3.1. Catalysts Characterization
Firstly, the as-prepared Hf/SBA-15(X) catalysts were subjected to XRD characterization, as presented in
Figure 1. Wide diffraction peaks around 20–30° were observed across all Hf/SBA-15 catalysts with different Si/Zr ratios, which indicated the amorphous structure of the obtained catalysts [
19]. No typical peaks corresponding to metal oxide (i.e., HfO
2) were found, further suggesting the homogeneous dispersion of Hf metal within the catalysts. In the small angle XRD spectra, all Hf/SBA-15 catalysts showed three diffraction peaks around 0.9°, 1.6°, and 1.8°, which can be assigned to the (100), (110), and (200) diffraction peaks of a two-dimensional (2D) hexagonal p6mm structure, respectively. They were also the characteristic patterns of SBA-15, furthering confirming the successful construction of the SBA-15 framework [
20]. In addition, the orderliness of the SBA-15 structure was well reserved after the incorporation of Hf metal, which might benefit the mass diffusion within the catalyst’s structure during the reaction.
The textural properties of the as-prepared Hf/SBA-15 catalysts were presented in
Table 1. The catalysts showed decreased Brunauer–Emmett–Teller (BET) surface area from 811.3 to 624.4 m
2/g, as well as the pore volume from 1.4 to 1.1 cm
3/g, along with the increase of Hf content in the catalyst (Si/Hf ratio decreased from 30 to 15). These results were predictable as more Hf metal would replace Si to form an SBA-15 framework, which had a different atom size or coordination mode as compared to Si that would alter the native SBA-15 structure. Furthermore, some of the small Hf oxide aggregates may be formed and block the pores, due to the introduction of more Hf metal. Aside from that, all the catalysts still possessed high BET surface areas over 600 m
2/g, which would be beneficial for the alkylation reaction (
Table 1).
XPS characterization of the representative catalyst Hf/SBA-15(20) was investigated, together with the metal oxide HfO
2 as a control sample, to provide detailed insights into the chemical states of the Hf within the catalysts. As shown in
Figure 2, the binding energy value around 16–22 eV was the Hf 4f peak of the Hf/SBA-15(20) catalyst, which can be further deconvoluted into Hf 4f
7/2 and Hf 4f
5/2, with the binding energies of 17.7 and 19.3 eV, respectively. For HfO
2, these two major binding energy values were located at 16.8 and 18.4 eV, respectively. By comparing the binding energies of Hf species in these two different catalysts, clear shifts of the binding energies to higher values were observed from HfO
2 to Hf/SBA-15, which was indicative of more electron deficiency of Hf metal in the Hf/SBA-15 catalyst. This was caused by the electron transfer from Hf to silica, due to the strong higher electronegativity of silicon [
21], thereby leading to a more positively charged Hf metal center. This would also promote the Lewis acidity of the Hf center and increase the catalyst’s reactivity toward the FC reaction.
The acid properties of Hf/SBA-15(20) and HfO
2 were investigated by Py-IR characterizations, as revealed in
Figure 3. For the Hf/SBA-15(20) catalyst, three peaks corresponding to different types of acids were observed. The peak at 1450 cm
−1 originated from the Lewis acidic sites, whereas the wide peak at 1540 cm
−1 represented the Brønsted acidity of the catalyst. The peak at 1490 cm
−1 was assigned to the mixed Brønsted and Lewis acidic sites. According to the quantitative analysis of the desorbed pyridine, the amount of Lewis acid sites of Hf/SBA-15(20) was calculated to be 176.3 µmol/g and the amount of Brønsted acid sites was 60.7 µmol/g, indicating the dominated Lewis acid types of the Hf/SBA-15(20) catalyst. For non-mesoporous HfO
2, only weak Lewis acidic sites were detected on the catalyst’s surface and the amount of acid sites was 11.4 µmol/g, much lower than that of mesoporous Hf/SBA-15(20). These results clearly showed that the immobilization of metal sites into the framework of mesoporous material can expose more metal sites, thereby increasing the total acid sites on the catalyst’s surface. Moreover, by using a silica framework, the electron transfer from Hf metal to silica, as indicated in the XPS result, also promoted the Lewis acidity of the Hf center, making Lewis acid sites the dominant acid type of the catalyst, which would also benefit the alkylation reaction. These characterizations further highlighted the importance of mesoporous silica frameworks for the preparation of highly efficient metal catalysts.
SEM and TEM characterizations of typical catalyst Hf/SBA-15 (20) were also provided to elucidate structure characteristics in
Figure 4. The catalyst had a rod shape with a mean size of 600 nm in length (
Figure 4a). TEM spectra showed 2D pore distribution in sight, along with darker areas within the framework of the catalyst (
Figure 4b). These results indicated that the Hf metal were homogeneously dispersed in the catalysts, without forming visible metal oxide. The catalyst was also subjected to the STEM-mapping characterization, and well-dispersed metal sites were observed, indicating the homogeneous immobilization of the metal within the structure of the silica framework.
3.2. Catalyst Screening
Firstly, benzyl alcohol and toluene were selected as the model compound for the FC alkylation reaction over a series of solid acid catalysts. The main reaction steps involved in the reaction were presented in
Scheme 1.
Table 2 showed the catalytic performances of different solid acid catalysts for FC reaction under neat reaction conditions at a fixed metal loading. The reaction conversion and BA yield were analyzed by GC and 1H-NMR (SI). For the blank test, pure SBA-15 material was subjected to the FC reaction, and no product was formed, indicating the necessity of employing metal sites for the reaction (
Table 2, Entry 1). When we subjected our previously used Zr/SBA-15 catalyst, a Lewis acidic catalyst that was successfully applied to the alkylation of 2-methylfuran with furfural to produce a C15 diesel precursor, to the FC reaction, an 89.3% yield of the target product was obtained (
Table 2, Entry 2). This result encouraged us to extend our research by exploring more metal-doped mesoporous catalysts for FC reactions. For W/SBA-15, an 83.5% yield of the target products could also be obtained (
Table 2, Entry 3). However, Fe/SBA-15 showed no catalytic reactivity. When we turned to the Hf/SBA-15 catalyst, product yields over 90% could be obtained across all the catalysts with different ratios (
Table 2, Entries 5–8). Particularly, when the Si/Hf ratio was 20, the catalyst exhibited the highest reactivity, offering an almost quantitative product yield (99.1%). By comparing the structure characteristics with other Hf/SBA-15 catalysts with different Si/Hf ratios, Hf/SBA-15(20) seemed to possess an approximate BET surface area or pore volume/diameter ratio that enabled better catalytic performance of Hf sites. According to the above Py-IR characterization, Lewis acidity of Hf/SBA-15 should be primarily responsible for the FC reaction, which activated the alcohol to generate a positively charged carbon center to undergo an electrophilic substitution reaction on aromatic rings. By comparing our results with the previously reported FC benzylation reactions with Lewis type catalysts, such as SiO
2-ZrO
2 and Al/SBA-15, Hf/SBA-15 possessed higher reactivity under relatively mild reaction conditions (
Table S1). This catalyst was further applied to the condensation of furfural and 2-methylfuran where a 96.6% product yield was achieved, higher than that with the Zr/SBA-15 catalyst (
Scheme S1), which further confirmed the higher reactivity of Hf over Zr when immobilized in the framework of SBA-15 supports [
16].
Additionally, various Hf catalysts (i.e., HfCl
4, HO
2, HfO(PO
4)
2, and HfO(OH)
2) were also tested for the FC reaction (
Table 2, Entries 9–12). However, all these catalysts showed almost no reactivity toward the FC reaction. These results were probably caused by the lower Lewis acid density or lower pore structure of the catalyst that was unable to activate benzyl alcohol for the reaction, which, on the other hand, emphasized the necessity of creating a mesoporous silica framework for promoting the Lewis acidity of Hf sites. Some commercially available acidic catalysts were also compared with our Hf/SBA-15(20). We found that only a moderate yield could be achieved over Amberlyst-15 and H-β, with even less yield achieved over H-ZSM-5, implying that the catalytic efficiency of Hf/SBA-15(20) in the FC alkylation of benzyl alcohol and toluene was very high.
The influence of substrate ratio was also evaluated by reducing toluene usage from 3 to 1.5 or 1 mL (
Table S6). A gradual decrease in the product yields was observed along with the decrease of toluene usage, which was common in the FC alkylation reaction. Fortunately, by further prolonging the reaction time, a satisfying product yield could be achieved even with the lower toluene usage. In terms of the recyclability of the aromatics and economics of the process, an excessive loading of aromatics was acceptable.
3.3. Reaction Pathway
The reaction pathway was investigated by monitoring the reaction over time.
Figure 5 showed the product evolution in the reaction over the Hf/SBA-15(20) catalyst. After a short reaction time of 1 h, the benzylated products could be detected with a total yield of 20%, accompanied by the side product, dibenzyl ether (DBE). This side product was generated by self-condensation of benzyl alcohol over acidic sites at elevated temperature, which is common in an FC reaction [
22]. Luckily, this side product could also act as the electrophile to undergo FC reaction by releasing a PhCH
2+ to attack the aromatic ring of toluene. As the reaction proceeded, the yield of DBE gradually decreased, while that of the target benzylated products continued to increase, reaching the highest yield of 99.1% at 6 h. These results clearly suggested that the etherification of benzyl alcohol and the direct FC reaction occurred in parallel, in the presence of the Hf/SBA-15 catalyst. The etherification product DBE was further activated by Hf sites and contributed to the alkylation reaction, as elucidated in route 2. To further asses the difference between Zr/SBA-15 and Hf/SBA-15 in catalyzing this reaction, we also traced the FC reaction by using a Zr/SBA-15 catalyst, which exhibited similar product evolution trends, but with slower reaction rates, offering 89.3% product yields within 6 h reaction time (
Figure S1). This was probably due to the stronger acidity of Hf/SBA-15 over Zr/SBA-15, as evidenced by the Pyridine-IR characterization (
Figure S2) that promoted the whole reaction rates.
3.6. Substrate Applications
After obtaining the optimal reaction condition, we also applied the catalytic system for the FC reaction with other substrates possessing different structures and functional groups (
Table 3). For active primary benzyl alcohol, it showed high reactivity toward the FC reaction (
Table 3, Entry 1). For a more specific substrate cinnamyl alcohol, the reaction can provide 63.1% of the alkylation products by prolonging the reaction time to 9 h (
Table 3, Entry 2). Common aliphatic alcohols, such as cyclohexanol and 1-hexanol, were also tested, but no reaction occurred under the standard reaction condition (
Table 3, Entries 3 and 4). This was probably attributed to the inert nature of these primary and secondary alcohols for the FC reaction, which were also mentioned in the reported literature [
15]. In particular, alkyl halide benzyl bromide was also tested as an electrophile, which provided 99.1% of the target product under optimal reaction conditions, indicating the broad applicability of the Hf/SBA-15(20) catalyst in the FC reaction (
Table 3, Entry 5). A series of aromatic substitutes, including ethylbenzene, mesitylene, m-xylene, and anisole, were also subjected to the reaction with benzyl alcohol, which also showed outstanding performances in the FC reaction, offering >96% product yields for all respective substrates, demonstrating the broad applicability of the Hf/SBA-15 catalyst in an FC reaction. However, for aromatics with electron-withdrawing functionalities, the reaction showed inferior results. It could only provide a 10.5% product yield when using bromobenzene as an electron-deficient aromatic compound for FC alkylation reaction. By increasing the reaction temperature to 160 °C, a 29% product yield could be obtained, which was also comparable to some of the previous work [
5,
12].
3.7. Catalyst Recycling
The robustness and recyclability of the Hf/SBA-15(20) catalyst in the FC reaction was also investigated under the optimal reaction condition using a five-run recycling test. After each run, the solid catalyst was separated and calcined to completely remove organic residues on the catalyst surface to obtain the calcinated catalyst.
Figure 8a presents the five-run test results, where the Hf/SBA-15(20) catalyst provided comparable product yields in each run. Slight decreases of the product yield in each run can be observed, which may be caused by the accumulated weight loss of the catalyst during the recycling process. To avoid the mask of catalyst deactivation due to excessive catalyst loading [
23], the recycling test was also conducted with a lower catalyst loading of 60 mg (
Figure S4), which showed a similar product variation trend to that from the optimal catalyst loading of 80 mg. This result clearly indicated that the optimal catalyst loading was appropriate to demonstrate the property and recycling ability of the catalyst. The calcinated catalyst was also subjected to SEM, TEM, and XRD characterizations to reveal structure changes during the recycling process. Very similar spectra were obtained as compared to the freshly prepared catalysts, suggesting the robustness of the Hf/SBA-15(20) catalyst.
During the reaction, we also noticed that the recovered catalyst without calcination showed inferior catalytic activity, offering an 87% yield of product. This was probably caused by the deposition of the carbonaceous species and the absorption of formed water molecules that prevented mass diffusion and deactivated the acidic sites. To verify this, TGA characterizations of the recovered Hf/SBA-15(20) and calcinated Hf/SBA-15(20) were conducted to reveal their difference (
Figure 9). A 9.4% weight loss was observed in the temperature ranging from 100 to 800 °C, for recovered Hf/SBA-15(20), whereas it was only 5.7% for the calcinated catalyst, indicating the organic carbonaceous species deposition on the catalyst surface. As for the influence of formed water, we carried out the experiments by adding a molecule sieve in the reaction system to remove the formed water (
Figure S3). Accordingly, the model reaction with the molecule sieve showed a better reaction yield as compared to that without a molecule sieve at a comparatively low catalyst loading, which implied that the water did deactivate the acidic sites of the catalyst. However, this effect was minimized when the catalyst loading was increased. Thus, the deposition of carbonaceous species formed in the reaction would be a primary reason for the deactivation of the recovered catalyst. Future work would be dedicated to the exploration of a new supportive silica structure to create more robust and efficient catalysts.