Synergy between Palladium and Potassium Species for Efficient Activation of Carbon Monoxide in the Synthesis of Dimethyl Carbonate

Article abstract of DOI:10.1002/cctc.201500317

A potassium ion containing Pd/NaY catalyst is introduced that effectively activates the CO molecule for the carbonation of methyl nitrite to synthesize dimethyl carbonate. The potassium ions play a dual role during the activation of CO. Doping the catalyst with potassium enhances the electron density of Pd active species, which strengthens the Pd-C atoms interaction and thus more strongly activates the CO molecule. Further, the charge-balancing potassium cations in the zeolite interact with the O atom of linear adsorbed carbonyls to form a Pd-CO...K⁺ structure, which further activates the C-O bond. Both experimental analyses and density functional theory calculations indicated that the combination of potassium and Pd species facilitates the activation of CO in the carbonylation reaction.

Full text of DOI:10.1002/cctc.201500317

DOI: 10.1002/cctc.201500317
Full Papers
Synergy between Palladium and Potassium Species for
Efficient Activation of Carbon Monoxide in the Synthesis
of Dimethyl Carbonate
[
a]
Yuanyuan Dong, Yongli Shen, Yujun Zhao, Shengping Wang, and Xinbin Ma*
A potassium ion containing Pd/NaY catalyst is introduced that
effectively activates the CO molecule for the carbonation of
methyl nitrite to synthesize dimethyl carbonate. The potassium
ions play a dual role during the activation of CO. Doping the
catalyst with potassium enhances the electron density of Pd
active species, which strengthens the PdÀC atoms interaction
and thus more strongly activates the CO molecule. Further, the
charge-balancing potassium cations in the zeolite interact with
+
the O atom of linear adsorbed carbonyls to form a PdÀCO···K
structure, which further activates the CÀO bond. Both experi-
mental analyses and density functional theory calculations indi-
cated that the combination of potassium and Pd species facili-
tates the activation of CO in the carbonylation reaction.
Introduction
Dimethyl carbonate (DMC) can be widely used in versatile
chemical syntheses owing to its great reactivity towards nucle-
A potassium promotion effect on supported noble-metal
catalysts has been widely reported in the literature. Doping
a Pd/C catalyst with potassium ions increased the rate of hy-
drogen production by 1–2 orders of magnitude from vapor-
[
1]
ophile molecules. In addition, it is nontoxic, safe, and excel-
lently soluble, thus can be extensively used as a solvent in the
paint and battery industry and as an oxygen-containing fuel
[
6]
phase formic acid decomposition. The addition of potassium
ions activates Pt absorbed on silica or alumina and, conse-
[
2]
additive. Vapor-phase carbonylation of methyl nitrite (MN)
catalyzed by supported Pd-based catalysts has been consid-
[
7]
quently, the low-temperature water gas shift reaction activity.
[3]
ered a promising path to produce DMC. Previous researches
reported that active-carbon-supported PdCl –CuCl is one of
The facilitation effects considered are deviations in the elec-
tronic and structural nature of the noble metals caused by the
doping potassium species, which is beneficial to the activation
of reactants. Therefore, the introduction of potassium ions into
the Pd/NaY catalyst probably further activates the CO molecule
for the carbonylation of MN to synthesize DMC. Additionally,
2
2
[1a]
the most active catalysts. However, the deactivation of the
catalyst caused by the loss of chloride is still the main bottle-
neck. In addition, chloride-containing systems usually suffer
from many problems such as the color of DMC, the corrosion
of reaction apparatus, and so on. Chloride-free Pd/NaY cata-
lysts, prepared by conventional liquid-ion exchange methods,
have attracted much attention on the account of their good
[
8]
Silvia Bordiga et al. reported that the electric field induced by
alkali metal cations polarized the CO molecule. This polariza-
tion is thus speculated to facilitate the activation of the CO
molecule to some extent. Aimed to effectively activate CO, it is
therefore of interest to combine potassium ions and Pd spe-
cies in the catalyst (e.g., by decorating the Pd/NaY catalyst
with potassium ions).The cooperative activation is expected to
promote CO for better participation in the reaction, thereby
obtaining excellent catalytic performance. In this way, it is con-
venient to control the reactivity by adjusting the amount of
noble metal and/or alkali metal.
[
3,4]
catalytic stability.
However, their catalytic activity is not as
good as that of PdCl –CuCl supported on activated carbon.
2
2
Therefore, the modification of Pd/NaY catalyst is necessary to
obtain better carbonylation reactivity.
For the carbonylation of MN to synthesize the DMC system,
[
5]
the activation of CO is one of the most important factors.
Hence, it is desirable to address this issue to improve carbony-
lation reactivity. Adding additives is usually considered as an
effective means. However, there have been limited studies on
the introduction of dopants into the Pd/NaY catalyst.
A series of potassium-promoted Pd/NaY catalysts were pre-
pared through liquid-ion-exchange technology. The catalysts
were evaluated under the same condition in a fixed-bed micro-
reactor. The results clearly demonstrated an improvement in
catalytic activities that were achieved by the addition of potas-
[
a] Y. Dong, Dr. Y. Shen, Dr. Y. Zhao, Prof. Dr. S. Wang, Prof. Dr. X. Ma
Key Laboratory for Green Chemical Technology of Ministry of Education
Collaborative Innovation Centre of Chemical Science and Engineering
School of Chemical Engineering and Technology
Tianjin University
sium to Pd/NaY catalyst. Low-temperature N adsorption, TEM,
2
and XRD studies were performed to detect the physical nature
of unpromoted and potassium-promoted catalysts. The chemi-
cal properties of all samples were determined by X-ray photo-
electron spectroscopy (XPS), CO temperature-programmed de-
sorption (TPD) and CO adsorption IR spectroscopy. Density
Tianjin 300072 (P.R. China)
E-mail: xbma@tju.edu.cn
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201500317.
ChemCatChem 2015, 7, 2460 – 2466
2460
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
functional theory (DFT) calculations were employed to uncover
the underlying causes of potassium promotion effect in more
detail.
images (Figure 1) depict the morphology of Pd active species
in the catalysts. The well-defined {200} and {111} facets identi-
[
9]
fy the particles on all catalysts. Active-species Pd particles
uniformly disperse in NaY zeolite, which provide a clue to the
high reactivity of these materials. Compared with the original
sample, the addition of potassium did not cause any relative
enlargement of the supported Pd particles. In other words, the
Pd particle size is found to be independent of the presence
and abundance of potassium species, excluding variations in
diameter as a possible cause of the differences in the observed
catalytic activity.
Results and Discussion
In Table 1 and Figure S1 (XRD results, Supporting Information),
the physical properties and crystal structures of the undoped
and doped K catalysts are shown. Only slight changes are ob-
served relative to NaY zeolite, indicating that doping insignifi-
cantly impacts the textural properties of the catalyst. To discuss
the effects of potassium on carbonylation performances, the
catalysts labeled as Pd/KNaY(i) (i represents the K/Pd molar
ratio) were evaluated under the same reaction conditions in
a fixed-bed microreactor. The results (Table 1) indicate that the
doped catalysts differ in activity from the undoped ones owing
to the amount of potassium. The space time yield (STY) of
It has been shown in the literature that the introduction of
potassium ions modifies the electronic and structural proper-
ties of supported noble metals. Thus, XPS was performed to
determine the changes in the electronic nature of Pd species
caused by the potassium dopant (Figure 2). The peaks at ap-
proximately 338 eV and approximately 335 eV are assigned to
À1 À1
DMC reaches up to 696 gL h ,
an increase of 25% in compari-
son with the STY achieved over
undoped catalyst. The selectivity
to DMC increases from 80.8% to
8
6.3%. Notably, the turnover fre-
quency (TOF, calculated by the
moles of DMC per mole Pd per
hour) increases steadily at low K/
Pd molar ratios, reaching a maxi-
mum at K/Pd=3.6, and then de-
creases as this ratio further in-
creases. The STY and selectivity
of DMC follow the same trend,
but the STY and selectivity of MF
exhibit a contrary tendency with
the potassium loading. The
DMM results are slightly different
because of the limitations of
zeolite pores.
To explore the potassium pro-
motion effect in the carbonyla-
tion reaction, a series of charac- Figure 1. a–d) TEM images of undoped and K-doped Pd/NaY catalysts and f) their crystal facets.
terizations were employed. TEM
[
a]
[b]
[c]
Table 1. The structural properties and catalytic activity of NaY zeolite and Pd/KNaY(i) catalysts.
[
d]
[e]
[f]
À1 À1
[g]
[h]
[i]
Catalysts
Pd loading
wt%]
Dispersity
[%]
TOF
[h
STY [gL
h
]
Sel. based on MN [%]
S
BET
2
Pore vol.
Pore width
[Š]
E
a
À1
À1
À1
À1
[
]
[m g
]
[mlg
]
[kJmol ]
DMC DMM MF
DMC
DMM
MF
Pd/NaY
0.99
91.8
88.9
90.3
91.1
89.6
167.5 550
180.3 613
196.9 696
162.1 567
153.0 519
33.6
37.8
35.9
38.6
37.3
32.5 80.8
22.7 83.8
21.5 86.3
31.5 80.0
33.3 78.6
6.2
5.9
5.4
9.3
8.1
13.0
10.3
8.3
10.7
13.3
699
710
730
685
650
0.29
0.33
0.32
0.31
0.29
8.5
8.0
7.7
7.5
7.3
60.3Æ1.2
42.5Æ1.0
33.4Æ0.9
37.6Æ0.8
37.0Æ1.4
Pd/KNaY(1.8) 0.98
Pd/KNaY(3.6) 0.97
Pd/KNaY(4.9) 0.97
Pd/KNaY(6.5) 0.98
À1
[
2
a] Reaction conditions: gas hourly space velocity (GHSV)=8000 h , CO/MN/N =3.3:20:76.7 (vol.%), T=383 K, P=0.2 MPa. [b] Physical properties of NaY
2
À1
À1
zeolite: SBET: 698 m g , Pore vol.: 0.31 mLg , Pore width: 8.6 Š. [c] Pd/KNaY(i) catalysts: i represents the molar ratio of K/Pd, KNaY denotes the K-doped
NaY zeolite as the carrier to load the Pd active species. [d] Determined by ICP–OES. [e] Determined by CO chemisorption characterization. [f] Calculated by
the moles of DMC per mole Pd per hour. [g] Pore volume was determined by the t-plot method. [h] Pore width was calculated using the Horvath–Kawazoe
a
method. [i] E was calculated based on Arrhenius equation.
ChemCatChem 2015, 7, 2460 – 2466
www.chemcatchem.org
2461
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
and (III) decrease, whereas the area of peak (II) increases. The
dominant moderate-level adsorption observed suggests that
the introduction of potassium modulates the CO adsorption
strength such that CO molecule is more able to participate in
[
12]
subsequent reaction. Moreover, the CO desorption peaks (I)
and (II) (Figure 3) move to higher temperature at the lower K/
Pd molar ratio until the ratio reaches 3.6, and then shifts in the
opposite direction. Peak (III) also drifts to higher temperatures
but disappears when K/Pd exceeded 3.6. These phenomena
were caused by changes in the Pd–CO interaction intensity.
Specifically, these results indicate that the Pd–CO interaction
strength is improved by adding a certain amount of potassium
ions, but it is counterproductive if added in excess.
Further experiments were conducted to understand the in-
fluences of potassium on the adsorption properties of CO mol-
ecule in more detail. In Figure 4, the FTIR spectra of CO ad-
sorbed onto the undoped and K-doped catalysts are shown.
Figure 2. Pd3d XPS spectra of undoped and K-doped Pd samples.
[10]
the oxidation and metallic states of Pd, respectively. The two
states of Pd species are clearly distinguished for the original
Pd/NaY catalyst. With an increasing potassium level, the peaks
of oxidation and metallic states of Pd begin to overlap and fi-
nally become a single peak. Additionally, the band energies of
Pd species gradually decline regardless of the oxidation and
metallic state, which suggests that the electron density of Pd
[11]
species is progressively enhanced. Therefore, the assumption
is proposed that the addition of potassium decorates the elec-
tronic nature of Pd active species and, consequently, the car-
bonylation reactivity.
For the MN carbonylation reaction system, the activation of
CO plays an important role. In the light of the observed effects
of potassium on the electronic properties of Pd species, we
speculate this function may further affect desorption and/or
adsorption of CO on the active species, and finally influences
the catalytic performances. The CO TPD was thus employed to
verify the conjecture. In Figure 3, the CO TPD spectra and the
results of peak fitting for undoped and doped catalysts are
shown. The spectrum of parent sample is composed of three
peaks. With increasing potassium levels, the areas of peak (I)
Figure 4. FTIR spectra of CO (1 atm) adsorbed onto Pd/NaY and K-promoted
À1
À1
Pd/NaY catalysts at 258C. a) 2200–2080 cm , b) 2082–2000 cm , c) 2008–
À1
À1
1800 cm , d) 1720–1800 cm
.
À1
The bands at approximately 2170, 2162, and 2100 cm are at-
+
+
tributed to carbonyls that are C-end-adsorbed on Na and K ,
+
[13]
and to those O-end-adsorbed on Na , respectively. Three
main categories of carbonyls on Pd particles can thus be iden-
[
8]
tified (Figure S3). (i) Linear carbonyls (top site) usually appear
À1
in the stretching region at approximately 2130–2000 cm ;
À1
À1
bands at approximately 2070 cm and 2028 cm are assigned
to CO adsorbed on terrace and edges or corners of particles,
[
14]
respectively. (ii) Two-fold bridged carbonyls are defined by
À1
a n(CO) in the interval at approximately 2000–1850 cm ;
À1
À1
À1
bands at 1990 cm and 1950 cm , 1910 cm represent CO
adsorbed on the {100} and {111} faces of Pd crystals, respec-
[
14b]
tively (Figure 1(f)).
(iii) Three-fold bridged carbonyls (hollow
site), are characterized by a CÀO stretching mode in the region
À1 [8,11]
at approximately 1920–1800 cm .
As shown in Figure 4,
the bands of the linear carbonyls adsorbed on the corners of
À1
Pd particles (ꢀ2026 cm ) redshift first and then blueshift as
the K/Pd atomic ratio increases. The bands of the carbonyls lin-
À1
early adsorbed on terrace sites (ꢀ2070 cm ) and those of 2-
Figure 3. CO TPD spectra of Pd/NaY catalysts with different K/Pd mole ratio.
and 3-fold bridged carbonyls did not seem to be affected by
ChemCatChem 2015, 7, 2460 – 2466
www.chemcatchem.org
2462
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
changes in the metal K/Pd atom ratio. Another noteworthy fea-
ture is that augmenting the potassium loading of the catalyst
caused a volcano-shaped CO adsorption peak (Figure 4b and
c).
To examine in more detail about the interaction among po-
tassium ions, Pd species, and CO molecule, we performed DFT
calculations by using the DMol3 code from Materials Studio
[
16]
3.0. A Pd13 cluster is selected as the active-site model of this
carbonylation reaction (Figure 6c). The well described 16 tetra-
hedral (T) cluster including two six-ring and three four-ring
structures is chosen as the Y-type zeolite support model (Fig-
ure 6b). (The computational details, selection of active sites,
and support models are given in the Experimental Section).
The charge imbalance owing to the presence of aluminum in
this 16-T cluster model of zeolite framework is compensated
by sodium cations.
The phenomena seen in Figure 4 are the result of the varia-
tions in the interaction intensity of Pd–CO. After decorating
the catalyst with potassium, the electron density of the Pd
active surface sites is enhanced, this is well evidenced by the
results of XPS (Figure 2). The increased electron density of Pd
species can enhance back-donation from Pd to antibonding or-
bitals of CO, which weakens the CÀO bond and thus induces
[
8]
a bathochromic shift of the vibration band (Figure 4b). If
excess potassium species are added, the electron density of Pd
is enhanced even further. The exorbitant electron density de-
creases the electron donation from the highest occupied mo-
lecular orbital (HOMO) of CO to the metal (s-bond) to hinder
[11]
the adsorption of molecular CO, resulting in the blueshifting
of the band of the linearly adsorbed CO and the disruption of
CO adsorption on the surface.
Additionally, in Figure 5 the relationships between the areas
of CO adsorption peaks in Figure 4 and the catalytic perform-
ances is shown. The TOF of DMC presents a \-shape with the
peak areas of 2- and 3-fold bridged carbonyls, whereas it mo-
notonically increases with the peak area of linear carbonyls ad-
sorbed on the corner site of Pd clusters. The case of the domi-
nant adsorption of linear carbonyls is unusual for carbonylation
Figure 6. Results from DFT calculations: a) Optimized structure of Pd cluster
embedded in the supercage of Y-type zeolite. b) Schematic view of the 16T
cluster model for the Y-type zeolite. c) The optimized structure of Pd/Y cata-
lyst (the single Pd atom was used to balance the charge in the Y-type zeolite
structure with Si/Al=2.2).
[
15]
reactions. It can be inferred that the linear carbonyls ad-
sorbed on the corner sites of Pd particles are more active par-
ticipators for this carbonylation reaction thanks to some inter-
action with the alkali metal cations than the 2- and 3-fold
bridged carbonyls. Cairon et al. studied adsorption of CO on
NaY and indicated that alkali metal cations interact with CO
The Si/Al ratio (2.2) and the number and locations of charge-
balancing cations are in accordance with experimental find-
ings. The sodium cations are located in the supercage: (1) at
the center of sodalite but displaced into the supercage; (2) at
the center of hexagonal prisms but migrated into the super-
+
[17]
through the C end and/or O end, in the form of M ···CO and/
cage. The positions of sodium cations after optimization are
+
[13]
or M ···OC adducts. Accordingly, the combination of potassi-
um and palladium species is speculated to be responsible for
the efficient activation of linearly adsorbed CO molecule. The
shown in Figure S2. The potassium-doped catalysts are formed
by gradually replacing the Na atom with the K atom, as illus-
trated in Figure S2d–f.
+
+
À1
appearance of PdÀCO···Na /K (ꢀ1766/1740 cm ) peaks in
A charge analysis was first performed for all optimized struc-
tures, which is listed in Table 2. The charges of Pd atoms are
positive in Pd/Y catalyst. As the doping amount of alkali metal
the spectra in Figure 4(d) may be a good indication of the syn-
[14b]
ergy between Pd and potassium species.
(
Na and/or K) increases, the negative charges on Pd atoms
gradually increase, especially for Pd atom numbers 2, 3, 4, 5
and 13. As these Pd atoms located on the top layer of cluster
(
seen in Figure S2) are relatively closer to alkali metal, the
doped alkali metal atoms mainly influence their electron densi-
ty. The enhanced negative charges on Pd verify that alkali
metal atoms donate electrons to Pd, which is consistent with
the XPS results. Additionally, it is clearly seen that the number
of negative charges on Pd atoms increases stepwise if the
sodium atoms are progressively substituted by potassium
atoms (Table 2), which suggests that the influence of potassi-
um on the electron density of Pd atoms is much larger than
that of sodium. This observation provides some insight into
the interpretation of the promotion behavior of potassium-
doped Pd/NaY catalyst.
To clarify the influences of enhanced electronic density of Pd
on reactivity, the properties of CO adsorption/desorption on
model catalysts were further studied in depth. According to
Figure 5. Relationship between the areas of CO adsorption peaks for the
three adsorption manners of a) hollow site, b) bridge site, and c) linear site
and the TOF of DMC for K-doped Pd/NaY catalysts.
ChemCatChem 2015, 7, 2460 – 2466
www.chemcatchem.org
2463
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Table 2. Charge distribution of Pd atoms for alkali metal doped/undoped
model catalysts.
[a]
[b]
Entry
Pd
2Na
0.012
3Na
K2Na
2KNa
3K
1
2
3
4
5
6
7
8
9
0.116
0.025
0.007
0.010
0.015
0.061
0.154
0.173
0.119
0.166
0.165
0.214
0.370
À0.059
À0.191
À0.046
À0.010
À0.016
0.023
0.169
0.018
0.097
0.129
À0.030
À0.192
À0.061
À0.025
À0.031
À0.004
0.158
0.004
0.085
0.107
0.129
À0.016
À0.186
À0.066
À0.026
À0.034
À0.021
0.165
À0.008
À0.210
À0.071
À0.034
À0.033
À0.005
0.158
0.001
0.082
0.099
0.121
À0.093
À0.020
0.002
0.001
0.071
0.176
0.095
0.112
0.137
0.150
0.263
0.034
À0.004
0.080
1
0
0.100
0.123
0.293
À0.003
1
1
0.142
0.272
0.003
1
1
2
3
0.301
À0.009
0.296
À0.020
[
a] The number represents the number of Pd atoms in the model cata-
lysts (Figure 6c). [b] Pd, 2Na, 3Na, K2Na, 2KNa, and 3K denote the cata-
lysts of Pd/Y, Pd/2NaY, Pd/3NaY, Pd/K2NaY, Pd/2KNaY, and Pd/3KY, re-
spectively. 2 and 3 denote the number of alkali metal atom in catalyst.
Figure 7. Changes of the CO bond order given as the concentration of alkali
cations on the model catalysts. Pd, 2Na, 3Na, K2Na, 2KNa, and 3K repre-
sent the model catalyst of Pd/Y, Pd/2NaY, Pd/3NaY, Pd/KNaY, Pd/2KNaY,
and Pd/3KY, respectively. 2 and 3 denotes the number of alkali metal atoms
in the model catalyst.
the charge distribution of Pd atoms and the principles of elec-
trostatic attraction, Atom numbers 2, 3, and 13 were chosen as
the adsorption sites of CO molecule. Owing to the three differ-
ent adsorption manners of the CO molecule, we investigated
the cases: top-site (linear adsorption), bridge-site, and hollow
site adsorption for each model catalyst. The optimized struc-
tures are illustrated in Figure S3. In Table 3 and Figure 7, the
adsorption properties of CO on undoped and doped model
catalysts are illustrated. The desorption energy of CO mole-
cules (Edes) for Pd/Y model catalyst is 1.58 eV consistent with
doping Na atoms. These observations echo the results of CO
TPD. With increasing alkali metal level, the bond order of ad-
sorbed CO clearly decreases in comparison with undoped Pd/Y
catalyst. These results indicate that CO is more strongly activat-
ed by additional doped alkali metal. The data of PdÀC bond
parameters also indirectly reflects the activation of CO. It is
noteworthy that the bond order of CO diminishes even more
when Na atoms are gradually replaced by K atoms, but it will
rebound after complete replacement with K atoms (Table 3
and Figure 7). These findings suggest that the introduction of
K atom more facilitate the activation of CO than the addition
of Na to Pd/Y model catalyst, which is a strong evidence of K
promotion effect in catalytic activities.
[18]
the data reported in the literature. If Na and/or K atoms are
added to Pd/Y model catalyst, the Edes of the CO molecule dis-
plays a volcano shape, which implies that the interaction
strength between Pd and CO increases first and then decreas-
es. The impact of doping K atoms on Edes is larger than that of
Table 3. Desorption energy of CO; bond length, and bond order of CÀO and CÀPd for alkali-doped Pd/Y model catalysts.
[
a]
Model catalysts
Adsorption
manner
CÀPd2
CÀPd3
CÀPd13
CÀO
Desorption
Energy[d] [ev]
[
b]
[c]
BL
BO
BL
BO
BL
BO
BL
BO
top
1.884
1.995
2.069
1.846
1.978
2.094
1.840
1.991
2.084
1.838
1.998
2.064
1.837
1.996
2.069
1.833
1.978
2.056
1.28
0.79
0.56
1.36
0.75
0.52
1.54
0.78
0.53
1.57
0.79
0.56
1.58
0.80
0.56
1.56
0.81
0.58
–
–
0.76
0.59
–
–
–
–
0.56
–
–
0.60
–
–
0.59
–
–
0.62
–
–
0.64
–
–
1.154
1.177
1.191
1.161
1.176
1.188
1.182
1.181
1.197
1.183
1.180
1.200
1.186
1.183
1.200
1.181
1.181
1.206
2.21
2.02
1.90
2.13
2.02
1.88
1.86
1.95
1.83
1.84
1.98
1.82
1.82
1.97
1.82
1.89
1.95
1.78
Pd/Y
bridge
hollow
top
bridge
hollow
top
bridge
hollow
top
bridge
hollow
top
bridge
hollow
top
bridge
hollow
2.000
2.063
–
1.999
2.062
–
2.010
2.073
–
2.007
2.091
–
2.013
2.103
–
1.999
2.109
1.58
1.63
2.23
2.25
2.29
2.12
2.080
–
–
2.074
–
–
2.074
–
–
2.068
–
–
Pd/2NaY
Pd/3NaY
Pd/K2NaY
Pd/2KNaY
Pd/3KY
0.78
0.60
–
0.80
0.61
–
0.80
0.58
–
0.79
0.57
–
2.065
–
–
0.80
0.57
2.063
0.64
[
a] CÀPd2 represents the bond between C atom and Pd atom No. 2. [b] BL indicates bond length. [c] BO indicates bond order. [d] CO desorption energy
calculated by the following equation: Edes =E(cat.)+E(CO)ÀE(cat.–CO), where E(cat.–CO), E(cat.), and E(CO) are the total energies of each structure.
ChemCatChem 2015, 7, 2460 – 2466
www.chemcatchem.org
2464
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Furthermore, the bond length and bond order of top-site-
adsorbed CO display \- and U-shape, respectively. However,
the data are irregular for CO adsorption on bridge-site and
hollow-site.
interaction between O and Na/K species are closed-shell (elec-
[
20]
trostatic) interactions. Potassium cations display the proper-
ties of Lewis acids in this case based on the definition of Lewis
acids. According to the above experimental (Figure 4d) and
theoretical analyses (Figure 8 and Figure S4), it can be conclud-
The U-shape of CO bond order for top sites as a function of
the potassium concentration implies that the activation extent
of CO first increases and then decreases. In other words, the
activation of CO is related to the concentration of potassium
atoms. Doped certain amount of potassium can enhance the
activation of CO, but excessive doping will lead to an opposite
result. These results correlate well with the trends observed in
activity data and characterization analyses (Figure 3 and 4).
Moreover, it is surprising to find that the bond orders of CO
adsorbed on top sites for alkali metal doped model catalysts
are atypically low (Figure 7), which distinguishes the case of
undoped Pd/Y model catalyst. This unusual situation suggests
that top-site-adsorbed CO molecules are much more activated
than bridge- and hollow-site-adsorbed ones, coinciding with
the prevailing linear carbonyls in Figure 4. Under the above
analyses, it is reasoned that the exceptional situation of dwin-
dling bond order for linear carbonyls is a cooperative conse-
quence of Pd species and alkali metal. To further confirm this
idea, a topological study on the electronic charge density dis-
tribution was conducted using the atoms in molecules (AIM)
+
+
ed that the structure of PdÀCO···Na /K indeed exists in K-pro-
moted Pd/NaY catalysts during activation of CO molecule. As
such, the activation of linearly adsorbed CO molecule is not
only ascribed to its interaction with Pd species but also to the
interaction with potassium cations as Lewis acid sites in zeo-
lite. This observed synergy between Pd and alkali metal is
absent for linearly adsorbed carbonyls on terrace sites as well
as for CO bound to bridge and hollow sites, reflected in the ir-
regularity of bond orders (Figure 7).
The CO molecule is activated and thus energetically partici-
pates in the carbonylation reaction. In turn, the apparent acti-
vation energies (E ) of MN carbonylation reaction should be re-
a
duced. To test this value, each catalyst was evaluated at differ-
ent temperatures (373 K, 383 K, and 393 K). The logarithm of
STY was correlated with the reciprocal of temperature to calcu-
late E based on Arrhenius equation. The E (Table 1) for doped
a
a
samples is, as expected, lower than that of the undoped sam-
[
5b]
ple. It is therefore concluded that adding a certain amount
of potassium facilitates the activation of CO and, consequently,
the rate of the carbonylation reaction.
[
19]
analysis. As shown in Figure 8 and S4, a bond critical point
BCP) is found between sodium/potassium and the O atom of
(
linear CO. For all CO linear adsorption structures, the local
topological properties (in a.u.) of the electronic charge density
distribution and the Laplacian of electron density are shown in
Conclusions
We have successfully enhanced the catalytic performance of
Pd/NaY catalyst for the carbonylation reaction of methyl nitrite
by adding potassium dopants to the zeolite structure. The
origin of the outstanding reactivity was unraveled through
a combination of experimental characterizations and DFT cal-
culations. Doping the structure with potassium enhanced the
electron density of Pd species in the catalyst, leading to an im-
provement in the activation of CO. Moreover, the potassium
cations, presenting the nature of Lewis acid sites in zeolite, in-
teracted with the O atom of CO linearly adsorbed onto the
corner site of the Pd cluster, further activating the adsorbed
CO entities. These two activation ways work together to allow
CO to easily participate in the carbonylation reaction, in turn,
promote the synthesis of dimethyl carbonate. This paper dem-
onstrated a synergy effect between potassium and Pd active
species in the carbonylation of methyl nitrate to dimethyl car-
bonate in detail. Understanding this cooperativity is conducive
to comprehending the mechanism for this carbonylation reac-
tion, and consequently to aiding in the design and optimiza-
tion of catalysts.
the Supporting Information, Table S1 and Figure S4. The posi-
2
tive values of ! 1 for this BCP indicate that the interatomic
b
Experimental Section
Preparation of catalysts
NaY zeolite was first calcined at 5508C for 5 h in air at a ramp rate
À1
of 28Cmin . The K-doped NaY support (KNaY) was prepared by
Figure 8. Molecular graphs for CO linearly adsorbed on a) Pd/3NaY, b) Pd/
K2NaY, c) Pd/2KNaY and d) Pd/3KY catalysts. Large circles correspond to the
nuclei. Lines connecting the nuclei indicate bond paths. Small dots on lines
are the BCP.
a conventional liquid-phase ion-exchange technique. NaY zeolite
powder was dispersed into a previously prepared KCl solution and
then stirred vigorously for 12 h at 808C. To remove excess chloride
ChemCatChem 2015, 7, 2460 – 2466
www.chemcatchem.org
2465
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
ions, the exchanged samples were thoroughly washed with deion-
ized water, dried at 110 8C for 2 h, and calcined at 5508C for 5 h.
The process of loading Pd active component was similar. Notably,
the Pd(NH ) Cl solution (0.01m) was added dropwise to KNaY
[3] Y. Dong, S. Huang, S. Wang, Y. Zhao, J. Gong, X. Ma, ChemCatChem
2013, 5, 2174–2177.
[
[
[
4] Y. Yamamoto, T. Matsuzaki, S. Tanaka, K. Nishihira, K. Ohdan, A. Naka-
mura, Y. Okamoto, J. Chem. Soc. Faraday Trans. 1997, 93, 3721–3727.
5] a) Y. Yamamoto, Catal. Surv. Asia 2010, 14, 103–110; b) A. Nakamura, T.
Matsuzaki, Res. Chem. Intermed. 1998, 24, 213–225.
3
4
2
slurry at 808C and then magnetically stirred for 8 h. The calcination
procedure was performed at 2008C for 4 h at a ramp rate of
6] D. A. Bulushev, L. Jia, S. Beloshapkin, J. R. H. Ross, Chem. Commun. 2012,
À1
0
.58Cmin in a muffle furnace. The samples were designated by
4
8, 4184–4186.
the label Pd/KNaY(i) with i representing the molar ratio of potassi-
um and palladium. The loadings of Pd and the K/Pd molar ratios
were analyzed by inductively coupled plasma optical emission
spectroscopy (ICP–OES). The details of catalysts characterization
method (BET, XRD, CO TPD, XPS, ICP and CO IR) and the catalytic
activity test are described in the Supporting Information.
[7] Y. Zhai, D. Pierre, R. Si, W. Deng, P. Ferrin, A. U. Nilekar, G. Peng, J. A.
Herron, D. C. Bell, H. Saltsburg, M. Mavrikakis, M. Flytzani-Stephanopou-
los, Science 2010, 329, 1633–1636.
[
8] C. Lamberti, A. Zecchina, E. Groppo, S. Bordiga, Chem. Soc. Rev. 2010,
9, 4951–5001.
9] B. Lim, M. Jiang, J. Tao, P. H. C. Camargo, Y. Zhu, Y. Xia, Adv. Funct. Mater.
009, 19, 189–200.
3
[
2
[
[
10] a) Z. S. Seddigi, A. Bumajdad, S. P. Ansari, S. A. Ahmed, E. Y. Danish, N. H.
Yarkandi, S. Ahmed, J. Hazard. Mater. 2014, 264, 71–78; b) D. Teschner,
J. Borsodi, A. Wootsch, Z. RØvay, M. Hävecker, A. Knop-Gericke, S. D.
Jackson, R. Schlçgl, Science 2008, 320, 86–89.
11] G. Pekridis, N. Kaklidis, M. Konsolakis, E. F. Iliopoulou, I. V. Yentekakis,
G. E. Marnellos, Top. Catal. 2011, 54, 1135–1142.
Details of DFT calculations
DFT calculations were performed using DMol3 code from Materials
Studio 3.0 with the exchange and correlation energy functional ex-
pressed in the Perdew–Wang (PW91) generalized gradient approxi-
[
[
12] P. Panagiotopoulou, D. I. Kondarides, J. Catal. 2009, 267, 57–66.
13] O. Cairon, A. Loustaunau, J. Phys. Chem. C 2008, 112, 18493–18501.
[21]
mation (GGA). A double numerical plus polarization (DNP) basis
set was used throughout the calculation. The convergence criteria
[14] a) R. Watanabe, Y. Sakamoto, K. Yamamuro, S. Tamura, E. Kikuchi, Y.
Sekine, Appl. Catal. A 2013, 457, 1–11; b) L. F. Liotta, G. A. Martin, G. De-
ganello, J. Catal. 1996, 164, 322–333; c) E. Jeroro, V. Lebarbier, A. Datye,
Y. Wang, J. M. Vohs, Surf. Sci. 2007, 601, 5546–5554.
[15] X. Zhao, Q. Lin, W. Xiao, Appl. Catal. A 2005, 284, 253–257.
[16] a) B. Delley, J. Chem. Phys. 1990, 92, 508–517; b) B. Delley, J. Chem.
Phys. 2000, 113, 7756–7764.
À5
À1
were set to be 2”10 Ha, 0.004 HaŠ , and 0.005 Š for energy,
force, and displacement convergence, respectively. A self-consis-
tent field (SCF) density convergence with a threshold value of 1”
À5
1
0
Ha was specified. All wave function (wfn) files used in this
[22]
work were generated by gaussian03 software. The topological
graphs were plotted by Multiwfn software.
[23]
[
17] a) A. Martin-Calvo, J. B. Parra, C. O. Ania, S. Calero, J. Phys. Chem. C
2014, 118, 25460–25467; b) S. Buttefey, A. Boutin, C. Mellot-Draznieks,
A. H. Fuchs, J. Phys. Chem. B 2001, 105, 9569–9575.
18] M. Hirsimäki, M. Valden, J. Chem. Phys. 2001, 114, 2345–2354.
19] R. F. W. Bader, Chem. Rev. 1991, 91, 893–928.
2
+
Initially exchanged Pd(NH3)4 ions were gradually reduced by CO,
migrated, agglomerated and eventually formed Pd clusters during
[
[
[
[4]
the carbonylation reaction. Using studies about the formation of
Pd particles in NaY and the pore size (13 Š) of FAU zeolite as
guides, Pd13 cluster (8.5 Š) was selected as the carbonylation active
20] M. F. Zalazar, N. M. Peruchena, J. Phys. Chem. A 2007, 111, 7848–7859.
[21] J. P. Perdew, J. A. Chevary, S. H. Vosko, K. A. Jackson, M. R. Pederson,
D. J. Singh, C. Fiolhais, Phys. Rev. B 1992, 46, 6671–6687.
[24]
site model. Additionally, owing to steric effects (Figure 6a), the
[17a,25]
locations of Pd cluster and charge-balanced cations,
scribed 16 tetrahedral (T) cluster consisting of two 6-ring and three
-ring structures was used as the model of Y-type zeolite. All dan-
a well de-
[22] Gaussian 03, Rev. E. 01, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, J. J. A. Montgomery, T. Vreven,
K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone,
B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsu-
ji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Na-
kajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P.
Hratchian, J. B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Strat-
mann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y.
Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakr-
zewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick,
A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G.
Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P.
Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. AlLaham, C. Y.
Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Pittsburgh
PA, 2004.
4
gling bonds were saturated with hydrogen atoms and the terminal
H atoms were oriented in the direction of next tetrahedral atom.
Acknowledgements
Financial support from the National Science Foundation of China
(
20936003, 21325626) is gratefully acknowledged and thanks
Prof. Xianggao Li for the detailed discussions about DFT part.
Keywords: carbonylation · density functional calculations ·
doping · palladium · potassium
[
23] T. Lu, F. Chen, J. Comput. Chem. 2012, 33, 580–592.
[24] T. Beutel, Z. Zhang, W. M. H. Sachtler, H. Knoezinger, J. Phys. Chem.
993, 97, 3579–3583.
25] a) Y.-X. Jiang, W.-Z. Weng, D. Si, S.-G. Sun, J. Phys. Chem. B 2005, 109,
1
[
[
[
1] a) N. Keller, G. Rebmann, V. Keller, J. Mol. Catal. A 2010, 317, 1–18; b) P.
Tundo, M. Selva, Acc. Chem. Res. 2002, 35, 706–716.
2] a) M. Richter, M. Fait, R. Eckelt, E. Schreier, M. Schneider, M. M. Pohl, R.
Fricke, Appl. Catal. B 2007, 73, 269–281; b) K. Inamoto, C. Hasegawa, K.
Hiroya, Y. Kondo, T. Osako, Y. Uozumi, T. Doi, Chem. Commun. 2012, 48,
7
2
637–7642; b) F. Li, Q. Zhang, Y. Wang, Appl. Catal. A 2008, 334, 217–
26.
Received: March 25, 2015
2
912–2914.
Published online on July 14, 2015
ChemCatChem 2015, 7, 2460 – 2466
www.chemcatchem.org
2466
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim