−Emmett−Teller (BET) areas and 50−60% lower pore volume than N-PtGa prepared using commercial γ-Al2O3 (Table 1). The mesopore size distributions of P-PtGa and PCPtGa were centered at 3.3 nm, whereas N-PtGa had a larger pore size distribution centered at 9.3 nm. The results imply that P-PtGa and PC-PtGa catalysts have structures significantly denser than that of N-PtGa. It is notable that P-PtGa and PC-PtGa have very similar structural properties (BET area, pore volume, and pore size) except for Ce doping in the latter. To understand the Ce dispersion in the PC-PtGa catalyst, we separately synthesized a Pt-free catalyst (i.e., γ-Al2O3 doped with 3 wt % Ga and 1 wt % Ce) via a sol−gel method because Pt and Ce are difficult to distinguish by high-angle annular dark field scanning transmission electron microscopy (HAADF-STEM). HAADFSTEM images showed that a major fraction of Ce is atomically distributed over the γ-Al2O3 without forming a bulk CeO2 domain (Figure 2a). Such a high dispersity of Ce on γ-Al2O3 was further confirmed by the blue shift of UV absorption of CeOx (Figure 2b).27,28 UV−vis diffuse reflectance spectroscopy showed that commercial CeO2 nanoparticles (20−50 nm by TEM, Figure S1 in the Supporting Information) exhibited a strong UV absorption at λ 400 nm (Eg = 3.1 eV). In comparison with unsupported CeO2, the present γ-Al2O3 doped with 3 wt % Ga and 1 wt % Ce showed a blue-shifted UV absorption at λ 365 nm (Eg = 3.4 eV), indicating extremely small Ce domain size (i.e., atomically dispersed, Figure 2b). HAADF-STEM images (Figure 3a−c) clearly revealed that Pt clusters are highly dispersed on all of the prepared catalysts. Surface-area-weighted mean cluster diameters (dTEM) were determined to be ca. 1.1 nm for all of the samples. Nevertheless, the CO chemisorption amount of N-PtGa

■

RESULTS AND DISCUSSION Two strategies were combined for suppressing Pt sintering: a “physical” approach and a “chemical” approach (Scheme 1). For the physical stabilization, the interfacial contact between Pt and γ-Al2O3 was maximized by adding a Pt precursor during a sol−gel synthesis of Ga-doped γ-Al2O3. In the case of Pt/γAl2O3 catalysts, it was reported that sol−gel synthesis can produce a Pt catalyst with lower H2 and O2 accessibility due to the partial entrapment of Pt in the γ-Al2O3 matrix.24 For further stabilization via a chemical strategy, we doped 1 wt % of cerium (Ce) on γ-Al2O3 by adding Ce(NO3)3·6H2O during the sol− gel synthesis. In automotive catalysts, Ce is known to stabilize and redisperse Pt species by Pt−O−Ce bond formation under oxidative conditions.25,26 Considering that Pt sintering mostly occurs during the oxidative regeneration steps due to the high mobility of PtOx, such a Pt−O−Ce interaction is expected to play an important role during repeated catalyst regeneration steps. The physically stabilized Pt catalyst is denoted as “PPtGa”, whereas the catalyst that is simultaneously stabilized physically and chemically is denoted as “PC-PtGa”. For comparison, an ordinary, nonstabilized PtGa/γ-Al2O3 (“NPtGa”) catalyst was also prepared with a commercial γ-Al2O3 support using the incipient wetness method reported by Sattler et al.15 The physical properties of catalysts are summarized in Table 1. All samples showed a typical XRD pattern for the γ-Al2O3 2820

DOI: 10.1021/acscatal.6b00329 ACS Catal. 2016, 6, 2819−2826

Research Article

ACS Catalysis Table 1. Physicochemical Properties of Catalysts composition (wt %)a Pt N-PtGa P-PtGa PC-PtGa

0.10 0.10 0.10

Ga 3.1 (67) 3.0 (66) 3.0 (66)

Ce

BET area (m2 g−1)

pore volume (cm3 g−1)b

pore size (nm)c

CO/Ptd

1.0

201 260 237

0.47 0.22 0.20

9.3 3.3 3.3

0.45 0.32 0.29

e

a

The composition was analyzed by ICP-AES. bDetermined at a P/P0 value of 0.99. cCalculated using the BJH method. dDetermined by CO chemisorption at 323 K. eThe values in parentheses indicate percentages (%) of tetrahedrally coordinated Ga determined by Ga K-edge XANES analysis.

Figure 1. (a) XRD patterns and (b) N2 adsorption−desorption isotherms of the catalysts measured at 77 K.

(CO/Pt = 0.45) was substantially higher than those of P-PtGa (CO/Pt = 0.32) and PC-PtGa (CO/Pt = 0.29) (Table 1). Such lower chemisorption values at similar Pt size distributions were previously attributed to the increased interfacial contact between the Pt cluster and γ-Al2O3 due to the partial entrapment of Pt within the γ-Al2O3 matrix.24 Pt L3-edge Xray absorption near edge structure (XANES) showed that the white line intensity increased in the order N-PtGa < P-PtGa < PC-PtGa, indicating that Pt becomes more electron deficient after physical and chemical stabilizations (Figure 3d). Such electron deficiency in supported metal catalysts frequently has been used to prove the existence of strong interfacial interactions between a metal and an oxide support.29 Because N-PtGa and P-PtGa have very similar chemical compositions (Table 1), the greater electron deficiency in P-PtGa in comparison to that in N-PtGa can be attributed to the increased interfacial contact between Pt and γ-Al2O3, as indicated by the CO chemisorption results. PC-PtGa doped with Ce showed even greater electron deficiency than P-PtGa, although both catalysts have very similar textural properties and chemisorption values (Figure 1b and Table 1). The enhanced Pt electron deficiency in PC-PtGa can be attributed to the presence of the doped Ce species, which can form Pt−O−Ce bonds.26 Additional analyses on Pt such as EXAFS and XPS did not provide meaningful and reliable structural information, due to the extremely low Pt loading (0.1 wt %) (Figures S2 and S3 in the Supporting Information). The initial alkane conversions of the catalysts during 20 consecutive reaction−regeneration cycles in propane and isobutane dehydrogenation are shown in Figure 4. At each cycle, the dehydrogenation reaction was carried out for 1 h, and then a catalyst was regenerated by air calcination for 30 min at the reaction temperature. In propane dehydrogenation at 893 K, the conversion on N-PtGa catalyst at the first cycle was ca. 60% (Figure 4a). The catalytic conversion of N-PtGa decreased

Figure 2. (a) HAADF-STEM image of γ-Al2O3 doped with 3 wt % Ga and 1 wt % Ce prepared by a sol−gel method. The white spots are Ce atoms. (b) UV−vis diffuse reflectance spectra of commercial CeO2 (black) and γ-Al2O3 doped with 3 wt % Ga and 1 wt % Ce prepared by a sol−gel method (red).

rapidly to 27% during the first 10 cycles, and thereafter it was fully stabilized. During the initial cycles, the selectivity to propene also slightly decreased from 98 to 95% (Figure 4b) due to increased formation of hydrogenolysis/cracking products (C1−C2). The overall catalytic behaviors are consistent with those reported by Sattler et al.15 It is notable that the industrially used PtSn/γ-Al2O3 catalyst (0.7 wt % Pt, 0.3 wt % Sn, and 0.8 wt % K; PtSn in Figure 4) showed much faster deterioration of catalytic performances over the reaction cycles (from 39 to 3% after 20 cycles). This is consistent with an earlier report showing that coke burning via simple oxidation (not via industrially used oxychlorination) causes significant sintering and a gradual change of the alloying state of the bimetallic catalysts, resulting in rapidly deteriorating catalytic performances.3 The results clearly imply that the Pt-promoted Ga/γ-Al2O3 catalyst (N-PtGa) is much more effectively 2821

DOI: 10.1021/acscatal.6b00329 ACS Catal. 2016, 6, 2819−2826

Research Article

ACS Catalysis

Notably, the PC-PtGa catalyst with 1 wt % Ce doping showed remarkably improved catalyst stability even in comparison with N-PtGa catalyst. The propane conversion was similar to that of N-PtGa (ca. 60%) at the first cycle, but the conversion was fully stabilized at a 50% conversion level only after five cycles. This is much higher than the stabilization level of N-PtGa (27%) (Figure 4a). Furthermore, the PC-PtGa showed higher and steadier selectivity behavior (>98% propene selectivity, Figure 4b). We additionally investigated the catalytic effect of the Ce doping amounts in PC-PtGa (i.e., 0, 0.5, 1, and 2 wt % Ce doping) (Figure S4 in the Supporting Information). The results showed that the PC-PtGa-type catalysts showed generally enhanced stability at increased Ce doping. Beyond 1 wt % Ce doping, however, the stability enhancement was only marginal, which indicated that 1 wt % Ce doping is sufficient for achieving the optimal catalytic performance. The P-PtGa catalyst showed behaviors intermediate between those of NPtGa and PC-PtGa. In isobutane dehydrogenation at 823 K (Figure 4c,d), all of the catalysts also showed trends similar to those in propane dehydrogenation. Among the three Ptpromoted Ga/γ-Al2O3 catalysts, PC-PtGa showed the steadiest catalytic conversion (conversion decreased from 57 to 51% after 20 cycles), while N-PtGa showed the fastest decrease of conversion (from 65 to 31%). To understand the differences in catalyst performance, all the Pt-promoted Ga/γ-Al2O3 catalysts in their fresh states and after 20 reaction/regeneration cycles in propane and isobutane dehydrogenations were analyzed by combining various complementary characterization techniques. Inductively coupled plasma atomic emission spectroscopy (ICP-AES) showed that all the catalysts retained their Ga content after 20 cycles in both reactions (Table S1 in the Supporting Information). This infers that Ga3+ was not reduced to volatile Ga2O or metallic Ga during the reactions. Ga 2p XPS investigations, before and after 1 h of propane dehydrogenation, clearly revealed that Ga3+ was not reduced during the reaction (Figure S5 in the Supporting Information). Temperature-programmed reduction (TPR) also showed no appreciable H2 consumption up to 1073 K (Figure S6 in the Supporting Information). According to earlier studies, it was proposed that tetrahedrally coordinated Ga3+ is an active site in dehydrogenation reactions.15,20 In order to analyze the coordination state of Ga species, Ga K-edge XANES of the catalysts were carefully analyzed (Figure 5 and Figure S7 and Table S2 in the Supporting Information). XANES spectra of the catalysts showed two white line peaks at 10375 and 10379 eV (Figure 5), which correspond to tetrahedral Ga (Ga(t)) and octahedral Ga (Ga(o)), respectively.30,31 For quantitative analysis, the XANES spectra were deconvoluted with two Gaussian curves for the white line and two arctangent curves for continuum absorption. The percentages of Ga(t) and Ga(o) were estimated from the ratio of the Gaussian peak areas at 10375 and 10379 eV, respectively.30,31 The results showed that the fractions of Ga(t) of all catalysts were in the range of 66− 68% before and after 20 propane and isobutane dehydrogenation/regeneration cycles (Table S2 in the Supporting Information). These results imply that there were only slight changes in Ga content and states after the dehydrogenation reaction/regeneration cycles, and thus they cannot explain the significant differences in the catalytic performance of the three catalysts. On the other hand, HAADF-STEM revealed that there are substantial differences in the Pt cluster sizes between the

Figure 3. HAADF-STEM images of (a) N-PtGa, (b) P-PtGa, (c) PCPtGa, and (d) Pt L3-edge XANES spectra of the catalysts. The insets in (a)−(c) show the histograms of the Pt particle size distributions. The surface-area-weighted cluster diameter, dTEM, was calculated from dTEM = ∑nidi3/∑nidi2.32

Figure 4. (a) Propane conversions and (b) propene selectivities over the 20 repeated propane dehydrogenation/catalyst regeneration cycles (reaction conditions: 893 K, 20 kPa propane, 80 kPa He, WHSV = 5.4 h−1; catalyst regeneration 893 K, He flow 30 min, air flow 30 min, He flow 30 min). (c) Isobutane conversions and (d) isobutene selectivities over the 20 repeated isobutane dehydrogenation/catalyst regeneration cycles (reaction conditions: 823 K, 20 kPa isobutane, 80 kPa He, WHSV = 7.1 h−1; catalyst regeneration 823 K, He flow 30 min, air flow 30 min, He flow 30 min). In the reaction/regeneration of PtSn/γAl2O3 (PtSn), 30 min of an H2 reduction step was added before each reaction.3

regenerated by simple oxidation in comparison to the PtSn/γAl2O3 catalyst, presumably because the main active site is GaOx rather than Pt.15 2822

DOI: 10.1021/acscatal.6b00329 ACS Catal. 2016, 6, 2819−2826

Research Article

ACS Catalysis

and stability even in comparison with a commercialized catalyst, specifically PtSn supported on γ-Al2O3. The Pt stabilizing strategies reported in this work may be applied to other metalcatalyzed reactions that require severe reaction/regeneration conditions.

■

EXPERIMENTAL SECTION Catalyst Preparation. P-PtGa and PC-PtGa were prepared by a one-pot sol−gel method. In a typical synthesis of PC-PtGa with 1 wt % Ce doping, 0.01 g of Pt(NH3)4(NO3)2 (Aldrich, >99.9%), 0.66 g of Ga(NO3)3 (Aldrich, >99.9%), and 0.19 g of Ce(NO3)3·6H2O (Samchun, >99.0%) were dissolved in 212 g of deionized H2O and homogenized at 358 K for 10 min. A 24 g amount of aluminum isopropoxide (AIP, Aldrich, >98.0%) was added to the solution with vigorous stirring at the same temperature. After the hydrolysis of AIP for 30 min, 2.4 g of concentrated nitric acid (HNO3, 61%) was added to the solution and the temperature maintained at 358 K with stirring. After H2O was fully evaporated, a mixed hydroxide powder was obtained. The resultant solid was calcined in dry air (flow rate 200 mL min−1 g−1) at 1023 K for 2 h. The P-PtGa catalyst was prepared in a similar way, except that Ce(NO3)3·6H2O was not added in the initial step. N-PtGa was prepared via conventional incipient wetness of an aqueous solution containing Pt(NH3)4(NO3)2 and Ga(NO3)3 into the commercial γ-Al2O3 (Strem Chemicals). In a typical synthesis, 6 g of γ-Al2O3 was impregnated with 3.6 mL of an aqueous solution dissolving 0.01 g of Pt(NH3)4(NO3)2 and 0.66 g of Ga(NO3)3. The impregnated sample was dried at 373 K for 24 h and calcined in dry air at 1023 K (200 mL min−1 g−1) for 2 h. PtSn/γ-Al2O3 (0.7 wt % Pt, 0.3 wt % Sn, and 0.8 wt % K) was synthesized by incipient wetness coimpregnation of the aqueous solution dissolving H2PtCl6 (Kojima Chemicals), SnCl2·2H2O (Aldrich), and KNO3 (Junsei) on a commercial γ-Al2O3 (Strem Chemicals). The impregnated sample was dried at 373 K for 24 h and calcined in dry air at 1023 K (200 mL min−1 g−1) for 2 h. Characterization. Elemental analyses were carried out by using ICP-AES using iCAP-6500 (Thermo). XRD patterns were recorded using a D2-phaser (Bruker) equipped with Cu Kα radiation (30 kV, 10 mA) and LYNXEYE detector. N2 adsorption−desorption isotherms were measured using a BELSorp-max (BEL Japan) volumetric analyzer at 77 K after degassing at 673 K. The BET areas were determined in a P/P0 range between 0.05 and 0.20. Pore size distributions were determined by using the Barrett−Joyner−Halenda (BJH) method. X-ray absorption spectra of Pt L3 edge and Ga K edge were measured in a transmission mode at Pohang Accelerator Laboratory (8C-Nano XAFS beamline). For XAFS measurements, 0.3 g of the calcined catalyst was pressed into self-supporting pellets (13 mm in diameter) at 200 bar. XANES spectra were acquired without pretreatment. After background removal using ATHENA, a deconvolution analysis of Ga Kedge XANES spectra was carried out by following the method reported in an earlier study.30,31 HAADF-STEM images were taken with Titan Cubed G2 60-300 (FEI Co.) at 300 kV after the samples were mounted on a copper grid (300 square mesh) using an ethanol dispersion. Pt particle size distributions were determined by counting at least 150 crystallites using the “ImageJ” program (Wayne Rasband, National Institutes of Health, USA). The surface-area-weighted mean cluster diameter was determined using HAADF STEM analysis and calculated from dTEM = ∑nidi3/∑nidi2, where ni is the number of crystallites having the diameter di.32 The TEM image of

Figure 5. Ga K-edge XANES spectra of (a, b) N-PtGa and (c, d) PCPtGa before (a, c) and after 20 cycles of propane dehydrogenation/ regeneration (b, d). XANES spectra were deconvoluted with two Gaussian curves (solid line) for the white line and with two arctangent curves (dashed line) for continuum absorption. White line peaks at 10375 and 10379 eV correspond to tetrahedral Ga (red) and octahedral Ga (blue), respectively.

catalysts after 20 cycles of propane and isobutane dehydrogenation/regeneration (Figures 6 and 7a). As described above, all of the catalysts possessed highly dispersed Pt clusters (dTEM = 1.1 nm) in their fresh states. However, after 20 cycles of propane dehydrogenation/regeneration, the catalysts showed evident Pt sintering. The degree of sintering increased in the order PC-PtGa (dTEM = 4.4 nm) < P-PtGa (dTEM = 8.2 nm) < N-PtGa (dTEM = 18 nm). Consistent with the HAADF-STEM analysis, CO chemisorption decreased in the order PC-PtGa > P-PtGa > N-PtGa (Figure 7b). As in the propane dehydrogenation, the catalysts also showed a similar sintering trend in isobutane dehydrogenation (Figures 6 and 7). The only difference is that Pt sintering is more severe in propane dehydrogenation because the reaction/regeneration temperature (893 K) is higher than that of isobutane dehydrogenation (823 K). After the isobutane dehydrogenation/regeneration cycles, PC-PtGa, P-PtGa, and N-PtGa showed average Pt cluster diameters (dTEM) of 2.4, 3.4, and 12 nm, respectively (Figures 6 and 7a). The results clearly revealed that the physical stabilization of the Pt catalyst can significantly increase the Pt sintering resistance under the harsh reaction/regeneration cycles. In addition, the synergistic combination of physical stabilization with Ce doping (chemical stabilization) could further stabilize the Pt catalyst against sintering. Therefore, the rapid deterioration of the catalytic performance of the N-PtGa catalyst in the initial reaction cycles and the relatively much more stable performance of the P-PtGa and PC-PtGa catalysts could be unequivocally attributed to the different Pt sintering behaviors of the catalysts.

■

CONCLUSIONS In summary, a PtGa/γ-Al2O3 catalyst with Pt stabilized by both physical and chemical methods showed significantly enhanced activity, light olefin selectivity, and long-term stability during the repeated dehydrogenation/regeneration cycles. We confirmed that the catalyst showed remarkably enhanced activity 2823

DOI: 10.1021/acscatal.6b00329 ACS Catal. 2016, 6, 2819−2826

Research Article

ACS Catalysis

Figure 6. HAADF-STEM images of (a−c) N-PtGa, (d−f) P-PtGa, and (g−i) PC-PtGa before reaction (a, d, g), after 20 cycles of isobutane dehydrogenation/regeneration (b, e, h), and after 20 cycles of propane dehydrogenation/regeneration (c, f, i).

(SKM) function, expressed by FSKM = (1 − R)2/2R = K/S, was used to represent spectra (R is the reflectance and K and S are the absorption and diffusion coefficients, respectively). As a reference, the UV−vis spectrum of commercial CeO2 (Junsei, >99.0%) was also recorded. The band gap width (Eg) of CeOx was determined from the Planck equation, Eg = hc/λ = 1240/λ, where h, c, and λ are Planck’s constant, the speed of light, and the absorption edge wavelength, respectively. X-ray photoelectron spectroscopy (XPS) measurements were carried out using a Kα (Thermo VG Scientific) analyzer equipped with a microfocused monochromator X-ray source. The binding energies were calibrated on the basis of the C 1s peak at 284.6 eV. TPR profiles were collected using a Belcat instrument (BEL Japan) equipped with a thermal conductivity detector (TCD). The temperature was increased from room temperature to 1073 K with a ramping rate of 10 K min−1 under 5% H2/Ar (flow rate 30 mL min−1). Catalytic Measurements. All of the reactions were carried out in a quartz plug-flow reactor (inner diameter 13 mm) connected to an online gas chromatograph. The catalysts were pressed into pellets without any diluents at 400 bar, crushed, and sieved (75−100 mesh) before the catalytic measurements. The catalysts were pretreated at the reaction temperature under

commercial CeO2 (Junsei, >99.0%) was collected with a Tecnai F20 instrument (Tecnai) at 200 kV acceleration voltage. Pt dispersions were determined using repeated CO chemisorptions, which were measured with an ASAP2020 instrument (Micromeritics). In the chemisorption experiments, CO was used as an adsorbate because the experiments with H2 were not highly reliable due to the extremely small Pt loading (0.1 wt %) and relatively much faster leaking speed of H2 in comparison to CO in the adsorption instrument. Before measurements, all of the samples were re-reduced at 723 K under H2 for 3 h, followed by evacuation for 3 h at the same temperature. Adsorption measurements were carried out at 323 K in the pressure range of 0.3−60 kPa. After the first adsorption was measured, the samples were evacuated for 1 h at 323 K, after which a second isotherm was measured under the same conditions. CO chemisorption amounts were determined from the difference between two adsorption isotherms extrapolated to zero pressure assuming a stoichiometric factor of unity. For the accuracy of the measurements, the CO chemisorption experiment was repeated three times for each sample. The UV−vis diffuse reflectance spectra (DRS) were recorded on a Lambda 1050 (PerkinElmer) spectrometer equipped with a diffuse reflectance attachment. The Schuster−Kubelka−Munk 2824

DOI: 10.1021/acscatal.6b00329 ACS Catal. 2016, 6, 2819−2826

Research Article

ACS Catalysis

Basic Science Research Program through the National Research Foundation of Korea (NRF-2011-0011392) and Advanced Biomass R&D Center (ABC) of Global Frontier Project (ABC2015M3A6A2066121) funded by the Ministry of Science, ICT & Future Planning. XANES experiments at PLS-II were supported in part by the MSIP and POSTECH.

■

(1) Sattler, J. J. H. B.; Ruiz-Martinez, J.; Santillan-Jimenez, E.; Weckhuysen, B. M. Chem. Rev. 2014, 114, 10613−10653. (2) Bhasin, M. M.; McCain, J. H.; Vora, B. V.; Imai, T.; Pujado, P. R. Appl. Catal., A 2001, 221, 397−419. (3) Iglesias-Juez, A.; Beale, A. M.; Maaijen, K.; Weng, T. C.; Glatzel, P.; Weckhuysen, B. M. J. Catal. 2010, 276, 268−279. (4) de Miguel, S. R.; Castro, A. A.; Scelza, O. A. Catal. Lett. 1996, 36, 201−206. (5) Siri, G. J.; Ramallo-Lopez, J. M.; Casella, M. L.; Fierro, J. L. G.; Requejo, F. G.; Ferretti, O. A. Appl. Catal., A 2005, 278, 239−249. (6) Shi, L.; Deng, G.-M.; Li, W.-C.; Miao, S.; Wang, Q.-N.; Zhang, W.-P.; Lu, A.-H. Angew. Chem., Int. Ed. 2015, 54, 13994−13998. (7) Weckhuysen, B. M.; Schoonheydt, R. A. Catal. Today 1999, 51, 223−232. (8) Puurunen, R. L.; Weckhuysen, B. M. J. Catal. 2002, 210, 418− 430. (9) Hakuli, A.; Kytokivi, A.; Krause, A. O. I. Appl. Catal., A 2000, 190, 219−232. (10) Kanazirev, V. I. Wide mesoporous alumina composites having trimodal pore structure. U.S. Patent 7,790,130, Sept 7, 2010. (11) Saito, M.; Watanabe, S.; Takahara, I.; Inaba, M.; Murata, K. Catal. Lett. 2003, 89, 213−217. (12) Blasco, T.; Nieto, J. M. L. Appl. Catal., A 1997, 157, 117−142. (13) Chen, K.; Xie, S.; Iglesia, E.; Bell, A. T. J. Catal. 2000, 189, 421− 430. (14) Chen, M.; Xu, J.; Cao, Y.; He, H. − Y.; Fan, K. − N.; Zhuang, J. − H. J. Catal. 2010, 272, 101−108. (15) Sattler, J. J. H. B.; Gonzalez-Jimenez, I. D.; Luo, L.; Stears, B. A.; Malek, A.; Barton, D. G.; Kilos, B. A.; Kaminsky, M. P.; Verhoeven, T. W. G. M.; Koers, E. J.; Baldus, M.; Weckhuysen, B. M. Angew. Chem., Int. Ed. 2014, 53, 9251−9256. (16) Siddiqi, G.; Sun, P.; Galvita, V.; Bell, A. T. J. Catal. 2010, 274, 200−206. (17) Filez, M.; Redekop, E. A.; Poelman, H.; Galvita, V. V.; Ramachandran, R. K.; Dendooven, J.; Detavernier, C.; Marin, G. B. Chem. Mater. 2014, 26, 5936−5949. (18) Zheng, B.; Hua, W.; Yue, Y.; Gao, Z. J. Catal. 2005, 232, 143− 151. (19) Jablonski, E. L.; Castro, A. A.; Scelza, O. A.; de Miguel, S. R. Appl. Catal., A 1999, 183, 189−198. (20) Chen, M.; Xu, J.; Su, F. − Z.; Liu, Y. − M.; Cao, Y.; He, H. − Y.; Fan, K. − N. J. Catal. 2008, 256, 293−300. (21) Otroshchenko, T.; Sokolov, S.; Stoyanova, M.; Kondratenko, V. A.; Rodemerck, U.; Linke, D.; Kondratenko, E. V. Angew. Chem., Int. Ed. 2015, 54, 15880−15883. (22) Waku, T.; Biscardi, J. A.; Iglesia, E. Chem. Commun. 2003, 14, 1764−1765. (23) Cortright, R. D.; Hill, J. M.; Dumesic, J. A. Catal. Today 2000, 55, 213−223. (24) Cho, I. H.; Park, S. B.; Cho, S. J.; Ryoo, R. J. Catal. 1998, 173, 295−303. (25) Hatanaka, M.; Takahashi, N.; Tanabe, T.; Nagai, Y.; Dohmae, K.; Aoki, Y.; Yoshida, T.; Shinjoh, H. Appl. Catal., B 2010, 99, 336− 342. (26) Hatanaka, M.; Takahashi, N.; Takahashi, N.; Tanabe, T.; Nagai, Y.; Suda, A.; Shinjoh, H. J. Catal. 2009, 266, 182−190. (27) Zaki, M. I.; Hussein, G. A. M.; Mansour, S. A. A.; Ismail, H. M.; Mekhemer, G. A. H. Colloids Surf., A 1997, 127, 47−56. (28) Bensalem, A.; Bozon-Verduraz, F.; Delamar, M.; Bugli, G. Appl. Catal., A 1995, 121, 81−93.

Figure 7. (a) Surface-area-weighted cluster diameter, dTEM, of Pt clusters and (b) CO chemisorptions of the catalysts before and after the 20 repeated propane and isobutane dehydrogenation/regeneration cycles. To ensure accuracy of the measurement, CO chemisorptions were repeated three times and averaged.

He. The reaction conditions are 893 K, 20 kPa propane, 80 kPa He, and WHSV = 5.4 h−1 for propane dehydrogenation and 823 K, 20 kPa isobutane, 80 kPa He, and WHSV = 7.1 h−1 for isobutane dehydrogenation. After 1 h of reaction, the coked catalysts were regenerated at the reaction temperature for 30 min under dry air (20% O2 in N2). Between the reaction and regeneration processes, the catalysts were flushed with He for 30 min. In the reaction/regeneration of PtSn/γ-Al2O3, 30 min of an H2 reduction step was added before each reaction. Before and after reduction step, the catalyst was flushed with He for 30 min. Without the reduction steps, the catalyst showed even inferior catalytic performances in both reactions. It should be noted that, in this experiment, PtSn/γ-Al2O3 catalyst showed a very poor recyclability because the catalyst was regenerated by a simple oxidation rather than an industrially used oxychlorination process.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acscatal.6b00329. Tables S1 and S2 and Figures S1−S7 as described in the text (PDF)

■

REFERENCES

AUTHOR INFORMATION

Corresponding Author

*E-mail for M.C.: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS SK Innovation actively participated in scientific discussions and supported the research. This work was also supported by the 2825

DOI: 10.1021/acscatal.6b00329 ACS Catal. 2016, 6, 2819−2826

Research Article

ACS Catalysis (29) Stakheev, A. Y.; Kustov, L. M. Appl. Catal., A 1999, 188, 3−35. (30) Shimizu, K.; Takamatsu, M.; Nishi, K.; Yoshida, H.; Satsuma, A.; Tanaka, T.; Yoshida, S.; Hattori, T. J. Phys. Chem. B 1999, 103, 1542− 1549. (31) Nishi, K.; Shimizu, K.; Takamatsu, M.; Yoshida, H.; Satsuma, A.; Tanaka, T.; Yoshida, S.; Hattori, T. J. Phys. Chem. B 1998, 102, 10190−10195. (32) Choi, M.; Wu, Z.; Iglesia, E. J. Am. Chem. Soc. 2010, 132, 9129− 9137.

2826

DOI: 10.1021/acscatal.6b00329 ACS Catal. 2016, 6, 2819−2826