The reaction of [Fe(pic)3] with hydrogen peroxide: A UV-visible and EPR spectroscopic study (Hpic = picolinic acid)

Article abstract of DOI:10.1039/b504897d

The Gif family of catalysts, based on an iron salt and O₂ or H₂O₂ in pyridine, allows the oxygenation of cyclic saturated hydrocarbons to ketones and alcohols under mild conditions. The reaction between [Fe(pic)₃] and hydrogen peroxide in pyridine under GoAgg^III (Fe(III)/Hpic catalyst) conditions was investigated by UV-visible spectrophotometry. Reactions were monitored at 430 and 520 nm over periods ranging from a few minutes to several hours at 20 °C. A number of kinetically stable intermediates were detected, and their relevance to the processes involved in the assembly of the active GoAgg^III catalyst was determined by measuring the kinetics in the presence and absence of cyclohexane. EPR measurements at 110 K using hydrogen peroxide and t-BuOOH as oxidants were used to further probe these intermediates. Our results indicate that in wet pyridine [Fe(pic)₃] undergoes reversible dissociation of one picolinate ligand, establishing an equilibrium with [Fe(pic) ₂(py)(OH)]. Addition of aqueous hydrogen peroxide rapidly generates the high-spin complex [Fe(pic)₂(py)(η¹-OOH)] from the labilised hydroxy species. Subsequently the hydroperoxy species undergoes homolysis of the Fe-O bond, generating HOO^. and [Fe(pic) ₂(py)₂], the active oxygenation catalyst. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b504897d

View Article Online / Journal Homepage / Table of Contents for this issue
F U L L P A P E R
The reaction of [Fe(pic)3] with hydrogen peroxide: a UV-visible and
EPR spectroscopic study (Hpic = picolinic acid)†
Sneh L. Jain* and Pravat Bhattacharyya
School of Chemistry, University of St Andrews, St Andrews, Fife, Scotland, UK KY16 9ST.
E-mail: slj2@st-andrews.ac.uk; Fax: (+44)-1334-463808
Received 11th April 2005, Accepted 22nd June 2005
First published as an Advance Article on the web 7th July 2005
The Gif family of catalysts, based on an iron salt and O2 or H2O2 in pyridine, allows the oxygenation of cyclic
saturated hydrocarbons to ketones and alcohols under mild conditions. The reaction between [Fe(pic)3] and hydrogen
peroxide in pyridine under GoAggIII (Fe(III)/Hpic catalyst) conditions was investigated by UV-visible
spectrophotometry. Reactions were monitored at 430 and 520 nm over periods ranging from a few minutes to
several hours at 20 ◦C. A number of kinetically stable intermediates were detected, and their relevance to the
processes involved in the assembly of the active GoAggIII catalyst was determined by measuring the kinetics in the
presence and absence of cyclohexane. EPR measurements at 110 K using hydrogen peroxide and t-BuOOH as
oxidants were used to further probe these intermediates. Our results indicate that in wet pyridine [Fe(pic)3] undergoes
reversible dissociation of one picolinate ligand, establishing an equilibrium with [Fe(pic)2(py)(OH)]. Addition of
1
aqueous hydrogen peroxide rapidly generates the high-spin complex [Fe(pic)2(py)(g -OOH)] from the labilised
hydroxy species. Subsequently the hydroperoxy species undergoes homolysis of the Fe–O bond, generating HOO• and
[Fe(pic)2(py)2], the active oxygenation catalyst.
precatalyst in GoAggIII, rather than [Fe(pic)3]. Barton’s synthesis
Introduction
of [Fe(pic)2Cl2]− is difficult to reproduce under the reported
The catalytic functionalisation of C–H bonds in alkanes sustains
conditions, although we have recently published a more efficient
extensive and ongoing interest worldwide, with systems mimick-
procedure.17 Nevertheless [Fe(pic)3] is readily prepared and func-
ing the dioxygen activation induced by iron-containing biolog-
tional in GoAggIII catalysis. When hydrogen peroxide is added
ical systems such as bleomycin, cytochrome P450 and methane
to [Fe(pic)3] in pyridine, rapid formation (t1/2 ≈ 1 min at 20 ◦C)
monooxygenase receiving particular attention.1 One attractive
of a purple solution occurs (kmax = 530 nm, e ≈ 1000 M−1 cm−1),
group of catalysts for the oxygenation of cycloalkanes to ketones
characteristic for the ligand–metal charge-transfer absorption
and alcohols is the Gif family of oxidants,2–5 within which GifIV
of a non-haem iron(III) hydroperoxy species.18 That this species
[a heterogeneous mixture of an iron(II) salt, dioxygen, acetic
is the predominant complex in solution is consistent with the
acid and zinc dust in pyridine] is the most efficient manifold,
results from numerous studies on the interaction of hydrogen
with turnover numbers for ketonisation exceeding 2000. The
peroxide with iron(II) and iron(III) precursors, which show that
mechanism of Gif systems has, since their inception, provoked
formation of the hydroperoxy species is very facile.18 Within Gif
fierce debate; the current consensus of opinion favours a radical-
chemistry iron(III) hydroperoxy species are the likeliest source of
based pathway, displacing Barton’s original proposal of iron(V)
the oxygen-based radicals mediating the oxidation process. Iron–
hydroperoxy complexes have been characterised by a battery
oxo intermediates.2,3
The GoAggIII system,6,7 comprising an iron(III) salt, picolinic
of spectroscopic techniques, but kinetic investigations of the
acid (Hpic) and aqueous hydrogen peroxide is a homogenous
processes leading to their formation and decomposition are rare.
counterpart of GifIV. Hpic is a powerful promoter of Gif
Herein we detail our kinetic studies on the construction and
chemistry, far more so than other nitrogen-containing het-
decay of the intermediates generated when aqueous hydrogen
eroaromatics, and the following observations are notable: (a)
peroxide is added to [Fe(pic)3] in pyridine.
Hpic reduces the requirement for pyridine and affords a higher
conversion rate; (b) the ideal ratio for efficient ketonisation
Experimental
of cycloalkanes is iron(III) : Hpic : hydrogen peroxide =
Reagents and solvents (Aldrich, AR grade) were used as sup-
plied, synthetic and spectroscopic procedures were conducted
under aerobic conditions. [Fe(pic)3] was prepared in methanol
from Hpic and iron(III) nitrate nonahydrate. Stock solutions of
[Fe(pic)3] in pyridine (0.5 mM) were protected from light and
used within three days of preparation. After this time solutions
darkened due to unidentified reactions and were unsuitable for
kinetic studies. Aqueous 30% w/w hydrogen peroxide was found
to be 9.46 M upon standardisation with cerium(III) sulfate
or potassium permanganate solutions.19 X-Band EPR spectra
were recorded in 4 mm o.d. tubes on a Bruker EMX 10/12
spectrometer operating at 9.5 GHz with 100 kHz modulation.
The absorbance of a 3.0 cm3 aliquot of [Fe(pic)3] solution
over the range 300–900 nm with varying concentrations of
hydrogen peroxide in 1◦cm path length quartz cuvettes was
measured at 20.0 0.1 C using a Shimadzu W-2101 PC UV-
vis spectrophotometer and a thermostatically controlled Lauda
RM6 water-bath.
1 : 4 : 4; (c) with two equivalents of Hpic to iron salt, in
situ dioxygen production (catalase activity) is suppressed.8–10
Pyridine can largely be replaced by acetonitrile,11 but cannot
entirely be removed without substantial loss of catalytic activity.
This reflects its importance in stabilising iron intermediates by
coordination, as well as its involvement in the radical chemistry
which underpins the Gif reaction. The use of GoAggIII catalysis
to selectively oxidise complex natural products underlines both
the mildness and versatility of this system.12–15
Barton’s investigations of GoAggIII using quantitative 13C
NMR spectroscopy8 confirmed an earlier result,16 that the
reaction between iron(III) hexahydrate chloride and Hpic in
pyridine–acetic acid generates [Fe(pic)2Cl2]−, mooted as the true
† Electronic supplementary information (ESI) available: Tables S1 and
S2: Kinetic data for the [Fe(pic)3]–hydrogen peroxide system. Fig. S1:
k2(obs) vs. [H2O2] plot. See http://dx.doi.org/10.1039/b504897d
2 6 9 6
D a l t o n T r a n s . , 2 0 0 5 , 2 6 9 6 – 2 7 0 1
T h i s j o u r n a l i s
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 5
©

View Article Online
Kinetic runs, studied by conventional means under pseudo
first-order conditions with >40 equivalents of hydrogen peroxide
(0.016–0.124 M) over [Fe(pic)3] (0.5 mM), were initiated by
injection of 5–40 ll aliquots of hydrogen peroxide into the
[Fe(pic)3] solution placed in the thermostatted cuvettes. Data
were averaged over two runs. The reaction of [Fe(pic)3] with 165
equivalents of hydrogen peroxide was initially monitored at 5 ◦C
in order to obtain time-resolved spectra at intervals of 2 min
(Fig. 1). The yellow solution rapidly became dark brownish-
purple (kmax 530 nm), after 12 min a second process was evident,
characterised by an isosbestic point at 520 nm. Thereafter a
further rise in absorbance over the entire wavelength range
characterised much slower changes, which at 20 ◦C were ac-
companied by evolution of dioxygen, resulting in a deep yellow–
brown solution after 12 h. For the detailed kinetic study, 430 nm
(significantly away from the isosbestic point) and 520 nm wave-
lengths were chosen for monitoring. A preliminary investigation
revealed that the rapid initial changes surrounding the isosbestic
point could be followed at 20 ◦C using time intervals of 2 s.
of cyclohexane (0.5 equivalents relative to hydrogen peroxide),
were investigated at 20 ◦C. Absorbance–time plots as a function
of [H2O2] at the isosbestic point (520 nm) over a 10 min
period appear in Fig. 2. There is a delay of approximately 5 s
between the addition of hydrogen peroxide and commencement
of the spectroscopic measurement. The effect on A0 is most
pronounced at higher concentrations, however the intercept on
the absorbance axis is not critical to the determination of the
pseudo-first-order rate constant k1(obs)
.
Fig. 2 Absorbance–time trace, [Fe] = 0.5 mM, [H2O2] = 0.016–0.124 M
(20 ◦C, 520 nm). Increasing [H2O2] leads to higher absorbance.
The curves in Fig. 2 have first-order kinetic profiles over
the first 10 min, from which k1(obs) for the formation of the
initial iron(III) product can be calculated. In the presence
of 0.5 equivalents of cyclohexane, absorbance–time traces at
520 nm when hydrogen peroxide was added showed the same
characteristics (Fig. 3) i.e. the profiles were largely unaffected
by addition of the hydrocarbon. The absorbance rise at 520 nm
for runs with and without cyclohexane were fitted to the single
exponential function [eqn. (1)], allowing k1(obs) to be obtained as
a function of [H2O2] (Fig. 4). The extent of absorbance change
increased with [H2O2], with evidence of saturation in A∞ at
higher concentrations. Taken together, these findings suggest
a first-order reversible equilibration reaction [eqn. (3)] with
the slope of the line in Fig. 4 characterising the forward rate
constant (k1) and the intercept, the rate constant for the reverse
process (k−1). Values of k1 and k−1 in the presence and absence
of cyclohexane (Table 1) demonstrate that adding the substrate
during this stage of the reaction leads to some diminution of the
rate (ca. 35%).
Fig. 1 Time-resolved spectra (Dt = 2 min) for [Fe(pic)3] (0.5 mM) with
165 equivalents of hydrogen peroxide in pyridine at 5 ◦C.
Absorbance–time plots at 430 and 520 nm were fitted to either
one or two successive exponentials using a non-linear least-
R
squares fitting procedure provided by GRAFITꢀ.20 For the fit
to a single exponential, eqn. (1) was employed:
At = A[1 − exp(–kt)] + A0
(1)
k
1
[Fe(pic)3] + H2O2 −ꢀ 1 (kmax530 nm)
ꢁ−
(3)
At is the absorbance at time t, A and k represent respectively
the extent of absorbance change and rate constant for the single
stage and A0 is the absorbance at t = 0. For the fit to two
exponentials, eqn. (2) was employed;
k
−1
During retention of the isosbestic point a second process is
seen, which results in fading of the absorption maximum at
530 nm for the first intermediate to give a shoulder at ca. 500 nm
At = A1[1 − exp(–k1t)] + A2[1 − exp(–k2t)] + A0
(2)
A1 and A2 represent the extent of absorbance change for the
first and second stages, respectively, k1 and k2 are the successive
first-order rate constants. Since the end of the second stage is
not fully defined due to subsequent reactions, evaluation of the
parameters in eqn. (2) was achieved by first fixing values for A1
and k1 based on data gathered at 520 nm, wherein only the first
stage contributes, and letting the program calculate values for
A2, k2 and A0 representing the best fit to the data (lowest value
of v2). The fixed values of A1 and k1 were then each varied in
succession until v2 was further minimised. These data appear in
Tables S1 and S2 (ESI).†
Results and discussion
Kinetic measurements
Fig. 3 Absorbance–time trace with cyclohexane (0.006–0.062 M)
added: [Fe] = 0.5 mM, [H2O2] = 0.016–0.124 M (20 ◦C, 520 nm).
Increasing [H2O2] leads to higher absorbance.
The kinetics of the reaction of [Fe(pic)3] with varying con-
centrations of hydrogen peroxide, in the absence or presence
D a l t o n T r a n s . , 2 0 0 5 , 2 6 9 6 – 2 7 0 1
2 6 9 7

View Article Online
Table 1 Rate constants for [Fe(pic)3]-hydrogen peroxide system
Data collected at 520 nm
Constant
Cyclohexane absent
Cyclohexane present
Data collected at 430 nm
k1/M−1 s−1
k−1/s−1
(9.46 0.32) × 10−2
(3.59 0.24) × 10−3
—
(8.08 0.69) × 10−2
(5.50 0.54) × 10−3
—
(9.74 0.66) × 10−2
(4.39 0.52) × 10−3
(1.00 0.05) × 10−4
k2/s−1
to 2) from the data obtained at 430 nm appear in Table 1. Our
kinetic data are consistent with eqns. (3) and (4).
k
2
1 ꢀ 2
(4)
The k2 value (1.0 × 10−4 s−1) is smaller in magnitude
than for the decomposition of iron(III) hydroperoxy com-
plexes bearing 4,5-pinene-2,2ꢀ-bipy or N-methyl-N,Nꢀ,Nꢀ-tris(2-
pyridylmethyl)propane-1,3-diamine co-ligands (2.0 × 10−3 s−1
at 0 ◦C in acetonitrile and 2.7 × 10−3 s−1 at 20 ◦C in methanol,
respectively),21,22 which are themselves consistent with k for the
unimolecular decomposition of FeIII–OOH in aqueous solution
(2.7 × 10−3 s−1).23 For data collected in the absence of cyclohex-
ane, k1 (ca. 9 × 10−2 M−1 s−1) is an order of magnitude smaller
than for a similar system,22 and contrasts markedly with values
for the [Fe(pic)2]–hydrogen peroxide reaction in pyridine–acetic
acid (2.0 × 103 and 2.0 × 102 M−1 s−1 when the iron : H2O2 ratios
are 1 : 1 and 1 : 200, respectively).24 Our k−1 (ca. 4 × 10−3 s−1)
agrees with data for an iron(III) hydroperoxy complex with an N-
benzyl-N,Nꢀ,Nꢀ-tris(2-pyridylmethyl)ethane-1,2-diamine ligand
(5.0 × 10−3 s−1 at 20 ◦C).25
Fig. 4 kobs vs. [H2O2] for the [Fe(pic)3]–hydrogen peroxide system
at 520 nm in the presence and absence of cyclohexane (triangles
represent data with cyclohexane added, filled circles are data without
cyclohexane).
and a general increase in the absorbance below 520 nm. To
obtain data characterising this second process the reaction was
monitored at 430 nm over 30 min with several hydrogen peroxide
concentrations (Fig. 5). For each run a rapid rise over the first
2 min is followed by a slower increase over the next 10 min,
subsequently the behaviour becomes more complex. The initial
two stages could be fitted to two successive exponentials using
eqn. (2). The first stage was found by the iterative treatment to
correspond to the same process which resulted in the formation
and retention of the isosbestic point. This evidence of a first-
order equilibration was found to depend linearly on [H2O2]. The
slower second process, however, was shown to be independent of
[H2O2]. For both processes the extent of absorbance change (A∞
value) increased with [H2O2] but showed evidence of saturation
at higher concentrations. These are the classic hallmarks of
equilibration processes. Rate constants k2(obs) obtained at 430 nm
as a function of [H2O2] (Table S2, ESI†) clearly illustrate the
independence of the second process on this concentration (Fig.
S1, ESI†). Thus it can be concluded that the isosbestic point
involves a further reaction taking place once the initial reaction
with hydrogen peroxide is complete. Values for k1 and k−1 (for
the first stage leading to 1) and k2 (for the second stage leading
Talsi et al. have proposed that in pyridine K[FeII(pic)3]
irreversibly loses Hpic to form [Fe(pic)2(OH)(py)].26 To ascer-
tain whether [Fe(pic)3] underwent similar changes, runs were
conducted in the presence of 10 equivalents of Hpic; should Hpic
dissociation occur prior to rate-determining formation of 1, a
decrease in k1(obs) would be expected at each hydrogen peroxide
concentration. The results (Fig. 6) show that the initial stage of
the reaction clearly is much slower cf. Fig. 2, with the rate of
the second stage largely unaffected. The effect of added Hpic
upon the kinetics strongly suggests that [Fe(pic)3] equilibrates
with an iron(III)-bis(picolinate) complex, which by analogy with
Talsi’s work is likely to be [Fe(pic)2(py)(OH)], formed by rapid
pyridine/water (HO−) coordination. This complex is a strong
candidate for the species reacting with hydrogen peroxide in the
rate-determining step to give 1 [eqns. (5) and (6)].
py,H
O
2
[Fe(pic)3] −−ꢀ [Fe(pic)2(py)(OH) + Hpic]
(5)
ꢁ−−
K
k
1
[Fe(pic)2(py)(OH)] + H2O −ꢀ 1 + H2O
2 ꢁ−
(6)
k
−1
Fig. 5 Absorbance–time trace, [Fe] = 0.5 mM, [H2O2] = 0.016–
0.124 M, (20 ◦C, 430 nm). Increasing [H2O2] leads to increased
absorbance.
Fig. 6 Absorbance–time trace, [Fe] = 0.5 mM, [H2O2] = 0.016–
0.124 M, with 10 equivalents of Hpic to iron (20 ◦C, 430 nm). Increasing
[H2O2] leads to higher absorbance.
2 6 9 8
D a l t o n T r a n s . , 2 0 0 5 , 2 6 9 6 – 2 7 0 1

View Article Online
To establish whether the retardation induced by added Hpic
was a general acid phenomenon, kinetic runs were performed
with 10 equivalents of p-toluenesulfonic acid added to the
solutions. A similar retardation occurred, albeit with a com-
pletely different absorbance–time profile, which could reflect
protonation of the pyridine and suppression of its ability to
displace Hpic in the initial step. Barton’s quantitative 13C
NMR experiments showed that [Fe(pic)3] did not undergo Hpic
dissociation in pyridine.8
EPR measurements (vide infra) confirm that upon dissolution
of [Fe(pic)3] in pyridine chemical change occurs but that
mononuclearity is retained i.e. there is no conversion into a
l-oxo bridged diferric complex. As the rate constants for the
second stage of the reaction were independent of [H2O2] and
[Hpic], this implies the conversion of one bis(picolinate) complex
1 to another 2. The saturation in A∞ as [H2O2] increases (Table
S1, ESI†) indicates that this reaction, as with the first stage,
is reversible. The A∞ values for the second stage are, however,
only estimates since after 10–30 min subsequent reactions occur,
whose onset is dependent on [H2O2].
When the reaction time exceeds 30 min a complex series of
reactions occurs, shown in Fig. 7 for [H2O2] = 0.042 and 0.126 M
with cyclohexane present and absent. At each concentration of
hydrogen peroxide, the addition of cyclohexane, added prior to
the H2O2, markedly decreased the overall absorbance during
a 10 h period. The rate reduction upon cyclohexane addition
indicates that the substrate reacts with the species responsible
for the absorbance at this time. After approximately 6 h
the curves for the 0.042 M runs intersect the traces for the
0.126 M aliquots, which reach a plateau after 3–4 h, leading
to higher absorbances for the former. We infer that there may
be precipitation of very minor quantities of iron(III) hydroxy
species at the higher peroxide concentration, although we have
obtained no direct evidence for this [n.b. formation of iron(III)
oxide/hydroxide particles was observed by Schuchardt27 during
GoAggIII cyclohexane oxygenation]. Indeed, for the 0.042 M
solution, the absorbance continues to rise steadily, even after
10 h, due to further reactions of the more soluble iron species.
Scheme 1
To parallel our UV-vis study, an approximately 0.3 cm3 aliquot
was taken from 3.0 cm3 of [Fe(pic)3] in pyridine (0.5 mM),
transferred to an EPR tube and dipped into a vessel containing
5 : 1 v/v isopentane–methylcyclohexane, itself immersed in
liquid nitrogen. This rapid freezing technique produced more
clearly resolved spectra compared to when solely liquid nitrogen
was used for cooling. The frozen solution was transferred to
the spectrometer, whose cavity was maintained at 110 K. The
X-band spectrum of the frozen glass showed a signal at g = 4.4
(Fig. 8(a)), compatible with the ms = 1/2 Kramers doublet of
a high-spin (S = 5/2) iron(III) centre in a rhombically distorted
environment, both in zero-field splitting (ZFS) and geometry.
The rhombicity arises from the mixed ligands and the constraints
imposed through the chelate effect with the picolinates. The
complexity of the spectrum suggests that several isomeric high-
spin species are formed, conceivably arising from reactions of
[Fe(pic)3] with pyridine or water involving reversible loss of pic−,
consistent with eqn. (5) (K ꢁ 1).
Fig. 7 Absorbance–time trace, [Fe] = 0.5 mM, [H2O2] = 0.042 and
0.126 M (20 ◦C, 520 nm). Dotted line = 0.126 M, full line =
0.042 M. Within each pair of curves the lower line represents when
cyclohexane is added.
Fig. 8 X-Band EPR spectra at 110 K for (a) 0.5 mM solution of
[Fe(pic)3] in pyridine. (b) Solution (a) with 170 equivalents of hydrogen
peroxide after 2 min (purple species) at room temperature then refrozen.
(c) Solution (b) warmed for a further 2 min and refrozen. (d) Solution
(c) warmed for a further 2 min and refrozen. All spectra are shown to
the same scaling along the absorbance axis.
Scheme 1 summarises the reaction of [Fe(pic)3] with hydrogen
peroxide in pyridine.
EPR Analysis
1
Talsi et al. used EPR and H NMR spectroscopies to analyse
The EPR tube was removed from the spectrometer, thawed
completely and its contents combined with the stock solution.
An aliquot of hydrogen peroxide (20 ll, ca. 170 equivalents
over iron) was added, giving a purple colour. After 2 min
a portion of the solution was transferred into a clean tube,
frozen as before and its EPR spectrum recorded (Fig. 8(b)).
An isotropic g = 4.4 signal appeared, whose sharpness and
intensity suggests an increase in molecular symmetry in the
the K[FeII(pic)3]–hydrogen peroxide system;26 in their study
[FeIII(pic)2(OH)(py)] (g = 4.25) was produced when 1 equivalent
of hydrogen peroxide was added to a 2 : 1 v/v pyridine–acetic
acid solution of K[Fe(pic)3]. Addition of a large excess of hydro-
gen peroxide generated a high-spin iron(III) hydroperoxo com-
plex, purportedly [Fe(pic)2(py)(OOH)]. However no EPR study
has been made of the corresponding reaction with [Fe(pic)3].
D a l t o n T r a n s . , 2 0 0 5 , 2 6 9 6 – 2 7 0 1
2 6 9 9

View Article Online
oxidised complex compared to [Fe(pic)3] and the presence of a
single major high-spin component. The appearance of only one
resonance upon hydrogen peroxide addition is consistent with
several related studies into the formation of iron–hydroperoxy
intermediate.18 Two consecutive cycles of removal from the
spectrometer, thawing for 2 min and refreezing (Fig. 8(c) and
(d)) suggest an initial increase and subsequent decrease in the
g = 4.4 signal, however double integration gave inaccurate
values of signal intensity as the baselines were insufficiently flat.
In Fig. 8(b) and (c) a low-field shoulder at the main signal
arises from the rhombicity parameter k ≈ 0.17 (k = E/D,
where D and E are the axial and rhombic ZFS parameters,
respectively) rather than k = 1/3, which is the maximum possible
rhombicity in ZFS. Whereas iron(III) hydroperoxy complexes
bearing polydentate aminopyridyl co-ligands are exclusively
low-spin,18 the more oxygen-rich environment about the iron
core in 1 favours a high-spin configuration, only the second such
example,28 which has important ramifications for its subsequent
reactivity (vide infra).
Further clues to the nature of 2 (and hence 1) came from
EPR studies of the reaction of [Fe(pic)3] in pyridine with t-
BuOOH, a reagent which can replace hydrogen peroxide in
Gif (GoAgg) chemistry and preserve the main reactivity profile.
The [Fe(pic)3]–t-BuOOH reaction was studied under analogous
conditions to the hydrogen peroxide experiment, i.e. with 170
equivalents of hydroperoxide. A deep purple solution forms
rapidly at 20 ◦C upon t-BuOOH addition to [Fe(pic)3]. The
g = 2 signal in the EPR spectrum of the frozen glass after
2 min of reaction (Fig. 9(b)) is due to t-BuOO•, consistent
with Fe–O cleavage of [Fe(pic)2(py)(OOBu-t)], producing the
EPR-silent high-spin iron(II) complex [Fe(pic)2(py)2]. The as-
signment is supported by the observations that treatment of
[Fe(pic)3] with hydrogen peroxide generates [Fe(pic)2(py)2],29 and
that K[Fe(pic)3] converts into [Fe(pic)2(py)2] when dissolved in
pyridine.26 The EPR silence of the complex arises from the large
ZFS, causing the potentially allowed transitions to be at too
high an energy to be excited by X-band microwave frequency.
Further corroboration is provided by Que’s findings that the
Fe–O bond cleaves in high-spin iron(III) butylperoxyl complexes
with 6-methyl substituted tris(2-pyridylmethyl)amine ligands,
whereas O–O homolysis occurs in low-spin analogues.30 The
g value for the iron(III) butylperoxyl species in Fig. 9(b) is
consistent with Que’s high-spin complexes (g = 4.3),31 although
such nitrogen-rich coordination environments typically favour
low-spin configurations.32–34 As in Fig. 8, the feature to low-field
of the g = 4.4 signal could belong to the main absorption with
differences in the ZFS parameters, though still with significant
rhombicity.
Since t-BuOOH and hydrogen peroxide give similar purple
species it is concluded, by extrapolation, that in the latter reac-
tions 1 is the iron(III) hydroperoxy complex [Fe(pic)2(py)(OOH)]
and 2 is [Fe(pic)2(py)2] along with HOO•; this radical species,
unlike t-BuOO•, has a short half-life in these solutions and
rapidly disproportionates to hydrogen peroxide and dioxygen.
Indeed the evolution of dioxygen is a feature of these solutions
as they warm to room temperature over several minutes.
Furthermore, despite the basicity of the solvent, we exclude the
2
possibility of an Fe(g -O2) or Fe2(l-O2) structure for 1, as the
EPR of the t-BuOOH reaction was essentially identical to
the H2O2 study.
Summary and conclusions
Barton has written that “. . . the hydrocarbon oxidation mech-
anisms of, say, Fe(PA)3 [PA = picolinate] compared with
FeCl3·6H2O + three picolinic acid could be different, as they
differ in their starting materials. . .”.10 Although it is difficult to
disagree with this statement, some similarities are evident. Since
[Fe(pic)3] is readily synthesised and does catalyse cycloalkane
oxygenation under GoAggIII, we undertook a kinetic study of
its chemistry with hydrogen peroxide.
The [Fe(pic)3]–hydrogen peroxide reaction in pyridine was
monitored using UV-vis and EPR spectroscopies. Upon dis-
solution of [Fe(pic)3] in non-dried pyridine, an equilibrium
reaction involving displacement of one picolinate ligand by
solvent and HO−/H2O is suggested by the dependence of the
kinetics on added Hpic. On addition of aqueous hydrogen
peroxide the formation of a hydroperoxy complex 1 takes place,
followed by a slower transformation, the generation of the active
catalyst 2 from 1. The rate constants for the formation and
subsequent reactions of 1 or 2 are typically smaller than in
iron(III) hydroperoxy complexes with polyamine ligands, due
to the N,O-donor environment. Over longer timeframes the
presence of cyclohexane dramatically affects the absorbance of
the solution, as 2 catalyses the oxygenation of the substrate.
The EPR spectrum of [Fe(pic)3] in pyridine at 110 K contains a
complex signal at g = 4.4, consistent with equilibration mixtures
of high-spin tris- and bis-(picolinate) iron(III) species, as detected
in our kinetic studies. The production of an isotropic signal
at g = 4.4 on reaction with hydrogen peroxide reflects the
presence of a single high-spin complex, the decrease of signal
intensity with time signifying conversion to an EPR-silent S =
2 component. t-BuOOH gives almost identical spectra under
these conditions, suggesting that similar high-spin FeIII–OOR
(R = H or t-Bu) complexes are formed. We conclude that 1 and
1
2 are [Fe(pic)2(py)(g -OOH)] and [Fe(pic)2(py)2], respectively.
It is feasible that pyridine could deprotonate the hydroperoxy
2
complex to give [Fe(pic)2(g -O2)]−,16 particularly as iron(III) per-
1
oxo complexes are high-spin,18,35,36 moreover g -hydroperoxo ↔
2
g -peroxo interconversions can be effected by acid/base
introduction.25,37 However, the similarity between the hydrogen
peroxide and t-BuOOH EPR experiments firmly refute that
deprotonation is occurring. That Fe–O bond cleavage occurs
in our high-spin FeIII–OOR species (R = H or t-Bu) mitigates
against the formation of high-valent iron–oxo intermediates in
GoAggIII oxidation using [Fe(pic)3], although such species are
readily invoked from low-spin precursors.
The formation of iron(II) species prior to any detectable
influence of added cyclohexane accords with the current body of
opinion, that Fenton/Haber–Weiss reactions involving radical
formation from iron(II)–hydrogen peroxide chemistry underpins
the mechanism of oxygenation by Gif systems. The involvement
of pyridine in stabilising bis(picolinate) intermediates in addition
to intercepting HO• makes it an integral component of the Gif
(GoAgg) protocol.
Fig. 9 X-Band EPR spectra at 110 K for (a) 0.5 mM solution of
[Fe(pic)3] in pyridine. (b) Solution (a) with 170 equivalents of t-BuOOH
after 2 min (purple species) at room temperature then refrozen.
2 7 0 0
D a l t o n T r a n s . , 2 0 0 5 , 2 6 9 6 – 2 7 0 1

View Article Online
18 J. J. Girerd, F. Banse and A. J. Simaan, Struct. Bonding (Berlin),
2000, 97, 146.
Acknowledgements
We acknowledge Dr W. John Ingledew, Dr Rona R. Ramsay and
Prof. John C. Walton (University of St Andrews) for assistance
with spectrometric measurements. We thank the EPSRC (P. B.)
and BBSRC (S. L. J.) for financial support and Dr David Collison
(The University of Manchester) for very helpful discussions.
19 A. I. Vogel, Textbook of Quantitative Inorganic Analysis, Longman,
London and New York, 4th edn., 1978.
20 GRAFITꢀR is a trademark of Erithacus Software, Leatherhead,
Surrey, UK. The package uses a standard Marquardt algorithm for
non-linear least-squares fitting.
21 Y. Mekmouche, H. Hummel, R. Y. N. Ho, L. Que, V. Schu¨nemann,
F. Thomas, A. X. Trautwein, C. Lebrun, K. Gorgy, J.-C. Lepreˆtre,
M.-N. Collomb, A. Deronzier, M. Fontecave and S. Me´nage, Chem.
Eur. J., 2002, 1196.
22 V. Balland, F. Banse, E. Anxolabe´he`re-Mallart, M. Ghiladi, T. A.
Mattioli, C. Philouze, G. Blondin and J. J. Girerd, Inorg. Chem.,
2003, 42, 2470.
23 J. DeLaat and H. Gallard, Environ. Sci. Technol., 1999, 33, 2726.
24 C. Sheu, S. A. Richert, P. Cofre´, B. Ross, A. Sobkowiak, D. T. Sawyer
and J. R. Kanofsky, J. Am. Chem. Soc., 1990, 112, 1936.
25 A. Hazell, C. J. McKenzie, L. P. Nielsen, S. Schindler and M. Wietzer,
J. Chem. Soc., Dalton Trans., 2002, 310.
26 A. P. Sobolev, D. E. Babushkin, A. A. Shubin and E. P. Talsi, J. Mol.
Catal. A, 1996, 112, 253.
27 U. Schuchardt, C. E. Z. Kra¨hembu¨hl and W. A. Carvalho, Stud. Surf.
Sci. Catal., 1994, 82, 647.
28 A. Wada, S. Ogo, S. Nagamoto, T. Kitagawa, Y. Watanabe, K.
Jitsukawa and H. Masuda, Inorg. Chem., 2002, 41, 616.
29 S. Kiani, A. Tapper, R. J. Staples and P. Stavropoulos, J. Am. Chem.
Soc., 2000, 122, 7503.
30 N. Lehnert, R. Y. N. Ho, L. Que and E. I. Solomon, J. Am. Chem.
Soc., 2001, 123, 12802.
31 Y. Zang, J. Kim, Y. Dong, E. C. Wilkinson, E. H. Appelman and L.
Que, J. Am. Chem. Soc., 1997, 119, 4197.
32 M. Rivera, G. A. Caignan, A. V. Astashkin, A. M. Raitsimring, T. K.
Shokhireva and F. A. Walker, J. Am. Chem. Soc, 2002, 124, 6077.
33 A. P. Sobolev, D. E. Babushkin and E. P. Talsi, J. Mol. Catal. A, 2000,
159, 233.
References
1 M. Costas, M. P. Mehn, M. P. Jensen and L. Que, Chem. Rev., 2004,
104, 939.
2 P. Stavropoulos, R. C¸ elenligil-C¸ etin and A. E. Tapper, Acc. Chem.
Res., 2001, 34, 745.
3 F. Gozzo, J. Mol. Catal. A, 2001, 171, 1.
4 D. H. R. Barton, Tetrahedron, 1998, 54, 5805.
5 D. H. R. Barton and D. Doller, Acc. Chem. Res., 1992, 25, 504.
6 D. H. R. Barton, E. Csuhai and D. Doller, Tetrahedron Lett., 1992,
33, 3413.
7 D. H. R. Barton, B. Hu, D. K. Taylor and R. U. R. Wahl, Tetrahedron
Lett., 1996, 37, 1133.
8 D. H. R. Barton, B. Hu, R. U. R. Wahl and D. K. Taylor,
New. J. Chem., 1996, 20, 121.
9 E. About-Jaudet, D. H. R. Barton, E. Csuhai and N. Ozbalik,
Tetrahedron Lett., 1990, 31, 1657.
10 D. H. R. Barton, B. Hu, D. K. Taylor and R. U. R. Wahl, J. Chem.
Soc., Perkin Trans. 2, 1996, 1031.
11 D. H. R. Barton and T. Li, Tetrahedron, 1998, 54, 1735.
12 D. Doller, S. Chackalamannil, A. Stamford, B. McKittrick and M.
Czarniecki, Bioorg. Med. Chem. Lett., 1997, 7, 1381.
13 D. H. R. Barton and D. Doller, Collect. Czech. Chem. Commun.,
1991, 56, 984.
14 I. Okamoto, W. Funaki, K. Nakaya, E. Kotani and T. Takeya, Chem.
Pharm. Bull., 2004, 52, 756.
15 S. B. Kim, K. W. Lee, Y. J. Kim and S. I. Hong, Bull. Korean Chem.
Soc., 1994, 15, 424.
34 M. V. Lobanova, K. P. Bryliakov, E. A. Duban and E. P. Talsi,
Mendeleev Commun., 2003, 175.
35 O. Horner, C. Jeandey, J. L. Oddou, P. Bonville, C. J. McKenzie and
J. M. Latour, Eur. J. Inorg. Chem., 2002, 3278.
16 D. H. R. Barton, S. D. Beviere, W. Chavisiri, D. Doller, G. Liu and
J. H. Riebenspies, New J. Chem., 1992, 16, 1019.
17 S. L. Jain, P. Bhattacharyya, H. L. Milton, A. M. Z. Slawin and J. D.
Woollins, Inorg. Chem. Commun., 2004, 7, 423.
36 A. J. Simaan, F. Banse, J. J. Girerd, K. Wieghardt and E. Bill, Inorg.
Chem., 2001, 40, 6538.
37 K. B. Jensen, C. J. McKenzie, L. P. Nielsen, J. Z. Pedersen and H. M.
Svendsen, Chem. Commun., 1999, 1313.
D a l t o n T r a n s . , 2 0 0 5 , 2 6 9 6 – 2 7 0 1
2 7 0 1