Skip to main content

Thank you for visiting nature.com. 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.

Anti-dissolution Pt single site with Pt(OH)(O3)/Co(P) coordination for efficient alkaline water splitting electrolyzer

An Author Correction to this article was published on 25 July 2022

This article has been updated

Abstract

As the most well-known electrocatalyst for cathodic hydrogen evolution in water splitting electrolyzers, platinum is unfortunately inefficient for anodic oxygen evolution due to its over-binding with oxygen species and excessive dissolution in oxidative environment. Herein we show that single Pt atoms dispersed in cobalt hydrogen phosphate with an unique Pt(OH)(O3)/Co(P) coordination can achieve remarkable catalytic activity and stability for oxygen evolution. The catalyst yields a high turnover frequency (35.1 ± 5.2 s−1) and mass activity (69.5 ± 10.3 A mg−1) at an overpotential of 300 mV and excellent stability. Mechanistic studies elucidate that the superior catalytic performance of isolated Pt atoms herein stems from optimal binding energies of oxygen intermediate and also their strong electronic coupling with neighboring Co atoms that suppresses the formation of soluble Ptx>4 species. Alkaline water electrolyzers assembled with an ultralow Pt loading realizes an industrial-level current density of 1 A cm−2 at 1.8 volts with a high durability.

Introduction

Development of efficient and durable electrocatalysts for both cathodic hydrogen evolution reaction (HER) and anodic oxygen evolution reaction (OER) is essential to fulfill the promises of water splitting electrolyzers1,2. More challenging between the two, OER suffers from particularly sluggish kinetics, thus demanding high-performance catalysts to lower the reaction overpotential for practical applications3,4. Over the past decades, noble metals, such as Ir5,6,7 and Ru8,9 have been extensively explored as potential OER catalysts, but their turnover frequency (TOF) and mass activity (MA) values remain less than satisfactory owing to their insufficient atom-utilization efficiencies, as exemplified in IrO2 (0.01 s−1)10, IrO2/GCN (0.17 s−1)11, Cr0.6Ru0.4O2 (0.229 A mgRu−1)12 and Ru1-Pt3Cu (0.779 A mgRu−1)13. Moreover, although significant progress has been made in rotating disk electrodes, much less is known of an integral device configuration14.

Platinum (Pt) is the best known elemental HER catalyst at the cathode of water splitting electrolyzers thanks to its optimal hydrogen absorption energy15,16. However, Pt is less efficient for OER at the anode due to its overly strong bonding with oxygen intermediates1,17. In addition, Pt is readily oxidized to soluble high-valence Ptx>4 derivatives (e.g., PtO3) in an excessively oxidative environment at the anode potential of >1.4 V versus a reversible hydrogen electrode (RHE)18,19,20. This in turn could trigger dissolution of active Pt species, resulting in a sharp decrease in catalytic activity18,20,21. Therefore, the rational design of a highly efficient and dissolution-resistant Pt-based OER catalyst requires fine control over the local coordination environment and electronic structure of Pt, lowering the OER overpotential and simultaneously suppressing the over-oxidation of active Pt species during electrolysis. To go further, considering the noble nature of Pt, downsizing the nanometre-scale to single-atoms could effectively decrease Pt usage and maximize atomic utilization efficiency. However, despite tremendous efforts have been dedicated to tuning the electronic structure and coordination of Pt in recent decades15,16,17, whether Pt sites can be tailored to exhibit true state-of-the-art Ir/Ru-like activity on a TOF or MA basis and to what degree they are stable in the OER still remains an open question.

Herein, we present a class of highly active and stable OER catalysts with atomically dispersed Pt sites embedded in cobalt hydrogen phosphate (CoHPO) support with Pt(OH)(O3)/Co(P) coordination. Over a long period of time, platinum has been neglected as a catalyst for the OER as its nanoparticle was tested to be almost inert1,21. In this work, surprisingly, the Pt-based single-site catalysts display 2–4 orders of magnitude higher TOF and MA than Pt/C and Pt nanoparticles on CoHPO and the state-of-the-art Ir/C electrocatalysts, and also show excellent stability. Combining a series of experimental studies and density functional theory (DFT) calculations indicate that the special coordination structure of monatomic Pt sites coupled with strong electronic coupling between isolated Pt atoms and surrounding Co atoms are the origin of much enhanced electrocatalytic activity and stability. Impressively, the catalysts also show remarkable HER performance with a TOF as high as 12.8 s−1 per Pt at −100 mV, about 41 times greater than that of commercial Pt/C. We further assemble an anion-exchange-membrane water electrolyser (AEMWE) using this bifunctional catalyst as both anode and cathode with an ultralow total mass loading of 0.029 mgPt cm−2. The AEMWE achieves an industrial current density of 1 A cm−2 at a low cell voltage of 1.8 volts without iR-correction and durability of more than 100 h, which delivers two-order of magnitude higher performance than commercial Pt/C and Ir/C in terms of mass density and precious metal utilization efficiency.

Results

Synthesis and structural characterizations

Atomic platinum was anchored on CoHPO support (Supplementary Figs. 1 and 2) by mean of an icing-assisted photochemical reduction. In this process, PtCl62− precursor was effectively inhibited to migrate, avoiding the nucleation and growth of platinum species into clusters or particles. Inductively coupled plasma (ICP) spectroscopy result verifies that the platinum concentration on Pt1/CoHPO is 0.57 wt.%. Due to low incorporation of platinum, the X-ray diffraction (XRD) pattern exhibits no perceptible change compared to original CoHPO (Supplementary Fig. 2b). Nanosheet-assembled flower-like Pt1/CoHPO is observed from the transmission electron microscopy (TEM) image (Fig. 1a), without any discernible metallic nanoparticles. Aberration-corrected high-angle annular dark-field (HAADF) scanning TEM (STEM) image (Fig. 1b and Supplementary Fig. 3) show platinum atoms (marked by dashed red cycles) are atomically dispersed on amorphous CoHPO (inset of Fig. 1b). Moreover, the atomic intensity profiles along with the direction of X-Y (Fig. 1c) uncover that each platinum atoms are separated by at least 0.37 nm, exceeding the platinum-effective atomic radius, further corroborating the single-site platinum on the supports. The composition analysis by STEM elemental mapping at larger scale reveals the homogenous dispersion of platinum species in Pt1/CoHPO (Fig. 1d).

Fig. 1: Morphology and fine-structure characterizations of the Pt1/CoHPO catalyst.
figure 1

a TEM image. Sale bar, 0.5 μm. b Representative magnified HAADF-STEM image, showing that only Pt single atoms are present in the CoHPO support (marked by red dotted circles). Inset, the FFT image. Sale bar, 2 nm. c Atomic line-scanning intensity profiles along the X-Y marked by 1 and 2 in b. d Elemental mapping. e k3-weighted Fourier-transformed (FT) of EXAFS spectra at R space. f Fitting curve of EXAFS spectra at R space. Inset, showing the optimized atomic model of Pt1/CoHPO, Pt (dark cyan), Co (blue), O (red), P (pink) and H (white). g XANES spectra at the Pt L3-edge. Inset, the fitted average valence state of Pt from XANES spectra.

X-ray absorption fine structure spectroscopy (XAFS) at Pt L3-edge was used to investigate the local coordination environment of single-atom Pt1 in Pt1/CoHPO. The Fourier transform (FT) of extended XAFS (EXAFS) oscillations (Fig. 1e) reveal that Pt1/CoHPO exhibits a major peak at ~1.62 Å, shorter than the Pt-O peak at ~1.68 Å for the PtO6 octahedra in PtO2, and slightly larger than the Pt-OH peak at ~1.58 Å for the Pt-OH6 octahedra in Pt(OH)4, which can thus be assigned to a mixture of Pt-O and Pt-OH bonding. The coexistence of O2− and OH species was also revealed by O 1 s X-ray photoelectron spectroscopy (XPS) spectrum (Supplementary Fig. 4). In addition, Pt1/CoHPO reveals two minor signals at ~2.22 and ~3.05 Å, attributed to Pt-P and Pt-Co bonding, respectively (Supplementary Fig. 5). Compared with Pt foil and PtO2, there are no characteristic peaks for Pt-Pt or Pt-O-Pt (clustered platinum oxides) scatterings in Pt1/CoHPO, consistent with the result of HAADF-STEM images. Based on the EXAFS best-fitting analyses (Supplementary Table 1), the Pt1 atomic environment within Pt1/CoHPO consists of three O atoms and one OH atom in first coordination shell, and four Co and P atoms in second coordination shell (inset of Fig. 1f). We further identified atomic-site structures of Pt1/CoHPO by X-ray absorption near-edge structure (XANES) simulations, which has high sensitivity to the three-dimensional arrangement of atoms22. The XANES profiles of Pt1/CoHPO can be reproduced fairly well by the structure of a Pt(OH)(O3)/Co(P) moiety embedded in a cobalt hydrogen phosphate support (Supplementary Fig. 9). The interatomic distances predicted by density Functional theory (DFT) are also fully consistent with the structural parameters obtained by FT-EXAFS fitting (Supplementary Fig. 8 and Table 1).

The XANES and XPS measurements were further performed to study the chemical state of Pt1 in Pt1/CoHPO and the electronic coupling between Pt1 and substrate. The normalized XANES spectra show that the white line intensity (black arrow) of Pt1/CoHPO is between platinum powder (0) and PtO2 (+4) (Fig. 1g). The fitted valence state of platinum from XANES spectra is +2.59 (inset of Fig. 1g), being in agreement with the Pt 4 f XPS analysis (Supplementary Fig. 10). Noticeably, a negative shift (0.5 eV) in the Co 2p XPS spectra (Supplementary Fig. 11) is observed upon Pt loading. The increase of Co valence state in Pt1/CoHPO can be attributed to the noble Pt with higher electronegativity, which is attracting electrons from Co through the Co-O-Pt bonds, being in accord with the fact that Pt possess a valance state lower than its initial salt K2PtCl6, implying the strong electronic interactions of Pt and Co atoms23.

As a control, other various single-atom metals anchored on CoHPO (M1/CoHPO), such as Ir1, Ru1, Ni1 and Fe1, were also synthesized using similar procedure with that of Pt1/CoHPO (Methods). The atomically dispersed nature of M1/CoHPO was demonstrated by high-resolution STEM and XAFS spectra (Supplementary Figs. 1215). Meanwhile, as a comparison, the CoHPO with Pt nanoparticles of ~3 nm (1.15 wt.%) was also synthesized via a NaBH4-reduction method (denoted as PtNP/CoHPO, Supplementary Fig. 16).

Electrochemical OER

The OER polarization curves in O2-saturated 0.1 M KOH (Fig. 2a) reveal that the Pt1/CoHPO exhibits an unexpected catalytic activity with a low overpotential of 246 mV (209 mV in 1 M KOH) at 10 mA cm−2, much lower than those of state-of-the-art Ir/C (344 mV) and other M1/CoHPO that include Ir1 (313 mV), Ru1 (355 mV), Ni1 (328 mV) and Fe1 (390 mV), whereas the nanoparticle counterparts Pt/C and PtNP/CoHPO (Supplementary Fig. 18) show very low catalytic activities. The overpotential vs. Tafel slope (Fig. 2b and Supplementary Fig. 19) and electrochemical impedance (Supplementary Fig. 20) further reveal the more favorable OER kinetics of Pt1/CoHPO, in which the Pt1/CoHPO shows the lowest Tafel slope (49.8 mV dec−1) and charge-transfer resistance (8.1 Ω). We note that the OER activity of CoHPO support is two orders of magnitude lower than that of Pt1/CoHPO (Supplementary Fig. 18), suggesting that the atomically dispersed Pt1 species plays a crucial role in the observed high OER activity. Moreover, Pt1/CoHPO exhibits a TOF value of 6.81 ± 0.13 s−1 per Pt atoms and MA of 13.5 ± 0.3 A mg−1Pt at 300 mV in 0.1 M KOH, four orders of magnitude larger than that of Pt/C (7.6 × 10−4 s−1, 1.5 × 10−3 A mg−1Pt) and 52-times larger than that of PtNP/CoHPO (0.13 s−1, 0.26 A mg−1Pt, Supplementary Fig. 18b), and also superior to those of other M1/CoHPO (Fig. 2c) and state-of-the-art Ir/C (0.027 s−1, 0.056 A mg−1Ir). When evaluated in 1 M KOH as the electrolyte, the TOF and MA values increase to as high as 35.1 ± 5.2 s−1 and 69.5 ± 10.3 A mg−1Pt at 300 mV (Fig. 2a and Supplementary Fig. 18c, d), respectively, which are superior to most of the state-of-the-art single-atom oxygen-evolving catalysts and Ir, Ru-based catalysts reported to date (Supplementary Table 3).

Fig. 2: Electrochemical OER performance.
figure 2

a Polarization curves of CoHPO and different M1/CoHPO catalysts tested in 0.1 M and 1 M KOH solutions. Reference samples of Ir/C and Pt/C are also included for comparison. b OER data analysis of overpotentials (obtained from the polarization curves at 10 mA cm−2 current density) and Tafel slopes in 0.1 M KOH. c TOFs curves of different M1/CoHPO catalysts based on the loading amounts of metals at different overpotentials. d Mass activities of Pt1/CoHPO with different Pt loadings at 1.53 V vs. RHE. e The polarization curves of Pt1/CoHPO and Pt/C before and after 5000 (for Pt1/CoHPO) and 50 (for Pt/C) potential cycles tested in 0.1 M KOH. f A high-resolution HAADF-STEM image of Pt1/CoHPO after the durability test, in which the bright contrast spots present the Pt single atoms. Scale bar, 2 nm. Error bars in b, d represent the average values (mean ± s.d., n = 3).

The atomic dispersion of platinum is vital to the high activity of Pt1/CoHPO in the OER. Figure 2d presents the MAs at 1.53 V as a function of the Pt loadings. For the catalysts with low Pt loadings, there are subtle differences in MAs until the loading reaches to ~0.95 wt.%, namely Pt1PtNP/CoHPO, with mixed Pt single atoms and nanoparticles (Supplementary Fig. 21). Specifically, MAs of single-atom samples with a low Pt loadings (0.12, 0.25, 0.43 and 0.57 wt.%) all exceed 13 A mgPt−1 at 1.53 V, decreases to 3.1 A mgPt−1 for Pt1PtNP/CoHPO, and further sharply decreases to 0.26 A mgPt−1 for PtNP/CoHPO (Supplementary Fig. 16). These results clearly show that the single atomic Pt1 is responsible for the high OER performance of Pt1/CoHPO.

We further found that Pt1/CoHPO was quite stable in long-term catalysis. The OER polarization curves obtained before and after the accelerated durability tests of 5000 potential cycles almost overlap, well surpassing that of Pt/C (Fig. 2e). The stability was further demonstrated by the negligible increase in potential during the 48 h test (Supplementary Fig. 22). Moreover, unlike the ~60% dissolution of Pt/C catalysts, no Pt species were detected in Pt1/CoHPO by ICP in electrolyte after the chronoamperometry test (Supplementary Table 4), suggesting the high structural stability of single atomic Pt1 in the CoHPO with strong resistance to dissolution under harsh OER conditions. Atomic-resolution HAADF-STEM images and elemental mappings (Fig. 2f and Supplementary Fig. 23) further confirm the uniformly distributed single atomic Pt1 on the CoHPO after stability test, without noticeable agglomeration of larger Pt species.

Insights into the underlying OER mechanism

The potential-dependent operando attenuated-total-reflection (ATR) Fourier-transform infrared (FTIR) spectroscopy was performed to probe the mechanistic pathway and possible role of single-atom Pt1 in Pt1/CoHPO. No absorption bands were detected on either Pt1/CoHPO or CoHPO at open circuit potential (OCP). For Pt1/CoHPO, no obvious band was observed at potentials of 1.0 and 1.2 V versus RHE, which are more negative than OER thermodynamic potential. A prominent absorption vibration band at ~1083 cm−1, assigned to the stretching vibration of superoxide species (-O-O-) on the Pt surface (refs. 24,25), appears at a potential of 1.4 V and gradually strengthens in intensity with the augment of the applied potentials (from 1.4 V to 1.6 V). When 1.2 and 1.4 V were reversely applied to Pt1/CoHPO, the characteristic bands intensity was identical with the original ones (Fig. 3a). These results demonstrate the formation of a surface intermediate superoxide OOHads species on the Pt1 sites during the OER26, affirming that single-atom Pt1 species in Pt1/CoHPO act as the dominating active site for the OER. Meanwhile, the emergence of OOHads intermediate suggests that an adsorbate evolution mechanism (AEM) pathway rather than a lattice oxygen activation mechanism (LOM) pathway dominated O2 generation over the Pt1/CoHPO, which improved its structural stability under harsh OER conditions (Supplementary Fig. 24, refs. 27,28). By contrast, for CoHPO, only a very weak band at ~1015 cm−1 was observed when potential was applied (Fig. 3b), which corresponds to the vibrational band of Co-OOHads (ref. 29), suggesting insufficient OOHads generation and sluggish kinetics on the cobalt sites.

Fig. 3: Operando ATR-FTIR and XAFS spectra of catalysts under different applied potentials.
figure 3

a, b Operando potential-dependent ATR-FTIR spectra for Pt1/CoHPO (a) and CoHPO (b). cf, Operando potential-dependent XAFS spectra of the XANES (c, e) and FT-EXAFS (d, f) at the Pt L3-edge (c, d) and Co K-edge (e, f) of Pt1/CoHPO.

We further applied operando XAFS techniques to investigate the potential-dependent oxidation of Pt1 during OER (Supplementary Fig. 25). With the potential increased from OCP to 1.6 V, the oxidation state of Pt1 in Pt1/CoHPO appeared relatively stable, lower than that of reference PtO2 (+4), as revealed by the essentially identical normalized white-line intensities (Fig. 3c) and local coordination (Fig. 3d) at the Pt L3-edge. In contrast, an appreciably shift to higher energy in the absorption threshold of the Co K-edge (Fig. 3e) indicates a decrease of d-band occupy states for Co atoms in the CoHPO substrate, probably attributed to the electron transferring from adjacent Co to Pt1. This is further corroborated by surface XPS analysis (Supplementary Fig. 25), in which Pt1 almost kept the initial oxidation state after running a chronoamperometry test, without transforming into an unstable phase of Ptx>4 derivative, considering the possible charge compensation from Co to Pt1 induced by the strong metal-support interactions, thereby preventing the over-oxidation and dissolution of Pt1. In addition, a shrinkage of the Co-O bond is also observed with increasing the potential (Fig. 3f), which could further fix the isolated Pt1 atoms on the surface of Pt1/CoHPO, avoiding possible migration and agglomeration during OER.

Theoretical investigations based on first-principles calculations were further implemented to rationalize the observed activity and dissolution resistance of single-atom Pt sites in Pt1/CoHPO. A four-electron AEM reaction pathway proposed in the operando ATR-FTIR spectroscopy is considered (Fg. 4a). The OH-covered Pt (111) surface is chosen to model the OER reaction at Pt (111) in an alkaline condition. On the clean Pt(111) surface, we find that the OER overpotential is determined by the adsorption of *OOH intermediate on the surface (Fig. 4b). The over-binding of *O intermediate on the surface yields a high overpotential UOP = 1.278 V. In a sharp contrast, *O binding on the OH-covered Pt(111) surface is too weak due to *O-OH interaction on the surface, which also leads to a high overpotential (UOP = 0.963 V). These results are consistent with the classic volcano picture — binding too strongly or too weakly lower the catalytic activity, which explains why the conventional Pt electrodes are inefficient for OER. On the other hand, thanks to the single-atom Pt sites with Pt(OH)(O3)/Co(P) coordination, *O binding on the Pt1/CoHPO surface is neither too strong or too weak, resulting in a much lower overpotential UOP = 0.378 V. To paint a complete physical picture, we also calculate the projected crystal orbital Hamilton population (pCOHP) for Pt-O (red) and Pt-OH (black) bonds on the three surfaces30,31. Since more bonding states and fewer anti-bonding states yield stronger bonding and vice versa, the pCOHP analysis is directly connected to the binding energy results (Fig. 4c). On clean Pt(111) surface, since pCOHP values near EF for the antibonding states of Pt-O bonds are very low, one would expect strong (or over)-binding of *O. In contrast, on OH-covered Pt(111) surface, a large number of antibonding states is present for the Pt-O bonds, leading to weak binding of *O. Interestingly, on Pt1/CoHPO surface, the pCOHP values for the antibonding states of Pt-O and Pt-OH bonds are comparable, suggesting similar binding energies of *O and *OH intermediates, which in turn gives rise to a lower overpotential. Therefore, the binding energy results are consistent with the premises of the d-band theory although the latter is not explicitly considered here.

Fig. 4: DFT simulations of catalytic activity and electronic structure.
figure 4

a The OER reaction pathway on Pt1/CoHPO. The adsorption structures of intermediates on the HO1-Pt1-O3/Co(P) sites are shown in the insets. Cyan, red, and white spheres represent Pt, O, and H atoms, respectively. b The OER free energy diagram for Pt (111), OH-covered Pt (111), and Pt1/CoHPO surfaces at equilibrium potential of 1.23 V. c The projected crystal orbital Hamilton population (pCOHP) for Pt-O bond in the *OH and *O adsorbed Pt (111), Pt (111) (*OH), and Pt1/CoHPO surfaces. Bonding and antibonding states are shown on the right and left, respectively. d Overpotential contour map in terms of the free energies of *OH and difference between *O and *OH.

Using the linear scaling relation ΔG*OOH = 0.97ΔG*OH + 3.24 fitted on the basis of the established M1/CoHPO models (Supplementary Fig. 27)32, we can further construct a volcano plot for OER (Fig. 4d). The constructed volcano is found to provide a reasonable estimate of OER activity on a number of similar catalysts. In particular, Pt1/CoHPO sits very close to the top (blue area) of the volcano. On the other hand, Pt(111), Ru1, Fe1 and Ir1/CoHPO surfaces constitute the left-leg (or over-binding) of the volcano, whereas Ni1/CoHPO is on the right-leg of the volcano (weak-binding). Pt1/CoHPO is predicted to be at the volcano top, agrees well with the experimental observations.

Finally, by Bader charge analysis, we find that the adsorption of an oxygen atom gains ~0.45 e from the Pt1/CoHPO surface, with ~0.32 e from the single-atom Pt site, and ~0.13 e from the CoHPO substrate. In other words, the substrate acts as an electron reservoir and donates electrons to the reaction intermediates, which can help suppress the over-oxidation and dissolution of Pt1 on the surface, giving rise to much improved stability of Pt1/CoHPO catalyst as observed in our experiments.

Performance in water electrolyzers device

In addition to its promising OER property, Pt1/CoHPO catalyst also achieves an exceptional low overpotential of 49 mV (in 0.1 M KOH) to deliver 10 mA cm−2, a TOF as high as 12.8 s−1 at −100 mV as well as superior durability for the HER (Supplementary Figs. 28 and 29). Our Pt1/CoHPO is no doubt one of the most efficient bifunctional catalysts toward both OER and HER, exceeding most of the reported state-of-art catalysts (Supplementary Tables 3 and 5). In a proof-of-principle demonstration of its application, we leveraged the bifunctional catalytic activity of Pt1/CoHPO, and used it as both anode and cathode for water electrolysis. We first assessed the catalytic properties in a two-electrode lab configuration. Compared to the benchmark Pt/C + Ir/C catalyst, it exhibits a 130 mV smaller potential to achieve 10 mA cm−2 as well as enhanced durability (Fig. 5a, Supplementary Fig. 30). The Faradaic efficiency of the production of hydrogen was nearly 100% (Supplementary Fig. 31).

Fig. 5: AEMWE performance.
figure 5

a Electrocatalytic overall water-splitting properties in a H-type cell of the Pt1/CoHPO as both anodic and cathodic catalysts. The benchmark Pt/C + Ir/C as a catalyst is also included for comparison. b Electrocatalytic water-splitting properties of the Pt1/CoHPO and the benchmark Pt/C + Ir/C measured in an alkaline AEMWE setup operating at 80 °C. Inset, a typical single AEMWE setup comprising a membrane electrode assembly (MEA) and bipolar plates (BP) with a flow field are presented, wherein the MEA comprises a gas diffusion layers with a Ti felt and a carbon fiber paper (CFP) at the anodic and cathodic sides, respectively, anodic and cathodic catalyst layers and an anion exchange membrane (AEM). c Mass activities comparisons of the Pt1/CoHPO and the commercial Pt/C + Ir/C at various cell voltages, showing over two orders of magnitude higher of mass activity for the Pt1/CoHPO compared with that of commercial Pt/C + Ir/C. d Stability tests of the Pt1/CoHPO-based MEA. Error bars in b, c represent the average values (mean ± s.d., n = 3).

We further integrated our Pt1/CoHPO into a membrane-electrode-assembly (MEA) to assemble an alkaline AEMWE device (Supplementary Fig. 32 and 33). Figure 5b shows the water-splitting performance of the Pt1/CoHPO-based and the benchmark Pt/C + Ir/C-based AEMWE setup operating in 0.1 M KOH at 80 °C (The error bars show the standard deviation for three independently integrated and tested cells). In addition to an industrial current density of 1 A cm−2 at a low cell voltage of 1.8 V (Fig. 5b and Supplementary Fig. 34), the Pt1/CoHPO-based MEA containing an ultralow total Pt loading of ~29 μg cm−2 (11.4 μg cm−2 for cathode and 17.1 μg cm−2 for anode) delivers a much higher current density than the MEA that uses benchmark Pt/C + Ir/C (0.2 A cm−2 at 1.8 V) in spite of its high precious metal loading of ~1 mg cmPt+Ir−2. Mass activities using Pt1/CoHPO-based MEA reach 15.7 ± 4.5 and 64.5 ± 5.8 A mgPt−1 at 1.7 and 1.9 V (Fig. 5c), respectively, which are more than two-order of magnitude higher than those of commercial Pt/C + Ir/C. Furthermore, the Pt1/CoHPO-based MEA can be stably operated for at least 100 h without apparent loss of the current density (Fig. 5d).

Discussion

In summary, we report a type of anti-dissolution Pt single site with Pt(OH)(O3)/Co(P) coordination embedded in cobalt hydrogen phosphate as a highly efficient bifunctional catalyst for both OER and HER catalysis. Studies of the mechanism by employing advanced operando ATR-FTIR and X-ray absorption spectroscopic techniques demonstrate that Pt single atoms provide predominant OER active sites and the oxidation state of Pt is kept under 4+ within the OER catalysis potential range. The DFT calculations reveal that the highly enhanced activity and stability on Pt1/CoHPO originate from the unique coordination of Pt(OH)(O3)/Co(P) sites, and from the strong electronic coupling between isolated Pt atoms and surrounding Co atoms. The resultant charge redistribution not only optimizes the binding energies of oxygenated intermediates, lowers the OER energy barriers, but also prevents the formation of soluble high valence Ptx>4 by tuning the redox behaviour of Pt. The Pt-based single-atom catalyst with an ultralow total Pt loading enables stable and high-performance water splitting using an alkaline AEM electrolyser as both anode and cathode. This work provides fundamental and technological insights for developing highly efficient catalysts with ultralow Pt loading for renewable energy conversion devices.

Methods

Preparation of M1/CoHPO

As-prepared CoHPO (20 mg, Supplementary methods) was first ultrasonically dispersed in deionized water (15 mL) in a beaker. A controlled amount of 2 mg mL−1 K2PtCl6 was then dropwise added into the solution, and stirred for 12 h. Subsequently, the solution was rapidly frozen by a liquid nitrogen (−196 °C) bath. The mixed solution could be frozen quickly to form ice, which ensures the homogeneous distribution of Pt precursor in the solution. The frozen ice was then irradiated in a 300 W Xe light (PLS-SXE300D) equipped with 420 nm and IR light filters. After irradiation for about 7 min, the ice layers were naturally melted into solution. The product was collected by centrifugation and washed via ethanol for three times, and dried in an oven at 60 °C overnight. The final Pt loading of Pt1/CoHPO was determined by ICP-AES.

The preparation of other metals (Ni, Fe, Ru and Ir) single atoms on the CoHPO support was similar to that of Pt1/CoHPO, except that NiCl2·6H2O, FeCl2·4H2O, RuCl3·xH2O, and IrCl3·xH2O solutions were used as Ni, Fe, Ru and Ir precursors, respectively.

Electrochemical measurements

Electrochemical tests were applied using a typical three-electrode cell configuration on CHI 660E instruments. A graphite carbon rod was used as the counter electrode, and a Hg/HgO electrode (1 M KOH) were used as the reference electrode. The preparation of the working electrodes containing the examined electrocatalysts can be found as follows. A suspension was prepared by dispersing as-prepared catalysts (2.5 mg) and carbon black (0.5 mg, Ketjen Black-300J) in a 0.5 ml mixed solvent containing 140 μL ultrapure water, 340 μL ethanol and 20 μL 5 wt.% Dupont Nafion D521 solution. The mixed solution was then ultrasonicated for 0.5 h to obtain a uniform catalyst ink. After that, a certain volume of dispersion was pipetted onto a piece of clean carbon fiber paper and then dried in air. The catalysts loading was determined as 1 mg cm−2. N2 and O2-saturated 0.1 M KOH solutions were chosen as electrolytes for the HER and OER, respectively. The polarization curves were collected at a scan rate of 5 mV s−1 with 95% iR-correction. Before collecting, several fast cyclic voltammograms (CVs, 500 mV s−1) were taken to clean and stabilize the electrocatalyst surface till steady-state was received. The Nyquist plots were performed at frequencies ranging from 100,000 Hz to 0.1 Hz with an amplitude voltage of 5 mV. The overall-water-splitting performances were collected in 0.1 M and 1 M KOH using the Pt1/CoHPO catalyst as both anode and cathode in a H-type electrolytic cell separated by an anion-exchange membrane (fumasep, 130 μm) with iR-correction. All recorded potentials (E) were converted to the reversible hydrogen electrode (RHE) based on the following equation: ERHE = EHg/HgO + 0.059 pH + E0Hg/HgO. The potential of reference electrode and the pH value of the used electrolyte were examined before the electrochemical measurement (Supplementary Fig. 17).

The turnover frequency (TOF, s−1) per metal site were calculated according to the following equation33,34:

$${{{{{\rm{TOF}}}}}}=j/(n\times F\times N)$$
(1)

where j is the geometric current density after 95% iR-correction, n is the number of electrons transferred in the reactions (4 for OER and 2 for HER), F is the Faraday constant, and N is the molar number of active sites. The number of active sites was calculated on the basis of the hypothesis that all single-atoms in M1/CoHPO acted as active centers and neglected the contribution of the CoHPO support4,5,15. Specifically, the active site number N was estimated via the total catalyst mass loading on the electrode (m, mg cm−2), the weight percent of active metals in the catalysts (w, wt.%, determined by the ICP measurement) and the molar mass (M, g mol−1) according to the following equation:

$$N=m\times w/M$$
(2)

Mass activity (MA, A g−1) was estimated via the catalyst loading (m, mg cm−2), weight percent of metal in the catalysts (w, wt.%) and the geometric current density (j, mA cm−2) under a certain overpotential based on the following equation:

$${{{{{\rm{MA}}}}}}=j/(m\times w)$$
(3)

Ex situ XAFS experiments and data processing

The XAFS spectra data (Pt L3-edge, Ru, and Ni K-edges) were performed at the 1W1B beamline of Beijing Synchrotron Radiation Facility (BSRF, 2.5 GeV, a maximum current of 250 mA, Si (311) double-crystal). The ATHENA module of the IFEFFIT software packages was applied to process the acquired XAFS raw data (background subtraction, normalization and Fourier transformation) according to standard procedures35. Least-squares curve-fitting of EXAFS data was performed using an ARTEMIS program36. Fourier transformation of k3-weighted EXAFS oscillations, k3χ(k), from k space to R space was performed to obtain a radial distribution function. The data were fitted in R-space with theoretical models constructed based on the crystal structure derived from DFT optimization. Many-body amplitude-reduction factor (S02) was fixed to 0.787 as determined from Pt foil fitting. Fitting range for Pt1/CoHPO was 3 ≤ k (/Å) ≤ 10.5 and 1.2 ≤ R (Å) ≤ 4.0.

Operando XAFS experiments

The Pt (11,564 eV) L3-edge and Co (7,709 eV) K-edge XAFS data were collected at the BL11B station in the Shanghai Synchrotron Radiation Facility (SSRF). A Si (111) double-crystal monochromator was utilized with the energy calibrated using Pt and Co foils. The operando XAFS experiment was conducted in a sensitive fluorescence mode using a homemade electrochemical cell and a computer-controlled electrochemical analyzer, and all the data were collected during one period of beam time. An as-prepared Pt1/CoHPO/thin carbon paper as the working electrode was in contact with a slip of copper tape and fixed with kapton tape to the exterior of the wall of the cell, with the Pt1/CoHPO layer facing inwards. A Pt foil and a Hg/HgO electrode were used as counter electrode and reference electrode, respectively, and a 0.1 M KOH solution was used as the electrolyte.

Operando ATR-FTIR measurements

Electrochemical operando ATR-FTIR spectroscopy was performed on a Bruker 70 V FTIR spectrometer with an attenuated-toal-reflection (ATR) configuration using a liquid nitrogen cooled MCT detector. Based on a specially-made three-electrode thin-layer FTIR cell configuration, a FITR radiation sequentially passed by the CaF2 window and thin-layer solutions, and then it was reflected by the electrode surfaces. Before each FTIR data collection, the 0.1 M KOH electrolyte was saturated with O2 gas, and several fast cyclic voltammograms (500 mV s−1) between 0.05 and 1.35 V versus RHE were first taken to obtain a stable current response. The FTIR spectra were collected at an open-circuit voltage, and a potential range of 1.0-1.6 V versus RHE with an interval of 0.1 V. Each infrared absorption spectrum was acquired at a resolution of 4 cm−1.

MEA fabrication and evaluation

Pt1/CoHPO catalyst was integrated into a MEA to assemble an AEMWE device, and evaluated its electrochemical performance using an 850e fuel cell test system equipped with 885-HS potentiostat. Alkymer (Ionomr Innovations Inc.) was used as the anion exchange membrane (AEM) and immersed in 1 M KOH solution for at least 24 h prior to being used to exchange Cl into OH. To prepare the catalyst ink, 20 mg of Pt1/CoHPO catalyst and 2 mg of Ketjen Black were homogeneously dispensed into 6 mL of mixed solvent containing ultrapure water, ethanol and Dupont D521 under sonication for about 1 h at room temperature. Both anodic and cathodic electrodes were fabricated by the catalyst-coated membrane (CCM) method. Specifically, membrane pieces were placed and taped flat onto a hotplate (P-2030, Polish) with an exposed membrane surface area of 1.0 cm2. The temperature of the hotplate was 80 °C, ensuring that the solvent evaporated quickly. In addition, vacuum was also used to confirm the membrane firmly adsorbed on the hotplate. The well dispersed catalyst ink was then sprayed with a spray gun (Mr. Hobby, Japan) until the loading was ~2 mg cm−2 (11.4 μgPt cm−2) for cathode. For the anodic electrode, the catalytic ink was coated on the opposite membrane side by the same process described above with the loading of ~3 mg cm−2 (17.1 μgPt cm−2). Commercial IrOx and Pt/C electrodes were also prepared in the above method using Ir oxide (Alfa Aesar) and Pt/C (20 wt.% Pt, Alfa Aesar), respectively. At the end, the as-prepared MEA was assembled into a homemade integrated AEMWEs device with titanium mesh and carbon paper (GDL-29BC, Sigracet) as the anodic and cathodic gas diffusion layer, respectively. Electrically insulating gaskets were also placed to prevent the liquid and gas from escaping through any space between flow fields. The torque applied when assembling the cell was ~5 Nm. 0.1 M KOH was circulated through the anodic side by a peristaltic pump. Polarization curves were collected from 1.0 V to 2.2 V at a water temperature from 25 °C to 80 °C under an ambient pressure. A stability test was carried out by potentiostatic electrolysis at a constant cell voltage of 1.9 V at about 30 °C. Before the stability curves were obtained, a potentiostatic electrolysis was held for 20 min to reach a steady state.

DFT calculations

DFT calculations were performed using the Vienna ab initio Simulation Package (VASP) with the projector-augmented wave (PAW) pseudopotentials37,38. The exchange-correlation interaction was described by using the Perdew-Burke-Ernzerhof (PBE) functional39. The plane-wave energy cut-off was taken as 400 eV. The Brillouin zone was sampled based on the Monkhorst-Pack scheme with a 3 × 3 × 1 k-point mesh40. The Pt1/CoHPO slab model was constructed based on the experimental results of the EXAFS spectra (Supplementary Fig. 7). The model structure was further examined by calculation of the XANES spectra of the Pt atoms using FEFF8 (Supplementary Fig. 9)41. The adjacent slabs were separated by a 15 Å vacuum in the normal direction. The contributions of zero-point energy, entropic and solvent corrections to the free energies of the reaction intermediates were considered42. The free energy change in each reaction step was calculated based on the computational hydrogen electrode (CHE) model32. In particular, the free energy of an electron-proton pair in the CHE model is computed as a half of the free energy of H2 molecule at the standard conditions, which is then shifted by –eU upon an applied external potential U.

Data availability

The data supporting the findings of this study are available within the article and its Supplementary Information. Source data are provided with this paper.

Change history

References

  1. Seh, Z. W. et al. Combining theory and experiment in electrocatalysis: Insights into materials design. Science 355, eaad4998 (2017).

    PubMed  Google Scholar 

  2. Kibsgaard, J. & Chorkendorff, I. Considerations for the scaling-up of water splitting catalysts. Nat. Energy 4, 430–433 (2019).

    ADS  Google Scholar 

  3. Guan, J. et al. Water oxidation on a mononuclear manganese heterogeneous catalyst. Nat. Catal. 1, 870–877 (2018).

    CAS  Google Scholar 

  4. Liu, C. et al. Oxygen evolution reaction over catalytic single-site Co in a well-defined brookite TiO2 nanorod surface. Nat. Catal. 4, 36–45 (2020).

    Google Scholar 

  5. Wang, Q. et al. Single iridium atom doped Ni2P catalyst for optimal oxygen evolution. J. Am. Chem. Soc. 143, 13605–13615 (2021).

    CAS  PubMed  Google Scholar 

  6. Wang, Q. et al. Ultrahigh-loading of Ir single atoms on NiO matrix to dramatically enhance oxygen evolution reaction. J. Am. Chem. Soc. 142, 7425–7433 (2020).

    CAS  PubMed  Google Scholar 

  7. Lai, W. H. et al. General pi-electron-assisted strategy for Ir, Pt, Ru, Pd, Fe, Ni single-atom electrocatalysts with bifunctional active sites for highly efficient water splitting. Angew. Chem. Int. Ed. 58, 11868–11873 (2019).

    CAS  Google Scholar 

  8. Li, P. et al. Boosting oxygen evolution of single-atomic ruthenium through electronic coupling with cobalt-iron layered double hydroxides. Nat. Commun. 10, 1711 (2019).

    ADS  PubMed  PubMed Central  Google Scholar 

  9. Cao, L. et al. Dynamic oxygen adsorption on single-atomic Ruthenium catalyst with high performance for acidic oxygen evolution reaction. Nat. Commun. 10, 4849 (2019).

    ADS  PubMed  PubMed Central  Google Scholar 

  10. Zhang, B. et al. Homogeneously dispersed multimetal oxygen-evolving catalysts. Science 352, 333–337 (2016).

    ADS  CAS  PubMed  Google Scholar 

  11. Chen, J. et al. Low‐Coordinate Iridium Oxide confined on graphitic carbon nitride for highly efficient oxygen. Evolution. Angew. Chem. Int. Ed. 58, 12540–12544 (2019).

    CAS  Google Scholar 

  12. Lin, Y. et al. Chromium-ruthenium oxide solid solution electrocatalyst for highly efficient oxygen evolution reaction in acidic media. Nat. Commun. 10, 1–13 (2019).

    ADS  Google Scholar 

  13. Yao, Y. et al. Engineering the electronic structure of single atom Ru sites via compressive strain boosts acidic water oxidation electrocatalysis. Nat. Catal. 2, 304–313 (2019).

    CAS  Google Scholar 

  14. Fan, J. et al. Bridging the gap between highly active oxygen reduction reaction catalysts and effective catalyst layers for proton exchange membrane fuel cells. Nat. Energy 6, 475–486 (2021).

    ADS  CAS  Google Scholar 

  15. Jiang, K. et al. Single platinum atoms embedded in nanoporous cobalt selenide as electrocatalyst for accelerating hydrogen evolution reaction. Nat. Commun. 10, 1743 (2019).

    ADS  PubMed  PubMed Central  Google Scholar 

  16. He, T. et al. Mastering the surface strain of platinum catalysts for efficient electrocatalysis. Nature 598, 76–81 (2021).

    ADS  CAS  PubMed  Google Scholar 

  17. Gao, R. et al. Pt/Fe2O3 with Pt–Fe pair sites as a catalyst for oxygen reduction with ultralow Pt loading. Nat. Energy 6, 614–623 (2021).

    ADS  CAS  Google Scholar 

  18. Favaro, M. et al. Elucidating the alkaline oxygen evolution reaction mechanism on platinum. J. Mater. Chem. A 5, 11634–11643 (2017).

    CAS  Google Scholar 

  19. Damjanovic, A. et al. Kinetics of oxygen evolution and dissolution on platinum electrodes. Electrochim. Acta 11, 791–814 (1966).

    CAS  Google Scholar 

  20. Arrigo, R. et al. In situ study of the gas-phase electrolysis of water on platinum by NAP-XPS. Angew. Chem. Int. Ed. 52, 11660–11664 (2013).

    CAS  Google Scholar 

  21. Reier, T. et al. Electrocatalytic oxygen evolution reaction (OER) on Ru, Ir, and Pt catalysts: a comparative study of nanoparticles and bulk materials. ACS Catal. 2, 1765–1772 (2012).

    CAS  Google Scholar 

  22. Fei, H. et al. General synthesis and definitive structural identification of MN4C4 single-atom catalysts with tunable electrocatalytic activities. Nat. Catal. 1, 63–72 (2018).

    CAS  Google Scholar 

  23. Hao, S. et al. Dopants fixation of Ruthenium for boosting acidic oxygen evolution stability and activity. Nat. Commun. 11, 5368 (2020).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  24. Briega-Martos, V. et al. Detection of superoxide anion oxygen reduction reaction intermediate on Pt (111) by infrared reflection absorption spectroscopy in neutral pH conditions. J. Phys. Chem. Lett. 12, 1588–1592 (2021).

    CAS  PubMed  PubMed Central  Google Scholar 

  25. Kukunuri, S. & Noguchi, H. In situ spectroscopy study of oxygen reduction reaction intermediates at the Pt/acid interface: Surface-enhanced infrared absorbance spectroscopy. J. Phys. Chem. C. 124, 7267–7273 (2020).

    CAS  Google Scholar 

  26. Cheng, W. et al. Lattice-strained metal-organic-framework arrays for bifunctional oxygen electrocatalysis. Nat. Energy 4, 115–122 (2019).

    ADS  CAS  Google Scholar 

  27. Zagalskaya, A. & Alexandrov, V. Role of defects in the interplay between adsorbate evolving and lattice oxygen mechanisms of the oxygen evolution reaction in RuO2 and IrO2. ACS Catal. 10, 3650–3657 (2020).

    CAS  Google Scholar 

  28. Grimaud, A. et al. Activating lattice oxygen redox reactions in metal oxides to catalyse oxygen evolution. Nat. Chem. 9, 457–465 (2017).

    CAS  PubMed  Google Scholar 

  29. Zhang, M. et al. Time-resolved observations of water oxidation intermediates on a cobalt oxide nanoparticle catalyst. Nat. Chem. 6, 362–367 (2014).

    CAS  PubMed  Google Scholar 

  30. Deringer, V. L. et al. Crystal orbital Hamilton population (COHP) analysis as projected from plane-wave basis sets. J. Phys. Chem. A 115, 5461–5466 (2011).

    CAS  PubMed  Google Scholar 

  31. Maintz, S. et al. LOBSTER: A tool to extract chemical bonding from plane-wave based DFT. J. Comput. Chem. 37, 1030–1035 (2016).

    CAS  PubMed  PubMed Central  Google Scholar 

  32. Man, I. C. et al. Universality in oxygen evolution electrocatalysis on oxide surfaces. ChemCatChem 3, 1159–1165 (2011).

    CAS  Google Scholar 

  33. Li, S. et al. Oxygen-evolving catalytic atoms on metal carbides. Nat. Mater. 20, 1240–1247 (2021).

    ADS  CAS  PubMed  Google Scholar 

  34. Zhang, B. et al. High-valence metals improve oxygen evolution reaction performance by modulating 3d metal oxidation cycle energetics. Nat. Catal. 3, 985–992 (2020).

    CAS  Google Scholar 

  35. Newville, M. IFEFFIT: interactive XAFS analysis and FEFF fitting. J. Synchrotron Radiat. 8, 322–324 (2001).

    CAS  PubMed  Google Scholar 

  36. Ravel, B. & Newville, M. ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using IFEFFIT. J. Synchrotron Radiat. 12, 537–541 (2005).

    CAS  PubMed  Google Scholar 

  37. Kresse, G. & Furthmüller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B 54, 11169 (1996).

    ADS  CAS  Google Scholar 

  38. Blöchl, P. E. Projector augmented-wave method. Phys. Rev. B 50, 17953 (1994).

    ADS  Google Scholar 

  39. Perdew, J. P. et al. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865 (1996).

    ADS  CAS  PubMed  Google Scholar 

  40. Monkhorst, H. J. & Pack, J. D. Special points for Brillouin-zone integrations. Phys. Rev. B 13, 5188 (1976).

    ADS  MathSciNet  Google Scholar 

  41. Ankudinov, A. et al. Parallel calculation of electron multiple scattering using Lanczos algorithms. Phys. Rev. B 65, 104107 (2002).

    ADS  Google Scholar 

  42. Calle-Vallejo, F. et al. Density functional studies of functionalized graphitic materials with late transition metals for oxygen reduction reactions. Phys. Chem. Chem. Phys. 13, 15639–15643 (2011).

    CAS  PubMed  Google Scholar 

Download references

Acknowledgements

This work was supported from the National Natural Science Fund for Distinguished Young Scholars (Grant No. 52025133 to S.G.), the National Natural Science Foundation of China (Grant No. 52002005 to L.Z., Grant No. 22102077 to Z.Z.), the Tencent Foundation through the XPLORER PRIZE, Sponsored by CNPC Innovation Found (Grant No. 2021DQ02-1002 to S.G.) and the China Postdoctoral Science Foundation (Grant No. 2020M680199 to L.Z.). The work at California State University Northridge was supported by the NSF-PREM program (Grant No. DMR-1828019 to G.L.). The authors thank the photoemission photoendstations BL11B in the Shanghai Synchrotron Radiation Facility (SSRF) and BL1W1B in the Beijing Synchrotron Radiation Facility (BSRF) for the help with characterizations.

Author information

Authors and Affiliations

Authors

Contributions

S.G. supervised the research. S.G. and L.Z. designed and performed most of the experiments and data analysis. G.L. and Z.Z. performed the DFT simulations. L.Z., K.S. and J.L. conducted the the X-ray absorption fine structure characterization and corresponding data analysis. L.Z., F.L. and S.L. carried out the membrane-electrode-assembly experiments. L.G., Q.Z., K.W., Y.S., Q.H. and Y.C. performed the electron microscopy characterization. Z.X. provided suggestions on the manuscript. S.G. and L.Z. co-wrote the manuscript. All authors discussed the results and assisted during manuscript preparation.

Corresponding authors

Correspondence to Gang Lu or Shaojun Guo.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Shi-Hsin Lin and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

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

Supplementary information

Source data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Zeng, L., Zhao, Z., Lv, F. et al. Anti-dissolution Pt single site with Pt(OH)(O3)/Co(P) coordination for efficient alkaline water splitting electrolyzer. Nat Commun 13, 3822 (2022). https://doi.org/10.1038/s41467-022-31406-0

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1038/s41467-022-31406-0

Comments

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.

Search

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