Capacity fade mechanism of CoSb3 intermetallic compound

Capacity fade mechanism of CoSb3 intermetallic compound

Journal of Electroanalytical Chemistry Journal of Electroanalytical Chemistry 568 (2004) 323–327 www.elsevier.com/locate/jelechem Capacity fade mech...

466KB Sizes 2 Downloads 23 Views

Journal of

Electroanalytical Chemistry Journal of Electroanalytical Chemistry 568 (2004) 323–327 www.elsevier.com/locate/jelechem

Capacity fade mechanism of CoSb3 intermetallic compound Jian Xie, Gaoshao Cao, Yaodong Zhong, X.B. Zhao

*

Department of Materials Science and Engineering, Zhejiang University, Zheda road 38, Hangzhou 310027, PR China Received 2 February 2004; received in revised form 2 February 2004; accepted 12 February 2004 Available online 20 March 2004

Abstract The intermetallic compound CoSb3 exhibits a larger Li-storage capacity compared with carbon-based materials, which are currently used as the anode materials in commercial lithium ion batteries. However, the CoSb3 electrode shows poor cycling stability during repetitive charge and discharge cycles. This behavior has precluded it from practical use in lithium ion batteries to date. In our present study, galvanostatic cycling, scanning electron microscopy (SEM), electrochemical impedance spectroscopy (EIS), and Fourier transition infra-red (FTIR) techniques were used to explore the capacity fade mechanism of CoSb3 electrodes. This preliminary study may supply a methodology for improving the electrochemical performances of the CoSb3 compound or other intermetallic compounds. Ó 2004 Elsevier B.V. All rights reserved. Keywords: Lithium ion batteries; Intermetallic compound; Electrochemical performances; Capacity fade; Anode material

1. Introduction The ever-increasing demand for high energy-storage systems has driven extensive search for new Li-storage materials with improved electrochemical properties. In recent years, there has been great interest in some intermetallic compounds due to their significantly larger capacity compared to carbon-based materials. The key drawback of these compounds is that they undergo substantial volume changes during charge and discharge cycling. It is well understood that the anode capacity of alloy electrodes is based on the alloying/de-alloying of the active component with Li rather than on the intercalation/de-intercalation of Li ions into/from the interlayer spacing in the graphite. The discharge can be described as a Li-ion insertion and inactive component extrusion process, during which, volume expansion is unavoidable, since there is a large difference in density between the intermetallic compound and its lithiation product. *

Corresponding author. Tel.: +86-571-8795-1451; fax: +86-5718795-1171/1403. E-mail address: [email protected] (X.B. Zhao). 0022-0728/$ - see front matter Ó 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.jelechem.2004.02.005

On the contrary, minor volume changes are observed in the case of the graphite electrode. The capacity loss for the graphite anode occurring in the initial cycles is associated with the formation of the solid electrolyte interface (SEI) layer [1–4], and only slight loss of capacity is seen in the subsequent cycles. The capacity fade of carbon materials upon long-term cycling can be ascribed to the exfoliation of the graphite particles and their amorphization, which result from the co-intercalation of the solvent molecules and their subsequent reduction within the graphite lattice [5–9]. There has been much work done in seeking an explanation for the capacity fade mechanism of carbonaceous materials for lithium ion batteries. Nevertheless, little has been reported in the literature concerning the exact fade mechanism of intermetallic compounds. The poorer electrochemical performances of CoSb3 electrodes compared to those of graphite electrodes suggest that the alloy electrode obeys a different capacity fade mechanism. In our present work, this mechanism was investigated using galvanostatic cycling, scanning electron microscopy (SEM), electrochemical impedance spectroscopy (EIS), and Fourier transition infrared (FTIR) techniques.

324

J. Xie et al. / Journal of Electroanalytical Chemistry 568 (2004) 323–327

2. Experimental

3. Results and discussion

th

Voltage / V

rd nd

4 3 2 1st

1.5

1.0

0.5

st

1 nd

th

2

4 0.0 0

100

200

300

400

500

600

700

800

-1

Capacity / mA h g

Fig. 1. Voltage profiles of the CoSb3 electrode for the initial four cycles.

which indicates that irreversible reactions (surface reduction and SEI formation reactions) take place during the first discharge process. These reactions can be written as [6]: EC þ 2e ! C2 H4 þ CO2 3

ð1Þ

þ CO2 3 þ EC þ 2Li ! ðCH2 OCO2 LiÞ2

ð2Þ

þ CO2 3 þ 2Li ! Li2 CO3

ð3Þ

The large irreversible capacity loss observed in the first cycle may be ascribed to these reactions since they consume some Li ions irreversibly. As a result, the SEI film is composed mainly of (CH2 OCO2 Li)2 and Li2 CO3 . Fig. 2 compares the charge (Li-ion removal) capacity of CoSb3 on the nickel foam and copper foil current collectors. In both the cases, the electrodes exhibit a rapid capacity fade during the initial cycles. As mentioned above, the CoSb3 electrode is subjected to a large volume expansion upon the first discharge due to the large density difference between CoSb3 (7.25 g cm3 )

-1

600

Capacity / mA h g

The CoSb3 intermetallic compound was prepared by levitation melting. The stoichiometric ratio of metallic Sb and Co were put in the copper crucible for evacuation under 0.1 Pa. The metals were levitated in the crucible under an Ar atmosphere and alloyed by the diffusion mixing effect of intensive stirring with the electromagnetic force produced by a high frequency eddy current. The as-prepared alloy ingot was annealed for at 600 °C for 7 days to homogenize it and was artificially ground into 50 lm fine powder. The alloy powder is single phased after annealing as identified by its powder XRD pattern [10]. The powder was further ball-milled to obtain an ultrafine electrode powder for the electrochemical tests. A slurry consisting of CoSb3 ultrafine powder, acetylene black as the conductivity agent, and polyvinylidene fluoride (PVDF) as the binder was pressed onto a nickel foam substrate or a copper foil as the working electrode. The weight ratio of CoSb3 active powder, conductivity agent and binder was 70:15:15. After being dried in vacuum for about 10 h at 120 °C, the electrodes were assembled into the half-cells in an Ar-filled glove box. The cells were mounted by using metallic lithium foil as both the counter electrode and reference electrode. A solution of 50 vol% ethylene carbonate (EC), 50 vol% dimethyl carbonate (DMC) and 1 M LiPF6 was used as the electrolyte solution and a polypropylene (PP) film Cellgard 2300 as the separator. The electrochemical properties of the alloy electrodes were measured by cycling the cells between 0.05 and 1.5 V with a constant current density of 20 mA g1 . EIS was used to characterize the electrode resistance during cycling. The measurements were carried out using a Solartron FRA 1250 frequency response analyzer combined with a Solartron SI 1287 electrochemical interface. The impedance spectra were recorded potentiostatically by applying an ac voltage of 5 mV amplitude in the 200 kHz–5 mHz frequency range after the cells had undergone the desired number of cycles (in the delithiated state) and had been left for 24 h to achieve equilibrium. The microstructure of the electrodes subjected to the desired number of charge and discharge cycles were observed by a Hitachi-S570 scanning electron microscope. The composition changes in the electrolyte solution were investigated by FTIR using a NEXUS670FTIR tester.

2.0

nickel foam current collector copper foil current collector

500 400 300 200 100 0

Fig. 1 shows the voltage profiles of the CoSb3 electrode for the initial four charge and discharge cycles. As shown in the figure, the curve shape of the first discharge process is different from that of the subsequent cycles,

0

2

4

6

8 10 12 14 16 18 20 22 24

Cycle number Fig. 2. Comparison of cycling behavior of CoSb3 electrodes on the nickel foam and copper foil current collectors, respectively.

J. Xie et al. / Journal of Electroanalytical Chemistry 568 (2004) 323–327

and Li3 Sb (3.34 g cm3 ), and the reaction process can be written as:

The results are consistent with those observed by Besenhard et al. [11], who reported that a major expansion in volume occurs in the first discharge process. After pressing, the porous structure in the nickel foam can support CoSb3 particles mechanically and electronically, and the exfoliation of active particles during the first discharge process is less drastic. As a result, the first charge capacity (550 mA h g1 ) of the electrode using the nickel foam substrate is approximately equal to the theoretical capacity (569 mA h g1 ) of CoSb3 alloy. On the contrary, the first charge capacity of the electrode using copper foil is only 413 mA h g1 . This means that the exfoliation of active material during the first discharge is significant. The smooth surface of copper foil may facilitate the exfoliation of active material. Based on the above analyses, we can conclude that the initial rapid capacity loss is caused mainly by the loss of active material. After the initial cycles, the microstructure of the electrode may be stabilized as is evident from the relatively sluggish capacity fade in the subsequent cycles. The shrinkage and the expansion effect become smaller in successive cycles than in the initial cycles, due to the high reversibility in the subsequent charge and discharge cycles. We attribute the capacity fade after the initial cycles to two major factors: (1) the continuous loss of active material, although it occurs at a relatively low rate; (2) the aggregation of the active species to large clusters, which results in an increase of the absolute volume change and loss of interparticle contact upon long-term cycling. The aggregation is an extremely critical drawback for alloy electrodes. It seems to be unavoidable during charge and discharge cycles since it is thermodynamically favorable for small-sized particles to congregate into large clusters to reduce their surface energy. The aggregation of the particles has a significant effect on the cycling stability of the alloy electrode. According to Winter and Besenhard [12], the large particles, because of the high absolute volume change, are more sensitive to cracking and crumbling than small ones. It is reasonable that the aggregation of particles makes the diffusion of Li ions in the active particles kinetically unfavorable. In other words, some of the active species become electrochemically inert toward Li ions. This should be true, because CoSb3 exhibits an improved cycling stability over pure antimony (Fig. 3), even though they follow the same Li-storage mechanism and show similar volume changes upon cycling. As expected, CoSb shows better cycling stability than CoSb3 due to its higher Co content. The better cycling stability can be attributed to the ‘‘divide’’ effect of the Co matrix, which

-1

ð4Þ

pure antimony CoSb CoSb2 CoSb3

500

Capacity / mA h g

CoSb3 þ 9Liþ þ 9e ¼ 3Li3 Sb þ Co

325

400 300 200 100 0 0

2

4

6

8

10

12

14

16 18

20

Cycle number Fig. 3. Comparison of cycling behavior of CoSb3 and pure antimony electrodes.

suppresses the aggregation of the active particles to some degree. Based on this concept, some intermetallic compounds [13–22], which generally contain an electrochemically active center and an inert matrix, are suggested as novel anode materials for secondary lithium ion batteries. The aggregation phenomena are further proved by SEM images. Fig. 4 shows the SEM images of CoSb3 electrodes on the copper foil current collector for freshly prepared electrodes, after 1 cycle and after 30 cycles, respectively. Note that, before cycling, CoSb3 with small particle size is finely dispersed by the acetylene black and is effectively bonded by the PVDF binder as indicated by arrows in Fig. 4(a). It can be seen that the particle size of CoSb3 after ball-milling is about 2 lm as indicated in Fig. 4(a2). After cycling, however, the agglomerate, which consists of small particles, can be seen in Fig. 4(b) and (c). The SEM also shows the exfoliation of the active particles when the surface images between Fig. 4(b1) and (c1) are compared. Fig. 5 shows the EIS plots of the CoSb3 electrode measured for different numbers of cycles (in the delithiated state). Before cycling, a well-defined semicircle and sloping line with a 45° angle can be seen in the plots. The large diameter of the semicircle implies that the alloy electrode has a large inherent charger transfer resistance before electrochemical activation. Two depressed semicircles and a sloping straight line are observed for all the EIS plots after different numbers of charge and discharge cycles. The first semicircle in the high-frequency range is correlated to Li-ion migration through the multilayer surface films, the second one in the medium-frequency range may correspond to the charge transfer between the surface films and the active material, and the sloping straight line at low frequency is characteristic of the solid state diffusion of Li ions in the active material, and at very low frequency, the impedance behavior is capacitive and it reflects the accumulation of lithium [23–25]. It is obvious that the

326

J. Xie et al. / Journal of Electroanalytical Chemistry 568 (2004) 323–327

Fig. 4. SEM images of CoSb3 electrodes on a copper foil current collector in the delithiated state for a fresh electrode (a), after the 1st cycle (b) and after 30 cycles (c), respectively. Images a–c2 are 10 the magnification of a–c1.

150

fresh electrode after 2 cycles after 10 cycles after 20 cycles

-Z'' / Ω

100

50

0 0

50

100

150

200

250

Z' / Ω Fig. 5. EIS plots of the CoSb3 electrode after different cycles in the delithiated state.

diameter of the second semicircle is much larger than that of the first one, indicating that the resistance of the electrode bulk is the major contributor to the total electrode resistance. It is clear that the electrode resistance develops with increasing cycle number, especially for long-term cycling. The increase of the electrode resistance is due mainly to the loss of contact between particles due to the exfoliation and the aggregation of the active particles. The cracks in the SEM images (see Fig. 4(b) and (c)) may be direct evidence of the loss of interparticle contact. The higher the cycle number, the broader are the cracks. The cracks may result from the exfoliation and aggregation of the active particles. On the contrary, no obvious changes of SEI layer resistance can be seen.

J. Xie et al. / Journal of Electroanalytical Chemistry 568 (2004) 323–327

327

LiPF6

δCH δCH

500

νC-O

1000

δCH νC=O

1500

(b)

The work is supported by National Natural Science Foundation of China (No. 50201014) and by PFDP of the Education Ministry of China (No. 20010335045).

(a)

References

νOH

LiPF6

Absorbance

Acknowledgements

νCH

2000

2500

3000 3500

4000

-1

Wavenumbers / cm

Fig. 6. FTIR spectra of electrolyte: (a) fresh solution; (b) after 20 cycles.

It seems that the electrolyte solution may become electrochemically inactive during cycling. This assumption stems from the fact that the originally transparent electrolyte solution becomes turbid and a yellowish surface layer was deposited on the Li counter electrode after continuous cycling, To investigate the effect of electrolyte solution on the electrochemical performances of the CoSb3 electrode, FTIR tests were performed. As indicated in Fig. 6, no obvious changes were observed in the composition of the electrolyte solution after 20 charge and discharge cycles. This agrees well with the EIS results, in which the almost fixed intercept of the high-frequency semicircle on the real axis of the EIS plots is indicative of only slight changes in the composition of the electrolyte solution. 4. Conclusions In our present work, we investigated preliminarily the capacity fade mechanism of the CoSb3 electrode using galvanostatic cycling, SEM, EIS, and FTIR techniques. It is found that the rapid capacity fade in the initial cycles can be attributed to the large volume changes upon lithiation of CoSb3 in the first discharge process and to the structure reconstruction within the electrode, which induce drastic exfoliation of the active material. The relatively low capacity loss in the subsequent cycles can be ascribed mainly to the aggregation of the active materials and to the gradual loss of active material (at a lower rate compared to the initial cycles). We also found that the electrolyte solution has little effect on the electrochemical behavior of the CoSb3 electrode.

[1] D. Aurbach, Y. Ein-Eli, B. Markovsky, Y. Carmeli, H. Yamin, S. Luski, Electrochim. Acta 39 (1994) 2559. [2] D. Aurbach, Y. Gofer, M. Ben-Zion, P. Aped, J. Electroanal. Chem. 339 (1992) 451. [3] D. Aurbach, B. Markovsky, A. Schechter, Y. Ein-Eli, H. Cohen, J. Electrochem. Soc. 143 (1996) 3809. [4] D. Aurbach, Y. Ein-Eli, O. Chusid, M. Babai, Y. Carmeli, H. Yamin, J. Electrochem. Soc. 141 (1994) 603. [5] M. Winter, J.O. Besenhard, M.E. Spahr, P. Novak, Adv. Mater. 10 (1998) 725. [6] D. Aurbach, M.D. Levi, E. Levi, A. Schechter, J. Phys. Chem. B 101 (1997) 2195. [7] G.C. Chung, S.H. Jun, K.Y. Lee, M.H. Kim, J. Electrochem. Soc. 146 (1999) 1664. [8] M. Winter, G.H. Wrodnigg, J.O. Besenhard, W. Biberacher, P. Novak, J. Electrochem. Soc. 147 (2000) 2427. [9] D. Aurbach, B. Markovsky, K. Gamolsky, E. Levi, Y. Ein-Eli, Electrochim. Acta 45 (1999) 67. [10] J. Xie, X.B. Zhao, G.S. Cao, Y.D. Zhong, M.J. Zhao, J. Electroanal. Chem. 542 (2003) 1. [11] J.O. Besenhard, J. Yang, M. White, J. Power Sources 68 (1997) 87. [12] M. Winter, J.O. Besenhard, Electrochim. Acta 45 (1999) 31. [13] I. Rom, M. Wachtler, I. Papst, M. Schmied, J.O. Besenhard, F. Hofer, M. Winter, Solid State Ionics 143 (2001) 329. [14] L. Monconduit, J.C. Jumas, R. Alcantara, J.L. Tirado, C. Perez Vicente, J. Power Sources 107 (2002) 74. [15] K.C. Hewitt, L.Y. Beaulieu, J.R. Dahn, J. Electrochem. Soc. 148 (2001) A402. [16] N. Tamura, R. Ohshita, M. Fujimoto, S. Fujitani, M. Kamino, I. Yonezu, J. Power Sources 107 (2002) 48. [17] O. Mao, J.R. Dahn, J. Electrochem. Soc. 146 (1999) 423. [18] J. Wolfenstine, S. Campos, D. Foster, J. Read, W.K. Behl, J. Power Sources 109 (2002) 230. [19] X.B. Zhao, G.S. Cao, Electrochin. Acta 46 (2001) 891. [20] O. Mao, R.A. Dunlap, J.R. Dahn, J. Electrochem. Soc. 146 (1999) 405. [21] M.M. Thackeray, J.T. Vaughey, C.S. Johnson, A.J. Kroppf, R. Benedek, L.M.L. Fransson, K. Edstrom, J. Power Sources 113 (2003) 124. [22] D. Larcher, L.Y. Beaulieu, O. Mao, A.E. George, J.R. Dahn, J. Electrochem. Soc. 147 (2000) 1703. [23] D. Aurbach, J. Power Sources 89 (2000) 206. [24] D. Aurbach, M.D. Levi, J. Phys. Chem. B 101 (1997) 4641. [25] M.D. Levi, D. Aurbach, Electrochim. Acta 45 (1999) 167.