PAN/PI functional double-layer coating for dendrite-free lithium metal anodes

Fei Shen a, Kaiming Wang a, Yuting Yin a, Le Shi a, Dingyuan Zeng a and Xiaogang Han *ab
aState Key Laboratory of Electrical Insulation and Power Equipment, School of Electrical Engineering, Xi'an Jiaotong University, Xi'an 710049, Shaanxi, China. E-mail: xiaogang.han@xjtu.edu.cn
bKey Laboratory of Smart Grid of Shaanxi Province, Xi'an, Shaanxi 710049, China

Received 13th December 2019 , Accepted 16th February 2020

First published on 18th February 2020


Li metal anodes attract intensive attention due to their high specific capacity, low density and the lowest electrochemical potential. However, inferior coulombic efficiency (CE) and safety hazards induced by uncontrollable Li dendrite growth are main obstacles against the commercialization of Li metal batteries (LMBs). Herein, we demonstrate that the polyacrylonitrile fiber/polyimide sphere (PAN fiber/PI sphere) double-layer coating can serve as an interfacial functional layer to guide uniform Li deposition and inhibit the formation of Li dendrites. Owing to polar groups and regularly arranged PI spheres of the coating layer, Li||PAN/PI@Cu cells display high coulombic efficiency and long cycle life at a current density of 1 mA cm−2, maintaining an average CE of ∼98.5% after 400 cycles with a deposition capacity of 1 mA h cm−2.


1. Introduction

Li metal batteries (LMBs) have been regarded as one of the most promising candidates for next-generation high energy density batteries due to high specific capacity (3861 mA h g−1), low density (0.53 g cm−3) and the lowest electrochemical potential (−3.04 V vs. standard hydrogen electrode) of Li metal.1,2 Batteries based on the Li metal anode can offer high specific energy to movable power systems. For example, Li–S and Li–O2 batteries can improve specific energies to ∼350–650 W h kg−1 and ∼500–950 W h kg−1 which are much higher than those of the present lithium-ion batteries (LIBs) whose specific energy is typically ∼250 W h kg−1. Therefore, the further utilization of metallic Li anodes becomes critical for the rising energy storage market.

However, LMBs have serious problems such as severe safety hazards induced by uncontrollable Li dendrite growth and inferior coulombic efficiency (CE) caused by the highly reactive nature of Li metal.3 It is worth noting that dendrite formation is mainly induced by the inhomogeneous Li nucleation and growth. Multiple factors, such as an uneven surface of the substrate, a large concentration gradient of Li ions and non-uniform surface electronic conductivity,4 may disturb the uniform deposition of Li metal and result in dendrites. In addition, the weak solid electrolyte interphase (SEI) layer is another non-negligible factor. The SEI layer is not tough and flexible enough to prevent dendrite penetration, which causes cracks to expose the fresh Li metal to the electrolyte repeatedly, leading to the continuous formation of a new SEI layer and a large drop in the CE.5 All of these drawbacks mentioned above greatly limit the commercialization of Li metal anodes.

In order to overcome the drawbacks of the Li metal anode, several strategies have been developed to suppress the formation of Li dendrites. Generally, all of these methods can be divided into three categories. The first is controlling the composition, additives or concentration of the electrolyte to regulate the properties of the SEI layer and the Li deposition behavior. For example, halogenated Li salts can significantly enhance the surface mobility of Li ions and the stability of Li deposition.6 Fluoroethylene carbonate additives are used to form a compact and stable LiF-rich SEI layer.7,8 RNO3 salts (R = Li, Na, K, Rb and Cs) are beneficial for strengthening the SEI layer due to products like Li3N. Moreover, the positively charged K+, Rb+ and Cs+ ions can accumulate around the tips of dendrites, forming a positively charged electrostatic shield to suppress further growth of Li dendrites.9,10 The second is structuring an artificial SEI layer to protect the Li metal anode. Artificial protective layers such as 3D Al2O3 networks,11 Mo6S8,12 Cu3N,13 and LLZTO/Li-Nafion14 possessing high mechanical strength, good flexibility, and high Li-ion conductivity can meet the strict requirements of the SEI layer and alleviate dendrite formation.15 The third is using functionalized interlayers to suppress Li dendrites. Given the fact that the rough surface of Li metal and uneven current density distribution can lead to inhomogeneous Li nucleation and Li dendrite growth, it is crucial to obtain a uniform Li-ion flux during deposition. Therefore, interlayers such as polyacrylonitrile (PAN) fibers with a large number of polar groups (C[double bond, length as m-dash]O, C[double bond, length as m-dash]N and so on) can act as efficient lithiophilic sites to guide even Li nucleation and growth.16–20 The microsphere layer is another type of functionalized interlayer that can form nano-scaled pores to improve the uniformity of Li deposition.5,21 Besides the above-mentioned strategies, modifying the surface conductivity and lithiophilicity of the anode can also affect Li deposition.22,23 According to the methods mentioned above, it can be concluded that the microsphere layer with polar groups and lone pairs of electrons can effectively lead to uniform Li deposition via nano-scaled pores and the adsorption force.

Herein, we demonstrated a polyacrylonitrile (PAN) fiber/polyimide (PI) sphere double-layer coating double-layer coating on a Cu substrate (PAN/PI@Cu) to regulate Li deposition behavior. PI containing imide polar groups (–CO–N–CO–) are one of the most comprehensive organic polymer materials in LMBs.24,25 We propose that the first layer (PI sphere), with dense and uniform pores, can guide uniform Li-ion diffusion and Li deposition. Besides, we introduced the second layer (PAN fiber) to protect the structure of the PI sphere and act as a buffer layer to distribute Li ions preliminarily.18 Unlike the bare Cu substrate, PAN/PI@Cu guides homogeneous Li deposition, as shown in Fig. 1. After binding with PAN fibers, the structure of the PI sphere layer in PAN/PI@Cu develops more endurance and will not be destroyed and loosened easily during repeated discharging/charging processes, compared with PI@Cu.


image file: c9ta13678a-f1.tif
Fig. 1 Schematic diagrams show Li deposition behaviors on different substrates. (a) Bare Cu has inhomogeneous Li ion distribution on the surface and induces the growth of Li dendrites; (b) PI@Cu can help redistribute surface Li ions and cause uniform Li deposition in the initial cycles, but the PI sphere layer will be loose and get damaged easily; (c) PAN/PI@Cu can guide Li deposition for a longer life-span, since the fibers can bind the spheres and increase the mechanical strength of the PI sphere layer.

2. Experimental section

Preparation of PAN/PI@Cu

The preparation process is illustrated in Fig. S1. Firstly, the PI dispersion was loaded into a 20 mL plastic injection syringe with a stainless-steel needle. A high voltage of 20 kV was applied during the process of electrospray, and Cu foil was placed 20 cm beneath the needle to obtain PI@Cu. Secondly, the PAN precursor solution was loaded into the same type of plastic injection syringe and the voltage was changed to 15 kV. After electrospinning, the as-prepared PAN/PI@Cu was dried in a vacuum at 75 °C for 24 hours to remove the residual solvent.

Structural characterization

The morphologies of the samples were characterized using a Phenom ProX scanning electron microscope (SEM) operating at 10.0 kV. Energy dispersive spectrometry (EDS) was also carried out using a Phenom ProX operating at 15.0 kV.

Electrochemical measurements

The Li||Cu, Li||PI@Cu, Li||PAN/PI@Cu cells were assembled using Li metal (10.0 mm in diameter) as the counter electrode and 1 M LiTFSI in DOL/DME (1[thin space (1/6-em)]:[thin space (1/6-em)]1 in volume) with 2% LiNO3 as the electrolyte. The cells were monitored in galvanostatic mode using a LAND battery tester at room temperature.

DFT calculations

DFT calculations were carried out using Abinit software package.26–28 Perdew–Burke–Ernzerhof (PBE) generalized gradient approximation (GGA)29 was adopted as the exchange correlation functional, and the projector-augmented-wave30 method was used to describe the electron–ion interaction. The cutoff energy was set to 20 Ha, and the k-point mesh size was set to <0.05 Å−1. All the structures were fully optimized to reach a force tolerance of 0.01 eV Å−1. For Li-ion adsorption on PI, a cube box with a side length of 25 Å was used to build the model system. For Li-ion adsorption on Li metal, {001}31 and {110}32 surfaces were chosen for their lower surface energies.33 Slab models with five layers of Li atoms and a 15 Å vacuum layer were adopted to prevent the interaction between slabs. The adsorption energy of Li atoms on PI and Li metal surfaces was calculated as follows:
Eads = Etot(PI/Li surface + Li) − Etot(PI/Li surface) − Etot(Li)
where Etot(PI/Li surface + Li) is the total energy of PI or Li metal surfaces with the adsorbed Li atom, Etot(PI/Li surface) is the total energy of PI or the Li metal surface, and Etot(Li) is the total energy of a single Li atom. For the charged system with Li ion adsorbed on PI, the adsorption energies were calculated as follows:
Eads = EadsorbedEfaraway
where Eadsorbed is the total energy of the system with Li-ions adsorbed on PI, and Efaraway is the total energy of the system with Li-ions faraway from PI.

3. Results and discussion

The morphologies of PAN/PI@Cu were investigated by SEM. Fig. 2a shows the top-view SEM image of PI spheres in the first layer ranging from 1 μm to 2 μm in diameter. Raw PI powder is yellow in color (Fig. S2a) and double concave disc-like in shape (Fig. S2b). After adding PI to DMF, a brown dispersion is obtained and the dispersion shows Tyndall effect, proving PI exists as particles in DMF (Fig. S2c and d). Double concave disc-like PI powders may be shrunk into micro-spheres (Fig. S2e) using electric field force and sprayed onto the Cu substrate during the electrospray process. The diameter of PAN fibers in the second layer of PAN/PI@Cu is uniform and is less than 500 nm as shown in Fig. 2b. PAN fibers can physically immobilize PI spheres and preliminarily homogenize the concentration of Li ions near the surface of the first layer.18,34 Therefore, PAN fibers can also fetter the continuous consumption of Li metal and electrolyte via hindering the accumulation of Li ions around the cracks of the SEI layer. As shown in Fig. 2c, the thicknesses of the first layer (PI sphere) and the second layer (PAN fiber) are uniform (around 2 μm and 1 μm, respectively), and a double-layer structure can be observed by peeling off the PAN fibers (Fig. 2d). Moreover, Fig. S3 clearly shows an explicit boundary between the two layers of the coating stripped from the Cu substrate which is completely identical to our design.
image file: c9ta13678a-f2.tif
Fig. 2 Morphologies of PAN/PI@Cu. Top-view SEM images of (a) the first layer (PI sphere) and (b) the second layer (PAN fiber); cross-sectional SEM images of (c) PAN/PI@Cu and (d) the first layer.

To investigate the influence of the double-layer coating on overall electrochemical performance, galvanostatic cycling tests of Li‖Cu, Li‖PI@Cu and Li‖PAN/PI@Cu cells were carried out at various current densities with different deposition capacities. Coulumbic efficiency (CE) is considered as an important indicator of cycling stability, representing the capacity ratio of Li striping versus Li plating in each cycle. The improved cycling stability of Li‖PAN/PI@Cu cells can be observed at all the current densities (Fig. 3a and b and S4). As shown in Fig. 3a, the CE of Li‖Cu presents a sharp drop before the 100th cycle at a current density of 1 mA cm−2 with a deposition capacity of 1 mA h cm−2, and eventually decreases to 52.8% after 180 cycles. After modifying bare Cu with the PI sphere layer, the Li‖PI@Cu can even maintain a CE of ∼98.72% for 200 cycles. Unfortunately, the CE of Li‖PI@Cu becomes unstable and shows a significant drop after 300 cycles, which is most likely caused by PI sphere shedding and layer structural damage. By contrast, a CE of higher than 98% is achieved for 500 cycles, proving that Li‖PAN/PI@Cu has a steady CE and shows better cycling stability than Li‖Cu and Li‖PI@Cu. When increasing the current density to 2 mA cm−2 and the deposition capacity to 2 mA h cm−2, Li‖PAN/PI@Cu still delivers a high CE of 97.3% for 130 cycles (Fig. 3b). EIS curves and corresponding equivalent resistances of Li‖Cu and Li‖PAN/PI@Cu after the 1st, 10th and 25th cycles are displayed in Fig. S5.Rct of Li‖Cu decreases rapidly upon cycling, which is attributed to the increase in the active surface area resulting from the formation of Li dendrites. While Rct of Li‖PAN/PI@Cu decreases much more slowly indicating that with a PAN/PI double-layer coating Li dendrite growth is prohibited and the SEI layer is more stable. For Li‖PAN/PI@Cu, even longer cycles with high CEs can be achieved if lower current density (0.5 mA cm−2) is applied, 98% for 600 cycles and 98.5% for 300 cycles with deposition capacities of 1 mA h cm−2 and 2 mA h cm−2, respectively (Fig. S4). In order to tightly bind PI spheres to the Cu substrate and prevent them from falling off, we electrospun an extremely thin PAN fiber layer onto the PI sphere layer. The performance is not good when using a PAN layer to modify the Cu substrate alone. As shown in Fig. S6, Li‖PAN@Cu exhibits an even shorter lifespan than Li‖Cu, probably due to the reaction between PAN fibers with electrolyte and Li metal. We suggest that the decrease in CE is attributed to tip growth of Li on the untreated substrate and breakage of the SEI layer. In Li‖Cu, after the SEI layer cracks, the newly deposited Li metal forms Li dendrites and consumes electrolyte very fast. The higher CE of Li‖PAN/PI@Cu indicates homogeneous Li nucleation, smooth Li plating and corresponding less SEI layer breakage, which can be attributed to the large number of polar groups of the double-layer coating.


image file: c9ta13678a-f3.tif
Fig. 3 Coulumbic efficiencies of Li‖Cu, Li‖PI@Cu and Li‖PAN/PI@Cu cells: (a) at 1 mA cm−2, and 1 mA h cm−2 and (b) at 2 mA cm−2, and 2 mA h cm−2. (c) SEM images and EDS mappings: PAN/PI@Cu with (i–iii) Li deposited and (iv–vi) Li stripped, respectively; bare Cu with (vii–ix) Li deposited and (x–xii) Li stripped, respectively. The Li deposition capacity is 2 mA h cm−2.

Compared with Li‖Cu, Li‖PAN/PI@Cu cells maintain almost the same ΔVs between plating and stripping plateaus (Fig. S7) and unchanged capacities upon cycling (Fig. S8), suggesting an extraordinarily stable interphase between the anode and electrolyte. For example, the ΔVs of Li‖PAN/PI@Cu for the 100th cycle at a current density of 1 mA cm−2 with a deposition capacity of 1 mA h cm−2 is about 34 mV and there is essentially no change during the following cycles. By contrast, the ΔVs of Li‖Cu under the same conditions exhibits a gradual increase from ∼34 mV for the 100th cycle to ∼70 mV for the 500th cycle, higher than that of Li‖PAN/PI@Cu. Besides, at the same discharge time (deposited same amount of Li) at different cycles, the voltage hysteresis values of Li‖PAN/PI@Cu are steadier than that of Li‖Cu (Fig. S9), which also proves the presence of a more stable electrolyte/electrode interphase in Li‖PAN/PI@Cu. Additionally, the ΔVs of Li‖PAN/PI@Cu at a current density of 1 mA cm−2 (Fig. S7) is slightly higher in the initial period but lower in the later stage than that of Li‖Cu. This is reasonable because Li ions need to diffuse through the double-layer polymer coating to deposit on the Cu surface, thus improving impedance. Owing to uniform and dendrite-free Li deposition in Li‖PAN/PI@Cu, the polarization maintains its the initial level, but the ΔVs of Li‖Cu increases remarkably upon cycling. Moreover, the stripping cycle capacity loss of Li‖Cu is larger than that of Li‖PAN/PI@Cu (Fig. S8), reflecting severe irreversible Li consumption in Li‖Cu.

To further reveal the effect of double-layer coating on the electrochemical performance, we examined the morphologies of the deposited Li on bare Cu and PAN/PI@Cu after the 5th plating or 5th stripping cycle at a current density of 1 mA cm−2 with a deposition capacity of 2 mA h cm−2. The SEM images are shown in Fig. 3c. Undoubtedly, homogeneous Li-ion diffusion can regulate the Li plating/stripping behavior. Fig. 3c(ii) shows that the plated Li on PAN/PI@Cu exhibits a dense morphology, whereas on the bare Cu electrode it shows a loose and porous structure, including Li dendrites (Fig. 3c(viii)). The elements on the surface of the electrode were then scanned via an energy dispersive spectrometer (EDS), as shown in Fig. 3c(iii, vi, ix and xii). It is clear that the blue areas in the cross section of PAN/PI@Cu, representing Li metal or Li oxides, disappear completely after the stripping cycle. However, the top-view and cross-sectional elemental map of the bare Cu electrode show that Li metal (green areas) cannott be completely stripped from the Cu substrate after stripping. Moreover, we also observed the morphology of the deposited Li after the 1st plating cycle at a current density of 1 mA cm−2 with a deposition capacity of 1 mA h cm−2, as shown in Fig. S10. Clearly, the double-layer coating can distribute Li ions uniformly and inhibit the growth of Li dendrites without deformation. In sharp contrast, fast growth of Li dendrites on the bare Cu electrode is observed. It can be seen from Fig. S11 that for even up to 30 cycles, the double-layer structure remains unchanged and maintains its function to guide smooth Li deposition.

Interactions between Li atoms/Li ions and PI with polar functional groups were evaluated through density functional theory (DFT) calculations. Fig. 4a shows the chemical formula and structural model of polyimide. In DFT calculations, a unit of PI with both ends saturated with hydrogen atoms was taken to construct the adsorption models. Fig. 4b and S12 show the studied adsorption sites of PI and the detailed adsorption models of Li atoms and Li ions on PI, respectively. Fig. 4c shows the comparison of the adsorption energies of Li ions and Li atoms on these chosen adsorption sites. Obviously, the adsorption energies of Li ions on all these sites are more negative than those of Li atoms, indicating that PI can adsorb Li ions easily and form a Li-ion-rich layer on the surface of the Cu substrate. The value of adsorption energies for Li atoms on Li metal surfaces is also shown in Fig. 4c and is more negative than that of Li atoms on PI. The adsorption models of Li atoms on the Li metal surface is shown in Fig. S13. The atomic absorption spectroscopy test (Table S1) does prove that PI has a certain adsorption effect on Li ions, which agrees with the above calculated result. We speculate that PI has the function of adsorbing Li ions and evenly distributing them on the surface of the Cu electrode. The proposed dendrite-inhibiting mechanism is shown in Fig. 4d. The lone pair of electrons of oxygen in the oxygen-containing functional groups (C[double bond, length as m-dash]O) of PI show strong attraction towards Li ions, and can uniformly adsorb Li ions on the surface of PI spheres. Once Li ions obtain electrons and turn into Li atoms, the adsorption between PI and Li atoms becomes very weak. Consequently, Li atoms tend to deposit uniformly on the surface of the electrode, leaving vacant adsorption sites. In this way, as a Li-ion transfer station with high Li-ion concentration, PI layer can implement uniform Li-ion distribution and dendrite-free deposition on the surface of the electrode.


image file: c9ta13678a-f4.tif
Fig. 4 (a) The chemical formula and molecular structure model of PI, (b) seven selected adsorption sites on PI, (c) comparison of adsorption energies for Li ions and Li atoms in the selected sites of PI and for Li atoms on the deposited Li metal and (d) the schematic diagram of the dendrite-inhibiting mechanism.

When a higher current density of 3 mA cm−2 is applied, the Li‖PAN/PI@Cu cell failed quickly after 60 cycles with a deposition capacity of 2 mA h cm−2, as shown in Fig. 5a. To investigate the reason for failure, we disassembled the cell in a glove-box and examined the morphologies of the PI sphere layer and separator. It can be clearly observed in Fig. 5b that the structure of the first layer has been severely damaged, aggravating uneven Li deposition and resulting in the destruction of the SEI layer. Many scallops appear on the PE separator surface (Fig. 5c), indicating that the separator may undergo a significant deformation while cycling at a high current density and capacity. It is clearly seen that some Li metal has also grown on the separator surface. The unwanted Li metal on the surface finally pierces the separator, causing inner short circuit and a CE of zero, as revealed in Fig. 5a. We believe that by further modifying the micro-structure of the coating layer, a better anode for Li metal for use at high current density can be achieved.


image file: c9ta13678a-f5.tif
Fig. 5 (a) Coulumbic efficiency of Li‖PAN/PI@Cu; SEM images of (b) cross-section of the PI sphere layer and (c) top-view of the PE separator.

4. Conclusion

In summary, we demonstrated a novel design of a PAN fiber/PI sphere coating via the cheap and facile electrospinning method to achieve a stable cycling performance of Li metal. This PAN fiber/PI sphere coating with a large number of polar groups enables homogeneous deposition of Li metal and effectively suppresses the growth of dendrites. Besides, the double-layer coating also satisfies many other requirements, such as ultralight mass, good mechanical and thermal stability and high flexibility. Moreover, an average CE of higher than 97.8% for 400 and 200 cycles were achieved at a current density of 1 mA cm−2 with deposition capacities of 0.5 mA h cm−2 and 1 mA h cm−2, respectively. The utilization of the double-layer coating contributes to improved safety and performance, accelerating the industrialization of Li metal anodes towards next generation high energy density battery systems.

Conflicts of interest

The authors declare no conflict of interest.

Acknowledgements

This work was supported by the National Key R&D Program of China (Grant No. 2018YFB0104300) and National Natural Science Foundation of China (Grant No. 51772241). X. Han would like to thank the Independent Research Project of the State Key Laboratory of Electrical Insulation and Power Equipment (Grant No. EIPE19111) for the financial support. F. Shen acknowledges the State Key Laboratory of Electrical Insulation and Power Equipment for financial support. Thanks to Guilin Electrical Equipment Scientific Research Institute for providing PI materials.

References

  1. D. Lin, Y. Liu and Y. Cui, Nat. Nanotechnol., 2017, 12, 194–206 CrossRef CAS PubMed.
  2. X. B. Cheng, R. Zhang, C. Z. Zhao and Q. Zhang, Chem. Rev., 2017, 117, 10403–10473 CrossRef CAS PubMed.
  3. X. Zhang, A. Wang, X. Liu and J. Luo, Acc. Chem. Res., 2019, 52, 3223–3232 CrossRef CAS PubMed.
  4. Q. Li, H. Pan, W. Li, Y. Wang, J. Wang, J. Zheng, X. Yu, H. Li and L. Chen, ACS Energy Lett., 2018, 3, 2259–2266 CrossRef CAS.
  5. W. Liu, W. Li, D. Zhuo, G. Zheng, Z. Lu, K. Liu and Y. Cui, ACS Cent. Sci., 2017, 3, 135–140 CrossRef CAS PubMed.
  6. J. Zheng, M. H. Engelhard, D. Mei, S. Jiao, B. J. Polzin, J.-G. Zhang and W. Xu, Nat. Energy, 2017, 2, 17012 CrossRef CAS.
  7. X.-Q. Zhang, X.-B. Cheng, X. Chen, C. Yan and Q. Zhang, Adv. Funct. Mater., 2017, 27, 1605989 CrossRef.
  8. X. Fan, L. Chen, O. Borodin, X. Ji, J. Chen, S. Hou, T. Deng, J. Zheng, C. Yang, S.-C. Liou, K. Amine, K. Xu and C. Wang, Nat. Nanotechnol., 2018, 13, 715–722 CrossRef CAS PubMed.
  9. Y. Zhang, J. Qian, W. Xu, S. M. Russell, X. Chen, E. Nasybulin, P. Bhattacharya, M. H. Engelhard, D. Mei, R. Cao, F. Ding, A. V. Cresce, K. Xu and J. G. Zhang, Nano Lett., 2014, 14, 6889–6896 CrossRef CAS PubMed.
  10. F. Ding, W. Xu, X. Chen, J. Zhang, Y. Shao, M. H. Engelhard, Y. Zhang, T. A. Blake, G. L. Graff, X. Liu and J.-G. Zhang, J. Phys. Chem. C, 2014, 118, 4043–4049 CrossRef CAS.
  11. R. Tian, X. Feng, H. Duan, P. Zhang, H. Li, H. Liu and L. Gao, ChemSusChem, 2018, 11, 3243–3252 CrossRef CAS PubMed.
  12. K. Lu, S. Gao, R. J. Dick, Z. Sattar and Y. Cheng, J. Mater. Chem. A, 2019, 7, 6038–6044 RSC.
  13. Y. Liu, D. Lin, P. Y. Yuen, K. Liu, J. Xie, R. H. Dauskardt and Y. Cui, Adv. Mater., 2017, 29, 1605531 CrossRef PubMed.
  14. R. Xu, Y. Xiao, R. Zhang, X.-B. Cheng, C.-Z. Zhao, X.-Q. Zhang, C. Yan, Q. Zhang and J.-Q. Huang, Adv. Mater., 2019, 31, 1808392 CrossRef PubMed.
  15. R. Xu, X.-B. Cheng, C. Yan, X.-Q. Zhang, Y. Xiao, C.-Z. Zhao, J.-Q. Huang and Q. Zhang, Matter, 2019, 1, 317–344 CrossRef.
  16. G. Wang, X. Xiong, Z. Lin, J. Zheng, Z. Fenghua, Y. Li, Y. Liu, C. Yang, Y. Tang and M. Liu, Nanoscale, 2018, 10, 10018–10024 RSC.
  17. L. Fan, H. L. Zhuang, W. Zhang, Y. Fu, Z. Liao and Y. Lu, Adv. Energy Mater., 2018, 8, 1703360 CrossRef.
  18. J. Lang, J. Song, L. Qi, Y. Luo, X. Luo and H. Wu, ACS Appl. Mater. Interfaces, 2017, 9, 10360–10365 CrossRef CAS PubMed.
  19. C.-H. Chang, S.-H. Chung and A. Manthiram, Adv. Sustainable Syst., 2017, 1, 1600034 CrossRef.
  20. X. B. Cheng, T. Z. Hou, R. Zhang, H. J. Peng, C. Z. Zhao, J. Q. Huang and Q. Zhang, Adv. Mater., 2016, 28, 2888–2895 CrossRef CAS PubMed.
  21. B. Zhu, Y. Jin, X. Hu, Q. Zheng, S. Zhang, Q. Wang and J. Zhu, Adv. Mater., 2017, 29, 1603755 CrossRef PubMed.
  22. Q. Zhang, J. Luan, Y. Tang, X. Ji, S. Wang and H. Wang, J. Mater. Chem. A, 2018, 6, 18444–18448 RSC.
  23. H. Zhang, X. Liao, Y. Guan, Y. Xiang, M. Li, W. Zhang, X. Zhu, H. Ming, L. Lu, J. Qiu, Y. Huang, G. Cao, Y. Yang, L. Mai, Y. Zhao and H. Zhang, Nat. Commun., 2018, 9, 3729 CrossRef PubMed.
  24. Y.-E. Miao, G.-N. Zhu, H. Hou, Y.-Y. Xia and T. Liu, J. Power Sources, 2013, 226, 82–86 CrossRef CAS.
  25. D. Chen, T. Liu, X. Zhou, W. C. Tjiu and H. Hou, J. Phys. Chem. B, 2009, 113, 9741–9748 CrossRef CAS PubMed.
  26. X. Gonze, J.-M. Beuken, R. Caracas, F. Detraux, M. Fuchs, G.-M. Rignanese, L. Sindic, M. Verstraete, G. Zerah, F. Jollet, M. Torrent, A. Roy, M. Mikami, P. Ghosez, J.-Y. Raty and D. C. Allan, Comput. Mater. Sci., 2002, 25, 478–492 CrossRef.
  27. X. Gonze, G. M. Rignanese, M. Verstraete, J. M. Beuken, Y. Pouillon, R. Caracas, F. Jollet, M. Torrent, G. Zerah, M. Mikami, P. Ghosez, M. Veithen, J. Y. Raty, V. Olevano, F. Bruneval, L. Reining, R. Godby, G. Onida, D. R. Hamann and D. C. Allan, Z. Kristallogr., 2005, 220, 558–562 CAS.
  28. X. Gonze, B. Amadon, P. M. Anglade, J. M. Beuken, F. Bottin, P. Boulanger, F. Bruneval, D. Caliste, R. Caracas, M. Côté, T. Deutsch, L. Genovese, P. Ghosez, M. Giantomassi, S. Goedecker, D. R. Hamann, P. Hermet, F. Jollet, G. Jomard, S. Leroux, M. Mancini, S. Mazevet, M. J. T. Oliveira, G. Onida, Y. Pouillon, T. Rangel, G. M. Rignanese, D. Sangalli, R. Shaltaf, M. Torrent, M. J. Verstraete, G. Zerah and J. W. Zwanziger, Comput. Phys. Commun., 2009, 180, 2582–2615 CrossRef CAS.
  29. J. P. Perdew, M. Ernzerhof and K. Burke, J. Chem. Phys., 1996, 105, 9982–9985 CrossRef CAS.
  30. P. E. Blochl, Phys. Rev. B, 1994, 50, 17953–17979 CrossRef PubMed.
  31. Y. Chen, X. Dou, K. Wang and Y. Han, Adv. Energy Mater., 2019, 1900019 CrossRef.
  32. R. Zhang, X. B. Cheng, C. Z. Zhao, H. J. Peng, J. L. Shi, J. Q. Huang, J. Wang, F. Wei and Q. Zhang, Adv. Mater., 2016, 28, 2155–2162 CrossRef CAS PubMed.
  33. L. Shi, A. Xu and T. Zhao, ACS Appl. Mater. Interfaces, 2017, 9, 1987–1994 CrossRef CAS PubMed.
  34. Z. Liang, G. Zheng, C. Liu, N. Liu, W. Li, K. Yan, H. Yao, P. C. Hsu, S. Chu and Y. Cui, Nano Lett., 2015, 15, 2910–2916 CrossRef CAS PubMed.

Footnotes

Electronic supplementary information (ESI) available. See DOI: 10.1039/c9ta13678a
The authors contributed equally to this paper.

This journal is © The Royal Society of Chemistry 2020