- Email: [email protected]
Molybdophosphoric Acid Adsorption on Titania from Ethanol–Water Solutions
JOURNAL OF COLLOID AND INTERFACE SCIENCE ARTICLE NO.
204, 256–267 (1998)
CS985552
Molybdophosphoric Acid Adsorption on Titania from Ethanol–Water Solutions AnalıB a Concello´n, Patricia Va´zquez, Mirta Blanco, and Carmen Ca´ceres 1 Centro de Investigacio´n y Desarrollo en Procesos CatalıB ticos (CINDECA), UNLP, CONICET, 47 No. 257, 1900, La Plata, Argentina Received December 15, 1997; accepted March 26, 1998
The equilibrium adsorption at 207C of molybdophosphoric acid solutions, using ethanol–water as solvent, on titania was studied. The molybdenum adsorption isotherm showed a sigmoidal shape; low values of molybdenum adsorbed were observed for final equilibrium concentrations lower than 50 mg Mo/ml, and for higher concentrations, the adsorbed molybdenum amount almost reached a plateau. From this isotherm it could be concluded that the solute–support interaction was not strong. UV–visible and NMR spectra of the solutions before and after the adsorption on titania showed that the species PMo12O 340– was present. This species also was observed by DRS in the wet samples and by NMR, FT-IR, and DRS, in the solid samples dried at room temperature and calcined at 255, 310, 365, and 4257C, showing that the thermal stability of molybdophosphoric acid adsorbed on titania is similar to that of the bulk acid. The impregnating solutions and the impregnated solid changed with the time to a bluish color as a consequence of the formation of heteropoly blues which presented Mo 6/ partially reduced to Mo 5/ . The XRD patterns indicated that the species adsorbed onto the support surface are highly dispersed like a noncrystalline form. q 1998 Academic Press Key Words: molybdophosphoric acid; titania; equilibrium adsorption impregnation.
1. INTRODUCTION
It is well known that the heteropolyacids (HPA) and their salts exhibit catalytic versatility arising from their combined acid and redox properties (1, 2). Also the use of heteropolyacids as catalysts for fine chemical synthesis processes is actually developing (3). However, a disadvantage of HPA as catalysts lies in their low stability, and various factors can cause it; for instance, in molybdophosphoric acid (MPA), problems appear with thermal decomposition, acidity changes (4), and loss of the active phase, the latter being due to formation of a volatile molybdic acid (5). HPA have also been supported using different carriers with the aim of improving their stability, which, in addition, 1 To whom correspondence should be addressed. E-mail: hds@nahuel. biol.unlp.edu.ar.
256
0021-9797/98 $25.00 Copyright q 1998 by Academic Press All rights of reproduction in any form reserved.
AID
JCIS 5552
/
6g46$$$521
depends on heteropolyacid concentration and nature of the solvent (6). Bartoli et al. (4) used supports such as SiC to take advantage of its high thermal conductivity and zirconium oxide and cerium oxide to prevent HPA over-reduction. The use of these supports improves the stability of the H4PMo11VO40 active phase in the oxydehydrogenation reaction of isobutyric acid. In the vapor phase synthesis of methyl tert-butyl ether, the activity of silica-supported MPA is higher than that of MPA supported on silica–alumina, alumina, and magnesium oxide (7). This is related to the acidic or basic characteristics of the supports. In particular, silica has a low-level interaction with the HPA, but reports indicate that MPA becomes stable up to 6007C when supported on SiO2 (8). Alkaline salts of HPA are also used as modifiers of the acidic and thermal properties of their corresponding heteropolyacids (9, 10). On the other hand, with the objective to obtain new generations of hydrodesulfurization (HDS) and hydrodenitrogenation (HDN) catalysts, our research group has begun studies about the adsorption on alumina of molybdenum and phosphorus integrating the same structure by using aqueous solutions of pentamolybdodiphosphate anion (11). In another paper (12) the impregnation of alumina was studied with solutions of other HPA, that present primary structure of Keggin type, with general formula [XM12O40 ] ( 80 n ) 0 (13). X is called heteroatom or central atom, usually P, Si, or As; M are called addenda atoms, usually Mo, W, V, or Nb; and n is the valence of X. As it is known, this structure consists of a central XO4 tetrahedron surrounded by twelve MO6 octahedra. Also, it was proved that the use of organic solvents (ethanol–water, dimethylformamide) stabilizes the primary Keggin structure. There have been attempts to optimize the performance of traditional hydrotreatment catalysts as alumina-supported CoMo, NiMo, or NiW. To this end, the use of other supports, notably carbon, titania, and zirconium oxide, was tested. Among these supports, titania has attracted attention in view of the higher HDS activities displayed by molybdenum catalysts supported on this oxide. Ng and Gulari (14) have found that low-loading titania-supported catalysts are more active, compared to alumina-supported catalysts, in the HDS
07-21-98 10:42:23
coida
MOLYBDOPHOSPHORIC ACID ADSORPTION ON TITANIA
reaction carried out at atmospheric pressure. Luck (15) reported that the specific HDS rate constant of a Co–Mo/TiO2 catalyst is greater than Co–Mo catalysts supported on other carriers, as Al2O3 , CeO2 , SiO2 , attapulgite, and carbon. It was also found that titania-based catalysts showed a high activity without presulfurization (16). This feature is one of the advantages of titania-based catalysts in commercial processes. Damyanova and Fierro (17), using the Co (Ni) salts of HPA supported on titania samples, studied the effect of the counterion on the surface and catalytic properties of these solids in the thiophene hydrodesulfurization reaction. They suggested that the highest activity of nickel-promoted catalysts may be caused by an increase in dispersion of molybdenum and its lighter reduction. The catalytic properties, and hence the efficiency, of supported heteropolyacid-based catalysts obtained by equilibrium adsorption impregnation techniques depend on the nature and subsequent stability of the species during the impregnation, drying, and calcination steps, so appropriate conditions must be found to control such species. An important factor is the adsorption behavior of the precursors of active species during the impregnation stage, i.e., the contact between the support and the solution. Thus, as a first study toward the obtainment of MPA on TiO2 catalysts, one of the objectives of this paper is the study of MPA adsorption from ethanol–water solutions using titania as support. Besides, it is presented a detailed analysis of the nature of the species present in the impregnating solutions as well as the characterization of the present species in the solids obtained by equilibrium adsorption impregnation. 2. EXPERIMENTAL
Adsorption studies. To obtain each experimental point of the molybdenum adsorption isotherm at a constant temperature of 207C, 1 g of support (TiO2 anatase, Riedel de Ha¨en, S BET : 9.8 m 2 /g), and 4 ml of MPA(H3PMo12O40r xH2O, MERCK p.a.) solution of known concentration were shaken, for a given time. Solvent employed was an 1:1 (in volume) mixture of water and ethanol Soria 96%. The pH of the impregnating solutions was that resulting of the MPA dissolution in the solvent. The solid was separated of the solution by centrifugation, dried at room temperature, and then calcined at 255, 310, 365, and 4257C. The adsorption isotherm was obtained using different MPA contents, the initial concentration range being 4–110 mg Mo/ml. However, to determine the optimum contact time required to achieve equilibrium between the support and the impregnating solution, adsorption kinetic measures were previously performed, but for only one initial concentration of MPA (110 mg Mo/ml), the contact time being varied from 5 min to 72 h. The amount of molybdenum adsorbed on the support,
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
257
C(ads) (g Mo/100 g TiO2 ), is calculated through a mass balance on the basis of the difference between the initial solute concentration in the solution, C(aq)i (mg Mo/ml), and the equilibrium or final solute concentration in the solution, C(aq) f (mg Mo/ml), and assuming negligible variation in the solution volume. The following expression was used: 01 C(ads) Å [C(aq)i 0 C(aq) f ]Ms r 01 M 01 , a 10
where Ms is the solution mass (g), r is the solution density (g/ml), and Ma , support mass (g). Total molybdenum concentration in the catalysts, CT (g Mo/100 g TiO2 ), was also calculated as CT Å C(ads) / C(occl). The occluded concentration, C(occl) (g Mo/100 g TiO2 ), was calculated taking into account the solution volume retained in the pores (Vr Å 0.74 ml) and C(aq) f , as 01 C(occl) Å C(aq)fVr M 01 . a 10
Quantitative molybdenum analysis. The molybdenum concentration in the solutions, both before and after being in contact with titania, was measured by atomic absorption spectrometry with an IL model 457 double-beam spectrophotometer (Instrumentation Laboratory Inc.). The analysis was carried out at the following conditions: wavelength, 313 nm; band width, 0.5 nm; lamp current, 7 mA; phototube amplification, 700 V; burner height, 7 mm; and acetylene/ air flame in 7:14 ratio. pH measurements. As standards for pH measurements, Tritisol buffer solutions (Merck) of pH 7.00 { 0.02 (phosphate) and of pH 2.00 { 0.02 (citrate-HCl) were employed. An Instrumentalia PH-MD digital equipment connected to an Orion Research model 91-03 pH-sensitive combined semielectrode was used. UV–visible spectroscopy. Spectra of solutions were obtained with a Varian Super Scan 3 double-beam UV–visible spectrophotometer with a built-in recorder, using quartz cells of 0.5 mm optical path, in the range of 200–600 nm. MPA solutions of different concentrations were analyzed before and after being in contact during different times with titania. Nuclear magnetic resonance. The MPA solutions, both before and after being in contact with titania, were analyzed by 31P NMR. A Bruker MSL-300 equipment linked to a ‘‘SOLIDCYC.DC’’ pulse program was utilized, using 5 ms pulses, a repetition time of 10 s, and a frecuency of 121.496 MHz for 31P at 247C, being the resolution of 3.052 Hz, per point. Phosphoric acid 85% was employed as external reference. The same equipment and similar conditions were used to obtain 31P MAS-NMR spectra of the impregnated solids
coida
´ N ET AL. CONCELLO
258
TABLE 1 Molybdenum Concentrations in Solution and in Solid for the Studied Samples C(aq)i C(aq)f C(ads) CT Sample (mg Mo/ml) (mg Mo/ml) (g Mo/100 g TiO2) (g Mo/100 g TiO2) M1 M3 M4 M5 M7 M8 M10 M11 M12
3.71 21.69 27.05 45.36 57.20 63.11 79.39 90.16 110.0
3.10 19.72 26.45 41.91 49.31 52.27 69.28 77.92 94.0
0.36 0.72 0.29 1.26 2.88 3.94 3.72 4.89 5.24
0.59 2.06 2.10 3.87 5.87 7.18 11.57 10.38 17.49
dried at room temperature and calcined at 255, 310, 365, and 4257C. A sample holder of 5 mm diameter and 10 mm in height was used, the spin rate being 2.1 kHz. The repetition time was 3 s, and several hundred pulse responses were collected. Fourier transform infrared spectroscopy. A Bruker IFS 66 instrument, pellets in KBr and a measuring range of 400– 1500 cm01 were used to obtain the FT-IR spectra of titania, bulk molybdophosphoric acid, and MPA on titania samples dried at room temperature and calcined at 255, 310, 365, and 4257C. Diffuse reflectance spectroscopy. The samples were studied with an UV–visible Varian Super Scan 3 equipment, fitted with a diffuse reflectance chamber with a BaSO4 inner surface. DRS spectra of titania, bulk molybdophosphoric acid, and wet, dried at room temperature, and calcined MPA on titania samples, in the range of 200–600 nm, were obtained. Each sample was compacted in a Teflon sample holder, which was then covered with a quartz circular window to obtain a sample thickness of 2 mm.
FIG. 2. 31P NMR spectra of MPA impregnating solutions, corresponding to M12 sample, before (a) and after (b) the adsorption on titania.
X-ray diffraction. The XRD patterns of titania and calcined samples were obtained by using a Phillips PW-1714 diffractometer. In this measurements, the following condi˚ ), nickel filter, tions were used: CuKa radiation (1.5417 A 30 mA and 40 kV in the high voltage source, scanning angle (2u ) from 5 to 607, and scanning rate of 17 per minute. 3. RESULTS
3.1. Adsorption Kinetics and Equilibrium Isotherms
For the contact between TiO2 and MPA solutions in ethanol–water, the adsorption kinetics was measured. To this end, an initial MPA concentration of 110 mg Mo/ml was used and measurements were carried out at contact times from 5 min to 72 h. The results showed that C(ads) remains constant from 10 min on. For this adsorption kinetics, a small decrease in pH was observed for the final solutions in relation to the initial one (pH 1). The pH values of the final solutions were between 0.8 and 1. Thus, unimportant variations of pH in the impregnating solution were observed, allowing to asume that the depolymerization of PMo12O 340– to PMo11O 739– according to the reaction [1] would not be significant: / PMo12O 340– / 3H2O r PMo11O 739– / MoO 20 4 / 6H .
FIG. 1. Mo adsorption isotherm on titania from MPA solutions in ethanol–water at 207C.
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
[1]
The hydroxyl groups of the titania surface in solution tends to be either positively or negatively charged below or
coida
MOLYBDOPHOSPHORIC ACID ADSORPTION ON TITANIA
259
FIG. 3. UV–visible spectra of 110 mg Mo/ml initial impregnating solution at different times.
above the IEP of titania (anatase: 6.2) (18), according to the following equilibrium: Ti–OH 2/ B Ti–OH B Ti–O 0 . Above the IEP of titania, the hydroxyl groups tend to be negatively charged, so that it is relatively hard for the anions to be adsorbed on the titania surface, due to the electrostatic repulsion. On the other hand, below the IEP of titania the adsorbed amount of the molydophosphate anion should be favored. Hence, in our working conditions, MPA adsorption on titania is favored by the low pH value of the initial solutions, resulting of the MPA dissolution in the solvent, and of the solutions resulting from the contact with the support. For the experiments that allowed us to obtain the adsorption isotherm, to ensure that the adsorption–desorption equilibrium had been reached, a contact time of 72 h for each point was selected. The C(aq)i , C(aq) f , C(ads), and CT values are summarized in Table 1. The molybdenum adsorption isotherm from solutions of H3PMo12O40 on titania at 207C, expressed as C(ads) in function of C(aq) f , is shown in Fig. 1. The isotherm has a sigmoidal shape; it presents very low initial values of C(ads) and later, for MPA concentrations above 50 mg Mo/ml in the final solutions, it tends to reach a plateau of C(ads) values. The initial slope shows that
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
adsorption becomes easier as concentration in solution increases. Hence, at low concentration, it is possible that the solute molecules meet strong competition, for surface support sites, from molecules of the solvent (19). At high concentration, the presence of one plateau indicates the adsorption of only one molybdenum–phosphorus species, possibly the molybdophosphate anion; this will nevertheless be confirmed in the characterization of solutions and solids. 3.2. Characterization of the Solutions
3.2.1. Nuclear Magnetic Resonance The 31P chemical shift ( d ) provides important information concerning the structure, composition and electronic states of heteropoly compounds. The chemical shift for 31P in aqueous solution at 03.9 ppm for PMo12O 340– (20) was correlated with the P{Oa bond strength [ n(P{Oa )] (21). The 31P NMR spectra of MPA solutions in ethanol–water, corresponding to M12 sample, both before and after the adsorption on titania are shown in Fig. 2. The chemical shifts observed in the spectra of initial and final solutions, 03.6 and 03.35, respectively, show that PMo12O 340– is neither de-
coida
´ N ET AL. CONCELLO
260
FIG. 4. UV–visible spectra of impregnating solutions before (a) and after (b) the adsorption on titania, for different samples and an impregnation time of 72 h.
polymerized nor degraded as a consequence of the contact with titania. Small solvent effect on 31P chemical shift of the heteropolyanion in ethanol–water with respect to that in water indicates that the central atom is protected by the four groups of three edge-shared MoO6 octahedra (M3O13 triad). 3.2.2. UV–Visible Spectroscopy For an MPA solution in ethanol–water of 110 mg Mo/ ml initial concentration, Fig. 3 shows the UV–visible spectra for different times, from 5 min to 72 h. In the range under study, an absorption band with a maximum at 310 nm belonging to the oxygen–metal charge transfer is observed, suggesting the presence of only one species, the intact Keggin phase. Likewise, spectra showing the same band were observed for the impregnating solutions analyzed after contact with the support. Papaconstantinou (22) showed that non-reduced HPA are characterized by oxygen-to-metal charge transfer bands in the near visible and UV regions and no absorption in the visible (all metals are Mo VI or W VI with d 0 configuration). Reduction
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
of HPA proceeds without substantial change of their structure with addition of a certain characteristic number of electrons. On reduction they are colored mainly blue, producing the socalled heteropoly blues (HPB) which have a broad absorption band at about 700 nm. Because the HPB contain metal ions in different oxidation states they are usually assigned to the category of mixed valence compounds. That band is attributed to metal-to-metal charge transfer (Mo 5/ r Mo 6/ ) which is responsible for the blue color of the compounds, and to d–d transitions of the d 1 metal ions (13). The equipment used records spectra up to a maximum of 600 nm so that the region around 700 nm is not discussed. Notwithstanding, color variations were observed with time both in the original and in the final solutions; they are initially yellow and become greenish as a consequence of HPB formation. Spectra of initial and final solutions are presented in Fig. 4a and b, respectively, for only one contact time with the support (72 h) and different concentrations of the impregnating solution. Both in original solutions and in final ones, the above-mentioned value of the absorption maximum (310 nm) holds, though the absorption intensity varies depending upon the concentration change. Hence, the technique allows
coida
MOLYBDOPHOSPHORIC ACID ADSORPTION ON TITANIA
261
FIG. 4—Continued
to observe that the H3PMo12O 40 is neither depolymerized nor degraded in the solution during the equilibrium adsorption, in agreement with NMR results and with the low pH presented by the impregnating solutions before and after contact with titania (pH 0.8–1). 3.3. Characterization of the Solids
3.3.1. Nuclear Magnetic Resonance Figure 5 shows 31P MAS-NMR spectra recorded for bulk molybdophosphoric acid and MPA-impregnated titania samples, dried at room temperature and calcined at 255, 310, 365, and 4257C. According to several authors, the chemical shift ( d ) of the bulk acid is found between 02.9 and 04.8 ppm (20, 23, 24). The spectrum obtained (Fig. 5a) exhibits two lines, one at 03.41 and another at 03.83 ppm, which can be assigned to the acid in different hydration levels (20). The spectra showed by MPA on titania samples M5 and M12, dried at room temperature, present only one line with maxima at 04.1 ppm (Fig. 5b) and 03.51 ppm (Fig. 5c), respectively. This suggests that, in the concentration range
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
under study, the structure of the supported molybdophosphate anion keeps intact. When the M12 sample is subsequently calcined at 255, 310, 365, and 4257C, the corresponding spectra also show only one line, which is observed at 04.1, 04.0, 04.37 and 04.1 ppm, respectively (Fig. 5d– g). The shifts shown by calcined samples of MPA on TiO2 in relation to the dried sample are 00.86 ppm at most and 00.54 ppm compared to the bulk acid lines. All observed chemical shifts can be assigned to the acid with undegraded Keggin structure. The slight shift toward higher field can be caused by an increase of the interaction of the acid with the support during calcination, as it was the case for another heteropolyacid ( H3PW12O40 ) on carbon ( 25 ) . 3.3.2. Fourier Transform Infrared Spectroscopy Each M3O13 triad in the Keggin structure has a common oxygen vertex connected to the central heteroatom X{Oa (M3O13 ). The other three classes of symmetric-equivalent oxygens in the a isomer of Keggin structure are (20) M{Ob{M, connecting two M3O13 units by corner sharing; M{Oc{ M, connecting two M3O13 units by edge sharing;
coida
´ N ET AL. CONCELLO
262
FIG. 5. 31P MAS-NMR spectra of bulk MPA (a), M5 sample dried at room temperature (b), M12 sample dried at room temperature (c), and M12 sample calcined at 255 (d), 310 (e), 365 (f ), and 4257C (g).
and terminal oxygens Od| M. The maxima of the stretching modes of vibration for IR absorption bands of MPA are P{Oa (1070 cm01 ), Mo|Od (965 cm01 ), Mo{Ob{Mo (870 cm01 ), and Mo{Oc{ Mo (790 cm01 ) (20). TiO2 spectra obtained for several calcination temperatures (310, 365, and 4257C) are presented in Fig. 6a, while the spectrum of the noncalcined bulk MPA is shown in Fig. 6b. The graphs show that the titania absorption band extends over a wide region with maxima in the 400–900 cm01 zone that mask the MPA bands at 790 and 870 cm01 . Two characteristic bands of the anion under study (1070 and 965 cm01 ) are observed in the remaining zone of the spectra. An outstanding feature of these spectra (Fig. 6a) is the widening of the TiO2 band toward higher wavenumbers as the temperature increases, the change being more considerable from 365 to 4257C. The thermal stability of TiO2 support is usually lower than that of g-Al2O3 and the particular note is the reducibility of TiO2 , which results in the formation of various stoichiometric and nonstoichiometric oxides, tak-
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
ing Ti ions several valence states (16). Consequently, shifts can be ascribed to structural changes in view of the low TiO2 thermal stability. For the solids obtained by impregnating titania for 72 h with an ethanol–water solution of MPA (initial concentration: 110 mg Mo/ml), Fig. 6b also presents the FT-IR spectra of dried and calcined samples, treated at the same temperatures used before for the support. For samples dried at room temperature and calcined below 3657C, no modifications are observed in the 1070 and 975 cm01 bands. It can be seen that the intensity of such bands decrease beyond 3657C, becoming hardly perceptible at 4257C because it is masked by the widening of the support band or, else, it vanishes owing to the start of Keggin unit degradation. However, bearing in mind NMR results, the latter possibility may be discarded. Figure 7 presents the FT-IR spectra of samples obtained from MPA solutions of different concentration and dried at room temperature. These spectra show the characteristic
coida
MOLYBDOPHOSPHORIC ACID ADSORPTION ON TITANIA
263
FIG. 6. FT-IR spectra of titania calcined at different temperatures (a) and M12 sample dried at room temperature and calcined at different temperatures and bulk MPA (b).
bands of the heteropolyacid, the intensity of which increases with the concentration. FT-IR results are in agreement with those arrived at by NMR for samples prepared with different concentrations and subjected to various thermal treatments. 3.3.3. DRS The charge-transfer absorption spectra of most non-reduced polyanions appear in the 200–500 nm region and
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
consist of bands which may be ascribed to oxygen-to-metal transfers. The tetrahedral Mo exhibits two absorption bands at 220 and 260 nm, whereas Mo in octahedral coordination presents, apart from those bands, another band at higher wavelength. Most polymolybdates and polytungstates contain octahedrally coordinated Mo 6/ and W 6/ but with the metal atom displaced from the center of the octahedra toward a corner or an edge. The short M{O bonds produced by these displacements imply the existence of significant metal–oxy-
coida
´ N ET AL. CONCELLO
264
FIG. 7. FT-IR spectra of MPA supported on titania samples with different concentrations, dried at room temperature.
gen p-bonding, and this feature is the key to the stabilities of all polyoxoanion structures. If the polyanion is reduced to a HPB, the intensity of the lowest energy charge-transfer band is diminished, and at the same time new bands (intervalence charge transfer) arise in the visible and near-infrared region (13). Figure 8 presents DRS spectra of TiO2 and bulk MPA without thermal treatment. This figure also includes spectra of wet samples resulting from TiO2 impregnation with MPA solutions of different concentrations. Wet samples spectra present one band at 220 nm and another that extends up to 480 nm assigned to ligand-to-metal charge transfer (LMCT) O 20 r Mo 6/ (13). Although this absorption band is partially overlapped with that of the support which appear in the same region up to 400 nm, it can nevertheless be observed that the band extends beyond wavelengths of 400 nm, suggesting the presence of the undegraded PMo12O 340– species. Besides, as the concentration of the species increases on the support, the intensity of the LMCT band also increases. In turn, for the TiO2 sample impregnated with an MPA solution of 57.2 mg Mo/ml initial concentration (M7), Fig. 9 shows DRS spectra of wet and room-temperature dried samples, and of those subsequently calcined at 255, 365 and
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
4257C. A sufficiently extended band suggests the presence of an intact Keggin structure up to 4257C, in agreement with previous results obtained here by other characterization techniques. 3.3.4. X-Ray Diffraction The XRD pattern of the support revealed an anatase structure. For the M12 sample, MPA/TiO2 system, X-ray diagrams were recorded for wet and dried samples, as well as for those calcined up to 4257C, and the patterns only diffraction lines of the support. This result may show that the MPA adsorbed onto the support surface is highly dispersed like a noncrystalline form or crystallites low enough to give diffraction lines. 4. DISCUSSION
Heteropolyacid stability in the impregnating solutions is a key factor for the preparation of titania-supported molybdophosphoric catalysts. For the original solutions and for those resulting from the contact with the support, pH measurements indicated low values (0.8–1), so conditions do not favor MPA
coida
MOLYBDOPHOSPHORIC ACID ADSORPTION ON TITANIA
265
FIG. 8. Diffuse reflectance spectra of bulk MPA, titania, and MPA supported on wet titania samples with different concentrations.
depolymerization. From NMR and UV–visible studies of such solutions (Figs. 2–4, respectively), it can be confirmed that MPA keeps stable during the equilibrium adsorption step, so the ethanol–water solvent used in the impregnant solutions stabilizes the primary Keggin structure. This behavior was unlike that observed for adsorption of MPA aqueous solutions on alumina, where the PMo12O 3– 40 species depolymerized to the PMo11O 7– 39 species (12). On the other hand, a color variation from strong yellow to emerald green in the impregnating solutions indicated the presence of Mo 5/ in the solution (13). According to Eguchi et al. (26) the reducibility of the heteropolyanion must be strongly regulated by its electronic structure and is affected by the central atom species. The study of the adsorption isotherms provides information about some characteristics of the interaction of impregnating solution and support, for example, the amount of adsorbed solute, solute–support interaction strength, and competition for surface adsorption sites between the solute and other species of the impregnant solution, among others. The hydroxyl groups bonded to titania surface are proton-
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
ated in acidic solutions, thereby creating a positively charged surface group which can bind a complex anion by electrostatic attraction. Therefore, the charge change in titania surface observed in a very strong acid medium favor the adsorption of PMo12O 340– ions. The isotherm shape is sigmoidal (Fig. 1) for C(aq) f values below 50 mg Mo/ml. Although S-shaped curves can result of surface precipitation, in this case it was observed no change of the solution pH, neither turbidity in the solution. On the other hand, for high C(aq) f values, the adsorbed concentration shows an asymptotic approach to a constant. This maximum in adsorption can be explained by a monolayer coverage if it is considered that the surface coverage of an MPA molecule is 1.13 nm2 . Then, this isotherm shape would be due probably to competition between MPA molecules and those of the mixture of polar solvents, for titania surface adsorption sites, i.e., the protonated hydroxyl groups. For example, Giles et al. (19) pointed out that monohydric phenols usually give S-shaped curves when adsorbed on a polar substrate, e.g., alumina, from a polar solvent such as water or ethanol.
coida
´ N ET AL. CONCELLO
266
FIG. 9. Diffuse reflectance spectra of bulk MPA, titania, and M7 sample wet, dried at room temperature, and calcined at 255, 365, and 4257C.
The adsorbed concentration plateau observed suggests that the equilibrium impregnation with highly concentrated solutions results essentially in the adsorption of only one species. Such species is the molybdophosphate anion, which adsorbs intact as indicated by DRS spectra of wet samples (Fig. 8). The similarity between the spectra of MPA adsorbed on support and bulk MPA shows that this ion is adsorbed without specific chemical reaction since the primary coordination sphere remains intact. Concerning the stability of MPA, in samples dried at room temperature after being impregnated on titania from MPA solutions of different concentrations, no degradation of the supported PMo12O 340– species was observed by NMR, FT-IR, and DRS (Figs. 5, 7, and 9, respectively). In a previous paper (27), it was concluded that the bulk MPA remains unaltered up to 3507C, and that its degradation starts above 4007C, according to the reaction H3PMo12O40 r 1/2P2O5 / 12MoO3 / 3/2H2O.
[2]
This conclusion was supported by the detection of MoO3
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
by XRD and by differential thermal analysis, which allowed an exothermic peak at 4307C to be observed. These results coincide with those reviewed by Okuhara et al. (20). In the same work (27), but on a different point, it was also concluded that MPA is degraded above 2507C when adsorbed on alumina from ethanol–water solutions. However, for the titania-supported samples studied in this work, the results obtained by NMR, FT-IR, and DRS indicated that the acid keeps the Keggin structure intact up to a calcination temperature of 4257C (see Figs. 5, 6, and 9). Therefore, the thermal stability of MPA adsorbed on titania is similar to that of the bulk acid but considerably higher than that observed when the acid is supported onto alumina. The explanation is given by the corresponding adsorption isotherms which showed that the interaction of MPA and alumina is stronger than that of MPA and titania (27). It must be pointed out that the presence of a partially reduced Keggin phase can be detected by the bluish coloration of the solids in contrast with the white color of solids prepared with alumina support (27). Polyanions that form
coida
MOLYBDOPHOSPHORIC ACID ADSORPTION ON TITANIA
heteropoly blues have polarograms with several reversible diffusion-controlled waves. It follows therefore that reduction must be accompanied by only minor structural changes. This can be achieved provided all MO6 octahedra, which make up the polyanion, have one terminal oxygen atom, since an electron added to M enters an orbital that is predominantly nonbonding, with minimal subsequent bond length alteration (13). On the other hand, in MPA/TiO2 samples, the species adsorbed onto the support surface are highly dispersed as a noncrystalline form, as it was stated by DRX. Therefore, catalyst based on molybdophosphoric acid and supported on titania can be obtained with low solute–support interaction and high dispersion of the heteropolyacid onto the support surface. Besides, as far as thermal stability is concerned, catalytic reactions can be conducted up to 4257C without modifications in the molybdophosphoric acid primary structure. ACKNOWLEDGMENTS We appreciate the technical assistance of Mrs. Graciela Valle and Mrs. Lilian Osiglio.
6. 7. 8. 9. 10. 11. 12.
13. 14. 15. 16. 17. 18. 19. 20.
REFERENCES 1. Ono, Y., ‘‘Perspectives in Catalysis’’ (J. M. Thomas and K. I. Zamaraev, Eds.), p. 431. Blackwell, London, 1992. 2. Izumi, Y., Urabe, K., and Onaka, A., ‘‘Zeolite, Clay, and Heteropolyacid in Organic Reactions’’, p. 100. VCH, Weinheim, 1992. 3. Kozhenikov, I. V., Catal. Rev. Sci. Eng. 37, 311 (1995). 4. Bartoli, M. J., Monceaux, L., Bordes, E., Hecquet, G., and Courtine, P., in ‘‘New Developments in Selective Oxidation by Heterogeneous Catalysis’’ (P. Ruiz and B. Delmon, Eds.), Studies in Surface Science and Catalysis, Vol. 72, p. 81. Elsevier, Amsterdam, 1992. 5. Watzenberger, O., Haeberle, Th., Lynch, D. T., and Emig, G., in ‘‘New Developments in Selective Oxidation’’ (G. Centi and F. Trifiro`, Eds.),
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
21. 22. 23. 24. 25. 26. 27.
267
Studies in Surface Science and Catalysis, Vol. 55, p. 843. Elsevier, Amsterdam 1990. Fournier, M., Thouvenot, R., and Rocchiccioli-Deltcheff, C., J. Chem. Soc., Faraday Trans. 87, 349 (1991). Matsuda, T., Igarashi, A., and Ogino, Y., J. Jpn. Petrol. Inst. 23, 30 (1980). Kasztelan, S., Payen, E., and Moffat, J. B., J. Catal. 125, 45 (1990). Black, J. B., Clayden, N. J., Gai, P. L., Scott, J. D., Serwicka, E. M., and Goodenough, J. B., J. Catal. 106, 1 (1987). Bruckman, K., Haber, J., Lalik, E., and Serwicka, E. M., Catal. Lett. 1, 35 (1988). Va´zquez, P., Ca´ceres, C.,and Blanco, M., Actas XIV Simp. Iberoamer. Cata´l. 3, 1437 (1994). Va´zquez, P., Gonza´lez, M. G., Blanco, M., and Ca´ceres, C., in ‘‘Preparation of Catalysts VI’’ (G. Poncelet, J. Martens, B. Delmon, P. A. Jacobs, and P. Grange, Eds.), Studies in Surface Science and Catalysis, Vol. 91, p. 1121. Elsevier, Amsterdam 1995. Pope, M., ‘‘Heteropoly- and Isopolyoxometallates.’’ Springer, Berlin, 1983. Ng, K. Y. S., and Gulari, E., J. Catal. 95, 33 (1985). Luck, F., Bull. Soc. Chim. Belg. 100, 781 (1991). Matsuda, S., and Kato, A., Appl. Catal. 8, 149 (1983). Damyanova, S., and Fierro, J. L. G., Appl. Catal. A 144, 59 (1996). Kim, D. S., Kurusu, Y., Wachs, I. E., Hardcastle, F. D., and Segawa, K., J. Catal. 120, 325 (1989). Giles, C. H., Mac Ewan, T. H., Nakhawa, S. N., and Smith, D., J. Chem. Soc. (London) 3973 (1960). Okuhara, T., Mizuno, N., and Misono, M., in ‘‘Catalytic Chemistry of Heteropoly Compounds’’ (D. D. Eley, W. O. Haag, and B. Gates, Eds.), Advances in Catalysis Vol. 41, p. 113. Academic Press, New York 1996. Massart, R., Contant, R., Fruchart, J. M., Ciabrini, J. P., and Fournier, M., Inorg. Chem. 16, 2916 (1977). Papaconstantinou, E., Chem. Soc. Rev. 18, 1 (1989). McGarvey, G. B., and Moffat, J. B., J. Mol. Catal. 69, 137 (1991). van Veen, J. A. R., Sudmeijer, O., Emeis, C. A., and de Wit, H., J. Chem. Soc., Dalton Trans. 1825 (1986). Kozhevnikov, I. V., Catal. Lett. 27, 187 (1994). Eguchi, K., Seiyama, T., Yamazoe, N., Katsuki, S., and Taketa, H., J. Catal. 111, 336 (1988). Castillo, M., Va´zquez, P., Blanco, M., and Ca´ceres, C., J. Chem. Soc., Faraday Trans. 92, 3239 (1996).
coida
204, 256–267 (1998)
CS985552
Molybdophosphoric Acid Adsorption on Titania from Ethanol–Water Solutions AnalıB a Concello´n, Patricia Va´zquez, Mirta Blanco, and Carmen Ca´ceres 1 Centro de Investigacio´n y Desarrollo en Procesos CatalıB ticos (CINDECA), UNLP, CONICET, 47 No. 257, 1900, La Plata, Argentina Received December 15, 1997; accepted March 26, 1998
The equilibrium adsorption at 207C of molybdophosphoric acid solutions, using ethanol–water as solvent, on titania was studied. The molybdenum adsorption isotherm showed a sigmoidal shape; low values of molybdenum adsorbed were observed for final equilibrium concentrations lower than 50 mg Mo/ml, and for higher concentrations, the adsorbed molybdenum amount almost reached a plateau. From this isotherm it could be concluded that the solute–support interaction was not strong. UV–visible and NMR spectra of the solutions before and after the adsorption on titania showed that the species PMo12O 340– was present. This species also was observed by DRS in the wet samples and by NMR, FT-IR, and DRS, in the solid samples dried at room temperature and calcined at 255, 310, 365, and 4257C, showing that the thermal stability of molybdophosphoric acid adsorbed on titania is similar to that of the bulk acid. The impregnating solutions and the impregnated solid changed with the time to a bluish color as a consequence of the formation of heteropoly blues which presented Mo 6/ partially reduced to Mo 5/ . The XRD patterns indicated that the species adsorbed onto the support surface are highly dispersed like a noncrystalline form. q 1998 Academic Press Key Words: molybdophosphoric acid; titania; equilibrium adsorption impregnation.
1. INTRODUCTION
It is well known that the heteropolyacids (HPA) and their salts exhibit catalytic versatility arising from their combined acid and redox properties (1, 2). Also the use of heteropolyacids as catalysts for fine chemical synthesis processes is actually developing (3). However, a disadvantage of HPA as catalysts lies in their low stability, and various factors can cause it; for instance, in molybdophosphoric acid (MPA), problems appear with thermal decomposition, acidity changes (4), and loss of the active phase, the latter being due to formation of a volatile molybdic acid (5). HPA have also been supported using different carriers with the aim of improving their stability, which, in addition, 1 To whom correspondence should be addressed. E-mail: hds@nahuel. biol.unlp.edu.ar.
256
0021-9797/98 $25.00 Copyright q 1998 by Academic Press All rights of reproduction in any form reserved.
AID
JCIS 5552
/
6g46$$$521
depends on heteropolyacid concentration and nature of the solvent (6). Bartoli et al. (4) used supports such as SiC to take advantage of its high thermal conductivity and zirconium oxide and cerium oxide to prevent HPA over-reduction. The use of these supports improves the stability of the H4PMo11VO40 active phase in the oxydehydrogenation reaction of isobutyric acid. In the vapor phase synthesis of methyl tert-butyl ether, the activity of silica-supported MPA is higher than that of MPA supported on silica–alumina, alumina, and magnesium oxide (7). This is related to the acidic or basic characteristics of the supports. In particular, silica has a low-level interaction with the HPA, but reports indicate that MPA becomes stable up to 6007C when supported on SiO2 (8). Alkaline salts of HPA are also used as modifiers of the acidic and thermal properties of their corresponding heteropolyacids (9, 10). On the other hand, with the objective to obtain new generations of hydrodesulfurization (HDS) and hydrodenitrogenation (HDN) catalysts, our research group has begun studies about the adsorption on alumina of molybdenum and phosphorus integrating the same structure by using aqueous solutions of pentamolybdodiphosphate anion (11). In another paper (12) the impregnation of alumina was studied with solutions of other HPA, that present primary structure of Keggin type, with general formula [XM12O40 ] ( 80 n ) 0 (13). X is called heteroatom or central atom, usually P, Si, or As; M are called addenda atoms, usually Mo, W, V, or Nb; and n is the valence of X. As it is known, this structure consists of a central XO4 tetrahedron surrounded by twelve MO6 octahedra. Also, it was proved that the use of organic solvents (ethanol–water, dimethylformamide) stabilizes the primary Keggin structure. There have been attempts to optimize the performance of traditional hydrotreatment catalysts as alumina-supported CoMo, NiMo, or NiW. To this end, the use of other supports, notably carbon, titania, and zirconium oxide, was tested. Among these supports, titania has attracted attention in view of the higher HDS activities displayed by molybdenum catalysts supported on this oxide. Ng and Gulari (14) have found that low-loading titania-supported catalysts are more active, compared to alumina-supported catalysts, in the HDS
07-21-98 10:42:23
coida
MOLYBDOPHOSPHORIC ACID ADSORPTION ON TITANIA
reaction carried out at atmospheric pressure. Luck (15) reported that the specific HDS rate constant of a Co–Mo/TiO2 catalyst is greater than Co–Mo catalysts supported on other carriers, as Al2O3 , CeO2 , SiO2 , attapulgite, and carbon. It was also found that titania-based catalysts showed a high activity without presulfurization (16). This feature is one of the advantages of titania-based catalysts in commercial processes. Damyanova and Fierro (17), using the Co (Ni) salts of HPA supported on titania samples, studied the effect of the counterion on the surface and catalytic properties of these solids in the thiophene hydrodesulfurization reaction. They suggested that the highest activity of nickel-promoted catalysts may be caused by an increase in dispersion of molybdenum and its lighter reduction. The catalytic properties, and hence the efficiency, of supported heteropolyacid-based catalysts obtained by equilibrium adsorption impregnation techniques depend on the nature and subsequent stability of the species during the impregnation, drying, and calcination steps, so appropriate conditions must be found to control such species. An important factor is the adsorption behavior of the precursors of active species during the impregnation stage, i.e., the contact between the support and the solution. Thus, as a first study toward the obtainment of MPA on TiO2 catalysts, one of the objectives of this paper is the study of MPA adsorption from ethanol–water solutions using titania as support. Besides, it is presented a detailed analysis of the nature of the species present in the impregnating solutions as well as the characterization of the present species in the solids obtained by equilibrium adsorption impregnation. 2. EXPERIMENTAL
Adsorption studies. To obtain each experimental point of the molybdenum adsorption isotherm at a constant temperature of 207C, 1 g of support (TiO2 anatase, Riedel de Ha¨en, S BET : 9.8 m 2 /g), and 4 ml of MPA(H3PMo12O40r xH2O, MERCK p.a.) solution of known concentration were shaken, for a given time. Solvent employed was an 1:1 (in volume) mixture of water and ethanol Soria 96%. The pH of the impregnating solutions was that resulting of the MPA dissolution in the solvent. The solid was separated of the solution by centrifugation, dried at room temperature, and then calcined at 255, 310, 365, and 4257C. The adsorption isotherm was obtained using different MPA contents, the initial concentration range being 4–110 mg Mo/ml. However, to determine the optimum contact time required to achieve equilibrium between the support and the impregnating solution, adsorption kinetic measures were previously performed, but for only one initial concentration of MPA (110 mg Mo/ml), the contact time being varied from 5 min to 72 h. The amount of molybdenum adsorbed on the support,
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
257
C(ads) (g Mo/100 g TiO2 ), is calculated through a mass balance on the basis of the difference between the initial solute concentration in the solution, C(aq)i (mg Mo/ml), and the equilibrium or final solute concentration in the solution, C(aq) f (mg Mo/ml), and assuming negligible variation in the solution volume. The following expression was used: 01 C(ads) Å [C(aq)i 0 C(aq) f ]Ms r 01 M 01 , a 10
where Ms is the solution mass (g), r is the solution density (g/ml), and Ma , support mass (g). Total molybdenum concentration in the catalysts, CT (g Mo/100 g TiO2 ), was also calculated as CT Å C(ads) / C(occl). The occluded concentration, C(occl) (g Mo/100 g TiO2 ), was calculated taking into account the solution volume retained in the pores (Vr Å 0.74 ml) and C(aq) f , as 01 C(occl) Å C(aq)fVr M 01 . a 10
Quantitative molybdenum analysis. The molybdenum concentration in the solutions, both before and after being in contact with titania, was measured by atomic absorption spectrometry with an IL model 457 double-beam spectrophotometer (Instrumentation Laboratory Inc.). The analysis was carried out at the following conditions: wavelength, 313 nm; band width, 0.5 nm; lamp current, 7 mA; phototube amplification, 700 V; burner height, 7 mm; and acetylene/ air flame in 7:14 ratio. pH measurements. As standards for pH measurements, Tritisol buffer solutions (Merck) of pH 7.00 { 0.02 (phosphate) and of pH 2.00 { 0.02 (citrate-HCl) were employed. An Instrumentalia PH-MD digital equipment connected to an Orion Research model 91-03 pH-sensitive combined semielectrode was used. UV–visible spectroscopy. Spectra of solutions were obtained with a Varian Super Scan 3 double-beam UV–visible spectrophotometer with a built-in recorder, using quartz cells of 0.5 mm optical path, in the range of 200–600 nm. MPA solutions of different concentrations were analyzed before and after being in contact during different times with titania. Nuclear magnetic resonance. The MPA solutions, both before and after being in contact with titania, were analyzed by 31P NMR. A Bruker MSL-300 equipment linked to a ‘‘SOLIDCYC.DC’’ pulse program was utilized, using 5 ms pulses, a repetition time of 10 s, and a frecuency of 121.496 MHz for 31P at 247C, being the resolution of 3.052 Hz, per point. Phosphoric acid 85% was employed as external reference. The same equipment and similar conditions were used to obtain 31P MAS-NMR spectra of the impregnated solids
coida
´ N ET AL. CONCELLO
258
TABLE 1 Molybdenum Concentrations in Solution and in Solid for the Studied Samples C(aq)i C(aq)f C(ads) CT Sample (mg Mo/ml) (mg Mo/ml) (g Mo/100 g TiO2) (g Mo/100 g TiO2) M1 M3 M4 M5 M7 M8 M10 M11 M12
3.71 21.69 27.05 45.36 57.20 63.11 79.39 90.16 110.0
3.10 19.72 26.45 41.91 49.31 52.27 69.28 77.92 94.0
0.36 0.72 0.29 1.26 2.88 3.94 3.72 4.89 5.24
0.59 2.06 2.10 3.87 5.87 7.18 11.57 10.38 17.49
dried at room temperature and calcined at 255, 310, 365, and 4257C. A sample holder of 5 mm diameter and 10 mm in height was used, the spin rate being 2.1 kHz. The repetition time was 3 s, and several hundred pulse responses were collected. Fourier transform infrared spectroscopy. A Bruker IFS 66 instrument, pellets in KBr and a measuring range of 400– 1500 cm01 were used to obtain the FT-IR spectra of titania, bulk molybdophosphoric acid, and MPA on titania samples dried at room temperature and calcined at 255, 310, 365, and 4257C. Diffuse reflectance spectroscopy. The samples were studied with an UV–visible Varian Super Scan 3 equipment, fitted with a diffuse reflectance chamber with a BaSO4 inner surface. DRS spectra of titania, bulk molybdophosphoric acid, and wet, dried at room temperature, and calcined MPA on titania samples, in the range of 200–600 nm, were obtained. Each sample was compacted in a Teflon sample holder, which was then covered with a quartz circular window to obtain a sample thickness of 2 mm.
FIG. 2. 31P NMR spectra of MPA impregnating solutions, corresponding to M12 sample, before (a) and after (b) the adsorption on titania.
X-ray diffraction. The XRD patterns of titania and calcined samples were obtained by using a Phillips PW-1714 diffractometer. In this measurements, the following condi˚ ), nickel filter, tions were used: CuKa radiation (1.5417 A 30 mA and 40 kV in the high voltage source, scanning angle (2u ) from 5 to 607, and scanning rate of 17 per minute. 3. RESULTS
3.1. Adsorption Kinetics and Equilibrium Isotherms
For the contact between TiO2 and MPA solutions in ethanol–water, the adsorption kinetics was measured. To this end, an initial MPA concentration of 110 mg Mo/ml was used and measurements were carried out at contact times from 5 min to 72 h. The results showed that C(ads) remains constant from 10 min on. For this adsorption kinetics, a small decrease in pH was observed for the final solutions in relation to the initial one (pH 1). The pH values of the final solutions were between 0.8 and 1. Thus, unimportant variations of pH in the impregnating solution were observed, allowing to asume that the depolymerization of PMo12O 340– to PMo11O 739– according to the reaction [1] would not be significant: / PMo12O 340– / 3H2O r PMo11O 739– / MoO 20 4 / 6H .
FIG. 1. Mo adsorption isotherm on titania from MPA solutions in ethanol–water at 207C.
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
[1]
The hydroxyl groups of the titania surface in solution tends to be either positively or negatively charged below or
coida
MOLYBDOPHOSPHORIC ACID ADSORPTION ON TITANIA
259
FIG. 3. UV–visible spectra of 110 mg Mo/ml initial impregnating solution at different times.
above the IEP of titania (anatase: 6.2) (18), according to the following equilibrium: Ti–OH 2/ B Ti–OH B Ti–O 0 . Above the IEP of titania, the hydroxyl groups tend to be negatively charged, so that it is relatively hard for the anions to be adsorbed on the titania surface, due to the electrostatic repulsion. On the other hand, below the IEP of titania the adsorbed amount of the molydophosphate anion should be favored. Hence, in our working conditions, MPA adsorption on titania is favored by the low pH value of the initial solutions, resulting of the MPA dissolution in the solvent, and of the solutions resulting from the contact with the support. For the experiments that allowed us to obtain the adsorption isotherm, to ensure that the adsorption–desorption equilibrium had been reached, a contact time of 72 h for each point was selected. The C(aq)i , C(aq) f , C(ads), and CT values are summarized in Table 1. The molybdenum adsorption isotherm from solutions of H3PMo12O40 on titania at 207C, expressed as C(ads) in function of C(aq) f , is shown in Fig. 1. The isotherm has a sigmoidal shape; it presents very low initial values of C(ads) and later, for MPA concentrations above 50 mg Mo/ml in the final solutions, it tends to reach a plateau of C(ads) values. The initial slope shows that
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
adsorption becomes easier as concentration in solution increases. Hence, at low concentration, it is possible that the solute molecules meet strong competition, for surface support sites, from molecules of the solvent (19). At high concentration, the presence of one plateau indicates the adsorption of only one molybdenum–phosphorus species, possibly the molybdophosphate anion; this will nevertheless be confirmed in the characterization of solutions and solids. 3.2. Characterization of the Solutions
3.2.1. Nuclear Magnetic Resonance The 31P chemical shift ( d ) provides important information concerning the structure, composition and electronic states of heteropoly compounds. The chemical shift for 31P in aqueous solution at 03.9 ppm for PMo12O 340– (20) was correlated with the P{Oa bond strength [ n(P{Oa )] (21). The 31P NMR spectra of MPA solutions in ethanol–water, corresponding to M12 sample, both before and after the adsorption on titania are shown in Fig. 2. The chemical shifts observed in the spectra of initial and final solutions, 03.6 and 03.35, respectively, show that PMo12O 340– is neither de-
coida
´ N ET AL. CONCELLO
260
FIG. 4. UV–visible spectra of impregnating solutions before (a) and after (b) the adsorption on titania, for different samples and an impregnation time of 72 h.
polymerized nor degraded as a consequence of the contact with titania. Small solvent effect on 31P chemical shift of the heteropolyanion in ethanol–water with respect to that in water indicates that the central atom is protected by the four groups of three edge-shared MoO6 octahedra (M3O13 triad). 3.2.2. UV–Visible Spectroscopy For an MPA solution in ethanol–water of 110 mg Mo/ ml initial concentration, Fig. 3 shows the UV–visible spectra for different times, from 5 min to 72 h. In the range under study, an absorption band with a maximum at 310 nm belonging to the oxygen–metal charge transfer is observed, suggesting the presence of only one species, the intact Keggin phase. Likewise, spectra showing the same band were observed for the impregnating solutions analyzed after contact with the support. Papaconstantinou (22) showed that non-reduced HPA are characterized by oxygen-to-metal charge transfer bands in the near visible and UV regions and no absorption in the visible (all metals are Mo VI or W VI with d 0 configuration). Reduction
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
of HPA proceeds without substantial change of their structure with addition of a certain characteristic number of electrons. On reduction they are colored mainly blue, producing the socalled heteropoly blues (HPB) which have a broad absorption band at about 700 nm. Because the HPB contain metal ions in different oxidation states they are usually assigned to the category of mixed valence compounds. That band is attributed to metal-to-metal charge transfer (Mo 5/ r Mo 6/ ) which is responsible for the blue color of the compounds, and to d–d transitions of the d 1 metal ions (13). The equipment used records spectra up to a maximum of 600 nm so that the region around 700 nm is not discussed. Notwithstanding, color variations were observed with time both in the original and in the final solutions; they are initially yellow and become greenish as a consequence of HPB formation. Spectra of initial and final solutions are presented in Fig. 4a and b, respectively, for only one contact time with the support (72 h) and different concentrations of the impregnating solution. Both in original solutions and in final ones, the above-mentioned value of the absorption maximum (310 nm) holds, though the absorption intensity varies depending upon the concentration change. Hence, the technique allows
coida
MOLYBDOPHOSPHORIC ACID ADSORPTION ON TITANIA
261
FIG. 4—Continued
to observe that the H3PMo12O 40 is neither depolymerized nor degraded in the solution during the equilibrium adsorption, in agreement with NMR results and with the low pH presented by the impregnating solutions before and after contact with titania (pH 0.8–1). 3.3. Characterization of the Solids
3.3.1. Nuclear Magnetic Resonance Figure 5 shows 31P MAS-NMR spectra recorded for bulk molybdophosphoric acid and MPA-impregnated titania samples, dried at room temperature and calcined at 255, 310, 365, and 4257C. According to several authors, the chemical shift ( d ) of the bulk acid is found between 02.9 and 04.8 ppm (20, 23, 24). The spectrum obtained (Fig. 5a) exhibits two lines, one at 03.41 and another at 03.83 ppm, which can be assigned to the acid in different hydration levels (20). The spectra showed by MPA on titania samples M5 and M12, dried at room temperature, present only one line with maxima at 04.1 ppm (Fig. 5b) and 03.51 ppm (Fig. 5c), respectively. This suggests that, in the concentration range
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
under study, the structure of the supported molybdophosphate anion keeps intact. When the M12 sample is subsequently calcined at 255, 310, 365, and 4257C, the corresponding spectra also show only one line, which is observed at 04.1, 04.0, 04.37 and 04.1 ppm, respectively (Fig. 5d– g). The shifts shown by calcined samples of MPA on TiO2 in relation to the dried sample are 00.86 ppm at most and 00.54 ppm compared to the bulk acid lines. All observed chemical shifts can be assigned to the acid with undegraded Keggin structure. The slight shift toward higher field can be caused by an increase of the interaction of the acid with the support during calcination, as it was the case for another heteropolyacid ( H3PW12O40 ) on carbon ( 25 ) . 3.3.2. Fourier Transform Infrared Spectroscopy Each M3O13 triad in the Keggin structure has a common oxygen vertex connected to the central heteroatom X{Oa (M3O13 ). The other three classes of symmetric-equivalent oxygens in the a isomer of Keggin structure are (20) M{Ob{M, connecting two M3O13 units by corner sharing; M{Oc{ M, connecting two M3O13 units by edge sharing;
coida
´ N ET AL. CONCELLO
262
FIG. 5. 31P MAS-NMR spectra of bulk MPA (a), M5 sample dried at room temperature (b), M12 sample dried at room temperature (c), and M12 sample calcined at 255 (d), 310 (e), 365 (f ), and 4257C (g).
and terminal oxygens Od| M. The maxima of the stretching modes of vibration for IR absorption bands of MPA are P{Oa (1070 cm01 ), Mo|Od (965 cm01 ), Mo{Ob{Mo (870 cm01 ), and Mo{Oc{ Mo (790 cm01 ) (20). TiO2 spectra obtained for several calcination temperatures (310, 365, and 4257C) are presented in Fig. 6a, while the spectrum of the noncalcined bulk MPA is shown in Fig. 6b. The graphs show that the titania absorption band extends over a wide region with maxima in the 400–900 cm01 zone that mask the MPA bands at 790 and 870 cm01 . Two characteristic bands of the anion under study (1070 and 965 cm01 ) are observed in the remaining zone of the spectra. An outstanding feature of these spectra (Fig. 6a) is the widening of the TiO2 band toward higher wavenumbers as the temperature increases, the change being more considerable from 365 to 4257C. The thermal stability of TiO2 support is usually lower than that of g-Al2O3 and the particular note is the reducibility of TiO2 , which results in the formation of various stoichiometric and nonstoichiometric oxides, tak-
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
ing Ti ions several valence states (16). Consequently, shifts can be ascribed to structural changes in view of the low TiO2 thermal stability. For the solids obtained by impregnating titania for 72 h with an ethanol–water solution of MPA (initial concentration: 110 mg Mo/ml), Fig. 6b also presents the FT-IR spectra of dried and calcined samples, treated at the same temperatures used before for the support. For samples dried at room temperature and calcined below 3657C, no modifications are observed in the 1070 and 975 cm01 bands. It can be seen that the intensity of such bands decrease beyond 3657C, becoming hardly perceptible at 4257C because it is masked by the widening of the support band or, else, it vanishes owing to the start of Keggin unit degradation. However, bearing in mind NMR results, the latter possibility may be discarded. Figure 7 presents the FT-IR spectra of samples obtained from MPA solutions of different concentration and dried at room temperature. These spectra show the characteristic
coida
MOLYBDOPHOSPHORIC ACID ADSORPTION ON TITANIA
263
FIG. 6. FT-IR spectra of titania calcined at different temperatures (a) and M12 sample dried at room temperature and calcined at different temperatures and bulk MPA (b).
bands of the heteropolyacid, the intensity of which increases with the concentration. FT-IR results are in agreement with those arrived at by NMR for samples prepared with different concentrations and subjected to various thermal treatments. 3.3.3. DRS The charge-transfer absorption spectra of most non-reduced polyanions appear in the 200–500 nm region and
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
consist of bands which may be ascribed to oxygen-to-metal transfers. The tetrahedral Mo exhibits two absorption bands at 220 and 260 nm, whereas Mo in octahedral coordination presents, apart from those bands, another band at higher wavelength. Most polymolybdates and polytungstates contain octahedrally coordinated Mo 6/ and W 6/ but with the metal atom displaced from the center of the octahedra toward a corner or an edge. The short M{O bonds produced by these displacements imply the existence of significant metal–oxy-
coida
´ N ET AL. CONCELLO
264
FIG. 7. FT-IR spectra of MPA supported on titania samples with different concentrations, dried at room temperature.
gen p-bonding, and this feature is the key to the stabilities of all polyoxoanion structures. If the polyanion is reduced to a HPB, the intensity of the lowest energy charge-transfer band is diminished, and at the same time new bands (intervalence charge transfer) arise in the visible and near-infrared region (13). Figure 8 presents DRS spectra of TiO2 and bulk MPA without thermal treatment. This figure also includes spectra of wet samples resulting from TiO2 impregnation with MPA solutions of different concentrations. Wet samples spectra present one band at 220 nm and another that extends up to 480 nm assigned to ligand-to-metal charge transfer (LMCT) O 20 r Mo 6/ (13). Although this absorption band is partially overlapped with that of the support which appear in the same region up to 400 nm, it can nevertheless be observed that the band extends beyond wavelengths of 400 nm, suggesting the presence of the undegraded PMo12O 340– species. Besides, as the concentration of the species increases on the support, the intensity of the LMCT band also increases. In turn, for the TiO2 sample impregnated with an MPA solution of 57.2 mg Mo/ml initial concentration (M7), Fig. 9 shows DRS spectra of wet and room-temperature dried samples, and of those subsequently calcined at 255, 365 and
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
4257C. A sufficiently extended band suggests the presence of an intact Keggin structure up to 4257C, in agreement with previous results obtained here by other characterization techniques. 3.3.4. X-Ray Diffraction The XRD pattern of the support revealed an anatase structure. For the M12 sample, MPA/TiO2 system, X-ray diagrams were recorded for wet and dried samples, as well as for those calcined up to 4257C, and the patterns only diffraction lines of the support. This result may show that the MPA adsorbed onto the support surface is highly dispersed like a noncrystalline form or crystallites low enough to give diffraction lines. 4. DISCUSSION
Heteropolyacid stability in the impregnating solutions is a key factor for the preparation of titania-supported molybdophosphoric catalysts. For the original solutions and for those resulting from the contact with the support, pH measurements indicated low values (0.8–1), so conditions do not favor MPA
coida
MOLYBDOPHOSPHORIC ACID ADSORPTION ON TITANIA
265
FIG. 8. Diffuse reflectance spectra of bulk MPA, titania, and MPA supported on wet titania samples with different concentrations.
depolymerization. From NMR and UV–visible studies of such solutions (Figs. 2–4, respectively), it can be confirmed that MPA keeps stable during the equilibrium adsorption step, so the ethanol–water solvent used in the impregnant solutions stabilizes the primary Keggin structure. This behavior was unlike that observed for adsorption of MPA aqueous solutions on alumina, where the PMo12O 3– 40 species depolymerized to the PMo11O 7– 39 species (12). On the other hand, a color variation from strong yellow to emerald green in the impregnating solutions indicated the presence of Mo 5/ in the solution (13). According to Eguchi et al. (26) the reducibility of the heteropolyanion must be strongly regulated by its electronic structure and is affected by the central atom species. The study of the adsorption isotherms provides information about some characteristics of the interaction of impregnating solution and support, for example, the amount of adsorbed solute, solute–support interaction strength, and competition for surface adsorption sites between the solute and other species of the impregnant solution, among others. The hydroxyl groups bonded to titania surface are proton-
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
ated in acidic solutions, thereby creating a positively charged surface group which can bind a complex anion by electrostatic attraction. Therefore, the charge change in titania surface observed in a very strong acid medium favor the adsorption of PMo12O 340– ions. The isotherm shape is sigmoidal (Fig. 1) for C(aq) f values below 50 mg Mo/ml. Although S-shaped curves can result of surface precipitation, in this case it was observed no change of the solution pH, neither turbidity in the solution. On the other hand, for high C(aq) f values, the adsorbed concentration shows an asymptotic approach to a constant. This maximum in adsorption can be explained by a monolayer coverage if it is considered that the surface coverage of an MPA molecule is 1.13 nm2 . Then, this isotherm shape would be due probably to competition between MPA molecules and those of the mixture of polar solvents, for titania surface adsorption sites, i.e., the protonated hydroxyl groups. For example, Giles et al. (19) pointed out that monohydric phenols usually give S-shaped curves when adsorbed on a polar substrate, e.g., alumina, from a polar solvent such as water or ethanol.
coida
´ N ET AL. CONCELLO
266
FIG. 9. Diffuse reflectance spectra of bulk MPA, titania, and M7 sample wet, dried at room temperature, and calcined at 255, 365, and 4257C.
The adsorbed concentration plateau observed suggests that the equilibrium impregnation with highly concentrated solutions results essentially in the adsorption of only one species. Such species is the molybdophosphate anion, which adsorbs intact as indicated by DRS spectra of wet samples (Fig. 8). The similarity between the spectra of MPA adsorbed on support and bulk MPA shows that this ion is adsorbed without specific chemical reaction since the primary coordination sphere remains intact. Concerning the stability of MPA, in samples dried at room temperature after being impregnated on titania from MPA solutions of different concentrations, no degradation of the supported PMo12O 340– species was observed by NMR, FT-IR, and DRS (Figs. 5, 7, and 9, respectively). In a previous paper (27), it was concluded that the bulk MPA remains unaltered up to 3507C, and that its degradation starts above 4007C, according to the reaction H3PMo12O40 r 1/2P2O5 / 12MoO3 / 3/2H2O.
[2]
This conclusion was supported by the detection of MoO3
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
by XRD and by differential thermal analysis, which allowed an exothermic peak at 4307C to be observed. These results coincide with those reviewed by Okuhara et al. (20). In the same work (27), but on a different point, it was also concluded that MPA is degraded above 2507C when adsorbed on alumina from ethanol–water solutions. However, for the titania-supported samples studied in this work, the results obtained by NMR, FT-IR, and DRS indicated that the acid keeps the Keggin structure intact up to a calcination temperature of 4257C (see Figs. 5, 6, and 9). Therefore, the thermal stability of MPA adsorbed on titania is similar to that of the bulk acid but considerably higher than that observed when the acid is supported onto alumina. The explanation is given by the corresponding adsorption isotherms which showed that the interaction of MPA and alumina is stronger than that of MPA and titania (27). It must be pointed out that the presence of a partially reduced Keggin phase can be detected by the bluish coloration of the solids in contrast with the white color of solids prepared with alumina support (27). Polyanions that form
coida
MOLYBDOPHOSPHORIC ACID ADSORPTION ON TITANIA
heteropoly blues have polarograms with several reversible diffusion-controlled waves. It follows therefore that reduction must be accompanied by only minor structural changes. This can be achieved provided all MO6 octahedra, which make up the polyanion, have one terminal oxygen atom, since an electron added to M enters an orbital that is predominantly nonbonding, with minimal subsequent bond length alteration (13). On the other hand, in MPA/TiO2 samples, the species adsorbed onto the support surface are highly dispersed as a noncrystalline form, as it was stated by DRX. Therefore, catalyst based on molybdophosphoric acid and supported on titania can be obtained with low solute–support interaction and high dispersion of the heteropolyacid onto the support surface. Besides, as far as thermal stability is concerned, catalytic reactions can be conducted up to 4257C without modifications in the molybdophosphoric acid primary structure. ACKNOWLEDGMENTS We appreciate the technical assistance of Mrs. Graciela Valle and Mrs. Lilian Osiglio.
6. 7. 8. 9. 10. 11. 12.
13. 14. 15. 16. 17. 18. 19. 20.
REFERENCES 1. Ono, Y., ‘‘Perspectives in Catalysis’’ (J. M. Thomas and K. I. Zamaraev, Eds.), p. 431. Blackwell, London, 1992. 2. Izumi, Y., Urabe, K., and Onaka, A., ‘‘Zeolite, Clay, and Heteropolyacid in Organic Reactions’’, p. 100. VCH, Weinheim, 1992. 3. Kozhenikov, I. V., Catal. Rev. Sci. Eng. 37, 311 (1995). 4. Bartoli, M. J., Monceaux, L., Bordes, E., Hecquet, G., and Courtine, P., in ‘‘New Developments in Selective Oxidation by Heterogeneous Catalysis’’ (P. Ruiz and B. Delmon, Eds.), Studies in Surface Science and Catalysis, Vol. 72, p. 81. Elsevier, Amsterdam, 1992. 5. Watzenberger, O., Haeberle, Th., Lynch, D. T., and Emig, G., in ‘‘New Developments in Selective Oxidation’’ (G. Centi and F. Trifiro`, Eds.),
AID
JCIS 5552
/
6g46$$$521
07-21-98 10:42:23
21. 22. 23. 24. 25. 26. 27.
267
Studies in Surface Science and Catalysis, Vol. 55, p. 843. Elsevier, Amsterdam 1990. Fournier, M., Thouvenot, R., and Rocchiccioli-Deltcheff, C., J. Chem. Soc., Faraday Trans. 87, 349 (1991). Matsuda, T., Igarashi, A., and Ogino, Y., J. Jpn. Petrol. Inst. 23, 30 (1980). Kasztelan, S., Payen, E., and Moffat, J. B., J. Catal. 125, 45 (1990). Black, J. B., Clayden, N. J., Gai, P. L., Scott, J. D., Serwicka, E. M., and Goodenough, J. B., J. Catal. 106, 1 (1987). Bruckman, K., Haber, J., Lalik, E., and Serwicka, E. M., Catal. Lett. 1, 35 (1988). Va´zquez, P., Ca´ceres, C.,and Blanco, M., Actas XIV Simp. Iberoamer. Cata´l. 3, 1437 (1994). Va´zquez, P., Gonza´lez, M. G., Blanco, M., and Ca´ceres, C., in ‘‘Preparation of Catalysts VI’’ (G. Poncelet, J. Martens, B. Delmon, P. A. Jacobs, and P. Grange, Eds.), Studies in Surface Science and Catalysis, Vol. 91, p. 1121. Elsevier, Amsterdam 1995. Pope, M., ‘‘Heteropoly- and Isopolyoxometallates.’’ Springer, Berlin, 1983. Ng, K. Y. S., and Gulari, E., J. Catal. 95, 33 (1985). Luck, F., Bull. Soc. Chim. Belg. 100, 781 (1991). Matsuda, S., and Kato, A., Appl. Catal. 8, 149 (1983). Damyanova, S., and Fierro, J. L. G., Appl. Catal. A 144, 59 (1996). Kim, D. S., Kurusu, Y., Wachs, I. E., Hardcastle, F. D., and Segawa, K., J. Catal. 120, 325 (1989). Giles, C. H., Mac Ewan, T. H., Nakhawa, S. N., and Smith, D., J. Chem. Soc. (London) 3973 (1960). Okuhara, T., Mizuno, N., and Misono, M., in ‘‘Catalytic Chemistry of Heteropoly Compounds’’ (D. D. Eley, W. O. Haag, and B. Gates, Eds.), Advances in Catalysis Vol. 41, p. 113. Academic Press, New York 1996. Massart, R., Contant, R., Fruchart, J. M., Ciabrini, J. P., and Fournier, M., Inorg. Chem. 16, 2916 (1977). Papaconstantinou, E., Chem. Soc. Rev. 18, 1 (1989). McGarvey, G. B., and Moffat, J. B., J. Mol. Catal. 69, 137 (1991). van Veen, J. A. R., Sudmeijer, O., Emeis, C. A., and de Wit, H., J. Chem. Soc., Dalton Trans. 1825 (1986). Kozhevnikov, I. V., Catal. Lett. 27, 187 (1994). Eguchi, K., Seiyama, T., Yamazoe, N., Katsuki, S., and Taketa, H., J. Catal. 111, 336 (1988). Castillo, M., Va´zquez, P., Blanco, M., and Ca´ceres, C., J. Chem. Soc., Faraday Trans. 92, 3239 (1996).
coida