Catalytic epoxidation of propylene over a Schiff-base molybdenum complex supported on a silanized mesostructured cellular foam

Article abstract of DOI:10.1007/s11164-020-04232-6

A Schiff-base molybdenum complex (MoO₂–salen) supported on mesostructured cellular foam (MCF) was initially prepared by an in situ synthesis method under acidic conditions. Following silanization modification, a MoO₂–salen?MCF-S sample with improved surface hydrophobicity was obtained. The ligand environment of molybdenum within the samples has been analyzed by Fourier-transform infrared spectroscopy, ultraviolet/visible spectroscopy, and X-ray photoelectron spectroscopy. Furthermore, the textural and structural properties of the corresponding materials have been characterized by nitrogen adsorption–desorption isotherms and transmission electron microscopy. Despite of the presence of fewer MoO₂ species, the results showed that MoO₂–salen?MCF-S has more active Mo centers than MoO₂–salen and MoO₂–salen?MCF on the basis of maintaining the mesoporous structure. The catalytic performances of the synthesized samples were assessed in the epoxidation of propylene with tert-butyl hydroperoxide (TBHP) as an oxidant, and the mechanism of propylene epoxidation under MoO₂–salen?MCF was given. The prepared MoO₂–salen?MCF-S material showed the best epoxidation performance with 1,2-dichloroethane as a solvent and a molar ratio of propylene to TBHP of 10:1 at 120 °C, giving a TBHP conversion of up to 100% after 1 h, with selectivities for propylene oxide and tert-butyl alcohol reaching 94.7% and 84.6%, respectively.

Full text of DOI:10.1007/s11164-020-04232-6

Research on Chemical Intermediates
https://doi.org/10.1007/s11164-020-04232-6
Catalytic epoxidation of propylene over a Schiꢀ‑base
molybdenum complex supported on a silanized
mesostructured cellular foam
1
1
1
1
Dawei Chen · Xiaoyu Zhang · Haoran Jiang · Xia Yuan
Received: 20 June 2020 / Accepted: 3 August 2020
©
Springer Nature B.V. 2020
Abstract
A Schif-base molybdenum complex (MoO –salen) supported on mesostructured
2
cellular ꢀoam (MCF) was initially prepared by an in situ synthesis method under
acidic conditions. Following silanization modiꢁcation, a MoO –salen@MCF-S
2
sample with improved surꢀace hydrophobicity was obtained. The ligand environ-
ment oꢀ molybdenum within the samples has been analyzed by Fourier-transꢀorm
inꢀrared spectroscopy, ultraviolet/visible spectroscopy, and X-ray photoelectron
spectroscopy. Furthermore, the textural and structural properties oꢀ the correspond-
ing materials have been characterized by nitrogen adsorption–desorption isotherms
and transmission electron microscopy. Despite oꢀ the presence oꢀ ꢀewer MoO spe-
2
cies, the results showed that MoO –salen@MCF-S has more active Mo centers than
2
MoO –salen and MoO –salen@MCF on the basis oꢀ maintaining the mesoporous
2
2
structure. The catalytic perꢀormances oꢀ the synthesized samples were assessed
in the epoxidation oꢀ propylene with tert-butyl hydroperoxide (TBHP) as an oxi-
dant, and the mechanism oꢀ propylene epoxidation under MoO –salen@MCF was
2
given. The prepared MoO –salen@MCF-S material showed the best epoxidation
2
perꢀormance with 1,2-dichloroethane as a solvent and a molar ratio oꢀ propylene to
TBHP oꢀ 10:1 at 120 °C, giving a TBHP conversion oꢀ up to 100% aꢀter 1 h, with
selectivities ꢀor propylene oxide and tert-butyl alcohol reaching 94.7% and 84.6%,
respectively.
Keywords Mesostructured cellular ꢀoam · Propylene epoxidation · Schif-base
molybdenum complex · Silanization modiꢁcation
*
Xia Yuan
yxiamail@163.com
1
School oꢀ Chemical Engineering, Xiangtan University, Xiangtan 411105, Hunan, China
Vol.:(012
1
3456789)
3

D. Chen et al.
Introduction
Propylene oxide (PO) is an important industrial raw material ꢀor the production
oꢀ waterprooꢀ materials, coatings, and other products [1]. Due to its high poten-
tial application value and economic value, the production process is constantly
developing and progressing [2–4]. Several methods are used ꢀor the conversion
oꢀ propylene to propylene oxide, including chlorohydrination, co-oxidation, and
hydrogen peroxide direct oxidation [5]. The chlorohydrin method has gradually
been eliminated due to pollution problems [6]. At present, the main process used
in industry is the co-oxidation method, which involves oxidation oꢀ a third sub-
stance to obtain a predetermined oxygen source, and then, the propylene epoxi-
dation reaction is carried out with this oxygen source to ꢁnally ꢀorm propylene
oxide and the corresponding by-product. Compared with the ethylbenzene co-
oxidation process using ethylbenzene hydroperoxide as the oxygen source, the
isobutane co-oxidation process with tert-butyl hydroperoxide (TBHP) as the
oxygen source is more mature [7]. The catalysts used in this process are mainly
homogeneous molybdenum-based systems, the most representative oꢀ which are
Mo(CO) /MoO (acac) used by Arco (USA) and (CH OH) /(NH ) Mo O used
6
2
2
2
2
4 6
7
24
by Texace (USA). The conversion rate oꢀ TBHP and PO selectivity over these two
catalysts can exceed 98% in the corresponding processes. However, most homo-
geneous molybdenum-based catalysts are diꢂcult to recover, the product is diꢂ-
cult to separate and reꢁne, and long-term use oꢀ the catalyst leads to deposition in
the reactor [8]. Heterogeneous catalysts have been a research hotspot concerning
the epoxidation oꢀ propylene due to their inherent advantage oꢀ easy separation
ꢀ
rom the reaction system [9]. Thereꢀore, the development oꢀ suitable heterogene-
ous catalysts ꢀor the epoxidation oꢀ propylene through a co-oxidation process has
high signiꢁcance and research value.
Schif-base metal complexes are active components with diverse composi-
tions and easily modiꢁable structure, and can be used to obtain highly eꢂcient
and stable heterogeneous catalysts by various loading methods. Due to the rich
variety oꢀ ligand groups, there are many types oꢀ corresponding complexes, and
they are easily modiꢁed [10–13]. Mo–salen (salen = N,N′-bis(salicylidene)eth-
ylenediamine), as a unique Schif-base metal complex, is not uncommon in the
catalytic epoxidation oꢀ oleꢁns [14–16]. Bagherzadeh et al. [17] synthesized a
Mo–salen polymer, which was used as a heterogeneous catalyst ꢀor the epoxi-
dation oꢀ oleꢁns and sulꢀoxidation, ofering selectivities oꢀ 99% to sulꢀoxides.
However, molybdenum complexes have the same problems as other metal com-
plexes, namely poor stability and diꢂculty in recycling, which greatly limit their
practical applications. The use oꢀ a carrier to immobilize Mo complexes, better
dispersing the active component oꢀ the catalyst, thereby improving its stabil-
ity and catalytic eꢂciency per unit active center, has proven to be an efective
strategy. Suitable carriers mainly include high molecular weight polymers, metal
oxides, metal–organic ꢀrameworks, and mesoporous silicon materials [18–21].
Long et al. [22] used MoO(O ) (DMF) as a Mo source, and Mo–salen was suc-
2
2
2
cessꢀully loaded onto halloysite nanotubes to obtain HNTs-APTMS-Mo-SL by
1
3

Catalytic epoxidation of propylene over a Schiꢀ‑base…
surꢀace chemical modiꢁcation and selꢀ-assembly. This research group used TBHP
as the oxygen source and evaluated the perꢀormance oꢀ the catalyst in cyclohex-
ene epoxidation as a probe reaction. The results showed that the conversion oꢀ
cyclohexene reached 95%, and the selectivity ꢀor epoxycyclohexane reached 96%.
Sharghi et al. [23] loaded Mo–salen on mesoporous SiO to obtain a supported
2
Mo catalyst. This catalyst was applied in the one-pot synthesis oꢀ 2,4,5-triar-
ylimidazoles via a three-component reaction using benzil or benzoin, aryl alde-
hydes, and ammonium acetate as a nitrogen source under mild conditions, and
showed higher activity and selectivity, giving yields in excess oꢀ 90%. Moreo-
ver, this catalyst maintained its structural integrity aꢀter ꢀour or ꢁve recoveries.
Bagherzadeh et al. [24] used MoO (sal-phz)(CH OH) as a molybdenum source
2
3
and loaded the active Mo component onto SBA-15 by an impregnation method to
aford MoO (sal-phz)(CH OH)/SBA-15. The catalyst exhibited good activity and
2
3
selectivity in the epoxidation oꢀ cyclic and long-chain oleꢁns. These authors also
ꢀ
ocused on the efect oꢀ solvents on the activity and selectivity oꢀ cyclooctene
epoxidation.
Mesoporous cellular silica ꢀoam (MCF) is an aerogel mesoporous material with
a three-dimensional interconnected pore structure oꢀ ꢃexible size [25–27]. The pore
structure ꢀacilitates the entry oꢀ reactants and active components, reduces mass-
transꢀer limitations when the product leaves, and ensures that the active sites are
evenly distributed. Compared with common molecular sieves such as SBA-15, it is
more conducive to the difusion oꢀ raw materials and products with larger molecular
size in the reaction system, and its preparation process is simple and permits easy
chemical modiꢁcation. Miao et al. [28] pointed out that Mo supported on MCF was
best oꢀ diferent types oꢀ silica supports tested ꢀor the same reaction using a hydrop-
eroxide oxidant. In this work, mesoporous silica ꢀoam has been used as a carrier, and
a Mo–Schif-base complex has been used to construct a mesoporous silica ꢀoam-
supported metal complex as a heterogeneous catalyst ꢀor PO production with TBHP
as the oxygen source. The aim oꢀ this study was to provide valuable data ꢀor the
development oꢀ heterogeneous catalytic systems ꢀor propylene epoxidation.
Experimental
Materials
PEO-PPO-PEO triblock copolymer (P123, MW=5800), isopropyl titanate (95%),
TBHP (70%), and hexamethyldisilazane (HMDS) were obtained ꢀrom Latin Biotech-
nology Co., Ltd. (China); molybdenum acetylacetonate (97%), mesitylene (TMB,
9
7%), ammonium ꢃuoride, and tetraethyl orthosilicate (TEOS) were obtained ꢀrom
Maclean Biotechnology Ltd. (China); propylene (99.8%) was obtained ꢀrom Shan-
dong Yue’an Chemical Co., Ltd. (China); and 1,2-dichloroethane, salicylaldehyde,
and ethylenediamine were obtained ꢀrom Tianjin Kemiou Chemical Reagent Co.,
Ltd. (China). It should be noted that all reagents were used without ꢀurther puriꢁca-
tion, and all solutions were prepared with distilled deionized water.
1
3

D. Chen et al.
Preparation of the catalyst
Preparation of MoO –salen
2
Salicylaldehyde (6.11 g) and ethylenediamine (1.50 g) were separately dissolved in
2
0 mL oꢀ absolute ethanol, and each solution was stirred ꢀor 30 min. Thereaꢀter, the
ethylenediamine solution was slowly dropped into the salicylaldehyde solution, and
the mixture was heated under reꢃux ꢀor 1 h at 79 °C under N atmosphere. It was
2
then allowed to cool to room temperature, beꢀore being ꢁltered. The collected solid
was washed two or three times with absolute ethanol, then recrystallized twice ꢀrom
absolute ethanol, and ꢁnally transꢀerred to a vacuum drying oven ꢀor 24 h at 79 °C.
The obtained golden-yellow ꢃaky solid was designated as salenH .
2
Next, molybdenum acetylacetonate (MoO (acac) , 1.63 g) was taken up in abso-
2
2
lute ethanol (75 mL) with stirring. A portion oꢀ the above-produced salenH (1.34 g)
2
was added, and the mixture was heated under reꢃux at 79 °C ꢀor 24 h under a nitro-
gen atmosphere to obtain a dark-brown solution. The mixture was then washed three
times each with absolute ethanol and acetone, ꢁltered, and the collected solid was
dried in vacuo ꢀor 12 h at 79 °C. The obtained dark-brown solid was designated as
MoO –salen. A schematic diagram oꢀ the preparation process is given in Scheme 1.
2
Preparation of MoO –salen@MCF
2
First, P123 templating agent (4.00 g) was weighed into a three-necked ꢃask, 1.6 M
aqueous HCl (150 mL) was added, and the mixture was gently stirred at 40 °C until
the P123 had completely dissolved. TMB (4.00 g) and NH F (0.05 g) were then
4
sequentially added and the mixture was stirred ꢀor 1 h at 40 °C. TEOS (8.80 g) was
then added and the mixture was maintained ꢀor 12 h at 40 °C. Finally, the requi-
site amount oꢀ MoO –salen was added, and stirring was continued ꢀor 12 h. The
2
dark-green mixture obtained was transꢀerred to a crystallization vessel and crystal-
lized ꢀor 24 h at 100 °C. Thereaꢀter, the product was allowed to cool naturally to
room temperature and then washed with deionized water until no obvious bubble.
The obtained dark-green solid was dried at room temperature ꢀor 24 h. Finally, the
O
N
N
H N
2
CH
HC
CH
C H OH
2
5
2
+
(salenH )
2
N , Reflux
2
OH
HO
OH H N
2
N
N
HC
O
CH
O
C H OH
2
5
+
MoO (acac)
2 2
(MoO -salen)
2
N , Reflux
2
Mo
O
O
Scheme 1 Illustration oꢀ preparation oꢀ MoO –salen
2
1
3

Catalytic epoxidation of propylene over a Schiꢀ‑base…
template was removed by Soxhlet extraction. Absolute ethanol (300 mL) mixed with
concentrated hydrochloric acid (10 mL) was used as a washing solution, and the
sample was washed ꢀor 48 h at 100 °C. Thereaꢀter, it was dried in a vacuum drying
oven ꢀor 12 h at 79 °C to obtain MoO –salen@MCF. The synthetic process ꢀor MCF
2
was the same as that described above, except that no MoO –salen was added. A
2
schematic diagram oꢀ the preparation process is shown as (1) in Scheme 2.
Silanization modiꢁcation of MoO –salen@MCF
2
MoO –salen@MCF (0.60 g) was weighed into a three-necked ꢃask, and the silylat-
2
ing agent HMDS (0.60 g) and toluene (50 mL) were added. The mixture was stirred
at room temperature ꢀor 1 h, then warmed to 110 °C and heated under reꢃux ꢀor
1
2 h. Aꢀter the mixture was cooled, it was washed with toluene and ꢁltered three
times. Finally, the obtained solid was dried in a vacuum drying oven at 110 °C ꢀor
1
2 h to obtain MoO –salen@MCF-S. A schematic diagram oꢀ the silylation process
2
is shown as (2) in Scheme 2.
Catalyst characterization
Fourier-transꢀorm inꢀrared (FTIR) spectroscopy was carried out on
a
Thermo Electron Corporation (USA) Nicolet-380 instrument in the region
−
1
4
000–400 cm , ꢀor which the sample was ground with KBr and pressed into a
thin waꢀer. Ultraviolet/visible (UV–Vis) absorption spectrophotometry was car-
ried out using a UV-2550 spectrophotometer manuꢀactured by Shimadzu Corpo-
ration (Japan); high-purity BaSO was used as a reꢀerence, and the measurement
4
range was 200–800 nm. X-ray photoelectron spectroscopy (XPS) was perꢀormed
on an ESCALAB 250Xi X-ray photoelectron spectrometer manuꢀactured by
Thermo Fisher Scientiꢁc, with an energy scan range oꢀ 0–5000 eV. The sur-
ꢀace areas and pore volumes oꢀ the prepared samples were calculated using the
Scheme 2 Illustration oꢀ preparation oꢀ silylated modiꢁed MoO –salen@MCF
2
1
3

D. Chen et al.
Brunauer–Emmett–Teller (BET) equation and the Barrett–Joyner–Halenda (BJH)
model, respectively, using a Quantachrome Instruments (USA) NOVA-2200e
surꢀace-area analyzer at − 196 °C. Prior to measurements oꢀ N adsorption, all
2
samples were out-gassed at 100 °C ꢀor 24 h. Transmission electron microscopy
(
TEM) was carried out with a JEM-2100 TEM instrument (JEOL, Japan) oper-
ated at 200 kV. The molybdenum content oꢀ the catalyst was determined using a
coupled plasma atomic emission spectrometry (ICP-AES) device (Thermo Jarrell
Asch, USA; IRIS Advantage 1000).
Evaluation of the propylene epoxidation reaction
The epoxidation reaction oꢀ propylene was carried out in an autoclave (100 mL)
placed in a temperature equilibrated oil bath. In a typical experiment, the mix-
ture oꢀ 21.35 g (in C H Cl , 16.23%) TBHP, a 3 cm rotor and 0.36 g catalyst
2
4
2
were placed into autoclave. Propylene in liquid ꢀorm (room temperature, 16.2 g)
was then introduced into the autoclave through an advection pump pressing to
1
0 MPa by the back-pressure valve installed in the pipeline oꢀ outlet oꢀ advection
pump. Leak detection was perꢀormed aꢀter the material had been introduced. The
autoclave was placed in an oil bath and heated at a set temperature ꢀor a certain
period oꢀ time, and the rotation speed was adjusted to 200 rpm. Aꢀter comple-
tion oꢀ the reaction, the vessel was cooled in an ice–water mixture, and when the
temperature was 10 °C or lower, unreacted propylene was discharged through an
exhaust valve and taken up in a suitable solvent. Aꢀter ensuring that no residual
gas remained in the reactor, the reaction solution was removed ꢀor qualitative and
quantitative analysis. A schematic diagram oꢀ the reaction process is shown in
Scheme 3.
PO and tert-butyl alcohol (TBA) were analyzed by an internal standard tech-
nique on a gas chromatograph (Agilent GC-6820 with ꢃame ionization detector),
equipped with a DB-1701 column, N as a carrier gas and n-pentane was used
2
as an internal standard. The conversion oꢀ TBHP was evaluated by iodometric
analysis based on Chinese National Standard GB/T 32102-2015. The conversion
oꢀ TBHP (X
) and the selectivities ꢀor PO and TBA (S , S ) were based on
PO TBA
TBHP
the amount oꢀ TBHP converted during the reaction.
OH
O
OH
O
Propylene
Tert butyl hydroperoxide
TBHP)
Propylene oxide
(PO)
Tert butyl alcohol
(TBA)
(
Scheme 3 Illustration oꢀ propylene epoxidation
1
3

Catalytic epoxidation of propylene over a Schiꢀ‑base…
Results and discussion
Characterization results
FTIR analysis
−
1
It can be seen ꢀrom Fig. 1 that the absorption peak oꢀ the MCF sample at 3430 cm
was mainly due to hydrogen bonding between Si–OH and water molecules [29].
In the low-wavenumber region, absorption peaks appearing at ν=1081, 803, and
−1
4
68 cm can be attributed to stretching vibrations oꢀ Si–O–Si in the SiO struc-
2
−
1
ture, and a small and sharp absorption peak at ν=966 cm can be assigned to
an≡ Si–OH stretching vibration. The spectral ꢀeatures indicate that the silicon source
was hydrolyzed to ꢀorm a skeleton oꢀ mesoporous silica ꢀoam. The absorption peak
oꢀ the P123 template was no longer seen, conꢁrming its complete removal.
−
1
Two adjacent absorption bands were seen at ν=903 and 933 cm , attributable
to Mo=O vibrations, which are characteristic oꢀ a cis-MoO group. The absorption
2
−
1
peak at ν=751 cm can be attributed to υ
, directly conꢁrming the coordination
Mo–O
oꢀ Mo by –OH in salenH and the successꢀul synthesis oꢀ MoO –salen [21, 30].
2
2
The FTIR spectra oꢀ MCF and MoO –salen@MCF were basically the same, indi-
2
cating no signiꢁcant changes in the structure oꢀ MCF aꢀter loading oꢀ the complex.
−
1
Except ꢀor the typical C–H vibration at 3000–2700 cm , no IR absorption peak
−
1
corresponding to MoO –salen was seen in the range ν=3500–500 cm due to the
2
shielding efect oꢀ the MCF skeleton.
The original absorption peaks oꢀ MoO –salen appeared at ν=933, 903, 751,
2
−1
and 629 cm . Aꢀter mechanical mixing oꢀ MoO –salen with MCF, the other peaks
2
attributed to Mo coordination disappeared, except ꢀor the peak intensity correspond-
ing to 751 cm− becoming lower, but the peaks in the range ν=1630–1260 cm−1
1
8
03
3
430
468
1
081
MoO -salen@MCF
2
Mo
6
29
9
33
903
7
51
MoO -salen
2
MCF+MoO
MCF
2-salen
4
68
803
9
66
4
000
3500
3000
2500
2000
1500
1000
500
-1
Wavenumbers(cm )
Fig. 1 FT-IR spectra oꢀ samples
1
3

D. Chen et al.
persisted, illustrating the masking efect oꢀ MCF on the inꢀrared response oꢀ
MoO –salen.
2
UV–Vis analysis
As can be seen in Fig. 2, MoO –salen gives rise to a peak at 480 nm, attributa-
2
ble to an O(p )→Mo(d ) charge-transꢀer transition. Aꢀter mechanical mixing oꢀ
π
π
MoO –salen with MCF, this peak still appeared. When the MoO –salen complex
2
2
was loaded on the MCF carrier in situ [16]; however, the peak shiꢀted to 321 nm,
indicative oꢀ interactions between the components. Hence, the coordination environ-
ment was diferent ꢀrom that in the simple complex.
XPS analysis
Figure 3 shows the photoelectron spectra oꢀ Mo, O, and Si in the MoO –salen@
2
MCF sample. Comparing MoO –salen with MoO –salen@MCF, it can be seen that
2
2
the two samples have the same Mo 3p and Mo 3d energy peaks, but the peak inten-
sity oꢀ the latter is much lower than that oꢀ the ꢀormer. This is because the ꢀorm oꢀ
Mo(VI) does not change during the loading oꢀ MoO –salen onto MCF by the in situ
2
method, but dispersion on the carrier reduces the relative content oꢀ Mo(VI) in the
sample. Two main peaks are seen in the Mo 3d spectrum oꢀ MoO –salen@MCF, at
2
2
32.2 and 235.2 eV, attributable to Mo 3d and Mo 3d , respectively. This con-
5
/2
3/2
ꢁ
rms that the active center is Mo(VI) [31]. The Mo 3d energy spectrum peak oꢀ
MoO –salen@MCF shiꢀts slightly to lower electron-binding capacity compared to
2
that oꢀ MoO –salen, which is mainly due to loading on the MCF carrier changing
2
the coordination state oꢀ Mo. The O 1 s energy peak oꢀ MoO –salen is located at
2
5
30.7 eV, which is attributed to the typical lattice oxygen in molybdenum oxide,
MCF+MoO -salen
2
480
MoO -salen
2
MoO
2-salen@MCF
321
MCF
200
400
600
800
Wavelength(nm)
Fig. 2 UV–Vis spectra oꢀ samples
1
3

Catalytic epoxidation of propylene over a Schiꢀ‑base…
Mo 3d
Mo 3p
2
32.9
2
36.0
(
a)
(
a)
b)
2
35.2
2
32.2
(
(b)
2
20
225
230
235
240
380
390
400
410
420
430
Binging Energy(ev)
Binding Energy(ev)
5
30.7
Si 2p
O 1s
a)
(
533
(
b)
(b)
5
24 526 528 530 532 534 536 538 540
94 96 98 100 102 104 106 108 110 112
Binding Energy(ev)
Binding Energy(ev)
Fig. 3 XPS diagram oꢀ samples. a MoO –salen, b MoO –salen@MCF
2
2
whereas the O 1 s peak at 533 eV ꢀor MoO –salen@MCF is associated with oxygen
2
in MCF support (Si–O–Si). The MoO –salen@MCF sample also showed a strong
2
Si 2p absorption peak at 104 eV, attributable to the silicon-containing component in
the MCF carrier. From the above analysis, it was concluded that MoO –salen@MCF
2
had been successꢀully produced by the in situ method.
N2 physical adsorption–desorption analysis
In Fig. 4, it can be seen that both MCF and MoO –salen@MCF show type IV
2
hysteresis loops in their N physical adsorption–desorption isotherms, indi-
2
cating mesoporous structures. From the pore size distributions, the pore size
oꢀ MoO –salen@MCF is smaller than that oꢀ MCF, which is consistent with
2
MoO –salen occupying the pores oꢀ the support.
2
TEM analysis
The pore structure oꢀ MCF and the distribution oꢀ active components oꢀ
MoO –salen@MCF were scrutinized by TEM. It can be seen ꢀrom Fig. 5a that the
2
1
3

D. Chen et al.
1
1
1
40
20
00
MCF
MoO2-salen@MCF
0
0
0.30
0
0
0
0
0
0
.40
.35
.25
.20
.15
.10
.05
.00
8
6
4
2
0
0
0
0
0
-0.05
0
10
20
30
40
Pore Diameter(nm)
0
.0
0.2
0.4
0.6
0.8
1.0
P/P0
Fig. 4 N physical adsorption–desorption isotherm curve and pore size distribution oꢀ the samples
2
Fig. 5 TEM oꢀ samples. a MCF b MoO –salen@MCF
2
prepared MCF had a theoretical “ꢀoamy” pore structure. Figure 5b shows that the
active component in the prepared MoO –salen@MCF was relatively well dispersed.
2
MoO –salen@MCF‑catalyzed epoxidation of propylene with an anhydrous oxygen
2
source system
The water in the aqueous TBHP solution was removed by azeotropic distillation.
The speciꢁc process is as ꢀollows: a certain proportion oꢀ 1,2-dichloroethane is
added to the TBHP source having a H O content oꢀ 30% to ꢀorm an azeotrope
2
with water. The majority oꢀ H O contained in TBHP was removed by azeotropic
2
distillation at 30 °C under a negative pressure (− 0.09 MPa), the remaining water
is removed by adding an excess oꢀ anhydrous Na CO ꢀor 12 h, and ꢁnally obtain-
2
3
ing a TBHP solution using C H Cl as a solvent. It can be seen ꢀrom Table 1
2
4
2
that propylene epoxidation is almost inactive without catalyst and the catalytic
1
3

Catalytic epoxidation of propylene over a Schiꢀ‑base…
Table 1 Samples catalyzed propylene epoxidation under diferent oxygen sources
Catalyst
Oxidant
Conversion/%
Selectivity
PO)/%
Selec-
tivity
(
(
TBA)/%
None
TBHP (in H O)
14.9
15.2
51.7
97.8
73.7
99.5
15.7
20.1
90.4
77.7
47.7
79.5
–
–
100
81.6
99.4
82.1
2
TBHP (in C H Cl )
2
4
2
MoO –salen
TBHP (in H O)
2
2
TBHP (in C H Cl )
2
4
2
MoO –salen@MCF
TBHP (in H O)
2
2
TBHP (in C H Cl )
2
4
2
Reaction conditions: oxidant 0.0385 mol, C H Cl 12.51 g, propylene 0.385 mol, m (catalyst)=1.69 wt%
2
4
2
(
relative to the initial amount oꢀ TBHP and solvent), Si/Mo=6.5, temperature 120 °C, time 1 h
activities oꢀ MoO –salen and MoO –salen@MCF in the epoxidation oꢀ propylene
2
2
using TBHP in C H Cl were higher than those in the aqueous TBHP system.
2
4
2
It is generally known that MoO –salen is sensitive to aqueous solution; water
2
can change the structure oꢀ the complex and deactivate the catalyst. Anhydrous
TBHP in C H Cl was thereꢀore used as the oxygen source in the ꢀollowing work.
2
4
2
MoO –salen@MCF catalyzes the epoxidation oꢀ propylene under the condition
2
oꢀ anhydrous oxygen source. The conversion oꢀ TBHP is 99.5%, reaching a very
high level, but the selectivity oꢀ PO is only 79.5%, which is related to the content
oꢀ molybdenum.
Eꢀect of silicon/molybdenum ratio on the preparation of MoO –salen@MCF
2
by the in situ method
The preparation process oꢀ MoO –salen@MCF by the in situ method was opti-
2
mized. The efect oꢀ silicon/molybdenum ratio on its catalytic perꢀormance was
investigated in detail. It can be seen ꢀrom Fig. 6 and Table 2 that an increase in
the ratio oꢀ silicon/molybdenum ratio (its range is 3.5–8.0, and the amount oꢀ
Mo was varied ꢀor a constant amount oꢀ silica support) did not destroy the pore
structure oꢀ MCF. All oꢀ the samples retained a type IV hysteresis loop, but the
amount adsorbed gradually decreased. This was because the volume oꢀ the MCF
was greatly reduced due to the excessive proportion oꢀ MoO –salen.
2
It can be seen ꢀrom Fig. 7 that the TBHP conversion rate in the propylene
epoxidation reaction increased with decreasing silicon/molybdenum ratio in the
in situ preparation process oꢀ MoO –salen@MCF, whereas the selectivity ꢀor PO
2
ꢁ
rst increased and then decreased. There was almost no change in the selectivity
ꢀ
or TBA. The experimental results showed that the optimal silicon/molybdenum
ratio oꢀ MoO –salen@MCF prepared by the in situ method was 4.5:1, which gave
2
a TBHP conversion oꢀ 99.7%, a PO selectivity oꢀ 84.8%, a TBA selectivity oꢀ
−
1
8
1.3%, and a TOF oꢀ 161.6 h .
1
3

D. Chen et al.
1
60
20
Si:Mo=3.5
Si:Mo=4.5
Si:Mo=5.0
Si:Mo=6.5
Si:Mo=8.0
1
0
0
0
0
0
0
.30
.25
.20
.15
.10
.05
8
4
0
0
0
0.00
0
10
20
30
40
50
60
Pore Diameter(nm)
0
.0
0.2
0.4
0.6
0.8
1.0
P/P0
Fig. 6 N physical adsorption–desorption isotherm curve and pore size distribution oꢀ samples under diꢀ-
2
ꢀ
erent silicon-to-molybdenum ratios
Table 2 Speciꢁc surꢀace area and pore structure oꢀ samples at diferent ratios oꢀ silicon to molybdenum
cn (Si: Mo)
aMo (wt%)
BET surꢀace area
Pore volume
bPore sizes (nm)
2
3
(
m /g)
(cm /g)
8.0
6.5
5.0
4.5
3.5
3.12
4.30
6.49
7.61
12.30
417
433
302
257
174
2.31
2.18
1.52
1.30
0.82
9.60
12.70
9.74
12.65
9.77
a
Determined by ICP
b
c
Catalyst pore sizes were determined ꢀrom the desorption branches oꢀ the isotherms
The mole ratio oꢀ Si/Mo in the synthesis solution
Investigation of liquid silylation process of MoO –salen@MCF
2
It can be seen ꢀrom Table 3 that the catalytic perꢀormance oꢀ MoO –salen@
2
MCF-S prepared using HMDS as the silylation reagent was improved compared
with that oꢀ the unsilylated catalyst. Indeed, the conversion oꢀ TBHP was 100%,
the selectivity ꢀor PO was 94.7%, and the selectivity ꢀor TBA was 84.6% with this
catalyst. Through silylation, hydrophilic silicone hydroxyl groups on the MCF
support were converted into hydrophobic alkoxy groups. The hydrophobic car-
rier environment ꢀacilitates protection oꢀ the active component oꢀ the catalyst and
rapid desorption oꢀ the polar product, which improves the stability and reaction
selectivity oꢀ the catalyst [32, 33].
1
3

Catalytic epoxidation of propylene over a Schiꢀ‑base…
7
6
5
00
00
00
1
00
X(TBHP)
S (PO)
S (TBA)
9
8
7
6
5
0
0
0
0
0
400
3
2
1
00
00
00
TOF
2
3
4
5
6
7
8
9
n(Si : Mo)
Fig. 7 Catalytic propylene epoxidation perꢀormance oꢀ samples at diferent silicon-to-molybdenum
ratios. Reaction conditions: TBHP (in C H Cl ) 0.0385 mol, C H Cl 12.51 g, propylene 0.385 mol,
2
4
2
2
4
2
m (catalyst)=1.69 wt%, temperature 120 °C, time 1 h. Turnover ꢀrequency (TOF) is calculated by the
−1
expression [n
]/[n ]×h
Mo
TBHP
Table 3 Efect oꢀ silanization on
propylene epoxidation catalyzed
by samples
Silanization Conversion/%
reagents
Selectivity Yield/%
(PO)/%
Selec-
tivity
(
TBA)/%
None
HMDS
99.7
100
84.8
94.7
84.5
94.7
81.3
84.6
Reaction conditions: TBHP (in C H Cl ) 0.0385 mol, C H Cl
2
4
2
2
4
2
1
2.51 g, propylene 0.385 mol, m (catalyst)=1.69 wt%, temperature
20 °C, time 1 h
1
Study on the recovery performance of MoO –salen@MCF‑S
2
From Fig. 8, it can be seen that the conversion oꢀ TBHP remained above 90%
aꢀter MoO –salen@MCF-S had been recovered and reused three times, but the
2
selectivity ꢀor PO dropped below 60%.
N physical adsorption and desorption curves ꢀor ꢀresh and recycled catalysts
2
are shown in Fig. 9, and speciꢁc surꢀace area and pore structure data are listed
in Table 4. From Fig. 9, it is evident that the MoO –salen@MCF-S sample aꢀter
2
graꢀting silanization with HMDS retained a type IV hysteresis loop aꢀter three
recoveries and that the original pore structure was not destroyed. From Table 4,
it can be seen that the speciꢁc surꢀace area increased and the pore volume var-
ied aꢀter each recovery and reuse, indicating a loss oꢀ active component during
the reaction. According to the ICP results in Table 4, with increasing number oꢀ
1
3

D. Chen et al.
X(TBHP)
S(PO)
S(TBA)
1
00
8
6
4
2
0
0
0
0
0
0
1
2
3
Number of Recovery
Fig. 8 Catalytic perꢀormance oꢀ propylene epoxidation aꢀter MoO2–salen@MCF-S recovery. Reac-
tion conditions: TBHP (in C H Cl ) 0.0385 mol, C H Cl 12.51 g, propylene 0.385 mol, m (cata-
2
4
2
2
4
2
lyst)=1.69 wt%, temperature 120 °C, time 1 h
60
MoO2-salen@MCF-S
R(1)
R(2)
R(3)
50
0.12
40
0.08
30
0.04
20
0.00
0
10
20
30
40
50
60
10
Pore Diameter(nm)
0
0.0
0.2
0.4
0.6
0.8
1.0
P/P0
Fig. 9 N physical adsorption–desorption isotherm curve and pore size distribution oꢀ MoO –salen@
2
2
MCF-S
recovery and reuse cycles, the Mo content gradually decreased, indicating that
the Mo active component was partly lost during the reaction.
In Table 5, the application oꢀ several supported Schif-base molybdenum com-
plexes in epoxidation reaction oꢀ small oleꢁns, such as 1-octene [21, 24] and propyl-
ene [34], was summarized and compared with the MCF molecular sieve supported
Schif-base molybdenum complexes MoO –salen@MCF prepared in this work.
2
1
3

Catalytic epoxidation of propylene over a Schiꢀ‑base…
Table 4 Speciꢁc surꢀace area
aMo (wt%) BET
bPore sizes (nm)
Catalyst
Pore
and pore structure oꢀ MoO –
2
surꢀace area volume
salen@MCF-S
2
3
(
m /g)
(cm /g)
MoO –
7.56
228
1.22
9.76
2
salen@
MCF-S
R (1)
c
7.48
6.68
5.32
305
315
341
1.51
1.55
1.47
9.73
9.77
9.76
R (2)
R (3)
a
Determined by ICP
b
Catalyst pore sizes were determined ꢀrom the desorption branches
oꢀ the isotherms
c
“
R” recovery, “1,2,3” numbers oꢀ recovery
Table 5 The comparisons oꢀ the catalytic results with literature data
Catalyst
Oleꢁn
Mole ratio
oꢀ oleꢁn to
oxidant
Tem./°C Time/h Epoxide TOF/h−1 Reꢀerences
yield/%
UiO-66-NH -SA-Mo 1-Octene 1:2
Reꢃux
Reꢃux
8
8
78
53
5
101
[21]
[24]
2
MoO (sal-phz)/
1-Octene 10:1
2
SBA-15
PBI·Mo
PGMA·AMP·Mo
Propylene 23:1
Propylene 23:1
Propylene 10:1
80
80
120
1
1
1
59
84
95
49
70
161
[34]
[34]
This work
MoO –salen@MCF
2
TBHP was used as oxidant; 1,2-dichloroethane was used as solvent
Although the evaluation conditions oꢀ the catalytic reaction were not exactly the
same, the catalysts prepared in this work had better reaction eꢂciency with higher
epoxide yield and higher TOF value.
Speculation on the mechanism of propylene epoxidation catalyzed by MoO2–
salen@MCF
The mechanism oꢀ propylene epoxidation under MoO –salen@MCF is shown
2
in Fig. 10. When (CH ) COOH difuses to the active components oꢀ the cata-
3
3
lyst in the reaction system, it encounters the Mo=O bond. Afected by its polar-
ity, (CH ) COOH is ꢀurther polarized, and the –OH attacks the π bond in the
3
3
Mo=O to turn it into Mo–O–O–C(CH ) . The three-dimensional structure oꢀ the
3
3
molybdenum complex is destroyed, the steric tension and the molecular energy
increase. Through hydrogen bonding, –OC(CH ) is eliminated to ꢀorm a three-
3
3
membered cyclic peroxygen active intermediate, and the π bond in the propyl-
ene double bond is attracted by the oxygen atom to approach the active complex,
1
3

D. Chen et al.
O
Mo
O
O
OH
O
OH
O
O
O O
Mo O*
H
Mo
O
O*
H
O
O
O
Mo O*
O
H
Fig. 10 The mechanism oꢀ propylene epoxidation under MoO –salen@MCF
2
ꢀ
orming propylene oxide with it. At the same time, the dropped –OC(CH ) bonds
3 3
ꢀ
orm tert-butanol with hydrogen, and the complex restores the Mo=O bond.
Conclusions
MoO –salen complex can be supported on the carrier oꢀ MCF by the method oꢀ
2
in situ hydrothermal synthesis. The obtained heterogeneous catalyst was evalu-
ated in the epoxidation oꢀ propylene with 1,2-dichloroethane solution oꢀ TBHP as
oxygen source. Aꢀter immobilization, the catalytic perꢀormance oꢀ MoO –salen@
2
MCF is better than that oꢀ MoO –salen. In addition, hydrophobic modiꢁcation
2
oꢀ the MoO –salen@MCF catalyst by a liquid-phase silanization method can
2
ꢀ
urther improve its catalytic perꢀormance and stability. The modiꢁed catalyst
MoO –salen@MCF-S showed the best perꢀormance in the epoxidation oꢀ propyl-
2
ene with 1,2-dichloroethane as solvent, and a molar ratio oꢀ propylene to TBHP
oꢀ 10:1 ꢀor 1 h at 120 °C. Under these conditions, the conversion oꢀ TBHP was
up to 100% and the selectivities ꢀor PO and TBA were 94.7% and 84.6%, respec-
tively. When the catalyst was recovered and reused three times, the selectivity ꢀor
PO decreased due to some loss oꢀ Mo. In addition, the mechanism oꢀ propylene
epoxidation catalyzed by MoO –salen@MCF was reasonably speculated.
2
Acknowledgements This study was ꢁnancially supported by the National Natural Science Foundation oꢀ
China (No. 21776237).
1
3

Catalytic epoxidation of propylene over a Schiꢀ‑base…
References
1
2
3
.
.
.
T.A. Nijhuis, M. Makkee, J.A. Moulijn, B.M. Weckhuysen, Ind. Eng. Chem. Res. 45, 3447 (2006)
E. Kertallia, J.C. Schoutena, T.A. Nijhuis, Appl. Catal. A 524, 200 (2016)
Q. Hua, T. Cao, X.-K. Gu, J.-Q. Lu, Z.-Q. Jiang, X.-R. Pan, F. Luo, W.-X. Li, W.-X. Huang, Angew.
Chem. Int. Ed. 53, 4856 (2014)
4
5
6
7
8
9
.
.
.
.
.
.
G.-J. Jin, G.-Z. Lu, Y.-L. Guo, Y. Guo, J.-S. Wang, X.-H. Liu, Catal. Today 93, 173 (2004)
X.-Q. Lu, H.-H. Wu, J.-G. Jiang, M.-Y. He, P. Wu, J. Catal. 342, 173 (2016)
V. Russo, R. Tesser, E. Santacesaria, M.D. Serio, Ind. Eng. Chem. Res. 52, 1168 (2013)
F.G. Resketi, S. Salehi, B. Bahramian, M. Mirzaee, J. Phys. Org. Chem. 32, 3921 (2018)
M. Kooti, M. Aꢀshari, Catal. Lett. 142, 319 (2012)
T. Luts, R. Frank, W. Suprun, S. Fritzsche, E. Hey-Hawkins, H. Papp, J. Mol. Catal. A: Chem. 273,
2
50 (2007)
1
1
1
0. A.M. Abu-Dieꢀ, I.M.A. Mohamed, Beni-Sueꢀ Univ. J. Basic Appl. Sci. 4, 119 (2015)
1. C.R. Bhattacharjee, C. Datta, G. Das, P. Mondal, Phase Trans. 85, 956 (2012)
2. I. Anis, M. Aslam, Z. Noreen, N. Aꢀza, A. Hussain, M. Saꢀder, A.H. Chaudhry, Acad. Sci. 5, 21
(
2013)
1
1
1
1
3. C.J. Dhanaraj, J. Johnson, J. Joseph, R.S. Joseyphus, J. Coord. Chem. 66, 1416 (2013)
4. T. Luts, W. Suprun, D. Hoꢀmann, O. Klepel, H. Papp, J. Mol. Catal. A: Chem. 261, 16 (2007)
5. A. Rezaeiꢀard, I. Sheikhshoaie, N. Monadi, M. Alipour, Polyhedron 29, 2703 (2010)
6. M. Moghadam, V. Mirkhani, S. Tangestaninejad, I. Mohammadpoor-Baltork, M. Moshreꢀ Javadi,
Polyhedron 29, 648 (2010)
1
1
1
2
7. M. Bagherzadeh, M. Zare, J. Coord. Chem. 66, 2885 (2013)
8. M. Aghajani, N. Monadi, J. Iran. Chem. Soc. 14, 963 (2017)
9. Y. Yang, S.-J. Hao, Y. Zhang, Q.-B. Kan, Solid State Sci. 13, 1938 (2011)
0. R. Mbeleck, M.L. Mohammed, K. Ambroziak, D.C. Sherrington, B. Saha, Catal. Today 256, 287
(
2015)
2
1. R. Kardanpour, S. Tangestaninejad, V. Mirkhani, M. Moghadam, I. Mohammadpoor-Baltork, F.
Zadehahmadi, J. Solid State Chem. 226, 262 (2015)
2
2
2
2
2. Y. Long, B. Yuan, J.-T. Ma, Chin. J. Catal. 36, 348 (2015)
3. H. Sharghi, M. Aber, M.M. Doroodmand, Mol. Divers. 19, 77 (2015)
4. M. Bagherzadeh, M. Zare, T. Salemnoush, S. Özkar, S. Akbayrak, Appl. Catal. A 475, 55 (2014)
5. P. Schmidt-Winkel, W.W. Lukens Jr., P.-D. Yang, D.I. Margolese, J.S. Lettow, J.Y. Ying, G.D.
Stucky, Chem. Mater. 12, 686 (2000)
2
2
6. X.-H. Yan, P. Xue, Res. Chem. Intermed. 7, 4295 (2018)
7. Y.-J. Zhao, M.-S. Liu, R.-Y. Zhao, F.-S. Liu, X.-P. Ge, S.-T. Yu, Res. Chem. Intermed. 12, 7711
(
2018)
2
8. Y.-X. Miao, G.-Z. Lu, X.-H. Liu, Y.-L. Guo, Y.-Q. Wang, Y. Guo, Microporous Mesoporous Mater.
1
22, 55 (2009)
2
3
9. K. Kannan, R.V. Jasra, J. Porous Mater. 18, 409 (2011)
0. H. Martinez, M.F. Caceres, F. Martinez, E.A. Paezmozo, S. Valange, N.J. Castellanos, D. Molina, J.
Barrault, H. Arzoumanian, J. Mol. Catal. A: Chem. 423, 248 (2016)
3
3
3
3
1. G.C. Behera, K.M. Parida, Appl. Catal. A 464, 364 (2013)
2. D.P. Serrano, R. Sanz, P. Pizarro, I. Moreno, Appl. Catal. A Gen. 435, 32 (2012)
3. Y. Yang, J.-Q. Guan, P.-P. Qiu, Q.-B. Kan, Appl. Surꢀ. Sci. 256, 3346 (2010)
4. M.M. Miller, D.C. Sherrington, S. Simpson, J. Chem. Soc. 2091 (1994)
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published
maps and institutional aꢂliations.
1
3