Full Papers
which is the logical consequence of byproduct formation and
leads to the dramatic increase of defect in the molar balances
(Figure 3a).
In opposition to what was observed in the base-free reac-
tion, the results in Figure 3b, c, and d indicate that, with the
addition of the base NaHCO3, a close match was obtained
within the entire range of time between the experimental and
estimated concentrations versus time data. This shows that the
model described in Scheme 1 represents a good approxima-
tion for the conversion of HMF into FDCA. The mechanism of
oxidation of aldehyde to carboxylic acid involves a base-cata-
lyzed hydration of aldehyde to geminal diol. According to the
calculated kinetic constants, and as expected from this mecha-
nism, increasing amounts of base not only neutralize the car-
boxylic functions but also accelerate the oxidation rates of the
aldehyde functions in the intermediates to the fully oxidized
FDCA (k4 and particularly k5). The model provides higher values
for k2 than for k1, which would suggest the predominant for-
mation of the DFF intermediate. However, one must consider
that the aldehyde function reacts very rapidly in the presence
of the base to yield FFCA and the maximum amount of DFF
formed is very low. This is consistent with the high reactivity of
DFF demonstrated by Verdeguer et al.,[15] who measured a turn-
over frequency of 48 hꢀ1 for DFF compared with 4 hꢀ1 for
HMF. In the presence of 4 equivalents of NaHCO3, the constant
k5, which determines the rate of transformation of FFCA to
FDCA, exhibits a four-times higher value than that with the
use of 1 equivalent of NaHCO3.
Figure 5. Evolution of the concentrations of HMF and the products DFF,
HMFCA, FFCA, and FDCA as a function of time (experimental data: symbols;
model estimates: lines) over 3.6% Pt/TiO2 in the presence of Na2CO3. Reac-
tion conditions: T=1008C, P=40 bar air, 0.1m HMF, HMF-to-Pt molar
^
ratio=100, Na2CO3/HMF=2. : HMF; ~: DFF; &: HMFCA; *: FFCA;
FDCA; +: molar balance; ꢂ: TOC; g: pH value.
:
*
complete over the entire course of the reaction. This indicates
that some waste products are formed under these basic condi-
tions that are not detected by HPLC; however, they are soluble
and not adsorbed on the solid catalyst. As mentioned previ-
ously (Scheme 2), under strong basic conditions (which seems
to be the case with 2 equivalents of Na2CO3), HMF degradation
occurs through the Canizzaro reaction disproportionation to
form HMFCA. Furthermore, under too basic conditions, HMFCA
is also not stable and degradation of HMFCA can take place. In
addition, the absence of DFF might be due to the basic
medium favoring the Canizzaro reaction of DFF. These results
confirm that a moderately basic reaction medium is preferable
for the Pt/TiO2 catalyst. The substrate HMF and the first inter-
mediates formed seem to be more sensitive to the basicity of
the reaction medium in the presence of Pt/TiO2 than Pt/C.
The kinetic constants in the presence of 2 equivalents of
Na2CO3 differ from those obtained with 4 equivalents of
NaHCO3 (Table 2). The oxidation reactions of the aldehyde
functions in the reactant and in the intermediates were drasti-
cally accelerated, as demonstrated by the larger values of the
k1, k4, and k5 constants, which were multiplied by around
a threefold factor. In particular, there was a significant accelera-
tion in the final oxidation of FFCA to FDCA. In contrast, the k2
and k3 constants that correspond to the oxidation of an alco-
hol function increased little. The higher performance of the
Effect of using Na2CO3 base for Pt/TiO2
As previously studied with a Pt/C catalyst,[17] the influence of
2ꢀ
the nature of the carbonate base (CO3 versus HCO3ꢀ, pKb =
3.65 and 7.65, respectively) was studied over the 3.6% Pt/TiO2
catalyst. The substitution of 4 equivalents of NaHCO3 by
2 equivalents of Na2CO3 improved significantly all oxidation
rates, as shown in Figure 5. The reaction was completed within
8 h with a final yield of FDCA of 91%. The effect was as shown
previously in the presence of Pt/C; that is, the addition of
2 equivalents of carbonate with respect to HMF increased sig-
nificantly the overall oxidation rate compared to the use of
4 equivalents of bicarbonate. These rate increases are essential-
ly due to the pH value, which remains at a value of more than
8 throughout the oxidation reaction.
2ꢀ
As compared with the reaction with 4 equivalents of
NaHCO3, it was also noted that the yield of the intermediate
oxidation product HMFCA was significantly increased at the ini-
tial reaction stage in the presence of 2 equivalents of Na2CO3.
catalyst with the addition of 2 equivalents of the CO3 base
ꢀ
compared with 4 equivalents of HCO3 base can be explained
by the strength of these bases.[40]
Analysis of the final reaction mixtures by inductively coupled
plasma optical emission spectroscopy (ICP-OES) showed that
a small amount of the platinum metal was leached (up to 3%).
HMFCA was formed in
a maximum concentration of
30 mmolLꢀ1 after 30 min. The HMFCA and FFCA maximum
concentrations were formed in almost the same amount. This
is different to the reactions performed in the presence of
NaHCO3, in which both HMFCA and DFF were observed in
small amounts. Also, no intermediate product DFF was detect-
ed during the reaction with Na2CO3 as the base. Moreover, the
HPLC and TOC balances did not fit. A defect in balance based
on HPLC analysis was noted, whereas the TOC balance was
Effect of the nature of the oxide support
The supports ZrO2, Y2O3ꢀZrO2, and La2O3ꢀZrO2 were used to
evaluate if the basic sites introduced in ZrO2 could cooperate
with the homogeneous base and allow the amount of added
base to be reduced. Catalytic experiments of the zirconia-
ChemSusChem 0000, 00, 0 – 0
7
ꢁ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ