Manipulating the diffusion energy barrier at the lithium metal electrolyte interface for dendrite-free long-life batteries

Constructing an artificial solid electrolyte interphase (SEI) on lithium metal electrodes is a promising approach to address the rampant growth of dangerous lithium morphologies (dendritic and dead Li0) and low Coulombic efficiency that plague development of lithium metal batteries, but how Li+ transport behavior in the SEI is coupled with mechanical properties remains unknown. We demonstrate here a facile and scalable solution-processed approach to form a Li3N-rich SEI with a phase-pure crystalline structure that minimizes the diffusion energy barrier of Li+ across the SEI. Compared with a polycrystalline Li3N SEI obtained from conventional practice, the phase-pure/single crystalline Li3N-rich SEI constitutes an interphase of high mechanical strength and low Li+ diffusion barrier. We elucidate the correlation among Li+ transference number, diffusion behavior, concentration gradient, and the stability of the lithium metal electrode by integrating phase field simulations with experiments. We demonstrate improved reversibility and charge/discharge cycling behaviors for both symmetric cells and full lithium-metal batteries constructed with this Li3N-rich SEI. These studies may cast new insight into the design and engineering of an ideal artificial SEI for stable and high-performance lithium metal batteries.

Reviewer #1 (Remarks to the Author): In this work, the authors designed a Li3N-rich SEI on Li metal anode for rechargeable Li metal batteries by an immersion method.The properties of the artificial SEI and the modified Li metal anode were investigated.The performance of Li metal anode was improved under the protection of the artificial SEI.Large revision is necessary to meet the high level of the journal.The following are some tips.
(1) Is LiTFSI-DOL/DME ether based liquid electrolyte suitable for cells when the voltage is above 4 V? (2) Fig. 5a and c should be enlarged to show the phenomenon of "short circuit".
(3) Many details are missing, such as the mass loading of LFP cathode, the electrolyte used for NMC-Li coin cell.Please enrich these information.(4) Many careless mistakes should be avoided, such as "f Li3N", "O2", "LiNO3".
(5) Some sentences should be polished to improve the fluency.
Reviewer #2 (Remarks to the Author): Pokharel et al reported a Li3N-rich SEI layer derived from the TEMED on the metallic Li surface via the solution-based process.The Li electrode employed with the generated phase-pure single crystalline Li3N displayed lower Li+ diffusion barrier and improved electrochemical performances, as revealed by simulations and experiments.Although the method of manipulating the diffusion barriers and kinetics of Li+ is very promising and interesting, there exist some critical problems and some contradictory data in the manuscript.I think the work can be published after addressing the following comments carefully.Comments: 1. it is not persuasive to conclude that the formed Li3N is phase-pure and single crystalline through the weak even disappeared XRD signals.How does the author rule out the nanosize effect of the Li3N?More specific and detailed characterizations should be provided to demonstrate the presence and formation of single crystalline.Meanwhile, the Li2O and LiNxOy are obviously observed according to the XPS results, and therefore the enhanced performances might not be simply attributed to the presence of phase-pure Li3N. 2. The statement that the single crystalline Li3N can decrease Li diffusion barrier in comparison to the polycrystalline one needs to be further verified.A series of comparisons from phase field simulations to electrochemical measurements are suggested to conduct before drawing this conclusion.3.In the cells, the Li diffusion along which direction will be accelerated?Vertical?If so, how can the Li dendrite growth be prevented and what is the mechanism for the improved spatial homogeneity?The authors are recommended to consider more about the lateral diffusion that benefits the uniform Li ion flux spreading and distribution.4.During the immersion process, organic moieties derived from TEMED are expected to be introduced and left on the Li surface, as no further cleaning procedure is performed.The TOF-SIMS and in-depth XPS measurements are suggested to carry out to rule out the presence of organic moieties.Or, maybe the organic residues have a positive effect on the lithium plating. 5.For practical application, the areal capacities of electrodes are usually above 3.0 mAh cm-2.Therefore, the stripping/plating capacity of 1.0 mAh cm-2 is far behind that for practical application and a capacity above 3.0 mAh cm-2 should be performed.The pouch cells based on large areal loading cathodes are also suggested to strengthen the argument.6.The Coulombic efficiency tests of the Li electrodes should be given under low and high stripping/plating capacity.Also, A much thinner Li chip of 50 um thickness is recommended to adopt in Li-Li cell and Li-Cu cells to demonstrate the superiorities.7.In figures 5a, 5c and S6-S9, it is very confusing that the TEMED treated Li electrodes possess the much higher overpotentials than the pristine Li electrode in the initial tens of cycles while it is opposite in Figure S14.In Figure 6a, the bare Li one shows better capacities that the treated one in the first 50 cycles while it is lower than TEMED treated one in rate performance (Figure 6b), why? 8. What is the R2 value in Fig S2 ?The error of fitting line is too large to accept.Please reperform the measurement.There are also many confusing descriptions lacking of details, such as the electrolyte selection (with LiNO3?Without LiNO3?); the electrode fabrication of LFP,NCM in experimental section, and so on.What is purpose of using LTO in the manuscript?What is the immersion time used for treated Li electrode in successive tests?9. Some recently published works on decreasing Li diffusion barriers by alloys or SACs in the lithium metal battery are suggested to be discussed in the introduction parts.
Reviewer #3 (Remarks to the Author): In this manuscript, phase-pure/single crystalline Li3N-rich SEI was constructed via a convenient TEMED dipping treatment.The Li3N-rich SEI exhibited low diffusion energy barrier, high Li+ transference number, and outstanding mechanical to achieve dendrite-free Li plating/stripping.The TEMED treated symmetrical cell shows outstanding plating/stripping cycles with reduced overpotential and the full cell exhibits remarkably improved cycling stability at high rates compared to bare Li.The results are interesting.I would recommend it for further consideration.Some suggestions for the authors are below.
(1)For the EIS analysis of symmetric cells treated with TEMED, the impedances of charge transfer in the Figure 2f (~400 Ohm) and Figure 2d (~200 Ohm) differ significantly.Please explain the reasons for this difference.
(2)Figure 1 show that the surface morphology of the top artificial SEI appears rough and mossy accompanied with some cracks, which is ascribed to the formation of Li3N-rich layer.Normally, the formation of cracks in SEI is not conducive to uniform the distribution of Li ions flux and the deposition of lithium.
(3)Did the content and crystal phase of Li3N in SEI change during cycling?(4)Please give a comparison of the performance of the Li3N-rich SEI with other artificial layers, especially Li3N protective layers formed by other methods.(5)The O 1s XPS peak attributed to Li2O is at ~531.5 eV in TEMED treaded Li, while for bare Li sample, the value is ~529.5 eV.The authors do not explain the reasons for such a big difference in binding energy.I strongly suggest the authors to adjust the X-axis of the two samples and reanalyze the high-resolution deconvoluted peaks.
Thank you for your consideration of our manuscript entitled "Manipulating Diffusion Energy Barrier at the Interface for Dendrite-free and Long-life Lithium Metal Batteries" (NCOMMS-22-30813).We are very grateful to the three reviewers; the comments expressed by the reviewers have been carefully considered, and all necessary additions/revisions have been made based on their suggestions or criticisms in the revised manuscript.
The following are our point-by-point responses to the reviewers' comments.

Reviewer(s)' Comments to Authors:
Reviewer #1: In this work, the authors designed a Li 3 N-rich SEI on Li metal anode for rechargeable Li metal batteries by an immersion method.The properties of the artificial SEI and the modified Li metal anode were investigated.The performance of Li metal anode was improved under the protection of the artificial SEI.Large revision is necessary to meet the high level of the journal.The following are some tips.

Authors' Response:
We appreciate the valuable and insightful comments on our study.The manuscript has been extensively revised and significantly improved by including additional experimental results and discussions in response to the reviewer's comments.
(1) Is LiTFSI-DOL/DME ether based liquid electrolyte suitable for cells when the voltage is above 4 V?

Authors' Response:
We would like to appreciate the reviewer's valuable comment.In practical application the electrolyte performance depends on several factors, including the specific electrode materials being used, the operating conditions of the cell as temperature and current density, and the desired performance characteristics of the cell such as energy density and cycle life.However, in general LiTFSI/DOL-DME ether-based liquid electrolyte is not suitable for operation above 4 V due to the potential for ether decomposition or instability of the electrolyte.To address this, the addition of lithium nitrate (LiNO3), fluroethylene carbonate (FEC) improves the stability of the LiTFSI/DOL-DME electrolyte at high voltages by reducing the formation of gas and the degradation of the electrode/electrolyte interface.
(2) Fig. 5a and c should be enlarged to show the phenomenon of "short circuit".

Authors' Response:
We would like to thank the reviewer for carefully reviewing our manuscript and proving the valuable suggestion.We have inserted the inset of the respective "short circuit" phenomenon in the related figures as R1 and R2.We have updated the figures in our manuscript accordingly.
We appreciate the reviewer for pointing out the missing content in our manuscript.We have made the necessary corrections and have provided the information of the mass loading for LFP !A %#$ 61 -6 B2 ) cathode and the electrolyte that had been used for the NMC (LiTFSI in DOL/DME with 1% LiNO3) based full cell.We have enriched that information in the revised manuscript.

Authors' Response:
We appreciate the reviewer for pointing out the typos for the elements.We have made all the necessary corrections in the revised manuscript.
(5) Some sentences should be polished to improve the fluency.

Authors' Response:
We would like to thank the reviewer for pointing out the fluency of the sentences, we have polished all sentences carefully and mistakes have been corrected accordingly.Various literature have reported that the metallic Li reacts readily with the trace amount of residual gases, resulting in the formation of surface contamination layers primarily composed of Li2CO3, Li2O and LiOH.According to literature, XPS spectra obtained without continuous sputtering exhibits notably significantly higher levels of surface contamination.Listed below are some papers with related discussion included.Even with continuous sputter-cleaning, minor presence of Li2O is still detected on the surface.It is noted that the whole synthesis and cell assembly were performed in a glovebox with <1ppm O 2 and H 2 O.And all experiments were repeated many times.
We do not think the superior performance of treated samples is due to the negligible presence of Li2O and LiNxOy.. We have updated the figure and revised the expressions accordingly in the revised manuscript.It should be emphasized that the simulation models are also based on this crystal structure and the calculated results are consistent with our experimental data, as shown in the revised manuscript.
We further performed the integrated phase field simulations to investigate the activation energy of Li + ion and its transport behavior at the interface on bare Li anodes, N2 treated Li and TEMED treated Li anodes with Li3N as an artificial solid-electrolyte interface (SEI) layer (Fig 6 a-c).The observations showed that on bare lithium, Li dendrites grew into filament-like structures with side branches, and on N2 treated lithium, dendrite growth was slower and the side growth was minimal.In contrast, on the TEMED treated lithium with the Li3N layer, the initial protrusion exhibited a dome-like morphology with a smooth electrode-electrolyte interface, and its growth rate was significantly reduced.This indicated that the artificial SEI layer with higher Li-ion diffusivity and transference number could suppress Li dendrite growth from the anode.
We further analyzed the Li-ion concentration across the dendrite tip and were able to find that Li-ion concentration increased sharply for bare Li anodes, while it increased gradually for treated Li anodes, supporting the notion that higher Li-ion concentration gradients at the dendrite tip facilitate dendrite growth (Fig 6 d-f).Further analyzing the electric field variation in the different Li anodes, electric field was maximum for bare Li at the tip as compared to N 2 treated and TEMED treated Li anodes.This was attributed to the sharper tip morphology and larger curvature, which led to a higher Li-ion concentration gradient and facilitated dendrite growth.In contrast, treated Li anodes exhibited less significant variations in the electric field at different time steps, indicating inhibition of Li dendrite growth.
These results suggest that the Li3N protective layer as an artificial SEI layer can effectively suppress Li dendrite growth by enhancing Li-ion diffusivity and reducing Li-ion concentration gradients at the dendrite tip, thereby inhibiting the self-accelerating process of dendrite growth on bare Li anodes.
3. In the cells, the Li diffusion along which direction will be accelerated?Vertical?If so, how can the Li dendrite growth be prevented and what is the mechanism for the improved spatial homogeneity?The authors are recommended to consider more about the lateral diffusion that benefits the uniform Li ion flux spreading and distribution.

Authors' Response:
We appreciate the reviewers' valuable comments.We agree with the reviewer that the diffusion directions are important for the uniform Li ion flux.In response to the reviewer's valuable comment, we have performed lateral and vertical diffusion calculations by using the density functional theory (DFT).
In the calculation models, we considered all possible migration pathways including lateral (or inplane; !c axis) and vertical (or out-of-plane; || c axis) diffusions (Fig. R6).For the lateral diffusion, a Li can diffuse along the Li(2)-N plane (path i) or the Li(1) plane (path ii), where the diffusion via path i shows a much lower energy barrier (0.01 eV vs. 1.0 eV).For the vertical diffusion, a Li can diffuse between the Li(2)-N plane directly (paths iii and iv) or passing through the Li(1) plane (path v).Paths iii and iv show ~0.6 eV of energy barrier, whereas path v shows 1.8 eV.Because of the lowest barrier, the lateral diffusion via path i is most likely dominant for Li diffusion in a-Li3N.Also, the results indicate that the Li diffusion passing through the Li (1) plane has a high barrier for diffusion, and thus Li in the Li(2)-N plane is responsible for the most diffusion.Hence, the lateral diffusion dominates the Li ion diffusion at the interface, which can largely benefit the uniform deposition of Li metal.During the immersion process, organic moieties derived from TEMED are expected to be introduced and left on the Li surface, as no further cleaning procedure is performed.The TOF-SIMS and in-depth XPS measurements are suggested to carry out to rule out the presence of organic moieties.Or, maybe the organic residues have a positive effect on the lithium plating.

Authors' Response:
We appreciate the reviewer's suggestions to improve the novelty and importance of our work.In response to the reviewer's valuable comment, the in-depth XPS measurement was performed to 7. In figures 5a, 5c and S6-S9, it is very confusing that the TEMED treated Li electrodes possess the much higher overpotentials than the pristine Li electrode in the initial tens of cycles while it is opposite in Figure S14.In Figure 6a, the bare Li one shows better capacities that the treated one in the first 50 cycles while it is lower than TEMED treated one in rate performance (Figure 6b), why?
Authors' Response: We appreciate the reviewer for pointing out the discrepancy between the figures.We would like to clarify that there is a mistake while naming the x-axis in Fig. S14.We have made the correction in the revised manuscript.The x-axis should be capacity rather than time as it shows the nucleation potential.For nucleation overpotential to perform the measurement we waited for  We appreciate reviewer for suggesting repeating the experiment as the fitting line was too large.
We would like to add that we performed the related experiment and have updated all the related experiments accordingly based on the activation energy obtained from the Arrhenius plot.We repeated the phase field simulation after repeating the experiment for the activation energy and have updated the revised manuscript.We apologize for the confusion regarding the use of electrolyte.We have used 1% LiNO3 in the LiTFSI-DOL/DME based electrolyte.We have added the details regarding the same in the updated manuscript as well.
The details regarding the cathode fabrications and the optimum emersion time regarding the TEMED treatment process have been enriched with more details in the updated manuscript.
We thank the reviewer for pointing out the presence of LTO in the manuscript.We have made corrections as it should have been NMC instead of the LTO. 9. Some recently published works on decreasing Li diffusion barriers by alloys or SACs in the lithium metal battery are suggested to be discussed in the introduction parts.

Authors' Response:
Considering the suggestion of the reviewer, we incorporated the recent publications on the decreasing Li diffusion barrier by alloys or SACs in the introduction section.Reviewer #3 (Remarks to the Author): In this manuscript, phase-pure/single crystalline Li3N-rich SEI was constructed via a convenient TEMED dipping treatment.The Li3N-rich SEI exhibited low diffusion energy barrier, high Li+ transference number, and outstanding mechanical to achieve dendrite-free Li plating/stripping.
The TEMED treated symmetrical cell shows outstanding plating/stripping cycles with reduced overpotential and the full cell exhibits remarkably improved cycling stability at high rates compared to bare Li.The results are interesting.I would recommend it for further consideration.
Some suggestions for the authors are below.
Response to the reviewer: We greatly thank the valuable comments from the review.The manuscript has been revised extensively based on the suggestions.
(1)For the EIS analysis of symmetric cells treated with TEMED, the impedances of charge transfer in the Fig. 2f (~400 Ohm) and Fig. 2d (~200 Ohm) differ significantly.Please explain the reasons for this difference.

Authors' Response:
We appreciate reviewer for carefully reading and raising this good point.
The EIS in Fig. 2d is the initial EIS for the bare lithium and TEMED treated Li, where the charge transfer resistance is ~400 ohms and 200 ohms respectively.(2) Fig. 1 show that the surface morphology of the top artificial SEI appears rough and mossy accompanied with some cracks, which is ascribed to the formation of Li3N-rich layer.Normally, the formation of cracks in SEI is not conducive to uniform the distribution of Li ions flux and the deposition of lithium.

Authors' Response:
We appreciate the reviewer's valuable and insightful comment on our study.
We agree that the formation of SEI layer is a complex electrochemical process, a rough and mossy SEI layer may have a non-uniform distribution of Li ion flux due to the presence of surface irregularities and variations in the thickness of the layer.
To examine the artificial SEI layer, we performed various experiments to compare TEMED treated In addition, the Arrhenius-plot showed that the activation energy of 0.723 eV for bare Li, N2 treated Li of 0.613 eV whereas TEMED treated Li showed an activation energy of 0.48 eV.This decrease in activation energy leads to a much higher mobility of Li + , which in turn decreases the concentration gradient to provide a more uniform surface for Li + migration and plating.
The SEM image after multiple cycles also supports the results from other measurements.The (3) Did the content and crystal phase of Li3N in SEI change during cycling?
The use of a Li3N-rich layer as a protective coating has shown promise as a solid-state ionic conductor for lithium metal batteries, primarily due to its high ionic conductivity at room temperature and high Young's modulus.Various techniques have been employed to analyze the effectiveness of the Li3N -rich layer as a protective coating.These include direct exposure to nitrogen gas, plasma treatment of the Li metal surface, gas reactions between liquid Li and N2 gas, and direct pressing of Li3N onto Li metal.These results have highlighted the potential of Li3N as an artificial SEI material for lithium metal batteries.The application of Li3N coatings has demonstrated their ability to reduce dendrite growth and enhance battery stability.However, concerns remain regarding the long-term stability and the achievement of higher capacity performance.
In our work, the TEMED treated Li based SEI have shown exceptional results at higher capacity and longer cycle life.We demonstrated long stable plating/stripping cycling up-to 3500 hours at 0.5 mA cm -2 and capacity 1mAh cm -2 , ~600 hours have been achieved for the current density of 5 mA cm -2 at the capacity of 3mAh cm -2 for 50um thick Li.Similarly full cell showed exceptional performance up-to 500 cycles at 1C rate.Comparative analysis has been shown in the table below.
Comparative performance analysis of the different Li3N-rich artificial layers as protective layers has also been incorporated in the manuscript.

Reviewer # 2 :
(Remarks to the Author): Pokharel et al reported a Li3N-rich SEI layer derived from the TEMED on the metallic Li surface via the solution-based process.The Li electrode employed with the generated phase-pure single crystalline Li3N displayed lower Li + diffusion barrier and improved electrochemical performances, as revealed by simulations and experiments.Although the method of manipulating the diffusion barriers and kinetics of Li + is very promising and interesting, there exist some critical problems and some contradictory data in the manuscript.I think the work can be published after addressing the following comments carefully.To clarify the controversial part of the Li2O and LiNxOy in our XPS data, an in-depth XPS measurement has been performed.In Fig R4, the XPS data of the N1s spectra with the etching time of 30 minutes show the formation of the Li3N, whereas the Li2O can be observed in Li 1s spectra.The presence of the oxides in Li 1s and O 1s in the in-depth XPS results suggest that the oxides are formed due to the exposure of the sample during the analysis process.

Fig R15 :
Fig R15: Coulombic efficiency of Li-Cu cell at the current density of 1mA cm -2 with the capacity of 3 mAh cm -2 with 50µm lithium chip.
Half cell TEMED Li-Cu Half cell the interface to be stable while for the symmetrical cell testing the cell were directly placed for measurement.In Fig.5a, 5c and S6-S9, we tested the cells upon assembly.The higher overpotentials at beginning can be attributed to the stabilization at the interface between the electrode and the electrolyte.After several cycles, the overpotentials drop quicky and lower compared to bare lithium.Hence, the difference in the voltage of the figures can be due to the different testing scenarios.We would like to clarify the confusion regarding fig.6(a) and fig.6(b).Fig. 6(a) represents the long term cycling performance of both the bare Li and TEMED treated Li at the C-rate of 1C.While Fig. 6(b) shows the different specific capacity of the Li and TEMED treated cell at the various C-rate.For Fig.6a, the lower capacity of the cell with TEMED treated lithium compared to the cell with bare lithium at the beginning can be attributed to the initial stabilization of the interfacial layer, which is similar to the overpotential discussion above related to Fig.5a.For the rate performance in Fig.6b, we initially performed the formation cycles at a very low current density to stabilize the SEI layer before carrying out tests with different rates.Hence, the capacity of the cell with TEMED treated lithium metal is higher from the beginning compared to the cell with bare lithium since the interfacial layer stabilization has been completed.

Fig R16 :
Fig R16: Correlation between ln D and reciprocal temperature (Arrhenius-plot) for TEMED treated Li and Bare-Li.The ionic conductivity of Li3N can be calculated using the equation & = 2 #/$% , in which, L is the thickness of Li 3 N, R is the resistance of Li 3 N and a is the area.

Fig. 2f represents
Fig.2frepresents the EIS before and after the polarization for the calculation of the transference number.We performed charge and discharge cycles at 0.01 mA cm -2 , with 4 hour charge, 30 minute rest and 4 hour discharge, with the process repeated 6 times.The cell was then polarized at 10 mV for 10 hours to ensure a steady state (Fig.2 e,f).EIS spectra before polarization and after the steady-state had been reached is shown in inset of Fig. 2 e,f.Hence, the testing conditions between Fig. 2f and Fig. 2d are quite different, showing a different charge transfer resistance.
Li and bare Li.AFM measurements showed the average RMS roughness of 242 and 157 nm, for Bare Li and TEMED respectively.The higher RMS value for bare Li implies uneven and rough surfaces that can create large protuberances on the electrode surface3 .These protuberances generate non-uniform electric fields during charge/discharge leading to inhomogeneous plating of Li.In contrast, the smooth surface of TEMED treated Li electrode based on the AFM tests provides a route for uniform Li plating.Similarly the an average Young's modulus values of 0.32 and 6.85 GPa, for Bare Li and TEMED treated Li indicating that the TEMED treated Li electrode can withstand mechanical forces, providing the desired mechanical stability during Li plating/stripping, as well as offering high resistance and sufficient strength to suppress Li dendrite growth4 .
unregulated and unprotected surface of the bare Li creates large protuberances generating a nonuniform electric field, leading to inhomogeneous plating of Li.Whereas the TEMED treated Li leads to a dense and nodule-like morphology with no lithium dendrites observed.This compact and nodular artificial layer serves as a physical protection barrier to inhibit penetration of organic electrolyte which could subsequently corrode the underlying Li electrode.Hence an artificial SEI achieved through TEMED-treated Li offers lower impedance with a lower energy barrier for Li + ion migration, further benefiting ion transport at the interface between electrode and electrolyte.This effectively facilitates transport of Li + ions across the electrode surface, leading to high transference number and excellent mechanical strength that tolerates volume change to enforce more uniform Li + ion flux.TEMED treated symmetrical cell also showed an outstanding plating/ stripping cycles with reduced overpotential and the full cell exhibits remarkably improved cycling stability and capacity retention as well as capacity utilization at high rates compared to untreated bare Li.

Table 1 .
Comparative Li3N-rich layer as an artificial protective layers.