HY catalysts

HY catalysts

Applied Catalysis A: General 241 (2003) 15–24 Rare earth-modified bifunctional Ni/HY catalysts Dao Li, Fang Li, Jie Ren, Yuhan Sun∗ State Key Laborat...

285KB Sizes 1 Downloads 36 Views

Applied Catalysis A: General 241 (2003) 15–24

Rare earth-modified bifunctional Ni/HY catalysts Dao Li, Fang Li, Jie Ren, Yuhan Sun∗ State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, The Chinese Academy of Sciences, 030001 Taiyuan, PR China Received 6 February 2002; received in revised form 24 May 2002; accepted 25 July 2002

Abstract Rare earth elements (La, Nd, Sm, Gd and Dy) were characterized by NH3 -TPD, Py-FTIR and CO-DRIFTS and tested for n-octane hydroconversion. The Bronsted acidity of modified catalysts increased as the rare earth ionic radius became longer. A similar relationship between metallic function and ionic radius was also observed. Among rare earth elements, lanthanum improved the nickel dispersion and then the ratio of metallic to acid functions. This led to the high catalytic activity and isomerization selectivity of Ni/HY catalyst. However, the catalysts modified by the other rare earth elements showed poor catalytic performance due to the reduction of Bronsted acidity and nickel dispersion, especially a lower ratio of metallic to acid functions. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Rare earth elements; Bifunctional catalyst; Metal/acid balance; n-Octane hydroconversion

1. Introduction Bifunctional catalysts are widely used in various industrial processes, such as hydroisomerization, hydrocracking, and dewaxing [1,2]. Usually, such reactions take place via (de)hydrogenation on metallic sites and isomerization/cracking on acid sites [3,4]. Thus, the ratio of metal to acid functions is crucial in determining the activity, selectivity and stability of these catalysts. Among bifunctional catalysts, rare earth ion-exchanged faujasites enjoy a position of considerable technological importance due to their superior catalytic properties. Rare earth elements could prevent aluminum loss from Y zeolite structure and could enhance the structural resistance to the severe hydrothermal conditions in the process of fluid catalytic cracking [5,6]. They ∗ Corresponding author. Tel.: +86-351-4041627; fax: +86-351-4041153. E-mail address: [email protected] (Y. Sun).

also influenced the acidity of catalysts inside the zeolite through their hydrolysis, which in turn changed the catalytic activity and the product distribution [7,8]. However, only lanthanum and cerium have been studied widely in relation to their catalytic performances, while others were seldom reported separately [9–11]. Recently, Sedran et al. reported a linear correlation between the ionic radius of rare earth cations and the strength of the hydroxyls associated to rare earth cations, which was caused by the differences of the hydrolysis over hydrated rare earth cations [7]. And the relative significance of hydrogen transfer reactions was also found to increase linearly as a function of the catalyst Bronsted acidity on various rare earth cation-exchanged NaY zeolites [12]. Furthermore, the positions of rare earths in zeolite matrix and the change of zeolite acidity caused by rare earths were studied thoroughly [13,14]. However, the influence of rare earth elements on bifunctionalty of the catalysts, especially on the balance between metal and acid functions, have not been elucidated clearly.

0926-860X/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 6 - 8 6 0 X ( 0 2 ) 0 0 4 5 4 - 4

16

D. Li et al. / Applied Catalysis A: General 241 (2003) 15–24

To shed more light onto the effect of different rare earths on the metal/acid balance of bifunctional catalyst, Ni/HY catalyst was modified by different rare earth elements (La, Nd, Sm, Gd and Dy) in the present work. The reactivity of catalysts was tested by n-octane hydroconversion, and the surface properties were investigated by ammonia temperature-programmed desorption (NH3 -TPD), Py-FTIR and CO-DRIFTS.

2. Experimental 2.1. Preparation

placed into the measurement cell with CaF2 windows. The evacuation at 773 K (ca. 10−4 Torr) for 4 h was carried out prior to adsorption of pyridine. IR spectra were recorded after subsequent evacuation at increasing temperatures from 423 to 723 K (1 h at each temperature). Diffuse reflectance infrared Fourier transform spectroscopy of carbon monoxide adsorption (CO-DRIFTS) was employed on Nicolet Magna 550 FTIR with a resolution of 8 cm−1 . The sample was finely grounded and pressed into the IR cell. After pretreatment and reduction, the sample was cooled down in Ar flow to room temperature, then CO gas was admitted, followed by purging with Ar flow. CO-FTIR spectra were all recorded at room temperature.

NaY with Si/Al ratio of 5.1 supplied by Research Institute of Petrochemical Process was used as starting material. HY zeolite was prepared by ion-exchange of NaY with 1 mol/l NH4 NO3 solution repeated three times. The material was then calcined at 773 K. So prepared HY zeolite was immersed in a solution of 0.1 mol/l rare earth nitrate (La, Nd, Sm, Gd and Dy, respectively) and then slowly evaporated to dryness. The resultant was dried at 383 K for 12 h and calcined at 773 K for 4 h to obtain the rare earth-modified zeolite. The catalysts were prepared by impregnating modified HY zeolites with an ammonia solution of nickel nitrate (pH 12), and then dried at 383 K overnight and calcined at 773 K for 4 h. All the catalysts contained 5.7 wt.% of nickel and 6 wt.% of rare earth oxides.

n-Octane hydroconversion was conducted in a fixed-bed downstream reactor (stainless steel tube with 12 mm i.d.). Prior to reaction, 0.80 g of catalyst was pretreated in Ar flow and then reduced in H2 flow at 773 K for 2 h. The reaction was performed at 533–593 K, 2.4 MPa and H2 /n-octane molar ratio of 8. The effluent was in situ analyzed with a gas chromatograph with a GDX-103 column and FID detector, a ov-101 capillary column and an FID detector for the liquid products.

2.2. Characterization

3.1. Acidity of catalysts

Ammonia temperature-programmed desorption was measured by a flow system with a thermal conductivity detector. The sample of 200 mg was outgassed in argon flow at 773 K for 30 min, which followed ammonia-saturation by flowing NH3 /Ar stream at 373 K for 10 min. After equilibration in argon flow for 1 h at 373 K, the catalyst was heated in a linear rate of 10 K/min to 773 K, and the detector signal of ammonia desorption was recorded. FTIR spectra of adsorbed pyridine were recorded on a Nicolet Magna 550 Fourier transform infrared spectrometer at 4 cm−1 resolution. The sample was finely grounded and pressed into a self-supporting wafer (10 mg/cm2 , diameter = 15 mm); this was then

Temperature-programmed desorption of ammonia was carried out to compare the acidity of catalysts. In order to obtain the acid distribution, the experimental curves were fitted by Gaussian deconvolution (see Fig. 1), in which three peaks were used to represent the weak, medium and strong acidity, respectively [15,16]. The total acidity and the acid distribution of the samples are listed in Table 1, according to NH3 -TPD profiles and deconvoluted peaks. Obviously, the modification led to the decrease in both total acidity and weak acidity. However, La-modified catalyst showed a maximum for medium acidity, and Laand Nd-modified catalysts showed a higher amount of strong acid centers.

2.3. Catalytic test

3. Results and discussion

D. Li et al. / Applied Catalysis A: General 241 (2003) 15–24

17

Fig. 1. NH3 -TPD curves and Gaussian deconvoluted peaks.

Table 1 Catalyst acidity: total acidity and Gaussian deconvoluted peaks Catalysts

Total acidity (mmol/g)

Gaussian deconvoluted peaks Weak acidity T

Ni/HY Ni-La/HY Ni-Nd/HY Ni-Sm/HY Ni-Gd/HY Ni-Dy/HY

1.73 1.47 1.04 0.72 0.89 0.82

(◦ C)

240 240 250 260 260 260

Medium acidity (◦ C)

A (mmol/g)

T

1.00 0.62 0.44 0.25 0.42 0.33

340 330 330 340 340 340

Strong acidity

A (mmol/g)

T (◦ C)

A (mmol/g)

0.55 0.62 0.35 0.31 0.30 0.33

410 410 410 410 410 410

0.18 0.23 0.25 0.15 0.17 0.16

18

D. Li et al. / Applied Catalysis A: General 241 (2003) 15–24

Fig. 2. Py-FTIR profiles of different catalysts.

Py-FTIR analysis was used to discriminate acid type in the supercages and out-surface of catalysts. As it revealed, the bands at 1453, 1545 and 1490 cm−1 (representing Lewis acid, Bronsted acid and both, respectively) were observed for all catalysts (shown in Fig. 2). Similar to the change of total acidity, Bronsted and Lewis acidity decreased with the introduction of rare earths. And it was interesting that the Bronsted and Lewis acidity of modified catalysts increased along with the rare earth ionic radius (see Fig. 3). The modification over the acidity that occurred when a rare earth was incorporated to an acid support might be due to (1) blockage of the pores and superficial sites by large-sized rare earth oxide particles that obstructed ammonia access to acid sites, and (2) deactivation of acid sites by the strong interaction between the rare earth cations and the catalyst surface [17]. In the present work, hydrated rare earth cations were restricted to supercages at room temperature because of their size [18,19]. In this case, little change in the structure of zeolites took place, since the interaction between hydrated cations and zeolite framework was generally weak. But on calcination at 773 K, the hy-

dration sheath would be activated and stripped from the cations, which allowed the cations to migrate from supercages to small cages through six-member rings [8]. As a result, some of rare earth cations would replace the framework protons to balance the negative framework charge by means of solid-state ion exchange [20] and the original acidity of different strength over HY zeolite surface disappeared to some extent. On the other hand, rare earth cations would be hydrolyzed to produce protons at high temperatures [21], which combined with bridged oxygen over the catalyst surface and created new Bronsted acid sites. The newly generated Bronsted acid sites associated with rare earths would compensate the original acidity partly, but not completely. Therefore, the acidity of rare earth-modified catalysts was less than that of Ni/HY catalyst (see Table 1). Sousa-Aguiar et al. [7] suggested that the increasing ionic radius of rare earth cations would facilitate the hydrolysis, which meant that the shorter the ionic radius of rare earth cation was, the less Bronsted acidity was generated. Therefore, Ni-La/HY catalyst possessed the largest acidity, since the hydrolysis of lan-

D. Li et al. / Applied Catalysis A: General 241 (2003) 15–24

19

Fig. 3. Influence of the ionic radius of rare earth cations on the acidity of catalysts.

thanum cation was much more effective than the others. And the Bronsted acidity of modified catalysts increased along with the rise of rare earth ionic radius (see Fig. 3). Moreover, the newly generated Bronsted acid sites presented medium and strong acidity, which led to the improvement of the ratio of medium and strong acidity for the modified catalysts (see Table 1). Thus, comparing with the weak acidity, one sees that the medium and strong acidity of modified catalysts decreased much more slowly. For lanthanum-modified catalyst, the amount of medium and strong acid sites was even larger than that of Ni/HY catalyst, though its total acidity decreased. 3.2. CO-DRIFTS

co-existed in equilibrium in each catalyst. As the formation of Ni(CO)4 indicated the high dispersion of nickel species, the amount of K–M function could represent the divergence of nickel species due to the same nickel content in all catalysts. Lanthanum was favorable to improve the nickel dispersion, as reported by others [23]. The other rare earth elements reduced the divergence of metal particles. Therefore, the value of K–M functions of modified catalysts decreased with the ionic radius of rare earth cations (see Fig. 5). As a matter of fact, finely dispersed nickel particles could be formed on impregnation of Ni(II)Y zeolite in aqueous ammonia with a pH between 9 and 12 [24,25]. However, as the introduction of rare earth was prior to that of nickel, some rare earth cations that were not hydrolyzed would interact with nickel as follows [26]:

In order to determine the nature and relative abundance of exposed nickel atoms at catalyst surface, the CO-DRIFTS spectra were recorded (see Fig. 4). Clearly, the absorbance between 2150 and 2000 cm−1 was caused by the formation of Ni(CO)4 , which included three types [22]: Ni(CO)4 in Td symmetry (2053 cm−1 ), ␣-Ni(CO)4 (2060 and 2011 cm−1 ) in C3v symmetry, and ␤-Ni(CO)4 (2135, 2074, 2031, and 1998 cm−1 ) in C2v symmetry. All these types

As mentioned above, the cations with higher ionic radius would be hydrolyzed more easily [7], which led to lower interaction of hydrated cations with nickel particles. As a result, the higher the ionic radius of the cations was, the more the exposed nickel particles remained, and the stronger the metallic function of the modified catalyst was (as shown in Fig. 5).

20

D. Li et al. / Applied Catalysis A: General 241 (2003) 15–24

Fig. 4. CO-DRIFTS profiles of the catalysts with different rare earth elements.

Fig. 5. Influence of the ionic radius of rare earth cations on K–M functions of catalysts.

D. Li et al. / Applied Catalysis A: General 241 (2003) 15–24 Table 2 Relative uptake of CO K–M function and ratio of K–M function to acidity and K–M function of catalysts Catalysts

Relative K–M function (%)

Ni/HY 100 Ni-La/HY 216 Ni-Nd/HY 26 Ni-Sm/HY 14 Ni-Gd/HY 5 Ni-Dy/HY 2

Medium (M) + strong (S) acidity

K–M/(M + S) acidity ratio

0.73 0.85 0.60 0.46 0.47 0.49

1.37 2.54 0.43 0.30 0.11 0.04

3.3. Effect of rare earth on the balance between metallic and acid functions Over bifunctional catalysts, the hydroconversion of alkanes involves (de)hydrogenation on metallic centers, isomerization and cracking on acid centers and diffusion of olefinic intermediates [19–27]. Therefore, the activity, selectivity and stability of catalysts depend on the characteristics of the acidity and the metal state, especially on the balance between metal and acid functions [28,29]. Table 2 gives the relative amount of CO bands, acidity and the ratio of K–M function to acidity for different catalysts, in which the value of CO absorbance on Ni/HY was taken as a reference and the combi-

21

nation of medium and strong acid centers that operate the transformation of carbenium ions [30] was used to represent acidity. It is evident that the addition of rare earth elements led to different effects on the acidity and the metal state of Ni/HY catalyst, and then changed the ratio of metallic to acid functions. Lanthanum-modified Ni/HY catalyst comprised a highly active hydrogenation component compared with the parent catalyst, which led to the rise of the ratio of metallic to acid functions, though the acidity of the catalyst increased to a moderate extent. For Nd-, Sm-, Gd- and Dy-modified Ni/HY catalysts, the decrease of medium and strong acidity was about 35% at most. However, the decrease of active metallic sites was much more than that of acid sites, which led to the decrease of metal/acid ratios. 3.4. n-Octane conversion over rare earth modified catalysts The conversion of n-octane was carried out to determine the relationship between the catalytic performances of catalysts and their surface properties (see Figs. 6–8). Table 3 gives the product distribution for all catalysts at the conversion of ca. 10%. As indicated in Fig. 6, lanthanum modification improved the activity of Ni/HY catalyst, and the improvement became more apparent with the rise of reaction

Fig. 6. Hydroconversion of n-octane: activity of different catalysts.

22

D. Li et al. / Applied Catalysis A: General 241 (2003) 15–24

Fig. 7. Hydroconversion of n-octane: isomerization selectivities.

Table 3 Product distribution for ca. 10% conversion Ni/HY Product distribution Methane Ethane Propane Butane Pentane Hexane Heptane Iso-octane Temp. (K) Conversion (%)

Ni-La/HY

Ni-Nd/HY

Ni-Sm/HY

Ni-Gd/HY

Ni-Dy/HY

– – 2.99 11.97 5.64 0.69 2.07 76.64

– – 1.68 6.72 3.14 0.45 0.90 87.11

– – 7.30 21.80 13.21 1.29 1.93 54.47

– – 4.37 12.93 6.69 0.54 1.43 74.04

– – 10.31 31.12 15.66 0.50 0.50 41.91

– – 7.56 26.16 14.28 – – 52.00

260 8.69

260 8.93

320 9.31

300 11.21

320 10.09

320 9.52

temperature. The other rare earth elements suppressed the activity. For Ni-La/HY catalyst, there were more intermediates generated on metallic sites due to its stronger metallic function, and more acid sites were provided to transform carbenium intermediates (see Table 2). Thus, the activity of Ni-La/HY catalyst increased, and the other rare earth-modified Ni/HY catalysts showed poor activity. The main reactions occurring here were isomerization and cracking. At a given reaction temperature, rare earth-modified catalysts showed a better tendency

to produce isomers (see Fig. 7). At a similar conversion, the isomerization selectivity, influenced by the balance between metallic and acid functions, followed a similar tendency to the activity pattern (see Table 3). The enhanced metal/acid ratio for Ni-La/HY catalyst guaranteed the sufficient (de)hydrogenation activity, which suppressed the probability of readsorbance of the intermediates on an acid site and then the further cracking of the carbenium ions. Thereafter, the cracking of n-octane was suppressed and the isomerization selectivity increased even at high conversions (see

D. Li et al. / Applied Catalysis A: General 241 (2003) 15–24

23

Fig. 8. Hydroconversion of n-octane: cracking product selectivities.

Fig. 7). However, the other rare earth-modified catalysts showed extremely low metallic functions, which slowed down the hydrogenation of the intermediates (see Table 2) and increased the successive transformations of the carbenium ions to some extent. Thus, the probability of cracking increased and the isomerization selectivities for these modified catalysts decreased (see Fig. 7).

The major hydrocracking products for all catalysts were C4 and C3 + C5, and a trace of C1 + C2 and the small amount of C6 + C7 were also observed (see Table 3). With respect to the composition of cracking products, the selectivities for propane, butane and pentane changed concordantly due to the protonated cyclopropane mechanism [31] (see Fig. 8). The changes in hexane and heptane selectivites were similar to

24

D. Li et al. / Applied Catalysis A: General 241 (2003) 15–24

those of K–M functions, since hexane and heptane, as well as methane and ethane, were the products from hydrogenolysis over metal surface [32]. 4. Conclusions Bifunctional Ni/HY catalysts modified by different rare earth elements (La, Nd, Sm, Gd and Dy) showed different performances. The characterization of NH3 -TPD, Py-FTIR and diffuse reflectance infrared Fourier transform spectroscopy of carbon monoxide (CO-DRIFTS) revealed that Bronsted acidity increased along with the rare earth ionic radius, and a similar relationship between metal dispersion and rare earth ionic radius was also observed. Among the additives, only lanthanum showed a favorable effect on metal dispersion, which then improved the ratio of metal to acid function. As a result, Ni-La/HY catalyst displayed a better activity and isomerization selectivity, while others showed poorer catalytic properties. Acknowledgements The authors gratefully acknowledge the funding of this project by National Key Fundamental Research and Development Projects of China (Project ID G1999022402). References [1] J.W. Ward, Fuel Process. Technol. 35 (1993) 55. [2] F. Alvarez, F.R. Ribeiro, G. Perot, C. Thomazeau, M. Guisnet, J. Catal. 162 (1996) 179. [3] P.B. Weisz, Adv. Catal. 13 (1962) 137. [4] T.F. Degnan, C.R. Kennedy, AIChE J. 39 (1993) 607. [5] J. Scherzer, Catal. Rev. Sci. Eng. 31 (1989) 215.

[6] J. Biswas, I.E. Maxwell, Appl. Catal. 63 (1990) 197. [7] E.F. Sousa-Aguiar, V.L.D. Camorim, F.M.Z. Zotin, R.L.C. dos Santos, Micro. Meso. Mater. 25 (1998) 25. [8] S. Sugunan, J. Bindhu, B. Jessey, Indian J. Chem. 36 (1997) 621. [9] A. Corma, J.B. Monton, A.V. Orchilles, Appl. Catal. 23 (1986) 255. [10] F. Lemos, F.R. Ribeiro, M. Kern, G. Gianetto, M. Guisnet, Appl. Catal. 29 (1987) 43. [11] F. Lemos, F.R. Ribeiro, M. Kern, G. Gianetto, M. Guisnet, Appl. Catal. 39 (1988) 227. [12] G. de la Puente, E.F. Sousa-Aguiar, F.M.Z. Zotin, V.L.D. Camorim, U. Sedran, Appl. Catal. A 197 (2000) 41. [13] L. Moscou, M. Lakeman, J. Catal. 16 (1970) 173. [14] T. Yamamoto, T. Tanaka, T. Matsuyama, T. Funabiki, S. Yoshida, J. Phys. Chem. B 105 (2001) 1908. [15] M.D. Romero, A. de Lucas, J.A. Calles, A. Rodriguez, Appl. Catal. 146 (1996) 425. [16] R.J. Taylor, R.H. Petty, Appl. Catal. A 119 (1994) 121. [17] H.H. Kung, Stud. Surf. Sci. Catal. 45 (1989) 73. [18] H. Winkler, K.H. Steinberg, Zeolites 9 (1989) 445. [19] J.G. Nery, Y.P. Mascarenhas, N.C. Mello, T.J. Bonagamaba, E.F. Souza-Aguiar, Zeolites 18 (1997) 44. [20] M. Jiang, S.J. Ye, S.C. Shan, Chin. J. Chem. Phys. 10 (1997) 377. [21] J.G. Kim, T. Kompany, R. Ryoo, T. Ito, J. Fraissard, Zeolites 14 (1994) 427. [22] H. Hobert, R. Marx, I. Weber, W.P. Reschetilowski, K.-P. Wendlandt, Z. Anorg. Allg. Chem. 539 (1986) 211. [23] Y. Guan, Y.N. Qing, L. Zhang, Chin. J. Catal. 18 (1997) 517. [24] M. Suzuki, K. Tsutsumi, H. Takahashi, Y. Saito, Zeolites 8 (1988) 284. [25] M. Sano, T. Maruo, H. Yamatera, M. Suzuki, Y. Saito, J. Am. Chem. Soc. 109 (1987) 5. [26] M. Briend-Faure, J. Jeanjean, G. Spector, D. Delafozze, F. Bozon-Verduraz, J. Chim. Phys. 79 (1982) 489. [27] V. Calemma, S. Peratello, C. Perego, Appl. Catal. A 190 (2000) 207. [28] M.J. Girgis, Y.P. Tsao, Ind. Eng. Chem. Res. 35 (1996) 386. [29] K.-C. Park, S.-K. Ihm, Appl. Catal. A 203 (2000) 201. [30] R.A. van Santen, G.J. Kramer, Chem. Rev. 95 (3) (1995) 637. [31] S. Tiong Sie, Ind. Eng. Chem. Res. 31 (1992) 1881. [32] S.M. Davis, G.A. Somorjai, in: The Chemical Physics of Solid Surfaces and Heterogeneous Catalysts, vol. 6, Springer, Berlin, 1984, p. 162.