The following is a copy of an article originally published by the Volta Foundation for their BatteryBits series that can be found here. It was written with support from the Electric Goddess team.
Understanding the cyclic corrosion processes that occur within a mass-produced lithium-ion cell plays a critical role in the design of a battery pack. While the redox reactions of the lithium and electrolyte with the anode and cathode during cycling are fundamentally important to cell operation, they are not the biggest threat to long-term reliability and safety. Further, production cells are quality tested and screened so that their specification and performance are well understood. Hence, the likelihood of serious cell defects making it into end-use products is uncommon. The leading threat to battery packs, however, is the corrosion that can occur on external surfaces of components and interconnect bussing of cells.
Corrosion in a battery pack refers to the slow electrochemical reaction of components with oxidants such as chlorides, sulfates, carbonates, oxygen, and hydroxides. A well known example of electrochemical corrosion is rusting, or the formation of iron oxides. The rusting of iron produces various oxide states and salts which show distinctive colouration from red-orange to black. When corrosion occurs in materials other than metals, such as ceramics or polymers, the term “degradation” is more commonly used.
Corrosion degrades the useful properties of materials and structures including strength, appearance and permeability to liquids and gases.
Condensate pathways are formed inside battery packs when humidity condenses onto internal surfaces as liquid droplets, which then wick into interfaces between conductive surfaces via capillary action. Corrosion is then driven by voltage gradients across these aqueous condensate pathways in cells, cell interconnects, thermal management interfaces, connectors, bussing, etc. Droplets wick proportionately to the surface energy of the interfaces to travel long distances and form connecting capillary pathway networks between metal-metal, polymer-metal, and polymer-polymer surface interfaces. Even if the condensate droplets are originally formed with low ion content, exposure to atmospheric CO2 (410 ppm) results in the formation of carbonic acid (H2CO3). This acid functions as an electrolyte that, when combined with a voltage difference, rapidly corrodes and consumes nickel, copper, aluminum, iron, tin, and more.
Even gold, platinum or tantalum plating of conductive elements in the battery assembly cannot protect against dissociation potentials over ~6 VDC The corrosion of these platings then exposes the base metals to EV battery electrolysis potentials. Since the corrosion process is slow, the liquid water may evaporate before the metal has progressed to failure. Still, the dissolved salts are left behind as salt residues that offer favorable pathways for wicking condensed water, reinforcing the corrosive pathway with each subsequent exposure.
If the salt residue-hydrate conversion cycle reaches the correct hydrate state, a molten ionic liquid can be created at room-temperature, which suddenly conducts electricity very effectively.
Additional vulnerable locations in batteries include circuit boards, connectors, flex circuits, sense tap and temperature sense harnessing, and any interfaces that offer both a voltage difference and the opportunity for condensates to wick across. During a battery's use, it can be exposed to atmospheric gases including water vapor, tire spray from salted winter roads, ocean mist, powerwashing, and cleaning detergents and solvents.
In addition, design characteristics of the battery pack can influence gas exchange and the rate of corrosion:
Corrosion is a gradual process that steadily removes material from the anode side and deposits corrosion byproducts on the cathode side. Since the voltages present in batteries greatly exceed the galvanic series, the influence of the metal composition is insignificant compared to actual voltages imposed directly across the ionic pathway.
Battery systems under-protected from corrosion can show the effects of failure weeks to months, or even years after a product enters the wild. The onset of failure depends on several factors, such as temperature, humidity, free gas volume, voltage gradient strength, available surface metal ions. Monitoring the chassis-to-cell string isolation resistance by the BMS is often the first available warning sign that a corrosion process has been taking place. In these situations, actions must be taken to prevent the risks of a fire hazard. It is much easier to discover evidence of corrosion-driven failures in the early stages. After corrosion has progressed to molten salt-bridging and fire, identifying root causes may be challenging.
Clear evidence of corrosion-induced failure can sometimes still be apparent post-event, but requires more tedious forensics, so it is always ideal to analyze isolation faults that may have appeared prior to a severe thermal event.
One method to conclusively determine if a chassis isolation fault is related to a corrosion path is to perform bi-polar cyclic voltammetry sweeps over a long time interval. For example, it might take 5 minutes to sweep up, and another 5 minutes to sweep down voltage between the chassis and each end of the battery string assembly. Non-linearities in the sweeps would be the signature of an electrolysis pathway. High resistance (>1 kΩ) isolation faults in EV packs are often induced by condensate paths and may cause the chassis to take on the voltage of the interconnect string along that path. These condensate pathways tend to become more conductive over time.
One common factor accelerating corrosion is the cooling of gas inside the battery assembly in the presence of high humidity or corrosive vapors. Testing at high temperatures in humidity or salt spray does not validate vapor ingress resistance, as the contraction of internal air volume in the pack enhances the mechanics of ingress due to the favorable pressure differential.
While the isolation fault remains a single-point failure, the battery can still be safely decommissioned. When the liquid path bridges between the chassis and any secondary voltage potential in the EV assembly, highly accelerated corrosion occurs, driven by the high pack voltages in EVs. These potentials can split the surface species into radical ion forms and/or re-bond radicals into new electrically-formed species. Even very reactive and short-lived species and compounds can be evolved steadily through sustained electrolysis.
Safe, reliable battery design solutions that mitigate against corrosion are critical to the electrification of transportation. A battery system that does not survive in the wild will ultimately be more expensive than a battery that can be used over its full intended lifetime. Designing to mitigate corrosion not only prevents recall and safety hazards, but also decreases waste across the industry. Second-life battery applications, which are of increasing interest to promote sustainability, are not practical or possible without batteries surviving their first-life.
While there are many battery safety test standards in use today, they fall short of replicating real-world battery operation and do not test in multi-metal ion salt fog blends while the pack undergoes temperature swings. For instance, the published Salt Fog Test standards from UL, SAE, ASTM all use 99.9% NaCl blended with distilled water as the test specification. The upside of this is a repeatable test anywhere in the world. The downside, however, is that this test is too easily passed, leading to a false sense of product integrity. When the product is exposed to the compositions of road salt mist that are actually sprayed onto the vehicle during use, thermal runaway can be induced by the many ions other than Na+ and Cl- that are more prone to forming molten salt bridges at low melting temperatures.
Safety standards and test conditions need to be updated to include the multi-ion salt mixes present in real road salt and marine sprays. These include Na+, Ca2+, K+, Mg2+, Al3+, paired with Cl-, CO32-, and SO42- ions. Endemic sources of metal ions in the battery assembly or BMS may include Fe+, Sn2+, and Cu2+ ions.
All of these salt compounds dissociate in molten form to free radical ions between 1.5 and 4.5 V of potentia.. While these compounds have negligible conductivity in their crystalline state, they become impressive conductors in their molten ionic state and are capable of rapidly discharging at 10’s or 100’s of kW from the battery assembly, the localized heating leading to runaway.
These are several methods to mitigate the risk of corrosion in battery pack design:
Voltage Design Considerations
The voltage gradient creates dielectric stress in the materials, which can lead to electromigration and dielectric decay. To improve heat transfer, packs are designed with minimal distance, often 200 μm or less between the cells and a common heat transfer system. In a 400VDC system, this creates a dielectric stress of 2MV/m. While this is not enough to instantly break down many dielectrics, it is sufficient to drive ionic migration and decomposition in many materials over time. Furthermore, when materials are saturated with humidity, effects induced by ionic migration occur much more rapidly. Polymers in the nylon family exhibit these effects quickly, and, while materials like carbonates, epoxides, and poly-amides tend to be slower to fail, arc tracking plasma heating and fire is the typical failure mode. Polypropylene, urethanes, HDPE, polyureas, and fluoropolymers tend to exhibit excellent long-term resilience to dielectric stress, but individual stackup material compatibility testing should be performed with the representative interface materials and increasing voltage stress, including stress-to-failure testing. It is important to understand that a hi-pot test, or any voltage stress testing that is completed in less than a day can not measure dielectric decay, but, rather, is testing for instantaneous dielectric breakdown, which is largely unrelated to its dielectric decay and the mechanics of ionic migration.
Electromigration occurs when metal atoms are transported along the paths of peak electrostatic field gradients induced by a voltage differential. Similarly, in a dielectric material, dielectric decay occurs when ions move through the material due to a voltage gradient caused by an electric field. Both of these processes exacerbate corrosion: electromigration forms pits that preferentially corrode whereas dielectric decay degrades formerly protective polymers and coatings. Migrated ions in the polymer tend to improve ionic conductivity, enabling additional ions to be ripped from the metallic surface and reinforcing the corrosion pathway.
Use Potting in Battery Packs
Since weight is a strong consideration for pack design in vehicles, utilizing a potting material in the battery pack and electronics can seem counter-intuitive as it adds weight and decreases the gravimetric energy density. Despite this drawback, potting is often the best option for ensuring pack reliability with the lowest size and weight penalties, and decreases the risk of corrosive failure and associated fire losses and/or recalls.
Potting materials that are endothermic fillers undergo a one-time thermo-chemical decomposition, which limits cell-to-cell heat transfer, enabling very tight cell spacing while maintaining outstanding resistance to thermal runaway propagation and limiting the penalties on pack energy density.
Conformal coatings designed specifically to protect circuit boards sometimes yield similar failure modes during corrosion testing, albeit with marginally longer time intervals when compared to a bare assembly. The ‘dam-and-fill’ method for protecting PCBAs is a repeatable method of ensuring defect-free encapsulation and yields excellent results when exposed to mixed salt sprays.
Not all potting materials are created equal. Some oxamine cure silicone based materials are not recommended as ionic migration can occur under prolonged DC voltage stress, whereas platinum cure silicones generally do not exhibit this effect. Potting is composed of a polymer material, which is injected into the battery pack between cells, or used to protect electronics. In applications dedicated to optimizing battery safety and weight such as aircraft, multi-layered potting can be used, with different layers carefully selected for mechanical strength, cell failure protection, and corrosion prevention.
The activation of CIDs in cylindrical cells is often a concern in potted battery packs. Based on many internal tests of CID activation in assemblies with and without potting, however, there is no statistically significant difference in the activation pressure. This is because internal activation pressures of the CID diaphragm are about 350-500 psi, significantly greater than the resistance offered by 500µm of displacement in the soft durometer potting compounds.
Since pouch cells expand as they age (6-12% for Gr/NCM), it is critical to account for this space in the pack at the beginning of life. Pouch cells must also have low-density displacing materials for each pouch that can displace the local potting material to accommodate gas headspace. For most pouches, this is the area where the tab is encapsulated in the pouch. This is the only meaningful gas volume inside the pack envelope where gas presence does not impact the cell’s electrochemical performance.
Without potting, the battery is exposed to air, humidity, and other environmental contaminants that can initiate and exacerbate corrosion. A common misconception is that using a robust seal is good enough to keep water and contaminants out of the pack. In reality, not only are the particles of water vapor smaller than droplets of liquid water, but water itself is a smaller and lighter molecule than the N2 and O2 composing the bulk of our atmosphere.
The higher the voltage gradient, or higher the battery voltage, the higher the force of these processes.
Low voltage batteries mitigate these issues by decreasing the voltage gradients and, thus, the risk of electromigration and dielectric decay failure. More information on low voltage battery design and other benefits are available here.
In industry, the reasons publicly stated for recalls are seldom related to the actual causes of the recall. Corrosion-induced failure modes and forensics are unfortunately often misdiagnosed by industry experts even today. Industry standards for safety testing must seek to assess susceptibility to corrosion-induced failures and develop test methods that can reliably recreate long-term failure modes, including sprays of several road salt blends. Such tests must be accelerated to enable multiple iterations between corrosion resistance design and testing validation. High-voltage architectures have increased corrosive driving forces, as well as dielectric decay and ionic migration risks, of which there is limited awareness in the industry today.
While the cell scientists in industry have taken efforts in taming the corrosion inside of the cell, the effects of corrosion outside of the cell is the leading reason for EV battery replacement today. Battery design that considers and addresses each corrosion risk area, from the cells to the connectors, as well as robust safety testing that mimics real-world environmental conditions are crucial for reducing recalls and fires.