Reac Kinet Mech Cat (2012) 107:157–165 DOI 10.1007/s11144-012-0451-3

Synthesis, characterization and catalytic performance of binary CeO2–MgO oxides in the dehydrogenation of ethylbenzene Van-Khoa Nguyen • Jung-Hyun Park • Chae-Ho Shin

Received: 3 October 2011 / Accepted: 22 April 2012 / Published online: 24 May 2012 Ó Akade´miai Kiado´, Budapest, Hungary 2012

Abstract Binary CeO2–MgO oxides with various Ce/Mg mol ratios were prepared by the co-precipitation method and were tested in the dehydrogenation of ethylbenzene to styrene. The obtained materials were characterized by N2 adsorption, X-ray diffraction, and X-ray photoelectron spectroscopy. The N2 adsorption showed that CeO2–MgO oxides were mesoporous materials and the BET surface area increased with increasing magnesia content. The fluorite type structure of ceria was observed in samples with low magnesia content while the mixture of the ceria phase and the magnesia phase existed simultaneously in samples with high magnesia content. The results showed that the Ce4?–Mg2?–O2- couple in CeO2/MgO catalyst could be supposed to be an active site in the dehydrogenation of ethylbenzene. Keywords

Ethylbenzene  Dehydrogenation  Styrene  CeO2–MgO

Introduction Styrene, one of the most important intermediates of the chemical industry, is produced commercially by the dehydrogenation of ethylbenzene using Fe–K oxide based catalyst in presence of steam [1, 2]. The working catalyst is a metastable mixture of KFeO2, K2Fe22O34, and Fe3O4. KFeO2 is formed under the ethylbenzene dehydrogenation process, non-stoichiometric Fe3O4 (with an excess of Fe2O3) and K2Fe22O34, which serves as a source of potassium [3, 4]. Addition of ceria into the K–Fe system to improve its activity and selectivity, since cerium oxide enhances the stability of catalyst by hindering the reducibility of Fe3? to Fe3O4, enhances the polarization of the Fe–O bond and increases the basicity of the catalyst surface. V.-K. Nguyen  J.-H. Park  C.-H. Shin (&) Department of Chemical Engineering, Chungbuk National University, 12, Gaeshin-dong, Heungduk-gu, Cheongju, Chungbuk 361-763, Republic of Korea e-mail: [email protected]

123

158

V.-K. Nguyen et al.

Besides, the addition of cerium decreases the activation energy for styrene formation and the catalyst activity is increased by higher concentrations of cerium oxide [3, 5, 6]. Despite the well-known positive effect of CeO2 in the activity of K–Fe based dehydrogenation catalyst, few basic studies have been carried out to investigate its role in this reaction. A study of the role and activity of the cerium oxide based catalyst was much needed. Recently, ceria has been shown to have potential applications in the removal of soot from diesel engine exhaust [7], catalytic waste oxidation for the removal of organics from waste waters [8], and as a component of combustion catalysts [7, 9]. The ability of ceria to enhance the removal of CO, hydrocarbons and NOx makes it a critical component of the high-performance three-way catalyst for the treatment of automotive exhaust gases. The importance of CeO2 in such systems originates mainly from the ability to act as an oxygen buffer, storing/releasing oxygen due to the Ce4?/Ce3? red-ox couple [7, 10]. But pure ceria is rarely used since it is known to have poor thermal stability and undergoes rapid sintering under high temperatures, thus greatly decreasing its oxygen storage capacity [9, 11, 12, 15]. To dissolve this problem, ceria should be combined with another metal oxide. The mixture of two different oxides offers an opportunity not only to improve the performance of the involved metal oxide but also to form new stable compounds that lead to totally different physical and chemical properties from the individual components. Actually, the addition of metal oxides such as: Co3O4, MnOx, Al2O3, and ZrO2 to ceria had enhanced the activity of catalyst in the oxidation of ammonia, the oxidation of the carboxylic acids, the treatment of automotive exhaust gases in comparison to the single ceria [11–14]. In this study, the metal oxide MgO was investigated to add to ceria. One of the problem of the styrene catalyst is its deactivation by carbon deposit. CeO2 can play an important role for soot oxidation processes. Therefore, in this study, with the purpose using the ceria containing oxides as catalyst for the ethylbenzene dehydrogenation, the series CeO2–MgO oxides with the various Ce/Mg molar ratios are prepared by the co-precipitation method and characterized by the N2 adsorption, X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS). The activity of catalyst is tested in the dehydrogenation of ethylbenzene to styrene.

Experimental Preparation of samples The CeO2–MgO mixed oxides are prepared by the co-precipitation method using cerium(III) nitrate hexahydrate, magnesium nitrate hexahydrate as precursors and aqueous ammonia solution as a hydrolyzing agent. The mixture of cations with a molar ratio of Ce3? to Mg2? = 1:9, 3:7, 5:5, 7:3, and 9:1 is dissolved in 500 ml distilled water with total cation concentration of 0.4 M at 5 °C for 1 h. Hydrogen peroxide (with volume ratio of H2O2/cation solution = 3:1) was added into the cation solution. The color of the solution changes rapidly to orange. The mixed solution is hydrolyzed by ammonia solution at pH = 10 and aged in the mother

123

Binary CeO2–MgO oxides in the dehydrogenation of ethylbenzene

159

liquor at room temperature for 24 h. The yellow solid is obtained after reflux at 100 °C for 2 h. The obtained solid was dried at 110 °C for 24 h and calcined at 700 °C for 4 h in the flowing air of 2 l/min. The final product was sieved to the mesh size below 210 lm. Characterization XRD experiment was performed on a Brucker diffractometer (Model D5005) using CuKa radiation generated at 40 kV and 40 mA. The crystal structure had been determined from wide-angle diffraction patterns in the 20–90° range in 2h. Reference patterns from the international center for diffraction data (ICDD) were used for identification of bulk phases. The specific surface area, total pore volume and average diameter of the samples were obtained by N2 sorption at 77 K with a Micromeritics ASAP 2010 analyzer. XPS measurements were performed in a VG ESCALAB 210 model equipped with an MgKa (1253.6 eV) source. The base pressure of analysis chamber was *10-10 Torr and the analyzed area was *2 mm2. The binding energy scale was calibrated to Au 4f7/2 line of 83.95 eV. For charge referencing, adventitious Si 2p peak set at 103.6 eV was used. Each data set was first corrected for the non-linear emission background. The data was then fitted with Gaussian function to find the peak position. Catalytic activity The dehydrogenation of ethylbenzene was carried out in the continuous flow microreactor system at the atmospheric pressure. Prior to the studies, the catalyst was routinely activated by a temperature program. First, the oven was heated up to 350 °C with a rate of 2.5 °C/min and steam was added to the feed at this temperature. Subsequently, the temperature was further increased up to 650 °C and the reaction temperature was kept at 600 °C after being maintained at 650 °C for 2 h. Then water (partial pressure of 76.4 torr) and ethylbenzene (partial pressure of 8.4 torr) in helium as a balancing gas (GHSV = 80000 h-1) was fed into a quartz reactor containing 0.15 g of catalyst. Ethylbenzene and products were analyzed using an on-line gas chromatography equipped with flame ionization detector and capillary column CP Chirasil-Dex CB DF = 0.25. Total conversion of ethylbenzene is defined as the percentage of ethylbenzene converted to hydrocarbon products, while styrene selectivity is the percentage of styrene in the observed product.

Results and discussion Characterization of Ce–Mg–O supports The N2 adsorption/desorption isotherms and pore size distributions (PSDs) of 5CeO2–5MgO at calcination in temperature range of 550–900 °C are given in Fig. 1. The other figures of the samples at different Ce/Mg molar ratios are similar

123

160

V.-K. Nguyen et al.

and not shown here. The desorption branches of these isotherms exhibited hysteresis and correspond to the type IV isotherms. The existence of the hysteresis loop in the isotherms is due to the capillary condensation of N2 gas occurring in the mesopores and therefore, the Type IV isotherm is considered as the characteristic feature of the mesoporous materials [15]. This fact could also be observed from the PSD profiles obtained for the sample (small figure in Fig. 1). The PSDs of all samples were in the mesopore range (pore size, 2–50 nm). The pore size of these materials increases with the temperature increase from 550–900 °C. This can be explained as follows: smaller pores collapsed during calcination at higher temperatures. The N2 adsorption/ desorption isotherms of all samples with temperature increase indicated that these materials were stability at higher temperature. The evolution of the textural properties of the Ce–Mg–O oxides as a function of the magnesia and the calcinations temperature in terms of specific surface area, pore volume is shown in Table 1. The BET surface area and the pore volume value of the samples calcined at 550 °C increases from 76.5 to 138.3 m2/g and from 0.208 to 0.339 cc g-1, respectively, with increasing magnesia content from 10 to 90 mol.%. The BET surface area of the pure magnesia is higher than that of the pure ceria. Hence, the addition of magnesia into the ceria makes a binary oxide, which its surface area is higher than that of the pure ceria and lower than that of the pure magnesia. The value of the BET surface area and the pore volume decreased with the higher thermal treatment as sintering of the samples and the destruction of the smaller pores occurred at higher thermal treatment. Actually, after calcination at 700 °C, a remarkable loss was noticed which the surface area and pore volume varied from 42 to 58 % and from 18 to 32 %, respectively. This effect was seen clearly with the samples calcined at 800 and 900 °C. The surface area and pore volume value decreased drastically. These values of the sample calcined at 900 °C

5Mg-5Ce o

5Mg-5Ce

180

0,020

o

550 C o 700 C o 800 C o 900 C

0,018

160 140 120 100

Pore Volume (cm³/g·nm)

Quantity Adsorbed (cm³/g STP)

200

0,016 0,014 0,012

550 C o 700 C o 800 C o 900 C

0,010 0,008 0,006 0,004 0,002

80

0,000 1

10

100

Pore Diameter (nm)

60 40 20 0 0,0

0,2

0,4

0,6

0,8

1,0

Relative Pressure (P/Po)

Fig. 1 N2 adsorption/desorption isotherms and pore size distribution of samples 5Ce–5Mg at different calcined temperature

123

Binary CeO2–MgO oxides in the dehydrogenation of ethylbenzene

161

Table 1 Textural properties of catalysts 550 °C

Sample

700 °C

SBET (m2 g-1)

800 °C

900 °C

Vpore (cc g-1)

SBET (m2 g-1)

Vpore (cc g-1)

SBET (m2 g-1)

Vpore (cc g-1)

SBET (m2 g-1)

Vpore (cc g-1)

CeO2

62.2

0.196

38

0.157

19.3

0.117

6.9

1Mg–9Ce

76.5

0.208

44

0.169

22.0

0.123

7.0

0.073

3Mg–7Ce

90.6

0.312

48

0.212

33.3

0.175

12.9

0.106

5Mg–5Ce

95.0

0.327

51

0.230

35.3

0.210

16.7

0.131

7Mg–3Ce

107.8

0.366

54

0.268

40.8

0.239

18.6

0.110

9Mg–1Ce

138.3

0.399

58

0.272

54.9

0.278

29.3

0.146

MgO

140.7

0.906

78

0.599

49.0

0.326

47.6

0.290

800 700 550 CeO2

30

40

50

2 theta (degree)

60

70

331

220

222

200

220

MgO

400

111

111

400

331 222

220

700 550 CeO2

CeO2 200

111

900 800

20

1Ce-9Mg

MgO

MgO

MgO

20

CeO2

200

400

222

331

220

900

(c)

3Ce-7Mg

MgO

200

111 200

Intensity (a.u.)

CeO2

Intensity (a.u.)

(b)

5Ce-5Mg

Intensity (a.u.)

(a)

0.063

900

800 700 550 CeO2 MgO

30

40

50

2 theta (degree)

60

70

20

30

40

50

60

70

2 theta (degree)

Fig. 2 XRD patterns of Ce–Mg–O samples with different Mg/Ce mol ratios: (a) 5Ce–5Mg, (b) 3Ce– 7Mg, and (c) 1Ce–9Mg

decreased from 79 to 90 % and from 60 to 65 % in comparison with those of the sample calcined at 550 °C. The XRD patterns of Ce–Mg–O catalysts (Fig. 2) showed that a fluorite type structure of CeO2 was seen in all samples. The characteristic peaks of ceria phase were located on 28.5°, 33.2°, 47.5°, 56.4°, 59.2°, and 69.7° could be indexed to the (111), (200), (220), (311), (222), and (400) planes of CeO2 structure [JCPDS 43-1002]. There are no reflections of the MgO phase observed with magnesia content up to 50 mol.%. However, three characteristic peaks of the periclase MgO structure at 37.0°, 42.8°, and 62.2° in 2h in correspondence with (111), (200), and (220) planes [JCPDS 43-1022] were observed with magnesia content higher than 50 mol.%. The disappearance of the MgO phase at low magnesia content could be assumed that MgO reacted to CeO2 in the calcinating process to form the solid Ce–Mg–O product. The XPS of Mg 2p (Fig. 3a) has confirmed this opinion. There is only one component at 49.7 eV obtaining in the samples with magnesia content up to 50 mol.%, while it presents two components at 49.1 and 49.7 eV in the samples with magnesia content upper 50 mol.%. The value at 49.7 eV is attributed to the Mg 2p in the Ce–Mg–O solid phase and the value at 49.1 eV may be ascribed to Mg in the free MgO. These bands shifted much lower than the value of pure MgO

123

162

V.-K. Nguyen et al.

(a)

(b) 49.7

881.3

49.1 (e) (d)

Intensity

Intensity

(c)

(e) (d)

(b)

u'''

u''

(c)

u'

u v'''

v''

v'

v

(b) (a)

(a)

53

52

51

50

49

48

47

Binding energy (eV)

46

45

930

920

910

900

890

880

870

Binding Energy (eV)

Fig. 3 XPS of Mg 2p (Fig. 4a), Ce 3d (Fig. 4b) of Ce–Mg–O samples with different Ce/Mg mol ratios: (a) 9Ce–1Mg, (b) 7Ce–3Mg, (c) 5Ce–5Mg, (d) 3Ce–7Mg, and (e) 1Ce–9Mg

(Eb = 49.8 eV) because of the formation of the solid phase. Besides, the free MgO on the surface interacts to the electron donor groups such as OH- groups, which render Mg2? cation less positive and shifts to the lower binding energy [16]. The complex multi-peaks spectra of Ce 3d labeled following Burroughs [17, 18] are shown in Fig. 3b. The peaks labeled v correspond to the Ce 3d5/2 and those labeled u represent the Ce 3d3/2 spin orbital components. The major difference of the Ce 3d XPS features between Ce(IV) oxide and Ce(III) oxide is that the Ce 3d XPS from Ce(IV) oxide consists of three pairs of spin–orbit doublets, whereas that from the Ce(III) oxide only two pairs of spin–orbit doublets [17–20]. In our case, the Ce 3d XPS of the CeO2–MgO samples shows clearly four-pairs of spin orbital doublets. The bands v, v00 , v000 and u, u00 , u000 are attributed to Ce4?, while v0 and u0 are represented Ce3?. Taken together, the mixture of Ce4? and Ce3? exists simultaneously in the CeO2–MgO samples. Besides, the main Ce 3d5/2 peak is observed at 881.3 eV and shifts to lower energy than that of pure CeO2 (Eb = 883.8 eV). This result confirmed the opinion that the MgO reacted to ceria at high temperature to form the solid phase. Therefore, the energy band of ceria is shifted to the lower energy in comparison with the pure CeO2. The morphologies of the CeO2–MgO mixed oxides were observed by TEM (Fig. 4). It is evident from the picture that the morphologies of particles appear as the uniform in the shape. It looks like the particle morphologies of ceria aggregated to make a string of particles. The estimated particle sizes are around 10–20 nm, consistent with the crystal sizes determined from the XRD pattern from 9.3 to 18.5 nm. Fig. 4b shows the TEM of 1Ce–9Mg–O sample and their lattice image of nano-particles, the inter-planar spacing for (111) and (200) planes are determined as 0.319 and 0.267 nm, respectively. The free MgO crystal also ˚ of the observes in the 1Ce–9Mg–O sample. The lattice parameter of 5.402 A Ce–Mg–O samples is calculated and this value is rather lower than that of pure ceria prepared in the same method of 5.4067 [JCPDS 43-1002]. This result confirms that MgO reacted to ceria to form the new solid product, resulting in a decrease in the lattice parameter.

123

Binary CeO2–MgO oxides in the dehydrogenation of ethylbenzene

163

Fig. 4 TEM images of (a) 9Ce–1Mg, (b) 1Ce–9Mg, and (c) fresh CeO2 prepared in the same method

Activity of Ce–Mg–O oxides The activities of pure MgO and CeO2 materials were tested under the same reaction conditions. The results showed that MgO was not active, while the ethylbenzene conversion of few percent with low selectivity of styrene was obtained with pure ceria catalyst. The ethylbenzene conversion catalyst is improved much with the mixture of ceria and magnesia (Fig. 5). Ethylbenzene conversion increases strongly from 16 to 35 % while the styrene selectivity decreases slightly from 70 to 65 % with decreasing ceria content from 90 to 10 mol.% because of increasing of the BET surface area (Table 1). The literature [3, 5, 6, 21] indicated that the ceria improved a few percent of styrene selectivity when it was added into the K–Fe based catalyst. This study showed that ceria can act as the active phase in the ethylbenzene dehydrogenation. Weiss [22] suggested that the reaction mechanism of the ethylbenzene dehydrogenation over a-Fe2O3 included both an acid–base and redox reaction, in which iron played as a Lewis acid site and oxygen site played as Brønsted basic site. Tope [23] agreed that the reaction mechanism of the ethylbenzene dehydrogenation over Fe– Co/Mg(Al)O catalyst included both an acid–base and redox reactions, in which the Fe3?–Co3?/2 played a role of active species in the Fe3?/Fe2? reduction–oxidation process, the Fe3? played a role as Lewis acid sites and the role of Mg2?–O2- was as base sites on the catalyst. Ethylbenzene was strongly adsorbed on the Fe3? acid sites via p-bonding and the dehydrogenation was initiated by a-H? abstraction from ethyl group on Mg2?–O2- basic sites. The a-H? abstraction by O2-(–Mg2?) was likely followed by b-H abstraction, leading to the formations of styrene and H2. The XPS of Ce 3d (Fig. 3b) shows that the Ce4?/Ce3? couple always exists in the Ce–Mg–O samples. The presence of the Ce4?/Ce3? couple was a main reason for the high mobility of the lattice oxygen in the ceria. Herein, Ce4? active species could play a role similar to Fe3? species, which Ce4?/Ce3? couple played a role as the red-ox species and the role of Ce4? was as Lewis acid sites, Mg2?–O2- played a role of base sites on the catalyst. Taken together, the Ce4?–Mg2?–O2- couples could be supposed to play a role as active sites in the dehydrogenation reaction like the iron catalyst do.

123

100

100

90

90

80

80

70

70

60

60

50

50

40

40

30

30

20

20

10

10

0 0

50

100

150

200

250

300

350

Styrene selectivity (%)

V.-K. Nguyen et al.

Ethylbenzene conversion (%)

164

0 400

TOS (min)

Fig. 5 Activity of the Ce–Mg–O samples with different mol ratios: 1Ce–9Mg (M), 3Ce–7Mg (w), 5Ce–5Mg (filled triangle), 7Ce–3Mg (open circle), and 9Ce–1Mg (filled square) versus time on stream

Conclusions Binary Ce–Mg–O oxides with different Ce/Mg molar ratios have the fluorite type structure of ceria with low magnesia content, while the mixture of the fluorite structure of ceria and the MgO phase was observed with higher magnesia content. The Ce4?–Mg2?–O2- couple in CeO2/MgO catalyst could be supposed to be an active site in the dehydrogenation of ethylbenzene.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

Cavani F, Trifiro F (1995) Appl Catal A 133:219–239 Kano Y, Ohshima M, Kurokawa H, Miura H (2010) Reac Kinet Mech Catal 100:79–83 Kotarba A, Bieniasz W, Kustrowski P, Stadnicka K, Sojka Z (2012) Appl Catal A 407:100–105 Li Z, Shanks BH (2009) Appl Catal A 354:50–56 Dulamita N, Maicaneanu A, Sayle DC, Stanca M, Craciun R, Olea M, Afloroaei C, Fodor A (2005) Appl Catal A 287:9–18 Rao KN, Reddy BM, Abhishek B, Seo YH, Jiang N, Park SE (2009) Appl Catal B 91:649–656 Aneggi E, Leitenburg C, Trovarelli A (2012) Catal Today 181:108–115 Kim KH, Ihm SK (2011) J Hazard Mater 186:16–34 Colussi S, Trovarelli A, Cristiani C, Lietti L, Groppi G (2012) Catal Today 180:124–130 Martins RC, Amaral-Silva N, Quinta-Ferreira RM (2010) Appl Catal B 99:135–144 Edwin Ntainjua N, Garcia T, Solsona B, Taylor SH (2008) Catal Today 137:373–378 Wang Z, Qu Z, Quan X, Wang H (2012) Appl Catal A 411–412:131–138 Cao H, Song W, Gong M, Wang J, Yan S, Liu Z, Chen Y (2009) J Nat Gas Chem 18:83–87 Liotta LF, Di Carlo G, Pantaleo G, Deganello G (2007) Appl Catal B 70:314–322 Bhagat SD, Kim YH, Yi G, Ahn YS, Yeo JG, Choi YT (2008) Micropor Mesopor Mat 108:333–339 Ardizzone A, Bianchi CL, Fadoni M, Vercelli B (1997) Appl Surf Sci 119:253–259 Swiatowska J, Lair V, Pereira-Nabais C, Cote G, Marcus P, Chagnes A (2011) Appl Surf Sci 257:9110–9119 Burroughs P, Hamnett A, Orchard AF, Thornton G (1976) J Chem Soc Dalton Trans 17:1686–1698

123

Binary CeO2–MgO oxides in the dehydrogenation of ethylbenzene

165

19. Lykhach Y, Staudt T, Streber R, Lorenz MPA, Bayer A, Steinru¨ck HP, Libuda JY (2010) Eur Phys J B 75(1):89–100 20. Fang J, Bi X, Si D, Jiang Z, Huang W (2007) Appl Surf Sci 253(22):8952–8961 21. Hirano T (1986) Appl Catal 28:119–132 22. Weiss W, Schlogl R (2000) Top Catal 13:75–90 23. Tope BB, Balasamy RJ, Khurshid A, Atanda LA, Yahiro H, Shishido T, Takehira K, Sulaiman SA (2011) Appl Catal A 407:118–126

123