Biomimetic oxidation with Fe-ZSM-5 and H2O2? Identification of an active, extra-framework binuclear core and an FeIII-OOH intermediate with resonance-enhanced Raman spectroscopy

Article abstract of DOI:10.1002/cctc.201402642

Developing inorganic materials that can mimic nature's ability to selectively oxidise inert C-H bonds remains a topic of intense scientific research. In recent years, zeolitic materials containing Fe and/or Cu have been shown to be highly active, heterogeneous catalysts for the selective oxidation of alkanes (including methane), amongst a range of other related oxidation challenges. By using resonance-enhanced Raman spectroscopy, we demonstrate that, following high-temperature pre-treatment (activation), Fe-containing ZSM-5 possesses an active binuclear core, and forms a key Fe-OOH intermediate upon activation with H₂O₂. Both factors are reminiscent of biological oxidation catalysts, and may account for the unique ability of this material to selectively oxidise methane to methanol at low temperature. A pre-ferryl cat: Through insitu resonance-enhanced Raman spectroscopy, we identify an active, binuclear Fe-O(H)-Fe core and an Fe^III-OOH intermediate in Fe-containing ZSM-5 following activation with H₂O₂. The pre-ferryl nature of this biomimetic intermediate may account for the unique ability of this solid catalyst to selectively oxidise methane to methanol under mild conditions.

Full text of DOI:10.1002/cctc.201402642

DOI: 10.1002/cctc.201402642
Communications
Biomimetic Oxidation with Fe-ZSM-5 and H O ?
2
2
Identification of an Active, Extra-Framework Binuclear
III
Core and an Fe ÀOOH Intermediate with Resonance-
Enhanced Raman Spectroscopy
Ceri Hammond,* Ive Hermans, and Nikolaos Dimitratos
[a]
[b]
[a]
Developing inorganic materials that can mimic nature’s ability
to selectively oxidise inert CÀH bonds remains a topic of in-
tense scientific research. In recent years, zeolitic materials con-
taining Fe and/or Cu have been shown to be highly active,
heterogeneous catalysts for the selective oxidation of alkanes
and oxo intermediates are formed, each possessing various
[4]
levels of oxidising ability, although much of the cycle can be
bypassed (shunted) by using more-activated oxygen donors,
[5]
such as H O .
2
2
Recently, a zeolite catalysts, CuFe-ZSM-5, was found to medi-
ate the selective and catalytic oxidation of methane to metha-
nol at high levels of activity and selectivity with H O as
(including methane), amongst a range of other related oxida-
tion challenges. By using resonance-enhanced Raman spectros-
copy, we demonstrate that, following high-temperature pre-
treatment (activation), Fe-containing ZSM-5 possesses an
active binuclear core, and forms a key FeÀOOH intermediate
upon activation with H O . Both factors are reminiscent of bio-
2
2
[6]
[7]
a green oxidant. In contrast to previous zeolitic systems,
this new peroxide-mediated system was both catalytic (turn-
over number >10000) and selective to the desired product,
methanol (selectivity >90%), even at high levels of methane
2
2
[6a]
logical oxidation catalysts, and may account for the unique
ability of this material to selectively oxidise methane to metha-
nol at low temperature.
conversion. Furthermore, in contrast to the most active and/
[1]
or selective methane oxidation catalysts reported to date,
this catalytic system was found to be highly active under
benign reaction conditions, and was able to catalyse the selec-
tive oxidation of methane to methanol in the aqueous phase,
in the absence of any acidic promoters, at temperatures be-
The activation and selective oxidation of CÀH bonds remains
one of the most elusive targets in catalysis, the most challeng-
ing reaction of this type being the selective oxidation of meth-
ane to methanol. Methane, the major constituent of natural
[6a]
tween 2 and 708C.
The key to this system appears to be the heterolytic activa-
tion of H O by active Fe species within the zeolitic micropores.
2
2
[
1]
gas, is a hugely abundant and promising feedstock but its
direct utilisation is hampered, given the difficulties associated
Through various spectroscopic and computational methods,
some of us recently demonstrated that the active Fe species
for this reaction are cationic, extra-framework Fe oligo-
mers, with the resting state of the catalyst most likely possess-
ing two m-hydroxo-bridged Fe atoms: [Fe (m -OH) (OH) -
with selectively upgrading this most unreactive of hydrocarbon
À1 [1]
substrates (DH =104 kcalmol ). Tantalisingly, naturally oc-
CÀH
curring metalloenzymes are able to perform this and related
oxidation reactions with ease even under mild reaction condi-
tions, with exquisite chemo- and stereoselectivity, and with
2
2
2
2
2
+ [6a,c]
(H O) ]
.
Upon activation by the oxidant, DFT calculations
2
2
predict that this resting state is converted into a bifunctional
active site possessing both iron–oxo and iron–hydroperoxo in-
[
2]
molecular oxygen as the sole, terminal oxidant. Typically,
these monooxygenase enzymes possess Fe or Cu active cen-
tres, and are able to oxidise substrates through the reductive
activation of dioxygen with stoichiometric reductive cofac-
[6a,b]
termediates.
Such intermediates are reminiscent of many
formed in biological systems, and raise the enticing prospect
that this catalytic system operates in a biomimetic peroxide
[
3]
[2]
tors. During this process, various superoxo, (hydro)peroxo
shunt-type process. Nevertheless, in situ characterisation of
the catalyst has not yet been achieved, without which defini-
tive identification of 1) the reactive intermediates, and 2) the
precise speciation of the active site(s) operating during the re-
action, cannot be obtained. This precludes further comparison
to biological oxidation catalysts.
[
a] Dr. C. Hammond, Dr. N. Dimitratos
Cardiff Catalysis Institute
Cardiff University
School of Chemistry, Main Building, Park Place
CF10 3AT, Cardiff (UK)
In contrast to well-defined homogeneous model catalysts,
the in situ spectroscopic analysis of a heterogeneous catalyst is
considerably more challenging because heterogeneous cata-
lysts do not contain a homogeneous distribution of a single
active site or species. Rather, a heterogeneous distribution of
species, each displaying various levels of activity and/or selec-
tivity, is typically found. As such, the percentage of active spe-
cies may be very low, and features arising from spectator spe-
E-mail: hammondc4@cardiff.ac.uk
[
b] Prof. Dr. I. Hermans
University of Wisconsin-Madison
Department of Chemistry
and Department of Chemical and Biological Engineering
1101 University Ave.
Madison, WI 53706 (USA)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201402642.
ChemCatChem 0000, 00, 0 – 0
1
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Communications
cies and/or competitive active sites may dominate or mask the
spectroscopic data of interest. Unfortunately, many spectro-
scopic techniques are unable to differentiate between the
active and spectator species, and therefore only provide an
averaged signal of many different species.
To overcome this limitation, we have extended our study of
this system by using resonance-enhanced UV/Raman spectros-
[
8]
copy. Recently, we demonstrated that the catalytic activity of
Fe-containing zeolites for methane oxidation correlated to the
fraction of Fe species experiencing electronic transitions be-
tween l( =250 and 350 nm, as determined by UV/Vis spec-
abs)
[
6d]
troscopy.
Indeed, more recent kinetic studies demonstrate
that the initial rate of methane oxidation is intimately linked to
the fraction of these extra-framework species, which are
formed through high-temperature pre-treatment of the as-syn-
thesised zeolite (Figure 1). By tuning the Raman excitation
Figure 2. Raman analysis of 0.5Fe-silicalite-1550 with A) 785, B) 514, and
C) 325 nm laser lines.
length into the UV region (l=325 nm) leads to four new
Raman-active bands that are not observed by using visible
(
l=514 nm) and IR (l=785 nm) Raman excitation sources.
Thus, in addition to the signals arising from the silicalite-
1
3
matrix, that is, heteroatom free MFI-type zeolite (at n˜ =290,
À1
80 and 800 cm ; Figure S1, Supporting Information), a rela-
À1
tively broad feature between n˜ =500 and 600 cm , with a max-
À1
imum at n˜ =521 cm , and additional features at n˜ =1013,
À1
1120 and 1165 cm are now observed. Previously, all of these
bands have been assigned to tetrahedrally coordinated, iso-
3
+
morphously substituted framework Fe species [Fe(OSiO ) ]
3
4
[9b]
[9c]
by Bonino et al. and Li and co-workers, suggesting that
3
+
Fe was present exclusively in the zeolite framework. Howev-
3
+
Figure 1. Initial rate of methane oxidation with various samples of A) 0.5Fe-
er, it is believed that such saturated framework Fe sites are
unable to coordinate reacting species, and are thus unable to
participate in catalytic reactions, which is seemingly at odds
with the known activity of this catalyst for methane oxidation
silicalite-1 and B) 0.5Fe-ZSM-5, each of which contained different amounts
3
+
3+
of extra-framework Fe . Increased formation of extra-framework Fe was
achieved by increasing the pre-treatment temperature employed (550–
9008C) and was determined by using UV/Vis spectroscopy.
[10]
and other reactions. Consequently, further Raman spectros-
3
+
copy analysis of the Fe speciation in Fe-containing zeolites
and particularly the evolution of these species as a function of
pre-treatment temperature, is required.
wavelength into this particular electronic transition feature
lexc =325 nm), it is possible to selectively enhance any Raman
(
3
+
vibrations associated with those electronic transitions, thereby
selectively enhancing only the species of interest, even if these
species only constitute a fraction of the total metal content.
It is well established that, despite the initial presence of Fe
in the framework of Fe-silicalite-1 and related materials after
hydrothermal synthesis, considerable changes to the specia-
3
+
Furthermore, given the low Raman cross-section of H O,
tion of Fe occur following the high temperature treatments
required to 1) remove the organic template (at 5508C), and
2) pre-activate the catalyst for reaction (typically at tempera-
2
Raman spectroscopy offers a convenient means of studying
the chemistry of this catalyst in situ, despite the presence of
the aqueous solvent.
[6c,11]
tures between 550 and 11008C).
In fact, recent studies by
Preliminary investigations focused on Fe-silicalite-1, an MFI-
type zeolite containing 0.5 wt% Fe and pre-activated at 5508C
our team and other research groups have found that, even
3
+
after removal of the template, a large percentage of Fe in
Fe-siliaclite-1 and Fe-ZSM-5 is located in extra-framework posi-
(denoted henceforth 0.5Fe-silicalite-1550), which is both an
[11]
active and selective catalyst for the oxidation of methane in
the absence of additional metal promoters. It is also an active
catalyst for other oxidation processes, such as that of benzene
tions of the zeolite material and is present typically as isolat-
ed or oligomeric (hydr)oxo clusters, or even bulk Fe oxides, de-
[6c]
pending on the exact procedure employed. To establish the
influence of this migration on the Raman spectrum of 0.5Fe-
silicalite-1, and to further corroborate the results obtained in
Figure 1, a series of catalysts at various stages of pre-treatment
were examined by using UV/Raman spectroscopy.
[
6a]
to phenol. Full characterisation of this material by conven-
tional methods [UV/Vis, X-ray absorption (XAS) and FTIR spec-
troscopies and XRD and BET analyses] was described recent-
[
6c]
ly. As can be seen in Figure 2, tuning the excitation wave-
&
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
2
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Communications
It is apparent at this stage that, despite the previous assign-
3
+
[9b,c]
ment of all four Fe bands to framework Fe species,
story is much more complex and the Raman spectrum of
.5Fe-silicalite-1 contains signals arising from multiple species.
We have recently established that, during the heat treatment
the
0
3
+
process, several changes in the speciation of Fe are observ-
[6c]
ed. Initially, there is a homogeneous distribution of frame-
3
+
work Fe after the synthetic procedure. During template re-
3
+
moval, a large amount of Fe migrates from the framework
and initially forms a broad distribution of isolated and/or oligo-
nuclear extra-framework cationic complexes within the zeolite
micropores, which are characterised by intense electronic tran-
sitions between l=250 and 350 nm. These species, which cor-
relate with catalytic activity for methane oxidation (Fig-
[6c,d]
ure 1),
are maximised following pre-treatment at 7508C,
Figure 3. Effect of pre-treatment on the Raman spectrum of 0.5Fe-silicalite-
. A) As-synthesised zeolite containing template, B) NH -form, C) H-form,
after calcination of NH -form at 5508C, D) H-form, after calcination of NH
form at 7508C and E) H-form, after calcination of NH -form at 9008C.
1
4
with increased temperatures thereafter (9008C) only causing
3
+
4
4
-
further agglomeration of Fe into bulk Fe O . To date, our
2
3
4
spectroscopic (XAS) and theoretical studies indicate that these
active, extra-framework oligomers are most likely hydroxo-
[6]
bridged binuclear Fe clusters, with the resting state of the
In good agreement with the known chemistry of these ma-
terials, considerable changes in the Raman spectra of 0.5Fe-sili-
calite-1 are observed during pre-treatment, suggesting major
changes in the speciation of Fe (Figure 3). The first stage in the
preparation of active samples of Fe-silicalite-1 involves the re-
moval of the organic template (tetrapropylammonium hydrox-
catalyst best described as a di-m-hydroxo-bridged Fe complex:
2
+ [6a,b]
[Fe (m -OH) (OH) (H O) ]
.
Notably, the self-organisation of
2
2
2
2
2
2
extra-framework metal complexes into bridged, binuclear com-
plexes was proposed recently by Pidko et al. to be spontane-
[13]
ous for many metallozeolites.
In light of these results, it is clear that the n˜ =1013 cm
band behaves as we would expect for a band arising from tet-
À1
ide) and conversion to the NH form. During this step, the vi-
4
3
+
brations arising from the template are lost, and the spectrum
already looks similar to that of 0.5Fe-silicalite-1550, possessing
rahedrally-coordinated framework Fe : during template re-
moval and pre-activation, the band is continually eroded as
3
+
the features of the silicalite-1 framework ( n˜ =290, 380 and
the fraction of framework Fe diminishes. The steady, albeit
À1
À1
8
00 cm ) and the Fe modes described above ( n˜ =521, 1013,
slower, erosion of the n˜ =1120 and 1165 cm bands indicates
À1
3+
1
120 and 1160 cm ).
that these bands are also related to the framework Fe com-
Although the vibrations of the organic template overlap
ponent. Their slower rate of erosion upon pre-treatment likely
3
+
with the Fe
features in the as-synthesised material (Fig-
relates to their (partial) reliance on neighbouring SiÀOÀSi vi-
[9c]
ure S2) and thus prohibit a full comparison, it is clear that the
brations; previous work has indicated that SiÀOÀSi vibra-
À1
3+
lower energy signal at n˜ =521 cm increases in intensity upon
tions in the vicinity of the framework Fe species partially
pre-treatment relative to the silicalite-1 framework modes.
Upon further pre-treatment of the catalyst, it is also clear that
drive this FeÀO vibration and thus account for some of its
spectral intensity. Nevertheless, the increase in the n˜ =
À1
À1
there is a steady increase in intensity of the n˜ =521 cm band,
521 cm band as a function of pre-treatment is contrary to
3
+
with the relative intensity of this band at its greatest following
pre-treatment at 7508C, after which the band broadens and
erodes. Conversely, it is also clear that with increasing pre-
what we would expect for framework Fe species. Indeed, its
steady increase in intensity, especially in comparison to the
negligible level found in the as-synthesised material, suggests
À1
3+
treatment temperature, the n˜ =1013 cm band is eroded
that this feature does not arise from framework Fe but from
À1
3+
steadily and experiences a red shift of Æ35 cm following
extra-framework Fe species, the proposed active species for
pre-treatment at 7508C. By 9008C, this band is lost almost
completely from the Raman spectrum. This indicates that the
species responsible for this signal decrease in concentration
with pre-treatment, which is in line with their assignment as
methane oxidation, which are known to increase in concentra-
[6c]
tion during heat treatment. Notably, this is the first systemat-
3
+
ic Raman spectroscopy study that follows the changes in Fe
speciation (framework vs. extra-framework) in Fe-containing
zeolites as a function of essential heat pre-treatment.
3
+
[9b,c]
framework Fe species.
The observed red shift also indi-
3
+
cates a relaxation in the coordinative strain of the Fe species
To further probe this hypothesis, 0.5Fe-ZSM-5550 [containing
0.5 wt% Fe and 0.6 wt% Al (SiO /Al O =90), pre-activated at
[
12]
responsible for this vibration, further indicating the removal
2
2
3
3
+
of Fe from the highly strained framework positions. Changes
550 8C] was studied subsequently by using UV/Raman spec-
troscopy. Recently, we established that this catalyst was sub-
stantially more active for methane oxidation than its Al-free an-
alogue 0.5Fe-silicalite-1550, and that it contained substantially
À1
in the position and intensity of the n˜ =1120 and 1165 cm vi-
brations are also evident and their steady decrease in intensity
suggests that these species are also related to tetrahedrally co-
3
+
3+
ordinated framework Fe species, in line with previous stud-
more extra-framework (and catalytically active) Fe species at
[
9b,c]
[6d]
ies.
any given pre-treatment temperature. In line with the hy-
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
3
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Communications
À1
3+
pothesis that the n˜ =521 cm band arises from the extra-
work Fe species, as proposed previously. In regards to the
3
+
À1
framework and catalytic Fe content of these materials, this
band is substantially more intense and prominent in 0.5Fe-
ZSM-5550 than in 0.5Fe-silicalite-1550 (Figure 4).
broadness of the n˜ =521 cm feature, which has a span of ap-
À1
proximately n˜ =100 cm , we propose that this catalyst likely
contains a number of bi- and/or oligonuclear Fe complexes of
various geometry and speciation, each of which would possess
slightly different stretching frequencies. Although this phe-
nomenon is unfortunately undetectable by using XAS, which
provides only an averaged signal of all the species present in
a heterogeneous catalyst, it further demonstrates the potential
of resonance-enhanced Raman spectroscopy for studying het-
erogeneous catalysts that contain complex active-site distribu-
tions.
To determine whether the species responsible for this
stretch is active in H O -based oxidations and if it may be the
2
2
active site for CÀH activation and methane oxidation, treat-
ment of the catalyst with aqueous H O was subsequently ex-
2
2
amined. As can be observed, this resulted in the complete loss
of the proposed FeÀO(H)ÀFe stretch from the Raman spectrum
and strongly suggested that this species was responsible both
for H O activation and subsequent methane oxidation (Fig-
2
2
À1
Figure 4. Comparison of A) 0.5Fe-ZSM-5 and B) 0.5Fe-silicalite-1, both after
calcination at 5508C. C) 0.5Fe-ZSM-5550 after treatment with H O .
2 2
ure 4C). The complete loss of the n˜ =521 cm vibration indi-
cated that, upon coordination and activation of H O , the
2
2
bridges between the two Fe atoms were cleaved, likely
through protonation following heterolysis of H O . This would
2
2
À1
3+
The assignment of the n˜ =521 cm band to extra-frame-
lead to the formation of two transient, mononuclear Fe sites,
which would operate in a cooperative manner in the remain-
ing catalytic cycle. Nevertheless, this cleavage of the
FeÀO(H)ÀFe core of the catalyst was reversible, as even moder-
ate irradiation under the UV laser regenerated the n˜ =
3
+
work Fe complexes is well-supported by Raman and IR spec-
troscopic analysis of model homogeneous compounds and
DFT calculations. For instance, the Raman-active (symmetric)
stretching mode of various bridged binuclear Fe complexes
À1[14]
À1
has been shown to appear between n˜ =400 and 600 cm
521 cm band to some extent, and reheating the sample to
and is known to be resonance-enhanced by low-wavelength
110 8C completely restored the same band (Figure S4). The
[
15]
lasers (l=300–400 nm). Indeed, Sanders-Loehr et al. identi-
fied that the exact location of this Raman active stretching
mode was highly dependent upon the precise FeÀO(H)ÀFe
shift between dimer and transient monomeric structure (and
3
+
vice versa) demonstrated the flexibility of the Fe sites for
mediating this complex catalytic mechanism. Similar “transient
self-organisation” also played a role in the activation of glucose
[
15]
angle of the binuclear Fe complex under study, with smaller
angles shifting the stretching frequency to higher energy. By
graphically plotting the data of Sanders-Loehr et al., a correla-
tion between the FeÀO(H)ÀFe angle and stretching frequency
2
+
[16]
by transient Cr dimers, as identified by Pidko et al.
Nevertheless, no new vibrations corresponding to potential
iron–(hydro)peroxo or iron–oxo complexes are observed in the
UV/Raman spectrum of H O /0.5Fe-ZSM-5550, preventing the
[
15]
was obtained. Our previous DFT and spectroscopic studies
have indicated that the potential active site for methane oxida-
tion is consistent with a di-bridged, m-hydroxo diiron complex,
2
2
acquisition of further mechanistic insight. This might be due to
the known instability of various metal–(hydro)peroxides and
metal–oxo species, which are notoriously difficult to character-
ise under the intense Raman beam. However, it is also possible
that the potential metal–oxygen intermediates formed in this
system do not undergo resonance enhancement under the
[6a,b]
possessing FeÀO(H)ÀFe angles between Æ95 and 1108.
comparison of this DFT-calculated FeÀO(H)ÀFe angle in 0.5Fe-
ZSM-55 with the systematic data of Sanders-Loehr et al. re-
A
[
6a,b]
50
veals that an FeÀO(H)ÀFe stretching frequency between n˜ =
À1
3+
5
39 and 580 cm should be observed in this material. As can
same conditions as the Fe species present within the Fe-con-
be seen, this is in excellent agreement with the experimental
data. To further validate this assignment, the Raman spectrum
taining zeolites themselves. Indeed, iron–oxo, iron–superoxo
and iron–(hydro)peroxo complexes typically experience intense
2
+
[17–21]
of a cation-exchanged [Fe (m -OH) (OH) (H O) ] complex was
electronic transitions in the visible and/or IR regions.
2
2
2
2
2
2
also calculated by DFT. These simulations predicted that two
Although shifting the excitation wavelength to the visible
À1
FeÀO(H)ÀFe stretching vibrations at 521 and n˜ =574 cm
region (lexc =514 nm) leads to the loss of Fe-specific modes at
À1
should be observed in 0.5Fe-ZSM-5550, in full agreement with
the experimental data (Figure S3 and Ref. [6b]). Thus, in combi-
n˜ =521, 1013, 1120 and 1160 cm through a lack of resonance
enhancement, two new H O -related features are observed
2
2
[
6a]
nation with our previous XAS results,
which imply that
under visible light excitation (Figure 5). The first of these fea-
a large fraction of dimeric Fe species is present in this materi-
tures arises from the n(OÀO) of physisorbed H O ( n˜ =
2
2
[
6a]
À1
À1
al, we propose that the n˜ =521 cm feature arises from bi-
nuclear Fe species and not from catalytically inactive frame-
874 cm ) and is unrelated to any Fe coordination (Figure 5C,
À1
Figure S1B). However, the vibration observed at n˜ =631 cm
&
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
4
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Communications
including the low-temperature oxidation of methane to metha-
nol. In the first part of this study, in depth characterisation of
these materials with resonance-enhanced Raman was per-
formed. Our full, systematic study of the heat pre-treatment
3
+
process has revealed that the active site speciation of Fe in
these materials is particularly complex and that a range of
3
+
both framework and extra-framework Fe species are present
in these materials. The precise ratio of these species is highly
dependent on the pre-treatment temperature. Notably, this
provides significantly greater levels of insight than previously
reported Raman studies of these materials, which did not fully
monitor the heat pre-treatment processes and consequently
3
+
proposed that a homogeneous distribution of framework Fe
sites was present. Most important amongst the Raman features
À1
Figure 5. Vis/Raman spectrum of A) 0.5Fe-ZSM-5550 B) H
and C) H /silicalite-1550
2
O
2
/0.5Fe-ZSM-5550
is a prominent, broad band at n˜ =521 cm , which correlates
2
O
2
.
favourably to the catalytic activity of these materials for meth-
ane oxidation. Comparison of the Raman spectra against litera-
ture studies of model compounds, our DFT calculations, our
previous X-ray absorption, and UV/Vis spectroscopy studies re-
veals that the species responsible for this feature is both extra-
framework and binuclear in nature. Treatment of the catalyst
with aqueous H O also reveals this species to be responsible
is consistent with a n(FeÀO) stretch, as demonstrated through
numerous sophisticated labelling studies, and is character-
[
18–21]
III
istic of model Fe ÀOOH complexes, such as [Fe(bbpc)-
(
(
[
(
MeCN)(OOH)]SbF6
(bbpc=N,N’-di(phenylmethyl)-N,N’-bis-
2
2
À1 [17]
2-pyridinylmethyl)-1,2-cyclohexanediamine, n˜ =632 cm ),
for H O coordination during methane oxidation.
2 2
2
+
À1 [19]
2+
(N Py)Fe(OOH)]
4
( n˜ =632 cm ),
[(H bppa)Fe(OOH)]
This work also provides the first in situ spectroscopic study
of the Fe-containing zeolites that are able to catalyse methane
oxidation in the aqueous phase. Our in situ Raman spectrosco-
py studies monitoring the treatment of Fe-ZSM-5 with H O
2
bppa=bis(2-picolyl)(2-hydroxy-3,5-di-tert-butylbenzyl)amide,
À1 [20]
2+
n˜ =621 cm )
and [(TMC)Fe(OOH)]
(TMC=1,4,8,11-tetra-
À1 [21]
methyl-1,4,8,11-tetraazacyclotetradecane, n˜ =658 cm ). This
2
2
III
indicates that an Fe ÀOOH intermediate is present in 0.5Fe-
(tert-butyl hydroperoxide) reveal that the analogous FeÀOOH
(FeÀOO(tBu)) intermediates are present during the reaction
cycle and that these intermediates are formed by heterolysis of
ZSM-55 following activation of the catalyst with aqueous
50
III
H O . Although the accompanying n(OÀO) stretch of Fe ÀOOH
2
2
À1
3+
is not discernible between n˜ =750 and 800 cm , this is likely
the peroxide by the binuclear Fe species. The identification
III
due to its extremely weak Raman scattering intensity com-
of an Fe ÀOOH intermediate in Fe-containing zeolites gives
[
18]
pared to the n(FeÀO) stretch and the unfortunate overlap
further support to our previously calculated reaction mechanis-
À1
[6a]
with the n˜ =800 cm vibration of the MFI framework, decreas-
m
and provides spectroscopic evidence that this catalytic
ing the detection limit.
system contains at least some similarities to biological oxida-
III
Nevertheless, to further verify the identity of the intermedi-
ate, we investigated the activation of other peroxides known
tion systems. Fe ÀOOH is known to be a key precursor to the
most potent CÀH activating oxidant in biological catalysts—
3
+ [22]
to undergo heterolysis in the presence of Fe
.
Upon substi-
highly electrophilic iron–oxo species—and its conversion to
tution of the H atom of H O for a tert-butyl group (tert-butyl
the iron–oxo moiety through OÀO homolysis or heterolysis is
2
2
À1
[2–4]
hydroperoxide), bands at n˜ = Æ495, 715 and 798 cm were
well-established.
We reason that the lack of evident signals
observed (Figure S5). These were consistent with those found
associated with the oxo species arises from 1) the overlap of
[
23]
À1
in other FeÀOO(tBu) complexes and confirmed the ability of
the Fe species in Fe-ZSM-5/Fe-silicalite-1 to mediate the forma-
tion of (hydro/alkyl)peroxides by heterolysis. These band have
been attributed to tert-butyl deformation containing considera-
the MFI framework feature at n˜ =800 cm and 2) its extremely
short lifetime under the conditions employed in this study (t<
[24]
7 s at 258C in water). Indeed, although H O is a relatively
2
benign solvent for catalytic applications, it severely restricts
the possibility of detecting the ultimate high-valent iron–oxo
species, which to date has been detected only at sub-ambient
À1
ble n(FeÀO) character ( n˜ =492 cm ) and the n(FeÀO) ( n˜ =
À1
À1
715 cm )
and
23]
n(OÀO)
stretches
( n˜ =798 cm )
of
[
[25]
FeÀOO(tBu).
Notably, these vibrations are not observed
temperatures with soluble catalysts. Identification of the elu-
upon treatment of Fe-free zeolites (such as silicalite-1 and H-
sive iron–oxo species, and detailed microkinetic analysis of the
reaction intermediates formed, are therefore required to gain
deeper insight into the potential biomimetic properties of this
catalytic system, and are the target of future research studies
with additional, advanced in situ spectroscopic techniques.
ZSM-5) with tert-butyl hydroperoxide, confirming that the pres-
3
+
ence of Fe is critical to the formation of this reactive inter-
mediate, in excellent agreement with our catalytic studies.
Conclusions
Fe-containing zeolites, such as Fe-silicalite-1 and Fe-ZSM-5, are
well-known catalysts for various selective oxidation processes,
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
5
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Communications
Rev. 1996, 96, 2841–2887; c) R. A. Himes, K. D. Karlin, Curr. Opin. Chem.
Biol. 2009, 13, 119–131.
Experimental Section
[
4] a) W. A. van der Donk, C. Krebs, J. M. Bollinger, Jr., Curr. Opin. Struct. Biol.
Catalyst preparation
2
010, 20, 673–683; b) B. J. Brazeau, R. M. Austin, C. Tarr, J. T. Groves,
Fe-containing ZSM-5 and silicalite-1 were prepared as described
J. D. Lipscomb, J. Am. Chem. Soc. 2001, 123, 11831–11837; c) C. Krebs,
D. G. Fujimori, C. T. Walsh, J. M. Bollinger, Jr., Acc. Chem. Res. 2007, 40,
484–492; d) J. T. Groves, J. Inorg. Biochem. 2006, 100, 434–447; e) C. E.
Tinberg, S. J. Lippard, Acc. Chem. Res. 2011, 44, 280–288.
[6a]
elsewhere. Tetraethylorthosilicate (Sigma–Aldrich, 99.999%, trace
metal basis), iron nitrate (Sigma–Aldrich, 98%) and sodium alumi-
nate (Sigma–Aldrich, 50–55 wt% Al O ) were used as the Si, Fe and
2
3
[
[
5] a) K. K. Andersson, W. A. Froland, S.-K. Lee, J. D. Lipscomb, New J. Chem.
Al sources, respectively. Tetrapropylammonium hydroxide (Sigma–
Aldrich, 20 wt% in water) was used as a structure-directing agent.
Crystallisation was performed in a stainless-steel autoclave at
1
991, 15, 410–415; b) Y. Jiang, P. C. Wilkins, H. Dalton, Biochim. Biophys.
Acta 1993, 1163, 105–112; c) A. A. Shteinman, Russ. Chem. Bull. 2001,
0, 1795–1810.
5
1
758C for 120 h. The as-synthesised materials were pre-treated at
3+
6] a) C. Hammond, M. M. Forde, M. H. Ab Rahim, A. Thetford, Q. He, R. L.
Jenkins, N. Dimitatos, J. A. Lopez-Sanchez, N. F. Dummer, D. M. Murphy,
A. F. Carley, S. H. Taylor, D. J. Willock, E. E. Stangland, J. Kang, H. Hagen,
C. J. Kiely, G. J. Hutchings, Angew. Chem. Int. Ed. 2012, 51, 5129–5133;
Angew. Chem. 2012, 124, 5219–5223; b) C. Hammond, R. L. Jenkins, N.
Dimitatos, J. A. Lopez-Sanchez, M. H. Ab Rahim, M. M. Forde, A. Thet-
ford, D. M. Murphy, H. Hagen, E. E. Stangland, J. M. Moulijn, S. H. Taylor,
D. J. Willock, G. J. Hutchings, Chem. Eur. J. 2012, 18, 15735–15745; c) C.
Hammond, N. Dimitatos, R. L. Jenkins, J. A. Lopez-Sanchez, S. A. Kon-
drat, M. H. Ab Rahim, M. M. Forde, A. Thetford, S. H. Taylor, H. Hagen,
E. E. Stangland, J. M. Moulijn, D. J. Willock, G. J. Hutchings, ACS Catal.
various temperatures (5508C–9008C) to induce migration of Fe
sites to the extra-framework, following prior occlusion of the or-
ganic template and ion exchange with NH NO .
4
3
Catalyst testing
Catalytic evaluation was performed in a 50 mL stainless-steel auto-
clave (working volume 35 mL). The vessel was charged with an
aqueous solution of H O (10 mL, 0.5m) and the desired amount of
2
2
2
013, 3, 689–699; d) C. Hammond, N. Dimitatos, R. L. Jenkins, J. A.
catalyst (27 mg). The reactor was pressurised with methane
Lopez-Sanchez, G. Whiting, S. A. Kondrat, M. H. Ab Rahim, M. M. Forde,
A. Thetford, H. Hagen, E. E. Stangland, J. M. Moulijn, S. H. Taylor, D. J.
Willock, G. J. Hutchings, ACS Catal. 2013, 3, 1835–1844.
(
(
3.05 MPa) and the autoclave heated to the reaction temperature
508C) with stirring at 1500 rpm once the desired temperature was
obtained. Following the appropriate reaction time, the vessel was
[7] a) P. Vanelderen, R. G. Hadt, P. J. Smeets, E. I. Solomon, R. A. Schoon-
heydt, B. F. Sels, J. Catal. 2011, 284, 157–164; b) V. I. Sobolev, K. A.
Dubkov, O. V. Panna, G. I. Panov, Catal. Today 1995, 24, 251–252.
cooled in ice (128C) and the resultant solution filtered and ana-
1
lysed by using H NMR spectroscopy. Analysis was performed on
[
8] a) H. Kim, K. M. Kosuda, R. P. Van Duyne, P. C. Stair, Chem. Soc. Rev. 2010,
a Bruker 500 MHz Ultra-Shield NMR spectrometer and reactant
products were quantified against a 1 vol% TMS/CDCl3 internal
standard.
3
9, 4820–4844; b) F. Fan, Z. Feng, C. Li, Chem. Soc. Rev. 2010, 39, 4794–
4801.
[
9] a) S. Bordiga, A. Damin, F. Bonino, G. Ricchiardi, C. Lamberti, A. Zec-
china, Angew. Chem. Int. Ed. 2002, 41, 4734–4737; Angew. Chem. 2002,
114, 4928–4931; b) F. Bonino, A. Damin, A. Piovano, C. Lamberti, S. Bor-
Catalyst characterisation
diga, A. Zecchina, ChemCatChem 2011, 3, 139–142; c) K. Sun, F. Fan, H.
Xia, Z. Feng, W.-X. Li, C. Li, J. Phys. Chem. C 2008, 112, 16036–16041.
10] A. Zecchina, M. Rivallan, G. Berlier, C. Lamberti, G. Ricchiardi, Phys.
Chem. Chem. Phys. 2007, 9, 3483–3499.
Raman spectroscopic analysis was performed on a Renishaw inVia
Raman microscope with laser excitation wavelengths of l=325,
[
5
14 and 785 nm. Solid sample measurements were performed
[11] a) J. Pꢂrez-Ramꢃrez, J. C. Groen, A. Brꢄckner, M. S. Kumar, U. Bentruip,
M. N. Debbagh, L. A. Villaecusa, J. Catal. 2005, 232, 318–334; b) S. Bordi-
ga, R. Buzzoni, F. Geobaldo, C. Lamberti, E. Giamello, A. Zecchina, G.
Leofanti, G. Petrini, G. Tozzola, G. Vlaic, J. Catal. 1996, 158, 486–501.
inside a Raman cell at 10 mW power, with a total of 256ꢁ2 s accu-
mulations. No organic species were present in the material. H O -
treated samples were measured at 1 mW power, with 50ꢁ2 s scans
performed. Metal contents were determined by using inductively
coupled plasma atomic emission spectroscopy to an accuracy of
2
2
[
[
12] C. Li, J. Catal. 2003, 216, 203–212.
13] E. A. Pidko, E. J. M. Hensen, R. A. van Santen, Proc. R. Soc. London Ser. A
2
012, 468, 2070–2086.
1
0%. Supporting characterisations have been described else-
[
14] a) K. Park, C. B. Bell III, L. V. Liu, D. Wang, G. Xue, Y. Kwak, S. D. Wong,
K. M. Light, J. Zhao, E. E. Alp, Y. Yoda, M. Saito, Y. Kobayashi, T. Ohta, M.
Seto, L. Que, Jr., E. I. Solomon, Proc. Natl. Acad. Sci. USA 2013, 110,
[6c,d]
where.
6
275–6280; b) P. C. Junk, B. J. McCool, B. Moubaraki, K. S. Murray, L.
Spiccia, J. D. Cashion, J. W. Steed, J. Chem. Soc. Dalton Trans. 2002,
024–1029; c) M. Kodera, Y. Kawahara, Y. Hitomi, T. Nomura, T, Ogura,
Acknowledgements
1
Y. Kobayashi, J. Am. Chem. Soc. 2012, 134, 13236–13239.
15] J. Sanders-Loehr, W. D. Wheeler, A. K. Shiemke, B. A. Averill, T. M. Loehr,
J. Am. Chem. Soc. 1989, 111, 8084–8093.
16] E. A. Pidko, V. Degirmenci, R. A. van Santen, E. J. M. Hensen, Angew.
Chem. Int. Ed. 2010, 49, 2530–2534; Angew. Chem. 2010, 122, 2584–
C.H. and N.D. appreciate the support of the UK Catalysis Hub
and the EPSRC (EP/K014714/1). C.H. is also indebted to the Royal
Society of Chemistry for financial support.
[
[
2588.
[
17] a) Y.-M. Lee, S. Hong, Y. Morimoto, W. Shin, S. Fukuzumi, W. Nam, J. Am.
Chem. Soc. 2010, 132, 10668–10670; b) F. Li, K. K. Meier, M. A. Crans-
wick, M. Chakrabarti, J. M. Van Heuvelen, E. Munck, L. Que, Jr., J. Am.
Chem. Soc. 2011, 133, 7256–7259; c) J. J. Braymer, K. P. O’Neill, J.-W.
Rohde, M. H. Lim, Angew. Chem. Int. Ed. 2012, 51, 5376–5380; Angew.
Chem. 2012, 124, 5472–5476.
Keywords: biomimetic catalysis · CÀH activation · in situ
analysis · methane oxidation · zeolites
[1] a) C. Hammond, S. Conrad, I. Hermans, ChemSusChem 2012, 5, 1668–
1686; b) H. Schwarz, Angew. Chem. Int. Ed. 2011, 50, 10096–10115;
Angew. Chem. 2011, 123, 10276–10297; c) S. R. Golisz, T. B. Gunnoe,
W. A. Goddard III, J. T. Groves, R. A. Periana, Catal. Lett. 2011, 141, 213–
[18] Y. He, C. R. Goldsmith, Chem. Commun. 2012, 48, 10532–10534.
[19] G. Roelfes, M. Lubben, K. Chen, R. Y. N. Ho, A. Meetsma, S. Genseberger,
R. M. Hermaut, R. Hage, S. J. Mandal, V. G. Young, Jr., Y. Zang, H. Kooij-
man, A. L. Speck, L. Que, Jr., Inorg. Chem. 1999, 38, 1929–1936.
[20] A. Wada, S. Ogo, S. Nagamoto, T. Kitagawa, Y. Watanade, K. Jitsukawa,
H. Massuda, Inorg. Chem. 2002, 41, 616–618.
221.
[
[
2] a) L. Que, Jr., W. B. Tolman, Nature 2008, 455, 333–340.
3] a) M. Costas, M. P. Mehn, M. P. Jensen, L. Que, Jr., Chem. Rev. 2004, 104,
939–986; b) M. Sono, M. P. Roach, E. D. Coulter, J. H. Dawson, Chem.
&
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
6
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Communications
[
[
[
[
21] L. V. Liu, S. Hong, J. Cho, W. Nam, E. I. Solomon, J. Am. Chem. Soc. 2013,
35, 3286–3299.
22] T. G. Traylor, S. Tscuchiya, Y.-S. Byun, C. Kim, J. Am. Chem. Soc. 1993, 115,
775–2781.
23] N. Lehnert, R. Y. N. Ho, L. Que, Jr., E. I. Solomon, J. Am. Chem. Soc. 2001,
23, 8271–8290.
24] O. Pestovsky, A. Bakac, J. Am. Chem. Soc. 2004, 126, 13757–13764.
[25] E. V. Kudrik, P. Afanasiev, L. X. Alvarez, P. Dubourdeaux, M. Clemancey, J.-
M. Latour, G. Blondin, D. Bochu, F. Albrieux, S. E. Nefedov, A. B. Sorokin,
Nat. Chem. 2012, 4, 1024–1029.
1
2
Received: August 13, 2014
Revised: October 29, 2014
Published online on && &&, 0000
1
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
7
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

COMMUNICATIONS
C. Hammond,* I. Hermans, N. Dimitratos
A pre-ferryl cat: Through in situ reso-
nance-enhanced Raman spectroscopy,
we identify an active, binuclear
&& – &&
III
Biomimetic Oxidation with Fe-ZSM-5
and H O ? Identification of an Active,
FeÀO(H)ÀFe core and an Fe ÀOOH in-
termediate in Fe-containing ZSM-5 fol-
2
2
Extra-Framework Binuclear Core and
lowing activation with H O . The pre-
2 2
III
an Fe ÀOOH Intermediate with
ferryl nature of this biomimetic inter-
mediate may account for the unique
ability of this solid catalyst to selectively
oxidise methane to methanol under
mild conditions.
Resonance-Enhanced Raman
Spectroscopy
&
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
8
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!