Oxide supports for 12-tungstosilicic acid catalysts in gas phase synthesis of MTBE

Abstract

A series of simple and mixed oxides: SiO₂, AlPO₄, SiO₂-Al₂O₃, TiO₂, kaolin and γ-Al₂O₃ differing by their basicities as characterised by effective negative charge on oxygen calculated according to Sanderson’s theorem was used for the preparation of supported H₄SiW₁₂O₄₀ catalysts. Using gas phase synthesis of MTBE as the catalytic test reaction it has been shown that on the supports of highest basicity, i.e. γ-Al₂O₃ and kaolin the heteropolyacid was decomposed resulting in the catalyst which exhibited only low activity. On the other hand, the least basic oxide SiO₂ gave the most active catalysts. However, the sequence of activities differed from that of basicities by the fact that the activity of TiO₂-supported catalyst was at the coverage with H₄SiW₁₂O₄₀ Θ=0.25 nearly as high as SiO₂ supported one and at Θ=1.0 TiO₂-supported catalyst was the best. This indicates that the basicity of the support is only one of the factors determining the activity of supported heteropolyacid catalysts in acid–base type reactions. On the other hand good correlation was obtained between the catalytic activity and the neutralisation heat of the acid sites on the surface of catalyst which was determined using thermometric titration with n-butylamine solution in toluene.

Introduction

Methyl-tert-butyl ether (MTBE) is widely used additive to automotive fuel which enabled elimination of poisonous tetraethyl lead as octane number enhancer. The majority of this product is obtained industrially by catalytic addition of methanol to isobutene in liquid phase on sulphonated resins of the type of Amberlyst-15 at 363–373 K and 1.5 MPa. Such catalysts are not stable above 373 K, they lose activity and slow emission of sulphuric acid becomes troublesome both because of the corrosion of the technical installations and environmental reasons. This stimulates the studies on the other types of catalysts more convenient for MTBE synthesis. As the acid catalysts necessary for this reaction zeolites [1], [2], [3], modified zeolites [4], [5], [6] and silicalites [7] were investigated in gas phase MTBE synthesis. Another group of catalysts which arouses the interest are heteropolyacids (HPA) belonging to the strongest inorganic acids. Dodecaheteropolyacids due to their high Brönsted acid strength comparable to that of the superacids are very good catalysts for numerous acid–base type reactions. Moffat [8] basing on his extended Hückel calculations concluded that the acid strength of HPA is inversely related to the magnitude of the negative charge on the outer oxygen atoms of the anions. These results were confirmed in further investigations on acidic and catalytic properties of solid HPAs carried by Highfield and Moffat [9], [10]. The presence of very loosely bonded protons enables the penetration of polar molecules such as H₂O, NH₃, amines and alcohols into the bulk of heteropolyacid crystallites where they get protonated forming so-called pseudoliquid phase [11].

Following an American patent [12] and a paper by Igarashi et al. [13] in which a possibility of MTBE synthesis on HPAs from methanol and isobutene has been signalled Ono and Baba [14] used this process as a gas phase test reaction (at 353 K) in their study of catalytic properties of H₃PMo₁₂O₄₀, H₃PW₁₂O₄₀ and H₄SiW₁₂O₄₀ supported on active carbon and also their Ag, Cu and Al salts. A systematic investigation of MTBE formation in gas phase over HPAs was undertaken by Shikata et al. [15] who investigated unsupported Keggin-type HPAs, H_nXW₁₂O₄₀ (X=P, Si, Ge, B and Co) and also Dawson-type H₆P₂W₁₈O₆₂. In this paper a unique dependence of reaction rate on the methanol vapour pressure was observed. On increasing its partial pressure the reaction rate at first increases but above a certain value depending on the kind of HPA it decreases systematically. This behaviour was interpreted as being due to the fact that at higher pressures methanol penetrating the bulk of HPA crystallites is forming protonated oligomers (CH₃OH)_nH⁺ (n>3) which in contrast to the protonated monomers are non-active in the catalytic reaction. Such protonation of methanol in the bulk of HPA crystallites has been described for the first time by Highfield and Moffat [16], [17] who studied photoacoustic FTIR spectra of CH₃OH in H₃PW₁₂O₄₀ and latter by other authors as, e.g. in [18], [19]. Shikata et al. [15] suggested that the presence of methanol which penetrated the bulk of HPA crystallites enables also the penetration of non-polar isobutene and the reaction occurs in the pseudoliquid phase. An alternative model has been proposed in [20], [21], [22] basing on sorption and kinetic experiments. In this model the reaction between isobutene adsorbed at the HPA crystallites surface and methanol supplied from the bulk has been assumed and loosely bonded protons responsible for the high acid strength of HPA played the role of catalytically active centres.

Dodecaheteropolyacids as the catalysts for MTBE formation were studied also by Maksimov and Kozhevnikov [23] by measuring isobutene absorption at 315 K in liquid methanol containing dissolved HPAs. Molnar et al. [24] described experiment carried out in an autoclave at 358 K and 5 MPa.

Besides dodecaheteropolyacids several authors studied Dawson-type H₆P₂W₁₈O₆₂ as the MTBE formation catalyst in gas phase [15], [25], [26], [27] and also in liquid one [23]. In all cases higher activity of Dawson-type catalyst than that of Keggin-type was observed. Here, it should be observed that synthesis of MTBE by etherification of methanol–tert-butanol mixture—the process of smaller industrial importance than synthesis from methanol and isobutene—has been proved to occur also on heteropoly oxometalates in [24], [28], [29], [30].

Most of the investigations described above were carried out using unsupported HPAs. Active carbon as the support for HPAs in MTBE synthesis was used in [14], [31] as well as nanoporous carbon membranes in [32]. Amberlyst-15 sulphonic resin is known as the industrial catalyst for MTBE synthesis. This resin used as the support for HPAs gave excellent catalysts [33]. The other authors used also modified clay [28], silica [26], [29], [34], silica and alumina [30], [34], [35], [36], MCM-41 [37] and also polyaniline [38], but no more ample comparative studies of the supports for HPAs as the MTBE synthesis catalysts were undertaken. A review of the supports for heteropoly oxometalates as the catalysts in different reactions has been given recently in [39].

In the search for the suitable supports for HPA in acid–base catalysis the expected behaviour of HPA protons should be taken into account. In fact HPA anion is a very weak base and we may expect that protons from heteropolyacid will migrate to the surface basic centres, where they may be fixed more strongly than in HPA molecule thus giving weaker acid centres less active in catalysis. In the case of oxidic supports the role of such basic centres is played by surface oxygen atoms. Hence, the less basic are surface oxygen atoms the better should be the support. There are two approaches which enable to estimate roughly the basicity of oxygen atom in the oxides: one of them is Sanderson’s theorem [40] which allows to estimate effective negative charge in the oxide basing on Sanderson’s electronegativities. Another approach is optical basicity proposed by Duffy [41] and developed by Lebouteille and Courtine [42]. As it will be shown latter both indexes in a series of simple and mixed oxides exhibit similar trends indicating the increase or decrease of basicity. In the present research an attempt has been undertaken to correlate the catalytic properties of the catalysts containing heteropolyacid supported on the most frequently used simple and mixed oxides. For this purpose the following oxides were used as supports for H₄SiW₁₂O₄₀: SiO₂, TiO₂, AlPO₄, SiO₂-Al₂O₃, kaolin, γ-Al₂O₃. The formation of MTBE from methanol and isobutene has been taken as the catalytic test reaction. However, it should be observed that both methods of characterising the basicity of the oxides have some obvious deficiencies: they do not take into account the structural factors and the calculated values characterise the bulk of the oxide crystallites while the situation at the surface, which is responsible for catalytic effects, may not necessarily be identical. It should be also noticed that besides the basicity of the supports also other factors may influence the activity of the supported catalyst among them the porosity of the support, the distribution of the active mass over the surface, the structural factors, etc.

This is why we characterised the resulting acidity of the acid sites at the catalyst’s surface also by the independent calorimetric measurements of their neutralisation heat determined in their reaction with n-butylamine. In some cases also the ${}^1\text{H}$ MAS NMR measurements were carried out.