Elongated Gilman cuprates: The key to different reactivities of cyano- and iodocuprates

Article abstract of DOI:10.1021/ja501055c

In the past the long-standing and very controversial discussion about a special reactivity of cyano- versus iodocuprates concentrated on the existence of higher-order cuprate structures. Later on numerous structural investigations proved the structural equivalence of iodo and cyano Gilman cuprates and their subsequential intermediates. For dimethylcuprates similar reactivities were also shown. However, the reports about higher reactivities of cyanocuprates survived obstinately in many synthetic working groups. In this study we present an alternative structural difference between cyano- and iodocuprates, which is in agreement with the results of both sides. The key is the potential incorporation of alkyl copper in iodo but not in cyano Gilman cuprates during the reaction. In the example of cuprates with a highly soluble substituent (R = Me ₃SiCH₂) we show that in the case of iodocuprates during the reaction several copper-rich complexes are formed, which consume additional iodocuprate and provide lower reactivities. To confirm this, a variety of highly soluble copper-rich complexes were synthesized, and their molecular formulas, the position of the equilibriums, their monomers and their aggregation trends were investigated by NMR spectroscopic methods revealing extended iodo Gilman cuprates. In addition, the effect of these copper-rich complexes on the yields of cross-coupling reactions with an alkyl halide was tested, resulting in reduced yields for iodocuprates. Thus, this study gives an explanation for the thus far confusing results of both similar and different reactivities of cyano- and iodocuprates. In the case of small substituents the produced alkyl copper precipitates and similar reactivities are observed. However, iodocuprates with large substituents are able to incorporate alkyl copper units. The resulting copper-rich species have less polarized alkyl groups, i.e. gradually reduced reactivities.

Full text of DOI:10.1021/ja501055c

Article
pubs.acs.org/JACS
Elongated Gilman Cuprates: The Key to Different Reactivities of
Cyano- and Iodocuprates
Maria Neumeier⊥ and Ruth M. Gschwind*
Institut fur Organische Chemie, Universitat Regensburg, Universitatsstraße 31, D-93053 Regensburg, Germany
̈
̈
̈
S
* Supporting Information
ABSTRACT: In the past the long-standing and very con-
troversial discussion about a special reactivity of cyano- versus
iodocuprates concentrated on the existence of higher-order
cuprate structures. Later on numerous structural investigations
proved the structural equivalence of iodo and cyano Gilman
cuprates and their subsequential intermediates. For dimethylcup-
rates similar reactivities were also shown. However, the reports
about higher reactivities of cyanocuprates survived obstinately in
many synthetic working groups. In this study we present an
alternative structural difference between cyano- and iodocuprates,
which is in agreement with the results of both sides. The key is
the potential incorporation of alkyl copper in iodo but not in cyano Gilman cuprates during the reaction. In the example of
cuprates with a highly soluble substituent (R = Me3SiCH2) we show that in the case of iodocuprates during the reaction several
copper-rich complexes are formed, which consume additional iodocuprate and provide lower reactivities. To confirm this, a
variety of highly soluble copper-rich complexes were synthesized, and their molecular formulas, the position of the equilibriums,
their monomers and their aggregation trends were investigated by NMR spectroscopic methods revealing extended iodo Gilman
cuprates. In addition, the effect of these copper-rich complexes on the yields of cross-coupling reactions with an alkyl halide was
tested, resulting in reduced yields for iodocuprates. Thus, this study gives an explanation for the thus far confusing results of both
similar and different reactivities of cyano- and iodocuprates. In the case of small substituents the produced alkyl copper
precipitates and similar reactivities are observed. However, iodocuprates with large substituents are able to incorporate alkyl
copper units. The resulting copper-rich species have less polarized alkyl groups, i.e. gradually reduced reactivities.
However, the iodocuprates gave us a surprise while varying
the cuprate substituents in our investigations of Cu(III)
intermediates18,22 from methyl groups to the larger, highly
soluble Me3SiCH2 moieties and using substoichiometric
amounts of organolithium compounds. With Me3SiCH2
substituents we suddenly detected copper rich Cu(I) complexes
consisting of combinations of Gilman cuprates and uncharged
copper alkyl species in diethyl ether. The corresponding
Me3Cu2Li complexes had been detected previously in reactions
of methyl copper with less than one equivalent of
methyllithium in Me2O or THF.28 But in diethyl ether methyl
copper is known to be not soluble because of the formation of
extended aggregates29,30 and we never had detected such
complexes for dimethylcuprates during nearly a decade of
NMR measurements in diethyl ether. In principle, this is
nothing unexpected, because it is well-known that uncharged
homoleptic organocopper compounds RnCun often exist as
highly aggregated species in the solid state and are insoluble in
common organic solvents in case of small substituents (e.g., R =
Me).29,30 Furthermore, it is well established that disaggregation
of these polymers can be achieved by bidentate ligands (e.g., 2-
INTRODUCTION
■
The long-standing and very controversial scientific discussion
about a special reactivity of cyanocuprates seemed to be solved
during the past decade.1 In addition, the structures of all
relevant copper complexes seemed to be elucidated.2 Only two
crystal structures showed the principal existence of homoleptic
higher-order structures [R3Cu]2‑ in the solid state.3,4 All others
reported for R2CuLi·LiCN systems exclusively cyano Gilman
type cuprates in their monomeric5−9 as well as dimeric
structures.10,11 The aggregation levels of the supramolecular
oligomers of cyano- and iodocuprates were found to differ-
entiate slightly12 and correlated to reactivity.13 But all the
structural studies on Cu(I) and Cu(III) cuprate intermediates
did not reveal any significant deviations for cyano- and
iodocuprates.14−22 In addition, Bertz et al. found in a direct
comparison of the reactivity profiles of R2CuLi·LiI and
R2CuLi·LiCN (R = Me, Bu) in Michael addition reactions
that cyanocuprates have similar or even reduced reactivities
compared to iodocuprates.23,24 Hence, the early e.g. Alexakis,25
Nakamura,26 van Koten27 and later on persistent but often oral
reports of better performances of cyanocuprates from various
groups in the synthetic community were attributed to variations
in the experimental conditions.
Received: February 3, 2014
Published: March 24, 2014
© 2014 American Chemical Society
5765
dx.doi.org/10.1021/ja501055c | J. Am. Chem. Soc. 2014, 136, 5765−5772

Journal of the American Chemical Society
Article
(dimethylamino)phenyl ligands31), bulky substituents32−34
(e.g., R = mesityl, Me3SiCH2) or coordinating solvents3,32,33
(e.g., dimethyl sulfide, tetrahydrothiophene). Thus, clearly a
gradual better solubility of copper rich iodocuprate complexes
is expected for R = Me3SiCH2 compared to R = Me. However,
to our surprise we found an extreme situation: not detectable
signals for copper rich cuprates for R = Me compared to a
quantitative formation in the case of R = Me3SiCH2.
This raised two main questions: First, does the presence of
copper rich iodocuprates influence the reactivity of iodocup-
rates and thus lead potentially to differences compared to
cyanocuprates? Second, can the large solubility differences of
copper rich iodocuprates depending on the size of their
substituents explain the confusions in the various reports about
better or equal performances of cyano- versus iodocuprates?
In principle an influence is highly probable (see Scheme 1):
In both conjugate addition reactions as well as cross coupling
reactivities of cyanocuprates,36−39 i.e. yields between 82 to
100% in reactions of (n-Bu)2CuLi·LiCN with unactiviated
secondary alkyl halides in THF, while previous reports of
reactions with (n-Bu)2CuLi and Me2CuLi led only to yields
between 12 and 21%29,40,41 and also, later on, better
stereoselectivities using cyano- instead of iodocuprates were
published.42 Indeed, the reactivity differences reported were
dependent of the size of the cuprate substituent. While for
Me2CuLi·LiX (X = I, SCN) similar yields were found,
significantly improved yields were reported for (n-Bu)2CuLi·
LiSCN and (n-Pr)2CuLi·LiSCN.43 Furthermore, primary alkyl
iodides were observed to react readily with R2CuLi·LiCN, and
for substitution reactions with poorer leaving groups (Br, Cl)
extremely mild reaction conditions could be applied, i.e. lower
temperatures than required for R2CuLi.44 Van Koten reported
higher reactivities for cyanocuprates with aminoaryl ligands in
substitution reactions van Koten27 and Alexakis obtained the
highest yields in the boron fluoride-promoted opening of
epoxides using n-butyl-cyanocuprates, e.g. Alexakis.25 Further-
more, House published a reactivity of Ph2CuLi·PhLi higher
than that of Ph2CuLi in coupling reactions with aryl iodide,45
which would be in agreement with a regeneration of Gilman
cuprate from copper-rich complexes. Drawbacks of all of these
relative reactivity reports are potential deviations in the
experimental conditions between cyano- and iodocuprates.
Therefore, Bertz and Ogle reported a direct comparison of
reactivity profiles for R2CuLi·LiI and R2CuLi·LiCN (R = Me,
Bu).23,24 In Michael addition reactions, similar or even reduced
reactivities of cyanocuprates were found. However, in the
substitution reaction of R2CuLi·LiX (R = Bu; X = I, CN) with
cyclohexyl iodide, the reaction which initialized the higher-
order discussion, Bertz and Ogle initially reported higher
reactivities for the cyano Gilman cuprate compared to the iodo
Gilman cuprate,24 which were later on brought into line by
several renormalization procedures taking side reactions into
account.46
Scheme 1. Formation of RCu in Cross-Coupling (A) and
Conjugate Addition Reactions (B); Further Reaction of RCu
Depending on the Salt and Substituent Used (C, D, E); (F)
28
Schematic Structure of [R3Cu2Li]2
reactions of iodo and cyano Gilman cuprates alkyl copper
compounds (RCu) are formed as byproducts (see Scheme 1A,
B).29,35 Next the fate of RCu depends on the salt used. In case
of cyano Gilman cuprates, LiCN and RCu form an heteroleptic
cuprate and the remaining cyanocuprates are unaffected (see
Scheme 1C). For iodo Gilman cuprates the size of the
substituent is decisive. RCu with small substituents precipitates.
Again the remaining iodo Gilman cuprates are unaffected (see
Scheme 1D). However, RCu with large substituents can form
soluble copper rich complexes with an additional iodo Gilman
cuprate (see Scheme 1E). As a result the respective iodo
Gilman cuprate is overstoichiometrical consumed during the
reaction and the reactivity of the copper rich complexes
becomes more and more important during the course of the
reaction. This should reduce the reactivity of such iodocuprates,
because copper rich complexes (see Scheme 1F) have less
polarized i.e. less reactive alkyl groups. To sum up, for small
substituents (e.g., dimethylcuprates) the reactivities of iodo and
cyano Gilman cuprates are expected to be similar, whereas for
large substituents the reactivity and possibly also the selectivity
of iodocuprates compared to that of cyanocuprates should be
significantly affected by the formation of copper-rich
complexes. This should be evident mainly in reactions requiring
maximal cuprate reactivities, e.g. cross-coupling reactions with
sterical hindered reaction partners.
In order to clarify whether the existence of copper-rich iodo
Cu(I) complexes in the case of a large cuprate substituent are
the reason for the puzzling and long-standing reports about
reactivity differences between cyano- and iodocuprates, we
synthesized a variety of highly soluble copper-rich complexes
R
1−4Cu1−3Li0,1X0,1 (R = Me3SiCH2; X = I, 13CN) and
investigated their structures and reactivity in cross-coupling
reactions with methyl iodide.
RESULTS AND DISCUSSION
■
Model System. To test the influence of subsequently
formed copper-rich complexes on the reactivity and selectivity
of organocuprate reactions, Gilman cuprates R2CuLi·LiX with
large substituents (R = Me3SiCH2; X = I, 13CN) were used.
Furthermore, organocopper complexes derived from varying
RLi:CuX ratios less than 2:1 were synthesized to yield copper-
rich complexes. Et2O was chosen as the solvent to be able to
apply in our over-the-years developed approach for the
structure elucidation of organocuprates in this sol-
vent.2,10−13,18,19 In order to compensate the low solubility of
copper-rich complexes in Et2O, the extended substituent R =
Me3SiCH2 was chosen. To reveal the influence of copper-rich
complexes on the reactivity of organocuprate reactions, as a test
system the cross-coupling reaction between the Gilman
cuprates R2CuLi·LiX (R = Me3SiCH2; X = I, 13CN) and
methyl iodide was chosen.47 As discussed above, reactivity
differences between cyano- and iodocuprates were reported
Having this filter at hand, we searched again the literature
results for reactivity differences between cyano and iodo
Gilman cuprates. Thus, Liphutz initially reported enhanced
5766
dx.doi.org/10.1021/ja501055c | J. Am. Chem. Soc. 2014, 136, 5765−5772

Journal of the American Chemical Society
Article
mainly in substitution reactions with sterically hindered alkyl
halides (critical reactivity) combined with coordinating solvents
or larger cuprate substitutents (improved solubility). To mimic
such a sterical hindered substitution reaction providing both
critical reactivity and sufficiently high solubility of copper-rich
complexes, we chose a small alkyl halide (CH3I) and a large
relatively unreactive cuprate substituent (R = Me3SiCH2),
which is usually used as a dummy ligand.48 In addition, all
NMR spectroscopic measurements were performed at low
temperatures between 170 and 185 K, to resolve the different
copper-rich complex species. At typical synthetic conditions28
(195−273 K) only broad or averaged signals are observed (for
details see SI).
Formation of Copper-Rich Iodocopper(I) Complexes.
Copper-rich iodocopper(I) complexes were synthesized by
adding seven different ratios of Me3SiCH2Li (RLi) to an
ethereal suspension of CuI at 170 K. The formation of the
respective soluble copper(I) complexes started immediately,
which was visible by specific color changes (vide infra) and was
spectroscopically investigated by NMR at 170 K (see, e.g., the
1H spectra in Figure 1).
copper complex, and the determination of their exact number is
not feasible. Therefore, in Scheme 2 and in the following text
the LiI units are omitted.
At overstoichiometric amounts of RLi (see RLi:CuI = 3.0:1
in Figure 1), clear and colorless samples are formed containing
R2CuLi and the excess of RLi unreacted in solution.49 In the
colorless sample with RLi:CuI = 2.0:1 exclusively R2CuLi was
detected. In the clear and pale-brown solution with RLi:CuI =
1.8:1, additionally the first copper-rich complex R3Cu2Li
(purple in Figure 1), was observed with two chemically
nonequivalent organic moieties in a 1:2 ratio. Lowering further
the organolithium to copper salt ratio to RLi:CuI = 1.6:1 a
clear and yellow sample indicated the formation of R3Cu2Li
and additionally R4Cu3Li (R = Me3SiCH2) (light green in
Figure 1), with two chemically nonequivalent organic moieties
in a 1:1 ratio. In addition, small amounts of the alkyl copper
compound RCu (R = Me3SiCH2) were detected (light blue in
Figure 1). From the ratio RLi:CuI = 1.4:1 onward, only
negligible amounts of R2CuLi were detected, and exclusively
the copper-richer complexes R3Cu2Li, R4Cu3Li and RCu were
observed. At RLi:CuI = 1.0:1 was RCu exclusively detected.
Below a 1:1 ratio (RLi:CuI = 0.6:1), the solution was still clear
and yellow, but CuI partly remained at the bottom of the NMR
tube and a signal of an additional complex (dark blue in Figure
1) was observed, which fits very well to an X-ray structure
reported by van Koten et al.,50 for R2Cu3I. However, an
independent, exact determination of the stoichiometry of
R2Cu3I is obscured by severe spectral overlap.
In summary, in seven different mixtures of RLi and CuI, a
variety of different Cu(I) complexes with the general molecular
formula R1‑4Cu1‑3Li0,1I0,1 were observed. At ratios of RLi to
CuI larger than 2:1, exclusively R2CuLi (R = Me3SiCH2) was
present. By reducing the content of RLi, the position of the
equilibrium was shifted toward several copper-richer complexes.
The more RCu units are incorporated in these complexes the
more downfield shifted appear their proton signals compared to
those of R2CuLi. This indicates a stepwise-reduced polarization
as well as reactivity in the row R2CuLi, R3Cu2Li, R4Cu3Li,
RCu, and R2Cu3I, which is well-known for R2CuLi and RCu
from the introduction of the Gilman cuprates.
Figure 1. Multiple copper-rich complexes are detectable at ratios
smaller than 2:1 using Me3SiCH2Li (RLi) and CuI from their CH2-
signals in the high-field section of H spectra in d6-diethyl ether at
1
Formation of Cyanocopper(I) Complexes in Diethyl
Ether. Next, the complex formation of Me3SiCH2Li (RLi) with
Cu13CN was studied, to solve the question whether copper-rich
complexes are also formed in the case of cyano copper
complexes but remained undetected so far. Therefore, six
samples were prepared with RLi to Cu13CN ratios from 3.0:1
to 0.3:1. These samples were investigated by H as well as
DOSY experiments at 170 K in diethyl ether (see Scheme 3 and
Figure 2, for details see SI).
170 K.
The signals of RLi and the Gilman cuprate R2CuLi·LiI were
assigned on the basis of previous investigations.12 The
molecular formulas of the copper-rich complexes were
determined by integral analysis of the signals downfield from
the Gilman cuprate (for details see SI). The monomeric
1
structures were determined with H,13C hetero multi bond
1
correlation (HMBC) spectra and aggregation trends were
studied with diffusion ordered spectroscopy (DOSY) measure-
ments. With these techniques, not only the Gilman cuprate
R2CuLi·LiI (R = Me3SiCH2) but also overall four copper-rich
complexes with the formal monomers R3Cu2Li·2LiI, R4Cu3Li·
3LiI, RCu·LiI, and R2Cu3I·2LiI were detected (see Scheme 2).
Regarding the formal stoichiometries of these complexes our
previous12 as well as actual (see below) DOSY investigations
show that the LiI units are at least partially separated from the
Scheme 3. Formation of Heteroleptic Cyanocuprates
Depending on the stoichiometric ratio, exclusively RLi,
homoleptic cyanocuprate R2CuLi·Li13CN (R = Me3SiCH2,
orange in Figure 2), and the soluble heteroleptic cyanocuprate
RCu13CNLi (magenta in Figure 2) were detected, but not any
trace of additional copper-rich complexes or higher-order
cuprates. Thus, in accordance with previous studies36 no
copper-rich complexes are detected in the case of cyano copper
complexes.
Scheme 2. Formation of Copper-Rich Iodocuprates
5767
dx.doi.org/10.1021/ja501055c | J. Am. Chem. Soc. 2014, 136, 5765−5772

Journal of the American Chemical Society
Article
Figure 2. Heteroleptic cyanocuprate but no copper-rich complexes are
detectable at ratios smaller than 2:1 using Me3SiCH2Li (RLi) and
Cu13CN from their CH2− (<−1 ppm) and CH3-signals (>−0.5 ppm)
Figure 3. Structure identification of R3Cu2Li as Gilman cuprate
1
extended by one RCu unit based on the high field sections of H (A)
and 1H,13C HMBC spectra (B,C) of a 1.8:1 mixture of RLi and CuI at
185 K in Et2O (Me3Si region (B) scaled down for the sake of clarity).
The signals of R3Cu2Li and R2CuLi·LiI in the 1H spectrum as well as
1
in the H spectra in d6-diethyl ether at 170 K.
Structures of Iodocopper(I) Complexes in Diethyl
Ether. Next, the structures of the copper-rich iodocopper(I)
complexes were addressed. From various NMR spectroscopic
investigations on organocuprates,5,6 π-complexes51 as well as
Cu(III) complexes18,22 and in agreement with theoretical
calculations,52−55 it is known that the partly covalent copper−
alkyl bond allows for a magnetization transfer via scalar
coupling, whereas the predominantly electrostatic lithium−alkyl
interaction acts as a spin system barrier. Accordingly, the
1
their JH,C cross signals are depicted in purple and red, respectively.
The decisive cross signals are color coded according to the inset.
Figure 4. Schematic structures of RCuR(CuR)nLi·LiI (R =
Me3SiCH2, n = 0, 1, and 2) based on 1H,13C HMBC, and 1H-
DOSY measurements at 185 and 170 K. For details see text and SI.
1
existence or lack of cross signals in H,13C HMBC spectra
indicate the kind of metal in between the two groups revealing
the monomeric structure of these copper-rich complexes. In
addition, DOSY measurements can be used to elucidate the
hydrodynamic ratios of organocuprate aggregates including salt-
and solvent-coordination.12 Therefore, the structures of the
copper-rich complexes R3Cu2Li, R4Cu3Li, and RCu were
Diffusion Measurement on Iodo- and Cyanocopper(I)
Complexes in Diethyl Ether. The aggregation level of the
iodocopper complexes R2CuLi·LiI, R3Cu2Li, R4Cu3Li, RCu
and of the cyano copper complexes R2CuLi·LiCN and
RCuCNLi was determined by 1H DOSY investigations, similar
to our previous studies of R2CuLi·LiX (X = I, CN) in Et2O.12
There, in highly concentrated samples (0.6 mol/L) of RCuLi·
LiI in Et2O at 240 K slightly larger aggregates than dimers were
found with a high number of solvent molecules attached
([(R2CuLi)2(LiI)2(Et2O)7.5]1.3). In order to avoid extended
oligomeric structures at 185 and 170 K the concentrations of
the samples were drastically reduced to c = 0.07−0.13 mol/L, at
which a comparable aggregation number of 1.3 was found for
[(R2CuLi)2(LiI)2(Et2O)7.5].
1
investigated by H,13C HMBC, and DOSY measurements at
185 and 170 K. To avoid extended supramolecular
aggregates,12,13 samples with low concentrations were used
(c = 0.07−0.13 mol/L).
1
In the H,13C HMBC spectrum of R3Cu2Li, cross signals
between the chemically nonequivalent CH2 groups and to their
corresponding Me3Si moieties were detected but none between
the two chemically equivalent CH2 groups (see Figure 3). This
1H,13C HMBC coupling pattern is in perfect agreement with an
“extended” Gilman cuprate monomer shown in Figure 3, which
was first proposed by Ashby28 merely on the basis of H
1
chemical shifts. In this “extended” Gilman cuprate monomer
the two chemically equivalent CH2 groups show a cross peak to
the single CH2 moiety (3JH,C) but no cross peak among
themselves (5JH,C). Adducts of Gilman cuprates with RCu and
LiI units can be excluded on the basis of this scalar coupling
pattern.
Accordingly, the complex R4Cu3Li was investigated (for
details see SI), and a similar structure was determined, which is
elongated by one more RCu moiety (for the dimeric core
structure see Figure 4). For (RCu)n the monomeric structure is
evidently RCu. Thus, the structures R3Cu2Li and R4Cu3Li can
be derived from R2CuLi by gradual elongation of the
homocuprate by one and two RCu units, respectively, and
may be considered as charged organocuprate analogue
structures, whereas (RCu)n is an uncharged alkyl copper
structure.
The DOSY study revealed for R3Cu2Li and R4Cu3Li similar
dimeric core structures and aggregation trends as previously
found for the Gilman cuprates R2CuLi (for details see SI). Just
the amounts of LiI units and solvent molecules are slightly
reduced, which goes along with the reduced polarity of the alkyl
groups in these extended cuprates. For RCu, the experimental
diffusion coefficient excludes a monomer as well as the
tetrameric structure reported in X-ray studies34 but indicates
a higher aggregation, which would fit an octameric linear
aggregate or a corresponding cyclic system. For visualization,
the dimeric core structures for R2CuLi, R3Cu2Li, and R4Cu3Li
are schematically shown in Figure 4.
For R2CuLi·Li13CN and RCu13CNLi, the corresponding
DOSY investigations indicate the formation of moderately
higher aggregates than dimers. This is in accordance with the
known slightly higher aggregation trends for cyanocuprates
5768
dx.doi.org/10.1021/ja501055c | J. Am. Chem. Soc. 2014, 136, 5765−5772

Journal of the American Chemical Society
Article
compared to iodocuprates12 and the low concentration of the
actual samples (for details see SI).
detect the existent copper species and to determine the reaction
yields by integration.56
In Figure 5 sections of representative H spectra of these
1
Reactivity of Iodo- and Cyanocopper(I) Complexes in
Cross-Coupling Reactions with Methyl Iodide. The more
RCu units are integrated within the copper-rich iodo
reactions are presented, the reaction equations are given in
1
complexes, the larger ppm values are found for the H signals
of their CH2 groups (see Figure 1). This indicates a decreased
polarization of these organic moieties compared to the Gilman
cuprates and suggests a lower reactivity of copper-rich iodo
complexes in C−C bond formations. In addition, during a
reaction the more reactive Gilman cuprate is overstoichio-
metrically consumed at the moment copper rich complexes are
formed. Both effects are expected to reduce the reactivity of
iodocuprates compared to cyanocuprates, in case copper-rich
iodo complexes are soluble. In order to observe such reactivity
differences in the yields of typical cuprate reactions and in a
NMR spectrometer without a rapid injection unit, two main
conditions have to be met. First, the copper-rich iodo
complexes have to be soluble, e.g. due to large alkyl substituents
or by a good coordinating solvent. Second, the reactivity
required for the reaction has to be critical for cuprates, i.e. the
slightly reduced reactivity of the copper-rich complexes should
be not sufficient for this reaction. Such reactivity differences
were mainly reported for substitution reactions using sterical
hindered substrates (vide supra), and the use of diethylether for
substitution reactions reduces further the reactivity. To be able
to transfer directly all of our structural investigations and in
addition to meet both criteria, here an “inverse” setup was used:
The highly soluble but inreactive cuprates R2CuLi·LiI and
R2CuLi·Li13CN (R = Me3SiCH2) were tested in cross-coupling
reactions with methyl iodide in diethylether at low temper-
atures without and with an excess of RLi (see Scheme 4 (1)
and (2)). In addition, several blind reactions were performed to
validate the results (see Scheme 4 (3), (4), and (5)).
Figure 5. Cross-coupling reactions with stoichiometric amounts of
iodocuprates R2CuLi·LiI (R = Me3SiCH2) and CH3I show the
formation of copper-rich complexes and reduced yields of RCH3 (A).
Cyanocuprates (B) as well as both cuprates plus excess of RLi (C, D)
show higher yields of RCH3 (For exact yields see Table 1). 1H spectra
at c = 0.05−0.06 mol/L R2CuLi·LiX (X = I, 13CN) at 170 K in
Et2O-d6.
Table 1. Yields of Iodo and Cyano Gilman Cuprates in
Cross-Coupling Reactions with CH3I without and with
Excess of Me3SiCH2Li; the Corresponding Reactions Are
Given in Scheme 4
a,b
ratio
yields [%]
Scheme 4. Educts and Products of Cross-Coupling and Blind
Reactions at 170 K in Et2O; Side Products up to 5% Are
Given in Brackets
cuprate
X = I
R2CuLi·
Me3SiCH2CH3
(Me3SiCH2)2
LiX:CH3I:RLi
1a
2a
3a
4a
1b
2b
3b
4b
5
1.0:1.0:0.0
1.0:1.0:1.0
1.0:0.0:0.0
1.0:0.0:1.0
1.0:1.0:0.0
1.0:1.0:1.0
1.0:0.0:0.0
1.0:0.0:1.0
0.0:2.0:1.0
57 (31)
1 (1)
5 (2)
95 (91)
−
−
1 (<1)
3 (1)
X = CN
95 (81)
2 (−)
4 (−)
1 (−)
3 (−)
97 (99)
−
−
17
blind
reaction
a
b
c = 0.05−0.06 mol/L. c = 0.13 mol/L in brackets, of R2CuLi·LiX (X
= I, 13CN) in diethylether at 170 K. All yields were determined by
integral analysis of the corresponding proton spectra.
First an exact one-to-one stoichiometry of Gilman cuprate
and alkyl halide (R2CuLi·LiI/R2CuLi·Li13CN:CH3I:CH3Li =
1.0:1.0:0.0) was selected for both iodo- and cyanocuprate to
elucidate the possible influence of the copper-rich Cu(I)
complexes on the reactivity. Second, the reaction was
performed additionally in the presence of one equivalent
excess of the MeLi (R2CuLi·LiI/R2CuLi·Li13CN:CH3I:CH3Li
= 1.0:1.0:1.0) to enable the regeneration of the reactive Gilman
cuprates during the reaction. For both conditions two
concentrations (0.05 and 0.13 mol/L) were chosen to test
the influence of the aggregate size. Each mixture was reacted
Scheme 4, and the yields are summarized in Table 1. After the
stoichiometric reactions of R2CuLi·LiI with CH3I, indeed
substantial amounts of the copper-rich complexes R3Cu2Li,
R4Cu3Li, and RCu are detected (see Figure 5A). As expected,
the identical reaction with R2CuLi·Li13CN leads only to
RCuCNLi (see Figure 5B). Surprisingly, considering the long
reactivity discussion about iodo- and cyanocuprates but in
accordance with our considerations about the influence of
soluble copper-rich complexes, the integrals of the main
product 2,2-dimethyl-2-silabutane RCH3 (see Scheme 4)
show a formidable higher yield for the cyanocuprate R2CuLi·
1
20 min at 170 K, then H NMR spectroscopy was used to
5769
dx.doi.org/10.1021/ja501055c | J. Am. Chem. Soc. 2014, 136, 5765−5772

Journal of the American Chemical Society
Article
Li13CN (95%) than for the reaction with the iodocuprate
R2CuLi·LiI (57%).
reaction (17% RCH3). Therefore, all interpretations have to be
made cautiously. However, only traces of the metal−halide
exchange are detected (see Figure 5D), suggesting that the
blind reaction plays a minor role in this case. In addition, the
absolute yields observed for cyanocuprates are higher than
those for iodocuprates, especially for the higher concentrated
sample. According to the structural investigations and the
theoretical calculations in this field there are two possible
explanations. First, Nakamura postulated a MeLi-bridged
species MeCu(CN)Li·MeLi as a possible approach to explain
the special reactivity of cyanocuprates.48,57 The overstoichio-
metric use of an alkyl lithium compound in reaction 2b is
expected to favor the formation particularly of such structures,
and exactly this combination leads to the highest yields
observed (see entry 2b in Table 1). Second, mixed supra-
molecular aggregates of R2CuLi and RCuCNLi units could
decrease the reactivity. In accordance with the experimental
results both effects of mixed aggregates are expected to be more
pronounced at higher concentrations. Therefore, the reactivity
of cyanocuprates may be fine-tuned by the formation of such
alkyl lithium-bridged cyanocuprates or mixed aggregates with
heteroleptic cuprates, whereas the reactivity of iodocuprates is
strongly reduced in case soluble copper-rich complexes can be
formed.
Bertz and Ogle reported in their reactivity study of iodo- and
cyanocuprates46 high amounts of the side products for the
iodocuprates and low amounts for cyanocuprates formed
during the cuprate preparation. Therefore, they renormalized
the reactivities of both cuprates to their effective concen-
trations.46 In our investigation, besides RCH3 only small
amounts of the side product 2,2,5,5-tetramethyl-2,5-disilabu-
tane R2 were detected (2% for R2CuLi·Li13CN and 1% for
R2CuLi·LiI). All other possible side products46 were below the
detection limit. In general R2 can be generated on two
pathways, first during the cuprate preparation (see e.g. reaction
3 in Scheme 4) and second by metal−halide exchange and
subsequent substitution reaction (see e.g. reaction 2a and 2b in
Scheme 4). Therefore, in Table 1 the amount of R2 before (see
reaction 3a and 3b) and after the reaction with CH3I (see
reaction 1a and 1b) is given. In our case the concentrations of
R2 before the reaction are negligible for both cuprates (both
1%). Therefore, a renormalization is not necessary, and the
yields can be directly correlated with the reactivity. Next, the
influence of the concentration, i.e., larger aggregates, was tested
using the identical reactions at higher cuprate concentrations
(0.13 mol/L). Again, a significantly higher yield was observed
for R2CuLi·Li13CN (81%) than for R2CuLi·LiI (31%). As
expected for substitution reactions, which are faster in THF
than in diethyl ether, the absolute yields decreased for both
cuprates at higher concentrations, i.e., the formation of larger
aggregates.
CONCLUSIONS
■
More than three decades ago Lipshutz reported higher
reactivities of cyano- versus iodocuprates especially for large
substituents, and later on many synthetic chemists used
cyanocuprates very successfully. Lipshutz’s structural explan-
ation, the existence of higher-order cuprates, was based on an
assumption of Ashby and later on disproven by numerous
structural investigations showing cyano Gilman cuprates in
solution. For the structural model system, the dimethylcuprates,
also nearly identical reactivities were shown for iodo- and
cyanocuprates. Therefore, better performances of other
cyanocuprates observed in syntheses remained reports without
structural explanation.
These results raised the question whether the reduced
reactivity of the copper-rich complexes R3Cu2Li, R4Cu3Li, and
RCu cause this drastic reactivity difference, which only occurs
in the case of iodocuprates in combination with large, highly
soluble cuprate substituents. To prove this, additional experi-
ments with one equivalent excess of RLi were performed,
which enable the regeneration of the Gilman cuprates during
the reaction. Under these overstoichiometric conditions the
yield with R2CuLi·LiI increased dramatically from 57% to 95%,
whereas in the case of R2CuLi·Li13CN only a slight increase
from 95% to 97% yield was observed. Also for the higher
concentrated samples a similarly pronounced trend was found
(for spectra see SI Figure 6). There, the yield for R2CuLi·LiI
increased by 60% up to 91% and in the case of R2CuLi·Li13CN
also a pronounced effect (+18%) with the highest yield of all
(99%) was detected. In order to exclude that these increased
yields are only achieved by the direct reaction between RLi and
CH3I, a blind reaction (RLi: 0.07 mol/L; CH3I: 0.13 mol/L)
using 2 equiv of CH3I was investigated at 162 K to ensure that
the effects of this blind reaction are not underestimated (see
reaction 5 in Table 1). As an upper limit for this direct reaction
of RLi and CH3I, thus, a yield of 17% of RCH3 could be
determined. This blind reaction yielded also the products of
metal−halide exchange CH3Li and RI (see SI Figure 7). In the
case of the iodocuprate, R2CuLi·LiI, indeed also these products
of metal−halide exchange CH3Li and RI are detected (see
Figure 5C), indicating that the blind reaction plays a role.
However, for iodocuprates plus one equivalent of RLi the yields
of RCH3 increase by far higher (+38% and +60%) than the
upper limit yield of the blind reaction (17%). This
demonstrates impressively that the formation of copper-rich
complexes leads to a reduced reactivity.
Here we present an alternative explanation, which is, to our
knowledge, in agreement with all of the fundamental
statements of both sides and, we feel, the solution to this
Gordian knot. The key to different reactivities of cyano- and
iodocuprates is not mainly the structure of the cuprates before
the reaction but the fate of the alkyl copper side product of the
iodocuprates after the reaction. In the case of cyanocuprates
always heteroleptic cuprates are formed. However, for
iodocuprates the situation differs, depending on the solubility
of the alkyl copper. In the case of small substituents, e.g. methyl
copper, this side product precipitates and does not affect the
further reaction. As a result the reactivities of cyano- and
iodocuprates are very similar. However, in the case of large
substituents, e.g. R = Me3SiCH2 as in this study, the neutral
alkyl copper side product is incorporated into the remaining
iodocuprate and forms “extended” Gilman cuprates (only
detectable at temperatures far below those usually applied in
synthesis). These copper-rich complexes not only have reduced
polarities of their alkyl groups and therefore a reduced
reactivity, but in addition, consume overstoichiometrically the
higher reactive iodo Gilman cuprate. As a result both situations
are possible and true: very similar reactivities of cyano- and
iodocuprates and substantially reduced reactivities for iodocup-
rates. Key is the solubility and incorporation of the alkyl copper
in iodocuprates. This incorporation, which is detrimental to
In case of the cyanocuprate, R2CuLi·Li13CN, the increased
yields of RCH3 (by 2% and 18%) are in the range of the blind
5770
dx.doi.org/10.1021/ja501055c | J. Am. Chem. Soc. 2014, 136, 5765−5772

Journal of the American Chemical Society
Article
Accordingly, the integral of the side product R2 was correlated to this
sum, and the yields of R2 (%) were calculated.
reactivity, can be modulated by the size of the substituent (i.e.,
small substituent precipitation/full reactivity, large substituent
complete incorporation/reduced reactivity as shown in this
study and all intermediate levels for medium-sized substituents)
as well as most probably also by variations of solvent properties,
additives, or temperature. In addition, our reactivity and
structural study shows that, to a minor extent, also the well-
known aggregation and proposed mixed aggregate effects of the
cuprates overlay these reactivties, which may have led to further
confusion.
As a rule of thumb this study shows that, in reactions
requiring critical reactivities and large cuprate substituents,
cyanocuprates are expected to perform better than iodocup-
rates.
ASSOCIATED CONTENT
* Supporting Information
■
S
Assignments, diffusion coefficients, additional spectra, inter-
pretations and details. This material is available free of charge
via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
Ruth.Gschwind@ur.de
■
Present Address
⊥Dr. Maria Neumeier, Eicherstraße 2, 85123 Karlskron,
Germany.
EXPERIMENTAL SECTION
■
Funding
Sample Preparation. The Me3SiCH2Li solution (1.0 M in
pentane) was purchased from Sigma-Aldrich. Et2O-d10 was freshly
distilled over K/Na alloy. All manipulations during the synthesis were
done under exclusion of moisture and air. All cuprate samples were
prepared by a method described by John et al.10 The synthesis was
directly done in Et2O-d10 to exclude protonated Et2O. The pentane
from the Me3SiCH2Li solution was removed before the addition to
the Cu salt suspension. The conversions with methyl iodide were
performed at 170 K. The NMR titrations were performed at
0.2−0.06 mol/L, DOSY measurement at 0.07−0.18 mol/L, and
reactions with CH3I at 0.05−0.06 and 0.13 mol/L.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work is dedicated to Prof. B. H. Lipshutz. We gratefully
acknowledge financial support from the DFG grant GS 13/2-1,
allowing for these results.
REFERENCES
(1) Krause, N. Angew. Chem., Int. Ed. 1999, 38, 79.
(2) Gschwind, R. M. Chem. Rev. 2008, 108, 3029.
■
NMR Data Collection and Processing. The NMR spectra were
recorded on a Bruker Avance 600 MHz spectrometer equipped with a
5 mm broadband triple resonance Z-gradient probe. H,13C HMBC
1
(3) Olmstead, M. M.; Power, P. P. J. Am. Chem. Soc. 1990, 112, 8008.
(4) Olmstead, M. M.; Power, P. P. J. Am. Chem. Soc. 1989, 111, 4135.
(5) Bertz, S. H. J. Am. Chem. Soc. 1991, 113, 5470.
measurements were carried out with a standard Bruker pulse program
(hmbcetgpl2nd) using 64 scans, 16 dummy scans, TD(F2) = 4k and
TD(F1) = 400 with a relaxation delay of 2 s. All diffusion
measurements were performed with a convection-suppressing pulse
program58 in pseudo-2D mode and processed with Bruker software
package t1/t2. For each experiment, 2 dummy scans and 8 actual scans
were used, with a relaxation delay of 10 s and a diffusion delay of
50 ms. The shape of the gradients was sinusoidal, with a length of 4
ms, and the strength was varied in 24 increments (5−95%) of the
gradient ramp created by Bruker software DOSY. The temperature for
all measurements was controlled by a Bruker BVTE 3900 temperature
unit.
Determination of Diffusion Coefficients. The diffusion
coefficients were determined using the Bruker software package t1/
t2 to fit the measured decline of intensity. The diffusion coefficients
given in SI Table 2 and SI Table 4 are the average results of 1−6
experiments of each measurement with sample concentrations
between 0.07 and 0.18 mol/L. Due to the low concentration range,
the experimental diffusion coefficient show higher experimental errors
than usual for DOSY measurements: 9% (R2CuLi·LiI, 3 experiments,
1 × 0.07 mol/L, 2 × 0.13 mol/L), 9% (R3Cu2Li, 4 experiments, 2 ×
0.07 mol/L, 2 × 0.13 mol/L), 11% (R4Cu3Li, 4 experiments, 2 × 0.07
mol/L, 2 × 0.13 mol/L), 7% (RCu, 4 experiments, 2 × 0.07 mol/L, 2
× 0.13 mol/L), 5% (R2CuLi·Li13CN, 1 experiment, 0.18 mol/L), and
6% (RCuLi13CN, 3 experiments, 3 × 0.07 mol/L).
Internal Viscosity Reference. Since we noticed a strong
dependence of the viscosity on the concentration, the temperature,
and the kind of Cu(I) complexes, a trace (2−3 drops) of benzene
(C6H6) was used as an internal reference.59 By a comparison of the
diffusion coefficients of the reference measured in pure Et2O-d10 and in
the cuprate sample, the viscosity correction factors for the Cu(I)
complex samples were determined, in order to get comparable
experimental diffusion coefficients.
Integral Analysis of the Reactions between R2CuLi·LiX (X = I,
13CN) and MeI. For the determination of yields (%), the sum of
integrals of the CH3 groups, which correspond to unreacted MeI, the
main product RCH3, and the product of metal−halide exchange,
H3CLi, were correlated to the added amount of MeI and set to 100%.
(6) Mobley, T. A.; Muller, F.; Berger, S. J. Am. Chem. Soc. 1998, 120,
̈
1333.
̈
(7) Bertz, S. H.; Nilsson, K.; Davidsson, O.; Snyder, J. P. Angew.
Chem., Int. Ed. 1998, 37, 314.
(8) Kronenburg, C. M. P.; Jastrzebski, J. T. B. H.; Spek, A. L.; van
Koten, G. J. Am. Chem. Soc. 1998, 120, 9688.
(9) Boche, G.; Bosold, F.; Marsch, M.; Harms, K. Angew. Chem., Int.
Ed. 1998, 37, 1684.
(10) Gschwind, R. M.; Rajamohanan, P. R.; John, M.; Boche, G.
Organometallics 2000, 19, 2868.
(11) Gschwind, R. M.; Xie, X.; Rajamohanan, P. R.; Auel, C.; Boche,
G. J. Am. Chem. Soc. 2001, 123, 7299.
(12) Xie, X.; Auel, C.; Henze, W.; Gschwind, R. M. J. Am. Chem. Soc.
2003, 125, 1595.
(13) Henze, W.; Vyater, A.; Krause, N.; Gschwind, R. M. J. Am.
Chem. Soc. 2005, 127, 17335.
(14) Bertz, S. H.; Carlin, C. M.; Deadwyler, D. A.; Murphy, M. D.;
Ogle, C. A.; Seagle, P. J. Am. Chem. Soc. 2002, 124, 13650.
(15) Bertz, S. H.; Cope, S.; Dorton, D.; Murphy, M.; Ogle, C. A.
Angew. Chem., Int. Ed. 2007, 46, 7082.
(16) Bertz, S. H.; Cope, S.; Murphy, M.; Ogle, C. A.; Taylor, B. J. J.
Am. Chem. Soc. 2007, 129, 7208.
(17) Hu, H.; Snyder, J. P. J. Am. Chem. Soc. 2007, 129, 7210.
(18) Gartner, T.; Henze, W.; Gschwind, R. M. J. Am. Chem. Soc.
̈
2007, 129, 11362.
(19) Henze, W.; Gartner, T.; Gschwind, R. M. J. Am. Chem. Soc.
̈
2008, 130, 13718.
(20) Bartholomew, E. R.; Bertz, S. H.; Cope, S.; Dorton, D. C.;
Murphy, M.; Ogle, C. A. Chem. Commun. 2008, 1176.
(21) Bertz, S. H.; Murphy, M. D.; Ogle, C. A.; Thomas, A. A. Chem.
Commun. 2010, 46, 1255.
(22) Gartner, T.; Yoshikai, N.; Neumeier, M.; Nakamura, E.;
̈
Gschwind, R. M. Chem. Commun. 2010, 46, 4625.
(23) Bertz, S. H.; Miao, G.; Eriksson, M. M. Chem. Commun. 1996,
815.
5771
dx.doi.org/10.1021/ja501055c | J. Am. Chem. Soc. 2014, 136, 5765−5772

Journal of the American Chemical Society
Article
(24) Bertz, S. H.; Chopra, A.; Eriksson, M.; Ogle, C. A.; Seagle, P.
Chem.Eur. J. 1999, 5, 2680.
temperatures is not possible due to the close melting point of Et2O
(157 K) and relatively small line widths necessary for reliable spectra
integration. In addition, extensive reactivity investigations of Bertz and
Ogle showed that it is extremely difficult to obtain reliable and highly
reproducible kinetic data for substitution reactions of organocuprates.
Bertz, S. H.; Chopra, A.; Eriksson, M.; Ogle, C. A.; Seagle, P. Org.
Chem. Eur. J. 1999, 5, 2680. Bertz, S. H.; Human, J.; Ogle, C. A.;
Seagle, P. Org. Biomol. Chem. 2005, 3, 392. Murphy, M. D.; Ogle, C.
A.; Bertz, S. H. Chem. Commun. 2005, 854. Therefore, in this study
only yields are discussed.
(25) Alexakis, A.; Jachiet, D.; Normant, J. F. Tetrahedron 1986, 42,
5607.
(26) Nakamura, E.; Yamanaka, M.; Yoshikai, N.; Mori, S. Angew.
Chem., Int. Ed. 2001, 40, 1935.
(27) Kronenburg, C. M. P.; Jastrzebski, J. T. B. H.; van Koten, G.
Polyhedron 2000, 19, 553.
(28) Ashby, E. C.; Watkins, J. J. Chem. Commun. 1976, 784.
(29) Whitesides, G. M.; Fischer, W. F.; SanFilippo, J.; Bashe, R. W.;
House, H. O. J. Am. Chem. Soc. 1969, 91, 4871.
(30) House, H. O.; Respess, W. L.; Whitesides, G. M. J. Org. Chem.
1966, 31, 3128.
(57) Nakamura, E.; Yoshikai, N. Bull. Chem. Soc. Jpn. 2004, 77, 1.
(58) Jershow, A.; Mueller, N. J. Magn. Reson. 1997, 125, 372.
(59) Cabrita, E. J.; Berger, S. Magn. Reson. Chem. 2001, 39, S142.
(31) Guss, J. M.; Mason, R.; Thomas, K. M.; van Koten, G.; Noltes, J.
G. J. Organomet. Chem. 1972, 40, C79.
(32) Gambarotta, S.; Floriani, C.; Chiesi-Villa, A.; Guastini, C. J. C. S.
Chem. Commun. 1983, 1156.
(33) Meyer, E. M.; Gambarotta, S.; Floriani, C.; Chiesi-Villa, A.;
Guastini, C. Organometallics 1989, 8, 1067.
(34) Lappert, M. F.; Pearce, R. J. C. S. Chem. Commun. 1973, 24.
(35) Pearson, R. G.; Gergory, C. D. J. Am. Chem. Soc. 1976, 98, 4098.
(36) Lipshutz, B. H.; Wilhelm, R. S.; Floyd, D. M. J. Am. Chem. Soc.
1981, 103, 7672.
(37) Lipshutz, B. H.; Kozlowski, J. A.; Wilhelm, R. S. J. Am. Chem.
Soc. 1982, 104, 2305.
(38) Lipshutz, B. H.; Wilhelm, R. S.; Kozlowski, J. A. Tetrahedron
Lett. 1982, 23, 3755.
(39) Lipshutz, B. H.; Wilhelm, R. S.; Kozlowski, J. A. Tetrahedron
1984, 40, 5005.
(40) Ashby, E. C.; Watkins, J. J. J. Am. Chem. Soc. 1977, 99, 5312.
(41) Corey, E. J.; Posner, G. H. J. Am. Chem. Soc. 1968, 90, 5615.
(42) Lipshutz, B. H.; Wilhelm, R. S. J. Am. Chem. Soc. 1982, 104,
4696.
(43) Lipshutz, B. H.; Kozlowski, J. A.; Wilhelm, R. S. J. Org. Chem.
1983, 48, 546.
(44) Lipshutz, B. H.; Parker, D.; Kozlowski, J. A.; Miller, R. D. J. Org.
Chem. 1983, 48, 3334.
(45) House, H. O.; Koepsell, D. G.; Campbell, W. J. J. Org. Chem.
1972, 37, 1003.
(46) Bertz, S. H.; Human, J.; Ogle, C. A.; Seagle, P. Org. Biomol.
Chem. 2005, 3, 392.
(47) The reactvities of iodocuprates in 1,4-addition reactions to
cyclohexenone are too high to make slight reactivity reductions
observable. Bertz, S. H.; Chopra, A.; Eriksson, M.; Ogle, C. A.; Seagle,
P. Chem.Eur. J. 1999, 5, 2680. Therefore, cross-coupling reactions
in diethylether were chosen to create reduced, i.e., observable
differences in reactivity.
(48) Yoshikai, N.; Nakamura, E. Chem. Rev. 2012, 112, 2339.
(49) Deviating from Ashby’s results, the integrals of RLi and R2CuLi
show exactly the expected ratio, and no indication for a higher-order
cuprate is found.
(50) van Koten, G.; Noltes, J. G. J. Organomet. Chem. 1975, 102, 551.
(51) Krause, N.; Wagner, R.; Gerold, A. J. Am. Chem. Soc. 1994, 116,
381.
(52) Snyder, J. P.; Spangler, D. P.; Behling, J. R.; Rossiter, B. E. J. Org.
Chem. 1994, 59, 2665.
(53) Snyder, J. P. S.; Tipsword, G. E.; Splanger, D. J. J. Am. Chem.
Soc. 1992, 114, 1507.
(54) Yamanaka, M.; Inagaki, A.; Nakamura, E. J. Comput. Chem.
2003, 24, 1401.
(55) Bohme, M.; Frenking, G.; Reetz, M. T. Organometallics 1994,
̈
13, 4237.
(56) Without rapid injection NMR technique McGarrity, J. F.;
Prodolliet, J. J. Org. Chem. 1984, 49, 4465. McGarrity, J. F.; Ogle, C.
H.; Brich, Z.; Loosli, H. R. J. Am. Chem. Soc. 1985, 107, 1810. this is
the minimum time necessary for temperature equilibration, tuning and
shimming. 1H NMR investigations at a later point of time showed that
the reactions were already completed after that minimal 20 minutes
(see SI Figure 6). A further reduction of the reaction rate by lower
5772
dx.doi.org/10.1021/ja501055c | J. Am. Chem. Soc. 2014, 136, 5765−5772