Nowadays, commercial lithium-ion batteries (LIBs) are widely used for energy storage in our daily lives because of their high energy density and good cycling performance. However, safety and environmental issues, which arise from inflammable and toxic organic electrolytes, hinder their large-scale applications in some fields^[1-4]. In contrast, aqueous sodium-ion batteries employ non-toxic aqueous electrolyte that is environmentally friendly and improves the safety of the batteries^[5-8]. Besides, worldwide sodium reserves are much larger than lithium reserves, which could potentially reduce the cost of the sodium-ion batteries. Hence, high-performance aqueous sodium-ion batteries can be considered a promising alternative in some specific large-scale applications.

Unfortunately, the energy density of the aqueous sodium-ion batteries is lower than that of LIBs. Both the narrow operational potential window and the limited capacity of the electrodes restrict the energy density of the batteries. Broadening the potential window would lead to either side reactions involving the aqueous electrolyte or irreversible phase transition of the electrodes. Therefore, exploring high-capacity electrodes at this limited potential window would be a more efficient in enhancing energy density of the batteries. In the search for high-capacity Na-ion cathodes, many types of active materials have been developed, such as oxides, polyanions, and Prussian blue analogs^[9-18]. Among them, Na_0.44MnO₂ has drawn researchers’ attention due to its large and stable S-shape and O-shape tunnels which facilitate Na ion insertion and extraction. The theoretical capacity of Na_0.44MnO₂ cathode is as high as 121 mAh/g. Besides, Mn-based oxides are low cost and environmentally benign materials^[19-20]. Whitacre’s work firstly demonstrated that a Na_0.44MnO₂ cathode could offer a long cycle life with a cycling capacity of 30-40 mAh/g at lower rates (0.2 C and 1.2 C) when charged and discharged from between −0.1 V and 0.7 V. This capacity would decrease to less than 20 mAh/g if the cathodes were cycled at a high rate of 18 C^[21]. Low electrical conductivity or low Na-ion diffusivity are considered as possible causes of poor cycling stability and rate capability. Increasing the interfacial area between the electrolyte and the electrode by decreasing the size or dimension of the Na_0.44MnO₂ cathode would be available to improve the conductivity and the Na-ion diffusivity.

In this work, Na_0.44MnO₂ with different particle sizes are prepared to investigate the size effect on its cathode performance. Among these Na_0.44MnO₂, the nanorod one is found improved performance. By using α-MnO₂ nanoflakes as precursors, the Na_0.44MnO₂ nanorods were successfully prepared with a width of 200 nm. Such a size could enhance the electrical conductivity and the Na-ion diffusivity of the cathode. The as-prepared cathode shows a higher cycling capacity and lower polarization at a low rate of 1 C and a high rate of 5 C. The initial discharging capacity of Na_0.44MnO₂ nanorods cathode reaches 60 mAh/g at 1 C, with 55 mAh/g retained after 200 cycles. A high capacity of 47 mAh/g is delivered after 200 cycles at a high rate of 5 C. The high stability of thin Na_0.44MnO₂ nanorods is due to the stable morphology and crystal structure after long cycles.

Synthesis of Na_0.44MnO₂ nanorods. KMnO₄ solution (0.01 mol/L) was heated at 100℃. Cetrimonium bromide (CTAB) solution (0.05 mol/L) was heated at 140℃. 100 mL of KMnO₄ solution was injected into 100 mL of CTAB solution. The color of the obtained solution gradually turned from purple to brown. 500 mL of acetone was added to the solution and brown precipitates were obtained. The precipitates were washed with deionized water by centrifugation and dried at 80℃ for 24 h. The dried powder was added to 30 mL aqueous NaBH₄ solution. The mixture was kept stirring at room temperature for 24 h. The obtained power was washed with deionized water by filtration and dried at 80℃ for 24 h. The dried powder was mixed and ground with NaHCO₃ in a mass ratio of 2:1, and then heated in air at 500℃ for 5 h. After cooling down, the obtained powered was ground again and then heated at 900℃ in air for 12 h.

Synthesis of Na_0.44MnO₂ microrods. The synthesis procedure was the same as that for Na_0.44MnO₂ nanorods, except 0.1 mol/L KMnO₄ solution was used, instead of 0.01 mol/L.

Synthesis of Na_0.44MnO₂ bulk. Mn₂O₃ and NaHCO₃ were mixed and ground with 10% excess of Na in stoichiometry with the molar ratio of manganese atoms to sodium atoms as 1:0.484. The mixture was heated in air at 500℃ for 5 h. After cooling down, the obtained powered was ground again and then heated at 900℃ in air for 12 h.

Material Characterizations. X-ray diffraction (XRD) patterns were collected on a Panalytical XPert Pro X-ray diffractometer with Cu Kα radiation. Field emission scanning electron microscopy (FESEM) analysis was carried out on ZEISS Supra 55. High-resolution transmission electron microscopy (HRTEM) analysis was carried out using a JEOL 2100F at 200 kV. N₂ (77 K) adsorption test was conducted on an ASAP Tristar Ⅱ 3020. Brunauer−Emmett−Teller (BET) was used for the surface area determination. Inductively coupled plasma optical emission spectrometer (ICP-OES) test was conducted on a PerkinElmer@Optima 8300.

Electrochemical Measurements. Three-electrode system was used for the evaluation of the Na_0.44MnO₂. The working electrode was prepared by mixing Na_0.44MnO₂, single-walled carbon nanotube (SWCNT), and polytetrafluoroethylene (PTFE) at 8:1:1, followed by pressing the mixture on a 4 cm² Ni foam using a hydraulic press machine with a pressure of 10 MPa. The active mass loading was about 30 mg. A Pt foil was employed as a counter electrode and an Ag/AgCl (saturated NaCl) electrode was employed as a reference electrode. The electrolyte was 1 mol/L Na₂SO₄ in deionized water. Cyclic voltammetry (CV) measurement was conducted in a voltage range of −0.2 V to 0.8 V vs. Ag/AgCl at a scan rate of 0.1 mV/s. For galvanostatic charge-discharge (GCD) tests, the voltage window was controlled between −0.1 V and 0.7 V. 1 C corresponds to 121 mA/g in this work. Electrochemical impedance spectroscopy (EIS) was tested in the frequency range of 10 mHz to 100 kHz, and the perturbation amplitude is 5 mV. CV, GCD, and EIS were all measured on a Biological potentiostat electrochemical station.

XRD was used to determine the structure of the prepared materials (Fig. 1(a)). All the diffraction peaks of the three materials can be assigned to the orthorhombic Na_0.44MnO₂ phase (Fig. 1(b)) ^[23]. The diffraction peaks are sharp and intense, which indicates the high crystallinity of the materials. HRTEM analysis shows interplanar spacing of 0.26 nm (Fig. 1(c)) and 0.45 nm (Fig. 1(d)) for Na_0.44MnO₂ nanorods and microrods, respectively. They correspond to (3 5 0) and (2 0 0) crystal planes of orthorhombic Na_0.44MnO₂, respectively. For Na_0.44MnO₂ bulk (Fig. 1(f)), interplanar spacings of 0.28 nm and 0.45 nm were determined, corresponding to (0 0 1), and (2 0 0) crystal planes, respectively. The ratios of Na/Mn in the three prepared materials were examined by ICP-OES (Table 1). They are all very close to 0.44. These demonstrate all the three prepared materials are Na_0.44MnO₂ with an orthorhombic structure.

The morphology and dimensions of the samples can be observed from SEM images displayed in Fig. 1(g), (h), and (i). Na_0.44MnO₂ bulk show varying dimensions with a wide size distribution. The bulk have lengths of 3−10 μm and widths of 0.5−1 μm. Na_0.44MnO₂ microrods have a smaller dimension than Na_0.44MnO₂ bulk, with a narrow size distribution and regular shapes. Its widths are smaller than 0.5 μm. A uniform rod shape can be observed in images of Na_0.44MnO₂ nanorods with widths of about 200 nm. The lengths of the nanorods range from 2 to 10 μm. The difference in morphology between the three samples may be caused by the using of CTAB as a surfactant. Na_0.44MnO₂ nanorods are prepared by using the largest CTAB/KMnO₄ ratio, and its widths are the smallest. Na_0.44MnO₂ microrods are prepared by using the smaller CTAB/KMnO₄ ratio, and its widths are larger than Na_0.44MnO₂ nanorods. CTAB is not needed when preparing Na_0.44MnO₂ bulk. The precursor of Na_0.44MnO₂ bulk, Mn₂O₃, has a wide size distribution and irregular shapes, leading to the inhomogeneous product after the solid-state reaction.

N₂ adsorption/desorption isotherms of the prepared materials are shown in Fig.1(c). All the samples display III-type isotherm curves with no hysteresis loop, indicating the materials are non-porous. The BET surface areas and average particle sizes based on this test are listed in Table 1. Na_0.44MnO₂ nanorods show the highest BET surface area of 4.1 m²/g and the lowest average particle size of 1.45 μm, while Na_0.44MnO₂ bulk have the lowest BET surface area of 0.9 m²/g and largest average particle size of 6.64 μm. The BET surface area of Na_0.44MnO₂ microrods is 2.7 m²/g, and its average particle size is 2.20 μm. The trend that the three materials show in the isotherms is consistent with the SEM images.

Fig. 2(a) shows the typical CV curves (6^th CV cycle) of the three materials. Five anodic peaks can be observed for all the three materials, i.e., peaks A, B, C, D, and E. Peak A corresponds to the extraction of Na in the Na3 site in the S-shaped tunnel (Fig. 1(b)). ^[24-26] Peaks B, D, and E are related to the 3-step Na extraction in the Na2 site in the S-shaped tunnel. Peak C is due to the extraction of Na in O-shaped tunnel (Na1 site). Multiple cathodic peaks are also observed, i.e., peaks A’ (A’^,1 and A’^,2), B’, C’, D’, and E’, corresponding to A, B, C, D, and E, respectively. There are two cathodic peaks (A’^,1 and A’^,2) corresponding to the anodic peak A, indicating the insertion/extraction of Na in the Na3 site is 2-step processes. The differences between the cathodic (2 peaks) and anodic (1 peak) processes for Na insertion/extraction in the Na3 site could be due to different kinetics during Na insertion/extraction. Only one anodic peak A observed is possibly because of the overlapping of the 2-step process.

Cycle performances of the three cathodes can be observed in Fig. 2(b) and (c). The corresponding galvanostatic charge/discharge (GCD) results of the three cathodes can be seen in Fig. 3. It is observed in GCD results that Na_0.44MnO₂ bulk cathode shows much higher polarization than the other 2 cathodes, especially at 5 C cycle. It should be attributed to the size differences of the materials. Na_0.44MnO₂ bulk has the largest particle size and the smallest BET surface area which could lead to the poorest conductivity and highest polarization among the three materials. Moreover, the initial discharging capacity of Na_0.44MnO₂ nanorods cathode is 60 mAh/g, and 55 mAh/g is retained after 200 cycles. Na_0.44MnO₂ microrods cathode delivers a capacity of 55 mAh/g in the 1^st cycle and maintains 44 mAh/g in the 200^th cycle. In contrast, the Na_0.44MnO₂ bulk cathode can only deliver a 46 mAh/g capacity in the first cycle, which decreases to 40 mAh/g after 200 cycles. The Na_0.44MnO₂ nanorods cathode exhibits a 37.5% higher capacity than Na_0.44MnO₂ bulk cathode due to the size difference of the active material. The polarization of Na_0.44MnO₂ bulk cathode increases severely at 5 C, while the polarization of Na_0.44MnO₂ nanorods cathode keeps stable in the first 50 cycles. The discharging capacity of Na0_.44MnO₂ nanorods cathode reaches 53 mAh/g at 5 C in the first cycle and maintains 47 mAh/g after 200 cycles. Na_0.44MnO₂ microrods cathode delivers a 44 mAh/g in the 1^st cycle, while its capacity gradually increases with cycling and attains 47 mAh/g in the 200^th cycle. This cycling capacity is the same as the Na_0.44MnO₂ nanorods cathode in the 200^th cycle. In contrast, Na_0.44MnO₂ bulk cathode can only deliver 32 mAh/g in the initial cycle when discharging, and this drops to 19 mAh/g after 200 cycles. The cycling capacity of Na_0.44MnO₂ nanorods at 5 C is 147 % higher than that of Na_0.44MnO₂ bulk cathode. The Na_0.44MnO₂ nanorods have the largest specific surface area and the smallest average particle size, which is good for the contact between the active material and the electrolyte. Hence, this cathode delivers the highest cycling capacity at 1 C and at 5 C.

The Na_0.44MnO₂ nanorods cathode exhibits correspondingly higher capacities in the rate test (Fig. 2(d)). The nanorod cathode delivers a very high capacity of 84 mAh/g at 0.1 C in the second cycle, 58 mAh/g at 1 C, 49 mAh/g at 5 C, 41 mAh/g at 10 C, respectively. It recovers to 56 mAh/g in the second cycle back at 1 C. All these capacities are the highest among the three cathodes. Na_0.44MnO₂ microrods cathode shows lower capacities than Na_0.44MnO₂ nanorods cathode and higher capacities than Na_0.44MnO₂ bulk cathode at each rate. Meanwhile, the bulk cathode can only deliver 63, 47, 33, and 18 mAh/g at 0.1, 1, 5, and 10 C, respectively. It can only recover to just 39 mAh/g when the rate is back to 1 C. All these capacities are the lowest among the cathodes. The excellent rate performance of Na_0.44MnO₂ nanorods cathode implies wide applications at different rates.

To reveal the cause of the higher cycling capacity and the higher rate capacity of Na_0.44MnO₂ nanorods cathode, EIS of the cathodes has been conducted. Fig. 4 shows the Nyquist plots of three pristine cathodes. All the Nyquist plots display a semicircle at medium frequency followed by a slope at low frequency. The value that the semicircle begins corresponds to R_s (ohmic resistance of the electrode and electrolyte). The diameter of the half-circle corresponds to R_ct (charge transfer resistance). The slope depends on the diffusivity of ions that are conducted in the electrode. The Na_0.44MnO₂ bulk cathode shows a R_s and R_ct values of 2.0 and 1.2 Ω, respectively. The Na_0.44MnO₂ nanorods cathode shows a R_s of 1.9 and a R_ct and 3.0 Ω, respectively. The Na_0.44MnO₂ microrods cathode shows a R_s of 1.5 and a R_ct and 0.9 Ω, respectively. The three cathodes have the close ohmic resistances. For R_ct, The Na_0.44MnO₂ nanorods cathode owns the highest R_ct among the three cathodes, while the Na_0.44MnO₂ microrods cathode owns lowest R_ct.

The diffusivity of Na ion (D_Na) on the cathodes can be calculated to check the kinetics of the cathodes (Table 2). The equation calculating D_Na is shown as following: D_Na=R²T²/2A²n⁴F⁴C²σ², where R is gas constant, T is the testing temperature, A is the surface area of the cathode, n is the number of transferred electrons, F is Faraday constant, C is the concentration of the Na ion in the electrode, σ is Warburg coefficient^[26]. By plotting Z_real and ω^−1/2 at low frequency, σ can be obtained by calculating the slope of the fitting line. D_Na of Na_0.44MnO₂ nanorods cathode is calculated to be 4.2×10⁻¹⁴ cm²s⁻¹. D_Na of Na_0.44MnO₂ microrods cathode is calculated to be 2.7×10⁻¹⁴ cm²s⁻¹ which is smaller than that of Na_0.44MnO₂ nanorods cathode. The Na_0.44MnO₂ bulk cathode has the smallest D_Na of 1.3×10⁻¹⁴ cm²s⁻¹. The largest D_Na of Na_0.44MnO₂ nanorods cathode demonstrates that the nanorods structure, with the highest BET surface area and the smallest particle size, can facilitate the Na ion transfer in the active materials during electrochemical reactions. This high Na ion can diffusivity contributes to the high capacity of this cathode. The relationship between cycling capacity, D_Na, and specific surface area is displayed in Fig. 5. It can be observed that the larger specific surface area corresponds to the higher D_Na and leads to the larger cycling capacity at 1 C. The similar trends of the two profiles in this figure reveal the potential influence of D_Na on the cycling capacity. It can be deduced that if the BET surface area of Na_0.44MnO₂ turns larger, its D_Na can be higher, and the cycling capacity of the cathode can be further enhanced.

The crystal information of the electrodes before and after the cycling test can be studied by XRD and HRTEM. XRD patterns of the samples and the corresponding cathodes before and after the cycling test are illustrated in Fig. 6(a), (c), and (e). No difference is shown in peak positions between the patterns of the cathodes before and after the cycling tests, respectively. It means no new phase could be observed in the XRD patterns. The crystal lattices of the three cathodes after 5 C cycling are found in HRTEM images, Fig. 6(b), (d), and (f), respectively. The SAED patterns can confirm the existence of the crystals. The interplanar spacing of 0.28 nm and 0.45 nm can be correlated to (0 0 1) and (2 0 0) crystal plane, while the interplanar spacing of 0.26 nm can be correlated to (3 5 0) crystal plane. These parameters are the same as those of the active materials. The single crystal morphology and the maintained crystal interplanar spacings demonstrate that after 200 cycles at a high rate of 5 C, these three cathodes keep the lattices, and no phase transition happens after the cycling tests.

In summary, Na_0.44MnO₂ with different sizes have been prepared by reducing KMnO₄ in hot water followed by heat treatment with NaHCO₃. They are used in cathodes for aqueous Na-ion batteries. The nanorods with a width of ~ 200 nm gives the best cycling capacity and rate capacity. The initial discharging capacity of Na_0.44MnO₂ nanorods cathode attains 60 mAh/g at 1 C and it can remain 55 mAh/g after 200 cycles which is 37.5% higher than Na_0.44MnO₂ bulk cathode. At 5 C, the discharging capacity of Na_0.44MnO₂ nanorods cathode reaches 53 mAh/g in the first cycle and maintains 47 mAh/g after 200 cycles which is 147% higher than that of Na_0.44MnO₂ bulk cathode. Besides, the original orthorhombic Na_0.44MnO₂ phase of the Na_0.44MnO₂ nanorods maintains well after a high-rate cycling test at 5 C. This study demonstrates that the size of Na_0.44MnO₂ matters in the cathode performance. It indicates that reducing the size of active materials can be a promising way to explore high-capacity cathodes for aqueous Na-ion batteries.