PCCP View Article Online
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
PAPER
Cite this: Phys. Chem. Chem. Phys., 2015, 17, 18629
View Journal | View Issue
Hydrothermal synthesis, evolution, and electrochemical performance of LiMn0.5Fe0.5PO4 nanostructures† Wei Xiang,a Yan-Jun Zhong,a Jun-Yi Ji,a Yan Tang,b HuiHui Shen,a Xiao-Dong Guo,ac Ben-He Zhong,a Shi Xue Douc and Zhi-Ye Zhang*a LiMn0.5Fe0.5PO4 (LMFP) materials are synthesized by the hydrothermal approach in an organic-free and surfactant-free aqueous solution. The phase and morphological evolution of the material intermediates at different reaction temperatures and times are characterized by XRD, SEM and TEM, respectively. The results show that during temperature increase, the solubility product (Ksp) of the precursors (Li3PO4, Fe3(PO4)2 and (Mn,Fe)3(PO4)2) is the decisive parameter for the precipitation processes. Once the temperature locates at the range of 100–110 1C, the unstable precursors dissolve quickly and then LMFP nuclei are formed, followed by a dissolution-reprecipitation process. As the reaction progresses, the primary particles selfaggregate to form rod or plate particles to reduce the overall surface energy through oriented attachment (OA) and the Ostwald ripening (OR) mechanism. Moreover, the resultant concentration of the precursor
Received 8th May 2015, Accepted 22nd June 2015 DOI: 10.1039/c5cp02665b
significantly affects the crystal size of LMFP by altering the supersaturation degree of solution at the nucleation stage. The carbon coated LMFP nanostructure synthesized at 0.6 mol L 1 (resultant concentration of PO43 ) delivers discharge capacities of 155, 100 and 81 mA h g 1 at 0.1, 5 and 20 C rate, respectively. The understanding of nanostructural evolution and its influence on the electrochemical
www.rsc.org/pccp
performance will pave a way for a high-performance LMFP cathode.
1. Introduction Since the pioneering studies of Goodenough et al., olivine-type LiMPO4 (M = Fe, Mn, Co, and Ni) compounds with high thermal stability, cycling stability and superior safety properties have been widely investigated as promising alternative cathode materials especially for large-scale batteries designed for application in electric vehicles (EV) or in grid scale energy storage.1–4 Among them, LiFePO4 (LFP) is the most attractive cathode material for industrial application, owing to its low price, non-toxicity, and excellent compatibility with commercial electrolytes. By decreasing the characteristic size of particles to the nanoscale level, coating particles with electrically or ionically conductive layers and supervalent cation doping, the low electrochemical performance of LFP caused by poor electronic/ionic conductivity has been extensively studied and successfully overcome as evident in the recent a
College of Chemical Engineering, Sichuan University, Chengdu 610065, P. R. China. E-mail: [email protected] b College of Materials Science and Engineering, Sichuan University, Chengdu 610065, P. R. China c Institute for Superconducting and Electronic Materials, University of Wollongong, Wollongong, NSW 2522, Australia † Electronic supplementary information (ESI) available. See DOI: 10.1039/ c5cp02665b
This journal is © the Owner Societies 2015
commercialization of LFP as the cathode for Li-ion batteries.5–7 However, LFP is known to have low energy density because of its lower voltage platform (3.5 V, Fe2+/Fe3+ vs. Li/Li+) as compared with the 4 V class LiMO2 materials and their derivatives.8 To further increase the energy density, the iso-structural LiMnPO4 (LMP) is a more exciting cathode material due to its superior redox potential with flat plateau at 4.1 V versus Li/Li+. The theoretical energy density of LMP (701 W h kg 1) is 1.2 times greater than that of LFP (586.5 W h kg 1).9–12 Compared with LFP, LMP has been much less explored due to its extremely poor electronic conductivity (o10 10 S cm 1), Jahn–Teller distortion caused by unstable Mn(III) [d4:t2g3eg1] of MnPO4 when delithiated and a huge volumetric change between LiMnPO4 and MnPO4 during intercalation/de-intercalation (ca. 10%).13–15 LiMn1 xFexPO4 solid solution systems which combine the advantages of relative high electrical conductivity of LFP and relative high voltage of LMP have been widely studied.10,16–20 Ab initio calculations have shown that Fe substitution could increase the solubility limits between LiMn1 xFexPO4 and Mn1 xFexPO4, resulting in an expanded one-phase region and a contracted two-phase region.21 As lithiation/delithiation reactions in the one-phase region evade the sluggish kinetics of nucleation and growth of a new phase, an expanded one-phase region can improve the rate of delithation/lithiation reactions.22 Furthermore, Fe doping
Phys. Chem. Chem. Phys., 2015, 17, 18629--18637 | 18629
View Article Online
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
Paper
also dilutes the concentration of Mn. The resulting decrease in Jahn–Teller distortion caused by Mn3+ could improve electron conduction in the bulk and consequently the rate performance.21 Numerous synthetic methods to obtain LiMn1 xFexPO4 with improved electrochemical performance have been extensively investigated. Martha et al. have reported LiMn0.8Fe0.2PO4/C nanoparticles prepared by a conventional solid-state method with improved capacity.10 Oh et al. have synthesized LiMn1 xFexPO4/C on the basis of (Mn1 xFex)3(PO4)2 micro-spherical by co-precipitation reaction, which exhibited outstanding cycle ability and rate performance.8 Saravanan et al. have obtained LiMn0.5Fe0.5PO4 nanoplates through solvothermal reaction with a reversible capacity of 91 mA h g 1 at 5 C.23 As a soft-chemical route, the hydrothermal process is efficient, inexpensive, and sufficiently flexible so that the material’s properties (e.g. cation distribution, particle size, and morphology) can be generally tailored by the synthesis conditions, and it has been widely used to synthesize LFP nanomaterial with superior electrochemical properties.6,24–26 Researches have shown that the nucleation and crystallization can determine the particle size, purity, morphology and crystal structure of the product, which further influence the electrochemical performance.24,25,27,28 However, the nucleation and crystallization of LiMn1 xFexPO4 under hydrothermal condition is rare studied, especially for LiMn0.5Fe0.5PO4 (LMFP), which provides the optimal energy density, charge capability and durability for potential practical large-scale applications.8,29 Here, LMFP was prepared by the hydrothermal method under organic-free and surfactant-free conditions to avoid the interference of organics with the nature of LMFP. The intermediates of products were collected during temperature rising and soaking period. The particle morphology, composition, structure and electrochemical performance of the prepared samples were monitored. By understanding the particle growth mechanism and its relationship with resulting structures and function, this study is expected to optimize the growth conditions for hydrothermal synthesis of LMFP.
2. Experimental 2.1
Sample preparation
Nanostructure LMFP materials were synthesized by a simple hydrothermal method. Stoichiometric amounts of MnSO4H2O, FeSO47H2O, H3PO4 (85 wt% solution), and LiOHH2O with a molar ratio of 0.5 : 0.5 : 1 : 3 were used. Briefly, H3PO4 (85 wt%) was slowly added into the LiOH aqueous solution (2.5 mol L 1) with strong stirring at room temperature. Subsequently, MnSO4 and FeSO4 solution (1.25 mol L 1) was introduced into the neutralized mixture under stirring, forming a suspension with resultant concentration of precursor (PO43 ) at 0.5 mol L 1. After vigorous stirring at room temperature for 2 min, the suspension was transferred into a Teflon-lined stainless steel autoclave, into which a thermal couple and a pressure gauge were inserted to record the in situ temperature and pressure. The reactor was sealed and heated up to 100–180 1C for 0–600 min under mechanic stirring. To avoid the oxidation of Fe2+ (Mn2+),
18630 | Phys. Chem. Chem. Phys., 2015, 17, 18629--18637
PCCP
the mixing and hydrothermal treatment steps were operated under the protection of N2 atmosphere. After ambient cooling to room temperature, products were centrifuged and washed several times with deionized water. To achieve the carbon coating LiMn0.5Fe0.5PO4 composites, the hydrothermal products were mixed with 20 wt% of glucose, and then carbonized at 650 1C for 6 h under argon atmosphere with a heating rate of 5 1C min 1. For comparison, the LMFP was also prepared by the same process with resultant concentration of precursor (PO43 ) at 0.25 mol L 1 and 0.6 mol L 1 (the highest resultant concentration can be reached at room temperature). The LMFP prepared under resultant concentration of 0.25 mol L 1, 0.5 mol L 1 and 0.6 mol L 1 were denoted as LMFP-0.25, LMFP-0.5 and LMFP-0.6, and the carbon coated LMFP were denoted accordingly as LMFP-0.25/C, LMFP-0.5/C and LMFP-0.6/C, respectively.
2.2
Materials characterization
The crystalline phase of the as-prepared materials was identified by X-ray diffraction (XRD, D/max-rB, Rigaku, Japan) measurements using Cu-Ka radiation. The data for Rietveld refinement were collected in the 2y range of 5–1351 at a continuous scan mode with a step size of 0.021 for a counting time of 10 s. The GASA Rietveld program was used to analyze the diffraction patterns. Particle morphologies were observed by scanning electron microscopy (SEM, SPA400 Seiko Instruments) and fieldemission transmission electron microscopy (TEM, JEOL 2100F). The carbon content in composites was measured by a carbon-sulfur analyzer (CS-902, Wanlianda Xinke, Beijing, China).
2.3
Electrochemical measurements
The electrochemical performances of the as-prepared composite were evaluated using CR2032 coin-type cells. To prepare the working electrodes, the active materials, acetylene black, and poly(vinylidene fluoride) were mixed at a weight ratio of 80 : 13 : 7 in N-methyl pyrrolidone. The mixed viscous slurry was uniformly spread on an Al foil current collector with loading mass of active materials fixed about 5 mg cm 2. The electrodes dried under vacuum at 120 1C for 12 h. Subsequently, cells were assembled in glove box filled with argon using Li metal foil as anode and a polypropylene micro-porous film as separator. Then, 1 mol L 1 LiPF6 dissolved in a mixture of ethylene carbonate and dimethyl carbonate (1 : 1 by volume) was used as the electrolyte. The cells were tested in a voltage range of 2.5–4.5 V (vs. Li/Li+) using a constant-current protocol at various rates (1 C = 170 mA h g 1) on a battery testing system (Neware BTS-610). Cyclic voltammetry (CV) was performed on a LK9805 electrochemical analyser in the voltage range of 2.5–4.5 V at a scanning rate of 0.1 mV s 1. Electrochemical impedance spectroscopy (EIS) measurements were performed on a Zahner electrochemical workstation over the frequency range of 105 to 10 2 Hz with an applied amplitude of 5 mV. The EIS data for the three samples were collected at stable open circuit voltage (about 3.4 V). The Zview program was used to fit the EIS data.
This journal is © the Owner Societies 2015
View Article Online
PCCP
3. Results and discussion
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
3.1
Formation and evolution of LMFP
3.1.1 Transformation of precursor and LMFP during temperature increase. The composition and crystallographic structure transformation of the product were traced back by sampling the mixture at 30, 100, 110, 120, 130 and 140 1C during the temperature increase. Fig. 1a shows the XRD pattern of the asprepared Li3PO4, which can be indexed to orthorhombic Li3PO4 (JCPDS No.71-1528). The SEM image (Fig. 2a) shows that the aggregated quasi-spheres Li3PO4 consist of 30–50 nm nanocrystallites. The formation of quasi-sphere Li3PO4 could be explained by the difference solubility products of Li3PO4 (Ksp1 = 3.2 10 9) and Li2HPO4 (Ksp2 = 4.0 10 1).30 When the transition metal solution was introduced into the neutralized mixture and stirred at room temperature for 2 min, the XRD pattern of the sample (Fig. 1b) shows that the peaks can be only indexed as Li3PO4, and the intensity of peaks decrease with the adjunction of Fe2+ and Mn2+. SEM and TEM examinations (Fig. 2b and b-1) reveal that the precursors consist of quasispheres Li3PO4 and leaflet-like nanostructures. It indicates that the reaction is proceeded by a dissolution-reprecipitation mechanism.30,31 Owing to the Ksp value of transition metal phosphate (denoted as MexPO4) [1.0 10 36 for Fe3(PO4)2 and 1.0 10 27 for Mn3(PO4)2] is much lower than that of Li3PO4,30,32 the very minor amount of free PO43 available in solution, which is determined by the Ksp value of Li3PO4, are more preferred to combine with Fe2+ and Mn2+ than Li+ to form MexPO4 precipitation. With the depletion of the free PO43 , the concentration product of Li+ and PO43 becomes below the critical value of Ksp1, resulting in a portion dissolution of Li3PO4 and decrease of peaks intensity. From the viewpoint of thermodynamics, the cycles of Li3PO4 dissolution followed by MexPO4 precipitation would be continued until equilibrium between MexPO4 and
Fig. 1 XRD pattern of as-prepared Li3PO4 and the intermediates sampled at 30 1C (a), 100 1C (b), 110 1C (c), 120 1C (d) and 140 1C (e) during the temperature rising period.
This journal is © the Owner Societies 2015
Paper
Me2+ is reached. However, the dissolution-reprecipitation cycles would be limited by their kinetics at such low temperature and pressure.33 Meanwhile, the precipitation time may not be enough to obtain the thermodynamic equilibrium states at the concentration. Thus, only a part of MexPO4 is precipitated or a major amount of MexPO4 exists in amorphous form (Fig. 2b-1-1), resulting in the absence of typical diffraction peaks of MexPO4. The XRD pattern of the intermediate sampled at 100 1C (Fig. 1c) indicates that the peak intensity of Li3PO4 decrease along with the detection of (Mn,Fe)3(PO4)24H2O (metaswitzerite, JCPDS No. 20-0713), as well as small quantities of Fe3(PO4)28H2O (viviantie, JCPDS No. 30-0662). The SEM and TEM images (Fig. 2c and c-1) shows that the leaflet-like nanostructure changes into larger and thicker sheet-like crystallite with regular morphology, accompanying with the sacrifice of Li3PO4. It can be speculated that the crystallization and growth of thermodynamically preferred precipitates (MexPO4) are easily fulfilled at higher temperature for the acceleration of the dissolution-reprecipitation cycles. In addition, the simultaneous detection of metaswitzerite and viviantie imply that a complete solid solution [(Mn0.5Fe0.5)3(PO4)2xH2O] is not formed, which is also found by Oh et al. in synthesis of LiMn0.5Fe0.5PO4 via (Mn0.5Fe0.5)3(PO4)2 precursor.8 At reaction temperature of 110 1C (Fig. 1d), the intensity of the characteristic Bragg peaks corresponding to the Li3PO4, (Mn,Fe)3(PO4)24H2O and Fe3(PO4)28H2O precursor fall sharply and some weak peaks corresponding to LMFP appear. In Fig. 2d, the quasi-spheres Li3PO4 and sheet-like crystallite are disappeared with the formation of grain-like nanocrystallite. The detection of LMFP verifies that the critical temperature for the initial nucleation LMFP locates at the range of 100–110 1C. As the dissolution of (Mn,Fe)3(PO4)24H2O and Fe3(PO4)28H2O, the amount of dissolved ions (Fe2+, Mn2+and PO43 ) sacrificed from (Mn,Fe)3(PO4)24H2O or Fe3(PO4)2PO4 are sufficient to bond the Li+, thus providing high concentrations of nutrient for nucleation of LFMP.24,27,28 With the temperature further increasing to 120 1C (Fig. 1e), all the reflections are indexed as an orthorhombic olivine-type structure (JCPDS No. 42-0580), and no additional impurity phases are observed, indicating the complete formation of LMFP. The XRD patterns of intermediate sampled at 140 1C (Fig. 1f) show stronger LMFP diffraction peaks, suggesting the fast growth of the crystal. Fig. 2e and f show the SEM images of intermediates extracted at 120 1C and 140 1C, respectively. Although the change of the crystallite morphology with respect to the number of grain boundaries per particle appear to be negligible, the size of the crystallite grow larger as the reaction temperature rising (Fig. 2e, e-1, f and f-1). 3.1.2 Evolution of the LMFP crystal under 180 8C at different times during the soaking period. Fig. 3a shows the XRD patterns of the intermediate extracted during the soaking period. All the diffraction patterns could be indexed to an olivine-type LMFP with orthorhombic crystal structure (space group Pnma), indicating reaching pure phase at 180 1C with different time spots. With the extension of the soaking period, the intensity of the diffraction peaks slightly increases. It is worth noting that the transformation of the relative intensity of the diffraction peaks is quite different from the results on the hydrothermal synthesis of LFP reported by Ou et al.27 In their study, the relative intensity of (200)/(121) keeps
Phys. Chem. Chem. Phys., 2015, 17, 18629--18637 | 18631
View Article Online
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
Paper
PCCP
Fig. 2 SEM images of the as-prepared Li3PO4 microspheres (a) and the intermediate extracted at 30 1C (b), 100 1C (c), 110 1C (d), 120 1C (e) and 140 1C (f) during temperature rising. TEM and corresponding SAED patterns of the intermediate extracted at 30 1C (b-1, b-1-1) and 100 1C (c-1, c-1-1). The corresponding particle size distribution of the intermediate extracted at 110 1C (d-1), 120 1C (e-1) and 140 1C (f-1).
increasing throughout the soaking period. Contemporaneously, the crystal growth orientates along the (200) and/or the (121) plane, resulting formation of nanoplates with large bc plane. However, there is no obvious change in the relative intensity of three main diffraction peaks [(111)/(021), (121)/(200), (131)] in our results, implying no transformation of morphology with respect to preferential crystal orientation.34 Rietveld refinement on the XRD patterns were shown in the Fig. 3b, which shows a good fit between observed and calculated patterns. The lattice parameters of LMFP samples were refined and listed in Table S1 (ESI†). It is found that the unit cell volume of LMFP extracted at 0 min is 297.59 Å3, and decrease to 295.42 Å3 as the reaction time reached to 600 min. Normally, the unit cell contraction of hydrothermally synthesized phosphate is the results of the decrease of antisite defects with the increase of temperature or time.35 Since Li+ ion diffusion in phosphate occurs along onedimensional channels, the reduction of Fe on Li site improves lithium transport in that channel,36 enhancing the capacity and rate capability. Meanwhile, the decrease of cell volume suggests the increased crystallinity of intermediates.28 Fig. 4 presents the size and morphological evolution of the intermediates extracted at different time spots. As the reaction duration extends, the primary crystallite increase in size with a concurrent transformation of morphology. For example, the crystals of intermediate extracted at 0 min are nanoparticles (Fig. 4a).
18632 | Phys. Chem. Chem. Phys., 2015, 17, 18629--18637
With increasing reaction durations (i.e., 60 min in Fig. 4c), the crystals of nanostructure become a mixture of rods and nanoparticles. At the longest reaction duration (600 min in Fig. 4d), significantly larger plates coexist with some nanoparticles and rods. It is considered that the nonuniformity in crystal morphologies arises from the multiple nucleation events, which always give rise to crystals of various growth extents.24 The iso-oriented nanorods could attribute to the aggregation of the primary nanoparticles via oriented attachment (OA), which forms a single crystal upon fusion of the nanoparticles.31,37 During the further crystallization process, the primary nanoparticles continue joining up with the increased crystals through Ostwald ripening (OR). Driven by surface energy minimization, thermodynamically unstable primary nanoparticles dissolve and larger particles (nanorods) grow into plates by adsorbing monomers from the dissolved nanoparticles.24,31,38,39 The size and morphology transformation imply that the crystals undergo a distinct growth process during the soaking period. It is worthwhile to note that the evolution of the intermediate is largely determined by the resultant concentration of precursor (PO43 ). When the resultant concentration of precursor is decreased to 0.25 mol L 1 while keeping other parameters constant, the formation and transformation of the precursor are hindered. The XRD patterns of the intermediates (Fig. S2, ESI†) show that the precipitation of (Mn,Fe)3(PO4)24H2O and Fe3(PO4)28H2O are substituted by separate formation of Fe3(PO4)28H2O and
This journal is © the Owner Societies 2015
View Article Online
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
PCCP
Paper
Fig. 3 (a) XRD patterns of the intermediate extracted at 180 1C during soaking period, (b) Rietveld refinement plots of the XRD patterns of the intermediate.
Mn3(PO4)21.5H2O at different temperature. The sequential precipitation of Fe3(PO4)28H2O and Mn3(PO4)21.5H2O could be illustrated by the dissolution-reprecipitation mechanism. Because the value of solubility product of Fe3(PO4)2 is lower than that of Mn3(PO4)2, the dissolved PO43 are more preferred to combine with Fe2+ than Mn2+. Under low concentration, the effects induced by the difference of solubility product may become notable. Compared with the formation of Fe3(PO4)2, more amount of Li3PO4 is required to be dissolved for initiating the Mn3(PO4)2 precipitation. Meanwhile, the formation of phase pure LMFP is postponed to the temperature of 130 1C. Large quantity of Mn3(PO4)21.5H2O and Li3PO4 still exist in the precipitation at the temperature of 120 1C. Typically, the nucleation and crystal growth under hydrothermal condition are governed by the concentration of ions and the kinetic factor.28,35 The amount of dissolved ions at a relatively low concentration solution is insufficient, which results in the deficiency of crystal nucleus and the slowness of the crystal growth rate. Furthermore, the transportation of species to nucleation sites may be hindered in relatively low concentration, resulting the limited crystallization process.35 3.2
Fig. 4 SEM images and the corresponding particle size distribution of the intermediate extracted at 180 1C during soaking period: 0 min (a. a-1), 10 min (b, b-1), 60 min (c, c-1), 600 min (d, d-1).
be mainly expressed by a five-step process. Initially, quasi-spheres Li3PO4 are formed by neutralization reaction between H3PO4 and LiOH solution at room temperature. When the Fe2+ (Mn2+) solution is added into the mixture, the leaflet-like MexPO4 is formed with the coexistence of Li3PO4. At the low temperature and pressure, the dissolution-reprecipitation cycle cannot proceed completely, resulting in the formation of amorphous metal phosphate precipitation. In the subsequent hydrothermal process, the size and crystallinity of metal phosphate increase with the acceleration of dissolution–reprecipitation cycle at higher temperature.
Mechanism for nucleation and crystal growth
On the basis of the XRD, SEM and TEM results, accompanied with confirmation of dissolution-reprecipitation, a potential reaction mechanism of the LFMP nanostructure is proposed as schematically illustrated in Fig. 5. The reaction process can
This journal is © the Owner Societies 2015
Fig. 5 Schematic illustration of the growth mechanism involved in the hydrothermal synthesis of LiMn0.5Fe0.5PO4 nanostructure.
Phys. Chem. Chem. Phys., 2015, 17, 18629--18637 | 18633
View Article Online
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
Paper
PCCP
With temperature further increasing, unstable compounds in the precursor begin to dissolve and hydrolyze quickly due to the increase of solubility at high temperature, providing high concentration nutrient for nucleation and growth of LFMP. Finally, the primary particles evolved from LFMP nuclei would self-aggregate to form rods and plates to reduce the overall surface energy through oriented attachment (OA) and Ostwald ripening (OR) mechanism. According to the nucleation and growth mechanism proposed above, the resultant concentration of precursor would have distinct impact on the characteristic size of the crystals. To verify whether this is the case or not, we loaded three independent autoclaves with the resultant concentration at 0.25 mol L 1, 0.5 mol L 1 and 0.6 mol L 1 to synthesize LMFP. The XRD patterns of the three samples are shown in Fig. S2 (ESI†). Morphologies of the three samples shown in Fig. 6 confirm that the improved resultant concentration significantly reduces the particle size of the crystals. It is generally acknowledged that increasing nucleation sites could decrease crystal size of the precipitation in wet chemical process.25 As the nucleation of LMFP is based on the dissolution of Li3PO4 and Me3(PO4)2, improving the result concentration of precursor significantly increase the concentration of free ions (Li+, Mn2+, Fe2+ and PO43+) at the dissolution stage, resulting higher supersaturation degree for nucleation. Consequently, the accelerated nucleation
Fig. 6
rate would favor a great number of nuclei and facilitate smaller particles. 3.3
Electrochemical performance
In order to associate the LFMP structure with the performance, electrochemical measurements were conducted for crystalline synthesized under three concentrations. Fig. 7 shows the second cyclic voltammetry of LiMn0.5Fe0.5PO4/C after ‘‘activation’’ at a scan rate of 0.1 mV s 1 at room temperature. The CV curves show two oxidation peaks and two reduction peaks, corresponding to the charge/discharge reactions of the Fe2+/ Fe3+ and Mn2+/Mn3+ redox couple, respectively. It is noted that a small redox peak (B3.6 V) is observed in the CV, which corresponds to the small discharge plateau at B3.6 V as seen in the discharge curves at 1 C rate in Fig. 8b–d. This small peak was also found in earlier reports concerning of LiFe1 xMnxPO4 solid solution,40–42 which guess that the peak could belong to a Mn2P2O7 impurity phase that could not be detected by XRD.41 For LMFP-0.6/C, the voltage hysteresis (DV, difference between the anodic and cathodic peak voltages) of Mn3+/Mn2+ and Fe3+/ Fe2+ are 260 mV and 140 mV, respectively. By contrast, those hysteresis are 270 mV and 150 mV for LMFP-0.5/C, 340 mV and 240 mV for LMFP-0.25/C. Besides, the peak currents of LMFP0.6/C are also higher than those synthesized under lower concentration. It is well known that smaller voltage hysteresis
SEM, TEM and the corresponding HRTEM images of LMFP-0.25/C (a, d and g), LMFP-0.5/C (b, e and h) and LMFP-0.6/C (c, f and i).
18634 | Phys. Chem. Chem. Phys., 2015, 17, 18629--18637
This journal is © the Owner Societies 2015
View Article Online
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
PCCP
Fig. 7 The representative CV profiles of LMFP-0.6/C, LMFP-0.5/C and LMFP-0.25/C at a scan rate of 0.1 mV s 1.
Fig. 8 (a) Galvanstatic cycling profiles of LMFP-0.25/C, LMFP-0.5/C and LMFP-0.6/C at 0.1 C. (b) Discharge profiles of LMFP-0.25/C at different rates. (c) Discharge profiles of LMFP-0.5/C at different rates. (d) Discharge profiles of LMCP-0.6/C at different rates. (e) Rate capability of LMFP-0.25/C, LMFP-0.5/C and LMFP-0.6/C at different rates. (f) Cycling performance of LMFP-0.25/C, LMFP-0.5/C and LMFP-0.6/C at a discharge rate of 1 C for 100 cycles.
as well as higher peak currents indicate better electrode reaction kinetics and thus better electrochemical performance.21,43 Considering the similar amounts of carbon (2.71, 2.73 and 2.68 wt% for LMFP-0.25/C, LMFP-0.5/C and LMFP-0.6/C, respectively), the enhanced kinetics for LMFP-0.6/C electrode during the lithiation and delithiatoin can be mainly attributed to the crystal size reduction. The results of cyclic voltammetry suggest that the LMFP synthesized at higher resultant concentration of precursor would provide a better charge/discharge performance. The prediction was further confirmed by the charge/discharge
This journal is © the Owner Societies 2015
Paper
measurements. The initial charge/discharge profiles at current rates of 0.1 C are shown in Fig. 8a. The discharge curves of LiMn0.5Fe0.5PO4/C contain two flat voltage plateaus at 4.1 V and 3.5 V, associating with the Mn3+/Mn2+ and Fe3+/Fe2+ redox reactions, respectively. LMFP-0.6/C exhibits a better specific discharge capacity of 155.7 mA h g 1 (91.6% of the theoretical capacity) relative to LMFP-0.25/C and LMFP-0.5/C, which deliver 139.1 and 145.1 mA h g 1, respectively. The voltage profiles of LMFP-0.25/C, LMFP-0.5/C and LMFP-0.6/C at various charge/ discharge rates from C/10 to 20 C are shown in Fig. 8b–d. The fact that high voltage plateau of discharge profile is decreasing under high discharge current may originate from the increased polarization of Mn2+/Mn3+ couple.41 Compared with LMFP-0.25/C and LMFP-0.5/C, LMFP-0.6/C sample achieves the highest capacity at each C-rate, demonstrating the low internal resistance in solid-sate Li+ diffusion.21 Fig. 8e shows the rate capabilities of LMFP-0.25/C, LMFP0.5/C and LMFP-0.6/C. As expected, the LMFP-0.6/C exhibits the most excellent rate performance, with discharge capacities of 144, 127, 116, 106, 100, 91 and 81 mA h g 1 at 0.2, 0.5, 1, 3, 5, 10 and 20 C, respectively. However, the LMFP-0.25/C and LMFP0.5/C with large crystal sizes show aggravated degradation behavior and lower discharge capacities. For example, LMFP0.25/C and LMFP-0.5/C only deliver 59 and 77 mA h g 1 of discharge capacity respectively at 20 C. The cycling performance recorded after the rate performance is shown in Fig. 8f. At the first cycle, cells with LMFP-0.25/C, LMFP-0.5/C and LMFP-0.6/C can deliver discharge capacities of 87, 104, 109 mA h g 1 respectively, at 1 C. After cycling for 100 times, the discharge capacity retention of the cell with LMFP-0.6/C is 90.7% of the initial value, which is significantly outperformed compared with those of LMFP-0.25/C and LMFP-0.5/C, with only 88.4 and 87.0% respectively. It is reasonable to assume that the difference in rate performance and cycling performance are mainly attributed to the different particle sizes of these samples. For LMFP-0.6/C, the reduced crystal size favors charge transfer and Li ions diffusion, improving the yieldable capacity at higher rates. The reason for the difference in electrochemical performances of three samples under mutative reactant concentrations were further investigated by electrochemical impedance spectroscopic (EIS) measurements. Fig. 9a shows the Nyquist plots of the three samples collected with a two electrode coin cell in the discharged state after 100 charge–discharge cycles. Each of the curves has a depressed semicircle in the high-tomedium frequency and an oblique line in the low-frequency. The intercept of the curve in the high frequency region relates to the ohmic series resistance (Rs), which includes the interparticle contact resistance, electrolyte resistance and other physical resistance. The following depressed semicircle in the medium-frequency region corresponds to the charge transfer resistance (Rct) at the electrolyte-electrode interface. The straight line in the low-frequency region associated with the Warburg impedance (Zw), related to the diffusion behavior of lithium ions within particles. By comparing the diameters of the semicircles, the impedance of LMFP-0.6/C electrode is
Phys. Chem. Chem. Phys., 2015, 17, 18629--18637 | 18635
View Article Online
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
Paper
PCCP
Fig. 9 (a) Nyquist impedance spectra and equivalent circuit of LMFP0.25/C, LMFP-0.5/C and LMFP-0.6/C at room temperature, (b) the relationship between Zre and o 1/2 in the low frequency region.
significantly smaller than those of LMFP-0.5/C and LMFP-0.25/ C electrode. The values of Rct for LMFP-0.6/C, LMFP-0.5/C and LMFP-0.25/C electrodes were fitted by the given equivalent circuit and found to be 71.7 O, 268.2 O, and 554.3 O, respectively. The Li+ diffusion coefficient (DLi+) was calculated for these samples according to the following equation: DLi+ = (R2T2)/(2A2n4F4C4s2)
(1)
here R is the gas constant (8.314 J K 1 mol 1), T is the absolute temperature (298.15 K), A is the surface area of the electrode, n is the number of the electrons per molecule during oxidization (here, n = 1), F is the Faraday constant (96 500 C mol 1), C is the concentration of Li ions (7.69 10 3 mol cm 3) and s is the Warburg factor (O s 1/2) associated with Zre (the imaginary part of cell impedance, ohm) and o (the frequency, Hz), calculated by equation: Zre p so
1/2
(2)
Therefore the Warburg factor can be obtained from the slope between Zre and the inverse square root of the angular frequency, as shown in the Fig. 9b. Accordingly, the apparent Li ion diffusion coefficients of LMFP-0.6/C, LMFP-0.5/C and LMFP-0.25/C were determined to be 8.23 10 14, 2.59 10 14 and 6.0 10 15 cm2 s 1, respectively. Considering the similar amounts of carbon, the lowest Rct value and enhanced mobility of Li+ ions of LMFP-0.6/C are attributed to the reduced crystal size. The larger specific surface area and smaller crystal size of LMFP-0.6/C provide a higher electrolyte contact interface and shorter diffusion distance of Li+ ions, which lowers the charge transfer resistance and facilitates the electrochemical reaction, leading to enhanced electrochemical performance.
4. Conclusion In summary, we systematically investigated the phase and morphological evolution of the intermediate in the hydrothermal synthesis of LiMn0.5Fe0.5PO4. It shows that LiMn0.5Fe0.5PO4 can be formed at the temperature range as low as 100–110 1C. The formation of LMFP occurs by a straightforward dissolution of Li3PO4, Fe3(PO4)28H2O and (Mn,Fe)3(PO4)24H2O. As primary particles evolved from the nuclei of LMFP, the driven force induced by the reduction of overall surface energy enables the primary particles
18636 | Phys. Chem. Chem. Phys., 2015, 17, 18629--18637
to self-aggregate to form rods and plates through the known oriented attachment (OA) and Ostwald ripening (OR) mechanism. We also demonstrated that the resultant concentration of precursor plays a decisive role in determining the crystal size of LMFP by altering the nucleation rate. Electrochemical characterization of LMFP materials discloses a discharge capacity of 155 mA h g 1 at 0.1 C. Crystals synthesized at higher concentration deliver superior overall electrochemical performances compared with those synthesized at lower concentrations. In this study, the clear and unambiguous understanding of the formation mechanism provides a profound guiding significance for the synthesis of LMFP or other olivine-type cathode materials with shorter Li+ diffusion paths and improved performances.
Acknowledgements The authors appreciate the support of Sichuan University Funds for Young Scientists (No. 2011SCU11081), the PhD Programs Foundation of Ministry of Education of China (No. 20120181120103), the Project funded by China Postdoctoral Science Foundation (No. 2014M562322), AutoCRC Project 1-111 ‘‘Development of Advanced Electrode and Electrolytes for LIB’’.
Notes and references 1 C. Masquelier and L. Croguennec, Chem. Rev., 2013, 113, 6552–6591. 2 A. K. Padhi, K. S. Nanjundaswamy and J. B. Goodenough, J. Electrochem. Soc., 1997, 144, 1188–1194. 3 C. Sung-Yoon, J. T. Bloking and C. Yet-Ming, Nat. Mater., 2002, 1, 123–128. 4 Z. H. Chen and J. R. Dahn, J. Electrochem. Soc., 2002, 149, A1184–A1189. 5 L.-X. Yuan, Z.-H. Wang, W.-X. Zhang, X.-L. Hu, J.-T. Chen, Y.-H. Huang and J. B. Goodenough, Energy Environ. Sci., 2011, 4, 269–284. 6 M. K. Devaraju and I. Honma, Adv. Energy Mater., 2012, 2, 284–297. 7 J. Wang and X. Sun, Energy Environ. Sci., 2012, 5, 5163–5185. 8 S.-M. Oh, S.-T. Myung, Y. S. Choi, K. H. Oh and Y.-K. Sun, J. Mater. Chem., 2011, 21, 19368–19374. 9 C. Delacourt, P. Poizot, M. Morcrette, J. M. Tarascon and C. Masquelier, Chem. Mater., 2004, 16, 93–99. 10 S. K. Martha, J. Grinblat, O. Haik, E. Zinigrad, T. Drezen, J. H. Miners, I. Exnar, A. Kay, B. Markovsky and D. Aurbach, Angew. Chem., Int. Ed., 2009, 48, 8559–8563. 11 V. Aravindan, J. Gnanaraj, Y.-S. Lee and S. Madhavi, J. Mater. Chem. A, 2013, 1, 3518–3539. 12 W. Xiang, Y. Zhong, Y. Tang, H. Shen, E. Wang, H. Liu, B. Zhong and X. Guo, J. Alloys Compd., 2015, 635, 180–187. 13 C. Delacourt, L. Laffont, R. Bouchet, C. Wurm, J. B. Leriche, M. Morcrette, J. M. Tarascon and C. Masquelier, J. Electrochem. Soc., 2005, 152, A913–A921. 14 G. Chen, J. D. Wilcox and T. J. Richardson, Electrochem. Solid-State Lett., 2008, 11, A190–A194.
This journal is © the Owner Societies 2015
View Article Online
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
PCCP
15 D. Choi, J. Xiao, Y. J. Choi, J. S. Hardy, M. Vijayakumar, M. S. Bhuvaneswari, J. Liu, W. Xu, W. Wang, Z. Yang, G. L. Graff and J.-G. Zhang, Energy Environ. Sci., 2011, 4, 4560–4566. 16 Y.-K. Sun, S.-M. Oh, H.-K. Park and B. Scrosati, Adv. Mater., 2011, 23, 5050–5054. 17 S.-M. Oh, S.-T. Myung, J. B. Park, B. Scrosati, K. Amine and Y.-K. Sun, Angew. Chem., Int. Ed., 2012, 51, 1853–1856. 18 H. Wang, Y. Yang, Y. Liang, L.-F. Cui, H. S. Casalongue, Y. Li, G. Hong, Y. Cui and H. Dai, Angew. Chem., Int. Ed., 2011, 50, 7364–7368. 19 K. M. O. Jensen, M. Christensen, H. P. Gunnlaugsson, N. Lock, E. D. Bojesen, T. Proffen and B. B. Iversen, Chem. Mater., 2013, 25, 2282–2290. 20 G. R. Gardiner and M. S. Islam, Chem. Mater., 2010, 22, 1242–1248. 21 B. Ding, P. Xiao, G. Ji, Y. Ma, L. Lu and J. Y. Lee, ACS Appl. Mater. Interfaces, 2013, 5, 12120–12126. 22 P. Gibot, M. Casas-Cabanas, L. Laffont, S. Levasseur, P. Carlach, S. Hamelet, J.-M. Tarascon and C. Masquelier, Nat. Mater., 2008, 7, 741–747. 23 K. Saravanan, V. Ramar, P. Balaya and J. J. Vittal, J. Mater. Chem., 2011, 21, 14925–14935. 24 X. Qin, J. Wang, J. Xie, F. Li, L. Wen and X. Wang, Phys. Chem. Chem. Phys., 2012, 14, 2669–2677. 25 X. Qin, X. Wang, H. Xiang, J. Xie, J. Li and Y. Zhou, J. Phys. Chem. C, 2010, 114, 16806–16812. 26 K. Dokko, S. Koizumi, H. Nakano and K. Kanamura, J. Mater. Chem., 2007, 17, 4803–4810. 27 X. Ou, H. Gu, Y. Wu, J. Lu and Y. Zheng, Electrochim. Acta, 2013, 96, 230–236. 28 X. Ou, L. Pan, H. Gu, Y. Wu and J. Lu, J. Mater. Chem., 2012, 22, 25493.
This journal is © the Owner Societies 2015
Paper
29 P. F. Xiao, B. Ding, M. O. Lai and L. Lu, J. Electrochem. Soc., 2013, 160, A918–A926. 30 M.-H. Lee, J.-Y. Kim and H.-K. Song, Chem. Commun., 2010, 46, 6795–6797. 31 J. Zhu, J. Fiore, D. Li, N. M. Kinsinger, Q. Wang, E. DiMasi, J. Guo and D. Kisailus, Cryst. Growth Des., 2013, 13, 4659–4666. 32 T.-H. Kim, H.-S. Park, M.-H. Lee, S.-Y. Lee and H.-K. Song, J. Power Sources, 2012, 210, 1–6. 33 M.-H. Lee, T.-H. Kim, Y. S. Kim, J.-S. Park and H.-K. Song, J. Mater. Chem., 2012, 22, 8228–8234. 34 H. Guo, C. Wu, J. Xie, S. Zhang, G. Cao and X. Zhao, J. Mater. Chem. A, 2014, 2, 10581–10588. 35 J. Chen, J. Bai, H. Chen and J. Graetz, J. Phys. Chem. Lett., 2011, 2, 1874–1878. 36 L. Gu, C. Zhu, H. Li, Y. Yu, C. Li, S. Tsukimoto, J. Maier and Y. Ikuhara, J. Am. Chem. Soc., 2011, 133, 4661–4663. 37 H. Uchiyama and H. Imai, Cryst. Growth Des., 2010, 10, 1777–1781. 38 P. Barpanda, K. Djellab, N. Recham, M. Armand and J.-M. Tarascon, J. Mater. Chem., 2011, 21, 10143–10152. 39 G. Qin, S. Xue, Q. Ma and C. Wang, CrystEngComm, 2014, 16, 260–269. 40 X. Zhou, Y. Xie, Y. Deng, X. Qin and G. Chen, J. Mater. Chem. A, 2015, 3, 996–1004. 41 R. von Hagen, H. Lorrmann, K.-C. Moeller and S. Mathur, Adv. Energy Mater., 2012, 2, 553–559. 42 Y.-J. Zhong, J.-T. Li, Z.-G. Wu, X.-D. Guo, B.-H. Zhong and S.-G. Sun, J. Power Sources, 2013, 234, 217–222. 43 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.
Phys. Chem. Chem. Phys., 2015, 17, 18629--18637 | 18637
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
PAPER
Cite this: Phys. Chem. Chem. Phys., 2015, 17, 18629
View Journal | View Issue
Hydrothermal synthesis, evolution, and electrochemical performance of LiMn0.5Fe0.5PO4 nanostructures† Wei Xiang,a Yan-Jun Zhong,a Jun-Yi Ji,a Yan Tang,b HuiHui Shen,a Xiao-Dong Guo,ac Ben-He Zhong,a Shi Xue Douc and Zhi-Ye Zhang*a LiMn0.5Fe0.5PO4 (LMFP) materials are synthesized by the hydrothermal approach in an organic-free and surfactant-free aqueous solution. The phase and morphological evolution of the material intermediates at different reaction temperatures and times are characterized by XRD, SEM and TEM, respectively. The results show that during temperature increase, the solubility product (Ksp) of the precursors (Li3PO4, Fe3(PO4)2 and (Mn,Fe)3(PO4)2) is the decisive parameter for the precipitation processes. Once the temperature locates at the range of 100–110 1C, the unstable precursors dissolve quickly and then LMFP nuclei are formed, followed by a dissolution-reprecipitation process. As the reaction progresses, the primary particles selfaggregate to form rod or plate particles to reduce the overall surface energy through oriented attachment (OA) and the Ostwald ripening (OR) mechanism. Moreover, the resultant concentration of the precursor
Received 8th May 2015, Accepted 22nd June 2015 DOI: 10.1039/c5cp02665b
significantly affects the crystal size of LMFP by altering the supersaturation degree of solution at the nucleation stage. The carbon coated LMFP nanostructure synthesized at 0.6 mol L 1 (resultant concentration of PO43 ) delivers discharge capacities of 155, 100 and 81 mA h g 1 at 0.1, 5 and 20 C rate, respectively. The understanding of nanostructural evolution and its influence on the electrochemical
www.rsc.org/pccp
performance will pave a way for a high-performance LMFP cathode.
1. Introduction Since the pioneering studies of Goodenough et al., olivine-type LiMPO4 (M = Fe, Mn, Co, and Ni) compounds with high thermal stability, cycling stability and superior safety properties have been widely investigated as promising alternative cathode materials especially for large-scale batteries designed for application in electric vehicles (EV) or in grid scale energy storage.1–4 Among them, LiFePO4 (LFP) is the most attractive cathode material for industrial application, owing to its low price, non-toxicity, and excellent compatibility with commercial electrolytes. By decreasing the characteristic size of particles to the nanoscale level, coating particles with electrically or ionically conductive layers and supervalent cation doping, the low electrochemical performance of LFP caused by poor electronic/ionic conductivity has been extensively studied and successfully overcome as evident in the recent a
College of Chemical Engineering, Sichuan University, Chengdu 610065, P. R. China. E-mail: [email protected] b College of Materials Science and Engineering, Sichuan University, Chengdu 610065, P. R. China c Institute for Superconducting and Electronic Materials, University of Wollongong, Wollongong, NSW 2522, Australia † Electronic supplementary information (ESI) available. See DOI: 10.1039/ c5cp02665b
This journal is © the Owner Societies 2015
commercialization of LFP as the cathode for Li-ion batteries.5–7 However, LFP is known to have low energy density because of its lower voltage platform (3.5 V, Fe2+/Fe3+ vs. Li/Li+) as compared with the 4 V class LiMO2 materials and their derivatives.8 To further increase the energy density, the iso-structural LiMnPO4 (LMP) is a more exciting cathode material due to its superior redox potential with flat plateau at 4.1 V versus Li/Li+. The theoretical energy density of LMP (701 W h kg 1) is 1.2 times greater than that of LFP (586.5 W h kg 1).9–12 Compared with LFP, LMP has been much less explored due to its extremely poor electronic conductivity (o10 10 S cm 1), Jahn–Teller distortion caused by unstable Mn(III) [d4:t2g3eg1] of MnPO4 when delithiated and a huge volumetric change between LiMnPO4 and MnPO4 during intercalation/de-intercalation (ca. 10%).13–15 LiMn1 xFexPO4 solid solution systems which combine the advantages of relative high electrical conductivity of LFP and relative high voltage of LMP have been widely studied.10,16–20 Ab initio calculations have shown that Fe substitution could increase the solubility limits between LiMn1 xFexPO4 and Mn1 xFexPO4, resulting in an expanded one-phase region and a contracted two-phase region.21 As lithiation/delithiation reactions in the one-phase region evade the sluggish kinetics of nucleation and growth of a new phase, an expanded one-phase region can improve the rate of delithation/lithiation reactions.22 Furthermore, Fe doping
Phys. Chem. Chem. Phys., 2015, 17, 18629--18637 | 18629
View Article Online
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
Paper
also dilutes the concentration of Mn. The resulting decrease in Jahn–Teller distortion caused by Mn3+ could improve electron conduction in the bulk and consequently the rate performance.21 Numerous synthetic methods to obtain LiMn1 xFexPO4 with improved electrochemical performance have been extensively investigated. Martha et al. have reported LiMn0.8Fe0.2PO4/C nanoparticles prepared by a conventional solid-state method with improved capacity.10 Oh et al. have synthesized LiMn1 xFexPO4/C on the basis of (Mn1 xFex)3(PO4)2 micro-spherical by co-precipitation reaction, which exhibited outstanding cycle ability and rate performance.8 Saravanan et al. have obtained LiMn0.5Fe0.5PO4 nanoplates through solvothermal reaction with a reversible capacity of 91 mA h g 1 at 5 C.23 As a soft-chemical route, the hydrothermal process is efficient, inexpensive, and sufficiently flexible so that the material’s properties (e.g. cation distribution, particle size, and morphology) can be generally tailored by the synthesis conditions, and it has been widely used to synthesize LFP nanomaterial with superior electrochemical properties.6,24–26 Researches have shown that the nucleation and crystallization can determine the particle size, purity, morphology and crystal structure of the product, which further influence the electrochemical performance.24,25,27,28 However, the nucleation and crystallization of LiMn1 xFexPO4 under hydrothermal condition is rare studied, especially for LiMn0.5Fe0.5PO4 (LMFP), which provides the optimal energy density, charge capability and durability for potential practical large-scale applications.8,29 Here, LMFP was prepared by the hydrothermal method under organic-free and surfactant-free conditions to avoid the interference of organics with the nature of LMFP. The intermediates of products were collected during temperature rising and soaking period. The particle morphology, composition, structure and electrochemical performance of the prepared samples were monitored. By understanding the particle growth mechanism and its relationship with resulting structures and function, this study is expected to optimize the growth conditions for hydrothermal synthesis of LMFP.
2. Experimental 2.1
Sample preparation
Nanostructure LMFP materials were synthesized by a simple hydrothermal method. Stoichiometric amounts of MnSO4H2O, FeSO47H2O, H3PO4 (85 wt% solution), and LiOHH2O with a molar ratio of 0.5 : 0.5 : 1 : 3 were used. Briefly, H3PO4 (85 wt%) was slowly added into the LiOH aqueous solution (2.5 mol L 1) with strong stirring at room temperature. Subsequently, MnSO4 and FeSO4 solution (1.25 mol L 1) was introduced into the neutralized mixture under stirring, forming a suspension with resultant concentration of precursor (PO43 ) at 0.5 mol L 1. After vigorous stirring at room temperature for 2 min, the suspension was transferred into a Teflon-lined stainless steel autoclave, into which a thermal couple and a pressure gauge were inserted to record the in situ temperature and pressure. The reactor was sealed and heated up to 100–180 1C for 0–600 min under mechanic stirring. To avoid the oxidation of Fe2+ (Mn2+),
18630 | Phys. Chem. Chem. Phys., 2015, 17, 18629--18637
PCCP
the mixing and hydrothermal treatment steps were operated under the protection of N2 atmosphere. After ambient cooling to room temperature, products were centrifuged and washed several times with deionized water. To achieve the carbon coating LiMn0.5Fe0.5PO4 composites, the hydrothermal products were mixed with 20 wt% of glucose, and then carbonized at 650 1C for 6 h under argon atmosphere with a heating rate of 5 1C min 1. For comparison, the LMFP was also prepared by the same process with resultant concentration of precursor (PO43 ) at 0.25 mol L 1 and 0.6 mol L 1 (the highest resultant concentration can be reached at room temperature). The LMFP prepared under resultant concentration of 0.25 mol L 1, 0.5 mol L 1 and 0.6 mol L 1 were denoted as LMFP-0.25, LMFP-0.5 and LMFP-0.6, and the carbon coated LMFP were denoted accordingly as LMFP-0.25/C, LMFP-0.5/C and LMFP-0.6/C, respectively.
2.2
Materials characterization
The crystalline phase of the as-prepared materials was identified by X-ray diffraction (XRD, D/max-rB, Rigaku, Japan) measurements using Cu-Ka radiation. The data for Rietveld refinement were collected in the 2y range of 5–1351 at a continuous scan mode with a step size of 0.021 for a counting time of 10 s. The GASA Rietveld program was used to analyze the diffraction patterns. Particle morphologies were observed by scanning electron microscopy (SEM, SPA400 Seiko Instruments) and fieldemission transmission electron microscopy (TEM, JEOL 2100F). The carbon content in composites was measured by a carbon-sulfur analyzer (CS-902, Wanlianda Xinke, Beijing, China).
2.3
Electrochemical measurements
The electrochemical performances of the as-prepared composite were evaluated using CR2032 coin-type cells. To prepare the working electrodes, the active materials, acetylene black, and poly(vinylidene fluoride) were mixed at a weight ratio of 80 : 13 : 7 in N-methyl pyrrolidone. The mixed viscous slurry was uniformly spread on an Al foil current collector with loading mass of active materials fixed about 5 mg cm 2. The electrodes dried under vacuum at 120 1C for 12 h. Subsequently, cells were assembled in glove box filled with argon using Li metal foil as anode and a polypropylene micro-porous film as separator. Then, 1 mol L 1 LiPF6 dissolved in a mixture of ethylene carbonate and dimethyl carbonate (1 : 1 by volume) was used as the electrolyte. The cells were tested in a voltage range of 2.5–4.5 V (vs. Li/Li+) using a constant-current protocol at various rates (1 C = 170 mA h g 1) on a battery testing system (Neware BTS-610). Cyclic voltammetry (CV) was performed on a LK9805 electrochemical analyser in the voltage range of 2.5–4.5 V at a scanning rate of 0.1 mV s 1. Electrochemical impedance spectroscopy (EIS) measurements were performed on a Zahner electrochemical workstation over the frequency range of 105 to 10 2 Hz with an applied amplitude of 5 mV. The EIS data for the three samples were collected at stable open circuit voltage (about 3.4 V). The Zview program was used to fit the EIS data.
This journal is © the Owner Societies 2015
View Article Online
PCCP
3. Results and discussion
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
3.1
Formation and evolution of LMFP
3.1.1 Transformation of precursor and LMFP during temperature increase. The composition and crystallographic structure transformation of the product were traced back by sampling the mixture at 30, 100, 110, 120, 130 and 140 1C during the temperature increase. Fig. 1a shows the XRD pattern of the asprepared Li3PO4, which can be indexed to orthorhombic Li3PO4 (JCPDS No.71-1528). The SEM image (Fig. 2a) shows that the aggregated quasi-spheres Li3PO4 consist of 30–50 nm nanocrystallites. The formation of quasi-sphere Li3PO4 could be explained by the difference solubility products of Li3PO4 (Ksp1 = 3.2 10 9) and Li2HPO4 (Ksp2 = 4.0 10 1).30 When the transition metal solution was introduced into the neutralized mixture and stirred at room temperature for 2 min, the XRD pattern of the sample (Fig. 1b) shows that the peaks can be only indexed as Li3PO4, and the intensity of peaks decrease with the adjunction of Fe2+ and Mn2+. SEM and TEM examinations (Fig. 2b and b-1) reveal that the precursors consist of quasispheres Li3PO4 and leaflet-like nanostructures. It indicates that the reaction is proceeded by a dissolution-reprecipitation mechanism.30,31 Owing to the Ksp value of transition metal phosphate (denoted as MexPO4) [1.0 10 36 for Fe3(PO4)2 and 1.0 10 27 for Mn3(PO4)2] is much lower than that of Li3PO4,30,32 the very minor amount of free PO43 available in solution, which is determined by the Ksp value of Li3PO4, are more preferred to combine with Fe2+ and Mn2+ than Li+ to form MexPO4 precipitation. With the depletion of the free PO43 , the concentration product of Li+ and PO43 becomes below the critical value of Ksp1, resulting in a portion dissolution of Li3PO4 and decrease of peaks intensity. From the viewpoint of thermodynamics, the cycles of Li3PO4 dissolution followed by MexPO4 precipitation would be continued until equilibrium between MexPO4 and
Fig. 1 XRD pattern of as-prepared Li3PO4 and the intermediates sampled at 30 1C (a), 100 1C (b), 110 1C (c), 120 1C (d) and 140 1C (e) during the temperature rising period.
This journal is © the Owner Societies 2015
Paper
Me2+ is reached. However, the dissolution-reprecipitation cycles would be limited by their kinetics at such low temperature and pressure.33 Meanwhile, the precipitation time may not be enough to obtain the thermodynamic equilibrium states at the concentration. Thus, only a part of MexPO4 is precipitated or a major amount of MexPO4 exists in amorphous form (Fig. 2b-1-1), resulting in the absence of typical diffraction peaks of MexPO4. The XRD pattern of the intermediate sampled at 100 1C (Fig. 1c) indicates that the peak intensity of Li3PO4 decrease along with the detection of (Mn,Fe)3(PO4)24H2O (metaswitzerite, JCPDS No. 20-0713), as well as small quantities of Fe3(PO4)28H2O (viviantie, JCPDS No. 30-0662). The SEM and TEM images (Fig. 2c and c-1) shows that the leaflet-like nanostructure changes into larger and thicker sheet-like crystallite with regular morphology, accompanying with the sacrifice of Li3PO4. It can be speculated that the crystallization and growth of thermodynamically preferred precipitates (MexPO4) are easily fulfilled at higher temperature for the acceleration of the dissolution-reprecipitation cycles. In addition, the simultaneous detection of metaswitzerite and viviantie imply that a complete solid solution [(Mn0.5Fe0.5)3(PO4)2xH2O] is not formed, which is also found by Oh et al. in synthesis of LiMn0.5Fe0.5PO4 via (Mn0.5Fe0.5)3(PO4)2 precursor.8 At reaction temperature of 110 1C (Fig. 1d), the intensity of the characteristic Bragg peaks corresponding to the Li3PO4, (Mn,Fe)3(PO4)24H2O and Fe3(PO4)28H2O precursor fall sharply and some weak peaks corresponding to LMFP appear. In Fig. 2d, the quasi-spheres Li3PO4 and sheet-like crystallite are disappeared with the formation of grain-like nanocrystallite. The detection of LMFP verifies that the critical temperature for the initial nucleation LMFP locates at the range of 100–110 1C. As the dissolution of (Mn,Fe)3(PO4)24H2O and Fe3(PO4)28H2O, the amount of dissolved ions (Fe2+, Mn2+and PO43 ) sacrificed from (Mn,Fe)3(PO4)24H2O or Fe3(PO4)2PO4 are sufficient to bond the Li+, thus providing high concentrations of nutrient for nucleation of LFMP.24,27,28 With the temperature further increasing to 120 1C (Fig. 1e), all the reflections are indexed as an orthorhombic olivine-type structure (JCPDS No. 42-0580), and no additional impurity phases are observed, indicating the complete formation of LMFP. The XRD patterns of intermediate sampled at 140 1C (Fig. 1f) show stronger LMFP diffraction peaks, suggesting the fast growth of the crystal. Fig. 2e and f show the SEM images of intermediates extracted at 120 1C and 140 1C, respectively. Although the change of the crystallite morphology with respect to the number of grain boundaries per particle appear to be negligible, the size of the crystallite grow larger as the reaction temperature rising (Fig. 2e, e-1, f and f-1). 3.1.2 Evolution of the LMFP crystal under 180 8C at different times during the soaking period. Fig. 3a shows the XRD patterns of the intermediate extracted during the soaking period. All the diffraction patterns could be indexed to an olivine-type LMFP with orthorhombic crystal structure (space group Pnma), indicating reaching pure phase at 180 1C with different time spots. With the extension of the soaking period, the intensity of the diffraction peaks slightly increases. It is worth noting that the transformation of the relative intensity of the diffraction peaks is quite different from the results on the hydrothermal synthesis of LFP reported by Ou et al.27 In their study, the relative intensity of (200)/(121) keeps
Phys. Chem. Chem. Phys., 2015, 17, 18629--18637 | 18631
View Article Online
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
Paper
PCCP
Fig. 2 SEM images of the as-prepared Li3PO4 microspheres (a) and the intermediate extracted at 30 1C (b), 100 1C (c), 110 1C (d), 120 1C (e) and 140 1C (f) during temperature rising. TEM and corresponding SAED patterns of the intermediate extracted at 30 1C (b-1, b-1-1) and 100 1C (c-1, c-1-1). The corresponding particle size distribution of the intermediate extracted at 110 1C (d-1), 120 1C (e-1) and 140 1C (f-1).
increasing throughout the soaking period. Contemporaneously, the crystal growth orientates along the (200) and/or the (121) plane, resulting formation of nanoplates with large bc plane. However, there is no obvious change in the relative intensity of three main diffraction peaks [(111)/(021), (121)/(200), (131)] in our results, implying no transformation of morphology with respect to preferential crystal orientation.34 Rietveld refinement on the XRD patterns were shown in the Fig. 3b, which shows a good fit between observed and calculated patterns. The lattice parameters of LMFP samples were refined and listed in Table S1 (ESI†). It is found that the unit cell volume of LMFP extracted at 0 min is 297.59 Å3, and decrease to 295.42 Å3 as the reaction time reached to 600 min. Normally, the unit cell contraction of hydrothermally synthesized phosphate is the results of the decrease of antisite defects with the increase of temperature or time.35 Since Li+ ion diffusion in phosphate occurs along onedimensional channels, the reduction of Fe on Li site improves lithium transport in that channel,36 enhancing the capacity and rate capability. Meanwhile, the decrease of cell volume suggests the increased crystallinity of intermediates.28 Fig. 4 presents the size and morphological evolution of the intermediates extracted at different time spots. As the reaction duration extends, the primary crystallite increase in size with a concurrent transformation of morphology. For example, the crystals of intermediate extracted at 0 min are nanoparticles (Fig. 4a).
18632 | Phys. Chem. Chem. Phys., 2015, 17, 18629--18637
With increasing reaction durations (i.e., 60 min in Fig. 4c), the crystals of nanostructure become a mixture of rods and nanoparticles. At the longest reaction duration (600 min in Fig. 4d), significantly larger plates coexist with some nanoparticles and rods. It is considered that the nonuniformity in crystal morphologies arises from the multiple nucleation events, which always give rise to crystals of various growth extents.24 The iso-oriented nanorods could attribute to the aggregation of the primary nanoparticles via oriented attachment (OA), which forms a single crystal upon fusion of the nanoparticles.31,37 During the further crystallization process, the primary nanoparticles continue joining up with the increased crystals through Ostwald ripening (OR). Driven by surface energy minimization, thermodynamically unstable primary nanoparticles dissolve and larger particles (nanorods) grow into plates by adsorbing monomers from the dissolved nanoparticles.24,31,38,39 The size and morphology transformation imply that the crystals undergo a distinct growth process during the soaking period. It is worthwhile to note that the evolution of the intermediate is largely determined by the resultant concentration of precursor (PO43 ). When the resultant concentration of precursor is decreased to 0.25 mol L 1 while keeping other parameters constant, the formation and transformation of the precursor are hindered. The XRD patterns of the intermediates (Fig. S2, ESI†) show that the precipitation of (Mn,Fe)3(PO4)24H2O and Fe3(PO4)28H2O are substituted by separate formation of Fe3(PO4)28H2O and
This journal is © the Owner Societies 2015
View Article Online
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
PCCP
Paper
Fig. 3 (a) XRD patterns of the intermediate extracted at 180 1C during soaking period, (b) Rietveld refinement plots of the XRD patterns of the intermediate.
Mn3(PO4)21.5H2O at different temperature. The sequential precipitation of Fe3(PO4)28H2O and Mn3(PO4)21.5H2O could be illustrated by the dissolution-reprecipitation mechanism. Because the value of solubility product of Fe3(PO4)2 is lower than that of Mn3(PO4)2, the dissolved PO43 are more preferred to combine with Fe2+ than Mn2+. Under low concentration, the effects induced by the difference of solubility product may become notable. Compared with the formation of Fe3(PO4)2, more amount of Li3PO4 is required to be dissolved for initiating the Mn3(PO4)2 precipitation. Meanwhile, the formation of phase pure LMFP is postponed to the temperature of 130 1C. Large quantity of Mn3(PO4)21.5H2O and Li3PO4 still exist in the precipitation at the temperature of 120 1C. Typically, the nucleation and crystal growth under hydrothermal condition are governed by the concentration of ions and the kinetic factor.28,35 The amount of dissolved ions at a relatively low concentration solution is insufficient, which results in the deficiency of crystal nucleus and the slowness of the crystal growth rate. Furthermore, the transportation of species to nucleation sites may be hindered in relatively low concentration, resulting the limited crystallization process.35 3.2
Fig. 4 SEM images and the corresponding particle size distribution of the intermediate extracted at 180 1C during soaking period: 0 min (a. a-1), 10 min (b, b-1), 60 min (c, c-1), 600 min (d, d-1).
be mainly expressed by a five-step process. Initially, quasi-spheres Li3PO4 are formed by neutralization reaction between H3PO4 and LiOH solution at room temperature. When the Fe2+ (Mn2+) solution is added into the mixture, the leaflet-like MexPO4 is formed with the coexistence of Li3PO4. At the low temperature and pressure, the dissolution-reprecipitation cycle cannot proceed completely, resulting in the formation of amorphous metal phosphate precipitation. In the subsequent hydrothermal process, the size and crystallinity of metal phosphate increase with the acceleration of dissolution–reprecipitation cycle at higher temperature.
Mechanism for nucleation and crystal growth
On the basis of the XRD, SEM and TEM results, accompanied with confirmation of dissolution-reprecipitation, a potential reaction mechanism of the LFMP nanostructure is proposed as schematically illustrated in Fig. 5. The reaction process can
This journal is © the Owner Societies 2015
Fig. 5 Schematic illustration of the growth mechanism involved in the hydrothermal synthesis of LiMn0.5Fe0.5PO4 nanostructure.
Phys. Chem. Chem. Phys., 2015, 17, 18629--18637 | 18633
View Article Online
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
Paper
PCCP
With temperature further increasing, unstable compounds in the precursor begin to dissolve and hydrolyze quickly due to the increase of solubility at high temperature, providing high concentration nutrient for nucleation and growth of LFMP. Finally, the primary particles evolved from LFMP nuclei would self-aggregate to form rods and plates to reduce the overall surface energy through oriented attachment (OA) and Ostwald ripening (OR) mechanism. According to the nucleation and growth mechanism proposed above, the resultant concentration of precursor would have distinct impact on the characteristic size of the crystals. To verify whether this is the case or not, we loaded three independent autoclaves with the resultant concentration at 0.25 mol L 1, 0.5 mol L 1 and 0.6 mol L 1 to synthesize LMFP. The XRD patterns of the three samples are shown in Fig. S2 (ESI†). Morphologies of the three samples shown in Fig. 6 confirm that the improved resultant concentration significantly reduces the particle size of the crystals. It is generally acknowledged that increasing nucleation sites could decrease crystal size of the precipitation in wet chemical process.25 As the nucleation of LMFP is based on the dissolution of Li3PO4 and Me3(PO4)2, improving the result concentration of precursor significantly increase the concentration of free ions (Li+, Mn2+, Fe2+ and PO43+) at the dissolution stage, resulting higher supersaturation degree for nucleation. Consequently, the accelerated nucleation
Fig. 6
rate would favor a great number of nuclei and facilitate smaller particles. 3.3
Electrochemical performance
In order to associate the LFMP structure with the performance, electrochemical measurements were conducted for crystalline synthesized under three concentrations. Fig. 7 shows the second cyclic voltammetry of LiMn0.5Fe0.5PO4/C after ‘‘activation’’ at a scan rate of 0.1 mV s 1 at room temperature. The CV curves show two oxidation peaks and two reduction peaks, corresponding to the charge/discharge reactions of the Fe2+/ Fe3+ and Mn2+/Mn3+ redox couple, respectively. It is noted that a small redox peak (B3.6 V) is observed in the CV, which corresponds to the small discharge plateau at B3.6 V as seen in the discharge curves at 1 C rate in Fig. 8b–d. This small peak was also found in earlier reports concerning of LiFe1 xMnxPO4 solid solution,40–42 which guess that the peak could belong to a Mn2P2O7 impurity phase that could not be detected by XRD.41 For LMFP-0.6/C, the voltage hysteresis (DV, difference between the anodic and cathodic peak voltages) of Mn3+/Mn2+ and Fe3+/ Fe2+ are 260 mV and 140 mV, respectively. By contrast, those hysteresis are 270 mV and 150 mV for LMFP-0.5/C, 340 mV and 240 mV for LMFP-0.25/C. Besides, the peak currents of LMFP0.6/C are also higher than those synthesized under lower concentration. It is well known that smaller voltage hysteresis
SEM, TEM and the corresponding HRTEM images of LMFP-0.25/C (a, d and g), LMFP-0.5/C (b, e and h) and LMFP-0.6/C (c, f and i).
18634 | Phys. Chem. Chem. Phys., 2015, 17, 18629--18637
This journal is © the Owner Societies 2015
View Article Online
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
PCCP
Fig. 7 The representative CV profiles of LMFP-0.6/C, LMFP-0.5/C and LMFP-0.25/C at a scan rate of 0.1 mV s 1.
Fig. 8 (a) Galvanstatic cycling profiles of LMFP-0.25/C, LMFP-0.5/C and LMFP-0.6/C at 0.1 C. (b) Discharge profiles of LMFP-0.25/C at different rates. (c) Discharge profiles of LMFP-0.5/C at different rates. (d) Discharge profiles of LMCP-0.6/C at different rates. (e) Rate capability of LMFP-0.25/C, LMFP-0.5/C and LMFP-0.6/C at different rates. (f) Cycling performance of LMFP-0.25/C, LMFP-0.5/C and LMFP-0.6/C at a discharge rate of 1 C for 100 cycles.
as well as higher peak currents indicate better electrode reaction kinetics and thus better electrochemical performance.21,43 Considering the similar amounts of carbon (2.71, 2.73 and 2.68 wt% for LMFP-0.25/C, LMFP-0.5/C and LMFP-0.6/C, respectively), the enhanced kinetics for LMFP-0.6/C electrode during the lithiation and delithiatoin can be mainly attributed to the crystal size reduction. The results of cyclic voltammetry suggest that the LMFP synthesized at higher resultant concentration of precursor would provide a better charge/discharge performance. The prediction was further confirmed by the charge/discharge
This journal is © the Owner Societies 2015
Paper
measurements. The initial charge/discharge profiles at current rates of 0.1 C are shown in Fig. 8a. The discharge curves of LiMn0.5Fe0.5PO4/C contain two flat voltage plateaus at 4.1 V and 3.5 V, associating with the Mn3+/Mn2+ and Fe3+/Fe2+ redox reactions, respectively. LMFP-0.6/C exhibits a better specific discharge capacity of 155.7 mA h g 1 (91.6% of the theoretical capacity) relative to LMFP-0.25/C and LMFP-0.5/C, which deliver 139.1 and 145.1 mA h g 1, respectively. The voltage profiles of LMFP-0.25/C, LMFP-0.5/C and LMFP-0.6/C at various charge/ discharge rates from C/10 to 20 C are shown in Fig. 8b–d. The fact that high voltage plateau of discharge profile is decreasing under high discharge current may originate from the increased polarization of Mn2+/Mn3+ couple.41 Compared with LMFP-0.25/C and LMFP-0.5/C, LMFP-0.6/C sample achieves the highest capacity at each C-rate, demonstrating the low internal resistance in solid-sate Li+ diffusion.21 Fig. 8e shows the rate capabilities of LMFP-0.25/C, LMFP0.5/C and LMFP-0.6/C. As expected, the LMFP-0.6/C exhibits the most excellent rate performance, with discharge capacities of 144, 127, 116, 106, 100, 91 and 81 mA h g 1 at 0.2, 0.5, 1, 3, 5, 10 and 20 C, respectively. However, the LMFP-0.25/C and LMFP0.5/C with large crystal sizes show aggravated degradation behavior and lower discharge capacities. For example, LMFP0.25/C and LMFP-0.5/C only deliver 59 and 77 mA h g 1 of discharge capacity respectively at 20 C. The cycling performance recorded after the rate performance is shown in Fig. 8f. At the first cycle, cells with LMFP-0.25/C, LMFP-0.5/C and LMFP-0.6/C can deliver discharge capacities of 87, 104, 109 mA h g 1 respectively, at 1 C. After cycling for 100 times, the discharge capacity retention of the cell with LMFP-0.6/C is 90.7% of the initial value, which is significantly outperformed compared with those of LMFP-0.25/C and LMFP-0.5/C, with only 88.4 and 87.0% respectively. It is reasonable to assume that the difference in rate performance and cycling performance are mainly attributed to the different particle sizes of these samples. For LMFP-0.6/C, the reduced crystal size favors charge transfer and Li ions diffusion, improving the yieldable capacity at higher rates. The reason for the difference in electrochemical performances of three samples under mutative reactant concentrations were further investigated by electrochemical impedance spectroscopic (EIS) measurements. Fig. 9a shows the Nyquist plots of the three samples collected with a two electrode coin cell in the discharged state after 100 charge–discharge cycles. Each of the curves has a depressed semicircle in the high-tomedium frequency and an oblique line in the low-frequency. The intercept of the curve in the high frequency region relates to the ohmic series resistance (Rs), which includes the interparticle contact resistance, electrolyte resistance and other physical resistance. The following depressed semicircle in the medium-frequency region corresponds to the charge transfer resistance (Rct) at the electrolyte-electrode interface. The straight line in the low-frequency region associated with the Warburg impedance (Zw), related to the diffusion behavior of lithium ions within particles. By comparing the diameters of the semicircles, the impedance of LMFP-0.6/C electrode is
Phys. Chem. Chem. Phys., 2015, 17, 18629--18637 | 18635
View Article Online
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
Paper
PCCP
Fig. 9 (a) Nyquist impedance spectra and equivalent circuit of LMFP0.25/C, LMFP-0.5/C and LMFP-0.6/C at room temperature, (b) the relationship between Zre and o 1/2 in the low frequency region.
significantly smaller than those of LMFP-0.5/C and LMFP-0.25/ C electrode. The values of Rct for LMFP-0.6/C, LMFP-0.5/C and LMFP-0.25/C electrodes were fitted by the given equivalent circuit and found to be 71.7 O, 268.2 O, and 554.3 O, respectively. The Li+ diffusion coefficient (DLi+) was calculated for these samples according to the following equation: DLi+ = (R2T2)/(2A2n4F4C4s2)
(1)
here R is the gas constant (8.314 J K 1 mol 1), T is the absolute temperature (298.15 K), A is the surface area of the electrode, n is the number of the electrons per molecule during oxidization (here, n = 1), F is the Faraday constant (96 500 C mol 1), C is the concentration of Li ions (7.69 10 3 mol cm 3) and s is the Warburg factor (O s 1/2) associated with Zre (the imaginary part of cell impedance, ohm) and o (the frequency, Hz), calculated by equation: Zre p so
1/2
(2)
Therefore the Warburg factor can be obtained from the slope between Zre and the inverse square root of the angular frequency, as shown in the Fig. 9b. Accordingly, the apparent Li ion diffusion coefficients of LMFP-0.6/C, LMFP-0.5/C and LMFP-0.25/C were determined to be 8.23 10 14, 2.59 10 14 and 6.0 10 15 cm2 s 1, respectively. Considering the similar amounts of carbon, the lowest Rct value and enhanced mobility of Li+ ions of LMFP-0.6/C are attributed to the reduced crystal size. The larger specific surface area and smaller crystal size of LMFP-0.6/C provide a higher electrolyte contact interface and shorter diffusion distance of Li+ ions, which lowers the charge transfer resistance and facilitates the electrochemical reaction, leading to enhanced electrochemical performance.
4. Conclusion In summary, we systematically investigated the phase and morphological evolution of the intermediate in the hydrothermal synthesis of LiMn0.5Fe0.5PO4. It shows that LiMn0.5Fe0.5PO4 can be formed at the temperature range as low as 100–110 1C. The formation of LMFP occurs by a straightforward dissolution of Li3PO4, Fe3(PO4)28H2O and (Mn,Fe)3(PO4)24H2O. As primary particles evolved from the nuclei of LMFP, the driven force induced by the reduction of overall surface energy enables the primary particles
18636 | Phys. Chem. Chem. Phys., 2015, 17, 18629--18637
to self-aggregate to form rods and plates through the known oriented attachment (OA) and Ostwald ripening (OR) mechanism. We also demonstrated that the resultant concentration of precursor plays a decisive role in determining the crystal size of LMFP by altering the nucleation rate. Electrochemical characterization of LMFP materials discloses a discharge capacity of 155 mA h g 1 at 0.1 C. Crystals synthesized at higher concentration deliver superior overall electrochemical performances compared with those synthesized at lower concentrations. In this study, the clear and unambiguous understanding of the formation mechanism provides a profound guiding significance for the synthesis of LMFP or other olivine-type cathode materials with shorter Li+ diffusion paths and improved performances.
Acknowledgements The authors appreciate the support of Sichuan University Funds for Young Scientists (No. 2011SCU11081), the PhD Programs Foundation of Ministry of Education of China (No. 20120181120103), the Project funded by China Postdoctoral Science Foundation (No. 2014M562322), AutoCRC Project 1-111 ‘‘Development of Advanced Electrode and Electrolytes for LIB’’.
Notes and references 1 C. Masquelier and L. Croguennec, Chem. Rev., 2013, 113, 6552–6591. 2 A. K. Padhi, K. S. Nanjundaswamy and J. B. Goodenough, J. Electrochem. Soc., 1997, 144, 1188–1194. 3 C. Sung-Yoon, J. T. Bloking and C. Yet-Ming, Nat. Mater., 2002, 1, 123–128. 4 Z. H. Chen and J. R. Dahn, J. Electrochem. Soc., 2002, 149, A1184–A1189. 5 L.-X. Yuan, Z.-H. Wang, W.-X. Zhang, X.-L. Hu, J.-T. Chen, Y.-H. Huang and J. B. Goodenough, Energy Environ. Sci., 2011, 4, 269–284. 6 M. K. Devaraju and I. Honma, Adv. Energy Mater., 2012, 2, 284–297. 7 J. Wang and X. Sun, Energy Environ. Sci., 2012, 5, 5163–5185. 8 S.-M. Oh, S.-T. Myung, Y. S. Choi, K. H. Oh and Y.-K. Sun, J. Mater. Chem., 2011, 21, 19368–19374. 9 C. Delacourt, P. Poizot, M. Morcrette, J. M. Tarascon and C. Masquelier, Chem. Mater., 2004, 16, 93–99. 10 S. K. Martha, J. Grinblat, O. Haik, E. Zinigrad, T. Drezen, J. H. Miners, I. Exnar, A. Kay, B. Markovsky and D. Aurbach, Angew. Chem., Int. Ed., 2009, 48, 8559–8563. 11 V. Aravindan, J. Gnanaraj, Y.-S. Lee and S. Madhavi, J. Mater. Chem. A, 2013, 1, 3518–3539. 12 W. Xiang, Y. Zhong, Y. Tang, H. Shen, E. Wang, H. Liu, B. Zhong and X. Guo, J. Alloys Compd., 2015, 635, 180–187. 13 C. Delacourt, L. Laffont, R. Bouchet, C. Wurm, J. B. Leriche, M. Morcrette, J. M. Tarascon and C. Masquelier, J. Electrochem. Soc., 2005, 152, A913–A921. 14 G. Chen, J. D. Wilcox and T. J. Richardson, Electrochem. Solid-State Lett., 2008, 11, A190–A194.
This journal is © the Owner Societies 2015
View Article Online
Published on 24 June 2015. Downloaded by University of Cambridge on 14/08/2015 02:51:49.
PCCP
15 D. Choi, J. Xiao, Y. J. Choi, J. S. Hardy, M. Vijayakumar, M. S. Bhuvaneswari, J. Liu, W. Xu, W. Wang, Z. Yang, G. L. Graff and J.-G. Zhang, Energy Environ. Sci., 2011, 4, 4560–4566. 16 Y.-K. Sun, S.-M. Oh, H.-K. Park and B. Scrosati, Adv. Mater., 2011, 23, 5050–5054. 17 S.-M. Oh, S.-T. Myung, J. B. Park, B. Scrosati, K. Amine and Y.-K. Sun, Angew. Chem., Int. Ed., 2012, 51, 1853–1856. 18 H. Wang, Y. Yang, Y. Liang, L.-F. Cui, H. S. Casalongue, Y. Li, G. Hong, Y. Cui and H. Dai, Angew. Chem., Int. Ed., 2011, 50, 7364–7368. 19 K. M. O. Jensen, M. Christensen, H. P. Gunnlaugsson, N. Lock, E. D. Bojesen, T. Proffen and B. B. Iversen, Chem. Mater., 2013, 25, 2282–2290. 20 G. R. Gardiner and M. S. Islam, Chem. Mater., 2010, 22, 1242–1248. 21 B. Ding, P. Xiao, G. Ji, Y. Ma, L. Lu and J. Y. Lee, ACS Appl. Mater. Interfaces, 2013, 5, 12120–12126. 22 P. Gibot, M. Casas-Cabanas, L. Laffont, S. Levasseur, P. Carlach, S. Hamelet, J.-M. Tarascon and C. Masquelier, Nat. Mater., 2008, 7, 741–747. 23 K. Saravanan, V. Ramar, P. Balaya and J. J. Vittal, J. Mater. Chem., 2011, 21, 14925–14935. 24 X. Qin, J. Wang, J. Xie, F. Li, L. Wen and X. Wang, Phys. Chem. Chem. Phys., 2012, 14, 2669–2677. 25 X. Qin, X. Wang, H. Xiang, J. Xie, J. Li and Y. Zhou, J. Phys. Chem. C, 2010, 114, 16806–16812. 26 K. Dokko, S. Koizumi, H. Nakano and K. Kanamura, J. Mater. Chem., 2007, 17, 4803–4810. 27 X. Ou, H. Gu, Y. Wu, J. Lu and Y. Zheng, Electrochim. Acta, 2013, 96, 230–236. 28 X. Ou, L. Pan, H. Gu, Y. Wu and J. Lu, J. Mater. Chem., 2012, 22, 25493.
This journal is © the Owner Societies 2015
Paper
29 P. F. Xiao, B. Ding, M. O. Lai and L. Lu, J. Electrochem. Soc., 2013, 160, A918–A926. 30 M.-H. Lee, J.-Y. Kim and H.-K. Song, Chem. Commun., 2010, 46, 6795–6797. 31 J. Zhu, J. Fiore, D. Li, N. M. Kinsinger, Q. Wang, E. DiMasi, J. Guo and D. Kisailus, Cryst. Growth Des., 2013, 13, 4659–4666. 32 T.-H. Kim, H.-S. Park, M.-H. Lee, S.-Y. Lee and H.-K. Song, J. Power Sources, 2012, 210, 1–6. 33 M.-H. Lee, T.-H. Kim, Y. S. Kim, J.-S. Park and H.-K. Song, J. Mater. Chem., 2012, 22, 8228–8234. 34 H. Guo, C. Wu, J. Xie, S. Zhang, G. Cao and X. Zhao, J. Mater. Chem. A, 2014, 2, 10581–10588. 35 J. Chen, J. Bai, H. Chen and J. Graetz, J. Phys. Chem. Lett., 2011, 2, 1874–1878. 36 L. Gu, C. Zhu, H. Li, Y. Yu, C. Li, S. Tsukimoto, J. Maier and Y. Ikuhara, J. Am. Chem. Soc., 2011, 133, 4661–4663. 37 H. Uchiyama and H. Imai, Cryst. Growth Des., 2010, 10, 1777–1781. 38 P. Barpanda, K. Djellab, N. Recham, M. Armand and J.-M. Tarascon, J. Mater. Chem., 2011, 21, 10143–10152. 39 G. Qin, S. Xue, Q. Ma and C. Wang, CrystEngComm, 2014, 16, 260–269. 40 X. Zhou, Y. Xie, Y. Deng, X. Qin and G. Chen, J. Mater. Chem. A, 2015, 3, 996–1004. 41 R. von Hagen, H. Lorrmann, K.-C. Moeller and S. Mathur, Adv. Energy Mater., 2012, 2, 553–559. 42 Y.-J. Zhong, J.-T. Li, Z.-G. Wu, X.-D. Guo, B.-H. Zhong and S.-G. Sun, J. Power Sources, 2013, 234, 217–222. 43 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.
Phys. Chem. Chem. Phys., 2015, 17, 18629--18637 | 18637
