Aqueous-phase methane oxidation over Fe-MFI zeolites; Promotion through isomorphous framework substitution

Article abstract of DOI:10.1021/cs400288b

Fe- and Cu-containing zeolites have recently been shown to be efficient catalysts for the one-step selective transformation of methane into methanol in an aqueous medium at only 50 C, using H₂O₂ as green oxidant. Previously, we have observed that Fe species alone are capable of catalyzing this highly selective transformation. However, further catalytic testing and spectroscopic investigations demonstrate that although these extra-framework Fe species are the active component of the catalyst, significant promotion is observed upon the incorporation of other trivalent cations, e.g., Al³⁺ or Ga³⁺, into the MFI-framework. While these additional framework species do not constitute active catalytic centers, promotion is observed upon their incorporation as they (1) facilitate the extraction of Fe from the zeolite framework and hence increase the formation of the active Fe species and (2) provide an associated negatively charged framework, which is capable of stabilizing and maintaining the dispersion of the cationic extra-framework Fe species responsible for catalytic activity. By understanding these phenomena and subsequently controlling the overall composition of the catalyst (Fe and Al), we have subsequently been able to prepare a catalyst of equal intrinsic activity (i.e., TOF) but five-times higher productivity (i.e., space-time-yield) compared with the best catalysts reported for this reaction to date.

Full text of DOI:10.1021/cs400288b

Subscriber access provided by Columbia Univ Libraries
Article
Aqueous phase methane oxidation over Fe-MFI zeolites;
promotion through isomorphous framework substitution
Ceri Hammond, Nikolaos Dimitratos, Antonio Lopez-Sanchez, Robert Jenkins, Gareth Whiting,
Simon Kondrat, Mohd Hasbi ab Rahim, Michael Forde, Adam Thetford, Henk Hagen, E.
Stangland, Jacob A. Moulijn, Stuart H. Taylor, David Willock, and Graham J. Hutchings
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/cs400288b • Publication Date (Web): 02 Jul 2013
Downloaded from http://pubs.acs.org on July 8, 2013
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street
N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 12
ACS Catalysis
1
2
3
4
5
6
7
8
9
Aqueous phase methane oxidation over Fe-MFI zeolites; pro-
motion through isomorphous framework substitution
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Ceri Hammond,*[a ] Nikolaos Dimitratos,[a‡] Jose Antonio LopezꢀSanchez,[a††] Robert L. Jenkins,[a]
†
Gareth Whiting,[a] Simon A. Kondrat,[a] Mohd Hasbi ab Rahim,[a] Michael M. Forde,[a] Adam Thetꢀ
ford,[a] Henk Hagen,[b] Eric E. Stangland,[c] Jacob M. Moulijn,[a] Stuart H. Taylor,[a] David J. Willock,[a]
Graham J. Hutchings*[a]
[a] Cardiff Catalysis Institute, School of Chemistry, Cardiff University, Main Building, Park Place, Cardiff, CF10 3AT (UK)
[b] Dow Benelux B. V., Herbert H. Dowweg 5, 4542 NM HOEK, Postbus 48, 4530 AA Terneuzen (The Netherlands)
[c] National Corporate R&D, The Dow Chemical Company, Midland, MI 48674 (USA)
† Current address: Department of Chemistry and Applied Biosciences, ETH Zürich, WolfgangꢀPauliꢀStr. 10, CHꢀ8093 Züꢀ
rich (Switzerland)
‡ Current addresses: University College London, 20 Gordon Street, WC1H 0AJ London, UK, and Research Complex at
Harwell, Rutherford Appleton Laboratory, Harwell Oxford, Didcot, OX11 0FA (UK)
†† Current address: Stephenson Institute for Renewable Energy, Chemistry Department, The University of Liverpool, Crown
Street, Liverpool, L69 7ZD (UK)
Eꢀmail: ceri.hammond@chem.ethz.ch; hutch@cardiff.ac.uk
KEYWORDS: Methane oxidation. Zeolite catalysis. Selective oxidation. Natural gas upgrading. Sustainable chemistry.
ABSTRACT: Fe and Cuꢀcontaining zeolites have recently been shown to be efficient catalysts for the oneꢀstep selective transforꢀ
mation of methane into methanol in an aqueous medium at only 50 ºC, using H2O2 as green oxidant. Previously, we have observed
that Fe species alone are capable of catalysing this highly selective transformation. However, further catalytic testing and spectroꢀ
scopic investigations demonstrates that although these extraꢀframework Fe species are the active component of the catalyst, signifiꢀ
cant promotion is observed upon the incorporation of other trivalent cations e.g. Al3+, Ga3+, into the MFIꢀframework. Whilst these
additional framework species do not constitute active catalytic centres, promotion is observed upon their incorporation as they (1)
facilitate the extraction of Fe from the zeolite framework, and hence increase the formation of the active Fe species, and (2) provide
an associated negativelyꢀcharged framework, which is capable of stabilizing and maintaining the dispersion of the cationic extraꢀ
framework Fe species responsible for catalytic activity. By understanding these phenomena, and subsequently controlling the overꢀ
all composition of the catalyst (Fe and Al), we have subsequently been able to prepare a catalyst of equal intrinsic activity (i.e.
TOF) but fiveꢀtimes higher productivity (i.e. spaceꢀtimeꢀyield) to the best catalysts reported for this reaction to date.
fossil reserves, a large fraction of the natural gas reserves is
1 Introduction
inconveniently located in various inaccessible parts of the
Natural gas is a highly abundant source of hydrocarbons that is
globe. In contrast to crude oil, the transportation of this volaꢀ
primarily composed of methane (ca. 85 vol. %), and represents
tile and flammable gas to existing technological sites presents
one of the major building blocks of the present day chemical
considerable safety and economic issues. The conversion of
industry.1ꢀ3 Nevertheless, its conversion to various chemicals
methane to more energy dense liquid derivatives, particularly
and fuels currently requires the intermediate manufacture of
at the point of origin, could lead to significant breakthroughs
synthesis gas, and subsequent conversion to higher hydrocarꢀ
in the utilisation of natural gas as a primary feedstock.
bons or commodity chemicals through FischerꢀTropsch type
chemistry.4 During various stages of these processes, extremeꢀ
The primary obstacle and more challenging aspect of this reacꢀ
tion stems from the fact that methane is the least reactive of all
ly high temperatures and pressures are required, thus resulting
hydrocarbons, with very high CꢀH bond strengths of 438.8 kJ
in high operational costs and significant capital investment.
molꢀ1. Consequently, conditions that are sufficient to activate
As such, the development of alternative and less economically
methane also have the undesired effect of activating the partial
intensive routes for the selective transformation of natural gas,
oxidation products toward deeper oxygenated products (COx),
or methane in particular, to various valueꢀadded products is of
since bond strengths in these oxygenated species are typically
particular interest. Of greatest promise is the transformation of
much lower (ꢁHCꢀH = 373.5 kJ molꢀ1 for methanol). Thus, at
methane to more energy dense liquid derivatives, such as
high temperatures the formation of deeper oxygenated prodꢀ
methanol, formic acid or midꢀrange hydrocarbons. Along with
ucts such as CO and CO2 is generally unavoidable, which limꢀ
being useful as chemical building blocks e.g. methanolꢀtoꢀ
its the overall reaction selectivity. In view of this, it is apparꢀ
olefin technology,5 these liquid derivatives are significantly
ent that the selective oxidation of methane can only be
easier and cheaper to transport around the globe; like many
ACS Paragon Plus Environment

ACS Catalysis
achieved by developing new catalytic systems that are able to
Page 2 of 12
1
2
3
4
5
6
7
8
oxidise methane at mild temperatures (≤ 200 ºC), as this may
provide some inherent selectivity to the partial oxidation prodꢀ
ucts by operating under kinetic rather than thermodymanic
control.
2 Results and Discussion
Material synthesis. To determine the precise role(s) of Al3+,
analogous samples of silicaliteꢀ1, ZSMꢀ5, Feꢀsilicaliteꢀ1 and
FeꢀZSMꢀ5 were first prepared by hydrothermal synthesis. In
particular, this was necessary in order to avoid issues associatꢀ
ed with comparing samples obtained from different sources
(commercial material vs. laboratory prepared material) and
that contained different Fe loadings. To achieve this compariꢀ
son, Alꢀonly (ZSMꢀ5), Feꢀonly (Feꢀsilicaliteꢀ1), Fe and Alꢀ
containing (FeꢀZSMꢀ5) and metalꢀfree (silicaliteꢀ1) analogues
were prepared by an identical, benchmarked20 hydrothermal
synthesis procedure (Table 1), and later screened for activity
(Table 2). We note here that the preꢀtreatment temperature
employed during activation of the catalyst is provided as a
subscript following the description of the catalyst, and that the
SiO2/Al2O3 ratio – where applicable – is provided in parentheꢀ
sis. Additionally, the Fe loading of the catalyst – where appliꢀ
cable – is provided prior to the catalyst description e.g. Feꢀ
containing ZSMꢀ5, containing 0.5 wt. % Fe and a SiO2/Al2O3
mole ratio of 84, and preꢀtreated at 550 ºC, is denoted 0.5Feꢀ
ZSMꢀ5 (84)550. As demonstrated by XRD analysis (Extended
Supplementary Information (ESI) Figure S1), each of the synꢀ
thesised samples possesses the crystalline MFI structure. Couꢀ
pled with the similar surface areas (± 330 m2 gꢀ1) and microꢀ
prorous volumes (± 0.13 cmꢀ3 gꢀ1), it can be concluded that
each synthesised solid possesses comparable physical properꢀ
ties. In view of this, each sample was subsequently evaluated
for catalytic activity following activation at 550 ºC.
Whilst a number of low temperature approaches have been
proposed in recent times, each approach is typically limited by
low catalytic rates i.e. low turnover frequencies/spaceꢀtimeꢀ
yields, or environmental issues concerning the solvent system
or chosen oxidant.1 For example, whilst the electrophillic actiꢀ
vation of methane by bipyramidal platinum complexes yields
methyl bisulphate at high selectivity (81 % at > 90 % converꢀ
sion), this and related systems are limited by the highly corroꢀ
sive solvent system (oleum), low intrinsic activity (TOF < 10
hꢀ1) and the lack of a fullyꢀclosed catalytic cycle.6,7 An alternaꢀ
tive approach has focused upon on the use of Fe(Cu)ꢀ
containing zeolites for the selective oxidation of methane with
N2O(O2).8ꢀ13 Whilst these materials do exhibit a remarkable
ability to activate methane, the product in both cases is a
strongly chemisorbed methoxy species that cannot readily
desorbed or be isolated without destruction of the catalytic
active sites, thus resulting in a nonꢀcatalytic process. Alternaꢀ
tive approaches using both encapsulated14 or supported15,16 Feꢀ
phthalocyanine complexes have also been proposed, but these
systems are limited by very low intrinsic activities, maximum
methanol selectivities of 50 %, and catalyst stability issues.17
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We have demonstrated that Feꢀcontaining MFIꢀtype zeolites
are capable of not only oxidising methane at high catalytic
rates (≤ 14,500 hꢀ1) and partial oxygenate selectivity i.e. selecꢀ
tivity to oxidation products not including CO and CO2 (≥ 90
%), but that this favourable transformation can be carried out
in an environmentally benign process at only 50 ºC, in the
aqueous phase and with hydrogen peroxide as the chosen oxiꢀ
dant.18 Moreover, although the Feꢀonly system was primarily
selective to formic acid, we have demonstrated that the addiꢀ
tion of Cu2+ to these highly reactive Feꢀcontaining zeolites
eliminates the methanol overꢀoxidation process, thereby alꢀ
lowing methanol to be obtained at over 90 % selectivity at
methane conversions of up to 10 %.18,19
Table 1. Physical and chemical properties of a series of Alꢀ
and Feꢀcontaining MFIꢀtype zeolites, prepared by hydrotherꢀ
mal synthesis.
Catalyst
Fe content Molar ratio
(wt. %)‡
SBET
(m2 gꢀ1)
VMICRO
(cm3 gꢀ1)
SiO2/
Fe2O3
SiO2/
Al 2O3
silicaliteꢀ1550
< 0.001
0.003
ꢀ
ꢀ
ꢀ
330
340
0.13
0.14
Recently, we demonstrated that the catalytic activity of these
materials corresponds to the formation of extraꢀframework
Fe3+ species that reside within the zeolite micropores.20 Whilst
bearing some resemblance to the solidꢀstate chemistry exhibitꢀ
ed by these materials during activation for N2Oꢀbased oxidaꢀ
tions, key evidence has indicated that the active species
formed in this system for selective and catalytic methane oxiꢀ
dation with H2O2 are fundamentally different to those found
for these other stoichiometric oxidation systems, given the
different preꢀactivation procedures employed.20 However, a
major question remains. Whilst Fe alone in Feꢀsilicaliteꢀ1 is
capable of catalysing the reaction, significantly higher turnoꢀ
ver frequencies have been observed for an Fe and Alꢀ
containing systems (such as commercial zeolite ZSMꢀ5), deꢀ
spite the catalytic inactivity of Al3+ and its related properties
for this reaction.18 In this publication we now focus on this key
aspect in order to develop a more detailed understanding of
this unique methane oxidation catalyst, and subsequently proꢀ
duce significantly more active catalysts for this challenging
reaction.
ZSMꢀ5 (86)
86
550
Feꢀsilicaliteꢀ
1550
0.52
250
254
ꢀ
330
310
0.13
0.14
FeꢀZSMꢀ5(84) 0.49
84
550
‡ Determined by ICPꢀOES. Values are accurate to ± 10 %.
In line with our previous studies, the synthesis of metalꢀfree
silicaliteꢀ1550 and Alꢀonly ZSMꢀ5 (86)550 does not lead to maꢀ
terials with any significant activity for this reaction.18ꢀ20 This is
in agreement with our previous observations that catalytic
activity for methane oxidation with H2O2 can only be achieved
when sufficient quantities of Fe3+ are present within the cataꢀ
lyst, and that the presence of Brønsted and Lewis acid sites
(from framework Al3+), and/or a microporous framework
alone, are insufficient for catalysis to be observed. In fact, we
have previously attributed the very minor catalytic activity of
our synthesised ZSMꢀ5 materials to be related to their low but
nonꢀnegligible Fe3+ content (Table 1).20
ACS Paragon Plus Environment

Page 3 of 12
ACS Catalysis
Table 2. Catalytic activity of analogous samples of a series of This raises the important question regarding the precise role(s)
of Al3+. Indeed, even for the more thoroughly established N2Oꢀ
based oxidations, the role(s) of Al3+ in the same or similar
materials is still the subject of much debate. For instance,
Hensen et al. have reported that only MFI materials containing
both Fe and Al exhibit catalytic activity for N2Oꢀbased oxidaꢀ
tions, as the active site in these cases is an extraꢀframework
mixed oxide (FeꢀOꢀAl).21 A number of other reports also conꢀ
clude that Al3+ itself, or the Brønsted acid sites associated with
framework Al3+ also constitute active catalytic centres for such
reactions.22 It has also been reported that Al3+ facilitates the
autoꢀreduction of Fe3+ to Fe2+, which is the active state of Fe
for both benzene hydroxylation and N2O decomposition. Furꢀ
thermore, the ability of Al3+ to aid the extraction of Fe3+ from
the framework of the zeolite (and thus form active extraꢀ
framework species) is also widely reported.23 Finally, it is also
1
2
3
4
5
6
7
8
Alꢀ and Feꢀcontaining MFIꢀtype zeolites, prepared by hydroꢀ
thermal synthesis.
Catalyst
Product amount (ꢂmol)
Sum of Oxy.
products sel.
(ꢂmol) (%)
MeOH HCOOH MeOOH CO2 (g)
silicaliteꢀ1550 0.0
0.0
0.0
0.0
0
ꢀ
ZSMꢀ5 (86) 1.5
0.8
2.3
1.6
6.2
74
9
550
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Feꢀsilicaliteꢀ 17.6
1550
56.1
11.0
3.0
6.4
91.1
93
92
FeꢀZSMꢀ
5(84) 550
20.1
158.5
15.8 197.4
ꢀ
known that the associated cationꢀexchange site of the AlO4
Reaction conditions; cat: various (27 mg); P(CH4): 30.5 bar;
[H2O2]: 0.5 M; temp: 50 °C; time: 30 min; stirring speed: 1500
rpm; Note: Each catalyst was calcined at 550 °C for 3 h in air
prior to use.
tetrahedron is able to stabilise cationic extraꢀframework comꢀ
plexes.24,25 Nevertheless, it should not be overlooked that the
activation procedures employed for this reaction are fundaꢀ
mentally different to those utilised for N2Oꢀbased oxidations,
and that different Fe species apparently catalyse these different
reactions.
The requirement for Fe3+ is well emphasised by comparing
entries 1 and 3 of Table 2, where it can be observed that the
incorporation of a low amount of Fe3+ (0.5 wt. %) into the
inactive silicaliteꢀ1 material leads to large increases in catalytꢀ
ic activity. As expected in the absence of a Cu2+ additive,
which we have shown to be critical for maintaining MeOH
selectivity,18ꢀ20 the major product formed with 0.5Feꢀsilicaliteꢀ
1 is HCOOH (at 62 % selectivity), though the selectivity to
partial oxygenates (MeOOH, MeOH and HCOOH) remains
high (93 %) and the selectivity to CO2 (7 %) remains very low.
Extraction of framework Fe3+. We have previously demonꢀ
strated that although a homogeneous distribution of frameꢀ
work Fe3+ is found within as synthesised zeolite, significant
changes to the speciation of Fe3+ are observed following the
two heat preꢀtreatment procedures required to (1) remove the
residual organic template, and (2) to further activate the mateꢀ
rial prior to catalysis (Scheme 1).20 Specifically, we observed
that during removal of the organic template (preꢀtreatment 1)
and further activation (preꢀtreatment 2), Fe3+ migrates from
coordinatively saturated framework sites, to form extraꢀ
framework Fe3+ cations that take up position within the zeolite
channels. Indeed, this migration to the extraꢀframework was
found to be a preꢀrequisite for attaining high levels of activity,
and a positive correlation (R2 = 0.92) between the fraction of
these species and catalytic activity was observed.20 Considerꢀ
ing this, it seemed important to us to establish whether the
migration of Fe3+ to the extraꢀframework is greater when Al3+
is also present in the structure. In fact, it has previously been
reported that along with Fe3+ being less stable than Al3+ in the
ZSMꢀ5 framework, the stability of Fe3+ within the MFI
framework is significantly lower in ZSMꢀ5 than in silicaliteꢀ
1.21
Nevertheless, in spite of the minimal catalytic activity of Alꢀ
only ZSMꢀ5 (Entry 2, Table 2), 0.5FeꢀZSMꢀ5 (84), again
comprising only 0.5 wt. % Fe, is over twoꢀtimes more active
than the analogous sample of 0.5Feꢀsilicaliteꢀ1, following the
addition of only 0.6 wt. % Al (Table 2, compare entry 4 vs. 3).
Additionally, although the catalyst is over twice as active, no
significant loss of partial oxygenate selectivity was observed
(92 %), though HCOOH selectivity was marginally higher (80
%), likely as a consequence of the increased conversion. Givꢀ
en that these samples were prepared by the same method, and
contain similar physical and chemical properties (Fe content,
surface areas and pore volumes), it is highly unlikely that the
reactivity differences observed can be related to changes in the
physical properties of the zeolite, or potential secondary facꢀ
tors such as diffusion. Indeed, it can be assumed that the only
difference in these samples is the Al3+ content, and hence the
simultaneous presence of Fe3+ and Brønsted/Lewis acid sites.
This conclusively demonstrates that the presence of Al3+ is
highly beneficial to the activity of the catalyst. Whilst it canꢀ
not be forgotten that the addition of Al3+ does increase the
hydrophilicity of the MFIꢀframework, the very dilute levels of
Al3+ in 0.5FeꢀZSMꢀ5 (84)550 (SiO2/Al2O3 = 84) does not maꢀ
jorly change the hydrophobic nature of the sample, and since
the diffusion of both a hydrophobic (methane) and hydrophilic
(H2O2) reactant is required for reaction, changes in the hydroꢀ
philꢀ/phobꢀicity of the samples is unlikely to be responsible for
the observed difference in catalytic activity. We stress here
that in agreement to our previous publications,18ꢀ20 each cataꢀ
lyst was found to be heterogeneous i.e. leaching of an active
homogeneous catalyst into solution did not occur (Figure S2).
Scheme 1. Activation procedures employed for 0.5Feꢀsilicaliteꢀ1
and 0.5FeꢀZSMꢀ5 (84). Removal of the template at 550 ºC and
ionꢀexchange with NH4NO3 leads to an NH4ꢀform zeolite. Further
activation (≥ 550 ºC) yields the Hꢀform of the zeolite. This
scheme was originally published in reference 20.
ACS Paragon Plus Environment

ACS Catalysis
The ligand to metal charge transfer (LMCT) bands (Fe3+ꢀO)
Page 4 of 12
Table 3. Deconvolution data for 0.5Feꢀsilicaliteꢀ1550 and
0.5FeꢀZSMꢀ5 (84)550, both containing ± 0.5 wt. % Fe3+, after
template removal and heat preꢀtreatment at 550 ºC.
1
2
3
4
5
6
7
8
within the UVꢀVis spectra of Feꢀcontaining zeolites is an ideal
method for studying the extraction of Fe3+ from the framework
of the zeolite, due to the distinct absorbances found for Fe
species in different geometrical positions and coordination
environments within the zeolite.26ꢀ31 There are of course some
inherent limitations in regards to utilising UVꢀVis as a fully
quantitative tool. For example, deviations in the precise molar
extinction coefficients (ε) of various absorbing species, and
the presence of (multiple) broad, overlapping bands, make full
quantification an extreme challenge. Nevertheless, for Feꢀ
containing zeolites, it has been shown that the ε values are
equal to the same order of magnitude,32 and that the multiple
bands can adequately be fitted by using single bands correꢀ
sponding to (1) framework Fe species (200ꢀ250 nm, λ1), (2)
isolated and oligomeric extraꢀframework Fe cations within the
zeolite channels (250ꢀ350 nm, λ2), (3) larger Fe clusters (350ꢀ
450 nm, λ3) and finally (4) bulk Fe oxides on the surface of the
zeolite (> 450 nm, λ4). Moreover, any deviations of these
factors will also be systematic over the entire series of preꢀ
treated catalysts, and will still provide empirical and semiꢀ
quantitative insights in regards to the precise changes in
speciation of Fe3+ in these materials as a function of preꢀ
treatment and/or Al3+ incorporation.
Catalyst
Relative contribution of each λ range (nm)
λ1
λ2
λ3
λ4
0.5Feꢀsilicaliteꢀ
1550
0.627
0.337
0.025
0.011
0.5FeꢀZSMꢀ5
(84) 550
0.422
0.542
0.027
0.009
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Although the aforementioned issues do not allow us to calcuꢀ
late an exact percentage of each type of Fe species, calculating
the relative contributions of each area of the UVꢀVis absorpꢀ
tion spectra demonstrates that whilst significant Fe3+ migration
is observed for the activated form of 0.5Feꢀsilicaliteꢀ1550, it is
evident that following identical heat preꢀtreatment procedures,
significantly more Fe3+ has migrated from the tetrahedral
framework sites in 0.5FeꢀZSMꢀ5 (84)550. In fact, following
template removal and activation at 550 ºC, the majority of Fe3+
can be assigned to extraꢀframework species in 0.5FeꢀZSMꢀ5
(84) 550, clearly displaying the remarkably low stability of
framework Fe3+ in this sample. This also demonstrates the
lower stability of framework Fe3+ in ZSMꢀ5 than in silicaliteꢀ
1, an observation that is supported by the available literature.21
Most notably, a significantly higher fraction of extraꢀ
framework Fe cations within the micropores is present in
0.5FeꢀZSMꢀ5 (84)550 than in 0.5Feꢀsilicaliteꢀ1550. This is highꢀ
ly significant, as we have previously demonstrated through
computational18 and spectroscopic20 studies that these Fe3+
species are most likely the species that are responsible for the
catalytic activity displayed by this catalyst the for selective
oxidation of methane. In fact, if one were to assume that all of
these Fe species in these samples were active, and that no othꢀ
er Fe species impact catalytic activity whatsoever, we would
expect 0.5FeꢀZSMꢀ5 (84) 550 to be around 1.6 times more acꢀ
tive than 0.5Feꢀsilicaliteꢀ1550 under these conditions, which is
in good, but not perfect, agreement to the observed activities;
0.5FeꢀZSMꢀ5 (90)550 being 2.1 times more active than 0.5Feꢀ
silicaliteꢀ1550 (Table 2). We stress here that through UVꢀVis
spectroscopy alone it is not possible to be more specific reꢀ
garding the nature and composition of these extraꢀframework
species, as there are a large number of potential species that
could contribute to this broad absorbance feature i.e. isolated
extraꢀframework Fe3+ species, dimers, trimers and small oliꢀ
gomers are all expected to absorb within this region. However,
whilst not the purpose of this publication, our previous
EXAFS studies and DFT calculations have suggest that the
active extraꢀframework clusters contain between one and three
Fe atoms, with the best match for experiment and theory being
obtained for Fe2(µ2ꢀOH)2(OH)2(H2O)2]2+, a binuclear active
site that comprises an overall +2 charge.18
Figure 1. Diffuse Reflectance UVꢀVis spectra for 0.5Feꢀsilicaliteꢀ
1550 (blue/solid) and 0.5FeꢀZSMꢀ5 (84)550 (red/dashed) both conꢀ
taining ± 0.5 wt. % Fe3+ and calcined at 550 ºC in air.
Given the increased levels of absorbance at wavelengths at or
above 250 nm ꢀ which correspond to extraꢀframework Fe3+
species – it is clear that following template removal and idenꢀ
tical activation procedures (550 °C, 3 h in air), more extensive
Fe3+ migration has indeed taken place in 0.5FeꢀZSMꢀ5 (84)550
versus 0.5Feꢀsilicaliteꢀ1550 (Figure 1). Given that our previous
studies correlated catalytic activity to extraꢀframework Fe
species within the zeolite micropores,20 it is logical that the
increased dislodgement of framework Fe3+ to the extraꢀ
framework observed for 0.5FeꢀZSMꢀ5550 at identical preꢀ
treatment temperatures would result in an increase in catalytic
activity.
It seems reasonable, therefore, that, since greater Fe3+ migraꢀ
tion is observed in 0.5FeꢀZSMꢀ5 (84)550 than in 0.5Feꢀ
silicaliteꢀ1550 following identical activation procedures, the Alꢀ
containing analogue would display higher levels of activity. In
such a case, Al3+ would not so much act as a catalytic promotꢀ
er, but behave as a structural promoter for increasing the probꢀ
ability of forming the active species. In such a case, it should
be possible to form an equally active sample of 0.5Feꢀ
The relationship between extraꢀframework Fe species within
the micropores and catalytic activity, especially in regards to
the disparate activities of 0.5FeꢀZSMꢀ5 (90) and 0.5Feꢀ
silicaliteꢀ1, is even more clearly displayed through full deconꢀ
volution of the observed UVꢀVis spectra into the four relevant
subꢀbands (Table 3).
ACS Paragon Plus Environment

Page 5 of 12
ACS Catalysis
This is further emphasized from the deconvolution analysis of
silicaliteꢀ1550 just by increasing the preꢀtreatment (activation)
temperature, in order to obtain a similar distribution of Fe3+
species. To verify whether the increased activity observed for
0.5FeꢀZSM5 (84)550 over 0.5Feꢀsilicaliteꢀ1550 was simply due
to an insufficient preꢀtreatment of Feꢀsilicaliteꢀ1 i.e. whether
Al3+ did indeed promote the catalyst, or if the preꢀtreatment of
Feꢀsilicaliteꢀ1 was simply not optimised to achieve the same
level of extraꢀframework Fe3+, and hence activity, both NH4ꢀ
form samples of 0.5FeꢀZSMꢀ5 (84) and 0.5Feꢀsilicaliteꢀ1 were
preꢀtreated at different temperatures, in order to determine the
maximum activity possible for each sample (Figure 2).
1
2
3
4
5
6
7
8
0.5FeꢀZSMꢀ5 (84) and 0.5Feꢀsilicaliteꢀ1, following heat preꢀ
treatment (ESI Table S1 and Figure S3 and reference 20, respecꢀ
tively). Previously, we have demonstrated that the activity of
0.5Feꢀsilicaliteꢀ1 correlates well with the amount of extraꢀ
framework Fe within the zeolite micropores i.e. λ2. A similar inꢀ
terpretation of the UVꢀVis data of 0.5FeꢀZSMꢀ5 (84) also producꢀ
es a positive correlation between the percentage of these species
and catalytic activity (Figure 3). Nevertheless, it is clear that per
‘active’ Fe species, 0.5FeꢀZSMꢀ5 (84) displays over 20 % higher
activity than 0.5Feꢀsilicaliteꢀ1.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Catalytic activity of 0.5Feꢀsilicaliteꢀ1 (blue bars/left)
and 0.5FeꢀZSMꢀ5 (84) (red bars/right), following high temperaꢀ
ture preꢀtreatment. The temperature of preꢀtreatment is denoted in
Figure 3. Relationship between the % of extraꢀframework Fe
species within the zeolite micropores (the proposed active speꢀ
cies) and catalytic activity for 0.5Feꢀsilicaliteꢀ1 (▲) and 0.5Feꢀ
ZSMꢀ5 (84) (ꢁ) following various preꢀtreatment procedures. The
data for 0.5Feꢀsilicaliteꢀ1 was previously published in reference
20.
parenthesis. Reaction conditions; cat: various (27 mg); P(CH4)
:
30.5 bar; [H2O2]: 0.5 M; temp: 50 °C; time: 30 min; stirring
speed: 1500 rpm; catalyst preꢀtreatment: various °C, 3 h, air.
As we previously demonstrated for 0.5Feꢀsilicaliteꢀ1, both
ammoniumꢀform samples of 0.5Feꢀsilicaliteꢀ1NH4 and 0.5Feꢀ
ZSMꢀ5NH4 show considerable levels of activity, due to the
migration of Fe3+ from the framework sites that occurs during
removal of the organic template (Scheme 1). Nevertheless,
considerable increases in catalytic activity of both materials
can be achieved following a second highꢀtemperature preꢀ
treatment. However, whilst it is clear that an activation temꢀ
perature of 550 ºC is optimal for 0.5FeꢀZSMꢀ5 (84)550, 0.5Feꢀ
silicaliteꢀ1 does not reach its maximum catalytic activity until
a preꢀtreatment temperature of 750 °C is utilised. At signifiꢀ
cantly higher preꢀtreatment temperatures (900 ºC), both 0.5Feꢀ
silicaliteꢀ1 and 0.5FeꢀZSMꢀ5 (84) decrease in activity, though
the observed decrease is significantly greater for Feꢀsilicaliteꢀ
1.
Role of cation-exchange sites. An additional promotional role
of Al3+ was inferred by visual inspection and kinetic evaluaꢀ
tion of the samples calcined at 900 °C; whilst the 0.5FeꢀZSMꢀ
5 (84)900 sample was still white in colour, and comparable in
activity to those samples preꢀtreated at 550 ºC, the 0.5Feꢀ
silicaliteꢀ1900 sample was a dark orange/brown colour and sigꢀ
nificantly less active than the samples calcined at 550 and 750
°C. This implies that excessive clustering to catalytically inacꢀ
tive bulk Fe oxides is more evident in the Feꢀonly sample, in
spite of the fact that Fe3+ is initially more easily extracted from
the Feꢀ and Alꢀcontaining system. We have previously identiꢀ
fied bulk Fe oxides to not only be incapable of methane actiꢀ
vation, but responsible for the formation of carbon oxide speꢀ
cies and increased nonꢀselective H2O2 decomposition.20 Thus,
the decrease in the percentage of ‘active’ Fe species, coupled
with the increased formation of ‘inactive’ and undesirable
spectator species would correlate favourably with the observed
activities.
As we recently reported,20 UVꢀVis analysis confirms that
0.5Feꢀsilicaliteꢀ1900 has ‘overꢀclustered’ to Fe oxide species
upon the excessive thermal preꢀtreatment (ESI Table S1). Inꢀ
deed, a remarkable increase in the fraction of undesirable largꢀ
er clusters and bulk oxides is observed following preꢀtreatment
at 900 ºC. This agrees well with the colour of the sample (orꢀ
ange/brown) and the significant decrease in observed activity
(Figure 2). It is apparent, therefore, that once a given amount
of Fe migrates from the framework of 0.5Feꢀsilicaliteꢀ1, the
formation of larger clusters and bulk Fe oxides is triggered,
and the catalyst rapidly decreases in activity.
Nevertheless, in spite of the fact that the optimal preꢀtreatment
of Feꢀsilicaliteꢀ1 is ca. 200 ºC higher than for FeꢀZSMꢀ5,
0.5Feꢀsilicaliteꢀ1 is always less active than 0.5FeꢀZSMꢀ5 (84),
even following optimisation of both catalyst preꢀtreatment
procedures. This is in direct contrast to previous studies focusꢀ
ing on Feꢀcontaining zeolites for N2O decomposition, which
demonstrated that the catalytic activity of Feꢀsilicaliteꢀ1 and
FeꢀZSMꢀ5 are identical for NOx decomposition following opꢀ
timisation of the preꢀtreatments.28 This clearly indicates that in
the case of Feꢀcontaining zeolites for aqueousꢀphase methane
oxidation, there is an additional promotional role of Al3+ in the
catalyst beyond simply facilitating the extraction of Fe3+ speꢀ
cies to the extraꢀframework.
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 12
In contrast, despite Fe3+ apparently being more readily extractꢀ
Table 4. Deconvolution data for 2.5Fe3+/silicaliteꢀ1550 and
2.5Fe3+/ZSMꢀ5 (86)550
1
2
3
4
5
6
7
8
ed from the MFI framework in 0.5FeꢀZSMꢀ5 (84) than 0.5Feꢀ
silicaliteꢀ1 to begin with, the preꢀtreated samples of 0.5Feꢀ
ZSMꢀ5 (90)750 and 0.5FeꢀZSMꢀ5 (84)900 do not possess signifꢀ
icantly different Fe speciation to the sample calcined at 550
ºC, and the fraction of ‘active’ extraꢀframework Fe3+ species
within the zeolite micropores does change significantly. This
correlates favourably with the miminal decrease in activity
observed with these samples. Although the preꢀtreatment of
0.5FeꢀZSMꢀ5 (84) at high temperatures (> 550 ºC) will also
lead to some dealumination i.e. the formation of extraꢀ
framework Al3+, the fraction of such octahedral Al3+ remained
low, in agreement with the known stability of framework Al3+
in MFI frameworks. Furthermore, the preꢀtreatment of ZSMꢀ5
(86) at such temperatures still did not lead to any relevant
catalytic activity, thus demonstrating that the changes in Fe3+
speciation are the dominant factor in the activity of these maꢀ
terials.
The observation that the ‘over’ꢀclustering of Fe3+ to larger
clusters and bulk Fe oxides is less prevalent for 0.5FeꢀZSMꢀ5
(84) than for 0.5Feꢀsilicaliteꢀ1 suggests that an alternative
promotional role of Al3+ may be its ability to stabilise or disꢀ
perse the active extraꢀframework cationic Fe species, as previꢀ
ously observed by groups focusing on gas phase oxidation of
higher hydrocarbons and NOx decomposition.21,29ꢀ31 Should
this be the case, it would be expected that this would particuꢀ
larly apply for active catalytic materials prepared by postꢀ
deposition methods; although we have shown these samples to
be less active per mole of Fe3+, they contain the entire Fe3+
fraction in extraꢀframework positions, and thus allow the disꢀ
crimination of Al3+ promotion through dispersion/stabilisation
and the increased extraction phenomena observed for the hyꢀ
drothermally prepared samples. Thus, samples of 2.5 wt. %
Fe3+/ZSMꢀ5 (86) and 2.5 wt. % Fe3+/silicaliteꢀ1 were prepared
by solidꢀstate ionꢀexchange and screened for catalytic activity.
The higher metal loading (2.5 wt. % vs. 0.5 wt. %) was chosen
given the lower intrinsic activity of postꢀsynthetic deposition
compared to hydrothermal incorporation. We note here that
postꢀsynthetic Fe3+ deposition is denoted Fe3+/ZSMꢀ5, whereas
hydrothermal incorporation is defined FeꢀZSMꢀ5.
.
Catalyst
Relative contribution of each λ range (nm)
λ1
λ2
λ3
λ4
2.5Fe3+/silicalite 0.204
ꢀ1550
0.164
0.236
0.395
2.5Fe3+/ZSMꢀ5
(86)550
0.169
0.512
0.192
0.127
9
Data calculated from UVꢀVis spectra in ESI Figure S4.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4 demonstrates that in spite of identical Fe3+ loadings
(2.5 wt. %) and preꢀtreatment conditions (550 °C, 3 h, static
air), the 2.5Fe3+/ZSMꢀ5 (86)550 sample is more than one order
of magnitude more active than the analogous 2.5Fe3+/silicaliteꢀ
1550. This clearly emphasises that there is a beneficial role of
Al3+ that is not related to its ability to facilitate the extraction
of Fe3+ from the framework of the zeolite, as both samples
were prepared by postꢀdeposition methods and did not thereꢀ
fore contain any (initial) framework Fe3+.
Following deconvolution of the relevant UVꢀVis spectra of
2.5Fe3+/silicaliteꢀ1550 and 2.5Fe3+/ZSMꢀ5 (86)550, it is clear that
2.5Fe3+/silicaliteꢀ1550 possesses a significantly lower fraction
of ‘active’ Fe species than 2.5Fe3+/ZSMꢀ5 (86)550 (Table 4).
Whilst we could expect 2.5Fe3+/silicaliteꢀ1 to thus be around
1/3 the activity of 2.5Fe3+/ZSMꢀ5, based on the relative conꢀ
tributions within each sample, it cannot be overlooked that the
UVꢀVis spectrum of 2.5Fe3+/silicaliteꢀ1 is dominanted by largꢀ
er Fe clusters and bulk Fe oxides, which are responsible for
nonꢀselective H2O2 decomposition (Entry 2).20 On the other
hand, the fraction of such undesirable spectator species in
2.5Fe3+/ZSMꢀ5 (86) is much lower. The lack of active sites,
coupled with the large amount of undesirable Fe species, reꢀ
sults in this sample showing little activity whatsoever. It is
clear therefore that Fe3+ is significantly more dispersed withꢀ
in/on the MFI material when Al3+ is also present in the zeolite
structure, as significantly less bulk oxides (and thus less clusꢀ
tering) is observed in 2.5Fe3+/ZSMꢀ5 (86)550 than in
2.5Fe3+/silicaliteꢀ1550
.
Based on these experiments, it is evident that the inclusion of
Al3+ within the zeolite leads to higher levels of activity, due to
both an increased extraction of Fe3+ from the zeolite frameꢀ
work (where applicable), and an increased dispersion of the
extraꢀframework Fe species, which maximises the fraction of
extraꢀframework Fe species within the zeolite micropores and
minimises the fraction of undesirable clusters and bulk oxides.
Nevertheless, the exact nature of this increased dispersion is
not yet evident. Previously, it has been proposed that Al3+ aids
the dispersion of Fe3+ in similar zeolite materials even when
deposited by postꢀsynthesis methods; from this, it has been
proposed that extraꢀframework Al3+ species are able to aid
dispersion.21 Alternatively, the cationꢀexchange sites associatꢀ
ed with framework Al3+ may also be responsible for disperꢀ
sion, given the ability of the negative lattice charge to coordiꢀ
nate and stabilize cationic complexes.24,25 Nevertheless, each
of these studies utilized Feꢀcontaining zeolites for highꢀ
temperature (> 250 ºC) gas phase oxidation chemistry, which
we have shown to be unrelated to our present system. Furꢀ
thermore, both reported catalytic systems required preꢀ
treatment at significantly higher temperatures to the catalysts
reported herein, and in a vacuum or an inert atmosphere, in
Figure 4. Catalytic activity of 2.5 wt. % Fe/ZSMꢀ5 (86) and 2.5
wt. % Fe/silicaliteꢀ1, prepared by solidꢀstate ionꢀexchange. Reacꢀ
tion conditions; cat: various (27 mg); P(CH4): 30.5 bar; [H2O2]: 0.5
M; temp: 50 °C; time: 30 min; stirring speed: 1500 rpm; Note:
Catalyst preꢀtreatment: 550 °C, 3 h, static air.
ACS Paragon Plus Environment

Page 7 of 12
ACS Catalysis
order to facilitate the formation of Fe2+ species that are responꢀ
(28)550) is still around 5ꢀ10 times lower than the highest obꢀ
1
2
3
4
5
6
7
8
sible for activity in such cases. Thus, the extrapolation of these
previous studies to the present system, which focuses on the
low temperature, aqueous phase methane oxidation with Fe3+
active sites cannot be presumed.
tained with commercial 0.014ZSMꢀ5 (30)550 under identical
reaction conditions,18 we reason that the probability of obtainꢀ
ing the highest percentage of active Fe sites whilst concurrentꢀ
ly minimising the formation of undesirable Fe oxides is likely
to be increased at lower Fe loadings. This will be described in
depth in the following section.
To probe whether framework or extraꢀframework Al3+ species
were responsible for the promotion, Al3+ was incorporated into
0.5Feꢀsilicaliteꢀ1 by three different techniques; namely hydroꢀ
thermal synthesis, solidꢀstate ionꢀexchange and impregnation.
From the data presented (Figure 5), it is clear that the addition
of Al3+ to 0.5Feꢀsilicaliteꢀ1550 by solidꢀstate ionꢀexchange or
impregnation does not lead to any improvements in catalytic
activity. It can thus be firmly concluded that extraꢀframework
Al3+ species do not promote the catalytic activity of Fe3+, and
that an extraꢀframework mixed oxide (FeꢀOꢀAl) is highly unꢀ
likely to be responsible for catalytic activity, as has previously
been proposed for N2Oꢀbased oxidations.21
To substantiate the proposed interaction of cationic extraꢀ
framework Fe3+ and framework Al3+, 2.5Fe3+/ZSMꢀ5 (86)550
was investigated with FTꢀIR spectroscopy. It is wellꢀknown
that when in the H+ꢀform, the cationꢀexchange sites associated
with framework T3+ atoms give rise to clear stretches between
3700 – 3600 cmꢀ1 , the exact values of which depend entirely
on the identity of the T3+ atom. If cationic Fe complexes are
indeed dispersed on the cationꢀexchange sites, the intensity of
the AlꢀO(H)ꢀSi stretch (3610 cmꢀ1) should diminish taking into
account that some (or all) of the protons would be replaced by
cationic Fe complexes. This was subsequently verified experꢀ
imentally, as the AlꢀO(H)ꢀSi band in ZSMꢀ5 (86) (Figure 6,
A/blue) is completely eliminated from the FTꢀIR spectrum
upon solidꢀstate ionꢀexchange with Fe3+ (B/red). This indicates
the substitution of all of the cationꢀexchange sites with Fe3+
during catalyst synthesis, and confirms the association of exꢀ
traꢀframework Fe3+ with the cationꢀexchange sites associated
with framework Al3+. Nevertheless, in spite of the excess of Fe
to Al in these samples, a partial restoration of the AlꢀO(H)ꢀSi
band is observed after calcination the final catalyst at 550 ºC.
The lack of 100 % exchange i.e. complete loss of OꢀH band, in
this sample, which is the active catalyst, is likely due to inꢀ
creased clustering of Fe3+ into larger clusters or bulky iron
oxides upon calcination.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Influence of Al3+ addition on the catalytic activity of
0.5Feꢀsilicaliteꢀ1550. For the postꢀsynthesis samples, Al3+ was
added to a preꢀsynthesised sample of 0.5Feꢀsilicaliteꢀ1. For 0.5Feꢀ
ZSMꢀ5 (X) prepared by hydrothermal synthesis, Fe and Al were
concurrently incorporated into the framework by the addition of
the relevant precursors to the synthesis gel.
poration, impregnation, ♦ solidꢀstate ionꢀexchange. Reaction
conditions; volume: 10 mL; time: 30 min; temp: 50 ºC: P(CH4)
ꢂ hydrothermal incorꢀ
▲
:
30.5 bar; [H2O2]: 0.5M; catalyst: 27 mg; stirring speed: 1500
rpm; Catalyst preꢀtreatment: calcination (550 ºC, 3 h, static air).
In contrast, the incorporation of Al3+ into the framework of
0.5Feꢀsilicaliteꢀ1550 during hydrothermal synthesis leads to
significant increases in catalytic activity, that are directly proꢀ
portional to the Al3+ content up to 2.2 wt. %. Given that this
method is the only method capable of incorporating Al3+ into
the zeolite framework, and is thus the only method capable of
increasing the number of cationꢀexchange sites, it can be proꢀ
posed that the increased cationꢀexchange site density is reꢀ
sponsible for the promotion displayed by Al3+, presumably by
the stabilisation and dispersion of the active Fe3+ cations onto
Figure 6. OꢀH stretching region of the FTꢀIR spectra of ZSMꢀ5
(30) (A/blue), uncalcined/as synthesised 2.5 wt. % Fe3+/ZSMꢀ5
(30) (B/red) and calcined 2.5 wt. % Fe3+/ZSMꢀ5 (30) (C/green).
Absorbances were normalised to the SiꢀOꢀSi stretches of the zeoꢀ
lite framework.
To further substantiate the hypothesis that framework Al3+
maximises the formation of the active Fe3+ species, a final
analogous sample of [Fe,Ga]ꢀsilicaliteꢀ1 was prepared by hyꢀ
drothermal synthesis. This substitution of Si4+ by Ga3+ also
gives rise to cationꢀexchange sites in a similar manner to Al3+,
and a significant increase in catalytic activity versus the Feꢀ
only analogue (Figure 7). This strengthens the identification of
cationꢀexchange sites as the promoting functionality of the
ꢀ
the negative framework charge associated with the AlO4 tetꢀ
rahedron. This would lead to a more significant interaction
between extraꢀframework Fe3+ and the zeolite framework (anꢀ
choring) and would limit the formation of bulk and undesiraꢀ
ble oxide species upon preꢀtreatment. We note here that whilst
the observed TOF for the most active sample (0.5FeꢀZSMꢀ5
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 12
zeolite in this process. Moreover, other similarities to the
Many important features are immediately evident from the
[Fe,Al]ꢀsystem were observed in that the incorporation of Ga3+
alone into the MFIꢀmaterial was insufficient for catalytic acꢀ
tivity, and that postꢀsynthesis deposition of Ga3+ was also not
beneficial to the activity of the catalyst (Table S2). It is again
clear that significant promotion is only observed when a trivaꢀ
lent heteroatom is incorporated into the zeolite framework,
thus leading to increased cationꢀexchange site density.
data shown in Figure 8. First, it is clear that as the Alꢀcontent
of the material increases, the activity of each catalyst at a givꢀ
en Fe loading is higher, likely due to the more facile extraction
of framework Fe3+ at higher Al contents. Furthermore, it is
clear that at higher Al contents, the optimal Fe loading deꢀ
creases. It is likely that due to the increased extraction, the
possibility of forming inactive larger clusters and bulk oxides
is enhanced at higher loadings. Finally, it is apparent that
above the optimal Fe loading of each FeꢀZSMꢀ5 (X) series,
catalytic activity decreases, yet the apparent decrease in activiꢀ
ty is much lower at higher Al3+ content. This is in excellent
agreement to our previous observations and confirms that Al3+
inhibits the formation of larger Fe clusters and Fe oxides, thus
maximising the amount of active Fe species, and minimising
the amount of undesirable side species. Each of these observaꢀ
tions fully supports the hypothesis that the key to attaining the
highest levels of activity is maximising the content of Al3+,
and subsequently optimising the Fe content. We note here that
we have previously proposed through EXAFS analysis and
DFT calculations that catalytic activity is due to an extraꢀ
framework species, Fe2(µ2ꢀOH)2(OH)2(H2O)2]2+, a binuclear
active site that comprises an overall +2 charge.18 The coordiꢀ
nation of such a species on to the cationꢀexchange sites would
require two exchange sites within a critical distance of ~ 5ꢀ6
Å, or two Al3+ atoms within the 10 membered MFI ring. The
probability of having two exchange sites within this distance
would improve significantly upon the incorporation of addiꢀ
tional Al3+ into the framework, and has been shown by Feng
and Hall to be very high for zeolites with a SiO2/Al2O3 mole
ratio approaching 40, but negligible for zeolites with a
SiO2/Al2O3 much greater than 100.34 This correlates favouraꢀ
bly with the observed activity.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. Catalytic activity of various Feꢀcontaining zeolites at a
fixed Fe loading (0.5 wt. %). Ga3+ and Al3+, where applicable,
were added to the material at an SiO2/M2O3 ratio of ± 85. Reacꢀ
tion conditions; cat: various (27 mg); P(CH4): 30.5 bar; [H2O2]: 0.5
M; temp: 50 °C; time: 30 min; stirring speed: 1500 rpm; Note:
Catalyst preꢀtreatment: 550 °C, 3 h, static air.
Optimised material synthesis. At this point, we have estabꢀ
lished that (1) the catalytic activity of (Fe)ꢀZSMꢀ5 correlates
to the presence of extraꢀframework Fe3+ species within the
zeolitic micropores (work herein and reference 20), (2) that the
clustering of extraꢀframework Fe3+ to larger clusters and bulk
oxides leads to a decrease in activity and Cꢀ and H2ꢀbased
selectivity,20 and (3) that the coꢀpresence of Al3+ or Ga3+ is
critical for maximising the activity of each Fe3+ atom, first by
facilitating its extraction from the zeolitic framework, and
secondly by providing cationꢀexchange sites which are capaꢀ
ble of acting as stabilising ligands to the cationic Fe3+ species,
thereby inhibiting the formation of undesirable side species
and maximising the fraction of active species. Nevertheless,
the TOFs exhibited by the optimal material so far, 0.5Feꢀ
ZSMꢀ5 (28)550, are still around 5ꢀ10 times lower than those
exhibited by commercial ZSMꢀ5 (30), containing only 0.014
wt. % Fe.18
By considering these factors, we reasoned that the extremely
high TOFs of the commercial sample might be due primarily
to its low Fe loading (0.014 wt. %) but high Al loading (2.3
wt. %); a large amount of Al3+ would maximise the extraction
of framework Fe3+ (thereby forming a greater number of active
sites), and provide sufficient cationꢀexchange sites to stabilise
all of the extracted Fe3+ (thereby inhibiting the formation of
undesirable side species). The low Fe loading would also maxꢀ
imise the percentage of ‘active’ Fe species, as the spatial disꢀ
tribution of Fe would be maximised, and the formation of bulk
oxides inhibited. In view of this, a number of Feꢀ and Alꢀ
containing ZSMꢀ5 samples were prepared (Figure 8).
Figure 8. Catalytic activity of various FeꢀZSMꢀ5 samples conꢀ
taining different Feꢀ and Alꢀcontents. Reaction conditions; cat:
ZSMꢀ5 (27 mg); P(CH4): 30.5 bar; [H2O2]: 0.5 M; temp: 50 °C;
time: 30 min; stirring speed: 1500 rpm; Note: Catalyst pretreatꢀ
ment: 550 °C, 3 h, static air.
It is important to stress here that by understanding the critical
role(s) of each component of the catalyst, we have been able to
optimise the catalytic activity of FeꢀZSMꢀ5 enormously. The
TOFs displayed by the FeꢀZSMꢀ5 (28) series – containing 2.2
wt. % Al – are two orders of magnitude higher than those of
the Feꢀonly series i.e. 0.5Feꢀsilicaliteꢀ1. Thus, the TOFs exhibꢀ
ACS Paragon Plus Environment

Page 9 of 12
ACS Catalysis
ited by this series are now very similar to the TOFs obtained
Silicaliteꢀ1, Feꢀsilicaliteꢀ1 and ZSMꢀ5 were also used as preꢀ
cursors for the incorporation of Al3+, Fe3+ and Ga3+ by postꢀ
synthesis deposition methods (impregnation, solidꢀstate and
aqueous ionꢀexchange). To this end, 2.5 wt. %Fe/ZSMꢀ5 was
prepared by impregnation through the addition of support
1
2
3
4
5
6
7
8
for the original commercial ZSMꢀ5 (30) catalyst, the most
intrinsically active catalyst to date. For example, 0.045Feꢀ
ZSMꢀ5 (28)550 and 0.095FeꢀZSMꢀ5 (28)550 oxidises methane at
a TOF of ca. 1,600 hꢀ1, comparable to the ca. 2,000 hꢀ1 that we
have observed for commercial ZSMꢀ5 (30)550. More imꢀ
portantly, however, is that these levels of TOF have been
maintained at significantly higher Fe loadings (0.095 wt. % vs.
0.014 wt. %). Thus, along with possessing similar intrinsic
activity, 0.095FeꢀZSMꢀ5 (28) is around 5ꢀtimes more producꢀ
tive in terms of volumetric productivity (1.92×10ꢀ8 mol
(product) cmꢀ3 sꢀ1 after 30 minutes) than commercial ZSMꢀ5
(30) (4.2×10ꢀ9 mol cmꢀ3 sꢀ1 after 30 minutes).
(1.95g, NH4ꢀZSMꢀ5, SiO2/Al2O3 molar ratio = 30, ZeolystTM
)
to an aqueous solution of Fe(NO3)3.9H2O (3.613mL, 13.828g
dissolved in 1L). The solution was stirred until a homogeneꢀ
ous solution was obtained. The slurry was dried (16 h, 110 ºC)
before calcination at 550 ºC for 3 h in static air. 2.5 wt. %
Fe/ZSMꢀ5 was also prepared by solidꢀstate ionꢀexchange by
adding the desired amount of Fe(acac)3 (0.158g, 0.45 mmol) to
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
NH4ꢀZSMꢀ5 (0.975g, SiO2/Al2O3 molar ratio = 30, ZeolystTM
)
prior to mechanical grinding for 30 minutes. Finally, aqueous
ionꢀexchange of was performed by the addition of NH4ꢀZSMꢀ
5 (2g, SiO2/Al2O3 molar ratio = 30, ZeolystTM) to an aqueous
solution of Fe(NO3)3.9H2O (30 mL, 0.02 M, 20.2g
Fe(NO3)3.9H2O dissolved in 250 mL deionised water). The
suspension was stirred vigorously (85 ºC, 24 h) under reflux.
The catalyst was filtered, washed with deionised water and
dried (16 h, 110 ºC). Prior to testing, the catalyst was activated
by calcination at 550 ºC for 3 h in static air.
Conclusions. Through catalytic measurements and spectroꢀ
scopic investigations, we have demonstrated that, whilst extraꢀ
framework Fe3+ species are the active component of Feꢀ
containing MFIꢀtype zeolites for selective aqueousꢀphase meꢀ
thane oxidation, significant promotion is observed upon the
incorporation of other nonꢀcatalytic trivalent cations (e.g. Al3+,
Ga3+) into the MFIꢀframework. We have rationalised this proꢀ
motion in terms of two coꢀoperative effects. Firstly, the coꢀ
addition of Al3+ or Ga3+ to the framework leads to an increased
migration of (initially) framework Fe3+ to the extraꢀframework
during heat preꢀtreatment. Concurrently, the cationꢀexchange
sites associated with framework M3+ species also are able to
stabilise and disperse the soꢀformed extraꢀframework Fe3+
species that are responsible for catalytic activity. In this case,
the dispersion of Fe3+ is a consequence of both the dispersed
nature of Al3+ within the zeolite (Lowenstein’s rules ensuring
maximum dispersion), and an ‘anchoring’ process whereby Fe
is inhibited towards agglomeration into bulk oxides due to the
stabilisation provided by the negative framework charge. By
understanding these key roles exhibited by each aspect of the
solid catalyst, significant improvements in catalytic activity
(one order of magnitude) have been obtained by the careful
and rationalised design of new catalysts. Optimal activity has
thus been obtained with a catalyst comprising of 0.095 wt. %
Fe, and 2.2 wt. % of Al3+. This catalyst performs this highly
desirable reaction at volumetric productivities of 1.92×10ꢀ8
mol cmꢀ3 sꢀ1, and at TOFs comparable to the commercial cataꢀ
lyst previously reported (ca. 2,000 hꢀ1), and is therefore the
most active catalyst reported for this challenging reaction to
date.
Catalyst characterisation. Powder Xꢀray diffraction (XRPD)
was performed using a PANalytical X’PertPRO Xꢀray diffracꢀ
tometer, with a CuKα radiation source (40 kV and 40 mA).
Diffraction patterns were recorded between 6ꢀ55° 2θ at a step
size of 0.0167° (time/step = 150 s, total time = 1 h). FTꢀIR
spectroscopy was performed by forming selfꢀsupporting waꢀ
fers from a small amount of sample and KBr. The spectra were
recorded on a Jasco FTꢀIR660 Plus over a range of 4000ꢀ400
cmꢀ1 at a resolution of 2 cmꢀ1. UVꢀVis analysis was performed
on an Agilent Cary 4000 UVꢀVisible Spectrophotometer
equipped with Diffuse Reflectance setꢀup. Samples were
scanned between 190 and 900 nm at a scan rate of 600 nm
minꢀ1. Na, Si and Al content was determined by Neutron Actiꢀ
vation Analysis (NAA). Metal contents were determined by
ICPꢀOES to an accuracy of ± 10 %.
Kinetic evaluation. Microꢀkinetic analysis was carried out in
a batch stainlessꢀsteel autoclave containing a Teflon liner vesꢀ
sel and a working volume of 35 mL. The vessel was charged
with an aqueous solution of H2O2 (10 mL, 0.5 M, 5000 ꢃmol)
and the desired amount of catalyst (typically 27 mg). After
evacuation of contaminant gases, the autoclave was heated to
the reaction temperature (typically 50 oC), and vigorously
stirred at 1500rpm once the desired temperature was obtained.
The vessel was cooled in ice (12 oC) following the appropriate
reaction time, and the resultant solution was filtered and anaꢀ
lysed. Experimental error was determined to be ± 7 %.
Experimental Section
Catalyst synthesis and pre-treatment. MFIꢀtype zeolites
containing various amounts of Fe, Al and Ga were prepared by
a hydrothermal synthesis method in a batch autoclave. The
procedure used is more completely described previously.18,20
Crystallisation was performed in a stainlessꢀsteel autoclave at
Analytical methods. Aqueousꢀphase products were identified
1
through HꢀNMR spectroscopy on a Bruker 500MHz Ultra-
Shield NMR spectrometer, and quantified against a 1 vol. %
TMS/CDCl3 internal standard, previously calibrated against
authentic standards. The detection limit was optimised to a
level corresponding to a product yield of 0.1 µmol.20 End H2O2
concentrations were determined by titration against acidified
Ce(SO4)2 solution. Gaseous phase products were quantified by
means of an FIDꢀGC (Varian 450-GC) fitted with a CPꢀSil
5CB capillary column (50m length, 0.32mm ID). The GC was
equiped with a methaniser unit, and CO2 was quantified
against a calibration curve constructed from commercial
standards (BOC).
o
175 C for 120 h. The as synthesised materials obtained were
o
o
calcined at 550 C (1 C minꢀ1), first in a flow of nitrogen (5
h), and later air (3 h) in order to remove the organic template.
The deꢀtemplated sample was subsequently ionꢀexchanged
o
threeꢀtimes with NH4NO3 (1.0 M) at 95 C, and later dried for
o
16 h at 110 C. Activation was achieved by calcination in
flowing air (30 mL minꢀ1) at 550 oC, 750 ºC or 900 ºC for 3 h.
Following this route, ferrigallosilicate ([Fe, Ga]), ferrisilicate
([Fe]), aluminosilicate ([Al]) and silicate (no heteroatom) were
also prepared.
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 12
(4)
Fiedler, E.; Grossmann, G.; Keresbohm, D. B.;
AUTHOR INFORMATION
Corresponding Authors
Weiss, G.; Witte, C. In Ullmann's Encyclopaedia of Industrial
Chemistry; WileyꢀVCH: Weinheim, 2011; Vol. 23, p 26.
1
2
3
4
5
6
7
8
(5)
Chang, C. D.; Silvestri, A. J. Journal of
* Dr. Ceri Hammond, Department of Chemistry and Applied Bioꢀ
sciences, ETH Zürich, WolfgangꢀPauliꢀStr. 10, CHꢀ8093 Zürich,
Switzerland.
Catalysis 1977, 47, 249.
(6)
Periana, R. A.; Taube, D. J.; Gamble, S.; Taube,
H.; Satoh, T.; Fujii, H. Science 1998, 280, 560.
ceri.hammond@chem.ethz.ch
(7)
Palkovits, R.; Antonietti, M.; Kuhn, P.; Thomas,
Prof. Graham J. Hutchings, Cardiff Catalysis Institute, School of
Chemistry, Cardiff University, Main Building, Park Place, Carꢀ
diff, CF10 3AT (UK)
A.; Schueth, F. Angewandte Chemie-International Edition 2009, 48,
6909.
(8)
Dubkov, K. A.; Sobolev, V. I.; Panov, G. I.
9
hutch@cardiff.ac.uk
Kinetics and Catalysis 1998, 39, 72.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(9)
Dubkov, K. A.; Sobolev, V. I.; Talsi, E. P.;
Rodkin, M. A.; Watkins, N. H.; Shteinman, A. A.; Panov, G. I.
Journal of Molecular Catalysis a-Chemical 1997, 123, 155.
Funding Sources
This work formed part of the Methane Challenge. The Dow
Chemical Company is thanked for their financial support. Compuꢀ
ting facilities for this work were provided by ARCCA at Cardiff
University, HPC Wales, and the UK’s HPC Materials Chemistry
Consortium (MCC). The MCC is funded by EPSRC
(EP/F067496).
(10)
Groothaert, M. H.; Smeets, P. J.; Sels, B. F.;
Jacobs, P. A.; Schoonheydt, R. A. Journal of the American Chemical
Society 2005, 127, 1394.
(11)
Vanelderen, P.; Hadt, R. G.; Smeets, P. J.;
Solomon, E. I.; Schoonheydt, R. A.; Sels, B. F. Journal of Catalysis
2011, 284, 157.
(12)
Himes, R. A.; Karlin, K. D. Current Opinion in
Chemical Biology 2009, 13, 119.
ABBREVIATIONS AND DEFINITIONS
Partial oxygenated products;
(13)
Himes, R. A.; Karlin, K. D. Proceedings of the
National Academy of Sciences of the United States of America 2009,
106, 18877.
(14)
Raja, R.; Ratnasamy, P. Applied Catalysis a-
methyl hydroperoxide (CH3OOH), methanol (CH3OH) and
formic acid (HCOOH).
General 1997, 158, L7.
(15)
Sorokin, A. B.; Kudrik, E. V.; Alvarez, L. X.;
Afanasiev, P.; Millet, J. M. M.; Bouchu, D. Catalysis Today 2010,
157, 149.
Total product formed; partial oxygenates + CO2.
Oxygenate selectivity;
(16)
Sorokin, A. B.; Kudrik, E. V.; Bouchu, D.
Chemical Communications 2008, 2562.
(17)
Forde, M. M.; Grazia, B. C.; Armstrong, R.;
Jenkins, R. L.; Ab Rahim, M. H.; Carley, A. F.; Dimitratos, N.;
LopezꢀSanchez, J. A.; Taylor, S. H.; McKeown, N. B.; Hutchings, G.
J. Journal of Catalysis 2012, 290, 177.
moles ꢀpartial oxygenatesꢁ
ꢂ 100
(18)
Hammond, C.; Forde, M. M.; Ab Rahim, M. H.;
moles ꢀtotal oxygenated productꢁ
Thetford, A.; He, Q.; Jenkins, R. L.; Dimitratos, N.; LopezꢀSanchez,
J. A.; Dummer, N. F.; Murphy, D. M.; Carley, A. F.; Taylor, S. H.;
Willock, D. J.; Stangland, E. E.; Kang, J.; Hagen, H.; Kiely, C. J.;
Hutchings, G. J. Angewandte Chemie-International Edition 2012, 51,
5129.
Turnover frequency;
moles (oxygenated species formed) molꢀ1 (Fe) hꢀ1.
Volumetric productivity;
(19)
Hammond, C.; Jenkins, R. L.; Dimitratos, N.;
LopezꢀSanchez, J. A.; ab Rahim, M. H.; Forde, M. M.; Thetford, A.;
Murphy, D. M.; Hagen, H.; Stangland, E. E.; Moulijn, J. M.; Taylor,
S. H.; Willock, D. J.; Hutchings, G. J. Chemistry-a European Journal
2012, 18, 15735.
(20)
Hammond, C.; Dimitratos, N.; Jenkins, R. L.;
moles (oxygenated species formed) cmꢀ3 (reactor volume) sꢀ1.
LopezꢀSanchez, J. A.; Kondrat, S. A.; Ab Rahim, M. H.; Forde, M.
M.; Thetford, A.; Taylor, S. H.; Hagen, H.; Stangland, E. E.; Kang, J.;
Moulijn, J. M.; Willock, D. J.; Hutchings, G. J. ACS Catalysis 2013,
Accepted for publication; DOI: 10.1021/cs3007999.
Volumetric productivity and TOFs were calculated based on
the final yield at the end of the reaction i.e. they are an averꢀ
aged value over the entire time scale of the reaction. Typically,
the initial productivities and TOFs i.e. productivities at 2 and 5
minutes of reaction, were one to two orders of magnitude
higher.
(21)
Hensen, E. J. M.; Zhu, Q.; Janssen, R. A. J.;
Magusin, P.; Kooyman, P. J.; van Santen, R. A. Journal of Catalysis
2005, 233, 123.
(22)
(23)
Notte, P. P. Topics in Catalysis 2000, 13, 387.
Hensen, E. J. M., Zhu, Q. J., Liu, P. H., Chao, K.
J., van Santen, R. A. Journal of Catalysis 2004, 221, 560.
(24)
Battiston, A.A., Bitter, J. H., Heijboer, W. M., de
REFERENCES
Groot, F.M.F., Koningsberger, D. C., Journal of Catalysis 2003, 215,
(1)
Hammond, C.; Conrad, S.; Hermans, I.
279.
Chemsuschem 2012, 5, 1668.
(25)
Battiston, A.A., Bitter, Koningsberger, D. C.,
(2)
Golisz, S. R.; Gunnoe, T. B.; Goddard, W. A.,
Journal of Catalysis 2003, 218, 163.
III; Groves, J. T.; Periana, R. A. Catalysis Letters 2011, 141, 213.
(26)
Zecchina, A.; Rivallan, M.; Berlier, G.;
(3)
Hammer, G.; Lubcke, T.; Kettner, R.; Pillarella,
Lamberti, C.; Ricchiardi, G. Physical Chemistry Chemical Physics
2007, 9, 3483.
M. R.; Recknagel, H.; Commichau, A.; Neumann, H.ꢀJ.; Paczynskaꢀ
Lahme, B. In Ullmann's Encyclopaedia of Industrial Chemistry;
WileyꢀVCH: Weinheim, 2006; Vol. 23, p 739.
(27)
Bordiga, S., Buzzoni, R. Geobaldo, F. Lamberti,
C., Giamello, E., Zecchina, A., Leofanti, G., Petrini, G., Tozzola, G.,
Vlaic, G. Journal of Catalysis 1996, 158, 486.
ACS Paragon Plus Environment

Page 11 of 12
ACS Catalysis
(28)
PerezꢀRamirez, J., Groen, G. C., Bruckner, A.,
(31)
Hensen, E. J. M.; Zhu, Q.; Hendrix, M.;
Kumar, M. S., Bentrup, U., Debbagh, M. N., Villaescusa, L. A.
Journal of Catalysis 2005, 232, 318ꢀ334.
Overweg, A. R.; Kooyman, P. J.; Sychev, M. V.; van Santen, R. A.
Journal of Catalysis 2004, 221, 560.
1
2
3
4
(29)
Kumar, A. S.; PerezꢀRamirez, J.; Debbagh, M.
(32)
PerezꢀRamirez, J. Journal of Catalysis 2004,
N.; Smarsly, B.; Bentrup, U.; Bruckner, A. Applied Catalysis B-
Environmental 2006, 62, 244.
227, 512.
(34)
Feng, X. B.; Hall, W. K. Journal of Catalysis
5
(30)
PerezꢀRamirez, J.; Groen, J. C.; Bruckner, A.;
1997, 166, 368.
Kumar, M. S.; Bentrup, U.; Debbagh, M. N.; Villaescusa, L. A.
Journal of Catalysis 2005, 232, 318.
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Insert Table of Contents artwork here
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
130x101mm (150 x 150 DPI)
ACS Paragon Plus Environment