Single-phase perovskite oxide with super-exchange induced atomic-scale synergistic active centers enables ultrafast hydrogen evolution

The state-of-the-art active HER catalysts in acid media (e.g., Pt) generally lose considerable catalytic performance in alkaline media mainly due to the additional water dissociation step. To address this issue, synergistic hybrid catalysts are always designed by coupling them with metal (hydro)oxides. However, such hybrid systems usually suffer from long reaction path, high cost and complex preparation methods. Here, we discover a single-phase HER catalyst, SrTi0.7Ru0.3O3-δ (STRO) perovskite oxide highlighted with an unusual super-exchange effect, which exhibits excellent HER performance in alkaline media via atomic-scale synergistic active centers. With insights from first-principles calculations, the intrinsically synergistic interplays between multiple active centers in STRO are uncovered to accurately catalyze different elementary steps of alkaline HER; namely, the Ti sites facilitates nearly-barrierless water dissociation, Ru sites function favorably for OH* desorption, and non-metal oxygen sites (i.e., oxygen vacancies/lattice oxygen) promotes optimal H* adsorption and H2 desorption.


11.
In the experimental section you should state the size/area of the working electrode to aid reproducibility. 12. A general note of the figures: The large half-filled circles used in many of the plots makes reading the data and seeing the error bars difficult. A small cross or similar would improve readability 13. Generally, this paper is well-written but there are several instances of poor/inaccurate grammar and wording which should be caught with a thorough proof-reading.
Overall, this is an excellent contribution to the literature and will be of significant interest to those working in electrocatalysis.
Reviewer #2 (Remarks to the Author): The authors investigated the performance and origin of the single-phase perovskite oxides for hydrogen evolution reaction in alkaline media demonstrating the superior catalytic activity and the stability. The authors can identify the super-exchange interactions by Ru dopant in SrTiO3 and the concentration of oxygen vacancies using experimental and theoretical calculation results. They showed the single-phase oxides can be more active than other reported composite materials possessing the different activate sites for reaction intermediates. The reviewer has following comments and questions that might bring discussion on the optimization and extensive understanding on their behavior.
Comments: 1. The authors confirmed that co-existence and modification of charge state on the surface of the catalyst can provide reaction sites for both OH-adsorption and water dissociation. In this regard, the activity or efficiency could be influenced by the number or density of two different sites on the surface. It would be also interesting to analyze and demonstrate that Ru dopants are equivalently occupied in the bulk region and subsurface region of the catalyst from the experimental or DFT calculations. 2. The authors clearly showed that Ru doping on SrTiO3 can be advantageous in presence of superexchange behavior between Ru and Ti. The reviewer is also wondering if there are clue and hint predicting the similar behavior of other dopants to Ru ions with Ti ions. For example, can we expand the property of each component such as the orbital alignment of d-orbitals between Ti and Ru to HER catalytic property or adsorption behavior?
Reviewer #3 (Remarks to the Author): In the manuscript, the authors have synthesized single-phase HER catalyst, SrTi0.7Ru0.3O3-δ (STRO) perovskite oxide. Induced by a unique super-exchange interaction, the STRO perovskite exhibits outstanding HER activity with a low overpotential of 46 mV at 10 mA cm-2 and Tafel slope of 40 mV dec-1 in 1 M KOH. The super-exchange interaction in the STRO perovskite was proved and the underlying alkaline HER mechanism on STRO perovskite oxide was investigated via DFT calculations. However, some important concerns look strange or wrong, which could result from less rigorous experimental protocols. As such the paper cannot be published in Nature Communications. Some detailed questions are listed below: 1.The onset overpotential of Pt/C is lower than STRO, however, as the current density reaches about 40 mA/cm2, the overpotential of Pt/C starts to be higher than STRO. Why would that happen? Moreover, the Tafel slope of Pt/C should be added to compare with other samples. 2.The shape of CV plots in the EDLC measurements (Figure S8a c) look strange/wrong. EDLC is expected to observe a rectangular CV profile, in which the current should not change with the potential. 3.Since the Ti3+ and Ru5+ induced by super-exchange effect could promote the HER activity of STRO, then why with the amount of Ru increased to 0.4, the HER activity turned out to be inhibited? 4.As the authors demonstrated, the H2O molecular is first absorbed on Ti3+, but it is usually supposed that O is easier to be bonded with metal ions with higher valence. Why in this system O is first bonded with Ti3+ instead of Ru5+? 5.The language needs additional polishing.

Point-to-Point Responses to Reviewers' Comments and Suggestions
First of all, we thank the reviewers for their valuable comments and suggestions, which have resulted in modifications that have improved the quality and clarity of this paper. To address the specific concern/point clearly, we have separated the referees' comments into question areas and answered them in turn as follows.

Reviewer #1
GENERAL COMMENTS: This work presents a very active and novel electrocatalyst for the HER. The primary claim of novelty is that their STRO material acts as a single-phase synergistic electrocatalyst. The activity of this electrocatalyst is the highest of any perovskite electrocatalyst in the literature and amongst the highest non-precious metal electrocatalysts reported. The data supporting this claim are very strong and this catalyst does appear to be very exciting. I can see no issue with the experimental approaches used to collect this data and am impressed with the breadth of analysis used. I believe this paper contributes not only a very good electrocatalyst, but also a strong mechanistic and theoretical basis for this activity.
The authors describe a new method through which one can optimise a perovskite electrocatalyst -that of synergistic coupling of two B-site cations and lattice oxygen / vacancies. For these reasons, I recommend publication of the manuscript in Nature Communications with some changes noted below.

Response:
We very much appreciate the Reviewer #1's highly positive comments and recommendation for publication. Our responses to his/her comments per point are as follows:

Response to C1:
We thank the reviewer for pointing this out. We have modified related statements in the revised manuscript (Line 87-88, Page 4): "The binding energy of OHshould be optimal for a high-performance catalyst, so that more active sites can be exposed for the full reaction to proceed. 20 " Comment 2: At the top of page 5 it is implied that this is a cost-effective catalyst. However, there is no cost comparison in the paper to Pt or other high-activity non-PGM electrocatalysts.
Given that the reported catalyst has activity close to that of platinum, it is highly likely that including a cost-based metric would make this material even more impressive.

Response to C2:
We thank the reviewer for this valuable suggestion. As suggested by the reviewer, we have calculated the price activity of STRO, SRO, and commercial Pt/C catalysts for cost comparison as shown in Fig. R1 (also Fig. S12). It can be seen that the STRO catalyst shows much higher price activity than the commercial Pt/C catalyst, highlighting its cost-effectiveness. Related data (Fig. S12) and discussions have been added in the revised manuscript as follows (please see the yellow-highlighted part in Line 217-218, Page 11 and Line 219-220, Page 12 in the revised manuscript): "Furthermore, the mass activity (MA) and price activity (PA) of SRO, STRO, and Pt/C was also calculated ( Supplementary Fig. 12). It can be seen that the STRO catalyst exhibits both much higher mass activity and price activity than the commercial Pt/C catalyst, demonstrating the cost-effectiveness of STRO catalyst in practical applications.

Comment 3:
The Tafel plot of the Pt/C electrocatalyst would be interesting as a comparison in Figure 3b. Generally, given that Pt/C is the current catalyst of choice, it would worth including it in D and E as well.

Response to C3:
We thank the reviewer for this valuable suggestion. As suggested by the reviewer, the Tafel plot of the commercial Pt/C has been added in the Sci., 2019, 12, 149;Energy Environ. Sci., 2019, 12, 2569Adv. Mater., 2018, 30, 1803676;Adv. Energy Mater., 2018, 8, 1801690;Adv. Funct. Mater., 2019, 29, 1901217). Regarding to the specific activity and TOF values of Pt/C, we would like to stress that the measurement of electrochemically active surface area (ECSA, required for specific activity and TOF) of metallic Pt is completely different from that of metal oxides based on previous reports (Chem. Soc. Rev., 2019Rev., , 48, 2518Science, 2009Science, , 324, 1302Nat. Commun., 2019, 10, 149). The ECSA of metal oxides is estimated from the double-layer capacitance measurement, whereas the hydrogen underpotential deposition method is always applied for determining the ECSA of Pt materials. In view of the great difference between these two methods for calculating ECSA, so it is unfair to compare the specific activity and TOF between STRO oxide and Pt metal.  Figure 3c is unnecessary and the trend it indicates appears to be rather weak.

Response to C4:
We thank the reviewer for pointing this out. As suggested by the reviewer, we have removed the arrow in the revised manuscript.
Comment 5: The figure caption for 3g states a 'cathodic current density of -10 mA cm -2 '. A cathodic current density by definition (at least according to IUPAC) is negative, so the minus sign is redundant.

Response to C5:
We thank the reviewer for pointing this out. As suggested by the reviewer, we have removed the minus in the revised manuscript.
Comment 6: Does the BET surface area give any useful activity? The BET measurement is made on the dry catalyst, whereas the electrocatalyst ink is a mixture of the perovskite, conductive carbon, Nafion. This would presumably greatly change the accessible surface area of the electrode. As such supplementary figure 9 which is normalized to the ECSA would perhaps be of more relevance that Figure 3d.

Comment 7:
There are no mass-normalised current densities presented in the work. In addition to current densities at given overpotentials, it is useful to provide mass activity (A g −1 ) to increase comparability with other materials.

Response to C7:
We appreciate for reviewer's valuable suggestions. As suggested by the reviewer, the mass-normalized activity comparison has been added in the revised manuscript as shown in Fig. R1 (also Fig. S12). It can be seen that the STRO catalyst shows much higher mass activity than the SRO and Pt/C catalysts.
Related data (Fig. S12) and discussions have been added in the revised manuscript as follows

Response to C8:
We thank the reviewer for pointing this out. As suggested by the reviewer, we have modified this error in the revised manuscript.

Comment 9:
The downward arrow in Figure 6b doesn't add anything to the figure and makes reading the plot harder.

Response to C9:
We thank the reviewer for pointing this out. As suggested by the reviewer, we have removed the downward arrow in the revised manuscript.

Response to C10:
We thank the reviewer for this valuable suggestion. As suggested by the reviewer, the red and blue octahedra (RuO 6 and TiO 6 ) have been addressed in the figure caption in the revised manuscript.

Comment 11:
In the experimental section you should state the size/area of the working electrode to aid reproducibility.

Response to C11:
We thank the reviewer for this valuable suggestion. As suggested by the reviewer, the size of working electrode (5 mm in diameter) has been provided in the revised manuscript.

Comment 12:
A general note of the figures: The large half-filled circles used in many of the plots makes reading the data and seeing the error bars difficult. A small cross or similar would improve readability

Response to C12:
We thank the reviewer for this valuable suggestion. As suggested by the reviewer, we have changed the large half-filled circles into small crosses in the revised manuscript for improving readability.
Comment 13: Generally, this paper is well-written but there are several instances of poor/inaccurate grammar and wording which should be caught with a thorough proof-reading.

Response to C13:
We thank the reviewer for pointing this out. As suggested by the reviewer, we have carefully polished the language in the revised manuscript.

Reviewer #2
GENERAL COMMENTS: The authors investigated the performance and origin of the single-phase perovskite oxides for hydrogen evolution reaction in alkaline media demonstrating the superior catalytic activity and the stability. The authors can identify the super-exchange interactions by Ru dopant in SrTiO 3 and the concentration of oxygen vacancies using experimental and theoretical calculation results. They showed the single-phase oxides can be more active than other reported composite materials possessing the different activate sites for reaction intermediates. The reviewer has following comments and questions that might bring discussion on the optimization and extensive understanding on their behavior.

Response:
We very much appreciate the Reviewer #2's valuable comments to improve our manuscript.
According to the raised suggestions, more experiments and calculations have been made to clarify the issues and our point-by-point responses are as follows.

Comment 1:
The authors confirmed that co-existence and modification of charge state on the surface of the catalyst can provide reaction sites for both OHadsorption and water dissociation. In this regard, the activity or efficiency could be influenced by the number or density of two different sites on the surface. It would be also interesting to analyze and demonstrate that Ru dopants are equivalently occupied in the bulk region and subsurface region of the catalyst from the experimental or DFT calculations.

Response to C1:
We thank the reviewer for the valuable comments and suggestions. In experiments, we firstly performed high-solution HAADF-STEM to confirm the homogeneous distribution of Ru elements. As seen from the HAADF-STEM (Fig. R3), the Ru dopants homogeneously distribute in the surface and bulk regions. In addition, previous theoretical study suggested that Ru is thermodynamically stable on the surface and it can still be kinetically stabilized in sub-surface regions as well (Energy Environ. Sci. 2018, 11, 1762. As such, there will be a sizable and stable concentration of Ru centers within the topmost layers of the SrTiO 3 host to ensure a consistent and stable activity for STRO. Fig. R3 (also Fig. 2f). HAADF-STEM and the corresponding elemental mapping images of STRO.

Comment 2:
The authors clearly showed that Ru doping on SrTiO 3 can be advantageous in presence of super-exchange behavior between Ru and Ti. The reviewer is also wondering if there are clue and hint predicting the similar behavior of other dopants to Ru ions with Ti ions.
For example, can we expand the property of each component such as the orbital alignment of d-orbitals between Ti and Ru to HER catalytic property or adsorption behavior?

Response to C2:
We thank the reviewer for the insightful comment and question. This is a truly valid point, as it allows for the generalization of the principles that dictate the presence of Ti 3+ and M 5+ on the surface of SrTiO 3 . To check the possible universality, we have performed additional charge analysis for several other metal dopants on the surface of STO, e.g., Ir, Mo, Nb and Pt (Fig. R4). These metal ions tend to be high oxidation states in perovskite lattice. We found that the charges on Ti are 2.10, 2.04, 2.03 and 2.05 |e| for Ir, Mo, Nb and Pt dopants, respectively, which are very similar to the case (2.10 |e|) of Ru dopant, suggesting a charge transfer from the metal dopants to the surface Ti sites. This result could be a clear signal of the presence of Ti 3+ along with M 5+ surface species. While we have not carried out explicit calculations on the alkaline HER mechanism on these additional systems, which will constitute part of a broader experimental and computational screening work in the future. Fig. R4 (also Fig. S23) To further experimentally confirm the positive role of Ti 3+ /M 5+ exchange behavior in alkaline HER activity, we tried to prepare another Mo-doped SrTi 0.7 Mo 0.3 O 3-δ perovskite (denoted as STMO) and compared its activity with STO and SrMoO 4 (SMO). STMO exhibits higher catalytic activity than STO and SMO (Fig. R5), suggesting the super-exchange behavior between Mo and Ti (similar to STRO) can also boost alkaline HER. Based on above theoretical and experimental results, we believe that constructing Ti 3+ /M 5+ couples in perovskites via super-exchange effect may be a universal way for boosting alkaline HER. Related data (Fig. S23-25 (Supplementary Fig. 23)

Reviewer #3
GENERAL COMMENTS: In the manuscript, the authors have synthesized single-phase HER catalyst, SrTi 0.7 Ru 0.3 O 3-δ (STRO) perovskite oxide. Induced by a unique super-exchange interaction, the STRO perovskite exhibits outstanding HER activity with a low overpotential of 46 mV at 10 mA cm -2 and Tafel slope of 40 mV dec -1 in 1 M KOH. The super-exchange interaction in the STRO perovskite was proved and the underlying alkaline HER mechanism on STRO perovskite oxide was investigated via DFT calculations. However, some important concerns look strange or wrong, which could result from less rigorous experimental protocols.
As such the paper cannot be published in Nature Communications. Some detailed questions are listed below:

Response:
We sincerely appreciate the constructive comments raised by the Reviewer#1 to improve our manuscript. Following the suggestions, we have made more experiments and explanations to clarify the issues and our point-by-point responses are as follows.

Comment 1:
The onset overpotential of Pt/C is lower than STRO, however, as the current density reaches about 40 mA/cm 2 , the overpotential of Pt/C starts to be higher than STRO.
Why would that happen? Moreover, the Tafel slope of Pt/C should be added to compare with other samples.

Response to C1:
We thank the reviewer for raising this question. We also thank the reviewer for drawing our attention to this phenomenon that "The onset overpotential of Pt/C is lower than STRO, however, as the current density reaches about 40 mA/cm 2 , the overpotential of Pt/C starts to be higher than STRO." In fact, such phenomenon was widely reported in previous studies (Chem, 2018(Chem, , 4, 1139Nat. Commun., 2018, 9, 2452Adv. Mater., 2014, 26, 2683Adv. Funct. Mater., 2020, 30, 2000551;J. Mater. Chem. A, 2020, 8, 10831;Angew. Chem. Int. Ed., 2020, 59, 1659J. Mater. Chem. A, 2020, 8, 9239;Angew. Chem. Int. Ed., 2016, 128, 703;J. Mater. Chem. A, 2020, 8, 17202;Small, 2017, 13, 1701648), which can be can be explained by the underwater superaerophobic surface of STRO (Chem, 2018(Chem, , 4, 1139Nat. Commun., 2018, 9, 2452Adv. Mater., 2014, 26, 2683J. Mater. Chem. A, 2020, 8, 10831). Experimental result shows that the contact angle of the gas bubble on STRO in 1 M KOH solution is 151.2±2.0°, which is much larger than that (91.5±3.0°) on the Pt/C electrode (Fig. R6). Such a superaerophobic surface of STRO can significantly weaken the bubble effect that is especially prominent at large current densities, promote the mass transfer process during HER (i.e., facilitate the quick leaving of as-formed hydrogen bubbles and the electrolyte diffusion), and thereby result in a better catalytic activity at large current densities than Pt/C. Because of the fast kinetics of the HER on platinum, the current density is usually limited by the mass-transport of H 2 generated under high overpotentials. (Chem. Sci., 2019, 10, 9165;J. Phys. Chem. B, 1997, 101, 5405;J. Electrochem. Soc., 2015, 162, F190). This may explain why the activity of STRO dramatically exceeds that of Pt/C at high current densities. Besides, as suggested by the reviewer, the Tafel plot of the commercial Pt/C has been added in Fig. 3b in the revised manuscript.

Comment 3:
Since the Ti 3+ and Ru 5+ induced by super-exchange effect could promote the HER activity of STRO, then why with the amount of Ru increased to 0.4, the HER activity turned out to be inhibited?

Response to C3:
We thank the reviewer for raising this question. As demonstrated in our work, the STRO catalyst is endowed with multiple catalytic sites for affecting overall alkaline HER activity, including Ti 3+ , Ru 5+ , and oxygen vacancy, which is not only associated with Ru 5+ . In other words, the overall alkaline HER activity of STRO is dominated by multiple active sites. With the amount of Ru increased to 0.4, it means the amount of active sites Ti 3+ and oxygen vacancies will be reduced, which would have negative influence on both the water dissociation and H* adsorption steps during HER. Therefore, the HER activity turned out to be inhibited for STR0.4O. The optimal HER activity of STR(0.3)O among STRO system could be due to the perfect combination in the amount of multiple catalytic sites (e.g., Ti 3+ , Ru 5+ , and oxygen vacancy/lattice-oxygen).  2017, 121, 8378;Surf. Sci., 2015, 633, 38;J. Mater. Chem., 2011, 21, 18983).

Specifically in our case, the enhanced O (H 2 O) binding of Ti 3+ than Ru 5+ in STRO
perovskite may stem from the way the TM d-orbitals are split and occupied. The metal environments are shown in Fig. R7, which depict the octahedral arrangement with a point group symmetry of O h with the usual t 2g /e g splitting of d-orbitals as well as the surface TM sites where the point group symmetry is C 4v . For Ti 3+ which is a 3d 1 system, the electron occupies a d yz orbital which favors interaction with H 2 O, whereas, the Ru 5+ which is a 4d 3 system has its frontier states occupying a d xy orbital which does not provide an optimal overlap the p orbitals of an incoming O or H 2 O molecule (J. Phys. Chem., 1983Chem., , 87, 2960Angew. Chem. Int. Ed., 1987, 26, 846).