Hydrocarbon chlorination promoted by manganese and iron complexes with methylated derivatives of bis(2-pyridylmethyl)-1,2-ethanediamine

Article abstract of DOI:10.1016/j.molcata.2010.11.006

Non-heme iron halogenases, such as SyrB2 and CytC3, catalyze the regioselective chlorination and bromination of aliphatic C-H bonds. Reported here is the hydrocarbon chlorination promoted by manganese and iron complexes with methylated derivatives of bis(2-pyridylmethyl)-1,2-ethanediamine (bispicen). The reactions between these coordination compounds and meta-chloroperbenzoic acid generate oxidants capable of oxidizing weak C-H bonds to C-Cl bonds. This chemistry is regioselective, with a strong preference for activating C-H bonds on secondary carbons over weaker C-H bonds on tertiary carbons. The reactivity is consistent with the methyl groups on the ligands preventing more sterically encumbered substrates from accessing the reactive portions of a [M^IV(L^Men)(O)Cl₂] oxidant. The iron compounds promote more hydrocarbon chlorination than their manganese analogs.

Full text of DOI:10.1016/j.molcata.2010.11.006

Journal of Molecular Catalysis A: Chemical 335 (2011) 24–30
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Hydrocarbon chlorination promoted by manganese and iron complexes with
methylated derivatives of bis(2-pyridylmethyl)-1,2-ethanediamine
∗
Christian R. Goldsmith , Cristina M. Coates, Kenton Hagan, Casey A. Mitchell
Department of Chemistry & Biochemistry, Auburn University, Auburn, AL 36849, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Non-heme iron halogenases, such as SyrB2 and CytC3, catalyze the regioselective chlorination and bromi-
nation of aliphatic C–H bonds. Reported here is the hydrocarbon chlorination promoted by manganese
and iron complexes with methylated derivatives of bis(2-pyridylmethyl)-1,2-ethanediamine (bispicen).
The reactions between these coordination compounds and meta-chloroperbenzoic acid generate oxidants
capable of oxidizing weak C–H bonds to C–Cl bonds. This chemistry is regioselective, with a strong pref-
erence for activating C–H bonds on secondary carbons over weaker C–H bonds on tertiary carbons. The
reactivity is consistent with the methyl groups on the ligands preventing more sterically encumbered sub-
Received 2 September 2010
Received in revised form 4 November 2010
Accepted 5 November 2010
Available online 13 November 2010
Keywords:
Chlorination
Regioselectivity
C–H activation
IV Me
strates from accessing the reactive portions of a [M (L n)(O)Cl2] oxidant. The iron compounds promote
more hydrocarbon chlorination than their manganese analogs.
©
2010 Elsevier B.V. All rights reserved.
1
. Introduction
also been observed [16,17]. Alternative halogenation systems have
also displayed a similar lack of chemoselectivity [18–20].
The conversion of chemically inert C–H bonds to more use-
Halogenase enzymes, conversely, chemo- and regioselectively
activate C–H bonds without over-oxidation [21–24]. The active oxi-
dants in the enzymatic cycles are more geometrically complex than
halogen radicals, which enables them both to differentiate their
substrate from other compounds and to distinguish inequivalent
C–H bonds within the same molecule. One class of halogenases con-
tain mononuclear non-heme iron metal centers in their active sites
[21,25–27]. The non-heme iron halogenases are notable for using
O2 as a terminal oxidant and being capable of oxidizing aliphatic
ful functional groups remains a significant challenge in synthetic
chemistry [1–8]. A large amount of effort has been devoted to find-
ing conditions and reagents that (1) halt the oxidation at a product
that can potentially be further oxidized and (2) activate C–H bonds
within a substrate molecule selectively.
The conversion of aliphatic C–H bonds to C–Cl or C–Br bonds
commonly proceeds through radical processes [9]. The best known
method, free radical halogenation, relies on halogen atom radi-
cals to oxidize the hydrocarbon substrate. These oxidants are small
and highly symmetric, and they thereby lack the means to enable
either regio or stereosymmetric C–H activation at levels suitable for
preparative synthetic chemistry [10–12]. Due to the inherent lack of
selectivity and the high reactivity of chlorine and bromine radicals,
free radical halogenations often yield polyhalogenated compounds
as byproducts unless the substrate is present in excess [13,14]. Iso-
lating desired monohalogenated products from these reactions can
thereby be labor-intensive. An additional problem is that halogen
radicals are capable of multiple modes of reactivity. With allylic
substrates, for instance, the radical oxidants can either halogenate
the C–H bonds ˛ to the olefin or add across the olefin to yield dihalo-
genated alkanes [9,15]. Halogenation of aromatic C–H bonds has
C–H bonds. Haloperoxidases, in contrast, use H O2 as a terminal
2
oxidant and typically react with electronically activated C–H bonds
[21,24].
The non-heme iron halogenase SyrB2 has been structurally
characterized in its Fe(II) state (Scheme 1) [27]. The coordination
sphere around the iron ion in the active site includes two histidine
residues, a water molecule, a halide ligand, and ␣-ketoglutarate
(␣KG), which serves as a sacrificial reductant in the enzymatic
cycle. The oxidant that reacts with the hydrocarbon substrate in
the enzymatic cycle is believed to be a ferryl species, Fe(IV)(Cl)(O),
which abstracts a hydrogen atom from the substrate to yield an
organic radical and Fe(III)(Cl)(OH) [25,27,28]. A formal chlorine
atom transfer from the ferric species to the organic radical yields the
organochloride and Fe(II). The ferryl oxidants have recently been
characterized by Mössbauer and X-ray absorption spectroscopies
[
28,29] but the ferric species have not been observed.
Reported in this manuscript are the chlorination activities of
∗ Corresponding author at: Department of Chemistry & Biochemistry, 179 Chem-
istry Building, Auburn University, Auburn, AL 36849, United States.
Tel.: +1 334 844 6463.
two series of metal complexes with bis(2-pyridylmethyl)-1,2-
ethanediamine (L, bispicen) and five methylated derivatives (L n,
Me
E-mail address: crgoldsmith@auburn.edu (C.R. Goldsmith).
1
381-1169/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2010.11.006

C.R. Goldsmith et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 24–30
25
ether (ether) was purchased from Fisher Scientific, degassed,
and stored over 4 A˚ molecular sieves. Cyclohexene and chloro-
cyclohexane were purchased from Sigma–Aldrich and stored in
a glovebox free of moisture and oxygen. Toluene, cumene, and
ethylbenzene were bought from Sigma–Aldrich and distilled
prior to use. Meta-chloroperbenzoic acid (MCPBA) was acquired
from Sigma–Aldrich and stored in a refrigerator until needed.
Chloroform-d and acetonitrile-d were purchased from Cambridge
3
Isotopes and used without further purification. The syntheses
of N,N-bis(2-pyridylmethyl)-1,2-ethanediamine (bispicen, L)
ꢀ ꢀ
37], N,N -dimethyl-N,N -bis(2-pyridylmethyl)-1,2-ethanediamine
[
(
Me
ꢀ
2) [38], N,N -bis(2-pyridylmethyl)imidazolidine [39], N-
L
Scheme 1. Active site of SyrB2.
ꢀ
(LMe1)
methyl-N,N -bis(2-pyridylmethyl)-1,2-ethanediamine
ꢀ
[
39,40], N,N -bis(6-methyl-2-pyridylmethyl)-1,2-ethanediamine
Me
ꢀ
(L
2 )
[41]
and
cis-(N,N-bis(2-pyridylmethyl)-1,2-
Scheme 2). The N-donor ligands are meant to approximate the
endogenous coordination provided by the SyrB2 enzyme [27] and
are also electronically similar to ligands that have previously sup-
ported analogous hydrocarbon hydroxylation [30,31]. The bispicen
framework, while certainly an imperfect reproduction of the enzy-
matic coordination sphere, is attractive due to the ease of its
synthesis and modification. The methyl groups installed on the
pyridines and secondary amines allow us to assess the impact of
ligand sterics on the oxidative reactivity of the associated metal
complexes.
ethanediamine)-dichloromanganese(II) ([Mn(L)Cl ]) [42] have
2
been reported previously by others. We are detailing the syntheses
of the other ligands and metal complexes in another manuscript
that focuses on the conformational flexibility and dynamics of the
bispicen framework [43].
2.2. Instrumentation
One series contains iron(II) ([Fe(LMen)Cl ]); whereas, the other
2
All 1H and 13C nuclear magnetic resonance (NMR) spectra were
Me
contains manganese(II) ([Mn(L n)Cl ]). Manganese has previ-
2
recorded on a 400 MHz AV Bruker NMR spectrometer at 294 K;
internal standards were used to reference all NMR resonances. Gas
chromatography (GC) was performed on either a ThermoScientific
Trace GC Ultra or a Hewlett Packard 5890 gas chromatograph with
a flame ionization detector (FID). Tanden GC/mass spectrometry
ously substituted for iron in hydrocarbon oxidations with equal or
superior activity [32–36], but has not been systematically investi-
gated for its ability to chlorinate hydrocarbon C–H bonds. Reaction
of the M(II) compounds with meta-chloroperbenzoic acid (MCPBA)
generates oxidants that are capable of chlorinating hydrocarbon
substrates with weak C–H bonds.
(GC–MS) was performed on either a Hewlett Packard 5890 gas
chromatograph with a Fissons Instruments electrospray mass spec-
trometry detector or an Agilent 6890N gas chromatograph with
a Trio-2000 mass spectrometer. Optical spectroscopy data were
acquired on a Cary 50 spectrophotometer.
2
. Experimental
2.1. Materials
Acetonitrile
(MeCN),
iron(II)
chloride
tetrahydrate
•
(
FeCl2 4H O), and manganese(II) chloride (MnCl ) were bought
2
2
2.3. Mass spectrometry
from Sigma–Aldrich and used as received. Anhydrous diethyl
Samples were analyzed using an Ultra Performance LC Sys-
tems (ACQUITY, Waters Corp., Milford, MA, USA) coupled with
a quadrupole time-of-flight mass spectrometer (Q-TOF Premier,
Waters) with electrospray ionization (ESI) in positive ESI-MS and
ESI-MS/MS modes operated by the Masslynx software (V4.1). The
ion source voltages were set at 3 kV, the sampling cone was set at
3
7 V and the extraction cone was at 3 V. The source and desolvation
◦
temperature was maintained at both 80 C with the desolvation gas
flow at 200 L/h.
Samples containing the metal oxidant generated from
Me
[
Fe(L 2)Cl ] were directly injected into the ESI source at a
2
flow rate of 100 ␮L/min with a 50% acetonitrile, 0.1% formic
◦
acid mobile phase. The sample spent 6 s at 12 C before it was
ionized into gas phase. The TOF MS scanning was from 350
to 450 m/z at 1 s with 0.1 s inter-scan delay using extended
dynamic range acquisition with centriod data format. For real
time mass calibration, direct infusion of sodium formate solution
(
10% formic acid/0.1 M NaOH/aceonitrile) at a ratio of 1:1:8 at
s/10 s to ion source at 2 ␮L/min was used for a single point mass
calibration.
Ions of interest were analyzed for their elemental composition
1
using accurate mass (less than 5 ppm error) and isotope model-
ing to identify the formula. Collision-induced dissociation (CID)
by argon on precursor ions resulted in structural fragments that
further assisted the identification.
Scheme 2.

2
6
C.R. Goldsmith et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 24–30
2.4. Reactivity studies
◦
All oxidation reactions were run in MeCN at 23 C under N .
2
For the reactions with cyclohexene, 10 mM of the metal complex,
0 mM of MCPBA, and 400 mM of cyclohexene were mixed at once.
1
After stirring for 30 min, the solutions were filtered through a plug
of silica gel to remove the metal complex and analyzed through
GC. Products were identified both through GC–MS and by compar-
ison of their retention times with those of commercially available
standards, such as cyclohexene oxide and 3-chlorocyclohexene.
Yields were calculated through comparison to an internal stan-
dard, chlorocyclohexane, which was found to be unreactive under
these conditions. Parallel reactions in MeCN-d3 corroborated the
identities and ratios of the products.
The reactivity assays with the benzylic substrates proceeded in
an analogous fashion. In the first set of experiments, the starting
concentrations of MCPBA, the metal complex, and the benzylic sub-
strate were set at 10 mM. In the second set of reactivity assays, the
starting concentrations of MCPBA and hydrocarbon were 100 mM
and 400 mM, respectively. The reaction times with the benzylic
substrates were extended to 60 min. All reactions were repeated
at least three times to ensure reproducibility; all reported product
yields are theaverages of the results of these independent reactions.
Scheme 3.
present. A recent report from Fukuzumi and Nam’s groups found
that cyclohexene could initiate the generation of a ferryl oxo oxi-
dant from Fe(II) species and O2 [46]. A similar reaction involving
the [M(L n)Cl ] complexes and cyclohexene could explain the
additional oxygenated products in Table S1, given that the reaction
workup was done under air.
Despite these two complications, three observations are noted
from the cyclohexene data. First, 1,2-dichlorocyclohexane is never
found as an oxidation product. Second, iron consistently promotes
chlorination to a greater extent than manganese, at parity of ligand.
Third, the bispicen derivative with methyl groups installed on the
two amines, L 2, leads to the most chlorinated product for both
iron and manganese, with further methylation reducing the yield
of 3-chlorocyclohexene.
Me
2
Me
3
. Results
3.3. Reactivity of benzylic substrates
The oxidative reactivities of the manganese and iron complexes
were investigated, with a focus on their abilities to promote the
chlorination of hydrocarbon substrates. The M(II) complexes are
unstable in the presence of excess chloride, as assessed by optical
spectroscopy. In order to simplify the analysis, the present stud-
ies are therefore limited to stoichiometric chlorination, with no
additional chloride ion present beyond the two equivalents initially
associated with the metal complex.
The chlorination chemistry was subsequently investigated with
three substrates containing benzylic C–H bonds: toluene, ethyl-
benzene, and cumene (Scheme 3). In these substrates, the C–H
bonds most thermodynamically susceptible to oxidative attack
are on primary, secondary, and tertiary carbons, respectively. The
different steric environments around the benzylic C–H bonds pro-
vide a means of assessing the regioselectivity of the chlorination
reactions. The bond dissociation energies of the benzylic C–H
−
1
−1
3.1. Selection of oxidant
bonds are 88 (± 1) kcal mol
for toluene [47], 85 (± 1) kcal mol
−
1
for ethylbenzene, and 83 (± 1) kcal mol for cumene [48]. Based
on these bond dissociation energies, cumene would be antici-
pated to be the most active substrate. Unlike cyclohexene, these
substrates do not react extensively with MCBPA. In the control
experiments, ethylbenzene, and cumene react with MCPBA to form
1-hydroxyethylbenzene and 1-hydroxycumene in less than 10%
yield (Table 1).
Meta-chloroperbenzoic acid (MCPBA) was selected as the termi-
nal oxidant. Hydrogen peroxide, tert-butylhydroperoxide, oxone,
and iodosobenzene were also investigated but were not found to be
competent terminal oxidants for hydrocarbon chlorination by these
metal complexes. Although commercially available peracetic acid
was found to promote hydrocarbon oxidation to a greater extent
than MCPBA, the reactivity was found to be non-discriminate, and
the organic products included many compounds that are not read-
ily identifiable. The reactivity assays with MCPBA, conversely, led
to fewer organic products, all of which were easily identified. In the
The chemistries of the best chlorinating agents in the cyclo-
Me
hexene studies, [Fe(L 2)Cl2] and FeCl2 were compared to that of
Me
the bulkiest iron complex, [Fe(L 4)Cl2]. The reactivities of MnCl2,
Me
Me
[Mn(L 2)Cl2], and [Mn(L 4)Cl2] were also further explored in
order to corroborate the second observation noted in the cyclohex-
ene reactivity. In the first series of reactivity runs, 10 mM of the
metal complex was mixed with 10 mM of MCPBA and 10 mM of
the benzylic substrate. The only reaction that produces an oxidized
Me
absence of an organic substrate, the L n ligands undergo oxida-
tive degradation, as indicated by both thin layer chromatography
and NMR analysis.
3.2. Cyclohexene reactivity
product in more than trace (0.5%) quantities is that between ethyl-
Me
benzene and [Fe(L 2)Cl ], which produces 1-chloroethylbenzene
2
The reactivity was initially screened using cyclohexane and
in 15% yield (Scheme 4). No oxygenated or polychlorinated ben-
zylic products are observed in any of the reactions using 1 equiv. of
terminal oxidant.
In the second series of reactivity runs, excesses of oxidant and
substrate were used in attempts to increase the yields of the chlo-
rinated products. Although the yields of the chlorinated benzylic
compounds do increase, the reactions become much less selective
with the excess oxidant and generate more oxygenated side-
products. As was observed in the cyclohexene reactivity assays,
cyclohexene as substrates. Cyclohexane was found to be com-
pletely inert under the investigated conditions, in contrast to
similar systems previously reported by Que and Comba [44,45].
Me
When 400 mM cyclohexene reacts with 10 mM of a [M(L n)Cl2]
metal complex and 10 mM MCPBA in MeCN, the product mixtures
include cyclohexene oxide, 3-cyclohexenol, 3-cyclohexenone, and
-chlorocyclohexene (Table S1). Cyclohexene is a poor substrate for
these systems in that the oxygenated products form readily from
the reaction of MCPBA with the olefin, even in the absence of metal
ions. Furthermore, oxygenated products continue to form during
the reaction workup, and the yields of these products sometimes
exceed the theoretical maxima based on the amount of MCPBA
3
FeCl is the most active benzylic chlorinating agent when MCPBA is
2
present in excess (Table 1). The manganese complexes continue to
produce fewer chlorinated organic products than their iron analogs
at parity of the other reaction conditions. Mixtures of MCPBA with

C.R. Goldsmith et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 24–30
27
Table 1
Reactivity of [M(LMen)Cl2]-derived oxidants with benzylic susbstrates.
Substrate
Toluene
Complex
Oxygenated products
Chlorinated products
None
None
None
MnCl2
Benzyl alcohol (0.3 mM)
Benzaldehyde (0.3 mM)
Benzaldehyde (0.8 mM)
None
Benzaldehyde (0.9 mM)
None
Benzyl alcohol (0.9 mM)
Benzaldehyde (1.0 mM)
1-Hydroxyethylbenzene
Benzylchloride (0.4 mM)
FeCl2
Benzylchloride (9.1 mM)
Benzylchloride (0.7 mM)
Benzylchloride (2.0 mM)
None
[
[
[
[
Mn(LMe2)Cl2]
Me
Fe(L 2)Cl2]
Me
Mn(L 4)Cl2]
Fe(LMe4)Cl2]
Benzylchloride (2.0 mM)
Ethylbenzene
None
FeCl2
None
(
7.7 mM)
1-Hydroxyethylbenzene
6.0 mM)
1-Hydroxyethylbenzene
0.4 mM)
1-Hydroxyethylbenzene
3.3 mM)
1-Hydroxyethylbenzene
0.3 mM)
1-Chloroethylbenzene (17.7 mM)
1-Chloroethylbenzene (0.2 mM)
1-Chloroethylbenzene (8.3 mM)
1-Chloroethylbenzene (0.3 mM)
(
[
[
[
[
Mn(LMe2)Cl2]
Fe(LMe2)Cl2]
Mn(LMe4)Cl2]
Fe(LMe4)Cl2]
(
(
(
None
1-Chloroethylbenzene (6.0 mM)
Cumene
None
FeCl2
1-Hydroxycumene (5.0 mM)
1-Hydroxycumene (42.0 mM)
None
None
None
None
None
1-Chlorocumene (9.0 mM)
[
[
[
[
Mn(LMe2)Cl2]
None
None
None
None
Me
Fe(L 2)Cl2]
Me
Mn(L 4)Cl2]
Fe(LMe4)Cl2]
The initial concentrations of reagents are as follows: metal complex (if present), 10 mM; MCPBA, 100 mM; benzylic substrate, 400 mM. All reactions were run for 60 min at
◦
2
3
C in MeCN under N2.
Fe(LMe2)Cl ] and [Fe(L 4)Cl ] chlorinate toluene to the same
Me
[
2
2
extent, although the iron complex with the tetramethylated lig-
and does tend to promote the formation of more oxygenated
products (Table 1). When ethylbenzene is used as a substrate,
the iron compound with the dimethylated ligand leads to more
1
-chloroethylbenzene than that with the tetramethylated ligand,
reminiscent of what was observed with cyclohexene (Table S1).
Me
The additional methyl groups in [Fe(L 4)Cl ] appear to hinder the
2
ability of the oxidant to activate C–H bonds on secondary carbons.
When cumene is employed as a substrate, only FeCl2 promotes
the formation of any oxidized products above the lower limit of
Me
detection; the oxidants derived from [M(L n)Cl ] precursors do
2
not react with cumene, despite this substrate having the weakest
C–H bond in the series.
In each case, the yield of meta-chlorobenzoic acid (MCBA), the
reduced form of the terminal oxidant, exceeds the yields of oxidized
benzylic products. In the reaction between FeCl2 and ethylben-
zene, the yield of MCBA is twice that of 1-chloroethylbenzene.
This demonstrates that alternate pathways for MCPBA reduction
Fig. 1. Reaction between 0.32 mM [Fe(LMe2)Cl2] and 0.36 mM MCPBA in MeCN at
◦
Me
23 C. Four scans shown: (A) [Fe(L 2)Cl2] prior to addition of oxidant, (B) Fe(II)
are operable. These pathways include ligand oxidation when a
complex plus MCPBA 5 s after addition, (C) reaction mixture 30 s after addition,
and (D) reaction mixture 60 s after addition. An expanded version of this figure is
available in the supporting information.
Me
[
M(L n)Cl ] complex is present.
2
3.4. Mechanistic analysis
(Fig. 1), where optical features associated with Fe(IV) species often
Reactions were performed in the absence of substrate in order
appear [28,49–51]. The absorbance quickly drops, signifying that
the species completely vanishes by 30 s. Parallel studies using
mass spectrometry detect a transient species with a peak m/z
ratio of 412.0538 (Fig. 2); this feature matches that predicted for
to gain insight into the nature of the active oxidant. The reaction
between [Fe(LMe2)Cl ] and MCPBA at 23 C generates a short-
◦
2
lived species, with heighted absorbance in the 550–650 nm region
Scheme 4.

2
8
C.R. Goldsmith et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 24–30
Fig. 2. Mass spectrum of [Fe(LMe2)Cl2]/MCPBA reaction mixture at 6 s. Identified species include [Fe(LMe2)Cl2] at 396.0580, [Fe(LMe2)(OH)Cl] at 378.0920, and [Fe(LMe2)(O)Cl2]
at 412.0544. A mass spectrum of the same sample acquired 5 min after the start of the reaction is included in the supporting information.
[
FeIV(LMe2)(O)Cl ]. As with the optical feature in Fig. 1, the intensity
zene, and cumene are on primary, secondary, and tertiary carbons,
respectively. When the concentration of MCPBA is limited to 1
equiv. relative to the metal complex, only the reaction between
ethylbenzene and the iron complex with the dimethylated lig-
2
of the 412.0538 feature decreases quickly with time.
4
. Discussion
Me
and with L 2, leads to any substrate oxidation (Scheme 4). The
The reactivity of the [M(LMen)Cl ] complexes was investigated
yield of the reaction, which converts ethylbenzene exclusively to 1-
chloroethylbenzene, is relatively low. As in the cyclohexene assays,
iron outperforms manganese as a chlorination agent and the iron
2
with a focus on the chlorination chemistry. Hydrocarbon oxy-
genation reactions with metal chloride compounds often generate
traces of organochlorides as by-products [52]. In certain cases,
stoichiometric [45,53] or catalytic halogenation [19] has been
observed. Our primary aim with the methylated bispicen ligands
and their metal complexes was to investigate the factors that mod-
ulate halogenation chemistry, in particular to determine whether
manganese is capable of substituting for iron as it does in other
oxidative systems. A secondary aim was to determine whether the
chlorination chemistry could be made regiospecific through the
incorporation of modest steric bulk onto the ligand framework.
Me
Me
complex with L 2 outperforms that with L 4.
The steric environments around substrate C–H bonds have been
found to impact the ability of moderately bulky oxidants to access
and activate them [54]. Of the three substrates, toluene has the
highest benzylic C–H bond dissociation energy (BDE), with most
−
1
estimates of the bond strength at approximately 88 kcal mol [47].
The BDE of the benzylic C–H bonds in ethylbenzene are estimated to
−
1
be 85 kcal mol ; whereas, that of the weak C–H bond in cumene is
−
1
approximately 83 kcal mol [48]. If neither substrate nor oxidant
sterics are important, one would expect cumene to be oxidized to
the greatest extent. The results suggest that the oxidant generated
4.1. Reactivity with cyclohexene
Me
from [Fe(L 2)Cl ] and MCPBA is sufficiently sterically hindered to
2
preclude the activation of C–H bonds on tertiary carbons, yet the
For each metal compound, the oxidation reaction with cyclo-
hexene yields a mixture of cyclohexene oxide, 3-cyclohexenone,
-cyclohexenol, and 3-chlorocyclohexene. The chlorinated product
is a minor component in the product distributions of most of the
reactions (Table S1). Much of the oxygenated products likely results
from the direct interaction of MCPBA with the olefin. The lack of 1,2-
dichlorocyclohexane in the product mixtures, however, is notable
since this is the major product in the free radical halogenation of
cyclohexene [15]. The observed chlorination in this system is more
consistent with a metal-based oxidant than chlorine radicals.
oxidant is not strong enough for toluene oxidation. The oxidant
Me
derived from [Fe(L 2)Cl ] shows the best reactivity with ethyl-
3
2
benzene, which is intermediate with respect to both the strength
and accessibility of the benzylic C–H bonds. The lack of reactivity
Me
with [Fe(L 4)Cl ] suggests that the additional methyl groups may
2
discourage the activation of C–H bonds on secondary, as well as
tertiary, carbons.
4.2.2. Reactivity with an excess of terminal oxidant
Chlorinated products result from the other metal com-
plex/substrate mixtures when large excesses of MCPBA and
substrate are used. In most cases, the chlorination product is formed
in under 40% yield (0.8 turnovers), given the maximum of 20 mM
that can be formed with the chloride present in solution. This
activity is rather modest relative to previously reported systems
[19,45]. As with these systems, the increase in the amount of ter-
minal oxidant added heightens the overall oxidation activity at
the cost of selectivity, as evident by the increased incidence of
oxygenated products [44,45]. Consistent with the other reactivity
assays reported in this manuscript, the oxidants derived from the
manganese chloride complexes chlorinate benzylic substrates to a
much lesser degree than their iron analogs. Although manganese
4
4
.2. Reactivity with benzylic substrates
.2.1. Reactivity with a stoichiometric amount of terminal
oxidant
The oxidation chemistry was subsequently investigated with
benzylic substrates (Scheme 3), which display less background
reactivity with the MCPBA oxidant (Table 1). With these substrates,
Me
the studies focused on the complexes with L 2, which showed the
most extensive chlorination activity in the cyclohexene assay, and
Me
L
4, which should be the bulkiest ligand. The benzylic substrates
provide a means to assess the regioselectivity of the metal-based
oxidants, given that the weak C–H bonds in toluene, ethylben-

C.R. Goldsmith et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 24–30
29
has proven superior to iron in other oxidation reactions [32–34,55],
this does not appear to be the case with chlorination, at least with
the reaction conditions investigated thus far.
Ethylbenzene continues to be the most active substrate with
respect to chlorination. Notably, even with the excess terminal oxi-
dant, ethylbenzene is not oxidized past 1-chloroethylbenzene. We
speculate that the additional steric bulk provided by the installed
chlorine atom precludes further reactivity. The iron compound
with the tetramethylated ligand enables chlorination of the ethyl-
benzene to 1-chloroethylbenzene to a lesser degree, but with no
oxygenated byproducts. When cumene and toluene are used as
substrates, however, the yields of the chlorinated products asso-
ciated with the two discrete iron complexes are equivalent within
error. Even with excess oxidant, neither of the systems contain-
Scheme 5.
and thereby prevent the oxidation of cumene. The lessened reac-
Me
Me
Me
Me
ing [Fe(L 2)Cl ] or [Fe(L 4)Cl ] can oxidize cumene, despite this
tivity of [Fe(L 4)Cl ] relative to [Fe(L 2)Cl ] is consistent with
2
2
2
2
substrate having the weakest C–H bond in the series of hydrocar-
bons.
the additional two methyl groups further impeding the substrate
oxidation.
With the excess MCPBA, the oxidant(s) formed from FeCl2 are
The primary obstacle to catalytic halogenation is that the ligands
are themselves susceptible to oxidation, as indicated by NMR and
TLC. The picolinylic C–H bonds have comparable bond dissociation
energies to the benzylic substrates [56]. Other work from our group
has found that the bispicen framework is much more dynamic than
previously thought [57,58]. The flexibility of the ligand appears to
correlate to a higher degree of self-oxidation, and consequently a
lower degree of substrate oxidation, than those seen in systems
using more constraining ligand backbones [44,45]. A rigid or oxida-
tively robust ligand therefore seems essential to efficient substrate
chlorination.
the most active chlorination agents for each benzylic substrate.
Me
Unlike the oxidants generated from the [Fe(L n)Cl ] complexes,
2
the FeCl /MCPBA mixture is capable of activating benzylic C–H
2
bonds on primary, secondary, and tertiary carbons (Table 1).
With cumene, FeCl2 catalyzes the oxygenation of the substrate
to 1-hydroxycumene, achieving 0.9 turnovers with respect to the
chlorination and 4.2 turnovers with respect to the hydroxylation.
Given the high conversion of MCPBA to MCBA (Scheme 4), the active
species is likely an iron complex with one or more equivalents
of deprotonated MCBA. One possibility is that the more anionic
coordination sphere provided by these carboxylate ligands may
promote chlorination. Increasing the negative charge around the
metal ion should weaken the bonds between the metal and mon-
odentate anionic ligands, such as hydroxide and chloride. Another
factor that could explain the superior chlorination activity of FeCl2
is its lack of an organic ligand. The bispicen deriviatives act as com-
peting substrates and decompose during the reactions. Elimination
of ligand degradation pathways may direct more of the MCPBA into
substrate oxidation.
5. Conclusion
The chlorination activity of manganese and iron complexes with
methylated derivatives of the ligand bispicen is reported. The bispi-
cen framework supports modest chlorination that is limited to
benzylic and allylic C–H bonds. The identity of the metal and the
extent of ligand methylation are found to have significant impacts
on the reactivity. Iron appears to enable more extensive chlorina-
tion than manganese. The methylated bispicen ligands appear to
discourage the chlorination of more sterically congested carbon
Subtle modifications in the ligand structure thereby appear
to modulate the regiospecificity of the iron-mediated chlorina-
tion. In the absence of a tetradentate ligand, the oxidants formed
can activate weak allylic or benzylic C–H bonds, regardless of the
centers, likely by restricting substrate access to the reactive por-
IV
tions of the generated [M (L)(O)Cl ] species. Although the changes
2
substrate’s steric character. With FeCl , the extent to which the
2
to the ligand framework are ostensibly minor, they endow the chlo-
rination reactions with a significant degree of regioselectivity.
hydrocarbon is oxidized scales with the BDE of its weakest C–H
bond(s). As methyl groups are installed on the ligand, the oxidants’
abilities to chlorinate C–H bonds on tertiary and secondary carbons
are curtailed. Ethylbenzene is most susceptible to chlorination by
Acknowledgements
Me
Me
[
Fe(L 2)Cl ] and [Fe(L 4)Cl ]-derived oxidants partly due to its
2 2
This work was funded by Auburn University and a Doctoral New
Investigator grant from the American Chemical Society-Petroleum
Research Fund. K. H. was the recipient of a summer research
scholarship provided by the Auburn University Cell and Molecular
Biosciences Program (NSF EPS-0814103).
weakened C–H bonds, relative to toluene, and partly due to the
greater accessibility of these bonds, relative to cumene.
4
.3. Mechanistic studies
When MCPBAisadded tothe [Fe(LMe2)Cl ]inthe absenceofsub-
strate, a transient species is observed spectrophotometrically. At
room temperature, the intermediate completely decays within 30 s.
2
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.molcata.2010.11.006.
Parallel studies with mass spectrometry lead us to speculate that
IV Me
the transient species is [Fe (L 2)(O)Cl ] and that this interme-
2
diate is responsible for the hydrocarbon chlorination (Scheme 5).
The presence of the second chloride in the oxidant distinguishes
this system from another iron-based halogenation system recently
reported by Comba, in which aliphatic chlorination is believed to
proceed through an [Fe (L)(O)Cl] species [44]. Results from Que’s
group suggest that the additional chloride may destabilize the
Fe(IV) oxidation state [51].
References
[
[
[
1] M. Lersch, M. Tilset, Chem. Rev. 105 (2005) 2471–2526.
2] H.M.L. Davies, R.E.J. Beckwith, Chem. Rev. 103 (2003) 2861–2904.
3] D. Fiedler, D.H. Leung, R.G. Bergman, K.N. Raymond, Acc. Chem. Res. 38 (2004)
349–358.
IV
[
[
[
4] C. Jia, T. Kitamura, Y. Fujiwara, Acc. Chem. Res. 34 (2001) 633–639.
5] D. Balcells, E. Clot, O. Eisenstein, Chem. Rev. 110 (2010) 749–823.
6] C. Copéret, Chem. Rev. 110 (2010) 656–680.
The steric bulk provided by the methyl groups are believed to
impede substrate access to the reactive portions of the oxidant
[7] A. Gunay, K.H. Theopold, Chem. Rev. 110 (2010) 1060–1081.

3
0
C.R. Goldsmith et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 24–30
[8] I.A.I. Mkhalid, J.H. Barnard, T.B. Marder, J.M. Murphy, J.F. Hartwig, Chem. Rev.
[33] A. Murphy, T.D.P. Stack, J. Mol. Catal. A 251 (2006) 78–88.
1
10 (2010) 890–931.
[34] G. Dubois, A. Murphy, T.D.P. Stack, Org. Lett. 5 (2003) 2469–2472.
[35] R. Gupta, A.S. Borovik, J. Am. Chem. Soc. 125 (2003) 13234–13242.
[36] A.S. Borovik, Acc. Chem. Res. 38 (2005) 54–61.
[
9] R.C. Larock, Comprehensive Organic Transformations, Wiley-VCH, New York,
NY, 1999.
[
[
[
[
[
10] G.A. Russell, J. Am. Chem. Soc. 79 (1957) 2977–2978.
11] G.A. Russell, J. Am. Chem. Soc. 80 (1958) 4987–4996.
12] C. Walling, M.F. Mayahi, J. Am. Chem. Soc. 81 (1959) 1485–1489.
13] D.D. Tanner, J.E. Rowe, A. Potter, J. Org. Chem. 51 (1986) 457–460.
14] S.R. Jensen, W.A. Brown, E. Heath, D.G. Cooper, Biodegradation 18 (2007)
[37] H. Toftlund, E. Pedersen, S. Yde-Andersen, Acta. Chem. Scand. A 38 (1984)
693–697.
[38] J. Glerup, P.A. Goodson, A. Hazell, R. Hazell, D.J. Hodgson, C.J. McKenzie, K.
Michelsen, U. Rychlewska, H. Toftlund, Inorg. Chem. 33 (1994) 4105–4111.
[39] C. Baffert, M.-N. Collomb, A. Deronzier, S. Kjaergaard-Knudsen, J.-M. Latour,
K.H. Lund, C.J. McKenzie, M. Mortensen, L.P. Nielsen, N. Thorup, Dalton Trans.
(2003) 1765–1772.
[40] A.L. Nivorozhkin, E. Anxolabéhère-Mallart, P. Mialane, R. Davydov, J. Guilhem,
M. Cesario, J.-P. Audière, J.-J. Girerd, S. Styring, L. Schussler, J.-L. Seris, Inorg.
Chem. 36 (1997) 846–853.
7
03–717.
[
[
15] M.L. Poutsma, J. Am. Chem. Soc. 87 (1965) 2161–2171.
16] M.L. Alegre, M. Geronés, J.A. Rosso, S.G. Bertolotti, A.M. Braun, D.O. Mártire, M.C.
Gonzalez, J. Phys. Chem. A 104 (2000) 3117–3125.
[17] D.O. Mátire, J.A. Rosso, S. Bertolotti, G.C. Le Roux, A.M. Braun, M.C. Gonzalez, J.
Phys. Chem. A 105 (2001) 5385–5392.
[41] G.R. Newkome, Y.A. Frere, F.R. Fronczek, V.K. Gupta, Inorg. Chem. 24 (1985)
1001–1006.
[42] B. Chiswell, Inorg. Chim. Acta 12 (1975) 195–198.
[
[
[
[
18] T. Hori, K.B. Sharpless, J. Org. Chem. 44 (1979) 4204–4208.
19] D.T. Sawyer, J.P. Hage, A. Sobkowiak, J. Am. Chem. Soc. 117 (1995) 106–109.
20] Y. He, C.R. Goldsmith, SynLett 21 (2010) 1377–1380.
21] F.H. Vaillancourt, E. Yeh, D.A. Vosburg, S. Garneau-Tsodikova, C.T. Walsh, Chem.
Rev. 106 (2006) 3364–3378.
22] L.C. Blasiak, C.L. Drennan, Acc. Chem. Res. 42 (2009) 147–155.
23] S. Keller, T. Wage, K. Hohaus, M. Hölzer, E. Eichhorn, K.-H. van Pée, Angew.
Chem. Int. Ed. 39 (2000) 2300–2302.
24] V.M. Martínez, G.D. Cremer, M.B.J. Roeffaers, M. Sliwa, M. Baruah, D.E. De Vos,
J. Hofkens, B.F. Sels, J. Am. Chem. Soc. 130 (2008) 13192–13193.
25] C. Krebs, D.G. Fujimori, C.T. Walsh, J.M. Bollinger Jr., Acc. Chem. Res. 40 (2007)
[43] C. R. Goldsmith, C. M. Coates, K. Hagan,C. A. Mitchell, manuscript submitted.
[44] P. Comba, S. Wunderlich, Chem. Eur. J. 16 (2010) 7293–7299.
[45] R.A. Leising, Y. Zang, L. Que Jr., J. Am. Chem. Soc. 113 (1991) 8555–8557.
[46] Y.-M. Lee, S. Hong, Y. Morimoto, W. Shin, S. Fukuzumi, W. Nam, J. Am. Chem.
Soc. 132 (2010) 10668–10670.
[47] J. Berkowitz, G.B. Ellison, D. Gutman, J. Phys. Chem. 98 (1994) 2744–2765.
[48] I.W.C.E. Arends, P. Mulder, K.B. Clark, D.D.M. Wayner, J. Phys. Chem. 99 (1995)
8182–8189.
[
[
[
[
[
[
[
[
[49] L. Que Jr., Acc. Chem. Res. 40 (2007) 493–500.
4
84–492.
26] C. Wong, D. Galoni c´ Fujimori, C.T. Walsh, C.L. Drennan, J. Am. Chem. Soc. 131
2009) 4872–4879.
27] L.C. Blasiak, F.H. Vaillancourt, C.T. Walsh, C.L. Drennan, Nature 440 (2006)
68–371.
28] D.P. Galoni c´ , E.W. Barr, C.T. Walsh, J.M. Bollinger Jr., C. Krebs, Nat. Chem. Biol.
(2007) 113–116.
[50] N.D. Jabre, L. Hryhorczuk, J.J. Kodanko, Inorg. Chem. 48 (2009) 8078–8080.
[51] J.-U. Rohde, A. Stubna, E.L. Bominaar, E. Münck, W. Nam, L. Que Jr., Inorg. Chem.
45 (2006) 6435–6445.
[52] Y. Fujii, F. Ebina, M. Yanagisawa, H. Matsuoka, T. Kato, J. Inorg. Organomet.
Polym. 4 (1994) 273–288.
(
3
[53] T. Kojima, R.A. Leising, S. Yan, L. Que Jr., J. Am. Chem. Soc. 115 (1993)
11328–11335.
3
29] M.L. Matthews, C.M. Krest, E.W. Barr, F.H. Vaillancourt, C.T. Walsh, M.T. Green,
C. Krebs, J.M. Bollinger, Biochemistry 48 (2009) 4331–4343.
[54] M.S. Chen, M.C. White, Science 327 (2010) 566–571.
[55] A. Murphy, G. Dubois, T.D.P. Stack, J. Am. Chem. Soc. 125 (2003) 5250–5251.
[56] J.S. Roberts, M. Szwarc, J. Chem. Phys. 16 (1948) 981–983.
[57] G.J.P. Britovsek, J. England, A.J.P. White, Dalton Trans. (2006) 1399–1408.
[58] C. Hureau, G. Blondin, M.-F. Charlot, C. Philouze, M. Nierlich, M. Cesario, E.
Anxolabéhère-Mallart, Inorg. Chem. 44 (2005) 3669–3683.
[
[
30] S.V. Kryatov, E.V. Rybak-Akimova, Chem. Rev. 105 (2005) 2175–2226.
31] M. Costas, M.P. Mehn, M.P. Jensen, L. Que Jr., Chem. Rev. 104 (2004) 939–
9
86.
[32] A. Murphy, A. Pace, T.D.P. Stack, Org. Lett. 6 (2004) 3119–3122.