Synthesis of Ferrocene Oxazoline N,O ligands and Their Application in Asymmetric Ethyl- and Phenylzinc Additions to Aldehydes

Article abstract of DOI:10.1021/acs.joc.5b01766

The synthesis of a range of novel gem-disubstituted ferrocene-oxazoline ligands and their application in both the asymmetric ethyl- and phenylzinc additions to aldehydes is reported. These studies reveal that gem-disubstitution of i-Pr-containing ferrocene oxazoline ligands results in increased enantioselectivity compared to their unsubstituted counterparts. Utilizing zinc catalysis, these ligands provided a wide range of secondary alcohols in yields of up to 93% with ee's of up to >99%. An interesting crystal structure of a ferrocene oxide-lithium tetramer showing lithium-nitrogen coordination in the solid state is also presented.

Full text of DOI:10.1021/acs.joc.5b01766

Subscriber access provided by UNIV OF LETHBRIDGE
Article
Synthesis of Ferrocene Oxazoline N,O ligands and their Application
in the Asymmetric Ethyl- and Phenylzinc Additions to Aldehydes.
Chris Nottingham, Robert Benson, Helge Müller-Bunz, and Patrick Jerome Guiry
J. Org. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.joc.5b01766 • Publication Date (Web): 01 Sep 2015
Downloaded from http://pubs.acs.org on September 7, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
The Journal of Organic Chemistry is published by the American Chemical Society.
1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 42
The Journal of Organic Chemistry
1
2
3
4
5
6
Synthesis of Ferrocene Oxazoline N,O ligands and their
Application in the Asymmetric Ethylꢀ and Phenylzinc Additions to
Aldehydes.
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Chris Nottingham, Robert Benson, Helge MüllerꢀBunz, and Patrick J. Guiry*
Centre for Synthesis and Chemical Biology, School of Chemistry and Chemical Biology, University College
Dublin, Belfield, Dublin 4, Ireland.
KEYWORDS: Asymmetric Catalysis, Ferrocene, Oxazoline, Ligand, Diethylzinc, Diphenylzinc
ABSTRACT: The synthesis of a range of novel gemꢀdisubstituted ferroceneꢀoxazoline ligands
and their application in both the asymmetric ethylꢀ and phenylzinc additions to aldehydes is
reported. These studies reveal that gemꢀdisubstitution of iꢀPrꢀcontaining ferrocene oxazoline
ligands results in increased enantioselectivity compared to their unꢀsubstituted counterparts.
Utilizing zinc catalysis, these ligands provided a wide range of secondary alcohols in yields of
up to 93% with ee’s of up to >99%. An interesting crystal structure of a ferroceneꢀoxideꢀlithium
tetramer showing lithiumꢀnitrogen coordination in the solid state is also presented.
Introduction
1ꢀ3
Oxazoline ligands such as PHOX represent a privileged ligand scaffold in transition metal
4ꢀ8
catalysis.
In most cases, tꢀBuꢀPHOX ligands provide the highest ee’s, however, the
synthesis of these ligands involves starting from the expensive, nonꢀnatural amino acids (S)ꢀ
9
and (R)ꢀtꢀleucine, which severely limits their application in asymmetric catalysis. With this in
mind, Paquin recently reported the synthesis and application of 5,5ꢀ(dimethyl)ꢀiꢀPrꢀPHOX as a
10,11
cheaper, practical equivalent of the expensive tꢀBuꢀPHOX.
This gemꢀdisubstitution effect
1
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 2 of 42
12,13
14
has recently been explored further by both our group
and the group of Stoltz with
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
excellent results in both cases. We sought to further investigate the general applicability of the
gemꢀdisubstitution effect with different oxazoline ligand architectures rather than the
conventional PHOX type backbone. As an example, we have recently prepared series of a
gemꢀdisubstituted iꢀPr HetPHOX ligands 1 and the gemꢀdisubstituted FcPHOX ligand 2 and
applied them in the asymmetric intermolecular Heck reaction, furnishing products in up to 98%
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
12
yield with ee’s of up to 97% (Figure 1). In addition, we have prepared tridentate
bis(oxazoline) ligands 3 and applied them in the zincꢀcatalyzed asymmetric FriedelꢀCrafts
13
reaction, providing products with ee’s of up to 95% (Figure 1).
Figure 1. HetPHOX, FcPHOX and bis(oxazoline) ligands employing gem-disubstitution.
Continuing on from these studies, we wished to investigate the synthesis and application of a
series of ferrocenyl hydroxyoxazoline ligands which are gemꢀdimethylꢀsubstituted analogues
15,16
of ligands previously developed by the groups of Bolm and Butenschön (Figure 2).
Ph
Ph
OH
O
O
R
N
OH
Ph
N
Fe
Fe
Ph
Rp)ꢀ4: R = tꢀBu (Bolm15)
(Sp)ꢀ4 (Bolm15)
(
16
(Rp)ꢀ5: R = iꢀPr (Butenschön )
2
ACS Paragon Plus Environment

Page 3 of 42
The Journal of Organic Chemistry
Figure 2. Bolm and Butenschön’s ferrocenyl hydroxyoxazoline ligands.
1
2
With this in mind we designed disubstituted ferrocene ligands (R )ꢀ6 and (S )ꢀ6 (Figure 3). We
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
p
p
planned to synthesize these N, O bidentate ligands from the cheap, natural, amino acid (S)ꢀ
valine. While both ligands contain the same central chirality, they differ in planar chirality. We
also envisioned that the addition of a bulky TMS group ortho to the oxazoline ring could further
influence the conformation of the oxazoline ring and may even direct the iꢀPr group closer to
the reaction center. This inspired the design of trisubstituted ligands (R )ꢀ7 and (S )ꢀ7, which
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
p
p
would allow us to examine the conformational effects of an additional TMS group (Figure 3).
Figure 3: Ligand class 6 - 7 and key chiral precursor 8
Furthermore, while many disubstituted ferrocene ligands have been applied in asymmetric
17,18
catalysis, relatively few trisubstituted ligands have been examined.
Previously, we have
shown that trisubstituted ferrocenyl pyrrolidine ligands 9aꢀd containing a TMS group perform
18
differently to their disubstituted counterparts (Figure 4).
3
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 4 of 42
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 4. Ferrocenyl pyrrolidine ligands 9a-d.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Results and discussion
Synthesis of ligand class 6 and 7
We have recently reported the high yielding synthesis of key intermediate 8 starting from
12
readily available (S)ꢀvaline methyl ester hydrochloride (Figure 3). Based on this and previous
19
work by Sammakia, directed ortho-lithiation of 8 with secꢀBuLi and chelating agent
tetramethylethylenediamine (TMEDA), followed by quenching with benzophenone provided
ligand (R )ꢀ6 in 83% yield with a diastereomeric excess (de) of >99% (Scheme 1).
p
Scheme 1: Synthesis of ligands (R )ꢀ6, (S )ꢀ6 and (S )-7.
p
p
p
Conditions: (a) secꢀBuLi, TMEDA, Et O, ꢀ78 °C, 10 min (b) Benzophenone, 0 °C ꢀ rt, 10 min (c) TMSCl, 0 °C ꢀ
2
rt, 10 min (d) nꢀBuLi, Et O, ꢀ78 °C, 10 min (e) TBAF, THF, reflux, 72 h.
2
4
ACS Paragon Plus Environment

Page 5 of 42
The Journal of Organic Chemistry
Directed ortho-lithiation of 8 followed by quenching with TMSCl provided oxazoline 10 in 93%
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
yield as a single diastereomer by H NMR spectroscopy which was then subjected to a second
ortho-lithiation with nꢀBuLi and subsequent quench with benzophenone to provide the
trisubstituted ferrocene oxazoline ligand (S )ꢀ7 in 88% yield with 99% de. The TMS group was
p
then removed from ligand (S )ꢀ7 by exposure to tetrabutylammonium fluoride (TBAF) to
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
p
provide disubstituted ferrocene oxazoline ligand (S )ꢀ6 in 79% yield with 99% de (Scheme 1).
p
Originally we thought that employing two equivalents of a lithium base would allow successful
deprotonation at the remaining ortho-position in ligand (R )ꢀ6 to provide ligand (R )ꢀ7 (Scheme
p
p
1
). However, this was not successful and is likely due to the nitrogen of the oxazoline ring
forming a seven membered chelate ring with the lithiumꢀalkoxide formed upon alcohol
deprotonation. This would lock the nitrogen in place and prevent directed ortho-metalation
(DOM) by a second equivalent of base. The next logical step was to protect the alcohol group
to determine if this allowed successful lithiation. However, the sterically hindered tertiary
20
alcohol group proved unreactive using various protection protocols. This was also the case
18
with similar ligands (9a-d) where protection proved unsuccessful. The choice of protecting
group was also largely limited due to the fact that subsequent deꢀprotection would have to take
place without concomicant removal of the installed TMS group afterwards. Due to these
limitations we decided to investigate an alternative route to ligand (R )ꢀ7 that avoided
p
protecting the hydroxy group (Scheme 2).
Scheme 2. Synthesis of ligand (R )-7
p
5
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 6 of 42
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
O
O
I
O
I
N
N
N
TMS
Fe
a, b
TMS
c
Fe
Fe
9
0%
97%
10
11
12
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
TMS
Fe
O
TMS
N
N
OH
Ph
a, d
e, f
Fe
Ph
8
0%
8
1%
de > 99%
13
(R )ꢀ7
p
Conditions: (a) nꢀBuLi, Et O, ꢀ78 °C, 10 min (b) I , 0 °C ꢀ rt, 20 min (c) TBAF, THF, reflux, 6 h (d) TMSCl, 0 °C ꢀ
2
2
rt, 10 min (e) secꢀBuLi, TMEDA, Et O, ꢀ78 °C, 10 min (f) Benzophenone, 0 °C ꢀ rt, 10 min.
2
Ortho-lithiation of disubstituted ferrocene oxazoline 10 with nꢀBuLi and subsequent quenching
with freshly sublimed iodine provided trisubstituted ferrocene 11 in 90% yield before
subsequent deꢀprotection utilizing TBAF to provide iodoꢀferrocene 12 in 97% yield. Iodoꢀ
ferrocene 12 was then subjected to lithiumꢀhalogen exchange with nꢀBuLi and subsequently
quenched with TMSCl to provide disubstituted ferrocene oxazoline 13 in 81% yield. A final
directed ortho-lithiation with secꢀBuLi and TMEDA followed by quenching with benzophenone
provided ligand (R )ꢀ7 in 80% yield and >99% de (Scheme 2).
p
Revised synthesis of ligand (R )-7
p
During the time period of our work detailed above Clayden and Arnott reported a method of
reversing the preferred sense of diastereoselectivity in ferrocene oxazoline directed ortho-
21
lithiations. This was accomplished using tridentate chelating agent 14 in place of TMEDA to
provide the other diastereomer in 75% de (Scheme 3).
21
Scheme 3 Access to both diastereomers of ferrocene iꢀPr oxazoline using different chelating agents.
6
ACS Paragon Plus Environment

Page 7 of 42
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We sought to apply this to our own work in order to shorten the synthetic sequence of ligand
(
R )ꢀ7 to just two steps from ferrocene oxazoline 8. In the original synthesis of chelating agent
p
14, the ditosylated intermediate 15 was treated with sodium tꢀbutoxide to provide 14 in 28%
21
yield. We found that substituting sodium tꢀbutoxide for the more nucleophilic and less basic
lithium tꢀbutoxide provided the chelating agent 14 in an improved yield of 64% (Scheme 4).
Scheme 4. Synthesis of chelating agent 14
With chelating agent 14 in hand we then carried out a directed ortho-lithiation on ferrocene
21
oxazoline 8 utilizing conditions described. Unfortunately, only a modest de of 37.5% was
obtained (Scheme 5). While the 37.5% de of 13 obtained may not be useful synthetically, it
does provide a hint as to the conformation of the iꢀPr group in our gemꢀdimethyl substituted
ligand precursor 8.
Scheme 5. Lithiation of ferrocene 8 with chelating agent 14
7
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 8 of 42
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Clayden and Arnott attempted diastereoselective lithiations of both iꢀPr and tꢀBu substituted
ferrocene oxazolines and found that, while the iꢀPr oxazoline provided product with 75% de,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
21
the tꢀBu substituted oxazoline yielded only a 1:1 diastereomeric mixture. Thus, comparing
the de values hints that the iꢀPr group of our ferrocene oxazoline 8 has adopted a
conformation halfꢀway inꢀbetween that of an iꢀPr group and a tꢀBu group.
Ferrocene oxazoline lithiation studies
During the course of our research we sought to examine the lithiation complex formed upon
deprotonation of ferrocene oxazoline 8 by an organolithium reagent. A nitrogen coordination
mechanism has been postulated and Sammakia has conducted studies in this area by
22
synthesizing a tethered ferrocene oxazoline and subjecting this to lithiation. The product
obtained from this reaction indicates a lithiumꢀnitrogen coordinated intermediate. Furthermore,
Sebesta recently reported a computational study on the lithiation of ferrocene oxazolines via a
nitrogen coordination pathway and showed that for ferrocene iꢀPrꢀoxazoline with TMEDA, the
ꢀ1
R lithiation transition state is more stable than the S by 19 kJ mol due to steric interactions
p
p
23
between the TMEDAꢀlithium chelate complex and the iꢀPr group. However, despite the many
publications utilizing ferrocene oxazolines and DOM, such a Nꢀcoordinated lithiation
intermediate has not been isolated. Therefore, we attempted to isolate and crystallize our
lithiated ferrocene intermediate by cooling the lithiated reaction mix to ꢀ78 °C for 3 days in
24
various solvent mixtures. We proposed that a solvated dimer such as 16 would be formed
upon deprotonation and thus hoped to isolate and crystallize this lithiated intermediate
(Scheme 6).
8
ACS Paragon Plus Environment

Page 9 of 42
The Journal of Organic Chemistry
Scheme 6. Lithiation and attempted crystallization of ferrocene oxazoline 8
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
After numerous attempts under various conditions this approach proved unsuccessful.
However, after adventitious exposure of the reaction mixture to air, crystallization occurred
rapidly, providing crystals as small red blocks. Xꢀray analysis of these crystals provided the
structure 18 shown in Scheme 6 and Figure 5.
9
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 10 of 42
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
8
Figure 5. X-ray crystal structure of oxidized lithiation intermediate 18.
This is a tetramer of an oxidized form of the proposed lithiation intermediate. We postulate that
dimer 17 may have formed upon addition of molecular oxygen to the proposed lithiated
intermediate 16 (Scheme 6). The oxidized intermediate 17 can then dimerize and crystallize as
the tetrameric oxygenꢀcontaining species 18 (Scheme 6, Figure 5). Despite the fact that this is
an oxidized intermediate of what we hoped to isolate, this is still the first example of any
ferrocene oxazoline showing lithiumꢀnitrogen coordination in the solid state. It should be noted
that the sense of preferred diastereoselection observed experimentally is in place for each
ferrocene in this structure. This structure adds further weight that the proposed lithiumꢀnitrogen
coordination mechanism is correct, supporting both Sammakia’s lithiation studies and
Sebesta’s computational work. Upon addition of water, the monomeric hydrolysis product of 18
1
could be detected by H NMR spectroscopy but unfortunately undergoes rapid decomposition
and could not be isolated.
10
ACS Paragon Plus Environment

Page 11 of 42
The Journal of Organic Chemistry
Diethylzinc Additions to Aldehydes
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
16
15,25
As both the iꢀPr (R )ꢀ5 and tꢀBu
(R )ꢀ and (S )ꢀ4 derivatives of the ferrocenyl
p p
p
hydroxyoxazoline ligands have been applied in the dialkylzinc addition to aldehydes, we chose
this as a suitable model reaction to evaluate our ligand class. This reaction has been
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
extensively studied and so is an excellent testing ground for new concepts in ligand
26,27
design.
Ligand (R )ꢀ6 was applied to diethylzinc additions utilizing benzaldehyde as a
p
model substrate in order to find optimal reaction conditions. Hexane provided superior ee’s to
toluene and was chosen as a suitable solvent while lower temperatures of up to ꢀ20 °C proved
optimal. (Scheme 7, Table 1). It is noteworthy that toluene was the optimal solvent for the tꢀBuꢀ
25
analogue (R )ꢀ4, with hexane reducing enantioselectivity. Applying 5 mol% of ligand (R )ꢀ6,
p
p
using hexane as solvent at ꢀ20 °C provided (R)ꢀ1ꢀphenylꢀ1ꢀpropanol with 88% ee (Table 1,
entry 6; Scheme 7).
Scheme 7
Table 1. Diethylzinc additions to benzaldehyde catalyzed by ligand (R )-6.
p
a
entry
solvent
toluene
toluene
hexane
hexane
hexane
hexane
hexane
temp [°C]
mol (%) (R
p
)ꢀ6
time [h]
yield [%]
96
ee [%]
1
2
3
4
5
6
7
0
10
10
10
10
10
5
4
80
ꢀ20
0
4
94
83
4
95
83
ꢀ20
ꢀ20
ꢀ20
ꢀ40
5
90
87
b
5
95
87
24
24
93
88
5
69
88
11
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 12 of 42
a
b
All reactions carried out on a 1 mmol scale Determined by HPLC analysis using a Chiralcel® OD column.
3
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
25
equiv. of diethylzinc were used. Configuration assigned by comparison with literature retention times.
We next investigated the performance of the remaining ferrocene ligands (S )ꢀ6, (R )ꢀ7 and
p
p
(
S )ꢀ7 in the transformation using our optimized reaction conditions to determine the
p
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
importance of planar chirality and the effect of an additional TMS group (Scheme 8).
Scheme 8. Diethylzinc addition to benzaldehyde catalyzed by ligands 6-7.
All reactions carried out on a 1 mmol scale. Ee’s were determined by HPLC using a Chiralcel® OD column.
From the results obtained it is clear that planar chirality is the dominant factor controlling
asymmetric induction with reversal of planar chirality between ligands (R )ꢀ6 and (S )ꢀ6
p
p
resulting in a 55% drop in ee with a coincident 61% drop in yield. It is noteworthy that both
R )ꢀ6 and (S )ꢀ6 provide similar results to the corresponding tꢀBu ligands (R )ꢀ4 and (S )ꢀ4,
(
p
p
p
p
25
which provided product in 83% yield, 93% ee and 55% yield, 35% ee respectively. The
additional TMS group present in ligands (R )ꢀ7 and (S )ꢀ7 has a pronounced effect on
p
p
catalysis, lowering both ee and yield in the case of ligand (R )ꢀ7 and providing nearly racemic
p
product in the case of ligand (S )ꢀ7. We had initially hoped that the increased steric bulk
p
12
ACS Paragon Plus Environment

Page 13 of 42
The Journal of Organic Chemistry
provided by the TMS group would result in higher enantioselectivities; however, it appears the
increased steric bulk is not beneficial. With these initial results we sought to further probe our
ligand design. From consideration of our proposed transition states (vide infra) we wanted to
investigate the effects of increased steric bulk beside the oxygen donor atom. From a brief
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
16,18
literature survey we found that this has previously resulted in increased ee’s.
To examine
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
this we synthesized triferrocenyl ligand (R )ꢀ19 in 64% yield via DOM of ferrocene oxazoline 8
p
and subsequent quench with diferrocenylketone (Scheme 9). Diferrocenylketone itself was
synthesized in 90% yield via a modified literature procedure which originally yielded only
28
2
7%.
Scheme 9. Synthesis of ligand (R )-19 via DOM of 8.
p
Applying 5 mol% of ligand (R )ꢀ19 using toluene as solvent at 40 °C provided (R)ꢀ1ꢀphenylꢀ1ꢀ
p
propanol in 79% yield with 93% ee after 7 h (optimized conditions, Table 2). The ee of 93% is
an improvement upon the 88% ee from ligand (R )ꢀ6 and matches the ee of Bolm’s tꢀBu FcOx
p
25
(R )ꢀ4 for benzaldehyde. With both optimal ligands (R )ꢀ6 and (R )ꢀ19 in hand, a variety of
p
p
p
other aldehydes were tested in order to examine the substrate scope (Table 2, Scheme 10).
Scheme 10. Diethylzinc addition to aldehydes
Table 2. Diethylzinc addition to various aldehydes catalyzed by ligands (R )-6 and (R )-19.
p
p
13
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 14 of 42
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(R )ꢀ6 conditions, hexane, ꢀ20 °C. (R )ꢀ19 conditions, toluene, 40 °C. ee’s were determined by chiral HPLC or
p
p
SFC, configuration determined by comparison with HPLC literature values, or tentatively assigned by assumption
of an identical reaction pathway.
Comparing these results with those obtained from the corresponding tꢀBu (R )ꢀ4 and iꢀPr FcOx
p
(
R )ꢀ5 analogues provides insight into how effective the gemꢀdimethyl iꢀPr unit is at mimicking
p
15,16,25
a tꢀBu substituent.
With benzaldehyde and pꢀanisaldehyde, our gemꢀdimethyl based
ligand (R )ꢀ6 provides ee’s almost exactly inꢀbetween those provided by the iꢀPr and tꢀBu
p
analogues (R )ꢀ5 and (R )ꢀ4. With 4ꢀchlorobenzaldehyde, no significant changes in ee were
p
p
noted while for transꢀcinnamaldehyde the results are closer to those of the iꢀPr analogue (R )ꢀ
p
5
. With ferrocene carboxaldehyde, the ee from our gemꢀdimethyl ligand (R )ꢀ6 is nearly
p
identical to that of the tꢀBu based ligand (R )ꢀ4. From these results it is clear that the
p
conformation of ligand (R )ꢀ6’s iꢀPr group is different to those of both the iꢀPr and tꢀBu
p
analogues (R )ꢀ5 and (R )ꢀ4. This information corresponds with earlier lithiation studies, where
p
p
the diastereoselectivity upon lithiation of our gem-dimethyl ligand fell exactly half way between
those observed for the corresponding iꢀPr and tꢀBu analogues (vide supra, Scheme 5).
14
ACS Paragon Plus Environment

Page 15 of 42
The Journal of Organic Chemistry
Triferrocenyl ligand (R )ꢀ19 provides higher ee’s than ligand (R )ꢀ6 and effectively mimics
p
p
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Bolm’s tꢀBu analogue (R )ꢀ4 for benzaldehyde, p-anisaldehyde and transꢀcinnamaldehyde
p
while offering a significant improvement in enantioselectivity for 4ꢀchlorobenzaldehyde. For
ferrocene carboxaldehyde, the sterically congested ligand (R )ꢀ19 provides slightly lower ee’s
p
than both ligand (R )ꢀ6 and the tꢀBu analogue (R )ꢀ4. Prompted by these results, we also
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
p
p
wanted to examine our ligands performance with an ortho-substituted aldehyde and an
18,26,27
aliphatic aldehyde, both of which generally provide lower levels of enantioselectivity.
With 1ꢀnaphthaldehyde, ligand (R )ꢀ6 provided product in 81% ee while ligand (R )ꢀ19 gave
p
p
9
1% ee showing that high levels of enantioselectivity were possible with orthoꢀsubstituted
aldehydes. With cyclohexane carboxaldehyde, ligand (R )ꢀ6 provided an impressive ee of 96%
p
while ligand (R )ꢀ19 furnished product with only 38% ee. The high ee obtained with
p
cyclohexane carboxaldehyde is noteworthy, as in most cases aliphatic aldehydes provide
much lower ee’s than aromatic aldehydes.
Diphenylzinc Addition to Aldehydes
Encouraged by these results and wanting to further examine our ligands performance, we
decided to employ our ligand class in the more challenging phenylzinc addition to
18,29ꢀ32
aldehydes.
This transformation facilitates the asymmetric synthesis of diarylmethanols,
the main alternative to which is asymmetric reduction of diaryl ketones, however, this reduction
33
methodology is highly substrate dependent. Bolm and coꢀworkers reported that using a
mixed ethylphenylzinc reagent formed in-situ from diethylꢀ and diphenylꢀzinc resulted in
29
selective phenyl transfer and provided increased ee’s compared with diphenylzinc alone.
With this in mind, ligand (R )ꢀ6 was applied in several phenyl additions utilizing 4ꢀ
p
chlorobenzaldehyde as a model substrate in order to find optimal reaction conditions (Table 3,
Scheme 11).
Scheme 11. Phenyl addition to 4-chlorobenzaldehyde catalyzed by ligand (R )ꢀ6
p
15
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 16 of 42
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 3. Phenyl addition to 4-chlorobenzaldehyde catalyzed by ligand (R )-6.
p
a
entry
temp [°C]
mol % ligand
Ph Zn/ Et Zn equiv.
Yield [%]
ee [%]
2
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
10
10
10
5
1.5/0
83
82
88
89
72
5
0.65/1.3
0.65/1.3
92
10
5
92
b
RT
0.65/1.3
91
a
All reactions carried out on a 0.2 mmol scale for 24 h Determined by SFC using a Chiralpak®
b
IB column. Room temperature was ~18 °C. Configuration determined by comparison of the
34
optical rotation with literature value.
Optimal conditions were found by applying 5 mol% of ligand (R )ꢀ6 with a mixed
p
ethylphenylzinc reagent using toluene as solvent at 10 °C, providing 1ꢀ(4’ꢀchlorophenyl)ꢀ1ꢀ
propanol in 82% yield with 92% ee (Table 3, entry 2). We next investigated the performance of
the remaining ferrocene ligands (S )ꢀ6, (R )ꢀ7, (S )ꢀ7 and (R )ꢀ19 using our optimized reaction
p
p
p
p
conditions to determine the effect of changing planar chirality and the presence of an
additional TMS group (Scheme 12).
Scheme 12. Phenyl addition to 4-chlorobenzaldehyde catalyzed by ligands 6 and 7.
16
ACS Paragon Plus Environment

Page 17 of 42
The Journal of Organic Chemistry
O
OH
Ph
1
2
3
4
5
6
7
8
9
ZnEt /ZnPh (0.65/1.3 equiv.)
2
2
Ligand (5 mol%)
H
Toluene, 10 °C, 24 h
Cl
Cl
Sp
Sp
O
TMS
O
Ph
OH
Ph
Ph OH
5
Ph
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
5
1
4
N
N
N
4
5
5
N
OH
Ph
1
1
3
2
OH
Ph
4
4
3
2
O
O
Fe
Fe
3
2
3
2
TMS
Ph
Fe
Fe
Ph
Rp
Rp
(
R )-6
(R )-7
p
p
(S )-6
(S )-7
p
p
8
9
2% yield
2% ee
48% yield
85% ee
81% yield
ꢀ68% ee
32% yield
ꢀ30% ee
All reactions carried out on a 0.2 mmol scale. ees were determined by SFC using a Chiralpak® IB column.
As with the ethylzinc addition, it is clear that planar chirality is the dominant factor controlling
asymmetric induction with reversal of planar chirality between ligands (R )ꢀ6 and (S )ꢀ6
p
p
reversing the stereochemical outcome of the reaction. This is particularly interesting as in the
diethylzinc addition to aldehydes reversal of the ligands planar chirality did not reverse the
stereochemical outcome, but merely lowered the yield and ee. In terms of comparison, ligand
(
R )ꢀ6 yields similar results to the corresponding tꢀBu ligand (R )ꢀ4, which provided product in
p p
29
9
2% yield with 95% ee under the same conditions. Unfortunately, ligand (S )ꢀ6’s tꢀBu
p
analogue (S )ꢀ4 was not tested and thus, a direct comparison cannot be made. The additional
p
TMS group present in ligands (R )ꢀ7 and (S )ꢀ7 has a pronounced effect on catalysis, lowering
p
p
both ee’s and yield’s compared to their disubstituted counterparts (R )ꢀ6 and (S )ꢀ6. Similar to
p
p
the diethylzinc addition, it appears that the increased steric bulk from the TMS group is in fact
detrimental. Bulky triferrocenyl ligand (R )ꢀ19 provides an unusual result, forming product in
p
9
3% yield but with no enantioinduction. However, this is in agreement with another
16
triferrocenyl system prepared by Butenschön and tested by Bolm. We next tested a variety of
17
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 18 of 42
other aldehydes in the ethylphenylzinc addition with optimal ligand (R )ꢀ6 in order to examine
p
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the substrate scope (Table 4, Scheme 13).
Scheme 13. Phenyl addition to aldehydes catalyzed by ligand (R )ꢀ6
p
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 4. Phenyl addition to various aldehydes catalyzed by ligand (R )-6.
p
OH
Ph
OH
Ph
OH
Ph
OH
Ph
MeO
(R )ꢀ6 80%, 94% ee
Cl
72%, 93% ee
86%, 98% ee
78%, 92% ee
86%, 97% ee
65%, 87% ee
97%, 90% ee
p
29(
R )ꢀ4 82%, 98% ee
p
OH
OH
Ph
Br
OH
Ph
Ph
OH
Ph
Fe
(
29(
R )ꢀ6 68%, 82% ee
77%, 90% ee
79%, >99% ee
53%, 99% ee
p
R )ꢀ4 64%, 91% ee
p
Ee’s were determined by chiral HPLC or SFC, configuration determined by comparison with HPLC literature
values, or tentatively assigned by assumption of an identical reaction pathway.
Again we will compare our results for ligand (R )ꢀ6 with those obtained from the corresponding
p
tꢀBu analogue (R )ꢀ4 to provide insight into how effective the gemꢀdimethyl iꢀPr group is at
p
29
mimicking the tꢀBu substituent (Table 4).
For pꢀanisaldehyde, pꢀtolualdehyde, 4ꢀ
chlorobenzaldehyde and transꢀcinnamaldehyde, ligand (R )ꢀ6 provides very similar results to
p
its tꢀBu analogue (R )ꢀ4. With ortho-substituted 2ꢀbromobenzaldehyde, ligand (R )ꢀ6 provides
p
p
a lower ee than its tꢀBu counterpart (R )ꢀ4 (82% vs. 91%). However, with 1ꢀnaphthaldehyde,
p
ligand (R )ꢀ6 provides product in 90% ee, showing that high levels of enantioselectivity are
p
18
ACS Paragon Plus Environment

Page 19 of 42
The Journal of Organic Chemistry
possible with some ortho-substituted aldehydes. With ferrocene carboxaldehyde and
cyclohexane carboxaldehyde, which gave the best results in the diethylzinc addition, we
obtained excellent ee’s of >99% and 99% respectively.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
X-Ray Crystal Structures
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
To enable a better understanding of our results and to fully examine the impact of gemꢀ
disubstitution, we attempted to obtain the solidꢀstate structures of all of our ligands. However,
despite numerous attempts, a single crystal of (R )ꢀ7 suitable for xꢀray analysis could not be
p
obtained. The crystal structure of ligands (R )ꢀ6, (S )ꢀ6, (S )ꢀ7 and (R )ꢀ19 are shown in Figure
p
p
p
p
6
.
(R )ꢀ6
(R )ꢀ19
p
p
(
S )ꢀ7
p
(
S )ꢀ6
p
19
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 20 of 42
Figure 6. X-ray crystal structures of; ligand (R )-6 (top left), ligand (R )-19 (top right), ligand (S )-6 (bottom
p
p
p
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
left) and ligand (S )-7 (bottom right).
p
In all structures we observe a seven membered chelate ring with the hydroxy proton strongly
coordinated to the oxazoline nitrogen as expected. A detailed analysis of ligand (R )ꢀ6 was
p
deemed important as solidꢀstate structures of both the iꢀPr and tꢀBu analogues (R )ꢀ5 and
p
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
16,25
(
R )ꢀ4 have been reported.
By comparing the three structures it is clear that the iꢀPr group
p
in ligand (R )ꢀ6 has adopted a solidꢀstate conformation inꢀbetween that of an iꢀPr and a tꢀBu
p
with the slight distortion of the oxazoline ring forcing the iꢀPr group slightly closer to the metal
center (Figure 7).
Figure 7. Overlay of crystal structures showing distortion of the oxazoline ring and rotation of the i-Pr
25
group in ligand (R )-6 (left), and ligand (S )-7 (right). Order: (R )-4 (Blue, bottom) , (R )-5 (Green,
p
p
p
p
16
middle) , (R )-6 (left) or (S )-7 (right) (Red, top).
p
p
This conformation is closer to that of an iꢀPr than a tꢀBu group but whether this is dominant in
solution is unknown. A similar motif is present in ligands (S )ꢀ6 and (R )ꢀ19 with both solidꢀ
p
p
state structures possessing a slightly distorted oxazoline ring and iꢀPr group. In ligand (S )ꢀ7
p
the introduction of an additional TMS group had a dramatic effect on the conformation of the iꢀ
Pr group compared to the corresponding disubstituted ligand (S )ꢀ6. The oxazoline ring has
p
distorted only slightly but this minor distortion has resulted in the iꢀPr group adopting a
conformation similar to that of a tꢀBu group, with the iꢀPr group facing directly into the axial
phenyl group (Figure 7). This is noteworthy as it shows how a small change in the
conformation of the oxazoline ring can result in the iꢀPr group adopting a dramatically
20
ACS Paragon Plus Environment

Page 21 of 42
The Journal of Organic Chemistry
different conformation. It is possible that in solution the ligands adopt conformations similar to
this, rather than those observed in the solid state, which would explain why the enantiomeric
excesses obtained by ligand (R )ꢀ6 are so similar to those induced by the tꢀBu ligand (R )ꢀ4.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
p
p
However, from preliminary 2D NOESY experiments no interactions that could help elucidate
the solution phase conformation of the oxazoline ring were found.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Transition states
Based on our Xꢀray crystal structures, we propose similar transition states to those proposed
39
40
41,42
43
by previous investigators such as Noyori , Norrby , Corey
sense of asymmetric induction in the addition of ethyl and phenyl groups to aldehydes (Figure
).
and Houk to rationalize the
8
Figure 8. Possible transition states for the zinc alkoxides of our (R ) chiral ligands (R )ꢀ6, (R )-7 and (R )-
p
p
p
p
19. The terms syn and anti define the relationship between the transferring alkyl and the ferrocene
backbone of the bidentate ligand. Cis and trans define which aldehyde lone pair coordinates to the
21
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 22 of 42
catalytic zinc chelated by the amino alcohol ligand. Inv implies inversion of configuration on the catalytic
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
zinc.
For ligands (R )ꢀ6 and (R )ꢀ19, we propose anti-cis is the dominant transition state and the
p
p
major source of the (R)ꢀalcohol while syn-trans is partially blocked due to steric repulsion
between the ferrocene backbone and the aldehyde. All four inv transition states are blocked to
some extent by the bulky aryl group on the top face of the alkoxide oxygen preventing inv
coordination of the dialkylzinc, although when Ar = Ph, as in ligands 6 and 7, this blocking may
be imperfect. As syn-cis is completely blocked by the ferrocene backbone this leaves the anti-
trans transition state as a main source of (S)ꢀalcohol. This helps explain why having a tꢀBu
group, rather than an iꢀPr group, at C(4) of the oxazoline ring generally leads to higher ee’s of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(R)ꢀalcohol. In this case the steric repulsion between the C(4) group of the oxazoline and the
R’ group of the aldehyde is the only inhibitor to this transition state. Thus, a bulkier group at
C(4) will more successfully suppress this transition state and lead to increased ee’s of the (R)ꢀ
alcohol. These transition states also explain why ligand (R )ꢀ19 provides better ee’s than
p
ligand (R )ꢀ6. The inv transition states leading to the (S)ꢀalcohol are less hindered than those
p
leading to the (R)ꢀalcohol due to the position of the aldehydes R’ group, thus, the inv transition
states are likely an overall contributor to (S)ꢀalcohol formation. However, in ligand (R )ꢀ19 the
p
bulky ferrocenyl groups will block the inv transition states much more effectively then the
phenyl groups in ligand (R )ꢀ6, suppressing (S)ꢀalcohol formation and resulting in increased
p
ee’s for ligand (R )ꢀ19 (93% compared to 88% for ethyl transfer). For ligands (S )ꢀ6 and (S )ꢀ7
p
p
p
we propose similar transition states but with reversal of planar chirality on the ferrocene’s Cp
ring (Figure 9).
22
ACS Paragon Plus Environment

Page 23 of 42
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 9. Possible transition states for the zinc alkoxides of our (S ) chiral ligands (S )ꢀ6 and (S )-7. The
p
p
p
terms syn and anti define the relationship between the transferring alkyl and the ferrocene backbone of
the bidentate ligand. Cis and trans define which aldehyde lone pair coordinates to the catalytic zinc
chelated by the amino alcohol ligand. Inv implies inversion of configuration on the catalytic zinc.
For the diethylzinc addition, ligand (S )ꢀ6 provided (R)ꢀalcohol in 33% ee, this can be
p
explained by the mismatched blocking resulting from the iꢀPr group now being syn to the axial
phenyl group. Ideally, the blocking provided by the ferrocene backbone and the C(4)
substituent on the oxazoline should be matched, as in the (R ) ligands. For ligands (S )ꢀ6 and
p
p
(S )ꢀ7 the iꢀPr group and phenyl group block all of the inv transition states to some extent with
p
the inv states leading to the (S)ꢀalcohol encountering more steric hindrance due to the
aldehydes bulky R’ group facing the iꢀPr group of the ligand. Syn-cis is completely blocked by
the lower ferrocene ring, leaving the unhindered anti-trans as the major source of (R)ꢀalcohol.
Anti-cis is a minor source of (S)ꢀalcohol with steric repulsion between the C(4) group of the
oxazoline and the R’ group of the aldehyde suppressing this transition state. However, the
major source of (S)ꢀalcohol comes from syn-trans, which experiences only a small amount
23
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 24 of 42
of steric repulsion between the lower ferrocene ring and the aldehydic proton. Overall, this
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
results in a small ee of the (R)ꢀalcohol in the case of ligand (S )ꢀ6 and nearly racemic product
p
in the case of the more sterically congested ligand (S )ꢀ7.
p
The results from the mixed ethylphenylzinc addition are particularly interesting because in this
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
reaction ligands (S )ꢀ6 and (S )ꢀ7 provided predominantly (S)ꢀalcohol in good (68%) and
p
p
moderate (30%) ee’s respectively (Scheme 12). This inversion of stereochemistry simply by
changing the R group on the zinc is unprecedented. Based on previous studies we assume a
mixed ethylphenylzinc regent is formed and that the catalytic zinc is still bonded to an ethyl,
36ꢀ38
rather than a phenyl group.
We postulate that this bulkier mixed regent must destabilize
the anti-trans conformation and suppress formation of the (R)ꢀalcohol. Ligand (R )ꢀ19 also
p
behaves in a curious manner, providing excellent ee’s in the diethylzinc addition but yielding
only racemic product in the mixed ethylphenylzinc addition. From the crystal structure of ligand
(R )ꢀ19 we can see that the ferrocenyl group equatorial to the ligand backbone is pointing
p
forward, directly into the space where the stoichiometric zinc needs to coordinate (Figure 6).
We propose that when the bulkier ethylphenylzinc reagent is employed this undergoes steric
interactions with the equatorial ferrocenyl group and leads to destabilization of the transition
states hindering effective chiral induction.
Finally, we wanted to examine the effect of replacing the ferrocene backbone of our ligand with
a simple phenyl ring. Previous reports indicate that this lowers ee’s compared to the parent
44
ferrocenyl backbone but Bolm and coꢀworkers reported that this had little effect on their tꢀBu
25
FcOx ligand (R )ꢀ4. This led to the design of ligand 20, which is a phenylꢀbased derivative of
p
ligands (R )ꢀ6 and (S )ꢀ6. Not only would its use in catalysis provide insight into our transition
p
p
states, but, an Xꢀray crystal structure of such a ligand would be inherently interesting for
comparative purposes. Ligand 20 was synthesized in an acceptable yield for testing purposes
24
ACS Paragon Plus Environment

Page 25 of 42
The Journal of Organic Chemistry
from the known gem-dimethyl precursor 21 via lithium/halogen exchange followed by
quenching with benzophenone (Scheme 14).
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 14. Synthesis of ligand 20 via lithium/halogen exchange of 21
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Application of ligand 20 in the diethylzinc addition to benzaldehyde under the same conditions
as ligand (R )ꢀ6 provided 1ꢀphenylꢀ1ꢀpropanol in 83% yield but with only 67% ee. As the
p
25
corresponding phenylꢀtꢀBuꢀoxazoline ligand provided product in 92% ee this result was
disappointing, but did correlate with our transition states, which assign an important role to the
ferrocene backbone in suppressing certain conformations (Figure 8). To provide further insight
an Xꢀray crystal structure of ligand 20 was obtained (Figure 10).
Figure 10. X-ray crystal structure of ligand 20, front view (left) and side view (right).
From this structure we can see that the gemꢀdimethyl effect is not apparent in the solid state,
furthermore, the oxazoline ring is highly distorted. This is likely due to the more acute angle
between the two substituents bonded to the phenyl ring compared to the Cp ring of ferrocene
(
123.51º between C14ꢀC19ꢀC20 in ligand 20 vs 127.16º between C2ꢀC1ꢀC6 in ligand (R )ꢀ6).
p
This results in the nitrogen and oxygen donor atoms being closer together in 20, which is
disfavored due to the large size of the seven membered chelate ring. In order to
25
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 26 of 42
accommodate this, the oxazoline ring distorts to provide some extra distance between the
donor atoms resulting in the conformation we see in Figure 10.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Conclusions
We have reported the design, synthesis and Xꢀray characterization of a total of six new ligands
and their application in the ethyl and/or phenyl addition to both aromatic and aliphatic
aldehydes providing secondary alcohols with excellent ee’s of up to >99%. Application of the
gemꢀdimethyl effect was successful, with our more economical ligands inducing similar levels
of enantioselectivity to the corresponding expensive tꢀBu ligands in the majority of cases. From
our studies we have determined that planar chirality is the dominant factor controlling
asymmetric induction with reversal of planar chirality lowering ee’s in the ethylzinc addition and
reversing stereoselectivity in the phenylzinc addition. We have also demonstrated that
trisubstituted ferrocene oxazoline ligands performed worse than their disubstituted
counterparts in terms of both yield and ee. Furthermore, our lithiation studies yielded an
interesting crystal structure of a ferrocenylꢀoxideꢀlithium tetramer that shows for the first time,
the proposed lithiumꢀnitrogen coordination in the solid state. The excellent ee’s of up to >99%
obtained for phenyl addition to both aromatic and aliphatic aldehydes showcase our ligands as
among the best available for these transformations while being readily synthesized from cheap
starting materials.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Experimental Section
General Information. Unless otherwise noted, all commercial reagents were used as received
without further purification. Anhydrous diethyl ether (Et O), tetrahydrofuran (THF) and
2
dichloromethane (CH Cl ) were obtained from a dry solvents dispenser. Aldehydes were
2
2
purified either by distillation, formation of the bisulfite derivative and regeneration with TMSCl
45,46
or simply by washing with 10% aqueous Na CO .
2
3
26
ACS Paragon Plus Environment

Page 27 of 42
The Journal of Organic Chemistry
General procedure 1 (GP1) for ferrocene oxazoline lithiation. Ferrocene oxazoline
1.0 equiv) was added to a dried, nitrogen flushed Schlenk tube containing a magnetic stirring
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
bar and dissolved in anhydrous Et O (2 mL/mmol). TMEDA (1.2 equiv) was added where
2
required and the reaction mixture cooled to ꢀ78 °C before being treated dropwise with the
appropriate organolithium reagent (1.1 equiv). After being stirred for 10 min at this
temperature, the reaction was warmed to 0 °C and quenched with the electrophile (1.0 equiv).
The resulting solution was allowed to warm to RT over 10 ꢀ 15 min before being quenched with
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
saturated aqueous NaHCO and the layers separated. The organic extracts were washed with
3
brine, dried (MgSO ), filtered and concentrated.
4
1
-[(S)-4-Isopropyl-5,5-dimethyl-oxazolinyl]-2-(S )-(trimethylsilyl)ꢀferrocene - 10. GP1 was
p
followed utilizing ferrocene oxazoline 8 as the substrate on a 0.65 mmol scale, secꢀBuLi as the
organolithium reagent, TMEDA as a chelating ligand and TMSCl as the electrophile.
Purification by column chromatography (pentane:EtOAc, 19:1) provided the disubstituted
1
ferrocene oxazoline 10 (239.6 mg, 93%, one diastereomer by HꢀNMR) as a viscous red oil
20
D
which solidified on standing: Mp 92 – 94 °C (Pentane); [α]
= +89.4 (c = 1.00, CHCl ); IR
3
ꢀ1 1
(film) νmax 3108 (m), 2970 (m), 1655 (s), 1458 (m) cm ; HꢀNMR (400 MHz, CDCl ) δ 4.92 –
3
4
1
.86 (m, 1H), 4.46 – 4.37 (m, 1H), 4.27 – 4.21 (m, 1H), 4.18 (s, 5H), 3.26 (d, J = 8.6 Hz, 1H),
.89 – 1.76 (m, 1H), 1.52 (s, 3H), 1.37 (s, 3H), 1.12 (d, J = 6.5 Hz, 3H), 0.99 (d, J = 6.5 Hz,
13
1
3H), 0.33 (s, 9H); C{ H}ꢀNMR (101 MHz, CDCl ) δ 164.2, 85.9, 80.9, 77.0, 76.7, 73.2, 72.7,
3
7
1.6, 69.5, 29.27, 29.26, 21.6, 21.1*, 0.8. (*HSQC accounts for overlap of both Me carbons
+
from the iꢀPr group showing at 21.1); HRMS: (ESIꢀTOF) calculated for C H FeNOSi [M + H]
21
32
398.1603, found 398.1620. TLC (pentane:EtOAc, 5:1), Rf: 0.60 (UV, Vis).
-[(S)-4-Isopropyl-5,5-dimethyl-oxazolinyl]-2-(S )-(trimethylsilyl)-5(S )-
1
p
p
(
diphenylmethanol)-ferrocene - (S )-7. GP1 was followed utilizing ferrocene oxazoline 10 as
p
the substrate on a 0.22 mmol scale, nꢀBuLi as the organolithium reagent and benzophenone
as the electrophile. Purification by column chromatography (pentane:EtOAc, 19:1) provided
27
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 28 of 42
the trisubstituted ferrocene oxazoline (S )ꢀ7 (111.6 mg, 88% yield, 99% de) as a viscous red oil
p
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
which solidified on standing. Recrystallization from pentane at RT led to the formation of red
20
D
crystals that proved suitable for Xꢀray crystal analysis: Mp 131 – 135 °C (Pentane); [α]
=
+
193.7 (c = 1.02, CHCl ); IR (film) ν 3058 (m), 2960 (m), 2900ꢀ3100 (br), 1644 (s), 1411 (m)
3 max
ꢀ1 1
cm ; HꢀNMR (400 MHz, CDCl ) δ 9.66 (s, 1H), 7.54 – 7.47 (m, 2H), 7.36 – 7.19 (m, 3H), 7.16
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
–
7.05 (m, 5H), 4.29 (s, 5H), 4.11 (d, J = 2.6 Hz, 1H), 3.71 (d, J = 2.5 Hz, 1H), 3.22 (d, J = 8.7
Hz, 1H), 1.46 (s, 3H), 1.20 – 1.06 (m, 1H), 0.81 (s, 3H), 0.75 (d, J = 5.1 Hz, 3H), 0.73 (d, J =
13
1
5.4 Hz, 3H), 0.30 (s, 9H); C{ H}ꢀNMR (101 MHz, CDCl ) δ 167.1, 149.7, 147.0, 127.9, 127.7,
3
1
2
27.3, 127.2, 126.6, 126.3, 105.6, 87.9, 86.6, 79.9, 76.8, 74.3, 74.2, 72.4, 70.7, 29.0, 28.5,
+
0.9, 20.7, 20.7, 1.2; HRMS: (ESIꢀTOF) calculated for C H FeNO Si [M + H] 580.2334,
34
42
2
found 580.2347. TLC (pentane:EtOAc, 5:1), Rf: 0.73 (UV, Vis).
-[(S)-4-Isopropyl-5,5-dimethyl-oxazolinyl]-2-(S )-(diphenylmethanol)-ferrocene - (S )-6.
1
p
p
Ferrocene oxazoline ligand (S )ꢀ7 (288 mg, 0.477 mmol, 1.0 equiv) was added to a dry,
p
nitrogen flushed, 25 mL Schlenk tube equipped with a magnetic stirring bar and dissolved in
1
0 mL of THF. TBAF (5 mL, 1 M solution in THF, 10 equiv) was then added and the resulting
solution heated to reflux with stirring for 72 h before being cooled to RT, partitioned with H O
2
(
10 mL) and separated before back extracting the H O with Et O (15 mL). Purification by
2 2
column chromatography (pentane:Et O, 9:1) provided the disubstituted ferrocene oxazoline
2
ligand (S )ꢀ6 (199 mg, 79%, 99% de) as a viscous red oil which solidified on standing.
p
Recrystallization from pentane at RT led to the formation of red crystals that proved suitable for
20
D
Xꢀray crystal analysis: Mp 123 ꢀ 124 °C (Pentane); [α]
= +225.6 (c = 1.25, CHCl ); IR (film)
3
ꢀ1
1
νmax 3087 (m), 2968 (m), 2921 (m), 3000 (br), 1644 (s), 1447 (m) cm ; HꢀNMR (400 MHz,
CDCl ) δ 9.49 (s, 1H), 7.55 – 7.48 (m, 2H), 7.35 – 7.18 (m, 3H), 7.17 – 7.00 (m, 5H), 4.71 (s,
3
1
H), 4.27 (s, 5H), 4.21 (s, 1H), 3.67 (s, 1H), 3.32 (d, J = 6.1 Hz, 1H), 1.42 (s, 3H), 1.35 – 1.31
13
1
(
m, 1H), 1.13 (s, 3H), 0.74 (d, J = 6.6 Hz, 3H), 0.31 (d, J = 6.5 Hz, 3H); C{ H}ꢀNMR
(
101 MHz, CDCl ) δ 166.3, 149.5, 146.6, 127.9, 127.7, 127.3, 127.2, 126.6, 126.4, 101.1,
3
28
ACS Paragon Plus Environment

Page 29 of 42
The Journal of Organic Chemistry
8
6.6, 79.2, 77.2, 74.9, 70.6, 70.2, 67.7, 67.1, 29.6, 28.8, 21.1, 21.0, 18.4; HRMS: (ESIꢀTOF)
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
+
calculated for C H FeNO [M + Na] 530.1758, found 530.1752. TLC (pentane:EtOAc, 5:1),
31 33
2
Rf: 0.85 (UV, Vis).
-[(S)-4-isopropyl-5,5-dimethyl-oxazolinyl]-2-(R )-(diphenylmethanol)-ferrocene - (R )-6.
1
p
p
GP1 was followed utilizing ferrocene oxazoline 8 as the substrate on a 2.32 mmol scale, secꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
BuLi as the organolithium reagent, TMEDA as a chelating ligand and benzophenone as the
electrophile. Purification by column chromatography (pentane:Et O, 9:1) provided the
2
disubstituted ferrocene oxazoline (R )ꢀ6 (0.982 g, 83%, >99% de) as a viscous red oil which
p
solidified on standing. Recrystallization from pentane at RT led to the formation of red crystals
20
D
that proved suitable for Xꢀray crystal analysis: Mp 127 – 130 °C (Pentane); [α]
= ꢀ197.4 (c =
ꢀ1 1
0.98, CHCl ); IR (film) νmax 3085 (m), 2970 (m), 3000 (br), 1644 (s), 1447 (m) cm ; HꢀNMR
3
(400 MHz, CDCl ) δ 9.21 (s, 1H), 7.49 – 7.42 (m, 2H), 7.26 – 7.11 (m, 3H), 7.11 – 6.98 (m,
3
5
H), 4.61 (s, 1H), 4.20 (s, 5H), 4.13 (s, 1H), 3.60 (s, 1H), 2.71 (d, J = 9.0 Hz, 1H), 1.79 – 1.65
13
1
(m, 1H), 1.25 (s, 3H), 1.19 (s, 3H), 1.01 (d, J = 6.5 Hz, 3H), 0.82 (d, J = 6.6 Hz, 3H); C{ H}ꢀ
NMR (101 MHz, CDCl ) δ 166.0, 149.4, 146.4, 127.9, 127.5, 127.2, 127.1, 126.6, 126.2,
3
101.3, 86.7, 79.5, 77.3, 74.8, 70.7, 70.1, 67.7, 67.4, 28.9, 28.5, 21.5, 21.3, 20.8; HRMS: (ESIꢀ
+
TOF) calculated for C H FeNO [M + H] 508.1939, found 508.1957. TLC (pentane:EtOAc,
31
34
2
5
:1), Rf: 0.82 (UV, Vis).
1-[(S)-4-Isopropyl-5,5-dimethyl-oxazolinyl]-2-(R )-(iodo)-5-(S )-(trimethylsilyl)-ferrocene -
p
p
11. GP1 was followed utilizing ferrocene oxazoline 10 as the substrate on a 0.20 mmol scale
and nꢀBuLi as the organolithium reagent. Freshly sublimed I was used as the electrophile. No
2
TMEDA was added. An addition wash with 10% aqueous Na S O was performed to remove
2
2
3
any remaining I . Purification by column chromatography (pentane:Et O, 9:1) provided the
2
2
trisubstituted ferrocene oxazoline 11 (93.1 mg, 90%) as a viscous orange oil which quickly
20
D
solidified on standing: Mp 96 – 98 °C (Pentane); [α]
= +6.0 (c = 0.96, CHCl ); IR (film) ν
3
max
ꢀ1
1
3
087 (m), 2924 (m), 1646 (s), 1422 (m), 1265 (m) cm ; HꢀNMR (400 MHz, CDCl ) δ 4.67 –
3
29
ACS Paragon Plus Environment