Fast charging is a compromise. The problem in real charging is: Temperature. What do you do with a car that arrives at SuC station with a battery that has 50°F?
Tesla encounters the problem by high initial current to heat up the cell by its (relatively high) internal resistance. But apparently they underestimated Li-plating a bit, especially for cells beyond BOL. They corrected now thermal management in a very agressive way, which helps to save what can be saved at the risk of a higher cathode degradation.
I posted the graph to help users to understand better the way they should charge in order to mantain longevity of their packs.
And for the versions of NCR18650. Forget about the rumour that Tesla has added a miracle powder to prevent Li-plating. They were using a (V)HE cell with graphite anode, thats all. No room for miracles!
And please don't bother that only Tesla batteries show Li-stripping after longterm storage. This is a common mechanism.
Thanks for the detailed reply! I'm not disagreeing with you at all. I just wanted to show the relevant charts for supercharging from the paper as well and that the exact charging protocol can make quite a difference to cycle performance. I might have misinterpreted that you wanted to say the 3A CC-CV curve was equivalent to a SuC session, which it is obviously not. Sorry for that.
I agree that Tesla was probably pushing the boundaries of supercharging too far, which is par for the course with Elon - he pretty much has the opinion that if you're not failing from time to time, you're not trying hard enough. Unfortunately, an unmanned rocket blowing up on landing does not carry as much concequence as tens of thousands of customer cars... I just hope they have learned from that. The 250kW charging on the Model 3 smells fishy.
I'm not saying they performed any sort of miracle. Just that they stress at every possibility that the chemistry in the cars is different from any cell on the free market. And improvements to internal resistance are comparatively easy compared to capacity increases. The NCR18650PD came out in 2011, there have been some significant incremental improvements in that field since then. Just look at the power output of Tesla's respective flagships. The original P85D could hardly manage to deliver more power than the good old P85 (see: 691gate), but soon after there were new battery revisions and a current P100DL absolutely crushes the P85D. Of course the 100 has 18% more battery cells, but the power increase is much more than that. So at least the power delivery has improved massively, I think it is reasonable to suggest that an uptick in charging capability came with it.
I never disputed that Li-stripping was a common mechanism, I just pointed out that Tesla cells (and their relatives...) are apparently quite good at it, compared to some other chemistries (for example the Samsung 30Q).
May I quote the relevant section from the dissertation:
"
Recovery Effects
Different recovery effects have been observed in the three aging studies presented in this thesis. The capacity recovery of cells stored or cycled at high SoC only could be explained by lithium diffusion into the overhang areas of the anode. The amount of lithium stored in the overhang areas becomes inaccessible for charge-discharge cycling. However, a considerable amount of lithium returns when the cells are at low or medium SoC for a longer time. In contrast to this regeneration mechanism, the recovery effects after severe degradation owing to lithium plating have not been fully understood yet. The mechanisms behind short-term capacity recovery and long-term recovery of electrode-specific storage capabilities should be subject to further research."