- Email: [email protected]
The measurement of specific capacitances of conducting polymers using the quartz crystal microbalance
Available online at www.sciencedirect.com Journal of
Electroanalytical Chemistry Journal of Electroanalytical Chemistry 612 (2008) 140–146 www.elsevier.com/locate/jelechem
Short Communication
The measurement of specific capacitances of conducting polymers using the quartz crystal microbalance Graeme A. Snook a
a,b,*
, George Z. Chen
a,*
School of Chemical and Environmental Engineering, University of Nottingham, University Park, Nottingham NG7 2RD, UK b CSIRO Division of Minerals, Clayton South, Victoria 3168, Australia Received 24 April 2007; received in revised form 30 August 2007; accepted 31 August 2007 Available online 8 September 2007
Abstract Mass specific capacitance (F g1) of conducting polymers is an important factor in selecting these materials for use in supercapacitors. Presented in this work is a new method for measuring the mass specific capacitance of conducting polymers using the electrochemical quartz crystal microbalance (EQCM). It can be done by one of two ways: (1) using one thin deposit, calculating the capacitance and dividing by the deposition mass, or (2) conversion of the capacitance per deposition charge (for thick deposits) to capacitance per mass using Faraday’s Law and the EQCM data (for thin deposits). Using the new method, specific capacitances of polyaniline, polypyrrole and poly[3,4-ethylenedioxythiopene] comparable to those quoted in the literature are found. Such obtained mass specific capacitance should effectively be the theoretical maximum capacitance that can be extracted from a thick film. The standard plots of mass vs. potential were converted to ‘‘massograms’’ by differentiating the mass with respect to time. As the mass variation rate, dM/dt, is proportional to the current, I, the plot is directly comparable to the voltammogram. The onset of polymerisation and nucleation loops are made more apparent using the massogram plot. Using this data, the ideal deposition potentials for the conducting polymers can be determined. Ó 2007 Elsevier B.V. All rights reserved. Keywords: Supercapacitor; Mass specific capacitance; Conducting polymers; Electrochemical quartz crystal microbalance; Massogram
1. Introduction The electrochemical quartz crystal microbalance (EQCM) is a powerful technique for studying solid state reactions [1,2]. This technique involves using quartz crystals with metal discs (typically gold) coated on both sides with one side used as a working electrode. This device utilises the piezoelectric properties of quartz crystals to measure the attached mass (down to nanogram levels) on the electrode surface. A change in the resonant frequency can be related to a change in the mass according to the Sauerbrey equation [3].
*
Corresponding authors. Address: School of Chemical and Environmental Engineering, University of Nottingham, University Park, Nottingham NG7 2RD, UK. Tel.: +44 115 9514171; fax: +44 115 9514115. E-mail addresses: [email protected] (G.A. Snook), george.chen@ nottingham.ac.uk (G.Z. Chen). 0022-0728/$ - see front matter Ó 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jelechem.2007.08.024
The EQCM has been used extensively to study conducting polymers such as polypyrrole [4–12], poly(3,4-ethylenedioxythiopene) [12–17] and polyaniline [18–27]. This technique can be used for three main purposes in studying conducting polymers. Firstly, it can be used to monitor the deposition of the conducting polymer from the monomer solution. Secondly, the EQCM can be used to study the intercalation of ions into the deposited conducting polymer layer. Finally, something that few publications calculate, the EQCM can be used to calculate the mass specific capacitance (Cspec, F g1) of the polymer layer. This can firstly be done by capacitance measurements (Cthin) on a single thin deposit (of measured mass, Dmthin): C thin C spec ¼ ð1Þ Dmthin Alternatively, it can be extended to thicker layers by converting the capacitance per deposition charge (F C1)
G.A. Snook, G.Z. Chen / Journal of Electroanalytical Chemistry 612 (2008) 140–146
for thick deposits to Cspec using Faraday’s Law. i.e. Q = nNF = n(M/Meq)F, where F is Faraday’s constant, Q the deposition charge, n the number of electrons, N the number of moles reacted, M the mass of the deposited polymer and Meq the equivalent polymer mass per electron transferred during polymerisation: C C F C spec ¼ ¼ ð2Þ M Q M eq where C/Q is the slope of a plot of capacitance vs. deposition charge for thick deposits. The value of Meq can be calculated from the EQCM data from deposition of a thin layer, again using Faraday’s Law: M eq ¼
F Dmthin Qthin
141
and the Maclab 8s system. All experiments were carried out at ambient temperatures (20.0 ± 2 °C). The reference half cell was Ag/AgCl in 3.0 M KCl(aq). Working electrodes were 13 mm diameter AT-cut a-quartz crystals (CH Instruments). These crystals had gold disks (5 mm diameter) vapour deposited on each side and resonated in air at 8 ± 0.05 MHz. The crystals used for Fig. 2 resonated at 10 ± 0.05 MHz (Bright Star Crystals, Rowville, Victoria, Australia). The auxiliary electrode was a platinum wire. AC impedance of the thicker deposits was recorded using the Autolab PGSTAT30 potentiostat using a 1.6 mm diameter platinum disc as the working electrode for PPy and PEDOT and using a 6.5 mm diameter graphite disc as the working electrode for PAni.
ð3Þ
where Dmthin and Qthin are respectively the deposition mass and charge of the thin layer. This calculation of capacitance per gram of material is very important when evaluating conducting polymers as possible supercapacitor materials. An accurate calculation of this value allows the usefulness of a variety of conducting polymers as supercapacitor materials to be determined. None of the reports cited in this paper uses massograms [28] to investigate the conducting polymer systems. Massograms are the plots of differentiated mass vs. potential, which are directly comparable to voltammograms as dðDMÞ / I. In this paper, massograms will be used to invesdt tigate the deposition of the polymers. Nucleation loops [29] are made more apparent in the massogram plot and confirm that the growth of the polymers is via a nucleation and growth mechanism. This observation has been previously reported for polypyrrole [4]. Also, massograms more clearly show the optimal potential for deposition of the polymers so that over-oxidation is avoided. 2. Experimental Pyrrole (Py) (98%, Aldrich), aniline (Ani) (98%, Aldrich), 3,4-ethylenedioxythiophene (EDOT) (Bayer), tetrabutylammonium bromide (>99%, Fluka), lithium chloride (99%, May & Bayer) and potassium chloride (>99%, Sigma) were all used as purchased. Distilled water was used for preparation of electrolytic solutions. Polypyrrole (PPy) was deposited from either an aqueous solution of 0.1 M Py and 0.5 M lithium chloride or from an aqueous solution of 0.5 M Py and 0.5 M potassium chloride. Polyaniline (PAni) was deposited from an aqueous solution of 0.25 M Ani and 1.0 M HCl. Poly-(3,4-ethylenedioxythiophene) (PEDOT) was deposited from an acetonitrile solution of 0.2 M EDOT and 0.5 M lithium perchlorate. Voltammograms and massograms were recorded using the CH Instruments 440 Potentiostat or the EG&G PAR Model 173 Potentiostat coupled with the EG&G PAR Model 175 Universal Programmer (Princeton Applied Research Corp., New Jersey, USA). Data acquisition was assisted by the Elchema EQCN (Elchema, Potsdam, NY)
3. Results and discussion 3.1. Mass data for polymer depositions The EQCM results are presented in Fig. 1 for the deposition of PEDOT from an acetonitrile solution of 0.2 M EDOT and 0.5 M lithium perchlorate. There are significant currents flowing between 0.5 V and 0.8 V (vs. Ag/AgCl). However, no significant mass change occurs in the same potential range, which means these currents are background currents resulting in no deposition of PEDOT onto the electrode surface. Previous work [30] suggests this is the reaction of gold adsorbed EDOT species. The mass plot shows that significant deposition begins to occur at 1.0 V (vs. Ag/AgCl). There is a slight current loop, which is indicative of nucleation and growth [29] of the polymer. The massogram plot of dM/dt vs. potential reveals a more apparent mass loop, that further confirms that polymer growth is via a nucleation and growth mechanism. In addition, it can be seen from the massogram that significant deposition begins to occur at 1.0 V. Hence, the authors recommend a constant deposition potential of 1.0 V to avoid over-oxidation and to deposit a polymer layer of sufficient thickness. Table 1 indicates this as a suitable potential for constant potential deposition of thicker deposits. Using constant potential deposition data (Qthin and Dmthin), it was found Meq = 87 ± 16 g mol1 e1. A list of values is given in Table 2. According to theory, Meq = (Mpolymer + Mcounterion)/(2 + c), where c is the molar ratio of dopant ions and monomeric units (i.e. dopant level) in the oxidised polymer and (2 + c) is the number of electrons withdrawn from one monomeric unit during anodic deposition of the polymer [30]. The typical values of c are given in Table 2 [31]. Given c 0.3 for PEDOT, Meq = 74 g mol1 e1. Consequently, the QCM data matches closely with theory, taking into account possible solvation of the cation and experimental errors. As long as the deposition is reasonably thin, the equivalent mass per electron calculated should be very accurate. The equivalent mass per electron was also calculated for a thick layer (30 C cm2) which was removed from a platinum electrode and weighed using a mechanical balance. The value
142
G.A. Snook, G.Z. Chen / Journal of Electroanalytical Chemistry 612 (2008) 140–146
Fig. 1. (a) Cyclic voltammogram (50 mV s1), (b) mass plot and (c) massogram for the deposition of PEDOT from a 0.2 M EDOT/0.5 M LiClO4 (ACN) solution.
Table 1 Recommended deposition potentials Polymer
Deposition potential (V)
PPy PAni PEDOT
0.7 0.9 1.0
calculated was approximately 100 g mol1, which agrees qualitatively with the EQCM measurement. The EQCM results for the deposition of PPy from an aqueous solution of Py/Cl have been previously published by this group [4]. Fig. 2 shows the deposition using 0.1 M pyrrole/0.5 M LiCl(aq) and sweeping to more negative potentials than the previous publication. There is no
significant or obvious background current. The voltammetric current rapidly rises at around 0.7 V (vs. Ag/AgCl). The mass plot confirms a significant deposition of mass at this potential. The authors recommend a deposition potential of 0.7 V to avoid over-oxidation of the polymer and for sufficiently rapid coating of the electrode. There is a reverse peak observed when the voltammetry is carried out back to more negative potentials. This is the de-intercalation and reduction of the polymer. This was not seen in Fig. 2 for PEDOT, as the voltammetry was not extended to a sufficiently negative range. Multiple scans are shown, in which the magnitude of the polymer redox chemistry (current) increases with continuing deposition. The equivalent mass per electron calculation, using the constant potential deposition data, yields a value of 40 ± 5 g mol1 e1 compared
Table 2 Constant voltage deposition data Polymer
Capacitance per charge (F C1)
Measured Meq(g mol1 e1)
Theoretical Meq(g mol1 e1)
ca
Mass specific capacitance (F g1)
PPy PAni PEDOT
0.10 0.15 0.083
40 27 87
34 44 74
0.33 0.5 0.3
240 530 92
a
Cited from Ref. [31].
G.A. Snook, G.Z. Chen / Journal of Electroanalytical Chemistry 612 (2008) 140–146
143
Fig. 2. (a) Cyclic voltammogram (20 mV s1), (b) mass plot and (c) massogram for the deposition PPy from a 0.1 M pyrrole/0.5 M LiCl(aq) solution (Cycles 2–3).
with the theoretical equivalent mass of 34 g mol1 e1 (see Table 2). Again the EQCM measurement is higher than the theoretical value but the difference is within experimental errors. Fig. 3 shows the 8th cycle of the deposition of PAni as measured by the EQCM using an aqueous solution of 0.25 M aniline and 1.0 M HCl. Again there is a reverse (reduction) peak (this time much closer to the deposition potential) at 0.55 V (vs. Ag/AgCl). The reverse (reduction) peak for the other two polymers occurs at below 0.0 V (vs. Ag/AgCl). The mass plot shows a mass increase, which becomes significant at 0.9 V (vs. Ag/AgCl). This is the suggested deposition potential for growing thicker deposits of the polymer as shown in Table 1. The nucleation loop is not apparent as the scan has been reversed at too high a potential. It is observed, once the potential direction is reversed, that the mass decreases at below approximately 0.8 V (vs. Ag/AgCl). This is attributed to the expulsion of chloride from the oxidised polymer. The massogram supports this interpretation as the massogram peak matches the voltammetric reduction peak (both occurring at around 0.55 V). The equivalent mass per electron is calculated (using constant potential deposition data) as 27 ± 6 g mol1 e1 (compared with a theoretical equivalent
mass per unit of 44 g mol1). The value of Meq calculated from QCM data is significantly lower than theory. This can be explained by the observation of the formation of coloured/soluble oligmers of Ani [18] that contribute to the current/charge but not the deposited mass. 3.2. Alternative calculations of equivalent mass from cyclic voltammetry measurements Calculations of the equivalent mass, Meq, in the previous section (and shown in Table 2) are all from constant potential deposition. In this section, a number of calculations will be done for Meq from the cyclic voltammetry data. These are summarised in Table 3 along with the constant potential data (last column). The first calculation, Meq(MPP), is the calculation of equivalent mass using the mass (from the mass potential plot [MPP]) and the charge from the integrated cyclic voltammogram using Eq. (3). The next calculation, Meq(MQ), is using the slope from a plot of mass vs. charge (effectively gives a straight line) and converting to Meq by multiplying by Faraday’s constant. Fig. 4a shows the plot of mass vs. charge from which Meq(MQ) is calculated. To the best of our knowledge, this plot has not been used before to calculate the
144
G.A. Snook, G.Z. Chen / Journal of Electroanalytical Chemistry 612 (2008) 140–146
Fig. 3. (a) Cyclic voltammogram (20 mV s1), (b) mass plot and (c) massogram for the deposition PAni from a 0.1 M aniline/1.0 M HCl(aq) solution (Cycle 8).
Table 3 Values of equivalent polymer mass per electron transferred in electrochemical polymerisation (Meq, g mol1 e1) calculated from different dataa Polymer
Meq(MPP)
Meq(MQ)
Meq(MG)
Meq(dMdtI)
Meq(BG)
Meq(CP)
PPy PAni PEDOT
29 28 44
25 31 38
27 26 38
29 35 42
– – 62
40 27 87
a The original data used for the calculation of Meq are from the mass potential plot (MPP), slope of mass vs. Q plot (MQ), massogram plot (MG), slope of dM/dt vs. I (dMdtI), massogram plot minus background (BG), and constant potential deposition (CP).
equivalent mass per electron. This particular plot is using PPy but only scanning in the range where deposition occurs (ie 0.0 V to 0.7 V vs. Ag/AgCl). The third calculation, Meq(MG), employs the value of dM/dt from the massogram plot and the current from the cyclic voltammogram (at the switching potential) using the following equation: M eq ¼
F dM=dt I
ð4Þ
The fourth calculation is Meq (dMdtI), by calculating the value of the slope of a plot of dM/dt vs. I (which again gives a straight line) and equating to molecular weight by multiplying by Faraday’s constant. Fig. 4b shows the example plot for this calculation. The final calculation, done for PEDOT, subtracts the background current (IBG) from the cyclic voltammogram (I) to calculate the equivalent mass according to Eq. (5):
M eq ¼
F dM=dt I I BG
ð5Þ
Looking at the data for PPy, the cyclic voltammetric data consistently gives lower values of Meq compared to the constant potential data. The most likely explanation for this is the different mode of deposition resulting in different morphology and hence a different c value (dopant level). The PAni cyclic voltammetry data are also consistent for the different methods of calculation of Meq and this time match well with the value obtained under constant potential conditions. This suggests growth is very similar in both types of depositions and that similar proportions of oligimers are lost to the solution giving a lower equivalent mass than expected. The PEDOT cyclic voltammetry data are also consistent for the different methods of calculation but give
G.A. Snook, G.Z. Chen / Journal of Electroanalytical Chemistry 612 (2008) 140–146
145
Fig. 4. (a) Plot of mass vs. charge for PPy, and (b) plot of dM/dt vs. current for PPy (deposition from 0.5 M Py/ 0.5 M KCl at 50 mV s1, between 0.0 V and 0.7 V vs. Ag/AgCl).
considerably lower values than the constant potential method. This is most likely due to the background currents discussed previously. In the column Meq(BG) the background current has been subtracted in the calculation of equivalent mass and gives a value much closer to the Meq values as calculated theoretically or from constant potential data. The value could still be too low, due to different morphologies growing using the different methods of deposition. It has been shown in earlier work [30] that constant potential deposition gives the most favourable morphology for thick growth and rapid kinetics in PEDOT.
3.3. Calculation of mass specific capacitance of the polymers Using Eqs. (2) and (3), the capacitance per deposition charge (F C1) can be converted to the mass specific capacitance (Cspec, F g1) using Faraday’s Law and the EQCM (constant potential) data. Essentially, we use the previously derived Meq values to calculate Cspec for thick deposits. Fig. 5 shows the plot of capacitance vs. deposition charge for thick deposits of PAni (galvanostatic deposition), PPy, and PEDOT (potentiostatic deposition). The lines on the graph are the linear parts of the depositions. So for PPy the line only goes through the points up to around 5 C cm2 as the capacitance tails off at the higher deposition charges. Table 2 shows the calculated value of capacitance per deposition charge for each polymer, Meq and the resulting Cspec. The PEDOT shows Cspec = 92 F g1, compared with PPy with 240 F g1 and PAni with a very large 530 F g1. The low Cspec value of PEDOT is mainly due to the molecular weight per monomer unit (138 g mol1) being very large compared to PPy (67.1 g mol1) and PAni (93.1 g mol1). PAni has a high value due to its high dopant level (high c value, as seen in Table 2). The values of PEDOT and PPy agree with the literature and the large values for PAni have been seen before. These Cspec values
Fig. 5. Capacitance vs. deposition charge as calculated using AC impedance (amplitude = 10 mV) for PAni, PPy deposits (bias potential = 0.5 V vs. Ag/AgCl), and PEDOT deposits (bias potential = 0.4 V vs. Ag/AgCl). Including in the graph is the calculated slope of the capacitance vs. deposition charge (2 significant figures).
generally agree with the literature and hence validate the calculated Meq values. The thin deposit calculation (from Eq. 1) was done with PPy [4]. Using 0.5 M KCl(aq) as the electrolyte a 11.5 lg deposit gave an average (over different scan rates) of 2.7 ± 0.1 mF. Dividing by the mass deposited (11.5 lg) gives Cspec = 235 ± 9 F g1. This agrees with the previous calculation. For the tetrabutylammonium bromide, a capacitance of 2.5 ± 0.2 mF was obtained, corresponding to Cspec = 217 ± 15 F g1. Therefore, the larger counter ion results in a lower Cspec value most likely due to the slow kinetics due to steric effects impeding fast and deeper intercalation. This limited depth of penetration has been seen previously for solid organometallic compounds abrasively attached onto electrode surfaces [32].
146
G.A. Snook, G.Z. Chen / Journal of Electroanalytical Chemistry 612 (2008) 140–146
4. Conclusions Using the EQCM and massograms, the optimal potentials to deposit three conducting polymers are suggested. A new method of calculating mass specific capacitance (Cspec, F g1) from the QCM data is demonstrated. Calculation of Cspec using this method results in accurate values that agree with the previous literature. Two new types of QCM plots were used to calculate the equivalent polymer mass per electron (Meq, g mol1 e1), namely M vs. Q and dM/dt vs. I. These new plots may lead to new and interesting interpretation of EQCM data. Acknowledgements The authors thank the EPSRC for funding and Dr. Greg Wilson (CSIRO) for valuable discussions and Mr. Michael Horne (CSIRO) for comments on the manuscript. References [1] G. Inzelt, in: A.J. Bard (Ed.), Electroanalytical Chemistry, Vol. 18, M..Dekker, New York, 1994, p. 89. [2] D.A. Buttry, in: A.J. Bard (Ed.), Electroanalytical Chemistry, Vol. 17, M. Dekker, New York, 1991, p. 1. [3] G. Sauerbrey, Z. Phys. 155 (1959) 206. [4] G.A. Snook, G.Z. Chen, D.J. Fray, M. Hughes, M. Shaffer, J. Electroanal. Chem. 568 (2004) 135. [5] S. Song, D. Han, H. Lee, H. Cho, S. Chang, J. Kim, H. Muramatsu, Synth. Met. 117 (2001) 137. [6] S. Bruckenstein, K. Brzezinska, A.R. Hillman, Electrochim. Acta 45 (2000) 3801. [7] S. Bruckenstein, K. Brzezinska, A.R. Hillman, Phys. Chem. Chem. Phys. 2 (2000) 1221. [8] M. Song, J. Park, I. Yeo, H. Rhee, Synth. Met. 99 (1999) 219. [9] S. Suematsu, Y. Oura, H. Tsujimoto, H. Kanno, K. Naoi, Electrochim. Acta 45 (2000) 3813.
[10] J.M. Kim, S.M. Chang, H.W. Lee, Y.S. Kwon, Y.H. Oh, Synth. Met. 85 (1997) 1371. [11] A. Bund, A. Baba, S. Berg, D. Johannsmann, J. Lubben, Z. Wang, W. Knoll, J. Phys. Chem. B 107 (2003) 6743. [12] W. Paik, I.H. Yeo, H. Suh, Y. Kim, E. Song, Electrochim. Acta 45 (2000) 3833. [13] F. Blanchard, B. Carre, F. Bonhomme, P. Biensan, H. Pages, D. Lemordant, J. Electroanal. Chem. 569 (2004) 203. [14] A. Bund, M. Schneider, J. Electrochem. Soc. 149 (2002) E331–E339. [15] L. Niu, C. Kvarnstrom, S. Dong, A. Ivaska, Synth. Met. 121 (2001) 1389. [16] G.A. Sotzing, J.R. Reynolds, P.J. Steel, Adv. Mater. 9 (1997) 795. [17] A. Lisowska-Oleksiak, K. Kazubowska, A. Kupniewska, J. Electroanal. Chem. 501 (2001) 54. [18] M.Q. Wu, G.A. Snook, V. Gupta, M. Shaffer, D.J. Fray, G.Z. Chen, J. Mater. Chem. 15 (2005) 2297. [19] G. Inzelt, Electrochim. Acta 45 (2000) 3865. [20] A. Baba, S. Tian, F. Stefani, C. Xia, Z. Wang, R.C. Advincula, D. Johannsmann, W. Knoll, J. Electroanal. Chem. 562 (2004) 95. [21] S.J. Choi, S.M. Park, J. Electrochem. Soc. 149 (2002) E26. [22] S.Y. Cui, S.M. Park, Synth. Met. 105 (1999) 91. [23] M.M. Ayad, N. Salahuddin, M.A. Shenashin, Synth. Met. 142 (2004) 101. [24] R.M. Torresi, S.L.D. Maranhao, J. Electrochem. Soc. 146 (1999) 4179. [25] S. Pruneanu, E. Csahok, V. Kertesz, G. Inzelt, Electrochim. Acta 43 (1998) 2305. [26] M.M. Ayad, A.H. Gemaey, N. Salahuddin, M.A. Shenashin, J. Colloid Interface Sci. 263 (2003) 196. [27] R.M.Q. Mello, R.M. Torresi, S.I. Cordoba de Torresi, E.A. Ticianelli, Langmuir 16 (2000) 7835. [28] G.A. Snook, A.M. Bond, S. Fletcher, J. Electroanal. Chem. 526 (2002) 1. [29] S. Fletcher, D. Gates, C.S. Halliday, T. Lwin, G. Nelson, M. Westcott, J. Electroanal. Chem. 159 (1983) 267. [30] G.A. Snook, C. Peng, D.J. Fray, G.Z. Chen, Electrochem. Commun. 9 (2007) 83. [31] K. Lota, V. Khomenko, E. Frackowiak, J. Phys. Chem. Solids 65 (2004) 295. [32] G.A. Snook, A.M. Bond, S. Fletcher, J. Electroanal. Chem. 554 (2003) 157.
Electroanalytical Chemistry Journal of Electroanalytical Chemistry 612 (2008) 140–146 www.elsevier.com/locate/jelechem
Short Communication
The measurement of specific capacitances of conducting polymers using the quartz crystal microbalance Graeme A. Snook a
a,b,*
, George Z. Chen
a,*
School of Chemical and Environmental Engineering, University of Nottingham, University Park, Nottingham NG7 2RD, UK b CSIRO Division of Minerals, Clayton South, Victoria 3168, Australia Received 24 April 2007; received in revised form 30 August 2007; accepted 31 August 2007 Available online 8 September 2007
Abstract Mass specific capacitance (F g1) of conducting polymers is an important factor in selecting these materials for use in supercapacitors. Presented in this work is a new method for measuring the mass specific capacitance of conducting polymers using the electrochemical quartz crystal microbalance (EQCM). It can be done by one of two ways: (1) using one thin deposit, calculating the capacitance and dividing by the deposition mass, or (2) conversion of the capacitance per deposition charge (for thick deposits) to capacitance per mass using Faraday’s Law and the EQCM data (for thin deposits). Using the new method, specific capacitances of polyaniline, polypyrrole and poly[3,4-ethylenedioxythiopene] comparable to those quoted in the literature are found. Such obtained mass specific capacitance should effectively be the theoretical maximum capacitance that can be extracted from a thick film. The standard plots of mass vs. potential were converted to ‘‘massograms’’ by differentiating the mass with respect to time. As the mass variation rate, dM/dt, is proportional to the current, I, the plot is directly comparable to the voltammogram. The onset of polymerisation and nucleation loops are made more apparent using the massogram plot. Using this data, the ideal deposition potentials for the conducting polymers can be determined. Ó 2007 Elsevier B.V. All rights reserved. Keywords: Supercapacitor; Mass specific capacitance; Conducting polymers; Electrochemical quartz crystal microbalance; Massogram
1. Introduction The electrochemical quartz crystal microbalance (EQCM) is a powerful technique for studying solid state reactions [1,2]. This technique involves using quartz crystals with metal discs (typically gold) coated on both sides with one side used as a working electrode. This device utilises the piezoelectric properties of quartz crystals to measure the attached mass (down to nanogram levels) on the electrode surface. A change in the resonant frequency can be related to a change in the mass according to the Sauerbrey equation [3].
*
Corresponding authors. Address: School of Chemical and Environmental Engineering, University of Nottingham, University Park, Nottingham NG7 2RD, UK. Tel.: +44 115 9514171; fax: +44 115 9514115. E-mail addresses: [email protected] (G.A. Snook), george.chen@ nottingham.ac.uk (G.Z. Chen). 0022-0728/$ - see front matter Ó 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jelechem.2007.08.024
The EQCM has been used extensively to study conducting polymers such as polypyrrole [4–12], poly(3,4-ethylenedioxythiopene) [12–17] and polyaniline [18–27]. This technique can be used for three main purposes in studying conducting polymers. Firstly, it can be used to monitor the deposition of the conducting polymer from the monomer solution. Secondly, the EQCM can be used to study the intercalation of ions into the deposited conducting polymer layer. Finally, something that few publications calculate, the EQCM can be used to calculate the mass specific capacitance (Cspec, F g1) of the polymer layer. This can firstly be done by capacitance measurements (Cthin) on a single thin deposit (of measured mass, Dmthin): C thin C spec ¼ ð1Þ Dmthin Alternatively, it can be extended to thicker layers by converting the capacitance per deposition charge (F C1)
G.A. Snook, G.Z. Chen / Journal of Electroanalytical Chemistry 612 (2008) 140–146
for thick deposits to Cspec using Faraday’s Law. i.e. Q = nNF = n(M/Meq)F, where F is Faraday’s constant, Q the deposition charge, n the number of electrons, N the number of moles reacted, M the mass of the deposited polymer and Meq the equivalent polymer mass per electron transferred during polymerisation: C C F C spec ¼ ¼ ð2Þ M Q M eq where C/Q is the slope of a plot of capacitance vs. deposition charge for thick deposits. The value of Meq can be calculated from the EQCM data from deposition of a thin layer, again using Faraday’s Law: M eq ¼
F Dmthin Qthin
141
and the Maclab 8s system. All experiments were carried out at ambient temperatures (20.0 ± 2 °C). The reference half cell was Ag/AgCl in 3.0 M KCl(aq). Working electrodes were 13 mm diameter AT-cut a-quartz crystals (CH Instruments). These crystals had gold disks (5 mm diameter) vapour deposited on each side and resonated in air at 8 ± 0.05 MHz. The crystals used for Fig. 2 resonated at 10 ± 0.05 MHz (Bright Star Crystals, Rowville, Victoria, Australia). The auxiliary electrode was a platinum wire. AC impedance of the thicker deposits was recorded using the Autolab PGSTAT30 potentiostat using a 1.6 mm diameter platinum disc as the working electrode for PPy and PEDOT and using a 6.5 mm diameter graphite disc as the working electrode for PAni.
ð3Þ
where Dmthin and Qthin are respectively the deposition mass and charge of the thin layer. This calculation of capacitance per gram of material is very important when evaluating conducting polymers as possible supercapacitor materials. An accurate calculation of this value allows the usefulness of a variety of conducting polymers as supercapacitor materials to be determined. None of the reports cited in this paper uses massograms [28] to investigate the conducting polymer systems. Massograms are the plots of differentiated mass vs. potential, which are directly comparable to voltammograms as dðDMÞ / I. In this paper, massograms will be used to invesdt tigate the deposition of the polymers. Nucleation loops [29] are made more apparent in the massogram plot and confirm that the growth of the polymers is via a nucleation and growth mechanism. This observation has been previously reported for polypyrrole [4]. Also, massograms more clearly show the optimal potential for deposition of the polymers so that over-oxidation is avoided. 2. Experimental Pyrrole (Py) (98%, Aldrich), aniline (Ani) (98%, Aldrich), 3,4-ethylenedioxythiophene (EDOT) (Bayer), tetrabutylammonium bromide (>99%, Fluka), lithium chloride (99%, May & Bayer) and potassium chloride (>99%, Sigma) were all used as purchased. Distilled water was used for preparation of electrolytic solutions. Polypyrrole (PPy) was deposited from either an aqueous solution of 0.1 M Py and 0.5 M lithium chloride or from an aqueous solution of 0.5 M Py and 0.5 M potassium chloride. Polyaniline (PAni) was deposited from an aqueous solution of 0.25 M Ani and 1.0 M HCl. Poly-(3,4-ethylenedioxythiophene) (PEDOT) was deposited from an acetonitrile solution of 0.2 M EDOT and 0.5 M lithium perchlorate. Voltammograms and massograms were recorded using the CH Instruments 440 Potentiostat or the EG&G PAR Model 173 Potentiostat coupled with the EG&G PAR Model 175 Universal Programmer (Princeton Applied Research Corp., New Jersey, USA). Data acquisition was assisted by the Elchema EQCN (Elchema, Potsdam, NY)
3. Results and discussion 3.1. Mass data for polymer depositions The EQCM results are presented in Fig. 1 for the deposition of PEDOT from an acetonitrile solution of 0.2 M EDOT and 0.5 M lithium perchlorate. There are significant currents flowing between 0.5 V and 0.8 V (vs. Ag/AgCl). However, no significant mass change occurs in the same potential range, which means these currents are background currents resulting in no deposition of PEDOT onto the electrode surface. Previous work [30] suggests this is the reaction of gold adsorbed EDOT species. The mass plot shows that significant deposition begins to occur at 1.0 V (vs. Ag/AgCl). There is a slight current loop, which is indicative of nucleation and growth [29] of the polymer. The massogram plot of dM/dt vs. potential reveals a more apparent mass loop, that further confirms that polymer growth is via a nucleation and growth mechanism. In addition, it can be seen from the massogram that significant deposition begins to occur at 1.0 V. Hence, the authors recommend a constant deposition potential of 1.0 V to avoid over-oxidation and to deposit a polymer layer of sufficient thickness. Table 1 indicates this as a suitable potential for constant potential deposition of thicker deposits. Using constant potential deposition data (Qthin and Dmthin), it was found Meq = 87 ± 16 g mol1 e1. A list of values is given in Table 2. According to theory, Meq = (Mpolymer + Mcounterion)/(2 + c), where c is the molar ratio of dopant ions and monomeric units (i.e. dopant level) in the oxidised polymer and (2 + c) is the number of electrons withdrawn from one monomeric unit during anodic deposition of the polymer [30]. The typical values of c are given in Table 2 [31]. Given c 0.3 for PEDOT, Meq = 74 g mol1 e1. Consequently, the QCM data matches closely with theory, taking into account possible solvation of the cation and experimental errors. As long as the deposition is reasonably thin, the equivalent mass per electron calculated should be very accurate. The equivalent mass per electron was also calculated for a thick layer (30 C cm2) which was removed from a platinum electrode and weighed using a mechanical balance. The value
142
G.A. Snook, G.Z. Chen / Journal of Electroanalytical Chemistry 612 (2008) 140–146
Fig. 1. (a) Cyclic voltammogram (50 mV s1), (b) mass plot and (c) massogram for the deposition of PEDOT from a 0.2 M EDOT/0.5 M LiClO4 (ACN) solution.
Table 1 Recommended deposition potentials Polymer
Deposition potential (V)
PPy PAni PEDOT
0.7 0.9 1.0
calculated was approximately 100 g mol1, which agrees qualitatively with the EQCM measurement. The EQCM results for the deposition of PPy from an aqueous solution of Py/Cl have been previously published by this group [4]. Fig. 2 shows the deposition using 0.1 M pyrrole/0.5 M LiCl(aq) and sweeping to more negative potentials than the previous publication. There is no
significant or obvious background current. The voltammetric current rapidly rises at around 0.7 V (vs. Ag/AgCl). The mass plot confirms a significant deposition of mass at this potential. The authors recommend a deposition potential of 0.7 V to avoid over-oxidation of the polymer and for sufficiently rapid coating of the electrode. There is a reverse peak observed when the voltammetry is carried out back to more negative potentials. This is the de-intercalation and reduction of the polymer. This was not seen in Fig. 2 for PEDOT, as the voltammetry was not extended to a sufficiently negative range. Multiple scans are shown, in which the magnitude of the polymer redox chemistry (current) increases with continuing deposition. The equivalent mass per electron calculation, using the constant potential deposition data, yields a value of 40 ± 5 g mol1 e1 compared
Table 2 Constant voltage deposition data Polymer
Capacitance per charge (F C1)
Measured Meq(g mol1 e1)
Theoretical Meq(g mol1 e1)
ca
Mass specific capacitance (F g1)
PPy PAni PEDOT
0.10 0.15 0.083
40 27 87
34 44 74
0.33 0.5 0.3
240 530 92
a
Cited from Ref. [31].
G.A. Snook, G.Z. Chen / Journal of Electroanalytical Chemistry 612 (2008) 140–146
143
Fig. 2. (a) Cyclic voltammogram (20 mV s1), (b) mass plot and (c) massogram for the deposition PPy from a 0.1 M pyrrole/0.5 M LiCl(aq) solution (Cycles 2–3).
with the theoretical equivalent mass of 34 g mol1 e1 (see Table 2). Again the EQCM measurement is higher than the theoretical value but the difference is within experimental errors. Fig. 3 shows the 8th cycle of the deposition of PAni as measured by the EQCM using an aqueous solution of 0.25 M aniline and 1.0 M HCl. Again there is a reverse (reduction) peak (this time much closer to the deposition potential) at 0.55 V (vs. Ag/AgCl). The reverse (reduction) peak for the other two polymers occurs at below 0.0 V (vs. Ag/AgCl). The mass plot shows a mass increase, which becomes significant at 0.9 V (vs. Ag/AgCl). This is the suggested deposition potential for growing thicker deposits of the polymer as shown in Table 1. The nucleation loop is not apparent as the scan has been reversed at too high a potential. It is observed, once the potential direction is reversed, that the mass decreases at below approximately 0.8 V (vs. Ag/AgCl). This is attributed to the expulsion of chloride from the oxidised polymer. The massogram supports this interpretation as the massogram peak matches the voltammetric reduction peak (both occurring at around 0.55 V). The equivalent mass per electron is calculated (using constant potential deposition data) as 27 ± 6 g mol1 e1 (compared with a theoretical equivalent
mass per unit of 44 g mol1). The value of Meq calculated from QCM data is significantly lower than theory. This can be explained by the observation of the formation of coloured/soluble oligmers of Ani [18] that contribute to the current/charge but not the deposited mass. 3.2. Alternative calculations of equivalent mass from cyclic voltammetry measurements Calculations of the equivalent mass, Meq, in the previous section (and shown in Table 2) are all from constant potential deposition. In this section, a number of calculations will be done for Meq from the cyclic voltammetry data. These are summarised in Table 3 along with the constant potential data (last column). The first calculation, Meq(MPP), is the calculation of equivalent mass using the mass (from the mass potential plot [MPP]) and the charge from the integrated cyclic voltammogram using Eq. (3). The next calculation, Meq(MQ), is using the slope from a plot of mass vs. charge (effectively gives a straight line) and converting to Meq by multiplying by Faraday’s constant. Fig. 4a shows the plot of mass vs. charge from which Meq(MQ) is calculated. To the best of our knowledge, this plot has not been used before to calculate the
144
G.A. Snook, G.Z. Chen / Journal of Electroanalytical Chemistry 612 (2008) 140–146
Fig. 3. (a) Cyclic voltammogram (20 mV s1), (b) mass plot and (c) massogram for the deposition PAni from a 0.1 M aniline/1.0 M HCl(aq) solution (Cycle 8).
Table 3 Values of equivalent polymer mass per electron transferred in electrochemical polymerisation (Meq, g mol1 e1) calculated from different dataa Polymer
Meq(MPP)
Meq(MQ)
Meq(MG)
Meq(dMdtI)
Meq(BG)
Meq(CP)
PPy PAni PEDOT
29 28 44
25 31 38
27 26 38
29 35 42
– – 62
40 27 87
a The original data used for the calculation of Meq are from the mass potential plot (MPP), slope of mass vs. Q plot (MQ), massogram plot (MG), slope of dM/dt vs. I (dMdtI), massogram plot minus background (BG), and constant potential deposition (CP).
equivalent mass per electron. This particular plot is using PPy but only scanning in the range where deposition occurs (ie 0.0 V to 0.7 V vs. Ag/AgCl). The third calculation, Meq(MG), employs the value of dM/dt from the massogram plot and the current from the cyclic voltammogram (at the switching potential) using the following equation: M eq ¼
F dM=dt I
ð4Þ
The fourth calculation is Meq (dMdtI), by calculating the value of the slope of a plot of dM/dt vs. I (which again gives a straight line) and equating to molecular weight by multiplying by Faraday’s constant. Fig. 4b shows the example plot for this calculation. The final calculation, done for PEDOT, subtracts the background current (IBG) from the cyclic voltammogram (I) to calculate the equivalent mass according to Eq. (5):
M eq ¼
F dM=dt I I BG
ð5Þ
Looking at the data for PPy, the cyclic voltammetric data consistently gives lower values of Meq compared to the constant potential data. The most likely explanation for this is the different mode of deposition resulting in different morphology and hence a different c value (dopant level). The PAni cyclic voltammetry data are also consistent for the different methods of calculation of Meq and this time match well with the value obtained under constant potential conditions. This suggests growth is very similar in both types of depositions and that similar proportions of oligimers are lost to the solution giving a lower equivalent mass than expected. The PEDOT cyclic voltammetry data are also consistent for the different methods of calculation but give
G.A. Snook, G.Z. Chen / Journal of Electroanalytical Chemistry 612 (2008) 140–146
145
Fig. 4. (a) Plot of mass vs. charge for PPy, and (b) plot of dM/dt vs. current for PPy (deposition from 0.5 M Py/ 0.5 M KCl at 50 mV s1, between 0.0 V and 0.7 V vs. Ag/AgCl).
considerably lower values than the constant potential method. This is most likely due to the background currents discussed previously. In the column Meq(BG) the background current has been subtracted in the calculation of equivalent mass and gives a value much closer to the Meq values as calculated theoretically or from constant potential data. The value could still be too low, due to different morphologies growing using the different methods of deposition. It has been shown in earlier work [30] that constant potential deposition gives the most favourable morphology for thick growth and rapid kinetics in PEDOT.
3.3. Calculation of mass specific capacitance of the polymers Using Eqs. (2) and (3), the capacitance per deposition charge (F C1) can be converted to the mass specific capacitance (Cspec, F g1) using Faraday’s Law and the EQCM (constant potential) data. Essentially, we use the previously derived Meq values to calculate Cspec for thick deposits. Fig. 5 shows the plot of capacitance vs. deposition charge for thick deposits of PAni (galvanostatic deposition), PPy, and PEDOT (potentiostatic deposition). The lines on the graph are the linear parts of the depositions. So for PPy the line only goes through the points up to around 5 C cm2 as the capacitance tails off at the higher deposition charges. Table 2 shows the calculated value of capacitance per deposition charge for each polymer, Meq and the resulting Cspec. The PEDOT shows Cspec = 92 F g1, compared with PPy with 240 F g1 and PAni with a very large 530 F g1. The low Cspec value of PEDOT is mainly due to the molecular weight per monomer unit (138 g mol1) being very large compared to PPy (67.1 g mol1) and PAni (93.1 g mol1). PAni has a high value due to its high dopant level (high c value, as seen in Table 2). The values of PEDOT and PPy agree with the literature and the large values for PAni have been seen before. These Cspec values
Fig. 5. Capacitance vs. deposition charge as calculated using AC impedance (amplitude = 10 mV) for PAni, PPy deposits (bias potential = 0.5 V vs. Ag/AgCl), and PEDOT deposits (bias potential = 0.4 V vs. Ag/AgCl). Including in the graph is the calculated slope of the capacitance vs. deposition charge (2 significant figures).
generally agree with the literature and hence validate the calculated Meq values. The thin deposit calculation (from Eq. 1) was done with PPy [4]. Using 0.5 M KCl(aq) as the electrolyte a 11.5 lg deposit gave an average (over different scan rates) of 2.7 ± 0.1 mF. Dividing by the mass deposited (11.5 lg) gives Cspec = 235 ± 9 F g1. This agrees with the previous calculation. For the tetrabutylammonium bromide, a capacitance of 2.5 ± 0.2 mF was obtained, corresponding to Cspec = 217 ± 15 F g1. Therefore, the larger counter ion results in a lower Cspec value most likely due to the slow kinetics due to steric effects impeding fast and deeper intercalation. This limited depth of penetration has been seen previously for solid organometallic compounds abrasively attached onto electrode surfaces [32].
146
G.A. Snook, G.Z. Chen / Journal of Electroanalytical Chemistry 612 (2008) 140–146
4. Conclusions Using the EQCM and massograms, the optimal potentials to deposit three conducting polymers are suggested. A new method of calculating mass specific capacitance (Cspec, F g1) from the QCM data is demonstrated. Calculation of Cspec using this method results in accurate values that agree with the previous literature. Two new types of QCM plots were used to calculate the equivalent polymer mass per electron (Meq, g mol1 e1), namely M vs. Q and dM/dt vs. I. These new plots may lead to new and interesting interpretation of EQCM data. Acknowledgements The authors thank the EPSRC for funding and Dr. Greg Wilson (CSIRO) for valuable discussions and Mr. Michael Horne (CSIRO) for comments on the manuscript. References [1] G. Inzelt, in: A.J. Bard (Ed.), Electroanalytical Chemistry, Vol. 18, M..Dekker, New York, 1994, p. 89. [2] D.A. Buttry, in: A.J. Bard (Ed.), Electroanalytical Chemistry, Vol. 17, M. Dekker, New York, 1991, p. 1. [3] G. Sauerbrey, Z. Phys. 155 (1959) 206. [4] G.A. Snook, G.Z. Chen, D.J. Fray, M. Hughes, M. Shaffer, J. Electroanal. Chem. 568 (2004) 135. [5] S. Song, D. Han, H. Lee, H. Cho, S. Chang, J. Kim, H. Muramatsu, Synth. Met. 117 (2001) 137. [6] S. Bruckenstein, K. Brzezinska, A.R. Hillman, Electrochim. Acta 45 (2000) 3801. [7] S. Bruckenstein, K. Brzezinska, A.R. Hillman, Phys. Chem. Chem. Phys. 2 (2000) 1221. [8] M. Song, J. Park, I. Yeo, H. Rhee, Synth. Met. 99 (1999) 219. [9] S. Suematsu, Y. Oura, H. Tsujimoto, H. Kanno, K. Naoi, Electrochim. Acta 45 (2000) 3813.
[10] J.M. Kim, S.M. Chang, H.W. Lee, Y.S. Kwon, Y.H. Oh, Synth. Met. 85 (1997) 1371. [11] A. Bund, A. Baba, S. Berg, D. Johannsmann, J. Lubben, Z. Wang, W. Knoll, J. Phys. Chem. B 107 (2003) 6743. [12] W. Paik, I.H. Yeo, H. Suh, Y. Kim, E. Song, Electrochim. Acta 45 (2000) 3833. [13] F. Blanchard, B. Carre, F. Bonhomme, P. Biensan, H. Pages, D. Lemordant, J. Electroanal. Chem. 569 (2004) 203. [14] A. Bund, M. Schneider, J. Electrochem. Soc. 149 (2002) E331–E339. [15] L. Niu, C. Kvarnstrom, S. Dong, A. Ivaska, Synth. Met. 121 (2001) 1389. [16] G.A. Sotzing, J.R. Reynolds, P.J. Steel, Adv. Mater. 9 (1997) 795. [17] A. Lisowska-Oleksiak, K. Kazubowska, A. Kupniewska, J. Electroanal. Chem. 501 (2001) 54. [18] M.Q. Wu, G.A. Snook, V. Gupta, M. Shaffer, D.J. Fray, G.Z. Chen, J. Mater. Chem. 15 (2005) 2297. [19] G. Inzelt, Electrochim. Acta 45 (2000) 3865. [20] A. Baba, S. Tian, F. Stefani, C. Xia, Z. Wang, R.C. Advincula, D. Johannsmann, W. Knoll, J. Electroanal. Chem. 562 (2004) 95. [21] S.J. Choi, S.M. Park, J. Electrochem. Soc. 149 (2002) E26. [22] S.Y. Cui, S.M. Park, Synth. Met. 105 (1999) 91. [23] M.M. Ayad, N. Salahuddin, M.A. Shenashin, Synth. Met. 142 (2004) 101. [24] R.M. Torresi, S.L.D. Maranhao, J. Electrochem. Soc. 146 (1999) 4179. [25] S. Pruneanu, E. Csahok, V. Kertesz, G. Inzelt, Electrochim. Acta 43 (1998) 2305. [26] M.M. Ayad, A.H. Gemaey, N. Salahuddin, M.A. Shenashin, J. Colloid Interface Sci. 263 (2003) 196. [27] R.M.Q. Mello, R.M. Torresi, S.I. Cordoba de Torresi, E.A. Ticianelli, Langmuir 16 (2000) 7835. [28] G.A. Snook, A.M. Bond, S. Fletcher, J. Electroanal. Chem. 526 (2002) 1. [29] S. Fletcher, D. Gates, C.S. Halliday, T. Lwin, G. Nelson, M. Westcott, J. Electroanal. Chem. 159 (1983) 267. [30] G.A. Snook, C. Peng, D.J. Fray, G.Z. Chen, Electrochem. Commun. 9 (2007) 83. [31] K. Lota, V. Khomenko, E. Frackowiak, J. Phys. Chem. Solids 65 (2004) 295. [32] G.A. Snook, A.M. Bond, S. Fletcher, J. Electroanal. Chem. 554 (2003) 157.