An ultra-stable oxoiron(iv) complex and its blue conjugate base

Article abstract of DOI:10.1039/c3sc52755g

Treatment of [Fe^II(L)](OTf)₂ (4), (where L = 1,4,8-Me₃cyclam-11-CH₂C(O)NMe₂) with iodosylbenzene yielded the corresponding S = 1 oxoiron(iv) complex [Fe ^IV(O)(L)](OTf)₂ (5) in nearly quantitative yield. The remarkably high stability of 5 (t_1/2 ≈ 5 days at 25°C) facilitated its characterization by X-ray crystallography and a raft of spectroscopic techniques. Treatment of 5 with strong base was found to generate a distinct, significantly less stable S = 1 oxoiron(iv) complex, 6 (t _1/2 ～ 1.5 h at 0°C), which could be converted back to 5 by addition of a strong acid; these observations indicate that 5 and 6 represent a conjugate acid-base pair. That 6 can be formulated as [Fe^IV(O)(L-H)] (OTf) was further supported by ESI mass spectrometry, spectroscopic and electrochemical studies, and DFT calculations. The close structural similarity of 5 and 6 provided a unique opportunity to probe the influence of the donor trans to the Fe^IVO unit upon its reactivity in H-atom transfer (HAT) and O-atom transfer (OAT), and 5 was found to display greater reactivity than 6 in both OAT and HAT. While the greater OAT reactivity of 5 is expected on the basis of its higher redox potential, its higher HAT reactivity does not follow the anti-electrophilic trend reported for a series of [Fe^IV(O)(TMC) (X)] complexes (TMC = tetramethylcyclam) and thus appears to be inconsistent with the two-state reactivity rationale that is the prevailing explanation for the relative facility of oxoiron(iv) complexes to undergo HAT. The Royal Society of Chemistry 2014.

Full text of DOI:10.1039/c3sc52755g

Chemical
Science
View Article Online
View Journal | View Issue
EDGE ARTICLE
An ultra-stable oxoiron(IV) complex and its blue
conjugate base†
Cite this: Chem. Sci., 2014, 5, 1204
Jason England,‡a Jennifer O. Bigelow,a Katherine M. Van Heuvelen,§a
a
b
Erik R. Farquhar,{ Marlene Martinho,k Katlyn K. Meier,b Jonathan R. Frisch,a
´
b
a
*
*
Eckard Munck and Lawrence Que, Jr
¨
Treatment of [FeII(L)](OTf)2 (4), (where L ¼ 1,4,8-Me3cyclam-11-CH2C(O)NMe2) with iodosylbenzene
yielded the corresponding S ¼ 1 oxoiron(IV) complex [FeIV(O)(L)](OTf)2 (5) in nearly quantitative yield. The
remarkably high stability of 5 (t1/2 z 5 days at 25 ꢀC) facilitated its characterization by X-ray
crystallography and a raft of spectroscopic techniques. Treatment of 5 with strong base was found to
generate a distinct, significantly less stable S ¼ 1 oxoiron(IV) complex, 6 (t1/2 ꢁ 1.5 h at 0 ꢀC), which could
be converted back to 5 by addition of a strong acid; these observations indicate that 5 and 6 represent a
conjugate acid–base pair. That 6 can be formulated as [FeIV(O)(L–H)](OTf) was further supported by ESI
mass spectrometry, spectroscopic and electrochemical studies, and DFT calculations. The close
structural similarity of 5 and 6 provided a unique opportunity to probe the influence of the donor trans
to the FeIV]O unit upon its reactivity in H-atom transfer (HAT) and O-atom transfer (OAT), and 5 was
found to display greater reactivity than 6 in both OAT and HAT. While the greater OAT reactivity of 5 is
expected on the basis of its higher redox potential, its higher HAT reactivity does not follow the anti-
electrophilic trend reported for a series of [FeIV(O)(TMC)(X)] complexes (TMC ¼ tetramethylcyclam) and
thus appears to be inconsistent with the two-state reactivity rationale that is the prevailing explanation
for the relative facility of oxoiron(IV) complexes to undergo HAT.
Received 2nd October 2013
Accepted 23rd December 2013
DOI: 10.1039/c3sc52755g
www.rsc.org/chemicalscience
a notion that has been experimentally veried for several
enzymes.3–8 Spurred by a desire to understand the fundamental
Introduction
properties of such oxoiron(IV) centers and harness their oxida-
tive properties for synthetic purposes, a relatively large family of
oxoiron(IV) complexes has been synthesized and extensively
characterized, the overwhelming majority of which are inter-
mediate-spin (S ¼ 1).9–11 Out of the six X-ray structures of
mononuclear oxoiron(IV) complexes published thus far12–17 four
are established to have an S ¼ 1 ground state.12–15 Furthermore,
of these four structures, only the tetracarbene complex recently
reported by Meyer and co-workers contains a supporting ligand
with non-N donor atoms.15 None contain O-atom donors, which
are ubiquitous in nature.
Oxygen activating nonheme monoiron enzymes are responsible
for catalyzing a staggeringly diverse array of biologically
important oxidative transformations, despite having similar
active sites with the majority containing a 2-His-1-carboxylate
facial triad structural motif.1,2 Much of this chemistry is
proposed to proceed via formation of oxoiron(IV) intermediates,
aDepartment of Chemistry and Center for Metals in Biocatalysis, University of
Minnesota, 207 Pleasant St. SE, Minneapolis, Minnesota 55455, USA. E-mail:
larryque@umn.edu
bDepartment of Chemistry, Carnegie Mellon University, 4400 5th Ave., Pittsburgh, PA
15213, USA. E-mail: emunck@cmu.edu
Electronic effects on the reactivity of S ¼ 1 oxoiron(IV)
complexes have been investigated. The most signicant
contributions were achieved using oxoiron(IV) complexes con-
taining a solvent occupied labile coordination site, either cis or
trans to the oxo unit, allowing for exchange with various
ligands.18–21 Metathesis studies of this type have been per-
formed for a number of systems, but achieved its apotheosis in
[FeIV(O)(TMC)(CH3CN)]2+ (1-CH3CN, Chart 1).20 This is partly
due to the high stability of this complex, but also because
ligands trans to the oxo unit have been found to exert a greater
inuence than those ligated cis.21 This difference was attributed
to the fact that trans ligands can interact with both s- and p-type
† Electronic supplementary information (ESI) available: Experimental procedures,
X-ray crystallographic data, kinetic data, ESI-MS data, additional XAS and
electrochemical data, and supplementary DFT information. CCDC 964341 and
964342. For ESI and crystallographic data in CIF or other electronic format see
DOI: 10.1039/c3sc52755g
‡ Current address: Max Planck Institute for Chemical Energy Conversion,
¨
Stistrasse 34-36, D-45470, Mulheim an der Ruhr, Germany.
§ Current address: Department of Chemistry, Harvey Mudd College, 301 Platt
Blvd, Claremont, California 91711.
{ Current address: Case Center for Synchrotron Biosciences, Brookhaven
National Laboratory, Upton, NY 11973.
´
k Current address: Aix-Marseille Universite and CNRS, UMR7281, Marseille,
France.
1204 | Chem. Sci., 2014, 5, 1204–1215
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
all cases, presumably due to a combination of steric hindrance
around the binding site and mass action (CH3CN was used as
the solvent). On this basis, it was reasoned that appending
neutral donors to the ligand framework would promote coor-
dination to the iron center. This supposition was tested by
replacing one of the methyl substituents of TMC by a 2-pyr-
idylmethyl group, affording the ligand TMC-py (Chart 1).
This ligand was shown to be capable of supporting the desired
[FeIV(O)(TMC-py)]2+ complex (3),14 in which coordination of the
pyridine donor trans to the oxo group was conrmed by X-ray
crystallography.
Chart 1
orbitals involved in the Fe]O bond but cis ligands cannot.
Herein we extend this approach by introducing a N,N-
Interestingly, whereas the complexes [FeIV(O)(TMC)(X)]+ (1-X, dimethylacetamido donor to yield 1,4,8-Me3cyclam-11-CH2C(O)
where X ¼ monoanion) were found to display increased reac- NMe2 (L) (Chart 1). Although there has been a handful of reports
tivity in O-atom transfer (OAT) with increasing electrophilicity, detailing the copper chemistry of this ligand,28–32 its iron
the reverse trend was found in hydrogen atom transfer (HAT). chemistry has yet to be explored. Gratifyingly, L was found to
The latter counterintuitive ‘anti-electrophilic’ trend was ratio- support an iron(II) complex, [FeII(L)](OTf)2 (4), which could be
nalized using a Two-State Reactivity (TSR) model, which has converted to [FeIV(O)(L)](OTf)2 (5), an oxoiron(IV) complex of
become the prevailing explanation of HAT reactivity in low-spin unprecedented stability. Characterization of the latter by X-ray
oxoiron(IV) systems.22–25 In this DFT-derived model, the transi- crystallography provided the rst example of an oxoiron(IV)
tion state on the S ¼ 2 excited state potential energy surface is at complex with an axial O-atom donor. When treated with strong
signicantly lower energy than its S ¼ 1 ground state analog. base, 5 converts to its conjugate base 6, which exhibits an
Hence, the rate of reaction increases as the contribution of the S unusual blue chromophore. Complexes 5 and 6 represent the
¼ 2 excited state to reactivity increases, which is in turn rst conjugate acid–base pair in oxoiron(IV) chemistry, and
inversely correlated with the energy gap between the two spin provide a means to assess the inuence of trans-axial donor
states. This energy gap decreases as the donor strength of X basicity upon HAT and OAT reactivity using two complexes of
increases. Consequently, the contribution of the S ¼ 2 state similar structure.
increases, which is manifested in enhanced rates of reaction,
thereby yielding the observed ‘anti-electrophilic’ trend in HAT.
Results and discussion
Subsequent studies have hinted at a more complicated
Synthesis and characterization of complexes 4 and 5
reaction surface for S ¼ 1 oxoiron(IV) complexes than these
simple models imply. For example, it has been found that
replacing the N-methyl substituents of TMC by benzyl groups to
give S ¼ 1 [FeIV(O)(TBC)(CH3CN)]2+ (2),26 which might be
expected to sterically inhibit reaction, instead resulted in HAT
and OAT reactivity that is two orders of magnitude greater than
that observed for 1-CH3CN.27 The accompanying DFT study
suggested that both HAT and OAT reactions for 2 proceeded on
the S ¼ 2 potential energy surface because the corresponding
transition states on the S ¼ 1 reaction coordinates were at
prohibitively high energy. Furthermore, the energy gap for the S
¼ 1 ground and S ¼ 2 excited states was found to be smaller for
2 than for 1-CH3CN because of the weakened equatorial ligand
eld associated with TBC due to the greater steric bulk of the
benzyl substituents. Therefore, relatively minor structural
changes can have a large and unexpected impact upon the
reactivity of oxoiron(IV) complexes. This has wide ranging
implications because one must consider the structural inu-
ence of donors added to complexes such as 1-X, in addition to
their basicity. By extension, a direct comparison of the inuence
of donor basicity upon reactivity would require a rigid ligand
framework and the use of structurally near-identical donor
moieties of signicantly different basicity.
Combination of the pentadentate ligand L (ref. 28) with
Fe(OTf)2(CH3CN)2 yielded the expected iron(II) complex
[FeII(L)](OTf)2 (4). The X-ray structure of 4 displayed a 5-coordi-
nate geometry with s ¼ 0.51 (ref. 33) (Fig. 1 le). The cyclam ring
exhibits a trans-I conguration similar to that seen in all the
previously reported X-ray structures of iron(II) TMC
complexes,27,34 and the appended dimethylacetamido donor
coordinates to the iron center via its O-atom. The associated
iron-ligand bond lengths (Table 1) are typical of 5-coordinate
Fig. 1 Thermal ellipsoid drawings of 4 (left) and 5 (right), showing 50%
probability ellipsoids. Hydrogen atoms, counterions, and solvent
molecules have been omitted for clarity. Selected bond lengths are
given in Table 1. Atom color scheme: C, gray; N, blue; O, red; Fe,
During our efforts to test the scope of the TSR model, we
sought to expand the series of 1-X complexes available by dis-
placing the solvent ligand in 1-CH3CN using a number of
neutral donors (e.g. pyridine N-oxide21). This approach failed in magenta. Selected bond lengths are given in Table 1.
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 1204–1215 | 1205

View Article Online
Chemical Science
Edge Article
Table 1 Comparison of selected structural parameters for published S ¼ 1 oxoiron(IV) and related complexes
˚
Bond Lengths (A)
O]Fe–Xaxial bond
Complex
Fe]O
Avg. Fe–Nequatorial
Fe–Xaxial
angle (ꢀ)
4
5
—
2.164
2.0605
2.103
2.04a
2.102
2.03a
2.091
2.133
2.083
2.136
1.957
2.010
2.000(2)
1.9807(12)
2.024
—
1.6579(13)
1.646
1.63
1.667
1.67
1.646(3)
1.640
1.667(3)
1.649
1.639(5)
1.661(3)
175.57(6)
173.79
—
177.31
—
178.9(1)
—
169.77(13)
—
5 (DFT)
5 (EXAFS)
6 (DFT)
6 (EXAFS)
1-CH3CNb
1-CH3CN (DFT)b
3c
1.944
2.058(3)
2.099
2.118(3)
2.135
2.033(8)
2.105(4)
3 (DFT)
7d
178.6(3)
176.65(17)
8e
a
b
c
The resolution of this technique is not sufficient to distinguish between Ncyclam and Oamide scatterers. From ref. 12. This X-ray structure
d
displayed disorder due to co-crystallization of enantiomers and only the data for the major enantiomer is presented herein; from ref. 14. 7 ¼
[FeIV(O)(N4Py)]2+ (N4Py ¼ N,N-bis(2-pyridylmethyl)-N-bis(2-pyridyl)-methylamine); from ref. 13. 8 ¼ [FeIV(O)(LNHC)(EtCN)]2+ (LNHCH4
¼
e
3,9,14,20-tetraaza-1,6,12,17-tetraazoniapenta-cyclohexacosane-1(23),4,6(26),10,12(25),15,17(24),21-octaene); from ref. 15.
1
high-spin iron(II). Furthermore, the H NMR spectrum of this
Complex 5 is the most stable oxoiron(IV) complex published
complex in CD3CN (Fig. S1†) displays ve paramagnetically thus far. It exhibits a half-life (t1/2) at 25 ꢀC in CH3CN solution of
shied methyl resonances, assigned by peak integration, approx. 5 days, which is about one order of magnitude longer than
showing that the appended donor is not displaced by solvent for 1-CH3CN (t1/2 at 25 C ¼ 10 h)18 and 3 (t1/2 at 25 C ¼ 7 h)14
ꢀ
ꢀ
and that the solid-state structure is retained in solution.
and approximately twice that for [FeIV(O)(N4Py)]2+ (7, t1/2 at
Reaction of 4 with excess iodosylbenzene gave a pale brown 25 ꢀC ¼ 60 h, N4Py ¼ N,N-bis(2-pyridylmethyl)-N-bis(2-pyridyl)-
colored solution of a new species 5, exhibiting an electronic methyl-amine).38 This remarkable stability allowed growth of
spectrum with a NIR band (lmax ¼ 810 nm; 3max ¼ 270 Mꢂ1 crystals suitable for X-ray analysis. In the resultant structure
cmꢂ1) characteristic of S ¼ 1 oxoiron(IV) complexes (Fig. 2, (Fig. 1B), the ligand retains the amide O-atom bound penta-
Table 2).35 Previously, features of this type have been assigned as dentate trans-I coordination mode seen in its iron(II)
arising from ligand eld transitions, with the accompanying precursor 4. As a consequence, it is the only existing structure of
well dened ne structure being vibronic in nature.36,37 an oxoiron(IV) complex with a supporting ligand having an
Consistent with assignment of 5 as an oxoiron(IV) complex, its oxygen donor atom.
electrospray mass spectrum exhibited an ion fragment at m/z ¼
As might be expected, oxidation of 4 to 5 is accompanied by a
548.2 (Fig. S12†), with a mass and isotope distribution pattern signicant contraction of the average Fe–Nequatorial bond
in agreement with its formulation as {[FeIV(O)(L)](OTf)}+ lengths by ca. 0.1 A (Table 1). In contrast, the axial Fe–Oamido
˚
˚
(calculated m/z ¼ 548.2).
shortens by only ca. 0.02 A, likely due to the strong trans effect of
the oxo donor atom in the latter complex. Given this minor
change in Fe–Oamido bond length, it is not surprising that the
amide C]O and C–NMe2 distances are virtually identical in 4
˚
(1.267(4) and 1.309(4) A, respectively) and 5 (1.274(2) and
˚
1.308(2) A, respectively). These C]O and C–NMe2 bond lengths
are intermediate between values typical of single and double
bonds,39 a feature characteristic of amide functionality that
stems from the resonance structure implied by canonical forms
A and B (Chart 2).40,41 The importance of canonical form B is
reected by the planarity of the dimethylamido N-atom in both
4 and 5, for which the sum of its bond angles are 360ꢀ, and by
the deviation of the N-atom from the plane dened by the three
˚
adjacent C-atoms is effectively zero (0.003 and 0.002 A in 4 and
5, respectively).
˚
The Fe]O distance of 1.6579(13) A seen for 5 is, within
experimental error, indistinguishable from those found in the
other four published S ¼ 1 oxoiron(IV) crystal structures
Fig. 2 The electronic spectra of 5 (solid line) at 25 ꢀC and 6 (dashed
line) at ꢂ40 ꢀC in CH3CN. Inset: expansion of the visible and near-IR
regions.
(Table 1).12–15 The average Fe–Nequatorial distance of 2.06 A in 5 is
˚
shorter than those of the other TMC-supported complexes
1206 | Chem. Sci., 2014, 5, 1204–1215
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
Table 2 Spectroscopic properties of selected TMC-based oxoiron(IV) complexes
lmax, nm
Complex (3max, Mꢂ1 cmꢂ1
)
n(Fe]O) (cmꢂ1
)
D (cmꢂ1
)
Ax,y,z/gnbna (T)
DEQ (mm sꢂ1
)
h
d (mm sꢂ1
)
E0 (eV) Pre-edge areab
7125.9 30.4(3)
7125.8 23
5
810 (270)
833
27(2)
ꢂ22(2), ꢂ22(2), ꢂ3(2) +1.10(1)
ꢂ21.7, ꢂ21.5, ꢂ4.8a +1.07
ꢂ23.5 (2), ꢂ23.5 (2), 0.51(1)
ꢂ2 (2)
ꢁ0
0.13(1)
5 (DFT)
6
0.18 0.145
352 (1830), 588 (460), 823
806 (270), 1005 (230)
29(2)
0.4 0.13(1)
6 (DFT)
ꢂ21.5, ꢂ21.8, ꢂ4.7a ꢂ0.22
0.78 0.15
0.5 0.17
ꢂ3.0 0.15
0.2 0.16
1-CH3CNc 282 (10000), 824 (400) 839
29
31
30
ꢂ23, ꢂ18, ꢂ3
ꢂ22, ꢂ22, ꢂ2
ꢂ24, ꢂ20, ꢂ4
1.23
0.16
0.60
7124.5 32.8
7125.4 21.7
7125.4 29.1
1-OHd
830 (100), 1060 (110)
387 (3500), 850 (200), 820
1010 (170)
350 (3100), 845 (150), 822
1010 (130)
407 (3600), 850 (130), 814
1050 (110)
834 (260)
460 (1300), 570 (1100), n.d.
860 (230)
n.d.
1-NCSd
1-NCOd
31
29
ꢂ23, ꢂ23, ꢂ5
ꢂ24, ꢂ20, ꢂ5
0.42
0.70
0.1 0.16
0.1 0.17
0.18
7125.3 30.5
7125.3 29.6
d
1 N3
3e
9d
826
1.08
ꢂ0.22
7124.0 f 34 f
7125.1 20.0
35
ꢂ23, ꢂ22, ꢂ5
0
0.19
a
Calculated A-tensor was obtained by taking the experimentally determined Aiso ¼ ꢂ19.4 T and adding the spin-dipolar term obtained from DFT.
b
Weighted pre-edge areas scaled to values expected for pure oxoiron(IV) complex, which assumes that other impurities present do not contribute
signicantly to the pre-edge. c From ref. 12 and 35. d [FeIV(O)(TMC)(X)]+ (1-X), where X ¼ OH, NCS, and N3 and [FeIV(O)(TMCS)]+ (9); data from ref.
35. e From ref. 14. f This work.
˚
atoms themselves from this plane in 5 (0.174 A) is much larger
˚
than in either 3 or 1-CH3CN (0.027 and 0.004 A, respectively).
In any case, the availability of a third structure for an
[FeIV(O)(TMC)(X)]n+ complex with a neutral axial ligand X allows
us to discern a trend that emerges from these data. The
Chart 2
progressive shortening of the axial ligand bond from 3 to 1 to 5
parallels an increase in the lifetime of the oxoiron(IV) unit, sug-
gesting that a more strongly bound neutral axial ligand stabilizes
the FeIV]O unit. In contrast, when the axial ligand X is mono-
anionic, the opposite trend in thermal stability is observed, with
the more basic ligand affording a shorter lived oxoiron(IV)
complex.35 These opposing trends highlight complexities in the
chemistry of nonheme oxoiron(IV) complexes that have yet to be
understood.
˚
˚
1-CH3CN (2.09 A) and 3 (2.08 A). Given that the complexes are
otherwise identical, these differences in average Fe–Nequatorial
distance must originate from the axial ligand. Commensurately,
the most obvious difference among these three TMC oxoiron(IV)
structures is the Fe-axial donor bond length, which increases in
the order of 5 < 1-CH3CN < 3 (Table 1). The short Fe-axial donor
length in 5 is likely due to the greater donor strength of the
dimethylacetamido donor due to some contribution from the
Treatment of oxoiron(IV) complex 5 with strong base
canonical form B (Chart 2). However, the fact that the Fe–Naxial
˚
bond length in 1-CH3CN is 0.06 A shorter than that of 3, even Interestingly, treatment of 5 with excess tetraalkylammonium
though the acetonitrile ligand in the former is a much weaker hydroxide or methoxide instantaneously yielded a new species 6
Lewis base than pyridine in the latter, suggests that steric that is signicantly less stable. At 0 ꢀC, it decayed over a period
factors may also play a role. A comparison of the three TMC- of 1.5 h, representing a difference in thermal stability of about
derived structures shows that the O]Fe–Xaxial bond angle three orders of magnitude. Complex 6 not only exhibited NIR
increasingly deviates from linearity in the order of 1-CH3CN features characteristic of an S ¼ 1 oxoiron(IV) complex, yet
(178.90ꢀ) < 5 (175.57ꢀ) < 3 (169.77ꢀ). The deviations in complexes distinct from those of 5, but also a band centered at 588 nm
3 and 5 may arise not only from the increasing steric bulk of the (3max ¼ 460 Mꢂ1 cmꢂ1) that gives this species its unusual blue
axial donor, but also from the geometric constraints of teth- color (Fig. 2). Notably, there are signicant optical differences
ering a donor to the cyclam ring. Such constraints are likely to between 6 and [FeIV(O)(TMC)(OH)]+ (1-OH, Table 2), for which
cause signicant distortions of the cyclam ring and are most no visible absorption features were observed,35 which suggests
likely the primary origin of the observed variation of the average that hydroxide/alkoxide does not coordinate to the metal center
Fe–Nequatorial distance. Indeed, although the iron center sits in 6 and instead performs an alternate function. Neutralization
close to the mean plane of the NTMC donors in 1-CH3CN, 3 and 5 of the hydroxide used to generate 6, by addition of an equimolar
˚
(it only deviates towards the oxo donor by 0.033 A in 1-CH3CN quantity of perchloric acid, regenerated the parent complex 5 in
˚
and 5, and by 0.071 A in 3), the average deviation of the NTMC approximately 80% yield, suggesting that 5 and 6 simply
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 1204–1215 | 1207

View Article Online
Chemical Science
Edge Article
comprise a Bronsted conjugate acid–base pair. This notion is
strongly supported by the electrospray mass spectrum of 6
(Fig. S13†), which is dominated by an ion fragment with m/z ¼
397.9 and an isotope distribution pattern consistent with its
assignment as {[FeIV(O)(L)]–H}+ (calculated m/z ¼ 398.2), the
product of deprotonation of 5. Given that the most probable
site of deprotonation in 5 is the relatively acidic methylene
linker of the pendant donor arm (Scheme 1), 6 is formulated as
[FeIV(O)(L–H)](OTf). This assignment is supported by the
spectroscopic and DFT studies detailed in subsequent
sections.
Spectroscopic comparison of 5 and 6
Spectroscopic features show 5 and 6 to be closely related, but
distinct oxoiron(IV) complexes. For instance, the 1H NMR
spectra of 5 and 6 (Fig. 3) both show a set of resonances heavily
skewed towards negative chemical shis (the reverse is true of
the iron(II) starting complex 4), which is typical of S ¼ 1 oxoir-
on(IV) complexes.13,42 Most importantly, these two spectra are
composed of a set of resonances distinct from one another and
consistent with the presence of a single C1-symmetric species.
(Full spectral assignments, plus some information on how they
were arrived at, are detailed in the ESI.†) This indicates that
treatment of 5 with base results in complete conversion to 6. On
the basis of peak integration, it is clear that the spectrum of
5 contains paramagnetically shied resonances corresponding
to ve methyl groups, thereby conrming continued penta-
dentate coordination of the supporting ligand. In contrast, only
paramagnetically shied peaks belonging to the three methyl
substituents of the cyclam ring are observed in the spectrum of
6, presumably because the resonances corresponding to the
N,N-dimethylacetamido donor have shied into the 0–10 ppm
region, where there are intense peaks associated with the
solvent, NBu4+, H2O and PhI. The reduced chemical shi of the
methyl substituents in the latter is suggestive of decreased spin
density on the NMe2 unit in 6 relative to 5. Such a scenario
might be expected due to decreased interaction of the NMe2
group with the iron center in the enolate-like structure of the
former, presumably because of the reduced importance of
canonical form B in Chart 2, consistent with the formulation of
6 provided in Scheme 1.
1
Fig. 3 H NMR spectra of 5 (top) and 6 (bottom) in CD3CN at ꢂ40 ꢀC
with methyl resonances indicated by dots. Acquisition time ¼ 0.064 s;
relaxation delay ¼ 0.03 s; line broadening factor ¼ 30 Hz. Compre-
hensive peak assignments are provided in the ESI.†
interactions with parameters typical of S ¼ 1 FeIV]O complexes
(Fig. 4B–D). We have analyzed the data for 5 using the S ¼ 1 spin
Hamiltonian,
2
^
2
^
2
^
^
^
^
H e ¼ D(Sz ꢂ 2/3) + E(Sx ꢂ Sy) + bS$g$B + S$A$I
^
ꢂ gnbnB$I + H Q
,
2
2
2
^
^
^
H Q ¼ (eQVzz/12)[3Iz ꢂ 15/14 + h(Ix ꢂ Iy )]
(1)
where all symbols have their conventional meanings. Within
resolution, the zero-eld splitting (ZFS) tensor, described by the
parameters D and E, was found to be axial (E ¼ 0). The value of
Mossbauer studies corroborate the 1H NMR ndings. The
¨
¨
zero eld Mossbauer spectrum of 5 recorded at 4.2 K exhibits a
doublet representing ca. 93% of the iron with quadrupole
splitting DEQ ¼ 1.10 mm sꢂ1 and isomer shi d ¼ 0.13 mm sꢂ1
(Fig. 4A). Applied magnetic elds, B, elicit magnetic hyperne
Fig. 4 Mossbauer spectra of 5 (A–D) and 6 (E–H) recorded at 4.2 K (A–
¨
C, E–G) and 100 K (D, H) in applied fields as indicated. Red lines are
simulations for 5 (ꢁ93% of Fe) and 6 (ꢁ86% of Fe), obtained by using
eqn (1) and the parameters listed in Table 2. The sample for 6, has a
ꢁ10% likely diferric contaminant with DEQ z 0.51 mm sꢂ1 and d z
Scheme 1
0.52 mm sꢂ1
.
1208 | Chem. Sci., 2014, 5, 1204–1215
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
D was obtained by tting the variable eld data obtained at 4.2 10 cmꢂ1 lower in energy than that for 5 implies a slight weak-
K, for which the splitting along molecular direction i is roughly ening of the Fe]O bond, consistent with the expected greater
linear in (Ai/D), and taking into account that the 100 K/8.0 T donor strength of the anionic enolate donor in the former
spectrum is essentially independent of D. The resulting value relative to the neutral N,N-dimethylacetamide donor in the
D ¼ 27 ꢃ 2 cmꢂ1 for 5 is typical of oxoiron(IV) complexes of the latter. Of particular note is that the n(Fe]O) in 6 is not only
TMC ligand.35
resonance enhanced by excitation at 407 nm, as seen for other
¨
Obtaining Mossbauer spectra of 6 was not straightforward TMC-ligated oxoiron(IV) complexes, but also at 568 nm. (In fact,
due to bleaching of the blue chromophore of 6 and reformation the greater enhancement in intensity is observed using light
¨
of 5 (established by Mossbauer spectroscopy) upon freezing of with the latter wavelength.) This observation indicates that the
acetonitrile solutions of 6. This problem was circumvented by electronic transition in 6 centered at 588 nm, and possibly also
using toluene solutions of tetraoctylammonium hydroxide43 in that at 352 nm, involve occupation of orbitals associated with
place of tetrabutylammonium hydroxide in order to minimize the Fe]O unit (vide infra). The greater enhancement of the
the amount of water present in the system. This necessitated n(Fe]O) mode observed with 568 nm excitation likely arises
¨
use of the more lipophilic solvent butyronitrile. The Mossbauer from the photoinstability of the complex upon near-UV irradi-
spectrum of an unbleached sample of 6 conrmed formation of ation, as noted previously for other FeIV ¼ O complexes.14,35
a new iron(IV) complex in 86% yield with #5% residual 5 and a
The Fe K-edge X-ray absorption spectra of 5 and 6 exhibit
ferric impurity making up the balance. In zero eld we observed virtually superimposable normalized XANES regions (Fig. 6A),
a doublet with DEQ ¼ 0.51 mm sꢂ1 and d ¼ 0.12 mm sꢂ1 with respective edge energies of 7125.91 and 7125.83 eV and
(Fig. 4E). While the d-values of 5 and 6 are the same, their DEQ pre-edge energies of 7114.18(1) and 7114.14(1) eV (Table S3†),
values are quite distinct. The DEQ observed for 6 is in the range suggesting that the differences in iron(IV) environment inferred
associated with [FeIV(O)(TMC)(X)]+ (1-X) complexes with basic from Mossbauer spectroscopy do not signicantly impact the
¨
axial ligands such as SCNꢂ, NCOꢂ and N3ꢂ (1-NCS, 1-NCO and transitions comprising both the rising K-edge and the near-edge
1-N3, respectively; Table 2),35 but is signicantly larger than that regions of the XAS spectrum. Both the edge and pre-edge
reported for 1-OH, which supports the notion that formation of
6 is unlikely to be due to the displacement of the pendant donor
arm of 5. From the analysis of the applied eld spectra we
obtained the parameters listed in Table 2. As for 5, we used
E/D ¼ 0 for 6.
The resonance Raman spectra of complexes 5 and 6 exhibit
n(Fe]O) of 833 and 823 cmꢂ1 (Fig. 5), respectively. In accor-
dance with expectations from Hooke's law for a diatomic
oscillator, these features downshi by 35 and 36 cmꢂ1 upon
18O-labelling of 5 and 6, respectively. That the n(Fe]O) of 6 is
Fig. 6 (A) Comparison of Fe K-edge XANES spectra and pre-edge
Fig.
5 Top: resonance Raman spectra of 5 (black line) and its features (inset) of 5 (solid line) and 6 (dashed line). (B) Fits to the
18O-labeled isotopomer (red line), recorded in frozen CH3CN solution Fourier transforms of the Fe K-edge EXAFS data (k3c(k)) and unfil-
with samples mounted on a brass cold finger (lex ¼ 407 nm, power ¼ tered EXAFS spectra (k3c(k)), insets, for 5 and 6. Experimental data is
20 mW). Bottom: resonance Raman spectra of 6 (black line) and its shown with dotted lines, while fits are shown with solid lines. Fourier
18O-labeled isotopomer (red line) obtained in CH3CN solutions at transformation ranges are as follows: k ¼ 2.0–15 A˚ꢂ1 (5) and k ¼ 2.0–
ꢂ20 ꢀC (lex ¼ 568.2 nm, power ¼ 90 mW). Asterisks designate features 14.0 A˚ꢂ1 (6). Fit parameters are shown in bold italics in Table S5 and
from the CH3CN solvent.
S6 of the ESI.†
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 1204–1215 | 1209

View Article Online
Chemical Science
Edge Article
energies are in line with values obtained previously for other reected by respective changes in the C–C(O)NMe2, C]O, and
oxoiron(IV) complexes (Table 2).35,44 Indeed, the only notable C–NMe2 bonds, from 1.515, 1.296 and 1.328 A in 5 to 1.369,
˚
˚
difference between the near-edge spectra of these two iron(IV) 1.360 and 1.375 A in 6. The negative charge on the dimethyla-
species is that the area of the pre-edge feature is moderately cetamido oxygen donor atom in 6 leads to a shortening of the
˚
˚
larger for 5 than that of 6 (30.4 and 23 units, respectively). The Fe–Oamide bond by ca. 0.08 A to 1.944 A and an increase in the
˚
˚
intensity of this feature, which originates from 1s-to-3d transi- Fe]O distance by ca. 0.02 A to 1.667 A, but crucially does not
tions at the iron center, generally reects the extent to which perturb the average Fe–NTMC distance, which differs by only
˚
distortion from centrosymmetry facilitates 4p mixing into the 0.001 A between 5 and 6. Hence, any differences between the
3d orbitals and increases the probability of transition. Hence, reactivity and spectroscopic properties of these two species can
the smaller pre-edge area of 6 reects a more centrosymmetric be attributed solely to the change in the axial donor trans to the
coordination environment, stemming from stronger coordina- FeIV]O. It should be noted that the changes in the Fe-donor
tion by the pendant donor.
atom bond lengths predicted by DFT are consistent with the
In good agreement with its X-ray structure, EXAFS analysis of structural parameters obtained by EXAFS (Tables 1 and S4–S6†).
˚
¨
The calculated and experimental Mossbauer parameters for
5 yielded a best t with 5 O/N scatterers at 2.04 A, attributable to
the 5 donor atoms of the ligand, and a further O/N scattering 5 agree very well. In particular, the calculated spin dipolar part
˚
atom at 1.63 A that corresponds to the Fe]O unit (Fig. 6B, of the A-tensor and the electric eld gradient (EFG) tensor have
Tables S4 and S5†). A similar best-t EXAFS analysis is obtained their largest component along z, in agreement with the
IV
˚
¨
for 6, consisting of 5 O/N scatterers at 2.03 A and a single O/N Mossbauer analysis of eqn (1) (the zero-eld splitting of Fe ]O
˚
scatterer at 1.67 A (Table S6†). This data serves to further complexes is generally axial around the Fe]O bond, the chosen
conrm that the Fe]O moiety is retained in 6, with the slight z axis.). The calculated A-tensor of 6 also agrees well with the
elongation in its length due to the increased basicity of the experimental data. However, DEQ differs from the experimental
trans-oriented donor. Attempts to resolve a shortened Fe–Oamide data by 0.7 mm sꢂ1. Close inspection of the output DFT les of
distance in the EXAFS of 5 and 6 by splitting the principal O/N 5 and 6 did not reveal any particular feature that allowed us to
shell were unsuccessful, on the basis of a poorer overall good- identify the source for the difference in the calculated DEQ's of
ness of t. However, it should be noted that the Debye–Waller 5 and 6. We note that for 6 the calculated spin-dipolar part of A
factor associated with this shell in 6 is considerably larger than for 6 agrees well with the data, and this part is proportional to
that of its precursor 5, which presumably reects greater static the valence part of the EFG. We thus would expect the valence
disorder in the range of bond lengths comprising this shell and part of the EFG to be well represented by the calculations, and
is consistent with a decreased Fe–Oamide distance in 6.
the deviations might be rooted in covalency effects. We note
Taken together, the spectroscopic data presented here show that calculated DEQ values oen differ substantially from
that 5 and 6 t well into the series of [FeIV(O)(TMC)(X)] experimental data, for reasons not well understood, and thus
complexes that have been characterized over the past decade the disagreement for 6 is thus not surprising. In fact the good
(Table 2).35 Complex 5 represents only the third complex in the agreement for 5 may be a bit fortuitous.
series where X is a neutral ligand. Complex 6 is the conjugate
In order to investigate the origin of the 588 nm feature in the
base of 5, where one of the relatively acidic protons on the UV-visible spectrum of 6, spin-unrestricted time-dependent
methylene linker of the pendant donor arm is abstracted by the DFT (TD-DFT) studies of 5 and 6 were conducted using the PBE0
addition of base. This step is postulated to convert the neutral functional. These calculations gave rise to the frontier molec-
N,N-dimethylacetamide ligand into an enolate anion as shown ular orbital energy diagrams for 5 and 6 shown in Fig. 7, which
in Scheme 1. When the spectroscopic properties of 6 are can be used to understand the spectroscopic differences
compared with those of the other complexes in Table 2, the between 5 and 6. In the case of 5, the spin-up (a) and spin-down
n(Fe]O) and DEQ values of 6 closely resemble to those of 1-NCS (b) highest occupied molecular orbitals (HOMO) are primarily
and 1-NCO, while the XAS pre-edge area of 6 approaches that of (Fe dxz + O px) in character and the occupied ligand-derived MOs
1-OH. These comparisons suggest that the axial donor of 6 has a are lower in energy. Upon deprotonation of the methylene
basicity intermediate between NCSꢂ/NCOꢂ and hydroxide.
group of the pentadentate ligand L, the ligand-based MOs are
destabilized and become the a and b HOMOs in 6 (Fig. S18†).
The TD-DFT generated UV-visible spectrum of 5 is shown in
Fig. 8. Examination of the electron density difference map
DFT studies
DFT calculations using the B3LYP/6-311g functional and basis (EDDM) corresponding to state i reveals that the near-IR band at
set were carried out to gain further insight into the differences 810 nm can be assigned as a (Fe dyz)-to-(Fe dx2–y2) ligand–eld
in electronic structure between 5 and 6. Given the tendency of transition, while state ii arises from an oxo-to-iron charge
DFT to overestimate metal–ligand bond lengths, the agreement transfer transition. This assignment is corroborated by the
between the DFT geometry optimized and X-ray structures of 5 resonance Raman data presented in Fig. 5, which demonstrates
is excellent, with the calculated Fe]O and the average Fe– that laser excitation into the near-UV feature results in reso-
˚
N/OTMC distances being ca. 0.012 and 0.04 A, respectively, nance-enhancement of the n(Fe]O) vibration. It should be
longer than the experimental values (Fig. S16,† Table 1). DFT noted that electronic spectra of this type, incorporating ligand
analysis of the proposed structure of 6 yields a geometry opti- eld transitions at ꢁ800 nm and an intense charge transfer
mized structure in which the expected enolate-like character is band in the near-UV spectral region, are characteristic of
1210 | Chem. Sci., 2014, 5, 1204–1215
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
Fig. 7 Top: spin-down molecular orbitals of 5 and 6 generated using
the single-parameter PBE0 functional. For comparison, orbital ener-
gies are plotted relative to the spin-down LUMO. The electronic
transition that gives rise to the 588 nm optical feature of 6 is marked
with an arrow. Bottom: isosurface plots of the spin-up HOMOs of 5
(left) and 6 (right).
Fig. 8 Top: calculated (PBE0) and experimental electronic absorption
spectra of 5 and 6. The intensity of the simulated spectrum of 6 was
scaled by a factor of 2/3 to facilitate comparison with experimental
data. Bottom: electron density difference maps (EDDMs) illustrating
the loss (grey) and gain (green) of electron density.
[FeIV(O)(TMC)(X)]+ complexes.35 Computational results indicate
that acetamide-to-Fe]O transitions should occur at high
energy (>33 000 cmꢂ1), outside of our experimental detection
limit.
in the near-IR region (Table 2).36,37 Additionally, the pseudoha-
lide-ligated oxoiron(IV) complexes 1-NCS, 1-OCN and 1-N3 display
moderately intense charge transfer bands in the near-UV that
were assigned as being pseudohalide-to-Fe]O in origin.35 The
HOMOs of these pseudohalides are at lower energy than those of
the enolate-like donor in 6, so these charge transfer transitions
occur at higher energy. It should be noted that an oxoiron(IV)
complex with a chromophore centered in the visible region is not
without precedent, with the [FeIV(O)(TMCS)]+ (9) complex dis-
playing two such bands at 460 and 570 nm that also render this
complex blue and are presumed to originate from thiolate-to-
Fe]O transitions.45
Consistent with experimental data, the calculated UV-visible
spectrum of 6 exhibits multiple electronic transitions in the
visible region that are absent in 5. As is typical of S ¼ 1 oxoir-
on(IV) complexes, the lowest energy feature in the near IR region
(state iii) is predominantly ligand eld in character. Likewise,
the intense band in the near-UV region of 6 (state vii) may be
assigned as an oxo-to-iron charge transfer transition. However,
in contrast to the analogous spectral features observed in 5, the
dimethylacetamido moiety contributes to each of these elec-
tronic transitions by acting as an electron donor. Similarly, the
EDDM associated with state v indicates that the unique 588 nm
feature in 6 is a ligand-to-iron charge transfer transition in
which electron density from the enolate-like moiety is donated
to the vacant b(Fedyz + Opy)* orbital. The destabilization of the
Electrochemistry and reactivity
ligand-based orbitals results in a red-shi of acetamide-to-iron The addition of a conjugate acid–base pair to the family of TMC-
transitions relative to 5 and gives rise to the novel electronic ligated nonheme oxoiron(IV) complexes provides a unique
absorption features observed for 6.
opportunity to compare their redox behavior with that of other
Taken as a whole, the assignment of the features of the members of the family. Previous studies of four [FeIV(O)(TMC)X]
electronic spectra of 5 and 6 are analogous to those previously complexes (X ¼ NCCH3, O2CCF3, N3, SR) by cyclic voltammetry
described for TMC-supported complexes 1-X (X ¼ CH3CN, in MeCN showed an irreversible reduction peak for each of
SCNꢂ, NCOꢂ and N3ꢂ), with intense oxo-to-iron charge transfer these complexes in the range of ꢂ0.3 – ꢂ1 V vs. the ferrocenium/
transitions located in the UV region and ligand eld transitions ferrocene redox couple (Fc+/o).20 The relative order of Ep,c values
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 1204–1215 | 1211

View Article Online
Chemical Science
Edge Article
obtained was rationalized by the electron donating properties of
The oxidative reactivity of 3, 5, and 6 was compared with that
the axial donor, X. Not surprisingly, potentials spanning the of the [FeIV(O)(TMC)(X)] series (X ¼ NCMe, O2CCF3, N3, and SR)
same range were observed in cyclic voltammetric measure- studied by Sastri et al. at 0 C.20 For 3 and 5, the reaction with
ꢀ
ments of 3 (Fig. S19†), 5 and 6 (Fig. 9) and provide additional PPh3 produced OPPh3 in 93% and 90% yield, while the reaction
points that can be used to make correlations with oxidative of 6 with PMePh2 produced OPMePh2 in 83% yield. All iron(IV)-
reactivity data.
oxos reacted with 1,4-cyclohexadiene to produce benzene in 60–
Cyclic voltammetry of 3 in MeCN at room temperature 80% yield. The second order rate constant, k2, of 3 for oxygen
revealed a reduction peak at ꢂ0.48 V vs. Fc+/o, similar to that of atom transfer (OAT) to PPh3 was found to be 3.8 Mꢂ1 sꢂ1
,
[FeIV(O)(TMC)(O2CCF3)]+. In the case of 5, however, two comparable to that for 1-CH3CN. However, the OAT value for 5
comparably sized reduction peaks were observed: one at was 0.19 Mꢂ1 sꢂ1, more than an order of magnitude smaller
ꢂ0.63 V and a second at ꢂ1.16 V (Fig. 9 bottom, red line). At than for 1-CH3CN, and that for 6 was too slow to be distin-
ꢀ
ꢂ40 C, 5 (Fig. 9, top) can be compared to 6 by the addition of guished from its self-decay. In order to obtain a more quanti-
Bu4N(OH) to a solution of 5 (Fig. 9 top, blue line), which tative reactivity comparison of 5 and 6, we carried out OAT
resulted in an increase in the current of the lower potential peak reactivity studies with PMePh2 and found 6 to react about 20-
and the suppression of the higher potential peak. On the other fold more slowly than 5 (Table 3). This order of decreasing OAT
hand, the addition of pyridinium triate increased the intensity rates appears to correlate with the increasing basicity of the
of the current associated with the Ep,c peak at ꢂ0.63 V and axial ligand, as noted previously for the [FeIV(O)(TMC)(X)] series
suppressed the Ep,c peak at ꢂ1.16 V (Fig. 9 bottom). To ratio- studied by Sastri et al.20 This trend is visually demonstrated by a
nalize the observed behavior above, we attribute the feature at plot of log k2 vs. Ep,c (Fig. 10 top) where the data points for 3, 5,
ꢂ0.63 V to the reduction of 5 and the feature at ꢂ1.16 V to the and 6 lie close to the trendline dened only by data points
reduction of 6, which is supported by the effects of adding base associated with the [FeIV(O)(TMC)(X)] series studied by Sastri
or acid. The appearance of the second reduction wave in the et al., reecting the decreasing electrophilicity of the FeIV]O
cyclic voltammetry of 5 suggests that 6 can be generated spon- unit with an increasingly basic axial ligand. A similar linear
taneously. We propose that the reduction of 5 produces a highly correlation between OAT rates and redox potential noted for a
basic FeIII–Oꢂ moiety, which in turn deprotonates some series of complexes supported by pentadentate N5 ligands
remaining 5 in solution to form 6. Protonation of the one- reinforces this conclusion.46,47
electron-reduced FeIII–Oꢂ species (formed from the reduction
In contrast, 5 and 6 exhibit a much smaller difference in the
of 5) by addition of pyridinium triate prevents 6 from being rates of the H-atom transfer (HAT) reactions. The k2 value for
formed, thus decreasing the current for the second peak the reaction of 5 with 1,4-cyclohexadiene (CHD) is only twofold
at ꢂ1.16 V.
larger than for 6, values that are in turn an order of magnitude
smaller than those of 1-CH3CN and 3 (Table 3). However, unlike
for OAT, the HAT rates of all [FeIV(O)(TMC)(X)] complexes do not
display a consistent trend with redox potential (Fig. 10, bottom).
In the subset of [FeIV(O)(TMC)(X)] complexes studied by Sastri
et al.,20 an anti-electrophilic trend was observed, where the HAT
rate constant decreased with increasing redox potential. A
similar trend was found in another study of FeIV(O) complexes
of a different tetragonal N4 ligand L8py2 (L8py2 ¼ N,N'-bis-
(2-pyridylmethyl)-1,5-diazacyclooctane) where the axial ligand
trans to the oxo group was pyridine N-oxide;15 variation of the
4-substituent on the pyridine N-oxide showed higher HAT
reactivity for the more electron donating substituents. This
counter-intuitive trend has been accounted for by Shaik and
Table 3 Rate constants and Ep,c values vs. Fc+/Fc in CH3CN for tet-
raalkylcyclam-ligated oxoiron(IV) complexes
k2 [Mꢂ1 sꢂ1] at 0 ꢀ
C
Ep,c (V vs. Fc+/Fc)
Complex t1/2
CHDa PPh3
PMePh2 RTe
ꢂ40 ꢀ
C
1-CH3CNb 10 h (25 ꢀC) 0.12
5.9
—
—
5.9(3)
ꢂ0.32
ꢂ0.48
—
—
3
5
6
7c h (25 ꢀC) 0.22(1) 3.8
5 days (25 ꢀC) 0.037(2) 0.19
ꢂ0.63 ꢂ0.74
Fig. 9 Cyclic voltammograms of 5 in CH3CN, scan rate: 100 mV sꢂ1
.
ꢁ1.5 h (0 ꢀC) 0.016(1) <0.004d 0.32(2) ꢂ1.16 ꢂ1.23
1,4-Cyclohexadiene. b From ref. 18. c From ref. 14. d Indistinguishable
Top: before (red) and after (blue) the addition of 1.4 eq. NBu4(OH) at
ꢂ40 ꢀC. Bottom: before (red) and after the addition of 0.1 eq. (blue)
and 0.2 eq. pyridinium triflate (magenta) at room temperature.
a
from self-decay. e Room temperature.
1212 | Chem. Sci., 2014, 5, 1204–1215
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
to the TMC ligand framework overcame any inherent thermody-
namic barriers associated with coordination of this neutral donor
and resulted in pentadentate coordination of L in complex 4.
Furthermore, oxidation of 4 with PhIO yielded the ultra-stable
oxoiron(IV) complex 5, with a half-life of a 5 days at 25 ꢀC that is
approximately an order of magnitude longer than that of
1-CH3CN (t1/2 ¼ 10 h)18 and the longest thus far reported for any
oxoiron(IV) complex. This stability allowed isolation of crystals
suitable for X-ray analysis, with the resultant structure being the
rst of an oxoiron(IV) complex supported by a ligand containing
an oxygen atom donor and only the h with a S ¼ 1 spin-
state.12–15 Interestingly, treatment of 5 with strong base leads to
deprotonation of the methylene group that links the appended
donor to the cyclam ring and formation of a second, much less
stable, oxoiron(IV) complex [FeIV(O)(L–H)](OTf) (6). The latter
converts to the former in high yield upon treatment with acid,
making 5 and 6 the rst example of a conjugate acid–base pair in
non-heme oxoiron(IV) chemistry. A particularly notable spectro-
scopic feature of the enolate-like complex 6 is its blue color, which
on the basis of TD-DFT calculations is assigned as a LMCT
transition originating from the high energy lled MO of the
Fig. 10 Correlation of rate constants for OAT (PPh3, top) and HAT
(1,4-cyclohexadiene, bottom) reactions with Ep,c values of various
[FeIV(O)(TMC)(X)] complexes. Filled circles represent data obtained in
this study. Squares represent data from ref. 20 and the solid lines are
trend lines created using only this data. The dashed line is the trend
line created using the data collected for the complexes in this study
plus that for 1-CH3CN.
enolate-like pendant acetamide donor to the vacant b(Fe dyz
+
Opy)* orbital. This assignment was corroborated by resonance
Raman spectroscopy, wherein resonance enhancement of the
n(Fe]O) feature was observed upon excitation at wavelengths
near the LMCT band at 588 nm. (In contrast, near UV excitation
was required to observe the n(Fe]O) features of other nonheme
oxoiron(IV) complexes.)
co-workers with a Two-State Reactivity (TSR) model.22,23 This
model recognizes the proximity of an excited S ¼ 2 state that is
much more reactive for HAT than the ground S ¼ 1 state. As C–H
bond cleavage progresses, the reaction coordinate will cross over
from the ground triplet surface to the more reactive quintet
surface, with an earlier transition being more favorable for HAT.
Thus the reactivity of an S ¼ 1 FeIV]O unit is governed by the size
of the triplet-quintet gap and spin-orbit coupling; when the
triplet-quintet gap decreases as the axial ligand becomes more
basic, the spin-orbit coupling increases. For this 1-X subset, the
parent complex with X ¼ CH3CN becomes more reactive when X
is replaced by an anionic ligand (X ¼ O2CCF3, N3, SR); in the case
of the complex with X ¼ SR (9), HAT from 9,10-dihydroanthracene
was 40-fold faster than for 1. Contrary to our expectations based
on the Sastri results, both 5 and 6 were in fact less reactive than 1,
even though both have lower redox potentials than 1. The axial
enolate ligand of 6 should be of comparable basicity as thiolate
based on their Ep,c values, but the HAT rate of 6 with CHD as
substrate is in fact 7-fold slower than for 1 (Table 3) and is esti-
mated to be two orders of magnitude slower than that for 9. From
the above discussion, as well as studies of HAT reactivity on two
other sets of oxoiron(IV) complexes,46,47 it is clear that HAT reac-
tivity can be affected by other factors besides TSR. This provides
impetus for developing a new theoretical model that can account
for the wide difference observed in the reactivity of iron(IV)-oxo
complexes, especially those with similar ligand sets.
The chemical inertness inferred from the remarkably long
half-life of 5 was conrmed by observation of sluggish inter-
molecular OAT and HAT reactivity, with k2 values approximately
an order of magnitude slower than 1-CH3CN in reactions with
PPh3 and CHD. Interestingly, the signicantly less stable
complex 6 is an even less facile oxidant than 5 in both OAT and
HAT. It is clear that this series of complexes defy the anti-elec-
trophilic trend in HAT reactivity established for the closely
related 1-X series of complexes.20 Given that the only structural
parameters that differ signicantly between the DFT optimized
geometries of 5 and 6 are those associated with the axial donor,
it seems likely that changes in the basicity of the axial donor
alone are not enough to yield enhanced reactivity in HAT. This
nding adds another facet to the increasingly complex reaction
landscape of S ¼ 1 oxoiron(IV) complexes, and serves to high-
light the shortcomings of our mechanistic understanding. It is
clear that further systematic studies are required in order to
unify all mechanistic observations into a single model that will
allow qualitative prediction of reactivity of S ¼ 1 oxoiron(IV)
complexes as a function of supporting ligand.
Acknowledgements
This work was supported by grants from the National Institutes
of Health (GM-33162 to L.Q. and EB-001475 to E.M.) and the
National Science Foundation (CHE1058248 to L.Q. and CHE
1305111 to E.M.) and a postdoctoral fellowship from the
Conclusions
In this study, we have demonstrated that, in an analogous fashion National Institutes of Health (GM-093479 to K.M.V.H.). XAS
to the TMC-py ligand, appending a N,N-dimethylacetamide donor data were collected on beamlines 7-3 and 9-3 at the Stanford
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 1204–1215 | 1213

View Article Online
Chemical Science
Edge Article
Synchrotron Radiation Laboratory (SSRL), a national user facility 17 D. C. Lacy, R. Gupta, K. L. Stone, J. Greaves, J. W. Ziller,
operated by Stanford University on behalf of the U.S. Department
of Energy, Office of Basic Energy Sciences. The SSRL Structural
M. P. Hendrich and A. S. Borovik, J. Am. Chem. Soc., 2010,
132, 12188–12190.
Molecular Biology Program is supported by the Department of 18 J.-U. Rohde and L. Que Jr, Angew. Chem., Int. Ed., 2005, 44,
Energy, Office of Biological and Environmental Research, and by 2255–2258.
the National Institutes of Health, National Institute of General 19 J.-U. Rohde, A. Stubna, E. L. Bominaar, E. Munck, W. Nam
Medical Sciences (including P41GM103393). We thank Drs and L. Que Jr, Inorg. Chem., 2006, 45, 6435–6445.
Matthew Latimer, Allyson Aranda, and Erik Nelson for their 20 C. V. Sastri, J. Lee, K. Oh, Y. J. Lee, J. Lee, T. A. Jackson,
¨
excellent technical support of our XAS experiments. We are
grateful to Prof. Thomas C. Brunold at the University of Wis-
consin–Madison for the use of his computer clusters and Prof.
K. Ray, H. Hirao, W. Shin, J. A. Halfen, J. Kim, L. Que Jr,
S. Shaik and W. Nam, Proc. Natl. Acad. Sci. U. S. A., 2007,
104, 19181.
Philippe Buhlmann for discussions on the electrochemical 21 Y. Zhou, X. Shan, R. Mas-Balleste, M. R. Bukowski,
experiments. X-ray crystallographic data collection and structure
solutions were conducted by Dr Victor G. Young, Jr at the X-Ray
Crystallographic Laboratory, Department of Chemistry, Univer-
sity of Minnesota.
A. Stubna, M. Chakrabarti, L. Slominski, J. A. Halfen,
¨
E. Munck and L. Que Jr, Angew. Chem., Int. Ed., 2008, 47,
1896–1899.
22 H. Hirao, D. Kumar, L. Que Jr and S. Shaik, J. Am. Chem. Soc.,
2006, 128, 8590–8606.
23 H. Hirao, L. Que Jr, W. Nam and S. Shaik, Chem.–Eur. J.,
2008, 14, 1740–1756.
Notes and references
1 C. Krebs, D. Galonic Fujimori, C. T. Walsh and 24 S. Shaik, H. Chen and D. Janardanan, Nat. Chem., 2011, 3,
J. M. Bollinger Jr, Acc. Chem. Res., 2007, 40, 484–492.
19–27.
2 E. G. Kovaleva and J. D. Lipscomb, Nat. Chem. Biol., 2008, 4, 25 D. Usharani, D. Janardanan, C. Li and S. Shaik, Acc. Chem.
186–193. Res., 2013, 46, 471–482.
3 J. C. Price, E. W. Barr, B. Tirupati, J. M. Bollinger Jr and 26 M. S. Seo, H. G. Jang, J. Kim and W. Nam, Bull. Korean Chem.
C. Krebs, Biochemistry, 2003, 42, 7497–7508. Soc., 2005, 26, 971–974.
4 L. M. Hoffart, E. W. Barr, R. B. Guyer, J. M. Bollinger Jr and 27 S. A. Wilson, J. Chen, S. Hong, Y.-M. Lee, M. Clemancey,
C. Krebs, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 14738–
14743.
5 B. E. Eser, E. W. Barr, P. A. Frantom, L. Saleh, J. M. Bollinger
R. Garcia-Serres, T. Nomura, T. Ogura, J.-M. Latour,
B. Hedman, K. O. Hodgson, W. Nam and E. I. Solomon,
J. Am. Chem. Soc., 2012, 134, 11791–11806.
Jr, C. Krebs and P. F. Fitzpatrick, J. Am. Chem. Soc., 2007, 129, 28 C. Bucher, E. Duval, J.-M. Barbe, J.-N. Verpeaux, C. Amatore
11334–11335.
and R. Guilard, C. R. Acad. Sci., Ser. IIc, Chim., 2000, 3, 211–
222.
29 C. Amatore, J.-M. Barbe, C. Bucher, E. Duval, R. Guilard and
J.-N. Verpeaux, Inorg. Chim. Acta, 2003, 356, 267–278.
6 D. P. Galonic, E. W. Barr, C. T. Walsh, J. M. Bollinger Jr and
C. Krebs, Nat. Chem. Biol., 2007, 3, 113–116.
7 M. L. Matthews, C. M. Krest, E. W. Barr, F. H. Vaillancourt,
C. T. Walsh, M. T. Green, C. Krebs and J. M. Bollinger Jr, 30 D. Tschudin, A. Basak and T. A. Kaden, Helv. Chim. Acta,
Biochemistry, 2009, 48, 4331–4343.
1988, 71, 100–106.
8 A. J. Panay, M. Lee, C. Krebs, J. M. Bollinger Jr and 31 D. Tschudin and T. A. Kaden, Pure Appl. Chem., 1988, 60,
P. F. Fitzpatrick, Biochemistry, 2011, 50, 1928–1933. 489–493.
9 A. R. McDonald and L. Que Jr, Coord. Chem. Rev., 2013, 257, 32 T. A. Kaden, D. Tschudin, M. Studer and U. Brunner, Pure
414–428. Appl. Chem., 1989, 61, 879–883.
10 J. Hohenberger, K. Ray and K. Meyer, Nat. Commun., 2012, 3, 33 A. W. Addison, T. N. Rao, J. Reedijk, J. Van Rijn and
720.
G. C. Verschoor, J. Chem. Soc., Dalton Trans., 1984, 1349–
1356.
34 E. K. Bareeld, Coord. Chem. Rev., 2010, 254, 1607–1627.
11 S. P. de Visser, J.-U. Rohde, Y.-M. Lee, J. Cho and W. Nam,
Coord. Chem. Rev., 2013, 257, 381–393.
12 J.-U. Rohde, J.-H. In, M. H. Lim, W. W. Brennessel, 35 T. A. Jackson, J.-U. Rohde, M. S. Seo, C. V. Sastri, R. DeHont,
¨
¨
M. R. Bukowski, A. Stubna, E. Munck, W. Nam and L. Que
A. Stubna, T. Ohta, T. Kitagawa, E. Munck, W. Nam and
Jr, Science, 2003, 299, 1037–1039.
L. Que Jr, J. Am. Chem. Soc., 2008, 130, 12394–12407.
13 E. J. Klinker, J. Kaizer, W. W. Brennessel, N. L. Woodrum, 36 A. Decker, J.-U. Rohde, L. Que Jr and E. I. Solomon, J. Am.
C. J. Cramer and L. Que Jr, Angew. Chem., Int. Ed., 2005,
44, 3690–3694.
14 A. Thibon, J. England, M. Martinho, V. G. Young Jr,
Chem. Soc., 2004, 126, 5378–5379.
37 A. Decker, J.-U. Rohde, E. J. Klinker, S. D. Wong, L. Que Jr
and E. I. Solomon, J. Am. Chem. Soc., 2007, 129, 15983–15996.
¨
J. R. Frisch, R. Guillot, J.-J. Girerd, E. Munck, L. Que Jr and 38 J. Kaizer, E. J. Klinker, N. Y. Oh, J.-U. Rohde, W. J. Song,
¨
F. Banse, Angew. Chem., Int. Ed., 2008, 47, 7064–7067.
15 S. Meyer, I. Klawitter, S. Demeshko, E. Bill and F. Meyer,
Angew. Chem., Int. Ed., 2013, 52, 901–905.
A. Stubna, J. Kim, E. Munck, W. Nam and L. Que Jr, J. Am.
Chem. Soc., 2004, 126, 472–473.
39 F. H. Allen, O. Kennard, D. G. Watson, L. Brammer,
A. G. Orpen and R. Taylor, J. Chem. Soc., Perkin Trans. 2,
1987, S1–S19.
¨
16 J. England, Y. Guo, E. R. Farquhar, V. G. Young Jr, E. Munck
and L. Que Jr, J. Am. Chem. Soc., 2010, 132, 8635–8644.
1214 | Chem. Sci., 2014, 5, 1204–1215
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
40 L. G. Wade, Organic Chemistry, Pearson Prentice Hall, Upper 45 M. R. Bukowski, K. D. Koehntop, A. Stubna, E. L. Bominaar,
¨
Saddle River, NJ, 2006.
J. A. Halfen, E. Munck, W. Nam and L. Que Jr, Science, 2005,
41 F. A. Carey, Organic Chemistry, McGraw Hill, New York, 2006.
310, 1000–1002.
¨
42 K. Ray, J. England, A. T. Fiedler, M. Martinho, E. Munck and 46 D. Wang, K. Ray, M. J. Collins, E. R. Farquhar, J. R. Frisch,
L. Que Jr, Angew. Chem., Int. Ed., 2008, 47, 8068–8071.
43 L. Toib, S. Dermaik, M. Michman and Y. Sasson, Synlett,
1995, 245–246.
L. Gomez, T. A. Jackson, M. Kerscher, A. Waleska,
P. Comba, M. Costas and L. Que Jr, Chem. Sci., 2013, 4,
282–291.
44 J.-U. Rohde, S. Torelli, X. Shan, M. H. Lim, E. J. Klinker, 47 S. Hong, Y.-M. Lee, K.-B. Cho, K. Sundaravel, J. Cho,
J. Kaizer, K. Chen, W. Nam and L. Que Jr, J. Am. Chem.
Soc., 2004, 126, 16750–16761.
M. J. Kim, W. Shin and W. Nam, J. Am. Chem. Soc., 2011,
133, 11876–11879.
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 1204–1215 | 1215