Lithium-ion battery chemistry is challenged for long-duration energy storage applications.
What do we mean by long duration energy storage applications?
Currently the grid consists of a diverse array of generation assets that provide power as baseload, power during peaks and ancillary services to the grid. Increasing the share of renewables on the grid could drastically reduce the carbon intensity of the world's energy supply. To do so requires some method to manage the inherent intermittency of renewable generation. Energy storage is one way to achieve this transformation. It would allow loads to be served independent of the time of generation, by charging when resources are available and discharging when they are needed. Energy storage can also replace many services currently provided to the grid by natural gas turbines including ancillary services and capacity reserves. Some aspects of renewable integration (ramp control, etc.) as well as some ancillary services such as frequency regulation that require less than 30 minutes of power at any given time can be performed by batteries, such as Li-ion, that are economic in power-focused applications. However, to truly change the way the world is generating and managing the electrical grid, we believe that long duration technologies that can address over 4 hours of energy storage are required. This extended duration would allow energy storage to replace capacity currently provided by gas turbines and alleviate many of the issues impeding significant penetration of renewables to the grid.
Which storage technologies are being deployed successfully now?
The two most successfully deployed technologies on the grid are Sodium Sulfur and Lithium Ion (Li-ion). (We'll talk about Sodium Sulfur another time.) Li-ion has made significant inroads into grid energy storage given its strong performance in high power grid applications such as frequency regulation, black start and spinning reserves.
Over 200 MWs of Li-ion is deployed for grid applications globally, including several installation by AES.
However, Li-ion chemistries fall short when evaluating the requirements of long-duration, stationary energy storage applications, like microgrids and grid-scale storage.
There are three key concerns with Li-ion chemistry that support this finding:
2. Cost: Path to cost-down for high performing systems is a tough road
3. Cost/Performance: Cheaper Li-ion chemistries fall short on performance
1. Lithium-Ion Battery Safety
As we’ve seen increasingly in the news, Li-ion is an inherently unstable chemistry- one of the most recent examples being the Samsung Galaxy Note 7 battery fires. However, there are also stationary storage Li-ionbattery fires like the one in July of 2013 in the Pacific Northwest (just one example of many such event).
Additionally, issues with the Boeing 787 Dreamlinershave illustrated serious safety concerns with Li-ion chemistries. The organic liquid solvent-based electrolytes used in Li-ion batteries are extremely flammable and, as a result, significant thermal events involving Li-ion batteries often result in robust fires.
These types of events are going to continue to occur and, as the installations get larger, the magnitude of the problems will increase. The only way to avoid fires like this with Li-ion battery installations is to implement very costly and complex fire mitigation technology that may be cost effective for high power/low energy installations, but is cost prohibitive for large multihour storage systems.
2. Cost: The path to cost-down for high performing systems is a tough road
Of all of the industry-relevant objective analyses on battery pack pricing, we find the Advanced Automotive Batteries group, led by Mennahem Anderman, to be the most detailed and credible. In reading through this carefully-researched content, with input from 16 major battery producers and over 20 automotive producers, it becomes clear that the future pricing of Li-ion battery packs is not as optimistic as the values quoted by many battery manufacturers.
According to AAB's 2010 report, the projected average cell pricing in 2020 is about $325/kWh for long range EV's (and much higher for high power cells), and that the pack-level balance of systems will take the total pricing of the units to above $400/kWh. Assembling Li-ion into reliable high voltage strings can result in higher costs for both safety and reliability issues. Specifically this chemistry requires sub-string level battery management electronics to keep the string constituents working properly in concert. Cell degradation is also known to be very thermally sensitive, so significant thermal management is also required to keep all cells within a large pack in a similar thermal environment.
3. Cost/Performance: Cheap Li-ion alternatives fall short on performance
Less expensive Li-ion batteries do not have good cycle life or thermal performance. Multiple Li-ion manufacturers offer a low-cost alternative to automotive-specific cells, however these higher energy cells also have inherent cycle life limitations because they have thicker electrode structures. In general, those Li-ion batteries that exhibit excellent cycle life and temperature performance have thin electrode structures and costly electrolyte blends, raising the price per unit energy of the cells significantly.
The reality is that the excellent cycle life observed in thin electrode Li-ion cells with optimal electrolyte blends, costing well over $500/kWh, cannot be mapped directly to lower-priced alternatives optimized for energy content.
An interesting example is the Tesla battery electric vehicle battery pack strategy, where large format battery packs are made from low cost high energy “18650” cells.
It is an excellent choice for this application, since these batteries that are only actually deeply cycled hundreds of times, not even thousands of times; most drivers will use under 50 miles of driving a day, which is a fraction of the total range/capacity of the pack. These daily shallow cycles are better for this chemistry since the full capacity of the battery is rarely exercised. According to the spec sheet for the batteries used by Tesla, significant capacity is lost after only 500 deep cycles. This would not be acceptable for high energy, long duration stationary applications, where the battery installations will be very deeply cycled every day.