Identification of Active and Spectator Sn Sites in Sn-β Following Solid-State Stannation, and Consequences for Lewis Acid Catalysis

Article abstract of DOI:10.1002/cctc.201500545

Lewis acidic zeolites are rapidly emerging liquid-phase Lewis acid catalysts. Nevertheless, their inefficient synthesis procedure currently prohibits greater utilization and exploitation of these promising materials. Herein, we demonstrate that Sn^IV-containing zeolite beta can readily be prepared both selectively and extremely rapidly by solid-state incorporation (SSI) method. Through a combination of spectroscopic (XRD, UV/Vis, X-ray absorption, magic-angle spinning NMR, and diffuse reflectance infrared Fourier transform spectroscopy) studies, we unambiguously demonstrate that site-isolated, isomorphously substituted Sn^IV sites dominate the Sn population up to a loading of 5 wt % Sn. These sites are identical to those found in conventionally prepared Sn-beta, and result in our SSI material exhibiting identical levels of intrinsic activity (that is, turnover frequency) despite the threefold increase in Sn loading, and the extremely rapid and benign nature of our preparation methodology. We also identify the presence of spectator sites, in the form of Sn^IV oligomers, at higher levels of Sn loading. The consequences of this mixed population with regards to catalysis (Meerwein-Pondorf-Verley reaction and glucose isomerization) are also identified.

Full text of DOI:10.1002/cctc.201500545

DOI: 10.1002/cctc.201500545
Full Papers
Identification of Active and Spectator Sn Sites in
Sn-b Following Solid-State Stannation, and Consequences
for Lewis Acid Catalysis
Ceri Hammond,*[a] Daniele Padovan,[a] Abbas Al-Nayili,[a] Peter. P. Wells,[b, c]
Emma K. Gibson,[b, c] and Nikolaos Dimitratos[a]
Lewis acidic zeolites are rapidly emerging liquid-phase Lewis
acid catalysts. Nevertheless, their inefficient synthesis proce-
dure currently prohibits greater utilization and exploitation of
these promising materials. Herein, we demonstrate that SnIV-
containing zeolite beta can readily be prepared both selective-
ly and extremely rapidly by solid-state incorporation (SSI)
method. Through a combination of spectroscopic (XRD, UV/
Vis, X-ray absorption, magic-angle spinning NMR, and diffuse
reflectance infrared Fourier transform spectroscopy) studies,
we unambiguously demonstrate that site-isolated, isomor-
phously substituted SnIV sites dominate the Sn population up
to a loading of 5 wt% Sn. These sites are identical to those
found in conventionally prepared Sn-beta, and result in our SSI
material exhibiting identical levels of intrinsic activity (that is,
turnover frequency) despite the threefold increase in Sn load-
ing, and the extremely rapid and benign nature of our prepa-
ration methodology. We also identify the presence of spectator
sites, in the form of SnIV oligomers, at higher levels of Sn load-
ing. The consequences of this mixed population with regards
to catalysis (Meerwein–Pondorf–Verley reaction and glucose
isomerization) are also identified.
Introduction
Lewis acid catalysts have found widespread applicability for
a number of liquid-phase oxidation and isomerization reac-
tions.[1,2] Amongst this class of catalysts are Lewis acidic zeo-
lites, which are important catalysts in the area of sustainable
chemistry.[3–5] These crystalline, porous, inorganic Lewis acids
possess several major advantages over conventional homoge-
neous analogues, such as AlCl3 and ZnCl2. In addition to the
practical benefits of solid catalysts, which simplify downstream
processing and various aspects of process intensification, the
incorporation of Lewis acidic centers into hydrophobic frame-
works inhibits hydrolysis and deactivation of the Lewis acid in
the aqueous phase, and thereby allows these promising mate-
rials to be utilized as catalysts for a range of emerging aque-
ous-based reactions.[6]
Of particular interest is Sn–zeolite-b. With promising results
obtained by many groups for the isomerization of glucose to
fructose,[7–12] the Baeyer–Villiger (BV) oxidation of ketones to
lactones using H2O2 as green oxidant,[4,13,14] Meerwein–Pon-
dorf–Verley (MPV) transfer hydrogenations,[15,16] and various
other emerging catalytic transformations,[17–19] the potential of
Sn-b has attracted significant industrial interest. Nevertheless,
some significant practical hurdles currently curtail industrial ex-
ploitation. Amongst these limitations are 1) the lengthy and
complicated hydrothermal synthesis procedure, 2) the low
amount of active metal incorporated per kilogram of final cata-
lyst, resulting in low space–time yields, and 3) the large crystal-
lite sizes obtained via typical hydrothermal synthesis, resulting
in mass-transfer issues.
Given the lack of a widely applicable and scalable prepara-
tion methodology, significant academic and industrial research
has recently focused on the development of new methodolo-
gies for preparing Sn-b.[19–24] Recently, Hammond and co-work-
ers[20] demonstrated how simple, post-synthetic modifications
of commercially available zeolites could readily enhance the
availability and practicality of Lewis acidic zeolites for sustaina-
ble catalytic transformations.[20] Dealumination of a commercial-
ly available b zeolite with HNO3 was found to lead to a highly
siliceous framework possessing vacant tetrahedral (T)-sites,
into which SnIV could be placed by means of solid-state incor-
poration (SSI) with tin(II) acetate. Not only does this approach
avoid the long synthesis timescales associated with conven-
tional Sn-b synthesis procedures (the total time for mechano-
chemical and heat-treatment steps is only 8 h), but it also
allows for the synthesis of a material with significantly smaller
[a] Dr. C. Hammond, D. Padovan, A. Al-Nayili, Dr. N. Dimitratos
Cardiff Catalysis Institute, Cardiff University School of Chemistry
Main Building, Park Place, Cardiff, CF10 3AT (UK)
E-mail: hammondc4@cardiff.ac.uk
[b] Dr. P. P. Wells, Dr. E. K. Gibson
UK Catalysis Hub, Research Complex at Harwell
Oxon, OX11 0FA (UK)
[c] Dr. P. P. Wells, Dr. E. K. Gibson
Department of Chemistry, University College London
Gordon Street, London, WC1H 0AJ (UK)
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201500545. Open data for this article are
at http://dx.doi.org/10.17035/d.2015.100100.
ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.
This is an open access article under the terms of the Creative Commons
Attribution License, which permits use, distribution and reproduction in
any medium, provided the original work is properly cited.
ChemCatChem 2015, 7, 3322 – 3331
3322ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
crystallite sizes, increased crystallinity, greater metal content,
and improved catalytic performance. Indeed, SSI Sn-b was pre-
viously found to be up to one order of magnitude more active
on a space-time-yield scale than conventionally prepared Sn-
b both for BV oxidations and the conversion of dihydroxyace-
tone into ethyl lactate.[20] Furthermore, aqueous/liquid solu-
tions of Sn are also avoided, and complicated/sensitive vapour
deposition methods are avoided.
lysts and their preliminary catalytic activity for MPV transfer hy-
drogenation are presented. Complete dealumination of Al-
b (Zeolyst, experimentally determined SiO2/Al2O3 =26.5) was
achieved by treatment in HNO3 following established literature
protocols (13m HNO3, 1008C, 20 h, 20 mLgÀ1).[20] In agreement
to our prior studies, treatment of zeolite b with HNO3 does not
affect the b structure, as indicated by the lack of major
changes in micropore volumes (Vmicro, Table 1 and Table S1 in
the Supporting Information), and the powder XRD (pXRD) pat-
tern (Figure 1 and Figure S1), and results in a highly siliceous
In our original communication,[20] characterization of SSI Sn-
b was investigated by means of UV/Vis and Raman spectrosco-
py, and preliminary catalytic data focusing on BV oxidation and
ethyl lactate synthesis was obtained. However, for a more thor-
ough evaluation of SSI Sn-b, and to fully compare its catalytic
performance relative to the conventional analogue, further ki-
netic studies of additional catalytic reactions, such as the iso-
merization of glucose and MPV transfer hydrogenation, are re-
quired. Additionally, advanced characterization of SSI Sn-b with
more sensitive spectroscopic techniques, such as X-ray absorp-
tion spectroscopy (XAS) and solid-state magic-angle spinning
(MAS) NMR, is essential. Only by performing these studies can
a full comparison between conventional Sn-b, and that pre-
pared by SSI, be obtained. In this manuscript, we extend our
investigations of this particular catalytic material, and demon-
strate both through catalytic and spectroscopic methodologies
that SSI Sn-b possesses identical site speciation and activity to
conventional Sn-b at loadings up to 5 wt%, which are a factor
of three higher than can be obtained by conventional hydro-
thermal synthesis methods. Nevertheless, some inactive sites
are formed at higher loadings of Sn. The identity of these spec-
tator sites, and their consequences for Lewis acid catalysis, is
rationalized with XAS, MAS NMR, UV/Vis, XRD and diffuse re-
flectance infrared Fourier transform spectroscopy studies
(DRIFTS).
Figure 1. pXRD patterns of various Sn-b catalysts prepared by SSI, and con-
taining different Sn contents. The inset gives an enlargement of the indicat-
ed spectral region of 10Sn-b catalyst.
b zeolite material (SiO2/Al2O3 >1000) possessing vacant frame-
work sites, within which SnIV can be incorporated through
post-synthetic methodologies. SnIV was subsequently intro-
duced into these vacant sites by solid-state incorporation, by
mechanochemical treatment of the dealuminated b zeolite
with the appropriate quantity of tin(II) acetate for 10 min in
a pestle and mortar. Calcination of the resulting homogeneous
powder at 5508C subsequently yielded the final catalytic
powder. Metal loadings were determined by inductively cou-
pled plasma mass spectrometry (ICP–MS), and are consistent—
within experimental error—to the nominal loadings used
during preparation. Porosymmetry measurements (Table 1,
Table S1) reveal that each of the synthesized materials possess-
es similar micropore volumes (Vmicro) to conventional Sn-b, al-
though a small but steady decrease in Vmicro is observed as the
metal content is increased beyond 5 wt%. Clearly, increased Sn
concentrations beyond this point lead to partial pore blockage,
and to decreased accessibility of the micropores of zeolite b.
Preliminary characterization of the synthesized catalysts was
performed with pXRD (Figure 1) and UV/Vis spectroscopy (Fig-
ure S2). Analysis of the ligand-to-metal charge transfer bands
of metal-containing zeolites with diffuse reflectance UV/Vis
spectroscopy (DRUV/Vis) provides a great deal of insight of the
metal site speciation.[25] The UV/Vis spectra obtained for the
Results and Discussion
According to a modified version of our original synthesis pro-
tocol,[20] we prepared a range of Sn-b catalysts with nominal
loadings between 2 and 10 wt% Sn by SSI. In Table 1, the vari-
ous physical and chemical properties of the synthesized cata-
Table 1. Physical and catalytic properties of Sn-containing zeolites pre-
pared by SSI.
Catalyst
Sn wt%[a] (Si/Sn)[b]
Vmicro
XCyO
[c]
[cm3 gÀ1
]
[%][d]
Al-b
–
0.23
0.23
0.23
0.21
0.19
0.18
0.20
–
0
deAl-b
–
2Sn-b
5Sn-b
8Sn-b
10Sn-b
conventional Sn-b[e]
1.8 (75)
4.5 (29)
7.3 (18)
9.2 (15)
Æ1.5
100
100
95.3
86.3
95.4
[a] Determined by ICP–MS; [b] Experimentally determined Si/Sn molar
ratio; [c] Micropore volume, calculated by the t-plot method; [d] MPV re-
action conditions: 1008C, 1 h, 10 mL cyclohexanone in 2-butanol (0.2m),
CyO/Sn molar ratio of 100; [e] From Ref. [15].
ChemCatChem 2015, 7, 3322 – 3331
www.chemcatchem.org
3323ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
synthesized samples (containing between 2 and 10 wt% Sn) all
exhibit a sharp maximum at approximately 219 nm, indicative
of isolated, tetrahedral SnIV species that are isomorphously
substituted into the zeolite framework. Each spectrum is also
significantly blue-shifted with respect to bulk SnO2, which ex-
hibits a broad, characteristic absorbance at 280 nm, further in-
dicating the presence of site-isolated SnIV sites.[16,20] Although
the absorbance maxima remains constant in each material,
there is, however, a clear shift in absorbance edge following
the incorporation of greater quantities of Sn. The absorbance
edge appears to shift from approximately 250 nm/5.0 eV for
2Sn-b, to approximately 282 nm/4.4 eV in 10Sn-b. Although
the absorbance edge falls well short of that observed for bulk
SnO2 (approximately 350 nm/3.7 eV), this shift in band edge in-
dicates that at least some extraframework Sn species are pres-
ent in these samples, although DRUV/Vis is not sufficiently sen-
sitive to provide a quantitative value of their presence, particu-
larly since the extinction coefficients are unknown. However,
given the relative sharpness of each peak, any extraframework
Sn species are likely to be oligomeric (SnxOy) as opposed to
oxidic (SnO2).
Figure 2. Catalytic activity of Sn-b catalysts prepared by SSI, and containing
different Sn contents, for MPV hydrogenation of CyO. Reaction conditions:
1008C, 1 h, 10 mL cyclohexanone in 2-butanol (0.2m), CyO/Sn molar ratio of
100.
tively negligible level. With the exception of 10Sn-b, each Sn-
b catalyst prepared by SSI demonstrated comparable activity
and selectivity to conventional Sn-b prepared by hydrothermal
synthesis, as reported by Corma and co-workers,[15] indicating
that the SnIV site speciation in each catalyst is comparable to
that of conventional Sn-b.
As described above, the dealumination of Al-b leads to no
major changes in the pXRD pattern, suggesting that the
b framework remains fully intact after acidic treatment in
HNO3. However, a shift in the calculated d-spacing following
dealumination (Æ3.93 to 3.90 Š for the main reflection in Al-
b and deAl-b, respectively (Figure S1B)) suggests that there is
a contraction in unit cell volume following the removal of Al
(Figure 1). Upon incorporation of SnIV into the vacant frame-
work sites of this material by SSI of tin(II) acetate, no obvious
changes in diffraction pattern are observed, although a slight
re-expansion of the unit cell volume can be observed (d-spac-
ing of 3.93 Š for 2Sn-b), which strongly indicates that Sn has
occupied the vacant framework sites. SnO2 produces a very
clear pXRD pattern, and can readily be observed by pXRD at
levels of <1 wt% (detailed description in Figure S1A). The
complete absence of any SnO2 reflections at 26.7 and 51.88 2q
in all the catalytic materials, even those loaded with 10 wt%
Sn, suggests that either SnO2 is not present in the materials, or
that any extraframework Sn domains are so small (<5 nm) and
heterogeneous that they do not possess the sufficient long-
range order to produce a clear diffraction pattern. Clearly, char-
acterization of the materials with more sensitive techniques,
such as MAS NMR and XAS, is required to clarify the presence
or absence of extraframework Sn species.
Nevertheless, more detailed kinetic analysis (Figure 2) shows
this not to be the case, and demonstrates that the various cat-
alysts possess markedly different catalytic activity. For these
catalytic reactions, an identical quantity of Sn (1 mol% relative
to CyO) was utilized per catalytic experiment by varying the
mass of catalyst, to accurately compare specific SnIV site activi-
ty. In all cases, cyclohexanol and 2-butanone were the sole re-
action products, and carbon balances were above 95%, sug-
gesting that SnIV loading does not impact reaction selectivity
(that is, product distribution) or any adsorption effects. By ex-
amining the full kinetic profile, and particularly the early time
course of the reaction, it becomes clear that that as the SnIV
loading of Sn-b is increased systematically from 2 to 10 wt%,
and the amount of Sn present in the reactor is kept constant,
catalytic activity per mole of Sn decreases. Indeed, 8Sn-b and
10Sn-b are significantly less active per mole of Sn than 2Sn-
b and 5Sn-b, which are almost identical in catalytic activity.
This decrease in activity at elevated loadings strongly suggests
that above Sn loadings of 5 wt%, inactive, that is, spectator,
SnIV sites are produced, effectively decreasing the amount of
active Sn present in the reactor, and resulting in decreased cat-
alytic activity. To gain further critical insight into the influence
of SnIV loading and catalytic activity, we also explored the cata-
lytic activity of Sn-b for catalytic glucose isomerization, which
is also catalyzed by framework SnIV sites in zeolite b, reportedly
through an intramolecular 1,2-hydride shift mechanism.[8,26,27]
In Figure 3 the time online kinetic profile for glucose isomer-
ization over the SSI Sn-b series is presented, whilst in Table 2
glucose conversion, fructose and mannose yields, and carbon
balance for each catalytic reaction is given at levels close to
iso-conversion. Reactions were performed under identical con-
To evaluate the activity of the synthesized catalysts, we first
focused on the MPV transfer hydrogenation of carbonyl com-
pounds such as cyclohexanone (CyO), as this is a key reaction
catalyzed by Sn-b zeolites and other Lewis acidic analogues.
Preliminary kinetic analysis under conditions comparable to
those of Corma et al.[15] initially revealed (Table 1) that each of
the synthesized catalysts exhibits similar catalytic activity for
the MPV transfer hydrogenation of cyclohexanone. In all cases,
cyclohexanol and 2-butanone were the only products detect-
ed, and carbon balances above 95% were obtained, suggest-
ing that any potential byproducts produced below the detect-
ability limit of the analytical instrument are present at a rela-
ChemCatChem 2015, 7, 3322 – 3331
www.chemcatchem.org
3324ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Sn species could be differentiated through catalytic means[10]
(by monitoring the epimerization/isomerization ratio if per-
forming the glucose reaction in methanol), more recent studies
have shown that the isomerization pathway dominates regard-
less of solvent choice,[11] and that the modifications in selectivi-
ty observed in previous cases was owing to contamination of
samples with alkali metal salts, such as Li+ and K+, and not dif-
ferences in site speciation. Thus, no clear trends between Sn
loading and selectivity were identified (Table 2) despite the dif-
ferences in activities observed. In all cases, fructose was the
main reaction product, and low levels of mannose formation
were also observed. As such, identifying the active site ensem-
bles present in these materials relies upon advanced spectro-
scopic study (see below).
To gain further insight into this decrease in activity at elevat-
ed Sn content, we calculated the initial TOF per mole of Sn in
each SSI Sn-b catalyst for both catalytic reactions. Given the
difficulties associated with sampling such rapid catalytic pro-
cesses at such short time intervals, the experimental errors de-
termined for these samples is considerably higher (Æ10%)
than those experienced during most of our catalytic protocols
(typically Æ5%). Nevertheless, it is clear from Figure 4 that as
Figure 3. Catalytic activity of various Sn-b catalysts prepared by SSI, and
containing different Sn contents, for the isomerization of glucose. Reaction
conditions: 110 8C, 0.5 h, 10 wt% glucose solution (Æ0.61m); glucose/Sn
molar ratio of 50.
Table 2. Glucose conversion (X), fructose and mannose yields (Y), and
carbon balances (C) obtained during catalytic isomerization with various
SSI Sn-b samples. Data is presented close to iso-conversion.[a]
Catalyst
Time
[min]
XGluc
[%]
YFru
[%]
YMan
[%]
Cbal
[%]
2Sn-b
5Sn-b
8Sn-b
10Sn-b
10
10
10
15
22.3
22.9
20.0
20.1
16.4
15.1
13.1
13.8
4.4
3.2
2.9
3.1
98.6
95.4
96.0
96.7
[a] Reaction conditions: 110 8C, 0.5 h, 10 wt% glucose solution (Æ0.62m);
glucose/Sn molar ratio of 50.
ditions to those described by Roman-Leshkov et al.[9] Only cata-
lytic data up to Æ25% conversion is displayed, as a significant
increase in byproduct (humin) formation and a large decrease
in carbon balance was observed above this level of conversion.
In good agreement to the MPV reaction, both 2Sn-b and 5Sn-
b catalyze the reaction to a similar extent, with >25% conver-
sion readily achieved in 15 min of reaction time under the con-
ditions employed. The initial turnover frequencies (TOF, calcu-
lated as moles of substrate converted per mole of Sn, as deter-
mined by ICP–MS, per hour) determined in both of these cases
(112 and 102 hÀ1, respectively) are almost identical, if not
slightly higher, to the value calculated from the original report
of Roman-Leshkov et al. (96 hÀ1).[9] The almost identical initial
TOFs observed for SSI Sn-b at loadings of up to 5 wt%, com-
pared to open literature values, indicate that the SSI materials
contain an identical active site distribution to that found in the
conventional analogue. This is in good agreement with the
MPV results and our initial spectroscopic data. Nevertheless, as
the loading is elevated to 8 wt%, catalytic activity per mole of
Sn decreases, and at a loading of 10 wt%, 10Sn-b is considera-
bly less reactive, strongly suggesting that spectator sites are
a component of these materials at these loadings. Although it
was originally proposed that framework and extraframework
Figure 4. Initial TOF for each Sn-b catalyst as a function of total Sn loading
for both glucose isomerization (left axis, calculated at 5 min reaction time)
and MPV transfer hydrogenation (right axis, calculated at 2 min reaction
time). TOF calculated as moles converted per mole Sn per hour.
Sn loading increases above Æ5 wt%, there is a considerable
decrease in TOF, strongly indicating that inactive SnIV sites are
present in the higher loaded materials. Indeed, as the TOF of
the 10 wt% sample in both cases is approximately one half of
that of the 5 wt% sample, there are strong indications that po-
tentially all of the additional Sn loading in the higher loaded
materials is present in a catalytically inactive, extraframework
form, or that the additional Sn loading decreases activity by
other means, for example, loss of micropore access, as de-
scribed above.
Noticeably, the relationship between initial activity (TOF) and
Sn loading (Figure 4) is remarkably similar for both MPV trans-
fer hydrogenation and glucose isomerization experiments. The
ChemCatChem 2015, 7, 3322 – 3331
www.chemcatchem.org
3325ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
similar TOF plots obtained strongly indicates that 1) the same
active sites and species are responsible for catalytic activity in
both reactions and that 2) inactive SnIV sites, possessing lower
(or no) Lewis acidity, are formed in Sn-b catalysts as the load-
ing is increased substantially beyond 5 wt%. It is thus evident
that as the Sn content of SSI Sn-b is elevated above 5 wt%,
spectator sites are formed during the calcination process,
which negatively impact catalytic activity. We note that this is
in contrast to our original communication, which had indicated
that similar SnIV site activity was observed between 5 and
10 wt% samples. This deviation may be caused by several fac-
tors, but likely arises from 1) the modified synthesis procedure
(see Experimental Section), which has improved the relative
TOFs of the 2 and 5 wt% samples, but may change the overall
Sn site speciation at higher loadings; 2) the different reactions
undergoing study; and 3) the different source of zeolite pre-
cursor utilized.
Figure 5. A) EXAFS k3 weighted c data and B) magnitude of the non-phase-
corrected Fourier transform, of dehydrated 5Sn-b (blue) and SnO2 (black).
Clearly the synthesis of Sn-b by SSI can lead to a material
containing multiple Sn species at very high Sn loadings, al-
though catalytic activity—and hence Sn site speciation—is re-
markable similar to that found in the conventional analogue
even at significantly higher, that is, a factor of three, loadings
(Æ5 wt%). To gain further insight into the active site speciation
of Sn-b as a function of loading, and to gain further insight
into the nature of the active and spectator sites present, we
subsequently focused upon the characterization of the Sn site
speciation of SSI Sn-b with a variety of spectroscopic tech-
niques. X-ray absorption fine structure (XAFS) analysis is a pow-
erful technique for determining the local structure of substitut-
ed metal sites within framework architectures, with the near-
edge structure (XANES) also providing information on the oxi-
dation state. In Figure S3 the normalized Sn K edge XANES
data of 2Sn-b, 5Sn-b, and 10Sn-b zeolite samples are shown
alongside reference spectra of SnO2 and Sn foil.[27] The maxima
of the first derivative of the XANES data are used to determine
the position of the absorption edge. Sn foil and SnO2 referen-
ces have edge positions of 29200.0 eV and 29205.1 eV, respec-
tively. All samples have an edge position consistent with that
of SnO2, suggesting the presence of SnIV after the 5508C pre-
treatment.
14 ŠÀ1. The first features present in the FT data are assigned to
SnÀO scattering interactions, whilst SnÀSn scattering paths can
be identified by the relatively intense features between 2.5
and 4 Š. SnO2 is also readily identified by the characteristic
À1
splitting of the oscillation around 10 Š in the EXAFS c data
(Figure 5A). Clearly the first peak in the FT spectrum is shifted
to a lower radial distance for 5Sn-b relative to SnO2. This is re-
flected in change in first-shell oxygen distance (SnÀO) from
Æ2.05 Š in SnO2 to 1.95 Š in 5Sn-b, which is fully consistent
with a change in geometry from octahedral to tetrahedral. This
is also consistent with Sn atoms that are substituted into the
vacant sites of the b framework.[28] Between 2.5 and 4 Š there
are a number of intense scattering peaks arising from SnÀSn
scattering interactions in SnO2. Despite their very low level, it
is clear that there is a nonnegligible SnÀSn component in 5Sn-
b following calcination at 5508C. This can arise either from the
presence of extraframework, oligonuclear Sn clusters present
within the zeolite micropores (SnxOy species), a bulky SnO2
component on the external zeolite surface, or potentially
through Sn–Sn pairing.[29] Given the complete absence of SnO2
in all the pXRD patterns (Figure 1), the narrow full width at
half maximum of the UV/Vis data (Figure S2), and the relatively
low intensity of the SnÀSn scattering pattern in the FT spec-
trum, it is likely that Sn–Sn interactions arise from a minor, but
nonnegligible extraframework, oligomeric SnxOy component in
5Sn-b, although a possible contribution from SnÀSn pairs
cannot be completely excluded. The SnÀSn coordination num-
bers, as derived from the EXAFS fitting, are also consistent
with this hypothesis (Tables 2–3). We note that the observation
of oligomeric SnxOy species in Sn-b is not unexpected, given
that EXAFS and TEM analysis of conventional Sn-b revealed
that even at low loadings (<2 wt%), conventionally prepared
Sn-b contains an extraframework Sn component.[29] Indeed,
comparison of our experimental spectra to those presented by
Bare et al.[29] for conventional Sn-b indicates that an identical
Sn site population is present in our material, despite the three-
fold increase in Sn loading. From this initial analysis, it appears
that the majority of the Sn atoms in 5Sn-b are substituted into
the crystalline zeolite framework, although a minor fraction of
extraframework SnxOy is potentially present in this material.
The Sn K edge extended X-ray absorption fine structure
(EXAFS) technique provides information about the local struc-
ture of the Sn within these samples. For completeness and
clarity, we analyzed the SSI Sn-b samples both in the hydrated
and dehydrated form. Whilst the majority of EXAFS analysis on
metal substituted zeolites are performed following dehydration
(to minimize the influence of water coordination), we feel that
presenting the EXAFS data in both dehydrated and hydrated
form is more suitable. The dehydrated data allows us to fully
compare our results with the open literature, whilst the hydrat-
ed form is more representative of the catalytic material that
operates in the liquid/aqueous phase.
The EXAFS spectra of dehydrated 5Sn-b (our optimal materi-
al in terms of loading and TOF) and the SnO2 reference are
shown in Figure 5A, whilst Figure 5B presents the k3 weighted
Fourier transform spectra of the same two samples. The spec-
tra are of excellent quality up to and beyond a distance of
ChemCatChem 2015, 7, 3322 – 3331
www.chemcatchem.org
3326ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
The tabulated fitting parameters for hydrated and dehydrated
2–10Sn-b are shown in Tables 3 and 4, respectively.
Table 3. EXAFS fitting data obtained for hydrated SSI Sn-b catalysts.[a]
Catalyst
Abs Sc
N
R
[Š]
2s2
[Š2]
Ef
[eV]
R-factor
From Figure 6, it is evident that all the hydrated samples
show the presence of SnIV substituted within the zeolitic frame-
work alongside the copresence of SnxOy, which is present at
various levels in each material. Two SnÀO scatters are present
in the data. The shortest SnÀO scattering path around 1.9–
2.0 Š is assigned to SnÀO scattering interactions arising from
Sn substituted into the b-zeolite framework. The longer SnÀO
scattering path at a distance of 2.1 Š is associated either with
SnÀO scattering interactions from SnxOy, or water bound to the
Sn within the b-zeolite framework. Upon dehydration, there is
a clear decrease in intensity for the main peak, and a contrac-
tion in the first shell SnÀO distance from 2.0 to 1.95 Š (Fig-
ure S4). There is also a noticeable decrease in intensity for (or
complete elimination of) the second SnÀO distance, and a de-
crease in the mean square disorder parameter for the oxygen
shells upon dehydration. Each of these features indicates that
the longer SnÀO scattering path in hydrated SSI Sn-b catalysts
arises predominantly from water coordination to framework
SnIV sites. However, the need for a second SnÀO scattering in-
teraction even following dehydration (Figure 5) indicates the
copresence of some SnxOy clusters in all the samples. This is in
line with the observed SnÀSn scattering pattern, and likely
confirms that the SnÀSn scattering path arises predominately
from SnxOy oligomers, and less likely from SnÀSn pairing. How-
ever, more sophisticated methods are required to completely
eliminate the possibility of extensive SnÀSn pairing in these
materials.
2Sn-b
SnÀO
SnÀO
SnÀSn
SnÀSn
SnÀO
SnÀO
SnÀSn
SnÀSn
SnÀO
SnÀO
SnÀSn
SnÀSn
3.7 (3)
1.6 (3)
0.7 (3)
0.8 (4)
3.3 (4)
2.1 (4)
0.6 (3)
0.9 (4)
3.8 (8)
1.4 (9)
0.9 (5)
1.4 (7)
2.01 (1)
2.15 (2)
3.24 (3)
3.73 (2)
2.00 (1)
2.13 (2)
3.24 (3)
3.74 (2)
2.02 (1)
2.13 (4)
3.23 (4)
3.75 (2)
0.002 (2)
0.002 (1)
0.005 (2)
0.004 (2)
0.002 (1)
0.002 (2)
0.005 (2)
0.004 (2)
0.002 (2)
0.002 (1)
0.006 (4)
0.003 (1)
8 (1)
8 (1)
8 (1)
0.02
5Sn-b
0.02
0.02
10Sn-b
[a] Fitting parameters: S20 =1 as deduced by SnO2 standard; Fit range
2.5<k<13, 1<R<4; Number of independent points=19. Values in pa-
renthesis give the experimental error.
The comparable spectroscopic data obtained for our material
in reference to the benchmark material is in excellent agree-
ment with our observed catalytic data (Figures 2–4), which
demonstrate that SSI Sn-b is equal to conventional Sn-b in
terms of specific Sn site activity (TOF).[9] The comparable speci-
ation and activity of our material compared to the benchmark
Sn-b material prepared by hydrothermal synthesis highlights
the efficacy of our approach to preparing Sn-b, particularly
since a much higher Sn loading can be achieved.
The k3 weighted Fourier transform of hydrated 2, 5, and
10Sn-b, along with the simulated fit, is shown in Figure 6. The
c data for these samples, with simulated fit is shown in Fig-
ure S4. k3 weighted Fourier transform of dehydrated and hy-
drated 5 and 10Sn-b is in Figure S5, along with a comparison
in k3 weighted FT data for dehydrated and hydrated samples.
Further information on the degree of SnÀSn interaction for
each sample can be obtained from the SnÀSn scattering paths
at Æ3.2 and 3.7 Š, and the characteristic splitting of the oscilla-
À1
tion around 10 Š in the k3 weighted c data shown in Fig-
ure S3. The inset of Figure 6 shows the overlaid FT data be-
tween 2 and 4 Š for 2, 5 and 10Sn-b. It is evident that the
magnitude of the SnÀSn scattering path is dependent on the
degree of hydration, as much more intense SnÀSn interactions
are observed for the hydrated materials. Despite this, it is still
possible to compare the relative intensities between all the hy-
drated samples. It is clear that the magnitude of SnÀSn scatter-
ing in 2 and 5Sn-b is very similar, and this strongly suggests
that both catalysts contain a comparable SnxOy component.
This is in agreement with their observed activity, and particu-
larly their very similar TOF values both for glucose isomeriza-
tion and the MPV reaction. The comparable SnIV site speciation
is also indicated by the c data, which clearly shows that that
there is only a minor splitting of the 10 AÀ1 oscillation in these
two samples. This further confirms that SnxOy is only a minor
component of these materials. We note that the extent of
splitting, the magnitude of SnÀSn interactions, and the coordi-
nation numbers for SnÀSn interactions in both samples are
almost identical (Table 4). This demonstrates that the SnxOy
component in both catalysts is of a similar magnitude, and
that the SnxOy component in both materials is of comparable
nuclearity and composition.
Figure 6. Comparison of the magnitude of the Fourier transform (FT) of hy-
drated Sn-b catalyst containing different loadings. From bottom to top;
2Sn-b, 5Sn-b, and 10Sn-b. The fitted data is overlaid onto each spectrum;
2Sn-b (red), 5Sn-b (blue), and 10Sn-b (green). The 2–4 Š region of each
spectrum is overlaid and magnified in the inset.
As the loading is increased beyond 5 wt%, it is clear that
the SnÀSn interactions, and hence the quantity of extraframe-
ChemCatChem 2015, 7, 3322 – 3331
www.chemcatchem.org
3327ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Table 4. EXAFS fitting data obtained for dehydrated SSI Sn-b catalysts.[a]
Catalyst
Abs Sc
N
R
[Š]
2s2
[Š2]
Ef
[eV]
R-factor
2Sn-b
SnÀO
SnÀO
SnÀSn
SnÀSn
SnÀO
SnÀO
SnÀSn
SnÀSn
SnÀO
SnÀO
SnÀSn
SnÀSn
SnÀO
2.9 (2)
1.4 (2)
0.7 (3)
0.4 (3)
2.8 (2)
1.3 (2)
0.4 (2)
0.4 (3)
2.5 (2)
2.0 (2)
0.7 (2)
0.7 (2)
2.0 (2)
1.95 (1)
2.11 (2)
3.24 (3)
3.74 (2)
1.95 (1)
2.11 (2)
3.25 (5)
3.74 (4)
1.95 (1)
2.11 (2)
3.25 (5)
3.74 (4)
2.11 (2)
0.002 (1)
0.003 (2)
0.007 (2)
0.002 (1)
0.002 (1)
0.002 (2)
0.008 (5)
0.004 (4)
0.002 (2)
0.002 (1)
0.005 (2)
0.002 (1)
0.002 (1)
7 (1)
8 (1)
8 (1)
0.01
5Sn-b
0.01
0.02
10Sn-b
[a] Fitting parameters: S20 =1 as deduced by SnO2 standard; Fit range
2.5<k<13, 1<R<4; Number of independent points=19. Values in pa-
renthesis are the experimental error value.
Figure 7. 119Sn MAS NMR spectra of Sn-b samples containing 2 (red), 5
(blue), and 10 (green) wt% Sn. From bottom to top; bulk SnO2, 10% Sn(O2)-
b, 2Sn-b, 5Sn-b, and 10Sn-b.
work SnxOy species, increases, evidenced by increased SnÀSn
coordination numbers, an increase in the coordination number
of the longer SnÀO distance, and an increase in the splitting
structure shown in the c data. This indicates that a greater
amount of bulkier SnxOy clusters, potentially even SnO2, is
formed at elevated loadings, accounting for at least some of
the decrease in intrinsic activity observed for this sample. How-
ever, the crystallite size of this SnO2 must still be below Æ4–
5 nm, because of the absence of any SnO2 reflections in the
pXRD pattern and the position of the absorbance edge in the
UV/Vis spectra.
resonance at d=À602 ppm is also dependent upon the
degree of SnxOy oligomerization, with SnO2 producing a MAS
NMR profile with an intense resonance at d=À602 ppm, and
further, less intense resonances between d=À500 and
À700 ppm. Despite containing only natural quantities of 119Sn
and thus exhibiting low signal-to-noise ratios, several impor-
tant pieces of information can be gained from the MAS NMR
spectra of the 2, 5, and 10Sn-b samples (Figure 7). As can be
seen, both 2Sn-b and 5Sn-b give similar MAS NMR spectra,
suggesting that both materials possess a similar distribution of
Sn sites. The major resonance at d=À688 ppm demonstrates
that hydrated, framework SnIV species predominate in these
material, although the small resonances observed at d=
À602 ppm clearly demonstrates that some SnxOy domains are
present in these catalysts. The absence of additional resonan-
ces between d=À500 and À700 ppm indicates that the SnxOy
domains are of low nuclearity. Each of these observations is
fully consistent with, and strongly supports, our XAS analysis
(see below). As the loading is further increased to 10%, it is
evident that the extraframework Sn component increases in
concentration, as evidenced by a significant increase in intensi-
ty of the d=À602 ppm resonance. At this loading, extraframe-
work Sn apparently constitute a larger percentage of the Sn
site speciation, although the current analysis method does not
yet allow us to quantify the exact concentration of active and
spectator species in this material.
In the absence of full linear combination analysis, precise
quantification of the framework/extraframework Sn ratio is not
possible. Even so, it is evident that the increase in SnxOy forma-
tion at loadings above 5 wt% Sn is not substantial enough to
fully account for the significant decrease in observed activity,
that is, TOF. We reason that the discrepancy between the in-
creased level of SnxOy formation in 10Sn-b, and the observed
TOF decrease (50%) likely arises from a secondary effect; at
a loading of 10 wt%, the micropore volume of Sn-b decreases
by some 20%, and access to the micropores, and hence the
most active Sn atoms, may become somewhat restricted. This
would prohibit access to some of the active Sn component, re-
sulting in a greater-than-expected decrease in activity. Thus,
not only do SnxOy clusters lead to a decreased number of
active sites, but they also prohibit access to the existing active
sites in the zeolite.
To further elucidate the nature of Sn species in the SSI sam-
ples, 119Sn MAS NMR spectroscopy was used. 119Sn MAS NMR
spectroscopy is a sensitive technique for probing the Sn spe-
cies present in Sn-containing zeolites,[10,11] and by now it is well
accepted that several important pieces of information can be
gained from the MAS NMR spectral analysis. Hydrated, frame-
work SnIV sites give rise to a clear resonance at d=À688 ppm,
with this resonance shifting to d=À440 ppm upon in situ de-
hydration. Extraframework SnxOy clusters and SnO2 exhibit
a major resonance at d=À602 ppm. The magnitude of this
By now it is clear that the Sn site speciation in SSI Sn-b cata-
lysts remains relatively uniform up to a loading of 5 wt%, al-
though some minor SnxOy formation is observed at these
levels of Sn loading. At higher Sn loadings, increased SnxOy for-
mation is observed, and larger nuclearity clusters, potentially
even some SnO2, are also formed. To further determine wheth-
er this increased oligomer/oxide formation impacts the Lewis
acidity, and hence resulting catalytic activity, of Sn-b, we subse-
quently investigated the Lewis acidity of the catalysts with
in situ DRIFT spectroscopy. CD3CN is a useful molecule for
ChemCatChem 2015, 7, 3322 – 3331
www.chemcatchem.org
3328ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
creases in overall TOF observed at these Sn loadings (Figure 5).
This accounts for the lower activity of the higher loaded sam-
ples and further clarifies the roles of the active and spectator
sites. We note that the pronounced asymmetry in the n˜ =
2311 cmÀ1 band upon desorption, in particular for 2Sn-b, may
indicate that two possible bands, corresponding to two differ-
ent Sn active sites with different catalytic properties, may be
present in the samples. Previous FTIR studies have proposed
that these may arise from “closed” (n˜ =2308 cmÀ1) and “open”/
“hydrolyzed” (n˜ =2316 cmÀ1) Sn species that are still present in
the beta framework. Given that our procedure employs dealu-
minated beta zeolite, which evidently contains a high density
of networked silanol groups and remaining vacant framework
sites, it might be expected that a preference towards „open“
Sn species may be present in this material. This may account
for the apparent asymmetry of the n˜ =2311 cmÀ1 band. How-
ever, further spectroscopic study is required to confirm this
effect, particularly as recent work has suggested that the pack-
ing structure of CD3CN may inadvertently lead to the presence
of multiple bands in the n˜ =2311 cmÀ1 region.[26]
Figure 8. In situ CD3CN desorption profile for 2Sn-b (bottom) and 10Sn-
b (middle). CD3CN was desorbed under a dynamic vacuum at various tem-
perature intervals, increasing from top to bottom, and all spectra are back-
ground-referenced against the dehydrated zeolite sample. A reference spec-
trum of physisorbed CD3CN/2Sn-b is provided (dashed line).
Conclusions
probing the Lewis acid speciation in metal containing zeolites.
It has been demonstrated that isolated Lewis acid sites within
the zeolite framework interact with CD3CN, resulting in a peak
in the FTIR spectrum at n˜ = Æ2310 cmÀ1, which is not ob-
served for extraframework species.[29–33] The adsorption/desorp-
tion profile for 2Sn-b and 10Sn-b is presented in Figure 8. Fol-
lowing adsorption of CD3CN, two major adsorption features
are observed, at n˜ = Æ2270 and Æ2311 cmÀ1. The first feature
at n˜ =2275 cmÀ1 is related to physisorbed CD3CN, weakly
bound to the sample, as evidenced by its very rapid desorp-
tion upon heating. The presence of the second feature at n˜ =
2311 cmÀ1 in both samples arises from the CD3CN–Lewis acid
site interaction, and demonstrates that Lewis acidic SnIV cen-
ters are present in both materials. Upon heating (up to a maxi-
mum of 2008C), both physisorbed CD3CN and the CD3CN coor-
dinated onto the Lewis acidic centers, gradually desorb, al-
though the decrease in intensity for the chemisorbed species
is evidently much slower. The comparable rate of desorption in
both SSI samples strongly indicates that the active species in
both materials possess similar levels of Lewis acid strength.
Despite this, it is evident that the normalized area of the
CD3CN–Lewis feature, that is, the area of CD3CN–Lewis band
divided by wt% Sn, is considerably lower for 10Sn-b than for
2Sn-b. Despite the semiquantitative nature of this normaliza-
tion, the large difference in normalized area clearly demon-
strates that the SnxOy species identified in 10Sn-b by EXAFS
and MAS NMR spectroscopy are unable to activate the Lewis
base. Thus, they are spectator species to Lewis acid catalysis,
and hence decrease the number of Lewis acid sites in the cata-
lyst. Though our spectroscopic methods do not readily allow
for quantification of the amount of spectator sites at this time,
inactive sites likely account for a significant fraction of the Sn
population at loadings above 5 wt%, given the significant de-
Sn-b is a material of considerable industrial interest, as it has
been shown to be an important catalyst for several established
and emerging sustainable chemical transformations. In this
manuscript, we further demonstrate that solid-state incorpora-
tion (SSI) is a suitable alternative methodology for preparing
highly active Sn-b catalysts. Not only are the resulting catalysts
identical in activity to the conventional analogues, but they
are prepared very rapidly, very cleanly (without, for example,
solvent) and very selectivity.
In this manuscript, we have combined both catalytic and
spectroscopic studies to demonstrate that isolated, framework
SnIV sites dominate the Sn population of these materials, even
at loadings of up to 5 wt%, which is a loading approximately
two-to-three times greater than can be obtained by conven-
tional hydrothermal methods. Identical catalytic activity to the
established material is also observed, even at these significant-
ly elevated loadings. This consequently leads to significant in-
creases in the space–time yield of each catalyst, that is, the
productivity per gram of catalyst. At higher levels of Sn loading
(8–10 wt%), spectator sites in the form of Sn oligomers, and
likely some SnO2, are also formed. Though they do not partici-
pate in the main catalytic reactions or competitive side reac-
tions, these Sn oligomers decrease the effective concentration
of active Sn in the material, and result in a catalytic powder ex-
hibiting lower values of TOF. Spectroscopic analysis with a vari-
ety of sophisticated methods, including extended X-ray ab-
sorption fine structure, magic-angle spinning NMR, and diffuse
reflectance infrared Fourier transform spectroscopy, allowed to
us to obtain an overview of the active site distribution, and fur-
ther allowed us to rationalize the observed catalytic trends.
ChemCatChem 2015, 7, 3322 – 3331
www.chemcatchem.org
3329ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
X-ray absorption spectroscopy analysis
Experimental Section
Sn K-edge XAFS studies were performed on the B18 beamline at
the Diamond Light Source, Didcot, UK. Measurements were per-
formed using a QEXAFS setup with a fast-scanning Si (311) double
crystal monochromator. The time resolution of the spectra report-
ed herein was 2 min/spectrum (kmax =14, step size 0.5 eV), on aver-
age three scans were acquired to improve the signal-to-noise level
of the data for transmission measurements. All solid reference sam-
ples were diluted with cellulose and pressed into pellets to opti-
mize the effective edge-step of the XAFS data and measured in
transmission mode using ion chamber detectors. All Sn-substituted
zeolite samples were prepared as undiluted pellets, with the
amount of sample optimized to yield a suitable edge step and
measured in transmission mode using ion chamber detectors. All
XAFS spectra were acquired concurrently with the appropriate foil
placed between It and Iref. XAFS data processing was performed
using IFEFFIT[34] with the Horae package[35] (Athena and Artemis).
The amplitude reduction factor, S20, was derived from EXAFS data
analysis of a known reference compound, SnO2, (with known coor-
dination numbers which were fixed during analysis) to be 1.0,
which was used as a fixed input parameter.
Catalyst synthesis and pretreatment
Commercial zeolite Al-b (Zeolyst, NH4-form, SiO2/Al2O3 =25) was
dealuminated by treatment in HNO3 solution (13m HNO3, 1008C,
20 h, 20 mLgÀ1 zeolite). The dealuminated powder was washed ex-
tensively with water (Æ500 mLgÀ1 catalyst), and dried overnight at
110 8C. Solid-state incorporation was performed by a modified pro-
cedure of Refs. [<litr16] and [<litr16], by grinding the appropriate
amount of tin (II) acetate with the necessary amount of dealumi-
nated zeolite for 10 min in a pestle and mortar. Following this pro-
cedure, the sample was heated in a combustion furnace (Carbolite
MTF12/38/400) to 5508C (108C minÀ1 ramp rate) first in a flow of
N2 (3 h) and subsequently air (3 h) for a total of 6 h. Gas flow rates
of 60 mLminÀ1 were employed at all times. The sample was held
horizontally in an alumina combustion boat (10 mL capacity), and
a quartz tube was used to seal the sample environment and
permit gas flow.
Kinetic evaluation and analytical methods
Catalyst characterization
Batch MPVO reactions were performed in a 50 mL round bottom
flask equipped with a reflux condenser, which was thermostatically
controlled by immersion in a silicon oil bath. The vessel was
charged with a 10 mL solution of cyclohexanone in 2-butanol
(0.2m), which also contained an internal standard (biphenyl,
0.01m), and was subsequently heated to the desired temperature
(1008C internal temperature). The reaction was initiated by addi-
tion of an appropriate amount of catalyst, corresponding to
1 mol% Sn relative to cyclohexanone. The solution was stirred at
Æ600 rpm with an oval magnetic stirrer bar. Aliquots of reaction
solution were taken periodically for analysis, and were centrifuged
prior to injection into a GC (Agilent 7820, 25 m CP-Wax 52 CB). Re-
actants were quantified against a biphenyl internal standard.
Powder X-ray diffraction analysis was performed on a PANalytical
X’PertPRO X-ray diffractometer, with a CuKa radiation source
(40 kV and 40 mA). Diffraction patterns were recorded between 6–
558 2q at a step size of 0.01678 (time/step=150 s, total time=1 h).
DRIFT spectroscopy was performed in a Harrick praying mantis cell.
The spectra were recorded on a Bruker Tensor spectrometer over
a range of 4000–650 cmÀ1 at a resolution of 2 cmÀ1. In situ CD3CN
measurements were performed on pretreated zeolite powders
(5508C, 1 h under flowing air, 60 mLminÀ1) as follows: following
pretreatment, the sample was dosed with CD3CN vapor at room
temperature for 5 min, and one spectrum was recorded. The
sample chamber was subsequently evacuated under dynamic
vacuum (approximately 10À4 mbar), and spectra were recorded at
25, 50, 100, 150, and 2008C. All spectra were background subtract-
ed against the pretreated zeolite. UV/Vis analysis was performed
on an Agilent Cary 4000 UV/Visible spectrophotometer in diffuse
reflectance mode. Samples were scanned between 190 and
900 nm at a scan rate of 600 nmminÀ1. Sn contents were deter-
mined by energy dispersive X-ray spectrometry. Specific surface
area was determined from nitrogen adsorption using the BET equa-
tion, and microporous volume was determined from nitrogen ad-
sorption isotherms using the t-plot method. Porosymmetry meas-
urements were performed on a Quantachrome Autosorb, and sam-
ples were degassed prior to use (2758C, 3 h). Adsorption isotherms
were obtained at 77 K. MAS NMR experiments were performed at
Durham University through the EPSRC UK National Solid-State
NMR Service. Samples were measured under conditions identical
to those reported by Bermejo-Deval and co-workers.[11] Nonen-
riched Sn-b samples were measured on both a Varian VNMRS spec-
trometer, and a Bruker Avance III HD spectrometer with compara-
ble performances. Both spectrometers possess operating frequen-
cies of 400 and 149 MHz for 1H and 119Sn, respectively. Approxi-
mately 60–100 mg of sample was packed into a 4 mm rotor. Meas-
urements were performed in direct excitation mode (spin-echo
90x-t-180y), with a recycle delay of 2 s. Samples were spun at
Æ12000 Hz, and approximately 50000 repetitions were typically
employed for each sample.
d-glucose (Sigma–Aldrich, ꢀ99%) isomerization experiments were
performed in 15 mL thick-walled glass reactors (Ace pressure tube,
Sigma–Aldrich) that were heated in a temperature-controlled oil
bath. The reactor was charged with 5 mL of an aqueous solution
of glucose (10 wt%, 0.61m) and an appropriate amount of catalyst
corresponding to a 1:50 metal/glucose molar ratio. Once the oil
had reached the desired temperature (1108C), the reaction was ini-
tiated by vigorous stirring with a magnetic stirrer bar (600 rpm).
The reactor was stirred for an appropriate length of time, and time
online samples were obtained by periodically quenching the reac-
tion by rapidly cooling the reactor in an ice bath. Aliquots of solu-
tion were extracted with a syringe, centrifuged to remove solid
particulates, and were subsequently analyzed by HPLC (Agilent
1220). The compounds were separated with a Ca Hi-Plex column
(6.5”300 mm, 8 m particle size, Agilent), which was isothermally
held at 608C. Ultrapure water was used as the mobile phase, at
a flow rate of 0.3 mLminÀ1. The compounds were detected by use
of diode array and refractive index detectors.
Acknowledgements
C.H. is grateful to The Royal Society for research grant funding
(RG140754). C.H., N.D., P.P.W. and E.K.G. wish to acknowledge the
ChemCatChem 2015, 7, 3322 – 3331
www.chemcatchem.org
3330ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[14] M. Renz, T. Blasco, A. Corma, V. FornØs, R. Jensen, L. Nemeth, Chem. Eur.
J. 2002, 8, 4708–4717.
[15] A. Corma, M. E. Domine, L. Nemeth, S. Valencia, J. Am. Chem. Soc. 2002,
124, 3194–3195.
[16] P. Wolf, C. Hammond, I. Hermans, Dalton Trans. 2014, 43, 4514–4519.
[17] M. S. Holm, S. Saravanamurugan, E. Taarning, Science 2010, 328, 602–
605.
[18] W. R. Gunther, Y. Wang, Y. Ji, V. K. Michaelis, S. T. Hunt, R. G. Griffin, Y.
Romµn-Leshkov, Nat. Commun. 2012, 3, 1109.
[19] B. Tang, W. Dai, G. Wu, N. Guan, L. Li, M. Hunger, ACS Catal. 2014, 4,
2801–2810.
[20] C. Hammond, S. Conrad, I. Hermans, Angew. Chem. Int. Ed. 2012, 51,
11736–11739; Angew. Chem. 2012, 124, 11906–11909.
[21] W. N. P. van der Graaff, G. Li, B. Mezari, E. A. Pidko, E. J. M. Hensen,
ChemCatChem 2015, 7, 1152–1160.
support of the UK Catalysis Hub (EP/K014714/1). P.P.W. and E.K.G.
acknowledge the EPSRC for research funding (EP/I019693/1). The
Diamond Light Source and RCaH are thanked for the provision of
beamtime (SP8071) and the support of their staff. D.P. is grateful
to the University of Bologna (Italy) for a Toso Montanari Fellow-
ship. A.A.-N. is grateful to Al-Qadisiya University (Iraq) for the
provision of a PhD scholarship. C.H. and N.D. thank Prof. Stefania
Albonetti (University of Bologna) for continued support. C.H.
thanks Dr. David Apperley, Dr. Fraser Markwell and Dr. Eric
Hughes of the ESPRC UK National Solid-State NMR Service at
Durham University for performing the MAS NMR experiments,
and for useful discussions.
[22] J. Dijkmans, D. Gabriels, M. Dusselier, F. de Clippel, P. Vanelderen, K.
Houthoofd, A. Malfliet, Y. Pontikes, B. F. Sels, Green Chem. 2013, 15,
2777–2785.
[23] C. C. Chang, H. J. Cho, Z. Wang, X. Wang, W. Fan, Green Chem. 2015, 17,
2943–2951.
Keywords: biomass · heterogeneous catalysis · Lewis acids ·
tin · zeolites
[24] P. Y. Dapsens, C. Mondelli, J. Jagielski, R. Hauert, J. Perez-Ramirez, Catal.
Sci. Technol. 2014, 4, 2302–2311.
[25] C. Hammond, N. Dimitratos, J. A. Lopez-Sanchez, R. L. Jenkins, G. Whit-
ing, S. A. Kondrat, M. A. 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.
[26] R. S. Assary, L. A. Curtiss, J. Phys. Chem. A 2011, 115, 8754–8760.
[27] S. Bordiga, E. Groppo, G. Agostini, J. A. van Bokhoven, C. Lamberti,
Chem. Rev. 2013, 113, 1736–1850.
[28] S. R. Bare, S. D. Kelly, W. Sinkler, J. J. Low, F. S. Modica, S. Valencia, A.
Corma, L. T. Nemeth, J. Am. Chem. Soc. 2005, 127, 12924–12932.
[29] C. M. Osmundsen, M. S. Holm, S. Dahl, E. Taarning, Proc. R. Soc. London
Ser. A 2012, 468, 2000–2016.
[30] R. Roy, K. Bakhmutsky, E. Mahmoud, R. F. Lobo, R. J. Gorte, ACS Catal.
2013, 3, 573–580.
[31] M. Boronat, P. Concepción, A. Corma, M. T. Navarro, M. Renz, S. Valencia,
Phys. Chem. Chem. Phys. 2009, 11, 2876–2884.
[32] M. Boronat, P. Concepción, A. Corma, M. Renz, S. Valencia, J. Catal.
2005, 234, 111 –118.
[33] J. Chen, J. M. Thomas, G. Sankar, J. Chem. Soc. Faraday Trans. 1994, 90,
3455–3459.
[1] Y. Romµn-Leshkov, M. E. Davis, ACS Catal. 2011, 1, 1566–1580.
[2] M. Moliner, Dalton Trans. 2014, 43, 4197–4208.
[3] S. Bordiga, F. Bonino, A. Damin, C. Lamberti, Phys. Chem. Chem. Phys.
2007, 9, 4854–4878.
[4] A. Corma, L. T. Nemeth, M. Renz, S. Valencia, Nature 2001, 412, 423–
426.
[5] C. Hammond, R. L. Jenkins, N. Dimitratos, J. A. Lopez-Sanchez, M. H.
Ab Rahim, M. M. Forde, A. Thetford, D. M. Murphy, H. Hagen, E. E. Stang-
land, J. M. Moulijn, S. H. Taylor, D. J. Willock, G. J. Hutchings, Chem. Eur.
J. 2012, 18, 15735–15745.
[6] R. Gounder, Catal. Sci. Technol. 2014, 4, 2877–2886.
[7] M. Moliner, Y. Roman-Leshkov, M. E. Davis, Proc. Natl. Acad. Sci. USA
2010, 107, 6164–6168.
[8] R. Bermejo-Deval, R. S. Assary, E. Nikolla, M. Moliner, Y. Roman-Leshkov,
S.-J. Hwang, A. Palsdottir, D. Silverman, R. F. Lobo, L. A. Curtiss, M. E.
Davis, Proc. Natl. Acad. Sci. USA 2012, 109, 9727–9732.
[9] Y. Romµn-Leshkov, M. Moliner, J. A. Labinger, M. E. Davis, Angew. Chem.
Int. Ed. 2010, 49, 8954–8957; Angew. Chem. 2010, 122, 9138–9141.
[10] R. Bermejo-Deval, R. Gounder, M. E. Davis, ACS Catal. 2012, 2, 2705–
2713.
[11] R. Bermejo-Deval, M. Orazov, R. Gounder, S.-J. Hwang, M. E. Davis, ACS
Catal. 2014, 4, 2288–2297.
[12] G. Li, E. A. Pidko, E. J. M. Hensen, Catal. Sci. Technol. 2014, 4, 2241–
2250.
[34] M. Newville, J. Synchrotron Radiat. 2001, 8, 322–324.
[35] B. Ravel, M. Newville, J. Synchrotron Radiat. 2005, 12, 537–541.
[13] M. Boronat, A. Corma, M. Renz, G. Sastre, P. M. Viruela, Chem. Eur. J.
2005, 11, 6905–6915.
Received: May 16, 2015
Published online on September 1, 2015
ChemCatChem 2015, 7, 3322 – 3331
www.chemcatchem.org
3331ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim