Decarboxylation of Lactones over Zn/ZSM-5: Elucidation of the Structure of the Active Site and Molecular Interactions

Article abstract of DOI:10.1002/anie.201704347

Herein, we report the catalytic decarboxylation of γ-valerolactone (GVL) over Zn/ZSM-5 to butene, followed by aromatization at high yield with co-feeding of water. An evaluation of the catalytic performance after prolonged periods of time showed that a water molecule is essential to maintain the decarboxylation and aromatization activities and avoid rapid catalyst deactivation. Synchrotron X-ray powder diffraction and Rietveld refinement were then used to elucidate the structures of adsorbed GVL and immobilized Zn species in combination with EXAFS and NMR spectroscopy. A new route for the cooperative hydrolysis of GVL by framework Zn?OH and Br?nsted acidic sites to butene and then to aromatic compounds has thus been demonstrated. The structures and fundamental pathways for the nucleophilic attack of terminal Zn?OH sites are comparable to those of Zn-containing enzymes in biological systems.

Full text of DOI:10.1002/anie.201704347

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: A New Route for De-carboxylation of Lactones over Zn/ZSM-5:
Elucidation of Structure and Molecular Interactions
Authors: Edman Shik Chi Tsang, Lin Ye, Ben Lo, Qi Song, Junlin
Zheng, Dejing Kong, . Claire A Murray, and Chiu Tang
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.201704347
Angew. Chem. 10.1002/ange.201704347
Link to VoR: http://dx.doi.org/10.1002/anie.201704347
http://dx.doi.org/10.1002/ange.201704347

10.1002/anie.201704347
Angewandte Chemie International Edition
DOI: 10.1002/anie.200((will be filled in by the editorial staff))
Biomass Catalysis
A New Route for De-carboxylation of Lactones over Zn/ZSM-5:
Elucidation of Structure and Molecular Interactions
L. Ye[a]†, Q. Song[b]†, B.T.W. Lo[a], J. Zheng[b], D. Kong[b], C.A. Murray[c], C.C. Tang[c] and S.C.E. Tsang[a]*
Abstract: Here, we report the catalytic decarboxylation of gamma-
valerolactone, GVL (the biomass derived lactone as an example), over
Zn/ZSM-5 to butene, followed by aromatization at high yield with co-
feeding of water. The evaluation of catalytic performance at a prolonged
time shows that a water molecule is essential to maintain the decarboxylation
and aromatization activities without rapid catalyst deactivation. Synchrotron
X-ray powder diffraction and Rietveld refinement are then used to elucidate
the adsorbed structures of GVL and the immobilized Zn-species with the
support of EXAFS and NMR. A new route of cooperative hydrolysis of GVL
by the framework Zn-OH and Brønsted acid sites to butene and then to
aromatics is for the first time demonstrated. The structures and fundamental
pathways for the nucleophilic attack of terminal Zn-OH site are comparable
to the reported Zn-containing enzymes in biological systems.
conversion process, we note that the Zn incorporation into H-ZSM-5
gives much superior catalytic activity than the pure H-ZSM-5. One
would be interested to find out how the Zn/ZSM-5 structure provides for
such efficient conversion. It is interesting to note that nature also selects
the element Zn to catalyze acid-base hydrolysis reactions in cellular and
extra-cellular environments. This is because of its higher effective
nuclear charge (Zeff) among the transition metals, rapid water-exchange
and no complications from the redox properties. For example, carbonic
[7]
anhydrase (CA) enzymes containing Zn2+
are used to catalyze the
reversible hydration of CO2. Similarly, Zn2+ containing beta-lactamase
(BL) enzymes[8] can rapidly hydrolyze the lactam groups in penicillins.
With the specific Zn structure in these enzymes, their turnover
frequency can reach 106-107 s-1 (the diffusion rate limit of water),
making them the most ‘active’ and kinetically ‘perfect’ of all enzymes[9]
.
Currently, fuels and chemicals are mainly produced from non-renewable
oil, coal and gas[1]. However, their depleting supplies, geographically
uneven distributions and associated CO2 emissions are of significant
concerns. As a result, identifying an alternative sustainable carbon
source for fuel and chemical production is the key role for industry. In
fact, biomass has long been regarded as a potential alternative to
mitigate these problems. The present low oil prices, as opposed to the
high cost of processing biomass, does not yet offer immediate economic
incentives. As a result, vigorous research activities are currently
Due to the intrinsic shape and size selectivity as well as single-site
confinement, zeolite catalysts have long been regarded as ‘solid
enzymes’[10]. In addition, two or more catalytically active sites may
work in a cooperative manner, giving unusual properties[11]. The
introduction of Lewis metal ion sites into zeolites containing excess
Brønsted acid sites may create a synergistic effect in catalysis. In the
present case, GVL contains a reactive lactone group with a carbonyl that
could be susceptible to ring opening as the first step in hydrolysis by
these sites. In order to understand this new conversion, the information
on the structure of the active Zn sites with respect to the Brønsted acid
sites and their interactions with the GVL in the Zn/ZSM-5 is crucial.
Recently, the significant advancement of modern diffraction facilities
using high brightness synchrotron has provided chemists with a
powerful tool in studying porous but crystalline materials. For example,
high resolution synchrotron X-ray powder diffraction (SXRD) data was
used to determine the structure of various zeolites[12,13]. We have also
recently reported the use of SXRD combined with Rietveld refinement
to elucidate organic adsorbate structures in zeolites. The alteration in
scattering parameters of the modified framework atoms by the
molecules enables the probing of adsorption geometries and interactions
sought[2]
. Lactones, such as gamma-valerolactone (GVL) can be
favourably produced from lignocellulosic biomass[3]. However, only few
catalytic studies on the conversions of lactones to useful products have
been reported[4,5]. Using the Brønsted acid sites of SiO2/Al2O3, aqueous
GVL can undergo ring opening and decarboxylation to form butene.
After separation from water, butene gas can be processed into heavier
alkenes (for fuels) using amberlyst in a second catalyst bed[5]. Following
this work, the direct conversion of GVL to aromatics as platform
chemicals via butene without water has also been briefly studied over
the Brønsted acid sites in zeolites. However, the low liquid product yield
with rapid catalyst deactivation pose technical problems[6]. Furthermore,
it would be advantageous to carry out the catalysis without the need for
water separation (see SI).
with the Brønsted acid sites in terms of atomic distances and angles[14,15]
.
Thus, the environment of Zn and the adsorption geometry of GVL were
Here, we report that a high yield of aromatic compounds can be
produced from a GVL/steam mixture via butene at 450 oC by feeding an
aqueous solution of GVL directly over Zn-modified ZSM-5. The
process remains stable for at least 310 hours.
1
elucidated by SXRD in combination with EXAFS and H MAS NMR.
The obtained structures are compared to that of ‘perfect’ CA II enzyme
with a view of rationalizing this unique catalysis and providing hints for
further improvement.
While we realize the interesting potential of this new chemical
The Brønsted acidic sites in H-ZSM-5 may further catalyze butene to
aromatic compounds[6,16]. Thus, some pure and metal-modified ZSM-5
catalysts were tested using a flow reactor in SINOPEC (SRIFT) by
feeding a continuous liquid flow of 40 wt% GVL/H2O with weight
[a]
Dr Lin Ye[†]
, Benedict T.W. Lo, Prof S.C. Edman Tsang*
o
hourly space velocity (WHSV) of 1.0 h-1 to a catalyst zone at 450 C
Department of Chemistry, University of Oxford, Oxford, OX1 3QR (UK),
E-mail: edman.tsang@chem.ox.ac.uk;
(with a preheated zone) under industrial applicable conditions. As seen
in Figure 1a, the catalytic performances of some ZSM-5 based catalysts
are compared (SI Section 2). The GVL conversions are generally found
to be greater than 99% (complete conversion) within 12 hours of
reaction time for all the catalysts studied (Figure S1b). For the
unmodified H-ZSM-5 sample, the total aromatics (benzene, toluene,
xylene (BTX) and higher alkyl aromatics) selectivity is about 62% with
the light alkane/alkene selectivity reaching 38%. As stated, GVL can
[b]
[c]
[*]
Dr Qi Song[†], Dr Junlin Zheng and Prof Dejing Kong
Shanghai Research Institute of Petrochemical Technology SRIPT-
SINOPEC, Shanghai 201208 (China); [†] co-first authorship
Dr Claire A. Murray, Prof Chiu C. Tang
Diamond Light Source Ltd, Harwell Science and Innovation Campus,
Didcot, Oxfordshire, OX11 0DE (UK)
The authors wish to thank EPSRC, UK and Diamond Light Source Ltd and
SRIFT-SINOPEC for the financial support of this collaborative work and
are grateful to the Office of China Postdoctoral Council to grant
fellowship to LY to work at Oxford.
a
undergo ring opening and decarboxylation to form the initial butene[5,6]
but H-ZSM-5 alone appears to be insufficient to generate aromatics at
,
Supporting information is given via a link at the end of the document.
1
This article is protected by copyright. All rights reserved.

10.1002/anie.201704347
Angewandte Chemie International Edition
high selectivity. The addition of particularly Sn, Ga and Zn can
significantly increase the aromatics selectivity at the expense of light
alkane/alkene, presumably due to the introduction of other active sites
through these metal ions doping. Among the screened samples,
Zn/ZSM-5 gives the highest aromatics selectivity of 91.4% which is
about 30% higher than the unmodified H-ZSM-5. The carbon balance
was found to be 99 ± 1%, while coke and gaseous products accounted
for 2.4% and 7.2% of the total carbon detected, respectively. The major
gas phase product was found to be 1-butene (4.8%, see SI Section 2).
Figure 1b shows the long-term evaluation of the catalytic
performance of the Zn/ZSM-5 catalyst. There is a small but progressive
decrease of GVL conversion over 310 h due to ‘carbon’ deposition but
with a marginal change in total aromatics selectivity. It is interesting to
note that feeding the same amount of GVL in nitrogen without water
gives a much lower aromatic yield (ca. 50%). It also shows an extremely
fast deactivation, indicating the important role of water in this catalyzed
decarboxylation/aromatization reaction. Since this effect was not yet
reported in previous studies[6], further investigation is hereby conducted.
The evaluation of the Zn/ZSM-5 catalyst under repeated, accelerated
deactivation and re-activation steps (switching nitrogen carrier to air and
o
calcining the sample at 600 C for 1 h) shows that the activity of the
Zn/ZSM-5 is totally regained (Figure S2). The high catalytic activity,
good lifetime and recyclability of Zn/ZSM-5 make it a promising
candidate for the decarboxylation reaction of lactones (GVL) in water in
future bio-processing. Perhaps, there are a number of key questions that
remain to be answered: why does Zn2+ modification give the best
activity for the formation of aromatics via butene from GVL/H2O over
the tested catalysts? How does the active site interact with GVL? Why
does H2O appear to be essential for the catalysis? How do the structure
and catalytic mechanism of the decarboxylation reaction by this Zn-
containing zeolite differ from those of Zn enzymes in nature? It is noted
that structure of enzymatic catalyst and pathway can be elucidated
through the growth of corresponding single crystal enzyme structure and
substrate-enzyme complex. However, it would not be applicable to
industrial powder zeolite based catalysts even if they are regarded as
'solid enzymes'. On the other hand, the structure of Zn/ZSM-5 and the
substrate-active site interaction in the zeolite, if known, could provide
important hints to the catalytic activity. As a result, SXRD combined
with Rietveld refinement was used, for the first time, to monitor the
coordination environment of the Zn species in the zeolite. The Zn
environment reflects the nature of the interaction between Zn, the GVL
molecule and water during the first step in aromatization over the
Zn/ZSM-5 catalyst. All the data were refined by the Rietveld method
using TOPAS Academic refinement software (Bruker AXS, version 5),
see analytical details in SI Section 3.
Figure 1. (a) Product distribution over various ZSM-5 based catalysts (5 g) by feeding
40 wt% aqueous GVL solution to the preheated catalyst bed with WHSV 1.0 h-1 at 10
bar after 12 h (the time to collect the liquid products for analysis) at complete GVL
conversions. The mole % yield is based on mol. GVL to products (see SI Section 1). (b)
Long term catalytic performance of Zn/ZSM-5 with WHSV 1.0 h-1 at 450 oC.
First, by comparing the SXRD patterns of fresh unmodified H-ZSM-5
with and without Zn incorporation, the relative intensity of the peaks
changed between the two samples without peak broadening or formation
of other ZnO crystalline peaks (Figure S3). As mentioned in the method
section, more than 4,000 independent hkl reflections have been used for
the refinement calculations, which allow a greater number of structural
variables to be refined in satisfactory manner. The quality of the
Rietveld refinements of SXRD data was confirmed with low R-weighted
pattern (Rwp) values and goodness-of-fit factor (Tables S1-S5). The
structural details and Brønsted acid site (at the framework T6 site)
information of our H-ZSM-5 have been characterized in our previous
Figure 2. The refined structures of Zn/ZSM-5 catalyst derived from SXRD and
refinements. (a) Fresh (Figure S4) and regenerated Zn/ZSM-5 (Figure S6); (b) GVL
adsorbed on fresh Zn/ZSM-5 (Figure S5); (c) A close-up view of Zn(2)-OH interacting
with the carbonyl of GVL; (d) A published crystal structure showing a complex
intermediate of Zn-OH of T199A-CA II enzyme with adsorbed CO2 for comparison
(PDB: 1CAM) [Crystal Database from National Center for Biotechnology
Information][19]. In all the refined structures, the stick/ball model is used O = red, Si =
grey, Al = green, Zn = blue, C = black. In (c) and (d), only the activated complexes with
carbonyl adsorbates are colored. A mirror plane symmetry has been disregarded for
clarity, view down the x-axis (the sinusoidal channel).
work[14,15]
.
Figure 2a (also Figure S6) and the refinement result clearly show the
space group of the fresh and regenerated Zn/ZSM-5 from 310 h by air
treatment remains as Pnma as in a typical H-ZSM-5. Two isolated Zn
species with electron rich regions have been identified within the zeolite
cavity. One Zn2+ (named Zn-1) is found located at the intersection
region of the sinusoidal and straight channels.
By including a hexa-aqua Zn-1 species into the Rietveld refinement, a
good fit with the experimental data has been clearly achieved. This
entrapped Zn-1 with the 6 equivalent non-framework O as a typical
hexa-aqua Zn2+ gives an average Zn-O(H)2 distance of 2.12(3) Å (Table
S6 and SI). Its location in the porous cavity of Zn/ZSM-5 is likely
maintained by the surrounding negatively charged framework oxygens
through electrostatic interaction by ion-exchange with the Brønsted acid
2
This article is protected by copyright. All rights reserved.

10.1002/anie.201704347
Angewandte Chemie International Edition
sites. The site occupancy factor (SOF) of the Zn-1 is 0.104(2). Taking
the symmetry of the Pnma space group into account, there are 0.104 × 4
= 0.42 Zn-1 per unit cell of Zn/ZSM-5. Interestingly, a tetrahedrally
coordinated Zn (named Zn-2) attached to the 5-membered ring of ZSM-
5 through the wall-oxygens next to Al(T6) is also identified. The SOF of
Zn-2 is 0.071(3) corresponding to 0.071 × 8 = 0.57 Zn-2 per unit cell.
Thus, a total of 0.99 Zn per unit cell is obtained, which is in agreement
with the value derived from ICP-MS of 0.97 Zn per unit cell. The k3-
weighted Fourier transformed EXAFS data of the fresh Zn/ZSM-5 were
measured at the Zn K-edge, using ZnO as a standard (Table 1 and Figure
S7). A k1-k2-k3 fitting procedure in K- and R-space was carried out to
identify the presence of any Zn-O or Zn-Zn interactions in the first shell
and Zn-O in the second shell. The fitting result shows an averaged bond
length and coordination number (CN) of the Zn/ZSM-5. It is clear that
the average bond length of 2.07(1) Å is related to the backscattering
from the oxygen atoms to the directly bonded Zn. There is no evidence
for the possible Zn-Zn nearest bond distance of 2.3 Å or second shell
Zn-O bond distance of 3.0 Å of the ZnO lattice in the Zn/ZSM-5
EXAFS spectrum. This result undoubtedly indicates that the Zn species
in Zn/ZSM-5 are isolated without any ZnO aggregates or clusters, which
is consistent with the SXRD result. The average CN of 5.2(2) from the
EXAFS fitting matches the calculated average CN of 4.8(2) derived
from 57% four coordinated Zn (Zn-2) and 42% six coordination Zn (Zn-
1), according to the SOFs of the SXRD refinement. Also, the derived
Zn-O distance of 2.07(1) or 2.06(1) Å from EXAFS data is close to the
averaged bond distance of 2.11(2) Å of the two Zn species from the
SXRD data (Tables S2-S6).
approach cannot apply to those GVL-Zn-1 species if they are non-
crystallised and hence cannot rule out their existence. Despite the
polarisation by Zn2+ it is however envisaged that the non-dissociated
H2O molecule cannot act as a strong nucleophile as that of Zn2+-OH- for
the rapid GVL attack. Similarly, without the presence of Zn2+, the BAS
sites are also unable to polarize a water molecule to asymmetric OH- and
H+. The key question is how this Zn-OH site can be regenerated during
the catalysis with water co-feeding. It should be noted that the terminal
Zn-OH is located adjacent to the original Brønsted acid site in other
asymmetrical units (T6). Previous DFT calculations indicated one of the
coordinated water molecules on divalent metal cations such as Zn2+ can
dissociate when they are introduced into the framework of H-ZSM-5 via
ion-exchange[17,18]. This water dissociation is energetically favored,
resulting in the formation of a hydroxyl group bonded to the metal
cation and the regenerated proton residing on a nearby ionized Brønsted
acid site (see SI for further information). Thus, the cooperative
dissociative activation of water molecule that regenerates Zn-OH with a
deprotonated Brønsted acid site to complete the catalytic cycle is
plausible, based on our structural findings. Figure 2b depicts the crystal
structure of adsorbed GVL on Zn/ZSM-5 by SXRD and Rietveld
refinement. Despite there are the two Zn species (i.e. Zn-1 and Zn-2) in
our Zn/ZSM-5, the carbonyl of GVL molecule only aligns itself to the
terminal hydroxyl group of Zn-2. A closer inspection of the adsorption
geometry (Figure 2c) suggests the formation of a strong dative bonding
between the GVL and Zn-OH with a short O-Ccarbonyl distance of 1.82(4)
Å. In contrast, no interaction of GVL with the Zn-1 species in the
structure is found. We also note that the Zn-OH at the Zn-2 site in
Zn/ZSM-5 is structurally comparable to the active site of the CA II
enzyme (Figure S4d 1CA2). The active Zn2+ site is similarly
tetrahedrally coordinated, but with three histidine groups and a terminal
-O(H)[19]. During bio-catalysis, the -OH of the amino acid in site 199
near the Zn2+ is known to activate the incoming water molecules. Hence,
terminal Zn-O(H) and H+ through hydrogen bonding interaction are
formed. They have an analogous proton accepting role like our ionized
T6 site in a similar atomic spatial position in Zn/ZSM-5 as previously
discussed. The nucleophilic OH- can then attack CO2 to form a Zn-
hydrogen carbonate species. Thus, this enzymatic Zn2+ catalyzes the
According to the refined structure of the tetrahedral Zn-2 site
surrounded by four oxygens (Figure 2a), it is clear that this Zn2+ is
directly immobilized with the framework oxygens: O6, O18 and O9
sites following the ion-exchange and subsequent calcination of Zn/ZSM-
5. The derived bond distances of Zn-2-O6, Zn-2-O18 and Zn-2-O9
correspond to 1.92(2), 1.95(2) and 2.19(3) Å, respectively from the
SXRD and refinement. Although H atoms cannot be located by SXRD
due to their low electron density, the remaining non-framework oxygen
of 2.32(5) Å with Zn is identified in the form of Zn-2-O(H) as
1
demonstrated by our H-NMR and Raman spectroscopy, as shown later.
Thus, this hydrated Zn2+ after ion-exchange and calcination likely loses
some of its coordinated water. It is then immobilized onto the
framework oxygens which contain one of the acid sites. The formation
of the terminal Zn-OH initially looks puzzling, as Zn2+ itself cannot
activate water molecule at moderate temperature.
conversion of CO2 and H2O to HCO3 and H+ with extremely high
–
TOF[20]. Figure 2d shows the published crystal structure of the T199A
CAII enzyme-CO2 complex from the crystal database of the National
Center for Biotechnology Information[19]. Interestingly, the adsorption
geometry of GVL in the zeolite by the terminal Zn-OH as shown in
Figure 2c and the configuration of the Zn-containing enzyme-CO2
complex are alike. Scheme S2 (SI Section 8) demonstrates the proposed
mechanism for butene production by the cooperative ring opening and
hydrolytic decarboxylation of gamma-valerolactone (GVL) and water
activation over Zn/ZSM-5, followed by regenerating Zn-OH and
Brønsted acid site. As seen in the proposed mechanism (SI Sections S8-
10), a water molecule directly participates in the hydrolysis mechanism
for the decarboxylation of GVL, in a similar manner to that seen in the
‘perfect’ Zn-enzymes. This mechanism is consistent with our labelled
experiment showing the formation of C16/18O2 when H218O was used (SI
Section 5). Thus, water is not only required to reduce ‘carbon’
deposition, but also acts as a catalyst to produce CO2 and butene from
GVL over the Zn-OH and Brønsted acid site in a cooperative manner.
The butene, once formed, is known to undergo aromatization over H-
ZSM-5 based materials to give aromatics; this has been well studied and
Table 1. Comparison between the refined structural models from SXRD and best-fit
EXAFS data.
RSXRD
(Å)
REXAFS
(Å)
Debye-
Waller (Å)
Zn/ZSM-5
CNSXRD CNEXAFS
Enot (eV)
Catalyst
Fresh
Zn-O 4.8(2) 5.2(2)
2.11(2)
2.11(2)
2.07(1) 0.007(1)
2.06(1) 0.009(1)
3.4
4.6
Regenerated Zn-O 5.0(2) 5.0(1)
Statistical errors are displayed in brackets; CN stands for coordination number; R
represents derived atomic distance (see SI Section 4).
Nevertheless, the strong nucleophilic nature of terminal Zn-OH [5a] (i.e.
the new active site) if formed can attack the carbonyl group of GVL.
Through ring opening, CO2, butene and water are produced via Zn-OH
nucleophilic attack This offers an alternative to the simple acid
hydrolysis route by the Brønsted acid sites (framework or extra-
framework) that equally forms butene and CO2 as a result [5b]. We
observe higher rate and selectivity for GVL conversion to aromatics via
butene over Zn/ZSM-5 than H-ZSM-5 as shown in Figure 1. It clearly
suggests that Zn-OH is more active and selective for such conversion.
However, we cannot exclude the possibility of GVL that may also
interact with Zn-1 (presumably via one of its coordinated saturated H2O
ligand) and contribute to the overall catalytic activity. It is because our
will not be further discussed here[21]
.
As mentioned, the Zn/ZSM-5 catalyst is a typical microporous acidic
zeolite. It exhibits a slow but progressive deactivation during 310 h of
on-stream testing. However, the deactivated Zn/ZSM-5 catalyst can be
o
fully regenerated through air calcination at 600 C for 1 h. The same
concentration of Zn (0.99 Zn per unit cell) is obtained, by summing the
SOFs of Zn-1 and Zn-2 (Table S5). Similarly, the derived coordination
numbers and corresponding Zn-O distances from SXRD refinement and
3
This article is protected by copyright. All rights reserved.

10.1002/anie.201704347
Angewandte Chemie International Edition
EXAFS (Table 1 and Table S6) of the regenerated sample also clearly
indicate the structural integrity of the active site in Zn/ZSM-5 after the
prolonged testing (see SI Sections 6, 7 and Figure S9). The structure and
ability for nucleophilic attack of Zn/ZSM-5 and Zn-containing enzymes
resemble to each other. Thus, this suggests that the immobilized
tetravalent Zn2+ with the terminal -OH is the common active site.
Raman spectroscopy was also employed to study the adsorbed structure
of GVL on Zn/ZSM-5. Figure 3b shows a much larger and broader peak
with Raman frequencies at the 1100-2240 cm-1 region which
encompasses -C=O of GVL at ca. 1600 cm-1 and symmetrical O-C-O at
[24]
1360 cm-1
upon GVL adsorption on Zn/ZSM-5. In addition, the
A1(TO) mode of Zn-O at 374 cm-1
is clearly visible upon the
[25]
adsorption of GVL on the terminal Zn-OH in Zn/ZSM-5.
In conclusion, using SXRD combined with Rietveld refinement, we
demonstrate for the first time an anchored Zn-OH structure in Zn/ZSM-
5 can adsorb GVL via a strong Lewis acid-base interaction. The
regenerative Zn-OH and a neighbour Brønsted acid site in H-ZSM-5 by
dissociative adsorption of water molecule work in a cooperative manner.
They provide new catalytically active sites for a cascade of reactions to
form aromatics via butene. The structure and mechanism for the
nucleophilic attack of the active Zn-OH site appears are comparable to
that of the reported CA II enzyme. Hence, it can be regarded as an
efficient ‘solid enzyme’ for this reaction. Thus, the immobilized Zn2+ for
catalytic hydrolysis may provide inspirations to the chemical industry on
how to harness biomass to produce useful products.
Keywords: Zn/ZSM-5 • gamma-valerolactone • decarboxylation • synchrotron X-
ray powder diffraction • Rietveld refinement
[1] M. R. Rahimpour, M. Jafari, D. Iranshahi, Appl. Energ. 2013, 109, 79-93.
[2] W. R. H. Wright, R. Palkovits, Chemsuschem 2012, 5, 1657-1667.
[3] D. M. Alonso, S. G. Wettstein, J. A. Dumesic, Green Chem. 2013, 15, 584-595.
[4] I. T. Horvath, H. Mehdi, V. Fabos, L. Boda, L. T. Mika, Green Chem. 2008, 10,
238-242.
[5] (a) A. L. Allred, J. Inorg. Nucl. Chem. 1961, 17, 215-221.
(b) J. Q. Bond, D. M. Alonso, D. Wang, R. M. West, J. A. Dumesic, Science
2010, 327, 1110-1114.
Figure 3. (a). 1H MAS NMR of fresh Zn/ZSM-5 at 30 oC, after vacuum pre-treatment at
300 oC for 8 h, experimental data (black line), simulated data (red dot) and the
corresponding simulated partial contributions (blue, green, violet, dark yellow and light
blue line). (b) Raman spectra of various samples. They show clearly a characteristic
large peak and wide broadening of Raman absorption frequencies near 1100-2240 cm-1
which encompasses -C=O of GVL at ca. 1600 cm-1 and symmetrical O-C-O at 1360 cm-
1 upon the GVL adsorption on Zn/ZSM-5. In addition, the A1 (TO) mode of Zn-O at 374
cm-1 is visible on the terminal Zn-OH in Zn/ZSM-5.
[6] H. Xia, J. Zhang, X. Yan, S. Xu, L. Yang, J. Fuel. Chem. Technol. 2015, 43, 575-
580.
[7] in Carbonic Anhydrase: Mechanism, Regulation, Links to Disease, and Industrial
Applications, Vol. 75 (Eds.: S. C. Frost, R. McKenna, S. C. Frost, R. McKenna),
2014.
[8] J. Brem, S. S. van Berkel, W. Aik, A. M. Rydzik, M. B. Avison, I. Pettinati, K.-D.
Umland, A. Kawamura, J. Spencer, T. D. W. Claridge, M. A. McDonough, C. J.
Schofield, Nature Chem. 2014, 6, 1084-1090.
However, they should not be taken for a direct like-for-like comparison
in mechanism. The initial step of nucleophilic attack of GVL by Zn-OH
during hydrolysis requires a bond cleavage with a higher activation
barrier than that of the dehydration/hydration of the more labile CO2 in
the enzymatic reaction without bond breakage. In addition, the Zn2+ is
immobilized by the ‘O’ ligands (Si-O-Si) of the rigid zeolite framework.
It is hence electronically and structurally different from those of the
flexible prosthetic hydrophobic protein chain in the case of Zn-
containing enzymes. We understand that even if this heterogeneous
catalyst contains sites that are geometrically akin to enzymes, the
reaction temperature, conditions, interactions with the framework,
adsorption and desorption processes still make this system significantly
different. Nevertheless, the tetrahedral structure of Zn-2 with a terminal
nucleophilic Zn-OH for the hydrolysis and its regeneration by a
neighbouring H+ acceptor site, which activate a water molecule in a
cooperative manner, is clearly comparable in both cases. It is also noted
that the nucleophilicity of M-OH over our tested catalysts also follows
the order of Zn > Ga > Sn. This trend can be rationalised by the order of
electronegativity: Zn(1.65) < Ga(1.81) < Sn(1.96). Hence, the electron
density of the ‘O’ of the M-OH group follows the trend: Zn-OH > Ga-
OH > Sn-OH.[5a] Since the regenerative formation of nucleophilic Zn-
OH by GVL and water in this enzyme-mimic Zn/ZSM-5 is an important
key feature, it is crucial to experimentally verify the existence of the Zn-
[9] D. N. Silverman, S. Lindskog, Acc. Chem. Res. 1988, 21, 30-36.
[10] D. H. Olson, G. T. Kokotailo, S. L. Lawton, W. M. Meier, J. Phys. Chem. 1981,
85, 2238-2243.
[11] C. P. Grey, F. I. Poshni, A. F. Gualtieri, P. Norby, J. C. Hanson, D. R. Corbin, J.
Am. Chem. Soc. 1997, 119, 1981-1989.
[12] R. E. Morris, S. J. Weigel, N. J. Henson, L. M. Bull, M. T. Janicke, B. F.
Chmelka, A. K. Cheetham, J. Am. Chem. Soc. 1994, 116, 11849-11855.
[13] P. Norby, J. Am. Chem. Soc. 1997, 119, 5215-5221.
[14] B. T. W. Lo, L. Ye, J. Qu, J. Sun, J. Zheng, D. Kong, C. A. Murray, C. C. Tang,
S. C. E. Tsang, Angew. Chem. Int. Ed. 2016, 55, 1-5.
[15] L. Ye, B. T. W. Lo, J. Qu, I. Wilkinson, T. Hughes, C. A. Murray, C. C. Tang, S.
C. E. Tsang, Chem. Commun. 2016, 52, 3422-3425.
[16] H. Kitagawa, Y. Sendoda, Y. Ono, J. Catal. 1986, 101, 12-18.
[17] G. Yang, Y. Wang, D. H. Zhou, X. C. Liu, X. W. Han, X. H. Bao, J. Mol. Catal.
A: Chem. 2005, 237, 36-44.
[18] A. N. Subbotin, G. M. Zhidomirov, I. R. Subbotina, V. B. Kazansky, Kinet. Catal.
2013, 54, 744-748.
[19] Y. F. Xue, A. Liljas, B. H. Jonsson, S. Lindskog, Proteins: Struct., Funct., Genet.
1993, 17, 93-106.
[20] Z. Fisher, J. A. H. Prada, C. Tu, D. Duda, C. Yoshioka, H. Q. An, L.
Govindasamy, D. N. Silverman, R. McKenna, Biochemistry 2005, 44, 1097-1105.
[21] Y. Ono, H. Kitagawa, Y. Sendoda, J. Chem. Soc., Faraday Trans. 1987, 83, 2913-
2923.
[22] L. Heeribout, C. Doremieux-Morin, J. P. Nogier, R. Vincent, J. Fraissard,
Microporous Mesoporous Mater. 1998, 24, 101-112.
1
OH active site. Figure 3a shows five resolved H MAS NMR signals of
[23] J. F. Wu, S. M. Yu, W. D. Wang, Y. X. Fan, S. Bai, C. W. Zhang, Q. Gao, J.
Huang, W. Wang, J. Am. Chem. Soc. 2013, 135, 13567-13573.
[24] E. Garand, T. Wende, D. J. Goebbert, R. Bergmann, G. Meijer, D. M. Neumark,
K. R. Asmis, J. Am. Chem. Soc. 2010, 132, 849-856.
the fresh Zn/ZSM-5 at 1.21, 1.88, 2.66, 4.15 and 5.61 ppm assigned to
Zn-OH, Si-OH, Al-OH, H+ of Brønsted acid site and the H of adsorbed
H2O on Brønsted acid site, respectively. The latter four signals have
been previously detected in various aluminosilicate-based zeolites[22] but
the former peak is reflected by the characteristic Zn-OH in Zn/ZSM-5[23]
[25] Y. Dong, C. Feng, P. Jiang, G. Wang, K. Li, H. Miao, Rsc Adv. 2014, 4, 7340-
7346.
.
4
This article is protected by copyright. All rights reserved.

10.1002/anie.201704347
Angewandte Chemie International Edition
Entry for the Table of Contents (Please choose one layout)
Layout 1:
COMMUNICATION
A new cooperative ring-opening
route for decarboxylation of
Lin Ye, Qi Song, Benedict T.W. Lo,
Junlin Zheng, Dejing Kong, Claire A.
Murray, Chiu C. Tang, Shik Chi Edman
Tsang*
gamma-valerolactone
(biomass
derivative lactone) with water
molecule activation over Zn/ZSM-5
catalyst by the regenerative Zn-OH
and Brønsted acid site to CO2 and
1-butene. The structures and
Page No. – Page No.
A New Route for De-carboxylation of
Lactones over Zn/ZSM-5: Elucidation
fundamental
pathways
are
of
Structure
and
Molecular
comparable to that of reported Zn-
containing enzymes in biological
systems.
Interactions
5
This article is protected by copyright. All rights reserved.