Facile synthesis of CuMAl (M = Cr, Mn, Zn, and Co) with highly dispersed Cu and tailorable surface acidity for efficient 2-methylpyrazine synthesis

Article abstract of DOI:10.1039/c7ra08349a

Synthesis of 2-methylpyrazine (2-MP) from 1,2-propylene glycol (PG) and ethylenediamine (ED) was investigated in the presence of multifunctional catalytic systems (CuMAl) possessing acidic and metallic functional sites. Catalytic systems were prepared from mixed CuMAl-layered double hydroxides (CuMAl-LDHs, M = Cr, Mn, Zn, and Co) via their thermal decomposition. CuMAl-LDH were prepared from Cu(NO₃)₂, M(NO₃)_x and Al(NO₃)₃ and NaOH/Na₂CO₃ as a precipitating agent. X-ray diffraction (XRD), N₂ adsorption-desorption, temperature-programmed desorption with ammonia (NH₃-TPD), N₂O chemisorption, transmission electron microscopy (TEM), and high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) were used to characterize the physical and chemical properties of the catalysts. The results showed that the nature of the secondary metal M inserted into the LDH structure significantly affected the crystalline structure, the dispersion of copper nanoparticles, and the density of surface acidic sites of the catalysts. The as-prepared CuMAl catalysts displayed promising catalytic performances towards the synthesis of 2-MP. Among them, CuCoAl showed the highest PG conversion (97%) and 2-MP selectivity (55%). These high catalytic activities were found to be associated with the ultra-small Cu nanoparticles (～2 nm) and high surface acidity (2433 μmol g^-1).

Full text of DOI:10.1039/c7ra08349a

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Facile synthesis of CuMAl (M ¼ Cr, Mn, Zn, and Co)
with highly dispersed Cu and tailorable surface
acidity for efficient 2-methylpyrazine synthesis†
Cite this: RSC Adv., 2017, 7, 48662
Jindou Hou,‡ Wen. Luo,‡ Shizhong Luo, Chao Lin, Ping Liu, Xuemei Liao, *ab
a
b
b
c
a
b
c
FangLi Jing* and Xiaopeng Li*
Synthesis of 2-methylpyrazine (2-MP) from 1,2-propylene glycol (PG) and ethylenediamine (ED) was
investigated in the presence of multifunctional catalytic systems (CuMAl) possessing acidic and metallic
functional sites. Catalytic systems were prepared from mixed CuMAl-layered double hydroxides (CuMAl-
LDHs, M ¼ Cr, Mn, Zn, and Co) via their thermal decomposition. CuMAl-LDH were prepared from
Cu(NO
3
)
2
, M(NO
3
)
x
and Al(NO
3
)
3
and NaOH/Na
2
CO
3
as a precipitating agent. X-ray diffraction (XRD),
2
-TPD), N O
N
2
adsorption–desorption, temperature-programmed desorption with ammonia (NH
3
chemisorption, transmission electron microscopy (TEM), and high-angle annular dark-field scanning
transmission electron microscopy (HAADF-STEM) were used to characterize the physical and chemical
properties of the catalysts. The results showed that the nature of the secondary metal M inserted into
the LDH structure significantly affected the crystalline structure, the dispersion of copper nanoparticles,
and the density of surface acidic sites of the catalysts. The as-prepared CuMAl catalysts displayed
promising catalytic performances towards the synthesis of 2-MP. Among them, CuCoAl showed the
highest PG conversion (97%) and 2-MP selectivity (55%). These high catalytic activities were found to be
associated with the ultra-small Cu nanoparticles (ꢀ2 nm) and high surface acidity (2433 mmol gꢁ1).
Received 28th July 2017
Accepted 9th October 2017
DOI: 10.1039/c7ra08349a
rsc.li/rsc-advances
intermediate (Scheme 1), the well-designed catalysts should
possess dual (i.e., acidic and metallic) functional active sites.
1
. Introduction
Conversion of biomass into biodiesel has gained importance as Acidic sites are required to catalyze-cyclization/dehydration of
an alternative way to utilize renewable biological sources. ED and PG, while metallic sites (e.g., copper, nickel, or cobalt,
However, the byproduct of this process (i.e., glycerol) accounts among others) allow dehydrogenation of the piperazine inter-
for around 10 wt% of the total bio-diesel production. Since it mediate to 2-MP, with the metal particle size playing a major
has no practical use, bio-glycerol is accumulated, thus role in this latter transformation. Therefore, the synthesis of 2-
increasing the storage or waste costs of biodiesel companies. MP can be enhanced upon increasing the number of acidic
Therefore, the conversion of bio-glycerol into value-added active sites and decreasing the metal particle size via rational
chemicals has received great interest in recent years. 2-Meth- design of a bifunctional catalyst.
2
,3
ylpyrazine (2-MP) is an intermediate compound in the synthesis
Up to now, various catalysts such as Cu–Zn–Si–O, Cu–Zn–
4
5
8
of 2-aminopyrazine, a well-known bacteriostatic and antitu- Al–O, Cu–Zn–Cr–Al–O and Cu/Co/Ni–Al–O have been
bercular drug. Up to now, 2-MP is mostly synthesized via prepared and used in the synthesis of 2-MP. These Cu-based
cyclization of ethylenediamine (ED) and 1,2-propylene glycol materials containing alumina as an acidic material were
1
(
(
PG). Since the catalytic synthesis of 2-MP involves cyclization found to be efficient catalysts. However, these catalysts were
i.e., dehydration) and dehydrogenation of
a
piperazine mostly prepared by conventional methods such as impregna-
tion, co-precipitation and sol–gel techniques, which typically
a
College of Food and Biological Engineering, Xihua University, Chengdu, China.
E-mail: xmliao@mail.xhu.edu.cn
b
Department of Chemical Engineering, Sichuan University, Chengdu, China. E-mail:
fangli.jing@scu.edu.cn
c
CAS Key Laboratory of Low-Carbon Conversion Science and Engineering, Shanghai
Advanced Research Institute, Chinese Academy of Sciences, Shanghai, 201210,
China. E-mail: lixp@sari.ac.cn
†
1
‡
Electronic supplementary information (ESI) available. See DOI:
0.1039/c7ra08349a
Scheme 1 Reaction pathway for the synthesis of 2-MP from ED and
PG.
5
Both authors contributed equally.
48662 | RSC Adv., 2017, 7, 48662–48669
This journal is © The Royal Society of Chemistry 2017

View Article Online
Paper
RSC Advances
resulted in inhomogeneous agglomeration of copper species, with energy-dispersive X-ray spectroscopy (EDS). The results in
particularly in the case of the impregnation approaches. Table S1† showed that the actual Cu loading on CuAl and
Usually, increasing the metal loading may result in the forma- CuMAl catalysts was 43% and 21 ꢃ 2%, respectively; while the
7
tion of large copper particles and thus poor catalytic activity.
loading amount of M (Co, Zn, Cr, and Mn) was approximately
Thermal decomposition of layered double hydroxides 20%. In general, the actual loading of Cu and M are adjacent to
(
LDHs) is an attractive approach to obtain uniformly distributed their respectively theoretical values. The specic surface area,
7
–11
metal nanoparticles. Venugopal et al. prepared LDHs-derived total pore volume, and average pore diameter were determined
Cu/Zn–Cr–O catalysts and used them in the synthesis of 2,6- by N
dimethylpyrazine. However, the low surface area (50 m g
ꢂ
2
adsorption–desorption isotherms at ꢁ196 C, using
2
ꢁ1
)
a Quantachrome NOVA 1000e apparatus. Before each
and the large CuOx crystalline size (5.8 nm) of the catalyst measurement, the sample was degassed under vacuum at
ꢂ
9
ꢂ
resulted in a poor glycerol conversion (9%) at 375 C. The 300 C for 3 h. TPR experiments were carried out in a xed-bed
formation of the LDH precursor can be signicantly altered by reactor. 50 mg of catalysts were reduced under owing a 5% H /
2
ꢁ1
ꢁ
1
ꢂ
ꢂ
changing the secondary metal element (M) in CuMAl catalysts,
N (30 mL min ) from 100 to 900 C at a rate of 10 C min .
2
thereby potentially leading to higher metallic Cu dispersion. The hydrogen consumption was analyzed on-line by a SC-200
However, to date, a systematical study of the impact of M gas chromatograph provided with a thermal conductivity
species on the dispersion of Cu in CuMAl catalysts for 2-MP detector (TCD). The surface acidity and acid strength of the
synthesis is lacking in the literature. Moreover, strategies catalysts were determined by NH -TPD. All the samples were
3
allowing simultaneous control of the surface acidity and active pre-reduced before evaluating their acidity. 100 mg of the
metal dispersion in these catalysts have not been developed yet. catalysts were introduced into the reactor and the temperature
ꢂ
ꢂ
ꢁ1
In this work, we investigated a series of M doped (M ¼ Cr, Mn, was subsequently increased to 400 C at 10 C min under an
ꢁ
1
Zn and Co) CuMAl catalysts derived from LDHs. The catalysts Ar ow (30 mL min ). Finally, the catalysts were reduced in situ
ꢁ
1
ꢂ
were characterized by a variety of techniques including N
adsorption–desorption, X-ray diffraction (XRD), H -temperature- subsequently cooled to room temperature (RT) under a ow of
programmed reduction (H -TPR), NH
programmed desorption (NH
2
under a stream of pure H (15 mL min ) at 400 C for 1 h and
2
2
ꢁ
1
2
3
-temperature- Ar (30 mL min ) to remove the adsorbed H
2
. A stream of NH
3
ꢁ
1
ꢁ1
3
-TPD), transmission electron (25 mL min ) was subsequently, a ow of Ar (25 mL min ) was
microscopy (TEM), high-angle annular dark-eld imaging introduced to remove the NH3 physically adsorbed on the
(
(
HAADF) and scanning transmission electron microscopy catalyst surface. When the signals were stabilized, the temper-
ꢂ
ꢂ
ꢁ1
STEM). The effects of the M species on the Cu dispersion and ature was increased from 100 to 700 C at a rate of 10 C min
.
surface acidity of the CuMAl catalysts and their catalytic perfor- Integration of the TPD curves was used to quantify the amount
mance towards the 2-MP synthesis will be unraveled. of desorbed NH . Note that this method is valid for obtaining
3
the total acidity of the samples and it does not allow to obtain
the exact nature of the adsorption sites (e.g. Brønsted or Lewis).
The dispersion of Cu metal was measured by using dissociative
2. Experimental
2.1 Catalyst preparation
N O chemisorption. The experiment was carried out in a TPR
2
setup, using the sample mass 100 mg. The amount of the
consumed hydrogen in the rst prereduction reduction was
denoted as X. Subsequently, catalyst beds were purged with He
CuMAl LDH precursors (M ¼ Co, Zn, Cr, and Mn) were prepared
7
2 3
by a conventional Na CO /NaOH co-precipitation method.
Typically, 100 mL of a mixed metal solution, containing 1 M of
cations (Cu(NO $6H O, M(NO $yH O, and Al(NO $9H O),
ꢁ1
ꢂ
(
30 mL min ) and cooled to 50 C. Following this 10% N O/He
2
3
)
2
2
3
)
x
2
3
)
3
2
ꢁ
1
with the ow rate of 50 mL min was introduced to the catalyst
bed at 50 C for 10 min to perform the dissociative adsorption
with the desired molar ratio and 100 mL of a mixed basic
ꢂ
ꢁ
solution consisting of NaOH and Na CO with [OH ] ¼
2
3
ꢁ1
ꢁ
2ꢁ
and corresponding oxidation of Cu surface to Cu O. In order to
2
2
.0 mol L and [OH ]/[CO ] ¼ 2 were simultaneously added
3
remove excess oxidant, the samples were again ushed with
dropwise to 100 mL of distilled water under vigorous mechan-
ꢁ
1
a highly pure He (30 mL min ) and cooled to the temperature
ical stirring. The mixed solution was kept at constant pH (10 ꢃ
ꢂ
ꢂ
at 50 C under the same atmosphere. As the signal baseline
0
.5). The resulting suspension was aged at 70 C for 12 h. The
stabilized, the second temperature programmed reduction was
performed on the samples. The amount of the consumed
hydrogen in the second temperature programmed reduction
was denoted as Y. Hence, the dispersion of Cu is calculated as
resulting solid were obtained thoroughly washed with distilled
ꢂ
ꢂ
water, dried at 70 C for 12 h and further at 120 C overnight.
ꢂ
Finally, the precursors were calcined at 500 C for 4 h.
12,13
follows: D ¼ 2Y/X.
The morphology of the samples was
observed by TEM and it was conducted on an FEI Tecnai
2.2 Characterizations
0
All the XRD experiments were performed on a Philips X pert F20 transmission electron microscope at an acceleration
PRO powder diffractometer with a graphite-monochromated Cu voltage of 200 kV.
Ka operating radiation in the Bragg–Brentano parafocusing
ꢂ
geometry. The patterns were collected between 10 and 80
(
ꢂ
ꢁ1
2.3 Catalytic test
2q range), with a scanning speed of 10 min . The actual
element loadings of Cu and M (Co, Zn, Cr, and Mn) were The catalytic activity tests were carried out at atmospheric
measured by scanning electron microscope (SEM), equipped pressure in a continuous xed-bed reactor, 0.5 g of catalyst were
This journal is © The Royal Society of Chemistry 2017
RSC Adv., 2017, 7, 48662–48669 | 48663

View Article Online
RSC Advances
Paper
loaded in the middle of the tube (Quartz, inner diameter ¼ 8
mm, length ¼ 300 mm). Before reaction, the catalysts were
reduced under owing H /N gas mixture (H /N ¼ 1, molar
2
2
2
2
ꢂ
ratio) at 400 C for 1 h. The liquid feed, prepared by mixing ED
and PG in a mole ratio of 1 : 1, diluted with deionized water (to
5
5
0 wt%), was injected into the reactor by a micro pump at
mL min and N (25 mL min ) was supplied as a carrier gas.
2
ꢁ1
ꢁ1
The liquid mixture form the outlet of the reactor (including
remained raw materials and products) were collected in an ice-
water condenser and subsequently analyzed by gas chroma-
tography (GC-112A) with a capillary column (cross-linked SE-30
gum, 0.33 mm ꢄ 30 m) and a ame ionization detector (FID).
Different concentrated solutions of standard samples (such as
ED, PG, 2-MP, pyrazine, 2-methylpiperazine etc.) were prepared
to make the calibration curve for quantication of the chem-
icals in the products. The identication of the liquid products
was done by GC-MS (Agilent Technol. 6890N/Agilent Technol.
5
5973 Network, Mass Selective Detector).
3. Results and discussion
3
.1 X-ray diffractions of Cu, M, Al-LDHs and CuMAl oxides
Fig. 1(a) shows the XRD patterns of the Cu, M, Al-LDHs
prepared by co-precipitation using NaOH/Na CO as a precipi-
tant agent. The Cu, M, Al-LDHs showed the typical hydrotalcite-
2
3
ꢂ
ꢂ
like pattern with sharp diffraction peaks at 2q ¼ 11.8 , 23.6 and
ꢂ
3
4.9 , corresponding to the reection of (003), (006) and (012)
crystalline planes, respectively. The CuAl, CuZnAl and CuCoAl
ꢂ
samples showed an additional diffraction peak at 2q ¼ 37.5
corresponding to Al(OH) (JCPDS le 033-0018). A pure
3
hydrotalcite-like layered double hydroxide phase has been re-
III
III
II
ported to form only for M /(M + M ) ratios between 0.2 and
14,15
3+
3+
2+
2+
2+
0.4.
However, in our case, the Al /(Al + M (Cu , Co ,
2
+
and Zn )) ratio was considerable higher (0.67). As previously
reported for Co(Zn)–Al LDH samples, excess aluminum segre-
gates to form additional phases such as gibbsite AlO(OH) at 2q
Fig. 1 XRD patterns of Cu, M, Al-LDHs (a), after calcination (b) and
after reduction (c).
ꢂ
16,17
¼
37.4 ,
instead of incorporating into the hydrotalcite-type
Hence, these results support our results showing
7,18
structure.
2
2
the formation of an Al(OH) phase from excess Al in the CuAl, crystalline planes of a metallic Cu phase, respectively. This
3
ꢂ
CuZnAl, and CuCoAl samples. While for the case of CuCrAl result indicated that the reduction pretreatment at 400 C for
nor hydrotalcite-like crystallite phases 1 h allowed the reduction of Cu from the CuAl O spinel to
3 2 4
i.e., amorphous CuCrAl) was found, probably due to the low form a new phase of Cu . Similarly, the CuCrAl sample showed
2
+
sample neither Al(OH)
(
kinetic crystallization constant of Cr .
0
3+
19
a diffraction patterns with only one metallic copper phase.
The XRD patterns of the Cu, M, Al-LDHs aer being calcined at However, CuZnAl sample displayed weaker Cu diffractions
0
ꢂ
5
00 C for 4 h are presented in Fig. 1(b). All the patterns showed peaks accompanied with a spinel phase of ZnAl
2
O
4
. In case of
ꢂ
ꢂ
ꢂ
weak and broad diffraction peaks at 2q ¼ 37 , 45 , and 65 , cor- the CuCoAl sample, no X-ray diffraction peaks corresponding to
responding to a spinel phase structure. However, it was extremely metallic Cu were observed. Instead, diffraction peaks at 2q ¼
ꢂ ꢂ ꢂ
36 , 45 and 65 were detected and ascribed to a Co–Al spinel
difficult to distinguish the exact spinel structure of CuAl
2 4
O ,
2 4 2 4
CoAl O and ZnAl O because of their tiny difference in lattice phase. Considering that the copper loading of the catalyst (18%)
20
constant. As shown in Fig. 1(b), no diffraction peaks corre- is above the typical detection limit of the diffractometers, we
0
sponding to metal oxides such as CuO were detected. No ZnO and can therefore conclude that Cu was highly dispersed on the
Cr
2 3
O phases were observed for all the calcined samples, probably catalyst surface. As described below, this result was further
0
because these phases were highly dispersed on the catalyst and/or conrmed by TEM. The CuMnAl catalyst showed very weak Cu
were present in the form of Cu or Zn spinels.
2
1
diffraction peaks aer reduction, and no evidences of the
0
Fig. 1(c) shows the XRD patterns of the reduced CuMAl formation of Mn and/or MnO were obtained in Fig. 1(c).
ꢂ
ꢂ
catalysts. The CuAl sample showed peaks at 2q ¼ 43 , 50 and
2
N adsorption–desorption was utilized to characterize the
ꢂ
74 , which can be ascribed to the (111), (200) and (220) textural properties of the calcined catalysts prepared from the
48664 | RSC Adv., 2017, 7, 48662–48669
This journal is © The Royal Society of Chemistry 2017

View Article Online
Paper
RSC Advances
LDH as precursors. The Brunauer–Emmett–Teller (BET) specic
surface area, pore volume and pore diameter are summarized in
Table 1. It can been seen that aer thermal decomposition
under air, all the samples showed high surface areas, ca.
2
ꢁ1
centered at 200 m
g
except for CuCrAl, that showed an
2
ꢁ1
specic surface area of only 98 m g . According to the litera-
15,23
ture,
the mixed oxide samples decomposed from hydro-
talcite precursors with similar chemical compositions as our
catalysts usually present specic surface areas lower than
2
ꢁ1
1
00 m
g
. For examples, CoAl and CuCuAl mixed oxides
ꢂ
prepared by calcination of hydrotalcite precursors at 500
showed a surface area of ca. 100 m
respectively.
C
2
ꢁ1
2
ꢁ1
g
and 70 m g ,
1
5,23
3+
3+
2+
However, in our case, we used a Al /(Al + M
2
+
2+
2+
(
Cu , Co and Zn )) ratio (0.67), higher than the nominal
3
+
3+
2+
M /(M + M ) ratio, which led to the formation of hydrotalcite
and Al(OH) phases. Thus, Al(OH) in hydrotalcite-like
3
3
compounds decomposed into alumina oxides during the
thermal pretreatment, thereby resulting in higher surface areas
15,23
as compared to the reported LDH compounds.
3.2
H
2
-TPR of the calciend CuMAl catalysts
H
2
-TPR was used to evaluate the reducibility of the calcined
samples and the hydrogen consumption of the catalysts, and
the results are shown in Fig. 2. Obviously, the nature of the Fig. 2 TPR profiles of the calcined CuMAl samples.
secondary M species strongly inuenced the reducibility of the
calcined samples. In order to establish the relationship between
the reducibility and the catalytic performance, the main over- indicating the formation of a ZnAl O phase, as shown in
2
4
lapped peak was deconvoluted using a Gaussian function. For Fig. 1(c). When Mn was co-precipitated with CuAl, two over-
ꢂ
CuAl sample, there were two overlapped reduction peaks lapped reduction peaks were observed at 256 and 272 C,
ꢂ
centered at 273 and 320 C. The peak at low temperature cor- although their area was lower as compared to those of the CuAl
0
24–26
responded to the reduction of CuO to Cu ,
while the peak at and CuZnAl samples. In this case, the hydrogen consumption
high temperature was ascribed to the hydrogen consumption process at low temperature can be ascribed to the reduction of
process involving the CuAl O spinel in agreement with the CuO and CuAl O , while the one at high temperature can be
2
4
2 4
27–29
results reported by Faungnawakij et al.
nAl catalyst showed two overlapped reduction peaks at slightly / MnO.
Similarly, the CuZ- produced by the following reduction process: Mn O / Mn O
2 3 3 4
27,32
lower temperatures as compared to the CuAl sample (265 and
In case of the CuCoAl catalyst, the reduction was centered at
ꢂ
ꢂ
2
90 C), and also attributed to the reduction of CuO and a very low temperature of ca. 208 C. This enhanced CuO
30,31
CuAl
2
O
4
, respectively.
However, it is worth noting that, reduction compared to the other CuMAl catalysts indicated that
compared to the CuAl catalyst, the reduction peak of CuZnAl at the CuO nanoparticle might be highly dispersed on the mixed
high temperatures showed a signicantly lower area, thereby
Table 1 Textural properties of the calcined CuMAl catalysts and acidic properties of the reduced CuMAl catalysts
3
NH -TPD
N
2
physisorption
Acid site distribution and density
c
Surface
area (m g
Pore
diameter (nm)
Pore volume
a-Weak acidic
sites (%)
b-Moderate acidic
sites (%)
g-Strong acidic
sites (%)
Desorption amount
a
2
ꢁ1
b
3
ꢁ1
d
d
d
ꢁ1
Samples
)
(cm g
)
of NH
3
(mmol g
)
CuAl
168
185
201
227
98
7.6
5.3
4.3
7.8
8.1
0.32
0.25
0.21
0.45
0.20
19.7
43.6
17.9
29.4
17.8
62.1
56.4
57.1
51.4
58.9
18.2
—
25.0
19.2
39.5
1668
2433
1742
1454
598
CuCoAl
CuZnAl
CuMnAl
CuCrAl
a
b
c
BET specic area. Average pore diameter calculated by the Barrett–Joyner–Halenda (BJH) method. Total pore volumes were obtained at
d
P/P
0
¼ 0.975. Determined by deconvolution of the NH
3
-TPD curves.
This journal is © The Royal Society of Chemistry 2017
RSC Adv., 2017, 7, 48662–48669 | 48665

View Article Online
RSC Advances
Paper
3
4,35
metal oxides or the presence of cobalt promoting CuO reduction unsaturated electrons from the metal oxide.
Hence, as
ꢂ
due to their strong interaction. The broad high temperature shown in Fig. 3, the desorption peak at 180 C was ascribed to
ꢂ
peak from 400 to 800 C can be attributed to the reduction of the NH desorbed from Brønsted sites, while the desorption peak at
3
23,33–35
ꢂ
With respect to the CuCrAl sample, 270 C can be attributed to NH desorbed from Lewis Al –O
3
3+
2ꢁ
Co–Al spinel phase.
the hydrogen consumption peak at 200–400 C could be sites. Desorption of NH
ascribed to the reduction of CuCrO to Cu and Cr . In can be ascribed to NH adsorbed on strong Lewis acid sites (i.e.,
general, the nature of M element played a major role on the tetrahedral Al , from the amorphous Al phase formed upon
was desorbed
ꢂ
ꢂ
at high temperatures (above 500 C)
3
4
2
O
3
3
3
+
x y
O
32
reducibility of the calcined CuMAl catalysts. Except for the excess amount of Al ). In the case of CuCoAl, NH
3
ꢂ
sample CuCrAl, the other secondary M metals promoted the below 450 C probably because of the presence of excess Al
reducibility of CuO to a lower reduction temperature compared forming an octahedral Co–Al spinel (not decomposed aer
ꢂ
to that of CuAl catalyst. According to the TPR proles, the reduction at 400 C) providing strong Lewis acid sites for NH3
reducibility of the calcined CuMAl catalysts followed the trend adsorption. In contrast, the CuZnAl catalyst, despite forming
2
+
CuCoAl > CuMnAl > CuZnAl > CuAl > CuCrAl.
a Zn–Al spinel phase, showed Zn remaining sites with strong
Lewis acid properties. Therefore, the CuZnAl exhibited a strong
ꢂ
NH
3
desorption peak at high temperatures 539–551 C.
3
3.3 NH -TPD of reduced CuMAl catalysts
Since the acidic properties of the catalysts also play a key role in
catalyzing the cyclization/dehydration reaction to form the
3.4 Catalytic properties of CuMAl catalysts
important intermediate 2-methylpiperazine, TPD experiments The catalytic performances of the reduced CuMAl catalysts
using NH
as a base probe molecule were performed for all the towards the synthesis of 2-MP as a function of temperature are
3
reduced samples (Fig. 3). A wide distribution of acidic surface shown in Fig. 4. The conversion of PG and ED gradually
sites was obtained as revealed by the large temperature range of increased with temperature. Since the synthesis of 2-MP is an
the desorption peaks. Three NH
2
3
desorption peaks at 170–187, endothermic reaction and therefore the temperature positively
54–293 and 539–551 C were observed for the CuMAl catalysts affected the reaction rate. However, there were some differences
ꢂ
(
M ¼ Zn, Mn, and Cr), whereas the CuCoAl sample, only showed in the selectivity towards 2-MP. The selectivity of 2-MP increased
ꢂ
ꢂ
two desorption peaks at 170 and 257 C. NH
3
typically adsorbs with temperature, reaching a maximum value at 400 C and
ꢂ
on the surface of the CuMAl catalysts over two kinds of acid sites slightly decreasing thereaer (420 C, Fig. 4), probably due to
namely, Brønsted acid (low temperature) and Lewis acid (high the formation of polymerization by-products resulting from
temperature) sites. The Brønsted acid sites correspond to NH
3
side reactions promoted at high temperatures. The secondary M
interacting with surface hydroxyl group, while the Lewis acid species in CuMAl catalysts played a major role in the catalytic
sites correspond to the nitrogen atoms interacting with the performances of these materials. Among all the samples,
CuCoAl provided the highest PG and EG conversions in values
ꢂ
of 63% and 82% at temperature of 340 C, respectively. In
contrast, CuCrAl displayed the poorest catalytic performance
(
i.e., PG and ED conversions lower than 40% and a 2-MP
ꢂ
selectivity of 10–20% at 420 C). Moreover, CuAl, CuMnAl, and
CuZnAl catalysts showed nearly similar PG and ED conversions
and these values were higher than that of CuCrAl, but lower
than that of CuCoAl.
The synthesis of 2-MP via cyclization of ED and PG involves
two steps namely, dehydration and dehydrogenation. Several
groups have proposed a mechanism for the formation of 2-MP
36–38
37
over mixed oxide catalysts.
As reported by Basak et al., the
surface acidity originated from the metal oxides catalyzes the
dehydration reaction, while the metal sites catalyze the dehy-
drogenation of the 2-methylpiperazine intermediate into 2-MP.
Hence, according to the plots shown in Fig. 4, the acid site
distribution and density exposed on the catalyst surface was
calculated and summarized in Table 1. The acid site density of the
CuMAl catalysts as a function of the conversion of ED and PG at
ꢂ
3
60 C is shown in Fig. 5. The density of acidic sites of the catalyst
exhibited a good linear relationship with the conversion of EG
and PG, thereby indicating that the density of acidic sites posi-
tively affected the conversion of reactants. In our case, CuCoAl
ꢁ
1
showed the highest density of acidic sites (2433 mmol g ), and
thus dehydration of ED and PG was favored on acidic sites
present on the catalyst surface, which matched with the highest
3
Fig. 3 NH -TPD profiles of the reduced CuMAl catalysts.
48666 | RSC Adv., 2017, 7, 48662–48669
This journal is © The Royal Society of Chemistry 2017

View Article Online
Paper
RSC Advances
ꢂ
catalytic performance of this catalyst at 360 C. On the contrary,
CuCrAl catalyst exhibited the lowest PG and ED conversion due to
ꢁ1
its lowest surface acidity (598 mmol g ). This result was consis-
tent with the aforementioned mechanism of the synthesis of
2-MP.
The dehydrogenation has a strong dependence on the Cu
dispersion. The TPR results provided circumstantial evidence
that CuCoAl may possess the highest Cu dispersion among all
2
the CuMAl samples. Therefore, N O chemisorption experiment
was further conducted to verify the Cu dispersion. As shown in
Fig. 6, the CuCoAl sample exhibited the highest dispersion of
copper species, which was as high as 73%, and that the
dispersion of Cu particle increased in the following order:
CuCoAl > CuMnAl > CuAl > CuZnAl > CuCrAl. It is generally
acknowledged that higher dispersion of metal corresponds to
38,39
its smaller particle size.
Similarly, we correlated the Cu
dispersion of the reduced CuMAl catalysts with the selectivity
towards 2-MP. As shown in Fig. 6, the selectivity to 2-MP fol-
lowed a positive linear relationship with Cu dispersion. As
discussed above, the dispersion of metal has a major-effect on
the dehydrogenation of the 2-methylpiperazine intermediate
into 2-MP. The higher Cu dispersion led to the higher 2-meth-
ylpiperazine dehydrogenation activities and 2-MP yield (Fig. S1a
and b†). In general, among the investigated samples, CuCoAl
with the highest Cu dispersion and largest acidic site density,
and this material therefore, displayed the highest selectivity to
2
-MP (ca. 55%), in contrast, CuCrAl only showed 15% of 2-MP
selectivity. Interestingly, CuZnAl showed higher selectivity to
-MP as compared to the rest of samples, which was probably
2
produced by the high selectivity of ZnO to the dehydrogenation
6
reaction. Furthermore, the TOF of the catalysts in Fig. 7 showed
that CuCoAl catalyst exhibited the highest TOF, and decreased
values can be found on CuAl, CuCrAl, CuZnAl, and CuMnAl
samples. Among them, CuCrAl samples showed the lowest TOF.
This trend is similar as the trends of PG and ED conversion, as
well as 2-MP yield on these catalysts, indicating that CuCoAl has
the highest intrinsic catalytic performances.
Fig. 4 Catalytic performance of the CuMAl catalysts in the synthesis of
2-MP at different temperatures.
To further conrm the existence of tiny Cu nanoparticles in
CuCoAl, micro-structure characterization by STEM-HAADF and
Fig. 5 Acidic site density versus conversion of ED and PG for the
CuMAl catalysts during the synthesis of 2-MP (reaction conditions:
Fig. 6 2-MP selectivity as a function of the Cu dispersion for the
ꢂ
0
.5 g of catalysts, temperature ¼ 360 C).
reduced CuMAl catalysts (reaction conditions: 0.5 g of catalysts,
ꢂ
temperature ¼ 400 C).
This journal is © The Royal Society of Chemistry 2017
RSC Adv., 2017, 7, 48662–48669 | 48667

View Article Online
RSC Advances
Paper
4. Conclusions
In this work, well-dispersed copper over CuMAl as the bi-
functional catalysts (M ¼ Cr, Mn, Zn, and Co) derived from
LDH compounds were prepared and applied in the synthesis of
2-MP. The combined characterization results revealed that the
nature of M metal remarkably inuenced the Cu dispersion and
surface acidity of the catalysts, and thus played a major role on
the catalytic performance compared with the CuAl metal oxides.
Among them, CuCoAl catalyst displayed the highest 2-MP yield
ꢂ
(
56%) at 400 C, whereas CuCrAl sample exhibited the lowest 2-
ꢂ
MP yield (6%) at 400 C. The explanation for the higher catalytic
Fig. 7 TOF comparison over CuMAl catalysts (reaction conditions, behaviors achieved over CuCoAl catalyst with compared to the
ꢂ
0
.5 g catalyst, temperature 340 C).
other samples can be ascribed to (i) more surface acid density
ꢁ
1
(2433 mmol g ), which facilitated to cyclization/dehydration of
ED and PG, (ii) highly dispersed Cu (ꢀ2 nm) favored to dehy-
TEM were conducted. Fig. 8(a) showed the STEM-HAADF and drogenation of the piperazine intermediate into 2-MP. Our
TEM images of a reduced CuCoAl with porous feature. The results offer new hints for rational design of bifunctional cata-
corresponding EDS mapping in Fig. 8(b–d) revealed the lyst for 2-MP synthesis.
relatively homogeneous distribution of Al and O. Co also
distributed throughout the particle, with slightly higher
concentration in the particle center. By contrast, Cu metal Conflicts of interest
existed as nanosized domain. High resolution TEM provided
more structure details. According to the deduced lattice
There are no conicts to declare.
distance values from the metal particles (Fig. 8(j)), we conrmed
the formation of metallic Cu. Interestingly, the reduction of the
calcined CuCoAl sample resulted in the presence of metallic Co
Acknowledgements
and CoO nanoparticles (Fig. 7(i and k)). The calculated average
particle size of Cu nanoparticles based on HRTEM character-
ization was as small as ꢀ2 nm.
We gratefully acknowledge nancial support from the spring
plan of the Ministry of Education (Z2016137), the Sichuan
Education Department (17ZB0410), the Open Research Subject
of key laboratory (Research Base) of automotive engineering
(szjj2017-075), Science & Technology Department of Sichuan
Province (2016GZ0027, 2015GZ0127) and the opening project of
State key laboratory of Polymer Materials Engineering (Sichuan
University, Grant No. sklpme2016-4-35). P. Liu acknowledges
the Open Research Subject of key laboratory (Research Base) of
food biotechnology (szjj2013-046). X. Li and X. Liao acknowl-
edge the nancial support from the National Natural Science
Foundation of China (NSFC, No. 21403280, 21706216).
References
1
2
3
4
5
6
7
V. Akula, R. Sarkari, A. Chatla, K. Vankudoth and
K. K. Mandari, Appl. Catal., A, 2012, 441, 108–118.
I. Park, J. Lee, Y. Rhee, Y. Han and H. Kim, Appl. Catal., A,
2003, 253, 249–255.
I. Park, Y. Rhee, J. Lee, Y. H. Han, J. Jeon and H. Kim, Res.
Chem. Intermed., 2003, 29, 575–587.
F. L. Jing, W. Chu, Y. Y. Zhang, Y. Q. Chen and S. Z. Luo,
Chin. Chem. Lett., 2008, 19, 752–755.
F. L. Jing, Y. Y. Zhang, S. Z. Luo, W. Chu, H. Zhang and
X. Y. Shi, J. Chem. Sci., 2010, 122, 621–630.
F. L. Jing, Y. Y. Zhang, S. Z. Luo, W. Chu and W. Z. Qian, Appl.
Clay Sci., 2010, 48, 203–207.
Fig. 8 Morphology of the reduced CuCoAl catalyst. (a–c), and (e–f)
are the STEM-HAADF results, while (d), (g), (h), (j), (k), and (i) correspond
to the TEM and HRTEM results. (l) Cu particle size distribution.
W. Luo, F. L. Jing, X. P. Yu, S. Sun, S. Z. Luo and W. Chu,
Catal. Lett., 2012, 142, 492–500.
48668 | RSC Adv., 2017, 7, 48662–48669
This journal is © The Royal Society of Chemistry 2017

View Article Online
Paper
RSC Advances
8
R. Sarkari, C. Anjaneyulu, V. Krishna, R. Kishore, 24 M. N. Barroso, M. F. Gomez, J. A. Gamboa, L. A. Arrua and
M. Sudhakar and A. Venugopal, Catal. Commun., 2011, 12,
067–1070.
A. Venugopal, R. Sarkari, C. Anjaneyulu, V. Krishna,
M. C. J. Abelloa, Phys. Chem. Solids, 2006, 67, 1583–1589.
25 A. J. Marchi, D. A. Gordo, A. F. Trasarti and C. R. Apestegu ´ı a,
Appl. Catal., A, 2003, 249, 53–67.
1
9
M. K. Kumar, N. Narender and A. H. Padmasri, Appl. 26 Y. Tanaka, T. Utaka, R. Kikuchi, T. Takeguchi, K. Sasaki and
Catal., A, 2014, 469, 398–409. K. Eguchi, J. Catal., 2003, 215, 271–278.
0 V. Krishna, S. N. Kumar, S. Reema, A. H. Padmasri, 27 K. Faungnawakij, N. Shimoda, T. Fukunaga, R. Kikuchi and
1
K. V. R. Chary and A. Venugopal, Appl. Catal., A, 2014, 488,
75–284.
1 C. A. Antonyraj, M. Gandhi and S. Kannan, Ind. Eng. Chem.
Res., 2010, 49, 6020–6026.
K. Eguchi, Appl. Catal., A, 2008, 341, 139–145.
28 S. Tanasoi, N. Tanchoux, A. Urd ˘a and I. C. Marcu, Appl.
Catal., A, 2009, 363, 135–142.
29 Q. Song, W. Liu, C. D. Bohn, R. N. Harper, E. S. Scott and
S. A. Dennis, Energy Environ. Sci., 2013, 6, 288–298.
2
1
1
2 V. Dasireddy and B. Likozar, Fuel, 2017, 196, 325–335.
1
3 S. Hajduk, V. Dasireddy, B. Likozar, G. Drazi and Z. Orel, 30 S. Zhang, Q. Liu, G. Fan and F. Li, Catal. Lett., 2012, 142,
Appl. Catal., B, 2017, 211, 57–67. 1121–1127.
4 S. Velu, K. Suzuki, S. Hashimoto, N. Satoh, F. Ohashi and 31 J. Beiramar, A. Constant and A. Khodakov, ChemcatChem,
S. Tomura, J. Mater. Chem., 2001, 11, 2049–2060. 2014, 6, 1788–1793.
5 M. Gabrovsk, R. Edreva-Kardjieva, K. Tenchev, P. Tzvetkov, 32 Y. Tanaka, T. Takeguchi, R. Kikuchi and K. Eguchi, Appl.
1
1
A. Spojakina and L. Petrov, Appl. Catal., A, 2011, 399, 242–
51.
Catal., A, 2005, 279, 59–66.
33 S. Velu, K. Suzuki and T. Osaki, Catal. Lett., 2000, 69, 43–50.
2
1
1
6 T. Tsuchida and N. Ichikawa, React. Solids, 1989, 7, 207–217. 34 C. Rudolf, B. Dragoi, A. Ungureanu, A. Chirieac, S. Royer,
7 K. J. D. Mackenzie, J. Temuuijin and K. Okada, Thermochim.
Acta, 1999, 327, 103–108.
A. Nastro and E. Dumitriu, Catal. Sci. Technol., 2014, 1,
179–189.
1
1
2
2
2
2
8 M. E. Perez-Bernal, R. J. Ruano-Casero, F. Benito and 35 L. J. I. Coleman, W. Epling, R. R. Hudgins and E. Croiset,
V. Rives, J. Solid State Chem., 2009, 182, 1593–1601. Appl. Catal., A, 2009, 363, 52–63.
9 C. M. S. Polato, C. A. Henriques, A. C. C. Rodrigues and 36 A. Auroux and A. J. Gervasini, J. Phys. Chem., 1990, 94, 6371–
J. L. F. Monteiro, Catal. Today, 2008, 133–135, 534–540. 6379.
0 J. Cheng, J. Yu, X. Wang, L. Li, J. Li and Z. Hao, Energy Fuels, 37 J. Basak, N. Hardia, S. Saxena, R. Dixit, R. Dwivedi,
2
008, 22, 2131–2137.
S. Bhadauria and R. Prasad, Ind. Eng. Chem. Res., 2007, 46,
7039–7044.
1 E. L. Rodrigues, A. J. Marchi, C. R. Apesteguia and
J. M. C. Bueno, Appl. Catal., A, 2005, 294, 197–207.
38 A. S. Bennici, Appl. Catal., A, 2005, 281, 199–205.
2 B. R. Strohmeier, D. E. Levden, R. S. Field and 39 Z. Yuan, L. Wang, J. Wang, S. Xia, P. Chen, Z. Hou and
D. M. Hercules, J. Catal., 1985, 94, 514–530.
X. Zheng, Appl. Catal., B, 2011, 101, 431–440.
3 V. Rives, A. Dubey and S. Kannan, Phys. Chem. Chem. Phys.,
2001, 3, 4826–4836.
This journal is © The Royal Society of Chemistry 2017
RSC Adv., 2017, 7, 48662–48669 | 48669