J. CHEM. SOC. FARADAY TRANS., 1994, 90(8), 1183-1189 Oxidation of Carbon Monoxide on LaMn, -xCuxO, Perovskite-type Mixed Oxides Hiroyuki Yasuda, Yoshiko FujiwaraJ Noritaka MizunoS* and Makoto Misono* Department of Synthetic Chemistry, Faculty of Engineering, The University of Tokyo, Hongo, Bunkyo-ku, Tokyo 113,Japan Catalytic oxidation of carbon monoxide has been investigated over a series of LaMn,-,Cu,O, (x = 0-0.5) and La,CuO, catalysts having perovskite-type and perovskite-related structures. LaMn, -,Cu,O, catalysts were pre- pared by a freeze-drying method and showed uniform compositions. The oxidation states of Cu and Mn of the perovskite catalysts as well as several other properties relating to the reactivity of oxygen, such as temperature- programmed desorption of oxygen and carbon monoxide, the reducibility of the catalysts and the adsorption of carbon monoxide and dioxide were measured.La, -,Sr,MnO, (y = 0.2-1 .O) catalysts were also studied for com- parison. A remarkable synergistic effect on the catalytic activities for the oxidation was found with Mn and Cu. The synergistic effect and the deactivation process were discussed based on the above properties. The effect was attributed to the combination of two functions, namely the activation of oxygen by Mn oxide and that of carbon monoxide by Cu ions, and the deactivation towards adsorption of carbon dioxide. In the perovskite structure, ABO, , where the total charge on the cations is +6e and the ionic radius of the B-site cation is much smaller than those of the A-site cation and oxide ion, it is possible to control the valency of B-site ions and the amount of oxygen vacancies by changing the A- or B-site ions or by partial substitution with ions of different valency without affecting the fundamental structure.'-, The enhance- ment of the catalytic activity by A-site substitution has been studied extensively, but little is known on the effect of B-site element sub~titution.~-~ Zhang et a/., reported that the activ- ity of La,~4Sr,~,Mn,&o,~203 for the oxidation of butane was about five times higher than that of LaMnO, .However, we found that substitution of Co by Mn in LaCoO, led to only slight enhancement of the catalytic activity for the oxi- dation of propane when the catalyst composition was made uniform by freeze-drying.' Carbon monoxide (CO) oxidation catalysts have been used to clean up industrial, automotive and domestic emis-sions.'0-'2 For example, La,,,Ce,,,Co0,, a perovskite-type mixed oxide, is commercially used to oxidize CO and organic compounds in the emission from ovens by utilizing its high thermal stability, low cost and high catalytic activity.', We have also reported that La,~,Sr,,,CoO, shows higher cata- lytic activity for the oxidation of CH, and C,H, than Pt/Al,O,. 14,15 It has been reported that LaMn,~,Cu,,,O, shows a high catalytic activity for the oxidation of CO.', We have also steady-state activities, and discussed possible reasons for the pronounced synergistic effect and the deactivation based on the observed properties of catalysts.Experimental Catalysts LaMn, -,Cu,03 (x = 0-0.5) catalysts were prepared by freeze-drying solutions containing the component metal ace- tates.' LaMn,~,Cu,~,O, was also prepared from the metal nitrates in the same manner. La,CuO, and La,-,Sr,MO, (M = Co, Mn, y = 0.2-1.0) catalysts were prepared by evapo- rating the mixed acetates solutions, as described previously. " LaFe,,,Cu,~,O, and LaCo,~,Cu,~,O, catalysts were pre-pared via a citrate proce~s.'~ Each precipitate or precursor obtained was first decomposed in air at 573 K for 3 h and then calcined in air at 1123 K for 2-5 h. In this paper they are abbreviated by La,~,Sr,M,~,Cu,O, (M = Mn, Co, Fe), although the actual compositions may generally be non-stoichiometric with respect to oxygen.5% Pt/Al,O, and 0.7% Pt/A1,0, promoted by CeO, were commercially obtained. The crystal structure of the prepared catalysts was deter- mined by powder X-ray diffraction (XRD) (Rigaku Denki, Rotaflex, RU-200) using Cu-Kcr radiation. The specific surface areas were measured by means of the BET method using N, adsorption at 77 K after the pretreatment of the reported that the initial catalytic activity of L~M~,~,CU,~~O, samples in an He stream at 573 K. for the oxidation of CO was very high owing to a remarkable synergistic effect of Mn and Cu (ca. 102-103 times), although the activity declined rapidly with reaction time.' However, little is known of the mechanism of the synergistic effect and the catalyst deactivation.Hence, the elucidation of the syn- ergistic effect and the inhibiting effect of CO, may be of inter- est from the viewpoint of catalytic chemistry. In the present paper, we have extended our previous work," by measuring the steady-state catalytic activity and attempting to correlate the activity with the properties related to the adsorption and the reactivity of oxygen. As described below, we confirmed the synergistic effect for the Present address: Toshiba Co., Shin-isogo-cho, Isogo-ku, Yoko-hama 235, Japan.1Present address : Catalysis Research Center, Hokkaido Uni-versity, Sapporo 060, Japan. XPS Measurements Self-supporting discs (about 100 mg , 1 cm in diameter) were used, and the spectra were recorded with a JEOL JPS-90SX spectrometer using an Mg-Ka source (1253.6 eV).The pres- sure in the chamber was kept in the range 10-8-10-9 Torr. The binding energies were corrected by using the value of 285.0 eV for the C 1s peak resulting from carbon contami- nation. The surface atomic ratios of the samples were calcu- lated based on the equation described in the previous paper,,' using the integrated intensities of the La 3d,,,, Mn 2p3,, ,Cu 2p3,, and 0 1s photoelectron lines. Approximately the same La : Mn : Cu ratio was obtained when the peak intensities of the La 3d,/,, Mn 2p1,, and Cu 2p,/, lines were used. The oxidation states of copper on the surface of LaMn, -,Cu,03 catalysts were estimated from the ZsaJZmain ratios (Imsinand Isatare the peak intensities of the main and satellite signals of Cu 2p3!,, respectively) and the Cu L,VV Auger peak, as in the previous paper.,' Reduction of Catalysts by CO The reduction of the catalyst by CO was conducted by a pulse method at 573 K as in the previous study." The cata-lysts (25-50 mg) were heated in an 0, stream for 1 h at 673 K and then cooled to 573 K in 0, prior to the reaction.The flow rate of carrier gas (He) was 30 cm3 min- and the size of each pulse was 0.1 cm3. Products were analysed by gas chro- matography using a silica gel column. Adsorption of CO The adsorption of CO was measured volumetrically in a closed recirculation system. The catalysts (0.3-0.5 g) were evacuated at 773 K for 2 h and exposed to CO at 298 K.The equilibrium pressure was ca. 110 Torr. The amount of CO uptake was determined by the pressure decrease measured with a Baratron pressure gauge, giving the total amount of CO adsorbed. After evacuation of the sample for 1 h at 298 K, the amount of uptake was measured again. This was used to determine the amount of reversible adsorption. The amount of irreversible adsorption of CO was the difference between the total amount of CO adsorbed and the amount of reversible adsorption. Temperature-programmed Desorption (TPD) of 0,,CO, and CO TPD of O,, CO, and CO was carried out with a flow system using He as a carrier gas, as in the previous study.18 The oxygen impurity in He was removed by a molecular sieve 5A trap kept at 77 K.Prior to each run, the sample (ca. 0.5 g) was pretreated in an 0, stream (30 cm3 min-') at 1123 K for 1 h and was cooled to 298 K in 0,. In the case of TPD of CO, or CO, the sample was further pretreated in pure CO, or CO streams (30 cm3 min-') at 298 K for 30 min after the pretreatment with 0,. The temperature of the sample was raised from 298 to 1123 K at a constant rate of 20 K min-' in an He stream (30 cm3 min- '), and the gases desorbed were detected by use of a quadrupole mass spectrometer (NEVA, NAG-531). The rates and amounts of O,, CO, and CO desorbed were calculated from the concentration of the eluent gas. The reproducibility of the TPD curves was confirmed by repeating the TPD run after the same pretreatments.Catalytic Oxidation of CO The catalytic oxidation of CO was carried out at 473-873 K in a fixed-bed flow reactor at atmospheric pressure by feeding a gas mixture of CO (1.3%), 0, (1.3%) and N, (balance) at a flow rate of ca. 200 cm3 min- ' over a mixture of 3-20 mg of catalyst and 200 mg of Sic (GHSV = 1.8 x lo6 h-' for 10 mg of LaMnO~,CuO~,O,) after pretreatment of the catalyst in an 0, stream at 873 K for 1 h. The gas composition was analysed by gas chromatography using molecular sieve 5A and Porapak Q columns. Results Structures and Specific Surface Areas of Catalysts The XRD patterns of LaMn,-,Cu,O, (x = 0-0.5) and LaM,~,Cu,~,O, (M = Co and Fe) are shown in Fig. 1. The crystal structures and the specific surface areas of catalysts prepared in this study are summarized in Table 1. The XRD J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 I I I 1 30 40 50 28/d egrees Fig. 1 XRD powder patterns of LaMn, -,Cu,O, (x = 0-0.5) and LaMo,6Cuo,,0, (M = Co and Fe). (a) LaMnO,; (b) LaMn0.8Cu0.203; (') LaMn0.6Cu0.40~~ (4 LaMn0,5Cu0,50,;LaCoo,6Cuo,,0, and (f)LaFeo.6Cuo,,0,. Bar: lo3 counts s-l. patterns of LaMn, -,Cu,03 (x = 0-0.4), La, -,Sr,MnO, (y = 0.2-0.4) and La,~,Sr,~,CoO, catalysts showed only the perovskite-type structure. The XRD patterns of La,CuO, and SrMnO, showed the orthorhombic K,NiF,-type struc-ture and the four-layer hexagonal SrMnO, structure,21 respectively. Additional phases such as La,CuO, and CuO were observed for LaMn,~,Cu,~,O, and LaM,~,Cu,~,O, (M = Co and Fe) besides the perovskite phase, and a small amount of SrMnO, phase was present for La,~,Sr,~,MnO,.The structures of LaMn, -,Cu,O, catalysts were rhombo- hedral for x = 0-0.2 and cubic for x = 0.3-0.5, while those of La, -,Sr,MnO, (y = 0.2-0.6) were orthorhombic. Table 1 Structures and surface areas of LaM,-,Cu,O, (M = Mn, Co, Fe; x = 0-OS), La,CuO, and La,-,Sr,MO, (M = Mn, Co; y = 0.2-1.0) catalysts surface area catalyst" /m' g-' structure LaMnO, 3.1 P(Wb LaMn0.8Cu0.203 3.8 P(R) LaMn0.7Cu0.303 4.2 P(C)f LaMn0.6Cu0.403 2.3 P(C) LaMn0.6Cu0.403(N)d 0.8 P(C) + unidentified phase (tr.)e 2.8 P(C) + CuO(tr.) + La2Cu0,(0.14) La,CuO, 1.2 K(OIf5.0 P(R) + La,CuO,(tr.)LaCo0.6Cu0.403 LaFe0.6Cu0.403 4.9 P(C) + La,Cu0,(0.25) La0.8Sr0.2Mn03 8.5 P(0) La0.6Sr0.4Mn03 9.6 P(0) La0.4Sr0.6Mn0 3 7.6 P(0) + SrMnO, SrMnO, 1.4 SrMnO,(HB) Lao.8Sro.2CoO 3 2.8 P(C) " Calcined at 1123 K unless noted otherwise. P(R),rhombohedra1 perovskite phase. P(C), cubic perovskite phase. Prepared from the mixed-metal nitrates and calcined at 1373 K. The numbers in par- entheses are the intensity ratios of the impurity phase to the per- ovskite phase; tr., trace. K(O), orthorhombic K2NiF, phase. H, hexagonal. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 The specific surface areas of catalysts calcined at 1123 K varied from 1.2 m2 g-' for La,CuO, to 9.6 m2 g-' for La0.6Sr0.4Mn03 . Surface Properties of LaMn, -xCux03 (x = 0-0.5) and La,CuO, Catalysts The binding energies of the La 3d,,, , Mn 2p3,, ,Cu 2p3,, and 0 1s peaks, together with the IsaJlmainratios of Cu 2p3,,, for the LaMn,-,Cu,O, (x = 0-0.5) and La,CuO, catalysts are shown in Table 2.The XP spectra have intense shake-up satellite peaks in the Cu 2p regions. The IsaJZmainratios of Cu 2p3,, were between 0.50 (x = 0.5) and 0.62 (x = 0.3). These values agreed generally with the value of 0.53 reported for CuO.,, Only one L,VV Auger peak assignable to Cu2+ was observed for LaMn, -,CuXO3 (x = 0.3-0.5) and La,CuO, in the range 917.2-917.8 eV, while the Auger peak could not be detected for LaMn,~,Cuo~,O, probably due to the very weak intensity. The binding energies of Mn 2p3,, peaks for LaMn,-,Cu,O, were in the range 641.8-642.2 eV.The 0 1s spectra of LaMn,-,Cu,O, consisted of two peaks around 529 and 531 eV, as in the literat~re.~~ The former peak is assigned to the lattice oxide ions, while the latter peak has been assigned to hydro~ide,~~,~~ adsorbed water,,, adsorbed ~xygen~~.~~and/or carbonate oxygen.27 The surface compositions of LaMn, -,CuXO3 (x = 0-0.5) and La,CuO, catalysts determined by XPS are summarized in Table 3, together with those of the bulk calculated from the quantities of starting materials. It has previously been con- firmed by elemental analysis that the La : Mn atomic ratio in LaMnO, was 0.99 : The surface compositions of LaMn, -xCu,03 catalysts prepared from the metal acetates agreed fairly well with those of the bulk, except that the copper content on the surface was slightly greater than that in the bulk for x = 0.5.When LaMno&u0.60, was prepared from the metal nitrates, the La : Mn : Cu ratio of the surface was 1.0: 0.62: 0.68, different from that of the bulk Table 2 XPS binding energies (eV) of LaMn,-xCux03 (x = 0-0.5) and La,CuO," 0 641.8 834.9 531.4 529.2 0.2 934.2 0.58 642.1 834.6 531.0 529.5 0.3 934.3 0.62 642.2 834.4 531.0 529.4 0.4 934.2 0.55 642.2 834.1 531.1 529.3 0.5 934.2 0.50 641.8 835.0 531.6 529.3 La,CuO, 934.3 0.52 835.3 531.6 529.3 'The binding energies were corrected by using the value of 285.0 eV for the C 1s peak resulting from carbon contamination. Error limits are k0.3 eV. Intensity ratios of the satellite peak to the main peak of Cu 2p,/, .Table 3 Surface compositions of LaMn, -xCux03 (x = 0-0.5) and La,CuO, surface compositionb bulk composition X0 La Mn Cu 0 La Mn Cu 0 0 0.19 0.15 0.66 0.20 0.20 0.60 0.2 0.17 0.13 0.03 0.67 0.20 0.16 0.04 0.60 0.3 0.16 0.13 0.09 0.63 0.20 0.14 0.06 0.60 0.4 0.17 0.11 0.09 0.63 0.20 0.12 0.08 0.60 0.5 0.18 0.06 0.14 0.62 0.20 0.10 0.10 0.60 La,CuO, 0.23 0.12 0.65 0.20 0.14 0.60 LaMn, -xCu,03 (x = 0-0.5) catalysts were prepared by freeze-drying the mixed acetates solutions. Estimated by XPS peak intensity using the inte- grated areas of the La 3d,/, ,Mn 2p3/,,Cu 2p3,, and 0 1s photoelectron lines. 1185 (La : Mn : Cu = 1.0 :0.60 : 0.40), suggesting that the per-ovskite phase was not formed uniformly in this case.Reduction of LaMn, -xCux03by CO The reduction of catalyst by CO pulses in the absence of oxygen reflects the reducibility of the catalyst, that is, the oxi- dation power of the catalyst. The results of reduction of LaMn,-,Cu,O, (x = 0-0.5) by CO pulses at 573 K are shown in Fig. 2. The product was only CO,. The values in Fig. 2 are the amounts of CO, formed by the first pulse of CO per unit surface area. The reducibility of LaMn, -xCu,03 increased with x, reached a maximum at x = 0.4, and then decreased. The amount of CO, formed by the first pulse over LaMn,~,Cu,~,O, was 2.5 times the surface monolayer of oxygen, which was calculated on the assumption that the concentration of the surface oxide ion is 8.0 x 10l8 ions m-' according to ref.26. Adsorption of CO The amounts of irreversible adsorption of CO at 298 K on the LaMn, -,CuXO3 (x = 0-OS), La,CuO, and La, -,Sr,MnO, (y = 0.2-0.6) catalysts are shown in Fig. 3. As for LaMn, -,Cu,03, the amounts increased with x, reached 4.0 1 r I / I L I I 1 I I0.0 I J 0.0 0.2 0.4 0.6 X Fig. 2 Reduction of LaMn,-xCu,O, (x = 0-0.5) by the first CO pulse at 573 K. CO pulse size, 0.1 cm3. b 0.0 0.2 0.4 0.6 La2Cu04 x, Y Fig. 3 Amount of irreversible adsorption of CO on LaMn, -xCux03, La,CuO, and La,-,Sr,MnO,. (0)LaMn, -,Cu,03 (x = 0-0.5) and La,CuO,; (0)La, -,Sr,MnO, (y = 0.2-0.6); adsorption temperature, 298 K. 1186 a maximum at x = 0.4, and then decreased. The amounts of CO adsorbed varied in parallel with the reducibility of the catalysts by CO.On the other hand, the variation of the amount of CO adsorption was small for La,-,Sr,MnO, and the amount on La,~,Sr,~,MnO, was much smaller than that on LaMn,~,Cu,~,O,. A similar variation was observed for the amount of total or reversible adsorption of CO. TPD of Oxygen, CO, and CO Fig. 4 shows the TPD profiles of oxygen from LaMn, -,Cu,03 (x = 0-0.5) and La,CuO, in the tem-perature range 298-1123 K. Broad and small peaks appeared in a relatively low temperature range (473-823 K) for LaMn, -,Cu,03 (x = 0-0.5). Only small peaks were observed for La,CuO, . The amounts of oxygen desorbed below 823 K are summarized in Table 4.It was shown pre- viously that this amount for La, -,Sr,MnO, correlated well with the catalytic activity for the complete oxidation of propane." The highest value was obtained for LaMn,~,Cu,~,O,, but the amount was only 0.4 times that of a surface monolayer of oxygen.These trends are similar to those noted for the La,-,Sr,MnO, system,I8 and are in con- trast with those of the Co and Fe system^.^^,^^ Above 823 K, LaMnO, showed one large peak at 1060 K, which has been assigned to the desorption of the excess oxygen of LaMnO,,, accompanied by the reduction of Mn4+ to Mn3+.l8 Similar large peaks or ascents were observed for LaMn, -xCu,03 (x = 0.2-0.5). These results also resemble those found for the La, -,Sr,MnO, system.18 r& 1.0 I I 5 0.8 EL 0.6 i-II ternperature/K Fig.4 0, TPD profiles from LaMn,-,Cu,O, (x = 0-0.5) and La,CuO,. (-) LaMnO,; (--) LaMn,&u,,,O,; (-* -) LaMn,,,Cu,,,O,; (----) LaMn,,,Cu,~,O,; (-..-) LaMn,~,Cu,.,O,; and (. . .* .) La,CuO,. Table 4 Degrees of deactivation, amounts of 0, desorbed in the TPD of 0, and amounts of CO, desorbed in the TPD of CO, for LaMn, -xCu,03 (x = 0-0.5) and La,CuO, amount of OZb amount of CO,' X deactivation' degree of /lo-, mol g-' (<823 K) /lo-, mol m-' (<1123 K) 0 0.7 0.5 3.0 0.2 1.5 2.9 0.3 2.1 2.8 0.4 0.6 5.8 5.8 0.5 0.4 15.4 5.9 La,CuO, 0.2 4.5 9.2 Ratio of the steady-state rate to the initial rate at 473 K. The rate after 1 min was used as a measure of the initial rate. Amount of 0, desorbed in the range 298-823 K in the TPD experiment of 0,.Amount of CO, desorbed in the range 298-1123 K in the TPD experiment of CO, . J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 4.0 II II 3.0 r I!' ' & I i11 + 4.01 C .-4-g 3.0 v) -0 2.0 1 .o 0.0 273 473 673 873 1073 ternperature/K Fig. 5 CO, TPD profiles from LaMn,-,Cu,O, (x = 0-0.5) and La,CuO,. (-) LaMnO,; (--) LaMn,~,Cu,,,O,; -)(-a LaMn,,,Cu,,,O,; (-- - -) LaMn,,,Cu,.,O,; (-.--) LaMn,~,Cu,,,O,; 1 *and (. * .) La,CuO,. The TPD profiles of CO, are shown in Fig. 5. Several peaks, including two major ones at around 400 and 550 K, were observed for LaMn, -,Cu,03 (x = 0-0.5) and La,CuO, . Only a small amount of CO, was desorbed above 973 K.Tejuca et aL31 have assigned the peaks at around 390 and 540 K of LaMnO, to the desorption of a monodentate and bidentate carbonate, respectively. The amounts of CO, desorbed from 298 to 1123 K are summarized in Table 4. The amounts increased above x = 0.3. The coverages of CO, cal-culated from the data in Table 4, assuming that the cross- section of CO, is 17 were 0.29 and 0.94 for LaMn,. ,CU~.~O~ and La,CuO, ,respectively. The TPD profile of CO from LaMn,~,Cu,~,O, is shown in Fig. 6. For the desorption of CO, one peak and one ascent were observed at around 400 K and above 750 K, respec-tively. In addition to CO, CO, and 0, were also desorbed: 0, was desorbed in a similar way to the desorption in the 0, TPD (which was shown in Fig.4), but the amount of 0, desorbed below 823 K was smaller. As for CO,, only one desorption peak was observed at around 420 K. Similar results were obtained for x = 0, 0.3 and 0.5. The amounts of CO, CO, and 0, desorbed in CO TPD from LaMn,-,Cu,O, (x = 0-0.5) are summarized in Table 5. The differences between the data in Fig. 3 and Table 5 are due mainly to the different pretreatment (evacuation and oxidation). Catalytic Oxidation of CO When a reactant gas was introduced onto LaMn, -,Cu,O, (x = 0-0.5) and La,CuO, catalysts at 473 K after the pretreatment of 0, at 873 K for 1 h, the conversion decreased greatly in the initial stage and reached a constant value after ca. 2 h. It was confirmed for LaMn,,,Cu,~,O, J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 5.0 Y N 4.0 -0f 3.0 C .-'i1.0 p -0 0.0 273 473 673 a73 1073 temperature/K Fig.6 CO TPD profiles from LaMno,,Cuo,,03. (-) Desorption of CO; (---) desorption of CO,; (-* -) desorption of 0, in CO TPD; and (-.. . .) desorption of 0,in 0,TPD. that the conversions (<40%) at the steady state were pro- portional to the weight of catalyst. The degrees of deactiva- tion expressed by the ratios of the steady-state rate to the initial rate (after 1 min) at 473 K are shown in Table 4. The ratio decreased in the following order, LaMnO, z LaMn,~,Cu,,,O, > LaMn,~,Cu,~,O, > La,CuO,. A similar order of the deactivation has previously been found by the experiment using a recirculation system.l7 When the reaction temperature was raised to 873 K and then lowered, little hysteresis was observed for the conversion at the steady state for LaMn, -xCux03 (x = 0-OS), indicating the absence of irreversible deactivation of the catalysts. The apparent activation energies for the oxidation of CO on LaMn, -xCu,03 were 58 (x = 0), 43 (x = 0.2), 49 (x = 0.3), 55 (x = 0.5) and 48 (La,CuO,) kJ mol- ',respectively, and are in general agreement with the value of 52 kJ mol- ' reported for La,~,Sr,~,CoO,. l4 The steady-state rates of LaMn,-,Cu,O, (x = 0-0.5) and La,CuO, for the oxidation of CO at 573 K are shown in Fig. 7. Note that the ordinate is in the logarithmic scale. The rate showed a maximum at LaMn,,,Cu,~,O,, and the rate was about 25 and 15 times greater than those of LaMnO, and La,CuO, ,respectively.A similar trend was also observed at 473 K. Thus, the pronounced synergistic effect was confirmed for the steady-state activity, although, owing to the deactiva- tion, the magnitude of the effect was smaller than that for the initial activity reported previou~ly.'~ The steady-state rates of La,-,Sr,MnO, (y = 0.2 -1.0) at 573 K are also shown in Fig. 7. Upon Sr substitution, the rate also increased and reached a maximum at y = 0.6. However, the increase in the activity by Sr substitution was much smaller than that obtained by the Cu substitution. In Table 6, the steady-state CO oxidation rate of LaMn,~,Cu,,,O, is compared with those of several other catalysts, which have been reported to have high catalytic Table 5 Amounts of CO, CO, and 0, desorbed in the TPD of CO for LaMn, -xCu,03 (x = 0-0.5) x amount of CO" mol m-' amount of COZb/lop6mol m-' amount of 0,' mol g-' 0 1 .o 1.9 0.1 0.3 1.4 2.0 co.1 0.4 2.2 4.2 1.4 0.5 1.5 3.3 8.6 a Amount of CO desorbed in the range 298-1123 K in the CO TPD experiment.Amount of CO, desorbed in the range 298-1123 K in the CO TPD experiment. Amount of 0, desorbed in the range 298-823 K in the CO TPD experiment. 100 -E-N I 1 I I I I Id 0.0 0.2 0.4 0.6 La2Cu04 SrMn03 x, Y Fig. 7 Rates of CO oxidation over LaMn,-,Cu,O,, La,CuO, and La, -,Sr,MnO, at steady state. (0)LaMn, -xCu,03 (x = 0-0.5) and La,CuO,; (0)La, _,Sr,MnO, (y = 0.2-1.0);reaction temperature, 573 K.activities for the oxidation of CO. Among them, Ce0,- promoted Pt catalyst and La-Co perovskite-type mixed oxide are used commercially for the oxidation of CO, hydro- carbons et~.'~,~~Note that the activity of the LaMn,~,Cu,,,O, catalyst was the highest among the cata- lysts tested, even higher than those of the Ce0,-promoted Pt catalyst and La,~,Sr,~,CoO, . Furthermore, the calculated turnover frequency of LaMn,~,Cu,~,O, C1.4 molecules s-' (surface Mn and Cu atom)-'] was about five times higher than that of the Pt-Ce/Al,O, catalyst C0.31 molecules s-l (surface Pt atom)- '3 at 473 K. LaMn,~,Cu,,,O, prepared from an aqueous solution of metal acetates was about four times more active than LaMn,,,Cu,~,O, prepared from an aqueous solution of metal nitrates (both were prepared by a freeze-drying method).Discussion Structure and Surface Properties of Catalysts LaMn, -,CuXO3 prepared by the present freeze-drying method had the single perovskite phase in the range 0 6 x 6 0.4, and small amounts of La,CuO, and CuO were additionally observed for x = 0.5. Gallagher et ~1.'~reported the formation of the single phase in the range x = 0-0.6 by the calcination at 1273-1373 K in 0,. Rojas et uLt3 also reported the single perovskite phase up to x = 0.6 for Table 6 Rates of CO oxidation at 473 K catalyst conversion (yo) rate/cm3 min-' g-' 2.2 (2.8)" 19 (1.4)b LaMn0.6Cu0.403 1.6 (4.3) 9.5 (0.57) La0.8Sr0.2C003 La0.4Sr0.6Mn03 2.0 (6.0) 8.2 (0.18) 2.6 (10.1) 6.4 (0.21) LaCo0.6Cu0.403 2.8 (10.0) 7.0(0.24)LaFe0.6Cu0.403 0.5% Pt/Al,O,' 3.8 (0.11) 5% Pt/Al,O, 1.4(5.0) 6.8 (0.02) 0.7% Pt/Ce-A1 ,03 6.0 (10.2) 15 (0.31) a The numbers in parentheses are the catalyst weights (mg).The numbers in parentheses are the turnover frequencies calculated by assuming that the concentration of the surface transition-metal ions of the perovskite-type mixed oxide is 2.67 x 1OI8 ions m-' and that Pt is fully dispersed (dispersion = l),according to ref. 26,33 and 34, respectively. Unit: molecules s-' (transition metal or Pt atom)-'. 'Taken from ref. 16. samples prepared by the decomposition of the amorphous citrate complex. In the present work, the range of the single perovskite phase was somewhat narrower, probably owing to the relatively lower calcination temperature (1123 K) or dif- ferent reactivity of the precursors. As shown in Table 3, the surface compositions of the ele- ments of LaMn, -,CuXO3 (x = 0-OS), which were prepared from the metal acetates by a freeze-drying method, and La,CuO, agreed well with those of the bulk.Therefore, the surface properties of these LaMn, -,Cu,O, catalysts would reflect well those of the bulk. On the other hand, the use of the metal nitrates for freeze-drying gave less satisfactory agreement between the surface and bulk compositions as described previously.' Therefore, the greater activity of LaMn,~,Cu,~,O, prepared from metal acetates is probably due to its more nearly uniform composition between the surface and the bulk.The binding energies of Cu 2p3,, of the LaMn,-,Cu,O, (x = 0.2-0.5) catalysts were 934.2-934.3 eV, in agreement with the value for La,CuO,, in which the oxidation number of copper was two. It is known that Cu+ gives no satellite peaks in the Cu 2p regions, while Cu2+ has intense shake-up satellite peaks. The XP spectra of LaMn,-,Cu,O, (x = 0.2-0.5) had intense satellite peaks and the ZsaJZmain ratio of Cu 2p3,, and the Cu L,VV Auger peak agreed with those report- ed for Cu0.22936 These results show that copper ions on the surface of LaMn, -,Cu,03 (x = 0.2-0.5) are present as Cu2 +. On the other hand, in the case of Mn it is difficult to dis- tinguish the oxidation state from the binding It has been reported that LaMnO, perovskite has an oxidative non-stoichiometry, LaMnO,., (or defects at A and B sites), that is, a mixed valency of Mn3+ and Mn4+.38 When Mn3+ (or Mn4+) is partly substituted for Cu2+ in LaMnO,, the concentration of Mn4 + should increase and/or the concentra- tion of excess oxygen should decrease. Rojas et aL2, have reported, on the basis of the results of H, reduction, that the non-stoichiometry (6) of oxygen in LaMn, -,Cu,03 +d decreased with the increase of x up to 0.6, and manganese ions were present almost always as Mn4+ for 0.4 d x < 0.6. Therefore, the structural changes of LaMn, -,Cu,03 from rhombohedra1 (x = 0-0.2) to cubic (x = 0.3-0.5) phase could be explained; distortion due to the Jahn-Teller effect of Mn3+ is released by the oxidation of Mn3+ to Mn4+ upon Cu2+ substitution. Thus, the valency control of Mn by Cu substitution was achieved, as in the case of La, -,Sr,MnO, .Synergistic Effect for the Oxidation of CO As shown in Fig. 7, a pronounced synergistic effect was found for the coexistence of Mn and Cu of LaMn,-,Cu,O, cata-lysts for the oxidation of CO (more than 10 times enhancement). The effect was even greater for the initial activ- ity,17 owing to the absence of deactivation by CO, which is discussed in the next section. The catalytic activity for the oxidation of CO also increased by the substitution of Sr for La in LaMnO, . However, the increase in the activity for the Sr substitution was much less. We have previously reported that the amounts of 0, desorbed from La,-,Sr,MnO, below 823 K increased from y = 0 to y = 0.6 and this trend was well correlated with their catalytic activities for the complete oxidation of C,H, .A similar trend was observed for the desorption of 0, from LaMn, -,Cu,O, catalysts (Fig. 4 and Table 4). The amounts of 0, desorbed became much smaller for CO TPD, owing to the consumption of oxygen by oxidation of CO to CO, (Fig. 6 and Table 5). These results suggest that the 0, desorbed below 823 K for the Mn systems is very reactive for the oxi- dation of CO and therefore probably reflects the intrinsic J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 nature of the Mn oxide system or oxide ion in the neighbour- hood of the Mn ion.No such 0, desorption was observed for La,CuO, ,supporting this idea. Part of the enhancing effect of the Cu substitution of LaMnO, may thus be ascribed to the increased reactivity of the oxide ion of LaMnO, . However, the increase in the cata- lytic activity caused by the Cu2+ substitution was much greater than the increase in the activity caused by the Sr2+ substitution for La, -,Sr,MnO, ,although the increases in 0, desorption were very similar for both cases. The additional enhancement may be explained by the influence of the Cu ion on CO activation as discussed below. For LaMn, -,Cu,03, the amount of CO adsorption increased with x, reached a maximum at x = 0.4 and then decreased (Fig. 3), in parallel with the change of the reducibility measured by CO reduction (Fig.2). The parallel change indicates that the CO adsorption is an important factor controlling the reducibility of LaMn, -xCu,03 cata-lysts. Similarly, the increase of CO adsorption with the amount of Cu up to x = 0.4 suggests that Cu is the active site for CO adsorption. The idea is further supported by the much lower increase in the CO adsorption for Sr2+ substitut- ion (Fig. 3), although 0, TPD changed similarly for Cu and Sr substitution. The decrease in the activity from x = 0.4to 0.5 in Fig. 7 may be due to the formation of La,CuO,, which was less active for CO adsorption, as shown in Fig. 3. Therefore, it is very probable that the synergistic com- bination of the two properties, that is, the activation of 0, by Mn as in La,-,Sr,MnO, and of CO by Cu ion, brought about the great increase of the catalytic activity of LaMn, -,Cu,O,.The fact that the increase was much less for the oxidation of propane" also supports the specific activa- tion of CO by the Cu ion. Thus, the coexistence of Mn and Cu in the neighbourhood is essential for the enhancement of the catalytic activity. In other words, the role of Cu in LaMn,-,Cu,O, catalysts is the valency control of Mn, as in the case of Sr substitution for La in LaMnO, , and the activa- tion of CO. Inhibiting Effect of CO, The ratios of the steady-state rate (after ca. 2 h) to the initial rate decreased in the order of LaMnO, x LaMn,,,Cu,,,O, > LaMn,~,Cu,~,O, > La,CuO, (Table 4).This order is nearly the reverse order of the amounts of CO, adsorption; LaMnO, < LaMn,~,Cu,~,O, x LaMn,,,Cu,~,O, < La,CuO, (Table 4). The larger was the amount of CO, adsorption, the more greatly the activity decreased. This correlation indicates that the deactivation is caused by CO, adsorption. Similar results were obtained for the reaction carried out in the recirculation system; the rate of CO oxidation was expressed by eqn. (l).' The rate constant (k) reached a maximum at x = 0.4, while the b value was in the order of La,CuO, z LaMn,~,Cu,~,O, > LaMn,~,Cu,,,O, > LaMn,.,Cu,,,O, > LaMn,~,Cu,~,O, > LaMnO,. This order was in general agreement with the orders of the extent of the catalyst deacti- vation and the amount of CO, adsorption measured by CO, TPD, as shown in Table 4.In addition, the initial rate of LaMn, -,Cu,03 declined very rapidly and was greatly sup- pressed by the preadsorption of CO,. These results also support that the catalyst deactivation was caused by the adsorption of CO, . As shown in Fig. 5 and Table 4, the amount of CO, desorbed generally increased with the substitution of Cu, and the significant desorption of CO, from La,CuO, needed high J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 1189 temperatures such as above 570 K. Thus, the inhibiting effect of COz observed in LaMn, -xCux03 is mainly caused by the presence of Cu. Here it must be remarked that the pro- nounced synergistic effect observed in the study cannot be explained by the difference in the degree of this inhibiting 15 16 17 T.Nakamura, M. Misono and Y. Yoneda, J. Catal., 1983, 83, 151. P. K. Gallagher, D. W. Johnson Jr. and E. M. Vogel, J. Am. Ceram. SOC., 1977,60,28. N. Mizuno, Y. Fujiwara and M. Misono, J. Chem. SOC., Chem. Commun., 1989,316. effect, since the degree of deactivation was not in parallel with 18 T. Nitadori, S. Kurihara and M. Misono, J. Catal., 1986, 98, the steady-state catalytic activity (Fig. 7). As discussed in the previous section, the high activity most probably resulted from the synergistic effect of the Mn and Cu components; the ability for both the activation of 0, by Mn and of CO by Cu. 19 20 221. H-M. Zhang, Y. Teraoka and N. Yamazoe, Chem. Lett., 1987, 665. N. Mizuno, M. Yamato, M.Tanaka and M. Misono, Chem. Mater., 1989, 1, 232. This work was supported in part by a Grant-in-Aid for Scien- 21 22 T. Negas and R. S. Roth, J. Solid State Chem., 1970,1,409. D. C. Frost, A. Ishitani and C. A. Mcdowell, Mol. Phys., 1972, tific Research from the Ministry of Education, Science and Culture of Japan. 23 24,861. M. L. Rojas, J. L. G. Fierro, L. G. Tejuca and A. T. Bell, J. Catal., 1990, 124,41. 24 B. J. Tan, K. J. Klabunde and P. M. A. Sherwood, J. Am. Chem. References 1 2 3 4 5 6 7 8 9 10 11 12 13 R. J. H. Voorhoeve, in Advanced Materials in Catalysis, ed. J. J. Burton and R. L. Garton, Academic Press, New York, 1977, p. 129. L. G. Tejuca, J. L. G. Fierro and J. M. D. Tascon, Ah. Catal., 1989,36,237. M. Misono, in Future Opportunities in Catalytic and Separation Technology, ed.M. Misono, Y. Moro-oka and S. Kimura, Else- vier, Amsterdam, 1990, p. 13. N. Yamazoe and Y. Teraoka, Catal. Today, 1990,8,175. D. W. Johnson, Jr., P. K. Gallagher, F. Schrey and W. W. Rhodes, Am. Ceram. SOC. Bull., 1976,55520. H-M. Zhang, Y. Teraoka and N. Yamazoe, Hyomen Kagaku, 1987,8, 23. Y. Teraoka, T. Nobunaga and N. Yamazoe, Chem. Lett., 1988, 503. Y. Teraoka, H. Fukuda and S. Kagawa, Chem. Lett., 1990,l. T. Ichiki and M. Misono, 50th National Meeting of the Chemical Society of Japan, Tokyo, April, 1985, Abstr. No. 1Y25. M. Haruta, N. Yamada, T. Kobayashi and S. Iijima, J. Catal., 1989,115,301. S. D. Gardner, G. B. Hoflund, B. T. Upchurch, D. R. Schryer, E. J. Kielin and J. Schryer, J. Catal., 1991, 129, 114. M. Misono and N. Nojiri, Appl. Catal., 1990,64, 1. K. Tabata and M. Misono, Catal. Today, 1990,8,249. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 SOC.,1991, 113, 855. L. G. Tejuca, A. T. Bell, J. L. G. Fierro and M. A. Pena, Appl. Surf Sci., 1988,31, 301. M. A. Pena, J. M. D. Tascon, J. L. G. Fierro and L. G. Tejuca, J. Colloid Interface Sci., 1987, 119, 100. E. A. Lombardo, K. Tanaka and I. Toyoshima, J. Catal., 1983, 80,340. N. Mizuno, M. Tanaka and M. Misono, J. Chem. SOC., Faraday Trans., 1992,88, 91. T. Nakamura, M. Misono and Y. Yoneda, Bull. Chem. SOC.Jpn., 1982,55, 394. T. Nitadori and M. Misono, J. Catal., 1985,93,459. L. G. Tejuca, A. T. Bell, J. L. G. Fierro and J. M. D. Tascon, J. Chem. SOC., Faraday Trans. I, 1987,83,3149. J. M. D. Tascon and L. G. Tejuca, J. Chem. SOC., Faraday Trans. I, 1981, 77, 591. R. Barth, R. Pitchai, R. L. Anderson and X. E. Verykios, J. Catal., 1989, 116,61. H. C. Yao, Appl. Surf: Sci., 1984, 19, 398. H. C. Yao and Y. F. Yu Yao, J. Catal., 1984,86,254. G. Schon, Surf. Sci., 1973,35,96. M. Oku, K. Hirokawa and S. Ikeda, J. Electron Spectrosc. Relat. Phenom., 1975,7,465. B. C. Tofield and W. R. Scott, J. Solid State Chem., 1974, 10, 183. 14 T. Nakamura, M. Misono, T. Uchijima and Y. Yoneda, Nippon Kagaku Kaishi, 1980, 1679. Paper 3/07031J; Received 26th November, 1993