Influence of tetrahydrofuran on reactivity, aggregation, and aggregate structure of dimethylcuprates in diethyl ether

Article abstract of DOI:10.1021/ja055085o

The comprehension of factors influencing the reactivity of organocuprates is still far from enabling a rational control of their reactions. Especially the degree of aggregation and structures of organocuprates are the focus of discussion about the factors affecting their reactivity. Therefore, this study combines kinetic measurements and NMR investigations to elucidate the influence of disaggregation via addition of tetrahydrofuran (THF) on the reactivity and aggregate structure of Gilman cuprates. As model systems, Me₂CuLi-Lil (1-Lil) and Me₂CuLi-LiCN (1-LiCN) in diethyl ether (DEE) were chosen; as model reaction, the 1,4-addition to 4,4-dimethylcyclohex-2-enone. The kinetic data show for 1-Lil a pronounced acceleration effect upon addition of distinct amounts of THF, whereas the reactivity of 1-LiCN continuously decreases with the addition of THF. Series of NMR diffusion measurements as well as ¹H-⁷Li heteronuclear Overhauser effect (HOE), and ¹H-¹H nuclear Overhauser effect (NOE) spectra show different structural influences of THF on 1-Lil and 1-LiCN. For 1-Lil, small salt units are separated from the cuprate aggregate by THF. In contrast to this, THF disaggregates the oligomeric structures of 1-LiCN, while the core structures remain intact with salt attached. Thus, the reactivity of 1-Lil seems to be fine-tuned through distinct amounts of salt or THF, whereas the decreasing reactivity of 1-LiCN correlates with the disaggregation of oligomers via THF. Thus, for synthetic chemists with reactivity problems in specific reactions of iododialkylcuprates, the addition of small amounts of THF might be useful to enhance the reactivity. In addition to these structure-reactivity studies, the CN^- group is shown to be directly attached to the cuprate moiety via a combination of ¹H-¹³C HOE- and ¹H-¹H NOEs. This represents the first direct experimental evidence in solution for the position of the CN^- group relative to the cuprate moiety in cyano-Gilman cuprates.

Full text of DOI:10.1021/ja055085o

Published on Web 11/12/2005
Influence of Tetrahydrofuran on Reactivity, Aggregation, and
Aggregate Structure of Dimethylcuprates in Diethyl Ether
Wolfram Henze,† Armand Vyater,‡ Norbert Krause,*,‡ and Ruth M. Gschwind*,§
Contribution from the Kekule´-Institut fu¨r Organische Chemie und Biochemie, Rheinische
Friedrichs-Wilhelms-UniVersita¨t Bonn, Gerhard-Domagk-Strasse 1, D-53121 Bonn, Germany,
Organic Chemistry II, Dortmund UniVersity, Otto-Hahn-Strasse 6, D-44221 Dortmund,
Germany, and the Institut fu¨r Organische Chemie, UniVersita¨t Regensburg, UniVersita¨tsstrasse
31, D-93053 Regensburg, Germany
Received July 27, 2005; E-mail: ruth.gschwind@chemie.uni-regensburg.de; norbert.krause@uni-dortmund.de
Abstract: The comprehension of factors influencing the reactivity of organocuprates is still far from enabling
a rational control of their reactions. Especially the degree of aggregation and structures of organocuprates
are the focus of discussion about the factors affecting their reactivity. Therefore, this study combines kinetic
measurements and NMR investigations to elucidate the influence of disaggregation via addition of
tetrahydrofuran (THF) on the reactivity and aggregate structure of Gilman cuprates. As model systems,
Me2CuLi‚LiI (1‚LiI) and Me2CuLi‚LiCN (1‚LiCN) in diethyl ether (DEE) were chosen; as model reaction, the
1,4-addition to 4,4-dimethylcyclohex-2-enone. The kinetic data show for 1‚LiI a pronounced acceleration
effect upon addition of distinct amounts of THF, whereas the reactivity of 1‚LiCN continuously decreases
with the addition of THF. Series of NMR diffusion measurements as well as 1H-7Li heteronuclear Overhauser
1
effect (HOE), and H-1H nuclear Overhauser effect (NOE) spectra show different structural influences of
THF on 1‚LiI and 1‚LiCN. For 1‚LiI, small salt units are separated from the cuprate aggregate by THF. In
contrast to this, THF disaggregates the oligomeric structures of 1‚LiCN, while the core structures remain
intact with salt attached. Thus, the reactivity of 1‚LiI seems to be fine-tuned through distinct amounts of
salt or THF, whereas the decreasing reactivity of 1‚LiCN correlates with the disaggregation of oligomers
via THF. Thus, for synthetic chemists with reactivity problems in specific reactions of iododialkylcuprates,
the addition of small amounts of THF might be useful to enhance the reactivity. In addition to these structure-
reactivity studies, the CN- group is shown to be directly attached to the cuprate moiety via a combination
1
1
of H-13C HOE- and H-1H NOEs. This represents the first direct experimental evidence in solution for
the position of the CN- group relative to the cuprate moiety in cyano-Gilman cuprates.
finalized.1,4,5 Then different degrees of aggregation depending
on the kind of solvent and salt6,7 were observed and the core
structure of the aggregates was identified to be a homodimer
1. Introduction
The systematic control of chemical reactions via the com-
prehension of structure-reactivity and structure-selectivity
correlations is still a long-term objective for many chemical
reactions. Especially for organometallic reagents forming com-
plex aggregates as reactive species, the comprehension of the
factors influencing their reactivity and selectivity is still in its
infancy. A famous example is the group of lithium organocu-
prate reagents. There, despite their wide application in carbon-
carbon bond-forming reactions and intensive theoretical and
structural investigations, the structures of the reactive species
are not yet fully determined.1-3 During the first long-standing
scientific discussion about structure-reactivity correlations of
organocuprates, different reactivities of cyano- and iodocuprates
were explained by different monomeric structures for a long
time. However, in the end the “lower order” alkylcuprates were
in DEE.8 This put the aggregation, previously indicated by 15
N
NMR,9 mass spectroscopy studies,10 crystal structures,6,11 col-
ligative measurements,12-14 and extended X-ray absorption fine
structure (EXAFS) studies15 into the focus of interest as a
possible explanation for different reactivities16 despite similar
(4) Bertz, S. H. J. Am. Chem. Soc. 1990, 112, 4031-4032.
(5) Bertz, S. H. J. Am. Chem. Soc. 1991, 113, 5470-5471.
(6) John, M.; Auel, C.; Behrens, C.; Marsch, M.; Harms, K.; Bosold, F.;
Gschwind, R. M.; Rajamohanan, P. R.; Boche, G. Chem. Eur. J. 2000, 6,
3060-3068.
(7) Gschwind, R. M.; Rajamohanan, P. R.; John, M.; Boche, G. Organome-
tallics 2000, 19, 2868-2873.
(8) Gschwind, R. M.; Xie, X.; Rajamohanan, P. R.; Auel, C.; Boche, G. J.
Am. Chem. Soc. 2001, 123, 7299-7304.
(9) Bertz, S. H.; Nilsson, K.; Davidson, O¨ .; Snyder, J. P. Angew. Chem., Int.
Ed. 1998, 37, 314-317.
(10) Lipshutz, B. H.; Keith, J.; Buzard, D. J. Organometallics 1999, 18, 1571-
1574.
† Rheinische Friedrichs-Wilhelms-Universita¨t Bonn.
‡ Dortmund University.
(11) Olmstead, M. M.; Power, P. P. Organometallics 1990, 9, 1720-1722.
(12) Pearson, R. G.; Gergory, C. D. J. Am. Chem. Soc. 1976, 98, 4098-4104.
(13) Ashby, E. C.; Watkins, J. J. J. Am. Chem. Soc. 1977, 99, 5312-5317.
(14) Gerold, A.; Jastrzebski, J. T. B. H.; Kronenburg, C. M. P.; Krause, N.; van
Koten, G. Angew. Chem., Int. Ed. Engl. 1997, 36, 755-757.
(15) Huang, H.; Liang, C. H.; Penner-Hahn, J. E. Angew. Chem., Int. Ed. 1998,
37, 1564-1566.
§ Universita¨t Regensburg.
(1) Krause, N. Angew. Chem., Int. Ed. 1999, 38, 79-81.
(2) Nakamura, E.; Mori, S. Angew. Chem., Int. Ed. 2000, 39, 3750-3771.
(3) Krause, N., Ed. Modern Organocopper Chemistry; Wiley-VCH: Wein-
heim, Germany, 2002.
9
10.1021/ja055085o CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 17335-17342
17335

A R T I C L E S
Henze et al.
core structures. A first aggregation reactivity study comparing
very different degrees of aggregation in DEE and THF identified
contact ion pairs as reactive species.6 Then we found different
degrees of aggregation for the contact ion pairs in DEE,
standard conditions for synthetic procedures. The results show
that addition of THF has a different influence not only on
reactivity and disaggregation but also on the aggregate structure
of 1‚LiCN and 1‚LiI. Thus, for the first time different reactivities
of Gilman cuprates can be directly correlated to different
aggregate structures. Additionally, a direct dipolar interaction
with the CN- moiety in aggregates of 1‚LiCN is presented. In
principle, with NMR spectroscopy the structures of transition
states cannot be detected, but the investigation of the real
existing structures of reagents and intermediates is an essential
prerequisite to choose the ideal starting structures for theoretical
calculations.
depending on steric hindrance and the kind of salts.17
A
comparison with kinetic data derived by Bertz et al.16 suggested
an increase of reactivity through disaggregation by THF down
to the dimeric contact ion pair. However, such more refined
aggregation reactivity correlations lack direct comparability of
structural and kinetic data, which up to now have always been
performed under different experimental conditions. Additionally,
the influence of small amounts of THF on the aggregation and
the aggregation structure of organocuprates is unknown.3,7,17,18
As a second important point concerning organocuprates, the
core structure of the salt-containing dimethylcuprates in the
reactive species again came up as a subject in the discussion
during recent years. In the known X-ray structures of lithium
dialkylcuprates crystallized from salt-containing solutions6,11 and
in our NMR investigations of the pure reagents in solution,8
only homodimeric structures were found. Also, a recent study
about the control of electron transfer versus alkylation pathway
in the reaction of the Gilman reagent Me2CuLi‚LiI (1‚LiI) shows
that the lithium halide is not differentially involved in the
competition between these two reaction types.19 In contrast,
numerous theoretical calculations proposed heterodimeric core
structures20-22 of the Gilman reagents 1‚LiX (X ) I, CN) and
a recent theoretical publication suggested besides the involve-
ment of heterodimeric structures even the contribution of “higher
order” Lipshutz cuprates in the transition state of the addition
reaction of 1‚LiCN to acetylene.23 The main problem in the
discussion mentioned above is the lack of experimental evidence
about the position of the salt unit (LiI or LiCN) relative to the
organocuprate moiety in solution. The only information experi-
mentally available in solution is the coordination of the salt by
Li+, confirmed by previous 15N NMR9 and IR spectroscopic24
studies. The crystal structures,25-27 highly valuable in other
respects, showing different positioning possibilities of the salt
in organocuprates, are less meaningful for the aggregate structure
in solution, since significant differences of the aggregate
structures in the solid state and in solution were found.6,17
In this study experimental results for both subjects are
presented, concerning the correlation between disaggregation
via THF and reactivity of Gilman cuprates as well as regarding
the position of the salt units in solution. Thus, for the first time
kinetic studies and NMR structural investigations were per-
formed under identical experimental conditions, which are
2. Results and Discussion
2.1. Kinetic Data. The kinetic studies were carried out as
described previously.28 As model system we selected the 1,4-
addition of the Gilman cuprates 1‚LiX (X ) I, CN) to 4,4-
dimethylcyclohex-2-enone (2) in DEE at 210 K (Scheme 1).
The reaction mixture contained equimolar concentrations of the
reactants in the range of 0.025-0.1 M, as well as variable
amounts of THF and tetradecane as internal standard. The
progress of the reaction was determined by removing aliquots
at specific time intervals, which were hydrolyzed and analyzed
by gas chromatography. The substrate concentration showed an
exponential decrease against the reaction time (see Figure S1
in Supporting Information for an example), and the measured
values were fitted by use of the exponential function c ) c0
exp(-kt).28 The resulting first-order rate constants k typically
showed errors within 10% of the value of k.
Scheme 1. 1,4-Addition of the Gilman Cuprates 1‚LiX (X ) I, CN)
to 4,4-Dimethylcyclohex-2-enone (2) in DEE at 210 K Used as
Representative Model Reaction in the Kinetic Studies
In accordance with the “traditional” solvent effects in the
conjugate cuprate addition,6 we observed a continuous decrease
of the reaction rate for the 1,4-addition of the cyano-Gilman
cuprate 1‚LiCN to enone 2,1 as shown exemplarily in Figure 1.
This behavior proved to be independent of the concentration of
the reactants within the range examined here (see above and
Figure S2 in Supporting Information). In striking contrast to
this, the corresponding 1,4-addition of the copper iodide-derived
Gilman cuprate 1‚LiI to the standard enone 2 showed a
pronounced acceleration effect. At 0.1 M concentration of the
reactants, addition of just 0.25 equiv of THF caused a 2.5-fold
rate acceleration, and even in the presence of 1.0 equiv of THF
the reaction was slightly faster than in pure DEE (Figure 2).
This effect is in agreement with the observations of Bertz et
al.;16 however, they could not quantify the acceleration. We also
observed a strong dependence of the acceleration effect on the
concentration of the reactants: at 0.05 M, only a slight rate
increase (ca. 1.5-fold) was observed with the maximum at 1.0
(16) Bertz, S. H.; Chopra, A.; Eriksson, M.; Ogle, C. A.; Seagle, P. Chem. Eur.
J. 1999, 5, 2680-2691.
(17) Xie, X.; Auel, C.; Henze, W.; Gschwind, R. M. J. Am. Chem. Soc. 2003,
125, 1595-1601.
(18) Mori, S.; Nakamura, E. Chem. Eur. J. 1999, 5, 1534-1543.
(19) Yang, J. K.; Cauble, D. F.; Berro, A. J.; Bauld, N. L.; Krische, M. J. J.
Org. Chem. 2004, 69, 7979-7984.
(20) Yamanaka, M.; Nakamura, E. J. Am. Chem. Soc. 2005, 127, 4697-4706.
(21) Mori, S.; Nakamura, E.; Morokuma, K. Organometallics 2004, 23, 1081-
1088.
(22) Yamanaka, M.; Nakamura, E. Organometallics 2001, 20, 5675-5681.
(23) Nakamura, E.; Yoshikai, N. Bull. Chem. Soc. Jpn. 2004, 77, 1-12.
(24) Huang, H.; Alveraz, K.; Liu, Q.; Barnhart, T. M.; Snyder, J. P.; Penner-
Hahn, J. E. J. Am. Chem. Soc. 1996, 118, 8808-8816 and 12252
(correction).
(25) Kronenburg, C. M. P.; Jastrzebski, J. T. B. H.; Boersma, J.; Lutz, M.; Spek,
A. L.; van Koten, G. J. Am. Chem. Soc. 2002, 124, 11675-11683.
(26) Boche, G.; Bosold, F.; Marsch, M.; Harms, K. Angew. Chem., Int. Ed.
1998, 37, 1684-1686.
(27) Kronenburg, C. M. P.; Jastrzebski, J. T. B. H.; Spek, A. L.; van Koten, G.
J. Am. Chem. Soc. 1998, 120, 9688-9689.
(28) Canisius, J.; Gerold, A.; Krause, N. Angew. Chem., Int. Ed. 1999, 38, 1644-
1646.
9
17336 J. AM. CHEM. SOC. VOL. 127, NO. 49, 2005

THF Influence on Dimethylcuprates in Diethyl Ether
A R T I C L E S
one of the conditions used in the kinetic experiments (0.1 M,
210 K), an identical series of diffusion coefficients was
determined at 239 K, the temperature used in previous structural
studies.8,17 Diffusion studies at concentrations lower than 0.1
M could not be performed because of the too low signal-to-
noise ratio in the convection compensated diffusion experiments.
From the diffusion coefficients obtained and the aggregation
models shown in Figure 3, the degree of aggregation of the
investigated dimethylcuprates was calculated as described
previously.17 In the following, the homodimeric core structure
of 1‚LiX with no salt attached is denominated as homodimer
and the salt-containing aggregate is denoted as dimer with salt.
Figure 1. Rate constants of the 1,4-addition of 1‚LiCN (0.025 M) to 2
(0.025 M) at 210 K in pure DEE and in solvent mixtures of DEE and
different equivalents of THF.
In Figure 4a the results of the diffusion measurements for
1‚LiI are presented. Lower diffusion coefficients of 1‚LiI and
of THF at 210 K compared to 239 K indicate higher aggregates
at lower temperatures. Furthermore, at both temperatures, the
diffusion coefficients of THF (DTHF ) 1.15-1.45 × 10-9 m2/
s) are remarkably reduced in comparison to that of THF in pure
DEE (DTHF ) 2.16 × 109 m2/s) but much higher than those of
1‚LiI. This indicates that THF is partially attached to the
organocuprate aggregates. The diffusion coefficient (D ) 0.74
× 109 m2/ s) of 1‚LiI in pure DEE at 210 K correlates to 1.3
times a homodimer or a dimer with one LiI attached. Surpris-
ingly, the addition of up to 4 equiv of THF does not lead to an
increase of the diffusion coefficients beyond the error range.
This indicates that the average hydrodynamic radius of the
organocuprate aggregates is not significantly affected by these
amounts of THF in the case of 1‚LiI. However, small changes
in the aggregation structure cannot be detected by these PFG
NMR diffusion experiments, since the scattering of the diffusion
coefficient values due to the low concentration induces an error
of (5%.
The corresponding results for 1‚LiCN are given in Figure
4b. Similar diffusion coefficients of the cuprate at 210 and 239
K show a lower temperature dependence of aggregation for 1‚
LiCN than 1‚LiI. Again the diffusion coefficients of THF [DTHF
) (0.89-1.52) × 10-9 m2/s] are between the value of free THF
and those of the cuprate, indicating a partial binding of THF to
the organocuprate aggregates. As expected from previous
studies,17 the diffusion coefficient (D ) 0.52 × 109 m2/s) of
1‚LiCN in pure DEE indicates higher oligomeric aggregates of
2-2.5 dimers with salt at 210 K. In contrast to 1‚LiI, a clear
disaggregation is observed for 1‚LiCN with increasing amounts
of THF indicated by rising diffusion coefficients. By addition
of 4 equiv of THF, a disaggregation to a dimer with one salt
attached is calculated for the diffusion coefficient D ) 0.8 ×
10-9 m2/s at 210 K. For the diffusion coefficients of THF, a
similar trend with rising values is observed upon addition of
different equivalents of THF. This indicates that the averaged
Figure 2. Rate constants of the 1,4-addition of 1‚LiI (0.2 M) to 2 (0.1 M)
at 210 K in pure DEE and in solvent mixtures of DEE and different
equivalents of THF.
equiv of added THF. However, it is noticeable that the reaction
rate of 1‚LiCN without any THF is remarkably higher than for
the corresponding 1‚LiI, which is contrary to the observations
of Bertz et al.16
From these kinetic data a recommendation for syntheses with
dialkylcuprates can be derived. For syntheses with cyanocuprates
it is not advisable to add THF. However, if the reactivity of
iodocuprates is too low in a specific reaction, the addition of
up to 1 equiv of THF or the use of cyanocuprates is worthwhile
trying.
2.2. Diffusion Studies. To investigate whether the described
decrease of the reaction rate for 1‚LiCN or the acceleration effect
for 1‚LiI upon addition of THF is correlated with a change in
the degree of aggregation of these organocuprates in solution,
pulsed field gradient (PFG) NMR diffusion experiments were
performed. In DEE as solvent, series of diffusion coefficients
of 1‚LiI, 1‚LiCN, and THF were measured with increasing
amounts of THF up to 4 equiv (see Figure 4). In addition to
Figure 3. Structure models of salt-containing cuprate aggregates containing homodimeric core structures and salt units connecting the core structures with
(a) four and (b) six solvent molecules per homodimer.
9
J. AM. CHEM. SOC. VOL. 127, NO. 49, 2005 17337

A R T I C L E S
Henze et al.
Figure 4. Diffusion coefficients of 1‚LiI and THF (a) and of 1‚LiCN and THF (b) in pure DEE and solvent mixtures of DEE with different equivalents of
THF at 239 and 210 K.
hydrodynamic radii of the complexes containing THF decrease
with increasing amounts of THF.
aggregates. The bar charts of the cross-peak volume integrals
between the methyl group of the cuprates and the two different
CH2 groups of THF are shown in Figure 6. In the respective
1H-1H NOESY spectra the volume integrals were calibrated
in relation to the cross-peak volume integral between the two
chemically nonequivalent CH2 groups of THF. Therefore, the
volume integrals shown in the bar charts of Figure 6 are directly
related to the actual amount of THF.
These studies for the first time allow us to directly compare
the reactivity and aggregation of organocuprates at identical
experimental conditions. In the case of 1‚LiCN, the decline in
the reaction rates with increasing amounts of THF seems to be
directly correlated to a disaggregation of the oligomeric
structures of 1‚LiCN. However, in the case of 1‚LiI the
surprising maxima in reactivity upon addition of 0.25-1 equiv
of THF are not reflected in the diffusion coefficients of this
system.
1
In the case of 1‚LiI, the relative H-1H NOE cross-peak
volume integrals between the cuprate and THF decrease upon
the addition of THF. Based on our calibration relative to the
internal THF signal, this indicates a reduced fraction of THF
being bound to the cuprate aggregates with increasing equiva-
lents of THF. Further to this, change of the 1H-1H NOE volume
integrals is confirmation of the structural modification caused
2.3. 1H-7Li HOE and 1H-1H NOE Studies with Different
Equivalents of THF. Since the observed enhancements in the
reaction rates of 1‚LiI are not caused by an obvious change of
aggregation, it was investigated whether small structural changes
induced by the addition of THF can be detected. Therefore, 1H-
7Li heteronuclear Overhauser effect spectroscopy (HOESY) and
1H-1H nuclear Overhauser effect spectroscopy (NOESY)
experiments of 1‚LiX (X ) I, CN) in mixtures of DEE with 1,
2, and 4 equiv of THF were recorded at 0.1 M concentration
and 210 K. In Figure 5, exemplary 1H-7Li HOESY spectra of
1‚LiI (a) and 1‚LiCN (c) in a mixture of DEE with 4 equiv of
THF are shown. Additionally, bar charts of the HOE cross-
peak volume integrals between the lithium signal and the protons
of DEE (CH2 DEE/CH3 DEE) and THF (CH2 THF/CH2O THF)
of 1‚LiI (b) and 1‚LiCN (d) at different equivalents of THF are
given in Figure 5. There, the volume integrals are displayed
relative to the volume integrals of the cross-peak between
lithium and the methyl groups of the cuprates in each spectrum.29
In the case of 1‚LiI, the volume integral of the cross-peaks
of THF increases in a superproportional manner upon the
addition of THF, while the DEE cross-signal volume integrals
remain constant or increase only slightly. This trend indicates
that DEE is not replaced by THF, but additional coordination
sites are generated by increasing the amount of THF.
1
by the addition of THF, indicated previously by the H-7Li
HOE results. In contrast to that, the 1H-1H NOE volume
integrals of 1‚LiCN show similar integrals at increasing amounts
of THF. This mirrors a similar fraction of THF being bound to
the cuprate at different equivalents of THF. Again, the constancy
1
of the H-1H NOE volume integrals corroborates the results
1
from the H-7Li HOE measurements. It confirms that, in the
case of 1‚LiCN, the core structure of the aggregates is not
affected by increasing amounts of THF.
From these results we can build up a new model of salt-
containing organocuprates 1‚LiX (X ) I, CN) in mixtures of
DEE and THF (Figure 7). The kind of salt seems to influence
the disaggregation behavior induced by the addition of THF.
In the case of 1‚LiI, the diffusion studies show no obvious
1
disaggregation by THF. However, the more sensitive H-7Li
1
HOE and H-1H NOE studies show a structural change upon
the addition of THF. The coordination sites on lithium are
increased and simultaneously THF is not only bound to the
organocuprate. These facts suggest a separation of small salt
units from the homodimeric core structure not observed by the
less sensitive diffusion measurements. This means, at larger
amounts of THF, the THF is bound not only to lithium inside
the homodimeric cuprate but also to additional lithium species
separated from the cuprate. Previously published structural
suggestions for these additional lithium species are solvent-
separated ion pairs,6 Li-I-Li+ moieties,25,30 and dimeric (LiI)2
clusters.30 Additionally, this model complies with the observa-
tion that LiI itself is very well soluble in DEE and THF. Since
there was recently a revival of heterodimeric structures in the
In the case of 1‚LiCN, the volume integral of the cross-peaks
of THF increases proportionally to the added equivalents of
THF, while the cross-peak volume integral of DEE dramatically
decreases. This indicates a replacement of DEE by the addition
of THF at a similar number of coordination sites.
1
To prove these hypotheses, H-1H NOESY spectra were
recorded to investigate the closeness of THF to the cuprate
(29) The absolute integral of the reference cross-peak remains constant for 1.-
LiI at different amounts of THF within experimental error. In the spectra
of 1.LiCN the absolute volume integral of the reference cross-peak increases
according to the general signal-to-noise improvement of all signals in the
spectra of 1.LiCN caused by disaggregation.
(30) Bertz, S. H.; Vellekoop, A. S.; Smith, R. A. J.; Snyder, J. P. Organometallics
1995, 14, 1213-1220.
9
17338 J. AM. CHEM. SOC. VOL. 127, NO. 49, 2005

THF Influence on Dimethylcuprates in Diethyl Ether
A R T I C L E S
1
Figure 5. Representative for a series of spectra in different solvent mixtures, the H-7Li HOESY spectra of 1‚LiI (a) and 1‚LiCN (c) are shown in DEE
1
with 4 equiv of THF at 210 K. For mixtures of DEE and 1, 2, and 4 equiv of THF, the volume integrals of the H-7Li HOE cross-peaks between lithium
and the protons of DEE and THF are summarized as bar charts for 1‚LiI (b) and 1‚LiCN (d).
Figure 6. Bar charts summarizing the 1H-1H NOE volume integrals of the cross-peaks between the methyl group of the cuprate and the two different CH2
groups of THF for 1‚LiI (a) and 1‚LiCN (b) in mixtures of DEE and 1, 2, and 4 equiv of THF.
discussion of the intermediate structures of organocuprates upon
addition to enones,23,31,32 we also tried to interpret our data in
terms of heterodimeric structures. However, considering het-
erodimeric oligomers, a separation of Li moieties would lead
to a severe reduction of the 1H-7Li HOE cross-peaks between
the Li moieties and the methyl groups of the cuprates as well
as to a significant reduction of the hydrodynamic radius, that
is, a considerable increase in the diffusion coefficient. Both
factors are not supported by the experimental data.
(31) Murphy, M. D.; Ogle, C. A.; Bertz, S. H. Chem. Commun. 2005, 854-
856.
(32) Bertz, S. H.; Carlin, C. M.; Deadwyler, D. A.; Murphy, M. D.; Ogle, C.
A.; Seagle, P. J. Am. Chem. Soc. 2002, 124, 13650-13651.
In the case of 1‚LiCN, a clear disaggregation of the oligomeric
chains in combined homodimer/salt aggregates can be observed
9
J. AM. CHEM. SOC. VOL. 127, NO. 49, 2005 17339

A R T I C L E S
Henze et al.
Figure 7. Postulated aggregate structures of 1‚LiI and 1‚LiCN at 0.1 M concentration in pure DEE and their change through addition of 4 equiv of THF
derived from combined results of diffusion, NOE, and HOE studies.
by the addition of THF, but this does not significantly induce
either additional coordination sites or a structural change in the
aggregation structure (see Figure 7). The continuity of the
combined organocuprates’ salt aggregation structure by disag-
gregation is also congruent with the fact that LiCN is poorly
soluble in DEE or/and THF.33
the position and role of CN- in the structures and reactive
intermediates of lithium diorganocuprates in solution. In contrast
to the large number of theoretical studies regarding this
subject,21,23,34 from experimental studies up to now only indirect
evidence was given for the position of CN- in the different
cuprate species in solution (for details see Introduction). With
a special sample preparation and the utilization of 13C-labeled
LiCN salt, it was now possible for the first time to detect a
direct 1H-13C HOE to the CN- moiety in organocuprate
aggregates.
Now the kinetic data can be compared with the NMR studies
again. In the case of 1‚LiI, a distinct amount of salt and/or THF
attached to the homodimeric core structure seems to enhance
the reactivity. The reason for this could be that an optimal tuning
of the Lewis acidity of the solvated Li+ cations is reached. For
1‚LiCN the higher Lewis basicity of THF, compared to DEE,
induces a disaggregation of the oligomeric species and a
decrease of the Lewis acidity of the Li+ cations, and both effects
are correlated with a significant decrease in the reactivity.
Interestingly, the combined HOE/NOE studies show that for
1‚LiCN the disaggregation is not accompanied by a separation
of the salt unit from the core structure, but the combined cuprate
and salt unit persists. Thus, the presented kinetic and NMR
spectroscopic data point out that for cyanocuprates the short
oligomeric structures and the Lewis acidity of the lithium
cations, induced by the oligomeric structure, affect the reactivity
more than the presence of the cyanides alone. An estimation of
the contribution of each factor to the reactivity of the cuprate
may be possible with the aid of high-level quantum chemical
calculations of the complete oligomeric cuprate clusters.
Due to the extremely low signal-to-noise ratio of the 13C
signal of CN- in samples of 1‚Li13CN with a 0.1 M concentra-
tion, a sample of 1‚Li13CN in DEE with a 3-fold higher
concentration (0.3 M) was prepared. Thereby 12 equiv of THF
was used to adjust a degree of aggregation comparable to the
studies discussed above. The diffusion coefficient (D ) 0.83
× 10-9 m2/s) of the resulting cuprate aggregate at 239 K
indicates an aggregation equivalent to a dimer with one salt
attached. Under these experimental conditions the usually fast
exchange of solvent molecules between a position bound to the
cuprate aggregate and the solvent bulk is slowed on the NMR
time scale. Thus, for the first time we observed the THF signals
being split into two sets of signals, one set with small line widths
for the bulk THF molecules (THF) and one set with broad line
widths for the THF molecules attached to the organocuprate
aggregates (THF*) (Figure S3 in Supporting Information). Via
the cross-peaks between the methyl group of the cuprate and
1
1
2.4. H-13C HOE and H-1H NOE Studies Concerning
the Position of CN-. Our results again raise the central question
in the long-standing discussion about the special reactivity of
LiCN-containing diorganocuprates, namely, the question about
1
the two different CH2 groups of THF* in the H-1H NOESY
spectrum, THF* can be unambiguously assigned to be bound
to the cuprate aggregate (Figure S4 in Supporting Information).
(33) LiCN synthesized by method 1: Markley, T. J.; Toby, B. H.; Pearlstein,
(34) Yamanaka, M.; Inagaki, A.; Nakamura, E. J. Comput. Chem. 2002, 24,
1401-1409.
R. M.; Ramprasad, D. Inorg. Chem. 1997, 36, 3376-3378.
9
17340 J. AM. CHEM. SOC. VOL. 127, NO. 49, 2005

THF Influence on Dimethylcuprates in Diethyl Ether
A R T I C L E S
Figure 8. (a) 1H-13C HOESY spectrum of 1‚LiCN (0.3 M) with 12 equiv of THF at 239 K. Two sets of signals are observed for THF in the solvent bulk
(THF) and THF bound to the cuprate aggregate (THF*). (b) Model structure for 1‚LiCN derived from combined results of diffusion, NOE, and HOE studies.
1
In the respective H-13C HOESY spectrum (Figure 8a), for
the first time cross-peaks to the 13CN- group are observed,
namely, cross-peaks between the 13CN- group and both THF*
signals in considerable signal-to-noise ratios. A cross-peak
between the methyl group of 1‚Li13CN and the 13CN- moiety
could not be detected despite an extensive experimental effort
of varying the parameters in a number of test spectra.
structure are found for 1‚LiI and 1‚LiCN. For 1‚LiI, the addition
of 0.25-1.0 equiv of THF led to a pronounced acceleration
effect. In contrast, the reaction rate of 1‚LiCN, being remarkably
higher in pure DEE, significantly decreases upon addition of
THF. The parallel NMR investigations show that, in 1‚LiI
starting from a dimer with salt attached, small salt units are
separated through the addition of THF. Thus, the reactivity of
1‚LiI seems to show a fine-tuning through distinct amounts of
salt or THF influencing the Lewis acidity of the Li+ cations. In
striking contrast, 1‚LiCN starting from a short oligomeric
structure is disaggregated to dimers through THF without
structural change in its core structure; that is, the LiCN remains
attached. This is additionally confirmed via 1H-13C HOE
contacts to the CN- group, the first direct experimental evidence
for the position of the CN- group in small aggregates of
dialkylcuprates in solution. Thus, the decrease of reactivity of
1‚LiCN shows a direct correlation with the disaggregation of
its oligomeric structure, indicating simultaneously a decreasing
Lewis acidity of the Li+ cations.
In Figure 8b a structure model is depicted based on the
combined results of the diffusion measurements and H-13C
1
HOESY and 1H-H NOESY spectra. The diffusion coefficient
indicates a dimer with one salt attached. The H-1H NOESY
1
spectrum shows a direct contact between the cuprate and THF*.
The strong contact in the H-13C HOESY spectrum between
1
THF* and the CN- in conjunction with the absence of a cross-
peak between the CN- group and the methyl group of the
cuprate suggest the shown orientation of the CN-. Although
this agrees with models from infrared spectroscopy,24 the shown
orientation cannot be postulated as fixed due to the relatively
small isomerization barrier in LiCN35 and known crystal
structures.36 Thus, for the first time it is experimentally shown
that in 1‚LiCN the CN- group is attached to the cuprate
aggregate even at a low degree of aggregation. The not
In the long-standing scientific discussion about the special
reactivity of LiCN-containing organocuprates, the role of LiCN
for the reactivity was controversially discussed. Recently
Nakamura and Yoshikai23 again proposed a special contribution
of LiCN to the transition state. Since our NMR studies show
that LiCN remains directly attached to the homodimeric cuprates
through all kinds of aggregation, a contribution of the LiCN
cannot be excluded. However, our work shows that disaggre-
gation of short oligomers significantly decreases the reactivity
of cyanocuprates despite the permanent presence of cyanides
in the structures. This indicates that the oligomeric structure of
cyano dialkylcuprates and/or the Lewis acidity of Li+ gained
by this oligomeric structure influence the reactivity of LiCN-
containing cuprates rather than the presence of single LiCN units
alone.
detectable H-13C HOE contact between the CN- group and
1
the methyl groups suggests that the carbon of the CN- group
is aligned away from the cuprate units.
3. Conclusion
Kinetic measurements and NMR investigations performed
under identical experimental conditions, which are standard
conditions of synthetic protocols, allow us for the first time to
correlate directly the structure of LiI- or LiCN-containing
Gilman cuprates in DEE with their reactivity in the 1,4-addition
to an enone. To confirm or to invalidate previous hypotheses
about aggregation/reactivity correlations in these reactions, the
influence of different equivalents of THF on the reactivity and
aggregation structure was investigated. Surprisingly, different
influences of THF on the reactivity as well as on the aggregate
For synthetic chemists with reactivity problems in specific
reactions of iodo- and cyanodialkylcuprates, a recommendation
can be derived from the presented data. Whereas for cyanocu-
prates the addition of THF is unfavorable, in the case of
iodocuprates it is worthwhile to try the addition of small amounts
of THF to enhance the reactivity.
(35) Arranz, F. J.; Benito, R. M.; Borondo, F. J. Chem. Phys. 2004, 120, 6516-
6523 and references therein.
(36) Markley, T. J.; Toby, B. H.; Pearlstein, R. M.; Ramprasad, D. Inorg. Chem.
1997, 36, 3376-3378.
9
J. AM. CHEM. SOC. VOL. 127, NO. 49, 2005 17341

A R T I C L E S
Henze et al.
a convection-suppressing pulse program44 in pseudo-2D mode and
processed with Bruker software package t1/t2. For each experiment,
16 dummy scans and 16 actual scans were used, with a relaxation delay
of 3 s and a diffusion delay of 30 ms. The shape of the gradients was
sinusoidal, with a length of 2 ms, and the strength was varied in 10
increments (5-95%) of the gradient ramp created by Bruker software
DOSY. The diffusion coefficients were determined by using the Bruker
software package t1/t2 to fit the measured decline of intensity. The
obtained diffusion coefficients are the average results of 2-3 repetitions
of each measurement.
4. Experimental Section
Sample Preparation. All samples were prepared by a method
described by John et al.6 The ratio of DEE-d10/DEE, the equivalents of
THF, and the concentration of the samples were adjusted by comparing
the integrals of the proton spectra obtained by a single 90° pulse as
well as the values gained from the synthesis. The mixture of DEE-
d10/DEE is denoted as pure DEE solvent throughout this paper.
Internal Viscosity Reference. The viscosity value of nondeuterated
DEE at 239 and 210 K was obtained by nonlinear fitting to literature
data in the temperature range of 173-298 K.37 The viscosity value of
the mixed solvent (DEE-d10/DEE) used in this study was defined to be
1.1 times of the value of the nondeuterated DEE at the same
temperature.38,39
Since we noticed a strong dependence of the viscosity on the
concentration, the temperature, and the kind of cuprates, a trace (2-3
drops) of benzene (C6H6) was used as an internal reference.40 To check
influences of the added amounts of THF on the viscosity, diffusion
measurements of a sample with the internal viscosity reference and
THF (0.2 M) in DEE-d10 were performed. The respective diffusion
coefficients showed no influence of these amounts of THF on the
viscosity. By a comparison of the diffusion coefficients of the reference
measured in pure DEE-d10 and in the cuprate sample, the viscosity
correction factors for the cuprate samples were determined.
Theoretical Diffusion Coefficients. As described in our previous
publication,17 the theoretical diffusion coefficients were calculated from
hard-sphere increments41 and the Stokes-Einstein theory42 modified
by a shape factor.43
1
1
For the H-7Li HOESY and the H-13C HOESY measurements,
1
the standard Bruker pulse program45 was used. H-7Li HOESY was
carried out with 16 number of scans, TD(F2) ) 8k, TD(F1) ) 256, 3
s of relaxation delay for acquisition, and a mixing time of 1.3 s. The
data were processed with SI(F2) ) 2k and SI(F1) ) 256. For the 1H-
13C HOESY measurements, 256 scans, TD(F2) ) 16k, TD(F1) ) 170,
2 s of relaxation delay, and a mixing time of 900 ms were applied.
The processing parameters were SI(F2) ) 4k and SI(F1) ) 128.
1
The H-1H NOESY measurements were also carried out with the
standard Bruker pulse program46 using 16 scans, TD(F2) ) 16k, TD-
(F1) ) 512, 5 s of relaxation delay for acquisition, and a mixing time
of 800 ms. Here the data were processed with SI(F2) ) 4k and SI(F2)
) 512. The temperatures for all measurements were calibrated with
the Bruker methanol sample and the AU program calctemp and were
controlled by a Bruker BVT 3000 temperature unit.
Acknowledgment. We gratefully acknowledge financial
support from the Fonds der Chemischen Industrie. Also we
thank Dr. R. Kerssebaum of Bruker Analytik GmbH for many
useful discussions on PFG diffusion experiments.
NMR Data Collection and Processing. The NMR spectra were
recorded on a Bruker DRX500 spectrometer equipped with a 5 mm
broadband triple-resonance Z-gradient probe (maximum gradient
strength 53.5 G/cm). All diffusion measurements were performed with
Supporting Information Available: Comparison between
calculated and experimental diffusion coefficients, additional
kinetic curves, and NMR spectra. This material is available free
of charge via the Internet at http://pubs.acs.org.
(37) Hodgman, C. D., Ed. Handbook of Chemistry and Physics; The Chemical
Rubber Publishing Company: Cleveland, OH, 1963; p 2260.
(38) Haupts, U.; Maiti, S.; Schwille, P.; Webb, W. W. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 13573-13578.
(39) Su, Q.; Klinman, J. P. Biochemistry 1999, 38, 8572-8581.
(40) Cabrita, E. J.; Berger, S. Magn. Reson. Chem. 2001, 39, 142-148.
(41) Ben-Amotz, D.; Willis, K. G. J. Phys. Chem. 1993, 97, 7736-7742.
(42) Edward, J. T. J. Chem. Educ. 1970, 47, 261-270.
JA055085O
(44) Jerschow, A.; Mueller, N. J. Magn. Reson. 1997, 125, 372-375.
(45) Yu, C.; Levy, G. C. J. Am. Chem. Soc. 1984, 106, 6533-6537.
(46) Wagner, R.; Berger, S. J. Magn. Reson., Ser. A 1996, 123, 119-121.
(43) Bloomfield, V. A. I. On-Line Biophysics Textbook, Schuster, T. M., Ed.;
2000; Volume Separation and Hydrodynamics; pp 2001-2005.
9
17342 J. AM. CHEM. SOC. VOL. 127, NO. 49, 2005