Activity patterns of metal oxide catalysts in the synthesis of N-phenylpropionamide from propanoic acid and aniline

Article abstract of DOI:10.1039/c4cy01504e

The reactivities of various commercial and lab-made oxide samples (e.g., γ-Al₂O₃, CeO₂, ZrO₂ and TiO₂) in the heterogeneous catalytic synthesis of N-phenylpropionamide (T, 383 K) from aniline and propanoic acid have been investigated. All the materials studied drive the direct synthesis of the amide to an extent depending on both the chemical and structural properties. A 0th-order kinetic dependence on the substrate concentrations suggests that the reaction proceeds via a Langmuir-Hinshelwood (L-H) pathway under kinetic control of the adsorption-desorption steps (the rate determining step, r.d.s.). The comparative analysis of the activity data on the basis of the relative surface specific kinetic constant discloses a superior surface reactivity of TiO₂, CeO₂ and ZrO₂ over the γ-Al₂O₃ system, and also highlights marked differences in the catalytic functionality of the titania samples. IR spectroscopic studies of the carboxylic acids and amine adsorption and interaction patterns show the formation of the bidentate, bridging, and unidentate carboxylate intermediates accounting for the different amidation functionalities of the studied materials.

Full text of DOI:10.1039/c4cy01504e

Catalysis
Science &
Technology
View Article Online
View Journal | View Issue
PAPER
Activity patterns of metal oxide catalysts in the
synthesis of N-phenylpropionamide from
propanoic acid and aniline†
Cite this: Catal. Sci. Technol., 2015,
5, 1911
a
b
a
b
Francesco Arena,* Chiara Deiana, Agata F. Lombardo, Pavlo Ivanchenko,
a
a
b
Yuriy Sakhno, Giuseppe Trunfio and Gianmario Martra*
2 3 2 2 2
The reactivities of various commercial and lab-made oxide samples (e.g., γ-Al O , CeO , ZrO and TiO ) in
the heterogeneous catalytic synthesis of N-phenylpropionamide (T, 383 K) from aniline and propanoic acid
have been investigated. All the materials studied drive the direct synthesis of the amide to an extent
depending on both the chemical and structural properties. A 0th-order kinetic dependence on the sub-
strate concentrations suggests that the reaction proceeds via a Langmuir–Hinshelwood (L–H) pathway
under kinetic control of the adsorption–desorption steps (the rate determining step, r.d.s.). The comparative
analysis of the activity data on the basis of the relative surface specific kinetic constant discloses a superior
Received 17th November 2014,
Accepted 17th December 2014
2 2 2 2 3
surface reactivity of TiO , CeO and ZrO over the γ-Al O system, and also highlights marked differences
in the catalytic functionality of the titania samples. IR spectroscopic studies of the carboxylic acids and
amine adsorption and interaction patterns show the formation of the bidentate, bridging, and unidentate
carboxylate intermediates accounting for the different amidation functionalities of the studied materials.
DOI: 10.1039/c4cy01504e
www.rsc.org/catalysis
purification steps resulting in a poor atom economy, while
large amounts of harmful waste, requiring special disposal
1
. Introduction
1
,2
The synthesis of amidic groups is a fundamental class of
organic reactions for manufacturing a variety of biological
compounds and innovative materials; in particular, besides
widespread use in the agrochemical industry, ca. 65%
of drug molecules include at least an amide unit, while
amidic groups also represent the backbone of common poly-
treatments, contribute to high E-factors.
An alternative
method involves the condensation of carboxylic acids with
amines at high temperatures (>453 K), but is unfeasible for
2
many functionalized substrates. On this account, the synthe-
2
sis of amides has been recently indicated as a major green
2
–17
chemistry issue.
1
,2
meric materials.
Apart from the recognized effectiveness of microwave irra-
diation on the condensation of carboxylic acids with
Although organic chemistry offers several methods for the
lab-scale synthesis of amides, the main industrial
manufacturing processes still rely on the reaction of amines
with pre-activated carboxylic acid precursors, like anhydrides
or acid chlorides. In turn, the latter are obtained from thio-
nyl or oxalyl chloride or from carboxylic acids using stoichio-
4
,6
amines, and some catalytic redox routes using alternative
5
,8,9,16
reagents,
the most attractive option for amide bond for-
mation is the heterogeneous catalytic condensation of
2
,3
10–15
amines with carboxylic acids.
Many solid acids such as
7
3+
10
boron–organic compounds, Fe /K10 montmorillonite,
2
11
metric amounts of coupling reagents (i.e., carbodiimides).
FeCl , ZnCl , zeolites, and silica-based catalysts, sulphated-
3 2
tungstate, MCM-41, and Zr-salts, were shown to be
active in the synthesis of amides using toluene under
1
2
13
14
All such methods suffer from severe environmental,
safety, and economic drawbacks, including the use of nox-
ious and/or dangerous reagents, and multiple reaction-
1
–14
“azeotropic distillation” reflux conditions (383 K),
while
CeO showed the highest activity among many oxides in vari-
2
1
5
a
ous transamidation reactions. In addition, Comerford et al.
recently reported that silica and SBA materials are highly
active in the synthesis of N-IJphenyl)-phenylacetamide at
Dipartimento di Ingegneria Elettronica, Chimica e Ingegneria Industriale,
Università degli Studi di Messina, Viale F. Stagno D'Alcontres 31, 98166 Messina,
Italy. E-mail: Francesco.Arena@unime.it; Fax: +39 090 391518;
Tel: +39 090 393134
1
7
423 K in a continuous-flow reactor at short contact times,
b
Department of Chemistry and Centre for “Nanostructured Interfaces and
while mechanistic issues of the condensation of formic and
acetic acid with 1-pentanamine on titania have been recently
Surfaces-NIS”, University of Torino, via P. Giuria 7, 10125 Torino, Italy.
E-mail: Gianmario.Martra@unito.it
1
8
addressed. In this context, Grosjean et al. also emphasized
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
c4cy01504e
the role of the reaction system, showing significant positive
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol., 2015, 5, 1911–1918 | 1911

View Article Online
Paper
Catalysis Science & Technology
1
7
effects of heat input on the amidation kinetics because of
faster rates of water removal.
tests. The samples were treated at 723 K under a dynamic
3
vacuum for 1 h, then 6 mbar of O was admitted at the same
2
Therefore, this work is aimed at providing a comparative
view of the reactivity patterns of various lab-made and com-
mercial oxide catalysts (e.g., γ-Al O , TiO , CeO , ZrO ) in the
temperature (for 1 h) to restore the original oxidation states
of the oxides. Subsequently, the samples were cooled down to
473 K, still under O and then, to room temperature under
2
3
2
2
2
2
condensation reaction of aniline with propanoic acid to
N-phenylpropionamide using toluene as a solvent (T, 383 K).
The kinetic analysis of the activity data reveals that the
amidation functionality depends on both the chemical and
structural properties of the materials studied, promoting the
formation of different intermediates, as probed by IR spectro-
scopic measurements.
vacuum. After the collection of the spectra of the treated cata-
lysts in vacuum (background), high-purity vapors of
propanoic or formic acid and then of 1-pentanamine (Sigma
Aldrich) were added onto the catalysts for IR spectra collec-
tion. In the case of the sequential adsorption of propanoic
acid and 1-pentanamine a set of measurements was further
carried out after heating the pre-saturated samples at 383 K
for 30 minutes. The spectra were reported in absorbance after
subtraction of the background spectra, while the adsorbed
species were analyzed by High Resolution Mass Spectrometry
2
Experimental
2
.1. Catalyst preparation
Commercial TiO (TiO
Merck) and γ-Al O (000-1.5E Akzo Nobel) powders were used
(
HR-MS). After the spectroscopic measurements the pelletized
2
2 2
P25 Evonik, ex Degussa, and TiO
samples were ground and suspended in 0.5 mL of Milli-Q
water. Subsequently, the suspension was shaken in a Vortex
2
3
in their “as received” forms. The CeO
by heating an aqueous solution of a CeIJNH
2
sample was prepared
IJNO precur-
4
mixer for 15 min and then centrifuged at 10 rpm for 5 min.
4
)
2
3 6
)
The supernatant was then removed and the solid was treated
a second time in 0.5 mL of Milli-Q water. Then, the two
obtained solutions were mixed and analyzed in a LTQ
Orbitrap mass spectrometer (Thermo Scientific) equipped
with an atmospheric pressure interface and an electrospray
ionization (ESI) source. The source voltage was set to 4.48 kV.
The heated capillary temperature was maintained at 538 K.
The tuning parameters adopted for the ESI source were:
capillary voltage, 0.02 V; tube lens, 24.77 V; for the ion optics:
multipole 0 offset, −4.28 V; lens 0 voltage, −4.36 V; multipole
sor (Sigma Aldrich, >99%) under continuous stirring and
refluxing at 373 K for 4 h. The resulting precipitate was fil-
tered, washed, dried at 373 K for two days and finally cal-
1
9
cined at 723 K for 3 h. The ZrO sample was prepared by
2
dissolving a ZrOCl
in distilled water, and adding a NaOH solution (0.25 M) at
98 K under vigorous stirring. After digestion, the precipitate
2 2
Ĵ8H O precursor (Sigma Aldrich, >99.5%)
2
was filtered, washed until the chlorides completely
disappeared, dried at 373 K overnight and further calcined in
air at 873 K for 3 h. A list of the samples studied with their
relative codes and specific surface area values is reported in
Table 1.
0
−
offset, −4.28 V; lens 1 voltage, −13.69 V; gate lens voltage,
8.84 V; multipole 1 offset, −18.69 V; front lens voltage, −5.09 V.
The mass accuracy of the recorded ions (vs. calculated) was
±
1 mmu (without internal calibration). The samples, added
2
.2. Catalyst characterization
to 100 μL of a 0.1 M HCOOH aqueous solution, were delivered
directly to the mass spectrometer via a Hamilton microliter
syringe at a constant flow (10 μl min ).
Specific surface area (SSA) values were determined from
nitrogen adsorption isotherms (77 K) obtained using a ASAP
−
1
2
010 (Micromeritics Instrument) gas adsorption device.
Before measurements were conducted the samples were
outgassed at 423 K until a residual pressure of ca. 0.2 mbar
was achieved. The isotherms were elaborated by the BET
method for the SSA calculations.
IR spectroscopy measurements were performed using a
Bruker Vector 22 spectrometer equipped with a DTGS detec-
2.3. Catalyst testing
Catalytic tests were carried out in a magnetically stirred
three-neck glass flask reactor, operating in batch mode and
equipped with a thermometer, a reflux condenser, and a
Dean-Stark device, for continuous water removal by
3
“
azeotropic distillation”. The reactor was loaded with 10 mL
−
1
−2
tor (4 cm resolution); self-supporting pellets (ca. 10 mg cm )
of the catalyst powders were placed into an IR cell with CaF
of toluene (solvent), 0.1 mL of n-octane (internal standard),
mmol (0.45 g) of propanoic acid and 0.2 g of the powdered
catalyst (d < 0.1 mm). The suspension was heated to 383 K,
2
6
−
5
windows and connected to vacuum lines (P < 10 mbar) to
carry out in situ thermal treatments and adsorption–desorption
p
then a stoichiometric amount of aniline (6 mmol, 0.56 g) was
slowly added. Evidence for the absence of internal mass-
transfer resistances is given by the Weisz–Prater criterion
Table 1 List of the studied catalysts
2
.
2
−1
2
rp

r
.

 310
3
Catalyst code
γ-Al (000-1.5E)
Supplier
SSA (m g
)
CWP  
.
s

(1)
DAn/Tol CAn
O
2 3
Akzo Nobel
Merck
Evonik
Lab-made
Lab-made
261
10
55
90
21
TiO
TiO
2
_Merck
_P25
(
where r
p
and (−r) are the particle radius (cm) and the rate
per unit of volume of catalyst (mol cm
(0.09 cm h ) and C (0.55 mmol cm ) are the diffusivity
2
−
3
−1
h ), while D
An/Tol
CeO
ZrO
2
cat
−3
2
−1
2
An
1912 | Catal. Sci. Technol., 2015, 5, 1911–1918
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
and concentration of aniline, respectively), as under the
adopted conditions C
smaller than one.
is always orders of magnitude
W–P
2
0
Reaction mixture samples were withdrawn from the reac-
tor and analyzed by a GC (7890A, Agilent Technologies)
equipped with a capillary column (Restek, Rxi-1ms cross-
bond), connected to a FID for the determination of the
propanoic acid and aniline conversion (±3%). Furthermore,
after 24 h of reaction time the reaction mixture was cooled
and filtered by a sintered glass funnel, and the catalyst was
washed with hot ethanol for the complete solubilization of
the amide, the yield of which was determined by gravimetric
analysis after ethanol evaporation.
3
Results and discussion
3
.1. Catalytic activity
The results of the catalytic tests in the synthesis of the
N-phenylpropionamide at 383 K are summarized in Table 2
in terms of aniline and propanoic acid conversion and amide
yield values after 24 h of reaction time. A preliminary test in
the absence of any catalyst confirms the lack of reactivity of
the substrates, while all the oxide catalysts drive the conden-
sation reaction with amide yields between 4 and 47%, in sat-
isfactory agreement with the conversion degrees of both
reagents (Table 2). In particular, the lowest and highest yields
Fig. 1 (A) Average conversion of propanoic acid and aniline at 383 K
at different stoichiometric concentrations on the TiO
2
_P25 catalyst
−1
(
C
cat, 18 g L ); (B) log-plot of rate data vs. reagent concentrations.
2 2
refer to the TiO _Merck and TiO _P25 samples respectively,
while alumina (24%) and ceria (27%) show intermediate
activities, which are significantly better than ZrO (9%).
2
These data preliminarily indicate that chemical, structural
and textural properties concur to shape the reactivity of the
various oxide systems, although a proper assessment of
their amidation functionality requires the knowledge of their
kinetic parameters.
specific to the TiO _P25 system as no changes in the reaction
2
rate were found using double concentrations of the sub-
2
1
−1
−1
strates (1.1 mol L ) and the CeO
material (36 g L ), while
a constant N-IJphenyl)-
2
Comerford et al. indicated
phenylacetamide yield (0.34–0.39 g g
cat−
1
−1
h ) using a K60 sil-
The kinetic dependence has been properly ascertained by
a series of tests conducted at different (stoichiometric)
reagent concentrations (0.28–1.10 mol L ) and a constant
ica catalyst in a flow reactor (T, 423 K) at various contact
1
7
times for conversion degrees between 9 and 46%. In partic-
−
1
−3
ular, this yield value corresponds to a rate of ca. 1.50 × 10 mol
−
1
−1
−4
TiO _P25 catalyst load. The results in Fig. 1A show straight-
g
g
cat
h
which is comparable to the value of 5.4 × 10 mol
2
cat−
1
−1
2
_P25 at 383 K (Fig. 1B).
line conversion trends up to an extent of ca. 80%, indicating
a very small, if any, kinetic effect of substrate concentration
on the reaction rate and, thus, a 0th-order kinetic depen-
dence (Fig. 1B). Despite the lack of systematic kinetic data
for the other catalysts, this peculiar rate law seems not to be
h
herein found for TiO
This evidence provides a convincing argument that the
0th-order kinetic dependence has general validity
a
irrespective of the catalytic material, likely as a consequence
of the low operating temperature and characteristics of the
“
batch” reaction system (i.e., a high reagent-to-catalyst
2
2
ratio).
Table
2 Conversion data of aniline and propanoic acid, and
Besides, in the presence of (apolar) toluene as a solvent,
N-phenylpropionamide yields with the various catalysts at 383 K (CPA
=
−1
−1
an acid–base interaction of the reacting species drives the
CAn, 0.55 mol L ; Ccat, 18 g L )
3,4
primary formation of an adduct,
a
Conversion (%)
aniline, p. acid
b
Catalyst
Amide yield (%)
+
−
C H NH + C H COOH → C H NH ⋯ OCOC H
(2)
6
5
2
2
5
6
5
3
2
5
No (blank test)
0, 0
0
γ-Al
TiO
TiO
CeO
ZrO
2
O
3
_000-1.5E
22, 25
5, 4
47, 49
28, 27
8, 10
24
4
47
27
9
which undergoes a subsequent intramolecular rearrangement
leading to the formation of the C–N bond and water
elimination.
2
_Merck
_P25
2
2
2
a
b
From GC analysis. From gravimetric analysis.
C H NH ⋯H⋯OCOC H ⇆ C H –NH–CO–C H + H O (3)
6
5
2
2
5
6
5
3
5
2
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol., 2015, 5, 1911–1918 | 1913

View Article Online
Paper
Catalysis Science & Technology
−
1
−1
The latter requires a proper activation step in order to trigger
the nucleophilic attack of the C-atom by the N-atom of the
amine species, this being the rate determining step (r.d.s.) of
where k (h ) and k (h ) are the kinetic constants of the
5 6
−
1
relevant steps, K (L mol ) is the equilibrium constant of the
adduct adsorption and ijα ] (mol g ) is the catalyst density
0
of the active sites. Eqn (8) proves that the experimental
0th-order kinetic dependence depends on large values of the
4
adsorption constant K which means, in turn, an extensive
4
+
−1
4
the amidation reaction. In this context, the role of the solid
catalyst consists of the adsorption–activation of the adduct
+
on a Lewis acid center (α ) to restore the nucleophilic charac-
ter of the nitrogen atom by a weakening of the “N⋯H” inter-
coverage of the active sites by reagents, products, and/or
intermediates.
1
5,18,23
action,
as shown in Scheme 1.
Having ascertained the extensive adsorption of carboxylic
acids, amines, and amides on the surface acidic sites of many
On account of such evidence, the 0th-order constant
(k , mol L h ) has been calculated from the conversion
−
1
−1
0
1
5,18,23
oxide materials,
it can be speculated that the heteroge-
data in Table 2, while, by dividing k by the catalyst concen-
0
neous catalytic amidation proceeds via a classic Langmuir–
w 0
tration (k , k /Ccat) and the latter by the surface area values
2
2
Hinshelwood (L–H) reaction pathway, although a dual-site
(kSSA, k /SSA), the specific constants of the various catalysts
w
1
5
mechanism, like that of amide transamidation, formalde-
per mass and SSA units were obtained, respectively (Table 3).
Constant values different by more than one order of mag-
2
4
hyde to methyl formate dismutation, and triacylglyceride to
2
5
−2
−4
−1 −1
fatty methyl ester transesterification, cannot be ruled out in
light of spectroscopic findings documenting the presence of
nitude (k , 1.0 × 10 –9.2 × 10 mol L h ) confirm the
0
strong influence of the surface structure on the amidation
1
5
24,26
acid–base pairs on ceria, titania,
surfaces.
zirconia and alumina
functionality of the most (TiO
2
_P25) and least (TiO
active titania samples. However, the influence of the sur-
face area on k (Fig. 2A) shows that the TiO _P25 and γ-Al
2
_Merck)
2
4
26
In turn, the multi-step pathway of the catalytic reaction
hinders a direct assessment of the r.d.s. because adsorption–
desorption processes at the temperature of the catalytic
w
2
2 3
O
systems are significantly more and less active than expected
from the linear relationship depicted by TiO _Merck, ZrO
2
2
1
5,22,23
tests (383 K) can affect the reaction kinetics.
Then,
and CeO
γ-Al (1.2 × 10 mol mcat h ) and TiO
mol mcat h ) have the worst and best functionalities
respectively, while TiO _Merck, ZrO and CeO are character-
2
, respectively. Indeed, the kSSA values prove that
−
6
−2 −1
−5
at the first instance, product desorption is also considered to
be a r.d.s.
2
O
3
2
_P25 (1.0 × 10
−
2
−1
2
2
2
−
+
+
R–COO …NH
3
–R′ + α ⇆ (RCOO–H
3
NR′)ads
(4)
(5)
(6)
ized by an intermediate reactivity, as proven by similar kSSA
values, ca. five times larger than γ-Al O (Fig. 2B).
2
3
(
RCOO–H
3
NR′)ads → (RCO–NHR′)ads + H
2
O↑
The origin of the different reactivities of the studied
oxides is addressed by IR measurements.
+
(
RCO–NHR′)ads → RCO–NHR′ + α
3
.2. IR spectroscopy and mechanistic evidence
In particular, the above simplified reaction mechanism
predicts equilibrium conditions for the adsorption of the
adduct on a Lewis acid site (4), while the amide formation (5)
and desorption (6) steps are considered to be r.d.s. (e.g., rate =
r = r ) and irreversible due to continuous water evaporation
The amidation functionality of the various catalysts has been
probed by a systematic IR study of propanoic acid adsorption
and its reaction pattern toward 1-pentanamine (taken as a
model amine for the sake of safety). The spectral data of the
various oxides are shown in Fig. 3, while a summary of the
various signals is reported in Table S1 in the ESI.† In all cases
the adsorption of propanoic acid (blue lines) yields the for-
mation of carboxylate species, according to the νasymCOO
and νsymCOO modes in the 1650–1550 and 1450–1400 cm
ranges, respectively.
In the case of TiO _Merck and γ-Al O the νCO compo-
nent of the carboxylic group in C H COOH molecules in the
2 5
5
6
2
3
and slow amide desorption, respectively. The active site
balance at steady-state conditions
+
+
−
[
α ]
0
= [α ] + [(RCOO–H
3
NR′)ads] + [(RCO–HNR′)ads
]
(7)
−
−1
2
7
leads to the following kinetic equation
k  K  RCOOH NR
2
2 3


1
1
5
4
3

rate

mol gcat
h


  


0

k 
5
−1
1
 K  1


RCOOH NR

non-dissociated form is also present (ca. 1670 cm ).
4



3
k6

(
8)
0 w
Table 3 0th-order (k ) and specific 0th-order constants (k , kSSA) of the
studied catalysts
−
1
−1
cat−1 −1
(mol g h )
cat−2 −1
Catalyst
TiO _P25
k
0
(mol L
h
)
k
W
k
SSA (mol m
h )
−
−
−
−
−
02
03
04
03
03
−04
−05
2
1.0 × 10
2.1 × 10
5.4 × 10
1.2 × 10
5.1 × 10
3.4 × 10
3.0 × 10
1.0 × 10
5.5 × 10
5.1 × 10
3.9 × 10
1.2 × 10
−04
−05
−04
−04
−06
−06
−06
−06
ZrO
TiO
CeO2
γ-Al
2
2
_Merck 9.2 × 10
Scheme
1
Mechanism of the amidation reaction on the oxide
6.2 × 10
5.5 × 10
catalysts.
2
O
3
1914 | Catal. Sci. Technol., 2015, 5, 1911–1918
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
Fig. 2 Influence of the surface area on: (A) the specific 0th-order
constant (k
w
), and (B) the specific surface constant (kSSA) of amide
2 2 2 2 2 3
Fig. 3 IR spectra of TiO _P25, ZrO , TiO _Merck, CeO , and γ-Al O in
formation (T, 383 K) of the various catalysts.
contact with propanoic acid and outgassed at r.t. (blue lines: a, b, c, d, e);
and after the addition of 1-pentanamine (grey lines: a′, b′, c′, d′, e′.). The
black curve of TiO
2
_P25 results from a′−a curve subtraction.
2
Moreover, the pattern in the case of CeO is visibly shifted to
a lower frequency, probably reflecting the peculiar electronic
2
8
properties of the lanthanides.
The subsequent addition of 1-pentanamine on TiO
Fig. 3a′) causes a decrease in components due to propanoate
markedly lower than that of an ionic carboxylate like
2 5
C H COONa (Table 4), suggesting, thus, the prevalent forma-
2
_P25
(
tion on this system of a bidentate carboxylate structure
−
1
27
species and the appearance of a tail in the 1650–1560 cm
range due to the νCO mode of the amide species.
formation of the amide was confirmed by HR-MS data of the
(Scheme 2).
29,30
The
The lack of straight correlations in the evolution of sub-
bands at high and low frequencies hinders the recognition of
−
−
washing solution of the TiO _P25 sample, showing a signal
νasymCOO and ν COO pairs in the other cases. Besides, it
2
sym
at m/z = 144.14 due to the protonated N-pentylpropanamide
can be considered that i) the angle between the two branches
of –COO oscillators is expected to be larger than 90°, and ii)
−
(Fig. 3, top-left inset). For all the other systems the adsorp-
tion of amine results in the superimposition of its IR pattern
and no N-pentylpropanamide was detected in the washing
solutions by HR-MS.
in both the bridging and bidentate structures such oscillators
can be reasonably assumed to be sufficiently equivalent in
order to apply the well-known relationship
Furthermore, the IR signals of the adsorbed propanoic
acid due to the carboxylate moieties appear to be composed
of sub-bands, the position and relative intensity of which are
2
(Isym/Iasym) = ctg (θ/2)
(9)
3
1
characteristic for each oxide. Although the broadness of the
where θ is the angle between the two branches. As a conse-
quence, in each pair of signals the one due to the νasymCOO
−
−
overall νasymCOO patterns of TiO
2
_P25, TiO
2
_Merck and
CeO
2
hinders any reliable deconvolution of the various com-
mode should exhibit a higher intensity with respect to its
−
ponents, the difference between the spectra of TiO _P25,
after and before the amidation, indicates that the νasymCOO
and νsymCOO components which vanish during the reaction
are centered at 1500 and 1440 cm , respectively (Fig. 3, curve
νsymCOO partner. On such a basis, the most intense compo-
2
−
−
nents of the νsymCOO patterns observed for the adsorption
−
of propanoic acid on catalysts other than TiO
paired to the most intense νasymCOO ones. The calculated Δν
2
_P25 should be
−
1
−
−
1
a′−a). These data provide a splitting value of only 60 cm ,
splitting values (Table 4) indicate that the prevailing fraction
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol., 2015, 5, 1911–1918 | 1915

View Article Online
Paper
Catalysis Science & Technology
−
−
Table 4 Splitting between the νasymCOO and νsymCOO modes of
propanoate species on the various catalysts and in the sodium salt for
comparison
The addition of 1-pentanamine causes a decrease of the
components due to formates and the appearance of signals
−
1
in the range 1700–1600 cm , due to the νCO mode of the
−
−
29,30
Catalyst
TiO _P25
ν
asymCOO
ν
symCOO
Δν
amide species,
except for the γ-Al O , in which the addi-
2 3
tion of the amine basically gives rise only to the appearance
of the spectral patterns of newly adsorbed molecules. On the
2
1500
1540
1556
1535
1576
1560
1440
1440
1443
1425
1450
1430
63
ZrO
TiO _Merck
2
CeO
2
100
113
110
126
130
contrary, TiO _P25 shows the strongest decrease of the
2
2
−
−
νasymCOO and ν COO signals. Notably, on the whole these
sym
γ-Al
2
O
5
3
data are in a fairly good agreement with the activity scale
2
C H
COONa
(
TiO _P25 > ZrO ≈ TiO _Merck ≈ CeO > γ-Al O ) of
2
2
2
2
2 3
amidation tests (Fig. 2).
As already observed in the case of propanoic acid, the
spectral features of the formate species also result from the
overlapping of sub-bands, despite the fact that in several
cases they are better resolved. In addition, the absence of the
signal due to the deformation modes of the –CH and –CH
2
3
Scheme 2 Carboxylate structures stabilized on an oxide surface with
−
groups allows for a clearer recognition of the ν
COO and
n+
4+
4+,
4+
3+
asym
M
being the metal (e.g., Ti , Zr Ce , Al ) cation (ref. 27).
−
ν
symCOO patterns. Thus, a deconvolution analysis of the for-
mate patterns has been performed, setting the position of
the various components by the 2nd-derivative method
3
2
−
of carboxylate groups should be of the bridging type
Scheme 2) on ZrO , TiO _Merck, CeO and γ-Al O . The
(Fig. 5). For the broad and featureless νasymCOO band
(
observed on TiO _Merck no significant information was pro-
2
2
2
2
3
2
formation of different amounts of carboxylate species on the
titania samples is still to be related to the surface features of
vided by such a method; in such a case the overall contribu-
tion of the formate species with respect to the HCOOH ones,
2
6
−1
the TiO _P25 and TiO _Merck materials.
responsible for the component at 1674 cm , was evaluated,
2
2
−
Although different intermediates and activity evidence
confirm the superior amidation functionality of the TiO _P25
gathering all possible νasymCOO signals in one component
−
1
2
(centered at 1598 cm ). As a consequence, the possibility of
−
sample, these data do not provide information on differences
among the other catalysts, as they are all inactive toward
amidation under the “model” in situ conditions and at room
temperature.
establishing correlations between specific νasymCOO and
On this account, the catalysts subjected to the adsorption
of propanoic acid and 1-pentanamine were heated to 383 K
and further cooled down to r.t. for IR measurements (Fig. S1
of the ESI†). TiO
2
_P25 shows a further consumption of the
components of the carboxylate moieties, while the weak band
−
1
at 1630 cm rises in intensity, due to the νCO mode of the
2
9,30
amide group.
γ-Al further prove its poor catalytic functionality, while a
strong decrease of all the signals hinders any reliable assess-
ment of the various species on the ZrO , TiO _Merck and CeO
samples, likely as a consequence of significant desorption
Negligible changes in the spectral pattern of
2 3
O
2
2
2
(and condensation on colder points of the IR cell) phenomena.
Then, a series of measurements was carried out using
formic acid, since it is expected to be more reactive because
1
8
of the lack of inductive effects of the alkyl substituent. The
spectral features of formic acid adsorption (Fig. 4, blue lines)
account for the formation of the carboxylate species on
−
−
TiO _P25, ZrO and CeO2 (typical νasymCOO and νsymCOO
2
2
−
1
patterns in the 1600–1500 and 1400–1300 cm
ranges,
respectively), and are shifted to a lower frequency for CeO
vide supra). The components of the HCOOH molecules in
non-dissociated forms in the spectra of TiO _Merck and
γ-Al are also evident. A summary of the assignment of
2
(
2
2 3
O
2 2 2 2 2 3
Fig. 4 IR spectra of TiO _P25, ZrO , TiO _Merck, CeO , and γ-Al O in
the signals on the various catalysts is reported in Table S2
contact with formic acid and outgassed at r.t. (blue lines: a, b, c, d, e);
of the ESI.†
and after the addition of 1-pentanamine (grey lines: a′, b′, c′, d′, e′).
1916 | Catal. Sci. Technol., 2015, 5, 1911–1918
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
ZrO2 and CeO constitute a small fraction of the overall
2
−
νasymCOO pattern in comparison with TiO _P25. Finally, the
2
−
1
−1
narrowest Δν splitting of 202 cm (1593–1391 cm ) calcu-
lated for γ-Al O indicates the presence of bridging formates
2
3
2
7
whilst bidentate species can be excluded. Also the main
asymCOO component at 1629 cm can be attributed to
bridging species if paired with the ν COO component at
391 cm (Δν = 238 cm ), or even to unidentate species
Scheme 2) if νsymCOO at lower frequencies are considered.
For TiO _Merck the ν COO at higher frequencies is cen-
tered at 1403 cm , while the broad νasymCOO pattern
extends down to 1515 cm , indicating that bidentate for-
mates might be present. However, considering the maximum
−
−1
ν
−
sym
−
1
−1
1
(
−
−
2
sym
−
1
−
−
1
−
1
−1
at 1598 cm , a Δν of 195 cm suggests a prevalent presence
of bridging formates. As for propanoic acid, the different
amounts of the various types of formates on TiO _P25 and
2
TiO _Merck are still related to their significantly different
2
2
6
surface structures. Additional evidence on the different
structures of formate on the two titania catalysts are obtained
from the ratio between the sum of the integrated intensities
−
of all the νsymCOO components (Isym) and the analogous for
−
νasymCOO (Iasym). The values obtained were 0.3 and ca. 0.2
2 2
for TiO _P25 and TiO _Merck respectively, representing the
Fig.
5 IR spectra (Fig. 4; blue lines) and deconvolution analysis
average Isym/Iasym values for the various types of formate pres-
ent in the two cases. It is worthwhile to note that the data-
base for such calculations (Table S3 in the ESI†) was obtained
by minimizing the contribution of the δO–C–H mode in the
(
grey lines) with the relative predicted spectra (dotted black lines).
−
−1
νsymCOO components was prevented; yet, two limit pairing
possibilities of the νasymCOO components at lower frequen-
cies can be considered, always taking into account the crite-
rion based on the relative intensities of the νasymCOO and
deconvolution of the 1450–1300 cm pattern in the case of
−
TiO
TiO
2
_Merck and, vice versa, maximizing it in the case of
_P25. The above values of the Isym/Iasym ratio can then be
2
−
considered as the closest ones within the variability of the
fitting procedure. The average angles between the C–O moie-
−
νsymCOO signals reported above:
−
i) the pairing with the νsymCOO component at the lowest
2 2
ties of the formates on TiO _P25 and TiO _Merck are found
3
1
frequency, resulting in the maximum Δν with respect to
νasymCOO ;
to be ca. 120° and 130°, respectively. These values are con-
−
sistent with those reported by Deacon and Phillips for a
−
28
ii) the pairing with the νsymCOO component at the
series of acetate complexes (Fig. S2 in the ESI†). In fact,
highest frequency, resulting in the minimum Δν with respect
only bridging species feature angles larger than 123°; then,
also considering the differences between acetates and for-
mates, it can be argued that bridging species might be preva-
−
to νasymCOO .
−
On this account, the type i) pairings of the νasymCOO
components at 1541 and 1519 cm obtained for TiO _P25
with the νsymCOO component at 1361 cm provide a Δν
splitting of 180 and 158 cm ; these can be related to bridg-
ing/bidentate and bidentate formate species respectively, as
−
1
lent on TiO _Merck. Angle values in the range of 118–123°
2
2
−
−1
appeared common to bridging and bidentate acetates, while
the latter selectively exhibit angles between 118 and 109°;
−
1
thus, an average value of 120° for formates on TiO _P25
2
−
1 27
HCOONa exhibits a Δν splitting of 201 cm .
In the case of ZrO , the recognition of the components
accounts for a mixture of formate structures containing sig-
nificant amounts of bidentate formates.
2
due to the bridging/bidentate formates might result only
−
from type ii) pairings between the ν
COO component at
asym
−
1
−
1
550 cm and the highest frequency νsymCOO component at
4
Conclusions
−
1
−1
1
370 cm (Δν, 180 cm ).
The same occurs by considering components at 1548 and
• The reactivity of various catalytic oxide materials (T, 383 K)
in the direct synthesis of N-phenylpropionamide from
propanoic acid and aniline has been assessed.
−
1
−1
1
371–1361 cm
2
for CeO (Δν, 177–187 cm , bridging/
bidentate); the presence of the bidentate formate might be
−
indicated by the pairing between the ν
COO components
•
Chemical and surface properties determine the
asym
−1
−
−1
at 1509 cm with the νsymCOO components at 1371–1361 cm
amidation functionality of the studied systems:
−
1
−
(
Δν, 138–148 cm ). In any case, the ν
COO components
asym
of the bridging/bidentate and bidentate formates on both
TiO
2
_P25 > ZrO
2
≈ CeO
2
≈ TiO
2 2 3
_Merck > γ‐Al O
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol., 2015, 5, 1911–1918 | 1917

View Article Online
Paper
Catalysis Science & Technology
•
A 0th-order kinetic dependence is related to a Langmuir–
13 K. Komura, Y. Nakano and M. Koketsu, Green Chem., 2011,
13, 828.
14 C. L. Allen, A. R. Chhatwal and J. M. J. Williams, Chem.
Commun., 2012, 48, 666.
15 M. Tamura, T. Tonomura, K. Shimizu and A. Satsuma, Green
Chem., 2012, 14, 717.
Hinshelwood reaction path under kinetic control of the
product-desorption step.
•
The reactivity of the studied oxides matches their ability
to stabilize bidentate carboxylate intermediates.
Bidentate carboxylates are much more reactive towards
•
the nucleophilic attack of amines than both bridging and
unidentate species.
16 J. Soulè, H. Miyamura and S. Kobayashi, J. Am. Chem. Soc.,
2011, 133, 18550.
1
1
1
2
7 J. W. Comerford, T. J. Farmer, D. J. Macquarrie,
S. W. Breeden and J. H. Clark, ARKIVOC, 2012(vii), 282–293.
8 C. Deiana, Y. Sakhno, M. Fabbiani, M. Pazzi, M. Vincenti
and G. Martra, ChemCatChem, 2013, 5, 2832.
Acknowledgements
The University of Torino and the Compagnia di San Paolo
are acknowledged for funding (project no. ORTO11RRT5).
Dr. Marco Pazzi is gratefully acknowledged for the HR-MS
analysis.
9 H. Zou, Y. S. Lin, N. Rane and T. He, Ind. Eng. Chem. Res.,
2
004, 43, 3019.
0 H. S. Fogler, Elements of Chemical Reaction Engineering,
Pearson Education, Inc., 4th edn, 2006, p. 839.
2
2
1 G. Busca, Phys. Chem. Chem. Phys., 1999, 1, 723.
2 I. Chorkendorff and J. W. Niemantsverdriet, Concepts of
Modern Catalysis and Kinetics, WILEY-VCH GmbH & Co.
KGaA, Weinheim, 2005.
Notes and references
1
2
3
K. Ishihara, Tetrahedron, 2009, 65, 1085.
V. R. Pattabiraman and J. W. Bode, Nature, 2011, 480, 471.
C. Grosjean, J. Parker, C. Thirsk and A. R. Wright, Org.
Process Res. Dev., 2012, 16, 781.
23 L.-F. Liao, C.-F. Lien, D.-L. Shieh, F.-C. Chen and J.-L. Lin,
Phys. Chem. Chem. Phys., 2002, 4, 4584.
24 A. Parmaliana, F. Frusteri, F. Arena, A. Mezzapica and
V. Sokolovskii, Catal. Today, 1998, 46, 117.
25 C. Cannilla, G. Bonura, F. Arena, E. Rombi and F. Frusteri,
Catal. Today, 2012, 195, 32.
26 G. Martra, Appl. Catal., A, 2000, 200, 275.
27 K. Nakamoto, Infrared and Raman Spectra of Inorganic and
Coordination Compounds, Wiley, New York, 4th edn, 1986,
pp. 232–233.
4
5
6
L. Perreux, A. Loupy and F. Volatron, Tetrahedron, 2002, 58,
2
155.
A. Khalafi-Nezhad, A. Parhami, M. N. Soltani Rad and
A. Zarea, Tetrahedron Lett., 2005, 46, 6879.
E. Gelens, L. Smects, L. A. J. Sliedregt, B. J. van Steen,
C. G. Kruse, R. Leurs and R. V. A. Orru, Tetrahedron Lett.,
2
005, 46, 3751.
7
8
9
K. Arnold, A. S. Batsanov, B. Davies and A. Whiting, Green
Chem., 2008, 10, 124.
A. Kumar, H. K. Akula and M. K. Lakshman, Eur. J. Org.
Chem., 2010, 2709.
28 G. B. Deacon and R. J. Phillips, Coord. Chem. Rev., 1980, 33,
227.
29 N. B. Colthup, L. H. Daly and S. E. Wiberley, Introduction to
Infrared and Raman Spectroscopy, Academic Press, New York,
2nd edn, 1975, pp. 321–225.
B. Gnanaprakasam and D. Milstein, J. Am. Chem. Soc., 2011,
1
33, 1682.
30 X. G. Chen, R. Schweitzerstenner, S. A. Asher, N. G. Mirkin
and S. Krimm, J. Phys. Chem., 1995, 99, 3074.
31 P. S. Braterman, Metal Carbonyl Spectra, Academic Press,
London, 1975, p. 64.
1
1
0 K. V. N. S. Srinivas and B. Das, J. Org. Chem., 2003, 68, 1165.
1 J. W. Comerford, J. H. Clark, D. J. Macquarrie and
S. W. Breeden, Chem. Commun., 2009, 2562.
1
2 P. S. Chaudhari, S. D. Salim, R. V. Sawant and
K. G. Akamanchi, Green Chem., 2010, 12, 1707.
32 Handbook of Vibrational Spectroscopy, ed. J. M. Chalmers
and P. R. Griffiths, Wiley, 2001, vol. 3, p. 64.
1918 | Catal. Sci. Technol., 2015, 5, 1911–1918
This journal is © The Royal Society of Chemistry 2015