NaX zeolite was obtained from Tong Xing Corp (Shanghai, China). The CsX and KX zeolites were prepared from NaX following three ion exchange processes with aqueous solutions (1 mol/L) of CsOH and KOH, respectively. The exchanged samples were dried at 85 °C and calcined in air at 520 °C for 3 h before being rehydrated and rinsed with distilled water. Deuterated methanol-d₄ (CD₃OD) and deuterated toluene-d₈ (C₆D₅CD₃) were provided by Aladdin Industrial Corporation (Shanghai, China).

The mechanism of the side-chain alkylation of toluene was initially investigated using CD₃OD as an isotope tracer. The crude product was analyzed by GC-MS and ²H-NMR. The ²H-NMR spectrum of the reaction mixture (Fig. 1(a)) contained peaks at 2.23 and 1.18 ppm, which were assigned to the deuterated methyl groups of toluene and ethylbenzene, respectively.

Fig. 1. (a) 2H-NMR spectrum of the reaction mixture; (b) Representation of the faujasite structure showing only the tetrahedral atoms (SI, SI', SII, SII', SIII, and SIII' are the positions of the cationic sites); (c) Adsorption isotherms of methanol; (d) Deuterium distribution density.

Notably, the ion mechanism postulated by Marco Rep (as shown in Eqs. (1-3)) could not be used to explain why most of D atoms were located on toluene, whereas very few were attached to the β-C atoms of ethylbenzene. According to Eqs. (1-3), the β-C atoms of the styrene and ethylbenzene products resulting from the ion mechanism should contain deuterium. However, the results showed that only a small amount of D was incorporated on to the β-C atoms of ethylbenzene and that most of D were observed on the toluene. This result therefore suggested that the H-D exchange reaction between the methyl groups of CD₃OD and toluene was occurring by a free radical reaction (Eqs. (4-7)).

C₆H₅CH₃ ⇄ C₆H₅CH₂^• +H^• (4)

CD₃OD ⇄ ^•CD₂OD + D^• (5)

C₆H₅CH₂^•+ D^• ⇄ C₆H₅CH₂D (6)

^•CD₂OD +H^• ⇄ CHD₂OD (7)

As indicated in the above equations, the toluene would be cracked to give C₆H₅CH₂^• and H^• and CD₃OD would dissociate to give ^•CD₂OD and a deuterium radical (D^•). The C₆H₅CH₂^• and D^• radicals would then combine to form deuterated toluene-d₁ (C₆H₅CH₂D), whilst ^•CD₂OD would react with H^• to give trideuteromethanol (CHD₂OD).

Seemingly, H-D exchange between methyl group of CD₃OD and toluene also could proceed by ion reaction (Eqs. (8-11)).

C₆H₅CH₃ ⇄ C₆H₅CH₂^- + H⁺ (8)

CD₃OD ⇄ CD₂OD^- + D⁺ (9)

C₆H₅CH₂^- + D⁺ ⇄ C₆H₅CH₂D (10)

CD₂OD^- + H⁺ ⇄ CHD₂OD (11)

Quantum chemical calculations were performed at the B3LYP/6-31+G level using the Gaussian 09 program system to verify the route of the H-D exchange reaction. Scheme 1 provides a visual insight into the energies associated with the different reactants, as well as the energies of the transition states and products. The transition state of the radical mechanism was found to be about 0.149 Hartree higher than that of methanol. In contrast, the transitions state of the ion mechanism was about 0.657 Hartree. The energy barrier associated with the radical route was therefore shown to be much lower than that of the ion route, which provided a strong indication that the H-D exchange reaction between the methyl group of CD₃OD and toluene would have proceeded via the radical route.

Scheme 1. (a) Relative energies of H/D exchange between toluene and methanol (Hartree) in radical mechanism; (b) Relative energies of H/D ex- change between toluene and methanol (Hartree) in radical mechanism.

Although the results of the H-D exchange reaction can be used to confirm the presence of free radicals during the reaction, they cannot be used to demonstrate that the side-chain alkylation of toluene proceeds via a free radical reaction. To examine whether the side-chain alkylation of toluene is a free radical reaction, we investigated the reaction of p-nitrotoluene (NO₂-Ph-CH₃) with methanol instead of toluene. Aromatic nitro-groups generally behave as strong electron acceptors, as evidenced by the fact that the methyl hydrogen of NO₂-Ph-CH₃ (pK_a = 20.4) dissociates much more readily to give H⁺ than that of toluene (pK_a = 43), with the resulting p-nitro-benzyl anion (NO₂-Ph-CH₂^-) being much more stable than the benzyl anion (Ph-CH₂^-). If the side-chain alkylation of toluene did proceed via an ionic mechanism, then the yield of the product resulting from the side-chain alkylation of p-nitrotoluene would increase compared with that of toluene. Furthermore, nitrobenzene compounds generally behave as radical free radical scavenger because aromatic nitro groups can react with active radicals to form more stable radicals (Eq. (12)).

(12)

However, if the side-chain alkylation did proceed via a free radical reaction mechanism, then the methyl side-chain of p-nitrotoluene would not be alkylated. In practice, the results of this experiment revealed that there was no product resulting from the side-chain alkylation of p-nitrotoluene, which indicated that the side-chain alkylation of toluene proceeded via a free radical reaction.

Overall, the isotope trace experiments demonstrated that a D atom from the methyl group of CD₃OD was exchanged with a hydrogen atom from the methyl group of toluene. Furthermore, the results of the quantum chemical calculations showed that the energy barrier for the H-D exchange reaction was 0.148640 Hartree for the radical route and 0.656555 Hartree for the ionic route. Taken together, these results provided strong evidence in support of the H-D exchange reaction between the methyl groups of CD₃OD and toluene proceeding via a radical mechanism. Further proof of this process occurring via a radical mechanism came from the observation that the side chain of p-nitrotoluene was not alkylated with methanol. Instead, the C₆H₅CH₃/C₆H₅CH₂^• reacted with CH₃^•/CH₃OH to generate 2-phenyl-ethanol (C₆H₅CH₂CH₂OH), which underwent a dehydration to produce styrene (Eqs. (13) and (14)). Furthermore, the reaction of C₆H₅CH₂^• with CH₃OH and the reaction of CH₃^• with C₆H₅CH₃ would give ethylbenzene (Eqs. (15) and (16)). Ethylbenzene was also produced by the hydrogenation of styrene with H^• (Eq. (17)).

C₆H₅CH₂^•+CH₃OHC₆H₅CH₃CH₂OH ⇄ C₆H₅CH =CH₂ (13)

C₆H₅CH₃+^•CH₂OHC₆H₅CH₃CH₂OH ⇄ C₆H₅CH=CH₂ (14)

C₆H₅CH₂^•+ CH₃OHC₆H₅CH₂CH₃ (15)

C₆H₅CH₃ + CH₃^•C₆H₅CH₂CH₃ ΔH = 50.6 kJ/mol (16)

C₆H₅CH =CH₂ + 2H^• ⇄ C₆H₅CH₂CH₃ (17)

CH₃OH + H^• ⇄ CH₃^• + H₂O (18)

CH₃OH + H^• ⇄ H₂ + ^•CH₂OH (19)

Toluene could also react with CH₃^• to generate ethylbenzene (Eq. (16)), and the CH₃^• radical could react with CH₃OH to generate CH₄ (Eq. (20)), CO (Eq. (21)), and several other by-products. The radical mechanism could also be used to explain the formation of CO and CH₄ (Eqs. (20) and (21)). The enthalpy changes associated with Eqs. (16), (20), and (21) were 50.6, −27.9, and −27.0 kJ/mol, respectively, and therefore very different. Based on the theory of thermodynamics, CH₃^• would be better suited to form CH₄ and CO than styrene and ethylbenzene. To improve the usage of CH₃OH, it would therefore be necessary to reduce the probability of a collision between CH₃^• and CH₃OH/formaldehyde (CH₂O), while improving the probability of a collision between CH₃^• and C₆H₅CH₃ (or C₆H₅CH₂^•).

CH₃^• + CH₃OH ⇄ CH₄ + ^•CH₂OH ΔH = -37.9 kJ/mol (20)

CH₃^• + CH₂O ⇄ CH₄ + CO + H^• ΔH = -27.0 kJ/mol (21)

X-type zeolites are composed of faujasite and β cages (Fig. 1(b)). The diameters of the 12- and 6-ring orifices of the faujasite and β cage structures are around 0.74 and 0.25 nm, respectively. The diameter of toluene is about 0.65 nm, which is much bigger than the diameter of the 6-ring orifice of the β cage structure of the NaX. With this in mind, toluene should be able go inside the faujasite cages, but not inside the β cages. Basler et al. ^[19] demonstrated that NH₃ can enter the β cages of NaX. Given that the diameter of CH₃OH is similar to that of NH₃, it was envisaged that CH₃OH would also be able to access to the β cages of NaX.

The adsorption isotherms of CH₃OH were measured to determine the validity of this hypothesis and the results are shown in Fig. 1(c). The results revealed that NaX had a similar adsorption capacity (ca. 8500 molecules per unit cell) to KX. The adsorption capacity of NaX was found to be much greater than that of CsX (ca. 6300 molecules per unit cell). This result indicated that the β cages would be blocked when SII and SII′ were occupied by Cs⁺ instead of K⁺ because the radius of Cs⁺ is larger than that of K⁺ and CH₃OH would therefore be prevented from entering the β cages. In this case, the CH₃^• radical would only be able to react with CH₃OH/CH₂O to form side-products in the β cages. Furthermore, the CH₃^• radicals would collide with toluene molecules to generate styrene and ethylbenzene in the faujasite cages. The blocking of the β cages with Cs⁺ ions would therefore improve the probability of collisions between the CH₃^• radicals and toluene (or C₆H₅CH₂^•). Taken together, these results show that CsX is more efficient than KX for catalyzing the reaction of CH₃OH with toluene to give styrene and ethylbenzene.

As mentioned above, the side-chain alkylation of toluene occurs via a free radical mechanism. With this in mind, the process responsible for the formation of the free radicals represents an important step in this reaction. The activation of C₆H₅CH₃ to give the corresponding radical species C₆H₅CH₂^• was evaluated with a variety of different catalysts based on the H/D exchange reaction of a 10:1 (mol/mol) mixture of C₆H₅CH₃ and C₆D₅CD₃. The resulting distribution of D in the products was analyzed by GC-MAS and the results are shown in Fig. 1(d).

The deuterium distribution density results (Fig. 1(d)) showed that C₆D₅CD₃ existed as the major species in the reaction mixture when the catalytic process was conducted with active carbon. This result indicated that it was difficult to convert C₆H₅CH₃ to the corresponding C₆H₅CH₂^• radical using activated carbon as a catalyst. Furthermore, the experiments involving the side-chain alkylation of toluene showed that the conversion of toluene was about 0.1% when the reaction was conducted over active carbon (Table 1). Taken together, these results show that the catalytic effect of activated carbon towards the side-chain alkylation of C₆H₅CH₃ was in agreement its catalytic effect towards the conversion of C₆H₅CH₃ to C₆H₅CH₂^•. In contrast, when the reaction was conducted in the presence of a X-type zeolite (KX or CsX), the amount of C₆D₅CD₃ decreased sharply, whereas the amount deuterated toluene-d₁ increased considerably. This result demonstrates that the zeolite was having a significant effect on activating toluene towards free radicals. The effect of the zeolite could be attributed to a captodative effect involving the synergistic effect of electron-withdrawing (captor) and electron-releasing (donor) groups attached to the radical center, which could make the radical much more stable ^[19]. As the toluene molecules entered the cavities of the zeolite, the cations would form complexes with the aromatic rings, and the resulting complexes would attract electrons. At the same time, the anions within the zeolite framework would donate an electron to the methyl group of toluene (Fig. 2). Based on these interactions, the zeolite would be able to stabilize the C₆H₅CH₂^• radical.

Fig. 2. The captodative effect of metal cation and anion on the framework.

Table 1 Toluene side-chain alkylation data. Catalyst Conversion of toluene (%) Yield of styrene (%) Yield of EB(%) Selectivity of styrene (%) Conversion of methanol (%) Yield of methane (%) Yield of CO (%) C 0.75 0.15 0.15 — — — — KX 3.60 1.32 1.37 — 99.4 43.5 49.3 CsX 16.60 2.89 12.67 17.4 98.2 34.2 48.2 CsX (CO2 atmosphere) 13.80 8.20 7.80 59.4 — — — YCO = NCO, out /NMeOH, in, Ymethane = Nmethane, out /NMeOH, in, where Ni is calculated from the GC response signal (area of the FID or TCD peak) using the response factors for individual components and converted into mols.

Table 1 Toluene side-chain alkylation data.

Compared with the KX catalyst, more of the deuterium atoms present in the mixture existed as toluene-d₁ following the CsX-catalyzed process. The fact that more deuterium atoms were exchanged from deuterated toluene to toluene using the CsX catalyst demonstrated that the H/D exchange process occurred to a much greater extent over the CsX catalyst than it did over the KX catalyst and that the C₆H₅CH₂^• radical was much more stable over CsX. Cs can lose an electron more readily than K, which means that the CsX zeolite has a weaker Cs⁺ cation (electron-withdrawing) and stronger framework anion (electron-releasing) than the KX zeolite. The cooperativity of the acceptor and donor moieties of the CsX zeolite therefore possesses better captodative effects to increase the stability of the C₆H₅CH₂^• radical. Furthermore, CsX exhibited higher catalytic activity towards the side-chain alkylation of toluene than KX. The levels of toluene conversion using the CsX and KX zeolites were found to be 16.6% and 3.6% (Table 1), respectively. This result therefore demonstrates that the higher catalytic activity of CsX towards the conversion of C₆H₅CH₃ to C₆H₅CH₂^• was critical to its higher catalytic activity towards the side-chain alkylation of toluene compared with KX.

Several experiments were designed to improve the selectivity of the side-chain alkylation reaction of toluene based on the radical reaction mechanism. Notably, CO₂ was used as a carrier gas instead of N₂. Under an atmosphere of N₂, there are three possible reaction routes for H^• (Fig. 3(a)), including: (1) the reaction of H^• with CH₃OH following route S to generate ^•CH₂OH, which would be subsequently converted to styrene; (2) the reaction of H^• with CH₃OH according to route E₁ to generate CH₃^•, which would subsequently react with toluene to give C₆H₅CH₂CH₃; and (3) the reaction of H^• with C₆H₅CH=CH₂ following route E₂ to give C₆H₅CH₂CH₃. Under an atmosphere of CO₂, H^• would react with CO₂ to form ^•COOH. The high bonding energy of the C-H bond in formic acid (ca. 468 kJ/mol) indicates that the ^•COOH radical could readily abstract a H atom from the methyl group of CH₃OH to form HCOOH as an intermediate, which could decompose to give H₂ and CO₂ at the end of the reaction (Fig. 3(b)). It would be much more difficult for the ^•COOH radical to obtain an OH moiety from CH₃OH to form H₂CO₃ because of the low bonding energy of the C-OH bond in H₂CO₃. CH₃OH would therefore be more inclined to form styrene following route S because the reaction of CH₃OH with ^•COOH would be relative facile and energetically favorable. The presence of CO₂ would lead to a decrease in the concentration of H^•, which would lead to a further decrease in the production of C₆H₅CH₂CH₃ from the reaction between styrene and H^•. Thus, the selectivity of the reaction for styrene would be increased. Pleasingly, the experimental results of this study were found to be consistent with the design strategy, as shown in Table 1, with the conversion of toluene decreasing slightly and the selectivity for styrene increasing from 17.4% to 59.4% when the reaction was conducted under CO₂.

Fig. 3. (a) Reaction in atmosphere of N2; (b) Reaction in atmosphere of CO2.