Rechargeable metallic lithium batteries have attracted great attentions to industrial community in recent years due to the promising future of solid-state batteries,^[1–3] lithium–sulfur,^[4] and lithium–air^[5] batteries. The obstacles of the application for metallic lithium anodes include at least five factors: 1) inhomogeneous deposition and dissolution of lithium; 2) side reactions between lithium and liquid electrolyte during storage and electrochemical reactions; 3) large volume variation of electrodes; 4) low melting temperature of metallic lithium; 5) soft and sticky properties. The inhomogeneous stripping and plating of lithium will lead to the uneven volume variation, which will result in the fresh lithium exposing into and reacting with the electrolyte, eventually leading to the failure of the cell. Li et al.^[5] have studied the effect of electrochemical deposition/dissolution order and found that the problem of non-uniform Li-ion plating/stripping during charge/discharge is very serious. In order to improve the electrochemical performance and safety property, numerous strategies such as tuning the electrolytes and salts,^[6–15] searching the additives,^[16–21] treating the separators,^[22,23] using the lithium alloy anodes,^[24] using ceramic/polymer solid electrolytes,^[25,26] and coating buffer layers^[27–36] have been proposed. Coating methods have been considered as one of the most direct and effective strategies among them. The coating precursors can be solid, liquid and gas. Recently Wen^[28] grew an Li₃N coating layer on the lithium anode through an in-situ gas reaction method, exhibiting excellent performance. Since the gas is pervasive and can fully react with lithium in all the active sites, thus a dense and uniform buffer layer can be obtained by gas reaction method. Therefore, gas reaction method is regarded as simple and producible, which is suitable for large scale application in energy storage field.

Herein, four kinds of gases include Ar, CO₂–O₂ (2:1), N₂, and CO₂–O₂–N₂ (2:1:3) were used to react with metallic lithium anode and their effects on electrode performance were studied. Hereafter, three different coating layers on lithium tablets were obtained, which may mainly contain Li₂CO₃, Li₃N,^[28] and the mixture of Li₃N and Li₂CO₃ respectively. The SIMS data shows the coating layer on lithium tablets have different density and uniformity. The micro conductive performance of the four as-prepared lithium tablets was analyzed by SPM. Li/Li symmetrical cells were assembled to study the stripping/plating performance. In addition, the SEM and SIMS characterizations before/after cycle in Li/S cell were compared. The electrochemical testing results indicate that a SEI-like composition formed on the lithium tablets can modulate the interfacial stabilization between the lithium foil and the ether electrolyte.

Figure 1 shows the XRD pattern of the four kinds of gases treated lithium tablets. It can be seen that under different gases, three peaks centered at 36.173°, 52.086°, and 65.186° are mainly appeared, which all belong to the reflection peaks of metallic lithium. However, the Ar and CO₂–O₂ (2:1) treated samples only show peaks at 36.173° and 65.186° while the N₂ and CO₂–O₂–N₂ (2:1:3) treated samples show all of the peaks, regardless the slight different of peak intensity. It is assumed that the different treatments may change the arrangement of lithium atoms, resulting in various prefer orientations. In spite of this, there is no new peak appearing in the reflections, which is possible due to that the coating layers are too thin to be detected by the XRD.

Fig. 1.

	Figure Option ViewDownloadNew Window
	Fig. 1. (color online) XRD of lithium tablets treated by argon (black line), CO2–O2 (2:1) (red line), N2 (green line), and CO2–O2–N2 (2:1:3) (blue line) gases respectively.

In order to identify the existence of nitrogen in the coating layers, SIMS was applied as it is capable of detecting elements in a high resolution. Figure 2 shows the SIMS spectra of the treated lithium tablets. From the spectra, it can be seen that nitrogen and oxygen do exist in all of the samples, which is probably because the difficult of removing all the N₂ in the Glove box. In addition, because of the high reaction activity, the lithium will react with residual N₂ or O₂ in the instruments during the transfer and fabrication process. Figure 2(a) shows the Ar-treated sample has a relative thick interlayer. Figure 2(b) shows a sharpest interface and thinnest interlayer, which reveals the CO₂–O₂ (2:1)-treated sample shows a thinner coating layer because of the good passivation effect of the Li₂CO₃. The N₂-treated sample has a wider nitrogen distribution, which indicates the coverage of Li₃N on the lithium tablet is non-uniform. The CO₂–O₂–N₂ (2:1:3) treated sample shows a sharper interface than the N₂ treated one, which manifests the dense and even film has been formed on the lithium tablet.

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. (color online) SIMS of lithium tablets treated by argon (a), CO2–O2 (2:1) (b), N2 (c), and CO2–O2–N2 (2:1:3) (d) gases respectively.

The SPM was used to evaluate the density and uniformity of the formed layer on the lithium tablets. From the PF-TUNA mapping (Fig. 3), it can be observed that the Ar treated lithium tablet has a relative uniform and high micro-current distribution. The CO₂–O₂ (2:1)-treated sample reveals even current distribution, which may be due to the uniformly formed Li₂CO₃ coating layer on the lithium tablets. The N₂-treated sample shows an uneven current distribution, indicating the non-uniformity of Li₃N on the lithium tablet. And the CO₂–O₂–N₂ (2:1:3)-treated sample shows a moderate current distribution between the treatment situations by N₂ and CO₂–O₂ (2:1).

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. (color online) Current distribution (obtained by SPM) of lithium tablets treated by argon (a), CO2–O2 (2:1) (b), N2 (c), and CO2–O2–N2 (2:1:3) (d) gases respectively.

To further analyze the current distribution and the property of different gases treated lithium tablets, NanoScope Analysis software (1.80 (Build R1.126200), Copyright©2016 Bruker Corporation) is used to count the current in the scanning area. A data process of shifting the minimum value of the current to zero is used to get the following histogram (Fig. 4). Figure 4 displays that the micro-currents of the CO₂–O₂ (2:1), N₂, and CO₂–O₂–N₂ (2:1:3)-treated lithium metal anode significantly reduce than the Ar-treated one. The CO₂–O₂ (2:1)-treated lithium tablet shows the smallest average current, which indicates the Li₂CO₃ layer may forms on the lithium surface and has a high resistance. This infers that the layer is very dense or thick, preventing the lithium tablet from the electron contact with the current collector. The current of N₂-treated lithium tablet only slightly shifts to the low current region, which means the Li₃N is film is no denser than Li₂CO₃. The CO₂–O₂–N₂ (2:1:3)-coated sample shows a moderate current distribution between the CO₂–O₂ (2:1) and N₂-treated sample, which implying that the mixture gases may react with lithium tablet and Li₂CO₃ and Li₃N compounds are formed on the surface, leading to a middle current distribution.

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. (color online) Current histogram of lithium tablets treated by argon (black line), CO2–O2 (2:1) (red line), N2 (green line), and CO2–O2-N2 (2:1:3) (blue line) gases respectively.

Figure 5 shows the performance of Li/Li symmetrical cell with gas treatment. The cell with Ar-treated sample shows a very large polarization (over 100 mV), which reveals unstable electrochemical stability. The cell with CO₂–O₂ (2:1)-treated lithium tablet shows a much higher stability and smaller polarization than those of the Ar-treated sample, but still higher than the other two cases. The N₂-treated sample shows a small polarization, but it is not as stable as the condition in CO₂–O₂ (2:1) treatment. Among them, the cell with CO₂–O₂–N₂-(2:1:3)-treated lithium tablet possesses both the relative small polarization as well as high stability. Considering lithium will continuously react with N₂ to result in uneven reaction sites, N₂-treated lithium tablet could not be well protected. The Li₂CO₃ is relative more dense and stable than Li₃N, but the cell with lithium tablet coated by Li₂CO₃ has a bigger polarization which is not preferred in the batteries. Hence, the CO₂–O₂–N₂ (2:1) treatment process is more applicable in real cell systems.

Fig. 5.

	Figure Option ViewDownloadNew Window
	Fig. 5. (color online) Electrochemical performances of lithium symmetrical cell with lithium tablets treated by argon (black line), CO2–O2 (2:1) (red line), N2 (green line), and CO2–O2–N2 (2:1:3) (blue line) gases respectively.

Figure 6 shows the electrochemical properties of Li/S cell with different gases treated lithium tablets. The CO₂–O₂–N₂ (2:1:3)-treated sample shows the best capacity retention and the most stable Coulombic efficiency according to Figs. 6(b) and 6(c). Figure 6(a) shows CO₂–O₂ (2:1)-treated sample has a large overcharge. It is assumed that the dense Li₂CO₃ has a large resistance and poly-sulfides will accumulate near the surface on the lithium tablet. Moreover, the Li₂CO₃ component is not stable with poly-sulfides. However, the Li₃N buffer layer can block the shuttle of the poly-sulfides effectively. Thus the CO₂–O₂–N₂ (2:1:3)-treated sample combines the advantages of both Li₂CO₃ and Li₃N coating layers and shows better performance.

Fig. 6.

	Figure Option ViewDownloadNew Window
	Fig. 6. (color online) Electrochemical performances of Li/S cell with lithium tablets treated by argon (black line), CO2–O2 (2:1) (red line), N2 (green line), and CO2–O2–N2 (2:1:3) (blue line) gases respectively. (a) The 1st cycle curve; (b) the Coulombic efficiency; (c) the discharge capacity.

Figure 7 shows the morphology evolution of the gas protected lithium before and after cycling. No obvious difference could be observed among the un-cycled samples, indicating the morphology does not change largely and the passivated film is very thin. However, these 4 kinds of lithium metal anodes exhibit different surface morphology evolutions after electrochemical striping and plating in Li/S cells. In Ar-treated lithium anode, lots of cracks appear, and the similar morphology evolution was observed from CO₂–O₂ (2:1)-and N₂-treated lithium anodes. It is also displayed that CO₂–O₂ (2:1)-treated sample has a number of clusters of cracks and holes, while N₂-treated samples only has isolated holes. As expected, the CO₂–O₂–N₂ (2:1:3)-treated anode inherits the surface morphology features from the CO₂–O₂ (2:1) and the N₂ treated sample, which shows both cracks and isolated holes. If N₂ gas contributes to the Li₃N formation and the CO₂–O₂ (2:1) gas provides a denser and more uniform passivated Li₂CO₃ film, thus the CO₂–O₂–N₂ (2:1:3)-treated sample possesses the merits from both the CO₂–O₂ (2:1) and N₂ treating methods. As is well known, Li₂CO₃ has a lower electronic and ionic conductivity compared to that of Li₃N, as a result, the CO₂–O₂–N₂ (2:1:3)-treated sample has a dense, uniform, and ionic passable protection layer. It is also in consistent with the electrochemical cycling performance and XRD data.

Fig. 7.

	Figure Option ViewDownloadNew Window
	Fig. 7. (color online) Morphology evolution before (a), (c), (e), (g), and after (b), (d), (f), (h) cycle of Li/S cell with lithium tablets treated by argon [(a) and (b)], CO2–O2 (2:1) [(c) and (d)], N2 [(e) and (f)], and CO2–O2–N2 (2:1:3) [(g) and (h)] gases respectively.

As is known that F is the main element in the SEI composition. The redox of polysulfide on the lithium anode is the main reason of poor property. In order to further identify the protection effect of the gas treatment, we analyzed the F and S element distributions on the three protected samples after cycling (Fig. 8). Figure 8(a) shows the CO₂–O₂ (2:1) treated sample contains more F after one cycle. The N₂ treated sample shows the lowest F amounts. The small peak of red curve in Figure 8(b) shows that S was accumulated in the inner layer of the CO₂–O₂ (2:1)-treated sample, which is consistent with the poor cycle performance. The CO₂–O₂–N₂ (2:1:3)-treated sample shows the lowest S content especially in the inner-layer. It means there is little poly-sulfides decomposed on the CO₂–O₂–N₂ (2:1:3)-treated sample.

Fig. 8.

	Figure Option ViewDownloadNew Window
	Fig. 8. (color online) SIMS depth profile of lithium tablets treated by CO2–O2 (2:1) (red line), N2 (green line), and CO2–O2–N2 (2:1:3) (blue line) gases respectively. (a) F depth profile after cycle; (b) S depth profile after cycle.

Considering the lithium surface process, the possible chemical reaction equations among lithium, N₂, O₂, and CO₂ are given as follows: ; ; . According to the Gibbs energy variation, we can obtain the of each reaction equations. They are −308.3 kJ/mol, −1475.27 kJ/mol, −374 kJ/mol respectively, which means the Li₂CO₃ is easier to nucleate on the lithium surface than Li₃N. Besides, the existence of N₂ decreases the collision probability among the lithium atom, CO₂ and O₂ molecules, and also helps to form Li₃N simultaneously. The competition reaction mechanism finally makes the formation of mixed Li₂CO₃ and Li₃N component film on the lithium tablet.

In summary, we successfully fabricated three kinds of gases treated coating layers on lithium metal tablet: Li₃N, Li₂CO₃, and mixed buffer layer. The coated buffer layers effectively improve the storage and cycling stability of the lithium metal anode. The mixed buffer layer with Li₃N and Li₂CO₃ combines the advantages of both components and shows the best stability and lowest polarization in Li/S cell system. The reaction between Li and N₂ is uneven, and the coverage of Li₃N buffer layer is not so uniform. With the help of Li₂CO₃, a denser and more stable buffer layer is formed. The mixed buffer layer has relative high ion conductivity compared with pure Li₂CO₃ buffer layer. The SEM and SIMS together prove mixed buffer layer has a higher stability than the single component buffer layer during the cycling.

The developed gas protected electrode demonstrates good property and also allows the large scalable production in reality. We believe with the further purification of lithium surface and optimization of gas treatment process, metallic lithium will make further step towards the practical application in lithium batteries.