3.2 Hydrogenation of cyclohexyl acetate to cyclohexanol
Since the hydrogenation of cyclohexyl acetate to cyclohexanol has also not been explored to the best of our knowledge, we developed the Cu/ZnO/SiO2 catalysts for this purpose. The number of catalyst systems allowing the production of alcohols from ester hydrogenation is limited. The catalysts of choice, which allow the highly selective conversion of esters to alcohols, are mostly based on Cu.27–29 Besides, the Cu-based catalysts are more cost-effective than the noble metal-based catalysts for this purpose,30 which is an important factor to be considered for a reaction with significant industrial background. ZnO is capable of dispersing and stabilizing Cu nanoparticles to replace the conventional copper chromite catalysts that are hazardous to the environment.27,31 SiO2 was adopted as the support, since it is not only able to disperse Cu,27 but also weakly acidic in nature and would not lead to considerable undesired intramolecular/intermolecular dehydration or isomerization of the hydrogenation products.
Firstly, we systematically examined the effects of the contents of Cu and the Cu/Zn molar ratios on the catalytic performances of the Cu/ZnO/SiO2 catalysts in the hydrogenation of cyclohexyl acetate. As summarized in Table S3, over the Cu0Zn1Si2 catalyst without Cu, the hydrogenation reaction did not occur. When increasing the contents of Cu while fixing the Zn: Si molar ratio at 1: 2, the conversion of cyclohexyl acetate increased drastically, reaching a maximum over the Cu1Zn1Si2 catalyst, manifesting that Cu is the active center for ester hydrogenation. When further increasing the contents of Cu, the conversion of cyclohexyl acetate declined monotonically, which is attributed to the agglomeration of excessive Cu. A similar effect of the content of Cu on the hydrogenation activity was observed on the Cu/ZnO/Al2O3 catalysts for the hydrogenation of ethyl acetate.32 When altering the Cu/Zn molar ratio while fixing the (Cu + Zn): Si molar ratio at 2: 2, the conversion of cyclohexyl acetate evolved in a volcanic shape. At low Cu/Zn ratio, there was no sufficient active centers for ester hydrogenation, while at high Cu/Zn ratio, there was no sufficient ZnO to disperse Cu, thus an optimal Cu/Zn molar ratio occurred at 1: 1. An optimal Cu/Zn ratio was also identified for the CuZn–SiO2 bimetallic catalysts in the hydrogenation of methyl acetate.33 Among the Cu/ZnO/SiO2 catalysts investigated, the Cu1Zn1Si2 catalyst displayed the highest conversion of 99.5% and the highest cyclohexanol selectivity of 98.3% under identical reaction conditions of 483 K, 5.0 MPa, H2/ester molar ratio of 30, and weight hourly space velocity (WHSV) of 0.5 h–1. As anticipated, there were only trace amounts of the acid-catalyzed by-products including ethyl acetate, cyclohexane, and cyclohexyl ether.
To further improve the selectivity, we in situ modified the Cu1Zn1Si2 catalyst during the co-precipitation process by adding 0.1 molar fraction of La relative to Cu (denoted as Cu1Zn1Si2La0.1). The modification effects of lanthanum oxide have been reported to be originated from several aspects, such as the textural change in catalyst to form basic active sites and the improvement of the surface area of the catalyst.34,35 The basic physicochemical properties of the Cu1Zn1Si2La0.1and Cu1Zn1Si2 catalysts are presented in Table S4. As anticipated, the basicity of lanthanum oxide further suppressed the occurrence of the acid-catalyzed reactions during the hydrogenation of cyclohexyl acetate, which improved the cyclohexanol selectivity up to 99.7%, while did not impose an adverse effect on the conversion (Table S3 and Figure 4). Moreover, increasing the WHSV from 0.5 to 1.1 h–1 did not change significantly the conversion and selectivity over the Cu1Zn1Si2La0.1catalyst (Table S3).
According to Table S4, the specific surface area (S BET) of the Cu1Zn1Si2La0.1catalyst is larger than that of the Cu1Zn1Si2 catalyst. The XRD patterns of the Cu1Zn1Si2 and Cu1Zn1Si2La0.1catalysts in Figure 5A show that aside from the broad peak at 2θof ~25o from amorphous SiO2, only the diffraction peaks due to metallic Cu (fcc Cu, JCPDS 04-0836) were identified, indicating that ZnO is in the amorphous state on both catalysts. The phase relating to La on the Cu1Zn1Si2La0.1catalyst was also not discerned, which is attributed to its low loading and/or high dispersion. The crystallite sizes of Cu calculated by the Scherrer formula in terms of X-ray line broadening are 18.0 nm and 14.8 nm for the Cu1Zn1Si2 and Cu1Zn1Si2La0.1catalysts, respectively, which may account for the high conversion of cyclohexyl acetate when increasing the WHSV on the Cu1Zn1Si2La0.1catalyst. The HRTEM image of the Cu1Zn1Si2La0.1catalyst in Figure 5B shows only the Cu(111) lattice fringes with the interplanar spacing of 2.08 Å, while the lattice fringes of ZnO were not identified, which is consistent with the XRD result. The TEM image and particle size distribution histogram of the Cu1Zn1Si2La0.1catalyst in Figure 5C shows that the average particle size of Cu is 15.8 nm, which agrees well with the crystallite size derived from XRD.
Figures 5D–5I present the scanning transmission electron microscopic-energy dispersive spectroscopic (STEM–EDS) images of the Cu1Zn1Si2La0.1catalyst. Figure 5D is the HAADF–STEM image, Figures 5E–5H are the corresponding EDS mappings of Si, Cu, Zn, and La, and Figure 5I is the overlapping of Figures 5E–5H. Comparison of Figure 5D with Figure 5F readily leads to the conclusion that the bright particles in Figure 5D are originated from Cu. Comparison of Figure 5G with Figures 5D and 5F reveals that Zn is located preferentially on or in vicinity to Cu rather than distributed randomly on SiO2, signifying that ZnO interacts strongly with Cu. This conclusion is additionally substantiated by Figure 5I. On the other hand, the distribution of La is homogeneous and does not show such a preference as that of Zn.
From the experimental results demonstrated above, we figured out the overall atom economy of 99.4% and the yield of cyclohexanol per pass of 34.6% for the novel cyclohexene esterification–hydrogenation process, as also demonstrated in Figure 1. The former is comparable to that of the most selective cyclohexene hydration process, whereas the latter is more than twice of that of the most efficient phenol hydrogenation process. It is evident that the cyclohexene esterification–hydrogenation process is highly advantageous in producing cyclohexanol in a safe, efficient, and green manner.