Introduction

Fossil fuels bring a range of environmental pollution problems, and more and more new energy sources are being used in response to sustainable development requirements [1,2,3,4,5]. Wind and solar energy is limited by time and location and requires continuous and efficient energy storage systems to provide a sustainable energy supply [6,7,8,9]. Secondary electrochemical battery is a pivotal technology for large-sized energy storage systems [10,11,12,13]. Among them, lithium-ion batteries are suitable for a variety of electronic devices and large-sized energy storage, such as smart phones, electric bicycles, and so on [14,15,16,17,18]. In addition, the demand of power storage systems such as electric vehicles and hybrid vehicles is also increasing [19, 20]. Lithium resources are being consumed more and more, leading to its high price and resource lack [21, 22]. Therefore, it makes sense to research and develop other new battery systems [23].

Sodium-ion batteries are expected to buffer excessive demand for lithium-ion batteries [24,25,26], which have attracted widespread attention from battery researchers. The cost of sodium-ion batteries is low due to abundant sodium reserves and wide range of sources [27]. Low-cost batteries also contribute to the further application and development of energy storage systems [28]. As a result, reports on sodium-ion batteries have increased incredibly in last decades [29, 30]. Many electrode materials have been studied, such as carbon materials [31, 32], TiO2 [33], SnS2 [34], Na2Ti3O7 [35], NaVO2 [36], LiTi2(PO4)3 [37], and NaTi2(PO4)3 [38] as anode materials and NaxMnO2 [39], NaxCoO2 [40], Na3V2(PO4)3 [41], and NaFePO4 [42] as cathode materials. The working process of sodium-ion batteries is realized by transfer of electrons and the intercalation/deintercalation of sodium ions. They are “rocking chair type” batteries [43, 44] like lithium-ion batteries. But in terms of ionic radius, Na is larger than Li, which tends to cause greater expansion in volume during insertion process, further resulting in unsatisfactory cycling stability and inferior specific capacity [45]. Hence, to meet increasing demand, the development of suitable electrode materials for sodium storage remains a huge challenge.

NaTi2(PO4)3 possesses Na superionic conductor (NASICON) structure and is a class of fast ion conductor material [46, 47]. The [TiO6] octahedron and [PO4] tetrahedron in the NaTi2(PO4)3 crystal are interconnected by oxygen atoms at the apex to form a [Ti2(PO4)3]− polyanion structure [48, 49]. It has large channels inside, which are very suitable for the rapid transmission of lithium ions and sodium ions. However, its intrinsic conductivity is low, and this shortcoming is usually improved by combining with conductive additives or carbon [50, 51]. For instance, Zuo et al. [52] have proposed and manufactured reduced graphene and carbon co-modified NaTi2(PO4)3 sample by hydrothermal process and high-temperature calcination. Graphene and coated carbon can form a conductive network in this composite, which effectively improves the conductivity and sodium storage performance. At 0.1 C, the corresponding discharge capacity of modified NaTi2(PO4)3 can reach 129 mAh g−1. Another way to improve conductivity is ion doping [53], including anion doping and cation doping. Su et al. [54] have proposed a novel nanostructured Na1-2xTi2(PO4)3-xFx anode for sodium-ion batteries. Appropriate fluorine doping increases the ion/electron transport speed, resulting in ultra-long cycle life. Capacity decay of Na2.9Ti2(PO4)2.95F0.05/C at 10 °C is only 10% after 1000 cycles.

In this paper, crystal doping strategy was employed to raise the sodium storage performance of NaTi2(PO4)3/C compounds for sodium-ion battery. We used potassium instead of sodium partially and phenolic resin as carbon resource to construct Na1-xKxTi2(PO4)3/C anode materials. An easy sol-gel approach was utilized to prepare composites, roughly as depicted in Fig. 1. And the electrochemical behavior of all samples is reported in this paper.

Fig. 1
figure 1

Schematic diagram of the synthesis process of Na1-xKxTi2(PO4)3/C composites

Full size image

Experimental

Synthesis

Na1-xKxTi2(PO4)3/C (x = 0, 0.01, 0.03, 0.05) compounds were designed and compounded as following the steps. By stirring to form initial solution, 1.7202 g Ti(OC4H9)4 and 15 mL ethanol were mixed. H3PO4, CH3COONa∙3H2O, and CH3COOK with stoichiometric ratio were added to the solution. After that, 2-mL concentrated hydrochloric acid and 0.17-g phenolic resin were added successively and then stirred for 3 h at 55 °C in closed condition and dried for 10 h at 80 °C in open condition. Each precursor was calcined for 5 h at 750 °C in argon-filled tube furnace. Na1-xKxTi2(PO4)3/C (x = 0, 0.01, 0.03, 0.05) compounds were abbreviated as NC, NC/K-1, NC/K-3, and NC/K-5, respectively.

Characterizations

Crystal type of four samples were examined by X-ray diffraction (XRD), conducted on X-ray diffractometer (D/MAX2500PC) with the radiation of Cu-Kα. Morphology features of as-prepared samples and electrodes after cycling testing were investigated through S-4800 scanning electron microscopy (SEM) manufactured in Hitachi. Each cycled electrode was obtained by disassembling the cell. And the electrode was rinsed by absolute ethanol a few times. Then it was dried for 8 h at 80 °C.

Electrochemical measurements

Composite powder, polyvinylidene fluoride (PVDF), and super P were blended based on a mass ratio of 7:1.5:1.5 and then dissolved in N-methyl-2-pyrrolidone (NMP). The liquid mixture was coated on Cu foil. After drying for 6 h at 80 °C, working electrode was got after cutting Cu foil into small wafer with diameter of 14 mm. The mass load of electrode was 1.5–2.0 mg cm−2. CR2016 half cell used metallic sodium as counter electrode, glass fiber as membrane, and prepared electrode as working electrode. Electrolyte was 1 M NaClO4 dissolved in mixed organic solvents, including DMC, EC, and EMC with volume ratio of 1:1:1. In addition, 5% FEC was added to the electrolyte to make it more stable. All cells were assembled in the glove box with oxygen and water < 0.1 ppm. Chenhua electrochemical workstation (CHI660E) was employed to conduct electrochemical impedance spectroscopy (EIS) and cyclic voltammetry (CV) tests for cells. The scan rate and voltage window for CV tests were 0.2 mV s−1 and 0.01–3 V, respectively. The frequency and amplitude for EIS tests were 0.01–100,000 Hz and 5 mV, severally. EIS tests for cells were performed after activating at 0.04 A g−1 for 5 cycles and then charging to 2.1 V for 2 h. Rate and cycling performance were conducted on Neware battery testing system (CT-3008W), which were tested by galvanostatic charge-discharge mode. The rates for rate performance varied from 0.04 to 3 A g−1. Cycling performance was tested at current of 1.2 A g−1 for 1000 cycles.

Results and discussion

Figure 2 displays XRD pattern of all samples. Clear diffraction peaks reflect crystal form characteristics of NASICON-type NaTi2(PO4)3. Four sets of similar diffraction peaks indicate similar crystal structure of four samples, all of which are ascribed to the rhombohedral structure of NaTi2(PO4)3 (JCPDS No. 01–084-2008). The obvious diffraction peaks of impurity do not appear, indicating that each synthesized complex has the pure phase.

Fig. 2
figure 2

XRD patterns of all samples

Full size image

As seen in Fig. 3, SEM images exhibit the surface morphology comparison of NC and NC/K-3. It can be seen that the overall morphology of two materials is loose, with some agglomeration. As shown in high-magnification images, nano-sized particles exist in both materials. Small nanoparticles and loose structure can help the transport of Na ions between electrolyte and crystal. Pristine and K-doped NaTi2(PO4)3 demonstrate no obvious difference on the dispersion and particle size. It shows that the doping of K ion on Na site does not have a momentous influence on morphology of composite. Good stability of morphology is beneficial to get the reason of improved performance of K-doped sample.

Fig. 3
figure 3

SEM images for NC (a, b) and NC/K-3 (c, d) at different magnifications

Full size image

Cyclic voltammetry tests of NC and NC/K-3 were conducted on half cells, and cyclic voltammetry curves are shown in Fig. 4. Two sharp redox peaks indicate that both NC and NC/K-3 electrodes can realize the reversible intercalation/deintercalation of sodium ions. The peaks located at 1.89–2.35 V and 0.13–0.61 V correspond to redox reactions of Ti4+/Ti3+ and Ti3+/Ti2+, respectively. From Fig. 4a, oxidation and reduction peak current densities at about 2.1 V for NC are 0.21 and 0.19 A g−1, severally. And those for NC/K-3 are 0.29 and 0.28 A g−1 (Fig. 4b). Peak current densities increase obviously for NC/K-3, demonstrating that the K-doping strategy is effective. Furthermore, peak current densities at around 0.4 V for NC/K-3 are also higher than those of NC. Cyclic voltammetry data illustrate that sodium storage properties of NC/K-3 are better than those of NC coming from the efficient crystal doping, as K doping cannot affect the morphology of composites, which is confirmed by the SEM results.

Fig. 4
figure 4

Cyclic voltammetry curves of NC (a) and NC/K-3 (b) composites at scan rate of 0.2 mV s−1

Full size image

Aiming at discussing the kinetic process of charge transfer in electrode, EIS tests are carried out on half cells, and spectra are displayed in Fig. 5. As seen from Nyquist plot, the low-frequency line means Warburg impedance (Zw), and middle-frequency semicircle means charge transfer resistance (Rct). The larger slope of line for NC/K-3 indicates that sodium ions diffuse faster in the NC/K-3 electrode than NC electrode [55]. It is noted that Rct for NC and NC/K-3 are 201.31 Ω and 106.27 Ω, severally. The smaller Rct data illustrate that NC/K-3 electrode has a faster intercalation/deintercalation kinetics of sodium ions. This may be due to the increase in unit cell volume of NC/K-3 after K doping, which provides a wider intercalation/deintercalation channel for sodium ions and accelerates the migration of sodium ions [56]. At high frequency, two curves both have an intercept on Z-axis, corresponding to ohmic resistance (Rs). Two composites exhibit the similar Rs, which implies that K doping has no obvious effect on electrical conductivity.

Fig. 5
figure 5

Electrochemical impedance spectra for NC and NC/K-3

Full size image

The rate performances of four samples are given in Fig. 6. Three samples with K doping have a significant increase in discharge capacity, and NC/K-3 reaches the highest discharge capacity in four samples. NC/K-3 delivers discharge capacities of 206.65, 139.14, and 94.45 mAh g−1 at 0.1, 1.2, and 3 A g−1, severally, which are 118.54, 100.09, and 69.53 mAh g−1 higher when compared with the bare NC. When current density goes from 3 to 0.1 A g−1, NC/K-3 can keep discharge capacity at 197.75 mAh g−1. This value almost reaches 95.8% of its original value, indicating excellent reversibility of NC/K-3 electrode. The improved properties of K-doped composite are probably attributed to that K doping can lead to the increase in unit cell volume. This can also be seen in previous reports. Xia et al. [57] prepared K-doped Na3Fe2(PO4)3 cathode materials for sodium-ion battery and found lattice parameters indeed increase after K doping. The unit cell volume of K-doped samples increases slightly compared with blank Na3Fe2(PO4)3. Excessive doping may cause slight damage to the lattice structure, resulting in poor electrochemical performance. Charge-discharge profiles at different current densities for NC and NC/K-3 are given in Fig. 6b and c. All profiles exhibit planus plateaus for charging and discharging at 2.1 and 0.4 V approximately, which correspond with two pairs of redox peaks in CV tests. All plateaus for NC/K-3 are longer and more stable, and this advantage is more pronounced for plateaus around 0.4 V. With the increase of current, the stable existence of the charging and discharging plateaus indicates good rate performance for NC/K-3.

Fig. 6
figure 6

Rate performance of four samples (a), charge-discharge profiles for NC (b), and NC/K-3 (c) at different rate

Full size image

Figure 7 displays long-term cycling performance of two samples and coulombic efficiency for each cycle. The initial discharge capacity and coulomb efficiency of NC/K-3 are higher than those of NC. Moreover, after charge-discharge process for 1000 cycles, NC/K-3 delivers discharge capacity of 89.5 mAh g−1. In contrast, NC releases lower discharge capacity with value of 50.3 mAh g−1. Discharge capacities of NC/K-3 are always higher than those of NC. Apparently, NC/K-3 demonstrates excellent cycling performance at high current density. This is in accord with the rate results. From Fig. 7b, coulombic efficiency for NC and NC/K-3 is about 100% after increasing in first few cycles, implying stable and high-efficient nature of cells.

Fig. 7
figure 7

Cycling performance (a) and coulombic efficiency (b) for NC and NC/K-3 at 1.2 A g−1

Full size image

In order to investigate the structure stability of NC/K-3, we compared the morphology of NC/K-3 after 100 and 1000 cycles at 1.2 A g−1. SEM images of NC/K-3 after different cycles are shown in Fig. 8. As can be seen from Fig. 8, there is no significant etching and structural comminution on the electrode surface. The stable structure helps to output stable capacity. This also explains the outstanding cycle performance of NC/K-3.

Fig. 8
figure 8

SEM images of NC/K-3 after 100 (a) and 1000 (b) cycles at 1.2 A g−1

Full size image

Conclusion

Na1-xKxTi2(PO4)3/C (x = 0, 0.01, 0.03, 0.05) compounds were synthesized by simple sol-gel way. A small amount of doped K does not change crystal form of NASICON NaTi2(PO4)3 and surface morphology of composites. However, doping of K on Na site for NaTi2(PO4)3 has obvious effect on its electrochemical performance and has been shown to be positive. NC/K-3 delivers the highest discharge capacity at different rates in four composites (206.65, 139.14, and 94.45 mAh g−1 at 0.1, 1.2, and 3 A g−1). As comparison, NC delivers 88.11, 39.05, and 24.92 mAh g−1 at 0.1, 1.2, and 3 A g−1, severally. Moreover, NC/K-3 also has long cycling life and high capacity (89.5 mAh g−1 after 1000 cycles at 1.2 A g−1). These elementary electrochemical data imply that K-doped NaTi2(PO4)3 is a potential competitive anode material in sodium-ion battery.