Materials Science and Engineering B 135 (2006) 65–73

Preparation and characterization of polyurethane/poly(vinylidene fluoride) composites and evaluation as polymer electrolytes P. Santhosh a , T. Vasudevan a , A. Gopalan a,∗ , Kwang-Pill Lee b a

b

Department of Industrial Chemistry, Alagappa University, Karaikudi 630 003, India Department of Chemistry Education, Kyungpook National University, Daegu 702-701, South Korea Received 7 May 2006; received in revised form 2 July 2006; accepted 15 August 2006

Abstract Polymer composites (PU/PVdF) comprised of polyurethane (PU) and poly(vinylidene fluoride) (PVdF) were prepared. Polymer electrolytes were prepared with different amounts of lithium perchlorate loading. Differential scanning calorimetry and impedance spectroscopy were used to monitor the changes in the thermal characteristics and bulk conductivity of PU/PVdF composites, respectively. Fourier transform infra-red spectroscopy was employed to identify the modifications in molecular interactions associated with PU and PVdF. The bulk conductivity of PU/PVdF composites with different PVdF content was determined and the influence of PVdF on that was brought out. PU/PVdF composites possess sufficient electrochemical and thermal stability. A laminated cell was constructed and the performance of PU/PVdF composite was assessed. © 2006 Elsevier B.V. All rights reserved. Keywords: Composite polymer electrolyte; Polyurethane; Poly(vinylidene fluoride); DSC; FT-IR; AC impedance; TGA; Electrochemical studies

1. Introduction Since 1973 [1,2], considerable research has been directed toward the development of polymer electrolytes having high ionic conductivity at room temperature. Polymer electrolytes so far reported possess low dielectric constants and have high degree of ion association. Ultimately, it could be possible to have low concentrations of charge carriers in the polymer matrix which could result in low ionic conductivities at room temperature. Further, many of the polymer electrolytes suffer from poor mechanical stability and often creep under pressure when they are used in electrochemical devices. Polyurethane (PU) based polymer electrolytes have been developed for rechargeable lithium ion batteries. PUs are composed of a polyether or polyester soft segment (SS) and an isocyanate based hard segment (HS) and characterized by two-phase (soft and hard) morphology [3]. The rubbery soft phase dissolves alkali metal salts without formation of ionic clusters. Furthermore, the low glass transition temperature (Tg ) and hence higher segmental motion of the polyether SS leads to the higher mobility of the dissolved ions [4]. The HS

∗

Corresponding author. Tel.: +91 4565228836; fax: +91 4565225202. E-mail address: [email protected] (A. Gopalan).

0921-5107/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.mseb.2006.08.033

domains which are distributed or interconnected throughout the rubbery phase of the SS act as a physical cross-links and filler to the SS matrix and hence contribute to the dimensional stability of the polymer electrolyte [5]. The domain formation is derived from the strong intermolecular hydrogen bonding between the hard–hard segments dissolved in the SS phase [6]. PU synthesized from 4,4 -methylene bis(phenyl isocyanate) as HS and poly(propylene glycol) as SS and ethylene diamine as chain extender, doped with LiClO4 exhibited a low conductivity in the order of ∼10−8 S cm−1 at 40 ◦ C, even though the polymer electrolyte was mechanically stable [7,8]. Ferry et al. [9] reported a low value of conductivity (∼10−9 S cm−1 ) at room temperature for PU complexed with LiClO4 . Carvalho et al. [10] investigated the polymer electrolyte comprising of three different PEO of various molecular weights, crosslinked with urethane functions. Of the different PEO used, PEO2000 with a molar concentration ratio O/Li = 43 showed a maximum conductivity of 10−5 S cm−1 at room temperature. These earlier reports [3–10] revealed that PU based polymer electrolytes containing lithium salts have lithium ion conductivities in the range of ∼10−8 S cm−1 and these values are deplorably very low for effective operation of the battery at room temperature. Hence, different approaches were developed

66

P. Santhosh et al. / Materials Science and Engineering B 135 (2006) 65–73

to improve the ionic conductivity of the PU based polymer electrolytes. Preparation of composite of PU with other plasticizing polymers is effectively used as one of the approaches. Composite gel polymer electrolyte (GPE) based on PU and polyacrylonitrile (PAN) was reported [11,12]. An impedance study correlating ion transport mechanisms to the conductivity behavior of PU/PAN composite electrolyte was reported [13]. Composite GPEs based on PU blended with linear PEO incorporating different plasticizers have been developed [14,15]. It is known from several reports that use of poly(vinylidene fluoride) (PVdF) in electrolyte systems can have several advantages [16–19]. The PVdF based polymer plasticized by liquid electrolyte showed relatively high ionic conductivity in the order of 10−3 S cm−1 at room temperature [20–22]. Further, PVdF itself has a high dielectric constant (ε = 8.4) for a polymer, which can assist greater ionization of lithium salts and provide high concentration of charge carriers. Composite materials based on the combination of PU and PVdF could possess properties of the individual components with a synergistic effect. Hence, it would be worthwhile to prepare composites of polymer electrolytes of PU with PVdF and analyze the modifications in morphology, thermal characteristics and conductivity of the composite. In the present study, composites of PU with different proportions of PVdF were prepared. Morphology, thermal transitions and conductivity of the composites prepared with various proportions of PVdF and PU were analyzed with differential scanning calorimetry (DSC), Fourier transform infra-red (FTIR) spectroscopy and AC impedance measurements. Also, the electrochemical characteristics and applicability of the composite polymer electrolytes in lithium batteries were investigated.

was increased to 85 ◦ C and two drops of the catalyst, dibutyl tin dilaurate were added to the reactor. Then, 0.25 mol TDI (5:1 NCO/OH) was added stepwise to the reaction mixture. DMF was used to control the viscosity of PUs during the polymerization.

2. Experimental

Differential scanning calorimetric (DSC) experiments were carried out using a DSC 2010 Differential Scanning Calorimeter (TA Instruments, USA) over a temperature range −150 to 150 ◦ C at a scan rate of 10 ◦ C/min. Fourier transform infra-red spectra were recorded at ambient temperature using Perkin-Elmer Rx1 instrument with a wave number resolution of 4 cm−1 . Samples for FT-IR were made by casting the polymer–salt mixture directly on KBr pellets and then simultaneously dried at 120 ◦ C for 48 h. Thermogravimetric analysis (TGA) of the samples were performed under nitrogen atmosphere using Perkin-Elmer TGA 7/DX Thermal Analyzer with a scan rate of 20 ◦ C/min. Impedance measurements of the composite electrolytes were performed using EG & G PAR 6310 Potentiostat/Galvanostat controlled by the frequency response analysis (FRA) under an oscillation potential of 10 mV. For measurement of ionic conductivity, the samples of 100 ␮m thickness were sandwiched between two stainless steel (SS 304) electrodes.

2.1. Materials Poly(ethylene glycol) (PEG; molecular weight 400; E.Merck) was dehydrated under reduced pressure at 80 ◦ C for 24 h before use. Toluene diisocyanate (TDI), 1,4-buatane diol and poly(vinylidene fluoride) (PVdF; molecular weight 1.5 × 105 , Aldrich) were used as received. LiClO4 (Aldrich) was dehydrated at 120 ◦ C under reduced pressure for 72 h. All the other reagents and chemicals were used without further purification. 2.2. Synthesis of PU PUs were prepared with polyols, PEG-400 as soft segment, 1,4-butane diol as chain extender and TDI as hard segment. PUs were synthesized in a batch reactor consisting of a 2000 ml, fournecked, round-bottom flask with an anchor stirrer, a nitrogen inlet and outlet and a thermocouple connected to the temperature controller. The polyols, PEG-400 and the chain extender, 1,4butane diol were kept in a vacuum oven at 80 ◦ C for 1 day to remove moisture from the chemicals. The polyols (0.05 mol) and 0.20 mol chain extender were introduced in the reaction flask and mixed well at 50 ◦ C. After 30 min, the temperature

2.3. Preparation of composites of PU with PVdF PVdF was dried under vacuum at 50 ◦ C for 48 h. PU and PVdF were blended physically in various compositions by solution casting. The mixture was dissolved in DMF and stirred vigorously for 1 h using a homo-mixer. Then, the solutions were coated on a polypropylene plate and dried under vacuum at 50 ◦ C for 48 h. The films were then stored in a dry box. The thickness of the films was controlled between 50 and 100 ␮m. Compositions of PU containing 2.5, 5.0, 7.5% PVdF (w/w) were made and designed as A, B and C, respectively. Composites of higher PVdF content (>7.5%) did not give homogeneous films. 2.4. Preparation of the polymer electrolytes Electrolyte films were prepared by mixing the composite in DMF with suitable LiClO4 /DMF solution. After homogeneous mixing, the solution was cast into teflon plates. Solvent removal was done at 80 ◦ C under reduced pressure for 72 h. The films were then kept inside a glove box before further experiments. The undoped film was also made in the same way but without LiClO4 . Polymer electrolytes containing 0.1, 0.2 and 0.3 mM LiClO4 were designed as X1, X2, X3 (X = A, B and C), respectively. 2.5. Characterization

2.6. Cell assembly and performance characteristics Cell was assembled in dry argon atmosphere inside a glove box. Cyclic voltammetry (CV) and linear sweep voltammetry (LSV) were performed with a cell made by using SS 304 as working electrode and lithium metal as counter as well as reference electrode. A sweep rate of 1 mV s−1 was used in both cases.

P. Santhosh et al. / Materials Science and Engineering B 135 (2006) 65–73

67

For DC polarization measurements, the cell was constructed by sandwiching the electrolyte film between two symmetrical lithium metal electrodes. A small constant potential difference of about 10 mV was applied across the cell and the current was measured as a function of time until it reaches a constant value. By considering the potential drop occurring at surface layers on the electrode, the cation transport number, t+ was determined. AC impedance spectra were recorded before and after the current relaxation measurement without interruption of the DC bias. For charge–discharge cycling tests, a cell was fabricated using LiCoO2 and lithium as cathode and anode, respectively. The area of both electrodes was fixed as 2.0 cm2 . Li/(PU/PVdF)/LiCoO2 laminated cells were assembled by pressing Li, PU/PVdF and LiCoO2 , sealed by polyethylene film and laminated by an aluminium foil. The charge–discharge cycling tests of the laminated cell were conducted under galvanostatic conditions in a dry argon atmosphere. Also, the discharge curves were obtained at different current rates in order to obtain the rate capacity of the cell at room temperature. 3. Results and discussion 3.1. Differential scanning calorimeter The degree of miscibility of PU with PVdF in the PU/PVdF composite was determined by using DSC analysis. Fig. 1(A) shows the glass transition temperature (Tg ) of PU and PU/PVdF composite. The Tg s are shown by arrows. As the hard and soft segments of PUs are thermodynamically incompatible, separate thermal transitions are visible for both hard and soft segments. PU shows a low temperature endothermic transition at −85.2 ◦ C and endothermic peak at 18.2 ◦ C (Fig. 1A(i)). Whereas, the composite (2.5% PVdF, w/w) shows an endothermic peak at 20.1 ◦ C and a low temperature transition at −55.4 ◦ C (Fig. 1A(ii)). These two transitions correspond to the melting of the crystalline region (Tm ) and Tg of amorphous PEG in the soft segment (Tg SS) [23]. The addition of PVdF into the PU matrix shifts the Tg SS to higher temperature. These changes in Tg inform that there might be interactions between the groups present in PU and PVdF. Several researchers reported on the phase-mixed state of PU as a result of interactions between hard (C O and –NH groups) and soft (C(O)–O–C) segments through hydrogen bonding [11–15]. Fig. 1B shows the DSC thermograms of PU/PVdF polymer electrolytes doped with LiClO4 . The doping was done with 0.3 mM LiClO4 for various PVdF contents (2.5, 5.0 and 7.5%, w/w). The effect of LiClO4 on the thermal behavior of the composites is presented in Table 1. The Tg SS does show an increasing trend with increase in PVdF content. When these composites are doped with various concentrations of LiClO4 , lithium ions form complexation with segments in the composites and showed an increase in Tg SS. The increase in Tg SS is attributed to the formation of transient cross-links. The interaction of the Li+ ions restricts the segmental motion of the polymer chains, leading to an increase in Tg . However, the trend of variation of Tg s with the addition of LiClO4 (Table 1) depends on the content of the electrolyte (PVdF or PU). This is evident after an analysis of the

Fig. 1. DSC thermograms of: (A) PU (i) and PU/PVdF composite (2.5% PVdF, w/w) (ii); (B) PU/PVdF composites doped with 0.3 mM LiClO4 for various PVdF contents: (i) 2.5%, (ii) 5.0% and (iii) 7.5% (w/w). Table 1 DSC data for PU/PVdF composites with different LiClO4 concentration Film

Tg SS (◦ C)

Tg /C (◦ C/mM g)

A0 A1 A2 A3 B0 B1 B2 B3 C0 C1 C2 C3

−55.4 −30.7 −25.4 −20.1 −42.6 −21.4 −18.4 −15.2 −30.1 −13.2 −11.5 −9.4

– 15.6 18.5 23.4 – 17.5 14.8 11.2 – 14.2 18.2 16.5

PU with PVdF; 2.5% (A), 5.0% (B) and 7.5% (C) (w/w), (A0, B0, C0: undoped composite electrolyte; 1, 2 and 3 represent 0.1, 0.2 and 0.3 mM LiClO4 , respectively).

68

P. Santhosh et al. / Materials Science and Engineering B 135 (2006) 65–73

Fig. 2. FT-IR spectrum of PU/PVdF composite (2.5% PVdF, w/w) inset shows the deconvoluted C O and –NH stretching region of PU/PVdF composite.

normalized Tg data with respect to the LiClO4 concentration, Tg /C (Table 1). For the composite A, the increase in Tg /C of the SS was almost linear with LiClO4 concentration. In the case of composite B, a decrease in Tg /C with increasing LiClO4 concentration was noticed. This may be attributed to the plasticizing effect from the formation of charge-neutral contact ion pairs with increasing LiClO4 concentration [24]. The composite C showed a different behavior in comparison to A and B. Initially, Tg /C increases with increase in LiClO4 concentration up to 0.2 mM and thereafter decreases. This may be due to the plasticizing effect of the ion pair aggregations. 3.2. FT-IR spectroscopy FT-IR spectra of the PU/PVdF composites with different LiClO4 concentration provide evidence for the changes in molecular interactions between the groups present in PU/PVdF

and Li+ ions. The vibrational bands corresponding to –NH, C O and ether stretching modes were deconvoluted and compared for the changes in peak frequencies and band areas. A representative FT-IR spectrum of PU/PVdF composite is presented (Fig. 2). The deconvolution of the –NH spectral region (3000– 3600 cm−1 ) was made by using the best fits by Gaussian Lorentzian sum to make contributions from free –NH stretching, overtone of C O stretch modes, hard–hard (HH) and hard–soft (HS) segment hydrogen bonded –NH stretching modes [25]. The changes in band positions and band areas were analyzed between LiClO4 doped and undoped PU/PVdF composites. Table 2 presents the deconvolution results of –NH stretching mode of composites A, B and C. In general, the frequency shift of H-bonded –NH stretching represents the strength of H-bonding in PUs [26]. The results (Table 2) reveal the shift in the frequency of band corresponding to H-bonding stretching of both HH and HS segments is in the order; composite C > composite B > composite A. A close analysis of the –NH stretching frequency and band area (Table 2) indicates that there is shift in the position of –NH stretching band arising from the hydrogen bonding between hard segment –NH and C O groups in the presence of LiClO4 . Since the band position is related to the strength of –NH bond, the shift in frequency upon addition of LiClO4 informs the change in bond strength of –NH. This change in –NH bond strengths can be viewed due to localization of the electron rich oxygen as a result of coordination with Li+ ions via hydrogen bonded species. As a result of it, the hydrogen bonding between –NH and C O could be weakened. This is evident from the observed lower band area (25.6) for hydrogen bonded –NH stretch in composite A3 in comparison with composite A0 (32.7) [27]. This is attributed to the more release of the hydrogen bonds in composite A3 as a result of interaction with Li+ ions. The band position of hydrogen bonded –NH to ether oxygen (SS) showed shift from 3265 cm−1 for composite B0 to 3241 cm−1 for composite B3. The shift in lower frequency by the addition of LiClO4 informs that Li+ ions coordinate to non-bonded electrons in ether oxygen and make the –NH bond weaker. A similar analysis with respect to molecular interaction

Table 2 Deconvolution data of FT-IR spectra of PU/PVdF composites in the –NH stretching region Sample

A0 A1 A2 A3 B0 B1 B2 B3 C0 C1 C2 C3

LiClO4 (mM)

– 0.1 0.2 0.3 – 0.1 0.2 0.3 – 0.1 0.2 0.3

Peak positions (cm−1 )

Percent areaa

1

2

3

4

1

2

3

4

3401 3385 3362 3360 3427 3412 3405 3387 3395 3382 3385 3375

3390 3381 3374 3371 3372 3365 3359 3351 3369 3358 3351 3348

3332 3332 3325 3318 3338 3325 3321 3319 3339 3335 3330 3327

3252 3247 3235 3230 3265 3258 3251 3241 3265 3259 3247 3249

30.1 28.4 25.4 20.8 25.4 24.8 22.4 21.5 17.4 18.4 19.1 19.5

15.8 14.5 13.1 12.0 18.5 17.4 19.2 20.1 27.4 22.1 20.1 18.7

32.7 30.1 28.4 25.6 35.4 28.4 25.6 24.1 33.2 28.4 29.4 27.8

21.4 27.0 33.1 41.6 20.1 29.4 32.8 34.3 21.9 31.1 31.4 34.0

Peak 1: free –NH bonding; peak 2: overtone of C O; peak 3: hard–hard segment H-bonding; peak 4: hard–soft segment H-bonding. a Peak areas are based on total –NH stretching band area, PVdF: (A) 2.5%, (B) 5.0% and (C) 7.5% (w/w).

P. Santhosh et al. / Materials Science and Engineering B 135 (2006) 65–73

69

Table 3 Deconvolution data of FT-IR spectra of PU/PVdF composites in the C O stretching region Sample

A0 A1 A2 A3 B0 B1 B2 B3 C0 C1 C2 C3

LiClO4 (mM)

– 0.1 0.2 0.3 – 0.1 0.2 0.3 – 0.1 0.2 0.3

Peak positions (cm−1 )

Percent areaa

1

2

3

1

2

3

1762 1759 1755 1752 1760 1758 1755 1751 1755 1753 1750 1746

1756 1751 1749 1748 1752 1749 1746 1742 1749 1735 1730 1729

1733 1730 1729 1725 1730 1728 1725 1721 1729 1726 1721 1719

44.5 41.4 39.8 39.2 43.1 42.1 40.2 38.7 40.1 40.7 40.7 40.4

40.1 39.4 35.2 32.7 42.4 45.4 40.1 38.5 45.4 42.4 42.6 40.4

15.4 19.0 25.0 28.1 14.5 12.5 19.7 22.6 14.5 16.9 16.6 19.2

Peak 1: free carbonyl; peak 2: disordered H-bonded carbonyl; peak 3: ordered H-bonded carbonyl. a Peak areas are based on total C O stretching band area, PVdF: (A) 2.5%, (B) 5.0% and (C) 7.5% (w/w).

in terms of changes in –NH band frequency and area was extended to the other composites (Table 2). The presence of 5.0 and 7.5% PVdF makes significant changes in the molecular interactions between hard and soft groups in PU. Also, the molecular interactions between the groups in PU are different in composites B and C in comparison with composite A [28]. Addition of 0.3 mM LiClO4 alters the hard and soft segment interactions. Also, the presence of different extent of PVdF (5.0 and 7.5%) in the composite provides a modified environment for LiClO4 to have different type of interactions. Hence, the added LiClO4 can be present as dissociated ions and ion pairs with varying proportions in composites B and C. The interactions of Li+ ions would also influence the band positions and areas of carbonyl stretch in the PU/PVdF composites. To investigate this, the C O stretching regions of the composites were deconvoluted into free urethane carbonyl, ordered and disordered hydrogen bonded carbonyl stretch bands [29] and the results are presented (Table 3). The band positions of hydrogen bonded carbonyl (both ordered and disordered) showed shifting towards lower frequen-

cies. This indicates the weakening of hydrogen bonded C O bond strength. On increasing LiClO4 concentration, free urethane carbonyl shifts from 1760 cm−1 (B0) to 1751 cm−1 (B3). The shift in frequency from 1760 to 1751 cm−1 suggests that the ionic coordination between the free urethane carbonyl and Li+ ion increases with increasing LiClO4 concentration. In addition, the peak area decreases with increasing LiClO4 concentration. This may be due to the non-availability of free urethane carbonyl group. A similar analysis with respect to molecular interaction in terms of changes in carbonyl band frequency and area was extended to composites A and C (Table 3). Here too, the presence of various proportion of PVdF makes significant changes in the molecular interactions. However, no definite trend could be observed by the addition of the LiClO4 . Table 4 shows the deconvolution results of ether stretching region for the PU/PVdF composites. The deconvoluted FT-IR spectral data represent three characteristic vibrational modes; C–O–C stretch of PEG, C(O)–O–C stretch of the urethane and hydrogen bonded C–O–C stretch of PEG [29]. A shift to lower frequency region was noticed with the addition of LiClO4

Table 4 Deconvolution data of FT-IR spectra of PU/PVdF composites in the ether stretching region Sample

A0 A1 A2 A3 B0 B1 B2 B3 C0 C1 C2 C3

LiClO4 (mM)

– 0.1 0.2 0.3 – 0.1 0.2 0.3 – 0.1 0.2 0.3

Peak positions (cm−1 )

Percent areaa

1

2

3

1

2

3

1155 1151 1157 1148 1148 1145 1141 1138 1145 1147 1142 1140

1122 1120 1118 1115 1121 1115 1111 1108 1112 1110 1110 1109

1115 1114 1115 1113 1110 1109 1107 1107 1108 1107 1108 1107

45.2 42.5 43.4 42.0 39.4 35.6 32.4 31.4 35.4 33.4 31.4 30.1

38.0 37.5 36.4 35.1 37.8 35.4 33.8 31.9 36.4 35.1 34.2 33.4

16.5 20.0 20.2 22.8 22.7 28.8 33.8 36.7 28.2 31.5 34.4 36.5

Peak 1: C–O–C stretch of PEG; peak 2: C(O)–O–C stretch of the urethane; peak 3: H-bonded C–O–C stretch of PEG. a Peak areas are based on total ether stretching band area, PVdF: (A) 2.5%, (B) 5.0% and (C) 7.5% (w/w).

70

P. Santhosh et al. / Materials Science and Engineering B 135 (2006) 65–73

or Vogel–Tamman–Fulcher, VTF (Eq. (2)) relationship depending on whether the ionic mobility is coupled with segmental motion of the polymer or not.   E (1) σ(T ) = A exp − kB T   B σ(T ) = AT 1/2 exp − (T − T0 ) (2) kB

Fig. 3. Temperature dependence of ionic conductivity of PU/PVdF composites (PVdF: (A) 2.5%, (B) 5.0% and (C) 7.5%, w/w) doped with different amount of LiClO4 ; 1, 2 and 3 represent 0.1, 0.2 and 0.3 mM LiClO4 , respectively.

(Table 4). This is consistent with the earlier observations [30] of a MDI/PTMO/BDO based PU system. The addition of LiClO4 into PU/PVdF composites alters the molecular interactions with consequent modifications in the microstructure of the system. 3.3. Ionic conductivity Fig. 3 represents the variation of ionic conductivity for PU/PVdF composites A, B and C with various LiClO4 concentrations as a function of the temperature. Conductivity variations with temperature were explained either with Arrhenius (Eq. (1))

where A is a constant, E the activation energy, B the pseudoactivation energy related to polymer segmental motion, kB the Boltzmann constant and T0 is a reference temperature usually associated with the ideal Tg at which free volume disappears or the temperature at which the configurational entropy becomes zero. The value of T0 may be lesser than Tg . The application of the VTF relationship to ion transport requires the coupling of the charge carriers with the segmental motion of the polymer chains. However, the Arrhenius relationship is applicable when the charge carriers are decoupled from the polymer host. Composites A and B (Fig. 3A and B) follows Arrhenius relationship for the ion transport, i.e., ion movement occurs by an activated hopping mechanism. Whereas composite C follows VTF relationship (Fig. 3C) which indicates that ion transport is coupled with polymer segmental motion. The ionic conductivity of composite C was found to be approximately one order of magnitude higher than that of composite A with equivalent LiClO4 concentration. This may be due to the presence of higher amount of PVdF, which facilitates easier dissociation of LiClO4 and leads to an increased number of charge carriers. Another important feature in composite B (Fig. 3B) is the variation of the ionic conductivity with LiClO4 concentration. At lower temperature, composite B1 shows lower conductivity than composite B2. However, at higher temperature, an opposite trend was observed. The slopes of the two curves indicate that they intersect each other at a higher temperature. Composite B2 showed higher conductivity because of an increasing number of charge carriers, but the Tg SS was also simultaneously increased, and this reduced the mobility. Hence, at higher temperature region, the composite B2 containing higher amounts of LiClO4 (0.2 mM), though containing an increased number of charge carriers, showed lower conductivity because the mobility was restricted on account of the higher Tg value of the soft segment. According to the FT-IR results, some of the Li+ ions may be coordinated to the hard-segment urethane groups and so will not contribute conductivity unless the temperature becomes close to the hard-segment Tg . This might be another reason for the low conductivity of sample B2. 3.4. Cyclic voltammetry studies The electrochemical reversibility of PU/PVdF composite was evaluated by cyclic voltammetry. Fig. 4 shows cyclic voltammogram of lithium deposition and stripping process at composite electrolyte. The sweep rate was kept as 1 mV s−1 . An anodic stripping peak at ∼0.75 V with a cathodic deposition peak at ∼−0.6 V could be seen. Hence, lithium deposition-stripping process is considered to be reversible in the composite elec-

P. Santhosh et al. / Materials Science and Engineering B 135 (2006) 65–73

Fig. 4. Cyclic voltammogram of PU/PVdF composite (2.5% PVdF, w/w).

trolyte environment. The above results illustrate that PU/PVdF composites have good electrochemical reversibility, which is a sufficient requirement for lithium batteries. 3.5. Transport number The transport number of an electrolyte is an important index of its conductive behavior. Since the electrochemical process in lithium batteries involves the intercalation and de-intercalation of lithium cations throughout the host compound lattice, polymer electrolytes with cation transference number (t+ ) approaching unity are desirable for avoiding a concentration gradient during repeated charge–discharge cycles. Thus, the evaluation of t+ is of great importance for the characterization of the polymer electrolytes. In general, t+ was expressed as the ratio of the steady state, Iss to the initial current, Ii . Significant errors resulted from neglect of kinetic resistance at the electrode/electrolyte interface. Hence, by considering the potential drop occurring at surface layers on the electrode, the cation transport number, t+ was determined by using the relation [31], t+ =

71

Fig. 5. Current relaxation plot for transference number measurement of PU/PVdF composite (C1) (Inset: complex impedance plot; initial resistance and steady-state resistance measured before and after the current transient).

The average t+ for the PU/PVdF composite was determined as 0.48. 3.6. Electrochemical stability Fig. 6 shows the current–voltage curve of the SS ‘blocking’ working electrode in composite electrolyte cell. The potential was scanned from 0 to 5.5 V versus Li. The trend of the curve of Fig. 6 informs the anodic stability of the composite electrolyte. From the magnitude of the current response, the decomposition voltage of PU/PVdF composites on SS electrodes was found to be 4.75, 5.0 and 4.75 V versus Li (for A1, B1 and C1, respectively). The stability of the electrolytes is influenced partially by the weight of the PVdF content in the

Iss (V − Ii Ri ) Ii (V − Iss Rss )

where Ii and Iss are the initial and steady-state currents, Ri and Rss the initial and steady-state resistance of the passivating layers and V is the applied potential. AC impedance spectra were therefore recorded before and after the current relaxation measurement without interruption of the DC bias, to permit Ri and Rss to be evaluated. Cation transference number, t+ of the composite electrolyte was determined by the application of 10 mV DC potential across the test cell (Li/(PU/PVdF)/Li). The current decays immediately and asymptotically approaches steady state (Fig. 5). The impedance response of test cell (inset of Fig. 5) was monitored at an initial time and at a steady-state current condition. The inset in Fig. 5 shows an expansion in the impedance spectra semi-circle.

Fig. 6. Linear sweep voltammograms of PU/PVdF composites showing the electrochemical stability window. PVdF content: (A) 2.5%, (B) 5.0% and (C) 7.5% (w/w).

72

P. Santhosh et al. / Materials Science and Engineering B 135 (2006) 65–73

Fig. 7. Thermograms of PU and PU/PVdF composites. PVdF content: (A) 2.5%, (B) 5.0% and (C) 7.5% (w/w).

composites. The composite electrolyte of PVdF (5.0%, w/w) shows the maximum electrochemical stability of about 5.0 V versus Li. 3.7. Thermal analysis Fig. 7 represents the thermograms of PU and PU/PVdF composites. PU/PVdF composites show lower thermal stability than that of the PU, which may be due to the incorporation of PVdF into the PU matrix. Composite C1 showed higher initial decomposition temperature in comparison with composites B1 and A1. This demonstrates that the composite with lower PVdF content (2.5%, w/w) may have better thermal stability than the other composites.

Fig. 8. (A) Charge–discharge profiles of Li/PU/PVdF(C1)/LiCoO2 and Li/ (LiClO4 -PC)/LiCoO2 . (B) Rate capacity profiles of Li/PU/PVdF(C1)/LiCoO2 cell at various rates.

3.8. Charge–discharge cycling tests Li/LiCoO2 cell was fabricated with PU/PVdF composite as electrolyte. To access and compare the performance of PU/PVdF composite, simple cell with 1 M LiClO4 in propylene carbonate (PC) as liquid electrolyte was used. The charge–discharge curves of Li/(PU–PVdF)/LiCoO2 and Li/LiClO4 -PC/LiCoO2 cell at ambient temperature are shown in Fig. 8A. This cell was charged at C/10 rate between 2.5 and 4.3 V. The initial capacity was slightly lower of about 112 mAh/g during few cycles because of the initial interfacial contact between the electrolyte and the electrodes. A comparison among the discharge curves reveals that the capacity of PU/PVdF composite is slightly lower than that of liquid electrolyte (Fig. 8A). The reduced capacity is attributed to the lower diffusion rate of lithium ions in the PU/PVdF composite in comparison to the liquid electrolyte. The rated capacity of Li/(PU/PVdF)/LiCoO2 cell is shown in Fig. 8B. The cell at C/10 rate reached a value of 87% normal capacity. The reduced capacity at higher rates is due to the low value of the diffusion coefficient of lithium ions in the lattice

of LiCoO2 in the electrolyte environment. At the C/5 and C/1 rates, the cell delivered about 83 and 66% of the full capacity, respectively, at an average load voltage of 3.5 V. 4. Conclusions Polymer electrolytes based on PU/PVdF composites possess adequate properties for the use in lithium batteries. The soft and hard segment groups in PU form cross-links with Li+ ions and provide platform for the lithium ion conductivity. PVdF adds mechanical rigidity and improvements in ionic conductivity over the pristine PU. The ionic conductivity of the composite with PVdF 7.5% (w/w) was found to be approximately one order of magnitude higher than the corresponding composites with PVdF 2.5 and 5%, respectively, with equivalent LiClO4 concentrations. The PU/PVdF composite polymer electrolytes are electrochemically stable up to 5.0 V versus Li. The cell constructed with the composite polymer electrolyte showed adequate rated performance and proves to be suitable for lithium batteries.

P. Santhosh et al. / Materials Science and Engineering B 135 (2006) 65–73

References [1] D.E. Fenton, J.M. Parker, P.V. Wright, Polymer 14 (1973) 589. [2] M. Armand, J.M. Chabagno, M. Duclot, in: P. Vashishta, N.N. Mundy, G.K. Shenoy (Eds.), Fast Ion Transport in Solids, North-Holland, Amsterdam, 1979, p. 131. [3] P.E. Gibson, J.W.C. Van Bogart, S.L. Copper, J. Polym. Sci. B-Polym. Phys. 24 (1983) 885. [4] C.B. Wang, S.L. Copper, Macromolecules 16 (1983) 775. [5] G. Jiang, S. Maeda, H. Yang, Y. Saito, S. Tanase, R. Sakai, J. Power Sources 141 (2005) 143. [6] H.L. Wang, H.M. Kao, T.C. Wen, Macromolecules 33 (2000) 6910. [7] M. Watanabe, S. Oohashi, K. Sanui, N. Ogata, T. Kobayashi, Z. Ohtaki, Macromolecules 18 (1985) 1945. [8] M. Watanabe, K. Sanui, N. Ogata, Macromolecules 19 (1986) 815. [9] A. Ferry, P. Jacobson, J.D. van Heumen, J.R. Stevens, Polymer 37 (1996) 737. [10] L.M. Carvalho, P. Guegan, H. Cheradame, A.S. Gomes, Eur. Polym. J. 33 (1997) 1741. [11] T.C. Wen, H.H. Kuo, A. Gopalan, Solid State Ionics 147 (2002) 171. [12] H.H. Kuo, W.C. Chen, T.C. Wen, A. Gopalan, J. Power Sources 110 (2002) 27. [13] P. Basak, S.V. Manorama, R.K. Singh, O. Parkash, J. Phys. Chem. B 109 (2005) 1174. [14] Y.L. Du, T.C. Wen, Mater. Chem. Phys. 71 (2001) 62.

73

[15] T.C. Wen, W.C. Chen, J. Power Sources 92 (2001) 139. [16] C.L. Cheng, C.C. Wan, Y.Y. Wang, M.S. Wu, J. Power Sources 144 (2005) 238. [17] N.K. Chung, Y.D. Kwon, D. Kim, J. Power Sources 124 (2003) 148. [18] E. Quartarone, P. Mustarelli, A. Magistris, J. Phys. Chem. B 106 (2002) 10828. [19] M. Yoo, C.W. Frank, S. Mori, S. Yamaguchi, Chem. Mater. 16 (2004) 1945. [20] S.W. Choi, S.M. Jo, W.S. Lee, Y.R. Kim, Adv. Mater. 15 (2003) 2027. [21] S. Panero, F. Ciuffa, A. D’Epifano, B. Scrosati, Electrochim. Acta 48 (2003) 2009. [22] H.S. Choe, J. Giaccai, M. Alamgir, K.M. Abraham, Electrochim. Acta 40 (1995) 2289. [23] T.C. Wen, Y.J. Wang, T.T. Cheng, C.H. Yang, Polymer 40 (1999) 3979. [24] M. Salomon, M. Xu, E.M. Eyring, S. Petrucci, J. Phys. Chem. 98 (1994) 8234. [25] W.C. Chen, H.H. Chen, T.C. Wen, M. Digar, A. Gopalan, J. Appl. Polym. Sci. 91 (2004) 1154. [26] R. Mani, T. Mani, H.R. Stevens, J. Polym. Sci. A-Polym. Chem. 30 (1992) 2025. [27] T.C. Wen, H.H. Chen, J. Appl. Polym. Sci. 80 (2001) 935. [28] J. Herranen, J. Kinnunen, B. Mattsson, H. Rinne, F. Sundholm, L. Torell, Solid State Ionics 80 (1995) 201. [29] M. Digar, S.L. Hung, H.L. Wang, T.C. Wen, A. Gopalan, Polymer 43 (2002) 681. [30] J. Sun, D.R. Mac Farlane, M. Forsyth, Solid State Ionics 85 (1996) 137. [31] P.G. Bruce, C.A. Vincent, J. Electroanal. Chem. 225 (1987) 1.