J. MATER. CHEM., 1994, 4( l),47-50 Solid-state Reaction between Molybdena and Alumina: Effect of Water Vapour Pressure on the Dispersion and Nature of the Supported Phases Margarita del Arco: Silvia R. G. Carrazan; Vicente Rives,*a Francisco-Javier Gil-Llambiasb and Pilar Malet" a Departamento de Quimica Inorganica, Universidad de Salamanca, Facultad de Farmacia, 37007-Salamanca, Spain Departamento de Quimica, Universidad de Santiago de Chile, Chile " Departamento de Quimica Inorganica, Universidad de Sevilla, Facultad de Quimica, Sevilla, Spain The dispersion and nature of surface species formed upon calcination of Mo03-y-Al,0, mixtures at 770 K for 10 h and with different molybdena loadings under different water vapour pressures [P(H,O)] have been studied by X-ray diffraction (XRD), zero-point charge (ZPC) measurements and temperature-programmed reduction (TPR); P(H,O) was in the range 25-45 Torr with MOO, loadings up to 2 monolayers (1 monolayer=0.1681 g MoO,/g A1203). The intensity of the most intense XRD peak of MOO, at 326 pm decreased as P(H,O) increased, indicating an increase in the concentration of surface species different from bulk MOO,.Up to 0.7 monolayer MOO, loading, covering of the Al,03 surface with Mo-containing species increases with increasing P(H,O); ZPC at P(H20)=45 Torr was coincident with that of MOO,. P(H20)has no effect on the dispersion of high loaded samples (above 1 monolayer), where all the support surface is covered even for P(H,O) =25 Torr. TPR results indicate the presence of species with different reducibilities, depending on P(H,O) and MOO, loading; their nature has been assessed by comparison with previous results for samples obtained by impregnation; bulk MOO, (responsible for high-temperature reduction peaks) formation has been observed at high MOO, loadings. Catalysts containing molybdenum supported on y-A1203 are widely used in petroleum chemistry for hydrodesulfurization, hydrodenitrogenation and hydrodeoxygenation rea~ti0ns.l~~ The formation of molybdena monolayers or multilayers in Mo03/y-A1203 samples obtained by different impregnation techniques has been widely ~tudied;~?~ however, information on the distribution of molybdena on the alumina surface of samples prepared by mechanical mixing, especially when these samples are calcined in the presence or in the absence of water vapour, is scarce in the literature.With regard to catalysts prepared by mechanical mixing, Knozinger and co-workersc8 have reported that spreading and chemical transformation of the active oxide (MOO,) over the support (e.g. A1203,TiO,) are independent processes. By using ion scattering spectroscopy (ISS), they have demon- strated that the spreading process clearly occurs during ther- mal treatment both in the presence and in the absence of water vapour. Under the experimental conditions used by these authors (720 K and a water vapour pressure of 24 Torr), the spreading process seems to be complete on A120, as well as on TiO, in less than 5 or 10h, respectively. Similar results with respect to molybdena spreading have been reported by us9 from quantitative X-ray diffraction measurements, but electrophoretic measurements have shown that a calcination time of 8-10 h is not enough to achieve a constant percentage of dispersion on the support surface, either in the presence or in the absence of water vapour.On the other hand, by using laser Raman spectroscopy (LRS),Knozinger and co-workers' have shown that a surface transformation of MOO, into polymolybdate takes place only in the presence of water vapour, the rate of this process being strongly dependent on the water vapour pressure. Complete transformation of molybdena into polymolybdate species takes ca. 30 h. Similar results have been obtained by us9p12 in samples prepared by mechanical mixtures by using LRS and TPR techniques. In this paper we report additional results on the spreading process in samples prepared by mechanical mixtures and calcined in the presence of an O2flow saturated with different water vapour pressures (25, 35 and 45 Torr). The structures of the metal oxide overlayers formed by molybdenum on the alumina surface are analysed, taking into account the TPR and the ZPC results for the samples.Experimental An Aluminoxid C y-alumina from Degussa (batch RV005) with a BET specific surface area of 105 m2g-' was used as support, and was calcined overnight at 770K to eliminate adsorbed organic impurities. MOO, was prepared by decomposition of (NH4),Mo,0,4-4H,0 (AHM, from Carlo Erba) at 770 K in air for 5 h.The supported Mo samples were prepared by manually grinding the support and the amount of MOO, required for loading of 0.4,0.7,1 and 2 monolayer. The monolayer capacity (16.81 g Mo0,/100 y-A120,) was calculated from the specific surface area of the support (105 m2 g-') and the area covered by a 'molecule' of MOO,, 15 x lo4pm2,13 that is, the so-called 'geometrical monolayer'.14 These mixtures were hand-ground in an agate mortar for 20 min and then calcined at 770 K for 10h in a quartz reactor under an oxygen flow of 30cm3 min-(from Sociedad Castellana de Oxigeno, S.C.O., Spain, 99.95%, passed through a Superpure gas Purifier model H, from Alltech, to remove impurity traces) saturated with water vapour at 25,35 and 45 Torr.Chemical analysis of the calcined samples indicated that the calcination treatment led to no change in Mo content through sublimation. Naming of the samples, according to the molybdena content and the water vapour pressure used, are summarized in Table 1. Samples are designated as nMA-p, where M =MOO,, A =A1203, n =MOO, content (number of monolayers) and p =water vapour pressure (Torr). XRD patterns were recorded on a Siemens 11-500 diffractometer with graphite-filtered Cu-Ka, radiation ( 154.05 pm) and interfaced to a DACO-MP data-acquisition micro- processor provided with Diffract/AT software. Zeta potential Table 1 Dispersion percentage, isoelectric point (IEP), zero-point charge (ZPC) data and hydrogen consumption during temperature programmed reduction (TPR) of the samples studied sample initial" MOO, lost IEP ZPC H,/Moc -A1Z03 8.8 -MOO, 2.9 -3.10 0.4MA-25 6.72 67 4.5 -7.0 2.64 5.8 -6.6 -0.4MA-35 6.72 87 0.4MA-45 6.72 90 6.0 -6.0 2.29 0.7MA-25 11.76 47 5.5 -6.5 2.52 0.7MA-35 11.76 61 7.2 -3.9 -0.7MA-45 11.76 83 9.8 -3.1 2.37 1MA-25 16.81 64 10.8 -3.1 2.76 1MA-35 16.81 68 11.4 -3.1 -1 MA-45 16.81 74 12.5 -3.1 2.43 2MA-25 33.62 37 12.5 -3.1 2.65 2MA-35 33.62 37 12.5 -3.1 -2MA-45 33.62 37 12.5 -3.1 2.53 "Initial mass of MOO, (g/100 g y-Al,O,); bDispersion (ratio between mass of MOO, lost and initial mass of MOO,); 'Molar ratio between H, uptake and Mo content.(or Zero-point charge, ZPC) measurements were carried out in a Zeta-meter Tnc. instrument (Model ZM-77) using 200 mg of samples (average size ca.2 pm) dispersed in 200 ml of lo-, moll-' KC1 solutions. The pH was adjusted with either mol 1-' KOH or HCl solutions. The zeta potentials were obtained from electrophoretic migration rates using the Smoluchowski equation." TPR diagrams were recorded in a conventional apparatus with a catharometric detector, using a (S%)H,-Ar mixture (from S.C.O.) as carrier gas, with a flow of 50ml min-' and at heating rate of 10 K min-'. Good resolution of the different reduction steps under these experimental conditions was ensured by using sample weights containing ca. 100pmol of MoO,.I6 Results and Discussion The degree of dispersion of MOO, on the A1203 support has been calculated from XRD measurements, following the method previously reported in the literat~re.".'~ Table 1 lists the values of the dispersion percentage and 'lost MOO,' (hereafter LM, corresponding to MOO, species widely dis- persed on the surface and so undetectable by XRD as MOO, crystallites) after calcination of the samples for 10 h in the presence of different water vapour pressures (25, 35 and 45 Torr).The dispersion percentage steadily increases as the water vapour pressure does, indicating a continuous 'loss of MOO,'; such an increase is much more evident in samples with molybdena loadings of 0.4 and 0.7 monolayers. In samples with a higher MOO, content (1 monolayer), the amount of LM shows a slight increase with the water vapour pressure, while in samples 2MA it remains constant as the water vapour pressure changes.It is worth noting that, while for samples 0.4MA and 0.7MA, the amount of LM sharply increases with the water vapour pressure up to a value close to the initial content of MOO, existing in the sample, LM for samples 1MA only show a slight increase with the water vapour pressure, reaching a maximum value of 12.0 g MoO,/lOO g y-Al,O,. This value of LM remains constant for samples 2MA whichever the water vapour pressure, which indicates that the maximum dispersion capacity for these samples has been attained under these experimental conditions. It is also interesting to compare the value of LM for J. MATER. CHEM., 1994, VOL.4 different molybdenum loadings at a given water vapour pressure (Fig. 1). A slight increase in the amount of LM is observed on increasing the MOO, content from 0.4 to 0.7 monolayers for p=25 Torr. When the molybdena loading is increased from 0.7 to 1 monolayer, this value is twice as high, and still slightly increases from 1 to 2 monolayers. A similar behaviour is observed for samples treated at 35 Torr. However, samples prepared under a water vapour pressure of 45 Torr show a different behaviour: a steady increase in LM is observed from 0.4 to 1 monolayer, while the value remains constant from sample 1MA to sample 2MA. Note the high value for LM, even at 25 Torr, for sample 0.4MA; this fact must be attributed to the low MOO, content in these samples (6.72g MoOJ100g y-Al,O,), below the value corresponding to the upmost dispersion on Al,O, (12 g MoO,/lOO g y-A1203).10,17*18On the other hand, if dispersion data for samples containing 0.7 monolayer are compared with those found in MoO,/Al,O, samples prepared by mechanical mixtures and calcined in a static, uncontrolled atmosphere in an open crucible during 24 h,9 it is apparent that a calcination time of 10 h is not enough to attain the dispersion correspond- ing to the geometrical monolayer capacity even for samples treated at the highest water vapour pressure (45 Torr).Electrophoretic migration and TPR measurements have been used for a better understanding of the relationships existing between the dispersion of MOO, on Al,O, and the water vapour pressure.ZPC values for all Mo03/A1,03 samples as well as the isoelectric points (IEP) for MOO, and y-A1203 are included in Table 1; changes in the ZPC values as a function of the water vapour pressure and MOO, loading have been plotted in Fig. 2. The ZPC values for MA-25 samples (Fig. 2) show a steady decrease from the sample containing 0.4 monolayer to sample 0.7MA-25,while a further increase in the MOO, content produces a sharp decrease in the ZPC to 3.1, a value that remains constant for molybdena loadings ranging from 1 to 2 monolayers. In contrast to samples MA-25, those prepared at p=35 or 45 Torr show a sharp decrease in ZPC when the Mo loading increases from 0.4 to 0.7 monolayers.It is interesting to note that at the different water vapour pressures used in this work, values of 3.1 for the ZPC are achieved at Mo loadings of 0.7 or 1 monolayer, and always remain constant when the Mo content is further increased. Taking into account that the ZPC variation in all the samples lies between 7 and 3.1, that the IEP of unloaded y-Al,03 is 8.8, and assuming that bulk and supported MOO, 14 r c. -0 Fig. 1 Dependence of the amount of 'lost MOO,' (g/100 g y-Al,O,) for samples with different molybdena loading with the water vapour pressure during calcination: (0)25 Torr; (0)35 Torr; (V)45 Torr J. MATER. CHEM., 1994, VOL. 4 0.0 0.5 1.0 1.5 2.0 2.5 Mo content (monolayers) Fig.2 Change in ZPC of the samples with different molybdena loading calcined under different water vapour pressures: (0)25 Torr; (0)35 Torr; (V)45 Torr exhibit the same JEP (3.0), the decrease of the ZPC observed for the three series of samples should give a measure of the increasing coverage of the surface of alumina by molybdena.The behaviour observed in these samples is in agreement with the equation which relates the ZPC of a support-supported phase system with the isoelectric points of the pure support and the supported species.lg For samples 0.4-MA, the ZPC values show (Fig. 3) a steady decrease as the water vapour pressure increases, and the value calculated when the sample is calcined at 45 Torr for 10 h coincides with that found in samples with the same Mo loading, but calcined in an uncontrolled atmosphere for 8 h.' This dropping in the ZPC values can be interpreted as follows: the increase in the water vapour pressure favours the continuous formation of surface Mo species which are expected to have a value of IEP <<8.8, as it was evidenced in MoO3/A1,O3 samples with an Mo loading of 0.4 monolayer prepared by mechanical mixtures and calcined for 24 h in an uncontrolled atmosphere.In these samples, only polymolybdate species were detected by LRS.'O-'2 These results are in agreement with those found by XRD, which indicate a slight and continuous 'loss of bulk MOO,' as the water vapour pressure increases. A decrease in the ZPC values is also observed in 0.7MA samples when water vapour pressure is increased, although in 20 30 40 50 water pressure/mmHg Fig.3 Change in ZPC of the samples prepared under different water vapour pressures as a function of the molybdena loading: (0)0.4 monolayers; (e)0.7 monolayers; (V)1 and 2 monolayers this case the variations are sharper than in the case of 0.4MA samples, specially in the 25-35 Torr range. A value of 3.1 for sample 0.7MA-45 indicates that the y-Al,O, surface is com- pletely covered by surface Mo species, the surface concen- tration of which is larger than that of bulk MOO,, as it was put into evidence by the LM values calculated by XRD. The ZPC values for samples 1MA and 2MA remain con- stant for all these samples whatever the water vapour pressure used to prepare them (ZPC =3.1 ).This value of ZPC indicates that the y-A1203 surface is totally covered by surface oxo- molybdenum species and bulk MOO,, but the relative concen- tration of each species cannot be determined from the ZPC results because of the constancy in ZPC for all these samples. However, the differences in the amount of these species (amorphous MOO, and bulk MOO,) in 1MA samples with water vapour pressure can be put into evidence by XRD; the amount of LM increases with increasing water vapour pressure. The nature and relative amounts of dispersed species have been assessed by TPR, since it has been shown that the degree of dispersion changes the reducibility of supported molyb- dena.9 On the other hand, a knowledge of such a reducibility is important since when acting as catalysts in selective oxi- dation processes these systems usually undergo oxidation/ reduction reactions. The TPR profiles of the samples are shown in Fig.4, and the molar H, :Mo ratios, as calculated from the Mo content and the integrated hydrogen consumptions, are included in Table 1. Reduction of bulk molybdena under these experimen- tal conditions starts at 825 K9 with a total hydrogen uptake of 3.1 H,/Mo, thus corresponding to the total reduction of Mo6+ ions to Moo. In all Mo0,/A120, samples studied in this work, reduction starts at lower temperatures (ca. 673 K), mb0 n I I I J 600 800 1000 1200 TIK Fig. 4 Temperature programmed reduction profiles calcined under water vapour pressures of 25 Torr (solid line) or 45 Torr (dotted line): (a)2MA, (b)1 MA, (c) 0.7 MA, (d) 0.4 MA J.MATER. CHEM., 1994, VOL. 4 thus suggesting the presence of well dispersed oxomolyb- denum-containing species, more easily reducible than bulk molybdena. Hydrogen consumptions (2.5 f0.2) are always slightly lower than those expected for a total reduction from Mo6+ to Moo, although it should be noted that the TPR profiles do not recover the baseline at the highest temperature attainable by our experimental system, and so this low consumption can be tentatively ascribed to the existence of a small number of species only reducible at a very high temperature. A very simple TPR profile, with a low-temperature maxi- mum at 750K and a broad high-temperature maximum centred at 1150 K is recorded for sample 0.4MA-45, where a high dispersion degree of supported molybdena was concluded from XRD data.This profile is almost identical to that recorded’ for a 0.7 monolayer Mo03/A1,0, sample calcined at 770 K for 24 h in an uncontrolled atmosphere, where only polymolybdate species were detected by Raman The TPR maximum at 750 K is recorded in the profiles of all MoO3/A1,O3 samples studied here, and its intensity can be taken as a measurement of the amount of surface poly- molybdate species existing in the samples. Thus, when the 0.4 monolayer sample was pretreated under a water vapour pressure of 45 Torr, maxima at 750 and 1150 K have similar intensities to those recorded in the 0.4MA-25 sample.The lower dispersion in this sample, as detected by XRD, gives rise to a new maximum at ca. 1050 K in the TPR profile. This new maximum should be assigned to crystalline MOO,, detect- able by X-ray diffraction. The small changes in the intensity of the maximum at 750 K associated to changes in the water vapour pressure during calcination suggest that surface poly- molybdates are readily formed in samples containing 0.4 monolayer, even at the lowest water vapour pressures used furing pretreatment, in agreement with the small changes above reported in the ZPC values for these samples. On the contrary, the changes in the TPR profile of samples containing 0.7 monolayers can be readily related to changes in the water vapour pressure during calcination.Thus, the intensity of the 750 K maximum in the 45 Torr sample is at least twice than that for the 25 Torr sample. In addition, two maxima clearly recorded at ca. 920 and 1073 K in the TPR profile of the 0.7MA-25 sample vanish when the sample has been calcined under a water vapour pressure of 45 Torr, the latter maximum decreasing its intensity and shifting to 1050 K. The maximum at 920K has been previously ascribed’ to dispersed MOO,, while, in agreement with XRD data (which indicate 53% of crystalline MOO, in this sample) the broad maximum at 1073 K should be ascribed to crystalline MOO,. These data clearly suggest that crystalline MOO,, still present in the 0.7 monolayer samples calcined under a water vapour pressure of 25 Torr, is dispersed and forms surface polymolyb- dates when the samples are calcined under higher water vapour pressures.At the highest molybdena contents (1-2 monolayers), the intensity of the TPR maximum at 750K remains constant and almost independent of the water vapour pressure, thus confirming that in these samples the alumina surface is saturated of polymolybdates even when they have been cal- cined under the lowest water vapour pressure. Maxima at 1073 and 920 K indicate the coexistence of these polymolyb- dates with crystalline and dispersed MOO,, respectively. Small changes in the profiles of the 1 monolayer samples suggest that a small amount of surface polymolybdate is still formed at the expense of dispersed Moo3 when the water vapour pressure changes from 25 to 45 Torr, while the profiles of the 2 monolayer samples remain virtually unchanged at the different water vapour pressures.Authors are grateful for financial support from DGICYT (PB91-425) and Junta de Castilla y Leon (C. Cultura y Turismo, ref. 16/02/9 1). References 1 P. Grange, Catal. Rev. Sci. Eng.. 1980,21, 135. 2 F. E. Massoth and G. Muralidhar, in Proc. 4th Int. Conf. Chem. Uses Molybdenum, ed. H. F. Barry and P. C. H. Mitchell, Climax Molybdenum Co., Ann Arbor, MI, 1982, p. 343. 3 R. Thomas, E. M. Van Dors, V. H. J. de Beer, J. Medema and J. Moulijn, J. Catal., 1982,76,241. 4 J. Sonnemaus and P. Mars, J. Catal., 1973,31.209. 5 F. J. Gil Llambias, A. M. Escudey, A. Lopez Agudo and J.L. Garcia Fierro, J. Catal., 1984,90,323. 6 J. Leyrer, M. J. Zaki and H. Knozinger, J. Phys. Chem., 1986, 90,4775. 7 M. Margraf, J. Leyrer, H. Knozinger and E. Taglauer, Surf. Sci., 1987,189/190,842. 8 J. Leyrer, R. Margraf, E. Taglauer and H. Knozinger, Surf. Sci., 1988,201,603. 9 M. del Arco, S. R. G. Carrazan, V. Rives, F. J. Gil Llambias and P. Malet, J. Catal., 1993, 141,48. 10 M. del Arco, S. R. G. Carrazan, V. Rives and J. V. Garcia Ramos, Muter. Chem. Phys., 1992,31,205. 11 M. del Arco, S. R. G. Carrazan, V. Rives and J. V. Garcia Ramos, Spectrosc. Lett., 1992,25, 73. 12 M. del Arco, S. R. G. Carrazan, V. Rives and J. V. Garcia Ramos, J. Muter. Sci., 1992,27, 5921. 13 T. Fransen, P. C. Van Berge and P. Mars, Stud. Sur- Sci. Catal., 1976, 1,405. 14 G. C. Bond, S. Flamerz and R. Shukri, Furuday Discuss. Chem. SOC.,1989,87,65. 15 R. J. Hunter, Zeta Potential and Colloid Science Principles and Applications, Academic Press, New York, 1981, p. 72. 16 P. Malet and A. Caballero, J. Chem. SOC.,Faraday Trans. I, 1988, 84,2369. 17 Y. Xie, L. Gui, Y. Liu, B. Zhao, N. Yang, Y. Zhang, Q. Guo, L. Duan, H. Huang, X. Cai and Y. Tang, in Proc. 8th Znt. Congr. Catal., Berlin, Germany, Dechema, Frankfurt, 1984, vol. 5, p. 147. 18 P. Dufresne, E. Payen, J. Grimblot and J. P. Bonnelle, J. Phys. Chem., 1981,85,2344. 19 F. J. Gil-Llambias and A. M. Escudey-Castro, J. Chem. SOC., Chem. Commun., 1982,478. Paper 3/04258H; Receizied 20th July, 1993