mesoporous materials obtained by mordenite recrystallization

Microporous and Mesoporous Materials 164 (2012) 222–231

Contents lists available at SciVerse ScienceDirect

Microporous and Mesoporous Materials journal homepage: www.elsevier.com/locate/micromeso

Hydroisomerization of n-alkanes over Pt-modified micro/mesoporous materials obtained by mordenite recrystallization Stanislav V. Konnov a, Irina I. Ivanova a,b,⇑, Olga A. Ponomareva a,b, Vladimir I. Zaikovskii c a

A.V. Topchiev Institute of Petrochemical Synthesis, Russian Academy of Science, Leninskiy Prospect 29, 119991 Moscow, Russia Department of Chemistry, Moscow State University, Lenin Hills 1/3, 119991 Moscow, Russia c Boreskov Institute of Catalysis, Russian Academy of Sciences, Pr. Akademika Lavrentieva 5, 630090 Novosibirsk, Russia b

a r t i c l e

i n f o

Article history: Available online 23 August 2012 Keywords: Mordenite recrystallization Micro/mesoporous materials Hydroizomerization of n-hexane, n-octane and n-hexadecane

a b s t r a c t Hydroisomerization of n-hexane, n-octane and n-hexadecane has been studied over Pt-containing micro/ mesoporous materials obtained by mordenite (MOR) recrystallization in alkaline solution in the presence of cethyltrimethylammonium bromide. The degree of recrystallization was controlled by varying the concentration of NaOH. Micro/mesoporous nanocomposites MOR/MCM-41 with intermediate degree of recrystallization showed remarkably high activity and selectivity in hydroisomerization of n-octane and high selectivity in hydroisomerization of n-hexadecane. The effect is due to high accessibility of the active sites, improved transport of bulky molecules provided by mesopores and optimal size of Pt particles located in mesopores. Ó 2012 Elsevier Inc. All rights reserved.

1. Introduction Hydroisomerization of linear alkanes nowadays is among the most demanded technologies for paraffins transformation into high octane gasoline. However, while the processes for hydroisomerization of C4 and C5–C6 cuts are well established (PENEX, ISOTEX, TIP, HYSOMER, ISOFIN, SKIP, PAR-ISOM [1,2]), there is still no suitable technology for the conversion of longer-chain alkanes (C7– C8 cuts and higher). The major difficulty in the isomerization of n-alkanes with more than six carbon atoms is their pronounced tendency to cleave. Professor Jens Weitkamp has made a tremendous contribution to the understanding of the influence of an alkane chain length on its hydroisomerization/hydrocracking conversion over bifunctional catalysts [3–11]. He was among the first to show that high selectivity for hydroisomerization of n-alkanes with 7–15 carbon atoms can be achieved over Pt-containing zeolite catalysts, such as Pt/ CaY and Pt/USY [3–8]. Basing on these early studies he has introduced the term of ‘‘ideal’’ bifunctional hydroconversion of n-alkanes, which is characterized by low reaction temperature, high isomerization selectivity and high possibility of pure primary cracking [4]. This ‘‘ideal’’ hydroconversion can be achieved over bifunctional catalysts in which metal phase performs only hydrogenation–dehydrogenation reactions, while isomerization and cracking reactions are catalyzed exclusively by acid sites; the main ⇑ Corresponding author at: Department of Chemistry, Moscow State University, Lenin Hills 1/3, 119991 Moscow, Russia. Tel.: +7 495 939 3570; fax: +7 495 932 8846. E-mail address: [email protected] (I.I. Ivanova). 1387-1811/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.micromeso.2012.08.017

requirement for such catalysts being careful balance between acidic and hydrogenation functions. The detailed investigations of Weitkamp et al. [3–11] on the distribution of the products from long-chain alkane hydroisomerization and hydrocracking over Pt/ zeolites led to the important information on the mechanism of these reactions. Further studies allowed to quantify the ratio of acid and hydrogenation sites required for the ‘‘ideal’’ hydroisomerization catalyst. This ratio was found to be nA/nPt < 10 (i.e. 0.5–1 wt.% Pt per zeolite with Si/Al ratio of 7–45) [12,13]. Besides that it was demonstrated that high metal dispersion is required for high isomerization activity [14,15]. Numerous studies were devoted to the investigation of the influence of zeolite structure, pore geometry and acidity on the hydroisomerization activity [16–22]. Among the large number of the catalysts studied zeolites Y [3–8,16], BEA [17–19], ZSM-22, -23 [20] and zeotypes SAPO-11, -31, -41 [20–22] were found to be the most active and selective in hydroisomerization of n-heptane and n-octane. These results pointed that zeolites modified with nobel metals are among the most perspective catalysts for hydroisomerization n-alkanes with more than six carbon atoms. The main drawbacks related to these catalysts are rather low efficiency and limited selectivity at high conversion levels [23,24]. The kinetic studies performed for hydroisomerization of various alkanes and their mixtures over Pt/MOR catalyst pointed that for C5–C6 alkanes in low pressure range (<4 bar) the rate-limiting step is the isomerization reaction on acid sites, while at higher pressures of n-C6 and for n-alkanes with more than six carbon atoms the reaction is limited by diffusion [25]. The effectiveness factor for transformation of n-hexane over Pt/mordenite calculated

S.V. Konnov et al. / Microporous and Mesoporous Materials 164 (2012) 222–231

basing on Thiele model pointed that only 30–40% of zeolitic pore system is involved in the catalytic reaction [23,24]. Hydroisomerization of n-octane over Pt-Pd containing SAPO-11 and SAPO-41 catalysts was shown to occur at the pore mouths, at a short distance from the external surface of the crystal [21,22]. It has been demonstrated that the diffusion limitations can be reduced or even cancelled by the creation of mesopores in zeolitic crystals [23,24,26–30]. Thus, the creation of mesopores by dealumination procedure was shown to alleviate intracrystalline diffusion limitations for n-hexane in mordenite pore system and therefore resulted in fourfold increase of its hydroisomerisation activity [23,24]. Besides that, selectivity to monobranched isomers was increased due to easier desorption of products. Since 1992, when mesoporous molecular sieves have been discovered, numerous studies have been devoted to the investigation of n-alkane isomerization over mesoporous and composite micro/ mesoporous materials [31–34]. It has been observed that although zeolites Pt/ZSM-5, Pt/ZSM-22, Pt/H-b [34] with strong acid sites show relatively higher activity than mesoporous Pt/Al-MCM-41 catalysts, they give mainly hydrocracking products. At the same time, mesoporous Pt/Al-MCM-41 catalysts show considerably high selectivity to isoalkanes and the highest yield of multibranched isoalkanes due to their weaker acidity and transport properties [34]. The development of composite micro/mesoporous materials [26–30] opens new perspectives for the improvement of hydroisomerization activity of zeolitic catalysts. These materials combine the advantages of both zeolites and mesoporous molecular sieves, in particular, strong acidity, high thermal and hydrothermal stability and improved diffusivity of bulky molecules. It can be expected that the creation of secondary mesoporous structure, on the one hand, will result in the improvement of the catalytic activity in hydroisomerization process and, on the other hand, will lead to the decrease of the residence time of products and minimization of secondary reactions, such as cracking. In this paper we present the results on the effect of mesoposity on hydroisomerization of n-hexane, n-octane and n-hexadecane over Pt-containing micro/mesoporous materials obtained from commercially dealuminated mordenite. The composite micro/mesoporous materials with different degree of mesoporosity were prepared by mordenite recrystallization, the procedure, which has been successfully used for the synthesis of hierarchically structured porous materials from various zeolites [35–40].

2. Experimental Dealuminated mordenite (MOR) with Si/Al = 45 supplied by Zeolyst was used as a starting material. Composite micro/mesoporous materials were obtained by two-step procedure including partial destruction of zeolite in NaOH aqueous solution, followed by its hydrothermal treatment in the presence of cethyltrimethylammonium bromide (CTMABr) at 373 K [32]. The degree of recrystallization was adjusted by variation of NaOH concentration (0.5, 0.8, 1.6 M). The materials obtained were denoted as RM-1, RM-2 and RM-3. Products were recovered by filtration, washed thoroughly with distilled water and dried overnight. The organic template was removed by calcination in a dry air flow at 823 K for 24 h. Afterwards the samples were subjected to threefold ion-exchange with aqueous solution of 0.1 M ammonium nitrate. After ion-exchange the materials were washed, dried overnight and calcined in a dry air flow at 823 K. Platinum was deposited by incipient wetness impregnation with Pt(NH3)4Cl2 aqueous solution. The content of Pt was 0.7 wt.%. The chemical composition of the samples was determined by atomic emission spectroscopy (AES). X-ray diffraction (XRD) data

223

in the region of 5 < 2h < 80 were obtained on a DRON-3M powder diffractometer using CuKa radiation. The diffraction data in the range 0 < 2h < 5 were recorded at the station DICSY in Kurchatov Institute. Transmission electron microscopy images (TEM) were obtained using JEOL JEM 2010 electron microscope, operating at 200 kV. Sorption/desorption isotherms of nitrogen were measured at 77 K using an automated porosimeter (Micromeritics ASAP 2000). Micropore volumes (Vmicro) were determined using t-plot method. The total sorbed volumes (Vtot), including adsorption in the micropores and mesopores and on the external surface, were calculated from the amount of nitrogen adsorbed at relative pressure p/p0 of 0.95, before the onset of interparticle condensation. The acidic properties were studied by temperature programmed desorption of ammonia (TPD-NH3) and FTIR spectroscopy of adsorbed probe molecules. TPD experiments were carried out on chemisorption analyzer USGA-101 (UNISIT) in the temperature range of 293–1053 K in a flow of dry He (30 ml/min). The rate of heating was 8o/min IR spectra were recorded with a Nicolet Protégé 460 FT-IR spectrometer at 2 cm 1 optical resolution. Prior to the measurements, the catalysts were pressed in self-supporting discs and activated in the IR cell attached to a vacuum line at 723 K for 4 h. Adsorption of pyridine (Py) and 2,4,6-treemethylpyridine (TMPy) was performed at 423 K for 30 min. The excess of probe molecules was further evacuated at 423 K for 1 h. The adsorption – evacuation was repeated several times until no changes in the spectra were observed. Hydroisomerization of n-alkanes was performed in a fixed bed flow reactor under 2 MPa and in the temperature range of 430– 600 K. Weight hourly space velocity of n-alkane was 2.0 g/(g h) and the molar ratio of n-alkane: H2 = 1:5. The analysis of the products was performed on GC using 50 m capillary column SE-30.

3. Results and discussion 3.1. Characterization of micro/mesoporous catalysts 3.1.1. Structure and texture The typical procedure for the preparation of micro/mesoporous materials by zeolites recrystallization involves treatment of parent zeolite in NaOH aqueous solution followed by hydrothermal treatment in the presence of CTMABr [35–40]. The alkaline treatment induces partial dissolution and extraction of highly siliceous fragments of the structure from the zeolite resulting in perforation of zeolite crystals by mesopores. The following hydrothermal step in the presence of CTMABr leads to the assembling of these dissolved species into mesoporous phase, which, depending on the degree of zeolite dissolution, either just covers zeolite surface, or forms Zeolite/MCM-41 composites, or completely immerses the residual zeolite. Variation of dissolution level is the most important parameter for tuning the textural characteristics of recrystallized materials. It can be controlled by the variation of the time of alkaline treatment, temperature or alkalinity of the solution, the latter being the most important factor [36]. In the present work, by varying the NaOH concentration in the range of 0.5–1.6 M, a series of recrystallized mordenites with low, intermediate and deep degrees of recrystallization are synthesized. The characteristics of the parent and recrystallized materials are given in Table 1. The comparison of the chemical composition of the parent and recrystallized mordenites (Table 1) shows that the recrystallization does not affect significantly the overall Si/Al ratio, as it is observed in the case of alkaline-mediated mesoporous mordenites [41,42]. This effect is due to the involvement of dissolved siliceous species in the formation of mesoporous phase during recrystallization

224

S.V. Konnov et al. / Microporous and Mesoporous Materials 164 (2012) 222–231

Table 1 Characteristics of catalysts. Sample

MOR RM-1 RM-2 RM-3 a b

Si/Al

45 49 52 56

Pores volume, cm3/g

Phase composition a

Acidityb, lmol/g

Phase

Crystallinity , (%)

Vmic

Vmeso

MOR MOR MOR/MCM-41 MCM-41

100 110 77 1

0.16 0.16 0.08 –

0.03 0.17 0.48 0.74

250 220 160 130

Relative degree of crystallinity calculated from XRD data (normalized intensity of MOR (202) peak). Determined by TPD-NH3.

step. We can thus speculate that in recrystallized mordenites zeolitic fragments are enriched with Al, while mesoporous fragments – with silicon. The X-ray diffraction patterns point that MOR is the only crystalline phase detected in all the materials studied (Fig. 1). However, the intensity of the XRD peaks decreases after the treatment in alkaline solutions, indicating the disappearance of zeolite phase due to the partial dissolution in strong alkaline media (Table 1). Only after recrystallization in mild conditions (sample RM-1), the crystallinity is found to increase slightly. This unusual effect can be attributed to dissolution of some amorphous phase contained in commercially dealuminated mordenite and/or to the recovery of MOR phase due to the healing of some defects in the zeolite framework under recrystallization conditions. TEM images of the sample RM-1 (Fig. 2a) show that the surface of the crystals is covered by a thin film of mesoporous material. Besides that, dissimilar mesopores piercing zeolite crystals can be observed. The formation of mesopores is confirmed by low temperature nitrogen adsorption data (Fig. 3). While the parent zeolite MOR has reversible type-I adsorption/desorption isotherm with a steep rise at p/p0 < 0.01, typical for microporous solids, the recrystallized sample shows the appearance of feebly marked step at p/p0  0.35, corresponding to the existence of dissimilar mesopores. The volume of micropores does not change with respect to the parent zeolite, while the volume of mesopores increases significantly (Table 1). These results point that the recrystallization in mild conditions results in partial desilication of mordenite crystals leading to the formation of intracrystalline mesopores and in assembling of the siliceous fragments removed during desilication on the external surface of the crystal, resulting in formation of thin mesoporous layers. The second type of recrystallized materials (RM-2) contains two phases: zeolitic and mesoporous (Figs. 1 and 4 and Table 1). Small angle range XRD pattern involves a group of lines (1 0 0), (1 1 0) and

Fig. 1. Wide angle range XRD-patterns of parent mordenite and recrystallized materials.

(2 0 0) corresponding to hexagonal pore symmetry of MCM-41 (Fig. 4). Isotherm of N2 adsorption–desorption exhibits rather sharp step at p/p0  0.35, pointing to the formation of uniform mesopores with diameter of 3–4 nm (Fig. 3). The volume of micropores decreases by a factor of two, while the volume of mesopores increases 15 times with respect to the parent zeolite (Table 1). TEM data point to the formation of micro/mesoporous nanocomposites MOR/MCM-41, in which zeolite crystals so-crystallized with small mesoporous fragments are well distinguished (Fig. 2b). Completely recrystallized mordenite (RM-3) involves mainly mesoporous MCM-41 phase (Table 1). No signals in XRD spectrum corresponding to mordenite phase are observed (Fig. 1). On the contrary, the lines (1 0 0), (1 1 0) and (2 0 0) corresponding to hexagonal pore symmetry of MCM-41 become very intensive (Fig. 4). Tiny zeolite fragments can be detected only by TEM (Fig. 2c). 3.1.2. Acidic properties The amount and strength of the acid sites over micro/mesoporous materials was studied by TPD-NH3 (Fig. 5 and Table 1). The results show that the recrystallization of mordenite in mild conditions (RM-1) does not lead to significant changes in the amount and strength of acid sites, which is line with the structural data. On the contrary, the formation of micro/mesoporous nanocomposites MOR/MCM-41 (RM-2) is accompanied by the marked decrease of the amount and strength of acid sites as evidenced by the decrease of the high temperature TPD-NH3 peak and the appearance of the shoulder at lower temperatures corresponding to the acids sited with medium strength. The completely recrystallized mordenite (RM-3) shows rather week acidity. The shape of the TPD-NH3 profile of this sample is typical for MCM-41 materials [35]. For the investigation of the accessibility of Brønsted acid sites 2,4,6-TMPy was selected as a probe molecule. The size of the 2,4,6-TMPy molecule is about 7.4 Å, which is close to the diameter of big channels in MOR (6.7  7 Å). Although, theoretically the molecule of 2,4,6-TMPy can enter into the micropores of MOR, the diffusion inside the crystal should be very limited. The creation of transport mesopores in zeolitic crystal should increase the diffusion of bulky molecule and favor the accessibility of the sites inside the crystal. The concentration of Brönsted acid sites accessible to 2,4,6TMPy was measured by FTIR spectroscopy of adsorbed 2,4,6-TMPy and calculated basing on the published extinction coefficients [43]. Fig. 6 shows the accessibility of Brönsted acid sites estimated as the ratio of the amount of Brönsted acid sites accessible for 2,4,6-TMPy to the total amount of Brönsted sites measured by FTIR spectroscopy of adsorbed Py, assuming that all sites are accessible to Py in dealuminated mordenites [44]. The results point that only 60% of the acid sites of MOR accessible for pyridine can adsorb 2,4,6-TMPy. Recrystallization in mild conditions (RM-1) increases the accessibility up to 80%. Further recrystallization into RM-2 and RM-3 makes all the sites accessible, although the total amount of sites decreases (Table 1).

S.V. Konnov et al. / Microporous and Mesoporous Materials 164 (2012) 222–231

225

Fig. 2. TEM images for recrystallized materials: RM-1 (a), RM-2 (b), and RM-3 (c).

Fig. 3. Nitrogen adsorption/desorption isotherms obtained at 77 K over parent mordenite and recrystallized materials.

3.1.3. Dispersion and localization Pt in parent and recrystallized mordenites Since the balance between the dehydrogenation and acidic functions plays a key role for the synthesis of ‘‘ideal’’ bifunctional hydroisomerization catalyst [3–15], in this study we paid special attention to the dispersion of platinum in the catalysts prepared. The influence of preparation parameters on the dispersion of Pt in zeolite catalysts has been studied in detail by M’Kombe et al. [45]. Among various parameters studied including different loading techniques, calcination conditions (atmosphere, rate of heating, final temperature, duration of calcination) and reduction conditions (atmosphere and temperature), the temperature of sample calcination was shown to be the most important. Therefore, to ob-

Fig. 4. Small angle XRD-patterns of parent mordenite and recrystallized materials.

tain the catalysts with different platinum dispersion, MOR sample was loaded with 0.7 wt.% of Pt by incipient wetness impregnation with Pt(NH3)4Cl2 aqueous solution and oxidized in a flow of air under different temperatures: 670, 770 and 850 K. Afterwards the catalysts obtained were reduced in H2 flow at 670 K. The catalysts prepared were denoted as Pt/MOR-670, Pt/MOR-770 and Pt/MOR850. TEM images of the samples calcined at different temperatures are shown in Fig. 7. Oxidation of Pt precursor under moderate temperature (670 K) leads to the formation of tiny platinum particles

226

S.V. Konnov et al. / Microporous and Mesoporous Materials 164 (2012) 222–231

Pt deposition onto composite micro/mesoporous material (Pt/ RM-2) and completely recrystallized material (Pt/RM-3) leads to the formation of very uniform platinum particles of 3–4 nm, which are most probably localized in the pore system of mesoporous phase (Fig. 8b and c). Consequently, the creation of secondary mesoporous system in zeolite crystals promotes not only the improvement of diffusivity of bulky molecules and high accessibility of the acid sites but also leads to the regulation of Pt particles size.

3.2. Hydroisomerization of n-alkanes over parent Pt/MOR

Fig. 5. TPD-NH3 profiles for parent mordenite and recrystallized materials.

Fig. 6. Relative contents of Bronsted sites accessible for TMB in the series of parent and recrystallized zeolites.

(<1 nm) localized in porous system of the zeolite. The increase of the temperature up to 770 K results in larger particles in the range of 3–8 nm. Finally, calcinaion at 850 K leads to the formation of big Pt particles up to 20 nm and very broad distribution of their size (5–20 nm). The analysis of the literature data shows that tiny Pt particles located in the porous system of zeolites are usually electron-deficient due to the interaction with zeolite acid sites. This results in significant increase of alkanes hydrogenolysis [46,47]. Pt particles with the size larger than 3 nm lead to positive effect in hydroisomeriztion selectivity [47]. However, very large Pt particles are not favorable due to the disbalance of acidic and hydrogenation functions in bifunctional catalyst and therefore high contribution of cracking [12–15]. Therefore, for the modification of micro/mesoporous materials with Pt, the procedure involving calcination at 770 K was used. Fig. 8 displays TEM images of Pt-modified recrystallized mordenites. Pt/RM-1 shows dissimilar Pt particle distribution in the range of 1–8 nm as it was observed on parent mordenite sample (Fig. 7b). It can be seen that this sample contains tiny Pt particles (<1 nm) inside the micropores, medium-size Pt particles of 3–4 nm located in mesoporous layers covering zeolite crystals and big particles (up to 8 nm) located on the external surface of zeolite crystals.

3.2.1. Effect of the alkane chain length Three model compounds, n-hexane, n-octane and n-hexadecane, were selected for the investigation of the effect of the alkane chain length on the activity and selectivity Pt/MOR catalysts in hydromerization. For n-hexane feed Pt-modified dealuminated mordenite is considered to be among the best catalysts and an important number of the industrial units for hydroisomerization of C5–C6 cuts are based on this catalyst [1]. However, this catalyst could hardly be applied for the hydroconversion of n-alkanes with longer chain length due to significant diffusion limitations in the pore system of the mordenite [25]. To assess the influence of chain length on the activity and selectivity of Pt/MOR in hydroisomerization, the conversion of C6, C8 and C16 alkanes is plotted versus reaction temperature and the hydroisomerization selectivity – versus conversion (Fig. 9). The following general features of hydroisomerization can be recognized: – with increasing chain length the curves of conversion shift towards lower reaction temperature reflecting the increasing reactivities; – hydroisomerization selectivity decreases drastically with increasing chain length, indicating the increasing tendency for cleavage; – the main products of n-hexane hydroisomerization are dimethylbutanes (Table 2), while in the case of longer alkanes monobranched isomers are predominant (Table 3) due to easier cracking of polybranched isomers with longer chain. These observations are in line with those reported previously by Weitkamp et al. for Pt-containing large-pore zeolites [4,8]. Further insight into the chemistry of hydroconversion of n-alkanes over Pt/MOR can be gained from the examination of hydrocracking products (Tables 2–4). The nature of cracked products observed is indicative of carbenium ion mechanism of cracking. Methane and ethane are absent in the case of n-octane and n-hexadecane, which rules out their hydrogenolysis. Small amounts of ethane are detected only for n-hexane feed, pointing to some contribution of hydrogenolytic cracking in this case.nHexane and n-octane show predominant cleavage of the central C–C bond (Tables 2 and 3), indicating pure primary cracking as it

Fig. 7. TEM images for Pt/MOR samples oxidized at 670 K (a), 770 K (b), 850 K (c) and reduced at 670 K.

227

S.V. Konnov et al. / Microporous and Mesoporous Materials 164 (2012) 222–231

Fig. 8. TEM images for recrystallized materials: Pt/RM-1 (a), Pt/RM-2 (b), and Pt/RM-3 (c).

Fig. 9. Conversion of n-alkanes as a function of temperature (a) and selectivity to isomerization products as a function of conversion (b) over Pt/MOR (2 MPa, H2/ n-alkane = 5:1).

Table 3 n-Octane hydroconversion over parent and recrystallized mordenites.

Table 2 n-Hexane hydroconversion over parent and recrystallized mordenites. Catalyst

Pt/MOR

Pt/RM-1

Pt/RM-2

Catalyst

Pt/MOR

Pt/RM-1

Pt/RM-2

540 2 2

530 2 2

520 2 2

Temperature, K Pressure, MPa WHSV, h 1

580 2 2

580 2 2

600 2 2

Temperature, K Pressure, MPa WHSV, h 1

Conversion, % Yield of i-C6, wt.% Selectivity to i-C6, wt.%

79 74.7 94.5

80 77.1 96.4

79 66.1 83.7

Conversion, % Yield of i-C8, wt.% Selectivity to i-C8, wt.%

63 36.7 58.3

67 36.0 53.7

57 47.3 83.0

Product distribution, wt.% Methane Ethane Propane i-Butane n-Butane i-Pentanes n-Pentane Dimethylbutanes Methylpentanes Heptanes

– 0.3 2.7 0.7 0.4 0.6 0.3 64.9 29.6 0.5

– 0.2 1.0 0.7 0.5 0.5 0.2 68.1 28.3 0.5

– 0.8 5.7 3.3 1.9 2.7 1.5 57.9 25.8 0.4

Product distribution, wt.% Methane Ethane Propane n-Butane i-Butane n-Pentane i-Pentanes n-Hexane i-Hexanes Methylheptanes Dimethylheptanes Trimethylpentanes

– – 6.0 18.3 5.7 8.5 1.8 0.4 1.1 39.0 13.1 6.2

– – 5.6 21.4 6.0 9.2 1.4 0.0 1.6 37.5 10.5 5.7

– 0.1 2.4 7.1 2.2 2.9 1.1 0.0 0.8 56.2 18.5 8.3

was observed for large-pore zeolites [4,10,11]. However the degree of branching of cracked products is totally different in the case of n-octane pointing to the different mechanism of b-scission operating on mordenites. On large pore zeolites, cracking products are composed mostly of branched isomers, pointing to type A and type B modes of b-scission (Fig. 10) [4,10]. On the contrary, on mordenites n-alkanes give the main contribution to the products of cracking (Table 3), which is consistent with type C b-scission and points to the predominant b-scission of monobranched i-octanes. Similar observations were made over Pt/HMFI catalyst [9], which was accounted by shape selective effects in the porous system of MFI. The large channels of MOR are most probably sufficiently big for the isomerization of n-octane into monobranched isomers. However, mordenite does not have large cavities as in the case of FAU, in which further isomerization into di- and tribranched isomers and type A and type B modes of b-scissions can occur. There-

fore isomerization of n-octane into di- and tribranched isomers takes place on the external surface of mordenite or in its pore mouths, which leads to their faster desorption without cracking, while the main contribution of cracking products results from monobranched isooctanes. In the case of hexadecane, secondary cracking gives rather high contribution already at low conversions (Table 4). Most probably, the isomerization and primary cracking of bulky hexadecane takes place on the external surface and in the pore mouths of mordenite as it was suggested previously for n-octane hydroisomerization over SAPO-11 and SAPO-41 [21,22]. The products of primary cracking undergo further transformations in the porous system of mordenite. Thus, large pore one dimensional MOR shows non shape selective hydroisomerisation as in the case of large pore FAU, but shape selective hydrocracking as in the case of medium pore zeolites.

228

S.V. Konnov et al. / Microporous and Mesoporous Materials 164 (2012) 222–231

Table 4 n-Hexadecane hydroconversion over parent and recrystallized mordenites. Catalyst

Pt/MOR

Pt/RM-1

Pt/RM-2

Temperature, K Pressure, MPa WHSV, h 1

470 2 2

470 2 2

490 2 2

Conversion, % Yield of i-C16, wt.% Selectivity to i-C16, wt.%

32 6.7 20.9

23 5.5 23.9

24 8.1 33.7

Product distribution, wt.% Methane Ethane Propane i-Butane n-Butane i-Pentanes n-Pentane i-Hexanes n-Hexane i-Heptanes n-Heptane i-Octanes n-Octane i-Nonanes n-Nonane i-Decanes n-Decane i-Undecanes n-Undecane i-Dodecanes n-Dodecane i-Tridecanes i-Tridecane i-Hexadecanes

– – 2.3 18.5 5.6 17.3 2.6 11.2 2.0 6.0 1.3 4.1 0.8 2.2 0.4 1.8 0.3 1.7 0.3 0.6 0.2 0.0 0.1 20.9

– – 2.0 18.7 6.0 17.9 3.5 7.7 1.9 5.6 1.1 4.5 0.7 2.1 0.4 1.8 0.3 1.1 0.3 0.2 0.3 0.0 0.2 23.9

– – 2.7 16.8 4.0 12.6 3.2 8.4 2.3 4.1 0.8 3.6 0.6 2.2 0.4 1.7 0.3 1.2 0.3 0.7 0.3 0.0 0.1 33.7

Fig. 11. The influence of platinum oxidation temperature on the conversion of noctane and selectivity to iso-C8 (2 MPa, H2/n-alkane = 5:1).

– The conversion of n-octane only slightly increases with Pt dispersion. – On the contrary, significant differences are observed for the hydroisomerization selectivity. The highest selectivity is achieved over the sample Pt/MOR-770 with Pt particle size of 3–8 nm. The decrease of the size of Pt particles (Pt/MOR-670) leads to higher contribution of the products of hydrogenolysis (methane and ethane), while the increase of the size of particles (Pt/MOR-850), enhances contribution of cracking products (propane, butanes and pentanes). Thus the results show that selective hydroisomerization of long-chain n-alkanes is promoted by medium sized Pt particles in the range of 3–8 nm. These results are in line with the data published in the literature [47,48]. According to Namzek and Ryczkowski [47], the decrease of Pt particle size below 3 nm results in higher contribution of the products of hydrogenolysis in n-butane hydroconversion. The increase of platinum particle size above 3 nm leads to predominant isomerization. More detailed kinetic study performed by Brunelle et al. [48] for hydroisomerization of n-pentane shows that the ratio VH/VI, where VH is a rate of hydrogenolysis and VI is a rate of isomerization, decreases by a factor of 10 when the platinum particle size increases from 0.7 to 15 nm [48]. 3.3. Hydroisomerization of n-alkanes over Pt-containing micro/ mesoporous catalysts

Fig. 10. Types of b-scission in various alkylcarbenium ions.

3.2.2. The influence of platinum dispersion Although many reports have been published on the effect of Pt content [12,13] and Pt dispersion [14,15,46,47] on the hydroconversion of alkanes with less than 7 carbon atoms, there is a lack of information on the effect of Pt dispersion on the hydroconversion of higher n-alkane homologs. Therefore, we have concentrated our efforts on the investigation of the influence of Pt dispersion on the n-octane hydroconversion. This study was performed over a series of Pt/MOR samples characterized by different size of Pt particles (Fig. 7). The results on the activity and selectivity of these catalysts in hydroisomerization of n-octane are shown in Fig. 11. The following differences emerge:

The investigation of hydroconversion of n-hexane, n-octane and n-hexadecane over Pt-modified micro/mesoporous materials points that Pt/RM-1 and Pt/RM-2 catalysts have high catalytic activity and selectivity in hydroisomerization (Tables 2–4). On the contrary, Pt/RM-3 sample is completely inactive in hydroisomerization of n-alkanes in the temperature range studied (for nhexane below 610 K, for n-octane below 540 K and for n-hexadecane below 530 K), which is most probably due to very low acidity of this catalyst (Fig. 5). The comparison of recrystallized materials with parent mordenite shows significant differences, which are strongly influenced by the chain length of the feed. 3.3.1. Conversion of n-hexane The results on n-hexane hydroconversion are shown in Fig. 12 and Table 2. Recrystallization of dealuminated MOR into RM-1 does not affect neither n-hexane conversion, nor hydroisomerization selectivity, although the accessibility of the acid sites is improved significantly (Fig. 6). Furthermore, recrystallization into RM-2 leads even to the decrease of n-hexane conversion, which

S.V. Konnov et al. / Microporous and Mesoporous Materials 164 (2012) 222–231

229

Fig. 12. Conversion of n-hexane as a function of temperature (a) and selectivity to iso-hexanes as a function of conversion (b) over parent and recrystallized mordenites (2 MPa, H2/n-alkane = 5:1).

is most probably due to the decrease of the amount and strength of the acid sites (Fig. 5). These observations are in line with the earlier findings [23,24], pointing that hydroconversion of n-hexane over dealuminated mordenites is not limited by the diffusion and depends on the amount and strength of the acid sites. 3.3.2. Conversion of n-octane Different picture is observed for n-octane hydroconversion (Fig. 13 and Table 3). The curves of conversion as a function of temperature are shifted significantly to lower temperature, both in the case of Pt/RM-1 and Pt/RM-2 catalysts, indicating higher activity of these catalysts with respect to Pt/MOR. Besides that, the considerable improvement of hydroisomerization selectivity is observed over Pt/RM-2 catalyst (Fig. 13b). The enhancement of the activity is most probably due to the alleviation of the intracrystalline diffusion limitations for n-octane in the porous system of mordenite and to the improvement of the accessibility of its acid sites. The enhanced selectivity could be due two reasons. On the one hand, the better transport of bulky molecules in the pore system of Pt/RM-2 should result in the decrease of residence time of products and therefore minimization of secondary reactions, such as cracking. On the other hand, mesoporous structure of Pt/RM-2 favors narrow distribution of Pt particles in the range of 3–4 nm required for selective hydroisomerization (Fig. 11). The selectivity to mono- and polybranched isomerization products is shown in Fig. 14 for Pt/MOR and Pt/RM-2 catalysts. At low conversion levels monobranched isomers are exclusively formed. With increasing overall conversion the content of polybranched isomers first increases and then passes through a maximum. These results show that polybranched isomers are formed via consecutive steps from monobranched isomers and then are consumed in cracking reactions. Similar regularities are observed over Pt/ MOR and Pt/RM-2 catalysts, the main difference is in higher selectivities to both mono- and polybranched isomers over the latter. The analysis of the isomer distribution in monobranched i-octanes points that this fraction is mostly composed of methylhep-

Fig. 13. Conversion of n-octane as a function of temperature (a) and selectivity to iso-octanes as a function of conversion (b) over parent and recrystallized mordenites (2 MPa, H2/n-alkane = 5:1).

Fig. 14. Selectivity to monobranched and polybranched octanes as a function of noctane conversion over Pt/MOR and Pt/RM-2 (2 MPa, H2/n-alkane = 5:1).

tanes (MH), only traces of ethylhexanes are detected. The distribution of methyl isomers observed at high degrees of conversion (Fig. 15) roughly corresponds to those anticipated for thermodynamic equilibrium [4]. However, at low conversion levels it is kinetically controlled. The decrease of the content of 2-methylheptane at low conversions is consistent with the branching mechanism via protonated cyclopropanes as it was proposed previously by Weitkamp et al. [8] for non-selective hydroisomerization of long chain n-alkanes over Pt/CaY catalyst. Similar results observed over Pt/MOR and Pt/RM-2 catalysts (Fig. 15a and b) point to the same mechanisms of branching operating over these catalysts. Polybranched isomers are mostly composed of dimethylhexanes (DMH) (Table 3), among which isomers with one branching in 2 positions, such as 2,4-dimethylhexane and 2,5-dimethylhexane, are predominant. Trimethylpentanes (TMP) are formed in lesser amounts due to the constrains imposed in the pore system of mordenite. The comparison of the distribution of the products of n-octane hydrocracking over parent and recrystallized materials does not show any significant difference (Table 3), pointing to the same modes of b-scission over these catalysts. 3.3.3. Conversion of n-hexadecane The results of hydroconversion of n-hexadecane are presented in Fig. 16. As in the case of n-octane, the significant improvement of hydroisomerization selectivity is observed over Pt/RM-2 catalyst due to the improved transport of bulky molecules in the pore system of this catalyst and optimized size of Pt particles. On the contrary, the activity of Pt/RM-2 catalyst is lower with respect of parent mordenite (Fig. 16a). At present, we do not have straightforward explanation for this observation. Several reasons may account for it:

Fig. 15. Distribution of MH isomers as a function of n-octane conversion over Pt/ MOR and Pt/RM-2 (2 MPa, H2/n-alkane = 5:1).

230

S.V. Konnov et al. / Microporous and Mesoporous Materials 164 (2012) 222–231

Fig. 16. Conversion of n-hexadecane as a function of temperature (a) and selectivity to iso-hexadecanes as a function of conversion (b) over parent and recrystallized mordenites (2 MPa, H2/n-alkane = 5:1).

– the mesoporous phase, which covers zeolytic crystals (Fig. 2), may prevent bulky molecules of n-hexadecane to access zeolytic acid sites on the external surface of the crystals and in the pore mouths of mordenite; – the total amount of acid sites available on the external surface and in the pore mouths may decrease due to significant degree of recrystallization into mesoporous phase; – the desorption of products may become the limiting step of the reaction; – the competitive adsorption on the acid sites may play more important role on RM-2 materials. Although the information on the distribution of hexadecane isomers is not available due to analytical problems, the distribution of cracking products can be assessed in Table 4. The results show that mordenite recrystallization does not affect the distribution of the products of cracking and therefore it does not influence the mechanism of n-hexadecane cleavage. 3.3.4. Effect of mesoporosity on the hydroisomerization of n-alkanes with different chain length Fig. 17 depicts the comparison of the yields of isoalkanes obtained by hydroisomerization of n-alkanes with different chain length over micro/mesoporous and parent zeolites. In the case of n-hexane feed, the maximum yield of isohexanes is achieved over Pt/MOR and Pt/RM-1 catalyst. The yield of isohexanes is lower over Pt/RM-2 due to the decrease in the strength and the amount of acid sites. At the same time, the creation of secondary mesoporous structure allows to enhance the yields of isooctanes (Fig. 17b). The yield of isooctane increases almost by 20% over Pt/RM-2 with respect to Pt/MOR. In the case of n-hexadecane hydroisomerization, the yield of isohexadecanes only slightly increases due to selectivity enhancement (Fig. 17b). Thus, all the above observations suggest that the creation of the secondary mesoporosity by recrystallization has different effect on hydroisomerization of n-alkanes with different chain length:

Fig. 18. Comparison of maximal yields of iso-C8 (black) and DMH + TMP (white) over Pt/RM-2 with literature data for different catalysts [20,22,44–46].

– In the case of n-hexane and lower homologies, secondary mesoporous structure creation is not efficient since n-hexane hydroisomerization proceeds without obvious diffusion limitations in the porous system of dealuminated mordenite [25]. – In the case of n-octane and medium chain length hydrocarbons, the catalytic performance can be improved significantly due to the improved transport of bulky molecules, better accessibility of the acid sites and optimized Pt particle size. – For higher homologies such as n-hexadecane, which cannot diffuse in the porous system of Pt/MOR and are converted mainly on the external surface of the crystal or in zeolite pores mouths, selectivity can be improved mainly due to the controlled Pt particles size. 3.3.5. Comparison of the catalytic performance of Pt-modified micro/ mesoporous materials with the other Pt-containing catalysts Fig. 18 depicts the comparison of the maximal yields of i-C8 and TMP + DMH obtained in hydroisomerization of n-octane over Pt/ RM-2 and different catalysts presented in literature [20,22,49– 51]. Pt/WO42 /ZrO2 and Pt/SO42 /ZrO2 catalysts with strong acid sites show high activity in n-octane hydroconversion, however the main reaction pathway over these catalysts is hydrocracking. The yield of i-C8 does not exceed 24% [49]. Different zeolites and zeotypes show high hydroisomerization activity and selectivity to i-C8. However at any level of n-octane conversion the formation of MH isomers largely predominates. The highest yield of polybranched isooctanes is observed for Pt/ RM-2 catalyst.

Fig. 17. Yields of iso-C6, iso-C8 and iso-C16 as a function of reaction temperature (2 MPa, H2/n-alkane = 5:1).

S.V. Konnov et al. / Microporous and Mesoporous Materials 164 (2012) 222–231

4. Conclusions The results highlight that recrystallization in mild conditions leads to the creation of mesopores in zeolite crystals and their coating with thin mesoporous layers. The increase of the degree of recrystallization results in the formation of micro/mesoporous nanocomposites MOR/MCM-41, in which zeolite crystals are socrystallized with small mesoporous fragments. Deeply recrystallized mordenite involves mainly mesoporous MCM-41 phase, which immerses tiny zeolitic fragments. The creation of secondary mesoporous system promotes the formation of Pt particles with narrow distribution of size in the range of 3–4 nm, which favors selective hydroisomerization. Hydroisomerization of n-alkanes with various chain length over micro/mesoporous mordenites points to the following differences: – no effect is observed in the case of n-hexane hydroconversion proceeding without obvious diffusion limitation in the mordenite porous system; – in the case of n-octane hydroisomerization micro/mesoporous nanocomposites Pt/MOR/MCM-41 show remarkably high activity and selectivity due to high accessibility of the active sites, improved transport of bulky molecules provided by mesopores and optimal size of Pt particles located in mesopores; the yield of isooctanes increases almost by 20% with respect to parent mordenite; – in hydroisomerization of n-hexadecane micro/mesoporous nanocomposites Pt/MOR/MCM-41 show enhanced selectivity to isohexadecanes.

Acknowledgements The authors thank Federal Principal Program of Russian Ministry of Education and Science ‘‘Research and educational specialists of innovative Russia’’ for the financial support. The work was also supported by RFBR No. 10-03-00829-a. References [1] C. Martinez, A. Corma, Coord. Chem. Rev. 255 (2011) 1558–1580. [2] B. Ducoutry, G. Szabo, J.P. Dath, J.P. Gilson, J.M. Goupil, D. Cornet, Appl. Catal. A 269 (1/2) (2004) 203–214. [3] H. Schulz, J. Weitkamp, ACS Prepr. Div. Petr. Chem. 17 (4) (1972) 84. [4] J. Weitkamp, The Influence of Chain Length in Hydrocracking and Hydroisomerization of n-alkanes, in: J.W. Ward, S.A. Qader (Eds.), ACS Sympoeium Series 20, American Chemical Society, Washington, DC, 1975, pp. 1–27. [5] M. Steijns, G. Froment, P. Jacobs, J. Uytterhoeven, J. Weitkamp, Erdoel Kohle Erdgas Petrochem. Brennst. Chem. 31 (1978) 581. [6] M. Steijns, G. Froment, P. Jacobs, J. Uytterhoeven, J. Weitkamp, Ind. Eng. Chem. Prod. Res. Dev. 20 (1981) 654–660. [7] M. Steijns, G. Froment, Ind. Eng. Chem. Prod. Res. Dev. 20 (1981) 660–668. [8] J. Weitkamp, Ind. Eng. Chem. Prod. Res. Dev. 21 (1982) 550–558. [9] J. Weitkamp, P.A. Jacobs, J.A. Martens, Appl. Catal. 8 (1983) 123–141.

231

[10] J.A. Martens, P.A. Jacobs, J. Weitkamp, Appl. Catal. 20 (1986) 239–281. [11] J.A. Martens, P.A. Jacobs, J. Weitkamp, Appl. Catal. 20 (1986) 283–303. [12] G.E. Giannetto, G.R. Perot, M.R. Guisnet, Ind. Eng. Chem. Prod. Res. Dev. 25 (3) (1986) 481–490. [13] M. Guisnet, F. Alvarez, G. Giannetto, G. Perot, Catal. Today 1 (1987) 415–433. [14] M.E. Grillo, M.M. Ramirez de Agudelo, J. Mol. Catal. A: Chem. 119 (1997) 105– 112. [15] P. Liu, J. Wang, R. Wei, X. Ren, X. Zhang, Catal. Lett. 126 (2008) 346–352. [16] D. Karthikeyan, N. Lingappan, B. Sivasankar, Ind. Eng. Chem. Res. 47 (2008) 6538–6546. [17] A. Lugstein, A. Jentys, H. Vinek, App. Catal. A 176 (1999) 119–128. [18] C. Jimenez, F.J. Romero, R. Roldan, J.M. Marinas, J.P. Gomez, App. Catal. A 249 (2003) 175–185. [19] A. de Lucas, M.J. Ramos, F. Dorado, P. Sanchez, J.L. Valverde, Appl. Catal. A 289 (2005) 205–213. [20] H. Yunfeng, W. Xiangsheng, G. Xinwen, L. Silue, H. Sheng, S. Haibo, B. Liang, Catal. Lett. 100 (2005) 59–65. [21] P. Meriaudeau, V.A. Tuan, F. Lefebvre, T.V. Nghiem, C. Naccache, Micropor. Mesopor. Mater. 22 (1998) 435–449. [22] P. Mériaudeau, V.A. Tuan, F. Lefebvre, T.V. Nghiem, C. Naccache, Micropor. Mesopor. Mater. 26 (1998) 161–173. [23] M. Tromp, J.A. van Bokhoven, M.T. Garriga Oostenbrink, J.H. Bitter, K.P. de Jong, D.C. Koningsberger, J. Catal. 190 (2000) 209–214. [24] S. Van Donk, A. Broersma, O.L.J. Gijzeman, J.A. van Bokhoven, J. Catal. 204 (2001) 272–280. [25] A. Hollo, J. Hancsok, D. Kallo, Appl. Catal. A 229 (2002) 93–102. [26] J. Cejka, S. Mintova, Catal. Rev. Sci. Eng. 49 (2007) 457–509. [27] J. Pérez-Pariente, Y. Diaz, J. Agúndez, C. R. Chim. 8 (2005) 569–578. [28] Y. Tao, H. Kanoh, L. Abrams, K. Kaneko, Chem. Rev. 106 (2006) 896–910. [29] K. Egeblad, Cr.H. Christensen, M. Kustova, Cl.H. Christensen, Chem. Mater. 20 (3) (2008) 946–960. [30] M. Ogura, Catal. Surv. Asia 12 (2008) 16–27. [31] S.P. Elangovan, M. Hartmann, J. Catal. 217 (2003) 388–395. [32] K. Chaudhari, T.K. Das, A.J. Chandwadkar, S. Sivasanker, J. Catal. 186 (1999) 81– 90. [33] S.P. Elangovan, C. Bischof, M. Hartmann, Catal. Lett. 80 (2002) 35–40. [34] K.-C. Park, A.-K. Ihm, Appl. Catal. A 203 (2000) 201–209. [35] Yu.P. Khiteva, Yu.G. Kolyagina, b, I.I. Ivanovaa, b⁄, O.A. Ponomarevaa, b, F. Thibault-Starzykc, J.-P. Gilsonc, C. Fernandezc, F. Fajulad, Micropor. Mesopor. Mater. 146 (2011) 201–207. [36] I.I. Ivanova, E.E. Knyazeva, Patent RF, 2282587, 2006. [37] I.I. Ivanova, A.S. Kuznetsov, V.V. Yuschenko, E.E. Knyazeva, Pure Appl. Chem. 76 (2004) 1647–1658. [38] V.V. Ordomsky, V.Y. Murzin, Yu.V. Monakhova, Y.V. Zubavichus, E.E. Knyazeva, N.S. Nesterenko, I.I. Ivanova, Micropor. Mesopor. Mater. 105 (2007) 101–110. [39] Yu.P. Khitev, Yu.G. Kolyagin, I.I. Ivanova, O.A. Ponomareva, F. Thibault-Starzyk, J.-P. Gilson, C. Fernandez, F. Fajula, Micropor. Mesopor. Matter. 146 (2011) 201–207. [40] I.I. Ivanova, A.S. Kuznetsov, E.E. Knyazeva, F. Fajula, F. Thibault-Starzyk, C. Fernandez, J.-P. Gilson, Catal. Today 168 (2011) 133–139. [41] J.C. Groen, T. Sano, J.A. Moulijn, J. Perez-Ramırez, J. Catal. 251 (2007) 21–27. [42] V. Paixao, A.P. Carvalho, J. Rocha, A. Fernandes, A. Martins, Micropor. Mesopor. Mater. 131 (2010) 350–357. [43] F. Thibault-Starzyk, I. Stan, S. Abello, A. Bonilla, K. Thomas, C. Fernandez, J.-P. Gilson, J. Pérez-Ramírez, J. Catal. 264 (2009) 11–14. [44] N.S. Nesterenko, F. Thibault-Starzyk, V. Montouillout, V.V. Yuschenko, C. Fernandez, J.-P. Gilson, F. Fajula, I.I. Ivanova, Micropor. Mesopor. Mater. 71 (2004) 157–166. [45] C.M. M’Kombe, M.E. Dry, C.T. O’Connor, Zeolites 19 (1997) 175–179. [46] A.Yu. Stakheev, L.M. Kustov, App. Catal. A 188 (1999) 3–35. [47] D. Nazimek, J. Ryczkowski, React. Kinet. Catal. Lett. 40 (1) (1989) 137–143. [48] J.-P. Brunelle, A. Sugier, J.-F. le Page, J. Catal. 43 (1976) 273–291. [49] J.M. Grau, J.C. Yori, J.M. Parera, Appl. Catal. A 213 (2001) 247–257. [50] P. Sanchez et al., Appl. Catal. A 314 (2006) 248–255. [51] A. de Lucas, J.L. Valverde, P. Sanchez, F. Dorado, M.J. Ramos, Appl. Catal. A 282 (2005) 15–24.