2
microscopy (TEM) and energy-dispersive X-ray (EDX) spec-
troscopy (Figure S1b) demonstrated a lack of Ru aggregates in
the Ru-CTF. (TEM images for RuO2/C catalyst are also shown
in Figure S1a). Figure S1c shows the Fourier transform of the
loading was 5.1 ¯g/cm . An oxidation current can be observed
for both catalysts near the potential at which high-valency Ru
are formed, suggesting that the Ru-species was the active species
for glycerol oxidation, similar to the oxidation of benzyl alcohol
3
19
k -weighted extended X-ray absorption fine structure (EXAFS)
on Ru-based materials.
Ru-K edge oscillations for the synthesized Ru-CTF. The Ru-
CTF did not generate a peak corresponding to RuRu bonds,
whereas peaks assignable to RuN and RuCl bonds were
observed at 1.6 and 1.9 ¡, respectively. These results demon-
strate that atomically-dispersed Ru was present in the Ru-CTF,
as expected. The surface concentration of each element was also
determined by semi-quantitative X-ray photoelectron spectros-
copy (XPS), from which it was determined that the atomic Ru/N
ratios in the Ru-CTF and RuO2/C were 0.24 and 0.40, respec-
tively (Figure S1d). The Ru-CTF and RuO2/C particles were
separately loaded onto glassy carbon substrates by drop casting
to fabricate the electrodes.
In the case of the Ru-CTF, the glycerol oxidation current
increased beginning at 1.0 V and plateaued at approximately
1.2 V. The oxidation current was not observed for CTF without
Ru-loading (Figure S2), further supporting that Ru was the
active species for the glycerol oxidation. In contrast, the onset
potential for glycerol oxidation over the RuO2/C was approx-
imately 1.2 V and the associated current density was much lower
than that produced by the Ru-CTF. Figures 2c and d show the
chronoamperometry data acquired at 1.4 V in electrolytes with
various glycerol concentrations. When using the Ru-CTF, the
current increased as the glycerol concentration was increased
from 10 mM to 1 M (Figure 2c), while the current generated by
the RuO2/C plateaued at glycerol concentrations greater than
100 mM (Figure 2d).
The black curve in Figure 2a is the cyclic voltammogram
obtained in 0.1 M HClO4 for the Ru-CTF electrode, and exhibits
a pair of redox peaks at approximately 1.2 V. A previous study
demonstrated that high-valency Ru species were formed at
this potential in association with the oxidation of coordinated
It is known that this oxidation proceeds via the formation of
3
glyceraldehyde or dihydroxyacetone (Figure 1). Herein, C3
compounds resulting from the two- or four-electron oxidation of
glycerol (i.e., glyceraldehyde, glyceric acid, dihydroxyacetone
and hydroxypyruvaldehyde) are referred to as partial oxidation
products. Similarly, products obtained from additional down-
stream reactions are defined as deep oxidation products. In this
work, the partial oxidation products were analyzed by high
performance liquid chromatography (HPLC) (Figure S3) while
liquid chromatography-mass spectrometry (LC/MS) and differ-
ential electrochemical mass spectrometry (DEMS) were used
to verify the generation of specific deep oxidation products
(Figures S4 and S5). Among the partial oxidation products,
only glyceraldehyde and glyceric acid could be detected at
quantifiable levels, and the formation of dihydroxyacetone was
not confirmed. Among the potential deep oxidation products,
only glycolic acid and carbon dioxide were detected while the
amounts of tartronic acid, hydroxypyruvic acid, oxalic acid, and
formic acid were all below the detection limit of the method.
Figure 3a plots typical time courses of the accumulated
amounts of the partial oxidation products obtained using the
Ru-CTF. With increases in the Coulombic charge (Q), it is
evident that glyceraldehyde (blue) and glyceric acid (orange)
were gradually accumulated. The faradaic efficiency (FE) values
calculated from the data in Figure 3a are summarized in
Figure 3b. It should be noted that the contributions of catalysts
decomposition and oxygen evolution reactions are excluded
when calculating FE (please find details in the Supporting
Information). In the case of the Ru-CTF, the overall FE value for
the partial oxidation products was approximately 20%, while the
remaining FE was associated with the deeper oxidation products
including CO2 (Figure S5). It is also apparent that the FE
associated with the production of glyceraldehyde slightly
decreased (from 10% to 5%) with increases in Q. In addition,
the selectivity for glyceraldehyde generation was significantly
higher when employing the RuO2/C (Figures 3c and d), such
that the overall FE for the partial oxidation products reached
approximately 60%. As shown in Figures 2a and b, the onset
potential for glycerol oxidation over the Ru-CTF was approx-
2
1
water molecules. Redox peaks assignable to the oxidation of
coordinated water (i.e., the formation of high-valency Ru) were
also identified in the data obtained for the RuO2/C electrode
(Figure 2b, black). However, in contrast to the Ru-CTF,
repeating RuORu sites were evidently generated on the
RuO2/C, leading to competition with water oxidation via OO
bond formation at neighboring Ru sites.19 In fact, compared to
Ru-CTF, a large current increase due to oxygen evolution has
been observed when employing RuO2/C at potentials more
positive than 1.4 V.
Current density ( j) versus potential (U ) curves for the
Ru-CTF and RuO2/C electrodes obtained in 0.1 M HClO4
solutions containing 1 M glycerol are presented as red curves
in Figures 2a and 2b, respectively. In both catalysts, the Ru
Figure 2. (a) (b) Cyclic voltammetry of Ru-CTF (a) and
RuO /C (b) electrodes in 0.1 M HClO , without (black line)
2
4
¹
1
and with (red line) 1 M glycerol. Scan rate: 10 mVs . (c) (d)
Chronoamperograms at 1.4 V vs. RHE for Ru-CTF (c) and
RuO /C (d) electrodes in 0.1 M HClO , without (- - - - - - - - - -)
imately 0.2 V more negative than that when using the RuO /C.
2
4
2
and with 1 M (1), 100 mM (2), and 10 mM (3) glycerol.
Therefore, an analysis of the reaction products generated by the
© 2020 The Chemical Society of Japan