Oxidative dehydrogenation of propane over carbon nanofibers

Oxidative dehydrogenation of propane over carbon nanofibers

159 Studies in Surface Science and Catalysis, volume 159 Hyun-Ku Rhee, In-Sik Nam and Jong Moon Park (Editors) © 2006 Elsevier B.V. All rights reserve...

476KB Sizes 0 Downloads 70 Views

159 Studies in Surface Science and Catalysis, volume 159 Hyun-Ku Rhee, In-Sik Nam and Jong Moon Park (Editors) © 2006 Elsevier B.V. All rights reserved

745

Oxidative dehydrogenation of propane over carbon nanofibers Zhi-jun Sui, Hug-hong Zhou and Ying-chun Dai State Key Laboratory of Chemical Engineering, East China University of Science & Technology, Shanghai, 200237, P. R. China 1. INTRODUCTION As a new kind of carbon materials, carbon nanofilaments (tubes and fibers) have been studied in different fields [1], But, until now far less work has been devoted to the catalytic application of carbon nanofilaments [2] and most researches in this field are focused on using them as catalyst supports. When most of the problems related to the synthesis of large amount of these nanostructures are solved or almost solved, a large field of research is expected to open to these materials [3]. hi this paper, CNF is tested as a catalyst for oxidative dehydrogenation of propane (ODP), which is an attractive method to improve propene productivity [4]. The role of surface oxygen complexes in catalyzing ODP is also addressed. 2. EXPERIMENTAL The CNF was synthesized by catalytic deposition of 80 vol% CO/H2 mixtures on a 20 wt% NiFe (molar ratio 1:1) alloy catalyst. To remove metallic inclusions, the CNF was repeated washed in 2 mol/L HC1 over a period of 7 days. Then, the CNF sample was filtered, washed by a large amount of deionized water until the pH of filtrate was close to 7 and dried at 120 "C overnight. This sample was named CNF-R. Before testing as catalysts for ODP, CNF-R was calcined in air at 5001] for 2 h and designated as CNF-RA hereafter. To improve the graphitization extent of CNF, the as-grown CNF was treated at 1700 °C for 12 h under protection of argon, which was referred to CNF-HT. Prior to test as catalyst for ODP, the CNF-HT needed to be further oxidized. Three methods are used: 1) calcined in air at 600 "C for 1 h (CNF-HA); 2) immersed in the mixture of concentrated HNO3 and H2SO4 (1:1) for 1 day (CNF-HL); 3) CNF-HL was treated in Ar at 800 "C for 2 h; the sample temperature decreased to 60 "C under Ar protection; and then contact with air at 60 "C for 2 h (CNF-HB). Catalytic experiments were performed at 1 atm pressure in a conventional fixed bed flow reactor made of stainless steel. The CNF catalysts were charged (0.8 g, particle size < 0.18 mm) without inert diluents. The catalysts were placed on quartz wool in the isothermal zone of the reactor. Free volume of the reactor was packed with silica and quartz wool. The blank run results showed that the homogeneous reaction could be ignored under the experimental conditions used in this work. The reaction gases were composed of 4 vol % propane, 8 vol% oxygen and balance Ar (total flow rate is 100 ml/min). Analyses of reactants and products were carried out by using two separate on-line gas ehromatographs (Agilent 4890D) with

746

TCD detectors, using a HP-Plot Q capillary column (30m x 0.53mm) for hydrocarbons and a TDX packed column (2m) for permanent gases. Structures of the CNF were characterized by HRTEM (JOEL JSM2010, Japan), The textural properties had been obtained from N2 adsorption-desorption isotherms (ASAP 2010, Micromeritics, USA) at - 1 9 6 1 after out-gassing the samples at 190 t and 1 mmHg for 6 h. X-ray diffraction (XRD) was performed on a Rigaku D/Max2550VB/PC (Rigaku, Japan, Cu Ka radiation). Thermal gravimetry (TG), using air as carrier gas, was used to characterize the structure stability of the CNF and determine the ash content. Temperature-programmed surface reaction (TPSR) was performed to determine the role of surface oxygen complexes on catalyzing ODP, The samples were first temperature-programmed desorption in Ar to 650 "C, cooled in Ar to 15013, and then the sample temperature was raised at a constant ramping rate of 10 "U/min in a gas mixture of 8 vol% QjHg/Ar (total flow rate, 50 ml/min). Concentrations of Ar, CjHg, C3H& C2H4, CH4, COj and CO were calculated by the signal intensities of mass 40, 39, 41, 27, 16, 44 and 28 respectively. No C1-C2 hydrocarbon is detected by quadrupole mass spectrometer (Questor, ABB Extrel, USA). 3, RESULTS AND DISCUSSION 3.1. Catalyst Structure HRTEM picture (Fig.l(a)) shows that CNF-R is estimated of a diameter of 3O~40 nm and has a hollow core. The graphene layers are about 15~20 ° inclining to the axis. After heat treatment, CNF-HT, the structure remained, but graphene layers stack more regularly. The BET surface area of CNF-R and CNF-HT are 152.5 and 141.6 m2/g respectively. After oxidation, the surface areas increase a little (160-170 m2/g). All the samples have small micropore volumes (<0.008 cm3/g). XRD results imply a 0.339 nm dW2 spacing for CNF-HT, which is smaller than that of CNF-R (0.341 nm). The onset weight loss temperature (temperature needed to reach 5 % burn-off) of the CNF-HT is 660 V, while the CNF-R is only 540 t . These results indicate the graphitizaton extent of CNF increases after heat treatment. The ash content of CNF-HT is 0.06 wt%. So, the side effects of the metal on catalytic performances can be ruled out.

10 nm

10 nm

(a) Fig. 1. HRTEM images of CNF-R(a) and CNF-HT(b)

(b)

3,2, Catalytic performances The catalytic products for all the CNF catalysts include C3H6, COx and trace amount of

747

C2H4. Propene selectivity as a function of propane conversion is studied on CNF-RA under different reaction conditions. The results are shown in Fig. 2. The CjHg selectivity decreases with the increase of CjHg conversion, which implies consecutive reaction of C3H6 degradation. Apparently, a high reaction temperature and Oi/CsHg ratio is favored for high propene yield. Extrapolating the results in Fig.2 to zero conversion shows that the theoretical propene selectivity of ODP on CNF-RA increases when the reaction temperature is raised. But the theoretical propene selectivity does not exceed 80% on this sample. Except that propane undergoes oxidative dehydrogenation to form propene and combustion of propane and propene to form COx, CNF gasification (when the temperature is above 450*6) should be included. As mentioned above, high propene yield and theoretic selectivity could be achieved at high reaction temperature and O2/C3H1 ratio, which are also benefit for CNF gasification and restrain achieving higher propene yields on CNF-RA. It is found that the CNF-HT has not catalytic activity for ODP. After oxidation, all the three samples show highly catalytic performances, which are shown in Fig.3. CNF-HL has the longest induction period among the three samples, and it has relatively low activity and propene selectivity at the beginning of the test. During the induction periods, the carbon balance exceeds 105% and then fall into 100±5%, which implies the CNF structure is stable and the surface chemistry of CNF reaches a dynamic equilibrium eventually. These results indicate that the catalytic activity of ODP can be attributed to the existence of surface oxygen complexes which are produced by oxidation. The highest propene yield(18.96%) is achieved on CNF-HL at a 52.97% propane conversion.

«5"C,O 1 K,H,=1:2 40(fC,O l K J H,=J:1 S7S*C,OJC H =2:1

3

8

S

12

15

Propane Conversion {%)

Fig.2. Propene selectivity as a function of propane conversion over CNF-RA

us

m

5 on stream (mln)

250 300

50

100 150 200 Time on stream (mini

Fig. 3. Evolution of catalytic performances with time for CNF catalysts; reaction temperature, 550 t : ; W/F=37.9 (mol CjH»)/g h

748

— 8.8

H 8,7 *

C3H8 C3H6 - — CO, CO

N / \A.""' /

Sample Temperature (°G)

(a)

360 400 SOD Sample Temperature {°G)

(b)

Fig. 4. TPSR results of CNF-RA(a) and CNF-HA(b)

3.3. TPSR results TPSR results are presented in Fig. 4. Propene is produced when the sample temperature is above 350 "C on both samples, which means converting of propane over CNF catalysts could occur without oxygen. The desorption products amounts are 0.35 and 0.26 mmol/g for CNF-RA and CNF-HA respectively while the percentages of propene in the desorption substances over these two sample are 51.4% and 87.7%. These results imply that the propene selectivity may increase, at least partly, due to restriction of oxidation of propane to COx by heat treatment at the cost of catalytic activity. It's worth pointing out that no products (propene and COx) is detected by QMS in TPSR runs under the following circumstances: 1) TPD of CNF-RA to 1Q00TC, then cooled it down to 150°C in Ar before the TPSR runs; 2) CNF-HT without calcinations in air; 3) no CNF sample is presented. Based on these results we can deduce that converting of propane over CNF occurred on their surface oxygen complexes containing C=O bonds (carbonyl-like groups and basic oxides), which could exist on CNF surface after TPD to 650 'C [5], These structures are thought to be responsible of catalyzing redox reactions on carbon surfaces [6]. 4. CONCLUSIONS CNF could be the effective catalyst for ODP, but the high propene yield can only be achieved at high reaction temperature and G2/C3H8 ratio. Heat treatment of CNF at 1700"C for 12 h could increase its graphitization extent and enables it to operate at 550 "C without apparent gasification. TPSR results show that carbonyl-like groups could be the active sites for ODP over CNF. Heat treatment also could restrain the side reaction of oxidation of propane to COx. ACKNOWLEDGEMENT The author would thank the National Science Foundation of China (NO. 20490200 and NO. 20376021) for the finical support REFERENCES [1] K.RDe Jong and J.W. Geus, Catal Rev.- Sci. Eng, 42 (2000) 48, [2] M. J. Ledoux, R. Vieira, and C. Pham-Huu, J. Catal, 216 (2003) 333. [3] P. Serp, M. Comas and P. Kalack, Appl Catal, A, 253 (2003) 337. [4] E Cavani and F. Trifira, Catal. Today, 24 (1995) 307. [5] J. L. Figueiredo, F. R. Perelra, M.M.A. Freitas and I, J. M. 6rSo, Carbon, 37 (1999) 1379. [6] C.A. Leon y Leon D. and L.I, Radovic, Chem, Phys. Carbon, 23 (1993) 213.