Study of Catalyst Deactivation in Liquid-Phase Hydrogenation of 3-Nitrostyrene Over Au/Al2O3 Catalyst in Flow Reactor

Article abstract of DOI:10.1007/s10562-016-1959-3

Abstract: For the first time, we report the results of the study of dynamics of the Au/Al₂O₃catalyst deactivation during the selective liquid-phase 3-nitrostyrene (3-NS) hydrogenation in a fixed bed flow reactor. It is shown that the

Full text of DOI:10.1007/s10562-016-1959-3

Catal Lett (2017) 147:572–580
DOI 10.1007/s10562-016-1959-3
Study of Catalyst Deactivation in Liquid-Phase Hydrogenation
of 3-Nitrostyrene Over Au/Al O Catalyst in Flow Reactor
2
3
1
1
1
1,2
1,2
A. L. Nuzhdin · S. I. Reshetnikov · G. A. Bukhtiyarova · B. L. Moroz
·
1
,2
1
E. Yu. Gerasimov · P. A. Pyrjaev · V. I. Bukhtiyarov
Received: 26 October 2016 / Accepted: 20 December 2016 / Published online: 12 January 2017
Springer Science+Business Media New York 2017
©
Abstract For the first time, we report the results of the
Graphical Abstract
study of dynamics of the Au/Al O catalyst deactivation
2
3
during the selective liquid-phase 3-nitrostyrene (3-NS)
hydrogenation in a fixed bed flow reactor. It is shown that
the formation of carbonaceous deposits is the main reason
of catalyst deactivation, while the leaching of gold from the
*
*
H2
H2
+
+
+
*
*
NO2
NO2
NO
NH2
NH2
Au/Al O catalyst or enlargement of the gold particles has
2
3
*
not been detected. The content of the carbon deposits on
the spent catalyst at the inlet of the reactor is higher than at
the outlet and increases as the initial concentration of 3-NS
grows. The character of carbonaceous deposits allocation
and analysis of reaction products point out to 3-vinylni-
trosobenzene as the main source of the Au/Al O catalyst
NO
Carbon deposits
surface active site of Au/Al O catalyst
* -
2
3
2
3
deactivation coinciding with the speculation arising from
the kinetic modeling.
Keywords Heterogeneous catalysis · Deactivation ·
Liquid-phase hydrogenation · Selective hydrogenation ·
Kinetic modeling · Au/Al O catalyst · 3-Nitrostyrene ·
2
3
Flow reactor
1
Introduction
Functionalized anilines are commercially important inter-
mediates in the manufacture of agrochemicals, pharma-
ceuticals, and other chemical products [1–3]. Conventional
techniques for their production are based on the reduction
of a nitro group in a corresponding substrate with stoichi-
ometric amounts of reducing agents, such as iron in acid
media (Bechamp process), H S, or NaHS. These processes
2
*
A. L. Nuzhdin
produce a large amount of harmful wastes and display low
selectivity for the reduction of a nitro group in the pres-
ence of other reducible groups such as C=O or C=C bonds,
both reasons restrict their use in practice [3]. In accord-
ance with a current trend to use atom economy methods in
organic synthesis, the selective reduction of nitro groups
anuzhdin@catalysis.ru
1
2
Boreskov Institute of Catalysis SB RAS, Lavrentieva Ave., 5,
Novosibirsk, Russia 630090
Novosibirsk State University, Pirogova Str., 2, Novosibirsk,
Russia 630090
Vol:.(1234567890)
1
3

Study of Catalyst Deactivation in Liquid-Phase Hydrogenation of 3-Nitrostyrene Over…
573
with hydrogen over heterogeneous catalysts has been high-
lighted as a promising route for the development of a new,
environmentally friendly technology for the production
of functionalized anilines [3–5]. The preferential reduc-
tion of a nitro group in the presence of C=C bond is not so
straightforward because of high reactivity of this function-
ality toward metal-catalyzed hydrogenation [3]. Catalytic
metals (e.g., Pt, Ru, and Ni), which are most widely used
for the hydrogenation of organic molecules, hardly provide
the selective reduction of nitro groups when carbon–car-
bon double bond is present in the same molecule, although
adjustment of the solvent, catalyst support, and amount of
noble metal, as well as the addition of modifiers can con-
tribute to success in certain cases [3, 6–8]. In the last dec-
ade, heterogeneous catalysts comprising monometallic Pt
3-, and 4-CNBs; 3-nitrostyrene (3-NS) and 4-nitroaceto-
phenone over the Au/γ-Al O catalyst to the corresponding
2
3
anilines at 60–110°C and 1.0 MPa in a three-phase packed-
bed reactor operating in the up-flow mode. One important
result found was that hydrogenation of 3-NS led to the
formation of 3-vinylaniline (3-VA) with selectivity of up
to 93% at substrate conversion exceeding 98%. However,
a smooth deactivation of the catalyst with time on-stream
was observed mainly due to the appearance of carbona-
ceous deposits.
Therefore, the objective of the present work is to study
the dynamics of the Au/γ-Al O catalyst deactivation dur-
2
3
ing 3-NS hydrogenation in a liquid-phase packed bed flow
reactor as a function of the inlet 3-NS concentration. The
allocation of carbon deposition along the reactor length
depending on the initial 3-NS concentration has been also
examined.
[
3, 6, 7], Ru [7], Ni [7], Au [6, 7, 9–13], Ag [14, 15], Mo
16], or bimetallic Au–Pt [17] and Pt–Zn [18] nanoparticles
[
were proposed for selective hydrogenation of 3- or 4-nitros-
tyrene, which are used most often as model compounds
containing both nitro-group and C=C bond. It has been
shown that supported gold nanoparticles demonstrate one
of the highest selectivity toward the nitro group hydrogena-
tion and produce the corresponding aromatic amine with
high yield [5–7, 9–13].
2 Experimental
2.1 Chemicals
Hydrogen tetrachloroaurate H[AuCl ]⋅aq (49.47 wt% Au)
4
It has been observed that the supported gold catalysts
lose their activity during liquid-phase or gas-phase hydro-
genation of aromatic nitro-compounds [9, 12, 19–24], but
the publications considering the reasons for this are still
scarce. Aggregation of the gold nanoparticles was assumed
as a main reason for a decrease in the 4-chloronitrobenzene
from Aurat (Russia) was used as the source of gold to pre-
pare the gold catalyst. 3-Nitrostyrene (Acros Organics, 97%
pure) was used as the reagent for catalyst testing. Tolu-
ene of 99.5% purity from “ECOS” (Russia) and n-decane
(>99% pure) from Acros Organics were employed for cata-
lytic experiments as solvent and internal gas chromatogra-
phy (GC) standard, respectively.
(
4-CNB) conversion over the Au/ZrO catalyst in succes-
2
sive batch-mode experiments using ethanol as a solvent
[
20]. A decline in the activity of the Au/Al O and Au/
2.2 Catalyst Preparation
2
3
TiO catalysts for nitrobenzene hydrogenation detected
2
after their recycling was also explained by sintering of gold
nanoparticles [21]. On the contrary, no enlargement of the
gold particles was observed in the continuous gas-phase
hydrogenation of 4-CNB to 4-chloroaniline over the Au/
Al O catalyst [19]. The authors established that blocking
As the catalyst support, we used γ-Al O (BET specific
2
3
2
−1
3
surface area, 215 m g , total pore volume, 0.61 cm
−
1
g , mean pore diameter, 9.6 nm) prepared by extrusion
of a paste consisting of aluminum oxide of Puralox TH
100/150 grade and aluminum hydroxide of Pural SB grade
(both produced by Sasolchemie GmbH, Germany) mixed
in a ratio of 30:70 by weight and added with an aqueous
solution of nitric acid as a peptizing agent. The cylindri-
cal Al O granules of 1.5 mm in diameter thus obtained
2
3
of the active sites via deposition of carbon was a cause of
temporal deactivation of the Au/Al O catalyst.
2
3
Generally, hydrogenation of aromatic nitro compounds
is performed in either liquid-phase batch reactors or gas-
phase flow reactors. Meanwhile, the use of three-phase flow
reactor allows an improvement of gas-liquid-solid inter-
action and reduction of the side products through a better
control over process variables [25–28]. Besides, carrying
out the reactions in fixed bed flow reactor makes it possible
to monitor the changes in catalytic activity and selectivity
with time-on-stream and to reveal the dynamics of catalyst
deactivation.
2
3
were dried at 110°C overnight and calcined in a flow of
air at 550°C for 4 h. Then they were crushed and sieved,
the grains of 250–500 µm in diameter were used for the
catalyst preparation. The support was dried at 110°C in
a drying oven for 8 h immediately before gold deposi-
tion. The Au/Al O catalysts were prepared through the
2
3
adsorption of anionic gold(III) hydroxychloride com-
plexes onto the alumina support from an aqueous solution
Recently, we observed [12] a highly selective hydro-
genation of several substituted nitrobenzenes, such as 2-,
of H[AuCl ]+NaOH at 70°C and pH=7 (the so-called
4
“deposition–precipitation technique”) followed by washing
1
3

5
74
A. L. Nuzhdin et al.
with distilled water, drying and calcining in air at 400°C,
as described earlier [29]. Afterwards, the final sample was
calcined at 600 or 800°C. As a result of this procedure, a
set of the Au/Al O catalysts containing 1.8–1.9 wt% of
of automated back pressure regulator. The fixed bed car-
tridge-like reactors of CatCart®30 with an inner diameter
3
of 4.0 mm and length of 30 mm (volume 0.38 cm ) were
used in the experiments. A more detailed description of the
H-Cube Pro™ setup has been published elsewhere [25, 30].
Before the experiment, the cartridge containing the cata-
lyst (0.185 g of 250–500 μm particles) was installed in a
heater unit, and the mixture of pure solvent with hydrogen
was fed to the reactor until the required reaction parame-
ters (temperature, pressure, flow rates of hydrogen and liq-
uid feed) were reached. It takes about 10 min. Afterwards,
the inlet was switched to the flask with the 0.035–0.1 M
stock solution of 3-NS in toluene containing n-decane as
the internal GC standard. This moment was chosen as the
starting point of the reaction. The catalytic tests were car-
ried out at 70°C, 1.0 MPa hydrogen pressure, solution
2
3
Au with the different mean diameters of supported gold
nanoparticles was obtained (d = 2.1± 0.4, 2.8± 0.6 and
l
5
.7± 1.5 nm).
2
.3 Catalyst Characterization
The gold content of the fresh and spent catalyst samples
was measured in vacuo by using the X-ray Fluorescence
(
XRF) technique on an ARL instrument equipped with a
Rh anode (X-ray tube voltage, 50 kV, tube current, 40 mA,
exposure time, 10 s). Transmission electron microscopy
(
TEM) studies were performed with a JEM-2010 (JEOL,
−
1
Japan) electron microscope with a lattice resolution of
and hydrogen feed rates of 0.5 and 60 mL min , respec-
tively. As a result of special tests, we found early that there
is no influence of external or internal mass transfer limi-
tations under the conditions used [12]. The first sample of
liquid products was collected 30 min after feeding of the
substrate; the following samples were taken every 30 min
thereafter. A fresh catalyst sample was used for each cata-
lytic run.
0
.14 nm and a 200 kV accelerating voltage. The mean
diameters (d ) of gold particles for each catalyst sample
l
were determined by counting over 300 particles in TEM
images taken with medium magnification. Textural char-
acteristics of the catalysts were determined from adsorp-
tion–desorption isotherms of nitrogen measured at 77 K by
using an automatic volumetric device ASAP 2400 (Micro-
metritics). The specific surface area was calculated by the
standard Brunauer–Emmett–Teller (BET) method. The
mean pore diameter and total pore volume were obtained
by Barrett–Joyner–Halenda (BJH) analysis. Thermal anal-
ysis was conducted using an STA 449 Jupiter instrument
After the reaction was finished, the reaction flow
(mixture of H and the solution of 3-NS in toluene) was
2
−1
switched to the flow of pure toluene (0.5 mL min ), and
washing of the catalyst was proceeding at the reaction tem-
perature during 30 min. After the reactor was cooled to
room temperature, the catalyst was withdrawn carefully out
of the cartridge and divided into three equal samples taken
at the reactor inlet, at the middle part of the reactor, and
at the reactor outlet. To remove toluene and to prepare the
catalyst for the further characterization, each sample was
placed in the vacuum heat chamber at room temperature,
evacuated to 5 mbar and then slowly heated up to 80°C
with the subsequent drying at this temperature for 3 h.
Reaction products were identified by GC–MS using an
Agilent 7000B Triple Quad System. The concentrations of
substrate and products were measured with GC (Agilent
6890 N instrument with a 19091S-416 HP 5-MS capillary
column of 60.0 m×320 μm×0.25 μm in size). The conver-
(
NETZSCH). The sample (40 mg) in a corundum cruci-
−
1
ble was heated from 25 to 600 °С (ramping 10°C min )
under a flow of air. As a reference sample, calcined silica
was used. The simultaneous TG–DSC–MS measurement
was performed in an apparatus consisting of a STA 449 F1
Jupiter thermal analyzer and a QMS 403D Aëolos quadru-
pole mass spectrometer (NETZSCH, Germany). The meas-
urements were made in a “synthetic air” flow (80 vol% Ar
and 20 vol% O ) in the temperature range of 30–800°C
2
−1
using the heating rate of 10°C min , the gas flow rate of
1
2
5 mL min− and open Al O crucibles. The sample weight
2
3
is 25 mg.
2
.4 Hydrogenation Procedure
sion of 3-NS, X (%), and the selectivity, S (%) calculated
i
by using Equations (1) and (2), respectively:
Hydrogenation of 3-NS in toluene was performed under
continuous-flow conditions in an H-Cube Pro™ setup
ꢀ
ꢁ
0
3NS
X = 1 − C3NS∕C
× 100%,
(1)
(2)
(
Thalesnano, Hungary) equipped with a module for auto-
Ci
matic control of operational parameters (reaction tempera-
ture and pressure, flow rates of liquid feed and hydrogen).
The liquid feed was introduced into the reactor by an HPLC
pump in up-flow mode after mixing with the hydrogen
flow, generated by a built-in electrolytic cell. The hydro-
gen pressure was maintained at the set-point level by means
Si = ∑
× 100%
Cj
0
where C
and C
are the inlet and outlet concentra-
3
NS
3NS
tions of 3-NS, respectively, C is the outlet concentration
i
of the determined product; ΣC is the total concentration of
j
1
3

Study of Catalyst Deactivation in Liquid-Phase Hydrogenation of 3-Nitrostyrene Over…
575
reaction products. The error of determination of 3-NS con-
version and the selectivity to 3-VA is ±1–2%.
by-products, which were formed during hydrogenation of
nitrobenzenes to anilines such as benzylhydroxylamines,
azoxybenzenes, azobenzenes and hydrazobenzenes, were
detected in our experiments.
3
Results and Discussion
The data presented in Table 1 let us to see, that the
catalyst with the mean diameter of the Au nanoparticles
3
.1 Experimental Results of 3-NS Hydrogenation
(d ) equal to 2.1 nm displays the highest activity in terms
l
of 3-NS conversion. The increase in the mean diameter of
the Au particles to 2.8 nm leads to a drop in the catalytic
activity. The catalyst with the highest mean diameter of the
nanoparticles (5.7 nm) shows a significantly lower activity
in comparison with the other samples. These results are in
agreement with published data that indicates an increase of
the catalytic activity of Au/Al O catalysts in the chemose-
The Au/Al O catalysts with the mean diameters of Au
2
3
nanoparticles ranging from 2.1 to 5.7 nm were investigated
in the fixed bed flow reactor in the liquid phase hydrogena-
tion of the 3-NS. The physico-chemical characteristics and
catalytic properties of the different samples are presented
in the Table 1. The products typical of 3-NS hydrogena-
tion over the Au/Al O catalysts are shown in Scheme 1. In
2
3
lective hydrogenation of nitroarenes with the decrease of
gold particle size to less than 3 nm [5, 11, 21]. On the other
hand, the selectivity towards 3-VA is practically unaffected
by the size of the gold nanoparticles (Table 1).
2
3
addition to 3-vinylaniline (3-VA), undesired 3-vinylnitroso-
benzene (3-VNSB), 3-ethylaniline (3-EA) and 3-ethylnitro-
benzene (formed in trace amounts) were observed among
the reaction products. The conversion of 3-NS can proceed
via two routes, i.e., through the desired 3-vinylaniline gen-
eration, or via the 3-nitrostyrene hydrogenation to 3-ethyl-
nitrobenzene leading further to 3-EA as the fully hydrogen-
ated product [18]. No one of any previously reported [3, 5]
In accordance with the obtained results, the Au/Al O -
2
3
2.1 catalyst with the mean particle sizes of 2.1± 0.4 nm
was chosen for the further study of the liquid-phase 3-NS
hydrogenation. Figure 1 illustrates the TEM image for the
Au/Al O -2.1 catalyst. The small black dots represent the
2
3
metal gold particles up to 3.5 nm in diameter. TEM data
shows mainly Au particles of 1.5–2.5 nm in diameter; but
there are also a few smaller (d = 1.0–1.5 nm) and larger
i
(
d = 2.5–3.5 nm) particles. The particle size distribution
i
has a maximum at approximately 2 nm (d = 2.1 nm) with a
l
standard deviation (σ) equal to 0.4.
Figure 2 shows the experimental results of time-on-
stream dependence of the 3-nitrostyrene conversion at the
different initial concentration of 3-NS in toluene: 0.035,
0
.05, 0.075, and 0.1 M. It is seen that increase in the 3-NS
concentration in the reaction mixture leads to a sharper
decrease in the reagent conversion with time-on stream
and, as a result, to the faster catalyst deactivation.
3
.2 Modeling of Catalyst Deactivation
The rate of the first-order reaction over the fresh catalyst
Scheme 1 Reaction pathway for the hydrogenation of 3-nitrosterene
without deactivation (r (C,T)) can be presented by the fol-
(
3-NS) to the products: 3-vinylnitrosobenzene (3-VNSB), 3-vinylani-
o
line (3-VA) and 3-ethylaniline (3-EA)
lowing equation:
Table 1 Effect of the mean Au
particle size on 3-nitrostyrene
conversion and product
selectivity
Tcalc (°С)a
Entry
Catalyst
Au (%)
dl ± σ (nm)
X (%)
S (%)
3-VA
3-VNSB
3-EA
1
2
3
Au/Al O -2.1
400
600
800
1.85
1.82
1.85
2.1± 0.4
2.8± 0.6
5.7± 1.5
94
86
27
95
93
96
2
3
3
3
4
1
2
3
Au/Al O -2.8
2
3
Au/Al O -5.7
2
3
3
-nitrostyrene (0.05 M), T=70 °С, p(H )=1.0 MPa. The performance of the catalysts was evaluated by
2
averaging the 2 samples taken at 30–33 min and 33–36 min after the start of the experiment
a
Calcination temperature
1
3

5
76
A. L. Nuzhdin et al.
0
where Ф (С
,t) is the empirical expression also called
3
NS
“
relative activity” or “activity function”, which depends on
the inlet 3-NS concentration and time-on-stream.
The activity function is defined as the ratio of the current
reaction rate r(t) and a reaction rate r at the initial period
0
of the process, r(t) → = r :
t
0
0
Φ(t) = r(t)∕r = k(t)∕k
0
(5)
0
When the catalyst deactivation begins, the activity func-
tion Ф(t) decreases with time-on-stream.
In the absence of knowledge on the detailed mecha-
nism of the active site deactivation a wide range of empiri-
cal correlations is used for catalyst deactivation modeling
[
33]. On the basis of experimental observations, Froment
and Bischoff [34] proposed three main types of activity
function, as follows: linear Φ = 1−γt, exponential Φ =
1
2
Fig. 1 TEM data of the fresh Au/Al O -2.1 catalyst
2
3
−1
−1
exp(−γt), and hyperbolic Φ =Φ =(1−γt) , where γ—
3
1
−
1
the empirical deactivation constant (min ).
The dependencies of the activity function on time-
0
on-stream at different С
values are non-linear in our
3
NS
case; therefore, the Φ function is not appropriate for their
1
description. Further attempts are undertaken to describe
these dependencies by using the exponential Φ or hyper-
2
bolic Φ function. The log-lin plots of Ф versus time-on-
3
2
0
stream for different С
values are not the straight lines,
3
NS
thus, another function should be used. Applying of the
hyperbolic function results in a good description of experi-
mental results, since the Φ function depends linearly on
3
t at any inlet 3-NS concentrations that have been used
(
Fig. 3). Based on the fitting results, the following formula
can be suggested for the activity function in our case:
ꢀ
ꢁ
1
0
Φ C , t =
ꢀ
ꢁ × t
,
3
NS
0
(6)
1
+ k C
d
3NS
Fig. 2 Time dependence of 3-NS conversion over the Au/Al O -
where k (C 0) is parameter of deactivation depending on
2
3
d
3NS
2
.1 catalyst at different initial concentration (T = 70 °С; p(H ) = 1.0
2
the inlet 3-NS concentration.
MPa)
0
The values of k at different C
can be determined by
d
3NS
fitting the function (6) to the Φ values calculated from the
experimental data by Equations (3) and (5) (the correla-
tion between the fitting curves and the experimental data is
shown in Fig. 4).
r0(C, T) = k(T) × C3NS,
(3)
where k(T) is the first-order rate constant. To take into
account the influence of the catalyst deactivation on the
reaction performance, a so-called “separable deactiva-
tion kinetics” approach is usually applied [31, 32]. In this
approach, the rate of the catalytic reaction is expressed as
a superposition of the function of the initial reaction rate of
the fresh catalyst and a function characterizing the change
of the catalyst activity by deactivation.
Thus, the kinetic model consisting of Equations (4) and
(6) that was developed for the reaction of 3-NS hydrogena-
tion over the Au/Al O catalyst fully describes the dynam-
2
3
ics of the catalyst deactivation (Fig. 3). The dependence of
the parameter of deactivation is fitted well by a parabolic
0
2
−1
function k = B(C
) at B = ∼12.7 min (Fig. 4).
d
3NS
To explain the similar empirical form of Eq. (6) during
isomerization of α-pinene to camphene [32] from a mech-
anistic point of view the authors have assumed that the
formation of carbon deposits and deactivation of the Au/
Al O catalysts occur as a result of interaction between two
The kinetic model of 3-NS hydrogenation expresses the
reaction rate as the function of the main reaction kinetics
(
Eq. 3) and a function of Ф that is related to the current
catalytic activity:
r(C, T, t) = r (C, T) × Φ C , t ,
2
3
ꢀ
ꢁ
adsorbed molecules of the reagent. Taking into account this
0
3NS
0
(4)
1
3

Study of Catalyst Deactivation in Liquid-Phase Hydrogenation of 3-Nitrostyrene Over…
577
Au/Al O -2.1 catalysts using XRF, TEM, low-temperature
2
3
nitrogen adsorption and thermal analysis, respectively.
Analysis of Au content, particle size, and carbon content
was done for the catalyst samples tested in the reaction at
the different initial concentrations of 3-nitrostyrene in tolu-
ene, 0.035, 0.05, 0.075, and 0.1 M (Fig. 2). The character-
istics of the spent catalyst used in the experiment with the
highest 3-NS concentration are provided in the Table 2.
TEM data of the Au/Al O samples taken at the reactor
2
3
inlet and outlet after 3-nitrostyrene (0.1 M) hydrogenation
at 70°C are shown in Fig. 5. The comparison of the mean
particle size of gold in the catalysts taken after reaction
with that in the fresh catalyst has not elucidated noticeable
enlargement of the gold particles in any part of the reac-
tor. An increase in the d value for Au particles, which was
l
observed early after two successive catalytic runs in the
3
-NS hydrogenation [12], is most likely to be caused by the
intermediate regeneration of the Au/Al O catalyst in air
Fig. 3 Time dependence of activity function Φ at different inlet
2
3
3
3
concentrations of 3-NS. The symbols represent Φ values calculated
flow at 400°C.
from the experimental data by Eqs. (3) and (5) in linear scale
Thus, the comparison of characteristics of the fresh and
spent catalysts allows us to conclude that Au content, Au
particle size, and textural properties of Au/Al O catalyst
2
3
have not been changed visibly after 2.5 h on-stream even
after experiments using the highest concentration of rea-
gent (Table 2). This observation agrees with the results of
the investigation of gas-phase hydrogenation of 4-CNB to
4
-chloroaniline in a continuous flow fixed bed reactor over
the Au/Al O catalyst [19]. Authors have not detected any
2
3
noticeable changes in the textural properties or the mean
size of gold particles and suggested that the formation of
carbonaceous deposits is the main reason for the Au/Al O
2
3
catalyst deactivation.
In order to evaluate the contents of carbonaceous depos-
its, TG-DTA analysis of the fresh and spent catalysts was
performed in air atmosphere. The samples of the Au/Al O -
2
3
2
.1 catalyst taken from the different parts of reactor (inlet,
Fig. 4 Dependence of deactivation parameter (k ) as a function of
d
middle part and outlet) after experiments with various ini-
tial concentrations of 3-NS were examined to evaluate the
influence of 3-NS concentration on the carbon content and
elucidate the allocation of carbon deposition along the reac-
tor length. TG/DTG results for the fresh Au/Al O catalyst
inlet 3-NS concentration
consideration, we can speculate, that in our case reagent
(
3-NS) or the intermediates formed on the initial stages of
2
3
reaction are responsible for the formation of carbon depos-
its on the catalysts surface. In order to examine this spec-
ulation, the analysis of the spent catalysts taken from the
different part of reactor (inlet, middle part, and outlet) is
performed.
are presented in Fig. 6a. The differential thermogravimetric
(DTG) curve exhibits the sharp peak with a maximum at
about 90°C, which is explained by the removal of physi-
cally adsorbed water. Additionally, the slow weight loss is
observed at a temperature higher than 200°C, which is usu-
ally ascribed to the gradual removal of residual hydroxyl
groups from the matrix of γ-alumina. The DTA (data not
shown) and DTG curves of the fresh sample show no signs
of any exothermal reaction (Fig. 6a).
3
.3 Study of the Spent Catalysts
To elucidate the impact of the different factors on the cata-
lyst deactivation we compared the gold content, the mean
particle size of gold nanoparticles, textural properties, and
the carbonaceous deposition content in the fresh and spent
Figure 6b displays the DTG curves of the samples of
the Au/Al O -2.1 catalyst taken from the different parts
2
3
of the reactor after experiment using the solution with the
1
3

5
78
A. L. Nuzhdin et al.
Table 2 Physico-chemical characteristics of the fresh and spent Au/
Al O -2.1 catalyst
2
3
Fresh
Spenta
1.82
Content of Au (wt%)
Mean particle size (d ) (nm)
1.85
b
2.1± 0.4
1.8± 0.4 ,
.7± 0.4c
194e
l
1
BET specific surface area (m g−1)
Total pore volume (cm g
Mean pore diameter (nm)
2
215 , 197
d
3
−1
d
e
)
0.61 , 0.58
0.56
d
e
9.6 , 9.4
9.3
a
After 2.5 h on-stream in toluene at 70°C, p(H )=1.0 MPa, 3-nitros-
2
tyrene (0.1 M)
b
Reactor inlet
Reactor outlet
c
d
γ-Al O support
2
3
e
Averaged catalytic sample
Fig. 6 a TG/DTG curves of the fresh Au/Al O -2.1 catalyst; b DTG
2
3
curves of the samples of the Au/Al O -2.1 catalyst after testing for
2
3
3
-NS hydrogenation at 70 °C (C = 0.1 M): at reactor outlet (1), at
middle part of reactor (2), and at reactor inlet (3)
initial concentration of 3-NS equal to 0.1 M. For these
samples, the DTG curves exhibit two weight loss regions:
the first is caused by the removal of adsorbed water (not
shown) and the second (250–520 °C) could be explained
by combustion of carbonaceous species on the catalyst
surface that was confirmed by TG–DSC–MS data [32,
3
5]. The position of the exothermic peak on the DSC
curve as well as carbon dioxide, nitrogen monoxide and
water peaks on the MS curves corresponds to weight loss
region on the DTG curves (Figs. 6b, 7). Thus, carbon
deposits containing some amount of hydrogen and nitro-
gen are formed on the catalyst surface during the liquid-
phase 3-NS hydrogenation.
Fig. 5 TEM data of the Au/Al O -2.1 catalyst after testing for 3-NS
2
3
hydrogenation at 70 °C, C = 0.1 M, p(H ) = 1.0 MPa. a Reactor inlet,
2
b reactor outlet
1
3

Study of Catalyst Deactivation in Liquid-Phase Hydrogenation of 3-Nitrostyrene Over…
579
Fig. 8 Effect of the initial 3-NS concentration on carbon deposit con-
tents in the catalyst samples at the different parts of reactor. Time of
reaction run is 150 min, T = 70 °C
converted over Au/Al O catalyst at the same conditions. In
2
3
both of the above examples, no carbon deposits were found
on the catalyst surface after the experiment. Therefore it
is unlikely that 3-nitrostyrene is a source of carbon depos-
its. In accordance with the generally accepted mechanism
the reduction of nitroarenes to anilines occurs in several
steps, either by the direct route via nitroso and hydroxy-
lamine intermediates or via a condensation route [3, 5]. In
our case, during a hydrogenation of 3-NS to 3-VA over Au/
Al O catalyst, only the 3-vinylnitrosobenzene formed in
Fig. 7 TG–DSC–MS data for the sample of the Au/Al O -2.1 cata-
2
3
lyst (at middle part of reactor) after testing for 3-NS hydrogenation at
7
0 °C, C = 0.1 M
2
3
the first stage of reaction was detected in solution. Taking
into account that nitrosobenzene can be strongly coordi-
nated via –N=O group to the surface of oxides containing
Lewis acidic sites [36, 37], we suggest 3-vinylnitrosoben-
zene to be the main source of carbon deposits. Thus, the
following scheme for the catalyst deactivation can be pro-
posed (Scheme 2).
Figure 8 shows the dependence of carbonaceous deposit
content at the reactor inlet, at the middle part of the reac-
tor, and at the reactor outlet on the initial 3-NS concentra-
tion. The amount of carbon deposits in the catalyst was
determined as follows: the corresponding weight loss of
the fresh catalyst equal to 1.6% was subtracted from the
amount of relative weight loss of the spent catalyst in the
temperature range 250–520°C (TG curves). It is clearly
seen, that the content of carbonaceous deposits in the sam-
ples during the same time of reaction is increased with
the increasing of the initial 3-NS concentration (Fig. 8).
Besides, the content of the carbonaceous deposits on the
catalyst surface at the inlet of the reactor is higher than
that at the outlet. This observation leads us to conclusion
that the carbon-containing deposits on the surface of the
catalyst are obtained starting from 3-nitrostyrene or the
intermediates formed at the initial stages of 3-nitrostyrene
hydrogenation to 3-vinylaniline.
4 Conclusions
Dynamics of the Au/Al O catalyst deactivation in liquid-
2
3
phase chemoselective hydrogenation of 3-nitrostyrene was
studied in the fixed bed flow reactor. It was found that for-
mation of carbon-based deposits was the main reason of
deactivation of the Au/Al O сatalyst, while the leaching
2
3
of Au from the catalyst surface and increase in the mean
gold particle size were not detected. The content of the
carbonaceous deposits on the spent catalyst at the inlet of
the reactor is higher than that at the outlet and increases as
the initial concentration of 3-NS grows. The character of
carbonaceous deposits allocation and analysis of reaction
The conversion of 3-NS was not observed under reac-
tion conditions (T=70°C, p(H )=1.0 MPa) over Al O . If
2
2
3
nitrogen was used instead of hydrogen, 3-NS also was not
1
3

5
80
A. L. Nuzhdin et al.
1
1. Shimizu K, Miyamoto Y, Kawasaki T, Tanji T, Tai Y, Satsuma A
(
2009) J Phys Chem C 113:17803–17810
1
2. Nuzhdin AL, Moroz BL, Bukhtiyarova GA, Reshetnikov SI,
Pyrjaev PA, Aleksandrov PV, Bukhtiyarov VI (2015) ChemPlu-
sChem 80:1741–1749
1
3. Matsushima Y, Nishiyabu R, Takanashi N, Haruta M, Kimura H,
Kubo Y (2012) J Mater Chem 22:24124–24131
1
1
4. Shimizu K-I, Miyamoto Y, Satsuma A (2010) J Catal 270:86–94
5. Mitsudome T, Mikami Y, Matoba M, Mizugaki T, Jitsukawa K,
Kaneda K (2012) Angew Chem Int Ed 51:136–139
Scheme 2 Proposed mechanism of carbon deposits formation during
the selective liquid-phase hydrogenation of 3-nitrostyrene over Au/
Al O catalyst (where asterisk is the surface active site of the catalyst)
1
1
1
6. Reis PM, Royo B (2009) Tetrahedron Lett 50:949–952
7. Serna P, Concepcion P, Corma A (2009) J Catal 265:19–25
8. Yarulin A, Berguerand C, Yuranov I, Cardenas-Lizana F,
Prokopyeva I, Kiwi-Minsker L (2015) J Catal 321:7–12
2
3
products point out to 3-vinylnitrosobenzene as the main
source of the Au/Al O catalyst deactivation coinciding
19. Cardenas-Lizana F, Lamey D, Li M, Keane MA, Kiwi-Minsker
L (2014) Chem Eng J 255:695–704
2
3
2
2
0. He D, Shi H, Wu Y, Xu B-Q (2007) Green Chem 9:849–851
1. Hartfelder U, Kartusch C, Makosch M, Rovezzi M, Sa J, van
Bokhoven JA (2013) Catal Sci Technol 3:454–461
with the speculation arising from the kinetic modeling.
Acknowledgements The authors thank Russian Science Foundation
2
2
2
2
2. Wang X, Perret N, Delgado JJ, Blanco G, Chen X, Olmos CM,
Bernal S, Keane MA (2013) J Phys Chem C 117:994–1005
3. Perret N, Wang X, Delannoy L, Potvin C, Louis C, Keane MA
(
Grant 14-23-00146) for the financial support of this study.
(
2012) J Catal 286:172–183
4. Wang XD, Perret N, Keane MA (2012) Chem Eng
10:103–113
5. Irfan M, Glasnov TN, Kappe CO (2011) ChemSusChem
J
References
2
1
.
Downing RS, Kunkeler PJ, van Bekkum H (1997) Catal Today
4
:300–316
3
7:121–136
2
2
6. Hessel V (2009) Chem Eng Technol 32:1655–1681
2
3
.
.
Blaser H-U (2006) Science 313:312–313
7. Baxendale IR, Brocken L, Mallia CJ (2013) Green Process Synth
Blaser H-U, Steiner H, Studer M (2009) ChemCatChem
2
:211–230
1
:210–221
2
2
8. Wiles C, Watts P (2014) Green Chem 16:55–62
9. Moroz BL, Pyrjaev PA, Zaikovskii VI, Bukhtiyarov VI (2009)
Catal Today 144:292–305
4
.
Ananikov VP, Khemchyan LL, Ivanova YuV, Bukhtiyarov VI,
Sorokin AM, Prosvirin IP, Vatsadze SZ, Medvedko AV, Nuriev
VN, Dilman AD, Levin VV, Koptyug IV, Kovtunov KV, Zhi-
vonitko VV, Likholobov VA, Romanenko AV, Simonov PA,
Nenajdenko VG, Shmatova OI, Muzalevskiy VM, Nechaev MS,
Asachenko AF, Morozov OS, Dzhevakov PB, Osipov SN, Voro-
byeva DV, Topchiy MA, Zotova MA, Ponomarenko SA, Borsh-
chev OV, Luponosov YuN, Rempel AA, Valeeva AA, Stakheev
AYu, Turova OV, Mashkovsky IS, Sysolyatin SV, Malykhin VV,
Bukhtiyarova GA, Terentev AO, Krylov IB (2014) Russ Chem
Rev 83:885–985
3
0. Jones RV, Godorhazy L, Varga N, Szalay D, Urge L, Darvas F
(
2006) J Comb Chem 8:110–116
3
3
1. Froment GF, Bischoff KB (1961) Chem Eng Sci 16:189–201
2. Solkina YuS, Reshetnikov SI, Estrada M, Simakov AV, Murzin
DY, Simakova IL (2011) Chem Eng J 176–177:42–48
3. Satterfield CN (1984) Heterogeneous catalysis in practice.
McGraw-Hill Book Company, New York
3
3
3
4. Froment GF, Bischoff KB (1979) Chemical reactor analysis and
design. Wiley, New York
5
6
7
.
.
.
Cardenas-Lizana F, Keane MA (2013) J Mater Sci 48:543–564
Serna P, Boronat M, Corma A (2011) Top Catal 54:439–446
Corma A, Serna P, Concepcion P, Calvino JJ (2008) J Am Chem
Soc 130:8748–8753
5. Nuzhdin AL, Artiukha EA, Bukhtiyarova GA, Zaytsev SYu,
Plyusnin PE, Shubin YuV, Bukhtiyarov VI (2016) RSC Adv
6
:88366–88372
3
3
6. Meijers S, Ponec V (1996) J Catal 160:1–9
8
.
Motoyama Y, Lee Y, Tsuji K, Yoon S-H, Mochida I, Nagashima
H (2011) ChemCatChem 3:1578–1581
7. Richner G, van Bokhoven JA, Neuhold Y-M, Makosch M, Hun-
gerbuchler K (2011) Phys Chem Chem Phys 13:12463–12471
9
.
Corma A, Serna P (2006) Science 313:332–334
1
0. Boronat M, Concepcion P, Corma A, Gonzalez S, Illas F, Serna
P (2007) J Am Chem Soc 129:16230–16237
1
3