English

• Green energy storage devices have been extensively developed nowadays [1,2]. Specifically, alkali metal (Li, Na, K) batteries have become a research hotspot due to their high theoretical capacities (3861, 1166 and 685 mAh/g) and low redox potentials (−3.04, −2.71 and −2.93 V vs. standard hydrogen electrode (SHE)), which can meet the growing demand for high energy density in rechargeable energy storage devices [3-6]. However, dendrite growth and interfacial issues have hindered the practical application of solid-state alkali metal anodes in rechargeable batteries [7-10]. Dendrites may induce safety concerns of short circuits as separator piercing, and the formation of dead metal would bring irreversible capacity fading. In addition, interfacial issues may cause thermal runaway, poor cycle stability and low coulombic efficiency due to the continuous formation of unstable solid electrolyte interfaces (SEI) [11-14].

  Various strategies have been devised to address these problems. One of them is the electrolyte modification for SEI stabilization, which mainly studied the concentration and additives of electrolytes [15-18]; another route is the artificial interface construction by designing SEI from atomic/molecular layer deposition or chemical reactions [19,20]. In addition, the application of solid-state electrolytes is also an effective approach [21-25]. Another widely used strategy is preparing nanostructured anode hosts, which aim to build carbon-based or porous metal current collectors [26-28] or metalphilic interlayers [29,30]. However, there are still challenges and large room for improvement of these anodes due to the complex composition of SEI, the interface issues of solid-state electrolytes and the complication of combining the metal anodes and the nanostructure hosts.

  Compared with solid metal anodes hindered by the issues mentioned above, liquid metal anode (LMA) is a promising alternative since it is fluidly, dendrite-free, self-healable, and has fast ion transport with high energy density [25,31-38]. Currently, among LMAs, a commercialized application is high-temperature sodium-sulfur (Na-S) batteries, which could be operated at above 300 ℃. However, the high operating temperature and expensive thermal management system limit its application range [39], and studying LMAs at room temperature (RT) is necessary.

  When the sodium metal content is between 9.2 wt% and 58.2 wt%, the sodium-potassium (Na-K) alloy remains liquid at RT, making it a research hotspot in recent years [40-46]. However, it is hard to serve as an electrode alone due to the high surface tension, fluidity, infinite interface formation and volume change. Thus, the study of Na-K LMA mainly focuses on immobilizing the alloy with porous hosts to make it practical. In 2016, Goodenough and co-workers first proposed using carbon paper to absorb Na-K alloy as an anode for battery systems with a high operating temperature of 420 ℃ [47]. In 2018, metal foams (Cu, Al and Ni) were introduced as Na-K alloy hosts by vacuum infiltration at RT [48]. In 2019, Zhang et al. reported a highly graphitized carbon paper as a host for Na-K alloy by forming a graphite intercalation composite [49]. Although significant progress has been achieved, strategies for constructing composite with liquid metal at RT for practical applications are still highly needed.

  Herein, in this work, we have designed an amorphous hollow carbon film (AHCF) as a flexible host for liquid Na-K alloy anode. Specifically, two components are selected for constructing the AHCF: hollow carbon spheres (HCS) and carbon fibers from electrospinning (Scheme S1 in Supporting information). Specifically, HCS was prepared by the hard-template method with SiO₂ and phenol-formaldehyde resin (PF) [50,51]. The as-prepared HCS was then added and dispersed in the polyacrylonitrile (PAN) and DMF solution as precursors for electrospinning. After collecting and annealing the obtained film in the N₂ atmosphere, the AHCF can be obtained.

  When immersing the AHCF into the Na-K alloy, the liquid metal can be rapidly adsorbed and infiltrated into the hollow host, forming a Na-K@AHCF electrode. More experimental details can be found in the supporting materials. The designed AHCF host is expected to play a significant role in the following aspects. First, as the spheres are hollow and amorphous with abundant defects, the AHCF possesses a rapid adsorption ability to Na-K alloy. Namely, the liquid alloy can be infiltrated inside the hollow carbon spheres according to the surface defects adsorption and capillarity phenomena, forming the Na-K@AHCF electrode. In addition, Scheme 1 shows the schematic diagram of pure Na-K alloy and Na-K@AHCF during the deposition/stripping cycling processes. The hollow structure could offer sufficient space for Na-K liquid alloy volume change during charge/discharge processes. The separated hollow carbon spheres can also isolate the liquid metal droplets, providing better tolerance for electrode failure from volume changing. Further, the carbon-based matrix can improve its coulombic efficiency by providing skeletons to stabilize SEI, avoiding the continuous formation of interfaces between LMA and electrolytes. As a result, the Na-K@AHCF electrode could improve electrochemical performance compared to the K metal. When matching with cathode material of sulfurized polyacrylonitrile (SPAN), the obtained K-S battery could exhibit good cycle stability and rate performance. The design would provide a strategy for promoting the development of liquid metal anode.

  Scheme 1

  

  Scheme 1.  Schematic diagram of the metal liquid electrodes during the deposition/stripping process. (a) Na-K alloy; (b) Na-K alloy immersed into AHCF host.

  DownLoad: Full-Size Img PowerPoint

  To identify the morphologies of the synthesized HCS, AHCF and Na-K@AHCF, scanning electron microscopy (SEM) and transmission electron microscopy (TEM) with an energy dispersive spectrometer (EDS) were used. From Fig. 1, it can be found that most HCS and precursors have spherical morphologies, uniform size (the diameter is about 350 nm) and similar wall thickness (about 20 nm). It should be noted that from Fig. 1g, the walls of HCS are full of open pores, which could offer paths for the liquid alloy to be infiltrated into the cavity. Besides, from Figs. 1d and h, after electrospinning and carbonization, the AHCF presented a uniform distribution of hollow spheres in and on the fiber, presenting an open and hollow grapes-like structure. With the contribution of HCS, there exist plenty of pores and defects of AHCF on the surface. Apart from these, morphologies and internal structures of the carbonized PAN fibers were also characterized by SEM and TEM (Fig. S1 in Supporting information).

  Figure 1

  

  Figure 1.  SEM and TEM images of (a, e) SiO₂@PF/SiO₂, (b, f) SiO₂@C/SiO₂, (c, g) HCS and (d, h) AHCF. (i, j) TEM images of Na-K@AHCF. (k, l) STEM images of Na-K@AHCF and corresponding elemental mapping images of C, K and Na.

  DownLoad: Full-Size Img PowerPoint

  After adsorbing Na-K alloy at RT, it can be verified that liquid Na-K alloy has been successfully absorbed on and into the cavities of AHCF and HCS, which could be observed from the TEM images in Figs. 1i and j. To further prove that Na-K alloy can be absorbed into the cavity, the scanning transmission electron microscopy (STEM) at different magnifications (Figs. 1k and l) and their corresponding elemental mapping was presented for the Na-K@AHCF. In Fig. 1k, the carbon wall and Na-K alloy can be distinguished with different contrasts in the STEM image. Combined with the uniform distribution of elemental C, K and Na, it can also be inferred that Na-K alloy is distributed evenly in the cavities.

  XRD was applied to identify the composition and crystal structures of the prepared carbon hosts, as shown in Fig. 2a. The AHCF exhibits a similar pattern to the carbonized HCS and carbonized PAN film, which are all amorphous carbon. It suggests that there are plenty of defects on the surface of AHCF cavities, which could lead to the adsorption and infiltration abilities to the liquid alloy. Nitrogen adsorption-desorption isotherms (Fig. 2b) and corresponding pore size distributions (Fig. 2c) are carried out for the hollow materials. The adsorption-desorption curve of HCS shows a typical type IV curve, and the specific surface area can reach 1038.29 cm³/g. It can be seen from the pore size distribution curve that HCS has a large number of mesopores with diameters of 2–10 nm and microporous. However, the specific surface area of AHCF is only about half that of HCS, 511.88 cm³/g, and it has a large number of microporous. This is mainly because the precursor HCS accounts for only a component for building AHCF. The carbonized PAN fibers could offer the film better mechanical strength, and the pores and the surface defects of the samples are conducive to liquid infiltration and adsorption. It also should be noted that the addition of HCS does not significantly reduce the flexibility of AHCF compared with carbonized PAN film. Both the two samples are free-standing and bendable, as shown in Figs. 2d and f.

  Figure 2

  

  Figure 2.  (a) XRD patterns of HCS, AHCF and carbonized PAN. (b) Nitrogen adsorption-desorption isotherm and (c) the corresponding pore size distribution of HCS and AHCF. Photographs of bent (d) carbonized PAN and (f) AHCF. Photographs of (e) PAN, (g) AHCF contacting with Na-K alloy.

  DownLoad: Full-Size Img PowerPoint

  The carbonized PAN film and AHCF are then emersed into the liquid Na-K alloy to explore the adsorption ability of sample cavities to Na-K alloy. The corresponding photos are shown in Figs. 2e and g, respectively. When contacting with the high-surface-tension liquid alloy, the carbonized PAN cannot be wetted, while AHCF is infiltrated quickly, forming Na-K@AHCF (Videos S2 and S3 in Supporting information). This phenomenon is attributed to the amorphous HCS component, its rich defects and large cavities.

  The symmetric cells were assembled with a K-based electrolyte to investigate the electrochemical performance of the Na-K@AHCF. Fig. 3a shows the rate performance of Na-K@AHCF//Na-K@AHCF and K//K metal cells, stripping and depositing at various current densities (0.1, 0.2, 0.5, 1, 2 and 5 mA/cm²). The Na-K@AHCF cell presents a more stable voltage curve and a lower overpotential compared with the K metal one. The symmetrical cell of K failed after 22 cycles, which may be caused by the unstable interfaces of K metal. In contrast, the Na-K@AHCF exhibited enhanced cycling stability and rate performance. At a high current density of 2 mA/cm², a similar phenomenon also can be observed in Fig. 3b, in which Na-K@AHCF//Na-K@AHCF exhibits enhanced cycle stability than the K//K electrodes.

  Figure 3

  

  Figure 3.  Electrochemical performance of symmetrical cells cycled in 2 mol/L KFSI EC/DEC electrolytes. (a) Rate performance of Na-K@AHCF//Na-K@AHCF and K//K cells. (b, c) Cycling performance of the cells at different current densities.

  DownLoad: Full-Size Img PowerPoint

  Moreover, the contributions of the hollow structure can also be reflected by the cycle life up to over 400 h at the current density of 0.1 mA/cm² (Fig. 3c), showing good cycling stability. However, it should be noted that the polarization values become smaller at about 200 h and 350 h. It is probably due to the liquid volume changes in the hollow cavities, and some liquid will seep outside the carbon. Therefore, the electrode conductivity would change during cycling, causing variations in the polarization potentials.

  In order to characterize the full cell performance, a cathode material of SPAN with a high capacity and long cycle stability was prepared [52-54]. Fig. 4a shows a schematic diagram of a K-S battery using Na-K@AHCF anode and SPAN cathode. The corresponding schematic structure of SPAN is also shown in Fig. 4b. The survey scan and corresponding high-resolution spectrum of C 1s and S 2p are shown in Fig. S4 (Supporting information) and Fig. 4c, proving the successful preparation of the SPAN. The rate performance and cycle stability of this K-S battery are studied in a KFSI in EC/DEC electrolyte with a cycling voltage window from 0.5 V to 3 V (Figs. 4d and e). Under the current density of 0.1, 0.2, 0.5, 1 and 2 A/g, the discharge capacity of the cell can reach 400.4, 380.2, 341.3, 304.4 and 252.4 mAh/g, respectively. At a current density of 0.5 A/g, the capacity did not decrease significantly after 80 cycles. The galvanostatic charge/discharge curves at various current densities are also shown in Fig. 4f, according to a typical K-SPAN battery. Thus, it can be confirmed that when matched with cathode materials of SPAN, Na-K@AHCF anode exhibits excellent electrochemical performance, showing practical application prospects.

  Figure 4

  

  Figure 4.  Electrochemical performance of full cells of K-S battery with SPAN and Na-K@AHCF. The structure schematic of (a) the K-S batteries and (b) SPAN. (c) High-resolution XPS spectrum of S 2p in SPAN. (d) Rate performance and (e) cycle stability of the K-S battery. (f) Galvanostatic charging/discharging curves at various current densities.

  DownLoad: Full-Size Img PowerPoint

  In summary, we designed and presented a flexible and free-standing film with a hollow carbon structure, which could enable fast adsorption of Na-K liquid alloy as a host. The designed Na-K@AHCF composite can tolerate the volume expansion of the metal anode during charge/discharge processes and form a more stable interface. As a result, both symmetrical cells of Na-K@AHCF and full K-S battery can exhibit excellent rate performance and cycle stability. The concept of the host design could offer inspiration for advanced alkali metal batteries.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  Acknowledgment

  This work was financially supported by the National Natural Science Foundation of China (No. 51702155).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2022.107767.

  
  
• 1. [1]

     F. Xu, Y. Zhai, E. Zhang, et al., Angew. Chem. Int. Ed. 59 (2020) 19460–19467. doi: 10.1002/anie.202005118

  2. [2]

     M. Shao, C. Sun, T. Chen, et al., J. Energy Chem. 65 (2022) 127–132. doi: 10.1016/j.jechem.2021.05.037

  3. [3]

     J. Xiang, L. Yang, L. Yuan, et al., Joule 3 (2019) 2334–2363. doi: 10.1016/j.joule.2019.07.027

  4. [4]

     H. Wang, D. Yu, C. Kuang, et al., Chem 5 (2019) 313–338. doi: 10.1016/j.chempr.2018.11.005

  5. [5]

     B.Y. Ma, P. Bai, Adv. Energy Mater. 12 (2022) 2102967. doi: 10.1002/aenm.202102967

  6. [6]

     G. Liang, X. Li, Y. Wang, et al., Nano Res. Energy 1 (2022) 9120002. doi: 10.26599/NRE.2022.9120002

  7. [7]

     X. Cao, X. Ren, L. Zou, et al., Nat. Energy 4 (2019) 796–805. doi: 10.1038/s41560-019-0464-5

  8. [8]

     Y. Zhao, K.R. Adair, X. Sun, Energy Environ. Sci. 11 (2018) 2673–2695. doi: 10.1039/c8ee01373j

  9. [9]

     S. Liu, J. Mao, Q. Zhang, et al., Angew. Chem. Int. Ed. 59 (2019) 3638–3644.

  10. [10]

      X. Liang, Q. Pang, I.R. Kochetkov, et al., Nat. Energy 2 (2017) 17119. doi: 10.1038/nenergy.2017.119

  11. [11]

      C. Qin, D. Wang, Y. Liu, et al., Nat. Commun. 12 (2021). 7184 doi: 10.1038/s41467-021-27494-z

  12. [12]

      X. Li, J. Liu, C. Chen, et al., Chin. Chem. Lett. 32 (2021) 983–989. doi: 10.1016/j.cclet.2020.09.018

  13. [13]

      R. Chen, A.M. Nolan, J. Lu, et al., Joule 4 (2020) 812–821. doi: 10.1016/j.joule.2020.03.012

  14. [14]

      F. Jiang, S. Yang, X. Cheng, et al., J. Energy Chem. 72 (2022) 158–165. doi: 10.1016/j.jechem.2022.05.005

  15. [15]

      P. Jaumaux, J. Wu, D. Shanmukaraj, et al., Adv. Funct. Mater. 31 (2021) 2008644. doi: 10.1002/adfm.202008644

  16. [16]

      H. Sun, G. Zhu, X. Xu, et al., Nat. Commun. 10 (2019) 3302. doi: 10.1038/s41467-019-11102-2

  17. [17]

      S. Choudhury, S. Wei, Y. Ozhabes, et al., Nat. Commun. 8 (2017) 898. doi: 10.1038/s41467-017-00742-x

  18. [18]

      N. Xiao, W.D. McCulloch, Y. Wu, J. Am. Chem. Soc. 139 (2017) 9475–9478. doi: 10.1021/jacs.7b04945

  19. [19]

      Y. Zhao, L.V. Goncharova, A. Lushington, et al., Adv. Mater. 29 (2017) 1606663. doi: 10.1002/adma.201606663

  20. [20]

      Y. Zhao, L.V. Goncharova, Q. Zhang, et al., Nano Lett. 17 (2017) 5653–5659. doi: 10.1021/acs.nanolett.7b02464

  21. [21]

      I.D. Seymour, A. Aguadero, J. Mater. Chem. A 9 (2021) 19901–19913. doi: 10.1039/d1ta03254b

  22. [22]

      N. Wu, P.H. Chien, Y. Qian, et al., Angew. Chem. Int. Ed. 59 (2019) 4131–4137.

  23. [23]

      T. Famprikis, P. Canepa, J.A. Dawson, et al., Nat. Mater. 18 (2019) 1278–1291. doi: 10.1038/s41563-019-0431-3

  24. [24]

      Y. Shen, Y. Zhang, S. Han, et al., Joule 2 (2018) 1674–1689. doi: 10.1016/j.joule.2018.06.021

  25. [25]

      X. Guo, J. Bae, Y. Ding, et al., Adv. Funct. Mater. 31 (2021) 2010863. doi: 10.1002/adfm.202010863

  26. [26]

      D. Lin, Y. Liu, Z. Liang, et al., Nat. Nanotechnol. 11 (2016) 626–632. doi: 10.1038/nnano.2016.32

  27. [27]

      G. Zheng, S.W. Lee, Z. Liang, et al., Nat. Nanotechnol. 9 (2014) 618–623. doi: 10.1038/nnano.2014.152

  28. [28]

      S. Liu, S. Tang, X. Zhang, et al., Nano Lett. 17 (2017) 5862–5868. doi: 10.1021/acs.nanolett.7b03185

  29. [29]

      L. Ye, M. Liao, T. Zhao, et al., Angew. Chem. Int. Ed. 58 (2019) 17054–17060. doi: 10.1002/anie.201910202

  30. [30]

      B. Sun, P. Li, J. Zhang, et al., Adv. Mater. 30 (2018) 1801334. doi: 10.1002/adma.201801334

  31. [31]

      X. Guo, L. Zhang, Y. Ding, et al., Energy Environ. Sci. 12 (2019) 2605–2619. doi: 10.1039/c9ee01707k

  32. [32]

      K. Wang, K. Jiang, B. Chung, et al., Nature 514 (2014) 348–350. doi: 10.1038/nature13700

  33. [33]

      Z. Li, J.T. Zhang, Y.M. Chen, et al., Nat. Commun. 6 (2015) 8850. doi: 10.1038/ncomms9850

  34. [34]

      G. Liu, J.Y. Kim, M. Wang, et al., Adv. Energy Mater. 8 (2018) 1703652. doi: 10.1002/aenm.201703652

  35. [35]

      X. Lu, G. Li, J.Y. Kim, et al., Nat. Commun. 5 (2014) 4578. doi: 10.1038/ncomms5578

  36. [36]

      X. Guo, Y. Ding, G. Yu, Adv. Mater. 33 (2021) 2002577.

  37. [37]

      X. Guo, Y. Ding, H. Gao, et al., Adv. Mater. 32 (2020) 2000316. doi: 10.1002/adma.202000316

  38. [38]

      Y. Ding, X. Guo, Y. Qian, et al., Adv. Mater. 32 (2020) 2002577. doi: 10.1002/adma.202002577

  39. [39]

      S. Wei, S. Xu, A. Agrawral, et al., Nat. Commun. 7 (2016) 11722. doi: 10.1038/ncomms11722

  40. [40]

      D. Guo, J. Qin, Z. Yin, et al., Nano Energy 45 (2018) 136–147. doi: 10.1016/j.nanoen.2017.12.038

  41. [41]

      X. Guo, Y. Ding, L. Xue, et al., Adv. Funct. Mater. 28 (2018) 1804649. doi: 10.1002/adfm.201804649

  42. [42]

      A.C. Baclig, G. McConohy, A. Poletayev, et al., Joule 2 (2018) 1287–1296. doi: 10.1016/j.joule.2018.04.008

  43. [43]

      M. Huang, B. Xi, Z. Feng, et al., Small 15 (2019) 1804916. doi: 10.1002/smll.201804916

  44. [44]

      L. Qin, W. Yang, W. Lv, et al., Chem. Commun. 54 (2018) 8032–8035. doi: 10.1039/C8CC03545H

  45. [45]

      L. Zhang, Y. Li, S. Zhang, et al., Small Methods 3 (2019) 1900383. doi: 10.1002/smtd.201900383

  46. [46]

      C. Liu, H.Y. Chen, W.T. Deng, et al., J. Phys. Chem. Lett. 12 (2021) 9321–9327. doi: 10.1021/acs.jpclett.1c02248

  47. [47]

      L. Xue, H. Gao, W. Zhou, et al., Adv. Mater. 28 (2016) 9608–9612. doi: 10.1002/adma.201602633

  48. [48]

      L. Xue, W. Zhou, S. Xin, et al., Angew. Chem. Int. Ed. 57 (2018) 14184–14187. doi: 10.1002/anie.201809622

  49. [49]

      L. Zhang, S. Peng, Y. Ding, et al., Energy Environ. Sci. 12 (2019) 1989–1998. doi: 10.1039/c9ee00437h

  50. [50]

      H. Zhang, O. Noonan, X. Huang, et al., ACS Nano 10 (2016) 4579–4586. doi: 10.1021/acsnano.6b00723

  51. [51]

      J. Ding, H. Zhang, H. Zhou, et al., Adv. Mater. 31 (2019) 1900429. doi: 10.1002/adma.201900429

  52. [52]

      T.H. Hwang, D.S. Jung, J.S. Kim, et al., Nano Lett. 13 (2013) 4532–4538. doi: 10.1021/nl402513x

  53. [53]

      Z. Li, J. Zhang, Y. Lu, et al., Sci. Adv. 4 (2018) eaat1687. doi: 10.1126/sciadv.aat1687

  54. [54]

      S. Li, H. Dai, Y. Li, et al., Energy Storage Mater. 18 (2019) 222–228. doi: 10.1016/j.ensm.2018.09.012

• 

  Scheme 1  Schematic diagram of the metal liquid electrodes during the deposition/stripping process. (a) Na-K alloy; (b) Na-K alloy immersed into AHCF host.

  下载: 全尺寸图片 幻灯片
  

  Figure 1  SEM and TEM images of (a, e) SiO₂@PF/SiO₂, (b, f) SiO₂@C/SiO₂, (c, g) HCS and (d, h) AHCF. (i, j) TEM images of Na-K@AHCF. (k, l) STEM images of Na-K@AHCF and corresponding elemental mapping images of C, K and Na.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a) XRD patterns of HCS, AHCF and carbonized PAN. (b) Nitrogen adsorption-desorption isotherm and (c) the corresponding pore size distribution of HCS and AHCF. Photographs of bent (d) carbonized PAN and (f) AHCF. Photographs of (e) PAN, (g) AHCF contacting with Na-K alloy.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Electrochemical performance of symmetrical cells cycled in 2 mol/L KFSI EC/DEC electrolytes. (a) Rate performance of Na-K@AHCF//Na-K@AHCF and K//K cells. (b, c) Cycling performance of the cells at different current densities.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Electrochemical performance of full cells of K-S battery with SPAN and Na-K@AHCF. The structure schematic of (a) the K-S batteries and (b) SPAN. (c) High-resolution XPS spectrum of S 2p in SPAN. (d) Rate performance and (e) cycle stability of the K-S battery. (f) Galvanostatic charging/discharging curves at various current densities.

  下载: 全尺寸图片 幻灯片
•