Polydentate pyridyl ligands and the catalytic activity of their iron(II) complexes in oxidation reactions utilizing peroxides as the oxidants

Article abstract of DOI:10.1016/j.molcata.2013.03.006

The paper describes the synthesis of iron(II) complexes bearing new polydentate N,O-coordinating pyridyl ligands and their catalytic application in oxidation reactions employing peroxides as the oxidants. The tridentate N,O,N (10) and N,N,O (11) ligands, the tetradentate N,N,O,N ligand 12 and the pentadentate N,N,N,O,N-coordinating ligand 16 were synthesized, and obtained as oils or solids in 74-93% isolated yields. The ligands were subsequently converted to the iron complexes [Fe(10)₂](OTf)₂, [Fe(11)₂](OTf)₂, [Fe(12)(OTf)₂] and [Fe(16)(OTf)](OTf), which were obtained as tan powders in 90-94% yield and characterized by various instrumental techniques. Preliminary screening experiments revealed that all complexes are catalytically active in the oxidation of secondary alcohols and benzylic methylene groups to the corresponding ketones. Optimization experiments with the complex [Fe(12)(OTf)₂] yielded a system that provided under mild condition ketones from benzylic methylene groups and secondary alcohols in 63-90% isolated yields (3 mol% catalyst loading, 3 equiv. H₂O₂ in CH ₃CN for 2 h at room temperature). Similar conditions utilizing environmentally friendly acetone as the solvent and 4 equiv. ^tBuOOH resulted in 36-65% isolated yields for some of the substrates, indicating a somewhat lower catalytic activity in that solvent. For the complexes [Fe(10)₂](OTf)₂ (two tridentate ligands), [Fe(12)(OTf)₂] (one tetradentate ligand) and [Fe(16)(OTf)](OTf) (one pentadentate ligand), the product formation for a test reaction was followed over time at significantly reduced catalyst loading to determine activities. Under these conditions, the complex [Fe(10)₂](OTf)₂ exhibited a somewhat lower catalytic activity compared to the other two complexes. Thus, the denticity seems to have an impact on catalytic activity although it is not dramatic, and a higher denticity appears to be beneficial for catalysis.

Full text of DOI:10.1016/j.molcata.2013.03.006

Accepted Manuscript
Title: Polydentate pyridyl ligands and the catalytic activity of
their iron(II) complexes in oxidation reactions utilizing
peroxides as the oxidants
Author: <ce:author id="aut0005"> Matthew Lenze<ce:author
id="aut0010"> Sergey L. Sedinkin<ce:author id="aut0015">
Eike B. Bauer
PII:
S1381-1169(13)00107-6
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.molcata.2013.03.006
MOLCAA 8698
To appear in:
Journal of Molecular Catalysis A: Chemical
Received date:
Revised date:
Accepted date:
15-12-2012
16-2-2013
10-3-2013
Please cite this article as: M. Lenze, S.L. Sedinkin, E.B. Bauer, Polydentate pyridyl
ligands and the catalytic activity of their iron(II) complexes in oxidation reactions
utilizing peroxides as the oxidants, Journal of Molecular Catalysis A: Chemical (2013),
http://dx.doi.org/10.1016/j.molcata.2013.03.006
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

*Graphical Abstract (for review)
Graphical Abstract
Highlights:
 New polydentate N,O-coordinating ligands were synthesized
 The ligands were converted to their respective iron complexes, which were thoroughly
characterized
 The new iron complexes were successfully employed in the oxidation of benzylic
methylene groups and secondary alcohols to the corresponding ketones under mild
reaction conditions, utilizing peroxides as the oxidants.
Page 1 of 49

*Research Highlights
Highlights:
 New polydentate N,O-coordinating ligands were synthesized
 The ligands were converted to their respective iron complexes, which were thoroughly
characterized
 The new iron complexes were successfully employed in the oxidation of benzylic
methylene groups and secondary alcohols to the corresponding ketones under mild
reaction conditions, utilizing peroxides as the oxidants.
Page 2 of 49

1
Polydentate pyridyl ligands and the catalytic
activity of their iron(II) complexes in oxidation
reactions utilizing peroxides as the oxidants
a
a
a*
Matthew Lenze, Sergey L. Sedinkin, Eike B. Bauer
a
University of Missouri - St. Louis, Department of Chemistry and Biochemistry,
One University Boulevard, St. Louis, MO 63121, USA
Tel.: +1-314-516- 5311; Fax: +1-314-516-5342.
bauere@umsl.edu
Revised Version
Keywords: Iron Catalysis, Oxidation Reactions, Polydentate Ligands, Green
Chemistry, Biomimetic Catalysis
Abstract
The paper describes the synthesis of iron(II) complexes bearing new polydentate
N,O-coordinating pyridyl ligands and their catalytic application in oxidation reactions
Page 3 of 49

2
employing peroxides as the oxidants. The tridentate N,O,N (10) and N,N,O (11) ligands,
the tetradentate N,N,O,N ligand 12 and the pentadentate N,N,N,O,N-coordinating ligand
16 were synthesized, and obtained as oils or solids in 74 to 94% isolated yields. The
ligands were subsequently converted to the iron complexes [Fe(10) ](OTf) ,
2
2
[Fe(11) ](OTf) , [Fe(12)(OTf) ] and [Fe(16)(OTf)](OTf), which were obtained as tan
2
2
2
powders in 90 to 94% yield and characterized by various instrumental techniques.
Preliminary screening experiments revealed that all complexes are catalytically active in
the oxidation of secondary alcohols and benzylic methylene groups to the corresponding
ketones. Optimization experiments with the complex [Fe(12)(OTf) ] yielded a system
2
that provided under mild conditions ketones from benzylic methylene groups and
secondary alcohols in 63 to 90% isolated yields (3 mol% catalyst loading, 3 equivalents
H O in CH CN for 2 h at room temperature). Similar conditions utilizing
2
2
3
t
environmentally friendly acetone as the solvent and 4 equivalents BuOOH resulted in 36
to 65% isolated yields for some of the substrates, indicating a somewhat lower catalytic
activity in that solvent. For the complexes [Fe(10) ](OTf) (two tridentate ligands),
2
2
[Fe(12)(OTf) ] (one tetradentate ligand) and [Fe(16)(OTf)](OTf) (one pentadentate
2
ligand), the product formation for a test reaction was followed over time at significantly
reduced catalyst loading to determine activities. Under these conditions, the complex
[Fe(10) ](OTf) exhibited a somewhat lower catalytic activity compared to the other two
2
2
complexes. Thus, the denticity seems to have an impact on catalytic activity although it is
not dramatic, and a higher denticity appears to be beneficial for catalysis.
Page 4 of 49

3
Introduction
The development of iron-based catalyst systems has become an intense area of
research in recent years [1]. Compared to other transition metals typically employed in
catalysis, iron is abundant, cheap, relatively non-toxic, and environmentally benign.
Intense research efforts in the past decade have produced a number of iron-based
catalysts that have been employed in carbon-carbon bond forming reactions [2], carbon-
heteroatom bond forming reactions [3], hetero-heteroatom bond forming reactions [4],
polymerizations [5], reductions [6] and other reactions [7].
In addition, iron catalysts have also been employed in oxidation reactions
[8,9,10]. Research efforts have been directed towards mimicking iron-containing
enzymes, that occur in nature and that catalyze oxidation reactions in living organisms
[11]. These enzymes are roughly categorized as heme and non-heme systems. An
example of a heme enzyme is Cytochrome P450 [12]. The Rieske dioxygenase [13]
(Figure 1) or methane monooxygenase [14] are non-heme enzymes [15] which all
catalyze a variety of oxidation reactions in nature. Research efforts in the past decades
have been directed towards mimicking the active, iron-containing sites of these enzymes
to create artificial systems that effect oxidation reactions similar to those found in nature
utilizing peroxides or related compounds as terminal oxidants [10,15a].
Page 5 of 49

4
The development of iron-based catalyst systems for oxidation reactions that
utilize peroxides is of high academic and industrial interest. Such catalyst systems follow
some of the principles of “Green Chemistry” [16], as only non-toxic substrates would be
involved in the corresponding oxidation reaction. The employment of less safe oxidants,
such as stoichiometric chromium reagents as in the commonly applied Jones reaction
[17], would not be needed, which would be of special interest in the pharmaceutical
industry [18]. Furthermore, a more mild oxidation protocol might also suppress unwanted
side reactions, such as C-C bond fissions, which are sometimes observed when strong
oxidants are employed [19].
As a consequence, a notable number of publications in recent years have dealt
with the synthesis and characterization of iron complexes [20], their catalytic application
in oxidation reactions and the corresponding mechanistic investigations [21]. Iron
complexes were applied in the oxidation of hydrocarbons or activated methylene groups
to alcohols or ketones [22], of alkenes or alkynes to diols or epoxides [23], of alkenes to
aldehydes [24], of aromatics to phenol derivatives [25] and of alcohols to ketones or
aldehydes [26]. Enantioselective versions for sulfide oxidation reactions [27] or
epoxidations [28] have been reported as well. A number of iron catalysts perform very
well in oxidation reactions compared to other metal complexes. Most of the iron
complexes to be employed in catalysis bear multidentate, nitrogen-based ligands, such as
BPMEM (1, Figure 2) and its derivatives [29], tris(picolyl)amine (tpa, 2) [30] or related
systems [31,32]. We have recently reported that iron complexes bearing monodentate
phosphoramidite ligands [33], bidentate phosphinooxazoline ligands (PHOX) [34],
bidentate α-imino-pyridine ligands (3, Figure 2) [35] and tridentate bis(pyridyl)amine
Page 6 of 49

5
ligands (4, Figure 2) are efficient catalysts for the oxidation of benzylic methylene groups
or alcohols to the corresponding ketones [36].
However, there is still room for improvement [37]. Some of the catalyst systems
reported in the literature require a large excess of the substrate over the oxidant to
achieve selectivity [38], especially when aliphatic hydrocarbons are the substrates. Iron
catalyzed oxidation reactions can follow a radical pathway, where basically oxygen is the
terminal oxidant. The other suggested mechanistic pathway proceeds through an iron oxo
species Fe=O and provides higher selectivity. It is however, not easy to predict which
mechanism a particular iron complex would follow [39]; the pathway also depends on the
reaction conditions such as the pH value [40] and slow addition of the oxidant might
suppress the formation of freely diffusing radicals [41,42].
As part of our long-standing research interests in establishing structure-activity
relationships between ligands and their corresponding metal complexes [43] we were
interested in determining how polydentate N,O-coordinating ligands perform in catalytic
Page 7 of 49

6
oxidation reactions when employed as ligands on iron centers. The majority of iron
catalysts bear polydentate ligands with nitrogen donor atoms. These complexes, however,
do not accurately reflect, e.g., the active center of the non-heme Rieske dioxygenase
enzyme (Figure 1), where an iron(II) metal center is coordinated to two histidine residues
and a bidentate aspartate group [13]. Still, it has been our experience that multidentate
ligands perform better than mono- or bidentate ones [36] and it appears that polydentate
ligands are frequently employed in the development of new catalyst systems [10,44]. It
has been argued that pentadentate ligands are more suitable for Fe(III) catalyzed
oxidation reactions, proceeding through iron-centered Fe=O species rather than a free
radical mechanism [45]. Accordingly, we were interested in employing N,O coordinating
polydentate ligands to improve oxidation reactions beyond those achieved in pure
nitrogen-coordinating ligands. Herein, we wish to describe the synthesis of a set of
N,N,O-, N,O,N-, N,N,O,N- and N,N,N,O,N-coordinating ligands, which we subsequently
applied in the synthesis of their corresponding iron(II) complexes. We employed these
t
new iron complexes as catalysts in a number of oxidation reactions utilizing BuOOH and
H O as the oxidant in CH CN. We present also a catalytic system, that oxidizes
2
2
3
benzylic methylene groups and alcohols to ketones employing acetone as the solvent;
compared to commonly employed pyridine or CH CN solvents, acetone represents a
3
“green” alternative, making the process described herein a safe and environmentally
benign oxidation process.
Page 8 of 49

7
Ligand and iron complex syntheses
First, we sought access to multidentate, N,O-coordinating ligands. Pyridyl units
are known to coordinate well to iron(II) centers, and the commercial picolyl chloride salt
9 is a convenient building block for the incorporation of pyridyl rings into a ligand
architecture. The commercial aminoalcohol (S)-5 was envisaged as a common starting
material, whose O- and N-alkylation allows for access to a variety of multidentate
nitrogen and oxygen donor atoms (as outlined in Scheme 1). We sought to avoid –OH or
–NH units in the ligands, as they could increase in acidity upon coordination to a metal
center, complicating the coordination chemistry of the ligands and the workup of their
iron complexes. Accordingly, we “blocked” those groups through benzylation in cases
where no further functionalization was targeted.
Page 9 of 49

8
As –OH units are more acidic than –NH units and, thus, easier to alkylate under
basic conditions, the synthetic pathway started with their alkylation (Scheme 1). The
aminoalcohol (S)-5 was first stirred with NaOH in DMSO to provide the corresponding
alkoxide anion, which then was alkylated either with benzyl chloride or the picolyl salt 9
to give the ethers 6 and 7 in 94% and 87% yield, respectively [46].
The O-alkylated aminoalcohols 6 and 7 were then converted to the corresponding
N,O,N ligand 10 and the N,N,O ligand 11 by N-benzylation and N-picolylation. When the
aminoalcohol precursors 6 and 7 were stirred for 24 h in CH Cl with benzylbromide or
2
2
the picolylbromide 8, respectively, N-alkylation took place and the ligands 10 and 11
were isolated by column chromatography in 74% and 82% yield, respectively. The
picolyl bromide 8 was obtained by neutralization of the commercial 2-
(bromomethyl)pyridine hydrobromide (13) with K CO in a H O / CH Cl mixture
2
3
2
2
2
(Scheme 2). The bromide 8 turned out to be not very stable. It was, thus, immediately
used in alkylation reactions without further purification.
In order to obtain a tetradentate ligand, both the alcohol and amine unit in the
aminoalcohol 5 were alkylated utilizing the 2-(chloromethyl)pyridine hydrochloride 9 in
a NaOH-basic DMSO solution under conditions similar to the syntheses of the ligands 10
and 11 in Scheme 1. Double alkylation afforded the tetradentate N,O,N,N ligand 12 in
87% isolated yield after purification by column chromatography (Scheme 3).
Page 10 of 49

9
The pentadentate N,N,N,O,N ligand 16 was accessed by exhaustive alkylation of
the commercial 2-aminoethanol (14) with 2-(chloromethyl)pyridine hydrochloride (9)
using the NaOH/DMSO protocol described above for the ligands 10, 11 and 12 (Scheme
3). The hydrochloride 9 was first neutralized in NaOH/DMSO, and subsequently added to
the 2-aminoethanol (14) to give the ligand 16 in 85% isolated yield. In a related
procedure, the aminoalcohol 15 was exhaustively benzylated with benzyl bromide, giving
the tridentate ligand 17 in 93% isolated yield.
1
13
1
The ligands were characterized by H and C{ H} NMR spectroscopy, by MS or
HRMS and some by IR. All NMR spectra exhibited baseline purity and are given in the
Supporting Information. The NCH and OCH protons gave characteristic chemical
2
2
2
shifts between 2.5 and 5 ppm including the appropriate geminal J coupling constants
HH
Page 11 of 49

10
3
around 13 Hz and vicinal J coupling constants around 5 Hz. The ligands 10, 11 and
HH
12 were synthesized from the enantiopure aminoalcohol (S)-5. At this stage, we did not
plan for an enantioselective catalytic application of the ligands, and consequently, the
enantiopurity of the ligands and their intermediates were not established throughout the
course of their syntheses. However, we measured a specific optical rotation for the ligand
10, which was determined to be +67.4. Hence, this ligand shows optical activity. The
aminoalcohol starting material (S)-5 has a specific optical rotation of +57. Thus, ligand
10 is optically active and the chiral information from the starting material was not
completely lost during synthesis. However, the optical rotation alone does not allow for
determining an enantiomeric excess for ligand 10.
All the ligands described above were obtained in high overall yields in one- or
two-step procedures, and they could easily be synthesized on gram scales, making their
application in metal complex syntheses attractive. The ligands exhibit different
denticities, and the N- and O-donors take different positions in the ligand, making them
good candidates for catalytic investigations to probe the influence of the ligand
architecture on the activity of their respective iron complexes.
Iron Complex Syntheses
The ligands described in the previous section were subsequently converted to
their respective iron complexes. As an iron source, we selected the iron salt
[Fe(OTf) (CH CN) ], which had been previously employed by us [34] and others [47] in
2
3
2
the synthesis of iron complexes. It features the relatively non-coordinating OTf
–
counterion (OTf = CF SO ), which might be advantageous in potential catalytic
3
3
Page 12 of 49

11
applications, as an open coordination site can be created at the iron center, a precondition
for catalytic applications.
Accordingly, to a solution of [Fe(OTf) (CH CN) ] in CH CN, the ligands in
2
3
2
3
CH CN were slowly added with stirring, resulting in tan solutions. The solvent was
3
removed and the residues were washed with Et O to obtain the iron complexes
2
[Fe(10) ](OTf) , [Fe(11) ](OTf) , [Fe(12)(OTf) ] and [Fe(16)(OTf)](OTf) as tan solids
2
2
2
2
2
in 90 to 94% isolated yields. The ligand 17 gave under the reaction conditions employed
a powder, which was completely insoluble in any organic solvent, making
characterization efforts and potential applications meaningless. Its chemistry was, thus,
not pursued further.
The iron complexes were characterized by NMR, MS, HRMS, IR and elemental
analyses. Mass spectra and elemental analyses confirmed the coordination of two
tridentate ligands for the complexes [Fe(10) ](OTf) and [Fe(11) ](OTf) , and the
2
2
2
2
coordination of one ligand for the complexes [Fe(12)(OTf) ] and [Fe(16)(OTf)](OTf).
2
The complexes [Fe(11) ](OTf) , [Fe(12)(OTf) ] and [Fe(16)(OTf)](OTf) gave elemental
2
2
2
analyses in accordance with the coordinated number of ligands. The latter two gave
+
+
molecular ion peaks corresponding to the formula [Fe(12)(OTf)] and [Fe(16)(OTf)] .
2+
2+
For the dicationic species [Fe(10) ] and [Fe(11) ] , no molecular ions were observed,
2
2
but only monocationic fragments that resulted from ligand loss and coordination of one
+
OTf ion, such as [Fe(10)(OTf)] .
The NMR spectra of iron(II) complexes can be at times difficult to analyze, as
such samples can be paramagnetic or reside in the spin-crossover regime, resulting in
Page 13 of 49

12
significant line broadening and large paramagnetic shifts up to the invisibility of the
signals [29a,35]. However, the complexes [Fe(10) ](OTf) and [Fe(11) ](OTf) gave
2
2
2
2
reasonable NMR spectra with well-resolved signals, which are slightly broadened, either
1
due to spin-crossover phenomena or paramagnetic impurities. For both complexes, the H
NMR spectra of the complexes differed significantly from those of the free ligands,
indicating their coordination to the iron center. The XCH units (X = N,O) were shifted
2
around 0.5 ppm to higher or lower field with no clear trend observable. The complex
[Fe(10) ](OTf) exhibited somewhat more severe line broadening and the spectral
2
2
1
window in the H NMR was widened to 0 to 15 ppm, which was indicative either of
paramagnetic impurities or of dynamic processes in solution. Presumably for that reason,
13
1
the C{ H} NMR spectrum for the same complex showed a low signal-to-noise ratio, as
13
1
it has been our experience that acquisition of C{ H} NMR spectra of paramagnetic iron
complexes is difficult or impossible to obtain [29,35]. In turn, complex [Fe(11) ](OTf)
2
2
13
1
gave a well-resolved C{ H} NMR spectrum with a good signal-to-noise ratio. Again
the chemical shifts were different from the free ligands, indicating coordination. A more
1
complex H NMR spectrum was obtained for the complex [Fe(12)(OTf) ] that could not
2
be analyzed in detail.
All efforts to obtain X-Ray structures of the complexes have failed to date and
1
their exact coordination geometry, thus, has not been established. However, the H and
13
1
C{ H} NMR spectra of [Fe(10) ](OTf) and [Fe(11) ](OTf) showed only one set of
2
2
2
2
signals, suggesting that the ligands in the coordination sphere produce a C symmetric
2
complex. Possible coordination modes for these two complexes are given in Figure 3, and
Page 14 of 49

13
the NMR data were consistent with these structures. The number of peaks and their
intensities for [Fe(12)(OTf) ] did not allow for the determination of the exact
2
coordination geometry for that complex. Some line broadening also suggested that
perhaps more than one isomer was present in solution. Any efforts to separate isomers
were unsuccessful.
19
The complexes contain OTf counterions, which can be investigated by F NMR.
It has been reported that free OTf ions exhibit chemical shifts around –77 ppm, whereas
19
coordinated OTf exhibits a significant downfield shift to around 65 ppm [35,48]. The
F
NMR spectrum of the complex [Fe(11) ](OTf) showed a resonance at –75.9 ppm in the
2
2
non-coordinating solvent CDCl , which is in accordance with two tridentate ligands
3
coordinated to the iron center, and two non-coordinated OTF counterions. Only in
19
CD CN did complex [Fe(12)(OTf) ] give a reasonable F NMR spectrum, which
3
2
showed a resonance at –66.8 ppm. The resonance indicates also a non-coordinated OTf
counterion; in this case, the coordinating CD CN might have replaced the weakly
3
coordinating OTf, which has previously been observed in related systems [48]. In
Page 15 of 49

14
addition, IR spectra of the complexes [Fe(10) ](OTf) , [Fe(11) ](OTf) and
2
2
2
2
[Fe(12)(OTf) ] were recorded. They also showed typical stretches for the OTf
2
–1
–1
counterions around 1250 cm and at 1028 cm , and the former ones were slightly
desymmetrized. The desymmetrization of OTf stretches in IR spectra has been observed
before by us [34] and others [47], and indicates some bonding interaction between the
OTf ion and the iron complex in the solid state.
The analytical data available for the complexes confirm the general formula
[Fe(10) ](OTf) , [Fe(11) ](OTf) , [Fe(12)(OTf) ] and [Fe(16)(OTf)](OTf), and these
2
2
2
2
2
were subsequently applied in catalytic reactions.
Catalytic Applications
We subsequently utilized the previously described complexes as catalysts in
oxidation reactions utilizing peroxides as the oxidants. Aliphatic hydrocarbons can, under
these conditions, be oxidized to alcohols or ketones [22]; depending on the substrate, this
reaction might not be very selective, but the oxidation of cyclohexane is frequently
utilized to determine, whether the oxidation proceeds through a radical mechanism or
through an iron-centered oxidant [49]. A radical mechanism typically produces a low
alcohol/ketone product ratio during the oxidation of cyclohexane, and a metal-centered
oxidant typically provides high alcohol/ketone ratios [10]. Accordingly, for some of the
complexes described above, we first subjected cyclohexane to oxidation using H O or
2
2
t
BuOOH as the oxidant and a 3 mol% iron catalyst loading. The results are compiled in
Scheme 4.
Page 16 of 49

15
All complexes gave a low alcohol/ketone ratio between 0.5 and 1.1 for these test
reactions. It can, thus, be anticipated that the oxidation of hydrocarbons proceed through
a free radical mechanism, as reported previously for related systems [10].
To further investigate the involvement of radicals in the mechanism of the
reaction, we performed the experiments shown in Scheme 4 in the presence of the radical
scavengers 2,4,6-tri-tert-butylphenol (TTBP) and phenol. The TTBP scavenger did not
shut down the reactions, and in the presence of phenol, no reactions were observed.
However, these experiments are not conclusive, as it is known that TTBP can be oxidized
by peroxides [50] and phenol might coordinate through the oxygen atom to the iron
center, leading to catalyst deactivation. Still, the oxidation of alcohol might proceed
through different mechanisms, which we did not attempt to investigate further at this
time.
Page 17 of 49

16
To identify potential intermediates for the reactions in Scheme 4, we recorded
time-resolved UV-vis spectra for some of the new iron complexes under catalysis
conditions. Accordingly, we combined the substrates with each of the two iron complexes
[Fe(10) ](OTf) and [Fe(12)(OTf) ] in CH CN and recorded the UV-vis spectra. Then
2
2
2
3
we added the H O oxidant and recorded spectra in time intervals of one minute. The
2
2
results are compiled in Figure 4.
[Fe(10)2](OTf)2 in CH3CN, cyclohexane and
[Fe(10)2](OTf)2 before addition of
H2O2
H2O2
1 minute after addition of H2O2
3
2.5
2
2 minutes after addition of H2O2
3 minutes after addition of H2O2
4 minutes after addition of H2O2
5 minutes after addition of H2O2
6 minutes after addition of H2O2
1.5
1
0.5
0
330
430
530
630
730
Wavelength (nm)
3
2.5
2
[Fe(12)(OTf)2] in CH3CN, cyclohexane and H2O2
before addition of H2O2
1 minute after addition of H2O2
2 minutes after addition of H2O2
3 minutes after addition of H2O2
4 minutes after addition of H2O2
5 minutes after addition of H2O2
6 minutes after addition of H2O2
1.5
1
0.5
0
330
380
430
480
530
580
630
680
730
780
Wavelength (nm)
Figure 4. Time resolved UV-vis spectra of selected iron complexes in the
presence of substrate and H O . Details are given in the Experimental Section.
2
2
Page 18 of 49

17
Clearly, the two complexes behaved differently when the oxidant was added to
the solution. Complex [Fe(10) ](OTf) showed minimal spectroscopic changes when the
2
2
H O oxidant was added (Figure 4, top). However, for the complex [Fe(12)(OTf) ] the
2
2
2
UV-vis spectrum immediately changed after addition of the oxidant (Figure 4, bottom). A
new, broad band appeared around 580 nm, which decayed over the course of several
minutes. Bands around 580 to 620 nm were reported for iron peroxo species [Fe-O-O-R]
t
(R=H, Bu), which are potential intermediates for oxidation reactions [51]. However,
these are typically observed at low temperatures, where longer life times are observed.
Still, it appears that with [Fe(12)(OTf) ] a short lived species formed after addition of the
2
oxidant, whereas the complex [Fe(10) ](OTf) remained unchanged. In the complex
2
2
[Fe(12)(OTf) ], two open coordination sites are immediately available, as two of the
2
coordination sites are occupied by the weakly coordinating OTf counterions. Such open
coordination sites are not immediately available in [Fe(10) ](OTf) which might explain
2
2
why no reaction occurs right after the addition of the peroxide. The two complexes show
comparable activity as catalysts in oxidation reactions (vide infra). Thus, it cannot be
determined whether or not both catalysts operate by the same or different mechanisms
based on the UV-vis data in Figure 4.
Next, we were screening the new iron complexes as catalysts in the oxidation of a
t
number of test substrates utilizing BuOOH as the oxidant, and gas chromatography to
monitor the reaction. The goal of these screening efforts was to discover the scope of the
reaction and to identify potential “lead complexes”, which performed well. The screening
results are compiled in Table S1 (Supporting Information). We determined that all
Page 19 of 49

18
complexes were catalytically active in the oxidation of benzylic methylene groups (such
as in diphenylmethane) to the corresponding ketones. The oxidation of secondary
alcohols to the corresponding ketones was also successful. However, the oxidation of
primary alcohols to aldehydes or carboxylic acids was not observed under the reaction
conditions. Also, for any of the reactions, the GC yields for the different catalysts did not
differ significantly; in other words, for a given substrate, the yields were comparable for
the four catalysts. The employment of H O gave comparable results. Almost no
2
2
conversion from diphenylmethane to benzophenone was observed, when Fe(OTf) or an
2
in situ combination of Fe(OTf) and tmeda (= tetramethylethylenediamine, a bidentate
2
ligand) was employed as catalyst for the transformation, demonstrating that the
polydentate ligands are crucial for catalytic activity. On the other hand, an in situ
combination of Fe(OTf) and the ligand 12 in CH CN also showed almost no catalytic
2
3
activity. We speculate that under these dilute conditions coordination of the ligand to the
iron center is slow and/or incomplete, and consequently an insufficient catalytically
active iron species has accumulated in the solution.
As our screening efforts did not result in the identification of a “superior” lead
complex, we subsequently employed catalyst [Fe(12)(OTf) ] to determine isolated yields
2
of the oxidation products of a variety of substrates employing H O as the oxidant. The
2
2
results are compiled in Table 1. The conditions involved a three-fold excess of the
peroxide oxidant over the substrate and a precatalyst loading of 3 mol% in CH CN as the
3
solvent. These conditions have successfully been applied previously in related oxidation
reactions [34,35]. After shaking the reactants at room temperature for 2 hours, the
Page 20 of 49

19
corresponding products were chromatographically isolated in yields ranging from 63 to
90%. Full characterization data of the catalysis products are given in the supporting
information.
Page 21 of 49

20
Page 22 of 49

21
Under these conditions, benzylic methylene substrates were oxidized to ketones in
63 to 90% isolated yields (entries 1, 2, 4 and 5 in Table 1). Secondary alcohols were
oxidized to the corresponding ketones in 68 to 70% isolated yields (entries 7 to 9). Thus,
it appears that, under the conditions in Table 1, the oxidation of alcohols is slightly less
efficient than the oxidation of benzylic methylene groups.
The conditions described above involved CH CN as solvent, and preliminary
3
experiments in our laboratory showed, that pyridine solvent – also widely used for related
oxidation reactions [22] – can be employed as well (the reactions in Table 1, entries 6 and
10 were performed under these conditions). However, pyridine is deemed to be harmful,
and as such, an “undesired” solvent according to the principles of “Green Chemistry”
[16,52]. CH CN is categorized as “usable” but not “preferred” solvent [52]. We were,
3
thus, interested in determining whether another, more environmentally benign solvent
also could be employed in the title reaction. Acetone seemed to be a good choice. Indeed,
peroxides can detonate when in contact with acetone, especially at higher concentrations.
However, acetone is less harmful and environmentally more friendly than pyridine and is
considered a “green solvent” and categorized “preferred” [52].
Accordingly, we performed some oxidation reactions utilizing acetone as the
t
solvent and BuOOH as the oxidant, and the results are also compiled in Table 1. Under
conditions otherwise similar to those in CH CN, the corresponding oxidation products
3
were isolated in 36 to 65% yields. Again, benzylic methylene groups were oxidized to the
corresponding ketones (entries 1-3, 46 to 65% yield), and secondary alcohols to the
ketones in 36 to 39% yield (entries 11 and 12). Without catalyst, virtually no oxidation of
Page 23 of 49

22
diphenylmethane or cyclohexanol was observed. Once more, it appeared that the alcohols
were somewhat more difficult to oxidize compared to benzylic methylene groups, and the
oxidation reactions are less efficient in acetone compared to CH CN as the solvent.
3
Catalytic Activities
As mentioned above, the different iron complexes provided comparable yields in
test reactions. However, we were still interested in determining whether there are
differences in catalytic activities of iron complexes bearing ligands with different
denticities. In order to compare the catalytic activity of the iron complexes bearing either
a tridentate, tetradentate or pentadentate ligand, we followed the formation of the
benzophenone product over time for the oxidation of diphenylmethane to benzophenone
(entry 2 in Table 1). For this investigation, we employed the complexes [Fe(10) ](OTf)
2
2
(two tridentate ligands), [Fe(12)(OTf) ] (one tetradentate ligand) and [Fe(16)(OTf)](OTf)
2
(one pentadentate ligand). As previously reported [35,36], we reduced the amount of
catalyst to 0.3 mol% to slow down the reaction for easier monitoring and comparison.
Under these conditions, a pseudo-zero order rate law applies and observed rate constants
(k ) can be easily determined. The results are compiled in Figure 5.
obs
Page 24 of 49

23
60
50
40
30
20
10
0
kobs = 0.22 ± 0.01
kobs = 0.10 ± 0.02
[Fe(12)(OTf)2]
[Fe(10)2](OTf)2
[Fe(16)(OTf)](OTf)
kobs = 0.0073 ± 0.007
0
50
100
150
200
time / min
Figure 5. Determination of activities. The formation of benzophenone from
diphenylmethane (Table 1, entry 2) was followed at a significantly reduced catalyst load
of 0.3 mol% over time. Details see Experimental.
The complexes [Fe(12)(OTf) ] and [Fe(16)(OTf)](OTf) showed similar activities
2
–1
with observed rate constants of 0.22±0.01 and 0.1±0.02 min , respectively. The k
obs
value of complex [Fe(16)(OTf)](OTf) is lower, but - considering error bars - the overall
performance of the two complexes does not differ significantly. However, under the
reaction conditions in Figure 5, the complex [Fe(10) ](OTf) bearing two tridentate
2
2
–1
ligands exhibited a significantly lower k value (0.0073±0.007 min ) than the other
obs
two complexes. As mentioned previously, in catalytic reactions under the conditions in
Table 1 (where a 3 mol% catalyst load was employed), the new complexes described
herein did not give significant differences in yields. However, it appears that at a
Page 25 of 49

24
significantly reduced catalyst load, a small dependency of the activity on the denticity of
the ligands employed can be observed.
It has previously been argued that two tridentate ligands in the coordination
sphere of iron complexes might reduce their catalytic activity in oxidation reactions, as
all six coordination sites are occupied by ligands [8]. In turn, in the complexes
[Fe(12)(OTf) ] and [Fe(16)(OTf)](OTf), one or two coordination sites are occupied by
2
the weakly coordinated OTf ion, and in coordinating solvents, the triflate might be
replaced by solvent molecules. Accordingly, the complex [Fe(12)(OTf) ] gave in
2
19
coordinating CD CN F NMR signals consistent with free OTF ions, which had been
3
completely replaced by CD CN in solution (vide supra). Thus, the complexes
3
[Fe(12)(OTf) ] and [Fe(16)(OTf)](OTf) might more easily provide an open coordination
2
site, as required for catalysis, whereas such a site is not readily available in
[Fe(10) ](OTf) . Still, the performance of all complexes was comparable when a 3 mol%
2
2
catalyst load was employed (Table S1 of the supporting information). Under these
conditions, a significant amount of [Fe(10) ](OTf) might undergo a permanent chemical
2
2
change to another species, that is catalytically active, and that does not accumulate to a
considerable extent in solution, when only 0.3 mol% catalyst is employed.
It has been discussed previously that under oxidative reaction conditions the
ligands of transition metal catalysts can oxidatively decompose [53]. For example,
polypyridyl ligands in manganese-based oxidation catalysts can be oxidized to pyridine-
2-carboxylic acid [54]. A combination of pyridine-2-carboxylic acid and a manganese
source showed catalytic activity in alkene oxidation reactions similar to those observed
Page 26 of 49

25
for the original polypyridyl complexes. We have recently published an investigation of
ligand decomposition pathways in iron catalyzed oxidation reactions; we subjected
iron(II) complexes to oxidative conditions similar to those employed in catalysis,
displaced the ligand by addition of aqueous NaCN and analyzed the displacement
products by NMR and GC [36]. We determined that -amino pyridine ligands in iron
complexes are oxidized to a variety of amides. We have not observed pyridine-2-
carboxylic acid decomposition products in this study. However, it is known that iron(II)
salts in combination with pyridine-2-carboxylic acid are catalytically active in oxidation
reactions employing peroxide oxidants [55]. Preliminary studies to determine ligand
oxidation products were performed for the iron complex [Fe(12)(OTf) ] (see Supporting
2
Information). In this initial experiment, we observed a variety of ligand decomposition
products but we could not detect pyridine-2-carboxylic acid (however, it cannot be
excluded that it was lost during workup in this case). Furthermore, we were not able to
recover unoxidized iron catalyst starting material after catalysis. Thus, it cannot be
completely ruled out that the ligand decomposition product pyridine-2-carboxylic acid
forms the actual catalyst. On the other hand, as seen from Figure 5, different ligands
show different catalytic activities, and thus we assume that the catalytically active species
is not merely an iron(II) center in combination with pyridine-2-carboxylic acid derived
from ligand decomposition. Further studies will be performed in due course to identify
ligand decomposition products and the actual catalytic species in solution.
Iron complexes have been employed before in the oxidation of activated
methylene groups and alcohols. It has previously been shown that polydentate ligands
can have a beneficial effect on the catalytic activity of non-heme, iron-based catalyst
Page 27 of 49

26
systems in the oxidation of benzylic methylene groups [56]. The employment of iron salts
in corresponding oxidation reactions has been reported [22d, 55, 57], but often extended
reaction times or elevated reaction temperatures are required [22,57]. Iron-catalyzed
oxidations of secondary alcohols utilizing peroxides have also been documented [26];
some systems require heat [26a,d,g], and some systems perform the corresponding
transformations at room temperature [26i,l]. The catalysts reported herein perform the
oxidations of benzylic methylene groups or alcohols at room temperature within 2 to 8
hours. These are mild conditions compared to some other catalyst systems, however, it
requires more elaborate ligand syntheses as shown in Schemes 1 and 3.
Conclusion
Overall, the study revealed a set of iron complexes to be catalytically active in the
oxidation of benzylic methylene groups and alcohols to the corresponding ketones. It
seems that the ligand structure of the corresponding iron complexes has only a minor
impact on the yields of the reaction, as determined by screening experiments. The
oxidation of benzylic methylene groups and secondary alcohols in CH CN and H O as
3
2 2
the oxidant gave isolated yields between 63 and 90% under mild conditions. The yields
are comparable to other catalyst systems that have previously been employed in these
reactions. Acetone is a “greener” solvent compared to CH CN, but when employed in the
3
t
oxidation reactions employing BuOOH as the oxidant, this solvent tended to give lower
isolated yields of 36 to 65% compared to those obtained in CH CN. Still, the reactions
3
can be performed in acetone, making the process more environmentally benign and
Page 28 of 49

27
rendering it in accordance with the principles of “Green Chemistry”. Finally, we
investigated the activity of three complexes bearing ligands with different denticities in a
test reaction at significantly reduced catalyst load of 0.3 mol%. The complexes with a
tetradentate or a pentadentate ligand performed better than the complex bearing two
tridentate ligands in the coordination sphere. However, these differences in reactivity did
not translate to different yields in regular oxidation reactions employing a 3 mol%
catalyst load. The results presented herein add to the knowledge of iron-catalyzed
oxidation reactions employing peroxide oxidants, which become, overall, a serious
alternative to other oxidation reactions employing more toxic solvents and reagents.
Further investigations involving the new complexes described herein are underway.
Experimental Section
General Methods
Chemicals were treated as follows: THF, toluene, diethyl ether, distilled from
Na/benzophenone; CH Cl , distilled from CaH . All reactions were carried out under an
2
2
2
atmosphere of nitrogen applying Schlenk techniques if not stated otherwise.
Ethanolamine 14, 2-((2-aminoethyl)amino)ethanol (15), 2-(chloromethyl)pyridine-1-ium
chloride (9), NaH (60% suspension in mineral oil), 2-(bromomethyl)pyridine-1-ium
bromide (13), NaOH (all Aldrich), Silica gel, 200-400 mesh, 60 Å (Sigma-Aldrich), used
as received. (S)-2-(benzylamino)-1-phenylethanol (S)-5 (Aldrich, used as received) was
obtained free of charge from a local company. [Fe(OTf) (MeCN) ] was synthesized
2
2
following a literature procedure [58]. NMR spectra were obtained at room temperature on
1
a Bruker Avance 300 MHz or a Varian Unity Plus 300 MHz instrument ( H: 300.13
Page 29 of 49