C. Liu, et al.
CatalysisCommunications133(2020)105828
under N2 protection. The obtained solid was grinded into powder before
use.
A series of parallel experiments were performed to optimize the
reaction conditions (Fig. 3). 80 °C, as initially employed, was the pre-
ferable reaction temperature. The reaction speed slowed down at low
temperature, resulting in the incomplete conversion of ethene-1,1-
diyldibenzene and the relevantly decreased benzophenone yield
(Fig. 3a). Generation of over-oxidation by-products increased drama-
tically at the temperature over 90 °C (Fig. 3a). Besides 1,4-dioxane,
toluene was also a favourable solvent for the reaction, giving benzo-
phenone in 46.2% yield. EtOH, which was more environmental
friendly, was also fit for the reaction to produce benzophenone in
42.3% yield. Unexpectedly, the use of EtOAc resulted in a sharp de-
crease of the product yield, while a series of unidentified by-products
were generated and could be observed in TLC. DMC is an inflammable,
green and cheap chemical in industrial production and it was found to
be the most favourable solvent for the reaction among the candidates
listed in Fig. 3b. Since ICP analysis showed that the Se@Fe2O3 catalyst
contained ca 0.15 w/w% of Se, the TON of the reaction was calculated
to be 694 on the basis of the involved Se amount. The used catalyst
could be isolated from the reaction liquid by centrifugal separation and
was reusable for the next turn of reaction without obvious deactivation
(affording 50.3% product yield).
The substrate scope of the reaction was then examined. Besides
ethene-1,1-diyldibenzene, the prop-1-en-2-ylbenzene substrate could
afford the related product acetophenone in 67% yield. By introducing
methyl as an electron donation group, 1-methyl-4-(prop-1-en-2-yl)
benzene led to 1-(p-tolyl)ethan-1-one in elevated yield at 71%, while
the electron-deficient substrate 1-chloro-4-(prop-1-en-2-yl)benzene af-
forded 1-(4-chlorophenyl)ethan-1-one in decreased yield at 53%. Tri-
substituted ethene such as prop-1-ene-1,1-diyldibenzene led to the de-
sired product benzophenone in 51% yield. For ethene-1,1,2-triyl-
tribenzene bearing even larger steric hindrances, the benzophenone
yield decreased to 42%. Oxidation of 1,1,2,2-tetraphenylethene did not
occur under the Se@Fe2O3-catalyzed reaction conditions. The reactions
of (E)-1,2-diphenylethene and styrene were also tested, and they led to
benzoic acid in 72% and 78% yields respectively.
The mechanism of this interesting reaction was our next concern.
First, a blank reaction with unselenized Fe2O3 as catalyst was per-
formed under the standard conditions (see Section 2.3) and it produced
benzophenone in only 19% yield, while a lot of the starting ethene-1,1-
diyldibenzene was unconverted and could be recovered in 69% yield.
This result demonstrated that doping Se was essential to act as an
oxygen carrier catalyst for the transformation. Moreover, it has been
attested by 77Se NMR and X-ray photoelectron spectroscopy (XPS)
analysis that oxidation of low-valent Se by H2O2 could produce the
high-valent –SeO3H (Eq. (1)) [31,32], which was highly active and
participated the Se catalyzed oxidation reactions as an oxidative cata-
lytic species. Since GC–MS analysis detected the existence of 2,2-di-
phenyloxirane, it was suggested that the reaction proceeded via an
epoxidation step first.
Thus, on the basis of the experimental results as well as the refer-
ences reports including our previous works on organoselenium catalysis
[31–35], a plausible mechanism for this Se@Fe2O3-catalyzed oxidative
cracking reaction was given below (Fig. 4). Epoxidation of the starting
ethene-1,1-diyldibenzene initially afforded the intermediate 2,2-di-
phenyloxirane A. The hydration of A led to diol B [33], in which the
proximal hydroxyl was even more active and could be oxidized into
carboxyl of C in the presence of Se catalyst and H2O2 [34]. Further Se-
catalyzed Baeyer-Villiger oxidation reaction of C led to the intermediate
D, which produced the relatively stable benzophenone as the final
product [35]. Notably, the reaction afforded benzophenone in 5.6%
yield without H2O2 (Fig. 2a), showing that air might participate the
reaction as a supplementary oxidant, so that no excess H2O2 was re-
quired.
2.3. General procedure for the Se@Fe2O3-catalyzed oxidative scission of
C]C bond
1 mmol of alkene, Se@Fe2O3 catalyst and a piece of magnetic bar
were added into a reaction tube. A solution of 30 w/w% H2O2 in 1 mL of
solvent was then injected into the reaction tube and the obtained
mixture was stirred and heated for 24 h. After evaporation of the sol-
vent, the residue was subjected to preparative TLC to get the produced
benzophenone. Specific substrate, catalyst amount, H2O2 amount, sol-
vent and reaction temperature, were discussed in results and discus-
sions section vide infra. Characterization data and 1H and 13C NMR
spectra of products were given in supplementary data.
3. Results and discussions
Iron (Fe) is an abundant metal with extremely low cost for large-
scale applications. Moreover, references have demonstrated that Fe
could obviously enhance the catalytic activity of Se via a relay catalysis
route so that air can be used as partial oxidant to reduce the used
amount of H2O2 [29]. For these reasons, its oxide Fe2O3 was employed
as the support of Se catalyst. Bearing electron-withdrawing oxygen, the
Fe in Fe2O3 was electropositive to be attacked by HSe−, which, as a
strong nucleophile, was generated by reducing Se powder with NaBH4
and was used in situ without separation [30]. Fig. 1 shows the approach
for preparing Se@Fe2O3 catalyst and its micro-scale morphology in FE-
SEM image, which illustrates a variety of loose nanostructures within
200 nm, allowing sufficient specific surface area for sufficient contact
with the reactants.
The catalytic activity of Se@Fe2O3 for oxidative scission of C]C
was then tested. Oxidation of ethene-1,1-diyldibenzene was chosen as
the model reaction. Heating 1 mmol of ethene-1,1-diyldibenzene with
4 mmol of H2O2 (30 w/w%) in the presence of 30 mg of Se@Fe2O3 in
1,4-dioxane at 80 °C for 24 h, the desired oxidation product benzo-
phenone could be separated in 26.9% yield. The H2O2 dosage was then
examined and using H2O2/ethene-1,1-diyldibenzene at 2.0 mol/mol,
which was the theoretical H2O2 amount for C]C scission to produce
carbonyl, was screened out to be the preferable condition (Fig. 2a). The
product yield decreased with insufficient or excess H2O2 amount. The
former was caused by the incomplete conversion of the substrate, while
in the latter reaction, a series of over-oxidation by-products, such as
benzoic acid, phenol etc., were generated (as detected by GC–MS).
Dosage of the Se@Fe2O3 catalyst was also examined (Fig. 2b). Using
40 mg of Se@Fe2O3 for the reaction of 1 mmol of ethene-1,1-diyldi-
benzene was favourable, leading to an elevated benzophenone yield at
40.2%, but further enhancing the catalyst amount resulted in a sharp
decrease of product yield for elevated by-product generation.
Fig. 1. Diagram for the synthesis of Se@Fe2O3 and its FE-SEM image.
2
Liu, Chuang
