All solid lithium polymer batteries with a novel composite polymer electrolyte

All solid lithium polymer batteries with a novel composite polymer electrolyte

Solid State Ionics 159 (2003) 97 – 109 www.elsevier.com/locate/ssi All solid lithium polymer batteries with a novel composite polymer electrolyte Qi ...

740KB Sizes 0 Downloads 46 Views

Solid State Ionics 159 (2003) 97 – 109 www.elsevier.com/locate/ssi

All solid lithium polymer batteries with a novel composite polymer electrolyte Qi Li a, Takahito Itoh a, Nobuyuki Imanishi a, Atsushi Hirano a, Yasuo Takeda a, Osamu Yamamoto b,* a

Department of Chemistry for Materials, Faculty of Engineering, Mie University, Tsu, Mie 514-8507, Japan b Genesis Research Institute, Inc., Noritake-Shinmachi, Nishi-ku, Nagoya, 451-0051, Japan Received 2 September 2002; accepted 12 December 2002

Abstract A composite polymer electrolyte based on polyethylene oxide (PEO) with a hyperbranched polymer poly[bis(triethylene glycol)benzoate] capped with an acetyl group (HBP) and a ceramic filler, BaTiO3, was examined as the electrolyte in rechargeable lithium polymer batteries. The conductivity of the composite polymer electrolyte PEO – 10 wt.% HBP with Li(CF3SO2)2N – 10 wt.% LiPF6 as a lithium salt and 10 wt.% BaTiO3 was found to be 1.6  10 4 S/cm at 25 jC and 1.5  10 3 S/cm at 60 jC in a O/Li ratio of 10. The lithium rechargeable batteries consisted of this highly conductive composite polymer electrolyte and the 4 V class cathode, LiNi0.8Co0.2O2, showed excellent charge – discharge cycling performance. The initial cathode discharge capacity of 154 mA h/g declined only 0.1%/cycle during the first 30 cycles at 60 jC. D 2003 Elsevier Science B.V. All rights reserved. Keywords: Composite solid polymer electrolyte; Ionic conductivity; Aluminum corrosion; Cell performance

1. Introduction Recently, batteries for electric vehicles (EV) and engine/battery hybrid vehicles (HEV) have been the subject of active research and development. At this moment, nickel metal hydride batteries (Ni-MH) have an advantageous position as the batteries for EV and HEV because of its high reliability and good safety over competing lithium ion batteries. The disadvantages of Ni-MH are the low-energy density as well as

* Corresponding author. Tel.: +81-586-77-8820; fax: +81-58681-1885. E-mail address: [email protected] (O. Yamamoto).

the high cost. The low-energy density is of the most serious problem for EV application. Another candidate for the batteries in EV and HEV is lithium batteries, which have a potential of two or three times higher energy density and higher cost performance over NiMH [1,2]. While the lithium-ion batteries with liquid electrolytes have been seen significant progress over recent 10 years, all solid-state lithium batteries based on solid electrolytes have attracted much attention because of their advantages over lithium-ion batteries such as improved safety and easy of fabrication. In spite of the advantages of solid electrolyte batteries over liquid electrolyte batteries, the inherent low mobility of ionic species in solids has been the major limitation of the solid electrolyte batteries. The lith-

0167-2738/03/$ - see front matter D 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0167-2738(03)00004-3

98

Q. Li et al. / Solid State Ionics 159 (2003) 97–109

ium-ion conductivity solid electrolytes have been studied for more than 30 years. Many types of an inorganic solid electrolyte with a high lithium conductivity have been found in the oxide, nitride, and sulfide systems [1]. The highest conductivity of 10 3 S/cm at room temperature was found in Li3N by Alpen et al. [3]. However, the low decomposition potential of Li3N limits the applications for lithium batteries. The oxide systems showed a low lithium-ion conductivity at low temperatures. Some lithium sulfide glasses show a high lithium-ion conductivity of 10 3 S/cm at room temperature and unity lithium-ion transport number [4,5]. Cho and Liu [6] proposed a glass polymer composite electrolyte for lithium batteries. The composite, containing of a (0.56Li2S – 0.19B2S3 – 0.25LiI) glass and a PEO6Li(CF3SO2)2N polymer (13 vol.%), showed the conductivity of 3  10 4 S/cm at 25 jC and 1.4  10 3 S/cm at 80 jC. However, sulfide glass is extremely hydroscopic and it is a little difficult to make a thin composite film with a high conductivity (low content of polymer). The thickness of the conductivity measurement samples of the glass composite polymer was as thick as 0.6– 0.9 mm. The high conductivity and high lithiumion transport number in these composite electrolytes are quite attractive for the electrolyte in lithium batteries. Practical applications of the glass polymer composite electrolyte in the lithium batteries are for further study. At present, polymer lithium electrolytes are the most promising electrolyte for all solid lithium batteries because of the comparatively high conductivity, the ease to make a thin film, and the excellent mechanical and chemical stability. After Wright [7] reported that polyethylene oxide (PEO) with alkaline metal salts exhibits significant ionic conductivity, Armand et al. [8] recognized the potential of these materials in lithium batteries. During these two decades, many polymer systems have been reported as the lithium-ion conductor. PEO-based polymer electrolytes, which was first proposed by Wright, still keep the most promising position as the electrolytes in lithium batteries. The PEO-based electrolyte shows an acceptable conductivity only at a higher temperature. The basic structure of PEO – LiX polymer electrolytes involves PEO chains coiled around the Li+ cation, thereby separating them from the X counteranion [9]. Therefore, polymer electrolyte require local relax-

ation and segmental motion of the solvent (PEO) chain to allow ion transport, and this condition can only be obtained when the polymer is in an amorphous state, i.e., above 60 jC in PEO. Large research efforts have been devoted to enhance the low-temperature conductivity. The addition of a ceramic filler is effective to increase the low-temperature conductivity as well as the mechanical properties [10]. In the previous paper, we have reported that the ferroelectirc materials BaTiO3 in PEO is effective to increase the electrical conductivity, especially at a lower temperature, and to decrease the interface resistance between the lithium anode and the polymer electrolyte [11]. More recently, Itoh et al. [12] have found that the composite polymer electrolytes of PEO – BaTiO3 – Li(CF3SO2)2N with hyperbranched polymer poly[bis (triethylene glycol)benzoate] with terminal acetyl group (HBP) show a high electrical conductivity at room temperature as high as 10 4 S/cm. In this study, the high-conductivity composite polymer electrolytes with HBP have been examined as the electrolyte in the 4 V class lithium rechargeable batteries with the LiNi0.8Co0.2O2 cathode and the lithium metal anode. As the addition of 10 wt.% LiPF6 to PEO – Li(CF3SO2)2N – BaTiO3 enhanced the corrosion potential of aluminum and improved cycling performance [13], the composite lithium salt of 90 wt.% Li(CF3SO2)2N – 10 wt.% LiPF6 was used in the electrolyte.

2. Experimental A hyperbranched polymer poly[bis(triethylene glycol)benzoate] with terminal acetyl groups (HBP) was prepared according to the method reported previously [14] and its chemical structure is shown in Fig. 1, where a unit inside the bracket represents the repeat unit of the polymer that contains nine oxygen atoms. The structure of obtained HBP was confirmed by the IR and the NMR analysis. The molecular weight of HBP was determined by the gel permeation chromatography to be about 15,000. The obtained samples were dried completely under vacuum (0.1 Torr) at 60 jC for 24 h. The PEO-based composite polymer electrolytes with HBP were obtained by a solvent-casting technique using acetonitrile (AN) as a carrier solvent [15].

Q. Li et al. / Solid State Ionics 159 (2003) 97–109

99

Fig. 1. Molecular structure of poly[bis(triethylene glycol) benzoate] capped with an acetyl group (HBP).

High molecular weight PEO (Aldrich Chemical, 6  105 average molecular weight), Li(CF3SO2)2N (Li-imide) (Fluka Chemical), and LiPF6 (Aldrich Chemical) were used as received. Barium titanate (Sakai Chemicals, 0.5 Am average particle size) was dried under vacuum at 150 jC for 24 h. Preparation of the composite polymer electrolyte films involved the dispersion of the BaTiO3 powder and lithium salt mixture of Li-imide –10 wt.% LiPF6 in AN followed by the addition of PEO and HBP. The slurry was completely homogenized and then cast onto a flat polytetrafluroroethlene vessel. The solvent in the slurry was allowed to evaporate slowly under flow of nitrogen gas for 24 h at room temperature. Finally, the composite polymer electrolyte films were dried at 65 jC under vacuum for 48 h. These procedures yielded homogenous and mechanically stable membranes with an average thickness of about 200 Am. The electrical conductivity of the composite polymer electrolyte films and the interfacial resistance between the electrolyte and the electrode (Li metal anode and composite cathode) were measured by an ac impedance method using a Solartron 1260 frequency analyzer. Stainless steel-blocking electrodes were used for conductivity measurements, and symmetrical nonblocking lithium (or cathode composite electrode) electrodes were used to investigate interfacial phenomena. A 10 mV ac amplitude was applied, and the data were collected by recording 10 points/ decade over a frequency range from 1 Hz to 10 MHz. All solid cells were fabricated by stacking the lithium metal anode, the composite cathode, and the composite polymer electrolyte in a coin type cell. The composite cathode was prepared as follows: proper

amount of LiNi0.8Co0.2 (Kishida Chem., Japan) and acetylene black (Denkikagaku, Japan) were added to a (PEO – HBP) – (Li-imide – 10 wt.% LiPF6) and 10 wt.% BaTiO3 slurry in AN, and the mixture was strongly stirred for 24 h before casting on the aluminum substrate. After the cathode composite material was dried under vacuum at 100 jC, it was pressed into a thin film of about 40 Am thick. The weight ratio of LiNi0.8Co0.2O2 to acethylene black to the composite polymer electrolyte in the composite cathode was 65:20:15. The charge – discharge tests of the cells were performed galvanostatically at a constant current and at a regulated cutoff voltage. The current density was calculated from the active cathode area (0.5 cm2). The active area of the lithium anode was around 1.0 cm2, and the anode capacity was several times higher than that of the cathode. Mechanical properties of the composite polymer electrolytes were measured by use of a TP-101 tension tester (Sekegu, Japan) at 10 cm/min stretching speed in a temperature range 25 –80 jC. These measurements were carried out in a dry box. Thermal analysis was performed using a differential scanning calorimeter (DSC-8230, Rigaku). Samples (10 –20 mg) were hermetically sealed in Al pans in a glove box. Measurements were performed under a nitrogen gas flow from – 50 jC to 200 jC at a 10 jC/min heating rate.

3. Results and discussion Recently, the PEO-based composite polymer electrolyte with HBP and with BaTiO3 as the filler was

100

Q. Li et al. / Solid State Ionics 159 (2003) 97–109

found to show a high conductivity at lower temperatures by Itoh et al. [12]. The highest conductivity of 2.6  10 4 S/cm at 30 jC and 5.2  10 3 S/cm at 80 jC was found in the system of (PEO – 20 wt.% HBP)12(Li-imide) – 10 wt.% BaTiO3. The conductivity value is quit attractive for the electrolyte in the polymer lithium batteries. Sun et al. [11] reported the dependence of the salt on the electrical conductivity in the PEO – LiX system and found that Li-imide in the PEO – LiX system showed the high conductivity. However, the aluminum current collector undergoes serious corrosion in carbonate-based electrolyte solutions containing Li-imide. Yang et al. [16] found that a protective film formed on the aluminum surface when LiPF6 and LiBF4 was used as the salt in propylencarbonate (PC) and could also inhibit corrosion in Li-imide – PC. We have found that the addition of 10 wt.% LiPF6 in the PEO – Li-imide – BaTiO3 electrolyte showed good corrosion resistance to the aluminum current collector as observed in liquid electrolytes [13]. In this study, the electrical conductivities of the (PEO –HBP)– BaTiO3 composite polymer electrolyte with Li-imide and LiPF6 as the lithium salts have been examined. Fig. 2 shows the temperature dependence of the electrical conductivity as a function of O(PEO repeat unite)/Li ratios (a scale of

Fig. 2. Temperature dependence of electrical conductivity in [(PEO – 10 wt.% HBP)x(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3 and [PEO10(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3.

the concentration of lithium salt) in the polymer, where the content of HBP was fixed as 10 wt.% because the polymer with a high content of HBP was slightly sticky as shown in mechanical properties of this composite polymer. To compare the effect of the addition of HBP, the electrical conductivity of PEO10(Li-imide –10 wt.% PF6)– 10 wt.% BaTiO3 is shown also in Fig. 2. The electrical conductivity of the (PEO – HBP)10(Li-imide –  % LiPF6) – BaTiO3 system decreased with increasing the content of LiPF6. The addition of 10 wt.% LiPF6 was effective to enhance the aluminum corrosion resistance [13]. The conductivities of 1.6  10 4 at 25 jC and 4  10 3 S/cm at 80 jC in the polymer with 10 wt.% HBP were compared those of 1  10 5 at 25 jC and 7.4  10 4 at 80 jC in that without HBP, respectively. The conductivity maximum is found at the intermediate concentration of salt of O/Li = 10. This behavior is often explained in terms of the trade-off between increasing number of charge carries and ion migration and increased viscosity due to ionic cross-linking. The activation energy for conduction in the composite polymer electrolyte with HBP was calculated from Arrhenius plots in Fig. 2 to be 38.5 kJ/mol in a higher temperature range 50– 80 jC and 63.8 kJ/mol in a lower temperature range 25 – 50 jC. In the higher temperature range, the activation energy in the composite polymer with HBP is comparable that of 41.2 kJ/mole without HBP is comparable. In the lower temperature range, the activation energy of the polymer with HBP (O/Li = 10) is considerable lower than that of the polymer without HBP (O/Li = 10), which was estimated to be 137 kJ/mol. The low activation energy for conduction in the composite polymer electrolyte with HBP at lower temperature could be explained by the formation of amorphous phase near room temperature. DSC experiments were performed to determine the thermal behavior in the composite polymer electrolytes. Typical DSC traces of the composite polymer electrolytes with and without HBP are shown in Fig. 3. The traces of composite polymers with and without HBP show an endothermic peak at around 40 jC, which corresponds to the melting point of PEO with Li-imide –LiPF6. The melting enthalpy of the composite polymer with HBP was estimated to be 11 J/g compared to 31 J/g for that of without HBP. The low melting enthalpy suggests that the composite polymer

Q. Li et al. / Solid State Ionics 159 (2003) 97–109

101

Fig. 3. DSC traces of [(PEO – 10 wt.% HBP)10(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3 and [PEO10(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3.

with HBP is more amorphous near room temperature. An exothermic peak appears at around 150 jC, which can assigned to the decomposition of LiPF6. It was a little difficult to determine an exact glass transition point in this system. DSC curves showed a trace peak around –39 jC, which may be correspond to the glass transition. The electrical conduction in polymer electrolytes allows in an amorphous state [9]. Therefore, the rate of crystallization of the composite polymer electrolytes is an important factor to keep the high conductivity. Some composite polymer electrolytes with oxide filers showed a significant aging effect [17,18] The time dependence of the electrical conductivity of [(PEO – 10 wt.% HBP)10(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3 at room temperature is shown in Fig. 4, where the sample had no preheat treatment. No significant conductivity decline was observed within 300 h. In the characterization of the polymer electrolyte for lithium batteries, it is important not only to measure the ionic conductivity but also to determine the lithium-ion transport number. In this study, we measured the transport number of the lithium ion by a combination of ac impedance and dc polarization

measurements using a symmetrical cell having nonblocking lithium electrodes [19]. The temperature dependence of the lithium-ion transport number for

Fig. 4. Time dependence of electrical conductivity in [(PEO – 10 wt.% HBP)10(Li-imide – 10 wt.% LiPF6) – 10 wt.% BaTiO3 at 25 jC.

102

Q. Li et al. / Solid State Ionics 159 (2003) 97–109

(PEO – 10 wt.% HBP)10(Li-imide– 10 wt.% LiPF6) – 10 wt.% BaTiO3 is shown in Fig. 5. The transport number of the composite polymer without HBP was almost same to that for the polymer with HBP. The observed ionic transport number of about 0.1 is comparable that reported previously in the system of PEO – Li-imide [11]. The lithium-ion transport number slightly increases with increasing temperature. The mechanical properties of the polymer electrolyte are one of the important properties in the practical battery applications. Generally, polymers become softer by adding low molecular weight polymer to high molecular weight polymer PEO. The soft polymer electrolyte, such as the gel-type polymer, tends to easily form dendrite lithium in a charge– discharge cycling. The tensile strength of the composite polymer electrolytes was measured by a conventional tension tester in a temperature range from room temperature to 70 jC. Table 1 shows the tensile strengths at 100% elongation and the broken points for the composite polymer along those of PEO. The tensile strength of PEO decreased by dissolving a lithium salt and by the addition of HBP. However, the addition of 10 wt.% HBP brought a minimal decrease. The composite polymer can be handled easily. The main requisite for success in polymer electrolyte batteries is the optimization and control of the

Table 1 Tensile strength of the composite polymer electrolyte films at various temperatures Samples

Temperature 100% Broken (jC) Elongation point (MPa) (MPa)

[(PEO – 10 wt.% HBP)10 (Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3

30 40 50 60 70 30 60

2.7 1.7 1.4 0.56 0.33 3.4 0.6

40 60 60

0.33* 0.26* 0.67

[(PEO – 10 wt.% HBP)25 (Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3 [(PEO – 20 wt.% HBP)12 (Li-imide)] – 10 wt.% BaTiO3 [PEO10(Li-imide)] – 10 wt.% BaTiO3 PEO – 10 wt.% BaTiO3 PEO

30 60 30

3.6

1.3

17 10 14

* Maximum strength (yield point).

electrode/polymer electrolyte interface [20]. In some cases, the interface resistance between the electrolyte and the electrode was more than 10 times higher than that of the electrolyte. The interfacial impedance was measured for the cells, the Li/composite polymer electrolyte/Li, and the cathode mixture/composite polymer electrolyte/cathode mixture, annealed at 60 –80 jC (Table 2). The cathode mixture consisted of LiNi0.8Co0.2O2:polymer electrolyte:acetylene black (65:25:15 weight ration). Typical Cole –Cole plots of a symmetric cell Li/(PEO – 10 wt.% HBP) 10 (Liimide– 10 wt.% LiPF6)– 10 wt.% BaTiO3/Li at 60 jC are shown in Fig. 6. We have observed two semicircles: first, a high-frequency small semicircle and a low-frequency, large distorted semicircle. The Table 2 Interfacial resistances of cathode composite/composite electrolyte and lithium metal/composite electrolyte after annealed for 10 days

Fig. 5. Temperature dependence of lithium ion transport number in [(PEO – 10 wt.% HBP)10(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3.

Temperature

Cathode (V cm2)

Anode (V cm2)

80 70 60 50

15 40.3 62.5 85.2

62.5 175 500 1500

Cathode composite: LiNi0.8Co0.2O2:compsoite polymer electrolyte:acetylene black (70:15:15 weight ratio). Composite electrolyte: [(PEO – 10 wt.% HBP)10(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3.

Q. Li et al. / Solid State Ionics 159 (2003) 97–109

103

Fig. 6. Cole – Cole plots of the impedance response of the Li/[(PEO – 10 wt.% HBP)10(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3/Li cell at 60 jC. Electrode area: 0.5 cm2.

amplitude of the second one expands consistently with time. By fitting the semicircles trend with a proper equivalent circuit [21], one can refine the

analysis to obtain the bulk resistance, Rb, the interfacial resistance of the film, Ri, and the charge transfer resistance, Rct. Fig. 7 shows time dependence of these

Fig. 7. Time dependence of the bulk resistance, Rb, the interface resistance, Ri, and the charge transfer resistance, Rct, at 60 jC for the Li/[(PEO – 10 wt.% HBP)10(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3/Li cell.

104

Q. Li et al. / Solid State Ionics 159 (2003) 97–109

components at 60 jC. The bulk resistance is almost constant with time. The interface resistance increased with annealing time. The charge transfer resistance also increases with time, the value of which is comparable to that observed in the cell Li/PEO8Li(CF3 SO2)2N – 10 wt.% g –LiAlO2/Li at 70 jC by Borghini et al. [22]. The increase of Rct with time may be an increase of the coverage of the high-resistance film on lithium metal. The interface resistance is due to the formation of a high-resistance film at the interface of the composite electrolyte and the lithium electrode, which significantly decreased with increasing temperature. Fig. 8 shows the temperature dependencies of the impedance spectra for the cell Li/(PEO –10 wt.% HBP)10(Li-imide– 10 wt.% LiPF6)– 10 wt.% BaTiO3/ Li. The lithium interfacial resistances (including charge transfer resistance) after annealing at 80 and 60 jC for 240 h are about 140 and 500 V cm2, respectively. Appetecchi et al. [20] observed the lithium interface resistance of PEO12LiCF3SO3 –10 wt.% gLiAlO2 to be 200 V cm2 at 85 jC after annealing at 85 jC for 20 days. Our interfacial resistance results at 80 jC are comparable to these results. The lithium interfacial resistances increase

considerably with decreasing temperature. It may be due to the high activation energy for conduction in the interface film because the high-resistance film shows a high activation energy for conduction. The interfacial resistance of 800 V cm2 at 60 jC is acceptable at a low current drain as less than mA/cm2; the contribution of the cell voltage drop by the interface resistance is less than 1 V. To operate at a lower temperature, the lithium interfacial resistance will restrict the cell performance. In the lithium-ion batteries, the cathode interface resistance plays an important role for the long-term stability. Recently, Amine et al. [23] reported that the area-specific impedance of a lithium-ion cell, which consisted of a LiNi0.8Co0.2O2 positive electrode, a carbon negative electrode, and a LiPF6 in EC – DEC electrolyte, increased significantly from 40 V cm2 in the fresh cell to 100 V cm2 after 2 weeks of testing at 70 jC. They concluded that the bulk of the impedance rise is mainly due to the positive electrode. In the case of the composite polymer electrolyte, the interfacial resistance annealed at 70 jC was increased from 27.5 to 40.3 V cm2 for 240 h. The interfacial resistance is comparable to that in the lithium-ion cell.

Fig. 8. Cole – Cole plots of the impedance response of the Li/[(PEO – 10 wt.% HBP)10(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3/Li cell at various temperature after annealing for 240 h. Electrode area: 0.5 cm2.

Q. Li et al. / Solid State Ionics 159 (2003) 97–109

The layered phase LiNi0.8Co0.2O2 is an attractive cathode material for rechargeable lithium batteries because of its high specific capacity and its low cost as compared to LiCoO2 [24]. The Co doping level of LiNi0.8Ni0.2O2 can provide the best electrochemical properties [25].The LiNi0.8Co0.2O2 cathode shows a high initial capacity of 180 mA h/g and a capacity decrease with cycling to 160 mA h/g after 40 cycles at a cutoff voltage of 3.0 – 4.2 V. The capacity depends on the cutoff voltage. On previous studies of all solid-state batteries, the low-voltage cathode materials, such as VOx [26], LixMnO2 [27], and Cu0.1V2O5 [28], have been proposed. High cell voltage cathode materials, such as LiCoO2, LiNiO2, and LiMn2O4, have a much lower reversible lithium intercalation capacity in solid polymer electrolyte cells [29]. The poor rechargeability was assumed to result from the low decomposition potential compared to that of conventional liquid electrolyte. Xia et al. [27] estimated the decomposition potential of a typical polymer electrolyte to be 3.8 vs. Li/Li+. Also, Appetechhi et al. [20], Appetechhi et al. [28] and Borghini et al. [22] have reported that the decomposition process was observed at roughly 3.7 V vs. Li/Li+ in PEO20LiBF4 – 20 wt.% g-LiAlO2 by the sweep voltammetry. The decomposition potential of PEO-based polymer electrolyte depended on the lithium salt in PEO. The decomposition potential of PEO – LiClO4 – 1.4 wt.% BaTiO3 [30] and PEO – LiCF3SO3 – 10 wt.% g –LiAlO2 [22] was estimated to be more than 4.0 V vs. Li/Li+. These decomposition potentials were estimated by the sweep voltammetry. There are some uncertainties with regard to the determination of decomposition voltage of polymer electrolyte through this method. In a previous paper [31], the decomposition voltage of PEO19 – Li-imide – 10 wt.% BaTiO3 at 80 jC was assumed to be less than 4.0 V vs. Li/Li+ by the cyclic performance in a Li/electrolyte/LiNi0.8Co0.2O2 cell. The cell did not show a significant capacity fade at a cutoff voltage of 2.5 –3.9 V and a significant capacity loss at a cutoff voltage 2.5 – 4.0 V. In Fig. 9, cyclic voltammetry curves at 80 jC in Li/[(PEO –10 wt.% HBP)10(Liimide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3/Al are shown. The aluminum current collector undergoes serious corrosion in carbonate-based electrolyte solutions containing Li-imide. We have observed two peaks in the voltammetry curves. The first one

105

Fig. 9. Cyclic voltammetry curves at 80 jC of Li/[(PEO – 10 wt.% HBP)10(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3/Al. Scan rate: 10 mV/min.

corresponds to the aluminum dissolution. That peak is diminished by several cycling. Yang et al. [32] reported that a protective film is formed on the aluminum surface when LiPF6 and LiBF4 were used as salt in propylencarbonate (PC) and that could also inhibit corrosion in Li-imide/PC. In the case of PEO/ Li-imide, the aluminum dissolution voltage was enhanced up to 4.5 V vs. Li/Li+ by addition of 10 wt.% LiPF6 [13]. The second peak corresponds to the voltage of the decomposition of the composite polymer electrolyte, the value of which is estimated to be more than 4.0 V vs. Li/Li+. In Fig. 10, the charge–discharge cyclic performances in the Li/composite polymer electrolyte/LiNi0.8 Co0.2O2/Al cell at 60 jC are shown for the composite electrolyte with LiPF6 and without LiPF6, where the cutoff voltage was 2.5– 4.2 V. No significant capacity fade by cycling was observed at a cutoff voltage of 2.5– 3.9 V for the electrolyte with LiPF6 and without LiPF6. As shown in Fig. 10, a significant capacity fade by cycling is observed for the cell with the electrolyte without LiPF6 at a cutoff voltage of 4.2 – 2.5 V. The addition of LiPF6 in the polymer electrolyte is quite effective to expand the cutoff voltage. The protective film formed on the aluminum surface inhibited corrosion of aluminum by Li-imide, or a film formed on the lithium electrode depressed the decomposition of PEO. The charge– discharge cyclic performances for the cell Li/[(PEO – 10 wt.% HBP)10(Liimide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3/LiNi0.8 Co0.2O2/Al at 60 jC are shown as a function of the

106

Q. Li et al. / Solid State Ionics 159 (2003) 97–109

Fig. 10. Charge and discharge cycling performances of Li/[(PEO – 10 wt.% HBP)10(Li-imide) – 10 wt.% BaTiO3/LiNi0.8Co0.2O2 and Li/[(PEO – 10 wt.% HBP)10(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3/LiNi0.8Co0.2O2 at 60 jC. Current density: 0.2 mA/cm2.

cutoff voltage in Fig. 11. No significant capacity fade is observed by charging up to 4.4 V. On the other hand, a significant capacity fade is observed by

Fig. 11. Cutoff voltage dependence on the cathode capacity by cycling in Li/[(PEO – 10 wt.% HBP)10(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3/LiNi0.8Co0.2O2/Al at 60 jC. Current density: 0.2 mA/cm2.

charging up to 4.5 V. The increase of capacity in the first several cycles may be due to a good contact between the cell components. The total cell resistance (see Fig. 13) was less than 1000 V cm2, and the drop by cell resistance was less than 0.4 V. These results suggest that the decomposition potential of the composite polymer electrolyte is higher than 4.0 V at 60 jC. The capacity of LiNi0.8Co0.2O2 in the composite cathode, 150 mA h/g, is lower than that observed in the liquid electrolyte by a cutoff voltage of 4.2 – 2.5 V which is 180 mA h/g [13]. In the liquid electrolyte, the applied voltage is almost the same to the sum of the anode and the cathode electrode potentials because the voltage drop by the cell resistance is negligible. In the case of a cutoff voltage of 3.9– 2.5 V, the capacity of LiNi0.8Co0.2O2 was about 130 mA h/g. The capacity of 150 mA h/g corresponded to the cutoff voltage of 2.5 – 4.05 V. That is, LiNi0.8Co0.2O2 was charged up to 4.05 V vs. Li/Li+. Similar charge – discharge cycling performances were obtained at 50 jC. In Fig. 12, the cathode discharge capacity dependence on cycling are shown as a function of the cutoff voltage. The capacity of the cathode is slightly lower than that at 60 jC because of

Fig. 12. Cutoff voltage dependence on the cathode capacity by cycling in Li/[(PEO – 10 wt.% HBP)10(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3/LiNi0.8Co0.2O2/Al at 50 jC. Current density: 0.1 mA/cm2.

Q. Li et al. / Solid State Ionics 159 (2003) 97–109

Fig. 13. Cyclic performance of the cell Li/[(PEO – 10 wt.% HBP)10(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3/LiNi0.8 Co0.2O2/Al at 60 jC. Current density: 0.2 mA/cm2. Cutoff voltage: 2.5 – 4.4 V.

the high cell resistance. A stable capacity of 140 mA h/g was obtained at a current density of 0.1 mA/cm2. At a lower temperature such as 40 jC, the cell

107

resistance was observed to be more than 3000 V cm2 and the cathode capacity was found to be 140 mA h/g at a current density of 0.02 mA/cm2. The long cyclic performance of the all solid polymer batteries is an important factor for the batteries in EV and HEV. Recently, Croce et al. [33] have reported a long-life polymer electrolyte battery with the lithium metal anode, the LiFePO4 cathode, and the (PEO)20 LiCF 3 SO 3 – 10 wt.% Al 2 O 3 electrolyte. The cell showed an excellent cyclic performance at 105 jC. The initial cathode capacity of 140 mA h/g declined to 100 mA h/g after 260 charge– discharge cycling. However, low-temperature cell performance has not been reported. In Fig. 13, the long cyclic performance of the cell Li/[(PEO – 10 wt.%)10(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3/LiNi0.8Co0.2O2/Al at 60 jC is shown, where the cutoff voltage was 4.4 –2.5 V and the current density was 0.2 mA/cm2 (0.57 C). The initial cathode capacity of 150 mA h/g is declined to 74 mA h/g by 410 charge– discharge cycling. The capacity fade by cycling could be explained with the increasing cell resistance by cycling as shown in Fig. 14. The capacity fade rate is about 0.12%/cycling. To improve the cycling performance, the stability of the electrolyte/ electrode interface should be improved.

Fig. 14. Cole – Cole plots of impedance response for the cell, Li/[(PEO – 10 wt.% HBP)10(Li-imide – 10 wt.% LiPF6)] – 10 wt.% BaTiO3/ LiNi0.8Co0.2O2/Al at 60 jC before and after cycling tests. Open cell voltage = 3.2 V.

108

Q. Li et al. / Solid State Ionics 159 (2003) 97–109

4. Conclusion A novel lithium is conducted as polymer electrolyte with hyperbranched polymer. This polymer electrolyte showed a high conductivity of 10 4 S/ cm at room temperature. All solid lithium polymer batteries with the polymer electrolyte, the LiNi0.8 Co0.2O2 cathode, and the lithium anode have been examined at lower temperatures such as 50 –60 jC. A good cyclic performance is observed in the cell with the lithium anode, the composite polymer electrolyte [(PEO – 10 wt.% HBP)10(Li-imide – 10 wt.% LiPF6)] –10 wt.% BaTiO3, and the composite cathode of LiNi0.8Co0.2O2. The LiNi0.8Co0.2O2 cathode capacity of 150 mA h/g was obtained in the polymer cell which is comparable to that in liquid electrolyte cells. The energy density in the practical batteries using the composite electrolyte, the lithium anode, and the LiNi0.8Co0.2O2 cathode (30% of the calculated one from the anode and the cathode capacity) is estimated as high as 200 W h/g, the value of which clear the long-term target of the United States Advanced Battery Consortium (USABC), 150 –200 W h/kg [2]. The capacity fade by cycling was a little higher as 0.12%/cycle. The USABC target is 1000 cycles by 20% capacity degradation. The most important problem in this battery for EV and HEV applications is how to improve the cyclic performance. The cyclic performance can be improved by revising the electrode and the electrolyte interface with the addition of a filler or additives into the polymer electrolyte. It is now the course of our research.

Acknowledgements This work was carried out under the collaboration program of the Mie University and the Genesis Research Institute.

References [1] J.B. Goodenough, in: M. Wakihara, O. Yamamoto (Eds.), Lithium Ion Batteries, Fundamentals and Performance, Kodansha and Wiley – VCH, Tokyo, 1998, p. 1.

[2] M. Anderman, F.R. Kalhammer, D. MacArthur, Advanced Batteries for Electric Vehicles: An Assessment of Performance, Cost and Availability, State of California Air Resources Board, Sacrament, CA, 2000. [3] U.V. Alpen, A. Rabenau, G.H. Talat, Appl. Phys. Lett. 30 (1977) 621. [4] J.H. Kennedy, Z. Zhang, J. Electrochem. Soc. 131 (1994) 2441. [5] K. Takada, N. Aotani, K. Iwamoto, S. Kondo, Solid State Ionics 79 (1995) 284. [6] J. Cho, M. Liu, Electrochim. Acta 42 (1997) 1481. [7] P.V. Wright, Br. Polym. J. 7 (1975) 319. [8] M.B. Armand, J.M. Chabagno, M. Duclot, Extended Abstracts Second International Conference on Solid Electrolytes, St. Andrews, 1978. [9] P. Lightfoot, M. Metha, P.G. Bruce, Science 362 (1993) 883. [10] F. Croce, G.B. Appetecchi, L. Persi, B. Scrosati, Nature 394 (1998) 456. [11] H.Y. Sun, Y. Takeda, N. Imanishi, O. Yamamoto, H.-J. Sohn, J. Electrochem. Soc. 147 (2000) 2462. [12] T. Itoh, Y. Ichikawa, Y. Miyamura, T. Uno, M. Kubo, O. Yamamoto, Solid State Ionics 150 (2002) 337. [13] Q. Li, N. Imanishi, A. Hirano, Y. Takeda, O. Yamamoto, J. Power Sources 110 (2002) 38 – 45. [14] T. Itoh, M. Ikeda, N. Hirata, Y. Moriya, M. Kubo, O. Yamamoto, J. Power Sources 81/82 (1999) 824. [15] Q. Li, H.Y. Sun, Y. Takeda, N. Imanishi, J. Yang, O. Yamamoto, J. Power Sources 94 (2001) 201. [16] H. Yang, K. Kwon, T.M. Devine, J.W. Evans, J. Electrochem. Soc. 147 (2000) 4399. [17] F. Capuano, F. Croce, B. Scrosati, J. Electrochem. Soc. 138 (1991) 1917. [18] B. Kummar, L.G. Scanlon, R.J. Spry, J. Power Sources 96 (2001) 337. [19] J. Evans, C.A Vincent, P.G. Bruce, Polymer 28 (1987) 2324. [20] G.B. Appetecchi, F. Croce, G. Doutzenberg, M. Mastrogostino, F. Ronci, B. Scorsati, A. Zanelli, F. Alessandrini, P.P. Prosini, J. Electrochem. Soc. 145 (1998) 4126. [21] B.A. Boukamp, Solid State Ionics 20 (1985) 31. [22] M.C. Borghini, M. Mastragostino, S. Passerini, B. Scrosati, J. Electrochem. Soc. 142 (1995) 2118. [23] K. Amine, C. Hchen, J. Liu, M. Hammond, A. Jansen, D. Dees, I. Bloom, D. Vissers, G. Henriksen, J. Power Sources 97/98 (2001) 684. [24] C. Delmaa, J. Saadoune, A. Rougier, J. Power Sources 43/44 (1993) 593. [25] R.K.B. Grover, R. Kanno, M.J. Mitchell, M. Yonemura, Y. Kawamoto, J. Electrochem. Soc. 147 (2000) 4045. [26] M. Gauthier, M. Armand, L. Krause, 7th International Meeting on Lithium Batteries, Ext. Abst., Boston, USA, 1994, p. 177. [27] Y. Xia, K. Tatsumi, T. Fujieda, P.P. Prsini, T. Sakai, J. Electrochem. Soc. 147 (2000) 2050. [28] G.B. Appetecchi, F. Alessandrini, M. Carewska, T. Caruso, P.P. Prosini, B. Scrosati, S. Paserini, J. Power Sources 97/98 (2001) 790.

Q. Li et al. / Solid State Ionics 159 (2003) 97–109 [29] D.F. Fauteux, A. Massucco, M. McLin, M. van Buren, J. Shi, Electrochimica Acta 40 (1995) 2185. [30] H.Y. Sun, H.-J. Sohn, O. Yamamoto, Y. Takeda, N. Imanishi, J. Electrochem. Soc. 146 (1999) 1672. [31] Q. Li, Y. Takeda, N. Imanishi, J. Yang, H.-Y. Sun, O. Yamamoto, J. Power Sources 97/98 (2001) 795.

109

[32] H. Yang, K. Kwon, T.M. Devine, J.W. Evans, J. Electrochem. Soc. 147 (2000) 4399. [33] F. Croce, E.S. Flory, I. Persi, B. Scrosati, Electrochem. SolidState Lett. 4 (2001) A121.