Surface Properties and Catalytic Activity of a Mo-fixed Catalyst' Structure of the Active Site and Mechanism for Selective Oxidation of Ethyl Alcohol BY YASUHIRO IWASAWA," YASUO NAKANO AND SADAO OGASAWARA Department of Applied Chemistry, Faculty of Engineering, Yokohama National University, Ooka-cho, Minami-ku, Yokohama, Japan Received 13th March, 1978 The physico-chemical properties of the surface of a Mo-fixed catalyst obtained by a ready reaction between Mo(n-C3H5)4 and surface OH groups of SiOa were investigated by X-ray photoelectron spectroscopy, photoluminescence technique, U.V. diffuse reflectance spectroscopy, i.r. spectrometry and Hz/02 uptake. The difference between the surface properties of the fixed catalyst and the con- ventional impregnation catalyst is discussed.The Mo-ked catalyst showed higher activity and selectivity than the conventional impregnation catalyst in the ethyl alcohol oxidation. Selective oxidation to form acetaldehyde proceeded by a two-stage redox (Mo6+ + Mo4+) mechanism. Coordinatively unsaturated dioxomolybdenum of uniform nature in a tetrahedral coordination is the active species in the selective oxidation of ethyl alcohol. The surface properties of heterogeneous catalysts have been studied extensively.2 However, the surface state of an impregnation catalyst which has important effects on the catalytic activity, selectivity and reaction mechanism is in general complex and still indefinite: the nature of the catalyst surface strongly depends upon the kind of support, the method of preparation of catalysts, the preparation condition, etc.Consequently, the mechanism of surface reaction on a conventional impregna- tion catalyst becomes ambiguous and conflicting results may sometimes be drawn. summarized the method of preparation of supported cataIysts using metal n-ally1 complexes and their catalytic activities. Metal n-ally1 complexes selectively react with the OH groups of solid supports like SiO, to form stable surface complexes under oxygen-free atmospheres. reported high activity for the metal oxide or metal catalysts supported on metal n-ally1 or methallyl complex bases in ethylene hydrogena- tion, butadiene polymerization, metathesis and hydrogen oxidation. A synthetic catalyst with a well-defined surface is very suitable for obtaining clear information on the active site and reaction mechanism.The present study contributes to our understanding of the surface properties of molybdenum-fixed silica catalyst, its catalytic activity and the action of the active site in the selective oxidation of ethyl alcohol, chosen as a typical reactant, in com- parison with a conventional impregnation catalyst. Ballard Yermakov and coworkers EXPERIMENTAL PREPARATION OF Mo-FIXED CATALYST The Mo-fixed catalysts were obtained from the ready reaction between a tetrakis-n- allylmolybdenum complex, Mo(n-CSH&, and the OH groups on the surface of silica (silica gel, 30-60 mesh, Nishio Industry, Japan, surface area of 567 m2 g-'). All procedures 2968Y . IWASAWA, Y . NAKANO AND S. OGASAWARA 2969 for synthesis were carried out in a specially devised Pyrex-glass apparatus under vacuum (base pressure of 5 x lW5 Torr) or in a flow of high purity (99.999 %) argon at atmospheric pressure.M o ( ~ ~ - C ~ H ~ ) ~ was synthesized according to the following equation described by Wilke et al. ; 5 -lS°C, 12 h diethyl ether C3H5Cl + Mg (large excess) C3H5Mg C1 -35"C, 1,5h -78'C, 20 h -3OOC 4CSHSMgClf MoClj -+ -+ + Mo(z-C~H~)~. (2) diethyl ether pentane The Mo-fixed catalyst was prepared by a similar procedure to that described by Yermakov et aL6 as folIows ; +Si-O\ 0 2 , 20°C Mo e psi-o>Mo=O 011) (bluish black - black) ( 5 ) +Si-O/ H2, 600°C b S i - 0 s!-o>Mo<z (IV) (white). %Si-O\ 0 2 , 3OOOC +Si-o/ H2, 450°C bs1-0 Mo-0 + The procedure was carried out in a vacuum of base pressure 1 x Torr.Silica gel was heated at 550°C for 1.5 h in oxygen (70 Torr) with a liquid N2 trap and evacuated for 1 h at the same temperature prior to Mo fixation. The surface complex (I) formed was washed three times with pentane in vacuum, in order to remove the Mo-complex, with no chemical interaction with the silica surface. Some Mo remained in the pentane solution after the Mo-fixing and the washing solution was analysed by both gravimetry as Moo3 (requiring careful treatment because of easy sublimation) and colorimetry at 460nm using a thiocyanate- SnClz method. Thus, by subtracting the final M o ( ~ - C ~ H ~ ) ~ concentration left in the pentane from its initial concentration, the amount of molybdenum fixed on the silica gel surface was determined. PREPARATION OF CONVENTIONAL IMPREGNATION CATALYST The impregnation catalyst was obtained by an ordinal impregnation method using aqueous ammonium paramolybdate solution.REACTION AND PRODUCT ANALYSIS The decomposition of ethyl alcohol to acetaldehyde, diethyl ether, ethylene and water in the presence or absence of oxygen was investigated in a conventional closed-circulating system (dead volume : 170 cm3) using 0.39 g of catalyst. The reaction products were analysed by gas chromatography using a 2 m column of dioctyl sebacate at 70°C. SPECTROSCOPIC STUDY The i.r. absorption spectra were measured using a 36.5 mg disk of 20 mm Mo catalyst fixed on to aerosil, under various conditions, in a glass-cell directly connected to a closed circulating system using a JASCO IRA-2 i.r.spectrometer. The U.V. diffuse reflectance spectra were taken in the range 210-700 nm after the catalysts had been oxidized at 550°C for 2 h in a separate system. The X.P.S. emissions were monitored by means of McPherson ESCA 36 spectroscopy, where the binding energies were referred to 83.7 eV of the Au 4f712 level taking account of the charge effect. Photoluminescence was investigated in a quartz cell2970 PROPERTIES OF Mo-FIXED CATALYST directly connected to a vacuum system using a JASCO FP-4 spectrophotometer. The reproducibility of the peak position at maximum intensity was within 3 nm. The e.s.r. spectra were measured on a JEOL e.s.r. spectrometer after quenching the reaction by cooIing the catalyst to room temperature during the ethyl alcohol oxidation.RESULTS Mo-FIXING PROCESS Ballard and Yermakov et aL6 demonstrated that the metal n-ally1 complexes easily react with the surface OH groups of Si02 at room temperature. The decrease in the intensity of the OH stretching vibration in fig. 1 indicates that the Mo ions combined with the silica surface through the oxygen atom of the OH groups. wavenumber 1crn-l (b) SiOz, (c) the difference spectrum of (a)-@). FIG. 1.-Intensity change in the OH vibration band by Mo-fixing; (a) the fixed catalyst (IV), Yermakov et a1.’ suggested from an i.r. spectrum that the ligands of the surface complex (I) are of the 0-ally1 type. The reaction of Mo(n-C3H,), and the surface OH groups was more rapid when the silica containing the larger number of OH groups was employed. Thus the rate of Mo fixation depended upon the surface state of silica.At the silica surface with a low content of OH groups, in addition to eqn (3), the following step may also participate in the Mo-fixing process ;6 + Si-OH 20ec +Si-0 Si-0>Mo(c3H5)2 +C3H6(g) (7) * Si\, + Mo(Tc-C~H~)~ -+ + Si/ 3 Si-C3H5 REDOX BEHAVIOUR OF Mo-FIXED CATALYST The reversible cycle of reduction-oxidation between the fixed catalysts (11), (111) and (IV) was observed as shown in eqn (5) and (6). The oxidation of (11) to (111) by oxygen was very rapid at room temperature. The amount of oxygen consumed was 1.1 0-atoms per Mo atom. More than stoichiometric uptake of oxygen may be caused by adsorption of oxygen and/or partial over-oxidation of Mo2+ to Mo6+,Y . IWASAWA, Y . NAKANO AND S .OGASAWARA 297 1 although this is small. Further oxidation from (111) to (IV) proceeded slowly at 100°C and was very ready at 280-300°C. The catalyst (IV) was reduced to (111) with hydrogen at 450"C, at which stage the uptake of hydrogen molecules per Mo atom was unity with an error of 6%. (111) was converted to the reduced form (11) in 3 h at 600°C under 80 Torr of hydrogen, the ratio of the amount of dihydrogen uptake to the number of Mo atoms being 2k0.1 as shown in fig. 2. Thus the fixed catalysts, (II), (111) and (IV), can be produced individually by controlled reduction- oxidation steps. The irreversible surface reconstruction of the Mo-fixed catalyst employed was not observed during repeated redox treatments, judging from the redox behaviour of the catalyst and the reproducible catalytic activity.reduction time/h FIG. 2.-Amount of Hz consumed per Mo atom during reduction with H2 ; Mo/Si02 = 1.7 wt %. PHYSICOCHEMICAL PROPERTIES OF Mo-FIXED CATALYST (IV) The X.P.S. of the fixed catalyst (IV) showed characteristic doublet peaks of a hexavalent molybdenum ion. The values were found to be 235.7 &- 0.3 eV for the Mo 3 4 orbital and 232.7 k0.3 eV for the Mo 3 4 level and the full width at half maximum height of the Mo 3 4 emission peak was x 2.4 eV. MOO, as a standard sample gave the well-known doublet peaks at 235.5 and 232.5 eV for 3d3 and 3d3, respectively. E m . spectroscopy showed no signal at any temperatures, Mo5+ species not being contained in catalyst (IV). The U.V. diffuse reflection spectrum showed absorption peaks at 290 and 225 nm, neither peak nor shoulder at longer wavelengths being found.An illumination in the charge transfer band of the Mo=O double bond was followed by a strong emission in the energy range (20-25) x lo3 cm-l. The luminescence from the Mo-impregnated catalyst depended upon the energy of the exciting light as shown in fig. 3. The emissions are divided into two. Photoexcitationat (34.4-35.7) x lo3 cm-l generated an emission peak at 21.8 x lo3 cm-l, while illumination at (30.8-32.3) x lo3 cm-l was accompanied by luminescence at 22.9 x lo3 cm-'. On the other hand the emission energy of the Mo-fixed catalyst (IV) was independent of the exciting energy, indicating the presence of only one kind of molybdenum emitting species. The typical transmittance i.r.spectra of the fixed catalyst (IV) are given in fig. 4(a) and (b), where absorption peak at 916 cm-l was observed, SiO, having no peak near this position. The absorption is due to the stretching vibration of the Mo=O bond. The reactivity of this molybdenum-oxygen bond was checked by i.r. spectrometry. Fig. 4(c) is a variation of the band intensity of the Mo-0 bond vibration with ex- posure to ethyl alcohol, methyl alcohol or oxygen at given temperatures. The Mo=O double bond readily reacted with ethyl alcohol at 150°C ; the intensity markedly decreased from 1 to 2, in fig. 4(c), by reduction with ethyl alcohol in the absence of2972 PROPERTIES OF Mo-FIXED CATALYST oxygen. The absorption band at 916 cm-l was restored to its original level by oxidation with oxygen. The decrease in the band with ethyl alcohol was much faster than its increase with oxygen [l+2 and 2 4 3 in fig.4(c)]. Methyl alcohol behaved similarly to ethyl alcohol except at the higher temperature studied. The 2 20 emission energy/103 cm-l FIG. 3.-Photoluminescence of (a) fixed catalyst (IV) and (b) impregnation catalyst ; exciting light, 1, 285 ; 2,295 ; 3, 320 nm. 4 0 90 ,;- wavenumber /cm-' FIG. 4.-1.r. absorption spectra of the Mo-fixed catalyst ; (a) Difference spectrum between the Mo- fixed catalyst (IV) and Si02, where a reference Si02 disk of adequate weight was chosen by trial and error. (6) (1) fixed catalyst (IV), (2) Si02, (3) calculated difference spectrum between (1) and (2). (c) (1) fked catalyst (IV), (2) spectrum after reaction with ethyl alcohol for 20 min at 150°C in the absence of 02, (3) after oxidation with O2 for 1.2 h at 150°C after (2), (4) in a steady state under ethyl alcohol : Oz = 1 : 11 at 155"C, (5) after reaction with methyl alcohol at 180°C in the absence of 02.Y .IWASAWA, Y . NAKANO AND S . OGASAWARA 2973 steady-state level in the i.r. absorption band intensity depended upon the ratio of ethyl alcohol to oxygen. The behaviour of the 916cm-l absorption band was compatible with that expected from the kinetic data. DEHYDROGENATION OF ETHYL ALCOHOL IN THE ABSENCE OF OXYGEN The reactivity of the active sites, the Mo=O bonds, in the catalyst was studied in a closed circulating system in the temperature range 100-200°C in the absence of oxygen. The catalysts were treated at 550°C for 1.5 h under 70 Tom of oxygen with a liquid N2 trap followed by evacuation for 1 h prior to dehydrogenation; when necessary, the fixed catalyst (IV) was converted to catalysts (11) or (111) as shown in eqn (5) and (6).The products of the reaction were acetaldehyde, diethyl ether, ethylene and water. Evolution of hydrogen was not observed in the course of dehydrogenation. The formation of acetaldehyde on both the fixed (IV) and the impregnated catalysts are shown in fig. 5, where reactions on the fixed catalysts having tetra- and di-valent molybdenum structures are also shown. The dehydrogena- tion activity of the fixed catalyst (IV) was found to be much higher than those of the impregnated and other fixed catalysts. The fixed catalyst with Mo2+ was essentially inactive.Ethyl alcohol dehydrogenation on the conventional impregnation catalyst below 150°C was not significant, while on the fixed catalyst (IV) it took place even below 100°C. The activity of dehydrogenation decreased in the following order : 10 30 5 reaction timelmin FIG. 5.-Dehydrogenation activities of the fixed catalysts with three different structures and of the impregnation catalyst in the absence of oxygen ; reaction temp. = 150"C, catalyst = 0.39 g, Mo/SiOz = 1.7 wt %. (1) Fixed catalyst (IV), (2) fixed catalyst (III), (3) fixed catalyst a), (4) im- pregnation catalyst. Acetaldehyde was produced more selectively on the fixed catalyst than over the impregnated one as shown in table 1. The selectivity towards acetaldehyde formation on the fixed catalyst was 100 % at 100°C.A difference in the distribution of the inter- and intramolecular dehydration products on both catalysts was also observed : diethyl ether was predominantly produced on the fixed catalyst, while on the irn- pregnated one ethylene was mainly formed at higher temperatures. The activation energies of dehydrogenation were determined from the initial rates of reaction to be 51.0k5.4 kJ mol-' on the fixed catalyst (IV) and 46.0k5.4 kJ mol-1 on the im- pregnated one.2974 PROPERTIES OF Mo-FIXED CATALSYT The dependence of the initial rates of acetaldehyde formation upon the number of Mo6+ ions remaining in the fixed catalyst which could be determined from the amount of Mo6+ converted to Mo4+ by the controlled reduction mentioned above, is shown in fig.6 in which the initial rates change linearly as the amount of Mo6+ decreased. This result indicates that the active species is a hexavalent molybdenum of uniform nature. TABLE SE SELECTIVITY AND ACTIVATION ENERGY OF THE DECOMPOSITION OF ETHYL ALCOHOL ON THE MO-FIXED CATALYST (Iv) (A)* AND THE IMPREGNATION CATALYST (B)* IN THE reaction temperature 1°C 100 125 150 175 200 activation energy /kJ mol-l ABSENCE OF OXYGEN ylectivity/ % A B A B A B acetaldehyde diethyl ether ethylene - 0 0 100 - 0 - 96.2 I 3.8 - 86.7 82.6 8.5 9.6 4.8 7.8 82.7 73.5 11.3 11.8 6.0 14.7 67.4 - 14.2 18.3 - 51.0k5.4 46.0k5.4 75.2k7.5 62.7k7.5 - 81.1k8.4 * Catalyst = 0.39 g, Mo/SiOz = 1.7 wt%. X 4 4 3 - 2 0,- 0.5 e a x v Mo6+/total Mo FIG. 6.-Dependence of the initial rates of CHJCHO formation upon the amount of Mo6+ in the fixed catalyst (IV) at 150°C.OXIDATION OF ETHYL ALCOHOL WITH OXYGEN The rate of ethyl alcohol dehydrogenation in the absence of oxygen decreased with reaction time. When oxygen was admitted on to the catalyst during the reaction as shown in fig. 7, the oxidation activity was recovered. The oxidation of ethyl alcohol to form acetaldehyde in the presence of oxygen attained a steady state. The selectivity in the presence of oxygen was higher than in its absence. Table 2 shows a higher selectivity on the fixed catalyst (IV) than on the impregnation catalyst. The rate equation under experimental conditions was approximately expressed by the power rate law as follows; r + kP&2P&5, where PE and Po, represent the partial pressures of ethyl alcohol and oxygen, respectively. When acetaldehyde formation from ethyl alcohol in the presence of oxygenY .IWASAWA, Y . NAKANO AND S . OGASAWARA 2975 proceeds by the two-stage redox mechanism, the following scheme and rate equation for the steady state procedure may be expressed ; :$I:> Mo <: +C2H50H + k1 4si-o "-"> Mo=O + CH3CH0 + H,O (8) Mo=O+&O,+ :I:> Mo //o No (9) k2 + Si-O\ 3 Si-O/ PElr = l/ki +P&k2 J P . 2 (10) PE/r is plotted against P E I J c 2 in fig. 8, where both PE and Po, are varied and the rates measured. The correlation between both variables, PE/r and P,/,/X, was found to have good linearity. I .- reaction timelmin FIG. 7.-02 effect on CH&HO formation on 0.39 g catalyst (Mo/Si02 = 1.7 wt%) at 150°C; -: in the presence of 02, - - - : in the absence of 02.TABLE 2.-sELECTIVITY AND ACTIVATION ENERGY OF THE OXIDATION OF ETHYL ALCOHOL WITH OXYGEN ON THE FIXED CATALYST (Iv) (A)* AND THE IMPREGNATION CATALYST (B)* reaction selectivity/ % temperature acetaldehyde diethyl ether ethylene /"C A B A B A B 100 100 - 0 - 125 97.1 - 2.9 - 150 94.5 89.2 4.5 7.0 1 .o 2.8 175 93.5 87.2 5.3 7.6 1.2 5.2 200 - 85.1 - 8.1 - 6.8 - 0 0 - activation energy /kJ mo1-I 62.3+ 5.4 58.9+ 5.4 - 56.4+ 8.4 - - * Catalyst = 0.39 g, Mo/SiOz = 1.7 wt%. The rate constants, kl and k2, could be calculated from fig. 8. The values at 150°C are given in table 3. The initial rates of reduction of the Mo6+ species with ethyl alcohol [eqn (S)] and of oxidation of the Mo4+ species with oxygen [eqn (9)] were obtained independently by gas chromatography and by volumetry of O2 uptake, respectively.The rate constants determined from the rates thus obtained are also given in table 3.2976 PROPERTIES OF Mo-FIXED CATALYST PE/d% FIG. 8.-correlation of pE/r with pE/dK2 at 150°C. TABLE 3.-RATE CONSTANTS, kl AND kZ, IN STEPS (8) AND (g), RESPECTIVELY * 10kdmin-1 103k2/[m3 (s.t.p.)]+ min-1 1 1.7+ 0.3 7.6+ 0.8 2 1 . 5 k 0 . 3 7.1k0.9 * 1 : Obtained from eqn (10) ; 2 : independently obtained from the rates of steps (8) and (9). E.S.R. MEASUREMENT When the catalyst surface was exposed to ethyl alcohol at 165°C in the absence of oxygen, an asymmetric signal assigned to Mo5+ species (1.93, for g1 and 1.8g5 for gji) was observed. However, the maximum amount of the Mo5+ ions was found to be 0.5+0.12 % of the total molybdenum contained in the catalyst.The amount of Mo5+ ion in the steady state oxidation was only 0.3 %. On the other hand about 20 % of the Mo6+ ions under similar conditions were converted to Mo4+ judging from the amount of acetaldehyde formed. It seems that the Mo5+ species does not play a role in the oxidation of ethyl alcohol. DISCUSSION The conventional impregnation method followed by calcination for catalyst activation has been employed extensively in the preparation of supported catalysts. For a molybdenum-supported catalyst, inorganic acids or their salts, like ammonium paramolybdate in an aqueous solution, are generally used. Oxymolybdenic species in aqueous solution are in equilibrium between and (MoxOy)n-, depending upon the pH, temperature and concentration of the solution.Consequently, a non-uniform distribution of active sites and heterogeneous properties of the catalyst surface may be produced during the immersion and calcination. On the other hand the fixed catalyst was synthesized molecularly taking advantage of the ready reaction between Mo(n-C,H,), and the surface OH groups as expressed in eqn (3)-(6). The decrease in the peak intensity in the OH vibration region and the fact that Mo(n-C,H,), did not form a significant surface complex on the support with no OH groups available under the present experimental conditions, j- indicates that the surface OH groups participate in the Mo-fixing process. The Mo-fixation should be governed by the population of OH groups at the sup- port surface.Therefore the distribution of molybdenum could be controlled by the number and topography of the surface OH groups. The silica gel used was found to have nearly one OH group per 100 A2 on average, according to the method reported by Sat0 et aL9 using (C2H5),A1. Peri demonstrated, on the -f Reaction of MO(~T-C~H~)~ with the silyl ether bond of the SOz surface may occur, but its rate seems to be small.Y. IWASAWA, Y. NAKANO AND S . OGASAWARA 2971 basis of a stoichiometric reaction of AlCl, or SiC14 with surface OH groups, that silica gel largely holds paired OH groups even after the heat-treatment at 600°C. This indicates that Mo-Wng mainly proceeds by eqn (3) and that step (7) may play a lesser role in the fixing reaction under the present conditions.Consequently, the molybdenum ions may be atomically dispersed at the silica surface with chemical bondings through oxygen atoms of the silanol groups : the mean distance between the nearest-neighbour molybdenum ions in the fixed catalyst is estimated to be w 20& assuming ideal distribution of the paired OH groups. Indeed, all the molybdenum ions participated in the uptake of H2 or O2 in the stepwise redox process on the fixed catalysts, (IV) e (111) e (11), as shown in fig. 2 : the behaviour observed in the redox steps was very similar to that found by Yermakov et aL6 The molybdenum ion thus fixed [catalyst(IV)] was confirmed to be in a hexavalent level by X.P.S. emission data. The interaction of molybdic acid with the surface OH groups on the supports has been postulated ideally l1 on the basis of an acid-base interaction accompanied by the elimination of water as follows ; H H I I 0 0 + 0 0 - \& 0 ' ' 0 I H I H However, since an anion of paramolybdenic acid is usually used in the impregnation, the scheme according to Yamagata et aE.,13 must be modified as follows: H H I I + n OH'.anion of para- "-A paramolybdenic + (mdybdcnic acid) This form then follows activation by calcination. Yamagata l2 also demonstrated that molybdenum ions are supported on the sites at which anionic OH groups are located. The presence of acidic OH groups resulted in the precipitation and deposition of some molybdenum species. Castellan et aZ.I3 showed the presence of silicomolybdic acid on the Mo-impregnated silica catalyst where a tetrahedral molybdenum species was also found.With higher molybdenum content a poly- molybdate species was also observed. Thus these investigations show the hetero- geneous surface state of the impregnation catalysts. The situation of the Mo-fixed catalyst is different because Mo(n-C,H,), may react with the hydrogen (which is protonic) of the OH groups rather than the basic OH groups, followed by the removal of propene molecules. The Mo-fixed catalyst may have active sites in a different environment from the conventional impregnation catalyst. The dependence of the luminescence from the impregnation catalyst upon the excitation energy of the charge transfer band of the Mo=O bond, as shown in fig. 4, indicates that there must be at least two different emission centres.14 On the other hand the emission energy of the Mo-fixed catalyst (IV) was independent of the excital tion energy change, reflecting the homogeneity of the molybdenum emitting species at the surface.It is concluded from observations of the photoluminescence that2978 PROPERTIES OF Mo-FIXED CATALYST the active species in the Mo-fixed catalyst has a uniform environment, where the surface of the conventional impregnation catalyst was heterogeneous. The quench- ing behaviour of oxygen and propene molecules clearly support this.I5 The U.V. reflection spectrum of the fixed catalyst (IV) with absorption peaks at 290 and 225 nm shows that the Mo ion is situated in a tetrahedral position :I6 the 290 and 225 nm peaks are assigned to the transition of 3t, c tl and 3t2 - 2t,, respectively, in a tetrahedral c~mplex.~’ This is in agreement with the structure expected from the synthesis scheme.The absorption band at 916 cm-1 could be assigned to an asymmetric stretching vibration of the molybdenum-oxygen bond of either the tetrahedral or octahedral oxomolybdenum(6 + ) species. Mitchell et aZ.I demonstrated that the frequency of a Mo=O bond vibration depended upon the number of the coordination and the terminal oxygen atoms on the molybdenum : the oxomolybdenum species with two terminal oxygen atoms in tetrahedral coordination had a wavenumber near 920 cm-l. The frequency also depends upon the degree of distortion of the tetrahedral structure. Thus the absorption peak at 91 6 cm-’ indicates a tetrahedral dioxomolybdenum structure.Consequently, the collective results of U.V. reflectance spectroscopy, i.r. spectro- metry, the photoluminescence study, the X.P.S. emission peak and the volumetric H2/0, uptake confirm, as expected from the step-wise synthesis procedure for the Mo-fixed catalyst preparation, that a molybdenum ion in the fixed catalyst (IV) which is highly distributed in a hexavalent level has the dioxo-structure with homo- geneous character, and is in tetrahedral coordination at the silica surface, while a molybdenum ion in the conventional impregnation catalyst is heterogeneous in the surface chemical state. These differences in the surface properties of the fixed and impregnated catalysts should correlate with the differences in their catalytic activities and their selectivities in ethyl alcohol oxidation. Investigation of a fixed catalyst with a well-defined active site give clearer information on the reaction mechanism.Fig. 5 and table 1 show that the Mo-fixed catalyst (IV) is very active and selective in ethyl alcohol dehydrogenation compared with the conventionally impregnated catalyst. Lower-valent molybdenum ions showed much less activity. Fig. 6 shows that hexavalent Mo ions are active species for dehydrogenation. Again the linearity in the relation of the amount of Mo6+ with activity indicates that Mo ions in the fixed catalyst have uniform activity. The rate of oxidation to form acetaldehyde was reduced as the number of Mo6+ ions decreased in the absence of oxygen, while the activity was restored by the admission of oxygen into the system in the course of the ethyl alcohol dehydrogenation as shown in fig.7. Steady-state oxidation pro- ceeded in the presence of oxygen. Thus ethyl alcohol oxidation seemed to proceed by the two-stage redox mechanism expressed in the eqn (8) and (9). In the steady- state procedure the reaction rate is given by eqn (lo), which is confirmed by fig. 8. The fixed catalyst (111) with a tetravalent Mo ion could be produced as needed as above mentioned, while in the case of a conventional impregnation catalyst, controlled formation of uniform Mo4+ species may be difficult. The rate of oxidation of the Mo4+-fixed catalyst with oxygen was determined volumetrically. The rate of reduction of the Mo6+ species to Mo4+ was obtained under oxygen-free conditions.Accordingly, the rate constants, kl and k2, in eqn (8) and (9) could be determined independently. Table 3 shows good agreement in the values obtained from the steady-state eqn (10) and the individual measurements of the two steps. These results prove that ethyl alcohol oxidation took place in the redox reaction between the Mo6+ and Mo4+ species.Y. IWASAWA, Y. NAKANO AND S. OGASAWARA The reaction scheme may be shown in more detail as follows : 2979 0 0 ' \o I Si 1 Si /I\ /I\ 0 II 4 " O ' . I Si Si /I\ /I\ + C~HSOH H H I 0 (p: Si /I\ Si " f2H5 I 0 0 \I' /O Mo 0 / \o I Si I Si /I\ /I\ CH3CHO ( 3 ) Si Si /I\ /I\ Dissociation of the 0-H bond in alcohols on metal oxide catalysts is generally easy. Indeed, the i.r. spectra showed a rapid decrease in the intensity of the Mo=O vibration band with exposure to ethyl alcohol even below 100°C, at which tempera- tures formation of acetaldehyde was very slow.This implies that step (l), which is the dissociation of the OH bond of ethyl alcohol accompanied by the change of the Mo=O double bond to an Mo-0 single bond, is not rate-controlling. The absorp- tion bands in the steady-state oxidation of ethyl alcohol followed by evacuation at 150°C appear at 1446, 1385 and 1373 cm-1 in the region of the CH3 and CH2 de- formation modes besides the 2982,2935 and 2900 cm-l bands for the C-H stretching vibrations. These absorption peaks arise from the CH,CH,O-group.l The silica support showed no peak with significant intensity in this region under similar experimental conditions.Accordingly ethyl alcohol molecules dissociatively adsorb on dioxomolybdenum(6 + ) species forming ethoxide group. The ethoxide group in the i.r. spectrum decreased with reaction time at 150°C. The rate-determining step in oxidation of ethyl alcohol to form acetaldehyde should be the abstraction of a hydrogen atom (hydride) from the C-H bond, as shown in the step (3) of the scheme. The formation of water was also observed in the absence of oxygen, indicating that this process must occur prior to oxidation of the Mo4+ species by oxygen. No evolution of hydrogen was observed during the oxidation in spite of the presence or absence of oxygen. Consequently, ethyl alcohol oxidation takes place in conjunction with the redox cycle of Mo6+ + Mo4+, where the dioxomolybdenum(6+) structure is the active species.The oxidations of methyl alcohol and ethyl alcohol have been explained on the basis of a two-stage redox mechanism on MoO~,~O V205-M003,21 Fe-molybdate,22 Th-molybdate, 23Sn02-Mo03,24 e t ~ . ~ Our conclusion gives clear evidence for the mechanism of alcohol oxidation. Trifiro and Pasquon 26 demonstrated that the selective oxidation of methyl2980 PROPERTIES OF MO-FIXED CATALYST alcohol could be connected with the presence of a Mo=O terminal bond which has the character of a labilized double bond. The present results on a well-defined k e d catalyst prepared by stepwise synthesis clearly support this idea. Furthermore, the coordinatively unsaturated dioxomolybdenum(6 + ) which has uniform dis- tribution in a tetrahedral coordination is demonstrated to be the active species in the selective oxidation of ethyl alcohol.Ethyl alcohol molecules readily attack the unsaturated coordination sphere of the tetrahedral dioxomolybdenum ion and are activated for dehydrogenation. The structure and environment of the Mo6+ ion may affect the electronic property (a degree of c~valency)~’ of the Mo-0 bond as the active site : hence the abstraction of hydride from the methylene group of ethyl alcohol becomes easier on an Mo-fixed catalyst than on a conventional impregnation catalyst. The high activity of the coordinatively unsaturated dioxomolybdenum, which has uniform character in the tetrahedral position seems to be general in de- hydrogenation processes with various reactants.Investigation of the allyl-type oxidation of propene will confirm this.l5 The present study on an Mo-fixed catalyst with a well-defined surface state has many advantages for investigation of the essential factors of catalysis, the structure and environment of the active site, and the reaction mechanism. The authors thank Prof. Y. Kondo of Rikkyo University for measurements of the U.V. diffuse reflectance spectra and Dr. M. Soma of Tokyo University for measurements of the X.P.S. emission spectra. Part of this work was financially supported by the Asahi Glass Foundation. Y. I. thanks the Sakko-kai Foundation for a subsidy. The preliminary results were reported by Y. Iwasawa et al., Shokubai, 1977, 19,4. The term “ fixed catalyst ” is proposed to indicate a well-defbed supported catalyst with chemical bonding between the transition element and the support in order to distinguish this type of catalyst from conventional impregnation catalysts.For examples of Mo-containing catalysts, W. K. Hall and M. LoJacono, Proc. 6th Int. Congr. Catalysis (London, 1976), 1977,1,246 ; N. Giordano, J. C. J. Bart, A, Vaghi, A. Castellan and G. Martinott, J. Catalysis, 1975,36,81 ; T. Fransen, 0. Van Der Meer and P. Mars, J. Catalysis, 1976,42,79 ; G. T. Pott and W. H. J. Stork, in Preparation of Catalysts, ed. B. Delmon, P. A. Jacobs and G. Poncelet (Elsevier, Amsterdam, 1976), p. 548 ; V. M. Vorotyntsev, V. A. Shvets and V. B. Kazanskii, Kinetics and Catalysis, 1972, 12, 1108. G. H. Ballard, Advances in Catalysis (Academic Press, N.Y., 1973), vol.23, p. 263. Yu. I. Yermakov, Catalysis Rev.-Sci., 1976, 13, 77 ; Yu. I. Yermakov, B. N. Kuznetsov and Yu. A. Ryndin, J. Catalysis, 1976, 42,73, and references therein. G. Wilke, B. BogdanoviC, P. Hardt, P. Heimbach, W. Keim, M. Kroner, W. Oberkirch, K. Tanaka, E. Steinriicke, D. Water and H. Zimmermann, Angew. Chem. Ini. Edn, 1966,5, 151. Yu. I. Yermakov and B. N. Kuznetsov, Prep. 2nd Japan-Soviet Catalysis. Seminar, (Tokyo), 1973, p. 65. ’ Yu. I. Yermakov, B. N. Kuznetsov, L. G. Karaktchiev and S. N. Derbeneva, Kinetica i Kataliz, 1973, 14, 709. G. A. Tsigdinos and C. J. Hallada, Molybdenum Chemicals-Chemical Data Series Bulletin, 1969,14 Feb. ; A. K. Babko and G. I. Gridchina, Zhur. neorg. Khim., 1968,13,61(Eng). M. Sato, T. Kanbayashi, N. Kobayashi and Y.Shima, J. Catalysis, 1967, 7, 342. lo J. B. Peri and A. L. Hensley Jr., J. Phys. Chem., 1968, 72,2926. l1 M. Dufaux, M. Che and C. Naccache, J. Chim. phys., 1970, 67, 527 ; G. S. John, M. J. Den Herder, R. J. Mikorsky and R. T. Wasters, Advances in Catalysis (Academic Press, N.Y., 1957), vol. 9, p. 225. l2 N. Yamagata, Y. Owada, S . Okazaki and K. Tanabe, J. Catalysis, 1977, 47, 358. l3 A. Castellan, J. C. J. Bart, A. Vaghi and N. Giordano, J. Catalysis, 1976, 42, 162. l4 G. T. Pott and W. H. J. Stork, CataIysis Reviews (Marcel1 Dekker, 1976), p. 163 ; S. Coluccia, M. Deane and A. J. Tench, Proc. 6th Int. Cong. Catalysis (London, 1976), 1977, 1, 171 ; V. B. Kazansky, A. M. Gritscov and V. A. Shvets, Prep. 2nd Japan-Soviet Catalysis Seminar (Tokyo), 1973.Y . IWASAWA, Y . NAKANO AND S . OGASAWARA 298 1 l5 Y. Iwasawa and S. Ogasawara, to be published. l6 M. Che, F. Figueras, M. Forissier, J. McAteer, M. Perrin, J. L. Portefaix and H. Praliand, Proc. 6th Int. Cong. CatalysisLondon (1976), 1977, 1,261 ; G. N. Asmolov and 0. V. Krylov, Kinetics Catalysis, 1971, 11, 847. l7 M. Wolfsberg and L. Helmholz, J. Chem. Phys., 1952, 20, 837. l 8 P. C. H. Mitchell and F. Trifiro, J. Chem. SOC. (A), 1970, 3183. l9 Y. Soma, T. Onishi and K. Tamaru, Trans. Faraday SOC., 1969, 65, 2215 ; R. G. Greenler, 2o G. K. Boreskov, B. I. Popov, V. N. Bibin and E. S. Rozischnikova, Kinetika i Kataliz, 1968,9, 2 1 R. S. Mann and M. K. Dosi, J. Catalysis, 1973, 28, 282. 22 J. Edwards, J. Hicolaidis, M. B. Cutlip and C. 0. Bennett, J. Catalysis, 1977, 50, 24; M. Carbucicchio and F. Trifiro, J. Catalysis, 1976, 45, 77 ; P. Jiru, B. Wichterlova and J. Tichy, Proc. 3rd Znt. Cong. Catalysis (Amsterdam), 1965, 1, 199. 23 V. Srihari and D. S. Vismanath, J. Catalysis, 1976, 43,43. 24 M . Niwa, M. Mizutani and Y. Murakami, Chem. Letters, 1975, 1295. 25 J. Novakova, P. Jiru and V. Zavadil, J. Catalysis, 1971, 21, 143 ; P. L. Villa, A. Szabo, F. 26 F. Trifiro and I. Pasquon, J. Catalysis, 1968, 12,412. 27 G. K. Boreskov, Proc. 5th Int. Congr. Catalysis, 1972, 2, 981. J. Chem. Phys., 1962,37,2094. 796 ; J . Novakova, P. Jiru and V. Zavadil, J. Catalysis, 1970,17, 93. Trifiro and M. Carbucicchio, J. Catalysis, 1977, 47, 122. (PAPER 8 1462)