Oxygen-tolerant electroproduction of C2 products from simulated flue gas

Yi Xua, Jonathan P. Edwardsa, Junjie Zhonga, Colin P. O’Briena, Christine M. Gabardoa, Christopher McCalluma, Jun Liab, Cao-Thang Dinhb, Edward H. Sargentb and David Sinton*a
aDepartment of Mechanical and Industrial Engineering, University of Toronto, 5 King's College Road, Toronto, ON M5S 3G8, Canada. E-mail: sinton@mie.utoronto.ca
bDepartment of Electrical and Computer Engineering, University of Toronto, 10 King's College Road, Toronto, ON M5S 3G4, Canada

Received 23rd September 2019 , Accepted 23rd December 2019

First published on 23rd December 2019


The electroreduction of carbon dioxide (CO2) to C2 products is a promising approach to divert and utilize CO2 emissions. However, the requirement of a purified CO2 feedstock decreases the economic feasibility of CO2 electrolysis. Direct utilization of industrial flue gas streams is encumbered by low CO2 concentrations and reactive oxygen (O2) impurities. We demonstrate that pressurization enables efficient CO2 electroreduction of dilute CO2 streams (15% v/v); however, with the inclusion of O2 (4% v/v), the oxygen reduction reaction (ORR) displaces CO2 reduction and consumes up to 99% of the applied current in systems based on previously-reported catalysts. We develop a hydrated ionomer catalyst coating strategy that selectively slows O2 transport and stabilizes the copper catalyst. Applying this strategy, we convert an O2-containing flue gas to C2 products at a faradaic efficiency (FE) of 68% and a non-iR-corrected full cell energetic efficiency (EE) of 26%.



Broader context

The electrochemical conversion of CO2 into C2 chemicals from industrial point sources could reduce CO2 emissions while creating valuable large-market fuels and feedstocks. Most reports in the field of CO2RR have used pure CO2 streams; however, the capital and operational costs associated with purifying CO2 from industrial sources are significant and reduce the economic feasibility of CO2 electrolysis. Direct flue gas utilization addresses this concern; however, dilute CO2 concentrations and the presence of reactive impurities, especially O2, complicate the process. We demonstrate that pressurization overcomes the mass-transport limitations posed by dilute CO2 feeds, thereby enabling a more efficient conversion of CO2. When O2 is also present, the parasitic oxygen reduction reaction dominates CO2 conversion in systems based on previously reported catalysts. We then proceed to design an O2-tolerant hydrated ionomer catalyst support layer which selectively slows the ingress of O2, thereby enabling the selective and stable generation of C2 products. This report illustrates the potential for the direct electrocatalytic conversion of flue gas CO2 streams.

Introduction

The primary source of anthropogenic CO2 emissions is the combustion of fossil fuels (coal, natural gas, and petroleum). Major industrial point sources of CO2 are candidates for carbon capture and conversion, with a large fraction of point sources containing 5–15% v/v CO2 concentrations.1–3

The decreasing cost of renewable electricity, currently at US$0.03 kW h−1 from onshore wind power,4 is one key enabler of the electrochemical conversion of CO2.5,6 The electrochemical CO2 reduction reaction (CO2RR) has been shown to produce carbon monoxide,7–9 formic acid,10–12 methanol,13–15 ethylene,16–18 acetate,19 and ethanol.20,21 Ethylene and ethanol are of particular interest due to their large existing markets and high market values per ton.6,22 The market values for these commodities also set the total cost ceiling for commercial CO2RR, including the sum of input electricity, product separation, and source CO2 feedstock cost.

The majority of CO2RR studies have reacted pure CO2 streams;23,24 however, the cost to purify CO2 streams from flue gas sources is in the range of $70–$100 per ton of CO2 purified.25,26 Technoeconomic studies have demonstrated approximately 30% of the operating costs of a CO2 to ethylene plant will be from CO2 feedstock costs.6

Direct intake of flue gas as CO2RR feedstock could lower the system costs of a CO2 electrolyzer (Fig. 1a).27,28 Coal-fired power plant gases contain relatively low concentrations of CO2 in flue gas (15% v/v),1 and utilizing these streams directly in previously-reported CO2RR systems results in low CO2RR efficiency due to mass transport penalties.16,27,29 Certain oxidized components of flue gas, such as SOx and NOx, also influence CO2RR catalysts, but their concentrations are low (typically on the order of hundreds of ppm), and their effects are detailed elsewhere.30,31 O2, a major component of flue gas (4% v/v),1 is costly to remove32 and highly reactive on copper—the main C2-evolving CO2RR catalyst.33 Parasitic ORR is of particular concern since the thermodynamic voltage of ORR is more than 1 V more positive than CO2RR, and thus could displace CO2RR activity.34,35


image file: c9ee03077h-f1.tif
Fig. 1 Effect of pressure and the presence of O2 on product selectivity for CO2RR with flue gas CO2 concentrations on a Cu catalyst. (a) Schematic of electrochemical flue gas recycling. (b) The CO2RR FE for 15% CO2 (v/v) stream product with and without the presence of O2 at 1 bar and 15 bar. (c) The HER FE for 15% CO2 (v/v) stream product with and without the presence of O2 at 1 bar and 15 bar absolute pressure. (d) The total FE of both HER and CO2RR.

Here we present a strategy that directly converts simulated flue gas to C2 products in the presence of O2 impurities. We determine that pressurization enables the conversion of low CO2 concentrations, avoiding the parasitic H2 evolution reaction (HER). However, when O2 impurities are present, pressurization accelerates ORR, leading to a compromise between H2 and O2 evolution side-reactions – each of which competes with desired CO2RR. To slow O2 transport selectively and enhance the stability of copper (Cu) catalysts, we develop a catalyst comprised of an ionomer with hydrophilic nanopores and titanium dioxide (TiO2) nanoparticles. Operating this electrode structure at 10 bar, we convert O2-containing flue gas concentrations of CO2 to C2 products with 68% selectivity and 26% energetic efficiency, competitive with reactors that convert pure CO2.

Results and discussion

Electroreduction of dilute CO2 feedstocks

Employing a sputtered Cu catalyst for C2 production, we prepared an initial CO2 feedstock with argon as an inert diluent (15% CO2 v/v). In agreement with other reports,16,27,29 we found that decreasing the CO2 concentration from 100% to these flue gas levels significantly reduced the selectivity toward CO2RR products, especially at high current densities (Fig. 1b). Due to the low CO2 availability, the selectivity toward CO2RR products reduced to 36% FE at a full cell voltage of −2.50 V (compared to 81% FE at the same voltage using the 100% CO2 feedstock). Consistent with other reports,36 in the absence of CO2RR, electrons were diverted to the undesired HER: when reacting 15% CO2 (v/v) streams, the selectivity toward H2 exceeded 55% across cell voltages (Fig. 1c).

Pressurizing flue gas concentrations of CO2 increases CO2 availability at the catalyst by elevating the CO2 partial pressure. With the same 15% (v/v) CO2 feed gas at 15 bar, the selectivity toward CO2RR improved to 91%, while that of H2 decreased to less than 7% FE at −3.00 V (Fig. 1b and c). These results indicate that pressurization enables the direct conversion of streams with CO2 concentrations characteristic of major flue gas sources at industrially relevant current densities. While pressurization requires some energy input, the level of pressurization here is low by gas industry standards:37 the energy required to pressurize to 10 bar represents ∼3% of the energy required to perform efficient CO2RR to C2 products (calculations in ESI).

Influence of O2 impurities on CO2RR

O2 is the highest concentration reactive impurity in flue gas and is challenging to remove.32 The lower thermodynamic and kinetic requirements for ORR,38,39 compared to CO2RR,22 will greatly favour ORR on the cathode and thereby displace CO2RR.

To assess the effect of O2 impurities on CO2RR, we carried out experiments at 15 bar with the addition of O2 impurities (4% O2 v/v and 15% CO2 v/v diluted in argon). The O2 diminished the selectivity toward CO2RR (Fig. 1b) and HER (Fig. 1c). The dominant product of ORR is water, which is abundant in the aqueous electrolyte, and thus difficult to quantify as a reaction product. We associate the complement of the total selectivity (i.e. 100% less the total selectivity) with ORR selectivity. Only 9% of products could be quantified at −2.50 V when O2 was added (Fig. 1d, missing FE shading), suggesting that ∼90% of the electrons supplied to the reaction were diverted to the parasitic ORR.

To quantify the impacts of O2 on the pressurization strategy, we performed experiments at a range of pressures (1, 5, 10 and 15 bar), all with the same simulated flue gas mixture of CO2 (15% v/v) and O2 (4% v/v). The current densities were similar at all pressures for cell voltages between −2.50 V and −3.50 V (Fig. 2a). The FE toward HER was suppressed with increased flue gas pressure (Fig. 2b). Similar to the case without O2 impurities, the decrease in HER with pressure demonstrates that higher CO2 availability on the Cu catalyst is sufficient to suppress HER. However, the combined total selectivity of HER and CO2RR decreased with pressure when O2 was present (Fig. 2c). The increased ORR at higher pressures is attributed to the increased O2 partial pressure and the associated increase of O2 mass transport to the catalyst surface. While pressurization reduces undesired HER, it enhances parasitic ORR.


image file: c9ee03077h-f2.tif
Fig. 2 Pressure characterization for C2 product generation on a Cu catalyst in 1 M KOH with a gas feed of 15% CO2 (v/v) and 4% O2 (v/v). (a) Current–voltage characteristics as a function of pressure. (b) The FE toward HER at pressures of 1–15 bar. (c) The total FE of both HER and CO2RR products at pressures of 1–15 bar. (d) The FE of C2 products at pressures of 1–15 bar. (e) The EE and partial current density of C2 products at pressures of 1–15 bar. (f) The stability performance of bare Cu catalyst at −3.00 V cell voltage and 10 bar.

Peak C2 selectivities were seen at higher cell voltages when the pressure was increased (Fig. 2d). To maximize energy efficiency (EE), CO2 electrolyzers will operate at low cell voltages and high selectivities toward the products of interest (see ESI, for calculations). Both the 5 and 10 bar cases exhibit similar maximum EEs (non-iR compensated full cell) of approximately 23%, but only 10 bar operation allowed desirably high partial current densities (>150 mA cm−2) (Fig. 2e). This intermediate pressure was sufficient to reduce the unwanted side reactions (ORR at high pressures and HER at low pressures) and enable higher CO2RR selectivities for C2 products over short run times. Over longer 90 minute runs at these conditions (10 bar, −3.00 V), the selectivity toward C2 products declined rapidly, while that toward H2 and C1 products increased progressively (Fig. 2f), and the current was relatively steady (Fig. S1, ESI) – a trend typical of Cu catalysts made using practices previously reported in literature.16

Ionomer design strategy for stable and efficient CO2 consumption from O2-containing streams

Catalyst supports have been shown to enhance electrochemical stability.40,41 We first took the approach of depositing a mixture of Nafion-bound carbon nanoparticles – as was found previously to enhance the stability of Cu catalysts under CO2RR conditions.16 Compared to the performance achieved using bare Cu, this support layer reduced the C2 production (Fig. S2a, ESI) and total detectable products, with up to 99% missing FE attributed to ORR (Fig. S2b, ESI). These results indicate that the established Nafion-carbon Cu support layer fails with the use of O2-containing CO2 streams. We hypothesized that a stabilizing layer that selectively impeded O2 transport could allow the direct consumption of flue gas streams (Fig. 3a).
image file: c9ee03077h-f3.tif
Fig. 3 Effect of support layer ionomer on the performance of pressurized (10 bar) CO2RR with 15% CO2 (v/v) and 4% O2 (v/v) feedstocks, 1 M KOH electrolyte, and TiO2 support particles. (a) Schematic illustration of the Cu-PTFE GDE. (b) High magnification schematic of the GDE coated with the hydrophobic nanoporous ionomer. (c) High magnification schematic of the GDE coated with the hydrophilic nanoporous ionomer. (d) Current–voltage characteristics with different ionomer coatings on Cu-PTFE GDE. (e) The total FE of both HER and CO2RR products. (f) Missing current density for the different ionomers calculated using the complement of the total FE. (g) The FE toward C2 products for different ionomers.

Ionomers used as binders in CO2 electrolysis can be divided into two categories based on their nanoporous wetting properties: hydrophobic and hydrophilic. Hydrophobic nanopores provide a gas-phase diffusive pathway for CO2 and O2 (Fig. 3b). The second category of ionomers provides a hydrophilic environment within the nanopore network. Ionomers with hydrophilic nanopores are expected to be completely filled with electrolyte during operation, forcing gas phase reactants to reach the catalyst only in dissolved form (Fig. 3c).

We hypothesized that ionomer selection could influence CO2RR efficiency in the presence of O2 impurities. To assess computationally the O2 transport in these two ionomer types, we modelled O2 mass transport for both cases (calculations in ESI). For the hydrophobic nanoporous network, O2 molecules must first degas from the surrounding electrolyte film before diffusing into the hydrophobic gas nanopores wrapped around the catalyst (Fig. 3b). Due to the volatile nature of O2 in aqueous solutions, the high degassing rate of O2 enables a relatively high mass flux in the hydrophobic nanopore networks (7.7 × 10−3 mol m−2 s−1) as quantified by the Wilke–Chang correlation.42 However, the O2 mass flux through an electrolyte-wet hydrophilic nanopore, as predicted from the Bosanquet relation,43 is one order of magnitude less (4.4 × 10−4 mol m−2 s−1). Hydrophilic nanopore networks thus present an opportunity to reduce O2 mass flux to the catalyst surface.

Although promising for reducing O2 transport, sufficient CO2 penetration is needed in these nanopore networks to perform CO2RR at the catalyst surface. To assess the CO2 transport in these two types of ionomer networks, we employed the same computational thermodynamic approach. We predicted that dissolved phase CO2 can be delivered to the catalyst surface through the hydrophobic nanopores with a mass flux (2.7 × 10−2 mol m−2 s−1) that is very similar to that achieved in hydrophilic nanopores (3.6 × 10−2 mol m−2 s−1). Unlike O2, the transport of CO2 is not greatly enhanced by the de-gassing phenomena found in the hydrophobic nanopores, a finding we ascribe to the relatively low volatility of CO2. The result is similar CO2 fluxes for both ionomer types. These flux estimations indicate that the electrolyte in the hydrophilic nanopores could selectively impede O2 mass transport without a significant impact on CO2 transport, suggesting a means to achieve a stable, O2-tolerant catalyst structure for CO2RR.

In order to test this mass transfer hypothesis, we carried out a group of ionomer characterization experiments at 10 bar (corresponding to peak C2 EE previously), using ionomers with different nanopore behaviour: two hydrophobic and two hydrophilic ionomers. Each ionomer was mixed with TiO2 nanoparticles, which are electrocatalytic support materials used in fuel cell and CO2RR applications.44–48 TiO2 is commonly integrated into ion exchange membranes to promote hydration and prevent dry-out.49,50 In keeping with our hypothesis to suppress O2 mass transport by encouraging hydrophilic nanopore networks, we selected TiO2 as the support particle due to its hydrophilic nature. The ionomer and nanoparticle mixture was airbrushed onto the Cu-based gas diffusion electrode (GDE, details in the experimental section, Fig. S3a (ESI) with cross-sectional image in Fig. S4, ESI).

The performance of the four ionomers correlated with their wetting properties. The ionomers with more hydrophilic nanopores showed lower total current densities (Fig. 3d) than those coated with hydrophobic nanopores. The total FEs of the hydrophilic samples were much higher than the hydrophobic samples, indicating less ORR for the hydrophilic ionomers (Fig. 3e). The hydrophilic samples exhibit even higher total FEs than the unstable bare Cu catalyst tested previously (Fig. S5b, ESI). At a cell voltage of −3.00 V the hydrophilic samples exhibited total FEs approximately 70% higher than hydrophobic samples.

To estimate the ORR partial current for the different ionomers, we plotted the missing current (Fig. 3f) using the complement of the total FE (Fig. 3e) and the total current density for each ionomer sample (Fig. 3d). For a given ionomer, the missing current remained relatively constant with cell voltage, suggesting that ORR was mass transport limited in the range of cell voltages tested (−2.50 to −3.50 V). Consistent with the trend in total FE, the hydrophilic ionomers allowed several times less ORR than the hydrophobic ionomers. This slowing of oxygen transport is largely independent of the ion conducting nature of the ionomer support layer since non-ionomer based hydrophobic layers also generated more missing current than hydrophilic ones (Fig. S6, ESI). To verify the model, ionomer-specific O2 mass fluxes were calculated from the experimental data (calculations in ESI). The experimentally-determined O2 mass fluxes are in agreement with the theoretical ones: the O2 mass flux in the hydrophilic nanopores is an order of magnitude less than that of the hydrophobic case (Table S1, ESI). Control experiments demonstrate that the limiting current density for CO2RR is similar for both ionomer families (Fig. S7, ESI), indicating that both ionomers possess similar CO2 transport behaviour in keeping with our hypothesis.

The ionomers with hydrophilic nanopores demonstrated the highest selectivity toward C2 products (Fig. 3g) due to reduced ORR activity (Fig. 3f). The hydrophilic samples demonstrated more than 60% higher FE toward C2 products compared to the hydrophobic samples operated at the same full cell voltage of −3.00 V. The highest value of FEC2, 68%, was achieved with Hydrophilic-1 (Sustainion) ionomer at −3.00 V corresponding to the highest EEC2 of 26% (total product distribution displayed in Table S2, ESI).

To investigate the stability of the new catalyst during prolonged operation, the TiO2 nanoparticle layer bound with Hydrophilic-1 ionomer was coated on the Cu-GDE and operated at 10 bar. The current density and selectivity toward C2 products were stable for 10 hours of operation at a cell potential of −3.00 V in 1 M KOH (Fig. 4). Scanning electron microscopy (SEM) images of the GDE before and after operation show minimal change in surface morphology (Fig. S3, ESI).


image file: c9ee03077h-f4.tif
Fig. 4 Stability performance test of CO2 reduction to C2 products for Cu-PTFE GDE coated with Hydrophilic-1 ionomer/TiO2 support particles at 10 bar absolute pressure in 1 M KOH electrolyte, utilizing 15% CO2 (v/v) and 4% O2 (v/v) mixture gas feedstocks.

Comparison to other reports

We report an EEC2 of 26% in the presence of a dilute CO2 stream with O2 impurities. This EEC2 is the highest reported for CO2 electroreduction at currents above 150 mA cm−2 (Fig. S8, ESI). The financial costs of pressurization to 10 bar ($15 per ton of CO2 equivalent calculations in ESI) are lower than the costs of CO2 purification ($100 per ton of CO2 equivalent).26,51

Conclusions

In summary, we developed a catalyst for O2-tolerant electroproduction of C2 products from simulated flue gas streams. Pressurization was necessary to react the dilute CO2 concentrations found in flue gas streams. However, in the presence of O2 impurities at typical concentrations, up to 99% of the current was diverted from CO2RR toward ORR – a trend exacerbated at higher pressures. While pressurization was required to minimize parasitic HER, it enhanced parasitic ORR. We designed a stabilizing support layer that selectively slowed O2 transport to the catalyst. The layer comprised ionomer containing hydrophilic nanopores bound with TiO2 nanoparticles atop the Cu catalyst. Leveraging the differences in gas volatility between CO2 and O2, this strategy converted O2-containing dilute CO2 feeds with a FEC2 of 68% and an EEC2 of 26% over 10 hours of stable operation. This performance is competitive with the best past studies employing pure reactant CO2, and suggests that this strategy effectively removes the penalty associated with low CO2 levels and O2 impurities in flue gas.

Experimental

Reagents

Potassium hydroxide (>85%) was purchased from Sigma Aldrich. All reagents were of analytical grade. No further purification was used. All solutions were prepared by utilizing Milli-Q grade DI water (18.2 MΩ).

Electrode preparation

Nafion and Aquivion ionomer (selected as Hydrophobic-1 and Hydrophobic-2 respectively) possess tetrafluoroethylene backbone groups (–CF2–CF2–) and are thus capable of creating strongly hydrophobic nanoporous networks.52,53 Sustainion and Fumion ionomer were selected as Hydrophilic-1 and Hydrophilic-2 respectively due to their hydrophilic environment within their nanopore network. Sustainion possesses a hydrophilic poly(4-vinylbenzyl alkyl-imidazolium chloride) unit,54,55 while Fumion contains a quaternization unit with strong hydrophilic properties.56–58

Slight differences in performance were observed amongst the same family of ionomers likely due to minor variations in molecular structure. For example, at larger cell voltages Hydrophobic-1 samples exhibited slightly higher total currents than Hydrophobic-2 samples (Fig. 3d). These differences in performance may stem from the side chain length, as documented for fuel cells.59,60

The polytetrafluoroethylene (PTFE) based electrode used was fabricated in a manner similar to previous reports.9,16,61 Approximately 300 nm of Cu catalyst was sputtered onto the PTFE substrate using an AJA International ATC Orion 5 Sputter Deposition System (Toronto Nanofabrication Centre, University of Toronto). The catalyst support layers were airbrushed by hand onto the sputtered Cu-PTFE electrode. Airbrush inks were prepared with 40 mg of support particle, 4 mg of ionomer, and 1.25 mL of methanol. Dispersions of Aquivion D79-25BS (Sigma-Aldrich 802565, 25 wt% in water), Fumion (Fuel Cell Store FAA-3-SOLUT-10, 10 wt% in N-methyl-2-pyrrolidone), Nafion (Sigma-Aldrich 527084, 5 wt% in aliphatic alcohols and water), and Sustainion (Dioxide Materials XA-9, 5 wt% in ethanol) were used as ionomer solutions. Titanium dioxide (Evonik Aeroxide TiO2 P25, 35–65 m2 g−1) and carbon black (Alfa Aesar 39724, 75 m2 g−1) were tested as support particles. The resulting mixture was then sonicated for 20 minutes prior to airbrushing. The samples were sprayed until a loading of 0.18 mg cm−2 was attained. Cathodes were characterized using SEM on a Hitachi S-5200 instrument (The Centre for Nanostructure Imaging, University of Toronto) apparatus. A commercially available nickel foam (MTI Corp. EQ-BCNF-80um) was used for the anode.

Operation of the electrochemical flow cell

The pressurized electrochemical flow cell was operated in a similar manner to previous reports.9 Teledyne Isco pumps were used to pressurize the gas and liquid electrolytes to the appropriate pressures. The pressure in the cell was maintained by a back-pressure regulator. Pressures reported in this report are absolute pressures. The electrolyte was pumped once through the cell at a rate of 0.7–1.3 mL min−1 while the gas flow rate, controlled via a needle valve downstream of the cell, was supplied at a rate of approximately 40 standard cubic centimetres per minute (sccm). A single electrolyte stream was supplied to the cell with no ion exchange membrane located between cathode and anode. Unless otherwise noted, electrochemical tests were performed by running one fresh cathode sample at multiple voltages of interest sequentially (−2.50 V, −2.75 V, −3.00 V, −3.25 V, −3.50 V). Experiments started at the lowest magnitude voltage and after approximately 10 minutes of operation and product sampling, the voltage was increased by 0.25 V and the process was repeated. The voltages reported are full cell voltages with no iR compensation. These sequential tests were repeated a second time with a fresh cathode sample to obtain two sets of data for each operating condition.

Product analysis

The gas products from CO2 reduction were analysed in 1 mL volumes using a gas chromatograph (PerkinElmer Clarus 680) possessing a thermal conductivity detector (TCD) and a flame ionization detector (FID). The gas chromatograph, using argon as the carrier gas (Praxair, 99.999%), was equipped with a Molecular Sieve 5A capillary column and a packed Carboxen-1000 column. The flow rate of the gas was measured before each 1 mL volume was collected. For tests at atmospheric conditions, the sampling current was taken to be the average current over the previous 60 s whereas for pressurized tests the current was averaged over the previous 3 minutes.

The liquid products were quantified using nuclear magnetic resonance spectroscopy (NMR). 1H NMR spectra of freshly acquired samples were collected on an Agilent DD2 500 spectrometer using water suppression mode with dimethyl sulfoxide (DMSO) as an internal standard. Ten-second relaxation time between the pulses was used to allow for complete proton relaxation.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors acknowledge support from the Natural Sciences and Engineering Research Council (NSERC) of Canada. Support from Canada Research Chairs Program is gratefully acknowledged, as is support from an NSERC E. W. R. Steacie Fellowship to D. S. Y. X. acknowledges NSERC for their support through graduate scholarships. J. P. E. thanks NSERC, Hatch, and the Government of Ontario for their support through graduate scholarships. C. P. O. thanks the Province of Ontario for their funding toward graduate scholarships. C. M. G. would like to thank NSERC for support in the form of a post-doctoral fellowship award. J. L. acknowledges the Banting Postdoctoral Fellowships. The authors acknowledge Centre for Nanostructure Imaging at the University of Toronto and Dr Ilya Gourevich for sample SEM characterization.

References

  1. D. M. D’Alessandro, B. Smit and J. R. Long, Angew. Chem., Int. Ed., 2010, 49, 6058–6082 CrossRef PubMed.
  2. Y. J. Heo and S. J. Park, Sci. Rep., 2017, 7, 1–9 CrossRef PubMed.
  3. M. E. Boot-Handford, J. C. Abanades, E. J. Anthony, M. J. Blunt, S. Brandani, N. Mac Dowell, J. R. Fernández, M.-C. Ferrari, R. Gross, J. P. Hallett, R. S. Haszeldine, P. Heptonstall, A. Lyngfelt, Z. Makuch, E. Mangano, R. T. J. Porter, M. Pourkashanian, G. T. Rochelle, N. Shah, J. G. Yao and P. S. Fennell, Energy Environ. Sci., 2014, 7, 130–189 RSC.
  4. I. Renewable Energy Agency, IRENA: Renewable Power Generation Costs in 2017, 2018.
  5. O. S. Bushuyev, P. De Luna, C. T. Dinh, L. Tao, G. Saur, J. van de Lagemaat, S. O. Kelley and E. H. Sargent, Joule, 2018, 2, 825–832 CrossRef CAS.
  6. M. Jouny, W. Luc and F. Jiao, Ind. Eng. Chem. Res., 2018, 57, 2165–2177 CrossRef CAS.
  7. B. A. Rosen, A. Salehi-khojin, M. R. Thorson, W. Zhu, D. T. Whipple, P. J. A. Kenis and R. I. Masel, Science, 2011, 334, 643–645 CrossRef CAS PubMed.
  8. C. E. Tornow, M. R. Thorson, S. Ma, A. A. Gewirth and P. J. A. Kenis, J. Am. Chem. Soc., 2012, 134, 19520–19523 CrossRef CAS PubMed.
  9. C. M. Gabardo, A. Seifitokaldani, J. P. Edwards, C.-T. Dinh, T. Burdyny, M. G. Kibria, C. P. O’Brien, E. H. Sargent and D. Sinton, Energy Environ. Sci., 2018, 11, 2531–2539 RSC.
  10. J. J. Lv, M. Jouny, W. Luc, W. Zhu, J. J. Zhu and F. Jiao, Adv. Mater., 2018, 1803111, 1–8 Search PubMed.
  11. M. N. Mahmood, D. Masheder and C. J. Harty, J. Appl. Electrochem., 1987, 17, 1223–1227 CrossRef CAS.
  12. D. T. Whipple, E. C. Finke and P. J. A. Kenis, Electrochem. Solid-State Lett., 2010, 13, B109 CrossRef CAS.
  13. K. P. Kuhl, T. Hatsukade, E. R. Cave, D. N. Abram, J. Kibsgaard and T. F. Jaramillo, J. Am. Chem. Soc., 2014, 136, 14107–14113 CrossRef CAS PubMed.
  14. Y. Yan, E. L. Zeitler, J. Gu, Y. Hu and A. B. Bocarsly, J. Am. Chem. Soc., 2013, 135, 14020–14023 CrossRef CAS PubMed.
  15. P. Hirunsit, W. Soodsawang and J. Limtrakul, J. Phys. Chem. C, 2015, 119, 8238–8249 CrossRef CAS.
  16. C.-T. Dinh, T. Burdyny, M. G. Kibria, A. Seifitokaldani, C. M. Gabardo, F. P. García de Arquer, A. Kiani, J. P. Edwards, P. De Luna, O. S. Bushuyev, C. Zou, R. Quintero-Bermudez, Y. Pang, D. Sinton and E. H. Sargent, Science, 2018, 360, 783–787 CrossRef CAS PubMed.
  17. F. S. Roberts, K. P. Kuhl and A. Nilsson, Angew. Chem., Int. Ed., 2015, 54, 5179–5182 CrossRef CAS PubMed.
  18. Z. Q. Liang, T. T. Zhuang, A. Seifitokaldani, J. Li, C. W. Huang, C. S. Tan, Y. Li, P. De Luna, C. T. Dinh, Y. Hu, Q. Xiao, P. L. Hsieh, Y. Wang, F. Li, R. Quintero-Bermudez, Y. Zhou, P. Chen, Y. Pang, S. C. Lo, L. J. Chen, H. Tan, Z. Xu, S. Zhao, D. Sinton and E. H. Sargent, Nat. Commun., 2018, 9, 1–8 CrossRef CAS PubMed.
  19. J. Wu, S. Ma, J. Sun, J. I. Gold, C. Tiwary, B. Kim, L. Zhu, N. Chopra, I. N. Odeh, R. Vajtai, A. Z. Yu, R. Luo, J. Lou, G. Ding, P. J. A. Kenis and P. M. Ajayan, Nat. Commun., 2016, 7, 1–6 Search PubMed.
  20. T.-T. Zhuang, Z.-Q. Liang, A. Seifitokaldani, Y. Li, P. De Luna, T. Burdyny, F. Che, F. Meng, Y. Min, R. Quintero-Bermudez, C. T. Dinh, Y. Pang, M. Zhong, B. Zhang, J. Li, P.-N. Chen, X.-L. Zheng, H. Liang, W.-N. Ge, B.-J. Ye, D. Sinton, S.-H. Yu and E. H. Sargent, Nat. Catal., 2018, 1, 421–428 CrossRef CAS.
  21. Y. Zhou, F. Che, M. Liu, C. Zou, Z. Liang, P. De Luna, H. Yuan, J. Li, Z. Wang, H. Xie, H. Li, P. Chen, E. Bladt, R. Quintero-Bermudez, T. K. Sham, S. Bals, J. Hofkens, D. Sinton, G. Chen and E. H. Sargent, Nat. Chem., 2018, 10, 974–980 CrossRef CAS PubMed.
  22. M. G. Kibria, J. P. Edwards, C. M. Gabardo, C. Dinh, A. Seifitokaldani, D. Sinton and E. H. Sargent, Adv. Mater., 2019, 1807166, 1807166 CrossRef PubMed.
  23. B. Endrődi, G. Bencsik, F. Darvas, R. Jones, K. Rajeshwar and C. Janáky, Prog. Energy Combust. Sci., 2017, 62, 133–154 CrossRef.
  24. D. M. Weekes, D. A. Salvatore, A. Reyes, A. Huang and C. P. Berlinguette, Acc. Chem. Res., 2018, 51, 910–918 CrossRef CAS PubMed.
  25. M. T. Ho, G. W. Allinson and D. E. Wiley, Energy Procedia, 2009, 1, 763–770 CrossRef CAS.
  26. R. F. Service, Science, 2016, 354, 1362–1363 CrossRef CAS PubMed.
  27. B. Kim, S. Ma, H. R. M. Jhong and P. J. A. Kenis, Electrochim. Acta, 2015, 166, 271–276 CrossRef CAS.
  28. S. Verma, S. Lu and P. J. A. Kenis, Nat. Energy, 2019, 4, 466–474 CrossRef CAS.
  29. T. Hatsukade, K. P. Kuhl, E. R. Cave, D. N. Abram and T. F. Jaramillo, Phys. Chem. Chem. Phys., 2014, 16, 13814–13819 RSC.
  30. W. Luc, B. H. Ko, S. Kattel, S. Li, D. Su, J. G. Chen and F. Jiao, J. Am. Chem. Soc., 2019, 141, 9902–9909 CrossRef CAS PubMed.
  31. E. J. Dufek, T. E. Lister and M. E. McIlwain, J. Electrochem. Soc., 2011, 158, B1384 CrossRef CAS.
  32. R. Jones, K. McIntush and C. Wallace, GPA Annu. Conv. Proc., 2010, 1, 322–336 Search PubMed.
  33. A. Bagger, W. Ju, A. S. Varela, P. Strasser and J. Rossmeisl, ChemPhysChem, 2017, 18, 3266–3273 CrossRef CAS PubMed.
  34. X. Lu, Z. Jiang, X. Yuan, Y. Wu, R. Malpass-Evans, Y. Zhong, Y. Liang, N. B. McKeown and H. Wang, Sci. Bull., 2019, 64, 1890–1895 CrossRef.
  35. J. Wang and F. Ciucci, Small, 2017, 13, 1–15 Search PubMed.
  36. Y. J. Zhang, V. Sethuraman, R. Michalsky and A. A. Peterson, ACS Catal., 2014, 4, 3742–3748 CrossRef CAS.
  37. J. Falbe and H. Bahrmann, J. Chem. Educ., 2009, 61, 961 CrossRef.
  38. S. Polani, N. Kanovsky and D. Zitoun, ACS Appl. Nano Mater., 2018, 1, 3075–3079 CrossRef CAS.
  39. A. Qaseem, F. Chen, X. Wu and R. L. Johnston, Catal. Sci. Technol., 2016, 6, 3317–3340 RSC.
  40. S. Sharma and B. G. Pollet, J. Power Sources, 2012, 208, 96–119 CrossRef CAS.
  41. S. Y. Huang, P. Ganesan, S. Park and B. N. Popov, J. Am. Chem. Soc., 2009, 131, 13898–13899 CrossRef CAS PubMed.
  42. C. R. Wilke and P. Chang, AIChE J., 1955, 1, 264–270 CrossRef CAS.
  43. D. Shou, J. Fan, M. Mei and F. Ding, Microfluid. Nanofluid., 2014, 16, 381–389 CrossRef CAS.
  44. Y. Shao, J. Liu, Y. Wang and Y. Lin, J. Mater. Chem., 2009, 19, 46–59 RSC.
  45. H. Lv and S. Mu, Nanoscale, 2014, 6, 5063–5074 RSC.
  46. S. Ma, Y. Lan, G. M. J. Perez, S. Moniri and P. J. A. Kenis, ChemSusChem, 2014, 7, 866–874 CrossRef CAS PubMed.
  47. S. Hossain, J. Saleem, S. Rahman, S. Zaidi, G. McKay, C. Cheng, S. S. Hossain, J. Saleem, S. Rahman, S. M. J. Zaidi, G. McKay and C. K. Cheng, Catalysts, 2019, 9, 298 CrossRef.
  48. J.-X. Lu, J.-J. Zhang, J. Yuan, H. Wang, W.-J. Meng and M.-P. Yang, Catalysts, 2018, 8, 171 CrossRef.
  49. N. Abdullah and S. K. Kamarudin, J. Power Sources, 2015, 278, 109–118 CrossRef CAS.
  50. E. Bakangura, L. Wu, L. Ge, Z. Yang and T. Xu, Prog. Polym. Sci., 2016, 57, 103–152 CrossRef CAS.
  51. G. T. Rochelle, Science, 2009, 325, 1652–1654 CrossRef CAS PubMed.
  52. H. L. Lin, T. L. Yu and F. H. Han, J. Polym. Res., 2006, 13, 379–385 CrossRef CAS.
  53. O. Danyliv, C. Gueneau, C. Iojoiu, L. Cointeaux, A. Thiam, S. Lyonnard and J. Y. Sanchez, Electrochim. Acta, 2016, 214, 182–191 CrossRef CAS.
  54. Z. Liu, H. Yang, R. Kutz and R. I. Masel, J. Electrochem. Soc., 2018, 165, J3371–J3377 CrossRef CAS.
  55. R. L. Weber, Y. Ye, S. M. Banik, Y. A. Elabd, M. A. Hickner and M. K. Mahanthappa, J. Polym. Sci., Part B: Polym. Phys., 2011, 49, 1287–1296 CrossRef CAS.
  56. A. Konovalova, H. Kim, S. Kim, A. Lim, H. S. Park, M. R. Kraglund, D. Aili, J. H. Jang, H. J. Kim and D. Henkensmeier, J. Membr. Sci., 2018, 564, 653–662 CrossRef CAS.
  57. K.-D. Kreuer, Fuel Cells: Selected Entries from the Encyclopedia of Sustainability Science and Technology, Springer Science + Business Media, 2012 Search PubMed.
  58. Y. Sakai, Y. Sadaoka, M. Matsuguchi and H. Sakai, Sens. Actuators, B, 1995, 25, 689–691 CrossRef CAS.
  59. S. Shahgaldi, I. Alaefour and X. Li, Appl. Energy, 2018, 217, 295–302 CrossRef CAS.
  60. K. Talukdar, P. Gazdzicki and K. A. Friedrich, J. Power Sources, 2019, 439, 227078 CrossRef CAS.
  61. C.-T. Dinh, F. P. García de Arquer, D. Sinton and E. H. Sargent, ACS Energy Lett., 2018, 3, 2835–2840 CrossRef CAS.

Footnotes

Electronic supplementary information (ESI) available. See DOI: 10.1039/c9ee03077h
Authors contributed equally.

This journal is © The Royal Society of Chemistry 2020