Introduction

Lithium-ion battery has been widely used in electronic devices such as cell phones, digital cameras, and laptops, owing to its performances superior to other secondary batteries, but its energy density needs to be improved for the large-scale applications such as in electric vehicles [1, 2]. The energy density of lithium-ion battery can be improved by using high-voltage cathodes. Spinel-structured LiNi0.5Mn1.5O4 is one of the most promising candidates because it provides higher working voltage (4.75 V) [3, 4] than commercial cathode materials such as LiFePO4 (3.4 V) [5], LiCoO2 (3.9 V) [6], and LiMn2O4 (4.1 V) [7, 8]. However, there are issues that need to be overcome before LiNi0.5Mn1.5O4 is put into use in practice: (1) the dissolution of transition metal ions in LiNi0.5Mn1.5O4 into electrolyte, especially at elevated temperature [9–12], (2) the decomposition of electrolyte on cathode under high voltage [13–16], (3) the crystallographic distortion of LiNi0.5Mn1.5O4 to Jahn-Teller effect [3, 4, 17], and (4) the low electronic conductivity of LiNi0.5Mn1.5O4 [18]. These issues lead to poor cyclic stability and rate capability of LiNi0.5Mn1.5O4 [19].

Coating with oxides on and doping transition metals in LiNi0.5Mn1.5O4 have been applied to improve the cyclic stability of LiNi0.5Mn1.5O4, but the improvement is not satisfying and at the expense of a capacity loss [20–23]. New solvents such as sulfone were used to replace the conventional carbonate solvents to improve the electrolyte stability, but these solvent have high viscosity and do not facilitate the formation of SEI on graphite [24–26]. Electrolyte additives are found to be effective to suppress the electrolyte decomposition and prevent cathode from destruction, but the use of additives usually causes the increase in interfacial impedance, which is detrimental to the rate capability of the cathode [27]. Nano-size particles have been demonstrated to be able to improve the rate capability of LiNi0.5Mn1.5O4, but the synthesis is complicated [28].

Thiophene has been used as an electrolyte additive to improve the rate capability of LiCoO2 [29], LiNi1/3Co1/3Mn1/3O2 [30], LiCoPO4 [31], and Li3V2(PO4)3 [32] due to the electronic conductivity of polythiophene that is formed during charging process. In this work, we reported a new finding that thiophene was able to improve simultaneously the cyclic stability and the rate capability of LiNi0.5Mn1.5O4 cathode. Electrochemical measurements and physical characterizations indicated that protective cathode film and conductive polymer were formed when thiophene was used, which contributed to the improved cyclic stability and rate capability.

Experimental section

Preparation

Battery-grade carbonate solvents and lithium hexafluorophosphate (LiPF6) were provided by Guangzhou Tinci Materials Technology Co., Ltd., China, and used without further purification. Thiophene was purchased from Aladin (Wuhan, China). LiPF6-containing ethylene carbonate (EC) (1.0 M) /dimethyl carbonate (DMC) (1:2, v/v) electrolytes with various contents of thiophene (0, 0.2, 0.5, and 1 wt%) were prepared in a highly pure argon-filled glove box (MBraun, Germany), in which the water and oxygen contents were less than 0.1 ppm. The contents of water and free acid (HF) in the electrolytes were determined by Karl Fischer 831 Coulometer (Metrohm, Switzerland) for H2O and Karl Fischer 798 GPT Titrino (Metrohm, Switzerland) and controlled lower than 20 and 50 ppm, respectively. The cathode was prepared with the mixture of 80 wt% LiNi0.5Mn1.5O4 (Sichuan Xingneng Co., Ltd., China), 10 wt% polyvinylidene fluoride (PVDF) binder, 5 wt% acetylene black, and 5 wt% Super P in N-methyl-pyrrolidone (NMP). The mixture slurry was coated on Al foil and dried at 85 °C for 2 h and then dried at 120 °C for 6 h under in vacuum. Li/LiNi0.5Mn1.5O4 2025-coin cells were assembled using Celgard 2400 as separator in the argon-filled glove box for the evaluation of electrochemical performances.

Electrochemical measurements

Charge/discharge test was performed on a Land cell test system (Land CT2001A, China). Cells were cycled between 3 and 4.95 V (vs. Li/Li+) at 25 or 55 °C with a charging protocol of constant current (CC) followed by a constant voltage (CV) step at 4.95 V till the current decreased to one tenth of the CC (1 C = 147 mA g−1). The cyclic voltammetry was performed on Solartron-1470E at 25 °C between 3.5 and 4.95 V at a sweep rate of 0.1 or 0.2 mV s−1. The electrochemical impedance of the cells was measured on Autolab (PGSTAT302N) over a frequency range of 105 to 0.01 Hz with the amplitude of 5 mV r.m.s−1. To make the results reproducible, all the electrochemical measurements were done with five samples; the reported value in this work was the average one with the error less than 5 %.

Physical characterization

The LiNi0.5Mn1.5O4 electrodes after cycling were rinsed with anhydrous DMC for three times to remove residual EC and LiPF6 salt precipitated on the surface and then evacuated overnight at room temperature before ex situ analysis. The morphology of LiNi0.5Mn1.5O4 was observed by scanning electron microscope (SEM, JSM-6380, Japan) and transmission electron microscopy (TEM, JEOL JEM-2100HR). The crystal structures of the materials were analyzed by X-ray diffraction (XRD, Bruker D8 ADVANCE, Germany) with Cu Kα radiation. The chemical composition of the electrode surface was determined by X-ray photoelectron spectroscopy (XPS, ESCALAB 250), using focused monochromatized Al Kα radiation under an ultrahigh vacuum, and Fourier-transformed infrared spectra (FTIR, Nicolet 6700) with attenuated total reflection (ATR) mode.

Results and discussions

Figure 1a compares the cycling stability of LiNi0.5Mn1.5O4 cathodes in the base and thiophene-containing electrolytes at 1-C charge/discharge rate at 25 °C. In base electrolyte, LiNi0.5Mn1.5O4 delivers an initial capacity of 124.2 mAh g−1 and keeps stable before 200 cycles. After the 200th cycle, however, the capacity decaying appears, resulting in capacity retention of 70.5 % after 400 cycles. When thiophene is used, the initial capacities and the capacity retention increase with increasing content of thiophene and reach the maximum values of 132.1 mAh g−1 and 89.2 %, respectively, for the electrolyte containing 0.5 wt% thiophene, indicating that thiophene can be used as electrolyte additive to improve the cyclic stability of LiNi0.5Mn1.5O4. Higher thiophene content does not yield better effect, as shown by the behavior of the electrolyte containing 1 wt% thiophene in Fig. 1a, which might be ascribed to the reduced ionic conductivity of the electrolyte due to the use of thiophene. Therefore, further investigations were performed with the electrolyte containing 0.5 wt% thiophene.

Fig. 1
figure 1

Cycling performance of LiNi0.5Mn1.5O4 in base and thiophene-containing electrolytes at 25 °C (a) and 55 °C (b) with 1-C current rate, coulombic efficiency at 55 °C (c), and rate capability (d)

Full size image

The improvement in cyclic stability becomes more significant, when LiNi0.5Mn1.5O4 is cycled at elevated temperature. As shown in Fig. 1b, LiNi0.5Mn1.5O4 in the base electrolyte experiences a serious capacity decaying, with capacity retention of only 37.2 % after 100 cycles at 55 °C. This should be ascribed to the electrolyte decomposition and the destruction of LiNi0.5Mn1.5O4. When applying 0.5 wt% thiophene into the base electrolyte, the discharge capacity retention of LiNi0.5Mn1.5O4 is significantly improved, to 78.8 %, suggesting that the electrolyte decomposition can be suppressed and LiNi0.5Mn1.5O4 can be protected by the application of thiophene. In fact, the electrode with thiophene has higher coulombic efficiency than that without thiophene, as shown in Fig. 1c.

The improved initial capacity of LiNi0.5Mn1.5O4 by using thiophene can be ascribed to the electronic conductivity of polythiophene formed from the oxidation of thiophene under charging process, which should benefit the rate capability. Figure 1c compares the rate capability of LiNi0.5Mn1.5O4 in the electrolytes with and without 0.5 wt% thiophene, which was obtained by charging at 0.5 C and discharging at various current rates. It can be seen from Fig. 1d that LiNi0.5Mn1.5O4 in thiophene-containing electrolyte exhibits better rate capability than that in base electrolyte. The discharge capacity at 10 C retains 85.2 % of that at 0.5 C for LiNi0.5Mn1.5O4 in thiophene-containing electrolyte, while only 66.4 % for that in the base electrolyte. The improved rate capability confirms that the application of thiophene contributes to the improved electronic conduction of the cathode.

In order to understand the effect of thiophene on the cyclic stability and the rate capability of LiNi0.5Mn1.5O4, the cycled electrodes were characterized with CV, EIS, SEM, and XRD. Figure 2a, c presents the cyclic voltammograms of LiNi0.5Mn1.5O4 at the selected cycles. It can be seen that the charging/discharging capacity decreases for both electrodes, but this decreasing is more seriously for the system without thiophene. The peaks for oxidation/reduction of Ni2+/Ni3+ and Ni3+/Ni4+ redoxs are not clear because the scan rate is not slow enough but still can be identified at about 4.8 and 4.9 V, respectively [10, 33]. Apparently, the redox peaks of nickel ions mainly contribute to the lithium ion insertion/extraction capacity of LiNi0.5Mn1.5O4. It can be observed that the differences in redox peak potentials increase slightly with increasing cycle number for both electrodes, but the increased magnitude is more significant for the electrode in base electrolyte than that in thiophene-containing electrolyte. The more distinct difference in redox peak potentials shows the larger internal resistance caused by the more severe parasitic reactions on the electrode. This result suggests that the stability of LiNi0.5Mn1.5O4/electrolyte interface is improved by using thiophene, which can be confirmed by the change in interfacial resistance, as shown in Fig. 2b, d. The interfacial resistances (estimated with the increase in the semi-arc of the Nyquist plots) change more slightly for LiNi0.5Mn1.5O4 cycled in thiophene-containing electrolyte than that in base electrolyte. The experimental results of the Nyquist plots induce that interfacial resistance has increased with cycle increases. It is also important to remark that these Nyquist plots can be simulated using appropriate software in order to attain the resistance parameter using an equivalent electrical circuit, as previously reported [34–36]. However, in this investigation, these Nyquist diagrams were qualitatively compared.

Fig. 2
figure 2

Cyclic voltammograms at a scan rate of 0.2 mV s−1 and electrochemical impedance spectra of LiNi0.5Mn1.5O4 after cycling in base (a, b) and thiophene-containing (c, d) electrolytes at 25 °C

Full size image

Figure 3 presents the morphologies of the LiNi0.5Mn1.5O4 after 400 cycles, with a comparison of the fresh electrode. LiNi0.5Mn1.5O4 particles (about 1 μm) can be observed in the fresh electrode, which adopt a highly crystal spinel morphology, as shown in Fig. 3a. After LiNi0.5Mn1.5O4 is cycled in the base electrolyte for 400 cycles, the spinel shape becomes indistinct, as shown in Fig. 3b. This observation confirms that the dissolution of transition metal ions from and Jahn-Tell crystallographic distortion happen in LiNi0.5Mn1.5O4, which cause the poor cyclic stability. With the same cycles in the thiophene-containing electrolyte, on the contrary, the particle maintains its structure integrity, as shown in Fig. 3c. Besides, the particles are covered with a deposit and connected with fibrous spices (as shown by the arrow in Fig. 3c). Apparently, thiophene is oxidized when charged during cycling, forming a protective cathode film on and electronically conductive polythiophene among LiNi0.5Mn1.5O4 particles. The former protects LiNi0.5Mn1.5O4 from destruction, and the latter provides electronic conductivity, contributing to the improved cyclic stability and rate capability.

Fig. 3
figure 3

SEM images of the fresh (a) and cycled LiNi0.5Mn1.5O4 electrodes in base (b) and thiophene-containing (c) electrolytes

Full size image

The contribution of thiophene to the protection for LiNi0.5Mn1.5O4 can be confirmed by XRD analyses. Figure 4 presents the XRD patterns of LiNi0.5Mn1.5O4 electrodes after 400 cycles, with a comparison of the fresh LiNi0.5Mn1.5O4 electrode. To avoid the effect of the lithium content in the samples and the sample position, the samples were taken from the cell hold at 3 V till the current is lower than C/100, and the XRD profiles were referred with that of aluminum. It can be seen that three electrodes possess cubic spinel structure (Fig. 4a); the changes in peak position and intensity are more significant for the cycled electrode in the base electrolyte than that in the thiophene-containing electrolyte, confirming that thiophene can provide a protection for LiNi0.5Mn1.5O4 (Fig. 4b). This result indicates that LiNi0.5Mn1.5O4 suffers the shrinkage of the crystal lattice [37, 38] when deep cycling is performed. In fact, the lattice parameters calculated from XRD by using Rietveld refinement [4] are 8.171, 8.163, and 8.144 Å for the flesh electrode and the cycled electrodes in thiophene-containing and base electrolytes, respectively. The shrinkage of the LiNi0.5Mn1.5O4 crystal lattice can be ascribed to the disproportionation of Mn3+ in LiNi0.5Mn1.5O4 forming the dissolvable Mn2+ into electrolyte and the remaining Mn4+ in lattice. Mn4+ has smaller ionic radius than Mn3+, leading to the lattice shrinkage. The less content of Mn3+ in the samples can be confirmed by smaller capacity of Mn3+/Mn4+ redox at around 4.1 V, as shown in Fig. 5. The capacity is about 15 mAh g−1 for the electrode in thiophene-containing electrolyte after 400 cycles, but it is only 9 mAh g−1 for the electrode in base electrolyte.

Fig. 4
figure 4

XRD patterns of the fresh and cycled LiNi0.5Mn1.5O4 electrodes in base and thiophene-containing electrolytes

Full size image
Fig. 5
figure 5

The 400th discharge profiles of LiNi0.5Mn1.5O4 in base and thiophene-containing electrolytes at 25 °C with 1-C current rate

Full size image

The formation of conductive polythiophene has been demonstrated in other cathodes, such as LiNi1/3Co1/3Mn1/3O2 [30], LiCoPO4 [31], and Li3V2(PO4)3 [32]. Polythiophene possesses a conjugated π network and exhibits excellent electronic conductivity. The electrooxidation polymerization of thiophene on LiNi0.5Mn1.5O4 was observed from a charging profile cyclic voltammogram, as shown in Fig. 6. Figure 6a presents the initial charging profiles of LiNi0.5Mn1.5O4 in base and thiophene-containing electrolytes. The small voltage plateau at about 4.5 V (vs. Li/Li+) for the LiNi0.5Mn1.5O4 in thiophene-containing electrolyte, as shown by the circle in Fig. 6a, is indicative of the electrooxidation polymerization of thiophene. The subsequent discharge profile (the inset in Fig. 6a) shows the contribution of the formed conductive polymer to the improved discharge capacity due to the increased electronic contact between LiNi0.5Mn1.5O4 particles and conductive carbon. Figure 6b presents the initial cyclic voltammograms of LiNi0.5Mn1.5O4 in base and thiophene-containing electrolytes, in which the larger oxidation current from 4.5 V (vs. Li/Li+) in the first cycle for the electrode in thiophene-containing electrolyte than that in base electrolyte is also indicative of the electrooxidation polymerization of thiophene.

Fig. 6
figure 6

Initial charge-discharge profiles at 1-C rate (a) and cyclic voltammograms at a scan rate of 0.1 mV s−1 (b) of LiNi0.5Mn1.5O4 in base and thiophene-containing electrolytes

Full size image

Figure 7 presents the TEM images of the LiNi0.5Mn1.5O4 particles after cycling in base and thiophene-containing electrolytes, with a comparison of the fresh LiNi0.5Mn1.5O4 particle (Fig. 7a). It can be seen from Fig. 7b that, after cycling in base electrolyte, the LiNi0.5Mn1.5O4 particle is covered by random deposits, as indicated by the arrows in the figure. These deposits may be ascribed to the decomposition of the electrolyte or the recrystallization of the dissolved active materials, which cannot form a compact film due to the interference from the evolution of carbon dioxide, one of the decomposition products of solvent carbonate. Differently, a very thin and uniform film is observed on the LiNi0.5Mn1.5O4 particles cycled in thiophene-containing electrolyte, as indicated by the arrows in Fig. 7c. Apparently, a protective film can be formed by the application of thiophene, which suppresses the electrolyte decomposition and protects LiNi0.5Mn1.5O4 from destruction and thus contributes to the improved cyclic stability [39]. Figure 7d provides the image of the electrode in base electrolyte after 400 cycles. It can be observed that without the protection, the particle is corroded and becomes porous after extensive cycles, while the electrode in the thiophene-containing electrolyte retains its integrity with the protective thin film, as shown in Fig. 7e.

Fig. 7
figure 7

TEM images of fresh LiNi0.5Mn1.5O4 electrode (a), the electrodes after 10 cycles in base (b) and thiophene-containing (c) electrolytes, and the electrodes after 400 cycles in base (d) and thiophene-containing (e) electrolytes

Full size image

The participation of thiophene in the formation of protective cathode film can be further confirmed by XPS and FTIR analyses. Figure 8 presents the XPS spectra of the cycled LiNi0.5Mn1.5O4 in base and thiophene-containing electrolyte, with a comparison of the fresh LiNi0.5Mn1.5O4. In the Mn 2p spectra, the Mn 2p3/2 peaks (642.1 and 643.4 eV) and the Mn 2p1/2 (653.9 eV) are clearly identified for the three electrodes. The weaker peak intensities for the LiNi0.5Mn1.5O4 cycled in thiophene-containing electrolyte than in base electrolyte confirm that LiNi0.5Mn1.5O4 particles have been covered with a film. Similar phenomenon is also observed in the Ni 2p spectra where there exist Ni 2p3/2 (855.1 eV) and Ni 2p1/2 (872.2 eV) peaks and in the O 1 s spectra where the peak at 529.6 eV is corresponding to metal oxide. Differently, S 2p spectra can be recorded for the LiNi0.5Mn1.5O4 cycled in thiophene-containing electrolyte but not for that in base electrolyte, confirming that the oxidation products of thiophene have been incorporated into the cathode film. Figure 9 compares the FTIR spectra obtained from the cycled LiNi0.5Mn1.5O4 in base and thiophene-containing electrolyte. Compared with the cycled LiNi0.5Mn1.5O4 in base electrolyte, there appear extra peaks, located at around 720, 1025, and 1125 cm−1, for the cycled LiNi0.5Mn1.5O4 in thiophene-containing electrolyte. These peaks are associated with the sulfur-oxygen bonds [40, 41], confirming further that the oxidation products of thiophene have been incorporated into the cathode film.

Fig. 8
figure 8

XPS patterns of the fresh and LiNi0.5Mn1.5O4 electrodes after 10 cycles in base and thiophene-containing electrolytes

Full size image
Fig. 9
figure 9

FTIR spectra of LiNi0.5Mn1.5O4 electrodes after 10 cycles in base and thiophene-containing electrolytes

Full size image

The contribution of thiophene to the suppression of electrolyte decomposition and the protection of LiNi0.5Mn1.5O4 was further evaluated by storage test. Figure 10 presents the photographs of the electrolytes contacting the cycled electrodes after storage at 55 °C for 24 h, with a comparison of fresh electrolyte. The electrodes were charged and discharged at 1-C rate at 25 °C for 10 cycles. It can be observed that the color of the electrolyte with the electrode cycled in base electrolyte turns to yellow deeply, suggesting that the severe parasitic reactions take place on the electrode/electrolyte interface. However, this phenomenon is not so significant in the electrolyte for the electrode with thiophene, indicating the suppression of electrolyte decomposition and the protection of LiNi0.5Mn1.5O4 by the thiophene.

Fig. 10
figure 10

Photographs of the electrolyte stored at 55 °C for 24 h: a without electrode, b with electrode after 10 cycles in base electrolyte, and c with electrode after 10 cycles in thiophene-containing electrolyte

Full size image

Conclusions

The cyclic stability and the rate capability of LiNi0.5Mn1.5O4 cathode for a high-voltage lithium-ion battery can be improved by using thiophene as an electrolyte additive. This improvement is attributed to the simultaneous formations of the protective cathode film on and the conductive polymer among LiNi0.5Mn1.5O4 particles. Thiophene is oxidized at about 4.5 V (vs. Li/Li+) during charging process, forming the protective cathode film and the conductive polymer. The conductive polymer improves the electronic contact between LiNi0.5Mn1.5O4 particles and conductive carbon, enhancing the discharge capacity and rate capability of LiNi0.5Mn1.5O4. The oxidation products of thiophene participate in the formation of the cathode film, which suppresses the electrolyte decomposition and protects LiNi0.5Mn1.5O4 from destruction, improving the cyclic stability of LiNi0.5Mn1.5O4. To our knowledge, it is the first time to report that thiophene can be used as an electrolyte additive for the formation of protective cathode film for LiNi0.5Mn1.5O4 cathode in lithium-ion battery.