Comprehensive Understanding of the Effects of Carbon

---

Subscriber access provided by NEW YORK UNIV

Article

Comprehensive Understanding of the Effects of Carbon Nanostructures on Redox Catalytic Properties and Stability in Oxidative Dehydrogenation Han Chang Kwon, Sunwoo Yook, Seokin Choi, and Minkee Choi ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.7b01742 • Publication Date (Web): 27 Jun 2017 Downloaded from http://pubs.acs.org on June 27, 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 44

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

Comprehensive Understanding of the Effects of Carbon Nanostructures on Redox Catalytic Properties and Stability in Oxidative Dehydrogenation

Han Chang Kwon, Sunwoo Yook, Seokin Choi, and Minkee Choi* Department of Chemical and Biomolecular Engineering, Korea Advanced Institute of Science and Technology, Daejeon 34141, Republic of Korea

ABSTRACT: The intrinsic redox catalytic properties of metal-free carbons have been widely investigated due to fundamental interest as well as potential practical applications. Although a large variety of nanostructured carbons are now available, the effects of carbon nanostructures on redox properties have not been comprehensively understood. In this work, the redox catalytic properties and thermochemical stabilities of 16 different types of carbons including activated carbon, carbon nanotubes, onion-like carbons, and microporous/mesoporous templated carbons were systematically investigated using n-butane oxidative dehydrogenation as a model reaction. The results demonstrate that the overall catalytic activity increases with increasing content of C=O active sites. However, with increasing C=O content, the activity per site (i.e., turnover frequency) gradually decreases, while the alkene selectivity increases due to the decreased reducibility of each C=O site. Since more C=O sites are present in a thermochemically less stable amorphous framework, the carbons generally exhibit a trade-off relationship between catalytic activity and stability. However, a graphitic carbon with ‘coin-stacking’ carbon layers showed exceptionally high activity and stability simultaneously. This is attributed to its unique carbon structure that simultaneously provides high graphitic order and abundant carbon edge sites where C=O active sites are grafted.

KEYWORDS: Carbon, Nanostructure, Oxidative Dehydrogenation, Reducibility, Stability

ACS Paragon Plus Environment

1

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 2 of 44

INTRODUCTION Carbons have been widely used as a support for metal catalysts in heterogeneous catalysis.1,2 In recent years, the intrinsic catalytic properties of metal-free carbons have also been studied due to fundamental interest as well as potential practical applications.1-5 In particular, oxidative dehydrogenation (ODH) of hydrocarbons has been investigated most extensively in order to understand the intrinsic redox functions of the carbon surface.4 Various types of nanostructured carbons such as activated carbons,6,7 carbon nanotubes (CNTs),8-13 carbon nanofibers (CNFs),14 onion-like carbons (OLCs),15 nanodiamonds,16 few-layered graphenes,17,18 and mesoporous carbons19,20 have shown promising catalytic activity and alkene selectivity in ODH reactions. Notably, some carbon materials have exhibited remarkable stability even under the harsh oxidative reaction conditions of ODH.9,10,16,21 Earlier studies using model catalysts,22,23 in situ X-ray photoelectron spectroscopy (XPS),9 and titrations12 clearly demonstrated that quinone-type C=O functional groups are the catalytic active sites responsible for activating the C–H bonds of hydrocarbons during ODH reactions through a redox cycle involving quinone-hydroquinone interconversion. Inherently, the number and types of oxygen functional groups in carbon materials are determined by various synthesis parameters such as the carbon precursor, carbonization temperature, and post-synthesis modification process (e.g., partial oxidation using various chemical oxidants).1,2 Under the oxidative reaction conditions of ODH, the carbon surface is also oxidized in situ, and additional oxygen functional groups are generated after certain induction periods.14 Compared with other catalytic materials such as metals and metal oxides, carbons can be synthesized to have a much wider range of nanostructures and surface functional groups due to recent developments in carbon material chemistry.2 Although a large variety of nanostructured carbons with different framework structures are available (e.g., activated carbon, CNT, OLC, templated carbons, etc.), a rather limited number of carbon types has been directly compared under the same ODH condition in previous studies. Depending on the study, the ODH of different hydrocarbons such as ethane,11 propane,10 butane,9,16-20 and ethylbenzene6-8,12-15,21-23 has been investigated under a wide range of reaction conditions

ACS Paragon Plus Environment

2

Page 3 of 44

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

(different reaction temperatures and O2/hydrocarbon molar ratios), which makes a direct comparison of the various carbon catalysts impossible. This makes it difficult to comprehensively understand the effect of the carbon nanostructure on redox catalytic properties and thermochemical stabilities. Such information is important not only for the rational design of advanced ODH carbon catalysts, but also for other catalytic applications of carbon-based materials. In the present work, to comprehensively understand the effects of carbon nanostructures on their intrinsic redox catalytic properties and thermochemical stabilities, various types of carbons, including activated carbon, CNTs, OLCs, and microporous/mesoporous templated carbons24-30 (Scheme 1) were investigated in n-butane ODH. Important structural parameters determining the catalytic activity, selectivity, and stability of the carbons are discussed. This study also reveals that a graphitic mesoporous carbon with a unique ‘coin-stacking’ graphitic alignment (i.e., a stacking of discoid carbon layers perpendicular to the direction of the carbon framework, Scheme 1e) has exceptionally high catalytic activity, selectivity, and stability simultaneously.

RESULTS AND DISCUSSION Synthesis and Characterization of the Carbon Materials. Various nanostructured carbon materials, including activated carbon, multi-walled CNTs, OLCs, and microporous/mesoporous templated carbons (Scheme 1), were prepared and investigated in n-butane ODH reaction. The activated carbon, CNT, and OLC were washed with an aqueous HCl solution to remove metal impurities. Elemental analysis using inductively coupled plasma optical emission spectrometry (ICP-OES) confirmed that all of these carbon materials were essentially metal-free (< 0.1% Fe, Table S1). The samples were denoted as ‘AC’, ‘CNT, and ‘OLC’, respectively. A microporous zeolite-templated carbon (denoted as ‘ZTC’) was synthesized by acetylene chemical vapor deposition in the micropores of a NaX zeolite followed by the dissolution of the zeolite framework with an aqueous HCl/HF solution.24-26 Hexagonally ordered mesoporous carbons (CMK-3-type materials28-30) were synthesized by the carbon replication of an ordered mesoporous silica (SBA-15) using sucrose and acenaphthene as two different carbon precursors. The use of high-oxygen-

ACS Paragon Plus Environment

3

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 4 of 44

content sucrose as a carbon precursor has been reported to generate mesoporous carbon with a highly amorphous framework.28 On the other hand, the use of the polyaromatic acenaphthene as the carbon precursor has been reported to produce a mesoporous carbon with a graphitic framework.30 The carbon materials synthesized from sucrose and acenaphthene were denoted as ‘CMK-3’ and ‘CMK-3G’, respectively. The N2 adsorption-desorption isotherms and corresponding pore structural properties of the carbon catalysts are provided in Figure S1 and Table S2. The pore size distributions estimated by a nonlinear density functional theory (NLDFT) method are also provided in Figure S2. The powder X-ray diffraction (XRD) patterns of the various carbon materials are shown in Figure 1. Most of the carbon materials exhibit a major peak at around 2θ = 26°, which can be attributed to the (002) diffraction of the graphitic structure (i.e., stacking of the two-dimensional carbon layers). The sharpness of this peak increased in the order of ZTC < CMK-3 < AC < CMK-3G ≈ OLC < CNT. The result indicates that CNT has the highest graphitic stacking order of the carbon layers. The transmission electron microscopy (TEM) image in Figure 2a confirms the highly graphitic arrangement of the carbon layers in the cylindrical carbon walls of the CNT. Similar to the CNT, the OLC also exhibits highly graphitic stacking of the carbon layers in its concentric carbon shells (Figure 2b). The relatively broader (002) diffraction of the OLC compared to the CNT appears to originate from its smaller graphitic domain size (i.e., a smaller number of stacked carbon layers) due to the Scherrer formula.31 Compared with the CNT and OLC, the AC is composed of a more amorphous carbon framework (Figure 2c), which is consistent with the XRD analysis. As revealed by the low-angle XRD (Figure S3) and TEM images (Figures 2d and e), the ordered mesoporous carbons, CMK-3 and CMK-3G, showed hexagonal arrangements of carbon rods with mesopores between them. As confirmed by the broad, wide-angle XRD peaks (Figure 1) and the TEM image (Figure 2d), CMK-3 has a completely amorphous carbon framework. In contrast, CMK-3G has a much more aligned arrangement of carbon layers (i.e., more graphitic) resulting in sharper wide-angle XRD peaks (Figure 1), although the graphitic order is somewhat lower than those of the CNT and OLC according to the TEM image (Figure 2e). Notably, the discoid carbon layers are stacked perpendicular to

ACS Paragon Plus Environment

4

Page 5 of 44

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 direction of the carbon framework (‘coin-stacking’ of the carbon layers) in CMK-3G, which is consistent with an earlier report.30 Electron diffraction patterns (insets in Figures 2d and e) also confirm that CMK-3 has an amorphous framework, whereas CMK-3G is composed of highly ordered stacking of carbon layers. The amorphous CMK-3 possesses a significantly larger microporosity (0.36 mL g-1, Table S2) and an apparent Brunauer-Emmett-Teller (BET) surface area (1130 m2 g-1) than those of the solely mesoporous graphitic CMK-3G (0.02 mL g-1 and 504 m2 g-1). Both materials possess similar mesoporosity (0.42 – 0.51 mL g-1). ZTC, synthesized by the carbon replication of a zeolite, exhibited extremely weak (002) diffraction in the wide-angle XRD pattern (Figure 1), indicating the absence of graphitic stacking in this material. As shown schematically in Scheme 1f, ZTC is composed of highly curved three-dimensional networks of graphene nanoribbons.24,26 Carbon deposition within the confined micropores of a zeolite does not allow ordinary graphitic stacking of the carbon layers. It has been reported that even the extremely weak (002) diffraction is due to the formation of minor carbon deposits on the external surfaces of the zeolite crystallites.25 On the other hand, ZTC exhibits a sharp XRD peak at 2θ = 6° (Figure 1), indicating the presence of structural regularity with a periodicity of about 1.4 nm. This is due to the ordered microporous structure of the ZTC generated by the replication of the NaX zeolite.24,25 The ordered micropore arrangement is clearly visible in the TEM image (Figure 2f). Due to its unique structure, ZTC has the largest microporosity (1.2 mL g-1) and apparent BET surface area (2600 m2 g-1) among all the carbon samples. It was reported that post-synthesis oxidation of highly graphitic carbon by HNO3 can result in enhanced ODH activity due to the increased generation of carbon defect structures and oxygen functional groups that can act as catalytic active sites for ODH.13 In this respect, the graphitic CNT and OLC samples were additionally oxidized by HNO3 treatment at a variety of temperatures (50 to 140 °C). The HNO3-treated samples are referred to as ‘oCNT-t’ and ‘oOLC-t’, where t indicates the treatment temperature. As shown in the TEM images (Figure S4), the graphitic carbon walls of the CNT were partly exfoliated and unzipped by the HNO3 treatment, the degrees of which increased with increasing treatment

ACS Paragon Plus Environment

5

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 6 of 44

temperature (t). Likewise, the concentrically stacked carbon shells of OLC were similarly damaged and showed an increased number of defect structures after HNO3 treatment.

n-Butane ODH on Nanostructured Carbons. The redox catalytic properties of the prepared carbon materials were investigated in n-butane ODH. To check the activation/deactivation behavior of the carbon catalysts under accelerated conditions, the reactions were carried out with an O2-rich reactant composition (O2/n-butane ratio of 2) and under a fixed space velocity (WHSV = 2.9 h-1) at 450 °C (Figure 3). The alkene formation rates of the highly graphitic CNT (Figure 3a) and OLC (Figure 3b) were initially very low but slowly increased up to certain reaction periods (~30 h). After induction periods, the catalytic activities of these catalysts were maintained for a period of time, after which they became gradually deactivated. These induction periods have been attributed to the in situ generation of defects and C=O active sites in the carbon structure during the ODH reaction.14 In the cases of the HNO3-treated oCNT-t (Figure 3a) and oOLC-t (Figure 3b) samples, the induction periods became shorter with increasing HNO3 treatment temperature (t). Other carbon materials, including the AC, ZTC, CMK-3, and CMK-3G samples (Figure 3c), exhibited no appreciable induction periods, indicating that these carbon materials initially possess sufficient catalytic active sites (C=O), or these sites are readily produced in the very initial stages of the reaction. The different activation behavior between the CNT-/OLC-type materials and other carbons could be attributed to the different initial availabilities of carbon edge sites where C=O groups can be grafted. The ratio between the edges and basal planes of a carbon catalyst was determined by the ratio of the areas of the D1 and G bands (ID1/IG) in the Raman spectra.32 The perfectly graphitic CNT and OLC had the smallest ID1/IG values among the carbon samples (Figures S5a and b), indicating the lowest availability of carbon edge sites. The ID1/IG ratios of the CNT-/OLC-type materials gradually increased with increasing harshness of the HNO3-treatment due to the generation of defect structures. Other carbons, including AC, ZTC, CMK-3, and CMK-3G (Figure S5c), showed high ID1/IG values, indicating highly abundant edge sites. These results indicate that the carbons initially containing abundant edge sites can be readily

ACS Paragon Plus Environment

6

Page 7 of 44

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

activated during the reaction, resulting in short induction periods. After induction periods, the carbons started to be deactivated with time-on-stream. The catalyst mass substantially decreases after the longterm reactions, indicating that catalyst deactivation is due to the carbon combustion. Notably, the carbon catalysts generally exhibited trade-off relationship between the initial alkene formation rate and catalyst stability; carbon materials with higher initial alkene formation rates also became deactivated more rapidly (Figure 3). In ODH reactions, alkene selectivity generally decreases with increasing alkane conversion due to enhanced over-oxidation. Therefore, for a comparison of the intrinsic activity and selectivity of the different carbon catalysts, catalytic data should be collected at various conversion levels.33 Accordingly, the n-butane ODH reactions with all carbon catalysts were also carried out at different WHSVs (Figures S6-S8). n-Butane conversions were typically below 10% (Figure S9). The n-butane conversion and alkene selectivity measured at a fixed 1 h time-on-stream were extrapolated to zero residence time (1/WHSV)9,20 in order to obtain primary alkene formation rate and selectivity. The results are summarized in Table 1.

Correlation between Carbon Nanostructures and Redox Catalytic Properties. Earlier studies using model catalysts,22,23 in situ XPS,9 and titrations12 clearly demonstrated that quinone-type C=O groups are the redox active sites for ODH reactions. With this regard, we carried out quantitative analysis of the oxygen functional groups in the carbon catalysts with O 1s X-ray photoelectron spectroscopy (XPS). Because substantial amounts of oxygen functional groups can be generated in situ during the initial period of the reaction,14 we analyzed the carbon catalysts by XPS after they had been used in ODH reactions for 1 h, and correlated the results with catalytic data obtained at the same time (Table 1). The O 1s spectra (Figure S10) were deconvoluted with peaks for C=O (531.4 eV), O=C–O (532.5), and C–O (533.7 eV).13,34 The C=O group densities are summarized in Table 1. Among the various carbon catalysts, ZTC exhibited the largest quantity of C=O groups (2.0 mmol g1

) due to its curved three-dimensional graphene nanoribbon structure (Scheme 1f) that provides the

maximum number of edge sites where functional groups can be located.24,26 On the other hand, the

ACS Paragon Plus Environment

7

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 8 of 44

amounts of C=O groups in the highly graphitic CNT and OLC were very small (0.18 and 0.15 mmol g-1, respectively). As mentioned above, this can be attributed to the fact that the CNT and OLC consist of concentric tubes and shells of graphenes (Schemes 1a and b). They possess highly accessible carbon basal planes but have only a limited number of edge sites only at the defects35,36. The numbers of C=O groups in the CNT and OLC samples, however, significantly increased after HNO3 treatment (oCNT-t and oOLC-t samples) due to the additional generation of defect structures. HNO3 treatment at higher temperature (t) led to a larger amount of C=O groups (up to 1.2 mmol g-1). The carbon materials possessing amorphous frameworks, such as AC and CMK-3, contain abundant C=O functional groups (~1.4 mmol g-1) even without the post-synthesis HNO3 treatment. These results indicate that carbon that is less graphitic generally possesses more C=O functional groups. It is remarkable, however, that the fairly graphitic CMK-3G exhibited a large number of C=O functional groups (1.0 mmol g-1), which is slightly smaller than amorphous AC and CMK-3, but remarkably larger than those of graphitic CNT and OLC. The large amount of C=O groups in CMK-3G, despite its high graphitic order, can be attributed to its unique framework structure composed of ‘coin-stacking’ carbon layers that provide a large number of edge sites (Scheme 1e). As shown in Figure 4, the C=O functional group density and ID1/IG ratio in Raman spectra (Figure S5) are linearly proportional to each other, verifying that C=O functional groups are indeed predominantly generated at edge sites of carbons. As shown in Figure 5a, the n-butane conversion rate increases monotonically with the amount of C=O groups in the carbons. It is notable, however, that the trend is not completely linear. The nonlinear correlation is attributed to the fact that the catalytic activity per C=O group (i.e., turnover frequency, TOF) gradually decreases with increasing C=O functional group density (Figure 5b). Earlier studies demonstrated that linear regressional analysis provides a good fit between the alkane conversion rate and the C=O group density, although the fitting line did not pass through the origin (i.e., positive yintercept).6,12 However, the positive y-intercept, which suggests ‘catalytic activity in the absence of active sites’, is difficult to scientifically rationalize unless alternative catalytic active sites are postulated. Considering the fact that the earlier studies used a small number of carbon samples with a relatively

ACS Paragon Plus Environment

8

Page 9 of 44

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

narrow range of C=O densities,6,12 the reported trends are actually in agreement with the present results. On the other hand, the primary alkene selectivity determined at zero residence time (Table 1) monotonically increases with increasing C=O density of the carbon catalysts (Figure 5c). These results indicate that, at higher C=O group densities, each C=O group exhibits a lower TOF but higher alkene selectivity. In metal oxide catalysts, it has long been postulated that catalyst reducibility is one of the most important factors in determining ODH activity and selectivity.37 Catalysts that are highly reducible generally exhibit high activity but low alkene selectivity.38-42 In other words, an oxide that is too difficult to reduce is also too inactive, while one that is easily reduced is active but non-selective. Bell and Iglesia pointed out that the kinetically relevant step in an ODH reaction is the C–H bond activation step (H abstraction) and that high catalyst reducibility increases the TOF.40 Madeira et al. showed that the activation energies for n-butane conversion decreased with increasing catalyst reducibility.43 The alkene selectivity decreased with increasing reducibility, indicating that catalysts that are too highly reducible can catalyze C–C bond activations that can lead to combustion reactions. We believe that a similar argument can be applied to carbon catalysts because earlier reports showed that the presence of oxygen functional groups significantly alter the electronic structures of carbon materials,44,45 which affect their reducibility. With this regard, the reducibility of each carbon catalysts, collected after 1 h ODH reaction, was investigated by temperature-programmed reduction (TPR) under H2. It should be noted that the measurements of TPR profiles of carbons requires special caution because various oxygen functional groups on the carbon surface can be thermally decomposed as H2O, CO, and CO2 during temperature ramping.3,46 Fortunately, quinone-type C=O active sites have the highest thermal stability (decompose only above 700 °C3,46) among the various oxygen functional groups, and thus, their H2-reductions could still be investigated before their thermal decomposition. In the present TPR experiments, the gases (H2O, CO, and CO2) generated by thermal decomposition were completely removed by adsorption with a specially prepared trap containing a mixture of CuCl-modified activated carbon47 and a zeolite (13X

ACS Paragon Plus Environment

9

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 10 of 44

molecular sieve) (Figure S11) so that H2 consumption could be selectively detected using a typical thermal conductivity detector. The TPR profiles of the carbon materials exhibit single H2 consumption peaks in the 575 – 700 °C range (Figure S12). The H2 consumption peak area increases with the amount of C=O functional groups (H2/C=O ~ 0.59, Figure S13a). Notably, the reduction temperature generally shifted toward higher temperatures as the C=O content of the carbon catalysts increased (Figure 5d and Table S2). The results verify that individual C=O group reducibility decreases with increasing C=O density on the carbon catalysts. We also performed TPR experiments after thermal treatment of the carbon catalysts at 700 °C under He to remove the majority of the other oxygen functional groups that are less thermally stable than the C=O groups.3,46 We obtained similar Tred values and only slightly reduced H2 consumptions (15% reduction on average, Figure S13b) compared to those obtained without the thermal pre-treatment. Therefore, we conclude that the H2 consumption observed during the TPR experiments is mainly due to the reduction of quinone-type C=O groups (i.e., the active sites for ODH), even though the reduction of other oxygen functional groups cannot be completely excluded. When alkene selectivity is plotted as a function of the Tred, alkene selectivity is observed to gradually increase with Tred (Figure 5e). On the other hand, TOF decreases with Tred (Figure S14). Therefore, we conclude that the lower reducibility of each C=O group at higher C=O density results in a lower TOF (Figure 5b) but a higher alkene selectivity (Figure 5c), which is similar to the general trends reported for metal oxide catalyst systems.38-42 Because the TOF for n-butane conversion and alkene selectivity change in opposite directions with C=O content, the alkene formation rate show a good linear correlation with C=O content (Figure 5f). From these results, it appears that various carbon nanostructures affect ODH activity and selectivity by providing different amounts of carbon edge sites where C=O active sites are grafted. As demonstrated by the TPR results, the overall H2 consumption and reducibility of each C=O active site are mainly affected by the density of the grafted C=O functional groups. The carbon framework itself does not show redox activity (H2 consumption is exactly proportional to the density of C=O groups) and consequently should not be directly involved in the ODH reaction.

ACS Paragon Plus Environment

10

Page 11 of 44

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

Correlation between Carbon Nanostructures and Thermochemical Stabilities. The deactivation of various carbon catalysts over time-on-stream (Figure 3) can be attributed to carbon combustion under the harsh oxidative conditions of ODH.7 Therefore, it is reasonable to expect that the catalyst lifetime is determined by the thermochemical stability of the carbon framework against combustion. Accordingly, we investigated the combustion behavior of the various carbon catalysts by thermogravimetric analysis (TGA) in air (Figures 6a, S15, and S16). The temperature when the carbon mass decreases to 90% of the original mass is defined as ‘Tburn-off’ (Table S2). The results show that the carbon stability against oxidation (or Tburn-off) increases in the order: ZTC < CMK-3 < AC < CMK-3G < CNT < OLC. This result confirms that the carbons with a more graphitic structure have higher oxidation stabilities. In the cases of the oCNT-t (Figure S15) and oOLC-t (Figure S16) samples, oxidation stability (i.e., Tburn-off) was observed to decrease with increasing HNO3 treatment temperature (t) (Table S2). This is consistent with earlier reports showing that increases in the number of defect sites can lead to decreases in stability against carbon oxidation.48 As shown in Figure 6b, the catalyst half-life (t1/2, which is defined as the time required for the alkene formation rate to drop to half of the initial alkene formation rate determined at 1 h) of the various carbon catalysts (Table 1) correlates strongly with Tburn-off determined by the TGA experiments. Considering that the sp2-hybridized C=O groups should be located at edge sites of the carbon layers rather than on a defect-free basal plane, maximizing the number of accessible carbon edge sites increases the number of C=O active sites and, hence, the alkene formation rate. In principle, these types of carbon can be prepared by increasing the surface area and/or lowering the crystallinity (graphitic ordering) of the carbon framework. However, such carbon materials are expected to be less resistant to framework combustion under oxidative conditions, which can lead to fast deactivation during ODH. This could be why most of the carbon catalysts exhibit a general trade-off relationship between alkene formation rate and catalyst half-life, which are all located on the same trend line (Figure 7). However, there is one notable exception that is not located near the trend line; CMK-3G (entry 15 in Figure 7) exhibits an exceptionally high alkene formation rate and stability simultaneously compared to the other carbon catalysts. This may be attributed to the unique framework structure of the CMK-3G

ACS Paragon Plus Environment

11

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 12 of 44

which consists of carbon layers stacked perpendicular to the direction of the nanorod-like carbon framework (Scheme 1e and Figure 2e). Due to the ‘coin-stacking’ of the carbon layers, CMK-3G can simultaneously possess a high graphitic order in the framework as well as a large number of carbon edge sites where C=O active sites are located. The present results suggest that nanostructured carbon materials with ‘coin-stacking’ graphitic alignments have great potential to achieve high redox catalytic performance and thermochemical stability simultaneously. In the field of carbon material science, extensive attention has been paid to the synthesis/application of nanostructured carbons with abundant carbon basal planes (e.g., CNT, OLC, and graphene); however, relatively very little attention has been given to carbon materials with ‘coin-stacking’ graphitic alignment and abundant carbon edge sites.30,49 We believe that nanostructured carbon materials with ‘coin-stacking’ graphitic alignments are the most promising materials for a variety of catalytic applications due to the abundance of edge sites that enable the incorporation of a variety of heteroatoms (such as N and S in addition to O), and their high thermochemical stabilities.

Effect of the O2/n-Butane Ratio on Catalyst Stability and Alkene Selectivity. With the CMK-3G catalyst (entry 15) exhibiting high activity, selectivity and stability simultaneously, we further investigated the effect of the reactant composition (i.e., the molar ratios of the O2/n-butane) on n-butane ODH. Two different initial n-butane conversion levels (~10 and 20%) were achieved by controlling the WHSV (Table 2). As shown in Figure 8, catalyst deactivation is slower at lower O2/n-butane ratios (i.e., less oxidative reactant composition) at both conversion levels. Under O2/n-butane ratios of 1 and 2, the catalyst showed appreciable deactivation after long time-on-stream, and the catalyst mass decreased significantly after the reaction (>57%) due to carbon combustion (Table 2). However, under O2-lean conditions (O2/n-butane ratio of 0.5), the catalyst showed relatively much more stable alkene yields with time-on-stream. Notably, the alkene yield at lower n-butane conversion level could be maintained up to 150 h without any noticeable deactivation (Figure 8a). After the reaction, the catalyst mass decreased only slightly (5%, Table 2), indicating that catalyst combustion is significantly suppressed under these

ACS Paragon Plus Environment

12

Page 13 of 44

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

conditions. At a higher conversion level (Figure 8b), very slow deactivation was observed even under O2lean conditions (O2/n-butane ratio of 0.5). Notably, the catalyst mass increased by 26% after the reaction, indicating that the observed deactivation is due to coke formation rather than combustion of the carbon catalyst. Under these conditions, the O2 reactant was completely consumed (100% O2 conversion, Table 2) and this appears to cause coke deposition. At a lower O2/n-butane ratio, n-butane conversion was slower (a smaller WHSV was used to achieve a similar conversion level) but alkene selectivity was higher due to the suppressed formation of COx. Consequently, at similar conversion levels, the carbon catalyst showed similar or slightly higher alkene yields at lower O2/n-butane ratios (Figure 8). The present results indicate that the catalyst lifetime and selectivity are significantly affected by the choice of reactant composition and reactant conversion level.

CONCLUSIONS In this work, the redox catalytic properties and thermochemical stabilities of a variety of nanostructured carbons were investigated in n-butane ODH. Regardless of the type of carbon, the nbutane conversion rate monotonically increased with the content of the C=O (quinone) active sites. With increasing C=O content, however, the activity per site (i.e., turnover frequency) gradually decreases while the alkene selectivity increases. This is attributed to the lower reducibility of each C=O group at higher C=O density. Because more C=O active sites are present in a thermochemically less stable amorphous carbon framework, the carbons generally exhibit a trade-off relationship between redox catalytic activity and thermochemical stability that are all located on the same trend line. However, a graphitic mesoporous carbon with a ‘coin-stacking’ arrangement of carbon layers exhibited both exceptionally high catalytic activity and stability simultaneously. This was attributed to its unique carbon structure that simultaneously provides high graphitic order in its framework and a large number of carbon edge sites where C=O active sites are grafted. The present study demonstrates that carbon materials can have a remarkably wide range of redox catalytic properties depending on their nanostructures, although they seemingly have similar compositions (mainly carbon). Understanding the correlation between carbon

ACS Paragon Plus Environment

13

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 14 of 44

nanostructures and redox properties/stabilities can provide important insights into the design of advanced carbon catalysts for ODH and other reactions.

EXPERIMENTAL SECTION Catalyst Preparation. As-received activated carbon (G-60, Sigma Aldrich) and CNT (MWCNT, Carbon Nano-material Technology Co., Ltd.) were dispersed in 200 mL of 6 M aqueous HCl solution followed by sonication for 1 h at room temperature to remove metal impurities. Onion-like carbon (OLC) was synthesized by heating a nanodiamond (NanoPure-G01, PlasmaChem) at 1800 °C (ramp: 5 °C min-1) for 6 h in a vacuum furnace. The resultant OLC was also similarly treated with aqueous HCl solution. Some portions of CNT and OLC samples were further treated with HNO3 to control the number of carbon defect sites and oxygen functional groups. Briefly, 1 g of carbon was treated in 100 mL of HNO3 (60 wt%, Samchun Chemical) for 2 h. The treatment temperature varied from 50 to 140 °C. After filtration, the samples were thoroughly washed with deionized water and dried at 100 °C overnight. ZTC was synthesized by acetylene chemical vapor deposition (CVD) on NaX zeolite.24-26 Briefly, 5 g of NaX zeolite was introduced into a plug-flow quartz reactor and heated to 550 °C (ramp: 2 °C min-1) under an Ar flow. At 550 °C, 2% acetylene in He was flowed (200 mL min-1) for 24 h. After the CVD, the sample was carbonized at 800 °C (ramp: 2 °C min-1) for 3 h under an Ar flow (200 mL min-1). The resultant carbon/zeolite composite was twice treated with 800 mL of aqueous HF/HCl solution (1.1/0.8 wt%) to remove the zeolite template. The resultant sample was collected by filtration, thoroughly washed with deionized water, and dried at 100 °C overnight. Hexagonally ordered mesoporous carbons28-30 were synthesized by carbon replication of SBA-15 ordered mesoporous silica. SBA-15 was prepared following the previously reported procedure using a SiO2/P123 molar ratio of 60 and a hydrothermal synthesis temperature of 100 °C.50 CMK-3 was synthesized with sucrose (99%, Junsei Chemicals) as the carbon precursor.28 Typically, 1 g of SBA-15 was infiltrated with 5 mL of an aqueous solution containing 1.7 g of sucrose and 0.2 g of sulfuric acid (95%, Samchun Chemical). After impregnation, water was dried at 100 °C for 3 h, after which the sucrose

ACS Paragon Plus Environment

14

Page 15 of 44

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

was polymerized at 160 °C for 6 h. After cooling to room temperature, the sample was infiltrated with an additional 5 mL of an aqueous solution containing 1.1 g of sucrose and 0.12 g of sulfuric acid followed by the same polymerization process. Carbonization was the carried out in a plug-flow quartz reactor at 1000 °C (ramp: 2 °C min-1) for 2 h under an Ar flow. The resultant silica/carbon composite was twice etched with 300 mL of aqueous HF solution (1.3 wt%) to remove the silica template. For the synthesis of CMK-3G,30 SBA-15 was first impregnated with an AlCl3 ethanol solution to obtain Al-SBA-15 (Si/Al ratio = 20). After drying at 80 °C, the sample was calcined in air at 550 °C for 6 h. 3 g of the resultant AlSBA-15 and 4.5 g of acenaphthene (99%, Tokyo Chemical Industry) were physically mixed in a mortar. The mixture was placed into a stainless steel autoclave (100 cm3). After purging with Ar gas, the autoclave was sequentially heated at 300 °C (ramp: 2 °C min-1) for 6 h and at 500 °C (ramp: 2 °C min-1) for 2 h. After cooling to room temperature, the resultant sample was loaded into a plug-flow quartz reactor and further carbonized at 1000 °C (ramp: 2 °C min-1) for 6 h under an Ar flow. The resultant carbon-silica composite was twice etched with 800 mL of aqueous HF/HCl solution (1.5/0.07 wt%) to remove the aluminosilicate template. The resultant sample was collected by filtration, thoroughly washed with deionized water, and dried at 100 °C overnight.

Physical characterization. Powder X-ray diffraction (XRD) patterns were recorded with a D2 PHASER instrument (Bruker AXS) operating at 30 kV and 10 mA with Cu Kα as the X-ray source. High-resolution transmission electron microscopy (TEM) images were taken with an FEI Titan Cubed G2 60-300 microscope operating at 200 kV (KAIST Analysis center for Research Advancement, KARA). Thermogravimetric analysis (TGA) data were collected with a TGA N-1500 instrument (Sinco) with a 10 °C min-1 heating rate under an air flow (30 mL min-1). N2 adsorption-desorption isotherms were measured with a BEL-sorp-max (BEL Japan) volumetric analyzer at -196 °C after degassing at 200 °C for 4 h. The micropore volume was calculated using the t-plot method, and the total pore volume was determined at a P/P0 of 0.95. The mesopore volume was calculated by subtracting the micropore volume from the total pore volume. The apparent surface area was determined in the P/P0 range of 0.05 – 0.15

ACS Paragon Plus Environment

15

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 44

using the Brunauer-Emmett-Teller (BET) equation. The pore size distributions were estimated from the adsorption branches of the isotherms using a non-local density functional theory (NLDFT) method. The metal impurities (Fe, Al, and Si) in the carbon catalysts were determined by inductively coupled plasma optical emission spectroscopy (ICP-OES) using an ICP-OES 720 instrument (Agilent). Raman spectroscopy was carried out using an ARAMIS (Horiba Jobin Yvon) spectrometer with a 514 nm laser source. O 1s X-ray photoelectron spectroscopy (XPS) was carried out on a Sigma Probe instrument (Thermo Fischer Scientific VG) equipped with a microfocused monochromator X-ray source. The binding energies were calibrated based on the C 1s peak at 284.5 eV. The deconvolution of the O 1s spectra was performed after subtraction of the Shirley background. The maximum peak positions for individual species were within ±0.1 eV and the full widths at half-maximum (FWHM) were fixed at 1.6 eV during the fitting procedure. To check the reducibility of the carbon catalysts, temperature programmed reduction (TPR) profiles were collected with a Belcat instrument (BEL Japan) equipped with a thermal conductivity detector (TCD). The sample after 1 h ODH reaction was loaded into a U-type quartz tube, and the temperature was increased from 30 to 900 °C (ramp: 10 °C min-1) under a 5% H2/Ar flow (30 mL min-1). To remove CO, CO2, and H2O gases generated by the thermal decomposition of the carbon, a trap containing a mixture of CuCl-modified activated carbon47 and 13X molecular sieves was used after cooling with an ice bath. To quantify the amount of H2 consumed during TPR, the reduction of CuO powder (99%, Sigma Aldrich) was used to calibrate the TCD response.

Catalytic measurements. The n-butane ODH reactions were carried out in a plug-flow tubular quartz reactor (8 mm inner diameter) under atmospheric pressure. If not specially mentioned, the reaction was carried out at 450 °C with an O2-rich reactant composition (1.2 kPa n-butane, 2.4 kPa O2, 96.4 kPa He). For the reaction, the carbon catalyst was placed in a reactor, and the temperature was increased to 450 °C (ramp: 5 °C min-1) under a He flow. After 1 h stabilization at the reaction temperature, the reactant gas mixture was introduced. To check the activation/deactivation behavior of the carbon catalysts, the reaction was carried out with 20 mg of the catalysts at a fixed WHSV (2.9 h-1). For comparisons of

ACS Paragon Plus Environment

16

Page 17 of 44

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

intrinsic activity and selectivity of the different carbon catalysts, catalytic data were also collected at various conversion levels by varying the WHSV. n-Butane conversions were typically below 10%. The nbutane conversion and alkene selectivity measured at a fixed 1 h time-on-stream were extrapolated to zero residence time (1/WHSV) in order to obtain the primary alkene formation rate and selectivity. To investigate the effect of the O2/n-butane ratio, the reactions were also carried out with the CMK-3G catalyst at O2/n-butane ratios of 0.5, 1, and 2. The experiments were performed at two different initial nbutane conversion levels (~10 and 20%) by altering the WHSV. The products were analyzed with an online gas chromatograph (Younglin, YL6000) equipped with FID and TCD detectors. A GS-GASPRO capillary column was used to separate the hydrocarbons, and a Carboxen 1000 column was used to separate the inorganic gases including O2 and COx. Because carbon catalysts can be gradually combusted during the reaction, COx can also be generated by the catalyst combustion in addition to the catalytic oxidation of n-butane. However, for all carbon catalysts tested, the rate of COx formation by n-butane oxidation was significantly faster than the COx formation by the catalyst combustion. In the present calculation of product selectivities, we focused on the n-butane conversion into alkenes and COx (we determined the n-butane conversion and the formation of alkenes with FID, and assumed that their difference is due to the formation of COx). Therefore, COx formation due to catalyst combustion was ignored in our calculation. Even though this calculation method does not precisely describe the real molecular composition that can be determined by the combination with TCD, it better describes the product composition formed via ‘true catalytic action’.

ASSOCIATED CONTENT Supporting Information This Supporting Information is available free of charge on the ACS Publications website at DOI:

ACS Paragon Plus Environment

17

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 18 of 44

N2 adsorption-desorption isotherms, pore size distributions, low-angle XRD patterns, TEM images, Raman spectra, additional data for catalytic reactions, O 1s XPS spectra, TPR profiles, TGA curves, and ICP-OES data of the carbon materials (PDF).

AUTHOR INFORMATION Corresponding Author *E-mail: [email protected]

ORCID Minkee Choi: 0000-0003-0827-2572

Notes The authors declare no competing financial interest.

ACKNOWLEDGMENT This work was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF-2017R1A2B2002346).

REFERENCES (1)

Rodríguez-reinoso, F. Carbon 1998, 36, 159-175.

(2)

Su, D. S.; Perathoner, S.; Centi, G. Chem. Rev. 2013, 113, 5782-5816.

(3)

Figueiredo, J. L.; Pereira, M. F. R. Catal. Today 2010, 150, 2-7.

(4)

Qi, W.; Su, D. ACS Catal. 2014, 4, 3212-3218.

ACS Paragon Plus Environment

18

Page 19 of 44

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

(5)

Su, D. S.; Wen, G.; Wu, S.; Peng, F.; Schlögl, R. Angew. Chem., Int. Ed. 2016, 55, 2-31.

(6)

Pereira, M. F. R.; Órfão, J. J. M.; Figueiredo, J. L. Appl. Catal., A 1999, 184, 153-160.

(7)

Pereira, M. F. R.; Órfão, J. J. M.; Figueiredo, J. L. Appl. Catal., A 2001, 218, 307-318.

(8)

Zhang, J.; Su, D.; Zhang, A.; Wang, D.; Schlögl, R.; Hébert, C. Angew. Chem., Int. Ed. 2007, 46,

7319-7323. (9)

Zhang, J.; Liu, X.; Blume, R.; Zhang, A.; Schlögl, R.; Su, D. S. Science 2008, 322, 73-77.

(10)

Frank, B.; Zhang, J.; Blume, R.; Schlögl, R.; Su, D. S. Angew. Chem., Int. Ed. 2009, 48, 6913-

6917. (11)

Frank, B.; Morassutto, M.; Schomäcker, R.; Schlögl, R.; Su, D. S. ChemCatChem 2010, 2, 644-

648. (12)

Qi, W.; Liu, W.; Zhang, B.; Gu, X.; Guo, X.; Su, D. Angew. Chem., Int. Ed. 2013, 52, 14224-

14228. (13)

Wen, G.; Diao, J.; Wu, S.; Yang, W.; Schlögl, R.; Su, D. S. ACS Catal. 2015, 5, 3600-3608.

(14)

Delgado, J. J.; Chen, X.-W.; Frank, B.; Su, D. S.; Schlögl, R. Catal. Today 2012, 186, 93-98.

(15)

Keller, N.; Maksimova, N. I.; Roddatis, V. V; Schur, M.; Mestl, G.; Butenko, Y. V; Kuznetsov,

V. L.; Schlögl, R. Angew. Chem., Int. Ed. 2002, 41, 1885-1888. (16)

Liu, X.; Frank, B.; Zhang, W.; Cotter, T. P.; Schlögl, R.; Su, D. S. Angew. Chem., Int. Ed. 2011,

50, 3318-3322. (17)

Schwartz, V.; Fu, W.; Tsai, Y. T.; Meyer, H. M.; Rondinone, A. J.; Chen, J.; Wu, Z.; Overbury,

S. H.; Liang, C. ChemSusChem 2013, 6, 840-846.

ACS Paragon Plus Environment

19

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

(18)

Page 20 of 44

Dathar, G. K. P.; Tsai, Y.-T.; Gierszal, K.; Xu, Y.; Liang, C.; Rondinone, A. J.; Overbury, S. H.;

Schwartz, V. ChemSusChem 2014, 7, 483-491. (19)

Liang, C.; Xie, H.; Schwartz, V.; Howe, J.; Dai, S.; Overbury, S. H. J. Am. Chem. Soc. 2009, 131,

7735-7741. (20)

Xie, H.; Wu, Z.; Overbury, S. H.; Liang, C.; Schwartz, V. J. Catal. 2009, 267, 158-166.

(21)

Delgado, J. J.; Su, D. S.; Rebmann, G.; Keller, N.; Gajovic, A.; Schlögl, R. J. Catal. 2006, 244,

126-129. (22)

Iwasawa, Y.; Nobe, H.; Ogasawara, S. J. Catal. 1973, 31, 444-449.

(23)

Zhang, J.; Wang, X.; Su, Q.; Zhi, L.; Thomas, A.; Feng, X.; Su, D. S.; Schlögl, R.; Müllen, K. J.

Am. Chem. Soc. 2009, 131, 11296-11297. (24)

Nishihara, H.; Yang, Q. H.; Hou, P. X.; Unno, M.; Yamauchi, S.; Saito, R.; Paredes, J. I.;

Martínez-Alonso, A.; Tascón, J. M. D.; Sato, Y.; Terauchi, M.; Kyotani, T. Carbon 2009, 47, 1220-1230. (25)

Choi, S.; Kim, H.; Lee, S.; Wang, Y.; Ercan, C.; Othman, R.; Choi, M. Chem. Eng. J. 2015, 280,

597-605. (26)

Choi, C. H.; Kim, M.; Kwon, H. C.; Cho, S. J.; Yun, S.; Kim, H.-T.; Mayrhofer, K. J. J.; Kim, H.;

Choi, M. Nat. Commun. 2016, 7, 10922-10931. (27)

Ryoo, R.; Joo, S. H.; Jun, S. J. Phys. Chem. B 1999, 103, 7743-7746.

(28)

Jun, S.; Joo, S. H.; Ryoo, R.; Kruk, M.; Jaroniec, M.; Liu, Z.; Ohsuna, T.; Terasaki, O. J. Am.

Chem. Soc. 2000, 122, 10712-10713. (29)

Lu, A. H.; Schüth, F. Adv. Mater. 2006, 18, 1793-1805.

(30)

Kim, T.-W.; Park, I.-S.; Ryoo, R. Angew. Chem., Int. Ed. 2003, 42, 4375-4379.

ACS Paragon Plus Environment

20

Page 21 of 44

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

(31)

Tomita, S.; Burian, A.; Dore, J. C.; Lebolloch, D.; Fujii, M.; Hayashi, S. Carbon 2002, 40, 1469-

1474. (32)

Sadezky, A.; Muckenhuber, H.; Grothe, H.; Niessner, R.; Pöschl, U. Carbon 2005, 43, 1731-

1742. (33)

Chen, C.; Zhang, J.; Peng, F.; Su, D. S. Mater. Res. Bull. 2013, 48, 3218-3222.

(34)

Okpalugo, T. I. T.; Papakonstantinou, P.; Murphy, H.; McLaughlin, J.; Brown, N. M. D. Carbon

2005, 43, 153-161. (35)

Murphy, H.; Papakonstantinou, P.; Okpalugo, T. I. T. J. Vac. Sci. Technol. B 2006, 24, 715-720.

(36)

Masheter, A. T.; Xiao, L.; Wildgoose, G. G.; Crossley, A.; Jones, J. H.; Compton, R. G. J. Mater.

Chem. 2007, 17, 3515-3524. (37)

Madeira, L. M.; Portela, M. F. Catal. Rev. 2002, 44, 247-286.

(38)

Patel, D.; Andersen, P. J.; Kung, H. H. J. Catal. 1990, 125, 132-142.

(39)

Owen, O. S.; Kung, M. C.; Kung, H. H. Catal. Lett. 1992, 12, 45-50.

(40)

Chen, K.; Bell, A. T.; Iglesia, E. J. Catal. 2002, 209, 35-42.

(41)

Zhao, Z.; Yamada, Y.; Ueda, A.; Sakurai, H.; Kobayashi, T. Catal. Today 2004, 93–95, 163-171.

(42)

Schwarz, O.; Habel, D.; Ovsitser, O.; Kondratenko, E. V.; Hess, C.; Schomäcker, R.; Schubert,

H. J. Mol. Catal. A: Chem. 2008, 293, 45-52. (43)

Madeira, L. M.; Portela, M. F.; Mazzocchia, C.; Kaddouri, A.; Anouchinsky, R. Catal. Today

1998, 40, 229-243.

ACS Paragon Plus Environment

21

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

(44)

Page 22 of 44

Hiroki, A.; Thomas, K.; Franco, C.; William, R. S.; Milo, S. P. S.; Alan, H. W.; Richard, H. F. J.

Phys. Chem. B 1990, 103, 8116-8121. (45)

Wang, X.; Sun, G.; Routh, P.; Kim, D.-H.; Huang, W.; Chen, P. Chem. Soc. Rev. 2014, 43, 7067-

7098. (46)

Figueiredo, J. L.; Pereira, M. F. R.; Freitas, M. M. A.; Órfão, J. J. M. Carbon 1999, 37, 1379-

1389. (47)

Hirai, H.; Wada, K.; Komiyama, M. Bull. Chem. Soc. Jpn. 1986, 59, 2217-2223.

(48)

Li, C.; Zhao, A.; Xia, W.; Liang, C.; Muhler, M. J. Phys. Chem. C 2012, 116, 20930-20936.

(49)

Rodriguez, N.; Chambers, a; Baker, R. Langmuir 1995, 11, 3862-3866.

(50)

Choi, M.; Heo, W.; Kleitz, F.; Ryoo, R. Chem. Commun. 2003, 12, 1340–1341.

ACS Paragon Plus Environment

22

Page 23 of 44

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

Scheme 1. Schematic illustration of the nanostructured carbons investigated. (a) Multi-walled carbon nanotube (CNT), (b) onion-like carbon (OLC), (c) activated carbon (AC), (d) hexagonally ordered mesoporous carbon with an amorphous framework (CMK-3), (e) hexagonally ordered mesoporous carbon with a ‘coin-stacking’ graphitic framework (CMK-3G), and (f) zeolite-templated carbon (ZTC).

ACS Paragon Plus Environment

23

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 24 of 44

Figure 1. Wide-angle XRD patterns of the nanostructured carbons.

ACS Paragon Plus Environment

24

Page 25 of 44

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

Figure 2. High-resolution TEM images of (a) CNT, (b) OLC, (c) AC, (d) CMK-3, (e) CMK-3G, and (f) ZTC. Insets in (d) and (e) are the electron diffraction patterns.

ACS Paragon Plus Environment

25

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 26 of 44

Figure 3. Alkene formation rates as a function of the time-on-stream in n-butane ODH. (a) CNTs, (b) OLCs, and (c) AC and templated carbons (ZTC, CMK-3, and CMK-3G). The ODH reaction was carried out with an O2-rich reactant composition (reaction condition: 450 °C, 1.2 kPa n-butane, 2.4 kPa O2, 96.4 kPa He; WHSV = 2.9 h-1).

ACS Paragon Plus Environment

26

Page 27 of 44

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

Figure 4. Correlation between the ID1/IG ratio in Raman spectra and density of C=O groups in the carbon catalysts. All data were obtained from the carbon catalysts collected after 1 h ODH reaction. The numbers inside the plot indicate an entry noted in Table 1.

ACS Paragon Plus Environment

27

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 28 of 44

Figure 5. Correlation between carbon nanostructures and redox catalytic properties. (a) n-Butane conversion rate, (b) turnover frequency (TOF) for n-butane conversion, (c) alkene selectivity, and (d) Tred (defined as the onset temperature where the H2 consumption rate reaches 10% of the maximum H2 consumption rate in TPR profiles) as a function of the density of the C=O (quinone) groups in the carbon catalysts. (e) Alkene selectivity as a function of Tred. (f) Alkene formation rate as a function of the density of the C=O groups in the carbon catalysts. All the data were obtained from the carbon catalysts after 1 h ODH reaction. The numbers inside each plot indicate an entry noted in Table 1.

ACS Paragon Plus Environment

28

Page 29 of 44

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

Figure 6. Correlation between carbon oxidation stability and catalyst lifetime in n-butane ODH. (a) Thermogravimetric analysis (TGA) profiles of the various nanostructured carbons measured in air. (b) Correlation between the catalyst half-life (t1/2) and Tburn-off (defined as the temperature when the carbon mass decreases to 90% of the original mass). The numbers inside the plot indicate an entry noted in Table 1.

ACS Paragon Plus Environment

29

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 30 of 44

Figure 7. Correlation between the initial alkene formation rate and the catalyst half-life (t1/2) in n-butane ODH. The alkene formation rates were determined at 1 h time-on-stream. The numbers inside the plot indicate an entry noted in Table 1.

ACS Paragon Plus Environment

30

Page 31 of 44

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

Figure 8. Effects of the reactant composition (O2/n-butane ratio) on n-butane ODH over CMK-3G at different n-butane conversion levels (reaction conditions: 450 °C, 1.2 kPa n-butane, 0.6 – 2.4 kPa O2, and He balance). The initial n-butane conversion levels were controlled to ca. (a) 10% and (b) 20% by changing the WHSV (see Table 2).

ACS Paragon Plus Environment

31

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 32 of 44

Table 1. Catalytic results of n-butane ODH and densities of the C=O active site for the various carbon catalysts.a Entry

Catalyst

n-Butane Conv. rateb (mmol g-1 h-1)

C=O densityc (mmol g-1)

TOFd (h-1)

Alkene formation rateb (mmol g-1 h-1)

t1/2e (h)

1

AC

5.9

1.4

4.2

Alkenes 75

COx 25

4.4

19

2

CNT

1.8

0.18

10

41

59

0.74

110

3

oCNT-50

2.0

0.33

6.1

46

54

0.92

77

4

oCNT-60

2.2

0.36

6.1

55

45

1.2

59

Selectivityb (%)

5

oCNT-80

3.3

0.57

5.8

60

40

2.0

46

6

oCNT-100

3.6

0.81

4.4

70

30

2.5

17

7

oCNT-140

5.2

1.2

4.3

75

25

3.9

9

8

OLC

1.7

0.15

11

38

62

0.65

118

9

oOLC-60

2.1

0.31

6.8

41

59

0.86

110

10

oOLC-80

2.3

0.35

6.6

48

52

1.1

95

11

oOLC-100

4.3

0.83

5.2

58

42

2.5

43

12

oOLC-120

5.1

1.1

4.6

65

35

3.3

15

13

oOLC-140

5.4

1.2

4.5

71

29

3.8

5

14

CMK-3

6.3

1.4

4.5

76

24

4.8

12

15

CMK-3G

4.2

1.0

4.2

77

23

3.2

83

16 ZTC 7.5 2.0 3.8 78 22 5.9 4 All data were collected after 1 h time-on-stream (reaction condition: 450 °C, 1.2 kPa n-butane, 2.4 kPa O2, 96.4 kPa He). bThe numbers were determined by extrapolating the catalytic data to zero residence time (1/WHSV) (Figures S6-S8). cSurface C=O density calculated by XPS analysis. dTurnover frequency (TOF) was calculated by normalizing n-butane conversion rate by the density of the C=O active sites. eCatalyst half-life (t1/2) was defined as the time required for the alkene formation rate to drop to half of the initial alkene formation rate determined at 1 h time-on-stream (Figure 3). a

ACS Paragon Plus Environment

32

Page 33 of 44

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

Table 2. Catalytic results of CMK-3G in n-butane ODH with different reactant compositions and nbutane levels.a O2/ n-butane 0.5

1

2

Alkene yield (%)

Alkene formation rate (mmol g-1 h-1)

∆ mb (%)

40

6.6

0.66

–5

54

46

10

0.33

+ 26

31

57

43

6.3

1.01

– 75

21

64

45

55

9.5

0.61

– 57

1.3

11

17

55

45

6.1

1.4

– 91

0.56

19

33

43

57

8.2

0.79

– 90

Selectivity (%)

WHSV (h-1)

n-Butane Conv. (%)

O2 Conv. (%)

Alkenes

COx

0.58

11

52

60

0.19

19

100

0.93

11

0.37

a

All catalytic data were collected after 5 h time-on-stream (reaction conditions: 450 °C, 1.2 kPa n-butane, 0.6 – 2.4 kPa O2, and He balance). bMass change of the carbon catalyst after the reaction time shown in Figure 8.

ACS Paragon Plus Environment

33

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 34 of 44

Table of Contents Graphic

ACS Paragon Plus Environment

34

Page 35 of 44

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

Scheme 1. Schematic illustration of the nanostructured carbons investigated. (a) Multi-walled carbon nanotube (CNT), (b) onion-like carbon (OLC), (c) activated carbon (AC), (d) hexagonally ordered mesoporous carbon with an amorphous framework (CMK-3), (e) hexagonally ordered mesoporous carbon with a ‘coin-stacking’ graphitic framework (CMK-3G), and (f) zeolite-templated carbon (ZTC). 165x85mm (300 x 300 DPI)

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

Figure 1. Wide-angle XRD patterns of the nanostructured carbons. 80x75mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 36 of 44

Page 37 of 44

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

Figure 2. High-resolution TEM images of (a) CNT, (b) OLC, (c) AC, (d) CMK-3, (e) CMK-3G, and (f) ZTC. Insets in (d) and (e) are the electron diffraction patterns. 165x94mm (300 x 300 DPI)

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

Figure 3. Alkene formation rates as a function of the time-on-stream in n-butane ODH. (a) CNTs, (b) OLCs, and (c) AC and templated carbons (ZTC, CMK-3, and CMK-3G). The ODH reaction was carried out with an O2-rich reactant composition (reaction condition: 450 °C, 1.2 kPa n-butane, 2.4 kPa O2, 96.4 kPa He; WHSV = 2.9 h-1). 80x150mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 38 of 44

Page 39 of 44

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

Figure 4. Correlation between the ID1/IG ratio in Raman spectra and density of C=O groups in the carbon catalysts. All data were obtained from the carbon catalysts collected after 1 h ODH reaction. The numbers inside the plot indicate an entry noted in Table 1. 80x80mm (300 x 300 DPI)

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

Figure 5. Correlation between carbon nanostructures and redox catalytic properties. (a) n-Butane conversion rate, (b) turnover frequency (TOF) for n-butane conversion, (c) alkene selectivity, and (d) Tred (defined as the onset temperature where the H2 consumption rate reaches 10% of the maximum H2 consumption rate in TPR profiles) as a function of the density of the C=O (quinone) groups in the carbon catalysts. (e) Alkene selectivity as a function of Tred. (f) Alkene formation rate as a function of the density of the C=O groups in the carbon catalysts. All the data were obtained from the carbon catalysts after 1 h ODH reaction. The numbers inside each plot indicate an entry noted in Table 1. 165x109mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 40 of 44

Page 41 of 44

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

Figure 6. Correlation between carbon oxidation stability and catalyst lifetime in n-butane ODH. (a) Thermogravimetric analysis (TGA) profiles of the various nanostructured carbons measured in air. (b) Correlation between the catalyst half-life (t1/2) and Tburn-off (defined as the temperature when the carbon mass decreases to 90% of the original mass). The numbers inside the plot indicate an entry noted in Table 1. 80x147mm (300 x 300 DPI)

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

Figure 7. Correlation between the initial alkene formation rate and the catalyst half-life (t1/2) in n-butane ODH. The alkene formation rates were determined at 1 h time-on-stream. The numbers inside the plot indicate an entry noted in Table 1. 80x80mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 42 of 44

Page 43 of 44

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

Figure 8. Effects of the reactant composition (O2/n-butane ratio) on n-butane ODH over CMK-3G at different n-butane conversion levels (reaction conditions: 450 °C, 1.2 kPa n-butane, 0.6 – 2.4 kPa O2, and He balance). The initial n-butane conversion levels were controlled to ca. (a) 10% and (b) 20% by changing the WHSV (see Table 2). 80x130mm (300 x 300 DPI)

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

Table of Contents Graphic 85x41mm (150 x 150 DPI)

ACS Paragon Plus Environment

Page 44 of 44