Quantitative production of butenes from biomass-derived γ-valerolactone catalysed by hetero-atomic MFI zeolite

Article abstract of DOI:10.1038/s41563-019-0562-6

The efficient production of light olefins from renewable biomass is a vital and challenging target to achieve future sustainable chemical processes. Here we report a hetero-atomic MFI-type zeolite (NbAlS-1), over which aqueous solutions of γ-valerolactone (GVL), obtained from biomass-derived carbohydrates, can be quantitatively converted into butenes with a yield of >99% at ambient pressure under continuous flow conditions. NbAlS-1 incorporates simultaneously niobium(v) and aluminium(iii) centres into the framework and thus has a desirable distribution of Lewis and Br?nsted acid sites with optimal strength. Synchrotron X-ray diffraction and absorption spectroscopy show that there is cooperativity between Nb(v) and the Br?nsted acid sites on the confined adsorption of GVL, whereas the catalytic mechanism for the conversion of the confined GVL into butenes is revealed by in situ inelastic neutron scattering, coupled with modelling. This study offers a prospect for the sustainable production of butene as a platform chemical for the manufacture of renewable materials.

Full text of DOI:10.1038/s41563-019-0562-6

Articles
https://doi.org/10.1038/s41563-019-0562-6
Quantitative production of butenes from
biomass-derived γ-valerolactone catalysed
by hetero-atomic MFI zeolite
Longfei Linꢀ ꢀ1, Alena M. Sheveleva1,2, Ivan da Silvaꢀ ꢀ3, Christopher M. A. Parlettꢀ ꢀ4,5,6, Zhimou Tang7,
Yueming Liu7, Mengtian Fan1, Xue Han1, Joseph H. Carter1, Floriana Tuna1, Eric J. L. McInnesꢀ ꢀ1,
Yongqiang Cheng8, Luke L. Daemen8, Svemir Rudić3, Anibal J. Ramirez-Cuestaꢀ ꢀ8, Chiu C. Tang6
and Sihai Yangꢀ ꢀ1*
The efficient production of light olefins from renewable biomass is a vital and challenging target to achieve future sustain-
able chemical processes. Here we report a hetero-atomic MFI-type zeolite (NbAlS-1), over which aqueous solutions of
γ-valerolactone (GVL), obtained from biomass-derived carbohydrates, can be quantitatively converted into butenes with a
yield of >99% at ambient pressure under continuous flow conditions. NbAlS-1 incorporates simultaneously niobium(v) and
aluminium(iii) centres into the framework and thus has a desirable distribution of Lewis and Brønsted acid sites with optimal
strength. Synchrotron X-ray diffraction and absorption spectroscopy show that there is cooperativity between Nb(v) and the
Brønsted acid sites on the confined adsorption of GVL, whereas the catalytic mechanism for the conversion of the confined
GVL into butenes is revealed by insitu inelastic neutron scattering, coupled with modelling. This study offers a prospect for the
sustainable production of butene as a platform chemical for the manufacture of renewable materials.
pproximately 400 million tonnes of light olefins (ethene, to the acid sites and forming acid–base adducts25,26. Thus, the
propene and butenes) are produced each year for polymer, design of an efficient, yet water-tolerant, catalyst for this conversion
A
chemical and pharmaceutical industries1–6. State-of-the-art remains a fundamental challenge. Recently, emerging niobium-
processes are based on the stream cracking of naphtha and thus based catalysts were shown to possess a marked water tolerance25–32
critically rely on fossil fuels, which possess significant environmen- and exceptional activities for the hydrodeoxygenation of biomass
tal impacts1. The catalytic production of light olefins from renew- under mild conditions33–35
.
able plant biomass is a highly promising target7, and many processes
Here we describe the development of a hetero-atomic MFI-
have been developed to bridge future gaps in the supply of com- type zeolite (denoted as NbAlS-1) and its catalytic activity for the
modity chemicals from biomass8–11. In particular, biomass-derived conversion of GVL into butenes. NbAlS-1 integrates both Nb(v)
γ-valerolactone (GVL), obtained from agricultural waste via low- and Al(iii) sites into the framework, and through fine-tuning of
cost, high-yield commercial processes (production scale at 27 the ratio of Nb:Al:Si, the optimal acidity and hence reactivity of
tonnes in 2018), was identified as a sustainable resource to produce the catalyst can be achieved to promote a quantitative production
butenes12, which, through well-established petroleum processes, of butenes (yield >99%) using a 30wt% GVL feedstock at 320°C
can be readily transformed into a wide spectrum of petrochemicals, and atmospheric pressure in the absence of an external source of
such as propene2,13, 1,3-butadiene14, aromatics15, liquid fuels2,9, poly- hydrogen. NbAlS-1 shows a superior water tolerance and catalytic
ethylene2,16 and polybutene2,17. In this respect, the development of stability over 180hours in a continuous-flow reaction. The under-
efficient catalytic processes to convert GVL into butenes is of vital lying catalytic mechanism was elucidated at a molecular level,
importance. This conversion involves the ring opening and decar- employing a combination of in situ synchrotron X-ray powder
boxylation reactions catalysed over solid acids, and amorphous diffraction (SXPD), X-ray absorption spectroscopy (XAS), inelas-
SiO2/Al2O3, ZSM-5, La/ZSM-5, Ni2P/MCM-41, Zn-AlPO-5 and tic neutron scattering (INS) and density functional theory (DFT)
Pd/Nb2O5 have been studied9,15,18–24. Using a 30wt% GVL feedstock calculations. These complimentary techniques revealed that the
at 375°C and atmospheric pressure, a butene yield of 75% has been cooperativity between Nb(v) and Brønsted acid sites plays a key role
achieved over the SiO2/Al2O3 catalyst9. Commercial supplies of GVL in the reactivity of GVL and, compared with conventional HZSM-5,
from biorefineries are aqueous solutions with GVL concentrations the absence of strong Brønsted acid sites in NbAlS-1 effectively hin-
between20and40wt%(ref. 15);however, thewaterinthereactioncan dered the formation of by-products, which results in the observed
partially or completely deactivate solid acid catalysts by coordinating selectivity for butenes.
1Department of Chemistry and Photon Science Institute, University of Manchester, Manchester, UK. 2International Tomography Centre SB RAS and
Novosibirsk State University, Novosibirsk, Russia. 3ISIS Facility, STFC, Rutherford Appleton Laboratory, Chilton, UK. 4School of Chemical Engineering and
Analytical Science, University of Manchester, Manchester, UK. 5University of Manchester at Harwell, Diamond Light Source, Harwell Campus, Didcot,
UK. 6Diamond Light Source, Harwell Science and Innovation Campus, Didcot, UK. 7Shanghai Key Laboratory of Green Chemistry and Chemical Processes,
School of Chemistry and Molecular Engineering, East China Normal University, Shanghai, China. 8The Chemical and Engineering Materials Division
(CEMD), Neutron Sciences Directorate, Oak Ridge National Laboratory, Oak Ridge, TN, USA. *e-mail: Sihai.Yang@manchester.ac.uk
NATuRE MATERIALS | www.nature.com/naturematerials

Nature Materials
Articles
Synthesis and characterization
Catalytic tests
The hydrothermal synthesis of ZSM-5 doped with varying loadings Conversions of pure GVL were firstly performed over a fixed-bed
of Nb(v) (0–5.58wt%) and Al(iii) (0–1.64wt%) was systematically packed with the catalysts under flow conditions at 320°C (weight
studied, and all the obtained materials exhibit the MFI-type frame- hourly space velocity=0.18h−1; Table 1). HZSM-5 with varying Al/Si
work (Supplementary Fig. 1), BET surface areas of 357–408m2 g−1 ratios shows GVL conversions of 75–92% and moderate butene
(Supplementary Table 1) and particle size distributions between yields of 33.1–44.1% (Table 1, entries 1–3). This is due to the forma-
300 and 500nm (Supplementary Fig. 2 and Supplementary Table 1). tion of considerable amounts of 2-cyclopentenone, aromatics and
All the zeolites studied here exhibit similar micropore volumes polycyclic aromatics as by-products and coke over strong acid sites
(Supplementary Table 1). An optimal loading of 3.81wt% Nb(v) (Supplementary Figs. 5 and 6). However, the lack of Brønsted acid
and 1.64wt% Al(iii) was identified for the synthesis of NbAlS- sites in NbS-1 gives very low butene yields (~7%) and GVL con-
1(0.027/0.04/1) (yield=95%). The atomic ratio of Nb/Al/Si was versions (32%) (Table 1, entries 4 and 5). In contrast, a 96% GVL
determined by energy-dispersive X-ray (EDX) analysis to be conversion with an 80.4% butene yield was achieved over NbAlS-1
0.027/0.04/1 with a homogeneous distributions of metal ions (0.027/0.04/1) (Table 1, entry 6). This suggests that the strength and
(Fig. 1a–e). The vibrational modes of the Nb–O–Si moieties36 in nature of the framework acidity are key parameters for optimizing
NbAlS-1(0.027/0.04/1) were observed by Fourier transform infra- its performance. By introducing Nb(v) as weak Lewis acid sites to
red (FTIR) spectroscopy (971cm−1; Fig. 1f) and Raman spectros- weaken the strong Al(iii) acid sites in NbAlS-1, the formation of by-
copy (921 and 973cm−1; Fig. 1g), which indicates the incorporation products was effectively inhibited, which leads to high selectivities
of Nb(v) centres as the framework T sites.
for butenes over NbAlS-1.
The successful incorporation of Nb(v) and Al(iii) into frame-
The catalytic performance of these zeolites was further tested using
work sites was further confirmed by electron paramagnetic reso- 30 and 60wt% aqueous solutions of GVL (Table 1 and Supplementary
nance (EPR) spectroscopy. Although NbS-1 (the isostructural Table 4); the former is of direct relevance to practical processes15. The
analogue with 0% Al loading) and NbAlS-1 are EPR silent due to butene yields from aqueous solutions of GVL over HZSM-5 with vary-
the diamagnetic Nb(v) and Al(iii) ions, characteristic paramag- ing Si/Al ratios decreased to 22.2–29.2% owing to the deactivation of
netic centres can be induced by γ-irradiation of the zeolites36. EPR acid sites in the presence of water (Table 1, entries 1–3, 11 and 13–15).
spectra of γ-irradiated NbS-1(0.027/1) and NbAlS-1(0.027/0.04/1) Significantly, a quantitative production of butenes (yield>99%) was
show a wide, structured signal that spreads over a magnetic field achieved over NbAlS-1(0.027/0.04/1) using a 30wt% GVL feed-
range of ~350mT (Fig. 1h) with a ten-line hyperfine structure, con- stock (Table 1, entry 16), which demonstrates the significantly
sistent with the coupling of an unpaired electron to a 93Nb nucleus improved water tolerance by Nb(v) doping. Other widely studied
(I=9/2, 100% natural abundance). The site has axial symmetry with Nb-containing catalysts, such as Nb2O5, niobic acid (Nb2O5·H2O) and
g values of g|| =1.866 and g⊥ =1.935, and hyperfine couplings of NbOPO4, were also tested for this reaction and, as expected, these
ANb|| =0.0289cm−1 and ANb⊥ =0.0148cm−1 (Supplementary Table 2), catalysts all exhibited excellent water tolerance (Table 1, entries 8–10
consistent with the previous characterization of Nb(iv) centres and 17–19). A butene yield of 78.3% was obtained from the con-
generated at the pseudo-tetrahedral NbO4 units in the framework version of 30wt% aqueous solutions of GVL over NbOPO4, which,
of silicalite-136, whereas this signal is not observed in γ-irradiated however, has a much lower utilization efficiency of Nb compared
HZSM-5 or Nb2O5-HZSM-5 mixtures (Fig. 1h), indicative that the with NbAlS-1(0.027/0.04/1) (46% and 3.8%, respectively, in terms
Nb(v) ion in NbAlS-1 is solely incorporated into framework sites, of Nb concentration). The performance of NbAlS-1(0.027/0.04/1)
rather than deposited as Nb2O5 phases. A second narrow signal compares favourably with state-of-the-art catalysts (Supplementary
with a hyperfine structure observed in the spectra of γ-irradiated Table 5). As previously observed in niobic acid systems, water can
HZSM-5 and NbAlS-1 (Fig. 1i and Supplementary Table 2) arises only partially coordinate to the Nb(v) centres, which leaves sufficient
from an electron hole defect on a framework oxygen atom in an positive charges on NbO4 moieties to bind substrates26,31. Indeed,
Al–O•–Si site37, where the structure arises from hyperfine cou- the formation of coke, 2-cyclopentenone and other by-products is
pling with the 27Al nucleus (I=5/2, 100% natural abundance). In effectively inhibited over NbAlS-1(0.027/0.04/1) by the presence of
γ-irradiated NbS-1, a sharp signal is observed, but the hyperfine water (Table 1 and Supplementary Fig. 7). The excellent stability of
structure is lost, consistent with the absence of Al(iii). Thus, EPR NbAlS-1(0.027/0.04/1) was demonstrated by both cycling and time-
measurements provide convincing evidence for the incorpora- on-stream tests, in which the yield of butenes remained at >99% after
tion of both Nb(v) and Al(iii) into framework sites in NbAlS-1, 180hours of reaction (Fig. 2a,b). The used catalyst was characterized
which enables the fine-tuning of its framework acidity and hence by N2 sorption, synchrotron XPD, EDX analysis, FTIR spectroscopy,
catalytic activity.
EPR spectroscopy (on 60Co γ-irradiation), XAS, 27Al NMR spectros-
Pyridine-adsorption FTIR (Py-IR) spectroscopy has been used copy, ammonia temperature-programmed desorption and Py-IR
to characterize the nature of acidity in these zeolites (Supplementary spectroscopy (Fig. 2c,d and Supplementary Figs. 8–12). Importantly,
Fig. 3 and Supplementary Table 3). NbS-1 exhibits solely Lewis all these results confirm that there is little change to the structure and
acidity due to the presence of Nb(v), whereas NbAlS-1 shows both acidic sites of used NbAlS-1, which demonstrates the excellent struc-
Lewis and Brønsted acidity, which arises from the dual functional- tural stability of the bifunctional zeolite. Interestingly, little change was
ity of Nb(v) and Al(iii) sites. Notably, the nature of the acidity in observed in the structure and acidic sites of the used HZSM-5(0.04/1)
NbAlS-1 is distinct to that of HZSM-5: the ratio of Lewis/Brønsted (Fig. 2c,d and Supplementary Figs. 11–13), which indicates that the
acid sites in NbAlS-1 is 4.69, 12 times that of HZSM-5. The acid- decrease of catalytic activity of HZSM-5 on exposure to water com-
ity was quantified by ammonia temperature-programmed desorp- pared with reactions using pure GVL (Table 1, entries 2, 11 and 14)
tion (Supplementary Fig. 4 and Supplementary Table 3). HZSM-5 is primarily due to partial poisoning of the acidic sites, consistent with
displays both strong and weak acid sites with a total amount of the literature25,26
0.4–0.63mmolg−1 and NbS-1 exhibits minor weak acid sites of
.
0.04mmolg−1. Interestingly, NbAlS-1 shows only weak acid sites, Studies of the coordination environment of Nb(v) sites
but their concentration is considerably higher (0.22mmolg−1) than To understand the effect of active Nb(v) sites, their local environ-
that of NbS-1. Thus, doping of pentavalent niobium sites into the ment was interrogated through Nb K-edge XAS. The absorption
zeolite can effectively tune the nature and distribution of framework edge positions for both NbAlS-1(0.027/0.04/1) and NbS-1(0.027/1)
acidity and provides an excellent platform to examine the resultant from X-ray absorption near-edge spectra are consistent with the
catalytic activities.
pentavalent oxidation state38 with as-prepared samples that exhibit a
NATuRE MATERIALS | www.nature.com/naturematerials

Nature Materials
Articles
a
c
b
d
NbAlS-1(0.027/0.04/1)
NbS-1(0.027/1)
HZSM-5(0.04/1)
Si
8 µm
8 µm
O
e
Al
Nb
8 µm
8 µm
0.4
0.8
1.2
1.6
2.0
2.4
2.8
3.2
Energy (keV)
g
f
NbAlS-1(0.027/0.04/1)
NbS-1(0.027/1)
NbAlS-1(0.027/0.04/1)
NbS-1(0.027/1)
HZSM-5(0.04/1)
HZSM-5(0.04/1)
921
973
971
1,400 1,300 1,200 1,100 1,000 900 800 700 600
Wavenumber (cm–1
1,400
1,200
1,000
Raman shift (cm–1
)
800
600
400
200
)
h
i
NbS-1(0.027/1)
NbS-1(0.027/1)
NbAlS-1(0.027/0.04/1)
HZSM-5(0.04/1)
NbAlS-1(0.027/0.04/1)
HZSM-5(0.04/1)
Nb2O5/HZSM-5(0.04/1)
Nb2O5/HZSM-5(0.04/1)
100 150 200 250 300 350 400 450 500 550
Magnetic field (mT)
300
310
320
330
340
350
360
Magnetic field (mT)
Fig. 1 | Physical characterization of catalysts. a, EDX spectra of HZSM-5(0.04/1), NbS-1(0.027/1) and NbAlS-1(0.027/0.04/1). The peaks in the regions
of 0.52, 1.49, 1.74 and 2.17ꢀkeV are related to the binding energies of O Kɑ, Al Kɑ, Si Kɑ and Nb Lɑ, respectively. b, Scanning electron microscopy image of
NbAlS-1(0.027/0.04/1). c–e, EDX maps of Si Kɑ (c), Al Kɑ (d) and Nb Lɑ (e) in NbAlS-1(0.027/0.04/1). f,g, FTIR (f) and Raman (g) spectra of HZSM-
5(0.04/1), NbS-1(0.027/1) and NbAlS-1(0.027/0.04/1) samples. h,i, X-band (~9ꢀGHz) EPR spectra of activated zeolites by γ-irradiation at 77ꢀK: the wide
magnetic field sweep (h) highlights the induced Nb(iv) signals in NbS-1(0.027/1) and NbAlS-1(0.027/0.04/1) and the narrow magnetic field-sweep (i)
highlights the induced electron hole defect Al–O•–Si signals in NbAlS-1(0.027/0.04/1) and HZSM-5(0.04/1). The weak features at ~304 and 356ꢀmT in
i are due to radiation-induced trapped hydrogen atoms. The experimental spectra are shown in black, red, blue and magenta. The simulated spectra of
Nb(iv) (h) and Al–O•–Si (i) defects are shown in grey. a.u., arbitrary units.
common weak pre-edge peak (Fig. 3a). The weak nature of this pre- in the average Nb=O bond length due to the presence of an adsor-
edgefeaturecanbeattributedtodisruptionofthepseudo-tetrahedral bate and a subsequent decrease in the double-bond nature, whereas
Nb environment via water coordination (Fig. 3b)38. On adsorption the siloxane bridges remain unaffected (Supplementary Table 6).
of GVL to the pores of NbAlS-1, the pre-edge position was further The effect was apparent in both the NbAlS-1 and NbS-1 systems,
influenced with an apparent blueshift (~0.8eV); in contrast, no shift and the NbAlS-1 system was perturbed to a significantly greater
was witnessed for NbS-1. This reflects both a greater electron dona- extent, 0.22 versus 0.06Å increase in bond length, respectively.
tion from GVL39 and the preferred adsorption geometry within the Furthermore, fitting of the second coordination shell discredited
bifunctional zeolite NbAlS-1, which facilitates a closer interaction the potential for bulk Nb2O5 formation, which confirmed the loca-
between the Nb(v) sites and the substrate. This conclusion was tion of Nb(v) to be solely within the framework, consistent with the
verified through extended X-ray absorption fine structure analysis EPR observation. Thus, the framework Nb(v) site plays a key role in
(Fig. 3c). Fitting of the first coordination shell revealed an increase the adsorption of GVL in NbAlS-1.
NATuRE MATERIALS | www.nature.com/naturematerials

Nature Materials
Articles
Table 1 | Summary of the GVL conversion and product yields over different catalystsa
Entry
Catalyst
GVL concentration
(wt%)
GVL conversion
(%)
Yield (mol %)
Butene
33.1
2-cyclopentenone
Aromatics
Othersb
21.7
21.2
21.6
23.7
24.4
9.2
1
HZSM-5(0.027/1)
HZSM-5(0.04/1)
HZSM-5(0.067/1)
NbS-1(0.027/1)
NbS-1(0.04/1)
NbAlS-1(0.027/0.04/1)
NbAlS-1(0.04/0.027/1)
Nb2O5
99
99
99
99
99
99
99
99
99
99
60
60
30
30
30
30
30
30
30
75
83
92
31
9.9
11.5
13.4
0.4
0.4
6.2
4
10.3
12.4
12.9
0
2
37.9
44.1
6.9
3
4
5
32
96
84
8
7.2
0
6
80.4
56.3
6.9
0.2
0.2
0
7
23.5
1.1
8
0
9
Nb2O5·nH2O
36
75
81
15.8
47.2
29.2
92.3
22.2
23.2
23.3
99.3
10.1
0.9
5.3
9.3
0.3
6.3
4.7
4.6
0
0
19.3
22.0
33.3
7.4
10
11
12
13
14
15
16
17
18
19
NbOPO4
0.5
9.2
0
HZSM-5(0.04/1)
NbAlS-1(0.027/0.04/1)
HZSM-5(0.027/1)
HZSM-5(0.04/1)
HZSM-5(0.067/1)
NbAlS-1(0.027/0.04/1)
Nb2O5
100
73
76
85
100
11
6.7
7.1
7.4
0
37.8
41.0
49.7
0.7
0
0
0.9
Nb2O5·nH2O
38
20.0
2.3
0
15.7
NbOPO4
88
78.3
0
0
9.7
aReaction conditions: catalyst, 2.0ꢀg; reaction temperature, 320ꢀ°C; atmospheric pressure; weight hourly space velocity, 0.18ꢀh−1; time-on-stream, 10ꢀh. bOther products include C2–C4 hydrocarbons (except
butene), pentenoic acid isomers, 1-(2-methylcyclopropyl)ethanone, 3,3-diethoxy-1-propyne, cyclobutanecarboxylic acid-2-propenyl ester, 3-methyl-2-butenoic acid, cyclobutyl ester (the latter four compounds
are mainly for entries 4 and 5), 4-ethyl-4-methyl-2-cyclohexene-1-one (detected for entries 9, 10, 18 and 19), 3-ethyl-1-isopropyl-1H-indene (detected for entries 10 and 19) and polycyclic aromatics.
Determination of adsorption domains for GVL
SiO(H)Al sites in LTA zeolite was studied using periodic DFT cal-
The binding domains of GVL within these zeolites were determined culations and INS40,41. Here, 29Si NMR, INS and DFT calculations
from in situ SXPD data (Supplementary Fig. 14 and Supplementary demonstrate the presence of NbO(SiO)nAl (n=1 or 2) species in
Notes (Interaction between GVL and Nb/Al/H sites)). All datasets NbAlS-1, which promotes the preferred adsorption of GVL in
enabled the direct observation of two independent binding sites for NbAlS-1 (Supplementary Figs. 16–18, Supplementary Tables 13–15
GVL (I and II) within the pores (weighted profile factor of 4–6%; and Supplementary Notes (Distribution of Nb/Al/H sites)). Thus,
Fig. 4, Supplementary Fig. 15 and Supplementary Tables 7–12) the excellent catalytic activity of NbAlS-1 for butene production
without a notable structural phase change of the framework originates from (1) the highly confined adsorption of GVL in the
(Supplementary Fig. 14). GVLI and GVLII are located near the inter- framework pores and (2) the greatly reduced Brønsted acidity of the
section of straight and sinusoidal channels, respectively (Fig. 4a–c). resultant framework (~70% compared to HZSM-5), which lead to
Importantly, the adsorbed GVL molecules showed three distinct the rapid desorption of butene products from the catalyst surface,
spatial orientations within the framework of HZSM-5, NbS-1 and and thus prevent their further reaction to form various cyclic by-
NbAlS-1, whichindicatesthepresenceofvaryinghost–guestinterac- products and cokes.
tions. InthestraightchannelofHZSM-5, GVLI interactswithabridg-
ing O(H) centre through its C=O group (C=O∙∙∙O=3.812(54)Å) Studies of reaction mechanism
via a weak hydrogen bond (Fig. 4d). In NbS-1, GVLI directly binds Combined INS and DFT calculations were applied to investigate
to framework T sites (NbO4 moieties) via its intra-ring O cen- the vibrational dynamics of GVL@NbAlS-1(0.027/0.04/1) in ope-
tre (T∙∙∙OC4 =4.005(39)Å) (Fig. 4e). Interestingly, the adsorbed rando. Recently, this technique was employed to investigate the
GVLI in NbAlS-1 forms host–guest interactions with both bridg- vibrational dynamics of 1-octene@ZSM-5 (ref. 42). The INS spec-
ing O(H) centres and T sites (C=O∙∙∙O=3.574(141)Å and tra of solid and adsorbed GVL were recorded and modelled via
Nb∙∙∙OC4 =3.789(75)Å) (Fig. 4f). Similar, but more notable, obser- DFT calculations (Fig. 5a,b, Supplementary Figs. 19–23,
vations were observed for GVLII (Fig. 4g–i). For example, GVLII Supplementary Table 16 and Supplementary Notes (Inelastic
forms a weak hydrogen-bond interaction to a bridging O(H) centre neutron scattering)), which allowed a full assignment of the spec-
(C=O∙∙∙O=3.210(38)Å) in HZSM-5 and a dipole interaction to tral features for GVL. Comparison of the INS spectra of adsorbed
the framework T site (Nb∙∙∙OC4 =3.560(40)Å) in NbS-1, whereas GVL (spectrum of empty cell has been subtracted) and solid
in NbAlS-1, it displays a combination of these two types of interac- GVL shows a number of observations (Fig. 5a). There is little change
tions (C=O∙∙∙O=3.055(57)Å and Nb∙∙∙OC4 =3.490(75)Å]. These in the low-energy modes (<100cm−1) of GVL on adsorption, which
results strongly indicate that Brønsted or Lewis acid sites alone in indicates the adsorbed GVL molecules are highly ordered and con-
HZSM-5 or NbS-1 can only serve as ‘monodentate’ binding sites fined within the framework. The peaks at 1,060cm−1 (assigned to
for GVL molecules, whereas the cooperativity between these two C4–O1 stretching) and at 1,127cm−1 (assigned to C1–O1–C4 asym-
types of acid sites results in the optimal adsorption of GVL via a metric stretching) decrease significantly in intensity and shift to
‘chelating’ mechanism in NbAlS-1. Recently, the distribution of 1,046cm−1 and 1,121cm−1, respectively, which indicates that the
NATuRE MATERIALS | www.nature.com/naturematerials

Nature Materials
Articles
a
b
100
100
90
80
70
60
50
40
30
20
10
0
90
80
70
60
50
40
30
20
10
0
NbAlS-1(0.027/0.04/1)
HZSM-5(0.04/1)
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15
0
20 40 60 80 100 120 140 160 180
Time on stream (h)
Cycle
c
d
NbAlS-1(0.027/0.04/1)
NbAlS-1(0.027/0.04/1) used
HZSM-5(0.04/1)
NbAlS-1(0.027/0.4/1)
NbAlS-1(0.027/0.4/1) used
HZSM-5(0.04/1)
HZSM-5(0.04/1) used
HZSM-5(0.04/1) used
4
6
8
10
12
14
16
18
20
1,420 1,440 1,460 1,480 1,500 1,520 1,540 1,560 1,580
Wavenumber (cm–1
2θ (°)
)
Fig. 2 | Catalyst stability. a, Comparison of butene yield over 15 cycles of reactions over NbAlS-1(0.027/0.04/1) at 320ꢀ°C using bulk GVL (99%) as the
feed. After each cycle, the catalyst is calcined at 550ꢀ°C under an air flow. b, Lifetime study of NbAlS-1(0.027/0.04/1) and HZSM-5(0.04/1) at 320ꢀ°C
using 30ꢀwt% GVL in aqueous solution as the feed. The yield of butene remained at >99% after 180ꢀh of continuous reaction. c,d, Comparison of SXPD
patterns (λꢀ=ꢀ0.82487(1)ꢀÅ) (c) and acidities determined by Py-IR spectroscopy (d) of HZSM-5(0.04/1) and NbAlS-1(0.027/0.04/1) before and after the
conversion of 30ꢀwt% GVL at 320ꢀ°C for 180ꢀh.
b
c
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
a
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0
Nb
NbO2
2.0
1.6
1.2
0.8
0.4
0
NbAlS-1+ GVL
0.8 eV
Nb2O5
18,970 18,980
Energy (eV)
NbS-1
NbAlS-1
NbAlS-1 (dry)
NbAlS-1
NbAlS-1
2.5
NbAlS-1+ GVL
18,970 18,990 19,010 19,030 19,050
18,970 18,975 18,980 18,985 18,990
1
1.5
2
3
Energy (eV)
Energy (eV)
Radial distance (Å)
Fig. 3 | Nb K-edge XAS for Nb-containing zeolites. a), X-ray absorption near-edge spectra highlighting both the edge position and pre-edge feature of
the Nb-containing samples. b, Shifts of pre-edge feature in NbAlS-1, with inset showing the first derivate in the presence of water and GVL substrates.
c, Phase-corrected, k3-weighted Fourier transform of the extended X-ray absorption fine structure spectra and fitting curves (dotted lines) for bare and
GVL-saturated NbAlS-1. deriv, derivative.
C4–O1 and C1–O1 bonds are activated on adsorption via O1 to confirms the restricted motion of the methyl group on adsorp-
the Nb(v) site. Meanwhile, the peaks at 151cm−1 (assigned to C4 tion (–CH3∙∙∙Ozeolite =2.032(65)Å). The rocking mode of C1=O2
out-of-plane wagging) and at 827cm−1 (assigned to ring distor- (512cm−1) disappears completely, consistent with the formation
tion) also decrease in intensity owing to the hindered motion of of hydrogen bonds to the Brønsted acid site. The peak at 181cm−1
O1 on adsorption. The modes of the C4–C5 torsion (273cm−1) (assigned to the twisting mode of O2=C1–C2) shifts to 173cm−1,
and rocking (317cm−1) are also affected by adsorption via the O1 which further confirms the adsorption of GVL via the O2 centre on
centre, and shift to lower energies at 234 and 308cm−1, respec- the Brønsted acid sites. These findings are highly consistent with the
tively. The marked redshift of Δ=39cm−1 for the torsion mode X-ray crystal structures.
NATuRE MATERIALS | www.nature.com/naturematerials

Nature Materials
Articles
a
b
e
h
c
f
i
GVLI
GVLII
d
4.005 Å
3.789 Å
GVLI
T8
T8
O25
O25
3.812 Å
3.574 Å
g
3.560 Å
T6
3.490 Å
T6
GVLII
O26
O26
3.055 Å
3.210 Å
Fig. 4 | Views of crystal structures of GVL-loaded HZSM-5(0.04/1), NbS-1(0.027/1) and NbAlS-1(0.027/0.04/1). All the models were obtained from
Rietveld refinements based on in situ SXPD data. Two distinct binding sites for GVL were observed in the straight channel (I) and the sinusoidal channel (II).
a–c, HZSM-5(0.04/1)·8GVL (Al3.692Si92.308O192·8C5H8O2) (a), NbS-1(0.027/1)·8GVL (Nb2.524Si93.476O192·8C5H8O2) (b) and NbAlS-1(0.027/0.04/1)·8GVL (Al3.599
Nb2.429Si89.972O192·8C5H8O2) (c). d–i, Detailed views of the host–guest binding of GVLI (d–f) and GVLII (g–i) in HZSM-5(0.04/1), NbS-1(0.027/1) and
NbAlS-1(0.027/0.04/1), respectively. The GVL molecules and the functional sites involved in the cooperative binding are highlighted by the use of an
amplified ball-and-stick model (Nb/Al/Si, green; C, grey; O, red; H, white). The O1···Nb interactions, O2···O(H) interactions are highlighted in cyan and
blue, respectively. Owing to the uncertainty on locations of protons, all hydrogen bonds in this report are described as the distance between the OGVL and
the Ozeolite centres.
Adsorbed GVL molecules on NbAlS-1(0.027/0.04/1) underwent adsorption of GVL occurs on Nb(v) and Brønsted acid sites in a
a first catalytic conversion at 280°C for three minutes (Fig. 5a). ‘chelating mode’ via O1 and O2 centres, followed by a ring-open-
The peaks at 1,046, 1,121 and 827cm−1 (all involve C4–O1 vibra- ing reaction of activated GVL via selective cleavage of the C4–O1
tional modes) further reduce in intensity, which indicates that the bond at elevated temperatures. Meanwhile, C1=O2 is protonated
C4–O1 bond is selectively activated. To promote further conver- via proton transfer from the Brønsted acid sites. On the addition
sion, a second reaction was carried out at 300°C for three min- of a H shift in the C5 skeleton, a β-scission occurs and C1–C2 bond
utes (Fig. 5a). All the INS features related to C4–O1 modes (1,046 is cleaved to yield butenes and an equimolar CO2. By contrast, on
and 1,121cm−1) and ring deformation (609, 655, 802, 827 and HZSM-5, both the ring-opening and decarboxylation reactions
899cm−1) disappeared, which suggests that the ring-opening reac- proceed on the Brønsted acid sites (Fig. 5d). The inability for the
tion occurs via selective cleavage of the C4–O1 bond in activated intermediate carbocation to desorb from the strong acid sites of
GVL. Meanwhile, O2 is protonated, which leads to the decrease of HZSM-5 results in its transformation into undesirable by-products
peak intensities at 530 and 173cm−1 (assigned to C1 and C2 out-of- of 2-cyclopentenone, aromatics and coke via dehydration, aromati-
plane wagging modes, respectively). A final reaction was conducted zation and coking, respectively.
at 320°C, and these two peaks (530 and 173cm−1) disappeared, con-
sistent with the decarboxylation reaction via cleavage of the C1–C2 Outlook
bond (Fig. 5a). Meanwhile, new features appeared at 212, 229, Powerful drivers exist for the development of efficient catalysts to
254, 366, 407 and 437cm−1, fully consistent with the formation of enable the sustainable production of light olefins from biomass to
butenes (Supplementary Fig. 24).
remit our reliance on fossil fuels. NbAlS-1, which integrates both
A full catalytic circle was constructed based on the structural, Nb(v) and Al(iii) sites into the MFI zeolite framework, demonstrates
dynamic and modelling results (Fig. 5c). Over NbAlS-1, the an excellent catalytic performance to transform biomass-derived
NATuRE MATERIALS | www.nature.com/naturematerials

Nature Materials
Articles
b
a
Neutron energy loss (meV)
0
20 40 60 80 100 120 140 160 180 200
GVL
Adsorbed GVL
O1
O1
O1
C5
O1
O2
O2
C5
C5
C5
O2
O2
O2
C5
O1
C1
C2
C4
C1
C1
C1
C2
C1
C4
C4
C4
C4
C3
Reaction 280 °C
Reaction 300 °C
Reaction 320 °C
C3
C2 C3
C2 C3
C2 C3
1,060 cm–1
(1,046 cm–1
1,127 cm–1
(1,121 cm–1
151 cm–1
(144 cm–1
273 cm–1
827 cm–1
)
)
)
(234 cm–1
)
O1
O1
O2
C5
O2
C5
O2
O2
C5
O2
C5
O1
C1
C5
C1
O1
C1
O1
C1
C1
C4
C4
C4
C3
C4
C4
C2
C2 C3
C3
C2
C3
C2
C2 C3
317 cm–1
(308 cm–1
512 cm–1
(173 cm–1
)
530 cm–1
181 cm–1
)
0
200 400 600 800 1,000 1,200 1,400 1,600
Neutron energy loss (cm–1
)
c
d
Al
O
O
Al
H
O
O
O
H
O
O
O
Si
Si
HO
O
O
O
Al
Nb
Si
O
O
O
O
O
O
Al
O
O
O
O
O
O
O
O
O
H
O
O
O
Si
Si
O
O
O
Nb
Si
O
O
O
Al
O
O
O
O
O
O
H
Al
H
H
Aromatics
O
H
O
H
O
O
O
Si
Si
O
O
O
Nb
O
Si
O
O
O
O
O
O
O
O
Al
OH
OH
Al
O
O
H
O
H
O
O
C
O
H
O
C
O
O
O
O
Al
O
Nb
Si
Al
O
Si
Si
O
O
H
O
Al
O
O
O
H
O
O
O
O
O
O
O
O
Si
Si
O
Si
O
Si
O
O
O
O
O
Nb
O
Si
O
O
O
O
O
O
O
Fig. 5 | INS spectra for NbAlS-1(0.027/0.04/1) on the adsorption and catalytic conversion of GVL and the proposed reaction mechanisms. All the spectra
shown here are after the subtraction of the INS spectrum of the empty cell. No abscissa scale factor was used for INS calculations throughout this article.
Where no error bars are visible, the error are smaller than the symbols used to represent the data points. a, Comparison of INS spectra for condensed solid
GVL, adsorbed GVL and reacted GVL on NbAlS-1(0.027/0.04/1). b, Selected vibrational modes of GVL. The wavenumbers of adsorbed GVL as observed by
experiments are shown in parentheses. For clarification, the hydrogen atoms (white) are only displayed in the first GVL model. c,d, The proposed reaction
mechanisms for the conversion of GVL over NbAlS-1 based on INS/SXPD experiments (c) and HZSM-5 based on literature reports (d)13,18,21
.
Received: 21 March 2019; Accepted: 11 November 2019;
Published: xx xx xxxx
GVL in aqueous solutions into butenes with a quantitative yield.
Nb(v) sites in the MFI framework play three vital roles: optimize
the acidity of zeolite to hinder the production of by-products,
cooperate with Brønsted acid sites for the confined adsorption
of GVL and thus activate C–O bonds and resist water poison-
ing and thus enhance the overall catalytic stability. On reaction
completion, the liquid-phase product contains only water and
the gas-phase product is an equimolar mixture of butenes and
CO2, which can be facilely separated based on the difference in
their boiling points (Δ=72 °C). While butenes are used as a plat-
form chemical, the recovered CO2 can be readily sequestrated. The
industrial-scale synthesis of NbAlS-1 could be achieved based on
the infrastructure widely used for the synthesis of ZSM-5, which,
coupled with its high stability, further demonstrates its potential for
practical applications.
References
1. Torres Galvis, H. M. & de Jong, K. P. Catalysts for production of lower olefns
from synthesis gas: a review. ACS Catal. 3, 2130–2149 (2013).
2. Bender, M. An overview of industrial processes for the production of
olefns—C4 hydrocarbons. ChemBioEng Rev. 1, 136–147 (2014).
3. Galvis, H. M. T. et al. Supported iron nanoparticles as catalysts for sustainable
production of lower olefns. Science 335, 835–838 (2012).
4. Amghizar, I., Vandewalle, L. A., Van Geem, K. M. & Marin, G. B. New trends
in olefn production. Engineering 3, 171–178 (2017).
5. Jiao, F. et al. Selective conversion of syngas to light olefns. Science 351,
1065–1068 (2016).
6. Zacharopoulou, V. & Lemonidou, A. A. Olefns from biomass intermediates:
a review. Catalysts 8, 2 (2018).
7. Tuck, C. O., Pérez, E., Horváth, I. T., Sheldon, R. A. & Poliakof, M.
Valorization of biomass: deriving more value from waste. Science 337,
695–699 (2012).
Online content
8. Bozell, J. J. Connecting biomass and petroleum processing with a chemical
bridge. Science 329, 522–523 (2010).
Any methods, additional references, Nature Research reporting
summaries, source data, extended data, supplementary informa-
tion, acknowledgements, peer review information; details of author
contributions and competing interests; and statements of data and
code availability are available at https://doi.org/10.1038/s41563-
019-0562-6.
9. Bond, J. Q., Alonso, D. M., Wang, D., West, R. M. & Dumesic, J. A. Integrated
catalytic conversion of γ-valerolactone to liquid alkenes for transportation
fuels. Science 327, 1110–1114 (2010).
10. Jing, Y., Guo, Y., Xia, Q., Liu, X. & Wang, Y. Catalytic production of
value-added chemicals and liquid fuels from lignocellulosic biomass. Chem 5,
2520–2546 (2019).
NATuRE MATERIALS | www.nature.com/naturematerials

Nature Materials
Articles
11. Corma, A., Iborra, S. & Velty, A. Chemical routes for the transformation of
biomass into chemicals. Chem. Rev. 107, 2411–2502 (2007).
12. Yan, K., Yang, Y., Chai, J. & Lu, Y. Catalytic reactions of gamma-
valerolactone: a platform to fuels and value-added chemicals. Appl. Catal. B
179, 292–304 (2015).
28. Zhang, Y. et al. Mesoporous niobium phosphate: an excellent solid acid for
the dehydration of fructose to 5-hydroxymethylfurfural in water. Catal. Sci.
Technol. 2, 2485–2491 (2012).
29. Takagaki, A., Tagusagawa, C. & Domen, K. Glucose production from
saccharides using layered transition metal oxide and exfoliated nanosheets as
a water-tolerant solid acid catalyst. Chem. Commun. 14, 5363–5365 (2008).
30. Carniti, P., Gervasini, A., Biella, S. & Auroux, A. Niobic acid and niobium
phosphate as highly acidic viable catalysts in aqueous medium: fructose
dehydration reaction. Catal. Today 118, 373–378 (2006).
31. Carniti, P., Gervasini, A., Bossola, F. & Dal Santo, V. Cooperative action of
Brønsted and Lewis acid sites of niobium phosphate catalysts for cellobiose
conversion in water. Appl. Catal. B 193, 93–102 (2016).
32. Nowak, I. & Ziolek, M. Niobium compounds: preparation, characterization,
and application in heterogeneous catalysis. Chem. Rev. 99, 3603–3624 (1999).
33. Serrano-Ruiz, J. C., Wang, D. & Dumesic, J. A. Catalytic upgrading of
levulinic acid to 5-nonanone. Green. Chem. 12, 574–577 (2010).
34. Xia, Q. et al. Direct hydrodeoxygenation of raw woody biomass into liquid
alkanes. Nat. Commun. 7, 11162 (2016).
35. Shao, Y. et al. Selective production of arenes via direct lignin upgrading over
a niobium-based catalyst. Nat. Commun. 8, 16104 (2017).
36. Prakash, A. M. & Kevan, L. Synthesis of niobium silicate molecular sieves of
the MFI structure: evidence for framework incorporation of the niobium ion.
J. Am. Chem. Soc. 120, 13148–13155 (1998).
37. Wichterlová, B., Nováková, J. & Prášil, Z. Structure of defects in γ-irradiated
ZSM-5 and Y zeolites: an ESR study. Zeolites 8, 117–121 (1988).
38. Corma, A., Llabrés i Xamena, F. X., Prestipino, C., Renz, M. & Valencia, S.
Water resistant, catalytically active Nb and Ta isolated Lewis acid sites,
homogeneously distributed by direct synthesis in a beta zeolite. J. Phys. Chem.
C 113, 11306–11315 (2009).
39. Duereh, A., Sato, Y., Smith, R. L. & Inomata, H. Analysis of the cybotactic
region of two renewable lactone–water mixed-solvent systems that exhibit
synergistic Kamlet–Taf basicity. J. Phys. Chem. B 120, 4467–4481 (2016).
40. Lemishko, T., Valencia, S., Rey, F., Jiménez-Ruiz, M. & Sastre, G. Inelastic
neutron scattering study on the location of Brønsted acid sites in high silica
LTA zeolite. J. Phys. Chem. C 120, 24904–24909 (2016).
41. Lemishko, T. et al. Inelastic neutron scattering study of the aluminum and
Brønsted site location in aluminosilicate LTA zeolites. J. Phys. Chem. C 122,
11450–11454 (2018).
42. Hawkins, A. P. et al. Investigation of the dynamics of 1-octene adsorption at
293K in a ZSM-5 catalyst by inelastic and quasielastic neutron scattering.
J. Phys. Chem. C 123, 417–425 (2019).
13. Lin, L. et al. Acid strength controlled reaction pathways for the catalytic
cracking of 1-butene to propene over ZSM-5. J. Catal. 309, 136–145 (2014).
14. Hong, E., Park, J.-H. & Shin, C.-H. Oxidative dehydrogenation of n-butenes
to 1,3-butadiene over bismuth molybdate and ferrite catalysts: a review.
Catal. Surv. Asia 20, 23–33 (2016).
15. Ye, L. et al. Decarboxylation of lactones over Zn/ZSM-5: elucidation of the
structure of the active site and molecular interactions. Angew. Chem. Int. Ed.
56, 10711–10716 (2017).
16. Shao, H., Lv, Z., Sun, Z., Liu, C. & He, A. Synthesis of spherical polyethylene/
poly(1-butene) reactor blends with two-stage sequence polymerization
technology. Polymer 144, 72–79 (2018).
17. Kim, M. S., Park, M. S., Seo, H. J. & Lee, S. H. Method for preparing
polybutene. US Patent 9683060B2 (2017).
18. Kellicutt, A. B., Salary, R., Abdelrahman, O. A. & Bond, J. Q. An examination
of the intrinsic activity and stability of various solid acids during the catalytic
decarboxylation of γ-valerolactone. Catal. Sci. Technol. 4, 2267–2279 (2014).
19. Bond, J. Q., Wang, D., Alonso, D. M. & Dumesic, J. A. Interconversion
between γ-valerolactone and pentenoic acid combined with decarboxylation
to form butene over silica/alumina. J. Catal. 281, 290–299 (2011).
20. Bond, J. Q., Martin Alonso, D., West, R. M. & Dumesic, J. A. γ-Valerolactone
ring-opening and decarboxylation over SiO2/Al2O3 in the presence of water.
Langmuir 26, 16291–16298 (2010).
21. Bond, J. Q., Jungong, C. S. & Chatzidimitriou, A. Microkinetic analysis of
ring opening and decarboxylation of γ-valerolactone over silica alumina.
J. Catal. 344, 640–656 (2016).
22. Yun, G.-N., Ahn, S.-J., Takagaki, A., Kikuchi, R. & Oyama, S. T.
Hydrodeoxygenation of γ-valerolactone on bimetallic NiMo phosphide
catalysts. J. Catal. 353, 141–151 (2017).
23. Lin, W.-C. et al. Zinc-incorporated microporous molecular sieve for mild
catalytic hydrolysis of γ-valerolactone: a new selective route for biomass
conversion. ChemSusChem 11, 4214–4218 (2018).
24. Serrano-Ruiz, J. C., Braden, D. J., West, R. M. & Dumesic, J. A. Conversion of
cellulose to hydrocarbon fuels by progressive removal of oxygen. Appl. Catal.
B 100, 184–189 (2010).
25. Okuhara, T. Water-tolerant solid acid catalysts. Chem. Rev. 102,
3641–3666 (2002).
26. Nakajima, K. et al. Nb2O5·nH2O as a heterogeneous catalyst with water-
tolerant Lewis acid sites. J. Am. Chem. Soc. 133, 4224–4227 (2011).
27. Zhang, Y. et al. Direct conversion of biomass-derived carbohydrates to
5-hydroxymethylfurfural over water-tolerant niobium-based catalysts. Fuel
139, 301–307 (2015).
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
© The Author(s), under exclusive licence to Springer Nature Limited 2019
NATuRE MATERIALS | www.nature.com/naturematerials

Nature Materials
Articles
factors and well-fitted patterns with reasonable isotropic displacement factors
within experimental errors.
Methods
Catalyst preparation. HZSM-5 samples were purchased from Alfa Aesar and
are denoted as HZSM-5(Al/Si mole ratio). NbAlS-1 and NbS-1 samples were
prepared by hydrothermal synthesis, and denoted as NbAlS-1(Nb/Al/Si mole
ratio) and NbS-1(Nb/Si mole ratio), respectively. In a typical synthesis, aluminium
isopropoxide (99.99+% trace metals basis; Sigma Aldrich) was frst dissolved in
deionized water, into which tetrapropylammonium hydroxide solution (TPAOH,
1.0M in H2O; Sigma Aldrich) as the structure-directing agent was added. Te
mixture was stirred at room temperature for 2h, then niobium ethoxide (99.95%
trace metals basis; Sigma Aldrich) was added and the mixture stirred for another
2h. Next, tetraethyl orthosilicate (98%; Sigma Aldrich) was added dropwise and
the mixture was stirred for another 2h, which resulted in a gel with a chemical
composition of 1Si:xAl:yNb:0.25TPAOH:15H2O (x and y were determined by the
target Al/Si and Nb/Si mole ratios, respectively). Te gel was transferred into a
46ml Tefon-lined stainless-steel autoclave, which was sealed and heated at 170°C
for 48h. Te solid products were centrifuged, washed with deionized water, dried
overnight at 80°C and fnally calcined at 550°C under air fow for 6h. NbS-1
samples were synthesized by the same procedure, but without the addition of
aluminium isopropoxide. Except for the yield (30%) of NbAlS-1(0.04/0.027/1),
typical yields of all the zeolites in this study were 90–95%. NbOPO4 was
synthesized by a hydrothermal method at pH=2 according to the literature28.
Nb2O5 and Nb2O5·nH2O samples were purchased from Sigma Aldrich and CBMM,
respectively. Te methods of zeolite characterization are described in detail in
Supplementary Methods (Catalyst characterization).
INS. INS spectra were recorded on the VISION spectrometer at Spallation Neutron
Source, Oak Ridge National Laboratory, as well as on the TOSCA spectrometer at
the ISIS Facility at the STFC Rutherford Appleton Laboratory. Both VISION and
TOSCA are indirect geometry crystal analyser instruments that provide a wide
dynamic range with high resolution. All the INS spectra were collected after the
sample was cooled and stabilized at temperatures below 15K.
The adsorption/reaction experiments were conducted at beamlines TOSCA
and VISION and the results obtained from both spectrometers are consistent. In a
typical experiment, the catalyst (~11g) was loaded into a flow-type stainless-steel
cell that can also be used as a static cell with all the valves closed. The sample was
heated to 450°C (5°Cmin–1 ramping) under He for 3h to remove any remaining
trace water before the experiment. Then 20mmol of GVL was injected into the cell
at 135°C. Before the data collection, the cell was flushed using dry He to remove
weakly bound GVL molecules. The samples were cooled to <15K during the
data collection. After each reaction, the cell was flushed with dry He, sealed and
cooled for INS collection to detect the presence of possible reaction intermediates.
The INS spectra of pure solid compounds for both starting material and reaction
products were collected at 5K. The methods of DFT calculations and modelling
of the INS spectra are given in Supplementary Methods (DFT calculations and
modelling of the INS spectra).
Data availability
Catalytic testing. Catalytic reactions were carried out in a stainless-steel
continuous-flow reactor (12.7mm internal diameter). The catalyst (2g) was
pressed, crushed and sorted into grains by 40–60 meshes, and the grains were then
activated at 450°C for 3h under a nitrogen flow before the reaction. GVL was
injected into the nitrogen flow (50mlmin−1) by a syringe pump (Cole-Parmer) and
passed through the reactor at the target temperature. The output liquid products
were collected and analysed with an Agilent 7890B gas chromatograph equipped
with a HP-5 column (30m×0.32mm×0.25µm) and an Agilent 6890N–Agilent
5973N gas chromatograph–mass spectrometer (GC–MS) equipped with a HP-5MS
column (30m×0.25mm×0.25µm). The output gas products were collected and
analysed with an Agilent Micro GC 490 equipped with a PoraPLOT U column
(length 10m) and Agilent 7890A–Agilent 5975C GC–MS equipped with a HP-
PLOT/Q column (30m×0.53mm×40µm). GVL conversion was calculated by
the equation GVL conversion=(GVLin –GVLout)/GVLin ×100%, where GVLin and
GVLout denote the moles of GVL in the feed and exit, respectively. The product
selectivity was calculated on a carbon basis, that is, Si =(ai ×ni)/(Σ(ai ×ni))×100%,
where ai and ni denote the carbon number and mole of product i, respectively. CO2
was excluded in the selectivity calculation. The product yield was calculated by the
equation yield=conversion×selectivity×100%. The total carbon balances for the
production from GVL typically closed to within 5%.
All the relevant data are available from the authors, and/or are included with
the manuscript.
References
43. Tompson, S. P. et al. Beamline I11 at Diamond: a new instrument for high
resolution powder difraction. Rev. Sci. Instrum. 80, 075107 (2009).
Acknowledgements
We thank EPSRC (EP/P011632/1), the Royal Society and the University of Manchester for
funding. We thank the EPSRC National Service for EPR Spectroscopy at the University of
Manchester. A.M.S. thanks the Russian Science Foundation (Grant no. 17–73–10320) and
Royal Society of Chemistry for funding. We are grateful to Oak Ridge National Laboratory
(ORNL), the ISIS Facility and Diamond Light Source (DLS) for access to the beamlines
VISION, TOSCA and I11, respectively. We acknowledge DLS for the provision of beamtime
at B18 (UK Catalysis Hub SP15151, SP24726) and G. Cibin and V. Celorrio for help at B18
beamline. We acknowledge the support of The University of Manchester’s Dalton Cumbrian
Facility (DCF), a partner in the National Nuclear User Facility, the EPSRC UK National
Ion Beam Centre and the Henry Royce Institute. We recognize R. Edge and K. Warren
for their assistance during the 60Co γ-irradiation processes. We thank A. Jentys from the
Technical University of Munich and ISIS Facility for the measurement of the INS spectrum
of isobutene as part of RB20053 experimental proposal. We thank C. Webb for help with
GC–MS, D. Moulding for help with Raman spectroscopy and M. Kibble for help at the
TOSCA beamline. The computing resources were made available through the VirtuES and
the ICE-MAN projects, funded by the Laboratory Directed Research and Development
programme and by Compute and Data Environment for Science (CADES) at ORNL.
XAS. Transmission Nb K-edge X-ray absorption spectra were collected on
beamline B18 at Diamond Light Source, using a water-cooled Si(311) double
crystal monochromator and double toroidal Pt-coated mirror for beam focusing.
Data collection spanned 200eV before the edge to 1,000eV above it, with a step
size of 0.5eV; for reference, Nb foil spectra were simultaneously collected to enable
energy calibrations. GVL-saturated samples were collected as a slurry, contained
within a sealed Kapton capillary. Data processing was carried out in the Demeter
open source software package, version 0.9.26, with XAS spectra processing
(normalization and background subtractions) and extended X-ray absorption fine
structure fitting conducted within the Athena and Artemis programs, respectively,
within Demeter. Reference niobium oxide standards, NbO2 and Nb2O5, were also
collected after dilution in boron nitride.
Author contributions
L.L. and M.F carried out the syntheses and characterization of the zeolite samples. L.L.
and X.H. carried out the catalytic tests. L.L., A.M.S., F.T. and E.J.L.M. collected and
analysed the EPR data. L.L. and C.M.A.P. collected and analysed the XAS data. L.L.,
J.H.C., I.D.S. and C.C.T. collected and analysed the synchrotron X-ray diffraction data.
Z.T. and Y.L. collected and analysed the Py-IR data. L.L., Y.C., L.L.D., S.R. and A.J.R.-C.
collected and analysed the neutron scattering data and carried out the DFT modelling.
S.Y. was responsible for the overall direction of the project and preparation of the
manuscript, with contributions from all authors.
High-resolution SXPD and Rietveld refinement. SXPD data were collected
on Beamline I11, Diamond Light Source (wavelength=0.824869(10)Å) at
room temperature. The zeolite powder was loaded in a 0.7mm borosilicate glass
capillary. High-resolution diffraction data were obtained from the samples using
the multi-analyser crystal detectors. The patterns were collected in the 2θ range
0–150° with 0.001° data steps. Using the TOPAS software, the SXPD patterns were
refined by the Rietveld methods. The Thompson–Cox–Hastings pseudo-Voigt
peak function43 was applied to describe the diffraction peaks. The scale factor
and lattice parameters were allowed to refine for all the diffraction patterns.
The refined structural parameters include the fractional coordinates (x, y, z) and
isotropic displacement factors for all the atoms, and the site occupancy factors
for all the Nb, Al, Si species and GVL molecules. The quality of the Rietveld
refinements was assured with low goodness-of-fit factors, low weighted profile
Competing interests
S.Y. and L.L. are inventors of a patent based on this work.
Additional information
Supplementary information is available for this paper at https://doi.org/10.1038/
s41563-019-0562-6.
Correspondence and requests for materials should be addressed to S.Y.
Reprints and permissions information is available at www.nature.com/reprints.
NATuRE MATERIALS | www.nature.com/naturematerials