Introduction

The increasing deployment of renewable energy sources such as solar and wind power requires a commensurate increase in energy storage capacity to integrate them into the grid. Batteries are good means of storing the electricity in the form of chemical energy. Owning to good safety, high ionic conductivity and low cost, aqueous rechargeable batteries are potentially advantageous over their organic counterparts for large-scale energy storage1.

Various aqueous batteries such as alkaline Zn-MnO2, lead-acid, nickel-metal hydride (Ni-MH) and nickel-metal (e.g., cadmium, iron, zinc and cobalt) are commercially available or under extensive research1,2,3,4,5,6. Alkaline Zn/MnO2 is a primary battery; lead-acid and Ni-Cd batteries suffer from serious environmental problems because of the highly poisoning metals of lead and cadmium; Ni-MH is expensive due to the utilization of rare earth elements; Ni-Co, Ni-Fe and Ni-Zn batteries have poor cycling stability. These drawbacks in above systems hinder their utilization for large-scale energy storage. Recently, a variety of aqueous metal-ion batteries (e.g., Li+, Na+, K+ and Zn2+ etc.) on a basis of metal-ion intercalation chemistry have garnered great interests7,8,9,10,11,12,13,14,15,16,17,18. Aqueous “rocking-chair” lithium-ion batteries using the cathode materials adopted from organic electrolyte cells (e.g., LiMn2O4/VO2, LiMn2O4/LiV3O8, LiMn2O4/TiP2O7, LiMn2O4/LiTi2(PO4)3 and LiFePO4/LiTi2(PO4)3 etc.) have been developed, but in general they exhibited limited cycle life7,8,9,10. A study by Xia's group showed that the cycling stability could be raised to a commercially level by eliminating O2, controlling the pH of electrolytes and cabon-coating the electrode materials10. Cui et al. reported a novel potassium-ion battery using carbon/polypyrrole hybrid (AC-PPy) anode and copper hexacyanoferrate (CuHCF) cathode, which offered long cycle life and good rate capability12. A unique zinc-ion battery based on Zn anode, α-MnO2 cathode and mild electrolytes was explored by Kang et al. It exhibited high capacity and fast charge/discharge capability13.

In contrast to lithium, potassium and zinc, sodium is most abundant and economical. Therefore, sodium-based energy storage system is considered as a promising technology for large-scale energy storage. However, only a couple of materials have been reported for the deintercalation/intercalation of sodium ion in aqueous media14,15,16,17,18,19,20. The reason is that several side reactions are involved in aqueous media, such as electrode materials reacting with water or O2, proton co-intercalation into the electrode materials parallel to the intercalation of sodium ion, H2/O2 evolution reactions and the dissolution of electrode materials in water. Whitacre et al. found that sodium ion could be reversibly inserted into Na0.44MnO2, owning to its unique tunnel structure14. But it only delivered a specific capacity of 45 mAh g−1 at 0.125 C rate. Yamaki et al. investigated the electrochemical behavior of NaTi2(PO4)3 in aqueous sodium electrolytes15. Its redox potential is −0.6 V vs. standard hydrogen electrode, indicating that NaTi2(PO4)3 can be utilized as an anode for aqueous sodium battery. The aqueous “rocking-chair” sodium-ion battery consisted of NaTi2(PO4)3 and Na0.44MnO2 was reported by both Whitacre's and Chiang's groups16,17. Cui et al. found that materials with the Prussian Blue crystal structure (copper and nickel hexacyanoferrate) contained large interstitial sites, which allowed for the insertion and extraction of Na+ and/or K+ (ref. 18,19). Their capacities were in the range of 50–60 mAh g−1. In a word, much more work on the exploration of new sodium-intercalated materials is needed to make aqueous sodium batteries viable in the field of large-scale energy storage.

Since sodium-intercalated materials suitable for aqueous media are limited, an innovative concept of Li+/Na+ mixed-ion electrolytes is employed to construct rechargeable batteries, as shown in Fig. 1. In such batteries, one side involves the immigration of Li+ between electrolytes and electrode and the other one refers to the exchange of Na+ between electrode and electrolytes. During charging and discharging, the total concentration of Li+ and Na+ is fixed to ensure the charge neutrality of the electrolytes, but the Li+/Na+ ratio is changed. They are unlike traditional “rocking-chair” lithium-ion battery on a basis of the immigration of Li+ between cathode and anode. Herein, two systems based on Li2SO4/Na2SO4 mixed electrolytes (LiMn2O4/Na0.22MnO2 and Na0.44MnO2/TiP2O7), which to our best knowledge have never been reported before, are demonstrated. The capacity, operating voltage and stability of such batteries are dependent on the electrolytes. A LiMn2O4/Na0.22MnO2 system to separate Li+ and Na+ based on the unique mechanism of mixed-ion battery is also validated.

Figure 1
figure 1

Schematics of Li+/Na+ mixed-ion battery.

Lithium-intercalation compounds and sodium-intercalation compounds are used for anode and cathode, respectively. During charging (or discharging), the storage (or release) of Li+ takes place at anode and the release (or storage) of Na+ occurs at cathode.

Full size image

Results

Characterizations of Na0.44MnO2, LiMn2O4 and TiP2O7

The detailed synthetic procedures of above three materials are described in Methods. Manganese oxides form a rich structural family enlisting materials with either 1D-, 2D-, 3D-type tunnel structure. Among them, Na0.44MnO2 belongs to the space group Pbam with orthorhombic structure (see Supplementary Fig. S1). The manganese ions are located in two different environments: all Mn4+ ions and half of the Mn3+ ions are in octahedral sites, while the other Mn3+ gathered in a square-pyramidal environment. The latter forms edge-linked chains linked to two double and one triple octahedral chain by the vertices, leading to the formation of two types of tunnels (S-shaped tunnel and pentagon tunnel). The sites in S-shaped tunnel are approximately half-filled and the sites in pentagon tunnel are almost fully occupied. Ab initio studies shows that half of sodium ions can be reversibly inserted into Na0.44MnO2 in the range of −0.10–0.85 V vs. saturated calomel electrode (SCE), resulting in a theoretical capacity of 60 mAh g−1 (ref. 19). X-ray diffraction (XRD) pattern of the Na0.44MnO2 powders in Supplementary Fig. S1 shows five strong peaks at 14.2°, 16.8°, 19.7°, 34.2° and 37.6°, responding to the reflections of (130), (140), (150), (280) and (151) plane of orthorhombic structure, respectively21,22. In addition, no peaks related to Mn2O3 impurity are detectable14,22. The particles are also well crystallized and have grown anisotropically, leading to nanorod shape. The nanorods show an excellent atomic ordering and always exhibit (110) crystal plane with interplanar distance of 0.87 nm (Supplementary Fig. S1).

Spinel LiMn2O4 has 3D tunnel structure with space group Fd3m. The Mn ions have an octahedral coordination to the oxygen atoms and the MnO6 octahedra share edges in a 3D host for the Li guest ions (Supplementary Fig. S2). Proton can also diffuse into the tunnel due to its small ion radius. XRD pattern of synthesized LiMn2O4 powders confirms a pure spinel phase. Scanning electron microscopy (SEM) analysis shows that the powders are composed of microspheres (Supplementary Fig. S2).

Pyrophosphate TiP2O7 has a cubic 3 × 3 × 3 superstructure with space group Pa3. It consists of TiO6 octahedral and PO4 tetrahedral sharing corners in a 3D network (Supplementary Fig. S3). Such framework can accommodate Li+ insertion and extraction, leading to a theoretical capacity of 121 mAh g−1. Carbon-coated TiP2O7 with a pure cubic phase is synthesized through solid state reaction method (see method and Supplementary Fig. S3). It exhibits a highly dispersed particulate morphology. High resolution TEM image of TiP2O7 particle displays that amorphous carbon with average thickness of 7 nm is uniformly coated on the surface of the particle (Supplementary Fig. S3). The amount of carbon in the materials is about 16%, which is estimated by thermogravimetric analysis.

Selectivity of Na0.44MnO2 for Na+ against Li+

The electrochemical properties of prepared Na0.44MnO2 in different electrolytes are investigated by cyclic voltammetry (CV) and galvanostatic techniques. Fig. 2a displays three reversible redox couples at ca. 0.05, 0.28 and 0.46 V vs. SCE in aqueous Na2SO4 electrolytes, corresponding to the insertion and extraction of Na+ (ref. 13 and 20). In Li2SO4/Na2SO4 mixed electrolytes, these redox couples still exist, but their potentials are different from the ones in Na2SO4 electrolytes. In Li2SO4 electrolytes, three irreversible oxidation peaks at 0.08 V, 0.31 V and 0.49 V, attributed to the removal of Na+ from Na0.44MnO2, are observed in the first cycle. Very weak peaks related to Li+ insertion and extraction are seen in the subsequent cycles. It concludes that only a very small amount of Li+ can be reversibly inserted/de-inserted into/from Na0.44MnO2, regardless of its smaller ionic radius (76 pm) compared to Na+ (102 pm). Moreover, Na+ can be selectively extracted and inserted from/into Na0.44MnO2 in the presence of Li+ and Na0.44MnO2 exhibits high selectivity of Na+ towards Li+. Between −0.1 and 0.65 V vs. SCE, the discharge capacities of Na0.44MnO2 in 1 M Na2SO4, 1 M Na2SO4 + 0.125 M Li2SO4, 1 M Na2SO4 + 0.25 M Li2SO4 and 1 M Na2SO4 + 0.5 M Li2SO4 electrolytes are 62.1, 60.7, 60.2 and 57.9 mAh g−1 at 0.25 C rate (1 C = 60 mAh g−1), respectively (Supplementary Fig. S4). These values are much larger than 45 mAh g−1 obtained by Whitacre's group and agree well with the theoretical capacity of 60 mAh g−1, which indicate that about half of sodium ions in Na0.44MnO2 can be reversibly intercalated into tunnels of Na0.44MnO2 in mixed electrolytes. The unique nanorod shape and good crystallinity of Na0.44MnO2 may account for its high capacity and good reversibility. It is also found that the capacity slightly decreases with an increase of the Li+/Na+ ratio.

Figure 2
figure 2

Cyclic voltammograms in different electrolytes.

(a), for Na0.44MnO2 electrode at a scan rate of 0.5 mVs−1. (b), for LiMn2O4 electrode at a scan rate of 0.1 mVs−1. (c), for carbon-coated TiP2O7 electrode at a scan rate of 0.3 mVs−1. Black line: in 1 M Na2SO4; red line: in 1 M Na2SO4 + 0.125 M Li2SO4; green line: in 1 M Na2SO4 + 0.25 M Li2SO4; blue line: in 1 M Na2SO4 + 0.5 M Li2SO4; cyan line: in 1 M Li2SO4.

Full size image

Selectivity of LiMn2O4 and TiP2O7 for Li+ against Na+

Fig. 2b describes the CV curves of LiMn2O4 in different electrolytes. In Li2SO4 electrolytes, two pairs of redox peaks locate at 0.81 V and 0.93 V (oxidation potential vs. SCE), indicating two-phase lithium extraction/insertion behavior inherent to spinel LiMn2O4. In mixed electrolytes, the redox peaks still exist, but their potentials and currents are relatively lower than the ones in Li2SO4 electrolytes. However, only two irreversible oxidation peaks are seen in Na2SO4 electrolytes, indicating that Na+ can hardly diffuse into bulk LiMn2O4. Meanwhile, the peak potential decreases by decreasing the Li+/Na+ ratio, as expected from kinetics consideration. The kinetics of the insertion and extraction are strongly influenced by lithium concentration. During Li+ insertion reaction, a decrease in lithium concentration results in a decrease of the limiting diffusion current and therefore an increase of the concentration overvoltage, which results in a negative shift of the peak potential. In terms of Li+ extraction reaction, a decrease in lithium concentration leads to an easier removal of Li+ from bulk LiMn2O4, which also shifts the peak potential towards smaller value. It is also observed that both the potential of charge/discharge plateau and the capacity decrease with decreasing the Li+/Na+ ratio (see Supplementary Fig. S4).

Similar lithium ion sieve property is observed for TiP2O7. Fig. 2c shows the CV curves of TiP2O7 in different electrolytes. In Li2SO4 electrolytes, a pair of Li+ intercalation and extraction peaks appear at −0.72V and −0.60 V vs. SCE23. In mixed electrolytes, the peak potential decreases with a decrease of the Li+/Na+ ratio. Similar trend is seen for the LiMn2O4 material. No redox peaks are observed in Na2SO4 electrolytes, implying that TiP2O7 can only store and release Li+. The charging/discharging capacities between 0 V and −0.9 V vs. SCE in mixed electrolytes are close to the one in Li2SO4 electrolytes, indicating that Na+ has no obvious effect on the capacity of TiP2O7 (Supplementary Fig. S4). The coulombic efficiency in 1 M Li2SO4, 1 M Na2SO4 + 0.5 M Li2SO4, 1 M Na2SO4 + 0.25 M Li2SO4 and 1 M Na2SO4 + 0.125 M Li2SO4 electrolytes are 86.7%, 81.3%, 91.3% and 82.5%, respectively. All of them are much lower than the theoretical one of 100%. It may be mainly caused by the decomposition of water and trace of oxygen in the electrolytes. According to Nernst equation, the theoretical H2 evolution potential at pH = 7 locates at −0.66 V vs. SCE, the same as the redox potential of Ti3+/Ti4+ in TiP2O7 (−0.66 V vs. SCE). It means that the evolution of H2 may be parallel to the insertion of Li+ into TiP2O7 during discharging. Negative electrode candidates for aqueous lithium-ion battery are found to be sensitive to oxygen and water on the discharged-state10. The potential of charge/discharge plateau decreases with decreasing the Li+/Na+ ratio, which is consistent with the phenomena observed from CV experiments.

LiMn2O4/Na0.22MnO2 and Na0.44MnO2/TiP2O7 batteries

Based on the above mixed-ion chemistry of Na0.44MnO2, LiMn2O4 and TiP2O7, we assembled two practical prototypes of mixed-ion batteries (LiMn2O4/Na0.22MnO2 and Na0.44MnO2/TiP2O7). Fig. 3a illustrates the principle of LiMn2O4/Na0.22MnO2 system (Na0.22MnO2 is obtained through electrochemical oxidation of Na0.44MnO2). One side mainly involves the immigration of Na+ between electrolytes and Na0.22MnO2 and the other side is related to the insertion/extraction of Li+ into/from LiMn2O4. Typical charge-discharge curves of LiMn2O4/Na0.22MnO2 battery in two different electrolytes are displayed in Fig. 3b and 3c. Such a battery can deliver a capacity of 29 mAh g−1 (0.25 C rate) based on the total mass of active materials between 0.2 V and 1.05 V. The coulombic efficiency of the battery after the first cycle is above 90%, indicating its good reversibility. Fig. 3d shows the plot of discharging capacity retention percentage vs. the cycle number. In both electrolytes, the fading rate of the capacity in the initial 10 cycles is much higher than the one after 10 cycles. After 45 cycles, 83% of the original capacity is remained in 1 M Na2SO4 + 0.125 M Li2SO4. But in 1 M Na2SO4 + 0.5 M Li2SO4, the capacity remains only 61% of the original one. Therefore, its cycling stability can be enhanced by decreasing the Li+/Na+ ratio. The rate performance of LiMn2O4/Na0.22MnO2 battery in 1 M Na2SO4 + 0.125 M Li2SO4 is also tested. The discharge capacities of 27.9, 25.5, 19.2, 11.7 and 7.7 mAh g−1 are obtained at 0.5, 1, 2, 4 and 8 C rate, respectively (see Supplementary Fig. S5).

Figure 3
figure 3

The LiMn2O4/Na0.22MnO2 mixed-ion battery.

(a), A schematic of the battery. (b), Charging-discharging curves at a rate of 0.25 C in 1 M Na2SO4 + 0.5 M Li2SO4. (c), Charging-discharging curves at a rate of 0.25 C in 1 M Na2SO4 + 0. 125 M Li2SO4. (d), Cycle life tests of the battery at a rate of 0.25 C in mixed electrolytes. (1 C = 60 mAh g−1).

Full size image

The principle of the mixed-ion battery comprised by a TiP2O7 anode and a Na0.44MnO2 cathode is described in Fig. 4a. As shown, the Li+ insertion and removal take place at anode and the release and storage of Na+ occur at cathode. Fig. 4b and 4c depict charge-discharge profiles of Na0.44MnO2/TiP2O7 battery in two different electrolytes. The operating voltage of the battery ranges from 0.65 V to 1.3 V. And the voltage of charging/discharging plateau in 1 M Na2SO4 + 0.125 M Li2SO4 is 30 mV–40 mV higher than the one in 1 M Na2SO4 + 0.25 M Li2SO4. Between 0.65 V and 1.3 V, the battery can give reversible capacities of 23.5 mAh g−1 in 1 M Na2SO4 + 0.25 M Li2SO4 and 25.0 mAh g−1 in 1 M Na2SO4 + 0.125 M Li2SO4 at 0.4 C rate. Meanwhile, it exhibits good coulombic efficiency after the first cycle (> 96%) in both electrolytes, which is indicative of good reversibility of the battery. Its cycling stability is studied as well. The fading rate of discharging capacity is much higher in the first 10 cycles than the one after 10 cycles (see Fig. 4d). Similar phenomena are also observed for LiMn2O4/Na0.22MnO2 system. In 1 M Na2SO4 + 0.125 M Li2SO4 electrolytes, 83% and 62% of the original discharging capacity are retained after 10 and 50 cycles, respectively. In the case of 1 M Na2SO4 + 0.25 M Li2SO4 electrolytes, only 76% and 58% of the original discharging capacity are remained. As a result, raising the Na+/Li+ ratio can improve the cycling stability of Na0.44MnO2/TiP2O7 battery. A discharge rate capability study of Na0.44MnO2/TiP2O7 battery in 1 M Na2SO4 + 0.125 M Li2SO4 is shown in Supplementary Fig. S5. The discharge capacities of 23.3, 20.6, 17.7, 14.5 and 11.1 mAh g−1 are found at 0.5, 1, 2, 4 and 8 C rate, respectively.

Figure 4
figure 4

The Na0.44MnO2/TiP2O7 mixed-ion battery.

(a), A schematic of the battery. (b), Charging-discharging curves at a rate of 0.4 C in 1 M Na2SO4 + 0.25 M Li2SO4. (c), Charging-discharging curves at a rate of 0.4 C in 1 M Na2SO4 + 0. 125 M Li2SO4. (d), Cycle life tests of the battery at a rate of 0.4 C in mixed electrolytes. (1 C = 60 mAh g−1).

Full size image

Elucidation on the mechanism of Li+/Na+ mixed-ion batteries

The feature of mixed-ion battery is the changes of Li+/Na+ ratio during charging and discharging. To experimentally demonstrate such changes, a micro electrochemical cell consisted of a Na0.22MnO2 anode and a LiMn2O4 cathode is made. We choose several electrolytes to study (see Table 1). Firstly, Na2SO4 and Li2SO4 solution are selected. Before charging, pristine Na2SO4 solution with the Li+/Na+ ratio of 1/83 is injected into the cell. After charging, the Li+/Na+ ratio goes up to 1/10. This is due to the release of Li+ from a LiMn2O4 cathode and selective capture of Na+ from electrolytes by a Na0.22MnO2 anode. In contrast, Li2SO4 solution replaces Na2SO4 solution for the discharging process. The Li+/Na+ ratio in electrolytes drops from 144/1 to 13/1 after discharging. The reason is that Li+ in electrolytes is intercalated into cathode and Na+ is diffused from anode to electrolytes during discharging. Similar changes are also observed for mixed-ion solutions. The Li+/Na+ ratio increases after charging and decreases after discharging. Therefore, LiMn2O4/Na0.22MnO2 system can enrich Li+ in electrolytes by charging and enrich Na+ by discharging.

Table 1 Atomic ratio of Li+/Na+ in the electrolytes during charging and discharging
Full size table

Discussion

Li+/Na+ mixed-ion batteries are based on the ion-selectivity of electrode materials. As demonstrated by the electrochemical data, Na0.44MnO2, LiMn2O4 and TiP2O7 exhibit specific ion-selectivity. Their ion-selectivity has also been confirmed by the X-ray photoelectron spectroscopy experiments (see Supplementary Table S1). In Li+/Na+ mixed-ion solutions, Na0.44MnO2 exhibits specific ion-selectivity towards Na+. The reason may be the dimensions of two tunnels in Na0.44MnO2 (S-shaped tunnel and pentagon tunnel) are too large for Li+ to be anchored. The mismatch leads to the free movement of Li+ in the tunnels. Both LiMn2O4 and TiP2O7 behave as ion sieves which exhibit very high ion-selectivity towards Li+. Their specific selectivity can be easily explained by the steric effect: the tunnels are too small for metal ions with ionic radii larger than Li+ to enter.

The theoretical capacities of Na0.44MnO2, LiMn2O4 and TiP2O7 are 60 mAh g−1, 148 mAh g−1 and 121 mAh g−1, respectively. The average working voltages of LiMn2O4/Na0.22MnO2 and Na0.44MnO2/TiP2O7 batteries are 0.6 V and 1.0 V. If the theoretical capacities of Na0.44MnO2, LiMn2O4 and TiP2O7 are fully used, batteries with specific energy of ca. 21 Wh kg−1 or 40 Wh kg−1 based on the total weight of active electrode materials could be desired. Thus the Na0.44MnO2/TiP2O7 system can be superior to CuHCF/AC-PPy and Na0.44MnO2/active-carbon systems for large-scale energy storage12,14. Actually, two batteries with specific energy of 17 Wh kg−1 and 25 Wh kg−1 are obtained in our work. How to achieve the theoretical capacities of these compounds and long cycle life of the systems still remains a challenge.

More and more lithium is needed to meet for the evermore demanding market of electric cars. As of 2012, most of the lithium (roughly 83%) is obtained from brine lakes and salt pans. Existing methods for purification of lithium (“line soda evaporation process”, amorphous Al2O3 process and ion exchange technique) are wet chemistry techniques24,25. As such methods require a large quantity of chemicals (e.g., Na2CO3, AlCl3, manganese oxides and HCl etc), they are expensive, slow and inefficient. Therefore, to explore new efficient methods for the recovery of lithium from brine is necessary. Based on the demonstrated mechanism of Li+/Na+ mixed-ion batteries, a new electrochemical method is proposed. Fig. 5 schematically shows the working principle of mixed-ion battery system for Li+/Na+ separation. The battery is based on the Li+-releasing cathode and Na+-capturing anode. During the first step (charging), the battery releases Li+ and captures Na+ to enrich Li+ in mixed-ion solution I. After charging, the solution I is replaced by another mixed-ion solution II. Then the battery is discharged to enrich Na+, as Li+ are captured by cathode and Na+ are released from anode. In the last step, mixed-ion solution II is exchanged with mixed-ion solution I for the next cycle. Throughout repetitive cycles, Li+-enriched and Na+-enriched solutions are finally obtained respectively. Compared with other purification methods as stated above, this method presents the following two evident advantages. Firstly, it is dual-functional. Li+/Na+ mixed-ion battery can be used for both purification of lithium and energy storage. Secondly, this electrochemical method is green and energy efficient. Unlike wet chemical methods, it does not require any chemicals and generates zero waste for disposal. To verify the feasibility of the method, the LiMn2O4/Na0.22MnO2 system is selected. Two mixed electrolytes are used as shown in Table 1. Firstly, Li2SO4+Na2SO4 solution with the Li+/Na+ ratio of 100/42 (solution I) is added into the cell. After charging, the Li+/Na+ ratio goes up to 100/35. Then solution I is removed and substituted by another solution (solution II) for the next step. After discharging, the Li+/Na+ ratio in solution II decreases from 47/100 to 41/100. Subsequently, the LiMn2O4/Na0.22MnO2 battery is subjected to the second cycle. After two cycles, the Li+/Na+ ratio in solution I increases to 100/29 and the Li+/Na+ ratio in solution II drops to 36/100. These results demonstrates that Li+/Na+ mixed-ion battery can be applied for separating Li+ and Na+.

Figure 5
figure 5

Schematic representation of the working principle on the separation of Li+ and Na+ by the mixed-ion battery.

Full size image

In summary, a unique “mixed-ion electrolytes” strategy is successfully used to establish Li+/Na+ mixed-ion batteries with the cathode/anode systems such as LiMn2O4/Na0.22MnO2 and Na0.44MnO2/TiP2O7. The resultant batteries are built on the ion-selective properties of Na0.44MnO2, LiMn2O4 and TiP2O7. The capacities, operating voltage and stability of both batteries are dependent on the Na+/Li+ ratio. Raising the Na+/Li+ ratio is beneficial to their stabilities. As well, LiMn2O4/Na0.22MnO2 system can be used to separate Li+ and Na+. In a word, Li+/Na+ mixed-ion batteries not only pave a new route to the energy storage system, but also give a prospective technique for Li+/Na+ separation.

Methods

Material syntheses

All chemicals were obtained from Sinopharm Chemical Reagent Co., Ltd (Shanghai, China). Na0.44MnO2 and LiMn2O4 were prepared by solid-state reaction method. To make Na0.44MnO2 nanorods, Na2CO3 was ball milled with MnCO3 in a 0.25∶1 ratio at 350 rpm for 8 h. The precursor mix was successively heated for 3 h at 300°C and for 9 h at 800°C. For spherical LiMn2O4, starting materials MnCO3 and Li2CO3 were thoroughly mixed in a molar ratio of 4∶1. The mixture was hand-grounded, pelletized and heated at 800°C in air for 12 h. The as-prepared TiP2O7 samples were also synthesized by solid-state reaction routine. The high purity TiO2 powder (up to 99.9%) and NH4H2PO4 were added into ethanol to form paste. This paste was ball-milled in a planetary ball mill at 400 rpm for 8 h. After ball milling, the mixture was dried at 80°C to evaporate ethanol. Then it was successively heated for 5 h at 300°C and for 12 h at 900°C under air with intermediate grindings. The obtained white TiP2O7 solids, referred as as-prepared TiP2O7, were homogenized in a mortar for further use. For carbon-coated TiP2O7, 2.5 g of as-prepared TiP2O7, 2 g of glucose and appropriate amount of ethanol were mixed and then ball-milled for 4 h. The solids were obtained by evaporating the ethanol at 80°C. After calcined at 700°C for 5 h in Ar, the solids became black, implying that TiP2O7 were coated by carbon.

Characterizations

Powder X-ray diffraction patterns were collected using an AXS D8 Advance diffractometer (Cu Kα radiation; receiving slit, 0.2 mm; scintillation counter, 40 mA; 40 kV) from Bruker Inc. The morphology and structure of samples were analyzed by a Hitachi S-4800 field emission scanning-electron microscope and an FEI Tecnai G2 F20 transmission electron microscope at an accelerating voltage of 200 kV. Thermal gravimetric analysis was performed on a Pyris Diamond thermogravimetric/differential thermal analyzer by Perkin-Elmer. Elemental analyses were carried out on inductively coupled plasma optical emission spectrometer (Perkin Elmer Optima 2100 DV).

Electrochemical measurements

Electrochemical measurements were carried on Solartron 1470E multi-channel potentiostats using either two-electrode or three-electrode cell. For three-electrode setup, saturated calomel electrode (SCE) and Pt gauze were employed as reference and counter electrodes, respectively. Both cathode and anode were prepared by casting slurries of active materials (75 wt%), Super P (15 wt%) and polyvinylidene fluoride (10 wt%) in n-methyl-2-pyrrolidinone on steel iron grid and air drying at 80°C for 12 h. Disks of diameter (1.3 cm) were cut for electrochemical tests. Mass loadings for the electrodes were determined by comparing the mass of the electrode with that of the original blank one. Na0.22MnO2 electrodes were fabricated through electrochemical oxidation of Na0.44MnO2 electrodes (constant current: 30 mAg−1, cut-off voltage: 0.65 V vs. SCE). Deaerated solutions obtained through Ar gas bubbling for 60 min were used for TiP2O7 electrodes. The mass ratios of anode to cathode materials are designed as 1.80 for LiMn2O4/Na0.22MnO2 system and 0.85 for Na0.44MnO2/TiP2O7 system, according to the capacities of Na0.44MnO2, LiMn2O4 and carbon-coated TiP2O7. A micro electrochemical cell consisted of a Na0.22MnO2 anode and a LiMn2O4 cathode was fabricated to test the change of Li+/Na+ ratio during charging and discharging. The distance between two electrodes was fixed at 2.2 mm. The volume of electrolytes was 1.0 mL and the total concentration of Li+ and Na+ was about 0.35 M. Inductively coupled plasma optical emission spectroscopy (ICP-OES) was used to determine the ratio of Li+/Na+.