60th Annual Report on Research 2015

Objective

  Renewable energy sources including wind, solar and tidal have been identified as one of the best solutions to address the current environmental problem and energy crisis. To enable the regulation and transmission of such intermittent power, the use of grid-scale energy storage system is crucial. The redox flow battery (RFB) is an emerging energy storage technology, which offers a cheap and sustainable alternative for integration of the renewable energy sources into the electricity production and distribution system through smart grids[1]. In RFB, Graphite felt (GF) electrodes are primarily used as the positive and negative porous electrode for their high surface area, high conductivity, wide operating potential range and reasonably low cost. Though interest in RFB technology has significantly increased recently, the performance and durability of GF electrodes, including the kinetics of the redox reactions, side reactions and corrosion mechanisms are still poorly understood.    The primary objective of the project is to develop a detailed transport and multiscale model to understand the kinetics of felt electrodes in a Thin Film Rotating Disk Electrode (TFRDE) set-up. The project also aims to develop a model to examine the surface heterogeneity and carbon corrosion that can lead to electrode damage and battery failure under high operating voltages. The knowledge gained from detailed interfacial studies of the GF electrode provides insight on the effects of various surface modification protocols reported in the literature and suggest new pathways for electrode performance and durability improvement

Research Findings

The rotating disk electrode (RDE) is a useful tool to characterize electrochemical reactions and mass-transfer rates with a well-studies flow field, which can mimic the half-cell electrochemical reaction with flow in RFB system. The analytical equation that relates the diffusion-controlled current to the rotation rate is the analytic Levich equation: , where n is the number of electrons transferred, F is the Faraday constant, A is the electrode area, D is the diffusion coefficient, 𝜔 is the angular rotation rate, 𝜐 is the kinematic viscosity and c is the concentration of the reactant. However, the Levich equation does not hold for porous electrode. PRDE experiments were conducted in acidic VOSO₄ solution (0.1mol/L) with different porous electrode thicknesses at different rotation rates. The experimental data is summarized in Fig.1. For low rotation rates, porous rotating disk (PRDE) reacts similarly as a RDE, with little flow inside the porous electrode. As the rotation rate increases above a critical rotation speed, the value of limiting current increases more quickly than predicted by the Levich equation and reaches a plateau eventually.

To understand the porous electrode in flow system better, a detailed physics-based model was developed. When the PRDE is fully perfumed by the reactant, it can be treated as a batch reactor, which explains the plateau behavior after a second critical rotation rate[3]. At high concentration, mass transfer is quicker compared to reaction rate, thus the rotation rate does not affect the limiting current much. A detailed comparision of model with experimental data will be published soon and the current predictions agree qualitatively with the experimental data (figure 2). While the model developed predicted the experimental data better than the Levich equation, it still did not  fit the data 100%. We attribute this to the corrosion of the electrode. A kinetic Monte Carlo (KMC) approach is presented here in an attempt to model the surface heterogeneity on a carbon felt electrode. KMC is a computationally expensive stochastic approach which considers discrete events within a system. Here we consider a carbon electrode surface as a two dimensional plane modeled as a 10 by 10 rectangular lattice to model the surface reaction: VO²⁺ + H₂O ↔ VO₂⁺ + 2H⁺ + e^-                      E⁰=1.00 V This model considers several events that can occur at each KMC transition:

1)      Adsorption of the reduced species (VO²⁺)

2)      Adsorption of the oxidized species (VO₂⁺)

3)      Desorption of the reduced species (VO²⁺)

4)      Desorption of the oxidized species (VO₂⁺)

5)      Surface diffusion of both species

6)      Surface reaction of the reduced species to oxidized species

7)      Surface reaction of the oxidized species to reduced species

8)      Oxidation of the carbon felt electrode Surface representation of a carbon felt electrode before cycling (left) and at the end of cycling (right) is plotted in figure 3. The dark blue cells represent empty sites, the cyan cells are adsorbed reduced species V(IV), the orange cells are the adsorbed oxidized specied V(V) , and red cells are where corrosion has occurred. Notice the increase in the adsorbed oxidized species and decrease in the reduced species and siginficant corrosion   In addition, during the progress of this project a more efficient numerical method of lines approach was developed for solving RDE  model equations and elliptic PDEs.⁵   Impact on the PI and personnel – This project helped the PI diversify his research expertise from lithium-ion batteries to flow batteries and experimental research. In addition, it helped support PhD students who benefitted from doing an industrially relevant project. This project partially helped the PI in getting a follow-on flow battery experimental design project from Sun-Edison.  

References

[1] Knehr, K. W., et al. "A transient vanadium flow battery model incorporating vanadium crossover and water transport through the membrane." Journal of The Electrochemical Society 159.9 (2012): A1446-A1459. [2] T. Nguyen, R. Savinell, Flow Batteries, in: The Electrochemical Society Interface, vol. 19, The Electrochemical Society, New Jersey, USA, Fall 2010, pp. 54-56.   [3] Bonnecaze, Roger T., et al. "On the behavior of the porous rotating disk electrode." Journal of the Electrochemical Society 154.2 (2007): F44-F47.   [4] Gattrell, M., et al. "Study of the mechanism of the vanadium 4+/5+ redox reaction in acidic solutions." Journal of the Electrochemical Society 151.1 (2004): A123-A130.   [5] Pathak M, M. Ramanathan and Subramanian V.R., “A Direct Numerical Method of Lines Approach for Predicting Primary and Secondary Current Density Distributions: Linear and Nonlinear Boundary Conditions”, ECS Fall meeting, 2015.

Fig 1: Experimental data for kinetics

Fig. 2 Model prediction for limiting currents

Fig. 3 Surface degradation of graphite electrodes predicted by KM models.