ISSN 0023-1584, Kinetics and Catalysis, 2017, Vol. 58, No. 3, pp. 311–320. © Pleiades Publishing, Ltd., 2017. Original Russian Text © Luu Cam Loc, Nguyen Tri, Dao Thi Kim Thoa, N.A. Gaidai, Yu.A. Agafonov, Ha Cam Anh, Hoang Tien Cuong, A.L. Lapidus, 2017, published in Kinetika i Kataliz, 2017, Vol. 58, No. 3, pp. 327–337.
Kinetics of n-Hexane Isomerization over Supported Palladium Catalysts Luu Cam Loca, b, Nguyen Tria, Dao Thi Kim Thoab, N. A. Gaidaic, *, Yu. A. Agafonovc, Ha Cam Anhb, Hoang Tien Cuonga, and A. L. Lapidusc aInstitute
of Chemical Technology, Vietnam Academy of Science and Technology, Ho Chi Minh City, Vietnam b Ho Chi Minh City University of Technology, Ho Chi Minh City, Vietnam c Zelinsky Institute of Organic Chemistry, Russian Academy of Sciences, Moscow, 119991 Russia *e-mail:
[email protected] Received October 31, 2016
Abstract—The steady- and unsteady-state kinetics of n-hexane isomerization over Ni- or Co-promoted Pd/HZSM-5 catalysts in the presence of hydrogen has been investigated. The kinetics of the reaction is described by similar fractional rational rate equations differing in the values of their constants. Hydrogen exerts a favorable effect on the isomerization rate. The catalysts have been characterized by the BET-N2 method, X-ray diffraction, transmission electron microscopy, H2-TPR, and NH3-TPD. The introduction of a promoter (Ni or Со) strengthens the adsorption of the reactants and increases the amount of reactants adsorbed. Keywords: isomerization, kinetics, n-hexane, Pd catalysts DOI: 10.1134/S0023158417030090
INTRODUCTION Environmental regulations impose stringent limits on the concentration of aromatics, olefins, and oxygen-containing compounds in gasoline, implying a decrease in its octane number. This decrease can be compensated for by blending gasoline with isomers of С5–С6 paraffins, which have a higher octane number than their straight-chain analogues [1]. It is, therefore, essential to investigate the conversion of straightchain paraffins into their branched isomers. The isomerization of C5–C6 paraffins is typically carried out over bifunctional catalysts containing a small amount of Pt or Pd supported on an acidic material. The most promising supports are various zeolites. Pt/mordenite catalysts were reported [2, 3] to be very active in n-hexane hydroisomerization. Platinum and palladium catalysts supported on Beta zeolites have been employed in the hydroisomerization of С6–С7 n-paraffins and their mixtures [4, 5]. A comparison of palladium catalysts supported on zeolites HY and HZSM-5 demonstrated that Pd/HZSM-5 is more active and selective in n-hexane hydroisomerization than Pd/HY owing to the smaller size of palladium clusters and the higher Pd dispersion therein [6]. The introduction of a second metal (e.g., Co, Ni, Fe, or Re) enhances the n-paraffin isomerization efficiency of the catalysts. Yoshioka et al. [7, 8] discovered that a Pt–Ni/HUSY catalyst is much more active, selective, and stable in n-hexane hydroisomerization than its
Ni-free counterpart. The same conclusions were drawn for the hydroisomerization reaction by examining Ni-containing versus Ni-free Pt/SAPO-5 and Pt/SAPO-11 catalysts [9]. We studied the isomerization of n-pentane and n-hexane in the presence of hydrogen over Pd/HZSM-5 catalysts doped with Co, Ni, Fe, Cu, or Re or not [10]. It was demonstrated that the introduction of Co or Ni enhances the activity and on-stream stability of the palladium catalysts in the hydroisomerization of these paraffins. Here, we report the kinetics of n-hexane isomerization over these catalysts. Various forms of rate equations for the isomerization of С5–С6 paraffins have been suggested in the literature. Ribeiro et al. [11] investigated the kinetics of n-hexane isomerization over Pt/HY catalysts at 230– 270°C, varying the Pt content of the catalyst between 0.3 and 6.7 wt %. The n-hexane conversion did not exceed 10%. The highest isomerization rate was observed over the catalyst containing 1% Pt. The following rate equation was obtained for low n-hexane conversions, when the reverse reaction could be neglected:
r =
kPn−С 6Н14 , PH 2 + k1Pn−С 6Н14
(1)
where r is the n-hexane isomerization rate, k and k1 are constants, Pn−С 6Н14 and PH 2 are the n-hexane and hydrogen partial pressures. This equation referred to the bifunctional mechanism according to which n-paraffin
311
312
LUU CAM LOC et al.
dehydrogenation to an olefin takes place on platinum sites; olefin isomerization, on acid sites of the zeolite; olefin hydrogenation, on metal sites. The same mechanism and rate equation were suggested in other works [12, 13]. Chiang and Bhan [12] carried out n-hexane isomerization over Pt-containing mordenite-, faujasite-, and Beta-type zeolites (Т = 250°C), and Tannous et al. [13] used faujasite-supported Pt–La catalysts (Р = 30 atm, Т = 260–330°C). The slow step of n-hexane isomerization was assumed to be the rearrangement of the adsorbed olefin [11–13]. Ribeiro et al. [11] studied the hydrogen pressure effect on n-hexane isomerization (250°C, 40 atm) over a catalyst containing 2.2% Pt and set up a power-law rate equation in which the order with respect to hydrogen was 0.85 and the order with respect to n-hexane was 0.7. Hollo et al. [14] obtained power-law equations for the hydroisomerization of С5–С7 n-paraffins and their binary mixtures over mordenite-supported platinum catalysts at Т = 180–220°C, a hydrogen pressure of 5–40 atm, and Н2/hydrocarbon = 1–20. The orders of the equations with respect to С5, С6, and С7 hydrocarbons at PH2 ≤ 4 atm were 0.33, 0.39, and 0.15, respectively, and those with respect to hydrogen (at PH 2 ≤ 36 atm) were –0.81, –0.65, and –1.07. The slow step was assumed to be olefin isomerization on acid sites. For the hydroisomerization of n-octane, n-decane, and n-dodecane over Pt/US-Y zeolites at 130–250°C, 5– 100 atm, and Н2/hydrocarbon = 10–150, Froment [15] proposed Langmuir–Hinshelwood type equations. As is clear from the above literature survey, most authors adhere to the bifunctional mechanism of n-paraffin isomerization. However, this mechanism does not account for many observations, including the positive order of the reaction with respect to hydrogen, and for the fact that the reaction yields no olefins when it is carried out over metal catalysts in the absence of acidic support. For making n-paraffin isomerization proceed stably and for enhancing the branched isomer selectivity, it is necessary that hydrogen be present in the reaction system [16], but data concerning the effect of hydrogen on the isomerization kinetics and mechanism are lacking. These data will be referred to below in the discussion of the n-hexane isomerization mechanism. In this work, we report both the steady- and unsteadystate kinetics of n-hexane isomerization as studied by the response method. EXPERIMENTAL The support for palladium catalysts was HZSM-5 prepared by the calcination of (NH4)ZSM-5 (CBV3024E, Zeolist International, United States) at 500°C for 3 h. Its specific surface area was 353 m2/g, and its Si/Al ratio was 15. Monometallic and bimetallic palladium catalysts of optimal composition, which contained 0.8 wt % Pd and 1.0% Co or Ni [17], were prepared by impregnating HZSM-5 with an aqueous
solution of Pd(NO3)2, Co(NO3)2 ⋅ 6H2O, and Ni(NO3)2 ⋅ 6H2O, respectively. In the case of the bimetallic catalysts, the support was initially impregnated with cobalt or nickel nitrate and then with a palladium salt. After the impregnation products were dried at 110°C (2 h) and 130°C (3 h), they were treated with air at 400°C (2 h). Prior to performing experiments, the catalysts were reduced with flowing hydrogen at 400°C for 2 h. Physicochemical properties of the catalysts were studied by the BET-N2 method, X-ray diffraction (XRD), transmission electron microscopy (TEM), temperature-programmed reduction with hydrogen (H2-TPR), and temperature-programmed desorption of ammonia (NH3-TPD). The specific surface area of the catalysts (SBET) was determined by nitrogen adsorption on an ASAP 2020 V4.01 analyzer (Micromeritics, United States). The phase composition of the catalysts was determined by X-ray diffraction on an XD-5A diffractometer (Shimadzu, Japan) using CuKα radiation. Particles were sized using a JЕМ-1400 electron microscope (JEOL, Japan). Н2-TPR experiments were carried out using a ChemBET 3000 analyzer (Quantachrome, United States). The heating rate was 10о/min. The initial mixture consisted of 10.0% Н2 and 90.0% N2. Before TPR measurements, the oxidized samples were kept in flowing He for 2 h at 200°C and cooled to 30°C, and the temperature was then elevated to 900°C. The catalyst samples to be characterized by NH3TPD were reduced at 400°C with a flowing 10.0% Н2 + 90.0% N2 mixture for 2 h and were then cooled to 100°C in flowing He. After NH3 adsorption and removal of physically adsorbed ammonia at 150°C for 1 h, the temperature was raised to 700°C at a rate of 10о/min. The steady-state kinetics of n-hexane isomerization was studied in the presence of hydrogen in a flow circulation system at atmospheric pressure and temperatures of 215–260°C. The partial pressure of n-hexane (Pn0-hexane ), hydrogen (PH02 ), and intentionally added 0 isohexane (Pisohexanes ) was varied in the 33.3–137, 307– 718, and 0–19.2 hPa range, respectively. The gas hourly space velocity (v 0 ) was 3500–56000 h–1. Under these experimental conditions, the n-hexane conversion (Х) varied between 0.12 and 0.75. The side reaction was cracking, and its proportion did not exceed 10.0%. The reaction mixture was analyzed on an Agilent 6890 Plus gas chromatograph (Agilent Technologies, United States) fitted with a flame-ionization detector and a capillary column (30 m × 0.32 mm, 6% cyanopropylphenyl 94% dimethylpolysiloxane stationary phase). The unsteady-state kinetics of n-hexane isomerization was investigated at atmospheric pressure in a
KINETICS AND CATALYSIS
Vol. 58
No. 3
2017
KINETICS OF n-HEXANE ISOMERIZATION
small flow-through unit having three independent lines. The unit was coupled with an MSKh-6 time-offlight mass spectrometer (Russia). Transitional processes leading to a steady state were studied by the response method. Relaxation curves were obtained as a response to an abrupt change of the concentrations of reaction mixture components. The residence time, i.e., the ratio of the volume of the reaction system to the flow rate did not exceed 6 s, and this fact was taken into account in the construction of the relaxation curves. The following masses were registered: 2 (hydrogen), 43 (isohexanes), and 57 (n-hexane). In each experiment, a single substance was quantified at 1-s intervals. Under the conditions of our experiments, the process took place in the differential reactor regime. The experiments were carried out on the following catalysts: 0.8% Pd/HZSM-5 (hereafter, Pd/HZSM-5), (0.8% Pd + 1.0% Ni)/HZSM-5 (Pd–Ni/HZSM-5), and (0.8% Pd + 1.0% Co)/HZSM-5 (Pd–Co/HZSM-5). The experiments were conducted under the following conditions: n-hexane partial pressure, 66 hPa; H2 : n-hexane molar ratio, 14.2; feed flow rate, 6 L/h; temperature, 250°C; catalyst weight, 1.0 g. Hydrogen and n-hexane adsorption and desorption experiments and experiments on their mutual displacement were carried out at 150°C. RESULTS AND DISCUSSION The X-ray diffraction pattern of the monometallic palladium catalyst and those of the Ni- or Сo-containing bimetallic samples were similar and showed only peaks characteristic of HZSM-5 at 2θ = 7.9°, 8.8°, 23.1°, 24.1°, and 29.1°. The crystalline phases of metal oxides were not detected because of their low concentration or high dispersion. Table 1 presents the specific surface area (SBET), palladium particle size (d), and acidity data for the palladium catalysts. As follows from these data, the introduction of Ni or Co into the palladium catalyst causes no significant changes in its specific surface area but considerably decreases the size of the metalcontaining clusters relative to that in the monometallic palladium catalyst; that is, a geometric effect is manifested here. The surface concentration of mediumstrength and strong acid sites (characterized by an ammonia desorption temperature of >280°C) in HZSM-5 was 47.9 mmol NH3/100 g. As was demonstrated by preliminary experiments, weak acid sites are difficult to quantify with a high accuracy, because in this case the amount of ammonia desorbed depends strongly on the pre-purging time. Moreover, the desorption peaks characterizing the weak acid sites are very large, making it difficult to characterize the medium-strength and strong ones. This is the reason why we present the total concentration of mediumstrength and strong acid sites (Table 1). The introduction of Pd increases this acidity, while the subsequent introduction of Ni or Со causes a decrease in acidity, KINETICS AND CATALYSIS
Vol. 58
No. 3
2017
313
Table 1. Specific surface area, palladium particle size, and acidity data for the palladium catalysts Catalyst Pd/HZSM-5 Pd–Ni/HZSM-5 Pd–Co/HZSM-5
SBET, m2/g 349 323 319
d, nm
Medium-strength + strong acidity, mmol NH3/100 g
7.4 5.1 4.6
66.5 57.1 54.9
and this is accompanied by a marked decrease in the size of palladium clusters. According to the literature, the introduction of nickel and cobalt into Pt and Pd catalysts supported on various aluminosilicates exerts different effects on their acidity. It was demonstrated [18] that the introduction of 5% Ni into Pt- and Pdcontaining samples eliminates the strong acidity of the aluminosilicate support. It was discovered [19] that the introduction of nickel into a Pd catalyst supported on zeolite Y (SiO2/Al2O3 = 6, 0.1% Pd, 0–0.5% Ni) increases the ammonia binding strength, decreasing the total and strong acidities. Thus, the introduction of Ni or Co into Pd catalysts decreases their acidity and the size of active phase clusters therein. Figure 1 shows the Н2-TPR profiles for the palladium catalysts and Ni/HZSM-5 and Co/HZSM-5 samples. The Н2-TPR profile for the monometallic catalyst Pd/HZSM-5 (curve 1) shows a single peak arising from the decomposition of palladium hydride. Co/HZSM-5 (curve 4) is characterized by a single peak due to Со3О4 reduction at 335°C [20]. The Н2-TPR profile for Ni/HZSM-5 (curve 2) shows two peaks at 350 and 490°C, indicating that reduction occurs in two steps. Similar conclusions were drawn by Lima Hydrogen uptake, arb. units 8 5 4
7 6
3 5 2
4 3
1
2 1 0
100 200 300 400 500 600 700 800 T, °C
Fig. 1. Н2-TPR profiles for (1) Pd/HZSM-5, (2) Ni/HZSM-5, (3) Pd–Ni/HZSM-5, (4) Со/HZSM-5, and (5) Pd–Со/HZSM-5.
314
LUU CAM LOC et al.
Table 2. n-Hexane conversion (Х), isomer selectivity (S), total isohexane yield (Y), and on-stream stability (τ) data for the palladium catalysts Catalyst Pd/HZSM-5 Pd–Co/HZSM-5 Pd–Ni/HZSM-5
X
S
Y
τ, h
0.57 0.60 0.60
0.86 0.93 0.90
0.49 0.56 0.54
4.0 >30 >30
Reaction conditions: Т = 250°C, Pn0-hexane = 77.0 hPa, PH0 2 = 616 hPa.
Table 3. Effect of the granule size of the Pd–Co/HZSM-5 catalyst on the n-hexane isomerization rate and proportions of isohexane isomers d, mm 0.25–0.60 0.60–1.00 1.00–1.41 1.41–2.00 Х 0.63 0.63 0.60 0.58 2,3-DMB : 2-MP : 3-MP* 1 : 62 : 36 1 : 62 : 36 1 : 69 : 39 1 : 73 : 45 * 2,3-DMB = 2,3-dimethylbutane, 2-MP = 2-methylpentane, and 3-MP = 3-methylpentane. Reaction conditions: Т = 250°C, Pn0-hexane = 77.0 hPa, PH0 2 = 616 hPa.
et al. [21] from TPR data for the Ni/H-BEA catalyst. The Н2-TPR profile contained two peaks: the first peak (445°C) characterized the reduction of Ni2+ ions in the tetrahedral anion ( AlO 4− ) framework, and the second peak (555°C) was due to the reduction of Ni2+ cations bound to the siloxy anion [SiO 4−]. It is likely that the peak at 400°C in the Н2-TPR profile of Pd– Ni/HZSM-5 (Fig. 1, curve 3) is due to the reduction of Ni2+ ions in the tetrahedral anion ( AlO 4− ) framework. The presence of Pd shifts the reduction peaks of Co and, to a lesser extent, Ni to the left; therefore, Pd facilitates the reduction of these metals, and this is accompanied by an increase in peak intensity (Fig. 1, curves 3, 5). In the bimetallic catalysts, electronic interaction takes place between the metals to shift d electrons of Pd to the second metal (Ni or Co), and this effect can somewhat change the Н2-TPR peak temperatures. Table 2 lists n-hexane conversion (Х), selectivity (S), total isomer (2-methylpentane (2-MP), 3-methylpentane (3-MP), and 2,3-dimethylbutane (2,3-DMB), yield (Y), and on-stream stability (τ) data for the palladium catalysts Clearly, the Pd–Co/HZSM-5 catalyst, which has a lower surface concentration of medium-strength and strong acid sites than the monometallic Pd catalyst or the Ni-promoted sample, affords the highest isomer yield and isomer selectivity. The n-hexane isomerization kinetics was studied for the promoted palladium catalysts, which operated stably for a sufficiently long time. At a fixed gas hourly space velocity and initial n-hexane and hydrogen concentrations, the rate of n-hexane isomerization over Pd–Co/HZSM-5 at 260°C was independent of the circulation rate; that is, there were no external diffusion limitations on the reaction rate. Table 3 presents data concerning the
effect of the Pd–Co/HZSM-5 granule size on the n-hexane isomerization rate and isohexane isomer distribution. Since the rate of the reaction and product composition did not change on passing from the 0.25– 0.6 mm granules to the 0.6–1.0 mm ones, it could be assumed that there were no internal diffusion limitations with
