Pd/C Catalyzed selective hydrogenation of nitrobenzene to cyclohexanone oxime in the presence of NH2OH·HCl: Influence of the operative variables and insights on the reaction mechanism

Article abstract of DOI:10.1016/j.apcata.2020.117570

We studied the influence of temperature, solvent, pressure, catalysts type on the selectivity of nitrobenzene hydrogenation to cyclohexanone oxime (COX) in the presence of NH₂OH. The best reaction conditions are: pressure 0.8 MPa, temperature 333 K, solvent ethers, and catalyst Pd/C5%. Other hydrogenation metal catalysts did not give comparable results. The amount of Pd/C influences the yield in COX, which rises above to 90 % at the highest load. The reaction profile shows that aniline is the reaction intermediate. Indeed, aniline as a substrate gives COX, though in lower yield than that achieved employing nitrobenzene. The NH₂OH parallel hydrogenation to NH₄Cl, influences positively the selectivity to COX. It has been observed that COX, cyclohexanone and N-cyclohexylideneaniline are in equilibrium in the reaction solution and all likely derive from nucleophilic substitutions to a common imine intermediate formed on the Pd surface, whose high activity does not need any further metal catalyst.

Full text of DOI:10.1016/j.apcata.2020.117570

Journal Pre-proof
Pd/C Catalyzed selective hydrogenation of nitrobenzene to
cyclohexanone oxime in the presence of NH2OH·HCl: influence of the
operative variables and insights on the reaction mechanism
L. Pietrobon, L. Ronchin, C. Sadraoui, R. Pontello, C. Tosetto, A.
Vavasori
PII:
S0926-860X(20)30163-0
DOI:
https://doi.org/10.1016/j.apcata.2020.117570
APCATA 117570
Reference:
To appear in:
Applied Catalysis A, General
Received Date:
Revised Date:
Accepted Date:
11 March 2020
6 April 2020
14 April 2020
Please cite this article as: Pietrobon L, Ronchin L, Sadraoui C, Pontello R, Tosetto C, Vavasori
A, Pd/C Catalyzed selective hydrogenation of nitrobenzene to cyclohexanone oxime in the
presence of NH2OH·HCl: influence of the operative variables and insights on the reaction
mechanism, Applied Catalysis A, General (2020),
doi: https://doi.org/10.1016/j.apcata.2020.117570
This is a PDF file of an article that has undergone enhancements after acceptance, such as
the addition of a cover page and metadata, and formatting for readability, but it is not yet the
definitive version of record. This version will undergo additional copyediting, typesetting and
review before it is published in its final form, but we are providing this version to give early
visibility of the article. Please note that, during the production process, errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal
pertain.
© 2020 Published by Elsevier.

1
Pd/C Catalyzed selective hydrogenation of nitrobenzene to cyclohexanone oxime in the presence
of NH2OH.HCl: influence of the operative variables and insights on the reaction mechanism
L. Pietrobon, L. Ronchin* ronchin@unive.it, C. Sadraoui, R. Pontello, C. Tosetto, A. Vavasori
Department of Molecular Sciences and Nanosystems, University Ca’Foscari of Venice, via Torino155
30170, Mestre Venezia Italy.
To whom correspondence should be addressed.
Graphical abstract
Consecutive, parallel reactions and equilibria affecting the selectivity in cyclohexanone oxime

2
Highlights
 Nitrobenzene hydrogenation catalyzed by Pd/C in the presence of NH2OH HCl forms Aniline
and cyclohexanone oxime.
 Catalysts based on different metal are not selective to the cyclohexanone oxime.
 The best solvent are ethers followed by dichloromethane and dimethylcarbonate
 NH2OH HCl hydrogenates to NH4Cl in the presence of Pd/C as a catalyst.
 Aniline in the presence of NH2OH HCl gives cyclohexanone oxime and the presence NH4Cl
favors the selectivity to cyclohexanone oxime.
 Cyclohexanone, cyclohexanone oxime and N-cyclohexylideneaniline are in equilibrium and
derive from a common imine intermediate on catalyst surface.
Abstract
We studied the influence of temperature, solvent, pressure, catalysts type on the selectivity of
nitrobenzene hydrogenation to cyclohexanone oxime (COX) in the presence of NH2OH. The best
reaction conditions are: pressure 0.8 MPa, temperature 333K, solvent ethers, and catalyst Pd/C5%.
Other hydrogenation metal catalysts did not give comparable results. The amount of Pd/C influences
the yield in COX, which rises above to 90% at the highest load. The reaction profile shows that aniline
is the reaction intermediate. Indeed, aniline as a substrate gives COX, though in lower yield than that
achieved employing nitrobenzene. The NH2OH parallel hydrogenation to NH4Cl, influences positively
the selectivity to COX. It has been observed that COX, cyclohexanone and N-cyclohexylideneaniline
are in equilibrium in the reaction solution and all likely derive from nucleophilic substitutions to a
common imine intermediate formed on the Pd surface, whose high activity does not need any further
metal catalyst.
Keywords: Nitrobenzene hydrogenation, Aniline Hydrogenation, NH2OH hydrogenation,
cyclohexanone oxime, Pd catalysts.
1. Introduction

3
Nitrobenzene and in general nitro-aromatic compounds reductions are really important and useful
reactions for several industrial processes [1]. Béchamps-Brimmeyr reductions method with Fe-FeCl2
system is still used mainly for the formation of the Black Fe3O4 inorganic pigment as by-product of
reduction of the nitrocompound [2]. Other stoichiometric reduction methods of nitrocompounds
employ Zn or hydrazine for achieving high selectivity with sensitive substituents or particular synthesis
two-step one pot synthesis [1, 3].
Hydrogenation, however, is actually the most important and more used method of reduction in aromatic
nitro-compounds [1, 4]. Among them, nitrobenzene hydrogenation to aniline is one of the most
important reaction, since the latter is used as an intermediate in many fields of application, for instance,
in the production of isocyanates, rubber processing chemicals, dyes and pigments, agricultural
chemicals and pharmaceuticals [1]. Industrial hydrogenation methods employ several type of catalysts
and engineering solutions and each one depends on its specific use [5]. For instance, supported Pt, Ni,
Pd are commonly employed as catalysts in gas/solid or liquid/solid multiphase reactors, as well as in
continuous or in batch operation [6].
The reduction of the nitro-group involves several steps following the well-known Haber reaction path,
where nitro-group is reduced step by step to the amine together with the formation of condensation
products between nitrosobenzene and phenyl hydroxyl amine (see Scheme 1) [7].

4
Scheme 1. Haber and Jackson Nitrobenzene reduction/hydrogenation mechanism.
Recently, for the catalytic hydrogenation mechanism, it has been suggested a similar path, but the
formation of nitrosobenzene has been related to a surface equilibrium phenyl-
hydroxylamine/nitrosobenzene rather than the formation of an actual reaction intermediate (see Scheme
1 Jackson route). Condensation of nitrosobenzene and phenyl hydroxyl amine occurs in the same way
of the Haber scheme [8, 9]. This has been clearly pointed out by the study of Jackson and co-workers
on the kinetics of the nitrosobenzene hydrogenation, they demonstrated indeed a different reaction path
for nitrosobenzene hydrogenation with respect to that of nitrobenzene [8].
Among the one-pot multi-steps processes the selective hydrogenation of nitrobenzene to 4-
aminophenol is an elegant example of an industrial application, which combine a noble metal catalyzed
reaction with a homogeneous acid catalysis [9-13]. Industrially, the reaction is carried out in continuous
stirred tank reactor in a biphasic reaction medium, in the presence of aqueous sulfuric acid and a
supported Pt catalyst. At first, nitrobenzene is hydrogenated to N-phenylhydroxylamine, and then the

5
acid catalyses the Bamberger rearrangement of the intermediate [3, 9-13]. From an environmental point
of view, the major drawback of the process is the neutralization of the sulphuric acidic phase, which
causes huge corrosion concern, with the consequent increased costs [3, 9-13]. However, the use a single
reactor system allows an increase of the sustainability of the process compared to other multistep ones
[9].
Besides, Corma and Co-workers studied the selectivity of aromatic nitrocompounds reductions and
they direct selectivity of Au/supported catalyst towards azobenzene. They recognized responsible for
such a selectivity the stabilization of surface nitrosobenzene-Au species [14].
The complexity of the reaction scheme of nitrocompounds reduction allow to take into consideration
(see Scheme 1) the possibility of designing catalyst or processes that gives, under particular
experimental conditions, quite selectively a particular intermediate [9]. For instance, formation of azo-
compounds can be achieved by combining the partial reduction of nitrobenzene condensation and
dehydration in basic media [15]. Other interesting reduction methods for nitro-aromatic compounds are
hydrogen transfer reactions [17-19]. For instance nitrobenzene can be easily reduced to azo, azoxy,
anilines, carbamates and ureas by using CO as reductant and Pd(II) complexes as catalysts, the
operating conditions will determine the selectivity to each product [17, 18]. Under similar reaction
conditions in the presence of mild oxidant, nitrobenzene is, at first glance, reduced to an intermediate,
which gives aniline in protic solvent, while in aprotic one the main products is isocyanate [17, 18]. The
reaction gives almost selectively diphenyl urea in the presence of Pd supported catalysts in liquid
biphasic conditions and aqueous NaOCHO as reductant, [20].
The further reduction of the aromatic amines and of aniline in particular, is relatively less studied
compared to the hydrogenation of the nitro-aromatic compounds [21-33]. Hydrogenation of aniline to
the cyclohexylamine, particularly, has been investigated from long time, and the best catalytic system is
mostly Ru supported catalysts at quite high temperature and pressure (373-473K, 1-10 MPa) [22-26].
Even though, less studied than that of nitrobenzene reduction, the mechanism of this reaction has been
already proposed by Greenfield in the early sixty. In this paper, the author has assumed the formation of
the imine (Scheme 2) as the key intermediate, which, due to its high reactivity, undergoes facile
condensation reactions with all the nucleophiles present in solution, as showed in Scheme 2 [21]. The
reactivity of such an intermediate drives the selectivity toward the formation of cyclohexyl amine, di-
cyclohexyl amine, and, in the presence of water, to cyclohexanone and cyclohexanol [21, 28-31]. The
highly reactive imine intermediate has been never isolated in the hydrogenation medium, however, the
final products observed strongly supports the goodness of such a hypothesis [21, 28-31]. Greenfield

6
suggested also that the imine specie should be only stabilized in the adsorbed state on catalyst surface,
and a direct coupling deamination of CHA can occur only at temperature higher than 200°C on Ni
catalysts [21].
Scheme 2. Reaction mechanism of aniline hydrogenation suggested by Greenfield [21].
Cho and Park have observed a strong influence of the fluid dynamic behavior of the reactor on the
selectivity of the aniline hydrogenation to cyclohexyl amine [31]. The authors suggested a non-
stationary concentration of hydrogen in the reactor, which cause different rates of reactions in the
various steps suggested in the Greenfield mechanism [31]. This mechanism suggests of using the
reactivity of the imine intermediate towards nucleophiles; for instance, the substitution with water gives
cyclohexanone, a fundamental intermediate, in the production of nylon-6 and in several fine chemistry
synthesis [18, 31-33]. In fact recently, appears some patents claiming a direct route from nitrobenzene
to cyclohexanone oxime by hydrogenation and direct oximation [34, 35]. In particular, Corma and
coworkers have carried out the hydrogenation of nitrobenzene to cyclohexanone oxime in the presence
of hydroxyl amine hydrochloride, in multiphase systems, catalyzed by Pd and Au supported catalysts
[35-36]. These authors have claimed that the presence of the two type of catalyst allows high selectivity

7
in the cyclohexanone oxime. The highest selectivity towards the oxime is achieved in the presence of a
physical mixture of the two catalysts with respect to the single catalytic system Pd/C or Au/C,
moreover, the latter leads to aniline only. Finally, the bimetallic Pd-Au/C catalyst is less selective than
the physical mixture of the two catalysts, as well [36]. The mechanism, suggested by the Authors (see
Scheme 3), is based on the hypothesis that cyclohexyl amine is in dehydrogenative equilibrium with the
cyclohexyl imine, which undergoes to the nucleophilic attack of the aniline, thus forming N-
cyclohexylideneaniline. In turn, the latter may undergo nucleophilic attack of the hydroxylamine, or the
hydrogenation to phenyl cyclohexyl amine and finally to dicyclohexyl amine. From a mechanistic point
of view, the Authors have suggested the involvement of the gold catalyst in the transformation the N-
cyclohexylideneaniline, which determines the overall selectivity to the various products [36]. In fact,
the nucleophilic attack of the NH2OH catalyzed by Au/C catalyst to the N-cyclohexylideneaniline lead
to cyclohexanone oxime and finally its hydrolysis may give cyclohexanone if water is present [36].
Besides, Pd/C is responsible for the formation of the di-cyclohexyl amine for the sequential
hydrogenation of N-cyclohexylideneaniline to N-cyclohexyl aniline, which is finally hydrogenated to
the di-alkyl amine [36]. In any case, in agreement with the mechanism suggested by Greenfield, the
formation of the imine intermediate is, also in this case, an important step, since its reactivity lead to
the formation of the N-cyclohexylideneaniline recognized by these Authors as the actual intermediate,
whose reactivity catalyzed by Au determines the high selectivity in the cyclohexanone oxime.
Scheme. 3. Mechanism of Nitrobenzene hydrogenation to cyclohexanone oxime suggested by Corma
[36].

8
In this work, we have studied the selective hydrogenation of nitrobenzene to cyclohexanone oxime
with different solvents, in the presence of various hydroxylammonium salts, and catalyzed by a
commercial Pd/C catalyst. The research will focus on the influence of the operative variables
(Temperature pressure and solvent type), of the amount and type of catalyst on the selectivity to the
oxime and to the various byproducts. Besides, a particular attention will be reserved to the study of the
likely intermediates found in traces amount in the reaction media as well as the parallel reaction of
NH2OH.HCl reduction whose products favors the selectivity in the cyclohexanone oxime.
2. Experimental
2.1. Materials
Nitrobenzene, aniline, cyclohexylamine, dicyclohexylamine and N-phenylcyclohexylamine (PCA)
were all Aldrich products, their purity were checked by the usual methods (melting point, TLC, HPLC,
GC and GC-MS) and employed without any purification, acetonitrile HPLC gradient grade was
supplied by BDH, 1,4-dioxane, methanol, nitromethane, diethyl ether, dioxane, tetrahydrofuran,
dimethylformamide and dimethyl sulfoxide are ACS reagent supplied by Aldrich. N-
cyclohexylideneaniline (PCNA) was synthesized by following the procedure of literature and the
characterization was carried out by GC-MS, 1H and 13C NMR [37]. Catalysts used were commercial
materials supplied by Engelhard (now Basf Catalysts): Pd/C 5% Escat 10, Pt/C 3%, Ru/C 5% Escat 40
and Rh/C 5%. Au/SBA 2% and Au/C 0.5% were gently furnished by Proff. M. Signoretto and F.
Menegazzo; synthesis and characterization of these catalysts has been reported elsewhere [38, 39].
2.2. Equipment
Products were identified by gas chromatography (GC) gas chromatography coupled mass spectrometry
(GC-MS) and high performance liquid chromatography (HPLC). GC and GC-MS analysis were carried
out with an Agilent 7890A equipped with FID or MS detector (Agilent 5975C and a HP 5 column (I.D.
320 m 30 m long). Helium was employed as carrier in the GC-MS analysis, while nitrogen in the GC
ones, in both cases the conditions were the following: injector 523 K, detector 543 K, flow 1 mL min-1,
oven 333 K for 3 minutes 523 K 15 K minute-1 and 523K for 15 minutes. In order to check the thermal
stability of the products, some analysis were carried out by HPLC (Perkin Elmer 250 pump, LC 235
diode array detector and a C 8, 5 m, 4 mm i.d. 25 cm long column) analysis were carried out with
CH3CN-H2O as mobile phase in isocratic 70% of CH3CN at 1 mL min-1. The comparison of GC and

9
HPLC analysis shows a good agreement suggesting both technique gives reliable analytical data. After
the test of the analytical reliability, conversion yield and selectivity are calculated by calibrations of the
1
GC analysis obtained by standard solutions of the pure products. H and 13C NMR. were carried out in
a Brucker Avance 400 II at 400MHz and 100MHz, respectively and using CDCl3 as a solvent.
2.3. Catalyst characterization
Catalysts employed are commercial catalysts (except Au ones) whose characterization has been already
described in previous papers [38-40].
2.4. Hydrogenation of nitrobenzene
The reactions have been carried out in several solvents (1,4-dioxane CH3NO2, CH3OH, (CH3)2NCHO
and (CH3)2SO) in an AISI 316 autoclave (80 mL internal volume) equipped with a PTFE insertion (in
order to avoid metal contamination), and a pressure regulator to maintain constant the reaction
pressure. A circulation bath ensure a stable temperature in the range 323-353 K (± 0.5 K). Catalyst
activation were preliminarily carried into the autoclave at the running temperature and pressure (0.3-2
MPa of H2), without reagents. Blank reactions in the absence of nitrobenzene and hydroxylamine have
been carried out in order to verify the stability of the solvent. Solvents employed showed no relevant
products of reduction at GC analysis. A detailed description of hydrogenation apparatus has been
described elsewhere [41]. In a typical experiment 20 mL of solvent with suspended the Pd/C catalyst
(typically 50 mg) and NH2OH.HCl (2.7 g), was introduced into the reactor. After that, the reactor was
closed, purged before with nitrogen and then with hydrogen, pressurized at 0.2 MPa, finally heated at
323-353 K under stirring for 1 h in order to activate the catalyst. The reactor was left to stabilize at the
temperature of reaction (323-353K), than the preheated solution of nitrobenzene (c.a. 2.5 mL,
concentration of nitrobenzene 4 mol L-1) was added to the reactor. After the addition of this solution,
the computing of the reaction time starts, and the liquid phase analysis is by GC after 20 h of reaction.
An electronic pressure transducer monitors hydrogen consumption, which measure the pressure drop in
a reservoir at known volume.
The reaction environment, at initial stage of the reaction, is a complex mixture composed of a gas
phase (mainly hydrogen and solvent), a liquid phase (solvent and nitrobenzene) and two solid phases:
hydroxylamine hydrochloride and the catalyst. Both solid are not soluble (negligibly soluble) in the
solvent, however, during the course of reaction the formation of water, due to nitrobenzene

10
hydrogenation to aniline, may cause an increase of the solubility of the hydroxyl ammonium salt, or of
other salts present and/or produced in the reaction vessel.
In all the experiments at the opening of the reactor, we found a solid mixture crust on the bottom of the
reactor made of the catalyst and various salts, strong agitation needs for restoring the homogeneity, it is
necessary to verify carefully the agitation of the system in order to avoid reproducibility problems.
Some kinetic profiles has been obtained by using a 250 mL autoclave employing 80 mL of volume
under the same concentration of reagent and at 0.8 MPa of hydrogen pressure the apparatus has been
detailed descripted elsewhere [41].
2.5. Determination of residual NH2OH.HCl
The residual hydroxyl ammonium salt employed in the reaction has been quantified after filtration and
washing of the solid with diethyl ether. Then the salt is dissolved in water diluted, added to an almost
equimolar solution of ferric sulfate, left to react for 30 minutes, and the excess of Fe (III) titrated
directly with a solution of KMnO4 (0.05 mol L-1) [42]. The method has been checked by ionic
chromatographic analysis, showing equivalent results. In some sample a residual percentage of CHA
has been observed after extraction with CH2CL2 of the KOH basified solution to pH>12, in any case,
the amount is generally lower than 1%, except in those experiments where the CHA were added in high
concentration as a reagent.
3. Results and discussion
In this study, we will propose a first attempt to analyze a complex multistep multiphase reaction
starting from the choice of the catalytic system, in order to verify what metals are able to hydrogenate
nitrobenzene (NB) to cyclohexanone oxime (COX) in the presence of NH2OH salts, selectively.
Actually, the reaction presents a recognized intermediate, which is aniline (AN), and some compounds,
recognized in the reaction media, which are candidates of being intermediate or byproducts [3, 36].
These compounds are cyclohexanone (CON), cyclohexylamine (CHA), and PCA, traces of PCNA,
N,N-dicyclohexyl amine and finally as byproduct of nitro-group reduction H2O (two moles per mole of
substrate).
3.1. Choice of the catalytic system
The main results of some supported precious metal hydrogenation catalysts employed under mild
conditions (333K 0.8 MPa), are reported in Tab. 1. Actually, the presence of the hydroxylamine

11
hydrochloride is necessary for achieving a selectivity to the COX, the absence of hydroxyl amine
means no selectivity to oxime (see supplementary materials). It is immediately evident the higher
selectivity of the Pd/C catalyst compared with the other ones. Furthermore, it is noteworthy the lower
selectivity of the Pd/Al2O3 catalysts, compared to that obtained by the Pd/C. Besides, other metals,
under the condition investigated, are low active in the nitrobenzene hydrogenation, for instance, in the
case of the Ru or Au catalysts show after 20 h of reaction 5% and 1% of nitrobenzene conversion,
respectively. The reason of this behavior is likely related to the specific activity of each catalyst,
however, the high activity of Pd/C in nitrobenzene hydrogenation is well known from the beginning of
the studies in catalysis, while other metal are active only in more severe conditions [21-27]. As regard
the selectivity in COX, only Pd shows a clear selectivity to this compound. This is related to its high
activity in the hydrogenation reactions, since the formation of COX must proceed via the reduction step
of aniline which is indeed a harsh substrate to be reduced [21-27].
Table. 1 Influence of the catalyst on the conversion and selectivity of nitrobenzene hydrogenation. Run
conditions: reaction volume 20 mL, nitrobenzene 0.5 mol L-1, hydrogen Pressure 0.8 MPa, NH2OH.HCl
30 mmol, T 333K, catalyst 50 mg, time of reaction 20h, solvent diethyl ether.
Catalyst
Conv.a Select.a Select.a Select.a Select.a Select.a Select.a
(%)
AN
(%)
30
COX
CON
CHA
PCA
Othersb
(%)
(%)
(%)
1
(%)
2
(%)
2
Pd/C (5%)
>99
99
>99
90
1
64
1
>1
/
Pd/Al2O3 (5%)
Pt/C (3%)
70
29
>1
>1
>1
/
>1
>1
>1
/
>1
>1
>1
/
>99
>99
99c
99c
>99
>99
/
/
/
/
/
/
Pt/Al2O3 (5%)
Au/C (0.5 %)
Au/SBA (2 %)
Ru/C (5%)
/
/
1
/
/
/
/
5
/
/
/
/
Rh/C (5%)
10
/
/
/
/
a) Conv. = conversion, Select. = selectivity.
b) GC-MS reveals traces of condensation products and N,N-dicyclohexylamine.
c) Traces.

12
We have also verified if the presence of a mixture of Pd/C and a second metal catalyst may help to
improve the selectivity toward the COX. The results, reported in the supplementary materials (Tab. S2),
suggest that there is no increases in COX selectivity, with a second hydrogenation catalyst. On the
contrary, the addition of a second catalyst diminishes the selectivity in COX, and only the Au based
ones seem to have comparable selectivity in COX with respect to that obtained with the pure Pd/C
catalyst, suggesting that these materials act practically as an inert, under these reaction conditions.
3.2. Influence of solvent type, hydrogen pressure and temperature
The presence of Pd catalysts and hydroxylamine as reagent is thus the key point for obtaining COX
directly from nitrobenzene hydrogenation but the second essential element is the solvent. In fact, the
solvent play a fundamental role for driving selectivity toward COX, this is evident from the results of
Tab.3, in which nitrobenzene is in practical quantitatively hydrogenated in all the solvent, but low or no
selectivity to COX is achieved by using EtOH, acetonitrile (ACN) and toluene (Tol). The highest
selectivity in COX is obtained by using THF as the solvent. The reason is not easy to explain because
of the different nature of these solvents (chemical and/or physical). As can be observed in Tab. 2 there
are no evident relationships between solvent dielectric constant and selectivity, even though it appears
that neither polar solvent nor completely non-polar one seems to favor selectivity to COX. Other
solvent parameters, such as the relative rate of nucleophilic substitution or the solvatochromic
parameters does not suggest any correlation between COX selectivity and the solvent parameter itself
[43]. This suggests the selectivity in COX is the result of a complex process where a cat-walk between
several parallel, consecutive reactions and equilibria determines the selectivity. As a matter of fact,
THF appear to be the best solvent followed by Et2O DCM and DMC. These solvents are in any case no
protic solvent and show an intermediate polarity between Tol and acetone (Act). No evidence of direct
correlation between LogPOW (partition n-octanol/water) of the solvent and their selectivity toward
COX, however the best results are obtained with solvent of intermediate hydrophilicity/hydrophobicity
values (see Tab. 2) [44]. In any case, in this work we used preferentially Et2O, because of its lower
cost.
Table. 2 Influence of the solvent on conversion and selectivity. Run conditions: reaction volume 20 mL,
nitrobenzene 0.5 mol L-1, hydrogen Pressure 0.8 MPa, NH2OH.HCl 30 mmol, T 333K, Pd/C 50 mg,
time of reaction 20 h.

13
Solvent
LogPOW
Conv. Select. Select. Select. Select. Select. Select.

(%)
AN
(%)
52
50
58
45
99
24
98
32
64
73
55
40
COX CON
CA
(%)
/
PCA
Others
(%)
47 a
38 a
41 a
55 a
<1 b
1 b
(%)
<1
22
/
(%)
(%)
/
EtOH
ACN
NMe
Act
-0.24
-0.33
-0.35
-023
2.5
24.6 > 99
35.9 97
/
/
/
/
35.9 >99
20.7 <99
2.38 95
<1
/
/
/
/
/
/
Tol
<1
71
/
<1
<1
/
<1
3
/
THF
THF/H2Oc
0.49
0.1
7.58 > 99
42.8 > 99
4.34 > 99
4.34 > 99
3.81 > 99
3.07 > 99
8.93 > 99
<1
traces traces
1
1 b
1 b
Et2O
0.85
0.85
1.9
64
32
21
40
55
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
2
2
5
3
3
Et2Od
iPr2O
DMC
DCM
0.23
1.3
1 b
1 b
a) Products of aniline alkylation [45].
b) Condensation products and N, N-dicyclohexylamine.
c) THF/H2O = 1/1 wt. [46].
d) In the presence of 200 mg of molecular sieve 4A.
Tab. 3 reports the influence of hydrogen pressure and temperature on conversion and selectivity. It
appears that hydrogen pressure does not influence the conversion after 20h of reaction which is
quantitative, but it influences COX selectivity showing a maximum in the range of 8-10 bar (see). The
reason is not evident because of the increasing of the pressure would suggest the increase of the
hydrogenation rate also of the aniline, which, on the contrary, seems to be poorly affected by the
pressure being CHA selectivity in the range of 0.1-1%. PCA is the only product whose selectivity raises
significantly as hydrogen pressure increases thus suggesting a competition between COX and PCA
formation, even though this could be only a secondary effect due to the involvement in the
hydrogenation of the NH2OH, which could be consumed faster at the higher pressure.
The influence of temperature does not affect the NB conversion, which is always quantitative, but it has
an actual impact on the selectivity in COX. It is clear that the selectivity to COX reach a maximum in
the range of temperature 333-343K (see also supplementary materials Fig. 4S). The hydrogenation at

14
323 K temperature shows a lower selectivity probably because the second step of the reaction, (AN
hydrogenation) that occurs much slowly at lower temperature, and very long time of reaction appears
not of practical interest for this type of processes. At the highest temperature, likely, the reaction is less
selective because of an increase of the rate of side of reactions.
Table. 3 Influence of the hydrogen pressure and temperature on selectivity (quantitative NB
conversion). Run conditions: reaction volume 20 mL, nitrobenzene 0.5 mol L-1, NH2OH.HCl 30 mmol,
T 323-353K, Pd/C 50 mg, time of reaction 20 h.
Entry Pressure Temp. Select. Select. Select. Select. Select. Select.
(Bar)
(K)
AN
(%)
48
COX CON
CHA
(%)
0.1
PCA
(%)
1
Othersa
(%)
<1
(%)
49
(%)
0.1
1
2
5
8
333
323
59
31
27
32
32
52
40
64
67
51
62
53
<1
1
<1
1
<1
2
<1
1
3
4
5
6
7
8
333
343
353
333
333
8
1
1
2
2
8
1
1
3
2
10
20
1
1
2
1
<1
1
3
1
a) Condensation products and N,N-dicyclohexylamine
3.3 Influence of the catalyst amount on conversion and selectivity
The reaction shows that the selectivity in COX grows as the catalyst amount increase (see Fig. 1). The
explanation of this fact is not easy, but it is likely that key intermediate determining the selectivity in
COX is on catalyst surface. The presence of a reactive surface intermediate derived from aniline
hydrogenation explains also the increase of byproducts with catalyst amount, thus suggesting all
intermediates and byproduct derive from aniline hydrogenation. This is clear, since nitrobenzene, under
the reaction conditions, is readily and quantitatively hydrogenated to AN, while the hydrogenation of
the AN ring is not easy. As a matter of fact, It is probably that AN has the tendency of being strongly
adsorbed on catalyst surface, since stable aniline-Pd(II) complexes are well known and characterized
thus justifying strong interaction between the surface Pd atoms and the AN molecule [47].

15
100
80
60
40
20
0
AN
COX
Other
0
20
40
60
140
160
Catalyst amount (mg)
Fig. 1. Influence of the catalyst amount on the yield of aniline and cyclohexanone oxime. Run
conditions: reaction volume 20 mL, nitrobenzene 0.5 mol L-1, Hydrogen Pressure 0.8 MPa,
NH2OH.HCl 30 mmol, T 333K, Pd/C 50 mg, time of reaction 20 h.
3.4 Time vs concentration profiles, hydrogen consumption and mass balance
In this study the experiments after 20 h of reaction were carried out in an 80 mL autoclave, while, the
reaction profile has been obtained by using a 250 mL reactor equipped with capillary for sampling
during the reaction course. The comparison of the conversion and the selectivity of the various products
after 20 h of reaction obtained with the two reaction systems are similar thus suggesting a reliable
comparison between the results of the two reactors. Fig. 2 shows the reaction profile of a typical
experiment. This reaction profile must be considered only an attempt to represent what happen in the
reaction environment, since a detailed kinetic study must take into account of the complex fluid
dynamic aspect of this reaction system, which is beyond the purpose of this work. However, it appears
a fast consumption of NB, during the first hours of reaction. At the end of this period, it begins the
formation of COX and of other byproducts as reported in Fig. 3. It is noteworthy that when
nitrobenzene is consumed, hydrogenation of aniline starts. CHA and CON are less abundant among the
byproducts remaining lower than 1% of yield. Besides PCA is not abundant and we observe traces of
PCNA and N,N-dicyclohexylamine after 20 h of reaction.

16
100
80
60
40
20
0
Nitrobenzene
AN
COX
0
200
400
600
800
1000
1200
1400
Time (minutes)
Fig. 2. Hydrogenation profile. Run conditions: reaction volume 75 mL, nitrobenzene 0.5 mol L-1,
Hydrogen Pressure 0.8 MPa, NH2OH.HCl 75 mmol, T 333K, Pd/C 60 mg, time of reaction 20 h.
The role of these byproducts as intermediates in COX formation could be related by their direct
oximation in the presence of the hydroxylamine. In the mechanism proposed by Corma and coworkers,
a dehydrogenation of CHA to cyclohexanimine occurs, starting from this suggestion, it could be useful
to verify, how reactants and intermediates influence the selectivity in COX. At a first sight, as can be
noticed by the reaction profile of Fig. 2, nitrobenzene looks like not relevant for the selectivity in COX,
because its formation starts when nitrobenzene is practically disappeared. This behavior suggests of
studying the hydrogenation of aniline, in order to verify its role, in the formation of COX, which will
be reported in the next section.

17
2
1
0
PCA
CHA
CON
0
200
400
600
800
1000
1200
1400
Time (minutes)
Fig. 3. Yields of low concentration products. Run conditions: reaction volume 75 mL, nitrobenzene 0.5
mol L-1, Hydrogen Pressure 0.8 MPa, NH2OH.HCl 75 mmol, T 333K, Pd/C 50 mg, time of reaction 20
h.
The measure of hydrogen consumption (Fig. 4) allows to test the mass balance of H2 by verifying if the
trend is compatible with that of Fig. 2, where the consumption of nitrobenzene and the formation of the
products are shown. It is noteworthy, the fast consumption of hydrogen, which follows practically a
straight line, during the first 60-80 minutes. This is in agreement with the first part of the
hydrogenation of nitrobenzene to aniline, whose reaction rate is fast and shows practically a linear
trend. In addition, it is likely, that under this reaction conditions: with high catalyst load, multiphase
system and low hydrogen pressure, diffusion phenomena may partially limits the reaction rate of
nitrobenzene reduction [47]. This is in agreement with that observed by carrying out a reaction with 5
mg of catalyst where the initial hydrogen consumption rate remain practically the same of that
measured with 50mg of catalyst (see supplementary materials Fig. 5S). This behavior can be ascribed
to the complex reaction system, where dissolution adsorption and re-precipitation of salts occur and
where it is likely that diffusion may influence hydrogen adsorption rate [48]. It is noteworthy, that also,
the second part of the hydrogenation of nitrobenzene (see Fig. 6) appears as a straight line but its slope
is lesser than that of the first part (Fig. 5, 0.019 mmol min-1 vs. 0.52 mmol min-1, respectively). Such a
behavior confirms the low activity of the catalyst in the AN hydrogenation compared with that
measured for nitrobenzene. Another important point, is that the overall mass balance with all products
of hydrogenation does not hold, in fact, after 70 minutes 35 mmol of hydrogen (30 mmol of hydrogen

18
is the stoichiometric for nitrobenzene to AN) has been adsorbed, thus suggesting that there are more
reactions consuming hydrogen. At first glance, the solvent reduction could have been a likely
candidate, but the reaction carried out with pure solvent does not shows any hydrogen consumption
(experimental section).
40
Slope: 0.019 mmol min-1
30
20
10
Slope: 0.52 mmol min-1
0
0
50
100
150
200
250
Time (minutes)
Fig. 4. Hydrogen consumption profile of nitrobenzene hydrogenation. Run conditions: reaction volume
20 mL, nitrobenzene 0.5 mol L-1, Hydrogen Pressure 0.8 MPa, NH2OH.HCl 30 mmol, T 333 K, Pd/C
50 mg.
Under such hydrogenating conditions, hydroxylamine hydrochloride, in the presence 1 mol L-1 of water
(that is the water produced during nitrobenzene reduction), gives ammonium salt by following Scheme
4.
Scheme 4
Fig. 5 shows a clear hydrogen consumption, thus suggesting a competitive parallel reduction of this
reagent during nitrobenzene hydrogenation as well as also during the hydrogenation of the aniline and
the formation of COX. The profile of hydrogen consumption is a straight line, and being this reaction
rate 20 times slower than that of the NB hydrogenation, it is likely that diffusion phenomena does not
influence this reaction rate. Besides, this evidence suggests that the reaction rate is not influenced by

19
the hydroxylammonium concentration (it results as a pseudo zero order in the salt concentration) thus
suggesting a strong adsorption on the catalyst surface. It is noteworthy, reactions carried out in the
presence of Au catalysts do not show any hydrogen consumption in the hydrogenation of NH2OH.HCl.
This intrinsic low catalytic activity of gold catalysts in the hydrogenation of this substrate, is, however,
in agreement with the low activity of these catalyst also in the NB hydrogenation (see Tab.1). Taking
into account the presence of precipitation-dissolution equilibria this does not allow any kinetic
consideration on the reaction order because of the complexity of the system. The NH2OH.HCl solution
adsorbs after 4h, 7 mmol of hydrogen, which is in agreement with the conversion of the NH2OH.HCl to
NH4Cl measured by titration of the solution obtained by extraction with water of the inorganic salts
present into the reaction environment. The reaction rate, measured in Fig. 7, is higher than that
observed in the second part of Fig. 6 (0.026 mmol min-1 vs 0.019 mmol min-1), since it is likely that the
presence of the aniline lows down kinetics and thus also that of the NH2OH.HCl reduction.
8
6
4
Slope: 0.026 mmol min-1
2
0
0
50
100
150
200
250
Time (minutes)
Fig. 5. H2 consumption profile for NH2OH hydrogenation. Run conditions: reaction volume 20 mL,
Hydrogen Pressure 0.8 MPa, NH2OH.HCl 30 mmol, H2O 20 mmol, T 333K, Pd/C 50 mg.
3.5. Aniline Hydrogenation to COX
Fig. 6 reports the reaction profile of AN hydrogenation. As expected the initial reaction rate is slower
than that of nitrobenzene, since AN is a harder substrate to hydrogenate [22-26]. However, after 20 h it
proceeds until 47% of conversion, with a high selectivity (COX selectivity > 96%). Despite of the high
selectivity in COX starting from AN, after 20h of reaction the overall yield of COX is lower (45%)
with respect of what found in nitrobenzene hydrogenation (yield 62%), but the different initial
condition and the course of the nitrobenzene hydrogenation may lead to a different reactivity of the

20
system, thus determining a lower yield in COX. Indeed, the conditions of the hydrogenation of pure
aniline are different with respect to what occurs at the end of nitrobenzene consumption during its
reduction, because of the presence byproducts of NB and NH2OH hydrogenation, that is water and
ammonium chloride salt (Scheme 4). When we start from nitrobenzene, the presence of this parallel
reaction may modify both the reaction environment and the catalyst reactivity thus inducing likely the
higher final yield. On the contrary, the reaction in the presence AN shows an overall yield in COX
lower even though the reaction show a higher COX selectivity. The byproducts (that is CHA, CON and
PCA) are in lower concentration (traces) than that measured in NB hydrogenation, this result suggests a
quite different reactivity of the system, when the starting reagent is nitrobenzene likely induced by the
variable water and NH2OH concentration.
100
80
60
40
20
0
Aniline
COX
other
0
200
400
600
800
1000
1200
1400
1600
Time (minutes)
Fig. 6. Reaction profile of aniline hydrogenation. Run conditions: reaction volume 80 mL, aniline 0.5
mol L-1, H2O 20 mmol, hydrogen Pressure 0.8 MPa, NH2OH.HCl 75 mmol, T 333K, Pd/C 50 mg.
Fig. 7 A shows hydrogen consumption trend for AN hydrogenation, after 240 minutes of reaction, the
trend is a linear consumption of hydrogen similar to that observed in the second part of Fig. 6. The
slope measured, however, is significantly lower than that of the second part of Fig. 4, (0.012 mmol min-
1 vs 0.018 mmol min-1) thus suggesting a strong adsorption of the AN. This is in agreement with the
very low rate of hydrogen adsorption observed in Fig. 7 B (that is 0.003 mmol min-1) in the absence of
NH2OH.HCl.

21
1.0
0.8
0.6
0.4
0.2
0.0
3,5
3,0
2,5
2,0
1,5
1,0
0,5
0,0
Slope: 0.003 mmol min-1
slope: 0.012 mmol min-1
0
50
100
150
200
250
0
50
100
150
200
250
Time (minutes)
time (min)
B
A
Fig. 7. Reaction profile of hydrogen consumption for AN hydrogenation. Run conditions: reaction
volume 20 mL, aniline 0.5 mol L-1, H2O 20 mmol, hydrogen Pressure 0.8 MPa, T 333K, Pd/C 50 mg.
A), reaction in the presence of 30 mmol of NH2OH.HCl; B) no NH2OH.HCl.
Such an evidence confirms the strong inhibiting effect of the AN on the hydrogenation capability of the
catalyst, this probably due to its tough surface adsorption coupled with a slow surface reaction kinetics.
On the contrary, the presence of NH2OH.HCl may favor the kinetics of AN reduction by diminishing
the AN adsorption, and likely, increasing surface mobility of the surface Pd-H species.
The presence of CHA increases NH2OH hydrogenation rate, as can be observed in Fig. 8 where the
initial hydrogen consumption rate is almost the triple of that measured without CHA (Fig. 6, 0.026
mmol L-1 vs. Fig. 8, 0.075 mmolL-1). This effect is possibly due to the basicity of the CHA, which
displaces the hydroxylammonium equilibrium to NH2OH (pKa CHA< pKa NH3 < pKa NH2OH < pKa
AN), that easily reacts with hydrogen by the way of Pd/C catalysis [49-51]. In fact, CHA disappears
from the solution because it changes to its ammonium salt, which precipitates. This is evident since the
solution obtained by, washing the solid with water, filtered away the Pd/C, raised the pH of the solution
with KOH at pH>12, and finally extracting the aqueous phase with CH2Cl2, the GC-MS analysis of
such a solution shows the presence of CHA, clearly. By analyzing the profile of hydrogen consumption
of Fig. 8, it is evident a non-linear trend with a decrease of the slope after 120 minutes of reaction and
10 mmol of adsorbed hydrogen. This trend suggests that the faster initial part of the reaction is due to
the hydrogenation of the free NH2OH, derived from its salt displaced by the CHA, when the reaction

22
begin going slower it is likely due to the lower concentration of the free NH2OH due to the
.
precipitation of the CHA hydrochloride thus reducing the displacement of NH2OH HCl to free NH2OH.
30
Slope: 0.075 mmol min-1
25
20
15
10
5
0
0
50
100
150
200
250
Time (min)
Fig. 8. H2 consumption profile for NH2OH hydrogenation in the presence of 0.5 mol L-1 CHA. Run
conditions: reaction volume 20 mL, Hydrogen Pressure 0.8 MPa, NH2OH.HCl 30 mmol, H2O 20
mmol, T 333K, Pd/C 50 mg, time of reaction 4 h.
The importance of the presence of water in the reaction environment is evident by observing the results
reported in Fig. 9. It shows the influence of water on the yield of COX in the AN hydrogenation. It
appears, the amount of water, until 80 mmol, promotes the formation of COX, thus suggesting an
influence of H2O on NH2OH availability. The reaction, however, occurs also in the presence of
molecular sieves 5A, as H2O scavenger, but in lower yield (about 13%), suggesting that the NH2OH
exists in solution also in the organic solvent, but the presence of small amount of water will aid the
increase of NH2OH availability. In fact, hydroxylammonium salt is sparingly soluble in organic
solvent, being, on the contrary, soluble in water. The system however is really complex even when we
start from aniline. The addition of 20 mmol of water (1 mol L-1) to emulate that produced in the
hydrogenation of nitrobenzene, is not sufficient to estimate the whole water present in the reaction
media since NH2OH is reduced in parallel giving water and ammonium salt. Furthermore, it is
noteworthy, that at the higher water amount, there will be the formation of another liquid phase since
20 mL of pure diethyl ether can solubilize 13 mmol of water [52]. The calculation for these multiphase
equilibria is beyond the scope of the present work but approximate values are reported in Tab S5 in the
supplementary information section. In any case, the role of water is evident in Fig. 9, which shows that

23
the increasing the amount of water until 80 mmol will cause the growth of the COX yield up to 55%.
This effect is likely due to the increase of the NH2OH concentration, which is the active specie in the
COX formation. The Increase after a certain amount of water may cause a decrease of AN
hydrogenation thus diminishing the overall yield in COX. This is agreement with what observed by
using a mixture H2O/THF as a solvent (see Tab. 4 entry 5) where the conversion of aniline is negligible
at very high water concentration. In addition, it is necessary take into account that the NH2OH is
reduced in parallel and this cause the diminishing of the NH2OH availability and then overall COX
yield.
60
50
40
30
20
molecular
sieve 5A
10
0
0
20
40
60
80
100
120
140
160
180
Water (mmol)
Fig. 9. Influence of H2O on COX yield. Run conditions: reaction volume 20 mL, AN 0.5 mol L-1,
Hydrogen Pressure 0.8 MPa, NH2OH.HCl 30 mmol, T 333K, Pd/C 50 mg, time of reaction 20 h.
Further consideration must be done on the presence of the ammonium salt. In fact, its presence is not
unbiased for the reactivity of the system, since it may influence all equilibria, and then, the overall
COX yield. It appears from Fig. 10 that the trend of the influence of NH4Cl amount on the COX yield
have a maximum at 10 mmol (COX yield 65%). This behavior suggests that ammonium salt favor the
formation of the oxime increasing the hydrogenation rate of AN. Indeed, hydrogen consumption rate
reported in Fig. 11 is 0.039 mmol min-1, which is higher than that measured in the presence of only
NH2OH.HCl (Fig. 9 A, 0.012 mmol min-1).

24
70
60
50
40
30
20
10
0
5
10
15
20
NH4Cl (mmol)
Fig. 10. Influence of NH4Cl amount on COX yield. Run conditions: reaction volume 20 mL, AN 0.5
mol L-1, Hydrogen Pressure 0.8 MPa, NH2OH.HCl 30 mmol, H2O 20 mmol, T 333K, Pd/C 50 mg, time
of reaction 20 h.
10
Slope: 0.039 mmol min-1
8
6
4
2
0
0
50
100
150
200
250
Time (minutes)
Fig. 11. Reaction profile of hydrogen consumption for AN hydrogenation in the presence of NH4Cl.
Run conditions: reaction volume 20 mL, AN 0.5 mol L-1, Hydrogen Pressure 0.8 MPa, NH2OH.HCl 30
mmol, H2O 20 mmol, NH4Cl 10 mmol, T 333K, Pd/C 50 mg.

25
After this discussion, it is quite evident the first part of the reaction, that is the hydrogenation of NB to
AN influences the reactivity of the system, mainly for the coproduction of water and ammonium salt,
the latter deriving from the parallel reaction of NH2OH.HCl reduction. Both byproducts promote the
yield in COX.
3.6. Influence of the intermediates and of the products on conversion and selectivity
As discussed in previous sections, nitrobenzene disappears quantitatively in about 60 minutes then
other compounds appear in the reaction media, quantitatively the most important one is AN. Starting
from this evidence we study the reactivity of AN in the presence of various solvent and the results are
in Tab. 4. As expected, it appears the needs of H2 in order to obtain the conversion of AN to COX. In
fact, the reaction carried out without H2 (see supplementary materials Tab. S3) does not show the
formation of COX. It is noteworthy the high selectivity in COX observed with EtOH, Et2O and THF as
solvent. Nevertheless, the reaction does not reach high conversion, lower than 50%, except when
CH3CN is the solvent giving, however, mainly products of aniline alkylation. This low hydrogenation
activity the very strong adsorption of AN on the Pd surface, thus inhibiting the reaction of
hydrogenation, being the aniline surface complex stable and impeding the adsorption of hydrogen. In
fact, aniline hydrogenations were generally carried at high hydrogen pressure exceeding 5 MPa [22-
26]. This strong adsorption of aniline agrees with its reactivity in the absence of NH2OH.HCl, it
appears the conversion reaches as low as 8% and the main product is PCA and N,N-di-
cyclohexylamine (see entry 6). The byproduct (CHA, CON and PCA) are in low concentrations and no
other products are observed. In agreement with the experiment carried out with NB as reagent (Tab. 2),
in the presence of very large amount of water (Tab. 4 entry 5) AN does not give COX or other products
deriving from its reduction
Table. 4. Reactivity of AN. Run conditions: reaction volume 20 mL, AN 0.5 mol L-1, NH2OH.HCl 30
mmol, H2 pressure 0.8 MPa, H2O 20 mmol, T 333K, Pd/C 50 mg, time of reaction 20 h.
Entry Solvent
Conversion Selectivity
COX
CHA CON PCA Others
(%)
47
49
12
100
(%)
98
97
95
20
(%)
0.4
0.3
(%) (%) (%)
1
2
3
4a
5b
Et2O
THF
EtOH
CH3CN
0.2
1.5
0.9
0.3
1.5
0.9
traces traces traces 4
0.1
/
0.1
/
0.7
/
78
100
THF/H2O 2
/

26
6c
a) Other is mainly AN alkylation.
Et2O
8
/
14
/
48
38
b) Reaction carried out phase in great excess of water: THF/H2O = 1/1 wt.
c) Reaction carried out without NH2OH.
CHA does not gives COX, neither in presence of hydrogen nor in its absence (Tab. 5 entries 1, 2). The
simultaneous presence of CHA and AN does not give COX (entry 3 and 4). This confirms the strong
inhibiting effect of the CHA on the formation of COX even though the role of CHA is not clear. In fact,
it is important to consider that after 20 h of reaction CHA practically disappears from the reacted
solution, since it is quantitatively transformed in its ammonium salt. This is confirmed by analyzing the
solution obtained from the solid washed with water, where a mixtures of ammonium salts of NH2OH,
NH3 and CHA is detected. This evidences suggests that CHA is not an intermediate but it is a
byproduct deriving from AN total hydrogenation, which is not in dehydrogenative equilibrium, under
this reaction condition, since its presence completely inhibits AN hydrogenation to COX or other
products that could be an intermediate in COX formation such as CON.
Table. 5. Reactivity of CHA. Run conditions: reaction volume 20 mL, substrate 0.5 mol L-1,
NH2OH.HCl 30 mmol, H2O 20 mmol, T 333K, Pd/C 50 mg, time of reaction 20 h.
Entry substrate
P
H2
Conversion Selectivity
COX
CON PCA Others
(%) (%) (%)
(bar)
5(N2)
8
(%)
1
1
4
1
(%)
1a
2a
3b
4b,
CHA
CHA
/
/
/
/
/
/
/
/
/
/
5
/
100
100
95
CHA+AN 8
CHA+AN 5(N2)
90
a) Other mainly products of CHA alkylations.
b) CHA/AN =1/1 conversion on aniline.
In Tab. 6 the reactivity of CON and PCNA are reported. It appears that CON reacts with NH2OH.HCl
quite easily since the reaction is fast being the conversion of CON after 0.5 h equal to that obtained
after 20 h (conversion of 70% entries 1- 4). The selectivity is higher without any catalyst (94% in COX
entry 2), this agrees with previous researches on the self-catalyzed transformation of some ketones to
amides where the oximation of the ketones occur without any catalyst [53]. As expected, in agreement
with what reported by Corma and co-workers PCA is poorly hydrogenated and mainly to N,N-
dicyclohexylamine (entry 5) [36]. This suggests PCA is a side product and it is not an intermediate of

27
the reaction. PCNA has been observed in solution during the reaction, only in traces amount (by a
careful analysis at the GC-MS) but we have synthesized the compound and tested its reactivity.
Actually, it reacts with practically all the nucleophiles giving fast substitution being AN the living
group. This is evident observing that PCNA hydrolyzes to CON and AN by reacting with the water
present in the solvent and/or hair moisture reaching an equilibrium after few hours (see
supplementary). However, in the presence of hydrogen NH2OH and Pd/C catalysts, PCNA gives COX
in high yield (93% entry 6). In absence of hydrogen but with Pd or Au catalysts COX formation is
practically equal to that observed in experiments carried out without catalysts (compare entry 7, 8, 14
and 15). This suggests the presence of the Pd/C catalyst hydrogenates AN, displaced by NH2OH, that
in turns forms COX in the presence of NH2OH. Au based catalyst are completely ineffective (entry 12
and 13) giving the same results of that observed with no catalysts and/or hydrogen (compare entries 7,
8 and 12-15). Starting from these results, it appears there are no needs of catalysts for the nucleophilic
substitution. In fact, these Au catalysts seems to be not effective in the formation of COX, in agreement
with their poor hydrogenation ability already stated in Tab. 1. The hydrogenation in the presence of
Pd/C of PCNA without NH2OH shows the formation of PCA and di-cyclohexyl amine (entry 16).
Entries 17 and 18 evidences the facile nucleophilic substitution of CHA towards aniline giving N,N-
cyclohexylen-cyclohexylamine in the absence of Pd/C or, in the presence of the Pd catalyst, N,N-
dicyclohexylamine as the result of the hydrogenation of N,N-cyclohexylen-cyclohexylamine.
Comparable results have been observed with CON and COX suggesting the easy nucleophilic
substitution due to CHA (see supplementary Tab. S4).
Table. 6. Reactivity of CON, PCA and PCNA. Run conditions: reaction volume 20 mL, Et2O as a
solvent, substrate 0.5 mol L-1, NH2OH.HCl 30 mmol, H2O 20 mmol, T 333K, catalyst 50 mg.
En Substrate
try
Catalyst
Time of P
Conv.a Selectivity
metal
(%)
/
reaction H2
COX
CHA CON PCA Others
(%) (%) (%) (%)
(h)
20
0.5
20
0.5
20
20
20
20
(bar) (%)
5(N2) 70
5(N2) 70
8
5(N2) 70
8
8
5(N2) >99
5(N2) >99
(%)
89
94
88
91
/
93
46
48
1
2
3
4
5
6
7
8
CON
CON
CON
CON
PCA
PCNA
PCNA
PCNA
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
11
6
12
9
100b
1
50c
48c
/
Pd/C(5)
Pd/C(5)
Pd/C(5)
Pd/C(5)
Pd/C(5)
/
72
6
>99
/
/
3
2
2
3
2
2

28
9
CON
Au/SBA(2) 0.5
Au/SBA(2) 20
Au/C (0.5) 0.5
Au/SBA(2) 20
Au/C (0.5) 20
Au/SBA(2) 20
Au/C (0.5) 20
5(N2) 70
5(N2) 70
92
89
48
47
/
/
/
/
/
/
/
/
/
8
/
/
/
2
2
2
2
10
8
8
/
/
/
1
2
1
2
72
/
8
11
11
10 CON
11 CON
8
8
8
70
>99
>99
12 PCNA
13 PCNA
14 PCNA
15 PCNA
16 PCNAd
17 PCNA+CHAd
50c
50c
50c
50c
10b
92e
80f
5(N2) >99
5(N2) >99
47
46
/
/
/
Pd/C
/
20
20
20
8
8
8
98
>99
>99
18 PCNA+CHAd Pd/C
4
a) Conv. = conversion.
b) Other is mainly N, N-dicyclohexylamine.
c) Mainly AN.
d) Reaction carried without NH2OH.
e) Mainly a mixture AN and N, N-cyclohexyliden-cyclohexane.
f) Mainly a mixture AN and N, N-dicyclohexylamine.
The above results state that the main intermediate is aniline and its hydrogenation is the key step.
Besides CON and PCNA are both intermediates since the nucleophilic attack of NH2OH to these
compound gives COX. On the contrary, the addition of CHA displaces AN from PCNA, thus forming
of N, N-cyclohexyliden-cyclohexane. In addition, the latter is also obtained from CON and COX (see
supplementary Tab. S4) by reacting with CHA. This behavior clearly states that of CHA is a side
product and not an intermediate, since no COX is obtained when CHA is present in relevant amount but
only product of CHA condensation. There are no evidence of a dehydrogenative equilibrium between
CHA and an imine intermediate, which in turn will react to give PCNA and finally COX since CHA
reacts readily as a nucleophile with PCNA giving N, N-cyclohexyliden-cyclohexane. Also imine
intermediate is not present in solution, however it is likely that such an imine intermediate is present on
the Pd surface, which is finally attacked by all nucleophiles present in solution such as NH2OH, H2O,
AN, and CHA.
3.7. Possible reaction path
Despite of the complexity of the reaction and the large number of equilibria and reactions, we can try to
suggest a likely reaction path (scheme 4). All the evidences suggest the first stage is NB hydrogenation
to AN. Likely, the second step is AN partial hydrogenation and formation of an imine surface
intermediate, which readily reacts with all the nucleophiles present in solution giving products signed
by dashed lines. The role of water is here not only the nucleophile, which allow the formation of

29
cyclohexanone, but it favors also the multiphase equilibria that give NH2OH thus favoring COX
formation. As a matter of fact, the product that is revenue in higher amount is COX because of the
higher nucleophilicity of NH2OH compared to that of H2O and AN [54]. In addition, all products
obtained from the nucleophilic reaction are in equilibrium (molecules circled with dashed line). On the
contrary, irreversible hydrogenation reactions (under the conditions investigated) occur giving the
molecules circled by solid line; that is CHA, PCA and N, N-dicyclohexylamine. At the conditions
investigated CHA and N, N-dicyclohexylamine are observed in low concentration (at most 1%), and
the latter only after long reaction time. CHA formation is very slow and if added in a relevant amount
change completely the products distribution (see Tab. 6 and supplementary materials Tab. S3). In fact,
if CHA in the presence of CON, COX and DCHA easily gives nucleophilic substitution, thus forming
N, N-cyclohexylen-cyclohexanamine, which is, in turn, hydrogenated to N, N-dicyclohexylamine (see
supplementary Tab. S3).
Scheme 4. Proposed reaction path for the direct synthesis COX via nitrobenzene hydrogenation.
4. Conclusions
All results states the complexity of the reaction path and the large number of factors it is necessary to
take into account for optimizing the yield in COX. The new insights for this process are the importance