MnO powder as anode active materials for lithium ion batteries

MnO powder as anode active materials for lithium ion batteries

Journal of Power Sources 195 (2010) 3300–3308 Contents lists available at ScienceDirect Journal of Power Sources journal homepage: www.elsevier.com/...

2MB Sizes 3 Downloads 93 Views

Journal of Power Sources 195 (2010) 3300–3308

Contents lists available at ScienceDirect

Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour

MnO powder as anode active materials for lithium ion batteries Kaifu Zhong 1 , Xin Xia 1 , Bin Zhang, Hong Li ∗ , Zhaoxiang Wang, Liquan Chen Institute of Physics, Chinese Academy of Sciences, Beijing 100190, China

a r t i c l e

i n f o

Article history: Received 23 July 2009 Received in revised form 25 November 2009 Accepted 27 November 2009 Available online 2 December 2009 Keywords: Manganese oxide anode Conversion reaction Solid electrolyte interphase

a b s t r a c t MnO powder materials are investigated as anode active materials for Li-ion batteries. Lithium is stored reversibly in MnO through conversion reaction and interfacial charging mechanism, according to the results of ex situ XRD, TEM and galvanostatic intermittent titration technique. A layer of the solid electrolyte interphase with a thickness of 20–60 nm is covered on MnO particles after full insertion. MnO powder materials show reversible capacity of 650 mAh g−1 with average charging voltage of 1.2 V. It can deliver 400 mAh g−1 at a rate of 400 mA g−1 . The cyclic performance of MnO is improved significantly after decreasing particle size and coating with a layer of carbon. Among observed transition metal oxides, MnO shows relatively lower voltage hysteresis (<0.7 V) between the discharging and the charging curves at 0.05 C. In addition to its environmental benign feature and high density (5.43 g cm−3 ), MnO seems a promising high capacity anode material for Li-ion batteries among transition metal oxides. However, the initial columbic efficiency is less than 65% and the voltage hysteresis is still too high. The origins of them are discussed. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Lithium ion batteries with high specific volumetric and gravimetric energy densities are in growing demand for advanced portable electronic devices and electric vehicles. Great efforts have been paid on increasing lithium storage capacities of anode and cathode materials. Besides the materials which contain active component to form alloy with lithium, such as Si or Sn, transition metal (TMX) compounds have also attracted wide attention as high capacity anode materials for Li-ion batteries. Poizot et al. [1] reported for the first time on 2000 that lithium can be stored reversibly in transition metal oxides (TMOs) through heterogeneous conversion reaction: Li + TMO ↔ Li2 O + TM (TM = Co, Fe, Ni, Cu).

(1)

Later, reversible lithium storage was also observed in transition metal fluorides [2–4], sulfides, nitrides and phosphides [5–8]. It is very interesting that very inert LiF or Li2 O can react with transion metal (TM) at room temperature. The enhanced electrochemical reactivity of LiF or Li2 O is mainly benefited from the nanocomposite microstructure where the converted LiX and TM phases show extremely small grain size (<5 nm) and large contact area [1,2,9,10].

∗ Corresponding author at: Institute of Physics, Chinese Academy of Sciences, Zhongguancun South 3rd Street No.8, Beijing 100190, China. Tel.: +86 10 82648067; fax: +86 10 82649046. E-mail address: [email protected] (H. Li). 1 These authors contribute equally to this work. 0378-7753/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jpowsour.2009.11.133

In order to increase the energy density of lithium ion batteries, anode materials should have low average charging (delithiation) voltage. Typical values of the average charging voltages of anode materials in lithium half cells at low rate are 0.3 V for graphite, 0.6 V for silicon, 0.7 V for Sn and 1.6 V for Li4 Ti5 O12 . In addition, small voltage hysteresis (the voltage difference between the charge and discharge curve) is preferable for batteries to achieve high energy efficiency. The anode materials, such as mesophase carbon microbeads (MCMB) and Li4 Ti5 O12 , show low voltage hysteresis less than 0.2 V. However, most of TMX materials show high voltage hysteresis about 1 V [11,12], due to slow kinetic processes in electrode reactions. It was noticed that a theoretical lithium storage capacity could be approached at room temperature only when the thermodynamic equilibrium voltage (or called electromotive force, emf) of the material for the conversion reaction is higher than 1 V [12]. This is due to high overpotential (∼1 V) for this type of lithium storage. Cr2 O3 and MnO are two transition metal oxides that stoichiometric lithium can be inserted after a full discharge (discharging to 0 V vs Li+ /Li) and their emf values (1.085 V and 1.032 V vs Li+ /Li respectively) are relatively lower [5,12]. In view of energy density, they are more suitable as anode materials for Li-ion batteries among TMO materials. Cr2 O3 has been studied comprehensively [11–17]. Unfortunately, the voltage hysteresis of the Cr2 O3 electrode at 0.1 C is about 0.7–1.7 V in the full voltage range [11] and Cr2 O3 is a toxic chemical to human. As for MnO, an early report by Poizot et al. [5] indicates that the MnO powder electrode could be discharged and charged under a very small current density (1/300 C) by apotentialstatic

K. Zhong et al. / Journal of Power Sources 195 (2010) 3300–3308

3301

intermittent titration technique (PITT) mode. Recently, we have demonstrated that the MnO thin film electrode shows high volumetric capacity (>3484 mAh cm−3 at 0.125 C), excellent rate performance (55% capacity retention at 6 C) and relatively low polarization (<0.7 V at 0.125 C) [18]. This new result means that the kinetic property of the MnO anode is not as poor as early investigation. In this paper, electrochemical reaction mechanism, kinetic and cyclic performances of the MnO powder electrodes in lithium batteries are reported.

2. Experimental Four powder samples were used for the study. Commercial MnO (Alfa Aesar, 99%, 200 meshes) in dark green color was used directly and named as MnO-L. The MnO-L powder was mechanically milled at a speed of 3000 r min−1 for 5 h by a mill (Union Process 01ADDM) to decrease particle size. Milled powder was named as MnO-S. Both the MnO-L and the MnO-S samples were milled with sugar further at a weight ratio of 5:1 at a speed of 300 r min−1 for 1 h by a planet mill (WL -IA , Tianjin, China). After that, they were sintered at 600 ◦ C for 6 h under argon atmosphere for pyrolyzing sugar. The carbon-contained samples were named as C/MnO-L and C/MnO-S respectively (5 wt% carbon was remained). The test electrode was composed of active material (82.5 wt%), carbon black (7.5 wt%) and polyvinylidene fluoride (PVDF) (10 wt%). Cu foil was used as the current collector and Li foil was used as the counter electrode and 1 M LiPF6 dissolved in a mixture of ethylene carbonate (EC) and dimethyl carbonate (DMC) (1:1 by volume) (Shanghai Topsol Ltd., H2 O <10 ppm) was used as the electrolyte. A Swagelok-type two-electrode cell was constructed for electrochemical test. The cell was assembled in an argon-filled glove box and cycled between 10 mV and 3 V using a Land automatic battery tester. Galvanostatic intermittent titration technique (GITT) experiment was performed using an automatic battery tester (Arbin Instruments, Tianjin Co., Ltd.) by charging/discharging the cell for 4.6 h at a rate of 35.5 mA g−1 (about C/20 of theoretical capacity)

Fig. 1. XRD patterns of MnO samples: (a) MnO-L, (b) MnO-S, (c) C/MnO-L, and (d) C/MnO-S.

for the first discharging and 3.2 h for the first charging and the second cycle. Since it is not clear how long the electrode approaches the equilibrium state, the relaxation period was set for different hours (2 h, 4 h, 8 h, 16 h, 32 h and 64 h) respectively in each GITT experimental. The material was characterized by a X’Pert Pro MPD Xray diffractometer (Philips, Holland) using Cu K␣1 radiation ( = 1.5405 Å), a scanning electron (SEM) microscope (XL 30 S-FEG, FEI Co., 10 kV) and a transmission electron (TEM) microscope (JEOL 2010, 200 kV). The ex situ XRD experimental was investigated by the X’Pert Pro MPD X-ray diffractometer. Before ex situ XRD test, the reacted electrode was taken out from the cell in the glove box and washed by anhydrous DMC. The electrode sheet was pasted on a special sample holder and covered by a layer of Mylar film and sealed in a container before XRD test. The procedure for TEM investigation on lithiated samples was similar as our previous report [15].

Fig. 2. SEM images of MnO samples: (a) MnO-L, (b) MnO-S, (c) C/MnO-L, and (d) C/MnO-S.

3302

K. Zhong et al. / Journal of Power Sources 195 (2010) 3300–3308

Fig. 3. GITT curves of the C/MnO-L electrode. (a and b) 8 h relaxation time, (c and d) 64 h relaxation time, Solid lines are GITT results and dash lines are open circuit voltage. Dot dash line is the emf value for reference.

3. Results and discussion 3.1. Sample characterization XRD patterns of four samples are shown in Fig. 1. All peaks for all samples can be indexed as cubic phase of MnO (JCPDS, 78-0424) without the presence of impurity. No preferential orientation is founded for all samples. Grain sizes of four samples are estimated roughly by Scherrer equation using Jade software. Grain size of MnO-L, MnO-S, C/MnO-L and C/MnO-S is 72 nm, 19 nm, 30 nm and 34 nm respectively. The grain size variation of the carbon-coated samples should be a result of milling and sintering. The MnO-L sample has a particle size of 50–70 ␮m as shown in Fig. 2a. Each particle is composed of small particles of 0.5–1.5 ␮m (see inset zoomed image). After milling for 5 h, initial large particles transform into sheet-like small particles. The sheet thickness is about 80–120 nm and the size is about 0.5–1 ␮m, as shown in Fig. 1b. The C/MnO-L and the C/MnO-S samples show particle size of 1–2 ␮m (Fig. 2c and d). The particle size of the C/MnO-L sample is decreased compared with the MnO-L sample due to milling effect. The particle size of the C/MnO-S sample is increased compared with MnO-S due to sintering effect during carbon coating treatment. 3.2. Electrochemical reaction mechanism of MnO with lithium It is expected that MnO will react with lithium through a conversion reaction: 2Li + MnO → Li2 O + Mn.

(2)

Theoretical capacity of lithium storage in MnO is 755 mAh g−1 for a complete conversion reaction. The voltage profile for this reaction should be a plateau at the emf value of 1.032 V vs Li+ /Li at 298 K if the reactant and the product are perfect bulk materials [5,12]. GITT technique has been used widely to study the thermodynamic and kinetic properties of the electrode material [19]. GITT

curves of the C/MnO-L electrode with relaxation periods at 8 h and 64 h are shown in Fig. 3a–d respectively. It can be seen that there is no much difference in the discharge and the charge capacity between two GITT tests. A “plateau” at a voltage range of 1.0–0.4 V with a capacity of about 900 mAh g−1 and a slope with a capacity of 200 mAh g−1 can be seen in the OCV curve at the first discharging for the C/MnO-L sample in Fig. 3c. In the first charging, the capacity of the sloped region below the plateau region is 200 mAh g−1 . Since decomposition of SEI on anode normally occurs at higher voltage than 1.2 V and the reverse conversion reaction from the Li2 O/Mn to MnO should occur above the emf value of 1.032 V vs Li+ /Li, reversible lithium storage in the sloped region below 1.2 V should be caused by an interfacial charging mechanism, where lithium ions and electrons are stored in the grain boundary regions between Li2 O and Mn grains. Similar phenomenon has been observed nearly in all TMX anodes for conversion reactions. This mechanism has been explained by Maier et al. [2,10,12,20] in details. The total Li-storage capacity at the first discharge is 1100 mAh g−1 . The conversion reaction and the interfacial charging account at most 955 mAh g−1 . Rest of capacity about 145 mAh g−1 should be related to the formation of the solid electrolyte interphase (SEI). The irreversible capacity loss at the first cycle is 400 mAh g−1 . It means that both conversion reaction and the SEI formation are irreversible partially. In addition, it can be seen clearly that the OCV curve at the second discharge is 0.1–0.4 V higher than that at the first discharging. Such a voltage difference could be related to the microstructure variation. The origin has been discussed in the case of RuO2 by Delmer et al. [21] in terms of surface energy and amorphorization effect. Structure variation of the MnO-L electrode after lithiation and delithiation is shown in Fig. 4. Due to the Mylar film protection and short time measurement, the variation of the electrode phase during measuring can be ignored. The peaks from the MnO phase can still be seen when 2 Li are inserted in the plateau region. They disap-

K. Zhong et al. / Journal of Power Sources 195 (2010) 3300–3308

Fig. 4. Ex situ XRD patterns of the MnO-L electrode: (a) MnO-L powder, (b) MnOL electrode, (c) 2 Li inserted, (d) 3.2 Li inserted, (e) 1.1 Li extracted, and (f) 1.8 Li extracted.

pear completely when the electrode was discharged to 0 V while 3.2 Li are inserted. The peaks of MnO appear again after 1.1 Li and 1.8 Li are extracted. It is noticed that the grain size of the converted MnO is about 18 nm (estimated from Scherrer equation) after extraction of 1.8 Li. According to the TEM observation, the grain size of the restored TMX phase is 10 nm for TiF3 [2], 5–10 nm for CuO [9], 5 nm for RuO2 [10] and 5–10 nm for Cr2 O3 [15]. In previous reports, ex situ or in situ XRD patterns for TMX anodes at delithiated states do not show any peaks from the initial TMX phase [9,10]. The grain size of the converted TMX phase is determined by the transport kinetics of either Xn− anion or Mn+ cation in the LiX/M nanocomposite or across the intermediate MX phase. This result implies weakly that the Li2 O/Mn nanocomposite may have better kinetic property for reverse conversion reaction compared to some other TMO systems. In order to get a closer insight into structural transformation, we performed high-resolution transmission electron microscopy (HRTEM) and selected area electron diffraction (SAED) on the MnOS electrode at different Li insertion/extraction levels. The initial MnO-S sample shows polycrystalline feature with grain size around 10–20 nm (Fig. 5). Ordered structure extends to the edge region. The stripe distance is 0.255 nm, closed to d-value of (1 1 1) plane of MnO (0.2564 nm, JCPDS 78-0424). After the insertion of 2 Li, all particles are covered by a layer of amorphous species with a thickness of 20–40 nm in Fig. 6a. This

3303

layer is not covered uniformly. The existence of the surface layer has been observed in many lithiated TMX anodes and represents the so-called SEI [1,9,10,13]. In zoomed image in Fig. 6b, the thickness of the SEI layer is 7–12 nm in this region. The SAED of this sample revealed a set of rings, which can be indexed as MnO mainly, see the pattern converted to Cu K␣ wavelength in Fig. 9. As shown in the SAED pattern, sharp diffraction rings become more diffuse in this state, indicating the decrease of grain size. Obvious disintegration of the MnO grains is observed clearly in Fig. 6b. Some tiny grains less than 5 nm are dispersed in amorphous regions. The stripe distance in tiny grains is 0.227 nm, it can be either assigned to (2 2 0) plane of MnO with a d-value of 0.222 nm or (2 2 0) plane of ␤-Mn (33-0887) with a d-value of 0.223 nm. Such microstructure is typical and has been observed in other images for the same state (not shown here). After full lithiation of 3.2 Li, the thickness of the SEI is about 20–60 nm as shown in Fig. 7a. The SAED rings become more diffuse and weak, indicating further amorphization. The converted pattern in Fig. 9 can still be indexed to MnO, although this phase cannot be distinguished in the XRD pattern in Fig. 4. It is not clear whether it is related to trace amount of unreacted MnO or due to short exposure to air during the sample transfer for TEM investigation. In the zoomed image in Fig. 7b, the SEI thickness is 19–30 nm in this region. Most of interior regions are amorphous (also in several other images for the same state, not shown here) and a few tiny grains (2–5 nm) can be seen. The stripe distance for these grains is 0.223 nm. As mentioned above, it can be assigned to MnO or Mn. After charging up to 3.0 V, the thickness of the low contrast in the edge region is still about 20–60 nm but decreases to 10 nm in a few of areas as seen in Fig. 8a. The SAED pattern becomes sharper compared to the fully lithiated sample. It can be indexed as MnO phase, see Fig. 9. In zoomed image in Fig. 8b, the thickness of the SEI layer is 8–11 nm in this region. The stripe distance of the interior ordered region is 0.258 nm. The grain size is about 8–10 nm. A large ordered region (30 nm) with stripe distance of 0.260 nm can also be seen in the interior region (image not shown here). It explains why the MnO phase can be observed at the charging state in ex situ XRD pattern in Fig. 4. Above TEM results suggest that the initial MnO with large grain size is decomposed into nanocomposite after full lithiation and the MnO phase is restored after partial delithiation. The SEI film is very thick when 2 Li is inserted and grows further after more lithium is inserted. It is decomposed slightly after charging to 3.0 V. The existence of the thick SEI film at fully charging state explains at least partially the origin of the large irreversible capacity loss at

Fig. 5. TEM images and SAED pattern of the initial MnO-S sample. SAED pattern is indexed to cubic MnO (78-0425).

3304

K. Zhong et al. / Journal of Power Sources 195 (2010) 3300–3308

Fig. 6. TEM images and SAED pattern of the MnO-S electrode after 2 Li insertion.

Fig. 7. TEM images and SAED pattern of the MnO-S electrode after 3.2 Li insertion.

Fig. 8. TEM images and SAED pattern of the MnO-S electrode after 1.8 Li extraction.

K. Zhong et al. / Journal of Power Sources 195 (2010) 3300–3308

Fig. 9. Selected electron diffraction patterns of the MnO-S electrodes at different lithiated states (a–d corresponding to the SAED patterns in Figs. 5–8 respectively) converted to the patterns with Cu K␣ wavelength by a Process Diffraction program [24].

the first cycle for the MnO anodes. The thick SEI films on the Cr2 O3 anodes have been identified as oligomer, lithium alkyl carbonate and polyethylene oxide species [22,23]. It has been confirmed that the oligomer and polyethylene oxide species (totally 8 wt% of electrode materials) formed at fully discharging state are decomposed significantly after charging to 3.0 V [15,23]. Relatively stable SEI film on the MnO anodes could be composed of different SEI components as Cr2 O3 , needing further identification. 3.3. Kinetic and cyclic performances GITT results of the C/MnO-L electrode shown in Fig. 3 also provide kinetic information. It can be seen that there is a voltage gap about 0.3–0.4 V in the OCV curves between the first charging and

3305

the second discharging even after 64 h relaxation. Recently, Cedar et al. [25] proposed an explanation for the voltage hysteresis of the FeF3 electrode at 1/200 C. The first principle calculations indicate that the reaction paths of lithium insertion and extraction could be different under non-equilibrium states by selecting kinetic favorable intermediate products Lix Fey F3 . In the case of MnO, it seems that no any intermediate Lix MnO compounds could exist. It is still believed that the appearance of this large voltage hysteresis in the OCV curves of the MnO electrode is simply related to very sluggish phase transition reaction. Fig. 10 shows the OCV change of the electrode at different lithiated states for different relaxation periods. It can be seen that the OCV in most of discharging states and in the end of charging states is not stable even after 64 h relaxation. This indicates that the rate of the conversion reaction is significantly slow. For comparison, the OCV curve of the LiFePO4 electrode is stable after 20 h relaxation [26]. It is also seen in Fig. 10 that the OCV can be stable in most of charging states after 16 h relaxation. It seems that the rate for lithium extraction is easier than that for lithium insertion. This may be related to asymmetrical structure variations. For a spherical MnO particle, the shell layer of MnO phase is converted into the Li2 O/Mn phase during lithium insertion. It leads to a 170% volume expansion at least. The lithium extraction will occur in a reverse situation. Therefore, it might be more difficult for volume expansion than volume contraction in view of kinetics, especially for the first cycle. Another phenomenon is noticed in Fig. 3 that the voltage value at the plateau region at the first charging and the second discharging is 0.4–0.8 V higher than that at the first discharging. This could be partially related to improved kinetics, partially related to the deviation of the formation energy of the reactants and the products from perfect bulk materials after the first insertion [21]. Both are related to the formation of nanostructured products. Similar phenomenon has been observed in the GITT curves and galvanostatic discharging/charging curves of nearly all TMX anodes.

Fig. 10. OCV change with different relaxation times (2 h, 4 h, 8 h, 16 h, 32 h, and 64 h) at different normalized lithiated states of the C/MnO-L electrode (denoted D or C means discharge and charge, number means charge/discharge level, 1 = full discharge or charge). (a) First discharging, (b) second discharging, (c) first charging, and (d) second charging.

3306

K. Zhong et al. / Journal of Power Sources 195 (2010) 3300–3308

Fig. 11. Charging and discharging curves of the MnO powder electrodes at a rate of 50 mA g−1 . (a) MnO-L, (b) MnO-S, (c) C/MnO-L, and (d) C/MnO-S.

Charging and discharging curves of four samples are compared in Fig. 11. The voltage profiles of four samples are similar. The average charging voltage for MnO anodes is 1.2 V, similar as 1.2 V for Cr2 O3 , much lower than 1.8 V for CoO [1], 2.0 V for Co3 O4 [27], 1.6 V for FeO [1], 1.7 V for Fe3 O4 [28–29], 1.9 V for NiO [1], 2.1 V for CuO [9] and 2.4 V for RuO2 [10]. It should be mentioned, due to significant contribution of the reversible capacity at the sloped region at low voltage range, the average charging voltage in different materials is depressed to different levels compared to their emf values for conversion reaction. The reversible capacities for four samples at the second cycle are about 600 mAh g−1 . The C/MnO-S sample shows the highest initial columbic efficiency of 65.3%, which is lower than 70% of the MnO thin film electrode [18], less than 75% of many other TMO compounds, much less than 98% for RuO2 [10]. As shown in Figs. 3 and 5–8, significant capacity loss for the MnO anode could be related to the irreversible conversion reaction and the formation of the SEI. It should be mentioned that the content of carbon in the composite is 5 wt%. The hard carbon material from sugar pyrolysis showed a reversible capacity of 200–300 mAh g−1 between 10 mV and 3 V (450–600 mAh g−1 between −20 mV and 3 V). Therefore, carbon in the C/MnO composite will contribute 10–15 mAh g−1 to total reversible capacity of 600–650 mAh g−1 . It is obvious that decreasing particle size and carbon coating are effective to improve the cyclic performance of MnO electrode. Fig. 12 shows an excellent cyclic performance of the C/MnO-L electrode in a half lithium cell. After 90 cycles, the capacity is increased gradually. This phenomenon has also been reported in the case of Cu2 O anode by Chen et al. [30]. The origin was suggested to be related to either the formation of the SEI film or forming high oxidation state products, needing further clarification. As shown in Figs. 3 and 11, the MnO powder electrodes show large voltage hysteresis. It is related to high overpotential. Overpotential () values for lithiation and delithiation can be roughly estimated from the GITT result by the difference between the cutoff voltage and the open circuit voltage measured after relaxation.

It can be seen from the Curves (1) and (2) in Fig. 13 that the overpotential of the C/MnO-L electrode is less than 0.4 V. The overpotential of the C/MnO-L electrode can also be estimated roughly by the voltage hysteresis (V) curves as shown in the Curve (4). The V curve is obtained by subtracting the discharging curve at the second cycle from the charging curve at the first cycle after normalization. It is the sum of the overpotentials of both charging and discharging at galvanostatic mode. As shown in the Curve (4), V of the C/MnO-L electrode is around 0.7 V nearly in all lithiation/delithiation range. The Curve (3) is the sum of the Curve (1) and the Curve (2), drawn for comparison. The slight difference between the Curve (3) and the Curve (4) should be caused by incomplete relaxation in the OCV measurement. It will be similar if the voltage gap of the OCV curves between the first charging and the second discharging in Fig. 3 is accounted. Compared the V curves of four samples, decreasing the particle size and carbon coating is not effective on decreas-

Fig. 12. Cyclic performance and coulombic efficiency of the C/MnO-L electrode.

K. Zhong et al. / Journal of Power Sources 195 (2010) 3300–3308

Fig. 13. Voltage polarization analysis of the MnO powder electrodes at a rate of 50 mA g−1 (∼0.08 C). Curve (1): Voltage difference between the cut-off voltage and the open circuit voltage for the second discharging in the GITT curve (64 h interval) for the C/MnO-L electrode (Fig. 3), Curve (2): Voltage difference between the cut-off voltage and the open circuit voltage for the first charging in the GITT curve for the C/MnO-L electrode. Curve (3): Sum of the Curve (1) and the Curve (2). Solid line: Curves (4)–(7), V curves of MnO-L, MnO-S, C/MnO-L, C-MnO-S electrode respectively. Dot line: Curve (8), V curve of the MnO thin film electrode (130 nm thick, 0.125 C) for comparison [18]. Capacity normalization here in Curve (2): 0.0 refers to full delithiation state charged to 3.0 V, 1.0 means starting of charging.

ing voltage polarization. The voltage hysteresis values of the MnO powder samples are quite closed to the MnO thin film electrode at 0.125 C [18], as shown in the Curve (8). As also observed in the case of Cr2 O3 , the voltage hysteresis values of the thin film electrode and powder electrodes are quite close and not influenced strongly by the morphology and the existence of conductive additive. This is probable that the phase transition kinetics is the rate-determining step (rds) in conversion reaction. Therefore, V curves could be used to compare the kinetic property of different TMX compounds roughly. The V curves of several materials are shown in Fig. 14. Among oxides, RuO2 has the highest electronic conductivity, however, it shows higher V value. Converted nitrides should have higher lithium ion conductivity since Li3 N has very high lithium ion conductivity at 25 ◦ C (1.2 × 10−3 S cm−1 perpendicular to c-axis, 1 × 10−5 S cm−1 along c-axis and 7 × 10−4 S cm−1 for polycrystalline) [31,32]. However, CrN shows much larger polarization than

Fig. 14. V curves of TMX powder electrodes at low rate of 0.05–0.2 C. Solid lines: CrN, RuO2 , Co3 O4 , and TiF3 , MnO-L. Dash lines: Cr2 O3 , Fe2 O3 , and Fe3 O4 . Rate for each material: CrN: 0.1 C, Cr2 O3 : 0.08 C, RuO2 : 0.1 C, Co3 O4 : 0.1 C, Fe2 O3 : 0.1 C, TiF3 : 0.08 C, MnO-L: 0.08 C, and Fe3 O4 : 0.2 C [29].

3307

Fig. 15. Rate performances of the C/MnO-L and the C/MnO-S electrode

Cr2 O3 (also see CrN thin film voltage profile in Ref. [33]). Polarizations for different oxides are very different and the C/MnO-L and the Fe3 O4 nanospindles [29] show relatively lower values. These preliminary results indicate that the transport of both electronic and lithium ions may not be the rate-determining step (rds) for conversion reaction. Since the conversion reaction is a typical solid state reaction from two phases, the transport of anion Xn− from LiX phase or Mn+ from M phase across converted MX intermediate phase has to occur, which could be very sluggish at room temperature and becomes the rate-determining step. A recent report on MgH2 anode occurring conversion reaction in lithium ion batteries indicates that the V for MgH2 electrode is as low as 0.3 V at 0.05 C at room temperature [34]. This consists with above discussion since it is plausible that the transport of H could be much faster than other Xn− or Mn+ . Further understanding on the origins of the voltage hysteresis for TMX anodes needs accurate kinetic and structure data. The rate performances of the C/MnO-L electrode and the C/MnOS electrode are shown in Fig. 15. For the MnO thin film electrode (130 nm), 55% value of the capacity at 0.125 C rate can be retained at a rate of 6 C [18]. For the MnO powder electrodes after carbon coating, 1/3 capacity at a rate of 0.08 C can be achieved at a rate of 2.7 C. The rate performance of the MnO powder electrodes shown here is not superior to spherical graphite (MCMB) or Li4 Ti5 O12 anode, but is a significant improvement compared to previous report. It could be improved further by decreas-

Fig. 16. Relationship between the voltage hysteresis and the current density. V is taken from the middle of the voltage plateau in conversion reaction region.

3308

K. Zhong et al. / Journal of Power Sources 195 (2010) 3300–3308

ing particle size or improving carbon coating effect or using 3D current collectors, as nicely demonstrated in the case of Fe3 O4 [28,29]. The voltage hysteresis values in the middle of the conversion reaction region at different current densities are also drawn in Fig. 16. The linear response indicates that the electrode polarization obeys traditional ohmic rule in this current density range. It can be also seen that the voltage hysteresis is not zero when the current density is closed to zero. This phenomenon needs further. 4. Conclusions Li-storage mechanism in MnO powder electrode is determined as conversion reaction. During lithiation, initial MnO grains in polycrystalline particles decompose gradually into the nanocomposite composed of tiny Mn grains (<5 nm) and amorphous Li2 O. The particles are covered by a layer of 20–60 nm SEI film gradually. Upon charging, MnO nanograins (10–30 nm) are formed and the SEI film is decomposed slightly. Decreasing particle size and carbon coating is effective to improve the cyclic performance, but not very effective to decrease polarization. Carbon-coated MnO powder electrodes can deliver 400 mAh g−1 at a rate of 400 mA g−1 . In view of high capacity, good cyclic performance and good rate performance, low average charging voltage and environment benign feature, MnO material is perhaps the best candidate among transition metal oxides as high capacity anode material for Li-ion batteries. However, low initial columbic efficiency and large polarization need significant improvement. Acknowledgements This work was supported by NSFC (50730005), CAS (KJCX2-YWW26), “863” project (2006AA03Z228, 2009AA033101) and “973” project (2007CB936501). The authors thank Dr. Yanyan Liu in this lab for the help of SEM investigation, Dr. Binkun Guo for proving charging data of Fe2 O3 and Co3 O4 , Dr. P. Poizot in CNRS in Amiens for valuable discussion on conversion reaction, Prof. Y.G. Guo in CAS in Beijing for proving charging data of Fe3 O4 anode for overpotential analysis.

References [1] P. Poizot, S. Laruelle, S. Grugeon, L. Dupont, J.M. Tarascon, Nature 407 (2000) 496–499. [2] H. Li, G. Richter, J. Maier, Adv. Mater. 15 (2003) 736–739. [3] F. Badway, N. Pereira, F. Cosandey, G.G. Amatucci, J. Electrochem. Soc. 150 (2003) A1209–A1218. [4] F. Badway, F. Cosandey, N. Pereira, G.G. Amatucci, J. Electrochem. Soc. 150 (2003) A1318–A1327. [5] P. Poizot, S. Laruelle, S. Grugeon, L. Dupont, J.M. Tarascon, J. Electrochem. Soc. 149 (2002) A1212–A1217. [6] N. Pereira, L.C. Klein, G.G. Amatucci, J. Electrochem. Soc. 149 (2002) A262–A271. [7] D.C.C. Silva, O. Crosnier, G. Ouvrard, J. Greedan, A. Safa-Sefat, L.F. Nazar, Electrochem. Solid-State Lett. 6 (2003) A162–A165. [8] Y. Wang, Z.W. Fu, X.L. Yue, Q.Z. Qin, J. Electrochem. Soc. 151 (2004) E162–E167. [9] A. Débart, L. Dupont, P. Poizot, J.B. Leriche, J.M. Tarascon, J. Electrochem. Soc. 148 (2001) A1266–A1274. [10] P. Balaya, H. Li, L. Kienle, J. Maier, Adv. Funct. Mater. 13 (2003) 621–625. [11] J.P. Sun, K. Tang, X.Q. Yu, J. Hu, H. Li, X.J. Huang, Solid State Ionics 179 (2008) 2390–2395. [12] H. Li, P. Balaya, J. Maier, J. Electrochem. Soc. 151 (2004) A1878–A1885. [13] J. Hu, H. Li, X.J. Huang, Electrochem. Solid State Lett. 8 (2005) A66–A69. [14] S. Grugeon, S. Laruelle, L. Dupont, F. Chevallier, P.L. Taberna, P. Simon, L. Gireaud, S. Lascaud, E. Vidal, B. Yrieix, J.M. Tarascon, Chem. Mater. 17 (2005) 5041–5047. [15] J. Hu, H. Li, X.J. Huang, L.Q. Chen, Solid State Ionics 177 (2006) 2791–2799. [16] L. Dupont, S. Grugeon, S. Laruelle, J.M. Tarascon, J. Power Sources 164 (2007) 839–848. [17] L. Dupont, S. Laruelle, S. Grugeon, C. Dickinson, W. Zhou, J.M. Tarascon, J. Power Sources 175 (2008) 502–509. [18] X.Q. Yu, Y. He, J.P. Sun, K. Tang, H. Li, L.Q. Chen, X.J. Huang, Electrochem. Commun. 11 (2009) 791–794. [19] W. Weppner, R.A. Huggins, J. Electrochem. Soc. 124 (1977) 1569–1578. [20] J. Jamnik, J. Maier, Phys. Chem. Chem. Phys. 5 (2003) 5215–5220. [21] O. Delmer, P. Balaya, L. Kienle, J. Maier, Adv. Mater. 20 (2008) 501–505. [22] G. Gachot, S. Grugeon, M. Armand, S. Pilard, P. Guenot, J.M. Tarascon, S. Laruelle, J. Power Sources 178 (2008) 409–421. [23] Y. Zeng, L.F. Li, H. Li, X.J. Huang, L.Q. Chen, Ionics 15 (2009) 91–96. [24] J.L. La ba r, Ultramicroscopy 103 (2005) 237–249. [25] R.E. Doe, K.A. Persson, Y.S. Meng, G. Ceder, Chem. Mater. 20 (2008) 5274–5283. [26] N. Meethong, Y.H. Kao, M. Tang, H.Y. Huang, W.C. Carter, Y.M. Chiang, Chem. Mater. 20 (2008) 6189–6198. [27] P. Poizot, S. Laruelle, S. Grugeon, L. Dupont, J.M. Tarascon, J. Electrochem. Soc. 149 (2002) A234–A241. [28] P.L. Taberna, S. Mitra, P. Poizot, P. Simon, J.M. Tarascon, Nature Mater. 5 (2006) 567–573. [29] W.M. Zhang, X.L. Wu, J.S. Hu, Y.G. Guo, L.J. Wan, Adv. Funct. Mater. 18 (2008) 3941–3946. [30] C.H. Chen, N. Ding, L. Wang, Y. Yu, I. Lieberwirth, J. Power Sources 189 (2009) 552–556. [31] B.A. Boukamp, R.A. Huggins, Phys. Lett. A 58 (1976) 231–233. [32] A. Rabenau, Solid State Ionics 6 (1982) 277–293. [33] Q. Sun, Z.W. Fu, Electrochem, Solid State Lett. 10 (2007) A189–A193. [34] Y. Oumellal, A. Rougier, G.A. Nazri, J.M. Tarascon, L. Aymard, Nat. Mater. 7 (2008) 916–921.