Studies in Surface Science and Catalysis 133 G.F. Fromentand K.C. Waugh (Editors) 9 2001 Elsevier Science B.V. All rights reserved.
Catalyst Development for Methanol and Dimethyl Ether Production from Blast Furnace Off Gas Jun-ichiro Yagia, Tomohiro Akiyama b and Atsushi Muramatsu a alnstitute for Advanced Materials Processing, Tohoku University 2-1-1, Katahira, Aoba-ku, Sendai, 980-8577 Japan. bDept. Chemical Engineering, Osaka Prefecture University 1-1, Gakuen-cho, Sakai, Osaka-fu, 599-8531 Japan For C u - Z n O - A 1 2 0 3 catalyst, compositional research was on two parameters, namely Cu/ZnO ratio (3/7-5/5) and content of A1203 (0-33.0 mol%). The influence of each parameter was estimated by DME and MeOH yields for the same catalyst mass. Some Cu-ZnO-AI203 catalysts synthesized DME more than MeOH, in which the DME activity was related to specific surface area and existence of broadened ZnO peaks in XRD pattems. In contrast, three other Cu-ZnO catalysts including another oxide (Cr203, ZrO2 or Ga203) synthesized only MeOH without DME. Content of A1203 was more influential on DME synthesis than Cu/ZnO ratio. Physically-admixed (hybrid) catalysts of Cu-ZnO-Ga203 for CH3OH synthesis and y-AI203 for its dehydration were experimentally studied, in which the influence of mixing ratio of the two catalysts on both yield and selectivity of DME was mainly examined for the same catalyst mass. The results showed that the developed hybrid catalyst is very effective in producing DME directly from BFG without equilibrium limit of methanol. Interestingly, the yield of DME had a significant dependence of mixing ratio, and the hybrid catalyst with only 5 mass% ~/-A1203 showed the highest yield with 99.3%selectivity of DME + methanol. This implies that methanol formation governs the rate of this series reaction (BFG-->CH3OH-->DME) due to fast dehydration of methanol to DME. In conclusion, the most active composition of Cu-ZnO-A1203 catalyst for DME synthesis was Cu/ZnO=4/6 with 14.3mo1% A1203. However catalysts mixture of C u - Z n O - G a 2 0 3 ( 9 5 % ) and ~,-A1203 (5%) showed higher activity.
Key Words: methanol, dimethyl ether, catalyst, blast furnace, energy, industrial waste gas, CO-CO2-H2, Cu-ZnO-A1203, 7-A1203. 1. INTRODUCTION Methanol (CH3OH, MeOH) is now attracting worldwide attention because it is not only an important industrial raw material for chemicals such as formaldehyde 1), but is also a clean fuel which burns without emission of NOx and SOx. Previously, Feasibility studies 2-5) of MeOH production from blast furnace offgas (BFG) have been carried out theoretically and experimentally for energy saving and environmental protection due to reduction of carbon dioxide emission. Since composition of BFG is quite different in comparison to industrial
436 synthesis gas (syngas) for MeOH synthesis, it leads to serious thermodynamic restriction of MeOH yield due to less hydrogen concentration and higher carbon dioxide concentration. Therefore, we need to overcome this problem to promote the effective use of BFG. Dimethyl ether ((CH3)20, DME) is also expected to be a clean energy source with large calorific value and excellent transportation properties, almost same as LPG 6). Industrially, DME is generally produced in a two-step process 6), namely, MeOH formation and its dehydration. It should be pointed out that its equilibrium yield is far beyond that of MeOH. Therefore the use of a bifunctional catalytic system7), that is, a combination of a MeOH synthesis component with a dehydration partner s 12) can avoid the equilibrium limit of MeOH. Direct synthesis of DME from syngas (CO+H2) has been reported in the literature. 7' 12, 13) The method has been directed so far, to only the typical syngas; carbon monoxide/hydrogen with or without a small content of carbon dioxide, but not yet to BFG with a large carbon dioxide content. In addition, influence of the mixing ratio in the hybrid catalyst has not been well investigated in spite of its importance for the process design and operation. Therefore, the aim of this study is to determine an optimum composition of Cu, ZnO and A1203 as high activity catalysts for high yield of oxygenates (MeOH + DME) from BFG, in which Cr203, Zr02 or Ga203 were also examined instead of A1203 and to investigate catalytic activity and mixing ratio of the hybrid catalyst which is a physical mixture of Cu-ZnO-Ga203 and y-A1203. 2. MATERIALS For synthesizing DME from COCOa-Ha mixtures such as BF gas, we prepared catalysts of Cu-ZnO-X (X= A1203, Cr203, ZrO2, Ga203) by the coprecipitation method. Chemical composition of the catalysts prepared is given in Table 1. Twelve catalysts of the Cu-ZnO-A1203 system were prepared, in which additive content of alumina was systematically changed under three different ratios of Cu/ZnO. In addition, three catalysts with different oxides (Cr203, ZrO2, Ga203) instead of alumina were also prepared for comparison. The preparation procedure of the catalysts was reported in detail elsewhere. 3' 5,6,) For the Cu-ZnO-A1203 catalyst, the procedure framework is as follows; (1) Three commercial chemicals of Cu(NO3)2"3H20, Zn(NO3)2"6H20 and Al(NO3)3"9H20 are first dissolved in distilled water at room temperature, to
1. Chemical composition of fifteen catalysts used.
Cu/ZnO/A 1203 system Cu/ZnO A 1203 mol 6.7 14.3 23. 1 33.0
Composition, mol% 30.0/70.0/0.0 28.0/65.3/6.7 25.7/60.0/14. 3 23.1/53.8/23.1 20. 0/47.0/33. 0
0.0 6.7 14.3 23. 1
40. O/GO.0/0. 0 37. 3/56. 0/6.7 34. 3/51.4/14. 3 30. 8/46.1/23. 1
6.7 14.3 23. 1
46. 7/46. 7/6.7 42.9/42.9/14.3 38. 5/38.5/23. 1
Cu/ZnO/Oxide (Cr203, ZrO2, 6a203) system Cu/ZnO Oxide mol~ ] Composit ion, mol~ 3/7 14. 3 I 25. 7/60.0/14. 3
437 obtain aqueous solutions of 1 mol/dm 3 with the desired molecular ratio of the metallic elements. (2) The obtained solution and 0.7 mol/dm3 sodium carbonate (Na2CO3-10H20) solution are mixed at 353 K all at once, to precipitate the carbonates of Cu, Zn and A1. The precipitate, recovered by filtration, is compressed to from a cylindrical sample and then sintered at 673 K after sufficient drying at 393 K. (3) The sinter is crushed in a mortar and particles of size 0.35-0.84 mm are separated by sieving for packing into a reaction tube. In preparing hybrid catalyst, two catalysts of CuZnO-Ga203 and y-A1203 are admixed at different mass ratios and then fixed into a reaction tube. (4) The packed catalyst is finally reduced by pure hydrogen at 653 K just before the experiments. 3. METHOD
Figure 1 shows a schematic diagram of an experimental system for DME synthesis, consisting of three parts: gas supplier, fixed-bed reactor and gas analyzer. The molar ratio of feed gas excluding nitrogen, CO/CO2/H2=4/4/1 (molar basis), corresponds to operating data 2) of a BF with natural gas injection of 50 Nm 3 per ton-hot-metal. Nitrogen was added to the feed gas to avoid CO2 liquefaction in the cylinder so the composition of the actual feedgas is CO/CO2/H2~2 =35/35/8.7/21.3. After the feedgas was introduced, outlet gas from the fixedbed reactor was analyzed by two kinds of gas chromatography: FID (Flame Ionization Detector) for DME, MeOH and hydrocarbon (HC, mainly methane), and TCD (Thermal Conductivity Detector) for CO, CO2, H2 and N2.
438 An enlargement of the reactor is shown in Figure 2 The reactor employed was made of SUS316 stainless steel tube with 10 mm outer diameter and 8 mm inner diameter. The prepared catalyst was fixed at the middle position of the uniform temperature zone, on which glass beads were placed for preheating the feed gas and for obtaining radically uniform flow. To measure temperature changes of the catalyst during the experiment, a CA type thermocouple, 1 mm in outer diameter, was inserted into the bed. The experimental conditions of this study were 523 K, 1.0 MPa inlet pressure and 1.0xl0 -6 m3/sec (STP) flow rate of the feed gas. These were determined based on our previous studies 3' 4). The measurements were continued until outlet gas composition reached a steady state. In the experiment for finding optimal mixing ratio of 7-A1203 to the hybrid catalyst, mass ratio of ~/-A1203 was changed from 0 to 100% keeping the hybrid catalyst mass 2.0g. Since the densities of 7- AlaO3 and Cu-ZnO-GazO3 are considerably different, the height of the hybrid-catalyst bed depends on their mixing ratio, resulting in much different residence times. Therefore, not only the catalyst mass(2.0g) but also the residence time have to be considered in the unit of C-mol%/g-cat s, for quantitatively evaluating yields of DME, MeOH and HC. They are defined as follows: 2CI:,MV. CO + c co2
0 + Cco2 CHC
W x (L / u )
(C~ + C~ )/ lo0 • w • (L / u)
4. RESULTS AND DISCUSSION All monitored components of the outlet gas (CO, CO2, He, MeOH, DME, HC, N2) reached constant values with negligible further temperature changes after approximately 6 hours. The obtained concentrations of MeOH and DME for each catalyst were then defined as their final yields. Figure 3 shows bar graphs of evaluated yields of DME and MeOH, for all of the catalysts prepared. Here, Cu/ZnO ratio in the catalysts in the horizontal direction of the graph is constant and the three rows of the graph are for different Cu/ZnO ratios: 3/7, 4/6 and 5/5. The bar height indicates the DME and MeOH yields. Obviously, some catalysts of Cu-ZnOA1203 gave good yields of DME with little MeOH synthesis although only MeOH is normally synthesized from CO-H2 gas as mentioned later. The results suggest significant influence of A1203 content on DME yields, in comparison to Cu/ZnO ratio. The highest yield of DME+MeOH was given by the catalyst of 14.3 % A1203 with Cu/ZnO=4/6. In contrast, addition of oxides of Cr203, ZrO2 and Ga203 instead of A1203 did not cause DME generation, although Cu-ZnO-Ga203 showed the highest yield of MeOH among all of the catalysts. Strangely, Cu-ZnO catalysts without A1203 showed little activity for DME and MeOH syntheses, in spite of experimental reports 8) claiming that Cu-ZnO is effective for synthesizing MeOH from mixtures of CO, H2 and a little CO2. In addition, they claimed no effect of Cu-
439 0. 2 0.15
0.1 "-" ~
-o 0 9~
0.1 O. 040
0.05 O. 004
14. 3 tool%
14. 3mo1% Ox i de
Fig.3 Measured yields of methanol and dimethyl ether for all catalysts used. ZnO-A1203 catalyst on DME synthesis at all, although the chemical composition of industrial catalysts (Cu: 12--66 mol%, ZnO: 17-62 mol%; A1203:4-38 mol%) covers this study's best catalyst (34.3/51.4/14.3). This is mainly due to syngas composition (CO-CO2-H2). Regarding the relationship between syngas composition and catalyst composition, a noteworthy result was recently reported by Fujita and coworkers a2). They examined the catalyst mechanism of Cu-ZnO under CO-H2 and CO2-H2 by means of diffuse reflectance FT-IR spectroscopy. The result demonstrated that Cu promotes CO2 conversion to MeOH and ZnO promotes CO conversion. This suggests that the CO/CO2 ratio of syngas would have significant influence on the optimum catalyst composition in the Cu/ZnO system. Therefore a similar study should be extended to CO-CO2-H2 mixtures for constructing an advanced catalyst theory. To make clear the effects of A1203 content and Cu/ZnO ratio in the Cu-ZnO-A1203 system, we redrew data in the following 0.1 :-. . . . . . . . .-.-. -. .h. . . . . . . . . . . . . . . . . I 9 DME i figure. Figure 4 shows the effect of A1203 content on DME and MeOH yields with Cu/ZnO=3/7. The sharp rise in DME yield indicates strong sensitivity to A1203 content, in comparison to the MeOH yield. Both peaks of DME and MeOH curves are identically located at 14.3 mol% A1203. Similar tendencies were also observed for the other Cu/ZnO ratios of 4/6 and 5/5. This is explained well by the concept 15) that DME is produced by a series reaction via MeOH synthesis (BF gas --> MeOH ~ DME).
o.o. I.............................. ./ ,.:,E 0.06
~= o.o4 il. . . .
-/-- . . . . . .
0 9 02 i....... _ .O. :::--. / ~
':":-""O ".. . . . . . . . .
.... . ............ ............ .....
15 20 25 30 35Cr~Oa ZrO~ Ga2Oa MoI% AI2Oa 14.3 mol%-Oxide
Fig.4 Effect of alumina content on dimethyl ether and methanol yields over catalysts of Cu/ZnO=3/7, together with other oxides
440 Figure 5 shows measured surface 01 . . . . I . . . . i . . . . i . . . . ~. . . . ~. . . . ! . . . . area of the catalysts of Cu/ZnO=3/7. I The measured surface area showed ,-. 501............................................................................................................................................ , .......... I strong A1203-content dependence. I Ga203 Interestingly, this result can explain ~ o 401 ........................................................................................................................................................ well the catalyst property of the CuZnO-A1203 system (see Fig.3). That is, we can conclude that the catalysts with large surface area cause high yields of 201....................../ .................. \ ............l..........'-..................-'-............................i DME. The surface area of 14.3 and 23.1 mol% A1203 catalysts (approximately 33 m2/g) are more three times higher in comparison to the 0 catalyst without A1203. Similarly, the 0 5 10 15 20 25 30 35 14. 3 mol~ Oxide 14.3 mol% Ga203 catalyst showed mo196 AI203 large surface area and good catalyst Fig.5 Effect of A1203 content and oxides on properties. However, this applied to specific surface area of catalysts with only MeOH synthesis, not DME molar ratio of Cu/ZnO=3/7 synthesis. Figure 6 shows XRD (X-ray diffraction) patterns of the Cu-ZnOA1203 catalysts used, in which three ~ ~~ II . . . . . o ZaO substances of Cu, ZnO and ZnA1204 m 9 ^ 33.0 mol% AI203 were detected as main peaks. Peaks of A1 or A1203 never appeared in spite of O ,, J ~ 23.1 tool% AI.O. our expectation. With increasing m.~, A V 9 o o'" 9 . ," ",~.,,~~,~,.~. A1203 from 0 to 14.3 mol%, peaks of o Cu and ZnO were broadened and their r l 14.3 tool% A120:3 ~9 ~ o D o o heights became smaller. In this range, the catalysts became more active for  I o ,  l ' 6.7 mol% AJ20. DME and MeOH synthesis with l,l~ l, n o g um o o " increasing A1203 (see Fig. 4). In contrast, 23.1 and 33.0 mol% A1203 ? I o catalysts showed ZnAl204 peaks o f , instead of ZnO peaks, however its 'D' t Ill 'i o o El 0rn~ l peak width was very broad. In i lU II~, i ' O O O I addition, the Cu peaks of the 33.0 mol% catalyst became very weak. 20 40 60 80 100 According to these results, we can 2 t9 ( d e g r e e , CuK a ) expect that the optimum composition of the catalyst would range between Fig.6 XRD patterns of catalysts with molar ratio of Cu/ZnO=3/7 14.3 and 23.1 mol% A1203. As for the mechanism of DME direct synthesis from CO-CO2-H2, A1203 microstructure must be a key point as mentioned above. Microscopic research of catalyst structure is also needed in the future. Figure 7 shows an example of changes in gas concentration with time, where the hybrid catalyst with 5mass% - A1203 was employed. It is clear that, DME is synthesized. This
i i ;i; . . . . . . . . . . . . . . . . .,. . . . . . . . . . . . . . . .
441 demonstrates that DME synthesis from BFG over the hybrid catalyst is a typical series reaction: BFG
>D M E
Concentration of BFG decreases with time at the initial stage, then, it remains steady after aboutl0 ks. Conversely, DME concentration increases and shows a peak, corresponding to the initial decrease of BFG concentration. In contrast, such a peak is not observed in MeOH concentration. It takes about 15 ks for the hybrid catalysts to reach steady state. The same tendency to form a DME peak in an initial stage is found in other hybrid catalysts. This is probably caused by first conversion of DME from MeOH, in comparison to MeOH synthesis from BFG.
. . . . ! . .......... . . .i . i...........
o.o6 ~t ....... ~
....... ::. . . . . .
Time / ks
Fig.7 Concentration-time curves for DME, MeOH and C0+CO2+H2 in outlet gas. 5mass% y-A12o3 + 95mass % _ Cu_ZnO_Ga203 _ _ _ _
T -A1203 / mass%
Cu-ZnO-Ga203 / mass%
Effect of mixing ratio in the hybrid Fig.8 Effect of hybrid catalyst on DME, catalyst on yields of DME, MeOH and HC MeOH and HC. is shown in Figure 8, together with changes (Cu-ZnO-Ga2Q3)+ y-A1203, 2.0g in bed height. Yield of DME has very strong W/F=3.44x 10 -~g-cat's/mo l, dependence on the mixing ratio, showing CO/CO2/HE/N2=35/35/8.7/21.3, peak of DME yield at 5 mass% y-AI203 1.0Mpa, 523K addition. In contrast, HC yield is negligible in all cases. The selectivity of DME+MeOH is 99.3% on carbon mole basis. The MeOH yield is almost constant in the region 50 to 90 mass% of Cu-ZnO-Ga203 in spite of an increase of the methanol-active catalyst. This was probably caused by the fact that DME synthesis is a series reaction via MeOH from BFG and intermediately synthesized MeOH is very quickly converted to DME. The reaction of MeOH synthesis from BFG was the rate controlling step under this experimental condition. It was, therefore, concluded that the hybrid catalyst
442 developed is quite promising for the direct conversion of BFG to DME, and very small additions of T-A1203 to a methanol-active catalyst of Cu-ZnO-Ga203 is enough because the rate of dehydration due to T-AhO3 is much larger. These results appealed to the possibility to promote effective use of BFG by overcoming the thermodynamic limitation of MeOH yield.
5. CONCLUSION Fifteen kinds of single component catalysts in Cu-ZnO-A1203 and Cu-ZnO-Oxide (Cr203, ZrO2, Ga203) systems and the hybrid catalyst of Cu-ZnO-Ga203 and y-A1203 were systematically studied for direct synthesis of DME from CO- CO2-H2 mixture. The experimental conditions were 523 K, 1.0 MPa pressure and 2.0 g catalyst weight. The following conclusions were drawn: (1)The most noteworthy finding is the significant catalytic activity of Cu-ZnO-AI203 in DME synthesis, not MeOH, from CO-CO2-H2 mixture. The catalyst of 34.3Cu-51.4ZnO-14.3 A1203 (Cu/ZnO=4/6 (mole basis)) showed the best activity for DME synthesis. (2)The catalytic mechanism of the system of Cu-ZnO-AI203 is still not well explained. However, catalysts showed strong relationship between DME activity and surface area, and XRD peaks of ZnO in active catalysts tend to be broadened without Zn and ZnA1204 peaks. (3)Maximum yield of DME+MeOH was given by the hybrid catalyst with 5mass% y-AI203 and 95mass%Cu-ZnO-Ga203. As for the BFG conversion to MeOH, the single catalyst of Cu-ZnO-Ga203 reached 75% of equilibrium conversion, while for the DME conversion, the hybrid catalyst of 5mass% y-A1203 had only 4% in spite of large yield. (4)Rate controlling step of DME synthesis in the hybrid catalyst system was MeOH formation. In other words, y-A1203 is better catalyst for the reaction from MeOH to DME. For enhancing the DME conversion, development of more active catalyst for MeOH synthesis is expected.
1. Y. Ohno: Personal communication (1996). 2. T. Akiyama, H. Sato, A. Muramatsu and J. Yagi: ISIJ Int., 33 (1993), 1136. 3. A. Muramatsu, H. Sato, T. Akiyama and J. Yagi: ISIJ Int., 33 (1993), 1144. 4. H. Sato, T. Akiyama, A. Muramatsu, T. Sugimoto and J. Yagi: J. Jpn. Pet. Inst., 38 (1995), 390. 5. H. Sato: Ph. D. Thesis, Tohoku University, (1996). 6. H. Hover: Ullmann's Encyclopedia of Industrial Chemistry, A8 (1989), 541. 7. A.C. Sofianos and M.S. Scurrell: Ind. Eng. Chem. Res, 30 (1991), 2372. 8. R.G. Harman, K. Klier, G.W. Simmons, B.P. Finn, J.B. Bulko and T.P. Kobylinski: J. Catal., 56, (1979), 407. 9. B. Peplinski, W.E.S. Unger and I. Grohmann: Appl. Sur. Sci., 62 (1992), 115. 10. R. Burch, S.E. Golunski and M.S. Spencer: J. Chem Soc. Faraday Trans., 86 (15), (1990), 2683. 11. I. Nakamura, T. Fujitani, T. Uchijima and J. Nakamura: J. Vac. Sci. Technol., A14 (3),
443 (1996), 1464. 12. S. Fujita, M. Usui, H. Ito and N. Takezawa: J. Catal., 157, (1995), 403. 13. Y. Han, K. Asami and K. Fujimoto: J. Japan Inst. of Energy, 75, (1996), 42. 14. K. Fujimoto, K. Asami, H. Saima, T. Shikada and H. Tominaga: Ind. Eng. Chem. Prod. Res. Dev., 25, (1986), 262. 15. S. Machida, T. Akiyama, A. Muramatsu and J. Yagi: ISIJ Int., 37 (1997), 536.