16
C. Dai et al. / Journal of Catalysis 356 (2017) 14–21
diffractometer (Cu K
a
radiation, k = 0.154 nm). The diffractometer
Pt). Due to the lower surface energy of gold, higher surface Au com-
positions were induced according to the methods mentioned
operating conditions were 40 kV, 30 mA at a 2q range of 10–80
with the scan speed of 3° minÀ1. Thermogravimetric analysis
(TGA) of AuPt/C nanoparticles was carried in a TGA/DSC 2 STAR
System (Mettler Toledo). AuPt/C nanoparticles were heated from
25 °C to 800 °C in air with a heating rate of 10 K/min. Trace metal
analysis of AuPt/C nanoparticles was performed using an Agilent
7700x ICP-MS (Agilent Technologies). Samples were digested using
Milestone Ethos one microwave digestion system before the
analysis.
above, from AuPt (64% Ptsurf) with 64 atom% Pt, to AuPt (29% Ptsurf
)
with 29 atom% Pt and AuPt (15% Ptsurf) with 15 atom% Pt. This
change in the composition converted a Pt-rich surface to an Au-
rich surface. Fig. 1b&c and S2 shows the representative TEM
images and particle size distribution of as-synthesized AuPt (90%
Ptsurf) nanoparticles, respectively. The particle size was found to
be 6.3 1.4 nm. This observation is in accordance with the previ-
ous work by Suntivich et al. [28], in which AuPt nanoparticles with
different surface compositions prepared by the same method have
a particle size range of 6–8 nm. As for size effect of carbon-
supported metal nanoparticles on electrochemical performance,
normally it can be observed for those particles with more signifi-
cant size difference [29,30]. Therefore, in such a small size varia-
tion range, the size effect of AuPt nanoparticles is limited. Fig. S3
shows the XRD analysis used to determine the structure of AuPt
nanoparticles, and no phase separation was detected. The loading
of AuPt nanoparticles on carbon was confirmed by the TGA tests,
which was approximately 20 wt% (shown in Fig. S4). The molar
ratio of Pt/Au calculated based on counts from ICP-MS for AuPt
2.5. Electrochemical oxidation of glycerol
The electrochemical oxidation of glycerol was performed in an
H-type cell utilizing a three-electrode system with a Hg/HgO (1
M KOH) reference electrode and a graphite counter electrode
(schematic illustration shown in Fig. 1a). The two compartments
were separated with an AMI-7001 anion exchange membrane
(Membranes International), and the counter electrode was sepa-
rated from the working and reference electrodes to prevent the
reduction of oxidation products at the counter electrode. The
membrane was immersed in the 1 M KOH solution overnight
before the test. Glycerol (0.5 M) was dissolved into a 1 M KOH
solution (or other concentrations when mentioned), and the solu-
tion was bubbled with Ar to purge air before and during experi-
ments. Each chamber was filled with 10 mL of 0.5 M glycerol & 1
M KOH solution. A graphite paper (20 mm  15 mm) with AuPt
(90% Ptsurf), AuPt (64% Ptsurf), AuPt (29% Ptsurf) and AuPt (15% Ptsurf
were 1.05, 1.06, 1.03 and 1.04, respectively (shown in Table S1).
)
3.2. Product analysis and reaction pathways
To compare the influence of the AuPt surface composition and
applied potential on the selectivity (molar percentage) and activity
of the glycerol oxidation (total electron transferred), the glycerol
oxidation was first operated at four different potentials, 0.45 V,
0.6 V, 0.9 V and 1.05 V, catalyzed by four types of as-synthesized
AuPt nanoparticles for 12 h. The potentials were selected from
the potential where the oxidation started for all AuPt catalysts in
the linear sweep voltammetry (LSV) (shown in Figs. 2a and S5),
to the potential inducing the surface segregation [27] and Pt disso-
lution [31]. Ar was bubbled during the whole reaction process to
purge air, as lower O2 concentration favors the formation of LA
from glycerol oxidation [12,13]. The Au and Pt monometallic
nanoparticles were also employed for comparison purpose. How-
ever, the glycerol oxidation over Au electrode was only performed
at 0.9 V and 1.05 V, as the oxidation of glycerol on Au was only
observed at potential higher than 0.6 V [32].
nanoparticles loading of 120 l
g/cm2 on both sides was used as
the working electrode. The reaction products were collected with
a syringe at fixed time during or after chronoamperometry tests
for the following product analysis. The catholyte was also analyzed
after electrolysis, and no electro-oxidation product was detected
using either NMR or HPLC methods.
2.6. Analysis of glycerol oxidation products
Chromatographic determination of glycerol oxidation products
was analyzed by an Agilent 1260 Infinity II HPLC (Agilent Tech-
nologies). The column used was an Aminex HPX87-H (Bio-Rad)
and the eluent used was 5 mM sulfuric acid. During the test, 20
lL mixture of 0.5 M H2SO4 and sample solution was injected into
the column and the temperature of the column was kept at 55
°C. The flow rate was 0.5 mL/min. The separated compounds were
detected with a refractive index detector (RID) and a multiple
wavelength detector (MWD). The expected products were also
analyzed by HPLC to perform a standard calibration curve. Both
1H-NMR and 13C-NMR spectrum were recorded using a Bruker
AV 300 MHz NMR spectrometer. 0.4 mL sample and 0.2 mL D2O
were mixed and used for each test.
In all reactions in alkaline conditions, the products were
obtained as salts, but they were marked as the acid forms for sim-
plicity and comparison. The products of glycerol electro-oxidation
were first qualitatively determined by NMR and HPLC.
Figs. 3a, b and 4 show the 1H, 13C NMR and HPLC analysis of oxida-
tion products of 0.5 M glycerol in 1 M KOH by AuPt (15% Ptsurf) at
1.05 V after 12 h, respectively. Fig. 3a and b shows that the prod-
ucts analyzed from the NMR include FA, acetic acid (AA) (concen-
tration too low to be detected in 13C NMR), GA, GLA, LA, TA, OA (no
proton in the salt form to be detected in 1H NMR) and the reactant,
glycerol. In Fig. 4, the peaks at the retention time of 7.701, 9.161,
12.555, 14.301, 14.741, 15.528, 16.315, and 17.858 min are attrib-
uted to OA, TA, GLA, GA, LA, glycerol, FA and AA, respectively. The
types of products are consistent with the results obtained from
NMR. In addition, since there are no aldehyde and ketone group
detected in NMR, the possible peak overlapping with glyceralde-
hyde (GLAD) (12.887 min) and dihydroxyacetone (DHA) (15.650
min) can be ignored. Moreover, the all-acids production is more
accessible for separation and purification compared with the mix-
ture of acids and aldehydes/ketones for industrial applications
[33].
3. Results and discussion
3.1. Catalyst characterizations
The as-synthesized and heat-treated AuPt nanoparticles were
first electrochemically characterized to determine the surface com-
position [27,28]. All potentials mentioned are vs. RHE, for simplic-
ity and comparison. Fig. S1 shows CVs of AuPt nanoparticles with
different surface compositions. The ESA of Pt was calculated by
integrating and averaging the capacity-corrected by hydrogen
underpotential adsorption/desorption in the range of 0.05–0.4 V
using the constant of 210
by integrating the reduction peak of gold oxide in the range from
ca. 0.95 V to 1.3 V using the constant of 340
C/cm2. The AuPt
(90% Ptsurf) nanoparticles possessed a Pt-rich surface (90 atom%
l
C/cm2. The ESA of Au was calculated
On the basis of products analyzed from the NMR and HPLC
results, the proposed reaction pathway for glycerol oxidation is
shown in Scheme 1. Glycerol is first oxidized to GLAD or DHA by
l