Mannich-like three-component synthesis of α-branched amines involving organozinc compounds: ReactIR monitoring and mechanistic aspects

Article abstract of DOI:10.1016/j.jorganchem.2013.02.032

The infra-red spectroscopic in situ examination of the Mannich-like three-component reaction between organozinc reagents, amines and aldehydes is described, revealing that two distinct pathways are favored, depending on the nature of the starting amine. While reactions involving primary amines proceed via in situ-formed imines, reactions involving secondary amines rather imply a hemiaminal as the key intermediate species.

Full text of DOI:10.1016/j.jorganchem.2013.02.032

Journal of Organometallic Chemistry 736 (2013) 27e35
Contents lists available at SciVerse ScienceDirect
Journal of Organometallic Chemistry
journal homepage: www.elsevier.com/locate/jorganchem
Mannich-like three-component synthesis of
a
-branched amines involving
organozinc compounds: ReactIR monitoring and mechanistic aspects
*
Erwan Le Gall , Stéphane Sengmany, Caroline Hauréna, Eric Léonel, Thierry Martens
Électrochimie et Synthèse Organique, ESO, Institut de Chimie et des Matériaux Paris-Est, ICMPE, UMR 7182 CNRSeUniversité Paris-Est Créteil Val-de-Marne, 2-8 rue Henri Dunant,
9
4320 Thiais, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 19 December 2011
Received in revised form
The infra-red spectroscopic in situ examination of the Mannich-like three-component reaction between
organozinc reagents, amines and aldehydes is described, revealing that two distinct pathways are
favored, depending on the nature of the starting amine. While reactions involving primary amines
proceed via in situ-formed imines, reactions involving secondary amines rather imply a hemiaminal as
the key intermediate species.
17 January 2013
Accepted 26 February 2013
Ó 2013 Elsevier B.V. All rights reserved.
Keywords:
Multicomponent reaction
In situ spectroscopy
Infra-red
Reaction mechanism
Organozinc reagents
ReactIR
1. Introduction
diarylethylamines, benzylamines, b-arylethylamines and a-amino
esters [13,14]. While some excellent yields were obtained with this
process, the reaction mechanism remains unclear, mainly due to
the difficulty encountered in identifying the intermediate species.
One of the main goals of our current research interests being
directed towards the examination of mechanistic features for this
three-component reaction, we therefore describe herein our first
results obtained via ReactIR investigations. We assume that po-
tential mechanistic pieces of information that would be gathered
could open the way to the optimization of the reaction conditions
and the possible development of stereoselective versions of the
reaction.
In the last decades, in situ spectroscopies have emerged as useful
tools in chemistry by enabling reaction monitoring and providing
mechanistic elucidations required for a deeper understanding of
chemical processes, their optimization and development. Among a
range of spectroscopic techniques available, e.g. NMR [1e4] or
stopped-flow UV [5,6], the in situ mid-infra-red spectroscopy has
emerged as a major technique in the area [7e12], providing access
to accurate mechanistic interpretations under simple and useful
(
i.e., realistic) conditions. However, while physical chemistry pro-
cesses provide many examples of investigated processes, the ex-
amination of modern organic reactions by FTIR techniques still
remains underexplored. In this respect, multicomponent reactions
would constitute especially challenging reaction systems for
further in situ spectroscopic examinations as they represent com-
plex chemical processes in organic chemistry, generally involving
the formation of multiple transient species.
We recently described a three-component Mannich-like reac-
tion between organozinc reagents, amines and aldehydes as a
useful tool for the diversity-oriented synthesis of a range of
nitrogen-containing structures like, e.g., diarylmethylamines, 1,2-
2. Results and discussion
The investigated three-component reaction system is displayed
in Scheme 1.
In this reaction, organozinc reagents are assumed to be formed
in situ and to react with transient electrophilic species formed upon
addition of the amine to the carbonyl compound. Therefore, the
reaction can be considered as an “organometallic” variant of the
Mannich reaction. However, while classical Mannich reactions are
known to feature the formation of an iminium or an imine inter-
mediate, depending on the nature of the starting amine, the system
described herein could be more complex. Indeed, some published
*
Corresponding author. Tel.: þ33 1 49781135; fax: þ33 1 49781148.
E-mail addresses: legall@glvt-cnrs.fr, legall@icmpe.cnrs.fr (E. Le Gall).
work indicates that Bruylants-type reactions of a-aminonitriles are
0
022-328X/$ e see front matter Ó 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jorganchem.2013.02.032

28
E. Le Gall et al. / Journal of Organometallic Chemistry 736 (2013) 27e35
possible when using organozinc reagents as nucleophiles [15e17].
Iminium ions could therefore constitute conceivable electrophiles
in the process. On the other hand, a possible alternate mechanism
has also been proposed by Fan and co-workers in a recent study on
the multicomponent reaction of allyl- and benzylzinc halides with
aldehydes and primary amines [18]. The authors of this study
suggest that the imine formed upon reaction between the amine
and the aldehyde is not the reactive electrophile in the process, but
instead that the hemiaminal formed during the course of the re-
action is deprotonated by the organozinc species and that the
resulting zinc-alcoholate is subsequently engaged in a putative
six-membered transition-state with a second equiv of the orga-
nozinc reagent to finally give rise to the formation of the three-
component coupling product. Both reaction pathways are depic-
ted in Scheme 2.
Scheme 1. Multicomponent reaction investigated in this study.
The preparation of the suggested reactive intermediates A and B
and the examination of their reactions with organozinc reagents
might constitute a possible, direct, experimental approach for the
determination of the mechanism for this reaction. Unfortunately,
while the formation of imines or iminium intermediates B (or
synthetic equivalents thereof) might be considered, the preparation
of acyclic hemiaminals such as A has not been described to date,
probably due to an intrinsic instability. Consequently, the use of in
situ FTIR spectroscopy is envisioned here as a practical experi-
mental alternative to the preparation of the very unstable in-
termediates most probably involved in the reaction, assuming that
these species could be detected and identified via specific vibra-
tions, even at low concentrations. The requisite condition for a
reliable study of the mechanism then calls for the unambiguous
identification of specific absorption bands assignable to each spe-
cies involved in the reaction (reagents, intermediates, products). In
particular, these bands should not be mixed up with other ab-
sorption bands as they will serve as references for monitoring the
evolution of the corresponding species in a qualitative (and semi-
quantitative) manner during the reaction.
Scheme 2. Proposed reaction pathways for the investigated multicomponent reaction.
Two sets of substrates were selected as models in this study.
First, the reaction of primary amines was investigated with aniline,
benzaldehyde, and benzylzinc bromide as the organozinc species.
For the monitoring of the reaction with secondary amines, piperi-
dine was chosen instead, due the lack of reactivity of N-substituted
aromatic amines under standard conditions. Furthermore, the
Scheme 3. Model reaction for the examination of the mechanism when using primary
amines.
Fig. 1. Overlay of spectra obtained for two of the reagents (benzaldehyde (red) and aniline (green)), final product 1 (greyeblue), and putative imine intermediate PhCH]NPh
orange). The spectrum obtained at the end of the reaction (violet) when monitoring the process is added for comparison. (For interpretation of the references to colour in this figure
legend, the reader is referred to the web version of this article.)
(

E. Le Gall et al. / Journal of Organometallic Chemistry 736 (2013) 27e35
29
Scheme 4. Reaction scheme corresponding to the experimental protocol when using
aniline.
benzylzinc reagent was replaced with an arylzinc reagent to
both overcome increased spectral interferences [19] and further
expand the scope of this work.
2.1. Monitoring of the reaction with primary amines
The mechanistic investigation was first realized with a primary
amine, aniline, a preformed benzylzinc reagent and an aldehyde,
benzaldehyde (Scheme 3). The reagents were chosen as “bare” as
possible (i.e., non functionalized) to obtain simpler infra-red
spectra and subsequently limit the number of absorption bands
and the odds for spectral interferences.
Fig. 3. Three-dimensional plot of the IR monitoring (3998e648 cm 1 area) corre-
ꢀ
sponding to the experimental reaction conditions summarized in Fig. 2.
Reagents and products of the reaction were individually charac-
terized by IR, as well as a putative imine intermediate, N-benzyli-
deneaniline (PhCH]NPh), that results from the in situ condensation
of the aldehyde with the amine. Specific absorption bands were
considered for each species, and selected as follows for the subse-
modifications for the absorption of bands corresponding to benz-
aldehyde and aniline, respectively. Although both hemiaminal A
and imine B could be plausible intermediates in the reaction, close
inspection of the accumulated spectra only revealed the formation
of imine B , no presence of OH bands attributable to the formation
of hemiaminal A being detected, as evidenced on the three-
1
1
1
ꢀ1
quent monitoring of the reaction: benzaldehyde: 1703 cm , aniline:
1
ꢀ
1
ꢀ1
ꢀ1
1
(
280 and 1630 cm , imine: 1580 cm , final product 1: 870 cm
Fig. 1).
The following experimental protocol was then applied: aniline
dimensional plot of the IR spectra monitoring (Fig. 3).
At t ¼ 37 min, the organozinc reagent was introduced dropwise
over a 10 min time period. As soon as the first drops of the solution
were introduced in the reaction medium, the absorbance due to the
imine and the coupling product 1 increased (mid-section in Fig. 2)
whereas the intensity of the C]O band rapidly decreased, associ-
ated to the consumption of benzaldehyde. It is worth noting that
when aniline and benzaldehyde had first been introduced in the
solvent (see previous paragraph), the reaction equilibrium between
these reagents and the imine had only moderately shifted towards
the imine formation, while the present introduction of the orga-
nozinc solution had the effect to significantly increase the imine
formation, either by Lewis acid activation of the carbonyl function
through the addition of zinc salts and/or by acidebase catalysis.
and benzaldehyde (1 equiv each) were allowed to react for several
minutes in acetonitrile, until preformed benzylzinc bromide [20]
(
2 equiv) was added dropwise to the solution (Scheme 4).
The monitoring, over time, of the absorbance associated to the
specific bands mentioned above is depicted in Fig. 2.
At t ¼ 2 min, aniline was introduced in the reaction medium
(
acetonitrile), followed by benzaldehyde at t ¼ 7 min. Until
t ¼ 37 min, these species were allowed to react without further
addition of reagents. During this time period, the formation of in-
termediates from aniline and benzaldehyde appears to be
extremely limited, as evidenced by the absence of noticeable
Fig. 2. Monitoring of the model three-component reaction when using aniline as a model primary amine; subsequent addition to acetonitrile used as a solvent of: (1) aniline (light-
ꢀ
1
ꢀ1
ꢀ1
green (1280 cm band) and green (1630 cm band) curves), (2) benzaldehyde (red curve (1703 cm band)), and (3) benzylzinc bromide; imine appearance is monitored at
1
ꢀ
1
ꢀ1
580 cm (orange curve); product 1 appearance is monitored at 880 cm (greyeblue curve). (For interpretation of the references to colour in this figure legend, the reader is
referred to the web version of this article.)

30
E. Le Gall et al. / Journal of Organometallic Chemistry 736 (2013) 27e35
ꢀ
1
ꢀ
1
Fig. 4. Effect of zinc salts on the formation of imine B (light-orange (1190 cm band) and orange (1580 cm band) curves); subsequent addition to acetonitrile used as a solvent
1
ꢀ
1
ꢀ1
ꢀ1
of: (1) aniline (light-green (1280 cm band) and green (1630 cm band) curves), (2) benzaldehyde (red curve (1703 cm band)), and (3) zinc bromide. (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of this article.)
1
Scheme 5. Formation of imine B .
In order to determine whether zinc salts contained in the so-
lution were responsible for the rate enhancement observed in the
imine formation, 0.5 equiv of zinc bromide were added to an
acetonitrile solution containing benzaldehyde and aniline. The
monitoring of this reaction is depicted in Fig. 4.
formation of the imine could be observed, along with a
consistent decrease of benzaldehyde and aniline concentrations,
demonstrating the rate-accelerating effect and equilibrium shift
induced by zinc salts in the imine formation, as depicted in
Scheme 5.
At t ¼ 3 min, aniline was introduced in the reaction medium
We then proceeded further with these reaction FTIR monitoring
in order to verify that a mixture containing both reagents and zinc
salts constitute a suitable reaction medium for the reaction of
subsequently-added benzylzinc reagent. Thus, to the above solu-
tion was added preformed benzylzinc bromide. The corresponding
monitoring curves are presented in Fig. 5.
(
acetonitrile), followed by benzaldehyde at t ¼ 5 min. A slight
1
increase in the imine B absorbance occurred, indicating that
the imine is formed even in the absence of Lewis acids, albeit
at low conversion. When 0.5 equiv of zinc bromide were added
to the mixture at
t
¼
8.5 min, a quite more significant
ꢀ
1
Fig. 5. Effect of the addition of benzylzinc bromide on an acetonitrile solution containing aniline (1 equiv, green curve (1630 cm band)), benzaldehyde (1 equiv, red curve
ꢀ
1
ꢀ1
(
1703 cm band)) and zinc bromide (0.5 equiv); product 1 appearance is monitored at 880 cm (greyeblue curve). (For interpretation of the references to colour in this figure
legend, the reader is referred to the web version of this article.)

E. Le Gall et al. / Journal of Organometallic Chemistry 736 (2013) 27e35
31
benzylzinc reagent. In addition, we could show that imine B
1
for-
mation is favored by the addition of an organozinc solution con-
taining zinc salts, which most probably act as Lewis acids, thus
providing a rate enhancement in the reaction leading to coupling
product 1. This finding is corroborated by some of our previous work
that had indicated that, under similar conditions, aromatic orga-
nozinc reagents do not react with in situ-generated imines, since
they probably do not exhibit sufficient reactivity [21]. Thus, path I in
Scheme 6 may be the most probable reaction pathway when pri-
mary amines are employed in this three-component reaction.
2.2. Monitoring of the reaction with a model secondary amine
An analogous mechanistic study was then carried out, starting
from secondary amines. For consistency, we had initially intended
to use a N-substituted secondary aniline, yet a different amine
structure had finally to be used, due to both a lack of reactivity
exhibited by N-substituted secondary anilines, and the observation
of more spectral interferences than in the previous study. In order
to circumvent these issues and obtain reliable results, the mecha-
nistic investigation was thus realized with piperidine, benzalde-
hyde and phenylzinc bromide (Scheme 7).
Scheme 6. Putative reaction pathway when primary amines are used.
Reagents and product of the reaction were individually charac-
terized by IR, as well as a commercial iminium salt, Eschenmoser’s
salt (Fig. 10). Specific absorption bands were identified for each
species, and selected as follows for the subsequent monitoring of
ꢀ
1
ꢀ1
the reaction: benzaldehyde: 1703 cm and piperidine: 780 cm
Fig. 6). Unfortunately, it was not possible to identify specific ab-
Scheme 7. Model reaction for the examination of the mechanism when using sec-
(
ondary amines.
sorption bands for product 2 whose formation was thus monitored
by GC analysis.
At t ¼ 35 min, the preformed organozinc reagent was added
dropwise over a 5 min time period. The intensity of the band
assigned to coupling product 1 raises quickly as soon as the addi-
tion starts and then reaches a maximum, concomitantly to a par-
allel decrease of aniline, benzaldehyde and imine absorbances, thus
confirming that product 1 is formed by consumption of imine B via
1
equilibrium shifting.
In summary, this first part of our study indicates that an imine is
likely the reactive electrophilic intermediate in the multicomponent
reaction between a primary amine, an aromatic aldehyde and a
Owing to the influence of the zinc salt as noticed in the first part
of this study, we first attempted to determine whether the corre-
sponding salts would also promote the formation of reaction in-
termediates when benzaldehyde is mixed with piperidine. The IR
monitoring of the corresponding experiment is presented in Fig. 7.
At t ¼ 4 min, piperidine was introduced in the reaction medium
(acetonitrile). The ensuing addition of benzaldehyde at t ¼ 7.5 min
resulted in the appearance of three weak bands at 1274, 1084 and
ꢀ
1
ꢀ1
1006 cm (Fig. 8) as well as of a large band at 3540 cm . When
zinc bromide (0.5 equiv) was added to the solution at t ¼ 12.5 min,
þ
Fig. 6. Overlay of spectra obtained for two of the reagents (benzaldehyde (red) and piperidine (green)), final product 1 (greyeblue), and model iminium salt CH
2
]N Me
2
(blue).
The spectrum obtained at the end of the reaction when monitoring the process is added (violet curve) for comparison. (For interpretation of the references to colour in this figure
legend, the reader is referred to the web version of this article.)

32
E. Le Gall et al. / Journal of Organometallic Chemistry 736 (2013) 27e35
ꢀ
ꢀ1
1
ꢀ1
Fig. 7. Effect of zinc salts on the reaction between piperidine (green curve (780 cm band)) and benzaldehyde (red curve (1703 cm band)) in acetonitrile; the origin for extra
ꢀ
1
ꢀ1
ꢀ1
curves (yellow (1274 cm band), light-orange (1084 cm band), orange (3540 cm band) and light brown (1006 cm band)) is discussed in the text. (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of this article.)
þ
the intensity of these bands further increased as the result of a
possible Lewis-acid type activation of the carbonyl double bond
and the resulting improved formation of a novel intermediate.
An examination of possible fits for the wavenumbers associated
to the above bands suggests that a hemiaminal is likely formed
upon mixing of piperidine and benzaldehyde. Indeed, hemiaminal
iminium intermediate B
ylation of A
bending vibrations have not been described for such species in the
2
PhCH]N(CH
2
)
5
via a further dehydrox-
2
should also be considered. As typical stretching and
ꢀ1
spectroscopic region of interest (i.e., 950e1300 cm ), we under-
went the examination of the infra-red spectrum of a commercial
iminium salt (Eschenmoser’s salt) in acetonitrile to determine
A
1
would be characterized by typical vibrations for tertiary amines
whether the observed vibrations could be attributed to iminium B
2
(Fig. 10).
An examination of the corresponding spectrum (Fig. 10) in-
dicates that no significant bands are present in the three areas of
interest. Thus, it can be concluded that experimental data strongly
support the idea that zinc salts mainly promote the formation of
ꢀ
1
ꢀ1
(
1000e1200 cm ) and secondary alcohols (1050e1170 cm and
ꢀ1
1250e1350 cm ), in addition to the typical large band at 3400e
ꢀ
1
3
600 cm that corresponds to the OH stretching of alcohols. It
can be noted that the appearance of this latter band is clear, as
attested by the three-dimensional plot corresponding to the full-
wavelength IR monitoring of the reaction (Fig. 9).
2 2
heminaminal A rather than of iminium salt B .Consequently, it is
Consequently, at first glance, it could be considered that zinc
salts promote the formation of a hemiaminal intermediate from
benzaldehyde and piperidine (Scheme 8).
rather difficult to consider the additional presence of an iminium
species in the reaction medium, even in slight amounts.
The addition of an organozinc reagent on the above interme-
diate was then evaluated according to the following experiment:
piperidine and benzaldehyde (1 equiv) were allowed to react for
2
Although the formation of hemiaminal A appears consistent
with the collected spectroscopic data, the in situ generation of an
ꢀ
1
Fig. 8. Focus on the three new bands observed at 1274, 1084 and 1006 cm (violet spectrum corresponding to the experiment’s end) when mixing piperidine (green) and
þ
benzaldehyde (red) in acetonitrile; the spectrum for model iminium salts H
legend, the reader is referred to the web version of this article.)
2
C] N(CH
3
)
2
(blue) is added for comparison. (For interpretation of the references to colour in this figure

E. Le Gall et al. / Journal of Organometallic Chemistry 736 (2013) 27e35
33
Scheme 9. Reaction scheme corresponding to the experimental protocol when using
piperidine as a model secondary amine.
at 1084, 1006 and 3540 cm 1, the band at 1274 cmꢀ1 suffering from
an overlap with other bands was not considered. The concentration
of the hemiaminal (light-orange, orange and light-brown curves)
jumped to a plateau until t ¼ 24 min, corresponding to the complete
addition of the first equiv of phenylzinc bromide. At t ¼ 24 min, the
second equiv of the organozinc was added quickly. As a result of this
addition, the concentration of the hemiaminal rapidly decreased
and the three-component coupling product 2 appeared over time, as
confirmed independently by gas chromatography. These results
support the preliminary formation of the intermediate hemiaminal
ꢀ
Fig. 9. Three-dimensional plot of the IR monitoring (3998e648 cmꢀ1 area) corre-
sponding to the experimental reaction conditions summarized in Fig. 7.
2
A . This intermediate could then be involved in several reaction
pathways. The first possibility would be to form the iminium ion by
dehydroxylation. However, as mentioned above, no iminium ion
could be noticed by FTIR. In addition, although it could be supposed
that this species would be instantly trapped by the organozinc,
chromatographic analyses of the reaction medium over time
revealed that, while the organozinc reagent is continuously
consumed upon addition to the reaction medium, no three-
component coupling product is obtained before one equiv of the
organozinc reagent had been added to the medium. Consequently,
these elements are more compatible with a second alternate reac-
tion pathway that would involve the consumption of the organozinc
in a process implying another intermediate. A plausible mechanism
for this consumption of the organozinc reagent could be envisaged
2
Scheme 8. Formation of hemiaminal A .
several minutes, and then preformed phenylzinc bromide [21]
2 equiv) was added dropwise to the solution (Scheme 9).
The monitoring, over time, of the absorbance for the bands
mentioned above, is depicted in Fig. 11.
(
At t ¼ 3 min, piperidine was added to acetonitrile followed by
benzaldehyde at t ¼ 7.5 min. Both reagents were allowed to react for
2
through a sequence in which A would be initially consumed by an
acidebase reaction with the organozinc to form the alcoholate A ’
2
1
0 min. As before, no formation of iminium species B
2
could be
noticed. Then, the first equiv of the organozinc was added dropwise
tothe reaction medium between t ¼ 17.5 min and t ¼ 24 min. As soon
as the first drops of the organozinc solution had been added, the
(Scheme 10).
Further observations support this hypothesis. Indeed, between
0 and 1 equivalent of added organozinc, benzaldehyde and piper-
idine are still present in the reaction mixture, implying that the
hemiaminal is regenerated and maintain the monitored
2
formation of hemiaminal A could be identified by the noticeable
appearance of the three previously identified representative bands
Fig. 10. IR spectrum of Eschenmoser’s salt.

34
E. Le Gall et al. / Journal of Organometallic Chemistry 736 (2013) 27e35
Fig. 11. Monitoring of the three-component reaction when using piperidine as a model secondary amine; subsequent addition to acetonitrile used as a solvent of: (1) piperidine
ꢀ
1
ꢀ
1
ꢀ
1
(
green curve (780 cm band)), (2) benzaldehyde (red curve (1703 cm band)), and (3) phenylzinc bromide; intermediate species appearance is monitored at 1006 cm (light-
ꢀ
1
ꢀ1
brown curve), 1084 cm (light-orange curve) and 3540 cm (orange curve). (For interpretation of the references to colour in this figure legend, the reader is referred to the web
version of this article.)
Scheme 10. Successive formation of a hemiaminal and an alcoholate in the proposed mechanism.
concentration constant. Later, when benzaldehyde and piperidine
are in insufficient concentration to counterbalance the consump-
a hemiaminal, although not in acetonitrile, in which case an imine
is produced as noticed in the present study. Consequently, the
solvent might play a critical role in the reaction fate and mecha-
nism. These findings are not surprising since we had already
noticed that the three-component reaction of benzyl bromine with
benzaldehyde and aniline is both quite more rapid (10e20 times
2
tion of the hemiaminal, the concentration of A quickly decreases as
ꢀ1
clearly evidenced in the 3500 cm area of Fig. 12.
2
At this stage, one equiv of the zinc alcoholate A ’ might have
been produced, hence additional organozinc introduced in the re-
action medium should serve as a nucleophile in the following
mechanistic steps. GC experiments confirm that the three-
component product starts to appear at this stage. Thus, it seems
plausible to envision a reaction of the organozinc with the alco-
faster) in CH
compounds is strongly inhibited in THF contrary to what is
observed in CH CN [14].
3
CN than in THF, and that the reaction of arylzinc
3
2
holate A via the 6-center transition state TS as postulated above
(Scheme 11), via a zinc-assisted nucleophilic displacement of the
oxygen by the organozinc and the final production of the coupling
product 2, along with zinc salts (path II in Scheme 11).
Consequently, this part of the study also provides useful infor-
mation about the multicomponent reaction between a secondary
amine, an aromatic aldehyde and an arylzinc reagent. Indeed, it can
be argued that, contrary to what was observed with primary
amines, the three-component reaction of secondary amines does
not involve the formation of a C]N double bond by dehydrox-
ylation and trapping by the organozinc reagent, but rather the
formation of an alcoholate and a phenyl transfer at the stage of a
6
-center transition state TS. This result is possibly consistent with a
limited formation of the iminium intermediate due to a rapid acide
base reaction between the hemiaminal A and the organozinc
2
agent.
Our results are finally to be compared with those described by
Fan et al. for similar multicomponent coupling reactions realized in
tetrahydrofuran. Indeed, it was supposed by these authors that,
when using primary amines, the key intermediate of the reaction is
ꢀ1
Fig. 12. Three-dimensional plot of the IR monitoring (3998e648 cm area) corre-
sponding to the reaction conditions indicated in Fig. 11.

E. Le Gall et al. / Journal of Organometallic Chemistry 736 (2013) 27e35
35
function, it is possible to automatically detect the formation or
disappearance of compounds or intermediates whose simulated
spectra are extracted from the overall signal). Reactions were
concomitantly monitored by gas chromatography (GC) using a
Varian 3400 chromatograph fitted with a capillary DB1 column
(l ¼ 5 m, Ø ¼ 0.32 mm, df ¼ 0.4
mm). Solvents and reagents were
purchased from commercial suppliers and used without further
purification. Before monitoring experiments, the IR spectra of sol-
vent, starting and final compounds were recorded under neat
conditions for comparison purpose. Monitoring experiments were
then conducted in a 100 mL tricol under argon atmosphere. In a
typical experiment, 30 mL acetonitrile were added to the flask,
followed by the amine and the aldehyde (10 mmol of each). The
organozinc compound was prepared independently on a 30 mmol
scale, in 30 mL acetonitrile (see Refs. [13,14,20,21] for details). Then
the filtered (syringe) solution was added dropwise to the medium
under constant IR monitoring.
Acknowledgements
The financial support of this work by the CNRS and Université
Paris-Est (PhD grant to CH) is gratefully acknowledged. The authors
also thank Pr. Michel Troupel for useful discussions.
Scheme 11. Putative reaction pathways when secondary amines are used.
3
. Conclusion
References
In conclusion, we have demonstrated by in situ spectroscopic
[1] B. Nelson, W. Hiller, A. Pollex, M. Hiersemann, Org. Lett. 13 (2011) 4438e4441.
[2] W.R.R. Harker, E.L. Carswell, D.R. Carbery, Org. Lett. 12 (2010) 3712e3715.
[3] M.B. Schmid, K. Zeitler, R.M. Gschwind, J. Org. Chem. 73 (2011) 3005e3015.
[4] A.W. Douglas, N.L. Abramson, I.N. Houpis, S. Karady, A. Molina, L.C. Xavier,
N. Yasuda, Tetrahedron Lett. 35 (1994) 6807e6810.
experiments using a ReactIR instrument that the Mannich-like
three-component reaction between amines, aromatic aldehydes
and organozinc reagents might proceed according to two distinct
pathways. While reactions involving primary amines proceed via in
situ-formed imines, the reactions involving secondary amines
rather feature a hemiaminal as the key intermediate. These results
will be used for the further optimization of the multicomponent
system and the design of stereoselective versions of the reaction.
[5] T. Paul, M.A. Hassan, H.-G. Korth, R. Sustmann, J. Org. Chem. 61 (1996)6835e6848.
[6] O.S. Tee, N.R. Iyengar, M. Paventi, J. Org. Chem. 48 (1983) 759e761.
[7] H.A. Stephani, J.M. Pena, F. Manarin, R.A. Ando, D.M. Leal, N. Petragnani,
Tetrahedron Lett. 52 (2011) 4398e4401.
[8] N.E. Leadbeater, Chem. Commun. 46 (2010) 6693e6695.
[9] G.B. Chernobay, Y.A. Chesalov, V.P. Baltakhinov, G.Y. Popova,
T.V. Andrushkevich, Catal. Today 164 (2011) 58e61.
[
[
[
10] M.C. Li, D.B. Bao, C.A. Ma, Electrochim. Acta 56 (2011) 4100e4104.
11] W. Zhao, E.M. Carreira, Org. Lett. 13 (2011) 5084e5087.
12] C.F. Carter, H. Lange, S.V. Ley, I.R. Baxendale, B. Wittkamp, J.G. Goode,
N.L. Gaunt, Org. Proc. Res. Dev. 14 (2010) 393e404.
4
. Experimental section
Reactions were monitored by Fourier transformed infra-red
[13] C. Haurena, E. Le Gall, S. Sengmany, T. Martens, M. Troupel, J. Org. Chem. 75
(
FTIR) in situ spectroscopy using a Mettler Toledo ReactIR iC10
(2010) 2645e2650.
[
14] E. Le Gall, C. Haurena, S. Sengmany, T. Martens, M. Troupel, J. Org. Chem. 74
fitted an articulated arm and a diamond probe. The ATR mode
(2009) 7970e7973.
(
attenuated total reflection) was used for the analyses, which were
[15] E. Le Gall, C. Gosmini, M. Troupel, Tetrahedron Lett. 47 (2006) 455e458.
[16] R. Moumne, S. Lavielle, P. Karoyan, J. Org. Chem. 71 (2006) 3332e3334.
[17] L. Bernardi, B.F. Bonini, E. Capito, G. Dessole, M. Fochi, M. Comes-Franchini,
A. Ricci, Synlett (2003) 1778e1782.
ꢀ
1
conducted in the 650e4000 cm
optical range. Spectral in-
ꢀ1
terferences were observed between 1950 and 2250 cm , due to the
intrinsic absorption of diamond in this area. The iC IR software was
used for the monitoring of the whole system and infra-red spectra
handling. In all cases, selection of the bands was performed
manually and the optional extraction and monitoring of “compo-
nents” by the iC IR software was not used (using this software
[
18] R. Fan, L. Pu, L. Qin, F. Wen, G. Yao, J. Wu, J. Org. Chem. 72 (2007) 3149e3151.
[19] The reactions have been undertaken with benzylzinc compounds as well and
provided similar results, although more spectral interferences were observed.
[
20] S. Sengmany, E. Le Gall, C. Le Jean, M. Troupel, J.-Y. Nédélec, Tetrahedron 63
2007) 3672e3681.
[21] E. Le Gall, M. Troupel, J.-Y. Nédélec, Tetrahedron 62 (2006) 9953e9965.
(