Synergistic Effect of Tungsten Nitride and Palladium for the Selective Hydrogenation of Cinnamaldehyde at the C=C bond

Article abstract of DOI:10.1002/cctc.201600184

Herein, a series of catalysts made of Pd-WN on various supports was synthesized by modifying supports with small-size WN NPs firstly and loading Pd subsequently. Their catalytic performances were evaluated for selective hydrogenation of cinnamaldehyde (CAL) to hydrocinnamaldehyde (HALD). Interestingly, it was found the synergistic effect between Pd and WN can improve conversion and selectivity of CAL to HALD. Among these catalysts, Pd-WN/SBA-15 shows best performance with highest conversion (99 %) and selectivity to HALD (97 %), which is superior to commercial 5 % Pd/C catalyst. The hydrogenation kinetics of Pd-WN/SBA-15 has been adequately represented by a standard pseudo-first-order approximation, and it discloses that the existence of WN can effectively decrease the activation energy (23.2 kJ mol^-1). The synergistic effect of Pd and WN results in enriching the electron density of Pd, increasing the ratio of surface Pd⁰ and decreasing the size of Pd on the Pd-WN/SBA-15 catalyst.

Full text of DOI:10.1002/cctc.201600184

DOI: 10.1002/cctc.201600184
Full Papers
Synergistic Effect of Tungsten Nitride and Palladium for
the Selective Hydrogenation of Cinnamaldehyde at the
C=C bond
Dong Wang,[a, b] Yujun Zhu,*[a] Chungui Tian,[a] Lei Wang,[a] Wei Zhou,[a] Yongli Dong,[a]
Haijing Yan,[a] and Honggang Fu*[a]
Herein, a series of catalysts made of Pd–WN on various sup-
ports was synthesized by modifying supports with small-size
WN NPs firstly and loading Pd subsequently. Their catalytic per-
formances were evaluated for selective hydrogenation of cin-
namaldehyde (CAL) to hydrocinnamaldehyde (HALD). Interest-
ingly, it was found the synergistic effect between Pd and WN
can improve conversion and selectivity of CAL to HALD.
Among these catalysts, Pd–WN/SBA-15 shows best per-
formance with highest conversion (99%) and selectivity to
HALD (97%), which is superior to commercial 5% Pd/C
catalyst. The hydrogenation kinetics of Pd–WN/SBA-15 has
been adequately represented by a standard pseudo-first-order
approximation, and it discloses that the existence of WN can
effectively decrease the activation energy (23.2 kJmolÀ1). The
synergistic effect of Pd and WN results in enriching the elec-
tron density of Pd, increasing the ratio of surface Pd0 and de-
creasing the size of Pd on the Pd–WN/SBA-15 catalyst.
Introduction
The hydrogenation of a,b-unsaturated aldehydes to a,b-unsa-
turated alcohols or saturated aldehydes is a critical step in the
synthesis of a large number of fine chemicals, such as pharma-
ceuticals, perfumes, and food additives.[1–3] Recently, hydrocin-
namaldehyde (HALD) was found to be an important intermedi-
ate in the synthesis of pharmaceuticals used in the treatment
of HIV.[4] Cinnamaldehyde (CAL), as typical a,b-unsaturated al-
dehyde, can result in various product distributions in its hydro-
genation route (Scheme 1), depending on the active metal, sol-
vent, and promoter. In general, Pd catalysts are in favor of the
hydrogenation of the C=C bond to produce HALD, and Pt-
based catalysts facilitate the production of cinnamyl alcohol
(COL) through the hydrogenation of the C=O bond.[5] Although
the reduction of the C=C group is favored thermodynamical-
ly,[6–11] it is still challenging to enhance the efficiency of Pd cata-
lysts and lessen Pd usage.[12,13]
Scheme 1. Reaction pathways for cinnamaldehyde hydrogenation.
tional catalysis and electrocatalysis.[14–16] It has been proven
that tungsten and tungsten nitride are effective co-catalysts of
Pt and Pd in hydrogen evolution reaction and oxygen reduc-
tion reaction.[15,17,18] Importantly, in our previous work it was
found that Mo2N could distinctively enhance the activity and
selectivity of Pt on the ternary Pt–Mo2N/SBA-15 catalyst for the
hydrogenation of CAL at the C=O bond to COL. Thus, we were
curious to investigate whether the synergistic effect between
Pd and WN could promote the catalytic performance of Pd for
the hydrogenation of cinnamaldehyde at the C=C bond to
HALD.
In recent years, metal nitrides have been studied intensely
owing to their synergistic effects and Pt-like behavior in tradi-
[a] D. Wang, Prof. Y. Zhu, Prof. C. Tian, Dr. L. Wang, Prof. W. Zhou, Y. Dong,
H. Yan, Prof. H. Fu
Key Laboratory of Functional Inorganic Material Chemistry
Heilongjiang University
Ministry of Education; School of Chemistry and Materials
Heilongjiang University
74 Xuefu Road, Harbin 150080 (P.R. China)
E-mail: fuhg@vip.sina.com
To strengthen the interaction of Pd and WN, it is essential to
gain small-sized and highly dispersed WN nanoparticles (NPs)
to increase the opportunity of WN contacting with Pd. Thanks
to our successful strategy to support small-size Mo2N NPs on
SBA-15,[19] on various supports with SiÀOH and AlÀOH groups,
we considered it more effective to use N-[3-(trimethoxysilyl)-
propyl]-ethylenediamine with two amino groups as a silane
coupling agent to immobilize H4[SiO4(W3O9)4](SiW12) clusters
and obtain tiny WN NPs after calcination in pure NH3.
yujunzhu@hlju.edu.cn
[b] D. Wang
School of Chemical and Environmental Engineering
Hanshan Normal University
Chaozhou 521041 (P.R. China)
Supporting Information for this article can be found under http://
dx.doi.org/10.1002/cctc.201600184.
ChemCatChem 2016, 8, 1718 – 1726
1718
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Based on above discussion, a series of ternary Pd–WN cata-
lysts on different supports were prepared by loading Pd on
WN-modified supports. These ternary catalysts showed superi-
or performance over the binary catalysts without WN for the
hydrogenation of CAL at the C=C bond; and Pd–WN/SBA-15
displayed the best selectivity to HALD with near 100% conver-
sion of CAL. The intrinsic synergistic effect between Pd and
WN was explored by investigating the influence of WN on the
structure, the hydrogenation kinetics, and the stability of the
Pd–WN/SBA-15 catalyst. It was found that the addition of WN
enriched the electron density of Pd, increased the ratio of Pd0
species, and greatly decreased Pd NPs size, which led to the
enhanced performance, lower active energy, and good stability.
In this work, we extend our previous work to the scope of syn-
ergistic effect between noble metal and nitrides on the hydro-
genation of CAL at the C=C bond to HALD. The study gives
a further insight to the nature of the synergistic effect between
noble metal and nitrides.
Figure 1. Low-angle XRD patterns of a) SBA-15, b) 1.5% Pd/SBA-15, c) WN/
SBA-15, and d) 1.5% Pd–WN/SBA-15.
Results and Discussion
WN/SBA-15, and 1.5% Pd/SBA-15 samples. In Figure 1, all the
samples exhibited prominent peaks corresponding to (100),
(110), and (200) reflection at 2q approximately 0.5–28, which
is characteristic of the hexagonal mesostructure of the SBA-15
host (Figure 1a).[20–22] This indicates that the loads of WN and
Pd do not destroy the well-ordered mesoporous structure of
SBA-15. However, the intensity of the (100) reflection showed
a clear decrease for WN/SBA-15 and 1.5% Pd–WN/SBA-15,
which depicts the decrease in the mesoporous ordering of
SBA-15. Moreover, compared with the pure SBA-15, the (100)
reflections of WN/SBA-15 and 1.5% Pd–WN/SBA-15 sample
shift to higher angles, indicating that WN and Pd NPs are scat-
tered into the pore channels of SBA-15.
The catalytic performances were investigated for the selective
hydrogenation of CAL over various catalysts including 1.5%
Pd–WN/SBA-15, 1.5% Pd–WN/MCM-41, 1.5% Pd–WN/Na-Beta,
0.75% Pd–WN/Al2O3, and 1.5% Pd–WN/SAPO-34 catalysts. The
results are listed in Table S1 in the Supporting Information. It
was found that HALD was the main product for all catalysts;
however, CAL conversion and selectivity to HALD were differ-
ent owing to the difference in the natures of supports. Notably,
the ternary catalysts with different supports showed higher
conversion and selectivity than the binary ones without WN.
Especially, for the 1.5% Pd–WN/SBA-15 catalyst, the selectivity
to HALD attained to 97.6%, corresponding 99.5% conversion
of CAL, which was superior to that of commercial 5% Pd/C cat-
alyst. Evidently, WN plays an important role in enhancing the
activity and selectivity of Pd NPs on various catalysts. This
result illustrates that the WN had a positive effect on Pd for
the selective hydrogenation of cinnamaldehyde at the C=C
bond. To clarify what the nature of the synergistic effect of Pd
and WN is, 1.5% Pd–WN/SBA-15 was taken as a typical catalyst
to investigate its structure and catalytic property.
The textural properties of the samples have been also dis-
closed by N2 adsorption–desorption analysis as listed in Table 1
and Figure S1 (Supporting Information), including WN/SBA-15,
1.5% Pd–WN/SBA-15, and 1.5% Pd/SBA-15. As expected, a sig-
nificant decrease in the surface area of all samples was ob-
served upon metal loading. Pore sizes were reduced for the
samples of WN/SBA-15 and 1.5% Pd–WN/SBA-15, which was
consistent with the low-angle XRD results. Pore volumes also
followed the similar trend, which indicates that nanoparticles
are dispersed within the pores of SBA-15.
Characterization of Pd–WN/SBA-15
The morphologies of the WN/SBA-15 and 1.5% Pd–WN/SBA-
15 were unambiguously characterized by TEM. The TEM image
for WN/SBA-15 composite (Figure 2a,b) clearly depicted a fairly
The low-angle XRD patterns are displayed in Figure 1 for 1.5%
Pd–WN/SBA-15 and the reference samples including SBA-15,
Table 1. Relevant physicochemical parameters of the different samples.
[a]
[a]
Sample
SBET [m2 gÀ1
]
Vpore [cm3 gÀ1
]
Pore size [nm][a]
dTEM [nm][b]
Pd content [wt%][c]
SBA-15
WN/SBA-15
1.5% Pd–WN/SBA-15
1.5% Pd/SBA-15
1.5% Pd–WN/SBA-15-r[d]
811
468
394
644
403
1.20
0.69
0.64
0.98
0.65
8.2
6.8
5.9
6.5
5.8
–
–
–
1.38
1.42
1.29
13.6/WN
16.8
12.2/Pd
16.4
[a] Derived from N2 adsorption–desorption isotherms. [b] Average particle sizes calculated from at least 100 individual crystallites in TEM images. [c] The
content of Pd determined by ICP–OES. [d] 1.5% Pd–WN/SBA-15 sample after five consecutive catalytic runs.
ChemCatChem 2016, 8, 1718 – 1726
www.chemcatchem.org
1719
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Figure 2. a,b) Low-magnification TEM images, c) HRTEM, and d) particle-size
distribution of the WN/SBA-15. The histogram in (d) depicts the particle-size
distribution and the average size of the particles determined by statistical
analysis of the TEM images, at least 100 individual crystallites were analyzed.
Figure 3. a) Low-magnification TEM images, b) particle-size distribution,
c) HRTEM, and d) FFT image of the 1.5% Pd–WN/SBA-15. The histogram in
(b) depicts the particle-size distribution and the average size of the particles
determined by statistical analysis of the TEM images, at least 100 individual
crystallites were analyzed.
good dispersion of tiny nanoparticles on the highly ordered
2D hexagonal mesostructure of SBA-15 with an average parti-
cle size of 13.6 nm. In the high-resolution (HR) TEM image (Fig-
ure 2c), the crystal plane spacing was measured as 0.250 nm,
corresponding to the (100) crystal facet of the hexagonal WN.
In addition, SEM element maps for WN/SBA-15 composite (Fig-
ure S2) also showed the uniform dispersion of WN on SBA-15.
No observable sintering could be found on WN/SBA-15, al-
though the WN content was approximately 8.9% according to
the analysis of inductively coupled plasma optical emission
spectroscopy (ICP–OES) and X-ray fluorescence analysis. In gen-
eral, the nitrides show large sizes above micrometer if they are
prepared by solid-state ion-exchange route.[23] Bazarjani and
co-workers have loaded WN NPs on porous silicon oxycarboni-
tride matrices.[24] The as-prepared nitrides have a large size of
approximately 40 nm, and their distribution on the support is
not uniform. In this work, the obtained WN/SBA-15 sample
with small sizes and good dispersion of WN particles was con-
tributed to the small size close to the nanometer and stable
structure of SiW12 clusters, the anchoring function of N-[3-(tri-
methoxysilyl)propyl]-ethylenediamine to SiW12 clusters, and the
mesoporous structure of the SBA-15 support. Thus, the ob-
tained small-size WN NPs possesses more chance to touch
with Pd.
15 in Figure 3c, it can be seen that WN and Pd were in close
contact with each other on the SBA-15. The large particles
were typical WN grains with the size of about 12 nm, whereas
the relatively small ones located around the WN grains can be
ascribed to Pd particles with the (200) crystal plane. The fast
Fourier transform (FFT) image (Figure 3d) confirms they are
indeed WN with (100) facet and Pd with (200) facet. Thus, the
images reveal the coexistence of Pd and WN on 1.5% Pd–WN/
SBA-15 catalyst. Notably, the images distinctly show the inti-
mate contact of Pd with WN, which could result in intensive in-
teraction. For the 1.5% Pd/SBA-15, the TEM image (Figure S4)
displayed polydispersity and the average particle size was
12.2 nm; even some much larger Pd particles of about 30 nm
were seen, indicating the sintering or aggregation of Pd parti-
cles in this catalyst. The sintering of pure Pd catalyst has been
found in previous reports by Abate et al. and Yang et al.[25,26] In
view of the TEM images for 1.5% Pd–WN/SBA-15 and 1.5% Pd/
SBA-15 catalyst, it can be concluded that the particle size and
dispersion of Pd can be controlled effectively by the modifica-
tion of WN NPs on SBA-15. This effect discloses the intensive
interaction between Pd and WN owing to their similar elec-
tronic structures.[16,27] In addition, the interaction of the nitrides
(carbides) with Pt has been discovered in our earlier works by
experiments and DFT calculations.[17,19,28] Therefore, it is rational
to suggest that the Pd NPs are before growing around the
highly dispersed WN grains on SBA-15. This point was proven
by the increase of the average particle size from 13.6 nm for
WN/SBA-15 composite to 16.8 nm for 1.5% Pd–WN/SBA-15
sample.
For the 1.5% Pd–WN/SBA-15 sample, the TEM image also
showed homogeneous dispersion (Figure 3a), but the average
particle size increased to 16.8 nm (Figure 3b). Although it is
difficult to distinguish Pd from WN in TEM owing to their simi-
lar appearance as dark spots in Figure 3a, it can be concluded
that the Pd NPs are uniformly dispersed on the WN/SBA-15
composite by corresponding element maps shown in Fig-
ure S3. Moreover, from the HRTEM image of 1.5% Pd–WN/SBA-
The compositions of the WN/SBA-15 and 1.5% Pd–WN/SBA-
15 were investigated by using wide-angle XRD and the pat-
ChemCatChem 2016, 8, 1718 – 1726
www.chemcatchem.org
1720
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Figure 4. Wide-angle XRD patterns and enlarged regional patterns between
348 and 458 (inset) for a) SBA-15, b) WN/SBA-15, c) 1.5% Pd–WN/SBA-15, and
d) 1.5% Pd/SBA-15.
Figure 5. Pd3d XPS spectra of a) 1.5% Pd–WN/SBA-15 and b) 1.5% Pd/SBA-
15.
terns are shown in Figure 4. The XRD patterns showed a diffuse
peak of amorphous SiO2 at approximately 22.58 for all samples
as pure SBA-15 (Figure 4a). For WN/SBA-15 (Figure 4b), the
peaks at 37.34 and 43.678 were assigned to the (100) and
(101) planes of hexagonal WN. The broad and weak WN dif-
fraction peaks indicate that the WN particles are well dispersed
on the composites as displayed in the TEM image (Figure 2a).
After loading Pd, the peaks belonging to WN disappeared in
the XRD pattern for the 1.5% Pd–WN/SBA-15. Furthermore, the
diffraction peak ascribed to Pd was no longer observed, and
the Pd content was 1.38% analyzed by ICP–OES (Table 1). How-
ever, for the 1.5% Pd/SBA-15 (Figure 4d), the diffraction peak
was clearly observed indexed as Pd (111) planes of a face-cen-
tered cubic (fcc) structure at 39.78, and the Pd content was ap-
proximately 1.42%. Compared with 1.5% Pd/SBA-15, the ab-
sence of the Pd peak for 1.5% Pd–WN/SBA-15 implies that the
existence of WN improves the dispersion of Pd on
Figure 6. W4f XPS spectra of a) WN/SBA-15 and b) 1.5% Pd–WN/SBA-15.
the SBA-15 support, which is consistent with the re-
sults revealed by TEM. This result is also similar to
the earlier works.[17,19] Although the high surface area
of SBA-15 is helpful for getting a high dispersion of
Pd, it is evident that WN NPs play a significant role in
improving the Pd dispersion and decreasing the par-
ticle size of Pd. The enhanced dispersion of Pd re-
flects that there is intensive interaction between WN
and Pd NPs. In addition, contrasted to WN/SBA-15,
the disappearance of WN peaks for the 1.5% Pd–
WN/SBA-15 also suggests this strong interaction.
To gain further insight into the surface structure of
the catalyst, X-ray photoelectron spectroscopy (XPS)
Table 2. XPS data of the Pd3d levels for the different samples.
Sample
Pd species
Pd species
BE [eV]
Content of Pd
species [%]
3d5/2
3d3/2
1.5% Pd–WN/SBA-15
1.5% Pd/SBA-15
Pd0
335.3
336.8
335.7
336.9
335.0
336.4
340.6
342.3
341.0
342.4
340.3
341.9
74
26
61
39
70
30
Pd2+
Pd0
Pd2+
Pd0
Pd2+
1.5% Pd–WN/SBA-15-r
analysis was conducted and the results are shown in Figure 5
and Figure 6, together with their deconvolution obtained by
fitting Gaussian peaks after Shirley background subtraction. As
shown in Figure 5, the Pd XPS spectra of 1.5% Pd–WN/SBA-15
and 1.5% Pd/SBA-15 presented a doublet corresponding to
Pd3d5/2 and Pd3d3/2, and the corresponding binding energies
(BE) are listed in Table 2. For 1.5% Pd–WN/SBA-15, the Pd3d5/2
peak at 335.3 eV was described to Pd0 (metallic palladium),
and the Pd3d5/2 peak at 336.8 eV was described to Pd2+ (palla-
dium oxide) (Figure 5a).[29] However, for 1.5% Pd/SBA-15 (Fig-
ure 5b), the peak at approximately 335.7 eV was attributed to
Pd0, and the peak at approximately 336.9 eV corresponded to
Pd2+ species. Thus, on one hand, there is a shift of 0.4 eV
toward low binding energy for 1.5% Pd–WN/SBA-15 with re-
spect to 1.5% Pd/SBA-15, which is ascribed to electron transfer
between WN and Pd.[19,30] Interestingly, this electron transfer
ChemCatChem 2016, 8, 1718 – 1726
www.chemcatchem.org
1721
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
was also proven by the shift of W4f BE for the WN/SBA-15
composite after the load of Pd. For WN/SBA-15 (Figure 6a), the
peak of W from WÀN located at 35.3 and 37.4 eV and the peak
at 39.6 eV were assigned to WO3, indicating low amount or
amorphous characteristics of WO3 on the surface.[17] However,
the W4f BEs from WÀN for ternary 1.5% Pd–WN/SBA-15 are
35.5 and 37.6 eV (Figure 6b). Clearly, there is a positive shift
(0.2 eV) relative to that of WN/SBA-15, confirming the electron
transfer from WN to Pd. This transfer is favorable to the in-
crease of the metallic characteristics of Pd by enriching Pd
electron density, and the strong electronic interaction can
therefore influence catalytic behavior.[31,32]
with that of 1.5% Pd/SBA-15. However, it showed much higher
TOF (757 hÀ1) and selectivity than that over 0.75% Pd/SBA-15
(451 hÀ1) and 1.5% Pd/SBA-15 (418 hÀ1). In other words, the ex-
istence of WN successfully increases the catalytic efficiency of
Pd on 0.75% Pd–WN/SBA-15. Herein, it further confirms that
the intensive interaction between Pd and WN favors the pro-
motion of the catalytic performance of Pd, and the WN as co-
catalyst could lessen the usage of Pd for the selective hydroge-
nation of CAL in some extent. This effect was also found for Pd
and Mo2N for 1.5% Pd–Mo2N/SBA-15 (Table 3). The CAL conver-
sion and selectivity to HALD was 88.7% and 80.8% over 1.5%
Pd–Mo2N/SBA-15, respectively. Clearly, the activity and selectiv-
ity of 1.5% Pd–Mo2N/SBA-15 is higher than that of 1.5% Pd/
SBA-15 without Mo2N, which indicates the universality of the
enhancement between Pd and different nitrides.
On the other hand, the XPS results also revealed the change
of Pd species caused by the interaction between Pd and WN
on the 1.5% Pd–WN/SBA-15 surface. The relative abundance of
the Pd0 (%) species of the samples has been estimated by con-
sidering the deconvolution peaks of Pd3d BE. As listed in
Table 2, the content of metallic Pd0 increased from 61% for
1.5% Pd/SBA-15 to 74% for 1.5% Pd–WN/SBA-15. It also sug-
gests the strengthening of metallic characteristics of Pd in
1.5% Pd–WN/SBA-15. Evidently, the nitride (carbide) could
largely improve the crystallinity of the noble metal because of
the intensive interaction of nitride (carbide) with noble
metal.[17,28]
The enhancement of conversion is related to the intensive
interaction of Pd and WN, which changes the surface electron-
ic structure of Pd as disclosed by XPS. Compared with 1.5%
Pd/SBA-15, the 1.5% Pd–WN/SBA-15 possesses more Pd0 spe-
cies with higher electron density on surface. It is well known
that the metallic Pd with high electron density would acceler-
ate its performance on chemoselective hydrogenation of a,b-
unsaturated carbonyls.[33]
Thus, the conversion rate of CAL was varied by the interac-
tion of Pd and WN. Given the same reaction conditions, the
1.5% Pd–WN/SBA-15 (Figure 7a) showed faster conversion rate
of CAL than the 1.5% Pd/SBA-15 (Figure 7b): it took 41 and
61 min to obtain a conversion of 50% for the 1.5% Pd–WN/
SBA-15 and 1.5% Pd/SBA-15, respectively. In addition, for the
1.5% Pd–WN/SBA-15, the selectivity to HALD slowly increased
from 90.6% at 10 min to 97.6% at 2 h, and the selectivity to
COL and HALC was distinctly lower than to HALD. However,
for 1.5% Pd/SBA-15, the selectivity to HALD reached 59.5% at
the beginning, and slowly decreased to 51.5% subsequently;
at the same time, the selectivity to HALC slowly rose up to
38.9%. Very little COL was detected at the present hydrogena-
tion conditions for the both catalyst, similarly to what was pre-
viously reported on Pd catalysts.[34,35] It can be seen that 1.5%
Pd/SBA-15 could convert more CAL to saturated alcohol
(HALC) than the 1.5% Pd–WN/SBA-15, which induces the de-
crease of the selectivity to HALD. But, this trend could be in-
hibited effectively over 1.5% Pd–WN/SBA-15. Conse-
Selective hydrogenation of CAL by Pd–WN/SBA-15
According to Table S1, 1.5% Pd–WN/SBA-15 showed superior
performance for the selective hydrogenation of cinnamalde-
hyde, and the activity of Pd–WN/SBA-15 with lower Pd content
(0.75%) was also studied with WN/SBA-15 and 1.5% Pd/SBA-15
as reference. As listed in Table 3, WN/SBA-15 gave trace con-
version (6.9%) of CAL, and CAL was mostly converted to the
side products (89.2%), indicating that WN shows poor activity
in the hydrogenation process. For 1.5% Pd/SBA-15, it per-
formed 76.9% CAL conversion and 51.5% selectivity to HALD,
whereas the selectivity to COL and HALC was 0.8% and 38.9%,
respectively. It can be seen that 1.5% Pd/SBA-15 facilitates the
production of completely hydrogenated hydrocinnamyl alco-
hol. However, 0.75% Pd–WN/SBA-15 displayed 70.6% CAL con-
version and 89.5% selectivity to HALD. This activity competes
quently, WN in the catalysts plays an important role
Table 3. Hydrogenation of cinnamaldehyde over different catalysts.[a]
in determining the course of CAL hydrogenation in
[b]
Conversion [%] TOF [hÀ1
]
Selectivity [%]
HALD COL HALC Others[c]
terms of both the rate and the selectivity toward
Sample
either C=C or C=O hydrogenation.
It has been proven that all HALC produced on Pd
catalyst is entirely produced through the further hy-
drogenation of the COL rather than HALD for the hy-
drogenation of CAL.[34,35] Thus, the higher selectivity
to HALC indicates that there is much COL formed
over the 1.5% Pd/SBA-15 by the hydrogenation of
C=O firstly, which is then hydrogenated to HALC im-
mediately. Little COL was detected during the pro-
cess because COL hydrogenation to HALC was 30
times faster than the hydrogenation of CAL to COL
on Pd.[34,35] It is well known the aldehyde reactant
WN/SBA-15
0.75% Pd/SBA-15
1.5% Pd/SBA-15
0.75% Pd–WN/SBA-15 70.6
1.5% Pd–WN/SBA-15 99.5
1.5% Pd-Mo2N/SBA-15 88.7
6.9
42.5
76.9
–
451
418
757
532
482
7.6
54.9
51.5
89.5
97.6
80.8
3.2
2.5
0.8
0.7
0.4
0.6
0
89.2
18.3
8.8
4.2
0.9
24.3
38.9
5.6
1.1
13.3
5.3
[a] Reaction conditions: 50 mg catalyst, 1.00 g CAL, 30 mL isopropanol, 10 bar H2,
313 K, 2 h. [b] Turnover frequency (TOF)=[moles of cinnamaldehyde reacted]/[(moles
of Pd loading)”(reaction time)]. [c] Includes 1-(3-propoxyprop-1-enyl)benzene, cin-
namyl formate, cinnamic acid, benzyl, cinnamate, 4,4-diphenylcyclohexa-1,5-dienyl
acetate, and other condensation products that could not be identified by GC–MS be-
cause of their large molecular masses.
ChemCatChem 2016, 8, 1718 – 1726
www.chemcatchem.org
1722
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Figure 7. Selective hydrogenation of cinnamaldehyde catalyzed by a) 1.5%
Pd–WN/SBA-15 and b) 1.5% Pd/SBA-15.
Figure 8. a) Plot of ln[CAL] versus reaction time for the conversion of CAL on
1.5% Pd–WN/SBA-15 at four different temperatures. b) Arrhenius plot of the
pseudo-first-order rate constants extracted from the data shown in (a). Reac-
tion conditions: 10 bar H2, 800 rpm stirring speed, 1.0 g CAL, 18 mg 1.5%
Pd–WN/SBA-15, 30 mL isopropanol.
can be adsorbed on a metal surface in a flat, planar adsorption
mode, parallel to the surface with both C=C and C=O bonds
involved in adsorption and in the mode of the carbonyl group
(C=O).[5] The former adsorption mode favors the hydrogenation
of C=C yielding HALD, whereas the later facilitates the hydro-
genation of C=O to produce COL.[36] It should be stressed out
that metal particle size is an important feature that could influ-
ence the selectivity toward saturated aldehyde, and smaller
a straight line. This indicates that the reaction rate fits to the
first-order reaction, consistent with others.[12,37] Arrhenius equa-
tion of the pseudo-first-order reaction constants was used to
calculate an apparent activation energy (Ea) at four different
particles of Pt, Pd, Ru, and Co result in higher selectivity to C= temperatures (Figure 8b). The Ea of 23.2 kJmolÀ1 was obtained
C hydrogenation. This particle size effect of the metal was con-
nected to the phenyl group, sterically hampering the approach
of the cinnamaldehyde molecule parallel to a flat metal surface
because of the steric repulsion of the aromatic ring.[5] However,
this steric effect does not operate on small particles on which
both the C=C and C=O bonds absorb on the surface. Thus,
this increased selectivity on the 1.5% Pd–WN/SBA-15 is as-
cribed to that the small Pd NPs promote the absorption of
CAL with both the C=C and C=O bonds, which facilitates the
hydrogenation of the C=C bond to HALD. However, for 1.5%
Pd/SBA-15 with relatively large Pd NPs, the parallel absorption
is inhibited and the CAL can absorb by its C=O group to form
COL, which is further hydrogenated to HALC immediately.
Therefore, 1.5% Pd–WN/SBA-15 performs constantly increasing
selectivity to HALD and much lower selectivity to HALC.
over 1.5% Pd–WN/SBA-15 in our case, which is lower than that
of the 1.5% Pd/SBA-15 (29.1 kJmolÀ1, Figure S5). Thus, it is
clearly confirmed that the synergistic effect between Pd and
WN does suppress the activation barrier of hydrogenation.
These results evidently depict the superiority of 1.5% Pd–WN/
SBA-15 over the 1.5% Pd/SBA-15 in respect to the catalytic
rate. The enhanced activity and reaction rate may be contribut-
ed to the decreased Pd NPs size and the modified electronic
structure of the surface Pd on 1.5% Pd–WN/SBA-15, which ori-
ginated from the intensive interaction of Pd and WN.[33,38]
The stability of the catalysts was also investigated in the se-
lective hydrogenation of CAL over the 1.5% Pd–WN/SBA-15
and 1.5% Pd/SBA-15. To our delight, the 1.5% Pd–WN/SBA-15
still maintained its original activity after the fifth run, and the
selectivity to HALD also showed no distinct loss within five re-
action cycles (Figure 9). After the fifth cycle, only negligible Pd
was leached for the 1.5% Pd–WN/SBA-15 catalyst as detected
by ICP–OES (Table 1). The TEM image of the used 1.5% Pd–
WN/SBA-15 (Figure S6) after the fifth run was almost the same
as that of the fresh catalyst, suggesting that it is aggregation-
free of Pd, which is consistent with the result of XRD pattern
(Figure S7). The XPS spectra also suggest that the electronic
More detailed kinetic runs of the CAL hydrogenation were
conducted over 1.5% Pd–WN/SBA-15. The kinetic curves of
CAL hydrogenation under different temperature (298, 308, 313,
and 323 K) and corresponding Arrhenius plots are displayed in
Figure 8. To extract the activation energy of CAL hydrogena-
tion, the data were gained at low conversions (<30%). As
shown in Figure 8a, plotting ln[CAL] against time showed
ChemCatChem 2016, 8, 1718 – 1726
www.chemcatchem.org
1723
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
ces were investigated for the chemoselective hydrogenation of
cinnamaldehyde (CAL) on the C=C bond. It was found that
modification of WN on the support improved the catalytic per-
formance of Pd as a result of the synergistic effect between
WN and Pd. Among these catalysts, Pd–WN/SBA-15 displayed
the highest yield of hydrocinnamaldehyde (HALD, 97.1%),
moreover, it was reused five times successively without signifi-
cant loss of catalytic activity and selectivity. The reaction kinet-
ics has been represented by a standard pseudo-first-order ap-
proximation, which indicates that the activation energy is
lower over Pd–WN/SBA-15 (23.2 kJmolÀ1) than over Pd/SBA-15
(29.1 kJmolÀ1). The enhancement is attributed to the synergis-
tic effect of Pd and WN, which enriches the electron density of
Pd by transferring electrons from WN to Pd, increases the ratio
of Pd0 species, and decreases the Pd size on the Pd–WN/SBA-
15 catalyst. This synergistic effect between Pd and WN also re-
duces the usage of Pd to enhance the catalytic efficiency in
some extent. Furthermore, hydrogenation of various a,b-unsa-
turated aldehydes over the 1.5% Pd–WN/SBA-15 could obtain
the corresponding saturated aldehydes with high selectivity
and conversion. Thus, it opens a new way to design many
more powerful catalysts by interaction between noble metal
and nitrides and lessen the usage of noble metal in catalysis
field.
Figure 9. Stability test for the 1.5% Pd–WN/SBA-15 shown as performance
versus number of recycling runs. Reaction conditions: 50 mg 1.5% Pd–WN/
SBA-15, 1.00 g CAL, 30 mL isopropanol, 10 bar H2, 313 K, 2 h.
state and ratio of the Pd species on the surface of the fifth
reused catalyst are similar to that of the fresh catalyst (Fig-
ure S8, Table 2). Nevertheless, the 1.5% Pd/SBA-15 performed
a noticeable decrease of CAL conversion during the recycling
(Figure S9). At the forth cycle, only 48.3% CAL conversion was
obtained. Thus, the existence of WN clearly benefits to im-
prove the stability of the catalyst.
To further investigate the versatility of the 1.5% Pd–WN/
SBA-15 catalyst, various unsaturated aldehydes were used as
substrates under optimized conditions, and these results are
listed in Table 4. It was found that hydrogenation of various
a,b-unsaturated aldehydes over the 1.5% Pd–WN/SBA-15
could obtain the corresponding saturated aldehydes with high
selectivity and conversion. In view of the results, it indicates
that steric effects induced by the substituent groups on the C
atoms in the C=C bond have an influence on the intramolecu-
lar selectivity.
Experimental Section
Catalyst preparation
SBA-15 was prepared with the triblock copolymer Pluronic P123
and tetraethylorthosilicate (TEOS) according to a typical synthesis
route.[20] The synthesis of other supports is described in the Sup-
porting Information, including MCM-41, Al2O3, SAPO-34 and Na-
Beta.
The preparation is described for the Pd–WN/SBA-15 catalyst as ex-
ample; the other catalysts were synthesized accordingly. As shown
in Scheme 2, the WN/SBA-15 composite was prepared by anchor-
ing H4[Si(W3O10)4] cluster on N-[3-(trimethoxysilyl)propyl]-ethylene-
diamine modified SBA-15 (SBA-15-NH2), and followed by calcina-
tion in NH3. A typical procedure was as follow: SBA-15 (2.0 g) was
added to the mixed solution containing N-[3-(trimethoxysilyl)prop-
yl] ethylenediamine (4.0 mL) and ethanol (4.0 mL). After stirring for
Conclusions
The catalysts were prepared by loading Pd on different sup-
ports modified by small-size WN, and their catalytic performan-
Table 4. Selective hydrogenation of various unsaturated aldehydes over the 1.5% Pd–WN/SBA-15.[a]
Entry
1
Substrate
Product
t [h]
Conversion [%][b]
80.7
Selectivity [%][b]
70.1
0.5
2
2.5
89.7
85.3
3
4
5
6
2
2
3
5
99.5
94.3
76.5
95.3
97.1
85.4
77.0
83.9
[a] Reaction conditions: 50 mg 1.5% Pd–WN/SBA-15, 1.00 g substrate, 30 mL isopropanol, 10 bar H2, 313 K. [b] Determined by GC and GC–MS.
ChemCatChem 2016, 8, 1718 – 1726
www.chemcatchem.org
1724
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
sample holders. XPS measurements were performed on a Kratos-
AXIS ULTRA DLD with an AlKa radiation source. The binding ener-
gies (BE) of elements were referenced to the C1s peak at 284.8 eV.
The content of palladium and tungsten nitride was determined by
ICP–OES, which was performed by a PerkinElmer Optima 7000DV
analyzer. Before the measurement, each sample was dissolved in
a diluted HF and chloroazotic acid solution. The content of tung-
sten nitride was also analyzed by X-ray fluorescence spectrometry
using a Bruker S4 Explorer instrument.
Selective hydrogenation of cinnamaldehyde
In a typical reaction, cinnamaldehyde (1.0 g), catalyst (50 mg), and
isopropanol (30 mL) were precharged in a 100 mL autoclave. The
autoclave was purged with N2 three times at 5 bar and H2 three
times at 5 bar to remove air, and the mixture was then magnetical-
ly stirred at 800 rpm under 10 bar of H2 at 313 K for 2 h. After the
reaction, the products were analyzed by GC (Agilent 7820A) with
flame ionization detection and a HP-5 capillary column (30 m”
0.32 mm”0.25 mm). Benzyl alcohol was used as an internal stan-
dard. The products were further identified by GC–MS (Agilent
6890/5973N). For the test of the stability, the catalyst used above
was recovered by centrifugation, washed three times with isopro-
panol, and then used for the next run without any further treat-
ment.
Scheme 2. Synthetic scheme for the construction of the ternary Pd–WN/
SBA-15 catalyst.
24 h, the excess N-[3-(trimethoxysilyl)propyl]-ethylenediamine was
removed by washing with ethanol. Finally, the SBA-15ÀNH2 was
obtained followed by drying at 333 K for several hours (step 1).
Subsequently, SBA-15ÀNH2 (0.50 g) was pretreated in an oven at
473 K for 2 h, and then was impregnated with the solution(10 mL
containing 0.05 g SiW12). After stirring for 6 h, the mixture was fil-
tered and washed. After drying at 333 K for 12 h, the precursor
SiW12/SBA-15ÀNH2 (step 2) was heated at 793 K with a heating rate
of 3 KminÀ1 and maintained for 3 h under pure NH3 (80 mLminÀ1).
After slowly cooling to RT under a N2 atmosphere, the WN/SBA-15
composite was obtained (step 3).
Acknowledgements
This work was financially supported by the National Natural Sci-
ence Foundation of China (No. 21371053, 21401048, 21571054),
the China Postdoctoral Science Special Foundation (No.
2015T80374), International Science & Technology Cooperation
Program of China (2014DFR41110), Natural Sciences Fund of Hei-
longjiang Province (B2015009), the Application technology re-
search and development projects in Harbin (2013AE4BW051), In-
novative Research Project of Key Laboratory of Functional Inor-
ganic Material Chemistry (Heilongjiang University), Ministry of
Education.
The ternary Pd–WN/SBA-15 was formed by loading Pd on WN/SBA-
15 composite by
a
sodium borohydride (NaBH4) method
certain amount of PdCl2 solution
(step 4).[39,40] Typically,
a
(10 mmolLÀ1) and WN/SBA-15 composite (0.30 g) were dissolved in
distilled water (10 mL), and the mixture was stirred for 12 h. 1m
NaOH was added to adjust the pH of the solution to ꢀ10. Then,
a certain amount of NaBH4 solution (0.2 mollÀ1) was added to this
suspension, and the molar ratio of NaBH4 to PdCl2 was 5. Finally,
Pd–WN/SBA-15 was separated by filtrations, washed with distilled
water and absolute ethanol in sequence and dried at 333 K over-
night in a vacuum oven. The theoretical content of Pd was 1.5%
and 0.75% and the catalysts were denoted as 1.5% Pd–WN/SBA-15
and 0.75% Pd–WN/SBA-15, respectively. As references, Pd support-
ed on SBA-15 with 1.5% Pd content was also prepared under the
similar conditions, which was denoted as 1.5% Pd/SBA-15.
Keywords: electron transfer
palladium · tungsten
· hydrogenation · nitrides ·
[1] K. Bauer, D. Garbe, Common Fragrance and Flavor Materials, VCH, Wein-
heim, 1985.
[2] H. G. Manyar, B. Yang, H. Daly, H. Moor, S. McMonagle, Y. Tao, G. D.
Yadav, A. Goguet, P. Hu, C. Hardacre, ChemCatChem 2013, 5, 506–512.
[3] H. L. Liu, L. N. Chang, L. Y. Chen, Y. W. Li, ChemCatChem 2016, 8, 946–
951.
[4] T. Szumełda, A. Drelinkiewicz, R. Kosydar, J. Gurgul, Appl. Catal. A 2014,
487, 1–15.
[5] P. Gallezot, D. Richard, Catal. Rev. Sci. Eng. 1998, 40, 81–126.
[6] J. J. Shi, M. Y. Zhang, W. C. Du, W. S. Ning, Z. Y. Hou, Catal. Sci. Technol.
2015, 5, 3108–3112.
Catalyst characterization
Wide-angle X-ray diffraction (XRD) of the catalysts was recorded in
the 2q range of 10–808 on a Bruker D8 Advance X-ray diffractome-
ter with CuKa (l=1.5418 Š) radiation (40 kV, 40 mA). As for the
small-angle XRD, the scan range was 0.5–58, and the scanning rate
was 18minÀ1. Nitrogen adsorption–desorption isotherms were col-
lected on a Micromeritics Tristar 3020 surface area and a porosime-
try analyzer for mesoporous materials at 77 K. The BET equation
was used to calculate the specific surface area. Pore size distribu-
tions were obtained using the Barrett-Joyner-Halenda (BJH)
method from the adsorption branch of the isotherm. SEM mapping
was analyzed by scanning electron microscopy (SEM: Hitachi S-
4800) with an acceleration voltage of 5 kV. TEM characterization
and high-resolution TEM (HRTEM) were determined by using
a JEM-2100 electron microscope (JEOL, Japan) with an acceleration
voltage of 200 kV. Carbon-coated copper grids were used as the
[7] K. R. Kahsar, D. K. Schwartz, J. W. Medlin, J. Mol. Catal. A 2015, 396, 188–
195.
[8] M. Bidaoui, C. Especel, S. Sabour, L. Benatallah, N. Saib-Bouchenafa, S.
Royer, O. Mohammedi, J. Mol. Catal. A 2015, 399, 97–105.
[9] Z. B. Tian, Q. Y. Li, Y. Li, S. Y. Ai, Catal. Commun. 2015, 61, 97–101.
[10] X. W. Ji, X. Y. Niu, B. Li, Q. Han, F. L. Yuan, F. Zaera, Y. J. Zhu, H. G. Fu,
ChemCatChem 2014, 6, 3246–3253.
[11] Z. Guo, Y. T. Chen, L. S. Li, X. M. Wang, G. L. Haller, Y. H. Yang, J. Catal.
2010, 276, 314–326.
ChemCatChem 2016, 8, 1718 – 1726
www.chemcatchem.org
1725
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[12] X. Yang, D. Chen, S. J. Liao, H. Y. Song, Y. W. Li, Z. Y. Fu, Y. L. Su, J. Catal.
2012, 291, 36–43.
[13] T. Szumełda, A. Drelinkiewicz, R. Kosydar, J. Gurgul, Appl. Catal. A 2014,
487, 1–15.
[14] J. S. J. Hargreaves, Coord. Chem. Rev. 2013, 257, 2015–2031.
[15] W. F. Chen, J. T. Muckerman, E. Fujita, Chem. Commun. 2013, 49, 8896–
8909.
[16] D. A. Ruddy, J. A. Schaidle, J. R. Ferrell, J. Wang, L. Moens, J. E. Hensley,
Green Chem. 2014, 16, 454–490.
[27] S. T. Oyama, J. Solid State Chem. 1992, 96, 442–445.
[28] R. H. Wang, Y. Xie, K. Y. Shi, J. Q. Wang, C. G. Tian, P. K. Shen, H. G. Fu,
Chem. Eur. J. 2012, 18, 7443–7451.
[29] V. Z. Radkevich, T. L. Senko, K. Wilson, L. M. Grishenko, A. N. Zaderko,
V. Y. Diyuk, Appl. Catal. A 2008, 335, 241–251.
[30] J. Yang, Y. Xie, R. H. Wang, J. B. Jiang, C. G. Tian, G. Mu, J. Yin, B. Wang,
H. G. Fu, ACS Appl. Mater. Interfaces 2013, 5, 6571–6579.
[31] W. Chen, Z. Fan, X. Pan, X. Bao, J. Am. Chem. Soc. 2008, 130, 9414–
9419.
[17] H. J. Yan, C. G. Tian, L. Sun, B. Wang, L. Wang, J. Yin, A. P. Wu, H. G. Fu,
Energy Environ. Sci. 2014, 7, 1939–1949.
[18] J. Yin, L. Wang, P. Yu, L. Zhao, C. G. Tian, B. J. Jiang, D. D. Zhao, W. Zhou,
H. G. Fu, ChemElectroChem 2015, 2, 1813–1820.
[19] D. Wang, Y. J. Zhu, C. G. Tian, L. Wang, W. Zhou, Y. L. Dong, Q. Han, Y. F.
Liu, F. L. Yuan, H. G. Fu, Catal. Sci. Technol. 2016, 6, 2403–2412.
[20] D. Y. Zhao, J. L. Feng, Q. S. Huo, N. Melosh, G. H. Fredrickson, B. F.
Chmelka, G. D. Stucky, Science 1998, 279, 548–552.
[21] A. Pineda, A. M. Balu, J. M. Campelo, A. A. Romero, D. Carmona, F. Balas,
J. Santamaria, R. Luque, ChemSusChem 2011, 4, 1561–1565.
[22] M. Ojeda, A. M. Balu, V. Barrón, A. Pineda, A. Garcia, A. A. Romero, R.
Luque, J. Mater. Chem. A 2014, 2, 387–393.
[32] X. Pan, X. Bao, Acc. Chem. Res. 2011, 44, 553–562.
[33] Z. Z. Wei, Y. T. Gong, T. Y. Xiong, P. F. Zhang, H. R. Li, Y. Wang, Catal. Sci.
Technol. 2015, 5, 397–404.
[34] A. Hammoudeh, S. Mahmoud, J. Mol. Catal. A 2003, 203, 231–239.
[35] G. R. Cairns, R. J. Cross, D. Stirling, J. Catal. 1997, 166, 89–97.
[36] P. Claus, Top. Catal. 1998, 5, 51–62.
[37] W. O. Oduro, N. Cailuo, K. M. K. Yu, H. W. Yang, S. C. Tsang, Phys. Chem.
Chem. Phys. 2011, 13, 2590–2602.
[38] X. Yang, L. P. Wu, L. L. Ma, X. J. Li, T. J. Wang, S. J. Liao, Catal. Commun.
2015, 59, 184–188.
[39] Y. Li, X. Xu, P. Zhang, Y. Gong, H. Li, Y. Wang, RSC Adv. 2013, 3, 10973–
10982.
[23] A. Kafizas, C. J. Carmalt, I. P. Parkin, Coord. Chem. Rev. 2013, 257, 2073–
2119.
[40] Y. Wang, J. Yao, H. R. Li, D. S. Su, M. Antonietti, J. Am. Chem. Soc. 2011,
133, 2362–2365.
[24] M. S. Bazarjani, M. M. Müeller, H. J. Kleebe, C. Fasel, R. Riedel, A. Gurlo, J.
Mater. Chem. A 2014, 2, 10454–10464.
[25] S. Abate, P. Lanzafame, S. Perathoner, G. Centi, Catal. Today 2011, 169,
167–174.
Received: February 16, 2016
[26] X. Yang, L. Du, S. Liao, Y. Li, H. Song, Catal. Commun. 2012, 17, 29–33.
Published online on April 18, 2016
ChemCatChem 2016, 8, 1718 – 1726
www.chemcatchem.org
1726
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim