Evaluation of enantiopure and non-enantiopure Co(III)-salen catalysts and their counter-ion effects in the hydrolytic kinetic resolution (HKR) of racemic epichlorohydrin

Article abstract of DOI:10.1016/j.molcata.2012.08.013

Hydrolytic kinetic resolution (HKR) resolves racemic epoxides using water as the nucleophile and is most often catalyzed by chiral Co(III)-salens. ¹ Previous studies have shown that the counter-ion of the Co(III)-salen has a direct effect on rate of HKR; when catalyzed by a 50:50 mix of (R,R)-Co(III)-salen-OH and (R,R)-Co(III)-salen-SbF₆, the fastest HKR rates occurred. It has further been shown that the enantioselectivity is primarily associated with the reaction of (R,R)-Co(III)-salen-OH on the activated epoxide. Based on the aforementioned origin of selectivity, a catalyst containing a 50:50 mix of (R,R)-Co(III)-salen-OH and (±)-trans-Co(III)- salen-SbF₆ could, in principle, give high activities and enantioselectivities for HKR comparable to a mixed counter-ion system containing both (R,R)-Co(III)-salens. Here, a series of experiments are described that demonstrate that highly selective catalysis is only achieved using 100% enantiopure ligand, and mixtures of (R,R)-Co(III)-salen and (±)-trans- Co(III)-salen yield lower activity and selectivity. Control experiments demonstrate this is due to rapid counter-ion scrambling under reaction conditions, precluding the possibility of effectively co-utilizing enantiopure (expensive) and racemic (inexpensive) catalysts with differing counter-ions. The mechanistic investigations resolving the counter-ion scrambling are consistent with the currently accepted mechanism for catalysis, involving cooperative activity of two Co(III)-salen species that activate the epoxide and water in the reaction.

Full text of DOI:10.1016/j.molcata.2012.08.013

Journal of Molecular Catalysis A: Chemical 366 (2013) 1–7
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Editor’s Choice paper
Evaluation of enantiopure and non-enantiopure Co(III)-salen catalysts and their
counter-ion effects in the hydrolytic kinetic resolution (HKR) of racemic
epichlorohydrin
a
b
a,b,∗
Rebecca E. Key , Krishnan Venkatasubbaiah , Christopher W. Jones
a
School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, GA 30332-0400, USA
School of Chemical & Biomolecular Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0100, USA
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 1 June 2012
Received in revised form 14 August 2012
Accepted 16 August 2012
Available online 28 August 2012
Hydrolytic kinetic resolution (HKR) resolves racemic epoxides using water as the nucleophile and is most
often catalyzed by chiral Co(III)-salens. Previous studies have shown that the counter-ion of the Co(III)-
salen has a direct effect on rate of HKR; when catalyzed by a 50:50 mix of (R,R)-Co(III)-salen-OH and
1
(R,R)-Co(III)-salen-SbF6, the fastest HKR rates occurred. It has further been shown that the enantioselec-
tivity is primarily associated with the reaction of (R,R)-Co(III)-salen-OH on the activated epoxide. Based on
the aforementioned origin of selectivity, a catalyst containing a 50:50 mix of (R,R)-Co(III)-salen-OH and
Keywords:
Co(III)-salen
(± )-trans-Co(III)-salen-SbF6 could, in principle, give high activities and enantioselectivities for HKR com-
parable to a mixed counter-ion system containing both (R,R)-Co(III)-salens. Here, a series of experiments
are described that demonstrate that highly selective catalysis is only achieved using 100% enantiopure
ligand, and mixtures of (R,R)-Co(III)-salen and (± )-trans-Co(III)-salen yield lower activity and selectivity.
Control experiments demonstrate this is due to rapid counter-ion scrambling under reaction conditions,
precluding the possibility of effectively co-utilizing enantiopure (expensive) and racemic (inexpensive)
catalysts with differing counter-ions. The mechanistic investigations resolving the counter-ion scram-
bling are consistent with the currently accepted mechanism for catalysis, involving cooperative activity
of two Co(III)-salen species that activate the epoxide and water in the reaction.
Chirality
Counter-ion exchange
Polymer resin-support
Hydrolytic kinetic resolution (HKR)
©
2012 Elsevier B.V. All rights reserved.
1
. Introduction
makes enantiopure epoxides a useful chiral building block for
numerous reactions [12]. In addition, the coproduct of the reac-
tion, a chiral diol, can be a useful building block as well, being used,
for example, in the synthesis of active pharmaceutical ingredients
(APIs) [12,13]. Also with respect to pharmaceutical significance,
macrocyclic Co(III)-salen-PF6 has been used as a catalyst to syn-
thesize the chiral drugs (R)-mexiletine and (S)-propranolol via
HKR [14]. From a catalysis science perspective, the Co(III)-salen
catalyzed HKR reaction has become a useful test-bed for the devel-
opment of supported catalysts that enhance cooperative catalysis,
as the HKR reaction is now the prototypical cooperative bimolecu-
lar reaction catalyzed by a metal–ligand complex (vida infra) [15].
The HKR of epoxides was originally developed by Jacobsen,
Metal-salen catalysts are now ubiquitous in catalysis and
organic synthesis [1–3], being used in a variety of enantioselective
reactions, including olefin epoxidation [4–6], epoxide ring-opening
[
7–9], nitroaldol reactions [10], and conjugate additions [11],
2
as recent examples. The hydrolytic kinetic resolution (HKR) of
racemic epoxides is a well-established technique to resolve epox-
ides, specifically terminal epoxides, into a single enantiomer via a
ring-opening reaction using water as the nucleophile (Scheme 1)
[
7]. The HKR of terminal epoxides has attracted widespread interest
for a number of reasons. From a synthetic chemistry perspec-
tive, the strained, reactive, and three-membered ring functionality
using chiral Co(III)-salen-benzoate complexes [7] [Co(III)-salen =
ꢀ
(
R,R)-(N,N )-bis(3,5-di-tert-butylsalicylidene)-1,2-
cyclohexanediaminocobalt(III)] as the precatalyst (Fig. 1) [16].
∗ Corresponding author at: School of Biomolecular Engineering, Georgia Institute
of Technology, Atlanta, GA 30332-0400, USA. Tel.: +1 404 385 1683;
fax: +1 404 894 2866.
−
Using Co(III)-salen-X (X = nucleophilic counter-ions, such as Cl ,
−
OAc) complexes, Jacobsen and Blackmond showed that the
E-mail address: christopher.jones@chbe.gatech.edu (C.W. Jones).
Co(III)-salen = (R,R)-(N,N )-bis(3,5-di-tert-butylsalicylidene)-1,2-
reaction typically follows a second-order, cooperative bimetallic
pathway (where two cobalt centers participate in the rate-
determining step of the reaction) [16–19]. In this proposed
mechanism, one Co(III)-salen (presumed to have the original
1
ꢀ
cyclohexanediaminocobalt(III).
Abbreviations: HKR: hydrolytic kinetic resolution Co(III)-salen: (R,R)-(N,N )-
bis(3,5-di-tert-butylsalicylidene)-1,2-cyclohexanediaminocobalt(III).
2
ꢀ
1
381-1169/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcata.2012.08.013

2
R.E. Key et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 1–7
Scheme 1. HKR of racemic epichlorohydrin using a Co(III)-salen catalyst [7].
(
X =/ OH) via a nucleophilic attack (by the counter-ion X) on the
epoxide (Scheme 2, black pathway) [16]. If X is a good nucleophile,
such as Cl , the counter-ion will rapidly add to the epoxide
−
in the black pathway, and water is suggested to produce the
hypothetical Co(III)-salen-OH. With an accumulation of Co(III)-
salen-OH occurring in the HKR, the catalytic activity decreases
because Co(III)-salen-OH is a poorer Lewis acid than Co(III)-salen-
−
X (X =/ OH) [16]. If X = OTs, HKR catalysis is highly efficient, as
counter-ion addition is reversible, because Co(III)-salen-OTs is not
suggested to completely convert to Co(III)-salen-OH [24]. If X is
−
non-nucleophilic, such as SbF6 , there should be no nucleophilic
attack of the counter-ion on the epoxide, thus no generation of
Co(III)-salen-OH [16]. As a result, a cooperative, bimetallic pathway
is not suggested to take place to a significant extent (Scheme 2,
red pathway) [16], and the HKR is suggested to be catalyzed
primarily via a less selective, monometallic pathway (Scheme 2,
blue pathway) [17]. Based on the above mechanistic picture, it has
been demonstrated that with a 1:1 ratio of Co(III)-salen-OH and
Co(III)-salen-SbF , the fastest rates for the HKR occur, with the
6
rapid formation of the bimetallic transition state, thus allowing
the enantioselective, bimetallic pathway to dominate [16,17].
As mentioned previously, the salen catalysts used for HKR are
typically enantiopure, which is required to successfully produce
a highly enantiopure product. However, Nielsen et al. suggested
that each epoxide enantiomer bound to Co(III)-salen-X with the
same affinity [16]. This suggests that the binding of an epoxide
to Co(III)-salen-X is not a strongly enantiodiscriminating step and
that the high enantioselectivity of the HKR is predominantly due
to the selective reaction resulting from the nucleophilic Co(III)-
salen-OH attacking one of the Co(III)-salen-X-epoxide complexes
[
16]. This is consistent with the suggestion that the monometallic
Fig. 1. Co(III)-salen precatalysts used in this work, built from enantiopure (1,2) and
racemic (3,4) ligands [16].
path (Scheme 2, blue pathway) [17] is less enantioselective than
the cooperative pathway (Scheme 2, blue + red pathway) [16,17],
although the still moderate enantioselectivity achieved using only
non-coordinating counter-ions suggests (i) there is some selectivity
associated with epoxide binding, or (ii) the bimetallic pathway still
operates when only trace amounts of Co(III)-salen-OH species are
produced [16]. Based on these findings, it might be inferred that the
chirality of the Co(III)-salen-X (X = non-nucleophilic counter-ion) is
less important or even unimportant in determining the enantiose-
lectivity of the HKR reaction. To probe this hypothesis, we evaluate
here the HKR of epoxides using salen complexes constructed from
counter-ion, X) activates the electrophilic epoxide, while the other
−
Co(III)-salen (presumed to have an OH counter-ion) activates the
nucleophilic water (Scheme 2, red pathway) [16,18,19]. Nielsen
et al. demonstrated that the nucleophilicity of the Co(III)-salen
counter-ion (Fig. 1) [16,17] had a direct and important effect on the
rate of HKR [16]. At low catalyst loadings, it has been said that there
is a less-selective, monometallic pathway that competes with the
bimetallic pathway [20]. Taking the nature of the Co(III)-salen
counter-ion into consideration, Jain et al. suggested that a less
selective, monometallic pathway (Scheme 2, blue pathway) [17]
couples with the bimetallic pathway (Scheme 2, red pathway) [16]
when Co(III)-salens with highly non-nucleophilic counter-ions
are present in the reaction. For the bimetallic pathway, the hypo-
thetical Co(III)-salen-OH3 is formed in situ from Co(III)-salen-X
−
−
−
−
a variety of counter-ions (Cl , OH , SbF
enantiopure and racemic salen ligands. If each complex is stable
under reaction conditions, a catalyst containing a 50:50 mix of
and OAc) using both
6
(
R,R)-Co(III)-salen-OH and (± )-trans-Co(III)-salen-SbF may give
6
good activity and enantioselectivity for HKR of epoxides that is
orientation [22]. Due to this orientation, the Co(III)-salen-OH can attack one epoxide
enantiomer over the other [21]. Recently, Vinck et al. proposed that a small fraction
of the Co(III)-salen-OAc species in HKR is a Co(III)-bound phenoxyl radical that is
formed via air oxidation of Co(II) with acetic acid present (Scheme 3) [23]. This
radical, which has one to two acetates ligated to the complex, is proposed to have
formed via a two-electron, two-proton reduction of oxygen to hydrogen peroxide,
thus producing a delocalized phenoxyl radical that is stabilized by the bulky tert-
butyl groups located at the ortho- and para-positions of the phenolate [23]. The
radical was characterized by pulsed and continuous-wave electron paramagnetic
resonance (EPR), as well as resonance Raman (RR) [23].
3
The Co(III)-salen-OH is a hypothetical species, first proposed by Nielsen et al [16].
This active species has yet to be isolated or conclusively characterized. Kemper et al.
used in situ solution NMR and quantum chemical calculations to study coordination,
conformational, and electronic properties of Co(III)-salens to better understand the
enantioselectivity of HKR [21]. They determined that for the bimetallic transition
state exemplified by the two Co(III)-salens in HKR, the nucleophilic Co(III)-salen-
OH exists as a pentacoordinated, high spin, paramagentic complex, and the Co(III)-
salen-X (X =/ OH) that activates the epoxide exists as a hexacoordinated, low spin,
diamagnetic complex [21]. The two Co(III)-salens are aligned in a “head-to-tail”

R.E. Key et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 1–7
3
Scheme 2. Mechanism showing the presence of proposed monometallic (blue) and bimetallic (blue coupled with red) pathways in HKR [16,17]. (For interpretation of the
references to color in this scheme legend, the reader is referred to the web version of the article.)
O
OAc
-
HO2
N
O
N
O
O
Co
AcOH
N
O
N
O
N
O
N
O
N
O
N
+
O2
Co
H
-e
Co
Co
- O2
O
(R,R)-Co(II)-Salen*
AcOH
H O
2 2
OAc
H
AcO
AcO
(
R,R)-Co(III)-Salen-Bound
Phenoxyl Radical**
*(R,R)-Co(II)-Salen =
N
N
Co
O
O
*
*(R,R)-Co(III)-Salen-Bound =
Phenoxyl Radical; X = OAc
N X
Co
X
N
O
O
Scheme 3. Formation of the proposed Co(III)-bound phenoxyl radical that is formed in small amounts via air oxidation of Co(II)-salen in the presence of acetic acid [23].
comparable to a catalyst containing a 50:50 mix of (R,R)-Co(III)-
To this end, mixed (R,R)-Co(III)-salen-OH and (± )-trans-Co(III)-
salen-SbF6 catalysts are explored here for the HKR of racemic
epichlorohydrin. This approach is not found to yield highly enan-
tioselective catalysis, and the causes for the reduced selectivity are
salen-OH and (R,R)-Co(III)-salen-SbF . This 50% decrease in use
6
of the more expensive chiral ligand [25] could provide signifi-
cant practical advantages for Co(III)-salen catalyzed HKR reactions.

4
R.E. Key et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 1–7
enumerated here, shedding further insight into the nature of the
HKR reaction pathways.
2.2.3. Synthesis of polymer resin-supported
(R,R)-CO(III)-Salen-OAc (8)
Catalyst 8 was prepared according to previously published
procedures [26–28]. First, 1 equiv. of 7 was put into a 25 mL pear-
shaped flask equipped with a triangular stirbar. Dichloromethane
was then added to the flask (1 mL per 5 mg 7). Subsequently,
2
. Experimental
5
equiv. of glacial acetic acid was added until the solid changed
2
.1. General
from orange to light brown. The reaction mixture was then stirred
at room temperature open to air for 30 min. The dichloromethane
and excess acetic acid was removed using a rotovap (T < 60 C).
The recovered catalyst 8 was dried for 1 h under high vacuum and
immediately used for HKR reactions. For EA, catalyst 8 was dried
under high vacuum overnight.
Reagents were purchased from commercial vendors and used
◦
as received unless otherwise stipulated. All starting materials
were dried under high vacuum before use. Styrene and para-
dichlorobenzene were dried over calcium hydride and purified by
vacuum distillation before use. Divinylbenzene (DVB) was washed
with 5% NaOH (2×) and deionized water (2×) to remove polymer-
2.3. Synthesis of styrene/DVB resin
ization inhibitors, and then it was dried with anhydrous MgSO
4
1
prior to use. H NMR spectra were obtained using a Varian Mer-
Styrene/DVBresinwassynthesizedaccordingto previouslypub-
cury spectrometer (400 MHz); the chemical shits were expressed
in ppm and are referenced to the corresponding residual nuclei
in the deuterated solvents. FT-IR spectra were obtained using a
Bruker Vertex 80v (Ram II) by dispersion of samples in a KBr pellet.
Elemental analyses (EA) on samples were performed by Columbia
Analytical Services, Tuscon Laboratory (Tuscon, AZ, USA). The con-
version of racemic epichlorohydrin and enantiomeric excess (ee)
of (S)-epichlorohydrin were determined by a Shimadzu 17A GC
with an FID and Chiraldex ␥-TA column (40 m × 25 mm × 0.25 ␮m).
Either para-dichlorobenzene or mesitylene was used as an internal
standard.
lished procedures [27]. A two-neck round-bottom flask equipped
with a stirbar and reflux condenser was purged with nitrogen. Then,
1
reaction mixture was heated under nitrogen for 36 h at 80 C. After
cooling to room temperature, the white solid was then washed sev-
eral times with methanol and dichloromethane. After filtation, the
resin was allowed to dry overnight under high vacuum.
equiv. of styrene and 5 equiv. DVB were added to the flask, and the
◦
2.4. HKR of racemic epichlorohydrin
A pear-shaped flask equipped with a stirbar was charged
with the respective catalyst at the appropriate loadings (0.25 or
0
.5 mol%). The flask was then immersed in a water bath to keep the
2
2
.2. Catalyst preparation
reaction temperature constant. Next, 1 equiv. of racemic epichloro-
hydrin was added to the flask, followed by internal standard
.2.1. Synthesis of (R,R)- and (± )-trans-Co(III)-salen-Cl (1 and 3,
(para-dichlorobenzene or mesitylene). After stirring briefly, 2.5 ␮L
respectively)
of the mixture was passed through a silica plug and washed
with diethyl ether prior to the catalytic reaction. Deionized water
Catalysts 1 and 3 were prepared according to previously
published procedures [16]. In a separatory funnel, (R,R)- or (± )-
trans-Co(III)-salen-OTs was dissolved in dichloromethane (1 g
catalyst to 40 mL dichloromethane). The organic layer was washed
three times with equal volumes of saturated aqueous NaCl and then
dried over anhydrous Na SO . The organic layer was put under
(0.6 equiv.) was then added to start the reaction. For homogeneous
catalytic studies, 2.5 ␮L samples were taken until 99% ee or a con-
stant, maximum % ee was reached. For reactions using polymer
resin-supported catalysts, 2.5 ␮L samples were taken for 4 h. Sam-
ples from all reactions were passed through a plug of silica gel to
remove the catalyst and washed with diethyl ether before GC/FID
analysis.
4
2
4
reduced pressure and concentrated to obtain a dark green solid;
this solid was then resuspended in n-pentane, vacuum filtered, and
dried under high vacuum overnight to obtain catalyst 1 or 3. 1
H
NMR (DMSO-d , 400 MHz) ı 7.81 (s, 2H), ı 7.45 (s, 2H), ı 7.42 (s,
6
2
H), ı 3.63–3.51 (m, 2H), ı 3.09–2.98 (m, 2H), ı 2.03–1.95 (m, 2H),
2.5. Catalyst recovery
ı 1.95–1.82 (m, 2H), ı 1.72 (s, 18H), ı 1.61–1.53 (m, 2H), ı 1.28 (s,
1
8H).
Upon completion of HKR reactions with polymer resin-
supported catalysts, the reaction mixture was poured onto a frit
to isolate the solid catalyst, and this catalyst was washed with
dichloromethane until the liquid filtrate became clear and colorless.
The solid catalyst was dried under high vacuum overnight and then
analyzed via EA and FT-IR. The recovered filtrate was concentrated
under reduced pressure (T < 60 C) to recover the homogeneous cat-
alyst from the reaction. This catalyst was also dried overnight under
high vacuum and analyzed via EA.
2
4
.2.2. Synthesis of (R,R)- and (± )-trans-Co(III)-Salen-SbF6 (2 and
, respectively)
Catalysts 2 and 4 were prepared according to previously pub-
◦
lished procedures [16,17]. First, 1.2 equiv. of AgSbF6 was dissolved
in ∼55 mL dichloromethane in a round-bottom flask equipped with
a stirbar and wrapped in aluminum foil. Then, 1 equiv. of (R,R)- or
(
± )-trans-Co(II)-salen was added to the AgSbF /dichloromethane
6
3
. Results and discussion
solution. The flask was immediately stoppered, and the reaction
was stirred overnight at room temperature in the dark. The dark
green solution was then filtered (in darkness) through two fresh
pads of celite. The filtered solution was then concentrated under
As noted above, previous work has suggested that only the (R,R)-
Co(III)-salen-OH complex generated from (R,R)-Co(III)-salen-X
containing a nucleophilic counter-ion plays a key role in the enan-
tiodiscriminating step in the HKR [16]. To this end, an array of
reactions was first explored evaluating the ligand chirality and
reduced pressure and then rinsed with n-pentane. The dark green
catalyst was dried overnight under high vacuum. 1H NMR (DMSO-
d6, 400 MHz) ı 7.81 (s, 2H), ı 7.45 (d, J = 2.5 Hz, 2H), ı 7.42 (d,
J = 2.7 Hz, 2H) ı 3.60–3.56 (m, 2H), ı 3.08–3.01 (m, 2H), ı 2.01–1.95
(
1
m, 2H), ı 1.93–1.83 (m, 2H), ı 1.72 (s, 18H), ı 1.61–1.52 (m, 2H), ı
.28 (s, 18H).
4
Time it took to determine that 99% ee of HKR with catalyst mix 2 + 8 was reached.

R.E. Key et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 1–7
5
Table 1
a
HKR of racemic epichlorohydrin evaluating ligand chirality and role of counter-ion.
Entry
Catalyst
Total catalyst loading (mol%)
Overall initial TOF (min−1)
Time at ≥99% ee (min)
Total conversion at ≥99% ee (%)
krel b
1
2
3
4
5
1
2
0.25
0.25
0.50
0.50
0.50
1.78
0.89
3.23
2.73
2.79
240
52.1
71.6
56.8
69.3
71.9
>99
8.2
36.0
4.3
4.5
1740c (97%)d
180
1 + 2 (50:50 mix)
1 + 4 (50:50 mix)
2 + 3 (50:50 mix)
180c (77%)
d
187.5c (82%)d
a
HKRs were run with 1 equiv. (rac)-epichlorohydrin (782.1 ␮L, 10 mmol) and 0.6 equiv. of water (108.1 ␮L, 6 mmol). para-Dichlorobenzene (120.3 ␮L) was used as an
internal standard.
b
krel (selectivity factor) is determined by the equation ln[(1 − C)(1 − ee)]/ln[(1 − C)(1 + ee)], where C = conversion and ee = enatiomeric excess [29].
c
Time at which the % ee reaches a maximum.
d
Denotes maximum % ee reached by the reaction.
Table 2
Elemental analysis data from catalysts used for HKRa evaluating counter-ion exchange.
Entry
Catalyst
Carbon (wt.%)
Hydrogen (wt.%)
Nitrogen (wt.%)
Antimony (wt.%)
Cobalt (wt.%)
mol Co/mol N
mol Sb/mol Co
1
2
3
4
5
6
7
8
9
2, Fresh
7, Fresh
8, Fresh
8, Spent
48.63
80.81
78.31
74.24
68.38
33.81
72.43
35.03
90.51
90.09
37.22
6.19
8.96
7.52
7.70
6.90
5.81
6.87
6.17
7.83
7.83
5.69
3.04
1.88
1.89
1.50
1.39
0.19
1.54
0.14
0.94
0.72
0.54
12.92
<0.04
<0.30
<0.20
1.39
0.31
2.05
0.24
6.11
3.23
3.17
2.74
2.58
0.25
2.88
0.20
<0.03
<0.03
1.12
0.48
0.50
0.40
0.43
0.45
0.32
0.44
0.34
<0.01
<0.01
0.49
1.02
<0.01
<0.05
<0.04
0.26
0.60
0.34
0.58
N/A
b
c
2 + 8
c
c
Filtrate of 2 + 8 reaction
c
2 + 7
Filtrate of 2 + 7 reaction
d
Styrene/DVB Resin, Fresh
<0.03
<0.08
2.11
1
1
0
1
2 + Styrene/DVB resinc
N/A
0.91
Filtrate of 2 + styrene/DVB
resin reaction
c
a
b
c
HKRs were run with 1 equiv. (rac)-epichlorohydrin and 0.6 equiv. of water. Mesitylene was used as an internal standard.
Recovered after one HKR cycle.
Solid catalyst, 8, or Co(II)-salen-functionalized resin, 7, recovered from the catalyst mixes denoted above after one HKR cycle.
Blank polymer resin composed of styrene crosslinked 1:5 with DVB.
d
role of the counter-ion in the HKR of racemic epichlorohydrin
Next, reactions were run with a 50:50 mixture of enantiopure
and racemic salen ligand. First, in entry 4, the reaction results from
the use of (R,R)-Co(III)-salen-Cl (1) and (± )-trans-Co(III)-salen-
(
Tables 1 and 2, respectively). Specifically, both the (± )-trans-
salen and (R,R)-salen ligands were used in combination with
−
−
the non-nucleophilic SbF6 counter-ion, and the nucleophilic Cl
SbF (4) are described. If the enantiospecificity of the reaction was
6
counter-ion, which has been suggested to rapidly convert to the
OH counter-ion under reaction conditions [16] (Table 1).
largely determined by the chirality of the complex with the nucleo-
philic counter-ion, in this case complex 1, this reaction should
proceed with identical selectivity as entry 1 in the absence of
counter-ion exchange. However, as shown in the table, this cata-
lyst combination was actually less active and selective than all the
reactions described in entries 1–3, suggesting either the mechanis-
tic hypothesis above is incorrect or other factors were important. In
entry 5, catalysts 2 and 3 [(R,R)-Co(III)-salen-SbF6 and (± )-trans-
Co(III)-salen-Cl, respectively] were used, with opposite pairing of
ligands and counter-ions as the previous case (entry 4). Based on
the mechanistic hypothesis described above, this catalyst should be
the least selective of the combinations in the absence of counter-ion
exchange. However, as shown in Table 1, the catalyst combination,
while also unselective, gave nearly identical results to entry 4. These
last two runs, in combination, are consistent with the occurrence of
rapid scrambling of counter-ions under reaction conditions, as such
scrambling should produce the same average combination of cat-
alysts in both entries 4 and 5. If this scrambling were to occur, the
above results would then remain fully consistent with the overall
mechanistic picture described in Scheme 2 [16–19] and remove the
possibility of lowering the cost of the reactions by using a combina-
tion of enantiopure and racemic ligands. Therefore, the counter-ion
exchange hypothesis was explored in more detail to support or
refute this hypothesis.
−
3.1. HKR of racemic epichlorohydrin
The conversion of racemic epichlorohydrin and the resulting
enantiomeric excess (ee) of (S)-epichlorohydrin were monitored
by GC/FID using a chiral capillary column. The results are collected
in Table 1. In the first entry, the use of (R,R)-Co(III)-salen-Cl (1)
at 0.25 mol% catalyst loading is demonstrated to be an active and
selective catalyst over one use, providing a high turnover frequency
(
TOF) and good enantioselectivity of unreacted epichlorohydrin,
achieving 99% ee at 52.1% conversion (perfect selectivity would give
a conversion of 50%). Catalyst 1 gave an initial turnover frequency
of 1.78 min and displayed a selectivity marked by a very high krel5
−1
of >99. By comparison, at the same catalyst loading, (R,R)-Co(III)-
−
1
salen-SbF6 catalyst (2) gave a much lower initial rate (0.89 min
)
and is also less enantioselective, with a krel of 8.2 (Table 1, entry 2).
These results are consistent with the reaction pathways described
above, with the catalyst in entry 1 rapidly proceeding down the
selective pathway shown in Scheme 2 (red/blue pathway) [16],
and the catalyst in entry 2 being shuttled through the less selec-
tive blue pathway (Scheme 2) [17]. In entry 3, a 50:50 mixture of
the two counter-ions on the enantiopure ligand (catalysts 1 and 2)
was employed at 0.5 mol% catalyst loading. This combination also
gave a selective catalyst, and was the most active of all runs, as
exemplified in the literature [16,17].
3.2. Co(III)-salen-X counter-ion exchange
One way to determine if counter-ion exchange is present
in a mixed counter-ion Co(III)-salen system is to separate and
isolate one Co(III)-salen-X species from the other for direct
5
See Table 1 for definition of krel [29].

6
R.E. Key et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 1–7
Scheme 4. Reaction pathways for formation of resin-supported Co(II) and Co(III)-salens for HKR reactions [26–28].
−
characterization; this can best be explored by immobilizing one
of the Co(III)-salens onto a support. Previously, Co(III)-salen-
triflates have been immobolized onto gold colloids [30], and
our group has immobilized metallosalens on silica [31], soluble
polymers [26,32], polymer resins [27,31], and oxide-supported
polymer brushes [33]. (R,R)-Co(III)-salen-OAc (8) was supported
on a polymer resin with a flexible linker, as these salens have
previously been used to give good catalytic rates and selectivi-
ties, as reported by our group (Scheme 4) [26–28]. This catalyst
was combined with (R,R)-Co(III)-salen-SbF6 as a soluble homo-
geneous catalyst (Fig. 1, catalyst 2) [17]. Catalysts 2 (Fig. 1) [17]
and 8 (Scheme 4) [26–28] were prepared according to previously
published procedures [17,26–28] and characterized by elemental
analysis (EA) (Table 2, entries 1, 3, 4). HKR reactions were run with
the combination of resin-supported (R,R)-Co(III)-salen-OAc (8) and
homogeneous (R,R)-Co(III)-salen-SbF6 (2), along with the appro-
priate control reactions (Table 2, entries 1–4). Following HKR, the
resin-supported salens were recovered via vacuum filtration for
analysis by EA and FT-IR. If counter-ion exchange occurred dur-
ing the reactions, the spent supported catalysts should exhibit an
decrease in SbF6 relative to the fresh catalyst 2 (Table 2, entry 1).
Based on the data presented in Table 2 and Fig. 2, it can be concluded
that there is a degree of counter-ion exchange occurring, where
−
SbF6 is transferred from the homogeneous (R,R)-Co(III)-salen 2 to
the resin-supported (R,R)-Co(III)-salen 8, which originally included
−
OAc.
This exchange is further exemplified when an HKR reaction
is run with homogeneous (R,R)-Co(III)-salen-SbF6 and resin-
supported (R,R)-Co(II)-salen, which is orange in color (catalyst mix
2 and 7, respectively). A brown solid catalyst was recovered from
this reaction, indicative of a Co(III)-salen species supported on the
resin. The Sb/Co ratio for this catalyst was 0.34 (Table 2, entry
7), which is greater than the ratio that would be expected for
catalyst 7, which has no counter-ion coordinated to the Co cen-
−
ter (<0.01, Table 2, entry 2). This suggests that either the SbF6
species are transferred onto the supported Co(II) species, oxidizing
it to Co(III)-SbF , or homogeneous Co(III)-salen-SbF6 species are
6
−
1
increased Sb/Co ratio, and an Sb-F stretch near 660 cm should be
apparent in the FT-IR spectrum of the recovered solid, which was
initially Sb-free [34].
For the HKR reaction with a catalyst mix of 2 (Co(III)-salen-SbF6)
and 8 (supported Co(III)-salen-OAc), the supported spent catalyst,
8, when analzyed by EA, had an Sb/Co ratio of 0.26 (Table 2, entry 5).
Though the Sb/Co ratio observed for this catalyst mix is less than the
Sb/Co ratio of fresh catalyst 2 that is fully oxidized to Co(III)-SbF6
(
1.02, Table 2, entry 1), it is greater than the Sb/Co ratio of fresh and
−
spent catalyst 8, which has no SbF6 counter-ion (<0.05 and <0.04,
Table 2, entries 3 and 4, respectively). This suggests that some SbF6
−
counter-ion from homogeneous catalyst 2 has exchanged with the
counter-ion from supported catalyst 8. To verify the presence of
−
SbF6 in spent catalyst 8, the solid was examined by FT-IR. A peak
−
1
at 661 cm associated with an Sb-F stretch (Fig. 2, dot-dashed line)
clearly suggests the presence of SbF6 in the spent solid 8, whereas
−
fresh and spent catalyst 8 from regular reactions without catalyst
Fig. 2. FT-IR spectra of fresh (solid line) and spent resin-supported (R,R)-Co(III)-
salen-OAc (8) (after one HKR cycle, dotted line) and recovered solid catalyst 8
2
do not have this peak (Fig. 2, solid and dotted line, respectively).
(
dot-dashed line) from HKR with homogeneous (R,R)-Co(III)-salen-SbF6 + resin-
Analysis of the filtrate of the HKR reaction using catalyst mix 2
supported (R,R)-Co(III)-salen-OAc (2 + 8). The appearance of the Sb-F stretch at
and 8 (which includes homogeneous (R,R)-Co(III)-salen-SbF ) gave
−1
−
6
661 cm is indicative of SbF6 likely coordinating to the Co center of the resin-
an Sb/Co ratio of 0.60 (Table 2, entry 6), indicating thtat there is a
supported (R,R)-Co-salen.

R.E. Key et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 1–7
7
−
a nucleophile-delivering complex (e.g. OH ) and 50% racemic
ligand with a counter-ion designed to make the salen complex
−
a Lewis acidic, electrophilic complex (e.g. SbF6 ). In this work, it
was shown that such experiments do not lead to highly selective
catalysts due to counter-ion scrambling between the Co centers.
The counter-ion scrambling hypothesis was supported by control
experiments using polymer-supported Co-salen species and the
associated kinetic evidence, FT-IR spectra, and elemental analysis
results. This study also showed that oxidation of a Co(II) species on a
polymer support can be promoted via contact with a homogeneous
Co(III) species, which suggests a method for catalyst regeneration
after deactivation by counter-ion loss [28].
Acknowledgements
We thank the U.S. Department of Energy, Basic Energy Sciences
for financial support via Catalysis Science Grant No. DE-FG01-
Fig. 3. FT-IR spectra of fresh (solid line) resin-supported (R,R)-Co(II)-salen (7),
recovered solid catalyst (dotted line) from the HKR reaction using homogeneous
0
3R15459. We also thank Dr. Marcus Weck and Mr. Michael Kahn
(both from New York University) for helpful discussions, as well as
(R,R)-Co(III)-salen-SbF6 + resin-supported (R,R)-Co(II)-salen (2 + 7), and recovered
solid catalyst from the HKR reaction using homogeneous (R,R)-Co(III)-salen-SbF6
Dr. Yan Feng (Georgia Tech) for experimental assistance.
(
6
2) + styrene/DVB resin (dot-dashed line). The appearance of an Sb-F stretch at
61 cm is indicative of SbF6 likely coordinating to the Co center of the resin-
−
1
−
References
supported (R,R)-Co-salen.
[
1] D.K. Garner, L. Liang, D.A. Barrios, J.-L. Zhang, Y. Lu, ACS Catal. 1 (2011)
1083–1089.
trapped within the resin. In this case, the spent resin material, 7,
−1
[2] G.J. Kim, J.H. Shin, Catal. Lett. 63 (1999) 83–90.
also shows an Sb-F stretch at 661 cm in its FT-IR spectrum (Fig. 3,
dotted line), whereas the spectrum of the fresh catalyst 7 does not
show this peak (Fig. 3, solid line). The light brown filtrate from the
[
3] A.W. Kleij, Chem. Eur. J. 14 (2008) 10520–10529.
[
4] K. Srinivasan, P. Michaud, J.K. Kochi, J. Am. Chem. Soc. 108 (1986) 2309–2320.
[5] R. Irie, K. Noda, Y. Ito, N. Matsumoto, T. Katsuki, Tetrahedron Lett. 31 (1990)
345–7348.
6] J. Huang, X. Fu, Q. Miao, Appl. Catal. A: Gen. 407 (2011) 163–172.
[7] M. Tokunaga, J.F. Larrow, F. Kakiuchi, E.N. Jacobsen, Science 277 (1997)
936–938.
[8] X. Zhang, Y.-B. Jia, X.-B. Lu, B. Li, H. Wang, L.-C. Sun, Tetrahedron Lett. 49 (2008)
589–6592.
7
HKR reaction using catalysts 2 and 7 gives an Sb/Co ratio of 0.58
[
−
(
Table 2, entry 8), indicative of a partial SbF6 transfer from homo-
geneous (R,R)-Co(III)-salen to resin-supported (R,R)-Co(II)-salen.
−
To determine whether significant amounts of SbF6 may
6
become trapped within the polymer resin without reacting with
a Co-salen center, HKR reactions were run using catalyst 2 and a
polymer resin constructed of 1:5 crosslinked styrene/DVB, which
does not include any salen species. In evaluating the EA data, the
Sb content in both fresh and spent styrene/DVB resins were found
to be negligible (Table 2, entries 9 and 10). Also, there was no Sb-F
[
9] K. Venkatasubbaiah, X. Zhu, E. Kays, K.I. Hardcastle, C.W. Jones, ACS Catal. 1
(2011) 489–492.
[
[
10] R.I. Kureshy, A. Das, N.-u.H. Khan, S.H.R. Abdi, H.C. Bajaj, ACS Catal. 1 (2011)
529–1535.
1
11] N. Madhavan, T. Takatani, C.D. Sherrill, M. Weck, Chem. Eur. J. 15 (2009)
1186–1194.
[12] S.E. Schaus, B.D. Brandes, J.F. Larrow, M. Tokunaga, K.B. Hansen, A.E. Gould, M.E.
Furrow, E.N. Jacobsen, J. Am. Chem. Soc. 124 (2002) 1307–1315.
13] P. Kumar, V. Naidu, P. Gupta, Tetrahedron 63 (2007) 2745–2785.
14] A. Sadhukhan, N.-u.H. Khan, T. Roy, R.I. Kureshy, S.H.R. Abdi, H.C. Bajaj, Chem.
Eur. J. 18 (2012) 5256–5260.
15] N. Madhavan, C.W. Jones, M. Weck, Acc. Chem. Res. 41 (2008) 1153–1165.
16] L.P.C. Nielsen, C.P. Stevenson, D.G. Blackmond, E.N. Jacobsen, J. Am. Chem. Soc.
−1
stretch found at 661 cm in the FT-IR spectra (Fig. 3, dot-dashed
line). In addition, the dark green filtrate of the HKR reaction with the
salen-free styrene resin and 2 had a Co/N ratio of 0.49 and an Sb/Co
ratio of 0.91 (Table 2, entry 11), indicating that nearly all of complex
[
[
[
[
2
was recovered (Table 2, entry 1). Since an increase in Sb/Co ratio,
1
26 (2004) 1360–1362.
[17] S. Jain, K. Venkatasubbaiah, C.W. Jones, R.J. Davis, J. Mol. Catal. A: Chem. 316
2010) 8–15.
as well as the presence of the Sb-F stretch is observed in the spent
catalysts upon exposure to catalyst 2 and minimal to no amount of
catalyst 2 becomes trapped within the salen-free polymer resin, it
can be concluded that the Sb species found in spent catalyst from
entires 5 and 7 of Table 2 are indicative of the prevalence of counter-
(
[
[
18] P.G. Cozzi, Chem. Soc. Rev. 33 (2004) 410–421.
19] S. Jain, X. Zheng, C.W. Jones, M. Weck, R.J. Davis, Inorg. Chem. 46 (2007)
8887–8896.
20] J.M. Ready, E.N. Jacobsen, Angew. Chem. Int. Ed. 41 (2002) 1374–1377.
21] S. Kemper, P. Hroba ˇı rik, M. Kaupp, N.E. Schlörer, J. Am. Chem. Soc. 131 (2009)
[
[
−
−
ion exchange occurring between SbF6 and OAc. Thus, the similar
reactivity seen in Table 1, entires 4 and 5 demonstrates that rapid
counter-ion exchange occurs between the Co(III)-salen species dur-
ing catalysis and, therefore, mixed chiral-racemic ligand systems
cannot be used to reduce the cost of HKR reactions. It should fur-
ther be noted that the counter-ion scrambing is consistent with the
mechanism of Nielsen et al. [24].
4172–4173.
[22] E.N. Jacobsen, Acc. Chem. Res. 33 (2000) 421–431.
[
[
23] E. Vinck, D.M. Murphy, I.A. Fallis, R.R. Strevens, S. Van Doorslaer, Inorg. Chem.
9 (2010) 2083–2092.
4
24] L.P.C. Nielsen, S.J. Zuend, D.D. Ford, E.N. Jacobsen, J. Org. Chem. 77 (2012)
2486–2495.
[25] G.J. Kim, H. Lee, S.J. Kim, Tetrahedron Lett. 44 (2003) 5005–5008.
26] X. Zheng, C.W. Jones, M. Weck, Adv. Synth. Catal. 350 (2008) 255–261.
27] K. Venkatasubbaiah, C.S. Gill, T. Takatani, C.D. Sherrill, C.W. Jones, Chem. Eur. J.
[
[
15 (2009) 3951–3955.
4
. Conclusions
[28] K. Venkatasubbaiah, X.-J. Zhu, C. Jones, Top. Catal. 53 (2010) 1063–1065.
[
29] H.B. Kagan, J.C. Fiaud, Topics in Stereochemistry, John Wiley & Sons, Inc., 2007,
pp. 249–330.
The potential of combining chiral and racemic ligands in
[
30] T. Belser, E.N. Jacobsen, Adv. Synth. Catal. 350 (2008) 967–971.
the enantioselective HKR of epoxides based on the mechansitic
hypotheses of Nielsen et al. was explored [16]. Based on the cur-
rently accepted mechanism, high activity and enantioselectivity
might be achieved in reactions catalyzed with by 50% enantiop-
ure ligand with a counter-ion designed to make the salen complex
[31] C.S. Gill, K. Venkatasubbaiah, C.W. Jones, Adv. Synth. Catal. 351 (2009)
344–1354.
1
[
[
32] X.J. Zheng, C.W.M. Weck, Chem. Eur. J. 12 (2006) 576–583.
33] C.S. Gill, K. Venkatasubbaiah, N.T.S. Phan, M. Weck, C.W. Jones, Chem. Eur. J. 14
(2008) 7306–7313.
[
34] O.Q. Munro, S.C. Shabalala, N.J. Brown, Inorg. Chem. 40 (2001) 3303–3317.