- Email: [email protected]
Separation of faradaic and capacitive current regions in the redox transformation of poly(3-methylthiophene) with the exclusion of overoxidation processes
Acta, Vol. 42, No. 4, pp. 651457, 1997 Copyright 0 1996 ElsevierScience Ltd. Printed in Great Britain. All rights reserved 00134686/97 $17.00 + 0.00
Elecfrochimica
Pergamon PII: s0013-46%(%)00210-1
Separation of faradaic and capacitive current regions in the redox transformation of poly(3-methylthiophene) with the ex6lusion of overoxidation processes Cs. Visy,* M. Lakatos, A. Sziics and M. Novak Institute of Physical Chemistry, Jdzsef Attila University,
(Received 5 February
H-6701 Szeged, P.O. Box 105, Hungary
1996; in revised form 23 May 1996)
Abstract-The electrostability of the poly(3-methylthiophene) film prepared in nitrobenzene provided the opportunity to study its redox transformations in an extended anodic potential region with the exclusion of the overoxidation. From spectroelectrochemical results a capacitive region could be separated in the highly oxidized state of the layer which exhibited a large capacitance. This pattern-assumed to be related to charge separation/redistribution within the film at high potential field connected to solvent movements-is promising from the point of view of symmetric p-type/n-type superconductors. Copyright 0 1996 Elsevier Science Ltd Key words: Electronically capacitive current.
conducting
polymer, poly(3-methylthiophene),
INTRODUCTION Very soon after the discovery that electronically conducting polymeric !ilms can be repeatedly driven between their oxidized (doped, conducting) and neutral (reduced, undoped, insulating) forms (11, an interesting characteristic of the cyclic voltammograms was also revealed: at potentials beyond the anodic peak the current tends towards a constant, non zero value which is a function of the sweep rate [2]. This part of the current behaves more or less as a capacitive type, namely that the charge us the potential function in this region is linear. As an explanation, Feldberg proposed to distinguish two processes during the redox transformation of the polymeric films [3], a faradaic type in which the charge carriers are generated in a reversible electron transfer step, and a capacitive type connected to the movement of the carriers to the interface, a charge accumulation without creating new carriers. Since that time it has been shown that the transformation of the neutral film is not a single charge transfer process, but the problem of the faradaic and the capacitive current is not yet solved and returns from time to time in the literature [4-l 31, *Author to whom correspondence
should be addressed.
redox transformation,
faradaic/
partly because of the overlapping overoxidation processes, partly because of the inseparability of the current components in pure electrical measurements. On the basis of complex impedance measurements analyzed through equivalent circuits, the distinction between the two types of charge is connected to the doping counter ion. According to Albery [6] the interaction energy between the charge of the chain and the counter ion can be different, leading to strongly and loosely bounded ions. As another explanation two different diffusion rates are assumed for the doping ion, depending on whether the ions move into more closed aggregates or only in larger pores around the aggregates [8]. In contrast, Heinze suggested the assignment of the current in the critical region to subsequent redox processes connected to a series of polymeric segments of different length [5]. This theory, considering the current at least partially faradaic, was also supported in [9]. From chronopotentiometric results obtained with end-capped thiophene oligomers, Zotti opposed the previous conclusion [ 121and favoured the “molecular metal” model by Feldberg. In a recent work the capacitive pattern of the majority of the current even at the beginning of the oxidation of polypyrrole and polybithiophene was concluded based on
651
652
Cs. Visy et al.
electrochemical impedance spectroscopic measurements [13]. The current generally assigned to the doping process is assumed to create a capacitance and to behave as a capacitive current. However, the possibility of parallel charge transfer processes such as over-oxidation cannot be excluded which-as in other cases-limit the validity of the model applied in the simplified equivalent electronic circuit. Overoxidation is the summarizing name of the processes in the more positive potential region which cause the irreversible deactivation of the polymeric film. The oxidized segments of the chain react with nucleophiles in the solution, which leads to the partial or total degradation of the film. This should be distinguished from the temporary deactivation [14] caused by ion-pair formation in the film, which can be resolved at least partially by more positive polarization. Although in some cases the overoxidation influences positively the properties of the film [15] eg the increased permeability and the enhanced cation exchanger pattern of polypyrrole, attention should be paid to the solution composition to prevent unwanted degradation processes [ 161. The aim of the present work was to prepare a polymeric film which is stable at positive potentials, this provides an opportunity to study the character of the current in the potential region beyond the current peak without interference from overoxidative degradation processes. It was recently demonstrated that very stable and well-dopable polymeric film can be obtained by the electrochemical polymerization of thiophene type monomers in nitrobenzene solution [17]. From the point of view of the anodic stability, the 3-substituted derivatives of thiophene are more favourable. On this basis 3-methylthiophene (MT) was chosen as a model, because it proved to be the most suitable among the alkyl-substituted analogues [181. In order to follow the faradaic type processes during the redox transformation of the polymeric film (PMT), in situ spectroelectrochemistry seems to be applicable: the generation of the charge carriers coupled with ionic movements can be followed from the spectral changes [lo], and the region of these processes can be compared to the current and charge measured during the transformation.
MT (Aldrich) was vacuum-distilled, tetrabutylammonium hexafluorophosphate (Bu4NPF6, Aldrich) was dried in vacuum, nitrobenzene (NB, Aldrich) was used as received, acetonitrile (AN, Aldrich) was dried over activated alumina and kept under dried nitrogen atmosphere. The water content of the solution was controlled by coulometric Karl Fischer titration, and it was always kept below 30 ppm. The IT0 glass was treated by oxygen plasma in Harrick plasma generator before use. The cell and solutions were deaerated by dried nitrogen flow. The PMT film was prepared galvanostatically at 2.5 mA/cm2 current density for 50 mC/cm’ total charge from a solution of 0.1 M for both the MT and the BmNPF6 in NB. The film was then reduced, washed several times with acetonitrile to remove the monomer, and the cell was filled up with 0.1 M BudNPFe/AN solution. The redox transformation of the PMT film was studied by using cyclic voltammetric and potential step methods. In the first case the transient current of the voltammetric curve was integrated in different potential regions, and the spectra were taken at the end-potentials of the intervals. In the second type of measurements, the potential was changed stepwise, the chronoamperograms were recorded after each step of 100 mV, and the transferred charges were determined by integration. The stationary spectra belonging to the actual potential and charge value were recorded when the steady-state current had been reached.
RESULTS
AND DISCUSSION
Figure 1 shows the chronopotentiogram during the galvanostatic polymerization of MT in NB solution. The almost constant potential value throughout the process indicates the formation of a well conducting film [17]. The redox pattern of the PMT layer was studied in monomer-free BuNPFs/AN solution. As POTENTIAL/V
5ti
EXPERIMENTAL In situ spectroelectrochemical measurements were carried out in a one compartment cell placed horizontally into the beam-hole of the spectrophotometer. The working electrode was an IT0 glass (ca 20 n/n) with a surface area of 0.28 cm2. A gold and a silver electrode were used as counter and reference, respectively. The cell was controlled by an Electroflex (type 45 1) potentio/galvanostat, the spectra were registered by an HP8452A diode array spectrophotometer. The measurements were synchronized by a PC.
2,
1
0' 0
5
10
15
20
25
TiMEis
Fig. 1. Typical chronopotentiogram obtained during the galvanostatic polymerization of MT in nitrobenzene at i = 2.5 mA/cm* current density.
Separation of faradaic and capacitive current regions
I
I
0
I
500
I
1000
E/mV
Fig. 2. Subsequent cyclic voltammetric curves registered with PMT film in acetonitrile solution in an extended potential region from -0.4 to + 1.5 V.
can be seen in Fig. 2, the oxidation of the film gives a current peak at around 0.60.65 V, which is followed by a rather long plateau, and at 1.3 V a new electrochemical process starts. The subsequent scans reflect the decrease in the charge transferred reversibly during the oxidation-reduction cycles. As the spectral differences before and after the scannings also indicated the irreversible deactivation of the film, the current in the most positive potential region was connected to degradation process. When the potential scan was restricted to below 1.3 V, both the charge values and the spectral changes proved the reversible redox transformation of the polymer. Figure 3 shows the reproduction of the second voltammetric curve after 10 min scanning although the anodic endpoint of the potential scan was 0.65-0.7 V more positive than the potential of the anodic peak.
1
653
The stability of the film was supported also by the spectral measurements (Fig. 4(a) and (b)); the spectrum at the beginning of the cycling is identical with that obtained at the end, after subsequent switchings between the oxidized and reduced form. The unusual shape of the spectra in the 400-500 nm wavelength region is due to the fact that the reference spectrum has been registered before the polymerization in the slightly coloured NB, while the redox transformations were studied in AN. The spectrum belonging to the neutral form of the film has the absorbance maximum at -530 nm, which reflects a highly conjugated polymer [lo, 201. It can also be seen that the spectrum of the highly oxidized form is similar to that reported as exhibiting a quasi-metallic behaviour (IO, 191. On the basis of these results, it was concluded that overoxidation takes place only beyond 1.3 V, so in the potential range below 1.3 V there is no other faradaic process but the so called doping-undoping reaction which can be followed by the spectral changes. In order to see the relation of the faradaic type charge transfer process to the total charge consumed during a cycle, two characteristic absorption values were plotted us the charge determined from the integration of the voltammetric currentpotential curve. To monitor the change in the amount of the neutral and the well oxidized forms of the polymer, the absorbance values at 532 nm and 758 nm were chosen, respectively. Figure 5 clearly shows that most of the spectral changes occur during a ca 600-800 uC charge transfer, so the current below 0.8-0.9 V is mainly faradaic. However, a large amount of charge passes without, or relatively almost without, spectral changes, hence the slope of the curve is almost zero while the second half of the total charge is transferred. This behaviour may indicate a charge accumulation in a capacitor. The character of the absorbance-charge plot is rather similar during the reverse scan (Fig. 6) but in the opposite order. First mainly the “capacitor” seems to be discharged, which is followed by the decrease in the charge carriers. The parallel absorbance increase observed at the beginning of the reduction in both wavelength ranges is further evidence of the modified redox mechanism of thiophene type polymers [2&22] which interpreted this observation by a dissociative reduction of the dication: b2+ + e- = n + p+.
b’
500
1000
E/mV
Fig. 3. Stability of the current-potential curve of PMT from - 0.4 to + 1.3 V potential us Ag electrode in AN: the second sweep (1) and the one after 10 min scanning (2).
This reaction describes well the fact that at the beginning of the reduction of the totally oxidized film the absorption increase in the low wavelength range-indicating the re-formation of neutral species-is not accompanied by a decrease in the mid-gap excitations, in fact an increase can be seen. This maximum-type curve of the absorbance at 758 nm can be explained by the different absorption coefficients of the monocation and the dication forms
654
cs. visy et al
ABSORBANCE IO
8
6
(4
WAVELENGTHhm Fig. 4(a).
[22] both having in this wavelength range the same excitation from the valence band to the non-bonding mid-gap level [19]. The results obtained from the stationary spectral measurements following the potential steps gave further evidence to support the previous observations for the separation of the two types of processes. As Fig. 7 shows, the separation of the charge transfer regions connected to spectral changes and/or to no spectral modifications is still more expressed, although the technique is able to follow the spectral effect of the delayed, slow conformational changes, as well. On this basis we concluded that there is a potential region at least as wide as 1.1-1.3 V, where the current is exclusively capacitive, and in the preceding potential region this capacitive current may flow also simultaneously with the faradaic current. In order to see the separation and/or overlapping of the two types of the current, the charge values were plotted us the potential (Fig. 8). If we assume that the capacitance of the film is independent of the potential, the constant slope obtained in the potential range 1.1-1.3 V-considered as the “net” capacitive region-can be extrapolated to less anodic potentials. From this rough extrapolation it follows that the capacitor has been created from at least 0.75 V potential by the faradaic process, and it becomes
gradually more and more dominant. It is worth mentioning that in the present case the subsequent redox transformation of the differently conjugated segments [5] does not interfer, which is in accordance with the fact that there is no blue (hypsochromic) or red (bathochromic) shift of the absorbance of the neutral form during its redox transformations [21], but we have a rather uniform, well-conjugated film. The capacitance of the layer was determined from the slope of the Q-E data, and a value of ca 14 mF/cm2 was obtained. For the thickness an estimation from the polymerization charge can be applied. Assuming that 5 mC/cm2 gives a 10 nm thick film (231, a thickness of 10es cm and a capacitancedensity of 1400 F/cm3 is obtained [24]. The question is what structure could stabilize such a big amount of charge concentrated in the thin polymeric film. Recently, the significant swelling and contraction of the PMT/LiClOd film was reported during its oxidative and reductive transformation, respectively [25], which was delayed with respect to the charge transfer, and it was interpreted by the movement of the doping ions. Since these kinds of changes in the size of a polymeric film proved to be solvent dependent [26,27], and considering that in the case of Li-salts a cation repulsion may take place, the delayed swelling can be connected rather to solvent
655
Separation of faradaic and capacitive current regions
ABSORBANCE
8
6
4
2
0 400
500
(b).
600
700
800
WAVELENGTHhm
Fig. 4. Spectra registered during the cyclic voltammetric transformation (b) during the reduction section (l-9) of the scan.
penetration. Of course the character of the incorporated ion affects the uptake of the solvent through solvation interactions as concluded in [28]. In contrast, the penetration of solvent molecules has been observed during the undoping process [29].
of the PMT film: (a) during the oxidation (l-l 1);
The capacitance C is connected to the geometric and dielectric properties as C = &A/l where E is the (absolute) permittivity,
A the surface
ABSORBANCE
6
i
6
4
2
0 ”
4 CHARGEimC
Fig. 5. Absorbance changes at 532 nm (solid) and 758 mn (dotted) wavelengths us the charge transferred during the voltammetric oxidation of PMT film.
1
2
3
CHARGEimC
Fig. 6. Absorbance changes at 532 nm (solid) and 758 nm (dotted) wavelengths us the charge transferred during the voltammetric reduction of PMT film.
Cs. Visy et al. haviour of the so-called hyperelectronic polarization of EKA-conjugated substances [30] reported some 25 years ago. It is assumed that in the increasing electric field a charge separation or redistribution occurs within the doped film, which is related to the movement of solvent molecules. The question of whether it is an uptake or expulsion of the solvent molecules is still open and will be the topic of future studies. Taking into account the fact that PMT can be both anodically and cathodically doped [20], this polymeric film is a promising material for the realisation of a type III, symmetric p-type/n-type supercapacitor [31]. 3
2
3
4
CHARGEimC
CONCLUSIONS (b)
7
”
i
Fig. 7. Stationary absorbance IIS charge plots at 532 nm (solid) and 758 nm (dotted) wavelengths registered (a) after positive potential steps; (b) after negative potential steps.
area (0.28 cm2 in this case), I the thickness of the layer. From the previous data a relative permittivity .sr= 1.6 x lo6 is obtained, which recalls the be-
The electropolymerization of MT in NB solution resulted in a well dopable film which proved to be stable in a rather extended positive potential range up to 1.3 V, 0.6547 V more positive than the peak potential in the oxidative curve. With the exclusion of the overoxidation, the reversible redox transformation of the PMT could be studied up to a quasi-metallic character by spectroelectrochemical methods, There were no spectral changes during the oxidation of the film in the 1.1-1.3 V potential interval, from which fact the net capacitive pattern of the current in this region was concluded. In the less anodic potential range the two types of current are overlapping. By assuming a constant capacitance of the film, which is only a rough estimation, the appearance and increasingly dominant pattern of the capacitive current starts from at least 0.75 V. On this basis it is concluded that the large capacitor is created by the charge transfer generating carriers in the film. In parallel with the accumulation of charge-pairs (called doping) in the thin film, solvent movement occurs, which is assumed to facilitate the charge separation/redistribution within the film at high potential field, leading to a large capacitance. On the basis of simple but frequently applied estimations, the big capacitance has been connected to a large relative permittivity in the lo6 order of magnitude. The behaviour of the PMT film is promising from the point of view of the realization of a supercapacitor based on p-type and n-type forms.
ACKNOWLEDGEMENTS
0
10
5
15
POTENTIAL%
Fig. 8. Charge transferred during the positive scan USthe actual potential obtained for the oxidation of the PMT film. The linear extrapolation indicates the “net” capacitive charge by assuming a constant capacitance for the layer.
Financial support from National Research Found (OTKA no. T 016017) as well as from Ministry of Education (MKM 52192) and Foundation for Hungarian Higher Education (no. 101l/92) are gratefully acknowledged. The authors thank Prof. Jouko Kankare for the valuable discussions during the work.
Separation of faradaic and capacitive current regions REFERENCES I. A. F. Diaz and J. I. Castillo, J. Chem. Sot. Chem. Commun. 397 (1980).
2. A. F. Diaz, J. I. Castillo, J. A. Logan and W. Y. Lee, J. Electroanal. Chem. 129, 115 (1981). 3. S. W. Feldberg, J. Am. Chem. Sot. 106, 4671 (1984). 4. I. Rubinstein, E. Sabatani and J. Rishpon, J. Electrochem. Sot. 134, 3078 (1987). 5. J. Heinze, J. Mortensen, K. Mullen and R. Schenk, J. Chem. Sot. Chem. Commun. 701 (1987). 6. W. J. Albery, Z. Chen, B. R. Horrocks, A. R. Mount, P. J. Wilson, D. Bloor, A. T. Monkman and C. M. Elliot, Faraday Discuss. Chem. Sot. 88, 247 (1989). 7. R. S. Hutton, M. Kalaji and L. M. Peter, J. Electroanal. Chem. 270, 429 (1989).
8. J. Tanguy, J. L. Baudoin, G. Chao and M. Costa, Electrochem.
Acia 37, 1417 (1992).
9. P. Bluerle, Adu. Mafer. 4, 102 (1992). 10. J. Roncali, P. Marque, R. Garreau, F. Garnier and M. Lemaire, Macromolecules 23, 1347 (1990). Il. P. A. Christensen and A. Hamnett, Techniques and Mechanisms in Electrochemistry, p. 335. Blackie Academic & Professional, London (1994). 12. G. Zotti, G. Schiavon, A. Berlin and G. Pagani, Synrh. Metals 61, 81 (1993). 13. G. S. Popkirov and E. Barsoukov, J. Electroanal. Chem. 383, 155 (1995).
14. H.
Harada,
T. Fuchigami and T. J. Electroanal. Chem. 303, 139 (1991).
657
15. Z. Gao, M. Zi and B. Chen, J. Electroanal. Chem. 373, 141 (1994). 16. A. A. Pud, Synth. Metals 66, 1 (1994). 17. Cs. Visy, J. Lukkari and J. Kankare, J. Electroanal. Chem. 401, 119 (1996). 18. E. W. Tsai. S. Basak. J. P. Ruiz. J. R. Revnolds and K. Rajeshwar: J. Elecrrachem. Sot. 136, 36’83 (1989). 19. A. 0. Patil, A. J. Heeger and F. Wudl, Chem. Rec. 183 (1988). 20. Cs. Visy, J. Lukkari and J. Kankare, J. Electroanal.
Chem. 319, 85 (1991). Visy, J. Lukkari and J. Kankare, Macromolecules
21. Cs. 26, 22. Cs. 27,
3295 (1993).
Visy, J. Lukkari and J. Kankare, Macromolecules 3322 (1994).
23. J. Roncali, A. Yassar and F. Gamier, J. Chem. Sot. Chem. Commun. 581 (1988). 24. M.-N. Collomb-Dunard-Sauthier, S. Langlois and E. Genies, J. Appl. Electrochem. 24, 72 (1994). 25. M. Costa. F. Chao. E. Museux and C. Tian. J. Chim. Phys., Phys.-Chim.‘Biol. 92, 939 (1995). 26. T. F. Otero, J. Rodriguez, E. Angulo and C. Santamaria, Synth. Met. 57, 3713 (1993). 27. Q. Pei and 0. Inganiis, Synrh. Met. 57, 3718 (1993). 28. X. Chen and 0. Inganls, Synrh. Metals 74, 159 (1995). 29. A. Hamnett and A. R. Hillmann, Ber. Bunsenges. Phys. Chem. 91, 329 (1987). 30. J. R. Wyhof and H. A. Pohl, J. Polym. Sri. A 8, 1741 (1970).
Nonaka,
31. C. Arbizzani, M. Mastragostino Elecrrochim. Acta 41, 21 (1996).
and L. Meneghello,
Elecfrochimica
Pergamon PII: s0013-46%(%)00210-1
Separation of faradaic and capacitive current regions in the redox transformation of poly(3-methylthiophene) with the ex6lusion of overoxidation processes Cs. Visy,* M. Lakatos, A. Sziics and M. Novak Institute of Physical Chemistry, Jdzsef Attila University,
(Received 5 February
H-6701 Szeged, P.O. Box 105, Hungary
1996; in revised form 23 May 1996)
Abstract-The electrostability of the poly(3-methylthiophene) film prepared in nitrobenzene provided the opportunity to study its redox transformations in an extended anodic potential region with the exclusion of the overoxidation. From spectroelectrochemical results a capacitive region could be separated in the highly oxidized state of the layer which exhibited a large capacitance. This pattern-assumed to be related to charge separation/redistribution within the film at high potential field connected to solvent movements-is promising from the point of view of symmetric p-type/n-type superconductors. Copyright 0 1996 Elsevier Science Ltd Key words: Electronically capacitive current.
conducting
polymer, poly(3-methylthiophene),
INTRODUCTION Very soon after the discovery that electronically conducting polymeric !ilms can be repeatedly driven between their oxidized (doped, conducting) and neutral (reduced, undoped, insulating) forms (11, an interesting characteristic of the cyclic voltammograms was also revealed: at potentials beyond the anodic peak the current tends towards a constant, non zero value which is a function of the sweep rate [2]. This part of the current behaves more or less as a capacitive type, namely that the charge us the potential function in this region is linear. As an explanation, Feldberg proposed to distinguish two processes during the redox transformation of the polymeric films [3], a faradaic type in which the charge carriers are generated in a reversible electron transfer step, and a capacitive type connected to the movement of the carriers to the interface, a charge accumulation without creating new carriers. Since that time it has been shown that the transformation of the neutral film is not a single charge transfer process, but the problem of the faradaic and the capacitive current is not yet solved and returns from time to time in the literature [4-l 31, *Author to whom correspondence
should be addressed.
redox transformation,
faradaic/
partly because of the overlapping overoxidation processes, partly because of the inseparability of the current components in pure electrical measurements. On the basis of complex impedance measurements analyzed through equivalent circuits, the distinction between the two types of charge is connected to the doping counter ion. According to Albery [6] the interaction energy between the charge of the chain and the counter ion can be different, leading to strongly and loosely bounded ions. As another explanation two different diffusion rates are assumed for the doping ion, depending on whether the ions move into more closed aggregates or only in larger pores around the aggregates [8]. In contrast, Heinze suggested the assignment of the current in the critical region to subsequent redox processes connected to a series of polymeric segments of different length [5]. This theory, considering the current at least partially faradaic, was also supported in [9]. From chronopotentiometric results obtained with end-capped thiophene oligomers, Zotti opposed the previous conclusion [ 121and favoured the “molecular metal” model by Feldberg. In a recent work the capacitive pattern of the majority of the current even at the beginning of the oxidation of polypyrrole and polybithiophene was concluded based on
651
652
Cs. Visy et al.
electrochemical impedance spectroscopic measurements [13]. The current generally assigned to the doping process is assumed to create a capacitance and to behave as a capacitive current. However, the possibility of parallel charge transfer processes such as over-oxidation cannot be excluded which-as in other cases-limit the validity of the model applied in the simplified equivalent electronic circuit. Overoxidation is the summarizing name of the processes in the more positive potential region which cause the irreversible deactivation of the polymeric film. The oxidized segments of the chain react with nucleophiles in the solution, which leads to the partial or total degradation of the film. This should be distinguished from the temporary deactivation [14] caused by ion-pair formation in the film, which can be resolved at least partially by more positive polarization. Although in some cases the overoxidation influences positively the properties of the film [15] eg the increased permeability and the enhanced cation exchanger pattern of polypyrrole, attention should be paid to the solution composition to prevent unwanted degradation processes [ 161. The aim of the present work was to prepare a polymeric film which is stable at positive potentials, this provides an opportunity to study the character of the current in the potential region beyond the current peak without interference from overoxidative degradation processes. It was recently demonstrated that very stable and well-dopable polymeric film can be obtained by the electrochemical polymerization of thiophene type monomers in nitrobenzene solution [17]. From the point of view of the anodic stability, the 3-substituted derivatives of thiophene are more favourable. On this basis 3-methylthiophene (MT) was chosen as a model, because it proved to be the most suitable among the alkyl-substituted analogues [181. In order to follow the faradaic type processes during the redox transformation of the polymeric film (PMT), in situ spectroelectrochemistry seems to be applicable: the generation of the charge carriers coupled with ionic movements can be followed from the spectral changes [lo], and the region of these processes can be compared to the current and charge measured during the transformation.
MT (Aldrich) was vacuum-distilled, tetrabutylammonium hexafluorophosphate (Bu4NPF6, Aldrich) was dried in vacuum, nitrobenzene (NB, Aldrich) was used as received, acetonitrile (AN, Aldrich) was dried over activated alumina and kept under dried nitrogen atmosphere. The water content of the solution was controlled by coulometric Karl Fischer titration, and it was always kept below 30 ppm. The IT0 glass was treated by oxygen plasma in Harrick plasma generator before use. The cell and solutions were deaerated by dried nitrogen flow. The PMT film was prepared galvanostatically at 2.5 mA/cm2 current density for 50 mC/cm’ total charge from a solution of 0.1 M for both the MT and the BmNPF6 in NB. The film was then reduced, washed several times with acetonitrile to remove the monomer, and the cell was filled up with 0.1 M BudNPFe/AN solution. The redox transformation of the PMT film was studied by using cyclic voltammetric and potential step methods. In the first case the transient current of the voltammetric curve was integrated in different potential regions, and the spectra were taken at the end-potentials of the intervals. In the second type of measurements, the potential was changed stepwise, the chronoamperograms were recorded after each step of 100 mV, and the transferred charges were determined by integration. The stationary spectra belonging to the actual potential and charge value were recorded when the steady-state current had been reached.
RESULTS
AND DISCUSSION
Figure 1 shows the chronopotentiogram during the galvanostatic polymerization of MT in NB solution. The almost constant potential value throughout the process indicates the formation of a well conducting film [17]. The redox pattern of the PMT layer was studied in monomer-free BuNPFs/AN solution. As POTENTIAL/V
5ti
EXPERIMENTAL In situ spectroelectrochemical measurements were carried out in a one compartment cell placed horizontally into the beam-hole of the spectrophotometer. The working electrode was an IT0 glass (ca 20 n/n) with a surface area of 0.28 cm2. A gold and a silver electrode were used as counter and reference, respectively. The cell was controlled by an Electroflex (type 45 1) potentio/galvanostat, the spectra were registered by an HP8452A diode array spectrophotometer. The measurements were synchronized by a PC.
2,
1
0' 0
5
10
15
20
25
TiMEis
Fig. 1. Typical chronopotentiogram obtained during the galvanostatic polymerization of MT in nitrobenzene at i = 2.5 mA/cm* current density.
Separation of faradaic and capacitive current regions
I
I
0
I
500
I
1000
E/mV
Fig. 2. Subsequent cyclic voltammetric curves registered with PMT film in acetonitrile solution in an extended potential region from -0.4 to + 1.5 V.
can be seen in Fig. 2, the oxidation of the film gives a current peak at around 0.60.65 V, which is followed by a rather long plateau, and at 1.3 V a new electrochemical process starts. The subsequent scans reflect the decrease in the charge transferred reversibly during the oxidation-reduction cycles. As the spectral differences before and after the scannings also indicated the irreversible deactivation of the film, the current in the most positive potential region was connected to degradation process. When the potential scan was restricted to below 1.3 V, both the charge values and the spectral changes proved the reversible redox transformation of the polymer. Figure 3 shows the reproduction of the second voltammetric curve after 10 min scanning although the anodic endpoint of the potential scan was 0.65-0.7 V more positive than the potential of the anodic peak.
1
653
The stability of the film was supported also by the spectral measurements (Fig. 4(a) and (b)); the spectrum at the beginning of the cycling is identical with that obtained at the end, after subsequent switchings between the oxidized and reduced form. The unusual shape of the spectra in the 400-500 nm wavelength region is due to the fact that the reference spectrum has been registered before the polymerization in the slightly coloured NB, while the redox transformations were studied in AN. The spectrum belonging to the neutral form of the film has the absorbance maximum at -530 nm, which reflects a highly conjugated polymer [lo, 201. It can also be seen that the spectrum of the highly oxidized form is similar to that reported as exhibiting a quasi-metallic behaviour (IO, 191. On the basis of these results, it was concluded that overoxidation takes place only beyond 1.3 V, so in the potential range below 1.3 V there is no other faradaic process but the so called doping-undoping reaction which can be followed by the spectral changes. In order to see the relation of the faradaic type charge transfer process to the total charge consumed during a cycle, two characteristic absorption values were plotted us the charge determined from the integration of the voltammetric currentpotential curve. To monitor the change in the amount of the neutral and the well oxidized forms of the polymer, the absorbance values at 532 nm and 758 nm were chosen, respectively. Figure 5 clearly shows that most of the spectral changes occur during a ca 600-800 uC charge transfer, so the current below 0.8-0.9 V is mainly faradaic. However, a large amount of charge passes without, or relatively almost without, spectral changes, hence the slope of the curve is almost zero while the second half of the total charge is transferred. This behaviour may indicate a charge accumulation in a capacitor. The character of the absorbance-charge plot is rather similar during the reverse scan (Fig. 6) but in the opposite order. First mainly the “capacitor” seems to be discharged, which is followed by the decrease in the charge carriers. The parallel absorbance increase observed at the beginning of the reduction in both wavelength ranges is further evidence of the modified redox mechanism of thiophene type polymers [2&22] which interpreted this observation by a dissociative reduction of the dication: b2+ + e- = n + p+.
b’
500
1000
E/mV
Fig. 3. Stability of the current-potential curve of PMT from - 0.4 to + 1.3 V potential us Ag electrode in AN: the second sweep (1) and the one after 10 min scanning (2).
This reaction describes well the fact that at the beginning of the reduction of the totally oxidized film the absorption increase in the low wavelength range-indicating the re-formation of neutral species-is not accompanied by a decrease in the mid-gap excitations, in fact an increase can be seen. This maximum-type curve of the absorbance at 758 nm can be explained by the different absorption coefficients of the monocation and the dication forms
654
cs. visy et al
ABSORBANCE IO
8
6
(4
WAVELENGTHhm Fig. 4(a).
[22] both having in this wavelength range the same excitation from the valence band to the non-bonding mid-gap level [19]. The results obtained from the stationary spectral measurements following the potential steps gave further evidence to support the previous observations for the separation of the two types of processes. As Fig. 7 shows, the separation of the charge transfer regions connected to spectral changes and/or to no spectral modifications is still more expressed, although the technique is able to follow the spectral effect of the delayed, slow conformational changes, as well. On this basis we concluded that there is a potential region at least as wide as 1.1-1.3 V, where the current is exclusively capacitive, and in the preceding potential region this capacitive current may flow also simultaneously with the faradaic current. In order to see the separation and/or overlapping of the two types of the current, the charge values were plotted us the potential (Fig. 8). If we assume that the capacitance of the film is independent of the potential, the constant slope obtained in the potential range 1.1-1.3 V-considered as the “net” capacitive region-can be extrapolated to less anodic potentials. From this rough extrapolation it follows that the capacitor has been created from at least 0.75 V potential by the faradaic process, and it becomes
gradually more and more dominant. It is worth mentioning that in the present case the subsequent redox transformation of the differently conjugated segments [5] does not interfer, which is in accordance with the fact that there is no blue (hypsochromic) or red (bathochromic) shift of the absorbance of the neutral form during its redox transformations [21], but we have a rather uniform, well-conjugated film. The capacitance of the layer was determined from the slope of the Q-E data, and a value of ca 14 mF/cm2 was obtained. For the thickness an estimation from the polymerization charge can be applied. Assuming that 5 mC/cm2 gives a 10 nm thick film (231, a thickness of 10es cm and a capacitancedensity of 1400 F/cm3 is obtained [24]. The question is what structure could stabilize such a big amount of charge concentrated in the thin polymeric film. Recently, the significant swelling and contraction of the PMT/LiClOd film was reported during its oxidative and reductive transformation, respectively [25], which was delayed with respect to the charge transfer, and it was interpreted by the movement of the doping ions. Since these kinds of changes in the size of a polymeric film proved to be solvent dependent [26,27], and considering that in the case of Li-salts a cation repulsion may take place, the delayed swelling can be connected rather to solvent
655
Separation of faradaic and capacitive current regions
ABSORBANCE
8
6
4
2
0 400
500
(b).
600
700
800
WAVELENGTHhm
Fig. 4. Spectra registered during the cyclic voltammetric transformation (b) during the reduction section (l-9) of the scan.
penetration. Of course the character of the incorporated ion affects the uptake of the solvent through solvation interactions as concluded in [28]. In contrast, the penetration of solvent molecules has been observed during the undoping process [29].
of the PMT film: (a) during the oxidation (l-l 1);
The capacitance C is connected to the geometric and dielectric properties as C = &A/l where E is the (absolute) permittivity,
A the surface
ABSORBANCE
6
i
6
4
2
0 ”
4 CHARGEimC
Fig. 5. Absorbance changes at 532 nm (solid) and 758 mn (dotted) wavelengths us the charge transferred during the voltammetric oxidation of PMT film.
1
2
3
CHARGEimC
Fig. 6. Absorbance changes at 532 nm (solid) and 758 nm (dotted) wavelengths us the charge transferred during the voltammetric reduction of PMT film.
Cs. Visy et al. haviour of the so-called hyperelectronic polarization of EKA-conjugated substances [30] reported some 25 years ago. It is assumed that in the increasing electric field a charge separation or redistribution occurs within the doped film, which is related to the movement of solvent molecules. The question of whether it is an uptake or expulsion of the solvent molecules is still open and will be the topic of future studies. Taking into account the fact that PMT can be both anodically and cathodically doped [20], this polymeric film is a promising material for the realisation of a type III, symmetric p-type/n-type supercapacitor [31]. 3
2
3
4
CHARGEimC
CONCLUSIONS (b)
7
”
i
Fig. 7. Stationary absorbance IIS charge plots at 532 nm (solid) and 758 nm (dotted) wavelengths registered (a) after positive potential steps; (b) after negative potential steps.
area (0.28 cm2 in this case), I the thickness of the layer. From the previous data a relative permittivity .sr= 1.6 x lo6 is obtained, which recalls the be-
The electropolymerization of MT in NB solution resulted in a well dopable film which proved to be stable in a rather extended positive potential range up to 1.3 V, 0.6547 V more positive than the peak potential in the oxidative curve. With the exclusion of the overoxidation, the reversible redox transformation of the PMT could be studied up to a quasi-metallic character by spectroelectrochemical methods, There were no spectral changes during the oxidation of the film in the 1.1-1.3 V potential interval, from which fact the net capacitive pattern of the current in this region was concluded. In the less anodic potential range the two types of current are overlapping. By assuming a constant capacitance of the film, which is only a rough estimation, the appearance and increasingly dominant pattern of the capacitive current starts from at least 0.75 V. On this basis it is concluded that the large capacitor is created by the charge transfer generating carriers in the film. In parallel with the accumulation of charge-pairs (called doping) in the thin film, solvent movement occurs, which is assumed to facilitate the charge separation/redistribution within the film at high potential field, leading to a large capacitance. On the basis of simple but frequently applied estimations, the big capacitance has been connected to a large relative permittivity in the lo6 order of magnitude. The behaviour of the PMT film is promising from the point of view of the realization of a supercapacitor based on p-type and n-type forms.
ACKNOWLEDGEMENTS
0
10
5
15
POTENTIAL%
Fig. 8. Charge transferred during the positive scan USthe actual potential obtained for the oxidation of the PMT film. The linear extrapolation indicates the “net” capacitive charge by assuming a constant capacitance for the layer.
Financial support from National Research Found (OTKA no. T 016017) as well as from Ministry of Education (MKM 52192) and Foundation for Hungarian Higher Education (no. 101l/92) are gratefully acknowledged. The authors thank Prof. Jouko Kankare for the valuable discussions during the work.
Separation of faradaic and capacitive current regions REFERENCES I. A. F. Diaz and J. I. Castillo, J. Chem. Sot. Chem. Commun. 397 (1980).
2. A. F. Diaz, J. I. Castillo, J. A. Logan and W. Y. Lee, J. Electroanal. Chem. 129, 115 (1981). 3. S. W. Feldberg, J. Am. Chem. Sot. 106, 4671 (1984). 4. I. Rubinstein, E. Sabatani and J. Rishpon, J. Electrochem. Sot. 134, 3078 (1987). 5. J. Heinze, J. Mortensen, K. Mullen and R. Schenk, J. Chem. Sot. Chem. Commun. 701 (1987). 6. W. J. Albery, Z. Chen, B. R. Horrocks, A. R. Mount, P. J. Wilson, D. Bloor, A. T. Monkman and C. M. Elliot, Faraday Discuss. Chem. Sot. 88, 247 (1989). 7. R. S. Hutton, M. Kalaji and L. M. Peter, J. Electroanal. Chem. 270, 429 (1989).
8. J. Tanguy, J. L. Baudoin, G. Chao and M. Costa, Electrochem.
Acia 37, 1417 (1992).
9. P. Bluerle, Adu. Mafer. 4, 102 (1992). 10. J. Roncali, P. Marque, R. Garreau, F. Garnier and M. Lemaire, Macromolecules 23, 1347 (1990). Il. P. A. Christensen and A. Hamnett, Techniques and Mechanisms in Electrochemistry, p. 335. Blackie Academic & Professional, London (1994). 12. G. Zotti, G. Schiavon, A. Berlin and G. Pagani, Synrh. Metals 61, 81 (1993). 13. G. S. Popkirov and E. Barsoukov, J. Electroanal. Chem. 383, 155 (1995).
14. H.
Harada,
T. Fuchigami and T. J. Electroanal. Chem. 303, 139 (1991).
657
15. Z. Gao, M. Zi and B. Chen, J. Electroanal. Chem. 373, 141 (1994). 16. A. A. Pud, Synth. Metals 66, 1 (1994). 17. Cs. Visy, J. Lukkari and J. Kankare, J. Electroanal. Chem. 401, 119 (1996). 18. E. W. Tsai. S. Basak. J. P. Ruiz. J. R. Revnolds and K. Rajeshwar: J. Elecrrachem. Sot. 136, 36’83 (1989). 19. A. 0. Patil, A. J. Heeger and F. Wudl, Chem. Rec. 183 (1988). 20. Cs. Visy, J. Lukkari and J. Kankare, J. Electroanal.
Chem. 319, 85 (1991). Visy, J. Lukkari and J. Kankare, Macromolecules
21. Cs. 26, 22. Cs. 27,
3295 (1993).
Visy, J. Lukkari and J. Kankare, Macromolecules 3322 (1994).
23. J. Roncali, A. Yassar and F. Gamier, J. Chem. Sot. Chem. Commun. 581 (1988). 24. M.-N. Collomb-Dunard-Sauthier, S. Langlois and E. Genies, J. Appl. Electrochem. 24, 72 (1994). 25. M. Costa. F. Chao. E. Museux and C. Tian. J. Chim. Phys., Phys.-Chim.‘Biol. 92, 939 (1995). 26. T. F. Otero, J. Rodriguez, E. Angulo and C. Santamaria, Synth. Met. 57, 3713 (1993). 27. Q. Pei and 0. Inganiis, Synrh. Met. 57, 3718 (1993). 28. X. Chen and 0. Inganls, Synrh. Metals 74, 159 (1995). 29. A. Hamnett and A. R. Hillmann, Ber. Bunsenges. Phys. Chem. 91, 329 (1987). 30. J. R. Wyhof and H. A. Pohl, J. Polym. Sri. A 8, 1741 (1970).
Nonaka,
31. C. Arbizzani, M. Mastragostino Elecrrochim. Acta 41, 21 (1996).
and L. Meneghello,