Thermal stability of LixCoO2 cathode for lithium ion battery

Thermal stability of LixCoO2 cathode for lithium ion battery

Solid State Ionics 148 (2002) 311 – 316 www.elsevier.com/locate/ssi Thermal stability of LixCoO2 cathode for lithium ion battery Yasunori Baba, Shige...

142KB Sizes 2 Downloads 31 Views

Solid State Ionics 148 (2002) 311 – 316 www.elsevier.com/locate/ssi

Thermal stability of LixCoO2 cathode for lithium ion battery Yasunori Baba, Shigeto Okada, Jun-ichi Yamaki * Institute of Advanced Material Study, Kyushu University 6-1, Kasuga Kouen, Kasuga 816-8580, Japan

Abstract It is well known that charged LixCoO2 (x < 1) is metastable, and that oxygen evolution has been observed at temperatures above 200 jC. LixCoO2, delithiated by a chemical method using H2SO4, was investigated by means of differential scanning calorimetry (DSC) with/without an electrolyte (1 M LiPF6/ethylene carbonate (EC) + dimethyl carbonate (DMC)). The lithium content x in the delithiated LixCoO2 was determined by atomic absorption spectroscopy. The DSC profile of Li0.49CoO2 showed two exothermic peaks, one beginning at 190 jC and the other beginning at 290 jC. From high-temperature X-ray diffraction (XRD), it was found that the first peak, from 190 jC, was the phase transition from a monoclinic (R3¯m) to a spinel structure (Fd3m). The DSC measurements of Li0.49CoO2 with the electrolyte at various mixing ratios showed two exothermic peaks, one beginning at 190 jC and the other at 230 jC. The exothermic heat of each peak was proportional to the amount of Li0.49CoO2. The peak starting at 190 jC probably resulted from the decomposition of solvent due to an active cathode surface, and the peak starting at 230 jC was electrolyte oxidation caused by released oxygen from Li0.49CoO2. The exothermic heat from 190 to 230 jC based on cathode weight was 420 F 120 J/g, and that from 230 to 300 jC was 1000 F 250 J/g. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Li-ion batteries; Chemical delithiation; LixCoO2; Cathode; 1 M LiPF6/EC+DMC; Thermal stability

1. Introduction Recently, demand for Li-ion batteries has increased enormously as a power source for portable devices, due to their excellent characteristics of high voltage, high energy density, and light weight. Large Li-ion batteries are expected to be used in EV and similar applications in the future. However, the larger batteries have a serious safety problem: the risk of explosion resulting from either shorting or an external temperature increase [1 – 6]. Therefore, safety improvements to the large batteries are indispensable for their practical application. It is generally consid*

Corresponding author. Tel.: +81-82-583-7790; fax: +81-92583-7790. E-mail address: [email protected] (J.-i. Yamaki).

ered that the ‘‘thermal runaway’’ of Li cells occurs if the heat output exceeds the thermal diffusion [2,5]. Therefore, it is difficult to design large cells that can pass a safety-test criteria. The thermal stability or the thermal behavior of Li-ion cells has been investigated energetically by DSC or Accelerating Rate Calorimetry (ARC) [7 –11] in order to reduce the heat output. The thermal behavior of lithium nickel oxide [4,12 – 16,18], lithium manganese oxide [4,12,18,20], and lithium cobalt oxide [4,12,17 – 21], as the cathode materials of the Li-ion batteries, has been investigated. LixCoO2 is a widely used cathode material for Li ion cells. However, the thermal behavior of LixCoO2 is not clear. MacNeil et al. reported that the first reaction that occurs when LixCoO2 is heated in electrolyte can be modeled using a simple kinetic model that describes an auto-catalytic reaction [19], though they

0167-2738/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 2 7 3 8 ( 0 2 ) 0 0 0 6 7 - X

312

Y. Baba et al. / Solid State Ionics 148 (2002) 311–316

did not attempt to identify the chemical reaction. Recently, they found that the reaction of Li0.5CoO2 with EC + DEC solvent initiates at temperatures as low as 130 jC, which is much lower than the decomposition temperature of Li0.5CoO2 itself [21]. Oxygen loss from the material according to the reaction Li0.5CoO2 ! 0.5 LiCoO2 + 1/6 Co3O4 + 1/6O2 occurs at temperatures above 200 jC [18]. MacNeil et al. also reported that the amount of reaction heat generated by the reaction of LixCoO2 with electrolyte is independent of the electrode/electrolyte mass ratio [20], though their experiment is a rough estimation because they cannot remove the electrolyte, carbon black, and PVdF from the electrode when they measured the weight of the electrode. In this study, LixCoO2 was synthesized chemically with sulfuric acid. Using the LixCoO2 without the electrolyte, carbon black, and PVdF, the thermal behavior of this delithiated LixCoO2 with/without electrolyte was measured by DSC to elucidate the decomposition mechanism. In addition, the decomposition mechanism of LixCoO2 alone was investigated by high-temperature X-ray diffraction (XRD).

2. Experimental LiCoO2 was prepared by firing a mixture of Li2CO3 and Co3O4 at 850 jC for 24 h following firing at 500 jC for 5 h in air [22]. Chemical delithiation of LiCoO2 was carried out by stirring a suspension of 5 g of LiCoO2 in 250 ml of 0.5 M H2SO4 from 1 to 24 h [22,23]. The product was filtered and washed several times with acetone, and dried at 80 jC in a vacuum. The Li content of LixCoO2 was analyzed by atomic absorption spectroscopy (HITACHI, Z-5000). The electrochemical property of delithiated LixCoO2 was tested by measuring the quasi-open circuit voltage (QOCV). The QOCV was measured after the cell had been standing for about 1 h at zero current flow for every 0.025 Li/mol at a discharge current of 0.2 mA/cm2. The cathode disk contained 70 wt.% Li0.49CoO2, 25 wt.% acetyleneblack (Denki Kagaku), and 5 wt.% polytetrafluoroethylene (Polyflon TFE F-103, Daikin Industry). The thermal behavior of LixCoO2 with/without electrolyte was measured by differential scanning calorimetry (DSC) and thermogravimetry (TG) (Rigaku, Thermo plus). The sample for DSC measurement was packed in

a stainless steel case, which was then crimp-sealed in a glove box filled with argon. An open aluminum case was used to measure the TG in air. The heating rate of both DSC and TG was 5 jC/min. The electrolyte was a 1 M LiPF6 solution with equal volumes of ethylene carbonate (EC) and dimethyl carbonate (DMC). The Li0.49CoO2 was characterized by high-temperature Xray diffraction with CuKa radiation (Rigaku, RINT2500).

3. Result and discussion The Li content of chemically delithiated LixCoO2 decreased as reaction time increased. The minimum x value in LixCoO2 was 0.49 at a reaction time of 24 h. Fig. 1 shows the results of QOCV measurement for chemically and electrochemically delithiated Li0.49CoO2 during the first discharge. The QOCV was similar in both cases. The difference was probably a result of proton intercalation into the chemically delithiated LixCoO2 during the treatment with H2SO4 [22,23]. The thermal behavior (DSC) of chemically delithiated Li0.49CoO2 with acetyleneblack and polyterafluoroethylene was almost the same as the DSC profile obtained from electrochemical delithiated Li0.49CoO2. Fig. 2 shows the results of DSC measurement of chemically delithiated LixCoO2 (x = 0.49, 0.57, 0.64, 0.75, 0.84, and 0.94) alone, without any other components. The profiles show gradual exothermal heat

Fig. 1. QOCV profiles of chemically and electronically delithiated Li0.49CoO2 during the first discharge.

Y. Baba et al. / Solid State Ionics 148 (2002) 311–316

Fig. 2. DSC profiles of chemically delithiated LixCoO2 at a heating rate of 5 jC/min.

between 190 and 400 jC. The temperature of the exothermic reaction shifted to the low side, and heat flow increased as Li content decreased. Li0.49CoO2 showed two main peaks at 230 and 320 jC, corresponding self-heating temperatures of electrochemically delithiated Li0.5CoO2 from ARC at 140 and 240 jC reported by MacNeil et al. [21]. The exothermic heats of Li0.49CoO2 at the first peak and the second peak are about 50 and 10 J/g based on cathode weight, respectively. The exothermic heats of LixCoO2 at x = 0.53, 0.64, 0.75, 0.84, and 0.94 were about 40, 30, 20, 20, 10 J/g, respectively. Fig. 3 shows the results of TG measurement of chemically delithiated

313

LixCoO2 alone. The derivative, dM/dT, of Li0.49CoO2 showed sudden weight loss of 230 jC, which is a similar tendency to that of electrochemically delithiated Li0.4CoO2 with carbon black reported by Dahn et al. [18]. Oxygen loss from the material according to the reaction Li0.5CoO2 ! 0.5 LiCoO2 + 1/6 Co3O4 + 1/6O2 occurs at temperatures above 200 jC [18]. The exothermic reactions (Fig. 2) occurred at the same temperature as the onset temperature of weight loss measured by TG (Fig. 3), except for the case of Li0.49CoO2. Therefore, the exothermic reaction of Li0.49CoO2 starting at 190 jC can be another reaction or phase transition, without cathode weight change. Fig. 4 shows high-temperature XRD profiles of chemically delithiated Li0.49CoO2. The peaks shown as Pt in Fig. 4 were those from a sample holder made of Pt. The structure at room temperature was identified as a monoclinic (R3¯m) structure, which is the same structure of electrochemically delithiated Li0.49CoO2 [18,24]. The X-ray pattern of Li0.49CoO2 at 220 jC shows a spinel structure (Fd3m). These experimental results suggest that the exothermic reaction starting at 190 jC was probably caused by the phase transition from the layered rocksalt to the spinel structure. Indeed, Arai et al. [14,15] reported a similar exothermic reaction for LixNiO2. The XRD profile at 350 jC was identified as a mixture of hexagonal LiCoO2 (R3¯m) and nonstoichiometric spinel Co3O4 (Fd3m). The result at 350 jC is consistent with the result of electrochemically delithiated Li0.4CoO2 [18].

Fig. 3. DTG (derivative of sample weight versus temperature) of chemically delithiated LixCoO2 at a heating rate of 5 jC/min.

314

Y. Baba et al. / Solid State Ionics 148 (2002) 311–316

Fig. 4. High-temperature XRD profiles of chemically delithiated Li0.49CoO2.

Fig. 5 shows DSC profiles of chemically delithiated Li0.49CoO2 and 3 Al electrolyte (1M LiPF6 EC/ DMC). The cathode weight was altered from 0 to 3.9 mg. There were mainly two peaks, one starting at 190 jC and the other starting at 230 jC. The electrolyte without cathode showed a exothermic peak around 270 jC. However, the peak disappeared by the addition of 0.8 mg of cathode. The cathode inhibited the thermal decomposition of the electrolyte. The mechanism is not clear. Probably, PF5, which is created by a reaction of LiPF6 ! LiF + PF5 and is believed to react with solvents, may be consumed by reacting with the cathode. Fig. 6 plots exothermic heat against the cathode weight. Open circles show the exothermic

heat from 190 to 230 jC. Closed circles show the exothermic heat from 230 to 300 jC. Either exothermic heat is approximately proportional to the cathode weight. This result suggests that there is excess electrolyte in the sample case. The exothermic heat from 190 to 230 jC based on cathode weight was 420 F 120 J/g, and that from 230 to 300 jC was 1000 F 250 J/g. The exothermic reaction of chemically delithiated Li0.49CoO2 with EC/DMC without LiPF6 started at 170 jC and reached a broad peak at 220 jC. In this case, the exothermic heat based on cathode weight was about 1000 J/g, which is larger than the exothermic heat of electrolyte (with Li salt; 420 F 120 J/g). MacNeil et al. concluded that the

Fig. 5. DSC profiles of chemically delithiated Li0.49CoO2 with electrolyte for various cathode weight in 3 Al electrolyte.

Fig. 6. Observed exothermic heat for various weight of Li0.49CoO2. (o): Exothermic heat from 190 to 230 jC. (.): Exothermic heat from 230 to 300 jC.

Y. Baba et al. / Solid State Ionics 148 (2002) 311–316

reaction is inhibited in the presence of electrolyte because of a coating of the electrode particles by salt decomposition products [21]. The exothermic heat, they obtained, is 500 F 60 J/g for Li0.5CoO2 + EC/ DEC. Their value is smaller than ours. However, our value showed a better fitting to their calculated value of 1209 J/g for EC/DEC A roughly calculated value for our EC + DMC solvent is about 1.2  103 J/g. In addition, neither Co3O4 nor LiCoO2, both with the electrolyte, did not show any exothermic peaks at the temperature range from 190 to 230 jC. Therefore, the peak starting at 190 jC was probably a decomposition of solvent caused by active surface of the chemically delithiated Li0.49CoO2 as pointed out by MacNeil et al. [21]. MacNeil et al. reported that the total heat generated in the first exotherm of the Li0.5CoO2/electrolyte reaction is independent of the amount of electrolyte present, using ARC [20]. Their result is consistent with our result shown in Fig. 5, though MacNeil et al. could not study the precise relation because they used a Li0.5CoO2 electrode/electrolyte sample to mix with electrolyte. MacNeil et al. also reported that the first reaction that occurs when LixCoO2 is heated in electrolyte can be modeled using a simple kinetic model that describes an auto-catalytic reaction [19], though they did not attempt to identify the chemical reaction. We think that electrolyte may reacts with oxygen at the surface of Li0.49CoO2, and creates oxygen defects on the surface. The oxygen atoms on the surface become unstable by the defects and become more reactive with electrolyte. From this reason, the reaction may be auto-catalytic. Recently, MacNeil et al. reported that the reaction mechanism is Li 0.5 CoO 2 + 0.1C 3 H 4 O 3 (EC) ! 0.5LiCoO 2 + 0.5CoO + 0.3CO2 + 0.2H2O assuming full combustion of EC [21]. Our experimental results strongly support their results. The sharp peak starting at 230 jC is consistent with the sudden oxygen evolution temperature from DSC of Li0.49CoO2 (Fig. 2), and did not appear with the reaction between Li0.49CoO2 and the EC + DMC solvent from our experiment. If we use the consideration by MacNeil et al. [21], the reaction at higher than 230 jC was not observed because the reaction at the temperature range from 190 to 230 jC is not inhibited in the presence of the solvent, and all the Li0.49CoO2 was decomposed to LiCoO2, 0.5CoO and oxidation products of the solvent. Therefore, it is

315

considered that the exothermic heat starting at 230 jC is mainly caused by the oxidation reaction of the electrolyte, as already pointed out by MacNeil et al. [21].

4. Conclusion Chemically delithiated Li0.49CoO2 exhibited an exothermic reaction beginning at 190 jC, at which temperature oxygen evolution does not occur. From high-temperature XRD, it was found that the exothermic peak from 190 jC is caused by the structural change from layered R3¯m to spinel (Fd3m). The reaction of Li0.49CoO2 with the electrolyte exhibited mainly two exothermic peaks. The peak starting at 190 jC probably reflected the decomposition of solvent due to an active cathode surface, and the peak starting at 230 jC was electrolyte oxidation caused by released oxygen from Li0.49CoO2. The exothermic heat from 190 to 230 jC based on cathode weight was 420 F 120 J/g, and that from 230 to 300 jC was 1000 F 250 J/g.

Acknowledgements The authors wish to thank Rigaku for the measurement of high-temperature X-ray diffraction, and Mitsubishi Heavy Industries, for financial support.

References [1] S. Tobishima, Y. Sakurai, J. Yamaki, J. Power Sources 68 (1997) 455. [2] S. Tobishima, J. Yamaki, J. Power Sources 81 – 82 (1999) 882. [3] K. Kitoh, H. Nemoto, J. Power Sources 81 – 82 (1999) 887. [4] Ph. Biensan, B. Simon, J.P. Peres, A. de Guibert, M. Broussely, J.M. Bodet, F. Perton, J. Power Sources 81 – 82 (1999) 906. [5] S. Tobishima, K. Takei, Y. Sakurai, J. Yamaki, J. Power Sources 90 (2000) 188. [6] S. Passerini, F. Coustier, B.B. Owens, J. Power Sources 90 (2000) 144. [7] M.N. Richard, J.R. Dahn, J. Electrochem. Soc. 146 (1999) 2068. [8] U. von Sacken, E. Nodwell, A. Sundher, J.R. Dahn, J. Power Sources 54 (1995) 240. [9] U. von Sacken, J.R. Dahn, Solid State Ionics 69 (1995) 284.

316

Y. Baba et al. / Solid State Ionics 148 (2002) 311–316

[10] B.M. Way, U. von Sacken, The Electrochem. Soc. Meeting Abstracts, San Antonio, TX, Oct. 6 – 11 vol. 96-2, 1996. [11] H. Maleki, G. Deng, A. Anani, J. Howard, J. Electrochem. Soc. 146 (1999) 3224. [12] Z. Zhang, D. Fouchard, J.R. Rea, J. Power Sources 70 (1998) 16. [13] J. Cho, H. Jung, Y.C. Park, G. Kim, H.S. Lim, J. Electrochem. Soc. 147 (2000) 15. [14] H. Arai, S. Okada, Y. Sakurai, J. Yamaki, Solid State Ionics 109 (1998) 295. [15] H. Arai, S. Okada, Y. Sakurai, J. Yamaki, J. Electrochem. Soc. 144 (1997) 3117. [16] W. Li, J.C. Currie, J. Wolstenholme, J. Power Sources 68 (1997) 565. [17] Y. Sato, K. Kanari, K. Takano, T. Masuda, Thermochim. Acta 296 (1997) 75.

[18] J.R. Dahn, E.W. Fuller, M. Obrovac, U. von Sacken, Solid State Ionics 69 (1994) 265. [19] D.D. MacNeil, L. Christensen, J. Landucci, J.M. Paulsen, J.R. Dahn, J. Electrochem. Soc. 147 (2000) 970. [20] D.D. MacNeil, T.D. Hatchard, J.R. Dahn, J. Electrochem. Soc. 148 (2001) A663. [21] D.D. MacNeil, J.R. Dahn, J. Electrochem. Soc. 148 (2001) A1205. [22] R. Gupta, A. Manthiram, J. Solid State Chem. 121 (1996) 483. [23] E. Zhecheva, R. Stoyanova, J. Solid State Chem. 109 (1994) 47. [24] G.G. Amatucci, J.M. Tarascon, L.C. Klein, J. Electrochem. Soc. 143 (1996) 1114.