Decarboxylative palladium(II)-catalyzed synthesis of aryl amidines from aryl carboxylic acids: Development and mechanistic investigation

Article abstract of DOI:10.1002/chem.201301809

A fast and convenient synthesis of aryl amidines starting from carboxylic acids and cyanamides is reported. The reaction was achieved by palladium(II)-catalysis in a one-step microwave protocol using [Pd(O ₂CCF₃)₂], 6-methyl-2,2'-bipyridyl and trifluoroacetic acid (TFA) in N-methylpyrrolidinone (NMP), providing the corresponding aryl amidines in moderate to excellent yields. The protocol is very robust with regards to the cyanamide coupling partner but requires electron-rich ortho-substituted aryl carboxylic acids. Mechanistic insight was provided by a DFT investigation and direct ESI-MS studies of the reaction. The results of the DFT study correlated well with the experimental findings and, together with the ESI-MS study, support the suggested mechanism. Furthermore, a scale-out (scale-up) was performed with a non-resonant microwave continuous-flow system, achieving a maximum throughput of 11 mmol h^-1 by using a glass reactor with an inner diameter of 3 mm at a flow rate of 1 mL min ^-1. Aryl amidines with ease: A Pd^II-catalyzed decarboxylative synthesis of aryl amidines from aryl carboxylic acids and cyanamides was developed and investigated with the aid of ESI-MS studies and DFT calculations (see scheme).

Full text of DOI:10.1002/chem.201301809

FULL PAPER
DOI: 10.1002/chem.201301809
Decarboxylative Palladium(II)-Catalyzed Synthesis of Aryl Amidines from
Aryl Carboxylic Acids: Development and Mechanistic Investigation
Jonas Rydfjord,[a] Fredrik Svensson,[a] Alejandro Trejos,[a] Per J. R. Sjçberg,[b]
Christian Skçld,[a] Jonas Sꢀvmarker,[a] Luke R. Odell,[a] and Mats Larhed*[a]
Dedicated to Gunilla Eriksson on the occasion of her 65th birthday
Abstract: A fast and convenient syn-
thesis of aryl amidines starting from
carboxylic acids and cyanamides is
N
ed well with the experimental findings
and, together with the ESI-MS study,
support the suggested mechanism.
ACHTUNGTERNNFUNG urthermore, a scale-out (scale-up)
was performed with a non-resonant mi-
crowave continuous-flow system, ach-
A
a
A
ieving
a
maximum throughput of
trifluoroacetic acid (TFA) in N-methyl-
pyrrolidinone (NMP), providing the
corresponding aryl amidines in moder-
ate to excellent yields. The protocol is
very robust with regards to the cyan-
11 mmolhÀ1 by using a glass reactor
with an inner diameter of 3 mm at a
flow rate of 1 mLminÀ1.
Keywords: decarboxylation · densi-
ty functional calculations · mass
spectrometry · microwave chemis-
try · palladium
Introduction
the formation of the arylpalladium species. Aryl carboxylic
acids also present some distinct advantages over organobor-
on derivatives, the most commonly employed substrates in
palladium(II)-catalyzed cross-couplings, as they offer a more
environmentally benign and less-toxic class of arylating
agents.[14] Additionally, sterically congested aryl carboxylic
acids are often excellent substrates in palladium-catalyzed
transformations as ortho-substituents are known to facilitate
the decarboxylation process.[15]
Compounds with the amidine motif are of considerable
interest in drug discovery and have been indicated as poten-
tial agents for the treatment of Alzheimerꢀs disease,[16] ma-
laria,[17] and as inhibitors of acid-sensing ion channels,[18] pla-
telet aggregation[19] and serine proteases.[20] Amidines are
also useful precursors for the synthesis of a wide variety of
heterocyclic ring systems such as quinazolines,[21,22] quinazo-
linones,[23] pyrimidines,[24,25] triazoles,[26] and benzimida-
zoles.[27] Amidines are traditionally prepared by the nucleo-
philic addition of an amine to a nitrile[28] and they are also
accessible by nucleophilic substitution of thioamides and
imidates.[29] More recent methodologies include Pd0-cata-
lyzed three-component methods,[30–33] addition of alkylchlor-
oaluminum amides to nitriles,[34] aryne insertion into thiour-
eas[35] and ytterbium-catalyzed addition of amines to ni-
triles.[36]
During the past few years aryl carboxylic acids have
emerged as synthetically useful aryl-metal precursors.[1,2]
The formation of the aryl-metal species occurs through a
metal-mediated decarboxylative process, which, although
long known,[3] has only recently captured the attention of
synthetic chemists. This has led to the development of nu-
merous decarboxylative palladium(II)-catalyzed cross-cou-
pling,[1,4–9] 1,2-addition[10,11] and Heck-type reactions.[12,13]
Carboxylic acids are attractive coupling partners due to
their low cost, wide-spread availability, inert nature and
most importantly, CO2 is the only byproduct produced in
[a] J. Rydfjord, F. Svensson, Dr. A. Trejos, Dr. C. Skçld,
Dr. J. Sꢁvmarker, Dr. L. R. Odell, Prof. Dr. M. Larhed
Organic Pharmaceutical Chemistry
Department of Medicinal Chemistry, Uppsala University
Box-574, 751 23 Uppsala (Sweden)
E-mail: mats.larhed@orgfarm.uu.se
[b] Dr. P. J. R. Sjçberg
Department of Chemistry–BMC, Uppsala University
BMC, Box-599, 751 23 Uppsala (Sweden)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201301809.
We[11,37] and others[38–44] have previously developed palla-
dium(II)-catalyzed protocols for the 1,2-carbopalladation of
nitrile derivatives, generating aryl ketones through a keti-
mine intermediate (Scheme 1a). Recently, we demonstrated
that aryl amidines could also be conveniently accessed
ꢂ 2013 The Authors. Published by Wiley-VCH Verlag GmbH & Co.
KGaA. This is an open access article under the terms of Creative
Commons the Attribution Non-Commercial NoDerivs License, which
permits use and distribution in any medium, provided the original
work is properly cited, the use is non-commercial and no modifica-
tions or adaptations are made.
Chem. Eur. J. 2013, 19, 13803 – 13810
ꢂ 2013 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
13803

Table 1. Selecting the optimal reaction conditions.
Entry
Pd
[%]
Solvent
Yield
[%][a]
1
2
3
4
5
6
7
8
2
2
2
2
2
4
8
1
2
2
0
2
2
dioxane
toluene
DMF
DMA
NMP
NMP
NMP
NMP
NMP
NMP
NMP
NMP
NMP
68
<15[b]
88
92
96
98
96
94
92
42
Scheme 1. Palladium(II)-catalyzed protocols for the 1,2-carbopalladation
of nitrile derivatives.
9[c]
10[d]
11[e]
12[f]
13[g]
through a palladium(II)-catalyzed 1,2-carbopalladation of
cyanamides with aryltrifluoroborates (Scheme 1b).[45] To fur-
ther extend the utility of this direct amidine synthesis we de-
cided to investigate the viability of carboxylic acids as the
arylpalladium precursor. Herein, we present a fast and con-
venient protocol for the preparation of aryl amidines from
aryl carboxylic acids and cyanamides (Scheme 1c).
not detected
<15[b]
92 (11 mmolhÀ1
)
[a] Yield of the isolated product (>95% pure by 1H NMR spectroscopic
analysis). Reaction conditions: [Pd(O2CCF3)2], 6-methyl-2,2’-bipyridyl (4;
4/[Pd(O2CCF3)2], 1.5:1), TFA (1 mmol), benzoic acid 1a (1.1 mmol), cy-
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
anamide 2a (1 mmol) and solvent (3 mL), were MW heated in a sealed
vial at 1208C for 30 min; [b] yield determined by 1H NMR spectroscopy;
[c] 1.0 mmol of 1a; [d] no ligand (4) added; [e] no [PdACHTUNRGTNEUNG(O2CCF3)2] or
ligand (4); [f] no TFA; [g] continuous-flow scale-out example, 1 mLminÀ1
of the reaction mixture corresponding to 1 min in the heated zone (tem-
perature set at 1408C, 0.2m in NMP of yield-determining 2a, 1.1 equiv of
1a), the yield is based on the work-up of an aliquot of 5 mL with a theo-
retical yield of 1 mmol of 3a.
Results and Discussion
As a starting point for our investigation we conducted a sol-
vent screen by using our previously developed catalytic
system for aryl amidine synthesis from aryl trifluoroborates
([PdACHTUNGTRENNUNG(O2CCF3)2], 6-methyl-2,2’-bipyridyl (4) as the ligand
and trifluoroacetic acid (TFA) as the proton donor).[45] 2,4,6-
Trimethoxybenzoic acid (1a) and N-cyanopiperidine (2a)
were chosen as model substrates and the catalyst and ligand
loadings were set at 2 and 3%, respectively. The reactions
were then microwave (MW) heated[46] at 1208C for 30 min
using the appropriate solvent and 1 equivalent of TFA in
sealed vessels. As can be seen from Table 1, polar aprotic
solvents performed best, with N-methylpyrrolidinone
(NMP) giving 96% yield of the isolated product (Table 1,
entry 5), whereas dimethylformamide (DMF) and dimethy-
lacetamide (DMA) gave slightly lower yields of 88 and 92%
(Table 1, entries 3 and 4), respectively. The use of dioxane
gave the product in a reduced yield of 68% (Table 1,
entry 1), whereas toluene afforded less than 15% yield
(Table 1, entry 2).
Having identified NMP as the solvent of choice we next
set about exploring the remaining parameters (temperature,
time, palladium loading, ligand loading, and stoichiometry).
Table 1 reveals that the reaction between 1a and 2a is a
robust process as the majority of changes had only a minor
influence on the reaction outcome. Control experiments per-
formed without the addition of a ligand (Table 1, entry 10),
of 42%. Based on these results we chose our original NMP
conditions for further investigation (Table 1, entry 5).
Next, the scope of the addition reaction with regards to
the cyanamide partner was investigated and the results are
presented in Table 2. Cyanamide 2b furnished the primary
aryl amidine 3b in 76% yield (Table 2, entry 2) and tert-
butyl cyanamide 2c gave the corresponding aryl amidine 3c
isolated in an excellent yield of 90% (Table 2, entry 3)
showing that N,N-disubstitution is not required. Dimethyl
and diethyl cyanamide (2d and 2e) also performed well,
providing 93 and 98% of 3d and 3e, respectively (Table 2,
entries 4 and 5). The bulkier cyanamide derivative 2 f fur-
nished a lower yield of 3 f (64%, Table 2, entry 6), presuma-
bly due to unfavorable steric effects. Cyclic cyanamides 2g
and 2h were also effective substrates, affording 3g and 3h
in 93 and 74% yield, respectively (Table 2, entries 7 and 8).
Dibenzyl derivative 2i reacted smoothly, providing 68% of
3i (Table 2, entry 9). The reaction was unsuccessful with car-
boxybenzyl-protected substrate 2j and only trace amounts
of the hydrolyzed product 3b were detected by HPLC-MS.
Next, we extended the scope of the carboxylic acid sub-
strate to include electron-rich di- or trimethoxy-substituted
benzoic acid derivatives 1b–d. Unfortunately, these sub-
strates proved to be much less reactive than 1a, requiring
both elevated temperature (1408C) and an extended reac-
tion time (60 min) to achieve full conversion. Substrate 1b
[PdACHTUNGTRENNUNG(O2CCF3)2] (Table 1, entry 11) and TFA (Table 1,
entry 12) demonstrated that each component is crucial for a
productive reaction. Interestingly, omission of the ligand
was not deleterious and the reaction gave a moderate yield
13804
www.chemeurj.org ꢂ 2013 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Chem. Eur. J. 2013, 19, 13803 – 13810

Synthesis of Aryl Amidines from Aryl Carboxylic Acids
FULL PAPER
Table 2. Scope of 2,4,6-trimethoxybenzoic acid with different cyanamides.
The scope of the cyanamides was also investigat-
ed further with carboxylic acid 1b (Table 3). The
substrate was effectively reacted with various cyan-
AHCTUNGTREGUNaNN mides, however, the yields were in general lower
than that obtained with 1a, with the exception of
cyanamides 2 f and 2i, both affording 74% yield of
the corresponding products 3r and 3u (Table 3, en-
tries 8 and 11; compare Table 2, entries 6 and 9).
Finally, a scale-out (scale-up) of the reaction
Entry
1
R
Product
Yield
[%][a]
96
76
90
93
98
64
93
74
68
(Figure 1), utilizing
a non-resonant microwave
system for continuous-flow organic synthesis,[49–52]
was performed. With this new equipment for direct
scale-out we were able to achieve a high throughput
of 11 mmolhÀ1 (corresponding to 3 ghÀ1). The ex-
periment used a stock solution with 0.2m of yield-
determining 2a, compound 1a (1.1 equiv), along
2
3
4
5
6
7
8
9
with [PdACTHNUGTRNEUNG(O2CCF3)2] (2%), compound 4 (3%), and
TFA (1 equiv) in N-methylpyrrolidinone (NMP).
After a minor adjustment of the reaction condi-
tions, we achieved an excellent yield of 92% of 3a
(Table 1, entry 13) at 1 mLminÀ1 and 1208C, which
corresponds to 1 min in the heated zone.
The reaction of 1a proceeds smoothly with a
range of different cyanamide derivatives, containing
bulky groups, small groups, and unsubstituted cyan-
amide. However, the carboxylic acid substrate
scope is limited and requires electron-rich ortho-
substituted aryl carboxylic acids. Somewhat disap-
pointed with this narrow scope we proceeded to in-
vestigate the reaction mechanism by using density
functional theory (DFT) calculations to better un-
derstand our experimental findings.
Previously we have investigated the reaction path
and mechanism for the 1,2-carbopalladation of ni-
triles with aryl carboxylic acids.[11,53] Based on our
results from the nitrile study[11] we suggest that ami-
dine formation occurs through the catalytic path-
way outlined in Figure 2. Starting with the ligand
chelated PdII-complex A, coordination of the aryl
carboxylic acid 1 generates complex B. Next, decar-
boxylation occurs to form arylpalladium intermedi-
ate C, followed by coordination of the cyanamide
to form complex D. 1,2-Carbopalladation of the cy-
anamide generates complex E, followed by proto-
nation to afford the free amidine product 3 and a
catalytically active PdII-species. DFT calculations
were then performed, based on the proposed mech-
10
not observed
1
[a] Yield of the isolated product (>95% pure by H NMR spectroscopic analysis). Re-
action conditions: [Pd(O2CCF3)2] (0.02 mmol), 4 (0.03 mmol), TFA (1 mmol), benzoic
acid 1a (1.1 mmol), cyanamides 2a–j (1 mmol), and NMP (3 mL) were MW heated in
a sealed vial at 1208C for 30 min.
ACHTUNGTRENNUNG
provided a good yield (65%) of the corresponding amidine
3k (Table 3, entry 1), whereas 1c and 1d furnished yields of
only 9 and 8%, respectively (Table 3, entries 2 and 3). Sev-
eral other carboxylic acid derivatives known to undergo de-
carboxylative coupling reactions were also tested,[47] howev-
er, only traces of the desired products were detected by
HPLC-MS. Finally, the addition of silver or copper salts
were evaluated in an attempt to promote decarboxylation,[1]
and, disappointingly, these additives did not improve the
outcome of the reaction.[48]
anism in Figure 2. To shed light on the narrow substrate
scope, the calculations focused on comparing the energy
profiles for the most electron-rich and the least electron-rich
benzoic acid analogues in the series, 1a and 1c, respectively.
Compound 1b was also included to represent a benzoic acid
with electronic properties in between 1a and 1c. Only the
di-ortho-substituted carboxylic acids were included to avoid
having to account for the ortho-effect when interpreting the
results. In the calculations [Pd
palladium source, compound 4 as the ligand, and 2b was
(OAc)2][54] was used as the
ACHTUNGTRENNUNG
Chem. Eur. J. 2013, 19, 13803 – 13810 ꢂ 2013 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemeurj.org
13805

M. Larhed et al.
Table 3. Scope of various carboxylic acid derivatives with different cyanamides.
identified as the lowest-energy complex prior to de-
carboxylation. Thus, complex II was chosen as the
starting point of the catalytic cycle and the energies
of all catalytic intermediates are reported relative
to this. To proceed with decarboxylation, one coor-
dination site needs to be made available. Dissocia-
tion of one anionic ligand to the cationic Pd com-
plex III was associated with a significant increase in
energy, approximately half of the required energy
for the decarboxylation process, which may partly
explain the benefits of a polar solvent reaction
medium in the reaction. A change in binding mode
of 1 giving organopalladium cation IV provides a
suitable arrangement for decarboxylation over tran-
sition state TS-I. The DFT calculations show that
the required free-energy of decarboxylation is
110.3 kJmolÀ1 for 1a, compared with 138.5 kJmolÀ1
for 1c, which suggests that decarboxylation of 1a is
significantly more facile than for 1c. From the
formed arylpalladium complex V dissociation of
carbon dioxide gives s-complex VI. Association of
1 to this complex furnishes the most stable inter-
mediate VII prior to the 1,2-carbopalladation step.
To initiate 1,2-carbopalladation, compound 1 was
exchanged with 2b to give the ionic species VIII.
As in the case for decarboxylation, charge separa-
tion was calculated to account for approximately
half the required energy for the 1,2-carbopallada-
tion process. From VIII 1,2-carbopalladation via
transition state TS-II leads to the desired aryl ami-
dine IX. When comparing the energy profiles of 1a
and 1c, the former has a required free-energy of
110.8 kJmolÀ1 for 1,2-carbopalladation, which is
substantially lower than the 149.1 kJmolÀ1 calculat-
ed for 1,2-carbopalladation using 1c. Association of
1 gives the neutral complex X and a subsequent
product release by replacement of with 1, or in the
case of 1b, acetate, regenerates complex II and
gives an overall exergonic reaction.
Entry
1
Ar
Cyanamide
Product
Yield
[%][a]
2a
65
2
2a
9[b]
3
2a
2b
2c
2d
2e
2 f
2g
2h
2i
8[b]
36[b]
52
4
1b
1b
1b
1b
1b
1b
1b
1b
5
6
37[b]
7
48
8
74
9
63
Overall, the reaction profiles support a substan-
tially faster reaction rate for 1a compared with 1c
and could explain the more favorable experimental
outcome of this substrate. Comparison of the Gibbs
energy profiles employing 1a and 1b, shows a much
more subtle difference between these two ana-
logues. Decarboxylation proceeds with very similar
energy requirements, 110.3 and 111.1 kJmolÀ1, re-
spectively; however, the 1,2-carbopalladation step
for 1a is slightly more favored (110.8 kJmolÀ1)
compared with 1b (117.9 kJmolÀ1). For all three
benzoic acid analogues examined, the 1,2-carbopal-
10
11
32[b]
74
1
[a] Yield of the isolated product (>95% pure by H NMR spectroscopic analysis). Re-
action conditions: [Pd(O2CCF3)2] (0.08 mmol), 4 (0.12 mmol), TFA (1 mmol), benzoic
ACHTUNGTRENNUNG
acids 1b–d (1.1 mmol), cyanamides 2a–i (1 mmol), and NMP (3 mL) were MW heated
in a sealed vial at 1408C for 60 min; [b] purified by using preparative HPLC.
used as cyanamide. The results of the calculations are shown
in Figure 3 and relative energies for all complexes are pro-
vided in the Supporting Information.
The calculations start with the diacetate-chelated species I
with an associated ligand. Exchange of one acetate for 1b or
both acetates for 1a or 1c gave intermediate II, which was
ladation was found to be the rate-determining step. These
calculations, together with the experimental results, support
the requirement for an electron-rich benzoic acid analogue
in the current reaction.
The overall reaction path is highly similar to the corre-
sponding decarboxylative addition of benzoic acid analogues
13806
www.chemeurj.org ꢂ 2013 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Chem. Eur. J. 2013, 19, 13803 – 13810

Synthesis of Aryl Amidines from Aryl Carboxylic Acids
FULL PAPER
to acetonitrile.[53] The decarbox-
ylation process is identical and
1,2-carbopalladation using cy-
anamide is calculated to pro-
ceed in an analogous fashion
compared with the reaction
using acetonitrile. The 1,2-car-
bopalladation transition state is
presented in Figure 4. In the
previously reported investiga-
tion of nitriles, it was establish-
ed that a more polar reaction
media would be favorable.[53]
This was experimentally veri-
fied by using an increased
amount of water, which also
was needed for the hydrolysis
step. The same conclusion
should be valid in the present
study and a more polar reaction
media should reduce the energy
cost for charge separation. This
protocol is, however, incompati-
ble with water and to some
extent alcohols due to a com-
peting urea byproduct forma-
tion. In addition, it was ob-
served in the previous investi-
gation that a significant byprod-
uct when including water in the
reaction is the decarboxylated
aryl product, which would also
be a source for the decreased
yield in the reaction.[53]
Figure 1. Top: schematic illustration of continuous-flow setup utilizing a non-resonant MW-cavity for heating
of the reaction mixture. Bottom left: inside the reactor cavity, the glass reactor and MW antenna are displayed.
Bottom right: closed reactor cavity.
To further validate and sup-
port the proposed mechanism
and the assumptions made for the DFT calculations we per-
formed a mass spectrometry (MS) study by using electro-
spray ionization (ESI). ESI is considered a soft mass spec-
trometric ionization technique and is therefore a valuable
tool for analyzing sensitive organometallic species.[55,56] In
addition, this technique also allows for the identification of
palladium complexes present in an ongoing reaction. The re-
action samples were removed for ESI-MS analysis after ap-
proximately 30–70% conversion. By monitoring the pres-
sure developed during the reaction, due to the carbon diox-
ide liberation, we were able to abort the reaction once it
was midway through. ESI-MS detection afforded several
cationic complexes that were readily identified as single pal-
ladium species based on the isotopic pattern of palladium.
However, to determine the composition of these palladi-
AHCTUNGTREGuNNNU m(II) complexes we performed additional studies by sub-
stituting one component at a time.
This strategy allowed us to identify four cationic species
that were present in all of the reactions (Figure 5 and the
Supporting Information). The suggested structures were as-
signed and classified according to the labels in the reaction
Figure 2. Proposed catalytic cycle as adapted from Lindh et al.[11]
Chem. Eur. J. 2013, 19, 13803 – 13810 ꢂ 2013 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemeurj.org
13807

M. Larhed et al.
Figure 3. Calculated Gibbs energy profile of the reaction of carboxylic acids 1a (red), 1b (blue), and 1c (black) with cyanamide 2b to give product 3.
Full lines indicate that the complexes are verified to be directly connected. For 1a and 1c, complex II has two coordinated aryl acids, whereas for 1b it
has one acid and one acetate.
pathway (Figure 2). Suggestions were supported by MS/MS
studies along with neutral loss experiments. One interesting
finding was that the intermediate with palladium, ligand,
and carboxylic acid (B in Figure 2) produced a neutral loss
of 44 Da, corresponding to the molecular mass of carbon di-
oxide. This finding suggests an alternate structure corre-
sponding to the same molecular mass as B, in which the de-
carboxylation has already occurred and carbon dioxide is
bonded to the palladium center (Figure 5). Alternatively, de-
carboxylation may occur in the mass spectrometer and the
neutral loss of 44 Da is due to this process rather than an
isolated complex. Examining the complexes that were con-
sistently found in all reactions we were able to identify
three out of the four suggested intermediates in the catalytic
cycle (Figure 2). The complex prior to decarboxylation, B,
was identified along with the arylpalladium complex C that
is formed after decarboxylation. In addition, the s-complex
D, which is formed upon coordination of cyanamide to C,
was identified by neutral loss experiments. Complex E,
which is formed upon 1,2-carbopalladation, would have the
same molecular mass as species D, but could not be conclu-
sively identified. Taken together, the findings in this ESI-MS
study are consistent with the suggested catalytic pathway.
Conclusion
We have developed a direct palladium(II)-catalyzed proto-
col for the synthesis of aryl amidines from aryl carboxylic
acids and cyanamides. The 1,2-addition reaction proceeds
smoothly with good to excellent yields using different cyan-
AHCTUNGTREGaNNUN mides and the preferred coupling partner, 2,4,6-trimethoxy
benzoic acid 1a.
Presently, the scope is limited to aryl carboxylic acids con-
taining ortho-methoxy substituents (1a–d), which is in ac-
cordance with the known limitations associated with palla-
dium(II)-catalyzed decarboxylation reactions.[1] Further-
more, a continuous-flow scale-out was conducted with a pur-
pose built non-resonant tubular MW reactor, achieving
11 mmolhÀ1 (92% yield of the isolated product) after slight
adjustment of the reaction conditions. DFT calculations and
an accompanying ESI-MS study support the proposed reac-
13808
www.chemeurj.org ꢂ 2013 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Chem. Eur. J. 2013, 19, 13803 – 13810

Synthesis of Aryl Amidines from Aryl Carboxylic Acids
FULL PAPER
tion mechanism, and the improved outcome with di-ortho-
methoxy-substituted benzoic acids.
Experimental Section
General procedure for the synthesis of aryl amidines 3a–i: [PdACHTNURTGENUNG(O2CCF3)2]
(6.6 mg, 0.02 mmol), 6-methyl-2,2’-bipyridyl (5.1 mg, 0.03 mmol), and
NMP (3.0 mL) were added to a 2–5 mL process vial and the mixture was
stirred for 2 min before cyanamide, 2, (1 mmol), benzoic acid, 1a,
(1.1 mmol), and trifluoroacetic acid (114 mg, 1 mmol) were added. The
vial was instantly capped under air and then heated by using microwave
irradiation at 1208C for 30 min. The reaction mixture was then diluted
with aq. NaHCO3 (20 mL) and washed with diethyl ether (20 mL). The
organic phase was further extracted with aq. NaHCO3 (2ꢃ20 mL). The
combined aqueous phases were basified to pHꢀ14 by the addition of aq.
NaOH and extracted with CH2Cl2 (3ꢃ30 mL). The combined organic
phases were dried over a cotton plug and then concentrated and dried in
vacuo to provide the pure isolated product in the yield stated in Tables 1,
2, and 3.
Computational details: All calculations were performed by using Jaguar
version 7.8[57] with the B3LYP[58,59] functional employing the LACVP*
basis set.[60] The complexes were optimized in the gas phase and the opti-
mized structures were subsequently subjected to a single-point energy
calculation using the PBF solvation model[61,62] with parameters suitable
for NMP (dielectric constant, e=32.2 and probe radius=2.67). The dis-
persion correction was applied by using the DFT-D3 program.[63] Vibra-
tional analysis was performed on the optimized geometry in the gas
phase and the Gibbs free-energy in solvent was calculated by adding the
gas phase thermodynamic contribution from the vibrational analysis at
373.15 K, including zero-point energy, and the dispersion correction to
the single-point solution phase energy.
Figure 4. Optimized geometry for the 1,2-carbopalladation transition
state TS-II. Bond lengths [ꢄ] between Pd and coordinated atoms as well
À
as the newly forming C C bond are presented.
The presented transition states are shown to be connected to the present-
ed proceeding and preceding energy minima, respectively. The presented
Figure 5. ESI-MS scan of reaction mixture at approximately 30–70% conversion. Spectrum from reaction conditions: [Pd
pound 4 (0.03 mmol), TFA (1 mmol), benzoic acid 1a (1.1 mmol), cyanamides 2a (1 mmol), and NMP (3 mL). Also shown are the proposed PdII inter-
mediates. The cis and trans geometries are based on the corresponding DFT calculated configurations with [Pd(OAc)2].
ACHTNUGTNER(UNNG O2CCF3)2] (0.02 mmol), com-
AHCTUNGTRENNUNG
Chem. Eur. J. 2013, 19, 13803 – 13810 ꢂ 2013 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemeurj.org
13809

M. Larhed et al.
transition states and stationary minima are verified to have one and zero
imaginary frequencies, respectively.
[30] C. G. Saluste, S. Crumpler, M. Furber, R. J. Whitby, Tetrahedron
Lett. 2004, 45, 6995–6996.
[31] K. K. R. Tetala, R. J. Whitby, M. E. Light, M. B. Hurtshouse, Tetra-
hedron Lett. 2004, 45, 6991–6994.
[32] C. G. Saluste, R. J. Whitby, M. Furber, Tetrahedron Lett. 2001, 42,
6191–6194.
[33] C. Saluste, R. Whitby, M. Furber, Angew. Chem. 2000, 112, 4326–
4328; Angew. Chem. Int. Ed. 2000, 39, 4156–4158.
[34] R. Garigipati, Tetrahedron Lett. 1990, 31, 1969–1972.
[35] K. Biswas, M. F. Greaney, Org. Lett. 2011, 13, 4946–4949.
[36] J. Wang, F. Xu, T. Cai, Q. Shen, Org. Lett. 2008, 10, 445–448.
[37] M. Behrends, J. Sꢁvmarker, P. J. R. Sjçberg, M. Larhed, ACS Catal.
2011, 1, 1455–1459.
[38] K. Garves, J. Org. Chem. 1970, 35, 3273–3275.
[39] L. Zhao, X. Lu, Angew. Chem. 2002, 114, 4519–4521; Angew. Chem.
Int. Ed. 2002, 41, 4343–4345.
[40] C. Zhou, R. C. Larock, J. Am. Chem. Soc. 2004, 126, 2302–2303.
[41] B. Zhao, X. Lu, Tetrahedron Lett. 2006, 47, 6765–6768.
[42] C. Zhou, R. C. Larock, J. Org. Chem. 2006, 71, 3551–3558.
[43] B. Zhao, X. Lu, Org. Lett. 2006, 8, 5987–5990.
[44] A. Yu, J. Li, M. Cui, Y. Wu, Synlett 2007, 2007, 3063–3067.
[45] J. Sꢁvmarker, J. Rydfjord, J. Gising, L. R. Odell, M. Larhed, Org.
Lett. 2012, 14, 2394–2397.
Acknowledgements
We thank the Swedish Research Council and the Knut and Alice Wallen-
berg Foundation for financial support. We also thank Ashkan Fardost for
assistance with the graphics for Figure 1.
[1] N. Rodrꢅguez, L. J. Goossen, Chem. Soc. Rev. 2011, 40, 5030–5048.
[2] J. D. Weaver, A. Recio, A. J. Grenning, J. A. Tunge, Chem. Rev.
2011, 111, 1846–1913.
[3] A. F. Shepard, N. R. Winslow, J. R. Johnson, J. Am. Chem. Soc.
1930, 52, 2083–2090.
[4] A. Voutchkova, A. Coplin, N. E. Leadbeater, R. H. Crabtree, Chem.
Commun. 2008, 47, 6312–6314.
[5] J.-J. Dai, J.-H. Liu, D.-F. Luo, L. Liu, Chem. Commun. 2011, 47,
677–679.
[6] K. Ouchaou, D. Georgin, F. Taran, Synlett 2010, 2083–2086.
[7] Z. Duan, S. Ranjit, P. Zhang, X. Liu, Chem. Eur. J. 2009, 15, 3666–
3669.
[8] L. J. Goossen, G. Deng, L. M. Levy, Science 2006, 313, 662–664.
[9] L. J. Goossen, N. Rodrꢅguez, B. Melzer, C. Linder, G. Deng, L. M.
Levy, J. Am. Chem. Soc. 2007, 129, 4824–4833.
[10] Y. Luo, J. Wu, Chem. Commun. 2010, 46, 3785–3787.
[11] J. Lindh, P. J. R. Sjçberg, M. Larhed, Angew. Chem. 2010, 122,
7899–7903; Angew. Chem. Int. Ed. 2010, 49, 7733–7737.
[12] A. G. Myers, D. Tanaka, M. R. Mannion, J. Am. Chem. Soc. 2002,
124, 11250–11251.
[13] P. Hu, J. Kan, W. Su, M. Hong, Org. Lett. 2009, 11, 2341–2344.
[14] L. J. Goossen, N. Rodrꢅguez, K. Goossen, Angew. Chem. 2008, 120,
3144–3164; Angew. Chem. Int. Ed. 2008, 47, 3100–3120.
[15] L. J. Goossen, W. R. Thiel, N. Rodrꢅguez, C. Linder, B. Melzer, Adv.
Synth. Catal. 2007, 349, 2241–2246.
[16] B. Ojo, P. G. Dunbar, G. J. Durant, P. I. Nagy, J. J. Huzl, S. Periyasa-
my, D. O. Ngur, A. A. El-Assadi, W. P. Hoss, W. S. Messer, Bioorg.
Med. Chem. 1996, 4, 1605–1615.
[17] A. Kimura, H. Nishikawa, N. Nomura, J. Mitsuyama, S. Fukumoto,
N. Inoue, S.-I. Kawazu, Antimicrob. Agents Chemother. 2012, 56,
2191–2193.
[18] X. Chen, B. A. Orser, J. F. MacDonald, Eur. J. Pharmacol. 2010, 648,
15–23.
[19] C. P. Cannon, Clin. Cardiol. 2003, 26, 401–406.
[20] J. Kotthaus, T. Steinmetzer, A. van de Locht, B. Clement, J. Enzyme
Inhib. Med. Chem. 2011, 26, 115–122.
[46] P. Nilsson, K. Olofsson, M. Larhed, Top. Curr. Chem. 2006, 266,
103–144.
[47] Investigating the scope of the reaction we tested several different
carboxylic acids known to undergo decarboxylative coupling reac-
tions. For instance, 2,4,6-trimethylbenzoic acid, 3-methylthiophene-
2-carboxylic acid, 2,4-dimethoxybenzoic acid, benzoic acid, 2,6-di-
methoxypyridine-3-carboxylic acid, ortho-anisic acid, 2,6-dimethoxy-
3-nitrobenzoic acid, 3-bromo-2,6-dimethoxybenzoic acid, and 2,6-di-
fluoro-4-methoxybenzoic acid were tested. However, these sub-
strates only provided, at most, traces of the corresponding products
(detected by HPLC-MS analysis of the crude reaction mixture).
[48] Addition of 2 equiv of silver trifluoroacetate or copper(I) oxide was
tested with mesitylenecarboxylic acid as the substrate without im-
proving the conversion (HPLC-MS analysis of crude reaction mix-
ture).
[49] Wavecraft, http://www.wavecraft.se.
[50] For technical notes and setup, see the Supporting Information.
[51] P. ꢆhrngren, A. Fardost, F. Russo, J.-S. Schanche, M. Fagrell, M.
Larhed, Org. Process Res. Dev. 2012, 16, 1053–1063.
[52] A. Fardost, F. Russo, M. Larhed, Chim Oggi 2012, 30, 4–6.
[53] F. Svensson, R. S. Mane, J. Sꢁvmarker, M. Larhed, C. Skçld, Orga-
nometallics 2013, 32, 490–497.
[54] The product was obtained in 91% yield by using [Pd
palladium source. Compare with Table 1, entry 5.
[55] L. S. Santos, Eur. J. Org. Chem. 2008, 235–253.
ACHTUNGRTENUN(NG OAc)2] as the
[56] P. A. Enquist, P. Nilsson, P. Sjçberg, M. Larhed, J. Org. Chem. 2006,
71, 8779–8786.
[21] Y. Ohta, Y. Tokimizu, S. Oishi, N. Fujii, H. Ohno, Org. Lett. 2010,
12, 3963–3965.
[57] Jaguar, Schrçdinger LLC, New York, 2011.
[58] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[59] A. D. Becke, J. Chem. Phys. 1993, 98, 1372–1377.
[60] P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 299–310.
[61] B. Marten, K. Kim, C. Cortis, R. A. Friesner, J. Phys. Chem. 1996,
100, 11775–11788.
[22] M. A. McGowan, C. Z. McAvoy, S. L. Buchwald, Org. Lett. 2012, 14,
3800–3803.
[23] B. Ma, Y. Wang, J. Peng, Q. Zhu, J. Org. Chem. 2011, 76, 6362–
6366.
[24] E. D. Anderson, D. L. Boger, J. Am. Chem. Soc. 2011, 133, 12285–
12292.
[62] D. J. Tannor, B. Marten, R. Murphy, R. A. Friesner, D. Sitkoff, A.
Nicholls, M. Ringnalda, W. A. Goddard III, B. Honig, J. Am. Chem.
Soc. 1994, 116, 11875–11882.
[25] E. D. Anderson, D. L. Boger, Org. Lett. 2011, 13, 2492–2494.
[26] G. M. Castanedo, P. S. Seng, N. Blaquiere, S. Trapp, S. T. Staben, J.
Org. Chem. 2011, 76, 1177–1179.
[63] S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys. 2010,
132, 154104–154119.
[27] J. Peng, M. Ye, C. Zong, F. Hu, L. Feng, X. Wang, Y. Wang, C.
Chen, J. Org. Chem. 2011, 76, 716–719.
[28] P. J. Dunn, in Comprehensive Organic Functional Group Transfor-
mations (Eds.: A. R. Katritzky, R. J. K. Taylor), Elsevier, Oxford,
2005, pp. 655–699.
[29] K. A. DeKorver, W. L. Johnson, Y. Zhang, R. P. Hsung, H. Dai, J.
Deng, A. G. Lohse, Y.-S. Zhang, J. Org. Chem. 2011, 76, 5092–5103.
Received: May 11, 2013
Revised: July 7, 2013
Published online: August 28, 2013
13810
www.chemeurj.org ꢂ 2013 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Chem. Eur. J. 2013, 19, 13803 – 13810