The roles of metal-promoter interface on liquid phase selective hydrogenation of crotonaldehyde over Ir-MoOx/BN catalysts

Article abstract of DOI:10.1016/j.apcata.2021.118269

A series of MoO_x-promoted Ir/BN catalysts were tested for liquid phase selective hydrogenation of crotonaldehyde. The MoO_x-promotion could significantly improve the reactivity up to 5-fold. Such improvement was mainly due to the form

Full text of DOI:10.1016/j.apcata.2021.118269

Applied Catalysis A, General 623 (2021) 118269
Contents lists available at ScienceDirect
Applied Catalysis A, General
journal homepage: www.elsevier.com/locate/apcata
The roles of metal-promoter interface on liquid phase selective
hydrogenation of crotonaldehyde over Ir-MoO /BN catalysts
x
a
a
a
a
a
a,b
Cen Tang , Fangru Zhou , Yang Wen , Yumeng Xu , Aiping Jia , Mengfei Luo
,
a,b,
Jiqing Lu
*
a
College of Chemistry and Life Sciences, Zhejiang Normal University, Jinhua, 321004, China
b
Key Laboratory of the Ministry of Education for Advanced Catalysis Materials, Institute of Physical Chemistry, China
A R T I C L E I N F O
A B S T R A C T
A series of MoO -promoted Ir/BN catalysts were tested for liquid phase selective hydrogenation of crotonalde-
Keywords:
x
Selective hydrogenation
hyde. The MoO
mainly due to the formation of Ir-MoO
adsorption and activation of crotonaldehyde molecule as evidenced by a much lower activation energy (18.9 kJ
x
-promotion could significantly improve the reactivity up to 5-fold. Such improvement was
α
, β - Unsaturated aldehyde
x
interfacial sites in the promoted catalyst, which accelerated the
Adsorption and activation
Kinetics
ꢀ 1
ꢀ 1
mol on the 0.3Mo-3Ir/BN versus 38.4 kJ mol on the 3Ir/BN). The kinetic results revealed negative reaction
order of crotonaldehyde (up to -1.1) and high reaction order of H (up to 2.2), indicating the high coverage of
crotonaldehyde on the catalyst surface suppressed the adsorption of H due to its low solubility in the liquid
phase. Therefore, the adsorption and activation of H on the catalyst might be the rate determining step.
Reaction mechanism
2
2
2
However, the preferred adsorption of crotonaldehyde on the Ir-MoO
adsorption, which was another reason for the higher activity.
x 2
interface alleviated the suppression of H
1
. Introduction
the catalyst surface. Engineering of the catalyst surface is a promising
strategy to adjust the adsorption modes of the unsaturated aldehyde.
One approach is the use of reducible supports that present the strong
The selective hydrogenation of,
α, β - unsaturated aldehydes to their
corresponding unsaturated alcohols is of great importance in the syn-
thesis of fine chemicals such as fragrances, agrochemicals and phar-
2 2
metal support interaction (SMSI) effect. Oxides such as TiO [6] or CeO
[7] are good supports for the noble metals (e.g. Ir and Pt) to promote the
selectivity to the hydrogenation of crotonaldehyde. For example, it was
maceuticals [1,2]. The hydrogenation of
over heterogeneous catalysts employing H
α
, β - unsaturated aldehydes
2
as the reducing agent is very
reported that the Ir/TiO
2
catalysts gave high selectivity to crotyl alcohol
appealing because of its high atomic efficiency. Yet, it is not an easy task
to obtain high selectivity to the desired unsaturated alcohol as the hy-
in the hydrogenation of crotonaldehyde due to the formation of di-
σ
CO
–
2
mode adsorption of C O bond at Ir-TiO interface [8]. Such strategy of
–
–
drogenation tends to occur on the C
–
C bond in the aldehyde molecule,
the generation of interfacial sites could be extended to metal-promoter
[9–14] and metal-metal alloy [15–17] systems. Among the employed
noble metals, iridium (Ir) appears to be more selective than other metals
–
which is more thermodynamically favorable than that on the C
–
O bond
ꢀ 1
by about 35 kJ mol [3–5].
Efforts have been made to improve the catalytic performance for the
reaction, particularly the selectivity. It is well recognized that the dis-
tributions of the products are closely related to the adsorption geome-
tries of the unsaturated aldehyde molecules on the catalyst surface,
which consequently affects the hydrogenation behaviors. Since the
(e.g. Pt and Au) for
α, β - unsaturated aldehyde, especially for those small
molecules such as crotonaldehyde [10,18–22], due to its larger D-band
–
–
width resulting enhanced adsorption of C
O bond and suppressed
C bond [23,24]. It was reported that the Ir/SiO
catalyst [22] could give a selectivity of 77 % to crotyl alcohol in hy-
drogenation of crotonaldehyde, while on the Pt/SiO [25] and
Pd/NRGO [26] catalysts the dominant product is butanal. Moreover, by
addition of the second mental such as MoO , WO , NbO , FeO and ReO
could greatly enhanced the performance (both activity and selectivity)
–
–
adsorption of C
2
–
preferred adsorption of C
unsaturated alcohol while the adsorption of C
formation of saturated aldehyde, high selectivity to the unsaturated
–
O bond is beneficial to the formation of
2
–
–
C bond results in the
x
x
x
x
x
–
alcohol could be achieved if the C O bond predominantly adsorbs on
–
*
Corresponding author at: College of Chemistry and Life Sciences, Zhejiang Normal University, Jinhua, 321004, China.
E-mail address: jiqinglu@zjnu.cn (J. Lu).
https://doi.org/10.1016/j.apcata.2021.118269
Received 26 April 2021; Received in revised form 3 June 2021; Accepted 15 June 2021
Available online 17 June 2021
0
926-860X/© 2021 Elsevier B.V. All rights reserved.

C. Tang et al.
Applied Catalysis A, General 623 (2021) 118269
ꢀ
◦
of the Ir catalysts for hydrogenation of crotonaldehyde [27], due to the
generation metal-promoter interface which are beneficial for the reac-
mL min 1) at 300 C for 1 h. High-resolution transmission electron
microscopy (HRTEM) was performed on a JEM-2100 F microscopy
operated at 200 kV. Before the measurements, the catalysts were pre-
tion [18,27–29]. For example, the decoration of Ir by MoO
x
on TUD-1
O bond preferentially on
, which could react with the dissociated H species on Ir to form
–
ꢀ 1
◦
catalyst [29] resulted in the adsorption of C
–
reduced in a H
The reducibility of the catalyst was measured by hydrogen temper-
ature programmed reduction (H -TPR) experiment on a MicrotracBEL
2
flow (99.999 %, 26 mL min ) at 300 C for 1 h.
the MoO
x
crotyl alcohol. It is worth mentioning that although noble metals are the
main catalysts for the reaction, it was recently reported that non-noble
metal such as Co shows high performance in the selective hydrogena-
tion of various unsaturated aldehydes [30].
2
BELCAT II automatic chemisorption analyzer. 50 mg of the catalyst was
◦
loaded in a quartz tubular reactor, and was heated to 300 C in a pure O
2
ꢀ
1
◦
flow (30 mL min ) and then kept at 300 C for 0.5 h, followed by Ar
ꢀ
Kinetic study is helpful to understand the underlying reaction
mechanism. For gas phase selective hydrogenation of crotonaldehyde, it
purge (30 mL min 1) for 0.5 h to remove the gaseous or physically
adsorbed O . After the sample was cooled down to room temperature, it
was heated to 700 C at a rate of 10 C min in a 5 % H
2
◦
◦
ꢀ 1
was reported that the promoted activity on the Ir-FeO
related to the enhanced the adsorption of crotonaldehyde molecules at
Ir-FeO interface [12]. Regarding the liquid phase reaction, works per-
x
/BN catalyst was
2
+ 95 % Ar
mixture (30 mL min ). The signal was recorded by a thermal conduc-
tivity detector (TCD), and the H consumption was calibrated by the
ꢀ
1
x
2
formed in this area mainly addressed selectivity issues and few quanti-
tative kinetic data were reported. Reyes et al. [31] found near negative
second - order dependence on citral concentration and first - order
dependence on hydrogen pressure in liquid-phase hydrogenation of
reduction of CuO powder with a known amount.
X-ray photoelectron spectra (XPS) of the catalysts were recorded on
an ESCALAB 250Xi spectrometer equipped with a monochromatic Al
a
anode K radiation (1486.6 eV), and the charging effects were corrected
citral over Ir/TiO
2
catalysts, which revealed competitive adsorption
by setting the C1s binding energy of the adventitious carbon to 284.8 eV.
Prior to the measurement, the sample was reduced in a pretreatment
between citral and hydrogen and the hydrogen adsorption on the active
sites was the rate-limiting step. Koo-amornpattana and Winterbottom
ꢀ
1
◦
chamber in high purity hydrogen (> 99.99 %, 26 mL min ) at 300 C
for 1 h, and was cooled down to room temperature. Then the sample was
directly transferred to analysis chamber without being exposed to the
environment. The obtained XPS data were fitted by a XPS peak 4.1
software.
[
32] reported the kinetic of cinnamaldehyde hydrogenation on Pt- and
Pt-alloy catalysts and the authors found that the reaction was zero -
order in cinnamaldehyde concentration and first - order in hydrogen
partial pressure, which indicated that the reaction was significantly
transport controlled and gas absorption resistance was more significant
than liquid - solid mass transport. Also, liquid - phase selective hydro-
Diffuse reflectance infrared Fourier transform (DRIFT) spectra of CO
chemisorption on the samples were recorded using a Nicolet iS50 FT-IR
spectrometer equipped with a MCT detector and a PIKE DRIFT acces-
genation of crotonaldehyde over Au/Mg
2
AlO catalysts gave a reaction
order of 1.8 in respect to H partial pressure [33].
2
sory. Prior to the measurement, the sample was in situ reduced in a H
2
ꢀ
1
◦
A kinetic approach to a better mechanistic understanding of liquid
phase hydrogenation reactions has already proven to be valuable [34,
flow (99.99 %, 26 mL min ) at 300 C for 1 h and purged by a flow of N
2
ꢀ 1
◦
◦
(30 mL min ) at 300 C for 1 h. The sample was cooled down to 30 C
3
5]. Therefore, in the current work, we prepared a series of MoO
x
-
and the backgrounds were collected at temperature intervals (30–100
◦
ꢀ 1
2
C). After the sample was exposed to a 1 % CO in N (20 mL min ) for
promoted Ir/BN catalysts and tested them for liquid phase selective
hydrogenation of crotonaldehyde. The performance of the promoted
15 min and purged by N
2
◦
purge for another 1 h, it was heated from 30 to
◦
ꢀ 1
catalyst was compared with that of the bare Ir/BN, and the roles of MoO
x
100 C at a ramp of 10 C min , and the spectra were recorded with
subtracted the background at the corresponding temperature. In all
promotion were clarified by spectroscopic characterizations and kinetic
investigation.
ꢀ
1
cases the spectra were taken with a resolution of 4 cm and cumulative
4 scans.
In situ Fourier transform infrared (FTIR) spectroscopy of croto-
6
2
. Experimental
naldehyde adsorption and in - situ reaction was performed on the same
FTIR spectrometer (Nicolet iS50). The sample with a weight of about
2
.1. Catalyst preparation
1
5–20 mg was pressed into a 13 mm self-supported wafer and placed
◦
The supported Ir/BN catalyst was prepared by impregnating com-
into a quartz in situ IR cell. The catalyst was reduced at 300 C under a
2
ꢀ 1
ꢀ 1
mercial hexagonal BN (h-BN, Aladdin, 99.9 %, SBET = 18.5 m g ) with
an aqueous solution of H IrCl (Macklin, Ir content of 35 wt.% in HCl) at
H
2
flow (99.99 %, 26 mL min ) for 1 h and purged with a N
2
flow for
on the catalyst surface) then
2
6
another 1 h (to remove the adsorbed H
2
◦
room temperature. The suspension was continuously stirred for 4 h (rpm
cooled to 80 C, and the background was collected at certain tempera-
ture points. Crotonaldehyde vapor was introduced to the sample for 10
◦
=
450). After solvent evaporation by water bath at 90 C and drying at
◦
◦
◦
ꢀ 1
8
0 C overnight, the solid was heated to 500 C at a ramp of 5 C min
min by passing the N
2
flow through a saturated vapor generator con-
◦
and kept for 4 h in static air to obtain the catalyst. The MoO
x
promoted
taining liquid crotonaldehyde maintained at 0 C. After that the sample
Ir/BN catalysts were also prepared using an impregnation method. In
detail, a certain amount of calcined Ir/BN powder was added in an
was purged by N
2
for another 1 h in order to remove the gaseous and
physisorbed crotonaldehyde and then the spectrum was recorded. Then
◦ ◦
ꢀ 1
the catalyst was heated from 80 to 180 C at a ramp of 5 C min in the
aqueous solution of (NH
4
)
6
Mo
7
O
24 ⋅ 4H
2
O (Aladdin, 100 %) for 4 h at
◦
room temperature, followed by drying at 80 C overnight and calcina-
N
2
flow to monitor the desorption behaviors of crotonaldehyde. For the
◦
◦
tion at 500 C for 4 h to obtain the final catalyst. The nominal contents of
in-situ reaction, after the sample was reduced and cooled down to 30 C,
during which the backgrounds with different temperature points were
collected. Afterwards, the crotonaldehyde vapor was introduced to the
Ir in all catalysts were 3 wt.%. The catalysts were denoted as nMo-3Ir/
BN, with n referring to the weight percentage of Mo in the catalyst.
ꢀ
1
sample with a H
2
flow (20 mL min ). Then it was purged with a N
2
flow
ꢀ 1
2
.2. Catalyst characterizations
for 3 h, the gas was switched to H
2
(26 mL min ) and the sample was
◦
◦
ꢀ 1
heated from 30 to 80 C at a ramp of 2 C min , and the spectra were
collected with corresponding backgrounds at different temperatures.
The specific surface areas of the catalysts were measured by N
2
adsorption at 77 K on a Quantachrome Nova 4000e surface area
analyzer. X-ray diffraction (XRD) was performed on a Bruker D8
2.3. Catalytic reaction and kinetic study
a
ADVANCE powder X-ray diffractometer using Cu K radiation, which
was operated at 40 kV and 40 mA. The patterns were collected in a 2θ
The liquid phase crotonaldehyde hydrogenation was performed in a
100 mL stainless steel autoclave (custom-designed, Keli Chem ical
Equipment Co., China). Prior to the experiment, 100 mg of the catalyst
◦
◦
ꢀ 1
. Before the
range from 10 to 90 , with a scanning step of 0.02
s
measurements, the samples were pre-reduced in a H flow (99.99 %, 26
2
2

C. Tang et al.
Applied Catalysis A, General 623 (2021) 118269
2
was pretreated in a specially designed quartz tube under H (99.999 %,
ꢀ 1
◦
2
6 mL min ) at 300 C for 1 h. 14.5 mL of isopropanol was injected to
the tube under H flow and the catalyst was immersed by the solvent
2
until cooled down to room temperature. Then the suspension was
immediately transferred into the autoclave without exposure to air and
2
mixed with crotonaldehyde (6.0 mmol), followed by H recharge and
vent 11 times to remove the air in the autoclave. The reaction began
◦
when the temperature reached 80 C, while constant hydrogen (99.999
%
, kept at 0.6 MPa) gas was introduced to the reaction under stirring at
◦
8
0
C (800 rpm). The reaction was stopped after 3 h and the reaction
products and reactant were analyzed on line using a gas chromatog-
raphy (Shimadzu GC-2014) equipped with a DB-Wax capillary column
Fig. 1. XRD patterns of the pre-reduced Ir/BN catalysts.
(
30 m ×0.25 mm ×0.25
μ
m).
MoO
3
to MoO
2
[38], while the 3Ir/BN catalyst gives a reduction peak
The kinetic study was performed on two representative catalysts
3Ir/BN and 0.3Mo-3Ir/BN) on the same reactor as mentioned above.
◦
centered at 230 C due to the reduction of IrO
x
species. For the
(
x
MoO -promoted catalysts, two reduction peaks are observed. The peak
The effect of crotonaldehyde concentration was studied in the concen-
tration range from 2.4 to 6.0 mmol, while the effect of hydrogen partial
pressure was studied in the range from 0.4 to 1 MPa. In all cases, the
conversion levels were controlled below 15 % to ensure differential
mode reactions.
◦
centered at 230–290 C could be assigned to the reduction of IrO
the peak centered at ca. 450 C could be assigned to the reduction of
x
and
◦
x x
MoO . It is found that the low - temperature reduction peak of the IrO
◦
◦
(
230 C) slightly shifts to higher temperature (245–290 C) while the
◦
high-temperature reduction peak of the MoO
x
(560 C) slightly shifts to
◦
lower temperature (460 - 450 C) with increasing Mo content in the
3
. Results and discussion
catalyst. This observation implies that the IrO
more difficult reduction of IrO species but easier reduction of MoO
species due to the typical spillover effect.
x x
-MoO interaction lead to
x
x
3
.1. General characterizations
Fig. 4 shows the XPS spectra of the Ir and Mo species in the pre-
reduced catalysts. The Ir 4f 7/2 spectra (Fig. 4a) could be resolved to
two components at BEs of 61.2–61.4 and 62.3–62.6 eV, assigning to
2
ꢀ 1
The catalysts have similar surface areas (22.8–26.6 m g , Table 1).
The pre-reduced catalysts give typical diffractions of hexagonal BN (h-
BN, JCPDS No. 34-0421) and metallic Ir (JCPDS No. 46-1044), as shown
in Fig. 1. However, no distinct diffractions of Mo species could be
0
+
metallic Ir and cationic Irδ species, respectively [39]. It is found that
the BEs of the Ir species slightly shift to higher values by 0.1 - 0.3 eV after
observed, indicating the MoO
BN surface.
x
species are highly dispersed on the 3Ir/
the MoO
charge transfer also implies stronger Ir-MoO
agreement with the H -TPR results (Fig. 3). Similar observation was also
found in the Ir-FeO /BN catalyst [12]. For the Mo 3d spectra (Fig. 4b),
the catalyst with low Mo content (e.g. 0.1Mo-3Ir/BN) contains mostly
x
- promotion, indicating charge transfer from Ir to MoO
x
. Such
x
interaction, which is in
The morphological features of the catalysts were investigated by
HRTEM, and the results are shown in Fig. 2. Metallic Ir species are
confirmed in the pre-reduced 3Ir/BN, 0.1Mo-3Ir/BN and 0.3Mo-3Ir/BN
catalysts, with the measurement of lattice fringes featuring (111) plane
2
x
4+
6+
Mo although the signals are quite weak, but Mo species become
predominant when the Mo content is higher in the catalyst (e.g.
(
d Ir (111) =0.225 nm). For the MoO
ence of MoO is identified by the lattice fringe measurements (e.g. MoO
220) =0.233 nm in Fig. 2b and MoO (002) =0.211 nm in Fig. 2c).
Moreover, the Ir-MoO interaction is clearly observed in the 0.1Mo-3Ir/
BN and 0.3Mo-3Ir/BN catalysts, as evidenced by the contacted Ir and
MoO particles (Fig. 2b - c). Such interaction could be due to electro-
static interaction between the Irδ and the Mo
x
- containing catalysts, the pres-
x
3
3
Mo-3Ir/BN). Therefore, this observation suggests that the Ir-MoO
interaction facilitates the reduction of MoO to MoO due to the spill-
over effect, however, high Mo content leads to the formation of isolated
MoO entities, which are not easy to be reduced. Such spillover effect is
x
(
2
3
2
x
3
x
also confirmed by the Mo 3d XPS spectra of the 1Mo/BN reference
+
6ꢀ
24 during the prepa-
7
O
sample (Fig. S1). It is clearly that both fresh and the pre-reduced
ration process, in which Mo species will selectively deposit on Ir particle
6+
1
Mo/BN sample contain only Mo species, suggesting that the bare
to form the Ir-Mo interface [29]. The mean Ir particle sizes of the cat-
MoO
3
species are difficult to be reduced. Such comparison of the Mo 3d
interaction.
alysts are ca. 2–3 nm, which are smaller than those on SiO
20,28,36]). The smaller Ir particle size on the BN support is related to
the strong interaction between the metal atoms and defect sites in h-BN
e.g. boron vacancy and nitrogen edge) [37].
2
(e.g. 4–5 nm
XPS spectra again highlights the Ir-MoO
x
[
In situ CO-DRIFTS experiments were conducted to further probe the
electronic properties of the Ir species in the pre-reduFced catalysts. As
(
shown in Fig. 5, The 3Ir/BN catalyst shows overlapped bands centered at
Fig. 3 shows the H
2
-TPR profiles of different catalysts. The 3Mo/BN
ꢀ 1
2
049 and 2079 cm , assigning to linear adsorbed CO on metallic Ir
◦
catalyst gives a reduction peak centered at 560 C due to the reduction of
[
40–42]. The addition of MoO
x
in the catalyst results in a systematic red
shift of these bands, which can not be explained by the typical charge
enrichment of the Ir species as the XPS results (Fig. 4a) clearly suggest
the opposite. Instead, the red shift might be due to the changes of the Ir
particles. As the HRTEM images of the catalysts (Fig. 2) reveal that the Ir
particle sizes in the Mo-promoted catalysts are slightly smaller than that
in the un-promoted one, which results in higher surface electron density
on the smaller metal particles and consequently lower wave number of
Table 1
Physical properties of various Ir catalysts.
a
b
Catalyst
S
BET
2
/
ꢀ 1
d
Ir
/
D
Ir
/
Formation rate of CROL /
TOF /
ꢀ 7
-1 -1
ꢀ 1
m
g
nm
%
x 10 mol g
s
h
3
0
0
1
3
Ir/BN
.1Mo-
24.6
25.5
26.6
24.7
22.8
3.2 ±
.8
34.3
34.4
47.8
–
7.0
47.1
149.5
190.2
–
0
3.2 ±
0.5
22.3
39.4
36.1
17.9
x
the CO band [43]. On the other hand, in the MoO -promoted catalysts
3
Ir/BN
the surface Ir atoms are isolated by the MoO entities via Ir-MoO
x
x
.3Mo-
2.3 ±
0.5
interaction, which weakens the interactions between the adsorbed CO
molecules and thus results in a red shift [6]. Similar observation was also
3
Ir/BN
Mo-3Ir/
BN
–
reported in MoO
44] catalysts. Also, the band intensity gradually declines with
increasing Mo content in the catalyst, which qualitatively suggests less
exposed Ir surface atoms by MoO coverage. Such MoO -modification of
x x 2
-promoted Ir/TUD-1 [29] and NbO -promoted Ir/SiO
Mo-3Ir/
BN
3.1 ±
35.5
116.1
[
0.5
a
b
d
Ir = mean particle size of Ir (determined by TEM).
x
x
D
Ir (Ir dispersion) = 1.1/dIr
.
3

C. Tang et al.
Applied Catalysis A, General 623 (2021) 118269
Fig. 2. HRTEM images of 3Ir/BN, 0.1Mo-3Ir/BN and 0.3Mo-3Ir/BN catalysts.
Ir species becomes more significant on the 1Mo-3Ir/BN catalyst as it
shows no observable CO band, implying the surface Ir atoms are
the catalyst. Regarding the product selectivity, the 3Ir/BN catalyst give a
considerable selectivity to the desired crotyl alcohol (CROL, 76.6 %),
and the main by-products are butanal (BUAL), 1-butanol (BUOL) and
others such as 2,4,6-octatrienal. The addition of small amount of Mo
improves the selectivity up to ca. 88 %, and it seems that this high
selectivity is barely influenced by the Mo content in the catalyst.
Therefore, the formation rate of crotyl alcohol is improved by 5-fold on
x
dominantly covered by the MoO species. Regarding the CO desorption
behavior on the 0.3Mo-3Ir/BN at elevated temperatures (Fig. 5b), the
band intensity continuously declines and the band almost disappears at
◦
1
00 C. Meanwhile, a slight red shift is observed with increasing tem-
perature, due to the decrease in the coverage of the CO species as
observed on metallic Ir species [45].
ꢀ 7
-1 -1
the 0.3Mo-3Ir/BN (39.4 × 10 mol g s ) compared with that on the
ꢀ
7
-1 -1
3
Ir/BN (7.0 × 10 mol g s ), as shown in Table 1. Moreover, the
catalysts are quite stable during the recycling test. As shown in Fig. 6b,
the representative 0.3Mo-3Ir/BN catalyst does not show obvious loss of
activity and selectivity during 5 cycles. To further evaluate the intrinsic
activities of the two catalysts, turnover frequencies (TOF) of crotyl
alcohol formation based on Ir dispersions were calculated and summa-
3
.2. Catalytic performances of liquid- phase crotonaldehyde
hydrogenation
Fig. 6a shows the results of liquid-phase selective hydrogenation of
◦
crotonaldehyde over the catalysts. Under reaction conditions of 80 C,
pressure of 0.6 MPa and reaction time of 3 h, the crotonaldehyde
-
1
rized in Table 1. The 0.3Mo-3Ir/BN catalyst gives a TOF of (190.2 h ),
H
2
-
1
which is four times higher than that on the 3Ir/BN (47.1 h ). Our results
are also among the best results reported in literature (Table S1). Since
the isopropanol was used as the solvent, there could be possible
involvement of the hydrogen species from the isopropanol molecules as
conversions on the catalysts follow a volcano-type trend. The 3Ir/BN
catalyst gives a quite low conversion of 16.4 %, and the conversion
reaches a maximum (80.4 %) at the Mo content of 0.3 wt.% (0.3Mo-Ir/
BN). However, the conversion declines with increasing Mo content in
4

C. Tang et al.
Applied Catalysis A, General 623 (2021) 118269
recently reported by Ye et al. [46] via a catalytic transfer hydrogenation
(
CTH) process. However, in our case, we did not detect the presence of
acetone in the products. Therefore we believe that the CTH process may
not exist in the current case due to different catalyst systems.
3
.3. FTIR investigation on crotonaldehyde adsorption and in situ reaction
In order to investigate the adsorption properties and reaction be-
haviors of crotonaldehyde, in situ FTIR spectroscopic study was con-
ducted on some representative catalysts. We focus on the vibrations of
–
–
ꢀ 1
C
–
O and C
–
C bonds, which are in 1500 - 1800 cm region (Fig. 7a).
ꢀ
1
Gaseous crotonaldehyde gives two bands at 1712 and 1724 cm due to
–
v(C
ꢀ 1
–
–
C). The adsorption
–
O), and a weak band at 1650 cm due to v(C
of crotonaldehyde on the catalysts result in new bands at 1684, 1676
and 1643 cm-1 (Fig. S2). The bands at 1684 and 1676 cm are assigned
ꢀ 1
–
–
to v(C
tively, while the band at 1643 cm is assigned to v(C
43]. These results indicate that crotonaldehyde molecules dominantly
adsorb on the catalysts via carbonyl group. Morever, the bands gradu-
ally decline in intensity when the catalysts are purged by N at elevated
temperatures (from 353 to 423 K). The changes in the band area in the
–
O) in a
π
CO mode and v(C
–
O) in a di-
σ
CO mode [6], respec-
ꢀ 1
–
–
C) in a
πCC mode
[
2
ꢀ 1
range 1550 - 1750 cm on the three catalysts were plotted as a function
of temperature versus area percent of the corresponding band at 353 K
(
Fig. 7b). The band area on 0.3Mo/BN and 3Ir/BN decreased more
rapidly (a larger negative slope) in comparison to that on the
0
0
.3Mo-3Ir/BN. Considering the presence of Ir-MoO
x
interface in the
.3Mo-3Ir/BN catalyst, it could be concluded that the adsorption
strength of crotonaldehyde molecules adsorbed on the interfacial sites
are stronger than those on Ir and MoO sites. Similar observation was
catalyst by Tamura et al. [27].
x
also reported on Ir-ReO
x
/SiO
2
–
–
–
The changes of v(C
crotonaldehyde + H
–
O) and v(C
C) under reacton conditions
(
2
) were also investigated by in situ FTIR spectros-
copy, as shown in Fig. 7c. For the 3Ir/BN catalyst, the intensities of the v
–
ꢀ 1
(
C
–
O) bands at 1676 and 1684 cm decreases slowly with increasing
◦
temperature, but remains substantial at 80 C. In contrast, the desorp-
2
Fig. 3. H -TPR profiles of fresh catalysts.
tion rate of these bands are much faster on the 0.3Mo-3Ir/BN and 3Mo-
3
Ir/BN catalysts, as shown in Fig. 7d. It thus suggests that the reaction
Fig. 4. a) Ir 4f and b) Mo 3d XPS spectra of pre-reduced Ir catalysts.
5

C. Tang et al.
Applied Catalysis A, General 623 (2021) 118269
Fig. 5. DRIFT spectra of a) CO adsorption on species on pre-reduced catalysts at 303 K and b) CO desorption on pre-reduced 0.3Mo-3Ir/BN catalyst at elevated
temperatures.
Fig. 6. a) Reaction results of liquid phase hydrogenation of crotonaldehyde over various catalysts and b) recycling test of 0.3Mo-3Ir/BN catalyst. Reaction con-
ditions: 6.0 mmol crotonaldehyde, 0.6 MPa H
2
, 14.5 mL isopropanol, 100 mg catalyst, T =80 ◦C, Time =3 h.
rates on the MoO
x
-promoted catalysts are higher than that on the bare
3.4. Kinetic investigation
3
Ir/BN, which is in agreement with the overall activity (Fig. 6). Note
that the disappearance of these bands at elevated temperatures is not
due to the desorption of crotonaldehyde from the catalyst surface, as it is
Although the FTIR results in gas phase offered useful information on
the reaction behaviors of the catalysts, there might be deviation because
the reaction was conducted in liquid phase. Therefore, kinetic experi-
ments were conducted on the 3Ir/BN and 0.3Mo-3Ir/BN catalysts to
explore their intrinsic behaviors in liquid phase selective hydrogenation
of crotonaldehyde. As shown in Table S2, the reaction rate increases
◦
clear that substantial bands are observed at 80 C when the catalyst is
–
–
exposed to crotonaldehyde (Fig. 7a). Also, diminish of the v(C
bands is accompanied by the emergence of bands at 2879, 2924, 2940,
O)
ꢀ
1
2
963, 3027 and 3320 cm (Fig. S3). The first five bands are assigned to
ꢀ 1
aliphatic or olefinic v(C
–
H) bands [18], and the band at 3320 cm is
with increasing partial pressure of H
2
, while it decreases with increasing
and croto-
assigned to surface hydroxyl group. The observation of these bands
implies unsaturated alcohol is formed on the catalyst surface. These
concentration of crotonaldehyde. Reaction orders of H
2
naldehyde and apparent reaction constant were evaluated by plotting
logarithm of reaction rate versus logarithm of the partial pressure of
hydrogen and the content of crotonaldehyde (Fig. 8). For the formation
x
findings imply that on the MoO -promoted 3Ir/BN catalysts the hydro-
–
genation of C
–
O bond is fast and dominant, although hydrogenation of
–
–
◦
ꢀ 6
C
–
C bond also occurs as the v(C
–
C) band disappears at 80 C. The very
of crotyl alcohol, the power law rate expression is r = 3.26 × 10
–
ꢀ 1
-1.1
2.2
fast disappearance of the v(C
–
O) band at 1676 and 1684 cm with
[CRAL]
[H
2
]
(CRAL : crotonaldehyde) on the 3Ir/BN catalyst;
-
6
-0.6
1.4
increasing temperature indicates that 0.3Mo-3Ir/BN is more active than
the 3Mo-3Ir/BN. Therefore, the in situ FTIR results on the three catalysts
are in good agreement with the observed overall catalytic behaviors of
the catalysts (Fig. 6), that is, all catalysts show high selectivity and the
while it is r = 5.12 × 10 [CRAL]
[H
2
]
on the 0.3Mo-3Ir/BN
catalyst. The 0.3Mo-3Ir/BN catalyst has a higher apparent rate con-
stant than the 3Ir/BN, suggesting that it is intrinsically more active,
which is in line with the overall catalytic performance as shown in Fig. 6.
Moreover, the negative reaction orders of crotonaldehyde on both cat-
alysts suggest that the presence of crotonaldehyde prohibits the reaction
0
3
.3Mo-3Ir/BN possesses the highest activity while the 3Ir/BN and
Mo-3Ir/BN are less active.
6

C. Tang et al.
Applied Catalysis A, General 623 (2021) 118269
Fig. 7. a) Changes in FTIR spectra of crotonaldehyde adspecies on 0.3Mo/BN, 3Ir/BN, 0.3Mo-3Ir/BN at 353ꢀ 423 K. b) Band area changes of crotonaldehyde
adspecies on various catalysts. The band area (%) was normalized by the area at 353 K. c) In situ FTIR spectra of hydrogenation of crotonaldehyde over 3Ir/BN,
0
2
.3Mo-3Ir/BN and 3Mo-3Ir/BN at 273–353 K d) Band area change of crotonaldehyde adspecies on various catalysts. The band area (%) was normalized by the area at
73 K.
rate. In contrast, the reaction orders of H
2
on both catalysts are very high
the crotonaldehyde molecules could be easily dissolved in the solvent,
and thus readily adsorb on the catalyst surface. Therefore, the high
surface coverage of crotonaldehyde hinders the surface adsorption of
(
1.4–2.2), which can not be explained the dissociative adsorption of H
2
because otherwise the reaction order of H
2
would not exceed 0.5. The
kinetic investigations of gas phase selective hydrogenation of croto-
naldehyde over Ir catalysts [44] revealed that high reaction order of H
H
2
, which results in negative reaction order of crotonaldehyde. Tamura
et al. [27] also reported liquid phase hydrogenation of crotonaldehyde
over an Ir-ReO /SiO catalyst and found a zero order of crotonaldehyde
and 0.9 order of H indicating saturated coverage of crotonaldehyde on
the catalyst surface and difficult adsorption of H . This scenario in liquid
phase is different from the observations in gas phase. For example, the
kinetic study on Ir/BN and Ir-FeO /BN catalysts for gas phase hydro-
2
(
e.g. 0.6 - 0.9) is related to heterolytical dissociation of H
2
on the metal
x
2
+
surface (H
2
→ H + H-). However, it should be noted that in liquid phase
2
reaction, the solubility of H
observed kinetics. Indeed, our results are similar to those obtained on
Au/Mg AlO catalysts reported by Chen et al. [33] in which a negative
reaction order of crotonaldehyde (-0.55) and a positive reaction order of
(1.84) were reported. The authors attributed the high reaction order
of H to its low solubility in polar solvent (e.g. isopropanol). In contrast,
2
in the solvent is crucial to affect the
2
2
x
genation of crotonaldehyde reveals crotonaldehyde reaction orders of
0.1 - 0.4 [12], due to the competitive adsorption of gas phase croto-
H
2
2
2
naldehyde and H on the catalyst surface. In addition, comparison of the
7

C. Tang et al.
Applied Catalysis A, General 623 (2021) 118269
Fig. 8. Dependency of a) crotonaldehyde content and b) hydrogen pressure on formation rates of crotyl alcohol; c) Arrhenius plots of the formation of crotyl alcohol
over 3Ir/BN and 0.3Mo-3Ir/BN catalysts.
rate expressions of the 3Ir/BN and 0.3Mo-3Ir/BN catalysts leads to some
remarks. Regarding the reaction order of crotonaldehyde, the 3Ir/BN
gives a more negative value (-1.1) than the 0.3Mo-3Ir/BN (-0.6),
implying that the prohibiting effect of crotonaldehyde on the
comparison of the activation energies (Eas) of the two catalysts. As
-
1
shown in Fig. 8c, the Ea on the 0.3Mo-3Ir/BN (18.9 ± 4.3 kJ mol ) is
-
1
only half of that on the 3Ir/BN (38.4 ± 2.2 kJ mol ), strongly suggesting
a more facile reaction route on the 0.3Mo-3Ir/BN.
0
.3Mo-3Ir/BN is somehow alleviated compared to that on the 3Ir/BN.
Meanwhile, the reaction order of H on the 0.3Mo-3Ir/BN (1.4) is
significantly lower than that on the 3Ir/BN (2.2), indicating that the
reaction is less dependent on the partial pressure of H on the former (or
relatively easier adsorption of H on the catalyst surface). This is coin-
cident with weakened prohibiting effect of crotonaldehyde on the
.3Mo-3Ir/BN, suggesting that the adsorption geometry of crotonalde-
Based on the above findings, the reaction behaviors of the catalysts in
the liquid phase are illustrated in Fig. 9. First of all, we would like to
explicate the roles of catalyst components in the reaction. Since CRAL
can adsorb on the BN support (Fig. S1), the BN may participate in the
reaction because hydrogen species dissociated on metallic Ir species
could migrate and react with adjacent CRAL molecules adsorbed on BN.
Although the roles of BN could not be ignored, its contribution in the
reaction is quite limited because otherwise the reaction behaviors of 3Ir/
BN and the MoOx-containing catalysts (e.g. 0.3Mo-3Ir/BN) should not
have differed so much as they have similar CRAL adsorption capacity
(Fig. S1). Therefore, we believe that the reaction mainly take place on
the Ir species. Accordingly, the metallic Ir species are responsible for the
2
2
2
0
hyde on the 0.3Mo-3Ir/BN is rather different to that on the 3Ir/BN.
Indeed, the adsorption behaviors of crotonaldehyde on supported noble
metal catalysts have been extensively investigated [25,28]. On bare
metal (such as Pt and Ir), the adsorption of crotonaldehyde and H
place on the metal surface, and thus may result in a competition.
However, when metal is supported on reducible oxides (such as TiO and
Co ) or is promoted with other components (such as FeO , MoO ), the
adsorption of crotonaldehyde tends to occur at the metal-support or
metal-promoter interface while the adsorption of H occurs on the metal
surface, therefore, the adsorption of the two species are non-competitive
10]. Such changes in the adsorption patterns may explain the different
behaviors of the 3Ir/BN and 0.3Mo-3Ir/BN catalysts, that is, the Ir-MoO
interface provides new sites for crotonaldehyde adsorption, thus alle-
viates the competition adsorption of H and relatively lower reaction
order of H . The generation of such Ir-MoO interfacial sites also implies
that reaction pathway may be altered, which is justified by the
2
take
2
adsorption/activation/dissociation of H
2
molecules while the cationic Ir
3
O
4
x
x
species are responsible for the adsorption/activation of CRAL molecules
–
O bond) [12]. When the Ir catalysts were promoted
(particularly the C
by FeO , NbO and ReOx, the promoters act as additional adsorptio-
n/activation sites for the C
–
2
x
x
–
–
O bond [12,27,44], which leads to much
[
higher activity for the reaction. Secondly, we would like to highlight the
different reaction behaviors of the catalysts in the liquid phase. In our
previous work, we have investigated gas phase selective hydrogenation
x
2
x x
over FeO -promoted Ir/BN catalysts [12]. The presence of Ir-FeO
2
x
interface improves the catalytic performance, which highlights the
critical role of Ir-promoter interface in the reaction. However, different
Fig. 9. The reaction behaviors of 3Ir/BN, 0.3Mo-3Ir/BN and 3Mo-3Ir/BN catalysts in the liquid phase.
8

C. Tang et al.
Applied Catalysis A, General 623 (2021) 118269
behaviors in gas phase and liquid phase should be noted. Unlike in the
gas phase reaction, the rate of liquid phase reaction is controlled by the
Fund of Key Laboratory of the Ministry of Education for Advanced
Catalysis Materials and Zhejiang Key Laboratory for Reactive Chemistry
on Solid Surfaces, Zhejiang Normal University (KLMEACM201910).
2
H solubility in the solvent, while the adsorption of crotonaldehyde on
the catalyst surface readily takes place. On the bare 3Ir/BN catalyst,
adsorption of crotonaldehyde with C–
–
O bonds results in high selec-
References
tivity, but its high coverage also hinders the adsorption/dissociation of
on the Ir surface and thus low activity. On the 0.3Mo-3Ir/BN, the
crotonaldehyde may preferably adsorb on the Ir-MoO interfacial sites,
and thus easy adsorption/dissociation of H on the Ir surface, which
accounts for the much improved activity and selectivity. However, it
should be noted that at high content of MoO (e.g. 3Mo-3Ir/BN) such
promotion may lead to less exposed Ir atoms due to excessive coverage
of Ir by MoO species, which makes H adsorption/activation difficult
and thus leads to lowered activity compared to that on the
[
[
[
1] L.A. Saudan, Acc. Chem. Res. 40 (2007) 1309–1319.
2] W. Chen, A.M. Viljoen, South African J. Bot. 76 (2010) 643–651.
3] M.A. Vannice, J. Mol. Catal. 59 (1990) 165–177.
H
2
x
2
[4] P. M a¨ ki-Arvela, J. H ´a jek, T. Salmi, D.Y. Murzin, Appl. Catal. A Gen. 292 (2005)
1
–49.
5] Y. Tan, X. Liu, L. Zhang, F. Liu, A. Wang, T. Zhang, Chin. J. Catal. 42 (2020)
70–481.
[
x
4
[6] A. Dandekar, M.A. Vannice, J. Catal. 183 (1999) 344–354.
[
[
7] F. Delbecq, Y. Li, D. Loffreda, J. Catal. 334 (2016) 68–78.
8] A.P. Jia, Y.S. Zhang, T.Y. Song, Z.H. Zhang, C. Tang, Y.M. Hu, W.B. Zheng, M.
F. Luo, J.Q. Lu, W.X. Huang, J. Catal. 395 (2021) 10–22.
x
2
0
.3Mo-3Ir/BN. Generally, previous works on Ir-MO
7,29,36,44] for selective hydrogenation of CRAL mainly focus on the
interface, which is indeed
critical in the reaction process. However, less attention has been paid on
the activation of H . In the current work, it is found that H adsorptio-
n/activation could also be affected by the Ir-MO interface, which thus
x
systems [12,18,19,
[9] S. Bai, L. Bu, Q. Shao, X. Zhu, X. Huang, J. Am. Chem. Soc. 140 (2018) 8384–8387.
2
[10] N. Mahata, F. Gonçalves, M.F.R. Pereira, J.L. Figueiredo, Appl. Catal. A Gen. 339
(
2008) 159–168.
adsorption/activation of CRAL on the Ir-MO
x
[
11] S.C. Tsang, N. Cailuo, W. Oduro, A.T.S. Kong, L. Clifton, K.M.K. Yu, B. Thiebaut,
J. Cookson, P. Bishop, ACS Nano. 2 (2008) 2547–2553.
[12] X. Ning, Y.M. Xu, A.Q. Wu, C. Tang, A.P. Jia, M.F. Luo, J.Q. Lu, Appl. Surf. Sci. 463
2
2
(
2019) 463–473.
13] W. Lin, H. Cheng, X. Li, C. Zhang, F. Zhao, M. Arai, Chin. J. Catal. 39 (2018)
88–996.
x
[
provides some new information on the better understanding of the
reaction.
9
[14] M. Aoun, A. Benamar, M. Chater, Chin. J. Catal. 32 (2011) 1185–1190.
[
[
15] J. Shi, M. Zhang, W. Du, W. Ning, Z. Hou, Catal. Sci. Technol. 5 (2015) 3108–3112.
16] H. Guo, H. Li, K. Jarvis, H. Wan, P. Kunal, S.G. Dunning, Y. Liu, G. Henkelman, S.
M. Humphrey, ACS Catal. 8 (2018) 11386–11397.
4
. Conclusions
[
[
17] M. Luneau, J.S. Lim, D.A. Patel, E.C.H. Sykes, C.M. Friend, P. Sautet, Chem. Rev.
1
20 (2020) 12834–12872.
x
In summary, MoO - promotion could greatly improve the perfor-
18] M. Tamura, K. Tokonami, Y. Nakagawa, K. Tomishige, Chem. Commun. 49 (2013)
mance of Ir/BN catalysts for liquid phase hydrogenation of croto-
naldehyde due to the formation of Ir-MoO interface, on which the
7
034–7036.
x
[19] Q. Yu, K.K. Bando, J.F. Yuan, C.Q. Luo, A.P. Jia, G.S. Hu, J.Q. Lu, M.F. Luo, J. Phys.
Chem. C. 120 (2016) 8663–8673.
adsorption/activation of crotonaldehyde is facilitated. The kinetic re-
sults suggest that the reaction rate is highly dependent on the partial
[
20] P. Chen, J.Q. Lu, G.Q. Xie, G.S. Hu, L. Zhu, L.F. Luo, W.X. Huang, M.F. Luo, Appl.
Catal. A Gen. 433 (2012) 236–242.
pressure of H
tion/activation of H
mining step of the reaction. On the contrary, high surface coverage of
crotonaldehyde prohibits the adsorption of H and thus results in a
negative reaction order of crotonaldehyde. However, the crotonalde-
hyde molecules preferably adsorbed at the Ir-MoO interface makes the
adsorption easier, which accounts for the higher activity. These
2
due to its low solubility in the solvent, thus the adsorp-
[21] P. Reyes, M.C. Aguirre, G. Pecchi, J.L.G. Fierro, J. Mol, Catal. A Chem. 164 (2000)
2
on the surface Ir atoms may be the rate - deter-
245–251.
[
[
22] X. Hong, B. Li, Y.J. Wang, J.Q. Lu, G.S. Hu, M.F. Luo, Appl. Surf. Sci. 270 (2013)
3
88–394.
2
23] S. He, L. Xie, M. Che, H.C. Chan, L. Yang, Z. Shi, Y. Tang, Q. Gao, J. Mol. Catal. A
Chem 425 (2016) 248–254.
[
[
24] Y.A. Ryndin, C.C. Santini, D. Prat, J.M. Basset, J. Catal. 190 (2000) 364–373.
25] G. Kennedy, L.R. Baker, G.A. Somorjai, Angew. Chem. Int. Ed. 126 (2014)
x
H
2
3
473–3476.
[26] R. Nie, M. Miao, W. Du, J. Shi, Y. Liu, Z. Hou, Appl. Catal. B Environ. 180 (2016)
07–613.
27] M. Tamura, K. Tokonami, Y. Nakagawa, K. Tomishige, ACS Catal. 6 (2016)
600–3609.
findings in liquid phase reaction again highlight the roles of metal-
promoter interface in selective hydrogenation of crotonaldehyde, but
also reflect different features compared to the behaviors in gas phase
6
[
3
reaction. In liquid phase reaction, the adsorption and activation of H
2
[28] M. Englisch, A. Jentys, J.A. Lercher, J. Catal. 166 (1997) 25–35.
[
29] Y.M. Xu, W. Bin Zheng, Y.M. Hu, C. Tang, A.P. Jia, J.Q. Lu, J. Catal. 381 (2020)
becomes the limiting step because of its low solubility in solvent and
high surface coverage of crotonaldehyde on the catalyst.
2
22–233.
[
[
[
[
30] L. Xu, R. Nie, X. Lyu, J. Wang, X. Lu, Fuel Process. Technol. 197 (2020), 106205.
31] P. Reyes, H. Rojas, J.L.G. Fierro, Appl. Catal. A Gen. 248 (2003) 59–65.
32] W. Koo-Amornpattana, J.M. Winterbottom, Catal. Today. 66 (2001) 277–287.
33] H.Y. Chen, C.T. Chang, S.J. Chiang, B.J. Liaw, Y.Z. Chen, Appl. Catal. A Gen. 381
CRediT authorship contribution statement
(
2010) 209–215.
Cen Tang: Data curation, Formal analysis, Investigation, Writing -
original draft. Fangru Zhou: Investigation. Yang Wen: Investigation.
Yumeng Xu: Investigation, Formal analysis. Aiping Jia: Investigation.
Mengfei Luo: Conceptualization, Methodology, Resources. Jiqing Lu:
Conceptualization, Methodology, Resources, Supervision, Writing - re-
view & editing.
[34] U.K. Singh, M. Albert Vannice, J. Catal. 191 (2000) 165–180.
[35] U.K. Singh, M.A. Vannice, Appl. Catal. A Gen. 213 (2001) 1–24.
[
36] J.F. Yuan, C.Q. Luo, Q. Yu, A.P. Jia, G.S. Hu, J.Q. Lu, M.F. Luo, Catal. Sci. Technol.
6
(2016) 4294–4305.
[
37] S. Lin, X. Ye, R.S. Johnson, H. Guo, J. Phys. Chem. C. 117 (2013) 17319–17326.
[38] J. Vakros, K. Bourikas, Ch. Kordulis, A. Lycourghiotis, J. Phys. Chem. B. 107 (2003)
1
804–1813.
[
39] P. Reyes, M.C. Aguirre, I. Meli a´ n-Cabrera, M. L o´ pez Granados, J.L.G. Fierro,
J. Catal. 208 (2002) 229–237.
Declaration of Competing Interest
The authors report no declarations of interest.
Acknowledgements
[40] F. Solymosi, J. Rask o´ , J. Catal. 62 (1980) 253–263.
´
´
´
[
[
[
[
41] F. Solymosi, E. Novak, A. Molnar, J. Phys. Chem. 94 (1990) 7250–7255.
42] M. Haneda, T. Fujitani, H. Hamada, J. Mol. Catal. A Chem. 256 (2006) 143–148.
43] A. Bourane, O. Dulaurent, D. Bianchi, J. Catal. 195 (2000) 406–411.
44] M. Tamura, K. Tokonami, Y. Nakagawa, K. Tomishige, ACS Sustain. Chem. Eng. 5
(
2017) 3685–3697.
[
[
45] A. Bourance, M. Nawdali, D. Bianchi, J. Am. Chem. Soc. 106 (2002) 2665–2671.
46] H. Ye, H. Zhao, Y. Jiang, H. Liu, Z. Hou, ACS Appl. Nano Mater. 3 (2020)
12260–12268.
This work is financially supported by the National Natural Science
Foundation of China (Grant No. 21773212, 22072137), Open Research
9