ELECTROCHEMISTRY

Atomically dispersed Fe3+ sitescatalyze efficient CO2electroreduction to COJun Gu1, Chia-Shuo Hsu2, Lichen Bai1, Hao Ming Chen2*, Xile Hu1*

Currently, the most active electrocatalysts for the conversion of CO2 to CO aregold-based nanomaterials, whereas non–precious metal catalysts have shown low tomodest activity. Here, we report a catalyst of dispersed single-atom iron sites thatproduces CO at an overpotential as low as 80 millivolts. Partial current densityreaches 94 milliamperes per square centimeter at an overpotential of 340 millivolts.Operando x-ray absorption spectroscopy revealed the active sites to be discreteFe3+ ions, coordinated to pyrrolic nitrogen (N) atoms of the N-doped carbon support,that maintain their +3 oxidation state during electrocatalysis, probably throughelectronic coupling to the conductive carbon support. Electrochemical data suggestthat the Fe3+ sites derive their superior activity from faster CO2 adsorption and weakerCO absorption than that of conventional Fe2+ sites.

Electrochemical reduction of carbon dioxide(CO2) is a promising approach to store in-termittent renewable solar and wind en-ergy in carbon-based fuels and chemicals,leading to reduced anthropogenic CO2 emis-

sion (1). To achieve high energy efficiency andscalability, the reaction must occur rapidly andselectively at low overpotentials. Numerous elec-trocatalysts have been developed for CO2 reduc-tion (2), among which gold (Au) and, to a lesserdegree, silver (Ag) are the most efficient at lowoverpotentials. For example, the Faradaic effi-ciency of carbon monoxide (CO) formation canexceed 90% with Au- (3–5) and Ag-based (6, 7)catalysts. On certain Au nanostructures, thepartial current density of CO (denoted as jCO)reached 10 mA cm−2 at overpotentials evenlower than 300 mV (4, 5). Catalysts composedsolely of Earth-abundant elements typicallyhave low selectivity for CO2 reduction (8–12).Recently, many single-atom catalysts (13, 14)have been developed, in which numerous cat-alytic metal sites separated from each otherwere chemically and electronically constrainedon solid supports. These catalysts exhibit prop-erties and activity distinct from both nano-particles and molecular complexes of the samemetal elements. Among them, iron (Fe) (15, 16),cobalt (Co), (17) and nickel (Ni) (18–20) catalystswere reported to exhibit Faradaic efficiencyof CO formation comparable with those of Auand Ag catalysts. However, with these non–precious metal catalysts, much larger over-potentials were required to obtain the same

jCO. Here, we report a catalyst with dispersedsingle-atom Fe sites with ultrahigh activity forCO2 electroreduction to CO.The Fe catalyst (Fe3+–N–C) was prepared

through the pyrolysis of Fe-doped zinc (Zn) 2-methylimidazolate framework (ZIF-8) (21) underN2 at 900°C. The precursor adopts the samecrystal structure as that of undoped ZIF-8 (fig.S2A), with a mole ratio of Fe:Zn of 4:96 (fig.S2E). Fe ions occupy Zn sites and are coordi-nated by four pyrrolic-type nitrogens (N), as re-vealed by the fitting of the Fe K-edge extendedx-ray absorption fine structure (EXAFS) spec-trum (fig. S2, G to I, and table S1). Fe3+–N–C isporous, with a Brunauer-Emmett-Teller surfacearea of 772 m2 g−1 (fig. S3A) and an electrochem-ical (double-layer) surface area of 554m2 g−1 (fig.S3D). The porosity was confirmed by means ofhigh-angle annular dark-field scanning trans-mission electron microscopy (HAADF-STEM)(Fig. 1A). Inductively coupled plasmaoptical emis-sion spectrometry (ICP-OES) analysis showedthe weight fractions of Fe and Zn to be 2.8 and3.4%, respectively, corresponding to anearly equalmole ratio of Fe:Zn. A similar Fe:Zn mole ratiowas found by means of x-ray photoelectron spec-trometry (XPS) (fig. S3E) and energy dispersivex-ray spectroscopy (EDS) (fig. S4F)measurements.The majority of Zn ions in the ZIF-8 precursorwere presumably reduced to Zn particles, whichthen evaporated during pyrolysis. The x-ray dif-fraction (XRD) pattern of Fe3+–N–C (fig. S3G)showed a broad feature at ~25° correspondingto the interlayer distance of the carbon matrix(with a d value of ~0.35 nm). No diffraction peaksof any crystalline species of Fe and Zn were ob-served. Likewise, no nanoparticles were foundin the transmission electron microscopy (TEM)and high-resolution TEM (HRTEM) images;only curved fringes of the layered carbonmatrixwere observed (fig. S4, D and E). The EDS map-pings (Fig. 1, B and C) revealed the homoge-

neous distributions of Fe and N in the carbonmatrix. In the aberration-corrected HAADF-STEM image with atomic resolution (Fig. 1D),the bright spots with size of ~0.2 nm correspondto atomically dispersed Fe and Zn sites. TheFe 2p3/2 XPS spectrum (fig. S3F) and the FeK-edge x-ray absorption near-edge structure(XANES) spectrum (Fig. 1F) showed bindingand edge energies close to those of Fe2O3 andFe3+-tetraphenylporphyrin-Cl (Fe3+TPPCl), in-dicating that the Fe ions in the as-synthesizedFe3+–N–C were in the +3 oxidation state. Thus,the Fe ions were oxidized from +2 to +3 duringthe pyrolysis, which is in agreement with pre-vious reports of pyrolysis of Fe-containing or-ganic precursors (15, 22). The oxidants might beof the same species as protons or residual oxy-gens that oxidized the carbon skeleton of theZIF precursor. Fe K-edge EXAFS (Fig. 1H) sup-ported the atomic dispersion of Fe sites inFe3+–N–C. The fitting of the spectrum (tableS1) indicated that the Fe center adopts a planarFe–X4 (X = N or C) structure. The averagecoordination numbers of Fe–N and Fe–C pathswere 3.4 and 0.5, respectively. No Fe–Fe bondwas detected.As shown by the linear sweep voltammetry

(LSV) curve of Fe3+–N–C in CO2-saturated 0.5 Mpotassium bicarbonate (KHCO3) electrolyte (fig.S5), the onset potential was more positive than–0.20 V versus reversible hydrogen electrode(RHE). Compared with Fe3+–N–C, the currentdensity of the Fe-free control sample, Zn–N–C(prepared by pyrolysis of undoped ZIF-8), wasnegligible, indicating that the electrocatalyticactivity of Fe3+–N–C originates fromFe sites.Wefirst tested the electrocatalytic activity in CO2

reduction using carbon paper electrodes in anH-cell (fig. S6A). CO and H2 were the only gas-phase products, and no solution-phase productwas detected (fig. S6, B to D). CO was detectedafter electrolysis at –0.19 V versus RHE, equiva-lent to an overpotential of 80mV (fig. S6, E andF). The Faradaic efficiency of CO was higherthan 80% between –0.2 and –0.5 V versus RHE(Fig. 2A). The jCO reached 20mA cm−2 at –0.47 Vversus RHE (overpotential of 360 mV) (Fig. 2B).The rate of CO2 reduction might be limited bymass transport in an H-cell (23). Thus, we de-posited Fe3+–N–C on a gas diffusion electrode(GDE) (24). The electrolysis was then conductedin a flow cell with N2-saturated 0.5 M KHCO3 asthe catholyte, andCO2 gaswas fedbehind theGDE(fig. S7A). Ni–Fe layered double hydroxide (LDH)nanosheets (25) were used as the electrocatalystfor the anodic reaction (oxygen evolution). At–0.45 V versus RHE (overpotential of 340 mV),jCO reached 94 mA cm−2 (corresponding to1.75mmolCO hour

−1 cm−2) (Fig. 2B), with Faradaicefficiency of CO on the cathode higher than 90%(Fig. 2A). Good reproducibility was observedin measurements of three independently pre-pared samples, giving a small standard deviation(Fig. 2, A and B).We compare the jCO of Fe

3+–N–C to other state-of-the-art catalysts in Fig. 2C and fig. S8. The jCOof Fe3+–N–Cmeasured in anH-cell between –0.2

RESEARCH

Gu et al., Science 364, 1091–1094 (2019) 14 June 2019 1 of 4

1Laboratory of Inorganic Synthesis and Catalysis, Institute ofChemical Sciences and Engineering, Ecole PolytechniqueFédérale de Lausanne (EPFL), EPFL-ISIC-LSCI, BCH 3305,Lausanne CH 1015, Switzerland. 2Department of Chemistry,National Taiwan University, Taipei 10617, Taiwan.*Corresponding author. Email: haomingchen@ntu.edu.tw(H.M.C.); xile.hu@epfl.ch (X.H.)

on February 24, 2021

http://science.sciencem

ag.org/D

ownloaded from

and –0.5 V versus RHE is considerably higher thanthat attainedby other non–preciousmetal catalystsand even Ag catalysts (7), reaching comparablelevels with those of oxide-derived Au catalysts (3).The mole-normalized current of Fe3+–N–C is sig-nificantly higher than that of Au catalysts in thispotential range (fig. S8B). Assuming all Fe atomsto be catalytically active, the apparent turnoverfrequencies (TOFs) of Fe3+–N–C (Fig. 2D) arecomparable with those of Au catalysts (3, 4, 26)and greatly exceed those of other non–preciousmetal catalysts (15, 16, 19). Unlike for catalystsbased on copper (Cu), Ag, and Au (27), ultrapureelectrolyte solutions were not necessary forFe3+–N–C. When KHCO3 with a purity of 99.5%was used to prepare electrolyte, or even tapwater was used in place of deionized water(18.2 megohm cm), the Faradaic efficiency andjCO show no obvious change, and the perform-ance was stable for at least 12 hours (Fig. 2E).After 12 hours of electrolysis, the weight fractionof Fe in Fe3+–N–C (measured with ICP-OES) was2.6%, indicating no appreciable leaching of Feions from the catalyst. No aggregation of Fe or Znspecies was detected in TEM images, and a highdensity of discrete Fe and Zn atoms was stillobserved (fig. S9).The performance of Fe3+–N–C was stable be-

tween –0.2 and –0.5 V versus RHE, although at

potentials more negative than –0.5 V versus RHE,the activity became unstable (fig. S7B). As shownin fig. S7C, the current density at –0.41 V wasstable during a 28-hour chronoamperometrytest. Whereas at –0.51 V versus RHE, the initialjCO was much higher, it decreased rapidly to avalue similar to that obtained at –0.41 V versusRHE. This result indicates some changes ofFe3+–N–C around –0.5 V versus RHE. To ex-plore the nature of this change, we conductedoperandoXASmeasurements in theCO2-saturated0.5 M KHCO3 catholyte. Fe K-edge spectra wereobtained on dry samples and on samples thatwere loaded on glassy carbon electrodes and im-mersed in the electrolyte at open circuit potential(OCP) as well as at –0.1 to –0.6 V versus RHE(Fig. 3A). For Fe3+–N–C, the Fe K-edge showedno obvious shift between the dry powder andthe in situ sample at –0.4 V versus RHE. The edgeenergy was close to that of Fe3+TPPCl, indicatingthat the Fe ions in Fe3+–N–C remained in the +3oxidation state during CO2 electroreduction atpotentials as negative as –0.4 V versus RHE.Whenthe applied potential was shifted further nega-tive, to –0.5 V versus RHE and beyond, the FeK-edge shifted to lower energies, which werecomparable with that of FeO, suggesting thereduction of Fe3+ to Fe2+. This reduction processoccurred at the same potential as the above-

mentioned deactivation of Fe3+–N–C, implyingthat Fe3+ sites aremore active for generating CO.Moreover, the fitting of EXAFS spectra (fig. S10C)indicates that the reduction of Fe3+ sites is ac-companied by a change of local structure aroundthe Fe ions. Before the reduction of Fe3+ sites, thefirst shell coordination number of Fe (Fe–N andFe–C)was about 4, whereas as the Fe3+ sites werereduced to Fe2+ sites, the first shell coordinationnumber of Fe decreased to about 3.To investigate the origins of the improved ac-

tivity of Fe3+–N–C as compared with previouslyreported single-atom Fe catalysts, we measuredin situ Fe K-edge XANES spectra of Fe0.5d (fig.S11, E and F) (15). Under potentials between –0.2and –0.5 V versus RHE, the energy of the Fe K-edge was close to that of FeO, indicating a +2rather than +3 oxidation state for the Fe sitesduring CO2 reduction. This difference in oxida-tion state might be due to different ligand envi-ronments, particularly with regard to the N atoms.For Fe0.5d (15) and Fe–N–C (16), Fe ions coor-dinated with four pyridinic N were proposed asthe active sites. For Fe3+–N–C, XANES and XPSspectra of N indicate that the Fe ions were co-ordinated to pyrrolic N. In the N K-edge XANESspectrum (fig. S12A), p* and s* features of Fe3+–N–C were similar to those of a metal-porphyrinderivative (28). In the N 1s XPS spectrum (fig.

Gu et al., Science 364, 1091–1094 (2019) 14 June 2019 2 of 4

Fig. 1. Characterizationsof Fe3+–N–C. (A) HAADF-STEM image and thecorresponding EDSmappings of (B) Fe and(C) N of the regionenclosed by the red square.(D) Aberration-correctedHAADF-STEM imageand (E) EDS spectrumof the red square region.(F) Fe K-edge XANESspectra of Fe3+–N–C(black), Fe2O3 (blue dashed),Fe3+TPPCl (green dashed),FeO (pink dashed), andFe foil (orange dashed).(Inset) The enlargement ofthe main edges. (G) k-spaceand (H) R-space Fe K-edgeEXAFS spectra. Shownare data (black) and fittingcurves (red).

E F

HG

BA C D

RESEARCH | REPORTon F

ebruary 24, 2021

http://science.sciencemag.org/

Dow

nloaded from

S12B), the major peak at 398.6 eV was attributedto pyrrolic N coordinated to Fe, which is in ag-reement with the spectrum of Fe3+TPPCl. Thisassignment is consistent with the atomic frac-tions of N andmetals and the coordination num-ber of metal-N (table S4).To further test the above hypothesis, we di-

rectly compared the Fe3+–N–C catalyst with ananalogous Fe–N–C catalyst in which the Fe ionswere coordinated by pyridinic N atoms. Consid-ering that pyrolysis of Fe precursors containingeither pyridinic ligands or pyrrolic ligands seemedto conserve the pyridinic or pyrrolic nature ofthe N atoms, we prepared the reference sample(Fe2+–N–C) by pyrolysis of a composite contain-ing a Fe-phenanthroline complex at 700°C (22).Aberration-corrected HAADF-STEM (fig. S13D)and Fe K-edge EXAFS (fig. S14E and table S1)confirmed the single-atom nature of the Fe sitesin Fe2+–N–C. In theN 1s XPS spectrum (fig. S12C),the major feature at 399.7 eV was assigned topyridinic N coordinated to Fe, which is in agree-ment with the assignments of the spectra ofFe-phenanthroline complexes and previously re-ported metal-N-C catalysts with pyridinic N lig-ands (16). This assignment is also consistent withthe percentage of coordinated N measured withother methods (table S4). Thus, Fe2+–N–C con-

tained Fe ions coordinated by pyridinic N atoms.A XANES spectrum (fig. S14D) showed that ini-tially, the energy of Fe K-edge of Fe2+–N–C wasconsiderably higher than that of FeO. The Fe 2pXPS (fig. S14C) spectrum showed that the bindingenergy of Fe ion was similar to that of Fe2O3.These data suggested an important number ofFe3+ sites in the as-prepared sample of Fe2+–N–C.The in situ XANES (Fig. 3B) showed that Fe3+ inthe as-prepared Fe2+–N–C started to be reducedto Fe2+ at –0.1 to –0.2 V versus RHE. During CO2

electroreduction (at potentials more negativethan –0.2 V versus RHE), the energy of the FeK-edgewas slightly lower than that of FeO. Thus,the Fe sites in Fe2+–N–C under reaction condi-tions had an oxidation state of +2 or lower. TheTOF of CO production of Fe2+–N–C ismore thanan order of magnitude lower than that of Fe3+–N–C under the same potential (Fig. 2D). The cur-rent density of Fe2+–N–C decreased markedlyduring 2-hour choronoamperometry tests (fig.S6G), indicating its lower stability as comparedwith the Fe3+–N–C catalyst. These data suggestthat pyrrolic type ligands are important to keepFe sites in the +3 oxidation state during CO2

electroreduction and consequentlymaintain thehigh activity and stability of Fe3+ sites. This hy-pothesis is further supported by the different re-

activity of Fe3+–N–C and Fe2+–N–C towardNaBH4.The Fe3+ ions coordinated by pyridinic N ligandsin Fe2+–N–C could be reduced byNaBH4,whereasthose coordinated by pyrrolic N ligands in Fe3+–N–C could not (fig. S15).For Fe2+–N–C, the jCO at a fixed potential ver-

sus the standard hydrogen electrode (SHE) islargely independent of the concentration of theproton donor (fig. S16, A and B), indicating thatthe 1-electron reduction (adsorption) of CO2 isdecoupled from a proton transfer (29, 30). Atmodest overpotentials, the jCO of Fe2+–N–C hasa Tafel slope of 117 mV/decade (fig. S16E), sug-gesting that CO2 adsorption is slow and rate-limiting (supplementary materials, kinetic andmechanistic analysis). On the other hand, thejCO of Fe3+–N–C is approximately first-order inthe concentration of HCO3

– (fig. S16D) and hasa Tafel slope of 64-71 mV/decade at low over-potentials (fig. S16E). These kinetic data sug-gest that for Fe3+–N–C, the 1 electron reductionof CO2 is also decoupled from a proton transfer.Moreover, CO2 adsorption is fast, and the rate-limiting step is the protonation of the adsorbedCO2

– to form an adsorbed COOH intermediate(supplementary materials, kinetic and mecha-nistic analysis). These results indicate CO2 ad-sorption as a descriptor of catalytic activity at

Gu et al., Science 364, 1091–1094 (2019) 14 June 2019 3 of 4

Fig. 2. CO2 electroreduction performance.(A) Faradaic efficiency of CO (solid lines)and H2 (dashed lines) production and(B) jCO of Fe3+–N–C in an H-cell (red) andon a GDE (blue), and of Fe2+–N–C inan H-cell (black). Data from the H-cell wereobtained by means of chronoamperometry,whereas data from the GDE were obtained bymeans of chronopotentiometry. Each errorbar was the standard deviation determinedbased on tests of three individual electrodes.Loading was 0.6 mg cm−2 for Fe3+–N–Cand Fe2+–N–C; 2.5 mg cm−2 for Fe3+–N–C/DGE.(C and D) Comparison of (C) jCO and(D) apparent TOFs of CO production ofFe3+–N–C in an H-cell (red circles) and ona GDE (red stars) and of Fe2+–N–C inan H-cell (red squares), to that of otherreported catalysts: other Fe-N-C catalysts[Fe-0.5d (15) and Fe–N–C (16)], a Co–N–Ccatalyst with two coordinating nitrogenatoms (Co–N2) (17), atomically dispersedNi on nitrogen-sulfur codoped graphene(A–Ni–NSG) (19), oxide-derived Au electrode(OD-Au) (3), carbon black supportedAu nanowires with a length of 500 nm(C–Au-500) (4), needle-shape Au nanostruc-tures (Au needles) (5), nanoporousAg electrode (np-Ag) (7), and Au-polymer-multiwall carbon nanotubes compositeloaded on GDE (Au/GDE) (34) in bicarbonateelectrolytes. (E) Chronoamperometry curveand Faradaic efficiency of CO production(dots) by Fe3+–N–C in H-cell at –0.37 V versusRHE. The electrolytes were prepared from potassium carbonate (K2CO3) (99.999%) and deionized water (18.2 megohms cm) (black),KHCO3 (99.5%) and deionized water (red), and KHCO3 (99.5%) and tap water (blue), respectively.

A B

DC

E

RESEARCH | REPORTon F

ebruary 24, 2021

http://science.sciencemag.org/

Dow

nloaded from

low to modest overpotentials. They also reveala faster CO2 adsorption in Fe3+–N–C than inFe2+–N–C, which explains why Fe3+–N–C hasa lower onset overpotential. The CO2 electro-reduction was conducted in the presence ofCO (0.2 atm) (fig. S17). External CO did notinfluence the activity of Fe3+–N–C but largelydecreased the activity of Fe2+–N–C. This resultsuggests that at high overpotentials, CO de-sorption becomes rate limiting for Fe2+–N–C.Because CO desorption is a non-Faradaic step,once it becomes rate limiting, the rate will hardlyincrease with increasing overpotentials. At higheroverpotentials, the Tafel slope of Fe2+–N–C be-comes enormous (546 mV/decade) (fig. S16E),and jCO cannot exceed 2 mA cm−2. On the otherhand, the reaction at Fe3+ sites was not limited byCO desorption and could reach a very high currentdensity. Thus, the higher activity of Fe3+–N–Ccompared with Fe2+–N–C at high overpotentialscan be rationalized by a weaker CO bindingat an Fe3+ center than at an Fe2+ center.The spectroscopic data indicate that Fe3+–

N–C comprises pyrrolic N ligands,whereas Fe2+–N–C comprises pyridinicN ligands. The pyrrolicNligandsmay stabilize Fe3+ relative to Fe2+, whereasthe pyridinic N ligands have the opposite effect.Thermodynamically, the respective reductionpotentials support this hypothesis: The stan-dard reduction potential of [Fe(phen)3]

3+/2+ is1.06 V versus SHE, whereas that of Fe3+/2+ couplein Fe-porphyrin complexes can reach as low as–0.4 V versus SHE (31). Preservation of the +3oxidation state during CO2 electroreduction iscounterintuitive because the formation of highly

reduced, low-valent Fe species is necessary formolecular Fe catalysts (32). Once conjugated to aconductive carbonmatrix, however, the Fe3+ siteis electronically coupled to the conductive sup-port so that the Fe3+/2+ reduction potential movesto the same degree as the Fermi level of the car-bon support when applying an external bias (fig.S18, A and B). The Fe3+/2+ reduction potentialremains more negative than the Fermi level ofthe carbon support, leading to the stabilization ofthe Fe3+ ions. An analogous “strong-coupling”effect was formulated to explain the lack of re-dox chemistry on conjugated molecular sitesduring potential cycling (33). The reduction ofFe3+ sites in Fe3+–N–C at –0.5 V versus RHE isprobably enabled by a change of their coordi-nation environment. In situ EXAFS (fig. S10C)revealed that the Fe ion lost one pyrrolic N lig-and at this potential, possibly because of pro-tonation or hydrogenation of the ligand drivenby the electric field. The new coordination en-vironment increases the Fe3+/2+ reduction po-tential to be more positive than the Fermi levelof the carbon support, resulting in conjugatedFe2+ ions (fig. S18C).

REFERENCES AND NOTES

1. S. Chu, Y. Cui, N. Liu, Nat. Mater. 16, 16–22 (2016).2. D. D. Zhu, J. L. Liu, S. Z. Qiao, Adv. Mater. 28, 3423–3452

(2016).3. Y. Chen, C. W. Li, M. W. Kanan, J. Am. Chem. Soc. 134,

19969–19972 (2012).4. W. Zhu et al., J. Am. Chem. Soc. 136, 16132–16135 (2014).5. M. Liu et al., Nature 537, 382–386 (2016).6. B. A. Rosen et al., Science 334, 643–644 (2011).7. Q. Lu et al., Nat. Commun. 5, 3242 (2014).

8. T. Zheng, K. Jiang, H. Wang, Adv. Mater. 30, e1802066(2018).

9. P. P. Sharma et al., Angew. Chem. Int. Ed. 54, 13701–13705(2015).

10. H. Won et al., Angew. Chem. Int. Ed. 55, 9297–9300(2016).

11. M. Asadi et al., Nat. Commun. 5, 4470 (2014).

12. J. Gu, F. Héroguel, J. Luterbacher, X. Hu, Angew. Chem. Int. Ed.57, 2943–2947 (2018).

13. A. Wang, J. Li, T. Zhang, Nat. Rev. Chem. 2, 65–81 (2018).

14. Y. Chen et al., Joule 2, 1242–1264 (2018).

15. T. N. Huan et al., ACS Catal. 7, 1520–1525 (2017).

16. W. Ju et al., Nat. Commun. 8, 944 (2017).

17. X. Wang et al., Angew. Chem. Int. Ed. 57, 1944–1948(2018).

18. C. Zhao et al., J. Am. Chem. Soc. 139, 8078–8081 (2017).

19. H. B. Yang et al., Nat. Energy 3, 140–147 (2018).

20. T. Zheng et al., Joule 3, 265–278 (2019).

21. K. S. Park et al., Proc. Natl. Acad. Sci. U.S.A. 103, 10186–10191(2006).

22. W. Liu et al., J. Am. Chem. Soc. 139, 10790–10798 (2017).

23. M. L. Perry, J. Newman, E. J. Cairns, J. Electrochem. Soc. 145,5–15 (1998).

24. H.-R. M. Jhong, F. R. Brushett, P. J. A. Kenis, Adv. EnergyMater. 3, 589–599 (2013).

25. F. Song, X. Hu, Nat. Commun. 5, 4477 (2014).

26. Z. P. Jovanov et al., J. Catal. 343, 215–231 (2016).

27. Y. Hori et al., Electrochim. Acta 50, 5354–5369 (2005).

28. G. Polzonetti et al., Chem. Phys. 296, 87–100 (2004).

29. A. S. Varela et al., ACS Energy Lett. 3, 812–817 (2018).

30. A. Wuttig, Y. Yoon, J. Ryu, Y. Surendranath, J. Am. Chem. Soc.139, 17109–17113 (2017).

31. N. Silva et al., J. Electroanal. Chem. 819, 502–510 (2018).

32. C. Costentin, S. Drouet, M. Robert, J.-M. Savéant, Science 338,90–94 (2012).

33. M. N. Jackson et al., J. Am. Chem. Soc. 140, 1004–1010(2018).

34. S. Verma et al., ACS Energy Lett. 3, 193–198 (2018).

35. J.Gu, C.-S. Hsu, L. Bai, H. M. Chen, X. Hu, X-ray absorptiondata and microscopic images of “Atomically dispersed iron(3+)sites catalyze efficient CO2 electroreduction to CO”. Zenodo(2019); doi:10.5281/zenodo.2671703

ACKNOWLEDGMENTS

We thank C. Corminboeuf and K.-h. Lin (EPFL) for discussionof theoretical analysis. We thank J. Luterbacher and F. Héroguel(EPFL) for their help in physical adsorption experiments.Funding: This work was supported by the GAZNAT SA andthe European Research Council (grant 681292). We alsoacknowledge support from the Ministry of Science andTechnology, Taiwan (contract MOST 107-2628-M-002-015-RSP).Author contributions: J.G. performed the majority of thesynthesis, characterization, and electrochemical tests. L.B.contributed to the initial synthesis of catalysts and TEMmeasurement. C.-S.H. performed the in situ x-ray absorptionexperiments. J.G., C.-S.H., H.M.C., and X.H. analyzed the data.J.G. and X.H. wrote the paper, with input from all otherco-authors. H.M.C. and X.H. directed the research.Competing interests: European priority patent applications(nos. 18156529.2 and 18193304.5) titled “Fe-N-C Catalyst,method of preparation and uses thereof” were filed by GAZNATSA with J.G. and X.H. as inventors. Data and materialsavailability: All results are reported in the main text andsupplementary materials. EXAFS, XANES, and microscopydata files are deposited in Zenodo (35). Unless restricted bypatent, the materials are available for the purpose ofreproducing or extending the analysis.

SUPPLEMENTARY MATERIALS

science.sciencemag.org/content/364/6445/1091/suppl/DC1Materials and MethodsSupplementary TextFigs. S1 to S18Tables S1 to S4References (36–43)

22 January 2019; accepted 20 May 201910.1126/science.aaw7515

Gu et al., Science 364, 1091–1094 (2019) 14 June 2019 4 of 4

Fig. 3. Operando XAS characterization. (A and B) Fe K-edge XANES spectra (left) and the firstderivative of the spectra (right) of (A) Fe3+–N–C and (B) Fe2+–N–C as dry powder (black) andloaded on glassy carbon electrodes at open circuit potential (OCP) (blue), –0.1 V (light blue),–0.2 V (green), –0.3 V (dark green), –0.4 V (dark blue), –0.5 V (red), and –0.6 V (pink) versusRHE, with the spectra of Fe2O3 (blue dashed), Fe3+TPPCl (green dashed), FeO (pink dashed),and Fe foil (orange dashed) as references.

RESEARCH | REPORTon F

ebruary 24, 2021

http://science.sciencemag.org/

Dow

nloaded from

electroreduction to CO2 sites catalyze efficient CO3+Atomically dispersed FeJun Gu, Chia-Shuo Hsu, Lichen Bai, Hao Ming Chen and Xile Hu

DOI: 10.1126/science.aaw7515 (6445), 1091-1094.364Science 

, this issue p. 1091Sciencedispersed single iron ions in the +3 oxidation state.iron catalyst with activity equaling or exceeding that of the precious metals. The key proved to be stabilization of the

now report anet al.whereas more abundant, less expensive metals tend to require impractically high potentials. Jun Gu the greenhouse gas to commodity chemicals. Currently, gold and silver are the most active catalysts for this process,

to CO could be a promising first step in sustainable conversion of2Large-scale electrochemical reduction of CO2Three's a charm for iron and CO

ARTICLE TOOLS http://science.sciencemag.org/content/364/6445/1091

MATERIALSSUPPLEMENTARY http://science.sciencemag.org/content/suppl/2019/06/12/364.6445.1091.DC1

REFERENCES

http://science.sciencemag.org/content/364/6445/1091#BIBLThis article cites 41 articles, 4 of which you can access for free

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAAS.ScienceScience, 1200 New York Avenue NW, Washington, DC 20005. The title (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement ofScience

Science. No claim to original U.S. Government WorksCopyright © 2019 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of

on February 24, 2021

http://science.sciencem

ag.org/D

ownloaded from