Skip to main content
NCBI home page
iScience logoLink to iScience
. 2020 Oct 1;23(10):101636. doi: 10.1016/j.isci.2020.101636

Efficient Construction of a C60 Interlayer for Mechanically Robust, Dendrite-free, and Ultrastable Solid-State Batteries

Zhenlong Li 1,6, Siwei Zhang 1,6, Kun Qian 2,∗, Pengbo Nie 1, Shuxiao Chen 1, Xuan Zhang 1, Baohua Li 3, Tao Li 2,5, Guodan Wei 1,7,∗∗, Feiyu Kang 1,3,4
PMCID: PMC7569341  PMID: 33103075

Summary

Interfacial instability between solid electrolytes (SEs) and lithium metal remains a daunting challenge for solid-sate batteries. Here, a conformal C60 interlayer is efficiently constructed on Li1.5Al0.5Ge1.5(PO4)3 (LAGP) SEs by physical vapor deposition, and an ideal interfacial contact is achieved via forming an ionically conducting matrix of LixC60 with lithium metal. The obtained LixC60 is beneficial to hinder the growth of lithium dendrites at interface and release the local stress during the lithiation and delithiation. As a result, the Li/LAGP-C60/Li symmetric cells demonstrate ultra-stable cycling performance for more than 4,500 h at a current density of 0.034 mA cm−2. The Li/LAGP-C60/LiFePO4 full cells deliver a reversible capacity of 152.4 mAh g−1 at room temperature, and the capacity retention rate is 85% after more than 100 cycles. This work provides a feasible and scalable strategy to improve the SEs/Li interface for high-performance solid-state batteries.

Subject Areas: Energy Storage, Energy Systems, Energy Materials

Graphical Abstract

graphic file with name fx1.jpg

Open in a new tab

Highlights

  • •

    Ionically conducting LixC60 matrix is formed at Li-LAGP interface

  • •

    Li/LAGP-C60/Li cells display ultra-long cycle life of 6 months

  • •

    Li/LAGP-C60/LFP cells exhibit high capacity with good cycle performance

  • •

    Mechanical integrity of cycled LAGP-C60 is validated by X-ray CT

Energy Storage; Energy Systems; Energy Materials

Introduction

Solid electrolytes (SEs) are promising to address the problem of flammability in traditional liquid electrolytes, offering a forward-looking solution for safe and high-energy-density batteries (Bachman et al., 2016; Gao et al., 2018). However, the challenge of cycling failure in solid-state batteries (SSBs) severely restricts their further practical applications, which is predominantly caused by the degradation of solid-solid interfaces (Hao et al., 2019; Luntz et al., 2015). In fact, many studies have shown that lithium dendrites are formed in various types of SEs, such as Na super ion conductor phase (NASICON), e.g., Li1.5Al0.5Ge1.5(PO4)3 (LAGP) (Hou et al., 2018) and Li1.3Al0.3Ti1.7(PO4)3 (LATP) (Hao et al., 2019); garnet, e.g., Li7La3Zr2O12 (LLZO) (Krauskopf et al., 2019); and Li2S-P2S5 electrolytes (Han et al., 2018; Porz et al., 2017). Even in the very dense garnet-type SEs, the lithium dendrites are able to grow along the grain boundaries and voids (Krauskopf et al., 2019). NASICON-type LAGP SEs have outstanding merits, such as high ionic conductivity (10−4–10−3 S cm−1), stability with moisture, and wide electrochemical windows (Feng et al., 2010; Wang et al., 2019). However, the interface degradation is more complicated than garnet, not only because of the formation of dendrites but also because of the poor chemical compatibility caused by Ge4+ reduction to Ge2+ or even Ge0 (Hartmann et al., 2013; Lewis et al., 2019). The expendable electronic conducting interphase (MCI) has been observed at Li/SEs interface, which comprises a stoichiometrically changed LAGP and lithium oxide compounds, accompanied by widening microcracks and pulverization, eventually resulting in cell failure (Chung, and Kang, 2017). Tippens et al. (2019) in situ observed the growth of the crack network within LAGP during cycling by X-ray computed tomography (CT), which proved that the extent of fracture strongly correlates with increases in impedance.

The interlayer strategy is designed to address these tough interfacial problems in terms of interface protection, interfacial resistance reducing, lithium wettability enhancement, and cycling stability. A typical approach is to introduce an ion-conducting and passivated interlayer between the metallic lithium and SEs, such as polymer-matrix membranes (Zhang et al., 2017), carbon materials (Shao et al., 2018; Feng et al., 2020; Duan et al., 2019), and various types of metal or metal oxide/nitride thin films (Liu et al., 2018a, 2018b; Cheng et al., 2019; Fu et al., 2017), providing good interfacial compatibility via isolating metallic lithium from SEs. Various preparation methods have been developed for functional interlayer materials. For example, solution-based methods were widely used for polymer-matrix interlayers (Peng et al., 2017; He et al., 2019; Yu et al., 2019; Liu et al., 2020). Sputtering methods, such as magnetron sputtering and etching sputtering, were employed for metal, metal oxides/nitrides interlayers of ZnO (Hao et al., 2019), Cr (Cortes et al., 2019), Ge (Liu et al., 2018b), Cu3N (Huo et al., 2020); Polishing methods were reported for removing the surface impurities and simultaneously coating materials, such as MoS2 (Fu et al., 2019) and Si (Wu et al., 2018); Moreover, nano-structure ZnO (Wang et al., 2017), Al2O3 (Han et al., 2017) interlayers were successfully achieved on SEs by atomic layer deposition (ALD) method. However, challenges remain since the current coating methods are either involved in complex multi-step chemical-physical processes or are only applicable to flat SE pellets. Hence, suitable interlayer materials and compatible efficient fabrication technology are still being explored to improve process efficiency and scalability.

Fullerene (C60) is well known for its spheroidal geometry and unique optoelectronic properties and is theoretically electrically insulating but exhibits high electron affinity of 2.65 eV and excellent structure stability (Wang et al., 2012; Haddon, 2002). It can be adopted as a superior SE interlayer material because of its highly reversible electrochemical reaction with metallic lithium and the mechanically soft features to release the compressive stress. More importantly, dense and homogeneous C60 layers can be efficiently constructed on SEs surfaces in a controllable manner by one-step physical vapor deposition (PVD). Herein, we have demonstrated that the as-deposited C60 molecules can effectively fill the microscopic pores on the LAGP surface to achieve ideal interfacial contact and form ionically conducting matrix of LixC60. The obtained LixC60 exhibits high ionic conductivity but negligible electronic conductivity, which is beneficial to hinder the growth of lithium dendrites at interface and release the local stress during the lithiation and delithiation. As a result, the interfacial impedance of Li-Li symmetric cell is significantly reduced by 15 times. And the cell stably cycles for more than 4,500 h at a current density of 0.034 mA cm−2 and over 1,800 h at 0.1 mA cm−2. The Li/LAGP-C60/LiFePO4 cells also obtain a reversible capacity of 152.4 mAh g−1 at room temperature with a capacity retention rate of 85% after 100 cycles. By monitoring the mechanical and chemical changes after electrochemical cycling via CT, XPS depth analysis, and SEM, it is found that the C60 interlayer successfully suppresses the reduction of Ge4+, the propagation of fracture, and even the formation of dendrites. This work provides a controllable, scalable, and efficient technology to improve the interface stability, hinder the growth of lithium dendrites, and maintain the mechanical integrity of SEs, which may boost the commercial applications of high-performance solid-state batteries.

Results and Discussion

Figure 1A demonstrates the working process of constructing C60 interlayer by physical vapor deposition. The C60 power is placed on the evaporator at the bottom of a vacuum chamber, while the SEs were mounted on a top rotating table using a custom-designed mask. The free C60 molecules are thermally evaporated and deposited on the surface of SEs, resulting in a conformal layer of C60 thin film in nanoscale (Figure 1B). Notably, the physical vapor deposition method is highly efficient on the interlayer construction and there is no special demanding on the geometries of SEs. In our case, a 6 × 6 mask, as shown in Figure 2A, was employed for batch deposition and 36 LAGP pellets were coated with C60 thin films in 10–15 min. Figure 2B displays optical images of an LAGP pellet before and after batch deposition, in which we can see the color of the surface changed from white to light yellow. A polished reference LAGP pellet was placed with other unpolished LAGP samples for calibrating the actual thickness coating layer. Figures 2C and 2D show the surface morphology of the reference LAGP and its partially enlarged view, respectively. As can be seen, the C60 thin films have excellent conformality and uniformity.

Figure 1.

Figure 1

Open in a new tab

Schematic Diagram of the Construction of C60 Interlayer

(A) Schematic diagram of the physical vapor deposition process.

(B) C60 interlayer between Li metal anode and SEs.

Figure 2.

Figure 2

Open in a new tab

Characterization of Deposited C60 Interlayer on LAGP Pellet

(A) Optical image of a 6 × 6 mask for batch deposition.

(B) Optical images of an uncoated LAGP pellet and a coated one in a batch deposition.

(C–E) SEM images of a reference LAGP pellet with C60 coating layer; (C) the surface morphology and (D) its partial enlarged view, and (E) its cross-sectional image.

(F–H) SEM images of an unpolished LAGP pellet; (F) the uncoated surface, (G) the coated surface and (H) the cross-sectional image.

(I) Raman spectra of LAGP and LAGP-C60.

(J and K) XPS results of LAGP and LAGP-C60; (J) C 1s spectra and (K) Ge 3d spectra.

In the cross-sectional view (Figure 2E), the thin C60 layer can be distinguished and measured by SEM. The unpolished LAGP pellets exhibited very rough surfaces as shown in Figure 2F, which could result in poor contact after it is assembled with lithium metal anode without interface layer protection. Previous study has proved that polishing the surface can improve the contact and lower down the interfacial resistance (Wang et al., 2018a), but it has inevitably complexed its assembly process since it usually takes extremely time-consuming efforts for preparation, not to mention its inconsistency between pellet to pellet. Herein, the LAGP-C60 pellets do not need any further surface treatment, significantly improving their assembly efficiency. Figure 2G shows the SEM image after C60 coating; the LAGP surface has a conformal coverage. The C60 molecules are able to homogeneously fill the edges and pores of the LAGP layer in Figure 2G, and the rough LAGP surface has been effectively smoothed out from the cross-sectional view in Figure 2H, demonstrating excellent conformal coating of C60 film. Raman and XPS spectra of the bare LAGP and LAGP-C60 samples are compared in Figures 2I–2K. New Raman bands appeared in the LAGP-C60 sample (272, 496, 706, 773, 1,420, 1,466, 1,560 cm−1), which are consistent with previously reported Raman scattering results in C60 (Kuzmany et al., 1994). Also, the XPS C1s peak at 284.7 eV increased while the Ge 3d peak disappeared in LAGP-C60 sample, which together confirmed that C60 layers were constructed and fully covered on LAGP surface.

DFT simulations were carried out to gain a deep understanding of the reaction between C60 and lithium metal. Figure 3A displays the electronic structure calculation results for C60 and five LixC60 compounds (x = 1, 3, 4, 6, and 12). The calculated HOMO-LUMO energy gap is 2.5 eV for C60, which is too deep to get free electrons excited effectively, indicating negligible conductivity of pristine C60. In the matrix of LixC60, Li ions prefer to locate the octahedral and tetrahedral sites in the gap of densely packed C60 molecules (Figure S2). With the injection of Li ions into C60 molecules, the calculated LUMO-HOMO gap decreases. Even though the gap values are above 1.3 eV, which indicates that the LixC60 compounds have better electron transport ability than C60, nevertheless, the electronic conductivity is still very limited. The previous literature indicated that C60 can accept up to 12 Li ions and is electrochemically reversible (Allemand et al., 1991; Chabre et al., 1992). Here, we also calculated formation energy of Li with C60 as shown in Figure 3B. The negative value of the formation energy means that the chemical reaction is thermodynamically feasible, whereas a larger absolute value represents the reaction is easier to initiate. It can be seen that, as the Li content in LixC60 increases, the absolute value of formation energy increases, which suggests that Li metal can spontaneously react with C60 to form compounds with high lithium content.

Figure 3.

Figure 3

Open in a new tab

Electrochemical Behavior between Li Metal and C60 Interlayer

(A) DFT calculation of energy levels for C60 and LixC60 (x = 1, 3, 4, 6, 12).

(B) DFT calculation of formation energy for LixC60 (x = 1, 3, 4, 6, 12).

(C) The cyclic voltammetry results of the Li/LAGP/Li and Li/LAGP-C60/Li cells.

(D) The Nyquist plots of the EIS measurements on the Li/LAGP/Li and Li/LAGP-C60/Li cells.

Li-Li symmetric cells were assembled to evaluate the electrochemical properties with the introduction of C60 layer. Figure 3C compares the cyclic voltammogram of the Li/LAGP/Li and Li/LAGP-C60/Li cells. In the redox cycle for bare LAGP, no reduction peaks were observed. In contrast, for C60-coated LAGP, a pair of well-defined redox peaks appeared at around 1.0 V versus Li/Li+. The reversible peaks are ascribed to the insertion reaction of LixC60, where x = 4, based on a previous report. (Chabre et al., 1992) The CV results also suggest that the x in LixC60 is not static but fluctuates with potential during the electrochemical processes, proving the effective Li+ transport in the interlayer. Therefore, it is LixC60, not C60, that played as the actual interlayer, and the actual anode instead of the lithium metal. In the calculated Li4C60 structure, lithium atoms occupied both octahedral and tetrahedral sites (Figure S2). The center-to-center distance between neighboring C60 molecules is around 14.1 Å, whereas the calculated C60 diameter is 7.1 Å (Stephens et al., 1991; Heiney et al., 1991). Considering such a large space of 7 Å and the radius of lithium ion is only 0.76 Å, the distance of Li and C in LixC60 is longer than ordinary chemical bonds and thus their interaction is weak, which suggests that the diffusion of Li ions in the C60 interlayer is excellent.

The fresh symmetric cells were measured by EIS at 25°C to investigate the impedance change before and after coating. The Nyquist plots of the Li/LAGP/Li and Li/LAGP-C60/Li cells are shown in Figure 3D. In EIS results, the cell impedance includes the bulk impedance, grain boundary impedance, and interface impedance. The bulk resistance is determined from the high-frequency x-intercepts, which are negligible for both cells. A distinct arc was seen in middle-frequency area, which represents the overlap of grain boundary and interface impedances. The cell impedance of Li/LAGP/Li cell is as high as 4,739 Ω cm−2, whereas for the Li/LAGP-C60/Li cell, the cell impedance significantly decreased by 15 times to 290 Ω cm−2. The grain boundary impedance is identical for both cells, which is calculated as 52 Ω cm−2 in supporting information (Figure S3). So, the large cell impedance of Li/LAGP/Li cell is mainly originated from the interface impedance, which is likely caused by the poor contact between the rough LAGP surface and lithium metal. C60 and its further reaction product of LixC60 successfully re-constructed the interface and significantly reduced the interface impedance more than one order of magnitude, indicating effectively improved interfacial contact at LAGP-C60 cells.

Lithium plating-stripping measurements were carried out by galvanostatic cycling of the symmetric cells at a current density of 0.034 and 0.1 mA cm−2. The charge-discharge voltage profiles of Li/LAGP/Li and Li/LAGP-C60/Li cells are shown in Figure 4A. For the control cell of bare LAGP, the overpotential value increased rapidly upon cycling as shown in the top-left insert image. Only 40 h (20 cycles) of cycling was maintained at the current density of 0.034 mA cm−2 and the cycle life was even shorter at a higher current density of 0.1 mA cm−2 (top-left insert at Figure 4B). This unstable performance is caused by the continuous and irreversible reaction between LAGP and metallic lithium, which includes the dendrite formation and the reduction of Ge4+.23 In sharp contrast, the Li/LAGP-C60/Li symmetric cell has demonstrated ultrastable cycling performance, which achieved more than 4,500 h of cycling at 0.034 mA cm−2, whereas the overpotential was still lower than 1 V, as shown by the enlarged voltage profiles, indicating a stable and durable interface with effective lithium ion transport (Figure 4A). Even under a higher current density of 0.1 mA cm−2, 1,800 h of cycling was achieved and the cell maintained a small overpotential (Figure 4B). Table 1 compares the cycling performance of symmetrical cells in this work with previously reported results. As can be seen, our LAGP-C60 cells have achieved the longest cycle time so far, which strongly indicates that the LixC60 interlayer could effectively hinder the growth of lithium dendrites. The critical current density (CCD) of the Li/LAGP/Li and Li/LAGP-C60/Li symmetric cells were tested to study the Li dendrite suppression. Owing to the side reactions of Li metal and LAGP, the interface impedance is dramatically increased under high current, suggesting the rapid degradation of the Li-LAGP interface. As a result, the voltage of Li/LAGP/Li cell increased sharply in 4 h (Figure 4E), resulting in cell failure. In sharp contrast, the cyclic curves of Li/LAGP-C60/Li symmetric cell under different current densities are stable and the critical current density is higher than 1.0 mA cm−2 (Figure 4F). These results indicated that the introduction of C60 interlayer is conducive to Li dendrite suppression. The previous literature pointed out that the dendrites formation at the interface is caused by the inhomogeneous distribution of electric field, which induces Li nucleation at local position. (Huo et al., 2020) Here, the perfect contacts of LixC60 with LAGP plus the rapid Li ion migration inside the interlayer has fundamentally reduced chances for Li nucleation and dendrite growth.

Figure 4.

Figure 4

Open in a new tab

The Electrochemical Performance of Li-Li Symmetric Cells and Li-LiFePO4 Full Cells with and without C60 Interlayer

(A and B) Galvanostatic cycling performance of the Li/LAGP/Li and Li/LAGP-C60/Li cells under (A) 0.034 and (B) 0.1 mA cm−2; the insert images are their enlarged view at selected time.

(C) The first six charging and discharging curves of Li/LAGP/LiFePO4 cell and the first six and the 20th, 40th, 60th, 80th, and 100th charging and discharging curves of Li/LAGP-C60/LiFePO4 cell.

(D–F) (D) Cycle performance of Li/LAGP-C60/LiFePO4 cell at 0.1 (C) CCD measurement of Li/LAGP/Li (E) and Li/LAGP-C60/Li (F) symmetric cells.

(G) Rate performance of Li/LAGP-C60/LiFePO4 cell. All the tests were performed at 25°C.

Table 1.

Comparison of the Cycling Stability with Different Interlayer Structures

Type of SEs Layer Methods Cycle Time (h) Current Density (mA cm−2) Temperature Reference
LAGP Ge MS 200 0.3 RT (Liu et al., 2018b)
LAGP Cr MS 850 0.2 RT (Cortes et al., 2019)
Cr & Al2O3 ALD 1,200 0.2
LAGP Polymer 1,100 0.1 60°C (Zhang et al., 2017)
LATP ZnO MS 2,000 0.05 RT (Hao et al., 2019)
1,000 0.2 RT
LATP Al2O3 ALD 600 0.01 RT (Liu et al., 2018a)
LATP Polymer LTCVD 500 0.3 60°C (Cheng et al., 2019)
LLZO Al2O3 ALD 90 0.2 RT (Han et al., 2017)
LLZO Al EBE 41 0.2 RT (Fu et al., 2017)
LLZO Ge EBE 160 0.05 RT (Luo et al., 2017)
LLZO ZnO ALD 50 0.1 RT (Wang et al., 2017)
LLZO Graphite 1,000 0.3 RT (Shao et al., 2018)
LAGP C60 PVD 4,500 0.034 RT This work
1,800 0.1
Open in a new tab

To further verify the Li-ion transport capability across the LAGP/Li interface and the interfacial stability, Li/LiFePO4 full cells were assembled and cycled at room temperature. Figure 4C compares the charge-discharge curves of Li/LAGP/LiFePO4 and Li/LAGP-C60/LiFePO4 cells. Although under the same cycling rate of 0.1 C, the charge-discharge plateau gap of Li/LAGP-C60/LiFePO4 cell remains relatively small after 100 cycles. It is even comparable with the initial cycles of Li/LAGP/LiFePO4 cell, which indicates the small polarization of Li/LAGP-C60/LiFePO4 cell. As shown in Figure 4D, the Li/LAGP-C60/LiFePO4 cell achieved a high initial capacity of 152.4 mAh g−1 and a capacity retention rate of 85% after 100 cycles. Therefore, it can be concluded that Li ions are able to effectively transfer from LixC60 layer to LAGP and finally intercalate into the cathode and vice versa. The coulombic efficiency remains above 98% during 100 cycles, which indicates that the undesired reaction of Li metal and LAGP can be effectively suppressed by introducing a C60 layer, thereby achieving good stability of Li/LAGP-C60/LiFePO4 cell. The rate capability of the Li/LAGP-C60/LiFePO4 cell was measured to investigate the cell performance under high current density. The Li/LAGP-C60/LiFePO4 cell delivered specific discharge capacities of 158., 153.9, 145.5, 141.2 mAh g−1 at 0.1, 0.3, 0.5, and 0.6 C, respectively (Figure 4G). The cell capacity retention is 97.7% when the current density was adjusted back to 0.1 C, presenting a very stable performance even cycled under high current density. These results suggest that the C60 protected LAGP is promising for achieving fast charging in solid-state batteries. X-ray diffraction (XRD) and scanning electron microscope (SEM) measurement were further performed to study the cathodes. There is no obvious change in surface morphology and particle size between the pristine and the cycled LiFePO4 cathodes, whether it is disassembled from the Li/LAGP/LiFePO4 cell or the Li/LAGP-C60/LiFePO4 cell (Figures S5A–S5C). In addition, all the XRD patterns of the pristine and the cycled LiFePO4 cathodes show good structural integrity (Figure S5D). These results indicate that the enhanced battery performance is not contributed by the optimization of the LiFePO4 cathode but by the protection of the Li/LAGP interface.

CT was employed to study the mechanical deformation and internal features of LAGP pellets inside the symmetric cells. CT scanning is able to ex situ detect the defects, voids, and cracks without disassembly. The spatial resolution in our experiment is around 2 μm. Figure 5 displays the top view (XY slices) and side view (XZ slices) of a Li/LAGP/Li cell and the extracted SEs from 3D tomogram, the contrast of CT imaging depends on the atomic number, thickness, and density of materials. In Figure 5A, the small internal shadow describes the contour of a LAGP pellet, whereas the large round shadow comes from the stainless steel sheets, which can be easily confirmed in side view (Figure 5E). The lithium metal is invisible in X-ray because of its low atomic number, which explains the gap between solid-state electrolyte and two pieces of stainless steel sheets (Figure 5E). The Li/LAGP/Li and Li/LAGP-C60/Li cells were scanned before and after cycling at 0.034 mA cm−2. Figures 5B–5D displays one horizontal slice from pellets of the pristine LAGP, the cycled LAGP, and the cycled LAGP-C60, respectively, and Figures 5F–5H show their corresponding side slice images. The pristine LAGP shows a complete structure without internal cracks (Figures 5B and 5F).

Figure 5.

Figure 5

Open in a new tab

Two-Dimensional X-Ray CT Images of SSBs and LAGP Pellets Extracted from the 3D Tomogram

XY slices of (A) a 2032 type coin cell, (B) pristine LAGP pellet, (C) LAGP pellet after 20 cycles, and (D) LAGP-C60 pellet after 150 cycles. XZ slices of (E) the 2032 type coin cell, (F) pristine LAGP pellet, (G) LAGP pellet after 20 cycles, (H) LAGP-C60 pellet after 150 cycles.

Because the CT images for pristine cells with and without C60 are similar, only one representative set of images is displayed here. After lithium plating-stripping cycling, we observed crack network formed inside the bare LAGP pellet (Figures 5C and 5G). In side view of Figure 5G, some of the cracks penetrated the entire pellet and others formed a closed loop, reflecting the pulverization of the LAGP after cycling. The crack formation is driven by the stress residual, which is caused by the growth of lithium dendrites and the propagation of MCI. The pulverization of LAGP destroyed the lithium transport channels and well explains the rapid growth of overpotential during lithium plating tests. The LAGP-C60 cells have well maintained the mechanical integrality without internal cracks, as shown in Figures 5D and 5H, which is consistent with the ultrastable cycling performance as shown in Figure 4. Apparently, it owes to the protection of LixC60 interlayer, and its mechanism can be considered as the following three aspects. First, the LixC60 layer effectively suppressed the reaction of LAGP with metallic lithium, thereby preventing the formation of MCI. Second, the LixC60 layer eliminated the growth of lithium dendrites at the interface, which hindered the propagation of MCI layer. Third, the softness characteristic of C60 and LixC60 layer is beneficial to alleviate the residual stress and avoid the initiation of cracks and lithium dendrites (Wang et al., 2018b), which is different from the previously reported metal, oxides, or nitrides interlayers that usually have rigid interfacial contact. The mechanical properties of LAGP and LAGP-C60 surface were performed by atomic force microscopy (AFM) in peakforce tapping mode to verify the softness of C60. To meet the roughness requirement of AFM testing, the LAGP pellet surface was carefully polished before C60 coating and AFM testing. The AFM height maps show the dense and smooth surface of both LAGP (Figure S6A) and LAGP-C60 (Figure S6C). The mean Young's modulus of the LAGP surface (Figure S6B) is 1,128 MPa. This value is about 20-fold higher than the mean Young's modulus of the LAGP-C60 surface (Figure S6D), which is 57.3 MPa. It proves that the C60-coated surface is very soft, which can release the local stress during the lithiation and delithiation, avoid dendrites and cracks formation, thereby achieving a mechanically robust interface.

The cycled symmetric cells and a fresh control cell were then disassembled in the glove box, and the lithium electrodes were stripped to expose the surfaces of SEs. XPS and SEM characterization were carried out to examine the interfacial chemical states and morphology. All the samples were transferred using an air-tight transfer box to avoid contact and reaction with O2 or H2O. Figures 6A and 6B compare the valence change of Ge element in the pristine LAGP, the cycled LAGP, and the cycled LAGP-C60. A sharp peak at 32.9 eV was observed in the XPS Ge 3d spectrum of pristine LAGP, which corresponds to Ge4+ (Figure 6A). For cycled LAGP, a new peak appeared at 31.2 eV with low intensity (Figure 6B), which is ascribed to Ge2+, proving that metallic lithium reduced Ge4+ to Ge2+. The cycled LAGP-C60 shows very weak Ge 3d signal on surface because of the coverage LixC60 layer. We etched the surface to different depths with Ar+ beam and collected the Ge 3d spectra every 300 s Figure 6C shows the XPS spectra during 1,800 s of etching. With the removal of the snugly ensconced LixC60 layer, an increase of Ge 3d peak can be seen with the sputtering time, whereas the corresponding C 1s peak decreases as shown in supporting information (Figure S4). Notably, the peak position of Ge at different depths were consistent, without any offset, suggesting that the valence of Ge keeps at the same state. By fitting the Ge 3d spectrum after 1,800 s of etching, we confirmed the peak position is consistent with that of the pristine LAGP (Figure 6D), proving that the LixC60 interlayer successfully inhibited the reduction of Ge4+. In addition, no peak position shift was observed in the in-depth C 1s spectra of LAGP-C60 (Figure S4), which indicates that the LixC60 interlayer itself is chemically homogeneous without side reactions.

Figure 6.

Figure 6

Open in a new tab

Investigation of Interface Protection Mechanism

XPS spectra of Ge 3d and their fitted results for (A) pristine LAGP and (B) LAGP after 25 cycles. (C) XPS Ge 3d spectra with Ar+ sputtering at different times for LAGP-C60 after 25 cycles. (D) Ge 3d spectra of cycled LAGP-C60 after 1 800 s of sputtering and its fitted results. Surface morphology of (E) cycled LAGP and (F) cycled LAGP-C60. Cross-sectional image of (G) cycled Li/LAGP interface and (H) cycled Li/LAGP-C60 interface. (I) Optical image of lithium foil surface from cycled Li/LAGP/Li and Li/LAGP-C60/Li cells. The LAGP pellet without C60 coating was pulverized after cycling and some white particles adhered to the Li foil. Schematic demonstration of the interface protection mechanism: (J) Li/LAGP interface before and after cycling, (K) Li/LAGP-C60 interface before and after cycling.

In the SEM image of cycled LAGP (Figure 6E), we observed lots of reacted black areas, which is discontinuously distributed in the white surface of unreacted LAGP. As a comparison, the surface of cycled LAGP-C60 displays a very uniform color (Figure 6F). The section images of SE/Li interface were also collected to investigate the interfacial contact. Owing to the pulverization of uncoated LAGP, many small pieces fell off during sample preparation. In the remaining observable LAGP/Li interface, as shown in Figure 6G, some parts of metallic lithium were found mechanically separated from the LAGP. The cycled Li/LAGP-C60 interface still maintained excellent contact, highlighting the bonding functionality of LixC60 layer (Figure 6H). Figure 6I is the optical images lithium electrodes peeled from LAGP and LAGP-C60. Some pulverized LAGP pieces contaminated lithium surfaces from Li/LAGP sample, whereas a clean lithium surface was observed with C60 coating, proving the successful inhibition of side reactions by LixC60 interlayer. Figures 6J and 6K schematically summarizes the protective mechanism of C60 interlayer. For uncoated LAGP, the interface is dominated by a point-to-point reaction owing to the poor contact of LAGP and lithium metal. Owing to the inhomogeneous distribution of electric field, lithium dendrites were induced to grow in local area, which is accompanied by a continuous reaction between metallic lithium and LAGP, and the MCI with black color eventually formed as shown in the Figure 6E. During the cycle, the MCI continuously propagated to the bulk of LAGP owing to its relatively high electronic conductivity. And it caused the pulverization of LAGP pellet and cell impedance skyrocketing. With a C60 interlayer, as shown in Figure 6K, the coated C60 was first reacted with lithium metal to form a homogeneous LixC60 layer and re-construct the interface. The close contact made the LAGP/LixC60 interface dominated by a stable face-to-face reaction, which explains the uniform surface observed by SEM in Figure 6H. Thanks to the improvement on interfacial contact, the cell impedance was reduced by more than one order of magnitude and maintained a low increase after cycling. The tightly wrapped LixC60 layer successfully isolates the LAGP from metallic lithium, prevents the reduction of Ge4+ to Ge2+/Ge0, and achieves a chemically stable electrode-electrolyte interface. Besides, LixC60 exhibits high ionic conductivity but negligible electronic conductivity, which is beneficial to hinder the growth of lithium dendrites at the interface. The soft nature of LixC60 layer can also release the local stress during the lithiation and delithiation, avoiding dendrites and cracks formation, achieving a mechanically robust interface.

Conclusions

Overall, a feasible strategy was proposed to optimize the interface between SEs and metallic lithium in solid state batteries. A conformal C60 layer was efficiently built at the Li/SEs interface by PVD. The as-deposited C60 can react with lithium to form a stable LixC60 layer, which has high ionic conductivity but negligible electronic conductivity. This LixC60 layer effectively suppressed the reduction of LAGP by metallic lithium, significantly improved the interfacial contact, and successfully hindered the growth of lithium dendrites and the propagation of internal cracks. As a result, the assembled Li/LAGP-C60/Li symmetric cell can stably cycle for 4,500 h (more than 6 months) at 0.034 mA cm−2 and Li/LAGP-C60/LFP cell can stably cycle for more than 100 times at room temperature. This work provides a controllable and scalable method to improve the SEs/Li interface in one step, which is promising to be applied to other solid electrolyte systems.

Limitations of the Study

There are some limitations to this work. The rate performance of the full battery is limited due to the fact that the intrinsic ionic conductivity of the solid electrolyte used here is much lower than that of the traditional liquid electrolyte. When working at high current density, the transfer of lithium ions in the battery is restricted, which greatly limits the specific capacity and cycle performance of the battery at high current density. Exploring solid electrolyte materials with higher ionic conductivity is an effective method to solve this problem, which requires further research.

Resource Availability

Lead Contact

weiguodan@sz.tsinghua.edu.cn.

Materials Availability

All the materials used in this work are available, and the sourse is stated in the Transparent Method supplemental file.

Data and Code Availability

The data and codes used or analyzed in this work are available from the corresponding authors through reasonable request.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.

Acknowledgments

This research was supported by the National Natural Science Foundation of China (No. 51872157), Shenzhen Municipal Development and Reform Commission, and New Energy Technology Engineering Laboratory (Grant Number: SDRC [2016]172), Shenzhen Technical Plan Project (No. KQJSCX20160226191136, JCYJ20170412170911187, and JCYJ20170817161753629). Z.L. appreciates the help on basic skills from Qi Liu and Aihua Ran in the research.

Author Contributions

G.W., K.Q., B.L., and F.K. gave the initial idea of this work. Z.L. and K.Q. performed the experiments and wrote the manuscript. S.Z. contributed to the DFT calculations. P.N. helped draw schematics. S.C., X.Z., and T.L. discussed the results and participated in the preparation of the paper. All authors have given approval to the final version of the manuscript.

Declaration of Interests

There are no conflicts of interest to declare.

Published: October 23, 2020

Footnotes

Supplemental Information can be found online at https://doi.org/10.1016/j.isci.2020.101636.

Contributor Information

Kun Qian, Email: qiankun425@163.com.

Guodan Wei, Email: weiguodan@sz.tsinghua.edu.cn.

Supplemental Information

Document S1. Transparent Methods and Figures S1–S6
mmc1.pdf (974KB, pdf)

References

  1. Allemand P.M., Koch A., Wudl F., Rubin Y., Diederich F., Alvarez M.M., Anz S.J., Whetten R.L. Two different fullerenes have the same cyclic voltammetry. J. Am. Chem. Soc. 1991;113:1050–1051. [Google Scholar]
  2. Bachman J.C., Muy S., Grimaud A., Chang H.H., Pour N., Lux S.F., Paschos O., Maglia F., Lupart S., Lamp P. Inorganic solid-state electrolytes for lithium batteries: mechanisms and properties governing ion conduction. Chem. Rev. 2016;116:140–162. doi: 10.1021/acs.chemrev.5b00563. [DOI] [PubMed] [Google Scholar]
  3. Chabre Y., Djurado D., Armand M., Romanow W.R., Coustel N., McCauley J.P., Fischer J.E., Smith A.B. Electrochemical intercalation of lithium into solid fullerene C60. J. Am. Chem. Soc. 1992;114:764–766. [Google Scholar]
  4. Cheng Q., Li A., Li N., Li S., Zangiabadi A., Li T.-D., Huang W., Li A.C., Jin T., Song Q. Stabilizing solid electrolyte-anode interface in Li-metal batteries by boron nitride-based nanocomposite coating. Joule. 2019;3:1510–1522. [Google Scholar]
  5. Chung H., Kang B. Mechanical and thermal failure induced by contact between a Li1.5Al0.5Ge1.5(PO4)3 solid electrolyte and Li metal in an all solid-state Li cell. Chem. Mater. 2017;29:8611–8619. [Google Scholar]
  6. Cortes F.J.Q., Lewis J.A., Tippens J., Marchese T.S., McDowell M.T. How metallic protection layers extend the lifetime of NASICON-based solid-state lithium batteries. J. Electrochem. Soc. 2019;167:050502. [Google Scholar]
  7. Duan J., Wu W., Nolan A.M., Wang T., Wen J., Hu C., Mo Y., Luo W., Huang Y. lithium–graphite paste: an interface compatible anode for solid-state batteries. Adv. Mater. 2019;31:1807243. doi: 10.1002/adma.201807243. [DOI] [PubMed] [Google Scholar]
  8. Feng J.K., Lu L., Lai M.O. Lithium storage capability of lithium ion conductor Li1.5Al0.5Ge1.5(PO4)3. J. Alloys Compd. 2010;501:255–258. [Google Scholar]
  9. Feng W., Dong X., Zhang X., Lai Z., Li P., Wang C., Wang Y., Xia Y. Li/garnet interface stabilization by thermal-decomposition vapor deposition of an amorphous carbon layer. Angew. Chem. Int. Ed. 2020;59:5346–5349. doi: 10.1002/anie.201915900. [DOI] [PubMed] [Google Scholar]
  10. Fu K.K., Gong Y., Liu B., Zhu Y., Xu S., Yao Y., Luo W., Wang C., Lacey S.D., Dai J. Toward garnet electrolyte-based Li metal batteries: an ultrathin, highly effective, artificial solid-state electrolyte/metallic Li interface. Sci. Adv. 2017;3:e1601659. doi: 10.1126/sciadv.1601659. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Fu J., Yu P., Zhang N., Ren G., Zheng S., Huang W., Long X., Li H., Liu X. In situ formation of a bifunctional interlayer enabled by a conversion reaction to initiatively prevent lithium dendrites in a garnet solid electrolyte. Energy Environ. Sci. 2019;12:1404–1412. [Google Scholar]
  12. Gao Z., Sun H., Fu L., Ye F., Zhang Y., Luo W., Huang Y. Promises, challenges, and recent progress of inorganic solid-state electrolytes for all-solid-state lithium batteries. Adv. Mater. 2018;30:e1705702. doi: 10.1002/adma.201705702. [DOI] [PubMed] [Google Scholar]
  13. Haddon R.C. Electronic structure, conductivity and superconductivity of alkali metal doped (C60) Acc. Chem. Res. 2002;25:127–133. [Google Scholar]
  14. Han X., Gong Y., Fu K.K., He X., Hitz G.T., Dai J., Pearse A., Liu B., Wang H., Rubloff G. Negating interfacial impedance in garnet-based solid-state Li metal batteries. Nat. Mater. 2017;16:572–579. doi: 10.1038/nmat4821. [DOI] [PubMed] [Google Scholar]
  15. Han F.D., Yue J., Zhu X.Y., Wang C.S. Suppressing Li dendrite formation in Li2S-P2S5 solid electrolyte by LiI incorporation. Adv. Energy Mater. 2018;8:1703644. [Google Scholar]
  16. Hao X., Zhao Q., Su S., Zhang S., Ma J., Shen L., Yu Q., Zhao L., Liu Y., Kang F. Constructing multifunctional interphase between Li1.4Al0.4Ti1.6(PO4)3 and Li metal by magnetron sputtering for highly stable solid-state lithium metal batteries. Adv. Energy Mater. 2019;9:1901604. [Google Scholar]
  17. Hartmann P., Leichtweiss T., Busche M.R., Schneider M., Reich M., Sann J., Adelhelm P., Janek J. Degradation of NASICON-type materials in contact with lithium metal: formation of mixed conducting interphases (MCI) on solid electrolytes. J. Phys. Chem. C. 2013;117:21064–21074. [Google Scholar]
  18. He L., Sun Q., Chen C., Oh J.A.S., Sun J., Li M., Tu W., Zhou H., Zeng K., Lu L. Failure mechanism and interface engineering for NASICON-structured all-solid-state lithium metal batteries. ACS Appl. Mater. Interfaces. 2019;11:20895–20904. doi: 10.1021/acsami.9b05516. [DOI] [PubMed] [Google Scholar]
  19. Heiney P.A., Fischer J.E., McGhie A.R., Romanow W.J., Denenstein A.M., McCauley J.P., Jr., Smith A.B., Cox D.E. Orientational ordering transition in solid C60. Phys. Rev. Lett. 1991;66:2911–2914. doi: 10.1103/PhysRevLett.66.2911. [DOI] [PubMed] [Google Scholar]
  20. Hou G.M., Ma X.X., Sun Q.D., Ai Q., Xu X.Y., Chen L.N., Li D.P., Chen J.H., Zhong H., Li Y. Lithium dendrite suppression and enhanced interfacial compatibility enabled by an ex situ SEI on Li anode for LAGP-based all-solid-state batteries. ACS Appl. Mater. Interfaces. 2018;10:18610–18618. doi: 10.1021/acsami.8b01003. [DOI] [PubMed] [Google Scholar]
  21. Huo H., Chen Y., Li R., Zhao N., Luo J., Pereira da Silva J.G., Mücke R., Kaghazchi P., Guo X., Sun X. Design of a mixed conductive garnet/Li interface for dendrite-free solid lithium metal batteries. Energy Environ. Sci. 2020;13:127–134. [Google Scholar]
  22. Krauskopf T., Dippel R., Hartmann H., Peppler K., Mogwitz B., Richter F.H., Zeier W.G., Janek J. Lithium-metal growth kinetics on LLZO garnet-type solid electrolytes. Joule. 2019;3:2030–2049. [Google Scholar]
  23. Kuzmany H., Matus M., Burger B., Winter J. Raman scattering in C60 fullerenes and fullerides. Adv. Mater. 1994;6:731–745. [Google Scholar]
  24. Lewis J.A., Cortes F.J.Q., Boebinger M.G., Tippens J., Marchese T.S., Kondekar N., Liu X.M., Chi M.F., McDowell M.T. Interphase morphology between a solid-state electrolyte and lithium controls cell failure. ACS Energy Lett. 2019;4:591–599. [Google Scholar]
  25. Liu Y., Sun Q., Zhao Y., Wang B., Kaghazchi P., Adair K.R., Li R., Zhang C., Liu J., Kuo L.Y. Stabilizing the interface of NASICON solid electrolyte against Li metal with atomic layer deposition. ACS Appl. Mater. Interfaces. 2018;10:31240–31248. doi: 10.1021/acsami.8b06366. [DOI] [PubMed] [Google Scholar]
  26. Liu Y., Li C., Li B., Song H., Cheng Z., Chen M., He P., Zhou H. Germanium thin film protected lithium aluminum germanium phosphate for solid-state Li batteries. Adv. Energy Mater. 2018;8:1702374. [Google Scholar]
  27. Liu Q., Yu Q.P., Li S., Wang S.W., Zhang L.H., Cai B.Y., Zhou D., Li B.H. Safe LAGP-based all solid-state Li metal batteries with plastic super-conductive interlayer enabled by in-situ solidification. Energy Storage Mater. 2020;25:613–620. [Google Scholar]
  28. Luntz A.C., Voss J., Reuter K. Interfacial challenges in solid-state Li ion batteries. J. Phys. Chem. Lett. 2015;6:4599–4604. doi: 10.1021/acs.jpclett.5b02352. [DOI] [PubMed] [Google Scholar]
  29. Luo W., Gong Y., Zhu Y., Li Y., Yao Y., Zhang Y., Fu K., Pastel G., Lin C.F., Mo Y. Reducing interfacial resistance between garnet-structured solid-state electrolyte and Li-metal anode by a germanium layer. Adv. Mater. 2017;29:1606042. doi: 10.1002/adma.201606042. [DOI] [PubMed] [Google Scholar]
  30. Peng Z., Zhao N., Zhang Z., Wan H., Lin H., Liu M., Shen C., He H., Guo X., Zhang J.-G. Stabilizing Li/electrolyte interface with a transplantable protective layer based on nanoscale LiF domains. Nano Energy. 2017;39:662–672. [Google Scholar]
  31. Porz L., Swamy T., Sheldon B.W., Rettenwander D., Fromling T., Thaman H.L., Berendts S., Uecker R., Carter W.C., Chiang Y.M. Mechanism of lithium metal penetration through inorganic solid electrolytes. Adv. Energy Mater. 2017;7:1701003. [Google Scholar]
  32. Shao Y., Wang H., Gong Z., Wang D., Zheng B., Zhu J., Lu Y., Hu Y.-S., Guo X., Li H. Drawing a soft interface: an effective interfacial modification strategy for garnet-type solid-state Li batteries. ACS Energy Lett. 2018;3:1212–1218. [Google Scholar]
  33. Stephens P.W., Mihaly L., Lee P.L., Whetten R.L., Huang S.-M., Kaner R., Deiderich F., Holczer K. Structure of single-phase superconducting K3C60. Nature. 1991;351:632–634. [Google Scholar]
  34. Tippens J., Miers J.C., Afshar A., Lewis J.A., Cortes F.J.Q., Qiao H., Marchese T.S., Di Leo C.V., Saldana C., McDowell M.T. Visualizing chemomechanical degradation of a solid-state battery electrolyte. ACS Energy Lett. 2019;4:1475–1483. [Google Scholar]
  35. Wang H., He Y., Li Y., Su H. Photophysical and electronic properties of five PCBM-like C60 derivatives: spectral and quantum chemical view. J. Phys. Chem. A. 2012;116:255–262. doi: 10.1021/jp208520v. [DOI] [PubMed] [Google Scholar]
  36. Wang C., Gong Y., Liu B., Fu K., Yao Y., Hitz E., Li Y., Dai J., Xu S., Luo W. Conformal, nanoscale ZnO surface modification of garnet-based solid-state electrolyte for lithium metal anodes. Nano Lett. 2017;17:565–571. doi: 10.1021/acs.nanolett.6b04695. [DOI] [PubMed] [Google Scholar]
  37. Wang S., Wang J., Liu J., Song H., Liu Y., Wang P., He P., Xu J., Zhou H. Ultra-fine surface solid-state electrolytes for long cycle life all-solid-state lithium–air batteries. J. Mater. Chem. C. 2018;6:21248–21254. [Google Scholar]
  38. Wang X., Zeng W., Hong L., Xu W., Yang H., Wang F., Duan H., Tang M., Jiang H. Stress-driven lithium dendrite growth mechanism and dendrite mitigation by electroplating on soft substrates. Nat. Energy. 2018;3:227. [Google Scholar]
  39. Wang L., Hu S., Su J., Huang T., Yu A. Self-sacrificed interface-based on the flexible composite electrolyte for high-performance all-solid-state lithium batteries. ACS Appl. Mater. Interfaces. 2019;11:42715–42721. doi: 10.1021/acsami.9b12112. [DOI] [PubMed] [Google Scholar]
  40. Wu B.B., Wang S.Y., Lochala J., Desrochers D., Liu B., Zhang W.Q., Yang J.H., Xiao J. The role of the solid electrolyte interphase layer in preventing Li dendrite growth in solid-state batteries. Energy Environ. Sci. 2018;11:1803–1810. [Google Scholar]
  41. Yu Q., Han D., Lu Q., He Y.B., Li S., Liu Q., Han C., Kang F., Li B. Constructing effective interfaces for Li1.5Al0.5Ge1.5(PO4)3 pellets to achieve room-temperature hybrid solid-state lithium metal batteries. ACS Appl. Mater. Interfaces. 2019;11:9911–9918. doi: 10.1021/acsami.8b20413. [DOI] [PubMed] [Google Scholar]
  42. Zhang Z., Zhao Y., Chen S., Xie D., Yao X., Cui P., Xu X. An advanced construction strategy of all-solid-state lithium batteries with excellent interfacial compatibility and ultralong cycle life. J. Mater. Chem. C. 2017;5:16984–16993. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Document S1. Transparent Methods and Figures S1–S6
mmc1.pdf (974KB, pdf)

Data Availability Statement

The data and codes used or analyzed in this work are available from the corresponding authors through reasonable request.

Articles from iScience are provided here courtesy of Elsevier

RESOURCES