Over-discharging lithium ion cells and the effects

Found this good read on over discharging lithium cells as i recently had 1 volt accident:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4957210/
Could someone elaborate on the -12% SOC figure?

Without having RTFA

SoC endpoints - 0% even more so than 100% are arbitrarily defined, ideally by the end-use system designer or owner, based on their use case, preferences, desired longevity vs need for range etc.

Obviously the 0% isolated-at-rest point should be defined so that it's nowhere near any danger of causing any permanent damage to the cells.

The LVC point needs to be much higher than that, since very low C-rates will result in a DoD much lower than high current loads.

So when doing destructive testing pulling every last mAh out of the battery rendering it worthless scrap, that capacity will register as a negative number, relative to the defined "safe 0%".

Is that what you meant?

Depending on the cells. My A123 20ah cells have been used a couple times to zero. Yes drain. And came back with or no negative affect. 1,490 cycles. 6.5years still going strong. I still get 17
5ah bit I work up hill down hill.

LFP is a completely different behaving chemistry from the other types of LI.

A123 is famously top-notch quality, but rarely used anymore given the huge cost per kWh.

And they can also be completely destroyed by going dead flat.

If just say a well coddled set could go 10,000 cycles, and your event lost a thousand of them but you won't know for decades

would you say they were "damaged" or not?

john61ct said:

Without having RTFA

SoC endpoints - 0% even more so than 100% are arbitrarily defined, ideally by the end-use system designer or owner, based on their use case, preferences, desired longevity vs need for range etc.

Obviously the 0% isolated-at-rest point should be defined so that it's nowhere near any danger of causing any permanent damage to the cells.

The LVC point needs to be much higher than that, since very low C-rates will result in a DoD much lower than high current loads.

So when doing destructive testing pulling every last mAh out of the battery rendering it worthless scrap, that capacity will register as a negative number, relative to the defined "safe 0%".

Is that what you meant?

Click to expand...

More or less. So if manufacturer specifies brackets of 2.8V and 4.2V and within that range states a capacity of 2000mah at 1C, -12% would render voltage of 2.46V which, regarding article would render little to no risk of copper electrods being dissolved in electrolite thus no damage to cell?

agniusm said:

if manufacturer specifies brackets of 2.8V and 4.2V and within that range states a capacity of 2000mah at 1C, -12% would render voltage of 2.46V which, regarding article would render little to no risk of copper electrods being dissolved in electrolite thus no damage to cell?

Click to expand...

I include both lowered cycle lifetimes and increased risk of fire in my definition of "damage", don't you?

As for the "risk of instantly reducing performance" or "risk of rendering a pack into scrap", I am not qualified to judge, and consider the issue irrelevant.

But I'm sure the specific numbers vary a lot between say NCA and NMC chemistry cells, or between different brand formulations.

According to this article, up to -12% does nothing. I am speaking on single occurances not constant use.
I had brand new cells driven to 1V. I have charged them up and then discharged. They gave out 2340mah vs 2050mah declared by manufacturer. This got me into reading how it effects cells in real life. I know they are scrap, still good for prototyping

No not scrap if they CC-discharge test out, no loss of immediate capacity, fine even for production use. A before / after to see if ESIR is impacted would be interesting too.

In the long run we're all dead, a meteor could wipe out life on the planet before you actually get impacted by the lost cycles off the back end.

Just don't want people left with the impression it's OK to do intentionally. Of course stuff happens.

I dont have anything to measure IR except icharger but that probably is as good as taking a guess. Bummer as i have spare cells from the same batch. Maybe ▲ would still be interesting even if not accurate

The iCharger may well be fine, but you always need to standardize all the conditions for comparability.

Especially precise internal temperature

but also connector resistance, SoC, current rate, what else?

agniusm said:

More or less. So if manufacturer specifies brackets of 2.8V and 4.2V and within that range states a capacity of 2000mah at 1C, -12% would render voltage of 2.46V which, regarding article would render little to no risk of copper electrods being dissolved in electrolite thus no damage to cell?

Click to expand...

agniusm said:

According to this article, up to -12% does nothing. I am speaking on single occurances not constant use.
I had brand new cells driven to 1V. I have charged them up and then discharged. They gave out 2340mah vs 2050mah declared by manufacturer. This got me into reading how it effects cells in real life. I know they are scrap, still good for prototyping

Click to expand...

I had a similar case (forgot to turn off Controller over winter, no BMS, battery was at 1,5V) so i read the article.
It is not totally clear what they mean with "-12%", but from what i understand when looking at the graphs i think they can be interpreted like this:

100% SOC means 4,2V or fully charged, and 0% SOC means 0V.
If you have a cell with lets say only 0,5V left, it still will output a little bit of energy until it is at 0V.
It seems they are talking about one bad cell beeing part of large pack with more cells in series, which would mean that current will continue to flow through the cell and turning it's polarity (+ becomes -).

These -12% now could come from "current x time" or/and max voltage difference the cell ever had while beeing operating in such reverse polarity condition.
Above -12% (well less than), the copper starts to build bridges between cathode and anode making internal shorts.

Hm, i know about cells which can be charged back up from 0,00V (older Sony VT or VTC cells for instance), but i also know about cells which just get warm and take no charge (not so old cells with more mAh), but yet these never should have had negative voltage across them which makes me think that those internal shorts could also be caused by other factors.

OT I know but apparently LTO can go to dead dead flat without losing any of its 10000 lifetime cycles

Also great performance in below freezing weather, even charging.

But poor density.

madin88 said:

0% SOC means 0V

Click to expand...

I can't imagine even the most ivory tower disconnected from reality academic setting that hypothesis with those chemistries.

Would make every benchmark SoH cap test a destructive one, need a fresh new battery for every cycle?

With lead the industry standard across the board is 1.75

With all the various "LI" chemistries varying so widely, best to just accept (as a stressful maximum for testing only, not for normal usage) the maker's spec, often 2.5V,

hence "overdischarge" meaning going below that, and negative percentages make perfect sense.

Without having to get into disastrous polarity reversal territory

john61ct said:

madin88 said:

0% SOC means 0V

Click to expand...

I can't imagine even the most ivory tower disconnected from reality academic setting that hypothesis with those chemistries.

Click to expand...

You have right. 0% on this particular test means 3.4V

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4957210/#__sec1title

Below 3.4V, decomposition of SEI (solid electrolyte interface) starts to begin and this process generates gases and the cell already could suffer damage, but not by all means (as they were able to recharge it), however they mentioned something about increased internal impedance.

At around -7% it's crossing the 0V mark and with further current flow it changes it's polarity and decomposition of SEI i going on until -11% (-2.2V) where this layer is broken or "leaky" and from here on copper starts to build bridges between anode and cathode (internal shorts -> self discharge).

What about the knee at 1,2V? It wasn't explained but might tell something?