Electrochemical Activities In Li2mno3

Journal of The Electrochemical Society, 156 共6兲 A417-A424 共2009兲

A417

0013-4651/2009/156共6兲/A417/8/$25.00 © The Electrochemical Society

Electrochemical Activities in Li2MnO3 Denis Y. W. Yu,*,z Katsunori Yanagida,* Yoshio Kato, and Hiroshi Nakamura* Mobile Energy Company, SANYO Electric Company, Limited, Kobe, Hyogo 651-2242, Japan Li2MnO3 is shown to be electrochemically active, with a maximum charge capacity of ⬃350 mAh/g and a discharge capacity of ⬃260 mAh/g at 25°C. A total of 1 mole of Li can be extracted from Li关Li1/3Mn2/3兴O2, and the first cycle efficiency is ⬃66% regardless of state of charge. Larger charge-discharge capacity is obtained from materials with smaller particle size and larger amount of stacking faults. Composition and structural analyses indicate that Li are removed from both the Li and transitional metal layers of the material during charging. Results from X-ray-absorption fine-structure measurements suggest that the valence of Mn remains at 4+ during charging but is reduced during discharging. Charging is accompanied by gas generation: at 25°C, oxygen is the main gas detected, and the total amount accounts for ⬃1/8 mole of O2 generation from Li关Li1/3Mn2/3兴O2. At an elevated temperature, amount of CO2 increases due to electrolyte decomposition. Li2MnO3 shows poor cycle performance, which is attributed to phase transformation and low charge-discharge efficiency during cycling. Low first-cycle efficiency, gas generation, and poor cycle performance limit the usage of Li2MnO3 in practical batteries. © 2009 The Electrochemical Society. 关DOI: 10.1149/1.3110803兴 All rights reserved. Manuscript submitted November 21, 2008; revised manuscript received February 4, 2009. Published April 3, 2009.

The study of Mn-based Li-excess layered cathode materials with a formula Li关LixMnyMz兴O2 共where M represents one or more transition metal elements兲 has been gaining popularity because these materials can give a discharge capacity ⬎200 mAh/g,1-16 much higher than practical discharge capacity of common cathode materials, such as LiCoO2 and LiMn2O4. Successful application of these materials can potentially increase energy density of Li-ion batteries. One challenge that researchers face in the study of Li关LixMnyMz兴O2 materials is that there is almost an endless combination of systems, depending on the additional element 共or elements兲 M and the ratio of cations in the layered material. A change in composition is sure to affect the electrochemical performance of the cathode material. In addition, modifications such as acid treatment and surface coating can further change material behavior.4,8,17-20 It is therefore difficult to compare results from different publications due to differences in material composition and processing techniques. Though, all of the Li关LixMnyMz兴O2 materials share a common feature: a plateau at 4.5 V vs Li/Li+ during initial charging which is similar to the behavior of Li2MnO3. We therefore think it would be beneficial to reinvestigate the fundamental mechanism of Li2MnO3 to help further understanding of the properties of Mn-based Li-excess materials. Electrochemical properties of Li2MnO3 and acid-leached Li2MnO3 were previously reported in the literature.4,5,8,21-30 However, the described discharge capacity varies from publication to publication, with values ranging from a few to ⬃200 mAh/g. The inconsistency in results is probably due to differences in synthesis condition and test condition.25,29,30 In theory, Li2MnO3 has a capacity of ⬃458 mAh/g 共with respect to the initial mass of the sample兲 if all Li can be extracted. In this study, we investigated the reasons for the limited capacity observed in experiments and studied the effect of physical parameters of Li2MnO3 on the discharge capacity of the material by methods such as X-ray diffractometry 共XRD兲 and scanning electron microscopy 共SEM兲. It was originally thought that Li2MnO3 is electrochemically inactive because Mn has a valence of 4+, and it is unlikely that Mn will go to higher valence. Many researchers attributed the electrochemical activity of Li2MnO3 to the removal of Li2O from the active material.2,4,8,25-28 However, there are few discussions on what is the status of the oxygen atoms after charge-discharge. There is also a lack of quantitative analysis on the correlation between the amount of oxygen “lost” from the active material and the observed capacities, because direct measurement of oxygen stoichiometry is difficult. Armstrong et al.31 observed O2 gas generation from the initial

* Electrochemical Society Active Member. z

E-mail: [email protected]

charging of Li关Li0.2Ni0.2Mn0.6兴O2 material using an in situ differential electrochemical mass spectroscopy. Though, the amount of oxygen gas cannot be determined by their method. In this study, we used a quantitative approach aiming to understand the fundamental mechanism of Li2MnO3, by tracking the amount of Li in the active material by inductively coupled plasma 共ICP兲 atomic emission spectroscopy and the amount of oxygen gas generated by gas chromatography 共GC兲. The change in valence of Mn is studied qualitatively by X-ray absorption fine structure spectroscopy 共XAFS兲. The results are presented in this paper. In addition, the cycle performance of the active material and practical issues of Li2MnO3 will be discussed. Throughout the paper, the formula Li2MnO3 is used interchangeably with Li关Li1/3Mn2/3兴O2, a notation that gives more information about the structure of the material.

Experimental Li2MnO3 powders were synthesized by a solid-state reaction with LiOH · H2O and MnCO3 · nH2O 共n ⬃ 0.5兲 as precursors. The starting materials were thoroughly mixed by milling in acetone for 1 h and then dried at 60°C. The resulting precursors were then annealed at temperatures between 400 and 1000°C in air. Structure of the powders was studied by XRD using a Cu K␣ source 共50 kV, 300 mA兲. Morphology and particle size information were obtained from SEM. The surface area of the materials was determined using an AUTOSORB-1 equipment by Quantachrome: ⬃0.5 g of material was first dried at 150°C for 1.5 h and then the surface area was determined by a five-point Brunauer–Emmett–Teller 共BET兲 method with nitrogen as the adsorbate gas. ICP atomic emission spectroscopy was used to study the amount of Li and Mn in the material before and after charge-discharge. For electrochemical evaluation, Li2MnO3 active material was mixed with acetylene black and polyvinyl difluoride 共PVdF兲 in 1-methyl-2-pyrrolidone 共NMP兲 with a weight ratio of 80:10:10 to form a slurry. The slurry was then coated onto a roughened aluminum current collector using a doctor blade. The electrodes were rolled with a calender press to a packing density of about 2.2–2.3 g/cm3, with a typical thickness of 25–35 ␮m. These electrodes were assembled in a glove box with Ar atmosphere using Li metal as the counter and reference electrodes with a layer of separator to make a flat laminated test cell. 1 M LiPF6 in ethylene carbonate/diethylcarbonate 共EC/DEC兲 = 3:7 by volume was the electrolyte used in the experiments. The electrodes were sandwiched between glass plates to maintain contact of the electrodes, and the cells were typically charged and discharged at room temperature 共25°C兲 between 4.8 and 2 V vs Li/Li+ with a constant current rate

Downloaded 21 Apr 2009 to 129.49.51.226. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

Journal of The Electrochemical Society, 156 共6兲 A417-A424 共2009兲

A418

Table I. Electrochemical properties of Li2MnO3 synthesized at different temperatures.

a

Annealing conditions 共°C兲 共h兲

Charge capacity 共mAh/g兲

Discharge capacity 共mAh/g兲

Charge efficiency 共%兲

fwhm 18.7° peak 共001兲 plane

Avg. particle sizea 共nm兲

BET surface area 共m2 /g兲

400 共48兲 425 共10兲 600 共10兲 750 共10兲 800 共10兲 850 共10兲 900 共10兲 1000 共10兲

367.0 355.0 342.0 294.8 290.0 206.7 63.6 28.4

258.7 251.8 232.4 197.2 191.6 135.4 43.8 22.7

70.5 70.9 68.0 66.9 66.1 65.5 68.9 79.9

0.434 0.368 0.302 0.252 0.221 0.180 0.132 0.123

70 70 90 110 130 140 390 650

20.2 18.0 15.9 10.7 9.0 6.6 2.2 1.5

Measured from SEM images.

of 10 mA/g, unless stated otherwise. Charge-discharge capacities were calculated with respect to the mass of Li2MnO3 before charging. To study gas generation from the sample, an in situ measurement was performed to monitor the change in cell thickness during charge-discharge. The cells were disassembled after charging, and the composition and total volume of the generated gas from the cells were measured by GC. Structural information of the material at different state of charges was further investigated by an ex situ XAFS with the Mn–K edge in transition mode using the BL08 beamline at SPring-8 in Hyogo Prefecture, Japan. X-ray absorption near-edge spectroscopy 共XANES兲 was used to monitor the electronic structure of Mn and extended X-ray absorption fine-structure 共EXAFS兲 was used to monitor the interaction of Mn with surrounding atoms. To interpret the XAFS results of Li2MnO3, XAFS of LiNi1/3Co1/3Mn1/3O2 and LiMn2O4 were taken as references. Average particle size and BET surface area of the reference materials were 10 ␮m and 2.0 m2 /g for LiNi1/3Co1/3Mn1/3O2 and 13 ␮m and 0.35 m2 /g for LiMn2O4, respectively. Results and Discussion Effect of synthesis conditions on electrochemical behavior of Li2MnO3.— Annealing temperature was varied from 400 to 1000°C to study the effect of synthesis condition on the electrochemical performance of Li2MnO3 共see Table I兲. The materials were made into test cells, and their electrochemical properties were tested. First cycle charge-discharge profiles with a rate of 10 mA/g at 25°C of several Li2MnO3 samples are shown in Fig. 1. An initial charge plateau at ⬃4.5 V vs Li/Li+ is observed for all samples. Reducing the synthesis temperature lowers the potential of the charge plateau and leads to higher charge and discharge capacities. A maximum charge capacity of about 350–370 mAh/g is obtained, correspond5

1000°C

900°C

850°C

800°C

ing to the removal of ⬃1 mole of Li out of Li关Li1/3Mn2/3兴O2. In return, a maximum discharge capacity of about 250–260 mAh/g is obtained. First cycle efficiency 共FCE兲 is ⬃66% for most of the samples, showing a reinsertion of only 2/3 of the extracted Li. The discharge capacity is first of all limited by the charge capacity. Thus, a larger discharge capacity is obtained from a sample with larger charge capacity. In addition, the low FCE suggests that the discharge capacity 共i.e., the reinsertion of Li兲 is limited by the structure of the material. This will be further discussed in a later section together with other analytical results. To understand how the change in synthesis condition leads to a change in electrochemical behavior of Li2MnO3, we first studied the structure of the materials by XRD. XRD profiles of the different samples are shown in Fig. 2. All samples show a layered structure with a c2/m space group, with alternating Li layers and transitional metal 共TM兲 layers separated by oxygen layers. The superlattice peaks between 20 and 30° in the XRD profile 共with a Cu K␣ source兲, which are not observed in other layered materials, are due to the ordering of Li/Mn in the TM layers. Figure 2 shows that below a synthesis temperature of 600°C, only a single broad peak at 20.9° is observed. This broadening of the superlattice peak is attributed to an increase in the amount of stacking faults 共shifting of the TM layers兲 perpendicular to the layered 共001兲 direction with a lower synthesis temperature.32 The role of the stacking fault on the electrochemical properties of the active material is unclear. From thermodynamics, a material without defects has the lowest energy state. Introducing stacking faults and other defects in the material increases the energy state of the material, which may lower the activation barrier for Li diffusion and allow Li to be extracted at a lower potential. Further studies on the effect of defects on electrochemical performance are in progress. The effect of synthesis temperature on the particle size and morphology was studied by SEM. Figure 3 shows the SEM pictures of

(131)

(001)

600°C 400°C

(130)

4 3.5 3 2.5 2

400°C

Intensity (a.u.)

Potential (V vs. Li/Li+)

4.5

1000°C 900°C

850°C 800°C 600°C 400°C

600°C 800°C 850°C

10mA/g

900°C

first cycle

1000°C

1.5 0

100

200

300

400

Capacity (mAh/g)

Figure 1. 共Color online兲 First cycle charge-discharge profiles of Li2MnO3 synthesized at different temperatures.

15

25

35

45 2 θ (°)

55

65

75

Figure 2. 共Color online兲 XRD profiles of Li2MnO3 synthesized at different temperatures.

Downloaded 21 Apr 2009 to 129.49.51.226. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

Journal of The Electrochemical Society, 156 共6兲 A417-A424 共2009兲 2.0

900°C

1μm

Normalized intensity (a.u.)

1000°C

1μm

800°C

A419

400-600°C

LiNi1/3Co1/3Mn1/3O2

1.8 1.6 1.4 1.2 1.0

MnO6 connectivity

0.8

Li2MnO3 450°C 600°C 750°C 900°C

0.6 0.4 0.2

1s → 3d

0.0 6535

6540

6545

6550

6555

6560

6565

Energy (eV)

1μm

Figure 5. 共Color online兲 XANES profiles of Mn K-edge of Li2MnO3 synthesized at 450, 600, 750, and 900°C, all four profiles overlap on top of each other. XANES profile of LiNi1/3Co1/3Mn1/3O2 is also shown for comparison.

1μm

Figure 3. 共Color online兲 SEM images of Li2MnO3 synthesized at different temperatures.

the sample annealed at different temperatures. Average particle size of the samples was determined by measuring the dimension of the particles from the SEM images, and the results are shown in Table I. Samples synthesized at a lower temperature have smaller particles size: Li2MnO3 synthesized at 400°C has a particle size as small as 70 nm. Increasing temperature leads to larger particles. Significant grain growth is observed for samples made at a temperature of ⬎850°C, particle size is ⬃650 nm when the sample is annealed at 1000°C. This is consistent with the decrease in the full width halfmaximum 共fwhm兲 of the XRD peaks with increasing synthesis temperature 共see Table I兲. Figure 4 shows the relationship between discharge capacity and particle size, showing larger capacity from Li2MnO3 nanoparticles. We attribute this to a smaller diffusion path for smaller particles, which allows the material to be charged to a higher capacity. Another consequence of reducing the annealing temperature is an increase in surface area. The BET surface area of the Li2MnO3 samples was measured and the results are shown in Table I. An inverse relationship between surface area and particle size is obtained, the smaller the particle size, the larger the surface area. A simple geometrical calculation shows that surface area per unit mass 共A/m兲 is related to the diameter of the particle 共d兲 by A/m = 6/␳d for spherical particles, where ␳ is the density of the material. Our result is consistent with the inverse proportionality between A/m and d, suggesting that the increase in surface area of our Li2MnO3

0

5

BET surface area (m2/g) 10 15

20

25

Discharge capacity (mAh/g)

300

Particle size

250

BET

200 150 100 50 0 0

100

200

300

400

500

600

700

Particle size (nm)

Figure 4. 共Color online兲 Relationship between particle size, BET surface area, and discharge capacity of Li2MnO3 samples.

sample synthesized at a reduced temperature is a direct result of the decrease in particle size, and not from other contributions, such as surface roughening. The effect of annealing temperature on the local environment of Mn in as-made Li2MnO3 was studied by XAFS. Figure 5 shows the results of Mn K-edge XANES of Li2MnO3 samples synthesized at 450, 600, 750, and 900°C. The double-peak near 6540 eV is attributed to a 1s-to-3d transition, which is expected to be weak for an octahedral coordination of Mn.33,34 The shoulder at ⬃6548 eV is associated with the connectivity of the MnO6 octahedral.35 No significant change in the XANES Mn K-edge is observed with annealing temperature. Similarly, EXAFS, which gives information about the interaction of Mn with neighboring atoms, does not depend on synthesis condition 共EXAFS of one of the pristine samples is shown in Fig. 12兲. These results suggest that annealing temperature does not change the local environment of Mn. In summary, electrochemical activity of Li2MnO3 is enhanced by the reduction of particle size of the material and the increase in the amount of stacking faults in the structure. Effect of particle size is expected to be dominating since Li2MnO3 with poor crystallinity but only gives discharge capacity ⬍100 mAh/g has been reported elsewhere.29,30 Though the effect of stacking faults cannot be ruled out because it was not possible to make small particles without stacking faults with the current synthesis method. Li2MnO3 sample synthesized at 425°C, which can give a charge and discharge capacity of 350 and 250 mAh/g at 25°C, respectively, is used in subsequent studies to investigate the charge-discharge mechanism of Li2MnO3. Tracking Li atoms and structural change during charge-discharge of Li2MnO3 at 25°C.— Li2MnO3 cells were charged and discharged to different capacities, and the cells were disassembled and their compositions were determined by ICP to correlate the measured capacities from electrochemical tests with the amount of Li in the material. Test results are shown in Table II. ICP results show that pristine samples have a Li/Mn ratio of 2:1, according to the stoichiometry of our starting precursors. By assuming that the amount of Mn in the active material remains constant after charge and discharge, we can calculate the amount of Li in the samples from ICP measurements. Figure 6 shows the relationship between measured capacity 共cumulative兲 and the amount of Li in LixMn2/3O2. The dotted line in Fig. 6 represents the theoretical capacity at a given Li content if all extracted Li contribute to the capacity. When the electrochemical tests were performed at 25°C, measured capacities coincide with the Li contents in the material according to calculation. This directly shows that the electrochemical activity of Li2MnO3 is accompanied by Li extraction/insertion. When charging was performed at 60°C, the measured capacity is ⬃50 mAh/g larger than

Downloaded 21 Apr 2009 to 129.49.51.226. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

Journal of The Electrochemical Society, 156 共6兲 A417-A424 共2009兲

A420

Table II. Measured charge-discharge capacities and LiÕMn ratios.

a

Samples

Test temperature 共°C兲

Charge capacity 共mAh/g兲

Discharge capacity 共mAh/g兲

Cumulative capacitya 共mAh/g兲

Li/Mn 共ICP results兲

x in LixMn2/3O2 共calculated兲

Pristine A1 A2 A3 A4 A5 A6 A7 B1

— 25 25 25 25 25 25 25 60

— 50.0 99.7 150.0 248.8 337.5 336.2 335.4 429.2

— — — — — — 125.0 210.0 —

0 50.0 99.7 150.0 248.8 337.5 211.2 125.4 429.2

2.00 1.75 1.66 1.32 0.95 0.47 1.05 1.47 0.40

1.33 1.16 1.10 0.88 0.64 0.31 0.70 0.98 0.27

Cumulative capacity = charge capacity − discharge capacity.

that corresponding to Li extraction. Other mechanisms, such as electrolyte decomposition, contribute to the excess capacity at an elevated temperature. To further study the electrochemical behavior of Li2MnO3, the electrodes were charged to different capacities and then discharged. Figure 7 shows the charge and discharge curves of electrodes that are initially charged to 0, 50, 150, 250, and 350 mAh/g 共fully charged兲 共cutoff capacities during charging are marked by X兲. An electrode that is initially discharged 共without charging兲 gives a capacity of ⬃20 mAh/g, suggesting that the as-made material can accommodate some amount of Li. The amount of Li accommodated from initial discharge decreases with synthesis temperature, with a capacity of ⬍2 mAh/g for Li2MnO3 synthesized at 900°C. We suspect the initial discharge capacity originates from either a surface

Cummulative capacity (mAh/g)

500

∗ 60°C charge

electrolyte decompositio

450 400

25°C charge 25°C discharge

350 300

theoretical capacity of Li[Li1/3Mn2/3]O2

250 200 150

pristine sample x=4/3

discharge

100

charge

50 0 0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

x in LixMn2/3O2

Figure 6. 共Color online兲 Plot of cumulative capacity with respect to Li content in LixMn2/3O2 共dotted line represents the theoretical capacity at a given Li content兲.

effect, such as the presence of additional active sites on the large surface, or from accommodation of Li in the grain boundary or defects such as stacking faults. When the charge capacity is increased, the discharge capacity also increases accordingly. At 25°C, a maximum charge capacity of 350 mAh/g is obtained, which corresponds to the extraction of 1 mole of Li out of Li关Li1/3Mn2/3兴O2. Not all of the Li can be extracted from the active material. During discharge, a maximum capacity of 250 mAh/g is obtained, in which 229 mAh/g can be accounted for by the 4+/3+ transition of Mn. The remaining 21 mAh/g most likely corresponds to the capacity from Li accommodated on surface or defect sites, same as that obtained when the material is initially discharged. The discharge capacities of Li2MnO3 obtained from different states of charge during first cycle is summarized in Fig. 8; all data points lie on a straight line with a slope close to 0.66. This indicates that the charge-discharge mechanism of Li2MnO3 is the same throughout the entire first cycle, regardless of the state of charge. For every three Li extracted out of the lattice, only two can be reinserted back into the lattice. This suggests that there is a structural limitation on the amount of Li reinserted, instead of a chemical limitation from the redox reaction of Mn. In addition, Li extraction is most likely not uniform throughout the particles, resulting in “reacted” and “not reacted” regions, similar to a two-phase transformation. The plateau at ⬃4.5 V vs Li/Li+ observed during charging supports this hypothesis. The change in structure of Li2MnO3 during charging and discharging was monitored by XRD. A summary of the XRD profiles at different charged states is shown in Fig. 9. Upon charging, the overall intensity of XRD peaks decreases. The peak at 18.7°, which correspond to the 共001兲 plane, is shifted to a larger angle, indicating

300

Potential (V vs. Li/Li+)

X

X

4.5

1st discharge capacity (mAh/g)

5 X

X

1st cycle 10mA/g

4 3.5 3 2.5 2

0 50 mAh/g mAh/g

1.5 0

50

100

150 mAh/g 150

250 mAh/g 200

350 mAh/g 250

charged to

slope = 1 250 slope = 0.66

200 150 100 Li2MnO3 50 0

300

350

400

Capacity (mAh/g)

Figure 7. 共Color online兲 Charge-discharge curves of Li2MnO3 by limiting the state of charge to 0, 50, 150, 250, and 350 mAh/g 共marked by X兲.

0

50

100

150

200

250

300

350

400

1st charge capacity (mAh/g)

Figure 8. 共Color online兲 First cycle discharge capacity of Li2MnO3 by limiting the charge capacity.

Downloaded 21 Apr 2009 to 129.49.51.226. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

Journal of The Electrochemical Society, 156 共6兲 A417-A424 共2009兲

A421

C1 C2 C3 C4

discharge

17

19

21 35

37

39

41 43 2 θ (°)

Normalized intensity (a.u.)

charge

45 62

64

66

68

a decrease in lattice spacing between the layers with Li extraction. In addition, the intensity of superlattice peak 共between 20 and 25°兲 is reduced gradually upon charging, indicating that the Li/Mn ordering in the transition metal layer is disrupted. This suggests that Li is extracted not only from the Li layer but also from the TM layer during charging. Because the charge mechanism is thought to be the same throughout the charging process, as shown in Fig. 8, we suspect Li are extracted from both Li layer and TM layer at the same time. At a fully charged state, almost all of Li atoms in the TM layer are removed, as the superlattice can hardly be seen. During discharging, there is a general shift for all major peaks to smaller angles, indicating an increase in lattice constants with Li insertion. The fully discharged state has a different XRD profile than that of the pristine sample: e.g., 共001兲 peak of the fully discharged state is at a smaller angle and only one single peak is observed at 65°. These results suggest a change in lattice structure after discharge, with a XRD profile similar to a spinel structure. Intensity of the superlattice peak remains small after discharge. Because only 2/3 of the extracted Li can be reinserted, we suspect the atoms reenter the Li layer, leaving Li vacancies in the TM layer. Status of Mn during charge-discharge.— XAFS measurement was carried out to study the change in local environment of Mn during charging and discharging. LiNi1/3Co1/3Mn1/3O2 is used as a reference to help interpret the XANES profile of Li2MnO3. The reason for using LiNi1/3Co1/3Mn1/3O2 as a reference is that Li2MnO3 and LiNi1/3Co1/3Mn1/3O2 shows similar Mn K-edge XANES profiles because both materials have a layered structure with Mn in an octahedral coordination 共as shown in Fig. 5兲. After charge-discharge, the cells were disassembled and the cathode materials were removed from the Al collector for XAFS measurements. Figure 10 shows the change in XANES profile of the reference LiNi1/3Co1/3Mn1/3O2 material after charge and discharge. When charged to 4.3 V, disappearance of the shoulder at ⬃6548 eV and the shift in the peak top to higher energy are attributed to the change in the local environment of Mn from Li extraction, with the valence of Mn remaining 4+.34,36 The change in the shape and position of the Mn XANES profile is reversible upon discharge. Quantitative analysis of the XANES profile is difficult because of the shape change during charge-discharge. An absorption intensity 共␮兲 of 0.4 is marked in Fig. 10 as a qualitative guide for the eye to show the shift 共or the lack of shift兲 of the profiles. Figure 11 shows the corresponding XANES profile of Li2MnO3. During charging, a disappearance of the shoulder around 6548 eV and a gradual shift to higher energy of the peak top, similar to that of LiNi1/3Co1/3Mn1/3O2, are also observed. The edge position at ␮ = 0.4 hardly moves. These results suggest that the effect of charging of Li2MnO3 on the local environment of Mn is similar to that of

1.6 1.4

pristine LiNi1/3Co1/3Mn1/3O2

pristine charged * discharged

1.2 charge discharge

1.0 0.8

charge

0.6

discharge

0.4 0.0 6543

70

Figure 9. 共Color online兲 XRD profiles of Li2MnO3 at different stages of charge. C1, C2, C3, and C4 correspond to a state of charge of 0, 150, 250, and 350 mAh/g, respectively. D1 and D2 correspond to a discharge of 125 and 210 mAh/g from a fully charged electrode.

1.8

0.2

D1 D2

6548

6553

6558

6563

Energy (eV)

Figure 10. 共Color online兲 XANES profiles of Mn K-edge LiNi1/3Co1/3Mn1/3O2 during charging 共to 4.3 V兲 and discharging.

in

LiNi1/3Co1/3Mn1/3O2 共i.e., that the valence of Mn remains 4+兲. After discharge, as opposed to LiNi1/3Co1/3Mn1/3O2, the XANES profile of Li2MnO3 does not return to the original position. The overall edge 共both the peak top and the edge position at ␮ = 0.4兲 is shifted to a lower energy, indicating a decrease in valance of Mn during discharge. Shape of the discharged profile is similar to that of LiMn2O4, as opposed to Li2MnO3. At this point, further quantitative investigation on the change of valence of Mn is needed to further understand the charge-discharge mechanism. In addition, in situ XAFS measurements should be conducted to minimize reaction between the environment and the charged and discharged electrodes. EXAFS profiles of Li2MnO3 show that charging is accompanied with a reduction of both Mn-O and Mn–Mn Fourier transformation 共FT兲 peaks 共see Fig. 12兲 and this reduction in FT is also not reversible upon discharge. We attributed the results to a distortion around the Mn atoms associated with the removal of Li from the Li and TM layers and the reduction of coordination number due to removal of oxygen from the lattice. Attempts to model the EXAFS results by simulations codes, such as FEFF, are in progress to clarify the origin of the large change. Oxygen evolution during charge-discharge.— XAFS results suggest that Mn valence remains 4+ during the charging of Li关Li1/3Mn2/3兴O2. But ICP results show that Li atoms are removed from the active material. If charge compensation is accompanied by oxygen removal, as suggested by various researchers,2,4,8,25-28 the extraction of one mole of Li from Li关Li1/3Mn2/3兴O2 corresponds to a change in oxygen stoichiometry from 2 to 1.5. In this study, we investigated how much of the change can be explained by gas generation.

2.0 Normalized intensity (a.u.)

Intensity (a.u.)

2.0

1.8 1.6 1.4 1.2

pristine Li2MnO3

pristine charged * discharged LiMn2O4

LiMn2O4

1.0 0.8

charge

discharge charge

0.6

discharge

0.4 0.2 0.0 6543

6548

6553

6558

6563

Energy (eV)

Figure 11. 共Color online兲 XANES profiles of Mn K-edge in Li2MnO3 during charging 共to 4.8 V兲 and discharging.

Downloaded 21 Apr 2009 to 129.49.51.226. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

Journal of The Electrochemical Society, 156 共6兲 A417-A424 共2009兲

A422

50

Mn-O

12

pristine Mn-Mn

charge

charged * discharged

10 FT Magnitude

45 Detected gas (ml/g)

14

charge 8

discharge discharge

6 4

40

25°C 1M LiPF6 EC/DEC=3:7

45°C

60°C

35 H2 O2 CO CO2

30 25 20 15 10 5

2

0 0

150 0

1

2

3 R (Å)

4

6

5

Figure 12. 共Color online兲 EXAFS profiles of Mn K-edge 共FT兲 of Li2MnO3 during charging and discharging.

First, a simple in situ test to monitor the change in cell thickness of a flat laminated Li2MnO3 cell during charge-discharge was performed. Figure 13 shows the variation of cell thickness and applied current with time. The cell was initially charged at a rate of 30, 20, and 10 mA/g with intermittent rest. Afterward, the “not-yet-fullycharged” cell was discharged at 20 mA/g for 2 h and then recharged again for 4 h. During initial charging, thickness of the test cell increases with charging and remains unchanged during rest. The larger the applied current, the larger is the rate of thickness change 关i.e., rate of gas generation 共see Fig. 13兲兴. Thus, gas generation is an integral part of the charging process of Li2MnO3. Gas generation is only observed during the initial charge of the active material. During discharge, thickness does not change. Upon recharging, no change in thickness is observed for the first 2 h 共which corresponds to the removal of Li that was inserted in the previous discharge step兲. After 2 h, cell thickness increases again from the charging of the “not reacted” Li2MnO3. This supports the hypothesis that Li is not extracted uniformly throughout the particles during first charge, as suggested from the charge efficiency results in Fig. 8. To determine the type of gases and the total volume generated, cells were charged to a capacity of 150, 250, and 350 mAh/g at 25°C and the gas was collected and analyzed by GC. The results are shown in Fig. 14 and summarized in Table III. H2, O2, CO, and CO2 are detected within the collected gas. The amounts of H2, CO, and

Thickness change (mm)

Current (mA/g)

rest

rest

rest

rest

rest

30 20 10 0 -10 -20 2hrs

0.8 slope =0.240

0.6

slope =0.080

0.4 0.2 0

2hrs

slope =0.145

slope =0.145 0

2

4

6 8 Time (hour)

10

12

14

Figure 13. 共Color online兲 Change in Li2MnO3 cell thickness with charge and discharge.

250

350

350

350

Charge capacity (mAh/g)

Figure 14. 共Color online兲 Volume of gases collected from Li2MnO3 cells with 1 M LiPF6 in EC/DEC = 3:7 charged at 25, 45, and 60°C.

CO2 remain almost the same regardless of charge capacity, where as the amount of O2 increases with charge capacity. This confirms that oxygen gas is generated during charging of Li2MnO3. About 44 mL of gas per active-material mass is detected by GC when the electrode is charged to a capacity of 350 mAh/g 共equivalent to the evolution of ⬃1/8 mole of O2 gas during the removal of one mole of Li from Li关Li1/3Mn2/3兴O2兲. This corresponds to a change in oxygen stoichiometry from 2 to 1.75. Our result shows that the amount of oxygen gas can account for half of the charge compensation. The remaining oxygen atoms, if they are removed from the active material, may have reacted with electrolyte or formed Li2O at the anode and are therefore not accounted for in this study. A certain amount of proton exchange may have also occurred during charging, resulting in a smaller change in oxygen stoichiometry. A better tracing of oxygen and hydrogen atoms in the system is needed to clarify the full process. Effect of test temperature on charge mechanism of Li2MnO3.— Li2MnO3 cells were charged to 350 mA/g at 25, 45, and 60°C and the composition of the resulting gas was determined by GC to study the effect of test temperature on charge mechanism. The results are shown in Fig. 14 and summarized in Table III. At 25°C, the main gas after charging is O2. With increasing test temperature, a significant reduction in the amount of O2 and an increase in the amount of CO2 are observed. At 60°C, a small amount of CH4 is also detected, but the signal overlapped with the O2 signal and therefore cannot be isolated. CO2 generation is associated with electrolyte decomposition, and therefore, the test results suggest that there is a change in charge mechanism of Li2MnO3 with temperature, from an oxygenloss process to the decomposition of electrolyte at a higher temperature. The change in mechanism is also reflected in the material structure after charging. Li2MnO3 active materials were collected after charging at 25 and 60°C to 350 mAh/g and examined by XRD, and the profiles are shown in Fig. 15. Li2MnO3 charged at 25°C shows a slight shift in the main peak from 18.68 → 18.80° 关the 共001兲 plane兴. This shift corresponds to the shrinking of the lattice plane distance after Li is extracted from the layer. Material charged at 60°C shows a larger shift of the main peak from 18.68 → 19.02° together with the growth of a peak at 38.3° 共marked by “x” in Fig. 15兲. These features are characteristics of a P3 structure associated with hydrogen bonding between adjacent oxide layers,25,26 originated from the exchange of Li+ and H+ ions in the active material, that are also seen after acid treatment of Li2MnO3.4,8,37 The protons are likely to be generated from electrolyte decomposition. Note that the total amount of gas detected from the charged cell is smaller at a higher test temperature 共see Table III兲. This is consistent with electrolyte decomposition and proton exchange because the interchange of H+ and Li+ does not involve oxygen loss, and a larger portion of the gas is absorbed in the electrolyte since the solubility of CO2 is higher than that of O2.38,39

Downloaded 21 Apr 2009 to 129.49.51.226. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

Journal of The Electrochemical Society, 156 共6兲 A417-A424 共2009兲

A423

Table III. Volume of detected gas per active material mass from charged Li2MnO3 cells at 25, 45, and 60°C (bracketed values refer to the percentage). Charge temperature

45°C

60°Ca

Charge capacity

150 mAh/g

250 mAh/g

350 mAh/g

350 mAh/g

350 mAh/g

Total detected volume 共mL/g兲 H2 O2 CO CO2

16.3

30.5

44.2

41.3

22.7

1.2共7.2%兲 12.3共75.6%兲 0.8共4.8%兲 2.0共12.4%兲

1.4共4.5%兲 26.4共86.5%兲 0.8共2.6%兲 2.0共6.4%兲

0.7共1.5%兲 39.8共90.0%兲 0.9共2.0%兲 2.9共6.5%兲

1.4共3.4%兲 28.2共68.4%兲 1.5共3.7%兲 10.1共24.4%兲

1.3共5.8%兲 2.6共11.3%兲 3.3共14.7%兲 15.4共67.8%兲

Small amount of CH4 is measured at 60°C but the signal overlaps with O2 signal from GC.

Our results show that oxygen loss is the dominating process at 25°C and electrolyte decomposition is more prominent at a higher temperature, confirming the results of Robertson and Bruce.25 Both mechanisms occurs at a similar potential of 4.5–4.6 V vs Li/Li+, and therefore, a small change in test temperature can tip the balance in favor of one mechanism over another. We suspect the contribution from electrolyte decomposition will be reduced if a more oxidationresistant electrolyte is used in the cell and further tests are being done to verify this. Cycle performance of Li2MnO3 with different states of charge.— Li2MnO3 electrodes were tested with different state of charges 共fixed by charge capacity兲, and the cycle performance was monitored 共see Fig. 16兲. Electrodes charged to 4.8 V vs Li/Li+ or with a charge cutoff of 250 mAh/g show poor cycle performance, where the capacity reduces to below 100 mAh/g after 10 cycles. By reducing the state of charge to 150 mAh/g, the cycle life can be prolonged, but the discharge capacity eventually decays after 15 cycles. The cycle efficiency 共ratio of discharge and charge capacities兲 is ⬃90% when the electrode is charged to 150 mAh/g. This suggests that Li2MnO3 keeps on losing Li with cycling and eventually runs out of Li that leads to a capacity fade. Cycle performance can be improved by reducing the utilization of the electrode. For example, an electrode with a charge cutoff of 50 mAh/g shows good cycle performance with a cycle efficiency of close to 100%. However, this is not a practical solution to improve cycle performance because we end up sacrificing the capacity. The charge-discharge curves of a cycled electrode with a cutoff capacity of 150 mAh/g are shown in Fig. 17. Discharge capacity increases during the first few cycles, and the formation of discharge plateaus around 4.0 and 2.9 V vs Li/Li+ is observed. These plateaus are characteristics of spinel Mn, suggesting that there is a layered-

to-spinel phase transformation during cycling, consistent with other reports.25,29,30 The phase transformation is most likely facilitated by the poor crystallinity of our sample, together with the removal of oxygen from the lattice and the formation of Li vacancies in the TM layer during charge-discharge. Issues of Li2MnO3.— We have shown that Li2MnO3 made at a low temperature is electrochemically active, with a discharge capacity as large as 250–260 mAh/g. However, Li2MnO3 shows the following characteristics: first cycle efficiency is ⬃66%; gas is generated during the initial charging process; electrolyte decomposition is observed at elevated temperature; and cycle performance is poor. These limit the usage of phase-pure Li2MnO3 in practical Li-ion batteries. Methods to increase first cycle efficiency, reduce gas generation and improve cycle performance are needed. One such method that has been explored by numerous researches is to stabi-

300 Discharge capcaity (mAh/g)

a

25°C

4.8V cut

250

250mAh/g cut 200 150

150mAh/g cut

100 50mAh/g cut

50 0 0

5

10

15

20

25

Cycle number

19.02°

17

18

19

20 2θ

5

21

22

23

pristine 25°C charge 60°C charge

Potential (V vs. Li/Li+)

Intensity (a.u.)

x

Figure 16. 共Color online兲 Cycle performance of Li2MnO3 with limited charge capacity.

18.80°

Intensity (a.u.)

18.68°

10mA/g

1st

4.5

2nd 4

5th

8th

150mAh/g charge cutoff

3.5 3 2.5

5th

increase of plateau with cycling

2

1st 2nd

8th

1.5

15

25

35

45

55

65

75

2 θ (°)

Figure 15. 共Color online兲 XRD of Li2MnO3 materials charged at 25 and 60°C to 350 mAh/g.

0

50

100

150

200

Capacity (mAh/g)

Figure 17. 共Color online兲 Charge-discharge curves of Li2MnO3 with a 150 mAh/g cutoff.

Downloaded 21 Apr 2009 to 129.49.51.226. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

A424

Journal of The Electrochemical Society, 156 共6兲 A417-A424 共2009兲

lize the Li2MnO3 phase by the formation of a “composite,” similar to that in Li关LixMnyMz兴O2 materials. Yoon et al. reported that the presence of Co in the material allows some charge compensation through the generation of electron holes in the O 2p state;34 thus, the charge mechanism could potentially be changed by the addition of elements and components into the material. Further studies are being conducted to extend our knowledge of Li2MnO3 to Li关LixMnyMz兴O2 materials.

6. 7. 8. 9. 10. 11. 12. 13.

Conclusions Li2MnO3 with a maximum discharge capacity of ⬃260 mAh/g was synthesized by solid-state reaction. Electrochemical performance of the material is found to depend strongly on synthesis conditions. Reducing particle size and increasing the amount of stacking faults enable Li2MnO3 to be charged at a lower potential to a higher capacity. ICP measurements confirmed that Li is extracted and reinserted into the material, consistent with electrochemical capacity, during charge-discharge. Ex situ XANES results suggest that Mn valence remains the same during charging and is reduced during discharging. Further study to quantify the valence of Mn and investigate the stability of the charged/discharged electrodes in air is needed. GC measurements confirmed the generation of O2 during charging. The total amount of O2 gas detected corresponds to ⬃1/8 mole of O2 per mole of Li extracted from Li关Li1/3Mn2/3兴O2 at 25°C. This can account for half of the charge compensation. Some of the oxygen lost from the active material may exist in other forms other than oxygen gas in the cell. In addition, H+ /Li+ exchange is also expected to take part during charging, especially at high potential and at high temperature. On the basis of our results, we speculate that vacancies are left in the lattice after oxygen is removed from the active material. The structure after charging is not in an equilibrium state, and the rearrangement of atoms within the material leads to a gradual transformation to spinel structure with cycling. Further analyses are in progress to clarify the charge-discharge mechanism. Sanyo Electric Company, Limited assisted in meeting the publication costs of this article.

References 1. Z. Lu, D. D. MacNeil, and J. R. Dahn, Electrochem. Solid-State Lett., 4, A191 共2001兲. 2. Z. Lu and J. R. Dahn, J. Electrochem. Soc., 149, A815 共2002兲. 3. J.-S. Kim, C. S. Johnson, J. T. Vaughey, M. M. Thackeray, S. A. Hackney, W. Yoon, and C. P. Grey, Chem. Mater., 16, 1996 共2004兲. 4. C. S. Johnson, J.-S. Kim, C. Lefief, N. Li, J. T. Vaughey, and M. M. Thackeray, Electrochem. Commun., 6, 1085 共2004兲. 5. C. S. Johnson, N. Li, J. T. Vaughey, S. A. Hackney, and M. M. Thackeray, Electrochem. Commun., 7, 528 共2005兲.

14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39.

D. A. R. Barkhouse and J. R. Dahn, J. Electrochem. Soc., 152, A746 共2005兲. T. A. Arunkumar, Y. Wu, and A. Manthiram, Chem. Mater., 19, 3067 共2007兲. C. S. Johnson, J. Power Sources, 165, 559 共2007兲. S.-H. Kang, P. Kempgens, S. Greenbaum, A. J. Kropf, K. Amine, and M. M. Thackeray, J. Mater. Chem., 17, 2069 共2007兲. J.-M. Kim, S. Tsuruta, and N. Kumagai, Electrochem. Commun., 9, 103 共2007兲. C. W. Park, S. H. Kim, I. R. Mangani, J. H. Lee, S. Boo, and J. Kim, Mater. Res. Bull., 42, 1374 共2007兲. S.-H. Park, S.-H. Kang, C. S. Johnson, K. Amine, and M. M. Thackeray, Electrochem. Commun., 9, 262 共2007兲. M. Tabuchi, Y. Nabeshima, K. Ado, M. Shikano, H. Kageyama, and K. Tatsumi, J. Power Sources, 174, 554 共2007兲. M. M. Thackeray, S.-H. Kang, C. S. Johnson, J. T. Vaughey, R. Benedek, and S. A. Hackney, J. Mater. Chem., 17, 3112 共2007兲. Y. Wu and A. Manthiram, Electrochem. Solid-State Lett., 10, A151 共2007兲. J. Kikkawa, T. Akita, M. Tabuchi, M. Shikano, K. Tatsumi, and M. Kohyama, Electrochem. Solid-State Lett., 11, A183 共2008兲. S.-H. Kang, C. S. Johnson, J. T. Vaughey, K. Amine, and M. M. Thackeray, J. Electrochem. Soc., 153, A1186 共2006兲. Y. Wu and A. Manthiram, Electrochem. Solid-State Lett., 9, A221 共2006兲. Y. Wu, A. V. Murugan, and A. Manthiram, J. Electrochem. Soc., 155, A635 共2008兲. J. M. Zheng, J. Li, Z. R. Zhang, X. J. Guo, and Y. Yang, Solid State Ionics, 179, 1794 共2008兲. M. H. Rossouw and M. M. Thackeray, Mater. Res. Bull., 26, 463 共1991兲. Y. Shao-Horn, Y. Ein-Eli, A. D. Robertson, W. F. Averill, S. A. Hackney, and W. F. Howard Jr., J. Electrochem. Soc., 145, 16 共1998兲. P. Kalyani, S. Chitra, T. Mohan, and S. Gopukumar, J. Power Sources, 80, 103 共1999兲. C. S. Johnson, S. D. Korte, J. T. Vaughey, M. M. Thackeray, T. E. Bofinger, Y. Shao-Horn, and S. A. Hackney, J. Power Sources, 81-82, 491 共1999兲. A. D. Robertson and P. G. Bruce, Chem. Mater., 15, 1984 共2003兲. A. R. Armstrong, A. D. Robertson, and P. G. Bruce, J. Power Sources, 146, 275 共2005兲. C. Gan, H. Zhan, X. Hu, and Y. Zhou, Electrochem. Commun., 7, 1318 共2005兲. Y.-S. Hong, Y. J. Park, K. S. Ryu, and S. H. Chang, Solid State Ionics, 176, 1035 共2005兲. S.-H. Park, Y. Sato, J.-K. Kim, and Y.-S. Lee, Mater. Chem. Phys., 102, 225 共2007兲. S.-H. Park, H. S. Ahn, G. J. Park, J. Kim, and Y.-S. Lee, Mater. Chem. Phys., 112, 696 共2008兲. A. R. Armstrong, M. Holzapfel, P. Novak, C. S. Johnson, S. Kang, M. M. Thackeray, and P. G. Bruce, J. Am. Chem. Soc., 128, 8694 共2006兲. J. Bréger, M. Jiang, N. Dupré, Y. S. Meng, Y. Shao-Horn, G. Ceder, and C. P. Grey, J. Solid State Chem., 178, 2575 共2005兲. M. Morcrette, P. Barboux, J. Perrière, T. Brousse, A. Traverse, and J. P. Boilot, Solid State Ionics, 138, 213 共2001兲. W.-S. Yoon, C. P. Grey, M. Balasubramanian, X.-Q. Yang, D. A. Fischer, and J. McBreen, Electrochem. Solid-State Lett., 7, A53 共2004兲. Y. J. Park, Y.-S. Hong, X. Wu, M. G. Kim, K. S. Ryu, and S. H. Chang, Bull. Korean Chem. Soc., 25, 511 共2004兲. P.-Y. Liao, J. G. Duh, J.-F. Lee, and H. S. Sheu, Electrochim. Acta, 53, 1850 共2007兲. S. Venkatraman and A. Manthiram, J. Solid State Chem., 177, 4244 共2004兲. J. Read, K. Mutolo, M. Ervin, W. Behl, J. Wolfenstine, A. Driedger, and D. Foster, J. Electrochem. Soc., 150, A1351 共2003兲. P. Kolář, H. Nakata, J.-W. Shen, A. Tsuboi, H. Suzuki, and M. Ue, Fluid Phase Equilib., 228-229, 59 共2005兲.

Downloaded 21 Apr 2009 to 129.49.51.226. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp