Deactivation study of Pd/Al2O3 catalyst for hydrogenation of benzonitrile in fixed-bed reactor

Article abstract of DOI:10.1016/j.apcata.2017.03.030

The Pd/Al₂O₃ catalyst was prepared by the impregnation method to investigate the deactivation and mechanism of coke formation on metal catalysts for the selective hydrogenation of benzonitrile in a continuous fixed-bed reactor under

Full text of DOI:10.1016/j.apcata.2017.03.030

Accepted Manuscript
Title: Deactivation study of Pd/Al O catalyst for
2
3
hydrogenation of benzonitrile in fixed-bed reactor
Author: Chengyong Dai Fanghua Liu Wenxiang Zhang
Yonggang Li Chunli Ning Xueguang Wang Chunlei Zhang
PII:
S0926-860X(17)30129-1
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2017.03.030
APCATA 16189
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
7-12-2016
3-3-2017
19-3-2017
Please cite this article as: C. Dai, F. Liu, W. Zhang, Y. Li, C. Ning, X.
Wang, C. Zhang, Deactivation study of Pd/Al O catalyst for hydrogenation
2
3
of benzonitrile in fixed-bed reactor, Applied Catalysis A, General (2017),
http://dx.doi.org/10.1016/j.apcata.2017.03.030
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

1
2
Deactivation Study of Pd/Al O
Catalyst for
2
3
Hydrogenation of Benzonitrile in Fixed-bed Reactor
3
4
Graphical abstract
5
6
1
Page 1 of 22

7
8
Highlights
9
0



The 1500-h fixed-bed hydrogenation of benzonitrile with a continuous activity decline is performed
1
over Pd/Al O
2 3
for studying the mechanism of coke formation.
1
1
2
The Pd/Al
O
2 3
exhibits an initial conversion of ~100% and a high benzylamine selectivity of ~86% for
1
the selective hydrogenation of benzonitrile.
13
The deposited carbon species contains aromatic imine and tertiary amine groups which are
1
4
5
chemically-bound on Pd particles, leading to the irreversible loss of active sites.
1
2
Page 2 of 22

1
6
7
Deactivation study of Pd/Al O catalyst for hydrogenation of benzonitrile in
2
3
1
fixed-bed reactor
a,b
b
b
b
b
a,
18
19
20
21
22
23
24
25
Chengyong Dai , Fanghua Liu , Wenxiang Zhang , Yonggang Li , Chunli Ning , Xueguang Wang *,
b,
Chunlei Zhang *
a
State Key Laboratory of Advanced Special Steel, School of Materials Science and Engineering, Shanghai
University, Shanghai 200072, China
b
Technology Research Institute of Shanghai Huayi Group, Shanghai 200241, China
Corresponding author: Tel: +86-21-56338244; Fax: +86-21-56338244.
E-mail: wxg228@shu.edu.cn (X. G. Wang); zhangchunlei@shhuayi.com (C. L. Zhang)
3
Page 3 of 22

2
2
2
2
3
6
7
8
9
0
Abstract
The Pd/Al
mechanism of coke formation on metal catalysts for the selective hydrogenation of benzonitrile in a
continuous fixed-bed reactor under mild conditions. The Pd/Al catalyst exhibited the initial benzonitrile
conversion of ~100% with a high benzylamine selectivity of ~86%, and continuous deactivation during the
long-term reaction period of ~1500 h. The comparative investigation of the Pd/Al catalysts before and
after the benzonitrile hydrogenation demonstrated that the deactivation of the Pd/Al was mainly
2
3
O catalyst was prepared by the impregnation method to investigate the deactivation and
2 3
O
31
32
33
34
35
36
37
38
39
40
2 3
O
2 3
O
attributed to carbonaceous compounds deposited on the catalyst. The carbonaceous species contained the
aromatic ring conjugated with C=N double bond and tertiary amine groups, which were chemically bound
onto the surface of Pd particles, leading to the irreversible loss of metallic active sites. Moreover, the
carbon deposition also resulted in partial blockage of the support pores. This study is helpful to understand
the deactivation and formation mechanism of carbonaceous compounds on the surface of metal catalysts
for the hydrogenation reaction.
Keywords：Deactivation, Coke, Hydrogenation, Benzonitrile, Benzylamine
41
4
Page 4 of 22

42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
1. Introduction
Benzylamine (BA) is an important chemical compound that has found widespread applications in
chemical, textile and agrochemical industries [1−4]. The heterogeneous catalytic hydrogenation of
benzonitrile (BN) over supported metal catalysts is considered as an atom-efficient route for the production
of benzylamine. However, these metal catalysts usually exhibit poor selectivity of benzylamine in the batch
reactor, resulting in the low benzylamine yield of less than 80%, due to the formation of by-products
including N-benzylidenebenzylamine (DBI), dibenzylamine (DBA), and toluene (TOL) (Scheme 1) [5−11]
Compared with conventional batch liquid phase operation, the fixed-bed hydrogenation of benzonitrile has
recently attracted much attention because of lower consumptions of energy and catalyst,
lower labor intensity and friendlier environment. Supported metal catalysts like Cu [12], Ni [13], Pd
[10,14], have been investigated for the hydrogenation of benzonitrile in the continuous fixed-bed reactor,
2 3
and among them, the Pd/Al O catalysts exhibited the highest activity and selectivity. However, the
deactivation of metal catalysts was always inevitable, limiting the application of the catalysts for the
long-term hydrogenation.
BN
BI
NH
BA NH2
H2
H2
N
H2
-NH3
H2
BA
NH
BIBA
TOL
NH2
-
NH3
H2
- NH3
H2
NH
-BA
N
H2
5
5
7
8
DBI
DBA
Scheme 1. Probable reaction scheme in the hydrogenation of benzonitrile.
5
9
0
6
Several explanations have been put forward for the deactivation of metal catalysts for the hydrogenation
of nitrile compounds. Verhaak and Arai et al. [15,16] attributed the catalyst deactivation to the formation of
61
62
63
64
2 2
partially-dehydrogenated species and adsorbed amines on the surfaces of Ni/SiO and Pt/SiO catalysts in
the gas-phase hydrogenation of acetonitrile. Huang et al. [17,18] suggested that the metal agglomeration
and formation of the overlayer of carbon species were major factors of catalyst deactivation over Ru/NaY
5
Page 5 of 22

65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
and Pt/NaY in the hydrogenation of acetonitrile and butyronitrile. These indirect assumptions were drawn
mainly through comparing restored activities of deactivated catalysts by different treatment processes [15],
or initial activities in the absence or presence of additional amines [16], or in-situ thermogravimetric data
and the experimental results [17,18]. It is desirable to investigate textural changes of the catalyst itself and
2 3
the chemical nature of the deposits in detail. In the present study, we used Pd/Al O as a model metal
catalyst to systematically characterize the spent catalyst and the coke deposition on the catalyst surface, and
proposed the deactivation mechanism of metal catalysts for the hydrogenation of benzonitrile.
2. Experimental
2.1. Catalyst preparation
3 2
Pseudoboehmite (AlO(OH)) was obtained from CHALCO Shandong Co., Ltd. Pd(NO ) aqueous
solution was purchased from Heraeus Materials Technology Shanghai Ltd. Other chemicals were
analytical-grade, purchased from Sinopharm Chemical Reagent Co., Ltd. and used as received without
purification.
Al
containing 0.5 wt% Pd was prepared through the incipient wetness impregnation. In a typical process, an
aqueous solution of Pd(NO (0.047 mol/L, adjusted to pH = 1.5 with nitric acid) was dropped slowly into
10 g Al (40−60 mesh) with vigorous stirring at room temperature and continuously kept for 12 h. The
2 3 2 3
O support was prepared by calcination of pseudoboehmite at 500 °C for 6 h. The Pd/Al O catalyst
3 2
)
2 3
O
mixture was dried at 100 °C overnight and calcined at 450 °C in air in a muffle furnace for 4 h.
2.2. Catalyst test
Hydrogenation of benzonitrile was performed in a fixed-bed reactor with an internal diameter of 10 mm
and a length of 600 mm. The reaction gases and ethanol solution of benzonitrile were controlled using the
mass flow controllers and a HPLC pump, respectively. The temperature of the catalyst was monitored using
a thermocouple placed in the middle of the catalyst bed. 2.5 g catalyst was loaded in the center of the
reactor between two layers of quartz sands. Prior to the reaction, the catalyst was first reduced in situ at 300
o
o
C under an 80 mL/min H
2
flow for 4 h, and then cooled to 70 C for reaction. After this, the ethanol
solution of benzonitrile containing 20 wt% benzonitrile was introduced into the reaction system, and
meanwhile, the reaction pressure was adjusted to 1.5 MPa with H
6
2
. During the reaction, the weight hourly
Page 6 of 22

9
9
9
9
9
9
4
5
6
7
8
9
2
space velocity (WHSV) of benzonitrile was set at 0.5 g/gcat h and the molar ratio of H to benzonitrile was
kept at 20. The ethanol was confirmed to be inert as a solvent under the reaction conditions. The product
mixture was collected in a cold trap and analyzed by using a gas chromatograph equipped with a FID
detector with a DB-5 capillary column. After the reaction, the catalyst was washed in situ by replacing the
o
ethanol solution of BN with pure ethanol for 4 h, and then was dried in a N
2
flow at 70 C for 24 h. The
spent catalyst was separated carefully and gathered in nitrogen. The overall mass balance was more than
95% on the basis of carbon in the starting reactants.
1
1
1
00
01
02
Benzonitrile conversion (X) and product selectivity (S
i
) were calculated with the equations (Eqs. 1, 2),
shown as follows:
[
BN]  [BN]
in
out
1
03
04
（1）
（2）
X(%) 
Si (%) 
100
[
BN]in
[
Product]  Ni
i
1
100
[
BN] [BN]
in
out
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
Where [BN]in and [BN]out are the molar amounts of benzonitrile in the feed and in the effluent mixture,
respectively. [Product] is the molar amounts of product i (BA, DBA, TOL) in the effluent mixture and N is
i
i
the stoichiometric coefficient for product i.
2.3. Catalyst characterization
N
2
adsorption and desorption isotherms were measured using a Quantachrome Autosorb-iQ analyzer at
o
liquid nitrogen temperature (−196 C). The sample was outgassed at 250 °C under vacuum (0.2 Torr) for 16
h prior to analysis. The specific surface areas were evaluated using the Brunauer-Emmett-Teller (BET)
method in the P/P
desorption branch of the isotherms and the Barrett-Joyner-Halenda (BJH) method and pore sizes were
obtained by the BJH method. Pore volumes were calculated from the adsorbed amount at P/P = 0.990.
Pd contents in the Pd/Al catalysts were analyzed on a PerkinElmer Optima 7000 DV spectrometer
0
range from 0.05 to 0.25. Pore size distribution curves were calculated using the
0
2 3
O
by inductively coupled plasma optical emission spectroscopy (ICP-OES). Mass fractions of carbon,
hydrogen and nitrogen were determined using a vario MICRO cube elemental analyzer.
The X-ray diffraction (XRD) patterns were obtained using a Bruker D8 ADVANCE X-ray
diffractometer with Cu Kα radiation (λ = 0.15418 nm) at a voltage of 40 kV and a current of 40 mA.
7
Page 7 of 22

121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
Thermogravimetry and differential scanning calorimeter measurements (TG-DSC) measurements were
performed on a NETZSCH STA449F3 synchronous thermal analysis system. About 30 mg sample was first
treated at 50 °C for 30 min, and then heated to 1000 °C in a 50 mL/min flow of air or N
10 °C/min.
Fourier transform infrared (FT-IR) was carried on a Nicolet NEXUS 670 spectrometer with a resolution
2
at a heating rate of
−1
of 2 cm using the KBr method.
The morphology and microstructure of the samples were characterized on a Hitachi S4800 field emission
scanning electron microscope (SEM) with an energy dispersive X-ray spectrometer (EDS). TEM
micrographs were made with a JEOL JEM-2010F field emission microscope operating at 200 kV. The
sample was prepared by placing a drop of ethanol solution of a well-ground catalyst powder onto a
carbon-coated copper grid followed by evaporation of ethanol. Size distributions of the Pd particles were
determined by counting ca. 100–150 particles in TEM images.
Dispersion of metal Pd was determined by CO chemisorption on a Micromeritics AutoChem II 2920
2
using a thermal conductivity detector (TCD). 50 mg sample was reduced in a 10 vol% H /He flow (30
mL/min) at 300 °C for 1 h, and then cooled to 50 °C and kept in a He flow (30 mL/min) for 30 min. After
this, the sample was treated with a 10 vol% CO/He flow (30 mL/min) for 30 min, and then purged in
flowing He (30 mL/min) for 60 min. The sample was heated to 600 °C at a rate of 10 °C/min in the He flow.
The increase of the TCD signal area indicated the CO uptake chemisorbed. The metal dispersion was
calculated from the CO chemisorption data, where the stoichiometry of the CO adsorption on the metal
atoms was assumed to be 1.
3. Results and discussion
3.1. Stability test for the Pd/Al
2 3
O catalyst
2 3
For investigating the stability of the Pd/Al O catalyst, a long-term test was conducted for the BN
hydrogenation. Fig. 1 presents the BN conversion and the products selectivities as a function of reaction
time. It could be seen that the BN hydrogenation could be roughly divided into two stages: 0−200 h and
200−1500 h.
During the first stage, although the catalyst showed initial activity with almost 100% BN conversion and
86% BA selectivity, the selectivity to DBA increased from ~7% to 22% and the selectivity to toluene (TOL)
8
Page 8 of 22

150
151
152
153
154
155
156
157
158
159
160
decreased from ~16% to 2%. As shown in Scheme 1, DBA is formed through the condensation of
partially-hydrogenated intermediate (BI) of BN and BA, while the further hydrogenolysis and deamination
of DBA and BA generate TOL, which is regarded as the ultimate product of BN hydrogenation. Thus, the
dramatic increase in DBA selectivity and the declines in TOL and BA selectivity indicated that the
2 3
hydrogenation ability of the Pd/Al O was weakened, although the BN conversion showed little change
during the initial 200 h.
During the second stage, the BN conversion decreased from ~97% to 21% almost linearly with the
reaction time on stream at a rate of ~5.9% per 100 h. The BA and TOL selectivities decreased from ~75%
to 6%, and from ~2.0% to 0.2%, respectively, and meanwhile, the DBA selectivity increased from ~22% to
34%. These results indicated that the Pd/Al
2
O
3
had a significant deactivation after the continuous 1500-h
hydrogenation of BN in the fixed-bed reactor.
100
100
BN
80
60
40
20
0
BA
80
60
40
20
0
DBA
TOL
0
200 400 600 800 1000 1200 1400
Time on stream (h)
1
61
62
1
Fig. 1. Deactivation behavior of the Pd/Al
O
2 3
catalyst with reaction time on stream.
163
164
165
166
167
168
169
170
171
172
3.2. Compositions and textural properties of the catalysts
o
The bulk composition and textural properties of the freshly-prepared, reduced (at 300 C under an 80
mL/min H
for 4 h) are summarized in Table 1. The Pd content of the fresh catalyst was 0.45 wt%, close to the
theoretical value of 0.5 wt%. After the reaction, the Pd content in the spent Pd/Al significantly
decreased to 0.34 wt%. However, when the spent Pd/Al catalyst was re-calcined in air at 450 °C for 4 h,
2 2 3
flow for 4 h), spent and re-calcined Pd/Al O catalysts (spent catalyst calcined in air at 450 °C
2
O
3
2 3
O
the Pd content was restored to 0.46% again. These results demonstrated that there was organic carbon
species deposited on the catalyst, which could be fully removed by heat treatment and no loss of metal Pd
during the long-term BN hydrogenation.
9
Page 9 of 22

1
73
74
Table 1
1
Compositions and textural properties of the Pd/Al
2
O
3
catalysts.
2
3
Pd/Al
O
2 3
Pd content (wt%)
S
BET (m /g)
Pore volume (cm /g)
Pore size (nm)
Fresh
0.45
—
216
0.65
0.53
0.43
0.67
8.0
6.9
Reduced
Spent
190
0.34
0.46
93
9.5
Re-calcined
206
10.2
175
176
177
178
179
180
181
182
183
184
185
186
2
Fig. 2 illustrates the N sorption curves and pore size distributions of fresh, reduced, spent and
re-calcined Pd/Al
which were characteristic of mesoporous materials [19]. The hysteresis loop of the spent Pd/Al
at a high relative pressure, indicating an increased mesopore size. Table 1 lists the textural properties of the
Pd/Al catalysts. Compared with the fresh catalyst, the reduced Pd/Al showed slight declines in
surface area, pore volume and pore size, due to reduction of PdO to metal Pd on the surface [20,21]. After
the reaction, the surface area and pore volume of spent Pd/Al drastically decreased, but the pore size
showed an obvious increase. This result demonstrated that the pores of spent Pd/Al were partially
clogged with deposited carbon species during the long-term test. After the spent Pd/Al was re-calcined,
the surface area and pore volume were restored due to the removal of carbonaceous deposits. The increased
pore size of the re-calcined Pd/Al was likely due to the swelling effect generated by carbonization [22].
2
O
3
catalysts. All the samples exhibited type IV isotherms with apparent hysteresis loops,
2 3
O
started
O
2 3
2 3
O
x
2 3
O
2 3
O
2 3
O
2 3
O
a
Re-calcined
Spent
Reduced
Fresh
0.2
0.0
0.4
0.6
0.8
1.0
Relative pressure (P/P0)
187
1
0
Page 10 of 22

b
Re-calcined
Spent
Reduced
Fresh
25
5
10
15
20
30
Pore size (nm)
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
2 2 3
Fig. 2. (a) N adsorption-desorption isotherms, and (b) BJH pore size distributions of the Pd/Al O
catalysts.
3.3. Dispersion of palladium and TEM analysis
Table 2 summaries the Pd dispersions of reduced, spent and re-calcined Pd/Al
2
O
3
catalysts. It could be
to 9.4%,
3
O , the Pd dispersion was recovered to 17.8%. This result indicated that
seen that Pd dispersion of the spent Pd/Al
however, for the re-calcined Pd/Al
2 3 2 3
O decreased from 22.7% for the reduced Pd/Al O
2
the number of Pd active sites dramatically reduced after the long-term reaction likely due to growth of
metallic particles or coverage of carbon species on the surface of Pd particles. The TEM images in Fig. 3
2 3 2 3
revealed that compared with those of the reduced Pd/Al O , the Pd particle sizes of spent Pd/Al O showed
no obvious change in the range of 5−9 nm. It was concluded that the deposition of carbon species on the
catalyst surface was responsible for the reduction of active sites. This point was also confirmed by the
2 3
regain of most Pd dispersion on the re-calcined Pd/Al O catalyst in Table 2.
Table 2
Dispersion of metallic Pd on the Pd/Al
Pd/Al
2 3
O catalysts.
2
O
3
Pd dispersion (%)
Reduced
Spent
22.7
9.4
Re-calcined
17.8
11
Page 11 of 22

205
a)
b)
2
06
ꢀ
207
208
209
2 3
Fig. 3. TEM images of the Pd/Al O catalysts. (a) Reduced, and (b) Spent
210
2
11
12
3.4. XRD analysis
The XRD patterns of the fresh and spent Pd/Al
2
2 3
O catalysts were illustrated in Fig. 4. No diffraction
213
214
215
216
217
218
219
220
peaks associated with Pd were observed in both cases due to the low loading and high dispersion of metal
Pd. In the case of the fresh Pd/Al , the XRD patterns indicated that the Al support consisted of two
mixed crystalline phases of γ-Al (JCPDS no. 80-0956) and χ-Al (JCPDS no. 04-0880). It is known
that the XRD patterns of γ-Al and χ-Al are indistinguishable for most of the diffraction peaks due to
the peak broadening. However, for the spent Pd/Al , it was found that the characteristic diffraction peak
2
O
3
2 3
O
2
O
3
2 3
O
2
O
3
2 3
O
2 3
O
simply due to χ-Al
O
2 3
phase at 2θ = 43.5° disappeared, and meanwhile, the relative peak intensity at 2θ =
o
67.2° decreased and the one at 2θ = 46.1 increased clearly. These results implied that χ-Al
2
O
3
phase was
2 3
transformed to γ-Al O during the long-term BN hydrogenation [23].
1
2
Page 12 of 22


-Al O3
2
+
-Al O3



2






Spent



Fresh

10
20
30
40
2/
50
60
70
80
o
2
21
22
2
Fig. 4. XRD patterns of the fresh and spent Pd/Al
O
2 3
catalysts.
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
3.5. TG analysis
Fig. 5 illustrates the TG profiles of the spent Pd/Al
respectively. The curve in N (curve a) could be divided into three regions. Except a mass loss of ~2.5 wt%
2 3 2
O catalyst obtained in a flow of air and N ,
2
below 130 °C due to the desorption of water, the low-temperature region (< 280 °C) had a large weight loss
of ca. 14.5 wt% in the range of 130−280°C, which was ascribed to the decomposition or vaporization of
small organic molecules. The second region in the range of 280−600 °C corresponded to a mass loss of ~7
wt%, indicating a possible formation of carbon deposition and/or degradation of the oligomers. The
high-temperature region above 600 °C produced a slight mass loss of 1.5 wt% with a noticeable
endothermic peak at ca. 812 °C (not shown), which resulted from the possible transformation of the
crystalline structure.
The TG curve of the spent Pd/Al
O
2 3
in air (curve b) showed the similar behavior to in N
2
. In the
low-temperature region, although the sample has an approximate mass loss to that in N
2
, the degradation
rate in air was much faster. This result indicated the oxidative atmosphere had a significant promotion for
decomposition or desorption of small organic molecules. This case could also be found in the
medium-temperature region with a mass loss of ~11 wt%, which exhibited approximately 1.9 times faster
2
than the degradation rate of the oligomers in N . However, the curves in the high-temperature region
presented only a mass loss of 1 wt%, indicating the complete decomposition of the residue in the sample.
1
3
Page 13 of 22

241
242
243
The difference of mass losses between curve a and curve b demonstrated that the deposited organic carbon
compounds could be pyrolyzed in N , but completely be burned off in air, in combination with the Pd
in Table 1.
2
content in the spent Pd/Al
O
2 3
100
3%
2
.5%
95
90
85
80
75
70
14%
1
4.5%
a 5%
2%
11%
1.5%
b
1%
0
150 300 450 6o00 750 900
Temperature (C)
2
44
45
2
2 3 2
Fig. 5. TG profiles for the spent Pd/Al O catalyst under different atmospheres. (a) N , and (b) air.
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
3.6. Elemental analysis
Since the TG and TEM analyses indicated the presence of carbonaceous species on the spent Pd/Al
2 3
O ,
the elemental (C, H, N) analyses of the fresh and spent Pd/Al were performed to understand the nature
2 3
O
and composition of deposited carbon species. In order to eliminate the influence of light compounds
remained on the spent catalyst, the samples were evacuated at 150 °C for 10 h to remove the materials with
boiling points up to 280 °C [24]. The analytic results are shown in Table 3. In the fresh catalyst, the
contents of C, H and N were negligible. On the contrary, the total content of C, H, and N in the spent
catalyst was as high as ca. 20.44%, which was approximate to the TG results. The degassing treatment
resulted in a reduction in the C content of the spent catalyst from 17.43% to 14.78%. This result
demonstrated that most of the residues were heavy carbon components or strongly adsorbed intermediates.
The total mass loss of C, H and N due to the evacuation is about 3.15 wt%, much less than that in the
low-temperature region of the TG profiles in Fig. 5. This result indicated that the coating compounds were
strongly-chemisorbed on the spent catalyst. The C/H molar ratio in the heavy carbonaceous species on the
spent catalyst after the vacuum treatment was ~1.34, nearly twice as that in the light carbonaceous
compounds removed by vacuum treatment, revealing the lower polymerization of organic compounds. On
1
4
Page 14 of 22

262
263
264
265
266
267
268
269
270
271
the other hand, the fact that the C/H molar ratio was not so high implied that the condensation degree of
coking species was relatively low over the spent Pd/Al under the reaction conditions. Therefore, these
2 3
O
deposited carbonaceous species could be easily removed by calcination below 450 °C in air, as revealed in
Table 1 and the TG profiles of Fig. 5. By comparison with the empirical formulas of the residues based on
the elemental analysis data in Table 3 and the related compounds in Scheme 1, it could be proposed that the
carbonaceous compounds on the catalyst surface should be some oligomers, whose properties were of
intermediates between dibenzylamine and tribenzylamine.
Table 3
Elemental and proximate analysis data for the Pd/Al
O
2 3
catalysts
C (wt%)
17.43
Pd/Al
O
2 3
Vacuum degassing
H (wt%)
1.82
N (wt%)
1.19
Before
After
—
Spent
Fresh
14.78
1.51
1.00
0.54
0.62
0.07
272
Deposited
Relative fraction of
compounds (wt%)
16.4
Contents in spent catalyst (wt%)
C/H
Empirical
formula
compounds
C
H
N
(mol/mol)
0.71
a
Light
2.65
0.31
0.19
C
C
16
H
23
N
N
b
Heavy
83.6
14.24
0.89
0.93
1.34
H
18 13
a
Removed by vacuum treatment on the basis of the data of spent catalysts before and after vacuum
treatment.
b
Survived after the vacuum treatment on the basis of the data of spent catalyst after vacuum treatment
and fresh catalyst.
273
274
275
276
277
278
3.7. FT-IR spectra
The FT-IR spectroscopy is a powerful tool to investigate the chemical properties of the support and
carbonaceous species deposited on the catalyst surface. Fig. 6 illustrates the FT-IR spectra of the Pd/Al
catalysts before and after the long-term BN hydrogenation. The two samples both exhibited the main bands
observed commonly in Al support. Slight differences pointed to the carbonaceous compounds deposited
on the spent Pd/Al . The FT-IR spectra could be divided into three parts of interest, assigned to the
2 3
O
2 3
O
2 3
O
1
5
Page 15 of 22

279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
aromatic, methylene, and amine groups, respectively. Part I contained the features due to the symmetric and
−1
antisymmetric vibrational stretching mode of aromatic functional groups (three peaks of 3100−3000 cm
−1
−1
−1
−1
for aromatic C-H stretch, and four peaks at 1603 cm , 1570 cm , 1496 cm and 1452 cm for aromatic
−
1
C=C stretch). Moreover, the unique absorption band at 1570 cm indicated the existence of a hyper
2
conjugation effect with the benzene rings, showing the deposition of carbon compounds of sp
hybridization [25].
−1
−1
−1
Part II involved spectral absorption bands at 2925 cm , 2840 cm , and 2795 cm attributed to the C-H
vibrations of the methylene groups. However, the intensities of these bands were relatively weaker,
indicating less content of hydrocarbon parts in the residue. In the third part, no extra absorption bands were
−1
observed at 3300−3500 cm , where the number of the peaks could be used to identify the N-H binding
states, i.e., the primary amine corresponds to bimodal structure, secondary amine has a single peak, and
−1
tertiary amine shows no absorption [25]. In combination with the absorption band at 1297 cm caused by
the stretching vibration of (aromatic) C-N bond, this result indicated the possible existence of tertiary
−1
amine. Meanwhile, a unique absorption band at 1643 cm corresponding to the stretching vibration of
(conjugated) C=N band also revealed the possible existence of aromatic imines [26,27], such as
N-benzylidenebenzylamine (DBI), which was another kind of possible structure. No characteristic peaks at
−1
2220−2240 cm associated with C≡N functional group were observed, indicating the absence of cyanide
parts [25,26]. These results demonstrated that the carbonaceous compounds on the spent Pd/Al
BN hydrogenation contained tertiary amine and/or aromatic imine groups.
2 3
O for the
100
80
60
40
20
b
a
Fresh
Spent
0
4
000
3500
3000
2500
2000
1500
1000
-1
Wavenumber (cm )
298
1
6
Page 16 of 22

9
8
8
7
7
6
6
5
5
0
5
0
5
0
5
0
5
2795
Fresh
Spent
a
2840
2925
3
062
3
029
3
085
0
3200
3100
3000
Wavenumber (cm )
2900
2800
2700
-
1
299
9
8
8
7
7
6
6
0
b
Fresh
Spent
1
452
5
0
5
0
5
0
1
496
1
603
1297
1643
1570
1
700 1600 1500 1400 1300 1200 1100
-1
Wavenumber (cm )
Fig. 6. FT-IR spectra of the fresh and spent Pd/Al
3
00
01
3
2 3
O catalysts.
302
303
304
305
306
307
308
309
310
311
312
313
3.8. SEM analysis
SEM analysis was performed to evaluate the morphologies of the Pd/Al
long-term BN hydrogenation and the SEM images are shown in Fig. 7. The fresh Pd/Al
amorphous morphology with porous structures, similar to that of Al support (not shown). However,
polymer-like carbon and carbon filaments were both found on the spent catalyst [28]. Elemental
distribution maps of the fresh and spent Pd/Al catalysts in Fig. 8 demonstrated that Pd was clearly
O
2 3
catalysts before and after the
showed an
2 3
O
2 3
O
2 3
O
observed and homogeneously dispersed on the fresh catalyst. For the spent catalyst, except the conspicuous
O and Al signals, the carbon signal was clearly observed. However, the Pd signals were hardly detected on
the image. These results indicated that the Pd particles were most likely covered by carbonaceous species.
It could be speculated that the carbon species were selectively deposited on the surfaces of Pd particles
2 3
during the BN hydrogenation and there was no immigration of the carbon species to the Al O support. This
1
7
Page 17 of 22

3
14
15
result was consistent with the kinetic studies on Pd/Al
2 3
O catalysts, which revealed the diverse processes of
3
coke formation on the active sites and the support [29].
316
317
318
3
19
20
3
3
21
22
Fig. 7. SEM images of the Pd/Al
2
O
3
catalysts. (a) Fresh, and (b, c) Spent.
3
1
8
Page 18 of 22

1-(a)
1-(b)
2
um
O-Ka1
1-(d)
1-(c)
Al-Ka1
Pd-La1
2-(b)
323
2
-(a)
2
um
O-Ka1
2-(d)
2-(c)
Al-Ka1
C-Ka1
3
24
25
3
2 3
Fig. 8. SEM images and corresponding elemental analysis maps for the Pd/Al O catalysts. (1) Fresh, and
326
327
328
329
330
(2) Spent.
3.9. Proposed possible routes for carbonaceous species
According to the accepted mechanistic pathways [8,30,31], the catalyzed nitrile hydrogenation can
proceed via two different routes shown in Scheme 2, namely, the partially-hydrogenated intermediates
1
9
Page 19 of 22

331
332
333
334
335
336
(nitrene, carbene and imine) are bound to the metal surface through either a nitrogen atom (M-N route) or
an α-carbon atom (M-C route). In the case of the hydrogenation of benzonitrile containing π-system of the
aromatic ring, the nitrile hydrogenation proceeds via an unique multi-coordinated side-on adsorbed
intermediate (Scheme 2), i.e. σ/π-adsorbed imine preferentially interacts with the Pd surface through its
delocalized aromatic π-electrons [11,32]. Then, these conjugated intermediates generate higher imines,
2
leading to sp type residues (Scheme 1).
337
338
339
340
Scheme 2. Possible adsorption complexes in the first step of benzonitrile hydrogenation
4. Conclusions
341
342
343
344
345
346
347
348
349
350
351
352
The long-term stability test of Pd/Al
performed in a fixed-bed reactor. The catalytic activity exhibited continuous degradation as a function of
reaction time. The comparative investigation of the Pd/Al catalyst before and after the reaction
demonstrated that the deactivation of the Pd/Al was mainly attributed to carbonaceous species deposited
2 3
O catalyst for the selective hydrogenation of benzonitrile has been
2 3
O
2 3
O
on the surface of Pd particles. The carbonaceous species characterized by the conjugated functional groups
contained tertiary amine and/or aromatic imine groups, which were chemically-bound onto the surface of
2
Pd particles, leading to the irreversible loss of metallic active sites. These unique sp type residues are
proposed to generate via a multi-coordinated side-on adsorbed σ/π-adsorbed imine.
References
[] A. Ricci, Modern amination methods, New York, Wiley, 2000.
[2] S.A. Lawrence, Amines: Synthesis, Properties and Application, Cambridge University Press,
2
0
Page 20 of 22

353
354
355
356
357
Cambridge, 2004.
[3] H.A. Wittcoff, B.G. Reuben, J.S. Plotkin, Industrial Organic Chemicals, John Wiley & Sons, Inc., 2005.
[4] K. Weisserme, H.-J. Arpe, Industrial Organic Chemistry, Wiley-Vch Verlag & Co. KGaA, Wenham,
2003.
[5] D.B. Bagal, B.M. Bhanage, Adv. Synth. Catal. 357 (2015) 883−900.
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
[6] L. Hegedűs, T. Máthé, Appl. Catal. A 296 (2005) 209−215.
[7] H. Yoshida, Y. Wang, S. Narisawa, S. i. Fujita, R. Liu, M. Arai, Appl. Catal. A 456 (2013) 215−222.
[8] A. Chojecki, M. Veprekheijman, T. Muller, P. Scharringer, S. Veprek, J. Lercher, J. Catal. 245 (2007)
237−248.
[9] C. de Bellefon, P. Fouilloux, Cat. Rev. - Sci. Eng. 36 (2006) 459−506.
[0] M. Chatterjee, H. Kawanami, M. Sato, T. Ishizaka, T. Yokoyama, T. Suzuki, Green Chem. 12 (2010)
87−93.
[1] J.J.W. Bakker, A.G.v.d. Neut, M.T. Kreutzer, J.A. Moulijn, F. Kapteijn, J. Catal. 274 (2010) 176−191.
[12] R.K. Marella, K.S. Koppadi, Y. Jyothi, K.S. Rama Rao, D.R. Burri, New J. Chem. 37 (2013)
3229−3235.
[3] C.V. Rode, M. Arai, M. Shirai, Y. Nishiyama, Appl. Catal. A 148 (1997) 405−413.
[4] Y. Hao, M. Li, F. Cárdenas-Lizana, M.A. Keane, Catal. Lett. 146 (2016) 109−116.
[5] M.J.F.M. Verhaak, A.J. Vandillen, J.W. Geus, J. Catal. 143 (1993) 187−200.
[6] M. Arai, Y. Takada, Y. Nishiyama, J. Phys. Chem. B 102 (1998) 1968−1973.
[7] Y. Huang, V. Adeeva, W.M.H. Sachtler, Appl. Catal. A 196 (2000) 73−85.
[8] W.M.H. Sachtler, Y. Huang, Appl. Catal. A. 191 (2000) 35−44.
[19] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T. Siemieniewska,
Pure Appl. Chem. 57 (1985) 603−619.
[20] G.L. Liu, T. Niu, A. Cao, Y.X. Geng, Y. Zhang, Y. Liu, Fuel 176 (2016) 1−10.
[21] Y. Yang, X. Qi, X. Wang, D. Lv, F. Yu, L. Zhong, H. Wang, Y. Sun, Catal. Today 270 (2016)
101−107.
[22] S. Nomura, K. Thomas, Fuel. 75 (1996) 187−194.
[23] J. Khom-in, P. Praserthdam, J. Panpranot, O. Mekasuwandumrong, Catal. Comm. 9 (2008)
1955−1958.
2
1
Page 21 of 22

382
383
384
385
386
[24] J.M. Goodman, P.D. Kirby, L.O. Haustedt, Tetrahedron Lett. 41 (2000) 9879−9882.
[25] E. Pretsch, B. Philippe, M. Badertscher, Structure Determination of Organic Compounds: Tables of
Spectral Data (4th edition), Springer, 2009.
[26] J.L. de la Fuente, M. Ruiz-Bermejo, D. Nna-Mvondo, R.D. Minard, Polym. Degrad. Stab. 110 (2014)
241−251.
3
3
87
88
[27] H. Namli, O. Turhan, Spectrochim Acta A Mol Biomol Spectrosc. 64 (2006) 93−100.
389
390
391
392
393
394
395
396
397
398
399
400
401
402
[28] P. Albers, J. Pietsch, S.F. Parker, J. Mol. Catal. A 173 (2001) 275−286.
[29] E. Klemm, B. Amon, H. Redlingshöfer, G. Emig, Stud. Surf. Sci. Catal. 126 (1999) 97−104.
[30] P. Scharringer, T. Muller, J. Lercher, J. Catal. 253 (2008) 167−179.
[31] P. Schärringer, Kinetic and mechanistic studies on the liquid-phase hydrogenation of nitriles and
dinitriles over cobalt-based catalysts, Doctor Thesis, University of Munich, 2007.
[32] J. Krupka, J. Pasek, Curr. Org. Chem. 16 (2012) 988−1004.
403
2
2
Page 22 of 22