Novel types of lithium-ion polymer electrolyte batteries

Novel types of lithium-ion polymer electrolyte batteries

Solid State Ionics 143 Ž2001. 73–81 www.elsevier.comrlocaterssi Novel types of lithium-ion polymer electrolyte batteries G.B. Appetecchi a , F. Croce...

151KB Sizes 2 Downloads 56 Views

Solid State Ionics 143 Ž2001. 73–81 www.elsevier.comrlocaterssi

Novel types of lithium-ion polymer electrolyte batteries G.B. Appetecchi a , F. Croce a , R. Marassi b, S. Panero a , F. Ronci a , G. Savo a , B. Scrosati a,) a

Sezione di Electrochimica, Dipartimento di Chimica, UniÕersity ‘La Sapienza,’ P. le Aldo Moro, 00185 Rome, Italy b Dipartimento di Scienze Chimiche, UniÕersity of Camerino, Italy

Abstract Three new types of polymer electrolyte lithium-ion batteries are presented and discussed. These batteries have been prepared by using gel-type, polyŽacrylonitrile., PAN-based membranes as the electrolyte separators of various electrode ombinations. The latter include ‘standard’ materials such as a graphite Li xC 6 anode and a manganese spinel LiMn 2 O4 cathode, as well as more innovative electrodes such as KC 8 and SnO 2 anodes, and LiCr y Mn 2yyO4 and LiNi y Co 1yyO 2 cathodes. The results, although preliminary, are convincing in demonstrating the feasibility of these new concepts and ultimately, that PAN-based, gel-type membranes are suitable for the development of plastic-like, lithium-ion batteries of practical interest. q 2001 Elsevier Science B.V. All rights reserved. Keywords: Lithium; Electrolyte; Battery

1. Introduction In the round of few years, lithium-ion batteries passed from laboratory bench studies to commercial production. Indeed, until 1992 the consumer electronic market was dominated by the nickel–cadmium and the nickel–metal hydride batteries, while today the majority of the most popular mobile devices, such as cellular phones, relies on lithium-ion batteries w1x. The standard version of these batteries uses a carbonaceous Že.g. graphite or coke. anode, a lithium cobalt oxide Že.g. LiCoO 2 . cathode and a lithium-ion conducting liquid electrolyte Že.g. a solution of a lithium salt in an organic solvent mixture. w2,3x. Although a commercial reality with a production rate of several millions of cells per month, lithium-ion ) Corresponding author. Tel.: q39-06-446-2866; fax: q39-06491-769. E-mail address: [email protected] ŽB. Scrosati..

batteries are still the subject of intense research to further improve their properties and characteristics. Expected advancements in the lithium-ion technology include: Ži. the replacement of the conventional liquidlike electrolyte with a plastic membrane which may act as both the separator and the electrolyte, with the aim of improving battery’s design and reliability; Žii. the replacement of cobalt with nickel or manganese in the cathode structure, with the aim of reducing costs and environmental impact; Žiii. the replacement of graphite with other low-voltage Li-accepting compounds, with the aim of improving safety characteristics. Many attempts to reach these goals are presently underway worldwide. The polymer electrolyte con-

0167-2738r01r$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 2 7 3 8 Ž 0 1 . 0 0 8 3 5 - 9

74

G.B. Appetecchi et al.r Solid State Ionics 143 (2001) 73–81

cept was proposed and demonstrated by the Bellcore Laboratory in the US w4x and recently announced to be exploited for the commercial production by various battery companies in Japan w5x. Many R & D projects are currently in progress to demonstrate the feasibility of cathodes other than LiCoO 2 , such as LiM y Mn 2yyO4 Žwhere M s Li, Cr, Al, Ga, . . . . w6–8x and LiNi y Co Ž1yy.O 2 w9,10x. Finally, consistent research is underway for the development and the characterization of anodes alternative to graphite, such as metal or intermetallic oxides w11–14x. A research program on various lithium-ion battery materials is in progress in our laboratory as well. In particular, we have carried out a systematic study on various types of lithium-ion, gel-type polymer electrolytes formed by the immobilization of liquid solutions in a series of polymer matrices w15–20x. Work has also been directed to the characterization of new anode w21–23x and cathode materials w24x. Having acquired experiences in both electrolyte and electrode materials, we have proceeded with the testing of their reciprocal combinations. In particular, we have considered the use of our gel polymer electrolytes with alternative anodes, such as KC 8 or SnO 2 , and with alternative cathodes, such as LiCr y Mn 2yyO4 and LiNi y Co Ž2yy.O 2 , for the fabrication of innovative polymer lithium-ion battery structures. Three examples of these novel types of lithium-ion polymer electrolyte batteries are reported and discussed in this work.

2. Experimental The gel-type electrolyte membranes used here as separators in our advanced lithium-ion batteries have been prepared by immobilizing a solution of a lithium salt Že.g. LiClO4 or LiPF6 in an organic solvent mixture Že.g. an ethylene carbonate–dimethyl carbonate, EC–DMC mixture. in a polymer matrix Že.g. a polyŽacrylonitrile. PAN matrix. w25x. The preparation procedure involved a sequence of steps Žall carried out inside an environmentally controlled dry box. which included: Ži. the dissolution at room temperature of the selected lithium salt in the EC– DMC organic solvent mixture; Žii. the addition of the PAN polymer component and its homogeneous dispersion in the above solution by stirring for several

hours at room temperature; Žiii. the transfer of the so-formed slurry on an aluminum plate preheated at 908C and its stay at this temperature for a very short time Žabout 30 s. for promoting a fast but yet complete dissolution; Živ. the cooling of the new solution to room temperature to favor cross-linking in the membrane and gel formation. Dimensionally stable, elastomeric membranes, having a thickness ranging between 100 and 200 mm, were typically obtained with this procedure. For reasons of simplicity, the membranes are hereafter indicated by writing in sequence the lithium salt, the liquid solvent mixture, and the immobilising polymer, i.e. with the notations LiClO4 –EC–DMC–PAN or LiPF6 –EC–DMC–PAN. The composition of the two membranes Žin molar ratio. was: 4.5:56.3:23.0:16.0 and 4:60:20:16, respectively. The potassium-doped carbon KC 8 anodes were prepared by placing a graphite electrode Ž95 wt.% graphite with 5 wt.% PVDF binder addition, coated on a copper current collector. in a reaction tube, sealed under vacuum together with potassium metal. Complete conversion of graphite into KC 8 occurs after annealing for a few days at 2508C w21x. Tin oxide and SnO 2 powder samples were prepared by mixing an aqueous alkaline NH 4 OH solution with SnCl 4 . The precipitate was annealed at 8008C for about 4 h in air w23x. The tin oxide-based anodes were prepared by mixing the SnO 2 powder with 5 wt.% acetylene black ŽAC. and 5 wt.% polyvinylchloride ŽPVC. binder to form a slurry, which was sprayed onto a copper current collector substrate. The chromium-stabilized manganese spinel cathode samples, LiCr y Mn 2yyO 2 Žwhere y was typically 0.02., were kindly provided by Prof. M. Wakihara of Tokyo Institute of Technology in Japan. A very similar compound, kindly provided by Covalent Associates, was also used. The nickel-substituted lithium cobalt oxide cathode, LiCo y Ni 1yyO 2 Žwhere y was typically 0.2., was a Merck battery grade product. The cathode films were fabricated by spreading on an aluminum foil current collector a mixture of either one of these active components Žtypically around 90 wro.: acetylene black Ž5 wro. and a polyŽvinyl chloride. –polyŽtetra ethylene fluoride. PVC–PTFE binder Ž5 wro., using tetrahydrofuran as the carrier solvent.

G.B. Appetecchi et al.r Solid State Ionics 143 (2001) 73–81

The lithium-ion battery prototypes were assembled in an environmentally controlled dry box by pressing a laminate formed by the selected film anode, the electrolyte membrane and the selected film cathode. Occasionally, a third Žusually a lithium metal strip. electrode was inserted to act as a reference. Typical cell thickness was in the order of 200 mm. The cells were housed into appropriately sealed metal or plastic containers. The sealed cells were finally removed from the dry box for the electrochemical tests. These mainly included cycling voltammetry, charge and discharge curves, as well as cycling response. Established electrochemical instrumentation, e.g. EG and G, PAR potentiostatrgalvanostat and MACCOR cycler, was used for the measurements.

3. Results and discussion The main requirements for a successful operation of a plastic lithium-ion battery is the availability of a polymer electrolyte having a lithium-ion conductivity approaching that of common liquid electrolytes. We believe that the gel-type membranes formed by the combination of a liquid solution and a polyŽacrylonitrile. PAN polymer fulfill this requirement. Accordingly, this type of electrolyte membranes was used in all the batteries developed in this work. We have tested a variety of compounds as electrode materials, which include a ‘standard’ graphite Li x C 6 anode and a standard manganese spinel LiMn 2 O4 cathode, as well as more innovative electrode materials, such as the KC 8 and SnO 2 anodes, and the LiCr y Mn 2yyO4 and the LiNi y Co 1yyO 2 cathodes. The aim was to fabricate and test novel types of lithium-ion polymer batteries. Three significative examples of these batteries are described below.

75

trode by monitoring its potential–capacity response during the lithium intercalation–deintercalation process Ži.e. the x Liqq 6C q xeyl Li x C 6 process. in the present LiClO4 –EC–DMC–PAN electrolyte cell; Fig. 1 illustrates the result. The response is similar to that expected for liquid electrolytes, i.e. the voltage profile of the Li-intercalation process develops along a series of plateaus which corresponds to a progressive staging process and a similar trend is reproduced in the following Li-deintercalation process, with a total cyclable capacity which approaches the theoretical 372 mA h gy1 Ž x s 1. value w2,3x. As typical in liquid electrolyte cells, also in this case there is a formal excess capacity in the course of the initial charging cycle, this being due to the formation of the well-known passivating film on the graphite electrode surface w2,3x. Due to the protective action of this film, the charge–discharge efficiency of the graphite film electrode approaches 100% after the second cycle and remains at this value for the subsequent cycles. The favourable electrode kinetics of the graphite electrode are also demonstrated by the cyclic voltammetry results as shown in Fig. 2, which evidences the series of reversible intercalation–deintercalation peaks related to the progressive intercalation stages in the graphite structure. We have then proceeded with the evaluation of the response of the other electrode, i.e. the LiCr0.02 Mn 1.98 O4 composite film cathode. Fig. 3 illustrates the result w26x. Again the response is similar

3 .1 . T h e C r L iC lO 4 – E C – D M C – P A N r LiCr0 .0 2 Mn 1.9 8 O4 lithium-ion polymer battery The determination of the compatibility of electrode materials with the PAN-based membrane electrolyte was a priority test for evaluating the battery’s feasibility. In this regard, we have first measured the electrochemical response of the graphite film elec-

Fig. 1. Voltage vs. capacity response of a graphite-film electrode in a LiClO4 –EC–DMC–PAN gel-electrolyte cell: 1st cycle. Room temperature; counter electrode: lithium metal; cycling rate: Cr4.

G.B. Appetecchi et al.r Solid State Ionics 143 (2001) 73–81

76

3.2. The KC8 r LiClO4 –EC–PC–PAN r LiMn 2 O4 lithium-ion polymer battery

Fig. 2. Slow scan rate cyclic voltammetry of the reversible anodic and cathodic cycles of a graphite-film electrode in a LiClO4 –EC– DMC–PAN gel electrolyte. Room temperature; scan rate: 0.1 V sy1 .

to that expected in liquid- electrolyte cells both in terms of voltage profile of the Li-deintercalation–intercalation process and of total cyclable capacity. Once having ascertained the compatibility of the LiClO4 –EC–DMC–PAN polymer electrolyte with the electrodes, we proceeded with the test of the complete battery prepared by a sequential lamination of the electrode and the electrolyte film components w26x. The expected electrochemical process of this battery is the cyclic transfer of lithium-ions from the manganese spinel electrode to the graphite electrode, i.e.

A second example of novel lithium-ion polymer battery configuration was obtained by combining a KC 8 anode with a ‘standard’ LiMn 2 O4 composite cathode film. The basic electrochemical behavior of the KC 8 intercalation electrode was reported and described in previous papers w21,22x. It was demonstrated that upon anodic polarization, this electrode irreversibly deintercalates potassium, resulting in a graphite-like electrode that, on subsequent cycles, performs with enhanced fast kinetics. The first concern was again that of testing the response of the KC 8 in PAN-based polymer electrolyte cells. Fig. 5 shows the incremental capacity vs. potential, d qrd e curve, determined for the first anodic deintercalation–cathodic intercalation curves of the KC 8 electrode in this type of cells. The response is very similar to those obtained for the same electrode in liquid electrolytes. In fact, as in the liquid case, the first anodic curve clearly shows the peak corresponding to the irreversible deintercalation of potassium ions. Once these ions are released, only lithium-ions intercalate back, and thus the electrochemical process becomes associated with the reversible lithium intercalation–deintercalation in and out the graphite structure. This is clearly demonstrated by the series of the following reversible peaks, clearly recognizable as those of the staging intercalation process of graphite electrodes. Once

6C q LiCr0.02 Mn 1.98 O4lLi x C 6 q Li Ž1yx .Cr0.02 Mn 1.98 O4 .

Ž 1.

Fig. 4 shows the voltage profile of the initial two cycles of the battery reported, in terms of percentage of graphite full capacity vs. cycle number. The trend of the 1st cycle reveals the expected initial irreversible capacity due to the cited formation of the passivating film on the graphite electrode surface. Once the film is formed, the charge–discharge efficiency of the battery reaches 100% with a discharge delivery, which approaches 80% of graphite full capacity.

Fig. 3. Voltage vs. capacity response of a LiCr0.02 Mn 1.98 O4 composite film electrode in a LiClO4 –EC–DMC–PAN gel-electrolyte cell; 1st cycle. Room temperature; counter electrode: lithium metal; cycling rate: Cr10.

G.B. Appetecchi et al.r Solid State Ionics 143 (2001) 73–81

77

Fig. 4. Voltage vs. percentage of graphite full capacity of the initial charge–discharge cycles of a CrLiClO4 –EC–DMC–PANr LiCr0.02 Mn 1.98 O4 plastic-type lithium-ion battery. Room temperature; cycling rate: Cr12.5.

having ascertained its favorable electrochemical response, the KC 8 electrode was tested as anode in a polymer lithium-ion cell using a LiClO4 –PC–EC– PAN membrane electrolyte and a LiMn 2 O4 composite cathode. For this type of battery, assuming a 1:1

ratio KC 8rLiMn 2 O4 , the first activation process is expected to be, KC 8 ™ Kqq eyq 8C

Ž 2.

at the anode and, LiMn 2 O4 q eyq Liql Li 2 Mn 2 O4

Ž 3.

at the cathode. The charge balance in the electrolyte is assured by the fact that the amount of Liq intercalated into LiMn 2 O4 is compensated by the amount of Kq deintercalated from KC 8 . The following discharge process and all of the other cycles involve the cycling of Liq ions between the two electrodes, according to the typical Li-ion process: 8C q Li 2 Mn 2 O4 l 4r3LiC 6 q Li 2r3 Mn 2 O4

Fig. 5. Incremental capacityrpotential curve of the KC 8 electrode in the PAN–PC–EC–LiClO4 gel-like polymer-electrolyte membrane at room temperature.

Ž 4.

where the maximum achievable specific capacity, 372 mA h gy1 , is referred to the anode. It is to be noticed that the use of a cathode containing an excess of lithium assures that all the available graphite, resulting from the deintercalation of potassium, can be used to form LiC 6 . The above listed processes are confirmed by the observed response of the complete KC 8rLiMn 2 O4 polymer electrolyte battery w27x. Fig. 6 Župper part. shows the voltage profile of the battery and those of

78

G.B. Appetecchi et al.r Solid State Ionics 143 (2001) 73–81

the lithium-rich manganese oxide spinel phase, and in the final part somewhat extending into the 4-V region due to the partial LiMn 2 O4 excess. Subsequently, the KC 8rLiMn 2 O4 polymer electrolyte battery maintains an appreciable response by delivering a good fraction Ži.e. about 80% at 0.1 mA cmy2 and about 65% at 0.2 mA cmy2 . of the theoretical capacity referred to graphite w27x.

Fig. 6. Voltage profile of the KC 8 rPAN-basedrLiMn 2 O4 plastic lithium-ion battery. Upper part: first charge ‘activation’ and first discharge cycle; lower part: steady-state charge–discharge cycle. The single-electrode voltage profiles are also shown.

the single electrodes during the first charge cycle: the initial potassium deintercalation Žprocess Ž2.., the subsequent lithium intercalation at the anode, and the corresponding lithium deintercalation at the cathode Žprocess Ž3.. are all clearly shown. The voltage profile of the following steady-state cycle is shown in Fig. 6 Žlower part. and in agreement with scheme Ž4.. The process at the anode is now that typical of a Li x C 6 electrode, and the battery voltage plateau develops around the expected 3-V characteristic of

3.2.1. The SnO 2 r LiClO4 – EC – DMC – PAN r LiNi 0 .8 Co0 .2 O2 lithium-ion polymer battery. Considerable interest is presently devoted to crystalline andror amorphous metal oxides w28x, such as tin oxides and SnO y w29,30x, as alternative to graphite in lithium-ion batteries. These oxides operate via the formation and dissolution of a lithium alloy rather than by the lithium intercalation–deintercalation reaction, which is characteristic of most carbonaceous anodes. The alloy-forming decomposition reaction, in principle, offers a much higher specific capacity than that of carbon intercalation, i.e. 710 A h gy1 vs. 370 A h gy1 in the case of Li 4.4 Sn vs. LiC 6 . However, metal electrodes such as Sn cannot be repeatedly cycled due to the large volume changes which accompany alloy formation w31x. These changes, which may extend up to 300%, cause progressive cracking of the metal particles and thereby induces losses of contact between them. Therefore, although greatly appealing in terms of storage capacity, lithium alloy electrodes are difficult to use in practice. The replacement of the corresponding oxides as starting electrode materials has in part solved this problem. In fact, when an oxide, e.g. SnO 2 , is negatively polarized in a lithium cell, it undergoes a first irreversible reaction: SnO 2 q 4Li ™ Sn q 2Li 2 O,

Ž 5.

which leads to Sn metal particles which are finely dispersed in the Li 2 O matrix. This produces a favorable morphology, which facilitates the subsequent reversible lithium alloy formation process w28x, e.g, Sn q 4.4Li l Li 4.4 Sn.

Ž 6.

Apparently, the lithium oxide, by surrounding the tin particles, creates a sufficient amount of free volume to accommodate the mechanical stresses experienced by the metal during the alloy formation– decomposition process. This greatly improves the

G.B. Appetecchi et al.r Solid State Ionics 143 (2001) 73–81

79

Fig. 7. Cyclic voltammogram of a SnO 2 electrode in a LiClO4 –EC–DMC–PAN electrolyte. Lithium-counter and lithium-reference electrodes; room temperature; scan rate: 0.1 mV sy1 .

cycling performance albeit some capacity fade, due to the occurrence of tin particle aggregation, may still be observed over prolonged, i.e. several hundreds of cycle, tests w32x. Therefore, the use of convertible oxides may effectively lead to the practical use of alloy-type electrodes. Indeed, this has been practically demonstrated w29x, but to our knowledge, always in liquid-electrolyte cells. Thus, we have recognized the particular interest to combine this material with a polymer electrolyte, namely with an electrolyte medium, that, by its plastic nature, is expected to accommodate the volume changes of the contacting electrodes w23x. Therefore, in this third example we report the characteristics of a PAN-based polymer electrolyte lithium-ion battery, using a SnO 2 anode and a LiNi 0.8 Co 0.2 O 2 composite film cathode w23x. The overall process of this battery is expected to be: SnO 2 q yLiNi 0.8 Co 0.2 O 2™ 2Li 2 O q Sn q yLi yy 4r y Ni 0.8 Co 0.2 O 2 ;

Ž 7.

Sn q yLi yy 4r y Ni 0.8 Co 0.2 O 2lLi x Sn q yLi yy Ž4qx . Ni 0.8 Co 0.2 O 2 .

Ž 8.

As in the previous cases, the compatibility of the polymer electrolyte with the electrode materials was first investigated. Fig. 7 shows a typical cyclic voltammogram of a thin-film, SnO 2 electrode in a LiClO4 –EC–DMC–PAN electrolyte cell. The trend is qualitatively similar to those reported for the same electrode in liquid electrolytes w32x, namely, a first irreversible cycle at about 0.8 V vs. Li Žassociated with process Ž7.., followed by a series of peaks, which remain reproducible for all subsequent cycles Žassociated with process Ž8... The electrode compatibility study has been extended to the cathodic side, i.e. to the LiNi y Co 2yyO 2 electrode. Fig. 8, which reports the voltage trends of typical Li-deintercalation–intercalation cycles in a LiClO4 –EC–DMC–PAN polymer cell, clearly demonstrates that the behavior of this electrode is as expected, i.e. a cyclability at a voltage ranging around 4.0 V vs. Li, and a delivery of a capacity averaging 120 mA h gy1 w9,10x. The SnO 2 electrode was firstly activated into Li x Sn by a single discharge–charge cycle in a cell with a Li anode, then removed and used as anode in combination with the LiNi 0.8 Co 0.2 O 2 cathode in a LiClO4 –EC–DMC–PAN electrolyte cell. The cell was not optimized in terms of anode-to-cathode balance, but contained excess cathode capacity Ž y in Eq. Ž8.. in order to assure full cycling operation of

80

G.B. Appetecchi et al.r Solid State Ionics 143 (2001) 73–81

Fig. 8. Typical Li deintercalation–intercalation cycles of a LiNi 0.75 Co 0.25 O 2 electrode in a LiClO4 –EC–DMC–PAN polymer cell. Lithium counter electrode; room temperature; cycling rate: Cr10.

the Li x Sn electrode. Fig. 9 shows the voltage profile of a typical discharge–charge cycle of this cell, along with anode and cathode voltage excursions w23x. Clearly, the Li x Sn anode cycles, with a trend comparable to that usually obtained in more-conven-

tional liquid electrolyte cells w32x, deliver a reversible capacity of ; 250 mA h gy1 . This capacity level is retained upon further cycling albeit with a slightly progressive fade; this again being in similarity to the trend typically experienced in liquid-electrolyte cells.

Fig. 9. Typical voltage profile of a discharge–charge cycle of a SnO 2rLiClO4 –EC–DMC–PANrLiNi 0.8 Co 0.2 O 2 lithium-ion cell. The single anode and cathode voltage profiles are also shown. Room temperature; cycling rate: 0.25 mA cmy2 ; lithium reference. The capacity is referred to the SnO 2 anode.

G.B. Appetecchi et al.r Solid State Ionics 143 (2001) 73–81

4. Conclusions The main purpose of this work was to demonstrate the general operational capabilities of novel types of PAN-based electrolyte lithium-ion polymer batteries. In this preliminary study we have used prototypes which were far from being optimized in terms of composition and morphology of the electrolytes. Indeed, some decay in capacity was observed upon prolonged cycling tests, suggesting that only a fraction of the electrode masses, probably that contained in the external layers, was effectively involved in the electrochemical process, as typically expected for not-uniform electrode configurations. In addition, the mass ratio between anode and cathode also was not properly balanced in these preliminary, demonstrative prototypes, and this factor was also critical in influencing their cycling performance. Having clarified this, we believe that the results of the present work are promising in suggesting that PAN-based, gel-type polymer electrolytes are indeed suitable for the development of plastic-like, lithiumion batteries of practical interest. Present efforts are directed to the optimization of the structure of the electrodes and electrolytes in order to further substantiate this prevision.

Acknowledgements The authors would like to thank Prof. M. Wakihara of the Tokyo Institute of Technology, Japan, for having kindly provided the Cr-stabilized manganese spinel samples, and Prof. M. Mastragostino of the University of Palermo, for having kindly provided the LiMn 2 O4 manganese-spinel samples. Financial support from the following sources: NEDO ŽInternational Joint Research Program., Japan; C.N.R. ŽNational Council of Research, Progetto Finalizzato ‘Materiali Speciali per Tecnologie Avanzate II,’ contract nos. 97.00908.PF34 and 97.00952.34 and MURST cofinaziamento 1998, is also acknowledged.

References w1x M. Wakihara, O. Yamamoto ŽEds.., Lithium Ion Batteries, Kodansha, Tokyo, 1998, p. 218.

81

w2x S. Megahed, B. Scrosati, J. Power Sources 51 Ž1994. 79. w3x S. Megahed, B. Scrosati, Interface 4 Ž1995. 34. w4x F. Shokoohi, P.C. Warren, S.J. Greaney, M. Tarascon, A.S. Gozdz, G.C. Amatucci, Proceedings 35th Power Sources Conference, Cherry Hill, New Jersey, June 1996. w5x http:rrwww.mbi.panasonic.co.jp w6x M.M. Thackeray, J. Electrochem. Soc. 142 Ž1195. 2558. w7x L. Guohua, H. Ikuda, T. Uchida, M. Wakihara, J. Electrochem. Soc. 143 Ž1996. 178. w8x G. Pistoia, A. Antonini, R. Rosato, C. Bellitto, G.M. Ingo, Chem. Mater. 9 Ž1997. 1443. w9x C. Delmas, I. Saadounne, Solid State Ionics 53–56 Ž1992. 370. w10x U. Ueda, T. Ohzuku, J. Electrochem. Soc. 141 Ž1994. 2010. w11x J.O. Basenhard, J. Yang, M. Winter, J. Power Sources 68 Ž1997. 87. w12x T. Brousse, R. Retoux, U. Herterich, D.M. Schleich, J. Electrochem. Soc. 145 Ž1998. 1. w13x O. Mao, R.A. Dunlap, J.R. Dahn, J. Electrochem. Soc. 146 Ž1999. 405. w14x M.M. Thackeray, J.T. Vaughey, A.J. Kahaian, K.D. Kepler, R. Benedek, Electrochem. Comm. 1 Ž1999. 111. w15x G.B. Appetecchi, F. Croce, G. Dautzenberg, F. Gerace, S. Panero, F. Ronci, E. Spila, B. Scrosati, Gazz. Chim. Ital. 126 Ž1996. 415. w16x G.B. Appetecchi, F. Croce, F. Gerace, S. Panero, E. Spila, B. Scrosati, Gazz. Chim. Ital. 127 Ž1997. 325. w17x G.B. Appetecchi, F. Croce, B. Scrosati, J. Power Sources 66 Ž1997. 77. w18x G.B. Appetecchi, B. Scrosati, Electrochim. Acta 43 Ž1998. 1105. w19x G. Dautzenberg, F. Croce, S. Passerini, B. Scrosati, Chem. Mater. 6 Ž1994. 538. w20x G.B. Appetecchi, F. Croce, P. Romagnoli, B. Scrosati, Electrochem. Comm. 1 Ž1999. 83. w21x R. Tossici, M. Berrettoni, V. Nalimova, R. Marassi, B. Scrosati, J. Electrochem. Soc. 143 Ž1996. L64. w22x R. Tossici, M. Berrettoni, M. Rosolen, R. Marassi, B. Scrosati, J. Electrochem. Soc. 144 Ž1997. 186. w23x S. Panero, G. Savo, B. Scrosati, Electrochem. Solid-State Lett. 2 Ž1999. 365. w24x G.B. Appetecchi, F. Croce, R. Marassi, L. Persi, P. Romagnoli, B. Scrosati, Electrochim. Acta 45 Ž1999. 23. w25x B. Scrosati, in: M. Wakihara, O. Yamamoto ŽEds.., Lithium Ion Batteries, Kodansha and Wiley, Weinheim, 1998, p. 218. w26x G.B. Appetecchi, B. Scrosati, Denki Kagaku 66 Ž1998. 1299. w27x S. Sconocchia, R. Tossici, R. Marassi, F. Croce, B. Scrosati, Electrochem. Solid-State Lett. 1 Ž1998. 159. w28x I.A. Courtney, J.R. Dahn, J. Electrochem. Soc. 144 Ž1997. 2045. w29x Y. Idota, T. Kubota, A. Matsufuji, Y. Maekawa, T. Miyasaka, Science 276 Ž1997. 1395. w30x Y. Idota, M. Mishima, M. Miyaki, T. Kubota, T. Miyasaka, US Patent 5,618,640, April 8 Ž1997.. w31x R.A. Huggins, Solid State Ionics 113–115 Ž1998. 57. w32x T. Brousse, S.M. Lee, L. Pasquerau, D. Defives, D.M. Schleich, Solid State Ionics 113–115 Ž1998. 51.