M. Cabanillas, et al.
Molecular Catalysis 477 (2019) 110522
firstly reported for the conversion of levulinic acid and its esters by
Dumesic and co-workers [14] using ZrO as catalysts and 2-butanol as
2
solvent. Other Zr-based catalyst have been reported to catalyse the
transformation of alkyl levulinate to GVL following a MPV mechanism
been performed via a microinyector using a cyclohexanic solution of
the probe molecule (0.989 M in Py and 0.686 M in DMPY, respectively).
Prior to the analysis, the catalyst have been standardised in a dehy-
drated nitrogen flow of 50 mL/min during one hour at 300 °C.
including MOFs [15], basic zirconium carbonate [16], and ZrO
2
sup-
Diffuse Reflectance Infrared Fourier Transform (DRIFT) spectro-
scopy was employed to characterise the used catalysts using an ABB
MB3000 provided with a Pike diffuse reflectance accessory DiffusIR
ported materials on SBA-15 silicates [17,18]. Additionally, bifunctional
materials such as zirconium modified zeolites have been investigated
for the one-pot transformation of xylose to GVL [19,20].
The main difficulty that lignocellulosic biomass valorisation pro-
cesses must address relate to achieving high yields towards the desired
product. In this sense, continuous flow systems offer advantages over a
batch reactor such as better control of reaction conditions (conversion,
selectivity), easy scale-up and minimize catalyst manipulation [21–24].
−1
with 8 cm of resolution and 256 scans. The spectra were acquired in
−
1
the wavenumber range between 4000 and 600 cm
ference.
using KBr as re-
2.3. Catalytic activity
Herein, we report the synthesis of zeolites supported ZrO
2
catalysts
The catalytic activity of as-synthesised materials was evaluated in
hydrogen donor solvent, at 200 °C and 30 bar pressure. In a typical
experiment, 0.3 M methyl levulinate in 2-propanol was passed through
the catalyst bed (˜0.5 g) at a flow rate of 0.2 mL/min. Additionally, to
study the role of the acid sites in the system, a 0.3 M methyl levulinate
and 0.01 M solution in PY was passed through the catalyst bed under
same conditions mentioned before. Samples were withdrawn at 15, 30
and 60 min, once achieved steady state (corresponding to residence
times of the feed in the catalyst ca. 40, 80 and 160 s). The reaction
mixture was analysed using an Agilent 5890 series II gas chromato-
graph, fitted with a flame ionisation detector (FID) and a non-polar
fused silica ihydroxy column SUPELCO EQUITY TM-1 (60 m ×
0,25 mm × 0,25 μm).
via ball-milling (mechanochemistry) and evaluated the role that acid
sites as well as zirconium play in the continuous flow transfer hydro-
genation of biomass-derived methyl levulinate to GVL using 2-propanol
as hydrogen donating solvent.
2
. Materials and methods
2
.1. Materials preparation
The materials used to carry out this work were employed as re-
ceived without further purification. Methyl levulinate used as substrate
for the hydrogenation reaction was provided by Avantium as by-pro-
duct from the YXY process after further purification [25]. ZrO(NO
3 2
)
xH O, employed as metal salt precursor, was purchased from Sigma-
2
Aldrich. 2-propanol used as hydrogen donor solvent for the transfer
hydrogenation of methyl levulinate was purchased from Merck. Zeolites
with different frameworks (ZSM-5 and BEA) and Si/Al (30, 50 and 75)
molar ratios and frameworks were obtained from Zeolyst International
Inc. Zeolites transformed into their respective H-form by thermal
treatment at 600 °C with a heating rate of 1 °C/min.
3. Results and discussion
The zirconium loaded zeolites as well as the parent support mate-
rials investigated in this work are essentially microporous materials,
exhibiting textural properties typical for microporous materials. The
2
The preparation of supported ZrO on zeolites was carried out fol-
lowing a protocol previously reported by our research group [26].
According to this protocol, the preparation of supported metal oxide
nanoparticles takes place by grinding together the zeolite, used as
2
incorporation of ZrO on the different zeolites used as support leads to a
decrease in surface area ca. 30% as compared to their respective parent
zeolites (Table 1S). This is indicative of the different interactions that
take place during the milling process used for the preparation of the
catalysts such as dehydroxylation of the zeolite, the reaction between
the external silanol groups and the zirconium salt precursor, and,
support, and the appropriate amount of ZrO(NO
Zr salt precursor, to achieve a metal loading 10 wt. % at 350 r.p.m. for
0 min. The final catalyst was obtained by calcination at 400 °C for 4 h.
3 2 2
) xH O, employed as
1
The materials prepared in this way were: Zr10/HZSM-5(30), Zr10/
HZSM-5(50), Zr10/HBEA(75), where the numbers in brackets relate to
eventually, the partial blockage of the micropores in the zeolite by ZrO
particles.
2
SiO
2
/Al
2
O
3
molar ratio in the zeolites.
XRD diffractograms corresponding to the material Zr10/HBEA(75),
and its respective support are shown on Fig. 1 (XRD diffractograms for
the remaining materials are shown on Figs. S1, S2). The parent zeolites
used as catalytic support displayed the typical diffraction patterns of
these crystalline materials [27]. Such crystallinity of the zeolites em-
ployed as support slightly decreased upon the milling step used in the
preparation of the ZrO loaded materials, preserving the distinctives
2
diffraction lines characteristics for each zeolite. Additionally, the in-
corporation of zirconium was confirmed by the presence of diffraction
2.2. Materials characterisation
The textural properties of the materials prepared were evaluated by
N
2
adsorption/desorption measurement using a Micromeritics ASAP
000. Samples were outgassed at 130 °C for 24 h (P < 0.1 Pa) prior to
analysis. BET surface has been calculated in the interval P/Po = 0.05-
2
2
0.30, assuming that the nitrogen surface is 0.162 nm .
XRD diffractograms of the materials employed in this work have
2
lines corresponding to ZrO in its cubic phase. The particle size of zir-
been acquired in a Bruker model DISCOVER D8 diffractogram. The line
CuK scan in the range
conium oxide loaded on zeolites was estimated using the Scherrer
equation to be ca. 24.9 nm for Zr10%HBEA75. For the remaining Zr-
loaded materials, particle size evaluation as well as phase identification
was not possible due to the higher crystallinity of ZSM-5 zeolites, whose
diffraction lines overlap with the signal corresponding to the zirconium
phase as well as Zr content in the materials (Table 1).
The elemental composition of the samples was evaluated by EDS/
SEM, showing lower Zr loadings as compared to the theoretical ex-
pected (Table 1, typically below 5%). This fact could be explained by
the semi-quantitative character of the characterisation technique, to-
α
(λ = 1,5406 Å) has been used with a
1
0° < 2θ < 80° and a goniometer speed of 0.5°/min.
Elemental analysis and mapping of the elements in the synthesised
materials were performed using a scanning electron microscope JEOL
JSM 7800 F equipped with a X-max 150 microanalysis system, Si/Li
detector, with a detection range from boron to uranium and a resolu-
tion of 127 eV.
The surface acid properties were evaluated using a chromatographic
pulse method using pyridine (PY) and 2,6-dimethylpyridine (DMPY) as
bases, due to their selective adsorption over Brönsted and Lewis acid
sites and Brönsted acid sites, respectively. The acidity measurements
have been carried out at 300 °C. The pulses to the catalyst bed have
2
gether with the not uniform dispersion of ZrO nanoparticles over the
different zeolites, confirmed by mapping images for the different ele-
ments existing in the investigated samples (Fig. 2). As expected,
2
Cabanillas, Matilde
