The structural and electronic promoting effect of nitrogen-doped carbon nanotubes on supported Pd nanoparticles for selective olefin hydrogenation

Article abstract of DOI:10.1039/c3cy00097d

A high-performance Pd catalyst for selective olefin hydrogenation was synthesized by supporting Pd nanoparticles on nitrogen-doped carbon nanotubes (NCNTs). X-ray diffraction, hydrogen chemisorption, transmission electron microscopy and X-ray photoelectron spectroscopy (XPS) were used to characterize Pd supported on NCNTs and nitrogen-free oxygen-functionalized CNTs (OCNTs). The Pd nanoparticles were stabilized on NCNTs with narrower size distribution compared with OCNTs. The XPS analysis revealed that the nitrogen functional groups favor the reduction of Pd on CNTs suggesting an electronic promoter effect. The Pd/NCNT catalyst showed extraordinary catalytic performance in terms of activity, selectivity and stability in the selective hydrogenation of cyclooctadiene, which is related to the structural and electronic promoting effect of the NCNT support.

Full text of DOI:10.1039/c3cy00097d

View Article Online
View Journal
Catalysis
Science&Technology
Accepted Manuscript
This is an Accepted Manuscript, which has been through the RSC Publishing peer
Catalysis
review process and has been accepted for publication.
Science&Technology
Accepted Manuscripts are published online shortly after acceptance, which is prior
to technical editing, formatting and proof reading. This free service from RSC
Publishing allows authors to make their results available to the community, in
citable form, before publication of the edited article. This Accepted Manuscript will
be replaced by the edited and formatted Advance Article as soon as this is available.
www.rsc.org/catalysis
Volume 1 | Number 1 | 2011 | Pages 0001–0100
To cite this manuscript please use its permanent Digital Object Identifier (DOI®),
which is identical for all formats of publication.
More information about Accepted Manuscripts can be found in the
Information for Authors.
Please note that technical editing may introduce minor changes to the text and/or
graphics contained in the manuscript submitted by the author(s) which may alter
content, and that the standard Terms & Conditions and the ethical guidelines
that apply to the journal are still applicable. In no event shall the RSC be held
responsible for any errors or omissions in these Accepted Manuscript manuscripts or
any consequences arising from the use of any information contained in them.
ISSN 2044-4753
www.rsc.org/catalysis
Registered Charity Number 207890

Page 1 of 9
Catalysis Science & Technology
View Article Online
Graphical and textual abstract
Pd on NCNTs showed a narrower particle size distribution compared with OCNTs, and
offered a remarkable catalytic performance in selective hydrogenation.
1
0
0
0
0
0
.0 Pd/NCNT
Pd/OCNT
.8
.6
.4
.2
.0
1
,5-COD
COE
1
,3-COD
1
,5-COD
COE
COA
H2
H
2
k
3
k
5
k
1
k4
k
2
H
2
0
2
4
6
8
10 0
2
4
6
8
10
Time on stream (h)
Time on stream (h)
1
,3-COD

Catalysis Science & Technology
Page 2 of 9
Dynamic Article Links ►
Journal Name
View Article Online
Cite this: DOI: 10.1039/c0xx00000x
www.rsc.org/xxxxxx
ARTICLE TYPE
The structural and electronic promoting effect of nitrogen-doped carbon
nanotubes on supported Pd nanoparticles for selective olefin
hydrogenation
a
a
b
a
*a
Peirong Chen, Ly May Chew, Aleksander Kostka, Martin Muhler and Wei Xia
5
Received (in XXX, XXX) Xth XXXXXXXXX 20XX, Accepted Xth XXXXXXXXX 20XX
DOI: 10.1039/b000000x
A highꢀperformance Pd catalyst for selective olefin hydrogenation was synthesized by supporting Pd
nanoparticles on nitrogenꢀdoped carbon nanotubes (NCNTs). Xꢀray diffraction, hydrogen chemisorption,
transmission electron microscopy and Xꢀray photoelectron spectroscopy (XPS) were used to characterize Pd
supported on NCNTs and nitrogenꢀfree oxygenꢀfunctionalized CNTs (OCNTs). The Pd nanoparticles were
stabilized on NCNTs with narrower size distribution compared with OCNTs. The XPS analysis revealed that
the nitrogen functional groups favor the reduction of Pd on CNTs suggesting an electronic promoter effect.
The Pd/NCNT catalyst showed extraordinary catalytic performance in terms of activity, selectivity and
stability in the selective hydrogenation of cyclooctadiene, which is related to the structural and electronic
promoting effect of the NCNT support.
1
0
15
16
activity of Pd nanoparticles. However, the synthesis of highly
dispersed supported Pd nanoparticles with a high performance in
4
4
5
5
6
0
5
0
5
0
1
. Introduction
selective olefin hydrogenation remains a substantial challenge in
Carbonꢀsupported palladium (Pd/C) is one of the most important
catalysts in hydrogenation reactions. Pd as active phase has superior
activity and selectivity, while carbon as support has outstanding
resistance to demanding pH and temperature conditions. Numerous
studies have been devoted to clarifying the reaction mechanism and
17,18
practice.
In addition, as commonly observed in hydrogenation,
the chemical state of Pd has a significant influence on the
2
2
3
3
0
19–21
performance of Pd/C catalysts.
Under reaction conditions, the
1
,2
oxidation state and even the nearꢀsurface composition of the Pd
nanoparticles can be altered, rendering the stabilization of Pd in the
3
–5
further improving the catalytic performance of Pd/C catalysts.
3,22
most suitable oxidation state difficult.
Recently, carbon nanomaterials, and especially carbon nanotubes
(CNTs), used as support for Pd were found to achieve extraordinary
catalytic performance in selective hydrogenation in terms of
selectivity and stability.
modified by oxygen and nitrogen functionalization.
surface functional groups can be used to tune the catalytic
performance of supported Pd nanoparticles in hydrogenation
reactions. Indeed, several studies have focused on the modification
of the CNT support in Pd/CNT catalysts applied in the selective
Recent studies showed that highly dispersed metal nanoparticles
can be efficiently immobilized on nitrogenꢀdoped CNTs (NCNTs)
5
0
5
due to the strong anchoring ability of nitrogenꢀcontaining functional
1
,6–9
The surface properties of CNTs can be
23,24
groups on the CNT surface.
It is not clear so far whether the
10–12
These
enhanced metalꢀsupport interactions can affect the chemical state of
the supported metal species and their catalytic performance. Here,
we show that supporting Pd nanoparticles on NCNTs can stabilize
the Pd particle size and influence the oxidation state of Pd. The
obtained Pd/NCNT catalyst was tested in selective olefin
hydrogenation using 1,5ꢀcyclooctadiene as a model reactant and was
found to be significantly superior to Pd catalysts supported on
oxygenꢀfunctionalized CNTs (Pd/OCNTs) in terms of activity and
selectivity under continuous flow conditions.
13–15
hydrogenation of benzene, cyclooctadiene and anthracene.
It is known that the size of the Pd particles has a significant
influence on the catalytic performance with nanosized Pd particles
being highly active in C=C bond hydrogenation. The relatively easy
formation of weakly bound subsurface hydrogen, which determines
the hydrogenation activity, was found to be the origin of the high
2
. Experimental
2
.1. Sample synthesis
a
Laboratory of Industrial Chemistry, RuhrꢀUniversity Bochum
®
CNTs (Baytubes C 150 P, outer mean diameter 13ꢀ16 nm, inner
mean diameter 4 nm, loose agglomerate size 0.1ꢀ1 mm) supplied by
Bayer AG (Leverkusen, Germany) were first purified by washing in
Universitätsstr. 150, 44780 Bochum, Germany. Fax: (+)49
2
34 32 14115; Eꢀmail: wei.xia@techem.rub.de
MaxꢀPlanckꢀInstitut für Eisenforschung, MaxꢀPlanckꢀStrasse
, 40237 Düsseldorf, Germany
b
65
ꢀ
1
1
diluted HNO solution (1.5 mol l ) in order to remove the residual
3
‡
Electronic supplementary information (ESI) available: details
of the kinetic modeling using Athena software, XPS, XRD
and additional catalytic test results.
growth catalysts. The synthesis of OCNTs and NCNTs was
10,11
described elsewhere.
Briefly, OCNTs were prepared by treating
the purified CNTs in HNO vapor at 200 °C for 48 h, and the NCNTs
3
This journal is © The Royal Society of Chemistry [year]
[journal], [year], [vol], 00–00 | 1

Page 3 of 9
Catalysis Science & Technology
View Article Online
were subsequently obtained by the postꢀtreatment of OCNTs in
flowing NH (10 vol% NH in He) at 400 °C for 6 h.
positioning the C 1s peak at 284.5 eV. For the deconvolution of the
60 spectra of the functionalized CNTs all the parameter settings were
consistent with our previous reports.
3
3
14,27
The deposition of Pd nanoparticles on OCNTs and NCNTs was
2
5
achieved by a colloidal method. In detail, 10 mg of anhydrous
2
.3. Catalytic tests
5
0
5
palladium (II) acetate (Pd(OAc) ) were dissolved in 50 ml of
2
anhydrous methanol in a flask followed by continuous stirring for 10
min. Subsequently, OCNTs or NCNTs (0.5 g) were added and mixed
for 10 min with the precursor solution under stirring. Then, the 65 cyclooctadiene (1,5ꢀCOD), which is commonly used to obtain
Catalytic activity was tested in the gasꢀphase olefin hydrogenation at
50 °C and ambient pressure in both batch and continuous mode. 1,5ꢀ
solvent was removed in a rotary evaporator. The obtained powders
cyclooctene (COE) for the metathesis production of metathetical
2
8
1
were dried in air at 60 °C for 12 h. It is known that Pd nanoparticles
polymer building blocks, was selected as a model reactant. Prior to
each test, an activation at 50 °C with 10 vol% H in N using a total
2
5
of very small size can be obtained by this method.
The reduction of the asꢀprepared samples was performed in a
quartz tube reactor. The samples were first heated to 200 °C in He, 70 previously reduced catalyst. Typically, 5 mg of Pd/CNT catalyst
2
2
ꢀ
1
flow of 200 ml min was conducted in the reactor with the
ꢀ
1
then reduced with 10 vol% H in He (100 ml min ) for 2 h at that
reduced at 200 °C dispersed in 1g quartz sand were loaded in a glass
reactor for activity evaluation.
2
1
temperature. Then, the samples were cooled in He to room
temperature and slowly exposed to air. The samples are designated
as Pd/OCNT and Pd/NCNT.
A constant flow rate of 1,5ꢀCOD was achieved by twoꢀstage
vaporizing liquid 1,5ꢀCOD with N2 as carrier gas. The feed
composition for the continuous mode was 1 vol% 1,5ꢀCOD, 2 vol%
H and 97 vol% N . After a stable 1,5ꢀCOD flow was achieved under
75
2
.2. Characterization
2
2
Inductively coupled plasma optical emission spectroscopy (ICPꢀ
OES, PU701 UNICAM) was used to determine the achieved Pd
loading on the CNTs. Powder Xꢀray diffraction (XRD) patterns were
byꢀpass conditions, the feed gas was passed to the reactor to start the
catalytic experiments. By separating the feed gas supply, a batch
system with a total volume of 358 ml can be obtained and used for
20
25
30
35
40
recorded in a Philips X’Pert MPD system with Cu Kα radiation. 80 kinetics investigations. A pump was employed to circulate the
Scanning transmission electron microscopy (STEM) and
transmission electron microscopy (TEM) measurements were carried
out with a Jeol JEMꢀ2200FS instrument operated at 200 kV. The
specimens for STEM and TEM were prepared by ultrasonically
dispersing the powder samples in ethanol and dropping the 85 continuous operation mode, 20 mg of a commercial Pd/A O
dispersion on a carbonꢀcoated Au grid.
Hydrogen chemisorption measurements were performed using a
Quantachrome Autosorb Automated Gas Sorption system. Before the
measurement, the asꢀprepared sample without reduction was dried in
reactants in the batch operation mode. A higher H fraction of 6
2
vol% was applied in order to minimize the influence of the H partial
2
pressure. The evolved gas mixture during the reaction was monitored
by an onꢀline gas chromatograph (Shimadzu GCꢀ2014). In the
2
3
hydrogenation catalyst (0.5 wt% Pd loading, Engelhard BASF,
Germany) diluted in quartz sand were also tested for comparison.
The conversion of 1,5ꢀCOD (XCOD), the selectivity to COE (SCOE
)
and 1,3ꢀCOD (S1,3ꢀCOD) were calculated according to equations (2),
He at 110 °C for 1 h, and reduced in flowing H (10 vol% H in He, 90 (3) and (4), respectively. In the equations, C0,COD is the initial
2
2
ꢀ
1
1
00 ml min ) at 200 °C for 2 h. After reduction the catalyst was
degassed at 10 mbar and 200 °C for 2 h in order to eliminate
chemisorbed hydrogen and water. The isotherms were measured at
concentration of 1,5ꢀCOD, while CCOD, CCOE and C1,3ꢀCOD represent
the measured concentrations of 1,5ꢀCOD, COE and 1,3ꢀCOD at
reaction time t, respectively.
−6
3
0 °C and at very low H2 pressure to avoid the formation of
palladium hydride. The H/Pd ratios are based on the total amount of
adsorbed hydrogen rather than the irreversibly adsorbed hydrogen as
usually used at zero pressure determined by extrapolation of the
linear part of the adsorption isotherm. Estimated average particle
sizes and dispersions were based on a spherical geometry and an
adsorption stoichiometry of H/Pd = 1. The average Pd particle size d
was calculated from
C0, COD −CCOD
95
X
COD
=
×100%
×100%
(2)
(3)
C0, COD
C
COE
S
COE
=
C
0, COD −CCOD
6
M
ρ
surf
C1,3−COD
d =
(1)
S1,3−COD
=
×100%
(4)
ρPd N D
C0,COD −CCOD
A
45
50
55
where d is the average Pd particle size in nm, D is the dispersion
ꢀ
1
(
H/Pd), M is the atomic weight of Pd of 106.42 g mol , ρ is the Pd
surf
2
surface site density of 12.7 Pd atoms per nm , ρ is the Pd metal
density of 12.02 g cm , and N is the Avogadro constant amounting
to 6.02×10 mol yielding d = 1.12/D (nm).
1,5-COD
COE
COA
Pd
ꢀ
3
H2
k3
H2
A
23
ꢀ1
26
k5
Xꢀray photoelectron spectroscopy (XPS) measurements were
performed in an ultrahigh vacuum setup with a high resolution
Grammadata Scienta SES 2002 analyzer and a monochromatic Al
Kα Xꢀray source (1486.6 eV, operated at 14.5 KV and 35 mA). The
spectra were taken at 200 eV pass energy (high pass energy mode) at
k1
k4
k2
H2
1,3-COD
ꢀ
10
ꢀ10
a pressure between 3.5
during the measurements were mediated by applying a flood gun 100
SPECS). The CASA XPS program was employed in the analysis of
XPS data. Calibration of the measured data was achieved by
×
10 and 6
×
10 mbar. Charging effects
Scheme 1. Reaction network of 1,5ꢀCOD hydrogenation
(
2
|
Journal Name, [year], [vol], 00–00
This journal is © The Royal Society of Chemistry [year]

Catalysis Science & Technology
Page 4 of 9
View Article Online
Rate constants were estimated by simulating the experimental
data in the batch reactor mode with the Athena visual studio
software. The reactions shown in Scheme 1 were assumed to be all
detected, which are identified as the Pd(111), Pd(200) and Pd(220)
reflections, respectively. In contrast, only a weak and broad
contribution at around 40° corresponding to Pd(111) was observed
2
9
firstꢀorder. More information about the modeling can be found in 55 for Pd/NCNT, and the Pd(200) and Pd(220) reflections were not
5
the supplementary information (ESI).
detected. The Pd loading on NCNTs was slightly lower than that on
OCNTs, but this difference is not enough to cause such a significant
difference in the XRD patterns. Therefore, the diffraction patterns
indicate different particle sizes.
3
. Results
3
.1. Sample synthesis and characterization
6
6
7
7
8
0
5
0
5
0
The dispersion as well as the particle size of the Pd nanoparticles
on OCNTs and NCNTs was further evaluated by H chemisorption.
Both the OCNTs and NCNTs were used as substrates for Pd
nanoparticles, which were synthesized by a colloidal deposition
method using Pd(II) acetate as precursor and methanol as stabilizer
and reductant. The Pd loadings on OCNTs and NCNTs were
determined by ICPꢀOES to be 0.89 wt% and 0.63 wt%, respectively
2
H/Pd ratios (H adsorbed on surface Pd sites) were used to estimate
the metal dispersions and to calculate the average Pd particle sizes.
The results are summarized in Table 1. It can be clearly seen that a
higher Pd dispersion was obtained with NCNTs (0.37) than with
OCNTs (0.23) due to a smaller average Pd particle size on NCNTs
compared with 4.9 nm on OCNTs.
1
1
2
0
5
0
25
(
Table 1). The different Pd loadings on OCNTs and NCNTs can be
related to the differences in the total amount of anchoring sites
1
4
generated by the oxygen and nitrogen functionalization. It has to be
noted that not all the oxygen and nitrogen groups contribute equally
to the anchoring of metal species. Hence, a quantitative correlation of
the functional groups and metal loading cannot be easily derived.
Furthermore, the surface chemistry of 0the CNTs, e.g.,
a
c
e
b
d
f
3
hydrophobicity, polarity and proton affinity determining the
3
1
adsorption of charged metal precursor in solution was changed by
the nitrogen functionalization and therefore should also be taken into
consideration.
25
Table 1 Properties of the reduced Pd/CNT samples based on ICPꢀOES, H
chemisorption and TEM
2
ICPꢀOES
H chemisorption
2
TEM
sample
Pd loading
d
Dispersion
(nm)
d
(nm)
(%)
Pd/OCNT
Pd/NCNT
0.89
0.23
4.9
2.7±0.8
85
0.63
0.37
3.0
2.7±0.5
20
15
10
5
0
50
Pd(111) Pd(200)
Pd(220)
9
0
4
0
0
30
35
40
3
9
5
20
10
(
b)
0
0
1
2
3
4
5
6
0
1
2
3
4
5
6
Size (nm)
Size (nm)
1
00
Fig. 2 Representative STEM images of (a) Pd/OCNT and (b) Pd/NCNT;
representative TEM images of (c) Pd/OCNT and (d) Pd/NCNT; particle size
distributions of (e) Pd/OCNT and (f) Pd/NCNT. Both Pd/OCNT and
(
a)
20
40
60
80
2
theta (°)
2
Pd/NCNT were reduced at 200 °C in H .
Fig. 1 XRD patterns of (a) Pd/OCNT and (b) Pd/NCNT after reduction in H
at 200 °C
2
105
45
Further investigations using STEM and TEM indicate that all the
particles are located on the outer surface of the CNTs (Fig. 2, Fig. S1
and S2 in ESI), and confirmed the differences in Pd particle size
The XRD patterns of Pd/OCNT and Pd/NCNT samples reduced
at 200 °C in H are displayed in Fig. 1. The peaks at around 26° and
disclosed by XRD and H chemisorption. As can be seen from the
2
2
4
2° detected in both samples are the (002) and (111) reflections of 110 STEM images in Fig. 2 (a) and (b), highly dispersed Pd
50
the hexagonal graphite structure of the CNTs, respectively. For the
sample Pd/OCNT, three weak peaks at 40°, 47° and 68° were
nanoparticles were detected in both Pd/OCNT and Pd/NCNT
samples. By comparing different STEM images we noticed that large
This journal is © The Royal Society of Chemistry [year]
Journal Name, [year], [vol], 00–00 | 3

Page 5 of 9
Catalysis Science & Technology
View Article Online
Pd nanoparticles can often be observed on OCNTs, whereas on
NCNTs large Pd particles were hardly detected. Representative TEM
there are 4.8 at% oxygen on OCNTs, and 1.6 at% nitrogen as well as
2.7 at% oxygen on NCNTs.
images of Pd/OCNT and Pd/NCNT samples after reduction in H at 25
Asꢀprepared and reduced samples were measured and compared
to monitor the Pd oxidation state. It has to be pointed out that the
2
2
00 °C are shown in Fig. 2.(c) and (d). It can be seen that the Pd
5
nanoparticles on both supports are smaller than 5 nm. Statistical
analysis was carried out to evaluate the particle size distributions.
The obtained histograms of Pd/OCNT and Pd/NCNT are shown in
reduced samples had been treated in H at 200 °C and transferred in
2
air before entering the ultraꢀhigh vacuum XPS chambers. However,
0
no change of the Pd oxidation state is expected because Pd is stable
Fig. 2 (e) and (f), respectively. Most of the particles were found to be 30 under ambient conditions. The obtained Pd spectra were normalized
in the range of 2 nm to 4 nm for both samples. However, the particle
size distribution is clearly narrower for NCNTs than for OCNTs.
Although the Pd nanoparticles on both samples have the same
average particle size of 2.7 nm, the standard deviation of the average
to the corresponding C 1s spectra for quantitative comparison (Fig.
3). Three major contributions indicated by the dashed lines can be
distinguished in the Pd 3d5/2 spectra of the asꢀprepared and 200 °C
10
21,32
reduced Pd/OCNT and Pd/NCNT samples.
The contributions at
Pd particle size is higher for Pd/OCNT (0.8) than for Pd/NCNT (0.5). 35 ca. 335.5 eV and ca. 336.5 eV can be assigned to metallic Pd and to
2
+
21
Pd as in PdO, respectively, and the contribution at ca. 338 eV
32
originates from Pd(OAc)2, that is, the precursor used for synthesis.
The presence of the Pd(OAc) precursor in the asꢀprepared samples
(
a) Pd/OCNT
2
is ascribed to the incomplete reduction with methanol during the
colloidal synthesis. Significant variations regarding the intensities of
the different Pd species can be found in the spectra of the asꢀprepared
Pd/OCNT and Pd/NCNT samples (Fig. 3). The contribution at ca.
3
d3/2
3d5/2
40
3
38 eV is strong in the spectra of asꢀprepared Pd/OCNT, whereas
only a weak contribution was observed at the binding energy of
335.5 eV (Fig. 3 (a)). This observation indicates that in the asꢀ
prepared Pd/OCNT sample a large amount of Pd exists in the form of
as-prepared
4
5
5
6
6
7
5
0
5
0
5
0
2+
0
Pd(OAc) or Pd rather than as Pd . Clear differences were observed
2
when using NCNTs as support, where a much lower intensity at
higher binding energies and a significantly higher intensity at ca.
335.5 eV was observed (Fig. 3 (b)) indicating a much higher degree
of reduction in the asꢀprepared state.
0
Reduction at 200 °C caused a significant increase of Pd in
2
00 °C reduced
Pd/OCNT, and the contribution attributed to Pd(OAc)2 decreased
strongly. In contrast, the reduction did not cause significant changes
in the Pd 3d XP spectra of Pd/NCNT. The comparison of the Pd XP
3
46
344
342
340
338
336
334
BE (eV)
3
d spectra of the two reduced samples indicates that Pd is still in a
more reduced state on NCNTs than on OCNTs. The corresponding O
s and N 1s spectra are shown in Fig. S3 and S4 (ESI). A decrease of
1
(
b) Pd/NCNT
d3/2
the oxygen concentration after reduction was observed on both
samples. Nitrogen species in Pd/NCNT did not show significant
changes due to the reduction at 200 °C.
Atomic ratios of Pd to carbon for the asꢀprepared and 200 °C
reduced Pd/CNT samples were derived from the XPS analysis. The
overall Pd intensity is higher on the asꢀprepared Pd/OCNT sample
3
3d5/2
ꢀ
3
ꢀ3
(3.0 x 10 ) than on the asꢀprepared Pd/NCNT (2.8 x 10 ) obviously
due to the higher Pd loading. After reduction at 200 °C a decrease of
as-prepared
ꢀ
3
the Pd/C ratio was observed for both Pd/OCNT (2.5 x 10 ) and
ꢀ
3
Pd/NCNT (2.7 x 10 ). This decrease originates from the sintering of
Pd nanoparticles caused by the reduction process, as the dispersion
of supported metal particles influences the corresponding intensity in
33
XPS to a great extent. The Pd/C ratio of Pd/OCNT amounting to
ꢀ
3
ꢀ3
2
.5 x 10 is lower than that of Pd/NCNT amounting to 2.7 x 10
2
00 °C reduced
after reduction indicating more severe sintering of Pd nanoparticles
on the OCNT support in agreement with the XRD and TEM results.
3
46
344
342
340
338
336
334
BE (eV)
75
3.2. Catalytic tests
15
The activities of the Pd/OCNT and Pd/NCNT catalysts reduced at
Fig. 3 XP Pd 3d spectra of (a) Pd on OCNTs and (b) Pd on NCNTs asꢀ
prepared and after reduction in H at 200 °C.
2
00 °C were examined in the hydrogenation of 1,5ꢀCOD, which is
2
converted to 1,3ꢀCOD by isomerization, COE by partial
hydrogenation, and cyclooctane (COA) by full hydrogenation
To examine the surface chemistry of the supported Pd 80 (Scheme 1). The catalytic tests were performed in the batch and in
20
nanoparticles in the presence of different functional groups, the
Pd/OCNT and Pd/NCNT samples were investigated by XPS. The
quantitative XPS analysis of the support materials disclosed that
the continuous mode, and all the compounds shown in Scheme 1
were detected by onꢀline GC.
4
|
Journal Name, [year], [vol], 00–00
This journal is © The Royal Society of Chemistry [year]

Catalysis Science & Technology
Page 6 of 9
View Article Online
Batch operation was carried out in order to obtain kinetic results
of the hydrogenation reactions over Pd/OCNT and Pd/NCNT. The
temporal evolution of 1,5ꢀCOD, 1,3ꢀCOD, COE and COA were
found to be quite similar over Pd/OCNT (Fig. 4 (a)) and Pd/NCNT
(Fig. 4 (b)). In both tests 1,3ꢀCOD was formed initially and
subsequently consumed yielding COE. The COE concentration was
found to increase steeply at the initial stage of the reaction due to fast
isomerization and subsequent hydrogenation. The COE
concentration reached a maximum as soon as the full conversion of
1,5ꢀCOD and 1,3ꢀCOD was achieved, and decreased afterwards to
form COA by full C=C bond hydrogenation (Fig. 4). COA formation
was observed to be very slow at the initial stage of the reaction over
Pd/OCNT, and to be accelerated after all the 1,5ꢀCOD and 1,3ꢀCOD
had been consumed. The hydrogenation of COE is known to be
inhibited, because Pd catalysts adsorb 1,5ꢀCOD and 1,3ꢀCOD much
more strongly than COE. However, the inhibition of COA
formation is not obvious in case of Pd/NCNT. Comparing the
performance of both catalysts it is clear that the Pd nanoparticles
supported on NCNTs had a significant higher activity than those
supported on OCNTs despite a lower Pd loading on NCNTs. The full
conversion of 1,5ꢀCOD was achieved within 4 min for Pd/NCNT,
but needed 12 min for Pd/OCNT. In addition, the time to achieve full
hydrogenation over Pd/NCNT (18 min) was significantly shorter
than that over Pd/OCNT (76 min).
two Pd/CNT catalysts confirmed the superior activity of Pd
supported on NCNTs (Table S3ꢀS5 and Fig. S5, ESI).
1.2
5
0
5
0
5
(a)
1
.0
.8
0
1
1
,5-COD
,3-COD
0.6
1
1
2
2
COE
COA
0.4
0.2
0.0
3
4
0
20
40
60
80
Time (min)
5
5
1.2
(b)
1.0
0.8
0.6
0.4
0.2
0.0
1
,5-COD
1,3-COD
COE
COA
Table 2 Estimated rate constants of 1,5ꢀCOD hydrogenation over Pd/OCNT
ꢀ1
and Pd/NCNT catalysts (s ) obtained by modeling the experimental data
assuming firstꢀorder reactions.
sample
Pd/OCNT
0.256
0
Pd/NCNT
1
0
5
10
15
20
k1
k2
k3
k4
k5
Time (min)
0
Fig. 4 Measured (solid lines) and simulated (dashed lines) profiles of 1,5ꢀ
COD, 1,3ꢀCOD, COE and COA during the hydrogenation over (a) Pd/OCNT
and (b) Pd/NCNT (5 mg catalyst in 1g quartz sand, 1 vol% 1,5ꢀCOD and 6
0.006
0.027
0.001
0.008
0.119
0.008
60
vol% H
2
in the feed gas, N as carrier gas, 50 °C, ambient pressure). Offꢀsite
2
ꢀ1
reduction (200 °C, 10 vol% H in He with a total flow of 100 ml min , 2 h)
2
and in situ activation (50 °C, 10 vol% H
min , 20 min) were performed with the catalysts prior to the catalytic tests.
2
in N
2
with a total flow of 200 ml
ꢀ1
30
The experimental results were modeled using the Athena visual
studio software, which generated the simulated profiles of all the
products (Fig. 4) and the reaction rate constants for each catalyst
1
00 (a)
(b)
XCOD
(
Table 2). The experimental and simulated results show a good
SCOE
80
35
40
45
50
agreement in general (Table S1 and S2, ESI), which confirms the
assumptions about the reaction pathway and the reaction order. It can
S1,3-COD
be seen from Table 2. that k , which is the rate constant of the
60
40
20
0
1
reaction from 1,5ꢀCOD to 1,3ꢀCOD, and k , which is the rate
4
constant of the reaction from 1,3ꢀCOD to COE, are much larger than
the other rate constants suggesting that the isomerization of 1,5ꢀCOD
to 1,3ꢀCOD as intermediate is an indispensable step for the
hydrogenation reaction for both catalysts, and that the main reaction
route is from 1,5ꢀCOD to 1,3ꢀCOD and then to COE. The small
values of k , which is the rate constant of the hydrogenation of 1,5ꢀ
3
0
2
4
6
8
10 0
2
4
6
8
10
COD to COE, and k , which is the rate constant of the hydrogenation
5
Time on stream (h)
of COE to COA reflect the corresponding slow reaction steps, while
the zero value of k implies that the reverse isomerization reaction 65
2
from 1,3ꢀCOD to 1,5ꢀCOD did essentially not occur. Also the rate
constants show the difference in activities of Pd/OCNT and
Pd/NCNT: k and k of Pd/NCNT are higher than the corresponding
Fig. 5 1,5ꢀCOD conversion (XCOD), COE selectivity (SCOE) and 1,3ꢀCOD
selectivity (S1,3ꢀCOD) over (a) Pd/NCNT and (b) Pd/OCNT (5 mg catalyst in 1g
quartz sand, 1 vol% 1,5ꢀCOD and 2 vol% H
2 2
in the feed gas, N as carrier
1
4
gas, 50 °C, ambient pressure).
rate constants of Pd/OCNT indicating a higher catalytic activity of
Pd/NCNT in the hydrogenation of 1,5ꢀCOD. Additional tests using
This journal is © The Royal Society of Chemistry [year]
Journal Name, [year], [vol], 00–00 | 5

Page 7 of 9
Catalysis Science & Technology
View Article Online
selectivity were achieved with Pd/NCNT (5 mg, 0.63 wt% Pd)
compared with Pd/OCNT (5 mg, 0.89 wt% Pd) as can be seen in Fig.
5. For comparison, a commercial Pd/Al O hydrogenation catalyst
6
5
4
3
2
1
0
0
0
0
0
0
0
2
2
3
3
0
5
0
5
2
3
P d/N C N T
P d/A l O 3
2
(20 mg, 0.5 wt% Pd loading, Engelhard BASF, Germany) was tested
for a prolonged timeꢀonꢀstream of 50 h (Fig. 6). A higher 1,5ꢀCOD
conversion (XCOD = 35%) and COE selectivity (SCOE = 46%) were
achieved using the Pd/NCNT catalyst compared with the commercial
Pd/Al O catalyst (X = 25% and SCOE = 35%). The 1,3ꢀCOD
COD
2
3
selectivity (S1,3ꢀCOD) of Pd/NCNT (54%) was lower than that of the
8
6
4
2
0
0
0
0
0
0
0
0
0
0
commercial Pd/Al O catalyst (ca. 65%). These results demonstrate
2
3
that the Pd/NCNT catalyst with lower Pd loading outperformed the
commercial Pd/Al O catalyst to a great extent.
2
3
The Pd/NCNT catalyst after the longꢀterm hydrogenation at
50 °C for 50 h was examined by STEM and TEM (Fig. 7
)
.
8
6
4
2
Agglomeration of Pd nanoparticles was not detected by these studies.
The size distribution of the Pd nanoparticles displayed in Fig. 7 (c)
showed almost no change as compared to that of Pd/NCNT before
reaction implying that the Pd nanoparticles on NCNTs are stable
under the reaction conditions even for a prolonged time on stream.
0
10
20
30
40
50
T im e on stream (h)
4. Discussion
The beneficial effect of surface nitrogen groups on the metal
dispersion on carbon nanotubes has been reported in several studies.
Fig. 6 1,5ꢀCOD conversion (XCOD), COE selectivity (SCOE) and 1,3ꢀCOD
selectivity (S1,3ꢀCOD) of Pd/NCNT and commercial Pd/Al
,5ꢀCOD hydrogenation reaction (5 mg Pd/NCNT or 20 mg Pd/Al
2
O
3
in the selective 40 It was proposed that around the Nꢀrich sites some individual sections
1
2
O
3
in 1g
35
formed, which can efficiently anchor metal particles. A recent
5
2 2
quartz sand, 1 vol% 1,5ꢀCOD and 2 vol% H in the feed gas, N as carrier
gas, 50 °C, ambient pressure)
study revealed that Nꢀdoped CNTs possessed a higher amount of
surface nucleation sites allowing efficient anchoring of metal active
24
phase. As a consequence, a smaller particle size and a higher
dispersion of supported metals such as Pd can be achieved on
NCNTs than on pristine CNTs. The TEM, XRD, chemisorption and
XPS results confirm that the surface nitrogen species can indeed
enhance the dispersion of Pd. The nitrogen functionalization was
performed at 400 °C, and the resulting nitrogen species are known to
be more stable than surface oxygen groups at higher
4
5
5
6
6
7
5
0
5
0
5
0
a
b
1
1,27
temperatures.
Consequently, they can stabilize Pd nanoparticles
efficiently against sintering during the reduction.
In addition, it was demonstrated that the nitrogen species on the
CNT surface have a significant influence on the chemical state of the
supported Pd nanoparticles. Pd tended to be in a more reduced state
in asꢀprepared Pd/NCNT than in asꢀprepared Pd/OCNT (Fig. 3). On
the one hand, the adsorption of metal precursors is different on
4
3
2
1
0
0
0
0
0
c
31
surface oxygen or nitrogen species, which may influence the
reduction of Pd during synthesis using methanol as mild reducing
agent. On the other hand, the supported Pd nanoparticles on NCNTs
are assumed to interact with the nitrogen groups acting as electron
30
donor. Another reason for the more reduced state of Pd on NCNTs
can be related to PdꢀN coordination complexes formed on carbon
36
surfaces with nitrogen groups leading to a lower binding energy of
3
7
Pd on the NCNT support. As a result of these effects, Pd supported
on NCNTs is in a lower oxidation state than on OCNTs before
0
1
2
3
4
5
6
reduction. Even after reduction in H at 200 °C Pd is in a relatively
2
Size (nm)
more reduced state on NCNTs than on OCNTs pointing to an
electronꢀdonating effect of the nitrogen species. Stronger metalꢀ
support interactions as well as a stronger electronic interaction on
10
Fig. 7 Representative STEM (a) and TEM (b) images and particle size
distribution (c) of Pd/NCNT after 50 h of 1,5ꢀCOD hydrogenation reaction
2
4,36
NCNTs than on OCNTs have been suggested in earlier studies.
The nitrogen species can be considered as both structural and
electronic promoters, because both the particle size and the oxidation
state of the Pd nanoparticles were strongly influenced. The
Catalytic studies in the continuous mode confirmed the superior
15
performance of Pd/NCNT. Although both catalysts showed a high 75 promoting effect of the NCNT support is manifested in the
catalytic stability without detectable deactivation within the test
period of 10 h, a higher 1,5ꢀCOD conversion and a higher COE
outstanding catalytic performance of the supported Pd nanoparticles.
2
4,38
On the one hand, as already reported in previous studies,
more
6
|
Journal Name, [year], [vol], 00–00
This journal is © The Royal Society of Chemistry [year]

Catalysis Science & Technology
Page 8 of 9
View Article Online
dispersed Pd nanoparticles are more favorable for the shift and
hydrogenation of C=C bonds. Correspondingly, Pd nanoparticles on
NCNTs possess higher activity in the isomerization of 1,5ꢀCOD to
7
Y. Xiang, Y. Lv, T. Xu, X. Li and J. Wang, J. Mol. Catal. A:
Chem., 2011, 351, 70.
C. Amorim, G. Yuan, P. Patterson and M. Keane, J. Catal.,
2005, 234, 268.
8
1
,3ꢀCOD as indicated by the high k value obtained in the kinetic 60
1
5
0
5
0
simulation, and in turn a higher hydrogenation activity as indicated
9
C. PhamꢀHuu, N. Keller, M. J. Ledoux, L. J. Charbonniere and
R. Ziessel, Chem. Commun., 2000,19, 1871.
W. Xia, C. Jin, S. Kundu and M. Muhler, Carbon, 2009, 47,
919.
S. Kundu, W. Xia, W. Busser, M. Becker, D. A. Schmidt, M.
Havenith and M. Muhler, Phys. Chem. Chem. Phys, 2010, 12,
4351.
S. Sahin, P. MäkiꢀArvela, J.ꢀP. Tessonnier, A. Villa, S. Reiche,
S. Wrabetz, D. Su, R. Schlögl, T. Salmi and D. Y. Murzin,
Appl. Catal., A, 2011, 408, 137.
B. Yoon, H.B. Pan and C. M. Wai, J. Phys. Chem. C, 2009,
113, 1520.
W. Xia, O. F.ꢀK. Schlüter, C. Liang, M. W. van den Berg, M.
Guraya and M. Muhler, Catal.Today, 2005, 102-103, 34.
B. Yoon and C. M. Wai, J. Am. Chem. Soc, 2005, 127, 17174.
A. M. Doyle, S. K. Shaikhutdinov, S. D. Jackson and H.ꢀJ.
Freund, Angew. Chem. Int. Ed., 2003, 42, 5240.
A. Binder, M. Seipenbusch, M. Muhler and G. Kasper, J.
Catal., 2009, 268, 150.
by the high value of k . On the other hand, the presence of different
4
surface nitrogen species on NCNTs seems to modify the supported
Pd nanoparticles electronically as suggested by the XPS results and
10
to contribute to the high selectivity and durability in selective 65 11
1
9,39
1
1
2
catalytic hydrogenation reactions.
Additionally, the polarity can
40
influence the adsorption of reactant molecules in catalysis. In this
work, the polarity of the CNT surface was altered significantly by the
removal of surface oxygen groups during nitrogen functionalization,
which may also influence the catalytic performance. Furthermore, 70
some key factors in olefin hydrogenation catalysis, such as the
formation of the subsurface carbon and the diffusion of subsurface
12
13
14
4,16,41
hydrogen,
nitrogen groups. Further studies are in progress using in situ
characterization techniques to achieve a deeper understanding of 75 15
these promoting effects. 16
are probably altered in the presence of surface
1
7
5
. Conclusions
Oxygenꢀ and nitrogen functionalized CNTs were used as support for
the synthesis of Pd catalysts applied in the hydrogenation of 1,5ꢀ
cyclooctadiene. Pd nanoparticles supported on NCNTs showed a
narrower size distribution than those on OCNTs as disclosed by
TEM and XRD. XPS studies revealed that the nitrogen functional
groups on CNTs promote the reduction of Pd suggesting an electronꢀ
donating effect.
As compared to the Pd/OCNT catalyst, the Pd/NCNT catalyst
showed a significantly better catalytic performance in selective olefin
hydrogenation in terms of activity and selectivity. The extraordinary
catalytic performance of Pd/NCNT catalyst is attributed to the
structural and electronic promoting effect of the nitrogenꢀ
functionalized CNT support. The method employed here provides a
promising strategy to obtain highꢀperformance Pd/C catalysts for
selective olefin hydrogenation reactions.
8
8
9
9
0
5
0
5
18
M. T. Reetz and M. Dugal, Catal. Lett., 1999, 58, 207.
M. CrespoꢀQuesada, R. R. Dykeman, G. Laurenczy, P. J.
Dyson and L. KiwiꢀMinsker, J. Catal., 2011, 279, 66.
L. GómezꢀSainero, X. Seoane, J. Fierro and A. Arcoya, J.
Catal., 2002, 209, 279.
S. Ordóñez, E. Díaz, R. F. Bueres, E. AsedegbegaꢀNieto and
H. Sastre, J. Catal., 2010, 272, 158.
E. Díaz, S. Ordóñez, R. F. Bueres, E. AsedegbegaꢀNieto and
H. Sastre, Appl. Catal., B, 2010, 99, 181.
R. Arrigo, S. Wrabetz, M. E. Schuster, D. Wang, A. Villa, D.
Rosenthal, F. Girsgdies, G. Weinberg, L. Prati, R. Schlögl and
D. S. Su, Phys. Chem. Chem. Phys., 2012, 14, 10523.
K. Chizari, I. Janowska, M. Houllé, I. Florea, O. Ersen, T.
Romero, P. Bernhardt, M. J. Ledoux and C. PhamꢀHuu, Appl.
Catal., A, 2010, 380, 72.
1
9
25
30
35
2
0
21
2
2
2
3
2
4
25
P. D. Burton, T. J. Boyle and A. K. Datye, J. Catal., 2011,
280, 145.
Acknowledgments
2
2
6
7
J. J. F. Scholten, A. P. Pijpers and A. M. L. Hustings, Catal.
Rev., 1985, 27, 151.
C. Li, A. Zhao, W. Xia, C. Liang and M. Muhler, J. Phys.
Chem. C, 2012, 116, 20930.
Financial support from the German Federal Ministry of Education
and Research (BMBF) for the CarboKat project (03X0204D) within
the scope of the Inno.CNT alliance is gratefully acknowledged.
40
100
2
2
8
9
J. Mol, J. Mol. Catal. A: Chem., 2004, 213, 39.
A. Schmidt and R. Schomäcker, Ind. Eng. Chem. Res, 2007,
References
4
6, 1677.
1
2
3
R. S. Oosthuizen and V. O. Nyamori, Platinum Met. Rev.,
011, 55, 154.
Z. M. de Pedro, E. Diaz, A. F. Mohedano, J. A. Casas and J. J.
Rodriguez, Appl. Catal., B, 2011, 103, 128.
D. Teschner, Z. Révay, J. Borsodi, M. Hävecker, A. Knopꢀ
Gericke, R. Schlögl, D. Milroy, S. D. Jackson, D. Torres and
P. Sautet, Angew. Chem. Int. Ed., 2008, 47, 9274.
D. Teschner, J. Borsodi, A. Wootsch, Z. Revay, M. Havecker,
A. KnopꢀGericke, S. D. Jackson and R. Schlögl, Science,
3
0
R. Arrigo, M. Hävecker, S. Wrabetz, R. Blume, M. Lerch, J.
McGregor, E.P.J. Parrott, J.A. Zeitler, L.F. Gladden, A. Knopꢀ
Gericke, R. Schlögl and D.S. Su, J. Am. Chem. Soc., 2010,
132, 9616.
2
1
05
45
50
55
3
3
1
2
D. J. Suh and T. Park, Carbon, 1993, 31, 427.
4
J. F. Moulder, W. F. Stickle, P.E. Sobol and K. D. Bomben,
Handbook of Xꢀray Photoelectron Spectroscopy, Perkin Elmer
Corp., Eden Prairie, MN, USA, 1992.
J. W. Niemantsverdriet, Spectroscopy in catalysis: An
introduction. 3rd ed. WileyꢀVCH: Weinheim, Germany; 2007.
H. Arnold, F. Döbert and J. Gaube in Handbook of
Heterogeneous Catalysis, 2nd ed.; Ertl G, Knözinger H,
Schüth F, Weitkamp J. WileyꢀVCH: Weinheim, Germany:
110
2
008, 320, 86.
C. A. Hamilton, S. D. Jackson and G. J. Kelly, Appl. Catal., A,
002, 237, 201.
5
6
3
3
3
4
2
S. DominguezꢀDominguez, A. BerenguerꢀMurcia, B. Pradhan,
A. LinaresꢀSolano and D. CazorlaꢀAmoros, J. Phys. Chem. C,
115
2
008, 112, 3827.
2
008.
This journal is © The Royal Society of Chemistry [year]
Journal Name, [year], [vol], 00–00 | 7

Page 9 of 9
Catalysis Science & Technology
View Article Online
3
5
X. Lepró, E. Terrés, Y. VegaꢀCantú, F. J. RodríguezꢀMacías,
H. Muramatsu, Y. A. Kim, T. Hayahsi, M. Endo, M. Torres R.
and M. Terrones, Chem. Phys. Lett., 2008, 463, 124.
D. He, C. Zeng, C. Xu, N. Cheng, H. Li, S. Mu and M. Pan,
Langmuir, 2011, 27, 5582.
M. D. Sánchez, P. Chen, T. Reinecke, M. Muhler and W. Xia,
ChemCatChem, 2012, 4, 1997.
3
3
3
6
7
8
5
0
5
S. Kogan, M. Kaliya and N. Froumin, Top. Catal., 2006, 297,
2
31.
1
39
J. Amadou, K. Chizari, M. Houllé, I. Janowska, O. Ersen, D.
Bégin and C. PhamꢀHuu, Catal.Today, 2008, 138, 62.
J. H. Bitter, J. Mater. Chem., 2010, 20, 7312
M. Wilde, K. Fukutani, W. Ludwig, B. Brandt, J.ꢀH. Fischer,
S. Schauermann and H.ꢀJ. Freund, Angew. Chem. Int. Ed.,
2008, 47, 9289.
4
4
0
1
1
8
|
Journal Name, [year], [vol], 00–00
This journal is © The Royal Society of Chemistry [year]