1548
Znd. Eng. Chem. Res. 1995,34, 1548-1557
Production of Hydrogen from Methanol over Promoted Coprecipitated Cu-AI Catalysts: The Effects of Various Promoters and Catalyst Activation Methods Raphael 0. Idem and Narendra N. Bakhshi" Catalysis and Chemical Reaction Engineering Laboratory, Department of Chemical Engineering, University of Saskatchewan, Saskatoon, Saskatchewan, Canada S7N OW0
The influence of various promoters and activation methods on the low-temperature methanolsteam reforming performance of promoted Cu-Al catalysts was studied at atmospheric pressure in a microreactor over a reaction temperature range of 170-250 "C at a methanol space velocity (WHSV) of 16.7 h-l. Maximum methanol conversions of 93 and 99 mol % and H2 selectivities of 99 and 93 mol % were obtained from the Mn-promoted and vaporized feed-activated catalyst at 200 and 250 "C, respectively. The overall results showed that favorable catalyst characteristics were such that the promoter and the activation process not only conditioned the catalyst to produce optimum amounts of active Cuo,Cu20, CUI,and Bronsted ZnO base sites but also ensured that these catalyst sites were maintained during the reaction.
Introduction There is a strong desire t o use H2 as an energy vector in the transportation sector, especially in low-temperature applications such as the fuel cell, because of its high efficiency and also for environmental reasons (Amphlett et al., 1993). The difficulties in H2 storage, transportation, and handling which are associated with this mode of energy application are generally circumvented by the use of methanol as a hydrogen carrier. H2 can then be produced by the methanol-steam reforming reaction. However, the commercial CuO/ZnO and CuO-ZnO/A.l203 catalysts which are currently used for this process require temperatures as high as 280 "C for efficient operation. It is highly desirable to operate the reformer at low reaction temperatures if it is to be coupled t o a fuel cell. This is because, in addition to the temperature mismatch between the reactor and the fuel cell, high reaction temperatures also result in the production of carbon monoxide which is a potential poison for a typical platinum-catalyzed solid polymer fuel cell (Amphlett et al., 1991). In a recent study, Idem and Bakhshi (1994b) have established the optimum conditions for maximum H2 production from methanol using coprecipitated Cu-Al catalysts. According to these authors, maximum H2 production efficiency of 78 mol % was obtained in the methanol-steam reforming reaction at 250 "C using a catalyst containing 27.8 w t % copper and calcined at 700 "C. This same catalyst produced a H2 production efficiency of 60 mol % at a methanol-steam reforming temperature of 200 "C. Also, these authors showed that H2 production efficiency increased with the amount of Cu20 species in the activated catalyst. In addition, it was observed that the amount of this active Cu20 species was a strong function of both catalyst copper concentration and activation method (i.e., activation by reduction in a H2 atmosphere followed by mild reoxidation with either methanol alone or a 1:l molar mixture of methanol and water). These results appear to indicate that the performance of coprecipitated CuAl catalysts could be improved by manipulating both the catalyst copper concentration and its subsequent
* Author to whom correspondence should be addressed.
activation to obtain a large amount of active CuzO species in the activated catalyst. On the other hand, Klisurski (1970) has made a comparative study of the performance of a number of single metal oxide catalysts in the oxidation of methanol to carbon dioxide. This worker observed that the overall performance of these catalysts in terms of both the rate of methanol oxidation and selectivity for COSproduction increased with catalyst reducibility in the order V205 < Fez03 < NiO < Mn203 < C0304 < CuO. In addition to a similar superior performance by CuO over other single metal oxide catalysts for the methanol-steam reforming reaction, Kobayashi et al. (1976) showed that a synergistic behavior existed in the performance of binary metal oxides formed between CuO and certain metal oxides such as Cu-AI, Cu-Mn, Cu-Cr, Cu-Ni, Cu-Co, and Cu-Si binary metal oxides. Amphlett et al. (1990) have attempted to exploit this useful characteristic by incorporating oxides of Mn, Cr, and V promoters into a CuO-MgO catalyst. However, the resulting catalysts did not produce the desired high catalytic performance. It is not certain from their work whether the optimum concentrations at which the various promoters would exhibit their optimum promotion effects were achieved since these authors reported the catalytic performance for only one concentration level for each promoter. Also, it is known that, to obtain the maximum benefits of the synergistic effects of any promoter, it is highly desirable to employ promoters and unpromoted catalysts with high unmodified activities in the envisaged reaction. Unmodified CuO-MgO catalyst has been reported by Amphlett et al. (1985) to be highly effective in the methanol-steam reforming reaction. However, our recent work (Idem and Bakhshi, 199413)has shown that coprecipitated Cu-Al catalysts are highly active and effective for the low-temperature methanol-steam reforming reaction. This therefore suggests that the incorporation of active promoters such as Mn, Cr, and Zn (Kobayashi et al., 1976; Klisurski, 1970; Amphlett et al., 1988) could have tremendous effects on the performance of these coprecipitated Cu-Al catalysts. In this work, we have studied the effects of promoter type, promoter concentration, and catalyst activation
0888-588519512634-1548$09.00/0 0 1995 American Chemical Society
Ind. Eng. Chem. Res., Vol. 34, No. 5, 1995 1549 Table 1. Identities and Elemental Compositions of Calcined Promoted and Nonpromoted Coprecipitated Cu-AI Catalysts catalyst
promoter none Mn
Cu-AI catalyst precursor, w t % Cu 27.8 24.1 27.8 24.1
Cr
27.8 24.1
Zn
27.8 24.1
elemental composition, catalyst identity p-5C p-4C p-5CM1 p-5CM2 p-5CM3 p-4CM1 p-4CM2 p-5CM3 p-5CC1 p-5CC2 p-5CC3 p-4CC1 p-4CC2 p-4CC3 p-5cz1 p-5CZ2 p-5CZ3 p-4CZ1 p-4CZ2 p-4CZ3
wt%
Cu 46.6 38.0 44.6 39.1 33.2 36.5 30.5 26.5 40.7 40.7 25.3 30.9 30.4 18.2 45.5 36.7 31.2 35.2 27.2 24.0
Al
Ja
23.1 29.4 22.1 19.4 16.3 28.3 23.6 20.5 18.0 18.0 11.2 25.7 25.3 15.1 22.2 17.9 15.2 28.6 22.1 19.5
0 0 2.96 7.25 20.2 2.58 7.6 20.9 3.56 11.4 32.2 3.58 11.5 33.4 3.74 13.4 27.0 4.08 14.8 28.1
a J = M, C, and Z for Mn-, Cr-, and Zn-promoted catalysts, respectively.
method on the low-temperature performance of two of the best coprecipitated Cu-AI catalysts (24.1 and 27.8 wt % copper) from a previous study (Idem and Bakhshi, 1994b). This work also involved a thorough characterization study of both the calcined and activated promoted and nonpromoted catalysts (Idem and Bakhshi, 1994~).In addition, the activity and stability of the best catalyst were studied at various reaction temperatures as functions of catalyst time on stream. These results are presented in this paper.
Experimental Section Catalysts. Preparation and Characterization. Dried coprecipitated Cu-AI catalyst precursors containing 24.1 and 27.8 wt % copper were used for the preparation of the promoted catalysts. The preparation of these precursors is described in detail elsewhere (Idem and Bakhshi, 1994a). Mn, Cr, and Zn were incorporated into these dried catalyst precursors by impregnation techniques using aqueous solutions of manganous nitrate and zinc nitrate for incorporating Mn and Zn, respectively, and chromium acetate for the incorporation of Cr. Catalysts with three concentration levels of each promoter for each catalyst precursor were prepared, in addition to the two catalysts with zero promoter concentration. This procedure resulted in a total of 20 catalysts which were calcined at 700 "C.Their identities and metal compositions are given in Table 1. Each catalyst was activated in situ prior to performance studies. The activated catalysts were thoroughly characterized using the in situ TPR characterization technique recently reported by Idem (1995). Details concerning the preparation and characterization of promoted and nonpromoted catalysts are given elsewhere (Idem and Bakhshi, 1994~). Performance Studies. Equipment. The performance of the promoted and nonpromoted coprecipitated Cu-Al catalysts in the methanol-steam reforming reaction was studied in a stainless steel (SS-316) microreactor (10 mm i.d. and 460 mm overall length)
which was placed concentrically in an electrically heated furnace. Figure 1 shows the schematic diagram of the experimental rig used. The furnace temperature during a test run was controlled by a Series SR22 microprocessor-based autotuning PID temperature controller (supplied by Shimaden Co. Ltd., Tokyo, Japan) through a K-type thermocouple inserted into a heating block placed concentrically within the furnace. A separate thermocouple was used to monitor the temperature of the catalyst bed. This arrangement was capable of ensuring an accuracy of f l "C for the catalyst bed temperature. Operating Conditions. All runs were conducted at atmospheric pressure over a temperature range of 170250 "C at a methanol space velocity (WHSV)of 16.7 h-l. Each run was performed using 2 g of calcined catalyst mixed with 3 g of Pyrex glass all in the same size range of -8 to +10 mesh. This mixture was loaded on a stainless steel grid positioned centrally within the reactor. This arrangement produced a catalyst bed height to particle size ratio (LID,) greater than 40. This was done to ensure a condition as close as possible to plug flow. The flow rate of the liquid feed was regulated by a syringe pump (Eldex Model BlOO S). As mentioned previously, each catalyst was activated in situ prior to an actual test run. Two catalyst activation methods were tested. These were (i) the reduction of the catalyst in a H z - N ~mixture (3% Hd at 300 "C at a flow rate of 100 mumin for 2 h followed by treatment at 250 "C with a vaporized 1:l molar mixture of methanol and water for 1h and (ii)the mild reduction at 250 "C in a vaporized 1:l molar mixture of methanol and water for 1 h. A typical test run was performed as follows: The catalyst was loaded into the reactor and activated using the desired activation procedure before the actual test run at the desired reaction temperature. The 1:l molar mixture of methanol and water used for both the activation and the actual test run was pumped at a methanol space velocity (WHSV)of 16.7 h-l to the vaporizer maintained at about 250 "C. The vaporized feed mixture from the vaporizer entered the reactor in a stream of nitrogen gas (99.995% purity and obtained from Linde). The methanol-steam reforming reaction is an endothermic reaction. Therefore, it was necessary t o stabilize the reactor temperature before actual reaction data were taken. Details concerning reactor temperature stabilization are given elsewhere (Idem and Bakhshi, 1994b). The product mixture during the actual test run was condensed with chilled water to separate the gaseous and liquid products for separate analysis. Analysis of the Products. The gaseous product was analyzed on-line with a gas chromatograph (Model HP 5880) using a Porapak Q column in series with a Porapak R column, a thermal conductivity detector (TCD), and helium (99.995% purity and obtained from Linde) as the carrier gas. The liquid product was analyzed by manual injection into the same set of columns using the same gas chromatograph and analysis conditions as in the gaseous products.
Results and Discussion Studies were conducted for each of the 20 promoted and nonpromoted coprecipitated Cu-Al catalysts in order to evaluate the effects of promoter type and concentration as well as those of activation method on their low-temperature methanol-steam reforming performance.
1550 Ind. Eng. Chem. Res., Vol. 34, No. 5,1995
8
12
I 20
Figure 1. Schematic diagram of the experimental rig for the methanol-steam reforming reaction. Legend: 1, Nz; 2, 3% Hz in Nz; 3, pressure gauge; 4, odoff valves; 5, feed reservoir; 6, syringe pump; 7, deoxygenatioddehydration catalyst; 8, needle valve; 9, vaporizer; 10, temperature indicator; 11, temperature controller; 12, fured bed reactor; 13, catalyst bed 14, heating block; 15, electrically heated furnace; 16, insulator; 17, water-cooled condenser; 18, gadliquid separator; 19, gaseous products; 20, liquid products; 21, gas chromatograph; 22. fume hood.
Catalyst Activity. Catalyst activity was evaluated in terms of methanol conversion (mol %). The methanol conversions obtained in the methanol-steam reforming reaction over the 20 promoted and nonpromoted catalysts used in this study are given in Table 2 as a function of reaction temperature. Effects of the Promoters. Table 2 shows that, at some promoter concentrations, the activities of the various promoted catalysts in terms of methanol conversion were superior t o that of the nonpromoted coprecipitated Cu-Al catalysts. For example, in the methanol-steam reforming reaction at 250 "C, methanol conversion for the nonpromoted coprecipitated CuAl catalyst (p-5C)was 83 mol %whereas the conversions for Mn-, Cr-, and Zn-promoted catalysts (i.e., catalysts p-5CM2, p-5CC1, and p-5CZ2) were 94,90, and 92 mol %, respectively. This synergistic behavior in promoted catalysts was attributed to the presence of certain additional species such as CuMnOn, CuzCrz04, and ZnO in these catalysts resulting in the net enhancement of their catalytic potentials. This is discussed below using eqs 1-5 which describe various reaction steps usually involved in the methanol-steam reforming reaction in conjunction with the results of our earlier characterization studies on activated methanol-steam reforming catalysts (Idem and Bakhshi, 1994~).
Reaction Scheme. Equation 1represents the overall equation for this reaction (Amphlett et al., 1991). It CH,OH
+ H,O
= CO,
+ 3H,
(1)
is seen in eq 1that, while methanol is oxidized to COZ by a net loss of electrons, water is reduced to Hz by a net gain of electrons. On the other hand, eqs 2-5 represent a typical but more detailed reaction scheme employed by Su and Rei (1991), Kobayashi et al. (19761, and Idem and Bakhshi (1994b) to account for the products obtained in the methanol-steam reforming reaction. 2CH,OH = 2HCHO
+ 2H,
+ HCHO = HCOOCH, HCOOCH, = CO + CH,OH CO + H,O = CO, + H,
HCHO
(2)
(3) (4) (5)
Also, eqs 2-5 show a decrease in the electron cloud density around the carbon (C) atom as we proceed from methanol in eq 2 to carbon dioxide in eq 5 . This is evident by considering the structural formulas of the
Ind. Eng. Chem. Res., Vol. 34, No. 5, 1995 1661 Table 2. Methanol Conversions of Promoted and Nonpromoted Cu-AI Catalysts as a Function of Reaction Temperature methanol conversion: mol % catalyst optimum catalysts p-ZCJod Cu-A1 catalyst promoter precursor, w t % Cu reaction temp, "C p-1C p-ZCJ1 p-1CJ2 p-1CJ3 (iIb (iiy Mn 27.8 250 83 90 94 92 94 99 200 59 77 81 74 81 93 170 19 24 27 20 27 32 24.1 250 81 85 90 89 90 96 200 55 71 79 75 79 90 170 15 22 24 16 24 27 Cr 27.8 250 83 90 89 87 90 95 200 59 64 62 29 64 87 170 19 20 13 7 20 26 24.1 250 81 87 82 38 87 94 200 55 61 60 9 61 88 170 15 17 10 4 17 18 Zn 27.8 250 83 89 92 90 92 96 200 59 73 79 28 79 86 170 19 20 22 10 22 23 24.1 250 81 87 90 89 90 94 200 55 58 60 25 60 70 170 15 16 17 9 17 20 a 1 = 4 and 5 for precursors containing 24.1 and 27.8 wt % C, respectively. J = M, C, and Z for Mn-, Cr-, and Zn-promoted catalysts, respectively. i: catalyst activated by reduction in Hz followed by treatment in vaporized methanol-water mixture. ii: vaporized methanol-water mixture activated catalyst. 0 = 2,1, and 2 for Mn-, Cr-, and Zn-promoted catalysts containing the respective optimum promoter concentration.
species involved. Thus, it appears that catalyst activity would be enhanced if the electron transfer potential of the catalyst could be improved. It is known in catalysis literature (Biswas et al., 1988; Trimm, 1980) that catalyst activity could be influenced by the environment of the active site, especially the chemical nature of its nearest neighbors. However, improvement in catalyst performance due to such an influence has not been reported previously in the literature for the methanolsteam reforming reaction. In this study, the tremendous improvement in the activity of promoted coprecipitated Cu-AI catalyst was attributed to the ligand effect introduced into the catalyst by the incorporation of Mn, Cr, and Zn. As was discussed in our earlier work (Idem and Bakhshi, 1994c), the species present in activated unmodified coprecipitated catalysts were CuaO, Cuo,and AI203 whereas promotion with Mn, Cr, and Zn introduced CuMnOz and Mnz03, CuzCrz04 and Crz03, and ZnO species, respectively, into the activated promoted catalysts. The ligand roles of these species in the methanol-steam reforming reaction are discussed below. Mn and Cr Promoters. The presence of CuMnOp and CuzCrz04 species is responsible for the increased amount of copper in the f l oxidation state. This means an increase in the electron-accepting potential of copper and, consequently, an increase in catalyst activity in Mn- and Cr-promoted catalysts. In a previous study (Idem and Bakhshi, 1994b),it was established that methanol conversion in the methanolsteam reforming reaction increased with the Cuz0/Cuo weight ratio in the activated catalysts. The amount of CuzO species represents the contribution to methanol conversion from catalyst redox ability, especially in the 0-H bond cleavage (Wacks and Madix, 1978; Bowker and Madix, 1980; Chan and Griffin, 1986; Idem and Bakhshi, 199413). In the case of Cuo species, Matsukata et al. (19881, Amphlett et al. (19911, and Takezawa et al. (1982) have shown that a large amount of highly dispersed Cuo (i.e., high copper surface area) may result in high methanol conversions. However, in contrast to the role of CuzO species, it appears that the major function of small copper crystallites alone is the C-H bond scission (Biswas et al., 1988).
On the other hand, Agaras et al. (1988) have attributed the methanol-steam reforming activity of coppercontaining catalysts to the presence of both CUIand Cu'I species, although they did not determine the species in the activated catalysts. However, as was shown earlier by the results of our catalyst characterization studies, there were various types of CUIspecies in the activated promoted catalysts. This suggests that catalyst activity may be correlated with the Cul/Cuo weight ratio of species in the activated catalyst. In addition to the specific roles of CuzO and Cuo mentioned earlier, the Cul/Cuo weight ratio represents the electron-accepting potential of the activated catalysts under steady state methanol-steam reforming conditions. This potential can be maintained because it is possible for Mn"' and CrlI1 ions to alternate between a number of other oxidation states. This property enhances their ligand role of apparent acceptance of electrons from methanol and back donation of these electrons to water on behalf of CUI during the reaction. That is why Mn- and Crpromoted catalysts with promoter concentrations yielding Cul/Cuo weight ratios larger than those in the nonpromoted catalysts exhibited correspondinglylarger methanol conversions. Zn Promoter. In the case of Zn-promoted catalysts where the Zn" ions in ZnO species are incapable of changing their oxidation state, the increase in catalyst activity was attributed t o the behavior of ZnO as a Bronsted base. According to Chan and Griffin (1986), the function of ZnO is for its surface 02anions to stabilize the adsorbed methoxy species (CH30(a))by accepting the spillover proton (H(a))obtained from the dissociative adsorption of methanol (CH30H)on metallic copper. This role of ZnO prevents the re-formation of methanol by the combination of the adsorbed methoxy species with its dissociated H(*)partner. Thus, apart from the role of CuzO species mentioned previously, this process provides an additional route for the 0-H bond cleavage during the methanol-steam reforming reaction. On the basis of chemisorptiorddesorption studies as well as on the basis of reaction experiments, Su and Rei (1991) have shown that the 0-H bond cleavage is the rate-controlling step in the steam reforming of
1552 Ind. Eng. Chem. Res., Vol. 34,No. 5, 1995 100
80
60
W 250oC(i) H 200oC(i) 170 OC(i) H 250 OC(ii) c] 200 OC(ii) H 170 OC (ii)
40
20
0
p-5c
p-5CM2
p-5cc1
p-5CZ2
catalyst Figure 2. Methanol conversion of nonpromoted catalysts and catalysts containing the optimum promoter concentration as a function of activation method and reaction temperature. i, catalyst activated by reduction in Hz followed by treatment in vaporized methanol-water mixture; ii, vaporized methanol-water mixture activated catalysts.
methanol. Their result is consistent with the report of Krylov (1970) that, generally, the 0-H bond cleavage is the rate controlling step in the decomposition of alcohols. Therefore, it would appear that the enhancement of the rate-controlling step with the additional role played by surface 02-anions present in ZnO is responsible for the increased methanol conversions observed in Zn-promoted catalysts. Specific Effects of Promoter Type. Figure 2 shows the maximum conversions obtained from the optimum Mn-, Cr-, and Zn-promoted catalysts as well as from the nonpromoted catalysts for all three reaction temperatures and two activation methods used in this work. It is seen that the optimum Mn-promoted catalysts yielded maximum methanol conversions of 94 and 99 mol % in the methanol-steam reforming reaction at 250 "C when activated by initial reduction in a hydrogen atmosphere followed by treatment in vaporized feed and, simply, by treatment in the vaporized feed, respectively. Corresponding activation treatments used for the optimum Cr-promoted catalyst yielded maximum methanol conversions of 90 and 95 mol %, while those for the optimum Zn-promoted catalyst were 92 and 96 mol %, respectively. These results indicate that the net effects of promoter incorporation into the coprecipitated Cu-Al catalysts on methanol conversion were different for Mn, Cr, and Zn. As was discussed earlier, the promotion effects of Mn and Cr incorporation are due to their ability to keep large amounts of copper in its positive oxidation states. Also, Rase (1977),Krylov (1970),and Trimm (1980)have reported that oxides or other compounds of transition metals in which the metal ions exist in an electron configuration of dl to d4 or d6 to d9 are active for hydrogenatioddehydrogenation reactions as well as for oxidationheduction reactions. According to these authors, Cu2Cr204 is expected to show a greater promotion effect than CuMnO2 on account of the CrIII ion existing
Table 3. Product Distribution of Nonpromoted Catalysts and Catalysts Containing the Optimum Promoter Concentrationa catalyst catalyst reaction product distribution, mol % promoter identity temp, "C H2 C02 CO HCOOCH3 none p-5C 250 73.9 24.6 1.5 0 Mn
p-5CM2
Cr
p-5CC1
Zn
p-5CZ2
a
200 170 250 200 170 250 200 170 250 200 170
74.9 75 73.6 74.8 75 74.5 74.9 75 74.1 74.8 75
24.9 25 21.5 24.4 25 23.4 24.8 25 22.4 24.8 25
0.1 0 4.9 0.8 0 2.1 0.3 0 3.5 0.4 0
0 0 0 0 0 0 0 0 0 0 0
Vaporized methanol-water mixture activated catalysts.
in the d3 electron configuration compared with that of d4 exhibited by the MnlI1ions in the latter species. On the other hand, Klisurski (1970), Kobayashi et al. (1976), and Matsukata et al. (1988) have shown that, between the oxides of Mn and Cr, Mn2O3 is the more active in the methanol-steam reforming reaction. The presence of two types of promoter species (i.e., Mn2O3 and CuMnO2 and Cr203 and Cu2Cr204 for Mn and Cr, respectively) implies that the overall result of Mn or Cr incorporation into the catalyst is the combined effects of their behavior both as promoter and as catalyst. The methanol conversion results (Table 3) show that these composite effects produced a more favorable net result for the optimum Mn-promoted catalysts for both methods of activation. On the other hand, ZnII ions present in the Znpromoted catalysts exist in the d10 electron configuration which is not known to possess high activity for hydro-
Ind. Eng. Chem. Res., Vol. 34, No. 5, 1995 1553
I
-C-
b
--C
3 4 0
V
0 .I Y
250 OC(ii) 200oC (ii) 170 OC(ii) 25OoC (i) 200 oc (i) 170oC(i) .,
2
0.30
5 *
0.22
. I
0
h
0, a
E 1
2
0
- 0.
10
-G
'
P
H
ao
1 0
0.20
Mn concentration, wt % 0.15
! 0
"
-
I
0
I
I
.
,
.
10 20 30 Cr concentration, wt %
200 OC(ii) 170 OC(ii) 250 OC(i) 200 o c (i) 170oC(I)
I
1
-
80'
-f-
.
--C
.-1 6 0 1 I
200 OC(ii) 170 OC (ii) 250 OC (i) 200 OC(i)
5
3u
40-
". 0
10
20
1
I
,
I
10
15
20
25
30
35
Promoter Concentration, wt % Figure 4. Variation of Cul/Cuo weight ratio with promoter concentration. a, Mn-promoted catalysts; b, Cr-promoted catalysts.
40
c 100,
-I
,
5
30
Zn concentration, wt %
Figure 3. Variation of methanol conversion with promoter concentration in the methanol-steam reforming reaction. a, Mnpromoted catalysts; b, Cr-promoted catalysts; c, Zn-promoted catalysts; i, catalyst precursor containing 24.1 wt % Cu; ii, catalyst precursor containing 27.8 wt % Cu.
genatioddehydrogenation and oxidatiodreduction reactions (Rase, 1977). However, as was discussed previously, the effect of ZnO in Zn-promoted catalysts appears t o be due to the stabilizing effect of its surface 02-anions on the adsorbed methoxy species with the net result of facilitating the 0-H bond cleavage during the dissociative adsorption of methanol on metallic copper. Also, as in the case of the Cr-promoted catalyst, the maximum methanol conversion obtained from the optimum Zn-promoted catalyst was not as high as that of the corresponding Mn-promoted catalyst, especially a t the reaction temperature of 200 "C. Effects of Promoter Concentration. Figures 3a-c shows the relationships between methanol conversion and promoter concentration for the Mn-, Cr-, and Znpromoted catalysts. It is seen from Figure 3 that a
maximum exists for all three promoters. In the case of Mn-promoted catalysts, the maximum occurred at ca. 7.25 wt % Mn whereas Cr- and Zn-promoted catalysts had maximum methanol conversions at ca. 3.56 and 13.4 wt % Cr and Zn, respectively. The existence of a maximum in the relationship between methanol conversion and the respective promoter concentration in the catalyst is explained below. As was mentioned earlier, studies on in situ characterization of activated catalysts showed the presence of various CUI species. The relationship between Cul/Cuo weight ratio and promoter concentration reported earlier by Idem and Bakhshi (1994~)for Mn- and Crpromoted catalysts is shown in parts a and b of Figure 4, respectively. As was discussed earlier in the case of Mn- and Cr-promoted catalysts, the Cul/Cuo weight ratio represents the electron-accepting potential of the activated catalyst. A comparison of Figure 4 with Figure 3 which gives the relationships between methanol conversion and the corresponding promoter concentration shows that methanol conversion increases as the Cul/Cuoweight ratio (electron-accepting ability) in the activated catalyst increases. This shows that, in order to maximize methanol conversion, the amount of CUI species in the catalyst must be maximized. However, as was observed in Figures 3 and 4,there is a limit to the amount of each promoter that should be incorporated into the catalyst. As is seen in Figure 4, beyond these concentration limits, the Cul/Cuo weight ratio decreases. In the case of Mn and Cr promoters, an increase in Mn and Cr promoter concentrations will definitely result in an increase in the respective amounts of CuMnOz and CuzCr204 that will be present in the catalyst. However, it also means a decrease in the overall copper concentration in the catalyst. Consequently, this will result in a decrease in the amounts of both Cuo and Cu2O in the activated catalyst (Idem and Bakhshi, 1994~).The overall effect is the existence of a maximum ratio. This therefore implies that, beyond the concentration limit for each promoter, its ligand effect starts decreasing while its role as an active site for methanol reforming becomes significant as was seen by the increase in the total amounts of MnzO3 and Cr203 species with the respective promoter concentration (Idem and Bakhshi, 1994~).In comparison with copper species, Klisurski (19701, Kobayashi et al. (19761, and Matsukata et al.
1664 Ind. Eng. Chem. Res., Vol. 34, No. 5, 1995
(1988) have shown that these metal oxide species are less active if used solely as catalysts. This superior activity of copper species over Mn and Cr species can also be seen in Figure 3a,b. This figure shows that methanol conversions from catalysts prepared by promoter impregnation into the coprecipitated Cu-Al catalyst precursor containing 27.8 wt % copper were larger than those from corresponding promoted catalysts prepared by promoter impregnation into the precursor containing 24.1% copper for all promoter concentrations. In the case of Zn-promoted catalysts, it is expected that a large amount of ZnO will ensure a large extent of its contribution toward the O-H bond cleavage. However, this will mean a reduction in the amount of copper species in the catalyst and, consequently, a reduction in the catalyst roles that require catalyst redox ability and adsorbate C-H bond scission. Thus, beyond a certain Zn concentration, the increase in the potential for the stabilization of adsorbed CH30(a,species on metallic copper does not compensate for the decrease in the amount of this CH30(*)species produced from Cu20. Thus, as in the case of Mn- and Crpromoted catalysts, beyond the optimum Zn concentration, the promotion effects start decreasing while its role as an active site for methanol reforming becomes significant. Again, in comparison with copper species, ZnO species are less active if used solely as a catalyst (Waugh, 1992; Chan and Griffin, 1986). This can also be seen in Figure 3c which shows that methanol conversions from catalysts prepared by impregnation of Zn into the dried Cu-Al catalyst precursor containing 27.8 wt % copper were larger than those from catalysts prepared by impregnation into the precursor containing 24.1 wt % copper. Effects of Activation Method. Figure 2 shows the maximum methanol conversions obtained from the optimum Mn, Cr, and Zn-promoted catalysts for the two types of activation techniques used in the study (i.e., either by reduction of catalysts in a Hz atmosphere followed by treatment with the vaporized methanolwater mixture or simply by treatment of the catalyst with the vaporized methanol-water mixture). It is seen from the figure that methanol conversions obtained from vaporized methanol-water feed-activated catalysts are higher than those from catalysts initially reduced in a H2 atmosphere. This can be explained on the basis of the results of our earlier characterization studies which showed that the Cul/Cuoweight ratios in the vaporized methanolwater feed-activated catalysts were larger than those in corresponding catalysts activated by initial reduction in a H2 atmosphere. As was discussed previously, the Cul/Cuo weight ratio in activated catalysts represents the extent of methanol conversion that depends on catalyst redox ability. A large Cul/Cuo weight ratio in the activated catalyst means a strong redox ability of the catalyst and, consequently, a large contribution from this role of the catalyst to the overall methanol conversion (Chan and Griffin, 1986; Bowker and Madix, 1980; Wacks and Madix, 1978; Klisurski, 1970; Idem and Bakhshi, 1994b). Effects of Reaction Temperature. Table 2 also shows the variation of methanol conversion as a function of reaction temperature for the various optimum promoted catalysts activated by both methods of activation used in the study. The methanol conversion was observed to increase with reaction temperature. This
behavior is typical of the relationship between methanol conversion and reaction temperature and is consistent with the well-known effects of temperature on reaction rates. However, it is seen from Table 2 that the difference between methanol conversions obtained at low reaction temperatures (170 and 200 "C) are higher than those obtained at higher temperatures (200 and 250 "C), indicating a thermodynamic limitation in some of the roles of the catalyst in the methanol-steam reforming reaction at low reaction temperatures. This can be explained as follows. So far, our studies have shown that the activity of a copper-based catalyst in the methanol-steam reforming reaction is a composite of three roles. These are metallic copper area and dispersion (principally for C-H bond scission), redox ability (involved mainly in the O-H bond cleavage), and promoters (responsible for the ligand effects). While the effects of temperature on the role of copper area and dispersion are mainly kinetic, the effects on redox ability which is involved in the ratecontrolling step appear t o be thermodynamically controlled. This is because a minimum temperature exists below which catalyst reduction and reoxidation may not take place. The methanol conversion results show that, below the reaction temperature of 200 "C, the catalytic role involving the direct participation of the catalyst in the redox cycle during reaction is thermodynamically inhibited. As was discussed previously, methanol O-H bond cleavage is the rate-controlling step. It therefore means that any reaction condition (such as low reaction temperature) that is detrimental to this rate-controlling step will result in unusually low methanol conversions such as were observed in our studies. Product Distribution. The major products obtained from the methanol-steam reforming reaction using both promoted and nonpromoted catalysts were H2, CO, and C02. Methyl formate (HCOOCH3) was obtained only in trace quantities and a t low reaction temperatures. Generally, the reaction scheme given in eqs 2-5 can be used to account for these products. However, Takahashi et al. (19821, Su and Rei (19911,Jiang et al. (19931, and Minachev et al. (1989) have shown that the CO shift reaction may not be involved in the steam reforming of methanol. According to these authors, the methyl formate (HCOOCH3) intermediate obtained in eq 3 undergoes hydrolysis to produce a formic acid intermediate and methanol as given in eq 6. This formic acid intermediate then decomposes t o yield carbon dioxide and hydrogen (eq 7) without the involvement of the CO shift reaction. HCOOCH,
+ H 2 0 = HCOOH + CH,OH
HCOOH = CO,
+ H,
(6) (7)
However, the presence of CO in the product distribution given in Table 3 implies that either the decomposition of HCOOCH3 according to eq 4 is also involved in the methanol-steam reforming reaction over our promoted and nonpromoted catalysts or CO is produced from the reaction between C02 and metallic copper (eq 8) during the Cuo reoxidation process or both. 2cu
+ co, = c u 2 0 + co
(8)
The presence of Cu20 species in the activated catalysts (Idem and Bakhshi, 1994b-c) confirms the formation of CO by eq 8. The formation of CO by eq 4 was not verified in the present work. Since CO is the
Ind. Eng. Chem. Res., Vol. 34,No. 5, 1995 1666 Table 4. Hydrogen Production Efficiencies of Nonpromoted Catalysts and Catalysts Containing Optimum Promoter Concentrationsas a Function of Reaction Temperature and Catalyst Activation Method
Hz production efficiency, mol % reaction Hz selectivity, reaction efficiency, mol % mol % promoter identity temp, "C RE(i)a RE(iiib S(iC S(ii,b catalyst
none
p-5C
Mn
p-5CM2
Cr
p-5CC1
Zn
p-5CZ2
250 200 170 250 200 170 250 200 170 250 200 170
78 59 19 88 80 27 88 64 20 88 78 22
83 68 21 93 92 32 91 87 26 90 85 23
94.3 99.6 100 93.5 98.4 100 97.6 100 100 95.5 98.2 100
94.1 99.7 100 93.7 98.9 100 96.1 99.5 100 93.8 99.1 100
i: catalysts activated by reduction in Hz followed by treatment in methanol-water mixture. ii: vaporized methanol-water mixture activated catalysts.
undesired product, it was important to study how its concentration in the product could be minimized. Therefore, it was necessary to determine how the formation of this CO product would be affected by various operating variables. This is discussed below. Effects of Promoter Type. Table 3 shows that the CO concentration in the product stream was low ( < 5 mol %) for all catalysts. The literature (Rase, 1977; Ghiotti and Boccuzzi, 1987; Amphlett et al., 1988) indicates that the CO shift reaction may be enhanced over Zn-promoted catalysts. However, it was observed that the concentrations of CO from the methanol-steam reforming reaction over Mn-, Cr-, and Zn-promoted catalysts were all similar to the ones over the nonpromoted coprecipitated Cu-Al catalysts. This shows that the CO shift reaction was not enhanced on any of the promoted catalysts. Effects of Reaction Temperature. Table 3 shows that CO concentration in the product increased with reaction temperature. No measurable amount of CO was observed at the reaction temperature of 170 "C. This may be attributed to the typical behavior of products from the exothermic CO shift reaction, if it takes place at all. However, if the CO shift reaction was not involved in the methanol-steam reforming reaction as indicated earlier, the result shows more explicitly that either reaction 4 or the copper redox cycle or both did not proceed to an appreciable extent at low reaction temperatures over all types of catalysts. It is seen from Table 4 that an extremely high H2 selectivity (S > 93 mol %) was obtained for both promoted and nonpromoted catalysts in the methanolsteam reforming reactions at the three reaction temperatures used in this study. These results indicate that, within this temperature range and using the catalysts under consideration in this study, maximum HZproduction is not affected significantly by H2 selectivity. CO Shift Reaction Equilibrium. As was discussed previously, our results indicate that the presence of CO in the product stream is due to reaction 8. It may also be attributed to reaction 4. Nevertheless, it should be possible to eliminate this CO product if the CO shift reaction (eq 5) can be driven t o completion. However, there seems to be an unfavorable equilibrium for CO consumption. This is because of the presence of high
concentrations of CO2 and H2 (about 24 and 74 mol %, respectively) as compared to the extremely small concentration of CO (about 2 mol %) in the catalyst bed, thereby shifting the equilibrium of reaction 5 to the left. The concentrations of the various products can be seen in the product distribution given in Table 3.Jiang et al. (1993)have also shown that it is unlikely for the CO shift reaction to take place during the methanol-steam reforming reaction. These authors attribute this to active site blockage because methanol and other reaction intermediates adsorb more strongly on copper than CO. Thus, CO formed by whatever process will be incapable of participating in any shift reaction. The unpreventable production of CO from reaction 8 coupled with the unfavorable shift in equilibrium of reaction 5 as well as the occurrence of active site blockage for CO implies that it is not possible to completely get rid of CO in the product stream during the methanol-steam reforming reaction within the same reactor. For example, Amphlett et al. (1993)have reported the presence of CO in the dry product stream for a wide range of reactor conditions (including reaction temperatures as low as 150 "C) using various types of copper-based catalysts. Hydrogen Production Efficiency. If we wish to use methanol as H2 storage, it will be desirable to know the maximum amount of H2 that can be made by a catalyst under various operating conditions. This type of work called "hydrogen production efficiency" has recently been reported in the literature (Idem and Bakhshi, 1994b)for coprecipitated Cu-Al catalysts to study the effects of operating conditions such as type of feed, reaction temperature, catalyst copper concentration, and catalyst calcination temperature. In this work, we have studied the effects of catalyst activation method on the hydrogen production efficiency of promoted coprecipitated Cu-AI catalysts. Hydrogen production efficiency was evaluated in terms of H2 selectivity ( S ) to account for the quality of H2 production. In addition, hydrogen production efficiency was evaluated in terms of reaction efficiency (RE) in order to account for the quantity of H2 produced and to take catalyst activity into consideration. Typical H2 production efficiencies are given in Table 4 for vaporized feed-activated promoted catalysts as well as for corresponding promoted catalysts activated by initial reduction in a H2 atmosphere. Table 4 shows that H2 selectivities (8) for the vaporized feed-activated catalysts were similar to those from corresponding catalysts activated by reduction in a H2 atmosphere followed by mild treatment in a vaporized methanolwater mixture. In contrast, it is seen from Table 4 that reaction efficiencies (RE) from the vaporized feed catalysts were superior to those from corresponding catalysts activated by initial H2 reduction. Superior reaction efficiency in the former catalysts was attributed to greater catalyst activity. This is discussed below. As was shown earlier, the amount of CuzO was larger in the vaporized feed-activated catalysts than in corresponding catalysts activated by initial H2 reduction. This resulted in a larger Cul/Cuoratio in the catalysts activated using the former mode of activation. This increase in the Cul/Cuo ratio is responsible for the increase in methanol conversion and the consequent increase in reaction efficiency in the vaporized feedactivated catalysts. These results for promoted catalysts are consistent with those of Idem and Bakhshi (1994b)for nonpromoted catalysts.
1556 Ind. Eng. Chem. Res., Vol. 34,No. 5, 1995 Table 5. Activity Comparison of Catalysts Used by Various Workers for the Methanol-Steam Reforming Reaction W/Fm, g of catalyst reaction temp, methanol conversion, hydrogen selectivity, reference catalyst &mol of CH30H "C mol % mol % present study p-5CM2 115.6 250 99 93 99 200 93 83 94 Idem and Bakhshi (1994b) p-5C 115.6 250 Amphlett et al. (1991) C18HC (ICI) catalyst 146.2 250 80'" 1006 Su and Rei (1991) CuO/ZnO 51.84 280 58 100 NA Matsukata et al. (1988) cd&o3 120 250 65 a
Calculated from approximate rate equation given in reference. Assumed; all reactions performed at atmospheric pressure.
Catalyst Activity and Stability Studies. Activity and stability versus time on stream studies were conducted on the best catalyst (i.e., p-5CM2) at three reaction temperatures (170, 200, and 250 "C). The methanol-steam reforming reaction at 250 "C was performed continuously for 8 h each day over the catalyst. This procedure was repeated for a total of 5 days. The entire operation was carried out for each reaction temperature. Results obtained showed that, at each temperature, the catalyst activity remained stable throughout the entire 5-day period, and no decrease in methanol conversion appropriate to that temperature was observed. After these studies, the catalyst used a t each temperature was characterized using the in situ TPR characterization method. The TPR spectra were similar for the used and fresh activated catalysts. Activity Comparison. It is known in catalysis literature that various factors such as catalyst constituents and preparation variables affect catalyst performance. A comparison of different catalysts used by different workers for H2 production from methanol has recently been reported in the literature for both methanol decomposition and methanol-steam reforming reactions. With the improvement in methanol conversions we achieved for promoted catalysts in the methanolsteam reforming reaction, we decided to include the best catalyst from the present study in the activity comparison. This updated activity comparison involving the best catalyst in this study (p-5CM2) and the various catalysts reported in the literature is given in Table 5 for the steam reforming reaction. Although different reaction conditions have been used by different workers, Table 5 shows that the best catalyst from the present study (p-5CM2) provides the best performance. Conclusions 1. Promotion of coprecipitated Cu-Al catalysts with Mn, Cr, and Zn results in the introduction of CuMnO2, Cu2Cr204, and ZnO species, respectively, into these catalysts and their consequent enhancement of methanol conversion in the methanol-steam reforming reaction. 2. While the incorporation of Mn and Cr in the respective metal-promoted catalysts enhanced catalyst activity by maintaining optimum amounts of Cuo and CUI species, that of Zn functioned by the behavior of surface 02-anions in ZnO as a Bronsted base. 3. Methanol conversions obtained from catalysts activated by treatment in vaporized methanol-water feed were higher than those from corresponding catalysts activated by initial reduction in a H2 atmosphere. 4. Maximum methanol conversions of 93 and 99 mol % were obtained in the Mn-promoted, vaporized methanol- water feed-activated, coprecipitated Cu-AI catalyst at reaction temperatures of 200 and 250 "C, respectively.
5. At low reaction temperatures, there is a thermodynamic constraint in the rate-controlling step of the methanol-steam reforming reaction, resulting in unusually low methanol conversions at this low end of reaction temperature.
Acknowledgment The financial support provided by the Canadian Commonwealth Scholarship and Fellowship Plan and the Natural Sciences and Engineering Research Council of Canada is gratefully acknowledged.
Nomenclature
D,: catalyst particle diameter (m) F: molar flow rate (mol/min) J: type of promoter L: height of catalyst bed (m) 0: promoted catalyst containing optimum promoter concentration RE: methanol-steam reforming reaction efficiency (mol %)
S: hydrogen selectivity (mol %) W weight of catalyst (g)
Subscripts (a): adsorbed species (i): catalyst activated by reduction in hydrogen at 300 "C followed by treatment with vaporized methanol-water mixture (ii): vaporized methanol-water activated catalyst (m): methanol
Literature Cited Agaras, H.; Carrella, G.;Larborde, M. A. Copper catalysts for the steam reforming of methanol. Analysis of preparation variables. Appl. Catal. 1988,45, 53-60. Amphlett, J. C.; Mann, R. F.; McKnight, C. Steam reforming of methanol on CuO/MgO catalysts. In 2nd International Symposium on Hydrogen Produced fiom Renewable Energy; Handcock, 0 . G., Sheinkopf, K. G., Eds.; Florida Solar Energy Research Center: Cocoa Beach, FL, 1985; pp 91-100. Amphlett, J. C.; Mann, R. F.; Weir, R. D. Hydrogen production by the catalytic steam reforming of methanol. Part 3: Kinetics of methanol decomposition using C18HC catalyst. Can. J. Chem. Eng. 1988,66,950-956. Amphlett, J. C.; Mann, R. F.; Peppley, B. A.; Stokes, D. M. Preparation and characterization of three methanol steam reforming catalysts based on copper-magnesia with additions of manganese, chromium and vanadium oxides. In Proceedings of the 11th Canadian Symposium on Catalysis, Halifax, NS, Canada, July 15-18, 1990; Monnier, J., Ed.; "he Catalysis Division of the Canadian Societies for Chemistry and Chemical Engineering: 392-401. Amphlett, J . C.; Mann, R. F.; Peppley, B. A.; Stokes, D. M. Some design considerations for a catalytic methanol-steam reformer for a PEM fuel cell power generating system. In Proc. Zntersoc. Energy Convers. Eng. Conf. 1991,26 (3), 642-649. Amphlett, J . C.; Mann, R. F.; Peppley, B. A. "he steam reforming of methanol: mechanisms and kinetics compared to the
Ind. Eng. Chem. Res., Vol. 34,No.5, 1995 1667 methanol synthesis process. Presented a t the 3rd International Symposium on Natural Gas Conversion, Sydney, Australia, 1993. Biswas, J.; Bickle, G. M.; Gray, P. G.; Do, D. D.; Barbier, J. The role of deposited poisons and crystallite surface structure in the activity and selectivity of reforming catalysts. Catal. Rev.-Sci. Eng. 1988, 30 (2), 161-247. Bowker, M.; Madix, R. J. XPS,U P S and thermal desorption studies of alcohol adsorption on Cu(ll0). I. Methanol. Surf. Sci. 1980, 95, 190-206. Chan, L.; Griffin, G. L. Methanol decomposition on CdZnO oriented thin films. Surf. Sci. 1986,173, 160-175. Ghiotti, G.; Boccuzzi, F. Chemical and physical properties of copper based catalysts for CO shifi reaction and methanol synthesis. Catal. Rev.-Sci. Eng. 1987,29 (2 & 3), 151-182. Idem, R. 0. Production of hydrogen from the low-temperature steam reforming of methanol. Ph.D. Thesis, University of Saskatchewan, Saskatoon, Canada, 1995, pp 58-61,146-187. Idem, R. 0.;Bakhshi, N. N. Production of hydrogen from methanol. 1. Catalyst characterization studies. Znd. Eng. Chem. Res. 1994a, 33,2047-2055. Idem, R. 0.;Bakhshi, N. N. Production of hydrogen from methanol. 2. Experimental studies. Z n d . Eng. Chem. Res. 1994b,33,20562065. Idem, R. 0.;Bakhshi, N. N. Characterization studies of calcined, promoted and non-promoted methanol-steam reforming catalysts. Submitted for publication in Can. J. Chem. Eng., 1994c. Jiang, C. J.; Trimm, D. L.; Wainwright, M. S.; Cant, N. W. Kinetic study of steam reforming of methanol over copper-based catalysts. Appl. Catal. 1993, 93, 245-255. Klisurski, D. Relationship between the selectivity and activity of oxide catalysts in the oxidation of methanol and the initial temperature oftheir reduction. Kinet. Katal. 1 9 7 0 , l l (11,263264. Kobayashi, H.; Takezawa, N.; Minochi, C. Methanol reforming reaction over copper containing mixed oxides. Chem. Lett. 1976, 1347-1350.
Krylov, 0. V. Catalysis by nonmetals: Rules for catalyst selection; Academic Press: New York, 1970; pp 47-69, 115-139. Matsukata, M.; Uemiya, S.; Kikuchi, E. Copper-alumina spinel catalysts for steam reforming of methanol. Chem. Lett. 1988, 761-764. Minachev, K. M.; Kotyaev, K. P.; Lin, G. I.; Rozovskii, A. Y. Temperature-programmed surface reactions of methanol on commercial Cu-containing catalysts. Catal. Lett. 1989,3,299308. Rase, H. F. Chemical reactor design forprocessplants;Wiley Publications: New York, 1977; Vol. 1, pp 60-110. Su, T.; Rei, M. Steam reforming reactions over nickel and copper catalysts. J. Chin. Chem. SOC.(Taipei) 1991, 38, 535-541. Takahashi, K.; Takezawa, N.; Kobayashi, H. The mechanism of steam reforming of methanol over a copper-silica catalyst. Appl. Catal. 1982,2, 363-366. Takezawa, N.; Kobayashi, H.; Hirose, A.; Shimokawabe, K. Steam reforming of methanol on copper-silica catalysts: effects of copper loading and calcination temperature on the reaction. Appl. Catal. 1982,4, 127. Trimm, D. L. Design of industrial catalysts; Elsevier Scientific Publishing Co.: New York, 1980; pp 53-90. Wacks, I. E.; Madix, R. J. The selective oxidation of CH30H to HzCO on a copper (110) catalyst. Mechanism. J. Catal. 1978, 53, 208. Waugh, K. C. Methanol synthesis. Catal. Today 1992,15,51-75.
Received for review September 26, 1994 Accepted February 10,1995 * IE940563H
* Abstract published in Advance ACS Abstracts, April 1, 1995.
