X-ray absorption spectroscopy and reactivity of thiolate-ligated Fe III-OOR complexes

Article abstract of DOI:10.1021/ic100670k

The reaction of a series of thiolate-ligated iron(II) complexes [FeII([15]aneN₄)(SC₆H₅)]BF4 (1), [Fe ^II([15]aneN₄)-(SC₆H₄-p-Cl)]BF ₄ (2), and [Fe^II([15]aneN₄)(SC ₆H₄-p-NO₂)]BF₄ (3) with alkylhydroperoxides at low temperature (-78 °C or-40 °C) leads to the metastable alkylperoxo-iron(III) species [Fe^III ([15]aneN ₄)(SC₆H₅)(OOtBu)]BF₄ (1a), [Fe ^III ([15]aneN₄)(SC6H₄-p-Cl)(OOtBu)]BF ₄ (2a), and [Fe^III ([15]aneN₄)(SC ₆H₄-p-NO₂)(OOtBu)]BF₄ (3a), respectively. X-ray absorption spectroscopy (XAS) studies were conducted on the Fe^III-OOR complexes and their iron(II) precursors. The edge energy for the iron(II) complexes (～7118 eV) shifts to higher energy upon oxidation by ROOH, and the resulting edge energies for the Fe^III-OOR species range from 7121-7125 eV and correlate with the nature of the thiolate donor. Extended X-ray absorption fine structure (EXAFS) analysis of the iron(II) complexes 1-3 in CH₂Cl₂show that their solid state structures remain intact in solution. The EXAFS data on 1a-3a confirm their proposed structures as mononuclear, 6-coordinate Fe_III-OOR complexes with 4N and 1S donors completing the coordination sphere. The Fe-O bond distances obtained from EXAFS for 1a-3a are 1.82-1.85 a, significantly longer than other low-spin Fe_III-OOR complexes. The Fe-O distances correlate with the nature of the thiolate donor, in agreement with the previous trends observed for v(Fe-O) from resonance Raman (RR) spectroscopy, and supported by optimized geometries obtained from density functional theory (DFT) calculations. Reactivity and kinetic studies on 1a-3a show an important influence of the thiolate donor.

Full text of DOI:10.1021/ic100670k

9178 Inorg. Chem. 2010, 49, 9178–9190
DOI: 10.1021/ic100670k
X-ray Absorption Spectroscopy and Reactivity of Thiolate-Ligated
FeIII-OOR Complexes
Jay Stasser,‡ Frances Namuswe,† Gary D. Kasper,† Yunbo Jiang,† Courtney M. Krest,§ Michael T. Green,§
James Penner-Hahn,*,‡ and David P. Goldberg*,†
†Department of Chemistry, Johns Hopkins University, Baltimore, Maryland 21218, ‡Department of Chemistry,
§
University of Michigan, Ann Arbor, Michigan 48109, and Department of Chemistry, Pennsylvania State
University, University Park, Pennsylvania 16802
Received April 8, 2010
The reaction of a series of thiolate-ligated iron(II) complexes [FeII([15]aneN4)(SC6H5)]BF4 (1), [FeII([15]aneN4)-
(SC6H4-p-Cl)]BF4 (2), and [FeII([15]aneN4)(SC6H4-p-NO2)]BF4 (3) with alkylhydroperoxides at low temperature
(-78 °C or -40 °C) leads to the metastable alkylperoxo-iron(III) species [FeIII([15]aneN4)(SC6H5)(OOtBu)]BF4
(1a), [FeIII([15]aneN4)(SC6H4-p-Cl)(OOtBu)]BF4 (2a), and [FeIII([15]aneN4)(SC6H4-p-NO2)(OOtBu)]BF4 (3a),
respectively. X-ray absorption spectroscopy (XAS) studies were conducted on the FeIII-OOR complexes and their
iron(II) precursors. The edge energy for the iron(II) complexes (∼7118 eV) shifts to higher energy upon oxidation by
ROOH, and the resulting edge energies for the FeIII-OOR species range from 7121-7125 eV and correlate with the
nature of the thiolate donor. Extended X-ray absorption fine structure (EXAFS) analysis of the iron(II) complexes 1-3
in CH2Cl2 show that their solid state structures remain intact in solution. The EXAFS data on 1a-3a confirm their
proposed structures as mononuclear, 6-coordinate FeIII-OOR complexes with 4N and 1S donors completing the
˚
coordination sphere. The Fe-O bond distances obtained from EXAFS for 1a-3a are 1.82-1.85 A, significantly
longer than other low-spin FeIII-OOR complexes. The Fe-O distances correlate with the nature of the thiolate donor,
in agreement with the previous trends observed for ν(Fe-O) from resonance Raman (RR) spectroscopy, and
supported by optimized geometries obtained from density functional theory (DFT) calculations. Reactivity and kinetic
studies on 1a- 3a show an important influence of the thiolate donor.
Introduction
of biological systems. These systemsinclude the mononuclear
nonheme iron enzymes superoxide reductase,1-11 soybean lip-
oxygenase,12 extradiol13-16 and Rieske dioxygenases,13,15,16
1-aminocyclopropane-1-carboxylate oxidase,17 and the anti-
biotic bleomycin.16,18,19 Significant effort has gone into char-
acterizing iron-dioxygen adducts (Fe-OO(H or R)) in both
the biological systems and in biologically relevant small-
molecule model complexes. However, structural characteri-
zation of such species remains limited. The first X-ray
structure of a mononuclear nonheme iron-peroxide complex
The coordination and processing of dioxygen and its
reduced forms (superoxide, peroxide) at both heme and
nonheme iron centers is of critical importance to a number
*To whom correspondence should be addressed. E-mail: dpg@jhu.edu
(D.P.G.), jeph@umich.edu (J.P.-H.).
(1) Emerson, J. P.; Coulter, E. D.; Cabelli, D. E.; Phillips, R. S.; Kurtz,
D. M., Jr. Biochemistry 2002, 41, 4348–4357.
(2) Katona, G.; Carpentier, P.; Niviere, V.; Amara, P.; Adam, V.; Ohana,
J.; Tsanov, N.; Bourgeois, D. Science 2007, 316, 449–453.
(3) Kurtz, D. M., Jr. Acc. Chem. Res. 2004, 37, 902–908.
(4) Lombard, M.; Houee-Levin, C.; Touati, D.; Fontecave, M.; Niviere,
ꢀ
ꢁ
ꢀ
(11) Moura, I.; Pauleta, S. R.; Moura, J. J. G. J. Biol. Inorg. Chem. 2008,
13, 1185–1195.
(12) Skrzypczak-Jankun, E.; Bross, R. A.; Carroll, R. T.; Dunham,
W. R.; Funk, M. O., Jr. J. Am. Chem. Soc. 2001, 123, 10814–10820.
(13) Costas, M.; Mehn, M. P.; Jensen, M. P.; Que, L., Jr. Chem. Rev. 2004,
104, 939–986.
(14) Kovaleva, E. G.; Lipscomb, J. D. Science 2007, 316, 453–457.
(15) Kovaleva, E. G.; Neibergall, M. B.; Chakrabarty, S.; Lipscomb, J. D.
Acc. Chem. Res. 2007, 40, 475–483.
(16) Que, L., Jr.; Ho, R. Y. N. Chem. Rev. 1996, 96, 2607–2624.
(17) Rocklin, A. M.; Kato, K.; Liu, H. W.; Que, L., Jr.; Lipscomb, J. D.
J. Biol. Inorg. Chem. 2004, 9, 171–182.
V. Biochemistry 2001, 40, 5032–5040.
ꢁ
ꢁ
(5) Mathe, C.; Weill, C. O.; Mattioli, T. A.; Berthomieu, C.; Houee-Levin,
ꢀ
C.; Tremey, E.; Niviere, V. J. Biol. Chem. 2007, 282, 22207–22216.
ꢀ
ꢁ
(6) Niviere, V.; Lombard, M.; Fontecave, M.; Houee-Levin, C. FEBS
Lett. 2001, 497, 171–173.
(7) Horner, O.; Mouesca, J. M.; Oddou, J. L.; Jeandey, C.; Niviere, V.;
Mattioli, T. A.; Mathe, C.; Fontecave, M.; Maldivi, P.; Bonville, P.; Halfen,
ꢀ
ꢁ
J. A.; Latour, J. M. Biochemistry 2004, 43, 8815–8825.
(8) Huang, V. W.; Emerson, J. P.; Kurtz, D. M., Jr. Biochemistry 2007, 46,
11342–11351.
ꢁ
(9) Mathe, C.; Mattioli, T. A.; Horner, O.; Lombard, M.; Latour, J. M.;
ꢀ
Fontecave, M.; Niviere, V. J. Am. Chem. Soc. 2002, 124, 4966–4967.
(18) Stubbe, J.; Kozarich, J. W. Chem. Rev. 1987, 87, 1107–1136.
(19) Stubbe, J.; Kozarich, J. W.; Wu, W.; Vanderwall, D. E. Acc. Chem.
Res. 1996, 29, 322–330.
ꢁ
ꢀ
(10) Mathe, C.; Niviere, V.; Houee-Levin, C.; Mattioli, T. A. Biophys.
Chem. 2006, 119, 38–48.
r
pubs.acs.org/IC
Published on Web 09/14/2010
2010 American Chemical Society

Article
Inorganic Chemistry, Vol. 49, No. 20, 2010 9179
of any type appeared in 2001 with the crystallographic char-
acterization of an alkylperoxo-iron(III) moiety in soybean
lipoxygenase.12 Other mononuclear, nonheme iron-peroxide
adducts in proteins that have been crystallographically char-
acterized include naphthalene dioxygenase (peroxo-iron),20
homoprotocatechuate 2,3-dioxygenase (superoxo-, peroxo-,
and alkylperoxo-iron)14,15 and superoxide reductase (SOR)
(peroxo-iron).2 Given the paucity of structures available for
these biologically relevant iron centers, it is important to
obtain structural information on analogous model com-
plexes, although thus far there is no X-ray structure available
for a mononuclear nonheme iron-peroxide model complex.
However, X-ray absorption spectroscopy (XAS) has been
successfully employed to gain structural insights into such
model complexes. For example, Kovacs et al. have charac-
terized a rare example of a thiolate-ligated, FeIII-OOH
complex by XAS,21 and Que and co-workers have charac-
terized both high-spin and low-spin nonheme alkylperoxo-
iron(III) intermediates with various polydentate supporting
ligands, mostly derived from pyridyl and/or neutral amine N
donors.22-24
We have been successful in synthesizing a series of
(N4S(thiolate))iron(II) complexes and using them to generate
metastable, low-spin alkylperoxo-iron(III) intermediates.
These complexes were prepared as models of the reduced
(His4Cys)FeII SOR active site and (hydro)peroxo-iron(III)
intermediates that may be important during turnover. SOR
reduces superoxide to hydrogen peroxide as part of a natural
defense mechanism in anaerobic microorganisms, and evi-
dence suggests that a (hydro)peroxo-iron(III) intermediate
may form during turnover. This intermediate must decay
through Fe-O bond cleavage to release H2O2, and avoid
O-O bond cleavage to give unwanted ferryl-type (Fe(O))
species.
cently, we have also succeeded in probing the influence of
the N donors of the [15]aneN4 ligand, which led to the
characterization of some high-spin FeIII-OOR species.27
The vibrational data from RR spectroscopy on the former
series of low-spin FeIII-OOR species showed that the
identity of the thiolate ligand influences the strength of
the Fe-O stretching frequency (ν(Fe-O)), but does not
perturb the O-O vibration. As the thiolate becomes more
electron-donating, a weakening of ν(Fe-O) is observed,
consistent with a trans influence of the thiolate donor weak-
ening the Fe-O bond. These findings were relevant to the
proposed mechanism for SOR, which relies upon Fe-O
bond cleavage of the (hydro)peroxo intermediate. However,
concrete structural information on the FeIII-OOR com-
plexes was lacking.
Herein we have used XAS to structurallycharacterize these
complexes. Iron K-edge X-ray absorption spectra have been
obtained for six different complexes, the alkylperoxo species
[FeIII([15]aneN4)(SC6H5)(OOtBu)]BF4 (1a), [FeIII([15]ane-
N4)(SC6H4-p-Cl)(OOtBu)]BF4 (2a), and [FeIII([15]aneN4)-
(SC6H4-p-NO2)(OOtBu)]BF4 (3a), and the iron(II) precursor
complexes [FeII([15]aneN4)(SC6H5)]BF4 (1), [FeII([15]-
aneN4)(SC6H4-p-Cl)]BF4 (2), and [FeII([15]aneN4)(SC6H4-
p-NO2)]BF4 (3). Examination of the X-ray absorption near-
edge spectroscopy (XANES) and extended X-ray absorption
fine structure (EXAFS) regions of the spectra has provided
detailed structural information on both the iron(II) and the
alkylperoxo-iron(III) complexes, including confirmation of
the thiolate ligation in both systems in solution. These studies
have also led to insights regarding the influence of the thiolate
donor on the structures of the alkylperoxo species. Compari-
son with the limited structural information on FeIII-OOR(H)
complexes aids in our understanding of the selection by
Nature of ligand type for the generation of peroxo-iron
species with particular properties. The reactivity of the
FeIII-OOR species toward electrophiles, nucleophiles, and
proton donors has also been examined. Kinetic studies on the
decay of the FeIII-OOR species suggest a reactivity pattern
controlled by the identity of the thiolate donor.
In our earlier work the FeIII-OOR model complexes were
generated at low temperature (-78 °C) and characterized by
UV-vis, electron paramagnetic resonance (EPR), and reso-
nance Raman (RR) spectroscopies. On the basis of the
combined spectroscopic data, these species were formulated
as [FeIII([15]aneN4)(SC6H4-p-X)(OOR)]BF4 ([15]aneN4 =
1,4,8,12-tetraazacyclopentadecane; X = H, OMe, Cl, NO2
or a polyfluorinated substituent; R= t-butyl or cumenyl),
with a low-spin ferric center.25,26 These low-spin FeIII-OOR
complexes contained arylthiolate ligands of varying electron-
donating power, which allowed us to examine the influence
of the sulfur donor on the spectroscopic properties of an
FeIII-OOR species at parity of ligand environment. Re-
Experimental Section
General Procedures. All synthetic reactions were carried out
under an atmosphere of N2 or Ar using a drybox or standard
Schlenk techniques. Reagents were purchased from commercial
vendors and used without further purification unless noted
otherwise. Dichloromethane was purified via a Pure-Solv Sol-
vent Purification System from Innovative Technology, Inc.
All solvents were degassed by repeated cycles of freeze-
pump-thaw and then stored in a drybox. Tert-butylhydroper-
oxide was purchased from Aldrich as a ∼5.5 M solution in
decane over molecular sieves. The concentration of tBuOOH
was determined as previously described.26 Cumene hydroper-
oxide (CmOOH) was purchased from Aldrich as an 88% solution
in ether. The iron(II) complexes [FeII([15]aneN4)(SC6H5)]-
BF4 (1), [FeII([15]aneN4)(SC6H4-p-Cl)]BF4 (2), and [FeII([15]-
aneN4)(SC6H4-p-NO2)]BF4 (3) were synthesized according to
published procedures.26
(20) Karlsson, A.; Parales, J. V.; Parales, R. E.; Gibson, D. T.; Eklund,
H.; Ramaswamy, S. Science 2003, 299, 1039–1042.
(21) Shearer, J.; Scarrow, R. C.; Kovacs, J. A. J. Am. Chem. Soc. 2002,
124, 11709–11717.
(22) Bukowski, M. R.; Zhu, S. R.; Koehntop, K. D.; Brennessel, W. W.;
Que, L., Jr. J. Biol. Inorg. Chem. 2004, 9, 39–48.
(23) Rohde, J. U.; Torelli, S.; Shan, X. P.; Lim, M. H.; Klinker, E. J.;
Kaizer, J.; Chen, K.; Nam, W. W.; Que, L., Jr. J. Am. Chem. Soc. 2004, 126,
16750–16761.
(24) Shan, X.; Rohde, J. U.; Koehntop, K. D.; Zhou, Y.; Bukowski,
M. R.; Costas, M.; Fujisawa, K.; Que, L., Jr. Inorg. Chem. 2007, 46, 8410–
8417.
Physical Methods. Low-temperature UV-visible spectra
were recorded at -78 °C or -40 °C (for kinetics) on a Cary 50
Bio spectrophotometer equipped with a fiber-optic coupler
(Varian) and a fiber-optic dip probe (Hellma 661.302-QX-UV,
(25) Krishnamurthy, D.; Kasper, G. D.; Namuswe, F.; Kerber, W. D.;
€
Sarjeant, A. A. N.; Moenne-Loccoz, P.; Goldberg, D. P. J. Am. Chem. Soc.
2006, 128, 14222–14223.
(26) Namuswe, F.; Kasper, G. D.; Sarjeant, A. A. N.; Hayashi, T.; Krest,
(27) Namuswe, F.; Hayashi, T.; Jiang, Y. B.; Kasper, G. D.; Sarjeant,
€
€
C. M.; Green, M. T.; Moenne-Loccoz, P.; Goldberg, D. P. J. Am. Chem. Soc.
A. A. N.; Moenne-Loccoz, P.; Goldberg, D. P. J. Am. Chem. Soc. 2010, 132,
2008, 130, 14189–14200.
157–167.

9180 Inorganic Chemistry, Vol. 49, No. 20, 2010
Stasser et al.
2 mm path length) for low temperature, using custom-made
Schlenk tubes designed for the fiber-optic dip probe. GC-FID
analysis was performed on an Agilent 6890N GC equipped with
a 30 m, 5% DPS column and an FID detector. GC-MS analysis
was performed on a Shimadzu GC17A GC equipped with a 30 m,
5% DPS column and a QP5050A quadrupole MS detector.
Computational Methods. Density functional theory (DFT)
calculations were performed on a model of the FeIII-OOR com-
plexes 1a-3a in which R = CH3. Frequency calculations were
performed at optimized geometries. Initial coordinates were
taken from the X-ray crystal structures of the iron(II) complexes
1- 3, and both high-spin and low-spin iron(III) configurations
(S = 5/2 or 1/2) were evaluated. Calculations were performed
with Gaussian 03 at the B3LYP/6-311G level of theory.28
XAS Sample Preparation. Custom designed sample holders
consisted of an aluminum plate sealed with Kapton tape on
both sides. Solid samples of 1, 2, and 3 were prepared by mixing
5-10 mg of the FeII complexes with boron nitride under an inert
gas atmosphere. The solid mixture was then transferred to an
aluminum sample plate taped on one side. Sample preparation
was completed by sealing the open side of the plate with Kapton
tape. Samples of the alkylperoxoiron(III) complexes 1a, 2a, and
3a were generated by adding 200 μL of the respective iron(II)
precursors dissolved in CH2Cl2 (20 mM) to an aluminum sample
holder sealed with Kapton tape on both sides with a volume
capacity of 300 μL. The holders were cooled to -78 °C over dry
ice, and then 50 μL of tBuOOH (1.75 equiv) stock solution in
CH2Cl2 was added to each sample holder. The samples were
mixed by manually shaking them, and after a suitable reaction
time the desired chromophore was observed (∼ 5 min). The
samples were then immediately submerged in liquid nitrogen for
storage until XAS studies could be conducted.
fit over the range of the data and subtracted to give the final
EXAFS spectrum by using EXAFSPAK, made available by
Dr. G. George on the SSRL Web site. The data were converted
to k space via the equation
1=2
k ¼ ½2meðE - E0Þ=h2ꢀ
where E0 = 7112 eV. Data were fit, using Excurv98,31 to the
EXAFS equation
X
NsAasðkÞ
χðkÞ ¼
expð - 2σ2ask2Þ sin½2kRas þ φasðkÞꢀ
kR2as
S
where Ns is the number of scatterers with atom type s at a
distance of Ras. Aas(k) is the effective backscattering amplitude
function, σ2as is the Debye-Waller factor, and φas(k) is the phase
shift of the photoelectron wave traveling between the potentials
of the absorbing and scattering atoms. Fourier transforms of the
EXAFS data were calculated using k3-weighted data from k=
2.0-15 A-1. Multiple scattering paths, phase shifts, and ampli-
˚
tudes were calculated in the Excurv98 program. Data were k3-
weighted, and the individual fits were performed with the
coordination numbers fixed and the Debye-Waller factors held
constant at values suggested by Dimakis and Bunker,32 while the
R
as values were allowed to vary. This fitting process allowed us
to determine coordination number and Ras values with a mini-
mal number of parameters. After obtaining optimal fits in this
manner, the Debye-Waller factors were varied and shown to
refine to reasonable values. Distances of the outer shell carbons
in both the N4 and the S donors were initially set using the
crystallographic distances. Fits were constrained to have fixed
S-C, N-C, and C-C distances, but were allowed to have freely
variable angles between the metal center and the ligand, for a
total of 8 variable parameters each for the thiolate and the
macrocyclic N4 ligand. In addition to the variable distances, the
threshold energies were allowed to vary. The final structures
were analyzed to ensure that in no case were there significant
changes in the intramolecular angles. In addition (see below)
the Fe-O and Fe-S coordination numbers were varied for
some fits.
XAS. Data Collection and Analysis. Fe K-edge X-ray absorp-
tion spectra (XAS) were collected at the Stanford Synchrotron
Radiation Laboratory (SSRL) on beamline 9-3 at 3 GeV with
currents between 80 and 100 mA. The samples were kept at 10 to
12 K in an Oxford liquid helium flow cryostat. Contamination
of higher harmonics radiation was minimized using a Rh-coated
mirror upstream of the Si[220] double crystal monochromator;
the mirror was set to a 10 keV energy cutoff. Data were collected
in fluorescence excitation mode using a high-count-rate Canberra
30-element Ge array detector with maximum count rates held
below 120 kHz per channel. Fe K-edge spectra were col-
lected using 5 eV steps in the pre-edge (6900-7100 eV), 0.2 eV
Pre-edge peaks were isolated by fitting the edge and first
XANES peak (7100-7150 eV) background using the sum of an
arctan and Gaussian functions. This background was then
subtracted from the normalized data to isolate the pre-edge
peak.
steps from 7100 to 7140 and 0.05 A-1 steps in the EXAFS region
˚
Reaction of Alkylperoxo-Iron(III) Complexes with Electro-
philes and Nucleophiles. The compounds PPh3, cyclooctene, 2,6-
di-t-butylphenol, cyclohexadiene (CHD), 2-phenyl propional-
dehyde (2-PPA), cyclohexane carboxaldehyde (CCA), and
to k = 15 A-1. Data collection was integrated for 1 s in the
˚
preedge, 1.5 s in edge regions, and 1.5 to 25 s (k2-weighted for
-1
3
k = 2.6 to 8 A and k -weighted for k = 8 to 15 A-1) in the
EXAFS region for total scan time of roughly 45 min. A Mn filter
(absorption equivalent to ∼20 μm Mn metal) was placed in front
of the detector to reduce the elastic scatter peak, and Soller slits
were used to reduce the Mn KR fluorescence. Each sample was
scanned multiple times to give a total of ∼109 counts total in the
˚
˚
HþBF4 Et2O were reacted with the alkylperoxo-iron(III)
-
3
complexes generated in situ. In a typical reaction, a solution of
1 in CH2Cl2 (7 mL, 3.31 mM) was cooled to -78 °C, and
tBuOOH (101 μL, 5.79 mM) was added from a stock solution
in decane/CH2Cl2. An immediate color change from colorless
to dark red was noted. The reaction was allowed to stir until
the alkylperoxo-iron(III) species 1a was fully formed, as indi-
cated by maximal absorbance at 526 nm. The substrate 2-PPA
(922 μL, 200 equiv) was then added, and caution was taken to
prevent warming of the reaction mixture. A rapid color change
from red to yellow was observed. After the red color was com-
pletely discharged, the reaction mixture was warmed to room
temperature and analyzed by GC-FID.
Kinetic Studies. In a typical reaction, [FeII([15]aneN4)(SC6-
H4-p-X)]BF4 (X = H, Cl, NO2) (5-10 mg) was loaded in a Sch-
lenk flask and dissolved in CH2Cl2 to give a final concentration
of 1-3 mM. A fiber optic dip-probe for UV-vis measurements
windowed Fe KR peak at k = 15 A-1, requiring 3 to 10 scans
˚
depending on sample concentration.
X-ray energies were calibrated by simultaneous measurement
of a Fe foil absorption spectrum and assigning the first inflection
point of the absorption edge as 7111.2 eV. Data from each
detector channel were inspected for glitches before inclusion
in the final average. Pre-edge subtraction and normalization to
McMaster values29 were carried out using the program
MBACK.30 A four-region third-order polynomial spline was
(28) Frisch, M. J. et al. Gaussian 03, Revision C.02; Gaussian, Inc.:
Wallingford, CT, 2004.
(29) McMaster, W. H.; Del Grande, N. K.; Mallet, J. H.; Hubbell, J. H.
Compilation of X-Ray Cross Sections, Lawrence Livermore National Labora-
tory Report UCRL-50174, Sect. 2, Rev. 1, 1969.
(30) Weng, T. C.; Waldo, G. S.; Penner-Hahn, J. E. J. Synchrotron Rad.
2005, 12, 506–510.
(31) Binsted, N. EXCURV98; CCLRC Daresbury Laboratory: Warrington,
U.K., 1998.
(32) Dimakis, N.; Bunker, G. Phys. Rev. B 2002, 65, 201103.

Article
Inorganic Chemistry, Vol. 49, No. 20, 2010 9181
Scheme 1. Formation of FeIII-OOR Complexes at -78 °C
was inserted into the flask, and the solution was cooled to
-40 °C. The alkylhydroperoxide (tBuOOH or CmOOH, 1.5
equiv, in CH2Cl2) was added to the FeII complex. A color change
from colorless (yellow for 3) to burgundy was observed, indicat-
ing the formation of 1a-3a or 1b- 3b, which was confirmed by
UV-vis spectroscopy. The kinetics of decay of the alkylperoxo-
iron(III) species was monitored by UV-vis spectral changes at
529, 524, 522, 528, 530, or 521 nm. Isosbestic points were observed
throughout the decay of each complex. First-order rate con-
stants (kobs) were calculated by plotting ln[( At - Af)/( A0 - Af)]
versus time, where A0 = initial absorbance at time = 0, At =
absorbance at time t, and Af = final absorbance. Kinetic data
for each intermediate were collected at least in triplicate.
Figure 1. Normalized Fe K-Edge XANES of [FeII([15]aneN4)-
(SC6H5)]BF4 (1) as a crystalline solid dispersed in BN matrix (solid line)
and dissolved in CH2Cl2 (dotted line), and [FeIII([15]aneN4)(SC6H5)-
(OOtBu)]BF4 (1a) (dashed line) dissolved in CH2Cl2.
Results and Discussion
Preparation of the Alkylperoxo Complexes [FeIII([15]-
aneN4)(SC6H4-p-X)(OOtBu)]þ. The alkylperoxo-iron(III)
complexes [FeIII([15]aneN4)(SC6H5)(OOtBu)]BF4 (1a),
[FeIII([15]aneN4)(SC6H4-p-Cl)(OOtBu)]BF4 (2a), and
[FeIII([15]aneN4)(SC6H4-p-NO2)(OOtBu)]BF4 (3a) were
generated by addition of tBuOOH to their respective
iron(II) precursors at low-temperature, as shown in
Scheme 1. Previously these species, which are a dark
burgundy in color, were characterized by UV-vis spec-
troscopy and exhibited UV-vis bands between 520-
530 nm, which are indicative of an alkylperoxo-to-iron(III)
ligand-to-metal charge transfer (LMCT) band. RR spec-
troscopy confirmed their identities as alkylperoxo com-
plexes, with prominent peaks between 610-620 cm-1
typical of ν(Fe-O) and a peak near 800 cm-1 character-
istic of ν(O-O), both of which shifted with 18O substitu-
tion. These complexes were characterized as low-spin
FeIII (S = 1/2) by EPR spectroscopy. The analogous
series of alkylperoxo complexes 1b-3b were prepared in
the same manner (Scheme 1) and exhibited similar spec-
troscopic signatures in the UV-vis, RR, and EPR
measurements.26 However, direct characterization of the
solution state structures of either the iron(II) precursor
complexes or the alkylperoxo-iron(III) species were lack-
ing. Below we describe XAS studies aimed at obtaining
detailed structural information of these complexes.
Figure 2. Normalized Fe K-Edge XANES in CH2Cl2 of[FeII([15]aneN4)-
(SC6H5)]BF4 (1) (solid line), [FeII([15]aneN4)(SC6H4-p-Cl)]BF4 (2)
(dotted line), and [FeII([15]aneN4)(SC6H4-p-NO2)]BF4 (3) (dashed line).
Inset: expanded view of pre-edge features.
tion, Figure S1. The solution state data for all three FeII
complexes 1-3 are overlaid for comparison in Figure 2.
The energy of the rising Fe K-edge depends both on the
metal oxidation state and on the structure. In particular, a
decrease in Fe-ligand distance (for example, because of a
change in coordination number or a change in spin state)
typically results in an increase in edge energy. The edge
energy for crystalline 1 is 7116.9 eV, and falls in the
expected range for an iron(II) complex.21,33 The edge is
shifted to higher energy by ∼1.1 eV upon dissolution of
the complex in CH2Cl2, and the intensities of the first two
XANES peaks (∼7130-7135 eV) are altered, with the
second peak going from a weak, poorly resolved shoulder
to a pronounced peak. Nearly identical changes are seen
for 2 and 3 (Supporting Information, Figure S1). These
data demonstrate that a structural rearrangement of
XANES of the FeII Complexes and FeIII-OOtBu Inter-
mediates. The X-ray absorption spectra in the near-edge
region for the iron(II) precursor complex 1 as a crystalline
solid (dispersed in boron nitride (BN) matrix) and dis-
solved in solution (CH2Cl2) are shown in Figure 1, and
similar plots for 2 and 3 are given in Supporting Informa-
(33) Westre, T. E.; Kennepohl, P.; DeWitt, J. G.; Hedman, B.; Hodgson,
K. O.; Solomon, E. I. J. Am. Chem. Soc. 1997, 119, 6297–6314.

9182 Inorganic Chemistry, Vol. 49, No. 20, 2010
Stasser et al.
some sort takes place for 1- 3 upon dissolution, although
it is difficult to interpret the higher energy XANES
features, which are sensitive to small changes in multiple
scattering, in terms of quantitative structural changes.
While it is possible that changes such as these could arise
from partial oxidation of FeII to FeIII (complete oxidation
would induce a much larger shift, see the spectrum for
1a in Figure 1 for comparison), careful observation of
the color of the samples (colorless for FeII, brown upon
conversion to FeIII by exposure to air) showed no evi-
dence of oxidation. Similarly, the XANES changes
on dissolution (Figure 1 and Supporting Information,
Figure S1) could, in principle, arise from gross decom-
position of the complex; however, this possibility can be
excluded by the EXAFS results (vide infra) demonstrat-
ing that the core ligation remains unchanged, aside from
modest changes in bond length.
For comparison, the near-edge regions for 1-3 in
CH2Cl2 are plotted together in Figure 2. The shape and
position of the rising edge for all three compounds are
nearly identical (edge energy = 7117.9 eV (1); 7118.1 eV
(2); 7118.0 eV (3)), and all three show broad, weak pre-
edge transitions at ∼7111 eV (inset, Figure 2). Thus, it is
unreasonable to assume that the modest changes in the
edge regions of 1-3 upon dissolution arise from identical
amounts of air contamination and subsequent oxidation,
but rather these changes must arise from conformational
changes of the complexes. The pre-edge peaks appear
to arise from several overlapping peaks, which can be
assigned to formally parity-forbidden 1s f 3d transitions.
From the crystallographic data on these high-spin iron-
(II) complexes, the iron centers are best described as
distorted square pyramidal based on a normalized τ
value34 (τ=0.50 (1), 0.50 (2), 0.27 (3), where τ=1.0 for
trigonal bipyramidal and τ = 0 for square pyramidal).
The overlapping peaks observed in the pre-edge region
can be attributed to the multiple 1s f 3d transitions that
are available for 5-coordinate, high-spin FeII, which
contains up to four “holes” in the d shell with differing
energies.33
The near-edge regions of the XANES spectra for the
low-spin iron(III)-alkylperoxo complexes 1a-3a are
shown in Figure 3, and they are also overlaid against
their reduced iron(II) precursors in Figure 1 and Support-
ing Information, Figure S1 for comparison. The Fe
K-edge energies for 1a-3a are upshifted by 3-7 eV as
compared to the iron(II) precursors 1-3. This increase in
edge energy is consistent with an increase in oxidation
state from FeII to FeIII for 1a-3a. Although the edge
positions for 1a-3a vary by ∼3 eV, they all fall in the
Figure 3. Normalized Fe K-Edge XANES in CH2Cl2 of [FeIII([15]-
aneN4)(SC6H5)(OOtBu)]BF4 (1a) (solid line), [FeIII([15]aneN4)(SC6H4-
p-Cl)(OOtBu)]BF4 (2a) (dotted line), and [FeIII([15]aneN4)(SC6H4-p-
NO2)(OOtBu)]BF4 (3a) (dashed line). Inset: expanded view of pre-edge
features.
range of previously reported FeIII-OOR(H) com-
plexes.21-23,35-40 Given that 1a-3a are all FeIII com-
plexes, the change in the edge energies must arise from
differences in the coordination environments of the
metal centers, from differences in composition (e.g.,
partial decomposition), or from a combination of these
effects. Inspection of the trend in edge energies for 1a-3a
reveals that they correlate with the identity of the thiolate
donor, decreasing as the thiolate ligand becomes less
electron-donating (1a: 7124.6 eV > 2a: 7122.4 eV > 3a:
7121.4 eV). A reduction in the edge energy can be attri-
buted to a strengthening of the Fe-O bond, which satis-
fyingly corroborates our previous analysis.25,26 In our
earlier work, it was shown that the Fe-O stretching
frequency increases as the thiolate ligand becomes a
weaker donor (ν(Fe-O): 1a<2a<3a), which was inter-
preted as a strengthening of the Fe-O bond induced by a
reduction of the trans effect of the thiolate ligand. A
similar shift in edge energies was reported for a pair of
high-spin FeIII-OOtBu complexes [FeIII(L8Py2)(OOtBu)-
(X)]2þ (L8Py2 = 1,5-bis(pyridin-2-ylmethyl)-1,5-diazacy-
clooctane, X = CH3CN or pyridine N-oxide (PyO)), in
which the more weakly donating CH3CN ligand resulted
in a 2.1 eV downshift of the edge position. This result was
attributed to an increased covalency of the FeIII-OOR
bond for the CH3CN complex.24 Thus, the XANES data
for 1a-3a provide good evidence that the covalency, and
therefore the bond strength, of the Fe-O bonds in these
complexes is directly influenced by the nature of the
thiolate ligand.
(34) Addison, A. W.; Rao, T. N.; Reedjik, J.; van Rijn, J.; Verschoor,
G. C. Dalton Trans. 1984, 1349–1456.
(35) Koehntop, K. D.; Rohde, J. U.; Costas, M.; Que, L., Jr. Dalton
Trans. 2004, 3191–3198.
(36) Lehnert, N.; Ho, R. Y. N.; Que, L., Jr.; Solomon, E. I. J. Am. Chem.
Soc. 2001, 123, 8271–8290.
(37) Lehnert, N.; Neese, F.; Ho, R. Y. N.; Que, L., Jr.; Solomon, E. I.
J. Am. Chem. Soc. 2002, 124, 10810–10822.
Expanded views of the pre-edge regions for 1a-3a
(Figure 3), show the 1s f 3d transitions. The pre-edge
features for 1a and 2a are similar in energy and intensity,
while that for 3a appears to be shifted to lower energy by
∼1 eV and to be more intense. This apparently anomalous
behavior for 3a is, at least in part, the consequence of the
fact that the XANES spectrum for complex 3a appears to
have a significantly different background. Thus, once the
background is subtracted, it is clear that 1s f 3d transi-
tions (Figure 4) all occur at approximately the same
(38) Bautz, J.; Comba, P.; Que, L., Jr. Inorg. Chem. 2006, 45, 7077–7082.
(39) Jensen, M. P.; Costas, M.; Ho, R. Y. N.; Kaizer, J.; Payeras, A. M. I.;
€
Munck, E.; Que, L., Jr.; Rohde, J. U.; Stubna, A. J. Am. Chem. Soc. 2005,
127, 10512–10525.
(40) Roelfes, G.; Vrajmasu, V.; Chen, K.; Ho, R. Y. N.; Rohde, J. U.;
Zondervan, C.; la Crois, R. M.; Schudde, E. P.; Lutz, M.; Spek, A. L.; Hage,
€
R.; Feringa, B. L.; Munck, E.; Que, L., Jr. Inorg. Chem. 2003, 42, 2639–2653.

Article
Inorganic Chemistry, Vol. 49, No. 20, 2010 9183
Figure 4. Background-subtracted pre-edge features for (bottom) [FeII-
([15]aneN4)(SC6H5)]BF4 (1) (solid line) and [FeIII([15]aneN4)(SC6H5)-
(OOtBu)]BF4 (1a) (dotted line), (middle) [FeII([15]aneN4)(SC6H4-p-
Cl)]BF4 (2) (solid line) and [FeIII([15]aneN4)(SC6H4-p-Cl)(OOtBu)]BF4
(2a) (dotted line), and (top) [FeII([15]aneN4)(SC6H4-p-NO2)]BF4 (3)
(solid line) and [FeIII([15]aneN4)(SC6H4-p-NO2)(OOtBu)]BF4 (3a)
(dotted line).
Figure 5. Fourier transforms of Fe K-edge EXAFS data and EXAFS
spectra (insets) in CH2Cl2, experimental data (solid line) and theoretical
fits (dotted line) for (bottom) [FeII([15]aneN4)(SC6H5)]BF4 (1), (middle)
[FeII([15]aneN4)(SC6H4-p-Cl)]BF4 (2), and (top) [FeII([15]aneN4)(SC6-
H4-p-NO2)]BF4 (3).
energy. The pre-edge peaks for 1a and 2a are less intense
than those observed for the iron(II) precursor complexes.
This result is consistent with the iron centers of 1a and 2a
being 6-coordinate, compared to the 5-coordinate envi-
ronments of the iron centers in 1 and 2 since, in general,
1s-3d intensity decreases as coordination number in-
creases because the metal has become more nearly cen-
trosymmetric.33,41 Interestingly, the pre-edge feature for
the para-nitro complex 3a appears more intense than that
of its iron(II) precursor. While this may be due, in part, to
an increase in the strength of the Fe-O interaction (see
below), it may also reflect incomplete background re-
moval in this case. A stronger Fe-O interaction would
reduce the symmetry of the iron site and cause a gain in
intensity of the pre-edge features. This result also corre-
lates with the trend in Fe-O bond distances seen by
EXAFS (vide infra).
EXAFS Spectra and Fitting Results. Iron(II) Com-
plexes. The normalized, phase-shifted Fourier transforms
(FT) of the Fe K-edge EXAFS data and the EXAFS
spectra (insets) for the three high-spin, thiolate-ligated
iron(II) precursor complexes 1- 3 in frozen CH2Cl2
solution are shown in Figure 5. The best fits are overlaid
with the spectra, and the curve-fitting results are sum-
marized in Table 1. The analogous data for 1- 3 in the
solid state are presented in Supporting Information,
Figure S3 and Table S1. The most prominent feature in
the Fourier transforms for all three complexes in CH2Cl2
˚
broad peak of significant intensity at R ∼ 3.0 A. The
fitting of the EXAFS data for 1-3 are all similar and
therefore only the fitting for 1 is described here in detail.
The best fit includes a combination of four nitrogen/
oxygen scatterers and a single sulfur scatterer attached
to the FeII center. In fits 1 and 2 in Table 1, the Debye-
Waller factors were fixed so as to determine the impor-
tance of inclusion of the sulfur scatterer. Clearly, the
single sulfur scatterer is necessary for a good fit. Debye-
Waller factors were then refined to reasonable values, as
seen in fit 3 (best fit, bold). The fitted model corresponds
to that expected from X-ray diffraction (XRD), with four
nitrogen donors from the [15]aneN4 ligand and the sulfur
donor from the phenylthiolate ligand comprising the first
coordination sphere around the metal. Fitting of the
outer-sphere features was accomplished with the addition
˚
of 2 carbon shells: 1 C scatterer at 3.43 A and 2 C scat-
˚
terers at 4.16 A, attributed to the C atoms of the phe-
nylthiolate ligand in proximity to the S donor. In addi-
˚
˚
tion, shells of 8 C atoms at 2.96 A and 3 C atoms at 3.26 A
were included and attributed to the N-C and N-C-C
atoms, respectively, of the [15]aneN4 ligand. Inclusion of
these outer-sphere shells, in which multiple scattering
paths were included, led to a significant improvement in
the fit quality (i.e., the F index) as compared to fits that
included only single scattering shells.
˚
is a peak with a maximum at R ∼ 2.1 A. There is also a
The Fe-S distances from the EXAFS for 1-3 in
CH2Cl2 solution are close to those found by XRD
(41) Roe, A. L.; Schneider, D. J.; Mayer, R. J.; Pyrz, J. W.; Widom, J.;
Que, L., Jr. J. Am. Chem. Soc. 1984, 106, 1676–1681.

9184 Inorganic Chemistry, Vol. 49, No. 20, 2010
Stasser et al.
Table 1. Least-Squares Fits of the EXAFS Dataa for the Iron(II) Complexes 1-3 in CH2Cl2
b
b
b
σ2c
coord no.
σ2c
coord no.
σ2c
E0
Fd
˚
Ras (A)
˚
Ras (A)
˚
Ras (A)
complex
fit
coord no.
1
1
2
4N
1S
2.00
2.31
2.01
2.31
2.01
2.00
2.29
2.02
2.29
2.02
2.00
2.31
2.03
2.31
2.03
2.0
2.0
4.0
2.8
4.8
2.0
2.0
4.0
3.2
4.6
2.0
2.0
4.0
3.0
4.5
8C
1C
8C
1C
8C
8C
1C
8C
1C
8C
8C
1C
8C
1C
8C
2.93
3.43
2.96
3.43
2.96
2.93
3.49
2.96
3.49
2.96
2.93
3.49
2.97
3.49
2.97
3.0
3.0
10.0
10.2
12.1
3.0
3C
2C
3C
2C
3C
3C
2C
3C
2C
3C
3C
2C
3C
2C
3C
3.27
4.16
3.26
4.16
3.27
3.28
4.22
3.27
4.22
3.27
3.28
4.22
3.33
4.22
3.33
5.0
5.0
8.0
12.3
12.9
5.0
5.0
8.0
11.9
13.2
5.0
5.0
8.0
-8.93
-6.93
2.13
1.64
4N
1S
4N
4N
1S
4N
1S
4N
4N
1S
3
-6.93
1.60
2
3
1
2
-8.95
-7.14
1.85
1.18
3.0
10.0
10.1
11.9
3.0
3
-7.14
1.08
1
2
-8.95
-7.37
2.11
1.45
3.0
4N
1S
4N
10.0
11.1
12.2
3
11.9
13.0
-7.14
1.43
P
P
2
c
2
3
a Best fits shown in bold. b Absorber-scatterer distance. Debye-Waller factor (A ) ꢁ 10 . d F ¼
k6i ½χexpðiÞ - χcalcðiÞꢀ = kj6jχexpðjÞj. The accuracy
˚
˚
in the determination of the bond lengths is approximately 0.02 A, and the precision is 0.004 A.
˚
i
j
˚
(2.3197(12)-2.3316(11)A), but the average Fe-N dis-
tances appear shorter by EXAFS (XRD: Fe-Nave
=
˚
2.109-2.195 A). The best fits of the EXAFS for the
crystalline samples for 1-3 (Supporting Information,
˚
Table S1) also yield Fe-Nave distances (2.06-2.12 A)
that are somewhat shorter than those found by XRD.
Although the EXAFS data for both the solution state and
the solid state samples can only be fit with N4S ligation,
the EXAFS-derived distances are significantly shorter in
˚
˚
solution (0.02-0.04 A for Fe-N, 0.05-0.07 A for Fe-S).
The origin of this discrepancy is not known. In any case,
the EXAFS corroborates the findings from the XANES,
which indicate that there is a structural rearrangement
for the FeII complexes upon dissolution. Attempts to fit
the solution state EXAFS in Figure 5 with a sixth low Z
scatterer at bonding distance to the iron, such as a weakly
bonded BF4- anion or exogenous solvent molecule (e.g.,
H2O), led to poor agreement with the data and unrea-
sonably high Debye-Waller factors. Taken together,
the EXAFS spectra for 1-3 are in full agreement with
these complexes being 5-coordinate, and confirm that the
4N/1S coordination sphere remains intact in solution.
Alkylperoxo-Iron(III) Complexes. The EXAFS spectra
for the three t-butylperoxo-iron(III) complexes 1a-3a
together with the best fits are shown in Figure 6, and the
fitted parameters are given in Table 2. Fitting results were
similar for 1a-3a, so only 1a will be discussed in detail.
There are three prominent peaks apparent in the Fourier-
˚
transformed data at R ∼ 1.9, 2.1, and 3.0 A. The main
difference in the Fourier-transforms of the EXAFS for
1a-3a, as compared to their iron(II) precursors, is the
additional short shell near 1.9 A, which is attributed to
Figure 6. Fourier transforms of Fe K-edge EXAFS data and EXAFS
spectra (insets) in CH2Cl2, experimental data (solid line) and theoretical
fits (dotted line) for (bottom) [FeIII([15]aneN4)(SC6H5)(OOtBu)]BF4
(1a), (middle) [FeIII([15]aneN4)(SC6H4-p-Cl)(OOtBu)]BF4 (2a), and
(top) [FeIII([15]aneN4)(SC6H4-p-NO2)(OOtBu)]BF4 (3a).
˚
the addition of the t-butylperoxo ligand. Accordingly, the
best fits to the data for 1a-3a require 3 shells of scatterers
˚
consisting of one N/O scatterer at 1.82-1.85 A, 4 N/O
˚
this gave a worse fit. Slight improvements to the fit (∼2%
decrease in F ) were seen if the population of the S
scatterer was allowed to vary between 0.5-1.0. Less than
full occupancy of the S donor can be expected given the
likelihood of minor decay of these species during sample
preparation because of their extreme thermal instability.
Similar modest improvements to the fit were seen upon
scatterers at 2.06-2.10 A, and one sulfur scatterer at
˚
2.29-2.32 A. The innermost shell can be attributed to the
O atom of the alkylperoxo ligand, while the next shell
corresponds to the four nitrogen atoms of the macrocyclic
[15]aneN4 ligand in 1a-3a. The S donors from each of
the arylthiolate ligands make up the third shell. Attempts
were made to fit the data in the absence of the sixth S
donor (Table 2, fit 1), in the absence of the alkylperoxo
ligand (Table 2, fit 2), or with a lower Z scatterer (N/O) in
the sixth position (Table 2, fit 3), but in all of these cases
˚
allowing the O atom occupancy at ∼1.85 A to vary, also
consistent with possible minor decomposition of 1a
(compare Table 2, fits 4 and 5). Although the EXAFS

Article
Inorganic Chemistry, Vol. 49, No. 20, 2010 9185
Table 2. Least-Squares Fits of the EXAFS Dataa for the Iron(III)-OOtBu Complexes 1a-3a in CH2Cl2
b
b
b
˚
Ras (A)
˚
Ras (A)
˚
Ras (A)
complex
fit
1
coord no.
σ2c
coord no.
σ2c
coord no.
σ2c
E0
Fd
1a
1 O
4 N
1 S
4 N
2 O
4 N
1 O
1 S
1.83
2.07
2.30
2.06
1.87
2.17
1.85
2.30
2.07
1.85
2.29
2.07
1.85
2.29
2.07
1.84
2.30
2.10
1.82
2.32
2.06
1.0
4.0
2.0
4.0
8.0
4.0
1.0
2.0
4.0
1.0
2.0
4.0
2.6
6.2
4.6
3.5
7.2
4.2
4.1
9.0
4.0
-11.21
1.66
8 C
1 C
8 C
3.01
3.45
3.01
10.0
3.0
10.0
3 C
2 C
3 C
3.28
4.17
3.29
8.0
5.0
8.0
2
3
4
-12.29
-12.65
1.53
1.57
8 C
2.97
10.0
3 C
3.30
8.0
1 C
8 C
3.44
3.01
3.0
10.0
2 C
3 C
4.17
3.28
5.0
8.0
-10.29
-10.23
1.44
1.41
4 N
5
6
7
8
0.9 O
0.6 S
4 N
0.9 O
0.6 S
4 N
1 O
0.6 S
4 N
0.5 O
0.6 S
4 N
1 C
8 C
3.43
3.01
3.0
10.0
2 C
3 C
4.16
3.28
5.0
8.0
-10.23
-10.61
-9.94
1.41
1.63
1.44
0.6 C
8 C
3.43
3.01
11.3
11.9
1.2 C
3 C
4.16
3.28
14.3
13.8
2a
3a
0.6 C
8 C
3.40
3.02
9.6
11.6
1.2 C
3 C
4.12
3.28
12.9
12.5
0.6 C
8 C
3.42
2.96
10.9
12.3
1.2 C
3 C
4.14
3.27
12.5
13.3
P
P
2
c
2
3
a Best fits shown in bold. b Absorber-scatterer distance. Debye-Waller factor (A ) ꢁ 10 . d F ¼
k6i ½χexpðiÞ - χcalcðiÞꢀ = kj6jχexpðjÞj. The accuracy
˚
˚
in the determination of the bond lengths is approximately 0.02 A, and the precision is 0.004 A.
˚
i
j
data were measured to relatively high resolution, it is
difficult, in general, to reliably fit three independent shells
of nearest neighbors because of the possibility of inter-
ference between the different scattering shells. Thus, the
observation that the bond lengths are unchanged when
the Fe-O and Fe-S coordination numbers are allowed
to vary (compare fits 4 and 5) provides additional evi-
dence that all 3 shells are required. Only for 3a is there a
significant decrease in the refined Fe-O coordination
number, consistent with the suggestion from the XANES
that this complex may be slightly decomposed.
The inclusion of outer-sphere carbon shells led to a
further improvement of the fit. As found for 1-3, the
number and distance of the outer-sphere C scatterers
correspond to the carbon atoms closest to the N and
S donors. For all three complexes, the Debye-Waller
factors were allowed to vary after optimizing the fit for
coordination number, type of scatterer, and Ras values as
described above. As seen in the best fits 6-8, the Debye-
Waller factors refine nicely to reasonable values. Overall,
it is evident that EXAFS data for the three alkylperoxo
complexes 1a-3a require the presence of four nitrogen
donors from the tetraazamacrocycle together with one
oxygen and one sulfur donor, attributed to the alkylper-
oxo and thiolate ligands, respectively. Thus the EXAFS
data are in excellent agreement with the proposed struc-
tures of 1a-3a shown in Scheme 1.
are close to those predicted by DFT. These Fe-N dis-
tances are slightly elongated in comparison to other low-
spin FeIII complexes, where d(Fe-N) is usually closer to
˚
2.0 A. Thus steric constraints on the [15]aneN4 ligand
may cause the Fe-N bonds to be slightly elongated in
1a-3a. The Fe-O bond lengths from DFT and EXAFS
are in excellent agreement, and the Fe-S distances calcu-
lated by DFT are also reasonably close to those derived
from EXAFS, although the DFT distances are some-
˚
what longer by 0.07-0.08 A. Metal-ligand distances for
longer, weaker bonds are often overestimated by DFT.42
Given the observed conformational change for the
iron(II) complexes upon dissolution, it is worthwhile to
consider the possibility of “folding” of the macrocycle
in the alkylperoxo complexes, such that the alkylperoxo
and thiolate ligands are positioned cis to each other.
The [15]aneN4 ligand exhibits only a trans geometry for
all known trivalent metal complexes, unlike cyclam
([14]aneN4), which can in fact fold into a cis geometry.
Thus, given the spectroscopic data, trends in the EXAFS
(vide infra), and DFT-optimized structures, we conclude
that the alkylperoxo-iron(III) species are in a trans geo-
metry. We have commented in some detail on this issue
previously.26
An important finding from the EXAFS, and corrobo-
rated by the DFT calculations, is that the identity of the
arylthiolate ligand influences both the Fe-S and Fe-O
distances. As the para substituent of the arylthiolate
becomes more electron-withdrawing, the Fe-S distance
becomes longer (d(Fe-S): 3a>2a>1a), consistent with
the expected weakening of the Fe-S bond for a more
electron-poor thiolate donor. The opposite pattern is
seen for the Fe-O bond lengths measured by EXAFS
(d(Fe-O): 1a>2a>3a), which is also reflected in the
DFT calculations. Although the changes in the EXAFS
The optimized geometries of the three FeIII-OOR
complexes were calculated previously by DFT,26 and
the calculated iron-ligand bond distances are presented
in Table 3 for comparison with the EXAFS data. The
coordinates from the X-ray structures of the FeII pre-
cursors 1-3 were used as the initial models in the calcula-
tions, and the tBu group of the alkylperoxo ligand was
truncated to a methyl substituent to facilitate the compu-
tations. The optimized structure for the PhS- complex is
shown in Figure 7. The ground spin state for all three
complexes was found to be low-spin FeIII (S = 1/2), in
agreement with the EPR analysis of 1a-3a. As seen in
Table 3, the average Fe-N distances obtained by EXAFS
˚
Fe-O distance are small (ΔFe-O = 0.03 A) and close
˚
to the estimated accuracy of EXAFS (∼0.02 A), these
differences are significantly greater than the precision of
(42) Neese, F. J. Biol. Inorg. Chem. 2006, 11, 702–711.

9186 Inorganic Chemistry, Vol. 49, No. 20, 2010
Stasser et al.
Table 3. Comparison of Bond Distances for 1a-3a from EXAFS and DFT
comparison with the limited metrical information avail-
able for other FeIII-OOR complexes. The Fe-O distances
found for 1a-3a are significantly longer than almost all
other low-spin FeIII-OOR(H) complexes, which do not
contain thiolate donors and display d(Fe-O) = 1.76-
(B3LYP/6311G)
˚
˚
˚
Fe-N (A)
Fe-S (A)
Fe-O (A)
ν(Fe-O)
complex EXAFS DFT EXAFS DFT EXAFS DFT
cm-1
˚
1.81 A. The Fe-O distances for the low-spin alkylperoxo
1a
2a
3a
2.07
2.10
2.06
2.09
2.09
2.09
2.29
2.30
2.32
2.374
2.380
2.388
1.85
1.84
1.82
1.873
1.870
1.864
611
612
615
complexes 1a-3a are closer to those recently reported for
the high-spin FeIII-OOR species (d(Fe-O) = 1.86-
˚
1.96 A), where an elongation in metal-ligand distance
is expected compared to their low-spin analogues.24 The
only other low-spin FeIII-OOR(H) complex with a simi-
lar elongated Fe-O distance is þthe thiolate-ligated com-
plex [FeIII(SMe2N4(tren)(OOH)] , with d(Fe-O)=1.86 A
˚
derived from EXAFS.21 Thus thiolate ligation appears
to be critical for elongating and weakening the Fe-O
bond in this complex as well as in 1a-3a. Interestingly,
the thiolate donor is held cis to the (hydro)peroxo ligand
in [FeIII(SMe2N4(tren)(OOH)]þ, in contrast to the pro-
posed trans arrangement in 1a-3a, suggesting that per-
haps the presence of a strong anionic donor such as RS- is
the dominant factor in weakening the Fe-OOR(H)
interaction, while its relative positioning to the peroxo
binding site is of secondary importance. It should be
noted that a mutant form of superoxide reductase,
E114A-SOR, reacts with H2O2 to give an FeIII-OO(H)
adduct that has been characterized by XRD and exhibits
2
˚
an Fe-O distance of 2.0 A, in good agreement with the
model complexes.
Figure 7. Molecular structure of [FeIII([15]aneN4)(SPh)(OOMe)]þ
based on the optimized geometry from DFT (yellow, Fe; red, O; orange,
C; gray, H).
Reactivity of the Alkylperoxo-Iron(III) Species. Mono-
nuclear nonheme iron-(hydro)peroxo and alkylperoxo
complexes39,45-49 have been investigated as both electro-
philic and nucleophilic oxidants. The examination of such
reactivity has helped in delineating the possible roles that
such species might play in enzymatic or synthetic catalytic
reactions. Given this motivation, we tested the reactivity
of the alkylperoxo-iron(III) complexes toward organic
substrates to determine if they could function as either
electrophilic or nucleophilic oxidants (Scheme 2). The
nucleophilic substrates PPh3 and cyclooctene were added
to 1a at low temperature. These substrates are prone to
oxidation by electrophilic metal-oxo species, giving
OPPh3 or cyclooctene oxide, respectively. The burgundy
alkylperoxo complex 1a was generated in the usual man-
ner at low temperature, as seen by formation of the peak
at λmax 526 nm. After maximal formation of the peak
at 526 nm, excess PPh3 (∼50 equiv) or cyclooctene
(∼100 equiv) dissolved in CH2Cl2 was added slowly to
the reaction mixture at -78 °C, with care taken to avoid
warming the solution. The addition of either substrate did
not change the rate of self-decomposition of 1a. The
reaction mixtures were warmed to room temperature
and analyzed by GC to determine if substrate oxidation
occurs upon thermal decomposition of 1a. One pathway
for the decomposition of 1a might involve O-O cleavage
˚
EXAFS (∼0.004 A), and are consistent with the magni-
tude of the changes predicted by the DFT calculations.
Previously some of us had described a correlation
between the nature of the thiolate ligand and the Fe-O
stretching frequency (ν(Fe-O)) obtained from RR spec-
troscopy for 1a-3a.26 The ν(Fe-O) values of 1a-3a are
reproduced in Table 3 for comparison. On the basis of the
vibrational data, it was suggested that the strength of the
Fe-O bond increases as the thiolate ligand becomes a
weaker donor. However, the ν(Fe-O) vibration for other
FeIII-OOR complexes has been shown to be coupled to
other vibrational modes, preventing a straightforward
connection of bond strengths with vibrational frequen-
cies.24,36,43,44 For the FeIII-OOR complexes described
here, it was found that the isotopic shifts (Δν(Fe-16O/
18O)) for the 18O-labeled isotopomers matched those
predicted for simple diatomic oscillators, suggesting that
a direct correlation could be made between their vibratio-
nal frequencies and bond strengths. In the present study
we have demonstrated that the Fe-O distance, which is
necessarily related to bond strength, correlates with the
identity of the thiolate ligand as predicted by the vibra-
tional data. Thus, the EXAFS provides strong support
for the conclusion that a weakening of the Fe-S interac-
tion trans to the alkylperoxo ligand induces a concomi-
tant strengthening of the Fe-O bond.
(45) Seo, M. S.; Kamachi, T.; Kouno, T.; Murata, K.; Park, M. J.;
Yoshizawa, K.; Nam, W. Angew. Chem., Int. Ed. 2007, 46, 2291–2294.
(46) Park, M. J.; Lee, J.; Suh, Y.; Kim, J.; Nam, W. J. Am. Chem. Soc.
2006, 128, 2630–2634.
Further insight regarding the influence of a thiolate
donor on the Fe-O interaction can be seen from a
(47) Jensen, M. P.; Payeras, A. M. I.; Fiedler, A. T.; Costas, M.; Kaizer,
€
J.; Stubna, A.; Munck, E.; Que, L., Jr. Inorg. Chem. 2007, 46, 2398–2408.
(43) Lehnert, N.; Fujisawa, K.; Solomon, E. I. Inorg. Chem. 2003, 42,
469–481.
(48) Kaizer, J.; Costas, M.; Que, L., Jr. Angew. Chem., Int. Ed. 2003, 42,
3671–3673.
(44) Lehnert, N.; Ho, R. Y. N.; Que, L., Jr.; Solomon, E. I. J. Am. Chem.
Soc. 2001, 123, 12802–12816.
(49) Sobolev, A. P.; Babushkin, D. E.; Talsi, E. P. J. Mol. Catal. A 2000,
159, 233–245.

Article
Inorganic Chemistry, Vol. 49, No. 20, 2010 9187
Scheme 2. Reactivity Pattern of Thiolate-ligated FeIII-OOR Com-
plexes
substrate, cyclohexane carboxaldehyde (CCA). Thus the
large increase in the rate of self-decomposition of 1a with
aldehyde substrates suggest a reaction is occurring with
1a and these electrophilic substrates. However, analysis
of reaction mixtures by GC did not reveal the expected
deformylation products (acetophenone in the case of
2-PPA, or cyclohexenone in the case of CCA). These
results suggest that the alkylperoxide of 1a may function
in a nucleophilic capacity toward aldehyde substrates, but
the mechanism of substrate decomposition does not lead
to deformylation products. We note that even in the case
of heme- and nonheme (hydro)peroxo-iron species that
lead to deformylation products, the mechanism of defor-
mylation is not well understood. We also examined the
reactivity of one of the alkylperoxo complexes toward
proton donors (Scheme 2). The cumenylperoxo-iron(III)
to give a high-valent iron-oxo species (FeIV(O) or FeV-
(O)), which could potentially be intercepted by substrate
to give oxygenated product(s). For both the PPh3 and
cyclooctene reactions, no oxidation products (OPPh3 or
cyclooctene oxide) were observed over the background
reaction (substrate þ tBuOOH, in the absence of Fe
complex). Thus the alkylperoxo complex 1a is unreactive
toward these oxygen atom acceptors and does not seem to
function as an electrophilic oxidant. These results further
suggests that the thermal decomposition of 1a does not
proceed via a mechanism that leads to a significant build-
up of a high-valent Fe(O) complex.
The chloro-substituted complex 2a was employed to
test possible reactivity toward H-atom donors. The alkyl-
peroxo complex was generated at -78 °C as described
above, and then the H-atom donor 2,6-di-tert-butyl-
phenol or cyclohexadiene (CHD) was added in excess
(∼50 equiv) (Scheme 2). The addition of these potential
H-atom donor substrates did not cause a detectable
increase in the rate of disappearance of 2a at -78 °C as
monitored by UV-vis spectroscopy. Reaction mixtures
with CHD as substrate were warmed to room tempera-
ture and analyzed by GC, but the production of benzene
was not observed. These data support the conclusion that
the alkylperoxo complexes do not function as electro-
philic oxidants, and furthermore, do not decompose in
a way that leads to a build up of a reactive high-valent
Fe(O) species.
complex 1b was treated with HþBF4 Et2O (1 equiv) at
-
3
-78 °C, and the reaction was monitored by UV-vis spec-
troscopy in the usual manner. Upon injection of HþBF4
-
there was an immediate loss of the burgundy color and
decrease at 528 nm, resulting in a final brown solution.
This rapid acceleration of the self-decomposition of 1b in
the presence of Hþ is consistent with the alkylperoxo
species exhibiting nucleophilic character.
Decay Kinetics. The decay kinetics of the alkylperoxo
species 1a-3a and 1b-3b in CH2Cl2 at -78 °C are
complicated and do not follow a simple first-order pro-
cess. We therefore attempted to monitor the decay kine-
tics at a different temperature. Although the alkylperoxo-
iron(III) species decompose rapidly upon warming from
-78 °C to room temperature, we found that they could be
generated at -40 °C and have sufficiently long enough
lifetimes to allow us to monitor their decay by UV-vis
spectroscopy. As can be seen in Figure 8, the disappear-
ance of the six different alkylperoxo complexes 1a-3a
and 1b-3b were monitored by following the decrease in
the peaks between 519-530 nm. A smooth decay of these
bands was observed, together with the concomitant loss
of the red chromophore. The final spectrum in each case
corresponds to yellow-brown reaction mixtures that con-
tain unidentified iron(III) products. Significant efforts
have been made to crystallize these final product(s), but
we have been unable to obtain crystalline materials
suitable for characterization by X-ray diffraction.
It is known that heme and nonheme iron(III)-(hydro)-
peroxo complexes can function as nucleophilic oxidants
toward electrophilic substrates such as aldehydes.50,51 A
recent, related example of such reactivity comes from a
report on a structurally characterized tert-butylperoxo-
nickel(II) complex, which reacts with CO and substituted
benzaldehydes to yield a carbonato-bridged nickel dimer
or the corresponding benzoato complex, respectively.52
Interestingly, addition of 2-phenyl propionaldehyde
(2-PPA) in excess (200 equiv) to 1a at -78 °C resulted
in an immediate color change from burgundy to yellow,
accompanied by a loss of the peak for 1a at 526 nm
(Scheme 2). A similar rapid color change and loss of the
526 nm peak was seen upon addition of a second aldehyde
The natural log plots (insets, Figure 8) are linear to ∼3
half-lives, with the exception of 3b, which was monitored
for ∼1.5 half-lives. Thus the decay kinetics at -40 °C
follow good first-order behavior, in contrast to what is
seen at -78 °C, where the decay curves do not follow
simple first-order behavior. These data suggest that there
is one dominant decay pathway at -40 °C, whereas at
-78 °C there may be competing mechanisms of decay.
The slopes of the best-fit lines of the natural log plots in
Figure 8 yield first-order decay rate constants (kdec) for
each of the complexes as follows: kdec = 3.6(1) ꢁ 10-2 s-1
(1a), kdec = 7.5(2) ꢁ 10-3 s-1 (2a), kdec=7(2) ꢁ 10-3 s-1
(3a), kdec=1.6(2) ꢁ 10-2 s-1 (1b), kdec=4.0(7) ꢁ 10-3 s-1
(2b), and kdec = 2.0(3) ꢁ 10-3 s-1 (3b). The rate con-
stants for both the FeIII-OOtBu and the FeIII-OOCm
complexes follow a clear trend in relation to the nature
of the thiolate ligand trans to the alkylperoxo moiety,
showing that the alkylperoxo complexes are significantly
stabilized as the para substituent on the thiolate ligand
(50) Wertz, D. L.; Valentine, J. S. Struct. Bonding (Berlin) 2000, 97, 37–60.
(51) Annaraj, J.; Suh, Y.; Seo, M. S.; Kim, S. O.; Nam, W. Chem.
Commun. 2005, 4529–4531.
(52) Hikichi, S.; Okuda, H.; Ohzu, Y.; Akita, M. Angew. Chem., Int. Ed.
2009, 48, 188–191.

9188 Inorganic Chemistry, Vol. 49, No. 20, 2010
Stasser et al.
Figure 8. Time-resolved UV-visible spectra for the decay of [FeIII([15]aneN4)(SC6H4-p-X)(OOtBu)]BF4 (1a, 1.9 mM; 2a, 1.9 mM; 3a, 1.4 mM) (left)
and [FeIII([15]aneN4)(SC6H4-p-X)(OOCm)]BF4 (1b, 2.6 mM; 2b, 2.0 mM; 3b, 1.3 mM) (right). From top to bottom, X = H, Cl, NO2. Inset: plot of
ln[(At - Af)/(A0 - Af) versus time (open circles), and the best fit of the data (black line), giving slope = kdec
.
increases in its electron-withdrawing character. These
effects are highlighted by comparing the half-life (t1/2
stretching frequencies than their tert-butylperoxo-iron-
(III) analogues (Fe-OOCm: ν(Fe-O)=622-631 cm-1
)
;
of the most stable species 3b (t1/2 =347 s) which carries
the p-NO2-substituted thiolate, to the unsubstituted phe-
nylthiolate ligand in 1a which exhibits t1/2 = 43 s.
The trend observed for the rates of self-decomposition
versus the identity of the thiolate ligand for both the tert-
butylperoxo- and the cumenylperoxo-iron(III) complexes
is in excellent agreement with the vibrational data, EX-
AFS, and DFT calculations. Interestingly, a comparison
of the rate constants between the tert-butylperoxo and the
cumenylperoxo complexes also exhibits a pattern that
corresponds to our previous results from RR studies. The
cumenylperoxo complexes exhibit higher energy Fe-O
Fe-OOtBu: ν(Fe-O) = 608-623 cm-1). The rates of
decay for each of the cumenylperoxo complexes are
slower than the corresponding tert-butylperoxo com-
plexes, correlating nicely with the difference in ν(Fe-O).
There is limited information on the kinetics of decay
of FeIII-OOR species, but two important studies are
relevant to the current work. The kinetics of decay for a
low-spin FeIII-OOtBu complex, [FeIII(TPA)(OOtBu)-
(NCCH3)]2þ (TPA = tris(pyridyl-2-methyl)amine) was
measured at -36 °C in CH3CN and found to accelerate
in the presence of Lewis bases such as 4-substituted pyri-
dines and pyridine N-oxides. An enhanced effect was

Article
Inorganic Chemistry, Vol. 49, No. 20, 2010 9189
noted upon increasing the electron-donating ability of the
added Lewis base.48 The UV-vis spectral decay of the
alkylperoxo complex was accompanied by the appear-
ance of a spectrum indicative of an FeIV(O) complex,
suggesting a decay mechanism involving homolytic clea-
vage of the O-O bond of the alkylperoxo ligand. Thus in
this case the Lewis bases likely coordinate to the metal
and facilitate O-O bond cleavage through a “push
effect”. In contrast, the rates of decay for a series of
high-spin FeIII-OOtBu complexes [FeIII(L8Py2)(OOt-
Bu)(X)]þ, where X = OTf-, C6H5CO2-, 4-CH3-C6H5S-,
and pyridine-N-oxide, showed the opposite trend, in
which the stronger Lewis base X correlated with a slower
rate of decomposition.53 These data were rationalized via
a mechanism of decay involving homolytic Fe-O bond
cleavage to give FeII, which would be destabilized by
electron-donating ligands.
decomposition via this mechanism. Thus we suggest that
these thiolate-ligated FeIII-OOR complexes decay via
release of ROOH, in analogy to the mechanism for SOR.
However, the final fates of the alkyperoxo ligands or the
iron complexes are not known at this time. We previously
examined the decay products by EPR spectroscopy, but
did not observe any new high-spin or low-spin FeIII
species.25 Therefore the initial iron decay product may
decompose rapidly to EPR-silent species. More studies
are warranted to probe further details of the decay
mechanism.
Summary and Conclusions
The alkylperoxo-iron(III) complexes 1a- 3a were char-
acterized by XAS studies and compared with their iron(II)
precursors. The XANES spectra reveal edge positions that
correspond to FeII for 1-3 and FeIII for 1a- 3a, and pre-edge
features for 1a and 2a are consistent with conversion from
5- to 6-coordinate upon going from FeII to FeIII, although
for 3a changes in the pre-edge features are more complex.
Analysis of the EXAFS spectra show that 1a-3a are 6-
coordinate, and confirm the first coordination sphere is
composed of N4SO ligation. These data provide the first
direct structural characterization of these metastable species.
The EXAFS-derived bond lengths reveal that the Fe-O
bonds in thiolate-ligated 1a-3a are significantly elongated
compared to other low-spin FeIII-OOR complexes that do
not contain thiolate ligands. In addition, there is a subtle
trend observed for the Fe-O bond distances versus identity
of the thiolate donor, in which the Fe-O bond shortens
as the thiolate donor becomes less electron-rich. This trend
is in agreement both with the vibrational data from RR
spectroscopy and with the optimized geometries obtained
from DFT.
The reactivity of the alkylperoxo-iron(III) complexes was
examined with various substrates. It was shown that the
alkylperoxo complexes are not competent electrophilic oxi-
dants, failing to react with the nucleophilic substrates PPh3 or
cyclooctene. In contrast, addition of electrophilic substrates
such as aldehydes or Hþ greatly accelerated the rate of decay
of the alkylperoxo species. This observation suggests that
the alkylperoxo species function as nucleophilic oxidants, con-
sistent with the nucleophilic behavior of other metal-peroxo
complexes.50-52 These studies also suggest that thiolate-
ligated 1a-3a do not decompose via O-O bond cleavage
to give high-valent iron-oxo species. These results contrast
previous work on non-thiolate-ligated, low-spin, alkylperoxo-
iron(III) complexes, which decompose via homolytic O-O
bond cleavage to give FeIV(O) complexes.39,47,48 Further-
more, an increase in the electron-releasing nature of the
thiolate ligands in 1a-3a correlates with faster rates of decay
for these species, in agreement with a weakening of the Fe-O
bond as seen by RR and now EXAFS studies, and supported
by DFT. Taken together with the substrate reactivity, these
studies suggest that the thiolate-ligated alkylperoxo-iron(III)
complexes decompose via a mechanism dominated by Fe-O,
and not O-O, bond cleavage, and this decay pathway may
be facilitated by the presence of the thiolate donors. These
results are interesting in light of the proposals for the mech-
anism of SOR, which must release H2O2 via Fe-O bond
cleavage and avoid O-O cleavage. Some evidence, including
recent DFT calculations,54 suggests that an FeIII-OO(H)
For our own system we considered four possible decay
pathways (eqs 1-4):
FeIII- OOR f FeIVðOÞ þ •OR
FeIII- OOR f FeII þ •OOR
ð1Þ
ð2Þ
ð3Þ
ð4Þ
-
FeIII- OOR f FeVðOÞþ OR
-
FeIII- OOR f FeIIIþ OOR
The first pathway, eq 1, is seen for the previously
mentioned low-spin FeIII-OOR species, but is unlikely
for 1a-3a and 1b-3b because we do not observe the rise
of a peak between 700-900 nm that is characteristic of
an FeIV(O) complex. In addition, the lack of any evidence
of oxygen-atom-transfer or hydrogen-atom abstraction
chemistry during the decomposition of these species in the
presence of substrates such as PPh3 or CHD strongly
argues against the generation of an FeIV(O) complex.
Homolytic cleavage of the Fe-O bond would give FeII
(eq 2) as described for the high-spin FeIII-OOR species,
but in this case the trend in decomposition rates versus the
nature of the thiolate donor should be reversed. Thus our
data support either eq 3 or eq 4 as the likely mechanism of
decay for 1a-3a and 1b-3b. In eq 3, heterolytic O-O
bond cleavage gives an FeV(O) complex. We have seen no
spectroscopic evidence for the production of an FeV(O)
complex, although such a species may be too short-lived
to be observed. However, it is expected that an FeV(O)
species would be a powerful oxidant, and should react
rapidly with exogenous substrates such as PPh3 or cyclo-
octene. The observed lack of reactivity with these sub-
strates argues against the generation of an FeV(O) species.
The mechanism of decay in eq 4 involves heterolytic
Fe-O bond cleavage, releasing the alkylperoxide ligand
as either an anion (-OOR), or perhaps as the protonated
hydroperoxide (HOOR) with proton donation from
exogenous H2O or ROH. This mechanism is consistent
with the kinetic data, and matches nicely with trends in
the Fe-O bond lengths from EXAFS. The longer, and
therefore weaker, Fe-O bonds should favor enhanced
(53) Bukowski, M. R.; Halfen, H. L.; van den Berg, T. A.; Halfen, J. A.;
Que, L., Jr. Angew. Chem., Int. Ed. 2005, 44, 584–587.

9190 Inorganic Chemistry, Vol. 49, No. 20, 2010
Stasser et al.
intermediate in SOR should be high-spin, and the high-spin
state may help to favor release of H2O2.27 However, definitive
evidence regarding the structure and spin-state of peroxo
intermediate(s) in the native reaction of SOR with superoxide
remains to be obtained. The active site structure of SOR may
be carefully tuned by the choice of ligands (His4Cys), and in
particular the presence of the thiolate donor, to favor Fe-O
over O-O cleavage, producing H2O2 instead of a high-valent
Fe(O) species.
GM38047 to J.E.P.-H.). Portions of this research were
carried out at the Stanford Synchrotron Radiation Light-
source, a national user facility operated by Stanford
University on behalf of the U.S. Department of Energy,
Office of Basic Energy Sciences. The SSRL Structural
Molecular Biology Program is supported by the Depart-
ment of Energy, Office of Biological and Environmental
Research, and by the National Institutes of Health,
National Center for Research Resources, Biomedical
Technology Program.
Acknowledgment. This work was supported by the
National Institutes of Health (GM62309 to DPG and
Supporting Information Available: Additional EXAFS data
(Table S1) and XANES and EXAFS spectra (Figures S1-S3).
This material is available free of charge via the Internet at http://
pubs.acs.org.
(54) Surawatanawong, P.; Tye, J. W.; Hall, M. B. Inorg. Chem. 2010, 49,
188–198.