Poly [oxymethylene-oligo (oxyethylene)] network electrolytes Shao-Min Mai Robert A. Colley Jane H. Thatcher Frank Heatley Peter M. Budd* and Colin Booth Manchester Polymer Centre and Department of Chemistry University of Manchester Manchester UK M13 9PL Poly [oxymethylene-oligo(oxyethy1ene)l s have been prepared with a proportion of the chain units having pendant unsaturated groups Radical chemistry was used to crosslink either the polymers alone or mixtures of the polymers with LiClO Swelling and tensile measurements were used to determine the extent of crosslinking and differential scanning calorimetry and dynamic mechanical thermal analysis to determine glass transition and melting temperatures Conductivities of network-salt mixtures were measured by ac impedance spectroscopy The conductivities and thermal properties of crosslinked and uncrosslinked materials were essentially identicalIonically conducting polymer electrolytes based on lithium .salts and high-molar-mass poly(oxyethy1ene) (POE) have been studied extensively over the last two decades Unplasticised POE electrolytes have low conductivities at ambient tempera- ture (e g o< S cm ' at 25 "C) and a number of modified polymer-salt systems with relatively high ambient-temperature conductivities (o<10-5-10-4 S cm-' at 25 "C) have been developed the work has been well reviewed' One such polymer is poly[oxymethylene-oligo(oxyethy1ene)l (POMOE) The repeat unit is -OCH2(0CH20CH2),- where n an average value over a narrow distribution of oxyethylene sequence lengths is typically in the range 4-20 Starting from polyethylene glycol 400 (PEG400 n z9 l),poly-mers of number-average molar mass greater than 100000 g mol -'have been prepared corresponding to chains containing on average about 200 repeat units The oxymethylene groups (acetal groups) which link the oxyethylene sequences disrupt the crystallinity With a suitably chosen value of n these polymers can be non-crystalline at room temperature and all have low glass transition temperatures i e Tg 2 -65 "C In common with other elastomers with flexible chains e g poly(dimethylsi1oxane)s and poly( phosphazene)s non-crystal- line POMOE is a soft rubber and is subject to creep The present work was directed towards removing this disadvan- tage by including sites for crosslinking along the polymer chain thereby allowing formation of a true network after mixing the polymer with the salt in the usual way It was expected that the local environment in the polymer would be unchanged by limited crosslinking and so since the conduc- tivity of a polymer electrolyte depends on local viscosity (segmental mobility) it was expected that the conductivity would not differ significantly between crosslinked and conven- tional forms of the polymer electrolytes Indeed this was our experience when working with polymer electrolytes formed from highly branched POMOE,* and also the experience of Sloop et d9who subjected conventional POMOE to UV irradiation in the presence of benzophenone and found no significant change in conductivity in polymer electrolytes on crosslinking The normal reaction used in preparing POMOE is between PEG (eg PEG400) and dichloromethane (DCM) in the pres- ence of an excess of powdered KOH It proceeds through formation of a chloroether -OCH2CH20H+CH2C12-+ -OCH,CH20CH2Cl +HCl followed by very rapid reaction of the chloroether with a second hydroxy group to form an oxymethylene link __ --OCH2CH20CH2C1+HOCH2CH2-+ -OCH2CH20CH20CH2CH2-Because of the extreme difference in reactivities of the chloromethylene and chloroether groups the polymerisation proceeds effectively as an RA2 self-condensation," yielding high-molar-mass products without need for balancing the concentrations of reagents as would be necessary to achieve high molar masses in a conventional RA2 +RB2 polycondens- ation lo Indeed the difference in reactivities is so large that the DCM can be used as the solvent for the reaction ie at a molar ratio [Cl]/[OH] x 10 (see later).Reactions starting from PEG400 or PEG200 were adapted to our present purposes by the inclusion of a diol bearing an unsaturated group In the work described below the diol used was normally 2-methylenepropane-l,3-diol(MPD) which was included in the polymerisation recipe at levels of 1-10 mol% relative to PEG One polymer was made using a second diol hexa-1,5-diene-3,4-diol (HDD) The diol residues were incor- porated statistically in the chain but had an insignificant effect on the melting point since that was determined by the dis- ordered-block structure of the linked poly(oxyethy1ene) chain.In related work Alloin and co-workers'' l3 explored poly- condensations based on the reaction of PEG with 3-chloro-2- chloromethylprop-1-ene [CCMP ClCH2C(=CH2)CH2Cl] under similar conditions to those described above This method of introducing unsaturated units has the advantage of avoiding the acetal link which is sensitive to hydrolysis in the presence of Lewis acids Enhanced reactivity of the substituted CCMP was noted,13 but the molar mass obtained with PEG400 (ca 30000 g mol ') was significantly lower than that normally obtained for PEG400 reacted with DCM Moreover the double-bond content in the polymers prepared in the present work was not fixed by the choice of PEG precursor (ie by the average oxyethylene-sequence length) as it is in the CCMP- linked systems Finally the hydrocarbon content (as distinct from oxyethylene and oxymethylene content) in DCM-linked PEG400 is low (< 1 mass%) compared with ca 13 mass% for CCMP-linked PEG400 These advantages of DCM linked polymers may well outweigh the disadvantage of the lability of their acetal links which is a significant problem only for impure systems14 or for electrolytes based on trivalent cations Notation.The notation used for the present polymers indicates the molar mass of the PEG precursor the diol (M or H) and the mol% J Muter Chem 1996 6(7) 1099-1 106 1099 of diol (relative to PEG) used in the polymerisation z e P400-M5 denotes a crosslinkable polymer prepared from PEG400 with 5 mol% MPD Conventional polymer (no MPD) is denoted by eg P400 Networks are denoted by adding the letter N e g P400-M5-N Polymer electrolytes formed from the polymers by adding LiC10 are distinguished by adding a number denoting the mole ratio (O/Li) of salt in the material e g P400-M5-N30 is a network polymer electrolyte containing LiClO at O/Li =30 Experimenta1 Preparation of polymers Polyethylene glycols PEG400 and PEG200 (Fluka AG) were dried under vacuum (10-3mmHg RT 24 h) before use Dichloromethane (DCM Fison) was purified by fractional distillation KOH (85 mass% BDH) was powdered immedi- ately before use 2-methylenepropane- 1,3-diol (MPD Aldrich) and hexa-l,5-diene-3,4-diol (HDD Aldrich) were used as received The polymerisation reaction followed the general procedures described previously For example a sample of conven-tional poly [oxymethylene-oligo (oxyethylene)] P400 was pre- pared by reacting PEG400 with an excess of dichloromethane in the presence of an excess of finely powdered KOH Finely ground potassium hydroxide (50 g) was mixed with dichloro- methane (40cm3) under nitrogen at room temperature in a resin kettle equipped with a condenser To this was added PEG400 (50 g) and the whole was stirred until the viscosity became too high (ca 30 min) More DCM (10 cm3) was added to allow stirring for a further period (ca 15 min) The polymeric mass was allowed to stand under nitrogen for a further 16 h after which it was divided and dissolved in additional dichloro- methane (1 5 dm3) and filtered through compacted diato-maceous earth Residual potassium (determined by microanalysis) after this treatment was below 0 1 mass% The rubbery polymer was isolated by rotary evaporation before finally drying on a vacuum line (12 h mmHg) and storing at low temperature in a refrigerator.Crosslinkable polymers were prepared in a similar way but with inclusion of MPD or HDD in the recipe Hydroquinone (002 mol% based on diol) was added to certain reactions as a precaution against premature crosslinking All polymers were examined by gel permeation chromato- graphy (GPC) Tke system comprised 4 p-Styragel columns (porosity 500-106A) with N N-dimethylacetamide (DMA) at 70°C as eluent at a flow rate of 1 cm3 min-l Samples were injected through a 100mm3 loop at a concentration of 2g dm-3 The emerging polymer was detected by differential refractometry The system was calibrated with eight poly(oxy- ethylene) standards covering the molar mass range 1500-106 g mol-I The elution volume of water (present as an impurity in the undried solvent) was used as flow-rate marker.One broad polymer peak was found in the GPC curves of each of the samples with little or no signal attributable to low-molar- mass cyclics The broad peaks were characteristic of conden- sation polymers z e of most probable distributions lo It was convenient to characterise a polymer by the molar mass at the peak (Mpk)and examples of the values obtained are listed in Table 1 together with corresponding values of degree of polym- erisation (DP) As can be seen high-molar-mass polymers were consistently produced from PEG400 by our procedure with lower molar masses recorded for polymers prepared from PEG200 The approximate DP values listed in Table 1 show that the low molar masses of the 200 series are not solely a result of the lower molar mass of the starting glycol but reflect a lower efficiency of the reaction with respect to chain extension The general chain structure i e alternating oxymethylene 1100 J Muter Chem 1996 6(7),1099-1106 Table 1 Conventional and crosslinkable POMOE P400 18 450 P400-MO 5 18 450 P400-M 1 06 17 430 P400-M2 18 17 430 P400 M5 35 17 430 P200 -0 54 270 P200-M5 42 0 40 200 P200-M 10 74 0 55 280 P200-H5 32 0 34 170 Too low a level to measure and oligo(oxyethy1ene) chain units was confirmed by ‘H and 13C NMR spectroscopy much as described previously l5 In the case of the crosslinkable polymers NMR was also used to verify the incorporation of unsaturated groups into the chain The following labelling scheme was used for polymers prepared with MPD -OCH20CH2CH2[OCH2CH2]n-20CH2CH20CH20 abccc cb a CH2C(=CH2)CH20-defd The assignments (with chemical shifts taken from the spectrum of sample P400-M5 0) were a b C d e f 8 467 368 362 406 - 5 16 8 954 668 704-705 682 1419 1142 Comparison of integrals gave the mole percentages of MPD residues incorporated into the chain which are listed in Table 1 A similar analysis of I3C NMR spectra was carried out for the polymer prepared with HDD the chemical shifts of the carbons in the unsaturated side groups being 6 133 8 (-CH=) and 6 118 4 (=CH2) As can be seen 100% incorporation of diol into the polymer was not achieved but the proportion was generally high.Preparation of polymer networks When necessary the inhibitor used for protection during polymerisation was removed by precipitating the polymer from solution in toluene by adding heptane The precipitated poly- mer was separated by vacuum filtration redissolved in dichloromethane and finally dried under vacuum ( lop3mmHg RT >24 h) Crosslinking reactions were carried out using benzoyl per- oxide (BPO BDH Ltd ) or a,@’-azobisisobutyronitrile(AIBN BDH) as thermal initiators or photochemically using UV radiation and benzophenone (BzPh Fison) as sensitizer BPO or AIBN were added to the polymers at approximately 50 mol% based on double bonds This reflected the expected chemistry of the reaction z e two radicals formed per initiator molecule and crosslinking by combination BzPh was added at 0 5 mass% based on polymer regardless of the double-bond content ze at about one-tenth of the concentration used by Sloop et a/’ This was intended to reduce crosslinking by hydrogen abstraction from saturated chain units relative to crosslinking uza the unsaturated groups For the examples reported here all involving polymers based on PEG200 this concentration of BzPh corresponded to 5-15 mol% based on double bonds reflecting its role as a sensitizer rather than a primary source of radicals.Preliminary experiments served to define satisfactory con- ditions for crosslinking with the observation of limited swelling with retention of shape of a sample immersed in water being used as a simple indicator of formation of a satisfactory network. Crosslinking with AIBN was more satisfactory than that with BPO and that initiator was used for thermally initiated crosslinking in subsequent work In the experiments a solution of polymer (15 g) and the required quantity of AIBN or BzPh in dry acetonitrile (10 cm3) was poured into a PTFE dish under dry nitrogen and the solvent slowly evapor- ated to form a thin film (ca 1mm) before finally drying extensively in vacuum For thermal crosslinking films were heated to 70 "C under dry nitrogen for 24 h while for photoch- emical crosslinkingfilms were exposed under dry nitrogen to UV radiation from a Spectroline R-51/F (short wave UV 140 W) lamp at a distance of between 15 and 20 cm from the films for the same time period Films were turned once during exposure to promote an even cure.Polymers with no unsaturation (P400 P200) and that prepared from HDD (P200-H5) were not thermally crosslinked under the conditions employed All polymers (with or without unsaturation) could be crosslinked by UV radiation in the presence of BzPh but crosslinking was much improved if unsaturation was present For example polymer P200 was not crosslinked after exposure to the UV radiation for 72 h (I e soluble in water) and was poorly crosslinked after 96 h (I e highly swollen in water with loss of shape) On the other hand polymer P200-M5 was adequately crosslinked after exposure for 24 h (I e limited swelling in water with retention of shape) Network polymers in the form of swollen gels were exam- ined by NMR spectroscopy Swelling with CDCl resulted in slow degradation of the network presumably a result of acid- catalysed hydrolysis of the acetal links Swelling with D,O gave long-term stable systems Attention was focused on the resonances assigned to =CH eg at 6 117 for MPD-crosslinked systems (note 6 114 in CDC1,) The integral of this resonance was much decreased after crosslinking but typically a small fraction of double bonds (<20%) remained Preparation of polymer electrolyte networks Polymer electrolytes were prepared from samples (purified if necessary see above) which were dried extensively The salt was anhydrous LiC10 (Aldrich) dried by heating under high vacuum (10 mmHg 50 "C 48 h) immediately before use Acetonitrile yas heated to reflux and distilled from molecular sieve type 4A 4-8 mesh All operations were carried out in a dry nitrogen atmosphere in a dry box The procedures for film formation from the polymer-LiC10,-initiator solution In acetonitrile and its subsequent crosslinking followed those described above Polymer electro- lytes with mole ratios O/Li=30 and 50 were prepared for the samples based on PEG400 while a wider range of salt concen- trations (O/Li =10-100) was investigated for the samples based on PEG200 The O/Li ratio was calculated for all oxygens in the chain O/Li mole ratios in the range 25-50 (depending on temperature) are known to give maximum conductivities in the P400-LiClO system l6 Tensile properties .A strip of known area of cross-section was clamped at its upper end and suspended vertically in a constant temperature (+ 1 "C) enclosure Stress-strain measurements were made at 25 or 30 "C z e above the melting range The undeformed length between marks was determined before weights were clamped to the lower end of the strip and the strain (1 ratio of deformed to undeformed length) was measured over a range of applied stress for increasing load A steady reading was achieved after cu 20 min Selected samples were measured with both increasing and decreasing load in order to verify that equilibrium was achieved under the conditions employed Swelling Weighed strips of the network polymers were immersed in water at 25 "C for 3 days before removing surface drying and reweighing.Approximate volume fractions of water in the swollen gel were calculated from the densities of water and liquid poly(oxy- ethylene) at 25 "C (p=1 1g cm-,) Differential scanning calorimetry A Perkin-Elmer DSC-4 instrument was used A sample of polymer or polymer electrolyte (ca 10mg) were dried under vacuum sealed into aluminium pans under dry nitrogen and cooled rapidly (quenched) in the calorimeter to -100°C Starting at this temperature the sample was heated at +10°C min-' to 100 "C The samples were then quenched from +100 to -100°C and re-heated at +1O"C min-l Melting and glass-transition temperatures were obtained from the DSC curves as the temperature at the melting peak and the mid- point of the inflection respectively The correction for thermal lag at the heating rate used was -2"C as determined by experiments on standards at various heating rates Calibration of the power and temperature scales was with pure indium The temperature scale was checked in the temperature range of interest either by melting organic standards or against a sample of POMOE (P400) with established properties Dynamic-mechanical-thermal analysis Dynamic-mechanical-thermal analysis (DMTA) was carried out by means of a Polymer Laboratories MKII DMTA instrument used in the shear-sandwich mode Samples (ca 4mm diameter 1 mm thick) were cooled to -lOO"C and measurements of complex shear modulus (G") were made at frequencies in the range 0 3-50 Hz whilst heating from -50 to +100"C in steps of 5 "C allowing 10 min for equilibration at each temperature Samples were clamped at low tempera- tures to counteract shrinkage but were not further adjusted during heating which meant that values of log,,(G*/Pa) at the higher temperatures were known only to +O 3 Pa as judged by replicate measurements At low temperatures the modulus approached the limit of measurement for the instrument in the geometry used (z e G* z lo8 Pa) which meant that Tg could not be determined by our DMTA technique Conductivity Conductivities of solutions of LiC10 in selected polymers were determined over a range of temperatures (20-80 "C) by ac impedance spectroscopy A Schlumberger Model 1260 impedance/gain phase analyser was used over a frequency range of 5 Hz to 15 MHz The resistance was obtained as the point where the extrapolations of the semicircle and the inclined spike cut the real impedance axis A dried film was sandwiched between two gold-plated brass electrodes and held in place by springs within a cylindrical glass cell This operation took place under a dry nitrogen atmosphere The assembled cell maintained under a slight positive pressure of dry nitrogen was placed in a temperature-controlled oven (Buchi Model TO-51 & 1 "C) The temperature (measured by a thermocouple to lfI0 1 "C) was raised to 80 "C and resistances were deter- mined at several temperatures on cooling from 80 to 20°C About 20 min was allowed for thermal equilibration at each temperature of measurement Film thickness was monitored at all times either by means of a travelling microscope (+O 001 cm) or by mounting a Schlumberger SM3 linearly variable differential transducer (& 0 0001 cm) in parallel with the electrodes The conductivities were obtained from the resistance measurements using cell constants determined from the electrolyte thicknesses and electrode areas in each case J Mater Chern 1996 6(7),1099-1106 1101.Results and Discussion The results described below are for polymers based on PEG400 which were thermally crosslinked and for polymers based on PEG200 which were photochemically crosslinked This separa- tion of technique and polymer type had no particular signifi- cance In part it arose because photochemically crosslinked P400 (our notation) and related polymer-salt mixtures had been studied by Sloop et UE' Our own experiments designed to crosscheck our results (eg on photochemically crosslinked P400-M5) gave consistent results. Characterisationof networks by mechanical properties The tensile properties of several polymer networks were investi- gated in this way Polymer electrolyte networks were too hygroscopic to study by our technique which did not allow for a dry atmosphere As described in Section 2 4 stress-strain curves obtained for selected polymers with increasing and decreasing load showed little or no hysterisis The observation that the networks supported significant stress without creep was very satisfactory in view of the intended end-use of the materials. Plots of nominal stress (o=ratio of tensile force to unde- formed cross-sectional area) us strain function (A-A-~) for increasing load were used to obtain values of the shear modulus (G) from O=G(&3.-2) (1) Example of the plots are shown in Fig 1 and values of G are listed in Table 2 The values of G are much as expected for unreinforced crosslinked elastomers The kinetic theory of elasticity of an ideal affine network18 provided an approximate guide to the network characteristics I I I I 01 (d56 0 05 0 0 02 04 06 08 A-A-2 Fig.1 Nominal stress us strain function for POMOE networks (0) P400-MO 5-N (0)P400-Ml-N (m) P400-M2-N (G) P400-M5-N through the equations G/Pa= vkT and ~/cm-~=N,p/M (2) where v is the density of network chains (I e chains terminating in crosslinks) M is the molar mass of the network chains p is the density of the network and NAis the Avogadro constant For approximate values the density of a network was taken to be that of non-crystalline poly(oxyethy1ene) Values of the network parameters are listed in Table 2 the values of v being converted to concentration of network chains c (mol dm-3) in order to facilitate comparison with published results In fact the values of c obtained were similar to those reported previously' The present results show that low extents of crosslinking adequately support reversible elastic behaviour in these systems.The concentration of network chains (c,) is plotted against the concentration of double bonds (c=) in the precursor poly- mers (determined by NMR spectroscopy) in Fig 2 The approximation pz 1 1 g cm-3 was used for the density (see above) Because of the approximations inherent in deriving c via eqn (1) and (2) and because of the presence of unreacted unsaturated groups in the networks (see Section 22 NMR) the results in Fig 2 cannot be interpreted rigorously However within a realistic estimate of the uncertainty (the error bars in Fig 2 are _+25%) the number of network chains formed corresponds to the number of unsaturated groups introduced This is as it should be in the crosslinked systems since two unsaturated groups produce one crosslink and there are two network chains per tetrafunctional crosslink in the networks As can be seen in Table 2 the photochemically crosslinked networks based on P200 were also characterised by tensile I I 0 200 400 c,/mmol dm-3 Fig.2 Concentration of network chains (c,) us concentration of double bonds (c-) in the prepolymers determined from (0 a)mechanical properties or ( CI)from swelling The prepolymers were derived from (0)P400 and ( m 0)P200 The full line IS c =c-The error bars are f25% Table 2 Network characteristics tensile swelling network G/MPa c,/mol dm M,/g mol ha c,/mol dm AIBN P400-MO 5-N 0 052 0 021 53000 X P400-M 1-N 0 096 0 038 29000 P400-M2-N 0 17 0 067 16000 P400-M 5-N 0 28 0 11 10000 v' UV/BzPh P400-N P200-M5-N 0 14 0 42 0 056 0 17 20000 6500 v' 0 14 0 21 P200-M 10-N 0 59 0 24 4600 0 17 0 35 a J denotes swelling and retention of shape x denotes swelling and loss of shape 1102 J Muter Chem 1996 6(7),1099-1106 properties The moduli of the two sets of networks are in satisfactory correspondence as illustrated by the values of c for the two systems which are compared in Fig 2 As mentioned above sample P200-H5 proved impossible to crosslink thermally but could be crosslinked photochemically However the networks so prepared had irreproducible charac- teristics whether determined by mechanical properties or swelling Accordingly results for this system are not included in Table 2 and are not considered further Characterisationof networks by swelling .The polymer networks swelled in water the more tightly crosslinked samples swelling the least In general swelling was useful for qualitative characterisation of the polymer networks Samples which swelled and retained their shape were judged satisfactory Of the polymer networks studied in detail only the least crosslinked sample (P400-MO 5-N) failed this test (see Table 2) and even that sample (when dry) formed a satisfactory free-standing film with no detectable creep The results reported below for polymer electrolyte networks all refer to samples which had satisfactory swelling properties The highly swollen soft gels were difficult to handle and consequently the extent of swelling could not be determined with precision The values of 42 given in Table 2 are averages of a spread of results Given the extent of swelling the approximate equation for swelling of a network [eqn (3)],19 (3) was used to find the density of network chains In eqn (3) and q52 are volume fractions of solvent and network respect- ively V is the molar volume of the solvent (ca 18 cm3 mol-') and v is in cmP3 Malcolm and Rowlinson2' have shown that the parameter x for poly(oxyethy1ene)-water mixtures depends strongly on d2 eg increasing from 0 4 at q52=025 to 124 at 42=0 9 Here a constant value representative of the semi-dilute range (y=O45) was used in order to maintain comparison with other authors' Values of c determined in this way for photochemically crosslinked networks based on polymers pre- pared from PEG200 and MPD are listed in Table2 The values from swelling compare favourably with those obtained from mechanical properties a systematic difference being explicable in view of the approximations used in the treatments of the two sets of data The extents of crosslinking were similar to those reported by Sloop et al ,9 which emphasises the action of BzPh as a UV sensitizer for crosslinking via the double bonds since in the present work BzPh was added at one- tenth of the concentration used previously The results obtained for the photochemically crosslinked networks based on P200 are compared with those obtained for thermally crosslinked networks based on P400 in Fig 2 The data points for the P200 systems map well onto those for the P400 systems and are in correspondence with the line c,= c= ,consistent with crosslinking under our conditions originat- ing predominantly from the unsaturated groups Differential scanning calorimetry Thermal properties of linear POMOE and its mixtures with salts have been reported previously from both our laboratory and elsewhere The following picture has emerged for POMOE prepared from PEG4005 1621 23 The DSC curves of quenched samples of high-molar-mass polymer show a glass transition in the approximate range -60 to -65°C and complex melting behaviour in the approximate range -20 to +20 "C the latter effect being caused by premelting of unstable crystals followed by recrystallisation to a more stable form Annealing at a few degrees below T leads to a DSC curve with a single melting endotherm peaking at ca 15T7 Enthalpies of fusion measured over the full melting range are typically 50 J g-' or less indicative of an extent of crystallinity of 25-30% Addition of salt raises Tp and suppresses the rate of crystallisation leading to DSC curves of quenched samples showing cold-crystallisation exotherms above the glass trans- ition followed (at low salt concentrations only) by complex melting The extent of crystallinity is supressed on adding salt so much so that essentially complete supression is the rule at O/Li mole ratios <20-25 .The present samples based on PEG400 fit into this general picture Examples of DSC curves obtained for polymers networks and polymer electrolyte networks based on PEG400 are shown in Fig 3 Examples of the values of Tg T and Af,,,H obtained are given in Table 3 Crosslinking made no significant difference to the thermal properties A similar conclusion regarding networks of POMOE prepared from PEG400 was reached by Sloop et al However Alloin et a/ l1 l2 observed a small change in Tg (ca + 6 "C) and elimination of crystallinity for their networks which was probably related to a high crosslink density in their system although their networks were not characterised The thermal properties of POMOE based on PEG200 have been investigated less extensively l4 22 Glass transitions of the high-M polymers have been located consistently in the range -60 to -65 "C Evidence of crystallinity is less consistent with limited crystallinity (eg 15% at low T) and melting at -9 "C reported5 22 on the one hand and zero crystallinity reported14 on the other The DSC curves obtained in the present work for P200 and its crosslinkable and crosslinked derivatives indicated glass transitions at or about -65 C and negligible crystallinity see Fig 4 and Table 3 for examples We speculate that this dichotomy in the crystallinity of the P200 system originates in different chain-length distributions in the various PEG200 precursors In this respect it is that crystallinities of POMOEs prepared from uniform oligoethy- lene glycols are indeed sensitive to E-sequence length in the range E3 to E In Fig 5 values of reciprocal Tp are plotted us salt concen- tration for the P400 and P200 systems.The plot also includes I I I I I PWM5 N 2 loo I 50 I 0 I 50 I 100 0a I I P4WM5 N30 J I I 100 50 0 50 100 TI'C Fig.3 DSC curves for networks and network electrolytes based on PEG400 (a) POMOE networks P400-M1 N and P400-M5-N (h) POMOE network electrolytes (with LiClO,) P400-M5-N50 and P400-M5-N30 All samples were quenched from the melt to -100 C The power scales zeros and slopes are arbitrary The temperature scale is uncorrected for thermal lag Fig.4 DSC curves for networks and network electrolytes based on PEG200 (a) POMOE networks P200-M5-N and P200-M10-N (b) POMOE network electrolytes (with LiClO,) P200-M5-N75 P200- M5-N50 P200-MSN25 and P200-M5-N10 All samples were quenched from the melt to -100°C The power scales zeros and slopes are arbitrary The temperature scale is uncorrected for thermal lag results for a high-molar-mass POMOE (Mpkz 110000) pre- pared from tetraethylene glycol (M = 194 g mol-l) here denoted P194 The results for P400 are in good agreement with others reported for P400-LiC104 systems l6 21 As can be seen within the estimated error the effect of salt on Tg is the same for all systems Dynamic-mechanical-thermal analysis Because of the low modulus of the rubbery materials the DMTA was used in its shear-sandwich mode Consequently the equipment reached its limit at dynamic moduli in the range lO7-lO8 Pa,25 which meant that Tg could not be detected Because of this limitation DSC was the preferred method of thermal analysis and DMTA was used only for selected samples Plots of storage modulus for samples P400 and P400-M1-N are shown in Fig 6 The two plots are essentially identical since the difference at high modulus (low tempera- ture) could well result from a minor difference in sample mounting at the limit of the range of the instrument The large fall in the temperature range 0-20°C is caused by the melting transition The storage modulus at 30°C was ca lo6 Pa as expected for an flexible elastomer and in keeping with the shear modulus determined at that temperature Conductivity As shown in Fig 7(a) conductivities of mixtures of sample P400 with LiC104 at constant temperature (27 "C 300 K) reached a maximum at salt concentration 0 5 g (kg polymer)-' (I e OILiw45) before falling away at high concentrations Plots of this kind have been reported previously for POMOE electrolytes based on PEG400 and a range of salts l6 21 24 The effect has been explained by competition between an increase in concentration of charge carriers and an increase in Fig.5 Reciprocal glass transition temperature us LiClO concen-tration for POMOE chains and networks (0)Linear P400 (+) linear P194 from ref 24 (A) network P400-M2-N,(V) network P400-M5-N (0)network P200-M5-N (0)network P200-M10-N polymer and network systems is seen to extend over the full temperature range examined.I I 5' -;o 0 50 100 TIT Fig. 6 DMTA curves of log,,(storage modulus) us. temperature for POMOE chains and networks (0)linear P400; (A) network P400- M2-N. The estimated uncertainty in log,,G' is k0.2. I I I I As discussed in Section 2.1 it proved difficult to prepare linear POMOE of high molar mass from PEG200 and the prepolymers used to form networks had M,,(GPC) M 50000 g mol-l. As a result polymer electrolytes based on uncrosslinked P200 were less resilient than those based on high-M POMOE (e.g.those of the P400 system) and determination of their conductivities was not attempted. No such problem was enco- untered in measuring the conductivities of the network polymer electrolytes. Examples of Arrhenius plots of the conductivities of polymer electrolytes P200-M5-N25 and P200-M 10-N25 are shown in Fig. 8(u). These are similar one to another and are similar in shape to the corresponding curves shown for P400 systems in Fig. 7(b). Comparison can also be made with results reported24 for polymer electrolytes prepared from P194 i.e. the sample of POMOE prepared from tetraethylene glycol which had a high molar mass (MPkz110000 g mol-l). This comparison shown in Fig.8(b)for conductivities measured at a given temperature (27 "C,300 K) indicates that crosslinking had little if any effect on the conductivity. Conductivities at 27 "C of network polymer electrolytes based on linear P400 and network P200-MS are compared in Fig. 9. The lower conductivities found for the P200 system (compared to the P400 system) are not a consequence of crosslinking but are consistent with previous results concern- ing the effect of E-sequence length on ionic conductivity in POMOE electrolytes. Finally conductivities at 27 "C interpolated from Arrhenius plots published by other worker^^.'^ for network electrolytes based on PEG400 are listed in Table4. Close comparison is not possible since the electrolytes were formed from different salts at different concentrations.However we note that the local viscosity the latter effect being associated with the increase in the glass-transition The network polymer electrolytes based on PEG400 were not examined in the same detail. Instead two compositions within the region of the maximum in conductivity were exam- ined i.e. O/Li=30 and 50. These results are included in Fig. 7(u) and within experimental error fall on the curve established for the uncrosslinked system. A more detailed comparison over the complete temperature range investigated is shown in Fig. 7(b) i.e. via Arrhenius plots of log loc(CT= conductivity) us. reciprocal temperature. The general form of ratio O/Li=25 (0)network P200-M5-N25; (0)network P200-MlO- these curves is similar to that found in previous ~ork.~*~~v~~ N25.(b) Log,,(conductivity) us. LiC104 concentration for POMOE Fig. 7 (a) Log,,(conductivity) us. LiClO concentration for POMOE electrolytes based on (0)linear P400; (A) network P400-M2-N; (V ) network P400-M5-N. (b) Arrhenius plots for POMOE/LiC104 electro- lytes with mole ratio O/Li=30 (0)linear P400-30; (A) network P400-M2-N30; (V ) network P400-M5-N30. The good agreement between the conductivities of the linear electrolytes based on (0)linear P194; (0)network P200-M5-N. J. Muter. Chem. 1996,6(7) 1099-1106 1105 I I4t i I I I I 0 1 2 3 chol kg-l Fig. 9 Log,,(conductivity) us LiCIO concentration for POMOE electrolytes based on (0)linear P400 (0)network P200-M5-N Table 4 Comparison of conductivities reported for network polymer electrolytes based on PEG400 ref Salt O/Li or O/Na 0/10 ’S cm 9 NaClO 15 <o 1 50 1 to 2 12 LiTFSI 15 06 30 1 present work LiClO 50 30 4 4 levels of conductivity measured in the present work compare favourably with those measured for related systems Conclusions Network POMOE electrolytes can be prepared by inclusion of unsaturated groups in the polymer chain and after mixing with salt crosslinking by radical chemistry The thermal and dynamic-mechanical-thermal properties of the polymers and the thermal properties and conductivities of the polymer electrolytes are unchanged by crosslinking but creep of the polymer electrolyte is eliminated Unsaturated groups are essential for thermally crosslinking the polymer but not for photochemical crosslinking However the presence of unsatu- rated groups improves the rate of photochemical crosslinking We thank Mr S K Nixon and Mr M Hart for help with the experimental work Financial support came from the British Council and from Trigon Packaging Systems (UK) Ltd J H T and R A C held EPSRC Research Studentships References 1 Polymer Electrolyte Reviews-1 and 2 ed J R MacCallum and C A Vincent Elsevier Applied Science London vol 1 1987 and vol 2,1989 2 J R Owen in Comprehensive Polymer Science ed C Booth and C Price Pergamon Oxford 1989 vol 2 ch 21 3 P G Bruce and C A Vincent J Chem Soc Faraday Trans 1993 89,3187 4 F M Gray Solid Polymer Electrolytes VCH New York 1991 5 J R Craven R H Mobbs C Booth and J R M Giles Mukromol Chem Rapid Commun ,1986,7,81 6 C Booth C V Nicholas and D J Wilson in ref 1 vol 2 ch 7 7 C V Nicholas D J Wilson C Booth and J R M Giles Br Polym J 1988,20,289 8 Y Pang S-M Mai K-Y Huang Y-Z Luo J H Thatcher R A Colley C V Nicholas and C Booth J Muter Chem ,1995,5 831 9 S E Sloop M M Lerner,T S Stephens,A L Tipton,D G Paul1 and J D Stenger-Smith J Appl Polym Sci 1994,53,1563 10 P J Flory Principles of Polymer Chemistry Cornell UP Ithaca 1953,ch 3 11 F Alloin J-Y Sanchez and M Armand Solid State Ionics 1993 60 3 12 F Alloin J-Y Sanchez and M Armand J Electrochem Soc 1994 141,1915 13 F Alloin C R Herrero J-Y Sanchez D Delabouglise and M Armand Electrochim Acta 1995,40 1907 14 S Besner A Vallee G Bouchard and J Prud’homme Macromolecules 1992,25,6480.15 J R Craven C V Nicholas R Webster D J Wilson R H Mobbs G A Morris F Heatley C Booth and J R M Giles Br Polym J 1987,19,509 16 S Nagae M Nekoomanesh C Booth and J R Owen Solid State Ionics 1992,53-56 11 18 17 F T Simon and J M Rutherford J Appl Phys ,1964,35,82 18 J P Quesnel and J E Mark in Comprehensive Polymer Science ed C Booth and C Pnce Pergamon Oxford 1989 p 297 19 ref 18 p 300 20 G N Malcolm and J S Rowlinson Trans Faraday Soc 1957 53,921 21 J H Thatcher K Thanapprapasr S Nagae S-M Mai C Booth and J R Owen J Muter Chem 1994,4,591 22 B-X Liao Y-M Chen C Booth and Y-Z Luo Polym Commun 1991,32,348 23 P G Bruce F M Gray J Shi and C A Vincent Philos Mag 1991,64,1091 24 M Nekoomanesh H S Nagae C Booth and J R Owen J Electrochem Soc 1992,139,3046 25 Polymer Laboratories MKIIDMTA Specification Polymer Laboratories Ltd Thermal Sciences Division Loughborough UK. 26 J F LeNest A Gandini and H Cheradame Br Polym J 1988 20,253 27 G C Cameron M D Ingham and G A Sorrie J Chem Soc Faraday Trans I 1987,83,3345 28 G C Cameronand M D Ingham,mref 1,vol 2,ch 5 Paper 5/07610B Received 22nd November 1995 1106 J Muter Chem 1996,6(7) 1099-1106