Rapid hydrogenation of aromatic nitro compounds in supercritical carbon dioxide: Mechanistic implications via experimental and theoretical investigations

Article abstract of DOI:10.1002/adsc.201200103

An exceptionally rapid hydrogenation of nitrobenzene to aniline [TOF=252,000 h^-1] over palladium containing MCM-41 (Pd/MCM-41) with excellent yield of >99% can be achieved in supercritical carbon dioxide at 50 °C and a hydrogen pressure of 2.5 MPa. It has been observed that this promising method preferred a single phase between liquid substrate and carbon dioxide-hydrogen system. The ascendancy of the supercritical carbon dioxide medium is established in comparison with the conventional organic solvent and solvent-less conditions. Changes in the reaction parameters such as carbon dioxide and hydrogen pressure, temperature and the reaction time do not affect the selectivity. A combined experimental and theoretical study has elucidated the mechanism under the studied reaction condition because experimental observations revealed a direct conversion of nitrobenzene to aniline. However, density functional theory (DFT) calculation shows that the direct conversion is energetically unfavourable; hence, a stepwise mechanism has been proposed. Theoretical predictions and experimental observations suggested that the rate-limiting step of nitrobenzene conversion is different from that of the liquid phase hydrogenation. This catalytic process can also be successfully extended to the hydrogenation of other aromatic nitro compounds with different substituents. Easy separation of the liquid product from catalyst and the use of an environmentally friendly solvent make this procedure a viable and an attractive green chemical process. Copyright

Full text of DOI:10.1002/adsc.201200103

FULL PAPERS
DOI: 10.1002/adsc.201200103
Rapid Hydrogenation of Aromatic Nitro Compounds in
Supercritical Carbon Dioxide: Mechanistic Implications via
Experimental and Theoretical Investigations
Maya Chatterjee,a,* Abhijit Chatterjee,b Hajime Kawanami,a,* Takayuki Ishizaka,a
Toshishige Suzuki,a and Akira Suzukia
a
Research Center for Compact Chemical System, 4-2-1 Nigatake, Miyagino-ku, Sendai 983-8551, Japan
Fax : (+81)-22-237-5388; phone: (+81)-22-237-5213; e-mail: c-maya@aist.go.jp or h-kawanami@aist.go.jp
Accelrys KK, Kasumigaseki, Tokyu bldg. 17F, 3-7-1, Kasumigaseki, Chiyoda-ku, Tokyo, Japan
b
Received: February 14, 2012; Published online: && &&, 0000
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201200103.
Abstract: An exceptionally rapid hydrogenation of sity functional theory (DFT) calculation shows that
nitrobenzene to aniline [TOF=252,000 hꢀ1] over pal- the direct conversion is energetically unfavourable;
ladium containing MCM-41 (Pd/MCM-41) with ex- hence, a stepwise mechanism has been proposed.
cellent yield of >99% can be achieved in supercriti- Theoretical predictions and experimental observa-
cal carbon dioxide at 508C and a hydrogen pressure tions suggested that the rate-limiting step of nitro-
of 2.5 MPa. It has been observed that this promising benzene conversion is different from that of the
method preferred a single phase between liquid sub- liquid phase hydrogenation. This catalytic process
strate and carbon dioxide-hydrogen system. The as- can also be successfully extended to the hydrogena-
cendancy of the supercritical carbon dioxide medium tion of other aromatic nitro compounds with differ-
is established in comparison with the conventional ent substituents. Easy separation of the liquid prod-
organic solvent and solvent-less conditions. Changes uct from catalyst and the use of an environmentally
in the reaction parameters such as carbon dioxide friendly solvent make this procedure a viable and an
and hydrogen pressure, temperature and the reaction attractive green chemical process.
time do not affect the selectivity. A combined experi-
mental and theoretical study has elucidated the
mechanism under the studied reaction condition be- Keywords: density functional theory; hydrogenation;
cause experimental observations revealed a direct nitroaromatics; palladium (Pd/MCM-41); supercriti-
conversion of nitrobenzene to aniline. However, den- cal carbon dioxide
Introduction
liquid phase hydrogenation of nitrobenzene faces the
following difficulties. Firstly, low selectivity due to the
Aromatic amines are used as an important intermedi- formation of various hazardous by-products such as
ate in the manufacture of dyes, drugs, plastics and ag- hydroxylamine,[6,7] which is toxic and accumulation of
ricultural product such as pesticides. Aniline, a proto- this compound leads to exothermic decomposition. To
typical aromatic amine, is the precursor of many in- avoid this shortcoming an additive was used to in-
dustrial chemicals such as methylenediphenyl diiso- crease the selectivity, which creates a waste problem
cyanate, and also used in rubber processing, pigments, and, if additives cannot be recycled, the process be-
dyestuffs and drugs.[1] Aniline is generally prepared comes more expensive and suffesr from reduced effi-
by the gas phase or liquid phase hydrogenation of ni- ciency of the catalyst after certain runs. Hence, it is
trobenzene over Ni- or Cu-based catalysts.[2] This re- difficult to implement the process in the industrial
action is extremely facile and can easily occur under scale. Secondly, the formation of coloured condensa-
mild reaction conditions. Thus, a vast amount of liter- tion products, thirdly product separation and finally,
ature is available on the hydrogenation of nitroben- serious environmentalal issues due to harmful organic
zene using various types of supported metal (Ni, Pt, solvents are further disadvanatges. A recent research
Pd, Ru and Au) and non-metal catalysts.[3–5] However, has demonstrated a solvent-free route to the forma-
Adv. Synth. Catal. 0000, 000, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
ÞÞ
These are not the final page numbers!

Maya Chatterjee et al.
FULL PAPERS
tion of aniline using ultrafine Pt nanoparticles depos- and ainmed to achieve high performance (complete
ited on multi-walled carbon nanotubes as catalyst.[8] conversion and selectivity) after suitable adjustment
In fact, carbon nanotubes are still expensive to pro- of reaction conditions to the specific properties of the
duce and the structures are hard to control.[9] Further- reaction medium.
more, Pt/MCM, Pt on reduced graphene oxide and Pd
nanoparticles on an ionic liquid brush were also con-
sidered as selective hydrogenation system for nitroar- Results and Discussion
enes.[10]
Supercritical carbon dioxide (scCO2) (1c = Reactions in scCO2 are strongly affected by phase be-
0.466 gcmꢀ3, Tc =31.18C, Pc =73.8 bar) has been used haviour. It is relevant to determine the number of
as an alternative medium for a number of heterogene- phases present in the reaction system by visual obser-
ously catalyzed reactions, often with higher reaction vations. Figure 1 presents a view of the substrate, CO2
rates and different product distributions, as well as and H2 mixture, which revealed that at a low pressure
high selectivity, in comparison with those in conven- of 8 MPa, two phases (CO2-H2, substrate) were pres-
tional organic solvents.[11–17] It has a great potential to ent in the system, but the existence of a single phase
overcome the disadvantages associated with conven- (CO2-H2-substrate) was observed at or above 12 MPa
tional heterogeneous catalysts such as mass transfer CO2 pressure.
limitation, which frequently occurs between reactants
in the gas phase and liquid phase. Beneficial effects
can also arise from the high compressibility of CO2 Catalyst Screening
and by variation of the density with comparable small
changes in the reaction parameters such as pressure It is well-known that a noble metal possesses the abil-
and temperature.[18] A significant advantage behind ity to generate atomic hydrogen,[30] thus, initial
the hydrogenation in scCO2 is particularly the high screening was performed over different metal cata-
miscibility of the reacting hydrogen gas and high dif- lysts and the results are shown in the Table 1. All the
fusivity. In particular, hydrogenation driven by heter- catalysts mentioned were supported on MCM-41 with
ogeneous catalysts in scCO2 has attracted considera- same metal content of ~1 wt% and the average metal
ble attention because of the enhanced reaction rate, particle size varied within the range of 5–10 nm. The
high product selectivity and clean product separa- use of MCM-41 as support material is logical because
tion.[19–21]
according to the reaction path, the hydrogenation of
Supercritical carbon dioxide (scCO2) has also been nitrobenzene to aniline is associated with the forma-
used in the hydrogenation of nitroaromatics.[22] For ni- tion of water, thus, most acceptable supports should
trobenzene hydrogenation, although a high conver- be hydrophobic like carbon.[31] However, carbon sup-
sion could be readily achieved, the selectivity of ani- ports are not inert and therefore, sometime cause un-
line was low because of the formation of noticeable desired by-product formation which hampers the se-
amounts of by-products (conversion of nitroben- lectivity of aniline. Hence, inert support materials
zene=70%; aniline selectivity=84%).[21,23] However,
an improved selectivity of aniline was achieved over
Ni/g-Al2O3 catalyst (conversion=68–73%; selectivity
of aniline= >99%; reaction time=50 min), but the
conversion was quite low and the process required
product separation.[24]
Up to now, much of the research has focused on
the activity and selectivity of this reaction. Very little
information exists regarding the actual reaction mech-
anism. In 1898, Haber first proposed a reaction path
of the hydrogenation of nitrobenzene based on the
electrochemical reduction. After that there is still on-
going research in the scientific community to find out
the exact reaction mechanism depending on detecta-
ble intermediates.[25–29] Interestingly, no intermediate
has been detected under the studied reaction condi-
tions even with a reaction time of 1 min.
Therefore, in this work, we specifically focused on
a fundamental understanding of reaction pathways
best suited for scCO2 through the combined approach
Figure 1. Phase behaviour of nitrobenzene at 508C PH
=
2.5 MPa: (a) initial, (b) 8 MPa, (c) 11 MPa, and (d) 12 M2Pa
of experimental findings and theoretical investigations CO2 pressure.
2
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 0000, 000, 0 – 0
ÝÝ
These are not the final page numbers!

Rapid Hydrogenation of Aromatic Nitro Compounds in Supercritical Carbon Dioxide
Table 1. Catalyst screening.
Entry[a]
Solvent
Catalyst
Conversion [%]
1
scCO2
–
0.0
2
3
4
5
5
7
8
9
scCO2
scCO2
scCO2
scCO2
scCO2
scCO2
scCO2
ethanol
methanol
2-propanol
hexane
Pd/MCM-41
Pt/MCM-41
Rh/MCM-41
Ru/MCM-41
Au/MCM-41
Ir/MCM-41
Ni/MCM-41
Pd/MCM-41
Pd/MCM-41
Pd/MCM-41
Pd/MCM-41
100.0
83.0
1.6
0.7
4.3
2.3
0.5
86.8
52.2
46.5
31.2
10
11
12
Figure 2. Variation of CO2 pressure. Reaction conditions:
catalysts=0.1 g, substrate=16.2 mmol; PH =2.5 MPa, tem-
2
perature=508C and reaction time=5 min.
[a]
Reaction
conditions:
catalyst=0.1 g,
substrate=
16.2 mmol, PCO =12 MPa, PH =2.5 MPa, and tempera-
2
2
ture=508C and reaction time=10 min. Solvent used
(entry 9–12)=5 mL.
lying cause of change in conversion. At lower pres-
sure, the reaction mixture existed as a two phase
system; the gas-liquid (H2-CO2 and substrate) phase,
with hydrophobicity are particularly preferred.[32] where mass transfer limitation of H2 causes the reac-
Noteworthy, mesoporous MCM-41 type materials are tion to proceed with a low reaction rate (TOF=
inert and possess similar hydrophobicity to the acti- 70,000 hꢀ1). On the other hand, with increasing CO2
vated carbon.[33] Independent of the metal ion used, pressure, the solubility of nitrobenzene increased and
this reaction led to the formation of aniline as the the system reached to a single phase (H2-CO2-sub-
only detected product. Apparently, Pd and Pt cata- strate). As a result, there was a significant increase in
lysts are more efficient with highest overall activity the reaction rate (TOF=252,000 hꢀ1) due to the en-
related to the high conversion (Table 1; entries 2 and hanced the availability of the reactant. However,
3). On the other hand, Rh, Ru, Au and Ir all belong above 14 MPa CO2 pressure, a slight drop in the con-
to the lower efficiency category because of the very version as well as TOF (217,000 hꢀ1) was observed
low conversions of 2–5% (Table 1; entries 4 to 7) and might be attributed to a dilution effect. At a con-
under the present reaction conditions, and follow the stant volume, increased CO2 pressure forces more
order of Pd>Pt>Au >RhffiRuffiIr. It is unreasonable molecules of CO2 between the reactant molecules,
to compare the performance of different metal cata- thus mimicking the enhanced dilution effects of the
lysts based on the conversion, or even turnover fre- conventional solvent. The hydrogenation in the ab-
quency (TOF), because the reaction time is too short sence of CO2 gave only 2.3% conversion, while under
here, some of these catalysts might exhibit long induc- the same conditions when CO2 (4 MPa) was intro-
tion and activation periods. Additionally, in most duced the conversion was increased significantly to
cases the initial activity is very different from that at 23.6%. This observation illustrates that using scCO2
steady state. Here, a Pd catalyst was chosen to pro- as reaction solvent can have a significant advantage
ceed further because of its high efficiency under the toward the efficiency of the reaction. In contrast to
studied reaction conditions (PCO =12 MPa, PH = the previous report,[21] no other by-products were de-
2
2
2.5 MPa, temperature=508C, reaction time 10 min.). tected.
Notably, without catalysts as well as with only the
silica support, no reaction was observed.
Effect of Reaction Time
Effect of CO2 Pressure
Catalytic runs were performed at different reaction
times to get an idea about the kinetics of the reaction.
The influence of different reaction parameters on the Reactions were carried out with an identical amount
hydrogenation of nitrobenzene was investigated. The of substrate, catalyst and at a fixed temperature
CO2 pressure effect is relatively general; with increas- (508C) and pressure (PCO =12 MPa, PH =2.5 MPa).
2
ing CO2 pressure conversion increased (Figure 2). In The response of conversion or yield (%) t2o the altera-
these sets of screening, the highest conversion of tions in the reaction time is illustrated in Figure 3.
100% was reached at 12 MPa and then started to de- The results show that the conversion of nitrobenzene
crease when the pressure exceeded 14 MPa and sug- was increased from 77.2 to 100% as the time changed
gested that phase behaviour is likely to be the under- from 1 to 10 min, which indicated an exceptionally
Adv. Synth. Catal. 0000, 000, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
3
ÞÞ
These are not the final page numbers!

Maya Chatterjee et al.
FULL PAPERS
1 mole of aniline and 2 moles of water. Now, accord-
ing to the present reaction conditions, ca. 0.05 mole of
hydrogen, which corresponds to 2.5–3 MPa at 508C
would be necessary for the hydrogenation of
0.0162 mole of nitrobenzene. Interestingly, complete
conversion of nitrobenzene was attained within the
same pressure range (2.5–3 MPa). When the amount
of substrate decreased, the low concentration of hy-
drogen would be perfect to provide complete conver-
sion. Moreover, at the fixed substrate concentration
of 0.0162 mol, when the hydrogen pressure rose to
4 MPa (ca. 0.08 mol) or above, the reaction was much
faster but the selectivity of aniline decreased, which
corresponds well with literature reports.[21,23] Thus,
managing the hydrogen pressure might be an impor-
tant parameter to control the selectivity of desired
aniline. Hence, based on the substrate concentration,
2.5 MPa was chosen as optimum pressure under the
described reaction conditions.
Figure 3. Conversion (yield [%]) of aniline vs. reaction time.
Reaction conditions: catalysts=0.1 g, substrate=16.2 mmol;
PCO =12 MPa, PH =2.5 MPa, and temperature=508C.
2
2
Effect of Temperature
In addition to the hydrogen pressure, the effect of
temperature was also significant on the catalytic be-
haviour. As can be seen from Figure 5a, the reaction
became slow at a low temperature of 358C, but in-
creasing the temperature changed the reaction rate.
From the visual observation of phase behaviour at
358C, a single phase of CO2-H2-substrate was con-
firmed. Hence, one would expect better performance
even at a lower temperature if phase behaviour is the
only controlling factor. However, from the kinetic
view point, it is normal for the reaction rate to in-
Figure 4. Effect of H2 pressure on nitrobenzene conversion.
Reaction conditions: catalysts=0.1 g, substrate=16.2 mmol,
PCO =12 MPa, time=10 min and temperature=508C.
2
rapid reaction (TOF=252,000 hꢀ1) in comparison crease with temperature, which suggested that the re-
with hydrogenation of other compounds studied pre- action kinetics were in action here rather than phase
viously.[18] Although the conversion was influenced by behaviour. Thus, an optimum temperature of 508C
the reaction time, the selectivity of aniline remains was selected to achieve the highest overall activity for
unchanged throughout the reaction. No other by- the nitrobenzene hydrogenation.
product formation was observed during the course of
From the initial reaction rate an Arrhenius plot was
the reaction, even in the shortest reaction time of obtained and the apparent activation energy of
1 min.
15 kcalmolꢀ1 (temperature range 35 to 608C) was cal-
culated from the slope of the plot (Figure 5, bottom).
The resultant activation energy is comparable with
the literature data obtained in the liquid phase hydro-
genation of nitrobenzene using a Pd catalyst.[34,35]
Effect of H2 Pressure
The effect of hydrogen pressure on the reaction was
significant and investigated under the fixed tempera-
ture (508C) and pressure (PCO =12 MPa) (Figure 4). Comparison with Conventional Organic Solvents
2
Again, all reaction mixtures were in a single phase
(no effect of hydrogen pressure was observed on the As mentioned above, nitrobenzene was hydrogenated
phase behaviour) as confirmed by visual observation. rapidly to aniline in scCO2. To make direct compari-
At the low hydrogen pressure of 0.5 MPa, the conver- son with conventional organic solvents, the reaction
sion was low (ca. 20%) and then 100% conversion was conducted under similar reaction conditions. In
was reached when the pressure increased to 2.5 MPa. this case it can be seen from Table 1 (entries 9–12)
From the Scheme 1, stoichiometrically 1 mole of ni- that nitrobenzene conversion was lower for organic
trobenzene required 3 moles of hydrogen to produce solvents and solvent-less conditions after identical re-
4
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 0000, 000, 0 – 0
ÝÝ
These are not the final page numbers!

Rapid Hydrogenation of Aromatic Nitro Compounds in Supercritical Carbon Dioxide
Scheme 1. Reaction pathway of the nitrobenzene hydrogenation.
action times. For example, in hexane, a solvent with decreased in the order of p>m>o. Following the ni-
comparable density and polarity as scCO2 only ca. trotoluenes, nitroanisoles also behaved in a similar
40% conversion was observed. The significantly im- way (entries 4–6). In addition, the described method
proved efficiency in the CO2 medium can thus be re- can also be applied successfully to the hydrogenation
lated to an improved H2 availability, which may result of dinitro compounds such as 2,4-dinitroACTHUNGTRENNUNG
from the highly increased solubility of H2 in CO2.
ACHTUNGTRENNUNG
(82.0%), and 2,4-diaminotoluene (28.0%) (entry 7).
Chloronitrobenzene, a representative of halogenated
nitro compounds also performed well to form chlor-
oaniline with the selectivity of >90% (entries 8 and
Hydrogenation of Other Nitroaromatic Substrates
The Pd/MCM-41 catalyst can be successfully applied 9). Furthermore, the Pd/MCM-41 catalyst continued
to the hydrogenation of other nitroaromatic com- its scope to the hydrogenation of nitrobenzoic acid
pounds. Table 2 presents the hydrogenation of differ- (entries 10–12). In each case, independent to the con-
ent substrates under the optimized reaction conditions version, only amino compounds were formed. Thus,
(temperature=508C, reaction time=10 min, PCO
=
with applications of the Pd/MCM-41 catalyst its is fea-
12 MPa and PH =2.5MPa). The hydrogenations of 2o- sible to extend the hydrogenation to other nitroaro-
2
m- and p-nitrotoluenes were producing only their cor- matics with different functional groups under the
responding amines (entries 1–3), and the conversion present reaction conditions.
Adv. Synth. Catal. 0000, 000, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
5
ÞÞ
These are not the final page numbers!

Maya Chatterjee et al.
Table 2. Hydrogenation of different nitroaromatics.[a]
FULL PAPERS
Entry Substrate
Conversion
[%][e]
Product
(selectivity [%])
1
78.5
2
3
4
89.8
100.0
67.1
5
80.1
6
92.3
66.3
7[b]
8[c]
9[d]
10
95.5
46.8
46.3
Figure 5. ACHTUNGTRENNUNG(Top) Effect of temperature on nitrobenzene con-
version with time. Reaction conditions: catalyst=0.1 g, sub-
strate=16.2 mmol, PCO =12 MPa, PH =2.5 MPa and time=
2
2
10 min. (Bottom) Arrhenius plot.
Possible Reaction Pathway of Nitrobenzene
Hydrogenation
Here, we have attempted to shed light on the reaction
mechanism to explore a possible reaction path under
the present reaction conditions. In general, hydroge-
nation of nitrobenzene includes the sequence of (i)
a direct route and (ii) a condensation route which is
favoured under basic conditions as shown in
Scheme 1 top and bottom, respectively. To determine
the possible reaction pathway, the activity of the inter-
mediate is very important and sometimes provides
a direct proof of the reaction mechanism. As men-
tioned before, from the very beginning (reaction
time=1 min.) except for aniline no other intermedi-
ate compounds were detected (Figure 3). Further-
more, on the basis of the experimental results, no con-
densation intermediate was detected throughout the
11
66.9
90.9
12
[a]
Reaction
conditions:
catalysts=0.1 g,
16.2 mmol, PCO =12 MPa, PH =2.5 MPa, temperature=
508C and reaction time=10 min.
28% 2,4-diaminotoluene.
substrate=
2
2
[b]
[c]
[d]
[e]
9.8% aniline.
1.1% aniline.
Amount of the substrate converted into products in%
reaction, hence, only the direct route will be consid- arately under similar reaction conditions. Figure 6
ered. To understand the possible reaction path, the represents a comparison between the product profile
hydrogenations of the two intermediates (nitrosoben- of nitrobenzene and nitrosobenzene hydrogenations
zene and phenylhydroxylamine) were carried out sep- with time, which revealed that the yield (%) of ani-
6
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 0000, 000, 0 – 0
ÝÝ
These are not the final page numbers!

Rapid Hydrogenation of Aromatic Nitro Compounds in Supercritical Carbon Dioxide
tions. Interestingly, the rate is faster in comparison to
the formation of aniline from nitrobenzene, which is
attributed to the possibility of phenylhydroxylamine,
as an intermediate and nitrobenzene!phenylhydrox-
ylamine as possibly the slowest step of this reaction
sequence. Hence, from the experimental observations,
themost probable route to the aniline formation is (i)
nitrobenzene!aniline (as no intermediate compound
was detected even in the shortest reaction time) and
(ii) nitrobenzene!phenylhydroxylamine!aniline.
DFT Study on the Mechanism
Figure 6. Product distribution of hydrogenation of nitroben-
zene and nitrosobenzene with time.
At this point, it is difficult to understand the exact re-
action path only from the experimental findings. To
gain further insights into the reaction mechanism the-
line from nitrosobenzene is negligible in comparison oretical calculations were performed. At a first step
to nitrobenzene. In addition, calculation of the initial towards elucidating the reaction mechanism, we set
rate (TOF) of aniline formation from nitrosobenzene out to determine the adsorption energies on the Pd
(TOF=6264 hꢀ1)
and
nitrobenzene
(TOF= surface (Table 3) (for geometry see the Supporting In-
252,000 hꢀ1) provides a significant difference. The formation; Figure S3). The present results predicted
slowest rate of hydrogenation indicating that nitroso- that aniline possesses the lowest adsorption energy of
benzene could not be the possible intermediate spe-
E
ads =- 8.9 kcalmolꢀ1 (Table 3; entry 4). The calculat-
cies involved in the aniline formation. This is again ed adsorption energies of nitrobenzene and nitroso-
confirmed from the results of competitive hydrogena- benzene are very close to each other (Table 2 en-
tion between nitrobenzene and nitrosobenzene (1:1) tries 1 and 2), which suggested a competitive nature
and initial rate of the reaction was significantly re- of adsorption on the Pd surface as also observed from
duced (TOF=93,600 hꢀ1). On the other hand, phenyl- experiments. In comparison to the reactant and inter-
hydroxylamine was rapidly hydrogenated to aniline mediate, the adsorption energy of phenylhydroxyla-
(TOF=288,000 hꢀ1) under the same reaction condi- mine (ꢀ27 kcalmolꢀ1) is higher and corresponded
well with literature.[36] The strong adsorption of phe-
nylhydroxylamine on the hydrogen-saturated Pd sur-
face made it possible to obtain the highest selectivity
of aniline and the absence of phenylhydroxylamine in
the reaction product can also be explained.
Table 3. Adsorption energies of reactant/intermediates/prod-
uct.
Entry[a]
Molecule
Eads (kcalmolꢀ1)
ꢀ12.9
Having determined the adsorption energies of the
principal reactant, intermediate and the product spe-
cies on Pd (100) we can compute the overall energet-
ics of each individual step of the hydrogenation of ni-
trobenzene (Table 4). Calculations indicate that the
surface reaction steps of nitrobenzene to aniline are
highly exothermic and correspond well with the litera-
ture.[37] The formation of phenylhydroxylamine could
happen through the direct hydrogenation of nitroben-
zene or via nitrosobenzene. The energetic difference
1
2
ꢀ10.5
3
4
ꢀ27.8
ꢀ8.9
Table 4. Calculated reaction energies for different reaction steps of nitrobenzene hydrogenation under the studied reaction
condition.
Entry
Reaction
DEa react [kcalmolꢀ1]
DHa react [kcalmolꢀ1]
1
2
3
4
5
C6H5NO2 +H2!C6H5NH2 +H2O
C6H5NO2 +H2!C6H5NO+H2O
C6H5NO+H2!C6H5NHOH
19.6
18.8
10.6
8.2
ꢀ45.0
ꢀ11.0
+4.2
C6H5NHOH+H2!C6H5NH2 +H2O
C6H5NO2 +H2!C6H5NHOH
ꢀ55.0
16.9
ꢀ97.0
Adv. Synth. Catal. 0000, 000, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
7
ÞÞ
These are not the final page numbers!

Maya Chatterjee et al.
FULL PAPERS
between the two pathways is obvious. In the former, the rate-determining step, which is different from that
the individual steps of nitrobenzene!nitrosobenzene of the liquid phase hydrogenation.
and nitrosobenzene!phenylhydroxylamine are re-
In summary, the unique properties of scCO2 such as
quired to overcome activation barriers of 18.8 and mass transfer, high availability of reactant gas and
10.6 kcalmolꢀ1, respectively (Table 4; entries 2 and 3). easy product isolation are beneficial parameters,
Interestingly, the reaction step of nitrosobenzene to which provide a greener and simpler way to the for-
phenylhydroxylamine is mildly endothermic (DH= mation of aniline from nitrobenzene with a very high
4.2 kcalmolꢀ1). The high barrier along with the endo- TOF of 252,000 hꢀ1 without any by-product formation
thermicity makes this an unlikely path on Pd. On the or catalyst deactivation. Successful application of this
contrary, phenylhydroxylamine from nitrobenzene is method is also possible to the other mononitro-, dini-
exothermic (DH=ꢀ97 kcalmolꢀ1) with an activation troarenes, and halogenated nitroaromatic compounds
barrier of 16.9 kcalmolꢀ1 (Table 4; entry 5), and could for the formation of their corresponding anilines.
be the effective choice as an intermediate in the hy-
drogenation of nitrobenzene. Furthermore, calcula-
tion shows that the conversion phenylhydroxyl- Experimental Section
amine!aniline possesses the lowest activation barrier
of 8.2 kcalmolꢀ1 and could not be the rate-limiting Catalyst Synthesis
step as described by several authors in the liquid
The catalyst was prepared by an in-situ addition of 1 wt%
phase hydrogenation of nitrobenzene.[38,39] Therefore,
instead of phenylhydroxylamine!aniline, the slowest
step of nitrobenzene!phenylhydroxylamine is con-
sidered as the rate-determining step and DFT predict-
ed an activation barrier (16.9 kcalmolꢀ1), which is in
good agreement with the experimental value of
15 kcalmolꢀ1. Therefore, in an effort to understand
the reaction mechanism under the studied reaction
conditions, experimental obsevations coupled with
theoretical calculations provide ample evidence in
favour of the reaction path of aniline formation via
phenylhydroxylamine.
solution of Pd salt in a mixture of surfactant and silica pre-
cursor. Briefly, the sources of Si and Pd are tetraethyl ortho-
silicate (95%, TEOS) (Wako, Japan) and PdCl2 (98.5%, Al-
drich), respectively. Cetyltrimethylammonium bromide
(99%, Merck) was used as template to obtain the mesopo-
rous structure. Under stirring conditions, the Pd salt solution
followed by TEOS (5.0 g) were added to the starting solu-
tion containing template (2.39 g) and sodium hydroxide
(0.33 g) in deionized water (27.0 g). The stirring was further
continued for 2 h. Finally, the resultant gel was autoclaved
at 1408C for 48 h. The molar composition of the gel was:
1 TEOS: 0.27 CTABr: 0.34 Na2O: 62.5 H2O. The resultant
material was filtered, dried and then calcined at 5508C for
8 h in air. The total amount of metal loading was ~1%. All
the catalyst was primarily characterized by X-ray diffraction
(XRD), transmission electron microscopy (TEM) before
and after the reaction (Supporting Information; Figure S1
and Figure S2).
Conclusions
In this work, we have reported on the hydrogenation
of nitrobenzene to aniline over Pd/MCM-41 in scCO2.
It is demonstrated that application of scCO2 leads to
a remarkable change in the reaction rate, which
would be impossible in classical organic solvents.
From a particular point of view, scCO2 is an essential
medium because it enhances the catalytic activity sub-
stantially by lowering the mass transfer limitations be-
tween solid and gas phase. Furthermore, the reaction
Catalytic Activity
In a typical experiment, a 50-mL batch reactor containing
catalyst and the substrate was placed in an oven with fan
heater to maintain the constant temperature and the details
are described elsewhere.[18] The reactor was pressurized with
2.5 MPa of hydrogen followed by the addition of CO2 with
is free from any by-product formation and long-term desired pressure using
a high pressure liquid pump
(JASCO). The reaction mixture was stirred continuously
with Teflon coated magnetic bar during the reaction. After
the reaction, the reactor was cooled with ice/water and dep-
ressurized carefully by the back-pressure regulator. Finally,
the liquid product was separated from the catalyst simply by
filtration and identified by NMR (see Supporting Informa-
tion; Figure S4) GC/MS and analyzed quantitatively by GC
(HP 6890) equipped with flame ionization detector. Quan-
tification of the products was obtained by a multipoint cali-
bration curve for each product. For all results reported, the
catalytic activity can be achieved. Alterations of the
process variables like CO2, H2 pressure, temperature,
reaction time and substrate concentration serve to
affect the conversion of nitrobenzene but the selectiv-
ity to aniline remains same. Most strikingly, proper
management of the hydrogen pressure would be
a good choice to prevent the formation of by-prod-
ucts.
Theoretical calculations along with experimental
findings provided useful information about possible selectivty was calculated as follows:
interactions of substrates and products with the cata-
lyst in scCO2 medium and allowed a rationalization of
the reaction path. A very remarkable observation is
8
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 0000, 000, 0 – 0
ÝÝ
These are not the final page numbers!

Rapid Hydrogenation of Aromatic Nitro Compounds in Supercritical Carbon Dioxide
termediate and compare the stability of the intermediate.
Phase Behaviour
Once we see the order of the individual reactant product
and intermediate we may think about the adsorption energy
preference of those components over the Pd surface, the ap-
proximation calculation is performed with Pd atoms not on
a specific surface to compensate the calculation time and ac-
curacy.
Visual observation of the phase behaviour of nitrobenzene
under the studied reaction condition was made in a 10-mL
high pressure view cell fitted with sapphire windows. The
cell was placed over a magnetic stirrer for stirring the con-
tent and connected to a pressure controller, to regulate the
pressure inside the view cell. In addition to that a tempera-
ture controller was also used to maintain the desired tem-
perature of 508C. Nitrobenzene was introduced into the
view cell at a constant hydrogen pressure of 2.5 MPa while
the CO2 pressure was varied in the range of 7–14 MPa and
the system was monitored.
References
[1] Kirk-Othmer, Encyclopedia of Chemical Technology,
4th edn., Wiley-VCH, New York, 1997.
[2] a) K. Weissermel, H-J. Arpe, Industrial Organic
Chemistry, 3rd edn., Wiley-VCH, Weinheim, 1997,
pp 373–377; b) S. G. Diao, W. Z. Qian, G. H. Luo, F.
Wei, Y. Wang, Appl. Catal. A General 2005, 286, 30–35;
c) J. Wang, Z. Yuan, R. Nie, Z. Hou, X. Zheng, Ind.
Eng. Chem. Res. 2010, 49, 4664–4669.
Computational Details
All the calculations related to the transition state were per-
formed with density functional theory (DFT)[40] using
DMol3[41,42] code of Accelrys Inc., BLYP exchange correla-
tion functional[43,44] and DNP basis set.[45] The transition
state (TS) calculations were performed using the synchro-
nous transit methods as included in the DMol3 module of
Accelrys Inc. Complete linear synchronous transit (LST)/
quadratic synchronous transit (QST) begins by performing
an LST/optimization calculation. The TS approximation ob-
tained in that way is used to perform QST maximization.
From that point, another conjugate gradient minimization is
performed. The cycle is repeated until a stationary point is
located or the number of allowed QST steps is exhausted.
DMol3 uses the nudged elastic band (NEB) method for min-
imum energy path calculations. The NEB method introduces
a fictitious spring force that connects neighbouring points
on the path to ensure continuity of the path and projection
of the force so that the system converges to the minimum
energy path (MEP), which is as well called the intrinsic re-
action co-ordinate if the co-ordinate system is mass weight-
ed. We also calculated the vibration mode to identify the
negative frequency to confirm the transition state. In order
to consider the effect of solvent, we employed the conduc-
tor-like screening model (COSMO) salvation method within
the DFT formalism as in the program DMol3 of Accelrys
Inc.[46] In this method, the solute molecules form a cavity
within the dielectric continuum of permittivity e that repre-
sents the solvent. The charge distribution of the solute polar-
izes the dielectric medium. The response of the dielectric
medium is described by the generation of screening (or po-
larization) charges on the cavity surface. The cavity surface
is obtained as a superimposition of spheres centered at the
atoms, discarding all parts lying on the interior part of the
surface. The spheres are represented by a discrete set of
points, the so-called basic points; eliminating the parts of
the spheres that lie within the interior part of the molecule
thus amounts to eliminate the basic grid points that lie in
the interior of the molecule. The radii of the spheres are de-
termined as the sum of the van der Waals radii of the atoms
of the molecule and of the probe radius. The surviving basic
grid points are then scaled to lie on the surface generated
by the spheres of van der Waals radii alone. The basic
points are then collected into segments, which are also rep-
resented as discrete points on the surface. The screening
charges are located at the segment points.
[3] S. Xu, X. Xi, J. Shi, S. Cao, J. Mol. Catal. A: Chemical
2000, 160, 287–292.
[4] V. R. Choudhury, M. G. Sane, J. Chem. Technol. Bio-
technol. 1998, 73, 336–340.
[5] B. Li, Z. Xu, J. Am. Chem. Soc. 2009, 131, 16380–
16382.
[6] P. Baumeister, H.-U. Blaser, M. Studer, Catal. Lett.
1997, 49, 219–222.
[7] U. Siegrist, P. Baumeister, H.-U. Blaser, Chem. Ind.
1998, 95, 207–219.
[8] Z. Sun, Y. Zhao, Y. Xie, R. Tao, H. Zhang, C. Huang,
Z. Liu, Green Chem. 2010, 12, 1007–1011.
[9] P. Serp, M. Corrias, P. Kalck, Appl. Catal. A General
2003, 253, 337–358.
[10] a) H. Liu, G. Lu, Y. Guo, Y. Wang, Y. Guo, J. Colloid
Interface Sci. 2010, 346, 486–493; b) R. Nie, J. Wang, L.
Wang, Y. Qin, P. Chen, Z. Hou, Carbon, 2012, 50, 586–
596; c) K. Xu, Y. Zhang, X. Chen, L. Huang, R. Zhang,
J. Huang, Adv. Synth. Catal. 2011, 353, 1260–1264; d) J.
Li, X.-Y. Shi, Y.-Y. Bi, J.-F. Wei, Z.-G. Chen, ACS
Catal. 2011, 1, 657–664.
[11] a) Green Chemistry using Liquid and Supercritical
Carbon Dioxide, (Eds.: J. M. DeSimone, W. Tumas,),
Oxford University Press, Oxford, 2003; b) Modern Sol-
vents in Organic Synthesis, (Ed.: P. Knochel), Springer,
Berlin, Heidelberg, 1999.
[12] A. Baiker, Chem. Rev. 1999, 99, 453–477.
[13] A. Fꢁrstner, L. Ackermann, K. Beck, H. Hori, D.
Koch, K. Langeman, M. Liebl, C. Six, W. Leitner, J.
Am. Chem. Soc. 2001, 123, 9000–9006.
[14] K. Wittmann, W. Wisniewski, R. Mynott, W. Leitner,
C. L. Kranemann, T. Rische, P. Eilbracht, S. Kluwer,
J. M. Ernsting, C. J. Elsevier, Chem. Eur. J. 2001, 7,
4584–4586.
[15] P. Stephenson, B. Kondor, P. Licence, K. Scovell, S. K.
Ross, M. Poliakoff, Adv. Synth. Catal. 2006, 348, 1605–
1610.
[16] a) J. R. Hyde, M. Poliakoff, Chem. Commun. 2004,
1482–1483; b) J. R. Hyde, B. Walsh, P. Singh, M. Poliak-
off, Green Chem. 2005, 7, 357–361.
[17] A. Fꢁrstner, D. Koch, K. Langemann, W. Leitner, C.
Six, Angew. Chem. 1997, 106, 2562–2565; Angew.
Chem. Int. Ed. Engl. 1997, 36, 2466-2469.
We first optimize all the reactants and products in the
scCO2 medium and then run a TS calculation to find the in-
Adv. Synth. Catal. 0000, 000, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
9
ÞÞ
These are not the final page numbers!

Maya Chatterjee et al.
FULL PAPERS
[18] a) M. Chatterjee, A. Chatterjee, Y. Ikushima, Green
Chem. 2004, 6, 114–118; b) M. Chatterjee, Y. Ikushima,
T. Yokoyama, M. Sato, Adv. Synth. Catal. 2008, 350,
624–632.
[19] M. G. Hitzler, M. Poliakoff, Chem. Commun. 1997
1667–1668.
[30] F. Figueras, B. Coq, J. Mol. Catal. A: Chemical 2001,
173, 223–230.
[31] U. Birkenstock, B. Lachmann, J. Metten, H. Schmidt,
U.S. Patent 4,265,834, 1981.
[32] a) X. S. Zhao, Q. Ma, G. Q. Lu, Energy Fuels 1998, 12,
7266–7269; b) R. Glꢂser, R. Roesky, T. Boger, G. Ei-
genberger, S. Ernst, J. Weitkamp, Stud. Surf. Sci. Catal.
1997, 105, 695–702.
[20] R. Tschan, M. Wandeler, M. Schneider, M. Schbert, A.
Baiker, J. Catal. 2001, 204, 218–229.
[33] P. N. Raylander, in: Catalytic Hydrogenation in Organic
Synthesis, Academic Press, New York, 1979.
[34] J. M. Chapuzet, A. Lasia, J. Lessard, in: Electrocataly-
sis, 1st edn., Vol. 2, (Eds.: J. Lipkowski, P. N. Ross,),
Wiley-VCH, New York, 1998, pp 10–20.
[35] A. Metcalfe, M. N. Rowden, J. Catal. 1971, 22, 20–34.
[36] J. Wienaik, M. Klein, Ind. Eng. Chem. Prod. Res. Dev.
1984, 23, 44–50.
[21] F. Zhao, R. Zhang, M. Chatterjee, Y. Ikushima, M.
Arai, Adv. Synth. Catal. 2004, 346, 661–668.
[22] S. Ichikawa, M. Tada, Y. Iwasawa, T. Ikariya, Chem.
Commun. 2005, 924–926.
[23] a) F. Zhao, Y. Ikushima, M. Arai, J. Catal. 2004, 224,
479–483; b) F. Zhao, S. Fujita, J. Sun, Y. Ikushima, M.
Arai, Catal. Today 2004, 98, 523–528.
[24] X. Meng, H. Cheng, Y. Akiyama, Y. Hao, W. Qiao, Y.
Yu, F. Zhao, S. Fujita, M. Arai, J. Catal. 2009, 264, 1–
10.
[25] E. A. Gelder, S. D. Jackson, C. M. Lok, Chem.
Commun. 2005, 522–524.
[37] I. L. Obraztsova, N. K. Eremenko, Yu. N. Velyakine,
Kinet. Catal. 2008, 49, 401–406.
[38] F. Haber, Z. Elektrochem. 1898, 22, 506.
[39] E. Kiemm, B. Amon, H. Redingshofer, E. Dieterich, G.
Emig, Chem. Eng. Sci. 2001, 58, 1347–1353.
[26] I. A. Makaryan, V. I. Savchenko, Stud. Surf. Sci. Catal.
[40] W. Kohn, L. J. Sham, Phys. Rev.
A 1965, 140,
1993, 75, 2439–2442.
[27] M. Studer, S. Neto, H.-U. Blaser, Top Catal. 2000, 13,
205–212.
[28] A. Corma, P. Concepcion, P. Serena, Angew. Chem.
2007, 119, 7404–7407; Angew. Chem. Int. Ed. 2007, 46,
7266–7269.
[29] L. Pernoud, J. P. Candy, B. Didillon, R. Jacquot, J. M.
Basset, Stud. Surf. Sci. Catal. 2000, 130, 2057–2062.
A1133 A1138 .
[41] B. Delley, J. Chem. Phys. 1990, 92, 508–517.
[42] B. Delley, J. Chem. Phys. 1991, 94, 7245–7250.
[43] B. Delley, J. Chem. Phys. 2000, 113, 7756–7764.
[44] B. Delley, J. Phys. Chem. 1996, 100, 6107–6110.
[45] A. D. Beceke, J. Chem. Phys. 1988, 88, 2547–2553.
[46] A. Klamt, G. Schꢁrmann, J. Chem. Soc. Perkin Trans. 2
1993, 798–805.
10
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 0000, 000, 0 – 0
ÝÝ
These are not the final page numbers!

FULL PAPERS
Rapid Hydrogenation of Aromatic Nitro Compounds in
Supercritical Carbon Dioxide: Mechanistic Implications via
Experimental and Theoretical Investigations
11
Adv. Synth. Catal. 2012, 354, 1 – 11
Maya Chatterjee,* Abhijit Chatterjee, Hajime Kawanami,*
Takayuki Ishizaka, Toshishige Suzuki, Akira Suzuki
Adv. Synth. Catal. 0000, 000, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
11
ÞÞ
These are not the final page numbers!