Making Fe(BPBP)-catalyzed C-H and CC oxidations more affordable

Article abstract of DOI:10.1039/c3ob42249f

The limited availability of catalytic reaction components may represent a major hurdle for the practical application of many catalytic procedures in organic synthesis. In this work, we demonstrate that the mixture of isomeric iron complexes [Fe(OTf)₂(mix-BPBP)] (mix-1), composed of Λ-α-[Fe(OTf)₂(S,S-BPBP)] (S,S-1), Δ-α- [Fe(OTf)₂(R,R-BPBP)] (R,R-1) and Δ/Λ-β-[Fe(OTf) ₂(R,S-BPBP)] (R,S-1), is a practical catalyst for the preparative oxidation of various aliphatic compounds including model hydrocarbons and optically pure natural products using hydrogen peroxide as an oxidant. Among the species present in mix-1, S,S-1 and R,R-1 are catalytically active, act independently and represent ca. 75% of mix-1. The remaining 25% of mix-1 is represented by mesomeric R,S-1 which nominally plays a spectator role in both C-H and C=C bond oxidation reactions. Overall, this mixture of iron complexes displays the same catalytic profile as its enantiopure components that have been previously used separately in sp³ C-H oxidations. In contrast to them, mix-1 is readily available on a multi-gram scale via two high yielding steps from crude dl/meso-2,2′-bipyrrolidine. Next to its use in C-H oxidation, mix-1 is active in chemospecific epoxidation reactions, which has allowed us to develop a practical catalytic protocol for the synthesis of epoxides.

Full text of DOI:10.1039/c3ob42249f

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
PAPER
Making Fe(BPBP)-catalyzed C–H and CvC
oxidations more affordable†
Cite this: Org. Biomol. Chem., 2014,
12, 2062
Vital A. Yazerski,a Peter Spannring,a David Gatineau,a Charlotte H. M. Woerde,a
Sara M. Wieclawska,a Martin Lutz,b Henk Kleijna and Robertus J. M. Klein Gebbink*a
The limited availability of catalytic reaction components may represent a major hurdle for the practical
application of many catalytic procedures in organic synthesis. In this work, we demonstrate that the
mixture of isomeric iron complexes [Fe(OTf)2(mix-BPBP)] (mix-1), composed of Λ-α-[Fe(OTf)2(S,S-BPBP)]
(S,S-1), Δ-α-[Fe(OTf)2(R,R-BPBP)] (R,R-1) and Δ/Λ-β-[Fe(OTf)2(R,S-BPBP)] (R,S-1), is a practical catalyst for
the preparative oxidation of various aliphatic compounds including model hydrocarbons and optically
pure natural products using hydrogen peroxide as an oxidant. Among the species present in mix-1, S,S-1
and R,R-1 are catalytically active, act independently and represent ca. 75% of mix-1. The remaining 25% of
mix-1 is represented by mesomeric R,S-1 which nominally plays a spectator role in both C–H and CvC
bond oxidation reactions. Overall, this mixture of iron complexes displays the same catalytic profile as its
enantiopure components that have been previously used separately in sp3 C–H oxidations. In contrast to
them, mix-1 is readily available on a multi-gram scale via two high yielding steps from crude DL/meso-
2,2’-bipyrrolidine. Next to its use in C–H oxidation, mix-1 is active in chemospecific epoxidation reactions,
which has allowed us to develop a practical catalytic protocol for the synthesis of epoxides.
Received 12th November 2013,
Accepted 9th January 2014
DOI: 10.1039/c3ob42249f
www.rsc.org/obc
Introduction
Selective oxidation of hydrocarbons is still a challenging task
in modern organic synthesis,1 while the oxidative modification
of functionalized molecules poses to chemists even more
complex selectivity issues.2 Only a limited number of intelli-
gently designed homogeneous catalysts, answering this selecti-
vity call, are available.3 One of them is the single site catalyst
Λ-α-[Fe(S,S-BPBP)(MeCN)2](SbF6)2 (S,S-1SbF6) reported by
Fig. 1 Structurally related tetradentate ligands applied in preparative
Fe-catalyzed C–H and CvC oxidation reactions with H2O2.
White et al. (BPBP
= N,N′-bis(2-picolyl)-2,2′-bipyrrolidine)
(Fig. 1, Scheme 1).3a,4 This catalyst was shown to be predictive
and outstandingly selective in challenging aliphatic C–H oxi-
dations of numerous and structurally diverse natural com-
pounds using hydrogen peroxide as a convenient oxidant.
Moreover, enantiopure R,R- and S,S-1SbF6 complexes are now
commercially available, but at a price per gram at least 6-fold
higher than for instance the Ru-based Grubbs I catalyst. This
clearly indicates that the value of the optically pure BPBP
ligand has a dominant share in the production costs of this
iron catalyst.
The S,S-1SbF6 complex was reported as a superior C–H oxi-
dation catalyst compared to its predecessor – racemic Δ/Λ-
α-[Fe(BPMEN)(MeCN)2](SbF6)2, 2SbF6 (BPMEN = N,N′-dimethyl-
N,N′-bis(2-picolyl)-ethylene-1,2-diamine, Fig. 1) based on an
achiral and readily accessible ligand. For instance, in the oxi-
dation of 3, both catalysts were comparably active (41 vs. 42%
substrate was converted using 2SbF6 and S,S-1SbF6, respecti-
vely), while the reaction chemoselectivity towards 4 signifi-
cantly varied (62% against 90% using 2SbF6 and S,S-1SbF6,
respectively) (Scheme 1).3a The formation of by-products as
observed with 2SbF6 was attributed to Fenton-type processes,
even though the obtained tertiary alcohol 4, in fact, was dia-
stereopure (dr > 99/1) with both catalysts. The reactivity of
aOrganic Chemistry & Catalysis, Debye Institute for Nanomaterials Science,
Utrecht University, Universiteitsweg 99, 3584 CG Utrecht, The Netherlands.
E-mail: r.j.m.kleingebbink@uu.nl; Fax: +31 30 2523615; Tel: +31 30 2531889
bCrystal and Structural Chemistry, Bijvoet Center for Biomolecular Research,
Utrecht University, Padualaan 8, 3584 CH Utrecht, The Netherlands
†Electronic supplementary information (ESI) available: Experimental procedures
and characterization data. CCDC 969747. For ESI and crystallographic data in
CIF or other electronic format see DOI: 10.1039/c3ob42249f
2062 | Org. Biomol. Chem., 2014, 12, 2062–2070
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Scheme 1 Comparison of the catalytic performances of 1SbF6 and
2SbF6 in C–H oxidation reactions with hydrogen peroxide.3a,4a,5
enantiopure S,S-1SbF6 catalyst was illustrated in a seminal
report about catalytic hydroxylations of complex molecules.3a
However, no examples of chiral induction from the catalyst to
the product were provided, as all optically pure alcohol
products retained the initial configuration of their chiral
precursors.3a
Subsequent reports indicated the importance of catalyst–
substrate topology matching in the case of hyper-functiona-
lized molecule oxidations with R,R- and S,S-1SbF6, where, e.g.
in the oxidation of 5 to diketone 6, the main product yield
varied from 22 to 45% depending on the employed Fe catalyst
enantiomer (Scheme 1).4a In contrast, the oxidation of
(−)-ambroxide 7 – another complex substrate – proceeded
smoothly with both R,R- and S,S-1SbF6, yielding 80 and 76% of
Scheme 2 Two-step preparation of mix-1 starting from crude DL/
meso-2,2’-bipyrrolidine.
stereoselective olefin epoxidation,9 and dihydroxylation reac-
tions.3f The use of triflate instead of antimony hexafluoride
counterions could further decrease the catalyst cost without
largely altering its activity and selectivity.6b
Results and discussion
(+)-sclareolide, respectively. In the synthesis of S-α-acetoxy iso- A DL/meso-2,2′-bipyrrolidine mixture (dr ca. 50/50) is available
caprolactone 10 from S-α-acetoxy isocaproic acid 9a, the use of on a multi-gram scale.7a,b In our study, this mixture was con-
S,S-1SbF6 was clearly beneficial compared to R,R-1SbF6. On the verted in two steps into the target catalyst mix-1 with an overall
other hand, the oxidation of the corresponding esterified sub- yield of 57–66%, containing rac-1 (equimolar mixture of S,S-
strate 9b led to the same lactone 10 and the yield did not and R,R-1, 65–75%) and Δ/Λ-β-[Fe(OTf)2(R,S-BPBP)] (R,S-1,
depend on the catalyst chirality (Scheme 1).5 More reports con- 25–35%) (Scheme 2; see ESI†).
firmed that only structurally enhanced analogues of 1SbF6 or
The X-ray crystal structure of enantiopure S,S-1 was pre-
1, in which the catalyst bite angle was affected, can feasibly viously reported.3f The S,S-BPBP ligand wraps around the Fe-
alter the substrate-imposed regio- and chemoselectivity in C–H center in a cis-α coordination topology, causing the exchange-
oxidations.6
able triflates to be chemically equivalent, positioned cis to
We envisioned that the most costly, tricky and low yielding each other and trans to the pyrrolidine nitrogen donors. The
step in the synthesis of enantiopure 1SbF6 catalysts, i.e. the R,S-BPBP ligand in the racemic R,S-1 complex, obtained in an
resolution of the crude DL/meso-2,2′-bipyrrolidine mixture,7,8 independent manner (see ESI†), adopts the alternative cis-β
might be avoided if the catalyst is used for certain C–H oxi- coordination topology in the solid state, where the labile
dation reactions that occur without serious topological restric- cis-sites, occupied by triflates, are not equivalent (Fig. 2). One
tions on the catalyst–substrate interaction and where catalyst- labile site is positioned in a position trans to a pyrrolidine
to-substrate stereo-induction is not involved. We therefore nitrogen donor atom, while the second one is located trans to
set out to evaluate the catalytic properties of a mixture of iron a pyridine nitrogen donor.
triflate complexes, [Fe(OTf)2(mix-BPBP)] (mix-1), derived from
Such a structural disparity, in principle, might raise certain
a mixture of the BPBP ligand diastereomers, i.e. using a non- selectivity issues if both catalyst structures are present in mix-
resolved mixture of BPBP ligands (Scheme 2). The ferrous 1.10 In addition, the enantiopure catalyst performance can be
triflate S,S-BPBP complex Λ-α-[Fe(OTf)2(S,S-BPBP)] (S,S-1) is affected by the presence of its antipode if (poly)nuclear
a part of it and has previously been reported to catalyze clusters are involved in catalysis or are responsible for catalyst
hydrocarbon oxidation processes, e.g. C–H oxidations,3b,6b deactivation.11
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 2062–2070 | 2063

View Article Online
Paper
Organic & Biomolecular Chemistry
Table 1 Alkane hydroxylation with 1 and 2/H2O2/AcOHa
Main products,
yieldb [%]
Substrate
Fe-cat.
Remarks
13/12
(K/A)c
8.9
9.0
2.4
—
11
S,S-1
mix-1
2
27/3.0
26/2.9
11/4.6
0.3/0.3
R,S-1
15/16/17
(3°/2°)d
21
21
16
—
14
18
S,S-1
mix-1
2
19/0.5/2.2
16/0.7/1.6
8.1/0.7/0.8
0.2/0.1/0.1
Fig. 2 Molecular structures of S,S-1 (left)3f and R,S-1 (right, Δ-isomer),
drawn with displacement ellipsoids at the 50% probability level. Of the
triflate groups, only the coordinated oxygen atoms are shown, and
hydrogen atoms are omitted for clarity.
R,S-1
( )-19/( )-20
43/0.3
41/0.2
(RC, %)e
>99
>99
S,S-1
mix-1
2
22/0.2
>99
R,S-1
0.1/0.2
—
a Reaction conditions: Fe-cat./H2O2/substrate/AcOH 1 : 120 : 100 : 50,
0
°C. b Determined by GC. c Ketone over alcohol product ratio,
K/A = [mol 13]/[mol 12]. d 3°/2° = 3 × [mol 15]/([mol 16] + [mol 17]).
e Retention of configuration, RC, % = 100 × [mol ( )-19]/([mol ( )-19] +
[mol ( )-20]).
Cyclohexanone (13) was formed as the main product in
26–27% substrate-based yield in the oxidation of 11, regardless
of the used catalyst. Adamantane oxidation provided the ter-
tiary alcohol 15 in 19 and 16% yield with both S,S-1 and mix-1
catalysts, respectively. The latter substrate is poorly soluble in
the reaction medium, which might make the catalyst activity
interpretation under preparative conditions within the stan-
dard reaction period (16 min) less accurate. E-1,2-Dimethyl-
cyclohexanol 19 was obtained from 18 in 43 and 41% yield
using S,S- and mix-1, respectively, in accordance with the
previous reports.3b,6b
Scheme 3 Catalytic oxidation of cyclohexane (11), adamantane (14)
and Z-1,2-dimethylcyclohexane (18). Reaction conditions: Fe-cat./H2O2/
substrate/AcOH x : 120 : 100 : 50, where x = 0.050–4.00, at 0 °C.
Remarkably, the product distribution obtained with S,S-1
and mix-1 in these oxidations was almost identical. For
instance, in the oxidation of cyclohexane (11) the molar ratio
between ketone 13 and alcohol 12 products (K/A) was found to
be 8.9–9.0 for both catalysts. Thus, the mixed catalyst can sub-
stitute enantiopure S,S-1 in certain transformations, where the
formation of carbonyl compounds is desired.
Another property of 1 is its ability to preferentially oxidize
tertiary in the presence of secondary C–H bonds, which is
illustrated by the 3°/2° ratio in adamantane (14) oxidation. The
corresponding value was 21 for the 3°/2° regioselectivity for
both catalysts, i.e. the different catalytic sites present in mix-1
act coherently also in this case. From a mechanistic point of
view, such a high 3°/2° value indicates that the rate limiting
step in alkane oxidation by 1/H2O2 is hydrogen abstraction
from a substrate molecule.3c,13
We have compared the easily accessible BPBP iron complex
mixture mix-1 and its separate components (S,S-, R,R- and R,S-
1) as catalysts for aliphatic C–H oxidations firstly using the set
of cyclohexane (11), adamantane (14), and Z-1,2-dimethylcyclo-
hexane (18) as benchmark substrates to reveal the relations
between the catalyst composition and its performance
(Scheme 3).
There are several slightly different catalytic protocols
reported in the literature, which involve Fe(BPBP) complexes
for alkane oxidation reactions. The initially developed protocol
utilizes 5–25 mol% of S,S-1SbF6.3a,4,5 The catalyst amount was
further optimized and lowered to 1–3 mol% for S,S-1 and
several structurally related Fe-complexes.3b,6b,12 Both reaction
protocols operate under formal substrate limiting (preparative)
conditions, and use hydrogen peroxide as an oxidant and
acetic acid as an additive. Following the latter protocol, we
found only very minor deviations in the catalyst activity, oxidiz-
ing the benchmark substrates in the presence of 1 mol% of
either S,S-1 or mix-1 (Table 1).
The regioselectivity of this step is firstly determined by the
difference of the homolytic dissociation energy of the different
C–H bonds in the substrate, while the steric accessibility of
these C–H bonds is less crucial. The formed transient tertiary
carboradicals nearly instantly get converted into the corres-
ponding alcohols, thereby retaining the configuration of their
2064 | Org. Biomol. Chem., 2014, 12, 2062–2070
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
that non-heme iron catalysts containing tetradentate poly-
amine ligands (TPA,16 BPMEN17 and S,S-BPBP3b,6b,15) can oxi-
datively dimerize in the presence of oxidants and acetic acid
forming binuclear oxo-bridged dimers [FeIII2L2(O)OAc]3+,
which represent either inert or resting states of the catalyst.
Investigating the catalytic reaction mixtures containing R,S- or
S,S-1 by ESI-MS analysis, we found a clear difference in the fate
of these catalysts: while ESI-MS analysis of S,S-1 at the end of
the oxidation of 13 (S,S-1/13/H2O2/AcOH 1 : 100 : 120 : 50)
showed the iron-containing ions [Fe2L2(O)OAc]3+, [Fe2L2(O)-
+
OAc](OTf)2+ and [Fe2L2(O)OAc](OTf)2 with respective m/z
values of 277.1005 (calcd 277.1033), 490.1337 (calcd 490.1369)
and 1129.2134 (calcd 1129.2139), no iron-containing ions were
Fig. 3 GC yield of cyclohexanone (13), 1-adamantanol (15) and ( )-E- identified in the reaction carried out with R,S-1. The ligand
recovery protocol described by White et al.15 was applied
to both reaction mixtures and allowed for the isolation of
1,2-dimethylcyclohexanol (( )-19) vs. Fe-catalyst 1 loading in catalytic
oxidations of corresponding substrates. Fe-cat./alkane/H2O2/AcOH
x : 100 : 120 : 50, where x = 0.050–4.0, at 0 °C.
unchanged S,S-BPBP in 79% yield (95% originally reported)
and the anticipated R,S-BPBP in 74% yield. Using the same
reaction and recovery protocol, the mix-BPBP ligand was recov-
precursor, which was confirmed for the oxidation of 18. In this
case, 19 was obtained as a single diastereomer (RC > 99%)
using either the S,S-1 or mix-1 catalyst. Overall, these experi-
ments show that mix-1 retains the most important intrinsic
C–H bond oxidation characteristics of its separate single site
congender S,S-1.
ered in 69% yield and an improved dr (from 75/25 to 85/15),
i.e. the share of rac-BPBP ligand slightly increased. Only the
rac-BPBP ligand (dr > 98/2, 58–63% yield) was recovered from
precipitates formed after diluting the product mixtures
obtained with mix-1/H2O2 using diethyl ether as an anti-
solvent (see ESI†). This experiment confirmed that rac-BPBP
remains bound to Fe after catalysis and it is insoluble in
diethyl ether in this state, while R,S-1 under the same con-
ditions dissociates liberating ether-soluble oxidized R,S-BPBP.
In addition, S,S-BPBP, being coordinated to iron, is more oxi-
datively robust than its diastereomer: no oxidized ligand was
identified in ESI-MS traces of catalytic oxidations utilizing
H2O2 in a larger excess (S,S-1/18/H2O2/AcOH 1 : 100 : 1000 : 50).
In contrast, only ions with m/z 337.2148 (calcd for [L + O − H]+
337.2023) were observed after catalysis with R,S-1 under these
conditions.
Based on the experiments with 1 mol% loading of mix-1
and S,S-1, a nearly equal catalytic behaviour of all components
(S,S-, R,R- and R,S-1) in the mixed catalyst might be assumed.
Yet, catalytic experiments carried out with pure R,S-1 demon-
strated its complete inertness (Table 1). It was found that
increasing the loading of mix-1 from 1 to 4 mol% had almost
no effect on the formation of the main product in the oxi-
dation of both 11 and 14 (Fig. 3). In the case of 18, the yield in
( )-19 rose slightly from 41 to 47% and did not improve after
more oxidant was added. Actually, the mixed catalyst loading
can be lowered to 0.75 mol% for the oxidation of 11 and 18
and even to 0.25 mol% in the case of 14 without substantially
affecting the product yield (Fig. 3). Yet, minor changes were
observed in the product distribution upon changing the cata-
lyst loading. In the oxidation of 11, the K/A ratio rose from 6.6
to 15 when increasing the loading of mix-1 from 0.75 to
4.0 mol%. The 3°/2° value in the oxidation of 14, on the other
hand, dropped from 26 to 20, using 0.25 and 4.0 mol% of mix-
1, respectively. This trend implies a certain overoxidation
of the respective minor product 12 and major product 15
at higher catalyst loadings. In the oxidation of 18, we found a
persistently high product diastereopurity with the mix-1 cata-
lyst within a broad range of catalyst loadings. When the cata-
lyst loading was reduced to 0.050 mol%, the tertiary alcohol
( )-19, formed in ca. 5% yield, was still diastereopure (dr 97/3).
Such a high retention of the configuration (RC 97%) at extre-
mely low catalyst concentrations rules out the participation of
free OH-radicals in alkane oxidations carried out with mix-1.
The observed relation between the catalyst activity and its
loading supports the idea of bi- or multimolecular pathways of
catalyst deactivation.3b,6b,14,15 Several reports have proposed
Next, we turned to functionalized substrates investigating
the performance of mix-1 (Scheme 4 and Table 2). For
instance, the oxidation of 2-adamantyl acetate (21) with
methyltrifluoromethyl dioxirane is known to provide a mixture
of two diastereomeric Z-22 and E-23 alcohols in ca. 2.5 : 1
Scheme 4 Catalytic oxidation of 2-adamantyl acetate (21) and L-(−)-
menthyl acetate (24) with 1 and hydrogen peroxide.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 2062–2070 | 2065

View Article Online
Paper
Organic & Biomolecular Chemistry
Table 2 Oxidation of functionalized substrates by 1a
Main products,
Substrate
Fe-cat.
yieldb [%]
Remarks
22 + 23/(rSM, %)c
34e/(35)
Z/Eb,d
2.6
2.0
21
S,S-1
R,R-1
mix-1
30e/(42)
35e/(38)
2.2
25/26/27
22/2.1/1.5
C1/C8f,g
10
10
10
24
S,S-1
R,R-1
mix-1
21/1.9/1.6
19/1.9/1.6 [16/1.4/1.1]e
a See Table 1. b See Table 1. c rSM – recovered starting material (21).
d Z/E = [mol 22]/[mol 23]. e Isolated yield. f C1/C8 = [mol 25]/[mol 26].
g Determined by 1H NMR.
Scheme 5 Catalytic oxidation of (−)-ambroxide 7 and (+)-sclareolide 8.
ratio.18 This ratio was found to be 2.6, 2.0 and 2.2 : 1 using the
S,S-, R,R- and mix-1 catalysts, respectively, while the total iso-
lated yield of these products did not depend on the catalyst
choice and was ca. 30–35% for each of these catalysts. The
unreacted substrate was recovered in 35, 42 and 38% after oxi-
dations with S,S-, R,R- and mix-1, respectively. The hydroxy-
lation of L-menthyl acetate 24 was also carried out with mix-1
and with its separate components, following the standard reac-
tion protocol. One major (25) and two minor (26 and 27) reac-
tion products were identified in this process (Scheme 4,
Table 2). The optically pure Fe catalysts, S,S-1 and R,R-1,
turned out to be slightly more active, yielding 21–22% of the
main product 25, while the mixed catalyst (1 mol%) afforded
19%. The isomeric alcohols 25 and 26 were formed in a 10 : 1
ratio irrespective of the employed catalyst, which is in agree-
ment with the distribution previously obtained with S,S-
1SbF6.3a,15 The observed regioselectivity is driven by the dis-
tinct steric accessibility of electronically similar tertiary C–H
bonds at the C1 and C8 carbons in 24 (Scheme 4).3a Interest-
ing, the reaction between 24 and ozone, a small and electro-
philic oxidant, demonstrated the inverted regioselectivity
providing mainly 26.19
recovered, rising the reaction mass balance and selectivity up
to 99%. When the oxidant was delivered by a syringe pump
to the reaction mixture, the reported reaction selectivity
towards lactone 8 was 57, 66 and 58% with R,R-1SbF6, R,R-1
and S,S-1, respectively (Fe-cat./7/H2O2/AcOH 3 : 100 : 260 : 150),
and no remaining 7 was observed.6b Under the latter con-
ditions and in our hands, the R,R-1, S,S-1 and mix-1
catalyst afforded 70, 66 and 67% of 8, respectively (GC yield),
while the substrate conversion was complete. Subsequently,
we set up a 1.5 g oxidation of 7 and isolated both products
of this reaction: the lactone (8, 54%) and the hydroxy acid
(28, 29%) using a casual “pH switching” separation technique
(see ESI†).
Further (+)-sclareolide oxidation leading to a mixture of
(+)-oxo-sclareolides (29, 30, 31) was shown to be sensitive
to the catalyst topology.4a In our hands, the C2–30 and C3–31
oxo-products formed equally in 33, 37 and 34% combined
yield using either S,S-, R,R- or mix-1 with 1/8/H2O2/AcOH
3 : 100 : 360 : 150 (Scheme 5, Table 3). Thus, no notable differ-
ence in the reaction regioselectivity was found. The R,R-1SbF6
catalyst at 3 mol% loading was reported to afford 43% of
an equimolar ketone mixture in this transformation.6b Also a
The plausible ketone product 27 has been initially over-
looked (tests with S,S-1SbF6)3a and was just recently detected
and characterized among the oxidation products of 24 (tests
with S,S-1, R,R-1 and S,S-1SbF6); however, the formation of 26
was not mentioned.6b A large-scale oxidation of 24 (10 g) using
cheap mix-1 allowed us to confirm that indeed both by-
products are formed in this reaction and that their amounts
are comparable (26 : 27 = 1.3 : 1 (isolated yields) or 1.1 : 1 (GC
yields)).
(−)-Ambroxide (7) is another substrate that attracted our
attention. Despite the numerous sites available for oxidation,
this terpenoid was shown to form the corresponding lactone
(+)-sclareolide (8) in a high isolated yield of 80% using R,R-
1SbF6 in an “iterative reaction protocol”, where the catalyst
and reagents are added to 7 in three equal portions with
the ultimate component ratio 1SbF6/7/H2O2/AcOH of
15 : 100 : 360 : 150 (Scheme 1 and 5).4a This oxidation is con-
sidered as regiospecific as 19% of the reactant (7) was
Table 3 Oxidation of functionalized substrates by 1a,b
Substrate
Fe-cat.
Main products, Yieldc [%]
8/28
TONe
29.3
31.1
29.8
7a
S,S-1
R,R-1
mix-1
66 (50)d/nd
70 (51)d/nd
67 (54)d/29d
29/30/31
8/17/16
5/20/17
7/18/16
TONe
18.2
18.7
18.2
8b
S,S-1
R,R-1
mix-1
Reaction conditions: Fe-cat./H2O2/substrate/AcOH. a 3 : 260 : 100 : 150,
RT. b 3 : 360 : 100 : 150. c Determined by GC. d Isolated yield. e TON =
[mol epoxide]/([mol S,S-1] + [mol R,R-1]). The amount of R,S-1 (inert
part) in mix-1 was determined to be ca. 25% and was not considered
calculating the TON values.
2066 | Org. Biomol. Chem., 2014, 12, 2062–2070
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Scheme 6 Catalytic oxidation of substituted alkenes with Fe catalysts
and hydrogen peroxide.
Fig. 4 Kinetic profiles of Z-cyclooctene (32a) oxidation with 1 catalysts
at 0 °C. Fe-cat./32a/H2O2/AcOH x : 100 : 150 : y, where x and y are given
in parentheses in the legend; y is omitted for the experiments carried
out without AcOH. The oxidant addition was complete at 60 min.
(up to 70% v/v) to maintain high levels of their catalytic activity
in alkene epoxidations and to suppress diol formation.16,22,23
In numbers, Δ/Λ-α-[Fe(rac-PYBP)(MeCN)2](ClO4)2 (0.10 mol%)
was reported to afford 32b in 10 and 72% yield in the presence
of 50 and 500 mol% of AcOH, respectively, approaching the
activity of mix-1 in the presence of only 0.5 mol% of AcOH.
The R,S-1 complex was inert in preparative epoxidations
just like in C–H oxidations, while the structurally related
β-[Fe(OTf)2(6-Me2-BPMCN)] catalyst was reported to be active
in olefin syn-dihydroxylation reactions under oxidant limiting
conditions.22b
Subsequent epoxidation of several substituted alkenes
using mix-1/alkene/H2O2/AcOH 0.13 : 100 : 150 : 0.5 resulted
in incomplete substrate conversions (Scheme 6, Table 4). This
issue for the oxidation of 34a–37a was solved by increasing the
amount of mix-1 and AcOH. 1-Decene (33a) was found to
be the least reactive substrate with hydrogen peroxide among
the tested unfunctionalized alkenes. Its complete conversion
into epoxide 33b required more elevated loadings of the mix-1
catalyst and acetic acid (0.52 and 2.0 mol%, respectively). The
epoxidation of this substrate turned out to be challenging
also for the reported Fe(rac-PYBP) catalyst, where 500 mol%
(5 equiv. to 33a) of AcOH and 0.50 mol% of Fe(rac-PYBP) were
necessary to drive the reaction to completion.21
Based on these observations, more electron-rich internal
alkenes are more readily oxidized by mix-1/H2O2 than terminal
ones. Subsequently, binary mixtures of 1-decene and model
alkene substrates were oxidized using the catalytic system mix-
1/H2O2/1-decene/alkene/AcOH 1 : 300 : 200 : 200 : 8 to probe
the regioselectivity of mix-1 in the oxidation of olefins. In this
experiment, the reaction progress was monitored along with
the gradual delivery of hydrogen peroxide into the system, i.e.
over 1 h. In all cases, we observed simultaneous oxidation
of both alkenes at every stage of H2O2 addition; however,
1-decene was always consumed at a lesser extent. The most
pronounced reactivity difference was found oxidizing
Z-cyclooctene and 1-decene (32a/33a), where the epoxide 32b/
33b ratio decreased from 8.6 to 5.8 and 4.5 : 1 at combined
substrate conversions of 11, 52 and 70%, respectively.
Table 4 Alkene epoxidation with 1/H2O2/AcOHa
Alkene Fe-cat. (mol%) AcOH, mol% Yieldb [%]/Conv.b [%] TONc
32a
S,S-1 (0.50)
R,S-1 (0.50)
mix-1 (0.50)
mix-1 (0.67)
S,S-1 (0.10)
mix-1 (0.13)
mix-1 (0.13)
mix-1 (0.13)
mix-1 (0.52)
mix-1 (0.26)
mix-1 (0.26)
mix-1 (0.26)
mix-1 (0.26)
0
97/99
<1/<3
97/99
95/99
26/26
24/25
92d/98
29/29
89d/95
92d/99
95d/99
89d/98
90d/99
194
—
0–50
0.050
0
0
0
0.50
0.50
2.0
1.0
1.0
1.0
1.5
194
190
260
240
920
290
223
460
475
445
450
33a
34a
35a
36a
37a
a Reaction conditions: alkene (1.0 mmol), H2O2 (1.5 mmol over 60 min),
AcOH, Fe-cat., MeCN (5 mL), 0 °C, 90 min. b Determined by GC. c TON
= [mol epoxide]/([mol S,S-1] + [mol R,R-1]). The amount of R,S-1 (inert
part) in mix-1 was determined to be ca. 25% and was not considered
calculating the TON values. d Isolated yield.
small amount of the C1–29 ketone (ca. 5–8%) was always
formed.
The synthetic accessibility of the mix-1 catalyst is compar-
able to that of the structurally related Δ/Λ-α-[Fe(OTf)2-
(BPMEN)] 2 20 derived from an achiral ligand and to recently
21
reported
Δ/Λ-α-[Fe(rac-PYBP)(MeCN)2](ClO4)2
catalysts,
which are among the most easily synthesized and highly
active non-heme Fe-catalysts for alkene epoxidations (Fig. 1).
Optimizing the reaction conditions for mix-1 it was found
that Z-cyclooctene (32a) is smoothly converted into its
epoxide 32b within 90 min using mix-1/32a/H2O2/AcOH
(0.13–0.67) : 100 : 150 : (0–0.50) (Fig. 4, Table 4). No AcOH is
required to complete the oxidation using 0.67 mol% of mix-1
and just 0.5 mol% AcOH turned out to be sufficient to reach
full conversion of 32a with 0.13 mol% of the mixed catalyst. In
the latter case, the oxidant was introduced over either 60 or
6 min, showing very little difference in product yield (92 and
95%, respectively).
The oxidation of E-4-octene and 1-decene (34a/33a) turned
out to be the least regioselective with the 34b/33b product
distribution changed from 3.1 to 2.1 and 2.0 : 1 at combined
substrate conversions of 13, 57 and 69%. As a measure of
Remarkably, the reported non-heme iron complexes of
BPMEN and rac-PYBP require substantial amounts of AcOH
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 2062–2070 | 2067

View Article Online
Paper
Organic & Biomolecular Chemistry
regioselectivity, here we consider the approximate ratio of natural diene ( )-limonene 39a bearing both highly reactive
an epoxy alkene and 33b at 50% combined conversion of geminal di- and endocyclic tri-substituted CvC bonds was
both competing alkene substrates. Thus, the following alkene converted into a 1 : 3.4 mixture (54% yield) of 8,9-exo-(exo-39b)
reactivity order depending on the CvC bond substitution and 1,2-endo-cyclic (endo-39b) monoepoxides using mix-1/39a/
pattern can be drawn: mono-<E-di<Z-di≤gem-di<tri-<<cyclic H2O2/AcOH 1 : 1000 : 750 : 5. Much better regioselectivity was
1.0 : 2.2 : 3.1 : 3.3 : 3.5 : 5.8, based on the oxidation of 33a, 34a, achieved oxidizing 4-vinyl-cyclohexene (40a) (a diene with iso-
36a, 35a, 37a and 32a, respectively.
lated Z-disubstituted and terminal CvC bonds), where the
The reactivity trend is applicable to the prediction of endocyclic monoepoxide endo-40b was mainly formed (>15 : 1).
the product distribution in random couples of alkenes. For In addition, the endo-monoepoxide can be obtained in a good
instance, the equimolar mixture of Z-(35a) and E-4-octene yield of 79% using almost stoichiometric quantity of
(34a) tested under the same conditions yielded the corres- an oxidant (0.90 equiv.).
ponding epoxides 35b and 34b in a 1.5 Z/E ratio at 48% total
An excellent regioselectivity of the mix-1/H2O2 system was
conversion of the substrates, which is in agreement with the observed oxidizing S-(+)-carvone (41a), where 8,9-epoxycarvone
determined reactivity order. The geminal disubstituted alkene 41b was the only regioisomer formed (94% isolate yield,
(2-methylundecene-2, 36a) and Z-35a turned out to be compar- Scheme 8). The conjugated electron-poor CvC bonds in 41a
ably active being oxidized as a mixture, producing 1.0–1.1 36b and 41b remain untouched by the oxidant as long as some
epoxide for every Z-35b epoxide molecule.
starting material (41a) remains in solution. This allows the
Based on these data it was concluded that the mix-1 catalyst substrate conversion to go to completion without compromis-
is unable to efficiently distinguish disubstituted alkenyl ing on the reaction selectivity.
moieties under the applied epoxidation conditions. Indeed,
A brief study on the functional group tolerance was carried
the epoxidation of E,E,Z-1,5,9-cyclododecatriene (38a) with 0.6 out showing that the utilized catalytic system can be success-
equiv. of H2O2 yielded 51% (85% monoepoxide selectivity) of fully applied to the epoxidation of unsaturated primary alco-
the 1 : 2 mixture of the corresponding epoxy E,E-(symm-38b) hols and carboxylic acids (Scheme 8). For instance, the
and E,Z-5,9-cyclododecadiene (asymm-38b) (Scheme 7).
epoxidation of S-citronellol (42a) with mix-1/H2O2 proceeded
A
better regioselectivity is expected in the oxidation smoothly leading to quantitative epoxide 42b formation.
of dienes containing endo- and exo-cyclic CvC bonds. The ( )-Citronellyl acetate 43a was converted into the epoxide 43b
without any sign of the acetate group hydrolysis or 3° C–H
hydroxylation processes.
Potentially chelating cis-1,2,3,6-tetrahydrophthalic acid
(44a) was readily oxidized into a 1 : 1.2 mixture of the corres-
ponding moderately stable epoxides using mix-1 (0.26 mol%)
and H2O2 (1.3 equiv.).
Scheme 7 Catalytic oxidation of dienes with mix-1 on a preparative
scale. The isomeric monoepoxides were efficiently separated from by-
products and unreacted dienes via vacuum distillation (38b, 39b) and
column chromatography (40b). Substrate based yields are indicated in
all cases. The isomer ratio was determined by 1H-NMR for 38b and by
GC for the epoxides 39b and 40b.
Scheme 8 Catalytic epoxidation of functionalized molecules with mix-
1. The isomer ratio was determined by GC for 41b and by 1H-NMR for
44b.
2068 | Org. Biomol. Chem., 2014, 12, 2062–2070
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
ITN 215193). The National Research School Combination
Catalysis (NRSCC) is acknowledged for further financial
Conclusions
A rapidly growing number of publications discussing the appli- support. The X-ray diffractometer has been financed by the
cation of the promising small molecular catalyst Fe(BPBP) in Netherlands Organization for Scientific Research (NWO).
various C–H and CvC bond oxidation processes is an evident
sign of its appreciation by the scientific community. Despite
the efforts at making this catalyst more efficient in several
challenging C–H oxidations, its recently developed and more
Notes and references
advanced alternatives are comparable or less accessible from a
synthetic point of view and their showcase of preparative oxi-
dations usually affords submillimolar product quantities. In
contrast, the mix-1 catalyst (rac-/R,S-1 ca. 75/25) presented here
can be prepared on a multi-gram scale in two steps from a
crude 2,2′-bipyrrolidine mixture. The R,S-isomer component
plays only a spectator role in both C–H and CvC bond oxi-
dation reactions. The presence of this inactive catalyst isomer
does not affect the overall reactivity and selectivity of the mix-1
catalyst system. Most strikingly, the mix-1 catalyst retained
the most crucial selectivity properties of its optically pure
components, e.g. a high ketone over alcohol ratio in methylene
oxidations and high selectivity towards tertiary alcohols
along with complete retention of the initial tertiary carbon
center configuration in C–H oxidations. The mix-1 catalyst
is also applicable to the functionalization of more complex
substrates, e.g. the transformation of (−)-ambroxide into
(+)-sclareolide proceeds in a very similar manner with mix-1 as
with itself enantiopure S,S-, and R,R-congenders. The mix-1
catalyst is even useful in a challenging synthesis of oxo-sclareo-
lides, in which the product distribution depends on the chiral
catalyst shape. Importantly, the C–H oxidations mediated
by this simple iron complex in combination with hydrogen
peroxide have been proven to be scalable without compromis-
ing on the reaction selectivity.
1 (a) K. Chen and P. S. Baran, Nature, 2009, 459, 824;
(b) A. Mendoza, Y. Ishihara and P. S. Baran, Nat. Chem.,
2012, 4, 21.
2 A. Company, L. Gómez and M. Costas, in Iron-Containing
Enzymes, Versatile Catalysts of Hydroxylation Reactions
in Nature, ed. S. P. De Visser and D. Kumar, RSC,
Cambridge, 2011.
3 C–H bond oxidation: (a) M. S. Chen and M. C. White,
Science, 2007, 318, 783; (b) L. Gómez, I. Garcia-Bosh,
A. Company, J. Benet-Buchholz, A. Polo, X. Sala, X. Ribas
and M. Costas, Angew. Chem., Int. Ed., 2009, 48, 5720;
(c) Y. Hitomi, K. Arakawa, T. Funabiki and M. Kodera,
Angew. Chem., Int. Ed., 2012, 51, 3448; (d) R. V. Ottenbacher,
D. G. Samsonenko, E. P. Talsi and K. P. Bryliakov,
Org. Lett.,
2012, 14, 4310;
syn-dihydroxylation:
(e) E. N. Jacobsen, I. Marko, W. S. Mungall, G. Schroeder
and K. B. Sharpless, J. Am. Chem. Soc., 1988, 110, 1968;
(f) K. Suzuki, P. D. Oldenburg and L. Que Jr., Angew. Chem.,
Int.
Ed.,
2008,
47,
1887;
CvC
epoxidation:
(g) E. N. Jacobsen, W. Zhang, A. R. Muci, J. R. Ecker and
L. Deng, J. Am. Chem. Soc., 1991, 113, 7063;
(h) F. G. Gelalcha, B. Bitterlich, G. Anilkumar, M. K. Tse
and M. Beller, Angew. Chem., Int. Ed., 2007, 46, 7293;
(i) B. Wang, C. Miao, S. Wang, C. Xia and W. Sun, Chem.–
Eur. J., 2012, 18, 6750.
Moreover, the mix-1/H2O2 system is very promising in pre-
parative epoxidation reactions, where the amount of acetic
acid – a commonly used additive (or even a co-solvent) – can
be remarkably reduced to less than 2 mol% or even totally
eliminated. Under these conditions, a high catalyst activity
(TON approaching 1000) is obtained in the oxidation of elec-
tron-rich olefins, where the epoxides are the only isolated
products. A highly selective monoepoxidation of dienes is only
possible if the substrate molecule bears electronically different
CvC fragments. This mix-1/H2O2 epoxidation system tolerates
functional groups like primary alcohols and dicarboxylic acids.
Overall, we have shown the practical use of an easy
to prepare iron complex in a range of challenging organic
reactions, which now become much more affordable.
We believe that our practical observations on the Fe(BPBP)
complex reactivity will further contribute to the popularization
of this promising catalyst and will further broaden its appli-
cation in organic synthesis.
4 (a) M. S. Chen and M. C. White, Science, 2010, 327, 566;
(b) M. A. Bigi, S. A. Reed and M. C. White, Nat. Chem.,
2011, 3, 216.
5 M. A. Bigi, S. A. Reed and M. C. White, J. Am. Chem. Soc.,
2012, 134, 9721.
6 (a) P. E. Gormisky and M. C. White, J. Am. Chem. Soc., 2013,
135, 14052; (b) L. Gómez, M. Canta, D. Font, I. Prat,
X. Ribas and M. Costas, J. Org. Chem., 2013, 78, 1421.
7 Large scale synthesis and resolution protocol: (a) T. Oishi,
M. Hirama, L. R. Sita and S. Masamune, Synthesis, 1991,
789; (b) S. E. Denmark, J. Fu and M. J. Lawler, Org. Synth.,
2006, 83, 121; enantioselective synthesis of 2,2′-bipyrrol-
idines:
(c) A. Alexakis, A. Tomassini, C. Chouillet,
S. Roland, P. Mangeney and G. Bernardinelli, Angew.
Chem., Int. Ed., 2000, 39, 4093; (d) X.-N. Song and Z.-J. Yao,
Tetrahedron, 2010, 66, 2589.
8 Recent advances on diastereoselective synthesis of BPBP
and its derivatives: C. H. Müller, R. Fröhlich, C. G. Daniliuc
and U. Hennecke, Org. Lett., 2012, 14, 5944.
9 (a) O. Y. Lyakin, R. V. Ottenbacher, K. P. Bryliakov and
E. P. Talsi, ACS Catal., 2012, 2, 1196; (b) O. Cussó, I. Garcia-
Bosch, X. Ribas, J. Lloret-Fillol and M. Costas, J. Am. Chem.
Soc., 2013, 135, 14871.
Acknowledgements
This work was supported by 7PCRD EU funds from the Marie-
Curie Initial Training Network NANO-HOST (grant agreement
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 2062–2070 | 2069

View Article Online
Paper
Organic & Biomolecular Chemistry
10 (a) L. A. Evans, N. S. Hodnett and G. C. Lloyd-Jones, Angew. 17 M. C. White, A. G. Doyle and E. N. Jacobsen, J. Am. Chem.
Chem., Int. Ed., 2012, 51, 1526; (b) P. D. Oldenburg and Soc., 2001, 123, 7194.
L. Que Jr., Catal. Today, 2006, 117, 15; distinct reactivity 18 (a) M. E. González-Núñez, J. Royo, G. Castellano,
of racemic and meso-catalysts is often encountered in
olefin polymerization with ansa-metallocenes: (c) N. Naga,
T. Shiono and T. Ikeda, Macromol. Chem. Phys., 1999, 200,
1587; the use of racemic catalysts in synthesis is generally
C. Andreu, C. Boix, R. Mello and G. Asensio, Org. Lett.,
2000, 2, 831; (b) R. Mello, J. Royo, C. Andreu, M. Báguena-
Añó, G. Asensio and M. E. González-Núñez, Eur. J. Org.
Chem., 2008, 455.
associated with their preliminary enantiomer selective pre- 19 Y. Asakawa, R. Matsuda, M. Tori and T. Hashimoto, Phyto-
activation or poisoning: (d) K. Mikami, T. Korenaga, chemistry, 1988, 27, 3861.
Y. Matsumoto, Ma. Ueki, M. Terada and S. Matsukawa, Acc. 20 (a) M. Fujita and L. Que Jr., Adv. Synth. Catal., 2004, 346,
Chem. Res., 2000, 33, 39; (e) J. W. Faller, A. R. Lavoie and
B. J. Grimmond, Organometallics, 2002, 21, 1662.
190; (b) S. Taktak, S. V. Kryatov, T. E. Haas and E. V. Rybak-
Akimova, J. Mol. Catal. A: Chem., 2006, 259, 24;
11 Nonlinear effects in catalysis: T. Satyanarayana, S. Abraham
and H. B. Kagan, Angew. Chem., Int. Ed., 2009, 48, 456–494.
12 (a) I. Garcia-Bosch, Z. Codolà, I. Prat, X. Ribas, J. Lloret-Fillol and
M. Costas, Chem.–Eur. J., 2012, 18, 13269; (b) I. Prat, L. Gómez,
M. Canta, X. Ribas and M. Costas, Chem.–Eur. J., 2013, 19, 1908.
(c)
F.
D.
A.
Clemente-Tejeda,
Bermejo, Tetrahedron,
A.
López-Moreno
2013, 69,
and
2977;
(d) P. Spannring, V. Yazerski, P. C. A. Bruijnincx,
B. M. Weckhuysen and R. J. M. Klein Gebbink, Chem.–Eur.
J., 2013, 19, 15012.
13 D. H. R. Barton and D. Doller, in Reason and Imagination 21 E. A. Mikhalyova, O. V. Makhlynets, T. D. Palluccio,
Reflections on Research in Inorganic Chemistry, ed.
D. H. R. Barton, World Scientific, 1996.
A. S. Filatov and E. V. Rybak-Akimova, Chem. Commun.,
2012, 48, 687.
14 L. Gómez, Bioinspired iron and manganese catalysts for the 22 (a) K. Chen, M. Costas, J. Kim, A. K. Tipton and L. Que Jr.,
effective and selective oxidation of alkanes and alkenes,
Doctoral dissertation, University of Girona, 2010.
15 N. A. Vermeulen, M. S. Chen and M. C. White, Tetrahedron,
2009, 65, 3078.
J. Am. Chem. Soc., 2002, 124, 3026; (b) M. Costas and L. Que
Jr., Angew. Chem., Int. Ed., 2002, 41, 2179; (c) M. Costas,
A. K. Tipton, K. Chen, D.-H. Jo and L. Que Jr., J. Am. Chem.
Soc., 2001, 123, 6722.
16 J. Y. Ryu, J. Kim, M. Costas, K. Chen, W. Nam and L. Que 23 M. Fujita and L. Que Jr., Adv. Synth. Catal., 2004, 346,
Jr., Chem. Commun., 2002, 1288.
190.
2070 | Org. Biomol. Chem., 2014, 12, 2062–2070
This journal is © The Royal Society of Chemistry 2014