Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Accelerated discovery of CO2 electrocatalysts using active machine learning


The rapid increase in global energy demand and the need to replace carbon dioxide (CO2)-emitting fossil fuels with renewable sources have driven interest in chemical storage of intermittent solar and wind energy1,2. Particularly attractive is the electrochemical reduction of CO2 to chemical feedstocks, which uses both CO2 and renewable energy3,4,5,6,7,8. Copper has been the predominant electrocatalyst for this reaction when aiming for more valuable multi-carbon products9,10,11,12,13,14,15,16, and process improvements have been particularly notable when targeting ethylene. However, the energy efficiency and productivity (current density) achieved so far still fall below the values required to produce ethylene at cost-competitive prices. Here we describe Cu-Al electrocatalysts, identified using density functional theory calculations in combination with active machine learning, that efficiently reduce CO2 to ethylene with the highest Faradaic efficiency reported so far. This Faradaic efficiency of over 80 per cent (compared to about 66 per cent for pure Cu) is achieved at a current density of 400 milliamperes per square centimetre (at 1.5 volts versus a reversible hydrogen electrode) and a cathodic-side (half-cell) ethylene power conversion efficiency of 55 ± 2 per cent at 150 milliamperes per square centimetre. We perform computational studies that suggest that the Cu-Al alloys provide multiple sites and surface orientations with near-optimal CO binding for both efficient and selective CO2 reduction17. Furthermore, in situ X-ray absorption measurements reveal that Cu and Al enable a favourable Cu coordination environment that enhances C–C dimerization. These findings illustrate the value of computation and machine learning in guiding the experimental exploration of multi-metallic systems that go beyond the limitations of conventional single-metal electrocatalysts.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Fig. 1: Screening of Cu and Cu based compounds using computational methods.
Fig. 2: Schematic and characterization of de-alloyed Cu-Al catalyst.
Fig. 3: CO2 electroreduction performance on de-alloyed Cu-Al, porous Cu and deposited Cu catalysts on C-GDL substrates in 1 M KOH electrolytes.
Fig. 4: CO2 electroreduction performance on de-alloyed Cu-Al catalysts on PTFE substrates in alkaline electrolytes at different pH values.

Data availability

Source data to generate figures and tables are available from the corresponding authors on reasonable request.

Code availability

Code to generate figures and tables is available from the corresponding authors on reasonable request.


  1. 1.

    Lin, S. et al. Covalent organic frameworks comprising cobalt porphyrins for catalytic CO2 reduction in water. Science 349, 1208–1213 (2015).

    CAS  Article  ADS  Google Scholar 

  2. 2.

    Schreier, M. et al. Solar conversion of CO2 to CO using Earth-abundant electrocatalysts prepared by atomic layer modification of CuO. Nat. Energy 2, 17087 (2017).

    CAS  Article  ADS  Google Scholar 

  3. 3.

    Dinh, C. et al. Sustained high-selectivity CO2 electroreduction to ethylene via hydroxide-mediated catalysis at an abrupt reaction interface. Science 360, 783–787 (2018).

    CAS  Article  Google Scholar 

  4. 4.

    Li, C. et al. Electroreduction of carbon monoxide to liquid fuel on oxide-derived nanocrystalline copper. Nature 508, 504–507 (2014).

    CAS  Article  ADS  Google Scholar 

  5. 5.

    Lu, Q. et al. A selective and efficient electrocatalyst for carbon dioxide reduction. Nat. Commun. 5, 3242 (2014).

    Article  ADS  Google Scholar 

  6. 6.

    Gao, S. et al. Partially oxidized atomic cobalt layers for carbon dioxide electroreduction to liquid fuel. Nature 529, 68–71 (2016).

    CAS  Article  ADS  Google Scholar 

  7. 7.

    Li, Y. C. et al. Electrolysis of CO2 to syngas in bipolar membrane-based electrochemical cells. ACS Energy Lett. 1, 1149–1153 (2016).

    CAS  Article  ADS  Google Scholar 

  8. 8.

    Jeanty, P. et al. Upscaling and continuous operation of electrochemical CO2 to CO conversion in aqueous solutions on silver gas diffusion electrodes. J. CO 2 Util. 24, 454–462 (2018).

    CAS  Article  Google Scholar 

  9. 9.

    Hori, Y. et al. Selective formation of C2 compounds from electrochemical reduction of CO2 at a series of copper single crystal electrodes. J. Phys. Chem. B 106, 15–17 (2002).

    CAS  Article  Google Scholar 

  10. 10.

    Yano, H. et al. Selective electrochemical reduction of CO2 to ethylene at a three-phase interface on copper(I) halide-confined Cu-mesh electrodes in acidic solutions of potassium halides. J. Electroanal. Chem. 565, 287–293 (2004).

    CAS  Article  Google Scholar 

  11. 11.

    Peterson, A. A. et al. How copper catalyzes the electroreduction of carbon dioxide into hydrocarbon fuels. Energy Environ. Sci. 3, 1311–1315 (2010).

    CAS  Article  Google Scholar 

  12. 12.

    Kortlever, R. et al. Catalysts and reaction pathways for the electrochemical reduction of carbon dioxide. J. Phys. Chem. Lett. 6, 4073–4082 (2015).

    CAS  Article  Google Scholar 

  13. 13.

    Mistry, H. et al. Highly selective plasma-activated copper catalysts for carbon dioxide reduction to ethylene. Nat. Commun. 7, 12123 (2016).

    Article  ADS  Google Scholar 

  14. 14.

    Li, Y. et al. Structure-sensitive CO2 electroreduction to hydrocarbons on ultrathin 5-fold twinned copper nanowires. Nano Lett. 17, 1312–1317 (2017).

    CAS  Article  ADS  Google Scholar 

  15. 15.

    Lum, Y. et al. Optimizing C–C coupling on oxide-derived copper catalysts for electrochemical CO2 reduction. J. Phys. Chem. C 121, 14191–14203 (2017).

    CAS  Article  Google Scholar 

  16. 16.

    De Luna, P. et al. Catalyst electro-redeposition controls morphology and oxidation state for selective carbon dioxide reduction. Nat. Catal. 1, 103–110 (2018).

    Article  Google Scholar 

  17. 17.

    Liu, X. et al. Understanding trends in electrochemical carbon dioxide reduction rates. Nat. Commun. 8, 15438 (2017).

    CAS  Article  ADS  Google Scholar 

  18. 18.

    Tran, K. et al. Active learning across intermetallics to guide discovery of electrocatalysts for CO2 reduction and H2 evolution. Nat. Catal. 1, 696–703 (2018).

    CAS  Article  Google Scholar 

  19. 19.

    van der Maaten, L. Accelerating t-SNE using tree-based algorithms. J. Mach. Learn. Res. 15, 3221–3245 (2014).

    MathSciNet  MATH  Google Scholar 

  20. 20.

    Davis, L. E. et al. Handbook of Auger Electron Spectroscopy 2nd edn (Physical Electronics Industries, 1976).

  21. 21.

    Persson, K. A. et al. Prediction of solid-aqueous equilibria: scheme to combine first-principles calculations of solids with experimental aqueous states. Phys. Rev. B 85, 235438 (2012).

    Article  ADS  Google Scholar 

  22. 22.

    Montoya, J. H. et al. A high-throughput framework for determining adsorption energies on solid surfaces. npj Comput. Mater. 3, 14 (2017).

    Article  ADS  Google Scholar 

  23. 23.

    Xiao, H. et al. Atomistic mechanisms underlying selectivities in C1 and C2 products from electrochemical reduction of CO on Cu (111). J. Am. Chem. Soc. 139, 130–136 (2017).

    CAS  Article  Google Scholar 

  24. 24.

    Xiao, H. et al. Cu metal embedded in oxidized matrix catalyst to promote CO2 activation and CO dimerization for electrochemical reduction of CO2. Proc. Natl Acad. Sci. USA 114, 6685–6688 (2017).

    CAS  Article  ADS  Google Scholar 

  25. 25.

    Jain, A. et al. The Materials Project: a materials genome approach to accelerated materials innovation. APL Mater. 1, 011002 (2013).

    Article  ADS  Google Scholar 

  26. 26.

    Nørskov, J. K. et al. Trends in the exchange current for hydrogen evolution. J. Electrochem. Soc. 152, J23–J26 (2005).

    Article  Google Scholar 

  27. 27.

    Wang, L. et al. Electrochemical carbon monoixde reduction on polycrystalline copper: effects of potential, pressure, and pH on selectivity toward multicarbon and oxygenated products. ACS Catal. 8, 7445–7454 (2018).

    CAS  Article  Google Scholar 

  28. 28.

    Liu, M. et al. Enhanced electrocatalytic CO2 reduction via field-induced reagent concentration. Nature 537, 382–386 (2016).

    CAS  Article  ADS  Google Scholar 

  29. 29.

    Zeng, Z. et al. Stabilization of ultrathin (hydroxy)oxide films on transition metal substrates for electrochemical energy conversion. Nat. Energy 2, 17070 (2017).

    CAS  Article  ADS  Google Scholar 

  30. 30.

    She, Z. W. et al. Combining theory and experiment in electrocatalysis: Insights into materials design. Science 355, 146 (2017).

    Google Scholar 

  31. 31.

    Larrazábal, G. O. et al. Building blocks for high performance in electrocatalytic CO2 reduction: materials, optimization strategies, and device engineering. J. Phys. Chem. Lett. 8, 3933–3944 (2017).

    Article  Google Scholar 

  32. 32.

    Whipple, D. T. et al. Prospects of CO2 utilization via direct heterogeneous electrochemical reduction. J. Phys. Chem. Lett. 1, 3451–3458 (2010).

    CAS  Article  Google Scholar 

  33. 33.

    De Luna, P. et al. What would it take for renewably powered electrosynthesis to displace petrochemical processes? Science 364, eaav3506 (2019).

    Article  ADS  Google Scholar 

Download references


This work was supported financially by the Ontario Research Fund Research-Excellence Program, the Natural Sciences and Engineering Research Council (NSERC) of Canada, the Canadian Institute for Advanced Research (CIFAR) Bio-Inspired Solar Energy programme, the University of Toronto Connaught programme, and TOTAL American Services. M.Z. thanks the National Natural Science Foundation of China (grant number 91963121), and S. Tolbert from the University of California, Los Angeles for discussions of de-alloying. S.S. thanks the National Key Research and Development Program of China (grant number 2016YFB0700205) and the National Natural Science Foundation of China (grant number U1632273). We thank R. Morris and D. Sinton from the University of Toronto for discussions. We thank C. McCallum, R. Wolowiec, D. Kopilovic, S. Boccia, A. Ip, M. Liu, Y. Pang, M. Askerka, A. Seifitokaldani, T. T. Zhuang and Z. Liang from the University of Toronto, Canada and C.-W. Huang, L.-J. Chen from National Tsing Hua University, Taiwan, for their help during the course of study. We thank the beamline scientists from the Source optimisée de lumière d’énergie intermédiaire du LURE (SOLEIL) Synchrotron in France for performing in situ X-ray absorption analyses. This research used resources of the National Energy Research Scientific Computing Center, a Department of Energy (DOE) Office of Science User Facility supported by the Office of Science of the US Department of Energy under contract number DE-AC02-05CH11231. Computations were performed on the Southern Ontario Smart Computing Innovation Platform (SOSCIP) Consortium’s Blue Gene/Q computing platform. SOSCIP is funded by the Federal Economic Development Agency of Southern Ontario, the Province of Ontario, IBM Canada Ltd, Ontario Centres of Excellence, MITACS and 15 Ontario academic member institutions.

Author information




E.H.S. supervised the project. M.Z. and E.H.S. conceived the idea. M.Z. and C.W. designed and carried out the experiments. K.T., Z.Y. and Z.U. performed the machine learning studies. K.T., Z.Y., Z.U., Y.M., Z.W. O.V., P.D.L., M.A., M.Z. and E.H.S. discussed the machine learning results. Y.M., Z.W. O.V., P.D.L., M.A., A.S., F.C., K.T., Z.Y. and Z.U. carried out the DFT simulations. P.B. carried out the Auger electron spectroscopy analyses. S.S. and P.D.L. performed X-ray absorption spectroscopy measurements. C.-S.T. and S.-C.L. carried out the TEM analyses. C.-T.D., A.S.R., C.-S.T., M.A., M.L., A.S., Y.P. and A.I. contributed to the discussion of the results. All authors discussed the results and assisted during manuscript preparation.

Corresponding authors

Correspondence to Zachary Ulissi or Edward H. Sargent.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Nature thanks Hailiang Wang and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

This file contains Supplementary Methods, which includes Supplementary Figures 1-68 and Supplementary Tables 1-11.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Zhong, M., Tran, K., Min, Y. et al. Accelerated discovery of CO2 electrocatalysts using active machine learning. Nature 581, 178–183 (2020).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing