An active and stable multifunctional catalyst with defective UiO-66 as a support for Pd over the continuous catalytic conversion of acetone and hydrogen

Article abstract of DOI:10.1039/d0ra09217g

The one-pot synthesis of methyl isobutyl ketone (MIBK) and methyl isobutyl methanol (MIBC) from acetone and hydrogen is a typical cascade reaction comprised of aldol condensation-dehydration-hydrogenation. Pd loss and aggregation during long term operation are typical problems in industrial application. In this paper, an active and stable catalyst was achieved with defective UiO-66 as a support for Pd, which was synthesized with the ratio 15?:?1 of ZrOCl₂·8H₂O to ZrCl₄as Zr-precursors. The resultant Pd catalyst remained active for at least 1000 h with a MIBK + MIBC selectivity of 84.87-93.09% and acetone conversion of 45.26-53.22% in a continuous trickle-bed reactor. Besides the increased Br?nsted acid amount generated by the defect sites was favorable for the activity, the cavity confinement in the UiO-66 (R= 15?:?1) structure also efficiently prevented Pd loss and aggregation during the long term run. The contrast of the characterization of the fresh and used Pd/UiO-66 (R= 15?:?1) indicated that the deactivation of the catalyst was attributed to carbonaceous accumulation on the catalyst surface, which could be easily regenerated by calcination. This work supplied a new alternative for the design and utilization of industrial catalysts for MIBK and MIBC synthesis.

Full text of DOI:10.1039/d0ra09217g

RSC Advances
View Article Online
View Journal | View Issue
PAPER
An active and stable multifunctional catalyst with
defective UiO-66 as a support for Pd over the
continuous catalytic conversion of acetone and
hydrogen†
Cite this: RSC Adv., 2021, 11, 48
a
a
a
Yingjie Hu, Yuxin Mei, Baining Lin, Xuhong Du,a Fan Xu,a Huasheng Xie,b
*
Kang Wangb and Yonghua Zhou
*
a
The one-pot synthesis of methyl isobutyl ketone (MIBK) and methyl isobutyl methanol (MIBC) from acetone and
hydrogen is a typical cascade reaction comprised of aldol condensation-dehydration-hydrogenation. Pd loss
and aggregation during long term operation are typical problems in industrial application. In this paper, an
active and stable catalyst was achieved with defective UiO-66 as a support for Pd, which was synthesized
with the ratio 15 : 1 of ZrOCl2$8H2O to ZrCl4 as Zr-precursors. The resultant Pd catalyst remained active for
at least 1000 h with a MIBK + MIBC selectivity of 84.87–93.09% and acetone conversion of 45.26–53.22% in
a continuous trickle-bed reactor. Besides the increased Brønsted acid amount generated by the defect sites
was favorable for the activity, the cavity confinement in the UiO-66 (R ¼ 15 : 1) structure also efficiently
prevented Pd loss and aggregation during the long term run. The contrast of the characterization of the fresh
and used Pd/UiO-66 (R ¼ 15 : 1) indicated that the deactivation of the catalyst was attributed to
carbonaceous accumulation on the catalyst surface, which could be easily regenerated by calcination. This
work supplied a new alternative for the design and utilization of industrial catalysts for MIBK and MIBC synthesis.
Received 29th October 2020
Accepted 11th December 2020
DOI: 10.1039/d0ra09217g
rsc.li/rsc-advances
90% at 1–10 MPa and 120–150 C.4–7 Nevertheless, due to the
ꢀ
1. Introduction
poor thermal stability of the resin, the Pd loss and aggregation
during long term operation is inevitable, which leads to the
deactivation of catalyst. Aiming at solving this problem,
although lots of bifunctional catalysts, including Pd/metal
oxide,8,9 Pd/molecular sieve and Pd/resin,10,11 Ni-based,12,13
Cu-based,14 and Pt-based,15 have been reported during the
recent decades, the high active and stable catalyst is still
desirable.
Metal–organic frameworks (MOF), in particular Zr-based
MOF,16–18 have been extensively investigated for acid catalyzed
reactions such as citronellal cyclization,19 esterication,20 transfer
hydrogenation,21 and cross-aldol condensation.22 Besides the acid
property, the cage structure of MOF also allows it to act as the
excellent support for metal nanoparticles.23–25 UiO-66, as a kind of
traditional Zr-based MOF, contains metal nodes composed of
a zirconium oxide complex bridged by terepthalic acid ligands and
possesses the high thermal stability.26 In addition, it is well
recognized that the defect sites of UiO-66, mainly the missing
linkers, correlates closely with the catalytic performance, since the
acid–base property arises mainly from the defect sites. Therefore,
to achieve the high performance of UiO-66 in catalysis, more
defects are desirable to be created. In common, the defects in UiO-
66 can be tailored via adjusting the synthesis conditions, including
incorporation of modulators,27–29 thermal activation/dehydration,30
metal cation substitution31 and linker modication.32
Methyl isobutyl ketone (MIBK) is one of the most important
chemical products widely used in the coatings, pharmaceutical,
and petrochemical industries.1 Methyl isobutyl methanol
(MIBC), synthesized via MIBK hydrogenation, is an excellent
medium-boiling solvent and used as a raw material for organic
synthesis and mineral oaters. With the development of the
global economy, the demand for MIBK and MIBC in Asia,
especially China and South Korea, will increase further, due to
the pull of rubber antioxidant 4020 and high-solid coatings.
At present, MIBK is mainly manufactured by a one-pot
process, composed of three steps, say, acetone condensation
catalyzed by acid or base to diacetone alcohol (DAA), DAA
dehydration catalyzed by acid to mesityl oxide (MO), and MO
hydrogenation catalyzed by Pd to MIBK.2,3 Normally, MIBC
could also be generated in this one-pot process. To date, the
only commercialized multifunctional catalyst is Pd/resin,
which possess both the acid and Pd sites, which delivered an
acetone conversion of 25–50% with a MIBK selectivity of 70–
aCollege of Chemistry and Chemical Engineering, Central South University, Changsha
410083, China. E-mail: baininglin@csu.edu.cn; zhouyonghua@csu.edu.cn
bCangzhou Dahua Group Company, Ltd, Cangzhou 061000, China
† Electronic supplementary information (ESI) available. See DOI:
10.1039/d0ra09217g
48 | RSC Adv., 2021, 11, 48–56
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Paper
RSC Advances
In this paper, the utilizing of the mixture Zr-precursor of on JSM-7610FPlus. A structural characterization was obtained by
ZrOCl2$8H2O and ZrCl4 was rstly developed to create defect-rich X-ray diffraction (XRD). XRD patterns of catalysts were recorded
UiO-66 and then Pd was loaded thereon to fabricate the multi- between 5ꢀ and 50ꢀ (2q) on a Bruker D8 Advance A25 diffrac-
functional catalyst for acetone hydrogenation reaction. As expected, tometer with monochromatic Cu Ka radiation (l ¼ 0.154 nm).
the catalytic performance results exhibited that a high active and Thermogravimetric (TG) analysis was carried out in a TA Instru-
stable catalyst was achieved, with the at least 1000 h stability with ments TGA Q500 at a heating rate of 10 ꢀC minꢁ1 to 800 ꢀC in air.
MIBK + MIBC selectivity of 84.87–93.07% and acetone conversion of The measurements of the acidity and basicity of the catalysts
45.26–53.22%.
were accomplished by Autosorb-iQ-C equipped with a thermal
conductivity detector. The sample was pretreated in helium gas at
400 ꢀC for 2 h and then cooled down to 50 ꢀC. The surface of the
materials was saturated with NH3 (or CO2) for 60 min. TPD prole
2. Experimental
2.1 Materials
ꢁ1
ꢀ
of NH3 (or CO2) was recorded at the rate of 10 C min from
50 ꢀC to 800 ꢀC by a He ow. Pd mass content was measured with
a PerkinElmer Inductively Coupled Optical Emission Spectrom-
eter (ICP), model Optima 5300DV, equipped with a peristaltic
pump and across-ow nebulizer, coupled to a Ryton double pass
spray chamber of the Scott type. Before analysis, Pd/UiO-66 (R ¼
x : y) was immersed in aqua regia solution (v(HCl + HNO3)/vH2O
¼ 20 : 30) to dissolve Pd into solution. The CO chemisorption
measurements were performed in MICROMERITICS AutoChem
II 2920 equipment. Before analysis, the catalysts were reduced
All chemicals and reagents were obtained from commercial
sources and used as received, unless otherwise stated. Specically,
zirconium chloride (ZrCl4), zirconyl chloride octahydrate (ZrOCl2-
$8H2O), glacial acetic acid (HAc), palladium chloride (PdCl2) and
1,4-benzenedicarboxylic acid (BDC) were purchased from Aladdin
Reagent (Shanghai) Company of China, and anhydrous N,N-
dimethylformamide (DMF) sodium chloride (NaCl2) and meth-
anol was purchased from Sinopharm Chemical Reagent Co., Ltd.
ꢀ
under 10% H2/Ar ow at 200–250 C for 3 h, and subsequently
2.2 Catalyst preparation
under He (50 mL minꢁ1) ow at 150 ꢀC for 1 h to blow away H2
adsorbed on the surface of catalysts. Then, CO chemisorption
analysis using double isotherm methodology was performed at
35 ꢀC and metal dispersion (or the particle size) was calculated.
Potentiometric acid–base titrations were completed with a Met-
rohm Titrando 905 autotitrator equipped with Dosino 800 20 and
10 mL dosing units. X-ray photoelectron spectroscopy (XPS) was
performed using an ESCALAB250Xi instrument with an Al Ka X-
ray source (1486.6 eV), under about 2 ꢂ 10ꢁ9 mbar at room
temperature and a pass energy of 20 eV.
UiO-66 was prepared following a procedure previously reported33
with slight modications. In a typical synthetic process, 3.28 g of
ZrCl4 and 2.32 g of BDC were rst added in 160 mL of DMF. Then,
1 mL of deionized water and 24 mL of HAc were successively
dropped into the mixture and stirring was initiated until all
precursors were completely dissolved. The obtained mixture was
transferred to a 500 mL of hydrothermal reaction vessel with
Teon liner, and kept at 120 ꢀC for 24 h under static conditions.
Aer cooling to room temperature, the solid MOF were collected
via centrifugation and thoroughly washed with a mixture of
methanol and DMF (v/v ¼ 1 : 4) for three times, and methanol for
two times. Finally, the obtained sample was calcined in N2
atmosphere under 180 ꢀC for 3 h and denoted by UiO-66.
Synthesis of UiO-66 (R ¼ x : y). The synthesis technique of
UiO-66 (R ¼ x : y) was similar to that of UiO-66, except replacing
ZrCl4 with the mixture Zr-precursor. Typically, mix ZrOCl2-
$8H2O and ZrCl4 in a mass ratio of 0 : 1, 1 : 1, 3 : 1, 7 : 1, 9 : 1,
15 : 1, 19 : 1 and 1 : 0 respectively. The obtained samples were
denoted by UiO-66 (R ¼ x : y), and x : y represented the mass
ratio of ZrOCl2$8H2O to ZrCl4.
2.4 Catalytic testing
Acetone hydrogenation reaction was performed in a batch
reactor with an inner volume of 40 mL. The reactor was
equipped with a mechanically driven agitator, thermocouples,
and inlet and outlet gas purge lines. Typically, 5 mL of acetone
and experimentally desired amount of catalyst were added to
the reactor chamber and mixed. The reactor was tightly sealed
and purged three times with H2 to remove air. Then the reactor
was pressurized with H2 at 0.75 MPa, and the reaction pro-
ceeded at various temperatures and time. Aer the reaction
nished, cold-water quenching was applied to the reactor to
stop the reaction. The liquid product was separated from the
solid residue (if formed) and the catalyst by ltration.
The performance of acetone hydrogenation was also tested
in a continuous trickle-bed reactor using 6 mL catalyst. The
reaction conditions, reaction temperature, hydrogen pressure,
liquid hour space velocity (LHSV), and feed ratio of acetone to
H2, were optimized. The long-term run of the catalyst in the
trickle-bed reactor was performed under the following condi-
tions: 0.18 mL minꢁ1 of acetone ow, 12 mL minꢁ1 of hydrogen
Synthesis of Pd/UiO-66 and Pd/UiO-66 (R ¼ x : y) catalysts.
The Pd catalysts were prepared by a wet impregnation tech-
nique. At room temperature, UiO-66 or UiO-66 (R ¼ x : y) was
added into the Na2PdCl4 aqueous solution and the mixture was
subsequently stirred for 4 h, followed by reduction by a KBH4
solution for 30 min and washing with ethanol aqueous solution
(volume ratio 1 : 1) to neutral. Eventually, vacuum drying at
ꢀ
50 C was carried out. The obtained samples were denoted by
Pd/UiO-66 and Pd/UiO-66 (R ¼ x : y) catalysts, respectively.
2.3 Catalyst characterizations
Fourier transform infrared (FT-IR) spectra was measured on ow, 120 ꢀC, 2.0 MPa and LHSV of 1.8 hꢁ1. A Shimadzu GC-2010
a Nicolet 6700 Fourier transform spectrometer (Nicolet Magana gas chromatograph (GC) equipped with a ame ionization
Co.). Scanning electron microscopy (SEM) images was recorded detector (FID) and a 30 m ꢂ 0.25 mm ꢂ 0.25 mm Insert Cap
© 2021 The Author(s). Published by the Royal Society of Chemistry
RSC Adv., 2021, 11, 48–56 | 49

View Article Online
RSC Advances
Paper
WAX capillary column was used to determine the amounts of acetone conversion of 34.00%, and a MIBK + MIBC selectivity of
acetone, MIBK, MIBC, MO, DAA, and others in the product. 92.72%. This indicates that both UiO-66 and UiO-66 (R ¼ 1 : 0)
Total analysis time was 20 min. The denition of acetone possess the acid–base property for acetone condensation. By
conversion (xA), selectivity to MIBK + MIBK (SBK+BC) and MIBK + increasing the ratio of ZrOCl2$8H2O to ZrCl4 in the raw material
MIBC yield (YBK+BC) were calculated according to the following from 1 : 1 to 19 : 1, the acetone conversion of Pd/UiO-66 (R ¼
formulas.
x : y) catalyst increases rstly and then drops gradually. In
addition, the selectivity increases at the beginning and then
varies little. Among them, Pd/UiO-66 (R ¼ 15 : 1) exhibits the
highest MIBK + MIBC yield of 51.95% with an acetone conver-
sion of 54.44% and a MIBK + MIBC selectivity of 95.38%.
The catalytic performances of Pd/UiO-66 (R ¼ 15 : 1) catalysts
with different palladium loading are shown in Table 2. UiO-66
(R ¼ 15 : 1) support itself shows the acetone conversion of
29.51%, the MIBK + MIBC selectivity of 9.72%, and MO selec-
tivity of 68.13%. This indicates that UiO-66 (R ¼ 15 : 1) support
itself indeed only has acid–base function, without hydrogena-
tion function. When Pd loading is higher than 0.3 wt%, MO is
almost completely converted. With the increase of the amount
of Pd, there is a climax of the MIBK + MIBC yield. It is noted that
too much Pd is not benecial to the MIBK + MIBC yield due to
the decreased acetone conversion, probably due to the coverage
of the acid–base sites by Pd. Therefore, 0.5 wt% Pd/UiO-66 (R ¼
15 : 1) catalyst is chosen here for further investigation in the
trickle-bed reactor. We compare the performance of Pd/UiO-66
(R ¼ 15 : 1) with those reported in a batch liquid phase process
as shown in Table S1.† It is seen that Pd/UiO-66 (R ¼ 15 : 1)
moles of acetone in the product
moles of acetone in the feed
xA ¼ 1 ꢁ
ꢂ 100
moles of acetone converted to BK þ BC
SBKþBC
¼
ꢂ 100
moles of acetone converted
YBK+BC ¼ xA ꢂ SBK+BC/100
3. Results and discussion
3.1 Catalytic performance
The catalytic performance of the Pd catalysts loaded on various
UiO-66 (R ¼ x:y) for acetone hydrogenation in the batch reactor
are shown in Table 1. It is observed that Pd/UiO-66 delivers an
acetone conversion of 25.31%, and a MIBK + MIBC selectivity of
94.00%. Comparatively, Pd/UiO-66 (R ¼ 1 : 0) exhibits an
Table 1 Effect of ZrOCl2$8H2O/ZrCI4 ratio on the catalytic performance in the batch reactora
Selectivity (%)
Acetone conversion
MIBK +
Catalyst
(%)
MIBK
MIBC
MO
DAA
IPA
Others
MIBC yield (%)
Pd/UiO-66
25.31
11.18
20.46
37.12
39.02
54.44
34.64
34.00
92.67
85.38
87.54
91.32
91.96
90.90
90.88
90.82
1.33
0.46
0.89
2.93
2.56
4.48
1.96
1.90
0.31
0.10
0.81
0.93
1.04
0.64
0.93
0.65
2.77
7.54
3.08
1.24
1.35
0.56
1.88
1.80
1.52
5.35
4.80
1.45
1.26
0.31
1.50
3.00
1.40
1.17
2.88
2.13
1.83
3.11
2.85
1.83
23.79
9.59
Pd/UiO-66 (R ¼ 1 : 1)
Pd/UiO-66 (R ¼ 3 : 1)
Pd/UiO-66 (R ¼ 7 : 1)
Pd/UiO-66 (R ¼ 9 : 1)
Pd/UiO-66 (R ¼ 15 : 1)
Pd/UiO-66 (R ¼ 19 : 1)
Pd/UiO-66 (R ¼ 1 : 0)
18.09
34.98
36.88
51.95
32.16
31.52
a
T ¼ 120 ꢀC, reaction time ¼ 2 h. IPA is the abbreviation of isopropanol. Others is mainly C9+. 0.5 wt% Pd loading.
Table 2 Effect of Pd loading on the catalytic performance of Pd/UiO-66 (R ¼ 15 : 1) in the batch reactora
Selectivity (%)
Acetone conversion
MIBK +
Catalyst
(%)
MIBK
MIBC
MO
DAA
IPA
Others
MIBC yield (%)
UiO-66 (R ¼ 15 : 1)
29.50
47.07
53.64
54.44
52.67
52.22
38.62
31.71
9.51
80.95
87.60
90.90
88.60
90.71
90.72
90.60
0.21
3.37
4.63
4.48
4.86
2.52
2.54
2.03
68.13
6.94
2.59
0.64
1.07
1.91
2.11
2.21
2.41
0.80
0.77
0.56
0.67
0.70
1.34
1.34
0.32
0.24
0.07
0.31
0.28
0.34
0.23
1.47
19.42
7.70
4.34
3.11
4.52
3.82
3.06
2.35
2.87
39.69
46.99
51.95
49.22
48.68
36.02
29.37
0.3wt% Pd/UiO-66 (R ¼ 15 : 1)
0.4wt% Pd/UiO-66 (R ¼ 15 : 1)
0.5wt% Pd/UiO-66 (R ¼ 15 : 1)
0.6wt% Pd/UiO-66 (R ¼ 15 : 1)
0.7wt% Pd/UiO-66 (R ¼ 15 : 1)
0.8wt% Pd/UiO-66 (R ¼ 15 : 1)
0.9wt% Pd/UiO-66 (R ¼ 15 : 1)
a
+
T ¼ 120 ꢀC, reaction time ¼ 2 h. IPA is the abbreviation of isopropanol. Others is mainly C9
.
50 | RSC Adv., 2021, 11, 48–56
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Paper
RSC Advances
delivers the best performance among them under the mild
reaction conditions (Table S1†).34–37
The reaction conditions, including the reaction temperature,
hydrogen pressure, LHSV and mole ratio of C3H6O/H2, over
0.5 wt% Pd/UiO-66 (R ¼ 15 : 1) were optimized via the orthog-
onal experiments with four factors and four levels. The effects of
these factors on the catalytic performance are shown in Tables
S2 and S3.† It is seen that as for 0.5 wt% Pd/UiO-66 (R ¼ 15 : 1)
operated in a trickle-bed reactor, the reaction temperature is the
most signicant one among the four factors. More experiment
results in Table 3 indicate that the increase of reaction
temperature from 80 ꢀC up to 160 ꢀC leads to an increase of the
acetone conversion from 32.95% to 56.89%, accompanied by
a volcanic-changed MIBK selectivity. The decline of MIBK
selectivity at higher temperature is due to the formation of
+
Fig. 1 Stability of 0.5 wt% Pd/UiO-66 (R ¼ 15 : 1) catalyst. Reaction
conditions: volume of catalyst ¼ 6 mL, T ¼ 120 ꢀC, P ¼ 2.0 MPa, H2
heavier condensation products (C9 ).3 In addition, it is noted
that more MIBK is hydrogenated further to MIBC with the
increase of temperature. Except the reaction temperature, the
mole ratio of C3H6O/H2 is the second signicant factor on the
catalyst performance. Considering the yield of MIBK + MIBC,
the mole ratio of C3H6O/H2 should be maintained at 1 : 2,
consistent with the theoretic value. Therefore, the optimal
reaction conditions of Pd/UiO-66 (R ¼ 15 : 1) catalyst is reaction
temperature of 120 ꢀC, hydrogen pressure of 2 MPa, LHSV of 1.8
hꢁ1 and mole ratio of C3H6O/H2 of 1 : 2, which deliver the
highest MIBK + MIBC yield of 51.95%.
The long-term stability of 0.5 wt% Pd/UiO-66 (R ¼ 15 : 1)
catalyst in the trickle-bed reactor under the optimal conditions
was investigated (Fig. 1). Aer 1000 h run, the acetone conver-
sion drops by 7.96% from 53.22% to 45.26%, and the MIBK +
MIBC selectivity drops by 8.22% from 93.09% to 84.87%. Zhu4
reported that Pd/S (PS-GO) catalyst was tested for 1000 h and the
acetone conversion dropped from 50.85% to 39.71% by 10.14%
and the MIBK selectivity dropped from 77.86% to 73.47% by
4.39%. Duan3 showed that the TiO2–SiO2/SiO2 (F)-1&Pd/Cor
catalyst was active for 300 h with a MIBK selectivity of 83–93%
at 26–36% acetone conversion. Comparatively, the catalyst in
this paper displays the better stability than those reported in
literature, presenting the prospect for industrial application.
flow rate ¼ 12 mL minꢁ1 and LHSV ¼ 1.8 hꢁ1
.
3.2 Characterizations of catalysts
The XRD patterns of UiO-66, UiO-66 (R ¼ 15 : 1) and Pd/UiO-66
(R ¼ 15 : 1) in Fig. 2 exhibit the characteristic peaks of the
typical UiO-66 structure.16 No additional peaks can be observed,
indicating that no impure crystalline phases are formed. These
suggest that the variation of the Zr precursors and the Pd
loading process have no inuence on the bulk structure of UiO-
66. In addition, it is noted that no characteristic peak of Pd at 2q
¼ 40.1ꢀ and 46.7ꢀ can be detected due to the quite low loading
amount of Pd.
The SEM images of UiO-66 and UiO-66 (R ¼ 15 : 1) are shown
in Fig. 3. Obviously, two samples display different morphology.
The UiO-66 granules display the cubic octahedron with the edge
length size of around 150 nm (Fig. 3a and b). Comparatively,
UiO-66 (R ¼ 15 : 1) sample exhibits the amorphous morphology
composed of aky granules (Fig. 3c and d). It is noteworthy that
although the macroscopic morphology of UiO-66 and UiO-66 (R
¼ 15 : 1) seems quite different, both of them possess the same
crystal structure revealed by XRD. It is known that the
Table 3 Catalytic performance of 0.5 wt% Pd/UiO-66 (R ¼ 15 : 1) at different temperature in the trickle-bed reactora
Selectivity (%)
Temperature
(ꢀC)
Acetone conversion
(%)
MIBK +
MIBC yield (%)
MIBK
MIBC
MO
DAA
IPA
Others
80
90
32.95
35.13
40.43
49.13
54.44
54.59
54.53
55.15
56.89
87.03
86.11
87.08
91.63
90.90
89.11
87.69
85.43
82.69
1.47
1.60
2.89
3.72
4.48
4.53
4.55
4.91
4.97
5.70
6.95
4.80
0.63
0.64
0.62
0.59
0.57
0.56
3.38
2.48
1.18
1.11
0.56
0.52
0.49
0.40
0.41
0.12
0.11
0.52
0.61
0.31
1.68
2.87
3.40
3.87
2.30
2.75
3.53
2.30
3.11
3.54
3.81
5.29
7.50
29.16
30.81
36.37
46.84
51.95
50.11
50.30
48.82
49.87
100
110
120
130
140
150
160
a
The catalyst volume is 6 mL, P ¼ 2.0 MPa, H2 ow rate ¼ 12 mL minꢁ1, and time on stream ¼ 2 h. IPA is the abbreviation of isopropanol. Others is
+
mainly C9
.
© 2021 The Author(s). Published by the Royal Society of Chemistry
RSC Adv., 2021, 11, 48–56 | 51

View Article Online
RSC Advances
Paper
Fig. 4 N2 adsorption–desorption isotherms and pore distribution
curves of UiO-66 and UiO-66 (R ¼ 15 : 1).
Fig. 2 XRD patterns of UiO-66, UiO-66 (R ¼ 15 : 1) and Pd/UiO-66 (R
¼ 15 : 1).
suggesting the main existence of mesopores, which probably
are the secondary particle piled pores of formed by aky gran-
ules. This was further conrmed by the pore size distribution
calculated from the N2 isotherms. Accordingly, the BET specic
surface area of UiO-66 (R ¼ 15 : 1) is 1068 m2 gꢁ1, the total pore
volume is 1.03 cm3 gꢁ1 and the proportion of mesopores volume
increases.
The thermal stability of UiO-66 and UiO-66 (R ¼ 15 : 1)
determined by TG measurement is shown in Fig. 5. The TG
curve of UiO-66 reveal three broad areas representing weight
loss. The rst area is linked to the removal of water between
30 ꢀC and 100 ꢀC. And the second one is caused by the removal
of the residual solvent N,N-dimethylformamide and/or of the
unreacted terephthalic acid between 100 ꢀC and 480 ꢀC. The
third one is assigned to the decomposition of the framework
between 400 ꢀC and 600 ꢀC. The nal solid product aer
thermal treatment in air is 6ZrO2.41 It is observed from Fig. 5
that UiO-66 and UiO-66 (R ¼ 15 : 1) have almost the same TG
proles, suggesting their similar thermal stability.
crystalline process could be accelerated and small UiO-66
particles with several nanometers were formed when using
ZrOCl2$8H2O as Zr-precursor.37 Therefore, in this paper, the
particle size of UiO-66 (R ¼ 15 : 1) is much smaller than UiO-66.
This is benecial for the acetone hydrogenation reaction due to
the readily accessibility of active sites on UiO-66 (R ¼ 15 : 1).
The BET specic surface area and the porous structure of
UiO-66 and UiO-66 (R ¼ 15 : 1) are evaluated by N2 adsorption–
desorption experiment (Fig. 4 and Table S4†). UiO-66 exhibits
the type I isotherm in the region of low relative pressure, indi-
cating the main existence of micropores. The BET specic
surface area and the total pore volume are estimated to be 1677
m2 gꢁ1 and 0.88 cm3 gꢁ1, respectively. The narrow distribution
of the pore size justies that the micropores come from the
intrinsic structure of UiO-66. According to literature, the theo-
retical pore volume of UiO-66 is 0.77 cm3 gꢁ1 and the surface
area is 1160 m2 gꢁ1
.
Comparatively, UiO-66 (R ¼ 15 : 1)
17,33,38
displays the type II isotherm with the hysteresis loop,
The FT-IR spectra of UiO-66 and UiO-66 (R ¼ 15 : 1) in Fig. 6
show that both samples have the characteristic peaks of UiO-66.
Fig. 3 SEM images of UiO-66 (a and b) and UiO-66 (R ¼ 15 : 1) (c and
d).
Fig. 5 TG curves of UiO-66 and UiO-66 (R ¼ 15 : 1) in air.
52 | RSC Adv., 2021, 11, 48–56
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Paper
RSC Advances
Fig. 6 FT-IR spectra of UiO-66 and UiO-66 (R ¼ 15 : 1). (a) Mono-
dentate, (b) bidentate, (c) bridging.
The sharp peak at around 1400 cmꢁ1 represents the symmetric
stretching of the O–C–O carboxylate group of BDC ligands, and
the asymmetric stretching vibration is located at about
1587 cmꢁ1. The peak at 1670 cmꢁ1 is assigned to the stretching
vibrations of C]O of the BDC ligands. The peak at 1503 cmꢁ1
represents the C]C vibrations of benzene rings.16 At the lower
range of wavenumbers, the strong peaks at 848 cmꢁ1, 745 cmꢁ1
,
and 656 cmꢁ1 are attributed to the vibrations of O–H and C–H
bonds in the BDC ligands. These peaks ranged in 848–656 cmꢁ1
overlap with the Zr–O modes. In addition, the gap between the
two peaks of around 1400 cmꢁ1and 1500–1700 cmꢁ1 can give us
the information about the coordination modes of BDC and Zr
atoms.39,40 When the gap is larger than 200 cmꢁ1, BDC acts as
a monodentate ligand to Zr atoms. And the gap in the range of
130–200 cmꢁ1 indicates that BDC behaves as a bridging ligand.
Alternatively, when the gap is in the range of 50–150 cmꢁ1, BDC
acts as a bidentate ligand to Zr atoms.42,43 Fig. 6 shows that there
is a strong splitting of 187 cmꢁ1 with a weak one of 103 cmꢁ1 for
UiO-66 and UiO-66 (R ¼ 15 : 1). This indicates that the coordi-
nation modes of BDC in UiO-66 and UiO-66 (R ¼ 15 : 1) are
mainly consisted of bridging ligands with a small amount of
bidentate ligands.
The defects sites in the MOF materials could be identied by
acid–base titration method.44,45 The acid–base titration curves of
UiO-66 and UiO-66 (R ¼ 15 : 1) are measured as shown in Fig. 7.
To more accurately estimate the equivalence points in the
titration curves, the rst derivative curves are t to Lorentzian-
type peaks. It is observed that both samples display the titration
curves with three apparent inection points, which correspond
to three types of protons (m3-OH, Zr–OH2, and Zr–OH).46,47
Among them, Zr–OH2 and Zr–OH sites are generated as the
defects sites by missing linker in the synthesis process. There-
fore, the missing linker number can be further calculated from
Fig. 7. The calculation process is shown in Tables S5 and S6.†
And the missing linker numbers of UiO-66 and UiO-66 (R ¼
15 : 1) are estimated to be 1.58 and 1.89, respectively. This
suggests that more defect sites are introduced into UiO-66 (R ¼
15 : 1) than UiO-66 via utilizing the precursor mixture of ZrCl4
Fig. 7 Acid–base titration curve and first derivative curve for (a) UiO-
66 and (b) UiO-66 (R ¼ 15 : 1).
and ZrOCl2$8H2O. In addition, according to the calculation
results, the percentage of Zr–OH in defects increases from
65.9% to 83.4% in UiO-66 (R ¼ 15 : 1) compared with UiO-66.
To correlate the nature of defect sites of UiO-66 and UiO-66
(R ¼ 15 : 1) and the acid–base property required for acetone
condensation, the NH3-TPD and CO2-TPD were carried out as
shown in Fig. 8. It is well known that both Zr–OH and Zr–OH2
sites belong to typical Brønsted acid.48 The 6-fold-coordinated
Zr sites on the Zr6 nodes of UiO-66 are Lewis acid. In Fig. 8a,
the broad peak can be deconvoluted into three peaks. The peak
centered at about 290 ꢀC can be attributed to the Lewis acid of
Zr sites, and the other two peaks can be assigned to the
Brønsted acid sites of Zr–OH and Zr–OH2. It is noted that the
amount of Brønsted acid sites in UiO-66 (R ¼ 15 : 1) is obviously
increased, while the strength of Lewis acid sites is almost
constant compared with UiO-66. It was previously reported that
the change of specic coordination number of Zr has a minor
effect on the Lewis acid strength of the active site.49–52
As for base sites, the m3-O on the Zr6 nodes was normally
regarded as the Lewis base sites. When the missing linker
number is higher, the unsaturated coordinative number of Zr
atoms becomes higher and leads to the shortened length of Zr–
O bond, which further results in the weakening of the Lewis
© 2021 The Author(s). Published by the Royal Society of Chemistry
RSC Adv., 2021, 11, 48–56 | 53

View Article Online
RSC Advances
Paper
Fig. 9 TG curves of Pd/UiO-66 (R ¼ 15 : 1) before and after 1000 h
reaction in air.
this, the crystal structure of UiO-66 (R ¼ 15 : 1) has not changed
even aer 1000 h run, since the decomposition temperature of
ꢀ
the framework at 500 C is the same as the fresh one. In addi-
tion, the structure stability of UiO-66 (R ¼ 15 : 1) can also been
identied by the contrast of FT-IR spectra in Fig. S1.† The
change of the acid–base property of fresh and used Pd/UiO-66 (R
¼ 15 : 1) is revealed by NH3-TPD and CO2-TPD proles in
Fig. S2.† Aer 1000 h run, the Brønsted acid amount of the used
Fig. 8 NH3-TPD (a) and CO2-TPD (b) profiles of UiO-66 and UiO-66 Pd/UiO-66 (R ¼ 15 : 1) decreases, and the acidity and basicity is
(R ¼ 15 : 1).
strengthened as shown in Fig. S2b.† About the reason for the
change of acidity and basicity, we assume that the long chain
alcohol produced during the reaction may be co-adsorbed by
the adjacent zirconium atoms, leading to the reduction of the
acid sites. Then, the alcohol further deprotonate to form
alkoxide, and the proton is transferred to the m3 oxygen to
strengthen its basicity. Hence, compared with the fresh catalyst,
the CO2 desorption peak on the used Pd/UiO-66 (R ¼ 15 : 1)
shis to a higher temperature.
In addition, the surface chemical state of Pd, Zr and O on Pd/
UiO-66 and Pd/UiO-66 (R ¼ 15 : 1) was characterized by XPS as
shown in Fig. S3.† The existence of C, O, Pd and Zr elements on
Pd/UiO-66 and Pd/UiO-66 (R ¼ 15 : 1) is identied in Fig. S3(a).†
The Pd 3d and Zr 3p spectra are present in Fig. S3(b).† The two
basicity. This can be reected by the peak shi to lower
temperature from 410 ꢀC of UiO-66 to 355 ꢀC of UiO-66 (R ¼
15 : 1) in the CO2-TPD proles in Fig. 8b. In addition, the other
broad peak shiing to lower temperature could be attributed to
other kinds of basic sites related to oxygen in the framework.
In a word, compared with UiO-66, UiO-66 (R ¼ 15 : 1)
exhibits the obviously increased amount of Brønsted acid sites,
mainly Zr–OH and Zr–OH2 sites, caused by the increasing of
missing linker number. For acetone condensation and dehy-
dration reactions, except for base site, it is intensively reported
that Brønsted acid is the typical catalytic site for the reaction, for
instance sulfonate ion.5 In this paper, although the basicity of
UiO-66 (R ¼ 15 : 1) is weakened, the enhanced Brønsted acid
amount leads to the enhancement of the cascade reaction.
The comprehensive contrast of the characterizations of UiO-
66 and UiO-66 (R ¼ 15 : 1) has been presented in the above text,
which help us to understand the reason for the better perfor-
mance of Pd/UiO-66 (R ¼ 15 : 1). To further reveal the nature of
the activity drop of Pd/UiO-66 (R ¼ 15 : 1) aer 1000 h run, the
TG curves, FT-IR spectra, NH3-TPD and CO2-TPD proles of the
fresh and used catalysts are presented as follows. From the
contrast of the TG curves of the fresh and used Pd/UiO-66 (R ¼
15 : 1) in Fig. 9, it is seen that an increased weight loss 26%
strong peaks in Fig. S3(b)† are assigned to Zr 3p3/2 and Zr 3p1/2
,
respectively.53 And the two weak peaks at 335.6 and 341.1 eV can
be attributed to Pd 3d5/2 and Pd 3d3/2, respectively, which are
partially overlapped by the two peaks of Zr 3p. The similar XPS
results were previously reported by Guan53 and Yang.54 It is
observed in Fig. S3(b)† that the binding energy of Pd species on
both samples is slightly higher than the standard value of
335.3 eV of Pd metal, probably attributed to the interaction
between Pd and the UiO-66 framework.54 As for Zr 3d and O1s
spectra in Fig. S3(c and d),† no difference is observed between
Pd/UiO-66 and Pd/UiO-66 (R ¼ 15 : 1).
Furthermore, the Pd mass percent and Pd dispersion of fresh
Pd/UiO-66, fresh and used Pd/UiO-66 (R ¼ 15 : 1) were charac-
terized by ICP and CO chemisorption as shown in Table S7.†
ꢀ
ranged in 100–500 C is observed, suggesting the existence of
the carbonaceous accumulation on the used catalyst. Except
54 | RSC Adv., 2021, 11, 48–56
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Paper
RSC Advances
The two samples present the close Pd mass percent. And the carbonaceous accumulation on the catalyst surface. This
average particle size of Pd on Pd/UiO-66 is estimated to be belongs to the reversible deactivation and the catalyst could be
3.05 nm, slightly larger than that of 2.20 nm on Pd/UiO-66 (R ¼ easily regenerated by calcination. This work supplied a new
15 : 1). Considering the contrast of the Pd average particle size alternative for the design and utilization of industrial catalysts
and the surface chemical state between Pd/UiO-66 and Pd/UiO- for MIBK and MIBC synthesis.
66 (R ¼ 15 : 1), we can regard that there is no obvious difference
in view of the Pd particles loaded thereon. In addition, the
condensation of acetone is normally regarded as the rate- Conflicts of interest
determining step, instead of the hydrogenation step, for the
There are no conicts to declare.
overall cascade reactions. Therefore, the gap of catalytic
performance between Pd/UiO-66 and Pd/UiO-66 (R ¼ 15 : 1) can
be mainly ascribed to the difference of acid and base sites.
Acknowledgements
As for the fresh Pd/UiO-66 (R ¼ 15 : 1), the Pd mass percent is
0.486% with an average Pd particle size of 2.20 nm. Compara-
This work was supported by the National Natural Science
Foundation of China (Grant 21676303) and the Fundamental
tively, as for the used Pd/UiO-66 (R ¼ 15 : 1), the Pd mass
percent drops to 0.442% (by 9.05%) accompanied by the slight
Research Funds for the Central Universities of Central South
aggregation of Pd particle to 2.73 nm. The change of the Pd
University (Grant 2020zzts415).
mass and particle size is quite smaller than those reported in
literature,3,4 This suggests that the phenomena of metal loss
and aggregation of Pd particles can be efficiently inhibited by
References
loading Pd in the small cavities of UiO-66 (R ¼ 15 : 1), resulting
in the excellent stability in long term run. Therefore, the reason
for the deactivation of Pd/UiO-66 (R ¼ 15 : 1) can be mainly
attributed to the carbonaceous accumulation, instead of the Pd
loss and aggregation. The used Pd/UiO-66 (R ¼ 15 : 1) catalyst
aer 1000 h run was regenerated in our experiments. It was
1 N. Takarroumt, M. Kacimi, F. O. Bozon-Verduraz and
L. F. Liotta, J. Mol. Catal. A: Chem., 2013, 377, 42–50.
2 M. Mediavilla, L. Melo, Y. Dıaz, A. Albornoz and A. Llanos,
´
Microporous Mesoporous Mater., 2008, 116, 627–632.
3 D. Hui, W. Zhihui, C. Lifeng, L. Baining and Z. Yonghua, Ind.
Eng. Chem. Res., 2018, 57, 12358–12366.
4 Y. Zhu, B. Lin, Y. Hu, Z. Cai and Z. Yonghua, Mol. Catal.,
2019, 478, 110609.
5 W. Nicol and E. L. D. Toit, Chem. Eng. Process., 2004, 43,
1539–1545.
6 S. Talwalkar and S. Mahajani, Appl. Catal. A., 2006, 302, 140–
148.
calcined in air ow with 10 mL minꢁ1 of at 150 C for 1 h fol-
ꢀ
lowed by reduced in H2 ow with 10 mL minꢁ1 at 150 ꢀC for 1 h.
Then, the catalytic performance of the regenerated catalyst was
tested again under the same reaction conditions. The regen-
erated catalyst exhibited an acetone conversion of 53.23% and
a MIBK + MIBC selectivity of 94.19%, close to the initial
performance. This indicates the deactivation of Pd/UiO-66 (R ¼
15 : 1) is reversible and can be readily regenerated by calcina-
tion. This is valuable for the prolongation of the service life of
Pd catalyst.
7 E. D. Toit, R. Schwarzer and W. Nicol, Chem. Eng. Sci., 2004,
59, 5545–5550.
´
8 L. M. Gandıa, R. Malm, R. Marchand, R. Conanec, Y. Laurent
and M. Montes, Appl. Catal. A., 1994, 114, L1–L7.
9 J. J. Gamman, S. D. Jackson and F. A. Wigzell, Ind. Eng. Chem.
Res., 2010, 49, 8439–8443.
4. Conclusions
The cage structure, high thermal stability and the adjustability 10 S. M. Yang and Y. M. Wu, Appl. Catal. A., 2000, 192, 211–220.
of acid–base property of UiO-66 allowed it to act as an excellent 11 A. Chakrabarti and M. M. Sharma, React. Polym., 1993, 20, 1–
candidate to build multifunctional catalysts. In this paper, the
defective UiO-66 was synthesized with the optimal ratio 15 : 1 of 12 A. C. C. Rodrigues, J. L. F. Monteiro and C. A. Henriques, C.
ZrOCl2$8H2O to ZrCl4 as Zr-precursors rstly. Then, Pd was R. Chim., 2009, 12, 1296–1304.
loaded thereon and tested in a typical cascade reaction, the one- 13 R. Unnikrishnan and S. Narayanan, Mol. Catal., 1999, 144,
45.
pot synthesis of methyl isobutyl ketone and methyl isobutyl
173–179.
´
´
´
´
methanol from acetone and hydrogen. The Pd/UiO-66 (R ¼ 14 V. Chikan, a. Molnar and K. Balazsik, J. Catal., 1999, 184,
15 : 1) catalyst exhibited a stable run for at least 1000 h with
a MIBK + MIBC selectivity of 84.87–93.09% and acetone 15 G. Waters, O. Richter and B. Kraushaar-Czarnetzki, Ind. Eng.
conversion of 45.26–53.22% in a continuous trickle-bed reactor. Chem. Res., 2006, 45, 6111–6117.
Compared with pure UiO-66, the improvement of the activity 16 J. H. Cavka, S. Jakobsen, U. Olsbye, N. Guillou, C. Lamberti,
134–143.
can be attributed to the increased Brønsted acid amount caused
by the introduced defect sites. And the connement of UiO-66
S. Bordiga and K. P. Lillerud, J. Am. Chem. Soc., 2008, 130,
13850–13851.
(R ¼ 15 : 1) structure efficiently prevented the Pd loss and 17 H. Wu, Y. S. Chua, V. Krungleviciute, M. Tyagi, P. Chen and
aggregation during the long term run. The contrast of the T. Yildirim, J. Am. Chem. Soc., 2013, 135, 10525–10532.
characterizations of the fresh and used Pd/UiO-66 (R ¼ 15 : 1) 18 Y. Bai, Y. Dou, L. H. Xie, W. Rutledge, J. R. Li and H. Zhou,
indicated that the deactivation of the catalyst was attributed to
Chem. Soc. Rev., 2016, 45, 2327–2367.
© 2021 The Author(s). Published by the Royal Society of Chemistry
RSC Adv., 2021, 11, 48–56 | 55

View Article Online
RSC Advances
Paper
19 F. Vermoortele, M. Vandichel, B. V. d. Voorde, R. Ameloot, 37 P. Wang, S. Bai, J. Zhao, P. Su, Q. Yang and C. Li,
M. Waroquier, P. V. V. Speybroeck and P. D. Edition,
Angew. Chem. Int. Edit., 2012, 124, 4887–4890.
20 F. G. Cirujano, A. Corma and F. X. L. Xamena, Catal. Today,
2014, 257, 213–220.
21 A. H. Valekar, K. H. Cho, S. K. Chitale, D. Y. Hong, G. Y. Cha,
U. H. Lee, D. Hwang, C. Serre, J. S. Chang and Y. Hwang,
Green Chem., 2016, 18, 4542–4552.
22 J. Hajek, M. Vandichel and B. V. Voorde, J. Catal., 2015, 331,
1–12.
23 H. Fei and S. Cohen, Chem. Commun., 2014, 50(37), 4810–
4812.
ChemSusChem, 2012 Dec, 5, 2390–2396.
38 L. Hao, X. Li, M. J. Hurlock, X. Tu and Q. Zhang, Chem.
Commun., 2018, 54, 11817–11820.
39 L. Valenzano, B. Civalleri, S. Chavan, S. Bordiga,
M. H. Nilsen, S. R. Jakobsen, K. P. Lillerud and
C. Lamberti, Chem. Mater., 2011, 23, 1700–1718.
40 P. Ghosh, Y. J. Colon and R. Q. Snurr, Chem. Commun., 2014,
50, 11329–11331.
41 G. C. Shearer, S. Chavan, J. Ethiraj, J. G. Vitillo, S. Svelle,
U. Olsbye, C. Lamberti, S. Bordiga and K. P. Lillerud,
Chem. Mater., 2014, 26, 4068–4071.
24 L. Chen, R. Luque and Y. Li, Chem. Soc. Rev., 2017, 46, 4614– 42 F. Verpoort, T. Haemers, P. Roose and J. P. Maes, Appl.
4620. Spectrosc., 1999, 53, 1528–1534.
25 Q. Yang, Q. Xu and H. L. Jiang, Chem. Soc. Rev., 2017, 46, 43 K. B. Lausund and O. Nilsen, Nat. Commun., 2016, 7, 13578.
4774–4808.
26 C. A. Trickett, K. J. Gagnon, S. Lee, F. Gandara and
44 R. C. Klet, Y. Liu, T. C. Wang, J. T. Hupp and O. K. Farha, J.
Mater. Chem. A., 2016, 4, 1479–1485.
´
¨
H. B. Burgi, Angew. Chem., 2015, 128, 11314–11319.
45 T. J. Bandosz, M. Laskoski, J. Mahle, G. Mogilevsky and
G. W. Wagner, J. Phys. Chem. C, 2012, 116, 11606–11614.
46 T. G. Glover, G. W. Peterson, J. B. DeCoste and M. A. Browe,
Langmuir, 2012, 28, 10478–10487.
´
27 O. V. Gutov, M. G. Hevia, E. C. Escudero-Adan and A. Shar,
Inorg. Chem., 2015, 54, 8396–8400.
28 Z. Hu, I. Castano, S. Wang, Y. Wang, Y. Peng, Y. Qian, C. Chi,
X. Wang and D. Zhao, Cryst. Growth Des., 2016, 16, 2295– 47 G. Ye, D. Zhang, X. Li, K. Leng, W. Zhang, J. Ma, Y. Sun,
2301.
W. Xu and S. Ma, ACS. Appl. Mater. Inter., 2017, 9, 34937–
34943.
48 Y. Liu, R. C. Klet, J. T. Hupp and O. Farha, Chem. Commun.,
2016, 52, 7806–7809.
49 L. Hao, X. Li, M. Hurlock and X. Tu, Chem. Commun., 2018,
54, 11817–11820.
50 B. Bueken, N. V. Velthoven and T. Willhammar, Chem. Sci.,
2017, 8, 3939–3948.
29 L. Yuan, M. Tian, J. Lan, X. Cao and X. Wang, Chem.
Commun., 2018, 54, 370–373.
30 G. C. Shearer, S. Forselv, S. Chavan and S. Bordiga, Top.
Catal., 2013, 56, 770–782.
31 Z. Niu, Q. Guan, Y. Shi, Y. Chen, Q. Chen, Z. Kong, P. Ning,
S. Tian and R. Miao, New J. Chem., 2018, 42, 19764–19770.
32 Y. Wu, Q. Li, D. J. Rutstrom, M. Zhuravleva and C. Melcher,
Phys. Status Solidi Rapid Res. Lett., 2017, 1700403, 1–4.
33 D. Jiang, G. Fang, Y. Tong, X. Wu, Y. Wang, D. Hong,
´
¨
51 C. A. Trickett, K. J. Gagnon, S. Lee, F. Gandara, H.-B. Burgi
and O. M. Yaghi, Angew. Chem., 2015, 127, 11314–11319.
W. Leng, Z. Liang, P. Tu, L. Liu, K. Xu, J. Ni and X. Li, ACS 52 J. Hajek, B. Bueken and M. Waroquier, ChemCatChem, 2017,
Catal., 2018, 8, 11973–11978. 9, 2203–2210.
34 R. Ma, Y. Li, G. Wu, Y. He, J. Feng, Y. Zhao and D. Li, Chinese. 53 Q. Guan, B. Wang, X. Chai, J. Liu, J. Gu and P. Ning, Fuel,
J. Catal., 2018, 39, 1384–1394. 2017, 205, 130–141.
35 F. Al-Wadaani, E. F. Kozhevnikova and I. V. Kozhevnikov, J. 54 Y. Yang, D. Deng, D. Sui, Y. Xie, D. Li and Y. Duan,
Catal., 2008, 257, 199–205.
Nanomaterials, 2019, 9, 1698–1704.
36 D. Wang, Z. Wang, B. Zahng, H. Sun, P. Wu, Z. Gao and
W. Yang, Chem. React. Eng. Technol., 2017, 33(2), 89–297.
56 | RSC Adv., 2021, 11, 48–56
© 2021 The Author(s). Published by the Royal Society of Chemistry