Characterization of Carbon Felt Materials As Negative Electrode Material for Flow-Assisted Zn-Air Batteries

Abstract

Secondary alkaline Zinc-Air batteries hold several distinct advantages over other large-scale systems for energy storage. Such batteries use materials that are low cost: Zn is abundant in the earth’s crust and alkaline aqueous solutions are cheap and well understood. They also have low toxicity and high chemical stability. To make Zn-air batteries viable for large scale use, issues at the Zn electrode need to be addressed. Allowance must be made to prevent dendrite growth that eventually leads to cell shorting during charging. During the oxidation of Zn occurring in the discharge process in the battery, a sequence of chemical steps occurs producing a passive layer of ZnO on the electrode. Additionally, electrode morphology changes between each cycle produce inconsistent performance. These issues can be improved by using a three-dimensional support structure on which Zn is deposited. Carbon felt electrodes provide an ideal structure for the Zn anode in Zn-air systems as they are porous and highly conductive. These features allow for high surface area for deposition to occur, low mass transport constraints and low internal IR losses of the electrode. Furthermore, the carbon structure provides a consistent structure to return to minimize Zn morphology changes. These structures have been explored for their potential as the anode in Zn electrodes and showed significant promise. No issues with dendrite growth and passivation have been observed in the electrode during cycling. They can achieve 100mA/cm2 of current density at 20mV of overpotential during deposition and dissolution at a flow rate of 15ml/min of electrolyte through the electrode. Additionally, they have been shown to be stable for 100 cycles of cycling in a symmetric cell at 50mA/cm2 while maintain an average overpotential of less than 50mV during each step. Cycling at high flow rates improves the performance of the felt with an average step overpotential of 16mV at 28 ml/min and 28mV overpotential average at 10ml/min, all also at 4M NaOH saturated with Zincate. Increased state of charge of the electrode appears to increase the efficiency of the electrode. During charging steps in 10ml/min cycling experiments overpotential starts out near 40mV and trends towards 20mV during the step. Further, a 50mV CA step at 30ml/min shows current density increase from 100mA/cm2 at the start of the test to 400mA/cm2 at the end of 30 min. This is the result of the increased amount of Zn in the felt system providing more surface area of Zn for deposition to occur directly on. The potential during galvanostatic cycling experiments appears to be somewhat unstable indicating that the basic electrochemical performance of the electrode needs more study. Nucleation of Zn deposits occurs throughout the felt thickness regardless of the applied overpotential or current. As deposition progresses, the location of Zn growth depends on the current applied. At higher currents or overpotentials growth localizes near the membrane separator. It then begins to fill in back towards the current collector. The reverse occurs during discharge of the electrode, material is removed preferentially from the parts of the electrode near the current collector then proceeds towards the membrane separator. Experiments show that this is based on overpotential in the cell and under both potentiostatic and galvanostatic conditions; at low observed or applied potentials growth occurs throughout the cell felt and at higher overpotentials the growth becomes localized. IR drop measurements through the felt indicate that only 20% of the electrode is active during the growth phase of a 1hour, 40mA/cm2 deposition test. This suggests migration of the negatively charged zincate towards the membrane separator is dominant in these conditions. This ultimately reduces the efficiency of the electrode as it increases the overpotential required to achieved the desired current. Methods to counteract this effect with be discussed. Optimized electrode performance including flow rate, electrolyte conditions, and preconditioning of the electrode for cycling of symmetric and full cell setups will also be discussed further. We gratefully acknowledge the support of Dr. Imre Gyuk and the DOE Office of Electricity Delivery and Reliability. Figure 1