J. MATER. CHEM., 1994, 4(5), 713-717 Electret Behaviour of Di= and Tri-nuclear Iron Hydrazone- Hexacyanoferrate Compounds studied by the Thermally Stimulated Depolarization Current Technique Alex Bonardi: Rosanna Capelletti,*' Corrado Pelizzi*" and Pieralberto Tarasconia a lstituto di Chimica Generale ed Inorganica, Universita di Parma, Wale delle Scienze, 43100 Parma, Italy 'Dipartimento di Fisica, Universita di Parma, Wale delle Scienze, 43700 Parma, Italy A series of di- and tri-nuclear iron hydrazone-hexacyanoferrate complexes have been prepared and characterized by vibrational and electronic spectroscopies, thermal analysis and by the thermally stimulated depolarisation current (TSDC) technique. The TSDC spectra, measured in the range 110-300 K, have confirmed the reorientation of (1) dipoles in the terminal aromatic radicals, belonging to the molecule and (2) water molecules, whenever they are present in the compound or are adsorbed on it.A fraction of water molecules is loosely bound to the compound and can be removed by pumping on the sample; the related dehydration and hydration kinetics have been analysed. Binuclear and polynuclear metal complexes containing similar or different metal ions are of interest for practical purposes in different fields, i.e. in homogeneous catalysis' and biochem- istry,,,, and for ele~trical~,~ and magnetochemica16 studies. In particular, the use of polydentate Schiff as ligands to form binuclear metal complexes has made it possible to bring pairs of metal atoms into close proximity, thus favouring the production of peculiar magnetic, optical or electric properties.Our research work on the synthesis and characterization of homo- and hetero-nuclear metal complexes with polyfunc- tional nitrogen ligands,lO,ll with the aim of isolating new electronic and magnetic materials, has been recently devoted to the investigation of a series of di- and tri-nuclear iron hydrazone-hexacyanoferrate complexes.12 As an extension of this research program we now report the characterization of this series of complexes by means of infrared and ultraviolet absorption spectroscopy, thermal analysis and a dielectric technique (TSDC).13-15 This technique has been employed previously to study a variety of electric polarization phen- omena in insulating materials such as ionic crystals, polymers and biopolymers, where the electronic conductivity is negleg- ible.The dielectric processes which have been analysed pre- viously are reorientation of ionic dipoles, the space charge and the interfacial polarization phenomena originating in inhomogeneous dielectrics. In this case the technique is used to study polarization phenomena induced by reorientation of groups in the presence of an electric dipole moment belonging to the molecule itself (e.g. see Fig. 3, later) or present in the sample as a consequence of the synthesis process (from water or methanol molecules). In this way it is possible to monitor the changes in the dipole concentration and/or environment as a consequence of specific treatments.In fact, the reorien- tation of water molecules, as detected by TSDC peaks, has already been studied in a variety of materials such as ice,16 inorganic hydrated compounds such as La( S04),17 (where TSDC peaks shift as a consequence of deuteriation) and biopolymers (melanin,ls lysozyme,19 casein,,' homopeptides2' and keratin2,). Application of the TSDC technique to biopoly- mers enabled the identification of different water sites and the analysis of the dehydration kinetic^.^^^^^ The reorientation of water molecules adsorbed on solid surfaces, for instance Si0223 or the hygroscopic NaI surface,15 has also been detected by this technique. Experimentalt [Fe( H,daps)Cl,].H,O (l),[Fe( H2dappc)C1,]-2H20 I 2), and [Fe( H2dapt)C12]-2H20 (3) were synthesized and characterized as previously de~cribed.2~ The complexes K, [Fe(daps)] [Fe( CN),12.2MeOH (4), [Fe( H,dappc)], [Fe( CN),] C1-4H20 (5)and K[Fe(H,dapt)] [Fe(CN),].4MeOH (6)were prepared by combining separately methanolic solutions of the above reported compounds (1.0 mmol in 80 ml of MeOH 99.9%) with an aqueous solution of K,[Fe(cN),] (2.0 mmol in 80 ml of water).$ The solutions were stirred at room temperature for ca.30min. After the solutions had been filtered, the resultant green microcrystalline products were washed with water and methanol and dried in vacuum. In a similar way, when 1 was treated with K,[Fe(CN),] in a 2: 1 molar ratio, a fine green precipitate of K,[Fe(dap~)]~[Fe(CN),].2W,0 (7) was collected by filtration.The yields of the purified compounds were >80%. The IR absorption measurements were performed at room temperature using a Perkin-Elmer 283B and a Jasco 702 G recording spectrophotometer, both operating in the 4000-200 cm-' range. In order to monitor the absorption changes induced by the sample dehydration process, the pellets were assembled in a cell (with KRS5 windows) which could be evacuated. The UV absorption spectra were meas- ured at room temperature using a Jasco 505 recording spectro- photometer. The thermal analysis was performed using a Perkin-Elmer Delta series TGA7 thermobalance operating in the range 30-400 "C, usually at a heatinglcooling rate of 20 "C min-l. Elemental analyses were carried out using a Perkin-Elmer 240 and a Philips PU7450 ICP atomic emission spectrometer. The TSDC measurements were performed using a home- made apparatus, typically as follows.The sample was assembled in a gas-exchange cryostat, placed between two 7 H,daps =2,6-diacetylpyridine bis(salicyloy1h ydrazone), H2dappc = 2,6-diacetylpyridine bis(2-pyridinecarbonylhydrazone), H,dapt = 2,6-diacetylpyridine bis( 2-thenoylhydrazone). $Elemental analyses (C, H, N, S, and Fe) are in agreement with the proposed formulations. Anal. Calc. (found 1 for K6Fe3C37H27N1,06(4): C, 36.79 (36.50%); H, 2.25 (2.36%); N, 19.72 (20.17%); Fe, 13.87 (14.55%). Anal. Calc. (found) for Fe,C48H,6N20C10, (5):C, 46.72 (47.15%); H, 3.76 (4.15%); x, 22.70 (22.68%); Fe, 13.58 (13.65%).Anal. Calc. (found) for KFeZC29H33N1106S2(6):C, 41.15 (40.92%); H, 3.93 (4.12%); N, 18.20 (18.35%); S, 7.56 (8.20%) Fe, 13.19 (13.50%). Anal. Calc. (found) for K3Fe3C52H42N16010(7): C, 46.75 (47.05%); H, 3.17 (3.33%); N, 16.78 (16.42%); Fe, 12.54 (13.01%). J. MATER. CHEM., 1994, VOL. 4 metal electrodes and polarized by an electric field E, (ca. lo3V cm-l), at a temperature Tp (usually ca. 300 K) at which the electric dipoles are mobile. The sample was then cooled to a temperature (usually 110 K) at which the dipoles are no longer mobile; the electric field was then turned off. The polarization induced by the field is frozen in because the dipoles remain aligned along their preferred orientations.The sample was then connected to an electrometer, then warmed at a constant rate (ca. 0.1 K s-'). During this step the dipoles gain mobility and lose their preferred orientations, giving rise to a displacement depolarization current which is detected by the electrometer (Cary 401), that is able to monitor currents as low as 10-15A. In this way a plot is obtained of the thermally stimulated depolarization current (TSDC) us. temperature, which is displayed by a recorder (Leeds & Northrup Speedomax XL682 recorder). The elec- trometer signal and the voltage supplied by the chromel-con- stantan thermocouple (which monitors the sample temperature) are acquired by means of a Labmaster interface and transferred to an IBM personal computer for data processing.In the present case the TSDC spectra were moni- tored in the temperature range 110-300 K. For further details of the apparatus used for TSDC measurements see ref. 15. Blocking electrodes were obtained by inserting A-type samples (see below) in a PTFE box which was then placed between the electrodes. In order to avoid changes in the hydration level, the TSDC measurements were never performed in dynamic vacuum, but under a dry nitrogen atmosphere at a pressure of 100Torr. Between subsequent TSDC runs the nitrogen pressure was increased to 600 Torr. A clean dynamic vacuum (2 x Torr) for dehydrating the samples was obtained by using a diffusion pump followed by a trap cooled to liquid-nitrogen temperature. Different kinds of samples were used A-type samples, i.e.pure-compound pellets (4= 13mm, x~0.3mm); B-type samples, i.e. compound (ca. 1mg) dispersed in KBr (100 mg) pellets and C-type samples, i.e. solution of the synthesized compound in DMSO (lo-' mol 1-l). A-type samples were used for TSDC measurements; B-type samples were used for IR measurements, since they were transparent enough to allow detection of the absorption spectrum in the IR range; C-type samples were used for UV spectroscopy. Results and Discussion The IR spectra of 1, 2 and 3 and of the parent cyano- derivatives are very similar, suggesting that there is no substan- tial change in the coordination of the hydrazone ligand towards the iron@) atoms caused by the substitution reaction.The v(CN) bands in 4-7 are in the range 2060-2040 cm-l, in agreement with bands reported for terminal and non-linearly bridging CN gro~ps.'~-~~ The IR spectra of 5 and 6 exhibit v(NH) bands comparable with those of 2 and 3, respectively, while in the spectra of 4 and 7 these bands are not present, in agreement with the doubly deprotonated nature of the hydrazone ligands. The complexes, dissolved in DMSO (lo+ mol l-'), show UV-VIS peaks at 315 <A/nm <330 with shoulders in the range 350<A/nm<460. It is difficult to assign these bands unequivocally because transitions of different origin, such as iron d-d transitions, charge transfer (ct) from the ligand (CN-, H,daps, H'dappc or H,dapt) to the metal, and ct between the metals, are expected to occur in this range.Free K3[Fe(CN)J in DMSO has bands at 424 and 320 nm due to M+L(CN-) ct and d-d transitions of the iron(m) crystal- field, respectively. By analogy with these findings, the peak and the first shoulder present in the spectra of the cyano- complexes are tentatively assigned to cyanoferrate absorption. The thermogravimetric analysis (TG) of the cyano-complexes 4-7 shows a first weight loss in the range 50-170 "C which corresponds to two MeOH molecules for 4, four water and one HC1 molecule for 5 (see Fig. 1) one MeOH molecule for 6, and two water molecules for 7; a second weight loss, in the range 180-330 "C, corresponds to three, four and three HCN molecules for 4, 7 and 5, respectively, three MeOH and one HCN molecule for 6.Typical TSDC spectra of the compounds investigated in the present paper are shown in Fig. 2. They are related to pellets (A-type samples) polarized in the range Tp=290 K and T,= 110K (T, =polarization temperature, at which dipoles can be easily oriented by the field; T,=temperature at which the field is switched off, since the polarization is frozen in). Fig. 2(a) concerns 5, while Fig. 2(b) concerns 6:12 the former shows a broad band peaking at ca. 192 K and a rise towards the high-temperature side (T>250 K), while the latter shows a huge band at 210 K and a shoulder at ca. 150 K. Note that the Fig. 2(b) does not exhibit any rise on the high-temperature side. These results support those reported in ref.12 for a different sample of the same compound [described by Fig. 2(a) of the present work] and for 4. The TSDC plot of the hydrated compound 5 shows a peak at ca. 192 K and a current rise in 'O0I90 8ol 111, , , , , , , , , , Illlj-//lllltt50 0 100 200 300 400 T/"C Fig. 1 TG of 5: (a) weight loss 0s. temperature; (b)first derivative of weight loss us. temperature NI5 0.15 a I 0.10 : 125 2250.00La-150 175 200 LL-LU-d 250 TIK Fig. 2 TSDC plot for a pellet of (a) 5 and (b) 6 with Tp=290 K, T,= 110 K and V,= 100V J. MATER. CHEM., 1994, VOL. 4 the high-temperature region (above 230 K); such a rise is the beginning of the band peaking at ca. 290 K, shown in ref. 3 and not shown in the present figure.The compound, including MeOH molecules, does not show such a rise, but only a broad band at 210K (peaking at ca.220K) in agreement with that of 4 reported in ref. 12. When blocking electrodes were used, the spectra were qualitatively the same, even if their amplitude is reduced. This result rules out charge-carrier injection being responsible for the observed spectra. Previously12the 290 K band was attributed to the reorien-tation of water molecules embedded in the crystal structure (see formula of compound 5): in Fig. 2 the presence of the displacement current rise on the high-temperature side of Fig. 2(a) (compound 5) and its absence in the spectrum in Fig. 2(b) (compound 6)support this interpretation. The peaks at lower temperatures exhibited by both curves are attributed to the reorientation of dipoles associated to the terminal aromatic radicals2*(see Fig.3): for the compounds described in Fig. 2 they are picolinic and salicylic groups. Such low-temperature relaxations are absent in compounds in which the terminal aromatic groups do not possess a dipolar moment.12 Moreover, the intensity of the low-temperature peak has a linear dependence on the strength of 043 OH H2hPPC H2daps H2dapt P = 2.22D 1.60 D 0.54D Fig. 3 Bis(acy1hydrazone)s of 2,6-diacetylpyridine ligands and the related dipole moments' (expressed in D; 1D x3.335 64 x lop3'C m) 715 1.5 cu5 1.0 a 70 :F 0.5 0.0 ' ' ' ' 1 " 1 ' I '1 ' 1 I LI 100 150 200 250 300 77K Fig.5 Effect of the dehydration process induced on the TSDC plots by the exposure of a pellet of 5 to a dynamic vacuum at room temperature for increasing times t,: (a)0 min, (b) 15 min, (c) 30 min, (d)60 min, (e) 120min, (f) 225 min, fg)1240min. For all curves T,, Tf, and V, are the same as in Fig. 1. the applied electric field, as expected if the process responsible for the TSDC peak is the reorientation of non-interacting dip01es.l~The results of this analysis are shown in Fig. 4 for 5: the 180 K peak? increases with increasing applied voltage and the peak amplitude scales linearly with the strength of the applied electric field, at least up to ca. 2 x lo4V cm-' (insert to Fig. 4). Note that the TSDC peaks in Fig. 4 and Fig. 2 (a)are broad (a shoulder is also present at cu.215 K). This means that the peak is caused by dipoles that have a distribution of relaxation times'* rather than a unique relaxation time z. To obtain further support for the hypothesis that reorien-tation of water molecules is responsible for the rise on the high-temperature side of the spectrum, dehydration and rehy-dration experiments were performed at room temperature and the effects on the TSDC plots were monitored. Fig. 5 shows the results of the dehydration process, obtained by exposing a pellet of 5 at room temperature to dynamic vacuum for increasing times. The rise on the high-temperature side of the plot is progressively reduced until it is completely suppressed for long times of sample exposure, supporting its attribution to the reorientation of water molecules.The release of water molecules is confirmed also by the simultaneous decrease of the typical IR absorption bands, associated with the stretching modes of the water moleculesi2and by TG results (see above).v).= 1 In the present case water molecule desorption takes place h l::y,,j C3 even at room temperature (ca. 300 K), i.e. at temperatures 0 lower than that at which the first minimum in the TG 0 50 100 150 200 500 c. c 0" 250 g I,,, I, 0 50 100 150 (a) 200 250 77K Fig. 4 Dependence of the low-temperature TSDC peak on the applied electric field for a pellet of 5 ca. 0.1 mm thick kept for ca. 2 h in a dynamic vacuum at room temperature.The TSDC plots were all obtained by polarizing in the range 290-110 K (as for Fig. 1) at different voltages V,: (a) 5 V, (b) 10 V, (c) 25 V, (d) 50 V, (e) 100 V, 200V. The insert shows the amplitude of the peak 11s. the applied voltage. derivative plot occurs [Fig. l(b)], and the desorption is induced by the dynamic pumping on the sample. Note, moreover, that this release of water molecules occurs in a time that is much longer than that required for the TG scan. By increasing the time of exposure of the sample to dynamic vacuum, the low-temperature band, peaking at 192 K in the freshly prepared sample [Fig. 2(u)], shifts gradually towards low temperatures down to a saturation value [Fig. 6(u)J. The changes induced by the dehydration process at room tempera-ture on the TSDC spectra displayed in Fig. 5, are summarized in Fig.6(b)and (c), where the current density, taken at three characteristic temperatures Ti, T2 and & in the TSDC plot of Fig. 5, is plotted us. the time of exposure of the sample to the t The shift in the peak from 192 K (for the freshly prepared sample, see curve 1 of Fig. 2) to 180 K (see Fig. 4) is accounted for by having stored the sample for ca. 2 h in a dynamic vacuum (see below). J. MATER. CHEM., 1994, VOL. 4 o-201saturation value 0.15 160 0 50 100 150 200 250 0 50 100 150 200 250 annealing time in vacuum/min Fig.6 Time evolution of the TSDC plots along the dehydration process displayed by Fig. 4 for a pellet of 5 exposed to a dynamic vacuum at room temperature.(a) Peak position TM0s. t,; (b)current density j(&) measured at & (indicated by an arrow in Fig.4); (c) current densities j(T,) and j(T,) measured at TIand & (indicated by arrows in Fig. 4). dynamic vacuum. Fig. 6(b) shows the fast decay of j(q), related to the high-temperature side of the TSDC plot; note that the ordinate at the origin is as high as ca. 5000 and is not displayed in Fig. 6(b)to avoid compression of the curve. Fig. 6(c) shows how the decrease in the current at is accompanied by an increase in the current at q.This can again be ascribed to the role played by the water molecules: their release during the sample exposure to dynamic vacuum causes a partial reorganization of the crystal structure around the dipoles associated with the terminal aromatic radicals (see Fig.3), the reorientation of which is responsible for the low- temperature peak in Fig. 1 and 3. This rearrangement might induce a change of the dipole reorientation parameters (acti- vation energy E, and pre-exponential factor zo) in the dipole relaxation time z, given by z=zo exp(EJJk,T) (1) and therefore a change of the temperature TM at which the peak occurs. For non-interacting dipoles TMis given by where /3 is the heating rate used during the TSDC recording process.15 In the present case the dipole moments related to the terminal aromatic radicals are reoriented more easily when water is removed, since the low-temperature TSDC peak shifts to still lower temperatures.Similar behaviour was observed in polymers of biological interest, such as rnelanin,l8 lysozyme," casein2' and homopeptides,21 and in dinuclear copper complexes.29 The shift of the 192 K peak observed during the dehydration process at room temperature is reversed if the sample is rehydrated at room temperature, i.e. it is exposed to the ambient moist atmosphere. Fig. 7 shows the changes induced in the low-temperature peak shape with increasing time of exposure of the sample to the atmosphere and Fig. 8 shows the time evolution of the peak at position TM.This process is accompanied again by the fast growth on the high-temperature 125 150 175 200 225 250 TIK Fig. 7 Effect of the rehydration process induced on the TSDC plots by the exposure to moist atmosphere at room temperature of a pellet of 5, previously dehydrated (see Fig.4) for different exposure times ta: (a) 22 h 50 min, (b) 45 h 50 min, (c)265 h 50 min, (d) 331 h 50 min, (e) 718 h 50 min, (f) 2275 h 50 min. For all curves Tp, Tf and V, are the same as in Fig. 1. t 0 2000 4000 6000 tlh Fig.8 Time evolution of the TSDC peak position TM along the rehydration process at room temperature displayed by Fig. 7 for a pellet of 5 exposed to a moist atmosphere side of the TSDC spectrum and the recovery of the IR absorption spectra related to the H20 stretching modes,12 which provides further support for the observed changes being attributed to the rehydration process. However, by comparing Fig.5(a)(for a freshly prepared sample) with Fig. 70 (for the same sample submitted to an initial dehydration process for a long time and then exposed to the moist atmosphere for a very long time), it turns out that the recovery is not complete. Moreover, Fig. 8 shows that the peak does not shift back to the original TM=192 K, notwithstanding that the sample has been exposed to the moist ambient atmosphere for an extremely long time (ca. 6000 h). This suggests that the removal of molecules induced by exposure of the sample to a dynamic vacuum has caused, even to a limited extent, some irreversible change in the crystal structure. Irreversible changes are indeed expected due to a partial release of HCl molecules (suggested by the TG results; see above and Fig.1)induced by pumping. This behaviour is similar to that reported for Ni( H,dapa)13C1.H20,12 where iodine is released as a conse- quence of pumping on the sample. However, when HC1 and iodine are removed from the crystalline structure they cannot J. MATER. CHEM., 1994, VOL. 4 be replaced by exposing the sample to the ambient moist atmosphere, as the water molecules can. Note that TSDC measurements performed on the second hydrated complex (7),investigated in the present work, give quite similar results, confirming the above interpretation on the role played by the water molecules. In this case, however, the removal of water molecules induced by dynamic pumping of the sample at room temperature causes a shift in the low- temperature TSDC peak towards higher temperature.This means that the reorientation of the dipoles associated with the terminal aromatic radical becomes less easy when reor- ganization of the crystalline structure around the dipole occurs as a consequence of the dehydration process. The exposure of the compounds containing methanol mol- ecules, such as 4 and 6 to dynamic vacuum does not suppress any of the TSDC peaks (at variance with the hydrated compounds), at least in the temperature range investigated in the present work, but causes again a shift in the low-temperature TSDC peak, attributed to the reorientation of the dipoles associated with the terminal aromatic radicals. In fact the dynamic pumping removes a fraction of methanol molecules, as supported by TG and by IR spectra, again causing a rearrangement of the environment around the reorienting dipole and, as a consequence, a change in the reorient at ion parameters.Generally the TSDC peaks of all the compounds investi- gated in the present work are broad and can be described by a distribution of relaxation times rather than by a unique z. This is also true of the TSDC peaks of the samples exposed to dynamic vacuum, but the relaxation time distribution is obviously different from that for the freshly prepared com-pound. The existence of a relaxation time distribution suggests that there are slightly different environments around the reorienting dipole which affect the reorientation parameters and, as a consequence, the relaxation time given by eqn. (1).Conclusions The di- and tri-nuclear iron hydrazone-hexacyanoferrate com-pounds behave as electrets, i.e. they are able to store electric polarization. The TSDC technique, used to investigate this feature, is able to confirm the reorientation of dipolar species, even in complex compounds, and the changes induced on their reorientation parameters by changing the hydration level. The TSDC technique is a sensitive tool for monitoring even weak loss of water molecules which escape detection by the more frequently used TG technique. The authors are indebted to Mr. Carlo Mora for technical help in performing the TSDC measurements. This work was supported by the Minister0 dell'Universit8 e della Ricerca Scientifica e Tecnologica (MURST, Roma) and by the Consorzio Interuniversitario di Struttura della Materia.References 1 J. C. Bailar Jr., Coord. Chem. Rev., 1980,31, 53. 2 F. L. Urbah, in Metal Ions in Biological Systems, ed. H Siegel, Marcel Dekker, New York, 1984, vol. 13, p. 73. 3 R. C. Long and D. N. Hendrickson, J. Am. Chem. Sot., 1983, 105,1513. 4 B. M. Hoffman and J. A. Ibers, Ace. Chem. Res., 1989,16, 15. 5 R. Gross-Lannert, W. Kaim and B. Olbrich-Deussner Inorg.Chem., 1990,19,5046. 6 0.Kahn, Angew. Chem., Int. Ed. Engl., 1985,24, 834. 7 U. Casellato, P. A. Vigato and M. Vidali, Coord. Chem. Ref?., 1977, 23, 31. 8 H. Adams, N. A. Bailey, W. Daniel Carlisle, D. E. Fenton and G. Rossi, J. Chem. SOC.,Dalton Trans., 1990, 1271.9 P. A. Vigato, S. Tamburini and D. E. Fenton, Coord. Chem. Rev., 1990, 106, 25. 10 S. Ianelli, G. Minardi, C. Pelizzi, G. Pelizzi, L. Reverberi, C. Solinas and P. Tarasconi, J. Chem. SOC., Dalton Trans., 1991, 2113. 11 M. Carcelli, C. Ferrari, C. Pelizzi, G. Pelizzi, G. Predieri and C. Solinas, J. Chem. Soc., Dalton Trans., 1992,2127. 12 A. Bonardi, R. Capelletti, C. Pelizzi and P. Tarasconi, Proc. Int. Symp. on Electrets ed. R. Gerhard-Multhaupt, W. Kunstler, L. Brehmer and R. Danz, IEEE Press, Piscataway, NJ 1991, pp. 123-128. 13 C. Bucci, R. Fieschi and G. Guidi, Phys. Rev., 1966,148,816. 14 J. van Turnhout, in Electrets, ed. G. M. Sessler, Springer-Verlag, Berlin, 1989, p. 81. 15 R. Capelletti, in Defects in Solids, ed.A. V. Chadwick and M. Terenzi, Plenum Press, New York, 1986, p. 407. 16 P. Dansas, S. Mounier and P. Sixou, C. R. Acad. Sci. Paris, Ser. B, 1968,267,1223. 17 F. Rull, L. F. Sanz and J. A. de Saja, J. Electrostatics, 1980,8,221. 18 M. Bridelli, R. Capelletti and P. R. Crippa, Bioelectrochem. Bioenerg., 1981,8, 555. 19 M. Bridelli, R. Capelletti, G. Ruani and A. Vecli, Proc. 5th Int. Symp. Electrets, IEEE Press, New York, 1985. 20 A. Anagnostopoulou-Konsta and P. Pissis, J. Phys. D, 19x7, 20, 1168. 21 M. G. Bridelli, R. Capelletti, A. Vecli and M. Zaniboni, Prclc. Int. Symp. on Electrets ed. R. Gerhard-Multhaupt, W. Kiinstler, L. Brehmer and R. Danz, IEEE Press, Piscataway, NJ, 1991, pp. 720-725. 22 J. L. Leveque, J. C. Garson, and G. Bouduris, Biopolymers. 1977, 16, 1725. 23 F. Ehrburger and J. B. Donnet, J. Appl. Phys., 1979,50,1478. 24 A. Bonardi, C. Carini, C. Merlo, C. Pelizzi, G. Pelizzi, P. Tarasconi, F. Vitali and F. Cavatorta, J. Chem. Soc., Dalton Trans., 1990,2771. 25 M. Wicholas and T. Wolford, Inorg. Chem., 1974, 13,316. 26 0.P. Anderson, Inorg. Chem., 1975,14,730. 27 J. Metz and M. Hanack, J. Am. Chem. Soc., 1983,105,828. 28 C. P. Smyth, Dielectric Behavior and Structure, McGraw -Hill, New York, 1955. 29 A. M. Manotti Lanfredi, F. Ugozzoli, A. M. Camus, N. Mmich and R. Capelletti, Inorg. Chim. Acta, 1993,206, 173. Paper 3/06403D; Received 26th October, 1993