Effect of nanostructured carbon support on copper electrocatalytic activity toward CO2 electroreduction to hydrocarbon fuels

Article abstract of DOI:10.1016/j.cattod.2016.11.001

The effect of support on electrocatalytic activity of Cu nanoparticles (NPs) towards CO₂ electroreduction to hydrocarbon fuels (CH₄ and C₂H₄) is investigated for three types of nanostructured carbons: single wall carbon nanotubes (SWNT), reduced graphene oxide (RGO) and onion-like carbon (OLC). Cu/SWNT, Cu/RGO and Cu/OLC composite catalysts are synthesized and characterized by X-ray diffraction analysis, transmission electron microscopy and electrochemical surface area measurements. Electrocatalytic activities of the synthesized materials, as measured in an electrochemical cell connected to a gas chromatograph, are compared to that of Cu NPs supported on Vulcan carbon. All four catalysts demonstrate higher activity towards C₂H₄ generation vs CH₄, with production of the latter mostly suppressed on Cu NPs supported on nanostructured substrates. Onset potentials for C₂H₄ vs CH₄ generation demonstrate 200?mV positive shifts for Cu/SWNT, Cu/RGO, and Cu/OLC catalysts. The Cu/OLC catalyst is found to be superior to the other two nanostructured catalysts in terms of stability, activity and selectivity towards C₂H₄ generation. Its faradaic efficiency reached 60% at ?1.8?V vs Ag/AgCl. The enhanced stability and activity of the Cu/OLC catalyst can be attributed to the unique catalyst design, wherein a shell of OLC surrounds the Cu NPs. Such a configuration enables the outer layer to act as a filter that protects the Cu surface from adsorption of undesirable species, enhances its electrocatalytic performance, and improves its viability in CO₂ electroreduction reaction. The enhanced selectivity of the Cu/OLC catalyst towards the C₂H₄ production is most likely related to the enhanced electrocatalytic activity of the OLC support towards the CO₂ electroreduction to CO. Higher CO surface concentration from CO₂ electroreduction on the OLC support, rather than a greater number of available sites for CO dimerization, likely originated this behavior. CO molecules generate on surfaces of OLCs, dimerize on Cu NP (100) planes, and, subsequently, yield additional C₂H₄ molecules. Both the reduction of the CO₂ to CO on the surface of OLCs and reduction of “extra” CO molecules to C₂H₄ on the surface of Cu NPs are expected to lead to an increase in the local pH, which is also beneficial for the C₂H₄ production.

Full text of DOI:10.1016/j.cattod.2016.11.001

G Model
CATTOD-10436; No. of Pages9
ARTICLE IN PRESS
Catalysis Today xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Effect of nanostructured carbon support on copper electrocatalytic
activity toward CO2 electroreduction to hydrocarbon fuels
Olga Baturinaa,∗, Qin Lua,∗, Feng Xub,1, Andrew Purdya, Boris Dyatkinc, Xiahan Sangd,
Raymond Unocicd, Todd Brintlingere, Yury Gogotsif
a Chemistry Division, Naval Research Laboratory, Washington DC, USA
b Wolfson Catalysis Centre, Department of Chemistry, University of Oxford, Oxford OX1 3QR, UK
c NRC Post-Doctoral Researcher, Naval Research Laboratory, Washington DC, USA
d Center for Nanophase Materials Sciences, Oak Ridge National Laboratory, Oak Ridge, TN, USA
e Materials Science and Technology Division, Naval Research Laboratory, Washington DC, USA
f
A.J. Drexel Nanomaterials Institute and The Department of Materials Science and Engineering, Drexel University, Philadelphia PA, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 30 July 2016
Received in revised form 17 October 2016
Accepted 3 November 2016
Available online xxx
The effect of support on electrocatalytic activity of Cu nanoparticles (NPs) towards CO2 electroreduction
to hydrocarbon fuels (CH4 and C2H4) is investigated for three types of nanostructured carbons: single
wall carbon nanotubes (SWNT), reduced graphene oxide (RGO) and onion-like carbon (OLC). Cu/SWNT,
Cu/RGO and Cu/OLC composite catalysts are synthesized and characterized by X-ray diffraction analy-
sis, transmission electron microscopy and electrochemical surface area measurements. Electrocatalytic
activities of the synthesized materials, as measured in an electrochemical cell connected to a gas chro-
matograph, are compared to that of Cu NPs supported on Vulcan carbon. All four catalysts demonstrate
higher activity towards C2H4 generation vs CH4, with production of the latter mostly suppressed on Cu
NPs supported on nanostructured substrates. Onset potentials for C2H4 vs CH4 generation demonstrate
200 mV positive shifts for Cu/SWNT, Cu/RGO, and Cu/OLC catalysts. The Cu/OLC catalyst is found to be
superior to the other two nanostructured catalysts in terms of stability, activity and selectivity towards
C2H4 generation. Its faradaic efficiency reached 60% at −1.8 V vs Ag/AgCl. The enhanced stability and
activity of the Cu/OLC catalyst can be attributed to the unique catalyst design, wherein a shell of OLC
surrounds the Cu NPs. Such a configuration enables the outer layer to act as a filter that protects the Cu
surface from adsorption of undesirable species, enhances its electrocatalytic performance, and improves
its viability in CO2 electroreduction reaction.
Keywords:
CO2 electroreduction
Copper electrocatalytic activity
Copper nanoparticles
Nanostructured carbon support
Hydrocarbon fuels
The enhanced selectivity of the Cu/OLC catalyst towards the C2H4 production is most likely related to
the enhanced electrocatalytic activity of the OLC support towards the CO2 electroreduction to CO. Higher
CO surface concentration from CO2 electroreduction on the OLC support, rather than a greater number
of available sites for CO dimerization, likely originated this behavior. CO molecules generate on surfaces
of OLCs, dimerize on Cu NP (100) planes, and, subsequently, yield additional C2H4 molecules. Both the
reduction of the CO2 to CO on the surface of OLCs and reduction of “extra” CO molecules to C2H4 on the
surface of Cu NPs are expected to lead to an increase in the local pH, which is also beneficial for the C2H4
production.
Published by Elsevier B.V.
1. Introduction
activity and selectivity for CO2 conversion to hydrocarbons (CH4
and C2H4) was achieved for Cu NPs supported on high surface area
Carbon dioxide is a greenhouse gas that can be converted into
useful fuels such as CH4 and C2H4 [1–8]. In our prior work [9], high
supports such as Vulcan carbon (VC), Ketjen black (KB), and nanos-
tructured single wall carbon nanotubes (SWNTs). Carbon materials
have been found to not only serve as inert supports, but also enable
CO2 electroreduction on Cu NPs, most likely via a hydrogen spillover
effect [10].
Here, we investigate the effect of carbon supports on the prod-
uct distribution in CO2 electroreduction reaction on Cu NPs in
more detail by including reduced graphene oxide (RGO) and onion-
∗
Corresponding authors.
E-mail addresses: olga.baturina@nrl.navy.mil (O. Baturina), qin.lu@nrl.navy.mil
(Q. Lu).
1
Present address: College of Materials Science and Engineering, Fuzhou Univer-
sity, Fuzhou 350108, China
http://dx.doi.org/10.1016/j.cattod.2016.11.001
0920-5861/Published by Elsevier B.V.
Please cite this article in press as: O. Baturina, et al., Effect of nanostructured carbon support on copper electrocatalytic activity toward
CO2 electroreduction to hydrocarbon fuels, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.11.001

G Model
CATTOD-10436; No. of Pages9
ARTICLE IN PRESS
2
O. Baturina et al. / Catalysis Today xxx (2016) xxx–xxx
oxidized Cu species to Cu0, but also in the reduction of GO support
to graphene (which we call RGO to be consistent with the previ-
ous reports in the literature), as confirmed by X-Ray Photoelectron
Spectroscopy (XPS) analysis (see Fig. S3). Unlike other catalysts in
this study, Cu/OLC catalyst was synthesized in the air-free environ-
ment, which resulted in the absence of any significant amount of
bulk Cu oxides, as confirmed by XRD (Fig. S2). Consequently, the
heat-treatment in H2/Ar flow was not required for this catalyst.
Specimens for transmission electron microscopy (TEM) char-
acterization were prepared by immersing the Cu NPs on different
carbon supports in isopropanol, followed by sonication for a few
seconds. The solution was then dispensed on a carbon coated/gold
TEM grid. A Hitachi HF-3300, equipped with a cold-field emission
gun and operating at 300 kV, was used for high resolution transmis-
sion electron microscopy (HRTEM). From selected HRTEM images,
Fast Fourier Transform (FFT) analysis was used to determine the
structure of the Cu NPs [17].
like carbon (OLC) [11–13] nanostructured supports. Three of these
(SWNT, RGO, and OLC) share common properties, such as good
chemical and thermal stabilities and high specific surface areas
(SSAs). They display mostly external surface areas and, subse-
quently, allow high catalyst utilization, as none of the catalyst
particles become inaccessibly buried inside the pores. However,
geometrical configurations for these three supports are com-
pletely different. While RGO can be represented by large, flat
two-dimensional (2D) sheets, SWNTs are one-dimensional (1D)
cylinders with an extremely high length-to-diameter ratio, and
OLC consists of multiple zero-dimensional (0D) concentric stacked
graphitic spheres [12]. The geometry of each material influences the
accessibility of the Cu metals to the reagents and governs the num-
ber of CO2 chemisorption sites. Subsequently, our efforts explored
this parameter in order to maximize the CO2 to hydrocarbon con-
version efficiencies, reaction yields, and catalyst stabilities.
The XRD patterns were acquired using a Rigaku SmartLab X-Ray
diffractometer equipped with a Cu K␣ radiation source and a D-teX
linear Si strip detector.
2. Material and methods
Cu NPs on VC, SWNT, and graphene oxide (GO) supports were
synthesized by the reduction of CuCl2 with NaBH4 in aqueous ethy-
lene glycol solution in the presence of supporting material [9]. VC,
SWNT and GO were acquired from Cabot, Thomas Swan & Co Ltd,
and Graphene Laboratories, Inc., respectively. VC and GO were used
without further purification. Raw chemical vapor deposition grown
SWNTs were first treated by H2O steam at 900 ◦C for 4 h to remove
amorphous carbon [14]. After steam purification, the sample was
mixed with 6 M HCl and the mixture was refluxed for 2 h to remove
any metal catalyst particles.
Supported Cu nanoparticles (20 wt%) on different carbon mate-
rials were synthesized by first dispersing carbon support (80 mg)
in an aqueous ethylene glycol solution (20% v/v) with 1 h sonica-
tion. To this dispersion, CuCl2 2H2O (53.69 mg) was added and the
resulting mixture was stirred overnight under N2. The reaction flask
containing the mixture was then placed in an ice bath before slowly
adding a freshly made NaBH4 solution (100 mg NaBH4 in 5 mL H2O).
The flask was kept in an ice bath for another 2 h under N2 to ensure
the complete reduction of Cu2+ to Cu0. The Cu nanoparticles sup-
ported on different carbon materials were collected by filtering the
mixture through a Nylon membrane (0.2 ␮m) and cleaned with
copious amounts of water and ethanol sequentially. GO could also
be reduced to RGO during that treatment.
XPS measurements were performed with K-Alpha XPS system
with a monochromatic Al K␣ source operating at 225 W. High res-
olution C1s and Cu 2p spectra were acquired.
Gas sorption measurements were carried out using Quadrasorb
(Quantachrome Instruments, USA) and ASAP 2020 (Micromeretics
Corporation, USA) sorption analyzers. The instruments collected
gas adsorption and desorption isotherms using N2 gas in the
0.00075–0.9995 P/P0 range at 77 K (with a liquid N2 coolant bath).
The Brunauer-Emmett-Teller equation (BET) was used to calculate
the BET specific surface area (SSA) from P/P0 values in the 0.05–0.30
range [18]. Quenched Solid Density Functional Theory (QSDFT) and
Non-local Density Functional Theory (NLDFT) data reduction analy-
sis provided information on a secondary SSA value, cumulative pore
volume (cm3 g−1), and pore size distributions [19]. Quantachrome’s
Quadrawin software and the Micromeretics data analysis package
provided all BET calculations, DFT data modeling kernels, and com-
putational analysis.
Thermogravimetric analysis (TGA) of carbon supports and Cu
NPs/carbon composites was carried out using a Q50 TGA (Thermal
Analysis). Samples were loaded into an alumina crucible and heated
to 900 ◦C under 10 mL/min flowing air environment.
The electrochemical surface area (ECSA) of each catalyst was
determined by Pb underpotential deposition (UPD) [20–22] from
a solution of 0.1 M HClO4 + 0.001 M Pb(ClO4)2, as described else-
where [9]. Briefly, a charge consumed for the stripping of Pb
adatoms from the surface of Cu NPs during 10-min potential hold
at −0.375 V vs Ag/AgCl, was converted into the ECSA assuming a
monolayer of Pb adatoms and 2e− Pb oxidation. 310 ␮C cm−2 was
used as a conversion factor [21].
OLC (ca. 7 nm in size) particles were synthesized using 1800 ◦C
vacuum annealing (10−6 torr) of UD-90 nanodiamonds (NanoBlox,
Inc.) [15,16] in a graphite furnace (Solar Atmospheres) [11]. For
the 20 wt% Cu/OLC composite catalyst, 142 mg of the OLC was
mixed with 20 g H2O and 4 g ethylene glycol and sonicated for 3 h.
CuCl2·2H2O (77 mg) was added and degassed by stirring under vac-
uum and then stirred overnight under N2. Then, the mixture was
stirred in an ice bath under N2 and 240 mg NaBH4 dissolved in 5 mL
H2O was added via syringe. Following that, the mixture was stirred
for 2 h under N2 and then filtered in a N2 flow purge box, and the
solid pumped dry, isolating 170 mg of Cu/OLC.
The Cu loading on the supports was confirmed by thermogravi-
metric analysis (TGA). Representative TGA data for nominal 20 wt%
Cu/OLC shown in Fig. S1 confirms that the actual Cu loading was
19.3 wt.%, based on a 23 wt.% weight content of CuO formed after
decomposition of the OLC support in air at ca 650 ◦C.
Electrochemical experiments were conducted in a sealed three-
electrode electrochemical cell connected to a gas chromatograph
(GC, Bruker-450) [9]. Catalysts inks were prepared by dispersing
catalyst powders in a mixture of isopropanol and 5 wt% Nafion
ionomer solution (Ion Power, Inc.) by ultrasound treatment in an
ultrasonic bath for one hour. The volume ratio of isopropanol to
5 wt% Nafion solution in the ink was 1000:1, and the ratio of cat-
alyst to solvent was 2 mg/ml. An aliquot of ink was dropped on
the surface of a glassy carbon electrode to obtain a Cu loading of
20 ␮gCu/cm2.
Prior to conducting electrochemical experiments, all catalysts
except Cu/OLC were reduced in a flow of 10 vol% H2 (Air Liquide,
99.999%) in ultra-high purity Ar (Air Liquide, 99.999%) at 250 ◦C for
2 h in a tube furnace to ensure that all the oxides were reduced to
Cu0. Reduction of Cu bulk oxides to Cu0 was confirmed by X-Ray
Diffraction (XRD) analysis (see Fig. S2). For Cu NPs supported on
GO, heat-treatment in H2/Ar resulted not only in the reduction of
Thin films of supported catalysts with
a Cu loading of
20 ␮gCu/cm2 were deposited on the surface of mirror-polished
glassy carbon electrodes, and current vs. time curves were recorded
at constant potentials of −1.4, −1.6, −1.8, −2 and −2.2 V vs
Ag/AgCl in CO2-saturated 0.1 M KHCO3 solution at a rotation rate
of 1600 rpm. The flow of CO2 (Scott Specialty Gases, 99.997%,
Please cite this article in press as: O. Baturina, et al., Effect of nanostructured carbon support on copper electrocatalytic activity toward
CO2 electroreduction to hydrocarbon fuels, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.11.001

G Model
CATTOD-10436; No. of Pages9
ARTICLE IN PRESS
O. Baturina et al. / Catalysis Today xxx (2016) xxx–xxx
3
SFE grade) through the solution was maintained at 60 mL min−1
throughout the experiment. Catalysts’ thin films on glassy carbon
served as working electrodes (WE), while Pt mesh and Ag/AgCl in
3 M NaCl (BASi) were used as counter and reference electrodes,
respectively. GC analysis of gaseous reaction products (CH4, C2H4,
CO and H2) was conducted at 25, 45, and 65 min after applying
potential. More details about GC analysis of gaseous reaction prod-
ucts can be found in our prior work [9].
Electrocatalytic activities of VC, SWNT, OLC and RGO supports
towards the CO2 electroreduction were evaluated in the same man-
ner as for the supported Cu NPs. Current vs time curves were
recorded for thin films of VC, SWNT, and VC at constant potentials
of −1.2 V, −1.6 V, and −2.2 V. For the OLC support, the potentio-
static experiments were performed at −1.4 V, −1.6 V, and −2.2 V.
VC and SWNT supports were reduced in H2/Ar at 250 ◦C for 2 h
prior to conducting electrochemical experiments, while no heat-
treatment was utilized for the OLC support. Due to lack of original
material, RGO support was obtained in-situ by Cu stripping from
thin film of Cu/RGO catalyst in an electrochemical cell. The loading
of standalone supports was identical to their loadings in the films
of supported Cu NPs.
3. Results and discussion
Fig. 1. HRTEM images of (a) 20 wt% Cu/VC, (b) 20 wt% Cu/SWNT, (c) 20 wt% Cu/OLC
and (d) 20 wt% Cu/RGO catalysts. All catalysts except 20 wt% Cu/OLC, were H2-
reduced prior to microscopic characterization.
Table 1 compares the size of Cu NPs before and after high-
temperature reduction in H2/Ar for three different supports, and
their ECSAs. Although the Cu NPs initially had similar size on all
supports (9–15 nm) before the reduction by H2, their size increases
following the reduction was support-dependent. While Cu NPs sup-
ported on SWNTs undergo finite, but insignificant growth (from 16
to 19 nm), the size of NPs supported on VC and RGO more than
doubled from 9 to 33 and 15 to 36 nm, respectively.
The theoretical surface area calculated assuming a spherical
XRD-based diameter of the Cu NPs (Column 4, Table 1) varies from
18.6 m2/gCu for Cu/RGO to 51.5 m2/gCu for Cu/OLC. However, Pb
UPD provided ECSAs that were significantly smaller (negligible for
Cu/OLC and 3.4 0.1 m2/gCu for Cu/RGO). TEM imaging, shown in
Fig. 2, aimed to clarify the discrepancies between the experimen-
tally measured and theoretically derived SAs.
tribution. The presence of larger Cu NPs in the latter catalyst is
confirmed by an additional TEM image shown in Fig. S2.
A Fast-Fourier-Transform (FFT) analysis of several images was
performed to determine the crystalline structure of the Cu/OLC cat-
alyst (Fig. 2) in order to confirm its copper composition. Measured
lattice spacings of several nanoparticles reveal that the Cu NPs show
a mixture of metallic Cu and CuO. The lattice spacing of 0.20 nm
corresponds to Cu {111} and the spacings of 0.25 nm and 0.28 nm
corresponds to CuO {002} and CuO {110}, respectively. The CuO
likely forms a shell that protects the Cu core from oxidation. While
the 10–30 nm larger particles in Fig. 1c most likely identify oxidized
Cu NPs, the smaller 5–7 nm diameter spheres are actually the OLC
support.
Cu NPs supported on VC, shown in Fig. 1 appear to be
evenly distributed on the support, with a particle size between
4 and 25 nm. However, in other cases (Fig. S2) agglomerated
Cu NPs were observed on the VC surface. The HRTEM image of
20 wt% Cu/SWNT catalyst composite (Fig. 1b) reveals mostly well-
dispersed 10–25 nm Cu NPs evenly distributed on the SWNTs, but
a few larger (ca 40 nm in diameter) agglomerates are also present.
The carbon nanotubes in Fig. 1b are easily distinguished from the
Cu NPs. Two types of NPs somewhat complicate the analysis of
20 wt% Cu/OLC catalyst (Fig. 1c): larger (10–30 nm) NPs that exhibit
a uniform light contrast and density, and smaller (5–7 nm) NPs with
clearly resolved onion-like morphologies. The RGO support yields
the largest Cu NPs (Fig. 1d), with the broadest size distributions
(25–90 nm). The larger size and extent of agglomeration of Cu NPs
on RGO support is due to clustering of Cu NPs in as prepared cata-
lyst, prior to its reduction in H2/Ar (Fig. S3). This is clearly reflected
in the agglomerated NP morphology presented in the inset of Fig.
S3. As one would expect, heat-treatment of such clusters would
result in coalescence and coarsening of Cu NPs.
Of note, the Cu NPs dimensions in Cu/OLC exceeded that of
the OLC support, and such a ratio is not typically encountered in
supported metallic nanoparticles. Usually, the size of support par-
ticles is much larger than that of the catalyst NPs. For example, the
high-surface area VC support consists of primary 50 nm spheres
connected in a network of larger aggregates [23], while the size of
catalytic VC-supported Pt NPs used at the proton exchange mem-
brane fuel cell cathodes is on the order of 5 nm [24]. The Cu/OLC
catalyst can be more accurately represented as OLC nanoparticles
mounted on a Cu substrate. Furthermore, close inspection of the
Cu/OLC catalyst shows that Cu NPs are indeed surrounded by OLC
NPs.
From the close inspection of TEM images shown in Fig. 1, the
large difference between the ECSAs of Cu NPs determined by Pb2+
UPD and the ones derived theoretically (Table 1) can be rationalized
in terms of Cu NPs agglomeration. Another reason for the discrep-
ancies between experimentally and theoretically derived SAs may
be the partial oxidation of Cu NPs that prevents Pb2+ ions from
deposition on the Cu NPs surface in the UPD region.
Comparison of XRD-derived Cu NP dimeters with those from
TEM analysis shows that XRD-based diameters fall within the range
realized from TEM images for Cu/SWNT (19 nm), Cu/OLC (13 nm)
and Cu/RGO (33 nm) catalysts (Fig. 1b–d, respectively). For Cu/VC
catalyst, Cu NP diameters determined from TEM (Fig. 1a) are smaller
than that of XRD (33 nm). Since XRD measurements are dominated
by large particles, this indicates a non-uniform particle size dis-
It should be noted that gas sorption measurements summa-
rized in Table 2, show very high specific surface areas for the Cu/VC
(264 m2/g), Cu/SWNT (732 m2/g), and Cu/OLC (176 m2/g) compos-
ite, despite their low ECSAs (see Table 1). Comparison to standalone
SWNT and OLC substrates shows that metal catalysts, which are
relatively heavier than carbon, as expected, decrease the amount
Please cite this article in press as: O. Baturina, et al., Effect of nanostructured carbon support on copper electrocatalytic activity toward
CO2 electroreduction to hydrocarbon fuels, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.11.001

G Model
CATTOD-10436; No. of Pages9
ARTICLE IN PRESS
4
O. Baturina et al. / Catalysis Today xxx (2016) xxx–xxx
Table 1
XRD-based Cu nanoparticle size before and after reduction in H2 at 250 ◦C, and Cu surface area determined by Pb UPD and estimated assuming spherical Cu NPs with
XRD-based diameters.
Cu NP size, nm
Cu ECSA by Pb UPD, m2/gCu
Theoretical surface area with
XRD-based NP diameters,
m2/gCu
Before H2-reduction
After H2-reduction
20 wt% Cu/VC
9
1
33
19
–
2
2
1.8 0.1
2.8 0.2
Negligible
3.4 0.1
20.3
35.2
51.5
18.6
20 wt% Cu/SWNT
20 wt% Cu/OLC
20 wt% Cu/RGO
16
13
15
2
1
2
36
5
Fig. 2. HRTEM images of two areas of 20 wt% Cu/OLC catalyst selected for FFT analysis.
Table 2
erence spectra are ca. 2 eV apart, which makes a peak assignment
feasible. The peaks for the Cu/VC (916.7 eV), Cu/SWNT (916.7 eV),
Cu/OLC (916.7 eV) and Cu/RGO (916.6 eV) align with the peak for
the Cu2O reference (916.8 eV). This finding indicates that the sharp
peak at 932.5 eV in Fig. 3 is due to the presence of the Cu2O, but not
Cu0 on the surface of the four catalysts.
Comparison of BET specific surface areas (SSA) for the Cu/VC, Cu/SWNT, Cu/OLC and
Cu/RGO catalysts and corresponding supports. Data from manufacturers are shown
for the SWNT and RGO supports.
Sample
Cu/VC
Cu/SWNT
Cu/OLC
Cu/RGO
Catalyst SSA, m2/g
Support SSA. m2/g
264
197
732
>800
176
398
37
833
A smaller peak at 934.4 eV on the XPS spectrum of the Cu/OLC
catalyst (Fig. 3) is likely due to superposition of Cu 2p3/2 maxima
at 932.5 and 933.1 eV [25], assigned to Cu2O and CuO, respectively.
The presence of both oxides on the surface of the Cu/OLC catalyst
is in agreement with its XRD pattern (Fig. S2). Close inspection of
this pattern reveals the presence of tiny diffraction peaks at 35.6
and 38.8◦, typical for CuO (111), in addition to Cu2O (111) peak
at 36.4◦ (JSPDS cards No. 5-0661 and 74-1230, respectively). This
implies that, contrary to Cu/VC, Cu/SWNT and Cu/RGO catalysts, the
surface of Cu/OLC catalyst is oxidized to a greater extent, as both
Cu2O and CuO oxide are detected by XPS abd XRD analyses.
Electrocatalytic activities of thin films of the four composite
catalysts towards CO2 electroreduction to CH4 and C2H4 were eval-
uated in a sealed three electrode electrochemical cell, connected to
a GC.
Fig. 5 shows representative current density vs time curves for
CO2 electroreduction reaction on thin films of Cu/VC, Cu/SWNT,
Cu/OLC and Cu/RGO (inset in Fig. 5) catalysts measured at constant
potentials of −1.4 V, −1.6 V and −1.8 V. Almost identical current
densities are observed for Cu/VC, Cu/SWNT and Cu/OLC catalysts
at −1.4 V and −1.8 V. At −1.6 V, the current density on Cu/OLC
films exceeds those of Cu/VC and Cu/SWNT. Current densities on
Cu/VC, Cu/SWNT and Cu/OLC films acquire steady-state values
after ca. 33 min, when holding working electrodes at −1.4 V and
−1.6 V. At −1.8 V, a gradual decrease in current is visible up to
65 min, although it becomes less pronounced between 33 min and
65 min. Observable noise is caused by a visible molecular hydro-
of accessible surface area normalized by mass. The SSA of VC (ini-
tially, 197 m2/g) increased slightly to 264 m2/g after the addition of
Cu nanoparticles. It is possible that the introduced Cu nanoparti-
cles deagglomerated the non-porous, dense Vulcan carbon material
and made the surface more accessible to both N2 adsorbate and CO2
reagent.
However, the Cu/VC, Cu/SWNT and Cu/OLC composites still
exhibit a high SSA of over 150 m2/g. Contrary to the VC-, SWNT-
and OLC-supported catalysts, RGO-supported catalyst reveals a
relatively low surface area of 37 m2/g. This is likely due to agglom-
eration of high surface area (833 m2/g, according to manufacturer)
GO flakes during synthesis of the Cu/RGO catalyst.
XPS surface analyses shown in Fig. 3, identify the Cu species on
the surface of the four catalysts, and complement XRD bulk mate-
rials characterization. High resolution Cu 2p spectra (Fig. 3) reveal
one sharp Cu 2p3/2 peak at 932.5 eV and a smaller Cu 2p1/2 peak
at 952.0 eV for every catalyst. In addition, a broad peak at 934.4 eV
is observed for the Cu/OLC material, overlapping with the peak at
932.5 eV. A sharp peak at 932.5 eV can be assigned either to Cu0 or
cupreous oxide Cu2O, due to a very small (0.1 eV) difference in their
tabulated peak positions [25]. However, a clear distinction between
the Cu0 and Cu+ in Cu2O can be made based on the position of the
Cu LMM peaks in the Auger spectra. Cu LMM Auger spectra for the
four catalysts along with the reference spectra for Cu0, Cu2O and
CuO [25] are presented in Fig. 4. Cu0 and Cu2O peaks on the ref-
Please cite this article in press as: O. Baturina, et al., Effect of nanostructured carbon support on copper electrocatalytic activity toward
CO2 electroreduction to hydrocarbon fuels, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.11.001

G Model
CATTOD-10436; No. of Pages9
ARTICLE IN PRESS
O. Baturina et al. / Catalysis Today xxx (2016) xxx–xxx
5
Fig. 3. Cu 2p high resolution XPS spectra recorded for powders of 20 wt% Cu/VC,
20 wt% Cu/SWNT, 20 wt% Cu/OLC and 20 wt% Cu/RGO catalysts. All catalysts except
20 wt% Cu/OLC, were H2-reduced prior conducting XPS analysis.
gen evolution from the electrochemical reduction of water. More
rapid current decrease with time is observed for Cu/RGO catalyst.
Although initial current densities on Cu/RGO films (−10 mA/cm2
at −1.8 V, −6 mA/cm2 at −1.6 V and −5 mA/cm2 at −1.4 V) are
greater than those for the other three catalysts (−6.5 mA/cm2,
−3.3 mA/cm2 and −1 mA/cm2, respectively), lower or similar cur-
rent densities are observed at the completion of the experiment
(i.e. −3 mA/cm2 for Cu/RGO vs −5.3 mA/cm2 for Cu/VC, Cu/SWNT
and Cu/OLC at −1.8 V and 65 min).
Fig. 4. Cu LMM Auger electron spectra extracted from the XPS survey spectra for
20 wt% Cu/VC, 20 wt% Cu/SWNT, 20 wt% Cu/OLC and 20 wt% Cu/RGO catalysts.
Fig. 6a–d displays Faradaic efficiencies (FEs) as a function of
applied potential for the four catalysts. At −1.4 V, only H2 (40–60%
FE, black squares in Fig. 6) and CO (10–20% FE, red circles) are
produced as major reaction products. Total FEs calculated as a
sum of FEs of gas-phase reaction products, vary from 60% (Cu/VC
and Cu/SWNT catalysts) to 70% and 90% for Cu/RGO and Cu/OLC
catalysts, respectively. This finding indicates the generation of a
significant amount of liquid products (e.g. HCOOH) by the Cu/VC,
Cu/SWNT and Cu/RGO catalysts at this potential.
C2H4 is first detected on the four catalysts at −1.6 V, with
faradaic efficiencies ranging from 6% for Cu/SWNT (Fig. 6b) to 35%
for Cu/OLC (Fig. 6c). Formation of CH4 is delayed by 200 mV vs. C2H4
for all the catalysts, except for Cu/VC (Fig. 6a) where appreciable
amount of CH4 is detected at −1.6 V. Although CH4 is detected at
−1.6 V for Cu/VC, C2H4 is still a dominant hydrocarbon product,
with the FE doubling that of CH4 (20% and 6% FEs in Fig. 6a, respec-
tively). More positive onset potentials for C2H4 vs CH4 generation
accompanied by preferential generation of C2H4 over CH4 in the
potential region from −1.4 V to −1.9 V on the Cu/SWNT, Cu/OLC
and Cu/RGO materials are likely due to the fact that CO2 electrore-
duction reaction is structure-sensitive [1,26–29]. C2H4 formation
proceeds more favorably through the dimerization of CO inter-
mediates at Cu (100) facets [28]. In addition, activities of Cu(100)
facets can be promoted by creating of low-coordinated sites with
dangling bonds [26]. CH4 forms preferentially at (111) facets at
higher overpotentials [28]. Most likely, Cu NPs supported on SWNT
(Fig. 6b), OLC (Fig. 6c) and RGO (Fig. 6d) are characterized by a
greater number of low-coordinated sites or defects capable of pro-
moting activities of (100) Cu facets compared to VC-supported
catalyst (Fig. 6a), which leads to a less pronounced shift in the onset
potential for CH4 formation on the latter catalyst. This hypothesis
is indirectly supported by analysis of the XRD pattern of the Cu/VC
catalyst (Fig. S2). There is no Cu2O diffraction peak in the XRD pat-
tern of this catalyst, contrary to the other three catalysts. This may
Please cite this article in press as: O. Baturina, et al., Effect of nanostructured carbon support on copper electrocatalytic activity toward
CO2 electroreduction to hydrocarbon fuels, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.11.001

G Model
CATTOD-10436; No. of Pages9
ARTICLE IN PRESS
6
O. Baturina et al. / Catalysis Today xxx (2016) xxx–xxx
Significant CH4 generation is observed for Cu/VC (22% FE,
Fig. 6a), Cu/SWNT (18% FE, Fig. 6b) and Cu/RGO (20% FE, Fig. 6d)
at −1.8 V. The lowest FE for CH4 generation (8%, Fig. 6c) at this
potential is observed for the Cu/OLC catalyst. However, the low-
est FE towards CH4 generation is combined with the highest FE
for C2H4 generation (∼60%) for this catalyst. As a result, the Cu/OLC
material most selectively promotes CO2 reduction to C2H4, i.e. with
the highest ratio of 7.5 of FEs towards C2H4 (60%) and CH4 (8%)
generation.
The lowest activity for hydrocarbon generation is observed with
the Cu/RGO catalyst (Fig. 6d). Contrary to other catalysts, H2 is a
dominating gas-phase reaction product over the entire potential
region (−1.4 V to −2.2 V). FE for H2 generation does not decline
with a decrease in potential, but passes through a minimum at
ca. −1.6 V (Fig. 6d). A rapid increase in H2 generation at potentials
below −1.8 V is accompanied by a decrease in Cu/RGO’s activity
towards hydrocarbon generation. As a result, FE efficiency for C2H4
generation becomes negligible at −2.2 V, while the FE for CH4 gen-
eration drops to 8%. The heavily agglomerated nature of Cu NPs,
coupled with a broad particle size distribution, likely causes the
low electrocatalytic activity of the Cu/RGO catalyst towards CO2
electroreduction to hydrocarbons.
An important evaluative metric of a composite catalyst is sta-
bility of its performance vs time, where consistent performance vs
time is an indication of a persistently clean catalyst surface. Fig. 7
shows FEs as a function of time for (a) C2H4, (b) CH4, (c) CO and
(d) H2 generation on four different catalysts. FEs vs time for all four
reaction products on Cu/VC and Cu/SWNT establish similar trends.
Less than 5% decrease in C2H4, CH4 and CO generation over 40 min,
accompanied by a 6–7% increase in H2 generation are observed for
these two catalysts. For Cu/RGO catalyst, the FE for H2 generation
Fig. 5. Current density vs time curves measured on thin films of 20 wt.% Cu/VC,
20 wt.% Cu/SWNT, and 20 wt.% Cu/OLC catalysts while holding the working elec-
trodes at −1.4 V, −1.6 V and −1.8 V. Inset shows the results for the 20 wt.% Cu/RGO
catalyst. Conditions: CO2-saturated 0.1 M KHCO3, 1600 rpm, 20 ␮gCu/cm2.
indicate that the surface Cu2O film identified by XPS is thinner for
the VC-supported Cu NPs. This would lead to a fewer number of
low-coordinated sites on the surface of electrochemically reduced
Cu0 NPs. The significant effect of the film thickness on the activ-
ity and selectivity of the oxide-derived Cu catalysts towards the
CO2 electroreduction has been recently demonstrated for thermally
annealed Cu films [30].
Fig. 6. Faradaic efficiencies vs. potential for CH4, C2H4, CO and H2 generation on thin films of (a) 20 wt% Cu/VC, (b) 20 wt% Cu/SWNT, (c) 20 wt% Cu/OLC and (d) 20 wt% Cu/RGO
catalysts. Conditions: CO2-saturated 0.1 M KHCO3, 1600 rpm, 20 ␮gCu/cm2. GC sampling was performed at 25 min. (b) is reprinted with permission from Olga A. Baturina, Qin
Lu, Monica A. Padilla, et al., “CO2 Electroreduction to Hydrocarbons on Carbon-Supported Cu Nanoparticles”, ACS Catalysis 4 (2014) 3682. Copyright 2014 American Chemical
Society.
Please cite this article in press as: O. Baturina, et al., Effect of nanostructured carbon support on copper electrocatalytic activity toward
CO2 electroreduction to hydrocarbon fuels, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.11.001

G Model
CATTOD-10436; No. of Pages9
ARTICLE IN PRESS
O. Baturina et al. / Catalysis Today xxx (2016) xxx–xxx
7
Fig. 7. Faradaic efficiencies vs. time for (a) CH4, (b) C2H4, (c) CO and (d) H2 generation on thin films of 20 wt% Cu/VC, 20 wt% Cu/SWNT, 20 wt% Cu/OLC and 20 wt% Cu/RGO
catalysts at −1.8 V. Conditions: CO2-saturated 0.1 M KHCO3, 1600 rpm, 20 ␮gCu/cm2.
Fig. 8. Current density vs time measured on thin films of (a) VC, SWNT, OLC and RGO supports and (b) 20 wt% Cu/VC, 20 wt% Cu/SWNT, 20 wt% Cu/OLC and 20 wt% Cu/RGO
catalysts while holding the potentials at −1.6 V and −2.2 V. Conditions: CO2-saturated 0.1 M KHCO3, 1600 rpm.
magnifies with time, growing from 35 to 75% over 40 min. There-
fore, this catalyst likely degrades rapidly, losing activity towards
hydrocarbon generation. Indeed, a 23% decrease in the FE for C2H4
production, 10% decrease in CH4 production and 14% decrease in
CO production are observed within 40 min.
The most encouraging results are obtained for the Cu/OLC cata-
lyst. Within 40 min, Faradaic efficiencies for CH4, C2H4, CO and H2
are stable and even slightly increase with time. This is likely due to
the fact that the Cu ECSA increases with time. Highly stable perfor-
mance of the Cu/OLC catalyst may be also assigned to its unusual
structure, in which small OLCs create a shield around Cu NPs, pre-
venting the adsorption of undesirable reaction intermediates.
Several factors need to be considered in order to understand the
reasoning behind the high activity and superior selectivity towards
C2H4 production of the Cu/OLC catalyst. First, the role of the sup-
port cannot be ignored. Our prior work [9] demonstrated that VC
Please cite this article in press as: O. Baturina, et al., Effect of nanostructured carbon support on copper electrocatalytic activity toward
CO2 electroreduction to hydrocarbon fuels, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.11.001

G Model
CATTOD-10436; No. of Pages9
ARTICLE IN PRESS
8
O. Baturina et al. / Catalysis Today xxx (2016) xxx–xxx
and SWNT supports by themselves show electrocatalytic activities
towards CO2 electroreduction reaction to CH4. Small, but distin-
guishable amounts of CH4 on thin films of VC and SWNT held at
−2.2 V were confirmed by GC analysis. Although formation of C
C
bonds (and consequently C2H4) has not been reported for VC and
SWNT supports, it cannot be ruled out for the OLC support.
Second, surface Cu oxides, electrochemically reduced to Cu0 at
negative potentials of CO2 electroreduction reaction [31] tend to
show higher electrocatalytic activity and selectivity towards C2H4
production vs thermally reduced Cu0. However, all four catalysts,
including thermally reduced Cu/VC, Cu/SWNT and Cu/RGO carried
a layer of surface Cu2O prior to utilizing them for ink preparation
(Fig. 4). Cu2O and CuO are both present on the surface (Fig. 3), and
underscore significant oxidation of the Cu/OLC catalyst. However, a
recent study [32] shows that the oxidation state of a precursor does
not play an important role in selectivity of the Cu catalysts towards
C2H4 formation.
Third, the Cu particle size in Cu/OLC catalyst (13 nm in diameter)
is smaller vs that in Cu/SWNT (19 nm), Cu/VC (33 nm) and Cu/RGO
(36 nm). Although C2H4 formation most likely occurs through
CO dimerization on (100) facets of different size Cu NPs [26,28],
smaller nanoparticles are characterized by a greater number of low
coordination sites such as corners, edges and defects which help
activate C2H4 formation and inhibit CH4 formation on (100) facets
[3,26,29,33]. Indeed, an increase in the C2H4:CH4 ratio from 1.3 to 2
at −2 V was observed with the decrease in Cu nanoparticle size from
19 to 12 nm in our prior work [9]. However, for a similar size OLC-
supported Cu NPs in the present study (13 nm) this ratio is much
greater (7.5 at −1.8 V). It is unlikely that this tremendous increase
in selectivity towards C2H4 production on Cu/OLC thin films is due
to a particle size effect.
Because of a difference in the pretreatment procedure, similar
size Cu NPs can have different morphology. Contrary to 12 nm VC-
supported Cu NPs in Ref. [9], 13 nm Cu NPs supported on OLC have
never been reduced by heat-treatment in reductive environment.
The heat-treatment may lead to a loss of some potentially active
sites such as grain boundaries, defects, and low coordination sites
(such as edges and corners). However, one would not expect such
a significant effect of morphology on the selectivity of the similar
size NPs towards the CO2 electroreduction to C2H4.
Fig. 9. Faradaic efficiencies for H2, CO and CH4 production on thin films of VC, SWNT,
OLC and RGO supports held at −1.6 V and −2.2 V. Conditions: CO2-saturated 0.1 M
KHCO3, 1600 rpm. GC sampling was performed at 25 min.
bined with the inhibiting effect of CO adsorption on the HER, may
explain such unusual behavior. As it is demonstrated in Fig. 9, cur-
rent on the supports held at −1.6 V is due to the HER and the CO2
electroreduction to CO, while the current at −2.2 V is due to the HER
and the CO2 electroreduction both to CO and CH4. At −1.6 V, none
of the supports demonstrate electrocatalytic activity towards the
CH4 generation. Activities towards the CO generation, expressed in
terms of Faradaic efficiencies, are low (less than 2.8%) for the VC,
SWNT and RGO supports. A striking difference between the OLC
vs other supports is its higher activity (18.4% FE) towards the CO
production accompanied by the lower activity towards the HER
(80%). The same trend is observed at a more negative potential of
−2.2 V. The OLC support demonstrates the highest activity towards
the CO production (37.2% FE) and the lowest activity towards the
HER (64.5%) vs other supports. Activity towards the CH4 genera-
tion is the lowest for the OLC support (0.4% FE). The highest activity
towards the CH4 production is demonstrated by the SWNT sup-
port (8.3% FE). Note, that none of the supports showed appreciable
activity towards the C2H4 generation.
Although the Cu catalyst nanoparticles altered the total
accessible surface arears of their respective supports, the mate-
rials retained their structure and interparticle morphology. Pore
size distributions (Fig. S6 in SI) show distributions of inter-
particle slit sizes of OLC and VC supports and demonstrate
no noticeable changes after introduction of Cu. Subsequently,
substrate-dependent surface processes, rather than pore confine-
ment, drove CO2 electroreduction and altered both the rate and
yield of its conversion to hydrocarbons.
Based on information above, we propose, that the superior selec-
tivity of the Cu/OLC catalyst towards the C2H4 production is due
to generation of additional CO species by the OLC support. CO
molecules generated on surfaces of OLCs, dimerize on Cu NP (100)
planes [28,36], and, subsequently, yield additional C2H4 molecules.
The fact that the FE for C2H4 production is close to zero at −1.4 V
(Fig. 6c), when no CO is produced by the OLC support (not shown in
the Figure), supports this hypothesis. CO adsorption on Cu nanopar-
ticles is likely facilitated by the fact that the HER is suppressed on
the OLC compared to the other three supports. One of the impor-
tant characteristics of the catalyst/solution interface for reactions
in which intermediate species are involved is their residence time
[31]. Due to unusual catalyst design, CO species produced by OLC
Finally, the effect of local pH needs to be considered, as higher
local pH was found to be beneficial for the C2H4 production [34,35].
The presence of carbon supports is expected to increase local pH
due to their activity both towards the HER and CO2 electroreduction
at potentials more negative than −1.2 V and −1.6 V, respectively.
Both these reactions are accompanied by release of OH- groups
leading to an increase in local pH:
2H2O + 2e− → H2 + 2OH−
CO2 + 6H2O + 8e− → CH4 + 8OH−
(1)
(2)
In order to elucidate the effect of carbon support on electrocat-
alytic activities of Cu NPs towards CO2 electroreduction to C2H4,
electrocatalytic activities of the standalone supports were evalu-
ated following the same protocol as for the supported Cu NPs.
Fig. 8 compares current vs time curves measured on thin films of
VC, SWNT, OLC and RGO supports (a) and supported Cu NPs (b) held
at −1.6 V and −2.2 V in aqueous solutions of CO2-saturated 0.1 M
KHCO3. As expected, current on VC, SWNT and OLC supports are
lower than on the supported Cu NPs. A different pattern is observed
for the RGO support. Current density on the RGO films exceeds the
current density on the films of the RGO-supported Cu NPs. This is
likely due to the fact that Cu NPs are wrapped by RGO and produce
little catalytic action, but they block the surface of the support and
make it less efficient towards the HER. Considering that the BET
surface area of this catalyst is low (37.3 2.4 m2/g), this effect, com-
Please cite this article in press as: O. Baturina, et al., Effect of nanostructured carbon support on copper electrocatalytic activity toward
CO2 electroreduction to hydrocarbon fuels, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.11.001

G Model
CATTOD-10436; No. of Pages9
ARTICLE IN PRESS
O. Baturina et al. / Catalysis Today xxx (2016) xxx–xxx
9
become trapped at the interface, which allows sufficient time for
their adsorption on the Cu NPs. In other words, the increase in selec-
tivity of the Cu/OLC catalyst towards C2H4 production is primarily
due to increase in CO surface concentration, rather than increase
in the amount of the available sites for the CO dimerization. The
importance of the last factor has been highlighted recently by Dutta
et al. [31].
In addition to increase in the CO surface concentration, CO2
reduction to CO is expected to lead to an increase in local pH, which
is also beneficial for C2H4 production. Both the CO2 reduction to CO
on the OLC support, and the reduction of CO produced by the OLC
to C2H4 on Cu NPs involve releasing one OH− group per electron:
Sang were supported by the Fluid Interface Reactions, Structures
and Transport (FIRST) Center, an Energy Frontier Research Cen-
ter funded by the U.S. Department of Energy, Office of Science,
and Office of Basic Energy Sciences. B. Dyatkin was also funded
by an NRC Research Associateship award. We are grateful to Dr.
Brad Pate from NRL Chemistry Division for his help in discussion of
XPS results, and Dr. Kateryna Artyushkova from University of New
Mexico for collecting XPS spectra for the Cu/SWNT catalyst.
The NRL researchers thank the Office of Naval Research for finan-
cial support.
Appendix A. Supplementary data
CO2 + H2O + 2e−→ CO + 2OH−
2CO + 6H2O + 8 e−→ C2H4 + 8OH−
(3)
(4)
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.cattod.2016.11.
001.
Although the HER on VC, SWNT and RGO supports will also lead
to the release of one OH− group per electron, due to unusual cat-
alyst design, effect of increase in local pH is expected to be more
pronounced on the Cu/OLC catalyst.
Detailed mechanism for the enhanced electrocatalytic activi-
ties of OLC nanoparticles towards the CO2 electroreduction to CO
remains to be fully understood. The closed structure and convex
surface with pentagons and other defects may be responsible for
this phenomenon [13,15,37,38]. Further computational and exper-
imental research is needed to fully understand the process and
improve the catalytic performance, based on this understanding.
References
[1] Y. Hori, Electrochemical CO2 reduction on metal electrodes, in: C.E.A. Vayenas
(Ed.), Modern Aspects of Electrochemistry, vol. 42, Springer New York, 2008,
p. 89.
[2] M. Gattrell, N. Gupta, A. Co, J. Electroanal. Chem. 594 (2006) 1.
[3] M.R. Goncalves, A. Gomes, J. Condeco, T.R.C. Fernandes, T. Pardal, C.A.C.
Sequeira, J.B. Branco, Electrochim. Acta 102 (2013) 388.
[4] W. Vilestich, H.A. Gasteiger, A. Lamm, H. Yokokawa, Y. Hori, Handbook of Fuel
Cells – Fundamentals, Technology and Applications, John Wiley & Sons, 2010.
[5] Y. Hori, A. Murata, R. Takahashi, J. Chem. Soc.- Faraday Trans. I 85 (1989) 2309.
[6] J. Medina-Ramos, R.C. Pupillo, T.P. Keane, J.L. DiMeglio, J. Rosenthal, JACS 137
(2015) 5021.
[7] H. Mistry, A.S. Varela, C.S. Bonifacio, I. Zegkinoglou, I. Sinev, Y.W. Choi, K.
Kisslinger, E.A. Stach, J.C. Yang, P. Strasser, B.R. Cuenya, Nat. Commun. 7
(2016).
4. Conclusions
[8] W.H. Wang, Y. Himeda, J.T. Muckerman, G.F. Manbeck, E. Fujita, Chem. Rev.
115 (2015) 12936.
[9] O.A. Baturina, Q. Lu, M.A. Padilla, L. Xin, W. Li, A. Serov, K. Artyushkova, P.
Atanassov, F. Xu, A. Epshteyn, T. Brintlinger, M. Schuette, G.E. Collins, ACS
Catal. 4 (2014) 3682.
[10] W.C. Conner, J.L. Falconer, Chem. Rev. 95 (1995) 759.
[11] N.A. Vacirca, J.K. McDonough, K. Jost, Y. Gogotsi, T.P. Kurzweg, Appl. Phys. Lett.
103 (2013).
Composite catalysts of Cu nanoparticles on nanostructured
SWNT, RGO and OLC supports demonstrate significantly different
electrocatalytic activities towards CO2 electroreduction to CH4 and
C2H4, both among themselves and when compared to a Vulcan
carbon support as a reference. While all the composite catalysts
exhibit higher activity towards C2H4 vs CH4 generation, the activity
towards CH4 generation is significantly suppressed on nanostruc-
tured supports. Onset potentials for C2H4 vs CH4 generation are
shifted positively by 200 mV for Cu/SWNT, Cu/RGO and Cu/OLC
composite catalysts. The same onset potential for CH4 and C2H4
generation is observed for the Cu/VC catalyst. The Cu/OLC catalyst
is found to be superior to the other two nanostructured catalysts in
terms of stability, activity and selectivity towards C2H4 generation,
with faradaic efficiency as high as 60% at −1.8 V vs Ag/AgCl. To the
best of our knowledge, this is the highest reported to date faradaic
efficiency (and selectivity) towards C2H4 generation on a Cu-based
catalyst. We propose that the unique Cu/OLC catalyst design offers
superior stability because the nanostructured OLC NPs forms a shell
around the Cu NPs, acting as a filter that protects the catalyst surface
from undesirable adsorbed species. The enhanced selectivity the
Cu/OLC catalyst is most likely related to the enhanced electrocat-
alytic activity of the OLC support towards the CO2 electroreduction
to CO. It is due to increase in CO surface concentration from the CO2
electroreduction on OLC nanoparticles adjacent to Cu NPs, rather
than increase in the number of available sites for the CO dimeriza-
tion on (100) planes of Cu NPs. The CO2 electroreduction to CO on
the OLC support following by the CO reduction to CH4 and C2H4
on the Cu NPs are expected to lead to an increase in the local pH,
which is also beneficial for C2H4 production.
[12] Y. Gogotsi, MRS Bull. 40 (2015) 1110.
[13] J.K. McDonough, Y. Gogotsi, Electrochem. Soc. Interface 22 (2013) 61.
[14] G. Tobias, L. Shao, C.G. Salzmann, Y. Huh, M.L.H. Green, J. Phys. Chem. B 110
(2006) 22318.
[15] J.K. McDonough, A.I. Frolov, V. Presser, J.J. Niu, C.H. Miller, T. Ubieto, M.V.
Fedorov, Y. Gogotsi, Carbon 50 (2012) 3298.
[16] V.N. Mochalin, I. Neitzel, B.J.M. Etzold, A. Peterson, G. Palmese, Y. Gogotsi, ACS
Nano 5 (2011) 7494.
[17] H. Geng, J.W. Heckman, W.P. Pratt, J. Bass, F.J. Espinosa, S.D. Conradson, D.
Lederman, M.A. Crimp, J. Appl. Phys. 86 (1999) 4166.
[18] S. Brunauer, P.H. Emmett, E. Teller, JACS 60 (1938) 309.
[19] G.Y. Gor, M. Thommes, K.A. Cychosz, A.V. Neimark, Carbon 50 (2012) 1583.
[20] F. Nasirpouri, Ionics 17 (2011) 331.
[21] H. Siegenthaler, K. Juttner, J. Electroanal. Chem. 163 (1984) 327.
[22] S. Trasatti, O.A. Petrii, Pure Appl. Chem. 63 (1991) 711.
[23] K. Kinoshita, Carbon: Electrochemical and Physicochemical Properties, Wiley,
1988.
[24] H.A. Gasteiger, S.S. Kocha, B. Sompalli, F.T. Wagner, Appl. Catal. B 56 (2005) 9.
[25] XPS Knowledge database, Thermo Avantage v.5.938 software. Copyright @
1999–2014, Thermo Fisher Scientific portions © Micro Focus Ltd, in.
[26] Y. Hori, I. Takahashi, O. Koga, N. Hoshi, J. Phys. Chem. B 106 (2002) 15.
[27] K.J.P. Schouten, Y. Kwon, C.J.M. van der Ham, Z. Qin, M.T.M. Koper, Chem. Sci.
2 (2011) 1902.
[28] K.J.P. Schouten, Z.S. Qin, E.P. Gallent, M.T.M. Koper, JACS 134 (2012) 9864.
[29] W. Tang, A.A. Peterson, A.S. Varela, Z.P. Jovanov, L. Bech, W.J. Durand, S. Dahl,
J.K. Norskov, I. Chorkendorff, PCCP 14 (2012) 76.
[30] C.W. Li, M.W. Kanan, JACS 134 (2012) 7231.
[31] A. Dutta, M. Rahaman, N.C. Luedi, P. Broekmann, ACS Catal. 6 (2016) 3804.
[32] A. Eilert, F.S. Roberts, D. Friebel, A. Nilsson, J. Phys. Chem. Lett. 7 (2016) 1466.
[33] Y. Hori, I. Takahashi, O. Koga, N. Hoshi, J. Mol. Catal. A: Chem. 199 (2003) 39.
[34] K.J.P. Schouten, E.P. Gallent, M.T.M. Koper, J. Electroanal. Chem. 716 (2014) 53.
[35] A. Murata, Y. Hori, Bull. Chem. Soc. Jpn. 64 (1991) 123.
[36] Y. Hori, A. Murata, R. Takahashi, S. Suzuki, J. Chem. Soc. Chem. Comm. (1)
(1988) 17.
Acknowledgements
[37] M.E. Plonska-Brzezinska, L. Echegoyen, J. Mat. Chem. A 1 (2013) 13703.
[38] Y.V. Butenko, L. Siller, M.R.C. Hunt, Carbon onions, in: Y. Gogotsi, V. Presser
(Eds.), Carbon Nanomaterials, 2nd edition, CRC Press, 2013, p. 279.
A portion of electron microscopy imaging was conducted at
the Center for Nanophase Materials Sciences (CNMS), Oak Ridge
National Laboratory, which is a U.S. Department of Energy, Office
of Science, user facility. B. Dyatkin, Y. Gogotsi, R.R. Unocic and X.
Please cite this article in press as: O. Baturina, et al., Effect of nanostructured carbon support on copper electrocatalytic activity toward
CO2 electroreduction to hydrocarbon fuels, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.11.001