- Email: [email protected]
Alternative materials for negative electrodes in lithium systems
Solid State Ionics 152 – 153 (2002) 61 – 68 www.elsevier.com/locate/ssi
Review
Alternative materials for negative electrodes in lithium systems Robert A. Huggins * Faculty of Engineering, Christian Albrechts University, Kaiserstrasse 2, D-24143 Kiel, Germany Accepted 14 February 2002
Abstract There is interest in the feasibility of the replacement of the lithium-carbons that are currently employed as negative electrode reactants in lithium systems. Possible improvements might involve the ability to operate safely at higher current densities, less first cycle irreversible capacity loss, better cycling behavior, reduced specific volume, and lower cost. The surprise announcement by Fujifilm on the use of convertible oxide materials in lithium negative electrodes a few years ago has led to growing interest in possible alternative materials. In addition to oxides, other convertible precursors are being considered for this purpose, including metal alloys and semiconductors. A disadvantage of this approach is the need for extra internal lithiumcontaining material within the cell. However, it is also possible to do the conversion of these precursors outside of the electrochemical cell to alleviate this problem. Binary and ternary metal – metal alloys are also being considered as negative reactants, with efforts to reduce volume changes that contribute to capacity losses during cycling. A recent model explains the mechanism of decrepitation and the observation of a critical minimum particle size. This helps explain one of the advantages of the use of very fine particles. A different approach is the consideration of metal – metalloid alloys, including nitrides, borides, and silicides. In some cases, their relatively light weight mean that it is not necessary that they react with large amounts of lithium to achieve attractive specific capacities. This leads to less mechanical distortion, and potentially to better behavior upon cycling. The advantages and disadvantages of these several approaches will be discussed, and an overview of the critical parameters of some of the more attractive candidates is presented. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Negative electrodes; Anodes; Metal – metal alloys; Metal – metalloid alloys; External conversion
1. Introduction The first ambient temperature rechargeable lithium cells used elemental lithium as the negative electrode reactant. Due to severe safety problems and significant loss of capacity during cycling when such electrodes are used with organic solvent liquid electrolytes, this approach has been almost fully discarded. *
Tel.: +49-431-365-10; fax: +49-431-365-38. E-mail address: [email protected] (R.A. Huggins).
Following their commercialization by Sony Energytec in 1991, negative electrodes in the currently available small commercial rechargeable lithiumbased cells typically employ solid solutions of lithium in one or another form of carbon. Up to one atom of lithium per six carbon atoms can be intercalated into graphite, giving a nominal composition of LiC6. This amounts to a maximum theoretical capacity of 372 mA h/g, although the practical capacity values are typically between 300 and 350 mA h/g. There is considerable interest in finding alternative materials that might be more attractive than the lithium-
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 3 3 7 - 5
62
R.A. Huggins / Solid State Ionics 152 – 153 (2002) 61–68
carbons. Improvements might involve the ability to operate safely at higher current densities, less first cycle irreversible capacity loss, reduced loss of capacity upon cycling, smaller specific volume, and lower cost. This paper will discuss a number of strategies that have been pursued, or might be pursued, in the search for new negative electrode reactants.
2. The Fuji approach—convertible oxides The possibility of attractive alternate materials suddenly got attention when Fujifilm announced the use of amorphous metal oxides in negative electrodes [1,2]. During the first charging cycle the oxides are converted to a fine composite microstructure containing fine particles of lithium –metal (e.g. tin) alloys in a matrix of Li2O and glass by reaction with lithium. After this initial irreversible reaction, the electrochemical properties of these electrodes are essentially those of the resulting lithium alloys. This is illustrated in the charge–discharge behavior of a glass containing tin shown in Fig. 1. It is seen that good reversibility is obtained after the first lithiation reaction. Analagous behavior can be expected with a number of other metal oxides that can be reduced at a
Fig. 2. Discharge – recharge curve of SiO after the initial cycle [31].
high lithium activity to form lithium oxide and the corresponding metal. The extra irreversible lithium consumed in the first cycle must be supplied from inside the cell, and this constitutes a serious disadvantage of this approach. In order to reduce the amount of irreversible capacity, it is desirable to use a monoxide, rather than a dioxide. Then only two, rather than four, Li ions are irreversibly consumed in the first cycle. A particularly favorable example is the use of SiO [31,46]. The discharge–recharge behavior after the initial lithiation of this material is illustrated in Fig. 2.
3. Metal – metal alloys
Fig. 1. Example of the behavior of a tin-containing glass electrode [45].
Rather than forming them by the decomposition of an oxide, why not simply consider the use of alloys alone? There should be no initial irreversible capacity related to the formation of lithium oxide. Lithium alloys have been considered as alternatives to elemental lithium for a long time, initially in connection
R.A. Huggins / Solid State Ionics 152 – 153 (2002) 61–68
with high temperature batteries, but later at ambient temperatures. Information about this earlier work can be found in Refs. [3,4]. However, the Fuji announcement has led to significant interest in the possibility of alternate materials, primarily alloys. One obvious reason for the interest in the use of alloys is that some of them can potentially have much larger capacities than can be obtained from lithium-carbons. Even though it is not realistic to expect that these theoretical values can be obtained in commercial cells, they provide a significant incentive to look into this direction further. However, it has been found that metals and metal – metal alloys typically lose capacity rapidly upon cycling. This is associated with the large changes in volume that occur upon reaction with lithium. Volume changes cause strains in the microstructure, and some particles fracture and others are pushed around, so that they can lose electronic contact with the electrode or
63
contact with the electrolyte, and thus do not continue to contribute to the capacity. Another phenomenon also can occur if there are large volume changes upon lithium insertion. This is called decrepitation, or crumbling. Observations on analogous metal hydrides that undergo large volume changes have shown that this phenomenon does not continue indefinitely. Instead, there is a terminal critical particle size that is characteristic of a particular material. Particles with smaller sizes do not continue to fracture. Experiments have shown that the electrochemical cycling behavior of lithium alloy electrodes is much better if the initial particle size is already very small [5], as expected from the critical particle size phenomenon. The mechanism of decrepitation and the influence of the important parameters has been investigated theoretically [6]. Consideration of a simple onedimensional model allows the calculation of the con-
Fig. 3. Influence of the dilation strain parameter upon the critical size below which decrepitation fracture will not occur for materials with different values of the fracture toughness parameter [6].
64
R.A. Huggins / Solid State Ionics 152 – 153 (2002) 61–68
ditions under which fracture will be caused to occur in a two-phase structure due to specific volume mismatch. This model predicts that there will be a terminal critical particle size below which further fracture will not occur. The value of this characteristic dimension is material-specific, depending upon the magnitude of a strain parameter related to the volume mismatch and the fracture toughness of the lowerspecific-volume phase. For the same value of volume mismatch, the tendency to fracture will be reduced and the critical particle size will be larger the greater the toughness of the material. The results of this model calculation are shown in Fig. 3. This means that metallurgical factors can be important. This is consistent with work at Sanyo Electric that showed that the cycle life of lithium cells with Li– Al negative electrodes can be improved by more than a factor of three by alloying that provides matrix hardening [7,8]. It is also possible to consider convertible alloys, in which the initial material decomposes to form other phases upon reaction with lithium. Upon cycling, the properties are determined by those of the decomposition reaction product phase. An example is the reaction of lithium with alloys in the Cu – Sn system [9]. Another interesting approach was recently presented [10] in which binary Sn – M alloys, where M=Co, Ni, In, Pb, were electrodeposited directly onto sheet electrodes. They showed remarkably good capacity and cycling properties. The latter perhaps due to the mechanical constraint of the substrate.
4. Compound semiconductors and metal –metalloid alloys There has also been some interest in the potential use of binary semiconductor compounds with the zinc – blende structure as negative electrode reactants in lithium cells. One example is InSb, and the first paper on this topic was Ref. [11]. Structural changes evidently occur after the insertion of lithium. The details of the reaction mechanism have become controversial, and a different interpretation has been presented recently [12]. Instead of metal –metal alloys or compound semiconductors, alloys of metals and metalloids have
received some attention. These include carbides, nitrides, borides and silicides. 4.1. Nitrides There have been a number of papers on the use of ternary nitrides as negative electrode materials [13 – 20]. One group of these materials has the antifluorite structure, which is known to have high ionic mobility. A number of ternary nitrides Li – M– N where M is a transition metal such as Fe and Mn are known to have this structure [13], and have been investigated as possible negative electrode reactants in lithium cells [19,20]. They show rather good rechargeability, with stoichiometric changes involving Li mol fractions of 0.75 and 1.25, respectively. This leads to specific capacities of about 150 and 200 mA h/g. They have relatively flat charge and discharge curves, with potentials centered about 1.2 to 1.1 V vs. Li. The shape of these curves indicates that they undergo reconstitution reactions. Since it has been reported that the basic symmetry of their structures does not change, the two phases involved are probably different ordered distributions upon the lithium sites. Another group of nitrides is based upon the Li3N structure, which was shown some time ago to be a solid electrolyte with very high lithium ionic conductivity [21 – 24]. Some of the Li ions are replaced by transition metal ions. The best example is Li2.6Co0.4N, which undergoes a structural change when Li is removed during the first cycle, becoming amorphous. Thereafter it has good rechargeability, with a capacity of about 760 mA h/g at an average discharge potential of about 800 mV vs. Li. The discharge – recharge curve after the first cycle is shown in Fig. 4.
Fig. 4. Discharge – recharge curve of Li2.6Co0.4N after initial cycle.
R.A. Huggins / Solid State Ionics 152 – 153 (2002) 61–68
65
4.2. Borides It was shown some time ago that lithium will react with boron at intermediate [25 –29] and elevated [30] temperatures to form a number of intermediate phases. Screening experiments in Kiel [31] on a number of binary borides have shown that lithium does not react with them significantly at ambient temperatures. One exception was the phase SiB3, where a specific capacity of 440 mA h/g was observed. The discharge–recharge curve for this material is shown in Fig. 5. 4.3. Silicides Screening experiments have also been undertaken in Kiel [31] on a number of binary silicides. Most of these have shown that lithium does not react with them significantly at ambient temperatures. On the other hand, Mg2Si, which has the antifluorite structure, is an interesting negative electrode reactant at
Fig. 6. Discharge – recharge curve of Mg2Si [31].
both elevated and ambient temperatures [32 – 34]. It has a capacity over 400 mA h/g with an average potential about 0.35 V vs. Li. The discharge – recharge curve for this material is shown in Fig. 6. Amorphous elemental silicon has also been investigated, and was found to produce a reversible capacity as large as 1020 mA h/g, which is an unusually large value [31].
5. Composite microstructures
Fig. 5. Discharge – recharge curve of SiB3 after initial cycle [31].
Attention is also being given to the use of several types of composite microstructures in electrodes. One is to add an electrochemically inert electronically conducting phase to enhance the contact between the reactant material and the current collector. This typically involves the presence of amorphous, rather than graphitic, carbon. An alternative is to add a metal. For example, Morales and Sanchez [35] added molybdenum to SnO2 electrodes. This was found to increase the cycling behavior. Another recent example was the use of a Si/TiN composite, which showed a rather steady capacity after the initial cycle [36]. Mixed-conductors can act to transport the reacting species as well as enhancing electronic contact. This was originally done at high temperatures [37,38], but has subsequently been demonstrated at ambient temperatures using a non-reactive Li – Sn phase with Li – Cd [39,40] and nonreactive SnFe3C with Sn2Fe [41]. The charge– discharge curve of the lithium – cadmium system in the presence of a fast mixed-conducting phase in the lithium – tin system at a current density
66
R.A. Huggins / Solid State Ionics 152 – 153 (2002) 61–68
Fig. 7. Charge – discharge curve of the lithium – cadmium system in the presence of a fast mixed-conducting phase in the lithium – tin system at a current density of 0.1 mA/cm2 at ambient temperature [39].
of 0.1 mA/cm2 at ambient temperature is shown in Fig. 7. Another alternative is to have two reactive phases present. A recent example of this is the proposal to include some elemental Sn along with SnO2 [42]. The advantage is that there is less oxygen present to produce the initial irreversible capacity. Depending upon the microstructure, it may be possible to improve the cycling properties also in this manner. The inclusion of a solid electrolyte phase in the electrode microstructure is a very old strategy. It can lead to improved contact between the reactant and the electrolyte. A recent example is the inclusion of some Li2O in an electrode along with Sn [43].
current collector, mentioned above [10], is also an interesting variant.
7. Final comments From this abbreviated discussion, we can see that there are a number of different approaches to the question of alternative materials that might be used as negative electrode reactants in lithium cells. None of these is as yet fully investigated, and it is not clear
Table 1 Measured values of average potential and apparent reversible capacity of several materials Material
Capacity (mA h/g)
Average V vs. Li
Reference
Graphite 33% Silicon Mg2Si SiB3 Sn2BPO6 SnO2 Sb Sn glass SiO a-Silicon
300 – 340 300 410 440 450 500 560 650 850 1020
0.15 0.3 0.35 0.3 0.5 0.4 0.95 0.4 0.3 0.3
Practical [36] [31] [31] [45] [45] [31] [2] [46] [31]
6. Alternative configurations The behavior of electrodes may also be improved by the use of configurations different from those normally employed. It was recently shown [44] that the contact between the current collector and the active material can be greatly improved by replacing the normal sheet-shaped current collector by a fine polymer fabric coated with copper. The Sanyo approach to electrodeposited alloys directly onto the
R.A. Huggins / Solid State Ionics 152 – 153 (2002) 61–68
which, if any, will be sufficiently better to displace the carbons currently used. Most investigators place primary emphasis upon the magnitude of the reversible capacity that can be obtained. However, the potential at which this capacity is available is also important, for it influences the voltage of the cell output. The average operating potentials and capacities of a number of materials are shown in Table 1. The operating potential range, and thus the composition range, that is employed can also have a large influence upon the cyclability of many electrodes. An example of this effect is the reaction of lithium with InSb [12]. If the potential is maintained above about 0.65 V vs. Li, a reasonable capacity is obtained (about 250 mA h/g) that does not decay rapidly upon cycling. On the other hand, if the potential is driven lower by further reaction with lithium, a different reaction takes place that substantially reduces the reversibility.
Acknowledgements This work was partially supported by the European Commission under the Joule Program, Contract No. JOR3-CT98-0180.
References [1] Fujifilm, Internet, http://www.fujifilm.co.jp/eng/news _ e/ nr079.html (1996). [2] Y. Idota, et al., Science 276 (1997) 1395. [3] R.A. Huggins, in: J.O. Besenhard (Ed.), Handbook of Battery Materials, Wiley-VCH, New York, 1999, p. 359. [4] R.A. Huggins, J. Power Sources 81 – 82 (1999) 13. [5] J. Yang, M. Winter, J.O. Besenhard, Solid State Ionics 90 (1996) 281. [6] R.A. Huggins, W.D. Nix, Ionics 6 (2000) 57. [7] T. Nohma, S. Yoshimura, K. Nishio, T. Saito, in: G. Pistoia (Ed.), Lithium Batteries; New Materials, Developments and Perspectives, Elsevier, Amsterdam, 1994, p. 417. [8] K. Nishio, N. Furukawa, in: J.O. Besenhard (Ed.), Handbook of Battery Materials, Wiley-VCH, New York, 1999, p. 19. [9] K.D. Kepler, J.T. Vaughey, M.M. Thackeray, Electrochem. Solid-State Lett. 2 (1999) 307. [10] S. Fujitani, Presented at the Third Hawaii Battery Conference, Waikoloa (2001), p. 210. [11] J.T. Vaughey, J. O’Hara, M.M. Thackeray, Electrochem. SolidState Lett. 3 (2000) 13.
67
[12] K.C. Hewitt, L.Y. Beaulieu, J.R. Dahn, Presented at 198th Meeting of the Electro-chemical Society, Phoenix (2000), Abstract No. 128. [13] M. Nishijima, Y. Takeda, N. Imanishi, O. Yamamoto, M. Takano, J. Solid State Chem. 113 (1994) 205. [14] M. Nishijima, N. Tadokoro, Y. Takeda, N. Imanishi, O. Yamamoto, J. Electrochem. Soc. 141 (1994) 2966. [15] M. Nishijima, et al., Solid State Ionics 83 (1996) 107. [16] T. Shodai, S. Okada, S-i. Tobishima, J-i. Yamaki, Solid State Ionics 86 – 88 (1996) 785. [17] M. Nishijima, et al., 8th International Meeting on Lithium Batteries (1996) 402. [18] T. Shodai, et al., 8th International Meeting on Lithium Batteries (1996) 404. [19] Y. Nitta, J. Yamaura, M. Hasegawa, S. Tsutsumi, T. Shodai, Y. Sakurai, Presented at the 196th Meeting of the Electrochemical Society, Honolulu (1999), Abstract No. 236. [20] N. Imanishi, Y. Takeda, O. Yamamoto, in: M. Wakihara, O. Yamamoto (Eds.), Lithium Ion Batteries, Wiley-VCH, New York, 1998, p. 98. [21] B.A. Boukamp, R.A. Huggins, Phys. Lett. 58A (1976) 231. [22] B.A. Boukamp, R.A. Huggins, Mater. Res. Bull. 13 (1978) 23. [23] U. von Alpen, G.H. Talat, A. Rabenau, Appl. Phys. Lett. 30 (1977) 621. [24] A. Rabenau, Adv. Solid State Phys. XVIII (1978) 77. [25] S.D. James, L.E. DeVries, J. Electrochem. Soc. 123 (1976) 321. [26] S. Dallek, D.W. Ernst, B.F. Larrick, J. Electrochem. Soc. 126 (1979) 866. [27] L.E. DeVries, L.D. Jackson, S.D. James, J. Electrochem. Soc. 126 (1979) 993. [28] P. Sanchez, C. Belin, C. Crepy, A. DeGuibert, J. Appl. Electrochem. 19 (1989) 421. [29] P. Sanchez, C. Belin, C. Crepy, A. DeGuibert, J. Mater. Sci. 27 (1992) 240. [30] G. Mair, R. Nesper, H.G. von Schnering, J. Solid State Chem. 75 (1988) 30. [31] A. Netz, R.A. Huggins, W. Weppner, unpublished results (2000). [32] A. Anani, R.A. Huggins, J. Power Sources 38 (1992) 363. [33] R.A. Huggins, A.A. Anani, US Patent 4,950,566, August 21, 1990. [34] C.-K. Huang, S. Surampudi, A.I. Attia, G. Halpert, US Patent 5,294,503, March 15, 1994. [35] J. Morales, L. Sanchez, J. Electrochem. Soc. 146 (1999) 1640. [36] I.S. Kim, P.N. Kumta, G.E. Blomgren, Electrochem. SolidState Lett. 3 (2000) 493. [37] B.A. Boukamp, G.C. Lesh, R.A. Huggins, J. Electrochem. Soc. 128 (1981) 725. [38] B.A. Boukamp, G.C. Lesh, R.A. Huggins, in: H.V. Venkatasetty (Ed.), Proc. Symp. on Lithium Batteries, Electrochem. Soc., 1981, p. 467. [39] A. Anani, S. Crouch-Baker, R.A. Huggins, J. Electrochem. Soc. 135 (1988) 2103. [40] A. Anani, S. Crouch-Baker, R.A. Huggins, in: A.N. Dey (Ed.), Proc. of Symp. on Lithium Batteries, Electrochem. Soc., 1987, p. 382.
68
R.A. Huggins / Solid State Ionics 152 – 153 (2002) 61–68
[41] O. Mao, R.L. Turner, I.A. Courtney, B.D. Fredericksen, M.I. Buckett, L.J. Krause, J.R. Dahn, Electrochem. Solid-State Lett. 2 (1999) 3. [42] J. Wolfenstine, J. Sakamoto, C.-K. Huang, J. Power Sources 75 (1998) 183. [43] D.L. Foster, J. Wolfenstine, W.K. Behl, Electrochem. SolidState Lett. 3 (2000) 203.
[44] A.H. Whitehead, C.M. Hagg, M. Schreiber, Presented at Electrochemical Society Meeting, Phoenix, AZ (2000), Abstract No. 157. [45] I.A. Courtney, J.R. Dahn, J. Electrochem. Soc. 144 (1997) 2943. [46] B. Klausnitzer, PhD Thesis, University of Ulm (2000).
Review
Alternative materials for negative electrodes in lithium systems Robert A. Huggins * Faculty of Engineering, Christian Albrechts University, Kaiserstrasse 2, D-24143 Kiel, Germany Accepted 14 February 2002
Abstract There is interest in the feasibility of the replacement of the lithium-carbons that are currently employed as negative electrode reactants in lithium systems. Possible improvements might involve the ability to operate safely at higher current densities, less first cycle irreversible capacity loss, better cycling behavior, reduced specific volume, and lower cost. The surprise announcement by Fujifilm on the use of convertible oxide materials in lithium negative electrodes a few years ago has led to growing interest in possible alternative materials. In addition to oxides, other convertible precursors are being considered for this purpose, including metal alloys and semiconductors. A disadvantage of this approach is the need for extra internal lithiumcontaining material within the cell. However, it is also possible to do the conversion of these precursors outside of the electrochemical cell to alleviate this problem. Binary and ternary metal – metal alloys are also being considered as negative reactants, with efforts to reduce volume changes that contribute to capacity losses during cycling. A recent model explains the mechanism of decrepitation and the observation of a critical minimum particle size. This helps explain one of the advantages of the use of very fine particles. A different approach is the consideration of metal – metalloid alloys, including nitrides, borides, and silicides. In some cases, their relatively light weight mean that it is not necessary that they react with large amounts of lithium to achieve attractive specific capacities. This leads to less mechanical distortion, and potentially to better behavior upon cycling. The advantages and disadvantages of these several approaches will be discussed, and an overview of the critical parameters of some of the more attractive candidates is presented. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Negative electrodes; Anodes; Metal – metal alloys; Metal – metalloid alloys; External conversion
1. Introduction The first ambient temperature rechargeable lithium cells used elemental lithium as the negative electrode reactant. Due to severe safety problems and significant loss of capacity during cycling when such electrodes are used with organic solvent liquid electrolytes, this approach has been almost fully discarded. *
Tel.: +49-431-365-10; fax: +49-431-365-38. E-mail address: [email protected] (R.A. Huggins).
Following their commercialization by Sony Energytec in 1991, negative electrodes in the currently available small commercial rechargeable lithiumbased cells typically employ solid solutions of lithium in one or another form of carbon. Up to one atom of lithium per six carbon atoms can be intercalated into graphite, giving a nominal composition of LiC6. This amounts to a maximum theoretical capacity of 372 mA h/g, although the practical capacity values are typically between 300 and 350 mA h/g. There is considerable interest in finding alternative materials that might be more attractive than the lithium-
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 3 3 7 - 5
62
R.A. Huggins / Solid State Ionics 152 – 153 (2002) 61–68
carbons. Improvements might involve the ability to operate safely at higher current densities, less first cycle irreversible capacity loss, reduced loss of capacity upon cycling, smaller specific volume, and lower cost. This paper will discuss a number of strategies that have been pursued, or might be pursued, in the search for new negative electrode reactants.
2. The Fuji approach—convertible oxides The possibility of attractive alternate materials suddenly got attention when Fujifilm announced the use of amorphous metal oxides in negative electrodes [1,2]. During the first charging cycle the oxides are converted to a fine composite microstructure containing fine particles of lithium –metal (e.g. tin) alloys in a matrix of Li2O and glass by reaction with lithium. After this initial irreversible reaction, the electrochemical properties of these electrodes are essentially those of the resulting lithium alloys. This is illustrated in the charge–discharge behavior of a glass containing tin shown in Fig. 1. It is seen that good reversibility is obtained after the first lithiation reaction. Analagous behavior can be expected with a number of other metal oxides that can be reduced at a
Fig. 2. Discharge – recharge curve of SiO after the initial cycle [31].
high lithium activity to form lithium oxide and the corresponding metal. The extra irreversible lithium consumed in the first cycle must be supplied from inside the cell, and this constitutes a serious disadvantage of this approach. In order to reduce the amount of irreversible capacity, it is desirable to use a monoxide, rather than a dioxide. Then only two, rather than four, Li ions are irreversibly consumed in the first cycle. A particularly favorable example is the use of SiO [31,46]. The discharge–recharge behavior after the initial lithiation of this material is illustrated in Fig. 2.
3. Metal – metal alloys
Fig. 1. Example of the behavior of a tin-containing glass electrode [45].
Rather than forming them by the decomposition of an oxide, why not simply consider the use of alloys alone? There should be no initial irreversible capacity related to the formation of lithium oxide. Lithium alloys have been considered as alternatives to elemental lithium for a long time, initially in connection
R.A. Huggins / Solid State Ionics 152 – 153 (2002) 61–68
with high temperature batteries, but later at ambient temperatures. Information about this earlier work can be found in Refs. [3,4]. However, the Fuji announcement has led to significant interest in the possibility of alternate materials, primarily alloys. One obvious reason for the interest in the use of alloys is that some of them can potentially have much larger capacities than can be obtained from lithium-carbons. Even though it is not realistic to expect that these theoretical values can be obtained in commercial cells, they provide a significant incentive to look into this direction further. However, it has been found that metals and metal – metal alloys typically lose capacity rapidly upon cycling. This is associated with the large changes in volume that occur upon reaction with lithium. Volume changes cause strains in the microstructure, and some particles fracture and others are pushed around, so that they can lose electronic contact with the electrode or
63
contact with the electrolyte, and thus do not continue to contribute to the capacity. Another phenomenon also can occur if there are large volume changes upon lithium insertion. This is called decrepitation, or crumbling. Observations on analogous metal hydrides that undergo large volume changes have shown that this phenomenon does not continue indefinitely. Instead, there is a terminal critical particle size that is characteristic of a particular material. Particles with smaller sizes do not continue to fracture. Experiments have shown that the electrochemical cycling behavior of lithium alloy electrodes is much better if the initial particle size is already very small [5], as expected from the critical particle size phenomenon. The mechanism of decrepitation and the influence of the important parameters has been investigated theoretically [6]. Consideration of a simple onedimensional model allows the calculation of the con-
Fig. 3. Influence of the dilation strain parameter upon the critical size below which decrepitation fracture will not occur for materials with different values of the fracture toughness parameter [6].
64
R.A. Huggins / Solid State Ionics 152 – 153 (2002) 61–68
ditions under which fracture will be caused to occur in a two-phase structure due to specific volume mismatch. This model predicts that there will be a terminal critical particle size below which further fracture will not occur. The value of this characteristic dimension is material-specific, depending upon the magnitude of a strain parameter related to the volume mismatch and the fracture toughness of the lowerspecific-volume phase. For the same value of volume mismatch, the tendency to fracture will be reduced and the critical particle size will be larger the greater the toughness of the material. The results of this model calculation are shown in Fig. 3. This means that metallurgical factors can be important. This is consistent with work at Sanyo Electric that showed that the cycle life of lithium cells with Li– Al negative electrodes can be improved by more than a factor of three by alloying that provides matrix hardening [7,8]. It is also possible to consider convertible alloys, in which the initial material decomposes to form other phases upon reaction with lithium. Upon cycling, the properties are determined by those of the decomposition reaction product phase. An example is the reaction of lithium with alloys in the Cu – Sn system [9]. Another interesting approach was recently presented [10] in which binary Sn – M alloys, where M=Co, Ni, In, Pb, were electrodeposited directly onto sheet electrodes. They showed remarkably good capacity and cycling properties. The latter perhaps due to the mechanical constraint of the substrate.
4. Compound semiconductors and metal –metalloid alloys There has also been some interest in the potential use of binary semiconductor compounds with the zinc – blende structure as negative electrode reactants in lithium cells. One example is InSb, and the first paper on this topic was Ref. [11]. Structural changes evidently occur after the insertion of lithium. The details of the reaction mechanism have become controversial, and a different interpretation has been presented recently [12]. Instead of metal –metal alloys or compound semiconductors, alloys of metals and metalloids have
received some attention. These include carbides, nitrides, borides and silicides. 4.1. Nitrides There have been a number of papers on the use of ternary nitrides as negative electrode materials [13 – 20]. One group of these materials has the antifluorite structure, which is known to have high ionic mobility. A number of ternary nitrides Li – M– N where M is a transition metal such as Fe and Mn are known to have this structure [13], and have been investigated as possible negative electrode reactants in lithium cells [19,20]. They show rather good rechargeability, with stoichiometric changes involving Li mol fractions of 0.75 and 1.25, respectively. This leads to specific capacities of about 150 and 200 mA h/g. They have relatively flat charge and discharge curves, with potentials centered about 1.2 to 1.1 V vs. Li. The shape of these curves indicates that they undergo reconstitution reactions. Since it has been reported that the basic symmetry of their structures does not change, the two phases involved are probably different ordered distributions upon the lithium sites. Another group of nitrides is based upon the Li3N structure, which was shown some time ago to be a solid electrolyte with very high lithium ionic conductivity [21 – 24]. Some of the Li ions are replaced by transition metal ions. The best example is Li2.6Co0.4N, which undergoes a structural change when Li is removed during the first cycle, becoming amorphous. Thereafter it has good rechargeability, with a capacity of about 760 mA h/g at an average discharge potential of about 800 mV vs. Li. The discharge – recharge curve after the first cycle is shown in Fig. 4.
Fig. 4. Discharge – recharge curve of Li2.6Co0.4N after initial cycle.
R.A. Huggins / Solid State Ionics 152 – 153 (2002) 61–68
65
4.2. Borides It was shown some time ago that lithium will react with boron at intermediate [25 –29] and elevated [30] temperatures to form a number of intermediate phases. Screening experiments in Kiel [31] on a number of binary borides have shown that lithium does not react with them significantly at ambient temperatures. One exception was the phase SiB3, where a specific capacity of 440 mA h/g was observed. The discharge–recharge curve for this material is shown in Fig. 5. 4.3. Silicides Screening experiments have also been undertaken in Kiel [31] on a number of binary silicides. Most of these have shown that lithium does not react with them significantly at ambient temperatures. On the other hand, Mg2Si, which has the antifluorite structure, is an interesting negative electrode reactant at
Fig. 6. Discharge – recharge curve of Mg2Si [31].
both elevated and ambient temperatures [32 – 34]. It has a capacity over 400 mA h/g with an average potential about 0.35 V vs. Li. The discharge – recharge curve for this material is shown in Fig. 6. Amorphous elemental silicon has also been investigated, and was found to produce a reversible capacity as large as 1020 mA h/g, which is an unusually large value [31].
5. Composite microstructures
Fig. 5. Discharge – recharge curve of SiB3 after initial cycle [31].
Attention is also being given to the use of several types of composite microstructures in electrodes. One is to add an electrochemically inert electronically conducting phase to enhance the contact between the reactant material and the current collector. This typically involves the presence of amorphous, rather than graphitic, carbon. An alternative is to add a metal. For example, Morales and Sanchez [35] added molybdenum to SnO2 electrodes. This was found to increase the cycling behavior. Another recent example was the use of a Si/TiN composite, which showed a rather steady capacity after the initial cycle [36]. Mixed-conductors can act to transport the reacting species as well as enhancing electronic contact. This was originally done at high temperatures [37,38], but has subsequently been demonstrated at ambient temperatures using a non-reactive Li – Sn phase with Li – Cd [39,40] and nonreactive SnFe3C with Sn2Fe [41]. The charge– discharge curve of the lithium – cadmium system in the presence of a fast mixed-conducting phase in the lithium – tin system at a current density
66
R.A. Huggins / Solid State Ionics 152 – 153 (2002) 61–68
Fig. 7. Charge – discharge curve of the lithium – cadmium system in the presence of a fast mixed-conducting phase in the lithium – tin system at a current density of 0.1 mA/cm2 at ambient temperature [39].
of 0.1 mA/cm2 at ambient temperature is shown in Fig. 7. Another alternative is to have two reactive phases present. A recent example of this is the proposal to include some elemental Sn along with SnO2 [42]. The advantage is that there is less oxygen present to produce the initial irreversible capacity. Depending upon the microstructure, it may be possible to improve the cycling properties also in this manner. The inclusion of a solid electrolyte phase in the electrode microstructure is a very old strategy. It can lead to improved contact between the reactant and the electrolyte. A recent example is the inclusion of some Li2O in an electrode along with Sn [43].
current collector, mentioned above [10], is also an interesting variant.
7. Final comments From this abbreviated discussion, we can see that there are a number of different approaches to the question of alternative materials that might be used as negative electrode reactants in lithium cells. None of these is as yet fully investigated, and it is not clear
Table 1 Measured values of average potential and apparent reversible capacity of several materials Material
Capacity (mA h/g)
Average V vs. Li
Reference
Graphite 33% Silicon Mg2Si SiB3 Sn2BPO6 SnO2 Sb Sn glass SiO a-Silicon
300 – 340 300 410 440 450 500 560 650 850 1020
0.15 0.3 0.35 0.3 0.5 0.4 0.95 0.4 0.3 0.3
Practical [36] [31] [31] [45] [45] [31] [2] [46] [31]
6. Alternative configurations The behavior of electrodes may also be improved by the use of configurations different from those normally employed. It was recently shown [44] that the contact between the current collector and the active material can be greatly improved by replacing the normal sheet-shaped current collector by a fine polymer fabric coated with copper. The Sanyo approach to electrodeposited alloys directly onto the
R.A. Huggins / Solid State Ionics 152 – 153 (2002) 61–68
which, if any, will be sufficiently better to displace the carbons currently used. Most investigators place primary emphasis upon the magnitude of the reversible capacity that can be obtained. However, the potential at which this capacity is available is also important, for it influences the voltage of the cell output. The average operating potentials and capacities of a number of materials are shown in Table 1. The operating potential range, and thus the composition range, that is employed can also have a large influence upon the cyclability of many electrodes. An example of this effect is the reaction of lithium with InSb [12]. If the potential is maintained above about 0.65 V vs. Li, a reasonable capacity is obtained (about 250 mA h/g) that does not decay rapidly upon cycling. On the other hand, if the potential is driven lower by further reaction with lithium, a different reaction takes place that substantially reduces the reversibility.
Acknowledgements This work was partially supported by the European Commission under the Joule Program, Contract No. JOR3-CT98-0180.
References [1] Fujifilm, Internet, http://www.fujifilm.co.jp/eng/news _ e/ nr079.html (1996). [2] Y. Idota, et al., Science 276 (1997) 1395. [3] R.A. Huggins, in: J.O. Besenhard (Ed.), Handbook of Battery Materials, Wiley-VCH, New York, 1999, p. 359. [4] R.A. Huggins, J. Power Sources 81 – 82 (1999) 13. [5] J. Yang, M. Winter, J.O. Besenhard, Solid State Ionics 90 (1996) 281. [6] R.A. Huggins, W.D. Nix, Ionics 6 (2000) 57. [7] T. Nohma, S. Yoshimura, K. Nishio, T. Saito, in: G. Pistoia (Ed.), Lithium Batteries; New Materials, Developments and Perspectives, Elsevier, Amsterdam, 1994, p. 417. [8] K. Nishio, N. Furukawa, in: J.O. Besenhard (Ed.), Handbook of Battery Materials, Wiley-VCH, New York, 1999, p. 19. [9] K.D. Kepler, J.T. Vaughey, M.M. Thackeray, Electrochem. Solid-State Lett. 2 (1999) 307. [10] S. Fujitani, Presented at the Third Hawaii Battery Conference, Waikoloa (2001), p. 210. [11] J.T. Vaughey, J. O’Hara, M.M. Thackeray, Electrochem. SolidState Lett. 3 (2000) 13.
67
[12] K.C. Hewitt, L.Y. Beaulieu, J.R. Dahn, Presented at 198th Meeting of the Electro-chemical Society, Phoenix (2000), Abstract No. 128. [13] M. Nishijima, Y. Takeda, N. Imanishi, O. Yamamoto, M. Takano, J. Solid State Chem. 113 (1994) 205. [14] M. Nishijima, N. Tadokoro, Y. Takeda, N. Imanishi, O. Yamamoto, J. Electrochem. Soc. 141 (1994) 2966. [15] M. Nishijima, et al., Solid State Ionics 83 (1996) 107. [16] T. Shodai, S. Okada, S-i. Tobishima, J-i. Yamaki, Solid State Ionics 86 – 88 (1996) 785. [17] M. Nishijima, et al., 8th International Meeting on Lithium Batteries (1996) 402. [18] T. Shodai, et al., 8th International Meeting on Lithium Batteries (1996) 404. [19] Y. Nitta, J. Yamaura, M. Hasegawa, S. Tsutsumi, T. Shodai, Y. Sakurai, Presented at the 196th Meeting of the Electrochemical Society, Honolulu (1999), Abstract No. 236. [20] N. Imanishi, Y. Takeda, O. Yamamoto, in: M. Wakihara, O. Yamamoto (Eds.), Lithium Ion Batteries, Wiley-VCH, New York, 1998, p. 98. [21] B.A. Boukamp, R.A. Huggins, Phys. Lett. 58A (1976) 231. [22] B.A. Boukamp, R.A. Huggins, Mater. Res. Bull. 13 (1978) 23. [23] U. von Alpen, G.H. Talat, A. Rabenau, Appl. Phys. Lett. 30 (1977) 621. [24] A. Rabenau, Adv. Solid State Phys. XVIII (1978) 77. [25] S.D. James, L.E. DeVries, J. Electrochem. Soc. 123 (1976) 321. [26] S. Dallek, D.W. Ernst, B.F. Larrick, J. Electrochem. Soc. 126 (1979) 866. [27] L.E. DeVries, L.D. Jackson, S.D. James, J. Electrochem. Soc. 126 (1979) 993. [28] P. Sanchez, C. Belin, C. Crepy, A. DeGuibert, J. Appl. Electrochem. 19 (1989) 421. [29] P. Sanchez, C. Belin, C. Crepy, A. DeGuibert, J. Mater. Sci. 27 (1992) 240. [30] G. Mair, R. Nesper, H.G. von Schnering, J. Solid State Chem. 75 (1988) 30. [31] A. Netz, R.A. Huggins, W. Weppner, unpublished results (2000). [32] A. Anani, R.A. Huggins, J. Power Sources 38 (1992) 363. [33] R.A. Huggins, A.A. Anani, US Patent 4,950,566, August 21, 1990. [34] C.-K. Huang, S. Surampudi, A.I. Attia, G. Halpert, US Patent 5,294,503, March 15, 1994. [35] J. Morales, L. Sanchez, J. Electrochem. Soc. 146 (1999) 1640. [36] I.S. Kim, P.N. Kumta, G.E. Blomgren, Electrochem. SolidState Lett. 3 (2000) 493. [37] B.A. Boukamp, G.C. Lesh, R.A. Huggins, J. Electrochem. Soc. 128 (1981) 725. [38] B.A. Boukamp, G.C. Lesh, R.A. Huggins, in: H.V. Venkatasetty (Ed.), Proc. Symp. on Lithium Batteries, Electrochem. Soc., 1981, p. 467. [39] A. Anani, S. Crouch-Baker, R.A. Huggins, J. Electrochem. Soc. 135 (1988) 2103. [40] A. Anani, S. Crouch-Baker, R.A. Huggins, in: A.N. Dey (Ed.), Proc. of Symp. on Lithium Batteries, Electrochem. Soc., 1987, p. 382.
68
R.A. Huggins / Solid State Ionics 152 – 153 (2002) 61–68
[41] O. Mao, R.L. Turner, I.A. Courtney, B.D. Fredericksen, M.I. Buckett, L.J. Krause, J.R. Dahn, Electrochem. Solid-State Lett. 2 (1999) 3. [42] J. Wolfenstine, J. Sakamoto, C.-K. Huang, J. Power Sources 75 (1998) 183. [43] D.L. Foster, J. Wolfenstine, W.K. Behl, Electrochem. SolidState Lett. 3 (2000) 203.
[44] A.H. Whitehead, C.M. Hagg, M. Schreiber, Presented at Electrochemical Society Meeting, Phoenix, AZ (2000), Abstract No. 157. [45] I.A. Courtney, J.R. Dahn, J. Electrochem. Soc. 144 (1997) 2943. [46] B. Klausnitzer, PhD Thesis, University of Ulm (2000).