L. He et al.
Molecular Catalysis 508 (2021) 111584
2.6. Catalyst stability
Catalyst stability is an important factor to assess its catalytic per-
formance in a reaction. However, in this study, all of these heteroge-
neous catalysts were unexpectedly dissolved in the solution after the
reaction, although some of them exhibited a higher operating temper-
ature (Amberlyst-70, max. 190◦C, Nafion, max. 350◦C). The change in
the catalytic activity of the catalyst was studied by repeating the ex-
periments of Amberlyst-70. After each reaction, the solution was
collected and evaporated in a vacuum. The residue was washed thrice
with ethyl acetate to remove the products with high boiling points,
which was then dried at 60◦C for the next catalytic reaction. Fig. 5 shows
that the catalyst has a poor reusability, as the yields of both ML and DMS
decreased from 26% to 17% and 27% to 6% after three cycles respec-
tively. Later, the control experiments were performed to probe the
possible reason for the decreased in reactivity.
Reaction conditions of run 1–3: 1.5 mmol D-fructose, 75 mg
Amberlyst-70 or recovered catalyst, 15 mL methanol, 10 bar O2, 130◦C,
2 h. Control test 1: 75 mg Amberlyst-70 was pretreated at 130◦C for 2 h
in presence of 15 mL methanol and 10 bar O2, followed by adding 1.5
mmol D-fructose and recharging 10 bar O2. Then the autoclave was
reheated to 130◦C for 2 h; Control test 2: 75 mg Amberlyst-70 was
pretreated at 130◦C for 2 h in presence of 15 mL methanol and 10 bar O2,
followed by removing all solvents by vacuum drying. Then 1.5 mmol D-
fructose with 15 mL fresh methanol and 10 bar O2 was recharged, and
the autoclave was reheated to 130◦C for 2 h.
Fig. 6. FT-IR spectra of fresh and spent Amberlyst-70.
–
group [23]. The apparent vibration changes in the O H and C=O bands
of the spent catalyst demonstrated a partially disintegrated catalyst
–
structure after the reaction. Additionally, the broadened O H band and
strengthened C=O bands implied that the hydrogen-bonding and po-
larity of the spent catalyst were stronger than those of the fresh one,
which might attribute to its dissolution in methanol.
In control test 1, Amberlyst-70 was pretreated under optimal con-
ditions, and 1.5 mmol D-fructose and 10 bar O2 were loaded to restart a
new reaction. As a result, only low amounts of the desired products were
detected. Upon pretreatment of Amberlyst-70, two major side products,
namely methyl formate and dimethoxymethane, formed from methanol
were successfully identified by GC–MS and NMR. Later, these two
commercial side-products were separately added for the conversion of D-
fructose under standard reaction conditions (Table S1). The experi-
mental results indicated that the side products exhibited negative effects
on the conversion of D-fructose into DMS, especially dimethoxymethane.
Also, a better reaction result was obtained by removing methyl formate
and dimethoxymethane formed during the pretreatment of Amberlyst-
70 and addition of fresh methanol (control test 2). Moreover, the cata-
lytic differences within Run 1 and control test 2 also revealed that the
pretreatment of Amberlyst-70 resulted in its catalytic activity loss. Thus,
it is easy to conclude that the limited yield of DMS (~30% from D-
fructose) is possibly due to the negative effects of the side products and
deactivation of the catalyst.
To better understand the changes in Amberlyst-70, p-toluenesulfonic
acid, with a comparable catalytic activity with Amberlyst-70 (Fig. 2),
was chosen as the model catalyst. After the pretreatment of the p-Tol-
uenesulfonic acid in methanol at 130◦C for 2 h under 10 bar O2, the
solvent was evaporated, and the liquid residue was analyzed by
recording the 1H NMR spectrum (Fig. S3). Obviously, there are new
peaks between 7.5–8.0 ppm ascribed to the hydrogen shift of the ben-
zene ring. The sulfonic acid group might be partially esterified, which is
clear evident from the detection of methyl p-toluenesulfonate by
GC–MS. In addition, a trace amount of toluene was also detected in the
pretreated p-toluenesulfonic acid by GC–MS, suggesting the unstable
nature of the sulfonic acid group under the present reaction conditions.
The changes in the sulfonic acid group were stipulated to be partially
responsible for the catalyst dissolution and deactivation. Furthermore,
elemental analysis and acidity measurement of fresh and spent catalysts
were performed (Table S2). Unfortunately, only negligible changes in
the C, H and S contents within the fresh and spent Amberlyst-70 were
observed, suggesting the leaching effect is inconspicuous. The acidity
measurement showed unexpectedly a slight increase in the acidity of the
spent catalyst, probably attributing to the differences in solubility
(acidic groups can be easily and sufficiently titrated or not) [23].
The ATR-IR analysis of the fresh and spent Amberlyst-70 was also
performed (Fig. 6). The characteristic vibration bands of the catalysts at
1161 and 1033 cmꢀ 1 are assigned to the O=S=O stretching vibration
2.7. Plausible reaction mechanism
According to the literature, the one-pot production of ML from D-
fructose could be easily realized using an acidic catalyst [24–27]. In this
study, ML was indeed observed as the major product under N2 and O2
atmospheres (Table 1 entries 4, 8 and 13 at 0.5 h), however, ML was
gradually consumed under O2 atmosphere with time, and DMS was
simultaneously formed. This indicated that ML acted as an intermediate
for the formation of DMS, which was consistent with the results of the
studies [3,6,16]. Additionally, the HMF derivatives were observed at the
beginning (Fig. S4). Therefore, a plausible mechanism was proposed
based on the results and observations as follows (Scheme 1): Initially,
D-fructose was firstly dehydrated into HMF, which was made to undergo
etherification and/or acetalization of HMF followed by its alcoholysis to
form ML. Finally, the B-V oxidation of ML into DMS was performed in
the presence of molecular oxygen and Amberlyst-70. The side products,
namely DMM and DMF, were possibly produced via the B-V oxidation of
HMF, as discussed previously [18].
Fig. 5. Reusability of the catalyst on the conversion of D-fructose.
5