Inhibition of oxygen dimerization by local symmetry tuning in Li-rich layered oxides for improved stability

Li-rich layered oxide cathode materials show high capacities in lithium-ion batteries owing to the contribution of the oxygen redox reaction. However, structural accommodation of this reaction usually results in O–O dimerization, leading to oxygen release and poor electrochemical performance. In this study, we propose a new structural response mechanism inhibiting O–O dimerization for the oxygen redox reaction by tuning the local symmetry around the oxygen ions. Compared with regular Li2RuO3, the structural response of the as-prepared local-symmetry-tuned Li2RuO3 to the oxygen redox reaction involves the telescopic O–Ru–O configuration rather than O–O dimerization, which inhibits oxygen release, enabling significantly enhanced cycling stability and negligible voltage decay. This discovery of the new structural response mechanism for the oxygen redox reaction will provide a new scope for the strategy of enhancing the anionic redox stability, paving unexplored pathways toward further development of high capacity Li-rich layered oxides.


5)
indicates that R-Li2RuO3 has significantly more discharge capacity 300 mAh g-1 than does ID-Li2RuO3 at 230 mAh g-1. Does this mean without oxygen dimerization, the repondance mechanism cannot achieve high capacity in ID-Li2RuO3? Were all the benefits in rate and voltages in the cost of high capacity? Suggest to turn the inset of charge/discharge profiles into a full figure 4a and compare both cycling plots in figure 4b. They should also mention what the initial charge capacities were for both samples.
6) The quality of Figure  7) While DEMS measurements are valued techniques to detect gas generation, they cannot tell what gas generated by their own. So, they should be cautious to claim the telescopic O-Ru-O configuration to suppress oxygen release for two reasons. a) the actual effect of telescopic O-Ru-O configuration seemed suppress the dimer formation. If no oxygen dimer, there is no suppressing at all. b) the figure 6d and 6e were not exactly the same in the region of gassing. ID-Li2RuO3 showed higher increases after CO2 detected at 4.1V. The charging curves were hardly comparable between 6d and 6e. So, need to clarify the difference in the gas generation. Both 6d and e indeed showed some gas generated once charging to high voltage. Be fairly enough, the 6d only showed reduced gassing in the end of charging comparing to 6e. But this cannot rule out whether it is oxygen or not without further analysis. The figure quality should be improved as well.
8) Provide some discussions or guidance how their proposed repondance mechanism would like to work for light elements other than heavy elements like Ru in practical cathode. Will such a mechanism only work for second row transitional metals? It seems the telescopic O-TM-O configuration unlikely stable for first row TM.
Reviewer #3: Remarks to the Author: This paper reports comparative studies on ordered and disordered Li2RuO3 as oxygen-redox cathode materials, which I don't recommend for publication. The main claim 'telescopic O-Ru-O' with short and long Ru-O bonds of 1.6 and 3.0 A, which is against the classic but fundamental concept of 'ionic radius', is just a speculation based on the hypothetical calculations and (subjective, in my opinion) fittings of EXAFS. I could not find any convincing experimental evidence for their hypothesis to deny the fundamental concept of inorganic chemistry. The followings are my serious concerns.
1. Concerning the DFT part, the authors found a specific local structure, that is, short and long Ru-O bonds in disordered Li2RuO3. I'm highly suspicious of such a chemically counterintuitive short and long Ru-O bonds of 1.6 and 3.0 A, which are completely against the simple concept of ionic radius. I believe that it is necessary for the authors to re-consider the validity of the calculation models, especially for disordered one. influence the reversibility of anionic redox but which is highly relevant with our work. Therefore, for significance and implications, we have completely different initiatives when we were doing such a work.
2) Second, regarding the structural triggering mechanism, the TM-O bond shortening in Li 2 Ir 1-x Sn x O 3 was caused by antisite-vacancy introduced by TM migration during delihiation, which is undesirable since it causes capacity and voltage irreversibility. Moreover, it is also uncontrollable since TM migration is coupled with anionic redox itself. However, in our ID-Li 2 RuO 3 system, the telescopic O-TM-O configuration during charging was caused by designing special local symmetry around oxygen ions, which is highly controllable and beneficial for the reversibility of anionic redox. TM migration does not occur in our ID-Li 2 RuO 3 system, as demonstrated by XRD refinement of charged ID-Li 2 RuO 3 ( Figure S16, Table S8-9), and also supported by the results of formation energy of Ru antisite defect from DFT calculations ( Figure   S5), which is a good news for cycling stability.  Figure S19). In contrast, as for the R-Li 2 RuO 3 served as a control group, the irreversible local-range structural response with O 2 release that caused by O-O dimerization during oxygen redox processes is proved by in situ DEMS (Figure 6e), in accordance with previous studies. 5 5) Fifth, the long-range structural evolution is irreversible for Li 2 Ir 1-x Sn x O 3 in Hong's work, as indicated by XRD patterns. However, in our work, the long-range structural evolution is reversible for ID-Li 2 RuO 3 (Figure 6a-c, Figure S17), whereas the long-range structural evolution is irreversible for R-Li 2 RuO 3 ( Figure S18), as observed by XRD patterns. Combing with the longand local-range structural evolution, the ID-Li 2 RuO 3 shows excellent cycling stability.

Changes in the revised manuscript:
The detailed data and discussions have been added to the revised manuscript (Line 2-13, Page 8) and the Supplementary Materials as follows: As for the short Ru-O bonds, the crystal orbital overlap population (COOP) analysis was performed to study the interaction between Ru and O, as shown in Figure S1.

Comment 2
In Figure 2c, the authors highlight that the oxygen release energy is less favourable for the Ru1Li5 and Ru3Li3 configurations of the ID-Li2RuO3 structure than the Ru2Li4 environments in the R-Li2RuO3 structure, which leads to less O 2 loss in the former case. However, the authors do not include the oxygen release energy for the Ru2Li4 configuration of the ID-Li2RuO3 structure.
From the schematic in Figure 2b and TEM in Figure 3, these environments Ru2Li4 should also be present in the ID-Li2RuO3 structure. The authors should show whether there is a difference in the driving force for oxygen evolution between the Ru2Li4 environments in the ID-Li2RuO3 and R-Li2RuO3 structures and if there is not, explain why the ID-Li2RuO3 structure shows considerably less O2 evolution.

Response to Comment 2
Thanks for referee's kind suggestion. The oxygen release energies for the O center [Ru 2 Li 4 ] environments in the ID-Li 2-x RuO 3 systems are supplemented in Figure 2c. Generally, the oxygen release energies for the O center [Ru 2 Li 4 ] environments in ID-Li 2-x RuO 3 systems are between the two values for O center [Ru 1 Li 5 ] environments and O center [Ru 3 Li 3 ] environments in ID-Li 2-x RuO 3 . Within the same ID-Li 2-x RuO 3 system, the oxidation extent for oxygen ions in O center [Ru 2 Li 4 ] environments is between that for oxygen ions with O center [Ru 1 Li 5 ] environments and O center [Ru 3 Li 3 ] environments ( Figure S15b). The oxygen release energies at deep delithiation states (x＞1 for Li 2-x RuO 3 ) for O center [Ru 2 Li 4 ] environments in the ID-Li 2 RuO 3 (positive values) are higher than that for O center [Ru 2 Li 4 ] environments in the R-Li 2 RuO 3 (negative values), which is related to the total energy influenced by overall structural evolution of the systems. Thus, the driving force for oxygen evolution is weaker and oxygen is more stable in ID-Li 2 RuO 3 .

Changes in the revised manuscript
The oxygen release energies for the O center [Ru 2 Li 4 ] environments in the ID-Li 2 RuO 3 structure are supplemented in Figure 2c. And the corresponding discussions were given in the revised manuscript (Line 1-5, Page 9) as follows: the oxygen release energies for O center [Ru 1 Li 5 ] coordination (green dashed line), O center [Ru 2 Li 4 ] coordination (purple dashed line) and O center [Ru 3 Li 3 ] coordination (blue dashed line) in ID-Li 2 RuO 3 are all more positive than that for R-Li 2 RuO 3 after deep delithiation, which is related to the total energy influenced by overall structural evolution of the systems, indicating that the oxygen is more stable in ID-Li 2 RuO 3 .

Comment 3
The authors use a combination of X-ray diffraction and neutron diffraction to refine the crystal structures of the ID-Li 2 RuO 3 and R-Li 2 RuO 3 phases, however, they do not provide details of the fitting procedure which are necessary to properly assess the data. The occupancy of the 4e Li sites are fixed at 1. Did the authors try and refine the occupancy of Ru on the 4e site, i.e. antisite disorder? In the power X-ray diffraction pattern of ID-Li 2 RuO 3 in Figure S2, there is a broad bump between 20-30° 2θ which is characteristic of superstructure ordering of the Li1/3Ru2/3 honeycomb layers. Was this region included in the refinement of the structures?
Where the thermal parameters refined for each structure?
The structure of R-Li 2  Li and Ru may alter the long-range dimer structure and electrochemistry.

Response to Comment 3
Thanks for referee's careful check and constructive suggestions. i) First, as for the structure of Li 2 RuO 3 , the space groups of C2/c, 6-11 C2/m 12-14 and P21/m [12][13][14][15] were reported. The structure of Li 2 RuO 3 at room temperature is often described as adopting the C2/c unit cell, as was reported in most of the literature [6][7][8][9][10][11] . Owing to the significant influence of the Ru dimerization on the magnetism, heat capacity, and resistivity, its effect on the structure been more thoroughly explored by the physics community. Kakurai et al. 12 suggested that the structure adopts the space group of P21/m at low temperatures (< 540 K) and C2/m at high temperatures (> 540 K) by using the powder neutron diffraction. Wang et al. 14 had proposed that either C2/m-or P21/m-type single crystals, as well as P21/m-type polycrystalline Li 2 RuO 3 at temperatures below 300 K exist, but the synthesis conditions of C2/m-and P21/m-type single crystals described in the main text are the same. Maeno et al. 16 believed that the Rietveld refinement of neutron diffraction is equally good for the space groups C2/m and C2/c after comparing his work with Kakurai's work 12 . Indeed, due to the similarity of XRD patterns, further research is needed to confirm the space group of Li 2 RuO 3 . Considering the complexity of the space group mentioned above, the space group of Li 2 RuO 3 at room temperature still cannot be ascertained simply from temperature. All the three space group of C2/m, C2/c, and P2 1 /m are possible for R-Li 2 RuO 3 . Thus, the possibility of all the three space groups of C2/m, C2/c, and P2 1 /m are considered for both the R-Li 2 RuO 3 and ID-Li 2 RuO 3 samples here. As for the R-Li 2 RuO 3 sample, XRD patterns of the R-Li 2 RuO 3 structures copied from reference with space groups of C2/m 14 , P2 1 /m 14 , and C2/c 6 are simulated by Materials Studio software for comparison, as shown in Figure R1. The peaks at ~18.8° (corresponding to (10 1 ) peak of P2 1 /m) and ~44.6° (corresponding to (202) peak of C2/c) are absence for the R-Li 2 RuO 3 sample. R-Li 2 RuO 3 sample was fitted best to C2/m space group, although the refinements of the R-Li 2 RuO 3 are equally good for the space groups C2/m and C2/c, as is highlighted in references 12,16 . Similarly, the refinements of the ID-Li 2 RuO 3 are equally good for the space groups C2/m and C2/c. The space group of ID-Li 2 RuO 3 sample is confirmed by comparing the observed and simulated patterns of the selected area electron diffraction (SAED). Figure R2 shows the observed SAED patterns of ID-Li 2 RuO 3 sample (a, b), the simulated SAED patterns of C2/m-type ID- Li 2   ii) Second, the antisite of Ru on Li site in Li layer is considered in the revised manuscript.
The antisite of Ru in Li layer was refined to be ~ 0.02% of the total Li content in Li layer, as listed in Table S5 and Table S7 for XRD refinement and NPD refinement, respectively. Thus, the antisite of Ru was generally absent. The XRD patterns with different antisite concentrations were also simulated by using Materials studio software, as shown in Figure R3a. The XRD patterns were normalized by the intensity of (001) peak. As the antisite concentration increases, the peak A decreases while the peak B and C increases. The relative intensity of peak A and peak C ( Figure   R3b) show that the ratio of the antisite absent system fitted best with that of the observed case. Ru dimer arrangement are based on long term hexagon arrangement. 16 Considering that the space groups of P2 1 /m with Ru-Ru dimer and C2/m without Ru-Ru dimer are both possible for R-Li 2 RuO 3 at room temperature, as discussed above, the relatively pure effect of the Ru-Ru dimer on electrochemical performance can be studied by comparing C2/m-and P2 1 /m-type R-Li 2 RuO 3 samples. Fortunately, P2 1 /m type R-Li 2 RuO 3 was additionally synthesized here (XRD patterns shown in Figure R4a) to preliminary investigate the effect of Ru-Ru dimer on electrochemical performance. The main difference of synthesis condition is that the RuO 2 and RuO 2 .xH 2 O were used to prepare the C2/m-and P2 1 /m-type R-Li 2 RuO 3 , respectively. The galvanostatic charge/discharge tests of the two type of R-Li 2 RuO 3 were performed. As shown in Figure R4b, charge-discharge curves for C2/m-and P2 1 /m-type R-Li 2 RuO 3 cathodes are quite different. Both the two kinds of charge-discharge curves had been reported, the charge-discharge curves in some references 11 similar to our C2/m-type charge-discharge curve, while some other references 17 agree well with our P2 1 /m-type charge-discharge curve. However, the two kind of charge-discharge curves had not been compared in references. This difference might be related to Ru-Ru dimer formation, as the effect of Ru-Ru dimer on electrochemical performance was reported by Knight et al. 18 However, the cycling stability of both the C2/m type R-Li 2 RuO 3 sample and the P2 1 /m type R-Li 2 RuO 3 sample are poor, as shown in Figure R4c. The effect of Ru-Ru dimer on electrochemical performance might be a good topic for next work. In the present work, both the ID-Li 2 RuO 3 and R-Li 2 RuO 3 sample with C2/m space group do not contain Ru-Ru dimer, the effect of local symmetry tuned by intralayer disordering was the focus of our work.

Changes in the revised manuscript
The XRD and NPD refinements (Figure 3a-b, Figure S9, Table S4 The observed and simulated selected area electron diffraction (SAED) patterns ( Figure S8) were also given to analyze the structure on long-range scale. The ID-Li 2 RuO 3 and R-Li 2 RuO 3 structures with C2/m space group used for SAED simulation are taken from the XRD refinements. The observed SAED patterns of the as-prepared ID-Li 2 RuO 3 sample shown in Figure S8a  ( Figure S8d) zone axis, respectively. Therefore, the intralayer disordering is verified by SAED patterns on long-range scale. Neutron powder diffraction (NPD) patterns were also obtained to further analyze the structural properties of the ID-Li 2 RuO 3 sample. As shown in Figure S9, the results of NPD refinement (details are listed in Table S4, and S7) show Ru/Li-intralayer disordering, which is similar to XRD refinement. Hence, the TM/Li-intralayer disordered arrangement in the ID-Li 2 RuO 3 sample was further confirmed by NPD results.

Response to Comment 4
Cation migration is linked with anionic redox especially in the case of oxygen release. The migration is evaluated from DFT calculation and XRD refinement in the revised manuscript. The energy to form Ru antisite defects (Ru migrate to octahedral sites of Li layer) in deep delithiated ID-Li 2-x RuO 3 and R-Li 2-x RuO 3 (x=1.5, 1.75, 2) were calculated at the same defect concentration (1 out of 16 Ru ion migrated to octahedral site Li layer), as shown in Figure S5. Generally, the anti-site defects formation energy of ID-Li 2-x RuO 3 are much higher than that of R-Li 2-x RuO 3 , which means the Ru migration in ID-Li 2-x RuO 3 is much harder than that in R-Li 2-x RuO 3 , Moreover, the formation energy become negative at a low Li content (x = ~ 2.0) for R-Li 2 RuO 3 , which means Ru ion is easy to migrate, thus harm to electrochemical performance such as cycling stability.
On the other hand, the XRD refinement results of delithiated ID-Li 2-x RuO 3 is further analyzed to confirm cation migration ( Figure S16, Table S8, S9). The antisite defects of Ru in Li layer are about 0.023% and 0.025% of the total Li site in Li layer for pristine (Table S5) and (Table S9), respectively, which means the antisite of Ru in Li layer in both pristine and charged (4.8 V) ID-Li 2 RuO 3 are almost absent. Thus, the Ru migration during delithiation is negligible. The negligible Ru migration is consistent with the excellent cycling stability. Figure S5 was added to predict the TM migration from DFT calculation. Figure S16, Table   S8-S9 were given to confirm the absence of TM migration XRD refinement. Discussions were added in the revised manuscript (Line 9-13, Page 9; Line 12-22, Page 20; Line 1-2, Page 21) of page 18. Charges are as follows:

Changes in the revised manuscript
In addition, since TM migration to Li layer would be promoted by oxygen release, the energy to form antisite defects of Ru in Li layer is calculated ( Figure S5), which shows a much higher formation energy in ID-Li 2 RuO 3 than in R-Li 2 RuO 3 . Thus, the Ru migration should be much more difficult in ID-Li 2 RuO 3 than in R-Li 2 RuO 3 .
According to the refinement of XRD pattern of the 4.8 V charged ID-Li 2 RuO 3 , we find that ID-Li 2 RuO 3 kept in C2/m phase with lattice parameter changed during delithiation, as shown in Figure S16, Table S8 and S9. The β was changed from 108.5870° to 90.0097°, indicating that the layered structure was altered from O3-to O1-type C2/m phase. 12,36 As shown clearly in Figure   S17, the phase changed gradually from O3-to O1-type structure during charge process, then almost returned back to O3-type structure of the pristine during discharge process. Hence, the long-range structure of ID-Li 2 RuO 3 is reversible during charge and discharge processes. In addition, the migration of Ru to Li layer is almost absent according to the XRD refinement as the occupancies of Ru in Li layer are about 0.023% and 0.025% of the total Li site in Li layer for pristine and charged (4.8 V) ID-Li 2-x RuO 3 , respectively, which is consistent with the results of the formation energy of Ru anti-site defects ( Figure S5).

Comment 5
The properties of a range of delithated Li 2-x RuO 3 structures are calculated with DFT calculations, but the authors do not give details about how the delithated structures were generated. Were multiple Li orderings considered for each structure? The authors should provide more details about the structures were generated and ideally include the lowest energy structures in the SI.

Response to Comment 5
The multiple Li ordering of each delithated Li 2-x RuO 3 structures had been tested. The final structures are the lowest energy structures among the multiple Li ordering. The formation energies for Li removal in R-Li 2-x RuO 3 and ID-Li 2-x RuO 3 (x= 0, 0.5, 1.0, 1.5, 1.75, 2.0) are given in Figure   S4, and the lowest energy structures for R-Li 2-x RuO 3 and ID-Li 2-x RuO 3 are shown in Figure S2 and S3, respectively.

Changes in the revised manuscript
Detailed structures ( Figure S2, S3), formation energies ( Figure S4), and corresponding descriptions have been added to the revised manuscript (Line 18-20, Page 7). Changes are as follows: The final structures for R-Li 2-x RuO 3 and ID-Li 2-x RuO 3 (x = 0, 0.5, 1, 1.5, 1.75, 2) are shown in Figure S2 and S3, which were tested to be the lowest energy structures among the multiple Li ordering ( Figure S4).

Reply to Reviewer 2 General Comment
The author reported a new structural response mechanism that may inhibit O-O dimerization and then improve cycling and voltage stabilities without oxygen release. Solving oxygen release is an important but challenging task for Li-rich cathode. This is a very important piece of this work and may impact on a wider field. While there were some interesting findings and good improvements, such as suppressing O-O dimerization, the writing should be significantly improved, mainly in the quality of figures before publication.
Here are a few issues to fix.

Response to the General Comment
We thank the referee for considering that "there were some interesting findings and good improvements". And thank the referee for his/her useful comments on our work. We have revised the manuscript according to the referee's advices. The quality of figures was improved.

Comment 1
The telescopic O-TM-O configuration seems be something similar to the idea of electron hole used by Tarascon's, Bruce's and Ceder's groups. Please elaborate the conceptive novelty in the telescopic O-TM-O configuration with respect to or simply a realization of electron hole?

Response to Comment 1
We agree with the reviewer that there are two kinds of anionic redox behavior in reported

Response to Comment 2
Thank you for prompting us to improve the readability of the results. The total energies of the systems with oxygen loss at the three kinds of oxygen sites in ID-Li 2 RuO 3 and original R-Li 2 RuO 3 systems, as well as the reference energy for the calculation of oxygen release energy for delithiated ID-Li 2-x RuO 3 and R-Li 2-x RuO 3 are listed in Table S1 and Table S2, respectively. The Gibbs free energy of oxygen release was calculated as Equation S1 .
where − ( ) and ( ) of the O 2 gas phase under standard conditions were taken from previous studies. [5][6] and are defined as Equation S2 where, ( ) and ( ) indicate the calculated total energy of Li 2-x RuO 3 with and without oxygen vacancies, respectively, where ( ) was modified by the experimental formation energy of water as the binding energy of O 2 molecules from DFT calculations is overestimated. [4,5] The energy of the O 2 molecule was calculated based on Equation S3 and S4. Values used in the calculation of the total energy of the O 2 molecule were shown in Table S1.

Changes in the revised manuscript
All fonts in Figure 2 were increased. The oxygen release energy for three kinds of oxygen site in ID-Li 2 RuO 3 are supplemented in Figure 2c. Detailed definition of ΔG for oxygen release are given in Supporting information, and Table S1-S3 are add in the supporting information.  Table S1 and Table S2, respectively. The same oxygen vacancy concentration of 1/48 (2% of the total oxygen content) is considered for both ID-Li 2-x RuO 3 and R-Li 2-x RuO 3 . The Gibbs free energy of oxygen release was calculated as Equation S1 .  Table S3.

Comment 3
I do not understand "The Rietveld refinement indicates that the long-range structure of ID-Li2RuO3 is consistent with that of ideal ID-Li 2 RuO 3 ." On page 9. What is ideal ID-Li 2 RuO 3 referring to? Because they used ionic exchange method to synthesize the Ru/Li mixing layer, clarify if there is any residual Na in the sample that makes the electrochemistry different?

Response to Comment 3
The "ideal ID-Li 2 RuO 3 " structure was obtained by intralayer disordering the distribution of  Figure R5 (copy from reference 24 ). In the XRD pattern of our ID-Li 2 RuO 3 sample, no residue peaks from Na 2 RuO 3 , as shown in Figure 3a and Fiugure S7.  Thus, the intralayer disordered Li 2 RuO 3 was achieved successfully.

Comment 4
The quality of Figure 3 makes their XRD and TEM results difficult to understand at this stage.

Response to Comment 4
Thank you for your kind reminder, the quality of Figure 3 was improved in the revised manuscript.

Changes in the revised manuscript
The quality of Figure 3 is improved. showed a batter rate capacity with 145 mAh/g at 5C, while R-Li 2 RuO 3 can deliver only 93 mAh/g at 5C. Furthermore, the capacity retention for the cycle at 0.1C after the progressive charging and discharging test was 100% and 78.8% in the ID-Li 2 RuO 3 ( Figure 4e) and R-Li 2 RuO 3 (Figure 4f) systems, respectively, further confirming the excellent cycling stability of ID-Li 2 RuO 3 .
In order to compare the cycling stability of ID-Li 2 RuO 3 and R-Li 2 RuO 3 in the comparable initial discharge capacity, several voltage range had been tested for both R-Li 2 RuO 3 and ID-Li 2 RuO 3 . As is shown in Figure S11 As shown in Figure S11a, the capacity retention of ID-Li 2 RuO 3 is significantly higher than that of R-Li 2 RuO 3 in all cases, even when the initial specific discharge capacity of ID-Li 2 RuO 3 (260 mAh/g for 2.0-5.0 V) turns higher than that of R-Li 2 RuO 3 (246 mAh/g for 2.0-4.2 V).

Comment 6
The quality of Figure

Response to Comment 6
Thanks for referee's suggestions. The quality of Figure

Changes in the revised manuscript
The quality of Figure 5/6 was improved.

Response to Comment 7
In situ DEMS measurements were carried out to detect gas generation and what kind of gas is produced, similar to previous literature reports. 5,23,25,26 From the blue curve (m/z = 44) in Figure   6d and 6e, it can be seen that CO 2 release occurred once the charge voltage reached 4.1 V for both ID-Li 2 RuO 3 (5.600 mg active material) and R-Li 2 RuO 3 (4.356 mg active material), which is similar to the DEMS results in previous reports. 5,23,25,26 No oxygen was detected until the charge Indeed, the in situ DEMS results show some differences in CO 2 evolution for ID-Li 2 RuO 3 and R-Li 2 RuO 3 , but these differences are another evidence for the absence of O 2 release in ID-Li 2 RuO 3 . In the earlier stage of CO 2 evolution, CO 2 evolution of ID-Li 2 RuO 3 was higher than R-Li 2 RuO 3 , while in the later stage of CO 2 evolution, CO 2 evolution of ID-Li 2 RuO 3 was higher than R-Li 2 RuO 3 . First, as for the earlier stage of CO 2 evolution, the average voltage of ID-Li 2 RuO 3 during this stage is higher than that of R-Li 2 RuO 3 . The electrolyte decomposition rate would increase at a higher voltage during charging, thus the average gas generation rate of ID-Li 2 RuO 3 is higher than that of R-Li 2 RuO 3 . Second, as for the later stage of CO 2 evolution, although the average voltage of ID-Li 2 RuO 3 is still higher than that of R-Li 2 RuO 3 , the CO 2 evolution of ID-Li 2 RuO 3 turns lower than that of R-Li 2 RuO 3 . Since O 2 in the cell could promote the electrolyte decomposition, 25

Changes in the revised manuscript
The quality of Figure 6 was improved. Further analysis on DEMS results were added in the revised manuscript (Lines 11-15, Page 23) as follows: In addition, a sharp increase of CO 2 generation at ~ 4.3 V for R-Li 2 RuO 3 was occurred as the

Comment 8
Provide some discussions or guidance how their proposed repondance mechanism would like to work for light elements other than heavy elements like Ru in practical cathode. Will such a mechanism only work for second row transitional metals? It seems the telescopic O-TM-O configuration unlikely stable for first row TM.

Response to Comment 8
Thank you for the good advice. Theoretically

Response to the General Comment
Thanks for referee's careful and serious review. We checked the validity of the calculation models again and discussed in depth with other experts in inorganic chemistry and crystallography.

Comment 1
Concerning the DFT part, the authors found a specific local structure, that is, short and long Ru-O bonds in disordered Li2RuO3. I'm highly suspicious of such a chemically counterintuitive short and long Ru-O bonds of 1.6 and 3.0 Å, which are completely against the simple concept of ionic radius. I believe that it is necessary for the authors to re-consider the validity of the calculation models, especially for disordered one.

Response to Comment 1
Thanks for referee's careful review and kind suggestion. We checked the validity of the To start with, we want to reconcile the ionic radius problem. Table R1 listed Figure S1). The "covalent bond" effect would shorten the bonds 28 . Thus, the Ru-O bond length should be further decreased. Table R2 listed the covalent radii of O element (copy from references 30 discharging. The Ru-Ru distance of 3.05 Å in previous work 15 was also called as bond length. As shown in Figure R5, the coordination number of the O ion in O2 site is not one. Oxygen ion in O2 site is additionally bonded with another Ru (marked as Ru2 in Figure R5), with a bond length of 1.77 Å, which is close to the bond length of ~1.78 Å Ru 5+ -O 2simply based on the sum of ionic radii in Table R1. As discussed above, the bond lengths are affected by complex factors including the net charge on O n-, coordination number, and the "covalent bond" effect. 28 Therefore, the Ru-O bond lengths in our proposal are theoretically rational based on above discussions.   Figure S16, Table S8, Table S9 Figure S21) is given in Figure S23b with the detailed values listed in Table S10. Generally, the Ru-O bond length decreased during charging then increased during discharging. The total coordination number of the Ru-O bonds dramatically decreased when charged to high voltage. However, as shown in Figure 7g, the total coordination number of the Ru-O shell was not recovered to the pristine during the discharge process, indicating that the structural variation is irreversible during charge and discharge processes. This irreversible coordination number is related to O 2 release during charging, as is demonstrated by in situ DEMS measurement in Figure 6e.
As for the ID-Li 2 RuO 3 , the variation in the Ru-O shell from the fitting results ( Figure S22) is given in Figure S23b with the detailed values listed in Table S11.

Changes in the revised manuscript
COOP analysis ( Figure S1), ABF-STEM image (Figure 7a), and the EXAFS fitting of the control group R-Li 2 RuO 3 ( Figure S21, S23b, Table S10) Figure S21) is given in Figure   S23a with the detailed values listed in Table S10.        Table S4 and Table S5) and NPD ( Figure S9, Table S4, and S7) refinement results for ID-Li 2 RuO 3 both show Ru/Li-intralayer disordering.
Thus, the intralayer disorder arrangement of Ru and Li within TM layer are confirmed from both the long-range scale and local-range scale by XRD, NPD, SAED patterns, and HAADF-STEM images.

Changes in the revised manuscript
SAED patterns ( Figure S8) were supplemented. Corresponding discussions were added in revised manuscript (Line 14-22, Page 8) as follows: The observed and simulated selected area electron diffraction (SAED) patterns ( Figure S8) were also given to analyze the structure on long-range scale. The ID-Li 2 RuO 3 and R-Li 2 RuO 3 structures with C2/m space group used for SAED simulation are taken from the XRD refinements.
The observed SAED patterns of the as-prepared ID-Li 2 RuO 3 sample shown in Figure S8a

Comment 3
Please compare the charge-discharge curves of ID-Li 2 RuO 3 and R-Li 2 RuO 3 in Figure 4, rather than only focusing on the cycle stability. I believe that the comparison of dQ/dV plots would also be of interest to readers.

Response to Comment 3
Thank you for the good suggestions. The charge-discharge curves and dQ/dV plots of ID-Li 2 RuO 3 and R-Li 2 RuO 3 are compared in Figure 4a-b and Figure S10 in the revised manuscript, respectively. The charge−discharge curves of ID-Li 2 RuO 3 and R-Li 2 RuO 3 was tested by galvanostatic discharge−charge in the voltage range of 2.0-4.8 V at a current density of 30 mA/g, as shown in Figure 4a and 4b, respectively. ID-Li 2 RuO 3 delivers a specific capacity of 230 mAh g -1 in the first discharge, which is larger than the theoretical capacity of 164 mAh g -1 , estimated through the redox reaction of Ru 4+ /Ru 5+ . The extra capacity could be assigned to the contribution of the oxygen redox. The charge−discharge curves of R-Li 2 RuO 3 is consistent well with the previous work. 11,34 The charge−discharge curves of ID-Li 2 RuO 3 largely differs from that of R-Li 2 RuO 3 . Generally, there are two stages and three stages in the initial charge processes for ID-Li 2 RuO 3 and R-Li 2 RuO 3 respectively. Besides, the voltage of ID-Li 2 RuO 3 are higher than that of R-Li 2 RuO 3 in general, which can be seen clearly in dQ/dV curves ( Figure S10).

Changes in the revised manuscript
The charge-discharge curves and dQ/dV plots of ID-Li 2 RuO 3 and R-Li 2 RuO 3 were compared in Figure 4a-b and Figure S10, and corresponding discussions were given in the revised manuscript (Line 2-17) as follows: The electrochemical performance of the ID-Li 2 RuO 3 was tested by galvanostatic charge−discharge in the voltage range of 2.0-4.8 V at a current density of 30 mA/g, as shown in indicate that charge and discharge voltage platform of ID-Li 2 RuO 3 are both higher than that of R-Li 2 RuO 3 . Figure 4c compare the cycling performance of the ID-Li 2 RuO 3 and R-Li 2 RuO 3 electrodes. ID-Li 2 RuO 3 demonstrates a discharge capacity of 221 mAh/g with a capacity retention of 96% after 80 cycles, which are significantly higher than the 57 mAh/g discharge capacity and 20% capacity retention of R-Li 2 RuO 3 .    and explain why the differences occur. Then, the authors would be able to discuss the origin of the better cycle stability of ID-Li 2 RuO 3 . The present Figure 6 and relating part only report 'results'.

Response to Comment 5
The phase evolution of both ID-Li 2 RuO 3 and R-Li 2 RuO 3 are analyzed from XRD patterns.
According to the refinement of XRD pattern of the 4.8 V charged ID-Li 2 RuO 3 , we find that ID-Li 2 RuO 3 kept in C2/m phase with lattice parameter changed during delithiation, as shown in Figure S16, Table S8, and Table S9. The β was changed from 108.5870° to 90.0097°, indicating that the layered structure was altered from O3-to O1-type C2/m phase. 4,19 As shown clearly in Figure S17, the phase changed from O3-to O1-type structure gradually during charge process, then almost return back to O3-type structure of the pristine during discharge process. In addition, the migration of Ru to Li layer is almost absent according to the refinement. Hence, the long-range structure of ID-Li 2 RuO 3 is reversible during charge and discharge processes. More importantly, O 2 release was not occurred, demonstrated by in situ DEMS measurement (Figure 6d).
In contrast, the R-Li 2 RuO 3 undergoes an irreversible phase transition, as shown in Figure S18.
The XRD patterns variation of our R-Li 2 RuO 3 during charge and discharge processes are similar to the results that reported by Inaguma et al. 11 As revealed by Inaguma et al., the structure changed from C2/c (or C2/m) phase to a mixed phase of R3 _ and C2/c when charged to 3.8 V, then the structural transition with oxygen evolution occurs when further charged to 4.8 V, and the corresponding structure is unknown. 11 Similar to the reference, 11 the structure of R-Li 2 RuO 3 cannot be recovered to the pristine during discharge processes. The crystallinity is lowered after the first charge-discharge cycle. Hence, the long-range structure of R-Li 2 RuO 3 is irreversible during charging and discharging, which should be related to the oxygen evolution and TM

Changes in the revised manuscript
Comparative ex situ XRD analysis of ID-Li 2 RuO 3 ( Figure S16, S17, Table S8, S9) and R-Li 2 RuO 3 ( Figure S18) during charging/discharging was supplemented, corresponding discussions were given in the manuscript (Line 12-22, Page 20; Line 1-12, Page 21). Changes are as follows: According to the refinement of XRD pattern of the 4.8 V charged ID-Li 2 RuO 3 , we find that ID-Li 2 RuO 3 kept in C2/m phase with lattice parameter changed during delithiation, as shown in Figure S16, Table S8 and S9. The β was changed from 108.5870° to 90.0097°, indicating that the layered structure was altered from O3-to O1-type C2/m phase. 12,36 As shown clearly in Figure   S17, the phase changed gradually from O3-to O1-type structure during charge process, then almost returned back to O3-type structure of the pristine during discharge process. Hence, the long-range structure of ID-Li 2 RuO 3 is reversible during charge and discharge processes. In addition, the migration of Ru to Li layer is almost absent according to the XRD refinement as the occupancies of Ru in Li layer are about 0.023% and 0.025% of the total Li site in Li layer for pristine and charged (4.8 V) ID-Li 2-x RuO 3 , respectively, which is consistent with the results of the formation energy of Ru anti-site defects ( Figure S5). In contrast, the R-Li 2 RuO 3 undergoes an irreversible phase transition, as shown in Figure S18.     represent the first charge and discharge, respectively.