Carbon nanotube decorated NaTi₂(PO₄)₃/C nanocomposite for a high-rate and low-temperature sodium-ion battery anode

Abstract

A sodium super-ionic conductor structure NaTi₂(PO₄)₃ has been considered as a promising anode material for sodium-ion batteries. However, the inherent poor electronic and ionic kinetics leading to inferior rate and low-temperature performance severely restricts its extensive developments. In this work, we report a carbon nanotube decorated nano-NaTi₂(PO₄)₃/C anode composite to achieve high-rate capability (116.8 mA h g⁻¹ at 1C, 113.3 mA h g⁻¹ at 10C and 103.4 mA h g⁻¹ at 50C) and stable cyclability (about 98% capacity retention at 50C of 1000 cycles) as well as impressive low-temperature performance (about 65.2 mA h g⁻¹ at 10C at a temperature of minus 20 °C). The carbon nanotube network not only improved electrolyte infiltration to decrease the internal diffusion resistance, but also provides fast transport pathways for electrons to enhance the poor electronic conductivity of the NaTi₂(PO₄)₃ anodes. In view of the advantages of the electrode architecture design, we anticipate that the nanocomposites might be promising anode materials for long-life and low-temperature rechargeable sodium-ion batteries.

---

Introduction

Motivated by the tremendous development of portable electronic devices, hybrid electric vehicles and large-scale grid energy storage, advanced high-energy storage systems have attracted numerous attention.^1,2 According to the currently used energy-storage systems, lithium-ion batteries (LIBs) have been considered as potential energy storage devices due to their high energy density, long service life and environmental friendliness.^3,4 Unfortunately, there are many disadvantages regarding LIBs, including security hazards, poor cyclability and low-temperature performance.⁵ Furthermore, the scarce lithium reserves in the earth lead to the relatively high cost of LIBs, which highly restrict their mass production and practical applications.⁶ In this background, rechargeable sodium-ion batteries (NIBs) containing much cheaper sodium with similar physicochemical properties of lithium, possess relatively good low-temperature and safety performance and would become one of the most competitive candidates for next-generation rechargeable batteries.^7,8

Sodium super-ionic conductor (NASICON) structured NaTi₂(PO₄)₃ (NTP) has been attracted intensive interests for anode materials of sodium-ion batteries because of its open 3D framework, high specific capacity (133 mA h g⁻¹), and good thermal stability.^9–11 However, the low electronic conductivity leads to poor transfer kinetics between the active particles. As a result, NTP anode is suffering from significant challenges which hinder its extensive application at present, such as poor rate performance, cycle stability and low temperature performance.¹² Several strategies have been done to address the aforementioned issues, containing reducing the active particles size, designing surface conductive layer, doping other electroactive material and building conductive composite network.¹³ Among them, construction of interconnecting conductive network is becoming an appealing approach to solve the poor electronic conductivity problem, as the conductive network provides fast and contiguous pathways for electrons transportation.^14,15

For building such conductive network, carbon nanotube is one of the most commonly used materials.^16–18 Carbon nanotubes (CNTs) with particular properties mainly comprising high aspect ratio, corrosion resistance and excellent electronic conductivity have great potential to improve the electrical conductivity of electrode materials in the field of rechargeable batteries.¹⁹ Most recently, Zhou's group reported a LiFePO₄–CNTs composite material with lower charge transfer resistance, higher specific capacity and better rate performance, attributed to the interpenetrating conductive CNTs networks.²⁰ Wu's group synthesized the Li(Ni_0.5Co_0.2Mn_0.3)O₂ cathode modified by CNTs, achieving electrochemically active and effectively depolarized components. The conductive network reduced the polarization of the electrode effectively, which results in approximately 25% increased reversible capacity and significantly improved cycling performance.²¹ Therefore, it is not hard to see that the incorporation of CNTs within electrode materials through an appropriate way can effectively enhance the rate capability and cycling performance.

In the present work, we designed and synthesized a carbon-nanotube-decorated NTP/C nanocomposite electrode for the poor electronic conductivity of NTP anode by a two-step solvothermal combined hydrothermal method. The NTP/C–CNTs electrode material integrates multiple advantages of CNTs-interconnected network and NASICON structure for the electrochemical performance of NTP anode. The CNTs network can provide a continuous pathway to speed the electrons transport. The prepared NTP/C–CNTs nanocomposite electrodes demonstrate good high-rate capability (about 88% capacity retention at a high rate of up to 50C) and cycling stability (about 98% capacity retention at 50C of 1000 cycles). Furthermore, the low-temperature performance (about 65.2 mA h g⁻¹ at 10C in the temperature of minus 20 °C) is much developed compared to the NTP/C, indicating its superiority at the low temperatures. This novel structure will be extended to other promising electrodes for high performance rechargeable batteries.

Experimental

Synthesis of NTP/C–CNTs nanocomposites

NTP/C–CNTs was synthesized through two-step solvothermal combined hydro-thermal method. First, 0.16 g CH₃COONa and 0.68 g tetrabutyl titanate were separately dissolved in 6 ml H₃PO₄ and 40 ml C₂H₅OH. Then the above two solutions were mixed with continuously stirring. After that, the mixture was transferred into the autoclave and solvothermally heated in an oven at 160 °C for 3 h. The white precursor was washed by vacuum filtration using absolute ethyl alcohol. After drying, the products were ground and transferred to the tube furnace being heated at 600 °C for 4 h in a flowing Ar-atmosphere to achieve the NTP/C particles. Second, 0.2 g NTP/C nanoparticles and 0.01 g CNTs was dispersed into appropriate deionized water with magnetic stirring and ultrasonic treatment. The solution was poured in the autoclave and then put in an oven at 180 °C for 10 h. The NTP/C–CNTs was obtained after annealing treatment in Ar. The reaction temperature was 600 °C and reaction time was 4 h.

Characterization

The morphologies and sizes of the samples were studied by field-emission scanning electron microscopy (SEM, Hitachi SU8000, Japan), transmission electron microscopy (TEM, Hitachi S-7650, Japan) and high resolution transmission electron microscopy (HTEM, JEM-2100, Japan). The X-ray diffraction (XRD) spectra of the samples were characterized by a D/max-γB X-ray diffractometer (Rigaku, Japan) with Cu Kα radiation (λ = 1.54178 Å). The diffraction data was obtained scanning from 10° to 90° with a rate of 0.02° s⁻¹. Raman spectrum was measured by a Raman spectrometer (Renishaw in Via, Germany) under laser excitation at 532 nm. The thermogravimetric was used to measure the carbon content on an STA449F3 (NETSCH, Germany).

Electrochemical measurement

The slurry of the working electrodes was prepared by dissolving 80 wt% active materials, 10 wt% acetylene black and 10 wt% polyvinylidene fluoride (PVDF) in an appropriate amount of N-methyl-2-pyrrolidone (NMP). Then the slurry was spread onto a Cu foil current collector and dried in a vacuum oven for 12 hours at 100 °C. 1 M solution of NaClO₄ in ethylene carbonate (EC) and propylene carbonate (PC) (1 : 1 in volume) was used as the electrolyte, and a glass fiber as the separator. The electrochemical performance of materials was carried out by the CR2025 coin-type cell with metallic sodium foil as the counter electrode in an argon-filled glove box. The average mass of all the electrodes disks with a diameter of 1.4 cm was controlled approximate 2 mg cm⁻².²² The galvanostatic charge–discharge tests were performed at a voltage window of 1.5–3.0 V using the Neware Battery Testing System. Note that “nC” means that the charge/discharge current is set up to achieve the nominal capacity in “1/n” hours (e.g. 1C corresponds to a full charge/discharge in 1 hour with a current density of 133 mA g⁻¹). The specific capacity was calculated on the basis of the as-prepared composites. Both of cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) measurements were conducted on GHI 660E electrochemical workstation. ZSimpWin software can calculate and analyse the parameters of some components according to the equivalent circuit.²³ The frequency range of the EIS measures was from 100 kHz to 10 mHz with an applied amplitude of 5 mV. All of the electrochemical tests were performed at room temperature apart from low temperature tests.

Results and discussion

The NTP has a 3D open framework structure, consisting of isolated TiO₆ octahedra and PO₄ tetrahedral interlinked by sharing their corner oxygen atoms (Scheme 1a).¹⁴ Sodium ions can occupy two types of interstitial sites (A1 and A2). A1 would be preferably occupied at low sodium content (n ≤ 1) which is in elongated octahedral oxygen environment at the intersection of three conduction channels. When n > 1, the sodium ions would be among the A1 and three 8-coordinate sites environment at each bend of the conduction channels.¹¹ As a result, the sodium ions could migrate easily between A1 and A2.¹² Despite all this, the poor electronic conductivity hinders fast transport of sodium ions. The designed architecture of the NTP/C–CNTs nanocomposites is illustrated in Scheme 1b emerging two advantages. On one hand, the crosslinked CNTs form a network among the NTP/C particles, which enhance the dispersity of NTP/C particles to improve electrochemical accessibility with the electrolyte. Internal diffusion resistance could be decreased resulting in good electrochemical performance. On the other hand, the CNTs provides efficient transport pathways for electrons to enhance the poor electronic conductivity of the NTP/C anodes. Thus, it could be speculated that the development of the novel construction is useful to the rate capacity and cyclability.

Scheme 1 The structure schematic illustration of NTP/C phase (a) and NTP/C–CNTs nanocomposites (b).

Fig. 1a shows the XRD patterns of the NTP/C nanoparticles and NTP/C–CNTs. All of the sharp diffraction peaks of both the samples are consistent well with the standard NTP peaks (JCPDS no. 33-1296), indicating the good crystallinity. Particularly, there is no evidence of diffraction peaks for CNT, representing the presence of CNTs does not influence the crystal structure of NTP due to the low content.²⁴ The crystal lattice parameters are a = 8.4984 Å, b = 8.4984 Å, c = 21.8233 Å for NTP/C, and a = 8.5219 Å, b = 8.5219 Å, c = 21.9011 Å for NTP/C–CNTs. The calculation results of these parameters of the two samples are respectively approximate to the values of NTP (a = 8.4913 Å, b = 8.4913 Å, c = 21.7858 Å). The Raman spectrum of NTP/C–CNTs further determined the existence of CNTs, as shown in Fig. 1b. Two characteristic bands of the carbon materials are significantly observed, which are located at around 1342 (D-band) and 1574 cm⁻¹ (G-band), attributed to the defects in graphitized structure and graphitized carbon, respectively. The results can prove the existence of both amorphous carbon and CNTs in the NTP/C–CNTs sample.

Fig. 1 (a) XRD patterns of NTP/C and NTP/C–CNTs comparing to the standard NTP crystallographic pattern [JCPDS no. 33-1296]; (b) Raman spectrum of NTP/C–CNTs.

Fig. 2 shows the typical SEM images of the NTP/C and NTP/C–CNTs products at different magnifications. The NTP/C synthesized by solvothermal method has an irregular particle morphology with the average diameter of less than 100 nm (Fig. 2a and b). As shown in Fig. 2c and d, the NTP/C nanoparticles in the NTP/C–CNTs composite are wrapped evenly by intertwined and overlapping carbon nanotubes, indicating a good conductive network for a close electronically connect between the particles. It is worth noting that the addition of CNTs enhances the dispersion of NTP/C particles to improve the electrolyte infiltration.²⁴ In addition, the microstructure of NTP/C and NTP/C–CNTs samples was deeper investigated by TEM and HRTEM techniques. The voids between irregular NTP/C particles can be seen from Fig. S1a.† The clearly spacing of lattice fringes corresponding to 4.38 Å (Fig. S1b†), 1.63 Å (Fig. S1c†), 1.81 Å (Fig. S1c†), 1.18 Å (Fig. S1d†) and 2.55 Å (Fig. S1d†) consistent well with the (104), (318), (042), (520) and (018) planes of NTP, indicating the high crystallinity of NTP crystals in the synthesized NTP/C (see Fig. S1b–d†). Simultaneously, the carbon layer coating on the surface of NTP particles can also be clearly observed (see Fig. S1b–d†), which may derive from the organic titanium source in the synthesis process.^15,25–27 As to NTP/C–CNTs, obviously, carbon nanotubes are tightly bound to the NTP/C particles, as displayed in Fig. 3a and S2.† The spacing of lattice fringes is 0.44 and 0.35 nm, consistent with the (104) and (202) planes of NTP (see Fig. 3b). The abundant CNTs could be fast transmission channels for electrons transport. Furthermore, elemental mapping demonstrates that carbon dispersed uniformly in the NTP array (Fig. 3c), indicative of a uniformly carbon layer and CNTs distribution. In order to determine the content of amorphous carbon and CNTs, the thermogravimetric (TG) test was carried out, and the test result was shown in Fig. S3.† As revealed form the TG curves, the content of amorphous carbon in the NTP/C sample is 1.8 wt% and the total content of carbon in the NTP/C–CNTs is 7.4 wt%, suggesting the content of CNTs in NTP/C–CNTs is approximately 5.6 wt%, which is almost coincide to the addition ratio of original CNTs.

Fig. 2 SEM images of NTP/C (a and b) and NTP/C–CNTs (c and d) at different magnification.

Fig. 3 TEM (a), HRTEM (b) and bright-field STEM (c) images of NTP/C–CNTs and elements (Na, Ti, P, O, C) mapping of the same region.

Cyclic voltammetry (CV) measurements were used to discuss the sodium ion insertion/extraction progress. CV curves of NTP/C and NTP/C–CNTs were carried out at a fixed scan rate of 0.1 mV s⁻¹ (Fig. 4a), through assembling half cells with sodium metal which is equal to the counter and reference electrode. A couple of redox peaks of Ti⁴⁺ to Ti³⁺ is observed, corresponding to the sodium ion insertion/extraction reaction in the NTP crystal: NaTi₂(PO₄)₃ + 2Na⁺ + 2e⁻ ↔ Na₃Ti₂(PO₄)₃.²⁷ Meanwhile, the overpotential of the NTP/C–CNTs sample (330 mV) is lower than that of NTP/C (420 mV) between the anodic and cathodic peaks. The NTP/C–CNTs demonstrates narrower peak shape and higher peak current compared with the pure NTP/C, indicating higher electrochemical activity and lower electrochemical polarization. Thus, the CNTs network has great influence on electrode kinetics and the optimized NTP/C–CNTs nanocomposite electrode displays much better electrochemical performance as follow discussed.

Fig. 4 Electrochemical performance of NTP/C and NTP/C–CNTs electrodes. (a) The CV curves in a voltage range of 1.2 to 2.8 V at 0.1 mV s⁻¹. (b) Discharge–charge profiles at 0.5C. (c) The Nyquist plots in fresh cells. (d) Variations and fittings between Z_re and the reciprocal square root of the angular frequency in the low frequency region.

Fig. 4b compares the charge–discharge profiles of both NTP/C and NTP/C–CNTs in a voltage window of 1.5–3.0 V at the rate of 0.5C. There is a distinct discharge/charge profile with a flat plateau around 2.1 V, indicative of good reversibility for Na⁺ intercalation/deintercalation in accordance with CV curves. In addition, a narrower gap between the charge and discharge profiles of the NTP/C–CNTs electrode is observable in comparison to the NTP/C sample, since the CNTs modified structure enhances the electrochemical activity. These results reveal that addition of CNTs component could be responsible for the improved Na⁺ insertion–extraction kinetics.

The data of the electrochemical impedance spectroscopy (EIS) can provide two aspects of forceful information used to study the relationship between the structure of the electrode materials and electrochemical performance. Fig. 4c shows the Nyquist plots of the two electrodes in fresh cells with an equivalent circuit inset.¹³ A depressed semicircle in high-medium frequency and an oblique straight line in the low frequency are shown in each of the impedance spectra of the two samples, representing the diffusion of the sodium ion between the electrode and electrolyte and the sodium ion diffusion in the active electrode material, respectively. On one hand, the charge-transfer resistance (R_ct, Ω) value can be calculated by fitting the data of the semicircle area, based on the equivalent circuit. The fitting results exhibit that the R_ct value of the NTP/C–CNTs electrode is much smaller than that of the NTP/C electrode, as shown in Table S1.† In other words, the CNTs as additives could indeed improve the kinetics of the NTP/C material, attributed to the improved electron transfer process. On the other hand, the straight line is related to Na⁺ intercalation/de-intercalation in the interfaces of active materials. The values of Na⁺ diffusion coefficient D (cm² s⁻¹) can be calculated from the straight lines area through the following equation:

D = R²T²/2A²n⁴F⁴C²σ² (1)

Among them, R, T, A, n, F and C are constant terms representing the gas constant, the absolute temperature, the surface area of the electrode, the number of electrons per molecule during oxidization, the Faraday constant, the concentration and the Warburg factor, respectively.²⁸ Because σ is proportional relations with Z_re (Z_re ∝ σω^−1/2), the value of σ can be obtained through fitting line of Z_re and the reciprocal square root of the angular frequency ω. The fitting results indicate that the σ values of the NTP/C–CNTs electrode and the NTP/C electrode are 2384.9 Ω s^−1/2 and 6728.1 Ω s^−1/2, respectively (Fig. 4d). According to the above-mentioned equation, the order of Na⁺ diffusion coefficient was NTP/C–CNTs > NTP/C. Consequently, it sufficiently proves the effectiveness of the incorporation of the CNTs around the NTP/C particles in improving the reversible capacity of NTP/C anode with excellent conductivity.

Fig. 5a shows the discharge–charge profiles of NTP/C–CNTs electrode at various rates. As current density increasing from 1 to 50C, the discharge capacities of the NTP/C–CNTs nanocomposite electrodes are 116.8, 113.3, 110.0, 107.4, 104.9 and 103.4 mA h g⁻¹. Even under the ultrahigh current density of 50C, the discharge capacity could reach 78% of theoretical capacity demonstrating its excellent rate performance. Notably, the capacity fading of NTP/C–CNTs sample at high C-rates is minimal, which could be associated with the CNTs network. Because the CNTs may be capable of facilitating the electrons transfer due to a large amount of fast transport pathways and enhancing the accessibility between the NTP/C particles and electrolyte. There is an additional charge plateau located at around 2.0 V, especially under higher current rates. In the NASICON NTP framework, Na⁺ can occupy two different sites and migrate easily between each other as local heating induced by higher charge/discharge rates.³⁰ Such the phenomenon is also reported in other NASICON-type materials, which may be caused by the structural rearrangement due to temperature variation.^30,31

Fig. 5 (a) Discharge–charge profiles of NTP/C–CNTs electrode at various rates. (b) Comparison of the rate performance of the nanocomposite electrodes with different CNTs content (2%, 5%, 10%) and the NTP/C electrode from 0.5 to 50C. (c) Comparison of the cycling performance of NTP/C and NTP/C–CNTs electrodes at 1C. (d) Long-term cycling performance of the NTP/C–CNTs electrodes at a high current density of 50C for 1000 cycles.

While adding amount of CNTs in the electrode materials can influence on the electrochemical performance, we compared the discharge capabilities of NTP/C–CNTs electrodes with different CNTs content (2, 5 and 10 wt%) from 0.5 to 50C in order to investigate the most appropriate additional proportion (Fig. 5b). The discharge capacities of NTP/C–CNTs electrodes are 102.6 mA h g⁻¹ (2 wt%), 116.4 mA h g⁻¹ (5 wt%) and 102.2 mA h g⁻¹ (10 wt%) under a 0.5C rate respectively, while NTP/C about 83.6 mA h g⁻¹. Meanwhile, the discharge capacities of the samples with CNTs content of 2 wt% and 10 wt% have not a significant difference at low rates and the discharge capacity gap is wider as the current increases. The CNTs as additives can improve the kinetics and utilizations of NTP/C effectively, but excessive CNTs leads to decrease in discharge capacity because of substitute for too much NTP/C. It is worth noting that the NTP–CNTs sample (5 wt%) imports excellent high-rate capacity of 102.6 mA h g⁻¹ at 50C rate. These results exhibit that the most appropriate content of CNTs is 5 wt% to guarantee the electronic conductivity and capacity of the NTP/C–CNTs electrode.

The cyclic stability, as well as rate capability, is another vital parameter of the electrode material performance in practical applications.^32,33 By comparing the long-life cycling performance of the both samples (NTP/C and NTP/C–CNTs) at 1C, CNTs as conducting agent have great influence to cyclic stability. As shown in Fig. 5c, the discharge capacity of the NTP/C sample is decreased rapidly from 91.1 mA h g⁻¹ to 81.7 mA h g⁻¹. Conversely, the initial discharge capacity of NTP/C–CNTs is 112.3 mA h g⁻¹ and the final is 108.5 mA h g⁻¹, reducing about 3%. The specific capacity of the NTP/C–CNTs is weakened barely after 200 times of charging and discharging circulation. It can be seen, the microstructure and crystal structure maintains well after long-term cycling, indicative of its good cyclic stability (see Fig. S4 and S5†). In order to further study the stability of the NTP/C–CNTs electrodes, the ultralong-life cycling performance at 50C is investigated in Fig. 5d. After 1000 cycles, the NTP/C–CNTs retains 98% of its initial discharge capacity (96.0 mA h g⁻¹ and 94.2 mA h g⁻¹) with unconspicuous capacity decay. In addition, the coulombic efficiency of the NTP/C–CNTs reaches approximately 100% during the whole cycling process. The good cycling performance may be attitude to the CNTs network for providing an effective conductive with abundant voids to buffer the volumetric change during charge and discharge.

Good electrochemical performance of NTP anode material in low-temperature environment should be another serious challenge for its development. It is can be expected that the electrochemical performance of NTP become worse suffering from the low temperature due to the bigger diffusion resistance.³⁴ Thus, the CNTs decorated NTP/C anode will display greater advantages in this situation.

Fig. 6 displays the comparison of rate performance of the pure NTP/C and the NTP/C–CNTs electrodes with various low temperature (0, −10 and −20 °C). The average discharge capacities of the NTP/C–CNTs sample under 0.5C rate are about 117.9 mA h g⁻¹ at 0 °C, 117.1 mA h g⁻¹ at −10 °C and 113.2 mA h g⁻¹ at −20 °C, which are much higher than those of NTP/C. The NTP/C–CNTs sample exhibits an excellent rate capability due to its less irreversible capacity loss with the improved electrochemical kinetics. By comparing NTP/C and NTP/C–CNTs electrodes, the capacities of the form sample not only are far lower than the latter, but also decrease more sharply as lower temperature and higher current density. Even under 10C rate at −210 °C, the specific capacity of NTP/C–CNTs is 62.2 mA h g⁻¹, reducing 39% from 0 °C (106.9 mA h g⁻¹) to −20 °C. Nevertheless, the discharge capacity of the NTP/C sample is almost 0, representing the poor low-temperature property. However, when the current density is restored to 0.5C, the discharge capacities of the both samples can recover to the initial value representing the excellent cycle reversibility. According to the above-mentioned results, the CNTs network has obviously advantages in the low temperature environment. Thus, the NTP/C–CNTs exhibits an improved low-temperature performance for expand its application field.

Fig. 6 The rate performance of the pure NTP/C (a) and NTP/C–CNTs (b) electrodes at different low temperatures.

Fig. 7a compares the CV curves of the NTP/C–CNTs electrode under different low temperatures at a scanning rate of 0.1 mV s⁻¹. The curves demonstrate not only a pair of anodic and cathodic peaks, but also an additional small cathodic peak corresponding to another charging platform in consistent with previous observation.^29–32 It also finds that temperature has a significant impact on the electrode kinetics because of the increasing overpotential for C₁ and C₂ peaks with the decreasing temperature. Furthermore, the intercalation/de-intercalation of sodium ion is a typical diffusion process so that the electrode reaction can be assumed to be a semi-infinite diffusion. The peak current of the CV (I_p) follows the relationship with the Na⁺ diffusion coefficient (D, cm² s⁻¹):

I_p = 2.69 × 10⁵n^3/2AC₀D^1/2ν^1/2 (2)

Fig. 7 Electrochemical performance of the NTP/C–CNTs electrodes in different low temperature. (a) CV curves in a voltage range of 1.2 to 2.8 V at 0.1 mV s⁻¹. (b) Discharge–charge profiles at 0.5C. (c) EIS spectra in a frequency range of 0.1 to 100 kHz. (d) Variations and fittings between Z_re and the reciprocal square root of the angular frequency in the low frequency region.

Among them, n (for NTP, n = 1), A (cm²), C₀ (mol cm⁻³) and ν (mV s⁻¹) are constant terms representing the number of electrons transferred per molecule, the active surface area of the electrode, the concentration of Na⁺ and the scanning rate, respectively.²⁸ As can been seen from the CV curves, the order of I_p is 0 °C > −10 °C > −20 °C. Namely, the D value of 0 °C is the largest. The result suggests that the low temperature is not profitable to Na⁺ diffusion process. This may actually be the main influence factor for the decreasing specific capacity suffering from the low temperature.

The charging and discharging voltage curves of NTP/C–CNTs electrode under 0.5C at various low temperatures are shown in Fig. 7b. The voltage differences between charge and discharge platforms are increasing with the lowing temperature, which represents higher electrochemical polarization. Moreover, there are two charge platforms appeared at 0.5C in accordance with CV measure results. In the case of large fluctuations in temperature, the phenomenon of structural rearrangement could be more obvious.

The results of the Nyquist plots from AC impedance spectroscopy measurements demonstrated in Fig. 7c. Meanwhile, the inset shows the simplified equivalent circuit, R_s representing the ohmic resistance, R_ct denoting the charge transfer resistance and W signifying the Warburg resistance. As can be seen from the high-medium frequency, the diameter of semicircle at 0 °C is much smaller than that at −20 °C. Correspondingly, the value of R_ct at −20 °C is the biggest, representing the largest internal resistance during the diffusion of Na⁺ between the electrode and electrolyte. These results adequately prove that low temperature leads to growing charge-transfer resistance. In addition, compared with the NTP/C electrode, the values of R_ct are considerably larger than the composite electrode. After fitting the data from the low frequency region, the σ values of the NTP/C–CNTs electrode at various low temperature are 468.55 Ω s^−1/2 (0 °C), 734.70 Ω s^−1/2 (−10 °C) and 2710.18 Ω s^−1/2 (−20 °C), respectively (Fig. 7d).

Conclusions

In summary, we have designed and fabricated the CNTs decorated NTP/C composite electrodes to achieve excellent electrochemical performance. The CNTs network decreases the internal diffusion resistance and provides fast transport pathways for Na⁺ and electron to improve the poor electronic conductivity of the NTP/C anodes. By properly a series of electrochemical tests, the designed NTP/C–CNTs anodes exhibits remarkable high-rate capability (about 88% capacity retention at a high rate of up to 50C) and cycling stability (about 98% capacity retention at 50C of 1000 cycles). Moreover, the low-temperature performance is also superior to the NTP/C (about 65.2 mA h g⁻¹ at 10C in the temperature of 20 °C below zero). In view of the advantages of the electrode composite design, we anticipate that the nanocomposites might be promising anode materials for long-life and low-temperature rechargeable sodium-ion batteries.

Acknowledgements

This work was supported financially by the National Natural Science Foundation of China (No. 50974045) and the Fundamental Research Funds for the Central Universities (Grant No. HIT.NSRIF.2017024).

Notes and references

1. B. Kang and G. Ceder, Nature, 2009, 458, 190.
2. X. Lai, J. E. Halpert and D. Wang, Energy Environ., 2012, 5, 5604.
3. B. Wang, W. Al Abdulla, D. Wang and X. S. Zhao, Energy Environ. Sci., 2015, 8, 869–875.
4. T. Liu, J. Qiu, B. Wang, Y. Wang, D. Wang and S. Zhang, RSC Adv., 2015, 5, 100018–100023.
5. M. D. Slater, D. Kim, E. Lee and C. S. Johnson, Adv. Funct. Mater., 2013, 23, 947–958.
6. D. Wang, Q. Liu, C. Chen, M. Li, X. Meng, X. Bie, Y. Wei, Y. Huang, F. Du, C. Wang and G. Chen, ACS Appl. Mater. Interfaces, 2016, 8, 2238–2246.
7. S. I. Park, I. Gocheva, S. Okada and J.-i. Yamaki, J. Electrochem. Soc., 2011, 158, A1067–A1070.
8. G. Pang, P. Nie, C. Yuan, L. Shen, X. Zhang, J. Zhu and B. Ding, Energ. Tech., 2014, 2(8), 705–712.
9. X. Wu, Y. Cao, X. Ai, J. Qian and H. Yang, Electrochem. Commun., 2013, 31, 145–148.
10. X. Y. Wu, M. Y. Sun, Y. F. Shen, J. F. Qian, Y. L. Cao, X. P. Ai and H. X. Yang, ChemSusChem, 2014, 7, 407–411.
11. G. Pang, P. Nie, C. Yuan, L. Shen, X. Zhang, H. Li and C. Zhang, J. Mater. Chem. A, 2014, 2, 20659–20666.
12. G. Pang, C. Yuan, P. Nie, B. Ding, J. Zhu and X. Zhang, Nanoscale, 2014, 6, 6328–6334.
13. J. Yang, H. Wang, P. Hu, J. Qi, L. Guo and L. Wang, Small, 2015, 11, 3744–3749.
14. C. Wu, P. Kopold, Y. Ding, P. Aken, J. Maier and Y. Yu, ACS Nano, 2015, 9, 6610–6618.
15. Y. Jiang, J. Shi, M. Wang, L. Zeng, L. Gu and Y. Yu, ACS Appl. Mater. Interfaces, 2016, 8, 689–695.
16. T. Yamada, Y. Hayamizu, Y. Yamamoto, Y. Yomogida, A. Najafabadi, D. Futaba and K. Hata, Nat. Nanotechnol., 2011, 6, 296–301.
17. Y. Shang, X. He, Y. Li, L. Zhang, Z. Li, C. Ji, E. Shi, P. Li, K. Zhu, Q. Peng, C. Wang, X. Zhang, R. Wang, J. Wei, K. Wang, H. Zhu, D. Wu and A. Cao, Adv. Mater., 2012, 24, 2896–2900.
18. C. Yu, C. Masarapu, J. Rong, B. Wei and H. Jiang, Adv. Mater., 2009, 21, 4793–4797.
19. B. J. Landi, M. J. Ganter, C. D. Cress, R. A. DiLeo and R. P. Raffaelle, Energy Environ. Sci., 2009, 2, 638.
20. Y. Zhou, J. Wang, Y. Hu, R. O'Hayre and Z. Shao, Chem. Commun., 2010, 46, 7151–7153.
21. Z. Wu, X. Han, J. Zheng, Y. Wei, R. Qiao, F. Shen, J. Dai, L. Hu, K. Xu, Y. Lin, W. Yang and F. Pan, Nano Lett., 2014, 14, 4700–4706.
22. B. Wang, A. Liu, W. A. Abdulla, D. Wang and X. S. Zhao, Nanoscale, 2015, 7, 8819–8828.
23. B. Wang, B. Xu, T. Liu, P. Liu, C. Guo, S. Wang, Q. Wang, Z. Xiong, D. Wang and X. S. Zhao, Nanoscale, 2014, 6, 986–995.
24. W. Wu, J. Yan, A. Wise, A. Rutt and J. F. Whitacre, J. Electrochem. Soc., 2014, 161, A561–A567.
25. Y. Jiang, L. Zeng, J. Wang, W. Li, F. Pan and Y. Yu, Nanoscale, 2015, 7, 14723–14729.
26. A. I. Mohamed, N. J. Sansone, B. Kuei, N. R. Washburn and J. F. Whitacre, J. Electrochem. Soc., 2015, 162, A2201–A2207.
27. B. Zhao, Q. Wang, S. Zhang and C. Deng, J. Mater. Chem. A, 2015, 3, 12089–12096.
28. X.-L. Wu, Y.-G. Guo, J. Su, J.-W. Xiong, Y.-L. Zhang and L.-J. Wan, Adv. Energy Mater., 2013, 3, 1155–1160.
29. G. Yang, H. Song, M. Wu and C. Wang, J. Mater. Chem. A, 2015, 3, 18718–18726.
30. K. Saravanan, C. W. Mason, A. Rudola, K. H. Wong and P. Balaya, Adv. Energy Mater., 2013, 3, 444–450.
31. C. Masquelier, C. Wurm, J. Carvajal, J. Gaubicher and L. Nazar, Chem. Mater., 2000, 12, 525–532.
32. X. Li, X. Zhu, J. Liang, Z. Hou, Y. Wang, N. Lin, Y. Zhu and Y. Qian, J. Electrochem. Soc., 2014, 161, A1181–A1187.
33. B. Wang, T. Liu, A. Liu, G. Liu, L. Wang, T. Gao, D. Wang and X. S. Zhao, Adv. Energy Mater., 2016, 1600426,  DOI:10.1002/aenm.201600426.
34. X. H. Rui, Y. Jin, X. Y. Feng, L. C. Zhang and C. H. Chen, J. Power Sources, 2011, 196, 2109–2114.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra11042h

---