DEALUMINATION OF AMORPHOUS SILICA-ALUMINAS WITH ACETYLACETONE AND ITS INFLUENCE ON ACIDITY AND ACTIVITY

Get Complete Project Material File(s) Now! »

Applications of ASAs in catalysis

Aluminosilicates, such as zeolites and amorphous silica-alumina (ASA), have been extensively used in oil refining, petrochemistry and fine chemicals production and their catalytic activities are associated with their surface acidity properties.1,2 Zeolites are crystalline materials that possess an exceptional combination of properties such as high thermal stability, high Brønsted acidity, and microporosity. The presence of regular micropores of molecular dimensions is responsible of their unequalled shape selectivity in catalytic conversions and separations.3 However, in many applications, this microporosity induces diffusional limitations due to the restricted transport of molecules to the active sites.
ASA are mesoporous materials that lack long-range order and are therefore XRD amorphous. Their mesoporosity facilitates physical transport of reactants and products. In addition, ASA generally exhibit fewer and milder Brønsted acid sites than zeolites. Busca reviewed the application of these materials in industrial processes.4 Although zeolites have supplanted ASA in many catalytic processes thanks to their stronger Brønsted acidity, ASA are still preferred to zeolite as hydrocracking supports when higher selectivity for middle distillates is desired or for the conversion of bulky molecules. They also are still used in dehydrochlorination of halided hydrocarbon and as support of sulfide for hydrotreatment and of metal for ring opening reactions.4 Beside conventional silica-alumina, other amorphous aluminosilicates are used: Al rich silica-aluminas are possibly used in the production of dimethylether (DME) or the coproduction of methanol and DME from syngas,4 whereas silicate alumina (prepared by grafting TEOS on alumina) have been reported to be used in the dehydration of terbutyl alcohol to isobutylene.4 At the other side of the composition, aluminated silica (prepared by addition of alumina precursors to preformed silica) are likely to be the catalysts used for the cracking of Methyl tert-butyl ether (MTBE) to isobutylene.4 Another specificity of ASA is their high density of Lewis sites, which could also lead to their application in Lewis catalysed reaction (e.g. Meerwein–Ponndorf–Verley (MPV) reduction of carbonyl compounds). The strong Lewis acid sites of ASA have been reported to contribute to the Sanderson electronegativity of the support and hence to the increase in the electronegativity of supported metallic particles (e.g. Pt nanoparticles) and in turns to an increased activity of these particles in C-H bond cleavage and enhance their activity toward neopentane hydrogenolysis.
ASA could also find application in reactions related to the biomass conversion such glycerol dehydration that requires mild acid sites6 and for the conversion of lignocellulosic biomass in aqueous reaction media.

Brønsted and Lewis acidity in ASA

As emphasize above, Brønsted and Lewis acidic properties of ASA have numerous and different applications in catalysis. Therefore, knowing and controlling the acidic properties of the catalysts are of great importance to understand the reaction mechanism and also to design more efficient acid catalysts. Hence, acid properties of ASA, such as acid type (Brønsted vs. Lewis), concentration, strength, location as well as cooperative effect between Lewis and Brønsted acids, attract considerable attention.

Brønsted acid sites

Zeolite, which play a dominant role industrial catalytic processes, generally have a higher catalytic activity than ASA and the structure of their Brønsted acid sites is well understood based not only on experiments8,9 but also on theory10). The strong Brønsted acidity resides in the bridging hydroxyl group originating from the replacement of a Si4+ in their crystalline framework by an Al3+ [≡Al–(OH)–Si≡].11 Zeolite acidity depends on the local environment and the Al content. Indeed as the number of Al3+ atoms in the second coordination sphere (i.e. in the shell of 12 or less T atoms surrounding the Al(OSi)4 site)increases, the strength decreases gradually.12 Zeolites with high Al3+ content exhibit a high density of acid sites of low strength, while on the other side, dealuminated zeolites have a low density of acid sites, but of higher acid strength. In addition, high Al3+ content favour the tendency of zeolites to produce extraframework aluminium species (EFAL), thus generating Lewis acid sites.12 The presence of Lewis acid sites can also influence the Brønsted acidity, thanks to a Lewis/Brønsted acid sites synergy that remarkably enhances the Brønsted acid strength.13-15 The acidity of the zeolites is also influenced by the dimension of the cavities. It has been found that for similar chemical composition, the strength of acid sites in medium pore size zeolites is higher than that found in large pore size zeolites.
In contrast to zeolite, due to the amorphous nature of these materials, acid sites in ASA are much less defined, and their structure is still far to be understood despite important contributions in the literature on this subject. The original structural models of ASA proposed that Brønsted acidity arise from proton compensating the electronic charge of the surface.29 Recently, many efforts have been made to study the Brønsted acid sites of ASA, based on probe molecule adsorption.5,18,22,26,27 Trombetta et al.28 suggested that, in mesoporous aluminosilicates, BAS originate from non-bridging SiOH groups in the vicinity of framework Lewis aluminium species. In presence of a basic molecule, a bridge is formed between the silanol and the Lewis aluminium (see Figure 1-1).
In 2005, Gora-Marek et al.30 reported the detection, on the spectra of two ASAs, of a weak band located at ca. 3600-3610 cm-1. They concluded to the presence of Si-OH-Al bridges in ASA and assigned the strong Brønsted acidity of ASA to these sites. In order to explain the weak intensity of this band, they proposed that it was due to a very low absorption coefficient for this band.
One year later, Busca et al.31 invalidated this hypothesis based on the fact that they could observe, on aluminated silica, sites of similar strength and in similar amount, independently of the presence of this band. Hence, they proposed that SiOH, Al3+ pair form a drawbridge that can be either closed (zeolite) or open (ASA). The closed drawbridge requires the rigidity of the zeolite framework to be stabilized and can therefore not exist in ASA.
In 2006, Crépeau et al.20 also concluded that the Brønsted acidity of ASA arose from a Si-OH, Al3+ pair but ruled out the hypothesis of Trombetta et al.28 concerning the formation of the Si-O–Al bridge upon adsorption of a base (at least when the adsorbed base is CO) based on the fact that Δν(OH) vs. ν(CO) of CO adsorbed on the acid sites of ASA follows the same linear correlation as in the case of zeolites. They also proposed that the detection of Brønsted acid sites with different strength is related to the number of tetrahedral Al atoms in the vicinity of the acidic silanol.
Figure 1-2 Brønsted and Lewis acid sites on the theoretical ASA surface model32
In 2009, Chizallet and Raybaud19 have proposed the first theoretical model for ASA. This model is based on the coating of a γ-alumina surface by a monolayer of silica. This modified surface has been optimized by periodic DFT. They identified pseudo-bridging silanols (PBS) (i.e. in electrostatic interaction with acceptor Al (PBS–Al) or Si (PBS–Si) (Figure 1-2), but not covalently bonded to it as they are in zeolite). When a strong enough base is adsorbed on these silanols, the pseudo-bridge (so-called drawbridge by Busca et al.) becomes a bridge that stabilizes the silanolate. Interestingly, the calculations of Chizallet et al. reconciliate the observation of Crépeau et al.20 vs. Trombetta et al.28. Indeed, based on these calculations, the closing of the drawbridge requires a strong enough base (such as lutidine), whereas, upon adsorption of a weak base such as CO, the drawbridge remains open. On the other side, based on their calculation, Chizallet et al. discarded the hypothesis of Crépeau et al. who suggested that the Brønsted acidity arose from silanols directly connected to a Lewis aluminium, as they indeed observed the presence of such species on their calculated ASA but, based on their interaction with basic probe molecules, these sites should be classified as very weak Brønsted sites. Nevertheless, one must keep in mind that the model of ASA they used (atomic film of Si on an alumina surface) is very different from the ASA commonly studied (where Al and Si are more intimately mixed and which usually contain a high fraction of silica). Baiker group also concluded that tetrahedrally coordinated aluminium in the vicinity of silanol groups results in moderate Brønsted acidity.33 According to them, sites with high Brønsted acidity (sites whose strength is similar or higher to that of Si-OH-Al bridges of zeolite) are due to silanols interacting simultaneously with a tetracoordinated and a pentacoordinated.

Lewis acid sites

Lewis acid sites (LAS), are considered to play a key role in numerous reactions in heterogeneous catalysis either by themselves (e.g. MPV reaction34) or by playing a synergetic role on the Brønsted Lewis sites.35-38 Reports on the structures of LAS of ASAs are very scarce. Hence we will present here briefly the literatures on LAS of zeolite.
LAS are associated with coordinatively unsaturated sites, and, for zeolites, very often assigned to Extra-Framework Aluminium species (EFAl). EFAl may occur in different forms such as Al3+, AlO+, Al(OH)2+, [Al(OH)2]+, AlO(OH), Al(OH)3, and Al2O3.39 Among these compounds, the cationic extra-framework aluminium species Al3+, AlO+, and Al(OH)2+ have been reported to act as strong Lewis acid sites. The other species (Al(OH)3, Al2O3 and AlO(OH)) are associated to a mild Lewis acidity, similar to the one of alumina.
Regarding the synergetic effect of Lewis acid sites on Brønsted acid sites in zeolite, the origin of this phenomena is usually assigned to the interaction of the EFAl with the electron pairs of the framework oxygen atoms, which leads to a delocalization of the electron density near the Brønsted acid sites, and hence to an increase acidity of this site (Figure 1-3 and ref.40). Besides framework Al, the presence of framework Lewis sites has also been proposed in zeolites. Basically this framework Lewis site should be a tricoordinated Al (formed via the dehydroxylation of the bridging OH), but the presence of this species remains difficult to ascertain. Indeed, although XANES (at the Al Kedge) and XPS experiments (Bokhoven et al.41,42) have provided clear evidences for the formation of distorted Al species, the exact nature of these species remains difficult to ascertain and the formation of distorted tetrahedral Al species by interaction of a tricoordinated Al with an oxygen from the zeolite framework has been proposed.43 Moreover, tricoordinated silicon species, whose formation has been postulated but never experimentally proven, could also contribute to the strong Lewis acidity.
Although little is known regarding LAS of ASA, EFAl species described for zeolite are probably not the main LAS in ASA because they require the stabilisation of the zeolite cages, and LAS sites of ASA are probably mostly related to the second type of LAS, i.e. framework Lewis sites.

Formation of ASA

The amorphous nature (and hence complex surface structure) of ASAs makes very difficult not only to the understanding of the origin of their Brønsted and Lewis acidities, but also to the accurate control of the acidic properties of these materials. Added to this, ASA can be prepared via a wide range of synthesis procedures and the choice of the synthesis method, together with the composition (SiO2/Al2O3 ratio) strongly influence the distribution of silicon and aluminium atoms through the particles and can lead to different acidic properties. Moreover, the comparison of the acidic properties of ASA obtained by different procedures is made difficult by the fact that different techniques can be used to evaluate their acidity (e.g. NH3-TPD, pyridine FTIR or catalytic activity in a model reaction). Recently, important developments in the synthesis of ASA via the sol-gel method,18,20,22,47-49 coprecipitation method,18,23,50 homogeneous deposition–precipitation,22 grafting process,17,50-56 flame-spray pyrolysis33 have been reported.
Crépeau et al.20 compared the acidic properties (based on FTIR of adsorbed pyridine and adsorbed CO) of samples prepared by impregnation (impregnation of alumina with silica, Siral series from Sasol) to those of samples prepared by cogelification of a small amount of aluminium with a high amount of silicon precursors. They concluded that cogelation leads to more homogeneous samples with higher Brønsted acidity (both in terms of number and of strength). Hensen et al.22 studied the acidity (n-heptane, FTIR of adsorbed pyridine) of aluminosilicates prepared by homogeneous deposition precipitation (pH jump by urea thermal decomposition) of aluminium nitrate on silica. They concluded to the presence, in these ASA, of very small fraction of strong Brønsted acid sites, similar to those of zeolites. Moreover, they observed that a higher temperature of calcination results in a higher amount of strong Brønsted acid that they associated to the diffusion of Al atoms into the silica network. Baiker group33 investigated an original preparation route for the synthesis of ASA based on flame-spray pyrolysis of a solution of aluminium acetylacetonate and tetraethoxysilane. These ASA exhibit a tunable Brønsted acidity ranging from moderate to zeolite-like depending on the aluminium content.
Controlled grafting of aluminium alkoxide or silicon alkoxide precursors respectively onto a silica or alumina support is also an efficient synthesis method to prepare ASA. Mokaya et al.55 synthesized ordered mesoporous aluminosilicate (Al-MCM-41) by grafting aluminium isopropoxide on silica in nonaqueous media or by grafting aluminium chlorohydrate in aqueous solution. These materials exhibited higher Brønsted acid content compared to Al-MCM-41 prepared via the conventional direct synthesis method. Caillot et al. prepared Al/SiO2 and Si/Al2O3 by grafting aluminium (respectively silicon) alkoxides on silica respectively alumina under various conditions: liquid-phase (water free or water-added conditions) and gas-phase. They investigated the influence of several parameters on the deposition and came to the following conclusions: (1) for grafting in anhydrous liquid phase conditions, when the density of OH groups of the support is no longer the limiting factor, deposition is saturated when the surface of the support oxide has been completely covered with a monolayer of precursor molecules; (2) following a first grafting, new species can be grafted after a calcination step that allows recovering of hydroxyl groups; (3) temperature and water/precursor ratio can control the structure of the deposit.50 In another study of the same team, the acidity of Al/SiO2 prepared by grafting aluminium isopropoxide on silica was examined and results suggested that the nature of the Brønsted acid sites strongly depends on the method used for ASA synthesis.52 For synthesis performed in presence of added water, grafting was less controlled leading to alumina agglomerates possessing weak and medium Lewis acid sites but no Brønsted acidity.52 This team also found that ASAs prepared by cogelation has Brønsted acid sites with strength similar to those in zeolites, while grafted ASAs possess specific Brønsted acid sites that differ from those of zeolites.18 More recently, this team performed DNP NMR experiments and DFT calculation on ASA obtained by grafting either Al on silica or Si on alumina. Based on these analyses, they proposed a detailed atomic level description of the preparation procedure and formation of Brønsted acid sites: on silica and alumina surfaces, molecular aluminium and silicon precursors are, respectively, preferentially grafted on sites that enable the formation of AlIV and SiIV interfacial sites and the genesis of Brønsted acidity is related to the formation of SiIV-O-AlIV linkages.
This analysis of the literature related to the preparation of ASA highlights the complexity of these amorphous materials and clearly shows that the understanding the acidic properties of ASA (nature, strength, concentration) and their relation with the synthesis procedure is of great importance to develop a synthetic route for the design novel solid catalysts with improved acidic properties. It also indicates that this goal can probably only be achieved by a carefully spectroscopic analysis of the environment of the Al and Si atoms.

Acidity Characterization

Various techniques can be used for acidity characterization. The traditional methods performed in solution (such as Boehm titration7) may give information about acid strength and concentration but their applications is limited due to (i) the fact that the measurement is performed in water (and hence, the acidity of the solid evaluated in these conditions may differ significantly from the acidity in gas phase) (ii) the slow diffusion, in aqueous solution of the probe into the micro- or mesopores of the solid. They remain very interesting when it comes to characterize the acidity of aluminosilicates for application in aqueous solution.7 Characterisation of the acidity by adsorption of a basic probe molecule can also be performed in the gas phase. The probe molecules that are most usually used are, by increasing proton affinity (PA): CO (PA = 594 kJ/mol); ammonia (PA = 854 kJ/mol); pyridine (PA = 930 kJ/mol); lutidine (PA = 963 kJ/mol). Two approaches can be followed to detect the probe molecule: either a simple detection (TCD or mass spectrometry) of the molecule in the gas phase to follow its desorption during a temperature ramp, or a detection of the adsorbed probe molecules (mostly by FTIR). Probe molecules are especially useful for the quantification of the acid sites (either only the total amount, or a more detailed evaluation, e.g. BAS vs. LAS or strong vs. weak). Beside the characterization of the acidic properties of the solid by probe molecules, a direct spectroscopic characterization of the solid can also be performed. In this domain, beside FTIR (direct observation of the OH groups via the ν(OH) stretching band), high resolution solid state nuclear magnetic resonance (SSNMR) is a powerful tool for the determination of the geometry of the acid sites.

Characterization of the structure of the acid sites

SSNMR

Solid state NMR spectroscopy is undoubtedly one of the most powerful techniques for determining molecular-level structure of solids. Solid state NMR spectroscopy plays a key role thanks to the developments of methodologies appropriate to the investigation of quadrupolar nuclei (for more than 63% of the elements, the only NMR active isotopes are quadrupolar (absence of I = 1/2 isotope)). NMR is element selective, and the signal of the element will depend on its local environment. For aluminosilicates, the elements of interest are 1H (S = 1/2, natural abundance 99.98%), 27Al (S = 5/2, natural abundance 100%), 29Si (S = 1/2, natural abundance 4.7%). Another isotope of interest is 17O (S = 5/2) but, due to its very low natural (0.037%), its observation generally requires the preparation of enriched samples. To get an image of the acid sites, the observation of a single element is usually not enough and the recent developments in heteronuclear correlation bring new opportunities in the characterization of these sites.
27Al MAS NMR have been extensively applied to determine local structure of aluminium sites, providing various information about aluminium coordination numbers. 27Al with 100% natural abundance and a rather high gyromagnetic ratio, has a relatively high NMR sensitivity. But as a quadrupole nucleus with spin quantum number I = 5/2 and a relatively large quadrupolar moment (eQ = 14.66×10-30 m2), the central transition of the Al NMR line is, under MAS conditions, broadened by the second-order quadrupole interactions. As a consequence, the Al signals can be rather large, and a part of the signal can become invisible (so-called NMR invisible aluminium). Moreover, chemical shift in 1D spectrum may vary with the magnetic field strength. The magnitude of the coupling (and hence, the broadening of the NMR signal) of a quadrupolar nuclei depends on the symmetry of the electric field gradient around the nuclei and hence on the symmetry of the distribution of charge of mostly (but not only) the coordinating oxygen atoms. For perfect tetrahedral or octahedral symmetry the CQ value is 0 but it increases gradually with the decrease in the symmetry (Figure 1-4).
For high surface area aluminosilicates, the hydration level of the sample modifies dramatically the spectra. Figure 1-5 shows the 27Al MAS NMR spectra of an amorphous aluminosilicates recorded without any specific pretreatment (and hence partially hydrated, H-MSA), fully dehydrated (dried H-MAS) and fully hydrated (H-MSA EQ60d). On the spectra recorded after drying, a broad signal is observed whose intensity corresponds to only 20% of expected one (large fraction of “invisible Al”), and, on the spectra recorded without specific pretreatment, still 50% of the NMR signal is invisible. Only a complete rehydration (or the adsorption of a basic molecule such as NH3) allows a quantitative observation. This is the reason why, the aluminosilicates samples are usually rehydrated prior to the measurement of the 27Al MAS NMR spectra of aluminosilicates.
This loss of signal for fully dehydrated aluminosilicates (and its recovery by rehydration or adsorption of basic molecules) has also been observed for zeolites.59,60 The dramatic broadening of the peak associated to tetrahedral aluminium upon dehydration is assigned in zeolite to the formation of the Si-OH-Al bridge. Koller et al.61 reported that the electric field gradient that surrounds an Al atom in a zeolite (calculated using DFT), is considerably larger for the protonated site (Al(OHSi)(OSi)3 than for the unprotonated one ([Al(OSi)4-, H3O+]) due to the weakening of the Al-O bond upon protonation.
On the spectrum of most ASAs, three signals are usually observed:
1) The peak observed at around 60 ppm is assigned to tetrahedral aluminium species, but, at variance with zeolites, only a small fraction of the aluminium atoms in tetrahedral coordination show Brønsted acidic properties (see Figure 1-6). Because of this, the term “framework aluminium” that is often used to qualify tetrahedral aluminium in ASA should be avoided as it is misleading.
2) The band at ca. 0 ppm is assigned to octahedral aluminium. The presence of Al in octahedral coordination is sometimes associated to the presence of transitional alumina domains or at least to some polymeric form of Al, as octahedral Al is the dominant coordination in many transitional aluminas62; octahedral Al is therefore often described as “extraframework Al”. Although this assertion is mostly true for alumina-rich ASA, octahedral Al in silica-rich ASA has been shown partially25,63,64 or even fully (see Figure 1-5) convert to tetrahedral Al upon adsorption of NH3. The changes in the coordination of octahedral Al upon exposure to ammonia indicates that these Al are still connected to the framework via oxygen atom(s). Omegna et al.25 concluded that the flexible coordination is limited to the aluminium atoms which are part of an amorphous silica alumina character, as no change in Al coordination upon adsorption of NH3 is observed for γ-alumina. This reversibility in Al coordination in presence of NH3 was used for the evolution of the fraction of Al in silica alumina domains in ASA.25,65 3) The third signal is centered at 30 ppm. Two possible assignments have been initially proposed: either five coordinated Al or distorted tetrahedral species but MQ-MAS experiments (see below) have allowed assigning this signal to pentacoordinated Al. 5-coordinated Al are also present in gamma alumina and are associated in this oxide to surface Al atoms.66 In Al rich ASA, these AlV can be related to the presence of an alumina phase and be associated with surface sites of this alumina phase. However, these sites are also observed in Si-rich ASA. De Witte et al.67 associated the presence of AlV to the interface between the alumina and the silica-alumina phase. They observed, using 27Al MAS NMR spectroscopy, that the proportion of this phase increases upon dehydration of the sample and also upon calcination (for 27Al NMR spectra recorded on hydrated samples). However, Wang et al. came to a different conclusion as they observed, in ASA prepared by flame spray pyrolysis, a high population of homogenously dispersed AlV in the absence of any bulk alumina phase and these sites were accessible to guest molecules.68 Besides, these AlV have been shown to contribute to the Brønsted acidity of the ASA (based on 27Al DHMQC experiments that indicate a close proximity of NH4+ cations with these sites).
As mentioned above, 27Al MAS NMR spectroscopy provides information about the coordination of the aluminium species. However, the spectral resolution is reduced by the second-order quadrupolar interaction that can only be partially averaged by MAS. This interaction also shifts the position of the central transition resonance away from its isotropic chemical shift. In 1995, Frydman and Harwood70 introduced multiple quantum (MQ) 27Al MAS NMR spectroscopy. This experiments produces a 2D spectrum that allows to obtain, not only a spectrum free of quadrupolar interaction (F1 dimension), but also information on the quadrupolar parameters of the nuclei (F2 dimension). In this sense, what was the main drawback of quadrupolar nuclei, becomes an advantage as it allows differentiating atoms as a function of their quadrupolar parameters. Moreover, the values of the isotropic chemical shifts and quadrupolar parameters determined from the analysis of the MQ MAS NMR spectra can be used to simulate the 27Al MAS NMR spectra, leading to quantitative information on the aluminium species.Furthermore, as the quadrupolar parameters are related to the electric field gradient around the observed atom, it is possible to relate them to the local structure of this atom (usually based on theoretical calculation).
Figure 1-7 shows the 27Al MQ MAS spectrum of a dehydrated HZSM-5.71 This experiment allows differentiating 4 types of aluminium environments: two weak peaks associated to hexacoordinated and pentacoordinated Al and indicative of a slight extraction of framework aluminium and two types of tetracoordinated Al with close chemical shifts but very different CQ value. The peak with the higher CQ value (and hence the less symmetrical environment) can be assigned to a framework Al which is charge compensated by a H+; the peak with the lower CQ value (hence corresponding to the most symmetrical environment) can be assigned to a framework Al which is charge compensated by a cation other than H+. This large difference in CQ between these two tetrahedral environments is due to the fact that when the charge compensating cation is a proton, it corresponds to the formation of Si-OH-Al bridge and leads to one longer Al-O bond.
On highly dehydroxylated aluminosilicate, one could expect a forth type of coordination for Al with the formation of tricoordinated aluminium. However, the observation of these species by 27Al MAS NMR has not been reported until very recently. In a recent work Brus et al.73 proposed, based on MQ MAS experiments and theoretical calculations, that framework trigonal Al in zeolites would be associated with a peak with the following parameters: δiso = 65 ppm, CQ = 20 MHz and η = 0.1. However, this assignment is questionable as the CQ value is close to the one of the Al engaged in Si-OH-Al bridges, and the values of CQ reported by these authors for the Al of Si-OH-Al bridges (ca. 3.5 MHz) is not consistent with the numerous reports for this type of sites in the literatures.60,71,74 The 27Al spectra were probably misinterpreted in this work, likely due to an inappropriate sample preparation (adsorption of water molecules).
The properties of proton of surface OH groups of ASA are mainly influenced by local effects, such as oxygen coordination, bond geometry, and type of metal atoms in their local structure. 1H NMR chemical shifts in aluminosilicates are illustrated in Table 1-1.
Signals at δ1H = 1.2 − 2.3 ppm indicate the presence of SiOH, which may shift slightly due to interaction of their hydrogen with a neighboring oxygen (H-bond donor). Isolated terminal silanol groups appear at 1.9 ppm, and move to 2.1 ppm upon increasing temperature from 20 to 400oC.75 In zeolites, silanol nests (produced by dealumination with alkaline solution) appear at 2.2 ppm. In amorphous silica, the presence of three types of silanols was evidenced by 1H NMR, located at 1.8, 2.9 and 4.0 ppm. The first peak was assigned to isolated silanols, whereas the two others were assigned to H bonded (H-donors) silanols.76 Yesinowski established a linear relationship between the distance between the two oxygens involved in the H-bond (picometer) and the chemical shift of the proton (ppm): Δδ1H = 79.05 − 0.255 d(OH−O).77 Hence, very high 1H chemical shift can sometimes be observed in zeolite and silicates, indicative of very short H-bond. This is, for example, the case in phyllosilicates such as RUB-18 (Ilerite) for which a 1H peak is observed at a chemical shift of 10.4 ppm which is associated with a ≡Si-OH…-O-Si≡H bond of ca. 250 pm.78
In zeolites, AlOH group of extra framework Al are responsible for peaks at 0.4 − 0.6 or 2.6 − 3.6 ppm.84 Similar chemical shifts have been reported for surface OH in gamma alumina by Delgado et al: a chemical shift of about 0.4 ppm for terminal Al-OH (µ1Al-OH) and of about 2.7 for bridging Al-OH-Al (µ2AlOH). Similarly to silanols, the position of these peaks can be shifted toward higher chemical shift (low-field resonance shift) if they are involved in a H bond.
Based on 1H MAS NMR experiments, bridging Si-OH-Al hydroxyls have only been reported in zeolite and their presence has never been clearly ascertained in ASA. Zeolite pore channels influence the 1H chemical shift of the acidic proton. For instance, the chemical shift of Brønsted bridging OH groups located in large cages and channels is in the range 3.6 − 4.3 ppm, while, when located in small cages and channels, it is shifted to lower field.
Generally speaking, 1H chemical shift can be regarded as an index of acid strength: the higher the 1H chemical shift of hydroxyl groups over zeolite, the higher the acid strength, since the hydroxyl groups are isolated over zeolites (a higher acidity corresponds to a higher atomic charge on the hydrogen. This induces a reduction in the shielding of the external magnetic field on this hydrogen and hence a higher chemical shift91). The higher acid strength of silanol groups in the vicinity of framework aluminium atoms in mesoporous aluminosilicates in comparison with terminal SiOH groups is caused by the flexible local structure of these surface sites.
Adsorption of ammonia as probe molecule is a very suitable method to highlight and quantify strong Brønsted acid sites. Upon formation of ammonium ions (NH4+) by protonation of the ammonia molecules by strong enough acidic OH groups, a narrow 1H MAS NMR signal appears at δ1H = 6.5~7.0 ppm33,82,86 which cannot overlap with most of the OH signals and whose intensity is four times larger compared with that of the 1H NMR signal of the acidic OH in the unloaded material. This can be used to quantify the number of accessible acid sites with high accuracy.
In order to attribute signals in 1H NMR spectra of solid, exploring the correlation between 1H chemical shifts and the FTIR wavenumber of the O-H stretching vibration bands of hydroxyl groups over various zeolites is of significant meaning. Several studies on this subject have been reported in the literature.79,91,92 Based on these studies, a linear relationship between 1H chemical shifts and the wavenumber of the O-H was established for non-interacting hydroxyl groups: δ1H / ppm = 57.1 – 0.0147υOH/cm-1.
FTIR spectroscopy have received considerable attention mainly as a tool to discriminate hydroxyl groups. The frequencies of the IR bands of the O-H stretching vibrations in the range of 3800-3200 cm-1 are used for primary characterization of the types of hydroxyl groups. Generally, four main bands appear in the region of OH stretching vibrations:
1) The band at 3745 cm-1 is characteristic of silanols located on the outer surface.93-97
2) the bands in the range ca. 3530 to 3620 cm-1 are attributed to bridging Si-OH-Al groups (strong Brønsted acid sites).79,96
3) the very broad IR adsorption band around 3500 cm-1, is assigned to nest silanol groups94,95 or Si-OH groups strongly disturbed by multiple hydrogen bonding24,98 or H2O interacting with the oxygen atoms of the framework.
4) the bands at 3780, 3705 and 3670 cm-1 are attributed to OH groups bonded to extraframework aluminium.

Quantitative evaluation of the acid sites

TPD (NH3 and other bases)

The most direct way to evaluate the acidity of the sample is to adsorb a base on it and follow its desorption vs temperature (Temperature Programmed Desorption – TPD experiments). Ammonia is, by far, the most commonly used base, but the use of other basic molecules such as isopropylamine,5 pyridine65 and ethanol18,52,100 has also been proposed.
Ammonia TPD is usually used for primary characterization of acidic properties. Ammonia is a strong base which is an appropriate probe for most OH groups, even in zeolites with small cavities, thanks to its small size ( 3.70 × 3.99 × 3.11 Å3).101 Experimentally, its simplicity and its low cost, make ammonia TPD a convenient method for quick determination of the acidic properties. The desorbed amount of ammonia gives information about the amount of acid sites, whereas information on the strength are obtained by the analysis of the desorption temperature. Unfortunately, quantitative determination of the acid strength distribution may be complicated and measurements made using different experimental conditions are often not easily reconciled. The desorption peak has an unsymmetrical shape (related to the equilibrium between adsorbed NH3 and NH3 in the gas phase) and the peak shape and position depend on the experimental conditions (mass of sample, heating rate or gas flow), as well as on kinetic phenomena such as readsorption and/or slow diffusion. The fitting of the peaks is therefore very challenging.
Nevertheless, under identical experimental conditions, the peak position gives comparative information about the relative acid strength and the width of the peak about the distribution of strength. Ammonia is usually adsorbed at 150oC to eliminate the contribution of very weak acid sites and improve the spectrum resolution. Generally, two desorption peaks are observed, indicating the existence of at least two different types of adsorbed NH3.104,105 The first one, between 170 and 280oC is referred to as the low temperature (LT) and is ascribed to the desorption of ammonia from acidic hydroxyls106 or weak Lewis acid such as Al(OH)2+ and Al(OH)2+107, it is also sometimes assigned to over-adsorbed NH3 (i.e. NH3 molecules adsorbed on NH4+)102. It is usually considered as related to sites with no catalytic importance.106,108 The second one, located between 350 and 440oC, is called the high temperature peak (HT), and usually attributed to ammonia desorption from strong Brønsted and Lewis acid sites.
However, the overlapping of the LT and HT peaks introduces some confusion in the interpretation of the TPD curves. The two peaks have to be separated to obtain the concentration of each type of acid site. Karge et al.109 have reported a curve-fitting method to analyze TPD curve according to a simple Gaussian function based on the assumption of no readsorption of ammonia. However, the simulation is not robust and produces ambiguous results. Alternatively, the HT and LT peaks can be fitted with Haarhoff-Van der Linde (HVL) functions, but, although the use of this function allows a good fit for the experimental data, it can still lead to ambiguous results.
It is clear that ammonia TPD is a useful technique for detection of the total acidity. But further distinguishing BAS and LAS far exceeds its power. Despite some limitations, it is still an ideal method to make a first evaluation of the acidity.

FTIR of adsorbed probe molecules

What differs between a probe molecule adsorbed on a Lewis and a Brønsted site is the mode of interaction of the probe with the site. For a strong base (e.g. pyridine or NH3), the adsorption on a Brønsted site is associated with the protonation of the base (provided that the acid site is strong enough) or the formation of H-bond between the probe and the acid site (for probes with weak basic properties). On a Lewis site, the adsorption of a base leads to a coordination bond.
Figure 1-8 illustrates, with the example of pyridine, the different types of bonds that can be formed between the probe and the acid sites. The spectroscopic detection of the mode of adsorption of a base requires that different modes of adsorption have different signature. Although several other technics can be used (e.g. NMR or XPS), FTIR remains the choice spectroscopy for this application.
Pyridine is the most commonly used probe molecule for the investigation of the properties of acid sites. Pyridine presents the advantages of giving rise to very distinct vibrational bands to distinguish LAS from BAS and allows therefore to determine the concentration of Lewis and Brønsted sites using the absorption coefficients of these two bands.
Adsorption of pyridine is performed by saturation of the sample with pyridine and evacuation at increasing temperature, allowing therefore to also extract information on the strength of the acid sites (the stronger sites corresponding to the higher desorption temperature).
The bands due to the 19b and 8a vibration modes (Table 1-2) are very sensitive to coordination on the nitrogen and are commonly used to identify and quantify the sites of BAS and LAS: wavenumber of υ19a band at about 1450 cm-1 is characteristic of LAS while υ19b band at about 1550 cm-1 is for BAS. The shift of 8a mode is also sometimes used to measure the strength of LAS (the stronger the LAS, the higher the frequency).20 Activation temperature changes the nature of the acid sites of ASA, due to dehydroxylation reactions: the higher the temperature of activation, the more LAS and the less BAS.

Table of contents :

CHAPTER 1. BIBLIOGRAPHIC REPORT 
1.1 Applications of ASAs in catalysis
1.2 Brønsted and Lewis acidity in ASA
1.2.1 Brønsted acid sites
1.2.2 Lewis acid sites
1.3 Formation of ASA
1.4 Acidity Characterization
1.4.1 Characterization of the structure of the acid sites
1.4.1.1 SSNMR
1.4.1.2 FTIR
1.4.2 Quantitative evaluation of the acid sites
1.4.2.1 TPD (NH3 and other bases)
1.4.2.2 FTIR of adsorbed probe molecules
1.4.2.3 Quantification of Protons
1.5 Model reaction for Brønsted acidity: Isomerization of 33DMB1
1.6 Reactivity of aluminosilicates toward water
1.6.1 Role of water on aluminium coordination
1.6.2 Role of water to acidity and reactivity
1.7 Objective and scope of this work
CHAPTER 2. EXPERIMENTAL AND CHARACTERIZATION
2.1 Materials
2.1.1 Commercial ASA
2.1.2 Commercial silica gel
2.1.3 Preparation of high surface area silica
2.2 Characterization
2.2.1 XRF
2.2.2 XRD
2.2.3 TGA
2.2.4 N2 physisorption
2.2.5 NH3-TPD
2.2.6 FTIR
2.2.6.1 FTIR of adsorbed pyridine
2.2.6.2 FTIR of adsorbed CO
2.2.7 Solid state NMR
2.2.7.1 1D MAS NMR
2.2.7.2 Quantification Proton
2.2.7.3 Homonuclear NMR
2.2.7.4 Heteronuclear NMR
2.2.8 Isomerization of 33DMB1 reaction
CHAPTER 3. DEALUMINATION OF AMORPHOUS SILICA-ALUMINAS WITH ACETYLACETONE AND ITS INFLUENCE ON ACIDITY AND ACTIVITY 
3.1 Introduction
3.2 Experimental
3.2.1 Synthesis
3.2.1.1 Starting ASA
3.2.1.2 ASA dealuminated with acetylacetone
3.2.2 Characterization and catalytic test conditions
3.3 Results
3.3.1 Textural properties
3.3.2 Acidic Properties
3.3.3 Catalytic Performance
3.3.4 Aluminium coordination
3.3.5 Hydroxyl groups
3.4 Discussion
3.5 Conclusion
CHAPTER 4. DEALUMINATION OF AMORPHOUS SILICA-ALUMINAS WITH CITRIC ACID AND ITS INFLUENCE ON ACIDITY AND ACTIVITY 
4.1 Introduction
4.2 Experimental
4.2.1 Synthesis
4.2.1.1 Starting ASA
4.2.1.2 ASA dealuminated with citric acid
4.2.2 Characterization and catalytic test conditions
4.3 Results
4.3.1 Effect of dealumination on composition (Si/Al ratio) and texture
4.3.2 Acidic Properties
4.3.2.1 NH3-TPD
4.3.2.2 FTIR of adsorbed pyridine
4.3.2.3 FTIR of adsorbed CO
4.3.2.4 Characterization of the acidic properties: conclusion
4.3.3 Catalytic Performance for the isomerization of 3,3-dimethylbut-1-ene
4.3.4 Aluminium coordination
4.3.5 Hydroxyl groups
4.4 Discussion
4.4.1 Mechanism of dealumination with CA
4.4.2 Effect of dealumination on the acidic properties of the ASA
4.5 Conclusion
CHAPTER 5. IDENTIFICATION OF THE ACID SITES ON DEALUMINATED AMORPHOUS SILICA-ALUMINAS BY SOLID STATE NMR 
5.1 Introduction
5.1 Experimental
5.1.1 Synthesis
5.1.2 Characterization
5.2 Results
5.2.1 Brief overview of the characteristics of the selected ASAs
5.2.2 Characterization of the ammonium form ASA
5.2.2.1 27Al MAS NMR
5.2.2.2 1H MAS NMR spectra of ammonium ASA
5.2.2.3 Heteronuclear 27Al-1H MAS NMR
5.2.3 Characterization of H form ASA
5.2.3.1 27Al MAS NMR
5.2.3.2 1H MAS NMR
5.2.3.3 Heteronuclear 27Al-1H NMR
5.3 Summary of NMR results
5.3.1 NMR results on ammonium forms of the ASAs:
5.3.2 NMR results on H forms of the ASAs:
5.4 Discussion and conclusion
CHAPTER 6. SYNTHESIS OF ASA WITH IMPROVED ACIDITY BY GRAFTING OF AL PRECURSORS 
6.1 Introduction
6.1.1 Parameters relevant for a controlled grafting of isolated Al
6.1.1.1 Characteristics of the silica support
6.1.1.2 Choice of the Al precursor
6.1.1.3 Conditions for grafting
6.1.2 Objectives
6.2 Experimental
6.2.1 Synthesis
6.2.1.1 Grafting Al(OPri)3 and substituted aluminium alkoxides
6.2.1.2 Grafting of Triisobutylaluminium (TiBA)
6.2.1.3 Grafting of Diisobutylaluminium hydride (DiBAH)
6.2.2 Characterization and catalytic test conditions
6.3 Results
6.3.1 ASA grafted with Al(OPri)xL3-x
6.3.1.1 Composition and textural properties
6.3.1.2 Acidic Properties
6.3.1.3 Aluminium coordination
6.3.1.4 Hydroxyl groups
6.3.1.5 Conclusion
6.3.2 ASA Grafted TiBA
6.3.2.1 Composition and porous properties
6.3.2.2 Acidic Properties
6.3.2.3 Aluminium coordination
6.3.2.4 Hydroxyl groups
6.3.2.5 Conclusion
6.3.3 ASA Grafted DiBAH
6.3.3.1 Composition and porous properties
6.3.3.2 Acidic Properties
6.3.3.3 Aluminium coordination
6.3.3.4 Hydroxyl groups
6.3.3.5 27Al-1H D-HMQC 2D NMR
6.3.3.6 Conclusion
6.3.4 Catalytic Performance for the isomerization of 3,3-dimethylbut-1-ene
6.4 Discussion
6.5 Conclusion and perspectives

GET THE COMPLETE PROJECT