Lawrence Berkeley National Laboratory
Recent Work
Title
Enhanced Electrochemical Methanation of Carbon Dioxide with a Dispersible Nanoscale
Copper Catalyst
Permalink
https://escholarship.org/uc/item/4882v99t
Journal
Journal of the American Chemical Society, 136(38)
ISSN
0002-7863
Authors
Manthiram, Karthish
Beberwyck, Brandon J.
Alivisatos, A. Paul
Publication Date
2014-09-24
eScholarship.org Powered by the California Digital Library
University of California
Enhanced Electrochemical Methanation of
Carbon Dioxide with a Dispersible Nanoscale
Copper Catalyst
Karthish Manthiram†⊥
∥
, Brandon J. Beberwyck‡⊥
∥
, and A. Paul
Alivisatos*§⊥
∥
†
Department of Chemical and Biomolecular Engineering,
‡
Department of
Materials Science and Engineering,
§
Department of Chemistry, and
⊥
Kavli
Energy Nanosciences Institute, University of California, Berkeley, California
94720, United States
∥
Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley,
California 94720, United States
DOI: 10.1021/ja5065284
alivis@berkeley.edu
•
Abstract
Although the vast majority of hydrocarbon fuels and products are presently
derived from petroleum, there is much interest in the development of routes for
synthesizing these same products by hydrogenating CO
2
. The simplest
hydrocarbon target is methane, which can utilize existing infrastructure for natural
gas storage, distribution, and consumption. Electrochemical methods for
methanizing CO
2
currently suffer from a combination of low activities and poor
selectivities. We demonstrate that copper nanoparticles supported on glassy
carbon (n-Cu/C) achieve up to 4 times greater methanation current densities
compared to high-purity copper foil electrodes. The n-Cu/C electrocatalyst also
exhibits an average Faradaic efficiency for methanation of 80% during extended
electrolysis, the highest Faradaic efficiency for room-temperature methanation
reported to date. We find that the level of copper catalyst loading on the glassy
carbon support has an enormous impact on the morphology of the copper under
catalytic conditions and the resulting Faradaic efficiency for methane. The
improved activity and Faradaic efficiency for methanation involves a mechanism
that is distinct from what is generally thought to occur on copper foils.
Electrochemical data indicate that the early steps of methanation on n-Cu/C
involve a pre-equilibrium one-electron transfer to CO
2
to form an adsorbed
radical, followed by a rate-limiting non-electrochemical step in which the
adsorbed CO
2
radical reacts with a second CO
2
molecule from solution. These
nanoscale copper electrocatalysts represent a first step toward the preparation of
practical methanation catalysts that can be incorporated into membrane-
electrode assemblies in electrolyzers.
•
Introduction
The conversion of CO
2
into hydrocarbons is an alternative route for synthesizing
fuels and feedstocks that are typically derived from oil or natural gas,
representing one potential strategy to store electrical energy derived from
intermittent sources of clean energy, such as wind and solar.(1, 2) Although
electrosynthetic pathways for converting CO
2
into hydrocarbon products are not
economically feasible at present,(3) expected decreases in the price of electricity
derived from clean energy sources(4) and policy changes regarding greenhouse
gas emissions(5) may alter the economics of reducing CO
2
dramatically. In fact,
growing use of intermittent renewable energy sources in certain regions has
accelerated the deployment of small-scale electrical energy storage systems,
including pilot plants for methanizing CO
2
.(6) These pilot plants utilize a two-step
process, in which electrical energy is used to power an electrolyzer that splits
water to produce hydrogen and oxygen. The hydrogen is then used in the
Sabatier reaction,(7) in which CO
2
and H
2
are reacted over a heterogeneous
nickel catalyst at temperatures of 250–400 °C and pressures of 1–80 bar to
produce methane, which can be injected into existing natural gas networks. A
single-step electrochemical process that can directly convert CO
2
to methane
under conditions of ambient pressure and temperature may represent an
attractive alternative.
Of the metals explored as catalysts for electrochemical CO
2
reduction,(8) the
most active and selective identified to date are gold, silver, and bismuth,(9-14)
which produce CO as their terminal product. Copper is attractive in comparison,
as it produces more reduced hydrocarbon products.(8, 15-17) One of the
hydrocarbon products formed on copper electrocatalysts is methane, which forms
through the following half-reaction:
(1)
Because the reaction involves eight electron-transfer steps at 0.17 V (all
potentials reported versus reversible hydrogen electrode (RHE)) that can easily
bifurcate to form a wide range of products, the process exhibits poor selectivity
for any single product, forming a mixture of methane, ethylene, hydrogen, carbon
monoxide, and formic acid.(18, 19) The highest Faradaic efficiencies for methane
reported to date are 64% on a (210) copper single crystal(18, 20) and 73% on an
electrodeposited copper electrode.(21) Although studies conducted on high-purity
foils, single crystals, and electrodeposited materials have served as useful
benchmarks and provide fundamental insights into how copper catalyzes the
reduction of CO
2
, these model materials are impractical for electrolyzers as they
have low surface areas, cannot be incorporated into the membrane electrode
assemblies(22) that are needed to achieve high current densities with low ionic
resistances, or are expensive. From the point of view of cost and ease of
manufacturing, highly dispersed nanoparticle catalysts are much better suited for
electrolyzers.(23) Here, we demonstrate that well-dispersed copper nanoparticles
supported on glassy carbon show high activities and Faradaic efficiencies for
methanation, comparable to those of much more expensive single-crystal
electrodes. Systematic studies of nanoparticle loading on the glassy carbon
support and electrochemical analysis indicate that the altered reactivity of the
copper nanoparticles is due to distinct catalytic sites present on isolated
nanoparticle catalysts supported on glassy carbon.
•
Results and Discussion
We colloidally synthesized copper nanoparticles capped with
tetradecylphosphonate of diameter 7.0 ± 0.4 nm (Figure 1A,B).(24) These
particles were spin-coated onto glassy carbon plates (Figure 1C), hereafter
referred to as n-Cu/C, which served as the working electrode in a three-electrode
setup containing CO
2
-saturated 0.1 M sodium bicarbonate electrolyte, pH 6.8. As
a control, we also used high-purity copper foils as the working electrode. All
current densities for nanoparticle electrodes are surface-area normalized.
Figure 1. Morphological evolution of copper nanoparticles during the course of
electrochemical CO
2
reduction. Transmission electron microscopy (TEM) images
of as-synthesized copper nanoparticles of diameter 7.0 ± 0.4 nm at (A) low
magnification and (B) high magnification, showing that the initial particles are
highly polycrystalline. (C) Scanning electron microscopy (SEM) of n-Cu/C
electrode, consisting of copper nanoparticles supported on glassy carbon
substrate. (D) SEM of the same n-Cu/C electrode following polarization for 10
min at −1.25 V under CO
2
electroreduction conditions, demonstrating that the
average particle diameter grows to 23 ± 8 nm. TEM images of copper
nanoparticle transferred from glassy carbon substrate onto TEM grid at (E) low
magnification and (F) high magnification, in which it is evident that the particles
that form under polarization are highly polycrystalline. (G) SEM of trimethylsilyl
chloride-treated n-Cu/C electrode prior to polarization, in which particles have an
average diameter of 52 ± 21 nm. (H) SEM of the same electrode following
polarization for 10 min at −1.25 V, in which the particles that form are 25 ± 8 nm
in diameter.
Morphological Evolution
During the course of electrochemical CO
2
reduction, the morphology of the
copper nanoparticles changes significantly, growing in size to 23 ± 8 nm in
diameter (Figure 1D). The nanoparticles that form are highly polycrystalline, as
revealed using high-resolution transmission electron microscopy (HR-TEM,
Figure 1E,F). We find that irrespective of the initial size of the nanoparticles on
glassy carbon, the particles evolve in size to form particles which are ∼25 nm in
diameter, even if we begin with larger particles. For instance, if we treat the
initially cast particles (Figure 1C) with trimethylsilyl chloride, the
tetradecylphosphonate ligand is stripped off of the surface of the particles,
causing the particles to ripen to a diameter of 52 ± 21 nm prior to polarization
(Figure 1G). These large, irregular particles then evolve in size and shape during
the course of electrochemical CO
2
reduction to form smaller, uniform, roughly
spherical particles which are 25 ± 8 nm in diameter (Figure 1H). Similar changes
in size are also observed in the absence of CO
2
(Figure S1). The morphological
evolution observed, which may be due to a combination of particle coalescence
and dissolution–redeposition, points toward the importance of verifying if size
distributions are maintained in studies of size-dependent electrocatalysis.(25)
Catalytic Behavior
Although the n-Cu/C electrodes and copper foil electrodes exhibit comparable
current densities at lower overpotentials, the current densities for n-Cu/C
electrodes are over twice as high at more reducing potentials (Figure 2A). Of this
increased current, a much greater fraction from the n-Cu/C electrode goes toward
methane compared to the copper foil (Figure 2B). The Faradaic efficiency for
methane is improved at more reducing potentials for n-Cu/C, reaching 76% at
−1.35 V. This is significantly higher than the Faradaic efficiency of 44% achieved
on a polycrystalline copper foil at the same potential (Figure 2B). The combined
enhancement in both the overall current density and Faradaic efficiency for
methanation on n-Cu/C leads to partial current densities for methane that are four
times higher for n-Cu/C compared to the copper foil at −1.35 V (Figure 2C).
