CHEMCATCHEM
FULL PAPERS
ered as “electron-rich,” which create O2 vacancies at oxide sur-
faces. These O2 vacancies created at oxide surfaces interacted
strongly with the Au nanostructure, which resulted in a signifi-
cant rearrangement in the electronic structure of Au as well as
in covalent bonding between the Au nanoparticles and the
defect oxide support. The electrons that are stored in these
chemical bonds led to the formation of active Au atoms in the
vicinity of the Au–oxide perimeter interface and contributed to
the catalytic process by providing an additional adsorption site
for the reactant owing to the effective nucleophilic attack of
C=O substrate bonds by activated O2 (O*) at the Au–oxide pe-
rimeter interface.[18–28]
oxidation of the aldehyde group (ÀCHO) of glucose resulted in
the formation of gluconic acid, and the oxidation of both the
primary and the secondary alcohol to yield ketones is sup-
pressed if the catalytic system used in the reaction is highly se-
lective and active.
The support properties can tune the interaction with the
metal particles deposited on or in it. They can modify both the
electronic and structural properties of the catalyst as well as
provides different anchoring sites for the reactants, which act
as active and sometimes reactive surfaces if strong metal–sup-
port interaction effects are induced. Our results and most re-
ports by various researchers show that in addition to acid
strength, density and glucan sorption affinity, the textural char-
acteristics of the support and the electronic state of the nano-
particles as a result of electron transfer from the support to
the Au nanoparticles during catalyst reductions under a H2
flow at high temperatures critically determine the catalyst suc-
cess.
Furthermore, critical AuÀOÀTi bonds that coupled the Au
nanoparticles to the support were expected to be formed in
the metal–support interaction. The AuÀO bond was polarized
to give Au a partial positive charge, and this interaction led to
unique dual catalytic active sites for the adsorption and activa-
tion of O2.[29] In an equivalent view of the detailed mechanism
of O2 adsorption and activation, Green et al. reported a back
donation of an electron density phenomenon that creates
a unique Au–Ti site at the Au/TiO2 interface, which is critical in
the activation of O2 because this unique Au–Ti site at the pe-
rimeter interface of the metal–support interaction allowed for
electron transfer from Au to Ti and the subsequent electron
transfer to 2p* antibonding states of O2, aiding in OÀO bond
activation. In addition, the adsorption of O2 thus occurred via
diÀs bonding to Au and Ti to form an AuÀOÀOÀTi state. The
basic O/Au and O/Ti species residing at the perimeter interface
of Au–Ti extracted electron density from the substrate and as
such readily attacked bound C=O and C=C-containing species
and activate CÀH and OÀH bonds, which thus catalyzed
a range of partial and full oxidative reactions.[30]
This hypothesis was supported by DFT studies[28,31–34] that
postulated that O2 dissociation was sensitive to the arrange-
ment of the Au surface atoms and the most active sites for O2
dissociation are found at the metal–support interface and not
at the Au particle surfaces. It was thus reasonable to state that
O2 adsorbed on the edges, at the metal–support interface, and
then migrated or diffused to the surfaces of Au particles, and
as such, oxidation reactions occurred at the surfaces of Au
nanoparticles because of the presence of diffused O*, which
could initiate nucleophilic attacks on biomass substrates at the
Au particle surfaces, with the subsequent reaction leading to
the corresponding gluconic acid. Because theoretical studies
reported in the literature have found that the active site for O2
activation is the perimeter interface between Au nanoparticles
and the TiO2 support, a strong interaction should therefore be
indispensable to afford high catalytic activity for cellobiose oxi-
dation.[35]
Reaction parameter optimization for cellobiose oxidation
Further kinetic study was performed over the Au/TiO2 catalyst
to gain more insights into the reaction conditions to efficiently
convert cellobiose to gluconic acid. The effect of the reaction
time on cellobiose conversion and the distribution of oxidative
reaction products is shown in Figure 8a. The pressure in the
reactor was set at 0.5 MPa. Cellobiose conversion increases
with the increase in reaction time. The conversion increases
sharply at the initial stage, reaching 93% in 2 h, followed by
a gradual increase slowly, and reaching 100% after 12 h. The
selectivity of gluconic acid did not follow the same trend. Glu-
conic acid selectivity is the highest at 2 h and decreases signifi-
cantly with the increase in reaction time from 2 to 12 h; de-
tailed analysis reveals that gluconic acid further converted to
other undesirable products such as lower carbon polyols at
a longer reaction time.[34] The best gluconic acid selectivity of
73.7% was obtained at a reaction time of 2 h and an O2 pres-
sure of 0.5 MPa, which is higher than that reported previously
by An et al. (57%) and Wang et al. (63%). The difference can
be attributed to the different catalyst preparation method (im-
pregnation) and lower reduction temperature (3008C) used by
them.[8,35]
The effect of O2 pressure on cellobiose oxidation was stud-
ied by varying the pressure from 0.5 to 2.5 MPa at 1458C (Fig-
ure 8b). The conversion of cellobiose was constant regardless
of the O2 pressure. The increase in O2 pressure affects the dis-
tribution of reaction products to a large extent. At an O2 pres-
sure of 0.5 MPa, a gluconic acid selectivity of 73.7% with an
appreciable amount of glycolic acid (25%) was observed. The
formation of ethylene glycol and glycolic acid with a subse-
quent decrease in the selectivity of gluconic acid with the in-
crease in the O2 pressure from 1.5 to 2.5 MPa indicates that
the increase in the amount of dissolved O2 directly affects the
distribution of reaction products. This phenomenon was con-
sistent with the results reported by An et al., in which a de-
creasing trend in gluconic acid selectivity with an increase in
the O2 pressure from 0.5 to 1.0 MPa was observed, which fa-
The increased metal–support interaction that consequently
constituted active sites for O2 adsorption and activation readily
reacts with bound hydrocarbon intermediates for the oxidation
of C=O bonds contained in most biomass-derived compounds
and several carbonyl and aldehyde-containing compounds.[28]
Glucose, the main reaction intermediate that was oxidized to
yield gluconic acid, contains an aldehyde group (1C), a primary
4
alcohol group (6C), and a secondary alcohol group (2CÀ C). The
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2105 – 2114 2110