A Novel Catalyst for Synthesis of Styrene: Carbon ... - UM Repository

Journal of Nanoscience and Nanotechnology 7 (2007) 3495-3501

A Novel Catalyst for Synthesis of Styrene: Carbon Nanofibers Immobilized on Activated Carbon J.J. Delgado a,*, X.W. Chen a,*, D.S. Su a,† , Sharifah B Abd Hamid b, R. Schlögl a a

Fritz Haber Institute of the Max Planck Gesellschaft, Department of Inorganic Chemistry, Faradayweg 4-6, 14195 Berlin, Germany. b Combinatorial Technology and Catalysis Research Center, Universiti Malaya, 50603 Kuala Lumpur, Malaysia corresponding author: +49 30 8413 5406, Fax. +49 30 8413 44 05 E-mail: [email protected]

†

* These two authors has contribute equally to the present work

Abstract Carbon NanoFibers (CNFs) with hierarchically structure have been immobilized onto Activated Carbon (AC) by impregnation with an aqueous solution of Fe(CH3COO)2, reduction and subsequent chemical vapor decomposition of ethylene. The morphology of the CNFs can be modulated by adjusting the pH of the Fe(CH3COO)2 solution used for impregnating the AC. A stable yield of 35% in the oxidative dehydrogenation of ethylbenzene to styrene was obtained at a temperature of 673K, around 200K lower than the current industrial process. The immobilized CNFs on AC catalysts combine the catalytic properties of the carbon nanofibers and the suprastructure of the AC host. The final material is an easy to handle active catalyst, with an open structure of immobilized CNFs avoiding the pressure drop problem, which is typically observed for fine powder forms of CNFs. The immobilized CNFs on AC are attractive for gasphase fixed-bed industrial applications. Keywords: Carbon Nanofibers; Oxidative Dehydrogenation; Ethylbenzene; Styrene;immobilized nanocarbon; Activated Carbon 1. Introduction Carbon NanoTubes and NanoFibers (CNTs and CNFs) have become one of the most important research areas in current nanoscience and nanotechnology. Their unique physical and chemical properties attract great attention and make them attractive for applications in the development of electronics [1, 2], hydrogen storage [3, 4] and catalytic systems [5, 6]. The styrene monomer is involved in several polymer syntheses. Its synthesis is one of the ten largest industrial processes in the world today. Nowadays, the industrial production of styrene is carried out by the direct dehydrogenation of ethylbenzene at high temperatures (873-953 K) [7, 8] over a potassium promoted iron oxide catalyst in the presence of a large excess of overheated steam. This process deals with several drawbacks such as thermodynamic limitations, a large amount of wasted energy, an irreversible catalyst deactivation by coke deposition [8, 9]. During the last years, different new technologies have been developed as alternatives to the industrial process [10-12]. One of the most promising processes is the Oxidative DeHydrogenation (ODH), which is an exothermic and nonthermodynamically favoured reaction and allows operating at lower temperatures [9]. Carbon materials have been used for a long time in heterogeneous catalysis as catalyst support, or even as active components [13]. Previous studies have shown that carbon itself was catalytically active for the ODH of ethylbenzene [14-19] and recent works have reported that Preprint of the Department of Inorganic Chemistry, Fritz-Haber-Institute of the MPG (for personal use only) (www.fhi-berlin.mpg.de/ac)


high performance could be obtained using nanostratured carbons such as onion-like carbons, CNTs and CNFs [20-22]. The presence of carbonyl/quinone groups on the surface of carbon materials is the prerequisite for the activity of the catalyst [17, 20]. A redox mechanism involving quinone/hydroquinone groups on carbon surface is suggested for the ODH reaction [17-22], as it was recently confirmed by quasi-in-situ XPS experiments [23]. Nanocarbons have a promising future in ODH of ethylbenzene as they exhibit higher performance than graphite [21], and are more stable than carbon black [16-18, 21]. This may open a new horizon for designing more efficient industrial processes. However, their powdery form is unsuitable for use in fixed-bed reactors for industrial applications, because fine powders usually introduce high-pressure drops and lead to moving-bed phenomena. In addition, the fine powders nature of the nanocarbons, such as onion-like carbons, carbon nanotubes or nanofibers, make them difficult to handle. Health concerns also demand fixation of the nanoparticles in applications with gas streams. On the other hand, post synthesis compaction is not desirable because it may introduce changes of the chemical properties and inhibition of the reactant diffusion to the catalyst surface. A suprastructural design of the carbon nanomaterials can be achieved by anchoring them on a macroscopic support. It is desirable to use a carbon material as host as it will keep the continuity in chemical and physical properties, preventing deterioration of the overall compound properties and chemical instability. The similarities between the support and the nanocarbon would also favour the strong of the CNFs. An ideal substrate for immobilization of CNFs is activated carbon (AC) from natural sources. These materials are typically wellstructured in several dimensions, capable of chemical modifications and available in large amount. Activated carbons also exhibit a high surface for anchoring CNFs and have been successfully used as support in heterogeneous catalysis. In addition, anchoring CNFs on activated carbon combines the catalytic activity of CNFs and activated carbon. Figueiredo et al. has reported that activated carbon is active in the ODH of ethylbenzene [16-18]. This synergic effect makes more attractive as catalyst for styrene synthesis. Although the synthesis of supported nanostructured carbons onto different materials have been previously reported [24-26], the effect of the synthesis conditions on the textural and chemical properties of the anchored nanocarbons is a new challenge. The optimization of the CNFs characteristics by controlling the synthesis process is a crucial step to obtain final materials with high performance. The aim of this work is to demonstrate that it is possible to control the CNF growth on activated carbon and optimize their chemical and textural properties, which determinate the catalytic activity of the final material. Palm kernel, a waste product from palm oil production, was used as carbon source for the AC. The synthesis strategy consists of supporting iron on the initial AC and subsequently grow CNFs by Chemical Vapor Decomposition (CVD). Here we report on the optimization of the nanofiber synthesis by controlling the pH of the iron solution used to obtain the Fe/AC precursor. The final material combines the properties of the nanostructured carbon with the advantages of the macrostructured active carbon. The catalytic performance and stability of the immobilized CNFs for the ODH of ethylbenzene is also reported in this work. 2. Experimental section 2.1. Material synthesis The activated carbon from palm kernel shell used as substrate was supplied by NanoC, Malaysia. Before impregnation, the activated carbon was calcined at 673K for 4 h in air in order to enlarge the pore diameter of the activated carbon using the ash content as catalyst. The activated carbon suffered a weight loss of about 5.2% during the calcination process. Aqueous iron acetate solutions (0.09 M) were used for the Fe/AC catalyst synthesis. The pH value of the iron acetate aqueous solution was adjusted to 2.0 and 4.0 with concentrated acetic acid and an aqueous solution of NH3·H2O respectively. The 1 wt% Fe/AC catalysts were prepared by impregnation employing each solution. The Fe/AC precursors were dried at 333K for 12 h, treated in flowing N2 at 773K and subsequent reduced in H2 at 773K for 4 h. The Fe/AC catalyst (200 mg) was Preprint of the Department of Inorganic Chemistry, Fritz-Haber-Institute of the MPG (for personal use only) (www.fhi-berlin.mpg.de/ac)


placed in a vertical quartz reactor and flushed with a flow of He for 2 h at room temperature. Subsequently, a mixture of C2H4 and H2 (200 ml/min, C2H4/H2 is 1:1 ) was introduced into the reactor at atmospheric pressure. The temperature was increased to 973K and maintained for 2 h. After 2 h, the temperature was decreased to room temperature in helium flow. The BET surface area, CNF content and productivity of the Fe/AC are summarized in Table 1. The immobilized CNF samples are denoted as CNF-2 and CNF-4, indicating the pH 2 and 4 of the iron acetate impregnation solution used for the catalyst. 2.1. Characterization techniques The morphology of the catalytic materials was observed by SEM on a Hitachi S-4000 FEG in secondary-electron (SE) mode and backscattered-electron (BSE) mode at 15-kV accelerating voltage. TEM investigations were performed on a Philips CM 200 LaB6 operating at 200 kV. 2.3. Catalytic tests The catalytic tests were carried out at 673K using 120 mg of catalyst. An ethylbenzene partial pressure of 2666 Pa (i.e. 2.6 vol.%) and a total flow of 5 ml.min-1, were used. The oxygenethylbenzene ratio was 5:1. The reactants and reaction products were analyzed by an on-line Varian 3800 gas chromatograph. An ethylbenzene conversion lower than 1% was obtained in a blank test. It is pointed out that no steam was added as it is necessary in the actual industrial process. 3. Results and discussion Figures 1a and 1b show low magnification SEM images of the AC before and after calcination in air at 673K, showing that the mild oxidation removes carbon debris from the surface and cleans the small pores of the active carbon. A significant increase in the BET surface area is observed as a result of the treatment (Table 1), which is in good agreement with the enlargement of the pore size. The calcination conditions also remove the soft carbon present in the AC, preventing burning of the catalyst in subsequent ODH reaction.

Figure 1 : SEM images of the AC before calcination (a), after calcination (b). Samples CNF-2 (c) and CNF-4 (d) are also included. Preprint of the Department of Inorganic Chemistry, Fritz-Haber-Institute of the MPG (for personal use only) (www.fhi-berlin.mpg.de/ac)


Figures 1c and 1d show the samples obtained after exposing the Fe/AC catalysts to the Chemical Vapour Descomposition (CVD). It is observed that in both cases the CNFs were formed and covered the surface of activated carbon. The SEM images show that the CNFs are not only found on the outer surface of the host, but also fill the pores of the host. They exhibit a random orientation and are highly entangled. This homogenous suprastructure gives a large void volume between each nanofiber, preventing the pressure drops problems that loose CNFs may cause and thus decreases the appearance of hot spots during the exothermic ODH reaction. Table 1 Productivity and BET specific surface area of the CNFs/AC composites Sample

pH

AC AC-400 CNF-2 CNF-4

--2 4

Productivity (g C/g Fe) --57.5 95.4

CNFscontent (weight content %) --36.5 47.8

SBET( m2·g-1) 1081 1490 32.4 20.5

Figure 2: SEM of samples CNF-2 (a) and CNF-4 (b). TEM images of CNF microstructure of samples CNF-2 (a) and CNF-4 (b).

The main difference between the CNFs observed in each sample is the diameter. According to the high resolution SEM images in Figure 2 (a, b), the nanofibers in sample CNF-2 have a significantly smaller diameter (20-40 nm) than in the case of sample CNF-4 (250-500 nm). Therefore, using a solution with a lower pH value results in a catalyst with a higher relative surface area for the same CNF content. After CVD, the BET surface area decreases in comparison with that of the initial AC (Table 1). This can be explained by the fact that CNFs typical exhibit small values of surface area [19]. However, obtained surface areas were smaller that expected from a physical mixture of CNFs and AC. As example, a physical mixture of the Preprint of the Department of Inorganic Chemistry, Fritz-Haber-Institute of the MPG (for personal use only) (www.fhi-berlin.mpg.de/ac)


same amount of CNFs and AC would lead to BET areas higher than 540 m2·g-1, due to the contribution of the activated carbon. This phenomenon looks more important in the case of CNF2 what may be explained by the presence of iron particles in the support micropores, leading to pore-blockage of the carbon during the CVD process. The surface area loss is not desirable for catalytic application and it should be optimized. Nevertheless, different authors have demonstrated that during ODH of EB, the micropores of AC are blocked by polymerization of ST and do not contribute to the steady state performance [15, 18]. The transmission electron micrographs in Figures 2c and 2d reveal that the CNFs present a herringbone structure, exposing an abundance of prismatic faces to the gas phase as required for the formation of oxygen functional groups, which are necessary for applications as catalyst in ODH reactions [19, 21, 23]. No significant structural differences were found between CNF-2 and CNF-4, indicating that the pH of the iron solution only determine the CNF diameter and does not play an important role in the microstructure of the carbon nanofibers.

Figure 3: SEM images of immobilized CNFs. using SE (a, b) and BSE (c, d) modes for samples CNF-2 (a, c) and CNF-4 (b, d).

Figure 3 shows SEM images of immobilized CNFs using SE (a, b) and BSE (c, d) modes for all the samples. By comparison of the contrast, the iron catalyst particles can be located as bright objects in the backscattered electron micrograph. Some metal particles where found on the tips of the CNFs, although a few of them were also found in the middle of the carbon nanofibers. According to this, a “tip-growth” model may be suggested for the growth of CNFs, which has been previously proposed for the synthesis of similar materials [25]. The iron particles found in CNF-4 are larger than those found in the CNF-2, in agreement with the CNF diameters. Therefore, the pH adjustment determines the metal particle size, leading to smaller CNFs diameter during the CNF synthesis by CVD process. The fast and not well controlled precipitation of the iron hydroxides on the AC surface may be prevented by decreasing the pH value. It is well known that the pH also influence in the interaction of the support surface with solved Fe3+ cations, determining the iron anchoring on the AC surface. The catalytic tests show that both samples are active and selective catalysts even at low temperatures. The performances of the catalysts, in terms of ethylbenzene (EB) conversion and selectivity to styrene (ST), were recorded as a function of time on stream and plotted in Figure 5.a. An initial activation period was observed for both samples. At the beginning of the reaction, Preprint of the Department of Inorganic Chemistry, Fritz-Haber-Institute of the MPG (for personal use only) (www.fhi-berlin.mpg.de/ac)


the conversion increases significantly from values below 10% to 45% in the case of the sample CNF-4 and 55% for the sample CNF-2. A comparable behaviour was reported by G. Mestl et al. [19, 26] for CNFs and CNTs. According to the authors, the origin of the activation period was related with the combustion of soft coke observed over the graphite layers, where the active catalytic centers would be located. The amorphous carbon would inhibit the reactant diffusion to

Figure 4: a) EB conversion () and selectivity () as a function of time on stream obtained on different AC supported CNFs: CNF-2 (empty) and CNF-4 (filled). b) The inset figure shows the carbon balance of both samples.

the active centers. During the induction period the combustion of such soft carbon occurs and the activity increases. In our case, the amount of carbon in the stream was higher than what expected according to the initial EB concentration during the activation period, indicating partial combustion of the samples (Figure 4.b). After the induction period the carbon balance calculated with EB, ST, CO and CO2 was mainly close to 100%, indicating the stability of the catalyst under operating conditions. The carbon dioxide is due to the total combustion of EB. Our data can easily be explained by the fact that disordered carbonaceous impurities are usually formed in a CVD process and liable to burn at the beginning of the catalytic test. It is pointed out that, in both cases, a decrease in the styrene selectivity, from 95% to 66%, is observed. It is apparent from these results that the final graphitic active centers favour the combustion of EB indicating a higher oxidizing capacity of these active sites in comparison with the initial amorphous layer. After the activation period, the macroscopically supported nanocarbon catalyst was quite stable with time on stream, even under severe reaction conditions, i.e. at a temperature of 673K in the presence of excess oxygen, similar to the behaviours shown by ordered sp2- bound nanocarbons in powder form [20-22]. The present catalyst performance differs from carbon black, which progressively burned off and deactivated on stream [18, 21]. The main differences of the catalytic performances are the time required for the activation period and to obtain a stable conversion. The activation period of the sample CNF- 2 is only 5 hours, while for the sample CNF-4 is around 10-15 hours. The low temperature used for the catalytic test may be the reason for the long induction period, because of the slow combustion rates. Our data also indicate that CNF-4 contains a higher amount of amorphous carbon than CNF-2. The final selectivity of the two samples is similar, which can be explained by the similar microstructure observed after the activation period. In terms of conversion, the sample CNF-2 exhibited the highest activity, probably due to the smaller fiber diameter resulting in a higher relative surface area. Figure 5 shows SEM and TEM images of CNF-2 and CNF-4 after ODH reaction at 673K for 100 h. It can be seen that there is no significant change of the samples after reaction. The high resolution TEM images of CNF-2 and CNF-4 after reaction indicate that the fine structure of nanocarbon still remains intact after reaction, which means that the CNFs/AC composites, after Preprint of the Department of Inorganic Chemistry, Fritz-Haber-Institute of the MPG (for personal use only) (www.fhi-berlin.mpg.de/ac)


the activation period, are stable under the testing conditions. This opens a new horizon for the application of carbon nanofibers on AC for ODH of EB to ST.

Figure 5: SEM images of immobilized CNFs on AC after ODH reaction of ethylbezene (a) CNF-2, (b) CNF-4. TEM images of microstructure of samples CNF-2 (c) and CNF-4 (d).

4. Summary Immobilized CNFs on modified AC have been synthesized using chemical vapour deposition of ethylene on 1 wt% Fe/AC catalysts prepared with aqueous solutions of iron acetate at different pH values. It is found that the pH values of the iron acetate solutions have a great influence on the morphology and productivity of the CNF/AC composite. A higher pH value of the iron acetate solution leads to higher productivity and larger diameters of the CNFs on the AC. Both the tested samples were active in the ODH of EB to ST, exhibiting an activation period of among 5-15 hours, depending on the diameters of the CNFs. The catalytic tests carried out over a period of more than 4 days showed an exceptionally stable performance, in contrast with typical performances of AC, which initially shows a high performance that decreases with time in the stream to values around 20% [15, 18]. It should be pointed out that the present materials exhibit activities similar to those of CNFs supported on graphitic carbon previously synthesized in our lab [27]. However, the low cost of AC in comparison to graphitic supports is another important point in favor of the CNFs/AC catalyst. The CNFs/AC catalyst also exhibits a higher density which leads to the design of smaller reactors making them more attractive for industrial applications. Acknowledgments This work has been funded by the European Union FP6 program (CANAPE project). Preprint of the Department of Inorganic Chemistry, Fritz-Haber-Institute of the MPG (for personal use only) (www.fhi-berlin.mpg.de/ac)


References [1] S. Fan, M. G. Chapline, N. R. Franklin, T. W. Tombler, M. A. Cassell and H. Dai, Science 283, 512 (1999) [2] T. Rueckes, K. Kim, E. Joselevich, G. Y. Tseng, C. L. Cheung and C. M. Lieber, Science 289, 94 (2000) [3] A. C. Dillon, K. M. Jones, T. A. Bekkedahl, C. H. Kiang, D. S. Bethune and M. J. Heben, Nature 386, 377 (1997) [4] J. Lin, Science 287, 1929 (2000) [5] J. M. Planeix, N. Coustel, B. Coq, V. Brotons, P. S. Kumbhar, R. Dutartre, P. Geneste, P. Bernier and P. M. Ajayan, J. Am. Chem. Soc. 116, 7935 (1994) [6] N. M. Rodriguez, A. Chembers and R. T. K. Baker. Langmuir 11, 3862 (1995) [7] D.H. James and W.M. Castor, Styrene. In: Ullmann’s encyclopedia of industrial chemistry, John Wiley & Sons (1994) vol A25, p 329-44 [8] E. H. Lee, Catal Rev. 8, 285 (1973) [9] F. Cavani and Trifiro, Appl. Catal. A: Gen. 133(2), 219-39 (1995) [10] A. Sun, Z. Qin and J. Wang, Appl. Catal. A: Gen. 234(1,2), 179-89 (2002) [11] J. Matsui, T. Sodesawa and F. Nozaki, Appl. Catal. A:Gen. 67(2), 179 (1991) [12] S. S. E. H. Elnashae, B. K. Abdallah, S. S. Elshishini, S. Alkhowaiter, M. B. Noureldeen and T. Alsoudani, Catal Today 64(3-4), 151 (2001) [13] F. Rodriguez-Reinoso, Carbon 36(3), 159 (1998) [14] G. Emig and H. Hofmann, J. Catal. 84(1), 15 (1983) [15] A. Guerrero-Ruiz and I. Rodriguez-Ramos, Carbon 32(1), 23-29 (1994) [16] M. F. R. Pereira, J. J. M. Orfao and J. L. Figuereido, Appl. Catal. A: Gen. 184(1), 153 (1999) [17] M. F. R. Pereira, J. J. M. Orfao and J. L. Figuereido, Appl. Catal. A: Gen 196(1), 43 (2000) [18] M. F. R. Pereira, J. J. M. Orfao and J. L. Figuereido, Appl. Catal. A: Gen 218(1-2), 307 (2001) [19] G. Mestl, N. I. Maksimova, N. Keller, V. V. Roddatis and R. Schlögl, Angew. Chem. Int. Ed. 40(11), 2066 (2001) [20] N. Keller, N. I. Maksimova, V. V. Roddatis, M. Schur, G. Mestl, Y. V. Butenko, V. L. Kuznetsov and R. Schlögl, Angew. Chem. Int. Ed. 41(11), 1885 (2002) [21] D.S. Su, N. Maksimova, J. J. Delgado, N. Keller, G. Mestl, M. J. Ledoux and R. Schlögl, Catalysis Today 102-103, 110 (2005) [22] M. F. R. Pereira, J. L. Figueiredo, J. J. M. Órfão, P. Serp, P. Kalck, Y. Kihn, Carbon, 42(14), 2807 (2004) [23] J. A. Maciá-Agulló, D. Cazorla-Amorós, A. Linares-Solano, U. Wild, D. S. Su and R. Schlögl, Catalysis Today 102-103, 248 (2005) [24] M. J. Ledoux and C. Pham-Huu, Catalysis Today 102-103, 2 (2005) [25] D. S. Su, X. Chen, G. Weinberg, A. Klein-Hofmann, O. Timpe, S. B. A. Hamid and R. Schlögl, Angew. Chem. Int. Ed. 44, 5488 (2005) [26] N. I. Maksimova, V. V. Roddatis, G. Mestl, M. Ledoux, R. Schlogl, Eurasian ChemicoTechnological Journal 2(3-4), 231 (2000). [27] J. J. Delgado, R. Viera, G. Rebmann, D. S. Su, N. Keller, M. J. Ledoux and R. Schlögl, Carbon 44, 809 (2006)

Preprint of the Department of Inorganic Chemistry, Fritz-Haber-Institute of the MPG (for personal use only) (www.fhi-berlin.mpg.de/ac)