The Prepared and Electrochemical Property of Mg Doped LiMnPO₄ Nanoplates as Cathode Materials for Lithium-Ion Batteries

The LiMnPO₄ nanoplates were synthesized by a solvothermal method as reported elsewhere.¹⁰ With Magnesium acetate tetrahydrate (C₄H₆MgO₄·4H₂O, Alfa Aesar) as the dopant precursor, a sample of LiMn_0.95Mg_0.05PO₄ was prepared with the same procedure.

The chemical extraction of lithium in non-aqueous medium was carried out at room temperature by stirring the LiMn_0.95Mg_0.05PO₄ sample in an acetonitrile solution of NO₂BF₄ for 24 h in an argon-filled glove box. The products formed after the reaction were washed three times with acetonitrile under argon atmosphere to remove LiBF₄ and dried under vacuum at ambient temperature.

Powder X-ray diffraction (XRD) patterns of the products were recorded at room temperature with a Philips X-ray diffractometer equipped with Cu K_α radiation in a step-scanning mode over the range 10–90° with a step width of 0.02° and a step time of 5 s.

Scanning electron microscopy (SEM) was performed with a JEOL JSM-5610 and FEI Strata DB235 SEM/FIB microscope. The high-resolution transmission electron microscopy (HR-TEM) images and selected-area electron diffraction (SAED) patterns were carried out on a transmission electron microscope (JEOL 2010F).

The Electrochemical performances of the samples were evaluated with a standard CR2032 coin cell. The electrodes were prepared by the same procedure used previously for LiMnPO₄.¹⁰ The electrochemical performance of the samples was evaluated with an Arbin Instruments testing system (Arbin BT-2000) at room temperature and the electrochemical capacity of samples was evaluated based only on the active material weight. Electrochemical impedance spectroscopy (EIS) was carried out in a frequency range from 0.1 Hz to 100 kHz with an Ac signal of 5 mV. All tests were performed at room temperature.

XRD patterns for the LiMnPO₄ and LiMn_0.95Mg_0.05PO₄ samples are shown in Fig. 1. For comparison, the data of the position and the relative intensity of the peaks of a crystalline LiMnPO₄ sample (JCPDS card No. 97-009-7763) are shown in the bottom of the figure. From Fig. 1 we can see that the Mg-doped sample gave a very similar diffraction profile to that of LiMnPO₄ except for the shift of peaks to higher diffraction angles as illustrated by the (311) peak, inset of Fig. 1. All diffraction peaks of the two patterns could be fully indexed into an orthorhombic structure, space group Pnma, with the program Dicvol, and no impurities were detected.

Zoom In Zoom Out Reset image size

In order to further explore the structural change affected by the Mg doping, a Rietveld refinement was conducted for the two samples; the refinement results, shown in Fig. 2, led to a good agreement between the experimental and the calculated XRD pattern. The reasonably small Rwp factor (11.1% for LiMnPO₄ and 10.3% for LiMn_0.95Mg_0.05PO₄) suggests that triple-phase compounds were obtained; no impurity phases were detected under the resolution of our X-ray diffractometer. Because of the similar ionic radius of Mg²⁺ and Li⁺ in octahedral coordination, some Mg²⁺ could be expected to occupy the Li⁺ site. In order to determine whether there are Mg²⁺ occupying Li⁺ sites in our samples, refinements were also performed where we substituted some Mg atoms on the lithium sites and refined the isotropic temperature parameters and/or occupation numbers for Li and Mg atoms. Either negative values of thermal parameters of Li and Mg or higher R_wp (>15%) were obtained, which implies that there is no Mg occupying lithium sites in our samples. The structural parameters of LiMn_0.95Mg_0.05PO₄ obtained from the refinement are listed in Table I. In the orthorhombic Pnma structure, the dopant Mg occupies the 4c Mn site.

Zoom In Zoom Out Reset image size

The lattice parameters calculated by refining the XRD patterns are shown in Table II. Substitution of the smaller Mg²⁺ ion clearly results in a decrease of the cell volume from 304.51 ų for LiMnPO₄ to 302.94 ų for LiMn_0.95Mg_0.05PO₄. These observations indicate that the Mg substitution was successfully achieved to form a LiMn_0.95Mg_0.05PO₄ solid solution.

Fig. 3 presents the SEM images of LiMnPO₄ and LiMn_0.95Mg_0.05PO₄ samples. Both samples form nanoplates like bamboo leaves with a width of around 500 nm and length of around 2 um; no significant difference is observed in the morphologies of the two samples, but the LiMn_0.95Mg_0.05PO₄ samples are more compact. The EDS mappings (Fig. 4) indicate that like the elements Mn, P and O, the substituted Mg have a fairly uniform distribution in LiMn_0.95Mg_0.05PO₄.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Nanoplates were analyzed with TEM. Fig. 5 shows the HR-TEM images and SAED pattern of the LiMnPO₄ and LiMn_0.95Mg_0.05PO₄ nanoplates, respectively. The nearly identical interplanar spacing of 3.047 Å, 3.740 Å in Fig. 5a and 3.044 Å, 3.738 Å in Fig. 5c can be identified as the (020) and (011) lattice planes, respectively, indicating that the LiMnPO₄ and LiMn_0.95Mg_0.05PO₄ nanoplates have the same crystallographic facet, i.e. the (100) plane. The corresponding selected area electron diffraction patterns (Fig. 5b and 5d) also indicate the (100) plane in the platelets and the single-crystal nature of the nanoplates. However, the LiMn_0.95Mg_0.05PO₄ nanoplates present a more perfect crystal.

Zoom In Zoom Out Reset image size

Fig. 6 presents typical voltage profiles of the LiMn_0.95Mg_0.05PO₄ nanoplates at various charge/discharge rates between C/10 and 2C in the voltage range between 2.5 and 4.5 V. The voltage profiles clearly show a flat redox potential around 4.1 V vs. Li/Li⁺, indicating that the charge/discharge reaction proceeds via a first-order phase transition. As reported by Chio et al.,⁸ Mg doping improves the specific capacity; but with the b axis in the plane of the platelets, the capacity is smaller than can be achieved with LiFePO₄.

Zoom In Zoom Out Reset image size

In order to understand the structural change during extraction/ reinsertion of lithium, a series of Li_xMn_0.95Mg_0.05PO₄ compositions (0 ≤ x ≤ 1), prepared by chemical delithiation of LiMn_0.95Mg_0.05PO₄, were obtained. The x can be controlled by the initial concentration of NO₂BF₄ in the solution. The corresponding XRD patterns are shown in Figure 7. For comparison, the same figure shows the position and the relative intensity of the peaks of a crystalline LiMnPO₄ sample (JCPDS card no. 33-0804) and MnPO₄ sample (JCPDS card no. 77-0180) at the bottom and the top, respectively. According to the Laffont et al.' experimental method,⁹ the values of x were deduced from calibration curves of the I_LiMnPO4:(I_LiMnPO4 + I_MnPO4) ratios of the intensities of both (200) and (210) reflections established from XRD patterns of xLiMnPO₄ + (1−x)MnPO₄ mixtures (prepared by simply mixing LiMnPO₄ and MnPO₄ powders in a mortar). Integrations and decompositions of the diffraction peaks were made with the FullProf Suite. The extraction/reinsertion of lithium gives an obvious two phase structure. For the LiMn_0.95Mg_0.05PO₄ sample, the unit-volume is 302.94 ų, and for the Mn_0.95Mg_0.05PO₄, the value is 274.98 ų. The volume contraction is about 9.2%, which is smaller than 10% contraction in LiMnPO₄. The smaller volume contraction is favorable for the electrochemical performance.

Zoom In Zoom Out Reset image size

The cycle performance of the LiMn_0.95Mg_0.05PO₄ nanoplates is given in Fig. 8. In order to investigate the rate capability during cycling, various current rates of C/10, C/2, 1C and 2C were applied to the working electrode. It is clear that LiMn_0.95Mg_0.05PO₄ has a much higher reversible capacity than LiMnPO₄ at different discharge rates. The initial discharge capacity at C/10 is 116.6 mAh/g for the LiMn_0.95Mg_0.05PO₄. However for the pure LiMnPO₄, the value is only 96.2 mAh/g. Even at the higher discharge rate 2C, a discharge capacity of 69.8 mAh/g has been obtained for the LiMn_0.95Mg_0.05PO₄; for the pure LiMnPO₄, the value is only 55.3 mAh/g. The increase of capacity indicates substantially enhanced electrochemical kinetics of lithium-ion extraction/insertion from/into LiMn_0.95Mg_0.05PO₄, which was confirmed by the electrochemical impedance measurement as shown in Fig. 9. EIS plots of the LiMnPO₄ and LiMn_0.95Mg_0.05PO₄ cell were recorded at room temperature after the 1^st and 30^th cycles. Although the plots of both samples show a big change after the 30^th cycle, a much smaller semicircle was observed for the LiMn_0.95Mg_0.05PO₄ sample at the different cycles, which indicates a much smaller charge transfer resistance and greatly improved electrochemical kinetics of lithium ion extraction/insertion.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

In addition, Fig. 8 shows that LiMn_0.95Mg_0.05PO₄ shows a relatively stable capacity, especially for higher rates. At C/10 discharge rate, the discharge capacities for the 3rd and 20th cycles of pure LiMnPO₄ are 90.5 mAh/g and 85.1 mAh/g, respectively, and the mean fade of the capacity is 0.3% per cycle; the mean fade of the capacity of LiMn_0.95Mg_0.05PO₄ is only 0.2% per cycle. However, at higher rate, the cycle performance shows better improvement for the Mg-doped sample.

A recent study⁹ has shown that the Jahn-Teller active Mn³⁺ ions give a big structural misfit at the phase boundary between LiMnPO₄ and MnPO₄ (ca. 9.5–11.0% volume difference), which hinders motion of the interface.⁷ High-spin Mn³⁺ in an octahedral site has a twofold orbital degeneracy; site distortions bias the cooperative ordering of the occupied orbital as reported.⁹ Substitution of the Mn by cations such as Mg, Fe, Ni or Zr that do not have an octahedral-site d-orbital degeneracy lowers the energy of the cooperative orbital ordering and hence the magnitude of the component of the site distortion resulting from long-range orbital order. The resulting reduction of the volume and structural mismatch across the LiMnPO₄/MnPO₄ interface, which intersects the b axis, makes it easier to move the interface, thus increasing the rate capability. The more perfect crystallinity of the nanoplates may also enhance the interface mobility and extent of reversible motion.

Mg-doped LiMnPO₄ nanoplates with orthorhombic olivine structure can be prepared by a simple solvothermal method. The X-ray diffraction pattern and Rietveld refinement result show that Mg-doping does not change the crystal structure of LiMnPO₄ and the dopant Mg occupies the 4c Mn site to form a LiMn_0.95Mg_0.05PO₄ solid solution phase. High-resolution transmission electron microscopy images and selected area electron diffraction patterns show that, like the LiMnPO₄ nanoplates, the crystallographic facet of the LiMn_0.95Mg_0.05PO₄ nanoplates is the (100) plane. As a result, the b axis lies in the platelets and the LiMnPO₄/MnPO₄ interface intersects the Li⁺ pathway. This geometry means that the specific capacity is reduced where the interface motion is either blocked or too slow. Reduction of the volume and site-symmetry mismatch across the interface and a better crystallinity of the platelets increases the mobility of the interface to give a higher capacity at higher rates and smaller capacity fade on cycling.

This work was supported by the Assistant Secretary for Energy Efficiency and Renewable Energy, Office of Vehicle Technologies, U.S. Department of Energy, under Contract DE-AC02-05CH11231 through the Batteries for Advanced Transportation Technologies (BATT) Program Subcontract 6805919. The work was also supported by the Office of Basic Energy Sciences Office of Science, U.S. Department of Energy, under Contract DESC0005397.