Interpreting QuantumScape’s Safety Test Results

Hazard Levels

A common way to evaluate battery safety tests is a grading system known as Hazard Levels. Each Hazard Level (HL) describes the behavior of a cell when subjected to a particular test – the higher the number, the more serious the safety risk: from HL0 (no reaction) to HL7 (an explosion). Many tests, such as nail penetration and cell crush tests, are destructive in nature, and so a result of HL 1 or below may not be possible in every case. For electric vehicle batteries, HL 4 or lower is typically considered a passing result, whereas HL 5 or above is nearly always considered failing.

Test Protocols

There are many sets of tests used to demonstrate battery safety in a variety of situations. A few common examples are:

• UN 38.3: This is the international standard for testing potentially hazardous goods for shipment. It subjects batteries to conditions they might experience during shipment but does not necessarily reflect the safety performance of the battery in its intended application.
• EV-specific test packages: EV batteries may be subject to a variety of test standards depending on the automaker and jurisdiction. A few noteworthy examples are EUCAR, SAE J2464, and IEC 62660-3. The precise content and requirements of the tests varies from standard to standard, but there is also significant overlap between the types of tests performed.

Each of these standards requires a panel of different tests, but it’s worth looking at a few specific examples in more detail. Some examples of typical individual tests are:

• External short-circuit: This test simulates what might happen if the two terminals of the battery (positive and negative) are accidentally bridged, which would result in the immediate discharge of the cell and release of the cell’s stored energy. This would generate significant internal heat and might trigger thermal runaway under certain circumstances. This test does not necessarily stress the fundamental chemistry of the cell, since there are also ways to design a cell, such as adding an internal fuse, that can stop it from heating up in the case of an external short.
• Overcharge: This test examines what could happen in case of a failure in a car’s battery management system (BMS). In this test, the battery is charged above its rated voltage, which may cause the cell components to break down or produce hazardous byproducts.
• Crush: The crush test simulates the mechanical deformation a cell might experience in a severe collision or other major safety incident. In the real world, such a dramatic abuse condition represents a worst-case scenario, but it is important for safety testing to characterize cell behavior even in such events.

It would be impractical to give an exhaustive list of all the variations of all the safety tests required by every safety protocol in every jurisdiction. However, there are two tests in particular that are worth examining in more detail: nail penetration and thermal stability.

It’s important to preface these results with a few caveats. First, these results are from prototype cells on the pathway to our QSE-5 design, not our final QSE-5 cells. Since product safety is a function of the precise cell design, the safety profile of the final cells may be different from these prototype cells. This is because safety is a function of cell materials and design, and consequently, when a given cell design changes, even slightly, the cell itself will have different physical characteristics and potentially a different safety profile. Any new cell design must therefore be re-tested to establish its behavior under abuse conditions.

Additionally, safety testing requires many types of tests using many individual cells to ensure statistically reliable conclusions about safety performance. Nevertheless, we believe the results presented here are a promising early indicator of the strength of our safety performance.

Nail Penetration

Nail penetration might be the most well-known battery safety test. It is certainly quite visually dramatic: In this test, the battery is punctured by a sharp metal spike to create an internal short-circuit. The purpose of this test is not necessarily to see whether the cell safely resists a puncture, since in the real world such an occurrence is unlikely (and highly inadvisable). Rather, the goal is to simulate how the battery might respond if an internal short were generated by a manufacturing defect, such as misalignment or metal fragment.

However, this test is not as straightforward as it may seem. The cell’s reaction depends heavily on precisely how the test is performed: factors such as speed and depth of penetration and the nail geometry can all significantly affect the result of the test. Counterintuitively, nail penetration does not always actually generate a short-circuit in the cell, and if a short isn’t created, then the nail penetration test does not test the consequences of an internal short-circuit at all. Nail penetration results cannot be interpreted without verifying that an internal short was created, by observing the characteristic voltage dip in concurrent electrical testing.

With this in mind, using energy-dense 24-layer prototype cells based on our Alpha-2 design, the QS performance on the nail penetration test shows that our solid-state lithium-metal system delivers good safety results under nail penetration testing. The nail penetration successfully induces an internal short-circuit inside the cell, but the cell does not enter thermal runaway, instead demonstrating only a modest temperature rise. The result is a Hazard Level of just 2, an excellent result for this type of test. We believe this good performance is attributable to the mechanical and thermal stability of our ceramic separator which, in contrast to conventional polymer separators, does not melt during the test.

Thermal Stability

The thermal stability test is straightforward: the cell is placed in a test chamber and heated up to simulate how the cell reacts to elevated external temperatures. This is the main mechanism by which EV battery packs experience catastrophic fire, as thermal runaway in one cell propagates to neighboring cells by heating them up until they experience thermal runaway themselves.

While the thermal stability test is straightforward, it is also very challenging for a battery to endure. Heating the cell directly tests the fundamental chemical stability of the materials in a way that translates to the real-world safety profile of an EV pack. We believe this test, better than any other, shows the limitations of liquid-based lithium-ion batteries and their fundamental susceptibility to thermal propagation. High-quality commercial battery cells, which pass many other safety tests with low hazard, can experience catastrophic failure when heated above a certain temperature. At 184 °C, the conventional lithium-ion battery cell we tested experiences a Hazard Level 6 failure: fire or flame with ignition and sustained combustion.

In contrast, using energy-dense 24-layer prototype cells based on our Alpha-2 design, QuantumScape’s solid-state technology demonstrates improved safety characteristics. When heated all the way to 300 °C, the QuantumScape battery passes with a result of Hazard Level 3: minor venting of gas from the liquid catholyte is the only significant effect.

We believe this demonstrates the fundamental safety advantage of our ceramic solid-state electrolyte-separator: because the ceramic is nonflammable and noncombustible, it essentially creates a series of firewalls as part of the electrochemical stack itself. With 24 of these ceramic sheets layered into each full cell, we expect that thermal runaway should be less likely compared to a conventional lithium-ion cell, which is flooded by a single contiguous mass of liquid fuel across all the layers of the cell.

Conclusion

As mentioned above, these early results are from prototype cells, not our final QSE-5 cells, and rigorous safety testing methodology requires a full suite of tests on many such cells before we can draw any firm conclusions about the safety performance of the finished product. Still, we’re encouraged that early prototype cell designs nevertheless display good safety performance.

From a big-picture perspective, a battery is by its very nature a highly reactive system that contains a lot of energy, and so there can be no such thing as a perfectly safe battery. However, we believe it is possible to improve the safety profile of conventional lithium-ion battery technology. At QuantumScape, we believe our ceramic solid-state electrolyte-separator has the potential to simultaneously offer both an improved safety profile and better energy, power, fast-charging, and cycle life, when compared with existing high-energy EV batteries. We also believe there is room to improve our safety profile even further, and we will continue to build, test, and refine our battery designs to achieve these improvements.