Applied Catalysis A: General 253 (2003) 469–486
Activity, selectivity and stability of Ni–Pt loaded zeolite- and mordenite catalysts for hydroisomerisation of n-heptane I. Eswaramoorthi∗ , A. Geetha Bhavani, N. Lingappan Department of Chemistry, Anna University, Chennai-600025, India Received 17 March 2003; received in revised form 8 June 2003; accepted 2 July 2003
Abstract Zeolites  and mordenite were impregnated with 0.1 wt.% Pt and varying (0, 0.1, 0.3 and 0.5 wt.%) amount of Ni and reduced at 475 ◦ C. The line broadening XRD analysis indicates that the increasing Ni addition decreases the crystallinity of zeolites. The TEM analysis indicates that the average metal particle size increases with increasing Ni addition. The states of Ni and Pt were analysed by ESCA. The acidity measurements by both TPD of ammonia and pyridine adsorbed FT-IR spectroscopy show the occupation of some of the acid sites by the added Ni species. Hydroisomerisation of n-heptane was carried out in the temperature range 225–375 ◦ C and found that Ni addition up to 0.3 wt.% over 0.1 wt.% Pt/H- and 0.1 wt.% over 0.1 wt.% Pt/H-MOR enhance the n-heptane conversion, isomerisation selectivity and sustainability of the catalysts. Also, the multibranched (MTB) isomers selectivity and protonated cyclopropane (PCP) intermediate mechanism are favoured over the above catalysts. Further Ni addition above the threshold value leads to fall in activity and isomerisation selectivity drastically. Zeolite- based Ni-Pt catalysts always show higher activity and selectivity than mordenite based catalysts. © 2003 Elsevier B.V. All rights reserved. Keywords: Hydroisomerisation; n-heptane; Nickel; Platinum; Zeolite-; Mordenite
1. Introduction The increasing interest to promote the efficiency of the automotive motors encourages the development of new catalytic process for high-octane gasoline production. Hydroisomerisation and hydrocracking are considered as an efficient method, in which bifunctional catalysts combining hydrogenation–dehydrogenation and Brønsted acid functions are widely used [1,2]. The acid and metal site density and strength distribution are both important and their proper balance is critical in determining the activity of these catalysts [3–5]. It ∗ Corresponding author. Tel.: +91-44-22351126; fax: +91-44-22200660. E-mail address:
[email protected] (I. Eswaramoorthi).
was found that the introduction of a second metal influences the property of first dispersed metal due to the formation of metallic clusters [6]. Vazquez et al. [7] studied the n-heptane hydroisomerisation over a series of Ni-Mo/H-Y catalysts at 300–350 ◦ C and observed maximum activity over catalyst with Ni/(Ni + Mo) ≈ 0.5. They correlated the Ni/(Ni + Mo) ratio to the rate controlling step and found that when the ratio is <0.5, the dehydrogenation of n-heptane on the metal site is the controlling step and for catalysts with Ni/(Ni + Mo) ratio >0.5, the rearrangement of carbenium ion on the acid sites is the rate controlling step. Blomsma et al. [8] observed that the suppression of bimolecular mechanism (dimerisation–cracking) by the addition of 20 mol% of Pd to Pt/H- zeolite in n-heptane isomerisation. Further they reported that the addition of Pd to
0926-860X/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0926-860X(03)00538-6
470
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486
Pt/H- increases the metal dispersion as well as the stability of the catalysts. Le Van Mao and Saberi [9] introduced Al, Zn and Cd species on Pt/H-Y zeolite and applied for n-heptane isomerisation. They found that incorporation of the above species in small amounts results in enhanced yield of branched paraffins and decreased the formation of cracked products. Jao et al. [10] introduced Ni as second metal in Pt/MOR catalytic system and found that moderate amount (0.5 wt.%) of Ni addition to 0.26 wt.% Pt/MOR enhances the activity in n-heptane isomerisation and selectivity to multibranched (MTB) isomers with good suppression of fuel gas formation. Further addition of Ni led to decrease in activity as well as without suppression of fuel gas formation. Jordao et al. [11] observed that bimetallic Ni-Pt/HUSY catalysts containing 20–30% Pt show better activity and selectivity for high octane dibranched hexanes than Pt only containing catalysts. The higher activity of Ni-Pt catalysts were accounted by (i) the presence of Pt enhancing the reduction of Ni cations forming more metal particles, (ii) the Ni particles serving as support for the Pt atoms and (iii) the generation of higher superficial energy due to difficulty in accommodation of platinum atoms in a nickel metal particle. In our previous reports, the addition of Ni up to a threshold value over Pt loaded H-, H-MOR [12], H-Y [13], SAPO-5 and SAPO-11 [14] enhances the n-hexane conversion, isomerisation selectivity and sustainability of the catalysts. Further, the threshold Ni addition is found to enhance the high octane DMBs selectivity and protonated cyclopropane (PCP) intermediate mechanism. The enhanced activity was accounted in terms of better metal–acid balance between catalytically active bimetallic (Ni–Pt) nanoparticles formed and acid sites of the support. Catalysts with higher Ni content show poor activity due to pore blockage by larger bimetallic particles formed and incomplete reduction of nickel species. Dominguez et al. [15] reported alloy formation between Ni and Pt with a common fcc lattice structure suggesting some potential bimetallic interactions between these two metals and hence allowing for the catalytic active in hydrocarbon reforming reactions. Few reports are available in literature about Ni addition as the second metal to Pt, but the concentration of Ni loading kept at fairly high range (>0.5 wt.%). Hence, in the present study, nickel is introduced
as the second metal in low concentration range (<0.5 wt.%) to modify the catalytic properties of Pt supported zeolites  and mordenite (with different acidic properties as well as pore size arrangements) on n-heptane hydroisomerisation. Further the present results are correlated to rationalise the variations in the product distributions in n-heptane hydroisomerisation in terms of pore structure of acidic supports, amount of metal loading, reaction temperature and acidity.
2. Experimental 2.1. Catalyst preparation The sodium form of zeolite- (SiO2 /Al2 O3 = 10) supplied by United Catalyst India Ltd., India. Mordenite (SiO2 /Al2 O3 = 12.5) was synthesised hydrothermally using water glass and sodium aluminate as source of silicon and aluminium respectively and tetraethyl ammonium bromide as template. Both  and mordenite samples were converted into ammonium form by repeated ion exchange with aqueous solution of 1 M NH4 Cl at 80 ◦ C and then converted into H-form by calcination at 550 ◦ C for 5 h. H-form of zeolite- and mordenite were loaded with 0.1 wt.% Pt by incipient wetness impregnation (IWI) and the resulting materials are designated as catalysts A1 and B1 , respectively. Parts of catalyst A1 were separately impregnated by IWI method with 0.1, 0.3 and 0.5 wt.% Ni and the resulting samples are designated as A2 , A3 and A4 , respectively. Similarly 0.1, 0.3 and 0.5 wt.% Ni was impregnated over B1 and the product materials are designated as B2 , B3 and B4 , respectively. For comparison purposes, 0.3 wt.% Ni, 0.1 wt.% Pt/H-, and 0.1 wt.% Ni, 0.1 wt.% Pt/H-MOR were prepared by ion-exchange method (IE) and the resulting materials are designated as catalysts A5 and B5 respectively. Catalysts 0.4 wt.% Ni/H- and 0.2 wt.% Ni/H-MOR were also prepared by IWI and are designated as catalysts A6 and B6 , respectively. The metal loaded catalysts were dried at 120 ◦ C for 12 h. Each of the earlier mentioned catalysts (1.5 g each) were packed in a quartz reactor and activated at 550 ◦ C for 3 h under N2 atmosphere. Then the temperature was lowered to 475 ◦ C under hydrogen flow (30 ml/min/g) for 6 h in order to reduce the metal ions.
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486
2.2. Characterisation The purity of zeolite- and hydrothermally synthesised mordenite was analysed using Rigaku X-ray diffractometer with Ni filtered Cu K␣ radiation (λ = 1.54 Å) in the scan range of 2θ between 5 and 60◦ . The line broadening XRD patterns for the metal loaded zeolites were analysed using a Siefert diffractometer using Cu K␣ (λ = 1.54 Å) radiation. Transmission electron microscopy measurements for the reduced catalysts were carried out in JEOL 200 kV electron microscope. Specimens were enlarged using thin photographic paper. The size of all the metal particles visible in each photograph was measured manually and averaged. The state of metals in the reduced catalyst was determined by ESCA. The ESCA spectras were acquired with a surface analysis system (ESCALAB-MK11, VG Scientific) by using the Mg K␣ radiation (1253.6 eV) with pass energy of 50 eV. During the spectral acquisition the pressure of the analysis chamber was maintained at better than 1 × 10−7 Torr. The total acidity of catalysts was measured by TPD of ammonia. The experimental procedures for ammonia adsorption are presented in previous reports [12–15]. The extent of ammonia adsorbed over each catalyst was measured by TGA in a TA 3000 Mettler system. Nitrogen as purge gas was passed during desorption of ammonia. The TGA study was conducted at a heating rate of 10 ◦ C/min up to 600 ◦ C. The acidity of the catalysts was also measured by pyridine adsorbed FT-IR spectroscopy. The experimental procedures and pretreatment conditions described for TPD-NH3 were followed and the pyridine in vapour form was used as probe molecule instead of ammonia. The FT-IR spectra of the pyridine adsorbed catalyst was recorded in a Nicolet (AVATAR) spectrometer in absorbance mode using KBr pellet technique. Surface area measurements were carried out in sorptomatic 1990 CE Instrument following the BET procedure using N2 as adsorbent at liquid nitrogen temperature. 2.3. Catalytic studies The n-heptane hydroisomerisation was carried out at atmospheric pressure in a fixed bed continuous down flow quartz reactor. About 1.5 g of catalyst was packed and placed in a tubular furnace. The reactor
471
system was flushed with dry nitrogen for 3 h and then the metals were reduced at 475 ◦ C under hydrogen flow (30 ml/min/g) for 6–7 h. After reduction, the temperature was lowered to reaction temperature. The reactant n-heptane was fed into the reactor by a syringe pump at LHSV = 1.33 h−1 , and pure hydrogen gas at a flow rate of 20 ml/min/g was passed with reactant and were preheated. The reaction products were passed through condenser in ice cold condition attached to the end of the reactor and were collected in a trap kept in ice at a time interval of 1 h. The products were analysed by Hewlett Packard 5890 A gas chromatograph equipped with FID. The identification of the products was done by GC–MS (SHIMADZU QP5000). The material balance calculation shows that more than 95% of the feed was recovered as products.
3. Results and discussion 3.1. Characterisation 3.1.1. XRD The XRD pattern (not shown) of -zeolite exhibits the most intense diffraction peaks at 2θ = 20–32◦ and thus confirms the BEA structure of zeolite- as well as its good crystalline nature. The XRD pattern (not shown) of mordenite obtained show sharp peaks between the 2θ values 6–30◦ indicating the presence of MOR structure. Both are found to be in good agreement with standard references [16,17]. 3.1.2. Line broadening XRD Line broadening XRD patterns obtained for catalysts A3 and A4 (A series) and B2 and B3 (B series) are shown in Fig. 1a and b, respectively. The intensity of the XRD peaks is found to decrease with increasing Ni addition in both the systems. By using Cu K␣ X-ray, the intense peaks of Pt, Ni (cubic) and NiO (hexagonal) are expected at the 2θ values of 39.8, 44.5 and 43.3◦ , respectively. But in the present case, all the above peaks are missing may due to lower concentration of Pt as well as Ni species. The decrease in XRD peaks intensity with increasing Ni addition over both the systems can be accounted in terms of decrease in crystallinity and increasing pore blockage of support materials by the added Ni species.
472
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486
(Ni–Pt) particles on the surface of the supports. The average size of the particles on each support is determined and is presented in Table 1. It is observed that the average particle size is found to increase with increasing Ni addition. Similar observations were already made in our previous studies over Ni–Pt loaded zeolites [12,13] and SAPOs [14]. Jao et al. [10] from their TPR spectra characterised by a single peak of Ni (0.5) Pt (0.26)/H-MOR catalysts observed a decrease in Ni reduction temperature with increasing Pt concentration and presumed a catalytic reduction of Ni due to mobile platinum oxide particles colliding into each other by thermal migration and the nickel oxide particles catalytically reduced by prereduced Pt particles. The average particle sizes of catalysts A3 and B2 are found to be 4.55 and 3.71 nm, respectively and that of catalysts A4 and B3 are 10.78 and 5.42 nm, respectively. Such particles may be larger in size compared to the pores of zeolite supports (Pore size of  is (5.2 Å × 7.2 Å (12MR) and 6.5 Å × 5.6 Å) and that of MOR is 6.7 Å × 7.0 Å, respectively) with a chance for thermal mobility during reduction and hence may be located mainly outside the pores as reported by Canizares et al. [18] in Ni/H-MOR catalysts. Ostard et al. [19] analysed the TEM pictures of 0.9 and 5 wt.% Pt containing MOR and found that the average Pt particle size increases with increasing metal content as well as increasing metal reduction temperature. Canizares et al. [18] used TEM to study the effect of Ni loading technique on the particle size of the Ni/MOR catalysts and observed that large particles outside the zeolite crystals are always formed, regardless of metal loading technique (ion-exchange and impregnation). From the observation that the Ni–Pt particle of definite size grows further on the addition of Ni, is possible only because of the migration of Ni particles during the reduction.
Fig. 1. Line broadening XRD patterns of (a) catalysts A3 and A4 ; (b) catalysts B2 and B3 .
3.1.3. TEM analysis The TEM pictures of catalysts A3 , A4 , B2 and B3 , are in shown in Fig. 2a–d, respectively. The black dots seen on the support matrix are assumed as bimetallic
3.1.4. ESCA The ESCA spectra of Pt and Ni species in reduced catalysts A3 , A4 , B3 and B4 are shown in Fig. 3. In the case of Pt, two major peaks are observed irrespective of the support and the amount of nickel. The peaks with binding energy values of 71.0 and 74.5 eV are corresponding to the core level Pt 4f7/2 and Pt 4f5/2 transitions respectively indicating the presence of platinum in metallic state. However the presence of Pt in higher oxidation state can not be discarded because of
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486
Fig. 2. TEM pictures of catalysts (a) A3 ; (b) A4 ; (c) B2 ; (d) B3 .
473
474
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486
Table 1 Physiochemical characteristics of A and B series catalysts S. No.
1 2 3 4 5 6 7 8 9 10 11 12
Catalyst
A1 A2 A3 A4 A5 A6 B1 B2 B3 B4 B5 B6
Pt content (wt.%)
Ni content (wt.%)
BET surface area (m2 /g)
NH3 -TPD (mmol/g) LT-peak
HT-peak
0.1 0.1 0.1 0.1 0.1 – 0.1 0.1 0.1 0.1 0.1 –
– 0.1 0.3 0.5 0.1 0.5 – 0.1 0.3 0.5 0.1 0.2
565 527 477 476 483 492 463 405 360 326 412 426
0.775 0.763 0.710 0.575 0.723 0.718 0.728 0.644 0.634 0.625 0.662 0.653
0.189 0.172 0.140 0.130 0.138 0.155 0.416 0.212 0.177 0.172 0.195 0.224
Total acidity (mmol/g)
Particle size (nm)
0.964 0.935 0.850 0.705 0.861 0.873 1.144 0.856 0.811 0.797 0.857 0.877
– – 4.55 10.78 – – – 3.71 5.42 – – –
LT: low temperature peak, HT: high temperature peak.
the possibility of overlap with Al 2p transition of support [21]. The ESCA spectra of Ni 2p3/2 peaks have two peak maxima with binding energies 852.3 and 854.0 eV indicating the presence of metallic nickel and NiO, respectively. A broad peak seen around 857.0 eV
in catalysts A4 and B4 indicate the presence of Ni2+ and the formation of NiAl2 O4 from which the reduction of Ni2+ is very difficult (Ni metal: 852.3 eV; NiO: 853.3 eV; NiAl2 O4 : 857.2 eV in Phi ESCA data book). Minchev et al. [22] observed that remarkable amount
Fig. 3. ESCA spectra of catalysts A3 , A4 , B2 , and B3 .
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486
of nickel remaining unreduced in their XPS studies on the reduction of NiY-zeolites. Further, the XPS study on Ni–mordenite by Narayanan [23] also showed that the reduction of Ni in mordenite is rather difficult and during reduction multiple species of nickel are formed. Xiao and Meng [24] analysed the reduction and oxidation behaviour of Ni containing HZSM-5 by XPS and found many forms of nickel are exists on the surface of the zeolite due to the interaction between Ni2+ and zeolite. Further they observed the surface enrichment of aluminium and nickel. An enhanced reduction of Ni2+ to Ni0 by the addition of Pt was concluded by Malyala et al. [20] from their XPS study on Ni/Y-zeolite and Ni–Pt/Y-zeolite. Thus in the present case too the added Pt is supposed to favour the reduction of nickel cations in the region 0–0.3 wt.% Ni above which NiO is observed
475
(for 0.5 wt.%) in A series catalysts. In B series NiO formation is observed in adding 0.3 wt.% Ni itself. 3.1.5. TGA–TPD of NH3 The desorption of ammonia was carried out over A and B series catalysts by TGA method and the amount of ammonia desorbed from each of A and B series catalysts are presented in Table 1. The model TGA–TPD curves for the typical catalysts A3 and B2 are shown in Fig. 4a and b, respectively. The temperature of desorption and amount of ammonia desorbed are the indexes of strength and number of acid sites, respectively [25]. There are two weight losses occurring at two different temperature ranges may be due to the desorption of adsorbed ammonia on weak and strong acid sites. The first weight loss occurs from A and B series catalysts in the temperature range of
Fig. 4. Typical TGA–TPD curves of catalysts (a) A3 and (b) B2 .
476
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486
200–250 ◦ C. The second weight loss from A series catalysts occurs at 350–400 ◦ C and that in B series catalysts at 500–550 ◦ C. From the Table 1, it is found that low temperature desorption is always more than that at higher temperature. When Ni is added in increasing amount to the Pt loaded supports, the total acidity continuously falls in both the system. Jao et al. [10] in their TPD study of Pt/MOR systems observed the same trend in acidity and accounted in terms of occupation of some of the acid sites of mordenite support by Ni species when increasingly loaded. On similar lines, the current observation is explained in terms of increasing occupation of the acid sites on the surface of the supports by the increasing Ni addition which process may be parallel with Ni adding up itself to Pt particle and allowing the growth of Pt–Ni particles. When the Ni is brought into the Pt loaded zeolites by ion-exchange method, there is no significant change in acidity when compared with that of impregnated catalysts. This shows that Ni dispersion whether done by IWI or ion exchange, a definite amount of it occupy the acid sites and the rest contributing to the growth of Pt–Ni particles. For a definite amount of Ni added, the fall in acidity is more in mordenite than -zeolite. For example, it is observed that when 0.3 wt.% Ni is added to the supports containing 0.1 wt.% Pt, the acidity fall is more in mordenite material (0.303 mmol/g) than that over -zeolite (0.114 mmol/g). This observation may be attributed to the structural differences among the zeolites used. The same low temperature and high temperature ranges observed in the NH3 desorption process from all Pt, Ni and Pt-Ni loaded catalysts of the two zeolite supports lead to the conclusion that the acid strength of the acid sites on that support does not vary by the addition of Ni. 3.1.6. Pyridine adsorbed FT-IR spectroscopy The pyridine adsorbed FT-IR spectra of A (A1 , A3 and A4 ) and B (B1 , B3 and B4 ) series catalysts are shown in Fig. 5a and b, respectively. A sharp peak appeared around 1545 cm−1 on all the catalysts indicating the presence of pyridine adsorbed on Brønsted acid sites of zeolites. The pyridine adsorption on Lewis acid sites is indicated by an another sharp peak around 1455 cm−1 . The increasing addition of Ni in A and B series catalysts has no effect on the position of both Brønsted and Lewis acid site peaks but the intensity of the each peak was found to decrease indicating that
the number of both Brønsted and Lewis acid sites decreases with increasing Ni loading. The decrease in the number of acid sites may be due to the occupation of acid sites by Ni species when increasingly loaded as observed in TPD-NH3 studies. A broad peak obtained around 1490 cm−1 indicates the physically adsorbed pyridine. The BET surface area (Table 1) of the both A and B series catalysts show a decreasing trend with increasing Ni addition, which may reflect the observation made in acidity measurement studies. 3.2. Catalytic studies Hydroisomerisation of n-heptane was carried out over the above bifunctional catalysts at LHSV = 1.33 h−1 in the temperature range 225–375 ◦ C in steps of 50 ◦ C. It is observed that the monobranched isomers, 2-methyl hexane (2MH), 3-methylhexane (3MH) and 3-ethyl pentane (3EP), dibranched isomers, 2,3 dimethyl pentane (23DMP), 2,4 dimethyl pentane (24DMP), 2,2 dimethyl pentane (22DMP) and 3,3 dimethyl pentane (33DMP) and tribranched isomer 223 trimethylbutane (223TMB) are the major products due to skeletal rearrangement of carbon chain. Traces of cracked, aromatised and cyclised products were also observed. The product distributions in n-heptane hydroisomerisation over A and B series catalysts are presented in Tables 2 and 3, respectively. The percentage conversions of n-heptane over A and B series catalysts obtained at different reaction temperatures are shown in Fig. 6a and b, respectively. It is observed that n-heptane conversion generally increases with increasing reaction temperature over all the catalytic systems. An abrupt increase in n-heptane conversion is observed in the case of A series catalysts when the reaction temperature increases from 225 to 275 ◦ C compared with B series catalysts. At still higher temperatures, the corresponding changes are marginal. The selectivity to isomerised products slightly decreases with increasing reaction temperature. Among the A and B series catalysts, B series catalysts show maximum selectivity towards cracking may be due to the higher acidic strength, small and one dimensional pore structure of mordenite. Also, the crystal size of mordenite is usually larger than that of -zeolite and in the case of cracking it would be more favoured over mordenite because of the larger size of the crystals.
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486
Fig. 5. Pyridine adsorbed FT-IR spectra of (a) A series and (b) B series catalysts.
477
478
Products
225 ◦ C A1
2MH 3MH 3EP 23DMP 24DMP 22DMP 33DMP 223TMP Cracked products Conversion (wt.%) Mo/Mu I/C
275 ◦ C A3
A4
A5
10.0 11.5 8.0 8.3 – – 1.2 2.7 1.0 2.0 0.5 1.0 – – – – 3.0 3.5
12.5 9.5 – 4.5 3.3 1.5 – – 4.0
11.5 8.5 – 3.3 2.0 0.5 – – 4.8
11.0 8.2 – 3.0 1.7 – – – 5.8
23.7 29.0
35.3
30.6
29.7
6.6 6.9
A2
3.4 7.28
2.36 7.82
3.4 5.37
4.0 4.12
A6
325 ◦ C
A1
A2
A3
A4
A5
A6
A1
8.0 6.3 – 1.5 1.0 0.5 – – 5.5
16.7 14.0 1.0 4.2 2.7 1.0 – – 6.6
18.4 16.3 1.0 5.8 4.0 2.3 – – 7.5
19.5 16.2 2.3 7.0 4.7 3.0 1.0 – 7.8
18.2 14.5 2.0 5.0 3.7 2.0 – – 8.8
18.0 15.2 1.0 4.0 3.0 0.5 – – 10.5
12.5 9.5 – 2.8 2.0 – – – 9.8
22.3
46.2
55.3
61.5
54.2
52.2
36.6
4.76 3.05
4.01 6.0
2.95 6.37
2.42 6.88
3.24 5.15
4.56 3.97
4.8 2.73
375 ◦ C A3
A4
A5
A6
A1
A2
A3
A4
A5
A6
17.5 18.3 15.0 16.3 1.5 2.0 5.0 7.2 4.2 6.0 2.0 3.3 – 1.0 – 1.0 8.3 9.1
19.0 16.8 3.0 9.5 8.0 4.8 2.0 1.0 9.4
18.3 15.7 2.0 6.3 5.0 2.5 0.5 – 10.7
17.4 15.5 2.0 5.0 4.2 3.0 1.0 – 12.1
14.0 11.5 1.0 3.0 2.3 1.0 – – 12.7
17.2 15.3 2.0 6.0 5.2 2.5 1.0 1.0 10.0
18.7 16.5 2.0 9.2 6.8 4.8 1.8 1.5 10.8
18.5 17.0 2.7 11.0 8.2 5.6 3.0 2.0 11.3
18.3 15.8 1.5 7.5 6.4 4.2 1.0 – 12.5
17.0 14.8 2.5 6.5 5.0 3.0 1.5 – 15.7
14.8 12.0 1.0 4.5 4.0 2.0 1.0 – 15.2
53.5 64.2
73.5
61.0
60.2
45.5
60.2
71.6
77.3
67.2
66.0
54.5
3.0 5.4
A2
1.97 6.05
1.53 6.81
2.86 4.7
2.64 3.97
4.2 2.58
LHSV = 1.33 h−1 , weight of catalyst = 1.5 g, H2 flow = 20 ml/min/g, time on stream = 1 h, Mo: monobranched isomers, Mu: multibranched isomers.
2.51 5.02
1.54 5.62
1.28 5.84
1.86 4.37
2.14 3.2
2.41 2.58
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486
Table 2 Product distribution of n-heptane hydroisomerisation over A series of catalysts at different temperatures
Products
225 ◦ C B1
B2
B3
B4
B5
B1
B2
B3
B4
B5
B6
B1
B2
B3
B4
B5
B6
B1
B2
B3
B4
B5
B6
2MH 3MH 3EP 23DMP 24DMP 22DMP 33DMP 223TMP Cracked products Conversion (wt.%) Mo/Mu I/C
10.7 8.0 – 2.0 – – – – 3.7
15.6 10.0 0.5 3.6 1.6 – – – 4.5
15.3 9.2 1.0 2.4 1.0 – – – 5.6
13.5 7.3 – 1.5 1.0 – – – 5.5
13.0 7.5 – 1.5 1.0 – – – 6.0
8.6 6.0 – 1.0 – – – – 5.0
15.5 10.2 1.5 2.8 1.5 1.0 – – 6.5
17.5 12.0 1.5 4.8 3.0 1.5 1.0 – 6.5
16.5 11.4 1.5 3.5 2.0 1.0 – – 8.0
15.8 10.5 1.0 3.2 2.0 – – – 8.2
15.5 10.0 – 2.3 2.0 – – – 8.8
12.0 7.8 – 1.4 1.0 0.5 – – 8.3
18.5 13.7 2.5 3.5 2.0 1.0 – – 9.0
20.0 15.5 3.0 5.5 3.0 1.5 1.0 – 9.0
20.0 14.5 2.5 4.0 2.8 1.0 – – 10.0
18.5 12.0 2.0 4.0 2.7 – – – 9.8
16.2 12.8 1.0 2.8 2.0 – – – 10.5
13.0 9.6 1.5 2.0 1.2 – – – 12.0
19.7 13.5 3.0 4.7 3.5 2.0 1.2 0.5 11.5
20.5 15.0 3.5 6.5 4.2 2.3 1.8 1.0 12.5
20.5 13.8 3.0 5.0 3.5 1.5 – – 14.0
19.0 12.8 2.0 4.0 3.2 1.0 – – 13.5
17.5 13.0 1.6 3.3 2.0 1.5 – – 13.5
15.0 10.7 1.0 3.0 2.5 1.0 – – 15.5
24.4
35.8
34.5
28.8
29.0
20.6
39.0
47.8
43.9
40.7
38.6
31.0
50.2
58.5
54.8
49.0
45.3
39.3
59.6
67.3
61.3
55.5
52.4
48.7
3.0 6.2
4.5 4.5
9.35 5.6
275 ◦ C
5.02 6.95
7.5 5.16
8.32 4.23
8.2 3.83
B6
14.6 3.12
5.13 5.0
325 ◦ C
5.25 3.96
5.3 3.38
6.8 2.73
5.23 4.57
375 ◦ C
3.5 5.5
4.7 4.48
4.8 4.05
6.25 3.3
7.53 2.27
LHSV = 1.33 h−1 , weight of catalyst = 1.5 g, H2 flow = 20 ml/min/g, time on stream = 1 h, Mo: monobranched isomers, Mu: multibranched isomers.
3.09 4.18
2.46 4.38
3.73 3.37
4.12 3.11
4.72 2.88
4.10 2.14
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486
Table 3 Product distribution of n-heptane hydroisomerisation over B series of catalysts at different temperatures
479
480
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486 80
Conversion (wt%)
60
40
20
0 200
250
300 Temperature(˚C)
350
400
250
300 Temperature (˚C)
350
400
(a)
80
Conversion (wt%)
60
40
20
0 200 (b)
Fig. 6. Effect of temperature on n-heptane conversion over (a) A series and (b) B series catalysts. (䉬) A1 ; (䊏) A2 ; (䉱) A3 ; (×) A4 ; (∗) A5 ; (䊉) A6 , (䉬) B1 ; (䊏) B2 ; (䉱) B3 ; (×) B4 ; (∗) B5 ; (䊉) B6 .
The effect of Ni addition on n-heptane conversion was studied by comparing the n-heptane conversion over Ni containing and Ni free catalysts at a definite reaction temperature. The addition of 0.1 wt.% Ni over 0.1 wt.% Pt/H- enhances the n-heptane conversion to 71.6 from 60.2 wt.% at 375 ◦ C. The corresponding increase in conversion at 225 ◦ C is from 23.7 to 29.0 wt.%. Similarly, the same amount of Ni addition over 0.1 wt.% Pt/MOR increases the n-heptane conversion to 67.3 and 54.9 wt.% from 59.6 and 47 wt.%,
respectively, at the above said reaction temperatures. The increasing trend in conversion with Ni addition continues up to 0.3 and 0.1 wt.% over A and B series catalysts, respectively. Further increase of Ni content (catalyst A4 and B3 ) leads to a decreasing trend in n-heptane conversion at all the temperatures studied. A maximum conversions of 77.3 and 67.3 wt.% were obtained over catalysts A3 and B2 respectively at 375 ◦ C. Also, the Ni addition up to the above threshold values enhances the isomerisation selectivity. Catalysts A3 and B2 show the isomerisation selectivity of 85.3 and 84.4%, respectively, at 375 ◦ C which are considerably higher than that observed over catalysts A1 (83.4%) and B1 (80.7%), respectively. Further increase in Ni content over A and B series catalysts (catalysts A4 and B3 ) leads to a significant decrease in isomerisation selectivity. The selectivity towards the cracking is found to be under control up to the above threshold Ni addition at all the temperatures studied. The increase in n-heptane conversion, isomerisation selectivity and suppression of cracking with increasing Ni addition is similar to the observation made by Jao et al. [10]. They studied the n-heptane conversion over 0.26 wt.% Pt/H-MOR with varying Ni (0.5 and 1.5 wt.%) content and found that a maximum n-heptane conversion of 39.2 wt.% with high selectivity to isomerisation and simultaneous suppression of fuel gas formation over 0.5 wt.% Ni 0.26 wt.% Pt/H-MOR and further increase in Ni addition (>0.5 wt.%) leads to fall in conversion. Further they attributed the decreasing fuel gas formation with increasing Ni content in terms of increase in the metallic site/acid site (NM /NA ) ratio when increasing the Ni content. For a catalyst with higher (NM /NA ) ratio, the diffusion distance between the two metallic sites is shorter than a catalyst with lower (NM /NA ) ratio. In the present case, the initial increase in conversion with suppression of cracking by increasing addition of Ni may be due to the formation and growth of catalytically active Ni–Pt bimetallic particles of nanometre scale as well as better balance between the acid sites of support and bimetallic particles. The TEM studies show that the average particle size of bimetallic particles increases with increasing Ni loading. Catalysts A3 and B2 have an average particle size 4.55 and 3.71 nm by TEM analysis. Bimetallic particles with the above sizes are supposed to offer better synergism with acid sites of supports and favours the monomolecular mechanism
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486
of n-heptane isomerisation as reported by Weitkamp and Erdgas [26]. Also, the ESCA spectra of catalyst A3 shows the complete reduction of Pt and Ni to metallic state. The added Ni species may increase the total number of active metallic sites, i.e. the metallic sites/acid sites ratio may increased towards the optimum value for isomerisation reactions. The initial increase in hydroisomerisation selectivity with decreasing cracking with increasing Ni addition is due to the more availability of metallic sites in the vicinity of acid sites enabling rapid hydrogenation of the carbenium ions and desorbing them as alkanes before they undergo for cracking reactions, i.e. the probability for cracking the olefinic intermediate during the migration from one metallic site to another site is less [27]. The decreasing conversion trend with increasing cracking observed over catalysts with higher Ni content may be due to the formation of larger bimetallic particles and presence of unreduced Ni as NiO and NiAl2 O4 as evidenced by ESCA. Catalysts A4 and B3 have the average particle size of 10.78 and 5.42 nm by TEM analysis. Such large bimetallic particles may block the pores of zeolites and exert restriction to the movement of bulky reaction intermediates. Also, the presence of unreduced Ni species, which are inactive in hydrogenation–dehydrogenation step may disturb the synergism between metal–acid sites leading to lower activity. The decrease in crystallinity of the supports indicated by line broadening XRD is also responsible for the lower activity. The catalysts prepared by ion-exchange method (A5 and B5 ) show considerably lower activity with higher selectivity to cracked products compared to impregnated catalysts A3 and B2 at all temperatures studied. Catalysts A5 and B5 show the maximum n-heptane conversion of 66.0 and 52.4 wt.% with isomerisation selectivity of 76.2 and 74.2%, respectively, at 375 ◦ C, which are significantly lower than those over A3 and B2 , respectively. The lower activity of ion-exchanged catalysts than that of impregnated catalysts with same metal content was already reported for Pt/zeolite catalysts [28,29]. The Ni only loaded catalysts (A6 and B6 ) show very low n-heptane conversion with higher cracking tendency compared with catalysts A3 and B2 all with same amount of total metal content. The effect of temperature on the selectivity of individual heptane isomers over typical catalysts A3 and
481
B2 are shown in Fig. 7a and b, respectively. Generally, the selectivity of the monobranched isomers 2MH and 3MH is found to decrease with increasing temperature as well as conversion over all the catalytic systems. The fall in selectivity of 2MH is more than that of 3MH. Also, the selectivity of 2MH is found to be always higher than that of 3MH. The selectivity of dibranched isomers 23DMP, 24DMP, 22DMP and 33DMP and tribranched isomer 223TMB increases with increasing temperature. Among the dibranched isomers, the selectivity of 23DMP and 24DMP is found to be higher than that of other dibranched isomers. The 33DMP as well as the monobranched 3EP do not appear at low (225 ◦ C) temperatures over all the catalytic systems. The increasing formation of dibranched isomers with increasing temperature as well as conversion indicates that the dibranched isomers are secondary products formed from primary monobranched isomers by protonated cyclopropane intermediate formation [13]. The tribranched isomer 223TMB is found to appear only at higher temperatures. The decrease in the selectivity of monobranched isomers and increase in the selectivity of multibranched isomers as well as cracking with increasing temperatures suggest that the transformations are monobranched into dibranched and dibranched into tribranched isomers. The presence of 3EP over many of the catalysts even at very low conversion cannot be explained by protonated cyclopropane mechanism through which 3EP cannot be formed. Another possible reaction intermediate of cyclobutanic structure as reported by Fajula and Gault [30] and a rapid rearrangement by ethyl shift are considered to explain the 3EP formation. Zeolite  based catalysts always show higher amount of multibranched isomer formation than mordenite (B series) based catalysts eventhough the acid strength of zeolite  is lower than that of mordenite as evidenced by the acidity studies. The earlier observations indicate that the pore structure, size and kinetic diameter of the reactants as well as products and intermediates are responsible for the selectivity of multibranched isomers. Zeolite- has three dimensional linear channels of 12MR (5.7 Å × 7.5 Å) and tortuous channel (6.5 Å × 5.6 Å) with channel intersections [31]. Mordenite has one dimensional channels of 12MR (6.7 Å×7.0 Å) with side pockets (8MR) at the walls of main channels. The kinetic diameter
482
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486
Selectivity (%)
40
20
0 200
250
(a)
300
350
400
Temperature(˚C)
Selectivity (%)
40
20
0 200 (b)
250
300 Temperature (˚C)
350
400
Fig. 7. Effect of temperature on the selectivity of individual heptane isomers over catalyst (a) A3 and (b) B2 (䉬) 2MH; (䊏) 3MH; (䉱) 3EP; (×) 23DMP; (∗) 24DMP; (䊉) 22DMP; (+) 33DMP; ( ) 223TMB.
of n-heptane, monobranched and dibranched isomers are found to be 0.49, 0.56 and 0.70 nm, respectively. Hence the bulky olefinic intermediates formed in the zeolite channel as well as outside may not have access to the acid sites due to hindrance by larger sized metallic particles formed to the movement of intermediates.
Due to the unidimensional pore structure in mordenite, the bimetallic particles grow in size almost to fit the pores. Hence many of the acid sites inside the pores are inaccessible to the reaction intermediates. But in the case of -zeolite (three dimensional), even if one of the channels is blocked, the reactants and interme-
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486
diates can have access to metal or acid sites through the other channels. The effect of temperature on the ratios among the heptane isomers over A and B series catalysts are derived from product distributions presented in Tables 2 and 3 respectively. The thermodynamic equilibrium distribution of heptane isomers at 350 ◦ C reported by Vazquez et al. [7] has been included in order to find the deviations of the ratios among the isomers experimentally obtained. The thermodynamic equilibrium ratio of monobranched/multibranched (Mo/Mu) is 0.62 and that of 2MH/3MH is 0.81 at 350 ◦ C. It is observed from the Tables 2 and 3, that the ratio of Mo/Mu decreases with increasing reaction temperature. It is generally considered that the branching of olefins occurs through the protonated cyclopropane intermediate as per the comprehensive description of carbenium ion rearrangement and cleavage by Brouwer and Oelderik [33]. From the point of view of PCP mechanism, the isomerisation of monobranched to dibranched heptanes goes through secondary and tertiary carbocations and therefore it should be thermodynamically favourable. The Mo/Mu ratio is found to decrease with increasing temperature and Ni content up to the threshold value. However, the rate constant for the formation of multibranched (MTB) isomers of n-octane from monobranched (MB) is 1.5 lower than the rate of formation of monobranched isomers from n-octane [34]. Taking into account this fact and the influence of the hydrocarbon chain length on the relative rate of formation of mono and multi branched isomers, Vazquez et al. [7] expected that in the case of n-heptane, KMTB/KMB = 1/2 or KMB/KMTB = 2 and the isomerisation follows a scheme of the type n-heptane → MB → MTB. The Mo/Mu ratio over catalyst A3 is found to be in the range 2.36–1.28 at 225–375 ◦ C, which is slightly deviated from the expected value of 2. The Mo/Mu ratio on B2 is 5.0 at 225 ◦ C and 2.46 at 375 ◦ C. The increasing MTB formation with increasing Ni content as well as temperature indicates that the addition of Ni and higher reaction temperature favour the protonated cyclopropane intermediate mechanism which is considered as one slower than 1,2 alkyl hydride shift mechanism which leads to again the formation of monobranched isomer. The enhancement in PCP mechanism can be accounted in terms better balance between metal particle formed and acid sites.
483
Catalysts with high average metal particle sizes are found to be not favouring PCP mechanism may be due to the larger particle size which affects the development of better metal–acid synergism. The unreduced Ni (shown by ESCA) not a species active in dehydrogenation–hydrogenation step has too its effect on the entire isomersiation scheme of n-heptane and hence ratio among the isomers. It is well known from the comprehensive rules for the rearrangement of alkyl carbenium ion, that the formation of 2- and 3-methyl hexanes would only take place by protonated cyclopropane (PCP) mechanism, a 2MH/3MH ratio of 0.5 was expected by Vazquez et al. [7]. The 2MH/3MH ratio over all the catalytic systems was found to decrease slightly with increase in temperature and increase with increasing Ni content. Generally, 2MH/3MH ratio was found to deviate more from the expected value of 0.5 as well as thermodynamic equilibrium value of 0.81 at 350 ◦ C. The extent of deviation increases with increasing Ni content indicating that there is some inhibition in the formation of 3MH from 2MH by 1,2-alkyl hydride shift which is normally faster than the PCP mechanism. The decrease of Mo/Mu ratio and increase of 2MH/3MH ratio with increasing Ni content indicate that the Ni addition facilitates the MTBs formation through PCP intermediate and suppresses the faster 1,2 alkyl hydride shift which usually leads to the formation of 3MH from 2MH. According to Blomsma et al. [8], the formation of gem-dimethyl isomers (22DMP and 33DMP) is not possible by the dimerisation–cracking (bimolecular) mechanism. But in the present case, considerable amounts of 22DMP and 33DMP are formed over all the catalytic systems and found to increase with increasing reaction temperature. Hence, the n-heptane isomerisation is considered to follow the mechanism of carbenium ion rearrangement and cleavage. According to comprehensive description of rules for carbenium ion rearrangement and cleavage given by Brouwer [32], the carbenium ion isomerisations that lead to change in degree of branching occur through protonated cyclopropane (PCP) ring intermediate formation. Hence, the 2MH/23DMP ratio is considered as an indication of rate of branching of heptane by protonated cyclopropane (PCP) intermediate mechanism. The effect of Ni addition on the ratio of 2MH/23DMP is studied by comparing the ratios over Ni free and
484
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486
9 8 7 2MH/23DMP
6 5 4 3 2 1 0 0
0.2
(a)
0.4
0.6
0.4
0.6
Ni content (wt%)
10 9 8
2MH/23DMP
7 6 5 4 3 2 1 0 0 (b)
0.2 Ni content (wt%)
Fig. 8. Effect of Ni addition on 2MH/23DMP ratio over (a) A series and (b) B series catalysts at different temperatures (䉬) 225 ◦ C; (䊏) 275 ◦ C; (䉱) 325 ◦ C; (×) 375 ◦ C.
Ni containing catalysts at different reaction temperatures and the results are shown in Fig. 8a and b for A and B series catalysts respectively. It is observed that the 2MH/23DMP ratio decreases with increasing Ni addition up to 0.3 wt.% for A series, and 0.1 wt.% for B series catalysts at all temperatures indicating the operation of PCP mechanism during the threshold
addition of Ni. Further Ni addition above the threshold values over both series catalysts increases the 2MH/23DMP ratio at all the temperatures studied. The isomerisation/cracking (I/C) ratio of an isomerisation catalyst is an indication of its suitability to isomerisation reactions. A good isomerisation catalyst must have very high I/C ratio. In Tables 2 and 3,
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486
485
Conversion (wt%)
80
60
40
20 0
2
4 Time (h)
(a)
6
8
Conversion (wt%)
80
60
40
20 0
1
2
(b)
3
4
5
6
7
Time (h)
Fig. 9. Effect of time on stream on (a) A series and (b) B series catalysts in n-heptane hydroisomerisation.
an increasing trend in I/C ratio is observed up to the Ni addition of 0.3 wt.% in A series catalyst, 0.1 wt.% in B series catalysts. Further Ni addition decreases the I/C ratio indicating the higher cracking activity of larger bimetallic particles. The I/C ratio over all the catalytic system is found to decrease with increasing temperature indicating that higher temperatures favour cracking. The sustainability of A and B series catalysts in n-heptane hydroisomerisation reaction is studied by conducting the time on stream study for 6 h at 375 ◦ C
and the n-heptane conversions are presented in Fig. 9a and b, respectively. All the catalysts show a fall in conversion when increasing the time on stream. Catalyst A3 of A series and B2 of B series catalysts show the minimum fall in their activity. Catalysts without Ni (A1 and B1 ) and Ni only loaded (A6 and B6 ) show the maximum fall in activity during the time on stream. The higher sustainability of the catalysts A3 and B2 is explained in terms of catalytically active bimetallic particles formation and better balance between metal–acid sites. The fall in activity of the
486
I. Eswaramoorthi et al. / Applied Catalysis A: General 253 (2003) 469–486
catalysts can be accounted in terms of coke formation due to cracking, which occupy the active sites of the catalysts leads to fall in activity.
4. Conclusion The XRD patterns of commercial  zeolite and hydrothermally synthesised mordenite confirm their good crystallinity and structure. The line broadening XRD analysis shows that the crystallinity of the zeolite decreases with increasing Ni loading. The TEM analysis shows the formation of bimetallic (Ni–Pt) particles of nanoscale size. The ESCA study reveals that complete reduction of Ni up to 0.3 and 0.1 wt.%, over 0.1 wt.% Pt loaded  and mordenite, respectively. The acidity measurements by both TPD-NH3 and pyridine adsorbed FT-IR spectral studies show that some of the acid sites are occupied by the added Ni species. Because of the best metal–acid balance between bimetallic particles and acid sites of the support, catalysts 0.3 wt.% Ni 0.1 wt.% Pt/H- and 0.1 wt.% Ni 0.1 wt.% Pt/Mordenite show an enhanced activity, multibranched isomers selectivity and sustainability of the catalysts in n-heptane hydroisomerisation. Further Ni addition leads to a fall in activity of the catalysts. Also, the -zeolite based catalysts are found be more suitable catalysts for n-heptane hydroisomerisation than mordenite based catalysts.
Acknowledgements The authors are gratefully acknowledge the financial support from Department of Science and Technology (DST), New Delhi, India. References [1] I.E. Maxwell, W.H.J. Stork, Stud. Surf. Sci. Catal. 58 (1991) 571. [2] H.W. Kouwenhoren, W.C. Zijll Langhout, Chem. Eng. Prog. 67 (1971) 65. [3] R.J. Taylor, R.H. Petty, Appl. Catal. 119 (1994) 121. [4] A. Gil, A. Diaz, L.M. Gandia, M. Montes, Appl. Catal. 109 (1994) 167. [5] D.L. Hoang, H. Berndt, H. Miessner, E. Schereier, J. Volter, van Santen, Ind. Eng. Chem. Res. 34 (1995) 55. [6] J.M. Ward, Fuel Processing Technol. 32 (1993) 55.
[7] M.I. Vazquez, A. Escardino, A. Corma, Ind. Eng. Chem. Res. 26 (1987) 1495. [8] E. Blomsma, J.A. Martens, P.A. Jacobs, J. Catal. 165 (1997) 241. [9] R. Le Van Mao, M.A. Saberi, Appl. Catal. A: Gen. 199 (2000) 99. [10] R.M. Jao, T.B. Lin, J.-R. Chang, J. Catal. 161 (1996) 222. [11] M.H. Jordao, V. Simoes, A. Montes, E. Cardoso, Stud. Surf. Sci. Catal. 130 (2000) 2387. [12] I. Eswaramoorthi, N. Lingappan, Korean J. Chem. Eng. 20 (2) (2003) 207. [13] I. Eswaramoorthi, N. Lingappan, Catal. Lett. 87 (3–4) (2003) 133. [14] I. Eswaramoorthi, N. Lingappan, Appl. Catal. 245 (2) (2003) 119. [15] A.M. Dominguez, A. Vazquez, A.J. Renouprez, M.J. Yacaman, Surf. Sci. 75 (1982) 101. [16] J.M. Newsman, M.M.J. Treacy, W.T. Koetsier, C.B. de Gruyter, Proc. Royal Soc., London A420 (1988) 375. [17] M.M.J. Treacy, J.B. Higgins, Collection of Simulated XRD Powder Patterns of Zeolites, fourth ed., Elsevier, Amsterdam, 2001. [18] P. Canizares, A. de Lucas, F. Dorado, A. Duran, I. Asencio, Appl. Catal A: Gen. 169 (1998) 137. [19] D.J. Ostard, L. Kustov, K.R. Poeppelmeier, W.M.H. Sachtler, J. Catal. 133 (1992) 342. [20] R.V. Malyala, C.V. Rode, M. Arai, S.G. Hegde, R.V. Chaudhari, Appl. Catal. A: Gen. 193 (2000) 71. [21] M.A. Arribas, F. Marquez, A. Martinez, J. Catal. 190 (2000) 309. [22] Ch. Minchev, V. Knazirev, L. Kosova, V. Pechev, W. Grunsser, F. Schimidt, in: L.V.C. Rees (Ed.), Proceedings of the Fifth International Conference on Zeolites, Heydon, London, 1980, p. 335. [23] S. Narayanan, Zeolites 4 (3) (1984) 231. [24] S. Xiao, Z. Meng, J. Chem. Soc. Faraday Trans. 90 (17) (1994) 2591. [25] L.-J. Leu, L.-Y. Hov, B.-C. Kang, C. Li, S.-T. Wu, T.-C. Wu, Appl. Catal. 69 (1991) 49. [26] J. Weitkamp, E.K. Erdgas, Petrochemie 31 (1978) 13. [27] G.E. Giannetto, G.E. Perot, M.R. Guisnet, Ind. Eng. Chem. Prod. Res. Dev. 25 (1986) 481. [28] H.Y. Chu, M.P. Rosynek, J.H. Lunsford, J. Catal. 178 (1998) 352. [29] G. Boskovic, K. Mimic, P. Pavloic, P. Putanov, Catal. Today 65 (2001) 123. [30] F. Fajula, F.G. Gault, J. Am. Chem. Soc. 98 (1976) 7690. [31] J.B. Higgins, R.B. Lapierre, J.L. Schlenker, A.C. Rohrman, J.D. Wood, Zeolites 8 (1988) 446. [32] D.M. Brouwer, in: R. Prins, G.C.A. Schult (Eds.), Chemistry and Chemical Engineering of Catalytic Processes, Sijthof and Noordhoff, Germantown, MD, 1980. [33] D.M. Brouwer, J.M. Oelderik, Recl. Trav. Chim. Pays. Bas. 87 (1968) 721. [34] M.H. Baltanas, H. Vansina, G.F. Froment, Ind. Eng. Chem. Proc. Res. Dev. 22 (1983) 531.
