Nitrogen- and oxygen-functionalized carbon nanotubes supported Pt-based catalyst for the selective hydrogenation of cinnamaldehyde

Article abstract of DOI:10.1016/j.apcata.2015.08.023

Carbon nanotubes (CNTs) supported Pt catalysts were employed to study the influence of nitrogen- and oxygen-functionalized surface on catalytic performance in the liquid-phase hydrogenation of cinnamaldehyde. Nitrogen- and oxygen-functionalized CNTs were simply treated with the hydrothermal method using ammonia and hydrogen peroxide, respectively, and the metal catalyst was prepared using an impregnation-reduction-deposition method. Here, we studied and gave some insights into the influence of the functionalization on CNTs surface chemistry environment and Pt nanoparticles using XRD, Raman spectroscopy, TEM, XPS, and nitrogen sorption techniques. It was shown that the functionalization lead to the small size distribution of Pt particles and promoted the conversion from ca. 75% to ca. 95%. To obtain the high selectivity, transition metals, cobalt and nickel, were doped into the Pt catalyst and the 99.7% conversion with the 87.9% selectivity was achieved over the nitrogen-functionalized CNTs supported PtCo catalysts. The PtCo/N-CNTs remained high activity even at 40°C and achieved 94.2% conversion and 90.3% selectivity.

Full text of DOI:10.1016/j.apcata.2015.08.023

Applied Catalysis A: General 506 (2015) 134–142
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Nitrogen- and oxygen-functionalized carbon nanotubes supported
Pt-based catalyst for the selective hydrogenation of cinnamaldehyde
Zhengbin Tian, Chao Liu, Qingyang Li, Juying Hou, Yan Li∗, Shiyun Ai∗
College of Chemistry and Material Science, Shandong Agricultural University, 271018, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 25 April 2015
Received in revised form 9 July 2015
Accepted 17 August 2015
Available online 21 August 2015
Carbon nanotubes (CNTs) supported Pt catalysts were employed to study the influence of nitrogen-
and oxygen-functionalized surface on catalytic performance in the liquid-phase hydrogenation of cin-
namaldehyde. Nitrogen- and oxygen-functionalized CNTs were simply treated with the hydrothermal
method using ammonia and hydrogen peroxide, respectively, and the metal catalyst was prepared using
an impregnation-reduction-deposition method. Here, we studied and gave some insights into the influ-
ence of the functionalization on CNTs surface chemistry environment and Pt nanoparticles using XRD,
Raman spectroscopy, TEM, XPS, and nitrogen sorption techniques. It was shown that the functionaliza-
tion lead to the small size distribution of Pt particles and promoted the conversion from ca. 75% to ca.
95%. To obtain the high selectivity, transition metals, cobalt and nickel, were doped into the Pt cata-
lyst and the 99.7% conversion with the 87.9% selectivity was achieved over the nitrogen-functionalized
CNTs supported PtCo catalysts. The PtCo/N-CNTs remained high activity even at 40 ◦C and achieved 94.2%
conversion and 90.3% selectivity.
Keywords:
Functionalized
Carbon nanotubes
Cinnamaldehyde
Selective hydrogenation
© 2015 Elsevier B.V. All rights reserved.
1. Introduction
banD width. Generally, the larger the d-band, the stronger the
repulsive four-electron interaction of the metal with the C
C
Cinnamaldehyde (CAL) is a particularly important representa-
tive of ␣, ␤-unsaturated aldehyde that contains two unsaturated
bonds the carbonyl (C O) and olefinic (C C) group in molecule.
The hydrogenation of CAL can generate different products, the
partial one of which, cinnamyl alcohol (COL), is important inter-
mediates in pharmaceutical and fragrance industries [1,2]. Also,
it is considered as a good model for investing the catalytic
behaviors of the microstructures of heterogeneous catalysts. The
selective hydrogenation is a rather complex reaction network
involving several important intermediates and a number of series-
parallel reactions. The main competitive products of selective
hydrogenation can be produced through the hydrogenation of
the conjugated C C or C O bond over the supported metal
catalysts. However, obtaining the unsaturated alcohols is diffi-
cult, owing to the fact that C&9552;O bond presents a higher
binding energy than C&9552;C bond (715 kJ/mol and 615 kJ/mol,
respectively) [3], the reduction of the C&9552;C bond is thermody-
namically more favorable than that of C O bond. The selectivity
to the unsaturated alcohol can be correlated with the metal D-
double bond and the stronger the attractive interaction of the
metal surface with the C O ␲-system, which results in good
selectivity to the unsaturated alcohol (Ir, Os > Pt > Ru > Rh > Pd)
[4].
Although various kinds of metal (Pd, Ru and Au)-based catalysts
were employed in the hydrogenation reaction [5–7], Pt was most
commonly used for the purpose of C O hydrogenations in fact.
Maintaining high selectivity of Pt-based catalysts while increasing
their activity remains an urgent task for further development in this
field. Many studies have been devoted to enhance the selectivity
towards the unsaturated alcohol either by fine control of the sizes
[8] and/or the exposed facets of the active metal nanoparticles [9],
or by exploiting the steric constraints imposed by the environment
of the active site [10], or by decorating the primary metal with a sec-
ond metal component [11], or by adding promoters like an alkali
element [1]. However, most methods usually involved ingenious
design and complex process. Therefore, searching for appropriate
materials used as supports is a simple and feasible way to prepare
highly active and selective catalysts for the hydrogenation reaction.
The conventional carbon nanostructure materials, especially
carbon nanotubes (CNTs), have been widely employed as support
in heterogeneous catalysis, owing to the high electrical conduc-
tivity and corrosion resistance [12]. Their high surface area could
∗
Corresponding authors at: 271018, Daizong Street 61, Taian, Shandong, PR China.
Fax: +86 538 8247660.
E-mail addresses: liyan2010@sdau.edu.cn (Y. Li), ashy@sdau.edu.cn (S. Ai).
http://dx.doi.org/10.1016/j.apcata.2015.08.023
0926-860X/© 2015 Elsevier B.V. All rights reserved.

Z. Tian et al. / Applied Catalysis A: General 506 (2015) 134–142
135
Fig. 1. XRD patterns (a) and Raman spectroscopy (b) for Pt/CNTs, Pt/N-CNTs and Pt/O-CNTs.
lead to a high dispersion of metal nanoparticles and the perfect
2. Experimental
smooth graphitic surface could make all metal nanoparticles on
CNTs accessible to the reactants. Additionally, the transfer of ␲-
electrons from the graphitic planes of CNTs to the metal particles
leads to the increase in the charge density on the metal, which
decreases the probability of C C bond activation and favors the
formation of unsaturated alcohols [13,14]. Recently, functionalized
or doped CNTs are considered as one of the promising supports
due to the merits performing in the catalytic process. The function-
alization mainly focuses on two aspects, nitrogen and oxygen. The
functionalization treatments have excited chemists to renovate tra-
ditional catalysis materials and, actually, it has been proved that
they could significantly improve CNTs interaction with solvents
and dispersion, allow the grafting of nanoparticles [15,16] and pro-
mote the adsorption of energy molecules and modulate molecular
reactivity and selectivity [17,18]. Chizari et al. prepared N-doped
CNTs (synthesized using a Chemical Vapor Deposition process) sup-
ported Pd catalyst and investigated the influence of different types
of incorporated nitrogen species on the liquid-phase hydrogena-
tion of CAL [19]. Solhy et al. reported three activation procedures
of Multi-walled CNTs: nitric acid oxidation, ball-milling and air oxi-
dation and its influence on the catalytic performance of Pt/MWCNT
catalysts for selective hydrogenation [20]. However, the different
influence of nitrogen and oxygen species on the catalytic perfor-
mance has been not researched.
Here, nitrogen- and oxygen-functionalized CNTs were sim-
ply obtained with the hydrothermal method using ammonia
and hydrogen peroxide, respectively. Pt-based catalyst sup-
ported on those nanotubes was prepared with the traditional
impregnation-reduction-deposition method. The main aim of the
present work was to study and compare the effects of oxygen
and nitrogen species on the CNTs surface chemistry environ-
ment and Pt-based nanoparticles, and to evaluate the influence
of the functionalization on the catalytic performances of Pt-
based catalysts for selective hydrogenation. The functionalized
CNTs and prepared Pt-based catalysts were fully characterized
using X-ray diffraction (XRD), Raman spectroscopy, transmis-
sion electron microscopy (TEM), nitrogen sorption isotherms and
X-ray photoelectron spectroscopy (XPS) techniques. The perfor-
mances of different Pt-based catalysts were investigated and
compared under mild reaction conditions. We found that the high
conversion and selectivity could be achieved over the nitrogen-
functionalized CNTs supported PtCo catalysts even at the room
temperature.
2.1. Materials and chemicals
Multi-walled CNTs prepared by the chemical vapor deposi-
tion method with Fe-Co/Al2O3 as catalyst were obtained from
Boyu Gaoke New Material Technology Co., Ltd. (Beijing, China).
All reagents used in the experiment were analytical reagent
grade and used without further purification. H2PtCl6·6H2O,
CoCl2·H2O, CAL and NaBH4 were purchased from Aladdin. Ammo-
nia, hydrogen peroxide (30%) and concentrated HCl (12 mol/L)
were purchased from Kangde Chemical Reagent (Yantai, China).
Ethanol was purchased from Yongda Chemical Reagent (Tianjin,
China). Distilled deionized water was utilized through the experi-
ments.
2.2. Catalysts preparation
Prior to functionalization, the pristine CNTs were refluxed in 6 M
NaOH and 6 M HCl aqueous solution at 110 ◦C for 6 h, respectively,
in turn to remove catalysts. The purified CNTs were thoroughly
washed with distilled water until the filtrate was neutral, and
then dried at 70 ◦C overnight. A typical surface functionalization
procedure of CNTs is described as follows: 2 g purified CNTs and
60 mL ammonia was placed into a 100 mL Teflon-vessel, which
was sealed in the autoclave subsequently, and then moved to the
autoclave for steaming at the temperature of 180 ◦C. After the
steaming treatment for 12 h, the autoclave was cooled down to
the room temperature and the nitrogen-functionalized CNTs (N-
CNTs) was obtained and thoroughly washed with distilled water.
The oxygen-functionalized CNTs (O-CNTs) were obtained though
the same method with hydrogen peroxide as the functionalization
reagent instead.
The Pt/CNTs catalyst was prepared using an impregnation-
reduction-deposition method. The CNTs support was impregnated
in aqueous solution of and H2PtCl6 at 80 ◦C for 5 h in a round-bottom
flask. Then, 20 mL of 0.1 mol/L NaBH4 solution was subsequently
added dropwise to the flask under vigorous stirring for another
1 h. Filter the slurry and wash the solid with water until fil-
trate is neutral, then dry the solid at 50 ◦C under vacuum for
12 h. The nominal content of Pt in the catalyst was 2 wt.%.
The PtCo and PtNi bimetallic catalyst was prepared using the
same method with nominal Pt and content of 1 wt.%, respec-
tively.

136
Z. Tian et al. / Applied Catalysis A: General 506 (2015) 134–142
Table 1
Comparison of some parameters of the different CNTs supported Pt catalyst.
d
Catalysts
SBETa(m2/g)
SBJHb(m2/g)
Vmesobcm3/g
Vmicroccm3/g
Vtotalcm3/g
DBJH (nm)
Pt/CNTs
Pt/N-CNTs
Pt/O-CNTs
212.6
224.3
232.2
211.6
220.1
235.5
1.00
1.12
1.26
0.06
0.08
0.10
1.06
1.20
1.36
3.15
3.41
3.83
a
SBET: from the BET method.
SBJH and Vmeso: calculated from the BJH method.
Vmicro: calculated from the DFT methods.
b
c
d
DBJH: mesopore diameter calculated from absorption branch of nitrogen isotherms using BJH method.
Table 2
XPS surface Element content, binding energy values and relative concentration in the Pt4f7/2 for different CNTs supported Pt catalyst.
Catalysts
Element content (%)
Binding energy (eV)
Relative concentration (%)
C
N
O
Pt
Pt(0)
Pt(II)
Pt(IV)
Pt(0)
Pt(II)
Pt(IV)
Pt/CNTs
Pt/N-CNTs
Pt/O-CNTs
96.24
94.34
93.61
0.69
1.85
0.63
1.77
2.04
3.96
1.19
1.54
1.60
71.7
72.2
71.8
72.5
72.9
72.6
74.2
74.7
74.7
69.04
55.14
53.69
21.26
30.60
30.36
9.70
14.26
15.96
2.3. Characterization
the two treated samples had a broader full width at half maxi-
mum (FWHM) which was related to crystallite size, implying that
the treatment damaged the ordered crystal structure, resulting
in a decrease in crystallite size. This is also proved through the
Raman spectroscopy. The broad diffraction peaks at 40.2◦, 46.4◦
and 68.2◦ in good agreement with the metallic platinum (1 1 1),
(2 0 0), and (2 2 0) reflections (JCPDS card No. 04-0802), could be
observed in the Pt/CNT sample, which were consistent with the
face-centered cubic (fcc) structure of platinum [22]. However, the
peaks disappeared in the Pt/N-CNTs and Pt/O-CNTs due to the
smaller crystallite size of Pt, indicating functionalized CNTs support
favor the dispersion of metal nanoparticles. When the crystal-
lite size was less than 5 nm, the two adjacent peaks Pt(1 1 1) and
Pt(2 0 0) will broaden and overlap each other in amplitude rather
than intensity, which would lead to the weakening even disappear-
ance in peaks.
XRD patterns were measured at room temperature by using a
D8 Advance (Bruker, Germany) X-ray diffractometer with the Cu
K␣ radiation at 40 kV and 30 mA. Raman spectra of samples were
collected from 200 to 2000 cm−1 on a LabRAM HR800 Laser Confo-
cal Micro-Raman Spectroscopy (Horiba Jobin Yvon, Japan) using a
532 nm laser source. TEM was characterized by a by a FEI Tecnai G20
(USA) with an accelerating voltage of 200 kV. The textural proper-
ties of samples were performed by nitrogen sorption isotherms on
an ASAP 3020 instrument (Micromeritics, USA) at 77 K. The surface
area and pore volume were calculated using BET and BJH meth-
ods, respectively. XPS was measured with a K-Alpha XPS system
(Thermo Fisher Scientific, USA) using a monochromatic Al K˛ as the
excitation source (1486.6 eV). Survey (wide) scans spanned from
1000 to 0 eV binding energy, which were collected with an analyzer
pass energy of 100 eV at an interval of 1.00 eV.
Raman spectroscopy was employed to characterize carbona-
ceous materials for distinguishing ordered and disordered crystal
structures. All the materials in the Fig. 1b had same peak positions
of the G band (1576 cm−1), assigned to Raman-active E2 g mode for
the tangential in-plane stretching vibrations of the sp2-hybridized
bonds, the D band (1326 cm−1), corresponding to the breakdown
of translational symmetry produced by the microcrystalline struc-
ture, and D’-band (1612 cm−1), attributed to a double-resonance
Raman feature induced by disorder, defects or ion intercalation
between the graphitic walls [23]. After the functionalization, the N
and O element was doped into the ordered crystal structure of CNTs.
Therefore the lattice defect obviously increased and the D band
became strong for N-CNTs and N-CNTs. Typically, The ratio of the
integral intensities of the G and D bands (ID/IG) is used to probe the
degree of graphitization [24] and the in-plane crystallite size (equa-
tion: La = (2.4 × 10 − 10)ꢀ4laser(ID/IG)−1) [25] of samples. Among the
three samples, Pt/N-CNTs and Pt/O-CNTs have the higher ID/IG value
than Pt/CNTs and so they have the lower degree of graphitization
and smaller crystallite size, which is consistent with XRD.
2.4. Hydrogenation reaction
The hydrogenation reactions were carried out in a stainless
autoclave reactor with a 50 mL Teflon sleeve. In a typical pro-
cedure, 0.1 g catalyst was dispersed in 19 mL ethanol and then
8.2 mmol CAL was added into solution. The reactor was sealed,
purging with H2 for 3 times, respectively, and then pressurized
to 2.0 MPa. The reaction was conducted at 70 ◦C with a stirring
speed of 750 rpm for 2 h. The reaction mixture was analyzed by
Shimadzu GC-2010 gas chromatograph with a flame ionization
detector (FID) system equipped with a capillary column SE-54 [(5%-
phenyl)-95%dimethylpolysiloxane, 30 m × 0.32 mm × 0.5 ␮m].
3. Result and discussion
3.1. Catalysts characterization
Fig. 2 shows the TEM photographs and the corresponding his-
tograms of the particle size distribution of Pt NPs on the CNTs,
respectively. There was no obvious difference in the terms of
morphology and all the catalysts presented regular mesochannels
with the external diameter averaged 13 nm and inner diameter of
3.6 nm. It was observed that Pt NPs were highly dispersed on the
support with a narrow sizes distribution ranging from 2 to 6 nm,
regardless of the types of support. This was because NaBH4 reduc-
tion was able to produce small and uniform metal particles [26].
However, the maximum of Pt sizes present a slight difference for
these catalysts. The main Pt sizes in Pt/CNTs concentrated on 4-
Fig. 1a shows the powder XRD patterns of as-prepared catalysts.
The peaks at 26.1◦, 43.5◦, 54.1◦, 77.5◦ were attributed to the (0 0 2),
(1 0 1), (0 0 4) and (1 1 0) diffractions of hexagonal graphitic car-
bon (JCPDS 41-1487), respectively. The wide (0 0 2) peak showed
that the CNTs possess amorphous structure with small regions of
crystallinity and the regions of crystallinity are highly graphitic
[21]. The position and sharpness of the C (0 0 2) peak in the Pt/N-
CNTs and Pt/O-CNTs indicates that the graphite structure of the
CNTs was functionalized without significant damage. The lattice
structure (d) of region of crystallization was maintained well after
the functionalization. A puzzling phenomenon was observed that

Z. Tian et al. / Applied Catalysis A: General 506 (2015) 134–142
137
Fig. 3. Nitrogen adsorption–desorption isotherms of different CNTs supported Pt
catalyst.
tion (at lower p/p0) with H3-type hysteresis loop in the relative
pressure region of 0.85–1.0, indicating the existence of plenty of
mesopores in the catalyst. The surface area, pore volume and pore
diameter calculated from the N2 sorption isotherm are tabulated in
the Table 1. It can been seen from the Table 1 that the similar date
of BET surface area and BJH surface area was observed and meso-
pores volume was dominant in the total pore volume. Obviously,
the functionalized presented an increase in the surface area, pore
volume and pore diameter. This would promote the effective expo-
sure of active sites in the catalyst, resulting in the improvement of
catalytic activity (Fig. 3).
XPS was used to investigate the elemental composition, chemi-
cal state and electronic state of the elements existed in catalysts. On
the basis of the XPS analyses, it was found that the three catalysts
were composed of C, O, N, Pt elements and the element content
was tabulated in the Table 2. The inherent existence of N and O
elements was detected in the Pt/CNTs, which may come from the
prepared process of CNTs and the molecular N2 and O2 absorbed
on the surface. Fig. 4a reveals the C1s spectrum, which was com-
posed of the major sp2C fraction of carbon species, sp3 hybridized
C and minor contribution from different bonding configurations
of carbon with inherent oxygen or nitrogen. There is no obvious
difference among the C1s spectrum, but a hump around 291.5 eV
in the ammonia and hydrogen peroxide functionalized catalyst,
respectively, demonstrating that the carbon matrix was chemically
modified with nitrogen- or oxygen-containing functional groups.
The High-resolution N1s spectra of Pt/N-CNTs are shown in Fig. 4b
and was deconvoluted into three types: 398.9 0.1, 400.4 0.2,
and 401.2 0.1, corresponding to the pyridinic N (N1), pyrrolic
N (N2), and N-H (N3) with relative percentage of 28.6, 57.1 and
14.3% [27–29]. From an oxygen deconvolution curve (Fig. 4c), the
oxygen spectrum can be well-resolved to three peaks with bind-
ing energies at 531.9 0.2, 533.1 0.1, and 534.1 0.1 eV assigned
Fig. 2. TEM photographs of the supported Pt catalysts and Pt size distribution his-
tograms: (a) Pt/CNTs, (b) Pt/N-CNTs, (c) Pt/O-CNTs.
4.5 nm with a mean size of 4.3 nm. The Pt/N-CNTs and Pt/O-CNTs
catalyst had smaller Pt NPs with a mean size of 3.7 nm and 3.9 nm,
respectively. This result strongly suggested that the final particle
size distribution of Pt was dominated by the interaction between
Pt and the functionalized support surface. Most of the Pt NPs were
located on the external wall of CNTs, while a portion of the Pt NPs
with a small size were selectively present with a light color at the
inner wall in the Pt/N-CNTs and Pt/N-CNTs. Due to the limit of the
CNTs channels, the large Pt NPs could not enter into the pores and
were forced to locate on the external surface of the support, while
the small ones could enter and locate on the inner wall. This phe-
nomenon shows that the confinement effect also plays a minor role
in affecting the size distribution of Pt particles.
to C O (O1), C
O C (O2), and C O H (O3) species [28,30]. The
corresponding relative percentage of three peaks is 44.7, 32.8 and
22.5%, respectively.
In comparison with the N and O spectrum in the Pt/CNTs (Fig.
S2), the N3-type nitrogen and O3-type oxygen in the functional-
ized catalyst increased obviously and the changes of the others
were irregular. Although the details of the surface chemistry occur-
ring during NH3 and H2O2 functionalization remain elusive, these
changes clearly suggested that the NH3 and H2O2 functionalization
not merely increased the density of the N and O element but also
The porosity and surface area of the different catalysts were
characterized by N2 sorption measurements. All the samples
exhibit mixed isotherm of type II (at high p/p0) and type IV adsorp-

138
Z. Tian et al. / Applied Catalysis A: General 506 (2015) 134–142
Fig. 4. XPS patterns of (a) C1s, (b) N1s in Pt/N-CNTs, (c) O1s in Pt/O-CNTs and (d-f) Pt4f in Pt/CNTs, Pt/N-CNTs and Pt/O-CNTs, respectively.
dramatically modified its surface functional groups, both of which
were crucial for enhancing the catalytic activity.
From the Table 2, the surface content of Pt in the Pt/N-CNTs and
Pt/O-CNTs was higher than that in the Pt/CNTs. This was most due
to the interaction between PtCl2- 6and the functional groups on the
surface of N-CNTs and O-CNTs. Tuaev et al. [31] had proved that the
doped nitrogen had the stabilizing effect on Pt nanoparticles and
mediated the enhancement of platinum adsorption on the surface
of carbon materials. Through the interaction, the metal ions were
adsorbed on the surface of support and reduced in the process of
catalyst preparation, which undoubtedly improved the content and
dispersion of metal nanoparticles in catalyst.

Z. Tian et al. / Applied Catalysis A: General 506 (2015) 134–142
139
Fig. 5. Reaction pathways of the hydrogenation of CAL.
Fig. 6. Conversion and selectivity of CAL as a function of reaction time over three catalysts (0.1 g catalyst, 70 ◦C, 2.0 MPa).
Fig. 4d–f presented the Pt4f spectra of obtained catalysts, which
could be deconvoluted into three pairs of doublets attributed to
zerovalent Pt(0), Pt(II) and Pt(IV) species. The interval of binding
energy between the doublets was around 3.5 eV and the ratio of
integral peak area of Pt4f7/2 to Pt4f5/2 was close to a value of 4: 3. The
most intense doublet was attributed to metallic Pt, but the weakest
peaks could be assigned to the Pt(IV) species such as PtO2 [5,32]. The
third pair of peaks was most likely ascribed to Pt(II) chemical state
as in either PtO or Pt(OH)2 [1,33]. The binding energy and relative
concentration for Pt species are summarized in Table 2. The bind-
ing energy of Pt is affected by the functionalization agent. A shift of
ca. 0.5 eV and 0.2 eV toward high binding energy for Pt/N-CNTs and
Pt/O-CNTs were observed as compared with Pt/CNTs. The observ-
able shift of the binding energy of Pt4f confirmed there mostly was
an electron transfer between Pt and amino and oxygen groups in the
Pt/N-CNTs and Pt/O-CNTs catalysts, which was similarly observed
in Pt NPs supported on polypyrrole carbon black [34]. The distri-
bution of Pt(0), Pt(II) and Pt(IV) species is similar for the Pt/N-CNTs
and Pt/O-CNTs catalysts studied, but the obvious difference was
that the content of Pt(II) and Pt(IV) was much higher than that in
the Pt/CNTs.
unidentified products may be produced in small amounts particu-
larly over nonselective catalysts [35].
Different CNTs supported Pt catalysts were tested for the selec-
tive hydrogenation of CAL under the same mild conditions for the
investigation of the support effect and time effect. Fig. 6 shows the
variation of CAL conversion and product selectivity as a function
of reaction time for Pt/CNTs, Pt/N-CNTs and Pt/O-CNTs, respec-
tively. As shown in Fig. 6, the reaction time sharply influenced the
conversion of CAL, but had little effect on selectivity for Pt/CNTs
catalyst. The conversion of CAL reached 93.7% over Pt/N-CNTs cat-
alyst after reaction for 1.5 h; in comparison, only 82.7% conversion
was obtained over Pt/CNTs catalyst after reaction for 2.5 h. It was
observed that the conversion of CAL reached 90.8% over Pt/O-CNTs
catalyst after reaction just for 1.0 h. It is obviously the activity
of Pt/N-CNT was lower than that of Pt/O-CNT in the same reac-
tion conditions. No significant selectivity change was found on
Pt/N-CNTs and Pt/O-CNT in the reaction time before 2 h. How-
ever, the amount of HCOL increased sharply after reaction for 1.5 h.
Extending reaction time improved the equivalent opportunity that
hydrogen molecules accessed to the C C and C O in the reac-
tion medium, which resulted in an increase in conversion and a
stable selectivity. However, as the reaction time was extended,
the increased concentration of CAL and COL leaded to the fur-
ther hydrogenation and caused a sharp increase in HCOL especially
over Pt/CNTs and Pt/O-CNTs. The presence of the functionalization,
especially for oxygen-functionalization, in the final catalyst was
detrimental to the selective hydrogenation of CAL to COL in long
reaction time. Though the COL selectivity was not changed, the cat-
alytic activity in terms of conversion was enhanced greatly over
Pt/CNTs and Pt/O-CNTs. The enhancement was related to the high
Pt content and smaller Pt NPs size without a doubt. Specially, the
attendance of high content of reducible Pt species, like PtO and PtO2
3.2. Catalytic test
The hydrogenation of CAL typically produced a mixture of the
desired COL and the undesired products. Fig. 5 illustrated the
reaction pathways in hydrogenation of CAL. The hydrogenation
in C&9552;O bond can form COL while the hydrogenation in C
C
bond leaded to the formation of HCAL. Further hydrogenation of
these two inter-mediates resulted in the formation of hydrocin-
namyl alcohol (HCOL). Aldol condensation of acetal (ACE) and other

140
Z. Tian et al. / Applied Catalysis A: General 506 (2015) 134–142
Fig. 7. Effects of temperature (a) and hydrogen pressure (b) on CAL hydrogenation performance with the PtCo/N-CNTs catalyst (reaction conditions: (a) 0.1 g catalyst, 2.0 MPa,
1.5 h; (b) 0.1 g catalyst, 40 ◦C, 1.5 h).
was generally considered to improve the catalytic activity. That was
because there was a strong interaction between the carbonyl group
and positively charged center (PtOı+x) [36].
increasing selectivity of COL up to 88.5%. This result agreed with
the previous reports, which indicated that the addition of Co
leaded to the acceleration of the activation of C O group. On the
other hand, Ni-doped catalysts showed high selectivity in HCAL
which demonstrated that the hydrogenation of the C C group
was normally favored over Ni-doped catalysts. The promotion
of selectivity was related to the nature of doped metal. More-
over, the selectivity of COL may depend up on the nature of the
exposed crystal faces. It was observed that metal nanoparticles
presented most the (1 1 1) surface in the Co-doped catalyst and
Ni-doped catalyst showed partly the (2 0 0) surface (Fig. S3). Gen-
erally, the PtM(111) surface is compact and prefers the adsorption
of the C O bond, resulting in high selectivity to COL. The open
PtM(2 0 0) surface offers the nearly equivalent opportunity that
the C C and C O bond access to the metal surface, which leads
to the formation of HCAL [37,38]. Among the catalysts tested, the
good conversion was still observed in the presence of functional-
ized catalysts. Though oxygen-functionalized catalyst also showed
the good conversion, a large amount of HCOL was produced. So,
the nitrogen-functionalized catalyst, especially PtCo/N-CNTs, was
Although the high conversion was achieved, some measures
need to be performed due to the poor selectivity. Transition met-
als have been employed widely as co-catalysts in large variety of
hydrogenation reactions. The selectivity and the activity of the
catalysts could be considerably modulated by combining a tran-
sition metal with Pt metals [31], because the differences in the
lattice constants between Pt and 3d transition metals can lead
to the formation of strained PtM bimetallic structures [11]. The
compact lattice surface prefers the adsorption of the C O bond,
due to the steric effect. The adsorption of the C C bond is sub-
stituted and thus hindered. Nevertheless, the open lattice cannot
hindered C C bond adsorption and then multiplies the chances
for C C bond contacting with metal nanoparticles compared with
compact structure [37]. The representatives of transition met-
als, cobalt and nickel, were used in the selective hydrogenation
of CAL in this work, expected to acquire a high selectivity. As
seen from Table 3, addition of Co to Pt/CNTs brought about the

Z. Tian et al. / Applied Catalysis A: General 506 (2015) 134–142
141
Table 3
The influence of addition of transition metals on the selective hydrogenation of CAL.
Catalysts
Conversion (%)
Selectivity
COL
HCAL
HCOL
ACE
PtCo/CNT
PtNi/CNT
PtCo/N-CNT
PtNi/N-CNT
PtCo/O-CNT
PtNi/O-CNT
85.4
82.9
99.7
99.6
99.3
97.9
88.4
25.2
87.9
11.4
72.1
27.4
6.6
60.6
4.1
58.1
3.5
5.0
13.6
7.7
32.0
24.3
32.7
0.3
0.6
0.3
0.5
0.1
1.2
38.7
aReaction conditions: 8.2 mmol CAL, 19.0 mL C2H5OH, 0.1 g catalyst, 70 ◦C, 2.0 MPa, and 1.5 h.
selected as the optimal catalyst in the following studies. Addi-
tionally, the functionalized bimetallic catalysts showed poorer
selectivity of unsaturated product than unfunctionalized catalysts
after Co and Ni doping. This is because the functionalized bimetallic
catalysts had strong activity, which promoted the further hydro-
genation at the high temperature, decreasing the selectivity of
unsaturated product. Therefore, it is guessed that the functional-
ized bimetallic catalysts would show high activity and excellent
selectivity even at the low temperature.
showed a great promotion in the conversion of hydrogenation reac-
tion. However, the presence of the functionalization, especially
for oxygen-functionalization, in the final catalyst was detrimen-
tal to the selective hydrogenation of CAL to COL in long reaction
time. The addition of transition metals, cobalt, into the Pt cata-
lyst significantly increased the selectivity of COL. The research of
effects of temperature over PtCo/N-CNTs showed the high conver-
sion and selectivity was obtained over PtCo/N-CNTs catalyst even
at a low temperature. Thus, a highly active catalyst for the selective
hydrogenation of CAL could be obtained by supporting Pt-based
nanoparticles on nitrogen-functionalized CNTs, suggesting a new
route for the preparation of high-performance catalysts by modi-
fying the surface groups of the CNTs support.
In order to verify the guess, the temperature effect on the CAL
reaction the optimal PtCo/N-CNTs catalyst was studied and the
results are given in Fig. 7a. Wu et al. [39] reported the conversion
could reached 40% at 25 ◦C after reaction for 2 h, when Pt3Co-alloy
nanoparticles was used as efficient catalysts for the selective hydro-
genation of CAL. In this work, it could be seen that 52.1 % conversion
was achieved at 20 ◦C just after 1.5 h (Fig. 7a). This implied that
the PtCo/N-CNTs remained high activity even at the low tempera-
ture. The conversion was improved significantly with temperature
increase during the lower temperature and reached 94.2% at the
temperature of 40 ◦C. The higher selectivity was achieved at a
lower temperature. Elevating the temperature to 70 ◦C, the selec-
tivity to COL slightly decreased and the undesired HCOL selectivity
increased. It was concluded the COL was a less thermodynami-
cally stable product although a desired one and prone to transform
into HCOL under high temperature. To further elevate the temper-
ature to 80 ◦C, the COL selectivity really decreased to 84.7%, while
the quantity of HCOL remarkably increased to 11.8%. It was more
difficult to achieve high selectivity to COL under high temperature.
According to aforementioned temperature effect, the experi-
ments were performed at the temperature of 40 ◦C to research the
effect of hydrogen pressure. It was obviously the hydrogen pressure
could influence activity to a certain extent, but had little effect on
product selectivity. This was because high pressure could enhance
hydrogen solubility, which improved the opportunity that hydro-
gen molecules accessed to the substrate in the reaction medium
[5]. However the improvement was paralleled for the C O bond
and C O bond, resulting in the constant selectivity. As observed,
the conversion increased with the hydrogen pressure increase from
0.5 to 1.5 MPa (Fig. 7b). To further elevate the temperature, the
hydrogen dissolution reached saturation and hydrogen transport
was limited, resulting in no obvious conversion increase.
Acknowledgement
The authors gratefully acknowledge financial support from the
National Natural Science Foundation of China (No. 21203113).
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.apcata.2015.08.
023.
References
[1] S. Bhogeswararao, D. Srinivas, J. Catal. 285 (2012) 31–40.
[2] B.-H. Zhao, J.-G. Chen, X. Liu, Z.-W. Liu, Z. Hao, J. Xiao, Z.-T. Liu, Ind. Eng. Chem.
Res. 51 (2012) 11112–11121.
[3] K. Taniya, H. Jinno, M. Kishida, Y. Ichihashi, S. Nishiyama, J. Catal. 288 (2012)
84–91.
[4] P. Claus, Y. Önal, M. Kishida, Y. Ichihashi, S. Nishiyama, J. Catal. 288 (2012)
84–91.
[5] P. Claus, Y. Önal, Regioselective hydrogenations, in: G. Ertl, H. Knözinger, F.
Schüth, J. Weitkamp (Eds.), Handbook of Heterogeneous Catalysis,
Wiley-VCH, Weinheim, 2008, p. 3308.
[6] X. Yang, D. Chen, S. Liao, H. Song, Y. Li, Z. Fu, Y. Su, J. Catal. 291 (2012) 36–43.
[7] W. Lin, H. Cheng, L. He, Y. Yu, F. Zhao, J. Catal. 303 (2013) 110–116.
[8] Q. Yu, X. Zhang, B. Li, J. Lu, G. Hu, A. Jia, C. Luo, Q. Hong, Y. Song, M. Luo, J. Mol.
Catal. A Chem. 392 (2014) 89–96.
[9] Y. Zhu, F. Zaera, Catal. Sci. Tech. 4 (2014) 955–962.
[10] J. Serrano-Ruiz, A. López-Cudero, J. Solla-Gullón, A. Sepúlveda-Escribano, A.
Aldaz, F. Rodríguez-Reinoso, J. Catal. 253 (2008) 159–166.
[11] Z. Guo, C. Xiao, R.V. Maligal-Ganesh, L. Zhou, T.W. Goh, X. Li, D. Tesfagaber, A.
Thiel, W. Huang, ACS Catal. 4 (2014) 1340–1348.
[12] Z. Liu, X. Tan, J. Li, C. Lv, New J. Chem. 37 (2013) 1350–1357.
[13] Y. Li, P.-F. Zhu, R.-X. Zhou, Appl. Surf. Sci. 254 (2008) 2609–2614.
[14] L.J. Malobela, J. Heveling, W.G. Augustyn, L.M. Cele, Ind. Eng. Chem. Res. 53
(2014) 13910–13919.
[15] M.L. Toebes, F.F. Prinsloo, J.H. Bitter, A.J. van Dillen, K.P. de Jong, J. Catal. 214
(2003) 78–87.
[16] A. Hirsch, Angew. Chem. Int. Ed. 41 (2002) 1853–1859.
[17] G.G. Wildgoose, C.E. Banks, R.G. Compton, Small 2 (2006) 182–193.
[18] Z. Wang, R. Jia, J. Zheng, J. Zhao, L. Li, J. Song, Z. Zhu, ACS Nano. 5 (2011)
1677–1684.
[19] J. Zhang, D. Su, A. Zhang, D. Wang, R. Schlögl, C. Hébert, Angew. Chem. 119
(2007) 7460–7464.
4. Conclusions
The graphite structure of the CNTs was functionalized by
ammonia and hydrogen peroxide without significant damage.
The functionalization not merely increased the density of the N
and O element but also dramatically modified its surface func-
tional groups, which significantly promoted the content of metal
nanoparticles and lead to the small size distribution of Pt par-
ticles on the CNT surface. The N and O species increased the
distribution of positively charged Pt(II) and Pt(IV) species, which
favors the activation of carbonyl group. The functionalized CNTs
[20] K. Chizari, I. Janowska, M. Houllé, I. Florea, O. Ersen, T. Romero, P. Bernhardt,
M.J. Ledoux, C. Pham-Huu, Appl. Catal. A Gen. 380 (2010) 72–80.

142
Z. Tian et al. / Applied Catalysis A: General 506 (2015) 134–142
[21] A. Solhy, B. Machado, J. Beausoleil, Y. Kihn, F. Gonc¸ alves, M. Pereira, J. Órfão, J.
Figueiredo, J. Faria, P. Serp, Carbon 46 (2008) 1194–1207.
[31] Z. Tian, L. Zong, R. Niu, X. Wang, Y. Li, S. Ai, J. Appl. Polym. Sci. 132 (2015)
42057–42065.
[22] R. Chetty, S. Kundu, W. Xia, M. Bron, W. Schuhmann, V. Chirila, W. Brandl, T.
Reinecke, M. Muhler, Electrochim. Acta 54 (2009) 4208–4215.
[23] S. Jiang, L. Zhu, Y. Ma, X. Wang, J. Liu, J. Zhu, Y. Fan, Z. Zou, Z. Hu, J. Power
Sources 195 (2010) 7578–7582.
[32] X. Tuaev, J.P. Paraknowitsch, R. Illgen, A. Thomas, P. Strasser, Phys. Chem.
Chem. Phys. 14 (2012) 6444–6447.
[33] X. Ji, X. Niu, B. Li, Q. Han, F. Yuan, F. Zaera, Y. Zhu, H. Fu, ChemCatChem. 6
(2014) 3246–3253.
[24] M.S. Dresselhaus, A. Jorio, M. Hofmann, G. Dresselhaus, R. Saito, Nano Lett. 10
(2010) 751–758.
[34] X. Xiang, W. He, L. Xie, F. Li, Catal. Sci. Technol. 3 (2013) 2819–2827.
[35] S. Zhang, H. Wang, N. Zhang, F. Kong, H. Liu, G. Yin, J. Power. Sources 197
(2012) 44–49.
[36] J. Breen, R. Burch, J. Gomez-Lopez, K. Griffin, M. Hayes, Appl. Catal. A Gen. 268
(2004) 267–274.
[37] B. Coq, F. Figueras, Coord. Chem. Rev. 178 (1998) 1753–1783.
[38] F. Delbecq, P. Sautet, J. Catal 152 (1995) 217–236.
[39] G.A. Somorjai, Langmuir 7 (1991) 3176–3182.
[25] H.J. Lee, W. Cho, E. Lim, M. Oh, Chem. Commun. 50 (2014) 5476–5479.
[26] Z. Sun, Z. Rong, Y. Wang, Y. Xia, W. Du, Y. Wang, RSC Adv. 4 (2014) 1874–1878.
[27] A. Wang, X.Y. Liu, C.-Y. Mou, T. Zhang, J. Catal. 308 (2013) 258–271.
[28] L. Qie, W.M. Chen, Z.H. Wang, Q.G. Shao, X. Li, L.X. Yuan, X.L. Hu, W.X. Zhang,
Y.H. Huang, Adv. Mater. 24 (2012) 2047–2050.
[29] W. Li, Z. Zhang, B. Kong, S. Feng, J. Wang, L. Wang, J. Yang, F. Zhang, P. Wu, D.
Zhao, Angew. Chem. Int. Ed. 52 (2013) 8151–8155.
[30] L. Wang, P. Yu, L. Zhao, C. Tian, D. Zhao, W. Zhou, J. Yin, R. Wang, H. Fu, Sci.
Rep. 4 (2014).