Katherine Hess — 2011-12 Fellow
Advanced energy storage technologies are crucial to the integration of renewable electricity into the grid and improving grid reliability. Increased grid-scale energy storage will also allow for superior peak-shaving and load shifting strategies, as well as provide a more efficient means for operating long duration ancillary services. To meet these needs, the energy storage technology must be efficient, reliable, dynamically-matched, environmentally benign, versatile and cost effective. Although many candidate technologies exist, none yet meet all of these criteria. One promising technology is hybrid batteries.
In developing batteries for large-scale energy storage, it is essential that the cost per unit energy be extremely low. The battery costs are distributed across the active materials, non-active materials (porous separators, current collectors, packaging, etc.), and manufacturing. Although active materials are a large part of the cost, the non-active materials also represent a significant portion of the cost, which increases roughly proportionally with electrode area. This cost can be reduced by preparing thicker active material electrodes that require less non-active materials due to the reduced electrode area for the same volume. For example, a DOE-funded technoeconomic analysis1 shows that increasing Li-ion battery electrode thickness is the only way to make those batteries competitive for electric vehicles; even if active material cost were reduced by 90%! Unfortunately, thick electrodes severely hinder power and result in long charging times. This project focuses on improving the power and charging rates of thick electrode batteries for highly cost competitive grid-scale energy storage applications. We initially focus on the aqueous sodium hybrid battery that is being commercialized by Aquion Energy, Inc., our 2010 Pennsylvania Infrastructure Technology Alliance (PITA) grant industry collaborator.
Aqueous sodium hybrid batteries use a sodium intercalation positive electrode and an electric double layer capacitance (EDLC) negative electrode. The batteries use only low cost, benign materials, exhibit high durability and reliably sustain charge. However, in order to maintain low costs of non-active materials while reaching the high levels of energy storage needed for grid-scale applications, they require the use of ultra-thick EDLC electrodes (1 mm – 1 cm). The longer transport distances associated with these ultra-thick electrodes leads to higher losses through the electrode hindering power, slower charging/discharging times, and non-uniform electrode utilization.
This project is aimed at advancing this hybrid battery technology with fundamental studies, and optimization of transport and electrochemistry in ultra-thick electrodes. To this end, we will apply advanced diagnostics and computational modeling. Our novel diagnostic device, an electrode scaffold (ES), uses sensing layers to obtain through-plane distribution measurements of ionic and electric potential and species concentration through the electrode. The ES provides a structure for intersecting sensing materials with the electrode cross-section at discrete distances through its thickness. For example, thin electrolyte layers and microchannels coupled with external reference electrodes allow us to measure electrolyte potentials, while potentiometric sensing materials enable measurements of species concentration. In concert with computational modeling, we will use these diagnostics to answer several key questions: 1) how effectively are the thick electrodes utilized, 2) what are the dominant loss mechanisms, and 3) how can new materials and fabrication techniques be used to improve transport and meet performance targets?
Approach and Methodology
The key to the proposed through-plane diagnostics is the electrode scaffold (ES), a structure that constrains the sides of an electrode and provides access for sensing layers at discrete distances parallel to the direction of mass and charge transport. The ES bounds the side of the working electrode and sensing layers (or microchannels) contact the electrode at discrete distances through its thickness. Sandwiched between the sensing layers are non-porous, insulating layers that provide the spacing for distributed measurements. Using a variety of electrochemical methods and sensing layer materials, we can uniquely measure a wide array of property distributions, including ionic and electric potentials, charging rate and charge distribution, conductivity, and ion concentrations.
Over the next year we will use different configurations of the ES to obtain ionic potential distributions and Na+ concentrations as well as develop a computational model in order to further characterize the electrodes and begin assessing new fabrication techniques and materials for simultaneously meeting performance and cost targets. Each of these tasks is described in further detail below.
Electrolyte Potential Sensing – We will examine Ohmic voltage losses in the electrolyte by measuring electrolyte potential across the battery electrode’s thickness. The ES and specialized cell hardware also allow us to measure in-situ, through-plane conductivities, which are important properties for optimization. The measured potentials will also allow us to estimate local current and charging rate distributions using finite difference methods. This information is crucial to experimentally quantifying the electrode’s effectiveness.
Na+ Concentration Sensing – By measuring Na+ concentration distributions, we will investigate concentration polarization, which is the voltage loss due to reactant depletion. We will develop a unique measurement that uses electrically isolated Na0.44MnO2 as a potentiometric Na+ concentration sensor.2 In other words, the electric potential of the Na0.44MnO2 (versus a reference electrode) varies with the Na+ concentration in the adjacent solution with a sensitivity of 55 mV per decade of Na+ concentration.2
Materials and Fabrication – We will experiment with new materials and electrode fabrication techniques for dramatically improving transport rates without deteriorating energy density or increasing cost. Our concepts include high aspect ratio materials that can be aligned through the electrode to create straight pathways as well as devising strategies to introduce aligned electrolyte channels in the electrode for fast ion transport. We will also explore functionally grading the porosity (open space) through the electrode for improved uniformity.
Computational Modeling – In concert with our advanced diagnostics, we will construct a computational battery model using a discretized, representative elementary volume approach that resolves sub-grid phenomena with closed-form analytical solutions. Litster and his group have expertise in modeling porous electrodes and have the necessary computational resources, including COMSOL Multiphysics finite element software and high performance workstation (12 cores, 96 GB RAM). The geometric inputs for the electrode microstructure will be extracted from optical and scanning electron microscope images, gas adsorption, and mercury porosimetry.
1. Barnett, B., "PHEV Battery Cost Assessment," Proc. of the DOE Vehicles Technology Program Annual Merit Review, Washington, DC, June 8, 2010.
2. Sauvage et al. (2007), Sensors Actuators B, vol. 120, pp. 638-644.