Ion association tailoring SEI composition for Li metal anode protection

Ion association tailoring SEI composition for Li metal anode protection

Journal of Energy Chemistry 45 (2020) 1–6 Contents lists available at ScienceDirect Journal of Energy Chemistry journal homepage:

2MB Sizes 0 Downloads 3 Views

Journal of Energy Chemistry 45 (2020) 1–6

Contents lists available at ScienceDirect

Journal of Energy Chemistry journal homepage:

Ion association tailoring SEI composition for Li metal anode protection Yitao He a, Yaohui Zhang a,∗, Peng Yu b, Fei Ding c,∗, Xifei Li d,e,∗, Zhihong Wang a, Zhe Lv a, Xianjie Wang a, Zhiguo Liu a, Xiqiang Huang a a

School of Physics, Harbin Institute of Technology, Harbin 150001, Heilongjiang, China Beijing Aerospace Technology Institute, The Third Academy of CASC, Beijing 100074, China c Science and Technology on Power Sources Laboratory, Tianjin Institute of Power Sources, Tianjin 300384, China d Institute of Advanced Electrochemical Energy, Xi’an University of Technology, Xi’an 710048, Shaanxi, China e Center for International Cooperation on Designer Low-carbon & Environmental Materials (CDLCEM), Zhengzhou University, Zhengzhou 450001, Henan, China b

a r t i c l e

i n f o

Article history: Received 5 September 2019 Revised 30 September 2019 Accepted 30 September 2019 Available online 3 October 2019 Keywords: Ion association Lithium metal Electrolyte additive

a b s t r a c t Electrolyte additives play an important role in suppressing lithium dendrites through tailoring the composition/property of the SEI, however lacking of additives can achieve high performances both in ether and carbonate electrolytes hinders further enhancement of the high voltage lithium-metal batteries. Here, lithium perchlorate (LiClO4 ) has been presented as an excellent additive to meet the above requirements. An optimized chemical composition of SEI can be achieved through the formation of ionic association. Our results indicate that the LiClO4 behaves like a catalyst, which promotes LiTFSI to form a better SEI to inhibit further reaction. Superior coulombic efficiencies and cycling performances were obtained both in ether and carbonate electrolytes. This study paves a new pathway for designing bi-soluble additives for safe lithium metal batteries. © 2019 Science Press and Dalian Institute of Chemical Physics, Chinese Academy of Sciences. Published by Elsevier B.V. and Science Press. All rights reserved.

1. Introduction Lithium metal is considered as a desirable anode for the nextgeneration rechargeable batteries due to its high theoretical capacity (3860 mAh g−1 ) and low electrochemical potential (−3.04 V vs. the standard hydrogen electrode) [1–4]. However, the uncontrolled growth of Li dendrite plagues its practical application during repeated Li plating/stripping process and results in limited lifespan and relatively low coulombic efficiency [5,6]. Considerable efforts have been devoted to tackle these problems [7–10], such as some valuable works by Zhang’s group [11–14]. Among various methods, electrolyte additives are very promising to effectively inhibit dendrites by improving SEI composition of lithium metal surface [15,16]. In addition to organic additives such as FEC and VC, inorganic lithium salts, such as LiAsF6 in our previous research [17], also take great effect as additives for lithium metals. Among them, lithium nitrate (LiNO3 ) is a highly efficient electrolyte additive in lithium metal batteries and has been intensively studied in recent years [18–20]. As a donor of LiNO2 , LiNx Oy and ∗

Corresponding authors. E-mail addresses: [email protected] (Y. Zhang), [email protected] (F. Ding), xfl[email protected] (X. Li).

other substances with high lithium ion conductivity [21], LiNO3 can thus accelerate the transmission of Li+ in SEI. Inevitably, as a sacrificial additive, its lifespan is unsatisfactory as it will be consumed gradually with increased cycling numbers. Moreover, such kind of “sacrificial additives” cannot be applied in the field of high voltage lithium metal batteries for its poor solubility in the carbonate electrolyte. In view of this, inorganic salt additive besides LiNO3 is required to broaden the application in different kinds of electrolytes. As mentioned in the previous works, lithium perchlorate (LiClO4 ) has been discovered as main salt in the electrolyte of lithium ion batteries which demonstrates advantages in high lithium ion conductivity [22]. However, it has rarely been considered as additives for lithium metal batteries. Here, we reported LiClO4 as an effective electrolyte additive to improve the CE and cycling life for lithium metal batteries. Our theoretical and experimental results show that compared with other sacrificial additives, the main protective mechanism of LiClO4 is to optimize the chemical composition of SEI by promoting the decomposition of the main salt of LiTFSI via ionic association effect, while LiClO4 itself will not be consumed during cycles. LiClO4 exhibits high performance comparable to LiNO3 in ether electrolyte, and more importantly, excellent performance has also been achieved in carbonate electrolyte. These results are encouraging and, without doubt, will 2095-4956/© 2019 Science Press and Dalian Institute of Chemical Physics, Chinese Academy of Sciences. Published by Elsevier B.V. and Science Press. All rights reserved.


Y. He, Y. Zhang and P. Yu et al. / Journal of Energy Chemistry 45 (2020) 1–6

Fig. 1. The LSV tests of forward (a) and backward (b) in the range of 0.1–3 V; (c) the results of EIS tests in three-electrode system; (d) the CV plot of the electrolyte of 1 M LiTFSI DOL:DME with LiNO3 .

expand the applications of LiClO4 in electrolytes as well as provides an opportunity to explore additives in carbonate electrolyte for high performance lithium metal batteries. 2. Experimental 2.1. Materials Cathode NCA (LiNi0.8 Co0.15 Al0.05 O2 ) electrode laminates (~16.446 mg active material) were used to assemble Li-NCA cells in this work. The electrode laminates were punched into discs and further dried at ~60 °C under vacuum for 12 h. Batterygrade LiTFSI, EC and DEC were purchased from Aldrich and used as received. These chemicals were kept and handled in a glovebox (Bruker) circulated with high-purity Ar gas. The LiTFSI–LiClO4 electrolyte consisting of 1 M LiTFSI plus 7 wt% LiClO4 in the solvent mixture of EC and DEC (1:1 by volume) was blended inside the glovebox. A commercial electrolyte (that is, 1.0 M LiPF6 in the same EC-DMC solvents) was purchased and evaluated for comparison. The 1 M LiTFSI DOL:DME (v:v= =1:1) was used as ether-based electrolyte. The control electrolytes (1 M LiTFSI DOL:DME with various amounts of LiClO4 ) were also blended inside the glovebox.

2.3. Electrochemical testing Charge/discharge performances were measured using CR2032 coin-type batteries. Li metal batteries were constructed using an NCA positive electrode, a foil of Li metal anode, one piece separator (Celgard 2500), and the prepared electrolyte (80 μL in each battery). Cycling and rate performance tests were carried out with constant current and constant voltage mode using battery test system (Neware CT-4002). All the Li||NCA batteries were tested between 3.0 and 4.2 V. 1 C is equal to 190 mA g−1 , depending on the active mass loading of NCA cathode material. EIS was executed on an electrochemical working station (CHI600E, Chenhua) (±5 mV perturbation and 106 –10−2 Hz frequency range). 2.4. MD calculations The highest occupied molecular orbital (HOMO) and LUMO energies were calculated by density functional theory using the Gaussian 09 software [23]. Geometry optimization simulations were carried out using the keywords: # opt b3lyp/3–21 g scrf=(solvent=generic, read, pcm) geom=connectivity. The PCM continuum model was used to study the solvent effect on the ion associations.

2.2. Material characterizations 3. Result and discussions The morphology/microstructure and chemical component of the samples were characterized by Field-emission scanning electron microscopy (FESEM, S-4300, Hitachi, Japan), and X-ray photoelectron spectrometry (XPS, 250Xi, Thermo Scientifific, USA). The SEI film on copper sheet was prepared through CV test in the three electrodes system.

To verify the ability of LiClO4 on decomposition of LiTFSI, the linear sweep voltammetric curves in the forward/backward scan and the Nyquist plot of the electrolyte with increasing amount of LiClO4 were achieved under three-electrode system using lithium metal both as the counter and as the reference electrode and a

Y. He, Y. Zhang and P. Yu et al. / Journal of Energy Chemistry 45 (2020) 1–6


Fig. 2. The DFT optimized structures of LiTFSI (a), LiClO4 (b), LiTFSI–LiClO4 association (c) and their visual LUMOs (d)–(f); the calculated energy of LUMOs (g) and HOMOs (h); comparison of associations of LiTFSI–LiClO4 (i) and LiTFSI–ClO4 − (j).

copper sheet as the working electrode (Fig. 1). Considering that the formation of SEI is related to both of the oxidation and reduction of LiTFSI [24], the oxidation reaction and the reduction reaction had been analyzed separately. No additional peaks except the oxidation decomposition peak at ~2.7 V (Fig. 1(a)) and the reduction peak at 0.6 V (Fig. 1(b)) appeared when LiClO4 was added, and the peak intensity increased with the increase of LiClO4 both in the forward and backward scan. While in contrast, besides the peaks that caused by the LiTFSI, obvious reduction peaks appeared in electrolyte with sacrificial additive such as LiNO3 (Fig. 1(d)). It can be said that different from the sacrificial additives, LiClO4 accelerates the rate of SEI formation from electrolyte decomposition as an assistant catalytic additive, thus passivating lithium metal and keeping itself less involved in the formation of SEI. Furthermore, as denoted in Fig. 1(c), the slope of the straight line at lower frequency range and the radius of the semicircle at higher frequency range both decreased when increasing the content of LiClO4 . This indicates that with the adding of LiClO4 , the enhanced interactions among LiClO4 /LiTFSI/solvent molecules [25,26] and the increased viscosity of electrolyte induced by the ionic association formed between LiClO4 and LiTFSI [27] both can lead to more difficult migra-

tion of the Li+ because of the large ion transfer resistance, and that charges can be transferred from the electrode to the LiTFSI that adsorbed on the surface more easily, which increases the probability of receiving electrons and thus prevents further reactions between the electrolytes and the lithium metal. To thoroughly understand the decomposition reason of the LiTFSI promoted by LiClO4 , the optimal configuration of the system and the corresponding electronic structure when LiClO4 and LiTFSI formed the ionic association were obtained and analyzed using density functional theory (DFT) methods. Fig. 2(a)–(c) illustrates the optimal configurations of LiTFSI, LiClO4 , and LiTFSI–LiClO4 , respectively. LiTFSI has a symmetric structure, and the O atoms with lone-pair electrons in LiClO4 will bond with −CF3 groups due to the stronger electronegativity of F atoms. An atom of Li in LiClO4 will bond with two symmetrical O atoms in LiTFSI, thus forming the ionic association of CF3 OSO-Li+ (ClO4 )− –O(NLi)SOCF3 (see enlarged image in Fig. 2(c)). To further study the causes of the easier accepting of electrons by LiTFSI, we measured the LUMO energy levels of the molecules (Fig. 2(d)–(f)) and calculated the corresponding values (Fig. 2(g)) to analysis the reducibility changes of lithium-containing mate-


Y. He, Y. Zhang and P. Yu et al. / Journal of Energy Chemistry 45 (2020) 1–6

Fig. 3. XPS spectra of SEI: C 1s (a), F 1s (b), S 2p (c) and Cl 2p (d) before and after the adding of LiClO4 ; (e) The schematic diagram for promoting decomposition of LiTFSI.

rials in electrolytes. Based on the frontier orbital theory, reduction reactions are more likely to occur in materials with lower LUMO energy levels [28]. It is found that the LUMO is concentrated on each atoms of the corresponding molecules when no association is formed, and the calculated LUMO values of LiTFSI and LiClO4 are −2.26487 and −4.16754 eV, respectively. However, the majority of the LUMO has transferred to LiTFSI molecules rather than LiClO4 after the formation of LiTFSI–LiClO4 , and the calculated LUMO value of association is −3.34282 eV. It is reasonable to speculate that the association may contribute to the reduction of LiTFSI via reducing its LUMO levels, which makes it easier for electrons on the electrode transfer to LiTFSI, and thus promotes the reduction reaction. In contrast, the LUMO in LiTFSI–LiNO3 are concentrated on both LiNO3 and −OSOCF3 under same conditions (Fig. S1). Similar phenomena are also found on HOMO energy level of the molecules (Fig. 2(i)–(j)). It must be pointed out that the association that having the higher HOMO energy levels are LiTFSI–ClO4 − instead of LiTFSI–LiClO4 . As can be seen, the ClO4 − is inclined to bond with NSO2 CF3 rather than to form CF3 OSO–Li+ (ClO4 )− – O(NLi)SOCF3 due to the lack of Li+ . The illustrations in Fig. 2(i) and (j) reveal that the HOMO of LiTFSO–LiClO4 are concentrated on LiClO4 molecules, and that the HOMO of LiTFSI–ClO4 − are concentrated on LiTFSI molecules. According to previous analysis (Fig. 1(a) and (b)) that no new oxidation peaks appeared after adding LiClO4 , it can be speculated that LiTFSI tends to associate with ClO4 − because of the preferential transfer of electrons to LITFSI. Therefore, it can be concluded that LiClO4 only served as a catalyst for the decomposition of LiTFSI but without any changes during the process. The decomposition of LiTFSI will generate LiF, LiSO2 CF3 and other SEI component. The high-resolution XPS measurements are applied to investigate the effect of catalytic additive of LiClO4 on SEI composition. As can be seen in Fig. 3(a), C 1s spectra of SEI consisted of four peaks, among which the peaks at the binding energies of ~292.8 and 286.4 eV were corresponded to the CF3 group [29,30] and C−O bond [31], respectively. With increasing content of LiClO4 , the intensity of these two peaks both increased. According to the following decomposition equations of LiTFSI [21]:

LiTFSI + 2e− + 2Li+ → Li2 NSO2 CF3 + LiSO2 CF3


LiTFSI + ne− + nLi+ → Li3 N + Li2 S2 O4 + LiF + C2 Fx Liy


The CF3 groups belonged to LiTFSI or Li2 NSO2 CF/LiSO2 CF3 , suggesting that the adding of LiClO4 was favorable to the reaction (1). Moreover, high-resolution spectra of F 1s (Fig. 3(b)) clearly showed higher intensity of LiF in SEI after adding LiClO4 , further unambiguously confirming that the association generated by LiClO4 contributes to the decomposition reaction of LiTFSI, and that the obtained LiF plays a critical role in improving the ionic conductivity of Li+ in SEI. A similar phenomenon was also found for high-resolution spectra of S 2p (Fig. 3(c)). The relative content of S atoms was in lower level without adding LiClO4 , and the reduction products all had weak peak intensities. With increasing the content of LiClO4 , the content of RSO3 Li, Li2 S, Li2 Sx , and Li2 SO3 all increased significantly. According to the following equations [21,32]:

Li2 S2 O4 + 6e− + 6Li+ → 2Li2 S + 4Li2 O


Li2 S2 O4 + 4e− + 4Li+ → Li2 SO3 + Li2 S + Li2 O


The completely reducing of LiTFSI promoted by LiClO4 has led to the secondary reactions of Li2 S2 O4 [33], resulting in an increase of sulfur-containing compounds. It must be pointed out that the content of Cl in SEI remained very low with/without adding LiClO4 (Fig. 3(d)), again confirming that LiClO4 does not participate in the formation of SEI during catalytic decomposition of LiTFSI, which is consistent with previous LSV analysis results (Fig. 1(a) and (b)). Thus, we can write a proper reaction equation to describe the passivating process: LiClO4


LiTFSI → CF3 OSO − Li (ClO4 )− − O(NLi )SOCF3 → LiF + Li3 N + Li2 S2 O4 + C2 Fx Liy A schematic diagram describes the passivating process that promotes decomposition of LiTFSI as shown in Fig. 3(e). It is noted that the lifespan of lithium metal batteries is severely limited by the stability of the lithium anode [34]. Therefore, the cycling performance of Li−Li symmetrical cells needs to be carefully investigated (Fig. S2). Next, the Li−Cu cells with 1 M LiTFSI DOL:DME as the control electrolyte were assembled to test

Y. He, Y. Zhang and P. Yu et al. / Journal of Energy Chemistry 45 (2020) 1–6


Fig. 4. The results of Li−Cu cells using the electrolytes of 1 M LiTFSI DOL:DME (a)–(c) and 1 M LiTFSI EC:DEC (d) with various amounts of LiNO3 or LiClO4 ; (e) comparisons of the cycling life and (f) rate performances of NCA−Li cells using various electrolytes.

the coulombic efficiency. As shown in Fig. 4(a), the Li−Cu cell with 1 M LiTFSI DOL:DME (no additives) displayed a poor cycling performance of only 20 cycles with an average CE of ~90%. In contrast, when adding 1% LiClO4 (mass ratio), an obvious improvement on the CE stability of ~97.2% over 70 cycles was achieved, which is comparable to the most promising additive of LiNO3 (with a CE of only 93.1%). Inspired by the encouraging results, we further increased the content of LiClO4 to 3% for testing, a comparison of electrolyte with 3% LiNO3 was also prepared. As displayed in Fig. 4(b), the CE of Li−Cu cell in electrolyte with 3% LiNO3 over 90 cycles was 98.5%. In contrast, the CE in electrolyte with 3% LiClO4

was as high as 97.7% and remained stable over 116 cycles. Need to mention that LiNO3 saturated in ether electrolyte with an adding amount of 6.8%, and the CE in electrolyte with 7% LiClO4 was thus achieved for comparison. As shown in Fig. 4(c), the performance of the Li−Cu cell in electrolyte with 7% LiClO4 was encouraging since an almost comparable CE and lifespan were obtained to that of the saturated LiNO3 . It can be said that the LiClO4 is totally feasible to be an excellent alternative of LiNO3 for suppression of dendrite growth on repeated lithium stripping/plating deposition. Herein, the adding content of more than 10%, usually led to poor stability of Li−Cu cell (Fig. S3), was not considered due to the resulting


Y. He, Y. Zhang and P. Yu et al. / Journal of Energy Chemistry 45 (2020) 1–6

consequences of uneven SEI formation and unstable sediments. It is noting that compared with the poor solubility of LiNO3 in carbonate electrolyte, LiClO4 exhibits good solubility in carbonate electrolyte. In view of this, we assembled the Li−Cu cell in 1 M LiTFSI EC:DEC with 7% LiClO4 for testing, expecting to achieve a better performance. As clearly shown in Fig. 4(d), after adding 7% LiClO4 in the electrolyte, the cell showed significantly enhanced cycling performance (from 25 cycles to 95 cycles), illustrating good passivation and protection effect of LiClO4 on lithium metal when coupled with carbonate electrolyte. To further study the contribution of LiClO4 in carbonate electrolyte for electrochemical performance, we assembled the NCA−Li full cell for testing at a voltage range between 3 V and 4.2 V (Fig. 4(e)–(f)). As can be seen in Fig. 4(e), the cell with 7% LiClO4 in 1 M LiTFSI EC:DEC exhibited a much longer cycle life over 60 cycles than that without LiClO4 , which is even longer than that using the commercial electrolyte of 1 M LiPF6 EC:DMC (26 cycles). In addition, as shown in Fig. 4(f), stable specific capacity of 202 mAh g−1 was achieved upon gradually increasing the discharge/charge rate from 0.5 C to 1 C, superior than those using 1 M LiTFSI EC:DEC (~168 mAh g−1 ) and 1 M LiPF6 EC:DMC(~190 mAh g−1 ). Moving back to 0.5 C, the specific capacities of NCA−Li full cell returned to the initial rate value of ~233 mAh g–1 . Therefore, it can be concluded that the SEI with higher reduction degree, which was generated from the reduction of LiTFSI that catalyzed by LiClO4 , played an important role in passivating the lithium metal, thus enabling suppression of dendrite growth of Li metal with a high CE and long lifespan in plating/stripping cycles. 4. Conclusions In conclusion, LiClO4 was first used as a bi-soluble additive in ether and carbonate electrolytes with enhanced coulombic efficiencies and cycling performances in lithium batteries. Through formation of ionic association, LiTFSI could be fully decomposed under the catalyzing of LiClO4 , and an optimized chemical composition of SEI was obtained. Experiments results and theoretical analysis reveal that an increased amount of the LiF and Li2 S were achieved through catalyst effect of the LiClO4 to promote Li+ transport, which effectively prevented the dendrite growth. Thus, significant enhancement of the efficiency was achieved in ether electrolyte comparable to that of lithium nitrate. Furthermore, a remarkable extended lifespan was realized at high voltage of 4.2 V in carbonate electrolytes for lithium metal batteries. This study not only suggests new applications of LiClO4 as bi-soluble additive, but provides new insights into developing additives for high voltage electrolytes. Declaration of Competing Interest None. Acknowledgments The authors gratefully acknowledge financial supports from the Foundation of National Key Laboratory (No. 6142808180202), P.R. China and the Pre-Research Foundation (Nos. 61407210406, 61407210208, 41421080401), P.R. China. The authors would like to express our sincere thanks to Dr. Jing Hu, for her valuable comments on the writing of this article.

Supplementary material Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.jechem.2019.09.033. References [1] X.-B. Cheng, R. Zhang, C.-Z. Zhao, Q. Zhang, Chem. Rev. 117 (2017) 10403–10473. [2] H. Liu, X.-B. Cheng, J.-Q. Huang, S. Kaskel, S. Chou, H.S. Park, Q. Zhang, ACS Mater. Lett. 1 (2019) 217–229. [3] H. Liu, X.-B. Cheng, Z. Jin, R. Zhang, G. Wang, L.-Q. Chen, Q.-B. Liu, J.-Q. Huang, Q. Zhang, Energy Chem. 1 (2019) 10 0 0 03. [4] W. Xu, J. Wang, F. Ding, X. Chen, E. Nasybulin, Y. Zhang, J.-G. Zhang, Energy Environ. Sci. 7 (2014) 513–537. [5] J. Qian, W.A. Henderson, W. Xu, P. Bhattacharya, M. Engelhard, O. Borodin, J.-G. Zhang, Nat. Commun. 6 (2015) 6362. [6] Z.-i. Takehara, J. Power Sources 68 (1997) 82–86. [7] X.-B. Cheng, C. Yan, X.-Q. Zhang, H. Liu, Q. Zhang, ACS Energy Lett. 3 (2018) 1564–1570. [8] X. Ren, L. Zou, S. Jiao, D. Mei, M.H. Engelhard, Q. Li, H. Lee, C. Niu, B.D. Adams, C. Wang, J. Liu, J.-G. Zhang, W. Xu, ACS Energy Lett. 4 (2019) 896–902. [9] L. Wang, Z. Zhou, X. Yan, F. Hou, L. Wen, W. Luo, J. Liang, S.X. Dou, Energy Storage Mater. 14 (2018) 22–48. [10] W. Zhang, S. Zhang, L. Fan, L. Gao, X. Kong, S. Li, J. Li, X. Hong, Y. Lu, ACS Energy Lett. 4 (2019) 644–650. [11] X.-R. Chen, B.-Q. Li, C.-X. Zhao, R. Zhang, Q. Zhang, Small Methods 0 (2019) 1900177. [12] F. Wu, Y.-X. Yuan, X.-B. Cheng, Y. Bai, Y. Li, C. Wu, Q. Zhang, Energy Storage Mater. 15 (2018) 148–170. [13] X.-Q. Zhang, X. Chen, L.-P. Hou, B.-Q. Li, X.-B. Cheng, J.-Q. Huang, Q. Zhang, ACS Energy Lett. 4 (2019) 411–416. [14] C.-Z. Zhao, P.-Y. Chen, R. Zhang, X. Chen, B.-Q. Li, X.-Q. Zhang, X.-B. Cheng, Q. Zhang, Sci. Adv. 4 (2018) eaat3446. [15] J. Zheng, M.H. Engelhard, D. Mei, S. Jiao, B.J. Polzin, J.-G. Zhang, W. Xu, Nat. Energy 2 (2017) 17012. [16] J. Zheng, P. Yan, D. Mei, M.H. Engelhard, S.S. Cartmell, B.J. Polzin, C. Wang, J.-G. Zhang, W. Xu, Adv. Energy Mater. 6 (2016) 1502151. [17] X. Ren, Y. Zhang, M.H. Engelhard, Q. Li, J.-G. Zhang, W. Xu, ACS Energy Lett. 3 (2018) 14–19. [18] S. Liu, X. Xia, Y. Zhong, S. Deng, Z. Yao, L. Zhang, X.-B. Cheng, X. Wang, Q. Zhang, J. Tu, Adv. Energy Mater. 8 (2018) 1702322. [19] C. Yan, Y.-X. Yao, X. Chen, X.-B. Cheng, X.-Q. Zhang, J.-Q. Huang, Q. Zhang, Angew. Chem. Int. Ed. 130 (2018) 14251–14255. [20] S.S. Zhang, Electrochim. Acta 70 (2012) 344–348. [21] C. Xu, B. Sun, T. Gustafsson, K. Edström, D. Brandell, M. Hahlin, J. Mater. Chem. A 2 (2014) 7256–7264. [22] L. Yu, G.Z. Chen, Faraday Discuss 190 (2016) 231–240. [23] M. Frisch, J. Chem. Phys. 98 (1993) 5648. [24] K. Matsumoto, K. Inoue, K. Nakahara, R. Yuge, T. Noguchi, K. Utsugi, J. Power Sources 231 (2013) 234–238. [25] L. Suo, Y.-S. Hu, H. Li, M. Armand, L. Chen, Nat. Commun. 4 (2013) 1481. [26] Z. Zeng, V. Murugesan, K.S. Han, X. Jiang, Y. Cao, L. Xiao, X. Ai, H. Yang, J.-G. Zhang, M.L. Sushko, J. Liu, Nat. Energy 3 (2018) 674–681. [27] Z. Wang, B. Huang, R. Xue, L. Chen, X. Huang, J. Electrochem. Soc. 145 (1998) 3346–3350. [28] X. Chen, H.-R. Li, X. Shen, Q. Zhang, Angew. Chem. Int. Ed. 130 (2018) 16885–16889. [29] W. Eck, V. Stadler, W. Geyer, M. Zharnikov, A. Gölzhäuser, M. Grunze, Adv. Mater. 12 (20 0 0) 805–808. [30] D.K. Sarkar, M. Farzaneh, R.W. Paynter, Appl. Surf. Sci. 256 (2010) 3698–3701. [31] X. Zhang, Q. Zhang, X.-G. Wang, C. Wang, Y.-N. Chen, Z. Xie, Z. Zhou, Angew. Chem. Int. Ed. 130 (2018) 12996–130 0 0. [32] A.M. Andersson, M. Herstedt, A.G. Bishop, K. Edström, Electrochim. Acta 47 (2002) 1885–1898. [33] L. Madec, L. Ma, K.J. Nelson, R. Petibon, J.-P. Sun, I.G. Hill, J.R. Dahn, J. Electrochem. Soc. 163 (2016) A1001–A1009. [34] S. Jiao, X. Ren, R. Cao, M.H. Engelhard, Y. Liu, D. Hu, D. Mei, J. Zheng, W. Zhao, Q. Li, N. Liu, B.D. Adams, C. Ma, J. Liu, J.-G. Zhang, W. Xu, Nat. Energy 3 (2018) 739–746.