Ionics (2006) 12: 153–157 DOI 10.1007/s11581-006-0019-1

ORIGINA L PA PER

Jiangang Li . Xiangming He . Maosong Fan . Chunrong Wan . Changin Jiang . Shichao Zhang

Capacity fading of LiCr0.1Mn1.9O4/MPCF cells at elevated temperature Received: 26 January 2006 / Accepted: 19 April 2006 / Published online: 25 May 2006 # Springer-Verlag 2006

Abstract Capacity fading of LiCr0.1Mn1.9O4 /MPCF (mesophase pitch-based carbon fiber) cells was investigated at elevated temperature (55 °C). The cells showed very fast capacity fading, keeping only 60% of capacity retention at the 100th cycle at 55 °C. The cycled electrodes and the electrolyte were analyzed using electrochemical test, inductively coupled plasma, and X-ray diffraction. Results of the analyses indicated that LiCr0.1Mn1.9O4 exhibited good effects on restraint of Mn dissolution and stabilization of structure at 55 °C. The cycled LiCr0.1Mn1.9O4 electrode and the cycled MPCF electrode presented good electrochemical performance again with fresh electrolyte. Therefore, it was proposed that the cycling fading of LiCr0.1Mn1.9O4/ MPCF cells was mainly caused by decomposition of electrolyte upon LiCr0.1Mn1.9O4 electrode during cycling. It was found that the decomposition of electrolyte led to the formation of a surface layer comprised of Li2CO3, LixPFy, CH3OCO2Li or (CH2OCO2Li)2, polymeric ether etc. The formation of this film consumed active lithium ions, leading to fast capacity fading of LiCr0.1Mn1.9O4/MPCF cell at elevated temperature. Keywords LiCr0.1Mn1.9O4 . Capacity fading . Electrolyte decomposition . Elevated temperature . Li-ion batteries

J. Li . M. Fan School of Material and Chemical Engineering, Beijing Institute of Petrochemical Technology, Beijing 102617, People’s Republic of China J. Li . X. He (*) . C. Wan . C. Jiang Institute of Nuclear & New Energy Technology, Tsinghua University, Beijing 100084, People’s Republic of China e-mail: [email protected] S. Zhang School of Material Science and Engineering, Beihang University, Beijing 100083, People’s Republic of China

Introduction Spinel LiMn2O4, with economical and environmental advantages, is considered to be one of the most promising cathode materials for rechargeable lithium-ion batteries. Its drawback is the significant capacity fading during cycling at elevated temperature due to several probabilities, e.g., manganese dissolution, electrolyte decomposition, Jahn– Teller distortion, and so forth [1–7]. Partial replacement of Mn in LiMn2O4 by Ni, Co, Al, and Cr metal ions can effectively enhance its cyclability. LiCr0.1Mn1.9O4 was reported to have better cycling performance at room temperature and at 55 °C, compared with the undoped spinel LiMn2O4 [8–10]. In Zhang et al. [9], chromium-doped spinel LiMn2O4 showed good cycling performance. The improvement of the cycle performance was attributed to the stabilization of the spinel structure because of the strong CrO2 bond. In Yoshio et al. [10], the influences of the partial substitution of Mn in LiMn2O4 by Cr3+ and Li+ on their charge/ discharge profiles were quite different: Cr3+ affected the profiles only in the high-voltage region, while Li+ showed in the high- and the low-voltage regions. Either Cr3+ or Li+ doping significantly improved the storage and the cycling performance of spinel LiMn2O4 at the elevated temperature, especially the doped spinels. The large capacity fading in the first cycle is due to the loss of Li from the system to form inactive materials. In the successive cycles, as the graphite electrode does not completely lose the function of the Li insertion/extraction and Li is supplied in the system from the counter electrode, capacity recovery occurs. Unfortunately, most of the previous studies were based on lithium metal anode, which could supplement the loss of active lithium ions from the cathode materials, probably leading to misunderstand the reason of capacity fading during cycling. Aurbach [3] proposed two types of capacity fading mechanisms. One mechanism involves the loss of about one-third of the initial capacity during the first Li deintercalation–intercalation cycle in the 3.5- to 4.2-V range. A new, less symmetric, and more disordered phase is

154

LiCr0.1Mn1.9O4 was prepared according to our previous work [17], and MPCF (Nikko) was used as received. LiCr0.1Mn1.9O4/MPCF (18650 type) cells were assembled. The positive electrodes were comprised of LiCr0.1Mn1.9O4 (86%), graphite (6%), acetylene black (2%), and polyvinylidene difluoride (PVDF) binder (6%). The negative electrodes were comprised of MPCF (89%), acetylene black (3%) and PVDF (8%). Commercial Li-battery grade electrolyte solution comprised of EC–DMC–EMC (1:1:1 by volume) and 1 M LiPF6 was used. The separator was microporous membrane (Celgard 2400). LiCr0.1Mn1.9O4/MPCF cells were cycled between 3.0 and 4.3 V at 0.5 C using Solartron 1477 multichannel battery analyzer. The cells were dismantled in the glove box after 100 cycles. The electrolytes were collected for the analysis of the content of the dissolved manganese. The electrodes were removed from the cells, cut into small pieces, and reassembled into coin cells with lithium metal anode for further electrochemical testing. Li/LiCr0.1Mn1.9O4 and Li/MPCF cells were also assembled and tested for comparison. The structures of the electrodes were analyzed using M21X X-ray diffractometer. The surface studies of

Results and discussion As prepared LiCr0.1Mn1.9O4 is cubic spinel as determined by the powder X-ray diffraction and has a discharge capacity of 115 mAh/g. The LiCr0.1Mn1.9O4/MPCF cells (18650 type) have a discharge capacity of 1.2 Ah. The cyclabilities of LiCr0.1Mn1.9O4/MPCF and Li/ LiCr0.1Mn1.9O4 cells are shown in Figs. 1 and 2, respectively. LiCr0.1Mn1.9O4/MPCF cell presents low-capacity fading and a capacity retention of 93% at the 100th cycle at 25 °C, consistent with the cyclability of Li/LiCr0.1Mn1.9O4 cell. However, LiCr0.1Mn1.9O4/MPCF cell shows fast capacity fading than Li/LiCr0.1Mn1.9O4 and presents only 60% of capacity retention at the 100th cycle at 55 °C, but Li/LiCr0.1Mn1.9O4 still presents 91% of capacity retention. Therefore, LiCr0.1Mn1.9O4 presents different cycling behaviors in the Li/LiCr0.1Mn1.9O4 and LiCr0.1Mn1.9O4/ MPCF systems at 55 °C, giving an understanding that cycling fading is not mainly caused by the LiCr0.1Mn1.9O4 material itself. Logically, the reason of capacity fading can probably be caused by the degradation of anode and/or electrolyte. The capacity fading of the spinel was commonly ascribed to Mn dissolution from the spinel into the electrolyte solution at elevated temperatures. However, inductively coupled plasma (ICP) analysis of the collected electrolytes from the cycled cells indicates that the amount of dissolved Mn has no significant difference between the cells cycled at 25 and 55 °C. The amount of dissolved Mn was found to be less than 0.2 wt% in all the cases, which only correspond to less than 1.2 mAh/g capacity loss according to the proposed models [4, 5]. To clarify if the capacity fading of LiCr0.1Mn1.9O4/ MPCF cells resulted from the disordered crystal structure triggered as a result of Mn dissolution [18], both cycled cathode and anode were cut into small pieces and reassembled into coin-type cells with lithium metal 1.5 1.4

Ah

Experimental description

the electrodes were performed on PHI5300 X-ray photoelectron spectroscopy (XPS), using Mg Kα excitation source. Depth dependence elemental composition was obtained by Ar ion-beam sputtering (5 keV).

25

1.3 1.2

Capacity

thus formed, probably close to the surface of the active mass. However, this new, inactive mass does not considerably influence the kinetics of the remaining active mass. Another capacity fading mechanism occurs at elevated potentials (>4.4 V vs Li/Li+). It involves the dissolution of manganese, which is accelerated by the parallel oxidation of the solution, which occurs at a low rate in this potential range. This capacity fading mechanism also leads to a pronounced increase in the electrode’s impedance. It is reported that the severe capacity loss is induced by the deposition of Mn ions dissolved out of the LiMn2O4 electrode on the graphite-negative electrodes [11, 12], as confirmed by further studies [13, 14]. This is one of the reasons causing capacity fading. In addition, the solid– electrolyte interphase (SEI) film is formed on the surface of the LiMn2O4 electrode at starting cycles, consuming lithium ions, leading to irreversible capacity loss [15]. A recent study showed that the SEI layer thickness on a cycled electrode increases in proportion to a linear function of the number of cycles [16]. When the amount of lithium ions in the cell system is limited, the increase of SEI film thickness can cause the capacity fading during cycling. It is practical to understand the mechanism of the capacity fading of LiCr0.1Mn1.9O4 in cells with carbon anode because carbon materials are widely used as anode in the commercial Lithium-ion batteries. In this study, LiCr0.1Mn1.9O4/MPCF (mesophase pitch-based carbon fiber; 18650 type) cell was assembled; its cycling capacity fading was studied at elevated temperature. The cycled electrodes, with which the cell capacity was fading, were cycled with fresh electrolyte and lithium metal electrode, as we tried to further understand the capacity fading of LiCr0.1Mn1.9O4.

1.1 1.0

55

0.9 0.8 0

20

40

60

80

100

Cycle number

Fig. 1 Cyclability of LiCr0.1Mn1.9O4/MPCF at 25 and 55 °C, respectively

155 Capacity mAh g-1 C

120 25

110 100

55 90 0

20

40 0

6 60

80 0

100

Cycle number

Fig. 2 Cyclability of Li/LiCr0.1Mn1.9O4 at 25 and 55 °C, respectively

as counter electrode. The cells were cycled once at 0.1 C. The cutoff voltage range was 3.3∼4.3 V for the Li/LiCr0.1Mn1.9O4 cell and 0.005∼2.5 V for the Li/MPCF cell, as shown in Fig. 3. The open-circuit voltages of the cycled LiCr0.1Mn1.9O4 electrodes at 25 and 55 °C were 3.9 and 4.1 V, respectively. This probably indicates that the value for x in Li1−xCr0.1Mn1.9O4 for the cycled electrode at 25 °C is smaller than at 55 °C. Figure 3 shows that the charge–discharge characteristics of cycled MPCF at 55 °C were same as the second cycle of the pristine electrode, indicating that the MPCF electrode shows no degradation after 100 cycles at 55 °C. Figure 3 also shows that the charge capacity of cycled LiCr0.1Mn1.9O4 at 55 °C was only 0.75 mAh, which is less than that at 25 °C. It indicates that less lithium ions intercalated into the cycled LiCr0.1Mn1.9O4 electrodes at 55 °C than at 25 °C, consistent with the data of open-circuit voltages. The reason is that they are at different initial charge states, but the discharge capacities are almost the same. It indicates that the cycled LiCr0.1Mn1.9O4 electrodes at 25 and 55 °C show no capacity fading. It seems that when LiCr0.1Mn1.9O4/MPCF cell is discharged to 2.75 V, almost all lithium ions are deintercalated from MPCF, but the amount of lithium ions intercalated into LiCr0.1Mn1.9O4 decreases during cycling, consisting with the capacity fading of LiCr0.1Mn1.9O4/MPCF cells. However, when Li metal is used as counter electrode, the LiCr0.1Mn1.9O4 electrodes cycled at 55 °C also resumes above 96% of initial capacity, indicating that the structure of LiCr0.1Mn1.9O4 remains very stable, even when cycled at 55 °C. At 55 °C, LiCr0.1Mn1.9O4/MPCF cell exhibits comparatively fast capacity fading and low charge–discharge coulomb efficiency over initial 40 cycles. To clarify what

occurred on electrodes at elevated temperature, the charge– discharge coulomb efficiency of cathode and anode with Li as counter electrode during cycling at 55 °C are shown in Fig. 4. Li/MPCF cell presents above 99.3% efficiency after the second cycle, same as that at 25 °C. However, the efficiency of Li/LiCr0.1Mn1.9O4 cell is only about 93% at initial cycling and increases gradually to nearly 100% until about 40 cycles. Tendency of change for Li/LiCr0.1Mn1.9O4 cell is consistent with that for LiCr0.1Mn1.9O4/MPCF cell. The elevated temperature appears to promote electrolyte decomposition on LiCr0.1Mn1.9O4 electrode, consuming active lithium ions and leading to fast capacity fading of LiCr0.1Mn1.9O4/MPCF cell. According to the above results, we propose that the capacity fading of the LiCr0.1Mn1.9O4/MPCF cell is probably caused by the deficiency of active lithium ions during cycling at 55 °C. Further X-ray diffraction (XRD) analysis of the cycled cathodes discharged to 3.3 V, as shown in Fig. 5, agrees with the above viewpoint. No change is observed from the XRD patterns of the electrodes, and the position and intensity of all reflection lines are also unchanged. The results showed that LiCr0.1Mn1.9O4 exhibits very good effects on the restraint of Mn dissolution and the stabilization of structure at elevated temperature. The capacity fading of the LiCr0.1Mn1.9O4/MPCF cell is probably attributed mainly to the decrease of the active lithium ions transferred between the LiCr0.1Mn1.9O4 cathode and the MPCF anode. XPS is utilized for the qualitative analysis of the compounds on the surface of the cycled electrodes. Elemental analysis shows that all cathode surfaces contain Mn, Cr, Li, C, O, P, and F. As reported in some papers [19, 20], there is a layer of electrolyte decomposition products (SEI) formed on the spinel electrodes after cycling. Atomic percentage analysis shows that the surface of the electrode cycled at 55 °C presents lower content of Mn and Cr than that cycled at 25 °C. This indicates that SEI film, triggered as a result of the electrolyte decomposition, formed on cathode electrode, is thicker during cycling at elevated temperature than that at 25 °C because the XPS analysis works only on the outmost surface of the materials.

350

102

Voltage V

4 c

a

3 2

a

Li / LiCro.1Mn1.9 1. O4 at 55

1

b c

Li / MPCF PCF at 55 Li / LiCr Lii ro.1Mn1.9O4 at 25 2

b

300

100 0

250

98

200

96

150

94

100

92

Coulomb efficiency

Capacity p mAh h g -1

55

0 0.0 .

0.5

1.0 .

1.5

2.0 .

2.5

3.0 .

3.5

Capacity mAh

Fig. 3 Charge–discharge curves of cycled electrodes with Li as counter electrode

0

20

40

60

80

100

Cycle number

Fig. 4 Cycling capacity and coulomb efficiency of Li/MPCF (△, ▴) and Li/LiCr0.1Mn1.9O4 (▿, ▾)

156

c

b

a 10

20

30

40 0

50

60

70

2θ

Fig. 5 XRD patterns of LiCr0.1Mn1.9O4 electrode before cycle (a), after 100 cycles at 25 °C (b), and after 100 cycles at 55 °C (c)

Furthermore, depth analysis of F1s, O1s, and C1s XPS peaks, as shown in Fig. 6, obtained from the LiCr0.1Mn1.9O4 electrode cycled at 25 and 55 °C, respectively, was carried out to understand the impact of the temperature on the electrolyte decomposition on the LiCr0.1Mn1.9O4 electrode. The C1s spectrum presents a large broad band at 284.7 eV corresponding to the conducting carbon black, graphite, and carbon binder, a shoulder peak at 286.0∼287.0 eV corresponding to CH3OCO2Li or (CH2OCO2Li)2 or polymeric ether, and a weak peak at 290.0 eV corresponding to Li2CO3. Decrease of the peak intensity at 286.0∼287.0 and at 290.0 eV on sputtering shows that the surface film contains Li2CO3, CH3OCO2Li or (CH2OCO2Li)2, or polymeric ether etc. Compared with the electrode cycled at 25 °C, the electrode cycled at 55 °C has strong peak at 290.0 eV, indicating that more Li2CO3 formed at elevated temperature, which is probably produced by the severe electrolyte decomposition. Li2CO3 was reported as a product of further decomposition of CH3OCO2Li, (CH2OCO2Li)2 and polymeric ether [21]. The F1s spectrum has two bands: one band around 687.3 eV corresponds to LixPFy and another band, around 685.2 eV, corresponds to PVDF binder and LiF. The intensity of peak around 687.3 eV decreases with sputtering, indicating that LixPFy is also a main component of the surface film on the cycled electrodes. Because the intensity of peak around 685.2 eV increases with sputtering, PVDF mainly contributes to the change of the peak. For the C1s Sputtering e duration tion

5min 2min in 1min in 0min i 5min i 2min i 1min 0min i

291

F1s Sputtering duration ration (d) 5min 2min 1min

(b)

O1s Sputtering g duration (f) 5min 2min 1min 0min

0min

(a)

5min 2min

(c)

1min 0min

288

285 5

Binding Energy / eV

7 690 687 684 Binding Energy / eV

(e) 5min 2min 1min 0min

537

534

531

528

Binding Energy / eV

Fig. 6 XPS of LiCr0.1Mn1.9O4 electrode cycled at 25 °C (a, c, e) and at 55 °C (b, d, f)

electrode cycled at 25 °C, the intensity of the peak for LixPFy became weaker than that for PVDF on sputtering for 1 min. However, for electrode cycled at 55 °C, the intensity of peak for LixPFy still remains stronger than that for PVDF even after sputtering for 5 min. This indicates that cycling at a higher temperature causes a thicker surface layer on the LiCr0.1Mn1.9O4 electrode. The band around 529.5 eV in the O1s spectrum is characteristic of LiCr0.1Mn1.9O4. In addition, there is a strong band around 532.5 eV assigned to carbonate (e.g., EC or Li2CO3) or semicarbonate, and a weak band around 534.0 eV corresponding to ether oxygen, both of them are associated with the surface film. Compared with the electrode cycled at 25 °C, the electrode cycled at 55 °C exhibits slower decrease of the peak intensity at 531.0∼535 eV and increase of the peak intensity around 529.5 eVon sputtering. This also indicates that more severe electrolyte decomposition on LiCr0.1Mn1.9O4 electrode cycled at higher temperature produces a thicker layer of surface film. Therefore, higher temperature accelerates electrolyte decomposition upon LiCr0.1Mn1.9O4 electrode, causing thicker layer of surface film comprised of Li2CO3, LixPFy, CH3OCO2Li or (CH2OCO2Li)2, polymeric ether etc., consuming more active lithium ions, and leading to faster capacity fading of LiCr0.1Mn1.9O4/MPCF cell. Because the loss of active lithium ions can be compensated by lithium metal anode, Li/LiCr0.1Mn1.9O4 cell still exhibited good cycling performance at elevated temperature. The proposed mechanism responsible for the capacity of spinel upon cycling at elevated temperature in this study is different from many suggestions reported previously by other groups. Further study needs to be done.

Conclusions Cycled at 55 °C, LiCr0.1Mn1.9O4/MPCF cell shows much faster capacity fading than Li/LiCr0.1Mn1.9O4 and presents only 60% of capacity retention at the 100th cycle. ICP analysis of the collected electrolytes from the cycled cell indicates that the amount of dissolved Mn is less than 0.2 wt%, which corresponds to 1.2 mAh/g capacity loss. Further electrochemical studies and XRD analysis of cycled electrodes show that LiCr0.1Mn1.9O4 exhibits good effects on restraint of Mn dissolution and stabilization of structure at elevated temperature, but the amount of active lithium ions intercalated into LiCr0.1Mn1.9O4 decreases fast during cycling. XPS analysis of cycled LiCr0.1Mn1.9O4 electrodes reveals that elevated temperature can accelerate electrolyte decomposition upon LiCr0.1Mn1.9O4 electrode; higher temperature causes thicker layer of surface film comprised of Li2CO3, LixPFy, CH3OCO2Li or (CH2OCO2Li)2, polymeric ether etc., consuming more active lithium ions, and leading to faster capacity fading of LiCr0.1Mn1.9O4/MPCF cell. Therefore, to reduce the electrolyte decomposition upon LiCr0.1Mn1.9O4 is an effective way to improve the performance of the spinel at elevated temperature.

157

References 1. Kulova TL, Karseeva EI, Skundin AM, Kachibaya EI, Imnadze RA, Paikidze TV (2004) Russ J Electrochem 40:494 2. Song GM, Wang YJ, Zhou Y (2004) J Power Sources 128:270 3. Aurbach D, Levi MD, Gamulski K, Markovsky B, Salitra G, Levi E, Heider U, Heider L, Oesten R (1999) J Power Sources 81–82:472 4. Jang DH, Shin YJ, Oh SM (1996) J Electrochem Soc 143:2204 5. Xia Y-Y, Zhou Y-H, Yashio M (1997) J Electrochem Soc 144:2593 6. Morita M, Nakagawa T, Yamada O, Yoshimoto N, Ishikawa M (2001) J Power Sources 97–98:354 7. Gummow RJ, Kock A, Thackery MM (1994) Solid State Ion 69:59 8. He XM, Li JJ, Cai Y, Wang YW, Ying JR, Jiang CY, Wan CR (2005) J Power Sources (in press) 9. Zhang D, Popov BN, White RE (1998) J Power Sources 76:81 10. Yoshio M, Xia Y-Y, Kumada N, Ma SH (2001) J Power Sources 101:79

11. Kumagai N, Komaba S, Kataoka Y, Koyanagi M (2000) Chem Lett 29:1154 12. Tsunekawa H, Tanimoto S, Marubayashi R, Fujita M, Kifune K, Sano M (2002) J Electrochem Soc 149:A1326 13. Komaba S, Kumagai N, Kataoka Y (2002) Electrochim Acta 47:1229 14. Komaba S, Itabashi T, Ohtsuka T, Groult H, Kumagai N, Kaplan B, Yashiro H (2005) J Electrochem Soc 152:A937 15. Zhang SS, Xu K, Jow TR (2002) J Electrochem Soc 149:1521 16. Lei JL, Li LJ, Kostecki R, Muller R, McLarnon F (2005) J Electrochem Soc 152:774 17. Li JG, Tang ZY, Xue JJ, Liu CY (2001) Chinese J Appl Chem 18:802 18. Inoue T, Sano M (1998) J Electrochem Soc 145:3704 19. Aurbach D, Gamolsky K, Markovsky B, Salitra G, Gofer Y, Heider U, Oesten R, Schmidt M (2000) J Electrochem Soc 147:1322 20. Eriksson T, Andersson A, Bishop A, Gejke C, Gustafsson T, Thomas JO (2002) J Electrochem Soc 149:A69 21. Aurbach D, Markovsky B, Levi MD, Schechter A, Moshkovich M, Cohen Y (1999) J Power Sources 81–82:95