Catalytic properties of Thallium-containing mesoporous silicas

2.1 Catalysts preparation

The catalysts Tl-HMS-n (where n is the Si/Tl ratio in the precursor gel = 55, 35, 15) have been prepared following the pathway reported by Tanev et al. (1994). In a representative preparation, hexadecylamine (HDA) was added to a solution containing water and ethanol (EtOH) and the mixture was stirred until homogeneous. Then tetraethyl orthosilicate (TEOS) was added under vigorous stirring. The metal precursor Tl(NO₃)₃ dissolved in TEOS itself.

This solution was then stirred at room temperature for 24 h to obtain the products. The solids were recovered by filtration, washed with distilled water, and air-dried at 393 K.

Organic molecules occluded in the mesopores were removed by solvent extraction. The dried precursor was dispersed in ethanol (5 g/100 ml) containing a small amount of NH₄Cl (1 g/100 ml) and the mixture was refluxed under vigorous stirring for 2 h. The presence of ${\text{NH}}_4^{+}$ cations in EtOH was reported to be necessary to exchange protonated amines formed during the synthesis and balance the excess of negative charges resulting from the substitution of In^III for Si^IV (Tuel et al., 1998; Echchahed et al., 1997). The solid was then filtered and washed with cold ethanol. The extraction procedure was repeated twice before drying the samples at 393 K in an oven. Finally the samples were calcined at 823 K in air for 6 h.

2.2 Characterization of the samples

Powder X-ray diffraction patterns were recorded on SIEMENS D500 diffractometer with Cu-Kα radiation. The chemical compositions of the samples were determined by atomic absorption. The N₂ adsorption/desorption isotherms were measured on samples that were preliminarily outgassed at 523 K overnight using a Catasorb apparatus.

2.3 Catalytic testing

The benzylation of benzene by benzyl chloride has been used as a model reaction for Friedel–Craft alkylation catalytic properties. To avoid diffusional limitations, the solids were crushed below 50 μm and the reaction was carried out in a batch reactor between 333 and 363 K. The quantity of 100 mg of the solids was tested after an activation consisting of a heat treatment under air (2 l h⁻¹) up to 573 K. Directly after cooling, the catalysts were contacted under stirring with a solution of 25 ml of benzene and 6.48 or 2.16 ml of benzyl chloride to obtain a benzene to benzyl chloride mole ratio of 5 or 15. The conversion of benzyl chloride was followed by analyzing samples of the reaction mixture collected at regular intervals by gas chromatography using a gas chromatograph equipped with a flame ionization detector FID and a capillary column RTX-1 (30 m × 0.32 nm i.d.). The selectivity is expressed by the molar ratio of formed diphenylmethane to converted benzyl chloride.

3.1 Characterization

The results of the chemical composition and characteristics of the catalysts are given in the Table 1. The Thallium compositions of the solids corresponded relatively well to those fixed for the synthesis except at low Thallium content (Tl-HSM-55) where a loss of Thallium was observed.

Most of the values of the specific surface areas of the solids were larger than 1000 (m² g⁻¹), which was typical of mesoporous materials (Beck et al., 1992; Huo et al., 1994). When the Thallium content increased, they decreased slightly. The N₂ adsorption/desorption isotherms show that the different solids contain regular pores of ca. 3.7 nm diameter.

The X-ray powder diffraction patterns of the solids showed a broad peak at (2θ) = 2 (Fig. 1) characterizing a mesoporous material not well-crystallized. The intensity of the peak decreased very slightly when the Thallium content increased showing that the addition of Thallium has not a negative effect on the crystallinity. Furthermore, no peak is observed in the 10 to 80° (2θ) (Fig. 2). Indeed, no modification in the X-ray patterns or N₂ isotherms was observed throughout the series of samples, indicating that the materials do not lose their crystallinity with Tl incorporation.

Close

Figure 1

DRX patterns of the Tl-HMS-n catalysts in the domain of 1–10° (2θ). n = Si/Tl = (a) 55, (b) 35, (c) 15.

Export to PPT

Close

Figure 2

DRX patterns of the Tl-HMS-n catalysts in the domain of 30–80° (2θ). n = Si/Tl = (a) 55, (b) 35, (c) 15.

Export to PPT

3.2 Reaction kinetics

The kinetic data for the benzene benzylation reaction in excess of benzene (stoichiometric ratio Bz/BzCl = 15) over the Tl-HMS-35 catalyst could be fitted well to a pseudo-first-order rate law: $$\mathit{Log}[1/1-x]=(k_a/2.303)(t-t_0)$$ where k_a is the apparent first-order rate constant, x the fractional conversion of benzyl chloride, t the reaction time and t₀ the induction period corresponding to the time required for reaching equilibrium temperature. A plot of log [1/1-x] as a function of the time gives a linear plot over a large range of benzyl chloride conversions.

The effect of temperature on the rate was studied by conducting the reaction at 343, 348, and 353 K under the standard reaction conditions (stoichiometric ratio PhH/BnCl = 15 and 0.1 g catalyst). The results showed that the catalytic performances of the catalyst strongly increased with the reaction temperature (Table 2). Indeed, the time for 50% conversion of benzyl chloride and the apparent rate constant K_a changed from 120.6 min and 132 × 10⁻⁴ min⁻¹ at 343 K to, respectively, 20.1 min and 338 × 10⁻⁴ min⁻¹ at 353 K.

By contrast, the selectivity to diphenylmethane decreased from 100 to 73%. The activation energy estimated thus obtained was 98.5 kJ mol⁻¹.

Two Bz/BzCl ratios have been investigated. The results obtained are reported in Table 3. It appears that the stoichiometric ratio between benzene and benzyl chloride has a strong influence on the selectivity to diphenyl methane. With a low ratio, the secondary reactions to dibenzylbenzenes and tribenzylbenzene were favored.

The effect of substituents was investigated using several aromatic substrates, with the results reported in Table 4. If the reaction was acid catalyzed a correlation of the Hammett type would have been expected, i.e. log K_a = log K_a0 + σ⁺ρ, in which K_a0 is the rate constant for benzene and σ⁺ a coefficient representing the changes of reactivity due to the substituent, and ρ a constant related to the charge in the intermediate complex (March, 1985).

In the present case, only a small change of K_a was observed. This suggested a mechanism different from the usual acid mechanism. The high activity observed with these catalysts could then be ascribed to a different initiation of the reaction, for instance the homolytic rupture of the carbon-chlorine bond followed by the oxidation of the radical: $$\varphi -{\text{CH}}_2\text{Cl}\to \varphi -{\text{CH}}_2^{•}+\text{Cl}•\text{,}$$ $$\varphi -{\text{CH}}_2^{•}+{\text{Tl}}^{3+}\to \varphi -{\text{CH}}_2^{+}+{\text{Tl}}^{+}\text{,}$$ $${\text{Tl}}^{+}+{\text{Cl}}^{•}\to {\text{T}}^{3+}+{\text{Cl}}^{-}$$ Indeed, this homolytic rupture of the carbon–chlorine bond should be the rate-determining step. This mechanism is similar to that proposed earlier for the alkylation and acylation reactions (Choudhary and Jana, 2001; Bachari et al., 2010, 2011; Hentit et al., 2007; Tahir et al., 2008; Benadji et al., 2010; Merabti et al., 2010; Brio et al., 2001; Choudhary and Jana, 2002).

In order to rule out the influence of a steric effect on the rate of reaction, we have applied the Taft relation (March, 1985). According to this relation when a steric effect influences the reaction, there is a linear relation between the rate and the parameter E_s values considered to be representative of the size of the substituting group of the studied aromatic compounds. Using the E_s parameter tabulated by Charton (Charton, 1975) we have shown that such a relation did not exist.

It was interesting to compare the solids with TlO_X/zirconia (LS) Choudhary and Jana, 2001. Both systems reached a final conversion of 100% with complete selectivity to the monoalkyl group; the half reaction time was about 3.8 min for TlO_X/zirconia (LS) and was here about 20 min, so that Tl-HMS is less active probably due to the higher amount of Thallium in TlO_X/zirconia (LS).

Furthermore, the catalyst Tl-HMS-35 is always active and selective for larger molecules like naphthenic compounds such as methoxynaphthalene (Table 4). The large pores of the mesoporous support permit the conversion of these molecules that could not be done on other supports.

3.3 Catalytic performances of Tl-HMS materials in the alkylation of benzene

A comparison of the catalytic properties of the solids tested is presented in Table 5. The pure silicic compound (HMS) and the compound containing less Thallium (Tl-HMS-55) were totally inactive. The other compounds showed an activity increasing with increase of their Thallium content. However, the selectivity to diphenylmethane at complete conversion of benzyl chloride decreased while the Tl content increased.

3.4 Effect of water

Lewis acids are sensitive to water and the effects of water on the catalytic activity were investigated using Tl-HMS-35 at 353 K with a ratio PhH/BnCl = 15, adding different amounts of water. The results are reported in Table 6. A small addition of water had almost no effect on the catalytic properties of the compound whereas a larger addition had a drastic one with a decrease both of the activity and of the selectivity. Similar results have been obtained on supported Gallium oxides (Choudhary and Jana, 2002) and Iron-mesoporous materials (Tahir et al., 2008).

3.5 Recycling of the catalysts

The stability of the catalysts has been studied by running the reaction successively with the same catalyst (Tl-HMS-35) under the same conditions without any regeneration between the two runs. The reaction was first run under the standard conditions (benzene to benzyl chloride ratio of 15, 353 K) to the complete conversion of benzyl chloride. Then after a period of 8 min another quantity of benzyl chloride was introduced in the reaction mixture leading to the same benzene to benzyl chloride ratio. After the achievement of the second run, the same protocol was repeated a second time. The results, presented in Table 7, showed that the catalyst could be used several times in the benzene benzylation process without a significant change of its catalytic activity.