Environmentally friendly hydrogenation of nitrobenzene to p-aminophenol using heterogeneous catalysts

Article abstract of DOI:10.1016/j.jcat.2010.08.005

Nitrobenzene was converted to p-aminophenol at 353 K, using water as solvent and a bi-functional catalyst composed of a mechanical mixture of supported Pt catalyst with zirconium sulphate calcined at 773-923 K. The performance of this system is independent of the support used for Pt, and various supports, such as pure or sulphated zirconia and titania, carbon, MgLa mixed oxide, give similar results. At low Pt content, the reaction rate is first order relative to nitrobenzene, and the slow step is the partial hydrogenation of nitrobenzene to phenylhydroxylamine, which requires only minute amounts of Pt. At higher Pt loadings, the rate of hydrogenation of nitrobenzene to phenylhydroxylamine and consequently to aniline takes over that of the acid-catalysed Bamberger rearrangement of phenylhydroxylamine to p-aminophenol. The selectivity of this step depends critically on the solid acid: strong acids such as sulphated zirconia or zeolites give poor selectivities because they tend to decompose the hydroxylamine intermediate. This process does not require sulphuric acid or additives such as DMSO or alkylsulphides, thereby simplifying the downstream processing.

Full text of DOI:10.1016/j.jcat.2010.08.005

Journal of Catalysis 275 (2010) 250–256
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Environmentally friendly hydrogenation of nitrobenzene to p-aminophenol
using heterogeneous catalysts
b
b
b
Abhay Deshpande a, F. Figueras a, , M. Lakshmi Kantam , K. Jeeva Ratnam , R. Sudarshan Reddy ,
⇑
N.S. Sekhar b
a Institut de recherches sur la catalyse et l’environnement, 2 av. A. Einstein, 69626 Villeurbanne Cedex, France
b Inorganic and Physical Chemistry Division, Indian Institute of Chemical Technology, Hyderabad 500 607, India
a r t i c l e i n f o
a b s t r a c t
Article history:
Nitrobenzene was converted to p-aminophenol at 353 K, using water as solvent and a bi-functional cat-
alyst composed of a mechanical mixture of supported Pt catalyst with zirconium sulphate calcined at
773–923 K. The performance of this system is independent of the support used for Pt, and various sup-
ports, such as pure or sulphated zirconia and titania, carbon, MgLa mixed oxide, give similar results. At
low Pt content, the reaction rate is first order relative to nitrobenzene, and the slow step is the partial
hydrogenation of nitrobenzene to phenylhydroxylamine, which requires only minute amounts of Pt. At
higher Pt loadings, the rate of hydrogenation of nitrobenzene to phenylhydroxylamine and consequently
to aniline takes over that of the acid-catalysed Bamberger rearrangement of phenylhydroxylamine to
p-aminophenol. The selectivity of this step depends critically on the solid acid: strong acids such as
sulphated zirconia or zeolites give poor selectivities because they tend to decompose the hydroxylamine
intermediate. This process does not require sulphuric acid or additives such as DMSO or alkylsulphides,
thereby simplifying the downstream processing.
Received 12 June 2010
Revised 5 August 2010
Accepted 5 August 2010
Available online 9 September 2010
Keywords:
Nitrobenzene hydrogenation
p-Aminophenol
Phenylhydroxylamine
Bi-functional catalysis
Platinum catalysts
Solid acids
Zirconium sulphate
Sulphated zirconia
Ó 2010 Elsevier Inc. All rights reserved.
1. Introduction
of water [7]. The PAP thus formed needs to be isolated by neutralis-
ing the acid. This leads to the production of salts as effluents, with
The selective catalytic hydrogenation of nitrobenzene (NB) to
p-aminophenol (PAP) is a reaction of industrial importance since
PAP is an important pharmaceutical intermediate for the manufac-
ture of paracetamol, a widely used analgesic drug. Two-step pro-
cesses have been reported for the preparation of PAP, in which
nitrobenzene is first partially hydrogenated to phenylhydroxyl-
amine (PHA), followed by the Bamberger rearrangement of PHA to
PAP in acidic conditions. Platinum is considered as the best noble
metal, and carbon the best support, for the hydrogenation of nitro-
compounds [1]. Higher selectivity to PHA has been reported by the
additionofsulphurcompoundssuch asdimethylsulphoxide(DMSO)
[2], sulphides [3], or basic additives such as pyridine [4], phosphines
[5] or phosphates [6]. When basic additives are used, a two-step
process has to be used for the synthesis of PAP, involving the isola-
tion of PHA after the partial hydrogenation and addition of sulphuric
acid after the hydrogenation step [7] to effect the rearrangement.
Using DMSO or sulphides, the process can be converted to one pot
synthesis. After 3 h of reaction, a 75% conversion of nitrobenzene
has been reported with a PAP selectivity of 79% using 27 mL of
concentrated sulphuric acid for 57 mL of nitrobenzene in 406 mL
more complex separations. Most patents also report the addition
of surfactants [8] which increase the area of contact between nitro-
benzene and water, usually used as solvent.
A few attempts have been made to substitute sulphuric acid
by a solid acid. For instance, Chaudhari et al. [9] reported the
use of zeolites and acid resins for this reaction. The results were,
however, disappointing since starting from 93.4 mmol of nitro-
benzene, the conversion was 97% but the yield was only 14%.
Very recently, a bi-functional Pt on highly acidic sulphated zirco-
nia was used also with relatively poor results since the PAP yield
was below 30% [10]. Indeed, the authors investigated the Bam-
berger rearrangement separately and observed selectivity lower
than 45% at full conversion for this step, using highly acidic solid
acids. It is important to note here that as reported by Groscova
et al. [11] or Wang et al. [10], the Bamberger rearrangement can
yield not only PAP but also aniline and azoxybenzenes. There is
therefore an interest for processes, which would be highly selec-
tive towards PAP, would not produce salts as effluents, and could
avoid complex separation, particularly of the additives required
to promote selectivity.
It can be pointed out that this is a difficult task since the
coupling of hydrogenation with acid-catalysed rearrangement is
rather uncommon in organic synthesis. The formal reaction
⇑
Corresponding author. Fax: +33 472445399.
E-mail address: figueras@catalyse.cnrs.fr (F. Figueras).
0021-9517/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2010.08.005

A. Deshpande et al. / Journal of Catalysis 275 (2010) 250–256
251
Table 1
(PHA)
NHOH
Main characteristics of the acid catalysts.
NH2
NO2
(NB)
Catalyst
Supplier Calcination
Surface area ‘S’
content
H2SO4
k3
temperature (K) (m2 gÀ1
)
Pt/C
H2
+
(wt%)
S or N comp
k1
BEA (Si/Al = 22)
ZrSAD-3
Zeolyst
Home
made
Home
made
Home
made
Home
made
MEL
723
Uncalcined
720
2.7
–
17.8
OH
k2
(PAP)
ZrSAD-3
ZrSAD-3
ZrSAD-1
898
923
923
11
4.2
4.3
2.6
NH2
15
122
(AN)
MEL-999/1
MEL-922/1
Loba (sulphated)
898
318
309
20
1.4
–
0.4
MEL
Loba
Uncalcined
898
Scheme 1. Simplified reaction scheme.
Chemie
Home
made
Sulphated TiO2
773
110
1.9
scheme, reported in Scheme 1, has some common points with
the bi-functional isomerisation of paraffins, with the major dif-
ference that in paraffin isomerisation, the concentration of the
olefin intermediate is governed by the thermodynamic equilib-
rium of hydrogenation, while here all steps are controlled by
kinetics. An obvious conclusion that can be drawn from Scheme
1 is that the rate of rearrangement of PHA to PAP (k3) and that
of the complete hydrogenation of PHA to aniline (AN) (k2) are
the critical factors in determining the selectivity towards the for-
mation of PAP. Thus, for the catalyst system to be highly selec-
tive for PAP, k3 needs to be substantially higher than k2. On the
other hand, AN would be the main product if k2 is substantially
higher than k3.
The zirconium sulphate (ZrSAD-3) was calcined in air at differ-
ent temperatures between 823 and 923 K. The catalysts, Loba-S
and MEL-999/1, were used after calcination in air at 898 K for
2 h. The final calcination temperature was obtained with a linear
programming rate of 2 K/min. The BET surface areas and the sul-
phur contents of these solid acids are reported in Table 1.
Supported Pt catalysts were prepared using zirconia, sulphated
zirconia, titania, or a MgLa mixed oxide as carrier. In this series of
supports, zirconia and titania are amphoteric oxides, sulphated zir-
conia is a strong acid, and MgLa mixed oxide has been shown to be
a strong solid base [13]. Using aqueous solutions of chloroplatinic
acid on the oxide carriers and amine salt for sulphated carriers,
0.1–2% Pt was introduced by the method of impregnation at incip-
ient dryness. After impregnation, the resulting solids were dried
overnight at 383 K, calcined in air at 673 K, and finally reduced
in H2 flow at 523 K. A commercial 1% Pt/C was used as received.
We report here a novel reaction system, which permits the sub-
stitution of sulphuric acid and selectivity promotors and gives rea-
sonable yields with high selectivity [12].
2. Experimental
2.2. Catalyst characterisation
2.1. Catalyst preparation
The structures of the catalysts were checked by powder X-Ray
diffraction using a Bruker D5005 instrument. The S content was
determined at SCA (Service central de micro analyse du CNRS, Sol-
aise) by infrared spectroscopy analysis of the gases evolved upon
oxidation at 1623 K in oxygen. The surface properties of the solid
acids were characterised by measuring their surface area and acid-
ity. The surface area was determined from the adsorption isotherm
of nitrogen at 77 K on samples activated in N2 flow for 2 h at 523 K.
The acid properties were determined by adsorption of NH3 at 353 K,
using microcalorimetry coupled with volumetry. The solid taken in
a sample cell was first evacuated up to 10À5 torr at 673 K, then trans-
ferred to a Tian–Calvet microcalorimeter. Small increments of NH3
were then introduced into the cell, and differential heat of adsorp-
tion was measured. The differential heat represents the strength
of the ammonia interaction with the surface, therefore the acid
strength, whereas the amount of NH3 measures the number of acid
sites. The dispersion of Pt on zirconia was estimated from electron
micrographs obtained by TEM, using a JEOL 2010 LaB6 instrument,
and carbon replicas. The resolution of the apparatus is 0.19 nm and
permits the detection of particles 0.5 nm in size. The distribution of
particles size was obtained from the micrographs taken from
different zones of the sample. A surface average size was calculated
A series of sulphated zirconias were investigated. Two samples
(reference numbers 922-1 and 999/1) were supplied by MEL
Chemicals, UK, and one was supplied by Loba Chemie (Mumbai, In-
dia). These were sulphated with a 1 N solution of sulphuric acid. A
laboratory-made zirconia was prepared by hydrolysis of ZrOCl2
(Aldrich) at pH 10. This solid obtained after filtration was sulphat-
ed by treatment with a 1 N solution of sulphuric acid (15 mL 1 N
sulphuric acid for 1 g zirconia, stirred 15 min), followed by drying
at 393 K and calcination at 923 K with a ramp of 1–2 K/min (de-
noted as ZrSAD-1). The sample 999/1 was received in a sulphated
form and used as such.
Another sulphated catalyst, ZrSAD-3, was prepared as follows.
Forty-eight grams zirconyl chloride (ZrOCl2ÁxH2O, MW: 322.249)
was dissolved in 375 mL water, and Zr hydroxide was precipitated
at constant pH = 10 by adding ammonium hydroxide. The precipi-
tate was aged at 353 K for 12 h after which it was separated by
centrifugation and washed several times with hot water to remove
the excess ammonia and chloride ions. The wet precipitate was
transferred to a conical flask. A solution of 21 mL concentrated sul-
phuric acid in 150 mL water was prepared and added to the conical
flask. The slurry was stirred for 1 h at room temperature. Within
this time, most of the precipitate dissolved and a turbid solution
was obtained. This solution was fed to a rotary evaporator, and
water was evaporated at 353 K. At the end, a white precipitate
appeared almost instantaneously. The precipitate was then trans-
ferred to a glass pan and dried overnight at 393 K. After drying, it
was sieved and stored in a bottle. Before use, the sample was
calcined at the desired temperature with a ramp of 1–2 K/min.
from this distribution using the formula d =
R R
nidi3/ nidi2.
2.3. Catalytic reaction
The hydrogenation of nitrobenzene to p-aminophenol was
investigated in a 100 mL autoclave from Autoclave Engineers, filled

252
A. Deshpande et al. / Journal of Catalysis 275 (2010) 250–256
at about 80%. The reaction conditions were 353 K and hydrogen
pressure 10 bar, reported to be optimum for the process using sul-
phuric acid [14]. In a typical experiment, 75 mL of water and 3 mL
of nitrobenzene (NB) were introduced in the autoclave equipped
with an efficient gas induction agitator, temperature sensor, sam-
pling tube, and a baffle. To the above, 0.01 g of 2% Pt/ZrO2 and
0.5 g of ZrSAD-3A, previously calcined at 923 K for 4 h, were added
and the reactor closed. Nitrogen was fed to the reactor (2.5 bar)
and purged three times. Then, nitrogen was replaced with hydro-
gen (10 bar) and purged three times. The agitation was started
and maintained at 1200 rpm, and the autoclave was heated to
353 K, under 10 bar H2. Samples were periodically withdrawn to
monitor the progress of reaction. When using nitrobenzene/water
ratios higher than 3/75, the reaction medium formed two phases,
aqueous and organic; in that case, at the end of the reaction, the
reaction mixture was dissolved in ethanol in order to obtain a
homogeneous solution and analysed to determine the conversion
and selectivities. The reaction produces PAP, traces of PHA, and
aniline. The reactants and products were analysed by HPLC using
a C18 column and a solution of acetonitrile (30%) in water as elu-
ent. The mass balance was found to be better than 90%. The conver-
sion is classically defined as the fraction of NB disappeared from
the reaction mixture, and the selectivity as the molar fraction of
PAP in the products.
773 K due to desulphation. The loss in the range 500–1260 K
amounts to 57.6 wt% of the weight of the dry sample, matching
with the 56.5 wt% expected for the decomposition of Zr(SO4)2 into
ZrO2. The weight loss below 500 K, attributed to the loss of water,
permits to calculate a formula Zr(SO4)2, 3.8H2O, which is in agree-
ment with the XRD pattern. After calcination at 773 K and storage
in air, a small rehydration occurs, giving a ratio Zr/H2O about 1.
The characterisation of the acidities of the solid acids has been
reported previously [15], and the comparison between dehydrated
zirconium sulphate (ZRSAD-3-898 K) and sulphated zirconias is re-
ported in Fig. 3. The two sulphated zirconias have been calcined at
973 K and consist of ZrSAD-1, prepared by impregnation of zirconia
by sulphuric acid, and ZrSAD-3. It appears clearly that the modifi-
cation of the parameters of the preparation allows tuning of the
surface properties of sulphated zirconias. In particular, the
ZrSAD-3 sample calcined at 893 K shows a much lower number
of sites of weaker strength compared to sulphated zirconia.
Representative electron micrographs of two Pt/ZrO2 catalysts,
containing 1% and 2% Pt are reported in Fig. 4. More particles are
visible on the 2% Pt sample, but both samples show a similar dis-
tribution of particles sizes as illustrated in Fig. 5, the main differ-
ence being the presence of a few bigger particles on the 2%Pt
sample. The surface average size calculated from these distribu-
tions is 2.2 nm for the 1%Pt sample and 2.3 nm for the 2%Pt sample.
3.2. Definition of the conditions of reaction
3. Results and discussion
An exploratory study was first performed on the possible sub-
stitution of sulphuric acid by solid acids using ZrSAD-1 sulphated
zirconia calcined at 923 K and DMSO as promoter. The results ob-
tained for the reaction of 3 mL of NB in 75 mL of water are reported
in Table 2. In the presence of DMSO, PAP is formed both on Pt/C and
on Pt/TiO2 (entries 1 and 2), showing that these supports indeed
exhibit some acidity. The selectivity is significantly increased on
Pt/TiO2, and PHA is also detected with a selectivity of about 12%,
which indicates that the acid-catalysed rearrangement of PHA is
slow and controls the overall process.
The number of acid sites can be increased by adding a solid acid
in a mechanical mixture. This strategy was applied first using
DMSO as additive. The reaction rate is slightly promoted by the
addition of an acidic zeolite such as BEA (entry 3). The interesting
point is that using a sulphate as solid acid, DMSO can be omitted.
As illustrated in Table 2, the yield of PAP is significantly improved
3.1. Characterisation of the catalysts
The chemical compositions of the solid acids used here are
reported in Table 1. Except BEA zeolite and MEL-999/1, all solid
acids are mesoporous (average pore size about 3.8 nm), which lim-
its the possibility of intra-particle diffusional limitations. The
material called ZRSAD-3 exhibits the composition of zirconium
sulphate and the XRD pattern of Zr(SO4)2, 4H2O (not shown) after
drying at oven temperature, and of Zr(SO4)2 after calcinations at
773 K (Fig. 1). Sulphated zirconias (calcined above 923 K) show
by contrast the pattern of tetragonal zirconia.
The results of the thermal analysis of ZRSAD-3 are reported in
Fig. 2 for the fresh sample and a sample calcined at 773 K. The fresh
sample shows weight losses below 500 K corresponding to dehy-
dration in successive steps, followed by a large weight loss above
1100
1000
900
800
700
600
500
400
300
200
100
0
5
10
20
30
40
50
60
70
2-Theta - Scale
Fig. 1. Powder XRD pattern of ZRSAD-3 after calcinations at 623 K.

A. Deshpande et al. / Journal of Catalysis 275 (2010) 250–256
253
Fig. 2. Differential gravimetric analysis of ZrSAD-3 fresh or calcined at 923 K.
formed in ethanol as solvent, and this solvent was therefore inves-
tigated. The results were very poor, since the conversion in a
mixture of ethanol/water = 70:30 reached only 6.6% after 1 h of
reaction, with a selectivity to PAP of 77%, compared to 30% of con-
version and 83% selectivity in pure water. The isomerisation of PHA
is therefore much slower in ethanol, and this inhibition is attrib-
uted to the fact that ethanol masks the acidity of the catalyst by
virtue of stronger adsorption on the active sites, when compared
to water. The adsorption of water on the other hand gives rise to
enhanced Bronsted acidity, which is essential for the Bamberger
rearrangement.
140
120
100
80
ZrSAD1 973K
ZrSAD3 973K
ZrSAD3 898K
60
40
3.3. Effect of the amount of Pt on the catalytic properties
20
If the type of support has little influence on the catalytic results,
the balance between metallic/acid sites is very important. The Pt
loading on the support was varied from 0.1% to 2% using cationic
exchange from Pt amine. The activity was measured at 353 K, 3.8
bars, with different amounts of Pt/ZrS catalyst (0.01, 0.03 and
0.1 g) and ZrSAD-3A calcined at 823 K (0.1 g) as solid acid (Table
3). In order to obtain a homogeneous liquid phase, 1.5 mL of NB
in 50 mL of water was used. The initial rate was measured from
the conversion after 1 h of reaction, conditions in which the selec-
tivity to PAP was better than 95%. The results are reported in Fig. 6.
Using 0.1 g of acid catalyst, the rate first increases linearly with the
number of Pt atoms engaged in the reaction, then reaches a pla-
teau. This is the behaviour expected for bi-functional catalysis, in
which the overall rate is controlled by the hydrogenation of nitro-
benzene at the Pt surface for a low Pt content, then by the acid-cat-
alysed isomerisation of PHA, at the higher Pt loadings. Indeed,
when using 0.025 g of catalyst at 2%Pt, increasing the amount of
acid catalyst from 0.1 g to 0.25 g induces a proportional increase
in the conversion after 1 h from 18% to 36.7%, thus confirming that
the overall process is controlled by the acid-catalysed isomerisa-
tion of PHA at high Pt loading.
0
0
50
100
150
200
250
300
350
400
450
NH3 adsorbed (µmol.g-1)
Fig. 3. Comparison of the acid properties of zirconium sulphate (ZRSAD-3-898 K)
and sulphated zirconias.
(entries 4 onwards). This result suggests a partial reduction of sul-
phates in the conditions of reaction, resulting in a promotion of
PHA selectivity.
The coupling of reactions requires a respect of the balance of the
two functions. The comparison of results with different amounts of
Pt in the catalyst shows that a large amount of Pt favours the
hydrogenation of NB to aniline, with a large loss of selectivity to
PAP (comparison of entries 7 and 8). In general, the support of Pt
has little importance, as long as Pt is well dispersed. It can be cho-
sen from a large variety of supports, hydrophobic, such as carbon,
acidic or basic.
Another aspect was the choice of the solvent. The original work
of Rylander et al. [1] on the hydrogenation of NB to PHA was per-

254
A. Deshpande et al. / Journal of Catalysis 275 (2010) 250–256
Fig. 4. High-resolution electron micrographs of 1%Pt and 2%Pt/ZrO2 catalysts.
0,5
0,4
0,3
0,2
0,1
0,0
Table 2
Effect of DMSO on the hydrogenation of nitrobenzene at 353 K, at 10 bar H2, in the
presence of solid acids.
1% Pt
2% Pt
Entry Catalyst
Without DMSO
With DMSO
Conv.
(%)
PAP
sel.
Conv.
(%)
PAP
sel.
1
2
3
1%Pt/C (0.1 g)
Pt/TiO2 (0.1 g)
Pt/C (0.13 g) + BEA zeolite
(1 g)
1%Pt/C (0.01 g) + ZRSAD-3
(0.5 g)
2%Pt/ZrO2 (0.01 g) + ZrSAD-3
(0.5 g)
2%Pt/TiO2 (0.01 g) + ZrSAD-3
(0.5 g)
27.4a
36a
0
27
50.7a
26b
20
68
86
17.8a
4
5
6
7
8
62b
85b
78b
98
83
82
7.9
77
2% Pt/MgLa (0.05 g) + ZrSAD-3 98.9c
<.0,67
0,67-1,02
1,02-1,7
Particle size (nm)
1,7-2,38
2,38-3,06
>3,06
(0.5 g)
2% Pt/MgLa (0.01 g) + ZrSAD-3 75b
(0.5 g)
Fig. 5. Distribution of particles sizes for the two 1%Pt and 2%Pt/ZrO2 catalysts.
a
b
c
Reaction time: 7 h.
Reaction time: 6 h.
Reaction time: 1 h.
3.4. Kinetics of the reaction: effect of the concentration of nitrobenzene
The use of mesoporous supports eliminates the possibility of
diffusional limitations by intra-particle mass transfer. The use of
a Rushton turbine and hollow holder improves the homogeneity
of the reaction medium. The effect of the stirring rate was investi-
gated and it was observed that the reaction rate and selectivity in-
creased with the stirring rate and reached a plateau beyond
1000 rpm, so that the stirring rate was fixed at 1200 rpm.
The initial rate as a function of concentration of nitrobenzene
was measured using 2%Pt/ZrSAD-3, 2%Pt/MgLa, and 1% or 2%Pt/
ZrO2 catalysts, which give similar rates normalised to the amount
of Pt, mixed with ZrSAD-3 calcined at 823 K. At low NB concentra-
tion, the reaction medium is monophasic and it is possible to mon-
itor the conversion and selectivity as a function of time, but above
3 mL of NB, a biphasic system is observed in the absence of stirring.
In that case, the reaction was run for 1 h, the reaction medium was
dissolved into methanol and analysed. The initial rate was com-
Table 3
Effect of the amount of Pt in the mechanical mixture on the activity measured at
353 K, with different amounts of Pt/ZrSAD-1 catalyst mixed with ZrSAD-3 catalyst as
solid acid. Reaction conditions: H2 pressure 3.8 bars, 1.5 mL of NB in 50 mL.
Entry Amount of Pt
cat (g)
%Pt
loading
ZrSAD-3
(g)
Conv. (%) after PAP sel.
1 h
(%)
1
2
3
4
5
6
7
8
0.025
0.01
0.01
0.01
0.03
0.1
2
2
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.25
15.9
18.8
6
96.1
95
0.5
0.5
0.5
0.1
0.5
2
97.4
97.2
97.9
97.2
95.7
93.3
5.2
17.8
17.8
12.9
36.7
0.1
0.025

A. Deshpande et al. / Journal of Catalysis 275 (2010) 250–256
255
Table 4
0.60
0.55
0.50
0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
Composition of the reaction mixture for the hydrogenation of nitrobenzene at 353 K,
after 1 h in standard conditions (3 mL of nitrobenzene in 75 mL of water, 353 K,
hydrogen pressure 10 bars), using 0.5 g of different solid acids used in mechanical
mixture with 0.01 g of 2% Pt/MgLa.
0.25g ZrS
Entry
Solid acid
Products distribution (%)
PAP sel. (%)
NB
PHA
PAP
AN
1
2
3
4
5
6
7
8
9
ZrO2 922/1
71.8
24.4
74.7
57
79.1
37.5
68.8
76.3
18.9
5.6
13.9
5.0
0.4
0.0
5.2
3.6
2.1
13.6
2.2
2.2
5.6
36
19.7
0.5
7.1
2.7
0.4
20.3
20.3
14.7
6.4
7.9
2.2
ZrSAD-3 (calc. 973 K)
ZrSAD-3 (calc. 948 K)
ZrSAD-3 (calc. 923 K)
ZrSAD-3 (calc. 823 K)
TiO2 sulf
Sulf 922/1
HBEA zeol
MEL-999/1
22.2
83.6
94.1
0.9
22.8
11.5
0.5
1.1
56.8
20.5
18.8
66.5
0.0
0.1
0.2
0.3
0.4
0.5
0.6
2% Pt/ZrO2 of similar dispersions. This suggests that mass transfer
limitations are not important. The observation of a high selectivity
to PAP also favours this conclusion.
Pt amount (mgPt)
Fig. 6. Effect of the amount of Pt in the mechanical mixture on the activity
measured at 353 K, with different amounts of Pt/ZrS catalyst mixed with ZrSAD-3A
catalyst as solid acid. Reaction conditions: H2 pressure 3.8 bars, 1.5 mL of NB in
50 mL.
3.5. Comparison of different solid acids
It was reported earlier that the Bamberger rearrangement of
PHA to PAP was very sensitive to the nature of the solid acid; there-
fore, an influence could be expected for the bi-functional hydroge-
nation of nitrobenzene to PAP. The comparison of a series of solids
is reported in Table 4. The results are given after 1 h at 353 K, using
a reaction mixture consisting of NB (3 mL), water (75 mL), and
0.01 g of 2% Pt/MgLa mixed with 0.5 g of solid acid. As illustrated
in this table, the temperature of calcination has a significant
influence on the catalytic properties (entries 2–5). In particular,
calcination at high temperature, which induces strong acid proper-
ties, is unfavourable in this case, since an increase in conversion is
observed at the cost of a selectivity decrease. The low selectivity
over sulphated titania and HBEA zeolite was expected from their
behaviour in Bamberger rearrangement.
Using the best catalyst, ZrSAD-3A calcined at 823 K, the conver-
sion varies as a function of time as reported in Fig. 8. The final
result is comparable to those reported with sulphuric acid as acid
catalyst.
The results obtained here for the hydrogenation of NB can be
compared to those obtained for the isomerisation of PHA, summa-
rised in Table 5. In both cases, the most selective catalyst is ZrSAD-
180
160
140
120
100
80
3
4
1
60
2
40
20
1
0
0,0
0,5
1,0
1,5
2,0
NB Concentration (mol.L-1)
Fig. 7. Variation of the initial rate, as a function of the initial concentration, for
reactions at 353 K and 10 bars of hydrogen, using 2%Pt/ZrSAD-3 (point 1), 2%Pt/
MgLa (point 2), 1%Pt/ZrO2 (point 3), and 2%Pt/ZrO2 (point 4).
100
80
puted from the conversion after one hour to determine the reaction
order. The results at 353 K are reported in Fig. 7. They show the
same reaction rate for Pt on different supports and a first-order
dependence of this rate on the NB concentration. Tanielyan et al.
[16] reported a constant rate for H2 pressures above 3.8 bar, for a
reaction in water, catalysed by Pt/C and sulphuric acid, and it can
therefore be assumed that the reaction is zero order relative to
hydrogen at 20 bar. The turnover frequency (rate per surface Pt
60
40
Conversion of nitrobenzene
Selectivity into PAP
atom) deduced from this plot is 18 mol sÀ1
.
Boudart and Cheng [17] have applied the test proposed by Koros
and Nowak [18,19] for the liquid-phase hydrogenation of cyclo-
hexene on nickel to ensure that the measured rates were not influ-
enced by slow heat or mass transfer. The Koros–Nowak Criterion
requires that the values of the turnover rate of a reaction on cata-
lysts with the same metal particle size but different metal loading
be the same, for the measured rate to reflect only the chemical pro-
cess. Here, the same activity is measured when using either 1% or
20
0
0
1
2
3
4
5
6
Time (h)
Fig. 8. Evolution of the conversion and selectivity into p-aminophenol as a function
of time at 453 K.

256
A. Deshpande et al. / Journal of Catalysis 275 (2010) 250–256
Table 5
here after 8 h is comparable to that reported for the current indus-
trial process after 3 h [25]. It can be pointed out that the industrial
process is based on sulphuric acid and this higher activity can be ac-
counted for by the higher density of protons in sulphuric acid com-
pared to any solid acid. The lower rate of the all heterogeneous
process is however compensated by the significant reduction in
the e-factor induced by the elimination of the liquid acid and a
much simpler separation of the products.
Comparison of a few solid acids for the isomerisation of phenylhydroxylamine at
353 K. 109 mg of PHA added to 7 mL of water reacted in the presence of 0.1 g of solid
acid, giving PAP, o-aminophenol, aniline and nitrobenzene.
Sample
Reaction time (min)
PHA conv.
PAP sel.
BEA zeolite
30
30
60
30
60
30
60
30
30
60
49.1
100.0
33.6
35.1
47.3
23.2
40.0
100.0
21.8
39.1
61.6
29.8
94.8
99.9
96.8
97.0
92.1
99.6
94.5
95.2
Sulphated TiO2 (calc. 773 K)
ZrS (from LOBA zirconia)
ZrS-exZrOCl2 (calc. 893 K)
ZrSAD-1 (calc. 923 K)
4. Conclusions
ZrSAD-3 (calc. 893 K
ZrS MEL-999/1 (calc. 923 K)
An all heterogeneous process for the bi-functional conversion of
nitrobenzene into p-aminophenol has been described, using the
combination of zirconium sulphate and supported platinum. The
reaction kinetics is characteristic of a bi-functional mechanism,
in which the slow step is the hydrogenation of NB to PHA at low
Pt content and the isomerisation of PHA to PAP at high Pt loading.
This system requires only very small amounts of Pt, and the sup-
port of Pt is not important. The most critical point is the choice
of the solid acid, which must be selective for the Bamberger rear-
rangement. This rearrangement has been shown to be very
demanding, and the strong effect of the calcination temperature
illustrates this point. Strong solid acids are not good catalysts for
this process.
Table 6
Experiments in
sulphate, at 353 K, pressure 20 bar.
a
700 ml reactor, using a mixture of 2% Pt/ZrO2 and zirconium
Exp.
no.
NB
Water
(mL)
Time
(h)
NB conv.
(%)
PAP Sel.
(%)
AN Sel.
(%)
(mL)
1
2
3
4
5
6
40
40
56
56
56
56
490
350
490
490
490
490
6
8
6
6
8
8
33.8
97
27.2
22.8
84.5
86.7
91.9
86
93.1
90.8
87.0
88.1
8.1
14
6.9
9.2
13.0
11.9
Acknowledgments
Mechanical mixture of catalysts: experiments 1 and 2: 2% Pt/ZrO2 (0.1 g), ZRS (5 g);
experiments 3–6: 2% Pt/ZrO2 (0.14 g), ZRS (7 g).
We thank M.M. Aouine for the investigations by TEM and the
analytical service of IRCELyon and SCA-Solaise for chemical and
textural analysis. We also thank Cefipra for funding this work in
the frame of the Franco-Indian cooperation.
3 calcined at most at 923 K. This catalyst can reach high conversion
of PHA with high selectivity, and the same behaviour is observed in
the hydrogenation of NB. This comparison shows that the loss of
selectivity occurs in the second step, the isomerisation of PHA. It
is surprising that strong acids such as sulphated zirconias give poor
results in the Bamberger rearrangement, when the industrial pro-
cess uses sulphuric acid. It can be pointed out that the industrial
reaction is indeed performed in rather diluted sulphuric acid. It
has been reported long ago that Fe3+ at ppm levels catalyses the
decomposition of PHA by a redox mechanism [20]. It is also well
known that radicals can be formed on sulphated zirconias by oxi-
dation [21–24]. It can then be speculated that over sulphated zirco-
nias, the selectivity leak occurs by this oxidation path.
References
[1] P.N. Rylander, I.M. Karpenko, G.R. Pond, DE 2118369, 1971, to Engelhard.
[2] P.N. Rylander, I.M. Karpenko, G.R. Pond, DE 2118334, 1971, to Engelhard.
[3] D.C. Caskey, D.W. Chapman, EP 85890, 1983, to Mallinckrodt.
[4] J. Le Ludec, DE 2327412, 1973, to Rhone-Poulenc.
[5] J. Freudenreich, J. Stohrer, WO 9928289, 1999, to Consortium fur
Elektrochemische Industrie.
[6] A.H. Sharma, P. Hope, EP 147879, 1985, to AKZO.
[7] D.C. Caskey, D.W. Chapman, US 4571437, 1986, to Mallinckrodt.
[8] S.S. Sathe, US 4176138, 1979, to Mallinckrodt.
[9] R.V. Chaudhari, S.S. Divekar, M.J. Vaidya, C.V. Rode, US 6028227, 2000, to CSIR.
[10] S. Wang, Y. Ma, Y. Wang, W. Xue, X. Zhao, J. Chem. Technol. Biotechnol. 83
(2008) p. 1466.
[11] D. Groskova, M. Stolcova, M. Hronec, Catal. Lett. 69 (2000) 113.
[12] F. Figueras, A. Deshpande, L.K. Mannepalli, J.R. Katuri, S.R. Ragireddy, S.S.
Nujilla, EP 2058050, to CNRS, C.S.I.R, Vinati Organics Ltd., 2009.
[13] M.L. Kantam, H. Kochkar, J.M. Clacens, B. Veldurthy, A. Garcia-Ruiz, F. Figueras,
Appl. Catal. B 55 (2005) 177.
[14] P.N. Rylander, I.M. Karpenko, and G.R. Pond, FR 2092589, 1972, to
Engelhard.
[15] K.J. Ratnam, R.S. Reddy, N.S. Sekhar, M. Lakshmi Kantam, A. Deshpande, F.
Figueras, Appl. Catal. A. 348 (2008) 26.
3.6. Experiments in a larger reactor
The scale up of the results obtained in a small reactor to a larger
one offers better conditions for the determination of the material
balance and is fruitful only when diffusional limitations are
negligible; therefore, a few experiments have been performed in a
larger reactor (700 mL from Autoclave Engineers), using a concen-
tration of 1 mol LÀ1 of NB. Due to the separation of the organic/li-
quid phases at rest, the reactions were run for 6 or 8 h, then the
whole reaction mixture was transferred into a 1-l beaker, 100 ml
of methanol was added to obtain a homogeneous solution, which
was analysed by HPLC (ODH-C18 column, 250 mm, wave length
254 nm, 70:30 water/acetonitrile, flow 1 mL/min). The results re-
ported in Table 6 illustrate the reproducibility of the experiment
and show that a good selectivity can be reached at high conversion,
as was first observed in the small reactor. Indeed the yield obtained
[16] S.K. Tanielyan, J.J. Nair, N. Marin, G. Alvez, R.J. McNair, D. Wang, R.L. Augustine,
Org. Process Res. Dev. 11 (2007) 681.
[17] M. Boudart, W.C. Cheng, J. Catal. 106 (1987) 134.
[18] R.M. Koros, E.J. Nowak, Chem. Eng. Sci. 22 (1967) 470.
[19] R.J. Madon, M. Boudart, Ind. Eng. Chem. Fund. 21 (1982) 438.
[20] D. Mulvey, W.A. Waters, J. Chem. Soc. Perkin Trans. 2 (1977) 1868.
[21] A. Ghenciu, D. Farcasiu, J. Mol. Catal. A. 109 (1996) 273.
[22] A. Ghenciu, D. Farcasiu, Catal. Lett. 44 (1997) 29.
[23] A.F. Bedilo, M.A. Plotnikov, N.V. Mezentseva, A.M. Volodin, G.M. Zhidomirov,
I.M. Rybkin, K.J. Klabunde, Phys. Chem. Chem. Phys. 7 (2005) 3059.
[24] A.F. Bedilo, A.M. Volodin, Kinet. Catal. 50 (2009) 314.
[25] D.C. Miller, WO 9325515, 1993, to Mallinckrodt.