Solvent effect on the formation of active free radicals from H2O2 catalyzed by Cr-substituted PKU-1 aluminoborate: Spectroscopic investigation and reaction mechanism

Article abstract of DOI:10.1016/j.apcata.2019.117283

The nature of a solvent plays an important role in selectively catalytic oxidations, influencing the catalytic activity and product selectivity. Most work attribute such roles to the polarity, acidity/basicity, or coordination ability of the solvent. Here, we systematically investigated the generated active free radicals in five different organic solvents during the catalytic oxidation of styrene by H₂O₂. Cr-doped PKU-1 aluminoborate was selected as the heterogeneous catalyst in this system. The solvent effect was comprehensively understood along this perspective. Hydroxyl radicals (^[rad]OH) were preferred to be formed in CH₃CN and responsible for the formation of carbonyl compounds, while superoxide ion radicals (^[rad]O₂^?) were favored in DMF and beneficial to the epoxide production. The function of Cr³⁺ was correlated with the solvent effect and a tentative mechanism pathway was proposed accordingly.

Full text of DOI:10.1016/j.apcata.2019.117283

AppliedCatalysisA,General588(2019)117283
Contents lists available at ScienceDirect
Applied Catalysis A, General
journal homepage: www.elsevier.com/locate/apcata
Solvent effect on the formation of active free radicals from H2O2 catalyzed
by Cr-substituted PKU-1 aluminoborate: Spectroscopic investigation and
reaction mechanism
Hongwei Chen, Weilu Wang, Yao Yang, Pengfei Jiang, Wenliang Gao⁎, Rihong Cong, Tao Yang⁎
College of Chemistry and Chemical Engineering, Chongqing University, Chongqing 401331, People’s Republic of China
A R T I C L E I N F O
A B S T R A C T
Keywords:
Solvent effect
Active radicals
Molecule probe method
Styrene epoxidation
Reaction mechanism
The nature of a solvent plays an important role in selectively catalytic oxidations, influencing the catalytic
activity and product selectivity. Most work attribute such roles to the polarity, acidity/basicity, or coordination
ability of the solvent. Here, we systematically investigated the generated active free radicals in five different
organic solvents during the catalytic oxidation of styrene by H2O2. Cr-doped PKU-1 aluminoborate was selected
as the heterogeneous catalyst in this system. The solvent effect was comprehensively understood along this
perspective. Hydroxyl radicals (%OH) were preferred to be formed in CH3CN and responsible for the formation of
carbonyl compounds, while superoxide ion radicals (%O2−) were favored in DMF and beneficial to the epoxide
production. The function of Cr3+ was correlated with the solvent effect and a tentative mechanism pathway was
proposed accordingly.
1. Introduction
key intermediates directly.
In essence, the type and concentration of active free radicals gen-
Liquid-phase selective oxidations towards value-added organic
products become more and more important because of the high eco-
nomic benefit and wide applications in chemical industry and our daily
life [1,2]. A number of studies have demonstrated that the used solvents
exert a significant influence on the catalytic activity and product dis-
tribution [3,4], even the same reaction carried out in different solvents
may give completely different catalytic performances [5–7]. In general,
the solvent functions as a medium to transfer the mass and heat during
reactions, therefore the solvent property, such as dielectric constant,
acidity and basicity, coordination ability, polarity and dynamic visc-
osity, will display an indispensable role in the process of the catalytic
oxidations [8–10]. In addition, a synergistic effect between the co-sol-
vents, or between the solvent and the hydrophilicity-hydrophobicity
surface of catalysts, also influences the catalytic activity [11,12]. In
most cases, solvents may interplay with the active sites on catalyst
surfaces, thereby modify the reaction pathways and the product dis-
tribution. For example, Flaherty et al discovered the solvent molecules
with a high nucleophilicity restrain the H2O2 activation to form reactive
species through a strong competition with the involved active sites, and
thus give a lower activity [4]. Despite these achievements, a compre-
hensive understanding to the solvent effect is so far mysterious because
of the complexity in reaction synergy and the difficulty in observing the
erated determine the efficiency and pathways of catalytic reactions.
With this regards, we investigated the solvent effect on the formation of
free radicals during a particular organic oxidation. Selective oxidation
of olefins to obtain epoxides, particularly for the epoxidation of styrene,
have attracted great attentions during the past decades. As a commer-
cially important reaction, the epoxidation of styrene can provide a
useful organic intermediate-styrene oxide, which was commonly used
as versatile chiral building blocks in organic synthesis. Various catalysts
have been attempted in the epoxidation of styrene in previous studies,
such as molecular sieve [13–17], nano-scaled materials [18–23], tran-
sition-metal complexes [24–27], metal-organic framework (MOF)
[28,29], metal cluster [30–33], composite oxides [34,35]. Among these
investigations, only a few reports focused on the solvents effect. For
example, Corma et al discovered the dramatic differences in catalytic
activity using different solvents during the olifins oxidation, and they
believed this solvent effect may attribute to the poisoning of framework
aluminium, i.e. Brönsted acid sites, by the basic solvent molecules [17].
Zahra et al. investigated the solvent effect on catalytic activity in the
presence of manganese porphyrins catalyst, and they attributed the
large differences between the product distributions to the stability and
high valence manganese-oxo and Mn(III) porphyrins species in the
protic and aprotic solvents [27].
⁎ Corresponding authors.
E-mail addresses: gaowl@cqu.edu.cn (W. Gao), taoyang@cqu.edu.cn (T. Yang).
https://doi.org/10.1016/j.apcata.2019.117283
Received 26 July 2019; Received in revised form 19 September 2019; Accepted 4 October 2019
Availableonline11October2019
0926-860X/©2019ElsevierB.V.Allrightsreserved.

H. Chen, et al.
AppliedCatalysisA,General588(2019)117283
Although a great efforts have been made in literature, most of them
2.3. Detection of free radicals
paid close attention to the relationship between catalytic activity and
surface characters of catalytic materials, no emphases has been paid to
correlate the solvent effect and the so-produced active free radicals. In
this study, selectively catalytic oxidation of styrene using H2O2 oxidant
was employed as a model reaction. We interpret the solvent effects in
terms of the involved free radicals released from H2O2 and catalyzed by
Cr-doped PKU-1 aluminoborate with an octahedra-based framework
[36–44]. The types and concentration evolution of the radicals have
been evaluated in five different solvents in order to understand the
product distributions and the proposed reaction pathways. In CH3CN,
the majority of products were carbonyl compounds (95% selectivity) as
well as a few amount of epoxide. On the contrary, the selectivity to
epoxide reached 53% in DMF. Molecule probe methods proved a direct
relationship between the selectivity tendency and the involved free
radicals, indicating that %OH radicals were preferred to be formed in
CH3CN and responsible for the formation of carbonyl compounds, while
The revolved free radicals were determined by photoluminescence
(PL) spectroscopy and UV–vis absorption spectra. For details, PL spec-
troscopy was used to detect %OH radicals using terephthalic aicd (TA) as
a probe molecule. In a typical run, a solution containing 1 mmol
styrene, 6 mL solvent, 40 mg TA and 3 mmol H2O2 was mixed with
10%Cr-PKU-1 (30 mg) under stirring, then heated at 343 K for a spe-
cified interval of time. After a centrifugation to remove the powder
from the suspension, the upper solution was taken for the PL mea-
surement with an excitation irradiation at 328 nm on a Hitachi F4600
fluorescence spectrometer. PMT voltage was fixed to be 700 V, and the
width of excitation and emission slit were both set to 2.5 nm.
Superoxide ion (%O2−) was determined by UV–vis absorption spectra
using nitroblue tetrazolium (NBT) as a probe molecule. In a typical run,
a solution containing1 mmol styrene, 5 mg NBT, 6 mL solvent and
3 mmol H2O2 was mixed with 30 mg 10%Cr-PKU-1 and heated at 343 K
under stirring for a specified interval of time. The suspension was
centrifuged to remove the solid powder and the upper clear liquid was
measured on an UV–vis spectrometer (PUXI TU-1810, China).
%O2 ion radicals were favoured in DMF and beneficial to the epoxide
−
production. A tentative reaction mechanism was proposed to explain
the reaction pathways of styrene epoxidation based on our results.
2. Experimental section
3. Results and discussion
2.1. Catalysts preparation and characterizations
3.1. Characterizations of Cr-doped PKU-1 catalysts
Cr-PKU-1 was synthesized using boric acid as the flux without using
any organic template as described in previous report [44]. For simpli-
city, the as-synthesized samples were denoted as 0%, 5%, 10%, 15%
and 20%Cr-PKU-1. Powder X-ray diffraction data were collected at
room temperature on a PANalytical Empyrean powder diffractometer
The XRD patterns of the as-synthesized x%Cr-PKU-1 samples were
shown in the right inset of Fig. 1 (x = 5, 10, 15 and 20) and give almost
the same profile with the parent catalyst (0%Cr-PKU-1), hinting that Cr-
incorporation does not change their host structures. In addition, the
refined unit cell volumes give a linear dependence on the Cr-content.
The expansion of unit cell volumes comply with the fact that Cr3+ ion
has a larger radius (0.615 Å, coordination number (CN) = 6) than Al3+
(0.535 Å, CN = 6) [45]. The physicochemical properties of the as-syn-
thesized Cr-PKU-1 catalysts have been characterized by many mea-
surements, i.e. the morphology and homogeneity by SEM, the thermal
behavior by TG-DSC analysis, the chemical compositions and valence
states by XPS spectrum. Detailed discussions were provided in Sup-
porting information (SI) from Figs. S1–S3.
equipped with
a PIXcel 1D detector under Cu Kα radiation
(λ = 1.5406 Å). In order to refine the cell parameters, Le Bail fitting
was performed using the TOPAS software package. A field emission
scanning electron microscopy (JEOL, JSM-7800F) was used to examine
sample morphology by dispersing dry powder on double-sided con-
ductive adhesive tape. X-ray photoelectron spectra (XPS) were acquired
with UK Kratos Axis Ultra spectrometer with Al Kα X-ray source.
Electron binding energies were calibrated against the C 1s emission at
Eb = 284.6 eV to correct the contact potential differences between the
sample and the spectrometer. Combined thermogravimetric (TG) and
differential scanning calorimeter (DSC) analyses were performed on a
Mettler-Toledo TGA/DSC1 instrument at a heating rate of 10 K/min
from room temperature to 1173 K.
3.2. Catalytic activity
In general, the oxidation of styrene has two reaction pathways, one
is the epoxidation to styrene oxide (SO) and the other is the carbony-
lation to carbonyl compounds (CA), such as phenylacetaldehyde (PAC),
benzaldehyde (BZA), benzoic acid (BA). As evidenced in some
2.2. Catalytic reaction
Styrene and other all organic solvents were purified by distillation
method. In a standard run, 1 mmol styrene, 6 mL DMF, 30 mg 10%Cr-
PKU-1 catalyst and 3 mmol H2O2 were mixed in a 10-mL single-neck
round-bottom flask and then immersed into a silicon oil bath to be
heated under stirring. After reacting at 343 K for 10 h, a small amount
of mixture was extracted with a syringe to transfer into a centrifuge
tube, then centrifuged to separate solid powder. Finally, the remaining
transparent solution was analyzed qualitatively or quantitatively with
gas chromatography-mass spectrometry (GC–MS) or gas chromato-
graphy (GC), respectively.
GC–MS was performed to identify the structure of produced organic
compounds on an Agilent 7890 N gas chromatograph coupled with a
capillary column (DB-5: length, 50 m; inner diameter, 0.25 mm; film
thickness, 0.25 μm) and mass spectrum system (Agilent 5975C). GC was
used to determine the concentration of the substrates and products on a
SHIMADZU gas chromatograph (GC2010 plus) equipped with a flame
ionization detector (FID). A capillary column (DB-5: length, 50 m; inner
diameter, 0.25 mm; film thickness, 0.25 μm) was chosen to separate the
mixed components at the following condition: 393 K for capillary
column, 523 K for injector port, and 523 K for FID detector.
Fig. 1. XRD patterns of Cr-PKU-1 samples with different Cr3+ content, and the
corresponding reflection indices for 10%Cr-PKU-1were also labeled, as a re-
presentative.
2

H. Chen, et al.
AppliedCatalysisA,General588(2019)117283
product distributions. Therefore, the further elucidation of this point
needs more efforts in future studies.
As shown in Fig. 4, Cr-PKU-1 effectively activated H2O2 to form
%O2 and %OH radicals in DMF, because the peak intensity, reflecting
−
the concentration evolution of free radicals, became stronger and
stronger along with the reaction time in the presence of the catalyst,
while the peak intensity were quite weak without using Cr-PKU-1. This
experiment demonstrates that Cr3+ species in the catalyst play an in-
dispensable role in the pre-activation of H2O2. Cr-PKU-1 can promote
H2O2 to decompose into both %O2 ions and %OH radicals when DMF
−
was selected as reaction medium. For comparison, our previous work
proved that Cr-PKU-1 in CH3CN catalyzed the quick formation of active
%OH radicals from H2O2, however strongly suppressed the generation of
superoxide %O2 ions [44]. Such a solvent effect can also be further
−
clarified with the following experiments. In the mixed solvents of
CH3CN and DMF, the concentration of generated %O2− ion radicals was
greatly enhanced with decreasing V(CH3CN)/V(DMF) from 6:0 to 0:6
(see Fig. 5a), on the other hand, such an evolution of volume ratio
restrained the formation of %OH radicals as shown in Fig. 5b.
Fig. 2. Styrene conversion and the selectivity to styrene oxide (SSO) and car-
bonyl compounds (SCA) using different solvents over 10%Cr-PKU-1. Reaction
conditions: 1 mmol styrene, 30 mg catalyst, 6 mL solvent, 3 mmol H2O2, and
heated at 343 K for 10 h.
In previous sections, it is proved that CH3CN is favorable for the
production of carbonyl compounds (95% selectivity), but DMF solvent
gradually weakens the tendency for carbonylation production and is
growing to benefit epoxidation pathway. The most essential reason is
that %OH radicals as the critical active species can be largely produced
references [8–10], reaction medium was a remarkable factor to influ-
ence catalytic styrene activity, both in homogeneous and heterogeneous
systems. An appropriate solvent can effectively optimize the reaction
route and greatly alter the product distribution. Herein, five solvents,
including acetonitrile (CH3CN), 1,4-dioxane, tetrahydrofuran (THF),
acetone and dimethyl formamide (DMF) were chosen as reaction
medium due to their good miscibility with reactants. As shown in Fig. 2,
a completely different catalytic activity was obtained in these solvents.
In detail, a slight enhancement from 39.2% (in THF) to 58.0% (in DMF)
was observed in the styrene conversion, while the product distributions
were quite different. When using DMF as solvent, styrene was oxidized
simultaneously into both epoxide and carbonyl compounds (0.53/0.46
in molar ratio), however it preferred to generate much more carbonyl
in CH3CN and responsible for the carbonylation, while %O2 ion radi-
−
cals tend to be formed in DMF and are responsible for the epoxiadation.
To reinforce this statement, we purposely added radical quenchers to
monitor the catalytic selectivity changing tendency as indicated in
Fig. 6. A two-stage experiments for 3 batches were performed. In batch
1, the reaction system was kept at 343 K for 4 h in the first stage and
another 6 h in the second stage with no quencher added during the
reaction. For the second reaction stage, the selectivity to epoxide and
carbonyl compounds were ∼56.6% and 43.6%, respectively. Differ-
ently in batch 2, the quenching reagent for inhibiting %O2 radical
−
ions, benzoquinone (BQ), was added in the second stage, and the se-
lectivity to epoxide and carbonyl compounds during the last 6 h time
interval were determined to be ∼20.3% and 78.8%, respectively. Ap-
parently, the epoxide product was greatly inhibited due to the addition
of BQ, in contrast, the selectivity to carbonyl compounds was largely
strengthened. In batch 3, when t-butanol, a hydroxyl radical quencher,
was added into the reaction system in the second stage, the selectivity
of obtained carbonyl compounds were enormously reduced from 52.8%
(in the first stage) to 23.1% (in the second stage), in the meantime, the
selectivity of epoxide was promoted from 46.7% to 75.9%. In summary,
the above quenching experiments strongly support the statement that
the hydroxyl radical is responsible for the formation the carbonyl
compounds, while the superoxide ion radicals are the key species to
promote the epoxidation.
compounds in CH3CN, i.e. the selectivity to carbonyl compounds (SCA
reached ∼ 95%.
)
3.3. Detection of the involved active free radicals
In general, H2O2 decomposes into hydroxyl radicals (%OH) or su-
peroxide ions (%O2−) over some transition metal sites, subsequently the
formed radicals will act as the critical species in redox reactions
[46,47]. Molecular probe methods were usually applied to detect these
generated active species, in which NBT is a widely used superoxide ion
indicator for the formation of purple formazan with the absorbance at
%
λ = 560 nm [48,49], and TA is used to react with OH radicals to form
a highly fluorescent compound, 2-hydroxyterephthalic acid (TAOH)
[50,51]. Herein this work, the styrene epoxide tends to be produced in
DMF, while carbonyl compounds are easily formed in CH3CN. Fig. 3
gives the corresponding spectra obtained by molecular probe methods
3.4. Reaction mechanism
in different solvents. Interestingly, the highest concentration of %O2
In literature, it was pointed out that styrene can be oxidized into
carbonyl compounds through two catalytic pathways, one is to oxidize
sidechain alkyl of benzene to give benzaldehyde or other carbonyl
compounds; the other is to form styrene oxide firstly, then the formed
epoxide will suffer a nucleophilic attack to produce benzaldehyde
[15,52]. Nevertheless, it was surprising that the second pathway, i.e.
the epoxide-to-benzaldehyde transformation, proceeds in a very limited
degree. In other words, most benzaldehyde was achieved through the
breaking of C]C double-bond in the side chain of benzene. The above
abnormal results were further clarified as follows. Under the same re-
action conditions, styrene oxide, benzaldehyde and phenylacetaldehyde
were used separately as starting reactant to react with H2O2 in order to
determine the actual reaction pathway for styrene oxidation. The
quantitative results of the obtained products were determined with GC
(Fig. 7a). It clearly shows that both benzaldehyde and
−
ion was demonstrated in DMF, while the highest concentration of %OH
radicals was detected in CH3CN. Both results are consistent with the
observed catalytic selectivity in DMF and CH3CN (see Fig. 2). As two
key intermediates from the H2O2 decomposition, %O2 ions and %OH
−
radicals will inevitably compete with each other to react with reactants,
and their concentration fluctuation will play an important role on the
product selectivity. Different with DMF, other four solvents do not ex-
−
hibit any obvious preference in the production of %O2 ions. They all
have a weak capacity to initiate and maintain a low concentration of
%O2− ions in the reaction solution, therefore, the epoxide selectivity has
been greatly restrained. Indeed, the solvent effect in a specific catalytic
reaction is commonly complex, the generation of free radicals some-
times was influenced by the solvation process, and the solvation po-
tential energy of products or active intermediates will determine the
3

H. Chen, et al.
AppliedCatalysisA,General588(2019)117283
−
Fig. 3. Concentration variety of formed %O2 ion radicals and %OH radicals in different solvents monitored by (a) UV–vis absorption spectrum of the produced
formazan and (b) fluorescent spectrum of the produced TAOH. Operational condition: TA (40 mg) or NBT (5 mg), styrene (1 mmol), H2O2 (3 mmol), solvent (6 mL)
and 10%Cr-PKU-1 (30 mg).
Fig. 4. Concentration evolution along the reaction time for %O2− ion radicals (a) and %OH radicals (b) with and without 10%Cr-PKU-1 catalyst. Operating conditions:
TA (40 mg) or NBT (5 mg), styrene (1 mmol), H2O2 (3 mmol), DMF solvent (6 mL) and solid catalyst (30 mg).
Fig. 5. Concentration variety of so-formed %O2− and %OH radicals in different volume ratios V/V (CH3CN/DMF) monitored by (a) UV–vis spectrum of the produced
formazan and (b) fluorescent spectrum of the produced TAOH. Operating conditions: TA (40 mg) or NBT (5 mg), styrene (1 mmol), H2O2 (3 mmol), solvent (6 mL),
10%Cr-PKU-1 (30 mg) and heated at 343 K for 6 h.
phenylacetaldehyde have high conversions than styrene oxide, the very
small conversion for styrene oxide in DMF (2.0%) and in CH3CN (1%)
indicate that it may be a thermodynamically stable product in both
solvents, probably combining with DMF or CH3CN molecules to form a
strong solvation product and is difficult to be activated in the catalytic
active sites, therefore the epoxide-to-benzaldehyde transformation
needs a higher activation energy and is unfavorable under the per-
formed conditions. Based on the above discussion, the actual oxidation
process for styrene oxidation were presented in Fig. 7b.
The evolution of styrene conversion and the selectivity with
4

H. Chen, et al.
AppliedCatalysisA,General588(2019)117283
−
Fig. 6. A proof evidenced by adding BQ and t-butanol quencher to determine the formed %O2 ions and %OH active free radicals. The conversion and selectivity
shown in the second stage are determined only with the amounts changes in these 6 h time interval. Reaction conditions: 1 mmol styrene, 30 mg catalyst, 2 mg BQ or
0.1 mL t-butanol quencher, 6 mL DMF, H2O2 oxidant, and heated for 10 h.
Fig. 7. The experimental schemes designed to search for the actual reaction pathway (a) and two reaction pathway proposed (b) for styrene oxidation using Cr-PKU-1
catalyst in both DMF and CH3CN solvent. Reaction conditions: 1 mmol reaction substrate, 30 mg 10%Cr-PKU-1 catalyst, 6 mL DMF or CH3CN, 3 mmol H2O2, and
heated at 343 K for 10 h.
5

H. Chen, et al.
AppliedCatalysisA,General588(2019)117283
Fig. 8. Evolution of catalytic activity and selectivity with reaction time using DMF and CH3CN as reaction solvent. Reaction conditions: 1 mmol styrene, 30 mg
10%Cr-PKU-1 catalyst, 6 mL DMF (a) or CH3CN (b), 3 mmol H2O2 and heated at 343 K.
reaction time also gives a powerful evidence on the above-proposed
pathways. As the representatives, two solvents, DMF and CH3CN were
selected as reaction medium, and the samples have been collected at the
time interval of 1, 2, 4, 6, 8 and 10 h, then were further analyzed
quantitatively using GC. The obtained results shows that, with
prolonging reaction time, styrene conversion is rapidly increasing,
however the product selectivity exhibits an opposite tendency in DMF
and CH3CN solvents. As shown in Fig. 8a, when using DMF as reaction
medium, the selectivity to epoxide product (SSO) gives a slow increase
from 45.9 to 53.3% during 1 h and 10 h, while the ones of BZA has a
very dramatic decrease from 54 to 28.1%, which was mainly oxidized
into BA product. The above results means, completely different with
BZA, SO should be a thermodynamically stable substance in DMF sol-
vent, the formed solvation product or other analogues has a lower
potential energy, thus results into a higher activation energy and can
afford to the nucleophilic attack of the active radicals. On the contrary,
it is interesting that, when selecting CH3CN as reaction solvent, the
selectivity to SO and BZA show an opposite tendency. The epoxidation
product distributes a weak preference, and SSO has a great decline from
21.6 to 5.12%; while the SBZA turns favorable, increasing from 62.5 to
87.5%, accompanying with the drop of SBA from 15.8% to 7.05%. This
phenomenon largely distributes from the fact that SO is neither easy to
be produced nor degraded with a ring-open reaction in CH3CN solvent.
At the same time, BZA seems to be an end product in CH3CN solvent, it
is relatively stable and difficult to be oxidized into BA products,
therefore results into an obvious achievement in the overall selectivity
to BZA, as indicated in Fig. 8b.
In order to demonstrate the structure-activity relationship shown in
the catalytic oxidation of styrene, we further employed fluorescence
and UV–vis molecule probes to monitor the formation of %OH radicals
and %O2 ion radicals in the presence of Cr-PKU-1 catalysts with dif-
−
ferent Cr3+ concentrations. Fig. 9 gives the concentration dependency
of the formed %O2 ions radical and %OH radicals from the catalytic
−
decomposition of H2O2 by Cr-PKU-1 catalyst. It clearly exhibits the
generation rate of both %OH and %O2 were greatly influenced by the
−
amount of active Cr3+ sites. When Cr-free catalyst (0%Cr-PKU-1) was
used, the peak intensity were very weak, hinting the poor catalytic
capacity. With the increase of Cr content, the formation of %OH radicals
was greatly promoted (Fig. 9b). On the other hand, a raising-first-then-
falling curve was obtained to convert H2O2 into %O2− radicals (Fig. 9a),
the maximum concentration was achieved over 10%Cr−PKU−1 cata-
lyst. That is to say, the excessive Cr3+ active sites are detrimental to the
formation of %O2 radicals, and %O2 radicals are very reactive at the
higher Cr3+ concentration, and they are inclined to be quickly de-
composed into the more stable O2.
−
−
The above concentration fluctuation phenomenon was also ac-
cordant with our former catalytic results, it shows the Cr3+-in-
corporated concentration obviously played a vital role on the styrene
conversion and the product selectivity. As shown in Entry 3–6 in
Table 1, the higher concentration of Cr3+ will lead to the higher con-
version, changing greatly from 25.9% to 67.9% with increasing Cr-
content from 5% to 20%. However, as for the product distributions, the
selectivity to both epoxide and carbonyl compounds has the opposite
variety tendency. The total amount of carbonyl compounds is
−
Fig. 9. Concentration variety of formed %O2 ions radical and %OH radicals in the different Cr3+ concentration in Cr-PKU-1 catalysts monitored by (a) UV–vis
absorption spectrum of the produced formazan and (b) fluorescent spectrum of the produced TAOH. Operational condition: TA (40 mg) or NBT (5 mg), styrene
(1 mmol), H2O2 (3 mmol), DMF solvent (6 mL), Cr-PKU-1 (30 mg) and heated at 343 K for 6 h.
6

H. Chen, et al.
AppliedCatalysisA,General588(2019)117283
Table 1
Summarized results for styrene oxidation with H2O2 oxidant over Cr-PKU-1 catalyst.a
Entry
Temp. (K)
Catalyst
H2O2 equiv.
Conv. (%)
SSO (%)
SCA (%)
SBZA
TOFb (h−1
)
c
SBA
SPAC
Ssum
1
2
3
4
5
6
7
8
343
343
343
343
343
343
323
333
353
363
343
343
343
343
343
blank
3
3
3
3
3
3
3
3
3
3
0.5
1
2
4
5
0.41
6.41
25.9
58.0
62.6
67.9
12.5
25.8
66.2
79.3
15.2
31.5
43.8
65.2
68.5
6.20
28.2
57.9
53.3
45.7
41.5
29.5
44.2
42.3
31.4
60.1
57.5
55.6
48.7
44.1
90.6
67.8
24.5
28.1
34.1
37.0
60.1
41.3
35.9
44.9
26.4
26.9
27.8
30.4
33.5
2.10
2.60
13.8
15.1
16.8
18.5
7.58
11.5
17.5
19.7
10.3
11.5
12.6
17.1
19.3
—
92.7
70.8
40.6
46.1
54.0
58.0
68.3
54.6
56.5
67.4
38.8
40.8
43.0
50.2
54.6
0.001
0.024
0.207
0.427
0.395
0.389
0.051
0.158
0.387
0.344
0.126
0.250
0.336
0.439
0.417
0%Cr-PKU-1
5% Cr-PKU-1
10% Cr-PKU-1
15% Cr-PKU-1
20% Cr-PKU-1
10% Cr-PKU-1
10% Cr-PKU-1
10% Cr-PKU-1
10% Cr-PKU-1
10% Cr-PKU-1
10% Cr-PKU-1
10% Cr-PKU-1
10% Cr-PKU-1
10% Cr-PKU-1
0.38
2.30
2.90
3.05
2.50
0.62
1.85
3.07
2.75
2.07
2.42
2.58
2.70
1.85
9
10
11
12
13
14
15
a
Reaction conditions: 1 mmol styrene, 30 mg catalyst, 6 mL DMF, H2O2 oxidant, and heated for 10 h.
Turnover frequency (TOF) is calculated with the division of the moles of styrene oxide by the moles of used catalyst and reaction time.
Ssum refers to total selectivity of carbonyl compounds and is calculated by the addition of SBZA, SBA and SPAC
b
c
.
proportional to Cr3+ concentration probably due to the rising con-
centration of generated %OH radicals from H2O2, among which, the
amounts of benzaldehyde and benzoic acid were also becoming larger
with the increase of Cr concentration. On the other hand, the amounts
of styrene oxide gradually decreased, and its selectivity reduced from
57.9% to 41.5%.
experimental fact that epoxide was unable to undergo a deeper de-
composition to form benzaldehyde, whether using DMF or CH3CN
solvent. In this sense, we exclude the possibility of thermal decom-
position of cyclic peroxide radicals ⑤ into benzaldehyde and other
products, although the further elucidation of this point is still needed in
future studies. Therefore, it also needs to denote that a few amount of
%OH radicals will be inevitable to be formed in DMF, therefore resulting
in the formation of carbonyl compounds, such as benzaldehyde and
benzoic acid. Nevertheless, as for CH3CN solvent, the majority of H2O2
can be catalytically decomposed into %OH radicals and a trace amount
of superoxide ion radicals were produced, therefore it was not sur-
prising to observe the much higher selectivity to carbonyl compounds
in CH3CN.
In our previous work, Cr3+ active sites were proved to be an effi-
cient stimulus for activation of H2O2 into %OH radicals in the CH3CN
[44], whereas the formation of %O2− were inhibited in the meantime. A
Cr3+–Cr2+–Cr3+ circle was discovered with cyclic voltammetry ana-
lysis during the oxidation reaction, and a tentative mechanism was
proposed to demonstrate the functions of Cr3+ sites based on the ob-
tained results. We believe the oxidative transformation from styrene to
carbonyl compounds in CH3CN medium will also comply with the si-
milar catalytic mechanism. Therefore, in the following, we intend to
elucidate the reaction mechanism for styrene epoxidation using DMF
solvent over Cr-PKU-1. As far, the mechanism of catalytic epoxidation
of styrene with H2O2 using active transition metal is not completely
understood, therefore the proposed mechanism based on our finding as
well as the previously-reported evidences is tentative. As shown in
Scheme 1, DMF molecule has a good coordination ability and firstly
coordinates to Cr3+ to form a transition metal complexes DMF−Cr
(III)−OH (①), which has been evidenced in FTIR and UV–vis spectrum
[17,53]. After combining with H2O2, metal complexes ① will transfer
into superoxo-type of complex H+−DMF−Cr(II)−OO% (②) and un-
dergoes an oxidative addition to the C]C double-bond in the side chain
to give an intermediate ③, accompanying with a valence transition of
Cr3+→Cr2+ in the meantime [44]. Thereafter, a cyclic peroxide radical
⑤ is formed through a migration insertion of the intermediate ③, and
simultaneously metal complexes ① is regenerated through the transition
state ④. In the end, the obtained cyclic peroxide radical ⑤ can further
react with another styrene molecule to give styrene oxide.
Emphatically, 1n the previous studies [52,54,55], some authors
have reported that cyclic peroxide radicals ⑤ can undergo a thermal
decomposition to form benzaldehyde or formaldehyde. However, it
does not seem to be the case in our current catalytic system. It is well
known that epoxide will also suffer a nucleophilic attack to undergo a
ring-open catalytic process in a stronger oxidation environmental
system, the active intermediates, such as cyclic peroxide radical ⑤, were
believed to be formed and can further perform a thermal decomposition
[15,52]. In our case, although a large amount of carbonyl compounds
have been produced, they were mostly likely to come from the first
pathway, i.e. the oxidation of sidechain alkyl of benzene, based on our
4. Conclusions
Catalytic oxidation of styrene was performed using H2O2 as oxidant
and Cr-PKU-1 aluminoborate as catalyst in five different solvents, in-
cluding CH3CN, DMF, THF, acetone and 1,4-dioxane, to achieve a
comprehensive understanding to the solvent effect in terms of the
generated active free radicals. Under the employed reaction conditions,
the isolated Cr3+ sites in Cr-PKU-1 were proved to function as catalytic
sites to activate rapidly H2O2 to release two different active free radi-
cals, %OH radicals and %O2− ions radicals. The molecule probe methods
including fluorescence and UV–vis absorption spectra indicated Cr3+
catalytic sites promoted the activation of H2O2 into %OH radicals ex-
clusively and restrained the formation of %O2 ion radicals in CH3CN,
−
while the opposite tendency was observed in DMF. Furthermore, cat-
alytic results showed that an enormous difference in product selectivity
occurred in these different solvents. Carbonyl compounds were pre-
ferred to be formed probably due to the strong carbonylation effect of
hydroxyl radicals in CH3CN, but the styrene oxide was favorably pro-
duced in DMF solvent because of the obvious tendency towards the
epoxidation with the help of superoxide ion radicals.
Acknowledgments
This work was supported by the National Natural Science
Foundation of China (21671028, 21771027, 21805020), Natural
Science Foundation of Chongqing (cstc2019jcyj-msxmX0312,
cstc2019jcyj-msxmX0330), Postdoctoral Science Foundation of China
(2018M643402) and Chongqing Postdoctoral Science Special
Foundation (XmT2018004). We also acknowledge the support from the
7

H. Chen, et al.
AppliedCatalysisA,General588(2019)117283
Scheme 1. The tentative reaction mechanism for styrene epoxidation using DMF solvent and H2O2 oxidant over Cr-PKU-1 catalyst.
Fundamental Research Funds for the Central Universities
(2019CDXYHG0013 and 2019CDQYWL009).
[22] Z.Q. Shi, Z.P. Dong, J. Sun, F.W. Zhang, H.L. Yang, J.H. Zhou, X.H. Zhu, R. Li,
Chem. Eng. J. 237 (2014) 81–87.
[23] Y. Zhang, F.L. Yang, R.H. Gao, W.L. Dai, Appl. Surf. Sci. 471 (2019) 767–775.
[24] J. Park, Y.M. Lee, K. Ohkubo, W. Nam, S. Fukuzumi, Inorg. Chem. 54 (2015)
5806–5812.
Appendix A. Supplementary data
[25] R. Tan, D.H. Yin, N.Y. Yu, H.H. Zhao, D.L. Yin, J. Catal. 263 (2009) 284–291.
[26] K. Matsumoto, T. Oguma, T. Katsuki, Angew. Chem. Int. Ed. 48 (2009) 7432–7435.
[27] E. Zahra, R. Hajar, Z. Saeed, Appl. Organometal. Chem. 33 (2019) 1–6.
[28] J.F. Hui, H.M. Chu, W.L. Zhang, Y. Shen, W.Z. Chen, Y. Hu, W. Liu, C. Gao,
S.H. Guo, G.W. Xiao, S. Li, Y. Fu, D.D. Fan, W.N. Zhang, F.W. Huo, Nanoscale 10
(2018) 8772–8778.
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.apcata.2019.117283.
References
[29] J.W. Brown, Q.T. Nguyen, T. Otto, N.N. Jarenwattananon, S. Gloggler,
L.S. Bouchard, Catal. Commun. 59 (2015) 50–54.
[30] W. Gao, X.F. Chen, J.C. Li, Q. Jiang, J. Phys. Chem. C 114 (2010) 1148–1153.
[31] L. Zhou, C.F. Gorin, R.J. Madix, J. Am. Chem. Soc. 132 (2010) 434–435.
[32] Y.M. Liu, A.T. Tsunoyama, T. Tsukuda, Chem. Commun. 46 (2010) 550–552.
[33] D.S. Pinnaduwage, L. Zhou, W.W. Gao, C.M. Friend, J. Am. Chem. Soc. 129 (2007)
1872–1873.
[34] Z. Opre, T. Mallat, A. Baiker, J. Catal. 245 (2007) 482–486.
[35] V.R. Choudhary, R. Jha, P. Jana, Green Chem. 8 (2006) 689–690.
[36] J. Ju, J.H. Lin, G.B. Li, T. Yang, H.M. Li, F.H. Liao, C.K. Loong, L.P. You, Angew.
Chem. Int. Ed. 42 (2003) 5607–5610.
[1] C.N. Dai, J. Zhang, C.P. Huang, Z.G. Lei, Chem. Rev. 117 (2017) 6929–6983.
[2] D.S. Su, G.D. Wen, S.C. Wu, F. Peng, R. Schlogl, Angew. Chem. Int. Ed. 56 (2017)
936–964.
[3] L.S. Anne, Z.M. Inmar, S.D. Jonathan, J. Am. Chem. Soc. 128 (2006) 14272–14273.
[4] D.T. Bregante, A.Y. Patel, A.M. Johnson, D.W. Flaherty, J. Catal. 364 (2018)
415–425.
[5] S. Mukherjee, M.A. Vannice, J. Catal. 243 (2006) 108–130.
[6] N.M. Bertero, A.F. Trasarti, C.R. Apesteguia, A.J. Marchi, Appl. Catal. A 394 (2011)
228–238.
[37] T. Yang, A. Bartoszewicz, J. Ju, J.L. Sun, Z. Liu, X.D. Zou, Y.X. Wang, G.B. Li,
F.H. Liao, B. Martin-Matute, J.H. Lin, Angew. Chem. Int. Ed. 50 (2011)
12555–12558.
[38] H. Chen, J. Ju, Q.P. Meng, J. Su, C. Lin, Z.Y. Zhou, G.B. Li, W.L. Wang, W.L. Gao,
C.M. Zeng, C. Tang, J.H. Lin, T. Yang, J.L. Sun, J. Am. Chem. Soc. 137 (2015)
7047–7050.
[39] J. Ju, T. Yang, G.B. Li, F.H. Liao, Y.X. Wang, L.P. You, J.H. Lin, Chem.-Eur. J. 10
(2004) 3901–3906.
[40] T. Yang, J. Ju, G.B. Li, F.H. Liao, X.D. Zou, F. Deng, L. Chen, Y.X. Wang, J.H. Lin,
Inorg. Chem. 46 (2007) 4772–4774.
[7] B.S. Akpa, C. D’Agostino, L.F. Gladden, K. Hindle, H. Manyar, J. McGregor, R. Li,
M. Neurock, N. Sinha, E.H. Stitt, D. Weber, J.A. Zeitler, J. Catal. 289 (2012) 30–41.
[8] G. Dodekatos, J. Ternieden, S. Schunemann, C. Weidenthaler, H. Tuysuz, Catal. Sci.
Technol. 8 (2018) 4891–4899.
[9] W.B. Fan, P. Wu, T. Tatsumi, J. Catal. 256 (2008) 62–73.
[10] N. Miniajluk, J. Trawczynski, M. Zawadzki, Appl. Catal. A 531 (2017) 119–128.
[11] E. Kockrick, T. Lescouet, E.V. Kudrik, A.B. Sorokin, D. Farrusseng, Chem. Commun.
24 (2011) 1562–1564.
[12] P. Anangamohan, Ch. J. Narayan, B. Paula, Mol. Catal. 449 (2018) 49–61.
[13] Y.Q. Li, S.C. Zhao, Q.M. Hu, Z. Gao, Y.Q. Liu, J.K. Zhang, Y. Qin, Catal. Sci. Technol.
7 (2017) 2032–2038.
[14] M. Fadhli, I. Khedher, J.M. Fraile, J. Mol. Catal. A 420 (2016) 282–289.
[15] J.M. Fraile, N. Garcia, J.A. Mayoral, F.G. Santomauro, M. Guidotti, ACS Catal. 5
(2015) 3552–3561.
[16] H.T. Cui, Y. Zhang, Z.G. Qiu, L.F. Zhao, Y.L. Zhu, Appl. Catal. B 101 (2010) 45–53.
[17] A. Corma, P. Esteve, A. Martınez, J. Catal. 161 (1996) 11–19.
[18] D. Nepak, D. Srinivas, Appl. Catal. A 523 (2016) 61–72.
[19] X. Liu, J. Ding, X. Lin, R.H. Gao, Z.H. Li, W.L. Dai, Appl. Catal. A 503 (2015)
117–123.
[20] W. Jia, Y.X. Liu, P. Hu, R. Yu, Y. Wang, L. Ma, D.S. Wang, Y.D. Li, Chem. Commun.
51 (2015) 8817–8820.
[21] A.F. Zeng, Y. Li, S.P. Su, D.B. Li, B. Hou, N.Y. Yu, J. Catal. 319 (2014) 163–173.
[41] W.L. Gao, Y.X. Wang, G.B. Li, L.P. You, J.H. Lin, Inorg. Chem. 47 (2008)
7080–7082.
[42] W.L. Gao, Y. Jing, J. Yang, Z.Y. Zhou, D.F. Yang, J.L. Sun, J.H. Lin, R.H. Cong,
T. Yang, Inorg. Chem. 53 (2014) 2364–2366.
[43] W.L. Gao, T. Yang, Y.X. Wang, G.B. Li, F.H. Liao, J.H. Lin, J. Phys. Chem. B 109
(2005) 22775–22779.
[44] W.L. Wang, S.X. Hu, L.J. Li, W.L. Gao, R.H. Cong, T. Yang, J. Catal. 352 (2017)
130–141.
[45] R. Shannon, Acta Crystallogr. Sect. A 32 (1976) 751–767.
[46] Y. Zhang, N. Zhang, Z.R. Tang, Y.J. Xu, Chem. Sci. 4 (2013) 1820–1824.
[47] K.C. Yang, J.H. Zheng, Y.L. Chen, K.C. Lee, E.C. Cho, RSC Adv. 6 (2016)
53025–53033.
[48] H. Goto, Y. Hanada, T. Ohno, M. Matsumura, J. Catal. 225 (2004) 223–229.
8

H. Chen, et al.
AppliedCatalysisA,General588(2019)117283
[49] Y.N. Wang, K.J. Deng, L.Z. Zhang, J. Phys. Chem. C 115 (2011) 14300–14308.
[50] F. Raziq, Y. Qu, X.L. Zhang, M. Humayun, J. Wu, A. Zada, H.T. Yu, X.J. Sun,
L.Q. Jing, J. Phys. Chem. C 120 (2016) 98–107.
[51] J. Hu, J. Men, Y.Y. Liu, H. Huang, T.F. Jiao, RSC Adv. 5 (2015) 54028–54036.
[52] J. Sebastian, K.M. Jinka, R.V. Jasra, J. Catal. 244 (2006) 208–218.
[53] J.C. van der Waal, H. van Bekkum, J. Mol. Catal. A 124 (1997) 137–146.
[54] H.T. Cui, Y. Zhang, L.F. Zhao, Y.L. Zhu, Catal. Commun. 12 (2011) 417–420.
[55] S. Rahman, C. Santra, R. Kumar, J. Bahadur, A. Sultana, R. Schweins, D. Sen,
S. Maity, S. Mazumdar, B. Chowdhury, Appl. Catal. A 482 (2014) 61–68.
9