A synthesis detour to planar-diastereoisomeric ferrocene derivatives around an unexpected rearrangement of ortho-lithiated Kagan's template [S(S)]-(p-tolylsulfinyl)ferrocene

Article abstract of DOI:10.1002/hlca.200790082

Usually, ortho lithiation of Kagan's template 1 and quenching with electrophiles leads highly diastereoselectively to planar-chiral 1,2-disubstituted ferrocenes. Surprisingly, lithiation of 1 with lithium diisopropylamide (LDA) followed by addition of paraformaldehyde afforded regioisomer (+)-{[S(S)]-[4-(2-hydroxyethyl)phenyl]sulfinyl}ferrocene (2), which was converted to (+)-{[S(S)]-{4-{2-[(methylsulfonyl)oxy]ethyl}phenyl}sulfinyl} ferrocene (3) (Scheme 1). The desired diastereoisomer (l)-1-(hydroxymethyl)-2- (p-tolylsulfinyl)ferrocene (5) in turn could also be obtained by ortho lithiation of 1 with LDA but by quenching with DMF to yield aldehyde 4 first, which then was reduced with NaBH₄ to 5. Finally, target compound (l)-1-[(dimethylamino)methyl]-2-(p-tolylsulfinyl)ferrocene (6) was obtained by substitution of the OH group of 5 under mild conditions or directly by ortho lithiation of 1 with lithio-2,4,6-triisopropylbenzene (=2,4,6- triisopropylphenyl)lithium; LTP) followed by quenching with N,N- dimethylmethyleneiminium chloride. At low temperatures, reaction of 1 with LDA leads, via the preferred diastereoisomeric transition state 'exo'-7 and under extrusion of a (diisopropylamine)lithium complex of type 8, in a highly selective manner, to diastereoisomeric ortho-lithiated chelate (l)-9 (Scheme 2). The reaction of 1 to 2 is explained by a rearrangement of (l)-9 to {[S(S)][4-(lithiomethyl)phenyl]sulfinyl}ferrocene 10, which is acid-catalyzed by coordinated diisopropylamine in complexes of type 8. This rearrangement is not observed if LTP is used as base or, in case LDA is applied, if the electrophile is sufficiently reactive at low temperatures.

Full text of DOI:10.1002/hlca.200790082

820
Helvetica Chimica Acta – Vol. 90 (2007)
A Synthesis Detour to Planar-Diastereoisomeric Ferrocene Derivatives
around an Unexpected Rearrangement of ortho-Lithiated Kaganꢀs Template
[S(S)]-(p-Tolylsulfinyl)ferrocene
by Immo Weber*1), Frank W. Heinemann, and Ulrich Zenneck
Institut für Anorganische Chemie, Friedrich-Alexander-Universität Erlangen-Nürnberg,
Egerlandstrasse 1, D-91058 Erlangen
Usually, ortho lithiation of Kaganꢀs template 1 and quenching with electrophiles leads highly
diastereoselectively to planar-chiral 1,2-disubstituted ferrocenes. Surprisingly, lithiation of 1 with lithium
diisopropylamide (LDA) followed by addition of paraformaldehyde afforded regioisomer (þ)-{[S(S)]-
[4-(2-hydroxyethyl)phenyl]sulfinyl}ferrocene (2), which was converted to (þ)-{[S(S)]-{4-{2-[(methyl-
sulfonyl)oxy]ethyl}phenyl}sulfinyl}ferrocene (3) (Scheme 1). The desired diastereoisomer (l)-1-(hy-
droxymethyl)-2-(p-tolylsulfinyl)ferrocene (5) in turn could also be obtained by ortho lithiation of 1 with
LDA but by quenching with DMF to yield aldehyde 4 first, which then was reduced with NaBH4 to 5.
Finally, target compound (l)-1-[(dimethylamino)methyl]-2-(p-tolylsulfinyl)ferrocene (6) was obtained
by substitution of the OH group of 5 under mild conditions or directly by ortho lithiation of 1 with lithio-
2,4,6-triisopropylbenzene (¼2,4,6-triisopropylphenyl)lithium; LTP) followed by quenching with N,N-
dimethylmethyleneiminium chloride. At low temperatures, reaction of 1 with LDA leads, via the
preferred diastereoisomeric transition state ꢁexoꢀ-7 and under extrusion of a (diisopropylamine)lithium
complex of type 8, in a highly selective manner, to diastereoisomeric ortho-lithiated chelate (l)-9
(Scheme 2). The reaction of 1 to 2 is explained by a rearrangement of (l)-9 to {[S(S)][4-
(lithiomethyl)phenyl]sulfinyl}ferrocene 10, which is acid-catalyzed by coordinated diisopropylamine in
complexes of type 8. This rearrangement is not observed if LTP is used as base or, in case LDA is applied,
if the electrophile is sufficiently reactive at low temperatures.
1. Introduction. – Robust mono- [1a,b] and bidentate [1c][2] planar-chiral
ferrocenyl ligands gained a fundamental role in enantioselective catalysis over the
past decade. This is documented by the industrially important Syngenta (S)-Metola-
chlor process [3], for example. Usually 1,2-disubstituted ferrocene derivatives are
obtained via ortho lithiation of a ferrocene containing coordinating functionalities;
R2NÀCR2À and ROÀCR2À are the most common types. Usually, alkyllithium
reagents are used for this deprotonation reaction. Coordination of the lithium ion with
the ligating substituent at the ferrocene guides the carbanion to the ortho position,
where deprotonation can then occur under kinetic control, popularly defined as ꢁkinetic
acidityꢀ. This deprotonation can also occur under thermodynamic control, to which is
1
)
New address: Kunststoff- und Metallwaren-Fabrik (KUM) GmbH & Co. KG, Abteilung Forschung
und Entwicklung, Essenbacher Strasse 2, D-91054 Erlangen (phone: þþ 49-(0)9131-826823; fax:
þþ 49-(0)9131-788629; e-mail: immo.weber@kum.net).
ꢂ 2007 Verlag Helvetica Chimica Acta AG, Zürich

Helvetica Chimica Acta – Vol. 90 (2007)
821
popularly referred as ꢁthermodynamic acidityꢀ2), if the pKA value of the ortho-H-atom
at the h5-cyclopentadienyl (h5-Cp) unit is lower than the one of the conjugate acid of
the lithiating reagent. In this sense, the resulting chelation of the lithium ion with the
coordinating group stabilizes the ortho-lithiated ferrocene, in the first case in a
metastable or in the second in a truly thermodynamic fashion. To gain full benefit from
these coordination effects, THF, Et2O, or hexanes and mixtures thereof are the solvents
of choice, which coordinate only in a labile fashion or not at all with the lithium ion.
After ortho lithiation, quenching with an electrophile leads then to the desired 1,2-
disubstituted planar-chiral ferrocene. Enantiomerically pure 1,2-disubstituted ferro-
cenes are usually obtained by resolution of diastereoisomers via a temporarily
introduced chiral auxiliary as covalent substituent [4] and mostly established by
diastereoselective ortho lithiation of a ferrocene template containing a ꢁdiastereodir-
ectingꢀ chiral auxiliary group such as Ugiꢀs template (R)- or (S)-[1-(dimethylamino)-
ethyl]ferrocene (FA) [2][3][5]. Kaganꢀs template {[S(S)]-p-tolylsulfinyl}ferrocene (1)
[6] can also be ortho-lithiated with high diastereoselectivity, but BuLi, sec-BuLi, and t-
BuLi cannot be used as bases, because they substitute the chiral sulfinyl group, which
serves as the ꢁdiastereodirectingꢀ functionality. This can be avoided, if lithio-2,4,6-
triisopropylbenzene (¼(2,4,6-triisopropylphenyl)lithium; LTP) [6e,g] is used instead,
which is sterically too bulky for a nucleophilic attack on the sulfinyl group. Compared
to FA, the thermodynamic CÀH acidities of the h5-Cp moieties of 1 are enhanced to
such an extent by the inherently strong electron-withdrawing sulfinyl group that also
lithium diisopropylamide (LDA) can be used for ortho lithiation. Unfortunately, the
use of LDA is limited by the electrophiles added later, which must be inert against LDA
and diisopropylamine as well. Taking advantage of the aforementioned base
incompatibility, the sulfinyl group of the resulting ortho-substituted {[S(S)]-p-
tolylsulfinyl}ferrocene can be substituted by virtually any other electrophile by Li-
t
exchange with BuLi in a following step [6]. In this way Kaganꢀs template 1 offers a
larger synthetic variety than FA, so we chose 1 for our quest for planar-chiral 1,2-
disubstituted ferrocenyl ligands. Suitable synthetic pathways for the diastereoselective
introduction of a (dimethylamino)methyl group in ortho position of 1 were investigated
(Scheme 1), because the resulting (l)-1-[(dimethylamino)methyl]-2-(p-tolylsulfinyl)-
ferrocene (6) is a very promising intermediate for the simple preparation of Josiphos
analogs [2c – f]. Derivative 6 is also a prototype ligand for bidentate planar-chiral
ferrocenyl ligands containing a chiral sulfinyl group as second donor functionality. Such
sulfinyl groups can complex a metal by an O- or S-binding mode and are expected to
show additional steric effects due to their important geometric asymmetry. Therefore,
such ligands are very attractive for possible applications in enantioselective catalysis.
2. Results and Discussion. – The diastereoselective introduction of a hydroxymethyl
group in 1 leading to (l)-1-(hydroxymethyl)-2-(p-tolylsulfinyl)ferrocene (5), followed
2
)
The authors wish to express their concerns about the terminologies ꢁkineticꢀ and ꢁthermodynamicꢀ
acidity, because ꢁacidityꢀ as the dissociation capability of an acid (HA) into protons (Hþ resp. H3Oþ)
and into the conjugate base (AÀ) expressed by the pKA value correlating to the HÀA bond strength
is by definition exclusively thermodynamic! Only for a common transparency, the authors finally
decided to use these unfortunate terminologies.

822
Helvetica Chimica Acta – Vol. 90 (2007)
Scheme 1
a) 1. LDA (2.0 equiv.)/THF, À 788; 2. (H2CO)n (5.8 equiv.), À 788 to r.t. b) MeSO2Cl (1.2 equiv.), Et3N
(1.5 equiv.), CH2Cl2, 08 to r.t. c) With racemate of 1: 1. LDA (2.0 equiv.), THF, À 788; 2. DMF
(5.8 equiv.), À 788 to r.t. d) With not further purified crude 4: NaBH4 (ꢁ1.1 equiv.ꢀ), MeOH, r.t. e) 1.
Addition to preformed phosphine adduct from PPh3 (1.1 equiv.) þ CBr4 (1.2 equiv.), CH2Cl2, À 708; 2.
[Me2N]2CH2 (3.0 equiv.), AgBF4 (1.2 equiv.), r.t. f) With racemate of 1: 1. LTP (2.0 equiv.), THF, À 788;
2. [Me2N¼CH2]þ ClÀ (4.2 equiv.), À 788 to r.t.
b
a) In this and the following schemes, 1 refers to 86.4% e.e. of [S(S)]. ) Overall yield.
by linear functional-group interconversion, was anticipated first. Lithiation of 1 (86.4%
enantiomeric excess (e.e.), [S(S)]) with LDA and reaction with paraformaldehyde did
not lead to the expected 1,2-disubstituted ferrocene diastereoisomer 5 but exclusively
to regioisomer (2) in 49% yield; some residual, not racemized 1 was recovered. This
result is the more surprising because racemic 5 was obtained under similar reaction
conditions, but using for ortho lithiation ferrocenyllithium (FcLi) [6e]3) and LTP [6e,g]
for analog 1,2-substituted ferrocenes instead. While LTP deprotonates irreversibly at
the ortho position, diisopropylamine, the conjugate acid of LDA, is obviously not an
3
)
Note that FcLi is not an appropriate base for ortho lithiation of 1 because the sulfinyl group
racemizes by an intermolecular exchange equilibrium with FcLi [6b,f]. Nevertheless, racemic 5 was
obtained as pure (l)-diastereoisomer after 1 h reaction with paraformaldehyde at 08.

Helvetica Chimica Acta – Vol. 90 (2007)
823
innocent spectator in ortho-lithiation reactions. Both, the MeÀC(4) group at the
phenylsulfinyl part and the h5-Cp units are in conjugation with the electron-with-
drawing sulfinyl group of 1, but this Me group is also in s*-p resonance with the
benzene moiety, which is more electron-deficient than the h5-Cp units4). Therefore, the
thermodynamic CÀH acidity of this Me group is even more enhanced than the ones of
the h5-Cp units. On the first hand, this Me group with three equivalent protons is
expected to have also a three times higher statistical probability for proton abstraction,
and in this way also a stronger kinetic acidity than the two diastereotopic positions at
the h5-Cp units. If so, then lithiation of 1 followed by electrophile quenching should
lead generally to the MeÀC(4)-functionalized isomer under kinetic and thermody-
namic conditions as well, and independently from the base used. But this is not the case,
so these alleged contradictions require a clarifying explanation (Scheme 2).
i) Obviously the strong basicity of LDA (pKA diisopropylamine > 35) does not
determine the position of proton abstraction but only that the reaction itself can
proceed (Paths c and e in Scheme 2). ii) It is reasonable to assume that the
diastereoselectivity of the ortho-lithiation is driven by precomplexation of LDA via
coordination of the sulfinyl O-atom of 1 with Liþ in the sense of two possible
conformational and diastereoisomeric transition states ꢁexoꢀ-7 and ꢁendoꢀ-7 (Paths a
and b in Scheme 2). This precomplexation under kinetic control at À 788 is obviously
strong enough to outflank the kinetic and the thermodynamic acidity of the MeÀC(4)
group: It increases the basicity of the N-atom of LDA and ꢁguidesꢀ it simultaneously
and preferably closest to that diastereotopic ortho-H-atom of the h5-Cp unit which
allows then the p-tolyl group to adopt a sterically preferred ꢁexoꢀ position in this whole
arrangement. This leads finally via transition state ꢁexoꢀ-7 to the diastereoisomeric
ortho-lithiated chelate (l)-9 (Path c in Scheme 2) and finally by trapping with a strong
electrophile to the corresponding like diastereoisomer 4 (Path d in Scheme 2). If very
slow addition rates of LDA at À 788 are chosen, the formation of the sterically less
preferable transition state ꢁendoꢀ-7 leading to ortho-lithiated chelate (u)-9 can be
suppressed completely (Path b in Scheme 2). iii) Seebach and co-workers found in the
deuteration of various compounds with LDA that the deuterium incorporation is
sometimes incomplete because diisopropylamine is coordinated to Li-cations [7]. In
such complexes of the general type 8 (Path c in Scheme 2), diisopropylamine is then
transformed to a comparably strong acid, which can now act catalytically. Furthermore,
it is also possible for diisopropylamine to coordinate to the corresponding ortho-
lithioferrocene intermediate (l)-9 in analogy to the known [Li(Fc)(tmeda)] complex
(tmeda ¼ N,N,N’,N’-tetramethylethane-1,2-diamine) [8], giving rise to an intramolec-
ular active proton source. Therefore, ortho lithiation of Kaganꢀs template 1 becomes
now reversible. Consistent with this explanation, we found that ortho lithiation of 1
followed by electrophile quenching never leads to complete conversion to the desired
1,2-disubstituted ferrocene, even when up to 4 equiv. of LDA are used. In other words,
the formation of diastereoisomeric ortho-lithiated chelate (l)-9 (also of (u)-9 if addition
of LDA is not performed slowly enough) is not an irreversible but only a metastable
energy sink under these reaction conditions. iv) Paraformaldehyde as polyacetal is not
Note that ferrocene reacts 3 · 106 times faster than benzene in electrophilic aromatic substitution
reactions.
4
)

824
Helvetica Chimica Acta – Vol. 90 (2007)
Scheme 2. Diastereoselective ortho-Lithiation of Kaganꢀs Template 1 (blue) and Lithio Rearrangement
i
(red). Permutative ligand coordination around Liþ is symbolized with L ¼ THF, Pr2NÀ, LDA, HNiPr2
(green).

Helvetica Chimica Acta – Vol. 90 (2007)
825
well soluble, not reactive, and does not depolymerize when added at À 788. Reaching
slowly thermodynamic conditions, (l)-9 can now rearrange to the thermodynamically
favored product 10 (Path e in Scheme 2) because reaction with an electrophile cannot
occur at that moment. Later on, during warmup to room temperature and after decay of
paraformaldehyde into its reactive carbonyl monomer (Path f in Scheme 2), base 10 is
trapped to give finally 2 (Path g in Scheme 2). v) The diastereoselectivity of the ortho
lithiation of Ugiꢀs template (R)- or (S)-FA is widely explained by an equilibrium of its
corresponding ortho-lithiated diastereoisomers. In analogy to (R)- or (S)-FA, an
equilibrium of (l)-9 and (u)-9 should be in favor of (l)-9 due to the preferred ꢁexoꢀ
position of the p-tolyl group with decreasing temperature, of course. If (l)- and (u)-9
interconvert directly, then such an equilibrium must be trapped by the irreversible
formation of the thermodynamically favored product 10. This is not the case if ortho
lithiation is not performed with LDA [6e], establishing that without the presence of
complexed diisopropylamine, diastereoisomeric chelates of the type (l)- and (u)-9 are
not in an equilibrium. Therefore, such an explanation for the high diastereoselectivity
of the ortho lithiation of Kaganꢀs template 1 must be questioned in favor of the
previously discussed strong kinetic acidity of the ortho-H-atom at the h5-Cp unit
represented by the diastereomeric transition state ꢁexoꢀ-7.
Rearrangement product 2 (min. 86.4% e.e.) was then converted to mesylate
derivative 3 in 98% yield (Scheme 1), which might be of interest as suitable
electrophilic reagent for the introduction of terminal chiral ferrocenyl units on
dendrimers [9], for instance. Crystals of 3 suitable for X-ray crystal-structure
determination were obtained from a saturated CH2Cl2 solution with some drops of
AcOEt (Fig. 1, Table 1). In this way, the rearrangement and trapping reaction of 1 to 2
is established beyond doubt.
Fig. 1. Thermal ellipsoid plot (50% probality) of the molecular structure of the chiral ferrocenyl derivative
3 obtained from unexpected product 2. For selected bond distances and angles, see Table 1.
Racemic aldehyde 4 was then obtained by ortho lithiation of racemic 1 followed by
quenching with DMF. NMR Analysis of the crude product established that racemic 4
was obtained diastereoisomerically pure. This shows again that the regio- and
coherently also the diastereoselectivity is only dependent on the electrophileꢀs

826
Helvetica Chimica Acta – Vol. 90 (2007)
Table 1. Selected Bond Distances and Angles of 3
Distance
[Š]
Angle
[8]
S(2)ÀC(11)
S(2)ÀO(1)
S(2)ÀC(10)
S(3)ÀO(2)
S(3)ÀO(3)
S(3)ÀO(4)
S(3)ÀC(19)
O(2)ÀC(18)
C(10)ÀFe(1)
C(5)ÀFe(1)
1.797(2)
1.488(2)
1.777(2)
1.574(2)
1.429(2)
1.428(2)
1.758(2)
1.471(2)
2.028(2)
2.057(2)
C(11)ÀS(2)ÀO(1)
C(10)ÀS(2)ÀO(1)
C(18)ÀO(2)ÀS(3)
O(3)ÀS(3)ÀO(4)
O(2)ÀS(3)ÀC(19)
C(10)ÀFe(1)ÀC(1)
C(10)ÀFe(1)ÀC(2)
C(10)ÀFe(1)ÀC(3)
C(10)ÀFe(1)ÀC(4)
C(10)ÀFe(1)ÀC(5)
106.91(9)
108.1(1)
119.3(2)
119.5(2)
103.3(2)
125.36(9)
158.83(9)
160.49(9)
126.57(9)
111.92(9)
reactivity at À 788 under otherwise identical reaction conditions (Path d in Scheme 2).
Aldehyde 4 was previously also obtained by using LTP as base and ethyl formate as
electrophile [6g]. Crude 4 was then directly reduced with NaBH4 to 5 allowing then an
easy removal of residual 1 from the previous product mixture by purification with
column chromatography to obtain pure 5 in 66% overall yield (Scheme 1). Again, no
epimerization of the chiral sulfinyl group was observed by NMR analysis of the crude
product.
Due to the strong electron-withdrawing sulfinyl group, the usual dissociation
tendency of the OH group proved to be retarded for 5 compared to (hydroxyme-
thyl)ferrocene. Of course, nucleophilic substitution of the OH group was desired, which
is known to proceed in SN1 fashion via (h4 :h2)-fulvenium intermediates [2][10], but the
corresponding intermediate 11 of that type is expected to be destabilized (Scheme 3).
Therefore, activation of the OH group by esterification [2] was considered to be
insufficient. If possible at all, nucleophilic substitution in glacial acetic acid [2][3]
would presumably have led to epimerization of the chiral sulfinyl group. Having Salzer
and co-workerꢀs electrophilic activation of benzylic positions of (h6-arene)tricarbonyl-
chromium(0) complexes in mind [11], racemic 5 was activated in situ as bromide 12
with the preformed PPh3/CBr4 adduct (Scheme 3). Preformation was necessary to
avoid reduction of the p-tolylsulfinyl to the corresponding p-tolylthio group by
triphenylphosphine. No attempts were made to isolate 12. Bromide 12 in turn was then
treated in situ with AgBF4 and N,N,N’,N’-tetramethylmethanediamine as synthetic
equivalent for free dimethylamine to give, presumably via adduct 13, the desired
[(dimethylamino)methyl]ferrocene 6. Again, only the like diastereoisomer of 6 could
be detected in the crude product by NMR, but 6 was only obtained in 12% yield after
chromatography with this method. To the best of our knowledge, this reaction is the
first successful activation and substitution reaction at a fulvenic position of a 1,2-
disubstituted ferrocene containing a strongly deactivating group.
This reaction was then finally complemented with the diastereoselective ortho
lithiation of racemic 1 with LTP followed by quenching with N,N-dimethylmeth-
yleneiminium chloride (Eschenmoserꢀs salt) [12] to diastereoisomerically pure racemic
6 (Scheme 1) in a more practicable yield of 52%. After addition of Eschenmoserꢀs salt
to intermediate (l)-9, the deep orange color turned into clear yellow within minutes,

Helvetica Chimica Acta – Vol. 90 (2007)
827
Scheme 3. Presumable in situ SN1-Type Reaction Pathway of 5 to 6 via (h4 :h2)-Fulvenium Intermediate
11
demonstrating the strong electrophilic potential of Eschenmoserꢀs salt compared to
DMF and to the polyacetal paraformaldehyde. After applying a trick (see Exper. Part),
single crystals of racemic 6 suitable for X-ray crystal-structure determination could be
obtained from a saturated AcOEt solution. The crystal examined contained both
enantiomers of 6 (Fig. 2, Table 2). This confirms the diastereoselectivity of the ortho
lithiation and the configurational integrity vide infra. Conclusively, now the method for
preparing enantiomerically pure 6 is given.
3. Conclusion. – The rearrangement of ortho-lithiated intermediate (l)-9 to the
thermodynamically more stable base 10 can only occur if diisopropylamine complexes
of type 8 are present as catalysts in the reaction solution. The following crucial issues
should be considered for the preparation of diastereoisomerically pure 1,2-disubsti-
tuted ferrocenes via ortho lithiation of Kaganꢀs template 1: i) if LDA is chosen as base,
then ortho lithiation as well as electrophile quenching should be performed strictly at
À 788; furthermore, the subjected electrophile should not only be inert against LDA
and diisopropylamine but also highly reactive at this temperature. ii) If the electrophile
does not fulfill these requirements, then ortho lithiation should preferably be
performed with LTP. iii) Complementary, if functionalization of the MeÀC(4) group
is desired, lithiation must be performed with LDA and, before electrophile quenching,
the reaction solution must warm up over a sufficient time scale to ensure rearrange-
ment to intermediate 10.

828
Helvetica Chimica Acta – Vol. 90 (2007)
Fig. 2. Thermal ellipsoid plot (50% probality) of the arrangement of the two symmetrically independent
molecules representing the two enantiomers of 6 (like) as [S(S),M]-6 (left) and [S(R),P]-6 (right). Note
that due to the inversion center of space group P1 (no. 2), the corresponding inversion image of this
arrangement is also present in the unit cell. The two independent molecules differ only slightly in their
conformations. For selected bond distances and angles, see Table 2.
Table 2. Selected Bond Distances and Angles of racemic 6
Distance
[Š]
Angle
[8]
S(1)ÀC(11)
S(1)ÀO(1)
1.796(2)
1.496(2)
1.777(2)
1.506(3)
1.469(2)
1.458(2)
1.462(2)
1.792(2)
1.496(2)
1.779(2)
1.508(3)
1.477(2)
1.460(2)
1.459(2)
C(11)ÀS(1)ÀO(1)
C(10)ÀS(1)ÀO(1)
C(9)ÀC(18)ÀN(1)
C(18)ÀN(1)ÀC(19)
C(18)ÀN(1)ÀC(20)
106.65(8)
106.91(8)
113.1(2)
109.8(2)
110.5(2)
S(1)ÀC(10)
C(18)ÀC(9)
N(1)ÀC(18)
N(1)ÀC(19)
N(1)ÀC(20)
S(2)ÀC(31)
S(2)ÀO(2)
S(2)ÀC(30)
C(38)ÀC(29)
N(2)ÀC(38)
N(2)ÀC(39)
N(2)ÀC(40)
C(31)ÀS(2)ÀO(2)
C(30)ÀS(2)ÀO(2)
C(29)ÀC(38)ÀN(2)
C(38)ÀN(2)ÀC(39)
C(38)ÀN(2)ÀC(40)
106.60(8)
106.96(8)
113.4(2)
109.6(2)
110.6(2)
We propose the strong kinetic acidity of the ortho-H-atom at the h5-Cp unit as main
reason for the high diastereoselectivity of the ortho lithiation of 1 and not an
interconversion equilibrium of the diastereoisomeric chelate intermediates (l)- and
(u)-9. This strong kinetic acidity is given by a synergism of a simultaneous increase of
the N-atom basicity of precomplexed LDA and the preferable conformational ꢁexo-

Helvetica Chimica Acta – Vol. 90 (2007)
829
guidanceꢀ of the base to the ortho-H-atom as represented by the diastereomeric
transition state ꢁexoꢀ-7. To suppress the formation of the undesired chelate intermediate
(u)-9 completely, slow addition rates of LDA to 1 have to be chosen. For practical
reasons, ortho lithiation of 1 with LTP is recommended to be performed by slow inverse
addition, but due to the steric bulk of the base, complete diastereoselectivity is also
ensured here. Based simply upon different electrophile reactivities and base variations,
these estimations are qualitative in nature in the scope of our more synthesis-orientated
work. However, they clearly indicate that the mechanistic discussion of the highly
diastereoselective ortho lithiation of chiral ferrocene templates has to be continued.
The in situ activation and nucleophilic substitution of the OH group of 5 might be
developed further into a general procedure for the preparation of various bidentate
planar-chiral ferrocenyl ligands containing a chiral sulfinyl group as second donor
functionality.
Financial support by the Deutsche Forschungsgemeinschaft (SFB 583) and the Kunststoff- und
Metallwarenfabrik Erlangen (KUM) GmbH & Co. KG is gratefully acknowledged. We are indebted to
´
´
´
Professor Dr. Henri B. Kagan (Laboratoire des Reactions Selectives, Institut de Chemie Moleculaire
´
dꢀOrsay, Universite Paris-Sud) for very helpful discussions. We are grateful to Professor Dr. Peter
Gmeiner (Emil Fischer Zentrum-Lehrstuhl für Pharmazeutische Chemie und Lebensmittelchemie,
Friedrich-Alexander-Universität Erlangen-Nürnberg) and to Professor Dr. Andreas Hirsch (Institut für
Organische Chemie II, Friedrich-Alexander-Universität Erlangen-Nürnberg) for providing access to a
melting point apparatus and to a polarimeter.
Experimental Part
1. General. All reactions were carried out under dry N2 by using conventional Schlenk and septum
techniques. Liquid reagents were generally added with disposable plastic syringes, but reagent solns.
freshly prepared prior to use were transferred to reaction solns. from a Schlenk tube by using a double
canula with N2 overpressure. All workups were performed in air. Solvents and chemicals were purchased
from Acros, Aldrich, Strem, and Merck. Solvents and amines were dried and distilled under N2 prior to
use: THF over sodium/benzophenone; MeOH over magnesium; CH2Cl2, iPr2NH, and Et3N over CaH2.
The actual concentrations of the alkyllithium reagents were determined by inverse titration of
diphenylacetic acid prior to use. All other chemicals were used as received. LTP was prepared from 1-
bromo-2,4,6-triisopropylbenzene [6e] in a slightly modified fashion (see Exper. 6). Kaganꢀs template 1
was prepared as published [6b,d]. The enantiomeric excess (e.e.) of 1 was determined by HPLC [6b], and
the e.e. of the resulting diastereoisomeric 2 and 3 referenced to it as starting material. We did not achieve
a satisfactory separation of the enantiomers of 2 and 3 by HPLC on various chiral stationary phases.
N,N,N’,N’-Tetramethylmethanediamine was prepared according to [12a] and distilled from CaH2 under
N2 prior to use. Eschenmoserꢀs salt N,N-dimethylmethyleneiminium chloride was prepared by
Danishefskyꢀs method [12c], but the isolated product contained also polymeric {Me2N[CH2NMe2]nCH2N
Me2}Cl, which does react in the same way. Flash column chromatography (FC): silica gel F 60 from Fluka
or Merck. Thin layer chromatography (TLC): Merck F-60 silica plates with a 364-nm fluorescence
indicator. Melting points: Büchi 530 melting point apparatus; not corrected. Polarimetric measurements:
Perkin-Elmer-341 polarimeter. NMR Spectra: Jeol FT-JNM-EX 270 (270 MHz) and Bruker AMX-300
(300 MHz) spectrometers; in deuterated solvents and referenced to the residual proton signal of the
particular solvent. Mass Spectra: Varian MAT-212 spectrometer; in m/z (rel. %). Elemental analyses:
Carlo-Erba elemental analyzer, model 1108.
2. (þ)-{[S(S)]-[4-(2-Hydroxyethyl)phenyl]sulfinyl}ferrocene (2). An LDA soln. prepared from
iPr2NH (8.00 ml, 56.92 mmol) in THF (20 ml) at À 208 with 1.58m tBuLi in pentane (31.00 ml,
48.97 mmol) was transferred dropwise within 15 min to a well stirred soln. of 1 (7.82 g, 24.12 mmol;
86.4% e.e.) in THF (120 ml) at À 788. The soln. turned slowly deep orange and was stirred 90 min at

830
Helvetica Chimica Acta – Vol. 90 (2007)
À 788 (!deep orange suspension). To this suspension was added in one portion solid paraformaldehyde
(4.21 g, 140.13 mmol), and the mixture was stirred for 16 h in the cooling bath slowly defrosting to r.t.
(! yellow suspension). The suspension was poured into brine saturated with NH4Cl and extracted twice
with AcOEt. The combined org. layer was dried (MgSO4) and concentrated: 9.02 g of crude 2 as solid
orange foam containing also starting material 1 (by NMR). The crude product was purified by FC
(substance applied in a silica gel matrix, hexanes/AcOEt 1:3): 4.17 g (49%; Rf 0.13 (TLC)) of 2 and
151 mg (2%; Rf 0.43 (TLC)) of recovered of 1. 2. M.p. 139 – 1408. [a]2D3 ¼ þ231.3 (CH2Cl2, c ¼ 0.0029;
86.4% e.e.). 1H-NMR (CDCl3, 270 MHz): 7.52 (d, 3J ¼ 8.1, 2 Ho of C6H4); 7.28 (d, 3J ¼ 8.1, 2 Hm of C6H4);
4.59 (m, 1 H, Cp); 4.34 – 4.31 (s and m, 8 H, Cp and Cp’); 3.82 (t, 3J ¼ 6.6, CH2CH2OH); 2.85 (t, 3J ¼ 6.6,
CH2CH2OH); 1.82 (br. s, OH). 13C{1H}-NMR (CDCl3, 68 MHz): 143.91 (Cipso of C6H4); 141.74 (Cp of
C6H4); 129.56 (Cm of C6H4); 124.42 (Co of C6H4); 94.13 (C(1) of Cp); 70.04 (Cp); 69.94 (Cp and Cp’);
67.92 (Cp); 64.99 (Cp); 63.19 (CH2CH2OH); 38.96 (CH2CH2OH). FAB-MS: 186 (43, Fcþ with respect to
56Fe), 339 (26, [M À O]þ with respect to 56Fe), 355 (100, Mþ with respect to 56Fe). A correct elementary
analysis could not be obtained.
3. (þ)-{[S(S)]-{4-{2-[(Methylsulfonyl)oxy]ethyl}phenyl}sulfinyl}ferrocene (3). To a stirred soln. of
2 (4.036 g, 11.39 mmol; 86.4% e.e.) and Et3N (2.30 ml, 1.670 g, 16.50 mmol) in CH2Cl2 (70 ml) at 08, mesyl
chloride (1.10 ml, 1.592 g, 13.90 mmol) was added dropwise. The red clear soln. was stirred for 12 h in the
cooling bath defrosting to r.t. The soln. was poured into a sat. NaHCO3 soln. and the lower org. phase
washed once with brine, dried (MgSO4), and concentrated to afford 4858 mg (99%) of nearly pure 3
which was purified by FC (substance applied in eluent, AcOEt): 4.837 g (98%) of 3 as yellow
microcrystals. Crystallization from CH2Cl2 and some drops AcOEt gave crystals suitable for X-ray
analysis. M.p. 1118. [a]2D3 ¼ þ199.6 (CH2Cl2, c ¼ 0.0027; 86.4% e.e.). 1H-NMR (CDCl3, 270 MHz): 7.56 (d,
3J ¼ 8.1, 2 Ho of C6H4); 7.30 (d, 3J ¼ 8.1, 2 Hm of C6H4); 4.57 (m, 1 H, HÀC(2) or HÀC(5) of Cp); 4.39 (t,
3J ¼ 6.7, CH2CH2O); 4.37 – 4.31 (s and m, Cp’ and HÀC(3), HÀC(4), and HÀC(5), or HÀC(2),
HÀC(3), and HÀC(4) of Cp); 3.06 (t, 3J ¼ 6.7, CH2CH2O); 2.83 (s, Me). 13C{1H}-NMR (CDCl3,
68 MHz): 145.06 (Cipso of C6H4); 139.10 (Cp of C6H4); 129.49 (Cm of C6H4); 124.54 (Co of C6H4); 94.27
(C(1) of Cp); 70.08 (Cp); 69.95 (CH2CH2O and Cp’); 69.49 (Cp); 67.92 (Cp); 64.84 (Cp); 37.41 (Me);
35.39 (CH2CH2O). FD-MS (pos.; CH2Cl2): 433 (100, Mþ with respect to 56Fe). Anal. calc. for
C19H20FeO4S2 (432.34): C 52.78, H 4.66, S 14.83; found: C 52.88, H 4.80, S 14.63.
4. rac-(l)-1-Formyl-2-[(4-methylphenyl)sulfinyl]ferrocene (4) and rac-(l)-1-(Hydroxymethyl)-2-[(4-
methylphenyl)sulfinyl]ferrocene (5). a) Crude 4. To racemic 1 (7.212 g, 22.24 mmol) in THF (100 ml) at
i
t
À 788, a freshly prepared LDA soln. from Pr2NH (7.00 ml, 5.040 g, 49.81 mmol) and 1.69m BuLi in
pentane (26.00 ml, 43.93 mmol) in THF (20 ml, at À 208) was transferred dropwise within 15 min. The
soln. turned slowly to a deep orange suspension and was stirred for 130 min at À 788. The suspension was
quenched at À 788 with DMF (10.00 ml, 128.61 mmol), and after 5 min stirring, the suspension turned
into a clear orange soln., which was stirred overnight defrosting slowly to r.t. inside the cooling bath. The
soln. was poured into brine and extracted once with AcOEt, the org. layer washed five times with brine to
remove all DMF, dried (MgSO4), and concentrated: 7.448 g (95%) of diastereoisomerically pure racemic
4 containing 1. The crude product can be purified by FC (substance applied in a silica gel matrix, hexanes/
AcOEt 1:2; Rf(1) 0.40 and Rf(4) 0.28 (TLC), but contained still traces of 1, and the yield dropped to 74%
in another batch, so the crude product was directly taken to the next step and purified then. Spectroscopic
data of 4: in full agreement with the ones reported [6g].
b) Pure 5. All of the crude product 4 from above was dissolved in MeOH (160 ml), and then NaBH4
(841 mg, 22.23 mmol) was added in portions under a stream of N2 to the well stirred soln. at r.t. The soln.
was stirred for another 20 min until foaming and H2 evolution ceased. After quenching with AcOEt
(2 ml) and concentration, the residue was dissolved in AcOEt again and the soln. washed once with brine,
dried (MgSO4), and concentrated to give 7.358 g (93%) of racemic 5 as a brown oil containing small
amounts of 1 but no 4. The crude product was purified by FC (substance applied in a silica gel matrix,
gradient elution with Et2O/CH2Cl2 1:0 (! impurities and 1) and then with Et2O/CH2Cl2 1:0, 3 :1, 1:1
(! 5): 644 mg (9%; Rf 0.32 (TLC)) 1 and 5.167 g (66% overall ; Rf 0.19 (TLC)) of racemic 5 as yellow
microcrystalline powder. Spectroscopic data of 5: in full agreement with the ones reported [6e].
5. rac-(l)-1-[(Dimethylamino)methyl]-2-[(4-methylphenyl)sulfinyl]ferrocene (6). a) From 5. To a
clear soln. of PPh3 (2.940 g, 11.21 mmol) in CH2Cl2 (20 ml), CBr4 (3.957 g, 11.93 mmol) was added at

Helvetica Chimica Acta – Vol. 90 (2007)
831
À 658 in one portion (! immediately yellowish soln. and white precipitation). This suspension was stirred
for 20 min outside the cooling bath defrosting to r.t. The resulting clear soln. was canuled to a soln. of
racemic 5 (3.481 g, 9.83 mmol) in CH2Cl2 (30 ml) at À 708, and the Schlenk tube with the phosphine
adduct was washed out twice with a total of 10 ml of CH2Cl2. The washing solns. were transferred to the
orange reaction soln., which was stirred for 26 h outside the cooling bath at r.t. The soln. was
concentrated to half its volume by blowing off the solvent with N2. Then N,N,N’,N’-tetramethylmetha-
nediamine (4.00 ml, 29.32 mmol) and thereafter AgBF4 (2.283 g, 11.73 mmol) were added in this order
(!AgBr precipitation). After 5 min stirring at r.t., the suspension was poured into brine, the mixture
extracted twice with CH2Cl2, and the combined org. layer dried (MgSO4), and concentrated to give
8.856 g of crude product as a complex mixture containing diastereoisomerically pure 6 (by NMR). The
crude product was purified twice by FC (substance applied in a silica gel matrix, gradient elution with
hexanes/AcOEt þ 10% Et3N 1:1, then 2 :1): 462 mg (12%) of 6 (free of triphenylphosphine oxide) as a
red oil solidifying on standing. Single crystals suitable for X-ray-analysis were obtained by mixing the
purified product with AcOEt and with a few drops of hexanes and CH2Cl2 to a red slime, by adding one
1
sand corn, and by leaving the mixture at r.t. for several days. M.p. 1008 (racemate). H-NMR (CDCl3,
270 MHz): 7.63 (d, 3J ¼ 8.3, 2 Ho of C6H4); 7.27 (d, 3J ¼ 8.3, 2 Hm of C6H4); 4.52 (m, 1 H, Cp); 4.28 (m, 1 H,
Cp); 4.18 (s, 5 H, Cp’); 4.15 (m, 1 H, Cp); 3.56 (d, 2J ¼ 13.3, 1 H, CH2); 3.49 (d, 2J ¼ 13.3, 1 H, CH2); 2.39
(s, 3 H, MeC6H4); 2.08 (s, Me2N). 13C{1H}-NMR (CDCl3, 75 MHz): 141.57 (Cipso of C6H4); 141.16 (Cp of
C6H4); 129.05 (Cm of C6H4); 125.01 (Co of C6H4); 91.64 (C(2) of Cp); 85.34 (C(1) of Cp); 72.21 (Cp);
70.29 (Cp’); 68.89 (Cp); 68.35 (Cp); 56.03 (CH2); 44.87 (Me2N); 21.17 (MeC6H4). FD-MS (pos.; CH2Cl2/
AcOEt): 382 (100, Mþ with respect to 56Fe). Anal. calc. for C20H23FeNOS (381.32): C 63.00, H 6.08, N
3.67, S 8.41; found: C 62.96, H 6.16, N 3.84, S 8.19.
b) From 1. To a stirred soln. of 1-bromo-2,4,6-triisopropylbenzene (1.809 g, 6.39 mmol) in THF
t
(15 ml), 1.72m BuLi in pentane (7.40 ml, 12.77 mmol) was added dropwise at À 788 within 7 min. The
soln. was stirred for 4 h at À 788 (!orange and LiBr precipitation). The À 788 cold soln. of (2,4,6-
triisopropylphenyl)lithium (LTP) was directly used. To the freshly prepared LTP soln., a soln. of racemic
1 (1.033 g, 3.186 mmol) in THF (25 ml), held at ca. À 408, was transferred dropwise within 15 min. The
suspension was stirred defrosting to À 408 inside the cooling bath within 2 h (!deep red and clear soln.)
The soln. was cooled again to À 788, and solid N,N-dimethylmethyleneiminium chloride (1.245 g,
13.307 mmol) was added in one portion. The suspension was stirred inside the cooling bath for 15 h
slowly defrosting to r.t. Then the soln. was poured into brine and extracted twice with AcOEt, and the
combined org. layer dried (MgSO4) and concentrated to give 2.541 g of crude product as a deep red oil
containing diastereoisomerically pure racemic 6, 1,3,5-triisopropylbenzene, and some 1 (by NMR). The
crude product was purified by FC (gradient elution with hexanes/AcOEt þ 10% Et3N 3 :1 (! impurities
and 1: Rf(1) 0.50 (TLC)) then with 1:1, and finally with 1:2 (!6): 630 mg (52% yield) of racemic 6. TLC:
Rf 0.21.
6. Crystal Structure Determination5). Crystal parameters, data collection, and structure refinement
details are summarized in Table 3. Intensity data were collected at 100 K on a Bruker-Nonius-KappaCCD
diffractometer (MoKa radiation, l 0.71073 Š, graphite monochromator). All structures were solved by
direct methods [13][14] and refined by full-matrix least-squares procedures on F2. For 3, the absolute
configuration was determined by anomalous dispersion effects (least-square procedures, f’ and f’’)
[14a,b]. All non-H-atoms were refined with anisotropic displacement parameters. Absorption
corrections were performed by using either a numerical Gauss integration [14c] or on the basis of
multiple scans with SADABS [13c]. The positions of all H-atoms were taken from a difference Fourier
synthesis and were refined with a fixed common isotropic displacement parameter. The asymmetric unit
of 6 contains two symmetry-independent molecules differing slightly in their conformation, the two
enantiomers of the like diastereoisomer being present due to the centro symmetry of the space group.
5
)
CCDC-603699 (for 3) and CCDC-603700 (for 6) contain the supplementary crystallographic data
for this paper. These data can be obtained free of charge via http://www.ccdc.cam.ac.uk/
data_request/cif.

832
Helvetica Chimica Acta – Vol. 90 (2007)
Table 3. Crystallographic Data of 3 and Racemic 6. Measured at 100 K.
3
6
Empirical formula
Mr
C19H20FeO4S2
432.34
C20H23FeNOS
381.30
Color, shape
Crystal size [mm]
Crystal system
Space group
a [Š]
orange, irregular
0.21 ꢀ 0.10 ꢀ 0.07
orthorhombic
P212121 (no. 19)
7.977(1)
orange, irregular
0.23 ꢀ 0.18 ꢀ 0.14
triclinic
P1 (no. 2)
7.6883(7)
b [Š]
c [Š]
a, b, g [8]
V [Š3]
14.900(2)
15.213(2)
90, 90, 90
1808.2(4)
4
1.588
1.087
896
7.7400(4)
29.575(2)
86.231(6), 85.435(8), 89.663(6)
1750.6(2)
Z
4
1 [g/cm3] (calc.)
1.447
0.987
800
m [mmÀ1
F (000)
]
Abs. corr.
Tmin; Tmax
2q range [8]
Coll. refl.
multiple scans (SADABS)
0.843; 0.930
7.4 ꢁ 2q ꢁ 56.0
52022
numerical (Gauss integration)
0.836; 0.904
6.1 ꢁ 2q ꢁ 57.4
47120
Indep. refl.
4372
8984
Obs. refl. (I0 ꢂ 2s(I))
No. ref. param.
wR2 (all data)
3987
295
0.0582
6781
571
0.0793
R1 (I0 ꢂ 2s(I))
0.0278
1.048
0.0352
1.021
GooF (on F2)
absolute structure parameter
Max.; min. res. electr. density
0.03(2)
0.583; À 0.331
–
0.472; À 0.529
REFERENCES
[1] a) H. L. Pedersen, M. Johannsen, Chem. Commun. 1999, 2517; b) H. L. Pedersen, M. Johannsen, J.
Org. Chem. 2002, 67, 7982; c) M. Lotz, K. Polborn, P. Knochel, Angew. Chem., Int. Ed. 2002, 41, 4708
and ref. cit. therein.
[2] a) ꢁFerrocenesꢀ, Eds. A. Togni and T. Hayashi, VCH, Weinheim 1995; b) ꢁMetallocenesꢀ, Eds. A.
Togni and L. Hintermann, Wiley-VCH, Weinheim 1998; c) A. Togni, Angew. Chem. 1996, 108, 1581;
d) H.-U. Blaser, W. Brieden, B. Pugin, F. Spindler, M. Studer, A. Togni, Topics in Catalysis 2002, 19,
3; e) A. Schnyder, L. Hintermann, A. Togni, Angew. Chem., Int. Ed. Engl. 1995, 34, 931; f) U.
Burckhardt, L. Hintermann, A. Schnyder, A. Togni, Organometallics 1995, 14, 5415.
[3] H.-U. Blaser, H. P. Buser, K. Coers, R. Hanreich, H. P. Jalett, E. Jelsch, B. Pugin, H. D. Schneider, F.
Spindler, A. Wegmann, Chimia 1999, 53, 275; F. Spindler, H.-U. Blaser, Enantiomer 1999, 4, 557; R.
Dorta, D. Broggini, R. Stoop, H. Rüegger, F. Spindler, A. Togni, Chem.–Eur. J. 2004, 10, 267; R.
Dorta, D. Broggini, R. Kissner, A. Togni, Chem.–Eur. J. 2004, 10, 4546.
[4] L. Xiao, K. Mereiter, W. Weissensteiner, M. Widhalm, Synthesis 1999, 8, 1354; O. Riant, O. Samuel,
T. Flessner, S. Taudien, H. B. Kagan, J. Org. Chem. 1997, 62, 6733.
[5] C. R. Hauser, J. K. Lindsay, J. Org. Chem. 1957, 22, 906; G. W. Gokel, I. K. Ugi, J. Chem. Educ. 1972,
49, 294; D. Marquarding, H. Klusacek, G. Gokel, P. Hoffmann, I. K. Ugi, J. Am. Chem. Soc. 1970, 92,
5389; D. Lambusta, G. Nicolosi, A. Patti, M. Piattelli, Tetrahedron: Asymmetry 1993, 4, 919; G.
Nicolosi, R. Morrone, A. Patti, M. Piattelli, Tetrahedron: Asymmetry 1992, 3, 753.

Helvetica Chimica Acta – Vol. 90 (2007)
833
[6] a) D. Guillaneux, H. B. Kagan, J. Org. Chem. 1995, 60, 2502; b) O. Riant, G. Argouarch, D.
Guillaneux, O. Samuel, H. B. Kagan, J. Org. Chem. 1998, 63, 3511; c) P. Diter, O. Samuel, S. Taudien,
H. B. Kagan, Tetrahedron: Asymmetry 1994, 5, 549; d) F. W. Heinemann, I. Weber, U. Zenneck, J.
Chem. Crystallogr. 2007, 37, 165; e) N. M. Lagneau, Y. Chen, P. M. Robben, H.-S. Sin, K. Takasu, J.-S.
Chen, P. D. Robinson, D. H. Hua, Tetrahedron 1998, 54, 7301; f) D. H. Hua, N. M. Lagneau, Y. Chen,
P. M. Robben, G. Clapham, P. D. Robinson, J. Org. Chem. 1996, 61, 4508; g) L. Bernardi, B. F.
`
Bonini, E. Capito, G. Dessole, C. Femoni, M. Fochi, M. Comes-Franchini, A. Mincio, A. Ricci,
ARKIVOC 2004, 2, 72.
[7] T. Laube, J. D. Dunitz, D. Seebach, Helv. Chim. Acta 1985, 68, 1373.
[8] M. D. Rausch, D. J. Ciappenelli, J. Organomet. Chem. 1967, 10, 127; I. R. Butler, W. R. Cullen, J. Ni,
S. J. Rettig, Organometallics 1985, 4, 2196.
[9] C. Kçllner, B. Pugin, A. Togni, J. Am. Chem. Soc. 1998, 120, 10274; A. Togni, R. Dorta, C. Kçllner,
G. Pioda, Pure Appl. Chem. 1998, 70, 1477; R. Schneider, C. Kçllner, I. Weber, A. Togni, Chem.
Commun. 1999, 2415.
[10] U. Behrens, J. Organomet. Chem. 1979, 182, 89.
[11] U. Englert, R. Haerter, D. Vasen, A. Salzer, E. B. Eggeling, D. Vogt, Organometallics 1999, 18, 4390;
U. Englert, A. Salzer, D. Vasen, Tetrahedron: Asymmetry 1998, 9, 1867; D. Vasen, A. Salzer, F.
Gerhards, H. J. Gais, R. Stürmer, N. H. Bieler, A. Togni, Organometallics 2000, 19, 539.
[12] a) D. Landnicer, C. R. Hauser, Org. Synth., Coll. Vol V 1955, 434; b) J. Schreiber, H. Maag, N.
Hashimoto, A. Eschenmoser, Angew. Chem., Int. Ed. 1971, 10, 330; c) S. Danishefsky, S.
Chackalamannil, P. Harrison, M. Silvestri, P. Cole, J. Am. Chem. Soc. 1985, 107, 2474.
[13] a) Bruker-Nonius, COLLECT, Program for Data Collection, 2002; b) Bruker-Nonius, EvalCCD,
Program for Data Reduction, 2002; c) Bruker AXS, SADABS, Program for Absorption Correction,
2002; d) Bruker AXS, SHELXTL NT 6.12, Program for Structure Determination, 2002; e) Bruker
AXS, SHELXTL NT 6.12, Program for Structure Refinement, 2002; f) Bruker AXS, SHELXTL NT
6.12, Program for Molecule Projection, 2002.
[14] a) H. D. Flack, Acta Crystallogr., Sect. A 1983, 39, 876; b) ꢁInternational Tables for Crystallographyꢀ,
Ed. A. J. C. Wilson, Kluwer Academic Publishers, Dordrecht, 1992, tables 6.1.1.4 (500 – 502), 4.2.6.8
(219 – 222), 4.2.4.2 (193 – 199); c) P. Coppens, in ꢁCrystallographic Computingꢀ, Ed. F. R. Ahmed,
S. R. Hall & C. P. Huber, Copenhagen (Munksgaard), 1970, p. 255 – 270.
Received August 30, 2006