Conductive porous vanadium nitride/graphene composite as chemical anchor of polysulfides for lithium-sulfur batteries

Although the rechargeable lithium–sulfur battery is an advanced energy storage system, its practical implementation has been impeded by many issues, in particular the shuttle effect causing rapid capacity fade and low Coulombic efficiency. Herein, we report a conductive porous vanadium nitride nanoribbon/graphene composite accommodating the catholyte as the cathode of a lithium–sulfur battery. The vanadium nitride/graphene composite provides strong anchoring for polysulfides and fast polysulfide conversion. The anchoring effect of vanadium nitride is confirmed by experimental and theoretical results. Owing to the high conductivity of vanadium nitride, the composite cathode exhibits lower polarization and faster redox reaction kinetics than a reduced graphene oxide cathode, showing good rate and cycling performances. The initial capacity reaches 1,471 mAh g−1 and the capacity after 100 cycles is 1,252 mAh g−1 at 0.2 C, a loss of only 15%, offering a potential for use in high energy lithium–sulfur batteries.


Overall comments: This paper reports porous VN/graphene composite as chemical
anchor of polysulfides. The authors report that VN effectively acts as a chemical anchor for improving cell performance of Li-S batteries using polysulfide catholyte.
This would be the first report that using VN as a chemical anchor of polysulfides. The VN/G composite shows long cycling performance and relatively high rate performance. The result shown in this paper will be useful for the researchers of Li-S batteries. On the other hand, this paper includes many unsupported discussions. To be accepted in Nat. Commun., the authors should revise the manuscript based on following comments.
We thank this referee for the very positive comments and valuable suggestions.
Question 1: The material characterization seems to be not enough.  Supplementary Table 1), and also added corresponding discussion in the revised manuscript.

Question 2:
The density of electrode (VN/G composite) seems to be very low. Thus, the cell needs a large volume ratio of electrolyte. Is the porous structure shown in this paper effective form?
Response 2: The density of the VN/G composite macroform is 276 mg cm -3 measured by a balance (METTLER TOLEDO XS205) equipped with accessories for the density determination by the Archimedes principle, which is more than two times higher than that of graphene macrostructure prepared by the same freeze-drying method (<100 mg cm -3 in most cases, Energy Environ. Sci., 2015, 8, 1390and Sci. Rep. 2013, 3, 2975. Controlling the drying process is a simple but effective way to simultaneously tailor the structures and properties of graphene-based macrostructure, such as pore structure, density, mechanical strength, conductivity etc. According to the literature (Sci. Rep. 2013, 3, 2975and J. Phys. Chem. Lett. 2015, the density and porosity of graphene-based macrostructure materials can be tuned (the density as high as 1.58 g cm -3 ) by an evaporation-induced drying method of a graphene hydrogel. Therefore, the density of our VN / G composites can be adjusted according to different requirements. In addition, a large packing pressure (about 7MPa) during the assembly of a coin cell can compact the VN/G electrode, further increasing the density of the electrodes and reducing the amount of electrolyte used. Many studies 1-5 also show that graphene-based 3D macroscopic electrode exhibit a highly porous network structure and abundant electrically conducting pathways, which can be cut and pressed into pellets to be directly used as electrode without using a metal current collector, binder, and conductive additive. [5] S. Han, D. Wu, S. Li, F. Zhang, X. L. Feng, Porous graphene materials for advanced electrochemical energy storage and conversion devices, Adv. Mater. 2014, 26, 849. Question 3: The scientific novelty in this paper should be shown more clearly. e.g.
The scientific discussion about catalytic properties is not enough.
Response 3: Thank the referee for sonstructive suggestion. In order to highlight the scientific novelty of this manuscript, we have added the discussion about catalytic properties of VN as follows: According to recent reports, the presence of electrocatalyst (Pt or Ni) helps to convert the polysulfide deposits back to soluble long-chain polysulfide and hence enhances reaction kinetics and retains high Coulombic efficiency (J. Am. Chem. Soc. 137, 11542-11545 (2015)). The deposition of insulating polysulfide on electrode can impede the electron transfer at the electrode/electrolyte interface and results in an increase of internal resistance. It is known that Pt is promising but expensive as an electrocatalyst to convert short-chain to long-chain lithium polysulfides efficiently in a kinetically facile manner during charging. The catalytic properties of metal nitrides have been the subject of many experimental and theoretical investigations. In many instances the catalytic activities of VN resemble those of noble metals like Pt. Recent research shows that VN has an electrocatalytic activity similar to Pt (Sci. Rep. 5, 11351 (2015)). In our experiments, we found that the reduction peaks with the VN/G cathode (2.0 and 2.35V) appeared at higher potentials than those with the reduced graphene oxide cathode (1.88 and 2.24V) in the cyclic voltammetry profiles.
We therefore conclude that the improved polysulfide redox kinetics may be derived from the high electrical conductivity and catalytic activity of VN.
We have added the sentence "According to recent reports, Pt as an electrocatalyst can help to convert polysulfide deposits back to soluble long-chain polysulfide and hence enhance reaction kinetics and retain high Coulombic efficiency, and the catalytic activities of VN resemble those of noble metal Pt. These results suggest that VN has similar catalytic activity to that of precious metals, which can improve the redox reaction kinetics." after the sentence "The distinguishable positive shift in the reduction peaks and negative shift in the oxidation peaks of the VN/G cathode indicate the improved polysulfide redox kinetics by VN." and labeled in the manuscript. We have also added the literature (J. Am. Chem. Soc. 137, 11542-11545 (2015) and Sci. Rep. 5, 11351 (2015)) in revised manuscript as reference 34 and 35, respectively.

Response to Reviewer 2
Overall comments: It is quite impressive that the monolith of graphene/VN exhibits very strong adsorption/absorption capability towards polysulfide species, considering its relatively low specific surface area. The loading of the active mass in the cells is good, and the capacity, cycling performance as well as coulombic efficiency support the argument that VN is a strong anchoring composition in the cathode toward polysulfide. The preparation of a monolith by a hydrothermal reaction is innovative, followed by a nitrodition reaction under NH3. The article is well written, which can be accepted after a minor revision.
We thank this referee for the very positive comments and valuable suggestions. Response 1: A large number of literature [6][7][8][9][10][11][12] show that the heat treatment of graphene oxide under ammonia gas is an effective method to obtain nitrogen doped reduced graphene oxide, which is consistent with the opinion of the reviewers. During the annealing process, NH 3 reacts with certain oxygen functional groups in the as-prepared graphene oxide to form C-N bonds. Also, atomic N decomposed from NH 3 can combine with defects sites of graphene oxide, contributing to the formation of stable C-N bonding against high temperature. As suggested by the reviewer, we performed XPS measurements of reduced graphene oxide annealing under NH 3 . As shown in the figure below (also see Figure S7 in the supporting information), the C1s spectrum consists of peaks at 284.6, 285.2, 285.8, 286.7, and 289.0, attributed to the C=C, C-OH, C-N, C-O-C and C=O groups, respectively. The N 1s spectrum, ranging from 394 to 408 eV, comprises peaks corresponding to pyridine-like and pyrrolic-like nitrogen atoms. The atomic concentration of N of the N-doped reduced graphene oxide is 4.6% determined from the full-range XPS spectrum. The weight of the graphene oxide is also reduced during the ammonia treatment process due to the large loss of the oxygenated groups of the graphene oxide. By measuring the mass change of the graphene oxide samples before and after NH 3 treatment, the mass loss of graphene oxide during NH 3 annealing was about 16%.
High-resolution C1s and N1s XPS spectra of the reduced graphene oxide We have added the sentence "In addition, the VN/G composite also exhibits an electrical conductivity of ≈1150 S m -1 measured by the four-point probe method, which is over 4 times larger than that of RGO (about 240 S m -1 ), even though RGO contains doped nitrogen (about 4.6%) after NH 3 annealing ( Supplementary Fig.7).
Although N doped graphene can improve the performance of Li-S batteries, but the electrochemical performance of VN/G composite electrode was much better than that of RGO electrode in the same condition." before the sentence "As shown in Figure 5d, when the electrode was cycled at different rates of 0.2 C, 0.5 C, 1 C, 2 C and 3 C" and labeled in the manuscript.  reported Li-S cells with ether-based electrolytes, contain lithium nitrate in their electrolyte composition. In our study, we also tested the electrochemical performances of the materials in the electrolyte without LiNO 3 additive according to the comments of the reviewers. The results showed that the Coulombic efficiency of the electrode material is slightly reduced in the electrolyte without LiNO 3 , which confirms that the electrolyte additive is important for Li-S batteries. Therefore, the contribution of LiNO 3 to the performance improvement should not be ignored. However, the electrochemical performance of VN/G composite electrode was better than that of RGO electrode in the same electrolyte without LiNO 3 . These results demonstrate the advantages of VN as the host material for lithium sulfur batteries.
In order to express more accurately, we have added the sentence "The LiNO 3 additive in the electrolyte also has a positive effect on the Coulomb efficiency and cyclic performance of Li-S batteries." after the sentence "The VN/G cathode delivered an excellent initial discharge capacity of 1471 mAh g -1 with a Coulombic efficiency above 99.5%, and more importantly, it was able to maintain a stable cycling performance for 100 charge-discharge cycles at 0.2 C, indicating that dissolution of polysulfides into the organic electrolyte was effectively mitigated in the VN/G electrode." We have added the literature (J. Electrochem. Soc. 2009, 156, A694) in the revised manuscript as reference 36 and labeled in the manuscript.

Question 3: Please provide the theoretical conductivity information of VN in the revised text. What is the conductivity of the Graphene/VN composite?
Response 3: As shown in the table below (also see Supplementary Table 1), many metal nitrides have a high electrical conductivity comparable to their metal counterparts. The theoretical conductivity of VN is about 1.17×10 6 S m -1 , which is larger than that of reduced graphene oxide. The addition of VN can greatly improve the electrical conductivity of VN/G composites. Furthermore, the electrical conductivity of the VN/G composite is ≈1150 S m -1 measured by the four-point probe method, which is over 4 times larger than that of reduced graphene oxide (about 240 S m -1 ). Therefore, the electrons involved in the charging and discharging processes can transport very fast in the nested network of the VN/G cathode, which is favorable for fast polysulfide conversion. We have given the theoretical electrical conductivity of VN, and also added the sentence "In addition, the VN/G composite also exhibits an electrical conductivity of ≈1150 S m -1 measured by the four-point probe method, which is over 4 times larger than that of RGO (about 240 S m -1 )." after the sentence "The VN/G cathode has a smaller resistance (28 Ω) than that of the RGO cathode (95 Ω), which can be explained by the high electrical conductivity of metal nitrides comparable to their metal counterparts, as shown in Supplementary Table 1." in the revised manuscript. Figure 1a, the schematic for the GO should be revised to reflect the existence of oxygen containing groups.

Response 4:
According to the referee's suggestion, we revised the schematic of the graphene oxide, as shown below. In the modified schematic, we used red and cyan ball to represent oxygen and hydrogen atoms, respectively, which reflects the existence of oxygen functional groups in the graphene oxide more clearly.

The schematic of graphene oxide
We also replaced the original schematic in Figure 1 with a new schematic of graphene oxide in the revised manuscript. We thank this referee for the efforts on evaluating our work. Response 2: According to the referee's suggestion, we synthesized VO x /G composites using a process similar to that for the synthesis of VN/G composite except that the atmosphere of heat treatment was changed from ammonia to argon, and the electrochemical performances of the VO x /G cathode were tested. As can be seen from the figure below ( Figure R1), although VO x /G also has good cycling stability (the initial capacity was 951 mAh g -1 and retained 67% of the initial capacity after 100 cycles), which is due to the chemical adsorption of VO x for polysulfides, but the Coulomb efficiency (only 93%) is significantly reduced after 100 cycles. The low

Response to Reviewer 3
Coulombic efficiency probably resulted from the low converting rate of surface-bound sulfur species on VO x , which led to surface poisoning by unreacted polysulfides, prevented the subsequent adsorption, and weakened its suppression of polysulfide shuttle. The results further show that the highly conductive VN can achieve adsorption and rapid conversion of polysulfides, thus lead to the high electrochemical performance of cathode for the lithium sulfur battery.

Figure R1
(a) Cycling stability at 1C and (b) rate performance of the VO x /G cathode.
The Figure R1 was added in Supplementary Fig.9 in the revised manuscript. We have also added the sentence "In contrast, the VO x /G electrode displayed rapid capacity decay and low Coulombic efficiency (about 93% after 100 cycles), which probably resulted from the low conversion efficiency of polysulfides adsorbed on non-conductive VO x surfaces ( Supplementary Fig.9)." before the sentence "The excellent electrochemical performance of the VN/G cathode can be attributed to the following factors." in the revised manuscript.

Question 3:
The use of 70 wt% of porous graphene sponge requires a large amount of electrolyte, >30 ul/mg sulfur as the authors report. This lowers the volumetric energy density. Even still, there have been reports of using graphene based catholyte cells that apply 8.5 mg cm-2 of sulfur loading or more (Nature Commun. 2015, 6, 7760).

Good performance at higher loading have to be demonstrated in this manuscript.
Response 3: Thank the referee for constructive suggestion. The density of the VN/G composite macroform is 276 mg cm -3 , which is more than two times higher than that of graphene macrostructure prepared by the same freeze-drying method (<100 mg cm -3 in most cases, Energy Environ. Sci., 2015, 8, 1390and Sci. Rep. 2013, 3, 2975. Controlling the drying process is a simple but effective way to simultaneously tailor the structures and properties of graphene-based macrostructure, such as pore structure, density, mechanical strength, conductivity etc. Therefore, the density of our VN / G composites can be adjusted according to different requirements. In addition, a large packing pressure (about 7MPa) during the assembly of a coin cell can compact the VN/G electrode, further increasing the density of the electrodes and reducing the amount of electrolyte used. Therefore, it is possible to further increase the volumetric energy density of the battery by increasing the density of the electrode.
We agree with the reviewer's comments that the cathode with high sulfur loading is very important for the application of lithium sulfur batteries. As pointed out by the reviewers, there are significant progresses on Li-S batteries with high sulfur loading very recently. For example, we reported a 3D hybrid graphene hierarchical macrostructure as both a current collector and a host for sulfur in Li-S batteries, and this material can achieve remarkably high sulfur loading of 14.36 mg cm -2 and sulfur content of 89.4 wt% simultaneously. (Adv. Mater. 2016, 28, 1603-1609. In the article mentioned by the reviewer (Nature Commun. 2015, 6, 7760), the authors used the N,S-codoped graphene with high specific surface area as a 3D scaffold to accommodate high active material loading. In the above studies, the high specific surface area and high electrical conductivity of the host materials are necessary for achieving a high sulfur loading cathode. Compared to the N,S-codoped graphene (its specific surface area is 171.4 m 2 g -1 ) reported in the literature, our VN/G composites have a small specific surface area (37 m 2 g -1 ). The host materials with a low surface area are unable to adsorb a high amount of polysulfides, and therefore, it is very difficult to obtain a high content of sulfur. In our manuscript, we propose a new concept that uses the high conductivity and catalytic effect of metal nitrides to inhibit the shuttle effect and promote the kinetics of polysulfide conversion reactions. We believe that if we can prepare metal nitrides with a high specific surface or their composite materials, we can achieve high sulfur loading electrodes and obtain excellent electrochemical performance.
In addition, we also note that the sulfur loading of some recently reported cathodes containing an inorganic polar host material is relatively low, but they proposed a new method to suppress the shuttle effect and improve the performance of sulfur cathode, and opened a new direction to fabricate high-performance advanced Li-S batteries.  Response 4: According to the referee's suggestion, we evaluated the specific surface area of the RGO by nitrogen adsorption-desorption method. As shown in the figure below, the RGO has a high specific surface area of 296 m 2 /g and a hierarchical pore structure.

Figure R2
Nitrogen adsorption-desorption isotherm of the RGO. Inset: the pore size distribution obtained using the BJH method.
We included the nitrogen adsorption-desorption isotherm of the RGO ( Figure R2) as Supplementary Fig. 4, and have added the sentence "In contrast, the specific surface area of the RGO was as high as 296 m 2 g -1 (Supplementary Fig. 4)." after the sentence "The specific surface area of the VN/G was 37 m 2 g -1 with mesopores 18 nm in diameter ( Supplementary Fig. 3), which is consistent with the TEM observation." and labeled in the revised manuscript.
Question 5: In Figure S5, what is the condition of the cell for EIS? Pristine or in discharged/ charged states? This is important for claiming the affinity between VN and polysulfides.
Response 5: We thank the reviewer very much for the valuable suggestion. Our electrochemical impedance spectra are derived from pristine cells before cycling and recorded from 10 kHz to 100 MHz at open circuit voltage at room temperature. In order to describe more accurately, we modified the original figure caption of the Figure S5 to "Comparison of the electrochemical impedance spectra of the VN/G and RGO cathodes before cycling. The data was recorded from 10 kHz to 100 MHz at open circuit voltage at room temperature." in the revised manuscript.
Question 6: In Figure 4d, the rate capability of rGO cell has to be compared directly.
Response 3: For comparison, we have added the rate performance of the RGO electrode. As shown below, the RGO electrode shows a lower discharge capacity than the VN/G cathode. This result further indicates that the VNG electrode has better electrochemical performance.

Figure R3
Rate performance of the RGO cathode at different current densities.
We have revised the Figure 4d, and also added the sentence "In contrast, the RGO electrode exhibited lower discharge capacity and poorer stability under the same conditions." after the sentence "As shown in Figure 4d, when the electrode was cycled at different rates of 0.2 C, 0.5 C, 1 C, 2 C and 3 C, the cell was able to deliver discharge capacities of 1447, 1241, 1131, 953, 701 mAh g -1 , respectively." in the revised manuscript.
Question 7: In Figure 5, the binding energy of VN with Li2S6 has to be compared with N-doped graphene, instead of plain graphene, in order to correlate with the experiments.

Response 7:
We agree with the referee that the binding energy of VN with Li 2 S 6 should be compared with N-doped graphene, instead of pristine graphene, for a more precise correlation with our experiments. Actually, the binding energy between Li 2 S 6 and N-doped graphene ranges from 0.7 to 2.85 eV for different N-doping configurations, as reported in our recent theoretical results (L. C. Yin et al.，Nano Energy, 2016, 25, 203-210). Considering that the pyridinic-N is the dominant dopant in N-doped graphene synthesized in this work, as shown in the Supplementary Figure   6, we compare the binding energy between VN and Li 2 S 6 with that between Li 2 S 6 and pyridinic-N-doped graphene in the revised manuscript.
In order to make this point clearly, we changed the statements "For comparison, the binding energy between Li 2 S 6 and graphene was also considered, and was calculated to be 0.74 eV. In contrast, the binding energy between Li 2 S 6 and VN was calculated to be much larger (3.75 eV). This is mainly due to the much stronger polar-polar interaction between Li 2 S 6 and VN than the polar-nonpolar interaction between Li 2 S 6 and graphene. In comparison with the case of Li 2 S 6 on graphene (Fig. 5b), the strong polar-polar interaction between Li 2 S 6 and VN results in an obvious deformation of the Li 2 S 6 molecule (Fig. 5c), forming three S-V and one Li-N bonds." to "As shown in the Supplementary Figure 7, the pyridinic-N is the dominant dopant in N-doped graphene synthesized in this work. For comparison, the binding energy between Li 2 S 6 and pyridinic N-doped graphene was considered, and it has been reported to be 1.07 eV 38 . In contrast, the binding energy between Li 2 S 6 and VN was calculated to be much larger (3.75 eV). This is mainly due to the much stronger polar-polar interactions between Li 2 S 6 and VN than those between Li 2 S 6 and pyridinic N-doped graphene. In comparison with the case of Li 2 S 6 on pyridinic N-doped graphene (Fig.   5b), the strong polar-polar interaction between Li 2 S 6 and VN results in an obvious deformation of the Li 2 S 6 molecule ( Fig. 5c), forming three S-V and one Li-N bonds" in the revised manuscript. We also modified the Figure  Response 8: According to the referee's suggestion, we added the calculated bond lengths of V-S and Li-N for comparison in the revised manuscript. We modified the sentence "The bond lengths of these S-V and Li-N bonds are very close to those the corresponding bonds in bulk VS and LiNH 2 (2.42 Å and 2.06 Å)" to "The bond lengths of these S-V (2.49-2.61 Å) and Li-N (2.08 Å) bonds are very close to the corresponding bond lengths in bulk VS (2.42 Å) and LiNH 2 (2.06 Å)" in the revised manuscript.