Introduction

Nowadays, with the popularization of portable high-energy storage devices and electric vehicles, traditional commercial battery systems are difficult to meet the existing needs. Lithium (Li) anodes are considered to be quite promising due to their high theoretical capacity (3860 mAh/g) and low voltage (− 3.04 V vs. Standard Hydrogen Electrode) [1]. In the traditional commercial battery system, LiFePO4 (LFP) [2], LiNi1-x-yCoxMnyO2 (NCM), LiNi1-x-yCoxAlyO2 (NCA), and LiCoO2 (LCO) are mainly used as cathode, and graphite is used as anode. Based on the intercalation/deintercalation reaction mechanism, their low theoretical capacity is difficult to meet the high energy density demand of portable high-energy storage devices and electric vehicles (EVs). Therefore, it is imperative to develop high-energy battery systems such as lithium–sulfur battery (LSB) and lithium–air battery (LAB). Li anodes have been initially studied due to high theoretical capacity and low voltage. However, Li dendrites are usually generated during repeated Li plating and depleting process because of the low surface energy of Li and tip effect [3]. Moreover, the large volume change originated from Li plating would break the formed solid electrolyte interface (SEI) film, and thus, facilitating the Li dendrite growth in the cracked regions. Whereas, during Li stripping process, the SEI film would be fractured and caused the formation of “dead” Li, which would severely block the electronic conductivity of the electrode. With the increase of the cycling time, such a phenomenon would be more and more serious, and finally, a thick accumulated SEI film and excessive dead Li would be generated, which is the main reason for the fading of the capacity [1, 4].

A series of problems caused by dendritic Li has been a bottleneck in the industrialization of Li metal anodes. The development of effective protection technologies for Li anodes has become the key to push the development of lithium metal batteries. Thanks to the unremitting efforts of scientific researchers, several solutions for improving the stability of Li anodes have been gradually emerged. For example, the introduction of electrolyte additives (e.g., fluoroethylene carbonate (FEC) [5], bis(fluorosulfonyl)imide salt (FSI−), bis(trifluoromethanesulphonyl)imide salt (TFSI−), lithiumbis(trifluoromethanesulphonyl)imide (LiFSI) [6], dimethyl carbonate (DMC), bis(2,2,2-trifluoroethyl) ether (BTFE) [7], and 1,1,3,3-tetramethylurea (TMU) [8], etc.) can effectively induce uniform deposition of Li+. In another case, constructing three-dimensional (3D) current collectors (e.g., modified copper foam [9, 10], foamed nickel with carbon spheres sputtered [11], graphitized carbon fibers (GCF) [12], alumina (Al2O3) particles [13], etc.) can guide lateral growth of Li, thereby inhibiting the production of dead Li. Moreover, by employing “lithiophilic” component (e.g., frame porphyrin (POF) [14, 15], LiC6 [16], heteroatom doped carbon [17], aluminum lithium alloys [18], nanometer diamond materials [19], grapheme [20,21,22,23,24,25,26] etc.) can reduce the diffusion barrier of Li+ in deposition, thereby inducing uniform deposition of Li. However, the abovementioned methods are generally complex and hard to be realized for practical applications. Using ex-situ protective layers (e.g., polydimethylsiloxane (PDMS) [27], fluoroethylene carbonate (FEC) [28], lithium phosphorus oxynitride (LiPON) [29], lithium orthophosphate (Li3PO4) [30], polyimide (PI) and poly (vinylidene fluoride-co-hexafluoropropylene) (PVDF-HFP) [31, 32], inorganic ceramic and organic polymer materials [33], LiTFSI-LiNO3-Li2S5 salt [34], monolayer cross-linked amorphous hollow carbon nanospheres [35], hexagonal lithium nitride [36], etc.) is also an effective strategy for inhibiting the growth of metallic lithium dendrites. In comparison, the experimental requirement for this strategy is quite simple and easy for scale up. Nevertheless, further improvement of the ionic conductivity of the fabricated protective layers is still a big challenge.

In this work, a stable protective layer (SPL) was coated on the surface of cooper current collector by a simple process with poly(vinylidene-co-hexafluoropropylene) (PVDF-HFP) as the matrix, nano alumina (nano-Al2O3), lithium fluoride (LiF), and lithiumbis(trifluoromethanesulphonyl)imide (LiTFSI) as the additives. Here, LiF is beneficial to increase the surface tension of Li electrode for reducing the diffusion barrier of Li+ [5], nano-Al2O3 particles is beneficial to improve the mechanical strength of SPL, and LiTFSI is beneficial to improve the ionic conductivity of SPL. In the copper foil electrode modified with SPL, the Li stripping/plating process happened in the space between SPL and the copper foil and exhibited a much enhanced cycling performance compared with the one without modification, which did not show obvious capacity decay after 250 cycles at the current density of 0.5 mA/cm2 for 2 h. From scanning electron microscopy (SEM) and X-ray photoelectron spectroscopy (XPS) characterizations, it had to be found that the preset protective composite layer kept a stable structure during the charge and discharge process, which guaranteed the achieved stable cycling performance. In addition, a full battery system based on modified copper foil anode and LiFePO4 (LFP) [37] cathode also constructed and presented a good cycling performance.

Experimental

Preparation

In a typical procedure [38], 1.5 g of PVDF-HFP (AR, Sigma-Aldrich), 0.5 g of nano-Al2O3 (AR, Sigma-Aldrich), 0.5 g of LiF (AR, Sigma-Aldrich), and 0.15 g of LiTFSI (AR, Sigma-Aldrich) were firstly dissolved in 27 ml mixed solvent (acetone/dimethylacetamide (DMAC) (AR, Sigma-Aldrich) with a volume ratio of 2:1) followed by stirred for 24 h and ultrasonically dispersed for 2 h. After that, the dispersion was blade coated onto an acid-treated copper foil (by soaking the copper foil within 2 M HCl for 30 min and washed with plenty of distilled water), with a certain thickness of 300 μm. The obtained copper foil with coated slurry was left in the ambient environment for 2 h and then transferred into a vacuum oven treated at 80 °C for 12 h to obtained a solid SPL coated copper foil, which was cut into pieces with a diameter of 16.0 mm for assembling the projective CR2025 Li-Cu half-cells. For comparison proposes, the control cells used acid-treated cooper electrode and (PVDF-HFP)-modified copper electrode was also assembled and tested. The separator used in this study was polypropylene (PP) film (celgard 2500). The electrolyte used in this study was 1 M lithium hexafluorophosphate (LiPF6) dissolved in ethylene carbonate (EC)/dimethyl carbonate (DMC) (AR, Sigma-Aldrich) mixed solvent with volume ratio of 1:1.

In this work, the SPL was firstly coated on the surface of the copper current collector, and then a sandwich structure was realized by electroplating of Li into the space between SPL and copper disc. After discharging at the current density of 0.5 mA/cm2 for 20 h, the discharged cells were transferred into the glove box and disassembly to obtain the electroplated Li anode and defined as SPL/Li/Cu. Schematic diagram of SPL/Li/Cu preparation process is displayed in Fig. 1.

Fig. 1
figure 1

Schematic diagram of the SPL/Li structure preparation process

Full size image

CR2025 SPL/Li/Cu|LFP full-cells were assembled with the SPL/Li/Cu as the anode and LFP as the cathode. For comparison purposes, CR2025 Li/Cu|LFP full-cells used bare Cu foil anode and LFP cathode were also assembled and tested. The used separator and electrolyte were the same with the abovementioned ones. All the cell assembly processes were accomplished in an Ar-filled glove box. In LFP cathode, it contained 80 wt% of LFP (Likai Co. Ltd., Taiwan), 10 wt% of poly(vinylidene fluoride) (PVDF), and 10 wt% acetylene black with mass loading of ~ 1.4 mg/cm2.

Electrochemical measurements

During each galvanostatic cycle of the Li-Cu half-cell, it was discharged at a current density of 0.5 mA/cm2 for 2 h and charged with a cut-off voltage of 1.0 V (vs. Li/Li+) for lithium plating and stripping processes. During each galvanostatic cycle of the SPL/Li/Cu|LFP full cell, the cells were charged and discharged over a voltage range of 2.8 V to 4.2 V (vs. Li/Li+) where 1 C corresponds to 170 mAh. All the cycling performance tests were tested on a Neware Battery Testing System.

The AC impedance (EIS) measurement was performed by an electrochemical workstation (Chen Hua, CHl760E) with the frequency range from 105 to 10−2 Hz under the amplitude of 5 mV. The parameters of the equivalent circuit were fitted by computer simulations using the Zview software.

The ionic conductivity of SPL was tested by EIS measurement in a stainless steel (SS) | liquid electrolyte | SPL | SS configured system. The used electrolyte and test conditions were the same with the abovementioned ones. The conductivity of the prepared SPL can be calculated by the following equation:

$$ s=D/\left(10\times S\times {R}_{\Omega}\right) $$

where s (mS/cm) is the ionic conductivity of the membrane, D (μm) is the thickness of the membrane, S (cm2) is the effective working area of the membrane, and R (Ω) is the electrolyte bulk impedance.

Characterization

The morphological information of the samples were collected by scanning electron microscopy (SEM, Hitachi SU8000, Japan). The crystallinity information of the samples were collected by X-ray diffraction (XRD, Rigaku, Japan) measurements using an Al Kα radiation. The valence state and composition information of the samples were collected by X-ray photoelectron spectroscopy (XPS, PHI 5700 ESCA System, American Physical Electronics Corporation). To obtain more information of the Li disc in SPL/Li/Cu, the SPL was carefully removed from the electrodes prior to the XPS test. The XPS spectra was fitted using peakfit V4 software.

Results and discussion

Figure 2a displays the SPL surface image of SPL/Li/Cu obtained from the disassembled cells. It can be seen that SPL surface was uniform and showed a yellow color. The morphology information of the SPL after cycling test were carefully examined by SEM. As shown in Fig. 2b, c, the small particles on the surface of the SPL were uniformly distributed, which was the LiF component (Fig. 2b). As depicted in Fig. 2c, attributed to good stability of SPL, there was no significant change on the surface after 3 cycles at the current density of 0.5 mA/cm2 for 2 h. Figure 2d–f presents the top view and cross section SEM image of deposited Li in Li/Cu (10 mA h/cm2 of Li was deposited on bare Cu foil at a current density of 0.5 mA/cm2 after 3 or 20 cycles (0.5 mA/cm2 for 2 h)). From the top view SEM image of deposited Li (Fig. 2d), the flattened Li dendrites with distinct grain boundaries could be obviously observed, which was caused by the large assembly stress in the cells, had consistent with the previous reports [27]. Furthermore, a rough SEI film could also been clearly observed in the cross section SEM image of the deposited Li (Fig. 2e). Figure 2f presents the surface morphological of Li deposition on copper foil after 20 cycles. It can be seen that the plating Li exhibited a distinct porous and dendritic structure on the bare Cu foil (Fig. 2f). On the contrary, when it comes to SPL/Li/Cu (under the same experimental conditions), the surface of Li deposition layer (the SPL has already been peeled off) showed a smooth and flat surficial morphology without broken interface and severe Li dendrites (Fig. 2g). It confirmed that a uniform Li layer can be deposited on the copper foil with SPL protection. More SEM images of the deposited Li with SPL are displayed in Fig. S1 (Support Information). The cross section SEM image of the plating Li on SPL/Li/Cu is shown in Fig. 2h, which demonstrated that the Li plating process happened in the space between SPL and the copper foil with a continuous and stable “sandwich” structure; SPL, Li layer, and copper foil all could be observed in the cross section image. Figure 2i presents the surface morphological of Li deposition on copper foil with SPL after 20 cycles (the SPL has already been peeled off). It can be seen that the plating Li exhibited a relatively flat morphology compared with Fig. 2f. The above evidences manifest that the Li stripping/plating process happened in the space between SPL and the copper foil and exhibited a much enhanced cycling performance compared with the one without modification. The reasons could be concluded as follows: On the one hand, LiF is beneficial to increase the surface tension of Li electrode for reducing the diffusion barrier of Li+ [5]; on the other hand, nano-Al2O3 particles are beneficial to improve the mechanical strength of SPL, inhibiting the growth of Li dendrites, and thus it exhibits excellent stability during battery cycling, forming a stable interface. These further improve the uniformity of Li deposition and prevent it from piercing the film, ensuring the safety of batteries.

Fig. 2
figure 2

Morphology information of SPL/Li/Cu. (a) The photo of SPL/Li/Cu. The SEM image of SPL surface at (b) initial state and (c) after 3 cycles for Li deposition. (d) Top view and (e) cross section SEM image of deposited Li on bare copper after 3 cycles. (f) Top view SEM image of deposited Li protected without SPL after 20 cycles. (g) Top view and (h) cross section SEM image of deposited Li on SPL/Cu after 3 cycles. (i) Top view SEM image of deposited Li protected with SPL after 20 cycles. All samples were obtained from the disassembled cells

Full size image

As shown in Fig. 3a, the SPL obtained from the disassembled cells after 3 cycles consisted of crystalline phases of PVDF-HFP, Al2O3 (JCPDS No. 05-0712) and LiF (JCPDS No. 45-1460), indicating that after electrochemical cycling, Al2O3 and LiF could keep stably in SPL without being dissolved. The peaks of PVDF-HFP are reduced by the addition of LiF and nano-Al2O3, which manifests that the crystallinity degree of PVDF-HFP is reduced, benefiting for strong segment motion, thereby improving the mass transfer ability of Li+ [38]. Figure 3b, c presents the XPS spectra of the Li disc in SPL/Li/Cu and Li/Cu after the 20th cycle, where the peaks of Li1s, C1s, O1s, and F1s could be clearly observed in both figures. In the high-resolution C1s spectra of the two electrodes, peaks corresponding to –CO– species, Li-alkylcarbonate (ROCO2Li), and Li-carbonate (Li2CO3) were observed at 287, 288, and 290 eV [30, 34], respectively. With the SPL protection (Fig. 3d), it was observed that the peak intensities for these products were significantly reduced (Fig. 3e), which were considered formed by reductive decomposition of cyclic carbonates (ea. EC and DMC), suggesting that SPL can mitigate electrolyte decomposition by hindering the direct contact between the Li anode with the liquid electrolyte. The Li1s spectra of both electrodes is given in Fig. S2 (Support Information).

Fig. 3
figure 3

XRD and XPS characterization. (a) XRD patterns of the SPL after 3 cycles, pristine PVDF-HFP film, and standard diffraction peaks of LiF (JCPDS No. 45-1460) and Al2O3 (JCPDS No. 05-0712). The XPS characterization spectra of (b) bare Cu foil and (c) SPL modified electrode after 20 cycles at 0.5 mA/cm2. The C1s XPS spectra of the Li electrode (d) without SPL and (e) with SPL after 20 cycles at 0.5 mA/cm2

Full size image

In order to evaluate the effect of SPL on electrochemical performance, we conducted a series of electrochemical tests in CR2032 cells. Figure 4 displays the specific electrochemical performance. Coulomb efficiency (CE) is identified as a vital argument for long cycle life, indicating the ratio of the amount of Li stripped in each cycle to the amount of Li plated. The stable value of CE generally indicates a stable surface between the electrode and the electrolyte. The change curves of CE for different electrodes are shown in Fig. 4a. The CE fluctuated drastically after 60 cycles for the bare Cu foil electrodes in the galvanostatic cycling test (0.5 mA/cm2 for 2 h). In comparison, Li-SPL/Cu half-cell maintained a relatively stable CE (≈ 98.9%) for more than 250 cycles, indicating that copper foil with SPL protection exhibits a stable CE and a long cycle life at a certain current density of 0.5 mA/cm2. In addition, when the SPL is replaced by the PVDF-HFP film without the function additives, CE can keep stable at around 98.0% for 180 cycles, indicating a lower improvement in cycling performance compared with SPL. The stable CE and long cycle life of Li-SPL/Cu half-cells could be interpreted as follows: In the process of Li deposition, LiF increase the surface tension of Li electrode for reducing the diffusion barrier of Li+; meanwhile, nano-Al2O3 particles improve the mechanical strength of SPL, which promote lateral growth of Li crystals, instead of the growth of “tip,” forming a uniform Li layer. Thereby, the original SEI film could keep stable, without the formation of dead Li during Li plating and stripping process, which are the main reasons for the stable CE and long cycle life. The voltage–time curve provides a better indication of the stability for Li plating and stripping. The voltage–time curves of the Li-SPL/Cu half-cells for 200 cycles are shown in Fig. 4b. From the two blown-up five-cycle curves at cycle time of ~ 30 and ~ 660 h, it can be observed that the charge–discharge platform difference is further reduced, and a stable Li plating/stripping platform can be found, further demonstrating the protection effect of SPL. EIS measurements (Fig. 4c) were conducted to study the cell resistance during cycling for different electrodes [39]. The control group without SPL exhibits an increase in resistance (28–31 Ω), while Li protected with SPL keeps decreased the trend in EIS impedance (49–42 Ω) after 20 cycles. Owing to the introduction of SPL, the impedance of the Cu foil modified with SPL is larger than that of the bare Cu foil. LiTFSI in SPL was gradually activated, which would improve the mass and electron transfer process and suppress the Li dendrites, thus reducing the impedance. As to the electrode without SPL, after 20 cycles, the SEI film would be fractured and caused the formation of dead Li, which would severely block the electronic conductivity of the electrode, with an increased impedance. The details for the fitted result of the equivalent circuit can be found in Table S1 [40,41,42,43]. In addition, the SPL film has a good ionic conductivity value of 0.3470 mS/cm (Fig. S3 and Table S2 in Support Information), similar to commercial separators. Furthermore, the CE of SPL-modified electrode still remained above 90% over 100 cycles in the galvanostatic cycling test with higher current density (1 mA/cm2 for 1 h), as shown in Fig. 4d, indicating that copper foil with SPL protection could also exhibit a stable CE and a long cycle life at a higher current density condition.

Fig. 4
figure 4

Cycling performances and EIS spectra of bare Cu foil and SPL modified Cu foil. (a) Comparison of the coulombic efficiency of bare Cu foil, (PVDF-HPP)-modified and SPL-modified electrode at 0.5 mA/cm2 and 1 mA h/cm2 Li deposition. (b) The voltage–time curve of modified electrode with SPL in long cycle time (the inset is the voltage–time curve of this cell at cycle time of ~ 30 and ~ 660 h). (c) EIS spectra of bare Li electrode or SPL-modified electrode before and after 20 cycles. (d) Coulombic efficiency of SPL-modified electrode at 1 mA/cm2 and 1 mA h/cm2 Li deposition

Full size image

In order to understand the electrochemical performance in the actual configuration, the galvanostatic cycling tests of the SPL/Li/Cu|LFP full batteries were conducted. Figure 5a displays the morphology of LFP (Likai Co. Ltd., Taiwan) material used in this work, in which regular shaped LFP particles can be seen, benefiting for good electrochemical performance. As depicted in Fig. 5b, the SPL/Li/Cu|LFP full cell maintained 90% of the initial discharge capacity after 100 cycles at a current density of 1 C, which was better than the Li/Cu|LFP full cell. By comparing the charging and discharging curves of both cells for the 1st and 50th cycle, as shown in Fig. 5c, it can be seen that the voltage difference between the charging and discharging platforms of SPL/Li/Cu|LFP full cells at both 1st and 50th cycles are smaller than Li/Cu|LFP full cells indicating that SPL reduces the battery polarization. Meanwhile, the CE of SPL/Li/Cu|LFP full cell could maintain 98% over 100 cycles (Fig. 5d). This suggested that even in the full battery configuration, the designed battery system still had good cycling performance by introducing SPL. In summary, we have demonstrated that SPL for Li anodes could greatly improve their electrochemical performance. By simply coating the SPL, using a conventional carbonate electrolyte, the cycle coulombic efficiency can be stabilized at about 95% over 200 cycles. It is anticipated that the electrochemical performance could be further enhanced by incorporating SPL with other ongoing efforts, such as combining with functional electrolytes. Therefore, we expect this viable design through a convenient process to provide complementary methods for the development of high-performance anodes for Li metal batteries, such as lithium–sulfur and lithium–air batteries.

Fig. 5
figure 5

Electrochemical performance of SPL/Li/Cu|LFP full cells and Li/Cu|LFP full cells. (a) SEM image of LFP. (b) The cycling performance of SPL/Li/Cu|LFP full cells and Li/Cu|LFP full cells at 1 C. (c) The charge and discharge curves of SPL/Li/Cu|LFP full cells and Li/Cu|LFP full cells at 1st and 50th cycle. (d) The coulombic efficiency of SPL/Li/Cu|LFP full cells at 1 C

Full size image

Conclusions

In summary, we have firstly prepared a SPL with PVDF-HFP as the matrix, and nano-Al2O3, LiF, and LiTFSI as the additives, which was coated on the surface of cooper current collector through a simple process. Then, the electrochemical performance, morphology, and mechanism of SPL/Li/Cu obtained by electroplating of Li into the space between SPL and copper disc have been systematically studied. The plating Li on the SPL/Li/Cu exhibited a relatively flat morphology compared with Li/Cu. Moreover, the copper foil electrode modified with SPL exhibited a much enhanced cycling performance compared with the one without modification, which did not show obvious capacity decay after 250 cycles at the current density of 0.5 mA/cm2 for 2 h. In addition, a full battery system based on SPL-modified copper foil anode and LFP cathode also was constructed and presented a good cycling performance. The method of the current collector coated with SPL provides a feasible approach to inhibit the growth of Li dendrites. In the future work, the Li anode with preset artificial SEI will be combined with more cathode materials, which will help us further explore the application prospect of metal lithium battery.