Battelle/Concurrent Technologies Corporation technology positioning paper: Improving Li-ion battery safety without decreasing energy density
Ed. introduction: The following technology positioning paper is a joint effort by a team from Battelle and Concurrent Technologies Corporation (CTC). The paper outlines the technology landscape and the opportunities that exist in the area of improved Li-ion battery safety.
Energy is a common technology area on which both research organizations focus in different ways, noted Dr. Vicki Barbur, CTC Senior Vice President and CTO. The two have decided that Li-ion battery safety is an area of opportunity for each. Supported by an ARPA-E award, Battelle recently developed an optical sensor to monitor the internal environment of a lithium-ion battery in real-time. (Earlier post.) The organizations intend to pursue further efforts directed toward safety in relation to Li-ion battery technology. Interest and involvement from external clients would be welcomed.
One of the things that prompted the tone of the paper was that most of the things done right now to improve safety are also decreasing the energy density of the battery. That pushed us in the direction of: “How do we improve the safety without decreasing the energy density?”—Jim Saunders, Battelle
A Collaborative Focus and Joint Effort of
Battelle Memorial Institute and Concurrent Technologies Corporation
Given their capacity to hold a relatively large electrical charge per unit weight (a quantity known as “energy density”), lithium-ion batteries have been applied in a wide variety of electrical devices including cell phones, laptop computers and electric vehicles. Much effort is currently dedicated to development of lithium-ion batteries having greater energy density. These efforts include increasing the capacity of the anodes and cathodes, increasing the cell voltage, and decreasing separator thickness.
Recently publicized, high-temperature safety incidents associated with the use of lithium-ion batteries have stimulated development to improve the safety of these batteries, through improved design and manufacturing, as well as changes to the materials of construction. It is important to differentiate between efforts to improve safety, and efforts to improve durability. Efforts to improve safety focus on reducing the number and severity of events that lead to sudden or catastrophic failure of the battery, while efforts to improve durability focus on slowing the gradual degradation of the battery as it is cycled or is exposed to conditions near the limit of its safe operational window.
Generally, the improvements in lithium-ion battery safety have decreased the energy density of the battery system, especially when those improvements are related to internal faults and thermal runaway. For instance, virtually all safety improvements in the electrode materials have led to less active electrodes. Thermal management has added weight and complexity to the battery at the system level, even if it has not increased weight at the cell level.
Therefore, we see the fundamental question for battery development as: “How do we improve battery safety without decreasing energy density?” Although there are many possible answers to this question, we have catalogued the answers into four major categories:
- Sensors for local fault detection
- Diagnostics for local fault detection
- Enhanced local fault safety features
- Intrinsically safer materials
Below we will discuss each of these areas, including ongoing work and opportunities.
Sensors for local fault detection. Many internal battery faults (i.e., electrical shorts) originate locally, in very small regions of a single cell. For example, dendrites (localized metallic growths that form on one electrode and can grow to bridge two electrodes) may only cover an area of a fraction of a square millimeter, while the electrodes for an EV battery can cover an area of greater than 200,000 square millimeters. Similarly, a localized hot spot may only encompass a region on the order of a square millimeter. Typically, the sensors currently used in batteries only measure global parameters, such as the cell voltage or the average temperature of the cell. Temperature measurements on the outside of a battery module cannot detect a local internal hot spot before it propagates while measurements of cell voltage will not detect the presence of dendrites until a major shorting of the cell occurs. Thus, neither of these measurements is sensitive enough to detect local faults in the early stage before a significant safety event occurs.
Several approaches are under development that can detect local faults within a battery. For example, a technology is under development that uses the battery separator as an optical sensor. In this case the impingement of a dendrite into the separator would lead to a change in the optical transmission through the separator; this change can easily be detected. The technology could also detect the presence of local hot spots, although this has not yet been demonstrated. A second sensor technology is under development that incorporates a fiber optic sensor into the cell that is able to measure small variations in the local temperature across the cell.
Diagnostics for local fault detection. One of the most significant elements of a large-scale battery is the battery management system, which uses global data such as cell voltage to determine the health of the battery, maintain balance between the cells, and maintain the battery within the desired state-of-charge window.
Major efforts are underway to improve the battery management systems, to yield an increase in the useable charge window and better monitor battery health. Many of these efforts are focused on the algorithms underlying the management systems, improving the way the information is processed, and providing improved correlation between measured and estimated properties. By combining advanced model-based algorithms with global measurements, anomalies in battery performance can be detected that are reliable indicators of impending internal faults.
The possibility of obtaining more localized sensor information, from systems such as those just described, also opens new opportunities for diagnostics. Determination of properties such as the root mean square (RMS) deviation of local voltage from the global voltage provides information on the inhomogeneity of the impedance across any given cell. An increase in the inhomogeneity, or crossing a threshold value or the rate of change, may indicate a potential failure. Similarly, an increase in root mean square (RMS) deviation of the local temperature across a cell or from cell to cell may indicate an actual or pending short. These are just some examples of the improved diagnostics that may be used as information becomes available on the local state of the cell, especially when combined with advances in model-based diagnostics.
With better understanding and diagnosis of both local and global faults, comes the possibility of model-based control. For instance, it is well known that lithium plates on carbon electrodes at a potential very close to the potential needed to charge the carbon electrodes, providing an opportunity for plating to occur under off-design charging conditions. By combining advanced sensing and diagnostics, control algorithms may be able to manipulate the individual cell voltage and current to avoid operating regions where faults are likely to occur. In addition, this approach is likely to avoid damaging side reactions that reduce battery capacity and contribute to battery aging.
Enhanced local fault safety features. In addition to the previously discussed methods to detect faults at the local level, there is a need for better local safety features. One major safety feature that has been introduced into batteries is a shutdown separator, which uses the heat generated in a local shorting event to partially melt the separator and prevent ion transport. Although this can act to prevent some failures, it will not prevent all types of failures.
There are several challenges with the approach taken with shutdown separators. One is they can leave the battery in a fully charged state, which is hazardous for handling. Similarly, current interrupters may work too late for local events. Secondly, in many battery system designs, shutting down a cell may lead to the battery becoming inoperable. Third, by the time a cell becomes hot enough to trigger the shutdown event, thermal effects may already be propagating to adjacent cells.
There are several approaches that can enhance the local fault safety features of batteries. In one example, a shutdown separator is under development that slowly allows the charge in the battery to drain off after the shutdown is triggered, making the battery safe to handle after a thermal event has occurred. In another example, developments are underway to electrically disengage cells that are failing from the battery circuit. There are also efforts underway to thermally isolate cells so failures cannot propagate. In one example, phase change materials are being used to create thermal barriers between cells, to prevent thermal damage to cells adjacent to a failed cell. In a more extreme example, developments are underway to physically separate the cell from the rest of the battery structure in case of a failure.
Intrinsically safer materials. The fourth major area for safety research is in the development of intrinsically safer materials. A major concern with current lithium-ion batteries is that the materials themselves are intrinsically hazardous. The carbonate electrolyte is highly flammable, and will often ignite if it is expelled from the cell packaging. Another challenge is that the current electrolytes are being used close to the limit of their safe-operating voltage window, meaning application of over-voltages to the cell can result in physical damage to the cell, battery and/or surrounding structures.
There is significant work underway to develop electrolytes with larger safe-operating voltage windows with increased safety. The leading materials at present are ionic liquids, which are salts that are liquid at room temperature. Many of these materials have very low vapor pressure, and do not degrade at temperatures below 300 ˚C. A second area of research is development of cathode materials with higher thermal stability. The temperature at which the metal oxide cathode will decompose is highly dependent on the crystal structure. Recent work has shown several classes of metal oxides that have both high voltages and greatly increased decomposition temperatures. By combining these cathodes with ionic liquids, it will be much more difficult to create a runaway thermal event.
Adding to the safety concerns of these batteries is the interest and ongoing activities to use these batteries in second-life applications. As these batteries degrade to 70-80 percent of original energy capacity, they are insufficient for vehicle use. However, they could be repurposed for grid energy storage or other applications. Battery characterization and an understanding of degradation become critical to set the charge and discharge rates of these batteries in second-life applications. Several of the items discussed earlier will also have the potential to obtain better utilization of these batteries, which will greatly improve battery economics.
Battelle and CTC are certain energy storage will continue to play an important role in all aspects of our future from small personal electronic devices to larger applications such as the electrical grid. To ensure the safety, reliability and energy efficiency of lithium-ion batteries, we are collaborating to develop cost-effective solutions that address the above issues. This collaboration includes development of proof of concepts and jointly seeking funds to support further development of those concepts showing the greatest promise.