K3PO4-catalyzed carbonylation of amines to formamides

Article abstract of DOI:10.1016/j.apcata.2015.09.004

Synthesis of formamides from the catalytic carbonylation of amines with CO is of great interest due to their wide applications of formamides as synthetic intermediates and aprotic polar solvents. Up to now, the most commonly used catalysts are either expensive noble metal compounds or difficult-to-handle alkali metal alkoxides. We have found that a solid base, K₃PO₄ is a highly active and selective catalyst for the carbonylation of primary and cyclic secondary amines, producing corresponding formaldehydes in yields up to 99% at relatively mild conditions. Spectroscopic and quantum mechanical calculation results indicate that such high activity of K₃PO₄ is closely related to the strong hydrogen bonding ability of PO₄^3- with the amino group of an amine, thereby enhancing the nucleophilicity of the amino group enough to interact with a neutral molecule, CO and mediating proton transfer from the amino group to the carbonyl group. Characterization of spent catalyst by FT-IR and XRD implies that small portion of K₃PO₄ is converted into less active K₂HPO₄ during the carbonylation reaction.

Full text of DOI:10.1016/j.apcata.2015.09.004

Accepted Manuscript
Title: K PO -catalyzed carbonylation of amines to
3
4
formamides
Author: Young Jin Kim Ji Woo Lee Hyun Ji Lee Shuyao
Zhang Je Seung Lee Minserk Cheong Hoon Sik Kim
PII:
S0926-860X(15)30137-X
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2015.09.004
APCATA 15536
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
15-5-2015
29-7-2015
5-9-2015
Please cite this article as: Young Jin Kim, Ji Woo Lee, Hyun Ji Lee,
Shuyao Zhang, Je Seung Lee, Minserk Cheong, Hoon Sik Kim, K3PO4-
catalyzed carbonylation of amines to formamides, Applied Catalysis A, General
http://dx.doi.org/10.1016/j.apcata.2015.09.004
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. 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.

K PO -catalyzed carbonylation of amines to
3
4
formamides
1
1
Young Jin Kim , Ji Woo Lee , Hyun Ji Lee, Shuyao Zhang, Je Seung Lee, Minserk Cheong,
Hoon Sik Kim* khs2004@khu.ac.kr
Department of Chemistry and Research Institute of Basic Sciences, Kyung Hee University, 26
Kyungheedae-ro, Dongdaemun-gu, Seoul 130-701, Republic of Korea
*
Corresponding author. Tel.: +82 2 961 0458; fax: +82 2 966 3701.
1
These authors equally contributed to this work.
1
| P a g e

Graphical Abstract
2
| P a g e

Highlights
►
►
►
►
►
K
3
PO
4
showed superior catalytic activity to others for the carbonylation of amines.
PO at mild conditions.
is capable of forming a strong hydrogen bond with an amine.
functions as a proton shuttle as well as an amine activator.
PO was transformed into K HPO during the carbonylation.
Up to 99% of formaldehydes were produced using K
3
4
3
3
-
-
PO
4
4
3 4
of K PO
PO
Small portion of K
3
4
2
4
3
| P a g e

Abstract
Synthesis of formamides from the catalytic carbonylation of amines with CO is of great interest
due to their wide applications of formamides as synthetic intermediates and aprotic polar
solvents. Up to now, the most commonly used catalysts are either expensive noble metal
3 4
compounds or difficult-to-handle alkali metal alkoxides. We have found that a solid base, K PO
is a highly active and selective catalyst for the carbonylation of primary and cyclic secondary
amines, producing corresponding formaldehydes in yields up to 99% at relatively mild
conditions. Spectroscopic and quantum mechanical calculation results indicate that such high
3
-
activity of K
3
PO
4
is closely related to the strong hydrogen bonding ability of PO
4
with the
amino group of an amine, thereby enhancing the nucleophilicity of the amino group enough to
interact with a neutral molecule, CO and mediating proton transfer from the amino group to the
carbonyl group. Characterization of spent catalyst by FT-IR and XRD implies that small portion
3 4 2 4
of K PO is converted into less active K HPO during the carbonylation reaction.
3 4
Keywords: Amines; Carbonylation; Formamides; K PO ; Nucleophilicity
4
| P a g e

1
. Introduction
Formamides are unique solvents that can dissolve wide variety chemicals including
inorganic salts, polymers, natural products, and dyes because of their high dielectric constant and
electron-donor characteristics [1,2]. They are also being widely used as aprotic solvents for
chemical reactions and as intermediates in various organic syntheses. For instance, N,N-
dimethylformamide (DMF) is one of the few solvents suitable for dissolving polyacrylonitrile.
Formamides also find applications as extracting agents in several extraction processes such as
the selective separations of acetylenes from olefins, butadiene from C
containing gases, and aromatics from hydrocarbon mixtures [3-7].
4 2 2
raffinates, SO from CO -
Formamides can be prepared by reacting methylformate, prepared separately from CO and
methanol, with ammonia, primary amines or secondary amines as shown in equations (1) and (2).
The most striking feature of this process is that the reactions can be conducted in the absence of a
catalyst and thus any additional process to remove catalyst is not required. One drawback is that
the economic efficiency of the process is heavily dependent on the price of methylformate [8,9].
OH
O
R NH
+
R N
C
H
O
CH3
(1)
(2)
2
2
H
O
C
O
CH3
OH
C
O
C
R N
CH3
2
R N
H
+
CH OH
3
2
H
Catalytic N-formylation of amines by various formylating agents such as formic acid,
paraformaldehyde has also been reported. A number of catalytic systems have been developed
and as a result, significant advance has been made [10-14]. Nonetheless, the N-formylation
5
| P a g e

process needs to be further improved in terms of material costs and catalyst recovery for
practical application.
Industrially, formamides are being manufactured from the direct carbonylation reactions of
amines by CO in alcohol solvents in the presence of an alkali metal alkoxide such as NaOCH
3
and KOCH as shown in equation (3). The direct synthesis using alkali metal alkoxides is proven
3
to be highly effective for manufacturing formamides in high yields, but the handling of water-
sensitive alkali metal alkoxides requires extreme care because they are easily converted to
inactive alkali metal hydroxides and/or alkali metal formates by the reaction with water [15,16].
O
(3)
CO
+
HNR2
R N
2
C
H
As alternatives to alkali metal alkoxides, group 4-10 organometallic compounds have been
employed as catalysts for the carbonylation reactions of amines [17-22]. However, most of
transition metal complexes were found to exhibit high activities only at elevated temperatures
and pressures even with precious metal compounds [23-29]. Recently, oxidative carbonylation of
amines has been developed to prepare formamides at mild conditions without using transition
metal complexes [30]. This is an elegant process in that the carbonylation is catalyzed by
inexpensive bases such as K
2
CO
3
and 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) at relatively
as an oxidant and NaI as a promoter
o
low temperatures of 90 C, but the additional use of NaIO
4
could be a drawback of this process. As a means of avoiding the use of toxic carbon monoxide,
the reductive functionalization of CO with amines to formamides has been intensively
investigated, and as a result, highly active catalysts based on Ru and Fe have been developed
2
6
| P a g e

[31-34]. Although the advance has been significant, the performance of catalysts needs to be
further improved to compete with that of the conventional alkali metal alkoxide catalysts.
Recently, it has been demonstrated that imidazolium-based ionic liquids with a bicarbonate
-
anion ([HCO
3
] ) are highly active for the carbonylation of amines to produce formamides [35].
Nonetheless, much needs to be improved for their practical application, especially in terms of
cost and recyclability of ionic liquid catalysts.
In a continuation of our study to search for high performance catalysts for the
3 4
carbonylation of amines, we have found that a solid base catalyst, K PO can be used as a
promising alternative to alkali metal alkoxides because it is highly active, cost-effective, and
easily recyclable. Herein, we report on the catalytic activities of a series of alkali metal
phosphates for the carbonylation of amines. The mechanistic aspects of the carbonylation of
amines are also discussed on the basis of experimental and computational results.
2
. Experimental
2
.1. Materials
All the amines used for the carbonylation reactions were purchased from Aldrich Chemical
Co. and used as received. Solvents were reagent grades, and were used after distillation over
appropriate drying agents under a nitrogen atmosphere prior to use [36]. Carbon monoxide (>
9
9%) was obtained from Air Liquide Korea Co. Alkali metal phosphates were obtained from
o
Aldrich Chemical Co., and used after calcination at 500 C for 3 h.
2
.2. Instrumentation
7
| P a g e

Liquid products obtained from the carbonylation reactions were characterized on an
Agilent 6890 gas chromatograph (GC) equipped with a FID and a DB-5 capillary column, and on
an Agilent 6890-5973 MSD GC-mass spectrometer equipped with a HP-MS capillary column.
FT-IR spectroscopic analysis was made on a Nicolet 380 spectrophotometer (Thermo Electron
Co.). The X-ray diffraction (XRD) patterns were recorded on a Shimadzu X-ray diffractometer
o
(
XRD-6000) using nickel-filtered CuKα radiation with a 2θ angle from 10 to 40 . FT-IR spectra
were obtained using a Nicolet FT-IR spectrometer (iS10, USA) equipped with a SMART
MIRACLE accessory and a Specac Reaction Cell Top Plate (P/N GS 10507) for in situ analysis.
2
.3. Carbonylation reaction
Carbonylation reactions of amines were performed in an alcohol using a 100 mL high
pressure reactor equipped with a magnetic stirrer and an electrical heater. The reactor was loaded
with appropriate amounts of an amine (50 mmol), CH OH (20 mL), a catalyst (0.5 mmol) and
3
toluene (2 g) as an internal standard. The reactor was pressurized with 2.0 MPa of CO and then
heated to a specified temperature. The pressure inside the reactor was maintained at 4.0 MPa
throughout the reaction using a gas reservoir equipped with a high pressure regulator and a
pressure transducer. After the completion of the carbonylation reaction, the reactor was cooled to
room temperature and the reaction mixture was analyzed by GC, GC-MS, and NMR.
Comparison of fresh and spent catalyst was made by means of FT-IR and XRD.
2
.4. Catalyst Recycle
The catalyst recycle was conducted with K
cyclohexylamine (CHA, 100 mmol) as a substrate and methanol (20 mL) as a solvent. The molar
ratio of CHA/K PO was set at 100. After the reaction was completed, the reactor was cooled to
◦
3
4
PO at 140 C and 4.0 MPa of CO for 4 h using
3
4
8
| P a g e

room temperature, and the liquid product mixture was separated from the solid catalyst using a
porous metal filter, and analyzed by GC. The remaining solid catalyst inside the reactor was used
3
for the next cycle with a fresh charge of CHA and CH OH (20 mL).
2
.5. Computational calculations
The carbonylation reaction of amine in the presence of K
3
4
PO was theoretically
investigated using Gaussian 09 program [37]. For simplicity, methylamine was chosen as the
model amine. Geometries were fully optimized without any symmetry constraints using nonlocal
method B3LYP, a combination of Becke’s three-parameter hybrid exchange functional and
correlation functional of Lee, Yang and Parr. The 6-31G* basis set was used for all of the atoms.
Harmonic force constants were computed at the optimized geometries to characterize the
stationary points as minima by the absence of imaginary frequencies. Preliminary transition state
geometries were obtained by the PES scan functionality available in Gaussian and then they were
subjected to full TS optimizations using an eigenvector following quadratic synchronous transit
(QST) algorithm. For selected transition state geometries, confirmation calculations involving
intrinsic reaction path (IRC) calculations and frequency calculations with only one imaginary
frequency in the reaction coordinate were performed. For a better modelling in consistent with
the experimental results, the bulk solvation effect of methanol was also included in the
calculation through the self-consistent reaction field theory (SCRF) based on the Polarisable
Continuum Model (PCM).
3
. Results and discussion
9
| P a g e

3
.1. Activities of alkali metal phosphates for the carbonylation of amines
As an alternative to alkali metal alkoxides, alkali metal phosphates were tested as catalysts
for the carbonylation of cyclohexylamine (CHA) to manufacture cyclohexylformamide (CHF).
o
As can be seen in Table 1, the carbonylation reaction of CHA was found to proceed at 140 C
even in the absence of a catalyst, but the yield was as low as 6% (Table 1). However, when the
carbonylation of CHA was conducted in the presence of K
3
PO
4
, the yield of CHF increased
PO or Li PO as the catalyst
drastically to 96%. In contrast, the carbonylation of CHA using Na
3
4
3
4
produced CHF in much lower yields of 66% and 23%, respectively, implying that the catalytic
activity of an alkali metal phosphate is strongly dependent on the kind of alkali metal. It is well
established that alkali metal ions are regarded as Lewis acids, and the acidity decreases with
increasing ion size [38,39]. This is why the alkali metal salts bearing the same anion exhibit
different basicity with the variation of the alkali metal cation: the stronger the Lewis acidity of
the cation, the lower the basicity of the salt. For alkali metal phosphates, it is reported that the
increasing order of base strength expressed by a Hammet acidity function (H
7.2) > Na PO (H ≥ 9.3) > Li PO (H ≤ 7.1) [40]. This suggests that the nucleophilicity or the
basicity of phosphate anion (PO ) of alkali metal phosphates increases with the increasing size
o 3 4 o
) is K PO (H ≥
1
3
4
o
3
4
o
3
-
4
of the alkali metal ion. For an amine to be carbonylated by CO, there should be a strong
interaction between the amino of the amine and the carbonyl carbon of CO. In general, however,
the nucleophilicity of amines is not sufficiently high enough to interact with the carbonyl carbon
of a neutral molecule, CO. Therefore, for the strong interaction with CO, the amino group needs
to be activated by a catalyst that is capable of forming a hydrogen bond or bonds with the amino
group. In other word, for a catalyst to be active for the carbonylation of an amine, the catalyst
should have a strong tendency to form a hydrogen bond with the amino group, thereby enhancing
1
0 | P a g e

3 4
the nucleophilicity of the amino group. In this context, the highest activity of K PO among three
3
-
alkali metal phosphates can be rationalized because PO
4
interacts most weakly with the least
becomes most available toward the
+
3-
Lewis acidic cation K , and as a result, PO
4
of K
3
PO
4
nucleophilic interaction with amines. As listed in Table 1, the catalytic activities of K
KH PO were found to be much lower than that of K PO , affording CHF in yields of 30 and
7%, respectively. This is reasonable because smaller sized H would bind to PO
2 4
HPO and
2
4
3
4
+
3-
1
4
much more
will be less susceptible than that of
toward the nucleophilic interaction with the amino group of CHA. As can be conceivable
from their neutral characters, the carbonylation activities of KNO and K SO were much lower
than that of K PO , affording CHF in yields of 28 and 22%, respectively.
For comparison, the catalytic performance of previously reported RuCl
+
3-
strongly than K , and thus PO
4
2 4 2 4
of K HPO and KH PO
K
3
PO
4
3
2
4
3
4
3
·3H
·3H
under the experimental
2
O and
RhCl
RhCl
3
3
·3H
2
2
O were evaluated [26,27,29]. However, the catalytic activities of RuCl
O were found to be significantly lower than that of K PO
3
2
O and
·3H
3
4
condition, producing CHF in yields of 36 and 11%, respectively. The formation of large amounts
of dicyclohexylurea was also observed, suggesting that RuCl
suitable catalysts for urea synthesis.
3 2 3 2
·3H O and RhCl ·3H O are more
3
.2. Effect of temperature and pressure
Table 2 shows the change of the CHF yield with temperature and CO pressure for the
3 4
carbonylation of CHA conducted in methanol in the presence of K PO for 4 h. The yield of CHF
o
increased rapidly with the temperature rise up to 150 C, but decreased slowly on further increase
of the temperature. The selectivity of CHF was maintained at almost 100% in the temperature
o
range 80 ~ 140 C, but the formation of small amounts of alkylated CHA and alkylated CHF
1
1 | P a g e

o
were observed above 140 C. In contrast to the effect of temperature, the effect of CO pressure
on the carbonylation of CHA was not pronounced. When the CO pressure was raised from 1.0 to
o
6
.0 MPa at 140 C, the yield of CHF increased only by 6% from 93 to 99%.
3
.3. Effect of catalyst loading
3 4
The effect of catalyst loading on the carbonylation of CHA was investigated using K PO
o
at 140 C and 4.0 MPa of CO for 4 h by varying the molar ratio of K
3
PO
PO /CHA molar ratio of
.001. The CHF yield rapidly increased with the increase of the molar ratio up to 0.01, but
remained almost constant thereafter.
4
/CHA. As listed in
Table 3, CHF was produced in moderate yield of 48% even at the K
3
4
0
3
.4. Effect of solvent
3 4
K PO -catalyzed carbonlyation reaction of CHA was also conducted in various solvents
having different polarities. As can be seen in Table 4, the carbonylation reactions were very slow
in aprotic solvents, producing CHF in yield around 1%, irrespective of the solvent polarity. In
particular, in N-methylpyrrolidinon (NMP), the CHA conversion and CHF selectivity were only
1
4 and 1%, respectively. The carbonylation in alcohol solvents were strongly influenced by the
alkyl chain length and steric hindrance. The CHF yield was found to decrease with the increase
of chain length and the steric hindrance. For instance, in n-butanol and t-butanol, the K PO was
3
4
almost inactive. The reason for this is not fully understood at the moment, but it is assumed that
the hydrogen bond interaction between the hydroxyl group of an alcohol and an oxygen atom of
1
2 | P a g e

K
3
4
PO is extremely important, and such interaction is closely related to the acidity of an alcohol
[41] and also to the steric crowding around the hydroxyl group.
3
.5. Carbonylation of other amines
The activity of K PO was also evaluated for the carbonylation of various aliphatic and
alicyclic amines. As can be seen Tables 5 and 6, primary amines such as n-butylamine (pK
0.59) and n-hexylamine (pK = 10.56) [42,43] produced corresponding formamides in yields
over 90%. On the contrary, less basic 2-methoxyethylamine (pK = 9.20) and aniline (pK
.63) [43,44] afforded considerably lower yields of corresponding amides in 17 and 6%,
respectively, implying that the reactivity of amines is strongly affected by their basicities. Unlike
primary amines, secondary amines, di-n-butylamine (pK = 11.25) and di-n-hexylamine (pK
1.31) [42,43] gave much lower yields of formamides although their basicities are stronger than
primary amines, indicating that the basicity is not a sole factor in determining the reactivity of an
amine. As discussed earlier, the carbonylation of an amine in the presence of K PO is believed
to be strongly dependent on the hydrogen bond interaction between the amino group and PO
9]. In this regard, the higher reactivity of primary amines than that of secondary amines could be
rationalized because secondary amines possess one less active hydrogen atoms than primary
3
4
a
=
1
a
a
a
=
4
a
a
=
1
3
4
3
-
4
[
3
-
4
amines to interact with PO . Steric crowding around the amino groups exerted by one more
alkyl group could also be a reason for the lower reactivity of secondary amines. The effect of
steric crowding was evident in the carbonylation of hindered primary amines. As can be seen in
Table 3, i-butylamine (pK
reactivity toward the carbonylation although their basicities expressed in pK
same as that of n-butylamine (pK = 10.59) [42,43]. It is likely that the activation of N-H bonds
a
= 10.63) and t-butylamine (pK
a
= 10.55) showed considerably lower
a
values are almost
a
1
3 | P a g e

3
-
of i-butylamine and t-butylamine by PO
4
via hydrogen bond interaction and the following
interaction with CO seem to be greatly restricted by the presence of a bulky i-butyl or t-butyl
group.
It is worth to note that, 5- and 6-membered heterocyclic amines, pyrrolidine (pK
and piperidine (pK = 11.22) exhibited much higher reactivity than acyclic secondary amines, di-
n-butylamine and di-n-hexylamine with close pK values [42,43], producing corresponding
amides, N-formylpyrrolidine and N-formylpiperazine, respectively, in yields over 90% (Table 5).
This cannot be explained without considering front strain effect because the pK values of
a
= 11.27)
a
a
a
heterocyclic amines are almost same as those of acyclic secondary amines. The higher reactivity
of heterocyclic amines could be attributed to the relieved front strain (direct steric interference)
at the amino groups in the heterocyclic ring structures. With such relieved steric interference, the
3
-
amino groups become more susceptible to the hydrogen bond interaction with PO
result, the nucleophilic interaction of amines with CO can be more facilitated. On the contrary,
piperazine (pK = 9.8) and morpholine (pK = 8.36) with an additional electronegative atom in
4
and as a
a
a
the heterocyclic ring exhibited significantly lower reactivity toward the carbonylation although
they have similar structures to pyrrolidine. This is a strong indication that the reactivity of an
3 4
amine toward the carbonylation in the presence of K PO is determined by both its basicity and
the steric crowding around the amino group. Taken all these together (Tables 1, 3 and 5), it is
concluded that, for the facile carbonylation of amines, the amines must be less hindered as
possible and strongly basic enough to interact with carbon monoxide, and the catalyst must
possess sufficiently high nucleophilicity enough to activate the amino group via a hydrogen
bonding interaction.
1
4 | P a g e

3
.6. Catalyst characterization
3
.6.1. FT-IR spectroscopy
3 4
To gain insight into the origin of the higher activity of K PO for the carbonylation of
amines, the interactions of alkali metal phosphates with n-hexylamine (HA) were investigated by
FT-IR spectroscopy. As can be seen in Fig. 1-(a), HA shows two distinct absorption peaks
-1
corresponding to asymmetric and symmetric N-H stretching frequencies at 3367 and 3290 cm ,
-1
respectively [45]. Upon interaction with equimolar amount of K
3
PO
4
, the peak at 3367 cm
-1
-1
-1
shifted by 20 cm to a lower frequency at 3347 cm , whereas the peak at 3290 cm moved to
-1
-1
lower frequency only by 8 cm at 3282 cm (see Fig. 1-(b)). This might indicate that
3
-
asymmetric N-H movement is much more strongly influenced by the interaction with PO
-(c) is the spectrum showing the interaction of HA with Na PO . The absorption peaks
corresponding to the asymmetric and symmetric N-H stretching frequencies also moved to lower
4
. Fig.
1
3
4
-
1
frequencies at 3363 cm by the interaction with Na
3
PO
4
, but the degrees of shift were only 4 and
PO , the peak
associated with the asymmetric N-H stretching frequencies of HA at 3367 cm remained almost
unchanged on interaction with Li PO . These results strongly suggest that the degree of peak
-1
5
cm , respectively. As can be conceivable from the much lower activity of Li
3
4
-1
3
4
shift for the asymmetric N-H stretching frequency could be used as a barometer in evaluating the
catalytic activity of alkali metal phosphates for the carbonylation of amines.
Like N-H stretching frequency, the peaks associated with P=O stretching also moved to
lower frequencies when alkali metal phosphates were interacted with HA, but the degrees of
peak shift were much smaller compared with those of N-H stretching frequencies. This could be
3
-
attributed to the charge delocalization among four oxygen atoms of PO
4
. As can be seen in Fig.
-1
2
, the peak corresponding to the P=O stretching frequency of K PO at 988 cm [46] moved by 9
3
4
1
5 | P a g e

-1
-1
cm to 979 cm when K
3
PO
4
was contacted with HA. Similarly, the P=O stretching peak shifts
-1
were also observed for Na
3
4 3 4
PO and Li PO , but the shifts were found as only 3 and 1 cm ,
respectively. The degree of P=O stretching frequency shift is also in good agreement with the
order of catalytic activity of alkali metal phosphates. It is also worth to mention here that the
P=O stretching frequency of alkali metal phosphate decreases with the increase of alkali metal
-1
ion size. The P=O stretching frequencies appeared at 1015, 1009, and 991 cm for Li
Na PO , and K PO , respectively.
The interactions of K PO with HA were further investigated in the presence of methanol.
3 4
PO and
3
4
3
4
3
4
However, the N-H peak shifts were not clearly distinguished due to the peak overlap of N-H and
O-H stretching frequencies, and also possibly to the proton exchange. Moreover, the change in
3 4
P=O stretching frequency of K PO could not be observed due to the presence of strong C-O
-1
stretching frequency of methanol at 1022 cm (see Supplementary data, Fig. S1).
In situ FT-IR study was also carried out in a FT-IR Reaction Cell Top Plate (see
3
Supplementary data, Fig. S2 for a photo of the cell) with a reaction mixture of CH OH (0.5 g),
o
HA (1 g) and K
3
PO
4
(0.5 g) at 140 C under 4.0 MPa CO. Fig. 3 shows a progress of the
carbonylation reaction with time. The appearance of carbonyl peak of hexyl amide was clearly
observed with time. However, the change in absorption peaks associated with P=O and N-H
stretching and frequencies were difficult to observe due to the peak overlaps and to the low
3 4
solubility of K PO in the reaction media.
For this reason, we have attempted to characterize the catalyst recovered from the
o
carbonylation reaction. To this end, carbonylation reaction was conducted at 140 C and 4.0 MPa
3 3 4
of CO for 4 h with a mixture of HA (5.05 g, 50 mmol), CH OH (20 mL), and K PO (5 mmol,
1
.06 g). After the completion of the reaction, solid precipitates were isolated, washed with
1
6 | P a g e

o
CH
FT-IR spectrum of the recovered solid with that of fresh K
PO was transformed into unknown species, presumably K
3
OH, and dried under a vacuum at 100 C for 1 h. As shown in Fig. 4, the comparison of the
PO suggests that small portion of
HPO
3
4
K
3
4
2
4
.
3
.6.2. XRD analysis
To have a clue on the nature of the unknown species, X-ray diffraction (XRD) analysis was
conducted with the recovered solid. As can be seen in Fig. 5, XRD powder patterns strongly
suggest that the solid contains mostly consists of K
3
PO
4
and small amounts of K
(mol K PO /mol K HPO
HPO may imply that small portion
by the interaction with CH
2
HPO
4
on the
basis of JCPDS. In fact, a mixture of K PO and K HPO
3
4
2
4
3
4
2
4
= 5) showed
similar XRD patterns to that of the solid. The presence of K
of K PO is converted into K HPO and KOCH
experimental condition.
2
4
3
4
2
4
3
3
OH under the
3
.7. Recycle test
The catalyst recycle was conducted with CHA in the presence of K
◦
3
PO
4
at 140 C and 4
MPa of CO for 4 h. After the completion of the reaction, the liquid product mixture was removed
from the reactor by means of a porous metal filter, and the remaining solid catalyst mixture was
3
reused for the next cycle with a fresh charge of CHA and CH OH (20 mL). As listed in Table 7,
the CHF yield was found to decrease with the cycle, but the drop in yield of CHA after five
3 4
cycles was as small as 9%, demonstrating that K PO is fairly stable under the experimental
condition.
One may suspect that the carbonylation of amines could proceed in a homogeneous way by
dissolved potassium species in CH OH. The possible formation of homogeneous species was
3
1
7 | P a g e

o
3 4 3
investigated by treating K PO (5 mmol) with 4.0 MPa of CO in CH OH (20 mL) at 140 C for 4
h in a 100 mL stainless steel bomb reactor equipped with a porous metal filter, which was
connected to a valve attached to the sampling port of the reactor. After the period of the reaction,
o
the reaction mixture was filtered at 140 C under CO pressure by opening the valve connected to
the metal filter and the sampling port. The resulting transparent filtrate was loaded into another
o
1
4
00 mL stainless steel bomb reactor containing CHA (100 mmol), and carbonylated at 140 C for
2
h under 4.0 MPa of CO pressure. GC analysis of the product mixture showed that the yield of
CHF was 6%, which is practically the same as that obtained in the absence of a catalyst. This
may indicate that the amount of dissolved potassium species in CH OH under CO pressure is
negligible or the dissolved species is not active for the carbonylation of CHA. Nonetheless, the
loss of activity with the cycle strongly suggests that K PO is slowly transformed into less active
species, most plausibly K HPO with the progress of the cycles.
3
3
4
2
4
3
.6. Computational calculations and proposed mechanism
Fig. 6 shows the optimized structures of ground state (reactant), transition state,
intermediate, and product for the interactions among methylamine, K PO , and CO. The relative
energies to that of the ground state were calculated for each state. As can be seen in Fig. 6, the
optimized structure of the first transition state (TS1) shows that the amino group of CH NH
via a hydrogen bond (O···H interaction distance
3
4
3
2
3
-
interacts strongly with an oxygen atom of PO
4
of 1.32 Å). A strong interaction between the amino group and CO is also observed (N···C
interaction distance of 1.73 Å), implying that the nucleophilicity of the methylamine is greatly
3
-
enhanced by the hydrogen bond formation with PO
of the ground state (reactant), any interaction is not observed between methylamine and CO. This
8 | P a g e
4
. As can be seen in the optimized structure
1

is reasonable because the nucleophilicity of methylamine is not sufficiently high enough to
interact with a neutral molecule, CO. Therefore, the observation of a strong interaction between
the amino group and CO in the optimized structure of TS1 (N···C interaction distance of 1.73 Å)
indicates that the nucleophilicity of the methylamine is greatly enhanced by the hydrogen bond
3
-
-1
-1
4
formation with PO . The activation energy (∆
G‡TS1) of 16.0 kcal mol is lower by 4.5 kcal mol
3 3 2
compared with that of the CH OK-catalyzed carbonylation of CH NH [35]. In the optimized
structure of the intermediate shown in Fig 6, the formations of N-C and O-H bonds are clearly
3
-
observed, implying that PO
energy of formation of the intermediate (
formation is assisted by the strong proton accepting ability of PO
seems to transform immediately into the product via the second transition state (TS2) because the
4
could function as a proton mediator. The relatively low Gibbs free
-1
∆
G = -5.7 kcal mol ) suggests that the N-C bond
I
3
-
4
. The intermediate species
-
1
activation energy for TS2 (
∆
G‡TS2) is calculated as low as -0.2 kcal mol . In the optimized
3
-
structure of the product shown in Fig. 6, the proton transfer from the protonated PO
4
to the
carbonyl carbon atom is evident. Such a facile proton transfer can be attributed to the bi-
3
-
functional roles of PO
4
as proton acceptor and donor. The Gibbs free energy of formation of the
-1
product ( ) is calculated as -24.6 kcal mol . In consideration of ∆ , ∆ , ∆G , and ∆G
∆
G
p
G‡TS1 G‡TS2
I P
values, it is conceivable that the carbonylation of amines could proceed well in the presence of
PO
K
3
4
.
For comparison, computational calculation was also conducted for the carbonylation
reaction of amines in the presence of Li PO . Fig. 7 shows that Li PO -catalyzed carbonylation
proceeds in a similar manner to that catalysed by K PO . However, the activation energies
3
4
3
4
3
4
1
9 | P a g e

‡
G
-1
‡
G
-1
(
∆
TS1 = 27.4 kcal mol ,
∆
TS2 = 11.2 kcal mol ) and the Gibbs free energy of formation of the
-
1
-1
intermediate (
respectively, compared with those for the K
energies for the Li PO -catalyzed carbonylation can be ascribed to the extremely low activity of
Li PO
∆
G
I
= 2.0 kcal mol ) are significantly higher by 11.4, and 7.7 kcal mol ,
3
PO -catalyzed carbonylation. Such high activation
4
3
4
3
4
.
4
. Conclusion
PO was proven to be a highly active catalyst for the carbonylation of amines to produce
K
3
4
corresponding formamides in excellent yields, especially for unhindered aliphatic primary
amines and heterocyclic aliphatic amines without an additional electronegative atom. The
reactivity of amines toward the carbonylation is strongly dependent on the basicity and steric
crowding around amino groups. The reactivity of an amine increases with increasing basicity and
also with decreasing the steric hindrance around the amino group.
Quantum mechanical calculations on the K
3
4 3 2
PO -catalyzed carbonylation of CH NH reveal
3
-
that PO
bonding interactions, PO
CH NH to the carbonyl carbon to form N-methylformamide.
4
function as a proton shuttle as well as an amine activator. Through strong hydrogen
3
-
4
is capable of mediating proton transfer from the amino group of
3
2
Acknowledgement
This work was supported by the “Fusion Research Program for Green Technologies
NRF-2012M3C1A1054497)” through the National Research Foundation of Korea (NRF)
funded by Ministry of Education, Science and Technology. This work was also supported
(
2
0 | P a g e

by Korea CCS R&D Center grant funded by the Ministry of Science, ICT and Future
Planning (2014M1A8A1049262).
Appendix A. Supplementary data
Supplementary data associated with this article can be found in the online version at
http://dx.doi.org/10.1016/j.apcata.XXXXXXX
2
1 | P a g e

References
[
1]
2]
K. Weisseermel, H.-J. Arpe, Industrial Organic Chemistry, fourth ed., Wiley-VCH, 2003.
H. Bipp, H. Kieczka, Ullmann's Encyclopedia of Industrial Chemistry, 5 ed.,
Weinhein:VCH, 1989.
[
[3]
[4]
[5]
G.M. Radwan, S.A. AlMuhtaseb, M.A. Fahim, Fluid Phase Equil. 129 (1997) 175-186.
B.P. Baruah, P. Khare, Energy & Fuels 21 (2007) 2156-2164.
H.-W. Häring, Industrial gases processing, Wiley-VCH Verlag GmbH & Co, KGaA,
Weinheim, 2008.
[
6]
7]
S.H. Ali, D.M. Hamad, B.H. Albusairi, M.A. Fahim, Energy & Fuels 23 (2009) 5986-
5
994.
[
A.A. Gaile, V.N. Chistyakov, L.L. Koldobskaya, V.V. Kolesov, Chem. Technol. Fuels
Oils 48 (2012) 187-194.
[
[
[
[
[
[
[
[
[
[
[
8]
R. Seidler, H. Klein, I. G. Farben, DE 714311 (1937).
9]
G.M. Blackburn, W.P. Jencks, J. Am. Chem. Soc., 90 (1968) 2638-2645.
10] J.-G. Kim, D.O. Jang, Synlett 2010 ( 2010) 2093–2096.
11] X. Yang, Y. Zhang, Res. Chem. Intermed. (2012) 1–6.
12] S.P. Pathare, R.V. Sawant, K.G. Akamanchi, Tetrahedron Lett. 53 (2012) 3259–3263.
13] M. Hong, G. Xiao, J. Fluorine Chem. 146 (2013) 11–14.
14] L. Ma'mani, M. Sheykhan, A. Heydari, M. Faraji, Appl. Catal. A: Gen. 377 (2010) 64–69.
15] N. Ortega, C, Richter, F. Glorius, Org. Lett. 15 (2013) 1776–1779.
16] H. Tumma, N. Nagaraju, K.V. Reddy, J. Mol. Catal. A: Chem. 310 (2009) 121–129.
17] A. Mullen, New Synthesis with Carbon Monoxide, Springer-Verlag, Berlin, 1980.
18] B.D. Dombek, R.J. Angelici, J. Catal. 48 (1977) 433-435.
2
2 | P a g e

[
[
[
[
[
[
[
[
[
[
[
[
[
[
19] C.W. Bird, Chem. Rev. 62 (1962) 283-302.
20] T. Saegusa, S. Kobayashi, K. Hirota, Y. Ito, Tetrahedron Lett. 7 (1966) 6125-6129.
21] Y. Yoshida, S. Asano, S. Inoue, Chem. Lett. 13 (1984) 1073-1076.
22] T. Saegusa, S. Kobayashi, K. Hirota, Y. Ito, Bull. Chem. Soc. Jpn. 42 (1969) 2610-2614.
23] D. Durand, C. Lassau, Tetrahedron Lett. 10 (1969) 2329-2330.
24] J.J. Byerley, G.L. Rempel, N. Takebe, J. Chem. Soc. D (1971) 1482-1483.
25] Y. Tsuji, T. Ohsumi, T. Kondo, Y. Watanabe, J. Organomet. Chem. 309 (1986) 333-344.
26] G. Bitsi, G. Jenner, J. Organomet. Chem. 330 (1987) 429-435.
27] G. Jenner, G. Bitsi, E. Schleiffer, J. Mol. Catal. 39 (1987) 233-236.
28] G.W. Liu, M. Hakimifard, M. Garland, J. Mol. Catal. A: Chem. 168 (2001) 33-37.
29] G. Jenner, G. Bitsi, Appl. Catal. 32 (1987) 293-304.
30] C.J. Gerack, L. McElwee-White, Chem. Commun. 48 (2012) 11310-11312.
31] P.G. Jessop, Y. Hsiao, T. Ikariya, R. Noyori, J. Am. Chem. Soc. 118 (1996) 344-355.
32] C. Federsel, A. Boddien, R. Jackstell, R. Jennerjahn, P.J. Dyson, R. Scopelliti, G.
Laurenczy, M. Beller, Angew. Chem. Int. Ed. 49 (2010) 9777-9780.
[33] K. Kobayashi, T. Kikuchi, S. Kitagawa, K. Tanaka, Angew. Chem. Int. Ed. 53 (2014)
1
1813-11817.
[
34] A. Tlili, E. Blondiaux, X. Frogneux, T. Cantat, Green Chem. 17 (2015) 157-168.
35] Y.-S. Choi, Y.N. Shim, J. Lee, J.H. Yoon, C.S. Hong, M. Cheong, H.S. Kim, H.G. Jang,
J.S. Lee, Appl. Catal. A: Gen. 404 (2011) 87-92.
[
[
36] W.L.F. Armarego, C.L.L. Chai, Purification of laboratory chemicals, sixth ed., Elsevier
Inc., 2009.
[37] M.J. Frisch, Gaussian 03, Revision C. 02, Gaussian Inc., Wallingford, CT, USA, 2004.
2
3 | P a g e

[
[
[
[
[
[
[
[
[
38] P. Speers, K.E. Laidig, J. Chem. Soc., Perkin Trans. 2 (1994) 799-806.
39] T. Marino, N. Russo, M. Toscano, J. Phys. Chem. B 107 (2003) 2588-2594.
40] A. Tada, Bull. Chem. Soc. Japan 48 (1975) 1391-1393.
41] P. Ballinger, F.A. Long, J. Am. Chem. Soc. 82 (1960) 795–798.
42] E.F. da Silva, H.F. Svendsen, Ind. Eng. Chem. Res. 42 (2003) 4414-4421.
43] K.Z. Sumon, A. Henni, A.L.L. Ease, Ind. Eng. Chem. Res. 51 (2012) 11924-11930.
44] J. Bjerrum, Stability Constants, Chemical Society, London, 1958.
45] N.M. Rocher, R. Frech, M. Khan, J. Phys. Chem. B 109 (2005) 20697-20706.
46] K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coordination Compounds,
foutth ed., John Wiley & Sons Inc., 1986.
2
4 | P a g e

Figure Captions
Fig. 1. FT-IR spectra showing the change in N-H stretching frequency of n-hexylamine (HA) by
3 4 3 4
the interaction with alkali metal phosphate: (a) HA, (b) HA + K PO , (c) HA + Na PO and (d)
HA + Li PO
3
4
.
2
5 | P a g e

3 4 3 4 3 4 3 4 3 4
Fig. 2. FT-IR spectra of (a) K PO , (b) K PO + HA, (c) Na PO , (d) Na PO + HA, (e) Li PO
and (f) Li PO + HA.
3
4
2
6 | P a g e

o
Fig. 3. In situ FT-IR spectra showing the progress of the reaction with time at 140 C and 4.0
MPa of CO: (a) 0 min, (b) 10 min, (c) 20 min and (d) 30 min. Carbonylation reaction was
o
conducted at 140 C and 4.0 MPa CO with a mixture of HA (1 g), CH
3
OH (0.5 mL), and K
3
PO
4
(0.5 g).
2
7 | P a g e

Fig. 4. FT-IR spectra of (a) recovered catalyst, (b) K
3
4 2 4
PO and (c) K HPO . Catalyst was
o
recovered from the carbonylation reaction conducted at 140 C and 4.0 MPa of CO for 4 h with a
3 3 4
mixture of HA (5.05 g, 50 mmol), CH OH (20 mL) and K PO (5 mmol, 1.06 g).
2
8 | P a g e

Fig. 5. Comparison of XRD patterns of the recovered catalyst with those of K
3
4 2 4
PO and KH PO :
(
a) recovered catalyst, (b) K
2
HPO
4
and (c) K PO . Catalyst was recovered from the
3 4
o
carbonylation reaction conducted at 140 C and 4.0 MPa of CO for 4 h with a mixture of HA
3 3 4
(5.05 g, 50 mmol), CH OH (20 mL), and K PO (5 mmol, 1.06 g).
2
9 | P a g e