Halogen-Bonding-Assisted Iodosylbenzene Activation by a Homogenous Iron Catalyst

Article abstract of DOI:10.1002/chem.201503112

The iron(III) complex of hexadentate N,N,N′-tris(2-pyridylmethyl)ethylendiamine-N′-acetate (tpena^-) is a more effective homogenous catalyst for selective sulfoxidation and epoxidation with insoluble iodosylbenzene, [PhIO]_n, compared with soluble methyl-morpholine-N-oxide (NMO). We propose that two molecules of [Fe(tpena)]²⁺ cooperate to solubilize PhIO, extracting two equivalents to form the halogen-bonded dimeric {[Fe(tpena)OIPh]₂}⁴⁺. The closest intradimeric I...O distance, 2.56 ?, is nearly 1 ? less than the sum of the van de Waals radii of these atoms. A correlation of the rates of the reaction of {[Fe(tpena)OIPh]₂}⁴⁺ with para-substituted thioanisoles indicate that this species is a direct metal-based oxidant rather than a derived ferryl or perferryl complex. A study of gas-phase reactions indicate that an ion at m/z=231.06100 originates from solution-state {[Fe(tpena)OIPh]₂}⁴⁺ and is ascribed to [Fe^III(tpenaO)]²⁺, derived from an intramolecular O atom insertion into an Fe-tpena donor bond. Proposed ion pairs, {[Fe(tpena)OIPh]Cl}⁺ and {[Fe(tpena)OIPh]ClO₄}⁺, are more stable than native [Fe(tpena)OIPh]²⁺ ions, suggesting that halogen-bonding, as for the solution and solid states, operates also in the gas phase.

Full text of DOI:10.1002/chem.201503112

DOI: 10.1002/chem.201503112
Full Paper
&
Halogen Bonding
Halogen-Bonding-Assisted Iodosylbenzene Activation by
a Homogenous Iron Catalyst
David P. de Sousa,[a] Christina Wegeberg,[a] Mads Sørensen Vad,[a] Steen Mørup,[b]
Cathrine Frandsen,[b] William A. Donald,[c] and Christine J. McKenzie*[a]
Abstract: The iron(III) complex of hexadentate N,N,N’-tris(2-
pyridylmethyl)ethylendiamine-N’-acetate (tpenaÀ) is a more
effective homogenous catalyst for selective sulfoxidation
and epoxidation with insoluble iodosylbenzene, [PhIO]n,
compared with soluble methyl-morpholine-N-oxide (NMO).
We propose that two molecules of [Fe(tpena)]2+ cooperate
to solubilize PhIO, extracting two equivalents to form the
halogen-bonded dimeric {[Fe(tpena)OIPh]2}4+. The closest in-
tradimeric I···O distance, 2.56 Š, is nearly 1 Š less than the
sum of the van de Waals radii of these atoms. A correlation
of the rates of the reaction of {[Fe(tpena)OIPh]2}4+ with para-
substituted thioanisoles indicate that this species is a direct
metal-based oxidant rather than a derived ferryl or perferryl
complex. A study of gas-phase reactions indicate that an ion
at m/z=231.06100 originates from solution-state {[Fe(tpe-
na)OIPh]2}4+ and is ascribed to [FeIII(tpenaO)]2+, derived from
an intramolecular O atom insertion into an Fe–tpena donor
bond. Proposed ion pairs, {[Fe(tpena)OIPh]Cl}+ and {[Fe(tpe-
na)OIPh]ClO4}+, are more stable than native [Fe(tpe-
na)OIPh]2+ ions, suggesting that halogen-bonding, as for the
solution and solid states, operates also in the gas phase.
Introduction
system is amongst the most efficient in terms of rates for cata-
lytic systems utilizing PhIO as terminal oxidant, despite the
heterogeneity.
The hypervalent iodine compound, iodosylbenzene (PhIO), is
a convenient oxidant because it is easy to prepare and a rela-
tively stable solid. It suffers disadvantages, however, with re-
spect to implementation in homogenous catalysis owing to its
general insolubility. Efforts to overcome this have included
functionalizing the benzene group with solubilizing sulfonate
groups.[1] In an alternative approach, we have communicated
that the iron(III) complex of the carboxylato-containing hexa-
dentate ligand, N,N,N’-tris(2-pyridylmethyl)ethylendiamine-N’-
acetate (tpenaÀ), not only aids PhIO solubilization but also cat-
alyzes the selective sulfoxidation of thioanisole.[2] This reaction
is interesting as the oxidant is provided, as needed, during the
course of the reaction: Thus, substrates are not potentially del-
eteriously exposed to large amounts of excess oxidant and se-
lectivity is enhanced. Our initial observations suggest that this
TpenaÀ is a potentially hexadentate chelating ligand. In this
respect, it would seem at face value that the property of open
coordination sites, for enabling reactant(s) and product coordi-
nation, possibly simultaneously, is challenged. The X-ray crystal
structure of the as yet unique iron(III)–OIPh adduct, [Fe(tpe-
na)OIPh](ClO4)2,[2] shows, however, that the iron(III) is seven-co-
ordinated. Although hepta-coordination is common for high-
spin iron(III), this geometry is sometimes overlooked in catalyt-
ic mechanisms proposed for synthetic catalysts and enzymes
alike. Similar to other iron(III)-terminal oxidant adducts (super-
oxo,[3] peroxo,[4] hydroperoxo,[5] alkylperoxo,[6] hypochlorite[7]),
iodosylaryls[2,8] have been variously proposed as the direct cat-
alytically competent iron and manganese-based oxidants;
however, unlike the aforementioned superoxo and peroxo ad-
ducts, there is no general consensus for the assignment of the
salient spectroscopic fingerprints for metal–OIPh adducts. The
present work most inconveniently divulges few salient features
for recognizing this important moiety. It does, however, shed
light on the way in which iron(III) complexes can activate in-
soluble PhIO.
[a] D. P. de Sousa, C. Wegeberg, Dr. M. S. Vad, Prof. C. J. McKenzie
Department of Physics, Chemistry and Pharmacy
University of Southern Denmark, Campusvej 55, 5230 Odense M (Denmark)
E-mail: mckenzie@sdu.dk
[b] Prof. S. Mørup, Prof. C. Frandsen
Department of Physics, Technical University of Denmark
DK-2800, Kongens Lyngby (Denmark)
The use of PhIO as a terminal oxidant is not generally per-
ceived as desirable within the paradigms of green chemistry.
This situation is only worsened if soluble derivatives need to
be used. Thus, a strategy of employing a metal complex to
both labilize and activate PhIO, as described here, has merits.
Not only is the issue of solubility resolved, the presence of an
excess of solubilized oxidant in the solution is also avoided.
This will reduce the risk of the ubiquitous side-products often
[c] Dr. W. A. Donald
School of Chemistry, University of New South Wales
Sydney, NSW (Australia)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201503112.
Part of a Special Issue “Women in Chemistry” to celebrate International
Women’s Day 2016. To view the complete issue, visit: http://dx.doi.org/
chem.v22.11.
Chem. Eur. J. 2016, 22, 3810 – 3820
3810
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
observed in catalytic oxidation chemistry. An environmentally
friendly system for recycling PhIO from the cleanly generated
side-product PhI would be very interesting with respect to po-
tential applications for the efficient and selective catalytic oxi-
dation reactions.
Results and Discussion
The most convenient starting material for the chemistry de-
scribed here is the oxo-bridged [Fe2(m-O)(tpenaH)2](ClO4)4
(1(ClO4)4).[2] By dissolving 1(ClO4)4 in non-aqueous solvents
(acetonitrile, acetone), pseudo-dehydration [Eq. (1)] effects oxo-
bridge cleavage to give [Fe(tpena)]2+ (2). In this process, two
uncoordinated protonated pendant pyridines are deprotonat-
ed and the oxo bridge is concurrently protonated to form
water. This takes approximately 15 min at room temperature,
whereupon the color of the solution changes from yellow–
brown to red (Figure S2 in the Supporting Information).
By adding/diffusing diethyl ether at À408C into solutions of
2, solid and hygroscopic [Fe(tpena)](ClO4)2 (2(ClO4)2) can be iso-
lated in powder and/or crystalline forms. The powders appear
to be lighter orange in color compared with the red–orange
crystals—more than can be warranted by particle size differen-
ces. IR and Mçssbauer spectroscopy (see below) show that the
bulk powder and crystalline products are mixtures of 2(ClO4)2
and 1(ClO4)4. Photos of the red–orange crystals of 2(ClO4)2
show that they are dispersed with crystals of its brown
pseudo-hydrate, 1(ClO4)4 (Figure S1 in the Supporting Informa-
tion). Complex 2 can be expected to exist as fac-py3 (2fac; py=
pyridine) and mer-py3 (2mer) isomers (Figure 1) and the dehy-
dration of 1 [Eq. (1)] can potentially yield both. In contrast to
in the X-ray crystal structure of 1, tpenaÀ has fully realized its
potential as a hexadentate ligand in the X-ray crystal structure
obtained for the red–orange crystals of 2fac (Figure 1, Tables S1
and S2 in the Supporting Information). The iron(III)–donor
bonds are all less than 2 Š. When compared to iron(III) com-
plexes with as close as possible N5O coordination,[9] this would
seem consistent with a low-spin (S=1/2) d-electron configura-
tion. The relatively close intermolecular interactions and
a Fe···Fe distance of 6.078(1) Š between pairs of dicationic 2fac
is notable. This seems to suggest a strong association through
non-classic hydrogen-bonding in short intradimer a-pyCH···O
interactions involving the coordinated and non-coordinated
carboxylato O atoms and p–p stacking of pyridines, which can
overcome the coloumbic repulsions between the dications of
2fac. Four counter anions and acetonitrile solvate surround the
dimer. It is relevant to note that the closest intermolecular
M···M distances in cobalt(III)[10] and chromium(III)[11] analogues
are longer (8.844 Š and 7.517 Š) and there is no such distinctly
obvious pairing in the solid state.
Figure 1. The associated cations of 2fac in the X-ray crystal structure of
[Fe(tpena)](ClO4)2(CH3CN)1.5 and chemical diagrams of the fac-py3 (2fac) and
mer-py3 (2mer) geometrical isomers. Displacement ellipsoids are drawn at the
50% probability level.
Strong halogen-bonding enables isolation of an intrinsically
unstable PhIO complex
If solid PhIO is added to dry acetonitrile solutions containing 2,
an immediate color change from red to orange is seen, which
is indicative of the formation of [Fe(tpena)OIPh]2+ (3), which
can be isolated as its perchlorate salt by rapid precipitation on
addition of diethyl ether at À408C. We have previously com-
municated the X-ray crystal structure of 3(ClO4)2.[2] A compari-
son of the conformations of tpenaÀ in 2fac and 3 suggests that
the direct solution-state precursor for 3 is most likely the non-
structurally characterized 2mer isomer as the PhIO is inserted cis
to the three pyridines and the carboxylato ligands are more or
less trans to the two amine N atoms; this forms the seven-co-
ordinated iron(III) center in 3. The supramolecular halogen-
bonded association of pairs of dicationic complexes that are
evident in the X-ray crystal structure have not been previously
described (Figure 2). The iron atoms in the dimers are separat-
ed by 8.011(1) Š and the closest intermolecular I···O distance is
2.555(3) Š. At 0.94 Š less than the sum of the van der Waals
radii of the respective atoms (3.50 Š[12]), this distance indicates
very strong halogen-bonding. As depicted, the other intermo-
lecular I···O interactions are also well below 3.5 Š. For refer-
ence, the energy of a single I···O halogen-bond between iodo-
benzene and DMSO has been calculated to be about 3 kcal
molÀ1 [13]
The cumulative power of six such close supramolec-
.
ular intradimer halogen-bonded contacts clearly achieves the
stability needed for the crystallization of 3, which is otherwise
Chem. Eur. J. 2016, 22, 3810 – 3820
www.chemeurj.org
3811
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Dimerization quenches carboxylato vibrations
The IR spectrum of [PhIO]n(s) contains bands at 440, 491, and
595 cmÀ1. In 18O-labeled [PhI18O]n(s), these are shifted to 414,
482, and 564 cmÀ1, respectively (Figure S4 in the Supporting
Information). These are absent in the IR spectrum of 3(ClO4)2(s)
(Figure 3). Weak bands at 684 and 654 cmÀ1 are probably asso-
Figure 2. Halogen-bonded dimer formation in the X-ray crystal structure of
dicationic [3]2 showing intermolecular I···O distances. Displacement ellipsoids
are drawn at the 50% probability level.
expected to be highly reactive. Solid-state samples are stable
for months to years at À408C. PhIO and its soluble (e.g., sulfo-
nated) derivatives have been used to prepare iron(IV), manga-
nese(V), chromium(V) and cobalt(IV) high-valent metal–oxo
complexes[14] by oxygen atom transfer [Eq. (2)]; however, the
putative intermediate metal–OIPh species in these systems
have not been isolated. This could suggest that in most cases
metal complexes of PhIO are more reactive than the derivative
higher-valent metal–oxo products. In turn, this balance will de-
termine which of the metal-based oxidants, [LM=O] or [LM–
OIPh], are pertinent in the catalysis of oxidation reactions
using PhIO as the terminal oxidant in metal-catalyzed reac-
tions.
Figure 3. IR spectra of [3(ClO4)2]2 (gray) and 1(ClO4)4 (black).
ciated with the coordinated PhIO; however, these bands were
not shifted in 18O-labeled samples of 3 prepared by using
PhI18O. A very weak unassigned band at 470 shifts to 453 cmÀ1
in the spectrum of 18O-labeled 3(ClO4)2 (Figure S5 in the Sup-
porting Information). Thus, the unique solid-state metal–OIPh
complex 3(ClO4)2 does not reveal any particularly distinctive IR
features that might be used as markers for identifying metal-
coordinated PhIO. There is no doubt that the label is present;
however, if the 18O-labeled 3(ClO4)2 is dissolved in acetonitrile
containing one equivalent of thioanisole, this is oxidized and
18O-labeled methyl phenyl sulfoxide is produced (Figure S9 in
the Supporting Information). Similarly inconvenient, there are
(surprisingly) no distinctive salient features that distinguish
2(ClO4)2(s) and 3(ClO4)2(s) by IR spectroscopy and particularly
conspicuous in the spectra of both compounds is the absence
or low intensity of the tpenaÀ-derived carboxylate stretches at
nas(COO) =1661 cmÀ1 and ns(COO) =1340 cmÀ1. These bands would
normally be expected to dominate the spectra of complexes of
carboxylato ligands, and are typically used to determine the
coordination mode. These stretches are the most intense in
the spectrum of 1(ClO4)4. A low intensity feature at 1661 cmÀ1
in the spectra of 2(ClO4)2 and 3(ClO4)2, originating from un-
avoidable contamination by 1(ClO4)4, was variable from sample
to sample. This is backed up by traces of the absorption band
at 837 cmÀ1 assigned to the nas(Fe-O-Fe) stretch, which is shifted
to 771 cmÀ1 when labeled with 18O (Figure S6 in the Support-
ing Information). The unexpected quenching of nas(COO) in
2(ClO4)2(s) and 3(ClO4)2(s) is attributed to solid-state dimer for-
mation with strong intermolecular interactions between the di-
cations in both structures (Figures 1 and 2).[15] This generates
a pseudo-inversion center resulting in the carboxylato groups
The dimeric associations found in the solid state of 2fac and
3, along with the fast reaction between 2 and insoluble PhIO,
could point towards cooperativity between two molecules of 2
to break the polymeric structure of PhIO to form the unexpect-
edly relatively robust 3. The question arises: do the halogen-
bonded dimers of 3 observed in the solid state remain intact
in solution? When 3(ClO4)2 is dissolved in non-protic solvents,
its half-life, as measured by the disappearance of a shoulder at
approximately 430 nm in the UV/Vis spectrum, depends
strongly on the solvent. For example, the half-life of 3 decreas-
es in the order of acetonitrile>acetone>nitromethane with
1
t = values of 4.6, 3.6, and 0.25 h, respectively (Figure S3,
2
Table S3 in the Supporting Information). Thus, the relative sta-
bility of 3 in solution correlates with the decreasing acidity of
these non-protic solvents. In accordance, 3 cannot exist in
protic solvents, presumably because protons will outcompete
halogen-bonding for the non-coordinated carbonyl oxygen
atoms.
Chem. Eur. J. 2016, 22, 3810 – 3820
www.chemeurj.org
3812
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
becoming symmetrically equivalent and thereby lowering the
Table 2. Mçssbauer parameters derived from the solid-state spectrum in
Figure 4b for co-precipitated 1, 2fac, and 3.[18]
intensity of the formally IR active nas(COO)
.
d
DEQ
[mmsÀ1
FWHM
[mmsÀ1
T
[K]
Assignment Rel. area
[%]
Mçssbauer spectroscopy
[mmsÀ1
]
]
]
1(ClO4)4
light gray
2fac(ClO4)2
black
3(ClO4)2
medium gray 0.61
0.45
0.42
0.18
0.19
0.48
1.63
1.62
2.43
2.14
n.a.
0.36
0.34
0.25
0.59
4.94
2.38
18 high-spin
49
50
2
8
49
43
Mçssbauer spectra of bulk powdered samples containing
150 iron(III)
18 low-spin
150 iron(III)
18 high-spin
150 iron(III)
2(ClO4)2 and 3(ClO4)2 are shown in Figure 4. Consistent with
n.a.
2fac(ClO4)2. This arrangement produces a more symmetrical
electronic configuration compared with the anticipated ligand
conformation of 2mer for which a high-spin (S=5/2) electronic
configuration would be reasonable. A S=5/2 electronic config-
uration for seven-coordinate 3 can be expected and this is
consistent with the broad singlet observed in the spectrum
(Figure 4, right). Broad singlets are commonly observed for
high-spin iron(III) compounds with relatively long paramagnet-
ic relaxation times of the order of a nanosecond.[17] The para-
magnetic relaxation can have contributions from temperature-
independent spin–spin relaxation and temperature-dependent
spin–lattice relaxation. The line width decreases with increas-
ing temperature (Tables 1 and 2), indicating a decrease in the
relaxation time. In Figure 4 (right), a broad sextet with low in-
tensity and with a hyperfine field of about 51 T is visible at
18 K, indicating high-spin iron(III) with a relaxation time on the
order of several nanoseconds. This weak component was not
included in the fit. At 80 K it was not visible. This suggests that
it has transformed to a broad singlet because of the faster
spin–lattice relaxation at 80 K.
Figure 4. Solid-state Mçssbauer spectra at 18 K (bottom) and 150 K (top) of
solid samples containing the reactive high-spin iron(III) complexes. Left:
2mer(ClO4)2 (dark gray). Right: 3(ClO4)2 (medium gray). Both samples contain
significant contamination by 1(ClO4)4 (light gray) and 2fac(ClO4)2 (black).
the IR spectra and the equilibrium in Equation (1), both sam-
ples show a significant and unavoidable fraction of 1(ClO4)4
and 2fac(ClO4)2. The Mçssbauer parameters assigned to
1(ClO4)4,[16] the fac and mer isomers of 2(ClO4)2 (Table 1), and
3(ClO4)2 (Table 2) were extracted from fits to both spectra. The
broad singlet in the spectrum in Figure 4 (left) was assigned
2mer(ClO4)2, which we propose to be high-spin (S=5/2), and
the one doublet is assigned to low-spin (S=1/2) 2fac(ClO4)2.
Thus, Mçssbauer data support the low-spin electronic configu-
ration implied by the bond lengths and the arrangement of
the three p-acceptor pyridine ligands trans to the weak-field
carboxylato and amine donors in the X-ray crystal structure of
Solution spectroscopy: PhIO and NMO adducts
The electron spin resonance (ESR) spectrum of 3 in frozen ace-
tonitrile shows an isotropic signal at giso =4.50 consistent with
a high-spin (S=5/2) electronic configuration persisting in solu-
tion (Figure 5b). If allowed to stand for several hours at room
temperature, a rhombic signal for low-spin 2fac and a low in-
tensity signal for high-spin 2mer emerge; similarly, the UV/Vis
spectrum takes on characteristics of 2fac and 2mer (Figure S3 in
the Supporting Information). This suggests either reversibility
of PhIO binding [Eq. (2)] or that 3 has oxidized an unknown
species in solution, with the concomitant regeneration of 2. By
using UV/Vis and ESR spectroscopy, we have also identified
Table 1. Mçssbauer parameters derived from the solid-state spectrum in
a
transient 4-methyl-morpholine-N-oxide (NMO) adduct,
Figure 4a for co-precipitated species 1, 2fac, and 2mer [18]
.
[FeIII(tpena)NMO]2+ (4). Although rare, the oxidant NMO has
been more often structurally characterized as a ligand in coor-
dination compounds than PhIO.[18] However, none of these co-
ordination compounds contain iron, which in fact has only
rarely been observed to coordinate to aliphatic N-oxides.[19,20]
When a slight excess of NMO is added to 2 in aprotic solutions,
a subtle isobestic conversion takes place within approximately
17 s, forming the weakly yellow–red chromophore 4 (Fig-
ure 5a). Complex 4 is more unstable at room temperature
compared with 3 and decays within seconds to minutes to
d
DEQ
[mmsÀ1
FWHM[a]
T
[K]
Assignment Rel. area
[%]
[mmsÀ1
]
]
[mmsÀ1
]
1(ClO4)4
light gray 0.44
2fac(ClO4)2 0.18
black
2mer(ClO4)2 0.25
dark gray 0.30
0.46
1.71
1.72
2.26
2.16
n.a.
0.33
0.33
0.55
0.44
4.19
3.66
18 high-spin
25
25
41
40
34
35
150 iron(III)
18 low-spin
150 iron(III)
18 high-spin
150 iron(III)
0.14
n.a.
[a] Full width at half maximum.
Chem. Eur. J. 2016, 22, 3810 – 3820
www.chemeurj.org
3813
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
oxidants (Table 3). No diol formation was observed in the
olefin epoxidations. An exclusively sulfoxidized product is ob-
tained for the bifunctional substrate allyl phenyl sulfide, sug-
gesting that sulfoxidation is easier than epoxidation. Of partic-
ular note is the swiftness and yield for the selective oxygena-
tion of thioanisole. However, small amounts of the correspond-
ing sulfones were produced with the diphenyl sulfide, dibenzo-
thiophene, and allyl phenyl sulfide substrates. One explanation
for this difference in selectivity is the difference in the reaction
times required. The initial oxygenation of thioanisole to methyl
phenyl sulfoxide proceeds much more readily than the initial
sulfoxidation of dibenzothiophene to dithiophene sulfoxide;
by contrast, the corresponding oxygenations of methyl phenyl
sulfoxide and dibenzothiophene sulfoxide to their sulfones
proceed at similar rates.[21] Overlap with the second oxygena-
tion can occur during the prolonged time required for the ini-
tial monooxygenation of dibenzothiophene.
Is the metal-based oxidant an iron(III)–OIPh/iron(III)–NMO or
an iron(V)oxo species?
The selective and catalytic oxygenations of electron-rich sub-
strates described above occur through either a direct oxygen
atom transfer or by coupled electron transfer–oxygen transfer
from 3, 4, or an iron(V)oxo species derived from heterolytic
IÀO or NÀO cleavage in these oxidant adducts. The alternative
mechanism involving homolytic FeO–X (X=IPh or N of NMO)
C
cleavage products, an iron(IV)oxo species and/or the sister PhI
C
or N-methylmorphyline , which in turn activate an abundant O
donor, for example, water, can be disregarded for two reasons:
the absence of diol and carbonyl byproducts in the olefin
epoxidations, and more indirectly by the lack of any spec-
troscopic detection of the green iron(IV)oxo derivative,
Figure 5. a) UV/Vis spectra showing the formation of 4 from 0.5 mm 2 and
2 equiv NMO in CH3CN. Recorded over 18 s. b) ESR spectra (120 K in CH3CN)
of (C) 2 (5 mm), (B) 3 (5 mm, generated from mixing 2 and four equivalents
PhIO), (A) 4 (5 mm, generated from mixing 2 and ten equivalents NMO). The
weak signal observed around 3400 gauss is due to the coil of the thermal
probe.
[FeIVO(tpenaH)]2+[16] or PhI radicals. Moreover, as an aggressive
C
hydrogen atom abstractor, [FeIVO(tpenaH)]2+ is a promiscuous
oxidant and the presence of this species would result in many
side-products, which, as stated, is not observed.
form 1. We suggest that in contrast to 3, the self-decay of 4
occurs by oxidation of a CÀH bond in N-methylmorpholine.
The outcome of this reaction will be the production of water.
This will shift the equilibrium in Equation (1) towards 1, and
thereby inhibit subsequent NMO adduct formation owing to
the diminishing concentrations of 2. The decay can be slowed
for several minutes if the solutions are cooled to À408C, ena-
bling characterization by ESR spectroscopy (Figure 5b), which
When acetonitrile solutions of 3 are treated with one equiva-
lent of thioanisole at room temperature, the broad visible ab-
sorption at l=428 nm decays immediately in an exponential
fashion, leading to a 240-fold reduction in its half-life from
4.6 h to 69 s. The reaction rate is even higher when higher
concentrations of thioanisole are used. A second-order rate
constant of k2 =16.7”10À3 sÀ1 mÀ1 was determined (Figure 6a).
ESR spectra show that 3 converts cleanly into 2. During this
time, a transient high-spin species, which we speculate could
be an iron(III)–sulfoxide adduct, [Fe(tpena)OS(CH3)Ph]2+, is de-
tected (Figure S8 in the Supporting Information). This is sup-
ported by the generation of the same ESR signal when 2 is
treated with methyl phenyl sulfoxide.
shows an axial signal with bands at g =4.00 and g =4.83,
?
k
consistent with high-spin 4. The lower signal intensity in the
spectrum of 4 compared with that of 3 is likely due to partial
formation of ESR-silent 1. All attempts to isolate 4 have result-
ed in the re-isolation of 1.
Three direct metal-based oxidants are plausible: the detect-
able [3]2, an undetectable monomeric 3, or an iron(V)oxo spe-
cies derived from heterolytic IÀO cleavage. If it is monomeric 3
or [FeVO(tpena)]2+, it must react immediately after being
formed, eliminating any concentration build-ups. Likewise, to
account for the strong dependence of the rate on the sub-
strate concentrations, [FeVO(tpena)]2+ would need to be
Catalytic oxidation of sulfides and olefins
In acetonitrile solutions, 2 catalyzes the oxidation of the elec-
tron-rich substrates thioanisole, diphenyl sulfide, and dibenzo-
thiophene to the derived sulfoxides, as well as catalyzing con-
version of trans-stilbene, cyclooctene, n-octene, n-decene to
the derived epoxides by using both PhIO and NMO as terminal
Chem. Eur. J. 2016, 22, 3810 – 3820
www.chemeurj.org
3814
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
oxygen transfer process than
Table 3. Oxidative catalysis by 2 using PhIO and NMO as terminal oxidants.[a]
a
oxygen
atom
transfer,
Substrate
Oxidant Oxidant
loading
Catalyst Reaction Product
loading time [h]
Yield
[mol%]
TON[b]
per se.[24,26]
Based on the mechanistic
studies, we conclude that the
evidence is overwhelmingly in
favor of the active metal-based
oxidant being an iron(III)–oxi-
dant adduct and not a derived
high-valent iron–oxo intermedi-
ate. Scheme 1 shows a proposed
catalytic cycle. In this, we sug-
gest that a dimer of 2mer extracts
two equivalents of PhIO from
[PhIO]n. Driving this reaction is
the remarkable stability of [3]2,
which can be regarded as a pro-
tected form of the even more re-
active 3. As described, iron(III)
adducts of the oxygenated prod-
uct are also briefly detected.
Water inhibits the reaction by
hydrating 2 to form the pro-cat-
alyst oxo-bridged complex 1.
[mol equiv] [mol%]
PhIO
NMO
1.1 0.1
2.1 4.0
1
20
sulfoxide/sulfone
sulfoxide/sulfone
80/0
69/0
800
20
PhIO
PhIO
1.1
1.1
1.0
1.0
3
5
78/9
87
70
sulfoxide/sulfone
epoxide/diol
54/16
PhIO
NMO
1.1
2.1
1.0
3.0
16
20
95/0
80/0
95
27
PhIO
NMO
1.1
2.1
1.0
3.0
5
20
epoxide/diol
epoxide/diol
14[c]/0
0/0
14
0
PhIO
NMO
1.1
2.1
1.0
3.0
20
20
10/0
3/0
10
1
PhIO
PhIO
1.1
1.1
1.0
1.0
8
8
epoxide/diol
epoxide/diol
79/0
18/0
79
18
PhIO
1.1
1.0
1
sulfoxide/sulfone[d]
86/7
93
Gas-phase studies
[a] Reactions performed under nitrogen at 508C with 150 mm substrate and 0.15–6 mm 2 in CH3CN. [b] TON=
turnover number. [c] Product distributed as 2% cis-stilbene oxide and 12% trans-stilbene oxide. [d] No epoxide
products detected.
By obtaining a nanoelectrospray
ionization (nESI) mass spectrum
immediately after dissolving
3(ClO4)2 in acetonitrile at À48C,
ions corresponding to [3·ClO4]+
formed in a reversible reaction with 3 and PhI. As detailed by
Hong et al. in a recent communication,[8c] if the formation is re-
versible the rate should be markedly slowed down by the pres-
ence of excess PhI. We observed no significant changes in the
rate of thioanisole oxidation in the presence of 10–100 equiva-
lents of PhI (Figure S13 in the Supporting Information). This im-
plies that an iron–OIPh adduct is the metal-based oxidant.
Confirming the notion that it is a very electrophilic oxidant, re-
actions of 3 with 4-methoxythioanisole are greatly accelerated,
whilst the reaction with 4-(methylthio)benzonitrile is much
slower (Table S4 in the Supporting Information). When correlat-
ing the second-order rate constants (kx) with the electrophilic
constants (s+), a large negative Hammett parameter of 1=
À2.20 is obtained (Figure 6b). The 1 value is almost twice as
large as those commonly observed for iron(IV)oxo species, and
is much more in the range of those measured for iron(III)–
OIPh[8(c)] and iron(IV)–tosylamido[22] adducts (Table S5 in the
Supporting Information). Likewise, when the rate constants are
correlated with the formal oxidation potentials of the sulfides,
a Marcus-type plot is obtained with a large negative slope of
À6.56 VÀ1 (Figure 6c). The value reflects the degree of electron
transfer in the transition state, with the theoretical maximum
(m/z=765), [3·Cl]+ (m/z=701), and [3]2+ (m/z=333) were de-
tected (Figure 7). The stoichiometry of all the PhIO adduct ions
were confirmed by accurate mass measurements performed
on a Fourier transform ion cyclotron resonance mass spectrom-
eter (<1 ppm mass accuracy). In addition, ions originating
from the unavoidable hydrolysis of 2 and 3 corresponding to
[Fe2O(tpena)2]2+ ([1À2H]+, m/z=454), [FeIII(tpena)OH]+ (m/z=
463), [FeII(tpena)]+ (m/z=446), and the ferryl [FeIVO(tpena)]+
(m/z=462) were also formed. ESI-MS of 1(ClO4)4 and 2(ClO4)2
resulted in the formation and detection of the same ions (m/
z=454, 463, 446, and 462).
[Fe2O(tpena)2]2+ and [FeIII(tpena)OH]+ can be expected to be
generated from solutions of 1.[16] Further collision-induced dis-
sociation (CID) experiments performed on [FeIII O(tpena)2]2+ to
2
investigate the source of [FeII(tpena)]+ and [FeIVO(tpena)]+
(Figure S12 in the Supporting Information) show that they are
formed by a gas-phase disproportionation[25] of this ion.
Crucially, they are not derived from the solution state.
CID of [FeIII O(tpena)2]2+ also results in the generation of
2
[FeIII(tpena)OH]+ and most of the degradation products ob-
served in the ESI mass spectra of 1(ClO4)4, including those at
m/z=418 ([Fe(tpena-CO2)]+), 388 ([Fe(tpena-CH2CO2)]+), and
355 ([Fe(tpena-CHPy)]+). CID of the species at m/z=462
([FeIVO(tpena)]+) reveals that these product ions (m/z=418,
388, and 355) can result from its intramolecular gas-phase oxi-
for a pure electron-transfer process being 16.9 VÀ1 [23]
The large
.
slope thus indicates that the oxidations proceed through a re-
action that has more character of a coupled electron transfer–
Chem. Eur. J. 2016, 22, 3810 – 3820
www.chemeurj.org
3815
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Only on rare occasions was the dicationic 3 directly observed.
Isolation and storage of 3 for 10 s (with no intentional deposi-
tion of internal energy into the ion through CID) resulted in
spontaneous coulombic fission to form the ions corresponding
to [FeIVO(tpena)]+ and PhI +. This amounts to a facile gas-
C
phase homolysis of the OÀI bond of the doubly charged pre-
cursor ion [Eq. (4), Figure S11 in the Supporting Information].
In contrast, [3·ClO4)]+ is highly stable with respect to dissocia-
tion when isolated and can be stored for up to 10 s. A CID
mass spectrum of the anion adduct is shown in Figure S11b.
Upon CID, this ion dissociates through the loss of PhIClO4 and
HClO4 to form [FeIVO(tpena)]+ and [Fe(tpena-H)OIPh]+, respec-
tively [Eqs. (5) and (6)]. To summarize, the dicationic [Fe(tpe-
na)OIPh]2+ ion (3) decays significantly faster than its anion ad-
ducts. Also consistent is the fact that [Fe(tpena)OIPh]2+ is de-
tected in comparatively low abundance compared with the
anion adducts. Overall, these data suggest a structurally stabi-
À
lizing interaction between [3]2+ and ClO4 (and adventitious
ClÀ) in the gas phase. It seems unlikely that the anions should
be directly coordinated to the metal center to form an octa-
dentate coordination complex. Thus, we hypothesize that the
anions and [3]2+ interact through halogen-bonding I···OClO3 or
I···Cl interactions. This is in line with the observations of Wang
et al.,[8i–j] who have recently structurally characterized [Mn-
(salen){(OI(Cl)(Ph)}2]. In this complex, the chloride atoms can be
regarded as bonded to the iodine atom of the PhIO unit (aver-
age IÀCl, 2.675 Š).
Figure 6. Kinetics of the reaction of 3 with para-substituted thioanisoles.
a) Time-resolved UV/Vis spectra of the reaction of 3 with four equivalents of
thioanisole. The insert shows the time trace at l=428 nm and its exponen-
tial fitting to extract the pseudo-first-order rate constant. b) Hammett plot
of the logarithm of the second-order rate constants, kx, versus the electro-
philic constants[23] (1=À2.20, R2 =0.994). c) Marcus plot of log(kx/kH) versus
the oxidation potentials of the para-substituted thioanisoles
(slope=À6.56 VÀ1, R2 =0.998).
Figure 7. Left: nESI mass spectrum of 3(ClO4)2 dissolved in CH3CN.
L=tpenaÀ. Right: High-resolution mass spectrum of the “[FeVO(tpena)]2+
”
ion (black line) with its theoretical isotope pattern (thick gray line). The spec-
trum was obtained under conditions most optimal for detecting 32+. By
contrast, most spectra (but not actually this spectrum) show its anion ad-
ducts [3·ClO4]+ and [3·Cl]+ with a greater ion intensity (for example, see Fig-
ure S10 in the Supporting Information).
dative decomposition. The sequential decomposition reactions
With reference to the above analyses of the ESI and CID
mass spectra of 1(ClO4)4, 2(ClO4)2, and 3(ClO4)2, and consider-
ation of the gas- versus solution-phase sources of ions, it was
noteworthy that a low abundance ion with an accurate mass
value of m/z=231.06100 (versus the theoretical value
231.06088 for {Fe(tpena)+O}2+) in Figure 7 (right) appeared
of [FeIII O(tpena)2]2+ and [FeIVO(tpena)]+ are shown in Equa-
2
tions (3) and (3a).
Ions ostensibly corresponding to [FeIVO(tpena)]+ (m/z=462),
+
“[FeVO(tpena)]2+” (m/z=231), and [PhI] (m/z=204) found in
C
the spectrum of 3(ClO4)2 are particularly interesting (Figure 7).
Chem. Eur. J. 2016, 22, 3810 – 3820
www.chemeurj.org
3816
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
strates (MeOH, EtOH, CH4, and (CH3)2S) in attempted
gas-phase ion–molecule reactions (IMRs) point to the
[FeIII(tpenaO)]2+ formulations as we anticipate that
a perferryl species would oxidize these substrates. It
is relevant to note that this IMR technique was suc-
cessfully implemented for the gas-phase oxidation of
alcohols by using a S=2 iron(IV)oxo species.[28]
Conclusion
We have shown that the iron(III) complex of the po-
tentially hexadentate ligand tpenaÀ is capable of acti-
vating the oxidants PhIO and NMO through forma-
tion of seven-coordinated iron(III)–oxidant adducts.
Hexadenticity for ligands supporting catalysts is un-
usual as they are typically regarded as “saturating”,
particularly for first row metal ions. This might be
true for the crystallographically characterized low-
spin 2fac diastereoisomer of [FeIII(tpena)]2+ and that it
is the presumably more reactive high-spin diastereo-
isomer 2mer that can more readily expand its coordi-
nation number and bind the O-donor oxidants. In the
catalytic oxidation of aromatic sulfides and olefins
the unanticipated stability of the iron(III)–OIPh
adduct, [FeIII(tpena)OIPh]2+ (3), compared with the
analogous iron(III)–NMO adduct, [FeIII(tpena)(NMO)]2+
(4), is presumably important for rendering the former
a superior catalytic oxidizing system. Kinetic studies
and a correlation of the rates of the reaction of 3
with a series para-substituted thioanisoles suggest
Scheme 1. A proposed mechanism in which [2]2 solubilizes [PhIO]n to form [3]2. All the
structurally characterized [Fe(tpena)]2+ species (1, [2fac]2, [3]2) occur as dimers or in the
case of the pro-catalyst as an oxo-bridged complex, suggesting that pairs of complexes
work in unison to mobilize [PhIO]n.
only in the spectra of 3(ClO4)2: thus, crucially, it cannot be
“synthesized” in the gas phase, for example by the CID of
any of [3]2+, [3·A]+ (A=ClO4À, ClÀ), [Fe2O(tpena)2]2+, or
[FeIVO(tpena)]+ species. Although this m/z=231 ion is consis-
tent with the formulation for the elusive [FeVO(tpena)]2+ spe-
cies (Figure 8A), alternative structural formulations are plausi-
ble. In these, the ligand is oxygenated and the formula is
better represented by “[FeIII(tpenaO)]2+” (Figure 8B–D). There is
precedence for both in situ N-donors and aromatic oxygena-
tions in iron complexes of a pentadentate N4O ligand related
to tpenaÀ by replacement of a methylpyridyl group with
a methyl group,[20] and peracid complexes have been identi-
fied.[27] Although inconclusive, the lack of reactivity of the
m/z=231 ion towards potentially oxidizable gas-phase sub-
that the direct oxidants in the catalytic reactions are iron(III)–
OIPh species. An elegant duet by pairs of [FeIII(tpena)]2+ com-
plexes effects the solubilization of insoluble PhIO by coordina-
tion and concomitant self-protection to produce higher con-
centrations of the active oxidizing species. Our hypothesis was
that homogeneity would effect more efficient catalytic oxida-
tions by using NMO as the terminal oxidant compared with
those using highly insoluble PhIO, for which reactions are by
necessity heterogeneous. This proved not to be the case and
the decreased solution-state stability of 4 compared with 3
can be correlated with the lower effectivity.
The observations provide an interesting example of halo-
gen-bonding being strong enough in solution to be useful in
assisting in homogeneous catalysis. Evidence for supramolec-
Figure 8. Structural assignments for the m/z=231.06100 ion: a perferryl complex (A) and isobaric isomers of [FeIII(tpenaO)]2+ with variously oxygenated
tpenaÀ ligands (B–D).
Chem. Eur. J. 2016, 22, 3810 – 3820
www.chemeurj.org
3817
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
were isolated. If diethyl ether was diffused into the solution, red–
orange crystals of 2fac(ClO4)2·1.5CH3CN were deposited along with
brown crystals of 1(ClO4)4. Elemental analysis calcd (%) for
2(ClO4)2·H2O (FeC22H26Cl2N5O11): C 39.85, H 3.95, N 10.56.; found: C
39.58, H 3.87, N, 10.31.
ular halogen-bonding interactions between the I atom of the
coordinated PhIO and the electron-donating carbonyl groups
of another cation in dimeric structures or anions was found in
the solid, solution, and gas phases. Noteworthy is the quench-
ing of the otherwise strong carboxylato bands in the IR spectra
of 3(ClO4)2 caused by the symmetry imposed by the halogen-
bonded dimers. It is possible that potentially characteristic
metal–OIPh vibrations have been quenched by the same struc-
tural features.
PhI18O
Small well-polished cubes of sodium (69 mg, 3 mmol) were slowly
added over 30 min to ice-cold H218O (1000 mL) with stirring and
under N2. After this time, finely ground vacuum-dried PhI(OAc)2
(150 mg, 0.47 mmol) was carefully added. Rigorous stirring under
N2 was continued for 3 h. Solids sticking to the sides of the flask
were washed down with 300 mL H218O. Initially, the solid sticks to
the magnet, but as the color changes from white to creamy yellow
a much more homogeneous dispersion was obtained. The disper-
sion was carefully transferred to a centrifuge glass and centrifuged.
The isolated liquor was used to extract the last of the solid from
the reaction flask. The isolated solid was washed with unlabeled
water (until pH neutral), then with acetonitrile, and finally diethyl
ether. The pale-yellow solid was dried under vacuum for 17 h and
stored at À408C.
Through a series of CID and IMR experiments, we have es-
tablished that [FeIVO(tpena)]+ is formed by gas-phase reaction
only; there is no evidence for the presence of this species, and
hence its participation, in the solution-phase catalysis reac-
tions. Notably, an ion that might have been mistakenly as-
signed to the perferryl [FeVO(tpena)]2+ is ascertained to be de-
rived from solutions containing the PhIO adduct 3. The same
ion could not be generated in the gas-phase reaction. The lack
of gas-phase reactivity observed for this ion suggests that it is
more likely to be an iron(III) complex of oxygenated tpenaÀ.
Other 18O-labeled species
Experimental Section
[Fe(tpena)18OIPh](ClO4)2 was prepared similarly to its unlabeled
counterpart[2] by the in situ dehydration of 1(ClO4)4 in dry acetoni-
trile over 30 min, followed by the addition of excess PhI18O, filtra-
tion, and precipitation of the product at À408C by diethyl ether.
Materials
Commercially available reagents were purchased from Sigma–Al-
drich and used without further purification. Acetonitrile and diethyl
ether were dried over activated 3 Š molecular sieves. PhIO,[29]
[Fe2O(tpenaH)2](ClO4)4(H2O)2 (1(ClO4)4),[2] and [Fe(tpena)OIPh](ClO4)2-
(CH3CN)(H2O)0.5 (3(ClO4)4)[2] were prepared as previously described.
[Fe218O(tpenaH)2](ClO4)2 was prepared by recrystallizing 1(ClO4)4
from 95% H218O.
18O-methyl phenyl sulfoxide was prepared by either a direct reac-
tion of thioanisole with 18O-3(ClO4) or by first exposing 3(ClO4)2
(1.8 mg, 2 mmol) in dry acetonitrile (200 mL) to 95% H218O
(250 equiv, 9 mL) for 20 s and then adding 1 equiv neat thioanisole.
The solution was allowed to stand for 5 min at RT. After dilution,
the sample was analyzed by GC-MS, which showed a 91% yield
(Figure S9 in the Supporting Information).
Caution! Perchlorate salts are potentially explosive upon exposure
to excess heat or shock, and should be handled with care and only
in small quantities.
Instrumentation
1H NMR spectra were recorded on a Bruker AVANCE III 400 spec-
trometer. Gas chromatographic analyses were performed on an HP
6890 Series GC system equipped with a Carbowax/20m polyethy-
lene glycol capillary column and flame ionization detector. GC-MS
analyses were performed on a Bruker 451-GC SCION SQ MS
equipped with a ZB WAX 624 column. ATR-IR spectra were record-
ed as neat solids on a PerkinElmer Spectrum Two spectrometer. All
spectra have been ATR- and baseline corrected. UV/Vis spectra
were recorded on an Agilent 8453 spectrophotometer using 1 cm
quartz cuvettes. ESR spectra where recorded on a Bruker EMX Plus
CW spectrometer. Mçssbauer spectra were obtained with conven-
tional constant acceleration spectrometers with sources of 57Co in
rhodium. The spectrometers were calibrated by using a 12.5 mm
foil of a-Fe. Spectra were obtained at 18–150 K by using a closed
cycle helium refrigerator from APD Cryogenics. Powdered samples
were contained in plastic holders (16 mm in diameter). ESI-MS
spectra were recorded on either a nanospray Finnigan hybrid
linear quadrupole ion trap and a 7 T Fourier transform ion cyclo-
tron resonance (FT/ICR) mass spectrometer, or a nanospray orbi-
trap XL mass spectrometer.
Catalysis and kinetic experiments
The substrate (0.9 mmol) and oxidant (1.1 equiv, PhIO or NMO)
were dispersed in dry acetonitrile (4 mL). 2 (0.05–4 mol%) was
added as an aliquot from a 10 mm acetonitrile stock solution. The
reaction mixture was lowered into a pre-heated oil bath at 508C
and stirred under nitrogen for the prescribed amount of time. The
cooled solution was gravity filtered and divided between diethyl
ether (or dichloromethane, 15 mL) and water (2 mL). The organic
phase was evaporated in vacuo to yield the crude product. De-
pending on the boiling point of the substrate, product analysis
was performed by GC by using biphenyl as internal standard or by
1H NMR spectroscopy in CDCl3 (or CD3CN). It was found that the
catalyst could be recovered intact as 1(ClO4)4 (as judged by IR and
ESI-MS) by pooling the aqueous phases from several experiments
and evaporating them in vacuo. The resulting brown oils solidified
upon exposure to a small amount of 96% aqueous ethanol.
Kinetic experiments were carried out at 19–228C. Complex 3
(2000 mL, 1 mm, 2 mmol), taken from a freshly prepared 1 mm stock
of 3(ClO4)2 (1.8–2.4 mg) dissolved in the appropriate amount of
acetonitrile, was transferred to a cuvette. Whilst measuring the UV/
Vis spectrum of the solution, an aliquot from a 40 mm stock (5–
15 mmol) of the appropriate para-substituted phenyl methyl sulfide
was added quickly, and the change in the spectra was monitored
as a function of time. Pseudo-first-order rate constants, kobs, were
[Fe(tpena)](ClO4)2-2(ClO4)2
Diethyl ether was added at À408C into solutions of 1(ClO4)4 dis-
solved in dry acetonitrile (5 mm); the solutions were allowed to
stand for 30 min before cooling. Solid pale or orange powders
Chem. Eur. J. 2016, 22, 3810 – 3820
www.chemeurj.org
3818
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
´
[7] A. Draksharapu, D. Angelone, M. G. Quesne, S. K. Padamati, L. Gymez, R.
Hage, M. Costas, W. R. Browne, S. P. de Visser, Angew. Chem. Int. Ed.
2015, 54, 4357–4361; Angew. Chem. 2015, 127, 4431–4435.
extracted by exponential fitting of the extinction at l=428 nm
using the Agilent ChemStation software (Levenberg–Marquardt al-
gorithm). The second-order rate constant, k, was extracted by
linear regression of a plot of kobs values against the substrate con-
centration. Selected samples were evaporated in vacuo and sub-
jected to 1H NMR spectroscopic analysis in CDCl3 to confirm that
the corresponding sulfoxides had formed.
[8] a) J. P. Collman, A. S. Chien, T. A. Eberspacher, M. Zhong, J. I. Brauman,
Inorg. Chem. 2000, 39, 4625–4629; b) K. P. Bryliakov, E. P. Talsi, Chem.
Eur. J. 2007, 13, 8045–8050; c) S. Hong, B. Wang, M. S. Seo, Y.-M. Lee,
M. J. Kim, H. R. Kim, T. Ogura, R. Garcia-Serres, M. ClØmancey, J.-M.
Latour, W. Nam, Angew. Chem. Int. Ed. 2014, 53, 6388–6392; Angew.
Chem. 2014, 126, 6506–6510; d) K. P. Bryliakov, D. E. Babushkin, E. P.
Talsi, Mendeleev Commun. 2000, 10, 1–3; e) Y. Yang, F. Diederich, J. S.
Valentine, J. Am. Chem. Soc. 1990, 112, 7826–7828; f) K. P. Bryliakov,
D. E. Babushkin, E. P. Talsi, J. Mol. Catal. A 2000, 158, 19–35; g) K. P. Bry-
liakov, E. P. Talsi, Angew. Chem. Int. Ed. 2004, 43, 5228–5230; Angew.
Chem. 2004, 116, 5340–5342; h) C. Wang, T. Kurahashi, H. Fujii, Angew.
Chem. Int. Ed. 2012, 51, 7809–7811; Angew. Chem. 2012, 124, 7929–
7931; i) C. Wang, T. Kurahashi, K. Inomata, M. Hada, H. Fujii, Inorg. Chem.
2013, 52, 9557–9566.
Supporting Information
Details of structure determinations, UV/Vis monitoring of formation
and decays of reactive species, IR spectra of 18O-labeled com-
pounds, and ESR spectra, ESI and CID mass spectra are given along
with details of the kinetic analyses. CCDC 1414517 contains the
supplementary crystallographic data for this paper. These data are
provided free of charge by The Cambridge Crystallographic Data
Centre.
[9] N. Ortega-Villar, A. Y. Guerrero-Estrada, L. PiÇeiro-López, M. C. MuÇoz, M.
Flores-lamo, R. Moreno-Esparza, J. A. Real, V. M. Ugalde-Saldívar, Inorg.
Chem. 2015, 54, 3413–3421.
[10] M. S. Vad, A. Nielsen, A. Lennartson, A. D. Bond, J. E. McGrady, C. J.
McKenzie, Dalton Trans. 2011, 40, 10698–10707.
Acknowledgement
[11] D. P. de Sousa, J. Bigelow, J. Sundberg, L. Que, Jr., C. J. McKenzie, Chem.
Commun. 2015, 51, 2802–2805.
[12] A. Bondi, J. Phys. Chem. 1964, 68, 441–451.
[13] R. Glaser, N. Chen, H. Wu, N. Knotts, M. Kaupp, J. Am. Chem. Soc. 2004,
126, 4412–4419.
This work was supported by the Danish Council for Independ-
ent ResearchjNatural Sciences (grant 12-124985 to C.McK.) and
the University of Southern Denmark. We thank Dr. Simon
Svane for help with ESI-MS. The COST CM1003 action is ac-
knowledged for travel funding.
[14] a) J.-U. Rohde, J.-H. In, M. H. Lim, W. W. Brennessel, M. R. Bukowski, A.
Stubna, E. Münck, W. Nam, L. Que, Jr., Science 2003, 299, 1037–1039;
b) W. J. Song, Y. J. Sun, S. K. Choi, W. Nam, Chem. Eur. J. 2006, 12, 130–
137; c) J. T. Groves, W. J. Kruper, J. Am. Chem. Soc. 1979, 101, 7613–
7615; d) F. F. Pfaff, S. Kundu, M. Risch, S. Pandian, F. Heims, I. Pryjomska-
Ray, P. Haack, R. Metzinger, E. Bill, H. Dau, P. Comba, K. Ray, Angew.
Chem. Int. Ed. 2011, 50, 1711–1715; Angew. Chem. 2011, 123, 1749–
1753; e) S. H. L. Wang, B. S. Mandimutsira, R. Todd, B. Ramdhanie, J. P.
Fox, D. P. Goldberg, J. Am. Chem. Soc. 2004, 126, 18–19; f) H. Kotani, S.
Kaida, T. Ishizuka, M. Sakaguchi, T. Ogura, Y. Shiota, K. Yoshizawa, T.
Kojima, Chem. Sci. 2015, 6, 945–955; g) J. England, J. O. Bigelow, K. M.
Van Heuvelen, E. R. Farquhar, M. Martinho, K. K. Meier, J. R. Frisch, E.
Münck, L. Que, Jr., Chem. Sci. 2014, 5, 1204–1215.
Keywords: epoxidation
· gas-phase chemistry · halogen-
bonding · iodosylbenzene · iron · sulfoxidation
[1] D. Macikenas, E. Skrzypczak-Jankun, J. D. Protasiewicz, J. Am. Chem. Soc.
1999, 121, 7164–7165.
[2] A. Lennartson, C. J. McKenzie, Angew. Chem. Int. Ed. 2012, 51, 6767–
6770; Angew. Chem. 2012, 124, 6871–6874.
[3] a) A. R. McDonald, L. Que, Jr., Coord. Chem. Rev. 2013, 257, 414–428;
b) S. Hong, Y.-M. Lee, W. Shin, S. Fukuzumi, W. Nam, J. Am. Chem. Soc.
2009, 131, 13910–13911; c) A. Mukherjee, M. A. Cranswick, M. Chakra-
barti, T. K. Paine, K. Fujisawa, E. Münck, L. Que, Jr., Inorg. Chem. 2010,
49, 3618–3628; d) Y.-M. Lee, S. Hong, Y. Morimoto, W. Shin, S. Fukuzumi,
W. Nam, J. Am. Chem. Soc. 2010, 132, 10668–10670; e) M. M. Mbughuni,
M. Chakrabarti, J. A. Hayden, E. L. Bominaar, M. P. Hendrich, E. Munck,
J. D. Lipscomb, Proc. Natl. Acad. Sci. USA 2010, 107, 16788–16793; f) S.
Chatterjee, D. Sheet, T. K. Paine, Chem. Commun. 2013, 49, 10251–
10253; g) C.-W. Chiang, S. T. Kleespies, H. D. Stout, K. K. Meier, P.-Y. Li,
E. L. Bominaar, L. Que, Jr., E. Münck, W.-Z. Lee, J. Am. Chem. Soc. 2014,
136, 10846–10849; h) B. Chakraborty, S. Bhunya, A. Paul, T. K. Paine,
Inorg. Chem. 2014, 53, 4899–4912.
[15] Halogen-bonding does not quench carboxylate IR absorption bands per
se. For example, the carboxylate bands are strong in the IR spectra of
2-iodobenzoic and 2-iodosobenzoic acid (IBX), both of which contain
halogen-bonds (Figure S7 in the Supporting Information).
[16] M. S. Vad, A. Lennartson, A. Nielsen, J. Harmer, J. E. McGrady, C. Frand-
sen, S. Morup, C. J. McKenzie, Chem. Commun. 2012, 48, 10880–10882.
[17] a) H. H. Wickman, M. P. Klein, D. A. Shirley, Phys. Rev. 1966, 152, 345–
357; b) M. Blume, J. A. Tjon, Phys. Rev. 1968, 165, 446–456; c) S. Mørup,
in Mçssbauer Spectroscopy and Transition Metal Chemistry: Fundamentals
and Applications (Eds.: P. Gütlich, E. Bill, A. X. Trautwein), Springer,
Berlin, pp. 201–34.
[4] a) K. B. Jensen, C. J. McKenzie, L. P. Nielsen, J. Z. Pedersen, H. M. Svend-
sen, Chem. Commun. 1999, 1313–1314; b) J. Annaraj, Y. Suh, M. S. Seo,
S. O. Kim, W. Nam, Chem. Commun. 2005, 4529–4531; c) J. Cho, S. Jeon,
S. A. Wilson, L. V. Liu, E. A. Kang, J. J. Braymer, M. H. Lim, B. Hedman,
K. O. Hodgson, J. S. Valentine, E. I. Solomon, W. Nam, Nature 2011, 478,
502–505; d) Y.-M. Lee, S. Bang, Y. M. Kim, J. Cho, S. Hong, T. Nomura, T.
[18] The uncertainty of the isomer shifts, the quadrupole splittings, and the
line widths of 1(ClO4)4 and 2fac(ClO4)2 are about 0.02 mmsÀ1. The uncer-
tainties of the parameters of the broad singlets due to 2mer(ClO4)2 and
3(ClO4)2 are somewhat larger. Therefore, the uncertainties of the relative
areas of all components is also relatively large, of the order of 5–10%.
[19] a) A. J. Bailey, M. G. Bhowon, W. P. Griffith, A. G. F. Shoair, A. J. P. White,
D. J. Williams, J. Chem. Soc. Dalton Trans. 1997, 3245–3250; b) R. B.
Brown, M. M. Williamson, C. L. Hill, Inorg. Chem. 1987, 26, 1602–1608.
[20] A. Nielsen, F. B. Larsen, A. D. Bond, C. J. McKenzie, Angew. Chem. Int. Ed.
2006, 45, 1602–1606; Angew. Chem. 2006, 118, 1632–1636.
[21] M. B. Chambers, S. Groysman, D. Villagrµn, D. G. Nocera, Inorg. Chem.
2013, 52, 3159–3169.
[22] a) A. A. Humffray, H. E. Imberger, J. Chem. Soc. Perkin Trans. 1981, 2,
382–387; b) A. K. Vardhaman, S. Sikdar, C. V. Sastri, Ind. J. Chem. Sect. A
2011, 50, 427–431.
´
´
Ogura, O. Troeppner, I. Ivanovic-Burmazovic, R. Sarangi, S. Fukuzumi, W.
Nam, Chem. Sci. 2013, 4, 3917–3923.
[5] a) I. Bernal, I. M. Jensen, K. B. Jensen, C. J. McKenzie, H. Toftlund, J. P. Tu-
chagues, J. Chem. Soc. Dalton Trans. 1995, 3667–3675; b) A. Wada, S.
Ogo, S. Nagatomo, T. Kitagawa, Y. Watanabe, K. Jitsukawa, H. Masuda,
Inorg. Chem. 2002, 41, 616–618; c) S. Taktak, M. Flook, B. M. Foxman, L.
Que, Jr., E. V. Rybak-Akimova, Chem. Commun. 2005, 5301–5303;
d) O. V. Makhlynets, P. Das, S. Taktak, M. Flook, R. Mas-BallestØ, E. V.
Rybak-Akimova, L. Que, Jr., Chem. Eur. J. 2009, 15, 13171–13180; e) A.
Thibon, V. Jollet, C. Ribal, K. Senechal-David, L. Billon, A. B. Sorokin, F.
Banse, Chem. Eur. J. 2012, 18, 2715–2724.
[23] a) H. C. Brown, Y. Okamoto, J. Am. Chem. Soc. 1958, 80, 4979–4987; b) Y.
Goto, T. Matsui, S. Ozaki, Y. Watanabe, S. Fukuzumi, J. Am. Chem. Soc.
1999, 121, 9497–9502.
[6] J. Kim, E. Larka, E. C. Wilkinson, L. Que, Jr., Angew. Chem. Int. Ed. Engl.
1995, 34, 2048–2051; Angew. Chem. 1995, 107, 2191–2194.
Chem. Eur. J. 2016, 22, 3810 – 3820
www.chemeurj.org
3819
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[24] A. K. Vardhaman, P. Barman, S. Kumar, C. V. Sastri, D. Kumar, S. P. de Viss-
er, Angew. Chem. Int. Ed. 2013, 52, 12288–12292; Angew. Chem. 2013,
125, 12514–12518.
[25] L. Eberson, Electron Transfer Reactions in Organic Chemistry (Eds.: K.
Hafner, J.-M. Lehn, C. W. Rees, P. von RaguØ Schleyer, B. M. Trost, R. Zah-
radník) Reactivity and Structure Concepts in Organic Chemistry, Springer
Heidelberg, 1987.
[28] W. A. Donald, C. J. McKenzie, R. A. J. O’Hair, Angew. Chem. Int. Ed. 2011,
50, 8379–8383; Angew. Chem. 2011, 123, 8529–8533.
[29] P. Dauban, L. Sani›re, A. Tarrade, R. H. Dodd, J. Am. Chem. Soc. 2001,
123, 7707–7708.
[26] J. Park, Y. Morimoto, Y.-M. Lee, W. Nam, S. Fukuzumi, J. Am. Chem. Soc.
2011, 133, 5236–5239.
Received: August 7, 2015
Revised: September 22, 2015
Published online on November 24, 2015
[27] J. Nakazawa, S. Terada, M. Yamada, S. Hikichi, J. Am. Chem. Soc. 2013,
135, 6010–6013.
Chem. Eur. J. 2016, 22, 3810 – 3820
www.chemeurj.org
3820
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim