Effects of Li₂SO₄·H₂O amounts on morphologies of hydrothermal synthesized LiMnPO₄ cathodes

Abstract

Developing an industrial LiMnPO₄ cathode, featuring a high rate capability, is still a huge challenge because of its slow Li⁺ diffusion speed as well as low electronic conductivity. Here, we expect to address these inherent obstacles by controlling its morphology in a well-dispersed way. The effects of the added amounts of Li₂SO₄·H₂O on the morphologies of the LiMnPO₄ samples prepared via the hydrothermal method were studied using structural and morphological characterizations. Based on the additional independent experiments, a two-step reaction was tentatively proposed to understand the formation of the LiMnPO₄ samples and the dual functions of Li₂SO₄·H₂O. For synthesizing LiMnPO₄ samples, a high SO₄²⁻ concentration may be helpful for a fast nucleation rate, while a high Li⁺ concentration can be beneficial for a rapid growth rate. The results showed that the LiMnPO₄ sample prepared with 40 mmol Li₂SO₄·H₂O (denoted as the S4) exhibited the best electrochemical performances in terms of the discharge capacity, the rate capability and the cycling stability among all the LiMnPO₄ samples, which was reasonably attributed to its well-dispersed characteristics.

---

Introduction

Lithium ion batteries have been successfully applied in many areas due to their high power and energy density, long cycle life, no memory effect, etc.¹ Cathode materials are a key part of lithium ion batteries. Among candidates to replace the high cost and toxic cobalt based cathodes for lithium ion batteries, LiMnPO₄ with ordered olivine structure is a promising alternative cathode material owing to its low prize, low toxicity, environmental benefits and excellent cyclability.² However, the two major obstacles of LiMnPO₄ which are a slow Li⁺ diffusion speed (∼10⁻¹⁵ cm² s⁻¹) and a low electronic conductivity (<10⁻¹⁰ S cm⁻¹) significantly influence its electrochemical properties, especially its high rate capability.³ A number of strategies have been proposed to overcome the above obstacles, such as reducing the particle sizes, coating with the electronic conductive agents, doping with the alien ions, and so on.⁴

Recently, remarkable enhancements in rate capabilities of LiMnPO₄ cathodes were successfully achieved by directly controlling their morphologies. For examples, a series of LiMnPO₄ nanoparticles with different morphologies were obtained by various precursors composites tailoring with various Li : Mn : P molar ratios in the solvothermal mediums, and the hierarchical LiMnPO₄ assembled from nanosheets delivered the reversible capacity of 125.5 mA h g⁻¹ at 1C;⁵ the controllable morphology tailoring of LiMnPO₄ nanocrystals from large spindles into small plates was realized by solvothermal processes, and the plates delivered a discharge capacity of 106.4 mA h g⁻¹ at 1C;⁶ the plate-like LiMnPO₄ nanomaterials were prepared via a solvothermal process by experimental variables, including the reaction time, the reaction temperature, the reactant mole ratio, and the prepared LMP190-3 electrode exhibited a stable structure after high rate cycles;⁷ the stamen-like LiMnPO₄ nanostructures self-assembled with nanorods were synthesized via a hydrothermal method by adopting a water-EG binary solvent in a volume ratio of 1 : 1, which exhibited better electrochemical performance than that of the previously reported flower-like LiMnPO₄ nanostructures.⁸

Although the above-mentioned LiMnPO₄ cathodes exhibited remarkable achievements in improving performances, most of their morphologies were composed of self-assembled architectures or agglomerations. Little research has focused on the effect of the dispersive state of LiMnPO₄ samples on electrochemical performances. Additionally, LiMnPO₄ samples with well-dispersed states have large surface areas, which are beneficial to enlarging contact areas among the electrode, the electrolyte and Super P. As a result, both the iR potential drop and the Li⁺ diffusion flux at surface areas substantially decrease, leading to a low electrode polarization at a high current density.^9,10 Moreover, no literature has reported the use of simply varying the Li₂SO₄·H₂O concentration for improving the dispersive state of LiMnPO₄ samples. Accordingly, the function of Li₂SO₄·H₂O amounts in improving the dispersive state of LiMnPO₄ samples is far from being clearly understood.

Here, we obtained the LiMnPO₄ sample with the well-dispersed characteristics by varying Li₂SO₄·H₂O amounts in the hydrothermal mediums. The dual functions of Li₂SO₄·H₂O and the formation mechanism of the LiMnPO₄ samples were elucidated. The relationships between dispersive states and electrochemical properties of the LiMnPO₄ cathodes were discussed.

Experimental

Materials syntheses

All the reagents were analytical grade and were used as received without further purification. The LiMnPO₄ samples were synthesized by a facile hydrothermal method in the 40 mL Teflon-lined stainless-steel autoclaves. In a typical route, 16 mmol Na₂S·9H₂O, 20 mmol MnCl₂·4H₂O, 20 mmol NH₄H₂PO₄ and different added amounts of Li₂SO₄·H₂O were mixed in turn in the autoclave with 30 mL distilled water. After 30 minutes magnetic stirring, the obtained mixtures were sealed in the autoclave, and then heated to 200 °C for 10 h. When the autoclave was naturally cooled to the room temperature, the precipitates were collected by the centrifugation, and washed with the distilled water and the ethanol for several times. The obtained sample was dried at 60 °C overnight. In order to enhance the electronic conductivity, the as-prepared LiMnPO₄ sample was mixed with the sucrose, and then carbonized at 700 °C for 2 h under the nitrogen atmosphere to obtain the LiMnPO₄/C composite.

Conveniently, the LiMnPO₄ sample derived from 10 mmol Li₂SO₄·H₂O in the hydrothermal reaction was denoted as the S1. Likewise, the S2, the S3 and the S4 stood for the samples fabricated from 20, 30 and 40 mmol Li₂SO₄·H₂O, respectively.

In order to investigate the formation mechanism of LiMnPO₄ samples and the function of Li₂SO₄·H₂O, the additional independent experiments were performed by changing the reaction time and the reagent, respectively.

Materials characterizations

XRD patterns of the samples were recorded on an X-ray powder diffractometer (Rigaku D/max-rA diffractometer, Cu Kα radiation, λ = 1.5406 Å). Scanning electron microscopy images were obtained using a field-emission scanning electron microscope (FE-SEM, FEI Quanta 200F). Transmission electron microscopy image was obtained using a JEOL JEM-2100 microscope operated at 200 kV. Specific surface areas of the samples were measured by Brunauer–Emmett–Teller method (BET, ASAP 2010).

The electrochemical properties were evaluated by cycling of 2032 coin-type cells with the LiMnPO₄ cathode as the working electrode and the pure lithium foil as the counter electrode on a Land-CT2001A battery test system (Jinnuo Wuhan Corp., China). To fabricate the working electrode, the LiMnPO₄/C composite, Super P and polyvinylidene fluoride (PVDF) were thoroughly mixed together with a weight ratio of 70 : 20 : 10. The electrolyte was a solution of 1 M LiPF₆ in ethylene carbonate and dimethyl carbonate (EC and DMC, 1 : 1 in volume). The cells were assembled in an argon-filled glove box. The cells were charged at galvanostatic mode to 4.5 V, held at 4.5 V until 0.05C, and then discharged at galvanostatic mode to 2.4 V. Electrochemical impedance spectroscopy (EIS) was recorded on a CHI660E electrochemical workstation (Chenhua Instruments Shanghai Inc., China), and the ac voltage signal of 5 mV was used in frequency range of 0.01 to 100 000 Hz. In addition, all the electrochemical measurements were carried out at the room temperature.

Results and discussion

XRD patterns of the samples prepared with different added amounts of Li₂SO₄·H₂O ranging from 10 to 40 mmol are presented in Fig. 1. The JCPDS card pattern of LiMnPO₄ is also presented for comparison (the rod patterns at the bottom of the figure). As seen from Fig. 1, the pure LiMnPO₄ phase can be only observed for the S2, the S3 and the S4 (b, c and d in Fig. 1). For the S1 sample (a in Fig. 1), the observed impurity of Mn₃(PO₄)₂ phase appears along with the partially crystallized olivine phase.

Fig. 1 XRD patterns of the samples prepared with different added amounts of Li₂SO₄·H₂O at 200 °C for 10 h. (a) The S1 (10 mmol) (•: the impurity of Mn₃(PO₄)₂); (b) the S2 (20 mmol); (c) the S3 (30 mmol); (d) the S4 (40 mmol). The lowest panel shows the standard patterns of LiMnPO₄ phase.

SEM images of the as-prepared LiMnPO₄ samples (the S1, the S2 and the S3) are shown in Fig. 2. As shown in Fig. 2a and b, the S1 displays the micron-sized and disc-shaped agglomerations consisted of the tiny rods. The rod-agglomerated discs have a diameter of ca. 10 μm, and a thickness of ∼1 μm. As the Li₂SO₄·H₂O amount is increased to 20 mmol (Fig. 2c and d), the S2 can be readily obtained, which has a similar morphology to that of the S1. When the amount of Li₂SO₄·H₂O is added up to 30 mmol (Fig. 2e and f), the S3 is easily prepared. The S3 shows the hybrid morphologies of the rod-agglomerated discs and the dispersed rods.

Fig. 2 SEM images of the LiMnPO₄ samples prepared with different added amounts of Li₂SO₄·H₂O. (a and b) The S1 (10 mmol); (c and d) the S2 (20 mmol); (e and f) the S3 (30 mmol).

While the Li₂SO₄·H₂O amount is up to 40 mmol, the rod-agglomerated discs basically disappear as shown in Fig. 3. The S4 is consisted of the uniform and dispersed rods. The high magnification TEM and SEM images clearly exhibit that the dispersed rods with around 3 μm in length and about 400 nm in width have homogenous distributed particle sizes and uniform shapes.

Fig. 3 SEM images and TEM image of the S4 prepared with 40 mmol Li₂SO₄·H₂O.

According to the above observations, it can be safely concluded that both the dispersive state and the morphology of the four samples can vary with the added amounts of Li₂SO₄·H₂O. Thus, the correlation between the dispersive states of the LiMnPO₄ samples and the added amounts of Li₂SO₄·H₂O can be easily clarified.

To investigate the function of Li₂SO₄·H₂O in hydrothermal reaction and the formation mechanism of the LiMnPO₄ samples, the time-dependent experiments were conducted. The evolution of the phase of the samples was studied by XRD patterns as shown in Fig. 4. On the basis of the XRD results, the phase evaluation of the samples can be clearly divided into the two reaction stages as shown in eqn (1) and (2).

MnCl₂·4H₂O + NH₄H₂PO₄ → NH₄MnPO₄·H₂O + 2HCl + 3H₂O (1)

2NH₄MnPO₄·H₂O + Li₂SO₄·H₂O → 2LiMnPO₄ + (NH₄)₂SO₄ + 3H₂O (2)

Fig. 4 XRD patterns of the samples prepared with 40 mmol Li₂SO₄·H₂O at 200 °C for 15 min and 30 min, respectively.

For the first reaction stage, all the samples have an approximate process under a similar synthetic condition due to their equal amounts of MnCl₂·4H₂O and NH₄H₂PO₄ in the reaction mediums. Hence, the secondary reaction stage is probably responsible for the change in dispersive states owing to their different Li₂SO₄·H₂O amounts in the reaction mediums. To explain the reason of the change, we look into the relationships of the Li₂SO₄·H₂O concentration, the oversaturation, the nucleation rate as well as the growth speed of the crystals.

Under the hydrothermal condition, Mn²⁺ cations and PO₄³⁻ anions can be locally produced from NH₄MnPO₄·H₂O, and combine with Li⁺ cations to rapidly form LiMnPO₄ crystals once these ions concentrations exceed the maximum LiMnPO₄ solubility in the solution. Generally, SO₄²⁻ anions are considered to have an effect on the crystal nucleation, which have been reported in previous works.^11–13 With SO₄²⁻ anions increasing, more Mn²⁺ cations and PO₄³⁻ anions may be released from NH₄MnPO₄·H₂O to the solution, and a circumstance with a greater oversaturation degree may be provided for the hydrothermal reaction, leading to a higher nucleation rate for the LiMnPO₄ crystals. With Li⁺ anions increasing, a larger Li⁺ concentration in the reaction medium can associate with a higher growth rate of the LiMnPO₄ crystals.

The above discussion indicates that the variation in Li₂SO₄·H₂O amounts has the dual functions in controlling the nucleation rate and the growth rate of the LiMnPO₄ crystals under hydrothermal condition. Namely, a high SO₄²⁻ concentration in the solution may be helpful for a fast nucleation rate for LiMnPO₄ crystals, while a high Li⁺ concentration in the medium can be beneficial for a rapid growth rate. To confirm the postulate, an additional independent experiment was conducted where 20 mmol (NH₄)₂SO₄ and 20 mmol Li₂SO₄·H₂O instead of 40 mmol Li₂SO₄·H₂O were used as the reagents to remain unchanged for SO₄²⁻ and to halve the Li⁺ concentration, while other conditions were kept identical with those of the synthesis of the S4. In this case, the nucleation rate of the crystals might be similar to that of the S4; however, the growth rate of the crystals might be remarkably slowed down. The as-obtained sample exhibits the morphology of the rod-agglomerated discs as shown in Fig. 5. This observation suggests that Li⁺ should be the effective ion in control of the growth rate.

Fig. 5 SEM images of the LiMnPO₄ sample prepared with 20 mmol (NH₄)₂SO₄ and 20 mmol Li₂SO₄·H₂O instead of 40 mmol Li₂SO₄·H₂O.

Obviously, the increase of the added amount of Li₂SO₄·H₂O can increase Li⁺ cations and SO₄²⁻ anions in the solution. For the S4 sample, the nucleation rate and the growth rate of the LiMnPO₄ crystals can be enhanced, which cause that the crystals can undergo a fast and independent growth in a short period of time. Conversely, crystals for the samples of the S1, the S2 and the S3 may undergo a slow and gradual growth in a long period of time. The crystals with a slow and gradual growth may attach to each other to minimize its surface area and energy, resulting in the formation of the morphology of the rod-agglomerated discs.

On the basis of the above discussion, the strategy to control the morphologies of the LiMnPO₄ samples with the different dispersive states is summarized in Fig. 6. Furthermore, the results may provide us a strategy to optimize the dispersive state of samples by manipulating the reagent amount, which is suitable for practical applications.

Fig. 6 Schematic illustration of the reaction mechanism for the variation in Li₂SO₄·H₂O amounts.

The samples of the S1, the S2, the S3 and the S4 as the cathodes were assembled into the cells which were tested to evaluate the effect of the dispersive states of the samples on their electrochemical properties.

The rate capabilities and the cycling performances of the samples are shown in Fig. 7. It can be found from Fig. 7a that the S4 exhibits better rate capability than the S1, the S2 and the S3. The BET surface areas of the S1, the S2, the S3 and the S4 can help to explain the difference in rate capabilities. The BET surface areas are 15.3, 12.4, 10.6 and 10.1 m² g⁻¹ for the S1, the S2, the S3 and the S4, respectively. The S4 exhibits higher surface areas which provide larger electrode/electrolyte contact areas as well as more diffusion channels for Li⁺, resulting in a better rate capability. Fig. 7b shows the cyclic performances of the samples at the rate of 0.1C between 2.4 and 4.5 V. After the total 50 cycles, the S1, the S2, the S3 and the S4 exhibit discharge capacities of 130, 108, 98 and 96 mA h g⁻¹, which indicate 93.5%, 92.3%, 92.1%, and 90.9% retention in relations to their initial discharge capacities, respectively. All the samples have a discharge capacity loss during the cycles. The fading of the discharge capacity can be ascribed to some origins such as the electrolyte decomposition, the manganese dissolution and the structure volume change.^14,15

Fig. 7 (a) The rate capabilities and (b) the cycling performances of the LiMnPO₄ samples prepared with the different added amounts of Li₂SO₄·H₂O.

Fig. 8 shows the charge/discharge curves of the S3 and the S4 at various C-rates. The profiles clearly exhibit the flat redox potential plateaus around 4.1 V. For the S4, the discharge capacities can reach 145, 139, 132, 124, 115 and 102 mA h g⁻¹ at 0.05, 0.1, 0.2, 0.5, 1 and 2C, respectively. For the S3, the discharge capacities are 129, 117, 111, 104, 97 and 89 mA h g⁻¹ at 0.05, 0.1, 0.2, 0.5, 1 and 2C, respectively.

Fig. 8 The charge/discharge curves of the S3 (a) and the S4 (b).

On the basis of the above tested results, it can be safely demonstrate that the S4 exhibits higher electrochemical performances than those of the S1, the S2 and the S3. Additionally, the electrochemical performances of all the samples are low, though they can be comparable to those of the recently reported LiMnPO₄ materials synthesized by other methods.^7,8,16–18

The differential capacity versus voltage (dQ/dV vs. voltage) plots of the S3 and the S4 for the 1st cycle at a rate of 0.05C in the voltage area of 2.4–4.5 V are used to further analyze their electrochemical reactions as shown in Fig. 9. The two polarization peaks concerning oxidation and reduction potentials are used to reveal the voltage value of LiMnPO₄/MnPO₄ phase transition. The lower voltage gap between the charge and discharge plateaus, the better the kinetics seen in the reaction. The different polarization peaks of the S3 and the S4 are quite obvious. For the S4, the voltage of the oxidation peak is 4.1753 V during charging (LiMnPO₄ converting to MnPO₄), while the potential of its reduction peak is 4.0517 V during discharging (MnPO₄ back to LiMnPO₄). Thereby, the voltage gap of the S4 is 123.6 mV. In comparison, the oxidation/reduction peaks of the S3 are surrounded at 4.2152/4.0041 V, resulting in 171.1 mV of the voltage gap. The voltage gap of the S4 is remarkably reduced, indicating its significant-enhanced performance, and deriving from its well-dispersed characteristics.

Fig. 9 The dQ/dV vs. voltage plots of the S3 and the S4 for the 1st cycle at a rate of 0.05C.

To deeply understand the reason for the enhanced the electrochemical performances of the S4 compared with that of the S3, we analyzed electrochemical impedance spectroscopy. Before EIS tests, the cells were cycled for two cycles at a rate of 0.1C. Fig. 10 shows the Nyquist plots of the two samples. Obviously, the diameter of the semicircle for the S4 is much smaller than that of the S3 at the high frequency region, implying the S4 has a lower charge transfer resistance. At the low frequency region, a more vertical straight line of the S4 compared to that of the S3 is an evident associated with a faster Li⁺ diffusion behaviour in the S4. The EIS results make it clear that the electron transfer and the ion diffusion of the S4 are much faster than that of the S3, leading to the elevated electrochemical activities. Furthermore, the faster electron transfer and ion diffusion of the S4 are closely associated with its well-dispersed state.

Fig. 10 The Nyquist plots of the S3 and the S4.

Conclusions

We safely demonstrated that the LiMnPO₄ samples synthesized with the different added amounts of Li₂SO₄·H₂O in the hydrothermal processes took on the different dispersive characteristics and resulted in the different electrochemical performances. Based on the additional dependent trials, a two-step reaction mechanism of the LiMnPO₄ samples and the dual functions of Li₂SO₄·H₂O in the hydrothermal synthesis were proposed. On the basis of the electrochemical measurements, the S4 sample prepared with 40 mmol Li₂SO₄·H₂O exhibited higher discharge capacity, better rate capability and more stable cycling stability than those of the samples of the S1, the S2 and the S3. Based on the analyses of the BET tests, the EIS tests and the dQ/dV vs. voltage plots, better electrochemical performances of the S4 sample were reasonably ascribed to its morphology with a well-dispersed state.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (51564024) and the Major Science & Technology Program of Jiangxi Province (20143ACB21024) and the Science & Technology Research Program of Jiangxi Provincial Education Department (GJJ150762) and the Science & Technology Supporting Program of Jiangxi Province (20133BBE50010) and the Science & Technology Supporting Program of Jian City (JA2015-10-02) and the Doctoral Starting up Foundation of Jinggangshan University (JZB15005).

Notes and references

1. M. Armand and J. M. Tarascon, Nature, 2008, 451, 652–657.
2. A. K. Padhi, K. S. Nanjundaswamy and J. B. Goodenough, J. Electrochem. Soc., 1997, 144, 1188–1194.
3. W. T. Geng and T. Ohno, J. Phys. Chem. C, 2012, 117, 276–279.
4. H. Shu, X. Wang, W. Wen, Q. Liang, X. Yang, Q. Wei, B. Hu, L. Liu, X. Liu, Y. Song, M. Zho, Y. Bai, L. Jiang, M. Chen, S. Yang, J. Tan, Y. Liao and H. Jiang, Electrochim. Acta, 2013, 89, 479–487.
5. Z. Zhang, G. Hu, Y. Cao, J. Duan, K. Du and Z. Peng, RSC Adv., 2015, 5, 81461–81467.
6. H. Guo, C. Wu, L. Liao, J. Xie, S. Zhang, P. Zhu, G. Cao and X. Zhao, Inorg. Chem., 2015, 54, 667–674.
7. K. Zhu, W. Zhang, J. Du, X. Liu, J. Tian, H. Ma, S. Liu and Z. Shan, J. Power Sources, 2015, 300, 139–146.
8. L. Bao, G. Xu, H. Zeng, L. Li, R. Zhao, G. Shen, G. Han and S. Zhou, CrystEngComm, 2016, 18, 2385–2391.
9. K. Dokko, T. Hachida and M. Watanabe, J. Electrochem. Soc., 2011, 158, A1275–A1281.
10. Z. Gao, X. Pan, H. Li, S. Xie, R. Yi and W. Jin, CrystEngComm, 2013, 15, 7808–7814.
11. L. Bao, G. Xu, J. Wang, H. Zong, L. Li, R. Zhao, S. Zhou, G. Shen and G. Han, CrystEngComm, 2015, 17, 6399–6405.
12. B. P. Boudreau and D. E. Canfield, Geochim. Cosmochim. Acta, 1993, 57, 317–334.
13. M. Wilf and J. Ricklis, Desalination, 1983, 47, 209–219.
14. L. Wang, P. Zuo, G. Yin, Y. Ma, X. Cheng, C. Du and Y. Gao, J. Mater. Chem. A, 2015, 3, 1569–1579.
15. M. Pivko, M. Bele, E. Tchernychova, N. Z. Logar, R. Dominko and M. Gaberscek, Chem. Mater., 2012, 24, 1041–1047.
16. Y. Gu, H. Wang, Y. Zhu, L. Wang, Y. Qian and Y. Chu, Solid State Ionics, 2015, 274, 106–110.
17. J. Su, Z.-Z. Liu, Y.-F. Long, H. Yao, X.-Y. Lv and Y.-X. Wen, Electrochim. Acta, 2015, 173, 559–565.
18. K. Wu, G. Hu, Z. Peng, Z. Zhang, Y. Cao and K. Du, RSC Adv., 2015, 5, 95020–95027.

---