Introduction

Recently, various iron-based spinel composite metal oxides MFe2O4 (M = Cu [1], Zn [2–4], Mn [5], Ni [6, 7], Co [8–10], etc.) are being used as electrode materials for lithium-ion batteries (LIBs). Among those, zinc ferrite ZnFe2O4 (ZFO) has become one of the most widely used anode materials for LIBs, due to its non-toxicity, natural abundance in the earth, and high specific capacity of ~1000 mA h g−1 [2, 11–13]. However, ZFO has also suffered from some problems such as poor electrical conductivity and large volume change during lithiation/delithiation processes. These typical problems, accordingly, result in pulverization and cracking of electrode materials, as well as a rapid fade in discharge/charge capacities in LIBs [2, 14–20].

The electrochemical properties of ZFO anode materials can be generally improved by hybridizing with carbon-based materials, such as graphene [12, 16, 21–26], amorphous carbon [15, 27–30], and carbon nanotube [14, 31, 32]. Besides that, doping with heteroatom (e.g., Ag, Mn, Mg, and Ce) can also enhance the electrochemical properties of ZFO nanomaterials [33–38]. Especially, it has been reported that sulfur doping in carbonaceous materials is quite favorable for improving electronic conductivities and lithium electro-activities, as well as enhancing electrical performances for LIB anodes [39–44]. Nonetheless, to the best of our knowledge, doping with foreign anions (e.g., sulfur) in ZFO has not been reported, although a related sulfide material of yolk–shell-structured Zn–Fe–S with a 1:2 Zn/Fe molar ratio was presented very recently.

In this study, low-dose dopant of sulfur (less than 1.5 wt%) was incorporated into ZFO to form S-doped ZFO nanoparticles (NPs) through a hydrothermal process for sulfurization. The content of sulfur dopant was tuned from 0.70 to 1.26 wt% by controlling concentrations of Na2S solutions. To evaluate the potential of S-doping for enhancing electrochemical performances, the pristine ZFO electrode and S-doped ZFO electrodes were tested in 2032 coin-type half cells composed of electrode materials versus Li. As a result, the electrodes of S-doped ZFO NPs showed much higher rate performances than the pristine ZFO electrodes.

Experimental

All the chemical reagents used in this study were of analytical grade and were used without further purification.

Preparation of ZFO NPs

In a typical synthetic process, 2 mmol of Zn(CH3COO)2·2H2O and 4 mmol of Fe(NO3)3·9H2O were added into 100 mL of ethylene glycol under a magnetic stirring for 15 min to ensure a complete dissolution. The obtained solution was transferred into a round bottom flask and then refluxed at for 2 h 160 °C. After reaction, a yellow precipitate was collected, washed with absolute alcohol thoroughly, and dried for 12 h at 80 °C. Lastly, the precipitate was annealed in air for 4 h at 450 °C with a heating rate of 2 °C min−1 to obtain the final product of ZFO NPs.

Preparation of S-doped ZFO NPs

The as-prepared ZFO NPs (120 mg) were dispersed in 70 mL of Na2S solutions with various concentrations under a vigorous stirring for 10 min, and then transferred into a Teflon-lined stainless steel autoclave for hydrothermal treatments for 24 h at 120 °C. After reaction, the autoclave was cooled down to room temperature naturally. The collected precipitates were centrifuged, washed alternately with deionized water and anhydrous ethanol for three times, and then dried under vacuum for 12 h at 60 °C. The resulting S-doped ZFO NPs synthesized with a Na2S concentration of 29, 73, or 146 mmol L−1 were abbreviated as ZFO-S1, ZFO-S2, and ZFO-S3, respectively.

Characterizations

The crystal structures were characterized using a MaXima-7000 (Shimadzu, Japan) diffractometer with graphite-monochromatized Cu K α radiation (λ = 1.5418 Å). Morphologies were investigated using the field emission scanning electron microscope (FE-SEM, JEOL JSM-6510LV), transmission electron microscope (TEM, JEM-2100, Japan), and high-resolution transmission electron microscope (HRTEM, JEM-2100, Japan). Brunauer–Emmett–Teller (BET) method was applied to calculate specific surface areas, and Barrett–Joyner–Halenda (BJH) model was used to derive pore size distributions (ASAP 2020, USA). Compositions were acquired on a basis of the X-ray photoelectron spectroscopy (XPS, Thermo ESCALAB 250) and the energy dispersive X-ray spectroscopy (EDAX, Phillips).

Electrochemical measurements

The working electrodes were made by coating a slurry mixture of 70 wt% active materials, 20 wt% super P as conducting agent, and 10 wt% carboxyl methyl cellulose (CMC) as binder on copper foils, which were subsequently dried under vacuum for 12 h at 70 °C. Lithium foils were used as the counter electrode. The specific capacities and cycling properties of the electrodes were determined using CR2032-type coin cells with a solution of 1 M LiPF6 in ethylene carbonate/diethyl carbonate (1:1 in volume) as electrolytes, which were assembled in an argon-filled glove box (LS800, Dellix, China). The galvanostatic charge/discharge tests were carried out on a battery tester (Land CT2001A, China) with a voltage window in the range of 0.005–3.0 V (vs. Li/Li+) at room temperature. Cyclic voltammetry measurements were carried out on a CHI 660B (Shanghai, China) in a potential range of 0.005–3.0 V at a scan rate of 0.1 mV s −1. The electrochemical impedance spectra (EIS) were recorded on a CHI 660D (Shanghai, China) electrochemical workstation with a frequency range from 0.1 to 1.0 × 105 Hz.

Results and Discussion

The typical XRD patterns of the pristine ZFO NPs and S-doped ZFO NPs are shown in Fig. 1.

Figure 1
figure 1

XRD patterns of ZFO, ZFO-S1, ZFO-S2, and ZFO-S3

Full size image

For the pristine ZFO NPs, six clearly visible peaks, centered at 29.95°, 35.28°, 42.87°, 53.18°, 56.69°, and 62.24°, were, respectively, indexed to (220), (311), (400), (422), (333), and (440) crystal planes of ZFO cubic structure (JCPDS Card No. 89-1011). Moreover, no diffraction peaks derived from the impurities were detected, which implied the high-purity samples. The relatively broadened diffraction peaks of the pristine ZFO NPs and S-doped ZFO NPs indicated the relatively small crystallite sizes. According to the Scherrer’s equation D hkl  = kλ/(βcosθ) (λ was the incident X-ray wavelength, β was the full width at half maximum of diffraction peak, θ was the diffraction angle) [45], grain sizes of ZFO, ZFO-S1, ZFO-S2, and ZFO-S3 were calculated to be 21.7, 16.2, 12.2, and 12.7 nm, respectively. Obviously, grain sizes of all the samples decreased with increasing sulfur content in the samples. It could be ascribed to dissolution–recrystallization of NPs as a result of adsorption and incorporation of S2− on the surfaces or grain boundaries of ZFO NPs during hydrothermal process for sulfurization [46].

Surface morphologies of all the samples were investigated by the SEM technique. Figure 2a shows that the pristine ZFO consists of the agglomerated nanoparticles with a typical secondary particle size of ~30 nm. As shown in Fig. 2b–d, S-doped ZFO displayed a typical secondary particle size of ~50 nm, which were obviously larger than that of the pristine ZFO. Compared with the results of XRD, the observed secondary particle sizes in SEM images were approximately two to three times larger than the calculated values of primary particles by XRD technique.

Figure 2
figure 2

SEM images of a ZFO, b ZFO-S1, c ZFO-S2, and d ZFO-S3

Full size image

To obtain information about the microstructure and compositions of S-doped ZFO NPs, TEM, HRTEM, as well as EDAX elemental mapping were applied. As shown in Fig. 3a, ZFO-S3 was composed of a large amount of irregular primary particles, which attached together and assembled into porous structure. Observed sizes of primary particles fell into the range of 10–20 nm, which was consistent with XRD results. HRTEM image of ZFO-S3 (Fig. 3b) exhibited the lattice spacing of 0.298 and 0.486 nm, which corresponded well with the inter-planar distances of (220) and (111) crystal planes of ZFO cubic structure [14]. EDAX spectra of ZFO-S3 demonstrated the existence of Zn, Fe, O, and S elements (Fig. 3c). EDAX quantitative microanalysis revealed that the contents of sulfur in ZFO-S1, ZFO-S2, and ZFO-S3 were 0.70, 0.91, and 1.26 wt%, respectively. This suggested that the content of sulfur in S-doped ZFO NPs gradually rose with increasing Na2S concentrations from 29 to 146 mmol L−1 in the hydrothermal process for sulfurization. Furthermore, using the microscopic technique of EDAX elemental mapping (Fig. 3d–h), Zn, Fe, O, and S elements were directly viewed to be distributed uniformly in ZFO-S3. Meanwhile, a relatively low content of sulfur was observed because of the weak bright–dark contrast between the sample and blank region.

Figure 3
figure 3

The sample of ZFO-S3: a TEM image; b HRTEM image; c EDAX spectrum; d Scanning TEM image and the corresponding EDAX elemental mappings of e Zn, f Fe, g O and h S. The scale bars were 50 nm

Full size image

Specific surface areas and pore size distributions of the samples were analyzed by the nitrogen adsorption–desorption isotherms. All the samples exhibited type IV isotherms according to the International Union of Pure and Applied Chemistry classification (Fig. 4). The specific surface areas of ZFO, ZFO-S1, ZFO-S2, and ZFO-S3, estimated from the BET method, were 50.9, 54.9, 52.3 and 35.4 m2 g−1, respectively. While the samples of ZFO, ZFO-S1, and ZFO-S2 possessed the similar specific surface areas, only ZFO-S3 exhibited a smaller one, which might arise from more defect sites induced by more S-doping. The pore size distributions of ZFO, ZFO-S1, ZFO-S2, and ZFO-S3, calculated from the BJH model, were located mainly in the range of 5–50 nm (insets of Fig. 4), and the average pore sizes were 15.3, 14.1, 16.2, and 17.4 nm, respectively. The pore volumes of ZFO, ZFO-S1, ZFO-S2, and ZFO-S3, which mainly derived from a large number of the inter-nanoparticle voids in nanostructures (see the SEM images shown in Fig. 2), were 0.21, 0.20, 0.23, and 0.17 cm3 g−1, respectively. In this work, the porous nanostructure of the samples was beneficial for improving contacts of electrode–electrolyte and diffusions of Li ions during the charge–discharge process, which ensured the good electrochemical performance of the samples.

Figure 4
figure 4

Nitrogen adsorption and desorption isotherms of a ZFO, b ZFO-S1, c ZFO-S2 and d ZFO-S3. The insets were the corresponding BJH pore size distributions

Full size image

Compositions and oxidation states of S-doped ZFO NPs were further characterized by XPS. Figure 5 shows the high-resolution XPS spectra of ZFO-S3. In the Zn region (Fig. 5a), the signals at 1046.5 and 1023.3 eV were attributed to Zn 2p1/2 and Zn 2p3/2 of Zn2+ [47]. In the high-resolution spectrum of Fe 2p (Fig. 5b), the signals centered at around 712.2 and 725.0 eV were assigned to Fe 2p3/2 and Fe 2p1/2, which were caused by Fe3+ in S-doped ZFO NPs [47]. The O 1s spectrum of ZFO-S3 displayed one peak at the binding energy of 530.6 eV, indicative of oxygen in metal oxides (Fig. 5c). As shown in Fig. 5d, two discrete broad bands in the S energy region could be discerned, although the signal-to-noise ratio was small due to the relatively low content of sulfur in ZFO-S3. One broad band at ~168 eV could be ascribed to oxidized sulfur such as SO4 2− or SO3 2− which existed on the surface or grain boundary of S-doped ZFO NPs, while the other broad band at ~162 eV could be assigned to S 2p1/2 and S 2p3/2 of S2− which was more likely to be located in the crystal lattice. These results from XPS analyses clearly demonstrated that S atom had been successfully incorporated into ZFO NPs.

Figure 5
figure 5

XPS spectra of ZFO-S3: a Zn 2p spectrum, b Fe 2p spectrum, c O 1s spectrum, and d S 2p spectrum

Full size image

To identify the electrochemical reactions, cyclic voltammetry (CV) profiles of all the samples, for the first three cycles, were studied in a voltage range of 0.05–3.0 V at a scan rate of 0.1 mV s−1 (Fig. 6). For all the samples, a large reduction peak at around 0.5 V was observed during the first scan, which could be attributed to three aspects: (i) reduction of Fe3+ to Fe0 and Zn2+ to Zn0; (ii) an irreversible reaction related to the decomposition of the electrolyte; (iii) formation of solid electrolyte interface (SEI) films on the surface of the electrodes [2, 13, 16, 35]. In the subsequent cycles, the cathodic peak shifted to 0.8 V, which was relative to the reversible reduction reaction of Fe2O3 and ZnO. The initial anodic peak located at 1.6 V was associated with the oxidation of Fe0 to Fe3+ and Zn0 to Zn2+ [13, 16, 35], which did not shift evidently in the second and third cycles. Compared to the first cycle, the integrated areas of the subsequent cathodic and anodic peaks decreased, indicating a gradual fading of specific capacity. In brief, there were little differences between the pristine ZFO NPs and S-doped ZFO NPs in retention of reversible capacity for the first three cycles.

Figure 6
figure 6

Cyclic voltammetry curves of a ZFO; b ZFO-S1; c ZFO-S2; and d ZFO-S3 electrodes in a potential window between 0.005 and 3 V at a scan rate of 0.1 mV s−1

Full size image

The initial charge and discharge curves of all the samples measured at a constant current density of 100 mA g−1 are presented in Fig. 7a. The electrode of the pristine ZFO exhibited an obvious voltage plateau at 0.80 V, while the electrodes of S-doped ZFO NPs such as ZFO-S1 and ZFO-S2 displayed voltage plateaus with slightly smaller values at 0.75 V and 0.70 V in the first discharge curve. With further increment of S-doping, the electrode of ZFO-S3 presented a distinct discharge curve with a lower voltage plateau than that of ZFO-S1 and ZFO-S2. The consecutive decrease in the voltage plateau from ZFO to ZFO-S3 might result from S-doping, as well as the corresponding increase of S-doping capability [48]. As shown in Fig. 6b, the pristine ZFO electrode and S-doped ZFO electrodes almost delivered similar discharge capacities at 1200–1350 mA h g−1 in the first cycle. After continuous galvanostatic discharge–charge at a current density of 100 mA g−1 for 60 cycles, the discharge capacity of ZFO electrode rapidly dropped to 200 mA h g−1, while ZFO-S1, ZFO-S2 and ZFO-S3 electrodes still retained the reversible discharge capacities of 386, 461, and 604 mA h g−1, respectively. Also, the initial coulombic efficiency of ZFO-S3 electrode was kept at 97.3 %, overtly higher than that of the ZFO electrode (78.2 %). These results implied that S-doping was favorable to enhance the cycling stability. The irreversible capacity loss in the first several cycles might result from the irreversible insertion of Li+ into the crystal structure of ZFO or S-doped ZFO, and/or the formation of thick SEI films on the surface of the electrodes [2, 47].

Figure 7
figure 7

a The first discharge–charge curves of all the samples at a current density of 100 mA g−1; b Discharge capacities at a current density of 100 mA g−1 and coulombic efficiencies of the ZFO-S3 electrode; c Rate capabilities of all the samples at current densities of 0.1, 0.2, 0.5, 1, 2 and 0.1 A g−1; d Nyquist plots of all the samples after 60 cycles at a current density of 1 A g−1

Full size image

The rate performances of all the samples at different current densities from 0.1 to 0.2, 0.5, 1, and 2 A g−1 are given in Fig. 7c. The ZFO-S3 electrode was found to deliver the best rate capability among all the samples. It exhibited final discharge capacities of 934.1, 764, 571, 406, and 277 mA h g−1 at current densities of 0.1, 0.2, 0.5, 1, and 2 A g−1, respectively. When the rate recovered to 0.1 A g−1, a reversible capacity as high as 761.9 mA h g−1 can still be retained. But for the ZFO electrode, the final rate capacities were only 731, 351, 202, 130, and 53 mA h g−1 at current densities of 0.1, 0.2, 0.5, 1, and 2 A g−1, respectively. Subsequently, the final discharge capacity dropped to 286.9 mA h g−1, although the current density returned to 0.1 A g−1. It was also observed that the rate capacities were significantly improved with the gradual increase of sulfur dopant from ZFO-S1 to ZFO-S3. In short, these results of rate performances had proved that S-doping within a certain range played an important role in enhancing discharge/charge capacities and cycling stabilities of the ZFO electrodes.

To further understand the mechanism behind the better electrochemical performance of S-doped ZFO NPs, electrochemical impedance spectroscopy (EIS) was measured. Nyquist plots in Fig. 7d show the typical semicircles in the high to medium frequency range, which revealed an inter-facial charge transfer impedance (R ct), and a typical sloping line in the low-frequency range, which corresponded with the Li+ ion diffusion impedance in the electrodes. It was seen that the diameters of the semicircles of S-doped ZFO electrodes (15–50 Ω), especially that of ZFO-S3 (15 Ω), were considerably smaller than that of the pristine ZFO electrode (90 Ω), indicating that the inter-facial charge transfer resistance remarkably decreased, as well as reflecting that incorporation of sulfur into ZFO NPs could promote electron transfer and Li+ ion diffusion at the interface between electrode and electrolyte. Above all, the results of EIS analysis proved that S-doping within a certain range could enhance the Li+ ion diffusion and electron kinetics in the electrodes for lithium storage, which in turn improved rate capabilities and cyclabilities of ZFO electrodes.

Conclusions

To summarize, S-doped ZFO NPs were synthesized through sulfurizing ZFO NPs during a hydrothermal process using Na2S as the sulfur source. EDAX results revealed that the contents of sulfur dopant were tuned from 0.70 to 1.26 wt% by modifying concentrations of Na2S solutions from 29 to 146 mmol L−1. EDAX elemental mapping illustrated that trace S atoms had been homogeneously incorporated into ZFO NPs. More importantly, S-doped ZFO electrodes exhibited a much higher reversible capacity, better rate performance, and more excellent cycling stability than the pristine ZFO electrode. EIS analyses proved that the enhanced lithium storage capabilities of S-doped ZFO electrodes were caused by the improved Li+ ion diffusion and electron kinetics due to sulfur doping in the ZFO NPs. After 60 cycles, at a current density of 100 mA g−1, the electrode of S-doped ZFO NPs (1.26 wt%, ZFO-S3) delivered a reversible specific capacity of 604 mA h g−1, while the pristine ZnFe2O4 electrode only remained 200 mA h g−1.