Catalytic activity of liquid phase prepared Cu-Zn-Al catalyst for CO hydrogenation in a fixed bed reactor

Transcription

1 Indian Journal of Chemistry Vol. 51A, December 2012, pp Catalytic activity of liquid phase prepared Cu-Zn-Al catalyst for CO hydrogenation in a fixed bed reactor Chunhui Luan a, b, Anrong Zhang b, Xiaodong Wang a, Wei Huang a, * & Lihua Yin a a Key Laboratory of Coal Science and Technology Ministry of Education and Shanxi Province, Taiyuan 0300, Shanxi, China huangwei@tyut.edu.cn b College of Chemistry and Chemical Engineering, Taiyuan University of Technology, Taiyuan 0300, Shanxi, China Received 9 March 2012; revised and accepted 26 November 2012 Cu-Zn-Al catalyst prepared by a complete liquid phase technology for CO hydrogenation has been investigated in a fixed bed reactor in terms of catalytic performance and compared with catalysts prepared by a traditional method such as co-precipitation. The catalysts are characterized by XRD, NH 3 -TPD, H 2 -TPR and XPS. Results show that the activity of the liquid phase prepared catalyst is much lower in fixed bed reactor than in a slurry reactor. It was also much lower than that of the catalyst prepared by co-precipitation, although its stability is higher. XPS and XRD characterization reveal that there are obvious differences in the structure, morphology and surface state of the catalysts prepared by the two methods. In the catalyst prepared by liquid phase technology, Cu exists in the form of metallic Cu even though the catalyst is unreduced and there are more acid sites and more carbon on its surface than on the surface of the co-precipitation catalyst. It is concluded that there is a surface carbon film covering the liquid phase catalyst, and the carbon film is an important reason for the low activity when the catalyst is applied to fixed bed reactor. Coke burning off, an effective way for removing the surface carbon, improves the catalyst activity in a fixed reactor. Keywords: Catalysts, Liquid phase catalysts, Fixed bed reactor, Carbon monoxide hydrogenation, Copper, Zinc, Aluminium Much attention has been paid to synthesis of lowcarbon alcohol or ether from syngas due to its potential use as industrial organic chemicals and clean fuel substitute for LPG and diesel oil. Among the catalysts for synthesizing low-carbon alcohol or ether, the low cost Cu-Zn-Al catalysts, with high activity and fine selectivity for alcohol or ether, occupies an important place, especially those prepared by sol-gel and co-precipitation methods 1-5. To rapidly deactive the catalyst in a slurry reactor for one-step synthesis of dimethyl ether (DME) from syngas, in view of the concept that the formation and growth of catalyst material should take place in the same environment where the catalyst is finally used, a novel method for preparing catalyst to be used in a slurry reactor, i. e., a totally liquid phase technique, has been developed by our group. The catalyst displayed good performance in a slurry reactor, with similar activity and higher stability than the catalysts prepared by traditional methods, such as co-precipitation 6-8. The Cu Zn Al catalyst prepared by this novel method for direct synthesis of DME from syngas in slurry reactor showed excellent stability during the reaction period of 440 h (ref. 8). Theoretical investigations indicate that the preparation environment, in other words solvent properties, influence the structure of the catalyst, adsorption and desorption of reactants as well as the reaction mechanism 9, 10. Therefore, investigating the performance of the liquid phase prepared catalyst in fixed bed reactor is important for enriching the theory of catalyst preparation. Herein, the performance of Cu-Zn-Al catalyst prepared by liquid phase technology for CO hydrogenation in a fixed bed reactor has been systematically evaluated and compared with that of the catalyst prepared by the conventional co-precipitation method. Materials and Methods Aluminum isopropoxide (AIP), cupric nitrate trihydrate (Cu(NO 3 ) 2 3H 2 O), zinc nitrate hexahydrate (Zn(NO 3 ) 2 6H 2 O), nitric acid (HNO 3 ), aluminum nitrate nonahydrate (Al(NO 3 ) 3 9H 2 O), sodium carbonate (Na 2 CO 3 ) and caustic soda (NaOH) were procured from Kermel, China. All chemicals were of analytical grade.

2 1664 INDIAN J CHEM, SEC A, DECEMBER 2012 Preparation and characterization of the catalysts To prepare the liquid phase catalyst, AIP (55.1 g) was added to 300 ml of distilled water and subjected to hydrolysis at 353 K for 1.5 h. It was then peptized with 2.4 ml of nitric acid at 368 K with refluxing under vigorous stirring for 1 h to obtain a stable alumina sol. A mixed solution containing the desired amounts of Cu(NO 3 ) 2 3H 2 O (32.8 g) and Zn(NO 3 ) 2 6H 2 O (19.8 g) was added dropwise, maintaining the optimal molar ratio of the reactant (Cu:Zn:Al) at around 2:1:2. The mixture was then vigorously stirred for 10 h to form a blue gel, which was aged at room temperature for 10 days. After that the gel was dispersed into 0.5 ml of Span 80 and 300 ml of liquid paraffin in a mechanical emulsifier. The mixture was transferred to a 1000 ml threenecked flask, heated up to 573 K and kept for 10 h under nitrogen atmosphere and vigorous stirring. After cooling down to room temperature, the mixture was filtered, extracted with petroleum ether (boiling point range from 333 K K ) for 4 days and finally dried at 343 K for 5 h to obtain the liquid phase catalyst, which was labeled as A. The co-precipitation catalyst was prepared in a typical procedure using Al(NO 3 ) 3 9H 2 O as Al source and Na 2 CO 3 as precipitant. The Cu and Zn sources were the same as for A. The molar ratio of Cu:Zn:Al was the same as in the case of A and the molar ratio of CO 2-3 to total metal ions was in stoichiometric ratio. Na 2 CO 3 was added to a mixture of Cu(NO 3 ) 2, Zn(NO 3 ) 2, Al(NO 3 ) 3 solutions and a mixture of solution and precipitate was obtained, to which NaOH aqueous solution was added to adjust the ph value to ~8.0. After further reaction at 323 K for 4 h under vigorous stirring, the precipitate was filtered and washed with distilled water to remove residual Na +. The catalyst precursor was obtained and divided into two parts. One part was dried in an oven at 373 K for 4 h and finally calcined in air at 573 K. This sample was labeled as B. The other part was subjected to heat treatment in paraffin according to the procedure in complete liquid phase technology to obtain a catalyst between A and B, which was labeled as C. To remove the surface carbon, sample A was crushed, sieved, and loaded on the fixed bed reactor. Nitrogen containing 10% O 2 was passed through the fixed bed reactor at a flow rate of 100 ml/min and the reactor was heated up to 573 K for 10 h, to remove paraffin adsorbed on catalyst surface. The sample obtained by this method was labeled as AO. The XRD patterns of all the samples were recorded on a powder X-ray diffractometer (Shimadzu XRD- 6000, Japan) which was operated at 40 kv and 100 ma, using a Cu Kα radiation source (λ = nm) and 2θ range Acidity was determined by temperature programmed adsorption of ammonia, using a TP-5000 instrument (XQ, Tianjin) combined with a QIC-20 mass spectrometer (HEDIN, UK) at the temperature range from 323 K to 1073 K. In a typical analysis, a given amount of sample (0.1 g) was purged at 553 K for 30 min under He flow (30 ml/min) prior to pulse adsorption of ammonia at 323 K for 30 min. The sample was purged with helium for 30 min to remove the weakly and physically adsorbed NH 3 on the catalyst surface. Then the sample was heated from 323 K to 1023 K at a rate of 10 K/min under a flow of helium carrier gas (30 ml/min), and the amount of ammonia desorbed from the catalyst was then measured and recorded as a function of temperature. Temperature-programmed reduction (H 2 -TPR) experiments were performed to determine the reducibility of the oxides. TPR experiments were carried out on a TP-5000 instrument (XQ, Tianjin). Prior to the TPR experiments, the samples were pre-treated under He flow up to 423 K with a heating rate of 10 K/min and kept for 30 min to remove the adsorbed water and other contaminants, followed by cooling to 323 K. The reducing gas containing 5 % H 2 / 95 % N 2 mixture at a flow rate of 30 ml/min was passed over the samples. The temperature range investigated was from K with a heating rate of 10 K/min. The consumption of H 2 was quantitatively measured by a thermal conductivity detector (TCD). The chemical and bonding states were characterized by XPS spectra recorded on a Axis Ultra instrument (Thermo VG Scientific, UK) using monochromatic Al Kα radiation (150 W, 15 kv, 16.6 ev). The vacuum in the spectrometer was Pa. Binding energies were calibrated relative to Zn 2p 3/2 peak at ev. The contents of the surface elements were calculated by the area of characteristic peak and sensitivity factor. Catalytic activity Catalytic performance for CO hydrogenation was evaluated in a stainless steel fixed-bed reactor, using 2 ml of sample, which was crushed and sieved to grain size mesh. Prior to this, the catalyst was reduced in a stream of 10 % H 2 /N 2 at 553 K for 10 h under atmospheric pressure. Activity was tested under

3 LUAN et al.: CATALYTIC ACTIVITY OF LIQUID PHASE Cu-Zn-Al CATALYST FOR CO HYDROGENATION 1665 Table 1 Catalytic activity of the Cu-Zn-Al catalysts. [Reaction cond.: T = 553 K, P = 4 Mpa, GHSV = 3000h -1 ] Sample Time (h) Conv. of CO (%) S (DME) S (MeOH) S (EtOH) S (CnHm) S (CO 2 ) A B C AO the following conditions: T = 553 K; P = 4.0 MPa; GHSV=4500 h -1 ; molar feed composition of H 2 :CO = 2:1. The effluent products were analyzed by an online gas chromatograph (GC-2060) equipped with two columns, (i) TDX-01 connected to a thermal conductivity detector (TCD) for H 2, CO and CO 2, and, (ii) Porapack-Q connected to a flame ionization detector (FID) for flammable products such as methanol, dimethyl ether (DME) and hydrocarbon byproducts. CO conversion and product selectivity were calculated based on total carbon balance. The contents of various components were quantified by external standard method Results and Discussion Catalytic activity The catalytic performance of different samples for CO hydrogenation is shown in Table 1. Catalysts A and C, which were given heat treatment in liquid paraffin, exhibited low activity, with CO conversion less than 10 %, while CO conversion of catalysts B and AO were obviously higher, reaching 20 % or more. Furthermore, AO gave somewhat higher CO conversion than B. Comparison of the selectivity of the above catalysts shows that catalyst A and C generated larger amount of hydrocarbon and CO 2, and very little DME. Catalyst B mainly resulted in methanol with higher selectivity and some DME, while in the case of AO, the selectivity of methanol reached 50 % on the second day. The DME selectivity and total selectivity of a lcohol and ether were also higher than those for catalyst B. Additionally, the liquid phase heat-treatment catalysts A and C also Fig. 1 XRD patterns of catalysts prior to reduction. [1, A; 2, AO; 3, B; 4,C]. showed a relatively high selectivity to ethanol, which reveals that the liquid phase heat treatment favored ethanol production. As compared with catalyst A,catalysts B, C and AO were inferior in stability and their relative decrease in CO conversion was 9.51, 9.95, and 9. % per day, respectively, based on the CO conversion on the second day. Decrease in activity may be attributed to the sintering of copper 11, 12. Obviously, one of main advantages of liquid phase technology is its ability to overcome the problem of poor thermal stability. XRD patterns The XRD patterns of the catalysts prior to reduction are shown in Fig. 1. The active component copper in

4 1666 INDIAN J CHEM, SEC A, DECEMBER 2012 the catalysts prepared by different methods was present in different forms. For liquid phase prepared catalyst A, copper species existed entirely in the metallic form, which is not consistent with our previous work 8. For co-precipitation catalyst B and coke burnt off catalyst AO, cupric oxide was in the majority. For co-precipitation combined with liquid phase heat treatment catalyst C, the common features of A and B were reflected, i. e. copper and cuprous oxide coexisted, which was the result of partial reduction. Obviously, one of the most distinctive features of liquid phase prepared catalyst was the reduction of copper species during heat treatment, and the reduction was also closely related to the preparation methods of precursor, i. e., to the structure of the precursor and existing forms of the component. The precursor prepared by sol-gel method, i.e., catalyst A, was reduced totally after liquid phase heat treatment, but the catalyst prepared by co-precipitation method, i. e., catalyst C was not reduced. The grain size and interplanar spacing of copper calculated based on peak broadening using Scherrer equation were studied (Supplementary Data, Table S1). The difference in the existing forms of copper species makes it difficult to compare the grain size of copper species in the catalysts prepared by different methods. However, from the comparison of interplanar spacing with JCPDS X-Ray Powder Diffraction Standards Data Pile, it can be seen that the interplanar spacings of A and AO are larger than the JCPDS data and that of B are smaller, while the interplanar spacing of C is in between. These results reveal that copper species in the liquid phase prepared catalyst was surrounded by other atoms, most probably by carbon atoms, which proves further that the existing form of the catalyst was related closely with the preparation method and the method of heat treatment of the precursor. No diffraction peaks of alumina and zinc oxide were observed, suggesting the amorphous forms of these species. XPS characterization XPS survey scans for the catalysts are shown in Fig. 2. Notably, a shake-up peak of Cu 2p 3/2 was detected for catalysts B and AO. Generally, for copper species, the presence of the shake-up peak is considered to be an indication of Cu 2+ species 13. Thus, the peak at ev was assigned to CuO. Cu 0 species and Cu + species in Cu 2 O, which gave Cu 2p 3/2, had similar peak shape and binding energy value. The Auger parameter (AP) a (a =Eb(Cu 2p 3/2 ) + E k (CuLVV)) was therefore introduced to distinguish between the two 13, 14. The a value of catalyst C was ev, which was belonged to Cu + species, consistent with XRD results of the coexistence of Cu 0 and Cu 2 O in catalyst C, but Cu + species was more on the surface. The Auger peak lines of Cu species were not observed on catalyst A, which is attributed to the low content of Cu species on the surface of catalyst A. The surface compositions of different catalysts are listed in Table 2. The carbon content on the surface of sample A was much greater than on the surface of any other catalyst. The surface carbon of catalyst C was also much higher than that of B and AO, indicating that the heat treatment in paraffin resulted in carbon deposition on the surface and the deposition amount was related to the preparation method of precursor. The amount of carbon deposition of sol-gel precursor catalyst A was twice that of the co-precipitation precursor catalyst C. It can also be seen from Table 2 that the Cu/Zn and Cu/Al Fig. 2 Cu 2p 3/2 -XPS spectra of different catalyst samples. [1, A;2, AO; 3, B; 4,C]. Table 2 XPS analysis of Cu-Zn-Al catalysts Sample C O Zn Cu Al Cu/Al Cu/Zn Zn/Al A B C AO

5 LUAN et al.: CATALYTIC ACTIVITY OF LIQUID PHASE Cu-Zn-Al CATALYST FOR CO HYDROGENATION 1667 Fig. 3 H 2 -TPR profiles of Cu-Zn-Al catalysts. [1, A; 2, AO;3, B; 4, C]. atomic ratios of co-precipitation catalyst B were much larger than that of the other catalysts, indicating that preparation method had a large influence on distribution of copper on the surface but had little effect on distribution of zinc and aluminum. Comparing Cu/Al, Cu/Zn, Zn/Al ratios of catalyst A with those of catalyst AO, it is found that coke burning off caused an increase in Cu content and a decrease in Al content, and consequently caused increase in Cu/Al, Cu/Zn and Zn/Al ratios. Comparing Cu/Al, Cu/Zn, Zn/Al ratios of catalyst B with those of catalyst C, it is found that liquid phase heat treatment resulted in a decrease in Cu content and an increase in Zn content, and consequently resulted in decreased Cu/Al and Cu/Zn ratios and increased Zn/Al ratio. These results suggest that the preparation method and the heat treatment method of precursor influenced the contribution of components differently. TPR and acidity profiles H 2 -TPR profiles of the four catalysts are shown in Fig. 3. Catalyst A exhibited the minimum reduction peak and catalyst C possessed a slightly higher peak, in agreement with XRD results for the main presence of Cu 0 species in these catalysts. The low peak intensity was assigned to low content of Cu 2 O in the catalyst. Compared with A and C, catalysts B and AO exhibited much larger reduction peaks, attributed to the reduction of CuO, and the maximum peak for AO was related to the highest content of oxygen on the surface. Fig. 4 NH 3 -TPD-MS profiles of four Cu-Zn-Al catalysts. [1, A;2, AO; 3,B; 4, C]. Ammonia desorption profiles over the four catalysts measured by NH 3 -TPD-MS are shown in Fig. 4. The amount of ammonia uptake was calculated from the MS intensity of the fragment with m/e = 16. All samples showed comparable acid strength distribution, but the number of acidic sites was in the sequence of AO>C>A>>B, which suggests that the acidic sites are caused by the liquid phase heat treatment, probably due to the carbon injection into lattice of copper during the heat treatment. Catalyst AO possessed the greatest number of moderate and strong acid sites, which may be related to the oxidation process. The catalyst prepared by liquid phase method showed lower activity in fixed bed reactor than in slurry-bed reactor. The main reason for its lower CO conversion is that a carbon film covered the surface of the catalyst and thus resulted in mass transfer resistance to reactants. In the slurry reactor, the carbon film was in the paraffin form (paraffin was used as dispersant in a slurry reactor), and therefore, the exchange of paraffin film with syngas-containing paraffin dispersant eliminated mass transfer resistance to a large extent. Both processes of methanol synthesis and methanol dehydration were affected by the carbon film, and hence the CO conversion and selectivity of DME of catalyst A were very poor. After coke burning off, the activity of the catalyst was improved to a large extent, and CO conversion was even higher than that of the catalyst prepared by co-precipitation, which was due to the removal of the paraffin film.

6 1668 INDIAN J CHEM, SEC A, DECEMBER 2012 Conclusions The preparation and heat treatment methods of precursor affected the form of active metal present in catalyst and the catalyst surface state to a large extent, thereby influencing catalytic performance for CO hydrogenation. A suitable structure of catalysts may be obtained by correctly selecting preparation method and heat treatment method of precursor and operation conditions. Coke burning off was an effective way for use of the catalyst in a fixed bed reactor. Liquid phase technology can be applied for the preparation of catalysts to be used in different bed-type reactors. Supplementary Data Supplementary data associated with this article, i. e., Table S1, is available in the electronic form at Acknowledgement This work was supported by Major Project of Chinese National Programs for Fundamental Research and Development (973 Program, 2011CB211709), National Natural Science Foundation of China ( ), The Key Foundation of Science and Technology of Education Ministry of China (212021) and The Priority Development Field of Doctoral Foundation of Education Ministry of China ( ), PR China. References 1 Li J L, Zhang X G & Inui T, Appl Catal A, 147 (19) Guo X, Chen B, Bao G & Li L, Nat Gas Chem Ind, 28 (2003) 9. 3 Erena J, Garona R, Arandes J M, Aguayo A T & Bilbao J, Catal Today, (2005) Moradi G R, Nosrati S & Yaripor F, Catal Commun, 8 (2007) Prasad P S S, Bae J W, Kang S, Lee Y & Jun K, Fuel Process Technol, 89 (2008) Huang W, Gao Z, Hao L & Xie K, CN Patent No. ZL , 9 May, Liu L, Huang W, Gao Z & Yin L, Energy Sources A, 32 (2010) Fan J, Chen C, Zhao J, Huang W & Xie K, Fuel Process Technol, 91 (2010) Zuo Z, Sun L L, Huang W, Han P & Li Z, Appl Catal A, 375 (2010) Zuo Z, Huang W, Han P & Li Z, Appl Surf Sci, 256 (2010) Zhai X, Shamoto J, Xie H, Tan Y, Han Y & Tsubaki N, Chinese J Catal, 28 (2007) Yin Y, Xiao T, Su J, Wang, Hai T, Lu Y, Li J, Liu C, Zhu Q & Li S, Nat Gas Chem Ind, 25 (2000) Velu S, Suzuki K, Vijayaraj M, Barman S & Gopinath C S, Appl Catal B, 55 (2005) Velu S, Suzuki K & Gopinath C S. J Phys Chem B, 106 (2002)