J. MATER. CHEM., 1994,4(12), 1775-1783 Sulfonated Polyaniline Films as Cation Insertion Electrodes for Battery Applications Part 1 -Structural and Electrochemical Characterization Cesar Barbero, Maria C. Miras, Bernhard Schnyder, Otto Haas and Rudiger Kotz* Electrochemistry Section, Paul Scherrer Institute, CH-5232Villigen PSI, Switzerland Sulfonated polyaniline (SPAN) was synthesized by sulfonation of polyaniline (PANI) base with fuming sulfuric acid. Thin films were cast from polymer solutions in basic media. The polymer films were characterized by X-ray photoelectron spectroscopy (XPS), Fourier transform infrared (FTIR) spectroscopy, ultraviolet-visible-near-infrared spectroscopy, scanning electron microscopy (SEM) and cyclic voltammetry. XPS in combination with FTIR showed that the preparation procedure led to ca.47% sulfonation of an otherwise unchanged polyaniline backbone. The NIR spectra of SPAN films showed a polaron band at higher energies than with polyaniline. This is in agreement with the lower conductivity of SPAN as compared with polyaniline. SEM micrographs of the SPAN films showed a compact globular morphology. Electrodes modified with thin SPAN films exhibited two redox steps, both in aqueous and in non-aqueous electrolytes. The specific charge stored in SPAN films was found to be ca. 37 A h kg-' in aqueous solution (only the first redox step) and ca. 68 A h kg-' in non-aqueous media (both redox steps). A practical SPAN-Li battery could have 50% more specific energy than a PANI-Li battery.The optical spectra of SPAN films exhibited bands at 310, 450 and 750 nm, the intensities of which changed during the redox process. The absorption coefficients of SPAN (emeraldine base state) solutions had values of a =410 at 31 3 nm and a =239 at 563 nm. The suitability of SPAN for use as a cation-insertion material for battery and electrochromic applications is discussed. Polyaniline (PANT) is one of the most promising conductive polymers for battery applications owing to its rather high theoretical specific charge, high conductivity (> 1 S ern-') and stability and because it has an inexpensive monomer.' Although PANI-Li batteries are commercially available,' their specific energy is relatively low (<30 W h kg-l) as compared with that for conventional systems.During electrochemical oxidation/reduction of the PANT polymer matrix, positive charges are created/neutralized within the film. In order to maintain the film's electroneutral- ity, ions have to be exchanged with the electrolyte s~lution.~ Previous studies of PANI showed that anions and protons are exchanged in aqueous solution^,^,^ whereas in non-aque- ous media anions are mainly used as the charge compensating Pure cation exchange by the polymer is necessary for its use in cation-transfer batteries. The transformation of an anion-exchanging polymer into a ca tion-exchanging polymer can be achieved by irreversible incorporation of negatively charged groups into the polymer. Different procedures can be used, such as polymerization of the monomer in the presence of a polyelectrolyte [e.g.poly-(styrene ~ulfonate)~ J or the preparation of composites [e.g. PANI-Nafion']. In order to keep the stored specific charge at reasonably high values, the amount of mass added should be low. One way to achieve this is by covalent bonding of light anionic groups to the polymer backbone. It was shown that sulfonate groups can be linked to the PANI backbone by sulfonation of the polymerlo to give partially (50%) sulfonated polyaniline (SPAN). It was also demonstrated that films can be produced from polymer solu- tions." Some of the properties of the sulfonated polymer itself have already been described in the In this work we describe the physical and electrochemical characterization of SPAN films and solutions.The parameters in polymer synthesis and film formation that are relevant to applications are discussed. Information about stability, specific charge capacity and electroactivity is necessary when using the polymer in electrochemical systems, such as batteries and electrochromic devices. In further work,14 the ion exchange properties of SPAN as studied by probe beam deflection and electrochemical quartz crystal microbalance, will be reported in detail. Experimental Electrochemistry A conventional three-electrode cell was used to perform the electrochemical investigations. The counter electrode was a Pt plate, the reference electrode a saturated calomel electrode (SCE).The working electrodes were SPAN-modified Au films on glass or SPAN films on GC plates (glassy carbon), the reverse side of which was covered with an inert varnish (Lacomit). The electrochemical experiments were cor itrolled by use of a BAS lOOa (Bioanalytical Systems) potentiostat. The redox charge was obtained by integration of the current- potential data. Acetonitrile (Aldrich, HPLC grade) and propyl- ene carbonate (Fluka, puriss.) were used as received. The electrolyte salts were dried under vacuum. All aqueous solu- tions were prepared with ultrapure water and analytica 1 grade reagents. All potentials are quoted uersus the SCE. X-Ray Photoelectron Spectroscopy (XPS) The XPS spectra were measured with a Kratos IS 300 electron spectrometer using non-monochromatized Mg-Ka radiation (1253.6 eV).The fixed analysis transmission mode was chosen. The vacuum in the analysis chamber was always greater than 2 x Torr. The samples (SPAN films on Au) were investigated in the pristine state, because even slight Ar sputtering was found + to alter the composition significantly. Scattering cross-sections for quantitative analysis were taken from the 1iterat~re.l~ No charging effects were observed. Fourier-transform Infrared (FTIR) Spectroscopy A Perkin-Elmer (PE) 2000 FTIR spectrophotometer was used to measure the ex situ IR spectra. Two arrangements were used: transmission and reflectance. For transmission measure- ments the polymer was deposited onto an Si wafer.The spectra of these polymer films were recorded with a bare Si wafer as the reference using a PE automatic sample shuttle. For reflectance measurements the polymer was deposited onto Au film on glass. The measurements were made at 70" with p-polarized light in a Harrick variable-angle reflectance accessory. Electrochemically treated films were obtained by cycling the potential of modified electrodes in the electrolyte solution between -0.2 and 0.5 V us. SCE. In order to obtain films in different redox states, the electrodes were kept at the desired potential until the current had decayed to negligible levels, then they were withdrawn from the solution and their spectra measured. The solution was carefully deaerated by Ar bub- bling and then maintained under an argon blanket.The electrodes were always removed from the solution under potential control and quickly dried in an argon stream. Ultraviolet-VisibIeNear-infrared (UV-VIS-NIR) spectroscopy Spectra of SPAN solutions were recorded in quartz cells (Suprasil, Hellma) of 1 cm optical pathlength. Spectra of SPAN films were recorded with the polymer deposited onto quartz plates. For studies of the electrochromism of SPAN, thin films were deposited onto transparent electrodes [indium-tin-oxide (ITO), 25 SZ cmP2; Balzers]. The electrodes were washed with toluene and ethanol and the contact was made with electrodag. The electrodes were mounted in a Kel-F holder. All optical absorption measurements were carried out with a Cary 2400 UV-VIS-NIR spectrophotometer.Scanning Electron Microscopy (SEM) Scanning electron micrographs were recorded from SPAN films deposited onto a polished Si wafer. A Hitachi 9410 scanning electron microscope was used. Polymer Preparation All chemicals used were of reagent grade. PANI Polyaniline salt was prepared according to the method of McDiarmid et a1.16 A solution of aniline in 1 moll-' HC1 was mixed with an equimolar amount of the oxidant (ammonium persulfate). Polyaniline base (I j was obtained by treating the polyaniline salt with NH,OH solution (10%) for 24 h with continuous stirring. As an alternative route, commer- cial polyaniline (Versicon, Allied Signal Corp.) in salt form was subjected to the treatment with NH,OH solution to obtain polyaniline base.SPAN The polyaniline sulfonation procedures utilized have pre- viously been described in the and are only briefly summarized below. Emeraldine base 0.5 g (I)was finely ground and dissolved in 40 ml of cool (<5 "Cj fuming sulfuric acid (30% SO, in H2S04) to yield a dark purple solution (11). The sulfonation was carried out at 5°C with constant stirring in a 500 ml closed Erlenmeyer flask for 2 h. The colour of the solution changed from dark purple to dark blue when sulfonation was complete (111 in Scheme 1). The liquid was then added, in small portions, to 200ml of methanol that were kept cooled by an ice-bath. Upon neutralization of the acid, SPAN (IVj precipitates out. It was found that the addition of acetone, as recommended by Yue et al.," was J.MATER. CHEM., 1994, VOL. 4 detrimental to polymer precipitation. The solution was filtered and the filtrate was washed with several 50ml portions of methanol until the pH of the washings was neutral. The polymer (IV) was then dried under vacuum at 50 "C for 48 h and stored in a desiccator. The average yield was below 20%. It has been reported previo~sly,'~ that the yield and conduc- tivity of the sulfonated polymer reached maximum values after 1 h of sulfonation, then decreased. We tried sulfonation at ambient temperature (20°C) using a reaction time of 1 h, the average yield was ca. 60%. The resulting polymer dissolved faster in ammonia solution, and the resulting films were more homogeneous (optical microscopy).The sulfonation procedure is summarized in Scheme 1. First the PANI base was dissolved in concentrated sulfuric acid. The imine nitrogens became protonated with formation of polyaniline sulfate salt. The subsequent attack of the electro- philic agent (SO,) occurred in the uncharged units, which have higher electronic density. Therefore a limit of 50% sulfonation was obtained with this medium. Two side reactions could occur in the strongly acid media: (i) breaking of the nitrogen bonds with formation of quinones and (ii) cross-linking of different chains; both reactions would be promoted by the presence of SO,. Preliminary data on the molecular weight of SPAN" suggested that both reactions occur. The starting material had a narrow dispersion of molecular weight but yielded a mixture of SPAN polymers with larger dispersion than PANI.Longer sulfonation times would allow side reac- tions to occur. After sulfonation, the solution of SPAN in H2S04 was added to methanol. The purpose of this step was to neutralize the remaining SO, and H,SO, with the methanol by formation of methyl sulfates without excessive heat release. The resulting solid was SPAN in its zwitterion state, which was insoluble in methanol and therefore precipitated out. SPAN Solutions Two milligrams of SPAN (IV) were dissolved in 10ml of 0.1 moll-' (or 1 moll-lj NH,OH. As discussed before, poly- mer samples prepared by the conventional procedure dissolve slowly (1-2 h) whereas samples prepared by the modified procedure dissolve instantaneously.The polymer was con- verted to the salt, [C12HxN2] SO,-NH,+, which was soluble in water and gave a dark blue solution. The properties of the solutions changed on storage (see below). Therefore fresh solutions were prepared before use. Solutions of the leucoemeraldine form of SPAN were pre- pared by reduction of SPAN solution with hydrazine (1% in water). The slow reduction was accompanied by a loss of colour and was complete after 2 h. SPAN Films These were cast on appropriate substrates by pouring a portion (typically 0.1 ml) of the polymer solution onto the substrate and drying under an IR lamp. The ammonium salt of SPAN decomposed upon heating, releasing NH,, which was followed by evaporation of the solvent.The resulting film was obtained in its zwitterionic form ([C12H8N2H2]+ -SO,-j, which was insoluble in water (at pH <8). The films were stable in air for several months and had a reproducible electrochemical response. We found that films obtained by spontaneous drying of the solution in air (even in a dry atmosphere) have no reproducible electrochemical response. Results and Discussion XPS The XPS investigations were performed with SPAN films deposited onto gold substrates. The electrodes were first cycled J. MATER. CHEM., 1994, VOL. 4 (1) pdyaniline (emeraldine base) 30%SO3 In HSO,1 (11) polyanilinesolution (vW) more stable structure (separated radical cations) (111) SPAN solution (dark Hue) fd-"q=y&yj;so; (Iv) SPAN (green powder) 9+\/ \/ N-\/ *+ \/ Scheme 1 in 1 mol I-' HC1 to assure that the film composition corre- sponded to the electrochemically active material and not to the virgin state after sulfonation.For the sake of comparison, XPS of a virgin film was also carried out. In addition to the survey spectrum the emission peaks of S 2p, N Is, C Is, C12p and 0 1s were investigated. The resulting spectra are reproduced in Fig. 1, both for the sample as prepared and for the same sample after electro- chemical cycling. For the latter the electrode was removed from the electrolyte in its oxidized state. The main difference between the spectra are the reduced intensities of 0 Is, N 1s and S 2p emission in the spectrum recorded after electrochemi- cal cycling.This observation could be explained by the presence of residual (NH4),SO4 in the as-prepared film, which might originate from the sulfonation step and the subsequent precipitation. After electrochemical cycling the contamination was found to have disappeared from the film. The contribution of these species to the spectra of N 1s and S 2p was on the high binding energy side at binding energies of 402 eV for N and at 168.5 eV for S. These binding energies correspond to NH4+ and species." In addition, the presence of NH4+ species was indicated by the low C:N ratio of 3.2 (theoretical ratio 6) in the untreated film. The S :N ratio of the untreated film was higher than that of the sample after cycling, indicating additional SO,2-retention.For the sample after electrochemical cycling, only one S peak occurred at a binding energy of 167.8 eV, typical of SO3-groups." The nitrogen peak was observed at 399.5 eV with a shoulder towards higher binding energies at 401.7 eV, which is in agreement with previous investigations.20 This shoulder indicated positively charged nitrogen atoms. A weak C12p peak could be observed for some samples at a binding energy of 200.4 eV, indicating C1- retention.15 The S :N ratio, which corresponds to the degree of sulfon- ation, was determined for five independent samples. The average degree of sulfonation was about 0.47 0.07%, the average C1- :N ratio was between 0.0 and 0.1.The C :N ratio was found to be 6.5 f2, while the 0:S ratio was 4.5 & 1, indicating roughly one benzene ring per nitrogen atom and some oxygen, probably originating from H20 and from the SO3-groups. The XPS results clearly indicate up to 47% sulfonation of PANI. In the half-oxidized state most of the charge was compensated by the SO3-groups, and a small incorporation of C1- was necessary for compensation of the remaining charge. The corresponding formula for the polymer unit can therefore be written as -{ [(C6H3NH+*S03-)0.47-(C6H4NH+ -C1-)o.03]-(C6H,NH)o.51-. Because the films were removed from the electrolyte in the oxidized state, this corresponded to an almost total compensation of the radical cation charge by the sulfonate group.FTIR The structure of the polymer films was investigated by trans- mission and reflection FTIR in the range 450-6500cm-'. The transmission spectrum of the SPAN film (Fig. 2) agrees quite well with the one reported previously (limited to the region 1700-500cm-1) for bulk SPAN.l2 The assignment of the vibrational bands2' in SPAN is described in Table 1. J. MATER. CHEM., 1994, VOL. 4 as prepared cycled01s t I10.0 I: *. I6.0 2.0 ' i, 540 530 520540 530 N 1s11 3.0 520 .r 1 0 t . G' . 0 . It 410 400 390 410 400 390 c 1s 6.0 4.0 2.0 0 295 285 275 295 285 275 S 2P 2.0 1.5 I' 1.o ?f b 1. I 0.5 180 170 160 180 170 160 binding energylev Fig. 1 XPS spectra of SPAN film samples on Au substrate as prepared and after electrochemical cycling in 1moll-' HCI For comparison, a similar assignment was made for a PANI reflectance spectra of SPAN films in the reduced and oxidized film.The main difference between SPAN and PANI was the state. By comparing Fig. 3 with Fig. 2 it can be seen that most presence of bands at 1082cm-', 1024cm-' and 618 cm-l, parts of the spectra are identical. However, a band at ca. attributed to the SO3-group in SPAN. The occurrence of 1450cm-1 is present with the virgin sample that is not seen two bands, attributed to SO3-, between 1000cm-' and with the electrochemically treated sample. This band was 1100cm-' indicated binding of the SO3-group to the benzene assigned to the NH4+ions that remain in the polymer together ring2' The FTIR reflectance spectrum of a virgin SPAN film with trapped sulfuric acid.Upon cycling in aqueous solution (Fig. 2) was not significantly different from the transmission the salt dissolved and the band disappeared. The correspond-spectrum. ing band for NH,+ at 3400cm-' overlapped with the CH Electrochemically treated samples were cycled in 1 moll-' and NH bands of the polymer. The presence of these contami-HC1 solution before the FTIR measurements were made in nants was deduced from the XPS results described above. order to assure equilibration of the film. Fig. 3 shows the By comparing the two spectra in Fig, 3, it can be seen that J. MATER. CHEM., 1994, VOL. 4 0.30 0.20 8 s e8 a a 0.10 0.00 400 1400 2400 3400 wavenumbehm-' Fig.2 FTIR spectra of SPAN films cast from aqueous solution onto: (a) an Si wafer (measured by transmission); and (b) a gold electrode (measured by reflection, 70" with p-polarized light) Table 1 Assignment of 1R absorption bands in SPAN ' ~~~~~~~~~~ IR band/cm bond group vibration mode 3235.2 N-H" aromatic amine stretching 3066.7, 2862 C-H" aromatic stretching 1 600.3 C=Nb quinoneimine stretching 1508.0 C-cb aromatic stretching 1451.7 NH,' 1425.9 C=Cb aromatic stretching 1311.1 C-N' secondary aryl amine bending 1176.2 C-Hh in-plane bending aromatic 1081.8, 1024.2 s=o sulfonate stretching 823.7 C-H' out-of-plane bending aromatic 707.9 c-s aromatic stretching 617.8 s=o sulfonate stretching "Also present in PANI, but difficult to see in the spectra because bands overlap with the broad conduction band of PANI.'Also present in PANI. the same bands are present in the oxidized and reduced state. However, there are differences in the band intensities. The absorbance above 4000 cm-' is significantly higher in the oxidized than in the reduced state. Such absorption corre- sponded to the tail of the NIR band of the polaron,22 which was only present in the emeraldine state of the film. Differences could also be observed in the region 1700-1400 cm-' (see insert in Fig. 3). The intensity of the band at ca. 1600 cm-' increased on oxidation, while the intensity of the band at 1510 cm-' decreased.The band at 1510 cm-' was attributed to the C-C vibration of the benzenoid ring.21 The vibration at ca. 1600 cm-' was attributed to the C=N of the quinone- imine ring." On oxidation the amine (containing aromatic C-C bonds) was partially converted to quinoneimine (con- taining C=N bonds). The insert in Fig. 3 shows that the C=N band is located at ca. 1620cm-' in the reduced state but seems to shift on oxidation towards ca. 1580cm-'. It is not clear whether the band actually shifts or whether a third band appears at 1580cm-' while the one at 1620cm-' remains unchanged. A similar variation of band intensity with the oxidation state was observed for PANI.22 Under the experimental conditions described, the reflection spectra of PANI filrns was measured in the reduced and oxidized state.The band at 1510cm-', corresponding to the benzenoid ring, decreased on oxidation and the band at ca. 1600cm-' (quinoneimine) increased on oxidation. The apparent band shift is also present for PANI. The background absorption below 4000 cm-' increa.;ed on oxidation of PANI, as for SPAN. However, a new band appeared at ca. 0.5 eV (4000 cm-'), corresponding to free carrier absorption. In SPAN such a band exists at energies above 1eV (8000 cm-I). The difference could be explained by a higher localization of the charge carriers in SPAY due to the steric effect of the sulfonate group. The lower mobility of the charge carriers makes SPAN less conductive than PANI." The FTTR investigations led to the conclusion that sulfon- ation of PANI was' achieved with an otherwise unchanged polymer structure.UV-VIS-NIR Spectroscopy Solutions Optical spectra of the polymer salt form were recorded in the solution used for film formation (0.1 mol I-' NH,OH). The spectra revealed bands at 317 nm and 563 nm [Fig. 4(b)]. These bands corresponded to the anionic (-SO,-) form of the polymer. The spectra are similar to those previously reported for the sodium salt of SPAN in water." The band at 317 nm was assigned to the n+n* transition of the aniline ring.14 The band at 563 nm was assigned to the 'exciton' transition of the quinone and was related to intrachain hopping.23 To confirm this assignment, the reduced form of SPAN was produced in solution. The UV-VIS spectrum [Fig.4(a)] revealed a single band at ca. 320 nm. This agreed with the spectra for the leucoemeraldine base of PANI that contain only the band due to the n+n* transition. The optical absorption obeys the Beer-Lambert law between 20 and 200 pmol I-'. This fact allowed the calculation of the specific absorption coefficient of the polymer chmmo- phores. The band at 313 nm has an absorption coefficient of J. MATER. CHEM., 1994. VOL. 4 400 1200 2000 4000 6000 wavenurnber/crn-' Fig. 3 FTIR spectra of SPAN films at different redox states: (a) reduced (emersed at -0.2 V us. SCE); and (h)oxidized (emersed at +0.5 V 1:s. SCE). The electrolyte solution was 1 moll-' HCI. The insert shows the region between 1700 and 1400 cm-l.1.0 r 0.8 8 0.6 1 8 .na 0.4 0.2 0 300 400 500 600 700 800 wavelengthhm Fig. 4 UV-VIS absorption spectra of SPAN solutions: (a)leucoemer-aldine state (reduced): and (b)emeraldine state (oxidised) a =410, for that at 563 nm a =239. By assuming a molecular weight for the polymer unit of 133 g mol-I (one aniline unit with 50% sulfonation), the molar absorption coefficients are ~=5400at 313 nm and ~=3200 at 563nm. These values indicate forbidden transitions (e< 10OOO).24 The value is low compared with the extinction coefficient, c;1= 1.6 x lo5cm-(corresponding to E = 10000),25determined for solid-state PANI at 320 nm. Film Optical spectra of the polymer films were measured (in the range 200-2600nm) on thin films of SPAN deposited onto quartz plates from aqueous solution.The PANI film was deposited from an emeraldine salt dispersion.26 The spectra in the UV-VIS range agreed with those reported previously.',' The spectrum of SPAN exhibited bands at 320nm and 850 nm and a shoulder at 440 nm [Fig. 5(a)]. The band at waven umber/crn-' 10000 5000 4000 2.60 1.95 1.30 n a 0.65 0.00 200 lo00 1800 2600 wavelength/nrn Fig.5 UV-VIS-NTR spectra of (a) SPAN: and (h) PANI films deposited onto quartz plates. SPAN was deposited from aqueous solution, PANI from a propan-2-01 dispersion. 320 nm corresponded to the n+n* transition of the aromatic ring. The shoulder at 440 nm was associated with the localized radical cations of aniline units.The band at 850 nm was broad and related to absorption by the metallic polarons.2s The spectrum did not reveal the band at ca. 1600nm present in the spectrum of PANI [Fig. 4(6)]. It is likely that such a band was shifted to higher energies and overlapped with the band at 850 nm. SEM The SEM micrograph of a SPAN film on Si (Fig. 6) revealed a globular structure with features of cn. 20 pm in size. The morphology differed significantly from that of electrochemi- cally deposited PANI,27 which has a fibrillar structure. The SPAN-film morphology was similar to that of PANT films J. MATER. CHEM., 1994, VOL. 4 Fig. 6 Scanning electron micrograph of a SPAN film deposited onto an Si wafer deposited from aqueous dispersions28 and from solutions in N-methylpyrr~lidone.~~Based on these SEM results, the porosity of the SPAN films was probably lower than that of electrochemically deposited PANI.Electrochemically deposited PANI grows around nuclei, favouring fibre forma- tion, whereas deposition from solutions takes place over the whole surface by a precipitation reaction that creates com- pact layers. Cyclic Voltammetry Films prepared as described in the experimental part were tested by cyclic voltammetry in aqueous and non-aqueous solutions. Aqueous Solutions The electrochemical behaviour of films in aqueous solutions was studied using several concentrations (0.1-4 moll-') of strong acids (HC1, H2S0,, HClO,, CF,SO,H). After immer- sion of the electrode, several (> 10) cycles of between -0.2 and 0.5 V us.SCE at 100 mV s-' were performed in order to equilibrate the film with the electrolyte. A typical cyclic voltammogram recorded in 0.1 moll- HC1 between -0.2 and 0.85 V us. SCE is shown in Fig. 7. It agrees with those previously reported." SPAN exhibited two redox processes (at 0.15 and 0.7 FS. SCE) as did PANI. The maximum currents of both peaks were linear functions of the scan rate (10-250 mV s-I), indicating that the redox sites in the layer 0.8 4 0.4 E =at L2 0.0 -0.4 0.0 0.4 0.8 electrode potentiaVV vs. SCE Fig. 7 Cyclic voltammogram of a SPAN film on GC in 0.1 moll-' HCl. Scan rate: 50 mV s-'. -0.4 0.0 0.4 0.8 electrode potentialN vs. SCE Fig.8 Cyclic voltammogram of a SPAN film on GC in 0 1 mol I-' LiC104-ACN. Scan rate: 50 mV s-'. were in thermodynamic equilibrium with the electrode poten- tial during the voltammetric excursion. Although PANI showed no electroactivity in solutions of pH<4, SPAN showed a clear electroactivity up to pH 7. At pH 7 the two peaks overlapped yielding a broad wave at ca. 0.5 V us. SCE. Non-aqueous solutions The electrochemical response of SPAN in non-aqueous elec- trolytes was also investigated. After immersion of the electrode in the electrolyte. several (~20)cycles of between -0.2 and + 1.0 V us. SCE were performed to achieve equilibration of the polymer film with the solution. The redox charge increased significantly during this period, and the peaks became more clearly defined.Fig. 8 shows the cyclic voltammogram obtained in acetonitrile- LiC104. The two redox processes present at 0.2 and 0.95 V us. SCE were broader than those in aqueous solution Electrochemical Stability of the Films The stability of the polymer towards oxidation/reductmon was tested in aqueous solution, by recording the charge during cycling between -0.2 and 0.5 V us. SCE. Some loss of charge occurred, which depended on the pH of the soluticm. The percentage loss of redox charge after 1000 cycles (scan rate: 100 mV s-l) at different proton concentrations is shown in Fig. 9. The polymer stability decreased with increasing H+ concentration. It is known that the polymer is soluble in concentrated acids,' therefore dissolution of the films at high H+ concentration could explain their instability.Cycling the electrode in aqueous solution between -0.2 and 0.8 V us. 8 20 0)F2 10i0 3 0 1 ............. . ......._..........................................................a.......... 0.0001 0.1 1 2 4 proton concentration/mol r1 Fig.9 Percentage loss of redox charge of SPAN films after 10oO cycles in solutions with different proton concentrations SCE induced strong degradation with almost total loss of charge after 1000 cycles. In non-aqueous solutions the loss of charge during cycling between -0.25 and 1.0V us. SCE at 50 mV s-' was only 3% after 1000 cycles. Stability of the Solutions It was found that the properties of SPAN solutions changed during storage.When the solution was stored for more than one week, the bands in the UV-VIS spectra shifted towards higher energies. A possible explanation was breaking of some aminic bonds with shortening of the SPAN chains. The breaking of the chains could occur by hydrophilic attack of water molecules on the aminic bonds. Shorter chains have a reduced delocalization of electrons, which increases the transition energy.13 Possible changes occurring in aqueous SPAN solutions with time were also studied by casting films from fresh and old (3 weeks) solutions. The IR spectra showed no significant differences in the region between 1800-450 cm-l, suggesting that no new compound was formed.However, the intensity of the free carrier absorption above 4500 cm-decreased significantly. This decrease could be related to a lower conductivity of the film deposited from an old solution. The electrochemical response of films deposited from old solutions was characterized by a sluggish voltammogram with a large separation of anodic and cathodic peaks. Polymer solutions in propan-2-01 media suffered no notice- able changes, indicating that the stability problem might be resolved by film deposition from alcoholic solutions. The increased stability in alcoholic solutions supported the idea that the attack by water molecules (present only in low concentration in alcoholic solutions) caused the degradation reaction. Specific Charge The amount of polymer deposited could be controlled by the amount and concentration of the solution used to form the film.The calculation procedure for the specific charge assumed that none of the material deposited dissolves into the electro- lyte. If it did, the film mass would be overestimated and the specific charge underestimated. Therefore, the values of specific charge reported probably represent lower limits. Aqueous Solutions The variation of the charge with the amount of material deposited is shown in Fig. 10 for polymer prepared by the modified procedure. The value for specific charge was obtained from the slope, thus any contribution of the background charge was avoided. The redox charge was measured in 1 moll-' HCl solution in H20 by cycling the electrode between -0.2 and 0.5 V us.SCE at 10 mV s-'. Lower scan rates gave the same charge. A value of 37f2 A h kg-I was deduced for the first peak of SPAN (the second process could not be used for charge storage in aqueous solution because the polymer degraded significantly, as discussed above). The fact that the polymer prepared by the conventional procedure had a lower specific charge (22A hkg-')' suggested that further modifications of the preparation procedure might boost the value even further. Non-aqueous solutions The charge was measured by cycling a SPAN film (modified procedure'') between -0.5 and 1.0V us. SCE at 10mV s-' in non-aqueous electrolyte (ACN-LiC10,). The plot of redox charge against the amount of polymer was linear (Fig.10). J. MATER. CHEM., 1994, VOL. 4 0 20 40 60 80 polymer mass/pg Fig. 10 Dependence of redox charge of a SPAN film on the amount of polymer deposited. Aqueous solution (1 moll-HCl): charge measured during cycling between -0.2 and 0.5 V cs. SCE at a scan rate of 10 mV s-'. Non-aqueous solution (0.5 moll-' LiC10,-ACN): charge measured during cycling between -0.2 and 1.0 V z's. SCE at a scan rate of 10 mV s-'. (A)LiClO,-ACN; (0)HC'l-H20. 0.5 0.4 g 0.3 2P n(II 0.2 0.1 I I I I 300 400 500 600 700 800 wavelengthhm Fig. 11 Evolution of the UV-VIS spectra of a SPAN film on IT0 with applied potential (V us. SCE). Electrolyte: 0.4 mol 1-' NaC10,-0.1 moll-' HC10,. From the slope of the plot a value for the specific charge of 68k5 A h kg-' (both redox steps) was obtained.Films of polymer prepared by the conventional procedure" gave a value for specific charge of 35 A h kg-' in the same media. With the measured value of specific charge of SPAN films, the calculation for the specific energy of a SPAN-Li battery could be made; assuming that the SPAN film exchanged only cations during oxidation/reduction, as demonstrated else-where.14 The mass balances of the active battery materials and the electrolyte are summarized in Table 2. The amount of solvent was calculated for 2 moll-' LiC10, solution with Table 2 Mass balance for a rechargeable Li-SPAN battery (not including current collectors, separator and cell container) component mass equivalent/g (mol e-)-' SPAN 394 LiClO, (0.5 MW) 53.75 Li (0.5 MW) 3.5 electrolyte 60 (solvent +LiClO,) total 511.25 (ca.52 Ah kg-'+ca. 130 W h kg-I) J. MATER. CHEM., 1994, VOL. 4 a density of ca. 1.2 g cm-3 and with an approximated volume, which neglected any swelling of the polymer film upon immer- sion in the electrolyte. The overall specific charge was 52 A h kg-'. By using an average discharge voltage of 2.5 V, a 'semi-empirical' specific energy of ca. 130 W h kg-' was estimated. During previous work,6 a value of 87 W h kg-' was obtained for the specific energy of a PANI-Li cell (using both redox steps). The specific energy for a SPAN-Li battery was therefore ca. 50% higher than a PANI-Li battery.Although the measured specific charge of the polymer electrode was significantly lower, the reduced contribution of solvent to the total mass of the battery would permit a net gain of specific energy of the battery. Electrochromic Properties A potential application for conducting polymers is in elec- trochromic devices. While it is known that PANI can be used for that purpose,3o no quantitative study has previously been performed on SPAN. The evolution of the SPAN spectra with applied potential was investigated (Fig. 11). Three bands were observed at 750, 450 and 310 nm. The band at 310 nm was attributed to the n-+n* transition of the aniline ring23 and was hypsochromically shifted relative to that of polyaniline. The hypsochromic shift of the UV band in SPAN (relative to polyaniline) was attributed to the inductive effect of the sulfonic groups in the aromatic ring.'' The band at 450nm was assigned to the localized cation radical.25 The absorbance at this wavelength first increased with oxidation, then decreased.This corresponded to the initial formation of radical cations that disappeared upon further oxidation. The band at 750 nm corresponded to the metallic polaron.22 The absorbance at 750 nm increased continuously with oxidation and shifted to higher energies in the NIR region. The behav- iour of the bands in SPAN was very similar to polyaniline. The most widely used electrochromic material,31 tungsten oxide, bleaches on oxidation and cations are inserted; SPAN bleaches on reduction with cation expulsion.A 'rocking chair' configuration could therefore be assembled in which a tung-, sten oxide electrode would be the negative electrode and SPAN the positive electrode in the coloured state. As the electrochromic effects are complementary, a high-contrast device could be achieved. Conclusions SPAN was produced by chemical sulfonation of PANT. The best procedure for this used a short (1 h) reaction time at room temperature. The polymer obtained dissolved faster in basic media and more homogeneous films with a higher charge density could be cast from these solutions. XPS investigations suggested an average degree of sulfon-ation of 47%. FTIR results indicated that the sulfonation did not significantly alter the PANI backbone but induced a higher localization of the charge carriers. SPAN solutions in aqueous ammonia changed with time and this could be seen from a hypsochromic shift of the absorption bands, a decrease of absorption, a decrease in the intensity of the free carrier absorption in the IR and a more sluggish electrochemical response for films formed from old solutions. Alcoholic solutions were stable.SPAN was electroactive in non-aqueous media and in aqueous solvent up to pH 7. The charge density was at least 35 A h kg-' (first redox step) in aqueous solution and more than 65 A h kg-' (both redox steps) in non-aqueous solution. A SPAN-Li battery could yield up to 50% more specific energy than a PANI-Li battery. The films were distinctly electrochromic.The evolution of the UV-VIS spectra with applied potential was similar to that of PANT. Professor E. M. Genies is gratefully acknowledged for helpful discussions and making unpublished results available. The project was financed by the Swiss National Science Foundation, Grant No. 20-32504.91. References 1 E. M. Genies, M. Lapkowski and C. Tsintavis, New J. Chem., 1988, 12, 181. 2 T. Matsunaga, H. Daifuku, T. Nakajima and T. Kawagoe, Polym. Adv. Technol., 1990, 1, 33. 3 W. S. Huang, B. D. Humphrey and A. G. MacDiarmid, J. Chem. Soc., Faraday Trans., 1986,82,2385. 4 C. Barbero, M. C. Miras, 0. Haas and R. Kotz, J. Electrochem. SOC.,1991, 138,669. 5 H. Daifuku, T. Kawagoe, N. Yamamoto, T. Ohsaka and N.Oyama, J. Electroanal. Chem., 1989,274,313. 6 J. Desilvestro, W. Scheifele and 0. Haas, J. Electrochem. Soc., 1992,139,2727. 7 C. Barbero, M. C. Miras, R. Kotz and 0.Haas, Synth. Met., 1993, 55-57,1539. 8 Y. Kang, M-H. Lee and S. B. Rhee, Synth. Met., 1992,52,319. 9 D. Orata and D. A. Buttry, J. Electroanul. Chem., 1988,257,71. 10 J. Yue and A. J. Epstein, J. Am. Chem. Soc., 1990,112,2800. 11 J. Yue, Z. H. Wang, K. R. Cromack, A. J. Epstein and A. G. MacDiarmid, J. Am. Chem. Soc., 1991,113,2665. 12 J. Yue, A. J. Epstein and A. G. MacDiarmid, Mol. Cryst. Liq. Cryst., 1990, 189,255. 13 Y. Cao, S. Li, Z. Xue and D. Guo, Synth. Met., 1988,16,305. 14 C. Barbero, M. C. Miras, R. Kotz and 0. Haas, unpublished work. 15 J. J. Yeh and I.Lindau, Atomic Data and Nuclear Data Tables, 1985,32,1-155. 16 A. G. MacDiarmid, J. C. Chiang, J. C. Richter, A. F. Richter, N. L. D. Somarisi and A. J. Epstein, in Conducting Polymers, ed. L. Alcacer, Reidel, Dordrecht, 1987, pp. 105-120. 17 J. Yue, G. Gordon and A. J. Epstein, Polymer, 1992,33,4410. 18 E. M. Genies, L. Rebattet and 0.Gay, personal communication, 1993. 19 Perkin-Elmer Corporation, Handbook of X-Ray Photoelectron Spectroscopy, Eden Prairie, MN, 1977. 20 E. T. Kang, K. G. Neoh, Y. L. Woo and K. L. Tan, Polym. Commun., 1991,32, 41 3. 21 D. Lin-Vien, N. B. Colthup, W. G. Fateley and J. G. Grasselli, The Handbook of Infrared and Raman Characteristic Frequencies of Organic Molecules, Academic Press, London, 1991. 22 N. S. Sariciftci, H. Kuzmany and H. Neugebauer, J. Molec. Electron., 1987,3, 141. 23 J. M. Ginder and A. J. Epstein, Phys. Rev. B, 1990,41, 10674. 24 R. M. Silverstein, G. C. Bassler and T. C. Morrill, Spectrometric Identijication of Organic Compounds, Wiley, New York, 1981. 25 D. E. Stilwell and S-M. Park, J. Electrochem. Soc., 1988,135,2491. 26 B. Wessling, Synth. Met., 1991,41-43,907. 27 J. Desilvestro and W. Scheifele, J. Muter. Chem., 1993,3, 263. 28 B. Wessling, Adu. Maters., 1993, 5, 300, and references cited therein. 29 J. Michaelson, A. J. McEvoy and T. Shimidzu, Chimia, 1993, 47,490. 30 A. Kitani, J. Yano and K. Sasaki, J. Electroanul. Chem., 1986, 209, 227. 31 C. G. Granqvist, Appl. Phys. A, 1993,57, 3. Paper 4/02015D; Received 5th April, 1994