Bifunctional SO4/ZrO2 catalysts for 5-hydroxymethylfufural (5-HMF) production from glucose

Article abstract of DOI:10.1039/c3cy00409k

The telescopic conversion of glucose to fructose and then 5-hydroxymethylfurfural (5-HMF), the latter a potential, bio-derived platform chemical feedstock, has been explored over a family of bifunctional sulfated zirconia catalysts possessing tuneable acid-base properties. Characterisation by acid-base titration, XPS, XRD and Raman reveal that submonolayer SO₄ coverages offer the ideal balance of basic and Lewis-Bronsted acid sites required to respectively isomerise glucose to fructose, and subsequently dehydrate fructose to 5-HMF. A constant acid site normalised turnover frequency is observed for fructose dehydration to 5-HMF, confirming a common Bronsted acid site is responsible for this transformation. The Royal Society of Chemistry.

Full text of DOI:10.1039/c3cy00409k

Volume 4 Number 2 February 2014 Pages 247–570
Catalysis
Science &
Technology
www.rsc.org/catalysis
ISSN 2044-4753
COVER ARTICLE
Wilson et al.
Bifunctional SO4/ZrO2 catalysts for 5-hydroxymethylfufural (5-HMF)
production from glucose

Catalysis
Science &
Technology
View Article Online
View Journal | View Issue
PAPER
Bifunctional SO4/ZrO2 catalysts for
5-hydroxymethylfufural (5-HMF) production from
glucose†
Cite this: Catal. Sci. Technol., 2014,
4, 333
Amin Osatiashtiani,ab Adam F. Lee,bc D. Robert Brown,d Juan A. Melero,e
e
ab
*
Gabriel Morales and Karen Wilson
The telescopic conversion of glucose to fructose and then 5-hydroxymethylfurfural (5-HMF), the latter a
potential, bio-derived platform chemical feedstock, has been explored over a family of bifunctional sul-
fated zirconia catalysts possessing tuneable acid–base properties. Characterisation by acid–base titration,
XPS, XRD and Raman reveal that submonolayer SO4 coverages offer the ideal balance of basic and
Lewis–Brønsted acid sites required to respectively isomerise glucose to fructose, and subsequently dehy-
drate fructose to 5-HMF. A constant acid site normalised turnover frequency is observed for fructose
dehydration to 5-HMF, confirming a common Brønsted acid site is responsible for this transformation.
Received 15th June 2013,
Accepted 5th August 2013
DOI: 10.1039/c3cy00409k
www.rsc.org/catalysis
Specifically, new methods are required to selectively deoxygen-
ate the highly functional molecules obtained from cellulose,
Introduction
Concern over dwindling fossil fuel reserves, and the impact of
CO2 emissions on climate change, is driving the quest for
alternative feedstocks to reduce dependence on non-renewable
sources of fuels and chemicals. Biomass offers the only renew-
able source of organic molecules for the manufacture of bulk,
fine and speciality chemicals necessary to secure the future
needs of society. To be sustainable, so called ‘second genera-
tion’ biomass feedstocks must be sourced from non-edible
components of crops (such as stems, leaves and husks), cellu-
lose from agricultural or forestry waste, or high yielding short
rotation non-food crops such as Switch grass or Willow which
require minimal cultivation. While such lignocellulosic mate-
rials are attractive feedstocks for both fuel and chemical pro-
duction via bio- or thermochemical platforms,1,2 significant
catalyst development is essential to improve the efficiency with
which biomass derived building blocks can be processed.
contrasting with historic selective oxygenation routes devel-
oped for petroleum feedstocks.
In 2004, the US DoE identified a range of sugar-derived
platform chemicals obtainable via chemical or biochemical
transformation of lignocellulosic biomass,3 subsequently
revisited by Bozell and Petersen in 2010.4 Furanic compo-
nents such as 5-hydroxymethylfurfural (5-HMF) and furfural
were identified as key chemical intermediates that can be
synthesised from sugars.5 Potential applications include the
production of linear alkanes of the molecular weight desired
for transformation into diesel and jet liquid fuels, with
5-HMF also a precursor to valuable chemical building blocks
such as levulinic acid, 2,5-furandicarboxylic acid (FDA), 2,5-
diformylfuran (DFF), dihydroxymethylfuran and 5-hydroxy-
4-keto-2-pentenoic acid for the synthesis of diverse polymers
and plastics.6 Conversion of C6 sugars to 5-HMF is of signifi-
cant current interest, with liquid mineral acids including
H2SO4, HCl and H3PO4 frequently employed to catalyse the
necessary dehydration. However, commercial implementation
of 5-HMF as a chemical intermediate is impeded by high pro-
duction costs.7 A heterogeneously catalysed route to directly
convert glucose into 5-HMF in aqueous media thus remains
highly sought after. Most research has focused on the more
facile conversion of fructose (as a model saccharide) to
5-HMF,8–18 to circumvent the formation of side products
such as oligosaccharides and humins19 commonly reported
during acid catalysed glucose conversion. Many studies have
also resorted to non-aqueous solvents in an effort to improve
overall 5-HMF yield. While high 5-HMF yields are reported
following fructose dehydration in dimethyl sulfoxide
a Present Address: European Bioenergy Research Institute, School of Engineering
and Applied Sciences, Aston University, Aston Triangle, Birmingham, B4 7ET, UK.
E-mail: k.wilson@aston.ac.uk
b School of Chemistry, Cardiff University, Main Building, Cardiff, CF10 3AT, UK.
E-mail: OsatiashtianiA@cardiff.ac.uk
c Present Address: Department of Chemistry, University of Warwick, Coventry,
CV4 7AL, UKSchool of Chemistry, Monash University, Victoria 3800, Australia.
E-mail: A.F.Lee@warwick.ac.uk
d School of Applied Sciences, University of Huddersfield, Huddersfield, HD1 3DH,
UK. E-mail: d.r.brown@hud.ac.uk; Tel: 01484 473397
e Department of Chemical & Environmental Technology, Universidad Rey
Juan Carlos, C/Tulipan. s/n. E-28933, Móstoles, Madrid, Spain. E-mail: juan.
melero@urjc.es; Fax: +34 91-488-74-68
† Electronic Supplementary Information (ESI) available: Porosimetry, tables of
product data from catalytic reactions. See DOI: 10.1039/c3cy00409k
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 333–342 | 333

View Article Online
Paper
Catalysis Science & Technology
(DMSO)16 due to the aprotic solvent inhibiting levulinic acid
and humin formation,20 product separation is problematic
due to the formation of toxic sulfur compounds during high
temperature DMSO distillation.21 Ionic liquids are an attrac-
tive processing solvent for carbohydrates; partnership with a
homogeneous metal halide catalyst offers 5-HMF yields from
glucose of ~70%, however separation and catalyst recovery
issues remain.22
strength of SZ can be readily tuned to direct selectivity in liq-
uid phase terpene isomerisation.43 We thus hypothesised
that judicious control over sulfur loading content may enable
predictable tuning of the one-pot conversion of glucose to
5-HMF, by optimising the relative surface coverage of sulfate
acid and ZrO2 base sites arising from the parent support.
Here we demonstrate that systematic control over the Lewis–
Brønsted acid and base properties of SZ enables the tele-
scopic isomerisation of glucose to fructose, and subsequent
fructose dehydration to 5-HMF in aqueous media, employing
a single bi-functional heterogeneous catalyst.
Since biomass pre-treatment, such as steam explosion23 or
enzymatic24 and chemical (acid or base)25,26 promoted cellu-
lose hydrolysis, will ultimately produce aqueous sugar
sources, routes to convert the cheaper and more abundant
glucose to 5-HMF in water remain a challenge. Operation in Experimental
biphasic systems (e.g. water/MIBK), which allows reactive
extraction of 5-HMF from the aqueous phase thereby limiting
Catalyst preparation
2−
A series of SZ catalysts with different SO4 loadings were pre-
pared by impregnation of 50 g Zr(OH)4 (MEL Chemicals – XZO
880/01) with 500 ml H2SO4(aq) of molarity 0.01–0.5 M. The
slurry was stirred for 5 h at ambient temperature, filtered and
dried at 80 °C overnight, and then calcined at 550 °C for 3 h.
Catalysts were stored in air and used without pre-treatment.
side reactions, is a promising approach to continuous 5-HMF
production,8,27 when catalysed by liquid8,28,29 or solid
acids.30–32
A tandem homogeneous Lewis/Brønsted acid
catalysed process, utilising AlCl3 and HCl in a biphasic sys-
tem of water/2-sec-butylphenol to respectively isomerise glu-
cose to fructose and then dehydrate fructose to 5-HMF,
recently achieved a 62% 5-HMF yield.33 Other homogeneous
metal halides,34,35 including Cr(III) Zn(II) and Sn(IV) and more
water tolerant lanthanide chloride,36 can also effect glucose
conversion, but confer lower HMF selectivity of ~45–48%. In
practice, the use of soluble catalysts, particularly AlCl3 (which
reacts violently with water), remains a concern due to the
associated toxic waste levels that could result from scaled-up
processes, and thus at odds with Green Chemistry principles.
The application of Lewis acidic Sn-β zeolite in conjunction
with aqueous HCl can convert glucose to 5-HMF in a biphasic
system at 180 °C with ~60% HMF selectivity, however the use
of corrosive HCl remains undesirable.37 A tandem reaction
using solid base hydrotalcites and solid acid resins in a sin-
gle reactor is most promising, although reaction was
conducted in N,N-dimethylformamide instead of water.38
Practical and efficient glucose to 5-HMF conversion thus
awaits improved bi-functional solid acid and base catalysts
which can operate in the aqueous phase. In this regard, prog-
ress has been unfortunately hampered by a lack of detailed
analysis including mass balances, and systematic studies cor-
relating catalyst acid–base properties with performance.
The amphoteric properties of zirconia make it an attrac-
tive catalytic material to employ in such a bi-functional pro-
cess.39,40 Indeed zirconia has been reported as a catalyst for
the isomerisation of glucose to fructose at 200 °C,41 while
sulfated zirconia (SZ) is also an attractive strong solid acid
for alcohol dehydration. Initial reports relating to the perfor-
mance of SZ in aqueous phase catalysis are somewhat disap-
pointing, reflecting instability under high temperature
hydrothermal conditions, most likely associated with dissolu-
tion of multilayer sulfate species present at the high S con-
tents employed.16,42 The potential for tuning the acid
strength in SO4/ZrO2 and thereby imparting bi-functionality
at low sulfate contents for glucose conversion has been
neglected to date. Our previous work showed that the acid
Catalyst characterisation
Surface area and pore size analysis was performed by N2
physisorption on a Quantasorb Nova 2000 instrument, after
sample out gassing at 120 °C for 2 h. Surface areas were calcu-
lated using the Brunauer–Emmett–Teller (BET) method over
the range P/P0 = 0.03–0.18, where a linear relationship was
maintained. Pore size distributions were calculated using the
Barrett–Joyner–Halenda (BJH) model applied to the desorp-
tion branch of the isotherm. Bulk S contents were determined
by EDX employing an Oxford Instruments EVO SEM and the
Oxford Instruments Inca software. X-ray Photoelectron Spec-
troscopy (XPS) measurements were performed using a Kratos
Axis HSi photoelectron spectrometer equipped with a charge
neutralizer and a Mg Kα X-ray source (hν = 1253.6 eV). Spectra
were recorded at normal emission using an analyzer pass
energy of 20 eV and X-ray power of 225 W. XRD patterns were
recorded on a Panalytical X'pert-Pro diffractometer fitted with
an X'celerator detector, using Cu Kα (1.54 Å) sources with a
nickel filter, calibrated against Si standards. Raman spectra
were obtained on a Renishaw Ramascope fitted with a 785
and 514 nm lasers. The spectra were recorded in the range of
0–1350 cm−1 using 514 nm source, 5× lens, 2 second exposure
time, 100 accumulation and 100% laser power. Diffuse Reflec-
tance Infra-red Fourier Transform (DRIFT) spectra were
obtained using a Nicolet Avatar 370 MCT with Smart Collector
accessory, mid/near infrared source and mercury cadmium
telluride (MCT-A) photon detector at −196 °C (liquid N2). Sam-
ples were diluted with KBr powder (10 wt% in KBr) for analy-
sis then loaded into an environmental cell and subjected to
additional drying under vacuum at 110 °C for 10 min prior to
measurements to remove moisture physisorbed during air
exposure. Ex situ pyridine adsorption was performed by expo-
sure of samples to pyridine vapour in a desiccator overnight.
Excess physisorbed pyridine was removed in a vacuum oven
334 | Catal. Sci. Technol., 2014, 4, 333–342
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
prior to sample loading in the environmental cell, with spec-
tra recorded at 25 °C in vacuo.
In some cases, due to the formation of humins and
unidentified compounds, a ‘relative selectivity’ is used, to
permit comparison of the reaction selectivity towards the fol-
lowing known and calibrated products: cellobiose, glucose,
fructose, 1,6-anhydroglucose, lactic acid, formic acid, acetic
acid, levulinic acid, HMF and furfural. Relative HMF selectiv-
ity is defined as:
Acid site loadings were measured via NH3 pulse chemi-
sorption on a Quantachrome ChemBET 3000 system inter-
faced to an MKS Minilab mass spectrometer (MS). Samples
were outgassed at 150 °C under flowing He (20 ml min−1) for
2 h prior to pulse titration at 100 °C. The same procedure was
employed to measure the base site loading, except that CO2
was used to titrate sites at 35 °C. Acid strength measurements
were made via flow adsorption calorimetry using a flow
through differential scanning calorimeter (Setaram DSC111)
connected to gas flow and switching systems. Gas flow rates
were controlled by automated mass flow controllers. The sam-
ple (5–60 mg) was held on a glass frit in a vertical silica glass
sample tube in the calorimeter. A steady 5 ml min−1 flow of
He was maintained across the sample for 4 h at 150 °C to
effect activation. A sequence of probe gas pulses (1% NH3 in
He) were delivered to the carrier gas stream from a 0.5 ml
sample loop using a two position Valco valve with an auto-
mated micro-electric actuator. Heat output associated with
interaction between NH3 and sample was detected by DSC,
and the concentration of NH3 in the gas flow downstream of
the DSC measured with a HPR 20 Hiden MS gas analyser via
a heated capillary at 175 °C. A pulse delay (30 min) was
employed to allow reversibly adsorbed NH3 to desorb back
into the pure He stream and/or redistribute on the sample,
and for baselines to stabilise. pH measurements on aqueous
catalyst suspensions were performed by adding 0.1 g of each
catalyst to 20 ml of deionized water and stirring at room tem-
perature. After 30 min the solution pH was measured using a
Jenway 3305 pH meter.
HMF relative selectivity % 


Yield of HMF
100
Sum of yields of all identified products
Yield, selectivity and relative selectivity of other products
are calculated on the same basis. Carbon balance is calcu-
lated based on moles of carbon in the identified products,
relative to moles of carbon atoms in the glucose converted.
(Moles of C in products)
Cout


100
Cin Moles of C in glucose converted
Results and discussion
Catalyst characterisation
Bifunctional SZ catalysts require surface sulfation of the
Zr(OH)4 precursor within the monolayer (ML) regime, and
concomitant retention of a high accessible surface area. The
impact of zirconium hydroxide impregnation by 0.01–0.5 M
H2SO4 was probed by XPS and EDX to determine the SO4 sat-
uration monolayer coverage. Fig. 1 shows that increasing the
concentration of the impregnating acid solution results in a
steep initial rise in both the surface and bulk S content,
which subsequently attain a plateau at ~5 and 3 wt% sulfur
respectively. The transition between these regimes occurs at
[H2SO4] > 0.25 M, indicative of a saturated sulfate mono-
layer. The surface S content is consistently higher than that
of the bulk, confirming localisation of SO4 species at the
Catalytic reactions
Initial kinetic studies of glucose and fructose conversion were
conducted on a Radleys Starfish carousel under stirred batch
conditions at 100 °C to facilitate detailed reaction profiling
and minimise side reactions. Reactions were performed
using 0.1 g glucose or fructose (Sigma Aldrich), 0.1 g SZ cata-
lyst, and 20 ml deionised water, with such dilute sugar solu-
tions selected deliberately to minimise side reactions of
products. Samples were withdrawn periodically and filtered
prior to analysis on an Agilent 1200 series HPLC equipped
with RI and diode array detectors, and a Hi-Plex H column
for analysis. A 5 mM aqueous solution of sulphuric acid was
used as the eluent phase, with a flow rate of 0.6 ml min−1
and 65 °C column temperature. Product yields were calcu-
lated from response factors determined from multi-point cali-
bration curves. Yields and selectivity were calculated on a
carbon basis as below:
Moles of carbon in HMF product
HMF yield % 
100


Moles of carbon as glucose at t  0
Yield of HMF
HMFyield % 
100


Glucose conversion
Fig. 1 Dependence of SZ surface and bulk sulfur content upon [H2SO4].
Catal. Sci. Technol., 2014, 4, 333–342 | 335
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Catalysis Science & Technology
Zr(OH)4 surface. Defining the sulfate saturated monolayer
point as 5 wt% S enables a sulfate calibration scale to be
constructed for all SZ materials as implemented in Table S1.†
SO4 surface densities were calculated for comparison with lit-
erature, and are in good agreement with those of Morterra,44
who determined a monolayer coverage of ~4 SO4 per nm−2.
NH3 titration and calorimetry revealed that increasing SO4
coverage enhanced both the acid site loading and strength
(Table 1) up to one monolayer.
Porosimetry (Fig. S1†) reveals all samples in the series
exhibit a type IV isotherm with hysteresis loops indicative of
bottle necked mesopores. Corresponding BET surface areas
increase with SO4 coverage up to the monolayer point
(Table 1), with a subsequent decrease suggesting some struc-
tural collapse for the highest loading, e.g. formation of amor-
phous zirconium sulfate. Such surface area enhancements
have been previously reported for SZ materials, wherein
sulfation is reported to inhibit bulk crystallisation of the par-
ent Zr(OH)4 during calcination.45 A shift in the hysteresis
loop from P/P0 = 0.6–0.8 to 0.4–0.6 with increasing S content
reflects a decrease in the mean mesopore diameter in the
BJH pore size distribution shown in Fig. S2,† which falls from
5 nm for the unsulfated calcined Zr(OH)4 to 3.5 nm for sam-
ples impregnated with 0.01–0.025 M H2SO4. This meso-
porosity likely arises from interparticle voids between sulfate
functionalised ZrO2 crystallites. Higher acid loadings induce
additional microporosity, which we attribute to contraction
of these interparticle voids as a result of more uniform crys-
tallite packing as supported by XRD (see later), with [H2SO4]
> 0.5 M eliminating this microporosity and suppressing
mesoporosity, consistent with bulk sulfation.
Fig. 2 S 2p XP spectra of sulfated Zr(OH)4 as a function SO4 coverage.
2−
The presence of multiple SO4 species is supported by
DRIFTS measurements shown in Fig. 3, which show the pro-
gressive evolution of surface sulfoxy modes with increasing
acid site loading. Vibrational bands are observed attributable
to νs (S–O) at 1010, νas (S–O) at 1130, νs (SO) at 1260 and νas
(SO) at 1362 cm−1, consistent with bidentate or tridentate
2− 46–48
SO4
,
which grow monotonically with sulfate coverage
up to 0.5 ML. The high νas (SO) frequency indicates a
highly covalent sulfate species, as reported by Morterra et al.
for dehydrated samples measured in vacuo.44,49 Peak broad-
ening and poorer spectral resolution at higher coverage is
attributable to the presence of multiple sulfate species as the
monolayer is saturated.14,15,50 The transition between isolated
and polynuclear sulfate species is in good agreement with
that reported by Bensitel51 and Morterra44 at SO4 loadings
The evolution of surface S species was probed via high
resolution S 2p XP spectra (Fig. 2), which reveal a character-
istic SO4 peak between 168–172 eV for all SZ samples. In the
low coverage regime (θSO < 0.5 ML), this sulfate species
4
shifts from 168.5 to 169 eV and broadens with increasing
surface sulfation. These concomitant changes indicate the
genesis of multiple, co-existing SO4 species as surface
sulfation progresses, likely associated with a change in coor-
dination geometry from bidentate to monodentate43 and
diminishing charge withdrawal from the zirconia due to
lateral interactions.
>1.5 nm−2
.
The formation of crystalline species in calcined SZ sam-
ples was subsequently explored using powder XRD, Fig. 4.
The sample with 0.1 ML SO4 exhibited reflections arising
from both monoclinic10 (2θ = 24.7°, 28.4°, 31.6°) and tetrago-
nal11 (2θ = 30.3°, 35.3°, 50.7°, 59.9°, 60.6° and 63.5°) ZrO2
phases. The tetragonal phase progressively increases with
Table 1 Physical properties of SZ as a function of SO4 coverage
SO4 coveragea /ML
Surface areab /m2 g−1
SO4 densityc /nm−2
Acid loadingd /mmol g−1
−ΔHads (NH3)e /kJ mol−1
0.0
0.1
0.2
0.3
0.4
0.6
0.7
0.8
1.0
1.1
93
143
142
169
175
189
175
203
194
118
0
0.07
0.13
0.14
0.17
0.17
0.27
0.29
0.30
0.37
0.29
—
87
—
95
0.10
0.35
0.40
0.55
1.10
1.79
2.09
2.83
5.66
—
102
115
115
115
115
a
Assuming 1 ML corresponds to 5 wt% surface S content. b From BET. c Using S content from EDX. d NH3 TPD. e NH3 adsorption calorimetry.
336 | Catal. Sci. Technol., 2014, 4, 333–342
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
Fig.
5
Raman spectra of impregnated SZ catalysts showing the
evolution of monoclinic (■) and tetragonal (□) phases with bulk S
content.
456 and 645 cm−1 with increasing θSO are attributable to the
4
tetragonal phase. High sulfate loadings degrade spectral reso-
lution, resulting in poorer discrimination between the mono-
clinic and tetragonal modes as the monolayer point is
reached, reflecting the surface sensitivity of Raman,54 while
new features emerge ~1000 cm−1 characteristic of surface sul-
phate species (Fig. S3†).
Fig. 3 DRIFTS spectra of impregnated sulphated zirconia as a function
of bulk S content (spectra recorded in situ at 200 °C in vacuo).
At low SO4 loadings a single peak at 997 cm−1 is observed,
with a second feature emerging at 1029 cm−1 which grows
continuously above 0.2 ML to form a broad feature upon
completion of the monolayer. This intense feature is attrib-
uted to the symmetric sulphate stretching mode, with peak-
splitting suggesting a change in sulphate geometry. This tran-
sition in the Raman spectra occurs at θSO between 0.2 to
4
0.4 ML, precisely the point at which the tetragonal phase of
ZrO2 becomes stabilised. Hence we tentatively assign the
997 cm−1 and 1029 cm−1 Raman features to SO4 coordinated
to monoclinic and tetragonal surface sites respectively.
Zirconia is amphoteric, with the potential to exhibit
Lewis basicity but also varying degrees of Lewis or Brønsted
acidity depending on the crystalline phase, with monoclinic
ZrO2 generated via calcination reported to exhibit predomi-
nantly Lewis acidity.55–57 The evolution of basic and Lewis–
Brønsted acidic properties for the SZ materials was probed
by CO2 and NH3 chemisorption and pyridine titration. The
inset to Fig. 6 shows representative DRIFT spectra for pyri-
dine adsorbed on submonolayer and monolayer SZ samples
which exhibit bands at 1450, 1470, 1610 cm−1 attributed to
pyridine bound to Lewis acid sites, while those at 1490,
1540, 1610 are 1635 cm−1 are characteristic of a pyridinium
ion coordinate to Brønsted sites. The unique Brønsted–
Lewis features at 1540/1450–1470 cm−1 were integrated to
quantify the variation in Brønsted : Lewis ratio, which
increases with both acid strength and SO4 coverage (Fig. 6)
and also correlates directly with the variation in tetragonal :
monoclinic zirconia ratio (determined by integrating the
Fig.
4 Powder XRD analysis of impregnated sulphated zirconia
catalysts showing the evolution of monoclinic (■) and tetragonal (□)
phases as the bulk S content increases.
surface coverage, becoming the dominant phase for 0.75 ML
SO42−. A loss of crystallinity observed at higher S contents, is
most likely due to the formation of an amorphous bulk
Zr(SO4)2 species.52
Raman spectroscopy allows the clear discrimination of
monoclinic and tetragonal phases of ZrO2 (Fig. 5), and is in
accordance with the powder XRD. Losses observed at 180,
307, 337, 381, 476 and 618 cm−1 for ZrO2 are assigned to the
monoclinic phase,53 while bands evolving at 148, 271, 320,
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 333–342 | 337

View Article Online
Paper
Catalysis Science & Technology
The base site density decreases monotonically with
increasing sulphate coverage, mirroring the fall in pH of
aqueous suspensions of the SZ samples from pH 5 to 3,
confirming that the balance of SZ acid–base character can be
precisely tuned within the submonolayer regime (θSO < 1).
4
The preceding holistic characterisation enables us to con-
struct
a model of the SZ system wherein a coverage-
dependent transition occurs from isolated SO4 species
chemisorbed on monoclinic ZrO2 → SO4 islands on tetrago-
nal and monoclinic ZrO2 → a SO4 monolayer bound to tetra-
gonal ZrO2.
Glucose conversion to 5-HMF
Glucose conversion to 5-HMF is proposed to initiate via a
Lewis acid or base catalysed isomerisation to fructose,
followed by a Brønsted acid catalysed dehydration as shown
in Scheme 1. To establish the validity of this hypothesis, the
kinetics of glucose versus fructose conversion to 5-HMF were
compared at 100 °C. This mild temperature was employed to
minimise competing degradation reactions to levulinic acid
or humins, and thus permit accurate rate data to be obtained
across the series of SZ materials.
Fig. 6 Correlation between acid strength of SZ catalysts determined
from calorimetry and evolution of Brønsted : Lewis ratio determined
from pyridine titration (inset) and tetragonal : monoclinic ratios
determined from Raman.
Fig. 8 shows the resulting variations in glucose and fruc-
tose conversion as a function of surface sulfate coverage. The
first striking observation is that whereas fructose conversion
latter's fingerprint Raman bands at 270 and 380 cm−1
respectively from Fig. 5). These observations confirm that
ZrO2 morphology and acidity can be readily tuned by
submonolayers of surface sulfate, consistent with previous
reports that ZrO2 crystallisation is dependent on surface sul-
fate density.44 Calculations also show that ZrO2 is also
basic,48 hence CO2 titrations were employed to map the
base site density as a function of sulfate loading. Fig. 7
shows that the calcined parent Zr(OH)4, and submonolayer
sulphated SZ materials, possess appreciable base site densi-
ties, albeit significantly lower than the corresponding acid
site loadings (determined via NH3 titration).
increases modestly with θSO , glucose conversion shows a
4
large decrease for coverages above 0.25 ML. Glucose and fruc-
tose conversions and corresponding relative selectivities to
the main dehydration products (5-HMF and furfural) are
Scheme 1 Conversion of glucose to fructose and 5-HMF.
Fig. 7 Titration of acid and base site loadings of SZ catalysts as a
function of S content.
Fig. 8 Effect of SO4 coverage on glucose and fructose conversion in
water after 6 h reaction at 100 °C.
338 | Catal. Sci. Technol., 2014, 4, 333–342
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
Table 2 Conversion and product selectivity following glucose and fructose dehydration over SO4/ZrO2 at 100 °C
SO4 coverage/ML
0
Substrate
Conversiona /%
HMFa selectivity/%
Fructosea selectivity/%
Glucosea selectivity/%
Fructose
Glucose
Fructose
Glucose
Fructose
Glucose
Fructose
Glucose
Fructose
Glucose
Fructose
Glucose
Fructose
Glucose
Fructose
Glucose
Fructose
Glucose
Fructose
Glucose
14.7
20.9
16.2
22.7
18.6
23.3 (42)
19.1
20.9
18.2
17.8
20.8
7.5
19.8
5.4
25.4
7.2
23.3
11.6 (37)
16.6
5.5
11.7
1.9
13.3
3.0
17.9
3.7 (11)
19.9
4.8
19.3
4.6
29.9
10.5
23.8
9.9
32.6
10.0
26.6
9.4 (18.5)
28.2
7.7
—
84.2
—
81.3
—
81.0 (71)
—
78.2
—
79.8
—
63.9
—
65.3
—
25.0
—
26.4
—
17.8
—
16.2
—
20.0
—
8.1
—
7.4
—
8.1
—
0.1
0.2
0.3
0.4
0.6
0.7
0.8
1.0
1.1
77.2
—
70.2 (41)
—
8.4
—
9.0
—
81.5
a
After 6 h reaction at 100 °C; remaining products are other sugar isomers and furfural (see Table S2). Values in brackets for 6 h reaction at 120 °C.
summarised in Table 2. While we note that 5-HMF yields are
lower than those obtained in biphasic systems and ionic liq-
uids, where 5-HMF yields as high as 60–70% are observed
from glucose, it must be noted that these typically operate
under higher temperature conditions of 120–200 °C and use
homogeneous catalysts.5 Indeed we find here that raising the
reaction temperature to 120 °C almost doubles our conver-
sion and triples our HMF selectivity for the most active cata-
lyst. These trends can be rationalised in terms of the change
in acid–base character with sulfur loading by comparing the
associated product yields (Fig. 9). Glucose ↔ fructose iso-
merisation is a reversible reaction, catalysed by either basic
or Lewis acid sites.
therefore acid strength, consistent with a common active acid
site. SZ reactivity is thus predominantly controlled by the bal-
ance of their acid : base character; Fig. 10 shows a volcano
The high glucose conversion observed for pure ZrO2 and
SZ catalysts possessing low SO4 coverages correlates with high
fructose yields (Fig. 9a), and is thus a reflection of the corre-
spondingly significant Lewis acid–base properties of these
materials comprising predominantly monoclinic zirconia.
In contrast, high SO4 coverages (i.e. Brønsted acid site
densities) suppress glucose isomerisation to fructose in
favour of HMF production associated with enhanced dehy-
dration of the fructose intermediate. Hence loss of basicity
and Lewis acidity upon zirconia sulfation switches off glucose
↔ fructose isomerisation but promotes the Brønsted acid
catalysed fructose → 5-HMF pathway, in perfect agreement
with the proposal in Scheme 1. The requirement for Brønsted
acid character to produce 5-HMF is confirmed by inspecting
the yield of products obtained from fructose as a substrate
(Fig. 9b). Note that the resulting 5-HMF yields are three times
higher starting from fructose versus glucose, indicating that
glucose isomerisation to fructose is the rate-determining step
in 5-HMF formation. TOFs for 5-HMF formation from fruc-
tose, normalised to the surface sulfate density, are indepen-
dent of the Brønsted acid site loading/sulfur coverage and
Fig. 9 Yields of a) fructose and 5-HMF during SZ catalysed glucose
isomerisation and dehydration and b) glucose and 5-HMF during SZ
catalysed fructose isomerisation after 6 h reaction at 100 °C.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 333–342 | 339

View Article Online
Paper
Catalysis Science & Technology
surface of base monoclinic ZrO2 initiates this transformation
via proton abstraction to form the enol (akin to that pro-
posed over sodium aluminate60) which undergoes subse-
quent hydrogen transfer to form fructose. Spillover onto
neighbouring Brønsted acid sulfate moieties then catalyses
the stepwise dehydration of fructose to 5-HMF61 as illus-
trated in Scheme 2. Lewis acid sites also have the potential
to initiate glucose isomerisation via an intramolecular
hydride shift.62 However, the extent to which Lewis acid
routes are able to participate in aqueous phase reactions
remains contentious due to their likelihood of hydration to
Brønsted counterparts, although it has been suggested that
Lewis acidity may be retained at defect sites44 or when
hydrophobic supports such as Sn-β are employed63 wherein
a Meerwein–Ponndorf–Verley mechanism has been postu-
lated.64,65 As Fig. 10 highlights, careful tuning of the degree
of zirconia surface sulfation enables the successful genesis
of bi-functional catalysts possessing dual solid acid–base
character which facilitate the telescopic conversion of glu-
cose to 5-HMF under mild reaction conditions. To investi-
gate the kinetics of the glucose → fructose → 5-HMF
pathway dilute operating conditions using feed with only a
0.5 wt% (0.028 M) glucose concentration to minimize side
reactions of 5-HMF. We note this is significantly lower than
that reported in the literature, where typical concentrations
range from 10 wt%,28 12 wt%66 to 30 wt%.67 However, it
should be noted that for homogeneous processes, the con-
version of glucose is reported to be exhibit good first order
behaviour and is independent of concentration over the
range 0.01–0.16 M.68 Likewise, the kinetics of HMF forma-
tion from fructose are not reported to show a strong
concentration dependence when operating at a fructose con-
centration ~0.7 M.69 We thus expect that any changes in
reactivity at higher initial glucose concentration would result
in an increased probability of side reactions when more con-
centrated HMF containing mixtures are produced. In light of
Fig. 10 Turnover frequencies for HMF formation from glucose at
100 °C as a function of acid : base ratio of SZ catalysts.
dependence for 5-HMF productivity, reflecting the require-
ment for both Lewis acidic–basic monoclinic ZrO2 (to catalyse
glucose isomerisation) and Brønsted surface sulfate species
(to catalyse fructose dehydration to 5-HMF) in optimal cata-
lyst formulations. Indeed the overall rate of 5-HMF produc-
tion from glucose closely mirrors the mathematical product
of the rate of glucose isomerisation with the rate of 5-HMF
formed from direct fructose dehydration (Fig. S4†). To verify
the stability of the optimum catalyst, a recycle test was
performed following re-calcination of the spent catalyst at
550 °C, which revealed only a small decrease in absolute con-
version was observed after 6 h reaction (Fig. S5†).
We can hence advance a bifunctional catalytic surface
mechanism for glucose conversion to 5-HMF. The first step
of this is most likely the Lewis base catalysed transformation
of glucose into an enol intermediate (Scheme 2),58,59 with
subsequent protonation of the resulting CC yielding either
fructose or mannose. We propose that O2− sites on the
Scheme 2 Bi-functional surface catalysed mechanism for a) isomerisation of glucose to fructose over basic O2− sites of monoclinic ZrO2 (Lewis
acidic Zr4+ may help stabilise the enolate intermediate) and b) dehydration of fructose to 5-HMF over Brønsted acid sites present in submonolayer
SO4/ZrO2 catalysts.
340 | Catal. Sci. Technol., 2014, 4, 333–342
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
this working at particularly high glucose concentrations may
not be advisable and suggest operation in a biphasic system
under more dilute glucose concentrations conditions could
improve HMF selectivity under continuous processing condi-
tions. We are presently investigating this and the synthesis
of high area, nanoporous SZ analogues to afford even greater
control over ZrO2 crystallisation, and opportunities to tune
catalyst hydrophobicity.
Notes and references
1 M. J. Climent, A. Corma and S. Iborra, Green Chem.,
2011, 13, 520–540.
2 P. Gallezot, Chem. Soc. Rev., 2012, 41, 1538–1558.
3 T. Werpy and G. Petersen, Top Added Chemicals from
Biomass Volume 1—Results of Screening for Potential
Candidates from Sugars and Synthesis Gas, U. S.
Department of Energy Report, 2004.
4 J. J. Bozell and G. R. Petersen, Green Chem., 2010, 12,
539–554.
Conclusions
5 R.-J. van Putten, J. C. van der Waal, E. de Jong,
C. B. Rasrendra, H. J. Heeres and J. G. de Vries, Chem. Rev.,
2013, 113, 1499–1597.
6 A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107,
2411–2502.
7 B. F. M. Kuster, Starch/Staerke, 1990, 42, 314–321.
8 Y. Roman-Leshkov, J. N. Chheda and J. A. Dumesic, Science,
2006, 312, 1933–1937.
9 Y. Zhang, J. Wang, J. Ren, X. Liu, X. Li, Y. Xia, G. Lu and
Y. Wang, Catal. Sci. Technol., 2012, 2, 2485–2491.
10 J. Wang, W. Xu, J. Ren, X. Liu, G. Lu and Y. Wang, Green
Chem., 2011, 13, 2678–2681.
The impact of surface sulfation upon the physico-chemical
properties of calcined Zr(OH)4 has been systematically inves-
tigated by bulk and surface spectroscopies and chemical
probes. The unsulfated precursor forms predominantly
monoclinic zirconia possessing mixed Lewis acid and base
surface sites and is effective for glucose isomerisation to fruc-
tose but poor towards fructose dehydration to 5-HMF. Dilute
sulfuric acid pre-treatment of Zr(OH)4 and subsequent calci-
nation results exclusively in the formation of polydentate sur-
2−
face SO4
species, and concomitant stabilisation of
tetragonal ZrO2, conferring significant Brønsted acidity and
corresponding enhanced 5-HMF production from either glu-
11 Q. Cao, X. Guo, J. Guan, X. Mu and D. Zhang, Appl. Catal., A,
2011, 403, 98–103.
cose or fructose. Higher degrees of surface sulfation (θSO
>
4
0.25 ML), and attendant loss of surface basicity from exposed
zirconia, progressively switches off glucose ↔ fructose iso-
merisation, while continuing to promote fructose → 5-HMF.
Saturated sulfate monolayers present a distribution of mono
and polynuclear sulfate species chemisorbed over tetragonal
crystalline zirconia and/or amorphous zirconium sulfate, and
the resulting materials (which exhibit almost entirely
Brønsted acid character) are the least efficient for 5-HMF syn-
thesis from glucose. Submonolayer sulfate coverages of
approximately 0.3 ML afford the optimal mix of Lewis base
sites arising from accessible ZrO2, and co-existing Brønsted
acid sites arising from mono- or bidentate sulfate, required
for the tandem isomerisation of glucose to fructose and the
latter's subsequent dehydration to 5-HMF. The design of such
bi-functional catalysts capable of effecting one-pot telescopic
syntheses in aqueous media will become increasingly critical
to achieve atom-economical, selective transformations of bio-
derived molecules for sustainable chemicals and fuels.
12 Q. Zhao, L. Wang, S. Zhao, X. Wang and S. Wang, Fuel,
2011, 90, 2289–2293.
13 C. Carlini, P. Patrono, A. M. R. Galletti and G. Sbrana, Appl.
Catal., A, 2004, 275, 111–118.
14 A. J. Crisci, M. H. Tucker, J. A. Dumesic and S. L. Scott, Top.
Catal., 2010, 53, 1185–1192.
15 P. Carniti, A. Gervasini and M. Marzo, Catal. Commun.,
2011, 12, 1122–1126.
16 X. H. Qi, M. Watanabe, T. M. Aida and R. L. Smith, Catal.
Commun., 2009, 10, 1771–1775.
17 P. Carniti, A. Gervasini, S. Biella and A. Auroux, Catal.
Today, 2006, 118, 373–378.
18 X. H. Qi, H. X. Guo and L. Y. Li, Ind. Eng. Chem. Res.,
2011, 50, 7985–7989.
19 H. E. Van Dam, A. P. G. Kieboom and H. Van Bekkum,
Starch/Staerke, 1986, 38, 95–101.
20 Y. Nakamura and S. Morikawa, Bull. Chem. Soc. Jpn.,
1980, 53, 3705–3706.
21 C. Moreau, M. N. Belgacem and A. Gandini, Top. Catal.,
2004, 27, 11–30.
22 H. B. Zhao, J. E. Holladay, H. Brown and Z. C. Zhang,
Science, 2007, 316, 1597–1600.
23 W. G. Glasser and R. S. Wright, Biomass Bioenergy, 1998, 14,
219–235.
24 P. Alvira, E. Tomás-Pejó, M. Ballesteros and M. J. Negro,
Bioresour. Technol., 2010, 101, 4851–4861.
25 W. S. Mok, M. J. Antal and G. Varhegyi, Ind. Eng. Chem. Res.,
1992, 31, 94–100.
26 R. Rinaldi and F. Schüth, ChemSusChem, 2009, 2, 1096–1107.
27 C. V. McNeff, D. T. Nowlan, L. C. McNeff, B. W. Yan and
R. L. Fedie, Appl. Catal., A, 2010, 384, 65–69.
Acknowledgements
We thank the EPSRC for funding under (EP/K000616/1) and
MEL Chemicals for the supply of Zr(OH)4. KW acknowledges
The Royal Society for the award of an Industry Fellowship,
and AFL thanks the EPSRC for the award of a Leadership
Fellowship (EP/G007594/2). Financial support from the
Spanish Ministry of Economy and Competitiveness through
the project CTQ2011-28216-C02-01 is also kindly acknowl-
edged. This Article is published in celebration of the 50th
Anniversary of the opening of the Chemistry Department at
the University of York.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 333–342 | 341

View Article Online
Paper
Catalysis Science & Technology
28 J. N. Chheda, Y. Roman-Leshkov and J. A. Dumesic, Green
Chem., 2007, 9, 342–350.
48 F. Haase and J. Sauer, J. Am. Chem. Soc., 1998, 120,
13503–13512.
29 Y. J. Pagan-Torres, T. Wang, J. M. R. Gallo, B. H. Shanks and
J. A. Dumesic, ACS Catal., 2012, 2, 930–934.
49 C. Morterra, G. Cerrato, F. Pinna, M. Signoretto and
G. Strukul, J. Catal., 1994, 149, 181.
30 V. V. Ordomsky, J. van der Schaaf, J. C. Schouten and
T. A. Nijhuis, J. Catal., 2012, 287, 68–75.
50 T. Yamaguchi, T. Jin and K. Tanabe, J. Phys. Chem. B,
1986, 90, 3148–3152.
31 C. Moreau, R. Durand, S. Razigade, J. Duhamet, P. Faugeras,
P. Rivalier, P. Ros and G. Avignon, Appl. Catal., A, 1996, 145,
211–224.
32 P. Rivalier, J. Duhamet, C. Moreau and R. Durand, Catal.
Today, 1995, 24, 165–171.
33 Y. J. Pagan-Torres, T. F. Wang, J. M. R. Gallo, B. H. Shanks
and J. A. Dumesic, ACS Catal., 2012, 2, 930–934.
34 V. Choudhary, S. H. Mushrif, C. Ho, A. Anderko,
V. Nikolakis, N. S. Marinkovic, A. I. Frenkel, S. I. Sandler
and D. G. Vlachos, J. Am. Chem. Soc., 2013, 135, 3997–4006.
35 C. B. Rasrendra, I. G. B. N. Makertihartha, S. Adisasmito and
H. J. Heeres, Top. Catal., 2010, 53, 1241–1247.
36 T. F. Wang, Y. J. Pagan-Torres, E. J. Combs, J. A. Dumesic
and B. H. Shanks, Top. Catal., 2012, 55, 657–662.
37 E. Nikolla, Y. Roman-Leshkov, M. Moliner and M. E. Davis,
ACS Catal., 2011, 1, 408–410.
51 M. Bensitel, O. Saur, J. C. Lavalley and B. A. Morrow, Mater.
Chem. Phys., 1988, 19, 147–156.
52 J. R. Sohn, T. D. Kwon and S. B. Kim, J. Korean Chem. Soc.,
2001, 22, 1309–1315.
53 D. I. Torres and J. Llopis, Superlattices Microstruct., 2009, 45,
482–488.
54 M. Li, Z. Feng, G. Xiong, P. Ying, Q. Xin and C. Li, J. Phys.
Chem. B, 2001, 105, 8107–8111.
55 V. Bolis, C. Morterra, M. Volante, L. Orio and B. Fubini,
Langmuir, 1990, 6, 695–701.
56 V. Bolis, G. Cerrato, G. Magnacca and C. Morterra,
Thermochim. Acta, 1998, 312, 63–77.
57 W. Hertl, Langmuir, 1989, 5, 96–100.
58 D. W. Harris and M. S. Feather, J. Am. Chem. Soc., 1975, 97,
178–181.
59 H. S. Isbell, H. L. Frush, C. W. R. Wade and C. E. Hunter,
Carbohydr. Res., 1969, 9, 163–175.
60 A. J. Shaw Iii and G. T. Tsao, Carbohydr. Res., 1978, 60,
327–325.
61 M. J. Antal Jr., W. S. L. Mok and G. N. Richards, Carbohydr.
Res., 1990, 199, 91–109.
38 A. Takagaki, M. Ohara, S. Nishimura and K. Ebitan, Chem.
Commun., 2009, 6276–6278.
39 M. A. Aramendıa, V. Boráu, C. Jiménez, J. M. Marinas,
A. Marinas, A. Porras and F. J. Urbano, J. Catal., 1999, 183,
240–250.
40 K. Tanabe and T. Yamaguchi, Catal. Today, 1994, 20,
185–197.
41 X. Qi, M. Watanabe, T. M. Aida and R. L. Smith Jr., Catal.
Commun., 2008, 9, 2244–2249.
62 Y. Román-Leshkov, M. Moliner, J. A. Labinger and
M. E. Davis, Angew. Chem., Int. Ed., 2010, 49, 8954–8957.
63 M. Moliner, Y. Román-Leshkov and M. E. Davis, Proc. Natl.
Acad. Sci. U. S. A., 2010, 107, 6164–6168.
42 I. J. Dijs, J. W. Geus and L. W. Jenneskens, J. Phys. Chem. B,
2003, 107, 13403–13413.
43 M. A. Ecormier, K. Wilson and A. F. Lee, J. Catal., 2003, 215,
57–65.
64 J. F. Miñambres, M. A. Aramendía, A. Marinas, J. M. Marinas
and F. J. Urbano, J. Mol. Catal. A: Chem., 2011, 338, 121–129.
65 F. J. Urbano, M. A. Aramendía, A. Marinas and
J. M. Marinas, J. Catal., 2009, 268, 79–88.
44 C. Morterra, G. Cerrato, C. Emanuel and V. Bolis, J. Catal.,
1993, 142, 349–367.
45 G. D. Yadav and J. J. Nair, Microporous Mesoporous Mater.,
1999, 33, 1–48.
46 K. Nakamoto, J. Fujita, S. Tanaka and M. Kobayashi, J. Am.
Chem. Soc., 1957, 79, 4904–4908.
47 S. J. Hug, J. Colloid Interface Sci., 1997, 188, 415–422.
66 K. Lourvanij and G. L. Rorrer, Ind. Eng. Chem. Res., 1993, 32, 11–19.
67 A. Ranoux, K. Djanashvili, I. W. C. E. Arends and
U. Hanefeld, ACS Catal., 2013, 3, 760–763.
68 B. M. Kabyemela, T. Adschiri, R. M. Malaluan and K. Arai,
Ind. Eng. Chem. Res., 1997, 36, 1552–1558.
69 G. R. Akien, L. Qi and I. T. Horvath, Chem. Commun.,
2012, 48, 5850–5852.
342 | Catal. Sci. Technol., 2014, 4, 333–342
This journal is © The Royal Society of Chemistry 2014