Zr-USY zeolite: Efficient catalyst for the transformation of xylose into bio-products

Article abstract of DOI:10.1016/j.cattod.2017.08.031

A series of bifunctional USY zeolite catalysts with different Al/Zr ratios were synthesised. The incorporation of Zr into the structure was accomplished after partial dealumination of H-USY. Structural and spectroscopic characterization confirmed the preservation of the zeolite network as well as the isolated incorporation of Zr atoms in Al vacancies, with no evidence of large zirconia domains. The catalytic evaluation in the transformation of xylose in 2-propanol allowed to obtain interesting mixtures of bio-products, and to identify the presence of two competitive routes: the formation of GVL following alternating acid-driven and hydrogen-transfer (MPV) steps, and the retro-aldol condensation of xylose. The extent of each of these two competing reaction cascades is strongly dependent on the Zr loading. Thus, the catalyst with the lowest Al/Zr ratio favours the xylose retro-aldol condensation. When considering furfural as starting substrate, only products involved in the cascade to GVL are obtained. In this case, the incorporation of Zr in the catalyst favoured the MPV reactions, enhancing furfural conversion rate. Thus, increasing concentrations of products coming from the reduction of furfural–furfuryl alcohol, furfuryl 2-propyl ether, lactones-, were detected over Zr-USY samples, with the Zr-USY-3 yielding 13.5% of the final product, GVL. The catalysts are reusable after thermal regeneration at 550 °C.

Full text of DOI:10.1016/j.cattod.2017.08.031

CatalysisTodayxxx(xxxx)xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Zr-USY zeolite: Efficient catalyst for the transformation of xylose into bio-
products
Clara López-Aguadoa, Marta Paniaguaa, Jose Iglesiasa, Gabriel Moralesa,⁎, José L. García-Fierrob,
Juan A. Meleroa
a
School of Experimental Sciences and Technology, Universidad Rey Juan Carlos, C/Tulipán s/n, E-28933, Móstoles, Madrid, Spain
Institute of Catalysis and Petrochemistry (ICP), CSIC, C/Marie Curie 2, E-28049, Cantoblanco, Madrid, Spain
b
A R T I C L E I N F O
A B S T R A C T
Keywords:
A series of bifunctional USY zeolite catalysts with different Al/Zr ratios were synthesised. The incorporation of Zr
into the structure was accomplished after partial dealumination of H-USY. Structural and spectroscopic char-
acterization confirmed the preservation of the zeolite network as well as the isolated incorporation of Zr atoms in
Al vacancies, with no evidence of large zirconia domains. The catalytic evaluation in the transformation of
xylose in 2-propanol allowed to obtain interesting mixtures of bio-products, and to identify the presence of two
competitive routes: the formation of GVL following alternating acid-driven and hydrogen-transfer (MPV) steps,
and the retro-aldol condensation of xylose. The extent of each of these two competing reaction cascades is
strongly dependent on the Zr loading. Thus, the catalyst with the lowest Al/Zr ratio favours the xylose retro-aldol
condensation. When considering furfural as starting substrate, only products involved in the cascade to GVL are
obtained. In this case, the incorporation of Zr in the catalyst favoured the MPV reactions, enhancing furfural
conversion rate. Thus, increasing concentrations of products coming from the reduction of furfural–furfuryl
alcohol, furfuryl 2-propyl ether, lactones-, were detected over Zr-USY samples, with the Zr-USY-3 yielding 13.5%
of the final product, GVL. The catalysts are reusable after thermal regeneration at 550 °C.
Bifunctional catalyst
Zr-USY zeolite
Zirconium
Cascade reaction
Xylose
Furfural
1. Introduction
lignocellulosic biomass is the synthesis of different bio-products starting
from xylose, an abundant monosaccharide derived from the hemi-
Chemical industry, strongly dependent on crude oil as raw material,
is driving towards the search for new sustainable and efficient alter-
natives that can substitute fossil sources. Among the different renew-
able options, abundant lignocellulosic biomass has the potential to
produce fuels and valuable chemicals through the transformation of
sugars coming from the hydrolysis of cellulose and hemicellulose, major
components of this biomass [1]. Currently, a great variety of catalytic
processes involving the transformation of the biomass carbohydrate
fraction into value-added products have been proposed. Such processes
usually require complex multi-step transformations, suffering from low
overall yields, which can seriously compromise the economic effi-
ciency. This limitation can be partly ameliorated by using integrated
processes based on the direct one-pot transformation of lignocellulose
derivatives into value added chemicals. In such systems, the use of a
multifunctional catalyst, able to catalyse several steps in one pot
avoiding the isolation of intermediate platform molecules, would be
highly desirable [2].
cellulose fraction of plant biomass. An approximation to the sequence
of reactions leading to the production of the different interesting mo-
lecules is represented in Scheme 1, considering γ-valerolactone (GVL)
as the final product. When this transformation is performed in alcohol
media, some of the bio-products that can be obtained comprise furfural,
furfuryl alcohol, furfuryl ether, levulinic acid/levulinates, angelica
lactones and GVL. All of them have commercial value: furfuryl ethers
can be used as blending components of gasoline and as flavours [3,4];
levulinic acid is considered as a platform molecule that can be trans-
formed in a wide range of chemicals, polymers, fuel additives, agro-
chemicals, etc. [5]; levulinates have applications in perfume and fla-
vouring industries, as blending agents for diesel fuel formulation, and
as plasticizers or solvents [6]; angelica lactones, like levulinates, have
been proposed as flavourings and fuel additives [7]; finally, GVL is a
molecule identified as fuel additive, polymer precursor and starting
material for the production of other advanced biofuels [8,9].
The cascade process starts with the dehydration of the mono-
saccharide to furfural, promoted by an acid catalyst [10]. Furfural can
In this field, an interesting route for the valorisation of waste
⁎
Corresponding author.
E-mail address: gabriel.morales@urjc.es (G. Morales).
http://dx.doi.org/10.1016/j.cattod.2017.08.031
Received 28 April 2017; Received in revised form 1 July 2017; Accepted 14 August 2017
0920-5861/©2017ElsevierB.V.Allrightsreserved.
Pleasecitethisarticleas:López-Aguado,C.,CatalysisToday(2017),http://dx.doi.org/10.1016/j.cattod.2017.08.031

C. López-Aguado et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Scheme 1. Reaction network for xylose conversion in 2-propanol into bio-products through alternative acid catalysed and MPV reactions.
subsequently be reduced by hydrogenation to furfuryl alcohol and
etherified to furfuryl ether, which in alcohol media evolves by hydra-
tion and isomerization to a mixture of levulinic acid and alkyl levuli-
nate in the presence of an acid catalyst. Finally, the hydrogenation of
these compounds followed by lactonization produces GVL [11]. The
direct and selective transformation of xylose into the different above-
described bio-products through a single-pot reaction would be highly
desirable, as the overall investment and operation costs would be re-
duced. Therefore, the rational design of a multifunctional catalyst
which can drive the chain of reactions to the diverse desirable bio-
products is of paramount importance in this field. In this case, the
proposed reaction pathway requires of a bifunctional catalyst, since
acid-driven transformations need to be combined with Meerwein–-
Ponndorf–Verley (MPV) reductions, using the adequate sacrificing al-
cohol in the reaction media, to allow the reduction of carbonyl groups.
The use of such catalytic-hydrogen-transfer reaction avoids the use of
elevated H2 pressures and expensive precious-metal catalysts. In this
regard, catalysts containing Lewis acid sites have been successfully
applied to this kind of MPV carbonyl moieties reduction [12]. On the
other hand, xylose dehydration and furfuryl alcohol transformation into
alkyl levulinates are mainly promoted by Brønsted acid sites [13].
The process using furfural as starting point of the cascade reaction
has been studied in previous works in literature. Román-Leshkov et al.
used a mixture of catalysts, Zr-beta and Al-MFI zeolites, for the efficient
production of GVL [14]. The combination of Brønsted and Lewis acid
sites in one single zeolite catalyst was investigated by the group of
Valente, analysing several combinations of metal species and loadings,
e.g. Sn-Al and Zr-Al-containing beta zeolites. They reached a mixture of
useful bio-products, though without achieving the final product in the
cascade, GVL [15,16]. More recently, Song et al. prepared a hier-
archical meso-Zr-Al-beta zeolite that exhibited remarkable catalytic
activity in the production of GVL from furfural [17]. All these works
demonstrate the potential of Zr & Al bifunctional zeolites in this kind of
multi-step reactions from furfural. Taking a further step, we have re-
cently reported the direct successful transformation of xylose into GVL
over bifunctional Zr-Al-Beta zeolite, being to the best of our knowledge
the first work using a single catalytic system for the whole process [18].
In such system, the incorporation of the zirconium species is achieved
after dealumination of a parent Al-beta zeolite in order to fix the zir-
conium atoms in the lattice Al vacancies generated. The so-in-
corporated framework Zr exhibits high activity in the MPV steps of the
reaction cascade. The adjustment of the Brønsted/Lewis acid sites ratio
by controlling the loadings of both Al and Zr in the beta zeolite turned
out to be a key aspect in the promotion of the alternating MPV and acid
driven reaction steps. Additionally, both the pore size and the pore
structure of the zeolite support could be important factors. Hence, in
this research, we propose the incorporation of Zr species, as MPV sites,
into the structure of a partially dealuminated zeolite H-USY. It can be
expected that the relatively large size of FAU-type framework pores
help in the progression of the reaction cascade by facilitating the
formation of bulky intermediates.
2. Experimental
2.1. Catalysts preparation
The synthesis procedure of zirconium-modified USY zeolites (Zr-
USY) was slightly modified from that reported in literature for zeolite
beta [18,19]. Thus, as parent support, a commercial H-USY (8.4 Si/Al
atomic ratio) was purchased from Zeolyst international. Partial deal-
umination was carried out by treatment in nitric acid solution (65% aq.
HNO3, Scharlau): two samples were prepared with a concentration of
0.5 and 2 M at RT (1 h, 20 mL g−1); and a third sample was prepared
under harsher conditions (10 M HNO3, 100 °C, 20 h, 20 mL g−1). After
washing with deionized water and centrifugation to recover the solid
product, the resulting materials were dried overnight (110 °C). Zirco-
nium incorporation was accomplished by impregnation in water with
the appropriate amount of Zirconium (IV) nitrate (Chemical Point),
suspending the pre-dealuminated USY zeolite in deionized water. The
solvent was removed under vacuum from the resultant slurry, and the
solid was dried overnight and calcined in air at 200 °C for 6 h
(3 °C min−1 heating ramp) and then at 550 °C for another 6 h with the
same heating rate. The three so-synthesised Zr-modified USY zeolites
were named as Zr-USY-n, wherein n is from 1 (softer dealumination
conditions) to 3 (harsher dealumination conditions).
2.2. Catalysts characterization
Zirconium and aluminium contents were determined by Inductively
Coupled Plasma-Optical Emission Spectroscopy (ICP-OES) with
a
Varian Vista AX apparatus. Textural properties of the tested catalysts
were calculated from argon adsorption–desorption isotherms, which
were recorded at 87 K using an AutoSorb equipment (Quantachrome
Instruments). Average pore size was estimated using the NLDFT
method, and total pore volume was taken at P/Po = 0.98. X-ray powder
diffraction (XRD) patterns were collected in a Philips X-pert dif-
fractometer using the Cu Kα line in the 2θ angle range from 5° to 65°
(step size of 0.04°). Structural characterization was completed by means
of transmission electron microscopy (TEM) on a Philips Tecnai-20
electronic microscope operating at 200 kV. Solid state 27Al MAS-NMR
and 29Si MAS-NMR spectra were acquired using a Varian Infinity
400 MHz spectrometer. Acid properties of the materials were de-
termined by means of temperature programmed desorption of NH3 in a
Micromeritics 2910 (TPD/TPR) equipment fitted with a TCD detector.
Additionally, acidity of the materials was further analysed by means of
n-propylamine chemisorption and subsequent thermal analysis on a
Mettler Toledo TGA/DSC1 Star System under a helium purge gas
(60 cm3 min−1) interfaced to a ThermoStar TM GSD 301 T3 mass
spectrometer. X-ray Photoelectron Spectroscopy (XPS) measurements
were obtained using
a Specs Phoibos 150 9MCD photoelectron
2

C. López-Aguado et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Table 1
Main physicochemical properties of parent H-USY and modified Zr-USY zeolites: composition, textural properties and acidity.
Catalyst
Compositiona
Textural properties
Acid properties
SBET (m2/g)
Vp (cm3/g)
c
Dp (Å)
Acid sites concentratione (mmol H+/g)
TD (°C)
b
d
f
% Al
% Zr
Al/Zr
SEXT
Sμp
H-USY
5.56
1.43
1.03
0.33
0.00
9.42
9.81
10.46
–
308
393
418
391
372
279
350
473
0.43
0.46
0.51
0.61
6.0
4.2
5.7
6.0
0.650
0.334
0.309
0.137
307
291
288
247
Zr-USY-1
Zr-USY-2
Zr-USY-3
0.49
0.33
0.10
a
Measured by ICP-OES: % Al (wt%); % Zr (wt%); Al/Zr (atomic ratio).
Specific surface area using the BET equation. External surface area and micropores area calculated with the t-plot method.
Total pore volume at a relative pressure of 0.98.
Average pore size using the NLDFT method.
Total acid capacity determined by ammonia TPD.
b
c
d
e
f
Maximum desorption temperature in ammonia TPD.
spectrometer.
with standard stock solutions of pure commercially-available chemicals
using decane as internal standard. Catalytic results are shown either in
terms of absolute conversion of furfural (XFAL) or xylose (XXYL) or in
terms of yields towards the different products (Yi). The definitions of
these parameters are as follows:
2.3. Catalytic tests
Catalytic experiments were performed under the following reaction
conditions: catalyst loading 0.01 g mL−1
, substrate to 2-propanol
Reacted mol offurfural/xylose
Initial mol offurfural/xylose
XFAL/XYL
=
× 100
(Scharlau, 98%) molar ratio 1:50, substrate loading 4 g for xylose (XYL)
(Sigma Aldrich, 99%) and 2.5 g for furfural (FAL) (Sigma Aldrich,
99%). Decane (Acros Organics > 99%) was used as internal standard
(0.01 g mL−1). Catalytic runs were carried out in a stainless steel stirred
autoclave (200 mL) with temperature control and pressure gauge. After
sealing the reactor, stirring was fixed at 1000 rpm and a heating rate of
2.5 °C min−1 was established. Samples were taken periodically and
filtered into a vial before analysis.
Formed mol of i
Initial mol offurfural/xylose
Y =
× 100
i
3. Results and discussion
3.1. Catalysts characterization
2.4. Products analysis
The bifunctional zeolites were synthesised by partial dealumination
of a commercial H-USY zeolite and subsequent incorporation of zirco-
nium into the vacant tetrahedral sites. The resultant lattice Zr sites have
been previously proven to be active in the reduction of carbonyl com-
pounds through MPV reaction [18,21]. Thus, applying increasingly
severe dealumination conditions, three materials with decreasing Al/Zr
atomic ratio were prepared. Table 1 summarizes the main physico-
chemical properties of these materials, including the parent H-USY for
comparison.
Conversion of xylose was evaluated by HPLC with an Evaporative
Light Scattering Detector (ELSD) (Agilent Technologies 1260 Infinity),
using a Hi-PLEX H+ column (300 mm-long, 7.7 mm-i.d., 8 μm particle
size). Reaction samples were also analysed by gas chromatography
(GC), using a Varian 3900 gas chromatograph fitted with a ZB-WAX
Plus column (30 m × 0.25 mm, DF = 0.25 μm) and a FID detector.
Compounds detected by GC were furfural (FAL) (Sigma Aldrich 99%),
furfuryl alcohol (FOL) (Sigma Aldrich 98%), 2-propoxy glycol (POG)
(99%), isopropyl furfuryl ether (FE) (Manchester Organics 97%), iso-
propyl levulinate (ILEV) (98%, see below), levulinic acid (LA) (Sigma
Aldrich 98%), α/β-angelica lactones (AngL) (Sigma Aldrich 98%), γ-
valerolactone (GVL) (Sigma Aldrich 99%) and methyl-β-D-xylopyrano-
side (Sigma Aldrich 99%). Isopropyl levulinate was synthetized by es-
terification of levulinic acid in 2-propanol with sulfuric acid as catalyst.
Purification of the levulinate was accomplished following a previously
described procedure [20]. After removal of 2-propanol by distillation
under reduced pressure, the product mixture was dissolved in MTBE
and extracted 3 times with saturated aq. KHCO3 solution and 3 times
with distilled water. The solution was dried with anhydrous Na2SO4
and then filtered. Finally, the solvent was removed by distillation under
reduced pressure. Purification grade was assessed by 1H NMR and 13C
NMR spectroscopy. 1H NMR (400 MHz, CDCl3): δ 4.97 (sep, J = 6.3 Hz,
1H), 2.72 (t, J = 6.7 Hz, 2H), 2.52 (dt, J = 0.4, 6.7 Hz, 2H), 2.17 (s,
3H), 1.2 (d, J = 6.3 Hz, 6H) ppm; 13C NMR (100 MHz, CDCl3): δ
206.98 (CH3C(O)e); 172.45 (eC(O)OiPr), 68.18 (eC(CH2)CH2), 38.19
(CH3C(O)Ce), 30.07 (CH3C(O)CH2Ce), 28.58 (CH3C(O)e), 21.96
(eC(CH2)CH2) ppm. The identification of the above-mentioned bio-
products was checked by GC–MS using a Bruker 320-MS GC Quadru-
pole Mass Spectrometer equipped with a capillary column (BR-SWax,
30 m × 0.25 mm, DF = 0.25 μm), and He as the carrier gas. Product
quantification was based on a previous calibration of the analysis unit
Depending on the employed process of acid treatment, the removal
of non-framework aluminium is followed by that located into the fra-
mework and, eventually, the collapse of the zeolite framework is ob-
served [22]. Therefore, as expected, in contrast with Beta zeolite
[18,23], the synthesis of a completely Al-free USY sample is not pos-
sible without affecting the crystalline microstructure. In this work, the
most severe dealumination conditions used (10 M aq. HNO3, 100 °C,
20 h), yielded a material with as low as 0.33 wt% Al content, starting
from 5.56 wt% in the commercial H-USY sample. In order to verify the
structural integrity of the prepared materials, XRD powder diffracto-
grams were recorded (Fig. 1). As shown, all of the Zr-containing sam-
ples preserved the typical XRD pattern of the parent zeolite H-USY. For
comparison purposes, a typical XRD pattern of crystalline ZrO2 has also
been included, evidencing the absence of significant diffractions cor-
responding to crystalline zirconia in the Zr-USY materials. Therefore,
XRD results suggest that the substitution of aluminium by zirconium
species does not damage the zeolite network and, on the other hand, the
incorporation of Zr mainly occurs onto the Al vacancies generated by
dealumination, finding no evidence of large zirconia domains.
Textural properties measured by argon adsorption-desorption iso-
therms (Fig. S1) and the corresponding t-plots analyses (Table 1) con-
firm the absence of significant structure damage during the synthesis
procedure. Moreover, the obtained Ar isotherms (Fig. S1), as well as the
3

C. López-Aguado et al.
CatalysisTodayxxx(xxxx)xxx–xxx
ZrO2, indicating the preferential incorporation into the silica frame-
work. However, the presence of small ZrO2 particles outside the net-
work as increasing metal loading in the zeolite structure cannot be fully
discarded. Fig. 2b depicts the correlation between the number of Zr
atoms effectively incorporated in the materials (quantified via ICP-OES)
and the number of tetrahedrally coordinated Al vacancies generated
during the dealumination step (estimated through Al content and 27Al
NMR, see Table S1). As shown, there exists a close agreement between
both values, with Zr loadings slightly over the number of aluminium
lattice vacancies for samples Zr-USY-1 and Zr-USY-2. In the case of Zr-
USY-3, the sample with the higher zirconium loading, an almost perfect
agreement is obtained. This is further indication of the efficient in-
corporation of Zr atoms within the zeolite structure, occupying the
spaces left by leached framework Al atoms.
The acid properties of the series of Zr-modified USY zeolites have
been evaluated, through the quantification of the total number of acid
sites, using NH3-TPD (Table 1, Fig. S6) and quantifying the reactively-
formed propene desorption from chemisorbed n-propylamine by TGA-
MS (Table S2, Fig. S7). Ammonia desorption occurred between 200 and
500 °C, recording a maximum desorption temperature of 307 °C for USY
parent material. In the zirconium-modified materials, the removal Al
from the network resulted in a progressive decrease of the acid sites
loading of the materials. Furthermore, the strength of the remaining
acid sites decreases with the Zr content, as can be deduced from max-
imum desorption temperature values displayed in Table 1. Regarding
the propylamine TGA-MS analysis, for Zr-USY zeolites two main peaks
have been observed (Fig. S7), the first one in the range 275–350 °C
corresponding to the strongest acid sites (able to catalyse the decom-
position of n-propylamine into ammonia and propene at the lowest
temperature), usually considered strong Brønsted acid sites. The second
peak, much larger, appears in the range 450–500 °C, and it can be as-
signed to weak acid sites (either Bronsted or Lewis in nature). It can be
deduced that the removal of Al from the network, and the subsequent
incorporation of Zr, resulted in a progressive decrease of the acid sites
loading of the materials, which is consistent with the commented NH3-
TPD results (Table 1). Regarding the strength of the remaining acid
sites, it is observed that the population of strong acid sites keeps almost
constant in the three materials, and it may be attributed to the most
recalcitrant Al species that cannot be extracted from the USY frame-
work by dealumination without inducing the collapse of the structure.
On the other hand, the concentration of weak acid sites decreases in-
sofar as the Al/Zr ratio decreases.
Fig. 1. Powder X-ray diffraction patterns of H-USY, Zr-modified materials and crystalline
tetragonal ZrO2.
resultant pore size distributions (Fig. S2), are very similar for the whole
series of USY catalysts. However, total pore volume and surface areas
show a small increase with the dealumination process severity. This
effect likely comes from the fragmentation of the parent USY surface
during the dealumination step, indicating that the preparation method,
although not significantly altering the zeolite structure or causing pore
blockages, introduces some degree of surface alteration. Noticeably,
such an alteration in the surface might slightly alter the local crystal-
linity. Indeed, the most varying parameter is the surface area of mi-
cropores (Sμp), which increases from 279 m2/g in Zr-USY-1 to 473 m2/g
in Zr-USY-3. TEM analysis (Fig. S3) also confirms the adequate con-
servation of the structure integrity, though the small distortion effect on
the surface can likewise be inferred from the image blurring in the
modified material Zr-USY-1. Regarding the incorporation of zirconium,
TEM micrographs do not show any evidence of ZrO2 bulk agglomerates,
in consonance with a lattice incorporation of the Zr species.
Octahedral extra-framework Al and tetrahedral framework Al are
considered as the main aluminium species in conventional USY zeolites,
though other minor species (octahedral Al species within the zeolite
framework, distorted tetrahedral Al, and five-coordinated non-frame-
work Al) have also been reported [22,24]. Leaching with low con-
centrated nitric acid removes mainly the non-framework species, while
harsher conditions (higher acid concentration or temperature) lead to
tetrahedral framework Al removal from the lattice, leading to the for-
mation of a great number of silanol groups, where the zirconium species
can be allocated. In order to get a deeper understanding of the co-
ordination number of remaining Al atoms within the synthesised Zr-
modified zeolites 27Al solid-state MAS NMR spectra were collected. The
presence of both tetrahedral framework Al (δ ≈ 55–60 ppm) and oc-
tahedral extra-framework Al (δ ≈ 0 ppm) was observed [24] with a
different degree of removal for each type of Al species in the deal-
umination step (see Supporting Information, Table S1 and Fig. S4). The
ratio between the population of both species, i.e. tetra/octa ratio, in-
creases with the severity of the dealumination process, corroborating
the preferential leaching of extra-framework Al species. On the other
hand, 29Si solid-state MAS NMR confirm the removal of framework Al
from the silica structure, as evidenced by the reduction of the signal at
δ ≈ −105 ppm [Si(1Al)] and the corresponding increase of the signal
at δ ≈ −115 ppm [Si(0Al)] (Fig. S5).
3.2. Catalytic evaluation of Zr-USY zeolites
The Zr-containing USY zeolites, prepared as mentioned in the ex-
perimental, display two types of catalytic functional activities: Brønsted
acid sites, coming from charge-deficit compensating protons in Al tet-
rahedra, and Lewis acid sites provided by remaining extra-framework
aluminium species as well as from the zirconium atoms isomorphically
substituting the extracted aluminium species. The catalytic activity of
these bifunctional materials has been assessed in the transformation of
xylose and furfural into γ-valerolactone (GVL) in 2-propanol as reaction
media. The proposed reaction route to achieve these transformations,
previously reported in literature [15–18,25,26], involves a cascade of
reactions in which Brønsted acid-mediated transformations–dehydra-
tion, hydration, isomerisation, intramolecular transesterification- are
combined with hydrogen transfer reactions through a Meerwein-
Ponndorf-Verley (MPV) pathway for the reduction of carbonyl com-
pounds, in which Lewis acid sites, such as those present in Zr-USY
zeolites, are present. The result is the direct transformation of xylose -or
furfural- into GVL following the proposed reaction scheme (Scheme 1).
On the basis of this scheme, two distinct activity studies were carried
out: one starting from xylose, and the other one starting from furfural.
Circumventing the first dehydration step -xylose to furfural-avoids the
in-situ generation of water molecules, thus reducing the extension of
In order to assess the incorporation of Zr into the generated Al va-
cancies, the materials were further characterized by means of XPS
analysis (Fig. 2a). Spectra reveal a shift of the Zr 3d signals to lower
binding energies, closer to the signal attributed to zirconium oxide, as
the Zr content increases. These results suggest the different Zr en-
vironment in Zr-USY materials in comparison with pure and crystalline
4

C. López-Aguado et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Fig. 2. (a) Zr 3d XPS spectra of Zr-USY modified
zeolites and zirconia; (b) correlation between
amount of incorporated Zr and amount of extracted
framework Al.
several water-mediated side reactions such as humins generation [27].
In addition, reducing the production of water in the reaction media
limits the inactivation of zirconium species, as these tend to strongly
adsorb water molecules [28], thus reducing their capability to interact
with other ligands such as the reaction substrates.
propanol at room temperature [29] Commercial USY zeolite in acid
form was used as catalyst in first term, to serve as reference catalysts,
and to identify possible reaction pathways. This material displays rather
low initial catalytic activity, since only alkyl xylosides are detected
during the early stages of the reaction. Alkyl xylosides are easily pro-
duced under the tested reaction conditions [30], as even weak acid
catalysts can drive the alkoxylation of sugars. With time, xylose ethers
react to provide furfural, which turns to be the major product after 3 h.
Furfural is then reduced to furfuryl alcohol and etherified with 2-pro-
panol to the corresponding 2-propyl furfuryl ether. These results evi-
dence the net positive activity of commercial H-USY zeolites not only in
the acid-driven steps but also in the MPV reduction of furfural. This
3.2.1. Xylose to GVL
Reaction tests on the direct transformation of xylose into GVL were
carried out at 170 °C for 24 h, using Zr-USY zeolites as catalysts (Fig. 3).
2-propanol was used as solvent, and sacrificing alcohol in MPV reac-
tions, in a xylose to alcohol molar ratio of 1:50, which is that corre-
sponding to the maximum soluble concentration of this sugar in 2-
Fig. 3. Xylose conversion and yields to products in 2-propanol over a) parent H-USY; b) Zr-USY-1; c) Zr-USY-2; d) Zr-USY-3 zeolites. Reaction conditions: 170 °C, substrate: 2-propanol
molar ratio 1:50, catalyst loading 0.01 g mL−1
.
5

C. López-Aguado et al.
CatalysisTodayxxx(xxxx)xxx–xxx
transformation is probably being mediated by Lewis acid sites corre-
sponding to the extra-framework aluminium atoms stabilizing the FAU
structure of USY zeolites [31]. However, both furfuryl alcohol and 2-
propyl furfuryl ether are present in a very low concentration in the
reaction media, suggesting that their sequential transformation into 2-
propyl levulinate, the next transformation of the reactions cascade, is
faster than the production of these furan derivatives. 2-propyl levuli-
nate appears to be the most abundant product after furfural, but no
traces of GVL in the reaction media were detected under the used re-
action conditions. In this case, the lower reactivity of ketones in MPV
reactions, as compared to aldehydes [32], as well as the low intrinsic
catalytic activity of aluminium in MPV transformations, could be the
main factors explaining the absence of GVL.
Simultaneously to the main cascade of reactions, the existence of a
competitive side reaction leading to the production of α- and β-an-
gelica-lactones (AngL) from furfuryl ether (FE)/furfuryl alcohol (FOL)
may also be considered (Scheme 1). We have previously demonstrated
the occurrence of this reaction over H-beta zeolites [13]. Other authors
have also attributed the formation of AngL, observed over Zr-Al-Beta
zeolites [17], to a mechanism involving the enolisation of levulinic
acid/alkyl levulinates, followed by intramolecular esterification/trans-
esterification into the corresponding lactone [33,34]. Furthermore, the
conversion of the angelica-lactones into GVL is also plausible [29,35].
In this way, this route is actually, more than a competitive reaction, an
alternative route for the production of GVL from xylose. Nevertheless,
the formation of AngL is limited in the presence of the parent H-USY, so
the main reaction route in this case seems to proceed through 2-propyl
levulinate.
The formation of 2-propyl lactate (PL) and 2-propoxy glycol (POG)
has also been detected when treating the xylose/2-propanol mixtures in
the presence of H-USY zeolite. Such products do not come from the
proposed reaction scheme to form GVL, and thus the existence of a side
reaction should be considered. The formation of these products can be
explained by the retro-aldol condensation of xylose into 2-propoxy-
glycolaldehyde and glyceraldehyde [36]. The formation of 2-propoxy
glycol aldehyde is supposed to come from the retro-aldol condensation
of alkyl xylosides, which already show the isopropoxy moiety. These
primary products can be transformed, in subsequent reaction steps,
since 2-propoxy glycol aldehyde can derive into ethylene glycol mono-
isopropyl ether by MPV reduction. On the other hand, glyceraldehyde
can suffer a dehydration step to form an enol, which is transformed into
pyruvaldehyde. Hydration and intramolecular Cannizaro reaction can
lead to the transformation of this chemical into lactic acid, which un-
dergoes the esterification with the alcohol solvent [37]. A tentative
reaction scheme of these transformations has been proposed and in-
cluded in Scheme 2.
whereas the pathway starting with xylose dehydration was minimized.
A plausible explanation for this behaviour is the strong reduction ob-
served in the acidity present in Zr-USY zeolites when increasing the
metals substitution degree (see Table 1, Fig. S6, Table S2, Fig. S7). This
behaviour has been previously described by Taarning et al. [38], when
treating methanol solutions of xylose in the presence of Sn-Beta zeolite,
although products distribution was different in this case. These authors
found methyl lactate, although no methoxy glycol was observed, but
methyl-vinylglycolate (MVG) and methyl 2-hydroxy-4-methoxy-bu-
tanoate (MHMB) instead. Moreover, the isopropyl counterparts of these
chemicals have not been detected in our investigation. The dis-
crepancies between products distributions reported for Sn-Beta as
compared to those here presented could be ascribed to several factors,
including the nature of the metal species -Sn vs. Zr-, which avoids the
etherification of the starting xylose to form methyl xyloside. In this
way, no methyl glycol aldehyde is formed and thus, glycol aldehyde can
undergo an aldol condensation to form a tetrose, which is the origin of
MVG and MHMB. On the other hand, another important difference is
the ability of the alcohol media to serve as sacrificing alcohol. Though
primary alcohols, such as the methanol used by Taarning et al., can be
transformed by MPV reductions, these are much slower than if using
secondary alcohols. Thus, the reduction of glycol aldehyde to ethylene
glycol does not seem to take place in the presence of methanol.
The influence of Al substitution by Zr species in the promotion of
xylose retro-aldol condensation is depicted in Fig. 4. Products dis-
tribution obtained in the presence of USY material evidences the very
rapid transformation of xylose into 2-propyl xylosides and furfural, as it
corresponds to a strong acid catalyst such as H-USY zeolite. On the
contrary, dealumination of USY abruptly decreases the acidity of these
materials, and thus, the rate of xylose dehydration is slowed down. In
addition, dealuminated USY zeolites display a poor catalytic activity in
retro-aldol condensation, so that the most abundant product present
when treating xylose with this material is 2-propyl xyloside. In-
corporation of zirconium strongly modifies the catalytic behaviour of
these materials, as it is evident from the strong promotion of retro-aldol
reactions. These reactions have been described to be promoted by zir-
conia-based materials, showing both acid and basic catalytic sites as-
sociated to zirconium species [39]. To the best of our knowledge, this
investigation is the first one to prove the existence of these transfor-
mations in the presence of Zr-containing zeolites, showing no extra-
framework zirconium species.
As shown, the progressive substitution of aluminium species by
zirconium exerts a key role on the catalytic behaviour of Zr-USY ma-
terials, not only on the promotion of different reaction pathways, but
also on the extent of the different transformations taking place in the
cascade transformation from xylose to GVL. Thus, increasing Zr loading
leads to a lower production of furfural, but also to a noteworthy in-
crease in the production of 2-propyl levulinate. These results are con-
sequence of the different balance between MPV and acid driven reac-
tions insofar as the Al/Zr ratio changes. This substitution leads to, as
above described, the reduction of the acidity in Zr-USY zeolites as
compared to the parent material. As previously discussed, both the acid
All the above-discussed transformations, the formation of GVL and
the retro-aldol route, have also been detected when using Zr-USY zeo-
lites as catalysts, though the extent of the two competing reaction
cascades drastically changed insofar as the zirconium loading in-
creased. Thus, the substitution of aluminium species by zirconium led
to an increase of the extent of xylose retro-aldol condensation reactions,
Scheme 2. Proposed reaction scheme for retro-aldol condensation of xylose into 2-propyl lactate and 2-propoxy glycol.
6

C. López-Aguado et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Fig. 4. Yields to xylose ethers, furfural, and xylose derivatives (products coming from the retro-aldol condensation of xylose) obtained in 2-propanol over (a) H-USY, (b) De-aluminated
USY and (c) Zr-USY-3 zeolites. Reaction conditions: 170 °C, substrate: 2-propanol molar ratio 1:50, catalyst 0.01 ng mL−1
.
capacity and the strength were gradually reduced from USY parent
material to Zr-USY-3, which is the material showing the lower acidity
and higher Zr content among those tested in this investigation.
However, zirconium incorporation also leads to an enhancement of the
MPV transformations, as this is more active than aluminium in hy-
drogen-transfer reactions using secondary alcohols as hydrogen source.
In this way, the conversion of furfural into furfuryl alcohol, and derived
products, is enhanced as the zirconium content increases, thus yielding
lower amounts of furfural, but higher yields to 2-propyl levulinate.
Finally, though scarce, some production of GVL was detected in the
reaction media treated in the presence of Zr-USY zeolites, because of
the same already mentioned reasons. However, if quantifying the to-
talized product yields coming from the dehydration of xylose, it seems
that the reduction in acidity seems to be accompanied by a strong re-
duction in the catalytic activity of Zr-containing USY zeolites in the
desired reaction cascade. Xylose dehydration is the first stage of the
desired reaction pathway to form GVL, so that when this first reaction is
limited, the overall transformation to GVL is limited too.
as main reaction product. Nevertheless, the small amount of furfuryl 2-
propyl ether undergoing hydration and isomerisation was readily
transformed in GVL, yielding 13.5% of the starting furfural as the de-
sired lactone, the maximum value among those obtained with the dif-
ferent Zr-USY materials. In order to assess the advantages of a single
multifunctional catalyst, a physical mixture of two catalysts bearing
separately the Al and Zr functions, such as USY zeolite and ZrO2, was
tested for furfural conversion into GVL. The results of products yields
(Fig. S8) show the worst performance when a combination of the two
catalysts is used, implying how the incorporation of both Zr and Al sites
in the structure of a single material increases the accessibility to the
active sites, thus enhancing the cascade of reactions.
Reusability of the catalysts was evaluated by reutilization experi-
ments of the parent USY and one of the zirconium-modified zeolites, Zr-
USY-1. Catalysts were recovered by filtration after a first reaction cycle
of 24 h at 170 °C. The samples, once dried, presented dark colour,
especially in the case of using xylose as substrate. This can be attributed
to the adsorption of reaction products and non-identified compounds,
i.e. humins. In order to remove the organic deposits from the catalysts
surface, a thermal treatment in air (5 h at 550 °C) was applied.
Thereafter, the catalysts were used again in a second identical reaction
cycle. Fig. 6 includes the results in terms of yields to the identified
products. As shown, both materials displayed a very good reusability
from furfural, evidencing that the active sites keep their activity es-
sentially intact. It is especially remarkable the almost identical result
obtained in the two consecutive reaction cycles over the Zr-modified
zeolite with furfural as substrate. In contrast, a slight decrease in ac-
tivity is shown in the reaction with xylose, indicating a different nature
of the organic deposits formed on the catalysts.
3.2.2. Furfural to GVL
Starting from furfural instead of xylose in the production of GVL
avoids the first reaction stage -xylose dehydration-, but also prevents
the formation of side products coming from the retro-aldol condensa-
tion of the monosaccharide. In this way, Zr-containing USY zeolites
have been tested in the direct transformation of furfural into GVL, being
the obtained results depicted in Fig. 5.
As expected, no retro-aldol condensation reactions have been de-
tected when treating furfural with Zr-USY materials and, obviously,
only products involved in the reactions cascade from furfural to GVL
were detected. Under these conditions, the H-USY parent material
displayed a quite poor catalytic performance, since the reduction of
furfural to GVL was accomplished in a very slow way. This is a con-
sequence of the poorer catalytic performance of aluminium, as com-
pared to zirconium, in MPV reactions. Indeed, despite the scarce
amount of produced furfuryl alcohol, completely transformed into AngL
and 2-propyl levulinate, no traces of GVL were detected in the presence
of H-USY parent material, which is another piece of evidence of the low
catalytic performance of H-USY in MPV transformations. Incorporation
of zirconium led to the acceleration of hydrogen-transfer reactions. In
this way, the furfural conversion rate was increased with the zirconium
content isomorphically incorporated to the catalyst. Thus, increasing
concentrations of products coming from the reduction of furfural -fur-
furyl alcohol, furfuryl 2-propyl ether-, were detected over samples Zr-
USY-1 and Zr-USY-2. These products were accompanied by the pro-
duction of GVL, resulting from the reduction and lactonization of 2-
propyl levulinate. Nevertheless, when treating furfural with Zr-USY-3,
the material showing the lowest Al/Zr ratio (Table 1), product dis-
tribution was displaced towards the formation of furfuryl 2-propyl ether
4. Conclusions
In conclusion, within this work we show that bifunctional zeolite
USY with different Al/Zr ratios can be selective catalysts to produce a
diverse distribution of bio-products in the cascade transformation from
xylose/furfural to GVL. The incorporation of Zr into the zeolite struc-
ture, after partial dealumination of a parent USY zeolite, was success-
fully accomplished preserving the zeolite network integrity and allo-
cating the Zr atoms into generated Al vacancies, with no evidence of
large ZrO2 domains. The catalytic evaluation in the transformation of
xylose in 2-propanol allowed identifying the presence of two competi-
tive catalytic routes: the formation of GVL following alternating acid-
driven and hydrogen-transfer (MPV) catalytic steps, and the retro-aldol
condensation of xylose. The extent of each of these two competing re-
action cascades is strongly dependent on the Zr loading of the catalyst.
Starting from furfural, no retro-aldol condensation is present and only
products involved in the reactions cascade to GVL are obtained. In this
case, the incorporation of Zr in the catalyst led to the acceleration of
7

C. López-Aguado et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Fig. 5. Furfural conversion and yields to products in 2-propanol over a) parent H-USY, b) Zr- USY-1, c) Zr-USY-2, d) Zr-USY-3 zeolites. Reaction conditions: 170 °C, substrate: 2-propanol
molar ratio 1:50, catalyst 0.01 g mL−1
.
072709). We thank IMDEA Energy Institute (Madrid, Spain), especially
H. Hernando and A. Berenguer, for the collaboration with the analysis
of Temperature Programmed Decomposition of chemisorbed n-propy-
lamine by TGA-MS.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the
online version, at http://dx.doi.org/10.1016/j.cattod.2017.08.031.
References
[1] J.J. Bozell, G.R. Petersen, Green Chem. 12 (2010) 539.
[2] M.J. Climent, A. Corma, S. Iborra, M.J. Sabater, ACS Catal. 4 (2014) 870–891.
[3] A.J.J.E. Eerhart, W.J.J. Huijgen, R.J.H. Grisel, J.C. van der Waal, E. de Jong, A. de
Sousa Dias, A.P.C. Faaij, M.K. Patel, RSC Adv. 4 (2014) 3536–3549.
[4] Bart Vanderhaegen, Hedwig Neven, Luk Daenen, Kevin J. Verstrepen,
H. Verachtert, G. Derdelinckx, J. Agric. Food Chem. 52 (2004) 1661–1668.
[5] G.M. González Maldonado, R.S. Assary, J.A. Dumesic, L.A. Curtiss, Energy Environ.
Sci. 5 (2012) 8990.
Fig. 6. Reutilization of USY and Zr-USY-1. Yields to products starting from furfural (left)
and xylose (right) in two consecutive uses with intermediate thermal activation of the
catalysts (550 °C). Reaction conditions: 170 °C, 24 h, substrate: 2-propanol molar ratio
1:50, catalyst loading 0.01 g mL−1
.
[6] A. Démolis, N. Essayem, F. Rataboul, ACS Sustain. Chem. Eng. 2 (2014) 1338–1352.
[7] R. Cao, J. Xin, Z. Zhang, Z. Liu, X. Lu, B. Ren, S. Zhang, ACS Sustain Chem. Eng. 2
(2014) 902–909.
[8] I.T. Horváth, H. Mehdi, V. Fábos, L. Boda, L.T. Mika, Green Chem. 10 (2008)
238–242.
[9] D.M. Alonso, S.G. Wettstein, J.A. Dumesic, Green Chem. 15 (2013) 584.
[10] I. Agirrezabal-Telleria, A. Larreategui, J. Requies, M.B. Güemez, P.L. Arias,
Bioresour. Technol. 102 (2011) 7478–7485.
MPV reactions, increasing the production furfuryl alcohol, furfuryl 2-
propyl ether, and lactones. The catalysts showed a good reusability
after thermal regeneration in the reaction from furfural and xylose,
confirming their stability.
[11] X. Tang, X. Zeng, Z. Li, L. Hu, Y. Sun, S. Liu, T. Lei, L. Lin, Renew. Sustain. Energy
Rev. 40 (2014) 608–620.
Acknowledgements
[12] M. Moliner, Dalton Trans. 43 (2014) 4197–4208.
[13] M. Paniagua, J.A. Melero, J. Iglesias, G. Morales, B. Hernández, C. López-Aguado,
Financial support from the Spanish Ministry of Economy and
Competitiveness through the project CTQ2014-52907-R, as well as from
the Regional Government of Madrid through the project S2013-MAE-
2882 is gratefully acknowledged. C. López-Aguado thanks the Spanish
Ministry of Economy and Competitiveness for a FPI grant (BES-2015-
Appl. Catal. A Gen. 537 (2017) 74–82.
[14] L. Bui, H. Luo, W.R. Gunther, Y. Román-Leshkov, Angew. Chem. Int. Ed. 52 (2013)
8022–8025.
[15] M.M. Antunes, S. Lima, P. Neves, A.L. Magalhães, E. Fazio, A. Fernandes, F. Neri,
C.M. Silva, S.M. Rocha, M.F. Ribeiro, M. Pillinger, A. Urakawa, A.A. Valente, J.
8

C. López-Aguado et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Catal. 329 (2015) 522–537.
[26] H.P. Winoto, B.S. Ahn, J. Jae, J. Ind. Eng. Chem. 40 (2016) 62–71.
[16] M.M. Antunes, S. Lima, P. Neves, A.L. Magalhães, E. Fazio, F. Neri, M.T. Pereira,
A.F. Silva, C.M. Silva, S.M. Rocha, M. Pillinger, A. Urakawa, A.A. Valente, Appl.
Catal. B Environ. 182 (2016) 485–503.
[27] L. Yan, A.A. Greenwood, A. Hossain, B. Yang, RSC Adv. 4 (2014) 23492–23504.
[28] J. Wang, K. Okumura, S. Jaenicke, G.-K. Chuah, Appl. Catal. A Gen. 493 (2015)
112–120.
[17] S. Song, L. Di, G. Wu, W. Dai, N. Guan, L. Li, Appl. Catal. B Environ. 205 (2017)
[29] F. Montañés, A. Olano, E. Ibáñez, T. Fornari, AIChE J. 53 (2007) 2411–2418.
[30] J. Iglesias, J.A. Melero, G. Morales, M. Paniagua, B. Hernández, ChemCatChem 8
(2016) 2089–2099.
393–403.
[18] B. Hernández, J. Iglesias, G. Morales, M. Paniagua, C. López-Aguado, J.L. García
Fierro, P. Wolf, I. Hermans, J.A. Melero, Green Chem. 18 (2016) 5777–5781.
[19] J. Wang, S. Jaenicke, G.-K. Chuah, RSC Adv. 4 (2014) 13481.
[20] J. Geboers, X. Wang, A.B. De Carvalho, R. Rinaldi, J. Mol. Catal. A Chem. 388–389
(2014) 106–115.
[21] J. Iglesias, J.A. Melero, G. Morales, J. Moreno, Y. Segura, M. Paniagua, A. Cambra,
B. Hernández, Catalysts 5 (2015) 1911–1927.
[22] Z. Yan, D. Ma, J. Zhuang, X. Liu, X. Liu, X. Han, X. Bao, F. Chang, L. Xu, Z. Liu, J.
[31] L.D. Borges, J.L. de Macedo, Microporous Mesoporous Mater. 236 (2016) 85–93.
[32] V. Polshettiwar, R.S. Varma, Green Chem. 11 (2009) 1313.
[33] K.R. Enslow, A.T. Bell, J. Gmehling, C. Tyrsted, E.D. Bojesen, N. Lock,
S.J.L. Billinge, B.B. Iversen, D.G. Vlachos, Catal. Sci. Technol. 5 (2015) 2839–2847.
[34] J.-P. Lange, W.D. van de Graaf, R.J. Haan, ChemSusChem 2 (2009) 437–441.
[35] M.G. Al-Shaal, W. Ciptonugroho, F.J. Holzhäuser, J.B. Mensah, P.J.C. Hausoul,
R. Palkovits, D.A. Kriz, C. Guild, A. Gudz, S.L. Suib, Catal. Sci. Technol. 5 (2015)
5168–5173.
Mol. Catal. A: Chem. 194 (2003) 153–167.
[23] Y. Zhu, G.-K. Chuah, S. Jaenicke, J. Catal. 227 (2004) 1–10.
[24] A. Corma, V. Fornés, A.J. Martínez Sanz, Tetrahedral and octahedral extraframe-
work aluminum in ultrastable Y zeolites, implications in the cracking of gas oil, in:
Mario L. Occelli (Ed.), Fluid Catalytic Cracking, ACS Symposium Series, vol. 375,
1988, pp. 17–33 (Chapter 2).
[36] N. Paksung, R. Nagano, Y. Matsumura, Energy Fuels 30 (2016) 7930–7936.
[37] M.G. Manas, J. Campos, L.S. Sharninghausen, E. Lin, R.H. Crabtree, Green Chem. 17
(2015) 594–600.
[38] M.S. Holm, Y.J. Pagán-Torres, S. Saravanamurugan, A. Riisager, J.A. Dumesic,
E. Taarning, Green Chem. 14 (2012) 702.
[25] M.M. Antunes, P. Neves, A. Fernandes, S. Lima, A.F. Silva, M.F. Ribeiro, C.M. Silva,
[39] P. Wattanapaphawong, P. Reubroycharoen, A. Yamaguchi, RSC Adv. 7 (2017)
18561–18568.
M. Pillinger, A.A. Valente, Catal. Sci. Technol. 6 (2016) 7812–7829.
9