Published in


How to Prevent Electric Dreams from Going Up in Flames: The Challenge of Battery Safety

This article is contributed by Sriram Bharath and Shashank Sripad

  • While instances of fires are lower in EVs compared to ICE vehicles, incidents involving EVs tend to garner outsized public attention.
  • The industry is still working to understand the challenge of battery safety; OEMs and cell manufacturers cannot afford multiple safety-related recalls during this period of industry-wide transition.
  • OEMs need to take a robust, testing-driven approach to ensure that EV battery packs exceed the levels of safety outlined in regulations and standards.
  • Safety must be considered from the early stages of battery design ideation–it is easier to design safe batteries than it is to impart safety to an already-designed system.
  • Standards and regulatory frameworks need to be expanded for the EVTOL space, and unified across different markets for EVs.

Total Recall

Recalls of the Chevy Bolt and Hyundai Kona/Ioniq EVs in 2021, totaling more than 200,000 vehicles and over $2B, have put a glaring spotlight on battery safety. The year 2021 also marked the momentous occasion of EVs reaching 10% of all new car sales. As the number of EVs increases, basic statistics tells us that we will see more battery safety-related recalls, unless the battery industry improves its standards of safety. Battery safety incidents tend to gather outsized public attention, and repeated recalls have the potential to become electrification’s Achilles heel, contributing to EV hesitancy amongst the public–in addition to the costs of the recall and replacement, the industry may also have to deal with the possibility of losing new customers.

Battery safety incidents occur across the world and extend beyond EVs, including residential energy storage, grid storage, personal mobility devices, electronic cigarettes, air transport and even marine applications. The largest ever recall was, in fact, from the world of portable electronics, covering nearly 2 million units of the Samsung Note 7 a few years ago. One thing we can be certain about is that battery recalls always leave a bitter aftertaste.

A comparison of recalls prompted by battery safety concerns. Source: Eli Leland, Voltaiq.

To be fair, batteries are actually much safer than their combustion counterparts. Designing a battery pack involves trade-offs between energy density, electrochemical performance, lifetime, cost and safety–a trilemma of sorts in which improving on one of performance, cost, or safety means compromising on another. Battery materials and chemistries with high performance tend to have more severe failure energetics, while cutting costs in manufacturing may lead to lower quality controls and safety standards. No matter how good the materials and quality controls are, however, perfect batteries can never be manufactured. As Voltaire put it, “perfect is the enemy of good,” and, as an industry, we need to find ways to be good or better.

If you want to learn about how batteries fail and catch fire, check out this accompanying piece.

Safety considerations for batteries on the move

Collision-related fires account for only 5% of vehicle fires and this is true for EVs as well; rather, most failures observed in the field stem from manufacturing defects, conservatively estimated to be on the order of around 1 in 10M. In EVs, depending on the cell format, the number of cells used may range from a few hundred up to nearly 8,000. The approach taken by battery safety standards is to accept a finite chance of thermal runaway in a single-cell and focusing on reducing hazard severity by enabling graceful failures.

The problem with manufacturing defects in cells is that they are very difficult to fix once they have been installed in a pack, except by replacing the entire pack. With the Chevy Bolt, a software update attempted to address the issue by limiting the battery SOC between 30 and 90% but even this severe restriction on usage was unable to prevent further incidents. Software updates could not address the manufacturing defects (torn anode tab and folded separators) at the root of the issue, ultimately requiring full-pack replacements and the recall to be expanded to every Bolt that ever rolled off the line.

The important safety consideration for batteries is preventing cascading failures from one cell to another in a pack. A cascading failure is typically accompanied by a sustained fire (the battery safety police call this Hazard Severity Level 6), which further accelerates battery failure. Battery pack design, cell form factor, pack topology, and various other factors influence the safety of a battery pack. Going back to the Bolt example, the battery pack was designed with pouch cells for ease of packaging and higher energy density, but pouches offer little resistance to cascading/propagation failure. The Bolt pack also had certain modules that were stacked on top of one another, meaning that failures in some modules could ignite multiple adjacent modules as vent gases or flames impinged directly on other cells.

Failure characteristics and energetics are deeply dependent on cell chemistry. The recent resurgence of LFP for both EVs and stationary storage applications are motivated by both cost and safety considerations. Owing to higher cathode stability, manufacturers are tempted to manufacture LFP cells in larger formats, so the fraction of cells with manufacturing defects may be higher than cylindrical cells. Thus, when single-cell failures occur, the total energy released is still quite high. Additionally, owing to high fractions of H2 in the vent gases from LFP cells, they can be more flammable and explosive than other chemistries. Consequently, the problem of cascading failure still exists, although the failure may occur over a longer period of time. BYD’s blade cells, for example, were designed to fail without fire or flames even when the cell loses mechanical integrity, enabling the pack to be limited to a Hazard Severity Level of 3.

For batteries on two wheels, safety should ideally win the tradeoff, especially in Asian markets. If regulations do not exist, or are under development, for such products in a particular region, the manufacturer must be proactive in identifying the relevant safety standards available in other markets rather than waiting for local bodies to issue guidelines. Product development life cycles are typically much shorter for this segment, and replaceable batteries are common for products such as electrified bicycles. Consequently, counterfeit replacement batteries are very easily accessible in online marketplaces, and these tend to have very low manufacturing quality, leading to frequent failures. Furthermore, the danger of battery failures is also not well understood by the public and education is needed to ensure that they do not approach failing batteries that may be on the verge of exploding.

How to police battery safety?

Standards and regulations are the main methods of establishing safety expectations, but given the large variety in battery designs being explored, they may not be able to keep pace with rapidly developing technology and will still not prevent all safety-related issues. The GM and Hyundai vehicles that are being recalled, for example, passed the required regulatory testing but still failed in the field. These incidents show that current safety regulations and standards for EVs are limited in scope and have gaps that still need to be addressed. Several instances of cells modified specifically to pass safety standards have occurred, and this is an ethical concern that the industry should not tolerate.

On the other hand, failure propagation resistance, while recommended by standards/manuals such as SAE J2464, UL 2580, UL9540A, IEC 62619, SAND2017–6925, and RTCA DO-311, is not currently a regulatory mandate. In fact, batteries for EVs in the US require just self certification according to the UN/DOT 38.3, which pertains only to the shipping of lithium-ion batteries. In some standards, nuances such as the determination of the “most vulnerable” locations in a pack are left to the manufacturer, with little guidance on the parameters (energy release, likelihood of failure, number of cell failures, etc.) to define the worst case scenario, and these need to be brought into the fold of standardized regulation. The introduction of the updated UN Global Technical Regulations are expected to aid in aligning the regulatory frameworks between different markets including North America, Europe, China, and Japan.

Propagating failures are generally mitigated through passive mechanical strategies and the primary goal of these systems is to delay the failure until occupants have evacuated the vehicle. Common approaches to imparting propagation resistance include increasing inter-cell distance, the use of rigid separations and thermal insulators between battery modules, and enclosing the interstitial space between cells with intumescent or “potting” materials. BMSs typically cannot compensate for cell design faults and issues after the onset of a thermal event at the cell level. With sufficient instrumentation within the pack, however, the BMS can trigger warning signals to inform passengers about impending failures. On the other hand, BMS malfunctions have also been the cause of catastrophic failures.

Below are some recommendations for battery designers and cell manufacturers to improve safety across all applications and through the product lifetime.


  1. Understand the trade-offs of the use case well to choose the right cell for the application.
  2. Adhere to the voltage and temperature limits established by the cell manufacturer.
  3. Implement module/pack level safety measures based on cell-level protections (CID/PTC).
  4. Characterize cell failure using calorimetry to size mitigation systems and develop predictive tools.
  5. Develop robust test methodologies for cells, modules and packs by combining existing standards.
  6. Perform independent safety testing of cells either internally or at qualified laboratories.
  7. Perform exhaustive safety testing early in the design process to inform design decisions.
  8. Cover low-probability cases and worst-case scenarios to ensure the module/pack is resistant to single-cell failures and side-wall ruptures.
  9. Choose the appropriate passive propagation mitigation method based on cell format.
  10. Consider thermal management systems that can also mitigate thermal runaway propagation (e.g., immersion cooling).
  11. Ensure that vent gases from one cell in the pack do not impinge on other cells, and that there are no sources of ignition in the vent flow path.
  12. Perform pack-level testing with live packs to simulate field failures.
  13. Implement a sufficient number of sensors (temperature, voltage, humidity, isolation, vent gas concentration) to capture changes that may indicate the onset of thermal runaway.
  14. Implement early anomaly detection and monitor as many cells/groups as possible to detect thermal runaway well before it occurs. Examples of anomalies:
    a. Excessive self discharge or drop in block voltage during rest periods
    b. Long taper current charging times
    c. Noisy voltage profiles during charge and discharge
    d. Excessive cell heating near the end of charging
    e. Charge capacity being higher than discharge capacity, beyond typical losses
    f. Change in efficiency of charge/discharge over a short period of time
  15. Identify and prevent any sources of high-voltage arcs, and use fast-acting fuses.
  16. Publish standardized and detailed emergency response guides for first responders.

Cell Manufacturing

  1. Ensure that cells are designed specifically for the given application to avoid abuse (e.g., separator thickness and coatings are different for consumer electronics and EVs).
  2. In addition to component level calorimetry, characterize heat flow in a full cell to determine contributions of each component during thermal failure.
  3. Even if minor changes are made to cell chemistry, assess cell safety comprehensively.
  4. Firmly establish with customers the cell’s operational limits and failure boundaries.
  5. Develop robust systems to identify manufacturing issues that can affect safety behavior.
  6. Continually inspect manufacturing quality and subject production cells to periodic testing to remain vigilant of drifts.
  7. Apprise customers of chemistry/construction modifications with implications on safety.

Can batteries fly safely?

Electric aircraft require very high specific energies, given that weight is a crucial metric for aircraft design and operation. Electric vertical takeoff and landing (EVTOL) aircraft also have very high power requirements for takeoff and landing, and the use of fast charging is desired to reduce aircraft downtime and to maximize revenue. In aviation, several high performance characteristics of batteries that increase the risk of safety related events converge. On the other hand, aviation is also a sector where safety is paramount and arguably even more important than ground-based transportation. This makes the battery safety question for aviation unique and challenging, but since cost is not quite as strong of a driver in aviation, interesting approaches and solutions will be taken to achieve the safe design of batteries.

Traditionally, aviation safety regulations for batteries have focused on auxiliary power units which are <10kWh, and the regulations around batteries for propulsion are still currently under development. Current standards based on DO-311A penned by RTCA Inc. for rechargeable battery packs used in aircraft provide a useful reference; however, these are not necessarily intended for propulsion batteries, which are several fold larger than batteries for auxiliary power. Both The European Union Aviation Safety Agency (EASA) and The Federal Aviation Administration (FAA) are actively proposing new certification standards for batteries used in aviation.

New Chemistries, Stationary Storage and Second Life

The industry is also anticipating the arrival of solid state batteries (SSBs) and Li-metal anodes this decade. The absence of liquid electrolytes is expected to improve safety, with the total heat released during thermal runaway estimated to be around 30% of conventional LIB chemistries. Still, the danger of failure propagation continues to exist, and the effects on the statistics of failure occurrence are still unclear. For example, Mercedes recalled an entire fleet of e-Citaro buses equipped with SSBs after a major fire incident at the Stuttgart bus depot. The UN 38.3 regulation already mentions Li-metal anodes, but this actually refers to primary cells and thus not relevant to large rechargeable energy-storage systems such as EVs. Coupling Li-metal anodes with LFP cathodes or with non-flammable electrolytes are some approaches that can reduce the severity of failure in EVs, aviation and grid storage.

Failures of LIBs deployed in large, grid-scale storage systems may not put the public in danger of a fire risk, but they can pose explosion risks to first responders. Safety standards for stationary applications, UL 9540, NFPA 855 and others continue to evolve and several shortcomings have been identified. The primary goal, similar to EVs, is the mitigation of failure propagation between adjacent storage units. Na-ion batteries are also being touted for future grid-scale storage systems, although the safety considerations of these batteries are unclear at this time given that the electrolyte still contains flammable components such as alkyl carbonates.

Several efforts have been made to employ LIBs from EVs in second-life applications for stationary storage. While batteries become more thermally stable over time, the incidence of failure also tends to increase over time. Currently, the only standard that addresses repurposing batteries for second-life applications is UL 1974. The availability of automotive-grade battery modules in the market has also spawned an EV conversion segment. Safety in converted vehicles can be challenging to evaluate and adequate care must be taken to ensure that the original thermal management and propagation mitigation systems remain intact. This is especially a concern if modules are salvaged from vehicles that have been involved in a collision.

Prevention is better than cure

Battery safety is a complex problem involving both electrochemistry and thermal-fire science. Failure characteristics are highly dependent on cell chemistry, format, pack design, and the nature of the application. It is particularly challenging to predict failure because of the inherent variability and the time delay between when design decisions are made and when catastrophic events occur in the field. Battery safety must remain a critical consideration throughout the product life cycle, from manufacturing, transport, and application to second life and recycling. OEMs must take a robust testing-driven approach to ensure that EVs exceed the mandated safety standards and go beyond the bare minimum. Fires in EVs have impacts beyond the vehicles themselves, requiring us to rethink infrastructure such as parking garages.

Large-format cells such as 46800 and prismatics have more severe failure energetics, although fewer cells in a pack also offer the opportunity for more extensive cell-level monitoring to diagnose anomalies. Solutions such as non-flammable electrolytes and modifications to collector design should be given serious consideration given their potential for easy integration into existing manufacturing processes. While there will be trade-offs with performance, an inherently safe cell can reduce thermal management requirements and inspire novel packaging methods that can improve energy density. OEMs, based on data from just Tesla, must not be drawn into a false sense of security that all BEVs will show low battery failure rates. Given that safety-related incidents have partially contributed to bankruptcies in the past, the industry is well incentivized to lean on the side of safety when designing battery packs to avoid major recalls, and the associated financial and reputational pitfalls during this moment of industry-wide transition.

Sriram Bharath is a researcher at University of California, Berkeley. His area of research is in thermo-fluids, fire science, microgravity and lithium-ion battery safety. His work has been published in Physical Review Fluids, Science Advances, Combustion and Flame and the Proceedings of the Combustion Institute, and featured by Axios, the Discovery Channel, Cambridge University Press, NSF, APS and Smithsonian Magazine.

Shashank Sripad is a Presidential Fellow at Carnegie Mellon University advised by Venkat Viswanathan. His academic research focuses on modeling, simulation, and testing of Li-ion and Li-metal batteries for electrified terrestrial and aerial mobility systems. His research has been published in journals such as ACS Energy Letters, Journal of the Electrochemical Society, and Nature Energy. His work has been featured in Forbes, WIRED, MIT Technology Review, Quartz, Axios, and other media outlets where he also regularly provides commentary on developments in the electric mobility and Li-ion battery space.

Interested in publishing with BatteryBits? Apply at this link to become a contributor.

Join our discussion with other battery professionals at the community slack space Battery Street!



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store