- Email: [email protected]
Synthesis of methyl isobutyl ketone from acetone over metal-doped ion exchange resin catalyst
Applied Catalysis A: General 302 (2006) 140–148 www.elsevier.com/locate/apcata
Synthesis of methyl isobutyl ketone from acetone over metal-doped ion exchange resin catalyst Sandip Talwalkar, Sanjay Mahajani * Department of Chemical Engineering, Indian Institute of Technology Bombay, Powai, 400076 Mumbai, India Received 21 September 2005; received in revised form 29 December 2005; accepted 4 January 2006 Available online 14 February 2006
Abstract The kinetics of one-step synthesis of methyl isobutyl ketone from acetone was studied in the presence of the bifunctional commercial ion exchange resin, Amberlyst CH28 over a wide range of temperature, total pressure and catalyst loading in a batch reactor. An activity-based kinetic model is proposed to predict the observed results, with the non-idealities of the liquid phase being described using the UNIQUAC method. Formation of mesityl oxide was found to govern the overall rate of reaction. Low reaction rates were observed at higher conversion, possibly due to a pseudo-equilibrium caused by reversible deactivation of the catalyst as a result of formation of water in the reaction system. Simultaneous removal of water during the course of the reaction may result in an enhanced conversion. # 2006 Elsevier B.V. All rights reserved. Keywords: Methyl isobutyl ketone; Bifunctional catalyst; Hydrogenation; Ion exchange resin; Acetone; Mesityl oxide
1. Introduction Cation exchange resins are popular solid acid catalysts for liquid phase reactions conducted under relatively mild conditions. Some excellent reviews on catalysis by cation exchange resin have appeared in the past [1,2]. However, not much information is available on catalysis with a bifunctional ion exchange resin catalyst, as it is a relatively new field. Multiple reactions involving acid catalysts followed by hydrogenation or dehydrogenation and vice versa are commonly encountered in many industrial processes. If the conditions for both the reactions are overlapping, one can advantageously perform them in a single step with a bifunctional catalyst. Aldol condensation followed by hydrogenation of dehydrated aldol is an important class of these reactions [3–5] for which such a bifunctional catalyst can be a potential candidate. An industrially important reaction of aldol condensation of acetone, followed by hydrogenation to methyl isobutyl ketone, has been studied in the present work. Methyl isobutyl ketone is a widely used solvent in the pharmaceutical, coating and mining industries. It is also used in the manufacture of rubber antiozonants. Its synthesis from
* Corresponding author. Tel.: +91 22 25767246; fax: +91 22 25726895. E-mail address: [email protected] (S. Mahajani). 0926-860X/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2006.01.004
acetone in the presence of acidic and hydrogenation catalyst consists of three reaction steps. They are as follows: Reaction 1:
Reaction 2:
Reaction 3:
Several references in the form of communications, research articles and patents are available in the literature on the synthesis of methyl isobutyl ketone from acetone.
S. Talwalkar, S. Mahajani / Applied Catalysis A: General 302 (2006) 140–148
Nomenclature ai E8 k8 kwtr kact MCAT ni nAc,0 nAc,f r S T xi X
activity of component i activation energy for reaction (kJ/mol) rate constant for reaction (kmol/(s gcat)) adsorption coefficient of water adsorption coefficient of acetone mass of the catalyst (g) number of moles of component i initial number of moles of acetone number of moles of acetone at any time t rate of reaction (kmol/s) selectivity towards methyl isobutyl ketone Temperature (K) liquid mole fraction of component i conversion of acetone
Greek letters yi stoichiometric coefficient for component i F objective function for optimisation
However, the use of bifunctional catalyst for one-step synthesis is gaining importance and substantial work has been reported with different metals from Groups VIII and IB of the Periodic Table like Pd, Cu and Pt on acid supports. Table 1 gives a review on the previous studies on this system with bifunctional catalysts. It is clear from Table 1 that Pd is the most widely used metal for this system with almost all types of acid/base catalytic supports like alumiosilicates, zeolites, metal oxides and ion exchange resins. Different catalysts, like Pd-ZSM-5 [24], Pd–C–Nb2O5 [12,13] and Pd–CuO/MgO/SrO [8] in liquid phase and PdH-ZSM-5 [37], Cu–MgO [26] and Pd-CS-H-ZSM-5 [16] in gas phase, have been studied. They show conversions of acetone in the ranges 20–40% and 40–60% and selectivity towards methyl isobutyl ketone of 90% and 30–80%, respectively. It is reported that the catalyst used in the present work gives a conversion as high as 50% with 90% selectivity towards methyl isobutyl ketone in a liquid phase reaction [45]. The supports other than ion exchange resins need pre-treatment; indeed, the performance of these supports is sensitive to the method and conditions of the pre-treatment. Moreover, the reaction conditions are more severe than those with ion exchange resins without much improvement in conversions and selectivities. Ion exchange resins are more popular than traditional supports for low temperature operations due to the ease of separation, less pre-treatment and high acid strength. A commercial bifunctional ion exchange resin, Amberlyst CH28, has been designed for such performance and has been investigated in the present work. Ion exchange resin can be loaded with desired metal ions by contacting an aqueous solution of the metal ion with the hydrogen form of the cation exchange resin in a batch or continuous mode. Typically, the metal ion will be provided in
141
the form of a metal salt, such as chlorides, bromides, nitrates, sulphates and acetates. The detailed procedure for preparation for such catalysts can be found elsewhere [9]. In spite of a lot of literature on the present system, very little work is reported on commercial metal-ion exchange resins as a catalyst. Nicol and du Toit [45] have successfully performed reactions with the same catalysts in a laboratory-scale trickle bed reactor. A couple of patents from Catalytic Distillation Technologies [50–53] claim the use of catalytic distillation for enhanced conversion of acetone to methyl isobutyl ketone preferably catalysed by metal-ion exchange resin as a catalyst. Systematic kinetic studies and a reliable kinetic model are necessary to design an industrial reactor; the present work is undertaken to provide inputs in this regard. 2. Experimental 2.1. Materials Acetone (99.5%), methyl isobutyl ketone (99%) and methyl ethyl ketone (99.5%) were obtained from Merck Ltd., India. Rohm and Haas, France, supplied the catalyst Amberlyst CH28. The properties of catalyst are given in Table 2. Before its use, the catalyst was dried in an oven at 353 K for 3 h. 2.2. Apparatus and procedure A stainless steel autoclave from Parr Instrument Company, USA, with a capacity of 3 104 m3, equipped with an online temperature monitoring and control facility was used for conducting all the batch reactions. Before feeding to the reactor, each catalyst was washed with acetone in order to remove the moisture present, if any. The desired quantities of the catalyst and reactants (100 g of acetone in all runs) were charged to the reactor. The reactor and gas lines were flushed with hydrogen in order to remove the air present in the empty space in the reactor and lines. The reaction mixture was heated up to the desired temperature with slow stirring. As the reaction temperature was reached, the speed of agitation was increased up to the desired level and the corresponding time was regarded as the zero reaction time. The samples were withdrawn at different time intervals to study the kinetics of the reaction. In all the runs, the reaction volume was about 8 105 m3. 2.3. Analysis The reactants and products were analysed using a gas chromatograph (GC-MAK-911) equipped with a flame ionisation detector (FID). A 25 m long capillary column BP-1 (SGE, Australia) was used to separate the different components in the reaction mixture using methyl ethyl ketone as an external standard. The column temperature was maintained at 373 K isothermally. The various components in the reaction mixture and the separated products were characterized by authentic samples.
142
S. Talwalkar, S. Mahajani / Applied Catalysis A: General 302 (2006) 140–148
Table 1 Previous work for one-step synthesis of methyl isobutyl ketone from acetone Catalyst
Pressure (atm)
T (K)
Conversion (%)
Selectivity (%)
Reference
Pd-KUZ Pd–CuO/Al2O3/SiO2 Pd–CuO/MgO/SrO Pd–Nb2O5–Al2O3 Pd-oxides of Ti, Zr, Cr Pd-KUZFPP Pd–C–Nb2O5 Pd–C–Nb2O5 Pd-oxides of Ce, Hf, Ta Pd-IER Pd-CS-H-ZSM-5 Pd–Nb2O5 Pd-oxides of Zr Pd-KS-IER Pd-ZSM-5 Pd-IER Ni, Cu, Co-g-a-Al2O3 Ni–Al2O3 Pd-ZSM-5 SiO2/Al2O3 = 30 Ni–CaO Cu–MgO Pt-HMF Ni–CaO Pd–AIPO4-II, SAPO4 Ni–Al phosphate Pd-(Zn)-H-ZSM-5 Pd–Nb2O5–SiO2 Ni–CaO Pd–Al2O3 Pt-H-ZSM-5 Pt-NaX Pd-H-ZSM-5 Pd-IER Cu–MgO–Al2O3 Pd-zeolite Pd-IER Pt-Sn-H-ZSM-5 Pd–C Pd–Ca–Al2O3 Amberlyst CH28 Pd-IER Pd-MCM-56 Pd-MCM-49 Cu–MgO–Al2O3
20 1 20 – 10 30 20 10 10 40 1 – – 50 50 – – 1 6–60 – 1 1 – – – 5 – – 40–90 1 – 1 40–70 – – – 1 1-20 30 5–15 – – 1
393 423–503 433 – 413 353–388 413 413 413 363–393 523 – – 373–403 443 – 453 373 433 473 653 433 – – – 408–483 – – 413–473 433 713 473 353–373 – – 403 433 333 403–423 373–453 – – 513
50 30 38.5 28 33.9 – 39.5 30 33 41.9 41.8 27 40.25 – 53.9 – 41.24 70–80 60–80 – 60–80 – – 55 30–35 60–70 – – – 47.3 – 65.26 70 43.2 – – – 25–50 – 33.5 35.6 71.7
94.5 60 93.6 90 92.3 – 92.5 91.7 90.2 – 82.4 93.5 94.9 90 95.36 – 37.1 95 90.98 60 60–75 – 50–60 – – 83–94 88–92 70 – – 70 30.7 – 57.49 87.5 98.2 – – – 70–90 – 81.2 85 50.1
[6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49]
Table 2 Physical properties of Amberlyst CH28 [59]
2.4. Calculations for conversion and selectivity
Property
Value
Matrix Physical form Total capacity (mequiv./g dry basis) Palladium loading (% dry basis) Specific surface area (m2/gm) ˚) Pore diameter (A Temperature stability (K) Harmonic mean size (mm) Moisture content in wet form (%)
Styrene DVB polymer Spherical beads 4.8 0.7 36 260 403 0.85–1.05 52–58
The conversion of acetone was calculated as X¼
nAc;0 nAc;f nAc;0
(1)
while the selectivity towards methyl isobutyl ketone was calculated as S¼
2nMIBK nAc;0 nAc;f
(2)
S. Talwalkar, S. Mahajani / Applied Catalysis A: General 302 (2006) 140–148
143
3. Results and discussion 3.1. General course of the reaction As mentioned earlier, we define the zero reaction time as the time at which the desired temperature is attained. Hence, in all the kinetic runs, we observe a small extent of reaction that occurred before the desired temperature was reached. The extent of reaction during this heat-up period was found to be relatively higher at higher catalyst loadings and temperature. Fig. 1 shows typical profiles of mole fractions of acetone and methyl isobutyl ketone with respect to time. The concentration of water is the same as the concentration of methyl isobutyl ketone and is not shown in Fig. 1. It should be noted that the concentrations of intermediates, diacetone alcohol and mesityl oxide, were below the detectable limits. 3.2. Mass transfer effects In order to ensure that there is no significant external mass transfer resistance across the solid–liquid and gas–liquid interfaces, the reactions were performed at different stirrer speeds over a range 600–1700 rpm. It was observed that the reaction kinetics is insensitive to agitation above 1000 rpm, as shown in Fig. 2, and hence all the reactions were performed above 1000 rpm.
Fig. 2. Effect of speed of agitation on reaction kinetics. Temperature = 413 K; catalyst loading: 5% w/w of acetone; pressure = 30 bar; speed in rpm.
In all the experiments, we have observed very little or no intermediate products, i.e., mesityl oxide and diacetone alcohol. Therefore, it was assumed that rapid dehydration of diacetone alcohol leads to formation of mesityl oxide, which in turn instantaneously gets converted to methyl isobutyl ketone in the presence of hydrogen. To compare the rates of mesityl oxide
and methyl isobutyl ketone formation, we performed a run in the absence of hydrogen for 50 min. The samples were collected up to 50 min and hydrogen was introduced to the reactor at this time at 30 bar. Fig. 3 shows that, as soon as hydrogen was fed to the reactor, the concentrations of mesityl oxide and diacetone alcohol dropped instantaneously to almost zero and the methyl isobutyl ketone concentration builds up sharply. This suggests that the hydrogenation is much faster than both the intermediate reactions and practically no or very little intermediate components were observed in the range of parameters for which the kinetics of methyl isobutyl ketone formation from acetone was generated on the bifunctional Amberlyst CH28.
Fig. 1. General course of the reaction. Temperature = 393 K; catalyst loading: 5% w/w of acetone; pressure = 30 bar.
Fig. 3. Speed of hydrogenation reaction. Temperature = 393 K; catalyst loading: 5% w/w of acetone; hydrogen injection at 3000 s at 30 bar.
3.3. Comparisons of the rates of individual steps
144
S. Talwalkar, S. Mahajani / Applied Catalysis A: General 302 (2006) 140–148
Therefore, the overall reaction with the present catalyst is given by
at higher conversion levels of acetone. To find out the cause of catalyst deactivation, the catalyst was separated from the
Fig. 3 also indicates that the rate of mesityl oxide formation is very low even at low conversion levels, indicating that the reaction rates are highly affected by equilibrium. Thus, the simultaneous hydrogenation of mesityl oxide may help in shifting the reaction in the forward direction by keeping the mesityl oxide concentration at low level at any given time in the reactor. Thus, complete conversion of acetone was expected in the present work because of the removal of mesityl oxide in hydrogenation reaction to of methyl isobutyl ketone. However, low reaction rates were observed at higher conversion levels. This may imply that the hydrogenation reaction is equilibriumlimited and only equilibrium conversions can be obtained. However, Winter et al. [43] showed that 100% conversion mesityl oxide in liquid phase could be possible on Pd/CNF catalyst. This confirms that the observed low reaction rates on Amberlyst CH28 at higher conversion level are due to catalyst deactivation and not because of reaction equilibrium. The overall reaction is thus a pseudo equilibrium type as a result of deactivation of catalyst by the moisture formed. More evidence of the same can be found in the literature [54]. We confirmed this through the catalyst deactivation experiments described in the next section.
reaction mixture and dried in an oven at 353 K for 3 h and the same catalyst was reused in the next run. This procedure was repeated for next few runs. As shown in Fig. 4, the performance of the used catalyst was almost identical to that of the fresh catalyst. This indicates that the deactivation of catalyst is reversible and that is caused by the presence of polar components like water [1,2,55], which is produced in the dehydration of aldol. Polar compounds like water have strong affinity towards catalytic sites. In a relatively non-polar environment, water molecules are known to adsorb on the resin surface and cover the catalytic sites, thereby causing an ‘‘inhibition effect’’. As mentioned before, it should be noted that the reaction becomes very slow in this region and the kinetics, as shown in Fig. 1, looks similar to that of a reversible reaction approaching equilibrium. Since in this case, true thermodynamic reaction equilibrium is not attained, we call this state a ‘‘pseudo-equilibrium’’. A detailed investigation of the same has been reported in the literature for the reaction of acetone to mesityl oxide on a mono-functional ion exchange resin [54]. 3.5. Effect of moisture concentration
3.4. Cause of catalyst deactivation As discussed above, the catalyst gets deactivated during the course of reaction, as a result of which low rates were observed
In order to ascertain the exact role of water in the kinetics, we performed experiments with and without water. As shown in Fig. 5, we found that the presence of water in the reaction
Fig. 4. Cause of catalyst deactivation. Temperature = 393 K; catalyst loading: 5% w/w of acetone; pressure = 30 bar.
Fig. 5. Effect of water addition. Temperature = 413 K; catalyst loading: 5% w/ w of acetone; pressure = 30 bar water concentration in wt.%.
S. Talwalkar, S. Mahajani / Applied Catalysis A: General 302 (2006) 140–148
Fig. 6. Effect of temperature on reaction kinetics. Temperature in K; catalyst loading: 5% w/w of acetone; pressure = 30 bar.
system is the true cause of the observed reversible catalyst deactivation; hence, low rates of reaction are observed for the reaction with initial moisture. This implies that the rate of formation of mesityl oxide is affected due to inhibition of sulfonic acid groups by polar components like water. This phenomenon is well known for many reactions on ion exchange resins. However, mesityl oxide formation is still followed by its instantaneous hydrogenation to methyl isobutyl ketone. In short, the rate of formation of methyl isobutyl ketone is equal to the rate of formation of mesityl oxide as in the case without moisture. This suggests that simultaneous removal of mesityl oxide along with the removal of water by some means is necessary to enhance the reactor performance further. Therefore, one can regard this system to be a good candidate for multifunctional reactor system like reactive distillation or reactive adsorption. As mentioned before, except for few
145
Fig. 7. Effect of temperature on selectivity towards methyl isobutyl ketone. Temperature in K; catalyst loading: 5% w/w of acetone; pressure = 30 bar.
references [50–53] not much work has been reported in the literature on this aspect. 3.6. Effect of temperature The effect of temperature was studied over a range of 373–413 K. Fig. 6 shows the conversion of acetone at different temperatures. As expected, the conversion of acetone increases with an increase in temperature at the cost of reduced selectivity towards methyl isobutyl ketone, as shown in Fig. 7. At relatively high temperatures, mesityl oxide is likely to react with acetone to form heavy components like isophorone and other condensation products. Nicol and du Toit [45] evaluated the thermal stability of the present catalyst for the same reaction, and reported that the catalyst is stable up to 423 K.
Fig. 8. Effect of catalyst loading on (a) reaction kinetics (b) initial rate. Temperature = 393 K; catalyst loading: in w/w of acetone; pressure = 30 bar.
146
S. Talwalkar, S. Mahajani / Applied Catalysis A: General 302 (2006) 140–148
3.7. Effect of catalyst loading
4. Kinetic modeling
The reactions were performed at different catalyst loadings (2–12% w/w of acetone) and it was observed that the rate of reaction increases when the catalyst loading was increased from 2% to 5% w/w of acetone. However, the rate of reaction reaches saturation above the catalyst loading of 5% w/w of acetone, as shown in Fig. 8. Moreau et al. [56] have reported such non-liner influence of catalyst loading on the rate of hydrolysis of acetals over H-montmorillonite and strong cation exchange resin. It is expected that the number of catalyst sites above a particular catalyst loading becomes excessive and hence the rate becomes insensitive to the further increase in the catalyst loading. We believe this effect is not because of the diffusional limitations as variables such as stirrer speed etc. do not influence the overall rate of reaction under these conditions. In our opinion, such an effect can be explained only through the rigorous thermodynamic models of adsorption. A theoretical explanation for the same is lacking in the literature though some evidences have been reported [56].
O’Keefe et al. [57] proposed a mechanism for hydrogenation of mesityl oxide to methyl isobutyl ketone. The mechanism consists of five steps. In the first step, mesityl oxide adsorbs on two adjacent sites to form a di-adsorbed species. In the second step, hydrogen adsorbs dissociatively on two adjacent sites. The adsorbed hydrogen atoms are then added stepwise to the diadsorbed mesityl oxide to give methyl isobutyl ketone, which then desorbs from the catalyst surface to the bulk. The authors have confirmed this mechanism by pulse adsorption of mesityl oxide, which was monitored by diffuse reflectance infrared spectroscopy and they proposed a model based on the Langmuir–Hinshelwood approach. We assume that a similar mechanism holds good for the present system. Acetone gets adsorbed on two adjacent active acid sites and condenses to give diadsorbed dimer, which then dehydrates to form diadsorbed mesityl oxide to give methyl isobutyl ketone in an adsorbed state in a similar way to that discussed above. As discussed before, we observed that the reaction is pseudo-zero order with respect to the hydrogen concentration over the range of pressure studied. The rate equation may be given by
3.8. Effect of hydrogen pressure The solubility of hydrogen varies with pressure and hence it may influence the reaction kinetics. In order to study the effect of hydrogen pressure, we performed reactions over a range of 15–45 bar. Interestingly, it was observed that there is no effect of hydrogen pressure on the reaction kinetics, as shown in Fig. 9. This finding again confirms that the catalyst activity for hydrogenation is sufficient enough and that it is the formation of mesityl oxide that governs the overall rate of the reaction.
r¼
ka2act ð1 þ kwtr awtr þ kact aact Þ4
(3)
The adsorption term for the methyl isobutyl ketone was omitted from the model, as O’Keefe et al. [57] found that there is no effect on reaction kinetics when methyl isobutyl ketone was added to the reaction mixture at time equal to zero. Also, the parameter fitting with inclusion of this term gave us very low values for the adsorption coefficient of methyl isobutyl ketone. 5. Parameter estimation The mole balance for the batch reactor for all the components can be written as dni ¼ MCAT yi r dt
(4)
The temperature dependency of the rate constant can be expressed by the Arrhenius equation E k ¼ k0 exp (5) RT The objective function to be minimized is given as X min F ¼ ðxi;calculated xi;measured Þ2
(6)
All samples
Fig. 9. Effect of hydrogen pressure on the reaction kinetics. Temperature = 393 K; catalyst loading 5% w/w of acetone; pressure in bar.
For optimisation, a SQP (Sequential Quadratic Programming) approach from NAG library was used in the DIVA [58] simulation environment. The non-idealities of liquid phase were described using the UNIQUAC method and the thermodynamic parameters were obtained from the data bank of the commercial package ASPEN. Hydrogen solubility was predicted by using Henry’s law and interaction parameters to calculate Henry’s constant were also taken from the commercial package ASPEN. The estimated parameters for the kinetic
S. Talwalkar, S. Mahajani / Applied Catalysis A: General 302 (2006) 140–148
147
Table 3 Kinetic parameters for proposed model k8 (kmol/(g s)) E8 (kJ/mol) Kwtr Kact f
5.8714 107 97.60 4.22 2.58 2.3 103
Fig. 12. Model predictions for effect of moisture conversion of acetone. Temperature = 413 K; catalyst loading: 5% w/w of acetone; pressure = 30 bar, moisture concentration in wt.%.
Fig. 10. Comparison between modelled and measured conversion of acetone at different temperatures. Temperature in K; catalyst loading: 5% w/w of acetone; pressure = 30 bar.
model are given in Table 3. As shown in Figs. 10 and 11, the model captures the effect of parameters like temperature and catalyst loading. It should be noted that the model is applicable for catalyst loading up to 5% w/w of acetone, beyond which the rate change is non-liner with catalyst loading (see Fig. 8). The model was thus used to independently predict the performance of a batch reactor with initial moisture present. The predicted results are compared with the experimental ones in Fig. 12. The model slightly over-predicts the rate in the presence of initial moisture content. Another approach to capture the effect of moisture by considering Freundlich adsorption of water [55] was also applied. However, the fitting was worse by the latter approach. 6. Conclusions
Fig. 11. Comparison between modelled and measured conversion of acetone at different values of catalyst loadings. Temperature = 393 K; catalyst loading: in % w/w of acetone; pressure = 30 bar.
Kinetics of the one-pot synthesis of methyl isobutyl ketone from acetone was studied. The effects of various parameters like temperature and catalyst loading were investigated in detail. It was observed that catalyst undergoes reversible deactivation due to the formation of water in the condensation reaction. As a result, lower rates are observed at higher conversion levels of acetone, showing pseudo-equilibrium, which was confirmed by independent reactions. About 45% conversion of acetone with more than 95% selectivity towards methyl isobutyl ketone was realised in about 3.5 h at 393 K. This is promising, considering the performance of different reported catalysts, like Pd-ZSM-5 [24], Pd–C–Nb2O5 [12,13], Pd–CuO/MgO/SrO [8], Pd-H-ZSM-5 [37], Cu–MgO [26] and Pd-CS-H-ZSM-5 [16] with 20–60% conversion of acetone and 30–90% selectivity towards methyl isobutyl ketone under similar or even more harsh conditions. The pseudo-equilibrium due to adsorption of water may be considered as a possible
148
S. Talwalkar, S. Mahajani / Applied Catalysis A: General 302 (2006) 140–148
limitation of ion exchange resin catalyst. However, the simultaneous removal of water during the course of reaction will help to overcome this drawback. A kinetic model that considers the adsorption of water on the catalyst is proposed which satisfactorily explains the experimental observations. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29]
A. Chakrabarti, M.M. Sharma, React. Polym. 20 (1993) 1–45. M.M. Sharma, React. Funct. Polym. 26 (1995) 3–23. S. Veibel, J.I. Nielsen, Tetrahedron 23 (1967) 1723–1733. L. Ba¨rfacker, D. Tom, P. Eilbracht, Tetrahedron Lett. 40 (1999) 4031– 4034. F. King, G.J. Kelly, Catal. Today 73 (2002) 75–81. S. Guseva, A. Grigor’eva, T. Prozorovskaya, Neftepererabotka i Neftekhimiya 6 (1978) 30–32. S. Wen, Z. Li, Z. Qian, X. Mao, Jilin Daxue Ziran Kexue Xuebao 1 (1981) 99–108. Y. Too, T. Nakayama, JP62258335 (1987). Y. Higashio, K. Hamada, EP87-308156 (1987). T. Maki, T. Yokoyama, Y. Sumino, JP63096147 (1988). A.A. Grigor’ev, S.I. Guseva, E.A. Katsman, Khimicheskaya Promyshlennost 7 (1988) 392–394. T. Maki, T. Yokoyama, Y. Sumino, JP63068538 (1988). T. Maki, T. Yokoyama, Y. Sumino, JP63068539 (1988). T. Maki, T. Yokoyama, Y. Sumino, JP63096146 (1988). Y. Qian, G. Ding, H. Wu, Y. Zheng, J. Jiang, Huaxue Fanying Gongcheng Yu Gongyi 4 (1988) 89–93. H. Chin, K. Shu, Y. Chin, N. Cho, B. Hayashi, S. So, JP02069430 (1990). Y. Too, T. Nakayama, K. Shioda, JP03093743 (1991). T. Ushikubo, Y. Kakuno, H. Tsujiura, EP458216 (1991). J. Pasek, J. Hajek, J. Koubek, S. Zicha, P. Pavlas, P. Kralik, A. Kospach, A. CS273524 (1992). L. Chu, Z. Yao, CN1069674 (1993). K. Bian, L. Ding, J. Gao, H. Gu, Ziran Kexue 31 (1995) 63–67. K. Tao, X. Qiu, Y. Wang, CN1108151 (1995). S. Narayanan, R. Unnikrishnan, Appl. Catal. A: Gen. 145 (1996) 231–236. P. Chen, S. Chu, K. Wu, W. Lin, US Patent 5,684,207 (1997). B. Coh, J. Hur, H. Lee, Chem. Lett. 7 (1998) 583–584. V. Chikan, M. Arpad, B. Katalin, J. Catal. 184 (1999) 134–143. L. Melo, G. Giannetto, L. Cardozo, A. Llanos, L. Garcia, P. Magnoux, M. Guisnet, F. Alvarez, Catal. Lett. 60 (1999) 217–222. B. Coh, J. Hur, H. Lee, Stud. Surf. Sci. Catal. 121 (1999) 453–456. S. Yang, Y. Wu, Appl. Catal. A: Gen. 192 (2000) 211–220.
[30] M. Montes, M. Martin, A. Diaz, J. Benitez, J. Odriozola, Mater. Sci. Forum (2000) 325–326. [31] D. Bombos, G. Bozga, M. Bombos, A. Stefan, I. Stanciu, Chem. Papers 54 (2000) 171–176. [32] Y. Chen, B. Liaw, H. Tan, K. Shen, Appl. Catal. A: Gen. 205 (2001) 61–69. [33] H. Lee, B. Coh, Proceedings of the 221st ACS National Meeting, 2001. [34] Z. Lu, W. Ai, L. Song, L. Xu, H. Zhu, Faming Zhuanli Shenqing Gongkai Shuomingshu 9 (2001). [35] L. Melo, A. Llanos, M. Mediavilla, D. Moronta, J. Mol. Catal. A: Chem. 177 (2002) 281–287. [36] L. Mattos, F. Noronha, J. Monteiro, J. Catal. 209 (2002) 166–176. [37] K. Lin, A. Ko, J. Chin. Chem. Soc. 49 (2002) 935–942. [38] Z. Lu, F. Ai, L. Song, W. Liang, CN1331071 (2002). [39] Y. Shen, J. Wu, Z. Zhang, Shenyang Huagong Xueyuan Xuebao 16 (2002) 261–265. [40] D. Bombos, T. Juganaru, A. Stefan, M. Bombos, E. Juganaru, RO117154 (2001). [41] M. Vandersall, R. Weinand, EP321450 (2003). [42] R. Morales, L. Melo, J. Brito, A. Llanos, D. Moronta, L. Albornoz, E. Rodriguez, J. Mol. Catal. A: Chem. 203 (2003) 277–286. [43] F. Winter, A. van Dillen, K. de Jong, J. Mol. Catal. A: Chem. 219 (2004) 273–281. [44] X. Sun, D. Lu, X. Bu, Z. Wang, H. Liu, D. Zhou, Jilin Huagong Xueyuan Xuebao 21 (2004) 22–23. [45] W. Nicol, E. du Toit, Chem. Eng. Process. 43 (2004) 1539–1545. [46] F. Ai, Z. Lu, L. Song, H. Wang, S. Wang, Shunnong, CN1488617 (2004). [47] P. Yang, J. Yu, Z. Wang, M. Xu, Q. Liu, X. Yang, T. Wu, Catal. Commun. 6 (2005) 107–111. [48] P. Yang, Z. Wang, J. Yu, M. Xu, Q. Liu, S. Jing, T. Wu, Polish J. Chem. 79 (2005) 121–127. [49] Y. Shen, H. Zhang, J. Wu, Z. Zhang, Huagong Shengchan Yu Jishu 12 (2005) 9–11. [50] K. Lawson, B. Nkosi, WO9965851 (1999). [51] N. Saayman, G. Lund, WO0222542 (2002). [52] N. Saayman, G. Lund, US2003065227 (2003). [53] K. Lawson, MXPA00011793 (2002). [54] E. du Toit, R. Schwarzer, W. Nicol, Chem. Eng. Sci. 59 (2004) 5545–5550. [55] E. du Toit, W. Nicol, Appl. Catal. A: Gen. 277 (2004) 219–225. [56] C. Moreau, R. Durand, P. Geneste, S. Mseddi, J. Mol. Catal. 112 (1996) 133–141. [57] W.K. O’Keefe, M. Jiang, F.T.T. Ng, G.L. Rempel, Chem. Eng. Sci. 60 (2005) 4131–4140. [58] M. Mangold, A. Kienle, E. Gilles, K. Chem. Eng. Sci. 55 (2000) 441. [59] http://www.rohmhaas.com/ionexchange/IP/ch28_typical.htm, seen on 20 August 2005.
Synthesis of methyl isobutyl ketone from acetone over metal-doped ion exchange resin catalyst Sandip Talwalkar, Sanjay Mahajani * Department of Chemical Engineering, Indian Institute of Technology Bombay, Powai, 400076 Mumbai, India Received 21 September 2005; received in revised form 29 December 2005; accepted 4 January 2006 Available online 14 February 2006
Abstract The kinetics of one-step synthesis of methyl isobutyl ketone from acetone was studied in the presence of the bifunctional commercial ion exchange resin, Amberlyst CH28 over a wide range of temperature, total pressure and catalyst loading in a batch reactor. An activity-based kinetic model is proposed to predict the observed results, with the non-idealities of the liquid phase being described using the UNIQUAC method. Formation of mesityl oxide was found to govern the overall rate of reaction. Low reaction rates were observed at higher conversion, possibly due to a pseudo-equilibrium caused by reversible deactivation of the catalyst as a result of formation of water in the reaction system. Simultaneous removal of water during the course of the reaction may result in an enhanced conversion. # 2006 Elsevier B.V. All rights reserved. Keywords: Methyl isobutyl ketone; Bifunctional catalyst; Hydrogenation; Ion exchange resin; Acetone; Mesityl oxide
1. Introduction Cation exchange resins are popular solid acid catalysts for liquid phase reactions conducted under relatively mild conditions. Some excellent reviews on catalysis by cation exchange resin have appeared in the past [1,2]. However, not much information is available on catalysis with a bifunctional ion exchange resin catalyst, as it is a relatively new field. Multiple reactions involving acid catalysts followed by hydrogenation or dehydrogenation and vice versa are commonly encountered in many industrial processes. If the conditions for both the reactions are overlapping, one can advantageously perform them in a single step with a bifunctional catalyst. Aldol condensation followed by hydrogenation of dehydrated aldol is an important class of these reactions [3–5] for which such a bifunctional catalyst can be a potential candidate. An industrially important reaction of aldol condensation of acetone, followed by hydrogenation to methyl isobutyl ketone, has been studied in the present work. Methyl isobutyl ketone is a widely used solvent in the pharmaceutical, coating and mining industries. It is also used in the manufacture of rubber antiozonants. Its synthesis from
* Corresponding author. Tel.: +91 22 25767246; fax: +91 22 25726895. E-mail address: [email protected] (S. Mahajani). 0926-860X/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2006.01.004
acetone in the presence of acidic and hydrogenation catalyst consists of three reaction steps. They are as follows: Reaction 1:
Reaction 2:
Reaction 3:
Several references in the form of communications, research articles and patents are available in the literature on the synthesis of methyl isobutyl ketone from acetone.
S. Talwalkar, S. Mahajani / Applied Catalysis A: General 302 (2006) 140–148
Nomenclature ai E8 k8 kwtr kact MCAT ni nAc,0 nAc,f r S T xi X
activity of component i activation energy for reaction (kJ/mol) rate constant for reaction (kmol/(s gcat)) adsorption coefficient of water adsorption coefficient of acetone mass of the catalyst (g) number of moles of component i initial number of moles of acetone number of moles of acetone at any time t rate of reaction (kmol/s) selectivity towards methyl isobutyl ketone Temperature (K) liquid mole fraction of component i conversion of acetone
Greek letters yi stoichiometric coefficient for component i F objective function for optimisation
However, the use of bifunctional catalyst for one-step synthesis is gaining importance and substantial work has been reported with different metals from Groups VIII and IB of the Periodic Table like Pd, Cu and Pt on acid supports. Table 1 gives a review on the previous studies on this system with bifunctional catalysts. It is clear from Table 1 that Pd is the most widely used metal for this system with almost all types of acid/base catalytic supports like alumiosilicates, zeolites, metal oxides and ion exchange resins. Different catalysts, like Pd-ZSM-5 [24], Pd–C–Nb2O5 [12,13] and Pd–CuO/MgO/SrO [8] in liquid phase and PdH-ZSM-5 [37], Cu–MgO [26] and Pd-CS-H-ZSM-5 [16] in gas phase, have been studied. They show conversions of acetone in the ranges 20–40% and 40–60% and selectivity towards methyl isobutyl ketone of 90% and 30–80%, respectively. It is reported that the catalyst used in the present work gives a conversion as high as 50% with 90% selectivity towards methyl isobutyl ketone in a liquid phase reaction [45]. The supports other than ion exchange resins need pre-treatment; indeed, the performance of these supports is sensitive to the method and conditions of the pre-treatment. Moreover, the reaction conditions are more severe than those with ion exchange resins without much improvement in conversions and selectivities. Ion exchange resins are more popular than traditional supports for low temperature operations due to the ease of separation, less pre-treatment and high acid strength. A commercial bifunctional ion exchange resin, Amberlyst CH28, has been designed for such performance and has been investigated in the present work. Ion exchange resin can be loaded with desired metal ions by contacting an aqueous solution of the metal ion with the hydrogen form of the cation exchange resin in a batch or continuous mode. Typically, the metal ion will be provided in
141
the form of a metal salt, such as chlorides, bromides, nitrates, sulphates and acetates. The detailed procedure for preparation for such catalysts can be found elsewhere [9]. In spite of a lot of literature on the present system, very little work is reported on commercial metal-ion exchange resins as a catalyst. Nicol and du Toit [45] have successfully performed reactions with the same catalysts in a laboratory-scale trickle bed reactor. A couple of patents from Catalytic Distillation Technologies [50–53] claim the use of catalytic distillation for enhanced conversion of acetone to methyl isobutyl ketone preferably catalysed by metal-ion exchange resin as a catalyst. Systematic kinetic studies and a reliable kinetic model are necessary to design an industrial reactor; the present work is undertaken to provide inputs in this regard. 2. Experimental 2.1. Materials Acetone (99.5%), methyl isobutyl ketone (99%) and methyl ethyl ketone (99.5%) were obtained from Merck Ltd., India. Rohm and Haas, France, supplied the catalyst Amberlyst CH28. The properties of catalyst are given in Table 2. Before its use, the catalyst was dried in an oven at 353 K for 3 h. 2.2. Apparatus and procedure A stainless steel autoclave from Parr Instrument Company, USA, with a capacity of 3 104 m3, equipped with an online temperature monitoring and control facility was used for conducting all the batch reactions. Before feeding to the reactor, each catalyst was washed with acetone in order to remove the moisture present, if any. The desired quantities of the catalyst and reactants (100 g of acetone in all runs) were charged to the reactor. The reactor and gas lines were flushed with hydrogen in order to remove the air present in the empty space in the reactor and lines. The reaction mixture was heated up to the desired temperature with slow stirring. As the reaction temperature was reached, the speed of agitation was increased up to the desired level and the corresponding time was regarded as the zero reaction time. The samples were withdrawn at different time intervals to study the kinetics of the reaction. In all the runs, the reaction volume was about 8 105 m3. 2.3. Analysis The reactants and products were analysed using a gas chromatograph (GC-MAK-911) equipped with a flame ionisation detector (FID). A 25 m long capillary column BP-1 (SGE, Australia) was used to separate the different components in the reaction mixture using methyl ethyl ketone as an external standard. The column temperature was maintained at 373 K isothermally. The various components in the reaction mixture and the separated products were characterized by authentic samples.
142
S. Talwalkar, S. Mahajani / Applied Catalysis A: General 302 (2006) 140–148
Table 1 Previous work for one-step synthesis of methyl isobutyl ketone from acetone Catalyst
Pressure (atm)
T (K)
Conversion (%)
Selectivity (%)
Reference
Pd-KUZ Pd–CuO/Al2O3/SiO2 Pd–CuO/MgO/SrO Pd–Nb2O5–Al2O3 Pd-oxides of Ti, Zr, Cr Pd-KUZFPP Pd–C–Nb2O5 Pd–C–Nb2O5 Pd-oxides of Ce, Hf, Ta Pd-IER Pd-CS-H-ZSM-5 Pd–Nb2O5 Pd-oxides of Zr Pd-KS-IER Pd-ZSM-5 Pd-IER Ni, Cu, Co-g-a-Al2O3 Ni–Al2O3 Pd-ZSM-5 SiO2/Al2O3 = 30 Ni–CaO Cu–MgO Pt-HMF Ni–CaO Pd–AIPO4-II, SAPO4 Ni–Al phosphate Pd-(Zn)-H-ZSM-5 Pd–Nb2O5–SiO2 Ni–CaO Pd–Al2O3 Pt-H-ZSM-5 Pt-NaX Pd-H-ZSM-5 Pd-IER Cu–MgO–Al2O3 Pd-zeolite Pd-IER Pt-Sn-H-ZSM-5 Pd–C Pd–Ca–Al2O3 Amberlyst CH28 Pd-IER Pd-MCM-56 Pd-MCM-49 Cu–MgO–Al2O3
20 1 20 – 10 30 20 10 10 40 1 – – 50 50 – – 1 6–60 – 1 1 – – – 5 – – 40–90 1 – 1 40–70 – – – 1 1-20 30 5–15 – – 1
393 423–503 433 – 413 353–388 413 413 413 363–393 523 – – 373–403 443 – 453 373 433 473 653 433 – – – 408–483 – – 413–473 433 713 473 353–373 – – 403 433 333 403–423 373–453 – – 513
50 30 38.5 28 33.9 – 39.5 30 33 41.9 41.8 27 40.25 – 53.9 – 41.24 70–80 60–80 – 60–80 – – 55 30–35 60–70 – – – 47.3 – 65.26 70 43.2 – – – 25–50 – 33.5 35.6 71.7
94.5 60 93.6 90 92.3 – 92.5 91.7 90.2 – 82.4 93.5 94.9 90 95.36 – 37.1 95 90.98 60 60–75 – 50–60 – – 83–94 88–92 70 – – 70 30.7 – 57.49 87.5 98.2 – – – 70–90 – 81.2 85 50.1
[6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49]
Table 2 Physical properties of Amberlyst CH28 [59]
2.4. Calculations for conversion and selectivity
Property
Value
Matrix Physical form Total capacity (mequiv./g dry basis) Palladium loading (% dry basis) Specific surface area (m2/gm) ˚) Pore diameter (A Temperature stability (K) Harmonic mean size (mm) Moisture content in wet form (%)
Styrene DVB polymer Spherical beads 4.8 0.7 36 260 403 0.85–1.05 52–58
The conversion of acetone was calculated as X¼
nAc;0 nAc;f nAc;0
(1)
while the selectivity towards methyl isobutyl ketone was calculated as S¼
2nMIBK nAc;0 nAc;f
(2)
S. Talwalkar, S. Mahajani / Applied Catalysis A: General 302 (2006) 140–148
143
3. Results and discussion 3.1. General course of the reaction As mentioned earlier, we define the zero reaction time as the time at which the desired temperature is attained. Hence, in all the kinetic runs, we observe a small extent of reaction that occurred before the desired temperature was reached. The extent of reaction during this heat-up period was found to be relatively higher at higher catalyst loadings and temperature. Fig. 1 shows typical profiles of mole fractions of acetone and methyl isobutyl ketone with respect to time. The concentration of water is the same as the concentration of methyl isobutyl ketone and is not shown in Fig. 1. It should be noted that the concentrations of intermediates, diacetone alcohol and mesityl oxide, were below the detectable limits. 3.2. Mass transfer effects In order to ensure that there is no significant external mass transfer resistance across the solid–liquid and gas–liquid interfaces, the reactions were performed at different stirrer speeds over a range 600–1700 rpm. It was observed that the reaction kinetics is insensitive to agitation above 1000 rpm, as shown in Fig. 2, and hence all the reactions were performed above 1000 rpm.
Fig. 2. Effect of speed of agitation on reaction kinetics. Temperature = 413 K; catalyst loading: 5% w/w of acetone; pressure = 30 bar; speed in rpm.
In all the experiments, we have observed very little or no intermediate products, i.e., mesityl oxide and diacetone alcohol. Therefore, it was assumed that rapid dehydration of diacetone alcohol leads to formation of mesityl oxide, which in turn instantaneously gets converted to methyl isobutyl ketone in the presence of hydrogen. To compare the rates of mesityl oxide
and methyl isobutyl ketone formation, we performed a run in the absence of hydrogen for 50 min. The samples were collected up to 50 min and hydrogen was introduced to the reactor at this time at 30 bar. Fig. 3 shows that, as soon as hydrogen was fed to the reactor, the concentrations of mesityl oxide and diacetone alcohol dropped instantaneously to almost zero and the methyl isobutyl ketone concentration builds up sharply. This suggests that the hydrogenation is much faster than both the intermediate reactions and practically no or very little intermediate components were observed in the range of parameters for which the kinetics of methyl isobutyl ketone formation from acetone was generated on the bifunctional Amberlyst CH28.
Fig. 1. General course of the reaction. Temperature = 393 K; catalyst loading: 5% w/w of acetone; pressure = 30 bar.
Fig. 3. Speed of hydrogenation reaction. Temperature = 393 K; catalyst loading: 5% w/w of acetone; hydrogen injection at 3000 s at 30 bar.
3.3. Comparisons of the rates of individual steps
144
S. Talwalkar, S. Mahajani / Applied Catalysis A: General 302 (2006) 140–148
Therefore, the overall reaction with the present catalyst is given by
at higher conversion levels of acetone. To find out the cause of catalyst deactivation, the catalyst was separated from the
Fig. 3 also indicates that the rate of mesityl oxide formation is very low even at low conversion levels, indicating that the reaction rates are highly affected by equilibrium. Thus, the simultaneous hydrogenation of mesityl oxide may help in shifting the reaction in the forward direction by keeping the mesityl oxide concentration at low level at any given time in the reactor. Thus, complete conversion of acetone was expected in the present work because of the removal of mesityl oxide in hydrogenation reaction to of methyl isobutyl ketone. However, low reaction rates were observed at higher conversion levels. This may imply that the hydrogenation reaction is equilibriumlimited and only equilibrium conversions can be obtained. However, Winter et al. [43] showed that 100% conversion mesityl oxide in liquid phase could be possible on Pd/CNF catalyst. This confirms that the observed low reaction rates on Amberlyst CH28 at higher conversion level are due to catalyst deactivation and not because of reaction equilibrium. The overall reaction is thus a pseudo equilibrium type as a result of deactivation of catalyst by the moisture formed. More evidence of the same can be found in the literature [54]. We confirmed this through the catalyst deactivation experiments described in the next section.
reaction mixture and dried in an oven at 353 K for 3 h and the same catalyst was reused in the next run. This procedure was repeated for next few runs. As shown in Fig. 4, the performance of the used catalyst was almost identical to that of the fresh catalyst. This indicates that the deactivation of catalyst is reversible and that is caused by the presence of polar components like water [1,2,55], which is produced in the dehydration of aldol. Polar compounds like water have strong affinity towards catalytic sites. In a relatively non-polar environment, water molecules are known to adsorb on the resin surface and cover the catalytic sites, thereby causing an ‘‘inhibition effect’’. As mentioned before, it should be noted that the reaction becomes very slow in this region and the kinetics, as shown in Fig. 1, looks similar to that of a reversible reaction approaching equilibrium. Since in this case, true thermodynamic reaction equilibrium is not attained, we call this state a ‘‘pseudo-equilibrium’’. A detailed investigation of the same has been reported in the literature for the reaction of acetone to mesityl oxide on a mono-functional ion exchange resin [54]. 3.5. Effect of moisture concentration
3.4. Cause of catalyst deactivation As discussed above, the catalyst gets deactivated during the course of reaction, as a result of which low rates were observed
In order to ascertain the exact role of water in the kinetics, we performed experiments with and without water. As shown in Fig. 5, we found that the presence of water in the reaction
Fig. 4. Cause of catalyst deactivation. Temperature = 393 K; catalyst loading: 5% w/w of acetone; pressure = 30 bar.
Fig. 5. Effect of water addition. Temperature = 413 K; catalyst loading: 5% w/ w of acetone; pressure = 30 bar water concentration in wt.%.
S. Talwalkar, S. Mahajani / Applied Catalysis A: General 302 (2006) 140–148
Fig. 6. Effect of temperature on reaction kinetics. Temperature in K; catalyst loading: 5% w/w of acetone; pressure = 30 bar.
system is the true cause of the observed reversible catalyst deactivation; hence, low rates of reaction are observed for the reaction with initial moisture. This implies that the rate of formation of mesityl oxide is affected due to inhibition of sulfonic acid groups by polar components like water. This phenomenon is well known for many reactions on ion exchange resins. However, mesityl oxide formation is still followed by its instantaneous hydrogenation to methyl isobutyl ketone. In short, the rate of formation of methyl isobutyl ketone is equal to the rate of formation of mesityl oxide as in the case without moisture. This suggests that simultaneous removal of mesityl oxide along with the removal of water by some means is necessary to enhance the reactor performance further. Therefore, one can regard this system to be a good candidate for multifunctional reactor system like reactive distillation or reactive adsorption. As mentioned before, except for few
145
Fig. 7. Effect of temperature on selectivity towards methyl isobutyl ketone. Temperature in K; catalyst loading: 5% w/w of acetone; pressure = 30 bar.
references [50–53] not much work has been reported in the literature on this aspect. 3.6. Effect of temperature The effect of temperature was studied over a range of 373–413 K. Fig. 6 shows the conversion of acetone at different temperatures. As expected, the conversion of acetone increases with an increase in temperature at the cost of reduced selectivity towards methyl isobutyl ketone, as shown in Fig. 7. At relatively high temperatures, mesityl oxide is likely to react with acetone to form heavy components like isophorone and other condensation products. Nicol and du Toit [45] evaluated the thermal stability of the present catalyst for the same reaction, and reported that the catalyst is stable up to 423 K.
Fig. 8. Effect of catalyst loading on (a) reaction kinetics (b) initial rate. Temperature = 393 K; catalyst loading: in w/w of acetone; pressure = 30 bar.
146
S. Talwalkar, S. Mahajani / Applied Catalysis A: General 302 (2006) 140–148
3.7. Effect of catalyst loading
4. Kinetic modeling
The reactions were performed at different catalyst loadings (2–12% w/w of acetone) and it was observed that the rate of reaction increases when the catalyst loading was increased from 2% to 5% w/w of acetone. However, the rate of reaction reaches saturation above the catalyst loading of 5% w/w of acetone, as shown in Fig. 8. Moreau et al. [56] have reported such non-liner influence of catalyst loading on the rate of hydrolysis of acetals over H-montmorillonite and strong cation exchange resin. It is expected that the number of catalyst sites above a particular catalyst loading becomes excessive and hence the rate becomes insensitive to the further increase in the catalyst loading. We believe this effect is not because of the diffusional limitations as variables such as stirrer speed etc. do not influence the overall rate of reaction under these conditions. In our opinion, such an effect can be explained only through the rigorous thermodynamic models of adsorption. A theoretical explanation for the same is lacking in the literature though some evidences have been reported [56].
O’Keefe et al. [57] proposed a mechanism for hydrogenation of mesityl oxide to methyl isobutyl ketone. The mechanism consists of five steps. In the first step, mesityl oxide adsorbs on two adjacent sites to form a di-adsorbed species. In the second step, hydrogen adsorbs dissociatively on two adjacent sites. The adsorbed hydrogen atoms are then added stepwise to the diadsorbed mesityl oxide to give methyl isobutyl ketone, which then desorbs from the catalyst surface to the bulk. The authors have confirmed this mechanism by pulse adsorption of mesityl oxide, which was monitored by diffuse reflectance infrared spectroscopy and they proposed a model based on the Langmuir–Hinshelwood approach. We assume that a similar mechanism holds good for the present system. Acetone gets adsorbed on two adjacent active acid sites and condenses to give diadsorbed dimer, which then dehydrates to form diadsorbed mesityl oxide to give methyl isobutyl ketone in an adsorbed state in a similar way to that discussed above. As discussed before, we observed that the reaction is pseudo-zero order with respect to the hydrogen concentration over the range of pressure studied. The rate equation may be given by
3.8. Effect of hydrogen pressure The solubility of hydrogen varies with pressure and hence it may influence the reaction kinetics. In order to study the effect of hydrogen pressure, we performed reactions over a range of 15–45 bar. Interestingly, it was observed that there is no effect of hydrogen pressure on the reaction kinetics, as shown in Fig. 9. This finding again confirms that the catalyst activity for hydrogenation is sufficient enough and that it is the formation of mesityl oxide that governs the overall rate of the reaction.
r¼
ka2act ð1 þ kwtr awtr þ kact aact Þ4
(3)
The adsorption term for the methyl isobutyl ketone was omitted from the model, as O’Keefe et al. [57] found that there is no effect on reaction kinetics when methyl isobutyl ketone was added to the reaction mixture at time equal to zero. Also, the parameter fitting with inclusion of this term gave us very low values for the adsorption coefficient of methyl isobutyl ketone. 5. Parameter estimation The mole balance for the batch reactor for all the components can be written as dni ¼ MCAT yi r dt
(4)
The temperature dependency of the rate constant can be expressed by the Arrhenius equation E k ¼ k0 exp (5) RT The objective function to be minimized is given as X min F ¼ ðxi;calculated xi;measured Þ2
(6)
All samples
Fig. 9. Effect of hydrogen pressure on the reaction kinetics. Temperature = 393 K; catalyst loading 5% w/w of acetone; pressure in bar.
For optimisation, a SQP (Sequential Quadratic Programming) approach from NAG library was used in the DIVA [58] simulation environment. The non-idealities of liquid phase were described using the UNIQUAC method and the thermodynamic parameters were obtained from the data bank of the commercial package ASPEN. Hydrogen solubility was predicted by using Henry’s law and interaction parameters to calculate Henry’s constant were also taken from the commercial package ASPEN. The estimated parameters for the kinetic
S. Talwalkar, S. Mahajani / Applied Catalysis A: General 302 (2006) 140–148
147
Table 3 Kinetic parameters for proposed model k8 (kmol/(g s)) E8 (kJ/mol) Kwtr Kact f
5.8714 107 97.60 4.22 2.58 2.3 103
Fig. 12. Model predictions for effect of moisture conversion of acetone. Temperature = 413 K; catalyst loading: 5% w/w of acetone; pressure = 30 bar, moisture concentration in wt.%.
Fig. 10. Comparison between modelled and measured conversion of acetone at different temperatures. Temperature in K; catalyst loading: 5% w/w of acetone; pressure = 30 bar.
model are given in Table 3. As shown in Figs. 10 and 11, the model captures the effect of parameters like temperature and catalyst loading. It should be noted that the model is applicable for catalyst loading up to 5% w/w of acetone, beyond which the rate change is non-liner with catalyst loading (see Fig. 8). The model was thus used to independently predict the performance of a batch reactor with initial moisture present. The predicted results are compared with the experimental ones in Fig. 12. The model slightly over-predicts the rate in the presence of initial moisture content. Another approach to capture the effect of moisture by considering Freundlich adsorption of water [55] was also applied. However, the fitting was worse by the latter approach. 6. Conclusions
Fig. 11. Comparison between modelled and measured conversion of acetone at different values of catalyst loadings. Temperature = 393 K; catalyst loading: in % w/w of acetone; pressure = 30 bar.
Kinetics of the one-pot synthesis of methyl isobutyl ketone from acetone was studied. The effects of various parameters like temperature and catalyst loading were investigated in detail. It was observed that catalyst undergoes reversible deactivation due to the formation of water in the condensation reaction. As a result, lower rates are observed at higher conversion levels of acetone, showing pseudo-equilibrium, which was confirmed by independent reactions. About 45% conversion of acetone with more than 95% selectivity towards methyl isobutyl ketone was realised in about 3.5 h at 393 K. This is promising, considering the performance of different reported catalysts, like Pd-ZSM-5 [24], Pd–C–Nb2O5 [12,13], Pd–CuO/MgO/SrO [8], Pd-H-ZSM-5 [37], Cu–MgO [26] and Pd-CS-H-ZSM-5 [16] with 20–60% conversion of acetone and 30–90% selectivity towards methyl isobutyl ketone under similar or even more harsh conditions. The pseudo-equilibrium due to adsorption of water may be considered as a possible
148
S. Talwalkar, S. Mahajani / Applied Catalysis A: General 302 (2006) 140–148
limitation of ion exchange resin catalyst. However, the simultaneous removal of water during the course of reaction will help to overcome this drawback. A kinetic model that considers the adsorption of water on the catalyst is proposed which satisfactorily explains the experimental observations. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29]
A. Chakrabarti, M.M. Sharma, React. Polym. 20 (1993) 1–45. M.M. Sharma, React. Funct. Polym. 26 (1995) 3–23. S. Veibel, J.I. Nielsen, Tetrahedron 23 (1967) 1723–1733. L. Ba¨rfacker, D. Tom, P. Eilbracht, Tetrahedron Lett. 40 (1999) 4031– 4034. F. King, G.J. Kelly, Catal. Today 73 (2002) 75–81. S. Guseva, A. Grigor’eva, T. Prozorovskaya, Neftepererabotka i Neftekhimiya 6 (1978) 30–32. S. Wen, Z. Li, Z. Qian, X. Mao, Jilin Daxue Ziran Kexue Xuebao 1 (1981) 99–108. Y. Too, T. Nakayama, JP62258335 (1987). Y. Higashio, K. Hamada, EP87-308156 (1987). T. Maki, T. Yokoyama, Y. Sumino, JP63096147 (1988). A.A. Grigor’ev, S.I. Guseva, E.A. Katsman, Khimicheskaya Promyshlennost 7 (1988) 392–394. T. Maki, T. Yokoyama, Y. Sumino, JP63068538 (1988). T. Maki, T. Yokoyama, Y. Sumino, JP63068539 (1988). T. Maki, T. Yokoyama, Y. Sumino, JP63096146 (1988). Y. Qian, G. Ding, H. Wu, Y. Zheng, J. Jiang, Huaxue Fanying Gongcheng Yu Gongyi 4 (1988) 89–93. H. Chin, K. Shu, Y. Chin, N. Cho, B. Hayashi, S. So, JP02069430 (1990). Y. Too, T. Nakayama, K. Shioda, JP03093743 (1991). T. Ushikubo, Y. Kakuno, H. Tsujiura, EP458216 (1991). J. Pasek, J. Hajek, J. Koubek, S. Zicha, P. Pavlas, P. Kralik, A. Kospach, A. CS273524 (1992). L. Chu, Z. Yao, CN1069674 (1993). K. Bian, L. Ding, J. Gao, H. Gu, Ziran Kexue 31 (1995) 63–67. K. Tao, X. Qiu, Y. Wang, CN1108151 (1995). S. Narayanan, R. Unnikrishnan, Appl. Catal. A: Gen. 145 (1996) 231–236. P. Chen, S. Chu, K. Wu, W. Lin, US Patent 5,684,207 (1997). B. Coh, J. Hur, H. Lee, Chem. Lett. 7 (1998) 583–584. V. Chikan, M. Arpad, B. Katalin, J. Catal. 184 (1999) 134–143. L. Melo, G. Giannetto, L. Cardozo, A. Llanos, L. Garcia, P. Magnoux, M. Guisnet, F. Alvarez, Catal. Lett. 60 (1999) 217–222. B. Coh, J. Hur, H. Lee, Stud. Surf. Sci. Catal. 121 (1999) 453–456. S. Yang, Y. Wu, Appl. Catal. A: Gen. 192 (2000) 211–220.
[30] M. Montes, M. Martin, A. Diaz, J. Benitez, J. Odriozola, Mater. Sci. Forum (2000) 325–326. [31] D. Bombos, G. Bozga, M. Bombos, A. Stefan, I. Stanciu, Chem. Papers 54 (2000) 171–176. [32] Y. Chen, B. Liaw, H. Tan, K. Shen, Appl. Catal. A: Gen. 205 (2001) 61–69. [33] H. Lee, B. Coh, Proceedings of the 221st ACS National Meeting, 2001. [34] Z. Lu, W. Ai, L. Song, L. Xu, H. Zhu, Faming Zhuanli Shenqing Gongkai Shuomingshu 9 (2001). [35] L. Melo, A. Llanos, M. Mediavilla, D. Moronta, J. Mol. Catal. A: Chem. 177 (2002) 281–287. [36] L. Mattos, F. Noronha, J. Monteiro, J. Catal. 209 (2002) 166–176. [37] K. Lin, A. Ko, J. Chin. Chem. Soc. 49 (2002) 935–942. [38] Z. Lu, F. Ai, L. Song, W. Liang, CN1331071 (2002). [39] Y. Shen, J. Wu, Z. Zhang, Shenyang Huagong Xueyuan Xuebao 16 (2002) 261–265. [40] D. Bombos, T. Juganaru, A. Stefan, M. Bombos, E. Juganaru, RO117154 (2001). [41] M. Vandersall, R. Weinand, EP321450 (2003). [42] R. Morales, L. Melo, J. Brito, A. Llanos, D. Moronta, L. Albornoz, E. Rodriguez, J. Mol. Catal. A: Chem. 203 (2003) 277–286. [43] F. Winter, A. van Dillen, K. de Jong, J. Mol. Catal. A: Chem. 219 (2004) 273–281. [44] X. Sun, D. Lu, X. Bu, Z. Wang, H. Liu, D. Zhou, Jilin Huagong Xueyuan Xuebao 21 (2004) 22–23. [45] W. Nicol, E. du Toit, Chem. Eng. Process. 43 (2004) 1539–1545. [46] F. Ai, Z. Lu, L. Song, H. Wang, S. Wang, Shunnong, CN1488617 (2004). [47] P. Yang, J. Yu, Z. Wang, M. Xu, Q. Liu, X. Yang, T. Wu, Catal. Commun. 6 (2005) 107–111. [48] P. Yang, Z. Wang, J. Yu, M. Xu, Q. Liu, S. Jing, T. Wu, Polish J. Chem. 79 (2005) 121–127. [49] Y. Shen, H. Zhang, J. Wu, Z. Zhang, Huagong Shengchan Yu Jishu 12 (2005) 9–11. [50] K. Lawson, B. Nkosi, WO9965851 (1999). [51] N. Saayman, G. Lund, WO0222542 (2002). [52] N. Saayman, G. Lund, US2003065227 (2003). [53] K. Lawson, MXPA00011793 (2002). [54] E. du Toit, R. Schwarzer, W. Nicol, Chem. Eng. Sci. 59 (2004) 5545–5550. [55] E. du Toit, W. Nicol, Appl. Catal. A: Gen. 277 (2004) 219–225. [56] C. Moreau, R. Durand, P. Geneste, S. Mseddi, J. Mol. Catal. 112 (1996) 133–141. [57] W.K. O’Keefe, M. Jiang, F.T.T. Ng, G.L. Rempel, Chem. Eng. Sci. 60 (2005) 4131–4140. [58] M. Mangold, A. Kienle, E. Gilles, K. Chem. Eng. Sci. 55 (2000) 441. [59] http://www.rohmhaas.com/ionexchange/IP/ch28_typical.htm, seen on 20 August 2005.