1 Introduction

Spinel LiMn2O4 has attracted extensive interest as a cathode material for power batteries for electric vehicles and hybrid electric vehicles due to its high voltage, low cost and good safety performance [1,2,3]. However, the capacity fading and short cycle life hamper the cyclability and rate capability of LiMn2O4, especially at elevated temperatures [4]. A major reason is that Mn2+ is readily dissolved from the LiMn2O4 structure and is deposited on the graphite anode due to the presence of HF [5], which is derived from the decomposition of the salt LiPF6 in the presence of trace amounts of water [6,7,8,9]. To stabilize the structure of spinel LiMn2O4, bulk doping and surface coating on LiMn2O4 are widely used to circumvent its instability [10]. However, these methods usually mean the complex synthetic processes and high cost.

By contrast, electrolytes play the equally significant role in the cycling performance and storage performance of lithium-ion batteries [11,12,13,14]. Recently, increasing attention has been focused on optimum electrolytes and additives to suppress capacity fading and improve the cycling performance of LiMn2O4 [11]. Generally, the dosage of additive is less than 5 wt%, but it brings evident improvement to the cycling performance of LiMn2O4 [15]. Se et al. [16] studied LiBOB as an additive for the stabilization of a high-voltage cathode–electrolyte interface and found an improved electrochemical performance of Li/LiNi0.5Mn1.5O4 cells with 1 wt% LiBOB in the conventional electrolyte at 60 °C. Xu et al. [17] used LiDFOB as an electrolyte additive to improve the thermal stability of electrolytes for lithium-ion batteries. It is suggested that a 5 wt% LiDFOB addition can significantly improve the cycling stability at elevated temperatures. The discharge capacity retention increased from 15 to 66.4% after 200 cycles at 60 °C. In this context, LiDFBOP has the similarly chemical structure to LiBOB, but is seldom reported as the electrolyte addictive, especially as the high-temperature one. Moreover, Yang et al. [18] reported that a battery containing 1 wt% LiPO2F2 can form a more conductive and stable SEI film on the graphite anode by the decomposition of LiPO2F2, which led to a low cell impedance.

Generally, LiBOB is used to alleviate the capacity fading at high temperatures, and LiPF2O2 contributes to form a stable SEI, while LiDFOB and LiDFBOP possess similar chemical structures to that of LiBOB. However, comprehensive comparisons of these lithium salt additives in electrolytes for LiMn2O4 systems have rarely been reported. In this paper, effects of LiDFOB, LiBOB, LiDFBOP and LiPO2F2 on LiMn2O4 electrodes were studied systematically, and each was investigated by a combination of charge and discharge curves to evaluate cycling capacity retention under room and high temperatures. Additionally, the surface interactions between LiMn2O4 electrodes and electrolytes containing these additives were analysed by X-ray photoelectron spectroscopy (XPS) and X-ray diffraction (XRD).

2 Experimental

The routine electrolyte was concocted by dissolving 1 M LiPF6 into a (1:1:1 by volume ratio) solvent blend of ethyl carbonate (EC), ethyl (methyl) carbonate (EMC) and diethyl carbonate (DEC), containing 1.5 wt% of VC, which was labelled as the 1# electrolyte. On this basis, 1 wt% of LiDFOB, LiBOB, LiDFBOP or LiPO2F2 was separately added into the 1# electrolyte to prepare 2#, 3#, 4# and 5# electrolytes, respectively. All raw materials of 1#–5# electrolytes mentioned above were purchased from Dongguan Shanshan Battery Materials Co., Ltd, and the chemical structures of these electrolyte additives are shown in Fig. 1. Among these lithium salt additives, at least two B–O bonds exist in the structures of LiBOB and LiDFOB. As for LiDFBOP, which has a similar overall structure to that of LiBOB, P–O bonds exist in its structure. The additive LiPO2F2 was selected mainly to achieve the potency of the P–F bond.

Fig. 1
figure 1

Chemical structures of additives

Full size image

The spinel LiMn2O4 was purchased from the Hunan Shanshan New Material Company. The cathode material consisted of 80 wt% LiMn2O4 powders, 10 wt% acetylene black as a conducting agent and 10 wt% poly(vinylidene fluoride) as a binder. After being blended in N-methyl pyrrolidinone, the mixed slurry was spread uniformly on a thin sheet of aluminium foil and dried in a vacuum for 12 h at 120 °C. The counter and reference electrodes were lithium foil. The cathode electrode was then pressed to a 2.6 g cm−3 compacted density. The LiMn2O4/Li half-cells (CR2032) were assembled in a dry Ar-filled glove box using a polypropylene micro-porous film as the separator, and the 1#–5# electrolytes were added into the cell accordingly.

The charge–discharge tests of these LiMn2O4/Li cells were carried out between 3.0 and 4.3 V by a Land BT-10 Tester (Land, Wuhan, China). The CC–CV mode was applied for the test of the initial charge–discharge with a constant current of 0.1 C to 4.3 V, followed by a constant voltage of 4.3 V until the current decreased to 0.01 C, and discharged to 3.0 V at 0.1 C current rate. Subsequently, the cells were cycled between 3.0 and 4.3 V by 1 C at 25 °C and 55 °C.

A CH instrumental electrochemical work station (CHI660A) was used for cyclic voltammetry (CV) analysis, with a scan rate of 0.1 mV s−1. Electrochemical impedance spectroscopy (EIS) measurements were carried out with working frequencies from 0.01 Hz to 100 kHz.

To analyse the elemental composition and microstructure of the surface of the LiMn2O4 after cycling at a high temperature, the cathode electrodes were separated from the cell and subsequently rinsed with dimethyl carbonate (DMC) in order to remove the residual salts, and then the electrodes were dried under vacuum at room temperature. Afterwards, X-ray photoelectron spectroscopy (XPS, ESCA 2000, ThermoFischer), scanning electron microscopy (SEM, JSM-5600LV, JEOL) and X-ray diffraction (XRD, X’pertPowder, PANalytical) were carried out.

3 Results and discussion

Figure 2 shows the initial charge–discharge profiles of LiMn2O4/Li cells containing five electrolytes between 3.0 and 4.3 V by 1 C at 25 °C. There is no significant difference between the charge and discharge capacities of these five cells. The cell voltage platforms of charge and discharge are almost unchanged, indicating that these additives have a good electrochemical compatibility within the working voltage. The order of discharge capacity is 2# > 3# > 4# > 1# > 5#.

Fig. 2
figure 2

1st charge and discharge curve of LiMn2O4/Li cells with five electrolytes (0.1 C, 3.0–4.3 V and 25 °C)

Full size image

To investigate the reductive and oxidative decomposition of the additives, cyclic voltammograms of the LiMn2O4/Li cells with the five electrolyte additives were conducted (Fig. 3). As shown in Fig. 3a, the CV curves before cycling exhibits two pairs of clearly separated oxidation/reduction peaks, which reflects the typical oxidation processes of LiMn2O4 at 4.10 and 4.21 V (vs. Li/Li+) and reduction processes at 3.96 and 4.08 V [9, 19]. Further, the peak potential shifted to the right direction during the intercalation of lithium, which indicates a varying degree of polarization with the electrolyte additives. The polarization extent follows the sequence of 3# > 2# > 4# > 5# > 1#, which means the different types of SEI film formed.

Fig. 3
figure 3

Cyclic voltammograms of LiMn2O4/Li cells with five electrolytes at a scan rate of 0.1 mV s−1: a before cycling; b after cycling at 25 °C; c after cycling at 55 °C

Full size image

This result is consistent with those of the charge curves (Fig. 2). In addition, it is clearly observed that there are small oxidation peaks at approximately 3.05 V when using electrolyte additives. This phenomenon may be the result of the decomposition of lithium salts on the surface of LiMn2O4, forming an insoluble substance and causing the polarization. However, as indicated by CV after 300 cycles at 25 °C, the small oxidation peaks disappeared, and the peak current intensities for the cells with electrolyte additives, especially with LiBOB and LiDFOB as electrolyte additives, are greater than those for the ones without additives. As the temperature is elevated to 55 °C, some new small oxidation peaks appeared again with LiPO2F2 as the electrolyte additive, which also shows increased irreversible capacity and had the worst capacity retention at high temperature. However, when LiBOB and LiDFOB were used as electrolyte additives, they still display the improved reversibility for the LiMn2O4/Li cells in contrast with others at high temperatures.

Effects of these electrolyte additives on cell rate performance were investigated, and different rate performances of LiMn2O4/Li cells in 3.0–4.3 V at 25 °C were carried out, as shown in Fig. 4. When the discharge rate is less than 2 C, the specific capacities of the LiMn2O4 electrodes in all five electrolytes show very slight differences. By contrast, when the discharge rate increased to 5 C and 10 C, the specific capacities of LiMn2O4 with LiBOB as an additive were 89.8 mAh g−1 and 75.5 mAh g−1, respectively. However, with LiDFOB as an additive, the specific capacities of the LiMn2O4 were 97.5 mAh g−1 at 5 C and 86.5 mAh g−1 at 10 C. Based on the results of cycling, using LiBOB and LiDFOB as additives, the cells exhibit good cycle performance, but additive of LiBOB deteriorates the rate performance of the cell. Compared with the CV of the cells in the presence of LiBOB and LiDFOB additives as shown in Fig. 3, the latter has more intensive peaks, which indicates that the Li+ could be more easily intercalated and extracted from the LiMn2O4 in that case.

Fig. 4
figure 4

Rate performance of LiMn2O4/Li cells with five electrolytes at 25 °C

Full size image

To evaluate the cycle performance of the LiMn2O4/Li half-cells with 1#, 2#, 3#, 4# and 5# electrolytes accordingly, the charge and discharge cycling tests were carried out between 3.0 and 4.3 V by 1 C at 25 °C and 55 °C. Figure 5 shows the cycle performances at 25 °C. It can be observed in Fig. 5 that by using 2#, 3#, 4# and 5# as electrolytes, the discharge capacity retention at room temperature of the cell is superior to that of the routine 1# electrolyte. In 1# electrolyte, the discharge capacity retention was 83.6% after 300 cycles (detailed in Table 1). By contrast, the 3# electrolyte containing LiBOB displays the best effect on the cycling performance of the LiMn2O4/Li cells, which retains 93.2% after 300 cycles. For the other three additives, the discharge capacity retentions are very close, approximately 90% under the same cycle numbers. The severe capacity fading with routine electrolytes maybe due to Mn2+ dissolution, electrolyte decomposition and the Jahn–Teller effect [20]. Comparatively, when adding these lithium salts as electrolyte additives, they may have decomposed and formed a protective layer on the surface of LiMn2O4. In fact, it is the case that the cycling performance of LiMn2O4 at 25 °C can be improved dramatically using LiDFOB, LiBOB, LiDFBOP and LiPO2F2 as electrolyte additives (Table 2).

Fig. 5
figure 5

The cyclic performance of LiMn2O4/Li cells with five electrolytes between 3.0 and 4.3 V by 1 C at 25 °C

Full size image
Table 1 Discharge capacity retention of LiMn2O4/Li cells with five electrolytes between 3.0 and 4.3 V by 1 C at 25 °C
Full size table
Table 2 Discharge capacity retention of LiMn2O4/Li cells with five electrolytes between 3.0 and 4.3 V by 1 C at 55 °C
Full size table

With the cycle temperature elevating to 55 °C, the discharge capacity retention is shown in Fig. 6. It can be observed that the cell with the routine 1# electrolyte displayed 82.6% capacity retention and approximately 17.4% capacity loss by the end of the 300th cycle, while the cells with LiDFOB and LiBOB as electrolyte additives show 84.3% and 86.9% capacity retention, respectively, and less than 1.7% and 4.3% capacity loss after the same cycles. However, with the 4# and 5# electrolytes, the capacity retention is only 78.2% and 76.8%, respectively, which are less than that of the cell with routine electrolyte. Relative to that at 25 °C, the cycling life of the LiMn2O4/Li cells with 2# and 3# electrolytes is clearly improved dramatically at 55 °C. However, compared with the routine electrolyte, LiPO2F2 and LiDFBOP electrolyte additives dramatically decrease the capacity retention at 55 °C. So, only the LiBOB and LiDFOB electrolyte additives improve the cycle performance of the cell at high temperatures. The possible reason may be that LiBOB and LiDFOB decompose and form a cathode protective layer to alleviate electrolyte decomposition and reduce the corrosion of LiMn2O4 during cycling [20]. By contrast, the capacity loss is aggravated severely with other two additives, LiPO2F2 and LiDFBOP, possibly due to their less stable structures and severe decomposition or breakdown.

Fig. 6
figure 6

The cyclic performance of LiMn2O4/Li cells with five electrolytes between 3.0 and 4.3 V by 1 C at 25 °C

Full size image

To further investigate this possible reason and distinct effects of the different additives, the cathode electrodes of LiMn2O4 with different electrolytes after cycling were soaked and cleaned with DMC and analysed by XPS. The C1s, O1s, F1s, P1s and B1s spectra of LiMn2O4 with five electrolytes are shown in Fig. 7.

Fig. 7
figure 7

XPS spectra of LiMn2O4 electrodes after 300 cycles at 55 °C: a C1s, b O1s, c F1s, d P2p, e B1s

Full size image

It can be observed that the C1s spectrum consists of four main characteristic peaks. The first C–C bond at 284.1 eV is attributed to carbon black [15]. The second C–H bond at 286.0 eV is assigned to lithium alkyl carbonates (R–CH2OCO2–Li) or PVDF [15]. The peak at 289.6 eV is attributed to the C=O bond in lithium alkyl carbonates (R–CH2OCO2–Li) or poly carbonates [15]. Finally, the last peak at 290.5 eV corresponds to Li2CO3 [21]. The O1s spectrum shown in Fig. 7b has four main peaks dominated by the Mn–O bond at 529.6 eV from MnxOy or LiMn2O4 [15], the C–O bond at 531.3 eV from R–CH2OCO2–Li or ethers [15], Li2CO3 at 532.0 eV [21], and the C–O bond at 533.3 eV from lithium alkyl carbonates (R–CH2OCO2–Li) or polycarbonates [21]. However, there are no significant differences for the C1s, O1s spectra of the LiMn2O4 with five electrolytes, which indicates that the protection layer mainly consists of organic (R–CH2OCO2–Li) and inorganic (Li2CO3) components.

The F1s spectrum contains four main peaks: LiF (685.3 eV), LiPxOyFz (687.0 eV), PVDF (687.6 eV) and LiPxFy (688.3 eV) [15]. The intensities of LiF for LiMn2O4 with the routine electrolyte and with LiDFBOP electrolyte additive are much greater than those with the other three additives, suggesting that more LiF products covered the surface of the electrode. Comparatively, the peak intensity of LiF with the LiBOB additive is evidently much lower than that of the cathode cycled with other additives. As we already know, LiF, HF and several possible fluorine compounds can be generated via hydrolysis of LiPF6 as Eq. (1) shows. More LiF production corresponds to the more decomposed lithium salt (LiPF6). Moreover, HF can accelerate the corrosion of LiMn2O4 at high temperatures as demonstrated in Eq. (2), so the LiMn2O4 is changed, which results in a poor cycling performance.

$${\text{LiP}}{{\text{F}}_6}+{{\text{H}}_2}{\text{O}} \to {\text{LiF}}+{\text{PO}}{{\text{F}}_3}+2{\text{HF}}$$
(1)
$$2{\text{LiM}}{{\text{n}}_2}{{\text{O}}_4}+4{\text{HF}} \to 3{\text{Mn}}{{\text{O}}_2}({\text{s}})+{\text{Mn}}{{\text{F}}_2}+2{\text{LiF}}+2{{\text{H}}_2}{\text{O}}.$$
(2)

Furthermore, it can be seen from the P2p and B1s spectra that both the P–O bond at approximately 133 eV [15] for LiMn2O4 with the LiPF2O2 and the B–O bond at 191 eV for LiMn2O4 with the LiBOB and LiDFOB additives have more intense peaks than those for the routine electrolyte and others. The BOB derived from LiBOB and the dissolved Mn2+ cations are bonded with each other to form an insoluble and stable surface layer with a network structure, as shown in Fig. 8, and thus to protect the cathode from performance deterioration [20]. The network structure of the protective layer can be formed as shown in Fig. 8. It is inevitable that the solvents are chemically oxidized by the oxygen released from the cathode, generating H2O and CO2, and the resulting H2O further hydrolyses LiPF6 to form acidic products such as HF, which results in the dissolution of LiMn2O4 [20]. Therefore, if the protective layer was formed on surface of LiMn2O4, it prevents the LiMn2O4 from corrosion by HF.

Fig. 8
figure 8

Formation process of protective layer in network structure by the additives a LiBOB and b LiDFOB

Full size image

As discussed above, the cell with LiDFOB exhibits the superior cycling performance at high temperatures, and the improvement is also fairly well with LiBOB. Concurrently, because of the similar chemical structure of LiDFOB and LiBOB, the LiF peak intensity of F1s in Fig. 7c is less than that for the routine electrolyte, and the intensity of the peak associated with the B–O bond is greater than that for the routine electrolyte, which indicates that the B–O bond in the above-mentioned network structure is the key factor governing the high-temperature cycling performance with LiMn2O4. The network structure protective layer on the surface of LiMn2O4 is shown in Fig. 8. However, with LiDFBOP electrolyte additive, the F1s peak intensity for LiF, shown in Fig. 7c, is much greater than that of the others, and with LiPF2O2 as the additive electrolyte, the LixPOyFz bond peak at approximately 687.0 eV [15] and Li2CO3 bond peak at 290.5 eV [21] are the most intense among all these electrolytes, which proves that more inorganic products are on the surface of LiMn2O4. Hence, it is not hard to understand the deterioration of the high-temperature cycling performance with the other electrolytes. Based on the XPS analysis results, an effective protective layer on the surface of LiMn2O4 can be formed, and continuous decomposition of the electrolyte can be abated using LiBOB and LiDFOB as additive electrolytes. In addition, this protective layer on the surface of LiMn2O4 is effectively resistant to Mn2+ dissolution and maintains the spinel structure well.

The EIS of the LiMn2O4/Li half-cells with five kinds of electrolyte are shown in Fig. 9. The impedances are measured before charging (Fig. 9a) and after 300 cycles at 25 °C (Fig. 9b) and 55 °C (Fig. 9c). Generally, the EIS impedance of the cells consists of electrolyte resistance (Re), SEI resistance (Rf) and charge transfer resistance (Rct). From Fig. 9a, it is clear that the interfacial impedances of the cells consist only of Re and Rct without Rf, which demonstrated the protective layer is not formed on the electrodes. By contrast, Fig. 9b, c shows that the interfacial impedances of the cells comprise the Re, Rf and Rct, which indicates that the protective layers on the surface of the electrodes have already formed after cycling at 25 °C and 55 °C. However, after 300 cycles at 25 °C, the magnification of EIS as shown in Fig. 9b displays the array of the results for 3#, 4#, 1#, 5# and 2# electrolytes according to the quantity of the Re of the cells with different additives. Therefore, the Re of the cell with the 3# electrolyte is larger than that of the others, and the one with the 2# electrolyte is the smallest. As shown in Fig. 9c, the impedances increase noticeably after 300 cycles at 55 °C. When the cycle temperature is raised to 55 °C, as seen in the magnification in Fig. 9c, the value of Re decreases in the following order: 2# > 4# > 3# > 1# > 5#. Although the Re of the cell with the 2# electrolyte after cycling at 25 °C increases much less than that of the others, it increases more than that of the others at 55 °C. Probably, the fresh film layer is formed continuously due to the decomposition of LiDFOB, resulting in the film layer becoming much thicker, which can protect the LiMn2O4 electrode but increase Re. As for the 4# electrolyte, the impedance values after cycling both at 25 °C and 55 °C are the second highest, which indicates that the SEI is not stable and that it is difficult to protect the LiMn2O4 electrode. Moreover, the Re will increase due to the accumulation of substances on the surface of the LiMn2O4 electrode formed by continuous side reactions with cycling. Compared to that of the 2# electrolyte, the impedance of the cell with the 3# electrolyte changed much less after cycling at 55 °C because the stable SEI on the surface of LiMn2O4 electrode formed by LiBOB can protect the LiMn2O4 electrode from corrosion. In contrast, with the 1# and 5# electrolytes at 55 °C, the impedances change more apparently than those for the cells cycling at 25 °C. In those cases, the SEI may be not stable, causing the LiMn2O4 electrode to be corroded by the electrolytes.

Fig. 9
figure 9

EIS of LiMn2O4/Li cells with five electrolytes: a before charging at 25 °C; b after cycling at 25 °C; c after cycling at 55 °C

Full size image

To observe the morphology of the electrodes after 300 cycles at 55 °C, SEM was carried out, as shown in Fig. 10. It can be observed that the electrodes consisted of micro-sized LiMn2O4 and nanoscale conductive carbon black. The electrodes of the cells with the 2# and 3# electrolytes show smooth surfaces, similar to that of the pristine electrode, which may not have been corroded by electrolytes; this result proves that the additives LiDFOB and LiBOB have a beneficial effect on the LiMn2O4 electrode at 55 °C. However, on the electrode surfaces of the cells with the 1#, 4# and 5# electrolytes, there are many small, fluffy particles formed by the decomposition of the electrolytes, and moreover the scaly trails on the electrodes demonstrate that the LiMn2O4 electrodes may be corroded by electrolytes, which indicates that the LiDFBOP and LiPO2F2 have a positive effect on the LiMn2O4 electrode at 55 °C.

Fig. 10
figure 10

SEM images of the LiMn2O4 electrode before charging and after 300 cycles by 1 C at 55 °C: a, b pristine, c, d with routine electrolyte, e, f with LiDFOB, g, h with LiBOB, i, j with LiDFBOP, k, l with LiPO2F2

Full size image

To the verify spinel structure of LMn2O4 after cycling at high temperatures, XRD was carried out, as shown in Fig. 11. Compared with those of the pristine electrode, the diffraction peak intensities of the LiMn2O4 electrode with the routine electrolyte were weaker than those with the four types of electrolyte additives. The diffraction peak intensities of LiMn2O4 with LiDFOB and LiBOB were greater than those of the others, which may be related to the less extent of decomposition of the LiMn2O4 spinel structure. In addition, by magnifying the region between 17° and 37° for 2θ as shown in Fig. 11b, it can be seen that both the diffraction peaks from (111) and (311) planes of LiMn2O4 with LiDFOB and LiBOB barely shifted to lower angles after high-temperature cycling as compared to those with other electrolytes, indicating no evident expansion of the crystal lattice [4]. Based on the XPS and XRD analysis results, LiBOB and LiDFOB electrolyte additives can form an effective protective layer on the surface of LiMn2O4 and this layer can be formed and abated through continuous decomposition of the electrolyte. In addition, this protective layer on the surface of LiMn2O4 effectively prevents Mn2+ dissolution and maintains the spinel structure well. It is no wonder that the high-temperature cycling life of the LiMn2O4/Li cell is improved with LiBOB and LiDFOB electrolyte additives.

Fig. 11
figure 11

XRD patterns of LiMn2O4 electrode before charging and after 300 cycles at 55 °C (a) and the magnified region between 17° and 37° of 2θ (b)

Full size image

Combined with the cycling performances at high temperatures and the different rates of performance, the high-temperature cycling performances are evidently improved with LiBOB and LiDFOB salts in the electrolytes because they form a protective layer on the surface of LiMn2O4. Moreover, using LiBOB as an electrolyte additive shows more excellent high-temperature cycling performance than that of the other additives. If rate performance is considered, among all these additives, only LiDFOB shows not only the improvement of cycling performance, but also the better rate performance at high temperature as compared with the routine electrolyte.

4 Conclusions

This work demonstrates that the LiDFOB and LiBOB electrolyte additives improve the cycling performance of a LiMn2O4 electrode at 55 °C. In contrast, using LiDFBOP and LiPO2F2 as electrolyte additives have the negative effects on the cycling performance at 55 °C. Furthermore, XPS clearly shows that the bonding of BOB− anion and Mn2+ favours the formation of an insoluble and stable surface layer with a network structure on the surface of the LiMn2O4 electrodes when LiDFOB and LiBOB are used, which prevents decomposition of the lithium salt electrolyte (LiPF6). Furthermore, the protective layer effectively retards the continuous dissolution of Mn2+ and maintains the spinel structure well. Among all 4 additives, only LiDFOB improves not only the cycling performance but also the high-temperature rate performance as compared with the routine electrolyte.