Hi everyone,
I've been a longtime silent lurker, but this topic certainly has grabbed my attention a lot and so I decided to make a lengthy post in the German Tesla forum (tff-forum.de) a few days back. It was well recieved and several members suggested I also post my findings here, which is what I'm going to do now.
Short background about myself: I have a MSc in chemistry with focus on functional materials and while my research topic right now is not about batteries, I regularly read research about Li-ion technology and consider myself to beat least somewhat knowledgeable in the topic. I also tinker with some applications for 18650 cells as a hobby and regularly read some discussions in the Second Life Storage forum where members are making stationary solar storage and similar things from discarded laptop and tool batteries. Whether this makes sense or is safe might be worthy of discussion, however you cannot deny that the members have collected a lot of knowledge about failure modes of 18650 batteries over the years.
Anyway, all the talk about Li plating and dendrite formation reminded me about a well-known problem in that community. A few years back, many members started to notice that older Sanyo 18650 cells often start to heat up upon charging. The cell looks fine at first glance (normal voltage, low internal resistance, reasonable capacity) and starts the charging process normally. However, once the charge voltage reaches around 4.0V, the cell begins heating up, often leading to dangerously hot temperatures of 80°C or higher. The reason is a parasitic internal current that soon exceeds the charging current making a full charge impossible. At first, only Sanyo UR18650A cells seemed to be affected, however by now examples are known from nearly all older Sanyo cell types as well as some Panasonic, Sony and LG batteries. All reasoning about the underlying causes was speculation until one member who investigated the problem on his job posted one very interesting contribution:
Red Sanyo 18650 Cells Getting Hot While Charging - Page 10
Short summary: If charging conditions deviate a little bit from the norm (temperature too high, too much time spent at high SoC), micro-dendrites form between cathode and anode, leading to an internal short-circut. This is the exact failure mode of Li plating. The effect usually vanishes after cooling down and the cells stabilize between 3.95 and 4.05V. It might even disappear completely for a few cycles, but eventually always comes back. Also, the onset is relatively independent from age (both temporal and cycling) and can occur as early as after one year. The effect starts to manifest as an elevated self-discharge rate and worsens from there. As a safety measure, the original poster recommends charging only to 4.10V (corresponds to a "standard charge" in Tesla terms) and reducing charging current and temperature while charging. No failures were observed with these precautionary measures.
Let's just hypothetically assume, Tesla found exactly such a problem in some batteries. How would that manifest?
- Balancing problems and self-discharge at high SoCs lead to some cars being unable to charge to 100%
- High SoCs suddenly make the cooling system work much more than usual due to heat generation
- In the worst case, single cells cross the threshold to thermal runaway and cause cars to burn up
- Tesla takes notice and distributes new software that looks for elevated self-discharge and sets charging voltage to 4.1V or lower on those cars. DC charging speed is reduced for all cars as a precautionary measure.
Correct me if I' wrong, but this seems to be pretty much exactly what is happening.
I will tell you in all honesty: As a Tesla fan, having ordered a Model 3 and holding TSLA shares, I very much hope that Tesla will find a solution to the problem. But everything looks very sketchy. Not only for Tesla, but for a lot of others as well. Panasonic might have to take partial blame - they must have known about the safety problems with their UR18650 cells for at least ten years, but never even notified their customers. And most other battery giants have been shown to deliver cells with similar problems (albeit not in the magnitude seen at Sanyo). I wish Tesla would just come forward with a meaningful statement.
I want to make clear that I do absolutely not suggest that most older Tesla packs are defective or dangerous. If I had to guess I would think that in very few select cases, problems as described above were observed and Tesla, fearing a PR nightmare, decided to play it safe with all packs that even remotely look like they might develop that issue in the future while they are working on a tool to reliably weed out any seriously damaged packs. Tesla batteries have been shown to be able to lead long and safe lives with minimal degradation repeatedly. A few black sheep do not change that big picture, as long as Tesla is able to identify them correctly.
What are your opinions?