A recyclable cobalt(iii)-ammonia complex catalyst for catalytic epoxidation of olefins with air as the oxidant

Article abstract of DOI:10.1039/d0nj05466f

[Co(NH₃)₆]Cl₃and other ammonia complexes with different external anions or metal ions were synthesized to catalyze the epoxidation of α-pinene. The synthesized complexes were characterized using XRD, SEM, TGA, FTIR and UV spectra. With air as the oxidant, [Co(NH₃)₆]Cl₃exhibited excellent catalytic activity for the epoxidation of α-pinene among the prepared complexes. The conversion of α-pinene reached 97.4%, with 98.3% selectivity of epoxide when using a small amount of cumene hydroperoxide (CHP) as the initiator. The results revealed that a single Co(iii) system can also catalyze the epoxidation process in the absence of Co(ii), even showing better catalytic performance than single Co(ii). Recycling experiments showed that there was no significant drop in activity after 10 cycles, demonstrating that it is a stable and efficient heterogeneous catalyst for the epoxidation of α-pinene. The excellent recycling performance may be attributed to the stability of the coordination complex itself.

Full text of DOI:10.1039/d0nj05466f

NJC
View Article Online
View Journal | View Issue
PAPER
A recyclable cobalt(III)–ammonia complex catalyst
for catalytic epoxidation of olefins with air as the
oxidant†
Cite this: New J. Chem., 2021,
45, 2147
Chenlong Wang,a Hongju Zhan,b Xinhuan Lu,*a Run Jing,a Haifu Zhang,a Lu Yang,a
a
a
Xixi Li,a Fanfan Yue,a Dan Zhou
and Qinghua Xia
*
[Co(NH3)6]Cl3 and other ammonia complexes with different external anions or metal ions were
synthesized to catalyze the epoxidation of a-pinene. The synthesized complexes were characterized
using XRD, SEM, TGA, FTIR and UV spectra. With air as the oxidant, [Co(NH3)6]Cl3 exhibited excellent
catalytic activity for the epoxidation of a-pinene among the prepared complexes. The conversion of
a-pinene reached 97.4%, with 98.3% selectivity of epoxide when using a small amount of cumene
hydroperoxide (CHP) as the initiator. The results revealed that a single Co(III) system can also catalyze
the epoxidation process in the absence of Co(II), even showing better catalytic performance than single
Co(II). Recycling experiments showed that there was no significant drop in activity after 10 cycles,
demonstrating that it is a stable and efficient heterogeneous catalyst for the epoxidation of a-pinene.
The excellent recycling performance may be attributed to the stability of the coordination complex
itself.
Received 8th November 2020,
Accepted 31st December 2020
DOI: 10.1039/d0nj05466f
rsc.li/njc
(TBHP) are expensive and dangerous.12 In contrast, abundant and
readily available molecular oxygen, which is also environmentally
1. Introduction
Epoxy compounds are usually produced by oxidation of friendly, is undoubtedly the most suitable oxidant. Previous
olefins,1 which are important synthetic intermediates2 in studies on olefin epoxidation with molecular oxygen as an
organic reactions, and can be used to prepare epoxy resins, oxidant have been carried out.13 Our group has recently also
coatings,3 advanced building blocks, etc. Besides, alcohols, made considerable progress in enhancing the reactivity of air
hydroxy ethers4,5 and some other chemicals which have wide with a-pinene by using microwaves.14
applications need epoxides as the starting material for their
From an economic and feasibility point of view, it is more
synthesis. For example, a-pinene oxide, which in nature is attractive to use cheap and readily available transition metals
formed by the epoxidization of a-pinene in some plants, is an instead of Ag,15 Au16 or other noble metals as the catalyst for
important raw material used in the production of drugs.6 Thus, epoxidation reactions.3,17,18 Among the transition metals,
developing efficient and highly selective catalysts to catalyze the Ti, Mo, W, V, and Re have been early and extensively
epoxidation of olefins is a research hotspot in the field of researched.19–25 Recently, a series of catalysts based on Fe,26
catalysis.7
Co,27–29 Ni,30 and Mn31 have also been studied and developed
Traditionally, the epoxidation of olefins in industrial produc- to catalyze the epoxidation reaction. Since the complexes of
tion requires the addition of stoichiometric organic peroxides8 or transition metal and Schiff base have been considered as one of
peracids9 as oxidizing agents. However, using peracids as the the most efficient catalysts for epoxidation reaction,32–34
oxidant will produce acidic toxic wastes,10,11 which are difficult to numerous studies on the complexes of different metals with
dispose, and organic peroxides such as tert-butyl hydroperoxide the Schiff base have been carried out.35,36 However, metal
complexes are applicable to homogeneous systems under nor-
mal conditions,37–39 resulting in difficulty in the separation of
catalysts from the reaction system. Comparatively, heterogeneous
catalytic systems can overcome this shortcoming, thus the trans-
form of homogeneous catalysts to heterogeneous ones has always
been a research direction for catalytic scholars. To this end,
various supported metal complexes40,41 have been developed to
catalyze the epoxidation process. The transition metal-containing
a Hubei Collaborative Innovation Center for Advanced Organic Chemical Materials,
Ministry-of-Education Key Laboratory for the Synthesis and Application of Organic
Functional Molecules, Ministry-of-Education Key Laboratory for Green Preparation
and Application of Functional Materials, Hubei University, Wuhan 430062, P. R.
China. E-mail: xinhuan003@aliyun.com, xiaqh518@aliyun.com
b Jingchu University of Technology, Jingmen 448000, P. R. China
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
d0nj05466f
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021
New J. Chem., 2021, 45, 2147--2156 | 2147

View Article Online
Paper
NJC
MOF materials can be also directly modified42 to be heteroge- The product was directly used as a catalyst after drying at 80 1C
neous catalysts. However, the uneven distribution and easy for 24 h.
leaching of active sites lead to reduction in activity or even
deactivation, which is not suitable for continuous industrial CoCl2ꢀ6H2O (50.0 g) were dissolved using concentrated NH3ꢀH2O
production.23,43
(160 mL), then 30% H2O2 (40 mL) was added slowly into the
2.2.2. Synthesis of [Co(NH3)5Cl]Cl2. NH4C1 (10.0 g) and
Mascharak et al. have reported that olefin epoxidation is above solution under stirring (foaming phenomenon was
carried out by radicals and Co(III), which is an intermediate that observed during this period). Purplish red precipitate was generated
promotes the reaction in the Co(II) catalytic system.44 Possibly, while adding HCl (16 mL) slowly. After cooling, the precipitate
Co(III) can facilitate to promote the formation of ROꢀ radicals was filtered and washed with diluted hydrochloric acid and
during the reaction. Herein, we synthesize Co(III)-based ethanol. Purplish red [Co(NH3)5Cl]Cl2 was obtained after drying
[Co(NH3)6]Cl3 material as a heterogeneous catalyst, which can at 80 1C for 24 h.
avoid the problem of the difficulty in separating of the catalyst
2.2.3. Synthesis of [Co(NH3)6](NO3)3. The synthesis process
from the product.45 When introducing air as the oxidant, the of [Co(NH3)6](NO3)3 was similar to that of [Co(NH3)6]Cl3, except
as-prepared catalyst performed excellent catalytic activity on the that CoCl2ꢀ6H2O and NH4Cl were replaced by equimolar
epoxidation of a-pinene, giving 97.4% conversion and 98.3% amounts of Co(NO3)2ꢀ6H2O and NH4NO3.
selectivity of epoxide. The recycling performance of the catalyst
2.2.4. Synthesis of [Co(NH3)6](Ac)3. The synthesis process
is also tested, and the results show that the activity is still well of [Co(NH3)6](Ac)3 was similar to that of [Co(NH3)6]Cl3, except
maintained after 10 recycles, with no significant decrease in that HAc (Acetic acid) instead of concentrated HCl was added
activity.
into the filtrate. [Co(NH3)6](Ac)3 was obtained after cooling and
filtering.
2.2.5. Synthesis of [Co(NH3)5Cl]SO4. Concentrated H2SO4
and synthesized [Co(NH3)5Cl]Cl2 were mixed in the ratio of 1 : 3.
Subsequently, the solution was diluted with hot water (70 1C)
until bubble (HCl) generation ceased. Then [Co(NH3)5Cl]SO4ꢀ
2H2O could be obtained by cooling and filtering.
2. Experimental section
2.1. Materials
The main reagents including CoCl2ꢀ6H2O (AR, Aladdin),
Co(NO3)2ꢀ6H2O (Z99%, Aladdin), Coconut shell charcoal (Hainan),
NH4Cl (Z99.5%, Sinopharm), NH3ꢀH2O (25–28%, Sinopharm),
HCl (hydrochloric acid, 36–38%, Sinopharm), THF (tetrahydro-
furan, Z99.5%, Sinopharm), Ni(NO3)2ꢀ6H2O (Z99%, Sinopharm),
CuSO4ꢀ5H2O (Z99%, Sinopharm), ZnCl2 (Z99%, Kelong), CHP
(Cumene hydroperoxide, Z72%, Sinopharm), TBHP (tert-butyl
hydroperoxide, Z65%, Sinopharm), H2O2 (Z30%, Sinopharm),
a-pinene epoxide (Z95%, TGI), DMF (N,N-dimethyl formamide,
Z98.5%, Sinopharm), DMA (N,N-dimethyl acetamide, Z98.5%,
Sinopharm), a-pinene (Z98%, Macklin), b-pinene (Z98%,
Macklin), 1-dodecene (Z95%, Macklin), cyclooctene (Z95%,
Macklin), a-methylstyrene (Z99%, Macklin) and styrene (Z99%,
Sinopharm) were directly used without further purification.
2.3. Characterization of the complex materials
X-ray diffraction (XRD) patterns of catalysts were collected
using a Bruker D8A25 diffractometer with Cu Ka as the radia-
tion (l = 1.54184 Å). The test conditions were a scanning range
from 2y = 5 to 801, a speed of 101 minꢁ1 and operated at 40 kV
and 40 mA. Scanning electron microscope (SEM) images was
taken using a JEOL JSM-6510 A to investigate the surface
morphology and size of the catalyst. The IR spectra of samples
in the range of 400–4000 cmꢁ1 were recorded using a OPUS
Fourier Transform Infrared Spectrophotometer. The UV-Vis
spectra of materials were detected using a Shimadzu UV-2550
spectrometer with a scanning range from 800 to 200 cmꢁ1
.
X-ray photoelectron spectra (XPS) of the sample were recorded
using a Thermo Fisher Scientific ESCALAB 250Xi spectrometer
2.2. Preparation of catalysts
2.2.1. Synthesis of [Co(NH3)6]Cl3. [Co(NH3)6]Cl3 was pre- under 0.5Ev of energy resolution. Additionally, thermogravi-
pared using the method reported by Bjerrum and McReynold.46 metric study (TGA) of the catalysts was conducted using a
NH4Cl (4.0 g) was dissolved in a beaker with deionized water METTLER-TOLEDO TGA2 instrument, with N2 at a flow rate
(8.4 g) under stirring, then CoCl2ꢀ6H2O (6.0 g) and coconut shell of 40 mL minꢁ1 as the shielding gas and testing from 30 to
charcoal (0.4 g) were mixed into the above solution. Sub- 1000 1C at a heating rate of 10 1C minꢁ1. The results of the NH3
sequently, NH3ꢀH2O (12.5 mL) was slowly added into the beaker temperature-programmed desorption (NH3-TPD) analysis of the
in an ice bath and then the beaker was sealed with a plastic material were recorded using an automatic gas sorption analyzer
wrap until the temperature of the solution dropped to 0 1C. The (TP-5076).
solution was heated and kept at 60 1C for 30 min after 5% H2O2
(13.5 mL) was injected to the solution. After cooling the
2.4. Epoxidation reactions of a-pinene
solution to 0 1C, black precipitate was obtained by filtration. Typically, the reaction process comprised the following steps:
Then, the precipitate was dissolved in 60 mL of boiled water 5 mg of catalyst, 10 g of DMF, 3 mmol of a-pinene and 0.3 mmol
containing concentrated HCl (1.7 mL) and filtered immediately. of CHP were added into a two-necked glass reactor. Then, the
Concentrated HCl (6.7 mL) was slowly added to the above filtrate reactor was placed in a water bath at 90 1C with a cryogenic-
and the orange-yellow solid could be precipitated during cooling. liquid condenser. The flow rate of air was 40 mL minꢁ1. After
Then, [Co(NH3)6Cl3] was collected by vacuum suction filtration. reacting for 5 h under stirring, the product was separated by
2148 | New J. Chem., 2021, 45, 2147--2156
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021

View Article Online
NJC
Paper
centrifugation. Subsequently, the filtrate was collected and
analyzed using a gas chromatography (GC) instrument that
was equipped with a capillary column (Rtx-5, 30 m ꢂ 0.25 mm ꢂ
0.25 mm) and an FID detector, in which chlorobenzene was used
as the internal standard to analyze different components.
3. Result and discussion
3.1. Structural descriptions
The XRD patterns of the synthesized [Co(NH3)6]Cl3 are shown
in Fig. 1. The diffraction peaks of the material were exactly
consistent with the simulated peaks, showing obvious crystal
peaks at 2y = 15.01,15.51, 16.81, 17.81, 24.31, 24.91, 25.61, 34.01
and 36.21, which indicated that [Co(NH3)6]Cl3 had been success-
fully synthesized. Fig. S1 (ESI†) compares the differences in the
XRD patterns between [Co(NH3)6]Cl3 and other synthesized
cobalt compounds ([Co(NH3)5Cl]Cl2 and [Co(NH3)5Cl]SO4), in
which the different characteristic diffraction peaks indicated
that these synthesized catalysts exhibited an entirely different
crystal structure. Moreover, it could be observed that the inten-
sity of the diffraction peaks of [Co(NH3)6]Cl3 was obviously
higher than that of the [Co(NH3)5]3+ complexes, which might
endow it with considerable catalytic performance.
Fig. 2 The IR spectra of [Co(NH3)6]Cl3.
those reported in the literature.47,48 The IR spectra of different
cobalt–ammonia complexes are shown in Fig. S2 (ESI†).
[Co(NH3)5Cl]Cl2, [Co(NH3)6](Ac)3 and [Co(NH3)6]Cl3 exhibited
similar spectra, with characteristic absorption peaks appearing
at about 820, 1320, 1580 and 3120 cmꢁ1, consistent with those
reported in the literature.46 However, the IR spectra of
[Co(NH3)5Cl]SO4 and [Co(NH3)6](NO3)3 were obviously different
Fig. 2 shows the infrared spectra of the [Co(NH3)6]Cl3
material. It could be seen from the figure that there are four
main characteristic absorption peaks, including (1) the strong
absorption peak at 820 cmꢁ1 belonging to the Co–N rocking
vibration, (2) the sharp absorption peak at 1320 cmꢁ1, which is
assigned to the symmetrical bending vibration of the N–H
bond, (3) the weak broad peak at 1580 cmꢁ1, which is attributed
to N–H unsymmetrical bending vibration, and (4) the absorption
peak at 3120 cmꢁ1 corresponding to the symmetrical stretching
vibration of the N–H bond. All of them are in agreement with
from those of others. The absorption peaks at 610 and 1090 cmꢁ1 in
2ꢁ
the spectra of [Co(NH3)5Cl]SO4 were typical absorption of SO4
.
The absorption peaks at 1320 cmꢁ1 of [Co(NH3)6](NO3)3 seemed
wider and stronger than those of other complexes, which could
be ascribed to tꢁhe overlap interference between stretching
vibration of NO3 and symmetrical bending vibration of the
N–H bond.
Fig. 3 shows the results of SEM micro-morphology analysis
of different cobalt–ammonia complexes. From the figure, it
could be seen that all the complexes exhibited micron sized
bulk morphology with an uneven grain size. [Co(NH3)6]Cl3
(a and a0) possessed a relatively smooth surface and displayed
regular columnar morphology. Another complex with regular
morphology was [Co(NH3)5Cl]Cl2 (d and d0). However, the
morphology of [Co(NH3)6](Ac)3 (b and b0), [Co(NH3)6](NO3)3
(c and c0), and [Co(NH3)5Cl]SO4 (e and e0) was amorphous and
the three samples exhibited uneven sizes from a few microns to
tens. Fig. S3 (ESI†) compares the UV-Vis spectra of the synthe-
sized [Co(NH3)5Cl]Cl2, [Co(NH3)6](Ac)3, [Co(NH3)6]Cl3 and
[Co(NH3)6](NO3)3. According to the literature,49 the characteristic
peaks at 340 and 470 nm were typical absorption peaks of a low
spin octahedral coordination compound (d6[Co(NH3)6]3+), which
1
1
belong to A1g - T1g and A1g
-
1T2g, respectively. But the
absorption peaks of [Co(NH3)5Cl]Cl2 were at 370 and 550 nm,
ascribed to the different coordination environment of Co(III)
Fig. 1 The XRD patterns of [Co(NH3)6]Cl3 synthesized and the simulated one.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021
New J. Chem., 2021, 45, 2147--2156 | 2149

View Article Online
Paper
NJC
Fig. 3 SEM spectra of (a and a0) [Co(NH3)6]Cl3, (b and b0) [Co(NH3)6](Ac)3, (c and c0) [Co(NH3)6](NO3)3, (d and d0) [Co(NH3)5Cl]Cl2 and (e and e0)
[Co(NH3)5Cl]SO4.
between [Co(NH3)5Cl]2+ and [Co(NH3)6]3+. The differences men-
tioned above were also the reason for the difference in color ion, XPS analysis was conducted as shown in Fig. 5. It could be
between [Co(NH3)6]3+ and [Co(NH3)5Cl]2+ complexes.
seen that, within the Co2p3/2 energy level, there was only one
To investigate the surface constituents and valence of cobalt
The result of thermal stability analysis of [Co(NH3)6]Cl3 is peak appearing at 781.6 eV, attributed to characteristic peak of
shown in Fig. 4. From the TG curve, it could be seen there was Co(III).52 This result indicated that cobalt species in the synthesized
no weight loss before 235 1C, showing that the structure of [Co(NH3)6]Cl3 all exhibited the Co(III) coordinative state. In combi-
[Co(NH3)6]Cl3 was stable at this stage. However, [Co(NH3)6]Cl3 nation with the epoxidation reaction results, it could be known that
exhibited two obvious weight loss stages after 230 1C. Based on in the absence of Co(II), the single Co(III) system still showed
the literature,50,51 the first stage that occurred at 230–320 1C excellent catalytic activity.
was ascribed to the decomposition from [Co(NH3)6]Cl3 to
The acidity of the complexes was investigated using the NH3-
(NH4)2CoCl4 and CoCl2, with the release of NH3, N2 and HCl TPD method (Fig. 6). The NH3-TPD profiles of [Co(NH3)6]Cl3
(the weight loss rate was about 52%). The other weight loss mainly showed two peaks at 228 and 291 1C, which corresponded
stage from 320 to 700 1C was about 22%, at which all released to the weak or medium strong acid site of the material. By
species were NH3 and HCl. The good stability of [Co(NH3)6]Cl3 comparing the peak position and peak area, it was found that the
resulted in its considerable recycling performance.
acidity of [Co(NH3)5Cl]Cl2 and [Co(NH3)6]Ac3 was similar to that
2150 | New J. Chem., 2021, 45, 2147--2156
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021

View Article Online
NJC
Paper
Fig. 6 NH3-TPD analysis results of cobalt–ammonia complexes.
Fig. 4 TG curves of [Co(NH3)6]Cl3.
Fig. 7 The effect of different cobalt–ammonia complexes on the epoxidation
of a-pinene. Conditions: catalyst, 5 mg; a-pinene, 3 mmol; solvent, DMF 10 g;
initiator, CHP 0.3 mmol; temperature, 90 1C; flow rate of air, 40 mL minꢁ1; and
time, 5 h.
Fig. 5 XPS spectra of [Co(NH3)6]Cl3.
98.3% selectivity of epoxide. When using [Co(NH3)6](Ac)3 as the
catalyst, the conversion was close to the result of using
[Co(NH3)6]Cl3 (96.2%), while its low selectivity led to the low yield
of epoxide. [Co(NH3)6](NO3)3 only converted 10.8% of a-pinene.
When using [Co(NH3)5Cl]2+ class complexes with different anions,
the conversion of a-pinene exhibited a significant drop from
of [Co(NH3)6]Cl3. However, [Co(NH3)5Cl]SO4 possessed a stronger
acidity than [Co(NH3)6]Cl3, which showed one peak band at
433 1C (strong acid site), indicating the enhancement of acidity
by introducing SO42ꢁ. The complexes synthesized by introducing
different acid radical ions showed different reactivity in the
subsequent epoxidation reaction.
2ꢁ
93.5% to 37.1% after Cl2 was replaced by SO42ꢁ. It could be
seen that different anions had a great effect on the activity of the
cobalt–ammonia complexes, which resulted in a significant dif-
3.2. Catalytic epoxidation of a-pinene
The results of different cobalt–ammonia complexes in catalyzing ference in a-pinene conversion and epoxide selectivity. Among the
the epoxidation process of a-pinene are shown in Fig. 7. It could above catalysts, [Co(NH3)6]Cl3 gave apparently higher conversion
be seen that the different complexes obviously exhibited different of a-pinene and selectivity of target product than other cobalt–
catalytic activity. The [Co(NH3)6]Cl3 catalyst showed the best ammonia complexes. The NH3-TPD analysis shown in Fig. 6 also
catalytic performance with 97.4% conversion of a-pinene and supported the above conclusion. Fig. 7 also shows the TON value
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021
New J. Chem., 2021, 45, 2147--2156 | 2151

View Article Online
Paper
NJC
of various complexes under standard conditions. Among catalysts
of different cobalt–ammonia complexes, [Co(NH3)6]Cl3,
[Co(NH3)6](Ac)3 and [Co(NH3)5Cl]Cl2 exhibited a fairly high TON
value of 4140 in 5 h. Other cobalt–ammonia complexes exhibited
a low TON value ([Co(NH3)6](NO3)3, 22.5 and [Co(NH3)5Cl]SO4,
82.8). Moreover, Table S1 (ESI†) compares the catalytic perfor-
mance between the as-prepared [Co(NH3)6]Cl3 and other catalysts
reported; it could be seen that the [Co(NH3)6]Cl3 catalyst in this
study gave an excellent conversion for the epoxidation of a-pinene
under mild reaction conditions. As shown in Fig. S4 (ESI†), when
[Co(NH3)6]Cl3 was used as the catalyst, the conversion of a-pinene
reached 97.4%, with 98.3% selectivity of epoxide. The complexes
of [Ni(NH3)6]Cl2, [Cu(NH3)6]SO4ꢀH2O and (NH4)3Cl[ZnCl4] had poor
catalytic activity, and a-pinene conversion could only reach 27.9%,
19.5% and 6.4%. At the same time, the related TON value was also
in the descending order of [Co(NH3)6]Cl3 (156) 4 [Ni(NH3)6]Cl2
(44.8) 4 [Cu(NH3)6]SO4ꢀH2O (31.3) 4 (NH4)3Cl[ZnCl4] (10.3).
As shown in Fig. 8, the effect of different reaction tempera-
ture on the epoxidation of a-pinene was explored by gradually
increasing the temperature from 60 to 95 1C. It could be easily
found that the reaction temperature played an important role
in the epoxidation process. The conversion was only 9.1% at
60 1C but reached 92.3% at 85 1C. The yield of epoxide reached
a maximum when the temperature was further increased to
90 1C, with 97.4% conversion of a-pinene and 98.3% selectivity
of epoxide. However, the selectivity of epoxide declined to
86.9% as the reaction temperature was continually increased
to 95 1C, with no significant improvement of conversion. This
may be ascribed to the transformation from a-pinene and
epoxide to by-products while increasing the temperature,
further leading to a decrease in the selectivity of the target
product.
Fig. 9 The effect of reaction time on the epoxidation of a-pinene over
[Co(NH3)6]Cl3. Conditions: catalyst, 5 mg; a-pinene, 3 mmol; solvent, DMF
10 g; initiator, CHP 0.3 mmol; temperature, 90 1C; and flow rate of air,
40 mL minꢁ1
.
was improved apparently at the stage of increasing the reaction
time from 1 to 5 h (from 15.7% to 97.4%). However, after
extending the reaction time to 6 h, the epoxide was further
converted into by-products, resulting in a significant decrease
in the selectivity of epoxide. Moreover, it could be seen that the
conversion of a-pinene was elevated slowly at the early stage of
the reaction but increased rapidly with time being extended.
This may be attributed to the effect of the olefin cation radical
amount, which was generated during the epoxidation
process.53 Less-generated radicals at the early reaction stage
resulted in a slower increase rate of conversion.
As shown in Fig. 10, various initiators have been tested to
promote the epoxidation of a-pinene over [Co(NH3)6]Cl3. When
As depicted in Fig. 9, the influence of reaction time on the
results of epoxidation was investigated with the other reaction
conditions remaining unchanged. The conversion of a-pinene
Fig. 10 The effect of different initiators on the epoxidation of a-pinene
over [Co(NH3)6]Cl3. Conditions: catalyst, 5 mg; a-pinene, 3 mmol; solvent, DMF
10 g; initiator, 0.3 mmol; temperature, 90 1C; flow rate of air, 40 mL minꢁ1; and
time, 5 h.
Fig. 8 The effect of reaction temperature on the epoxidation of a-pinene
over [Co(NH3)6]Cl3. Conditions: catalyst, 5 mg; a-pinene, 3 mmol; solvent,
DMF 10 g; initiator, CHP 0.3 mmol; flow rate of air, 40 mL minꢁ1; and time, 5 h.
2152 | New J. Chem., 2021, 45, 2147--2156
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021

View Article Online
NJC
Paper
only air was used as the oxidant without any initiators, no
acceptable conversion or epoxide yield was obtained (only
18.2% conversion of a-pinene and 61.2% selectivity of epoxide).
When H2O2 was used as the initiator, the conversion of
substrate and the selectivity of target product reached 28.6%
and 87.4%, respectively. It could be observed that the selectivity
of epoxide was increased to 94.2% when H2O2 was replaced by
a-epoxy pinane, whereas the conversion was still low (34.9%).
When CHP or TBHP was chosen as the single initiator, the
conversion was greatly increased to 97.4% and 96.5%, respec-
tively. Moreover, the selectivity of epoxide was improved to
98.3% by using CHP, higher than that by using TBHP (93.4%).
From the above results, it can be seen that CHP showed the best
promoting performance for the epoxidation of a-pinene.
The effect of the amount of CHP on the reaction results is
demonstrated in Fig. S5 (ESI†). It can be easily seen that the
conversion of a-pinene and the selectivity of epoxide were both
low in the absence of initiator CHP, but greatly improved (from
18.4% to 84.6% in conversion and from 60.2% to 72.1% in
selectivity, respectively) by adding a small amount of CHP
(0.06 mmol). When the amount of CHP was increased from
0.06 to 0.3 mmol gradually, both conversion and selectivity
were elevated to 97.4% and 98.3%, respectively. However, the
variation trend showed some differences when further increasing
the amount of initiator to 0.4 mmol, which manifested as a
continuous increase in conversion but decrease in selectivity of
epoxide. From the above result, it could be seen that an appro-
priate amount of CHP (0.3 mmol) facilitated to maximize the yield
of epoxide.
Fig. 11 The effect of solvents on the epoxidation of a-pinene over
[Co(NH3)6]Cl3. Conditions: catalyst, 5 mg; a-pinene, 3 mmol; solvent, 10 g;
initiator, CHP 0.3 mmol; temperature, 90 1C; flow rate of air, 40 mL minꢁ1
;
and time, 5 h.
further promoted the epoxidation reaction of a-pinene with
air.28,36,54,55 Co(III) in the reaction system also played an important
role in transferring oxygen.44
The schematic illustration for the epoxidation of a-pinene is
shown in Fig. S7 (ESI†). It could be observed that there was one
main product a-pinene oxide (a) and several by-products (b–d).
It is worth noting that camphorenal aldehyde (b) was the
product of further isomerization of a-pinene oxide. Table 1
lists the results of a-pinene epoxidation over [Co(NH3)6]Cl3
catalyst with different amounts of substrate. As shown in Entry
1, when the amount of a-pinene was 3 mmol, 97.4% a-pinene
conversion and 98.3% epoxide selectivity were obtained. As the
substrate amount increased from 3 to 6 mmol, the conversion
and selectivity of epoxide still maintained around 80% and
490%, respectively. However, as the amount of substrate
increased to 10 mmol, the conversion decreased rapidly to
42.3%. The reason for the drop may be the lack of active sites.
The conversion reached 98.0% when 6 mmol of substrate was
continuously reacted for 8 h. On the other hand, it can be seen
As shown in Fig. S6 (ESI†), the influence of different flow
rates of air on the epoxidation of a-pinene over [Co(NH3)6]Cl3
was investigated. It could be seen that only 14.7% conversion of
a-pinene was obtained with no air being introduced. Then the
conversion and selectivity of epoxide were gradually elevated
when the flow rate of air was increased from 20 to 40 mL minꢁ1
.
However, a significant drop in the selectivity of epoxide was
observed when the flow rate of air was further increased to 50
mL minꢁ1, accompanied by a slight increase in conversion.
This may be attributed to the decomposition of epoxide, caused
by excessive oxygen. Therefore, the yield of target product
reached a maximum when the flow rate of air was 40 mL minꢁ1
.
Fig. 11 lists the influence of different solvents on the
epoxidation of a-pinene over [Co(NH3)6]Cl3. Obviously, the
reaction solvent significantly affected the epoxidation process.
Due to the weak ability of dissolving oxygen, non-polar toluene
gave only 7.3% conversion of a-pinene. THF seemed to be the
most unsuitable one and achieved o5% conversion of a-
pinene. When the polar solvents DMA and 1,4-dioxane were
used, though the conversion of a-pinene was higher than when
using non-polar solvents, as described above, the conversions
were only 21.5% and 19.0%. The conversion reached 97.4%
when strong polar DMF was selected as the reaction solvent.
Owing to the strong coordination with cobalt species and O2,
DMF played an important role in the epoxidation of a-pinene
with air. Previous studies had found that Co(II) could effectively
Table 1 The effect of substrate amount on the epoxidation of a-pinene
Selectivity (%)
Substrate amount
Entry (mmol)
Time Conversion
(h)
(%)
a
b
c/d
Others
1
2
3
4
5
6
7
8
3
4
5
6
6
6
6
10
5
5
5
5
6
7
8
5
97.4
93.0
91.5
80.0
90.0
95.8
98.0
42.3
98.3 0.9 0.2/0.5 0.1
93.9 2.0 1.4/2.1 0.6
92.1 2.3 1.8/3.0 0.8
92.8 2.4 1.2/2.9 0.7
88.1 4.6 2.3/4.3 0.7
87.6 5.9 1.9/3.4 1.2
86.5 8.7 1.1/2.4 1.3
92.5 4.3 1.2/1.5 0.5
Conditions: catalyst, [Co(NH3)6]Cl3 5 mg; substrate, a-pinene; solvent,
DMF 10 g; initiator, CHP 0.3 mmol; temperature, 90 1C; flow rate of air,
activate oxygen and form active intermediates with DMF, which 40 mL minꢁ1; and time, 5 h.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021
New J. Chem., 2021, 45, 2147--2156 | 2153

View Article Online
Paper
NJC
Fig. 13 The catalytic epoxidation of other bulky cycloalkenes. Conditions:
catalyst, [Co(NH3)6]Cl3 5 mg; substrate, 3 mmol; solvent, DMF 10 g;
Fig. 12 Results of recyclable performance of the [Co(NH3)6]Cl3 catalyst.
Conditions: catalyst, [Co(NH3)6]Cl3; a-pinene, 3 mmol; solvent, DMF 10 g;
initiator, CHP 0.3 mmol; temperature, 90 1C; flow rate of air, 40 mL minꢁ1
;
initiator, CHP 0.3 mmol; temperature, 90 1C; flow rate of air, 40 mL minꢁ1
;
and time, 5 h.
and time, 5 h.
The conversion of epoxidation reaction of several cycloalkenes
under the above-stated conditions is listed in Fig. 13. It can be
seen that [Co(NH3)6]Cl3 performed excellent catalytic properties on
the epoxidation of some cyclic olefins. Comparatively, the conver-
sion of alkenes declined in the order of a-pinene (97.4%) 4
from entries 4 7 that the selectivity of a-pinene oxide gradually
declined when the reaction time was extended. Moreover, the
distribution of products reflected that the proportion of cam-
phorenal aldehyde showed an obvious upward trend, which
could be attributed to the further isomerization of a-pinene
oxide as the reaction time extended.
a-methylstyrene (94.6%) 4 styrene (85.5%)
4 cyclooctene
(44.6%) 4 b-pinene (43.4%) 4 3-cyclohexene-1-carboxaldehyde
(11.6%) 4 limonene (8.6%). However, the epoxide selectivity
declined in the order of limonene (100%) = 3-cyclohexene-1-
The recycling experiments of the [Co(NH3)6]Cl3 catalyst on the
epoxidation of a-pinene were conducted. As shown in Fig. 12, the
activity of [Co(NH3)6]Cl3 (ten times) showed a slight decline during
the initial three recycles, but was basically unchanged in the
subsequent seven recycles. Negligible leaching of cobalt species from
[Co(NH3)6]Cl3 was detected by analyzing the ICP results of fresh and
recycled catalysts. The XRD patterns and IR spectra of fresh and 10th
recycled catalysts are compared in Fig. S8 and S9 (ESI†). It could be
observed that the recycled catalyst had almost the same peaks as
that of the fresh one in the XRD patterns or IR spectra. The favorable
stability and recycling performance may be ascribed to the stable
coordination environment of the Co species. Moreover, we also
conducted the recycling performance of [Co(NH3)5Cl]Cl2. As could
be seen from Fig. S10, though the catalytic activity of [Co(NH3)5Cl]Cl2
was lower than [Co(NH3)6]Cl3, its conversion reached 86.4% after
being recycled 5 times. From Fig. S11 (ESI†) it could be seen that the
XRD patterns of recycled [Co(NH3)5Cl]Cl2 was exactly consistent with
that of the fresh one, with only a little drop in the intensity of peaks.
The reaction mechanism of the epoxidation reaction catalyzed by
the [Co(NH3)6]Cl3 catalyst is shown in Fig. S12 (ESI†). Firstly, the
active intermediates formed by DMF and O2 were adsorbed by
[Co(NH3)6]Cl3 (Step I). Then, molecular oxygen was activated by
Co(III) to form active peroxy intermediate species (Step II). Subse-
quently, the a-pinene molecules interacted with peroxy intermediate
species and formed cyclic intermediate species (Step III), which
released epoxide molecules at last (Step IV), and then the process
restarted with Step I.
carboxaldehyde (100%)
= cyclooctene (100%) 4 a-pinene
(98.3%) 4 a-methylstyrene (81%) 4 styrene (74.7%) 4 b-pinene
(55.1%). The epoxidation of linear terminal 1-heptene and
1-dodecene with air was negligible. The possible explanation for
apparently different results on epoxidizing cyclic olefin and linear
olefin was that electrophilic cycloaddition activity was weakened
by the low p-electron density of linear olefin.56
4. Conclusions
A series of cobalt–ammonia complexes with different external
anions (Acꢁ, NO3ꢁ, SO42ꢁ, Clꢁ) were prepared using a simple
method and applied to catalyze the epoxidation of a-pinene to
a-epoxy pinane. The experimental results showed that
[Co(NH3)6]Cl3 exhibited the best catalytic activity on the epoxida-
tion of a-pinene. The cobalt–ammonia complex used, the catalyst
amount, the reaction solvents, the reaction temperature, the
reaction time and initiators exerted notable impacts on the
epoxidation reaction. The considerable reaction results demon-
strated that single Co(III) even showed better catalytic effect on
epoxidation than a single Co(II) system. The recycling experi-
ments indicated that [Co(NH3)6]Cl3 was a stable and recyclable
heterogeneous catalyst, which showed no significant deactiva-
tion after 10 recycles.
2154 | New J. Chem., 2021, 45, 2147--2156
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021

View Article Online
NJC
Paper
21 G. D. Faveri, G. Ilyashenko and M. Watkinson, Chem. Soc.
Rev., 2011, 40, 1722–1760.
22 X. Wang, S. Wu, Z. Li, X. Yang, H. Su, J. Hu, Q. Huo, J. Guan
and Q. Kan, Microporous Mesoporous Mater., 2016, 221,
58–66.
Conflicts of interest
There are no conflicts of interest to declare.
Acknowledgements
23 A. R. Yaul, G. B. Pethe and A. S. Aswar, Russ. J. Coord. Chem.,
2010, 36, 254–258.
The authors thank the National Natural Science Foundation of
China (No. 22072038, 21673069, U20A20122 and 21503074), the
Hubei Student Innovation and Entrepreneurship Foundation of
China (S201910512016 and S202010512082), and the Applied
Basic Frontier Project of Wuhan Science and Technology
Bureau (2019010701011415).
´
24 A. E. Gerbase, J. R. Gregorio, M. Martinelli, M. C. Brasil and
N. F. Mendes, J. Am. Oil. Chem. Soc., 2002, 79, 179–181.
25 G. S. Owens, J. Arias and M. M. Abu-Omar, Catal. Today,
2000, 55, 317–363.
26 S. B. Tian, Q. Fu, W. X. Chen, Q. C. Feng, Z. Chen, J. Zhang,
D. S. Wang and Y. D. Li, Nat. Commun., 2018, 9, 2353–2360.
27 S. Shit, D. Saha, T. G. Row and C. Rizzoli, Inorg. Chim. Acta,
2014, 415, 103–110.
References
28 B. Tang, X. H. Lu, D. Zhou, P. Tian, Z. H. Niu, J. L. Zhang,
X. Chen and Q. H. Xia, Catal. Commun., 2013, 31, 42–47.
29 B. Tang, X. H. Lu, D. Zhou, J. Lei, Z. H. Niu, J. Fan and
Q. H. Xia, Catal. Commun., 2012, 21, 68–71.
30 K. Huang, Z. L. Wang and D. F. Wu, J. Chem. Sci., 2018, 130,
62–74.
1 O. A. Wong and Y. Shi, Chem. Rev., 2008, 108, 3958–3987.
2 J. S. Jirkovsky, M. Busch, E. Ahlberg, I. Panas and P. Krtil,
J. Am. Chem. Soc., 2011, 133, 5882–5892.
3 U. N. Gupta, N. F. Dummer, S. Pattisson, R. L. Jenkins,
D. W. Knight, D. Bethell and G. J. Hutchings, Catal. Lett.,
2015, 145, 689–696.
31 S. Muratsugu, Z. H. Weng and M. Tada, ACS Catal., 2013, 3,
2020–2030.
4 J. W. Brown, Q. T. Nguyen, T. Otto, N. N. Jarenwattananon,
¨
S. Gloggler and L. S. Bouchard, Catal. Commun., 2015, 59,
32 L. Canali and D. C. Sherrington, Chem. Soc. Rev., 1999, 28, 85–93.
33 M. P. Fernandez-Ronco, J. Kluge and M. Mazzotti, Cryst.
Growth Des., 2013, 13, 2013–2024.
34 M. R. Maurya, S. Sikarwar and M. Kumar, Catal. Commun.,
2007, 8, 2017–2024.
50–54.
5 A. Ansmann, R. Kawa and M. Neuss, US Pat., 7083780 B2,
2006, assigned to Cognis Deutschland.
6 R. Chakrabarty, B. K. Das and J. H. Clark, Green Chem., 2007,
9, 845–848.
35 E. Zamanifar, F. Farzaneh, J. Simpson and M. Maghami,
Inorg. Chim. Acta, 2014, 414, 63–70.
7 A. K. Suresh, M. M. Sharma and T. Sridhar, Ind. Eng. Chem.
Res., 2000, 39, 3958–3997.
36 X. H. Lu, J. Lei, X. L. Wei, X. T. Ma, T. J. Zhang, W. Hu, D. Zhou
and Q. H. Xia, J. Mol. Catal. A: Chem., 2015, 400, 71–80.
37 R. Neumann and M. Dahan, Nature, 1997, 388, 353–355.
38 Y. Nishiyama, Y. Nakagawa and N. Mizuno, Angew. Chem.,
Int. Ed., 2001, 40, 3639–3641.
39 M. E. Davis, Nature, 2002, 417, 813–821.
40 M. Masteri-Farahani and S. Mirshekar, Colloids Surf., A,
2018, 538, 387–392.
8 Y. Qi, Y. Luan, J. Yu, X. Peng and G. Wang, Chem. – Eur. J.,
2014, 20, 1–10.
9 M. Farias, M. Martinelli and D. P. Botteg, Appl. Catal., A,
2010, 384, 213–219.
10 K. A. Jorgensen, Chem. Rev., 1989, 89, 431–458.
11 R. Mbeleck, M. L. Mohammed, K. Ambroziak,
D. C. Sherrington and B. Saha, Catal. Today, 2015, 256,
287–293.
41 Y. Qi, Y. Luan, J. Yu, X. Peng and G. Wang, Chem. – Eur. J.,
2014, 20, 1–10.
12 L. L. Bai, L. Yao, Y. H. Yang and J. M. Lee, Chem. Commun.,
2015, 51, 4259–4262.
42 J. Zhao, W. Y. Wang, H. L. Tang, D. Ramella and Y. Luan,
Mol. Catal, 2018, 456, 57–64.
13 E. Poli, J. M. Clacens, J. Barrault and Y. Pouilloux, Catal.
Today, 2009, 140, 19–22.
14 N. S. Patil, R. Jha, B. S. Uphade, S. K. Bhargava and 43 P. Neves, A. C. Gomes, T. R. Amarante, F. A. Paz,
V. R. Choudhary, Appl. Catal., A, 2004, 275, 87–93.
15 D. S. Pinnaduwage, L. Zhou, W. W. Gao and C. M. Friend,
J. Am. Chem. Soc., 2007, 129, 1872–1873.
M. Pillinger, I. S. Gonçalves and A. A. Valente, Microporous
Mesoporous Mater., 2015, 202, 106–114.
44 F. A. Chavez, J. M. Rowland, M. M. Olmstead and
P. K. Mascharak, J. Am. Chem. Soc., 1998, 120, 9015–9027.
16 G. J. Hutchings, Chem. Commun., 2008, 1148–1164.
17 H. Tanaka, H. Nishikawa, T. Uchida and T. Katsuki, J. Am. 45 P. Buranaprasertsuk, Y. Tangsakol and W. Chavasiri, Catal.
Chem. Soc., 2010, 132, 12034–12041.
Commun., 2007, 8, 310–314.
18 M. Guidotti, N. Ravasio, R. Psaro, E. Gianotti, L. Marchese 46 J. Bjerrum and J. P. McReynold, Inorg. Synth., 1946, 2, 216–221.
and S. Coluccia, Green Chem., 2003, 5, 421–424.
47 D. G. Hill and A. F. Rosenberg, J. Chem. Phys., 1956, 24,
19 J. P. Zhang, S. Nagamatsu, J. Du, C. L. Tong, H. H. Fang,
1219–1231.
D. H. Deng, X. Liu, K. Asakura and Y. Z. Yuan, J. Catal., 48 N. Tanaka and M. Nanjo, B. Chem. Soc. Jpn, 1964, 37,
2018, 367, 16–26. 1330–1332.
20 M. G. Topuzovaa, S. V. Kotova and T. M. Kolev, Appl. Catal., 49 K. Mech, P. Zabinski and R. Kowalik, Electrochim. Acta,
A, 2005, 281, 157–166. 2012, 87, 254–259.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021
New J. Chem., 2021, 45, 2147--2156 | 2155

View Article Online
Paper
NJC
50 G. W. Watt, Inorg. Chem., 1964, 3, 325–329.
51 Z. Zivkovic, J. Therm. Anal. Calorim., 1994, 41, 99–104.
54 G. Xu, Q. H. Xia, X. H. Lu, Q. Zhang and H. J. Zhan, J. Mol.
Catal. A: Chem., 2007, 266, 180–187.
52 B. J. Tufts, I. L. Abrahams and C. E. Caley, J. Am. Chem. Soc., 55 P. P. Tao, X. H. Lu, H. F. Zhang, R. Jing, F. F. Huang, S. Wu,
1990, 112, 5123–5136. D. Zhou and Q. H. Xia, Mol. Catal., 2019, 463, 8–15.
53 E. Bosch and J. K. Kochi, J. Am. Chem. Soc., 1996, 118, 56 K. H. Tong, K. Y. Wong and T. H. Chan, Org. Lett., 2003, 5,
1319–1329.
3423–3425.
2156 | New J. Chem., 2021, 45, 2147--2156
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021