RES EARCH

◥ surface of the catalyst, in the absence of liquid

REPO R T electrolyte, thus overcoming the gas diffusion
limitations of low-temperature systems. How-
ELECTROCHEMISTRY ever, high-temperature conditions and the ab-
sence of liquid electrolyte have thus far limited
CO2 electrolysis to multicarbon products at activities CO2 reduction to the production to CO (26).

greater than 1 A cm−2

Here, we present a hybrid catalyst design
that, by decoupling gas, ion, and electron tran-
sport, enables efficient CO2 and CO gas-phase
F. Pelayo García de Arquer1*, Cao-Thang Dinh1*, Adnan Ozden2*, Joshua Wicks1,3*, electrolysis at current densities in the >1–A cm−2
Christopher McCallum2, Ahmad R. Kirmani4, Dae-Hyun Nam1, Christine Gabardo2, Ali Seifitokaldani1, regime to generate multicarbon products. We
Xue Wang1, Yuguang C. Li1, Fengwang Li1, Jonathan Edwards2, Lee J. Richter4, Steven J. Thorpe3, exploit an ionomer layer that, with hydropho-
David Sinton2†, Edward H. Sargent1† bic and hydrophilic functionalities, assembles
into a morphology with differentiated domains
Electrolysis offers an attractive route to upgrade greenhouse gases such as carbon dioxide (CO2) to that favor gas and ion transport routes, con-
valuable fuels and feedstocks; however, productivity is often limited by gas diffusion through a liquid formally, over the metal surface: Gas transport
electrolyte to the surface of the catalyst. Here, we present a catalyst:ionomer bulk heterojunction (CIBH) is promoted through a side chain of hydropho-

Downloaded from https://www.science.org at University of Nevada Reno on November 13, 2024

architecture that decouples gas, ion, and electron transport. The CIBH comprises a metal and a bic domains, leading to extended gas diffusion,
superfine ionomer layer with hydrophobic and hydrophilic functionalities that extend gas and ion whereas water uptake and ion transport occur
transport from tens of nanometers to the micrometer scale. By applying this design strategy, we through hydrated hydrophilic domains (Fig. 1D).
achieved CO2 electroreduction on copper in 7 M potassium hydroxide electrolyte (pH ≈ 15) with an As a result, the reaction interface at which these
ethylene partial current density of 1.3 amperes per square centimeter at 45% cathodic energy efficiency. three components come together—gaseous re-
actants, ions, and electrons—all at catalytically

T
active sites, is increased from the submicrometer
he electrochemical transformation of catalyst surface (Fig. 1A) (17–19). Gas reactant regime to the several micrometer length scale.
gases into value-added products using diffusion in the catalyst layer becomes the mass We began by modeling the available gaseous
renewable energy is an attractive route transport–limiting step in the cathode, as ob- reactant in different gas-phase electrolysis
to upgrade CO2 and CO into fuels and served in the oxygen reduction reaction (ORR) scenarios (Fig. 1, E and F), building on pre-
chemical feedstocks (1–4) based on hy- in fuel cells. To improve ORR performance, fuel- viously established models (27) (see methods
drocarbons. The success of the approach will cell catalyst layers are designed to balance for more details). We explored how catalyst
rely on continued improvements in energy ef- hydrophobicity to help expel water and hydro- performance toward gas electroreduction would
ficiency to minimize operating costs and on philicity to maintain sufficient ion conductivity. be modified as the availability of the gas reactant
increasing current density to minimize capital In contrast with oxygen reduction, which varied at the gas-electrolyte interface. To do so,
costs (5, 6). This will require catalysts that facil- generates water as a product, CO2 reduction we introduced an intermediate surface channel
itate adsorption, coupling, and hydrogenation requires water as a proton source for hydro- of 20-nm thickness between the catalyst and
via proton-coupled electron transfer steps (7–9). carbon production. Thus, the catalyst layer is the electrolyte with an in-plane gas diffusion
In these reactions, water-based electrolytes hydrophilic and fully hydrated during the coefficient (D) appreciably different from that
act both as a proton source and as the ion con- reaction. In this configuration, CO2 electro- of bulk electrolyte (D0). As D/D0 increases, gas
ductive medium (10). However, the solubility chemical reactions occur within a gas-liquid- flow is promoted through this layer until the
of these gases in water is limited, leading to solid three-phase reaction interface (Fig. 1B) gas is converted at the catalyst surface or dif-
constrained gas diffusion as gas molecules col- (20). This volume, in which gaseous reactants fuses into the electrolyte (Fig. 1F), potentially
lide or react with their environment (11). The and electrolytes coexist at catalyst electroactive enabling CO2 diffusion on the scale of several
diffusion length of CO2 can be as low as tens of sites, decays rapidly into the electrolyte, partic- micrometers; whereas, for a standard catalyst
nanometers in alkaline aqueous environments ularly at the high pH used in alkaline electrol- configuration, CO2 is available only within
(12). This has limited the productivity of cata- ysis. The decay is further increased at high about 1 mm (Fig. 1E). As the diffusion in the
lysts in aqueous cells to current densities in current densities because of local OH– gen- layer increases, so too does the current avail-
the range of tens of milliamperes per square eration (21). A large fraction of the catalyst able for the electrochemical conversion of the
centimeter due to mass transport (13–16). is in contact with electrolyte in which CO2 gas reactant (Fig. 1G). A similar trend holds for
In a gas-phase electrolyzer, catalyst layers are availability is limited by its solubility (<2 mM other reactant gases such as O2 (fig. S5).
deposited onto hydrophobic gas-diffusion layers at pH 15). Because hydrogen evolution is a We sought to design and implement such
so that gas reactants need to diffuse only short competing reaction with CO2 reduction in an enhanced transport system experimentally.
distances to reach electroactive sites on the a similar applied potential range, the large We turned our attention to perfluorinated
fraction of catalyst surface area exposed to sulfonic acid (PFSA) ionomers, which combine
CO2-depleted electrolyte promotes undesired hydrophobic and hydrophilic functionalities
1
Department of Electrical and Computer Engineering, H2 generation (Fig. 1C). Whereas recent ad- along with ion transport (28–30). We hypo-
University of Toronto, 35 St. George St., Toronto, Ontario
M5S 1A4, Canada. 2Department of Mechanical and Industrial
vances in gas-phase CO2 reduction have led to thesized that their controlled assembly into
Engineering, University of Toronto, 5 King’s College Rd., partial current densities for CO2 reduction of distinct hydrophobic and hydrophilic layered
Toronto, Ontario M5S 3G8, Canada. 3Department of ≈100 mA cm−2 (12, 22, 23), other liquid-phase domains would offer differentiated pathways
Materials Science & Engineering (MSE), University of
electrochemical technologies such as water whereby gas transport is promoted through
Toronto, 184 College St., Toronto, Ontario M5S 3E4, Canada.
4
Materials Science and Engineering Division, National electrolysis achieve multi-amperes per square the hydrophobic domains and water and ion
Institute of Standards and Technology (NIST), Gaithersburg, centimeter (24, 25). transport are facilitated by the hydrophilic do-
MD 20899, USA. High-temperature solid oxide electrolysis of- mains (31–36) (Fig. 2A).
*These authors contributed equally to this work.
†Corresponding author. Email: [email protected] (E.H.S.); fers a strategy to achieve CO2 reduction at high PFSA ionomers such as Nafion contain –SO3−
[email protected] (D.S.) current density: CO2 diffuses directly to the (hydrophilic) and –CF2 (hydrophobic) groups.

García de Arquer et al., Science 367, 661–666 (2020) 7 February 2020 1 of 6

RES EARCH | REPOR T

Nafion, a widely used material in fuel cells as a ferent loadings (36, 38, 39). The hydrophobicity ilar contribution of the different Cu planes and
catalyst binder and membrane material, exhibits of the catalysts before and after ionomer mod- PTFE backbone support. CIPH samples, in ad-
strong structure-function–dependent properties ification was characterized using static contact dition, revealed weak scattering at 1.2 Å−1 from
(28, 37). In a polar solvent (i.e., methanol), PFSA angles: These yielded similar values of ≈121° the amorphous PFSA phase. The crystalline
ionomers form colloids with hydrophilic –SO3− to 122° (fig. S8). Scanning electron microscopy PFSA is masked by the PTFE support at 1.28 Å−1
groups exposed to solvent (28). When this PFSA (SEM) images revealed a homogeneous, con- (28). Attempts to quantify ≈10-nm thin-film
ionomer solution is coated on the metallic cata- formal ionomer coating over the entire catalyst ionomers using grazing-incidence WAXS were
lyst surface, we expect a configuration in which (Fig. 2, C and D). Cryo-microtomed cross- unsuccessful. Neutron scattering has revealed
–SO3− is preferentially exposed to hydrophilic sectional transmission electron microscopy lamellar arrangements in comparably thin PFSA
polycrystalline metal surfaces and electrolyte pro- (TEM) images revealed the presence of a 5- to layers (37, 40, 41).
vides continuous percolating hydrophobic paths 10-nm continuous and conformal ionomer layer Seeking to characterize the CIPH and the
through –CF2 hydrophobic domains (Fig. 2B). (Fig. 2, E to G), establishing a catalyst:ionomer ionomer configuration in its hydrated condi-
Seeking to promote the exposure of SO3− planar heterojunction (CIPH). tion, we designed a suite of ex situ and in situ
groups toward catalyst and electrolyte sur- To characterize the CIPH structural config- surface-enhanced Raman spectroscopy (SERS)
faces, we prepared ionomer solutions in polar uration, we carried out wide-angle x-ray scat- experiments (Fig. 2I and fig. S10). As-deposited
solvents, which we then spray-coated onto hy- tering (WAXS) measurements on PTFE/Cu/ ionomers on Ag catalysts exhibited strong char-
drophilic metal catalysts deposited on a porous ionomer samples (Fig. 2H and fig. S9). Both acteristic signals at 733 cm−1 (characteristic

Downloaded from https://www.science.org at University of Nevada Reno on November 13, 2024

polytetrafluoroethylene (PTFE) substrate at dif- reference and CIPH samples exhibited a sim- of –CF2 and C–C vibrations, table S5) and at

Fig. 1. Limiting current in gas-phase electrocatalysis and ionomer gas-liquid decoupled, the three-phase reaction interface can be extended so that all electrons
decoupled transport channels. (A) Flow-cell schematic. Reactant gases are fed participate in the desired electrochemical reaction. (E and F) Modeled gas reactant
through the back of a gas diffusion–electrode catalyst, facing an aqueous electrolyte. availability along the catalyst’s surface for standard (E) and decoupled (F) gas
An anion-exchange membrane (AEM) facilitates OH− transport from cathode transport into a 5 M KOH electrolyte, assuming an in-plane laminar gas diffusivity of
to anode. GDL, gas-diffusion layer. (B) In a gas-diffusion electrode (GDE), catalysts D||/DKOH = 1000 for the latter, where D|| is gas diffusivity parallel to catalyst surface.
are deposited onto a hydrophobic support from which gas reactants [G] diffuse. Depending on the gas diffusivity within the gas transport channel, gas availability
(C) The volume in which gas reactants, active sites, and water and ions coexist dramatically increases. L||, distance parallel to catalyst surface; L?, distance
determines the maximum available current for gas electrolysis. Catalyst regions with perpendicular to catalyst surface. (G) Modeled maximum available current density
limited reactant concentration promote by-product reactions such as hydrogen for CO2 reduction. D/DKOH manipulation enables entrance into the >1–A cm−2 regime
evolution. (D) When gas and electrolyte (water and ion source) transport is for CO2R. See methods for details on gas transport and reaction simulations.

García de Arquer et al., Science 367, 661–666 (2020) 7 February 2020 2 of 6

RES EARCH | REPOR T

1005 and 1130 cm−1 (associated with –SO3− The lack of a competing reaction to the ORR fusion of O2 relative to bulk electrolyte based on
modes) as well as a complex background set of at potentials more positive than the hydrogen Knudsen diffusion of the reacting gas through
features arising from other C–C (1386 cm−1), evolution reaction (HER) can be used to iden- CIPH hydrophobic domains (see methods).
C–F (1182 and 1300 cm−1), and S=O (1446 cm−1) tify gas-reactant depletion and its impact on the To assess whether ion transport was modi-
modes (42, 43). Hydrated samples retain char- limiting current. We built CIPH structures fied in metal-ionomer catalysts, we compared
acteristic –CF2, C–C and –SO3− spectral features consisting of spray-cast ionomer coatings over the ORR and HER performance of standard
but a notably increased relative contribution of PTFE/Ag substrates at different loadings, and Ag and Ag-CIPH samples for various electro-
–SO3− groups (1009 and 1131 cm−1) compared we monitored the ORR current (fig. S12) using lytes. Because the reactant in HER is in the
with –CF2 (730 cm−1). This trend is maintained a 5 M KOH water electrolyte and air as re- aqueous phase (water or hydrated proton), the
during operation in 1 M KOH electrolyte at actants. Unmodified Ag catalysts showed a cur- performance of the catalyst is not affected by
−2 V versus Ag or AgCl reducing potentials rent density limited to less than 30 mA cm−2. the gas-diffusion properties of the PFSA ionomer
and is also retained with the use of other cata- CIPH catalysts, on the other hand, exhibited layer; instead, catalyst performance is deter-
lyst metals such as Cu (fig. S11), suggesting that a considerably enlarged current density that mined by water availability and ion transport.
hydrated –SO3− groups tend to face the electro- peaks at 250 mA cm−2 under the same con- We found that CIPH samples exhibit similar
catalyst surface. ditions (Fig. 3A), with no H2 production ob- hydrogen evolution activity to bare catalysts
To assess the impact of the ionomer on gas served. In situ Raman measurements showed and increased ORR current across a wide range
availability, we evaluated the electrochemical a consistent increase in the presence of O2 near of electrolytes and pH (figs. S14 to S17). This

Downloaded from https://www.science.org at University of Nevada Reno on November 13, 2024

performance of the CIPH for different metals the catalyst surface at operating conditions result supports the notion that the enhanced
and reactions in 5 M KOH electrolyte (Fig. 3). (fig. S13). The observed enhancement can be ORR performance of the CIPH samples stems
In the ORR, oxygen is reduced into water (2). explained as being due to ≈600× increased dif- from extended gas transport.

Fig. 2. The catalyst:ionomer planar heterojunction. (A) Schematic of metal images of ionomer-coated copper catalysts. (E to G) Cryo-microtomed TEM
catalyst deposited onto a PTFE hydrophobic fiber support. A flat ionomer layer cross-sections of catalyst and ionomer revealing a laminar conformal overcoating.
conformally coats the metal. (B) Perfluorinated ionomers such as Nafion exhibit (H) WAXS spectra for reference and ionomer-modified catalysts. These reveal
differentiated hydrophilic and hydrophobic characteristics endowed by –SO3– features at 1, 1.28, and 2 A−1, associated with various PFSA and PTFE-support
and –CF2 functionalities, respectively. Laminar Nafion arrangements have been phases. (I) Raman spectra of reference and ionomer-modified catalysts revealing
reported depending on its thickness and substrate (37, 40). (C and D) SEM distinctive features of ionomer –CF2 and –SO3− groups (table S5).

García de Arquer et al., Science 367, 661–666 (2020) 7 February 2020 3 of 6

RES EARCH | REPOR T

We then focused on studying the perform- CO2RR current of 50 mA cm−2. This perform- the PFSA ionomer in the CIPH after reaction
ance of CIPH samples for CO2 and CO re- ance is consistent with the increased presence (fig. S22).
actants, investigating their reduction toward of adsorbed CO intermediates, as observed To query the impact of CIPH when applied
different products. We first screened Ag-CIPH using in situ Raman spectroscopy at similar to other gas reactants, we monitored the CO
samples for a CO2 reduction reaction (CO2RR) conditions (fig. S13) (46). Based on the model reduction reaction (CORR) on Cu under sim-
targeting CO production (7, 44) and observed a presented herein, the observed enhancement ilar reaction conditions—a system with activ-
CO2RR partial current density of 400 mA cm−2 can be explained by ≈400× increased diffusion ity limited by the poor solubility of CO in the
(Fig. 3B and fig. S18). By contrast, Ag reference of CO2 relative to bulk electrolyte. electrolyte (Fig. 3D and fig. S23). Cu-CIPH sam-
samples, limited by CO2 availability, exhibited The electrochemical surface areas of refe- ples yielded a CORR to ethylene partial current
a maximum CO2RR partial current density of rence and CIPH samples, as well as cell re- density of up to 340 mA cm−2. Bare Cu samples,
≈54 mA cm−2. This trend is maintained across sistances, were comparable (see methods), by contrast, showed a CORR limiting current of
different electrolytes (figs. S19 and S20). indicating that these were not causes of the 64 mA cm−2.
These observations translate as well to Cu- observed enhancement. These conclusions are To study the effect of the ionomer on the
CIPH catalysts targeting hydrocarbon genera- further supported by the similar hydropho- kinetics of the reaction, which could lead to
tion (Fig. 3C and fig. S21) (45). Cu-CIPH catalysts bicity of the catalysts before and after addition the difference in partial current densities ob-
exhibited a notable increase of CO2RR current. of the ionomer (fig. S8), consistent with the served, we carried out both ORR and CO2RR
At 800 mA cm–2, H2 generation remained below view that the enhanced gas reduction in CIPH in aqueous H-cell reactors. In this configura-

Downloaded from https://www.science.org at University of Nevada Reno on November 13, 2024

10% Faradaic efficiency (FE), whereas the FE samples originates from the extended gas dif- tion, gas transport to the entire surface of the
toward ethylene (C2H4) surpassed 60%. A CO2 fusion through the ionomer layer, rather than catalyst takes place through the electrolyte.
partial current density toward CO and ethylene from a redistribution of the gas or electrolyte In ORR, we observed a slight improvement
of 510 mA cm−2 was achieved (Fig. 3C). Bare Cu in the PTFE substrate pores. Postreaction in reaction kinetics, as indicated by a higher
catalysts, on the other hand, exhibited a limited SEM revealed the unmodified presence of current density at low overpotential for CIPH

Fig. 3. Increased limiting current and underlying mechanisms for CIPH (COR), Jlim ≈ 64 mA cm−2 for standard Cu, whereas CIPH achieves a maximum
catalysts. (A) ORR showing a 30–mA cm−2 limiting current (Jlim) for Ag 340–mA cm−2 current for the same reaction; H2 by-product generation is
reference catalysts as opposed to 250 mA cm−2 for a CIPH configuration. RHE, restrained below 15% FE at all currents. (E and F) Partial pressure COR studies in
reversible hydrogen electrode. (B) For CO2RR, standard Ag catalysts yield a CO|N2 mixes for CIPH (E) and standard (F) catalyst show that only at partial
Jlim of ≈54 mA cm−2 (remaining current used for hydrogen evolution). This is in pressures below 60% is Jlim observed for CIPH, whereas a sharp, steady
stark contrast with CIPH samples, which retain a FE above 85% for CO2 reduction decrease is observed for reference samples. At all partial pressures, CIPH
(CO2R) to CO up to ≈500 mA cm−2. (C) This trend is maintained for Cu CIPH exhibits an order of magnitude larger Jlim. Both reference and CIPH samples
catalysts and hydrocarbon production: Jlim toward ethylene (dominant product) exhibit comparable resistance and double-layer capacitance. Electrochemical
is 50 mA cm−2 at −0.7 V versus RHE for bare Cu but increases beyond experiments were carried out in 5 M KOH electrolyte with a 50–cm3 min−1 CO or
0.5 A cm−2 for CIPH (peak FE of 61% at 835 mA cm−2). (D) For CO reduction CO2 feedstock (in the case of 100% partial pressure).

García de Arquer et al., Science 367, 661–666 (2020) 7 February 2020 4 of 6

RES EARCH | REPOR T

samples (fig. S24). In CO2RR on Ag-based cata- properties that enable overcoming the lim- hancing reactant availability and thereby in-
lyst, both bare Ag and Ag-CIPH showed com- iting current density in gas-phase electrolysis. creasing current density.
parable current densities (fig. S25), with a slight To explore further the role of gas availability In light of these findings, we sought to
increase in CO FE (≈5%) for Ag-CIPH samples in the limiting current, we varied the gas avail- develop a catalyst design that took advantage
at low current density (<40 mA cm−2), a finding ability by tuning the partial pressure of the of the gas-electrolyte segregated transport be-
attributable to a change in local environment reactant in N2 mixtures (Fig. 3, E and F). A yond two dimensions. Ideally, such a catalyst
induced by the Nafion layer. No change in oxi- steep partial pressure dependence of limiting would maximize the triple-phase reaction in-
dation or coordination number of the metal current density for ethylene was observed in terface across an extended three-dimensional
active sites was observed during in situ x-ray CORR on Cu. Only at partial pressures below (3D) morphology, enabling efficient operation
absorption spectroscopy (XAS) (fig. S26). 60% was a limiting current observed for CIPH. in higher current regimes. We implemented
In the H-cell configuration, we observed At all CO partial pressures, Cu-CIPH exhibited a 3D catalyst:ionomer bulk heterojunction
similar limiting current densities for bare and an order of magnitude higher partial current (CIBH) consisting of Cu nanoparticles and PFSA
CIPH samples in ORR and CO2RR. These re- density compared with bare Cu. We observed blended and spray-cast on a PTFE/Cu/ionomer
sults indicate that although the presence of a similar trend in CO2RR with varying CO2 par- (CIPH) gas-diffusion layer support, forming
Nafion on the surface can change the reaction tial pressure (figs. S27 and S28). These results a 3D morphology with metal and ionomer per-
kinetics (47), it is its extended gas-transport further confirm the role of the ionomer in en- colation paths (Fig. 4A). Cross-sectional SEM

Downloaded from https://www.science.org at University of Nevada Reno on November 13, 2024

Fig. 4. 3D catalyst:ionomer bulk heterojunction for efficient gas-phase density at cathodic energy efficiencies >40% (fig. S30). Each sample and operating
electrochemistry beyond 1 A cm−2. (A) Schematic representation of metal- condition ran for at least 30 min. (F) Performance statistics of the highest
ionomer bulk heterojunction catalysts on a PTFE support. (B) Cross-sectional partial current configuration for eight Cu CIBH catalysts. The box plot corresponds
SEM of the CIBH catalyst. (C and D) TEM image of a cryo-microtomed CIBH (C) to Q1 to Q3 interquartile range, median, and average. The error bar represents
and elemental mapping of Cu and C revealing CIBH nanomorphology (D). ≈5.4 standard deviations. EE1/2, half-cell (cathodic) energy efficiency.
(E) Partial current density for total CO2RR reactions, with C2+ and C2H4 at (G) Performance of the best CIBH catalyst in an ultraslim flow cell consisting
maximum cathodic energy efficiency. The total CO2R current saturates at of a 3-mm-wide catholyte channel. A full-cell energy efficiency of 20% for C2+
1.3 A cm−2 before cathodic energy efficiency drops for CIBH thicknesses beyond products is estimated at 1.1–A cm−2 operating current. All CIBH electrochemical
6 mm. CIBH samples achieve more than a sixfold increase in partial current experiments were carried out in 7 M KOH with a 50–cm3 min−1 CO2 feedstock.

García de Arquer et al., Science 367, 661–666 (2020) 7 February 2020 5 of 6

RES EARCH | REPOR T

images revealed the different layers in the lates into 3D CIBH catalysts, which show im- 35. L.-C. Weng, A. T. Bell, A. Z. Weber, Phys. Chem. Chem. Phys.
CIBH catalyst (Fig. 4B). High-resolution cryo- proved performance arising from the combi- 20, 16973–16984 (2018).
36. R. Subbaraman, D. Strmcnik, V. Stamenkovic, N. M. Markovic,
microtomed cross-section images obtained using nation of accelerated CO2 reduction kinetics J. Phys. Chem. C 114, 8414–8422 (2010).
TEM and elemental energy-dispersive x-ray and extended mass transport through the 37. S. C. DeCaluwe, A. M. Baker, P. Bhargava, J. E. Fischer,
spectroscopy mapping further revealed the ionomer layer with increasing temperature J. A. Dura, Nano Energy 46, 91–100 (2018).
38. C.-H. Ma et al., Polymer (Guildf.) 50, 1764–1777 (2009).
presence of continuous Cu nanoparticle and (fig. S35). As a result, CIBH catalysts achieve 39. J. H. Lee et al., Sci. Rep. 8, 10739 (2018).
ionomer domains (Fig. 4, C and D). ≈1 V reduced overpotential and more than a 40. S. C. DeCaluwe, P. A. Kienzle, P. Bhargava, A. M. Baker,
We first optimized CIBH morphology by 50% increase in C2 productivity when oper- J. A. Dura, Soft Matter 10, 5763–5776 (2014).
41. A. Kusoglu et al., Adv. Funct. Mater. 24, 4763–4774 (2014).
tuning the deposition conditions as well as the ated at industrial electrolyzer-relevant tem- 42. J. Zeng, D. I. Jean, C. Ji, S. Zou, Langmuir 28, 957–964 (2012).
Cu:ionomer blend ratio, which we found opti- peratures of 60°C in a full-cell configuration, 43. X. Ling, M. Bonn, S. H. Parekh, K. F. Domke, Angew. Chem. Int. Ed.
mized for a 4:3 weight/ by weight configura- compared with the case of room temperature 55, 4011–4015 (2016).
44. C.-T. Dinh, F. P. García de Arquer, D. Sinton, E. H. Sargent,
tion. Using this configuration, with 7 M KOH operation (fig. S36).
ACS Energy Lett. 3, 2835–2840 (2018).
electrolyte and 50 cm3 min−1 of CO2 flow, we The phenomena described herein showcase 45. Y. Lum, J. W. Ager, Nat. Catal. 2, 86–93 (2019).
then explored the effect of catalyst layer thick- catalyst design principles that are not con- 46. M. C. Figueiredo, I. Ledezma-Yanez, M. T. M. Koper, ACS Catal.
ness. In an effective CIBH catalyst, CO2RR cur- strained by prior gas-ion-electron transport re- 6, 2382–2392 (2016).
47. J. H. Lee et al., Adv. Funct. Mater. 28, 1804762 (2018).
rent is expected to increase with catalyst loading strictions. The CIBH catalyst paves the way to 48. T. Burdyny, W. A. Smith, Energy Environ. Sci. 12, 1442–1453
until the length of the gas percolation paths the realization of renewable electrochemistry (2019).

Downloaded from https://www.science.org at University of Nevada Reno on November 13, 2024

through the ionomer phase reaches the gas for hydrocarbon production at operating cur- 49. J. Lu et al., Sci. Adv. 4, eaar5100 (2018).
50. F. P. García de Arquer, Arquer_et_al_microscopies_aay4217.zip,
reactant diffusion length. As we increased rents needed for industrial applications, as figshare (2020); doi: 10.6084/m9.figshare.11406927.
catalyst loading and corresponding thickness, has been achieved with syngas for solid oxide
we observed a monotonic increase in the total electrolyzers (48, 49). AC KNOWLED GME NTS

CO2RR current, which surpassed 1 A cm−2 for The authors thank D. Kopilovic and R. Wolowiec for electrochemical
cell design and setup; Z. Wang and Y. Lum for assistance and
a loading of 3.33 mg cm−2 (5.7 mm thickness) useful discussions; the Ontario Centre for the Characterization of
RE FERENCES AND NOTES
and which saturated at 1.32 A cm−2 for higher 1. S. I. Seneviratne, M. G. Donat, A. J. Pitman, R. Knutti,
Advanced Materials (OCCAM) for sample preparation and
loadings before energy efficiency dropped (Fig. characterization facilities; and T. P. Wu, Y. Z. Finfrock,
R. L. Wilby, Nature 529, 477–483 (2016).
G. Sterbinsky, and L. Ma for technical support at 9-BM beamline of
4E). The total partial current for C2+ products 2. Z. W. Seh et al., Science 355, eaad4998 (2017).
APS. J.W. gratefully acknowledges financial support from the
3. S. Lin et al., Science 349, 1208–1213 (2015).
(ethylene, ethanol, acetate, and propanol) re- Ontario Graduate Scholarship (OGS) program. A.S. thanks Fonds
4. C. Liu, B. C. Colón, M. Ziesack, P. A. Silver, D. G. Nocera,
ached 1.21 A cm−2 (fig. S29), which was achieved Science 352, 1210–1213 (2016).
de Recherche du Quebec–Nature et Technologies (FRQNT) for
support in the form of a postdoctoral fellowship award. C.-T.D. and
at a 45 ± 2% cathodic energy efficiency. The 5. S. Verma, B. Kim, H.-R. M. Jhong, S. Ma, P. J. A. Kenis,
A.S. are currently affiliated with Queen’s University and McGill
achieved C2+ partial current density represents ChemSusChem 9, 1972–1979 (2016).
University, respectively. Certain commercial equipment,
6. M. Jouny, W. Luc, F. Jiao, Ind. Eng. Chem. Res. 57, 2165–2177
a sixfold increase compared with previous best (2018).
instruments, or materials are identified in this paper in order to
reports at similar energy efficiencies (12, 22, 23) 7. Y. Hori, in Modern Aspects of Electrochemistry (Springer,
specify the experimental procedure adequately. Such identification
is not intended to imply recommendation or endorsement by the
(fig. S30 and tables S6 to S9). 2008), pp. 89–189.
National Institute of Standards and Technology, nor is it intended
8. R. Kortlever, J. Shen, K. J. P. Schouten, F. Calle-Vallejo,
The product distribution for optimal CIBH M. T. M. Koper, J. Phys. Chem. Lett. 6, 4073–4082 (2015).
to imply that the materials or equipment identified are necessarily
catalysts at different current densities in 7 M the best available for the purpose. Funding: This work was
9. Y. Zhang et al., ACS Catal. 7, 4846–4853 (2017).
financially supported by the Ontario Research Foundation,
KOH electrolyte reveals that H2 generation re- 10. V. R. Stamenkovic, D. Strmcnik, P. P. Lopes, N. M. Markovic,
Research Excellence Program; the Natural Sciences and
mains below 10% from 0.2 to 1.5 A cm−2 (fig.
Nat. Mater. 16, 57–69 (2017).
Engineering Research Council (NSERC) of Canada; the CIFAR
11. M. L. Perry, J. Newman, E. J. Cairns, J. Electrochem. Soc. 145,
S29). At the highest current operation, optimized Bio-Inspired Solar Energy program; and TOTAL S.A. This research
5 (1998).
used resources of the National Synchrotron Light Source II, which
catalysts exhibited a maximum productivity 12. C.-T. Dinh et al., Science 360, 783–787 (2018).
is a U.S. Department of Energy (DOE) Office of Science Facility,
13. Y. Chen, C. W. Li, M. W. Kanan, J. Am. Chem. Soc. 134,
toward ethylene with a FE in the 65 to 75% range, 19969–19972 (2012).
operated at Brookhaven National Laboratory under contract no.
a peak partial current density of 1.34 A cm−2 at 14. C. W. Li, J. Ciston, M. W. Kanan, Nature 508, 504–507 (2014).
DESC0012704, and synchrotron resources of the Advanced Photon
Source (APS) (XAS measurements), an Office of Science User
a cathodic energy efficiency of 46 ± 3% (Fig. 4F 15. M. Liu et al., Nature 537, 382–386 (2016).
Facility operated for the U.S. DOE Office of Science by Argonne
16. Y. Lum, B. Yue, P. Lobaccaro, A. T. Bell, J. W. Ager, J. Phys. Chem. C
and figs. S31 and S32). We implemented the National Laboratory, and was supported by the U.S. DOE under
121, 14191–14203 (2017).
best CIBH catalyst in an ultraslim flow cell contract no. DE-AC02-06CH11357 and the Canadian Light Source
17. S. Verma, X. Lu, S. Ma, R. I. Masel, P. J. A. Kenis, Phys. Chem.
and its funding partners. Author contributions: A.R.K. is a guest
(with no reference electrode and a minimized Chem. Phys. 18, 7075–7084 (2016).
researcher. F.P.G.d.A., C.-T.D., A.O., and J.W. designed and carried
catholyte channel of ≈3 mm, with water oxi- 18. D. M. Weekes, D. A. Salvatore, A. Reyes, A. Huang,
out all electrochemical experiments. F.P.G.d.A. and C.-T.D.
C. P. Berlinguette, Acc. Chem. Res. 51, 910–918 (2018).
dized at a Ni foam anode), leading to an esti- 19. D. Higgins, C. Hahn, C. Xiang, T. F. Jaramillo, A. Z. Weber, ACS
designed all remaining experiments and characterizations. C.M.
carried out diffusion-reaction simulations. C.G. and F.P.G.d.A.
mated full-cell energy efficiency toward C2+ Energy Lett. 4, 317–324 (2019).
carried out Raman spectroscopies. A.S. carried out SEM imaging.
products of 20% at 1.1 A cm−2 without the bene- 20. M. Jouny, W. Luc, F. Jiao, Nat. Catal. 1, 748–755 (2018).
A.R.K. performed WAXS measurements. J.E. carried out contact
21. K. Yang, R. Kas, W. A. Smith, J. Am. Chem. Soc. 141,
fit of iR compensation (i, current; R, resistance) 15891–15900 (2019).
angle measurements. All authors discussed the results and
(Fig. 4G). CIBH catalyst current and FE re- assisted during manuscript preparation. F.P.G.d.A., C.-T.D., D.S.,
22. S. Ma et al., J. Am. Chem. Soc. 139, 47–50 (2017).
and E.H.S. supervised the project. Competing interests: F.P.G.d.
mained stable over the course of a 60-hour 23. J.-J. Lv et al., Adv. Mater. 30, e1803111 (2018).
A., C.-T.D., A.O., J.W., D.S., and E.H.S. have filed provisional patent
24. A. J. Martín, G. O. Larrazábal, J. Pérez-Ramírez, Green Chem.
initial study implemented in a membrane elec- 17, 5114–5130 (2015).
application no. 62930229 regarding catalyst:ionomer devices.
trode assembly configuration (fig. S33). Data and materials availability: All data are reported in the main
25. Z. Liu et al., J. CO2 Util. 15, 50–56 (2016).
text and supplementary materials. Raw microscopy images are
Although CO2 reduction kinetics improve 26. Y. Song, X. Zhang, K. Xie, G. Wang, X. Bao, Adv. Mater. 31,
available at figshare (50).
1902033 (2019).
with increasing temperature, alkaline electro- 27. T. Burdyny et al., ACS Sustain. Chem.& Eng. 5, 4031–4040 (2017).
lyzers manifest worsened CO2 availability as 28. A. Kusoglu, A. Z. Weber, Chem. Rev. 117, 987–1104 (2017). SUPPLEMENTARY MATERIALS
temperature increases, and this fact curtails 29. T. D. Gierke, G. E. Munn, F. C. Wilson, J. Polym. Sci., Polym.
science.sciencemag.org/content/367/6478/661/suppl/DC1
Phys. Ed. 19, 1687–1704 (1981).
reaction productivity. We explored the effect Materials and Methods
30. K.-D. Kreuer, Chem. Mater. 8, 610–641 (1996).
of temperature on planar CIPH metal:ionomer Figs. S1 to S36
31. F. I. Allen et al., ACS Macro Lett. 4, 1–5 (2015).
Tables S1 to S9
catalysts and observed that CIPH catalysts re- 32. J.-P. Melchior, T. Bräuniger, A. Wohlfarth, G. Portale,
References (51–75)
quire lower overpotentials to attain similar FE, K.-D. Kreuer, Macromolecules 48, 8534–8545 (2015).
33. K.-D. Kreuer, G. Portale, Adv. Funct. Mater. 23, 5390–5397 17 June 2019; resubmitted 22 September 2019
in contrast with planar reference catalysts (fig. (2013). Accepted 23 December 2019
S34), when operated at 60°C. This effect trans- 34. M. Schalenbach et al., J. Phys. Chem. C 119, 25145–25155 (2015). 10.1126/science.aay4217

García de Arquer et al., Science 367, 661–666 (2020) 7 February 2020 6 of 6