Association and dissociation of lithium cyanocuprates in ethereal solvents

Article abstract of DOI:10.1021/om200625z

We use a combination of electrospray ionization mass spectrometry and electrical conductivity measurements to probe the ions present in ethereal solutions of lithium cyanocuprates. Electrospray ionization mass spectrometry shows that solutions of LiCuR

Full text of DOI:10.1021/om200625z

ARTICLE
pubs.acs.org/Organometallics
Association and Dissociation of Lithium Cyanocuprates in Ethereal
Solvents
Aliaksei Putau and Konrad Koszinowski*
Department Chemie, Ludwig-Maximilians-Universit€at M€unchen, Butenandtstr. 5-13, 81377 M€unchen, Germany
S Supporting Information
b
ABSTRACT: We use a combination of electrospray ionization mass spectro-
metry and electrical conductivity measurements to probe the ions present in
ethereal solutions of lithium cyanocuprates. Electrospray ionization mass
spectrometry shows that solutions of LiCuR2 LiCN and Li0.8CuR0.8(CN)
3
(R = Me, Et, nBu, sBu, tBu, and Ph) in diethyl ether contain Linꢀ1CunR2nꢀ and
Linꢀ1CunRn(CN)nꢀ anions, respectively. Analogous species are also observed
for solutions of LiCutBu2 LiCN and Li0.8CutBu0.8(CN) in 2-methyltetrahy-
3
drofuran, cyclopentyl methyl ether, and methyl tert-butyl ether and were
previously found for solutions of lithium cyanocuprates in tetrahydrofuran.
Although the change of solvent thus does not lead to the formation of any
major new cuprate anions, it has a strong effect on the association/dissociation
equilibria. As directly confirmed by the conductivity experiments, contact ion
pairs strongly predominate in solutions of lithium cyanocuprates in diethyl ether, whereas the more polar tetrahydrofuran gives rise
to larger amounts of solvent-separated ion pairs; a particularly high dissociation tendency is observed for the LiCutBu2 LiCN
3
reagent. Temperature-variant conductivity measurements of LiCuPh2 LiCN solutions in tetrahydrofuran do not show a significant
3
temperature dependence of the association/dissociation equilibria for this system. The present findings largely support the results of
previous NMR spectroscopic studies and help to explain how the solvent affects the reactivity of lithium cyanocuprates.
diethyl ether, they observed a strong coupling, which indicated a
small distance between the Li+ counterion and the organyl
substituent and thus pointed to the presence of a contact ion
pair. In contrast, such a coupling was almost completely absent in
the more polar tetrahydrofuran (THF), suggesting a shift of the
association/dissociation equilibrium toward solvent-separated
ion pairs (eq 1). In line with this rationalization, the structures
of LiCuR2 crystals grown from diethyl ether correspond to
homodimeric contact ion pairs, whereas LiCuR2 crystallized
from solvents of higher affinity to Li+ contain solvent-separated
ions.5b Notably, the inferred solvent dependence of the associa-
tion/dissociation equilibria also permitted a straightforward
interpretation of reactivity data of lithium organocuprates pre-
viously obtained by Bertz et al.2j,6
1. INTRODUCTION
Lithium organocuprates have long been known as valuable
reagents in organic synthesis.1 The great potential of these
reagents comes at the price of an enormous complexity, however,
which has made mechanistic studies difficult and slowed down
the optimization of reagents and synthetic protocols.2 Key
progress in the elucidation of the structure and reactivity of
lithium organocuprates has been only achieved since the role of
association and dissociation equilibria of these reagents has been
focused on. A first breakthrough came with theoretical calcula-
tions by Nakamura and co-workers.3 These calculations pointed
to the crucial importance of Li+ centers in the organocuprate
reagents for activating Michael-type acceptors in 1,4 addition
reactions. As the Li+ centers are only present in associated
contact ion pairs, but not in dissociated solvent-separated ion
pairs, only the former should be effective in Michael-type
additions. This conclusion explained the previously observed
detrimental effect of crown ethers on the reactivity of lithium
organocuprates,4 as these strongly coordinating compounds
break up contact ion pairs. A second breakthrough then came
with the finding of Gschwind and collaborators that the associa-
tion/dissociation equilibria of homoleptic lithium organocup-
rates are largely determined by solvation. Using 1H,6Li HOESY
NMR spectroscopy, these authors showed for LiCuR2 model
systems (R = Me and (Me3Si)CH2) that the coupling between
the Li+ counterion and the R protons of the organyl substituent
strongly depends on the polarity of the solvent.5 In the less polar
ðLiCuR2Þ2ðsolvÞ h 2LiþðsolvÞ þ 2CuR2
ð1Þ
ꢀ
Despite the progress made so far, important aspects of the
association/dissociation equilibria of lithium organocuprates
await clarification. In particular, it remains to be shown whether
the observed solvent dependence of the equilibrium really is a
general phenomenon seen for a larger series of lithium organo-
cuprates and for other solvents in addition to THF and Et2O.
Moreover, only very little is known about the association/
dissociation equilibria of related heteroleptic cuprates, such as
Received: July 12, 2011
Published: August 16, 2011
r
2011 American Chemical Society
4771
dx.doi.org/10.1021/om200625z Organometallics 2011, 30, 4771–4778
|

Organometallics
ARTICLE
LiCuR(CN). Here, we address these questions by a combination
of electrospray ionization (ESI) mass spectrometry and con-
ductivity measurements. ESI mass spectrometry7 selectively
probes the charged components of the solutions under investiga-
tion and gives unambiguous stoichiometric information. There-
fore, it is increasingly employed for analyzing organometallic
reagents and intermediates.8 Following up on pioneering work
by Lipshutz et al.,9 we have recently shown that ESI mass
spectrometry is a useful method for providing qualitative insight
into the structural motifs and aggregation states of the ionic
components of lithium cyanocuprates in THF.10 We now extend
these studies to solutions of lithium cyanocuprates in Et2O and,
for selected cases, also probe solutions in 2-methyltetrahydrofur-
an (MeTHF), cyclopentyl methyl ether (CPME), and methyl
tert-butyl ether (MTBE). However, a caveat with respect to the
use of ESI mass spectrometry is that the ESI process disturbs the
system in a not easily predictable way. In the course of the ESI
process, the solution forms charged nanodroplets, from which
the analyte ions are then ejected into the gas phase and become
amenable to mass spectrometric detection.11 As the analyte con-
centration in the nanodroplets is higher than in solution12
and also the effective temperature may change,13 association/
dissociation equilibria are likely to be shifted. Although these
shifts should not prevent qualitative insight14 and the compar-
ison of different systems, they do not permit absolute quantita-
tion. As an independent and well-established method, we,
therefore, employ electrical conductivity measurements as well.
Like ESI mass spectrometry, these measurements selectively
probe the ionic components of solutions. While they can provide
valuable quantitative information, they afford only limited qua-
litative insight. Thus, ESI mass spectrometry and conductivity
measurements complement each other very well15 and promise
to improve our understanding of the association/dissociation
equilibria of lithium cyanocuprates in different ethereal solvents.
HCT quadrupole ion trap mass spectrometer (Bruker Daltonik) by a
pump-driven gastight syringe (1 mL hꢀ1). As the organocuprate solutions in
Et2O proved to be even more susceptible to hydrolysis than those in
THF, particular care was taken to exclude moisture from the inlet line
connecting the syringe with the ESI source. To this end, the inlet line was
flushed with dry THF for g1 h before introducing the sample solution.
Whereas the macroscopic stabilities of solutions of LiCuR2 LiCN held
3
in the syringe at room temperature were fairly high, decomposition of
solutions of Li0.8CuR0.8(CN) was fast (e5 min) and produced black or
greenish precipitates, which then could lead to clogging of the inlet line.
To avoid this problem, solutions of Li0.8CuR0.8(CN) were analyzed as
quickly as possible.
The ESI source was operated with N2 as a sheath and drying gas
(0.7 bar backing pressure and 5 L minꢀ1 flow rate, respectively) at an ESI
voltage of 3 kV. To minimize decomposition reactions during the ESI
process, we applied mild conditions identical to those reported pre-
viously (60 °C drying gas temperature and low potential differences
along the path of the ions).10,15 The helium-filled quadrupole ion trap
(estimated pressure p(He) ≈ 2 mTorr) was operated at a trap drive of
20. While this rather low value leads to a bias toward the detection of ions
with low m/z ratios, it avoids unwanted fragmentation of the labile
organocuprate ions in the quadrupole ion trap. For deliberate fragmen-
tation, the mass-selected ions (isolation widths e 2 amu) were subjected
to excitation voltages with amplitudes of Vexc and allowed to collide with
the He gas.
2.3. Conductivity Measurements. Electrical conductivity mea-
surements were performed with a SevenMulti instrument (Mettler
Toledo) and a stainless steel electrode cell (InLab741, Mettler Toledo,
k
cell = 0.1 cmꢀ1) calibrated against a 0.1 M solution of aqueous KCl at
298 K. In a first set of experiments, the conductivity of solutions of
n
t
CuCN/(RLi)n (n = 1 or 2; R = Bu, Bu, and Ph) in THF or Et2O,
respectively, was measured at 258 K. To this end, we monitored the
conductivity after the successive addition of 1 and 2 equiv of LiR to a
suspension of CuCN. Iodometric titration17 of aliquots of the resulting
solutions showed that we were able to suppress hydrolysis quite
successfully (measured concentrations of c(LiCuR(CN)) = 97 (
2 mM and c(LiCuR2 LiCN) = 91 ( 4 mM compared to nominal
3
2. EXPERIMENTAL SECTION
concentrations of c = 104 mM; for more diluted solutions of LiCuPh2
LiCN, we measured c = 47 and 21 mM for nominal concentrations of
c = 52 and 26 mM). In a second set of experiments, we measured the
3
2.1. Sample Preparation. Standard Schlenk techniques were
employed for handling air- and moisture-sensitive substances through-
out. THF and Et2O were distilled from sodium/benzophenone;
MeTHF, CPME, and MTBE were dried over a molecular sieve (4 Å).
CuCN was dried by repeated heating under vacuum at 350 °C. Solutions
of organolithium compounds RLi were used as purchased: MeLi (1.49
conductivity of LiCuPh(CN) and LiCuPh2 LiCN in THF at various
3
temperatures. Besides probing separate sample solutions at different
(fixed) temperatures, we also recorded how increasing the temperature
in increments of 5 K (starting at 233 K) changed the conductivity of a
single sample solution. Iodometric titrations performed at the beginning
and end of the measurements again showed only very little hydrolysis
(measured concentrations c g 100 mM compared to nominal concen-
trations of c = 107 mM).
n
M) in Et2O, EtLi (0.42 M) in benzene/cyclohexane (90/10), BuLi
(2.37 M) in hexane, sBuLi (1.58 M) in cyclohexane, tBuLi (1.88 M) in
pentane, and PhLi (1.74 M) in Bu2O. The exact concentrations were
determined by titration of 1,3-diphenyl-2-propanone tosylhydrazone.16
Solutions of CuCN/(RLi)n were prepared by treating suspensions of
CuCN in the solvent of choice (THF, Et2O, MeTHF, CPME, or MTBE)
with RLi under argon at ꢀ78 °C. After stirring at this temperature for 1 h,
the CuCN completely dissolved for CuCN/(RLi)n (n = 1 or 2), forming
3. RESULTS
3.1. ESI Mass Spectrometry. LiCuR2 LiCN Solutions. The
3
LiCuR(CN) and LiCuR2 LiCN, respectively. In the case of the sample
3
negative ion mode ESI mass spectra of solutions of LiCuR2
3
solutions prepared for ESI mass spectrometric analysis, the smaller
volumes necessarily increase the likelihood of small errors in the measured
reagent quantities. As we have shown previously for solutions in THF,
such errors can be particularly detrimental to the analysis of LiCuR(CN).
To avoid an excess ofRLi in theprobed samples, we, therefore, added only
0.8 equiv of LiR. The resulting solutions of nominal Li0.8CuR0.8(CN)
composition are supposed to contain LiCuR(CN) because the excess
CuCN does not dissolve, as could also be directly seen from the presence
of a solid residue.
LiCN (R = Me, Et, nBu, sBu, tBu, and Ph) in Et2O are dominated
by complexes of the homologous series Linꢀ1CunR2nꢀ, n = 1ꢀ3
(Figure 1 and Figures S1ꢀS5 in the Supporting Information).
These species are identical to those observed in our previous
investigation of solutions of LiCuR2 LiCN in THF and, as
3
expected, also show the same fragmentation behavior (for a
detailed discussion of the gas-phase fragmentation of mass-
selected cuprate anions, see refs 10 and 18). Interestingly,
however, the distributions obtained for the Et2O solutions are
systematically shifted to higher aggregation states compared with
2.2. ESI Mass Spectrometry. Organocuprate solutions of
c ≈ 25 mM were continuously administered into the ESI source of an
4772
dx.doi.org/10.1021/om200625z |Organometallics 2011, 30, 4771–4778

Organometallics
ARTICLE
Table 1. Organocuprate Anions Observed upon ESI of Et2O
Solutions of LiCuR2 LiCN in High [++], Medium [+], and
3
Low/Negligible [ꢀ] Relative Abundancea,10
n
R = Me
Et
nBu
sBu
tBu
Ph
ꢀ
Linꢀ1CunR2n 1 ꢀ (ꢀ) ꢀ (ꢀ) ꢀ (+) ꢀ (++) + (++) + (++)
2 ++ (++) ꢀ (ꢀ) ꢀ (ꢀ) ꢀ (ꢀ) ++ (+) ꢀ (ꢀ)
3 ++ (++) ++ (++) ++ (++) ++ (++) ꢀ (ꢀ) ++ (++)
a For comparison, the relative abundances measured for the analogous
solutions in THF are given in parentheses.
Figure 3. Negative ion mode ESI mass spectrum of a 25 mM solution of
Li0.8CuꢀtBu0.8(CN) in Et2O: a = Cu2tBu2(CN)ꢀ, b = Li3Cu4tBu2(OH)2-
(CN)4
.
situation in THF. For the latter solvent, however, we also
+
+
detected the complexes Li(THF)3 and Li2(CN)(THF)3 ,
whereas ions with n = 3 solvent molecules attached are largely
missing in the mass spectra recorded for Et2O solutions. For
solutions of LiCuPh2 LiCN in Et2O, we also observe the
3
+
incorporation of Bu2O in the ions Li2(CN)(Et2O)2ꢀn(Bu2O)n
(n = 1 and 2) (Figure S13, Supporting Information), although
the fraction of Bu2O in solution (stemming from the preparation
of the reagent; see the Experimental Section) is lower than that
of Et2O by a factor of 60. While none of these cations contain a
Figure 1. Negative ion mode ESI mass spectrum of a 25 mM solution of
LiCutBu2 LiCN in Et2O. The ion “a” (m/z 605/607/609) corresponds
3
t
to Li3Cu4 Bu4(CN)4ꢀ, which presumably results from partial hydrolysis.
+
Cu center, we do find small amounts of Li2CuR2(Et2O)2 for
R = Me. The ESI mass spectra measured for solutions of LiCutBu2
3
LiCN in MeTHF, CPME, and MTBE are similar to those
+
obtained for Et2O and THF solutions in that Li(solv)n and
Li2(CN)(solv)n+ are the predominant cations observed (Figures
S14ꢀS16, Supporting Information).
Solutions of Li0.8CuR0.8(CN). As detailed in the Experimental
Section, samples of a nominal composition of Li0.8CuR0.8(CN)
contain LiCuR(CN) in the solution phase because the excess
CuCN does not dissolve. In line with this assessment, we have
previously shown that solutions of LinCuRn(CN) in THF
(n= 0.5, 0.8, and 1.0) yield very similar ESI mass spectra. Negative ion
mode ESI of solutions of Li0.8CuR0.8(CN) in Et2O affords
a multitude of different organocuprate anions (Figure 3 and
Figures S17ꢀS21, Supporting Information), most of which are
already known from our previous analysis of the corresponding
THF solutions.10 A first set of ions belongs to the homologous
series Linꢀ1CunRn(CN)nꢀ. Ions of this series or the correspond-
ing partially hydrolyzed species Linꢀ1CunRnꢀ1(OH)(CN)nꢀ are
observed in all cases (Table 2; the partial hydrolysis most likely
occurs in ionꢀmolecule reactions with background water pre-
sent in the ion trap. See the Addition/Correction in ref 10).
Compared to the situation in THF, the change to Et2O as the
solvent does not seem to result in a clear shift in the aggregatioꢀn
states. A second set of prominent ions comprises Linꢀ1CunR2n
complexes, which do not show the expected stoichiometry but
Figure 2. Positive ion mode ESI mass spectrum of a 25 mM solution of
LiCutBu2 LiCN in Et2O.
3
their THF counterparts (Table 1). Also note that the absolute
ESI signal intensities are considerably lower in the case of the
Et2O solutions. For R = tBu, we also probed the effect of other
ꢀ
ethereal solvents and again found Linꢀ1CunR2n anions, n = 1
and 2 (Figures S6ꢀS8, Supporting Information). While the
fraction of the dimeric complex LiCu2R4ꢀ is relatively small for
MeTHF solutions, this ion apparently prevails in CPME
and MTBE.
instead are typical of LiCuR2 LiCN solutions (see above). As we
3
have discussed previously,10 species such as LiCuR2 might
possibly form from LiCuR(CN) in Schlenk-type equilibria
(along with LiCu(CN)2). The remaining anions observed for
solutions of Li0.8CuR0.8(CN) in Et2O exhibit intermediate
stoichiometries and are of limited abundance only. For Li0.8Cut-
Bu0.8(CN) in MeTHF, CPME, and MTBE, we also find
Turning to the positive ion mode ESI mass spectra of solutions
of LiCuR2 LiCN in Et2O, we observe Li(Et2O)2+ and Li2(CN)-
3
+
(Et2O)2 as main species (Figure 2 and Figures S9ꢀS13,
Supporting Information). With the exception of R = Et, Li2-
+
(CN)(Et2O)2 predominates in all cases, thus resembling the
4773
dx.doi.org/10.1021/om200625z |Organometallics 2011, 30, 4771–4778

Organometallics
ARTICLE
Table 2. Organocuprate Anions Observed upon ESI of Et2O Solutions of Li0.8CuR0.8(CN) in High [++], Medium [+], and Low/
Negligible [ꢀ] Relative Abundancea,10
n
R = Me
Et
nBu
sBu
tBu
Ph
ꢀ
Linꢀ1CunRn(CN)n
2
3
4
5
3
4
1
2
3
++ (++)
ꢀ (+)
+ (++)
ꢀ (+)
+ (+)
+ (+)
ꢀ (+)
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ (+)
++ (++)
ꢀ (+)
ꢀ (ꢀ)
+ (++)
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ (+)
++ (+)
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ (++)
ꢀ (+)
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ (ꢀ)
++ (++)
++ (+)
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ (ꢀ)
++ (+)
ꢀ (ꢀ)
++ (ꢀ)
ꢀ (ꢀ)
ꢀ (ꢀ)
+ (ꢀ)
ꢀ (+)
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ (+)
ꢀ (+)
ꢀ (ꢀ)
ꢀ (ꢀ)
++ (++)
ꢀ (ꢀ)
+ (ꢀ)
ꢀ (ꢀ)
+ (++)
ꢀ (+)
ꢀ (ꢀ)
ꢀ (ꢀ)
+ (ꢀ)
ꢀ (ꢀ)
ꢀ (ꢀ)
++ (+)
ꢀ (+)
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ
Linꢀ1CunRnꢀ1(OH)(CN)n
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ (ꢀ)
++ (+)
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ (ꢀ)
ꢀ
Linꢀ1CunR2n
Cu2R2ꢀ(CN)ꢀ
Cu3R4
ꢀ
LiCu4R6
a For comparison, the relative abundances measured for the analogous solutions in THF are given in parentheses.
Table 3. Molar Electrical Conductivities Determined for
Solutions of LiCuR2 LiCN (c = 91 ( 4 mM) and LiCuR(CN)
3
(c = 97 ( 2 mM) in THF and Et2O at 258 K (Activity
Coefficients Are Neglected)
molar conductivity
molar conductivity
Λ(LiCuR(CN))/(S cm2 molꢀ1
)
Λ(LiCuR2 LiCN)/(S cm2 molꢀ1
)
3
R
THF
Et2O
THF
Et2O
nBu 13( 1
tBu 19( 1
1.00 ( 0.02
6.2 ( 0.1
0.3 ( 0.1
0.008 ( 0.005
0.7 ( 0.1
4.0 ( 0.5
Ph
8.2 ( 0.3
0.20 ( 0.05
0.42 ( 0.04
0.20 ( 0.05
lower the viscosity of the solvent, thus resulting in enhanced ion
mobilities. Taking into account Walden’s rule, which assumes that
the product of the molar conductivity Λ of a given electrolyte and
the viscosity of the solvent η is constant,19 we attempt to fit the
measured conductivities on the basis of the known temperature
dependence of η(THF).20 The obtained fit is reasonably good
(Figure 5), suggesting that the observed temperature dependence
of the molar conductivity indeed can be rationalized by the change
in the viscosity of the solvent.
Figure 4. Positive ion mode ESI mass spectrum of a 25 mM solution of
Li0.8CutBu0.8(CN) in Et2O.
ꢀ
ꢀ
Linꢀ1CunRn(CN)n and Linꢀ1CunR2n complexes as the pre-
dominating anions (Figures S22ꢀS24, Supporting Information).
The positive ion mode ESI mass spectra obtained for solutions
of Li0.8CuR0.8(CN) in Et2O in all cases show Li(Et2O)2+ as the
main peak (Figure 4 and Figures S25ꢀS29, Supporting In-
formation). Very similarly, the corresponding THF solutions
also afforded solvated Li+ ions as the predominant species.10 Less
abundant cations observed for the Et2O solutions are Li2(CN)-
LiCuR(CN) Solutions. The molar conductivities of solutions of
LiCuR(CN) in THF are significantly smaller than those of the
corresponding LiCuR2 LiCN solutions (Table 3). Again, a
3
strong dependence on the nature of the R substituent is notice-
able: Λ(LiCutBu(CN)) . Λ(LiCuPh(CN)) > Λ(LiCunBu-
(CN)). The conductivities in Et2O are even lower (also lower
(Et2O)2+ and Li2CuR(CN)(Et2O)2 . Solutions of Li0.8CutBu0.8-
+
(CN) in MeTHF also yield Li(solv)n+ ions as main species (n = 2
and 3), whereas Li2(CN)(solv)2+ prevails for CPME and MTBE
(Figures S30ꢀS32, Supporting Information).
than the conductivities measured for Et2O solutions of LiCuR2
LiCN) but show a similar trend. For LiCuPh(CN) in THF, we
3
3.2. Electrical Conductivity Measurements. LiCuR2 LiCN
observe a temperature dependence that, like in the case of its
3
Solutions. Solutions of LiCuR2 LiCN in THF display significant
LiCuPh2 LiCN counterpart, is reproduced by a simple fit that
3
3
electrical conductivities (Table 3). The determined molar con-
ductivities show a clear dependence on the nature of the R
substituent: Λ(LiCutBu2 LiCN) > Λ(LiCunBu2 LiCN) >
only considers the effect of the changed solvent viscosity (Figure
S34, Supporting Information).
3
3
Λ(LiCuPh2 LiCN). The conductivities of LiCuR2 LiCN in
3
3
4. DISCUSSION
Et2O are much smaller, but exhibit a very similar trend.
For LiCuPh2 LiCN in THF, we also investigated the con-
4.1. Equilibria Operative. We rationalize the detection of
3
centration and temperature dependence. At lower concentra-
tions, the molar conductivity increases (Figure S33, Supporting
Information). A rise in temperature also increases the conductivity
(Figure 5). This behavior is expected because higher temperatures
Linꢀ1CunR2nꢀ anions in LiCuR2 LiCN solutions by the opera-
3
tion of association/dissociation equilibria, asdepictedinScheme 1.
The proposed scenario essentially corresponds to the equilibrium
already suggested by Gschwind and collaborators5 (eq 1) but, in
4774
dx.doi.org/10.1021/om200625z |Organometallics 2011, 30, 4771–4778

Organometallics
ARTICLE
Scheme 1. Association/Dissociation Equilibria Proposed To
Be Operative in Ethereal Solutions of LiCuR2 LiCNa
3
Figure 5. Molar electrical conductivity of a solution of LiCuPh2 LiCN
3
in THF (c = 98 mM; activity coefficients are neglected) as a function of
temperature. The open symbols represent data points collected during a
single conductivity measurement, in which the temperature was raised
continuously from 233 to 298 K in 5 K increments. The filled symbols
represent data points collected independently for different samples at
fixed temperatures. The red line corresponds to a fit that only takes into
account the effect of the temperature dependence of the solvent
viscosity.
a The highlighted species have been observed by ESI mass spectrometry.
species in significant quantities. Instead, it results in a shift in the
addition, also accounts for the formation of higher ionic aggre-
gates. While ESI mass spectrometry cannot detect the neutral
homodimers Li2Cu2R4, the presence of these species in ethereal
association/dissociation equilibria for the LiCuR2 LiCN re-
3
agents. We first analyze whether the observed shift reflects
the situation in solution or whether it might mirror a different
behavior during the ESI process. THF and Et2O not only
differ in their polarity but also in their boiling point (Table 4).
Whereas the higher polarity of THF is likely to favor dissocia-
tion in solution, the lower boiling point of Et2O should
facilitate desolvation during the ESI process and thus could
also explain the higher propensity to association observed for
solutions of LiCuR2 LiCN has been proven by NMR spectros-
3
copy, X-ray scattering, and ebullioscopic methods.5,21
For LiCuR(CN) solutions, the situation is less clear because
the aggregation state of the neutral, undissociated component
[LiCuR(CN)] is not known precisely. Nonetheless, the experi-
mental findings point to the operation of association/dissocia-
tion equilibria, such as eq 2. Here, the incorporation of CNꢀ in
the cuprate anions explains why simple, cyanide-free Li+(solv)
this solvent. However, the results obtained for LiCutBu2
3
cations prevail, in contrast to the case of LiCuR2 LiCN. Associa-
3
LiCN in MeTHF, CPME, and MTBE clearly show that the
solvent polarity is the decisive factor (Table 4). It is obvious
that the Lewis-basic ethereal solvents do not interact with
both cations and anions in a similar way but that they bind to
tion of Li+ with neutral [LiCuR(CN)] rationalizes the formation
of Li2CuR(CN)+(solv) cations, eq 3.
ꢀ
n½LiCuRðCNÞꢁ h LiþðsolvÞ þ Linꢀ1CunRnðCNÞn
LiþðsolvÞ þ ½LiCuRðCNÞꢁ h Li2CuRðCNÞþðsolvÞ
ð2Þ
ð3Þ
+
the Lin(CN)nꢀ1 cations (n = 1 and 2) much more strongly
than to the cuprate anions. As a consequence, we do not detect
any microsolvated cuprate anions by ESI mass spectrometry,
whereas we exclusively find Lin(CN)nꢀ1(solv)x+ cations, x = 2
and 3. Note that the number of bound solvent molecules
observed for the gaseous ions presumably does not corre-
spond to the first solvation shell in solution but rather reflects
the relative interaction energies (too weakly bound molecules
will be lost upon energetic collisions during the ESI process).
The higher number of Li+-bound THF and MeTHF molecules
(x g 2) correlates very well with the higher macroscopic
polarity of these solvents and their effect of shifting the
equilibria toward dissociated ions.
A comparison between the Linꢀ1CunR2nꢀ anions charactꢀeristic
of LiCuR2 LiCN solutions and the Linꢀ1CunRn(CN)n ob-
3
served for LiCuR(CN) solutions suggests that the latter reagents
favor higher aggregation/association states, thus reducing the
concentration of free ions. The relatively lower electrical con-
ductivities measured for solutions of LiCuR(CN) fully corrobo-
rate this interpretation. Also note that the higher dissociation
tendency of LiCuR2 LiCN helps to rationalize why already small
3
amounts of this species present in solutions of LiCuR(CN) can
result in appreciable concentrations of Linꢀ1CunR2nꢀ anions and
their detection by ESI mass spectrometry. As discussed previously,10
the lower dissociation tendencies of the LiCuR(CN) reagents pre-
sumably result from the incorporation of the cyanide in the cuprate
species. CNꢀ can not only bridge different Cu centers but, owing to
its ambident nature, at the same time, also bind to a Li+ cation with
high affinity, thus causing the buildup of larger aggregates.
The conductivity data provide independent and unambiguous
evidence that THF favors the dissociation of LiCuR2 LiCN in
3
comparison to Et2O. For the LiCuR(CN) reagents, the trend is
much weaker. In line with this observation, the ESI mass spectra
measured for solutions of LiCuR(CN) in THF, on the one hand,
and Et2O, on the other, do not display notable differences.
4.3. Effect of the Organyl Substituent. Although all of the
4.2. Effect of the Solvent. The ESI mass spectrometric experi-
ments show that the transition from THF to Et2O or other
ethereal solvents does not lead to the formation of new ionic
LiCuR2 LiCN reagents sampled, as well as their LiCuR(CN)
3
counterparts, behave quite similarly, some differences are discernible.
4775
dx.doi.org/10.1021/om200625z |Organometallics 2011, 30, 4771–4778

Organometallics
ARTICLE
Table 4. Properties of Ethereal Solvents Sampled and Ag-
solutions of LiCu(CH2(Me3Si))2 in THF, John et al. concluded
that the association/dissociation equilibrium of this reagent is
strongly affected by temperature.5b According to the authors, the
formation of the solvent-separated ion pairs is enthalpically
favored, but entropically disfavored, because the enhanced
solvation of the free Li+ cations results in the loss of degrees of
t
gregation Tendencies of Linꢀ1Cun Bu2nꢀ Anions in These
Solvents As Determined by ESI Mass Spectrometry of Solu-
tions of LiCutBu2 LiCN
3
t
relative permittivity
boiling
I(LiCu2 Bu4ꢀ)/
I(CutBu2
)
ꢀ
freedom.5b For the LiCuPh2 LiCN/THF system, our present
solvent
ε (298 K)
point (K)
3
7.42a
6.97a
4.76b
4.24a
2.60b
338a
353b
379b
308a
328b
<1
<1
>1
>1
>1
measurements show only a modest increase of the conductivity
as a function of temperature. The observed increase can be fully
explained by the effect of the reduced viscosity and thus excludes
a pronounced temperature effect on the association/dissociation
equilibrium. This result does not directly disagree with the
conclusions of John et al., because the cuprate reagents probed
in both studies are different.
THF
MeTHF
CPME
Et2O
MTBE
a Reference 22. b Reference 23.
4.5. Comparison of Analytical Methods. Beyond providing
insight specific to the lithium cuprate reagents examined, the
present investigation, in conjunction with previous work, also
permits a comparison of different experimental methods used for
the analysis of ionic species in solution. NMR spectroscopy,
electrical conductivity measurements, and ESI mass spectro-
metry all agree that the association/dissociation equilibria of
lithium cyanocuprates are largely governed by the nature of the
solvent and, in particular, its Li+ affinity. While NMR spectros-
copy and conductometry constitute well-established techniques
and thus are expected to give the same results, the consistency of
the ESI mass spectrometric findings deserves some further
comments. As mentioned above, the mass spectrometrically
detectable ions do not originate directly from the sampled
solution, but from the intermediately formed nanodroplets.
The good agreement between the results obtained by ESI mass
spectrometry and by conventional analytical methods indicates
that the relative position of the association/dissociation equili-
bria is largely preserved in the nanodroplets. Note that the ESI
mass spectrometric experiments are sensitive to the nature of the
solvent even if the observed ions do not contain any solvent
t
As the conductivity measurements clearly show, the Bu
substituent favors dissociation and the formation of solvent-
n
separated ion pairs more than the Bu and Ph groups. In full
accordance with this observation, ESI mass spectrometry finds
tBu to be the only substituent for which the aggregation state of
ꢀ
the Linꢀ1CunR2n anions is limited to n e 2. We ascribe the
lower aggregation tendency of the tBu-containing cuprates to the
higher steric demands of this substitꢀuent, which apparently
prevents the association of >2 CutBu2 monomers. The con-
ductivity data moreover suggest that the nBu-bearing homoleptic
cuprates give somewhat higher fractions of dissociated ions than
their Ph-containing analogs. This finding seems to be at odds
with the ESI mass spectrometric results, which point to a slightly
higher aggregation tendency for the nBu-bearing cuprates. This
discrepancy may possibly arise from the different temperatures in
both experiments (258 K for the conductivity measurements and
approximately 298 K for the ESI mass spectrometric studies).
However, it could also be the case that the deviating behavior
observed by ESI mass spectrometry is due to a (though rather
small) perturbation of the system caused by the very ESI process.
The recorded molar conductivities Λ also comprise informa-
tion on the absolute fractions of dissociated ions. For calculating
the degree of dissociation R according to eq 4, the limiting molar
conductivity Λ0 must be known.
ꢀ
molecules, as the example of the Linꢀ1CunR2n complexes
demonstrates. The observed solvent-dependent shift in the
aggregation state of these ions rationalizes at the microscopic
level what the conductivity measurements find macroscopically.
The consistency between the ESI mass spectrometric and
conductometric results is not limited to the effect of the solvent
but also extends to the observation of a particularly high
R ¼ Λ=Λ0
ð4Þ
Although Λ0 could be derived by the extrapolation of experi-
mental data to c = 0, the very steep slope and the increased
susceptibility to inevitable hydrolysis reactions at lowest con-
centrations render such an approach unreliable. For a rough
estimation, we instead approximate the limiting molar conduc-
tivities of the lithium cuprates by known values of other electro-
lytes in THF. At 298 K, the limiting molar conductivities of
many diverse 1:1 electrolytes in THF all fall into the range of
dissociation tendency of the LiCutBu2 LiCN reagent. In con-
3
trast, a comparison of the data for LiCunBu2 LiCN and Li-
3
CuPh2 LiCN possibly points to some smaller deviations
3
between the two methods and thus seems to suggest that their
overall agreement is not absolutely perfect. Nevertheless, the
present results demonstrate the suitability of ESI mass spectro-
metry to probe the speciation of cuprate ions in solution and to
provide qualitatively correct insight into their association/dis-
sociation behavior. This assessment is in line with the conclu-
sions of several other recent studies that investigated the
performance of ESI mass spectrometry.14,15
75 < Λ0(298 K) < 135 S cm2 molꢀ1 20,24
which converts into
,
Λ0(258 K) = 65 ( 20 S cm2 molꢀ1 on the basis of Walden’s
rule.19 If, simplistically, we apply this value to the lithium
cuprates, we obtain effective degrees of dissociation of 0.09 e
R(THF) e 0.44 and 0.002 e R(Et2O) e 0.14 for solutions of
LiCuR2 LiCN at concentrations of c ≈ 100 mM at 258 K.25
3
5. CONCLUSIONS
These estimates indicate that, even in the more polar THF, the
lithium cuprates are far from being completely dissociated. This
assessment is also consistent with the ESI mass spectrometric
experiments, which show abundant Linꢀ1CunR2nꢀ aggregates in
all cases examined.
We have used a combination of ESI mass spectrometry and
electrical conductivity measurements to probe the speciation of
LiCuR2 LiCN and LiCuR(CN) reagents in ethereal solvents.
3
ESI mass spectrometry shows that the transition from THF to
4.4. Effect of the Temperature. From the temperature
Et2O does not give rise to the formation of any major new ions
dependence of the 1H,6Li HOESY coupling observed for
for the LiCuR2 LiCN systems but that Linꢀ1CunR2nꢀ anions are
3
4776
dx.doi.org/10.1021/om200625z |Organometallics 2011, 30, 4771–4778

Organometallics
ARTICLE
present in both solvents. However, the change to Et2O, that is,
the less polar solvent, causes a shift of the equilibria toward higher
aggregation states. In the case of the LiCuR(CN) reagents, we
consistently find Linꢀ1CunRn(CN)nꢀ anions. The relatively high
aggregation states of these complexes presumably arise from the
presence of cyanide ions that adopt bridging binding modes.
’ REFERENCES
(1) (a) Gilman, H.; Jones, R. G.; Woods, L. A. J. Org. Chem. 1952,
17, 1630–1634. (b) Lipshutz, B. H.; Wilhelm, R. S.; Kozlowski, J. A.
Tetrahedron 1984, 40, 5005–5038. (c) Lipshutz, B. H. Synthesis
1987, 325–341. (d) Rappoport, Z., Marek, I., Eds. The Chemistry of
Organocopper Compounds; Wiley: Hoboken, NJ, 2009.
Additional experiments on solutions of LiCutBu2 LiCN and
(2) (a) Lipshutz, B. H.; Wilhelm, R. S.; Floyd, D. M. J. Am. Chem. Soc.
1981, 103, 7672–7674. (b) Lipshutz, B. H. Synlett 1990, 119–128. (c)
Lipshutz, B. H.; Sharma, S.; Ellsworth, E. L. J. Am. Chem. Soc. 1990,
112, 4032–4034. (d) Bertz, S. H. J. Am. Chem. Soc. 1990, 112, 4031–
4032. (e) Stemmler, T.; Penner-Hahn, J. E.; Knochel, P. J. Am. Chem.
Soc. 1993, 115, 348–350. (f) Snyder, J. P.; Spangler, D. P.; Behling, J. R.;
Rossiter, B. E. J. Org. Chem. 1994, 59, 2665–2667. (g) Barnhart, T. M.;
Huang, H.; Penner-Hahn, J. E. J. Org. Chem. 1995, 60, 4310–4311. (h)
Stemmler, T. L.; Barnhart, T. M.; Penner-Hahn, J. E.; Tucker, C. E.;
Knochel, P.; Boehme, M.; Frenking, G. J. Am. Chem. Soc. 1995,
117, 12489–12497. (i) Snyder, J. P.; Bertz, S. H. J. Org. Chem. 1995,
60, 4312–4313. (j) Bertz, S. H.; Miao, G.; Eriksson, M. Chem. Commun.
1996, 815–816.
(3) (a) Nakamura, E.; Mori, S.; Morokuma, K. J. Am. Chem. Soc.
1997, 119, 4900–4910. (b) Mori, S.; Nakamura, E. Chem.—Eur. J. 1999,
5, 1534–1543. (c) Nakamura, E.; Yamanaka, M. J. Am. Chem. Soc. 1999,
121, 8941–8942.
(4) (a) Ouannes, C.; Dressaire, G.; Langlois, Y. Tetrahedron Lett.
1977, 18, 815–818. (b) Bertz, S. H.; Gibson, C. P.; Dabbagh, G.
Organometallics 1988, 7, 227–232.
(5) (a) Gschwind, R. M.; Rajamohanan, P. R.; John, M.; Boche, G.
Organometallics 2000, 19, 2868–2873. (b) 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. (c) Gschwind,
R. M.; Xie, X.; Rajamohanan, P. R.; Auel, C.; Boche, G. J. Am. Chem. Soc.
2001, 123, 7299–7304. (d) Xie, X.; Auel, C.; Henze, W.; Gschwind,
R. M. J. Am. Chem. Soc. 2003, 125, 1595–1601. (e) Henze, W.; Vyater,
A.; Krause, N.; Gschwind, R. M. J. Am. Chem. Soc. 2005, 127, 17335–
17342. (f) Gschwind, R. M. Chem. Rev. 2008, 108, 3029–3053.
(6) (a) Bertz, S. H.; Eriksson, M.; Miao, G.; Snyder, J. P. J. Am. Chem.
Soc. 1996, 118, 10906–10907. (b) Bertz, S. H.; Chopra, A.; Eriksson, M.;
Ogle, C. A.; Seagle, P. Chem.—Eur. J. 1999, 5, 2680–2691.
(7) (a) Yamashita, M.; Fenn, J. B. J. Phys. Chem. 1984, 88, 4451–
4459. (b) Yamashita, M.; Fenn, J. B. J. Phys. Chem. 1984, 88, 4671–4675.
(8) For selected reviews and highlights, see: (a) Plattner, D. A. Int. J.
Mass Spectrom. 2001, 207, 125–144. (b) Chen, P. Angew. Chem. 2003,
115, 2938–2954. Angew. Chem., Int. Ed. 2003, 42, 2832–2847. (c)
Henderson, W.; McIndoe, J. S. Mass Spectrometry of Inorganic, Coordina-
tion and Organometallic Compounds: Tools, Techniques, Tips; Wiley:
Chichester, U.K., 2005; pp 175ꢀ219. (d) M€uller, C. A.; Markert, C.;
Teichert, A. M.; Pfaltz, A. Chem. Commun. 2009, 1607–1618. (e)
Agrawal, D.; Schr€oder, D. Organometallics 2011, 30, 32–35. (f) Coelho,
F.; Eberlin, M. N. Angew. Chem. 2011, 123, 5370–5372. Angew. Chem.,
Int. Ed. 2011, 50, 5261–5263.
3
LiCutBu(CN) in MeTHF, MTBE, and CPME as well as con-
ductivity measurements confirm that the polarity of the solvent
and its Li+ affinity control the association/dissociation behavior
of lithium cyanocuprates. Thus, our findings agree with the results
from previous NMR spectroscopic studies and demonstrate the
generality of the relation between solvent polarity and association
tendency for ethereal solutions of lithium cyanocuprates.5 This
dependence also bears practical importance because only the
associated contact ion pairs of lithium cuprates are supposed to
undergo efficient Michael-type addition reactions.3,5
Besides the solvent, the nature of the organyl substituent also
influences the association/dissociation equilibria. As the example
of LiCutBu2 LiCN shows, increased steric hindrance results in
3
higher degrees of dissociation, whose absolute values can be
estimated on the basis of the measured molar conductivities. For
the case of LiCuPh2 LiCN in THF, we have also investigated the
3
temperature dependence of the molar conductivity. The modest
increase observed at higher temperatures can be simply explained
by the effect of the lowered solvent viscosity. In contrast to the
conclusions drawn by John et al. for LiCu(CH2(Me3Si))2 in
THF, we thus do not find any evidence of a significant tempera-
ture dependence of the association/dissociation equilibria of
lithium cyanocuprates.
In a broader context, the present work adds to a growing
number of studies that demonstrate the suitability of ESI mass
spectrometry for probing ion speciation in solution.14,15 Although
the ESI process most likely will shift association equilibria relative
to the situation in solution, qualitative trends in ion speciation
seem to be remarkably robust. Until this finding has been
further validated, however, the most reliable approach remains
the combination of ESI mass spectrometry with other, well-
established analytical methods, such as electrical conductivity
measurements or NMR spectroscopy.15,26,27
’ ASSOCIATED CONTENT
S
Supporting Information. Additional negative and posi-
b
tive ion mode ESI mass spectra and figures showing the
concentration and temperature dependence of the molar con-
ductivity of lithium cyanocuprate solutions. This material is
available free of charge via the Internet at http://pubs.acs.org.
(9) (a) Lipshutz, B. H.; Stevens, K. L.; James, B.; Pavlovich, J. G.;
Snyder, J. P. J. Am. Chem. Soc. 1996, 118, 6796–6797. (b) Lipshutz,
B. H.; Keith, J.; Buzard, D. J. Organometallics 1999, 18, 1571–1574.
(10) Putau, A.; Koszinowski, K. Organometallics 2010, 29, 3593–
3601. 2010, 29, 6841–6842 (Addition/Correction).
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: konrad.koszinowski@cup.uni-muenchen.de.
(11) (a) Iribarne, J. V.; Thomson, B. A. J. Chem. Phys. 1976,
64, 2287–2294. (b) Thomson, B. A.; Iribarne, J. V. J. Chem. Phys.
1979, 71, 4451–4463.
(12) (a) Wang, H.; Agnes, G. R. Anal. Chem. 1999, 71, 3785–3792.
(b) Wang, H.; Agnes, G. R. Anal. Chem. 1999, 71, 4166–4172. (c)
Wortmann, A.;Kistler-Momotova, A.; Zenobi, R.; Heine, M. C.;Wilhelm, O.;
Pratsinis, S. E. J. Am. Soc. Mass Spectrom. 2007, 18, 385–393. (d) Tsierkezos,
N. G.; Roithovꢀa, J.; Schr€oder, D.; Onꢁcꢀak, M.; Slavíꢁcek, P. Inorg. Chem. 2009,
48, 6287–6296.
(13) Luedtke, W. D.; Landman, U.; Chiu, Y.-H.; Levandier, D. J.;
Dressler, R. A.; Sok, S.; Gordon, M. S. J. Phys. Chem. A 2008,
112, 9628–9649.
’ ACKNOWLEDGMENT
We thank Prof. Herbert Mayr for his continuous generous
support and gratefully acknowledge funding from LMU
M€unchen (LMUexcellent), the Deutsche Forschungsgemeinschaft
(SFB 749), the Center for Integrated Protein Science Munich,
the Fonds der Chemischen Industrie, and the Dr. Otto R€ohm
Ged€achtnisstiftung.
4777
dx.doi.org/10.1021/om200625z |Organometallics 2011, 30, 4771–4778

Organometallics
ARTICLE
(14) For a recent instructive example, see: Schr€oder, D.; Duchꢀaꢁckovꢀa,
L.; Tarꢀabek, J.; Karwowska, M.; Fijalkowski, K. J.; Onꢁcꢀak, M.; Slavíꢁcek, P.
J. Am. Chem. Soc. 2011, 133, 2444–2451.
(15) Koszinowski, K. J. Am. Chem. Soc. 2010, 132, 6032–6040.
(16) Lipton, M. F.; Sorensen, C. M.; Sadler, A. C.; Shapiro, R. H.
J. Organomet. Chem. 1980, 186, 155–158.
(17) Krasovskiy, A.; Knochel, P. Synthesis 2006, No. 5, 890–891.
(18) (a) Rijs, N. J.; O’Hair, R. A. J. Organometallics 2010, 29, 2282–
2291. (b) Rijs, N. J.; Yates, B. F.; O’Hair, R. A. J. Chem.—Eur. J. 2010,
16, 2674–2678.
(19) Hamann, C. H.; Vielstich, W. Elektrochemie, 4thed.; Wiley-VCH:
Weinheim, Germany, 2005.
(20) Carvajal, C.; T€olle, K. J.; Smid, J.; Szwarc, M. J. Am. Chem. Soc.
1965, 87, 5548–5553.
(21) Pearson, R. G.; Gregory, C. D. J. Am. Chem. Soc. 1976, 98,
4098–4104.
(22) Lide, O. R., Ed. CRC Handbook of Chemistry and Physics, 88th
ed.; CRC Press: Boca Raton, FL, 2008.
(23) Watanabe, K.; Yamagiwa, N.; Torisawa, Y. Org. Process Res. Dev.
2007, 11, 251–258.
(24) (a) Bhattacharyya, D. N.; Lee, C. L.; Smid, J.; Szwarc, M. J. Phys.
Chem. 1965, 69, 608–611. (b) B€ohm, L. L.; Schulz, G. V. Ber. Bunsen-Ges.
Phys. Chem. 1969, 73, 260–264. (c) Chen, Z.; Hojo, M. J. Phys. Chem. B
1997, 101, 10896–10902.
(25) At higher and, thus, synthetically relevant concentrations, the
fraction of free ions will be lower because of increased ion pairing
according to the law of mass action. This effect can be directly seen in the
concentration dependence of the molar conductivity of LiCuPh2 LiCN;
3
see Figure S33 in the Supporting Information.
(26) Di Marco, V. B.; Bombi, G. G. Mass Spectrom. Rev. 2006, 25,
347–379.
(27) Belyakov, P. A.; Kadentsev, V. I.; Chizhov, A. O.; Kolotyrkina,
N. G.; Shashkov, A. S.; Ananikov, V. P. Mendeleev Commun. 2010, 20,
125–131.
4778
dx.doi.org/10.1021/om200625z |Organometallics 2011, 30, 4771–4778