On the role of PIII ligands in the conjugate addition of diorganozinc derivatives to enones

Article abstract of DOI:10.1002/chem.200400316

A mechanistic study on the conjugate addition of diethylzinc to cyclohexenone catalyzed by various chiral P^III ligands, provides new insights into its mechanism. Complete in situ conversion of the catalytic amount of Cu(OTf)₂ into Cu

Full text of DOI:10.1002/chem.200400316

FULL PAPER
On the Role of PIII Ligands in the Conjugate Addition of Diorganozinc
Derivatives to Enones
Tatjana Pfretzschner,[a] Lutz Kleemann,[b] Birgit Janza,[a] Klaus Harms,[a] and
Thomas Schrader*[a]
Abstract: A mechanistic study on the
conjugate addition of diethylzinc to cy-
clohexenone catalyzed by various
chiral PIII ligands, provides new insights
into its mechanism. Complete in situ
conversion of the catalytic amount of
Cu(OTf)2 into CuI species by excess
ZnEt2 is demonstrated by EPR spec-
troscopy. Experimental evidence is pre-
sented in favor of a critical ternary
1:1:1 complex between enone, Et2Zn
and catalyst, supporting a rate-limiting
reductive elimination or carbocupra-
tion from a preformed mixed Cu/Zn
cluster carrying one- or two-ligand
molecules. A crystal structure has been
obtained for a CuI·L2 complex, which
shows catalytic turnover on addition of
reagent and substrate. NMR spectro-
scopic analyses and VT experiments
reveal that steric hindrance may pre-
vent complexation of a second ligand
molecule on the CuI cation, leading to
extremely fast precatalysts based on
CuX·L species.
Keywords: asymmetric catalysis
·
conjugate addition · dialkylzinc ·
kinetics
Introduction
ly tunable biphenyl-based phosphoramidites,[7] and with
Pfaltzꢀ phosphate–oxazolinones.[8]
The copper-catalyzed conjugate addition of diorganozinc
compounds to enones has become a powerful synthetic tool
for asymmetric catalytic Michael additions.[1] Catalytic
routes to cycloalkanones as well as tandem and annulation
procedures are now available offering almost quantitative
chemical yields and excellent enantioselectivities.[2] Chiral
induction in this valuable process has been achieved mainly
by ligand acceleration with chiral phosphoramidites, phos-
phites and phosphonites.[3] Since the first report by Alexakis
in 1993, many other PIII ligands have been tested in this re-
action during the past decade.[4] These differ substantially
both in their reactivity and their enantioselectivity. System-
atic optimization of Feringaꢀs original catalyst[5] has led to a
ligand with a second chiral structural unit which usually
gives enantiomeric excesses >95% with enones.[6] High op-
tical yields are also achieved with Alexakisꢀ conformational-
However, in spite of the importance of this promising cat-
alytic process, very few mechanistic studies have been un-
dertaken to elucidate its mechanism. In the related area of
the conjugate addition of Gilman cuprates[9] early kinetic
studies by Krauss and Smith[10] and subsequent NMR studies
by Ullenius and Smith[11] clearly pointed to a mechanism,
where an intermediate formed in equilibrium with the start-
ing material goes through an irreversible step to give the
conjugate adduct. Further support came from mechanistic
studies by Krause, who identified such an intermediate, fol-
lowed by elegant investigations on kinetic isotope effects by
Singleton.[12,13] It became clear, that the rate-limiting step is
III
ꢀ
C C bond formation by reductive elimination of a Cu in-
termediate.[14] Recently, Krause could determine the activa-
tion parameters for this process.[15] Finally, a much more de-
tailed picture evolved from extensive density functional
studies by Nakamura on conjugate additions of lithium orga-
nocuprates.[16] Based on these mechanistic insights and own
observations, Feringa[17] and Alexakis[18] postulated a similar
mechanism for the organozinc addition in the past years.
Both assume an initial alkyl transfer from R2Zn to the
copper center followed by coordination of the hard Lewis
acid RZnX to the enone carbonyl and formation of a p-
complex between the soft Cu–alkyl species and the enone
double bond. Oxidative addition of the organocopper(i) re-
agent to the enone is presumably followed by a reductive
[a] Dipl.-Chem. T. Pfretzschner, Dipl.-Chem. B. Janza, Dr. K. Harms,
Prof. Dr. T. Schrader
Fachbereich Chemie, Universität Marburg
Hans-Meerwein-Strasse, 35032 Marburg (Germany)
Fax : (+49)6421-28-289-17
E-mail: schradet@mailer.uni-marburg.de
[b] Dipl.-Chem. L. Kleemann
Institut für Organische Chemie, Universität Wien
Währinger Strasse 38, 1090 Wien (Austria)
ꢀ
elimination with concomitant C C-bond formation, the
latter being the rate-determining step (Scheme 1). Product
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
6048
ꢁ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200400316
Chem. Eur. J. 2004, 10, 6048 – 6057

6048 – 6057
should observe a close correla-
tion between the overall reac-
tion rate and the structure of
the respective PIII ligand, or
more precisely, its electron den-
sity. With the most common
standard reaction, that is, the
conjugate addition of diethyl-
zinc to cyclohexenone, we
therefore carried out a system-
atic investigation on the degree
of ligand acceleration exerted
by various classes of PIII li-
gands.[21] Care was taken to
alter the nucleophilicity of the
phosphorus atom, by starting
from N3P systems and including
more and more electron-with-
drawing oxygen atoms as next
neighbours to phosphorus. In
addition, the steric demands of
the ligands were varied from
slim to bulky. To this end, we
Scheme 1. Postulated catalytic cycle for the copper(i)-catalyzed conjugate addition of organozincs to cyclohex-
enone. b) In the key steps, oxidative addition of the organocopper(i)-species 1b to the enone double bond is
followed by the rate-determining reductive elimination of the catalytic copper(i)-complex 2 with simultaneous
alkyl transfer to the g-carbon of the enone. The resulting enolate anion 3 can be further processed by quench-
ing with appropriate electrophiles.
systematically synthesized
a
series of new phosphorous
amides based on 2-aminome-
thylpyrrolidine 4. The new li-
gands were derived from the
parent compound 5 with a di-
inhibition is circumvented by liberation of the zinc enolate
in its thermodynamically stable dimeric form. Both, Alexa-
kis and Noyori assume that prior to this the CuR moiety
captures another R2Zn molecule to form a highly nucleo-
philic Cu/Zn cluster.[18] Noyori recently discovered that N-
benzylbenzenesulfonamide itself catalyzes the 1,4-addition
of dialkylzinc compounds to a,b-unsaturated ketones with a
Cu:Zn ratio of up to 1:10000.[19] A catalytic cycle was pro-
posed on the basis of a kinetic study and structural analysis
of the zinc–enolate product. For this special case, kinetic iso-
tope effects point to a concerted mechanism in the alkyl
transfer step.[20] However, most mechanistic assumptions on
the Cu-catalyzed conjugate organozinc addition with PIII li-
gands are only scarcely backed by experimental evidence.
We present in this paper a mechanistic study on the conju-
gate addition of diethylzinc to cyclohexenone catalyzed by
various chiral PIII ligands. Experimental evidence is present-
ed in favor of the above-outlined rationale, supporting a
rate-limiting reductive elimination or a carbocupration from
a preformed mixed Cu/Zn cluster carrying one or two ligand
molecules. A crystal structure has been obtained for a
CuI·L2 complex, which shows catalytic turnover on addition
of reagent and substrate. NMR spectroscopic analyses and
VT experiments reveal that steric hindrance may prevent
complexation of a second ligand molecule on the CuI cation,
leading to extremely fast precatalysts based on CuX·L spe-
cies.
methylamino group, which was replaced by a slim alkoxy
group (6a) and a sterically demanding aryloxy group (6b,
Scheme 2).[22] We then carried out kinetic investigations with
these ligands and
a representative BINOL- (7) and
TADDOL-phosphoramidite ligand[23] (8) used by Feringa
and us before.
Scheme 2. Structures of all ligands examined in this comparative kinetic
study.
The test reaction was carried out at ꢀ308C in toluene
with 1 mol% of CuII–triflate and 2 mol% of ligand as cata-
lyst system (Scheme 3). One might argue, that the oxidation
state of Cu(OTf)2 differs from that of the postulated cop-
per(i) species. However, numerous earlier investigations
have demonstrated that both CuI and CuII salts can be used
with equal success.[2] It has been generally argued that the
Kinetic studies: If the above-outlined assumption of a reduc-
tive elimination in the rate-limiting step is correct, one
Chem. Eur. J. 2004, 10, 6048 – 6057
www.chemeurj.org
ꢁ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6049

T. Schrader et al.
FULL PAPER
CuII salt is quantitatively reduced to the CuI form by the
large excess of diethylzinc present in the reaction mixture.[24]
In order to prove this assumption experimentally, we carried
out EPR measurements on the reaction of Cu(OTf)2 with a
25-fold excess of diethylzinc in toluene, which confirmed the
complete conversion of paramagnetic CuII into diamagnetic
CuI within 20 min (Scheme 3). This reduction is effected by
the dialkylzinc reagent and not by the phosphorus ligand.[25]
30% conversion after 3 h.[26] A very slow background reac-
tion has also been reported independently by Alexakis[27]
and Noyori,[19] with drastic accelerations by PIII ligands or
benzenesulfonamides. Likewise, in our experiments only
1 mol% of catalyst greatly increased the reaction rate in all
cases. Most interestingly, reactions with electron-rich PIII li-
gands containing two or three N atoms, proceeded much
more slowly than those with only one N atom. This becomes
especially impressive in our new ligand series based on 2-
aminomethylpyrrolidine. The systematic variation of the
linear substituent on phosphorus demonstrates two effects
which seem to be of general relevance: heteroatom ex-
change of N against O leads to a marked increase in reac-
tion rate (NMe2!OMe), which is again decelerated by in-
creasing steric bulk (OMe!O-a-naphthyl). With NMe2 sub-
stituents, the BINOL- (7) and especially the TADDOL-de-
rived phosphoramidites (8) are the most efficient catalysts
in this series, bringing the reaction to completion within less
than 1 h at ꢀ308C; thus, in spite of their enormous steric
bulkyness both phosphoramidites are much faster than the
phosphorous triamide. Close inspection of early comparative
studies directed by Alexakis, reveal that under identical con-
ditions the simple achiral ligand P(OEt)3 brings the Michael
addition to completion much earlier than P(NMe)3.[27] Since
the reductive elimination step requires electron donation to
the CuIII centre, electron-rich PIII ligands should be counter-
productive, while electron-withdrawing ligands should facili-
tate the process. This is exactly the case. The same argument
is also valid for a potential carbocupration mechanism,
which we cannot exclude a priori.[28] Sterically demanding li-
gands might further hinder the alkyl transfer step, which re-
quires a close contact between Rdꢀ and C-3d+ of the double
bond in the p-complex. On the contrary, electron-donating
ligands which increase the nucleophilicity of the attacking
alkyl Cu species, should greatly accelerate the oxidative ad-
dition step.[29] In the light of Kitamuraꢀs recent 12C/13C iso-
tope effect experiments on the CuOTf–sulfonamide
system,[20] indicating a concerted mechanism, this seems con-
tradictory, because for such a concerted alkylation step,
electron-donating ligands should be most effective. Howev-
er, it has to be kept in mind, that the catalytic effect origi-
nates in one case from a bridging sulfonamide anion on
zinc, whereas in the other case a PIII ligand on copper accel-
erates the reaction; this may lead to entirely different situa-
tions in the key alkyl transfer step.
Scheme 3. Top: Standard reaction protocol for the kinetic investigations.
Bottom: EPR spectra of 1, free Cu(OTf)2 with [Cu(OTf)2]=4.2mm in
toluene; 2, Cu(OTf)2:ligand 8 (1:2); 3, Cu(OTf)2:ligand 8 (1:2) with a 24-
fold excess of ZnEt2 in toluene after 6 min; 4, after 10 min; 5, after
18 min (300 K).
In all cases, we observed a clean conversion of starting
material to product (Figure 1), albeit at very different rates.
The reaction was generally monitored as decreasing enone
concentration, determined by calibrated GC analyses. Start-
ing with 1 min, after ZnEt2 addition was complete, 50 mL ali-
quots were drawn from the reaction mixture, quenched with
water and extracted into Et2O. The organic phases were di-
rectly examined by GC, with cyclohexanone as an internal
standard.
The uncatalyzed background reaction was very slow
(Figure 2), but not negligible, in our hands, leading to only
Kinetic order: With the final reductive elimination being the
rate-limiting step, all reactions should follow first order ki-
netics, since in a preceding rapid equilibrium the substrate,
catalyst and organozinc reagent are all assembled in one p-
complex which goes directly to the products. First order ki-
netics in substrate, organozinc and (bridging) sulfonamide
ligand were already observed by Noyori in a related CuI-sul-
fonamide system and pointed to a bissubstrate–uniproduct
system with the alkyl-transfer step limiting the turnover
rate.[19] In a logarithmic plot, this is also confirmed for all of
the above-examined ligands: first order kinetics are ob-
served, producing k values between 9”10ꢀ5 sꢀ1 and 2”
10ꢀ3 sꢀ1 for the fast TADDOL-based ligand 8 (uncatalyzed
Figure 1. Kinetic curves for the continuous conversion of enone into 3-al-
kylketone catalyzed by phosphorous acid diamide monoester 6b and
Cu(OTf)2. (GC analysis commenced after 1 min, when ZnEt2 addition
was complete. Cyclohexanone was used as the internal standard, added
stoichiometrically to the reaction mixture simultaneously with the
enone).
6050
ꢁ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 6048 – 6057

Conjugate Addition of Diorganozinc Derivatives
6048 – 6057
Figure 2. Kinetics of the conjugate ethylation of cyclohexenone by dial-
kylzinc catalyzed by CuI complexes of various PIII ligands, monitored as
decreasing enone concentration (GC analysis, the first injection was only
carried out after complete addition of diethylzinc and therefore does not
start at 100%).
background reaction 2”10ꢀ5 sꢀ1). These reaction rates are
comparable to Noyoriꢀs sulfonamide system. In agreement
with the postulated mechanism, the reaction becomes faster,
if more substrate and more catalyst are added, and also, if
an excess of diethylzinc is used. First order kinetics were
also established in experiments with varying ZnEt2 or cata-
lyst concentrations, while all other parameters were kept
constant. The reaction rates were determined over the full
time range and in a concentration range between 100 and
500mm ZnEt2 as well as 0.5 and 4 mol% catalyst (Figure 3b
and c). We conclude that the catalyst forms a ternary 1:1:1
complex with one equivalent of diethylzinc and enone, in
which alkyl transfer occurs. Product release regenerates the
catalyst for the next cycle. As a consequence of this mecha-
nism, the turnover rate must be limited by the alkylation
step and not by the product release.
Recent variations in the copper salts indicate the impor-
tance of charge-alternating bridging anions for the mixed
zinc cuprates.[2,18] Since especially aromatic Cu carboxylates
produce highly efficient catalysts, the degree of lipophilicity
seems to play some role, whereas the Lewis acidity is of
minor relevance. This assumption is strongly supported by
the above-mentioned mechanistic studies conducted by
Noyori.[19] Likewise, exchange of the triflate anion for iodide
greatly decelerated the conjugate addition in our system,
pointing to the necessity of bridging the Cu and Zn center
for efficient catalysis.
Figure 3. a) ln[enone] versus t plot with resulting k values demonstrating
first order kinetics for all above-shown reactions (200mm enone, 240mm
diethylzinc, 2mm catalyst, ꢀ308C, toluene). b) Relation between reaction
rate k and [ZnEt2] (200mm enone, 2mm catalyst, ꢀ308C, toluene/hex-
anes). First order kinetics in [ZnEt2] are observed.[30] c) Relation between
reaction rate k and [catalyst
= 1:2 mixture of Cu(OTf)2 and ligand]
(200mm enone, 240mm ZnEt2, ꢀ308C, toluene/hexanes). First order ki-
netics in [catalyst] are observed.[30]
TADDOL-thiolate, which catalyzes the somewhat related
Grignard conjugate addition to enones. However, this com-
plex is tetranuclear, and the thio-TADDOL acts as a mono-
dentate ligand bridging two Cu atoms.[33] Optimal conditions
were often found for a 2:1 ratio for monodentate, but a 1:1
ratio for bidentate phosphoramidite ligands; in connection
with moderate nonlinear effects this led many authors to the
conclusion, that in the stereodiscriminating step two ligand
molecules are operating together.[34] Preliminary experi-
ments with our new ligands confirmed that a 3:1 complex
was catalytically inactive, as opposed to 2:1 mixtures which
greatly accelerated the conjugate addition. Although we
tried to crystallize a 3:1 complex from CuI and the N3P
ligand 5, we obtained a crystal structure clearly showing a
ligand/Cu ratio of 2:1, with a trigonal planar arrangement
around Cu (Figure 4). Addition of a large excess of diethyl-
X-ray crystal structure:[31] Next we turned to the stoichiome-
try of the precatalyst formed between the CuI salt and the
PIII ligand. Feringa and others observed that a 3:1 ratio of
ligand and Cu salt produced a very stable, catalytically inac-
tive complex, presumably because alkyl transfer to the CuI
centre is hindered.[2] The only crystal structure known to
date shows exactly this a 3:1 complex with CuI.[5,32] Recently
Seebach et al. obtained
a crystal structure of a
CuI-
Chem. Eur. J. 2004, 10, 6048 – 6057
www.chemeurj.org
ꢁ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6051

T. Schrader et al.
FULL PAPER
ꢀ
bonds: although no dramatic changes are observed, P N
bond lengths tend to be shortened on complexation (1.75 to
1.70 Š, 1.68 to 1.66 Š), whereas NPN angles are slightly
widened by about three degrees (90 to 928, 102 to 1058, 106
to 1098). With great caution, we take this as a first indication
of electron density flowing from nitrogen via phosphorus to
the CuI center. In the complex, both phosphorous triamide
ligands form a quasi C2-symmetric arrangement with respect
to the CuI axis. The concave faces of the roof-like bicycles
are oriented antiparallel to each other.[35] Similar to the free
TADDOL-based ligand 8,[5] one of the NMe2 methyl groups
stands coplanar to the lone pair on phosphorus. This creates
steric hindrance for the copper complexation.
NMR spectroscopy: We finally carried out an NMR-spectro-
scopic comparative study of all complexes formed between
CuI and the kinetically examined PIII ligands. Interestingly
enough, in CDCl3 a drastic upfield-shift of roughly 20 ppm
is consistently produced in most 2:1-complexes. This result
ꢀ
is counterintuitive, but may be explained by the direct Cu P
n
bond (in all cases, the JP, H couplings broke down).[36] We
were again surprised by the TADDOL ligand 8, which ex-
clusively forms a 1:1 complex with CuI. A VT study revealed
a coalescence temperature for the ligand exchange of
~+248C, corresponding to a fast ligand exchange rate of
349 sꢀ1 with a low activation barrier of ~11.1 kcalmolꢀ1
(Figure 6).[37] The high preference for 1:1 complexes may ex-
plain TADDOLꢀs superior performance with respect to re-
action rate and stereoselectivity, since in the catalytically
active species, steric hindrance around the central copper
atom is minimized.
Figure 4. Crystal stucture of the 2:1 complex between the new N3P phos-
phorous triamide ligand 5 and CuI (from toluene/CH2Cl2). Perspectives
have been chosen comparable to those in Figure 5. Hydrogen atoms have
been omitted for clarity.
Furthermore, the dioxolane CH and CH3 groups turned
magnetically equivalent, with a large downfield shift for the
CH protons and a large upfield shift for the methyl group;
this may indicate a conformational change in the ligand ro-
tating certain groups in or out of the anisotropic shielding of
TADDOLꢀs benzene rings,[38] see Figure 6b.
zinc and cyclohexenone started the conjugate addition and
proved the efficiency of this 2:1 precatalyst. Since we could
also carry out an X-ray structural analysis on the free phos-
phorous triamide ligand (Figure 5), we were able to com-
ꢀ
pare the bond lengths and angles around the critical P N
Conclusion
Additional pieces of evidence
have been collected in support
of a reductive elimination or
carbocupration as the rate-lim-
iting step in the CuI-catalyzed
conjugate addition of organo-
zincs to enones. PIII ligands
firmly bound to CuIII seem to
lower the activation energy bar-
rier for this process, especially
if they carry a high number of
electron-withdrawing substitu-
[39]
ꢀ
ents, that is, P O bonds. 2:1
complexes between ligand and
CuI are predominantly formed,
with TADDOL being an excep-
tion, because it furnishes an ex-
tremely powerful 1:1 complex,
Figure 5. Crystal structure of phosphorous triamide ligand 5.
6052
ꢁ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 6048 – 6057

Conjugate Addition of Diorganozinc Derivatives
6048 – 6057
exclusively. In the future, we will use our very slow N3P
ligand to characterize the reaction intermediates (e.g., the
postulated
p-complex)
spectroscopically
(1H,
31P,
13C NMR).[40] In addition, kinetic isotope effects should fur-
ther substantiate the postulated reductive alkyl transfer as
the rate-limiting step as opposed to a potential carbocupra-
tion. In a cooperation, we also intend to calculate the whole
process on a DFT level, following a similar approach as in
Nakamuraꢀs seminal papers on Gilman cuprates.[16]
Experimental Section
Ligands
(1R,8S)-1-(Dimethylamino)-2-phenyl-2,7-diaza-1-phosphabicy-
clo[3.3.0]octane (5): (S)-2-anilinomethylpyrrolidine (2.00 g, 11.5 mmol)
was dissolved in anhydrous toluene (10 mL) under argon. After the addi-
tion of tris(dimethylamino)phosphine (1.88 g, 2.1 mL, 11.5 mmol) the
mixture was heated to reflux for 4 h. The reaction was monitored by
31P NMR. After cooling to room temperature, the solvent and the excess
of tris(dimethylamino)phosphine were distilled off. The remaining oil
was subjected to fractional distillation in a Kugelrohr apparatus (2.8”
10ꢀ1 mbar, 1808C) to obtain the title compound as a colorless liquid
(2.15 g, 86.2 mmol, 75.0%). The compound crystallized after being stored
at room temperature under argon for several months. M.p. 418C;
1H NMR (200 MHz, CDCl3): d=1.54–1.85 (m, 3H), 1.93–2.05 (m, 1H),
2.60 (d, 3JH,P =8.7 Hz, 6H, C(14)-H), 3.02–3.24 (m, 2H, C(6)-H), 3.37–
2
3
3
3.54 (m, 1H, C(8)-H), 3.77 (ddd, J=1.0, JH,P =9.0, J=7.2 Hz, 1H, C(3)-
H), 4.06–4.19 (m, 1H, C(3)-H), 6.76 (tt, 3J=7.2, 4J=1.0 Hz, 1H, C(12)-
H), 6.84 (m, 2H, C(11)-H), 7.21 (m, 2H, C(10)-H); 31P NMR (81 MHz,
CDCl3): d=119.3 (s); 13C NMR (50 MHz, CDCl3): d=25.4, 32.3, 36.7,
37.1, 50.1, 62.3, 114.4, 114.6, 117.6, 128.9; MS (70 eV): m/z (%): 136 (49)
[N-P-N-C6H5+], 205 (100) [M +ꢀNMe2], 249 (4) [M +]; ESI-HRMS: m/z:
calcd for C13H20N3P: 249.1390, found 249.1390.
Figure 6. 1H and 31P NMR spectra for a) the free phosphorous triamide
ligand 5 used for the above-discussed mechanistic investigations and its
2:1 complex with CuI, and b) the free phosphoramidite ligand 8 and its
1:1 complex with CuI, both in CDCl3.
(1R,8S)-1-Methoxy-2-phenyl-2,7-diaza-1-phosphabicyclo[3.3.0]octane
(6a): (S)-2-anilinomethylpyrrolidine (1.00 g, 5.71 mmol) was dissolved in
anhydrous toluene (10 mL) under argon. After the addition of tris(dime-
thylamino)phosphine (0.99 g, 1.1 mL, 6.1 mmol, 1.1 equiv) the mixture
was heated to reflux for 4 h. The reaction was monitored by 31P NMR.
After cooling to room temperature, the solvent and the excess of tris(di-
methylamino)phosphine were distilled off. The residue was diluted with
anhydrous toluene (10 mL), treated with anhydrous methanol (0.20 g,
0.25 mL, 6.1 mmol, 1.1 equiv) and again heated to reflux. The solvent
and the excess of methanol were distilled off in vacuo. The remaining oil
was subjected to fractional distillation in a Kugelrohr apparatus (2.5”
10ꢀ1 mbar, 1908C) to obtain the title compound as a colorless liquid
1
(0.80 g, 3.4 mmol, 60.0%). H NMR (200 MHz, CDCl3): d=1.54–1.85 (m,
3
3H), 1.95–2.08 (m, 1H), 3.08–3.25 (m, 2H, C(6)-H), 3.33 (d, JH,P
=
8.5 Hz, 3H, C(14)-H), 3.48–3.64 (m, 1H, C(8)-H), 3.75 (ddd, 2J=1.0,
3JH,P =9.0, 3J=7.2 Hz, 1H, C(3)-H), 4.10–4.22 (m, 1H, C(3)-H), 6.83 (tt,
3J=7.2, 4J=1.0 Hz, 1H, C(12)-H), 6.99–7.04 (m, 2H, C(11)-H), 7.23 (m,
2H, C(10)-H); 31P NMR (81 MHz, CDCl3): d=123.5 (s); 13C NMR
(50 MHz, CDCl3): d=26.1, 32.1, 48.3, 49.0, 54.1, 63.3, 114.5, 114.8, 118.8,
129.0; MS (70 eV): m/z (%): 167 (20) [NP(OMe)N-Ph], 205 (35)
Figure 7. Linear dependence of lnk for the TADDOL-based ligand 8 ex-
change on 1/T determined by lineshape analysis of the corresponding
PNMe2 1H NMR signals produced in VT experiments between 233 and
315 K.
Chem. Eur. J. 2004, 10, 6048 – 6057
www.chemeurj.org
ꢁ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6053

T. Schrader et al.
FULL PAPER
[M +ꢀOMe], 236 (74) [M +]; ESI-HRMS: m/z: calcd for C12H17N2OP:
1.6 mmol, 92%). GC analyses (tR =14.5 min) did not show any unwanted
side products. 1H NMR (200 MHz, CDCl3): d=0.91 (t, 3J=8.5 Hz, 3H,
C(8)-H), 1.31–1.42 (m, 3H), 1.58–1.72 (m, 2H) 1.89–1.96 (m, 1H, C(3)-
H), 2.01–2.08 (m, 2H), 2.14–2.38 (m, 3H); 13C NMR (50 MHz, CDCl3):
d=11.0 (C(8)), 25.2 (C(7)), 29.2, 30.8, 40.6 (C(3)), 41.4, 47.7, 211.9
(C(1)).
236.1072, found 236.1072.
(1R,8S)-1-(a-Naphthoxy)-2-phenyl-2,7-diaza-1-phosphabicyclo[3.3.0]oc-
tane (6b): (S)-2-anilinomethylpyrrolidine (1.00 g, 5.71 mmol) was dis-
solved in anhydrous toluene (10 mL) under argon. After the addition of
NMR spectroscopic characterization of 2:1 and 1:1 ligand complexes
with CuI
Free ligand 5: 1H NMR (200 MHz, CDCl3): d=1.54–1.85 (m, 3H), 1.93–
2.05 (m, 1H), 2.60 (d, 3JH,P =8.7 Hz, 6H), 3.02–3.24 (m, 2H), 3.37–3.54
(m, 1H), 3.77 (ddd, J=1.0, 3JH,P =9.0, J=7.2 Hz, 1H, C(3)-H), 4.06–4.19
(m, 1H), 6.76 (tt, J=7.2, J=1.0 Hz, 1H), 6.84 (m, 2H), 7.21 (m, 2H);
31P NMR (81 MHz, CDCl3): d=119.3 (s).
CuI-ligand 5 (2:1, dissolved crystals): 1H NMR (200 MHz, CDCl3): d=
1.40–1.51 (m, 1H), 1.72–1.97 (m, 3H), 2.55 (s, 6H), 2.98–3.10 (m, 1H),
3.44 (t, J=8.6 Hz, 1H), 3.80–3.87 (m, 2H), 6.52 (d, J=6 Hz, 2H), 6.74 (t,
J=16 Hz, 1H), 7.10 (t, J=16 Hz, 2H); 31P NMR (81 MHz, CDCl3): d=
99.6 (brs); Dd(31P NMR)=20.3 ppm.
tris(dimethylamino)phosphine (0.99 g, 1.1 mL, 6.1 mmol, 1.1 equiv) the
mixture was heated to reflux for 4 h. The reaction was monitored by
31P NMR. After cooling to room temperature, the solvent and the excess
of tris(dimethylamino)phosphine were distilled off. The residue was dilut-
ed with anhydrous toluene (10 mL), treated with a-naphthol (0.88 g,
6.1 mmol, 1.1 equiv) and again heated to reflux. The solvent was distilled
off in vacuo. The remaining gum-like residue was subjected to fractional
distillation in a Kugelrohr apparatus (1.9”10ꢀ1 mbar, 2508C) to obtain
the title compound as a yellow gum (0.85 g, 2.4 mmol, 42.0%). 1H and
31P NMR spectra indicated a small impurity of (1R,8S)-1-(dimethyla-
mino)-2-phenyl-2,7-diaza-1-phosphabicyclo[3.3.0]octane (2) as well as a-
1
Free ligand 6a: H NMR (200 MHz, CDCl3): d=1.54–1.85 (m, 3H), 1.95–
2.08 (m, 1H), 3.08–3.25 (m, 2H), 3.33 (d, 3JH,P =8.5 Hz, 3H), 3.48–3.64
3
(m, 1H), 3.75 (ddd, J=1.0, JH,P =9.0, J=7.2 Hz, 1H), 4.10–4.22 (m, 1H),
6.83 (tt, J=7.2, J=1.0 Hz, 1H), 6.99–7.04 (m, 2H), 7.23 (m, 2H);
31P NMR (81 MHz, CDCl3): d=123.5 (s).
1
naphthol, which could not be completely eliminated. H NMR (200 MHz,
CDCl3): d=1.35–1.47 (m, 1H), 1.61–1.83 (m, 3H), 3.00–3.20 (m, 2H),
3.41–3.76 (m, 3H), 6.86–6.95 (m, 2H), 7.11 (m, 2H), 7.20–7.46 (m, 7H),
7.71 (m, 1H); 31P NMR (81 MHz, CDCl3): d=125.4 (s); 13C NMR
(50 MHz, CDCl3): d=26.4, 31.8, 47.3, 53.8, 62.7, 115.3, 115.6, 115.8, 119.6,
122.8, 123.1, 125.1, 125.8, 127.3, 128.4, 129.2, 145.2, 150.4.
CuI-ligand 6a (4.79 mg, 0.025 mmol, 1 equivCuI; 11.81 mg, 0.05 mmol,
2 equiv 6a): 1H NMR (200 MHz, CDCl3): d=1.52–1.75 (m, 2H), 1.76–
2.04 (m, 3H), 3.15 (t, J=8.7 Hz, 1H), 3.37–3.50 (brs, 3H), 3.62–3.86 (m,
2H), 3.87–3.99 (m, 1H), 6.84 (t, J=14 Hz, 1H), 7.02–7.27 (m, 4H);
31P NMR (81 MHz, CDCl3): d=104 (brs); Dd(31P NMR)=19.5 ppm.
(1R,7R)-4-N,N-Dimethylamino-9,9-dimethyl-2,2,6,6-tetraphenyl-3,5,8,10-
Free ligand 6b: 1H NMR (200 MHz, CDCl3): d=1.35–1.47 (m, 1H),
1.61–1.83 (m, 3H), 3.00–3.20 (m, 2H), 3.41–3.76 (m, 3H), 6.86–6.95 (m,
2H), 7.11 (m, 2H), 7.20–7.46 (m, 7H), 7.71 (m, 1H); 31P NMR (81 MHz,
CDCl3): d=125.4 (s).
tetraoxo-4-phosphabicyclo[5.3.0]decane
(8):
TADDOL
(10.0 g,
21.4 mmol) was dissolved under argon in dry chloroform (50.0 mL) and
HMPT (4.2 g, 4.7 mL, 25.7 mmol) was added dropwise with a syringe.
The mixture was heated to reflux for 12–16 h. After 2 h the product
begins to precipitate. At the end of the reaction the mixture is cooled to
room temperature and the product is filtered off, washed with dry chloro-
form and dried in vacuo to furnish pure phosphoramidite (8.5 g, 73%). A
second crop (2.0 g, 18%) of product was obtained after partial removal
of the solvent. The product can be recrystallized from chloroform or di-
chloromethane. 1H NMR (200 MHz, CDCl3): d=0.29 (s, 3H), 1.26 (s,
3H), 2.72 (d, 3JH,P =10.5 Hz, 6H), 4.82 (d, J=8.5 Hz, 1H), 5.18 (dd, J=
3.2, J=8.5 Hz, 1H), 7.16–7.33 (m, 12H), 7.41 (d, J=7.7 Hz, 2H), 7.47 (d,
J=7.5 Hz, 2H), 7.59 (d, J=7.7 Hz, 2H), 7.73 (d, J=7.5 Hz, 2H);
13C NMR (75 MHz, CDCl3): d=25.29 (q), 27.51 (q), 35.18 (q), 35.44 (q),
81.17 (d, JP,C =7.3 Hz), 81.81 (s), 82.35 (d, JP, C =23.2 Hz), 82.45 (d), 111.71
(s), 127.09 (d), 127.25 (d), 127.44 (d), 127.65 (d), 128.07 (d), 128.67 (d),
128.74 (d), 128.95 (d), 141.77 (s), 142.09 (s), 146.43 (s), 146.85 (s);
31P NMR (81 MHz, CDCl3): d=139.9 (s).
CuI-ligand 6b (3.81 mg, 0.02 mmol, 1 equivCuI; 13.95 mg, 0.04 mmol,
2 equiv 6b): 1H NMR (200 MHz, CDCl3): d=1.18–1.80 (m, 4H), 2.70–
2.91 (m, 2H), 2.92–3.07 (dd, J=7.8 Hz, 1H), 3.27–3.39 (dd, J=7.8 Hz,
1H), 4.00–4.19 (brs, 1H), 6.74–6.88 (m, 2H), 7.01–7.61 (m, 9H), 7.71–
7.78 (m, 1H), 8.22–8.33 (m, 1H) ; 31P NMR (81 MHz, CDCl3): d=96.5
(brs); Dd(31P NMR)=29.1 ppm.
3
Free ligand 7: 1H NMR (200 MHz, CDCl3): d=2.54 (d, JH,P =9 Hz, 6H),
7.18–7.51 (m, 7H), 7.87–7.98 (m, 4H); 31P NMR (81 MHz, CDCl3): d=
149 (s).
CuI-ligand 7 (3.05 mg, 0.016 mmol, 1 equivCuI; 11.46 mg, 0.032 mmol,
1
2 equiv 7): H NMR (200 MHz, CDCl3): d=2.26 (brs, 6H), 7.16–7.46 (m,
7H), 7.73–7.90 (m, 5H); 31P NMR (81 MHz, CDCl3): d=128 (brs);
Dd(31P NMR)=21 ppm.
1
Free ligand 8: H NMR (200 MHz, CDCl3): d=0.29 (s, 3H), 1.26 (s, 3H),
3
2.72 (d, JH,P =10.5 Hz, 6H), 4.82 (d, J=8.5 Hz, 1H), 5.18 (dd, J=3.2, J=
General procedure for the conjugate addition of diethylzinc to a,b-unsa-
turated carbonyl compounds: copper(ii)-bis(trifluoromethylsulfonate)
(6.0 mg, 1.7 mmol, 1 mol%) were stirred with two equivalents of the re-
spective ligand (3.4 mmol, 2 mol%) or its solution in toluene at room
temperature for 2 h. The solution was cooled to ꢀ308C and treated with
the respective enone (1.7 mmol) or its solution. After stirring for 30 min,
8.5 Hz, 1H), 7.16–7.33 (m, 12H), 7.41 (d, J=7.7 Hz, 2H), 7.47 (d, J=
7.5 Hz, 2H), 7.59 (d, J=7.7 Hz, 2H), 7.73 (d, J=7.5 Hz, 2H); 31P NMR
(81 MHz, CDCl3): d=139.9 (s).
CuI-ligand 8 (3.77 mg, 0.02 mmol, 1 equivCuI; 10.95 mg, 0,02 mmol,
1 equiv 8): 1H NMR (200 MHz, CDCl3): d=0.45–0.71 (brd, J=32.0 Hz,
6H), 2.49 (brd, J=11.3 Hz, 6H), 5.28 (brs, 2H), 7.17–7.36 (m, 11H),
7.37–7.46 (t, J=7.5 Hz, 4H), 7.50–7.66 (m, 6H); 31P NMR (81 MHz,
CDCl3): d=102.6 (s); Dd(31P NMR)=37.3 ppm.
a
diethylzinc 15% solution in hexanes (0.25 g, 2.2 mL, 2.2 mmol,
1.1 equiv) was added dropwise within 5 min, so that the reaction temper-
ature did not rise above ꢀ308C. After 16 h the solution was warmed to
08C and treated with 1n aqueous HCl. The resulting product was extract-
ed with diethyl ether; then the combined organic phases were washed
with satd aqueous NaHCO3 and subsequently with aqueous NaCl and
dried over MgSO4. After filtration, the solvent was removed in vacuo
and the crude product was purified by chromatography.
Note: The oxygen and water sensitivity of the ligands required perform-
ance of their weighing procedure and subsequent transfer into NMR
tubes inside a glove box.
Gas chromatography: Gaschromatographic analyses were carried out
with a Hewlett-Packard gaschromatograph (HP 5890 Series II) equipped
with FID and a Hewlett-Packard Integrator (HP 3396 Series II). The sta-
tionary phase was contained in a Supelco “g-Dex 120” column (30 m”
0.25 mm”0.25 mm). The same temperature program was used repeatedly
(start T = 808C (5 min), rate: 48Cminꢀ1, final T = 1408C (5 min) (nitro-
gen carrier gas).
3-Ethylcyclohexanone: The reaction was carried out with 2-cyclohexe-
none (0.16 mL, 0.16 g, 1.7 mmol). The reaction went to completion within
16 h with all ligands. The product was purified over silica gel eluting with
n-hexane/ethyl acetate 4:1 (Rf =0.40) to yield the title compound (0.24 g,
Kinetics of the 1,4-addition of diethylzinc to 2-cyclohexenone
a) Dependence on [enone]: 1m reference solutions were prepared from
2-cyclohexenone and cyclohexanone as internal GC standard in dry tolu-
6054
ꢁ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 6048 – 6057

Conjugate Addition of Diorganozinc Derivatives
6048 – 6057
ene. TADDOL-phosphoramidite 8 and BINOL-phosphoramidite 7 were
used as solids, whereas 0.1m solutions were prepared from the other li-
gands. Copper(ii)-bis(trifluoromethylsulfonate) (7.2 mg, 0.02 mmol) was
stirred with the ligand (0.04 mmol) or its solution in a total of 2.4 mL tol-
uene for 16 h under argon at ambient temperature. Subsequent cooling
to ꢀ308C was followed by addition of a 1m solution of 2-cyclohexenone
(2 mL, 0.19 g, 2 mmol) and a 1m solution of cyclohexanone (2 mL, 0.20 g,
2 mmol). After 10 min, a 1m solution of diethylzinc in hexanes (2.4 mL,
0.30 g, 2.4 mmol, 1.2 equiv), which was held at a constant temperature of
08C, was added within 1 min. As soon as the addition was complete, a
50 mL sample was taken from the reaction mixture and treated with
water (1 mL) and diethyl ether (0.5 mL). Additional aliquots were taken
after 2, 4, 6, 8, 10, 15, 20, 30, 45, 60, 80, 100, 120, 150 and 180 min and
treated likewise. The ethereal phases were examined by GC. The reac-
tion was carried on for another 13 h and finally quenched with 1n aque-
ous HCl. Workup proceeded in the same manner as described above.
The enantiomeric excess was determined by 13C NMR spectroscopy after
derivatization with (R,R)-1,2-Diamino-1,2-diphenyl-ethane.
Graphical representation as stacked plots: Stacked plots (aliphatic
region) for the collected 1H NMR data (400 MHz, dry CDCl3) from the
above-described VT experiment of a 2:1-mixture between TADDOL-
PNMe2 (8) and CuI (Figure 8). The black frame indicates the P-NMe2
signal used for the line-shape analysis.
b) Dependence on [ZnEt2]: 1m reference solutions were prepared from
2-cyclohexenone and cyclohexanone as internal GC standard in dry tolu-
ene. Copper(ii)-bis(trifluoromethylsulfonate) (7.2 mg, 0.02 mmol) was
stirred with ligand 5 (0.4 mL, 0.04 mmol in 0.1m solution in toluene) in a
total of 1.6 mL–2.4 mL dry toluene and 0.0 mL–2.4 mL dry n-hexane for
1 h under argon at ambient temperature. Subsequent cooling to ꢀ308C
was followed by addition of a 1m solution of 2-cyclohexenone (2 mL,
Figure 8. Lineshape analysis.
0.19 g, 2 mmol) and
a 1m solution of cyclohexanone (2 mL, 0.20 g,
2 mmol). After 30 min, 0.8 mL–5.0 mL of a 1m solution of diethylzinc in
hexanes (0.05 g, 2.4 mmol, 0.4 equivto 0.625 g, 5.0 mmol, 2.5 equi)v,
which was held at a constant temperature of 08C, was added within
1 min (total capacity: 9.6 mL; total capacity for 2.5 equivdiethylzinc:
10.6 mL). As soon as the addition was completed, a 50 mL sample was
taken from the reaction mixture and treated with 1n aqueous HCl
(1 mL) and diethyl ether (0.5 mL). Additional aliquots were taken after
1, 2, 3, 4, 5, 6, (7), 8, (9), 10, (11), 12, (13, 14), 15, 20, (25), 30, (40, 50),
60, 120 min and treated likewise. The ethereal phases were examined by
GC.
Evaluation: Since the equilibrium is degenerated, it can be described by
Equation (1):
L þ CuI Á L0 Ð CuI Á L þ L0
ð1Þ
with L=L’. For the calculation of rate constants, the usual approxima-
tions were applied, depending on the time regime between fast and very
slow exchange. A very well resolved signal is found in the methyl group
of the P-NMe2 moiety, which does not overlap with any other signal. Its
linewidth was measured at half peak height at various temperatures and
used for the calculation of k values from the respective approximation.
Thus, the following values were obtained:
c) Dependence on [catalyst]: 1m reference solutions were prepared from
2-cyclohexenone und cyclohexanone as internal GC standard in dry tolu-
ene. 3.6 mg–18.0 mg copper(ii)-bis(trifluoromethylsulfonate) (0.01 mmol–
0.05 mmol) were stirred with 0.2 mL–1.0 mL (0.02 mmol–0.1 mmol in
0.1m solution in toluene) of ligand 5 in a total of 2.4 mL dry toluene and
0.8 mL dry n-hexane for 1 h under argon at ambient temperature. Subse-
quent cooling to ꢀ308C was followed by addition of a 1m solution of 2-
cyclohexenone (2 mL, 0.19 g, 2 mmol) and a 1m solution of cyclohexa-
none (2 mL, 0.20 g, 2 mmol). After 30 min, a 1m solution of diethylzinc
in hexanes (2.4 mL, 0.30 g, 2.4 mmol, 1.2 equiv), which was held at a con-
stant temperature of 08C, was added within 1 min (total capacity:
9.6 mL). As soon as the addition was complete, a 50 mL sample was
taken from the reaction mixture and treated with 1n aqueous HCl
(1 mL) and diethyl ether (0.5 mL). Additional aliquots were taken after
1, 2, 3, 4, 5, 6, (7), 8, (9), 10, (11), 12, (13, 14), 15, 20, (25), 30, (40, 50),
60, 120 min and treated likewise. The ethereal phases were examined by
GC.
a) Fast exchange (>10–158C above TC) [Eq. (2)]:
k ¼ pDno2=2ðheꢀhoÞ
ð2Þ
with Dno =peak separations [Hz] for spectra without exchange effects,
and ho =Full Width at Half-Height (FWHH, in Hz) for peaks showing no
exchange effects, as well as he =FWHH for peaks widened from exchange
effects.
At 315 K, Dno =157 Hz, he =25, and ho =12 Hz. From this, k=2978 sꢀ1
.
b) Coalescence temperature TC [Eq. (3)]:
1
k ¼ pDno=2=
2
ð3Þ
with Dno =peak separations [Hz] for spectra without exchange effects.
Note: With further increase of the ligand complex concentration the
complex was not soluble any more in the available solvent quantity.
Therefore a kinetic investigation was not possible beyond this point.
At 297 K, Dno =157 Hz. From this, k=349 sꢀ1
.
c) Intermediate exchange (below TC and
>
20% peak overlap)
Lineshape analysis and calculation of TC and DG°: A 2:1 mixture was
prepared from the TADDOL-phosphoramidite 8 and CuI, and dissolved
in dry CDCl3, as described above. 1H NMR spectra as well as 31P NMR
spectra were recorded in temperature intervals of 88C, starting from
233 K and ending at 315 K. While no change was observed within this
temperature range in the 31P NMR spectra, the 1H NMR spectra showed
dramatic changes corresponding to a dynamic exchange between the 1:1
complex 8·CuI and the free ligand itself. At elevated temperatures, fast
exchange leads to a simplified signal pattern with averaged shift values.
At 297 K, coalescence is reached, demonstrated by extremely broad
lines. Below 297 K, further cooling leads to the slow and very slow ex-
change regime, with the appearance of two sets of signals, one for the
free ligand and the other one for the 1:1 complex. At 233 K, lines have
become very sharp and correspond exactly to the superimposed NMR
spectra of isolated 1:1 complex and free ligand.
[Eq. (4)]:
1
1
2
k ¼ pðDno2ꢀDne2Þ 2 =2=
=
ð4Þ
with Dno and Dne =peak separations [Hz] for spectra without and with
exchange effects, respectively.
At 281 K, Dno =157 Hz, and Dne =96 Hz. From this, k=276 sꢀ1
.
d) Slow exchange (well resolved peaks with less than ~20% overlap)
[Eq. (5)]:
k ¼ p ðheꢀhoÞ
ð5Þ
At 273 K, he =38 Hz, and ho =12 Hz. From this, k=82 sꢀ1
At 265 K, he =29 Hz, and ho =12 Hz. From this, k=53 sꢀ1
.
.
Chem. Eur. J. 2004, 10, 6048 – 6057
www.chemeurj.org
ꢁ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6055

T. Schrader et al.
FULL PAPER
At 249 K, he =19 Hz, and ho =12 Hz. From this, k=22 sꢀ1
.
baseline, indicating complete conversion of the starting ma-
terial.
Since Equation (6):
lnk ¼ ꢀEa=RT þ lnA
ð6Þ
a plot of lnk v s 1T/ (see Figures 7 and 9 and Table 1) yields ꢀEa/R as its
slope (R=8.314 Jmolꢀ1 Kꢀ1). From this, Ea was calculated as 11.1 kcal
molꢀ1. Thus, the ligand exchange rate at TC is 349 sꢀ1 and the activation
Acknowlegments
We are indebted to the group of Prof. Sundermeyer for the reactions car-
ried out in their glovebox. We thank Dr. B. Neumüller, Fachbereich
Chemie, Universität Marburg, for very helpful discussions, especially on
the nature of the involved organocopper species.
energy barrier for this process amounts to DG° =46.4 or 11.1 kcalmolꢀ1
.
[1] N. Krause, A. Hoffmann-Rçder, Synthesis 2001, 171–196; B. Ferin-
ga, in Modern Organocopper Chemistry, (Ed.: N. Krause), Wiley-
VCH, Weinheim, 2002, pp. 224–258.
[2] Recent reviews with a focus on ligand structures: B. L. Feringa, Acc.
Chem. Res. 2000, 33, 346–353; A. Alexakis, C. Benhaim, Eur. J.
Org. Chem. 2002, 3221–3236.
[3] Ligand acceleration: D. J. Berrisford, C. Bolm, K. B. Sharpless,
Angew. Chem. 1995, 107, 1159–1171; Angew. Chem. Int. Ed. Engl.
1995, 34, 1059–1070; recent reviews about related PIII ligands: a) J.
Ansell, M. Wills, Chem. Soc. Rev. 2002, 31, 259–268; b) G. Chelucci,
G. Orrß, G. A. Pinna, Tetrahedron 2003, 59, 9471–9515; new mixed
ligands: G. Delapierre, J. M Brunel, T. Constanstieux, G. Buono,
Tetrahedron: Asymmetry 2001, 12, 1345–1352; I. J. Krauss, J. L.
Leighton, Org. Lett. 2003, 5, 3201–3203.
Figure 9. k versus T plot for the ligand exchange rate between CuI·8 and
free 8.
[4] A. Alexakis, J. Frutos, P. Mangeney, Tetrahedron: Asymmetry 1993,
4, 2427–2430; the same ligands were used earlier for the addition of
organocopper reagents to enones: A. Alexakis, S. Mutti, J. F. Norm-
ant, J. Am. Chem. Soc. 1991, 113, 6332–6334.
Table 1. Temperature dependence of rate constants for the exchange re-
action between CuI·8 and free 8.
T [K]
1/T
k [sꢀ1
]
lnk
315
297
281
273
265
249
0.00317
0.00336
0.00356
0.00366
0.00377
0.00402
2978
349
276
82
53
22
8.00
5.85
5.62
4.40
3.97
3.09
[5] A. H. M. de Vries, A. Meetsma, B. L. Feringa, Angew. Chem. 1996,
108, 2526–2528; Angew. Chem. Int. Ed. Engl. 1996, 35, 2374–2376.
[6] B. L. Feringa, M. Pineschi, L. A. Arnold, R. Imbos, A. H. M. de V-
ries, Angew. Chem. 1997, 109, 2733–2736; Angew. Chem. Int. Ed.
1997, 36, 2620–2623; a polymer-bound version: A. Mandolini, M.
Calamante, B. L. Feringa, P. Salvadori, Tetrahedron: Asymmetry
2003, 14, 3647–3650.
[7] A. Alexakis, S. Rosset, J. Allamand, S. March, F. Guillen, C. Ben-
haim, Synlett 2001, 1375.
[8] A. K. H. Knçbel, I. H. Escher, A. Pfaltz, Synlett 1997, 1429–1431;
Synlett (special issue) 1999, 835–842; I. H. Escher, A. Pfaltz, Tetra-
hedron 2000, 56, 2879–2888.
[9] For a recent review, see: S. Woodward, Chem. Soc. Rev. 2000, 29,
393–401; E. Nakamura, S. Mori, Angew. Chem. 2000, 112, 3902–
3924; Angew. Chem. Int. Ed. 2000, 39, 3750–3771.
[10] S. R. Krauss, S. G. Smith, J. Am. Chem. Soc. 1981, 103, 141–148.
[11] C. Ullenius, B. Christensen, Pure Appl. Chem. 1988, 60, 57–64; A. S.
Vellekoop, R. A. J. Smith, J. Am. Chem. Soc. 1994, 116, 2902–2913.
[12] a) J. P. Snyder, J. Am. Chem. Soc. 1995, 117, 11025–11026; b) D. E.
Frantz, D. A. Singleton, J. P. Snyder, J. Am. Chem. Soc. 1997, 119,
3383–3384.
EPR measurements: Cu(OTf)2 (14.4 mg, 0.042 mmol) was weighed in a
dry flask under argon (glovebox), treated with dry toluene (9.6 mL) and
stirred vigorously for 2 h. An aliquot of 1.0 mL (corresponding to
0.0042 mmol Cu(OTf)2, 1 equiv) of the resulting suspension was transfer-
red into a 5 mm EPR tube and an EPR spectrum was taken at ambient
temperature (ESP300E, Bruker, 9.2027 GHz, 300 K). In a second experi-
ment, TADDOL-phosphoramidite 8 (43.2 mg, 0.08 mmol) was added
under argon to Cu(OTf)2 (14.4 mg, 0.04 mmol) (glovebox), and dissolved
in dry toluene (9.6 mL) by intense stirring for 2 h. An EPR spectrum was
taken from a 1.0 mL aliquot (corresponding to 0.0042 mmol Cu(OTf)2,
1 equiv) of this (clouded) solution as well. Subsequent addition of ZnEt2
(0.05 mL, 1.0m in hexane, corresponding to 0.050 mmol, 25 equiv) to this
aliquot in the EPR tube initiated the reduction of CuII to CuI. After 6, 10
and 18 min, EPR spectra were taken directly from this mixture under
argon at 300 K (9.2332 GHz).
[13] N. Krause, R. Wagner, A. Gerold, J. Am. Chem. Soc. 1994, 116,
381–382; N. Krause, A. Gerold, Angew. Chem. 1997, 109, 194–213;
Angew. Chem. Int. Ed. 1997, 36, 186–204.
[14] Calculations predict the existence of CuIII intermediates in addition
and substitution reactions of dialkylcuprates: A. E. Dorigo, J.
Wanner, P. von RaguØ-Schleyer, Angew. Chem. 1995, 107, 492–494;
Angew. Chem. Int. Ed. Engl 1995, 34, 476–478.
Results
[15] J. Canisius, A. Gerold, N. Krause, Angew. Chem. 1999, 111, 1727–
1730; Angew. Chem. Int. Ed. 1999, 38, 1644–1646.
[16] E. Nakamura, S. Mori, K. Morokuma, J. Am. Chem. Soc. 1997, 119,
4900–4910; S. Mori, E. Nakamura, Chem. Eur. J. 1999, 5, 1534–
1543; E. Nakamura, M. Yamanaka, J. Am. Chem. Soc. 1999, 121,
8941–8942.
[17] L. A. Arnold, R. Imbos, A. Mandoli, A. H. M. de Vries, R. Naasz,
B. L. Feringa, Tetrahedron 2000, 56, 2865–2878.
[18] A. Alexakis, C. Benhaim, S. Rosset, M. Humam, J. Am. Chem. Soc.
2002, 124, 5262–5263.
The EPR spectrum of free suspended Cu(OTf)2 (Scheme 3)
shows the typical solid phase spectrum of a paramagnetic
CuII species with g1 =2.520, g2 =2.098 and g3 =2.098. This
spectrum markedly changes on addition of the phosphor-
amidite, indicating formation of the 2:1 CuII–phosphorami-
dite complex. After addition of excess diethylzinc immediate
alkyl transfer concomitant with reduction of paramagnetic
CuII to diamagnetic CuI takes place. Even without stirring,
the EPR spectrum displays after 20 min only one straight
[19] M. Kitamura, T. Miki, K. Nakano, R. Noyori, Bull. Chem. Soc. Jpn.
2000, 73, 999–1014.
6056
ꢁ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 6048 – 6057

Conjugate Addition of Diorganozinc Derivatives
6048 – 6057
[20] K. Nakano, Y. Bessho, M. Kitamura, Chem. Lett. 2003, 32, 224–225.
[21] N. Krause, Angew. Chem. 1998, 110, 295–297; Angew. Chem. Int.
Ed. 1998, 37, 283–285.
[22] T. Constantieux, J. M. Brunel, A. Labande, G. Buono, Synlett 1998,
49–50; G. Delapierre, T. Constantieux, J. M. Brunel, G. Buono, Eur.
J. Org. Chem. 2000, 2507–2511.
[31] CCDC-225164 (ligand) and -225165 (complex) contain the supple-
mentary crystallographic data for this paper. These data can be ob-
tained free of charge via www.ccdc.cam.ac.uk/conts/retrieving.html,
or from the Cambridge Crystallographic Data Centre, 12 Union
Road, Cambridge CB2 1EZ, UK; fax: (+44)1223–336033; or email:
deposit@ccdc.cam.ac.uk.
[23] TADDOL-based ligands in the Et2Zn addition to enones: a) E.
Keller, J. Maurer, R. Naasz, T. Schrader, A. Meetsma, B. Feringa,
Tetrahedron: Asymmetry 1998, 9, 2409–2413; b) A. Alexakis, J.
Burton, J. Vastra, C. Benhaim, X. Fournioux, A. van den Heuvel,
J.-M. LevÞque, F. MazØ, S. Rosset, Eur. J. Org. Chem. 2000, 4011–
4027;
[32] Recently, Feringa et al. published another crystal structure for a
(catalytically inactive) Rh–(phosphoramidite-7)4·BF4 complex. It
was used in support of a postulated mutual restriction of the ligandsꢀ
conformational flexibility producing an asymmetric coordination
sphere around the rhodium center: see ref. [29].
[33] A. Pichota, P. S. Pregosin, M. Valentini, M. Wçrle, D. Seebach,
Angew. Chem. 2000, 112, 157–160; Angew. Chem. Int. Ed. 2000, 39,
153–156.
[24] No R2Cu species with two alkyl ligands on a CuII atom is known:
D. K. Breitinger, W. A. Herrmann, in Synthetic Methods of Organo-
metallic and Inorganic Chemistry, Vol. 5 (Eds.: W. A. Herrmann, G.
Brauer), Thieme, Stuttgart, 1999.
[34] L. A. Arnold, R. Imbos, A. Mandolini, A. H. M. de Vries, R. Naasz,
B. L. Feringa, Tetrahedron 2000, 56, 2865–2878.
[25] It should be mentioned, that during this reduction an additional
equivalent of Lewis-acidic EtZnOTf is produced which might not be
an innocent spectator. Thus, several groups report faster reactions if
Cu(OTf)2 is used instead of CuOTf.
[26] The quantitative determination of rate constants for the uncatalyzed
background reaction is problematic, because an unspecific catalysis
by copper traces in glassware or reagents cannot a priori be exclud-
ed.
[35] This arrangement may offer an explanation for the disappointing
stereodiscrimination achieved with this new ligand (10–20% ee): the
chiral information is not transferred effectively into the coordination
sphere around the copper center. By introduction of strategically
placed substituents we are currently trying to improve the perform-
ance of this ligand class via rational design.
[36] Similar, but smaller upfield shifts have recently been observed for
CuI complexes with imino-thiophosphoramide ligands: M. Shi, W.
Zhang, Tetrahedron: Asymmetry 2004, 15, 167–176.
[27] A. Alexakis, J. Vastra, P. Mangeney, Tetrahedron Lett. 1997, 38,
7745–7748.
[37] For details, see Supporting Information.
[28] Unfortunately, it is experimentally difficult to detect the transient
postulated CuIII species which is, e.g., diamagnetic and thus invisible
in the EPR spectrum. To distinguish between these two alternatives,
kinetic isotope effects have to be determined for the PIII-catalyzed
reaction.
[29] Monodentate phosphoramidites were also found to dramatically ac-
celerate the rhodium(i)-catalyzed conjugate addition of aromatic
boronic acids to enones, if compared with the classical electron-
richer bidentate BINAP bisphosphane ligand: J. G. Boiteau, A. J.
Minnaard, B. L. Feringa, J. Org. Chem. 2003, 68, 9481–9484.
[30] Rigorous exclusion of humidity and oxygen by carrying out all steps
in a glove-box led to another considerable increase of the overall re-
action rate by ~one order of magnitude.
[38] For a comprehensive review about TADDOL chemistry, see: D.
Seebach, A. K. Beck, A. Heckel, Angew. Chem. 2001, 113, 96–142;
Angew. Chem. Int. Ed. 2001, 40, 92–138.
[39] Nakamura et al. recently also suggested this to be the role of chiral
PIII ligands in the cuprate adddition to enones: see ref. [11].
[40] In one case, a drastic 31P NMR upfield shift has been observed for a
putative Et–CuI species after addition of Et2Zn to a 2:1 mixture of
ligand and Cu(OTf)2: A. S. C. Chan, Chem. Commun. 1999, 11–12.
Received: March 30, 2004
Revised: July 30, 2004
Published online: October 28, 2004
Chem. Eur. J. 2004, 10, 6048 – 6057
www.chemeurj.org
ꢁ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6057