Identification of a higher-order organozincate intermediate involved in Negishi cross-coupling reactions by mass spectrometry and NMR spectroscopy

Article abstract of DOI:10.1002/chem.201101029

Negishi cross-coupling reactions were analyzed in solution by mass spectrometry and NMR spectroscopy to identify both the effect of LiBr as an additive as well as the purpose of 3-dimethyl-2-imidazolidinone (DMI) as a co-solvent. The results suggest that the main role of DMI is to facilitate a higher order bromozincate formation during the addition of LiBr. Copyright

Full text of DOI:10.1002/chem.201101029

FULL PAPER
DOI: 10.1002/chem.201101029
Identification of a Higher-Order Organozincate Intermediate Involved in
Negishi Cross-Coupling Reactions by Mass Spectrometry and NMR
Spectroscopy
Howard N. Hunter, Niloufar Hadei, Voislav Blagojevic, Pascal Patschinski,
George T. Achonduh, Stephanie Avola, Diethard K. Bohme, and Michael G. Organ*
[a]
Abstract: Negishi cross-coupling reactions were analyzed in solution by mass spec-
trometry and NMR spectroscopy to identify both the effect of LiBr as an additive
as well as the purpose of 3-dimethyl-2-imidazolidinone (DMI) as a co-solvent. The
results suggest that the main role of DMI is to facilitate a higher order bromo-
Keywords: additives
coupling · Negishi cross-coupling ·
organozincate · palladium · zinc
·
cross-
ACHTUNGTRENNUNGz incate formation during the addition of LiBr.
Introduction
role of additives and co-solvents remains speculative. Re-
cently, Koszinowski and co-workers investigated the forma-
tion of organozincate anions, prepared by the transmetala-
Palladium-catalyzed cross-coupling of organo halides with
organozinc reagents (Negishi reaction) is a selective and ver-
tion of an organolithium compound with ZnCl in THF, by
2
[1]
[10]
satile reaction for the construction of CꢀC bonds. Studies
have shown that the yield of such Pd-catalyzed cross-cou-
pling reactions is influenced by solvent, temperature, ligand
electrospray ionization mass spectrometry (ESI-MS).
More recently, our group has investigated the effect of dif-
3
3
ferent additives in sp –sp Negishi cross-coupling reactions
[2]
[3]
[11]
nature, and additives. Although additives have been
shown to play a major role in cross-coupling reactions, only
a few mechanistic and additive studies are known. Original-
ly, Negishi and co-workers discovered that 2-iodotoluene
could be coupled with lithium 1-heptynyltributylborate (an
(see Scheme 1). A wide variety of different salts has been
“ate” complex), establishing for the first time that organo-
boranes were effective transmetalating agents in Pd-mediat-
[4]
ed cross-coupling reactions. Neutral alkenyl–boranes were
next applied by Miyaura and Suzuki, who found that the ad-
dition of an exogenous alkoxide base was necessary to effect
3
3
Scheme 1. The sp –sp Pd catalyzed cross-coupling using the PEPPSI cat-
alyst (PEPPSI=pyridine-enhanced precatalyst preparation stabilization
and initiation).
[5]
the cross-coupling of these organometallic species. Mecha-
[6]
nistic aspects of this reaction by Matos and Soderquest
demonstrated the presence of a hydroxy–borate complex as
an active transmetalation species. Since then it has been re-
vealed that a palladium–hydroxo complex with a boronic
acid is involved in transmetalation. Later Knochel and
co-workers found that LiCl accelerated formation of or-
used as an additive in these coupling reactions, with LiBr
proving to be the most effective. Moreover, in these re-
ACHTUNGTRENNUNG
[7]
[8]
[11]
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
ganozinc reagents that were prepared by using a Br/Mg ex-
leads to good yields of the cross-coupled product. Further-
more, in a careful titration study we found that the addition
[9]
change method, while Oshima and co-workers suggested
that LiCl enhanced the reactivity of organozinc halides
toward transmetalation. Despite these studies, the actual
of ZnBr inhibited the formation of the active species. Here,
2
we investigate the effect of adding LiBr and ZnBr on the
2
equilibria involving the organozinc species that are central
to the Negishi cross-coupling reaction using a multipronged
interdisciplinary approach, including synthetic results, mass
spectrometry, and NMR spectroscopic methods. Further, the
interplay of these additives with 3-dimethyl-2-imidazolidi-
none (DMI), a critical solvent for this transformation, is also
investigated.
[
a] Dr. H. N. Hunter, Dr. N. Hadei, Dr. V. Blagojevic, P. Patschinski,
G. T. Achonduh, S. Avola, Prof. Dr. D. K. Bohme, Dr. M. G. Organ
Department of Chemistry
York University
4
700 Keele St. Toronto, M3J 1P3 (Canada)
Fax : (+1)416-736-5936
E-mail: organ@yorku.ca
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201101029.
Chem. Eur. J. 2011, 17, 7845 – 7851
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7845

Results and Discussion
A study of the complex collection of charged and neutral
species in this reaction requires complimentary methods of
analysis that can give insight into the identity of these com-
pounds and intermediates. Clearly mass spectrometry and
NMR spectroscopy seem best suited to this task. The ap-
proach taken was to prepare solutions for analysis that
mimic reaction solutions. Therefore, all mass spectrometry
and NMR experiments were performed by using concentra-
tions employed in the cross-coupling reaction (0.2m). The
solutions consisted of nBuZnBr and LiBr in the solvents
THF, DMI or N-methyl-2-pyrrolidone (NMP) (or combina-
tions of these solvents). Although positive-ion mass spectra
of nBuZnBr and LiBr under all conditions showed only var-
+
ious protonated solvent and Li adducts as the main species,
ꢀ
ꢀ
negative-ion mass spectra were far more interesting, with
Figure 2. Changes in the ratio of nBuZnBr2 to ZnBr3 ion intensity as a
function of LiBr (black symbols) and ZnBr equivalents added (white
symbols) in DMI. The starting amount of LiBr for the ZnBr2 study is
.5 equivalents.
ꢀ
the predominantly observed species being nBuZnBr and
2
2
ꢀ
ZnBr . Whereas no cross-coupling occurs in straight THF,
3
1
it proceeds very efficiently in DMI and thus mass spectrom-
etry data were initially obtained in this solvent. Initial obser-
vations indicated that increased addition of LiBr caused the
lowed by back-titration with ZnBr . The similarity in the
2
ꢀ
ꢀ
ꢀ
ratio of nBuZnBr2 to ZnBr3 to change from approximate-
two curves demonstrates that LiBr acts as a source of Br
assisting the reaction by promoting the formation of
nBuZnBr , while ZnBr acts as a Br sponge to decrease
2 2
the amount of Br (and hence nBuZnBr ) present.
2
[12]
ly 1:8 (Figure 1a) to 5:1 (Figure 1e).
Since ZnBr was known to inhibit the reaction,
[11b]
ꢀ
ꢀ
mass
2
ꢀ
ꢀ
spectra were collected using solutions with 1.5 equivalents
ꢀ
of LiBr spiked with 0.6, 1.0, and 2.0 equivalents of ZnBr .
The presence of a significant amount of nBuZnBr2 at
0.5 equivalents of LiBr suggests that this ion is unlikely to
be the active transmetalating species in the Negishi reaction,
since the coupling requires a minimum of 1.0 equivalents of
LiBr before significant reaction occurs. It is possible that a
higher order species, such as nBuZnBr3 , is the active inter-
mediate even though we are unable to directly observe
these di-anions in the mass spectra. Such multiply charged
species, although stable in the presence of counterions and
solvation effects in solution, charge-separate during the elec-
trospray process due to Coulombic repulsion. For example,
2
ꢀ
Figure 2 shows the variation in the ratio of nBuZnBr to
2
ꢀ
ZnBr3 as a function of the stepwise addition of LiBr fol-
2ꢀ
2ꢀ
ꢀ
ꢀ
nBuZnBr3 would produce nBuZnBr2 and Br , both of
which are observed in significant quantities in our mass
spectra. However, the observation of small amounts of the
ꢀ
anion complex [nBuZnBr Li] (shown in Figure 3), obtained
3
from the addition of 1.5 equivalents of LiBr, suggests that
this ion pair is a more stable form of the species in solution.
The spectrum of this minor component along with the calcu-
lated intensities is shown in Figure 4. Recently, the isolation
of a doubly charged anion of this nature has been accom-
plished by Clyburneꢁs group, thus providing very strong sup-
[13]
port for our proposal.
A second observation from the DMI-only solution, was
ꢀ
the presence of small amounts of nBu ZnO (shown in Fig-
2
2
ure 1d). It is interesting to note the presence of nBu Zn, de-
2
ꢀ
ꢀ
tected indirectly through its O adduct nBu ZnO . Unlike
nBuZnBr2 and ZnBr , nBu ZnO does not show a de-
2
2
2
ꢀ
ꢀ
ꢀ
3
2
2
pendence on LiBr and ZnBr2 concentrations, other than
2ꢀ
Figure 1. Mass spectra showing the change in the ratio of the nBuZnBr
ꢀ
being present only after LiBr is added. Because this ion is
to ZnBr
3
ion intensities as a function of LiBr. The data were collected
ꢀ
created by a gas-phase reaction with O , rather than form-
with nBuZnBr in DMI solvent with the following equivalents of LiBr
2
added: a) 0; b) 0.5, c) 0.75; d) 1.0; and e) 1.5.
ing during the electrospray process, the intensity is depen-
7846
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 7845 – 7851

Negishi Cross-Coupling Reactions
FULL PAPER
Scheme 2. The proposed chain of intermediates involved in the Negishi
cross-coupling reaction. a) The main equilibria involved in THF:DMI
2
:1. b) A significant contribution in a THF-only solution. THF:DMI is
the solvent mixture that we have found to be optimal for alkyl–alkyl Ne-
[
11]
gishi coupling reactions.
Figure 3. Mass spectrum of nBuZnBr in THF:DMI 2:1 with 2 equivalents
LiBr.
ꢀ
tively large amount of ZnBr3 presumably results from the
Schlenk equilibrium in THF-only solvent. We have used
Scheme 2 to demonstrate that, when THF is the only solvent
present, addition of LiBr most likely results primarily in the
formation of the inorganic salt LiZnBr . Although DMI-
3
only solutions can also produce LiZnBr , this solvent is
3
ꢀ
2ꢀ
better at solvating nBuZnBr2 and ultimately nBuZnBr3
which is not favored in THF. Collectively, DMI resembles
N-methylimidazole (NMI) and NMP in the ability to sta-
bilize charged species.
,
[15]
NMR spectroscopy has been the method of choice for
[16]
studies of alkyl–Zn halides, even though the results from
these investigations indicate that there is rapid exchange on
the NMR timescale. The rate of interconversion can be
slowed down by using low temperatures; however, this ap-
proach does not give a true representation of the solution
chemistry for typical room-temperature reaction conditions.
Therefore any NMR study must deal with the reality that
the observed resonances are not a snap-shot of what is oc-
curring in solution, but rather a time-averaged collage of the
summed total of all of the various equilibria present at any
given time.
Figure 4. Comparison of observed and theoretical (heavy vertical line)
2ꢀ
+
isotopic distributions of nBuZnBr
3
Li in the mass spectrum obtained
from 1.5 equivalents of LiBr with nBuZnBr in THF:DMI 2:1.
ꢀ
dent on O2 availability, rather than its concentration in so-
lution. Thus the presence of oxygen varies from sample to
sample and solvent to solvent. As a result, although
ꢀ
nBu ZnO2 can confirm the presence of nBu Zn in the solu-
2
2
tion, it cannot be used as measure of relative concentration,
To simulate reaction conditions without the Pd catalyst,
we have taken aliquots of reaction solutions and studied the
reagents in situ. Fortunately, there is little spectral overlap
between the butylzinc species and the solvents used.
Figure 5 shows the variability in the proton spectra of
ꢀ
ꢀ
as we have done with nBuZnBr2 and ZnBr .
If the ratio of the intensities of nBuZnBr2 to ZnBr3 ions
3
ꢀ
ꢀ
in the mass spectrum is an indication of a favorable environ-
ment for the formation of an active transmetalation inter-
2
3
mediate, such as nBuZnBr , we can look at some other sol-
nBuZnBr and nBu Zn in different solvent systems. The
2
vent systems in which the reaction is known to proceed
proton spectra for both alkyl–Zn species show an upfield
chemical shift (0.1–0.6 ppm) for the methylene protons at
C1. The next highest field shifts are for the methyl protons
C4 (0.85–1.2 ppm). Both the methylene protons at C1 and
methyl protons at C4 in these solvent systems have a smaller
associated multiplet adjacent to the main resonance (see
arrows in Figure 5), which is attributed to ordered aggregate
formation such as the tetrameric aggregates reported by
(
NMP) or not proceed (THF) and compare them to the
DMI solvent system. Cross-coupling is known to start with
addition of approximately 1.0 equivalents of LiBr in DMI.
The mass spectrum of nBuZnBr with 1.5 equivalents of LiBr
[
14]
ꢀ
in NMP
shows a 3.5:1 intensity ratio of nBuZnBr2 to
ꢀ
ZnBr3 similar to the ratio observed with approximately
.2 equivalents of LiBr in DMI. On the other hand, the
[12]
1
[16]
mass spectrum of nBuZnBr with 2.0 equivalents of LiBr in
Boersma. With THF as a solvent (Figure 5b) these pro-
tons appear as a triplet.
ꢀ
ꢀ
THF shows a small nBuZnBr2 to ZnBr3 intensity ratio of
.7:1, equal to the ratio observed with approx. 0.5 equiva-
0
In THF the presence of significant amounts of LiZnBr in
3
lents of LiBr in DMI. It is interesting to note that while
nBuZnBr2 is present to some extent in THF, no coupling
occurs even with more than 2 equivalents of LiBr. The rela-
the mass spectra was unexpected, since the Schlenk equilib-
ꢀ
rium was believed to favor nBuZnBr in the moderately
[16]
polar THF solvent.
The increase of the intensity of
Chem. Eur. J. 2011, 17, 7845 – 7851
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7847

M. G. Organ et al.
1
3
Figure 6. C NMR spectra of addition of ZnBr
:1.
2 2
to Bu Zn in THF:DMI
2
1
Figure 5. 600 MHz H NMR spectra of n-butylzinc bromide in: a)
[
17,18]
THF:DMI 2:1; and b) THF prepared from Rieke
zinc. 600 MHz
Zn in: c) THF:DMI 2:1 d) THF. All solutions
were externally referenced to C H in a sealed capillary inside the
1
H NMR spectra of nBu
2
6
D
5
NMR tube. The arrows indicate the resonances from ordered aggregates.
ꢀ
ZnBr3 with the addition of LiBr indicates that the equilibri-
um can shift to form dialkylzinc and the inorganic salt
LiZnBr , which is contrary to what is seen when DMI is
3
present. Further study of this equilibrium by NMR can es-
tablish the shift of the Schlenk equilibrium in THF:DMI.
This was accomplished by following the dissociation of
Figure 7. Titration of nBu
of the methylene C shift adjacent to the zinc (C1) is shown.
2
Zn in THF:DMI 2:1 with ZnBr
2
. The change
1
3
nBu Zn with the addition of ZnBr by using C NMR spec-
13
2
2
troscopy. The complete spectral series is shown in Figure 6,
1
3
while Figure 7 shows the change in chemical shift of the
C
adjacent to zinc (C1) with the addition of ZnBr . At room
To observe the change in equilibrium from nBuZnBr to
2
ꢀ
temperature the exchange brought about by the Schlenk
nBuZnBr2 we monitored the addition of LiBr to nBuZnBr.
1
3
1
equilibrium is rapid and the C signal of C1 gradually shifts
This change is easily followed by H NMR spectroscopy as
upfield as the concentration of ZnBr increases. The change
of the C shift is much smaller on addition of between 1.0
and 1.25 equivalents of ZnBr2 and after the addition of
shown in Figure 8. The increase of LiBr has an apparent
twofold effect upon the spectra. First, the multiplets belong-
ing to the methylene protons at C1 and the methyl protons
on C4 sharpen into triplets with the addition of LiBr.
Second, the proton chemical shift decreases from 0.35 to
0.25 ppm with the addition of LiBr. The change in the
proton shift with the addition of LiBr indicated that ex-
2
13
1
.75 equivalents, ZnBr no longer dissolves, indicating that
2
the Schlenk equilibrium favors the formation of nBuZnBr
over the nBu Zn in the solvent system THF:DMI 2:1. Since
2
the Schlenk equilibrium tends to favor nBuZnBr in both
[16]
ꢀ
THF and THF:DMI, the difference in reactivity between
the two solvent systems is thus attributed to the ability to
solvate the increasingly charged intermediates (in this case
change with Br is very rapid on the NMR timescale. Inter-
estingly, the ordered aggregate does not decrease significant-
ly in amount with increasing addition of LiBr, indicating
that the aggregate formation is consistently a prevalent
2ꢀ
we are proposing nBuZnBr3 ).
7848
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 7845 – 7851

Negishi Cross-Coupling Reactions
FULL PAPER
Figure 9. The diffusion constants for the solvents THF (white) and DMI
(
gray) and nBuZnBr
n
resonances (horizontal lines) and ordered aggre-
[21]
gate (dotted) nBuZnBr species in various solutions.
nBuZnBr in
THF:DMI 2:1 was monitored with the following equivalents of LiBr with
respect to nBuZnBr: a) 0; b) 0.25; c) 0.75; d) 1.25; and e) 1.5.
for simplicity we will call nBuZnBr ) and its minor, ordered,
n
aggregate resonance. The change of the diffusion constants
is consistent with the expected difference in size between
1
Figure 8. H NMR spectra of nBuZnBr in THF:DMI 2:1 are shown with
nBuZnBr (e.g., 0.28 ppm in Figure 8b) and the ordered ag-
n
decreasing equivalents of LiBr: a) 1.5; b) 1.25; c) 0.75; d) 0.25; and e)
gregate (e.g., 0.43 ppm in Figure 8b) also present in solu-
tion. In the absence of LiBr the THF:DMI solvent system
has similar diffusion constants for each co-solvent. However,
with the increased addition of LiBr, the average diffusion of
the DMI solvent (representing both free solvent and solva-
tion shell solvent) quickly approaches the diffusion constant
of the nBuZnBrn (e.g., compare Figure 9a with 9e). This
change may be attributed to the ability of the DMI mole-
[
19]
0
.
equilibrium process regardless of the nature of the zincate
species. Therefore the expectation that the benefit of adding
DMI as a co-solvent is solely for the ability to break up
complex aggregates is not completely accurate. It is also
1
worth noting that the H NMR spectra do not change appre-
cules to effectively solvate the charged LinBuZnBr com-
2
ciably on addition of between 1.25 and 1.5 equivalents of
plex from the nBuZnBr equilibrium. The presence of DMI
n
ꢀ
LiBr, indicating that the nBuZnBr2 species is not the active
evidently assists in the solvation of the zincates. The gradual
addition of LiBr indicates the change of the role of DMI as
a solvent as well as the increased size of the zincates present
in solution. As the ionic strength of the solution increases
from the addition of LiBr, the diffusion constant for the
DMI decreases significantly, suggesting that the DMI mole-
cules are becoming an integral part of the solvation shell
surrounding the zincates. Overall, the information afforded
from DOSY spectroscopy indicates the importance of DMI
to solvate the various charged species in solution.
2ꢀ
transmetalating species. It is more likely that nBuZnBr3 is
necessary for reaction to occur, but its formation is not
2ꢀ
highly favorable. Thus nBuZnBr3 does not appear in the
equilibrium in any significant amount until more than
ꢀ
1
.0 equivalents of Br is added ensuring the equilibrium is
ꢀ
first fully populated with the more stable nBuZnBr . As
2
coupling proceeds, the equilibrium is constantly shifting
2ꢀ
forming more nBuZnBr3 as it is consumed.
Relatively recent advances in NMR spectroscopy permit
additional insight into the nature of the solution species.
The implementation of pulsed field gradients in diffusion
Questions have been raised about the possible role of the
cation in these alkyl–alkyl Negishi coupling reactions; to ex-
amine this, a series of control experiments were performed
(see Table 1). When butylzinc bromide was prepared free of
[20]
NMR or diffusion-ordered NMR spectroscopy (DOSY)
provides a convenient method of following the time-aver-
aged behavior of the different components from the
[17]
LiX salt byproducts (entry 1)
no coupling occurs, but
1
H NMR spectra. Figure 9 presents a graphical comparison
when LiCl or LiBr (entry 2), or MgBr2 (entry 3) were
added, very high levels of conversion were obtained. The
notion that it is the hard nature of the Li or Mg cation
which is responsible for driving these coupling reactions was
examined by use of a soft/diffuse cation, tetrabutylammoni-
um (entry 4), and conversion proceeded unabated. Further,
when a strong and highly specific chelator for Li cation
([12]crown-4, entry 5) was used to sequester it from solution,
of the diffusion constants of the various species in each sol-
vent system. The relatively large diffusion constant for THF
in all systems reflects the low viscosity of THF when the
DMI co-solvent is present to solvate the zincate species.
Also apparent is the consistent relative difference in diffu-
sion constants between the main resonance (which will be a
ꢀ
weighted average of the nBuZnBr–nBuZnBr2 equilibrium,
Chem. Eur. J. 2011, 17, 7845 – 7851
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7849

M. G. Organ et al.
Table 1. Effect of the additive cation in alkyl–alkyl Negishi coupling re-
actions.
Experimental Section
All reagents were purchased from commercial sources and were used
without further purification. Dry NMP and DMI (stored over 4 ꢂ molec-
ular sieves) were purchased from Fluka and handled under Argon. THF
was distilled from sodium/benzophenone prior to use. LiBr and ZnBr
2
[
a]
Entry
Additive (2.0 equiv)
Conversion [%]
were stored and weighed in an argon-filled glovebox. All reaction vials
(screw-cap threaded, caps attached, 17ꢃ60 mm) were purchased from
Fisher Scientific. Screw thread NMR tubes were purchased from New
Era Enterprises Inc. through ACP Chemicals (Montreal QC).
1
2
3
4
5
6
–
0
LiBr/Cl
MgBr2
TBAB
LiBr/[12]crown-4
LiOAc
quantitative
92
79
61
0
[
6
D ]Benzene was also purchased from ACP Chemicals.
Mass spectrometry: Experiments were performed using an AB SCIEX
API 2000 triple quadrupole mass spectrometer. Solutions of reagents
[
a] Percent conversion was assessed by GC-MS analysis using undecane
2
(nBuZnBr, LiBr, ZnBr ) in DMI, THF, or NMP were infused into the
ꢀ
1
as a calibrated internal standard; reactions were performed in duplicate.
TurboIonSpray electrospray source at flow rates of 3–5 mLmin , typical-
ly at 0.2m concentrations. The electrospray voltage was typically ꢀ4200 V
in the negative ion mode and 5000 V in the positive ion mode. Nitrogen
ꢀ
1
was used as both curtain (flow rate ca. 2 Lmin ) and nebulazer gas (flow
coupling still proceeded very well. Finally, when a variety of
other Li salts were used (e.g., lithium acetate, entry 6), in all
cases zero coupling was observed. Thus, based on these ex-
periments, we feel confident that the role of the additive re-
volves around the anion (i.e., specifically chloride or bro-
mide) and not the cation. That said, we have observed that
the MX salt must be reasonably soluble, thus we believe
that the primary role of the cation is to bring the additive
into solution.
ꢀ1
rate ca. 8 Lmin ) to reduce the possibility of oxidation, and as the colli-
sion gas for MSxMS experiments. For the most part, the identification of
ions was possible from simple mass spectra. The presence of bromine and
zinc atoms provided very characteristic isotopic patterns. When necessa-
ry, MSxMS experiments were performed on selected ions to confirm
their identity. Use of nitrogen nebulazer gas resulted in some amount of
discharge around the electrospray needle in the negative ion mode. Typi-
cally air, with oxygen acting as electron scavenger, was used as the nebu-
ꢀ
lazer gas to prevent this. The discharge created O
2
ions that could pro-
ceed to react with other constituents of the electrospray plume and so
create adduct ions such as Bu ZnO
2
2
.
NMR spectroscopy: NMR spectra were recorded at room temperature
ꢄ
on a Bruker DRX 600 spectrometer equipped with an Accustar x,y,z
Conclusion
1
13
15
gradient amplifier. A 5 mm TXI ( H, C, N) z-gradient probe was used
1
13
for all NMR experiments. The chemical shifts for H and C NMR spec-
tra are given in parts per million (ppm) and referenced to the residual
In summary, we have used mass spectrometry to identify the
negatively charged zinc species that are formed upon addi-
tion of LiBr to nBuZnBr in THF, 3-dimethyl-2-imidazolidi-
none (DMI), N-methyl-2-pyrrolidone (NMP), and mixtures
of these solvents. In less polar solvent (i.e., THF) solvation
of nBuZnBr32 is not favored and the Schlenk equilibrium
shifts towards ZnBr and nBu Zn, which is not an active
1
3
proton and C triplet signal of the deuterated benzene solvent at 7.15
and 128.0 ppm respectively. A sealed capillary of [D ]benzene was added
6
to each NMR tube to act as both a lock solvent as well as an external ref-
erence standard for all experiments. All NMR experiments were per-
formed using the standard Bruker pulse program library without further
ꢀ
1
modification. The H NMR spectra were acquired from 8 scans with a
1
spectral width of 8865 Hz using 32k data points. The H spectra were pro-
2
2
1
3
cessed by using a line broadening of 0.3 Hz. The C NMR spectra were
acquired from approximately 300 scans with a spectral width of 42372 Hz
transmetalating species. In DMI or NMP, the ability to form
more highly charged organometallics is enhanced driving
1
3
using 64k data points. The C spectra were processed by using a line
broadening of 2.0 Hz. The diffusion NMR experiments were acquired
the
equilibrium
nBuZnBr2 !nBuZnBr
successively
through
nBuZnBr!
ꢀ
2ꢀ
as progressively more LiBr is
[23]
3
using the method described by Altieri et al with a big-delta value of
added to the reaction mixture. These conclusions are very
well corroborated by the NMR studies conducted, which in-
dicate that the highly polar solvent (i.e., DMI or NMP) is
essential for stabilizing the higher order zincates in solution
formed by the addition of LiBr. Alkyl–alkyl Negishi cross-
coupling reactions catalyzed by Ni complexes are known to
take place at room temperature in THF. However, while Pd
is believed to adhere strictly to the 0 and II oxidation states
throughout the catalytic cycle, Ni has been proposed to
cycle through I and III by means of modified mechanisms. It
is possible that if single-electron processes are operative
with Ni that THF is suitable to facilitate such cycles. Col-
lectively, these results put together the most current under-
standing of the transmetalation process in the Negishi re-
action, which will prove informative for the many users of
this central cross-coupling reaction.
150 ms and 4.0 ms for little delta. The gradient strength was ramped be-
tween 5 and 95% of the amplifier output using 8 scans for each of 64
steps (see Figure 3 in the Supporting Information for a sample plot of dif-
fusion data). Diffusion data analysis was performed using the
XWINNMR 3.0 (Rheinstetten) spectrometer software. All diffusion con-
stants were normalized to an average diffusion constant determined from
1
13
6 5
the C D H reference. The connectivity of the H and C resonances was
monitored using 2D HSQC and 2D TOCSY experiments (see Figures 4–7
in the Supporting Information for sample data). Both 2D experiments
were acquired with 1024ꢃ256 data points using 4 scans per increment.
One zero fill was performed in the indirect domain with the application
of shifted sinebell window functions.
Synthetic procedures
[22]
Synthesis of n-butylzinc bromide (1m solution in DMI or NMP): Zinc
dust (9.80 mg, 15 mmol, 1.5 equiv) was added to a 25 mL round-bottomed
flask under argon. Iodine (130 mg, 0.5 mmol, 0.05 equiv) was then added
to the flask under a cone of argon followed by DMI or NMP (9 mL).The
mixture was stirred at room temperature until the red color of iodine dis-
appeared (ca. 5 min). At that point n-butyl bromide ( 1 mL, 10 mmol,
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
1
.0 equiv) was added through a syringe to the flask and transferred to an
[
24]
oil bath set at 708C for 16 h.
7850
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 7845 – 7851

Negishi Cross-Coupling Reactions
FULL PAPER
Synthesis of n-butylzinc bromide (1m solution in THF): Naphthalene
(Eds.: H. Baltes, W. Gçpel, J. Hesse), Wiley-VCH, Weinheim, 1996,
pp. 10–20.
(
400 mg, 3.2 mmol, 0.24 equiv) was added to a 25 mL round-bottomed
flask, and then it was flame-dried under argon. Following the addition of
freshly distilled THF (5 mL), freshly cut lithium wire (0.22 g, 32 mmol,
[3] a) P. Knochel, M. I. Calaza, E. Hupe in Metal-Catalyzed Cross-Cou-
pling Reactions, Vol. 2, Wiley-VCH, Weinheim, 2004, pp. 619–670;
for a recent review on the Negishi reaction, see: b) C. Valente,
M. E. Belowich, N. Hadei, M. G. Organ, Eur. J. Org. Chem. 2010,
4343–4354; c) L. Jin, C. Liu, J. Liu, F. Hu, Y. Lan, A. S. Batsanov,
J. A. K. Howard, T. B. Marder, A. Lei, J. Am. Chem. Soc. 2009, 131,
16656–16657.
2
.4 equiv) was added to the flask. The solution changed from colorless to
dark green after a couple of minutes and it was stirred for 2 h. ZnBr
3.6 g, 16 mmol, 1.2 equiv) and THF (8 mL) were placed in a separate
0 mL flame-dried flask under argon. This ZnBr solution was then
2
(
1
2
added dropwise to the 25 mL flask over a period of one hour. After stir-
ring for one additional hour, n-butyl bromide (1.4 mL, 13.3 mmol,
[4] a) E.-i. Negishi in Aspects of Mechanism and Organometallic
Chemistry, Plenum, New York, 1978, pp. 285–294; b) E.-i. Negishi,
J. Organomet. Chem. 2002, 653, 34–40; c) S. Baba, E.-i. Negishi, J.
Am. Chem. Soc. 1976, 98, 6729–6731.
1
.0 equiv) was added to the flask and it was sealed and stirred at room
[
25]
temperature for 16 h.
Synthesis of dibutylzinc (0.6m solution in THF): A 50 mL round-bot-
tomed flask equipped with a stir bar, reflux condenser, and a rubber
septum was flame-dried under vacuum. The apparatus was then purged
with argon and allowed to cool to room temperature. A second 50 mL
round-bottomed flask with stir bar was also flame-dried using the same
procedure. Magnesium metal turnings (950 mg, 39 mmol, 1.96 equiv)
were added to the first flask under a cone of argon, which was then re-
sealed, purged with argon, and then THF (5 mL) was added. n-Butyl bro-
mide (2.1 mL, 20 mmol, 1.0 equiv) and THF (23 mL) were added to the
second round-bottomed flask. The reaction was then initiated by warm-
ing with a heat gun and the remaining halide was cannulated dropwise to
the first flask maintaining a steady exotherm. After 2 h, the Grignard so-
lution was cannulated to a 50 mL flame-dried flask under argon where
[
5] a) N. Miyaura, A. Suzuki, J. Chem. Soc. Chem. Commun. 1979, 866–
8
2
67; b) N. Miyaura, K. Yamada, A. Suzuki, Tetrahedron Lett. 1979,
0, 3437–2440; c) N. Miyaura, T. Ishiyama, H. Sasaki, M. Ishikawa,
M. Satoh, A. Suzuki, J. Am. Chem. Soc. 1989, 111, 314–321.
6] K. Matos, J. A. Soderquist, J. Org. Chem. 1998, 63, 461–470.
7] B. P. Carrow, J. F. Hartwig, J. Am. Chem. Soc. 2011, 133, 2116–2119.
8] a) A. Krasovskiy, V. Malakhov, A. Gavryushin, P. Knochel, Angew.
Chem. 2006, 118, 6186–6190; Angew. Chem. Int. Ed. 2006, 45, 6040–
[
[
[
6
2
044; b) H. Ren, G. Dunet, P. Mayer, P. Knochel, J. Am. Chem. Soc.
007, 129, 5376–5377; c) N. Boudet, S. Sase, P. Sinha, C.-Y. Liu. A.
Krasovskiy, P. Knochel, J. Am. Chem. Soc. 2007, 129, 12358–12359;
d) A. Metzger, M. Schade, P. Knochel, Org. Lett. 2008, 10, 1107–
1
110.
[9] H. Ochiai, M. Jang, K. Hirano, H. Yorimitsu, K. Oshima, Org. Lett.
008, 10, 2681–2683.
2
upon ZnBr (2.16 g, 9.6 mmol, 0.5 equiv) was added to the Grignard solu-
[
26]
tion and the reaction mixture stirred for 16 h. The inorganic salt pre-
cipitate in the reaction mixture was allowed to settle for 8 h at room tem-
perature leaving a clear solution of salt-free dibutyl zinc.
2
[
10] a) K. Koszinowski, P. Bçhrer, Organometallics 2009, 28, 100–110;
b) K. Koszinowski, P. Bçhrer, Organometallics 2009, 28, 771–779.
11] a) M. G. Organ, S. Avola, I. Dubovyk, N. Hadei, E. A. B. Kantchev,
C. J. OꢁBrien, C. Valente, Chem. Eur. J. 2006, 12, 4749–4755;
b) G. T. Achonduh, N. Hadei, C. Valente, S. Avola, C. J. OꢁBrien,
M. G. Organ, Chem. Commun. 2010, 46, 4109–4111.
1
3
General procedure for sample preparation for C NMR spectra depicted
in Figures 6 and 7 (dibutylzinc+ZnBr titration): Dibutyl zinc (0.3 mL,
.2 mmol, 0.6m in THF), distilled THF (0.1 mL), and DMI (0.2 mL) were
added to an oven-dried NMR tube under argon. ZnBr (0–180 mL, 0–
.5 equiv) was then added to this solution. These solutions were charac-
terized by 1D and 2D NMR at 600 MHz.
[
2
0
2
1
[
[
12] Although peak intensities are only semi-quantitative, the trend is
consistent over the concentration range tested.
1
General procedure for sample preparation for the H NMR spectra de-
picted in Figure 8 (n-butylzinc bromide with LiBr titration): n-Butylzinc
bromide (0.2 mL, 0.2 mmol, 1m in DMI or NMP or THF) and distilled
THF (0.4 mL; or DMI or NMP) were added to an oven-dried NMR tube
under argon. LiBr (0–180 mL, 0–1.5 equiv) was then added to this solu-
tion. These solutions were characterized by 1D and 2D NMR at
13] I. S. MacIntosh, C. N. Sherren, K. N. Robertson, J. D. Masuda, C. C.
Pye, J. A. C. Clyburne, Organometallics 2010, 29, 2063–2068.
14] See Figure 1 in the Supporting Information for sample data.
15] S. Inoue, Y. Yokoo, J. Organomet. Chem. 1972, 39, 11–16.
16] a) J. Boersma, J. G. Noltes, Tetrahedron Lett. 1966, 7, 1521–1525;
b) A. Guijarro in The Chemistry of Organozinc Compounds, Part 1;
Wiley, New York, 2006, pp. 85–94.
[
[
[
6
00 MHz.
[
17] The activated zinc used in these experiments was prepared by the
addition of a THF solution of anhydrous zinc bromide to a mixture
of lithium and naphthalene in THF; see the Experimental Section
for removal of LiBr during the synthesis of nBuZnBr.
Acknowledgements
[
18] R. D. Rieke, M. V. Hanson, J. D. Brown, J. Org. Chem. 1996, 61,
This work was supported by NSERC (Canada) and the ORDCF (On-
tario).
2
726–2730.
[
[
19] See Figure 2 in the Supporting Information for sample data.
20] a) K. F. Morris, C. S. Johnson Jr. , J. Am. Chem. Soc. 1992, 114,
3
2
139–3141; b) C. S. Johnson, Jr., Prog. NMR Spectrosc. 1999, 34,
03–256.
[
1] a) E.-i. Negishi in Handbook of Organopalladium Chemistry for Or-
ganic Synthesis, Vol. 1, Wiley-Interscience, New York, 2002, pp. 229–
[
[
21] See Figure 3 in the Supporting Information for sample data.
22] V. B. Phapale, D. J. Cꢅrdenas, Chem. Soc. Rev. 2009, 38, 1598–1607,
and references therein.
2
48; b) M. Uchiyama, T. Furuyama, M. Kobayashi, Y. Matsumoto,
K. Tanaka, J. Am. Chem. Soc. 2006, 128, 8404–8405.
[
2] It has been suggested that the catalysis could be occurring on the
nanoparticle surface of Pd; however, previous studies have shown
that the ligands are essential for coupling to occur; a) N. Hadei,
E. A. B. Kantchev, C. J. OꢁBrien, M. G. Organ, Org. Lett. 2005, 7,
[
23] A. S. Altieri, D. P. Hinton, R. A. Byrd, J. Am. Chem. Soc. 1995, 117,
7
566–7567.
[
[
24] S. Huo, Org Lett. 2003, 5, 423–425.
25] L. Zhu, R. M. Wehmeyer, R. D. Rieke, J. Org. Chem. 1991, 56,
0
II
3
805–3807. In this study, a variety of Pd and Pd sources known to
1
445–1453.
26] M. L. H. Green, S. L. J. Conway, L. H. Doerrer, Polyhedron 2005,
4, 1388–1403.
produce Pd nanoparticles were examined (in the absence of the N-
heterocyclic carbene) and zero coupling was observed despite the
evident presence of Pd black. The reactions performed for the cur-
rent study are all conducted at room temperature and were fully ho-
mogeneous; b) J. W. Grate, G. C. Frye in Sensors Update, Vol. 2
[
2
Received: April 5, 2011
Published online: June 8, 2011
Chem. Eur. J. 2011, 17, 7845 – 7851
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7851