Polyaniline as pH-sensitive component in plasticized PVC membranes

Journal of

Electroanalytical Chemistry Journal of Electroanalytical Chemistry 560 (2003) 69–78 www.elsevier.com/locate/jelechem

Polyaniline as pH-sensitive component in plasticized PVC membranes Tom Lindfors *, Sini Ervel€ a, Ari Ivaska Process Chemistry Group, c/o Laboratory of Analytical Chemistry,
Abo Akademi University, Biskopsgatan 8,
Abo, Turku 20500, Finland Received 12 April 2003; received in revised form 27 June 2003; accepted 4 July 2003

Abstract The pH and redox sensitivity of electrically conducting polymer films consisting of a mixture of polyaniline (PANI) and plasticized poly(vinyl chloride) (PVC) have been studied with potentiometry, UV–vis spectroscopy and energy dispersive X-ray analysis (EDXA). It is well known that PANI is highly Hþ -selective and can easily be dissolved in many organic solvents with functionalized organic acids. PANI is therefore very suitable to be used as a Hþ -sensitive component in plasticized PVC membranes. In this work, we have studied the behaviour of PANI as the Hþ -sensitive component in plasticized PVC membranes as well as the influence of added lipophilic salts, tridodecylmethylammonium chloride (TDMACl) and potassium tetrakis(4-chlorophenyl)borate (KTpClPB), on the pH and redox sensitivity of these membranes. The pH sensitivity was studied between pH 2 and 9. The electrode membranes were prepared according to the all-solid-state electrode configuration by placing the membrane directly on a glassy carbon substrate. The PANI content in the membranes was varied from 0% to 100% (m/m) and the TDMACl and KTpClPB content from 0% to 40% (m/m). PANI was dissolved in tetrahydrofuran (THF) with phosphoric acid dihexadecyl ester. It was found that appropriate amounts of plasticized PVC decreases the hysteresis effect of pure PANI that was observed in the potentiometric measurements. It is also shown that TDMACl facilitates the emeraldine salt (ES)–emeraldine base (EB) transition of PANI while KTpClPB hinders it and allows PANI to stay in the conducting ES form even at pH 9. Differences in the redox sensitivity of membranes containing TDMACl and KTpClPB will also be discussed. Ó 2003 Elsevier B.V. All rights reserved. Keywords: Soluble conducting polyaniline; Poly(vinyl chloride); Ion-selective electrode; pH and redox sensitivity; TDMACl; KTpClPB; Potentiometry; UV–vis spectroscopy; EDXA

1. Introduction Polyaniline (PANI) is one of the most studied electrically conducting polymers (CP) due to its high stability, easy preparation procedure from acidic aqueous solutions, relatively cheap monomer, solubility in many organic solvents and processability with commonly used bulk polymers such as polyethylene, polypropylene, poly(methyl methacrylate) and polyvinylchloride (PVC) [1–6]. There are numerous reports in the literature of different kinds of blends of polyaniline and nonconducting polymers, as well as on electrochemically prepared ÔpureÕ PANI [7,8]. In our group, PANI has been applied mainly in different potentiometric all-solid-state ion-selective (ISE) electrode configurations without an

*

Corresponding author. Tel.: +358-2215-4422; fax: +358-2215-4479. E-mail address: tom.lindfors@abo.fi (T. Lindfors).

0022-0728/$ - see front matter Ó 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.jelechem.2003.07.020

inner filling solution (solid contact electrodes). PANI has been mixed with plasticized PVC in the preparation of Ca2þ -selective electrodes (single-piece configuration) [9,10] or it has been used as such as a membrane matrix [11–14]. We have also recently studied the pH sensitivity of electrochemically prepared PANI and different ringand N -substituted polyanilines [7,8]. Poly(N -alkylanilines) were not pH-sensitive, which is a useful property in sensor applications [8], whereas the strong pH sensitivity of unsubstituted PANI – the only CP that is ion-selective by itself – may limit its use in many applications. The pH sensitivity of PANI has been utilized in different types of PANI-based potentiometric [15–18] and optical [19–24] pH sensors. PANI can easily be made soluble in many common organic solvents with sulfonic acids, e.g., camphorsulfonic acid (CSA) [1], or different phosphoric acid derivatives, e.g., bis(2-ethylhexyl)phosphoric acid [11,12], diphenylphosphoric acid [25] or bis[4-(1,1,3,3-tetrame-

70

T. Lindfors et al. / Journal of Electroanalytical Chemistry 560 (2003) 69–78

thylbutyl)phenyl]phosphoric acid (DTMBP-PO4 H) [13]. This is a definite advantage compared to the electrochemical preparation procedure because membranes containing different additives can easily be tailored. It is particularly important when the influence of different additives on the oxidation state and pH sensitivity of PANI are studied. It is also well-known that PANI has a more complicated electrochemistry than most of the other CPs (Scheme 1) due to its pH sensitivity, which is described by the reversible emeraldine salt (ES)–emeraldine base (EB) transition (acid–base reaction). It is generally accepted that PANI has three oxidation states (abbreviations of their base forms are given in the parentheses): the fully reduced leucoemeraldine (LEB), half-oxidized EB and fully oxidized pernigraniline (PNB) form. The ES form, which is the only electrically conducting form of PANI, can be obtained either by protonation of EB or by oxidation of LEB. The complex electrochemistry of PANI is not only challenging but opens also many possibilities from the basic research and application point of view. The present study focuses on the pH and redox sensitivity of plasticized PVC membranes containing different amounts of PANI as the pH-sensitive component. Plasticized PVC is commonly used as a membrane matrix in ISEs and was therefore chosen as the membrane material for the present study. The influence of lipophilic additives, tridodecylmethylammonium chloride (TDMACl) or potassium tetrakis(4-chlorophenyl)borate (KTpClPB), on the membrane properties will be studied. PANI was dissolved with phosphoric acid dihexadecyl ester (DHDP, Scheme 2) in THF and mixed with plasticized PVC. All pH measurements were made in buffer solutions with pH 2–9 containing rather high concentration of background electrolyte consisting of

Scheme 2. The chemical structure of phosphoric acid dihexadecyl ester (DHDP).

citric acid, tris(hydroxymethyl)aminomethane (Tris), KCl, KH2 PO4 and NaB4 O7 (see Section 2.1). The techniques used in this study were mainly potentiometry, UV–vis spectroscopy and energy dispersive X-ray analysis (EDXA).

2. Experimental 2.1. Chemicals EB powder of PANI was obtained from AC&T (Applications, Chemistry & Technologies, St. Egreve, France). High molar mass PVC and the plasticizer bis(2ethylhexyl)sebacate (DOS), TDMACl and KTpClPB were obtained from Fluka, DHDP from Tokyo Kasei. THF was purchased from Labscan Ltd (Dublin, Ireland). The buffer solutions used in both potentiometric and UV–vis measurements were prepared according to Perrin and Dempsey [26]. These solutions consisted of 25 mM citric acid (monohydrate) ( P 99.5%), 25 mM Tris (p.a. P 99.8%) and 25 mM KCl (p.a. P 99.5%) – obtained from Fluka – and 25 mM KH2 PO4 and 25 mM NaB4 O7
10 H2 O (Merck). The pH of the buffer solution was adjusted to the desired pH value either with HCl or NaOH. Separate buffer solutions for each pH were prepared covering the pH range from 2 to 9. All aqueous

Scheme 1. The redox mechanism between the LEB, ES and PNB forms of PANI (in acidic solutions). The pH dependent transition between the ES and EB forms is also shown in the scheme.

T. Lindfors et al. / Journal of Electroanalytical Chemistry 560 (2003) 69–78

solutions were prepared from distilled deionized water with resistance P18.2 MX. 2.2. Protonation of polyaniline Soluble and electrically conducting PANI was obtained by adding 100 mg of the EB powder of PANI and 300 mg DHDP to 6 ml THF. The molar ratio of DHDP to the PANI repeat unit was 0.50. The colour of the solution slowly became bluish green after the addition of PANI, turning slowly to dark green, due to protonation of EB to ES (Scheme 1). The solution was allowed to equilibriate for three days and the insoluble fraction of PANI was then removed by filtration with a blue ribbon filter paper (5893 , Schleicher & Schuell). The soluble fraction of PANI was then mixed with the other membrane components (PVC, DOS, TDMACl or KTpClPB) and was used for electrode preparation. The mass ratio between PVC:DOS was always 1:2. All membrane components were dissolved in THF. 2.3. Electrode preparation Glassy carbon (GC) disk electrodes (A ¼ 0:07 cm2 ) with a polytetrafluoroethylene (PTFE) body were used as electrode substrates. They were polished with 0.3 lm Al2 O3 powder and rinsed with deionized water prior to the electrode fabrication. Electrode membranes were then prepared by applying 2  15 ll of the membrane solution, containing all membrane components, on the top of the upside down standing GC electrodes. The PANI content in the membranes varied from 0% to 100% (m/m) and the TDMACl and KTpClPB content from 0% to 40% (m/m). THF was allowed to evaporate for 4 h from the membranes before they were conditioned in 0.1 M HCl overnight prior to the potentiometric measurements. The thickness of the electrode membranes, measured with a micrometer screw, was approximately 5 lm. 2.4. Potentiometric measurements 2.4.1. pH sensitivity The polymer electrodes were used as indicating electrodes and a saturated calomel electrode (SCE) as the reference electrode in all potentiometric measurements. The pH sensitivity of three identically prepared polymer electrodes was always measured in the buffer solutions described earlier. The potential readings were taken after 15 min (10 min stirring) in quiescent solutions in order to obtain stable potential readings. The electrodes were calibrated from pH 2–9 and then back to pH 2. Although the response times of the different electrode types were not analysed in detail, most of the electrode types reached stable potential readings within

71

1–2 min. The mean value of the potential readings at the different pH values taken after 15 min were used to calculate the slope of the calibration plot. All potentiometric measurements were performed with a computer controlled homemade multichannel potentiometer. 2.4.2. Redox sensitivity The redox sensitivity of three identically prepared polymer electrodes was measured in a solution containing 10 mM K3 Fe(CN)6 /K4 Fe(CN)6 redox couple, without any background electrolyte or with the buffer solution of pH 6 as the ionic background. The ratio of [Fe(III)]/[Fe(II)] was varied from 0.1 to 10. The potential readings were taken after 30 min with continuous stirring. 2.5. UV–vis spectroscopy 2.5.1. pH sensitivity The polymer films were prepared by drop casting (2  10 ll) on 4 mm thick quartz glasses covered with a thin layer of tin oxide (TO). THF was allowed to evaporate for 4 h from the membranes before they were conditioned in 0.1 M HCl overnight prior to the potentiometric measurements. The membrane thicknesses were not measured but were estimated to be slightly thinner than for the membranes used in the potentiometric measurements (5 lm). The UV–vis transmission spectra of the polymer films were then measured in different buffer solutions with pH 2–9. The pH sensitivity was first measured from pH 2 to 9 and then back to pH 2. The UV–vis spectra at each pH were recorded with a Hitachi U-2001 spectrophotometer at 2 min intervals over 16 min. A syringe was used to fill the cell manually with 1400 ll buffer solution and to remove the previous solution from the cell. 2.5.2. Redox sensitivity The polymer films were prepared and pre-treated before the UV–vis measurements as described above. The composition of the redox solutions was the same as in the potentiometric redox measurements and the UV– vis spectra were measured at 2 min intervals over 16 min. No background electrolyte was used in these measurements. A background spectrum was always recorded for each redox composition before measuring the UV–vis spectra of the polymer film. All experiments in this study were performed at 23  1 °C. 2.6. EDXA measurements The EDXA measurements were performed with the Thermo-Noran Vantage X-Ray Microanalysis System instrument (accelerating voltage: 15 kV).

72

T. Lindfors et al. / Journal of Electroanalytical Chemistry 560 (2003) 69–78

3. Results and discussion 3.1. pH sensitivity Calibration plots for PANI–PVC membranes containing 0 (only plasticized PVC), 10%, 50% and 100% PANI are shown in Fig. 1. It was observed that the pH sensitivity was strongly suppressed when the PANI content in the membrane was less than 40%. The PANI– PVC membrane with 10% PANI, however, still shows pronounced pH sensitivity between pH 2 and 6, but only minor pH sensitivity is observed for pure plasticized PVC between pH 2 and 4. The pH sensitivity was not measured beyond pH 9 because of the strong deprotonation of PANI, which results in hysteresis in the electrode response at high pH values [7]. The results of the potentiometric measurements indicate that plasticized PVC has a favourable influence on the hysteresis effect of PANI. Membranes consisting of pure PANI, curve (a) in Fig. 1, show rather strong hysteresis. The membranes containing 50% plasticized PVC, however, show less hysteresis (curve (b)). The pH sensitivity is simultaneously slightly improved. The reason for the lower hysteresis may be that PANI changes its chain conformation under the influence of plasticized PVC, which can be considered as a secondary dopant, from a compact coil to an extended coil structure [27]. Plasticized PVC probably also functions as a plasticizer for PANI, which increases the flexibility of the PANI chains, thus also facilitating the EB–ES transition in the electrode membrane. Travers and coworkers have observed that the addition of an external plasticizer (triphenyl and

Fig. 1. pH sensitivity of PANI–PVC membranes containing (a) 100%, (b) 50%, (c) 10% and (d) 0% PANI. The calibration plot for an electrochemically prepared PANI film (e) is also given in the figure.

tritolyl phosphate) to solution processed PANI films leads to dramatic changes in their transport properties [28]. In our study, the polymer films containing 50–70% PANI showed the lowest hysteresis of all electrodes studied and the hysteresis decreases, compared to the pure PANI membrane, up to an addition of 60% plasticized PVC in the polymer membrane. For comparison, the calibration plot of PANI(Cl), which was electrochemically polymerised in 1.0 M HCl, is also given in Fig. 1 [7]. The slopes of the calibration plots of PANI(Cl), curve (e), and the PVC–PANI membrane containing 50% PANI, curve (b), are 62.4  0.9 and 52.7  1.1 mV/pH, respectively. The reproducibility of both calibration plots was good. The difference in the slopes indicate that the ES–EB transition of the PANI–PVC membrane is slightly hindered by the bulky DHDP anion, which is trapped in the polymer structure. The smaller and more mobile Cl ions in electrochemically polymerised PANI can, on the other hand, penetrate the polymer structure much more easily. Polymer membranes consisting of plasticized PVC and PANI dissolved with CSA in chloroform, were also prepared but the membranes were not stable due to the water solubility of CSA. The pure PANI dissolved with CSA showed strong pH sensitivity with a super-Nernstian slope of approximately 80 mV/pH when calibrating the electrodes from pH 9 to 2, which is in good accordance with the results obtained by Karyakin et al. [18]. An almost Nernstian slope of 58.0  0.8 mV/pH was obtained when we performed the calibration from pH 2 to 9, although the calibration curve was not completely linear at low pH values. We have previously observed that the addition of TDMACl facilitates the oxidation and reduction processes of PANI dissolved with the Ca2þ -selective organophosphate DTMBP-PO4 H [14]. In order to fulfil the electroneutrality condition and to compensate for the negative charge of the anions (DTMBP-PO 4 ) of the protonic acid during the ES–EB transition of PANI, cations other than Hþ must be present in the PANI membrane at higher pH values. We have therefore studied the influence of the lipophilic additives TDMACl and KTpClPB on the pH sensitivity of the PANI–PVC membrane containing equal amounts of PANI and plasticized PVC (50%). The lipophilic additive (0–40%, m/m) was added to the membranes while keeping the mass ratio between PANI and plasticized PVC constant. Calibration plots of polymer membranes containing 0%, 20% and 40% TDMACl or KTpClPB are shown in Fig. 2. As can be seen in the figure addition of TDMACl decreases the pH sensitivity and increases the hysteresis of the polymer membranes. The pH sensitivity of the membrane with 20% TDMACl is rather good, although the hysteresis is quite pronounced. The performance of the membrane with 40% TDMACl is, however, rather poor.

T. Lindfors et al. / Journal of Electroanalytical Chemistry 560 (2003) 69–78

73

Fig. 2. pH sensitivity of PANI–PVC membranes containing (a) 0%, (b) 20% and (c) 40% TDMACl or KTpClPB.

The UV–vis spectra of the PANI–PVC membranes with 0% and 40% TDMACl are shown in Fig. 3. The contribution of plasticized PVC to the absorbance spectrum in Fig. 3 is negligible ( 750 nm [30,31]. A small maximum can be seen in the PANI– PVC membrane (0% TDMACl) at 420 nm even at high pH values indicating that a minor fraction of the material is still in the ES form. The maximum at 800–850 nm indicates that the conjugation length in PANI is short with localised charge distribution. EB has a characteristic maximum at 630–650 nm [32]. The absence of any absorbance maximum at 420 nm for the membrane with 40% TDMACl at high pH values indicates that TDMACl facilitates the ES–EB transition in PANI by compensating for the charge imbalance. The UV–vis measurements also confirm that the pH dependent ES–EB transition takes place between pH 6 and 9, even though the potentiometric measurements indicate that the membrane with 40% TDMACl is only very slightly pH sensitive in this pH range. This shows the advantage of using both potentiometry and UV–vis spectroscopy in the characterisation of PANI membranes. Based on the potentiometric and UV–vis measurements, the potentiometric response of the PANI–PVC– TDMACl membranes can be explained to be the result of two competing processes – Hþ and Cl -sensitivity of the membrane. The pH and Cl sensitivities originate from PANI and TDMACl, respectively. TDMACl is

known to be sensitive to Cl ions and is used as the active material in Cl -selective electrodes although its selectivity towards Cl is rather poor [33]. Because the buffer solutions contain 25 mM KCl the Cl -sensitivity becomes superior to the Hþ -sensitivity at high pH (>6), especially when the high concentration of TDMACl in the membrane is taken into consideration. This conclusion is further supported by the very high Cl -sensitivity (50–60 mV/decade) that was observed for the PANI–PVC films containing 20% and 40% TDMACl when they were calibrated in pure NaCl-solutions from 105 (pH ¼ 5.5) to 101 M (pH ¼ 4.9). The films with KTpClPB, which is a cation exchanger, showed only minor Naþ -sensitivity in the same solutions. Very interesting results were obtained with the PANI–PVC membranes containing KTpClPB. Almost no potentiometric pH sensitivity was obtained for membranes with 20% and 40% KTpClPB (Fig. 2) and the UV–vis measurements showed that the membranes were still in the electrically conducting and protonated ES form even at pH 9, which is surprising (Fig. 3). The characteristic ES–EB transition is almost absent – although a minor decrease in the absorbance maximum of the ES form at 420 nm can be observed at pH > 6. The green colour of the polymer membrane at pH 9 confirms that PANI is mostly in the electrically conducting ES form (the EB form has a deep blue colour). The same behaviour was also observed when the potentiometric pH sensitivity of the PANI–PVC–KTpClPB membranes were measured by changing the pH from 2 to 9 starting with a 0.01 M HCl (pH 2) solution that was titrated with NaOH. The solution pH was continuously monitored with a glass pH electrode during the titration. This

74

T. Lindfors et al. / Journal of Electroanalytical Chemistry 560 (2003) 69–78

Fig. 3. The UV–vis spectra of PANI–PVC membranes containing 0% and 40% TDMACl or KTpClPB. The UV–vis spectra were measured from pH 2 to 9 (a) and then back to 2 (b).

experiment was conducted in order to find out if ions in the buffer solution influence the pH sensitivity of PANI. Only minor pH sensitivity was observed between pH 2 and 4 in the titrimetric experiment. The exact reason for the high stability of the ES form in the presence of KTpClPB in the electrode membrane is not completely clear at the moment. We believe, however, that one possible explanation is that the TpClPB and the DHDP anions repel each other in the membrane and the electrical force that arises from this interaction hinders the DHDP anions in leaving the PANI chains when the pH increases. This hinders or slows down the ES–EB transition. The possibility of affecting the ES–EB transition and the electrical conductivity of PANI is advantageous in many applications. The influence of

different types of anions on the pH sensitivity of PANI is currently being studied in our laboratory and will be reported later. 3.2. Redox sensitivity The redox sensitivity of the polymer membranes containing TDMACl or KTpClPB is shown in Fig. 4. As can be seen in the figure, no redox sensitivity was observed for the PANI membranes containing 20% and 4 40% TDMACl in the Fe(CN)3 6 /Fe(CN)6 redox solutions (total concentration: 10 mM), which is in accordance with our earlier observations with pure PANI membranes containing TDMACl [12,13]. The same behaviour was observed also when 10 mM of the FeCl3 /

T. Lindfors et al. / Journal of Electroanalytical Chemistry 560 (2003) 69–78

75

Fig. 4. Redox sensitivity of PANI–PVC membranes containing (a) 0%, (b) 20% and (c) 40% TDMACl or KTpClPB. The total concentration of the redox couple is 10 mM.

FeCl2 redox couple was used instead of the Fe(CN)3 6 / Fe(CN)4 couple in the redox measurements and a pH 6 6 buffer solution was used as the background electrolyte. All PANI–PVC membranes with KTpClPB are redox sensitive, which clearly shows that PANI in these membranes is in the conducting ES form. The slopes of the potentiometric plots shown in Fig. 4 were 57.8  0.1 (0%), 57.8  0.4 (20% KTpClPB) and 55.8  1.5 mV/ decade (40% KTpClPB). The influence of the redox sensitivity of PANI on the electrode potential is much stronger than the pH sensitivity. The electrode potential is therefore determined by the redox couple and not by the solution pH. This was confirmed in a separate experiment (not shown here) where the redox sensitivity of the PANI–PVC film (0% TDMACl) was measured in a pH 6 buffer solution as the background electrolyte. In this experiment, the potentiometric slope of the calibration plot measured in the 4 Fe(CN)3 6 /Fe(CN)6 redox solutions was 58.5  0.2 mV/ 4 decade. The pH of the Fe(CN)3 6 /Fe(CN)6 redox solutions that were used in the redox measurements shown in Fig. 4 was 5.2–5.5 and the results of these measurements reflect therefore only the redox sensitivity of PANI. The effect of pH can be neglected. It should be pointed out that the redox sensitivity of PANI is characterised by the LEB–ES–PNB transition in acidic solutions (LEB–PNB at neutral pH) whereas the pH sensitivity is characterised by the ES–EB transition (Scheme 1). The UV–vis spectra of PANI–PVC–TDMACl mem4 branes measured in Fe(CN)3 redox solu6 /Fe(CN)6 3 tions, with the same ratios of [Fe(CN)6 ]/[Fe(CN)4 6 ]¼

1/10 ) 10/1 as in Fig. 4, are shown in Fig. 5. As can be seen in the figure, PANI is in the electrically conducting ES form in all membranes, i.e., showing an absorbance maximum at 800–850 nm. The change in the oxidation level of PANI, induced by the redox solutions during the measurements, can be seen in the spectra. This is especially well observed for the membrane containing only PANI and plasticized PVC (0% TDMACl). The absorbance maximum at 845 nm shifts to 805 nm when the 4 ratio of [Fe(CN)3 6 ]/[Fe(CN)6 ] is increased from 1/10 to 10/1 indicating that the oxidation level of PANI increases, which is in good accordance with the results of the potentiometric measurements (Fig. 4). Similar changes in the oxidation level can also be seen in the membranes with 20% and 40% TDMACl. The UV–vis spectrum of pure PANI has been reported to change in a similar way when the potential (oxidation level) of a PANI film electrode was increased [34]. The shift of the absorbance maximum at 845 nm to lower wavelengths (Fig. 5) may indicate that shorter conjugated PANI chain segments are oxidized and that the average conjugation length decreases [35]. It was observed after the redox measurements of the membranes containing 20% and 40% TDMACl, that a thin layer of the membrane surfaces, which had been in 4 contact with the Fe(CN)3 solutions, was 6 /Fe(CN)6 slightly yellowish. The membranes with 40% TDMACl were more yellowish than membranes with 20% TDMACl. The same observation was made with plasticized PVC membranes containing 20% and 40% TDMACl, but no PANI. The yellowish colour of the electrode membranes originates obviously from the

76

T. Lindfors et al. / Journal of Electroanalytical Chemistry 560 (2003) 69–78

Fig. 5. The UV–vis spectra of PANI–PVC membranes containing 0%, 20% and 40% TDMACl, which were measured during the redox measurements 4 3 4 with the Fe(CN)3 6 /Fe(CN)6 redox couple when the ratio of [Fe(CN)6 ]/[Fe(CN)6 ] was increased from 1/10 to 10/1. 4 Fe(CN)3 6 /Fe(CN)6 couple, which is extracted into the membrane phase. The redox couple may replace the Cl ions in the membrane (probably in the vicinity of the membrane–solution interface), which was confirmed with EDXA measurements. The EDXA spectra of the PANI–PVC film containing 40% TDMACl are shown in Fig. 6. After the preparation of the film (1) it was conditioned in 0.1 M HCl (2) prior to the redox measure4 ments in a 10 mM solution of the Fe(CN)3 6 /Fe(CN)6 redox couple (3). It can be seen in the spectra in Fig. 6 that only minor changes occur in the polymer membrane during conditioning in 0.1 M HCl. Fe, however, can be detected in the membrane at 6.4 keV after the redox

Fig. 6. EDXA spectra of the PANI–PVC membrane containing 40% TDMACl, which were measured after: (1) the preparation of the film, (2) conditioning in 0.1 M HCl and (3) redox measurements with the 4 Fe(CN)3 6 /Fe(CN)6 couple.

measurement. No Fe-signal could be observed in the membranes containing KTpClPB after redox measurements. The Cl-signal has also decreased in intensity indicating that the concentration of Cl ions in the membrane phase has decreased, obviously being ex4 changed to Fe(CN)3 6 and Fe(CN)6 ions. The reason for the increasing C-signal is not fully understood because of the complexity of the membrane matrix, but it 4 can be expected that the Fe(CN)3 couple 6 /Fe(CN)6 slightly increases the C-content of the membrane (nitrogen cannot be detected with EDXA). The nature of the interaction of the redox couple in the polymer films containing TDMACl is still unclear, but probably the lipophilicity of the redox couple facilitates its interaction with the film. It was also observed in cyclic voltammetric measurements that the films with 20% and 40% TDMACl almost completely lost their electroactivity after the redox measurements in contrast to the film without TDMACl, which showed improved electroactivity after the redox measurements. The strong Cl interaction of the TDMACl containing membranes might, on the other hand, explain why no redox sensitivity can be observed in redox measurements with the FeCl3 /FeCl2 redox couple. In this case, the high concentration of Cl in the measuring solution determines the membrane potential [36]. 4 The interaction of the Fe(CN)3 redox 6 /Fe(CN)6 couple with plasticized PVC membranes containing 0%, 20% and 40% TDMACl (no PANI) was also studied by measuring the time-dependent potential profiles of the membranes in a redox solution with the ratio 4 [Fe(CN)3 6 ]/[Fe(CN)6 ] ¼ 1/10 (Fig. 7). The membranes were conditioned overnight in 0.1 M HCl prior to the measurements. The potential profiles in Fig. 7 show that the film without TDMACl (0%) has a rather stable potential whereas the potentials of the membranes containing TDMACl drift continuously towards lower potentials during the time scale of the experiment (30 min).

T. Lindfors et al. / Journal of Electroanalytical Chemistry 560 (2003) 69–78

77

tion phase is established more quickly in membranes with higher TDMACl content. EDXA spectra of the PANI–PVC membranes with 40% TDMACl after being in contact with 10 mM Fe(CN)4 6 (spectrum 2) and 10 mM Fe(CN)3 6 (spectrum 3) in two different experiments are shown in Fig. 8. In both cases the membranes were conditioned in 0.1 M HCl overnight and then for 2.5 h in the iron containing solutions. The spectra in Fig. 8 show that Fe(CN)3 6 ions are extracted to a larger extent into the membrane than Fe(CN)4 6 ions. The EDXA measurements support the interpretation that the yellowish colour of the membrane surface originates from intercalation of the strongly yellow coloured Fe(CN)3 6 ions into the membrane.

4. Conclusions

Fig. 7. Time dependent potential profiles of plasticized PVC membranes containing (a) 0%, (b) 20% and (c) 40% TDMACl during the 4 redox measurements; [Fe(CN)3 6 ]/[Fe(CN)6 ] ¼ 1/10. 3=4

EDXA measurements confirmed that no Fe(CN)6 was taken up by the membrane with 0% TDMACl, but a minor Fe-signal was observed at 6.4 keV for the membrane containing 40% TDMACl. This indicates that the membranes with TDMACl interact with the 4 Fe(CN)3 6 /Fe(CN)6 redox couple. It is also obvious by comparing the potential profiles b and c in Fig. 7 that the chemical equilibrium between the membrane and solu-

Fig. 8. EDXA spectra of the PANI–PVC membrane containing 40% TDMACl, which were measured after: (1) the preparation of the film, (2) conditioning in 10 mM K4 Fe(CN)6 and (3) in 10 mM K3 Fe(CN)6 for 2.5 h. The films were conditioned overnight in 0.1 M HCl prior to the conditioning in the K3 Fe(CN)6 and K4 Fe(CN)6 solutions.

It is shown that addition of 20–60% (m/m) plasticized PVC to PANI membranes decreases the hysteresis of the electrode response during potentiometric pH measurements and improves the mechanical strength and elasticity of the membranes. The best pH sensitivity (52.7  1.1 mV/pH) was obtained with membranes consisting of 50% PANI and 50% plasticized PVC. Addition of TDMACl to these membranes increased the hysteresis probably due to the Cl -sensitivity of TDMACl, whereas membranes with 20% and 40% KTpClPB showed almost no potentiometric pH sensitivity, which was confirmed with UV–vis measurements. The reason for this behaviour is not yet fully understood, but one probable explanation may be that the negatively charged TpClPB and the DHDP ions repel each other in the electrode membrane and thus hinder the occurrence of the ES–EB transition. UV–vis measurements confirm, however, that membranes containing KTpClPB were still in the electrically conducting ES form at pH 9. The effect of negatively charged additives on the ES–EB transition of PANI is currently being studied in our laboratory. No potentiometric redox sensitivity was observed for the PANI–PVC membranes containing P 20% 4 TDMACl. The reason is that the Fe(CN)3 6 /Fe(CN)6 redox couple is extracted into the membrane phase during the redox measurements, which was shown with EDXA measurements. It was also shown that Fe(CN)3 6 ions are extracted to larger extent into the membrane than Fe(CN)4 6 ions. The exact nature of the interaction of the redox couple in the membrane phase is currently unknown.

Acknowledgements
bo Akademi This work is part of the activities of the A Process Chemistry Group within the Finnish Centre of

78

T. Lindfors et al. / Journal of Electroanalytical Chemistry 560 (2003) 69–78

Excellence Programme (2000–2005) sponsored by the Academy of Finland. The authors gratefully acknowledge Mr. Clifford Ekholm who did the EDXA-measurements. We also thank Dr. Carita Kvarnstr€ om and Dr. Johan Bobacka for fruitful discussions. References [1] Y. Cao, P. Smith, A.J. Heeger, Synth. Met. 48 (1992) 91. [2] C.Y. Yang, M. Reghu, A.J. Heeger, Y. Cao, Synth. Met. 79 (1996) 27. [3] M. Zilberman, G.I. Titelmann, A. Siegmann, Y. Hara, M. Narkis, D. Alperstein, J. Appl. Polym. Sci. 66 (1997) 243. [4] U.A. Sevil, O. G€ uven, S. S€ uzer, J. Phys. Chem. B 102 (1998) 3902. [5] S. S€ uzer, O. Birer, U.A. Sevil, O. G€ uven, Monatshefte f€ ur Chemie 132 (2001) 185. € sterholm, P. Smith, A.J. Heeger, J. Laska, M. [6] A. Pron, J.-E. O Zagorska, Synth. Met. 57 (1993) 3520. [7] T. Lindfors, A. Ivaska, J. Electroanal. Chem. 531 (2002) 43. [8] T. Lindfors, A. Ivaska, J. Electroanal. Chem. 535 (2002) 65. [9] J. Bobacka, T. Lindfors, M. McCarrick, A. Ivaska, A. Lewenstam, Anal. Chem. 67 (1995) 3819. [10] T. Lindfors, P. Sj€ oberg, J. Bobacka, A. Lewenstam, A. Ivaska, Anal. Chim. Acta 385 (1999) 163. [11] T. Lindfors, A. Ivaska, Anal. Chim. Acta 404 (2000) 101. [12] T. Lindfors, A. Ivaska, Anal. Chim. Acta 404 (2000) 111. [13] T. Lindfors, A. Ivaska, Anal. Chim. Acta 437 (2001) 171. [14] T. Lindfors, C. Kvarnstr€ om, A. Ivaska, J. Electroanal. Chem. 518 (2002) 131. [15] A.A. Karyakin, O.A. Bobrova, L.V. Lukachova, E.E. Karyakina, Sensors Actuat. B 33 (1996) 34. [16] C.A. Lindino, L.O.S. Bulh~ oes, Anal. Chim. Acta 334 (1996) 317. [17] G. Cui, J.S. Lee, S.J. Kim, H. Nam, G.S. Cha, H.D. Kim, Analyst 123 (1998) 1855.

[18] A.A. Karyakin, M. Vuki, L.V. Lukachova, E.E. Karyakina, A.V. Orlov, G.P. Karpachova, J. Wang, Anal. Chem. 71 (1999) 2534. [19] Z. Ge, C.W. Brown, L. Sun, S.C. Yang, Anal. Chem. 65 (1993) 2335. [20] E. Pringsheim, E. Terpetschnig, O.S. Wolfbeis, Anal. Chim. Acta 357 (1997) 247. [21] U.-W. Grummt, A. Pr on, M. Zagorska, S. Lefrant, Anal. Chim. Acta 357 (1997) 253. [22] S. de Marcos, C. Asensio, I. Uru~ nuela, F. Gallarta, J. Galban, J.R. Castillo, Quımica Analıtica 19 (Suppl. 1) (2000) 99. [23] Z. Jin, Y. Su, Y. Duan, Sensors Actuat. B 71 (2000) 118. [24] P.C. Pandley, G. Singh, Talanta 55 (2001) 773. [25] C.-T. Kuo, C.-H. Chen, Synth. Met. 99 (1999) 163. [26] D.D. Perrin, B. Dempsey, Buffers for pH and Metal Ion Control, Chapman & Hall, London, 1974, p. 48. [27] A.G. MacDiarmid, A.J. Epstein, Synth. Met. 65 (1994) 103. [28] P. Fedorko, B. Dufour, P. Rannou, M. Trznadel, J.P. Travers, International Conference on Science and Technology of Synthetic Metals (ICSM 2002, Shanghai), Oral presentation, Book of Abstracts, p. 231 (FriB2). [29] J. Workman Jr., Handbook of Organic Compounds, Academic Press, New York, 2001, p. 239. [30] W.S. Huang, A.G. MacDiarmid, Polymer 34 (1993) 1833. [31] D. Chinn, J. DuBow, J. Li, J. Janata, M. Josowicz, Chem. Mater. 7 (1995) 1510. [32] A. Malinauskas, R. Holze, J. Solid State Electrochem. 3 (1999) 429. [33] D. Wegmann, H. Weiss, D. Ammann, W.E. Morf, E. Pretsch, K. Sugihara, W. Simon, Mikrochim. Acta III (1984) 1. [34] S. Quillard, K. Berrada, G. Louarn, S. Lefrant, M. Lapkowski, A. Pr on, New J. Chem. 19 (1995) 365. [35] H.S. Nalwa, Handbook of Advanced Electronic and Photonic Devices, Vol. 8, Academic Press, San Diego, 2001, p. 233. [36] J. Bobacka, Z. Gao, A. Ivaska, A. Lewenstam, J. Electroanal. Chem. 368 (1994) 33.