- Email: [email protected]
Spectroscopic studies on NiO supported on ZrO2 modified with MoO3 for ethylene dimerization
Applied Catalysis A: General 317 (2007) 216–225 www.elsevier.com/locate/apcata
Spectroscopic studies on NiO supported on ZrO2 modified with MoO3 for ethylene dimerization Jong Rack Sohn *, Si Hyon Kwon, Dong Cheol Shin Department of Applied Chemistry, Engineering College, Kyungpook National University, Taegu 702-701, Republic of Korea Received 23 May 2006; received in revised form 28 September 2006; accepted 10 October 2006 Available online 17 November 2006
Abstract NiO supported on zirconia modified with MoO3 for ethylene dimerization was prepared by drying powdered Ni(OH)2–Zr(OH)4 with ammonium heptamolybdate aqueous solution, followed by calcining in air at high temperature. The characterization of prepared catalysts was performed using FTIR, Raman, XRD, and DSC. MoO3 equal to or less than 15 wt% was dispersed on the surface of catalyst as two-dimensional polymolybdate or monomolybdate, while for MoO3 above 15 wt%, a crystalline orthorhombic phase of MoO3 was formed, showing that the critical dispersion capacity of MoO3 on the surface of catalyst is 0.18 g/g NiO–ZrO2 on the basis of XRD analysis. From 600 8C of calcination temperature, zirconium molybdenum oxide, Zr(MoO4)2 was formed due to the reaction between ZrO2 and MoO3; its amount increased with the calcination temperature. The acid amount and the catalytic activity of catalysts increased with the amount of dispersed MoO3. The high acid strength and acid amount was responsible for the Mo O bond nature of the complex formed by the interaction between MoO3 and ZrO2. The catalytic activity for ethylene dimerization was correlated with the acid amount of the catalysts measured by the ammonia chemisorption method. # 2006 Published by Elsevier B.V. Keywords: NiO–ZrO2/MoO3 catalyst; Ethylene dimerization; Acidic properties; MoO3 dispersion; Modification with MoO3
1. Introduction Nickel oxide on silica or silica–alumina is an effective catalyst in olefin dimerization and isomerization as well as in the hydrocracking, hydrogenation and methanation process following reduction of the nickel component [1–3]. The dimerization of alkenes is an important method for the production of higher olefins, which are extensively used as industrial intermediates. A considerable number of papers have dealt with the nickelcontaining catalysts for ethylene dimerization [1–12]. One of the remarkable features of this catalyst system is its activity in relation to a series of n-olefins. In contrast to the usual acid-type catalysts, nickel oxide on silica or silica–alumina shows a higher activity for a lower olefin dimerization, particularly for ethylene dimerization [1–4,13]. It has been suggested that the active site for dimerization is formed by an interaction of a low-valent nickel ion with an acid site [10,14]. The dimerization activities of such catalysts are reported to be related to the acidic properties of
* Corresponding author. Tel.: +82 53 950 5585; fax: +82 53 950 6594. E-mail address: [email protected] (J.R. Sohn). 0926-860X/$ – see front matter # 2006 Published by Elsevier B.V. doi:10.1016/j.apcata.2006.10.015
surface and low-valent nickel ions. In fact, nickel oxide, which is active for C2H4–C2D4 equilibration, acquires an activity for ethylene dimerization upon addition of nickel sulfate, which is known to be an acid [15]. A transition metal can also be supported on zeolite in the state of a cation or as a finely dispersed metal. Transition metal ions like Ni+ or Pd+ can be active sites in catalytic reactions such as ethylene and propylene dimerization as well as in acetylene cyclomerization [16–18]. The previous papers from this laboratory have shown that NiO–TiO2 and NiO–ZrO2 modified with sulfate are very active for ethylene dimerization [19–21]. High catalytic activities in the reactions were attributed to the enhanced acidic properties of the modified catalysts, which originated from the inductive effect of S O of the complex formed by the interaction of oxides with sulfate. Zirconium oxide is a very interesting material because of its thermal stability, its mechanical properties and its basic, acidic, reducing, and oxidizing surface properties. Therefore, it is attracting interest in catalysis as a support material as well as a catalyst. The potential for a heterogeneous catalyst has yielded many papers on the catalytic activity of sulfated zirconia materials [22–26]. Moreover, catalytic functions have been improved by loading additional components. Sulfated zirconia
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
incorporating Fe and Mn has been shown to be highly active for butane isomerization, catalyzing the reaction even at room temperature [27,28]. The promotion of activity of the catalyst has been confirmed by several other research groups [29–31]. Coelho et al. have discovered that the addition of Ni to sulfated zirconia causes an activity enhancement comparable to that caused by the addition of Fe and Mn [32]. It has been reported by several workers that the addition of platinum to zirconia modified by sulfate ions enhances the catalytic activity in the skeletal isomerization of alkanes without deactivation when the reaction is carried out in the presence of hydrogen [33–35]. Recently, it has been found that a main group element Al can also promote the catalytic activity and stability of sulfated zirconia in ethylene dimerization and n-butane isomerization [36–38]. Recently, some workers reported zirconia-supported WO3 or MoO3 as alternative materials in reactions requiring strong acid sites [26,39–41]. Several advantages of tungstate or molybdate over sulfate as dopant include the facts that they do not suffer from dopant loss during thermal treatment and that they undergo significantly less deactivation during catalytic reaction [41]. As an extension of the study on the ethylene dimerization, these authors tried to prepare other catalyst systems by combining nickel hydroxide to give low-valent nickel after decomposition with ZrO2 modified by MoO3 which is known to be an acid [40,42], to improve the catalytic activity for ethylene dimerization and acidic properties. In this paper, spectroscopic studies on NiO–ZrO2/MoO3 catalysts and catalytic activity for ethylene dimerization are reported.
217
290 Torr. A fresh catalyst sample of 0.2 g was used for every run and the catalytic activity was calculated as the number of moles of ethylene consumed during the initial 5 min. Reaction products were analyzed by gas chromatography with a VZ-7 column at room temperature. FTIR spectra were obtained in a heatable gas cell at room temperature using a Mattson Model GL6030E spectrophometer. The self-supporting catalyst wafers contained about 9 mg/cm2. Prior to obtaining the spectra we heated the samples under vacuum at 500 8C for 1.5 h. The Raman spectra were recorded on a Spex Ramalog ˚ line from spectrometer with holographic gratings. The 5145-A a Spectra-Physics Model 165 argon-ion laser was used as the exciting source. The spectral shift width was typically 4 cm 1, and laser source powers of approximately 45 mW, measured at the sample, were used. Catalysts were checked in order to determine the structure of the support as well as that of molybdenum oxide by means of a Jeol Model JDX-8030 diffractometer, employing Cu Ka (Ni-filtered) radiation. DSC measurements were performed by a PL-STA model 1500H apparatus in air, and the heating rate was 5 8C per minute. For each experiment, 10–15 mg of sample was used. The specific surface area was determined by applying the BET method to the adsorption of N2 at 196 8C. Chemisorption of ammonia was also employed as a measure of the acid amount of each catalyst. The amount of chemisorption was determined based on the irreversible adsorption of ammonia [43–45].
2. Experimental 2.1. Catalyst preparation
3. Results and discussion
The coprecipitate of Ni(OH)2–Zr(OH)4 was obtained by adding aqueous ammonia slowly into a mixed aqueous solution of nickel chloride and zirconium oxychloride at room temperature with stirring until the pH of the mother liquor reached about 8. The coprecipitate thus obtained was washed thoroughly with distilled water until chloride ion was not detected, and then was dried at 100 8C for 12 h. The dried precipitate was powdered below 100 mesh. The catalysts containing various MoO3 contents were prepared by adding an aqueous solution of ammonium heptamolybdate[(NH4)6(Mo7O24)4H2O] to the Ni(OH)2– Zr(OH)4 powder, followed by drying and calcining at high temperatures for 1.5 h in air. This series of catalysts are denoted by their weight percentage of NiO and MoO3. For example, 10NiO–ZrO2/15-MoO3 indicates the catalyst containing 10 wt% NiO regarding only two components, NiO and ZrO2, and containing 15 wt% MoO3 regarding all three components, NiO, ZrO2 and MoO3.
3.1. Crystalline structures of catalysts
2.2. Procedure The catalytic activity for ethylene dimerization was determined at 20 8C using a conventional static system following a pressure change from an initial pressure of
The crystalline structures of ZrO2 and NiO–ZrO2/MoO3 catalysts calcined in air at different temperatures for 1.5 h were examined. For pure ZrO2, ZrO2 was amorphous to X-ray diffraction up to 300 8C, with a two-phase mixture of the tetragonal and monoclinic forms at 400–700 8C and a monoclinic form at 800 8C (this figure is not shown here). Three crystal structures of ZrO2: tetragonal, monoclinic and cubic phases, have been reported [46,47]. However, in the case of NiO–ZrO2 modified with MoO3, the crystalline structures of the catalysts were different from that of pure ZrO2. For the 10NiO–ZrO2/15-MoO3 catalysts, as shown in Fig. 1, ZrO2 was amorphous to X-ray diffraction up to 500 8C with tetragonal ZrO2 formed at 600–800 8C. The transition temperature of ZrO2 from amorphous to tetragonal phase was higher by 200 8C than of pure ZrO2. However, the higher the calcination temperature, the larger the amount of tetragonal phase of ZrO2. It is assumed that the interaction between MoO3 (or NiO) and ZrO2 hinders the transition of ZrO2 from amorphous to tetragonal phase [48,49]. The presence of MoO3 strongly influences the development of textural properties with temperature. No phase of MoO3 or NiO was observed at any calcination temperature, indicating a good dispersion of MoO3
218
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
Fig. 1. X-ray diffraction patterns of 10-NiO–ZrO2/15-MoO3 catalysts calcined at different temperatures for 1.5 h: (*) tetragonal ZrO2 and () orthorhombic MoO3.
or NiO on the surfaces of ZrO2 supports due to the interaction between them. XRD patterns of 10-NiO–ZrO2/25-MoO3 catalysts calcined at 400–800 8C for 1.5 h are shown in Fig. 2. X-ray diffraction data indicated the amorphous phase of ZrO2 at 400 8C, showing the tetragonal phase of ZrO2 at 450–800 8C. However, the higher the calcined temperature, the larger the amount of tetragonal phase of ZrO2. Moreover, as shown in Fig. 2, the orthorhombic phase of MoO3 was observed at 400–500 8C. These results are in good agreement with those of Raman (Fig. 4) and IR (Fig. 6) described later. It has been reported that MoO3 for catalysts containing low MoO3 loading is dispersed on the surface of zirconia as a monolayer state, that is, two-dimensional polymolybdate or monomolybdate, which gives rise to the characteristic broad band at 900–1000 cm 1 [50,51]. However, as mentioned above, when MoO3 loading exceeds a monolayer capacity on the surface of NiO–ZrO2, the residual crystalline phase of MoO3 can be detected by XRD, Raman, and IR. From the calcination temperature of 600 8C, a new species, i.e., Zr(MoO4)2 hexagonal phase, has been detected in the 10-NiO–ZrO2/25-MoO3 catalyst. As shown in Fig. 2, the amount of Zr(MoO4)2 phase increased with the calcination temperature. Chen et al. reported the formation of Zr(MoO4)2 from MoO3 and ZrO2 in high MoO3 catalysts calcined above 600 8C [52].
Fig. 2. X-ray diffraction patterns of 10-NiO–ZrO2/25-MoO3 catalysts calcined at different temperatures for 1.5 h: (*) tetragonal ZrO2; (~) monoclinic ZrO2; () orthorhombic MoO3; (&) hexagonal Mo(ZrO4)2.
The XRD patterns of 10-NiO–ZrO2/MoO3 catalysts containing different MoO3 contents and calcined at 450 8C for 1.5 h are shown in Fig. 3. XRD data indicated only tetragonal phase of ZrO2 at 5 wt% of MoO3. However, 10-NiO–ZrO2/10-MoO3 and 10-NiO–ZrO2/15-MoO3 catalysts were amorphous to X-ray diffraction, indicating that the increase of MoO3 loading hinders the phase transition of ZrO2 from amorphous to tetragonal, due to the interaction between MoO3 and ZrO2 [38,49]. In this case, no detection of orthorhombic phase of MoO3 means that all MoO3 is highly dispersed on the surface of NiO–ZrO2 as two-dimensional polymolybdate or monomolybdate, showing that these results are in good agreement with those of Raman (Fig. 4) and IR (Fig. 6). For the catalysts containing higher MoO3 loading than 15 wt%, as shown in Fig. 3, the orthorhombic phase of MoO3 was observed because MoO3 loading exceeds a monolayer capacity on the surface of ZrO2, showing that its intensity increases with the MoO3 content. When the loading amounts of MoO3 are low, the highly dispersed MoO3 species can not be detected by XRD, as shown in Fig. 3. Therefore, we can estimate the amount of crystalline MoO3 by the quantitative XRD analysis. So, we can plot the content of crystalline MoO3 as a function of the total content of
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
219
MoO3 in the 10-NiO–ZrO2/MoO3 catalysts containing different MoO3 contents and calcined at 450 8C (this figure is not shown here), from which the MoO3 critical dispersion capacity on NiO–ZrO2, 0.18 g MoO3/g NiO–ZrO2 is deduced. Since the surface area of 10-NiO–ZrO2/15-MoO3 catalyst is 154 m2/g, this capacity is equivalent to 0.12 g MoO3/100 m2 NiO–ZrO2, which is very similar with the monolayer capacity for MoO3 supported on ZrO2 [51]. It is expected that this dispersion capacity will be related to the surface area and acid amount of catalysts, as discussed below. 3.2. Raman and infrared spectra
Fig. 3. X-ray diffraction patterns of 10-NiO–ZrO2/MoO3 catalysts containing different MoO3 contents and calcined at 450 8C for 1.5 h: (*) tetragonal ZrO2 and () orthorhombic MoO3.
Fig. 4. Raman spectra of 10-NiO–ZrO2/MoO3 catalysts containing different MoO3 contents.
Raman spectroscopy is a valuable tool for the characterization of dispersed metal oxides and can detect vibrational modes of surface and bulk structures. The Raman spectra of MoO3 obtained by calcining ammonium heptamolybdate at 450 8C and 10-NiO–ZrO2/MoO3 catalysts containing different MoO3 contents under ambient conditions are presented in Fig. 4. The MoO3 structure is made up of distorted MoO3 octahedra. Bulk crystalline MoO3 shows the main bands, in good agreement with data previously reported [53]. The major vibrational modes of MoO3 are located at 995, 819, 667, and 290 cm 1, and have been assigned to the Mo O stretching mode, the Mo–O– Mo stretching mode, OMo3 vibration mode, and the Mo O deformation mode, respectively [53]. The molecular structure of the supported MoO3 species depends on the loading. Several authors observed that the nature of surface molybdenum species on SiO2, Al2O3, TiO2 and ZrO2 depends on the amount of MoO3. A band at about 825 cm 1, corresponding to Mo–O–Mo vibrations [50,54], and a band at 910–980 cm 1, assigned to Mo O vibrations in twodimensional polymolybdate, were observed in these samples [50,54]. Raman bands between 910 and 980 cm 1 are usually attributed to the Mo O vibration of Mo species in either octahedral or tetrahedral environment [55]. Generally, monomolybdate or tetrahedral molybdenum oxygen species have been assigned for low MoO3 loading catalysts [50,55,56], and two-dimensional polymolybdate or octahedral molybdenum– oxygen species with characteristic band around 950–980 cm 1, have been assigned for high MoO3 loading catalysts [50,55,56]. Therefore, the broad band observed in the 900–1000 cm 1 region in Fig. 4 will be interpreted as an overlap of two characteristic bands for two molybdenum oxide species. The bands around 350–370 cm 1 for 10-NiO–ZrO2/15-MoO3 and 10-NiO–ZrO2/10-MoO3 catalysts are assigned to Mo O bending mode [57]. For 10-NiO–ZrO2/10-MoO3 and 10-NiO–ZrO2/15-MoO3 catalysts in Fig. 4, most of the zirconia is amorphous to X-ray diffraction, showing tiny amount of tetragonal phase zirconia for the other. The Raman spectrum of amorphous zirconia is characterized by a very weak and broad band at 550–600 cm 1 [58]. Tetragonal zirconia is expected to yield a spectrum consisting of Raman bands with frequencies at about 148, 263, 325, 472, 608, and 640 cm 1, while monoclinic zirconia exhibits the characteristic features at 180, 188, 221, 380, 476, and 637 cm 1 [59,60]. As shown in Fig. 4, the 10-NiO–ZrO2/5-
220
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
Fig. 5. Raman spectra of 10-NiO–ZrO2/25-MoO3 catalysts calcined at different temperatures for 1.5 h.
MoO3 catalyst exhibits the characteristic features of tetragonal zirconia, indicating no transformation of ZrO2 from tetragonal to monoclinic. Fig. 5 shows Raman spectra of 10-NiO–ZrO2/25-MoO3 catalysts calcined at 400–800 8C for 1.5 h. For catalysts calcined at 400–500 8C, the vibrational modes of bulk MoO3 were observed at 995, 819, 667 and 290 cm 1 similarly to those of Fig. 4. All catalysts calcined at 450–800 8C except that at 400 8C also exhibit the characteristic features of tetragonal zirconia at 265, 325, 472, 608, and 640 cm 1, in good agreement with XRD data described above. However, from the calcination temperature of 600 8C, a new species, i.e., Zr(MoO4)2 phase, identified by its characteristic Raman bands at 748, 945, and 1000 cm 1 [50], has been detected in the 10NiO–ZrO2/25-MoO3 catalyst. As shown in Fig. 5, the amount of Zr(MoO4)2 phase increased with the calcination temperature. Chen et al. [52] reported the formation of Zr(MoO4)2 between MoO3 and ZrO2 in high loading catalysts calcined above 600 8C. For 10-NiO–ZrO2/15-MoO3 catalysts calcined at 400– 500 8C, the broad bands in the 900–1000 cm 1 region were observed by an overlap of two characteristic bands for two molybdenum oxide species (this figure is not shown here). The bands around 370 cm 1 assigned to Mo O banding mode were also observed. However, all 10-NiO–ZrO2/15-MoO3 catalysts calcined at 600–800 8C exhibit the characteristic features of tetragonal zirconia at 265, 325, 472, 608, and 640 cm 1, in good agreement with XRD data as seen in Fig. 1. Fig. 6 shows IR spectra of 10-NiO–ZrO2/MoO3 catalysts (KBr disc) with different MoO3 contents calcined at 450 8C for 1.5 h. Although with samples having less than 20 wt% of
Fig. 6. Infrared spectra of 10-NiO–ZrO2/MoO3 catalysts calcined at 450 8C for 1 h.
MoO3 no definite peaks were observed, the absorption bands at 994, 862, and 599 cm 1 appeared for 10-NiO–ZrO2/MoO3 catalysts having more than 15 wt% of MoO3. These bands are due to the crystalline MoO3. When MoO3 loading exceeds a certain value, the surplus MoO3 is present as crystalline form [50,51,61]. In view of Raman spectra of crystalline MoO3 in Fig. 4, for crystalline MoO3 the IR bands at 991, 870, and 596 cm 1 in Fig. 6 are assigned to the Mo O stretching mode, the Mo–O–Mo stretching mode, and OMo3 vibration mode, respectively. XRD measurements have also supported the presence of the above crystalline MoO3, as discussed above. The Raman and IR spectra in Figs. 4–6 have been taken in air using the pure powders. To examine the structure of NiO–ZrO2 modified with MoO3 under dehydration conditions, we obtained IR spectra of 10-NiO–ZrO2/15-MoO3 catalyst in a heatable gas cell after evacuation at 25 and 500 8C for 1.5 h (this figure is not shown here). For the 10-NiO–ZrO2/15-MoO3 catalyst, IR band at 997 cm 1 after evacuation at 500 8C is due to the Mo O stretching mode of the molybdenum oxide complex bonded to the ZrO2 surface [62]. This Mo O band due to molybdenyl species only appears on the evacuated catalysts, being shifted on the wet catalysts. The broad band at 944 cm 1 under ambient condition is assigned to the terminal Mo O stretching of the hydrated form of the surface molybdenyl species [62]. This can be rationalized by assuming that the adsorption of water causes a strong perturbation of the corresponding molybdenum oxide species, with a consequent strong broadening and down shift of this band. This isolated molybdenum oxide species is stabilized through multiple Mo– O–Zr bonds between each molybdenum oxide species and the
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
zirconia surface [50,52]. Upon dehydration at elevated temperature, the hydrated surface metal oxide species is unstable and decomposes to form dehydrated surface metal oxide species by direct interaction with the surface OH groups of support, giving the formation of metal–oxygen-support bond. 3.3. Thermal analysis In X-ray diffraction patterns, it was shown that the structures of NiO–ZrO2/MoO3 catalysts were different depending on the calcined temperature. To examine the thermal properties for the precusors of samples more clearly, we carried out their thermal analysis as illustrated in Fig. 7. For pure ZrO2 (Fig. 7(h)) the DSC curve showed a broad endothermic peak below 180 8C due to the elimination of adsorbed water, and an exothermic peak at 418 8C due to the ZrO2 crystallization [49,59]. For (NH4)6(Mo7O24)4H2O (Fig. 7(i)), the DSC curve shows three endothermic peaks below 400 8C, indicating that the decomposition of ammonium heptamolybdate occurs in three steps. In the case of NiO–ZrO2/MoO3 precursors, three additional endothermic peaks appeared in the region of 200–400 8C due to the evolution of NH3 and H2O decomposed from ammonium heptamolybdate. Also it is considered that an endothermic at 796 8C is responsible for melting of MoO3 [61] and for 10NiO–ZrO2/25-MoO3 and ZrO2/15-MoO3 precursors (Fig. 7(a) and (b)) an exothermic peak around 610 8C is due to the formation of Zr(MoO4)2 as described in X-ray diffraction patterns. However, it is of interest to see the influences of NiO and MoO3 on the crystallization of ZrO2 from amorphous to
221
tetragonal phase. As Fig. 7 shows, the exothermic peak due to the crystallization appears at 418 8C for pure ZrO2, while for NiO–ZrO2 precursors (Fig. 7(c)–(g)) and NiO–ZrO2/MoO3 precursor (Fig. 7(a)), it was shifted to higher temperatures. The shift increased with increasing NiO content up to 25 wt% of NiO. It is considered that the interaction between NiO (or MoO3) and ZrO2 delays the transition of ZrO2 from amorphous to tetragonal phase [38]. Consequently, the exothermic peaks appear at 429 8C for 3-NiO–ZrO2, 446 8C for 5-NiO–ZrO2, 493 8C for 10-NiO–ZrO2, 545 8C for ZrO2/15-MoO3, 551 8C for 10-NiO–ZrO2/25-MoO3, 562 8C for 17-NiO–ZrO2, and 575 8C for 25-NiO–ZrO2. 3.4. Surface properties It is necessary to examine the effects of molybdenum oxide on the surface properties of catalysts, that is, specific surface area, acid amount, and acid strength. The specific surface areas of 10-NiO–ZrO2/MoO3 catalysts calcined at 450 8C for 1.5 h are listed in Table 1. The presence of molybdenum oxide strongly influences the surface area in comparison with the pure ZrO2. Specific surface areas of 10-NiO–ZrO2/MoO3 catalysts are much larger than that of pure ZrO2 calcined at the same temperature, showing that surface area increases gradually with increasing MoO3 content up to 15 wt% of MoO3. The decrease of surface area for samples containing MoO3 above 15 wt% is due to the blocking of ZrO2 pore by the increased MoO3 loading. It seems likely that the interaction between MoO3 and ZrO2 protects catalysts from sintering. The acid amount of catalysts calcined at 450 8C, as determined by the amount of NH3 irreversibly adsorbed at 230 8C [5,8,38], is also listed in Table 1. The acid amount per surface area for these catalysts was calculated to be nearly constant values of 0.9–1.0 mmol/ m2, indicating that the variation of acid amount runs parallel to the change of surface area. We examined the effect of NiO addition on the surface area and acid amount of NiO–ZrO2/15MoO3 catalysts. The specific surface areas and acid amount of catalysts calcined at 450 8C are listed in Table 2. Both surface area and acid amount increased with increasing NiO content up to 10 wt%, indicating the promoting effect of NiO. The acid strength of the catalysts was examined by a color change method, using Hammett indicators in sulfurylchloride [38,40,49]. NiO–ZrO2/MoO3 catalyst was estimated to have a H0 14.5, indicating the formation of new acid sites stronger than those of single oxide components. Acids stronger than H0 11.93, which corresponds to the acid strength of 100% Table 1 Specific surface area and acid amount of 10-NiO–ZrO2/MoO3 catalysts containing different MoO3 contents and calcined at 450 8C for 1.5 h
Fig. 7. DSC curves of catalyst precursors: (a) 10-NiO–ZrO2/25-MoO3, (b) ZrO2/15-MoO3, (c) 25-NiO–ZrO2, (d) 17-NiO–ZrO2, (e) 10-NiO–ZrO2, (f) 5NiO–ZrO2, (g) 3-NiO–ZrO2, (h) ZrO2, and (i) (NH4)6(Mo7O24)4H2O.
MoO3 content (wt%)
Surface area (m2/g)
Acid amount (mmol/g)
5 10 15 20 25
128 140 155 138 124
112 134 141 129 121
222
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
Table 2 Specific surface area and acid amount of NiO–ZrO2/15-MoO3 catalysts containing different NiO contents and calcined at 450 8C for 1.5 h
contents also showed the presence of both Lewis and Bro¨nsted acid sites.
NiO content (wt%)
Surface area (m2/g)
Acid amount (mmol/g)
3.5. Ethylene dimerization over NiO–ZrO2/MoO3
3 5 10 15 20 30 40
139 143 155 138 128 121 89
78 119 141 137 121 95 78
H2SO4, are superacids [26,63,64]. Consequently, NiO–ZrO2/ MoO3 catalysts would be solid superacids. The superacidic property is attributed to the double bond nature of the Mo O in the complex formed by the interaction of ZrO2 with molybdate, in analogy with the cases of ZrO2 modified with tungstate, chromate, and sulfate ions [48,49,65]. Infrared spectroscopic studies of ammonia adsorbed on solid surface have made it possible to distinguish between Bro¨nsted and Lewis acid sites [5,20,38,40]. Fig. 8 shows the IR spectra of ammonia adsorbed on 10-NiO–ZrO2/15-MoO3 catalyst evacuated at 400 8C for 1 h. The band at 1448 is the characteristic peak of ammonium ion, which is formed on the Bro¨nsted acid sites and the absorption peak at 1609 cm 1 is contributed by ammonia coordinately bonded to Lewis acid sites [5,20,38,40], indicating the presence of both Bro¨nsted and Lewis acid sites on the surface of catalyst. Other samples having different MoO3
NiO–ZrO2/MoO3 catalysts were tested for their effectiveness in ethylene dimerization. It was found that, over 10-NiO– ZrO2/15-MoO3, 10-NiO–ZrO2/10-MoO3, and 3-NiO–ZrO2/ 15-MoO3 catalysts, ethylene was continuously consumed, as shown by the results presented in Fig. 9, where catalysts were evacuated at 450 8C for 1 h. Over NiO–ZrO2/MoO3, ethylene was selectively dimerized to n-butenes. In the composition of n-butenes analyzed by gas chromatography, 1-butene was found exclusively at the initial reaction time, with no cisbutene or trans-butene. However, the amount of 1-butene decreases with the reaction time, while the amount of 2-butene increases. Therefore, it seems likely that the initially produced 1-butene is also isomerized to 2-butene during the reaction time [2,5,20]. NiO–ZrO2/MoO3 catalyst was effective for ethylene dimerization, but as shown in Fig. 9, NiO–ZrO2 without MoO3 did not exhibit any catalytic activity because the sample without MoO3 does not possess enough acid amount or acid strength to catalyze ethylene dimerization. Therefore, it is expected that the amount of dispersed MoO3 will be related to the acid amount and catalytic activity of catalyst, as discussed below. ZrO2/15-MoO3 alone without NiO was also totally inactive for ethylene dimerization, as shown in Fig. 9. It is known that ZrO2 modified with MoO3 is an acid [26,40].
Fig. 8. Infrared spectra of NH3 adsorbed on 10-NiO–ZrO2/15-MoO3 catalyst: (a) background of 10-NiO–ZrO2/15-MoO3 catalyst after evacuation at 400 8C for 1 h and (b) NH3 adsorbed on (a); where gas was evacuated at 230 8C for 1 h.
Fig. 9. Time-course of ethylene dimerization over catalysts evacuated at 450 8C for 1 h: (*) 10-NiO–ZrO2/15-MoO3; (&) 10-NiO–ZrO2/10-MoO3; (~) 3NiO–ZrO2/15-MoO3; (*) ZrO2/15-MoO3; (&) 10-NiO–ZrO2.
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
The active site responsible for dimerization is suggested to consist of a low-valent nickel ion and an acid, as observed in the nickel-containing catalyst [2,19–21]. The term ‘low-valent nickel’ originated from the fact that the NiO–SiO2 catalyst was drastically poisoned by carbon monoxide, since a low-valent nickel is favorable to chemisorb carbon monoxide [66,67]. Previously, from the results of IR of CO adsorbed on NiSO4/gAl2O3 and XPS to identify low-valent nickel species, we reported that the active sites responsible for ethylene dimerization consist of a low-valent nickel, Ni+, and an acid [67]. Kazansky and coworkers concluded that Ni+ ions in NiCaY zeolite are responsible for the catalytic activity; they used EPR spectroscopy to identify low-valent nickel species [17]. Therefore, the nickel species should be active sites for ethylene dimerization. In fact, as shown in Fig. 9, ZrO2/MoO3 without NiO or NiO–ZrO2 without MoO3 is inactive for ethylene dimerization. The catalytic activities of 10-NiO–ZrO2/15-MoO3 catalysts were tested as a function of calcination temperature. The activities increased with the calcination temperature, reaching a maximum at 450 8C, after which the activities decreased. The decrease of catalytic activity after calcination above 450 8C seems to be due to the decrease of surface area and acid amount at high calcination temperatures. In fact, the surface area of 10NiO–ZrO2/15-MoO3 catalyst calcined at 450 8C was found to be 154 m2/g, but that of the catalyst calcined at 600 8C decreased to 109 m2/g. The acid amount values of the catalysts calcined at 450 and 600 8C were found to be 140 and 110 mmol/ g, respectively. Thus, hereafter, emphasis is placed only on the catalysts calcined at 450 8C. 3.6. Effects of dispersed MoO3 amount on acid amount and catalytic activity The forms of active components present in heterogeneous catalysts are of importance to catalysis. A great many oxides can disperse spontaneously onto the surfaces of supports to form a monolayer, because the monolayer is a thermodynamically stable form [61]. Dispersed MoO3 amount, surface area, acid amount, and catalytic activity for 10-NiO–ZrO2/MoO3 catalysts containing MoO3 content below 20 wt% are listed in Table 3. There are good correlations among the dispersed MoO3 amount, acid amount, and catalytic activity. Namely, the larger the dispersed MoO3 amount, the higher both the acid amount and the catalytic activity are. This can be explained by the suggestion that strong acid sites are formed through the bonding between dispersed MoO3 and ZrO2 and consequently catalytic activity increases due to the increased acid sites.
223
Fig. 10. Catalytic activity of 10-NiO–ZrO2/MoO3 catalysts for ethylene dimerization as a function of MoO3 content.
The catalytic activity as a function of MoO3 content is plotted in Fig. 10. As shown in Fig. 10, the maximum activity is obtained with the catalyst of 15 wt% MoO3, where the amount of dispersed MoO3 is maximum. It is known that the active sites responsible for dimerization consist of a low-valent nickel ion and an acid [2,19–21]. Therefore, it seems likely that the highest activity of the catalyst containing 15 wt% MoO3 is related to its acid amount and acid strength. The high acid strength and acid amount are responsible for the Mo O bond nature of complex formed by the interaction between MoO3 and ZrO2 [26,40]. As discussed in IR spectra, IR spectra of 10-NiO– ZrO2/15-MoO3 catalyst after evacuation at 500 8C showed the band at 997 cm 1 due to the Mo O stretching mode of the molybdenum oxide complex bonded to the ZrO2 surface [62]. This isolated molybdenum oxide species is stabilized through multiple Mo–O–Zr bonds between each MoO3 species and the ZrO2 surface [50,52]. As listed in Tables 1 and 3, the acid amount of 10-NiO–ZrO2/15-MoO3 catalyst is the most among the catalysts. Of course, the acid amount of catalysts is related to their specific surface area, as mentioned above. In fact, Tables 1 and 3 show that the specific surface area attained a
Table 3 Dispersed MoO3 amount, specific surface area, acid amount, and catalytic activity of 10-NiO–ZrO2/MoO3 catalysts containing MoO3 content below 20 wt% Catalyst
Dispersed MoO3 amount (MoO3 (g)/NiO–ZrO2 (g))
Surface area (m2/g)
Acid amount (mmol/g)
Catalytic activity (mmol/g 5 min)
10-NiO–ZrO2/5-MoO3 10-NiO–ZrO2/10-MoO3 10-NiO–ZrO2/15-MoO3
0.05 0.11 0.18
128 140 154
112 134 140
1.0 1.3 1.5
224
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
maximum when the MoO3 content in 10-NiO–ZrO2/MoO3 catalysts is 15 wt%. Although the catalyst without MoO3 was inactive as catalyst for ethylene dimerization, as shown in Figs. 9 and 10, the 10-NiO–ZrO2/MoO3 catalysts with MoO3 exhibited high catalytic activity even at room temperature. 3.7. Correlation between catalytic activity and acid amount As mentioned above, the active site responsible for dimerization is suggested to consist of a low-valent nickel ion and an acid, as observed in the nickel-containing catalyst [2,19–21]. It is known that for ethylene dimerization the variations in catalytic activities are closely correlated to the acid amount values of catalyst [38,43]. The acid amount values of some samples after evacuation at 450 8C are listed in Tables 1 and 2 together with their surface areas. Tables 1 and 2 and Fig. 10 show that the catalytic activities substantially run parallel to the acid amount values. The catalytic activities of NiO–ZrO2/MoO3 catalysts containing different NiO and MoO3 contents were examined; the results are shown as a function of acid amount in Fig. 11, where catalysts were evacuated at 450 8C for 1 h before reaction. Fig. 11 shows a good correlation between the catalytic activity and the acid amount. It is confirmed that the catalytic activity gives a maximum at 15 wt% MoO3. This seems to be correlated to the specific surface area and to the acid amount of the catalysts. The acid amount of the catalysts can be also correlated to the amount of
Fig. 11. Correlationship between catalytic activity for ethylene dimerization and acid amount: (A) 3-NiO–ZrO2/15-MoO3, (B) 40-NiO–ZrO2/15-MoO3, (C) 30-NiO–ZrO2/15-MoO3, (D) 10-NiO–ZrO2/5-MoO3, (E) 5-NiO–ZrO2/15MoO3, (F) 10-NiO–ZrO2/25-MoO3, (G) 20-NiO–ZrO2/15-MoO3, (H) 10NiO–ZrO2/20-MoO3, (I) 17-NiO–ZrO2/15-MoO3, (J) 10-NiO–ZrO2/10MoO3, and (K) 10-NiO–ZrO2/15-MoO3.
dispersed MoO3. As listed in Table 3, the maximum acid amount is obtained with the catalyst of 15 wt% MoO3, where the amount of dispersed MoO3 is maximum. The acid amount of NiO–ZrO2/MoO3 catalyst calcined at 450 8C was determined by the amount of NH3 irreversibly adsorbed at 230 8C [2,5,7,10]. As listed in Tables 1 and 2, the BET surface area and acid amount attained a maximum extent for the 10-NiO–ZrO2/ 15-MoO3 catalyst. As shown in Fig. 11, the higher the acid amount, the higher the catalytic activity. In this way one can demonstrate that the catalytic activity of NiO–ZrO2/MoO3 catalyst essentially runs parallel to the acid amount. Good correlations have been found in many cases between the acid amount and the catalytic activities of solid acids. It has been reported that the catalytic activity values of nickel-containing catalysts in ethylene dimerization as well as in butene isomerization are closely correlated with the acid amount values of the catalysts [2,5,7,38]. It is known that a low-valent nickel is favorable to chemisorb carbon monoxide [66,67], as shown in Fig. 12. This shows how the dimerization activity activated at 450 8C is poisoned by carbon monoxide. It is obvious that ethylene over 10-NiO– ZrO2/15-MoO3 catalyst is rapidly consumed without induction period. However, over 10-NiO–ZrO2/15-MoO3 catalyst preadsorbed with 1 mmol/g of CO, the pressure change with reaction time stops, indicating that the preadsorbed CO prevents the subsequent adsorption of ethylene. Therefore, it is clear that the adsorption site of CO, that is, a low-valent nickel, is also responsible for the adsorption site of ethylene. Therefore, the maximum activity for ethylene dimerization at 10 wt% NiO can be explained by suggesting that the adsorption
Fig. 12. Time-course of ethylene dimerization over 10-NiO–ZrO2/15-MoO3 catalyst evacuated at 450 8C for 1.0 h: (*) 10-NiO–ZrO2/15-MoO3 and (*) 10-NiO–ZrO2/15-MoO3 with preadsorbed CO of 1 mmol/g.
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
amount of ethylene on the surface of catalysts is maximum at 10 wt% of NiO. As listed in Table 3, the maximum acid amount at 10 wt% NiO is probably because a low-valent nickel, Ni+ acts as a Lewis acid site. 4. Conclusions NiO–ZrO2/MoO3 catalysts were prepared by adding an aqueous solution of ammonium metatungstate to the Ni(OH)2– Zr(OH)4 powder, followed by drying and calcining at high temperatures for 1.5 h in air. The interaction between MoO3 (or NiO) and ZrO2 influences the physicochemical properties of prepared catalysts with calcination temperature. The interaction between MoO3 (or NiO) and ZrO2 delayed the phase transition of ZrO2 from amorphous to tetragonal. Since the ZrO2 stabilizes the surface MoO3 species, for the catalysts equal to or less than 15 wt%, MoO3 was well dispersed on the surface of zirconia, but for the catalysts above 15 wt% the orthorhombic phase of MoO3 was observed at the calcination temperature of 450 8C. The acid amount of 10-NiO–ZrO2/ MoO3 catalysts increased with the amount of dispersed MoO3, showing the presence of Bro¨nsted and Lewis acid sites on the surface of the catalyst. The high acid strength and high acid amount are responsible for the Mo O bond nature of complex formed by the interaction between MoO3 and ZrO2. The catalytic activities of NiO–ZrO2/MoO3 catalysts for ethylene dimerization were correlated to their acid amount as measured by the ammonia chemisorption method. Acknowledgement We wish to thank Korea Basic Science Institute for the use of Raman and X-ray instruments. References [1] F. Bernardi, A. Bottoni, I. Rossi, J. Am. Chem. Soc. 120 (1998) 7770. [2] J.R. Sohn, A. Ozaki, J. Catal. 61 (1980) 29. [3] G. Wendt, E. Fritsch, R. Scho¨llner, H.Z. Siegel, Anorg. Allg. Chem. 467 (1980) 51. [4] G.F. Berndt, S.J. Thomson, G.J. Webb, J. Chem. Soc., Faraday Trans. 1 79 (1983) 195. [5] J.R. Sohn, J.S. Lim, Catal. Today 111 (2006) 403. [6] Y.I. Pae, S.H. Lee, J.R. Sohn, Catal. Lett. 99 (2005) 241. [7] J.R. Sohn, D.C. Shin, J. Catal. 160 (1996) 314. [8] J.R. Sohn, J.S. Han, Appl. Catal. A: Gen. 298 (2006) 168. [9] T.V. Herwijnen, H.V. Doesburg, D.V. Jong, J. Catal. 28 (1973) 391. [10] J.R. Sohn, W.C. Park, H.W. Kim, J. Catal. 209 (2002) 69. [11] J.R. Sohn, W.C. Park, Bull. Korean Chem. Soc. 21 (2000) 1063. [12] K. Urabe, M. Koga, Y.J. Izumi, J. Chem. Soc., Chem. Commun. (1989) 807. [13] G. Wendt, D. Hentschel, J. Finster, R.J. Scho¨llner, J. Chem. Soc., Faraday Trans. 1 (79) (1983) 2013. [14] K. Kimura, A.J. Ozaki, J. Catal. 3 (1964) 395. [15] K. Maruya, A. Ozaki, Bull. Chem. Soc. Jpn. 46 (1973) 351. [16] M. Hartmann, A. Po¨ppl, L. Kevan, J. Phys. Chem. 100 (1996) 9906. [17] I.V. Elev, B.N. Shelimov, V.B. Kazansky, J. Catal. 89 (1984) 470. [18] H. Choo, L. Kevan, J. Phys. Chem. B 105 (2001) 6353. [19] J.R. Sohn, H.J. Kim, J. Catal. 101 (1986) 428.
225
[20] J.R. Sohn, S.Y. Lee, Appl. Catal. A: Gen. 164 (1997) 127. [21] J.R. Sohn, H.W. Kim, M.Y. Park, E.H. Park, J.T. Kim, S.E. Park, Appl. Catal. A: Gen. 128 (1995) 127. [22] D.A. Ward, E.I. Ko, J. Catal. 150 (1994) 18. [23] S.R. Vaudagna, R.A. Comelli, S.A. Canavese, N.S. Figoli, J. Catal. 169 (1997) 389. [24] L.M. Kustov, V.B. Kazansky, F. Figueras, D.J. Trichit, J. Catal. 150 (1994) 143. [25] A. Sayari, Y. Yang, X. Song, J. Catal. 167 (1997) 346. [26] K. Arata, Adv. Catal. 37 (1990) 165. [27] C.Y. Hsu, C.R. Heimbuch, C.T. Armes, B.C. Gates, J. Chem. Soc., Chem. Commun. (1992) 1645. [28] T.K. Cheung, B.C. Gates, J. Catal. 168 (1997) 522. [29] V. Adeeva, H.W. de Haan, J. Janchen, G.D. Lei, V. Schunemann, L.J.M. van de Ven, W.M.H. Sachtler, R.A. van Santen, J. Catal. 151 (1995) 364. [30] K.T. Wan, C.B. Khouw, M.E. Davis, J. Catal. 158 (1996) 311. [31] X. Song, K.R. Reddy, A. Sayari, J. Catal. 161 (1996) 206. [32] M.A. Coelho, D.E. Resasco, E.C. Sikabwe, R.L. White, Catal. Lett. 32 (1995) 253. [33] T. Hosoi, T. Shimadzu, S. Ito, S. Baba, H. Takaoka, T. Imai, N. Yokoyama, Prepr. Symp. Div. Petr. Chem., American Chemical Society, Los Angeles, CA, 1988, p. 562. [34] K. Ebitani, J. Konishi, H. Hattori, J. Catal. 130 (1991) 257. [35] M. Signoretto, F. Pinna, G. Strukul, P. Chies, G. Cerrato, S.D. Ciero, C. Morterra, J. Catal. 167 (1997) 522. [36] W. Hua, Y. Xia, Y. Yue, Z. Gao, J. Catal. 196 (2000) 104. [37] J.A. Moreno, G. Poncelet, J. Catal. 203 (2001) 453. [38] J.R. Sohn, E.S. Cho, Appl. Catal. A: Gen. 282 (2005) 147. [39] M. Hino, K. Arata, J. Chem. Soc., Chem. Commun. (1987) 1259. [40] J.R. Sohn, E.W. Chun, Y.I. Pae, Bull. Korean Chem. Soc. 24 (2003) 1785. [41] G. Larsen, E. Lotero, R.D. Para, in: Proceedings of the 11th International Congress on Catalysis, Elsevier, New York, 1996, p. 543. [42] A.S.C. Brown, J.S.J. Hargreaves, S.H. Taylor, Catal. Lett. 57 (1999) 109. [43] J.R. Sohn, S.H. Lee, Appl. Catal. A: Gen. 266 (2004) 89. [44] J.R. Sohn, D.H. Seo, S.H. Lee, J. Ind. Eng. Chem. 10 (2004) 309. [45] J.R. Sohn, S.H. Lee, W.C. Park, H.W. Kim, Bull. Korean Chem. Soc. 25 (2004) 657. [46] G. Larsen, E. Lotero, L.M. Petkovic, D.S. Shobe, J. Catal. 169 (1997) 67. [47] P. Afanasiev, C. Geantet, M. Breysse, G. Coudurier, J.C. Vedrine, J. Chem. Soc., Faraday Trans. 190 (1994) 193. [48] J.R. Sohn, J. Ind. Eng. Chem. 10 (2004) 1. [49] J.R. Sohn, S.G. Cho, Y.I. Pae, S. Hayashi, J. Catal. 159 (1996) 170. [50] Z. Liu, Y. Chen, J. Catal. 177 (1998) 314. [51] B. Zhao, X. Wang, H. Ma, Y. Tang, J. Mol. Catal. A: Chem. 108 (1996) 167. [52] K. Chen, S. Xie, E. Iglesia, A.T. Bell, J. Catal. 189 (2000) 421. [53] M.R. Smith, U.S. Ozkan, J. Catal. 141 (1993) 124. [54] G. Mestl, T.K.K. Srinivasan, Cat. Rev. Sci. Eng. 40 (1998) 451. [55] P. Dufresne, E. Payen, J. Grimblot, J.P. Bonnelle, J. Phys. Chem. 85 (1981) 2344. [56] H. Hu, I.E. Wachs, J. Phys. Chem. 99 (1995) 10897. [57] R.D. Roark, S.D. Kohler, J.G. Ekerdt, D.S. Kim, I.E. Wachs, Catal. Lett. 16 (1992) 77. [58] C.H. Schild, A. Wokaun, R.A. Ko¨ppel, A. Baiker, J. Catal. 130 (1991) 657. [59] J.R. Sohn, I.J. Doh, Y.I. Pae, Langmuir 18 (2002) 6280. [60] M. Scheithauer, R.K. Grasselli, H. Kno¨zinger, Langmuir 14 (1998) 3019. [61] Y.C. Xie, Y.Q. Tang, Adv. Catal. 37 (1990) 1. [62] L. Litteti, I. Nova, G. Ramis, L. DallAcqua, G. Busca, E. Giamello, P. Forzatti, F. Bregani, J. Catal. 187 (1999) 419. [63] F.G.A. Olah, G.K.S. Prakash, J. Sommer, Science 206 (1979) 13. [64] K. Tanabe, M. Misono, Y. Ono, H. Hattori, New Solid Acids and Bases, Elsevier Science, Amsterdam, 1989, p. 185. [65] J.R. Sohn, S.G. Ryu, Langmuir 9 (1993) 126. [66] K. Kimura, H. A-I, A. Ozaki, J. Catal. 18 (1970) 271. [67] J.R. Sohn, W.C. Park, Appl. Catal. A: Gen. 239 (2003) 269.
Spectroscopic studies on NiO supported on ZrO2 modified with MoO3 for ethylene dimerization Jong Rack Sohn *, Si Hyon Kwon, Dong Cheol Shin Department of Applied Chemistry, Engineering College, Kyungpook National University, Taegu 702-701, Republic of Korea Received 23 May 2006; received in revised form 28 September 2006; accepted 10 October 2006 Available online 17 November 2006
Abstract NiO supported on zirconia modified with MoO3 for ethylene dimerization was prepared by drying powdered Ni(OH)2–Zr(OH)4 with ammonium heptamolybdate aqueous solution, followed by calcining in air at high temperature. The characterization of prepared catalysts was performed using FTIR, Raman, XRD, and DSC. MoO3 equal to or less than 15 wt% was dispersed on the surface of catalyst as two-dimensional polymolybdate or monomolybdate, while for MoO3 above 15 wt%, a crystalline orthorhombic phase of MoO3 was formed, showing that the critical dispersion capacity of MoO3 on the surface of catalyst is 0.18 g/g NiO–ZrO2 on the basis of XRD analysis. From 600 8C of calcination temperature, zirconium molybdenum oxide, Zr(MoO4)2 was formed due to the reaction between ZrO2 and MoO3; its amount increased with the calcination temperature. The acid amount and the catalytic activity of catalysts increased with the amount of dispersed MoO3. The high acid strength and acid amount was responsible for the Mo O bond nature of the complex formed by the interaction between MoO3 and ZrO2. The catalytic activity for ethylene dimerization was correlated with the acid amount of the catalysts measured by the ammonia chemisorption method. # 2006 Published by Elsevier B.V. Keywords: NiO–ZrO2/MoO3 catalyst; Ethylene dimerization; Acidic properties; MoO3 dispersion; Modification with MoO3
1. Introduction Nickel oxide on silica or silica–alumina is an effective catalyst in olefin dimerization and isomerization as well as in the hydrocracking, hydrogenation and methanation process following reduction of the nickel component [1–3]. The dimerization of alkenes is an important method for the production of higher olefins, which are extensively used as industrial intermediates. A considerable number of papers have dealt with the nickelcontaining catalysts for ethylene dimerization [1–12]. One of the remarkable features of this catalyst system is its activity in relation to a series of n-olefins. In contrast to the usual acid-type catalysts, nickel oxide on silica or silica–alumina shows a higher activity for a lower olefin dimerization, particularly for ethylene dimerization [1–4,13]. It has been suggested that the active site for dimerization is formed by an interaction of a low-valent nickel ion with an acid site [10,14]. The dimerization activities of such catalysts are reported to be related to the acidic properties of
* Corresponding author. Tel.: +82 53 950 5585; fax: +82 53 950 6594. E-mail address: [email protected] (J.R. Sohn). 0926-860X/$ – see front matter # 2006 Published by Elsevier B.V. doi:10.1016/j.apcata.2006.10.015
surface and low-valent nickel ions. In fact, nickel oxide, which is active for C2H4–C2D4 equilibration, acquires an activity for ethylene dimerization upon addition of nickel sulfate, which is known to be an acid [15]. A transition metal can also be supported on zeolite in the state of a cation or as a finely dispersed metal. Transition metal ions like Ni+ or Pd+ can be active sites in catalytic reactions such as ethylene and propylene dimerization as well as in acetylene cyclomerization [16–18]. The previous papers from this laboratory have shown that NiO–TiO2 and NiO–ZrO2 modified with sulfate are very active for ethylene dimerization [19–21]. High catalytic activities in the reactions were attributed to the enhanced acidic properties of the modified catalysts, which originated from the inductive effect of S O of the complex formed by the interaction of oxides with sulfate. Zirconium oxide is a very interesting material because of its thermal stability, its mechanical properties and its basic, acidic, reducing, and oxidizing surface properties. Therefore, it is attracting interest in catalysis as a support material as well as a catalyst. The potential for a heterogeneous catalyst has yielded many papers on the catalytic activity of sulfated zirconia materials [22–26]. Moreover, catalytic functions have been improved by loading additional components. Sulfated zirconia
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
incorporating Fe and Mn has been shown to be highly active for butane isomerization, catalyzing the reaction even at room temperature [27,28]. The promotion of activity of the catalyst has been confirmed by several other research groups [29–31]. Coelho et al. have discovered that the addition of Ni to sulfated zirconia causes an activity enhancement comparable to that caused by the addition of Fe and Mn [32]. It has been reported by several workers that the addition of platinum to zirconia modified by sulfate ions enhances the catalytic activity in the skeletal isomerization of alkanes without deactivation when the reaction is carried out in the presence of hydrogen [33–35]. Recently, it has been found that a main group element Al can also promote the catalytic activity and stability of sulfated zirconia in ethylene dimerization and n-butane isomerization [36–38]. Recently, some workers reported zirconia-supported WO3 or MoO3 as alternative materials in reactions requiring strong acid sites [26,39–41]. Several advantages of tungstate or molybdate over sulfate as dopant include the facts that they do not suffer from dopant loss during thermal treatment and that they undergo significantly less deactivation during catalytic reaction [41]. As an extension of the study on the ethylene dimerization, these authors tried to prepare other catalyst systems by combining nickel hydroxide to give low-valent nickel after decomposition with ZrO2 modified by MoO3 which is known to be an acid [40,42], to improve the catalytic activity for ethylene dimerization and acidic properties. In this paper, spectroscopic studies on NiO–ZrO2/MoO3 catalysts and catalytic activity for ethylene dimerization are reported.
217
290 Torr. A fresh catalyst sample of 0.2 g was used for every run and the catalytic activity was calculated as the number of moles of ethylene consumed during the initial 5 min. Reaction products were analyzed by gas chromatography with a VZ-7 column at room temperature. FTIR spectra were obtained in a heatable gas cell at room temperature using a Mattson Model GL6030E spectrophometer. The self-supporting catalyst wafers contained about 9 mg/cm2. Prior to obtaining the spectra we heated the samples under vacuum at 500 8C for 1.5 h. The Raman spectra were recorded on a Spex Ramalog ˚ line from spectrometer with holographic gratings. The 5145-A a Spectra-Physics Model 165 argon-ion laser was used as the exciting source. The spectral shift width was typically 4 cm 1, and laser source powers of approximately 45 mW, measured at the sample, were used. Catalysts were checked in order to determine the structure of the support as well as that of molybdenum oxide by means of a Jeol Model JDX-8030 diffractometer, employing Cu Ka (Ni-filtered) radiation. DSC measurements were performed by a PL-STA model 1500H apparatus in air, and the heating rate was 5 8C per minute. For each experiment, 10–15 mg of sample was used. The specific surface area was determined by applying the BET method to the adsorption of N2 at 196 8C. Chemisorption of ammonia was also employed as a measure of the acid amount of each catalyst. The amount of chemisorption was determined based on the irreversible adsorption of ammonia [43–45].
2. Experimental 2.1. Catalyst preparation
3. Results and discussion
The coprecipitate of Ni(OH)2–Zr(OH)4 was obtained by adding aqueous ammonia slowly into a mixed aqueous solution of nickel chloride and zirconium oxychloride at room temperature with stirring until the pH of the mother liquor reached about 8. The coprecipitate thus obtained was washed thoroughly with distilled water until chloride ion was not detected, and then was dried at 100 8C for 12 h. The dried precipitate was powdered below 100 mesh. The catalysts containing various MoO3 contents were prepared by adding an aqueous solution of ammonium heptamolybdate[(NH4)6(Mo7O24)4H2O] to the Ni(OH)2– Zr(OH)4 powder, followed by drying and calcining at high temperatures for 1.5 h in air. This series of catalysts are denoted by their weight percentage of NiO and MoO3. For example, 10NiO–ZrO2/15-MoO3 indicates the catalyst containing 10 wt% NiO regarding only two components, NiO and ZrO2, and containing 15 wt% MoO3 regarding all three components, NiO, ZrO2 and MoO3.
3.1. Crystalline structures of catalysts
2.2. Procedure The catalytic activity for ethylene dimerization was determined at 20 8C using a conventional static system following a pressure change from an initial pressure of
The crystalline structures of ZrO2 and NiO–ZrO2/MoO3 catalysts calcined in air at different temperatures for 1.5 h were examined. For pure ZrO2, ZrO2 was amorphous to X-ray diffraction up to 300 8C, with a two-phase mixture of the tetragonal and monoclinic forms at 400–700 8C and a monoclinic form at 800 8C (this figure is not shown here). Three crystal structures of ZrO2: tetragonal, monoclinic and cubic phases, have been reported [46,47]. However, in the case of NiO–ZrO2 modified with MoO3, the crystalline structures of the catalysts were different from that of pure ZrO2. For the 10NiO–ZrO2/15-MoO3 catalysts, as shown in Fig. 1, ZrO2 was amorphous to X-ray diffraction up to 500 8C with tetragonal ZrO2 formed at 600–800 8C. The transition temperature of ZrO2 from amorphous to tetragonal phase was higher by 200 8C than of pure ZrO2. However, the higher the calcination temperature, the larger the amount of tetragonal phase of ZrO2. It is assumed that the interaction between MoO3 (or NiO) and ZrO2 hinders the transition of ZrO2 from amorphous to tetragonal phase [48,49]. The presence of MoO3 strongly influences the development of textural properties with temperature. No phase of MoO3 or NiO was observed at any calcination temperature, indicating a good dispersion of MoO3
218
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
Fig. 1. X-ray diffraction patterns of 10-NiO–ZrO2/15-MoO3 catalysts calcined at different temperatures for 1.5 h: (*) tetragonal ZrO2 and () orthorhombic MoO3.
or NiO on the surfaces of ZrO2 supports due to the interaction between them. XRD patterns of 10-NiO–ZrO2/25-MoO3 catalysts calcined at 400–800 8C for 1.5 h are shown in Fig. 2. X-ray diffraction data indicated the amorphous phase of ZrO2 at 400 8C, showing the tetragonal phase of ZrO2 at 450–800 8C. However, the higher the calcined temperature, the larger the amount of tetragonal phase of ZrO2. Moreover, as shown in Fig. 2, the orthorhombic phase of MoO3 was observed at 400–500 8C. These results are in good agreement with those of Raman (Fig. 4) and IR (Fig. 6) described later. It has been reported that MoO3 for catalysts containing low MoO3 loading is dispersed on the surface of zirconia as a monolayer state, that is, two-dimensional polymolybdate or monomolybdate, which gives rise to the characteristic broad band at 900–1000 cm 1 [50,51]. However, as mentioned above, when MoO3 loading exceeds a monolayer capacity on the surface of NiO–ZrO2, the residual crystalline phase of MoO3 can be detected by XRD, Raman, and IR. From the calcination temperature of 600 8C, a new species, i.e., Zr(MoO4)2 hexagonal phase, has been detected in the 10-NiO–ZrO2/25-MoO3 catalyst. As shown in Fig. 2, the amount of Zr(MoO4)2 phase increased with the calcination temperature. Chen et al. reported the formation of Zr(MoO4)2 from MoO3 and ZrO2 in high MoO3 catalysts calcined above 600 8C [52].
Fig. 2. X-ray diffraction patterns of 10-NiO–ZrO2/25-MoO3 catalysts calcined at different temperatures for 1.5 h: (*) tetragonal ZrO2; (~) monoclinic ZrO2; () orthorhombic MoO3; (&) hexagonal Mo(ZrO4)2.
The XRD patterns of 10-NiO–ZrO2/MoO3 catalysts containing different MoO3 contents and calcined at 450 8C for 1.5 h are shown in Fig. 3. XRD data indicated only tetragonal phase of ZrO2 at 5 wt% of MoO3. However, 10-NiO–ZrO2/10-MoO3 and 10-NiO–ZrO2/15-MoO3 catalysts were amorphous to X-ray diffraction, indicating that the increase of MoO3 loading hinders the phase transition of ZrO2 from amorphous to tetragonal, due to the interaction between MoO3 and ZrO2 [38,49]. In this case, no detection of orthorhombic phase of MoO3 means that all MoO3 is highly dispersed on the surface of NiO–ZrO2 as two-dimensional polymolybdate or monomolybdate, showing that these results are in good agreement with those of Raman (Fig. 4) and IR (Fig. 6). For the catalysts containing higher MoO3 loading than 15 wt%, as shown in Fig. 3, the orthorhombic phase of MoO3 was observed because MoO3 loading exceeds a monolayer capacity on the surface of ZrO2, showing that its intensity increases with the MoO3 content. When the loading amounts of MoO3 are low, the highly dispersed MoO3 species can not be detected by XRD, as shown in Fig. 3. Therefore, we can estimate the amount of crystalline MoO3 by the quantitative XRD analysis. So, we can plot the content of crystalline MoO3 as a function of the total content of
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
219
MoO3 in the 10-NiO–ZrO2/MoO3 catalysts containing different MoO3 contents and calcined at 450 8C (this figure is not shown here), from which the MoO3 critical dispersion capacity on NiO–ZrO2, 0.18 g MoO3/g NiO–ZrO2 is deduced. Since the surface area of 10-NiO–ZrO2/15-MoO3 catalyst is 154 m2/g, this capacity is equivalent to 0.12 g MoO3/100 m2 NiO–ZrO2, which is very similar with the monolayer capacity for MoO3 supported on ZrO2 [51]. It is expected that this dispersion capacity will be related to the surface area and acid amount of catalysts, as discussed below. 3.2. Raman and infrared spectra
Fig. 3. X-ray diffraction patterns of 10-NiO–ZrO2/MoO3 catalysts containing different MoO3 contents and calcined at 450 8C for 1.5 h: (*) tetragonal ZrO2 and () orthorhombic MoO3.
Fig. 4. Raman spectra of 10-NiO–ZrO2/MoO3 catalysts containing different MoO3 contents.
Raman spectroscopy is a valuable tool for the characterization of dispersed metal oxides and can detect vibrational modes of surface and bulk structures. The Raman spectra of MoO3 obtained by calcining ammonium heptamolybdate at 450 8C and 10-NiO–ZrO2/MoO3 catalysts containing different MoO3 contents under ambient conditions are presented in Fig. 4. The MoO3 structure is made up of distorted MoO3 octahedra. Bulk crystalline MoO3 shows the main bands, in good agreement with data previously reported [53]. The major vibrational modes of MoO3 are located at 995, 819, 667, and 290 cm 1, and have been assigned to the Mo O stretching mode, the Mo–O– Mo stretching mode, OMo3 vibration mode, and the Mo O deformation mode, respectively [53]. The molecular structure of the supported MoO3 species depends on the loading. Several authors observed that the nature of surface molybdenum species on SiO2, Al2O3, TiO2 and ZrO2 depends on the amount of MoO3. A band at about 825 cm 1, corresponding to Mo–O–Mo vibrations [50,54], and a band at 910–980 cm 1, assigned to Mo O vibrations in twodimensional polymolybdate, were observed in these samples [50,54]. Raman bands between 910 and 980 cm 1 are usually attributed to the Mo O vibration of Mo species in either octahedral or tetrahedral environment [55]. Generally, monomolybdate or tetrahedral molybdenum oxygen species have been assigned for low MoO3 loading catalysts [50,55,56], and two-dimensional polymolybdate or octahedral molybdenum– oxygen species with characteristic band around 950–980 cm 1, have been assigned for high MoO3 loading catalysts [50,55,56]. Therefore, the broad band observed in the 900–1000 cm 1 region in Fig. 4 will be interpreted as an overlap of two characteristic bands for two molybdenum oxide species. The bands around 350–370 cm 1 for 10-NiO–ZrO2/15-MoO3 and 10-NiO–ZrO2/10-MoO3 catalysts are assigned to Mo O bending mode [57]. For 10-NiO–ZrO2/10-MoO3 and 10-NiO–ZrO2/15-MoO3 catalysts in Fig. 4, most of the zirconia is amorphous to X-ray diffraction, showing tiny amount of tetragonal phase zirconia for the other. The Raman spectrum of amorphous zirconia is characterized by a very weak and broad band at 550–600 cm 1 [58]. Tetragonal zirconia is expected to yield a spectrum consisting of Raman bands with frequencies at about 148, 263, 325, 472, 608, and 640 cm 1, while monoclinic zirconia exhibits the characteristic features at 180, 188, 221, 380, 476, and 637 cm 1 [59,60]. As shown in Fig. 4, the 10-NiO–ZrO2/5-
220
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
Fig. 5. Raman spectra of 10-NiO–ZrO2/25-MoO3 catalysts calcined at different temperatures for 1.5 h.
MoO3 catalyst exhibits the characteristic features of tetragonal zirconia, indicating no transformation of ZrO2 from tetragonal to monoclinic. Fig. 5 shows Raman spectra of 10-NiO–ZrO2/25-MoO3 catalysts calcined at 400–800 8C for 1.5 h. For catalysts calcined at 400–500 8C, the vibrational modes of bulk MoO3 were observed at 995, 819, 667 and 290 cm 1 similarly to those of Fig. 4. All catalysts calcined at 450–800 8C except that at 400 8C also exhibit the characteristic features of tetragonal zirconia at 265, 325, 472, 608, and 640 cm 1, in good agreement with XRD data described above. However, from the calcination temperature of 600 8C, a new species, i.e., Zr(MoO4)2 phase, identified by its characteristic Raman bands at 748, 945, and 1000 cm 1 [50], has been detected in the 10NiO–ZrO2/25-MoO3 catalyst. As shown in Fig. 5, the amount of Zr(MoO4)2 phase increased with the calcination temperature. Chen et al. [52] reported the formation of Zr(MoO4)2 between MoO3 and ZrO2 in high loading catalysts calcined above 600 8C. For 10-NiO–ZrO2/15-MoO3 catalysts calcined at 400– 500 8C, the broad bands in the 900–1000 cm 1 region were observed by an overlap of two characteristic bands for two molybdenum oxide species (this figure is not shown here). The bands around 370 cm 1 assigned to Mo O banding mode were also observed. However, all 10-NiO–ZrO2/15-MoO3 catalysts calcined at 600–800 8C exhibit the characteristic features of tetragonal zirconia at 265, 325, 472, 608, and 640 cm 1, in good agreement with XRD data as seen in Fig. 1. Fig. 6 shows IR spectra of 10-NiO–ZrO2/MoO3 catalysts (KBr disc) with different MoO3 contents calcined at 450 8C for 1.5 h. Although with samples having less than 20 wt% of
Fig. 6. Infrared spectra of 10-NiO–ZrO2/MoO3 catalysts calcined at 450 8C for 1 h.
MoO3 no definite peaks were observed, the absorption bands at 994, 862, and 599 cm 1 appeared for 10-NiO–ZrO2/MoO3 catalysts having more than 15 wt% of MoO3. These bands are due to the crystalline MoO3. When MoO3 loading exceeds a certain value, the surplus MoO3 is present as crystalline form [50,51,61]. In view of Raman spectra of crystalline MoO3 in Fig. 4, for crystalline MoO3 the IR bands at 991, 870, and 596 cm 1 in Fig. 6 are assigned to the Mo O stretching mode, the Mo–O–Mo stretching mode, and OMo3 vibration mode, respectively. XRD measurements have also supported the presence of the above crystalline MoO3, as discussed above. The Raman and IR spectra in Figs. 4–6 have been taken in air using the pure powders. To examine the structure of NiO–ZrO2 modified with MoO3 under dehydration conditions, we obtained IR spectra of 10-NiO–ZrO2/15-MoO3 catalyst in a heatable gas cell after evacuation at 25 and 500 8C for 1.5 h (this figure is not shown here). For the 10-NiO–ZrO2/15-MoO3 catalyst, IR band at 997 cm 1 after evacuation at 500 8C is due to the Mo O stretching mode of the molybdenum oxide complex bonded to the ZrO2 surface [62]. This Mo O band due to molybdenyl species only appears on the evacuated catalysts, being shifted on the wet catalysts. The broad band at 944 cm 1 under ambient condition is assigned to the terminal Mo O stretching of the hydrated form of the surface molybdenyl species [62]. This can be rationalized by assuming that the adsorption of water causes a strong perturbation of the corresponding molybdenum oxide species, with a consequent strong broadening and down shift of this band. This isolated molybdenum oxide species is stabilized through multiple Mo– O–Zr bonds between each molybdenum oxide species and the
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
zirconia surface [50,52]. Upon dehydration at elevated temperature, the hydrated surface metal oxide species is unstable and decomposes to form dehydrated surface metal oxide species by direct interaction with the surface OH groups of support, giving the formation of metal–oxygen-support bond. 3.3. Thermal analysis In X-ray diffraction patterns, it was shown that the structures of NiO–ZrO2/MoO3 catalysts were different depending on the calcined temperature. To examine the thermal properties for the precusors of samples more clearly, we carried out their thermal analysis as illustrated in Fig. 7. For pure ZrO2 (Fig. 7(h)) the DSC curve showed a broad endothermic peak below 180 8C due to the elimination of adsorbed water, and an exothermic peak at 418 8C due to the ZrO2 crystallization [49,59]. For (NH4)6(Mo7O24)4H2O (Fig. 7(i)), the DSC curve shows three endothermic peaks below 400 8C, indicating that the decomposition of ammonium heptamolybdate occurs in three steps. In the case of NiO–ZrO2/MoO3 precursors, three additional endothermic peaks appeared in the region of 200–400 8C due to the evolution of NH3 and H2O decomposed from ammonium heptamolybdate. Also it is considered that an endothermic at 796 8C is responsible for melting of MoO3 [61] and for 10NiO–ZrO2/25-MoO3 and ZrO2/15-MoO3 precursors (Fig. 7(a) and (b)) an exothermic peak around 610 8C is due to the formation of Zr(MoO4)2 as described in X-ray diffraction patterns. However, it is of interest to see the influences of NiO and MoO3 on the crystallization of ZrO2 from amorphous to
221
tetragonal phase. As Fig. 7 shows, the exothermic peak due to the crystallization appears at 418 8C for pure ZrO2, while for NiO–ZrO2 precursors (Fig. 7(c)–(g)) and NiO–ZrO2/MoO3 precursor (Fig. 7(a)), it was shifted to higher temperatures. The shift increased with increasing NiO content up to 25 wt% of NiO. It is considered that the interaction between NiO (or MoO3) and ZrO2 delays the transition of ZrO2 from amorphous to tetragonal phase [38]. Consequently, the exothermic peaks appear at 429 8C for 3-NiO–ZrO2, 446 8C for 5-NiO–ZrO2, 493 8C for 10-NiO–ZrO2, 545 8C for ZrO2/15-MoO3, 551 8C for 10-NiO–ZrO2/25-MoO3, 562 8C for 17-NiO–ZrO2, and 575 8C for 25-NiO–ZrO2. 3.4. Surface properties It is necessary to examine the effects of molybdenum oxide on the surface properties of catalysts, that is, specific surface area, acid amount, and acid strength. The specific surface areas of 10-NiO–ZrO2/MoO3 catalysts calcined at 450 8C for 1.5 h are listed in Table 1. The presence of molybdenum oxide strongly influences the surface area in comparison with the pure ZrO2. Specific surface areas of 10-NiO–ZrO2/MoO3 catalysts are much larger than that of pure ZrO2 calcined at the same temperature, showing that surface area increases gradually with increasing MoO3 content up to 15 wt% of MoO3. The decrease of surface area for samples containing MoO3 above 15 wt% is due to the blocking of ZrO2 pore by the increased MoO3 loading. It seems likely that the interaction between MoO3 and ZrO2 protects catalysts from sintering. The acid amount of catalysts calcined at 450 8C, as determined by the amount of NH3 irreversibly adsorbed at 230 8C [5,8,38], is also listed in Table 1. The acid amount per surface area for these catalysts was calculated to be nearly constant values of 0.9–1.0 mmol/ m2, indicating that the variation of acid amount runs parallel to the change of surface area. We examined the effect of NiO addition on the surface area and acid amount of NiO–ZrO2/15MoO3 catalysts. The specific surface areas and acid amount of catalysts calcined at 450 8C are listed in Table 2. Both surface area and acid amount increased with increasing NiO content up to 10 wt%, indicating the promoting effect of NiO. The acid strength of the catalysts was examined by a color change method, using Hammett indicators in sulfurylchloride [38,40,49]. NiO–ZrO2/MoO3 catalyst was estimated to have a H0 14.5, indicating the formation of new acid sites stronger than those of single oxide components. Acids stronger than H0 11.93, which corresponds to the acid strength of 100% Table 1 Specific surface area and acid amount of 10-NiO–ZrO2/MoO3 catalysts containing different MoO3 contents and calcined at 450 8C for 1.5 h
Fig. 7. DSC curves of catalyst precursors: (a) 10-NiO–ZrO2/25-MoO3, (b) ZrO2/15-MoO3, (c) 25-NiO–ZrO2, (d) 17-NiO–ZrO2, (e) 10-NiO–ZrO2, (f) 5NiO–ZrO2, (g) 3-NiO–ZrO2, (h) ZrO2, and (i) (NH4)6(Mo7O24)4H2O.
MoO3 content (wt%)
Surface area (m2/g)
Acid amount (mmol/g)
5 10 15 20 25
128 140 155 138 124
112 134 141 129 121
222
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
Table 2 Specific surface area and acid amount of NiO–ZrO2/15-MoO3 catalysts containing different NiO contents and calcined at 450 8C for 1.5 h
contents also showed the presence of both Lewis and Bro¨nsted acid sites.
NiO content (wt%)
Surface area (m2/g)
Acid amount (mmol/g)
3.5. Ethylene dimerization over NiO–ZrO2/MoO3
3 5 10 15 20 30 40
139 143 155 138 128 121 89
78 119 141 137 121 95 78
H2SO4, are superacids [26,63,64]. Consequently, NiO–ZrO2/ MoO3 catalysts would be solid superacids. The superacidic property is attributed to the double bond nature of the Mo O in the complex formed by the interaction of ZrO2 with molybdate, in analogy with the cases of ZrO2 modified with tungstate, chromate, and sulfate ions [48,49,65]. Infrared spectroscopic studies of ammonia adsorbed on solid surface have made it possible to distinguish between Bro¨nsted and Lewis acid sites [5,20,38,40]. Fig. 8 shows the IR spectra of ammonia adsorbed on 10-NiO–ZrO2/15-MoO3 catalyst evacuated at 400 8C for 1 h. The band at 1448 is the characteristic peak of ammonium ion, which is formed on the Bro¨nsted acid sites and the absorption peak at 1609 cm 1 is contributed by ammonia coordinately bonded to Lewis acid sites [5,20,38,40], indicating the presence of both Bro¨nsted and Lewis acid sites on the surface of catalyst. Other samples having different MoO3
NiO–ZrO2/MoO3 catalysts were tested for their effectiveness in ethylene dimerization. It was found that, over 10-NiO– ZrO2/15-MoO3, 10-NiO–ZrO2/10-MoO3, and 3-NiO–ZrO2/ 15-MoO3 catalysts, ethylene was continuously consumed, as shown by the results presented in Fig. 9, where catalysts were evacuated at 450 8C for 1 h. Over NiO–ZrO2/MoO3, ethylene was selectively dimerized to n-butenes. In the composition of n-butenes analyzed by gas chromatography, 1-butene was found exclusively at the initial reaction time, with no cisbutene or trans-butene. However, the amount of 1-butene decreases with the reaction time, while the amount of 2-butene increases. Therefore, it seems likely that the initially produced 1-butene is also isomerized to 2-butene during the reaction time [2,5,20]. NiO–ZrO2/MoO3 catalyst was effective for ethylene dimerization, but as shown in Fig. 9, NiO–ZrO2 without MoO3 did not exhibit any catalytic activity because the sample without MoO3 does not possess enough acid amount or acid strength to catalyze ethylene dimerization. Therefore, it is expected that the amount of dispersed MoO3 will be related to the acid amount and catalytic activity of catalyst, as discussed below. ZrO2/15-MoO3 alone without NiO was also totally inactive for ethylene dimerization, as shown in Fig. 9. It is known that ZrO2 modified with MoO3 is an acid [26,40].
Fig. 8. Infrared spectra of NH3 adsorbed on 10-NiO–ZrO2/15-MoO3 catalyst: (a) background of 10-NiO–ZrO2/15-MoO3 catalyst after evacuation at 400 8C for 1 h and (b) NH3 adsorbed on (a); where gas was evacuated at 230 8C for 1 h.
Fig. 9. Time-course of ethylene dimerization over catalysts evacuated at 450 8C for 1 h: (*) 10-NiO–ZrO2/15-MoO3; (&) 10-NiO–ZrO2/10-MoO3; (~) 3NiO–ZrO2/15-MoO3; (*) ZrO2/15-MoO3; (&) 10-NiO–ZrO2.
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
The active site responsible for dimerization is suggested to consist of a low-valent nickel ion and an acid, as observed in the nickel-containing catalyst [2,19–21]. The term ‘low-valent nickel’ originated from the fact that the NiO–SiO2 catalyst was drastically poisoned by carbon monoxide, since a low-valent nickel is favorable to chemisorb carbon monoxide [66,67]. Previously, from the results of IR of CO adsorbed on NiSO4/gAl2O3 and XPS to identify low-valent nickel species, we reported that the active sites responsible for ethylene dimerization consist of a low-valent nickel, Ni+, and an acid [67]. Kazansky and coworkers concluded that Ni+ ions in NiCaY zeolite are responsible for the catalytic activity; they used EPR spectroscopy to identify low-valent nickel species [17]. Therefore, the nickel species should be active sites for ethylene dimerization. In fact, as shown in Fig. 9, ZrO2/MoO3 without NiO or NiO–ZrO2 without MoO3 is inactive for ethylene dimerization. The catalytic activities of 10-NiO–ZrO2/15-MoO3 catalysts were tested as a function of calcination temperature. The activities increased with the calcination temperature, reaching a maximum at 450 8C, after which the activities decreased. The decrease of catalytic activity after calcination above 450 8C seems to be due to the decrease of surface area and acid amount at high calcination temperatures. In fact, the surface area of 10NiO–ZrO2/15-MoO3 catalyst calcined at 450 8C was found to be 154 m2/g, but that of the catalyst calcined at 600 8C decreased to 109 m2/g. The acid amount values of the catalysts calcined at 450 and 600 8C were found to be 140 and 110 mmol/ g, respectively. Thus, hereafter, emphasis is placed only on the catalysts calcined at 450 8C. 3.6. Effects of dispersed MoO3 amount on acid amount and catalytic activity The forms of active components present in heterogeneous catalysts are of importance to catalysis. A great many oxides can disperse spontaneously onto the surfaces of supports to form a monolayer, because the monolayer is a thermodynamically stable form [61]. Dispersed MoO3 amount, surface area, acid amount, and catalytic activity for 10-NiO–ZrO2/MoO3 catalysts containing MoO3 content below 20 wt% are listed in Table 3. There are good correlations among the dispersed MoO3 amount, acid amount, and catalytic activity. Namely, the larger the dispersed MoO3 amount, the higher both the acid amount and the catalytic activity are. This can be explained by the suggestion that strong acid sites are formed through the bonding between dispersed MoO3 and ZrO2 and consequently catalytic activity increases due to the increased acid sites.
223
Fig. 10. Catalytic activity of 10-NiO–ZrO2/MoO3 catalysts for ethylene dimerization as a function of MoO3 content.
The catalytic activity as a function of MoO3 content is plotted in Fig. 10. As shown in Fig. 10, the maximum activity is obtained with the catalyst of 15 wt% MoO3, where the amount of dispersed MoO3 is maximum. It is known that the active sites responsible for dimerization consist of a low-valent nickel ion and an acid [2,19–21]. Therefore, it seems likely that the highest activity of the catalyst containing 15 wt% MoO3 is related to its acid amount and acid strength. The high acid strength and acid amount are responsible for the Mo O bond nature of complex formed by the interaction between MoO3 and ZrO2 [26,40]. As discussed in IR spectra, IR spectra of 10-NiO– ZrO2/15-MoO3 catalyst after evacuation at 500 8C showed the band at 997 cm 1 due to the Mo O stretching mode of the molybdenum oxide complex bonded to the ZrO2 surface [62]. This isolated molybdenum oxide species is stabilized through multiple Mo–O–Zr bonds between each MoO3 species and the ZrO2 surface [50,52]. As listed in Tables 1 and 3, the acid amount of 10-NiO–ZrO2/15-MoO3 catalyst is the most among the catalysts. Of course, the acid amount of catalysts is related to their specific surface area, as mentioned above. In fact, Tables 1 and 3 show that the specific surface area attained a
Table 3 Dispersed MoO3 amount, specific surface area, acid amount, and catalytic activity of 10-NiO–ZrO2/MoO3 catalysts containing MoO3 content below 20 wt% Catalyst
Dispersed MoO3 amount (MoO3 (g)/NiO–ZrO2 (g))
Surface area (m2/g)
Acid amount (mmol/g)
Catalytic activity (mmol/g 5 min)
10-NiO–ZrO2/5-MoO3 10-NiO–ZrO2/10-MoO3 10-NiO–ZrO2/15-MoO3
0.05 0.11 0.18
128 140 154
112 134 140
1.0 1.3 1.5
224
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
maximum when the MoO3 content in 10-NiO–ZrO2/MoO3 catalysts is 15 wt%. Although the catalyst without MoO3 was inactive as catalyst for ethylene dimerization, as shown in Figs. 9 and 10, the 10-NiO–ZrO2/MoO3 catalysts with MoO3 exhibited high catalytic activity even at room temperature. 3.7. Correlation between catalytic activity and acid amount As mentioned above, the active site responsible for dimerization is suggested to consist of a low-valent nickel ion and an acid, as observed in the nickel-containing catalyst [2,19–21]. It is known that for ethylene dimerization the variations in catalytic activities are closely correlated to the acid amount values of catalyst [38,43]. The acid amount values of some samples after evacuation at 450 8C are listed in Tables 1 and 2 together with their surface areas. Tables 1 and 2 and Fig. 10 show that the catalytic activities substantially run parallel to the acid amount values. The catalytic activities of NiO–ZrO2/MoO3 catalysts containing different NiO and MoO3 contents were examined; the results are shown as a function of acid amount in Fig. 11, where catalysts were evacuated at 450 8C for 1 h before reaction. Fig. 11 shows a good correlation between the catalytic activity and the acid amount. It is confirmed that the catalytic activity gives a maximum at 15 wt% MoO3. This seems to be correlated to the specific surface area and to the acid amount of the catalysts. The acid amount of the catalysts can be also correlated to the amount of
Fig. 11. Correlationship between catalytic activity for ethylene dimerization and acid amount: (A) 3-NiO–ZrO2/15-MoO3, (B) 40-NiO–ZrO2/15-MoO3, (C) 30-NiO–ZrO2/15-MoO3, (D) 10-NiO–ZrO2/5-MoO3, (E) 5-NiO–ZrO2/15MoO3, (F) 10-NiO–ZrO2/25-MoO3, (G) 20-NiO–ZrO2/15-MoO3, (H) 10NiO–ZrO2/20-MoO3, (I) 17-NiO–ZrO2/15-MoO3, (J) 10-NiO–ZrO2/10MoO3, and (K) 10-NiO–ZrO2/15-MoO3.
dispersed MoO3. As listed in Table 3, the maximum acid amount is obtained with the catalyst of 15 wt% MoO3, where the amount of dispersed MoO3 is maximum. The acid amount of NiO–ZrO2/MoO3 catalyst calcined at 450 8C was determined by the amount of NH3 irreversibly adsorbed at 230 8C [2,5,7,10]. As listed in Tables 1 and 2, the BET surface area and acid amount attained a maximum extent for the 10-NiO–ZrO2/ 15-MoO3 catalyst. As shown in Fig. 11, the higher the acid amount, the higher the catalytic activity. In this way one can demonstrate that the catalytic activity of NiO–ZrO2/MoO3 catalyst essentially runs parallel to the acid amount. Good correlations have been found in many cases between the acid amount and the catalytic activities of solid acids. It has been reported that the catalytic activity values of nickel-containing catalysts in ethylene dimerization as well as in butene isomerization are closely correlated with the acid amount values of the catalysts [2,5,7,38]. It is known that a low-valent nickel is favorable to chemisorb carbon monoxide [66,67], as shown in Fig. 12. This shows how the dimerization activity activated at 450 8C is poisoned by carbon monoxide. It is obvious that ethylene over 10-NiO– ZrO2/15-MoO3 catalyst is rapidly consumed without induction period. However, over 10-NiO–ZrO2/15-MoO3 catalyst preadsorbed with 1 mmol/g of CO, the pressure change with reaction time stops, indicating that the preadsorbed CO prevents the subsequent adsorption of ethylene. Therefore, it is clear that the adsorption site of CO, that is, a low-valent nickel, is also responsible for the adsorption site of ethylene. Therefore, the maximum activity for ethylene dimerization at 10 wt% NiO can be explained by suggesting that the adsorption
Fig. 12. Time-course of ethylene dimerization over 10-NiO–ZrO2/15-MoO3 catalyst evacuated at 450 8C for 1.0 h: (*) 10-NiO–ZrO2/15-MoO3 and (*) 10-NiO–ZrO2/15-MoO3 with preadsorbed CO of 1 mmol/g.
J.R. Sohn et al. / Applied Catalysis A: General 317 (2007) 216–225
amount of ethylene on the surface of catalysts is maximum at 10 wt% of NiO. As listed in Table 3, the maximum acid amount at 10 wt% NiO is probably because a low-valent nickel, Ni+ acts as a Lewis acid site. 4. Conclusions NiO–ZrO2/MoO3 catalysts were prepared by adding an aqueous solution of ammonium metatungstate to the Ni(OH)2– Zr(OH)4 powder, followed by drying and calcining at high temperatures for 1.5 h in air. The interaction between MoO3 (or NiO) and ZrO2 influences the physicochemical properties of prepared catalysts with calcination temperature. The interaction between MoO3 (or NiO) and ZrO2 delayed the phase transition of ZrO2 from amorphous to tetragonal. Since the ZrO2 stabilizes the surface MoO3 species, for the catalysts equal to or less than 15 wt%, MoO3 was well dispersed on the surface of zirconia, but for the catalysts above 15 wt% the orthorhombic phase of MoO3 was observed at the calcination temperature of 450 8C. The acid amount of 10-NiO–ZrO2/ MoO3 catalysts increased with the amount of dispersed MoO3, showing the presence of Bro¨nsted and Lewis acid sites on the surface of the catalyst. The high acid strength and high acid amount are responsible for the Mo O bond nature of complex formed by the interaction between MoO3 and ZrO2. The catalytic activities of NiO–ZrO2/MoO3 catalysts for ethylene dimerization were correlated to their acid amount as measured by the ammonia chemisorption method. Acknowledgement We wish to thank Korea Basic Science Institute for the use of Raman and X-ray instruments. References [1] F. Bernardi, A. Bottoni, I. Rossi, J. Am. Chem. Soc. 120 (1998) 7770. [2] J.R. Sohn, A. Ozaki, J. Catal. 61 (1980) 29. [3] G. Wendt, E. Fritsch, R. Scho¨llner, H.Z. Siegel, Anorg. Allg. Chem. 467 (1980) 51. [4] G.F. Berndt, S.J. Thomson, G.J. Webb, J. Chem. Soc., Faraday Trans. 1 79 (1983) 195. [5] J.R. Sohn, J.S. Lim, Catal. Today 111 (2006) 403. [6] Y.I. Pae, S.H. Lee, J.R. Sohn, Catal. Lett. 99 (2005) 241. [7] J.R. Sohn, D.C. Shin, J. Catal. 160 (1996) 314. [8] J.R. Sohn, J.S. Han, Appl. Catal. A: Gen. 298 (2006) 168. [9] T.V. Herwijnen, H.V. Doesburg, D.V. Jong, J. Catal. 28 (1973) 391. [10] J.R. Sohn, W.C. Park, H.W. Kim, J. Catal. 209 (2002) 69. [11] J.R. Sohn, W.C. Park, Bull. Korean Chem. Soc. 21 (2000) 1063. [12] K. Urabe, M. Koga, Y.J. Izumi, J. Chem. Soc., Chem. Commun. (1989) 807. [13] G. Wendt, D. Hentschel, J. Finster, R.J. Scho¨llner, J. Chem. Soc., Faraday Trans. 1 (79) (1983) 2013. [14] K. Kimura, A.J. Ozaki, J. Catal. 3 (1964) 395. [15] K. Maruya, A. Ozaki, Bull. Chem. Soc. Jpn. 46 (1973) 351. [16] M. Hartmann, A. Po¨ppl, L. Kevan, J. Phys. Chem. 100 (1996) 9906. [17] I.V. Elev, B.N. Shelimov, V.B. Kazansky, J. Catal. 89 (1984) 470. [18] H. Choo, L. Kevan, J. Phys. Chem. B 105 (2001) 6353. [19] J.R. Sohn, H.J. Kim, J. Catal. 101 (1986) 428.
225
[20] J.R. Sohn, S.Y. Lee, Appl. Catal. A: Gen. 164 (1997) 127. [21] J.R. Sohn, H.W. Kim, M.Y. Park, E.H. Park, J.T. Kim, S.E. Park, Appl. Catal. A: Gen. 128 (1995) 127. [22] D.A. Ward, E.I. Ko, J. Catal. 150 (1994) 18. [23] S.R. Vaudagna, R.A. Comelli, S.A. Canavese, N.S. Figoli, J. Catal. 169 (1997) 389. [24] L.M. Kustov, V.B. Kazansky, F. Figueras, D.J. Trichit, J. Catal. 150 (1994) 143. [25] A. Sayari, Y. Yang, X. Song, J. Catal. 167 (1997) 346. [26] K. Arata, Adv. Catal. 37 (1990) 165. [27] C.Y. Hsu, C.R. Heimbuch, C.T. Armes, B.C. Gates, J. Chem. Soc., Chem. Commun. (1992) 1645. [28] T.K. Cheung, B.C. Gates, J. Catal. 168 (1997) 522. [29] V. Adeeva, H.W. de Haan, J. Janchen, G.D. Lei, V. Schunemann, L.J.M. van de Ven, W.M.H. Sachtler, R.A. van Santen, J. Catal. 151 (1995) 364. [30] K.T. Wan, C.B. Khouw, M.E. Davis, J. Catal. 158 (1996) 311. [31] X. Song, K.R. Reddy, A. Sayari, J. Catal. 161 (1996) 206. [32] M.A. Coelho, D.E. Resasco, E.C. Sikabwe, R.L. White, Catal. Lett. 32 (1995) 253. [33] T. Hosoi, T. Shimadzu, S. Ito, S. Baba, H. Takaoka, T. Imai, N. Yokoyama, Prepr. Symp. Div. Petr. Chem., American Chemical Society, Los Angeles, CA, 1988, p. 562. [34] K. Ebitani, J. Konishi, H. Hattori, J. Catal. 130 (1991) 257. [35] M. Signoretto, F. Pinna, G. Strukul, P. Chies, G. Cerrato, S.D. Ciero, C. Morterra, J. Catal. 167 (1997) 522. [36] W. Hua, Y. Xia, Y. Yue, Z. Gao, J. Catal. 196 (2000) 104. [37] J.A. Moreno, G. Poncelet, J. Catal. 203 (2001) 453. [38] J.R. Sohn, E.S. Cho, Appl. Catal. A: Gen. 282 (2005) 147. [39] M. Hino, K. Arata, J. Chem. Soc., Chem. Commun. (1987) 1259. [40] J.R. Sohn, E.W. Chun, Y.I. Pae, Bull. Korean Chem. Soc. 24 (2003) 1785. [41] G. Larsen, E. Lotero, R.D. Para, in: Proceedings of the 11th International Congress on Catalysis, Elsevier, New York, 1996, p. 543. [42] A.S.C. Brown, J.S.J. Hargreaves, S.H. Taylor, Catal. Lett. 57 (1999) 109. [43] J.R. Sohn, S.H. Lee, Appl. Catal. A: Gen. 266 (2004) 89. [44] J.R. Sohn, D.H. Seo, S.H. Lee, J. Ind. Eng. Chem. 10 (2004) 309. [45] J.R. Sohn, S.H. Lee, W.C. Park, H.W. Kim, Bull. Korean Chem. Soc. 25 (2004) 657. [46] G. Larsen, E. Lotero, L.M. Petkovic, D.S. Shobe, J. Catal. 169 (1997) 67. [47] P. Afanasiev, C. Geantet, M. Breysse, G. Coudurier, J.C. Vedrine, J. Chem. Soc., Faraday Trans. 190 (1994) 193. [48] J.R. Sohn, J. Ind. Eng. Chem. 10 (2004) 1. [49] J.R. Sohn, S.G. Cho, Y.I. Pae, S. Hayashi, J. Catal. 159 (1996) 170. [50] Z. Liu, Y. Chen, J. Catal. 177 (1998) 314. [51] B. Zhao, X. Wang, H. Ma, Y. Tang, J. Mol. Catal. A: Chem. 108 (1996) 167. [52] K. Chen, S. Xie, E. Iglesia, A.T. Bell, J. Catal. 189 (2000) 421. [53] M.R. Smith, U.S. Ozkan, J. Catal. 141 (1993) 124. [54] G. Mestl, T.K.K. Srinivasan, Cat. Rev. Sci. Eng. 40 (1998) 451. [55] P. Dufresne, E. Payen, J. Grimblot, J.P. Bonnelle, J. Phys. Chem. 85 (1981) 2344. [56] H. Hu, I.E. Wachs, J. Phys. Chem. 99 (1995) 10897. [57] R.D. Roark, S.D. Kohler, J.G. Ekerdt, D.S. Kim, I.E. Wachs, Catal. Lett. 16 (1992) 77. [58] C.H. Schild, A. Wokaun, R.A. Ko¨ppel, A. Baiker, J. Catal. 130 (1991) 657. [59] J.R. Sohn, I.J. Doh, Y.I. Pae, Langmuir 18 (2002) 6280. [60] M. Scheithauer, R.K. Grasselli, H. Kno¨zinger, Langmuir 14 (1998) 3019. [61] Y.C. Xie, Y.Q. Tang, Adv. Catal. 37 (1990) 1. [62] L. Litteti, I. Nova, G. Ramis, L. DallAcqua, G. Busca, E. Giamello, P. Forzatti, F. Bregani, J. Catal. 187 (1999) 419. [63] F.G.A. Olah, G.K.S. Prakash, J. Sommer, Science 206 (1979) 13. [64] K. Tanabe, M. Misono, Y. Ono, H. Hattori, New Solid Acids and Bases, Elsevier Science, Amsterdam, 1989, p. 185. [65] J.R. Sohn, S.G. Ryu, Langmuir 9 (1993) 126. [66] K. Kimura, H. A-I, A. Ozaki, J. Catal. 18 (1970) 271. [67] J.R. Sohn, W.C. Park, Appl. Catal. A: Gen. 239 (2003) 269.