J. MATER. CHEM., 1994, 4(12), 1785-1791 Preparation and Properties of stat-Copoly(oxyethylene/ 0xypropylene)-LiCIO, Polymer Electrolytes Mehdi Nekoomanesh H.,a David J. Wilson," Colin Booth*" and John R. Owenb a Manchester Polymer Centre and Department of Chemistry, University of Manchester, Manchester; UK M73 9PL Department of Chemistry, University of Southampton, Southampton, UK SO9 5NH High-molar-mass statistical copolymers of ethylene oxide and propylene oxide with compositions in the range 60-90 mol% oxyethylene units have been prepared, mixed with lithium perchlorate, and studied by differential, scanning calorimetry (DSC) and ac impedance spectroscopy. The observation by DSC of two glass transitions gave edidence of separation into salt-poor and salt-rich phases.A copolymer of 85 mol% oxyethylene mixed with LiCIO, at an 0:Li mole ratio of 24 :1 had a conductivity at 25 "C of OM 10-5 S cm-', i.e. lower than the room temperature conductivities recorded for comparable non-crystalline polymer electrolytes formed from some alternative high-molar-mass linear copolymers. Many polymers dissolve salts to form solutions (polymer electrolytes) which support ionic conductivity. The topic has been well reviewed.'-4 Polymers containing oxyethylene sequences in one form or another have yielded the most promising results. However, polymer electrolytes based on poly(oxyethy1ene) itself are not useful at room temperature, since crystallisation, either of polymer or of polymer-salt complex, is detrimental to conductivity.The useful solvent properties of poly(oxyethy1ene) are preserved in copolymers which contain oxyethylene sequences but which neither crys- tallise themselves nor form crystalline polymer-salt complexes at room temperature. Fortunately the crystallisation of poly- (oxyethylene) in helical conformation, whether of polymer or complex, is readily disrupted by any imperfections in the chain.5 Statistical copolymerisation of ethylene oxide with a comonomer represents a particularly simple, though not neces- sarily effi~ient,~.' way of suppressing crystallinity. The method is particularly attractive since high-molar-mass polymers are readily prepared by use of coordination catalysts.8p11 The Vandenberg catalyst' (triethylaluminium, water and acetyl- acetone) has been most frequently used in the preparation of high-molar-mass statistical copolymers for use as polymer electrolytes: e.g.copolymers of ethylene oxide with propylene oxide5,12,13 or other 1,2-epoxides. l4 Use of propylene oxide as comonomer is attractive, as it is readily and cheaply available, and poly(oxypropy1ene) itself has been much investigated as a polymer electrolyte. The disruption of the crystallisation of poly(oxyethy1ene) by incor- I I I 1 0.0 0.2 0.4 0.6 0.8 mole fraction E in feed Fig. 1 Calculated dependence of copolymer composition on feed composition in the statistical copolymerisation of ethylene 1 )xide and propylene oxide (rp=0.19, rE =2.4). The dashed curve is for il random copolymerisation, for which copolymer composition equals feed composition.stepwise addition of monomers to the living chain,' the result of composition drift is the formation of copolymcrs with oxyethylene content (E content) decreasing with distance along the chain from the initiator unit. For example, given the reactivity ratios listed above and an initial feed composi- tion (mole fraction ethylene oxide) xEo=0.55, cop01 ymer of composition (mole fraction E units) xE=0.78 is formed in the conversion interval 0-1 YO,but of composition xE=0.06 in the conversionporation of oxypropylene co-units is well e~tablished,~,'~,'~"~ and it is known that stat-copoly(oxyethylene/oxypropylene) with 30-40 mol% or more of oxypropylene (P) units is essentially non-crystalline at room temperature.In the present study, the Vandenberg catalyst' was used to prepare high-molar-mass copolymers. Reactivity ratios in the copolymerisation of ethylene oxide and propylene oxide (here denoted rEO and rpo) have been reported for a number of anionic and coordination catalyst system^.^^'^^" The value of rEOgenerally exceeds that of rpo by a factor of 10 or so: e.g. for equimolar triethylaluminium and water," rEO =2.4, rpo = 0.19. The present copolymerisations were generally taken to high conversion (ca. 90%) and the difference in reactivity of the two monomers was a source of composition drift. The magnitude of the drift can be judged by the plot of instan- taneous copolymer composition us.feed composition shown in Fig. 1. Given that the copolymerisation reaction involves interval 89-90%. The average composition of the copolymer recovered at 90% conversion is xE=0.61. The variation of composition along the chain is disadvantageous for suppression of crystallinity, but high conversion IS neces-sary for preparation of high-molar-mass copolymer. Since statistical copolymers used5,12-15 in polymer electrolytes have usually been prepared to high conversion, the problem of composition drift is a general one. Possible effects on the properties of the copolymers and their mixtures with LiC104 will be discussed below. Experimental Copolymers Diethyl ether (BDH, >99.5%) was refluxed with sodium wire and benzophenone for 2-3 h then distilled under nitrogen: bp 34-36 "C.Toluene (BDH, 99.5%) was similarly treated (bp J. MATER. CHEM., 1994, VOL..4 108-109 "C) and collected over type 4A molecular sieve. Acetylacetone (BDH, >98%) and triethylaluminium (Aldrich, 1.0mol dm-3 solution in hexane) were used as received. Prior to copolymerisation ethylene oxide (EO) (Fluka, >99.8%) was dried over calcium hydride for 1-2 days at 0°C and propylene oxide (PO)(BDH, >99.5YO)was dried over ground potassium hydroxide overnight at room temperature (rt). The initiator was triethylaluminium-acetylacetone-water in molar ratio 1.0 :0.5 :0.5. Triethylaluminium solution (7.8 cm3, 7.8 x lo-, mol AlEt,) was transferred under dry nitrogen to a 100cm3 three-necked flask by syringe, followed by dry diethyl ether (8 cm').The flask was placed in an ice-water bath and distilled water (70.2 mm3, 3.9 x lop3mol) was added slowly. After 0.5 h, acetylacetone (0.4 cm3, 3.9 x lop3 mol) was added and the resulting yellow solution was stirred for 16 h under nitrogen in the dark before use. Statistical copolymers were prepared from feeds of ethylene oxide with initial mole fractions in the range xE0~0.55-0.85. Copolymerisation took place in small reaction flasks sealed under vacuum by Teflon taps. The following procedure for the copolymerisation of a feed with xEo~0.55is typical. Dry toluene (40 cm3), EO (1.7 cm3, 1.52 g, 0.0346 mol) and PO (1.95 cm3, 1.67 g, 0.0288 mol) were distilled into the reaction flask, followed by injection of initiator solution (1.8 mmol of Al).The mixture was stirred at rt for 15 h. The resulting viscous mass was diluted by dichloromethane (200 cm3) before rotary evaporating it at 40°C to obtain a rubbery product, which was further dried under high vacuum for 24 h. The copolymer produced contained ca. 4 wt.% aluminium by microanalysis. Much of this was removed by equilibrating a dilute aqueous solution of the copolymer (2.5 g dm-3, 200 cm3) with Amberlite Resin IR-l20(H) (BDH, 20 g) under nitrogen. After filtration, the mixture was extracted with dichloromethane (3 x 100 cm3) and the copolymer recovered by rotary evaporation under reduced pressure at 30 "C. The resulting copolymer contained 0.3 wt. % Al. Further reduction of the aluminium content was not attempted.Details of the six samples prepared in this way are listed in Table 1: the notation adopted reflects the mol% of oxyethylene (E) units in the chains as found by NMR spectroscopy (see below). The molar masses of the samples were investigated by gel- permeation chromatography (GPC). The system comprised three PL-gel columns (each 0.75 cm id and 30 cm long) with porosities in the range 50-10' A,giving a wide range of resolution. The eluent was N,N-dimethylacetamide at 70 "C at a flow rate of 1 cm3 min-'. Solutions of samples (0.2 g drn-,) were injected through a 100 mm3 loop. Detection was by differential refractometry (Waters Model 410). The system was calibrated with poly(oxyethy1ene) standards.Molar mass distributions were wide (M,/M, >2) typical of coordination polymerisation. Molar masses 'as if poly(oxyethy1ene)' corre- sponding to the peaks of the GPC curve (Mpk)are listed in Table 1. All lay within the range Mp,=(3.5-5.5) x 10' g mol-'. Compositions were determined by I3C NMR spectroscopy. Spectra of copolymer samples dissolved in CDC13 were recorded using a Bruker AC-300 spectrometer operating at 75.5 MHz. Intensities of resonances assigned18 to backbone carbons (6, =68-71 and 72.5-75.5 for oxyethylene and oxy- propylene, respectively) served to define the overall composi- tions listed in Table 1, which were checked against the intensities of resonance of the methyl carbons (6,= 16-18). Resonances at 6,=72.64 and 73.11 assigned" to racemic and meso-PP diads were observed in the spectra of samples EP64 and EP65, with P contents greater than 30 mol%.Otherwise, within the limits of detection by our technique, P units were effectively separated by E sequences. In Table 1 it is shown that the compositions found by NMR are in satisfactory agreement with those expected for 90% conversion on the basis of the reactivity ratios. The ranges of compositions expected within each sample are also shown and it can be seen that only samples EP90 and EP85 can be regarded as even approximately homogeneous. Sample EP85 was not investigated by NMR and its composition was calculated, as indicated in Table 1. Copolymer -Sal t Mixtures Copolymer-salt mixtures were prepared by dissolving vacuum-dried copolymer (ca.0.1 g) and the required quantity of dry LiClO, in dry acetonitrile (ca. 5 cm3) under dry argon. Compositions were prepared for 0:Li =12, 18, 24 or 30, where 0:Li is the mole ratio of chain units (oxyethylene and oxypropylene) to LiClO,, which correspond to molalities in the range m=0.7 to 1.8 mol (kg polymer)-'. After shaking (1 h) the solution was transferred to a Teflon plate and the solvent evaporated under a dry argon flow, followed by mild heating under vacuum (40 "C, <0.01 mmHg, 24 h). Thermal Analysis Thermal analysis was by DSC by Perkin-Elmer DSC-7 (University of Lancaster) for the copolymers and by Perkin- Elmer DSC-4 for the polymer electrolytes. Melting T,, and glass transition q,temperatures were obtained from the DSC curves as the temperatures at the peaks and the mid-points, respectively.Enthalpies of fusion, A,,H, were obtained from peak areas. The temperature and power scales of the calor- imeters were calibrated by melting indium, and the tempera- ture calibration was checked by melting point standards. Under the conditions of use, the thermal lag of the DSC-4 was found to be about 2 K at a heating rate of 10 K min-'. Melting and glass-transition temperatures were corrected Table 1 stat-Copoly(oxyethylene/oxypropylene)s copolymer composition, xEa calculated feed calculated instantaneous composition, observed average (0-90% sample xEO EP64 0.55 EP65 0.60 EP72 0.70 EP79 0.75 EP85 0.80 EP90 0.85 xEk0.02 (estimated).M,, averagec (90% conversion) conversion) M,,/i05 g mol-' 0.64 0.61 0.78-0.06 5.4 0.65 0.66 0.81-0.15 4.5 0.72 0.76 0.86-0.42 4.6 0.79 0.81 0.89-0.54 4.0 -~ 0.85 0.91-0.65 3.7 0.90 0.89 0.93-0.75 4.5 + 10% (estimated). From NMR. J. MATER. CHEM., 1994, VOL. 4 accordingly. A similar thermal lag was assumed for the DSC-7. Experimental uncertainties were established by replicate measurements. 'As-prepared' samples of the copolymers (5-10 mg,k0.005 mg) were dried under vacuum (<0.0 1 mmHg, 25 "C, >12 h) before being sealed into aluminium pans in a dry box. These samples were cooled in the calorimeter to 0 "C and then heated at 10 K min-' to ca.80°C in order to detect the melting transition. The molten samples were then cooled at -1'0 K min-' to -100"C and the experiment repeated. Samples were also quenched (-320 K min-') from 20-30 K above their melting temperature to -100 "C, and re-heated at 10 K min-' to 80°C. Dry samples of the copolymer-salt mixtures (5-10 mg, &0.005 mg) were sealed into aluminium pans under dry conditions as described above. These samples were cooled in the calorimeter to -100"C and then heated at 10 K min-' either to ca. 80°C or, for certain samples, to 150°C. The molten samples were then quenched (-320K min-l) to -100"C and reheated. Conductivity Conductivities of the copolymer-salt mixtures were deter- mined over a range of temperatures by means of a Hewlett- Packard 4192A impedance analyser, operated over the range 5 Hz to 13 MHz. The resistance of an electrolyte was obtained as the point where semicircle and the extrapolation of the inclined spur cut the real impedance axis.A dry polymer- electrolyte film, prepared as described above, was sandwiched between two gold-plated blocking electrodes and held in place by springs within the measuring cell. This operation took place under a dry argon atmosphere with a slight positive pressure of argon maintained within the cell. The assembled cell was placed in a temperature-controlled oven (& 1 K) and conductivities were determined at several temperatures in the range 20-90°C, either on heating from 20°C or on cooling from 90°C.About 30 min was allowed for thermal equili- bration at each temperature. Film thickness was monitored at all times by means of a travelling microscope (+O.OOl cm). The reliability of the conductivities was checked by repetitive measurements of the same preparation (&2%)and of different preparations (k10%).Readings taken on heating and cooling showed no significant differences. Results and Discussion Thermal Properties Examples of DSC curves found for the copolymers themselves are shown in Fig. 2 and 3, and a summary of the results is given in Table 2. The DSC curves of all 'as-prepared' samples had multi-peaked endotherms in the temperature range 30-50 "C (e.g.Fig. 2). The enthalpies of fusion were small and consistent with extents of crystallinity in the range 1-20%, 1787 r I I I I I 20 40 60 TIT Fig.2 DSC curves obtained for as-prepared copolymers EP72 and EP85.The curves were obtained at a heating rate of 10 K min-l and are presented without correction for thermal lag. Ordinate scales and baseline slopes are arbitrary. r vEP7*t--50 0 50 TI'C Fig. 3 DSC curves obtained for copolymers EP72 and EP85 quenched from the melt. The curves were obtained at a heating rate of 10 K min-' and are presented without correction for thermal lag. Ordinate scales and baseline slopes are arbitrary. compared with fully crystalline p~ly(oxyethylene),'~ A,,H25210 J g-'. Besides evident glass transitions, the DSC curves of quenched samples EP72 to EP90 (e.g.Fig. 3) con-tained broad endothermic melting peaks at temperatures towards the bottom of the melting ranges found for 'as- prepared' samples. Melting peaks were not detected in the DSC curves of quenched samples EP65 and EP64, which is in agreement with the conclusions of other group^."^'^ The DSC curves of cooled samples (-10 K min-') were similar to those of quenched samples. The expected increase in Tp with increase in E content was obscured by experimental uncertain ties. Multiple melting of the as-prepared samples is attributable, at least in part, to non-uniform composition along the lengths of the copolymer chains. Similar effects have been observed Table 2 Thermal properties of stat-copoly(oxyethylene/oxypropylene)s copolymer T,/"c" (quenched) melting range/"C (quenched) melting range/"C (as prepared) AfusHIJ g-'(as prepared) EP64 -69 24-48 2 EP65 -73 - 27-46 2 EP72 -72 10-40 26-55 16 EP79 -70 10-46 30-56 11 EP85 -73 8-53 30-57 22 EP90 -65 5-55 30-58 38 a <k3 K (estimated).A& & 5 J g-' (estimated). in related systems: e.g. partially isotactic poly(oxypropylene).20 Evidence of cold crystallisation is difficult to pick out in Fig. 3 and is a consequence of its spread over a wide temperature range, much as expected for statistical copolymers with wide distributions of crystallisable sequence lengths.'l The effect of annealing quenched sample EP90 was investi- gated. The DSC curve of the quenched copolymer contained a small melting peak with maximum at T, =37 "C.Annealing at 35°C for 1h, followed by quenching in the DSC, resulted in a melting peak with a maximum at 42°C. Annealing this sample at 40°C for 0.5 h resulted in a further increase in the melting maximum to 48°C. After each step, the area of the melting peak had reduced, indicating a process of fractionation rather than crystal perfecting, which was a consequence of the wide composition distribution. All the copolymers were mixed with LiCIO, at a mole ratio 0:Li of 24: 1. These samples are denoted EP64-24 etc. Examples of DSC curves obtained for as-prepared samples are shown in Fig. 4. The curves obtained for samples EP64-24 to EP79-24 clearly showed two low-temperature transitions, which are assigned to distinct two glass transitions, denoted 19 EP65-24 -50 0 50 T/"C Fig.4 DSC curves obtained for copolymers EP65, EP79 and EP90 mixed with LiClO, (0:Li =24, m =0.85-0.92 mol kg- ').The curves were obtained at a heating rate of lOK mind' and are presented without correction for thermal lag. Ordinate scales and baseline slopes are arbitrary. J. MATER. CHEM., 1994, VOL. 4 Tgl and Tg2.The lower transition at Tgl was barely detectable in the DSC curve of as-prepared sample EP90-24 (see Fig. 4), but was more apparent in the DSC curve of a quenched sample (not shown). Melting peaks attributable to crystalline polymer were observed in the DSC curves of samples EP72-24 to EP90-24, but corresponding curves recorded for quenched samples showed no melting peaks.Values of Gl and Tg2,the melting range, and Afu,H are listed in Table 3. For all samples, Tgl was approximately constant at -7O"C, while values of T2were in the range -47 to -37 "C, depending on composition. Melting peaks in the temperature range 25-45°C were similar to those found for the copolymers alone. Copolymer EP85 was mixed with salt at four concentrations in the range 0:Li= 12-30, see Table 4. Crystallisation of as-prepared samples was largely suppressed at the highest salt concentration (0:Li =12). Two glass transitions were observed for samples EP85-30 and EP85-24. However, the DSC signal at the lower glass transition (GIz -70 "C) was markedly reduced as salt concentration was increased/and the transition was not observed with certainty for samples EP85-18 and EP85-12.As salt concentration was increased, the DSC signal at the higher glass transition increased, as did the value of q2itself. Two glass transitions in mixtures of poly(oxypropy1ene) and LiClO, in the concentration range 0 :Li= 11-20 have .~~been reported by Vachon et ~1 The two transitions are attributable to dilute (Tgl) and concentrated (Tg2) salt solu- tions, the dilute phase having a glass-transition temperature essentially the same as that of the parent copolymer, and the concentrated phase having its glass-transition temperature (Tg2) raised by the increased cohesion in the salt-copolymer mixture. The optical clarity of the copolymer-salt mixtures suggests microphase separation.22 Presumably, the effect in poly(oxypropy1ene) is a consequence of the low solubility of the salt compared with that in poly(oxyethy1ene).The effect in the present copolymers must also reflect a low solubility, which may well be accentuated by the wide composition distributions. ' The increase in Tp2 with increase in salt concentration is as expected. The effect of salt concentration on Tghas been much discussed in the literat~re.'-~*~~~~~ Because of the complication Table 3 Thermal properties of copolymer-LiCIO, mixtures: O= Li =24 LiClO,/ TgllCCacopolymer mol kg -' (quenched) EP64-24 0.85 -EP65-24 0.85 -71 EP72-24 0.87 -69 EP79-24 0.89 -EP85-24 0.90 -EP90-24 0.92 -" Tk 3 K (estimated). Afu,H& 5 J g-' copolymer LiC10,(0:Li) LiClO,/mol kg-' EP85 12 1.81 EP85 18 1.21 EP85 24 0.90 EP85 30 0.72 "T+3 K (estimated). AfusHt5 J g-' (estimated).T,2/K"(quenched) melting range/"C (as prepared) AfUSHlJ g-'(as prepared) 70 -37 - - -37 - - -45 30-50 10 70 -42 30-55 12 70 -47 25-55 20 71 -46 20-55 28 mixturesThermal properties of EP85-LiC10,Table 4 T,I/"C" T,,/K"(quenched) (quenched) melting range/"C (as prepared) AfusHIJ g -(as prepared) --23 -30 3 --35 20-55 15 -70 -47 25-55 20 -70 -48 35-55 16 Transition at Tgl uncertain. J. MATER. CHEM., 1994, VOL. 4 of phase separation, the present results do not add to our understanding of this aspect of polymer electrolyte chemistry.Conductivities The observation of two glass transitions, indicative of phase separation, pertains directly to low temperatures: T< -20 “C. However, solubilities of salts in poly(oxyethylene), poly(oxy- propylene) and their copolymers are known to have negative temperature coefficients,26327 as would be expected if strong solute-solvent interactions lead to a negative enthalpy of solution. Consequently detection of liquid-liquid phase separ- ation at low temperatures implies similar phase separation at high temperatures. Moreover the polymer electrolytes may be partly crystalline at temperatures below 50 “C. This compli- cated phase behaviour is not obviously reflected in the tem- perature dependences of conductivity of the present systems.This is illustrated in Fig. 5, in which the conductivities of copolymer EP85 mixed with LiC104 at three concentrations are plotted as log CJ us. T-l. These Arrhenius plots are curved, as is usual for polymer electrolytes. As illustrated in Fig. 6, linear plots were obtained by use of the Vogel-Tammann-Fulcher equation” with suitably chosen reference temperatures. The VTF equation is: CJ =o0 exp [-B/(T-To)] 2.8 3.0 3.2 3.4 lo3 WT Fig. 5 Arrhenius plot. Logarithm of conductivity (0)us. reciprocal temperature for copolymer EP85 mixed with LiClO, in the 0:Li mole ratios, of 12 (O), 18 (a)and 24 (A).The results (not shown) for 0:Li =30 overlap those for 0 :Li =24. r I I I I I 6 8 10 12 14 103 W(T-T~) 1789 where go=A/T”’, and A and B are parameters, assumed to be independent of temperature, reflecting the number of charge carriers and the apparent activation energy for segmental motion, respectively.According to the VTF cquation, assuming that parameter B is independent of salt concen-tration, the local viscosity is determined by the difference between the temperature of measurement and Tg. Plots based on T,= q2-20 gave adequate straight lines: e.g. Fig. 6. These and similar plots were interpolated to obtain isothermal conductivities at 298 or 330 K and isoviscous conductivities (i.e. conductivities adjusted to a constant value of T-T,) at q2+ 105 K, the latter corresponding to temperatures (330-355 K) above the melting range of the copolymer-salt mixtures.In carrying through the calculations in this way it was assumed that the conductivity in these systems was effected predominantly through the concentrated-sal t phase. In fact, the variation of q2with overall salt concentration found in the present work was similar to that found2’ for non-crystalline, single-phase, oxymethylene-linked-poly(oxy-ethylene)-LiClO, electrolytes, which lends support to this assumption. The effect of salt concentration on the Conductivities of EP85 electrolytes is shown in Fig. 7. As salt molality was increased beyond 1 mol kg- ’, the isothermal conductivity (330 K, 57°C) fell from a maximum value of cu. 1.5 x S cm-’, see Fig. 7(u).This fall at high salt concen- tration can be ascribed to an increase in local vis~osity.~~’~,~~~~~ Isoviscous conductivities increased regularly with salt concen- tration, see Fig.7(b).The straight line drawn in Fig. 7(b)has a slope of unity, consistent with a one-to-one correspondence between isoviscous conductivity and salt concentralion, as found for other polymer-electrolyte sy~tem~.~~~~~~~~~~ The effect of copolymer composition on conductivity is shown in Fig. 8. The results are for salt concentration 0 :Li = -0.2 0.0 0.2 -3.5:-4.0 IfI I I 1 Fig.6 VTF plot. Logarithm of T1’20 cs. l/(T-To) for copolymer EP85 mixed with LiC10, in mole ratio 0:Li at mole ratios of 12 (O), 18 (m)and 24 (A),The reference temperature Tois taken to be T,,-20, where Tg2is the temperature at the higher of the two glass transitions. The results (not shown) for 0:Li= 30 overlap those for 0:Li= 24.Fig. 7 (a) Logarithm of isothermal conductivity PS. log (salt rnolality) for copolymer PE85 mixed with LiClO,: T= 330 K. (b) Logarithm of isoviscous conductivity us. log (salt molality) for copolymer PE85 mixed with LiClO,: T= Tg+ 105 K; slope = 1. 0-4c 0 0.6 0.7 0.8 0.9 XE Fig. 8 Logarithm of conductivity us. mole fraction oxyethylene units (xF) for PE copolymers mixed with LiClO, (O:Li=24, mz0.9 mol kg-I): 0, T=330K (57°C); B,T=298 K (25°C) 24 (rnz0.9 mol kg-') and for temperatures T= 330 K (57 "C) and T=298 (25 'C). It is clear that conductivities were generally reduced by reducing the oxyethylene content of the copolymer.The trend in G2(Table 3) is towards higher values as xE is reduced, which is in keeping with the conductivity results. At the lower temperature there is an indication (but not outside our experimental error) of a maximum in conduc- tivity at ca. 80 mol% E, somewhat as found by Florjanczyk et al. for EP copolymer-NaI electrolytes.12 As a result of crystallinity, the conductivities of poly(oxyethy1ene)-based electrolytes are very low at, or about, room temperature, so a maximum in conductivity as the E-content approaches 100 mol% is expected. Comparison with Other Polymer Hosts Compared with other n on-crys talline, poly (ox yet hy1ene)- based polymer hosts, which are good solvents for LiClO,, the levels of maximum conductivity in the EP copolymer systems are significantly lower.Approximate maximum conductivities are set out in Table 5. There is a clear advantage in using a modified poly(oxyethy1ene) rather than an EP copolymer. In this respect, copolymer ED70 (Table 5) is essentially a modi- fied poly(oxyethylene), since the side chain contains two oxyethylene units: -CH2(OCH2CH2),CH3. The advantage of POMOE400 over copolymer ED70 lies in a more efficient suppression of ~rystallinity.~,~ Comparison with Low-molar-mass stat-Copoly (oxyethylene/oxypropylene) Cameron et al. have reported thermal and electrochemical properties for polymer electrolytes based on liquid stat-copoly (oxyethylene/oxypropylene)s: e.g. M, in the range 1700-33OOg mol-l, 50 or 75 wt.% E.27,30.31The concen-Table 5 Approximate maximum conductivities for non-crystalline PO-based polymer hosts T' C EP85" /to5 S cm-' ED70h/ lo5 S cm-' POM0E40Oc/ lo5 S cm-' 57 10 10 40 20 1 2 3 LiClO, plus copolymer EP85, Mp,(GPC)%4x lo5 g mol-' (present work); 'LiCF,SO, plus statistical copolymers of ethylene oxide and digol methyl glycidyl ether, M,,.(GPC)% 1 x lo6 g mol-';14 LiCIO, plus oxymet hylene-linked poly(oxyethy1ene) prepared from PEG400, Mpk(GPC)% 1.4 x lo5 g mol-'.25 J. MATER.CHEM., 1994, VOL. 4 tration dependence of molar conductivity established in their work has served as a guide to conductivities in other non- crystalline polymer-electrolyte system^.^^'^^ So far as the pre- sent study is concerned, the bFneficia1 effect on conductivity of a high oxyethylene content, seen in the present results (Fig.8), is equally seen in the results for liquid system^.^^.^' Cry stallisation and melting occurred at substantially lower temperatures than found for the present high-molar-mass samples: e.g. melting from -30 to +2O"C for a liquid copolymer with 80 mol% E (75 wt.% E) compared with +30 to +55 "C for sample EP79.31 Characterisation by 13C NMR of the commercial copolymers referred to in ref. 30 and 31 has shown32 that their sequence length distributions were identical to those expected in the absence of composition drift,I6 i.e. that their average E-sequence lengths were shorter than those of the present copolymers. This characteristic, coupled with the effect of chain ends, makes close comparison of results from the two systems difficult.Just one glass-transition temperature was found3' for mixtures of liquid copolymer (80 mol% E) with lithium perchlorate, which may be a consequence of the difference in sequence length distri- bution, but equally of the better miscibility expected in a low- molar-mass system. Concluding Remarks The present study touches on a number of aspects of the chemistry of polymer electrolytes based on high-molar-mass statistical copoly(oxyethylene/oxypropylene)s. Separation into salt-poor and salt-rich phases, similar to that found for poly(oxypropy1ene)-LiC104 electrolytes,22 was identified via DSC. This phase separation was most easily detected in electrolytes of low salt content formed from copolymers of low E content. Presumably, the conductivities of these electro- lytes were limited by this solubility effect.At the other extreme of high E content, 'room temperature' conductivity was limited by crystallisation. A maximum conductivity at 25 'C of G% lo-' S cm-' was found for a copolymer with E content = 85 mol% mixed with LiC10, at a 0:Li mole ratio of 24: 1. For that copolymer, EP85, the extent of phase separation was small, and both the glass-transition temperature assigned to the salt-rich phase (q2)and the conductivity depended on overall salt concentration in a manner consistent with results for comparable single-phase systems.21 Thanks are due to the Iranian Government and the Science and Engineering Research Council for financial help.Drs. C. V. Nicholas, F. Heatley, J. H. Thatcher and G. E. Yu gave valuable advice, and Mr. K. Nixon and Mr. M. Hart gave practical help. Dr. J. Ebdon kindly made arrangements for use of equipment at Lancaster University. Our referee drew attention to ref. 31, which appeared just as our paper was submitted, and suggested inclusion of a comment in our discussion. References Polymer Electrolyte Reviews ed. J. R. MacCallum and C. A. Vincent, Elsevier, London, 1987, Vol. I; 1989. vol. 2. J. R. Owen, in Covlzprehensior Poljwier Science, ed. C. Booth and C. Price, Pergamon Press, Oxford, 1989, vol. 2. ch. 21. F. M. Gray, Solid Polymer Electrolytes, VCH, Cambridge, 1991.P. G. Bruce and C. A. Vincent, J. Clzem. SOL..,Furudnj) Truns., 1993,89,3187. C. Booth, C. V. Nicholas and D. J. Wilson, Polymer Electrolytc~ Reviews, ed. J. R. MacCallum and C. A. Vincent, Elsevier. London, 1989, vol. 2, ch. 7. B-X. Liao, Y-M. Chen, C. Booth and Y-Z. Luo. Polym. Commun., 1991,32, 348. J. MATER. CHEM., 1994, VOL. 4 7 8 9 10 11 12 13 14 15 16 17 18 19 20 J. H. Thatcher, K. Thanapprapasr, S. Nagae, S-M. Mai, C. Booth and J. R. Owen, J. Muter. Chem., 1994,4,591. F. E. Bailey and H. G. France, J. Polym. Sci., 1960, 45,243. E. J. Vandenberg, J. Polym. Sci., Part Al, Polym. Chem., 1969, 7, 525. Ph. Teyssie, T. Ouhadi and J. P. Bioul, International Review of Science, Macromolecular Science, Physical Chemistry, ed.C. E. H. Bawn, Butterworths, London, 1975, series 2, vol. 8, p. 191. C. Booth, W. C. E. Higginson and E. Powell, Polymer, 1964, 5,479. Z. Florjanczyk, W. Krawiec, W. Wieczorek and J. Przyluski, Angew. Mukromol. Chem., 1991,187, 19. J. Przyluski and W. Wieczorek, Solid State Ionics, 1992, 53-56, 1071. D. G. H. Ballard, P. Cheshire, T. S. Mann and J. E. Przeworski, Mmromolecules, 1990, 23, 1256. P. Passiniemi, S. Takkumaki, J. Kankare, M. Syrjama, Solid State lonics, 1987,28-30, 1001. F. Heatley, G-E. Yu, C. Booth and T. G. Blease, Eur. Polym. J., 1991.27, 573. T. N. Kurengina, L. V. Alferova and V. A. Kroprachev, Vysokomol. Soedin., Ser. B, 1971, 13, 419: Chem. Abstr., 1972, 75,118631h.F. Heatley, Y-2. Luo, J-F. Ding, R. H. Mobbs and C. Booth, Macromolecules, 1988,21,2713. D. R. Beech and C. Booth, J. Polym. Sci., Part B, Polym. Phys., 1970,8, 731. C. Booth, C. J. Devoy, D. V. Dodgson and I. H. Hillier, J. Polym. Sci., Part A2, Polym. Chem., 1970,8, 519. 21 L. Mandelkern, Crystallisation of Polymers, McGraw -Hill, New York, 1964, ch. 8. 22 C. Vachon, M. Vasco, M. Perrier and J. Prud’homme, Macromolecules, 1993,26,4023. 23 J-F. LeNest, A. Gandini and H. Cheradame, Br. Polyw. J., 1988, 20, 253; H. Cheredame and J-F. LeNest, Polymer Electrolyte Reviews, ed. J. R. MacCallum and C. A. Vincent, Elsevier, London, 1987, vol. 1, ch. 5. 24 J. M. G. Cowie and A. C. S. Martin, Polymer, 1991,32,2411. 25 S. Nagae, M. Nekoomanesh H., C. Booth and J. R. Owen, Solid State lonics, 1992,53-56, 1118. 26 D. Teeters and R. Frech, Solid State Ionics, 1986, 18-19.271. 27 G. C. Cameron and M. D. Ingram, Polymer Electrolyte Reviews ed. J. R. MacCallum and C. A. Vincent, Elsevier, London, 1989, vol. 2, ch. 5. 28 H. Vogel, Phys. Z., 1921, 22, 645; G. Tammann and W. Hesse, 2.Anorg. Allg. Chem., 1926,156,245; G. S. Fulcher, J. Am. Cerum. SOC.,1925,8, 339. 29 M. Nekoomanesh H., S. Nagae, C. Booth and J. R. Owen, J. Electrochem. Soc., 1992,139, 3046. 30 G. C. Cameron, M. D. Ingram and G. A. Sorrie, J. Chem. SOC., Faraday Trans. I, 1987,83,3345. 31 G. C. Cameron, M. D. Ingham, M. Y. Qureshi, G. M. Russel and G. I. Wood, Polym. Znt., 1994,33, 347. 32 F. Heatley, personal communication. Paper 4/03157A; Received 26th hluy, 1994