Chinese Chemical Letters  2018, Vol. 29 Issue (6): 884-886   PDF    
Honeycomb-shaped PtSnNa/γ-Al2O3/cordierite monolithic catalyst with improved stability and selectivity for propane dehydrogenation
Shiyong Zhaoa, Bolian Xua,b, Lei Yua,b,c, Yining Fana,b    
a Key Laboratory of Mesoscopic Chemistry of Ministry of Education, School of Chemistry and Chemical Engineering, Nanjing University, Nanjing 210093, China;
b Nanjing University-Yangzhou Chemistry and Chemical Engineering Institute, Yangzhou 211400, China;
c School of Chemistry and Chemical Engineering, Yangzhou University, Yangzhou 225002, China
Abstract: A novel honeycomb-shaped PtSnNa/γ-Al2O3/cordierite monolithic catalyst (Pt 0.5%, Sn 0.9%, Na 1.0%, relative to Al2O3 weight) was developed and its catalytic performances in propane dehydrogenation were tested and compared with the classical granule catalyst with same Pt, Sn and Na contents under the conditions of 0.1 MPa, 590℃, C3H8/H2 at 3/1 (molar ratio) and gas hourly space velocity (GHSV) at 155 h-1. Interestingly, despite the generated coke amount and exposed Pt fraction, the honey combshaped structure of monolithic catalyst exerted important influences on its catalytic activities and led to the advanced catalytic performances over the granule catalyst.
Key words: Propane dehydrogenation     Propylene     Platinum     Cordierite     Monolithic catalyst    

Catalytic dehydrogenation of light alkanes is one of the most important processes in petrochemical industry for optimizing the petroleum resources [1, 2] to produce light olefins as basic raw chemical materials and hydrogen as the green energy. Yet, dehydrogenation of propane to propylene involves the oxidative dehydrogenation method and dehydrogenation in the presence of hydrogen. The former mainly solves the issues of poor propylene selectivity, while the latter focuses on improving the stability of the catalysts.

In propane dehydrogenation reactions in the presence of hydrogen, Pt, Pd, Rh, Ir, V, Cr [3-6], etc. were usually employed as the active components, while Sn, Ce, Zn, Ga [7-10], etc. were used as the promoters. Alkali metals, such as Na and K [11, 12] could be added as the alkali additives. Owing to their high catalytic activities, PtSn-based catalysts have attracted much attention. Usually, Al2O3, ZSM-5, SAPO-34, MgAl2O4, NaY, CIT-6 [13-18], etc. were employed as the supports for their versatile structures and acidity that could be utilized for catalytic activity optimizations by tuning the interactions between the active component and the support. In the catalyst fabrication processes, Sn was uploaded first and led to its oxide fixing over Al2O3 after the subsequent calcination. The fixed tin oxide could well divide Pt nanoparticles and facilitate the dispersion of Pt during the Pt uploading process. Moreover, the reactions could be improved through the reaction condition optimizations [19] and reaction atmosphere adjustment [20] as well.

Recently, we developed a novel honeycomb-shaped PtSnNa/γ-Al2O3/cordierite monolithic catalyst. The performances of the monolithic catalyst and the classical granular catalyst in propane dehydrogenation in the presence of hydrogen were compared and it was found that the unique structures of the monolithic catalyst led to its advanced catalytic performances over the granular catalyst. Herein, we wish to report our findings.

The PtSnNa/γ-Al2O3/cordierite monolithic catalysts were initially designed and prepared for examination (See experimental details in Supporting information). Less Al2O3 amount led to reduced active components and insufficient use of the monolithic support, but excess Al2O3 might block the tunnels of the cordierite honeycomb. Meanwhile, Al2O3 washcoat thickness on monolith was an important parameter for the mass transfer coefficients varying with the coating thickness [21] and the radial mass transfer processes in monolithic reactors. Therefore, it was critical to control the coating thickness of the catalysts. The coatings were fabricated by dipping processes, in which the monoliths were immersed in Al2O3/Al(NO3)3/NaCl/H2O slurry for multiple times and calcined, as shown in detail in experimental section of Supporting information. It was found that 0.6-1.0 g was the favorable coating weight range and 0.7 g was finally screened out to be the preferable coating weight for the excellent repeatability. The washcoat was fabricated by repeating the immersing procedures for two or three times.

The coated cordierites were then immersed in aqueous SnCl4 for certain times, and then dried in oven. The rest solution volumes were measured and Sn concentrations in them were determined by ICP to calculate the lost Sn weight (Fig. S1 in Supporting information). With the increased immersing time, Sn concentrations in solution gradually decreased, while the lost Sn weight was enhanced. The values tended to be stable after 1h, which was the preferable immersing time for Sn loading in catalyst preparation. ICP analysis of the material indicated that the Sn content of the catalyst was around 0.9% (weight content, vs. Al2O3). Pt could be uploaded through the similar immersing method with aqueous H2PtCl6/HCl and the favorable immersing time to upload Sn was 4h (Fig. S2 in Supporting information), with Pt content at ca. 0.5% (weight content, vs. Al2O3).

The monolithic catalyst was then characterized by scanning electron microscopy (SEM), transmission electron microscopy (TEM) (Fig. 1). The cordierite support contained Si, Mg, Al etc., while the coatings were γ-Al2O3 only. Therefore, coating thickness could be judged by the distributions of Si and Al elements. Fig. 1a represents the SEM image of the coating section, and Figs. 1b and c were mapping imagesof Si and Al on the same section respectively. As shown in Fig. 1a, the binding compactness of γ-Al2O3 coatings with cordierite matrix was good. Although the industrial Al2O3 contained a few Si impurities, which could be observed in Fig. 1b as the sparse and independent reddots, the borderbetweenthe Al2O3 coatings and cordierite could be observed clearly to judge the thickness of the washcoat. The arrow length in Fig. 1c deducted by that in Fig. 1b was the Al dispersion thickness, which was the coating thickness and was calculated to be ca. 60 μm. As show in TEM image (Fig. 1d), the Pt nanoparticles (white dots) dispersed uniformly on the material. The nano-scale Pt facilitated its contact with the reactants.

Download:
Fig. 1. SEM and TEM images of the materials: (a) SEM images of the γ-Al2O3 washcoat. (b) Si mapping of the γ-Al2O3 washcoat. (c) Al mapping of the γ-Al2O3 washcoat. (d) TEM image of the monolithic catalyst.

Space velocity represents the reactant processing ability of unit volumeof catalyst in unit time. The effectof propane space velocity on propane dehydrogenation was then investigated. The propane conversion and propylene selectivity of the PtSnNa/γ-Al2O3 monolithic catalysts decreased with the reaction time. The propane space velocity rose from 78h-1 to 337h-1 after 12h reaction and the propane conversion over monolithic catalyst decreased, while the propylene selectivity increased. The propylene yield reached its maximum value at 11.7% with the propane space velocity at 155h-1, which was the optimized condition and eliminated the effect of external diffusion affected by the catalyst size (Fig. S3 in Supporting information).

Performances of the monolithic and granule catalysts in propane dehydrogenation were then tested under the same optimized reaction conditions for comparison (Fig. 2). Generally, the propane conversions over monolithic catalyst were ca. 5% higher than those of the reactions catalyzed by the granule catalyst. The propylene selectivities over the monolithic catalyst exceeded 98%, while the same values over the granule catalyst decreased to 88% after 12h of reaction.Meanwhile, the selectivities of the by-products methane, ethylene and ethane over the monolithic catalyst were 0.4%, 0.5% and 0.5%, respectively, which were much lower than the values of the reactions over the classical granule catalyst (methane 1%, ethylene 1%, ethane 4%).

Download:
Fig. 2. The catalytic performances for propane dehydrogenation over the PtSnNa/Al2O3/cordierite monolithic catalyst (curvea)and the granule PtSnNa/Al2O3 catalyst (curve b).

The generated coke amounts of both monolithic and granule catalysts after reaction were determined by thermogravimetrymass spectrometry (TG-MS). The granule catalyst was formed by Al2O3 support and the active components, while the monolithic catalyst contained cordierite, Al2O3 coatings and the active components, and the weight of cordierite (15g) is much higher than that of the Al2O3 coatings (0.7g). Weight loss rate of the granule catalyst=coke deposition/(Al2O3+PtSnNa+coke deposition) (eq.1), while theweight loss rate of monolithic catalyst=coke deposition/(cordierite+Al2O3+PtSnNa+coke deposition) (eq. 2). Carbon deposition ratio of the monolithic catalyst was calculated to be 9% of the Al2O3 weight (coke deposition/Al2O3=9wt%) through eq. 2 according to the weight loss value in Fig. 3, higher than that of the granule catalyst (2wt%) calculated from the eq. 1.

Download:
Fig. 3. Thermogravimetry (TG) and mass spectrometry (MS) curvesof the catalysts: curve a, TG curve of the granule catalyst; curve b, MS curve of the granule catalyst; curve c, TG curve of the monolithic catalyst; curve d, MS curve of the monolithic catalyst.

Hydrogen adsorption experiments were conducted to calculate the exposed Pt fractions [22]. The hydrogen adsorption capacity of the fresh granule catalyst and the monolithic catalyst were the similar 47mL/g Pt and 50mL/g Pt respectively. After 12h reaction, the values fell to the similar 39mL/g Pt and 38mL/g Pt, respectively. The exposed Pt fraction of the granule catalyst and the monolithic catalyst were calculated to be 83% and 78% respectively and accordingly.

As shown above, the monolithic catalyst led to more coke and less exposed Pt than granule catalyst, but resulted in better catalytic performances contrarily. It was supposed that the superior catalytic performances of the monolithic catalyst might be attributed to its unique structure. In monolithic catalyst, the tunnel structure was regular, and had very good mass transfer performance toallow thepropane diffusing from the bulk fluid into the catalyst interface [23]. The generated propylene could easily diffuse intothebulk fluid from the catalyst interface to reduce deep dehydrogenations, while the coke could be transferred from the catalyst surface into the supports timely as well [24] (Fig. 4). By contrast, the granule catalyst was piled up disorderly, and the generated propylene inevitably contacted with another section of the catalyst to initiate deep dehydrogenations. Thus, although monolithic catalyst resulted in more coke generation and contained less Pt exposing, it was still more active than granule catalyst.

Download:
Fig. 4. Mass transfer schematic diagram of the monolithic catalyst.

In conclusion, we developed a convenient method to fabricate the novel PtSnNa/γ-Al2O3/cordierite monolithic catalyst, which was found to be more stable and active than classical granule catalyst, despite its increased coke deposition and reduced Pt exposing. The abnormal phenomenon was probably attributed to the regular tunnel structure of the catalyst, which allowed masstransfer more freely and quickly.

Acknowledgments

We thank Jiangsu Planned Projects for Postdoctoral Research Funds (No. 1301080C), Specialized Research Fund for the Doctoral Program of Higher Education (No. SRFDP-2012009111001), Key Science & Technology Specific Projects of Yangzhou (No. YZ20122029) and Yangzhou Nature Science Foundation (No. YZ2014040) for financial support.

Appendix A. Supplementary data

Supplementary data associatedwith this article can be found, in the online version, at https://doi.org/10.1016/j.cclet.2017.11.016.

References
[1]
W.W. Liu, Y.S. Feng, G.Y. Wang, et al., Chin. Chem. Lett. 27 (2016) 905-909. DOI:10.1016/j.cclet.2016.02.026
[2]
X.L. Ni, J. Liu, Y.Y. Liu, et al., Chin. Chem. Lett. 28 (2017) 1057-1061. DOI:10.1016/j.cclet.2017.01.020
[3]
F.T. Zangeneh, A. Taeb, K. Gholivand, S. Sahebdelfar, J. Energy Chem. 22 (2013) 726-732. DOI:10.1016/S2095-4956(13)60096-0
[4]
D.J. Childers, N.M. Schweitzer, S.M.K. Shahari, et al., J. Catal. 318 (2014) 75-84. DOI:10.1016/j.jcat.2014.07.016
[5]
U. Obenaus, F. Neher, M. Scheibe, et al., J. Phys. Chem. C 120 (2016) 2284-2291. DOI:10.1021/acs.jpcc.5b11367
[6]
S. Sokolov, M. Stoyanova, U. Rodemerck, D. Linke, E.V. Kondratenko, J. Catal. 293 (2012) 67-75. DOI:10.1016/j.jcat.2012.06.005
[7]
H.N. Pham, J.J.H.B. Sattler, B.M. Weckhuysen, A.K. Datye, ACS Catal. 6 (2016) 2257-2264. DOI:10.1021/acscatal.5b02917
[8]
H. Zhu, D.H. Anjum, Q. Wang, et al., J. Catal. 320 (2014) 52-62. DOI:10.1016/j.jcat.2014.09.013
[9]
Y. Zhang, Y. Zhou, L. Huang, et al., Chem. Eng. J. 270 (2015) 352-361. DOI:10.1016/j.cej.2015.01.008
[10]
M. Guisnet, N.S. Gnep, Catal. Today 31 (1996) 275-292. DOI:10.1016/S0920-5861(96)00018-1
[11]
S.R. De Miguel, S.A. Bocanegra, I.M.J. Vilella, A. Guerrero-Ruiz, O.A. Scelza, Catal. Lett. 119 (2007) 5-15. DOI:10.1007/s10562-007-9215-5
[12]
M.P. Lobera, C. Téuez, J. Herguido, M. Menéndez, Ind. Eng. Chem. Res. 47 (2008) 9314-9320. DOI:10.1021/ie800848c
[13]
S.J. Kim, Y. Liu, J.S. Moore, et al., Chem. Mater. 28 (2016) 4397-4402. DOI:10.1021/acs.chemmater.6b01458
[14]
Y. Wang, S. Wang, X. Guo, et al., Catal. Lett. 132 (2009) 472-479. DOI:10.1007/s10562-009-0119-4
[15]
P. Mériaudeau, A. Thangaraj, J.F. Dutel, C. Naccache, J. Catal. 167 (1997) 180-186. DOI:10.1006/jcat.1997.1531
[16]
P. Andy, M.E. Davis, Ind. Eng. Chem. Res. 43 (2004) 2922-2928. DOI:10.1021/ie030357m
[17]
J.J.H.B. Sattler, J. Ruiz-Martinez, E. Santillan-Jimenez, B.M. Weckhuysen, Chem. Rev. 114 (2014) 10613-10653. DOI:10.1021/cr5002436
[18]
S.F. Zhao, X.T. Yao, B.H. Yan, et al., Chin. Chem. Lett. 28 (2017) 1318-1323. DOI:10.1016/j.cclet.2017.03.023
[19]
F.T. Zangeneh, S. Sahebdelfar, M. Bahmani, Chin. J. Chem. Eng. 21 (2013) 730-735. DOI:10.1016/S1004-9541(13)60537-6
[20]
Z. Nawaz, X. Tang, Y. Chu, F. Wei, Chin. J. Catal. 31 (2010) 552-556. DOI:10.1016/S1872-2067(09)60071-1
[21]
R.E. Hayes, B. Liu, R. Moxom, M. Votsmeier, Chem. Eng. Sci. 59 (2004) 3169-3181. DOI:10.1016/j.ces.2004.05.002
[22]
L. Hu, K.A. Boateng, J.M. Hill, J. Mol. Catal. A:Chem. 259 (2006) 51-60. DOI:10.1016/j.molcata.2006.06.018
[23]
R. Guettel, J. Knochen, U. Kunz, M. Kassing, T. Turek, Ind. Eng. Chem. Res. 47 (2008) 6589-6597. DOI:10.1021/ie800377n
[24]
Z. Nawaz, X.P. Tang, Y. Wang, F. Wei, Ind. Eng. Chem. Res. 49 (2010) 1274-1280. DOI:10.1021/ie901465s