Manipulating the diastereoselectivity of ortholithiation in planar chiral ferrocenes

Article abstract of DOI:10.1021/ol4013734

The sense of asymmetric ortholithiation directed by a chiral oxazoline may be inverted simply by the choice of achiral ligand. Comparison of results with a number of ferrocenyl oxazoline derivatives suggests that lithiation takes place by coordination to the oxazoline nitrogen irrespective of the ligand used.

Full text of DOI:10.1021/ol4013734

ORGANIC
LETTERS
2013
Vol. 15, No. 13
3334–3337
Manipulating the Diastereoselectivity of
Ortholithiation in Planar Chiral Ferrocenes
Simon A. Herbert,†,‡ Dominic C. Castell,† Jonathan Clayden,‡ and Gareth E. Arnott*,†
Department of Chemistry and Polymer Science, University of Stellenbosch, Matieland,
7602, South Africa, and School of Chemistry, University of Manchester, Oxford Road,
Manchester M13 9PL, U.K.
arnott@sun.ac.za
Received May 16, 2013
ABSTRACT
The sense of asymmetric ortholithiation directed by a chiral oxazoline may be inverted simply by the choice of achiral ligand. Comparison of
results with a number of ferrocenyl oxazoline derivatives suggests that lithiation takes place by coordination to the oxazoline nitrogen irrespective
of the ligand used.
Ferrocenes have been extensively studied and are the
subject of much topical research, particularly since their
use in material1,2 and even medicinal chemistry3 has
been widespread. Significantly, the planar chiral aspect
of ferrocene substitution has been an important feature
of their success.4 Control of the absolute planar chiral
configuration of ferrocenyl systems has been achieved in
a number of ways, typically involving asymmetric lithia-
tion of the ferrocene through the use of chiral directing
groups.2a,5 One of the most common of these methods
has been the use of chiral oxazolines, first described in 1995
simultaneously by the groups of Richards,6 Uemura,7 and
Sammakia.8 This method involves the use of an alkyllithium
base directed by a chiral oxazoline to diastereoselectively
deprotonate one of the ferrocenyl ortho-protons. By quench-
ing the reaction with a wide range of electrophiles, diaste-
reomer 2a can be obtained, which can be further modified
downstream for a multitude of purposes, for example the
synthesis of the valuable ferrocenylphosphine (Fc-Phox)1a,9
and organosilanol10 ligands. Sammakia has developed
the method of choice, showing that sec-butyllithium in
conjunction with N,N,N0,N0-tetramethylethylenediamine
† University of Stellenbosch.
‡ University of Manchester.
(1) Selected review articles: (a) Sutcliffe, O. B.; Bryce, M. R. Tetra-
hedron: Asymmetry 2003, 14, 2297–2325. (b) Atkinson, R. C. J.; Gibson,
ꢀ
V. C.; Long, N. J. Chem. Soc. Rev. 2004, 33, 313–328. (c) Arrayas, R. G.;
Adrio, J.; Carretero, J. C. Angew. Chem., Int. Ed. 2006, 45, 7674–7715.
(d) Butler, I. R. Eur. J. Inorg. Chem. 2012, 4387–4406.
(2) Selected books: (a) Ferrocenes: Ligands, Materials and Biomole-
cules; Stepnicka, P., Ed.; Wiley: Chichester, 2008. (b) Chiral Ferrocenes in
Asymmetric Catalysis: Synthesis and Applications; Dai, L.-X., Hou, X. L.,
Eds.; Wiley-VCH: Weinheim, 2010. (c) Ferrocenes: Compounds, Proper-
ties & Applications (Chemical Engineering Methods and Technology);
Philips, E. S., Ed.; Nova Science Publishers: New York, 2011.
(3) Selected review articles: (a) Fouda, M. F. R.; Abd-Elzaher,
M. M.; Abdelsamaia, R. A.; Labib, A. A. Appl. Organomet. Chem.
2007, 21, 613–625. (b) Hillard, E. A.; Jaouen, G. Organometallics 2011,
30, 20–27. (c) Roux, C.; Biot, C. Future Med. Chem. 2012, 4, 783–797.
(4) Selected recent examples of planar chirality in ferrocenes: (a)
Metallinos, C.; John, J.; Zaifman, J.; Emberson, K. Adv. Synth. Catal.
2012, 354, 602–606. (b) Barreiro, E. M.; Broggini, D. F. D.; Adrio, L. A.;
White, A. J. P.; Schwenk, R.; Togni, A.; Hii, K. K. Organometallics
2012, 31, 3745–3754. (c) Buergler, J. F.; Niedermann, K.; Togni, A.
Chem.;Eur. J. 2012, 18, 632–640. (d) Zirakzadeh, A.; Schuecker, R.;
Gorgas, N.; Mereiter, K.; Spindler, F.; Weissensteiner, W. Organome-
tallics 2012, 31, 4241–4250. (e) Eitel, S. H.; Bauer, M.; Schweinfurth, D.;
(5) Clayden, J. Top. Organomet. Chem. 2003, 5, 251–286.
(6) (a) Richards, C. J.; Damalidis, T.; Hibbs, D. E.; Hursthouse,
M. B. Synlett 1995, 1995, 74–76. (b) Richards, C. J.; Mulvaney, A. W.
Tetrahedron: Asymmetry 1996, 7, 1419–1430.
(7) (a) Nishibayashi, Y.; Uemura, S. Synlett 1995, 79–81. (b)
Nishibayashia, Y.; Segawaa, K.; Arikawaa, Y.; Ohea, K.; Hidaib, M.;
Uemura, S. J. Organomet. Chem. 1997, 545ꢀ546, 381–398.
(8) (a) Sammakia, T.; Latham, H. A.; Schaad, D. R. J. Org. Chem.
1995, 60, 10–11. (b) Sammakia, T.; Latham, H. A. J. Org. Chem. 1995,
60, 6002–6003. (c) Sammakia, T.; Latham, H. A. J. Org. Chem. 1996, 61,
1629–1635.
(9) Geisler, F. M.; Helmchen, G. J. Org. Chem. 2006, 71, 2486–2492.
€
Deibel, N.; Sarkar, B.; Kelm, H.; Kruger, H.-J.; Frey, W.; Peters, R.
J. Am. Chem. Soc. 2012, 134, 4683–4693. (f) Gao, D.-W.; Shi, Y.-C.; Gu,
Q.; Zhao, Z.-L.; You, S.-L. J. Am. Chem. Soc. 2013, 135, 86–89.
€
(10) Ozc-ubukc-u, S.; Schmidt, F.; Bolm, C. Org. Lett. 2005, 7, 1407–
1409.
r
10.1021/ol4013734
Published on Web 06/25/2013
2013 American Chemical Society

(TMEDA) and a noncoordinating solvent gave a superb
diastereomeric ratio (dr) of >99:1 (Scheme 1).8b
The mechanism has been proposed to be controlled by
N-coordination of the oxazoline, but the wide range of
diastereoselectivities reported under different conditions11
suggests a much more complicated process in fine balance
between four factors: (1) the group attached to the chiral
center of the oxazoline, (2) the alkyllithium used, (3) the
choice of solvent, and (4) the choice of ligand (although in
the case of ferrocenes only TMEDA and glyme have been
examined).8b
Although the so-called ‘minor’ diastereomer (2b) is
accessible via a stepwise protection/deprotection method
using a trimethylsilyl group (Scheme 1),6b,12 we wondered
whether the diastereoselectivity of the lithiation itself
might be perturbed in such a way as to favor the direct
formation of 2b.13 Herein we report our results, which
show that this is indeed possible. They also allow us to
contribute significant observations to the topical discus-
sion of this mechanism.14
Figure 1. Ligands examined in this work.
Table 1. Ortholithiation Results with Oxazoline Ferrocene 1
with Various Ligands
It is thought that the reason TMEDA successfully
improves the diastereoselectivity of the ortholithiation of
chiral ferrocenes resides in its ability to chelate and break
down the alkyllithium complexes that exist in solution.15
That said, it seemed surprising that other ligands which are
known to do the same thing had not been explored in this
system, particularly given that different ligands have been
shown to have a profound effect in the diastereoselective
ortholithiation of a calixarene oxazoline.13c Thus a selec-
tion of chelating ligands 3ꢀ8 was examined for their ability
to influence the diastereoselectivity of the ortholithiation
of ferrocenyl oxazoline 1 (Figure 1).
Scheme 1. Current Methods for Obtaining Both Diastereomers
of Oxazoline Ferrocenes
a Quantified using HPLC traces; conversion is relative to starting
material. b At ꢀ86 °C with 2.5 equiv of ligand. c Synthesized via a
reported method.16
reported when n-butyllithium and TMEDA (3) were used
(Table 1, entry 1).17 This established a baseline dr value for
comparison with the other ligands.
Preliminary exploratory work found that pentane was
not a suitable general solvent, particularly when using
the other ligands. Toluene was found to be a compatible
noncoordinating solvent and returned a better dr to that
Ligands 4ꢀ7 (entries 2ꢀ5) were all less effective than
TMEDA in promoting high dr values, confirming the
singularly unique and fascinating relationship TMEDA
has had with alkyllithiums. Interestingly the diglyme
(DGME) ligand (8a) showed a promising, albeit small,
preference for the opposite diastereomer 2b (entry 6).
Attempting the same reaction but with sec- or tert-
butyllithium switched the selectivity back toward the
conventional diastereomer 2a (entries 7 and 8), leading
us to suspect that the reversal in diastereoselectivity
may be intimately tied to the steric environment of the
(11) Some examples for isopropyl ferrocenyl oxazoline: n-BuLi/THF
(2:1);8b n-BuLi/ether (6:1);8b s-BuLi/ether (39:1);7b s-BuLi/hexanes/
TMEDA (>99:1);8b t-BuLi/hexanes/TMEDA (28:1).8b
(12) Ahn, K. H.; Cho, C.-W.; Baek, H.-H.; Park, J.; Lee, S. J. Org.
Chem. 1996, 61, 4937–4943.
(13) Very few examples of reagent-induced reversal of diastereo-
selectivity in ortholithiation reactions exist. In a chromium arene, see:
(a) Overman, L. E.; Owen, C. E.; Zipp, G. G. Angew. Chem., Int. Ed.
2002, 41, 3884–3887. In a 1,10-bis-oxazoline ferrocene, see: (b) Park, J.;
Lee, S.; Ahn, K. H.; Cho, C.-W. Tetrahedron Lett. 1996, 37, 6137–6140.
In a calixarene, see: (c) Herbert, S. A.; Arnott, G. E. Org. Lett. 2010, 12,
4600–4603.
ꢀ
(16) Garcıa, J. I.; Garcıa-Marın, H.; Mayoral, J. A.; Perez, P. Green
´ ´ ´
(14) Chadwick, S. T.; Ramirez, A.; Gupta, L.; Collum, D. B. J. Am.
Chem. Soc. 2007, 129, 2259–2268.
Chem. 2010, 12, 426–434.
(17) nBuLi/TMEDA in either diethyl ether or hexanes has been
reported to give a 100:1 (2a:2b) diastereomeric ratio.8b
ꢀ
(15) Saa, J. Helv. Chim. Acta 2002, 85, 814–840.
Org. Lett., Vol. 15, No. 13, 2013
3335

alkyllithiumꢀligand complex. We therefore targeted sym-
metrical derivatives of DGME (8bꢀf) to perturb the steric
environment of the lithiating complex and investigate their
influence on the reaction. Ligand 8b was commercially
available, while ligands 8cꢀf could be synthesized using a
Williamson ether synthesison the bis-tosylate of diethylene
glycol (Scheme 2).18
Scheme 3. Reversal of Diastereoselectivity in a Oxazoline
Calix[4]arene System
Scheme 2. Synthesis of DGME-Based Ligands
Sammakia and Latham have used a conformationally
restricted ferrocenyl oxazoline to show that only an
N-coordination mode occurs during ortholithiation,8c
even though computationally the difference between the
two mechanisms is small.14 The results reported here
prompt the question as to whether the DGME ligands
are promoting a switch to an O-coordination mechanism,
which may account for the reversal of diastereosel-
ectivity.13c Information regarding N- vs O-coordination
was obtained by looking at differently substituted chiral
oxazolines, i.e. by studying the effect of increasing and
decreasing the size of the R-group on the chiral center, and
shifting its position from next to the N-atom to next to the
O-atom.21 To this end five ferrocenyl oxazolines were
synthesized (11ꢀ15, Scheme 4) in excellent yields, except
cis-diphenyl-oxazoline ferrocene 15 which appeared to be
prone to decomposition. These were then treated with
n-butyllithium and THF, TMEDA (3), or di-tert-butyl-
DGME (8e) as ligands (Table 2).
The results paint an intriguing picture. Decreasing the
R0-group’s steric bulk from tert-butyl, to isopropyl, to
methyl (entries 1ꢀ3, Table 2) shows a correlation with
decreased dr only when THF is used. With both TMEDA
and di-tert-butyl-DGME (8e) no simple relationship be-
tween selectivity and steric bulk was observed. Curiously,
with di-tert-butyl-DGME (8e) the change from ferrocenyl
tert-butyl oxazoline 11 to isopropyl oxazoline 1 results in a
significant change in dr, but what should amount to a
significant reduction of steric strain with methyl oxazoline
12 has no effect whatsoever on the dr. The yields of the
reactions with di-tert-butyl-DGME (8e) are also instruc-
tive, showing that decreasing the steric bulk on the oxazo-
line increases the rate of the reaction, this being consistent
with an N-coordination mechanism.22 Furthermore, it
would be expected that moving the methyl group from
being R to the N-atom (12) to being R to the O-atom (13)
a 8f synthesized from bis-mesylate.
Applying these ligands to the ortholithiation of oxazo-
line ferrocene 1 showed what initially appeared to be a
trend of increasing selectivity for diastereomer 2b as the
R-group on the DGME ligand increased in steric size
(entries 9ꢀ12). However bis(tri(3-ethylpentyl))-DGME
(8f) failed to effect further improvement in the dr (entry 13),
leading us to suggest that the reaction had reached the
limits of the steric strain imposed on the system.19 How-
ever, by optimizing the number of equivalents of ligand 8e
and lowering the temperature of the reaction toꢀ86 °C, we
were able to demonstrate an optimized 1:7 (= 12:88) ratio
in favor of diastereomer 2b (entry 14).20 Although this
ratio is a far cry from the >99:1 diastereoselectivity when
using sec-butyllithiumꢀTMEDA, it is remarkable that the
diastereoselectivity can be manipulated this far. Syntheti-
cally the switch from >99:1 to 12:88 allows either diaste-
reoisomer of a functionalized ferrocene product to be made
in up to 87% yield by just a change of an achiral ligand.
The effect is potentially general for other diastereoselec-
tive oxazoline-directed metalations. For example, switch-
ing TMEDA to 8e in the lithiation of an oxazoline
substituted calixarene (Scheme 3) led to an inversion of
the diastereoselectivity from 1:1413c to 9:1, again demon-
strating a remarkable reversal of selectivity based solely on
the choice of achiral ligand.
In principle the oxazoline may direct the ortholithiation
of the ferrocene via either the N- or O-atom.14 However,
(18) Low yields of the di-tert-butyl-DGME (8e) and bis(tri(3-
ethylpentyl))-DGME (8f) ligands were obtained due to a competing
elimination reaction giving the monovinyl ether. Whilst the divinyl ether
was presumed to have been formed, it was undetected owing to its very
low boiling point of 28 °C. Attempts to improve the yield by conducting
the reaction in diethyl ether or toluene failed, returning similar results to
those performed in THF.
(19) Attempting to correlate a linear free energy relationship with the
steric bulk of the R-groups on the diglyme ligands was also unsuccessful,
even when applying recent developments in this area; see: Harper, K. C.;
Bess, E. N.; Sigman, M. S. Nat. Chem. 2012, 4, 366–374.
(20) From a practical point of view it was also found that the ligand
could be isolated from the reaction mixture by protonating the ferrocene
products with 3 M HCl and extracting the ligand into diethyl ether. The
recovered ligand could then be distilled and reused.
(21) Whilst the tethered approach of Sammakia would seem ideal to
examine this, it suffers from a 13-step, low yielding (4%) synthetic
process, making it undesirable for multiple experiments. We were also
concerned with whether restricting the rotation of the oxazoline would
impede subtle conformational aspects that only manifest with free
rotation.
(22) It has been shown that the increased steric bulk of the group R to
the nitrogen atom reduces the reaction rate; see refs 8b and 23.
3336
Org. Lett., Vol. 15, No. 13, 2013

Table 2. Results of Mechanistic Probe Reactions
Scheme 4. Chiral Ferrocenyl Oxazolines Synthesized for a
Mechanistic Probe
a Conversion and dr quantified using HPLC traces; conversion is
relative to starting material. b Reactions were conducted in THF as
solvent. c Reaction reported in literature.8b
would improve the dr if an O-coordination mechanism was
important. Entry 4 (Table 2) shows the converse, with the
diastereoselectivity of the reaction being effectively shut
down in all cases. Lastly, ferrocenyl oxazolines 14 (Ph,Ph
trans) and 15 (Ph,Ph cis) have the same configuration R to
the N-atom but opposite configurations R to the O-atom.
They might therefore be expected to give very different
outcomes in the ortholithiation reaction should O-coordi-
nation be important. It was however found that the dr
values were essentially the same (entries 5 and 6), support-
ing an N-coordination mechanism.
is used; or (2) that the DGME nBuLi complex acts as a
3
bulky Lewis acid coordinating to the oxazoline nitrogen,
blocking rather than allowing deprotonation. The ortho-
hydrogen next to the oxazoline oxygen would then be free
for deprotonation.
In conclusion we have reported on a new ligand for
alkyllithium chemistry that in the case of oxazoline direc-
ted asymmetric ortholithiation shows a remarkable ability
to manipulate the diastereoselectivity from conventionally
accepted values. We have also lent further experimental
support to the N-coordination theory for these types of
reactions.
These results have shed some light on the mechanistic
aspects of this reaction. The strong evidence for the
absence of an O-coordination mechanism suggests that
the role of the DGME ligands is not to change to coordi-
nation site of the organolithium base. Rather, they must
exert their influence in another way: i.e., either (1) the
Acknowledgment. We thank Stellenbosch University
for funding; S.A.H. thanks Sasol, the Harry Crossley
Foundation, and the Commonwealth Scholarship for
funding. D.C.C. thanks the Harry Crossley Foundation
for funding.
initial DGME nBuLi23 species coordinates to the N-atom
3
of the oxazoline and forms a prelithiation complex which
then finds a lowest energy pathway for deprotonation that
is different (effectively opposite) from that when TMEDA
Supporting Information Available. Experimental pro-
cedures and characterization data are included. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(23) No data exists regarding the nature of this species, but it is more
than likely monomeric in solution. For DGME LiHMDS monomer
3
data, see: Lucht, B. L.; Bernstein, M. P.; Remenar, J. F.; Collum, D. B.
J. Am. Chem. Soc. 1996, 118, 10707–10718.
The authors declare no competing financial interest.
Org. Lett., Vol. 15, No. 13, 2013
3337