The NMR study of the mechanism of alkene arylation with anhydrides of aromatic acids

Article abstract of DOI:10.1023/A:1015364310409

The main steps of the catalytic cycle of the alkene arylation reaction with the participation of anhydrides of aromatic acids as arylation agents were studied by ³¹P NMR spectroscopy. The catalytic cycle of the reaction included the steps of ox

Full text of DOI:10.1023/A:1015364310409

Kinetics and Catalysis, Vol. 43, No. 2, 2002, pp. 195–198. Translated from Kinetika i Kataliz, Vol. 43, No. 2, 2002, pp. 215–218.
Original Russian Text Copyright © 2002 by Shmidt, Smirnov.
The NMR Study of the Mechanism of Alkene Arylation
with Anhydrides of Aromatic Acids
A. F. Shmidt and V. V. Smirnov
Irkutsk State University, Irkutsk, 664063 Russia
Received February 13, 2001
Abstract—The main steps of the catalytic cycle of the alkene arylation reaction with the participation of anhy-
drides of aromatic acids as arylation agents were studied by 31P NMR spectroscopy. In contrast to the mecha-
nism proposed earlier, palladium complexes containing benzoate anions as acidoligands were not found in the
reaction mixture. It was found that the catalytic cycle of the reaction includes the steps of oxidative addition of
Pd(0) formed in situ to the anhydride of acid, the substitution of acidoligand, and the elimination of the CO
molecule. Further transformations probably take place according to the usual steps of the Heck reaction. It was
shown that CO elimination is a limiting step.
INTRODUCTION
Using regioselectivity and kinetic data for the reac-
tion, a hypothetical mechanism of the reaction was pro-
posed (Scheme 1) [3]. This mechanism includes the
preliminary reduction of the initial Pd(II) complex, the
step of oxidative addition of Pd(0) to the molecule of
anhydride (1), the substitution of the chlorine anion for
benzoate ligand (2), CO elimination (3), and a further
reaction taking place in accordance with the usual steps
of the Heck reaction [2]. In this case, it was suggested that
catalyst activity is largely determined by the ability of pal-
ladium benzoyl complexes to eliminate CO (scheme 1,
step 3).
A new method of alkene arylation with the anhy-
drides of aromatic acids (I) was proposed in [1]. In con-
trast to the traditional Heck reaction (II) [2], this
method does not require the use of a base. Reaction (I)
has some advantages: the reagents are readily available,
the base is not present in stoichiometric amounts in the
reaction mixture, and it is possible to reuse acid result-
ing from the reaction in anhydride synthesis.
(PhCO)2O + H2C=CHPh
(I)
The goal of this work was to study reaction (I) by 31P
NMR spectroscopy and to check the mechanism of the
catalytic cycle proposed earlier [3] using kinetic data
(Scheme 1).
Pd(II)
PhCH=CHPh + PhCOOH + CO↑,
Pd(II)
PhI + H2C=CHPh
PhCH=CHPh + HI . (II)
base
Pd(0)L2 + 2Cl–
reduction
PdL2Cl2
+Cl–
–CO
(PhCO)2O + Pd(0)L2
PhCOPdL2OCOPh
PhCOPdL2Cl
PhPdL2Cl
–PhCOO–
+CO
1
A
Ph
2
B
3
C
Ph
PhPdL2Cl +
+ Pd(0)L2 + HCl
Ph
Scheme 1. Hypothetical mechanism of reaction (I).
EXPERIMENTAL
Bruker WP-200SY (200 MHz) and Varian VXR-500S
(500 MHz) spectrometers. Spectrometer stabilization
was performed using the signal of deuterium in CDCl3
(external standard); the chemical shifts are given with
respect to the signal of 85% H3PO4. The monitoring of
Reaction (I) and the synthesis of palladium phos-
phine complexes were carried out under anaerobic con-
ditions (in an argon atmosphere). All solvents and reac-
tants were thoroughly degassed and stored in an argon
atmosphere. 31P NMR spectra were recorded using reaction (I) was carried out by gas–liquid chromatogra-
0023-1584/02/4302-0195$27.00 © 2002 MAIK “Nauka /Interperiodica”

196
SHMIDT, SMIRNOV
phy (HP-4890 chromatograph, 15% polyphenylsilicon,
PhPd(PPh3)2Cl was obtained under analogous condi-
15 m, 100–200°C, flame-ionization detector; the inter- tions by the addition of 10 mmol of PhI and 0.32 mmol
nal standard was naphthalene). The purification of the of a reducing agent to the Pd(PPh3)2Cl2 solution. The
reagent was performed using standard procedures.
process was carried out until the complete conversion
of the initial complex and the appearance of a signal at
23.36 ppm. The chloride ligand in the complex was
exchanged with the iodide ligand at room temperature
in a 20-fold excess of (NBu4)I. The reaction of carbon-
ylation of the complex was carried out in a CO atmo-
sphere at room temperature (~30 min).
The Reaction of Styrene Phenylation
by Benzoic Anhydride
Styrene (10 mmol) and benzoic anhydride
(10 mmol) were dissolved in 10 ml of N-methyl pyr-
rolidone (NMP). This solution was introduced into a
retort with PdCl2 and Pd(PPh3)2Cl2 (0.16 mmol). The
retort was kept at a constant temperature of 140°C and
equipped with a magnetic stirrer. In some experiments,
The Preparation of PhCOPd(PPh3)2X
and PhPd(PPh3)2X from Pd(0) Complexes
PhCOPd(PPh3)2OCOPh
(22.23
ppm)
and
the reaction was carried out in the presence of LiCl PhPd(PPh3)2I (24.26 ppm) were prepared by analogy
(1.6 mmol). Reaction (I) occurred until product accu- with the known procedures [4, 5] using the reaction of
mulation discontinued.
oxidative addition of Pd(dba)2 (0.16 mmol) (dba is
trans-trans-dibezylidenacetone) to (PhCO)2O
(10 mmol) and PhI (10 mmol), respectively, in the pres-
ence of PPh3 (0.32 mmol). The ratio of reactants and the
conditions of decarbonylation and ligand exchange
were identical to the parameters described above.
The Preparation of PhCOPd(PPh3)2X
and PhPd(PPh3)2X from Pd(II) Complexes
For the preparation of PhCOPd(PPh3)2Cl, we added
10 mmol of (PhCO)2O and 0.32 mmol of a reducing
agent (NBu3 or HCOONa) to 10 ml of an NMP solution
containing 0.16 mmol of Pd(PPh3)2Cl2. The reaction
RESULTS AND DISCUSSION
To verify the mechanism presented in Scheme 1, we
mixture was stirred at 100°C until the signal of initial studied the composition and structure of intermediate
Pd(PPh3)2Cl2 completely disappeared (~40 min) and a palladium complexes formed in reaction (I) in the pres-
signal at 20.19 ppm probably corresponding to ence of two equivalents of PPh3 by 31P NMR spectros-
PhCOPd(PPh3)2Cl appeared. The same signal was copy. Preliminarily, we obtained complexes that could
observed under analogous conditions when be formed in the reaction mixture according to the
Pd(PPh3)2Cl2 reacted with PhI and a reducing agent in a mechanism (Scheme 1, complexes A, B, and C). For all
CO atmosphere. The exchange of chloride with the complexes, the values of chemical shifts (Scheme 2) in
benzoate ligand was carried out at room temperature by NMP solvent were measured (parameters of the 31P
the interaction of the complex with the equivalent NMR spectra of these complexes in NMR were not pub-
amount of AgOCOPh. The AgCl precipitate was lished earlier). We also studied the reactivity of these com-
decanted. The reaction of decarbonylation of the com- plexes. These complexes were synthesized by two meth-
plex was carried out at 100°C in a flow of argon ods: using Pd(II) compounds and using a conventional
(~120 min).
method [4–6] from Pd(0) complexes (Scheme 2).
reducing agent + CO + PhI
1
+LiCl
(PhCO)2O + 2PPh3
+(PhCO)2O
PhCOPd(PPh3)2Cl
PhCOPd(PPh3)2OCOPh
4
2
6
20.19 ppm
22.23 ppm
+AgOCOPh
–CO 7
–CO
+LiCl
8
PhPd(PPh3)2OCOPh
23.56 ppm
Pd(PPh3)2Cl2
24.64 ppm
5
Pd(dba)2
+CO
+LiCl
9
PhPd(PPh3)2Cl
23.36 ppm
PhPd(PPh3)2I
24.26 ppm
reducing agent + PhI
PhI + 2PPh3
3
+(NBu4)I
reducing agent = NBu3, HCOONa
Scheme 2. The formation, chemical shifts, and reactions of palladium phosphine complexes.
KINETICS AND CATALYSIS Vol. 43 No. 2 2002

THE NMR STUDY OF THE MECHANISM OF ALKENE ARYLATION
197
(PhCO)2O, PhI, or PhI in the presence of CO (1 atm)
(‡)
began to react with Pd(PPh3)2Cl2 only if a reducing
agent was present in the system (Scheme 2, steps 1, 2, 3).
We used a twofold excess of HCOONa or NBu3 as a
reducing agent (the ability of NBu3 to efficiently reduce
Pd(II) complexes was demonstrated in [7]). The reac-
tion formed σ-benzoyl (PhCOPd(PPh3)2Cl) and
σ-phenyl (Pd(PPh3)2Cl2) complexes with a yield
close to 100%. The 31P NMR spectra of these com-
plexes contain only one singlet. This indicates that
phosphine ligands have trans-configuration. The
presence of a chlorine atom and benzoyl fragment in
PhCOPd(PPh3)2Cl is confirmed by the reversible ligand
exchange in the presence of AgOCOPh and LiCl
(Scheme 2, step 4), by the reaction of CO elimination
(heating at 100°C in an argon flow), and by reverse car-
bonylation (Scheme 2, step 5). These reactions are typ-
ical of the complexes of the ArCOPd(PR3)2X type
[5, 6]. It is necessary to note that the reaction of decar-
bonylation required a higher temperature and is much
slower than carbonylation.
PhPd(PPh3)2Cl
Pd(PPh3)2Cl2
PhCOPd(PPh3)2Cl
PhCOPd(PPh3)2Cl
(b)
The oxidative addition of anhydrides of acids to
Pd(0) complexes was observed for the first time in [4].
The formation of PhCOPd(PPh3)2Cl from (PhCO)2O
and Pd(PPh3)2Cl2 in our case is evidence for the feasi-
bility of this step under the conditions of catalytic reac-
tion (I) when the Pd(0) complexes are generated in situ.
Alkene most probably is a reducing agent in reaction (I)
as in the usual Heck reaction (II) [2, 7]. The formation
of chloride complexes instead of benzoate complexes
observed in [4] can be due to the higher strength of a
Pd–Cl bond or with the occurrence of the reaction via
intermediate anionic complexes of Pd(0) as in [6].
Pd(PPh3)2Cl2
28
26
24
22
20 δ, ppm
31
Fig. 1. P NMR spectrum of the reaction mixture of
reaction (I) at the instant of maximal catalytic activity
(~20% conversion): (a) the reaction was carried out in the
σ-Benzoyl and σ-phenyl complexes with the ben-
zoate anion can be obtained from Pd(0) compounds
using the reaction of their oxidative addition to anhy-
dride, as was observed in [4], and the subsequent elim-
ination of CO (Scheme 2, steps 6 and 7). As was sup-
posed above, the chlorine anion in these complexes can
easily substitute for the benzoate anion in the interac-
tion with LiCl (Scheme 2, steps 4 and 8). The same pro-
cess is also typical of the complex with the iodine anion
(Scheme 2, step 9). In this process, complexes analo-
gous to those obtained above from Pd(PPh3)2Cl2 were
formed. The evidence for this is the coincidence of
respective chemical shifts.
presence of Pd(PPh ) Cl ; (b) PdCl (PPh was added
3 2
2
2
3
immediately before recording the spectrum).
It was shown in [3] that the application of LiCl addi-
tives improves the catalytic activity. According to NMR
spectroscopic data, the addition of LiCl is accompanied
by the almost complete disappearance of the benzoyl
complex in the reaction mixture (table, experiment 3).
The high reactivity of benzoyl complexes in the pres-
ence of LiCl can be the reason for this phenomenon.
Perhaps, this is due to a change in the palladium state in
PhCOPd(PPh3)2Cl because of the interaction of the Li+
cation with chlorine (analogously to [6]) or oxygen
atoms. Note that the nature of a cation of the added salt
strongly affects catalyst activity in reaction (I) [3].
Analysis of samples taken during the catalytic pro-
cess (I) showed that the reaction mixture does not con-
tain complexes with benzoate ligands (Fig. 1a). In this
process, up to 50% of palladium is found in a catalyti-
cally inactive form Pd(PPh3)2Cl2 (table, experiment 1).
This shows that the processes of Pd(0) formation are
not efficient enough. According to 31P NMR data, the
reaction is completed due to the transition of all palla-
dium into this catalytically inactive form (table, exper-
iment 2). Palladium oxidation in reaction (I) can be
caused by side processes of PhH and Ph–Ph formation.
The presence of phosphine in the reaction mixture
can understandably change the nature of the reaction
and does not permit us to draw an unambiguous conclu-
sion on the phosphine-free variant of the catalytic sys-
tem, which was used in [1, 3]. Therefore, we carried out
experiments where phosphine was added immediately
This is also typical of the conventional Heck reaction [7]. before recording the spectrum and after fast cooling of
KINETICS AND CATALYSIS Vol. 43 No. 2 2002

198
SHMIDT, SMIRNOV
Relative signal intensities of palladium complexes in 31P NMR spectra during reaction (I)
J, arb. units
Experiment
Catalytic system
PdCl2 + 2PPh3
Pd(PPh3)2Cl2
PhPd(PPh3)2Cl
PhCOPd(PPh3)2Cl
1
2
1.0
1.0
1.2
0
0.3
0
*
PdCl2 + 2PPh3
3
4
PdCl2 + 2PPh3 + 10LiCl
PdCl2 + 10LiCl**
1.0
1.0
0.6
0
0
2.7
* After reaction (I).
** PPh was added after reaching 20% conversion in reaction (I).
3
the reaction mixture. In these experiments, it was sup-
posed that palladium phosphine complexes are formed
rapidly by the substitution of more labile ligands (sol-
vent molecules or halogen anions). It was found that, in
the moment of maximal activity of the catalyst during
the reaction (I) in the reaction mixture, in addition to
Pd(PPh3)2Cl2, there is only PdCO(PPh3)2Cl (table,
experiment 4, Fig. 1b). This agrees with the hypothesis
that CO elimination is a limiting step, as suggested in
[3] on the basis of the study of the kinetics and regiose-
lectivity of reaction (I).
ACKNOWLEDGMENTS
This work was supported by the Russian Foundation
for Basic Research (project no. 99-03-32090). The
authors thank the German Service for Academic
exchanges (DAAD) and personally Professor W. Keim
(Technische Hochschule, Aachen, Germany) for the
possibility to carry out some experiments.
REFERENCES
1. Stephan, M.S., Teunissen, A.J., Verzijl, G.K.M., and De
Vries, J.G., Angew. Chem., Int. Ed. Engl., 1998, vol. 37,
no. 5, p. 662.
Thus, the results of the study of the kinetics and
regioselectivity [3], as well as the NMR monitoring of
reaction (I) show that the catalytic cycle of the reaction
includes the steps of the oxidative addition of Pd(0)
formed in situ to anhydride, the substitution of the halo-
gen anion for the benzoate acidoligand, and the elimi-
nation of the CO molecule. The influence of halide
additives on the catalytic activity is due to the substitu-
tion of the halogen anion for the benzoate acidoligand
in the intermediate complexes of the catalytic cycle.
One of the reasons for the activating influence of LiCl
is the lower activity of benzoyl palladium complexes.
2. Beletskaya, I.P. and Cheprakov, A.V., Chem. Rev., 2000,
vol. 100, no. 8, p. 3009.
3. Shmidt, A.F. and Smirnov, V.V., Kinet. Katal., 2000,
vol. 41, no. 6, p. 820.
4. Nagayama, K., Kawataka, F., Sakamoto, M., Shimizu, I.,
and Yamamoto, A., Chem. Lett., 1995, no. 5, p. 367.
5. Garrou, P.E. and Heck, R., J. Am. Chem. Soc., 1976,
vol. 98, no. 14, p. 4115.
6. Amatore, C. and Jutand, A., J. Organomet. Chem., 1999,
vol. 576, nos. 1–2, p. 254.
7. Shmidt, A.F. and Khalaika, A., Kinet. Katal., 1998,
vol. 39, no. 6, p. 875.
KINETICS AND CATALYSIS Vol. 43 No. 2 2002