- Email: [email protected]
1.2 Volt manganese oxide symmetric supercapacitor
Electrochemistry Communications 13 (2011) 1264–1267
Contents lists available at SciVerse ScienceDirect
Electrochemistry Communications journal homepage: www.elsevier.com/locate/elecom
1.2 Volt manganese oxide symmetric supercapacitor Fatemeh Ataherian, Nae-Lih Wu ⁎ Department of Chemical Engineering, National Taiwan University, Taipei 106, Taiwan
a r t i c l e
i n f o
Article history: Received 29 July 2011 Received in revised form 15 August 2011 Accepted 16 August 2011 Available online 26 August 2011 Keywords: Manganese oxide Electrochemical capacitor Capacitance fading Cycle life
a b s t r a c t Capacitance fading of MnO2 supercapacitor electrode under negative polarization below 0.0 V (versus Ag/ AgCl/sat. KCl(aq)) arises from extensive reduction of Mn(IV) to form inactive Mn(II) species, and this has typically limited the operating voltage window of an aqueous symmetric MnO2 supercapacitor to be no greater than 0.8 V. As this lower potential limit is close to the onset potential of MnO2-catalyzed oxygen reduction reaction (ORR), the fading problem can be alleviated by effectively passing the accumulated electrons in the oxide electrode to the dissolved oxygen molecules in electrolyte in order to avoid the formation of the surface Mn(II) species. This has been demonstrated by either increasing the dissolved oxygen content or using the Ti(IV)/Ti(III) redox couple in the electrolyte as a charge-transfer mediator to enhance the electrocatalytic activity of MnO2 for ORR. Therefore, a MnO2 symmetric supercapacitor showing remarkable cycling stability over an operating voltage window of 1.2 V has been achieved by using Ti(IV)-containing neutral electrolyte (1 M KCl(aq)). © 2011 Elsevier B.V. All rights reserved.
1. Introduction MnO2 has been intensively studied as a supercapacitor electrode material in neutral aqueous electrolytes. It possesses the advantages of low cost, sufficiently high specific capacitance, and environmentally friendly nature [1]. The operating voltage window is a critical parameter, considering the voltage square dependence of energy and power densities, of a capacitor cell. Unfortunately, the voltage window of the aqueous symmetric MnO2 supercapacitor has typically been limited to be less than 1.0 V [2]. For example, in the case of KCl(aq) electrolyte, it has been shown that MnO2 electrode suffers from obvious capacitance fading either below 0.0 V (versus Ag/AgCl/sat. KCl(aq)) under negative polarization, due to formation of inactive Mn(II) surface species, or above 1.0 V under positive polarization, owing to extensive oxygen oxidation along with electrode passivation [3,4]. Amorphous MnO2 electrode has a typical open-circuit potential (OCP) of 0.4 V in a neutral aqueous electrolyte, and therefore the overall workable voltage window of a symmetric MnO2 supercapacitor is limited to 0.8 V. Komaba et al. [5] once reported the success of suppressing capacitance fading down to −0.1 V by introducing small amount of either NaHPO4 or NaHCO3 into electrolyte. This might allow for a voltage window of 1 V for the symmetric MnO2 cell, although such a cell was not demonstrated. The present work provides new insight into the capacitance fading issue of MnO2 electrode under negative polarization and suggests a
⁎ Corresponding author. Tel.: + 886 223627158. E-mail address: [email protected] (N.-L. Wu). 1388-2481/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.elecom.2011.08.028
new approach to achieve symmetric MnO2 supercapacitor cell with significantly widened, 1.2 V, operating voltage window. 2. Experimental MnO2 particles were synthesized by redox reaction between KMnO4 and MnSO4 aqueous solutions with a Mn(VII)/Mn(II) molar ratio of 2:3 at 25 °C. After being thoroughly washed with de-ionized water, the particles were finally heated at 200 °C for 1 h in air. To prepare the electrode, slurry containing the oxide powder, acetylene black (AB) and polyvinylidene difluoride (PVdF) dispersed in N-methyl pyrrolidone (NMP), was coated onto Ti foils, and finally dried at 120 °C for 6 h in vacuum. On the dry basis, the oxide to AB ratio was at 7:3, while the binder had a weight composition of 14%. Cyclic voltammetry (CV) was carried out on an electrochemical analyzer (Eco Chemie PGSTAT30). A beakertype electrochemical cell which consists of a Pt counter electrode, a reference (Ag/AgCl/saturated KCl(aq); EG&G; 197 mV versus NHE at 25 °C), and a square shaped MnO2 electrode with surface area of 1 × 1 cm2 as working electrode was employed to characterize the electrochemical behaviors of single MnO2 electrode, i.e., the so-called “half-cell”. Symmetric full-cells were assembled with two square shaped MnO2 electrodes with surface area of 2.5 × 2.5 cm2 arranged face-to-face and a porous separator (BS0712, Coin Nano Tech) in between. The cell was first immersed in excess amount of electrolyte for 24 h and then removed from the electrolyte container and tested in an empty sealed container. This is intended to simulate the situation typically encountered in a practical cell, where the amount of electrolyte is limited. The electrolyte was aqueous 1 M KCl with pH = 7, while the potential scan rate was fixed at 50 mV s − 1. Voltage scan has been started and
F. Ataherian, N.-L. Wu / Electrochemistry Communications 13 (2011) 1264–1267
1265
ended at the lower-end of the voltage window. The capacitance of the electrode was determined from the voltammograms according to the following equation: Cavg ¼ ΔV
−1
−1
∫Is
dV
where Cavg is the average specific capacitance, ΔV is the scanned potential or voltage window, I is current, and s is voltage scan rate. Subsequent to cycling, the electrodes were washed with deionized water and dried at room temperature prior to x-ray photoelectron spectroscopy (XPS) analyses. XPS analysis employed an Al K X-ray source operated at 15 kV and 100 W, and used a beam size of 400 μm and pass energy of 20 eV for spectrum acquisition. Sputtering gun was operated at 3 kV and 1 μA with a sputtering area of 2 × 2 mm2 and sputtering time of 2 min. 3. Result and discussion The fresh MnO2 particles are agglomerates of nanoflakes with the widths in the range of 10–30 nm [3]. The cycling stability of single MnO2 electrode within the potential window of − 0.2 to 0.4 V was tested by using CV scanning, and the variations of specific capacitance under different conditions are summarized in Fig. 1a. First, in order to investigate the effect of oxygen dissolved in electrolyte, the electrolyte was purged with either argon or air for at least 1 h prior to measurements. Hereafter, the electrodes cycled in Ar-purged electrolyte, airpurged electrolyte and blank electrolyte (electrolyte without any prior gas-purging treatment) are denoted as S–Ar, S–O2 and S–B, respectively. The S–B electrode has an initial specific capacitance 65 Fg − 1 at 50 mV s − 1 based on the mass of entire active-layer. As shown in Fig. 1a, the capacitance of the S–B electrode initially decreases monotonously within the first 1000 cycles and then levels off at ca. 50% of the initial value. The capacitance of the S–Ar electrode, which has an initial specific capacitance of 58 Fg − 1, exhibits very similar fading pattern. They are consistent with our previous findings [3]. In great contrast, the S–O2 electrode, which has a slightly higher specific capacitance (77 Fg − 1), shows no fading up to 3000 cycles. Fig. 1b illustrates the voltammograms of selected cycles of the S–B electrode. One clearly sees a prominent reduction tail below 0.0 V for the initial cycles, while there is no additional oxidation counterpart occurring during the anodic scan. Upon continuous cycling, while the capacitance decreases, the intensity of the low-potential reduction tail also diminishes. The reduction tail is no longer recognizable when the capacitance fades to the plateau value after ca. 1000 cycles. Before the capacitance of the electrode reaches the plateau, the total amount of charge passed during each cathodic scan is always higher than that of the corresponding anodic scan, suggesting irreversible electrode reduction. As shown in our previous study [3], the average Mn valence on the electrode surface has been reduced to +2.0 after cycling. The data of the S–O2 electrode (Fig. 1a and c), on the other hand, clearly indicates that the capacitance fading problem is alleviated by increasing the oxygen content in the electrolyte. These results suggest that the dissolved oxygen be involved in maintaining the electrode activity. One notes that MnO2 materials have also been studied as alternative cathodic catalysts to platinum (Pt) in air-cathode microbial fuel cells (MFCs) for oxygen reduction reaction (ORR) [6,7]. In such a process, the reduced Mn(III) species transfer electron to the O2 molecule and itself is oxidized back to Mn(IV), i.e., −
2MnðIIIÞ þ 1=2O2 þ H2 O↔2MnðIVÞ þ 2OH
ð1Þ
This process is facilitated by the fact that the Mn(III)/Mn(IV) redox potential is close to the onset potential of the ORR [8]. It is inferred that the same ORR helps to pass electrons to the dissolved oxygen in the electrolyte when MnO2 is negatively charged to below 0.0 V in the O2-saturated electrolyte in the S–O2 cell so that the formation
Fig. 1. (a) Capacitance retention (CCo− 1), normalized based on the capacitance of the second cycle (Co), versus cycle number for the MnO2 half-cells containing electrolyte without additive under ambient atmosphere (S–B), Ar-purged electrolyte (S–Ar), airpurged electrolyte (S–O2) and electrolyte with Ti(IV) ions as additive (S–Ti) and cycled within the potential window of − 0.2 to 0.4 V versus Ag/AgCl/sat. KCl(aq) reference electrode; (b) cyclic voltammograms of selected cycles for the S–B cell electrode; (c) cyclic voltammograms for the S–Ti cell (scan rate = 50 mV s− 1; 1 M KCl(aq)).
1266
F. Ataherian, N.-L. Wu / Electrochemistry Communications 13 (2011) 1264–1267
of Mn(II) surface species and its associated capacitance fading problem is suppressed. Accordingly, the capacitance fading appearing in the S–B and S–Ar cells may be due to low efficiency of reaction (1) associated with low O2 content. Noel et al. [9] once demonstrated an indirect electrochemical pathway that is capable of increasing the charge-transfer efficiency of reaction (1). Metal ionic redox couples such as Ti(IV)/Ti(III) and Sn(IV)/Sn(III) were used as mediators. The metal ions were reduced from the higher oxidation state to lower oxidation state on the electrode surface and then reacted with the substrate. This greatly motivates us to study the influence of Ti(IV) in the electrolyte on ORR catalytic activity of MnO2 in order to improve cyclic stability of the electrode. To introduce the electrolyte with Ti(IV) ions without causing TiO2 solid precipitate, an electrolyte was prepared as follows. TiO2 powder was first added to 1 M KCl solution with pH = 2, and the solution was stirred for 1 day. Then the solution pH was adjusted to 7 and left for one more day in order to settle down any non-dissolved particulates. The upper clear solution was finally collected, filtrated and used for electrochemical test. The final solution was transparent and colorless. In order to ascertained the presence of Ti(IV) ions in the solution, H2O2 was added and pH decreased gradually. At pH = 5, the color of the solution changed to light yellow, confirming the presence of Ti (IV) ions, although the exact amount of the ions has not been determined [10]. The cycle performance of the MnO2 electrode in the Ti(IV)-containing 1 M KCl solution (denoted as the S–Ti electrode) is summarized in Fig. 1a, while Fig. 1c presents the voltammograms of selected cycles. Note that, the electrolyte has not been bubbled with oxygen. As shown, the capacitance of the electrode is very stable. It is suggested that the enhanced cycling stability is due to the increased efficiency of reaction (1) via Ti(IV)–Ti(III) mediated charge-transfer scheme: TiðIVÞ þ MnðIIIÞ↔TiðIIIÞ þ MnðIVÞ 2TiðIIIÞ þ 1=2O2 þ H2 O↔2TiðIVÞ þ 2OH
ð2Þ −
ð3Þ
It is also worth mentioning that the S–Ti electrode has a specific capacitance of 118 Fg − 1 at 50 mV s − 1, which is significantly higher than those of the rest electrodes. The mechanism for the capacitance enhancement will be investigated elsewhere. The cycled (≥ 3000 cycles between − 0.2 and 0.4 V) electrodes are subjected to XPS analysis for determining the oxidation state of Mn located at the electrode surfaces, based on the multiple splitting of the Mn 3 s core level spectrum [11]. The Mn valence of the fresh electrode is +3.7, while they are +2.0 and +2.7 for the S–B and S–Ti electrodes, respectively. For comparison, an electrode cycled between 0.0 and 0.8 V showing no capacity fading in the “blank” electrolyte (i.e., without Ti(IV)) is also +2.7 [3]. It is clear that permanent Mn reduction to Mn(II) caused by cycling below 0.0 V is effectively suppressed by the presence of Ti(IV) ions in the electrolyte. XPS analysis also detected Ti(2p3/2) signal at 458.0 eV on the S–Ti electrode, which is characteristic of Ti(III) [12,13]. This confirms the occurrence of Ti(IV) reduction on the surface of MnO2 electrode according to reaction (2). Fig. 2 summarizes the electrochemical characteristics of a MnO2 symmetric full-cell with addition of Ti(IV) ions in the electrolyte. Note that the cell contains only limited amount of electrolyte, as described in Section 2. When the full-cell is operating with a voltage range of 1.2 V, the positive and negative electrodes show essentially symmetric potential range from 0.4 to 1.0 V and from 0.4 to −0.2 V, respectively (Fig. 2a). Fig. 2b presents the CVs of selected cycles of the full-cell operation. The curves exhibit the typical rectangular profiles after initial cycles. The cell capacitance does not fade; instead, it slightly increases along the course up to 10,000 cycles (Fig. 2c). The cause to the capacitance rise may partly be attributed to the change in electrode morphology, from being nanoflake to petal-like, resulting
Fig. 2. (a) Cyclic voltammograms of the positive and negative electrodes of a symmetric MnO2 full-cell containing Ti(IV) ions in electrolyte and cycled within a operating voltage window of 1.2 V; (b) cyclic voltammograms of the full cell containing Ti as additive; (c) capacitance retention (CCo− 1), normalized based on the capacitance of the second cycle (Co), versus cycle number for the full cells either with (○) or without (Δ) Ti(IV) the electrolyte (1 M KCl(aq); scan rate = 50 mV s− 1).
in richer mesophorous structure, as previously observed [3]. For comparison, the cell capacitance of a symmetric cell without Ti(IV) addition in the electrolyte loses 80% of the initial capacitance under the
F. Ataherian, N.-L. Wu / Electrochemistry Communications 13 (2011) 1264–1267
same cycling conditions. This full-cell apparently exhibits a much more complex fading history than that of single electrode under negative polarization (Fig. 1a). This may suggest adverse interaction between the two electrodes in a full-cell configuration when the cell is running over unstable voltage range.
Acknowledgments
4. Conclusion
References
Capacitance fading of MnO2 supercapacitor electrode under negative polarization below 0.0 V arises from irreversible formation of inactive Mn(II) surface species. In view of the fact that this potential limit is close to the onset potential of MnO2-catalyzed ORR, the fading problem has been tackled by passing the accumulated electrons in the oxide electrode to the dissolved oxygen molecules via the ORR in order to suppress the formation of the Mn(II) species. This has successfully been demonstrated by either increasing the dissolved oxygen content in the electrolyte or using the Ti(IV)/Ti(III) redox as a charge-transfer mediator to enhance the electrocatalytic activity of MnO2 for ORR. As a result, the reduction potential limit has been lowered to −0.2 V, and a 1.2 V MnO2 symmetric electrochemical supercapacitor with remarkable capacitance retention over more than 10,000 cycles has been achieved by using Ti(IV)-containing neutral electrolyte (1 M KCl(aq)).
1267
This work is financially supported by the National Science Council of Taiwan, R.O.C. and by the National Taiwan University. One of the authors, F. A., would like to thank NTU post-doctor fellowship (NTU99R40044).
[1] H.Y. Lee, J.B. Goodenough, Journal of Solid State Chemistry 144 (1999) 220. [2] E. Raymundo-Piñero, V. Khomenko, E. Frackowiak, F. Béguin, Journal of the Electrochemical Society 152 (2005) A229. [3] F. Ataherian, K.T. Lee, N.L. Wu, Electrochimica Acta 55 (2010) 7429. [4] F. Ataherian, N.L. Wu, Journal of the Electrochemical Society 158 (2011) 1. [5] S. Komaba, A. Ogata, T. Tsuchikawa, Electrochemistry Communications 10 (2008) 1435. [6] F. Cheng, Y. Su, J. Liang, Z. Tao, J. Chen, Chemistry of Materials 22 (2010) 898. [7] W. Xiao, D. Wang, X.W. Lou, Journal of Physical Chemistry C 114 (2010) 1694. [8] I. Roche, K. Scott, Journal of Applied Electrochemistry 39 (2009) 197. [9] M. Noel, P.N. Anatharaman, H.V.K. Udupa, Journal of Applied Electrochemistry 12 (1982) 291. [10] J. Muhlebach, K. Muller, G. Schwarzenbach, Inorganic Chemistry 9 (1970) 2381. [11] M. Toupin, T. Brousse, D. Bélanger, Chemistry of Materials 14 (2002) 3946. [12] P. Stefanov, M. Shipochka, P. Stefchev, Z. Raicheva, V. Lazarova, L. Spassov, Journal of Physics Conference Series 100 (2008) 1. [13] A. Sandell, M.P. Andersson, M.K.J. Johansson, P.G. Karlsson, Y. Alfredsson, J. Schadt, H. Siegbahn, P. Uvdal, Surface Science 530 (2003) 63.
Contents lists available at SciVerse ScienceDirect
Electrochemistry Communications journal homepage: www.elsevier.com/locate/elecom
1.2 Volt manganese oxide symmetric supercapacitor Fatemeh Ataherian, Nae-Lih Wu ⁎ Department of Chemical Engineering, National Taiwan University, Taipei 106, Taiwan
a r t i c l e
i n f o
Article history: Received 29 July 2011 Received in revised form 15 August 2011 Accepted 16 August 2011 Available online 26 August 2011 Keywords: Manganese oxide Electrochemical capacitor Capacitance fading Cycle life
a b s t r a c t Capacitance fading of MnO2 supercapacitor electrode under negative polarization below 0.0 V (versus Ag/ AgCl/sat. KCl(aq)) arises from extensive reduction of Mn(IV) to form inactive Mn(II) species, and this has typically limited the operating voltage window of an aqueous symmetric MnO2 supercapacitor to be no greater than 0.8 V. As this lower potential limit is close to the onset potential of MnO2-catalyzed oxygen reduction reaction (ORR), the fading problem can be alleviated by effectively passing the accumulated electrons in the oxide electrode to the dissolved oxygen molecules in electrolyte in order to avoid the formation of the surface Mn(II) species. This has been demonstrated by either increasing the dissolved oxygen content or using the Ti(IV)/Ti(III) redox couple in the electrolyte as a charge-transfer mediator to enhance the electrocatalytic activity of MnO2 for ORR. Therefore, a MnO2 symmetric supercapacitor showing remarkable cycling stability over an operating voltage window of 1.2 V has been achieved by using Ti(IV)-containing neutral electrolyte (1 M KCl(aq)). © 2011 Elsevier B.V. All rights reserved.
1. Introduction MnO2 has been intensively studied as a supercapacitor electrode material in neutral aqueous electrolytes. It possesses the advantages of low cost, sufficiently high specific capacitance, and environmentally friendly nature [1]. The operating voltage window is a critical parameter, considering the voltage square dependence of energy and power densities, of a capacitor cell. Unfortunately, the voltage window of the aqueous symmetric MnO2 supercapacitor has typically been limited to be less than 1.0 V [2]. For example, in the case of KCl(aq) electrolyte, it has been shown that MnO2 electrode suffers from obvious capacitance fading either below 0.0 V (versus Ag/AgCl/sat. KCl(aq)) under negative polarization, due to formation of inactive Mn(II) surface species, or above 1.0 V under positive polarization, owing to extensive oxygen oxidation along with electrode passivation [3,4]. Amorphous MnO2 electrode has a typical open-circuit potential (OCP) of 0.4 V in a neutral aqueous electrolyte, and therefore the overall workable voltage window of a symmetric MnO2 supercapacitor is limited to 0.8 V. Komaba et al. [5] once reported the success of suppressing capacitance fading down to −0.1 V by introducing small amount of either NaHPO4 or NaHCO3 into electrolyte. This might allow for a voltage window of 1 V for the symmetric MnO2 cell, although such a cell was not demonstrated. The present work provides new insight into the capacitance fading issue of MnO2 electrode under negative polarization and suggests a
⁎ Corresponding author. Tel.: + 886 223627158. E-mail address: [email protected] (N.-L. Wu). 1388-2481/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.elecom.2011.08.028
new approach to achieve symmetric MnO2 supercapacitor cell with significantly widened, 1.2 V, operating voltage window. 2. Experimental MnO2 particles were synthesized by redox reaction between KMnO4 and MnSO4 aqueous solutions with a Mn(VII)/Mn(II) molar ratio of 2:3 at 25 °C. After being thoroughly washed with de-ionized water, the particles were finally heated at 200 °C for 1 h in air. To prepare the electrode, slurry containing the oxide powder, acetylene black (AB) and polyvinylidene difluoride (PVdF) dispersed in N-methyl pyrrolidone (NMP), was coated onto Ti foils, and finally dried at 120 °C for 6 h in vacuum. On the dry basis, the oxide to AB ratio was at 7:3, while the binder had a weight composition of 14%. Cyclic voltammetry (CV) was carried out on an electrochemical analyzer (Eco Chemie PGSTAT30). A beakertype electrochemical cell which consists of a Pt counter electrode, a reference (Ag/AgCl/saturated KCl(aq); EG&G; 197 mV versus NHE at 25 °C), and a square shaped MnO2 electrode with surface area of 1 × 1 cm2 as working electrode was employed to characterize the electrochemical behaviors of single MnO2 electrode, i.e., the so-called “half-cell”. Symmetric full-cells were assembled with two square shaped MnO2 electrodes with surface area of 2.5 × 2.5 cm2 arranged face-to-face and a porous separator (BS0712, Coin Nano Tech) in between. The cell was first immersed in excess amount of electrolyte for 24 h and then removed from the electrolyte container and tested in an empty sealed container. This is intended to simulate the situation typically encountered in a practical cell, where the amount of electrolyte is limited. The electrolyte was aqueous 1 M KCl with pH = 7, while the potential scan rate was fixed at 50 mV s − 1. Voltage scan has been started and
F. Ataherian, N.-L. Wu / Electrochemistry Communications 13 (2011) 1264–1267
1265
ended at the lower-end of the voltage window. The capacitance of the electrode was determined from the voltammograms according to the following equation: Cavg ¼ ΔV
−1
−1
∫Is
dV
where Cavg is the average specific capacitance, ΔV is the scanned potential or voltage window, I is current, and s is voltage scan rate. Subsequent to cycling, the electrodes were washed with deionized water and dried at room temperature prior to x-ray photoelectron spectroscopy (XPS) analyses. XPS analysis employed an Al K X-ray source operated at 15 kV and 100 W, and used a beam size of 400 μm and pass energy of 20 eV for spectrum acquisition. Sputtering gun was operated at 3 kV and 1 μA with a sputtering area of 2 × 2 mm2 and sputtering time of 2 min. 3. Result and discussion The fresh MnO2 particles are agglomerates of nanoflakes with the widths in the range of 10–30 nm [3]. The cycling stability of single MnO2 electrode within the potential window of − 0.2 to 0.4 V was tested by using CV scanning, and the variations of specific capacitance under different conditions are summarized in Fig. 1a. First, in order to investigate the effect of oxygen dissolved in electrolyte, the electrolyte was purged with either argon or air for at least 1 h prior to measurements. Hereafter, the electrodes cycled in Ar-purged electrolyte, airpurged electrolyte and blank electrolyte (electrolyte without any prior gas-purging treatment) are denoted as S–Ar, S–O2 and S–B, respectively. The S–B electrode has an initial specific capacitance 65 Fg − 1 at 50 mV s − 1 based on the mass of entire active-layer. As shown in Fig. 1a, the capacitance of the S–B electrode initially decreases monotonously within the first 1000 cycles and then levels off at ca. 50% of the initial value. The capacitance of the S–Ar electrode, which has an initial specific capacitance of 58 Fg − 1, exhibits very similar fading pattern. They are consistent with our previous findings [3]. In great contrast, the S–O2 electrode, which has a slightly higher specific capacitance (77 Fg − 1), shows no fading up to 3000 cycles. Fig. 1b illustrates the voltammograms of selected cycles of the S–B electrode. One clearly sees a prominent reduction tail below 0.0 V for the initial cycles, while there is no additional oxidation counterpart occurring during the anodic scan. Upon continuous cycling, while the capacitance decreases, the intensity of the low-potential reduction tail also diminishes. The reduction tail is no longer recognizable when the capacitance fades to the plateau value after ca. 1000 cycles. Before the capacitance of the electrode reaches the plateau, the total amount of charge passed during each cathodic scan is always higher than that of the corresponding anodic scan, suggesting irreversible electrode reduction. As shown in our previous study [3], the average Mn valence on the electrode surface has been reduced to +2.0 after cycling. The data of the S–O2 electrode (Fig. 1a and c), on the other hand, clearly indicates that the capacitance fading problem is alleviated by increasing the oxygen content in the electrolyte. These results suggest that the dissolved oxygen be involved in maintaining the electrode activity. One notes that MnO2 materials have also been studied as alternative cathodic catalysts to platinum (Pt) in air-cathode microbial fuel cells (MFCs) for oxygen reduction reaction (ORR) [6,7]. In such a process, the reduced Mn(III) species transfer electron to the O2 molecule and itself is oxidized back to Mn(IV), i.e., −
2MnðIIIÞ þ 1=2O2 þ H2 O↔2MnðIVÞ þ 2OH
ð1Þ
This process is facilitated by the fact that the Mn(III)/Mn(IV) redox potential is close to the onset potential of the ORR [8]. It is inferred that the same ORR helps to pass electrons to the dissolved oxygen in the electrolyte when MnO2 is negatively charged to below 0.0 V in the O2-saturated electrolyte in the S–O2 cell so that the formation
Fig. 1. (a) Capacitance retention (CCo− 1), normalized based on the capacitance of the second cycle (Co), versus cycle number for the MnO2 half-cells containing electrolyte without additive under ambient atmosphere (S–B), Ar-purged electrolyte (S–Ar), airpurged electrolyte (S–O2) and electrolyte with Ti(IV) ions as additive (S–Ti) and cycled within the potential window of − 0.2 to 0.4 V versus Ag/AgCl/sat. KCl(aq) reference electrode; (b) cyclic voltammograms of selected cycles for the S–B cell electrode; (c) cyclic voltammograms for the S–Ti cell (scan rate = 50 mV s− 1; 1 M KCl(aq)).
1266
F. Ataherian, N.-L. Wu / Electrochemistry Communications 13 (2011) 1264–1267
of Mn(II) surface species and its associated capacitance fading problem is suppressed. Accordingly, the capacitance fading appearing in the S–B and S–Ar cells may be due to low efficiency of reaction (1) associated with low O2 content. Noel et al. [9] once demonstrated an indirect electrochemical pathway that is capable of increasing the charge-transfer efficiency of reaction (1). Metal ionic redox couples such as Ti(IV)/Ti(III) and Sn(IV)/Sn(III) were used as mediators. The metal ions were reduced from the higher oxidation state to lower oxidation state on the electrode surface and then reacted with the substrate. This greatly motivates us to study the influence of Ti(IV) in the electrolyte on ORR catalytic activity of MnO2 in order to improve cyclic stability of the electrode. To introduce the electrolyte with Ti(IV) ions without causing TiO2 solid precipitate, an electrolyte was prepared as follows. TiO2 powder was first added to 1 M KCl solution with pH = 2, and the solution was stirred for 1 day. Then the solution pH was adjusted to 7 and left for one more day in order to settle down any non-dissolved particulates. The upper clear solution was finally collected, filtrated and used for electrochemical test. The final solution was transparent and colorless. In order to ascertained the presence of Ti(IV) ions in the solution, H2O2 was added and pH decreased gradually. At pH = 5, the color of the solution changed to light yellow, confirming the presence of Ti (IV) ions, although the exact amount of the ions has not been determined [10]. The cycle performance of the MnO2 electrode in the Ti(IV)-containing 1 M KCl solution (denoted as the S–Ti electrode) is summarized in Fig. 1a, while Fig. 1c presents the voltammograms of selected cycles. Note that, the electrolyte has not been bubbled with oxygen. As shown, the capacitance of the electrode is very stable. It is suggested that the enhanced cycling stability is due to the increased efficiency of reaction (1) via Ti(IV)–Ti(III) mediated charge-transfer scheme: TiðIVÞ þ MnðIIIÞ↔TiðIIIÞ þ MnðIVÞ 2TiðIIIÞ þ 1=2O2 þ H2 O↔2TiðIVÞ þ 2OH
ð2Þ −
ð3Þ
It is also worth mentioning that the S–Ti electrode has a specific capacitance of 118 Fg − 1 at 50 mV s − 1, which is significantly higher than those of the rest electrodes. The mechanism for the capacitance enhancement will be investigated elsewhere. The cycled (≥ 3000 cycles between − 0.2 and 0.4 V) electrodes are subjected to XPS analysis for determining the oxidation state of Mn located at the electrode surfaces, based on the multiple splitting of the Mn 3 s core level spectrum [11]. The Mn valence of the fresh electrode is +3.7, while they are +2.0 and +2.7 for the S–B and S–Ti electrodes, respectively. For comparison, an electrode cycled between 0.0 and 0.8 V showing no capacity fading in the “blank” electrolyte (i.e., without Ti(IV)) is also +2.7 [3]. It is clear that permanent Mn reduction to Mn(II) caused by cycling below 0.0 V is effectively suppressed by the presence of Ti(IV) ions in the electrolyte. XPS analysis also detected Ti(2p3/2) signal at 458.0 eV on the S–Ti electrode, which is characteristic of Ti(III) [12,13]. This confirms the occurrence of Ti(IV) reduction on the surface of MnO2 electrode according to reaction (2). Fig. 2 summarizes the electrochemical characteristics of a MnO2 symmetric full-cell with addition of Ti(IV) ions in the electrolyte. Note that the cell contains only limited amount of electrolyte, as described in Section 2. When the full-cell is operating with a voltage range of 1.2 V, the positive and negative electrodes show essentially symmetric potential range from 0.4 to 1.0 V and from 0.4 to −0.2 V, respectively (Fig. 2a). Fig. 2b presents the CVs of selected cycles of the full-cell operation. The curves exhibit the typical rectangular profiles after initial cycles. The cell capacitance does not fade; instead, it slightly increases along the course up to 10,000 cycles (Fig. 2c). The cause to the capacitance rise may partly be attributed to the change in electrode morphology, from being nanoflake to petal-like, resulting
Fig. 2. (a) Cyclic voltammograms of the positive and negative electrodes of a symmetric MnO2 full-cell containing Ti(IV) ions in electrolyte and cycled within a operating voltage window of 1.2 V; (b) cyclic voltammograms of the full cell containing Ti as additive; (c) capacitance retention (CCo− 1), normalized based on the capacitance of the second cycle (Co), versus cycle number for the full cells either with (○) or without (Δ) Ti(IV) the electrolyte (1 M KCl(aq); scan rate = 50 mV s− 1).
in richer mesophorous structure, as previously observed [3]. For comparison, the cell capacitance of a symmetric cell without Ti(IV) addition in the electrolyte loses 80% of the initial capacitance under the
F. Ataherian, N.-L. Wu / Electrochemistry Communications 13 (2011) 1264–1267
same cycling conditions. This full-cell apparently exhibits a much more complex fading history than that of single electrode under negative polarization (Fig. 1a). This may suggest adverse interaction between the two electrodes in a full-cell configuration when the cell is running over unstable voltage range.
Acknowledgments
4. Conclusion
References
Capacitance fading of MnO2 supercapacitor electrode under negative polarization below 0.0 V arises from irreversible formation of inactive Mn(II) surface species. In view of the fact that this potential limit is close to the onset potential of MnO2-catalyzed ORR, the fading problem has been tackled by passing the accumulated electrons in the oxide electrode to the dissolved oxygen molecules via the ORR in order to suppress the formation of the Mn(II) species. This has successfully been demonstrated by either increasing the dissolved oxygen content in the electrolyte or using the Ti(IV)/Ti(III) redox as a charge-transfer mediator to enhance the electrocatalytic activity of MnO2 for ORR. As a result, the reduction potential limit has been lowered to −0.2 V, and a 1.2 V MnO2 symmetric electrochemical supercapacitor with remarkable capacitance retention over more than 10,000 cycles has been achieved by using Ti(IV)-containing neutral electrolyte (1 M KCl(aq)).
1267
This work is financially supported by the National Science Council of Taiwan, R.O.C. and by the National Taiwan University. One of the authors, F. A., would like to thank NTU post-doctor fellowship (NTU99R40044).
[1] H.Y. Lee, J.B. Goodenough, Journal of Solid State Chemistry 144 (1999) 220. [2] E. Raymundo-Piñero, V. Khomenko, E. Frackowiak, F. Béguin, Journal of the Electrochemical Society 152 (2005) A229. [3] F. Ataherian, K.T. Lee, N.L. Wu, Electrochimica Acta 55 (2010) 7429. [4] F. Ataherian, N.L. Wu, Journal of the Electrochemical Society 158 (2011) 1. [5] S. Komaba, A. Ogata, T. Tsuchikawa, Electrochemistry Communications 10 (2008) 1435. [6] F. Cheng, Y. Su, J. Liang, Z. Tao, J. Chen, Chemistry of Materials 22 (2010) 898. [7] W. Xiao, D. Wang, X.W. Lou, Journal of Physical Chemistry C 114 (2010) 1694. [8] I. Roche, K. Scott, Journal of Applied Electrochemistry 39 (2009) 197. [9] M. Noel, P.N. Anatharaman, H.V.K. Udupa, Journal of Applied Electrochemistry 12 (1982) 291. [10] J. Muhlebach, K. Muller, G. Schwarzenbach, Inorganic Chemistry 9 (1970) 2381. [11] M. Toupin, T. Brousse, D. Bélanger, Chemistry of Materials 14 (2002) 3946. [12] P. Stefanov, M. Shipochka, P. Stefchev, Z. Raicheva, V. Lazarova, L. Spassov, Journal of Physics Conference Series 100 (2008) 1. [13] A. Sandell, M.P. Andersson, M.K.J. Johansson, P.G. Karlsson, Y. Alfredsson, J. Schadt, H. Siegbahn, P. Uvdal, Surface Science 530 (2003) 63.