Biomimetic iron-catalyzed asymmetric epoxidation of aromatic alkenes by using hydrogen peroxide

Article abstract of DOI:10.1002/chem.200800595

A novel and general biomimetic non-heme Fe-catalyzed asymmetric epoxidation of aromatic alkenes by using hydrogen peroxide is reported herein. The catalyst consists of ferric chloride hexahydrate (FeCl₃·OH ₂O), pyridine-2,6-dicarboxylic acid (H₂-(pydic)), and readily accessible chiral N-arenesulfonyl-N′-benzyl-substituted ethylenediamine ligands. The asymmetric epoxidation of styrenes with this system gave high conversions but poor enantiomeric excesses (ee), whereas larger alkenes gave high conversions and ee values. For the epoxidation of trans-stilbene (1a), the ligands (S,S)-N-(4-toluenesulfonyl)-1,2- diphenylethylenediamine ((S,S)-4a) and its N′-benzylated derivative ((S,S)-5a) gave opposite enantiomers of trans-stilbene oxide, that is, (S,S)-2a and (R,R)-2a, respectively. The enantioselectivity of alkene epoxidation is controlled by steric and electronic factors, although steric effects are more dominant. Preliminary mechanistic studies suggest the in situ formation of several chiral Fe-complexes, such as [FeCl(L*)₂-(pydic)] ·HCl (L* = (S,S)-4a or (S,S)-5a in the catalyst mixture), which were identified by ESIMS. A UV/Vis study of the catalyst mixture, which consisted of FeCl₃·6H₂O, H₂(pydic), and (S,S)-4a, suggested the formation of a new species with an absorbance peak at λ = 465 nm upon treatment with hydrogen peroxide. With the aid of two independent spin traps, we could confirm by EPR spectroscopy that the reaction proceeds via radical intermediates. Kinetic studies with deuterated styrenes showed inverse secondary kinetic isotope effects, with values of k _H/k_D = 0.93 for the β carbon and k_H/k _D=0.97 for the a carbon, which suggested an unsymmetrical transition state with stepwise O transfer. Competitive epoxidation of para-substituted styrenes revealed a linear dual-parameter Hammett plot with a slope of 1.00. Under standard conditions, epoxidation of la in the presence of ten equivalents of H₂¹⁸O resulted in an absence of the isotopic label in (S,S)-2a. A positive non-linear effect was observed during the epoxidation of la in the presence of (S,S)-5a and (R,R)-5a.

Full text of DOI:10.1002/chem.200800595

FULL PAPER
DOI: 10.1002/chem.200800595
Biomimetic Iron-Catalyzed Asymmetric Epoxidation of Aromatic Alkenes
by Using Hydrogen Peroxide
Feyissa Gadissa Gelalcha,*[a] Gopinathan Anilkumar,[b] Man Kin Tse,[c]
Angelika Brückner,[c] and Matthias Beller*[c]
Abstract: A novel and general biomim-
etic non-heme Fe-catalyzed asymmetric
epoxidation of aromatic alkenes by
using hydrogen peroxide is reported
herein. The catalyst consists of ferric
chloride hexahydrate (FeCl3·6H2O),
pyridine-2,6-dicarboxylic acid (H2-
alkene epoxidation is controlled by
steric and electronic factors,although
steric effects are more dominant. Pre-
liminary mechanistic studies suggest
the in situ formation of several chiral
proceeds via radical intermediates. Ki-
netic studies with deuterated styrenes
showed inverse secondary kinetic iso-
tope effects,with values of kH/kD =0.93
for the b carbon and kH/kD =0.97 for
the a carbon,which suggested an un-
symmetrical transition state with step-
wise O transfer. Competitive epoxida-
tion of para-substituted styrenes re-
vealed a linear dual-parameter Ham-
mett plot with a slope of 1.00. Under
standard conditions,epoxidation of 1a
in the presence of ten equivalents of
H218O resulted in an absence of the iso-
topic label in (S,S)-2a. A positive non-
linear effect was observed during the
epoxidation of 1a in the presence of
(S,S)-5a and (R,R)-5a.
Fe-complexes,such as [FeCl(L*)
(pydic)]·HCl (L*=(S,S)-4a or (S,S)-5a
in the catalyst mixture),which were
-
2
AHCTREUNG
ACHTREUNG(pydic)),and readily accessible chiral
N-arenesulfonyl-N’-benzyl-substituted
ethylenediamine ligands. The asymmet-
ric epoxidation of styrenes with this
system gave high conversions but poor
enantiomeric excesses (ee),whereas
larger alkenes gave high conversions
and ee values. For the epoxidation of
trans-stilbene (1a),the ligands ( S,S)-N-
(4-toluenesulfonyl)-1,2-diphenylethyl-
identified by ESIMS. A UV/Vis study
of the catalyst mixture,which consisted
of FeCl3·6H2O,H
2ACHTRE(UNG pydic),and ( S,S)-
4a,suggested the formation of a new
species with an absorbance peak at l=
465 nm upon treatment with hydrogen
peroxide. With the aid of two inde-
pendent spin traps,we could confirm
by EPR spectroscopy that the reaction
ACHTREUNGenediamine ((S,S)-4a) and its N’-benzy-
lated derivative ((S,S)-5a) gave oppo-
site enantiomers of trans-stilbene
oxide,that is,( S,S)-2a and (R,R)-2a,
respectively. The enantioselectivity of
Keywords: asymmetric catalysis
epoxidation · hydrogen peroxide ·
iron · radical reactions
·
Introduction
Optically active oxiranes represent key building blocks for
the synthesis of fine chemicals and pharmaceuticals.[1] Cata-
lytic asymmetric epoxidation represents an interesting tool
for their synthesis as an alternative to the more common ki-
netic resolution processes.[2] In this regard,the state-of-the-
art protocols are the Sharpless epoxidation of allylic alco-
hols with titanium tartarate complexes;[3] the Katsuki–Jacob-
sen epoxidation of unfunctionalized alkenes by using chiral
[a] Dr. F. G. Gelalcha
4900 East University Boulevard,609
Odessa,Texas 79762 (USA)
E-mail: fgadissa@yahoo.de
[b] Dr. G. Anilkumar
Anthem Biosciences Pvt Ltd
49 Canara Bank Road
Bangalore 560 099 (India)
[c] Dr. M. K. Tse,Dr. A. Brückner,Prof. Dr. M. Beller
Leibniz-Institut für Katalyse e.V. an der Universität Rostock
Albert Einstein Strasse 29a
{N,N’-bis(salicylidene)ethylenediamine}] catalysts;[4] or-
[MnACHTREUNG
ganocatalytic methods that utilize chiral ketone catalysts,as
pioneered by Shi et al.;[5] and the Julia epoxidation of
enones by using poly-l-leucine with bases as the catalysts.[6]
The development of less expensive and environmentally
more benign catalysts and oxidant systems that have a wider
18059 Rostock (Germany)
Fax : (+49)381-12815000
E-mail: matthias.beller@catalysis.de
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the authors.
Chem. Eur. J. 2008, 14,7687 – 7698
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
7687

scope of applicability remains a challenge in organic synthe-
sis. Hydrogen peroxide is one of the most practical terminal
oxidants in terms of cost and atom efficiency. In recent
years,this oxidant has become increasingly popular in orga-
nocatalytic oxidations[7] and in metal-catalyzed asymmetric
Scheme 1. General synthesis of optically pure substituted diphenylethyle-
[9]
epoxidations.[8] Thus,Ti-,
Pt-,[10] and Ru-based catalysts[11]
nediamine ligands. a) R1SO2Cl,DIPEA,CH 2Cl2, 08C–RT; b) ArCHO,
1
have been reported for asymmetric epoxidations of olefins
EtOH,reflux; c) NaBH 4,EtOH,RT. See Table 1 for details of R
and
R2.
by using hydrogen peroxide. Similarly,iron is a cheap,versa-
[12]
tile,and less toxic catalyst.
Indeed,a highly efficient,
though impractical,biomimetic asymmetric epoxidation of
styrene derivatives has been developed recently,which uses
iodosobenzene in dichloromethane as the oxidant and chiral
iron porphyrin complexes.[13] Similarly,the aerobic asymmet-
ric epoxidation of styrene derivatives,which is catalyzed by
Table 1. Iron-catalyzed asymmetric epoxidation of 1a by using (S,S)-4a–
c and (S,S)-5a–d.[a]
Entry Compound
Ligand
Conv[b] Yield[b] ee[c] [%],
a tris(d,d-dicampholylmethanato) iron
ACHTRE(UNG III) complex ([Fe-
R1
R2
H
[%]
[%]
abs conf[d]
ACHTREUNG
vent,[14] the asymmetric epoxidation of trans-methylstyrene
by employing Fe complexes with peptidelike ligands as the
catalysts,[15] and the iron-catalyzed asymmetric peracid epox-
idation of electron-poor alkenes[16] have been described. All
of these methods are less attractive from a practical point of
view.
1
2
3
4
5
6
7
A
4-MeC6H4
4-PhC6H4
4-tBuC6H4
4-MeC6H4
4-PhC6H4
4-tBuC6H4
100
86
28,
(ꢀ)-(2S,3S)
G
H
H
100
100
100
100
100
84
84
87
84
95
92
29,
(ꢀ)-(2S,3S)
29,
AHCTREUNG
(ꢀ)-(2S,3S)
U
Bn
Bn
Bn
42,
(+)-(2R,3R)
39,
(+)-(2R,3R)
40,
(+)-(2R,3R)
39
Recently,we reported the first breakthrough and a prom-
ising non-heme iron-catalyzed asymmetric epoxidation of ar-
omatic alkenes by using hydrogen peroxide,which not only
gives good-to-excellent isolated yields of epoxides but also
enantiomeric excess (ee) values of up to 97%.[17] Herein,we
report on the further extension of this novel catalytic asym-
metric epoxidation system to more commonly encountered
alkenes,preliminary mechanistic insights,and give a full ac-
count of experimental details of the reaction.
AHCTREUNG
AHCTREUNG
T
4-MeC6H4 4-MeBn 100
(+)-(2R,3R)
[a] Conditions: trans-stilbene (0.5 mmol),H 2O2 (1.0 mmol),FeCl 3·6H2O
(5.0 mol%),H 2A(pydic) (5.0 mol%),( S,S)-4 or (S,S)-5 (12 mol%), tert-
CTHREUNG
amyl alcohol,RT,1 h. [b] Conversion and yield determined by GC with
dodecane as the internal standard; error ꢁ1%. [c] The ee of 2a was de-
termined by HPLC on a chiral column; error ꢁ1%. [d] Absolute config-
uration determined by comparing the sign of the optical rotation of the
major enantiomer with known data. Optical rotations were measured by
using a chiral detector connected to a chiral HPLC.[18]
Results and Discussion
Ligand design and substrate scope: As stated previously,we
have investigated the epoxidation of trans-stilbene (1a) to
trans-stilbene oxide (2a) by using 30% hydrogen peroxide
as the oxidant. Our catalyst systems consisted of commer-
cially available iron salts (such as ferric chloride hexahy-
(S,S)-3 is used as the chiral source. This led us to systemati-
cally modify the ligand to produce a sterically diverse set of
structures,such as ( S,S)-4b,c and (S,S)-5a–d (Table 1). N-
Benzylation of the remaining amino group generally result-
ed in further enhancement of the product ee values by about
10% (Table 1,entries 4–7). It is obvious that among these
compounds the simple N-benzyl-substituted derivative (S,S)-
5a ((S,S)-BnTsDPEN) led to the most significant increase in
ee values and gave (+)-(2R,3R)-2a in 42% ee (Table 1,
entry 4). It is also striking that the N’-sulfonylated amines
((S,S)-4a–c) and the corresponding N-benzyl-N’-sulfonylated
ligands ((S,S)-5a–d) gave excesses of enantiomers with op-
posite absolute configurations. These results suggested the
existence of several catalyst species,the importance of spe-
cific spatial orientation of the chiral ligand imposed by a
rigid aromatic ring system,and the possible role of p–p in-
teraction and hydrogen bonding in controlling enantioselec-
tivities.
drate),pyridine-26,-dicarboxylic acid (H
2ACHTRE(UNG pydic)),and care-
fully chosen chiral N-sulfonylated ethylene diamine deriva-
tives.[17] The required ligands are easily accessible by the
monosulfonylation of optically pure C2-symmetrical 1,2-di-
amines,such as ( ꢀ)-(S,S)-1,2-diphenylethylenediamine (3),
with arenesulfonylchlorides in the presence of bases,such as
diisopropyl ethylamine (DIPEA),and,where necessary,re-
ductive N-alkylation of the resulting products (4) in the
presence of suitable aldehydes and sodium borohydride
(NaBH4; Scheme 1). The influence of various ligands 4 and
5 on the enantioselectivity of the reaction is summarized in
Table 1.
The observation that the commercially available ligand
(S,S)-N-(4-toluenesulfonyl)-1,2-diphenylethylenediamine
((S,S)-TsDPEN; (S,S)-4a) gave (ꢀ)-(2S,3S)-2a in 28% ee at
room temperature (Table 1,entry 1) is in stark contrast with
the racemic product observed when the unsubstituted ligand
A surprising effect of the concentration of (S,S)-5a (i.e.,
[(S,S)-5a]) on substrate conversion and on the ee value of
2a was observed,as detailed in Table 2. As expected,at a
7688
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2008, 14,7687 – 7698

Asymmetric Epoxidation of Aromatic Alkenes
FULL PAPER
Table 2. Effect of the concentration of (S,S)-5a on the ee of (R,R)-2a.
Table 3. Fe-catalyzed asymmetric epoxidation of aromatic alkenes.
Entry
Alkene
Conv[b]
[%]
Yield[b]
[%]
ee[c] [%],
Abs conf[d]
Entry
A
t [h]
Conv[c] [%]
Yield[c] [%]
ee[d] [%]
ACHTREUNG
[e]
1
2
3
100
100
100
84
61
82
8,–
[e]
1
2
3
4
5
6
7
8
2.5
2.5
5.0
5.0
1
16
1
16
1
1
1
1
22
71
64
12
52
48
76
83
87
88
91
–
38
[e]
–
[e]
95
41
39
42
36
31
8,–
10.0
12.0
15.0
20.0
100
100
100
100
21,
(ꢀ)-(R,R)
[a] Conditions: 1a (0.5 mmol),H 2O2 (30%,1.0 mmol,2 equiv),
FeCl3·6H2O (5.0 mol%),H 2A(pydic) (5.0 mol%),( S,S)-5a, tert-amyl alco-
CTHREUNG
[e]
hol (10 mL),RT. [b] Based on the substrate. [c] Conversion and yield de-
termined by GC with dodecane as the internal standard. [d] The ee value
of (R,R)-2a was determined by HPLC on a chiral column; the (+)-
(2R,3R) enantiomer was the major product. [e] Not determined.
4
5
6
7
100
100
72
73
52
60
62
11,–
[e]
26,–
low [(S,S)-5a] of 2.5 mol%,low substrate conversion was
observed after 1 h,whereas extending the reaction time to
16 h not only gave a higher conversion (71%) and yield
(52%),but also resulted in a significant ee of 38% (Table 2,
entries 1 and 2). By using only 5 mol% of (S,S)-5a,we
could achieve an ee of 41% after 16 h although the conver-
sion was not complete (95%). When the (S,S)-5a concentra-
tion was increased to 10mol% the conversion was complete
after only 1 h,although this did not change the ee signifi-
cantly; the maximum ee value was observed with 12 mol%
(S,S)-5a. Higher concentrations of (S,S)-5a (15 or 20 mol%;
Table 2,entries 7 and 8) deteriorated the ee values without
significantly altering the yields. This is of practical significant
because avoiding high concentrations of catalyst is a positive
contribution for the efficient use of resources and the cur-
rent desire for the development of environmentally greener
methodologies.
We have previously demostrated that,in the case of trans-
stilbenes 1b–f and 1h–i,for a given substituent the catalyst
activity increased on moving from the ortho to the para po-
sition of the substrate.[17] Thus,high enantioselectivities were
obtained with sterically bulky 4,4’-dialkyl-substituted trans-
stilbenes. The highest ee value (97%) was achieved with
(E)-2-(4-tert-butylstyryl)naphthalene (1j),whereas 1k gave
a lower ee value. Therefore,we have extended the use of
(S,S)-5a to the asymmetric epoxidation of more commonly
encountered alkenes (Table 3).
[e]
20,–
[e]
91
14,–
48,
8
9
100
100
57
49
42
33
(+)-(2R,3R)
29,
(+)-(2R,3R)
23,
10
(+)-(2R,3R)
53,
11
12
100
100
91[f]
90[f]
(+)-(2R,3R)
71,
(+)-(2R,3R)
[a] Conditions: alkene (0.5 mmol),H 2O2 (30%,1.0 mmol,2 equiv),
FeCl3·6H2O (5.0 mol%),H 2A(pydic) (5.0 mol%),( S,S)-5a (12 mol%),
tert-amyl alcohol,RT,1 h. TBDMS =tert-butyldimethylsilyl. [b] Con-
CTHREUNG
version and yield determined by gas chromatography with dodecane as
the internal standard. [c] Determined by HPLC on a chiral column.
[d] Absolute configuration determined by comparing the sign of the opti-
cal rotation of the major enantiomer with known data. Optical rotations
were measured by using a chiral detector connected to a chiral HPLC.
See also ref. [19] for Entries 3 and 7,ref. [20] for Entry 8,ref. [21] for
Entry 9,ref. [22] for Entry 10,and ref. [18] for Entry 12. [e] Not deter-
mined. [f] Isolated yield.
It is obvious from Table 3 that the catalyst still displays
considerable reactivity toward a wide range of alkenes.
However,the enantioselectivities of small alkenes,such as
the styrenes (Table 3,entries 1–7),are lower than those ach-
ieved with larger alkenes. Moreover,alkenes with similar
steric demands (e.g.,4-chlorostyrene and 4-fluorostyrene
versus 4-methylstyrene and 4-trifluoromethyl styrene) are
oxidized with different enantioselectivities; this suggests that
electronic effects are also important. Oxidation of cinnamyl
alcohol gave the corresponding epoxide in 46% ee and 49%
yield. Hydroxyl protection slightly reduced the yield and the
ee value decreased to 29%. This suggested that hydrogen-
bonding groups placed in appropriate positions could in-
crease the ee values (Table 3,entries 8–10). trans-4,4’-Diiso-
propyl stilbene (1g) gave the epoxide in 71% ee (Table 3,
entry 12),which is in agreement with the common trend
Chem. Eur. J. 2008, 14,7687 – 7698
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
7689

F. G. Gelalcha et al.
previously observed for trans-
4,4’-dialkyl-substituted stilbenes
1d and 1i.[17]
Preliminary mechanistic investi-
gations
ESIMS studies: The ESI mass
spectrum of the catalyst mix-
ture (FeCl3·6H2O,0.025 mmol;
H CHTRE(UNG pydic),0.025 mmol; ( S,S)-
2A
4a,0.06 mmol) in isopropanol/
[23]
water (50:50,v/v)
is dominat-
ed by the signal of the chiral
ligand (S,S)-4a ([M+H+]; m/z
367.1484) and a signal that cor-
Figure 1. UV/Vis spectra of the catalyst mixture (FeCl3·6H2O (4.17”
10ꢀ3 m),H (pydic) (4.17”10ꢀ3 m),( S,S)-4a (0.01m)) in tert-amyl alcohol
2ACHTREUNG
responds to its dimeric form
(TsDPEN)2 ([M+H+]; m/z
a) before the addition of H2O2 and b)–f) after the portionwise addition of
a dilute solution of aqueous H2O2 (33.32%,0.5 mmol; 0.1 mmol per por-
tion) in tert-amyl alcohol,added at 0,4,7,10,and 13 min,respectively.
733.28791). It also displays a
complex set of very weak sig-
nals,which were assigned to various Fe-containing species:
[Fe
(pydic)2TsDPEN]+ ([M+]; m/z 953.22649),[FeCl
(TsDPEN)2]·HCl+H+ ([M+H+]; m/z 1025.18204),and
[Fe{H2A(pydic)}(pydic)
(TsDPEN)2]+ ([M+]; m/z 1120.249).
An attempted crystallization of a pale yellow product that
was isolated from the catalyst mixture,which consisted of
A
([M+]; m/z 587.08394),[Fe-
G
(pydic)-
The absorbance at l=465 nm is probably due to a hydroper-
G
oxo-to-FeIII charge-transfer transition and lies in the range
[24]
ꢀ
G
reported for related low-spin LFeIII OOH complexes. On
the other hand,when the catalyst mixture was treated with
small portions of hydrogen peroxide in the presence of
trans-stilbene,the band at l=465 nm was no longer visible.
Instead,a new band appeared at l=422 nm that is likely to
be due to the transformation product of the species with an
absorbance at l=465 nm after oxygen transfer (Figure 2).
Raman and FTIR spectroscopic studies that were conducted
alongside the UV/Vis spectroscopic measurements gave
bands that corresponded to alkene and epoxide functionali-
FeCl3·6H2O,H 2ACHTREUNG(pydic),and ( S,S)-5a in a ratio of 1:1:2.1
stirred in tert-amyl alcohol at room temperature for 2 h,
failed. However,ESIMS analysis of this product revealed
the presence of BnTsDPEN ([M+H+]; m/z 457.3502) as the
dominant species,together with a less intense signal that
corresponds to the [FeACTHERNGU(BnTsPDEN)2ClACHTRE(UGN pydic)]·HCl ([M+
H+]; m/z 1205.2847) species,which is in full accordance
with the calculated spectrum. Very small signals at m/z
1300.3287 and 1556.248,which correspond to as-yet unchar-
acterized Fe-containing species,were also detected. Both re-
sults suggest the formation of complex equilibria,which sup-
ports the hypothesis that a chiral Fe–ligand complex is re-
sponsible for the asymmetric induction. However,which of
these precatalysts converts to the active catalyst upon treat-
ment with hydrogen peroxide is currently under investiga-
tion.
UV/Vis spectroscopy: When a dilute solution of aqueous hy-
drogen peroxide (33.32%,0.5 mmol) in tert-amyl alcohol
was added in five portions (5”0.1 mmol) to a catalyst mix-
ture (FeCl3·6H2O (4.17”10ꢀ3 m),H (pydic) (4.17”10ꢀ3 m),
2ACHTREUNG
(S,S)-4a (0.01m)) in the same solvent in the absence of sub-
strate at room temperature,a new band in the UV trace de-
veloped progressively at 465 nm (Figure 1,b–f),but did not
increase in intensity after addition of the last portion. The
identity of this species is currently under investigation. How-
Figure 2. UV/Vis spectra of the reaction mixture (FeCl3·6H2O
(0.025 mmol,4.17”10 ꢀ3 m),H (pydic) (0.025 mmol,4.17”10 ꢀ3 m),( S,S)-4a
2ACHTREUNG
III
III
ꢀ
ꢀ
ever,it is likely to be either an LFe
OOH or an LFe
(0.06 mmol,0.01 m), 1a (0.5 mmol,0.08 m)) in tert-amyl alcohol a) before
the addition of H2O2 and b)–f) after the portionwise addition of a dilute
solution of aqueous H2O2 (33.32%,0.5 mmol; 0.1 mmol per portion) in
tert-amyl alcohol.
III
ꢀ
ꢀ
OO Fe L complex. Homolysis or heterolysis of the O O
bond of either of these species would lead to high-valent
Fe=O species that are suggested to be the actual oxidants.
7690
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2008, 14,7687 – 7698

Asymmetric Epoxidation of Aromatic Alkenes
FULL PAPER
ties. We could not observe peaks that were diagnostic of
ꢀ
ꢀ
Fe O or O O bonds in the Raman and FTIR spectra,pre-
sumably because of the overlap of these bands with those of
the alkene.
EPR spectroscopy: To determine whether paramagnetic Fe
species are formed before or during the reaction and how
this might change upon the sequential addition of each com-
ponent,we followed the EPR spectra of each component as
it was added to the reaction solution in tert-amyl alcohol,
starting with the iron salt. FeCl3·6H2O (Figure 3, c) has a
sharp signal at g=1.98,which is comparable to the report-
ed[34] value of g=2.022 for FeCl3·6H2O at room temperature
Figure 4. Graphs showing the change in the EPR signals of the reaction
mixture (FeCl3·6H2O (0.025 mmol,0.0125 m),H
2ACHTRE(UNG pydic) (0.025 mmol,
0.0125m),( S,S)-4a (0.06 mmol,0.03 m), 1a (0.5 mmol,0.25 m)) in tert-amyl
alcohol a) before the addition of H2O2 and b)–g) after the portionwise
addition of aqueous H2O2 (33.32%,91.13 mL,1 mmol) in tert-amyl alco-
hol (total volume 1.0 mL); 5”100 mL,0.1 mmol; 1” =500 mL,0.5 mmol.
All spectra were recorded at 77 K.
Spin trapping with TEMPO: A trace amount of 2,2,6,6-tetra-
methylpiperidine-1-oxyl (TEMPO; 1.282”10ꢀ4 mmol) in
tert-amyl alcohol was added to the standard reaction mix-
ture (FeCl3·6H2O (0.025 mmol),H
2ACHTRE(UNG pydic) (0.025 mmol),
Figure 3. Graphs showing the change in the EPR signals of the reaction
(S,S)-4a (0.06 mmol), trans-stilbene (0.5 mmol)) in tert-amyl
alcohol (2 mL) and the EPR spectrum was recorded at
77 K. The resulting spectrum was dominated by the signal of
TEMPO (Figure 5A, c). After an aqueous solution of hy-
mixture upon the successive addition of H2ACHTRE(UGN pydic) (0.025 mmol,0.0125 m;
g),( S,S)-4a (0.06 mmol,0.03 m; b),and 1a (0.5 mmol,0.25 m; d)
to a base solution of FeCl3·6H2O (0.025 mmol,0.0125 m; c) in tert-
amyl alcohol (T=77 K).
and a broader signal at g=4.34,which is typical of a high-
spin FeIII species with rhombic distortion. After addition of
an equal amount of H
weakened in intensity,which suggests that complex forma-
tion between Fe and H2A(pydic) had occurred. When (S,S)-4a
2ACHTRE(UGN pydic) (Figure 3, g),both signals
CTHREUNG
(2.4 equiv with respect to Fe; Figure 3, b) was added,a
new signal at g=2.33 (g? ) appeared and the original signal
at g=1.98 almost disappeared. This suggests that formation
of a low-spin chiral (pydic)Fe–(S,S)-4a complex in solution
had occurred. Finally,when excess trans-stilbene (20 equiv
with respect to Fe; Figure 3, d) was added,only the signal
at g=4.34 remained. While it is not clear how the alkene
functionality contributes to the disappearance of the signal
at g? =2.33,we cannot rule out additional coordination to a
metal center or displacement of the chiral ligand from the
coordination sphere. Upon addition of only 0.1 mmol of hy-
drogen peroxide (33.32%,0.2 equiv with respect to the
alkene),all signals except that at g=4.34 almost completely
disappeared (Figure 4b). This suggested that formation of a
dimeric FeIII species had occurred,which could not be de-
tected due to the anti-ferromagnetic coupling of unpaired
electrons. We propose that the species with a g value of 4.34
is not an active catalyst for the epoxidation because neither
its magnitude nor its position changed much in the course of
the reaction.
Figure 5. The results of spin trapping experiments with A) TEMPO; c:
FeCl3·6H2O,H
(pydic),( S,S)-4a, 1a,TEMPO, T=77 K; b: after the
2A(pydic),
addition of H2O2 (0.5 mmol) and B) BPN; d: FeCl3·6H2O,H CHTREUNG
(S,S)-4a, 1a,BPN,H 2O2 (1.0 mmol), T=77 K; c: T=298 K. See the
Experimental Section for conditions.
drogen peroxide (30%,0.5 mmol) in
tert-amyl alcohol
(0.5 mL) was added to the reaction flask and stirred for a
few minutes,it took an induction period of about 30 s for
the expected brown coloration to appear. When we record-
ed the spectrum,the TEMPO signal had disappeared com-
pletely (Figure 5A, b). This finding is consistent with the
formation of transient carbon radicals during the reaction,
which could be trapped by the radical scavenger at a very
high rate (Scheme 2,left). The rate of combination of the
benzyl radical with TEMPO is known to approach diffusion
Chem. Eur. J. 2008, 14,7687 – 7698
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
7691

F. G. Gelalcha et al.
Kinetic studies
Direct kinetic measurement of
the asymmetricepoxidation of
styrene with hydrogen peroxide:
To obtain kinetic data for a rep-
resentative asymmetric epoxi-
dation at room temperature,
the epoxidation of styrene
(489.64 mm) with hydrogen per-
oxide (30%,25.0 m m) in the
Scheme 2. Possible paths for the spin trapping of a transient carbon radical by TEMPO (left) and BPN (right).
presence
of
(S,S)-5a
control.[25] However,the possibility of signal disappearance
due to other reactions of TEMPO,rather than the trapping
of radicals,under these reaction conditions cannot be ruled
out. Therefore,we repeated the experiment with a nitrone
spin trap,the radical adduct of which should show distinct
EPR signals.
(0.6001 mm),H
2ACTHRE(UNG pydic) (0.127 mm),and FeCl 3·6H2O
(0.250 mm) in tert-amyl alcohol was evaluated by using GC
with dodecane as the internal standard. The concentration
of styrene oxide against time was fitted by using first-order
exponential decay with the aid of a computer to obtain [sty-
rene oxide]1 (Figure 6). The observed pseudo-first-order
rate constant (kobs =3.972”10ꢀ4 sꢀ1) was obtained from a
plot of log([styrene oxide]1ꢀ[styrene oxide]t) versus time
(Figure 7),according to Equation (1):
Spin trapping with BPN: To support the argument above,
the spin-trapping experiment was repeated by using N-tert-
butylphenylnitrone (BPN; 0.5 mmol) as a radical trap,under
the same reaction conditions described for TEMPO
(Scheme 2,right). The EPR spectrum of this reaction mix-
ture at room temperature revealed no signals resulting from
nitroxyl radicals. Aqueous hydrogen peroxide (33.32%,
1.0 mmol) in tert-amyl alcohol was added in five portions
and the EPR spectrum was recorded after each addition.
The spectrum at 77 K was too broad for clear interpretation
(Figure 5B, d). At room temperature,however,the gradu-
al development of a triplet of doublets (the hyperfine cou-
pling constants for the nitrogen and hydrogen nuclei are
AN =15 and AH =3 G,respectively) that was assignable to a
radical adduct of BPN was observed (Figure 5B, c). The
signal recorded after the final addition is shown in Fig-
ure 5B. Although BPN is not specific to carbon radicals,this
complementary finding suggests that,under the experimen-
tal conditions,the formation of epoxides proceeds via the
formation of radical intermediates that can be trapped by
appropriate radical scavengers. The spin adducts are cur-
rently being characterized.
logð½styrene oxideŠ1ꢀ½styrene oxideŠtÞ ¼ ꢀkobstþconstant
ð1Þ
with which the catalysis rate constant (kcat =8.112”
10ꢀ4 mꢀ1 sꢀ1) was calculated from k
obs/ACHTREUNG
[styrene]init =kcat. Each
reaction was repeated three times.
To determine the relative ratios of (S,S)-5a and H
in the active catalyst,we plotted kcat versus [(S,S)-5a]/[H2-
(pydic)] (Figure 8; conditions: [FeCl3·6H2O]/[H2A(pydic)]=
1:1,[Fe] =0.25 mm,[styrene] init =489.64 mm,[H 2O2]init
30%,25.0 m m). This plot shows that the slowest measured
rate was at [(S,S)-5a]/[H2A
(pydic)]ꢂ2.0. When [(S,S)-5a]/[H2-
(pydic)]ꢂ2.5,a steady increase in reaction rate was ob-
served until a peak was reached at [(S,S)-5a]/[H2A(pydic)]
ꢂ3.5,which was followed by a slight fall at [( S,S)-5a]/[H2-
CHTREUNG
A
CHTREUNG
=
CTHREUNG
AHCTREUNG
CHTREUNG
(pydic)]ꢂ4.0. After this drop the rate increased again,
which suggests that more than one species was involved in
catalyzing this reaction.
A plot of kcat versus [H
2ACHTRE(UGN pydic)]/[Fe] (Figure 9; conditions:
[FeCl3·6H2O]/[(S,S)-5a]=1:2.4,[Fe] =0.25 mm,[styrene]
A
=
init
489.64 mm,[H 2O2]init =30%,25.0 m m) shows a maximum
Figure 7. Plot of log([styrene oxide]1ꢀ[styrene oxide]t) versus time,
which gives the pseudo-first-order rate constant as 3.972”10ꢀ4 sꢀ1 (y=
ꢀ3.972”10ꢀ4x+5.4342, R2 =9.9076”10ꢀ1).
Figure 6. A typical first-order exponential decay fitting of [styrene oxide]
versus time.
7692
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2008, 14,7687 – 7698

Asymmetric Epoxidation of Aromatic Alkenes
FULL PAPER
presence of FeCl3·6H2O/
A
asymmetric epoxidation of various para-substituted styrenes
(Table 4) was examined more closely. In this case,the
Table 4. Relative rate constants for the competitive epoxidation of para-
substituted styrenes with hydrogen peroxide and s values for the corre-
sponding substituents.
Figure 8. Plot of kcat versus [(S,S)-5a]/[H
(pydic)=1:1 and [Fe]=0.25 mm).
2ACHTRE(UGN pydic)] (FeCl3·6H2O/H2-
[b]
[b]
C
Entry
X
krel
logkrel
sJJ
smb
ACHTREUNG
1
2
3
4
5
H
Me
F
Cl
CF3
1.00
2.52
1.56
1.22
0.34
0
0
0.15
0
0.925
0.443
0.201
ꢀ1.077
ꢀ0.29
ꢀ0.24
0.11
ꢀ0.02
0.22
ꢀ0.01
0.49
[a] Conditions: FeCl3·6H2O (5 mol%),H
(12 mol%),H 2O2 (30%,0.5 mmol),
ref. [26]
2A
tert-amyl alcohol,RT. [b] See
energy of the transition state (TS) was expected to be affect-
ed by polar and spin-delocalization effects. These effects can
be quantified separately by employing the dual-parameter
Hammett correlation [Eq. (2)] developed by Jiang and Ji.[26]
Figure 9. A plot of kcat versus [H
1:2.4 and [Fe]=0.25 mm).
2ACHTRNEUG(pydic)]/[Fe] (FeCl3·6H2O/ACHRTE(UGN S,S)-5a=
logkrel ¼ 1JJCsJJCþ1mbsmb
ð2Þ
A straight line with a slope of 1.00 (R2 =0.999) was ob-
C
tained from
a
plot of logkrel versus 2.07sJJ ꢀ2.13smb
rate at [H
above,the rate dramatically decreased,which suggested that
the optimum [H2A(pydic)]/[Fe] ratio (X) in the active species
2A
N
(Figure 11). This linear free-energy correlation with polar
CHTREUNG
was 0.8ꢃXꢃ1.0. The linear dependence of kcat on
[FeCl3·6H2O] was established from a plot of kcat versus
[FeCl3·6H2O] (Figure 10).
Figure 11. Dual-parameter Hammett correlation for the epoxidation of
styrene and para-substituted styrenes,according to Table 4. y=1.0025x,
R2 =0.9994, j1mb/1JJ j=1.03.
C
and spin-delocalization effects suggested the formation of a
benzylic radical species in the TS of the arylalkene epoxida-
Figure 10. A plot of kcat versus [FeCl3·6H2O] (FeCl3·6H2O/
(pydic)=1:0.8:2.8).
ACHTRE(UGN S,S)-5a/H2-
ACHTREUNG
C
tion. The large positive 1JJ value of 2.07 suggests that the re-
action is promoted by spin delocalization at the radical
center,whereas the negative 1mb value of ꢀ2.13 is consistent
with the electrophilic nature of the active oxidant. However,
Competitive asymmetricepoxidation of para-substituted styr-
enes with hydrogen peroxide: To assess whether the forma-
tion of a carbon radical intermediate is involved in the epox-
idation of aromatic alkenes with hydrogen peroxide in the
C
the magnitude of j1mb/1JJ j,1.03,suggests that polar sub-
stituent effects are slightly more important than spin-deloc-
alization effects.[26]
Chem. Eur. J. 2008, 14,7687 – 7698
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
7693

F. G. Gelalcha et al.
Competitive asymmetricepoxidation of deuterated styrenes
with hydrogen peroxide: For the determination of the secon-
dary kinetic isotope effect (KIE),the competitive epoxida-
tions of styrene and deuterated styrenes (a-[D]styrene and
b-[D2]styrene) were conducted under the conditions shown
in Schemes 3 and 4.
Figure 12. Graph of the ee of 2a versus the ee of 5a,which shows a posi-
Scheme 3. The competitive epoxidation of styrene and b-[D2]styrene.
&
tive NLE. b: Theoretical curve, : measured data points.
a) FeCl3·6H2O (5 mol%),H
2ACHTRE(UGN pydic) (5 mol%),( S,S)-5a (12 mol%),
H2O2 (30%,57 mL,0.5 mmol), tert-amyl alcohol,RT.
smoothly to give a quantitative yield of the unlabeled epox-
ide (Scheme 5) in which no trace of 18O-labeled 2a was de-
tected. A trace amount of benzaldehyde was observed in
Scheme 4. The competitive epoxidation of styrene and a-[D]styrene.
a) FeCl3·6H2O (5 mol%),H
2ACHTRE(UGN pydic) (5 mol%),( S,S)-5a (12 mol%),
H2O2 (30%,0.5 mmol), tert-amyl alcohol,RT.
Scheme 5. Reaction scheme for the isotope dilution studies.
a) FeCl3·6H2O (5 mol%),H
2ACHTRE(UGN pydic) (5 mol%),( S,S)-5a (12 mol%),
H2O2 (50%,1.19 mmol),H 18O (10 mmol), tert-amyl alcohol,RT.
2
The observation of a significant inverse secondary KIE in
the case of b-[D2]styrene (kH/kD =0.937ꢁ0.0090) and only a
small KIE (kH/kD =0.970ꢁ0.0073) in the case of a-[D]sty-
which the label was partly found (16O/18O=88:22). These re-
sults suggest that formation of benzaldehyde might take
place in a consecutive reaction that starts from an inter-
mediate of the catalytic cycle. In addition,the results suggest
that either no high-valent Fe=O species are involved as the
active oxidant or,more likely,that the rate of 18O exchange
with water is too low compared with the rate of 18O transfer
to the organic substrate at this concentration.
rene is consistent with an unsymmetrical TS in which the
3
ꢀ
C O bond formation (i.e.,sp hybridization) is more ad-
vanced at the b-carbon than at the a-carbon,which implies
a stepwise mechanism. Assuming a high-valent Fe=O inter-
mediate as the active oxidizing species,this finding again
suggests that epoxidation under this system may involve the
formation of benzylic radicals in the rate-determining step.
The number and nature of the active oxidant(s) is current-
ly under investigation. However,based on the above obser-
vations and the wealth of accumulated literature that impli-
cates high-valent Fe=O species as the actual active oxygen
transfer agents in biological mono-oxygenations in both
heme and non-heme iron enzymes and model systems,[28a–g]
it is reasonable to assume that the formation of reactive
Nonlinear effect (NLE) studies: Since its discovery in the
1980s,the NLE has been successfully applied as a mechanis-
tic tool in the interpretation of many catalytic asymmetric
reactions.[27] The asymmetric epoxidation of trans-stilbene
was performed with mixtures of (S,S)-5a and (R,R)-5a with
varying ee values. A plot of the ee of 2a versus the ee of 5a
revealed a small but clearly positive NLE (Figure 12). Fol-
lowing Kaganꢁs ML2 and reservoir models,[27] the source of
the +NLE may be the formation of a less-reactive meso
dimer. This indirectly suggests the existence of complex
equilibria that involve several chiral Fe complexes. The rela-
tive concentrations and reactivities of these species deter-
mine the observed enantioselectivity of the reaction.
high-valent Fe species,such as [(pydic)(L*) Fe=O] (n=1–2),
n
has occurred.
Assuming further that a top-on approach of the alkene to
such a species is preferred over the alternative and more
popular side-on approach,the following mechanistic path-
way is tentatively proposed to explain the enantioselectivi-
ties. Accordingly,the Fe =O species generated by using non-
benzylated ligands (S,S)-4a–c and their N-benzylated coun-
terparts (R,R)-5a–d would be preferentially approached
from the Re,Re face (Scheme 6,bottom),whereas the spe-
cies generated by using the corresponding enantiomers of
these ligands is approached from the Si,Si face (Scheme 6,
18O isotope dilution studies: When the asymmetric epoxida-
tion of trans-stilbene was conducted under standard condi-
tions in the presence of a large excess of H218O (250 mmol,
500 equiv),no reaction took place. However,with a smaller
excess of H218O (10 mmol),the reaction takes places
ꢀ ꢀ
top) of the trans-aryl alkene. The resulting L*Fe O aryl
7694
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2008, 14,7687 – 7698

Asymmetric Epoxidation of Aromatic Alkenes
FULL PAPER
C
ty that OH radicals play any
significant role in the current
alkene epoxidation because of
the unusually high yields and ee
values of the epoxides. A UV/
Vis study of the catalyst mix-
ture with (S,S)-4a as the chiral
ligand revealed a species that
had an absorbance peak at l=
465 nm upon the addition of hy-
drogen peroxide,which is cur-
rently being characterized. The
epoxidation reaction appears to
proceed via radical intermedi-
ates that could be intercepted
by using radical scavengers. An
inverse secondary isotope effect
(KIE) of kH/kD =0.93 for the b-
carbon and kH/kD =0.97 for the
a-carbon of styrene suggests an
unsymmetrical TS in which
Scheme 6. Proposed mechanistic pathways to explain the enantioselectivity of this reaction.
radical species would collapse by a rebound mechanism to
give the experimentally observed optically active (S,S)- and
(R,R)-epoxides,respectively. The alternative pathway,in
which rotation precedes collapse,is unlikely because we did
not observe the corresponding diastereomeric cis-epoxides.
Therefore,enantioselectivity is determined by the efficiency
of the first step. While this mechanism seems plausible,
steric factors on the substrate can dramatically alter both
the reaction rate and face selectivity,as previously ob-
served.[17]
oxygen transfer is nonconcerted. The epoxidation of para-
substituted styrenes revealed a linear dual-parameter Ham-
mett plot with a slope of 1.00,which suggests that the for-
mation of a benzylic radical species in the TS had occurred.
Epoxidation of trans-stilbene in the presence of excess
H218O resulted in an absence of the label in the epoxide
product,which is probably due to a much slower rate of
label exchange compared with O transfer from a high-valent
Fe=O species. The reaction showed a small but significant
NLE,which suggested the participation of several chiral Fe
complexes in catalyzing the reaction,the relative concentra-
tions and reactivities of which determine the observed ee
value. Efforts to identify these species and study the mecha-
nism in greater detail are under way. Although much re-
mains to be done in terms of improving ee values and in-
creasing substrate scope,the achievement of a significant
asymmetric alkene epoxidation with simple iron salts and
hydrogen peroxide at room temperature serves as a useful
biomimetic functional model for non-heme iron alkene
mono-oxygenases[28h–i] and helps stimulate further research
in this field.
Conclusion
The Fe-catalyzed asymmetric epoxidation of aromatic al-
kenes with hydrogen peroxide as the oxygen source and a
chiral Fe catalyst generated from FeCl3·6H2O,H 2ACHTER(UNG pydic),
and the novel,readily accessible,chiral ( S,S)-5a has been
extended to sterically less demanding alkenes. The asymmet-
ric epoxidation of styrenes gives almost complete conver-
sions,but lower ee values compared with larger alkenes.
(S,S)-5a and (S,S)-4a (the non-benzylated precursor of
(S,S)-5a) give opposite enantiomers of trans-stilbene oxide
((R,R)-2a and (S,S)-2a,respectively). cis-Alkenes are gener-
ally poorer substrates than trans-alkenes for the current
asymmetric epoxidation,and steric factors are more impor-
tant than electronic factors in controlling the enantioselec-
tivity. Mass spectrum signals corresponding to Fe complexes,
Experimental Section
General: A Leco CHNS-932 analyzer was used for elemental analyses.
An AMD 402/3 mass spectrometer was used for recording mass spectra.
High-resolution mass spectrometry experiments were performed by using
an Agilent Series 1200 HPLC system and an Agilent 1969A time-of-flight
mass spectrometer. The TOF-MS conditions (positive-ion mode) with a
dual sprayer API-ES source were as follows: nebulizer and drying gas,ni-
such as [FeCl(L*)2ACHTREUNG(pydic)]·HCl in which L*=(S,S)-4a or
(S,S)-5a,were identified by ESIMS studies of the catalyst
trogen; nebulizer pressure=241.313 kPa,drying gas flow =12 Lminꢀ1
drying gas temperature=3008C,capillary voltage =4 kV,fragmentor
voltage=225 V,skimmer voltage =60 V,octopole voltage =250 V,and
,
mixture. Hydrogen peroxide may react with this species to
initially form [FeIII(L*)
2ACTHERNGU(OOH)ACHTRE(UNG pydic)]. Homolysis or heter-
ꢀ
olysis of the O O bond of such a species would generate
mass reference (m/z): 121.050873 and 922.009798. The mobile phase con-
sisted of 10% H2O (0.1% HCOOH) and 90% MeOH at a flow rate of
0.5 mLminꢀ1. GC analysis was performed by using a Hewlett Packard
HP 6890 gas chromatograph with a flame ionization detector (FID). UV/
IV
V
ꢀ
C
Fe =O+ OH radicals and Fe =O+ OH,respectively. The
high-valent Fe=O species are more likely oxidants than h2-
III
C
LFe O2H and OH radicals. We have ruled out the possibili-
Chem. Eur. J. 2008, 14,7687 – 7698
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
7695

F. G. Gelalcha et al.
Vis spectroscopy measurements were carried out by using an Avantes
fiber-optic spectrometer (AVASPEC) with a fiber-optic sensor in trans-
mission against the solvent (tert-amyl alcohol or acetonitrile) as the refer-
ence. FTIR spectroscopy measurements were conducted by using a fiber-
optic Diamant-ATR probe,coupled with an FTIR-spectrometer (Nicolet
Avatar 370). Raman spectroscopy measurements were carried out by
using a fiber-optical RXN spectrometer (Kaiser Optical Systems,laser
l=785 nm) with the laser power set at 100 mW. UV/Vis,FTIR and
Raman spectroscopic data were collected simultaneously. EPR investiga-
tions were conducted separately with a Bruker CW-EPR spectrometer
ELEXSYS 500-10/12 in X-Band with the following settings: microwave
frequency=9.5 GHz,power =6.3 mW,modulation frequency =100 kHz,
and amplitude=0.5 mT. Alkenes 1a, 1d,and 1l–s; ligands (S,S)-3, (S,S)-
4a,and ( R,R)-4a; pyridine-2,6-dicarboxylic acid; FeCl3·6H2O; hydrogen
peroxide;[32] the aldehydes and arenesulfonyl chlorides used for the syn-
thesis of the ligands; the spin traps and the dry solvents are commercial
products (Sigma-Aldrich,Fluka,Merck) and were used as received. Al-
kenes 1b, 1c,and 1e–i were synthesized by a McMurry[29] coupling of the
corresponding alkyl-substituted benzaldehydes and obtained in high
yields and purities. Alkenes 1j and 1v were synthesized by the Heck re-
was achieved after this time (determined by GC and/or TLC monitoring).
For preparative purposes,excess peroxide was eliminated by adding a sa-
turated aqueous solution of sodium sulfite (1 mL) and shaking well.
After addition of diethyl ether (10 mL),the phases were separated,and
the aqueous phase was extracted with diethyl ether (3”10 mL). The com-
bined organic phases were then dried over anhydrous MgSO4. After fil-
tration and solvent removal by rotary evaporator,the crude product was
either directly analyzed by chiral HPLC to determine its ee or purified
by silica gel chromatography on a short column (eluent: hexane/ethyl
acetate 20:1,1% Et N) for full characterization.
3
General procedure for catalyst NLEstudies of asymmetric epoxidation
with hydrogen peroxide: Compound (R,R)-5a (24.66 mg,0.054 mmol),
(S,S)-5a (2.74 mg,0.006 mmol),H
FeCl3·6H2O (6.76 mg,0.025 mmol),and
0.500 mmol) were heated in tert-amyl alcohol (9.0 mL) in
(pydic) (4.17 mg,0.025 mmol),
trans-stilbene (90.1 mg,
25 mL
a
Schlenk tube until a clear yellow solution formed (ꢂ1 min). After the re-
action mixture had cooled to RT,dodecane (GC internal standard,
100 mL) was added. Aqueous hydrogen peroxide (30%,113 mL,
1.0 mmol) in tert-amyl alcohol (887 mL) was added to this reaction mix-
ture over a period of 1 h by using a syringe pump. After this addition,ali-
quots were taken from the reaction mixture and subjected to GC analysis
for the determination of yield and conversion data. The reaction mixture
was then quenched with an aqueous solution of Na2SO3 (ꢂ10 mL),ex-
tracted with dichloromethane (2”10 mL),and washed with water
action of 4-tert-butylbromobenzene with 2-vinylnaphthalene or styrene,
[30]
respectively,by modifying the method of Chandrasekhar et al.
Alkenes
1k and 1t were synthesized by protecting trans-cinnamyl alcohol with tri-
phenylsilyl chloride or tert-butyldimethyl silylchloride,respectively,in the
presence of pyridine. All racemic epoxides 2,which were necessary refer-
ences for the chiral HPLC data,were synthesized by epoxidation with
metachloroperbenzoic acid (MCPBA). Typically,one to two equivalents
of a solution of MCPBA in dichloromethane (CH2Cl2) were added drop-
wise to an ice-cooled solution of the alkene in CH2Cl2. After stirring
overnight at RT,the solvent was removed under reduced pressure and
the residue was purified by flash chromatography to give high yields of
the corresponding epoxides (see the Supporting Information for analyti-
cal data).
General procedure for the synthesis of (S,S)-4b–g:[31] A solution of the
corresponding arene sulfonyl chloride (4.86 mmol,1.0 equiv) in CH 2Cl2
(5 mL) was added dropwise with stirring to a solution of diamine (S,S)-3
(1.032 mg,4.86 mmol) and diisopropylethyl amine (DIPEA,0.923 mL) in
CH2Cl2 (10 mL) at 08C over 30 min. Then the mixture was stirred at RT
for 5 h. The progress of the reaction was monitored by TLC. After com-
plete conversion,water (5 mL) was added and the phases separated. The
aqueous phase was then extracted twice with CH2Cl2 (5 mL). The com-
bined organic phases were dried,concentrated,and purified by silica gel
chromatography and characterized (see the Supporting Information).
A
evaporated under reduced pressure to give the crude epoxides,which
were then dissolved in n-hexane for HPLC measurement.
General procedure for the competitive asymmetric epoxidation of para-
substituted styrene with hydrogen peroxide: Compound (S,S)-5a
(27.42 mg,0.060 mmol),H
2ACHTRE(UGN pydic) (4.17 mg,0.025 mmol),and
FeCl3·6H2O (6.78 mg,0.025 mmol) were heated in tert-amyl alcohol
(9.0 mL) in a 25 mL Schlenk tube until a clear yellow solution formed
A
(2.5 mmol), p-methylstyrene (2.5 mmol),and dodecane (GC internal
standard,100 mL) were added. Aqueous hydrogen peroxide (30%,57 mL,
0.5 mmol) in tert-amyl alcohol (276 mL) was added to this reaction mix-
ture over a period of 1 h by using a syringe pump. After the addition,ali-
quots were taken from the reaction mixture and subjected to GC analysis
for the determination of yield and conversion data.
General procedure for the competitive asymmetric epoxidation of deuter-
ated styrenes with hydrogen peroxide
For styrene and b-[D2]styrene: Compound (S,S)-5a (27.42 mg,
0.060 mmol),H
(pydic) (4.17 mg,0.025 mmol),and FeCl 3·6H2O (6.78 mg,
25 mL
Synthesis of (R,R)-5a, (S,S)-5b–i: These ligands were synthesized by
treating the corresponding N-amino-N’-arenesulfonyl diphenyl ethylene
diamine ligands and benzaldehyde ((R,R)-5a, (S,S)-5b–g),or 4-methyl-
benzaldehyde ((S,S)-5i),according the reductive alkylation procedure
given for the synthesis of (S,S)-5a[17a] (see the Supporting Information for
analytical data).
0.025 mmol) were heated in tert-amyl alcohol (9.0 mL) in
a
Schlenk tube until a clear yellow solution formed (ꢂ1 min). After the re-
action mixture had cooled to RT,styrene (2.5 mmol), b-[D2]styrene
(2.5 mmol),and dodecane (GC internal standard,100 mL) were added.
To this reaction mixture,aqueous hydrogen peroxide (30%,57
mL,
0.5 mmol) in tert-amyl alcohol (276 mL) was added over a period of 1 h
by using a syringe pump. After the addition,aliquots were taken from
the reaction mixture and subjected to GC analysis for the determination
of yield and conversion data. The reaction mixture was then quenched
with an aqueous solution of Na2SO3 (ꢂ10 mL),extracted with dichloro-
methane (2”10 mL),and washed with water ( ꢂ20 mL). The combined
organic layers were dried over MgSO4 and evaporated under reduced
pressure to give the crude epoxides,which were then purified by silica
gel chromatography (70–230 mesh,neutralized with 1% Et 3N) with
hexane to hexane/ethyl acetate (100:3) as the gradient eluent. The selec-
tivity was determined by 1H NMR spectroscopy. Each reaction was re-
peated three times.
General procedure for the Fe-catalyzed asymmetric epoxidation of al-
kenes 1a–y: Pyridine-2,6-dicarboxylic acid (4.24 mg, 0.025 mmol), ferric
chloride hexahydrate (6.76 mg,0.025 mmol),ligands ( S,S)-3, (S,S)-4a–g,
or (S,S)-5a–i (0.06 mmol),and alkenes 1a–y (0.5 mmol) were mixed in
tert-amyl alcohol (9 mL) and stirred at RT for about 30 min. The result-
ing mixture usually assumed a pale yellow color. For the GC determina-
tion of yields and conversions,dodecane (100 mL) was added as the inter-
nal standard. After taking samples for GC analysis,aqueous hydrogen
peroxide (“30%”,1 mmol) in tert-amyl alcohol (1 mL) was added to this
mixture over one hour by using a syringe pump. [A generally accurate
volume of the “30%” solution can not be given in this case because the
peroxide content (and thus the density) of this material varies considera-
bly with time. Therefore,before each experiment we determined the per-
oxide content (%) by titration. For multiple runs we weighed the amount
corresponding to 10 mmol H2O2 into a 10 mL volumetric flask,filled with
the solvent to volume and withdrew the required amount (1 mL,1 mmol)
of the solution using by a syringe (see also ref. [32]). For single runs
1 mmol H2O2 may also be directly diluted with solvent to 1 mL before in-
troduction into the reaction flask.] In most cases,complete conversion
For styrene and a-[D]styrene: The procedure was the same as described
above,except that a-[D]styrene (2.5 mmol) was used instead of b-
[D2]styrene.
General procedure for asymmetric epoxidation with hydrogen peroxide
in the presence of H218O: Compound (S,S)-5a (27.42 mg,0.060 mmol),
H2ACHTRE(UGN pydic) (4.17 mg,0.025 mmol),FeCl 3·6H2O (6.78 mg,0.025 mmol),and
trans-stilbene (90.1 mg,0.500 mmol) were heated in tert-amyl alcohol
(9.0 mL) in a 25 mL Schlenk tube until a clear yellow solution formed
7696
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2008, 14,7687 – 7698

Asymmetric Epoxidation of Aromatic Alkenes
FULL PAPER
A
of 2 min. After each addition,the mixture was stirred well and the spec-
tra were recorded (Figure 4).
2
10.0 mmol) and dodecane (GC internal standard,100 mL) were added. To
this reaction mixture,an aqueous solution of hydrogen peroxide (50%,
68 mL,1.19 mmol) in tert-amyl alcohol (932 mL) was added over a period
of 1 h by using a syringe pump. After this addition,aliquots were taken
from the reaction mixture and subjected to GC-FID analysis for determi-
nation of yield and conversion data. The reaction mixture was then fur-
ther analyzed by GC-MS.
General procedure for spin trapping with BPN: FeCl3·6H2O (6.76 mg,
0.025 mmol),H
2A
S,S)-4a (22.01 mg,
0.06 mmol),and trans-stilbene (90.15 mg,0.5 mmol) were added to a
Schlenk flask and dissolved in tert-amyl alcohol (2 mL) by heating to
about 60–708C for 10 min with stirring. A sample was taken and the
EPR spectrum of this solution was recorded at 77 K and at RT. The
sample was then returned to the reaction vessel. BPN (88.5 mg,
0.5 mmol) was added to the reaction mixture and the EPR spectrum was
recorded again. No signal resulting from nitroxyl radicals was observed.
The EPR sample was then returned to the reaction flask. Aqueous H2O2
(30%,103.1 mL,1.0 mmol) in tert-amyl alcohol (total volume 1 mL) was
added in five portions (3”100 mL,0.1 mmol; 1”200 mL,0.2 mmol; 1”
500 mL,0.5 mmol) with stirring for a few minutes after each addition
before the EPR spectra were recorded. Figure 5B shows the spectra re-
corded at 77 K and RT after the last addition.
General procedure for the direct kinetic measurement of the asymmetric
epoxidation of styrene with hydrogen peroxide: Compound (S,S)-5a
(27.40 mg,0.060 mmol),H
FeCl3·6H2O (6.75 mg,0.025 mmol) were heated in tert-amyl alcohol
(10 mL) in a 50.00 mL volumetric flask until a clear yellow solution
formed (ꢂ1 min.). After the reaction mixture had cooled to RT,the cata-
lyst solution was diluted up to the 50.00 mL mark with tert-amyl alcohol.
Dodecane (GC internal standard,100 mL,73.77 mg,0.433 mmol),styrene
(573 mL,509.96 mg,4.896 mmol),and the catalyst solution (5.00 mL)
were added to a 10.00 mL volumetric flask. It was then diluted up to the
10.00 mL mark with tert-amyl alcohol. Aqueous hydrogen peroxide
(30%,29.0 mL,0.250 mmol) was added at once and the reaction mixture
was vigorously shaken. This solution was transferred to a GC vial for
GC-FID analysis,and the styrene oxide concentration was determined at
approximately 4.5 min intervals at RT ((23.9ꢁ0.99)8C). The concentra-
tion of styrene oxide versus time was then fitted with first-order expo-
nential decay by using the software Origin 7.5 SR5 [v 7.5870 (B870)][33] to
obtain [styrene oxide]1. The observed rate constant was obtained by
plotting log([styrene oxide]1ꢀ[styrene oxide]t) versus time (see [Eq. (1)]
General procedure for spin trapping with TEMPO: FeCl3·6H2O
(6.76 mg,0.025 mmol),H
(22.01 mg,0.06 mmol),and
2A
S,S)-4a
added to a Schlenk tube and dissolved in tert-amyl alcohol (2 mL) by
heating to about 60–708C for 10 min with stirring. A sample was taken
and the EPR spectrum of this solution was recorded at 77 K and at RT.
The sample was returned to the reaction mixture. A solution of TEMPO
(0.2 mL,6.41”10 ꢀ4 m,1.282”10 ꢀ4 mmol) in tert-amyl alcohol (prepared
by dissolving TEMPO (1 mg) in tert-amyl alcohol (10 mL)) was added to
the reaction mixture. A sample was then taken and the EPR spectrum,
which was dominated by the signal from TEMPO,was recorded. The
sample was returned to the reaction flask. An aqueous solution of H2O2
(30%,52 mL,0.5 mmol) in tert-amyl alcohol (0.5 mL) was added to the
reaction flask immediately and stirred for a few minutes. After an induc-
tion period of around 30 s the solution turned brown,as expected. A
sample was then taken and the spectrum was recorded at 77 K. The
TEMPO signal disappeared completely from the spectrum (Figure 5A).
and kobsACHTREUNG[styrene]init =kcat). Each reaction was repeated three times.
General procedure for the simultaneous UV/Vis, Raman, and IR spectro-
scopic investigation of the catalyst mixture in the absence of 1a:
FeCl3·6H2O (6.76 mg,0.025 mmol),H 2ACTHRE(UNG pydic) (4.24 mg,0.025 mmol),and
(S,S)-4a (22.01 mg,0.06 mmol) in tert-amyl alcohol (6 mL) were added to
a 10 mL glass beaker. The UV/Vis,Raman,and IR spectroscopic data
were recorded simultaneously by passing UV light through a quartz cuv-
ette mounted in the solution and by inserting an IR sensor into the sol-
vent bulk. Then aqueous H2O2 (33.32%,45.6 mmL,0.5 mmol) in tert-
amyl alcohol (total volume 0.5 mL) was added in five portions (5”
100 mL,0.1 mmol) in intervals of 2 to 3 min. After each addition,the mix-
ture was stirred well with a magnetic stir bar and the spectra were re-
corded (Figure 1).
Acknowledgements
The authors thank the Deutsche Forschungsgemeinschaft,the State of
Mecklenburg-Western Pomerania,and the Federal Ministry of Education
and Research (BMBF) for financial support and the analytical staff at
LIKAT,lead by Dr. C. Fischer,for recording spectra and chromatograms.
Dr. D. Goerdes and E. Schmidt at CELISCA are acknowledged for their
assistance in measuring the ESI MS spectra.
General procedure for the simultaneous UV/Vis, Raman and IR spectro-
scopic investigation of the catalyst mixture in the presence of 1a:
FeCl3·6H2O (6.76 mg,0.025 mmol),H
(S,S)-4a (22.01 mg,0.06 mmol),and 1a (90.15 mg,0.5 mmol) in tert-amyl
alcohol (6 mL) were added to 10 mL glass beaker. The UV/Vis,
Raman,and IR data were recorded were simultaneously by passing UV-
light through a quartz cuvette mounted in the solution and by inserting
an IR sensor into the solvent bulk. Then aqueous H2O2 (33.32%,
91.13 mL,1.0 mmol) in tert-amyl alcohol (total volume 1 mL) was added
in ten portions (10”100 mL,0.1 mmol) in intervals of approximately
2 min. After each addition,the mixture was stirred well with a magnetic
stir bar and the spectra were recorded. Results from the first five por-
tions are shown in Figure 2. No significant changes occurred upon the ad-
dition of portions six to ten,and the corresponding curves are omitted
for clarity.
[1] a) K. A. Jørgensen in Transition Metals for OrganicSynthesis (Eds.:
M. Beller,C. Bolm),Wiley-VCH,Weinheim,
b) K. Furuhashi in Chirality in Industry (Eds.: A. N. Collins,G. N.
Sheldrake,J. Crosby),John Wiley,Chichester, 1992,p. 167; c) U.
1998,Vol. 2,p. 157;
Sundermeier,C. Dçbler,M. Beller in Modern Oxidation Methods
(Ed.: J. E. Bäckvall),Wiley-VCH,Weinheim, 2004,p. 1.
[2] a) M. Tokunaga,J. F. Larrow,F. Kakiuchi,E. N. Jacobsen,
Science
1997, 277,936–938; b) A. Gayet,S. Bertilsson,P. G. Anderson, Org.
Lett. 2002, 4,3777–3779.
[3] For reviews on the asymmetric epoxidation of allylic alcohols,see:
a) T. Katsuki in Comprehensive AsymmetricCatalysis, Vol. 2 (Eds.:
Procedure for the EPR spectroscopic investigation of the reaction mix-
ture in tert-amyl alcohol: tert-Amyl alcohol (2 mL) was added to a
Schlenk flask and the EPR spectrum was recorded. Then FeCl3·6H2O
E. N. Jacobsen,A. Pfaltz,H. Yamamoto),Springer,Berlin,
1999,
(6.76 mg,0.025 mmol),H
(22.01 mg,0.06 mmol),and
2A
S,S)-4a
pp. 621–648; b) R. A. Johnson,K. B. Sharpless in CatalyticAsym-
metricSynthesis (Ed.: I. Ojima),Wiley-VCH,New York,
1993,
added sequentially with stirring,to make sure that every component had
been fully dissolved. Slight heating and cooling to RT was necessary. The
EPR spectra were recorded at RT and at 77 K,and the sample was re-
turned to the reaction mixture after every measurement. Figure 3 shows
spectra recorded at T=77 K. Then aqueous H2O2 (33.32%,91.13 mL,
1.0 mmol) in tert-amyl alcohol (total volume 1 mL) was added in 6 por-
tions (5”=100 mL,0.1 mmol each,then 1”500 mL,0.5 mmol) at intervals
pp. 103–158.
[4] For reviews,see: a) E. N. Jacobsen,M. H. Wu in
Comprehensive
AsymmetricCatalysis, Vol. II (Eds.: E. N. Jacobsen,A. Pfaltz,H. Ya-
mamoto),Springer,Berlin, 1999,pp. 649–677; b) T. Katsuki in Com-
prehensive Coordination Chemistry II, Vol. 9 (Eds.: J. A. McCleverty,
T. J. Meyer),Elsevier Science,Oxford, 2003,pp. 207–264.
[5] Y. Shi, Acc. Chem. Res. 2004, 37,488–496.
Chem. Eur. J. 2008, 14,7687 – 7698
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
7697

F. G. Gelalcha et al.
[6] a) For a recent review,see: D. R. Kelly,S. M. Roberts, Biopolymers
2006, 84,74–89; b) For related peptide-catalyzed asymmetric epoxi-
dations,see: B. List,D. Kampen, Synfacts 2007,0983.
[7] For recent reviews,see: a) F. G. Gelalcha, Chem. Rev. 2007, 107,
3338–3361; b) S. Jaroch,H. Weinmann,K. Zeitler, ChemMedChem
2007, 2,1261–1264; c) A. Berkessel, Pure Appl. Chem. 2005, 77,
1277–1284.
[22] Z.-X. Wang,Y. Tu,M. Frohn,J.-R. Zhang,Y. Shi, J. Am. Chem. Soc.
1997, 119,11224–11235.
[23] The use of tert-amyl alcohol as the solvent was found to be unsuita-
ble for the ESIMS measurements.
[24] a) M. L. Ferreira, Macromol. Biosci. 2003, 3,179–188; b) M. R. Bu-
kowski,S. Zhu,K. D. Koehntop,W. W. Brennessel,L. Que,Jr.,
Biol. Inorg. Chem. 2004, 9,39–48; c) M. Costas,M. P. Mehn,M. P.
Jensen,L. Que,Jr., Chem. Rev. 2004, 104,939–986.
J.
[8] I. W. C. E. Arends, Angew. Chem. 2006, 118,6398–6400; Angew.
Chem. Int. Ed. 2006, 45,6250–6252.
[25] M. Newcomb, Tetrahedron 1993, 49,1151–1176.
[9] Y. Sawada,K. Matsumoto,S. Kondo,H. Watanabe,T. Ozawa,K.
Suzuki,B. Saito,T. Katsuki, Angew. Chem. 2006, 118,3558–3560;
Angew. Chem. Int. Ed. 2006, 45,3478–3480.
[26] a) X.-K. Jiang, Acc. Chem. Res. 1997, 30,283–289; b) C.-J. Liu,W.-
Y. Yu,C.-M. Che,C.-H. Yeung, J. Org. Chem. 1999, 64,7365–7374.
[27] a) H. B. Kagan, Oil & Gas Science and Technology Rev. IFP 2007,
62,731–738; Gas Science and Technology Rev. IFP 2007,62,731–
738; b) H. B. Kagan, Synlett 2001,888–899; c) C. Puchot,O.
[10] M. Colladon,A. Scarso,P. Sgarbossa,R. A. Michelin,G. Strukul,
Am. Chem. Soc. 2006, 128,14006–14007.
J.
[11] M. K. Tse,C. Dçbler,S. Bhor,M. Klawonn,W. Mägerlein,H. Hugl,
Samuel,E. Dunach,S. Zhao,C. Agami,H. B. Kagan,
J. Am. Chem.
M. Beller, Angew. Chem. 2004, 116,5367–5372; Angew. Chem. Int.
Ed. 2004, 43,5255–5260.
[12] For an excellent review on the current status of Fe-catalyzed organic
Soc. 1986, 108,2353–2357.
[28] For recent leading references,see: a) W. Nam, Acc. Chem. Res.
2007, 40,522–531; b) X. Shan,L. Que,Jr., J. Inorg. Biochem. 2006,
100,421–433; c) J. T. Groves, J. Inorg. Biochem. 2006, 100,434–437;
d) A. Decker,M. D. Clay,E. I. Solomon, J. Inorg. Biochem. 2006,
100,697–706; e) I. V. Korendovych,S. V. Kryatov,E. V. Rybak-Aki-
mova, Acc. Chem. Res. 2007, 40,510–521; f) R. Mas-Ballestꢂ,L.
Que,Jr., J. Am. Chem. Soc. 2007, 129,15964–15972; g) S. H. Lee,
J. H. Han,H. Kwak,E. Y. Lee,H. J. Kim. ,J. H. Lee,S. N. Lee,Y.
Kim,C. Kim, Chem. Eur. J. 2007, 13,9393–9398; h) F. J. Small,S. A.
Ensign, J. Biol. Chem. 1997, 272,24913–24920; i) V. Champreda,
N. Y. Zhou,D. J. Leak, FEMS Microbiol. Lett. 2004, 239,309–318.
[29] J. E. McMurry, Chem. Rev. 1989, 89,1513–1524.
reactions,see: C. Bolm,J. Legros,J. Le Paih,L. Zani,
Chem. Rev.
2004, 104,6217–6254.
[13] E. Rose,Q.-Z. Ren,B. Andrioletti, Chem. Eur. J. 2004, 10,224–230.
[14] Q. F. Cheng,X. Y. Xu,W. X. Ma,S. J. Yang,T. P. You,
Chinese
Chem. Letters 2005, 16,1467–1470.
[15] M. B. Francis,E. N. Jacobsen, Angew. Chem. 1999, 111,987–991;
Angew. Chem. Int. Ed. 1999, 38,937–941.
[16] C. Marchi-Delapierre,A. Jorge-Robin,A. Thibon,S. Mꢂnage Chem.
Commun. 2007,1166–1168.
[17] a) F. G. Gelalcha,B. Bitterlich,G. Anilkumar,M. K. Tse,M. Beller,
Angew. Chem. 2007, 119,7431–7435; Angew. Chem. Int. Ed. 2007,
46,7293–7296; b) F. G. Gelalcha,G. Anilkumar,M. K. Tse,M.
Beller,Abstract of papers from the 234th ACS National Meeting &
Exposition, Boston, MA,August 19–23, 2007.
[30] S. Chandrasekhar,Ch. Narsihmulu,S. S. Sultana,N. R. Reddy,
Lett. 2002, 4,4399–4401.
[31] D. Xue,Y.-C. Chen,X. Cui,Q.-W. Wang,J. Zhu,J.-G. Deng,
Chem. 2005, 70,3584–3591.
Org.
J. Org.
[18] a) D. Yang,M.-K. Wong,Y.-C. Yip,X.-C. Wang,M.-W. Tang,J.-H.
Zheng,K.-K. Cheung, J. Am. Chem. Soc. 1998, 120,5943–5952;
b) I. Moretti,G. Torre, Tetrahedron Lett. 1969, 10,2717–2720.
[19] Z.-L. Wei,Z.-Y. Li ,G.-Q. Lin, Tetrahedron 1998, 54,13059–13072.
[20] Y. Gao,J. M. Klunder,R. M. Hanson,H. Masamune,S. Y. Ko,K. B.
Sharpless, J. Am. Chem. Soc. 1987, 109,5765–5780.
[32] “30%” aqueous H2O2 (Merck) was used as received. The peroxide
content varied considerably (0 to 40%),as determined by titration;
50% H2O2 was obtained from Aldrich.
[33] Origin 7.5 SR5 [v 7.5870 (B870)] see: https://www.originlab.com/.
[34] A. O. Huque, Pak.J. Sci. Res. 1967, 19,140–145.
[21] A. Vidal-Ferran,A. Moyano,M. A. Pericꢃs,A. Riera, J. Org. Chem.
1997, 62,4970–4982.
Received: March 31,2008
Published online: June 20,2008
7698
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2008, 14,7687 – 7698