1 Department of Chemistry, Technical University of Denmark2 Department of Energy Conversion and Storage, Technical University of Denmark3 Royal Veterinary and Agricultural University
This report describes the work on development of hydride forming alloys for use as electrode materials in metal hydride batteries. The work has primarily been concentrated on calcium based alloys derived from the compound CaNi5. This compound has a higher capacity compared with alloys used in today’s hydride batteries, but a much poorer stability towards repeated charge/discharge cycling. The aim was to see if the cycleability of CaNi5 could be enhanced enough by modifications to make the compound a suitable electrode material. An alloying method based on mechanical alloying in a planetary ball mill was developed. The parameters milling time, milling intensity, number of balls and form of the alloying metals were investigated. Based on this a final alloying technique for the subsequent preparation of electrode materials was established. The technique comprises milling for 4 hours twice possibly followed by annealing at 700°C for 12 hours. The alloys appeared to be nanocrystalline with an average crystallite size around 10 nm before annealing. Special steel containers was developed for the annealing of the metal powders in inert atmosphere. The use of various annealing temperatures was investigated. The hydrogen absorbing alloy CaNi5 has been prepared by mechanical alloying and the structure was confirmed by X-ray diffraction. Gas absorption as well as electrochemical experiments showed hydride forming properties as fully developed as for a corresponding commercial alloy. For the electrochemical investigations simple methods were developed for the manufacture of hydride electrodes as well as Hg/HgO reference electrodes. It is shown that CaNi5 is not corroded during currentless storage in 6M KOH in the uncharged state. The degradation process is closely related to charge and discharge. Upon cycling the capacity of the calcium containing electrodes decreased rapidly following an exponential-like curve. The decay curve levelled out around 30 mAh/g, the capacity only falling very slowly hereafter. This residual capacity was independent of the discharge current density in the range from 50 to 200 mA/g. A suggested explanation of the residual capacity is the oxidation of Ni(0) to Ni(II) at the surface of the nickel remaining after leaching of calcium. A number of alloys with an overall formula of CaNi5-xMx (x = 0, 0.5 or 1 and M = Al, Co, Cr, Cu, Fe, Mg, Mn, Sn or Zn) were prepared and electrodes of these were cycled at constant current (100 mA/g). Alloys with M = Cu, Mg or Zn had capacities between 321 and 390 mAh/g (CaNi5: 388 mAh/g). The rest of the alloys had even lower capacities. No alloys showed any significantly higher stability towards cycling. The AB5 phase (CaCu5 structure) was maintained for M = Cu, Zn and to some extend for M = Sn. In the other cases A2B7 and AB3 compounds with Gd2Co7- and PuNi3-structures were formed. It is shown that magnesium occupies positions normally occupied by calcium in calcium nickel alloys. Ca0.67Mg0.33Ni3 was prepared with PuNi3 structure. The compound can be regarded as CaNi3 with partial substitution of magnesium for calcium. The alloy had a capacity of 390 mAh/g and a little higher stability towards cycling than CaNi5. The alloys were also cycled “as milled” (without annealing). The capacities were then in general lower, but the stability somewhat higher. The higher stability is explained by a smaller volume expansion during charge. It is shown than sodium can substitute for calcium forming the compound Ca0.8Na0.2Ni5. The compound had CaCu5 structure and a capacity of 365 mAh/g but a poor electrochemical cycle life. The alloys Ca0.8Na0.2Ni4Mg0.5Cu0.5 and CaNi3.6Co0.7Mn0.4Al0.3 were prepared and tested and found to have capacities of 325 mAh/g and 147 mAh/g, respectively. The cycle lives were also poor for these alloys. It is concluded that despite substitutions calcium alloys are not suited as electrode materials in an alkaline aqueous electrolyte. The Mischmetal alloy MmNi3.6Co0.7Mn0.4Al0.3 was prepared with CaCu5 structure. The capacity and cycleability were a little poorer than a corresponding commercial alloy but activation was much faster. An amorphous magnesium nickel alloy with a capacity of 532 mAh/g at 18 mA/g was prepared. This capacity is at least at the level of the best results found in literature.