- Email: [email protected]
High capacity alkaline super-iron boride battery
Electrochimica Acta 52 (2007) 8138–8143
High capacity alkaline super-iron boride battery Xingwen Yu a,∗ , Stuart Licht b a
Department of Chemical and Biological Engineering, The University of British Columbia, 2360 East Mall, Vancouver, BC, Canada V6T 1Z3 b Department of Chemistry, University of Massachusetts, 100 Morrissey Blvd, Boston, MA 02125, USA Received 6 June 2007; received in revised form 12 July 2007; accepted 13 July 2007 Available online 19 July 2007
Abstract A high capacity alkaline redox storage chemistry is explored based on a novel Fe6+ /B2− chemistry. The alkaline anodes based on transition metal borides can deliver exceptionally high electrochemical capacity. Over 3800 mAh/g discharge capacity is obtained for the commercial available vanadium diboride (VB2 ), much higher than the theoretical capacity of commonly used zinc metal (820 mAh/g) alkaline anode. Coupling with the super-iron cathodes, the novel Fe6+ /B2− battery chemistry generates a matched electrochemical potential to the pervasive, conventional MnO2 –Zn battery, but sustains a much higher electrochemical capacity. Stability enhancement of super-iron boride battery is also studied. A zirconia coating effectively prevents both the decomposition of boride anodes and the passivation of Fe(VI) cathodes, and sustains facile both anodic and cathodic charge transfer. Reversibility of boride anodes is demonstrated with TiB2 and VB2 . It is shown that these two boride anodes exhibit the reversibility in a certain extent. © 2007 Elsevier Ltd. All rights reserved. Keywords: Boride anodes; Super-iron boride battery; Zirconia coating; Stabilization; Reversibility
1. Introduction Electrochemical batteries have been widely used as convenient power sources for various portable electronic devices and electric vehicles. Searching for advanced electrode materials is a continuous need for the development of higher energy density batteries. A number of new materials, such as metal hydride and intercalation compounds have been successfully applied to the high performance Ni-MH and Li-ion batteries [1,2]. However, aqueous primary batteries, predominately used in consumer market, are still based on MnO2 cathode and zinc metal anode. The battery chemistry remains almost the same as a century ago. A new battery type, super-iron battery based on the high Fe(VI) cathodic charge storage was reported in 1999 [3]. Followed the primary alkaline super-iron battery, recently, rechargeable thin layer super-iron cathode has been reported [4,5], and a high performance composite Fe(VI)/AgO composite cathode stabilized by a 1% zirconia coating has also been successfully developed [6]. In 2004 it was reported that metal
∗
Corresponding author. Tel.: +1 604 7288895. E-mail address: [email protected] (X. Yu).
0013-4686/$ – see front matter © 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.electacta.2007.07.022
borides could be used as anodic alkaline charge storage materials [7,8]. Representative transition metal borides include TiB2 and VB2 which can store several folds more charge than a zinc anode through multi-electron charge transfer [7]: TiB2 + 12OH− → Ti(amorphous) + 2BO3 3− + 6H2 O + 6e− (1) VB2 + 20OH− → VO4 3− + 2BO3 3− + 10H2 O + 11e−
(2)
However, two obstacles were evident towards implementation of this alkaline boride (MB2 , M = Ti or V) anodic chemistry. There is a significant domain in which the boride anode materials corroded spontaneously generating hydrogen gas, and the electrochemical potential of the boride anodes was more positive than that of zinc. Therefore a boride MnO2 cell was subject to decomposition, and secondly the voltage of this cell was low compared to the electrochemical potential of the pervasive Zn–MnO2 redox chemistry. In our recent communication, we introduced a novel Fe6+ /B2− battery chemistry in which the super-iron (Fe6+ ) cathode provides the requisite additional electrochemical potential for the boride (B2− ) anode [9]. Therefore, the Fe(VI)–MB2 couple generates a similar potential to the Zn–MnO2 battery. In addition, the obstacles of the boride
X. Yu, S. Licht / Electrochimica Acta 52 (2007) 8138–8143
anode decomposition are overcome by the applying a zirconia hydroxide-shuttle overlayer on the anode particle surface [9]. This paper is a continuous study of the Fe6+ /B2− redox chemistry. More boride anodes are studied in super-iron boride batteries. The reversibility of boride anodes is also investigated in alkaline electrolyte by coupling with NiOOH cathode. 2. Experimental Batteries studied in this paper are prepared as 1 cm button cell configuration. The cells are prepared with a saturated KOH electrolyte. Preparation of the cathodes and anodes will be detailed in Section 3. Cathode materials used in this paper include MnO2 (EMD, EraChem K60), NiOOH (taken from the commercial NiMH button cell (Powerstream® )) and lab synthesized K2 FeO4 (97–98% purity, according to our previous publication [10,11]). Anode materials TiB2 (10 m powder), VB2 (325 mesh powder) TaB (325 mesh), TaB2 (325 mesh), MgB2 (325 mesh), CrB2 (325 mesh), CoB2 (325 mesh), Ni2 B (powder, 30 mesh) and LaB6 (powder, 10 m) are from Aldrich® . Conductive medium used in cathode and anode preparation is 1 m graphite (Leico Industries Inc.). Graphite foil served as the current collector is from Alfa Aesar® . The cells are discharged at a constant load (will be indicated in Section 3). In the secondary battery studies (with boride anode and NiOOH cathode), cells are additionally charged at a constant current of 4 mA. Primary discharge or charge/discharge cycling is measured as the cell potential variation over time, and is recorded with LabView Acquisition on a PC, and cumulative discharge, as milliampere hours, determined by subsequent integration.
8139
Table 1 Open circuit potentials (OCP) of various boride anodes alkaline super-iron (K2 FeO4 ) batteries Anode
OCP (V)
TaB TaB2 CoB2 MgB2 CrB2 Ni2 B LaB6 VB2 TiB2
1.65 0.83 1.22 1.50 1.53 1.30 1.16 1.42 1.55
Electrolyte used is saturated KOH.
MnO2 + 21 H2 O+e− → 21 Mn2 O3 +OH− ,
E=0.35 V
(4)
Coupling with super-iron cathode (K2 FeO4 ), open circuit potentials of these super-iron boride batteries are listed in Table 1. In addition to the previously studied TiB2 and VB2 [7,9], in Table 1, only the first two (tantalum) boride salts exhibit a degree of anodic charge storage, each of the other borides did not exhibit significant primary discharge behavior due to their high solubilities in alkaline solution or their reaction with alkaline electrolyte. As previously reported [7,9], the alkaline reduction of the TiB2 anode produced amorphous titanium, and similarly we expect the reduction of TaB2 can yield tantalum. According to the half cell oxidation reaction of TaB2 (FW 202.57 g/mol) or TaB (FW 191.76 g/mol) to a Ta product, the theoretical (intrinsic) anodic reaction in an alkaline medium will yield: TaB2 + 12OH− → Ta + 2BO3 3− + 6H2 O + 6e−
(5)
3. Results and discussion
TaB + 6OH− → Ta + BO3 3− + 3H2 O + 3e−
(6)
3.1. Various boride anodes super-iron batteries
Intrinsic capacity of TaB2 and TaB would accordingly be 793.8 and 419.3 mAh/g, respectively. Discharge profiles of TaB or TaB2 anode, K2 FeO4 cathode button cells are shown in Fig. 1a. Button cells are prepared with excess cathode capacity and discharged at a low current to probe the anode’s limits and characteristics. As seen in the figure, the TaB2 exhibits an anodic storage capacity comparable to the widely used conventional alkaline zinc anode (820 mAh/g). Compared to their intrinsic capacity (793.8 mAh/g and for TaB2 and 419.3 mAh/g for TaB), 88% for TaB2 and 93% for TaB of the coulombic efficiency were obtained for these two anodes. The discharge of TiB2 , rather than TaB2 , to the amorphous metal product is comparable to a significantly higher gravimetric charge storage capacity due to the lighter weight of this metal (47.87 g titanium/mol compared to 180.95 g tantalum/mol), in accord with Eq. (1). As seen in Fig. 1b, a significant advantage of the titanium boride anode is the higher capacity compared to the conventional alkaline zinc anode (820 mAh/g). The TiB2 anode discharge is in excess of 1300 mAh/g at moderate discharge rates (a 3 k load over a 1 cm diameter electrodes) and is in excess of 2000 mAh/g at low discharge rates (a 100 k load discharge). In accord with Eq. (1), and a formula weight, W = 69.5 g/mol, TiB2 , has a net intrinsic 6 electron anodic capacity of 6F/W = 2314 mAh/g (F is the Faraday constant).
Many transition metal borides have thermodynamic parameters and electronic conductivities similar to those of the corresponding transition metals [12,13]. Therefore, from the electrochemical energy conversion viewpoint, transition metal borides may constitute a large class of promising electrochemically active materials for batteries [14,15]. In addition to TiB2 and VB2 , the use of, FeB and CoB as anode materials for batteries has also been previously mentioned [8]. But either FeB or CoB exhibits an anodic capacity less than TiB2 and VB2 [8]. In the study herein, various transition metal borides TaB, TaB2 , MgB2 , CrB2 , CoB2 , Ni2 B, and LaB6 are considered as the anodes for alkaline battery. Similarly as TiB2 and VB2 [7], electrochemical potentials of the borides are more positive than zinc metal, thus the cell voltage of MnO2 –boride batteries are much lower than the pervasive, conventional MnO2 –Zn battery. The alkaline thermodynamic potential of the three electron reduction of super-iron cathodes Fe(VI → III) via Eq. (3), is approximately 250 mV higher than the one electron reduction of MnO2 via Eq. (4), with potentials reported versus SHE (the standard H2 electrode): FeO4 − + 25 H2 O+3e− → 21 Fe2 O3 +5OH− ,
E=0.60 V
(3)
8140
X. Yu, S. Licht / Electrochimica Acta 52 (2007) 8138–8143
Fig. 2. ATR/FT-IR spectra of zirconia-coated and uncoated TiB2 , VB2 and K2 FeO4 compared to the pure ZrO2 spectra. Spectra of 5% coating are included for emphasis; a 1% zirconia coating exhibits evident, but proportionally smaller, 1396 and 1548 cm−1 peaks. Fig. 1. (a) Discharge profile of TaB2 (TaB)–K2 FeO4 button cells. The cells are discharged at constant load of 3000 . (b) Comparison of TiB2 anode cells to cells with conventional Zn or MH alkaline anodes. The anodes (10 mAh intrinsic capacity) are studied in cells with excess intrinsic cathode capacity, discharged under the indicated constant load conditions. The TiB2 anode is prepared with the same wt% composition as the TaB2 , TaB and VB2 anodes. As indicated the cells contain a variety of different cathodes, a (conventional) NiOOH or MnO2 cathode, or a K2 FeO4 (75% K2 FeO4 , 25% graphite) cathode. (c) Discharge of a comparable VB2 anode/K2 FeO4 cathode alkaline cells under the indicated constant ohmic load conditions.
While the titanium diboride exhibits greater intrinsic capacity than the tantalum diboride, a vanadium, compared to a titanium, diboride salt can yield even higher alkaline anodic capacity. Unlike TiB2 , the alkaline VB2 anode undergoes an oxidation of both the boron and the tetravalent transition metal ion, with a net 11 electron process. In accord with Eq. (2), VB2 , will have an intrinsic anodic capacity of 11F/(W = 72.6 g/mol) = 4060 mAh/g, rivaling the high anodic capacity of lithium (3860 mAh/g). As evident in Fig. 1c, this substantial capacity of VB2 is experimentally realized (3800 mAh/g) in the discharge of the alkaline super-iron VB2 cell. The VB2 sustains more efficient higher rate discharge (and at lower polarization loss) than the comparable TiB2 alkaline anode cell. As seen comparing the 3 k discharges in Fig. 1b and c, the TiB2 discharge sustains 56% of the intrinsic capacity, whereas the VB2 sustains 91% of the intrinsic capacity (which increases to 94% of the intrinsic 11e− capacity during a 10 k discharge). 3.2. Zirconia coating stabilized super-iron boride battery Most transition borides are unstable in alkaline media due to their high reaction activities on contact with alkaline solution. For example, to our experience, TiB2 visibly reacts with KOH electrolyte and evolves H2 gas. This chemical reaction is not only a loss of the electrochemical capacity of TiB2 anode, but
also induces the swollen or even crack problems for a sealed battery during storage due to the evolved gas. Zirconia is extremely stable in aqueous alkaline media [16,17], and based on our previous experience (Mn coating [18]), we recently developed a novel zirconia overlayer, which is derived from an organic soluble zirconium salt (ZrCl4 ) via an organic medium [6]. In this study, high capacity boride anodes are also modified with this novel zirconia coating methodology. Pure ZrO2 is prepared (as a colloid) for comparison. Similar to the coated cathode materials [6], attenuated total reflectance Fourier transform infrared (ATR/FTIR) (Fig. 2) analysis reveals the zirconia 1396 and 1548 cm−1 peaks on the coated TiB2 and VB2 coincide with the absorption spectra of pure ZrO2 /Zr(OH)4 depending on the extent of hydration [6,19] ZrO2 + 2H2 O ⇔ Zr(OH)4
(7)
As recently demonstrated, this zirconia overlayer provides an ionic conductive, alkaline stable coating which is capable of mediating hydroxide transport from the electrolyte to the electrode [6,9]. A zirconia modified K2 FeO4 /AgO composite cathode exhibits longevity and high charge storage capacity [6]. In a K2 FeO4 –TiB2 battery chemistry system, stability of not only the K2 FeO4 cathode but also the TiB2 anode dramatically improves with this zirconia coating. As seen in Fig. 3a, after storage for 7 days at room temperature, the uncoated superiron titanium boride cell generates only 10–15% of the 3 k discharge capacity of the fresh cell. One hundred percent of the charge capacity is retained, when zirconia-coated super-iron and zirconia-coated boride are utilized. In lieu of the uncoated electrodes, if either anode or cathode (but not both) is coated, then a fraction, but not all, of the charge capacity is lost. Also evident in the figure, the zirconia-coated super-iron vanadium boride cell also retains its substantial charge capacity after storage. The fundamental chemistry of conventional alkaline primary [20] and metal hydride rechargeable [1] batteries are understood, and con-
X. Yu, S. Licht / Electrochimica Acta 52 (2007) 8138–8143
8141
electrochemical storage chemistry was quite well understood. The full cell reaction of this battery system can be generalized as [21] Zn + 2MnO2 + H2 O → Mn2 O3 + Zn(OH)2
(8)
The range from maximum experimental (<160 mAh/g) to theoretical (2F per Zn + 2MnO2 = 222 mAh/g) charge storage capacities of the conventional alkaline Zn–MnO2 cell are shown as dashed vertical lines in Fig. 3a. The discharge of the complete super-iron boride redox chemistry is investigated in Fig. 3a using cells with balanced anode and cathode capacity. The theoretical capacities for the complete, balanced super-iron boride batteries can be calculated based on the intrinsic capacity of the anode and cathode components in accordance with the full cell reactions: TiB2 + 2FeO4 2− + 2OH− → Ti + Fe2 O3 + 2BO3 2− + H2 O (9) 6VB2 + 22FeO4 2− + 10OH− → 6VO4 3− + 11Fe2 O3 + 12BO3 2− + 5H2 O
(10)
tinue to be of widespread interest. The charge retention for these super-iron boride cells is already comparable to that observed in early alkaline primary cells, and is better than that of contemporary alkaline rechargeable cells. We observe that the vanadium boride anode exhibits higher stability than the titanium boride anode. Accelerated long-term stability is measured at higher temperature. As shown in Fig. 3b, without the zirconia overlayer, after 1 week storage the vanadium boride anode retains 90% of the original charge capacity at 45 ◦ C (100% after 1 week with the zirconia coating), and 65% of the charge capacity at 70 ◦ C (85% after 1 week with the zirconia coating).
Therefore, theoretical capacities of super-iron TiB2 (6F per TiB2 + 2K2 FeO4 ) and super-iron VB2 (11F per VB2 + (11/3)K2 FeO4 ) are respectively 345 and 369 mAh/g, much higher than that of MnO2 –Zn battery. In addition, as seen in Fig. 3a, the super-iron titanium boride cell combined anode and cathode capacity is experimentally in excess of 250 mAh/g, and that of the super-iron vanadium boride cell is over 310 mAh/g, which is almost twofold higher than that of the conventional alkaline battery chemistry (MnO2 /Zn). The theoretical and practical capacities of MnO2 –Zn, K2 FeO4 –TiB2 and K2 FeO4 –VB2 batteries, as well as the intrinsic capacities of their anode and cathode components are summarized in Table 2. As seen in Table 2, the super-iron boride chemistry exhibits significantly higher charge storage than conventional MnO2 –Zn alkaline primary storage chemistry in both theoretical and practical aspects. A further optimization of both the boride and super-iron salt particle size, and variation of the zirconia coating, should further enhance cell performance. In addition, alternate metal borides, as well as alternate super-irons will also affect characteristics of the super-iron boride cell capacity.
3.3. High capacity of super-iron boride battery
3.4. Reversibility of TiB2 and VB2 boride anodes
Aqueous Zn–MnO2 redox charge storage chemistry has been established for over a century and the mechanism of this primary
Investigation of the rechargeable character of alkaline titanium boride anode is probed using a secondary coin cell
Fig. 3. (a) Capacity (anode + cathode) of the super-iron boride alkaline battery to the conventional (MnO2 /Zn) alkaline battery. The super-iron boride cell contains either a TiB2 , or a VB2 anode, as indicated in the figure. The cathode is 76.5% K2 FeO4 , 8.5% AgO, 5% KOH and 10% 1 m graphite. Charge retention (stability) of the cells is compared freshly discharged, and after 1 week storage, with, or without, a 1% zirconia coating applied to the Fe(VI) or boride salts. (b) Enhancement of stability of VB2 anode by ZrO2 coating at elevated temperature. The button cells are prepared with MnO2 cathode and VB2 anode and discharged at constant load of 3000 .
Table 2 Capacity comparison of MnO2 –Zn, K2 FeO4 –TiB2 and K2 FeO4 –VB2 batteries
MnO2 –Zn K2 FeO4 –TiB2 K2 FeO4 –VB2
Intrinsic capacity of the anode (mAh/g)
Intrinsic capacity of the cathode (mAh/g)
Theoretical capacity of full cell (mAh/g)
Practical capacity of full cell (mAh/g)
820 2314 4060
308 406 406
222 345 369
<160 >250 >310
8142
X. Yu, S. Licht / Electrochimica Acta 52 (2007) 8138–8143
Fig. 4. (a) Charge/discharge of alkaline TiB2 anode–NiOOH secondary button cell. In each cycle, button cells are discharged at a constant current of 0.4 mA in coin cell configuration to a cutoff voltage of 0.6 V. Then button cells are charged at a constant current of 4 mA for 50 min. (b) Charge/discharge of alkaline VB2 anode–NiOOH secondary button cell. In each cycle, button cells are discharged at a constant current of 0.4 mA in coin cell configuration to a cutoff voltage of 0.8 V. Then button cells are charged at a constant current of 4 mA for 160 min.
configuration with an excess capacity NiOOH cathode, and a thin layer TiB2 anode. The anode is composed of 75 wt% of the TiB2 and 25 wt% 1 m graphite, prepared with 2.8 mg TiB2 /graphite powder and be compressed to a piece of graphite foil served as the current collector. In each cycle, the button cell is first discharged to a cutoff voltage of 0.6 V at constant current of 0.4 mA. Following discharge, the button cell is charged at a constant current of 4 mA for 50 min. As seen in Fig. 4a, discharge capacity decreases to about 50% of the original discharge capacity in the first cycle and then continuously decreases slowly after the second cycle. Investigation of the reversibility of alkaline vanadium boride anode is carried out as the same way as TiB2 , in a coin cell configuration with an excess capacity NiOOH cathode, and a thin layer VB2 anode. Preparation of the VB2 anode is the same wt% as TiB2 anode with 2.8 mg VB2 /graphite powder. As the intrinsic capacity of VB2 is much higher than that of TiB2 , the NiOOH–VB2 button cell is charged at a constant current of 4 mA for 160 min in each cycle, after being discharged to a cutoff voltage of 0.8 V at constant current of 0.4 mA. As seen in Fig. 4b, reversibility of VB2 anode is not as good as that of TiB2 by comparing with Fig. 4a. Fig. 4 reflects that TiB2 and VB2 anodes are rechargeable in a certain extent. Obviously, both of the two anodes do not exhibit their full 6e− (TiB2 ) and 11e− (VB2 ) charge transfer during the reversible cycles. The partial reversibility of these two anodes is supposed to be attributed to the composition and structure change during the recharge cycles. Ongoing reversibility studies on the TiB2 and VB2 anodes include the composition and structure analyses of TiB2 and VB2 after each charge/discharge cycle. Important issues to be addressed are the identification of the valent state change of titanium, vanadium or boron during the charge/discharge cycles and the determination of the processes that take place in the anodes during the electrochemical cycles. However, more efforts need to be accomplished to get a repeatable reversibility and to understand the rechargeable mechanism of these novel anode materials.
4. Conclusions A novel high capacity battery chemistry based on super-iron cathode and transition metal boride anode is presented. This Fe6+ /B2− redox chemistry generates a matched electrochemical potential to the pervasive, conventional MnO2 –Zn battery chemistry, but sustains a much higher electrochemical capacity. Stability enhancement of super-iron boride battery is also presented. A low level zirconia coating effectively prevents both the decomposition of boride anodes and the passivation of Fe(VI) cathodes, and sustains facile both anodic and cathodic charge transfer. Boride anodes TiB2 and VB2 also exhibit the reversibility in a certain extent. The super-iron boride chemistry exhibits significantly higher charge storage than conventional alkaline primary storage chemistry. Resources to prepare super-iron salts are plentiful and clean. Iron is the second most abundant metal in the earth’s core, and the Fe(VI) reduction product is non-toxic ferric oxide. Transition metal borides are a large family of interesting compounds for electrochemical charge storage. Most of the transition metal borides are easy to synthesize and are commercial available. Alternate metal borides, as well as alternate super-irons will also affect characteristics of the super-iron boride cell capacity. Further understanding of the charge transfer of the super-iron salts and transition metal borides will also impact charge transfer, retention, capacity of the new super-iron boride chemistry. References [1] [2] [3] [4] [5] [6] [7]
S.R. Ovshinsky, M.A. Fetcenko, J. Ross, Science 260 (1993) 176. G.M. Julien, Mater. Sci. Eng. R 40 (2003) 47. S. Licht, B. Wang, S. Ghosh, Science 285 (1999) 1039. S. Licht, Ran Tel-Vered, Chem. Commun. 6 (2004) 628. S. Licht, C. DeAlwis, J. Phys. Chem. B 110 (2006) 12394. S. Licht, X. Yu, D. Zheng, Chem. Commun. 41 (2006) 4341. H.X. Yang, Y.D. Wang, X.P. Ai, C.S. Cha, Electrochem. Solid-State Lett. 7 (2004) A212.
X. Yu, S. Licht / Electrochimica Acta 52 (2007) 8138–8143 [8] Y.D. Wang, X.P. Ai, Y.L. Cao, H.X. Yang, Electrochem. Commun. 6 (2004) 780. [9] S. Licht, X. Yu, D. Qu, Chem. Commun. (2007), doi:10.1039/b701629h. [10] S. Licht, V. Naschitz, B. Liu, S. Ghosh, N. Halperin, L. Halperin, D. Rozen, J. Power Sources 99 (2001) 7. [11] S. Licht, V. Naschitz, L. Halperin, N. Halperin, L. Lin, J. Chen, S. Ghosh, B. Liu, J. Power Sources 101 (2001) 167. [12] A.J. Bard, R. Parsons, J. Jordan, Standard Potentials in Aqueous Solutions, Marcel Dekker, New York, 1985. [13] W.M. Latimer, The Oxidation State of the Elements and Their Potentials in Aqueous Solutions, 2nd ed., Prentice-Hall, New York, 1952. [14] S. Amendola, US Patent 5,948,558 (1999).
8143
[15] S. Amendola, US Patent 6,468,694 (2002). [16] W.F. Linke, Solubilities of Inorganic and Metal-Organic Compounds, 4th ed., Van Nostrand, Princeton, NJ, 1958. [17] S. Hettiarachchi, P. Kedzierzawski, D.D. Macdonald, J. Electrochem. Soc. 132 (1985) 1866. [18] S. Licht, V. Naschitz, B. Liu, S. Ghosh, N. Halperin, L. Halperin, D. Rozen, Electrochem. Commun. 1 (1999) 527. [19] X. Fang, C. Yang, J. Chen, J. Chin. Ceram. Soc. 6 (1998) 732. [20] Z.-R. Tian, W. Tong, J.-Y. Wang, N.-G. Duan, V. Krishnan, S.L. Suib, Science 276 (1997) 926. [21] D. Linden, T.B. Reddy, Handbook of Batteries, 3rd ed., McGraw-Hill, New York, 2002.
High capacity alkaline super-iron boride battery Xingwen Yu a,∗ , Stuart Licht b a
Department of Chemical and Biological Engineering, The University of British Columbia, 2360 East Mall, Vancouver, BC, Canada V6T 1Z3 b Department of Chemistry, University of Massachusetts, 100 Morrissey Blvd, Boston, MA 02125, USA Received 6 June 2007; received in revised form 12 July 2007; accepted 13 July 2007 Available online 19 July 2007
Abstract A high capacity alkaline redox storage chemistry is explored based on a novel Fe6+ /B2− chemistry. The alkaline anodes based on transition metal borides can deliver exceptionally high electrochemical capacity. Over 3800 mAh/g discharge capacity is obtained for the commercial available vanadium diboride (VB2 ), much higher than the theoretical capacity of commonly used zinc metal (820 mAh/g) alkaline anode. Coupling with the super-iron cathodes, the novel Fe6+ /B2− battery chemistry generates a matched electrochemical potential to the pervasive, conventional MnO2 –Zn battery, but sustains a much higher electrochemical capacity. Stability enhancement of super-iron boride battery is also studied. A zirconia coating effectively prevents both the decomposition of boride anodes and the passivation of Fe(VI) cathodes, and sustains facile both anodic and cathodic charge transfer. Reversibility of boride anodes is demonstrated with TiB2 and VB2 . It is shown that these two boride anodes exhibit the reversibility in a certain extent. © 2007 Elsevier Ltd. All rights reserved. Keywords: Boride anodes; Super-iron boride battery; Zirconia coating; Stabilization; Reversibility
1. Introduction Electrochemical batteries have been widely used as convenient power sources for various portable electronic devices and electric vehicles. Searching for advanced electrode materials is a continuous need for the development of higher energy density batteries. A number of new materials, such as metal hydride and intercalation compounds have been successfully applied to the high performance Ni-MH and Li-ion batteries [1,2]. However, aqueous primary batteries, predominately used in consumer market, are still based on MnO2 cathode and zinc metal anode. The battery chemistry remains almost the same as a century ago. A new battery type, super-iron battery based on the high Fe(VI) cathodic charge storage was reported in 1999 [3]. Followed the primary alkaline super-iron battery, recently, rechargeable thin layer super-iron cathode has been reported [4,5], and a high performance composite Fe(VI)/AgO composite cathode stabilized by a 1% zirconia coating has also been successfully developed [6]. In 2004 it was reported that metal
∗
Corresponding author. Tel.: +1 604 7288895. E-mail address: [email protected] (X. Yu).
0013-4686/$ – see front matter © 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.electacta.2007.07.022
borides could be used as anodic alkaline charge storage materials [7,8]. Representative transition metal borides include TiB2 and VB2 which can store several folds more charge than a zinc anode through multi-electron charge transfer [7]: TiB2 + 12OH− → Ti(amorphous) + 2BO3 3− + 6H2 O + 6e− (1) VB2 + 20OH− → VO4 3− + 2BO3 3− + 10H2 O + 11e−
(2)
However, two obstacles were evident towards implementation of this alkaline boride (MB2 , M = Ti or V) anodic chemistry. There is a significant domain in which the boride anode materials corroded spontaneously generating hydrogen gas, and the electrochemical potential of the boride anodes was more positive than that of zinc. Therefore a boride MnO2 cell was subject to decomposition, and secondly the voltage of this cell was low compared to the electrochemical potential of the pervasive Zn–MnO2 redox chemistry. In our recent communication, we introduced a novel Fe6+ /B2− battery chemistry in which the super-iron (Fe6+ ) cathode provides the requisite additional electrochemical potential for the boride (B2− ) anode [9]. Therefore, the Fe(VI)–MB2 couple generates a similar potential to the Zn–MnO2 battery. In addition, the obstacles of the boride
X. Yu, S. Licht / Electrochimica Acta 52 (2007) 8138–8143
anode decomposition are overcome by the applying a zirconia hydroxide-shuttle overlayer on the anode particle surface [9]. This paper is a continuous study of the Fe6+ /B2− redox chemistry. More boride anodes are studied in super-iron boride batteries. The reversibility of boride anodes is also investigated in alkaline electrolyte by coupling with NiOOH cathode. 2. Experimental Batteries studied in this paper are prepared as 1 cm button cell configuration. The cells are prepared with a saturated KOH electrolyte. Preparation of the cathodes and anodes will be detailed in Section 3. Cathode materials used in this paper include MnO2 (EMD, EraChem K60), NiOOH (taken from the commercial NiMH button cell (Powerstream® )) and lab synthesized K2 FeO4 (97–98% purity, according to our previous publication [10,11]). Anode materials TiB2 (10 m powder), VB2 (325 mesh powder) TaB (325 mesh), TaB2 (325 mesh), MgB2 (325 mesh), CrB2 (325 mesh), CoB2 (325 mesh), Ni2 B (powder, 30 mesh) and LaB6 (powder, 10 m) are from Aldrich® . Conductive medium used in cathode and anode preparation is 1 m graphite (Leico Industries Inc.). Graphite foil served as the current collector is from Alfa Aesar® . The cells are discharged at a constant load (will be indicated in Section 3). In the secondary battery studies (with boride anode and NiOOH cathode), cells are additionally charged at a constant current of 4 mA. Primary discharge or charge/discharge cycling is measured as the cell potential variation over time, and is recorded with LabView Acquisition on a PC, and cumulative discharge, as milliampere hours, determined by subsequent integration.
8139
Table 1 Open circuit potentials (OCP) of various boride anodes alkaline super-iron (K2 FeO4 ) batteries Anode
OCP (V)
TaB TaB2 CoB2 MgB2 CrB2 Ni2 B LaB6 VB2 TiB2
1.65 0.83 1.22 1.50 1.53 1.30 1.16 1.42 1.55
Electrolyte used is saturated KOH.
MnO2 + 21 H2 O+e− → 21 Mn2 O3 +OH− ,
E=0.35 V
(4)
Coupling with super-iron cathode (K2 FeO4 ), open circuit potentials of these super-iron boride batteries are listed in Table 1. In addition to the previously studied TiB2 and VB2 [7,9], in Table 1, only the first two (tantalum) boride salts exhibit a degree of anodic charge storage, each of the other borides did not exhibit significant primary discharge behavior due to their high solubilities in alkaline solution or their reaction with alkaline electrolyte. As previously reported [7,9], the alkaline reduction of the TiB2 anode produced amorphous titanium, and similarly we expect the reduction of TaB2 can yield tantalum. According to the half cell oxidation reaction of TaB2 (FW 202.57 g/mol) or TaB (FW 191.76 g/mol) to a Ta product, the theoretical (intrinsic) anodic reaction in an alkaline medium will yield: TaB2 + 12OH− → Ta + 2BO3 3− + 6H2 O + 6e−
(5)
3. Results and discussion
TaB + 6OH− → Ta + BO3 3− + 3H2 O + 3e−
(6)
3.1. Various boride anodes super-iron batteries
Intrinsic capacity of TaB2 and TaB would accordingly be 793.8 and 419.3 mAh/g, respectively. Discharge profiles of TaB or TaB2 anode, K2 FeO4 cathode button cells are shown in Fig. 1a. Button cells are prepared with excess cathode capacity and discharged at a low current to probe the anode’s limits and characteristics. As seen in the figure, the TaB2 exhibits an anodic storage capacity comparable to the widely used conventional alkaline zinc anode (820 mAh/g). Compared to their intrinsic capacity (793.8 mAh/g and for TaB2 and 419.3 mAh/g for TaB), 88% for TaB2 and 93% for TaB of the coulombic efficiency were obtained for these two anodes. The discharge of TiB2 , rather than TaB2 , to the amorphous metal product is comparable to a significantly higher gravimetric charge storage capacity due to the lighter weight of this metal (47.87 g titanium/mol compared to 180.95 g tantalum/mol), in accord with Eq. (1). As seen in Fig. 1b, a significant advantage of the titanium boride anode is the higher capacity compared to the conventional alkaline zinc anode (820 mAh/g). The TiB2 anode discharge is in excess of 1300 mAh/g at moderate discharge rates (a 3 k load over a 1 cm diameter electrodes) and is in excess of 2000 mAh/g at low discharge rates (a 100 k load discharge). In accord with Eq. (1), and a formula weight, W = 69.5 g/mol, TiB2 , has a net intrinsic 6 electron anodic capacity of 6F/W = 2314 mAh/g (F is the Faraday constant).
Many transition metal borides have thermodynamic parameters and electronic conductivities similar to those of the corresponding transition metals [12,13]. Therefore, from the electrochemical energy conversion viewpoint, transition metal borides may constitute a large class of promising electrochemically active materials for batteries [14,15]. In addition to TiB2 and VB2 , the use of, FeB and CoB as anode materials for batteries has also been previously mentioned [8]. But either FeB or CoB exhibits an anodic capacity less than TiB2 and VB2 [8]. In the study herein, various transition metal borides TaB, TaB2 , MgB2 , CrB2 , CoB2 , Ni2 B, and LaB6 are considered as the anodes for alkaline battery. Similarly as TiB2 and VB2 [7], electrochemical potentials of the borides are more positive than zinc metal, thus the cell voltage of MnO2 –boride batteries are much lower than the pervasive, conventional MnO2 –Zn battery. The alkaline thermodynamic potential of the three electron reduction of super-iron cathodes Fe(VI → III) via Eq. (3), is approximately 250 mV higher than the one electron reduction of MnO2 via Eq. (4), with potentials reported versus SHE (the standard H2 electrode): FeO4 − + 25 H2 O+3e− → 21 Fe2 O3 +5OH− ,
E=0.60 V
(3)
8140
X. Yu, S. Licht / Electrochimica Acta 52 (2007) 8138–8143
Fig. 2. ATR/FT-IR spectra of zirconia-coated and uncoated TiB2 , VB2 and K2 FeO4 compared to the pure ZrO2 spectra. Spectra of 5% coating are included for emphasis; a 1% zirconia coating exhibits evident, but proportionally smaller, 1396 and 1548 cm−1 peaks. Fig. 1. (a) Discharge profile of TaB2 (TaB)–K2 FeO4 button cells. The cells are discharged at constant load of 3000 . (b) Comparison of TiB2 anode cells to cells with conventional Zn or MH alkaline anodes. The anodes (10 mAh intrinsic capacity) are studied in cells with excess intrinsic cathode capacity, discharged under the indicated constant load conditions. The TiB2 anode is prepared with the same wt% composition as the TaB2 , TaB and VB2 anodes. As indicated the cells contain a variety of different cathodes, a (conventional) NiOOH or MnO2 cathode, or a K2 FeO4 (75% K2 FeO4 , 25% graphite) cathode. (c) Discharge of a comparable VB2 anode/K2 FeO4 cathode alkaline cells under the indicated constant ohmic load conditions.
While the titanium diboride exhibits greater intrinsic capacity than the tantalum diboride, a vanadium, compared to a titanium, diboride salt can yield even higher alkaline anodic capacity. Unlike TiB2 , the alkaline VB2 anode undergoes an oxidation of both the boron and the tetravalent transition metal ion, with a net 11 electron process. In accord with Eq. (2), VB2 , will have an intrinsic anodic capacity of 11F/(W = 72.6 g/mol) = 4060 mAh/g, rivaling the high anodic capacity of lithium (3860 mAh/g). As evident in Fig. 1c, this substantial capacity of VB2 is experimentally realized (3800 mAh/g) in the discharge of the alkaline super-iron VB2 cell. The VB2 sustains more efficient higher rate discharge (and at lower polarization loss) than the comparable TiB2 alkaline anode cell. As seen comparing the 3 k discharges in Fig. 1b and c, the TiB2 discharge sustains 56% of the intrinsic capacity, whereas the VB2 sustains 91% of the intrinsic capacity (which increases to 94% of the intrinsic 11e− capacity during a 10 k discharge). 3.2. Zirconia coating stabilized super-iron boride battery Most transition borides are unstable in alkaline media due to their high reaction activities on contact with alkaline solution. For example, to our experience, TiB2 visibly reacts with KOH electrolyte and evolves H2 gas. This chemical reaction is not only a loss of the electrochemical capacity of TiB2 anode, but
also induces the swollen or even crack problems for a sealed battery during storage due to the evolved gas. Zirconia is extremely stable in aqueous alkaline media [16,17], and based on our previous experience (Mn coating [18]), we recently developed a novel zirconia overlayer, which is derived from an organic soluble zirconium salt (ZrCl4 ) via an organic medium [6]. In this study, high capacity boride anodes are also modified with this novel zirconia coating methodology. Pure ZrO2 is prepared (as a colloid) for comparison. Similar to the coated cathode materials [6], attenuated total reflectance Fourier transform infrared (ATR/FTIR) (Fig. 2) analysis reveals the zirconia 1396 and 1548 cm−1 peaks on the coated TiB2 and VB2 coincide with the absorption spectra of pure ZrO2 /Zr(OH)4 depending on the extent of hydration [6,19] ZrO2 + 2H2 O ⇔ Zr(OH)4
(7)
As recently demonstrated, this zirconia overlayer provides an ionic conductive, alkaline stable coating which is capable of mediating hydroxide transport from the electrolyte to the electrode [6,9]. A zirconia modified K2 FeO4 /AgO composite cathode exhibits longevity and high charge storage capacity [6]. In a K2 FeO4 –TiB2 battery chemistry system, stability of not only the K2 FeO4 cathode but also the TiB2 anode dramatically improves with this zirconia coating. As seen in Fig. 3a, after storage for 7 days at room temperature, the uncoated superiron titanium boride cell generates only 10–15% of the 3 k discharge capacity of the fresh cell. One hundred percent of the charge capacity is retained, when zirconia-coated super-iron and zirconia-coated boride are utilized. In lieu of the uncoated electrodes, if either anode or cathode (but not both) is coated, then a fraction, but not all, of the charge capacity is lost. Also evident in the figure, the zirconia-coated super-iron vanadium boride cell also retains its substantial charge capacity after storage. The fundamental chemistry of conventional alkaline primary [20] and metal hydride rechargeable [1] batteries are understood, and con-
X. Yu, S. Licht / Electrochimica Acta 52 (2007) 8138–8143
8141
electrochemical storage chemistry was quite well understood. The full cell reaction of this battery system can be generalized as [21] Zn + 2MnO2 + H2 O → Mn2 O3 + Zn(OH)2
(8)
The range from maximum experimental (<160 mAh/g) to theoretical (2F per Zn + 2MnO2 = 222 mAh/g) charge storage capacities of the conventional alkaline Zn–MnO2 cell are shown as dashed vertical lines in Fig. 3a. The discharge of the complete super-iron boride redox chemistry is investigated in Fig. 3a using cells with balanced anode and cathode capacity. The theoretical capacities for the complete, balanced super-iron boride batteries can be calculated based on the intrinsic capacity of the anode and cathode components in accordance with the full cell reactions: TiB2 + 2FeO4 2− + 2OH− → Ti + Fe2 O3 + 2BO3 2− + H2 O (9) 6VB2 + 22FeO4 2− + 10OH− → 6VO4 3− + 11Fe2 O3 + 12BO3 2− + 5H2 O
(10)
tinue to be of widespread interest. The charge retention for these super-iron boride cells is already comparable to that observed in early alkaline primary cells, and is better than that of contemporary alkaline rechargeable cells. We observe that the vanadium boride anode exhibits higher stability than the titanium boride anode. Accelerated long-term stability is measured at higher temperature. As shown in Fig. 3b, without the zirconia overlayer, after 1 week storage the vanadium boride anode retains 90% of the original charge capacity at 45 ◦ C (100% after 1 week with the zirconia coating), and 65% of the charge capacity at 70 ◦ C (85% after 1 week with the zirconia coating).
Therefore, theoretical capacities of super-iron TiB2 (6F per TiB2 + 2K2 FeO4 ) and super-iron VB2 (11F per VB2 + (11/3)K2 FeO4 ) are respectively 345 and 369 mAh/g, much higher than that of MnO2 –Zn battery. In addition, as seen in Fig. 3a, the super-iron titanium boride cell combined anode and cathode capacity is experimentally in excess of 250 mAh/g, and that of the super-iron vanadium boride cell is over 310 mAh/g, which is almost twofold higher than that of the conventional alkaline battery chemistry (MnO2 /Zn). The theoretical and practical capacities of MnO2 –Zn, K2 FeO4 –TiB2 and K2 FeO4 –VB2 batteries, as well as the intrinsic capacities of their anode and cathode components are summarized in Table 2. As seen in Table 2, the super-iron boride chemistry exhibits significantly higher charge storage than conventional MnO2 –Zn alkaline primary storage chemistry in both theoretical and practical aspects. A further optimization of both the boride and super-iron salt particle size, and variation of the zirconia coating, should further enhance cell performance. In addition, alternate metal borides, as well as alternate super-irons will also affect characteristics of the super-iron boride cell capacity.
3.3. High capacity of super-iron boride battery
3.4. Reversibility of TiB2 and VB2 boride anodes
Aqueous Zn–MnO2 redox charge storage chemistry has been established for over a century and the mechanism of this primary
Investigation of the rechargeable character of alkaline titanium boride anode is probed using a secondary coin cell
Fig. 3. (a) Capacity (anode + cathode) of the super-iron boride alkaline battery to the conventional (MnO2 /Zn) alkaline battery. The super-iron boride cell contains either a TiB2 , or a VB2 anode, as indicated in the figure. The cathode is 76.5% K2 FeO4 , 8.5% AgO, 5% KOH and 10% 1 m graphite. Charge retention (stability) of the cells is compared freshly discharged, and after 1 week storage, with, or without, a 1% zirconia coating applied to the Fe(VI) or boride salts. (b) Enhancement of stability of VB2 anode by ZrO2 coating at elevated temperature. The button cells are prepared with MnO2 cathode and VB2 anode and discharged at constant load of 3000 .
Table 2 Capacity comparison of MnO2 –Zn, K2 FeO4 –TiB2 and K2 FeO4 –VB2 batteries
MnO2 –Zn K2 FeO4 –TiB2 K2 FeO4 –VB2
Intrinsic capacity of the anode (mAh/g)
Intrinsic capacity of the cathode (mAh/g)
Theoretical capacity of full cell (mAh/g)
Practical capacity of full cell (mAh/g)
820 2314 4060
308 406 406
222 345 369
<160 >250 >310
8142
X. Yu, S. Licht / Electrochimica Acta 52 (2007) 8138–8143
Fig. 4. (a) Charge/discharge of alkaline TiB2 anode–NiOOH secondary button cell. In each cycle, button cells are discharged at a constant current of 0.4 mA in coin cell configuration to a cutoff voltage of 0.6 V. Then button cells are charged at a constant current of 4 mA for 50 min. (b) Charge/discharge of alkaline VB2 anode–NiOOH secondary button cell. In each cycle, button cells are discharged at a constant current of 0.4 mA in coin cell configuration to a cutoff voltage of 0.8 V. Then button cells are charged at a constant current of 4 mA for 160 min.
configuration with an excess capacity NiOOH cathode, and a thin layer TiB2 anode. The anode is composed of 75 wt% of the TiB2 and 25 wt% 1 m graphite, prepared with 2.8 mg TiB2 /graphite powder and be compressed to a piece of graphite foil served as the current collector. In each cycle, the button cell is first discharged to a cutoff voltage of 0.6 V at constant current of 0.4 mA. Following discharge, the button cell is charged at a constant current of 4 mA for 50 min. As seen in Fig. 4a, discharge capacity decreases to about 50% of the original discharge capacity in the first cycle and then continuously decreases slowly after the second cycle. Investigation of the reversibility of alkaline vanadium boride anode is carried out as the same way as TiB2 , in a coin cell configuration with an excess capacity NiOOH cathode, and a thin layer VB2 anode. Preparation of the VB2 anode is the same wt% as TiB2 anode with 2.8 mg VB2 /graphite powder. As the intrinsic capacity of VB2 is much higher than that of TiB2 , the NiOOH–VB2 button cell is charged at a constant current of 4 mA for 160 min in each cycle, after being discharged to a cutoff voltage of 0.8 V at constant current of 0.4 mA. As seen in Fig. 4b, reversibility of VB2 anode is not as good as that of TiB2 by comparing with Fig. 4a. Fig. 4 reflects that TiB2 and VB2 anodes are rechargeable in a certain extent. Obviously, both of the two anodes do not exhibit their full 6e− (TiB2 ) and 11e− (VB2 ) charge transfer during the reversible cycles. The partial reversibility of these two anodes is supposed to be attributed to the composition and structure change during the recharge cycles. Ongoing reversibility studies on the TiB2 and VB2 anodes include the composition and structure analyses of TiB2 and VB2 after each charge/discharge cycle. Important issues to be addressed are the identification of the valent state change of titanium, vanadium or boron during the charge/discharge cycles and the determination of the processes that take place in the anodes during the electrochemical cycles. However, more efforts need to be accomplished to get a repeatable reversibility and to understand the rechargeable mechanism of these novel anode materials.
4. Conclusions A novel high capacity battery chemistry based on super-iron cathode and transition metal boride anode is presented. This Fe6+ /B2− redox chemistry generates a matched electrochemical potential to the pervasive, conventional MnO2 –Zn battery chemistry, but sustains a much higher electrochemical capacity. Stability enhancement of super-iron boride battery is also presented. A low level zirconia coating effectively prevents both the decomposition of boride anodes and the passivation of Fe(VI) cathodes, and sustains facile both anodic and cathodic charge transfer. Boride anodes TiB2 and VB2 also exhibit the reversibility in a certain extent. The super-iron boride chemistry exhibits significantly higher charge storage than conventional alkaline primary storage chemistry. Resources to prepare super-iron salts are plentiful and clean. Iron is the second most abundant metal in the earth’s core, and the Fe(VI) reduction product is non-toxic ferric oxide. Transition metal borides are a large family of interesting compounds for electrochemical charge storage. Most of the transition metal borides are easy to synthesize and are commercial available. Alternate metal borides, as well as alternate super-irons will also affect characteristics of the super-iron boride cell capacity. Further understanding of the charge transfer of the super-iron salts and transition metal borides will also impact charge transfer, retention, capacity of the new super-iron boride chemistry. References [1] [2] [3] [4] [5] [6] [7]
S.R. Ovshinsky, M.A. Fetcenko, J. Ross, Science 260 (1993) 176. G.M. Julien, Mater. Sci. Eng. R 40 (2003) 47. S. Licht, B. Wang, S. Ghosh, Science 285 (1999) 1039. S. Licht, Ran Tel-Vered, Chem. Commun. 6 (2004) 628. S. Licht, C. DeAlwis, J. Phys. Chem. B 110 (2006) 12394. S. Licht, X. Yu, D. Zheng, Chem. Commun. 41 (2006) 4341. H.X. Yang, Y.D. Wang, X.P. Ai, C.S. Cha, Electrochem. Solid-State Lett. 7 (2004) A212.
X. Yu, S. Licht / Electrochimica Acta 52 (2007) 8138–8143 [8] Y.D. Wang, X.P. Ai, Y.L. Cao, H.X. Yang, Electrochem. Commun. 6 (2004) 780. [9] S. Licht, X. Yu, D. Qu, Chem. Commun. (2007), doi:10.1039/b701629h. [10] S. Licht, V. Naschitz, B. Liu, S. Ghosh, N. Halperin, L. Halperin, D. Rozen, J. Power Sources 99 (2001) 7. [11] S. Licht, V. Naschitz, L. Halperin, N. Halperin, L. Lin, J. Chen, S. Ghosh, B. Liu, J. Power Sources 101 (2001) 167. [12] A.J. Bard, R. Parsons, J. Jordan, Standard Potentials in Aqueous Solutions, Marcel Dekker, New York, 1985. [13] W.M. Latimer, The Oxidation State of the Elements and Their Potentials in Aqueous Solutions, 2nd ed., Prentice-Hall, New York, 1952. [14] S. Amendola, US Patent 5,948,558 (1999).
8143
[15] S. Amendola, US Patent 6,468,694 (2002). [16] W.F. Linke, Solubilities of Inorganic and Metal-Organic Compounds, 4th ed., Van Nostrand, Princeton, NJ, 1958. [17] S. Hettiarachchi, P. Kedzierzawski, D.D. Macdonald, J. Electrochem. Soc. 132 (1985) 1866. [18] S. Licht, V. Naschitz, B. Liu, S. Ghosh, N. Halperin, L. Halperin, D. Rozen, Electrochem. Commun. 1 (1999) 527. [19] X. Fang, C. Yang, J. Chen, J. Chin. Ceram. Soc. 6 (1998) 732. [20] Z.-R. Tian, W. Tong, J.-Y. Wang, N.-G. Duan, V. Krishnan, S.L. Suib, Science 276 (1997) 926. [21] D. Linden, T.B. Reddy, Handbook of Batteries, 3rd ed., McGraw-Hill, New York, 2002.