The Effects of Different Pulse Charging Techniques On Dendrite

The effects of different pulse charging techniques on dendrite growth in lithium-air batteries Introduction • Lithium air batteries have been predicted to have more than 3 times the storage capacity of lithium ion batteries[1] ▪ Results from a lithium/oxygen reaction at the cathode[2] • Lithium metal anodes result in higher dendrite growth[3], which: ▪ Are needle like structures of solid lithium that grow on the anode ▪ Can pierce the separator, causing short circuits and thermal runaway[4] ▪ Can reduce the storage and efficiency of a battery without piercing the separator • Growth can be explained with the deposition/ dissolution model[5] ▪ Solid lithium is deposited on the surface of the anode ▪ Deposited lithium ions form dendrites • Solutions proposed to solve dendrite growth include: ▪ Changing the electrolyte mix in the battery ▪ Changing the design and material of the separator ▪ Controlled cycling schemes, including pulse cycling • Present method of charging is Constant Current - Constant Voltage(CC-CV)[6] ▪ Current is gradually decreased when max voltage is approached • Focus of this study is on square wave pulses, a pulse cycling technique that demonstrates the principle of TOn and TOff cycles [7] Results All simulations use a reaction rate of 1600 μ/s unless stated otherwise. Fig. 1: Charge and rest time modified to make different duty cycles. When current normalized (scaled so the same amount of overall charge is applied), it is found that dendrite growth is directly proportional to duty cycle. Image 1: Diagram of the inner workings of a lithium air battery[8] Fig. 2: Charge and rest time modified, as well as number of pulses, to test the effects of different pulse lengths. While pulse length has an effect on dendrite growth, it is not directly proportional. Image 2: Picture of dendrite growth within a battery[9] Image 3: Diagram of a square wave pulse Methods • Computational models used to simulate diffusion-reaction: ▪ SPH(Smooth Particle Hydrodynamics) ▪ Handles multi-phase physics, complex geometry, moving boundary • Simulations modeled at mesoscopic scale using tens of thousands of particles • Required use of BU's supercomputer • Three parameters were modified: ▪ Length of Ton ▪ Length of Toff ▪ Reaction rate (proportional to current) • After the simulation was run, the data was: ▪ Visualized in a program called Ovito ▪ Run through python scripts to graph the dendrite height Fig. 3: 2 monotonical graphs with constant change of 100 μ/s per pulse. The graph with a negative gradient creates more growth. Fig. 4: Reaction rate increased for 3 pulses and decreased for 3 to create a triangle wave. Normalized graphs demonstrate that the lowest growth comes from the median gradient. Image 4: Frame from an Ovito visualization of the charging process Discussion • Simulations demonstrated that, on average, a longer rest time leads to less dendrite growth ▪ Supports the idea that lower duty cycle pulses will create less dendrite growth regardless of pulse length • To expand upon this finding, further studies could: ▪ Compare different duty cycle pulses normalized for current and run for the same time period ▪ Simulate each pulse technique being used to reach a total current goal rather than a time limit • These would illustrate the real world practicality of each charging cycle when charging a battery • Simulations indicate that longer rest cycles lead to less dendrite growth, but there is an exception: ▪ Second longest pulse created most dendrite growth in its graph ▪ Means that longer pulse does not reliably lead to lower dendrite growth ▪ Simulations that run pulse lengths between 4 and 2 can be used to determine why the 2 second pulse is an outlier • Triangle wave graphs and monotonical graphs: ▪ Found that a lower gradient creates more growth when current normalized than one with a larger gradient ▪ Indicate that for monotonical and triangle pulses, a shorter pulse with a larger gradient best suppresses dendrites • Study only focuses on dendrite growth, not the other effects of cycle charging, including: ▪ Overcharging ▪ Overheating ▪ Electrode degradation ▪ Battery charge time • Also does not consider other factors that affect dendrite growth, including: ▪ Temperature ▪ The presence of a separator ▪ The composition of the electrolyte mix • Findings should not be seen as a final solution to the question of pulse charging in lithium-air batteries ▪ Instead, should be seen as a contribution to the current work on dendrites and pulse charging • Most effective solution for lithium air batteries needs to: ▪ Suppress dendrite growth ▪ Decrease charge time ▪ Increase cycle life and battery safety • Solution would be a combination of different methods to improve li-air batteries, rather than one method trying to solve all problems