Introduction

Lithium-ion batteries (LIBs) are widely used in electric vehicles and hybrid vehicles due to their excellent energy storage characteristics [1,2,3,4,5]. However, facing the industry’s pursuit of lower cost and higher energy density in the future, the existing commercial cathodes such as LiFePO4 and LiCoO2 cannot meet the needs [6,7,8]. Lithium-rich layered oxides (LLO) xLi2MnO3·(1-x)LiMO2 (M = Ni, Co and Mn), which exhibit extremely high energy density (> 250 mAh g−1) owing to their higher lithium content and the participation of lattice oxygen in the charge and discharge process, are widely considered as one of the candidate cathode materials for next-generation LIBs. Among them, cobalt-free lithium-rich materials with the advantages like lower cost and others have attracted much attention [9,10,11,12,13]. But because of incomplete utilization during crystallization by roasting, part of excessive lithium will react with CO2 or H2O in the air to form Li2CO3 or LiOH when it sticks to the surface [14]. On the one hand, it leads to the waste of lithium resources; on the other hand, the CO2 generated by the decomposition of Li2CO3 starting from 3.8 V (vs. Li/Li+, same below) promotes the formation of a poorly conductive solid electrolyte interphase (SEI) film, which will further weaken the ionic and electronic conductivity of particles and decrease the electrochemical performance of LLO [15,16,17]. They both hinder the commercial application of lithium-rich materials.

The surface coating strategy can achieve the goal of weakening the influence of Li2CO3 and LiOH by building a functional film on the surface of the particle to isolate the surface from the electrolyte [18,19,20,21,22]. However, this method ignores the value of residual lithium by treating it simply as the burden of improving material performance. Considering that LiOH and Li2CO3 form a thin coating on the surface of the particles, it is feasible to play a positive effect on the residual lithium by functionalizing the layer [23]. SO3 can react with LiOH or Li2CO3 to form an excellent lithium-ion conductor Li2SO4, while reducing the pH value of the particle surface, which is an ideal material for transformation. It has been reported that Li2O under high temperature conditions is preemptively reacted to form Li2SO4 by mixed gas composed of SO2 and O2, but there exists a small amount of insufficiently oxidized Li2SO3 and Li2S2O4 because of the influence coming from reaction environment [14]. In order to avoid the interference of by-products, the calcination process will be carried out under high temperature oxidizing environment. Thus, it is still challenging to develop simple and effective approaches to realize the compositional control of the functionalized surfaces for LLO. More importantly, studies on the effect of in situ formed functionalized surfaces on the composition and structural evolution of LLO during the cycling process are warranted.

Herein, we report a simple strategy to utilize the residual lithium compounds on the surface of Co-free LLO oxides. Through the substitution reaction of SO3 with Li2CO3 and LiOH under high temperature oxidation environment and the deposition of SO3, a functional layer rich in Li2SO4 can be built on the surface of a typical Co-free LLO with the chemical formula Li1.2Mn0.6Ni0.2O2. Based on the composition/structure characterizations and cyclic evolution analysis, it is found that the in situ generated functional surface layer can effectively perform three functions: the first is to reduce the content of Li2CO3 and LiOH on the surface of the particles and form a new covering layer; the second is to guide the evolution of particles’ crystal structure; the third is to exert its excellent characteristics of conducting lithium ions and inhibit the generation of poor electronic conductors. As expected, the designed LLO cathode with a sulfur-containing surface layer exhibits a larger first coulombic efficiency of 84.4% (in contrast to the original sample of 81.0%), superior cycle capacity retention of 93.3% after 300 cycles at 250 mA g−1 (while the original sample is only 78.9%), and excellent rate capabilities. This strategy provides a new solution for the plight of the commercial application of Co-free LLO in LIBs.

Experimental

The Co-free Li-rich layered oxide Li1.2Mn0.6Ni0.2O2 was prepared by co-precipitation. SO3 gas comes from the thermal decomposition of K2S2O8, reaction equation as follows:

$${2\mathrm{K}}_{2}{\mathrm{S}}_{2}{\mathrm{O}}_{8}\left(\mathrm{s}\right)\to {2\mathrm{K}}_{2}{\mathrm{SO}}_{4}\left(\mathrm{s}\right)+{2\mathrm{SO}}_{3}\left(\mathrm{g}\right)+{\mathrm{O}}_{2}\left(\mathrm{g}\right)$$

The reaction process as shown in Fig. 1, 600 ℃, is set in order to obtain a surface with low LiOH or Li2CO3 content, while taking out in 100℃ for reducing the total amount of SO3 deposition. The degree of modification is controlled by the mass ratio of active substance and K2S2O8 (keep the active substance at 0.25 g, S-0.5 represents K2S2O8 mass at 0.125 g, while S-1.0 and S-2.0 are 0.25 g and 0.50 g, respectively). In addition, for eliminating the interference of influence factors like modification temperature, the comparison sample selects S-0 and named Reference.

Fig. 1
figure 1

Schematic diagram of sulfur modification process of the Co-free LLO

Full size image

The specific details of Li1.2Mn0.6Ni0.2O2 preparation and sample characterization are shown in the Supporting Information.

Results and discussion

Theoretical calculations were used to verify whether the reaction occurs as we expected. According to the standard molar Gibbs free energy of formation (ΔfGΘm) (298 K) data (Supporting Information, Table S1), the ΔfGΘm changes of the main reaction processes that may occur as follows:

$$\begin{array}{cc}{\mathrm{Li}}_{2}{\mathrm{CO}}_{3}\left(\mathrm{s}\right)+{\mathrm{SO}}_{3}\left(\mathrm{g}\right)\to {\mathrm{Li}}_{2}{\mathrm{SO}}_{4}\left(\mathrm{s}\right)+{\mathrm{CO}}_{2}\left(\mathrm{g}\right)& {\Delta }_{\mathrm{f}}{\mathrm{G}}_{\mathrm{m}}^{\Theta }=215.9 \mathrm{kJ} {\mathrm{mol}}^{-1}\end{array}$$
$$\begin{array}{cc}2\mathrm{LiOH}\left(\mathrm{s}\right)+{\mathrm{SO}}_{3}\left(\mathrm{g}\right)\to {\mathrm{Li}}_{2}{\mathrm{SO}}_{4}\left(\mathrm{s}\right)+{\mathrm{H}}_{2}\mathrm{O}\left(\mathrm{g}\right)& {\Delta }_{\mathrm{f}}{\mathrm{G}}_{\mathrm{m}}^{\Theta }=304.5 \mathrm{kJ} {\mathrm{mol}}^{-1}\end{array}$$
$$\begin{array}{cc}{\mathrm{Li}}_{2}{\mathrm{SO}}_{3}\left(\mathrm{s}\right)+1/2{\mathrm{O}}_{2}\left(\mathrm{g}\right)\to {\mathrm{Li}}_{2}{\mathrm{SO}}_{4}\left(\mathrm{s}\right)& {\Delta }_{\mathrm{f}}{\mathrm{G}}_{\mathrm{m}}^{\Theta }=232.7 \mathrm{kJ} {\mathrm{mol}}^{-1}\end{array}$$
$$\begin{array}{cc}{\mathrm{Li}}_{2}\mathrm{O}\left(\mathrm{s}\right)+{\mathrm{SO}}_{3}\left(\mathrm{g}\right)\to {\mathrm{Li}}_{2}{\mathrm{SO}}_{4}\left(\mathrm{s}\right)& {\Delta }_{\mathrm{f}}{\mathrm{G}}_{\mathrm{m}}^{\Theta }=391.7 \mathrm{kJ }{\mathrm{mol}}^{-1}\end{array}$$

The results show that not only Li2CO3 and LiOH, but also the Li2SO3 and Li2O, produced during the treatment process tends to combine with SO3 or O2 gas to form Li2SO4. This result is also demonstrated in the powder X-ray diffraction (XRD) patterns shown in Fig. 2a and b. No obvious shift is found in each peak after sulfur modification, indicating that the sulfur layer does not significantly affect the particles’ crystal structure. What can be also observed is that the diffraction peak intensity of Li2SO4 is positively correlated with the total amount of K2S2O8 in the part where the degree of incident angle 2θ is 20–30, indicating that our sulfur treatment method and the regulation strategy of Li2SO4 production are effective. It is worth noting that if there have no special treatment steps, the strategy of creating a specific gas atmosphere for the gas–solid reaction will inevitably expend the gas-phase reactants that maintain the gas concentration. Especially in an open system, the limited sulfur source will be consumed more to create the reaction environment. The amount of the consumed part and the sulfur source does not always maintain a fixed ratio, which explains why the intensity of the characteristic peak of Li2SO4 cannot be maintained at twice the ratio between S-1.0 and S-0.5 as shown in Fig. 2b (the intensity of S-0.5 is less than 1/2 of S-1.0).

Fig. 2
figure 2

XRD patterns with the degree of incident angle 2θ are (a) 10–90 and (b) 20–30, SEM images of samples (c, e) Reference, and (d, f) S-modified

Full size image

A scanning electron microscope (SEM) was used for observing the micron-level morphology of the material before and after sulfur modified, and the results are shown in Fig. 2c, d, e, and f. There is no significant difference in the integral and surface morphology of the particles; the reason is that the SO3 gas that comes from a very limited K2S2O8 decomposition process cannot cause significant changes in particles surface.

Transmission electron microscopy (TEM) was used to confirm the coating state of the particle surface. As shown in Fig. 3a and d, there is a flat thin layer form on the surface of the primary particle after treatment. Observing the edge area of the particles with a high-resolution transmission electron microscope (HRTEM) (Fig. 3b and e), it is obvious that the (003) interplanar spacing of the two particles has increased from the usual 0.47 to about 0.49 nm [24,25,26], which is related to the high processing temperature. But the difference between the two groups of samples is very small, which means that the SO3 substitution reaction does not affect the lattice spacing of the particles, which matches the XRD result. After measuring the thickness of the coating layer at different positions, the range is determined to be between 2.0 and 3.0 nm. In the fast Fourier transform (FFT) images of the red frame area containing the thin layer (Fig. 3c and f), it can be observed that (-111) (4.1 Å) and (201) (3.2 Å) crystals related to Li2SO4 appear in the S-modified sample [14], what proves the presence of Li2SO4 in the coating. Since the production of Li2SO4 is limited by the total amount of SO3, the composition of the thin layer is more complicated and requires further analysis.

Fig. 3
figure 3

TEM, HRTEM, and FFT images of (a, b, and c) Reference and (d, e, and f) S-modified, (g) STEM-EDS results of S-modified

Full size image

In addition, the energy dispersive X-ray spectroscopy image in the scanning transmission electron microscope mode (STEM-EDS) (Fig. 3g) was also used to analyze the processed particles to determine the distribution of S, Mn, Ni, and O elements in the particles. The conclusion that the sulfur-containing mixed layer effectively covers the particle surface can be obtained when the SEM-mapping images (Supporting Information, Fig. S1) were used to analyze together.

X-ray photoelectron spectroscopy (XPS) is used for detailed analysis of the sample surface after SO3 treatment, and the full spectrums of Reference and S-modified are shown in Fig. 4a. In the core spectra of S 2p (Fig. 4b), the peak position at the binding energy of 169.2 eV indicates that the sulfur on the surface is in the form of SO42− instead of S2− or SO32− [27]. And it is also confirmed in the O 1 s core spectrum (532.2 eV, including Li2SO4 and transition metal sulfate) as shown in Fig. 4c. These results fully demonstrate the effectiveness of this sulfur treatment strategy (the seemingly unchanged intensities of OH− (531.0 eV, mainly LiOH) and CO32− (531.5 eV, mainly Li2CO3) can be attributed to the test error of the two samples) [14, 27,28,29]. In addition, the peak at the binding energy of 533.3 eV in the O 1 s spectrum may correspond to surface adsorption. Besides, it should be noted that there is a small amount of SO42− and SO32− on the surface before treatment, which is related to the use of sulfate when preparing the precursor. But SO32− is almost non-existent after sulfur modification, indicating that the treatment strategy of eliminating by-product in an air environment is effective.

Fig. 4
figure 4

XPS spectrums of (a) full scan, (b) S 2p, (c) O 1 s, (d) Mn 2p, and (e) Ni 2p. (f) The relative percentages of different valence states of transition metals before and after treatment

Full size image

In the core spectra of Mn 2p and Ni 2p (Fig. 4d and e), both the height of Mn4+ (643.0 eV) and Ni3+ (856.3 eV) peak [30, 31] rise after SO3 treatment, especially Mn4+, which means that part of the transition metal on the surface reacted. The changes in the relative percentages of Mn3+/Mn4+ and Ni2+/Ni3+ before and after treatment confirm it. In a word, it can be considered that the main components of the thin layer on the surface of the treated particles are Li2SO4, Li2CO3, LiOH, and transition metal sulfate.

Generally, the surface treatment will affect the change process of samples’ crystal structure during the cycle. Therefore, exploring their changing tendency and the state of surface layer (not only sulfur modified membrane, but also cathode electrolyte interphase (CEI) film) under different cycle periods will help to understand the function of the coating layer after treatment and materials’ electrochemical behavior in more detail. The finished half-cells need to go through three processes successively: be placed for 6 h, then activate 3 cycles with a current density of 0.1 C (1 C = 250 mA g−1), and cycle at 1 C at last. The XRD test results of the two groups of cathode pieces under different conditions are shown in Fig. 5a and e. It is observed that the intensity of the lithium-rich characteristic peak of S-modified is always weaker than Reference in every state, which means that the phase evolution in S-modified particles performs more significantly. Moreover, the weakening of the peak intensity corresponding to the (003) crystal plane mainly occurs in activation process and first cycle at 1 C, which indicates that the layered structure has undergone major changes at this stage. In the enlarged XRD images with an incident angle 2θ of 20 ~ 30°, the characteristic peak of Li2SO4 in Reference that gradually increases can be attributed to the residual SO42− and combines with Li+ during the cycle (Fig. 5b). Nonetheless, the Li2SO4 peak of S-modified sample increases significantly during the activation and first cycle at 1C, and then remained basically unchanged (excluding the Li2SO4 existing in the particles) (Fig. 5f). The fact indicates that there is a significantly new combination process of Li2SO4 formed by SO42− and Li+ at this stage. The new reaction process may be related to the presence of sulfate ions that are not bound to lithium on the surface. In addition, a detailed analysis of the XRD test results shows that the two groups of samples exhibit different phase evolution trends during the cycle (Fig. 5c, d, g, and h). For example, it can be seen that the peaks correspond to the crystal planes of (101) and (103) and weaken faster after sulfur treatment. And there is also the appearance and disappearance of peaks near the two characteristic peaks. The more violent phase evolution degree of the S-modified sample may suggest that its crystal structure has changed greatly compared with the initial state.

Fig. 5
figure 5

XRD patterns in different states of (a, b, c, and d) Reference and (e, f, g, and h) S-modified, TEM, and HRTEM images after 300 cycles of (i, j) Reference and (k, l) S-modified

Full size image

The changes in the surface state and crystal structure of Reference and S-modified after 300 cycles at 1 C can be clearly observed in TEM images. In Fig. 5i, the CEI film can be clearly seen on the particle surface of Reference, while the surface coating layer of the sulfur-treated sample (Fig. 5k) has a significant thickness increase (from 2.0 ~ 3.0 to 20.0 ~ 30.0 nm). On the one hand, this phenomenon can be partly attributed to the combination of SO42− and Li+ during cycles; on the other hand, it is related to the formation of CEI film on the surface of the Li2SO4 coating layer. The images of HRTEM and the local FFT are shown in Fig. 5j and l. After cycles, Reference still maintains a layered structure with clearly visible lattice fringes of (003) crystal planes, while the main crystal plane of sulfur-treated sample cannot be easily observed after significant phase evolution. These results correspond to changes in peak intensity in XRD.

Figure 6a shows the complete charge and discharge curves about first cycle of Reference, S-0.5, S-1.0, and S-2.0 at 0.1 C. Partially, pre-oxidized of transition metal especially Ni2+ during the sulfur modification process can be observed in the XPS spectrum, and it is shown in the charging curves in the form of reduction in the capacity contributed by them (enlarged image in blue frame as shown in Supporting Information Fig. S2) [32,33,34]. Sulfur can combine with lattice oxygen during the delithiation process to inhibit the irreversible oxygen release as reported, which is manifested as a shortened platform length near 4.5 V in the first charging curves (Supporting Information Fig. S3) [27]. All in all, moderate sulfur modification helps to improve the first coulombic efficiency (FCE) of the particles (Reference is 81.0%, while S-0.5 and S-1.0 are 85.7% and 84.4%, respectively). Moreover, it also helps to improve the electrochemical stability of the material (Fig. 6b and Fig. S4 in Supporting Information). After the phase evolution process is relatively stable, the capacity released by S-1.0 changes from rapid decay to gradual increase (the reason why S-0.5 sample shows different tendency can be attributed to a low degree of vulcanization). Among them, the capacity retention rate of S-1.0 after 300 cycles is 93.3% (Reference is 78.9%).

Fig. 6
figure 6

a Initial charge–discharge curves, b cycle performance for 300 cycles at 1 C, charge–discharge profiles, and corresponding dQ/dV curves of (c, e) Reference and (d, f) S-1.0, and (g) dQ/dV profiles at the 300th cycle at 1 C

Full size image

Figure 6c and d show the charge and discharge curves of Reference and S-1.0 at different cycles, and the corresponding differential capacitance (dQ/dV) curves are shown in Fig. 6e and f. The new charging peak of the S-1.0 sample near 3.15 V also contributes the capacity of a similar value during the discharge process, which shows good reversibility. This explains why the generation of new phase will increase the capacity during charging and discharging. Meanwhile, only S-1.0 and S-2.0 samples with a unique capacity change law have new peaks near 3.15 V (charge) and 2.8 V (discharge), which further proves our inference on the reason for the capacity increase of materials (Fig. 6g).

In addition, it is obvious that new reaction has the fastest growth in the process of rapid phase evolution. Combining XRD and TEM images (Fig. 5c, d, g, h, j, and l), we can associate the rapid decay of the discharge-specific capacity of S-1.0 and S-2.0 in the initial stage with the violent evolution of the crystal structure in the process. Both of them are related to the activation process of the material under high current (1 C), as observed in other experiments [35,36,37,38]. That is, the processed sample cannot release the capacity that can be released in a stable state in the first cycle after the current density increases. The materials need to experience several cycles to reach a stable state at the new current density which is named activation process. Comparing with other treatment methods, the stable state of the sample after SO3 treatment shows a unique varying law of gradual increase in capacity. When the activation process is complete, new redox reactions slowly increase, and the phase evolution of the sample is no longer intense. Considering that the S-0.5 sample with low vulcanization degree did not show a similar change process, it can be deliberated that moderate sulfur treatment will change the initial stable state of the sample under high current and endows the sample with a more robust crystal structure in the subsequent cycle process.

The rate performances of four samples at rates of 0.1, 0.2, 0.5, 1, 2, and 5 C are shown in Fig. 7a. At low rates, the capacity decreases with the sulfur increases, which can attribute to a higher resistance caused by coating layer to delithiation process at low rates than positive effects to Li+ diffusion. At high rates like 1, 2, and 5 C, the gaps between S-modified and Reference are reduced, while the S-1.0 sample is significantly higher than Reference, suggesting that the contrast has reversed [39]. This partly explains why there is almost no difference in cycle performance between Reference and S-1.0 at 2 C (Supporting Information, Fig. S5). To verify the point about Li+ kinetics, we investigated the Li+ diffusion coefficient (\({D}_{{Li}^{+}}\)) values of Reference and S-1.0 by using galvanostatic intermittent titration technique (GITT) and the following calculation formula:

Fig. 7
figure 7

a Rate performances of four samples, b a representative single titration at 4.04 V during GITT measurement, GITT discharge test profiles, and corresponding \({D}_{{Li}^{+}}\) of (c, e) Reference and (d, f) S-1.0 after 1st cycle and 100th cycle

Full size image
$${D}_{{Li}^{+}}=\frac{4}{\pi \tau }{\left(\frac{{m}_{B}{V}_{M}}{{M}_{B}S}\right)}^{2}{\left(\frac{{\Delta E}_{s}}{{\Delta E}_{\tau }}\right)}^{2}$$

among them, mB, MB, VM, and S stand for the molecular weight, molecular mass, molar volume of LLO, and active surface area, respectively. Figure 7b shows the \({\Delta E}_{s}\) and \({\Delta E}_{\tau }\) of the GITT discharge process, while the GITT discharge test profiles of Reference and S-1.0 after 1st cycle and 100th cycle at 1 C as shown in Fig. 7c and d, respectively. The \({\Delta E}_{s}\) and \({\Delta E}_{\tau }\) values of GITT curves are shown in Table S2a and b (Supporting Information, Table S2), and the \({D}_{{Li}^{+}}\) calculation results obtained based on them are shown in Fig. 7e and f. After 1st cycle at 1 C, the difference of crystal structure between two samples is not significant, which means that the significantly higher \({D}_{{Li}^{+}}\) values of S-1.0 can be attributed to the surface functional layer. Although the difference of \({D}_{{Li}^{+}}\) between two samples shrinks after 100th cycle, S-1.0, which has undergone a crystal structure evolution, still has some advantages over Reference which maintains a layered structure. Both facts highlight the significant promotion effect of the functional layer on the diffusion of lithium ions.

In the LLO system, irreversible extraction of oxygen from particle lattice will induce the carbonate-based electrolyte to generate CO2 and other substances, resulting in the formation of a poorly conductive SEI film in the electrolyte/electrode interface and increasing the values of charge transfer resistance (\({R}_{ct}\)) [16]. In fact, Li2CO3, the reaction product of the remaining lithium on the surface, can also decompose itself during the cycle and cause similar results [17]. Surface sulfur treatment has been proven to effectively reduce the production of O2 and CO2 [14]. Therefore, it can be expected that the sulfur-modification samples will have lower \({R}_{ct}\) values during cycling. To verify it, electrochemical impedance spectra (EIS) were used to provide a view about the resistance of Reference and S-1.0 under different cycles at 1 C and the results are shown in Fig. 8a (all half-cells are charged to a voltage of 4 V before testing). The curve is consisted of two consequent semicircles and one pitched line while they all represent different meanings: \({R}_{\Omega }\) is expressed as the intercept of the first semicircle while the semicircle itself represents the resistance \({R}_{f}\) (Li+ migration in SEI and surface-modified layer) in the high frequency region; the second semicircle represents the resistance \({R}_{ct}\) in the middle frequency region; the bottom line in the low-frequency means Warburg impedance \({Z}_{w}\) which relates to the diffusion of Li+ in the particle bulk [18, 40, 41]. The \({R}_{ct}\) values and the equivalent circuits obtained by fitting the impedance diagram are shown in Fig. 8b. What can be clearly observed is that the S-1.0 has a slower \({R}_{ct}\) values and growth rate, which is in line with expectations.

Fig. 8
figure 8

a EIS profiles of Reference and S-1.0 after different cycles. b \({R}_{ct}\) values of Reference and S-1.0 after different cycles and the equivalent circuits

Full size image

Conclusion

In summary, a strategy for in situ utilization of surface residual lithium to form a functionalized sulfur-containing surface coating layer has been proposed for Co-free LLO. This in situ generated functional surface layer has been constructed by designated consumption of the Li2CO3 and LiOH through a simple sulfur treatment. Benefit from the lower charge transfer resistance, higher Li+ diffusion coefficient, and more stable evolutionary crystal structure, the LLO sample with the in situ functionalized sulfur-containing surface layer exhibits a higher first coulombic efficiency of 84.4% (in contrast to the original sample of 81.0%), excellent rate capabilities, and superior cycle capacity retention of 93.3% after 300 cycles at 250 mA g−1 (while the original sample is only 78.9%). These results reveal that the design of an in situ generated functional surface layer can effectively enhance the electrochemical performance and might provide a new solution for the plight of the commercial application of Co-free LLO in LIBs.