Selective Oxidation of Isobutane over NH 4 CsFePVMoAs Heteropolycompound Catalysts

Transcription

1 Journal of Natural Gas Chemistry 14(2005)54 60 Selective Oxidation of Isobutane over NH 4 CsFePVMoAs Heteropolycompound Catalysts Tiejun Cai, Zhengjun Fang, Qian Deng, Zhenshan Peng Department of Chemistry and Chemical Engineering, Hunan University of Science and Technology, Xiangtan , China [Manuscript received January 25, 2005; revised March 09, 2005] Abstract: A series of NH 4Cs 1.5Fe 0.08H xpvmo 11As ao y heteropolycompound catalysts for the selective oxidation of isobutane, having Keggin structure, were synthesized by using co-precipitation method. The catalysts were characterized by FT-IR, H 2-TPR, TG-DTA, SEM and XRD. Effects of the As content, reaction time, reaction temperature and molar ratio of isobutane to oxygen in feedstock on the activity and selectivity of the catalyst were investigated. The activation energy of the catalysts was measured by kinetics researches. Results showed the introduction of Cs + into the catalysts shortened the stable period of them and enhanced their catalytic activity for the selective oxidation of isobutane. The highest conversion of isobutane and the total selectivity to liquid products were 18.6% and 81.2%, respectively, which were obtained at 380 with a space velocity of 975 h 1 over the NH 4Cs 1.5Fe 0.08H xpvmo 11As 0.3O y heteropolycompound catalyst. It is confirmed that completely oxidized products were controlled well. Key words: heteropolycompound catalyst, isobutane, selective oxidation, methacrylic acid, methacrolein 1. Introduction A major challenge to petroleum and petrochemical industries is how to make full use of alkanes. Lower alkanes are of great interest due to the cheapness and abundance. The utilization of molecular oxygen for the oxidation of hydrocarbons is a rewarding goal because molecular oxygen has the highest content of active oxygen and forms no by-products. Therefore, the reaction of lower alkanes with molecular oxygen is very interesting. The catalytic function of heteropolycompounds in the solid state has attracted much attention because their redox and acidic properties can be controlled at atomic/molecular levels [1]. Recently, the heteropolycompound catalysts for the selective oxidation of alkanes are mainly those partially salified with cesium and partially reduced heteropoly acid catalysts. The high activity of the former may be related to their strong acidity and high specific area, because the rate determining step for the selective oxidation of alkanes is the apparent reaction between alkane molecular and lattice oxygen, and higher specific surface area can improve the catalytic activity [2]. While, the latter can increase the concentration of Mo 5+ to change the activity of catalyst [3]. Cavani et al. [4] observed the selectivity for methacrylic acid (MAA) under isobutane-rich conditions was higher than that under oxygen-rich conditions with (NH 4 ) 3 PMo 12 O 40, because alkane-rich conditions makes reduction of Mo 6+ to Mo 5+ possible. In the case, the catalyst had good performance for reaction time, which were much longer than 100 h, and the finial yield of MAA reached 2.6% 2.8%. In this paper, a series of NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 As a O y catalyst were used for the oxidation of isobutane to methacrylic acid in a single step, the influences of the content of As and the reaction conditions were also studied. Corresponding author. Tel: (0732) , tjcai@163.com Supported by Scientific Research Fund of Hunnan Provincial Education Department (03C515) and Hunan Provincial Natural Science Foundation of China (04jj6003)

2 Journal of Natural Gas Chemistry Vol. 14 No Experimental 2.1. Catalysts preparation The catalysts of heteropoly compound with Keggin structure were prepared according to our previous method [5]. The typical preparation procedure was as follows: a certain amount of MoO 3, P 2 O 5, As 2 O 5 and NH 4 VO 3 were mixed in distilled water, and the resultant mixture was refluxed at for 72 h to form a homogeneous orange yellow solution. Then Fe(NO 3 ) 3 (0.08 mol/dm 3 ) was added dropwise in a desired stoichiometry to the solution, followed by the addition of an aqueous solution of Cs 2 CO 3 (0.08 mol/dm 3 ) at 60. The resulting suspension was evaporated to dryness at 100, and its composition was Cs 1.5 Fe 0.08 H x PVMo 11 As 0.3 O y, NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 As 0.2 O y, NH 4 Cs 1.5 Fe H x PVMo 11 As 0.3 O y and NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 - As 0.4 O y Catalyst characterization IR spectra of the catalysts were taken on a Perkin- Elmer 2000 IR Spectrometer by KBr tablet method. The H 2 -TPR was carried out on a Chembet-3000 Adsorption Instrument. The samples were pretreated at 150 under pure N 2 atmosphere for 1 h. The TPR experiments were carried out under an atmosphere of mixed H 2 -N 2 containing 15%H 2 (v/v) with a temperature increasing rate of 10 /min. The characterization by TG and DTA was carried out on a PCT-IA diffenential thermobalance with a heating rate of 10 /min. XRD patterns of the catalysts were recorded by using a MSAL XD-2 diffractometer with Cu K α radiation. The 2θ angles were scanned from 5 o to 65 o at a rate of 4 o /min. The surface structure of the catalysts was studied by SEM on a KYKY-3800 scanning electron microscope. products. Reactants and reaction products were analyzed by on-line gas chromatography equipped with a porapak-q column. The experimental data were all collected after reacting for 8 h unless stated otherwise. 3. Results and discussion 3.1. Characterization of the catalysts The catalysts were characterized by FT-IR, TG- DTA, TPR, SEM and XRD, and their results are shown in Figure 1 Figure 5. Figure 1 shows the IR spectra of the as-prepared NH 4 Cs 1.5 Fe 0.08 H x PV- Mo 11 As 0.3 O y catalyst and that after reaction at 380 for 48 h. Four strong absorption peaks near 1061, 964, 864 and 791 cm 1, are attributed to the vibration of P O, Mo=O and Mo O Mo bonds of different locations, respectively. These results show that both of the catalysts contain heteropoly anion of the Keggin structure [6], and NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 As 0.3 O y has good thermal stability in the course of catalytic reaction. Compared with IR spectrum of H 3 PMo 12 O 40, NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 As 0.3 O y has a shoulder at 1074 cm 1 caused by a splitting of antisymmetric P O vibration, corresponding to the lowering of the symmetric of the central PO 4 tetrahedron due to the substitution of one Mo in [PMo 12 O 40 ] 3 by one V atom [7]. A slight decrease in the peak intensity at 864 and 791 cm 1 after reaction is probably due to the reduction of Mo 6+ to Mo 5+ in Keggin anion. The slight reduction is consistent with the change of catalyst color to blue green during the catalytic tests [8] Catalytic tests Catalytic reactions were carried out in a fixedbed stainless steel tubular microreactor and the fixed mass of the catalyst was 1.2 g. The feed consisted of a mixture of isobutane/oxygen/nitrogen/water with a space velocity of 975 h 1. All the catalysts were treated at 320 for 1 h under 10 ml/min N 2 atmosphere before reaction. Experiments were carried out in the temperature range of in order to achieve the highest selectivity to partial oxidation Figure 1. IR spectra of catalysts (1) NH 4 Cs 1.5 Fe 0.08 H xpvmo 11 As 0.3 O y(as-prepared) (2) NH 4 Cs 1.5 Fe 0.08 H xpvmo 11 As 0.3 O y (after reaction)

3 56 Tiejun Cai et al./ Journal of Natural Gas Chemistry Vol. 14 No The TG and DTA curves of four catalysts are shown in Figure 2. All the curves of TG have two weight loss peaks, the first one below 200, the other one in The first weight loss of Cs 1.5 Fe 0.08 H x PVMo 11 As 0.3 O y catalyst is more than that of NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 As 0.2 O y, NH 4 Cs Fe 0.08 H x PVMo 11 As 0.3 O y and NH 4 Cs 1.5 Fe 0.08 H x PV- Mo 11 As 0.4 O y catalysts, while the second weight loss of Cs 1.5 Fe 0.08 H x PVMo 11 As 0.3 O y catalyst is less than that of NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 As 0.2 O y, NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 As 0.3 O y and NH 4 Cs Fe 0.08 H x PVMo 11 As 0.4 O y catalysts. Figure 2. TG-DTA curves of heteropolycompound catalysts (1) Cs 1.5 Fe 0.08 H xpvmo 11 As 0.3 O y (2) NH 4 Cs 1.5 Fe 0.08 H xpvmo 11 As 0.2 O y (3) NH 4 Cs 1.5 Fe 0.08 H xpvmo 11 As 0.3 O y (4) NH 4 Cs 1.5 Fe 0.08 H xpvmo 11 As 0.4 O y The DTA curve can be compared with the weight loss on the TG curves. Four catalysts all have two endothermic peaks below 200, which are attributed to the dehydration of the molecular water and the crystallized water. The endothermic peaks of the latter three catalysts are not completely separated yet. The exothermic peaks between 400 and 500 are likely correlated to the loss, the oxidation of the remaining ammonia molecules and the decomposition of the catalysts [7]. Cs 1.5 Fe 0.08 H x PVMo 11 As 0.3 O y catalyst has only an exothermic peak in this field, while other three have two uncompletely separated exothermic peaks. These phenomena are caused by the existence of NH + 4, because heteropolycompounds have only less crystallized water when NH + 4 replaces H+ as countercation. It leads to the decrease in weight loss below 200. In addition, the decomposition or oxidation of NH + 4 above 400 explained that the weight loss peaks of NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 As 0.2 O y, NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 As 0.3 O y and NH 4 Cs Fe 0.08 H x PVMo 11 As 0.4 O y catalysts are more obvious than Cs 1.5 Fe 0.08 H x PVMo 11 As 0.3 O y catalyst relatively, and their corresponding DTA curves have two uncompletely separated exothermic peaks. By analysis of the TG-DTA curves of NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 As 0.2 O y, NH 4 Cs 1.5 Fe H x PVMo 11 As 0.3 O y and NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 - As 0.4 O y catalyst, it is found that the introduction of a small quantity of As into the catalyst may be not beneficial to the thermal stability of catalysts, which is coincident with our previous paper [9]. The TPR profiles of four catalysts are shown in Figure 3 and Table 1. The reduction peaks of these four catalysts appear at 644.5, 646.0, and 658.3, respectively. Because heteropolycompounds are already decomposed to oxides at these temperature, it is suggested that the reduction peaks are attributed to redox of mixed oxides. We analysed H 2 -TPR of four catalysts at different speeds of raising temperature and researched kinetics course. If the catalyst surface is assumed to be symmetrical and no absorption happens on it again, profiles can be illustrated through 2lnT m -lnβ to 1/T m. The results from H 2 -TPR and kinetics show that the reaction activation energy of catalysts can be calculated [10] shown in Table 1. Compared with the Cs 1.5 Fe 0.08 H x PVMo 11 As 0.3 O y catalyst, the activation energy of the NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 As 0.3 O y catalyst reduces 7 kj/mol after introducing NH + 4. According to the value of activation energy, we may assert that oxygen species of catalysts surface is O [11], so the redox mechanism probably is:

4 Journal of Natural Gas Chemistry Vol. 14 No Isobutane+O +Mo 6+ Oxidized product+mo 5+ + Mo /2O 2 O +Mo 6+ ( represents lattice oxygen vacancy) of the phenomenon is probably due to the sintering of the catalyst. Figure 3. H 2 -TPR profiles of heteropoly compound catalysts (1) Cs 1.5 Fe 0.08 H xpvmo 11 As 0.3 O y (2) NH 4 Cs 1.5 Fe 0.08 H xpvmo 11 As 0.2 O y (3) NH 4 Cs 1.5 Fe 0.08 H xpvmo 11 As 0.3 O y (4) NH 4 Cs 1.5 Fe 0.08 H xpvmo 11 As 0.4 O y Many researchers [12] indicate that lattice oxygen plays an important role in the selective oxidation of alkanes. Increasing the density of Mo 5+ can accelerate the migration of the lattice oxygen, and lead to high selectivity of goal products. Table 1. Reduction peak temperatures and reaction activation energy of four catalysts Catalyst T m/ E a/(kj/mol) Cs 1.5 Fe 0.08 H xpvmo 11 As 0.3 O y NH 4 Cs 1.5 Fe 0.08 H xpvmo 11 As 0.2 O y NH 4 Cs 1.5 Fe 0.08 H xpvmo 11 As 0.3 O y NH 4 Cs 1.5 Fe 0.08 H xpvmo 11 As 0.4 O y The SEM micrographs of the catalysts are shown in Figure 4. From Figure 4, we can see that the heteropolycompounds are porous and the crystals are relatively intact. Crystallites in Cs Fe 0.08 H x PVMo 11 As 0.3 O y are bigger than that in NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 As 0.3 O y due to the existence of NH + 4. Compared with micrograph (a), the size of particles in micrograph (c) increases, the case Figure 4. SEM micrographs of the catalysts (a) Cs 1.5 Fe 0.08 H xpvmo 11 As 0.3 O y magnified by , (b) NH 4 Cs 1.5 Fe 0.08 H xpvmo 11 As 0.3 O y (as-prepared) magnified by , (c) NH 4 Cs 1.5 Fe 0.08 H xpvmo 11 As 0.3 O y (after reaction) magnified by XRD patterns of the Cs 1.5 Fe 0.08 H x PVMo 11 As O y, the as-prepared NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 As O y and NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 As 0.3 O y after reaction for 48 h are shown in Figure 5. Similar patterns are shown in all cases. The appearance of peaks at 2θ=23 o could apparently be related to the presence of (NH 4 ) 4 PMo 11 VO 40. A slight change in the position of diffraction peaks but no new peak appears comparing pattern (1) and pattern (2). It indicates that NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 As 0.3 O y catalyst is steady

5 58 Tiejun Cai et al./ Journal of Natural Gas Chemistry Vol. 14 No in reaction. Therefore, it proves further that the catalyst has good thermal stability under reaction conditions Choosing the optimum reaction conditions The effects of reaction time, C 4 H 10 /O 2 ratio, and reaction temperature on NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 - As 0.3 O y catalyst were investigated. The optimum reaction conditions are obtained and the results are shown in Figure 6 Figure Reaction time Figure 5. XRD patterns of the catalysts (1) Cs 1.5 Fe 0.08 H xpvmo 11 As 0.3 O y, (2) NH 4 Cs 1.5 Fe 0.08 H xpvmo 11 As 0.3 O y (as-prepared), (3) NH 4 Cs 1.5 Fe 0.08 H xpvmo 11 As 0.3 O y (after reaction) Figure 6 shows the time course of the oxidation of isobutane at 380. It can be observed that the conversion and selectivity reached almost constant after 8 h. In this course, the conversion of isobutane decreases gradually until being steady, and declines by a large margin in the first 5 h. The selectivity of AcOH is on the rise, but drop to some extent after 5 h, while the selectivity of AA, MAL and MAA keep ascendant trend until being steady all the time Catalytic activity for the isobutane oxidation Inf luence of As content Selective oxidation of isobutane over NH 4 Cs Fe 0.08 H x PVMo 11 As a O y was tested and the results are summarized in Table 2. The main oxidized products are MAA, methacrolein (MAL), acrylic acid (AA), acetic acid (AcOH), and CO x. It is found that as the content of As increases, the conversion of isobutane rises gradually, while the selectivities for AcOH, AA, MAL and MAA take the lead in rising and then reducing. The optimal selectivity of MAA is achieved at the As content a=0.3. The structure of Keggin-type heteropolycompounds is distorted when As atom partially replaces P atom in Keggin anion, because the radius of As atom is greater than radius of P atom. a Table 2. Selective oxidation of isobutane over NH 4 Cs 1.5 Fe 0.08 H xpvmo 11 As ao y catalysts at 380 Conversion Selectivity (%) (%) MAA MAL AA AcOH C 4 /O 2 /N 2 /H 2 O=2/1/4/1 Figure 6. Time course of the oxidation of isobutane catalyzed by NH 4 Cs 1.5 Fe 0.08 H x- PVMo 11 As 0.3 O y at 380 (1) X(Isobutane), (2) S(AcOH), (3) S(AA), (4) S(MAL), (5) S(MAA) (C 4 /O 2 /N 2 /H 2 O=2/1/4/1) C 4 H 10 /O 2 ratio Effect of C 4 H 10 /O 2 ratio on the selective oxidation of isobutane is shown in Figure 7. It is found that with the increase of C 4 H 10 /O 2 ratio, the conversion of isobutane and the selectivities for CO x and AcOH are

6 Journal of Natural Gas Chemistry Vol. 14 No reduced, while the selectivities for goal products MAL and MAA are increased. The selectivities to MAL and MAA are the highest at C 4 H 10 /O 2 =2/1(mol/mol). It was reported that the concentration of isobutane in the feed gas influences the surface active sites of the catalyst [13]. Isobutane-rich condition can not only promote the reduction of Mo 6+, but also suppress the total oxidation of products. However, the isobutanerich condition results in a low conversion of isobutane though the selectivity of MAA may be higher. Figure 8. Effect of reaction temperature on oxidation of isobutane catalyzed by NH 4 Cs Fe 0.08 H xpvmo 11 As 0.3 O y (1) X(Isobutane), (2) S(AcOH), (3) S(AA), (4) S(MAL), (5) S(MAA) (C 4 /O 2 /N 2 /H 2 O=2/1/4/1) 4. Conclusions Figure 7. Effect of isobutane/oxygen ratio on oxidation of isobutane catalyzed by NH 4 - Cs 1.5 Fe 0.08 H xpvmo 11 As 0.3 O y at 380 (1) X(Isobutane), (2) S(AcOH), (3) S(AA), (4) S(MAL), (5) S(MAA) Temperature Figure 8 shows the temperature dependency of the conversion and selectivities upon the NH 4 Cs Fe 0.08 H x PVMo 11 As 0.3 O y catalyst. It is found that the catalytic activity and selectivity to products are greatly affected by temperature. An appropriate enhancement of temperature makes for the improvement of catalytic performance. The conversion of isobutane and the selectivities to AcOH and AA monotonously increased with increasing the reaction temperature. The selectivity to MAL monotonously decreased with increasing the reaction temperature, and the selectivity to MAA reached the maximum 48.7% at 380. NH 4 Cs 1.5 Fe 0.08 H x PVMo 11 As a O y Keggin-type heteropolycompound catalysts were used for the selective oxidation of isobutane. Results show that the catalyst has good stability. The introduction of NH + 4 into the catalysts leads to a decrease in the reaction activation energy. Because the catalyst partially was reduced with the decomposition of NH + 4, its redox properties was modified. The cesium forms an alkaline salt of HPA that plays a role of support. This enables an efficient dispersion of the active phase, which is a key parameter for the development of an active catalyst for the selective oxidation of isobutane [14]. The modification with As makes for the enhancement of catalytic activity, but excessive content of As accelerates deeply oxidation. Results prove totally oxidized products were controlled effectively, and the total selectivity of liquid products reached 81.2%, because large cations such as Cs + and NH + 4 replace H + as counter-cation, and As atom replaces P atom in Keggin anion. References [1] Joon-Seok M, Noritaka M. Catal Today, 2001, 66: 45 [2] Mao X, Yin Y Q, Zhong B K. Fenzi Cuihau (J Mol

7 60 Tiejun Cai et al./ Journal of Natural Gas Chemistry Vol. 14 No Catal), 2000, 14(6): 486 [3] Jiang H Sh, Mao X, Xie S J et al. J Mol Catal A: Chem, 2002, 185: 147 [4] Fabrizio C, Roberto M, Anne P et al. Surface chemistry and catalysis, 2000, 3: 523 [5] Cai T J, Yu Ch L, Deng Q et al. Chin J Chem, 2004, 22: 831 [6] Wang E B, Hu Ch W, Xu L et al. Introduction to Polyacid Chemistry. Beijing: Chem Ind Press, [7] Laronze N, Marchal-Roch C, Guillou N et al. J Catal, 2003, 220: 172 [8] Mizuno N, Tateishi M, Iwamoto M. J Catal, 1996, 163: 87 [9] Cai T J, Yu Ch L, Deng Q et al. Chin J Catal, 2003, 24(12): 954 [10] Chen S Y, Sun Y H, Ding Y J. Adsorption and Catalysis. Zhengzhou: Henan Sci Technol Press, [11] Lu G Zh, Wang R. Chem, 1993, 10: 35 [12] Zhang X, Wan H L, Weng W Zh et al. Catal Lett, 2003, 87: 233 [13] Jiang H Sh, Mao X, Xie S J et al. J Mol Catal A: Chem, 2002, 185: 143 [14] Sultan M, Paul S, Fournier M et al. Appl Catal A: Gener, 2004, 259: 141