Solid State Ionics 147 (2002) 141 – 155 www.elsevier.com/locate/ssi
Ion-transport phenomena in concentrated PEO-based composite polymer electrolytes D. Golodnitsky*, G. Ardel, E. Peled School of Chemistry, Sackler Faculty of Exact Science, Tel Aviv University, Tel Aviv, 69978, Israel Received 30 July 2001; received in revised form 10 December 2001; accepted 12 December 2001
Abstract The enhancement of the ionic conductivity of Li – P(EO)n-based polymer electrolytes by the addition of finely divided inorganic oxides is the subject of considerable discussion. The increase in ionic conductivity in concentrated composite solid polymer electrolytes (CSPE) is related both to the suppression of the formation of crystalline PEO and PEO-salt phases and to interfacial conduction. In the case of polycrystalline or polyparticle solid electrolytes, however, grain-boundary (GB) resistance, which is associated with the crossover of ions from particle to particle across grain boundaries orthogonal to the direction of current flow (or parallel to the field lines), must be considered. The conduction across grain boundaries in polymer appears to be ignored so far. In this work, we provide SEM and ECS experimental data on LiI – P(EO) – Al2O3 CSPEs. The effects of plasticizers, doping by CaI2, change in Li/EO ratio, and nanosize alumina concentration on the interior grain ionic conductivity and on the resistance of the orthogonal grain boundaries are addressed. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Ion-transport phenomena; Composite solid polymer electrolytes; Conduction; Plasticizers
1. Introduction The great majority of studies on batteries over the last two decades has been directed toward the development of highly conductive solid polymer electrolytes because of their promising applications in all-solidstate rechargeable lithium or lithium-ion batteries. A central aspect of research on these electrolytes deals with the mechanism of ion transport. Theoretical models for the conduction mechanism have been proposed [1,2]. Despite this activity, the exact conduction mechanism in polymer electrolytes is still a matter of some controversy, but it is generally agreed *
Corresponding author. E-mail address:
[email protected] (D. Golodnitsky).
that above the melting point of the eutectic (Tm), the effective ionic conduction occurs mainly in the molten phase and that both ionic species are mobile. It has been demonstrated by several research groups [3 –16] that dispersion of small-size ceramic powders in the polymer matrix produces composite polymer electrolytes (CPE) that show consistent improvements in both interfacial and transport properties. On the basis of DSC, SEM, XRD, NMR, and conductivity data [11,12], we have suggested that the addition of 15-nm-size aluminum and magnesium oxides increases the degree of local crystal disorder of the PEO – LiI complex by preventing agglomeration of the chains. This is followed by enhanced segmental motion. Laboratory prototypes of monoand bipolar lithium/pyrite long-cycle-life batteries
0167-2738/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 2 7 3 8 ( 0 1 ) 0 1 0 3 6 - 0
142
D. Golodnitsky et al. / Solid State Ionics 147 (2002) 141–155
containing diluted and concentrated LiI – P(EO)n CPEs have been developed [13,14]. Croce et al. [9,10] explained the great conductivity enhancement of LiClO4, LiBF4, LiCF3SO3 – P(EO)8 composite electrolytes by the combined action of the
large surface area and Lewis-acid character of the ceramic additive. The electrochemical properties of mixed-phase composite electrolytes based on PEO, different lithium salts, and ferroelectric materials such as BaTiO3, PbTiO3, and LiNbO3 have been studied in
Fig. 1. SEM micrographs of PbI2/Pb electrode for the concentration cell measurements.
D. Golodnitsky et al. / Solid State Ionics 147 (2002) 141–155
in [15]. The increase in ionic conductivity of the CPEs with 8 < n < 30 was rationalized by complex interplay of the association tendency of ions and the spontaneous polarization of the ferroelectric ceramics due to their particular crystal structure. Nagasubramanian et al. [5] showed that both the transport number and ionic conductivity are influenced by the particle size of alumina. The composite solid electrolyte developed at the Jet Propulsion Laboratory exhibited the highest lithium transport number reported for polymer electrolytes. Notwithstanding the fact that the dispersion of fine ceramic particles in a polymer matrix is the subject of considerable literature, the mechanism of ion conduction in composite polymer electrolytes, to our knowledge, is not clearly understood. Of special note are the phenomena of mass and charge transport along and across internal grain boundaries. It was shown [11,16] that the conduction processes for concentrated composite solid polymer electrolytes (CSPE, n V 3) at temperatures below or close to the melting of PEO are entirely different from those of the composite polymer electrolytes (CPE) with n > 3. We have demonstrated that in LiI – P(EO)n – Al2O3 CSPEs with n < 3, the conductivity jump in the Arrhenius plots and low Ea values may reflect an interfacial conduction process between the solid LiI – P(EO)3 complex and the ceramic nanoparticles [11,16 – 19]. High-resolution NMR for these high-concentration electrolytes indicated at least two lithium environments, one solvated by PEO and one in a LiI-ionic cluster at room temperature [12]. The conductivity jump was also associated [12] with a change in Li + bonding in CSPEs that favors cation mobility. The emphasis in this work will be on the study of the conduction across grain boundaries in concentrated composite polymer electrolytes. The effects of plasticizers, doping by CaI2, change in Li/EO ratio, and nanosize alumina concentration on the ionic conductivity and grain-boundary resistance are investigated. Determination of Li transference number by an EMF concentration cell method is addressed.
2. Experimental All materials were processed and cells were built inside VAC glove boxes. High purity, vacuum-dried
143
components were used for polymer-film preparation. The LiI –PEO – Al2O3 CPEs were cast from acetonitrile solutions and vacuum dried at 120 C. A JEOL SEM was used for the study of surface topology. All investigations were performed in a 1-cm2 cell which permitted the sandwiching of a 150 –200-mm CPE film between blocking gold-coated stainless steel electrodes. To improve the interfacial contacts, the CPEs were gold-coated by vacuum deposition. The cells were held under spring pressure inside a hermetically sealed glass vessel. The AC measurements were performed with a Solartron 1255-frequency response analyzer controlled by a 586 PC. The errors in the calculation of Rb were estimated to be about 20% at near-ambient temperatures and 5% at T > 70 C. The error in the RGB determination did not exceed 5% over the whole temperature range under investigation (25 – 120 C). The data presented are normalized to 100-mm thickness. Lead/lead iodide electrodes for the measurement of transference numbers [20] were first prepared by electrolysis in a slightly acidified 0.1 M KI solution. In order to get a uniform distribution of the PbI2 film, a low current density (0.1 mA/cm2) was used. The Pb/ PbI2 electrodes were washed with distilled water and vacuum dried at 90 C for 20 h. SEM micrographs (Fig. 1) indicated the formation of uniform film made of small, about 1-mm-size, PbI2 crystals. EDS analysis showed homogeneous distribution of Pb and I elements on the electrode surface.
3. Results and discussion 3.1. Characterization of polymer electrolytes by SEM From the SEM micrographs (Fig. 2), it is seen that highly concentrated LiI – P(EO)n solid polymer electrolytes are made up of units whose areas are hundreds of square microns. In the 1:6 composite polymer electrolyte, the grains are distinguished only at ( 1000) magnification and at n = 9, the grain structure is scarcely visible. Addition of alumina to the polymer electrolytes causes minor reduction of the grain size. In 1:6 CPE, alumina smears out the cell boundaries. As can be seen from the mapping of elements (Fig. 3), both oxygen and iodine are homogeneously distributed over the CSPE surface. The atomic concentration of
144
D. Golodnitsky et al. / Solid State Ionics 147 (2002) 141–155
Fig. 2. SEM micrographs of concentrated polymer and composite polymer electrolytes.
alumina in the grain-boundary area, however, was found to be smaller than that in the bulk of the unit. Addition of EC and PEGDME causes deterioration of the homogeneous distribution of alumina in the film. These additives do not change the unit size, but cause the accumulation of Al2O3 and formation of large aggregates on the surface of the film (Fig. 4). Doping of the CSPEs by CaI2 is followed by considerable embrittlement of the films. Although there was no visible effect on the size of the crystalline units, the cells seemed to be more separated than those of the non-doped CSPE. Hot pressing of the CSPE films significantly improves interparticle contact (Fig. 5). 3.2. Conductivity data The complex impedance plot of the Li/CSPE/Li cell at 120 C is represented by a broad arc. When the temperature is lowered to 90 C, the single arc separates into partially overlapping semicircles with
fmax of 500 and 3 kHz, and a capacitance of 0.2 nF/ cm2 and 0.1 mF/cm2, respectively (Fig. 6). Based on the previous studies [16], the medium frequency arc is attributed to the solid electrolyte interphase (SEI) formed on lithium. In CPEs with n = 9, the highfrequency arc appears only below 90 C. In CSPEs with n < 3, this arc is found in the Nyquist plot even at 130 C. The resistivity of the high-frequency arc is strongly temperature-dependent and varies from 10 to 100 kV/cm at 120 and 60 C, respectively. In 1969, Baurle [21] was the first to use the modern approach to the application of impedance spectroscopy to solid polycrystalline electrolytes. He found that the presence of different phases in dense material could lead to the introduction of a second time constant in the equivalent circuit due to grain-boundary (intra-grain) phenomena. For a single-phase polycrystalline, but very nonisotropic sodium b-alumina, Hooper [22] found that intergrain (bulk) conductivity has greater activation energy and the corresponding arc in the
D. Golodnitsky et al. / Solid State Ionics 147 (2002) 141–155
145
Fig. 3. EDS mapping of elements in Li – P(EO)1.5 6% (v/v) Al2O3 composite polymer electrolyte ( 100).
Nyquist plot disappears at high temperatures. Similar data were found for more isotropically conducting materials such as LISICON [23]. We believe that in the case of polycrystalline or polyparticle solid polymer electrolytes, grain-boundary (GB) effects may be associated with two different phenomena, or more specifically, with two conduction paths through the GB: parallel (rGBII), and orthogonal (rGB?) to the current flow. The former should make itself evident in the reduction of the bulk resistance and enhancement of ionic conductivity in polymer electrolyte. The latter is represented by the build up of
the total PE resistance and the occurrence of the additional resistance component (high-frequency semicircle). It may be associated with the crossover of ions from particle to particle through grain boundaries which are orthogonal to the current flow (or parallel to the field lines). In accordance with the above, we attribute the high-frequency semicircle in the Nyquist plots of LiI –P(EO)n CSPEs to the resistance of grain boundaries (GB?), orthogonal to current flow. The boundaries between two crystalline units of LiI – P(EO)3 complex, between excess of free LiI particles and
146
D. Golodnitsky et al. / Solid State Ionics 147 (2002) 141–155
Fig. 4. SEM micrographs of composite polymer electrolytes with additives.
LiI – P(EO)3 crystallites, between alumina particles and complex or LiI, contribute to this resistance term in CSPEs with n < 3. Ceramic-free polymer electrolytes are polycrystalline as well [9,10,17,24]. The GB arc in the impedance plane of these PEs may be associated with ion crossover through randomly distributed crystalline units built of clusters of lithium salt – PEO complex and of crystalline regions, formed by sections of ordered helices of PEO chains. As a consequence, there are two paths available to the current in CSPEs: through interior of the grains (rGin), and across grain boundaries (rGB?) and/or
along grain boundaries (rGBII) as depicted in the schema shown in Fig. 7. Depending on the relative magnitudes of rGin, rGB?, and rGBII, one of the two paths may dominate. Ionic conductivity of composite polymer electrolytes was calculated from the high-frequency intercept of the GB arc with axis X of the Nyquist plot. Fig. 8 shows the effect on ionic conductivity of the Li/EO ratio in LiI – P(EO)n 6% (v/v) Al2O3 composite polymer electrolytes. Two clearly pronounced features are seen from the r vs. temperature plots. The first is that the conductivity values of both CSPEs with n < 3 and CPEs with 3 < n < 9 at near-ambient temperature vary
D. Golodnitsky et al. / Solid State Ionics 147 (2002) 141–155
Fig. 5. Effect of hot pressing on the SEM images of CaI2-doped CSPEs.
147
148
D. Golodnitsky et al. / Solid State Ionics 147 (2002) 141–155
Fig. 6. Nyquist plot of: Li/CSPE/Li cell 90 C.
only slightly (from 0.06 to 0.1 mS/cm). The highest r at T < 66 C is found for CSPE with n = 2.5. Above the melting point of the PEO (T > 66 C), r of the CPEs increases dramatically and reaches 1.5 mS/cm (for n = 9); this is one order of magnitude higher than that of the CSPE with n < 3. Another feature is that the conductivity of 1:2.5 CSPE is the least temperaturedependent. Fig. 9 shows the effect on the ionic conductivity of the alumina concentration in the polymer electrolyte with n = 2.5. It is clear that addition of 6% (v/v) alumina enhances r over the whole temperature range investigated. Further increase of the inorganic filler concentration is followed by a sharp drop in conductivity. The conductivity of the CSPE with 18% (v/v) alumina is about half that of the electrolyte without Al2O3. This is in agreement with the work of Wieczorek [25] who showed that at high concentrations of
inorganic fillers, nonconducting dispersoid regions are formed, and these lower the ionic conductivity of polymer electrolyte. A similar effect of high concentrations of nanosize alumina was found in LiI –P(EO)n CPEs with 3 < n < 9. It is noteworthy that CSPEs doped by plasticizers such as PEGDME (0.0075 M) and EC (0.05 M) have two to three times lower r values than those of nonplasticized CSPEs. This effect was observed both at RT and at T >Tm. The addition of 0.05 M CaI2 makes the CSPE more rigid and has little or no effect on r. Hot pressing of composite solid electrolytes caused almost no change in conductivity in spite of a decrease of the glass transition temperature from 15 to 0.7 C (at n = 2.5) [26]. In agreement with previously published data [11,16], Arrhenius plots for CSPEs differ from those of CPEs. While the conductivity/temperature depend-
D. Golodnitsky et al. / Solid State Ionics 147 (2002) 141–155
149
Fig. 7. Schematic presentation of highly concentrated composite solid polymer electrolytes.
ence of the CPEs is characterized by an inflection point at about 80 C [11,16], the Arrhenius plots for CSPEs show a jump in conductivity over the temperature range of 70 –90 C, that is above the melting point of the eutectic and pure PEO (Fig. 10). The apparent activation energy of conduction at T < Tk (T knee) in 1:9 and 1:6 CPEs is 100 –120 kJ/mol; and 26– 40 kJ/ mol at T >Tm [11]. These high values may indicate that
ion migration occurs through interior of the grains (intergrain conduction) and across orthogonal grain boundaries (intragrain conduction), similarly to path 1 (Fig. 7). A completely different situation is encountered in more concentrated composite solid polymer electrolytes (n = 3). The apparent activation energies of conduction before and after conductivity jump are very
Fig. 8. Effect on the ionic conductivity of EO/Li ratio in Li – P(EO)n 6% (v/v) Al2O3 composite polymer electrolyte.
Fig. 9. Effect on the ionic conductivity of alumina content (%) in Li – P(EO)2.5 polymer electrolyte.
150
D. Golodnitsky et al. / Solid State Ionics 147 (2002) 141–155
Fig. 10. Arrhenius plots of LiI – P(EO)1.5 6% (v/v) Al2O3 and LiI – P(EO)2.5 6% (v/v) Al2O3 composite polymer solid electrolytes.
conductivity jump observed at about 70 –90 C, therefore, may be attributed to the change in Li + bonding in CSPEs that facilitates the cation mobility. The confluence of low Ea and the absence of conductivity jump observed in the Arrhenius plots of the 1:2.5 CSPE, containing 12% (v/v) of nanosize alumina may be explained by the formation of composite units of filler particle covered by a thin (about 0.1 mm) interface layer occupying the entire CSPE volume. This is in agreement with the model of Wieczorek et al. [7,25], based on effective-medium theory. In 1:1.5 CSPE, however, the activation energy at T < Tj is higher than in CSPEs with n = 2.5 and n = 3. This may be due to the fact that ion transport in concentrated electrolytes with different Li/EO ratios occurs via energetically different types of GB, for instance LiI – alumina, complex-alumina, or complexLiI. This suggestion is in agreement with our previous observation that the activation energy in a LiI– Al2O3 pellet is two to three times that in composite solid polymer electrolytes. It should be mentioned that no conductivity hysteresis was observed on thermal cycling of the CSPEcomposed cells. This is in agreement with Refs. [9 – 11]. A possible explanation [10] is that once the composite electrolytes are annealed at temperatures higher than the PEO crystalline to amorphous transition, the ceramic additive, because of its large surface area, prevents PEO chain reorganization and causes the conductivity enhancement at ambient temperature. 3.3. Determination of T+
similar and do not exceed 4– 10 kJ/mol. Such low Ea values indicate that in CSPEs, the interfacial conduction path along parallel grain boundaries (path 2, Fig. 7) is preferred over the conduction path through the interior of the grain. In Refs. [12,18], it was shown that CSPEs with n < 3 are completely solid up to 250 C. There was no evidence, from XRD and 127I NMR [12] measurements, for the presence of crystalline LiI even in a very concentrated LiI1.5 PEO 6% Al2O3 sample. As mentioned above, high-resolution NMR measurements on these electrolytes indicated at least two lithium environments, one solvated by PEO and one in a LiI-ionic cluster at room temperature. At 80 C, the lithium-polymer association vanishes and the lithium environment becomes more purely ionic [12]. The
The transference number T + is defined as the net number of faradays of charge carried across the reference plane by the cation constituent in the direction of the cathode during the passage of one faraday of charge across the plane [20,27]. In this work, T + in CSPE was calculated from the EMF of the concentration cells. Four types of CSPE-concentration cells of the following composition were investigated [28]: 1. 2. 3. 4.
Li/CSPE(1)/PbI2/Pb/PbI2/CSPE(2)/Li, Li/CSPE(1)/CSPE(2)/Li, Pb/PbI2/CSPE(1)/Li/CSPE(2)/PbI2/Pb, Pb/PbI2/CSPE(1)CSPE(2)/PbI2/Pb.
D. Golodnitsky et al. / Solid State Ionics 147 (2002) 141–155
CSPE compositions: CSPE(1): LiI – P(EO)2.5 – Al2O3, CSPE(2): LiI – P(EO)3.5 – Al2O3. The overall voltage change in cells #1 and #3 is simply the difference between the changes in the two separate cells, i.e., these concentration cells are without ion transference, whereas the concentration cells #2 and #4 involve ion transfer. The electrodes are consequently reversible to Li + and I . The major advantage of this method is that there is no need to know the activity coefficients of Li + or I . The cells were first equilibrated for 3 h at 90 C prior to EMF measurements. The experimental data of EMF measurements and calculated transference numbers of the four electrochemical cells are presented in Table 1. The calculations carried out according to the Eqs. (1) – (5) give us the experimental values of the lithium transference number 1.25. TI1 was found to be 0.04. E1 ¼ E3 ¼ RT =Flna1 =a2
ð1Þ
E2 ¼ T RT =Flna1 =a2
ð2Þ
E4 ¼ Tþ RT =Flna1 =a2
ð3Þ
Tþ ¼ E4 =E3
ð4Þ
T ¼ E2 =E1
ð5Þ
A value of T + close to unity indicates that almost all of the charge in the LiI –PEO – Al2O3 composite electrolyte is carried by lithium cations. Improvement
Table 1 EMF of CSPE-cells Time, h Cell 1
Cell 2
TI
EMF, mV EMF, mV 4 17 21 44
185 180 176 150
9 8 7.6 6.4
Cell 3
Cell 4
TLi +
EMF, mV EMF, mV 0.048 0.044 0.043 0.043
112 102.6 101.6 98
140 129 127 110
1.25 1.25 1.25 1.13
151
of the lithium transference number of polymer electrolytes over the 45 – 90 C temperature range by the addition of nanosize ceramic particles has been recently reported [5,9,10,16,19]. We believe that the high (>1) lithium transference number may be associated with the effect of ion –ion interactions on the macroscopic transport parameters, i.e., the positive charge transport of both lithium cations and positively charged complex aggregates. In addition, consideration must be given to a possible experimental error of the T + determination that may be in the range of 10– 20%. The only explanation proposed at present for the negative value of iodide-anion transference number is also based on the formation of aggregates. According to Ma et al. [29], when mobile positive triplet ions exist and positive species carry iodide anions and in the opposite direction to that of free I , the transference number T may be negative. In the lowconcentrated LiI –P(EO) composite polymer electrolytes, the transference number varies from 0.2 to 0.5 [18]. This supports our suggestion about different conduction mechanisms in highly concentrated and diluted polymer electrolytes. 3.4. Ion transport across orthogonal grain boundaries (GB? ) As was mentioned above, the total CSE resistance depends critically on the resistance of the orthogonally oriented grain boundaries. We have found that RGB? generally decreases with n and increases sharply as the temperature is lowered for all CSPEs (Fig. 11). In the CSPE with n = 1.5, the RGB? vs. temperature dependence is the least pronounced. We believe that this may be due to a large contribution of LiI/Al2O3 grain boundaries as compared to that of complex/ alumina, PEO/alumina, and PEO/complex grain boundaries. CGB? increases with n over the whole temperature range. The Arrhenius plot of conduction across orthogonal grain boundaries in the CSPEs with n = 3 shows that there is no hysteresis of conduction measured during cooling and heating cycles. For n = 1.5, however, about one order of magnitude difference of rGB? was observed. The apparent activation energy of conductance through orthogonal grain boundaries increases with n from 70.6 to 172 kJ/mol (Table 2). A possible explanation of two opposite phenomena of RGB? decrease and Ea increase with n is as follows.
152
D. Golodnitsky et al. / Solid State Ionics 147 (2002) 141–155
Fig. 11. Effect on the normalized resistance of orthogonal grain boundaries of EO/Li ratio in Li – P(EO)n 6% (v/v) Al2O3 composite polymer electrolyte.
The number of composite units (nanosize alumina particles coated by Li salt – PEO complex) increases with lithium iodide concentration in LiI – P(EO)n – Al2O3 electrolytes. Additional orthogonal grain boundaries are created and cause the RGB? increase. The reduction of apparent activation energy attendant on this process allows us to conclude that ion transport via these composite units is preferred over that through clusters of complex and PEO. In addition, it should be mentioned that normalized RGB? of LiI – Al2O3 pellets (up to 20% alumina) [30] and the apparent activation energy of conductance are similar to those of 1:1.5 CSPE. Addition of Al2O3 up to 18% (v/v) and doping by CaI2 (0.05 M) (Figs. 12 and 13) were found to increase RGB and to decrease CGB. This is a result of poorer interparticle contact and a decrease in contact area (H)
Fig. 12. Effect on the normalized resistance of orthogonal grain boundaries of alumina content in Li – P(EO)2.5 polymer electrolyte at temperature, C: (1) 60; (2) 70; (3) 90; (4) 120.
in agreement with the SEM micrographs. The Ea of ion conductance across orthogonal grain boundaries decreases from 146 to 106 kJ/mol with the addition of 6% Al2O3 to pure LiI – P(EO)2.5 polymer electrolyte and remains almost constant with further addition of alumina. Hot pressing and doping of CSPE were expected to improve the contact between crystalline units. Actually, hot pressing causes a fourfold drop in
Table 2 Effect of Li/EO ratio on the estimated apparent activation energy of conductance across orthogonal grain boundaries in CPEs and CPSEs Li/EO ratio
Ea, kJ/mol
1.5 2.5 3 6 9
70.6 106.6 117.9 172.2 172.2
Fig. 13. Effect on the normalized resistance of orthogonal grain boundaries in Li – P(EO)2.5 6% (v/v) Al2O3 composite solid polymer electrolyte (1) doped by CaI2 (2).
D. Golodnitsky et al. / Solid State Ionics 147 (2002) 141–155
Fig. 14. Effect of plasticizers on the normalized resistance of orthogonal grain boundaries in Li – P(EO)2.5 6% (v/v) Al2O3 composite polymer electrolyte: (1) without additives; (2) PEGDME; (3) EC.
RGB?. This effect is even more pronounced below 90 C. It was also found that RGB? in CSPEs at T < 90 C can be lowered by a factor of 3 – 10 by EC and PEGDME doping of a CSPE (Fig. 14). The apparent activation energy of conductance through orthogonally oriented grain boundaries for LiI – P(EO)n < 3 6% Al2O3 CSPEs was found to be less dependent than the RGB? value on doping and pretreatment of the CSPE. The Ea in CSPEs with n < 3 was in the range of 70– 106 kJ/mol, about one order of magnitude higher than that for interfacial conduction along paralleloriented grain boundaries. The bricklayer model [31 – 33], which treats the microstructure as an array of cube-shaped grains separated by flat grain boundaries, and its modification which considers sloping grain boundaries [32,33],
153
successfully explains the mechanism of ionic conduction in a polycrystalline solid system. The proposed schematic equivalent circuit for composite polymer electrolytes, shown in Fig. 15, divides the ionic current into three paths, one of which (across GB?) is blocked capacitatively, while the others are not. When the current is mainly carried along grain boundaries as in CSPEs with n < 3, the high ionic conductivity and low apparent activation energy are observed over the entire temperature range. The current passing along grain boundaries makes a ‘‘detour’’ that follows the ‘‘easier’’ path for increasing the rate of ion migration. From the schematic equivalent circuit, it is obvious that the larger the number of composite units, the less ion migration rate is affected by the resistance of the grains. Therefore, the finer the dispersed inorganic filler, the higher the conductivity of the CSPEs in agreement with previous publications [5,7,11,12]. In the CPEs with 3 < n < 9 at T < Tk, the current is flowing mainly through the bulk of grain and across orthogonal grain boundaries. This is reflected in the high Ea values of conduction. The indirect information about the topology of the grain-boundary phases can be obtained from the analysis of the Arrhenius plots. As was found in CSPEs with n V 3, the slopes of Arrhenius plots of conduction along parallel GBs differ essentially from those across orthogonal GBs. This indicates (in accordance with the bricklayer model) that in spite of easy interfacial paths coinciding with the current flow, there is a significant blocking effect of orthogonal grain boundaries impeding ionic transport. As predicted from the bricklayer model, the tetragonal structure is more appropriate for such materials. By contrast, the estimated values of apparent activation energy of conduction through orthogonal GB and via interior of the grain for CPE with 3 < n < 9 are close (170 and 120 kJ/mol) at T < Tk, thus
Fig. 15. Equivalent circuit of a composite solid polymer electrolyte.
154
D. Golodnitsky et al. / Solid State Ionics 147 (2002) 141–155
suggesting mainly cubic GB structure with discrete grain-boundary phase. The conductance through orthogonal grain boundaries was found to be one to three orders of magnitude smaller than that of the bulk of the grain and along parallel grain boundaries in all electrolytes at low temperatures. At about 60 C, rGB? and rGin become comparable for CPE with n = 6, but rGB? is still much lower for concentrated electrolytes. This study, unfortunately, leads us to conclude that the highly concentrated composite solid polymer electrolytes are useless for RT lithium batteries, if the resistance of orthogonal GBs is not low enough and whatever conductivity the CPE has. This conclusion does not necessary imply a failure of the model proposed in Ref. [10]. It does however cast serious doubts on the view that ionic conduction plays a prime role in the transport properties of highly concentrated composite polymer electrolytes at near ambient temperatures.
[3] [4] [5] [6] [7] [8] [9] [10] [11] [12]
[13] [14]
4. Summary
[15] [16]
Addition of 6% (v/v) nanosize Al2O3 to a polymer electrolyte increases the ionic conductivity to a maximum of 1 mS/cm at near-ambient temperatures. Different activation energies were found for the inter- and intragrain conductances for both CPEs and CSPEs. This suggests that essentially different physical processes are involved in the ion transport in these electrolytes. The interior and parallel grain-boundary resistance in CPEs and CSPEs are systematically (one to three orders of magnitude) smaller than orthogonal, showing that it is undoubtedly the conductivity path across orthogonally oriented grain boundaries that impede the total ion transport in polymer electrolytes at near-ambient temperature. The values of RGB? at T < 90 C can be decreased by a factor of 3 – 10 by minor amounts of EC and PEGDME plasticizers and hot pressing of the CSPE.
References [1] P.G. Bruce (Ed.), Solid State Electrochemistry, Cambridge Univ. Press, Great Britain, 1995. [2] M.A. Ratner, in: J.R. MacCallum, C.A. Vincent (Eds.), Poly-
[17] [18] [19]
[20] [21] [22] [23] [24] [25]
[26]
[27]
[28]
mer Electrolyte Reviews—1, Elsevier, London, 1987, pp. 173 – 236. F. Capuano, F. Croce, B. Srosati, J. Electrochem. Soc. 138 (7) (1991) 1918 – 1922. B. Kumar, L.G. Scanlon, J. Power Sources 52 (1994) 261 – 268. G. Nagasubramanian, A.I. Attia, G. Halpert, E. Peled, Solid State Ionics 67 (1993) 51 – 56. F. Croce, S. Panero, S. Passerim, B. Scrosati, Electrochim. Acta 39 (2) (1994) 255 – 263. W.W. Wieczorek, K. Such, Z. Florzanczyk, J. Stevens, J. Phys. Chem. 98 (1994) 6840. H.C. Angell, C. Lin, E. Sanchez, Nature 362 (1993) 137. F. Croce, G.B. Appetecchi, L. Persi, B. Srosati, Nature 394 (7) (1998) 456 – 458. G.B. Appetecchi, F. Croce, L. Persi, F. Ronci, B. Srosati, Electrochim. Acta 45 (2000) 1481 – 1490. D. Golodnitsky, G. Ardel, Yu. Rosenberg, Y. Lareah, E. Peled, J. Electrochem. Soc. 144 (10) (1997) 3484 – 3491. Y. Dai, Y. Wang, S. Bajue, S.G. Greenbaum, D. Golodnitsky, G. Ardel, E. Strauss, E. Peled, Electrochim. Acta 43 (10 – 11) (1998) 1557 – 1563. E. Peled, D. Golodnitsky, G. Ardel, J. Lang, Y. Lavi, J. Power Sources 54 (1995) 495. D. Golodnitsky, E. Peled, Electrochim. Acta 1 – 2 (45) (1999) 335 – 351. H.Y. Sun, Y. Takeda, N. Imanishi, O. Yamamoto, H.-J. Sohn, J. Electrochem. Soc. 147 (7) (2000) 2462 – 2467. G. Ardel, D. Golodnitsky, E. Strauss, E. Peled, Proceedings of the 37th Power Sources Conference (ARL), Institute of Electrical and Electronic Engineers, NY, USA, 1996, pp. 283 – 286. F.M. Gray, Solid Polymer Electrolytes, VCA, UK, 1991. E. Strauss, D. Golodnitsky, G. Ardel, E. Peled, Electrochim. Acta 43 (1998) 1315 – 1321. E. Peled, D. Golodnitsky, R. Cohen, C. Menachem, Y. Lavi, Israel Patent Application Number 105341 8/4/93. US Pat. Appl. No. 08 224, 470, 7/4/94. J. Evans, C.A. Vincent, P.G. Bruce, Polymer 28 (1987) 2324 – 2328. J.E. Baurle, J. Phys. Chem. Solids 30 (1969) 2657 – 2670. A. Hooper, J. Phys. D: Appl. Phys. 10 (1977) 1487 – 1497. P.G. Bruce, A.R. West, J. Electrochem. Soc. 130 (1983) 662 – 669. D. Golodnitsky, E. Peled, Electrochim. Acta 45 (2000) 1431 – 1436. W. Wieczorek, Composite Polyether-Based Electrolytes, Publishing House of the Warsaw University of Technology, Warsaw, Poland, 1995. G. Ardel, D. Golodnitsky, E. Strauss, E. Peled, The Joint Meeting of Electrochemical Society, Proceedings. Paris, France, The Electrochemical Society Inc. & International Society of Electrochemistry, USA, 1997, pp. 161 – 162. M. Spiro, in: A. Weissberger, B.W. Rossiter (Eds.), Techniques of Chemistry, Vol. I, Part IIA, Wiley, New York, 1970, pp. 23 – 45. Lithium Single-Ion-Conductor Film, USA AF Contract SPC96-4020, Final Report, 1997.
D. Golodnitsky et al. / Solid State Ionics 147 (2002) 141–155 [29] Y. Ma, M. Doyle, T.F. Fuller, M.M. Doeff, L.C. De Jonghe, J. Newman, J. Electrochem. Soc. 142 (1995) 6. [30] Y. Dai, S. Greenbaum, D. Golodnitsky, G. Ardel, Yu. Rosenberg, E. Peled, Solid State Ionics 106 (1998) 25. [31] N.M. Beekmans, L. Heyne, Electrochim. Acta 21 (1976) 303 – 310.
155
[32] J. Fleig, J. Maier, J. Electrochem. Soc. 145 (6) (1998) 2081 – 2089. [33] J.R. Macdonald (Ed.), Impedance Spectroscopy. Emphasizing Solid Materials and Systems, John Willey & Sons, Inc., USA, 1987, p. 197.
