Cause of heater-effect at around 4.0V and high self discharge at full voltage (4.2 ... 4.0V)

[paddy72]

Hi all,

you probably all know the problem of so-called "heater cells" that tend to get (very) hot when charging above a specific voltage? Typically this voltage (when heating starts) is around 4.0V - sometimes a bit lower, rarely higher. On the other hand there are cells, that don't keep their full charge at 4.2V for long but tend to drop in voltage down to 4.0 rather quickly and then settle at this voltage to be stable. The last ones must not necessarily be heaters but could be.

I wonder if someone had ever come across a scientific explanation of these effects and if they both have the same cause - what i would assume.
Quite obviously its a matter of high self-discharge and this, in most cases, is caused by small dentrites growing towards the other electrode, penetrating the separator - as far as i read in some studies. But there could also be other effects like degradation of electrolyte or separator. The very small dendrites normally are not dangerous as they "glow away" when hitting the other electrode. Bigger dendrites can lead to severe short circuit with all the bad things that can happen ...

I am really interested to know if there is any research on that topic and what you think about it.
From my experience most of the faster dropping cells from 4.2 towards 4.0 are not completely useless as they are stable below that voltage and still can have a high capacity. But i wonder if these could be more dangerous in some aspect and what distinguishes them from the heaters?

 

[Oberfail]

At nearly 4.0v, there is a big entropy change inside the cell chemistry, causing it to create more heat than usual till 4.1v

View: https://www.youtube.com/watch?v=9qi03QawZEk&t=3307s

Thats all i know for a fact about that phenomena.

One thing i stumbled upon is, that if you charge a heater till it gets warm, take it out the charger and leave it for 2 days and charge it again, it often does not heat anymore, if it does, repeat it 2-3 times. None has blown up so far, but i dont know how safe they are anymore and would consider them a rather bad cell and sort them out for non critical usecases.

//Edit: Toss heaters, if you "unheat" them, they will loose capacity rapidly when using them. After just 4 recharges, mine are down from 1800mAh to under 1000mAh.

So for my powerwall, i want to go for a 3.95v down to 3.4v cycle range.

Some other document that has a more in-depth view on how the cell works inside: https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwjqvObk3LfyAhU0gf0HHXg1BNwQFnoECAYQAQ&url=https://www.mdpi.com/2071-1050/11/23/6697/pdf&usg=AOvVaw3La1kXHOr0kSiNzy7gmwh6

 

[OffGridInTheCity]

I (personally) define heaters as too hot to touch - e.g. somewhere greater than 60C/140F type of thing. There are also 'abnormally warm' cells - not sure the exact temp range but that's not the common case. I toss both kinds.

The majority of heater cases for me were cells <1.0v to start with - e.g. abused / severely discharged cells. Further, the heat seems to have some correlation with significant self-discharge - e.g. as it nears 4.1v-4.2v it cannot keep the charge, and as charging continues to poor in it heats up.

I've read that certain cell types are 'know for high % of heater phenomenon' but have not run into that personally.

 

[paddy72]

@Oberfail: thanks for sharing, very interesting. Prof. Jeff Dahn is talking about LCO-cells and special additives ("violin carbonate"?) that helps against the heating of LCO-cells. Not sure if heaters are mostly LCO chemistry - will have a look after that.

I have the same experience with letting hot cells settle down for a day and charge again - they dont heat up now when charging above 4.0 V or only very little. So it seems like a mechanism as formation for an electrolyte capacitor where the isolation layer of AlO2 has to build up again. You can also charge these heaters very slowly (at less than 100 mA) and they won't get really hot.
But still i wonder if its the same effect that draws cells back below 4.0 V when fully charged? It might also help here to charge very slowly from 4.0 to 4.2 with very small current to let the cells chemistry settle.

I think its a very good choice to go from 3.4 to 3,95V on the average cycle - that would give you much more cycle life. Maybe its a good habit to go from 3.0 up to 4.15V for each 100.th cycle - to "stretch" the cell-chemistry a bit? ;-)

P.S. its "Venylene Carbonate", a typical additive to protect liion cells agains parasitic reactions that form the SEI. Li-ions get trapped over time in the SEI (solid electrolytic interface) and thus reduce coulombic effeciency - which is proportional to the cell degradation (capacity loss).

 

[Oberfail]

I think its a very good choice to go from 3.4 to 3,95V on the average cycle - that would give you much more cycle life. Maybe its a good habit to go from 3.0 up to 4.15V for each 100.th cycle - to "stretch" the cell-chemistry a bit? ;-)

I'll look into it, once my powerwall will be build.

 

[Wolf]

Ah the heater delema again. The number one cause of a "heater" cell is high IR usually >120mΩ. This by itself does not necessarily mean a cell is a heater. Not all high IR cells are heaters but all heaters have high IR. The highest manufacturer spec I have found is ≤100mΩ for some Samsung ICR18650-XXX models. I dare say that if anyone finds a Samsung ICR18650-XXX cell that has any life left at 100mΩ I will buy you breakfast. Yes all measurements are taken with either a YR1035+ or my trusty favorite the RC3563.
Now to some exploration.
Most 18650 cells typically have a current-limiting PTC (positive temperature coefficient) device installed in the cell cap.
The PTC resistance increases sharply with temperature. When high current is applied to a cell, the elevated currents cause the PTC
to self-heat and move to a high-resistance state in which most of the cell voltage is across the PTC but the current is significantly reduced.
The PTC, is a thin ring consisting of a specially irradiated polyethylene laminated with a metal on both sides.
I am wondering if this PTC ring could be the cause of some of these "heaters"? Maybe over time the chemistry of the PTC matrix and laminate start to break down and cause some cells to become heaters The infamous Sanyo reds for example. Could it be that after a charge/heat cycle the PTC "normalizes" and then the cell will charge without heating? IDK. I haven't found a study on PTC longevity as I suppose they are designed for 300 to 600 cycles (maybe) and then the cell is "supposedly" done.
Additionally a little known fact is that the chemical reaction that takes place during charging of a Lithium chemistry cell is endothermic (the reaction absorbs heat). Ha bet you didn't see that one coming.
The reason "good" cells get warm on chargers like the OPUS, LiitoKala, etc is the internals of the charger give off heat. That heat is transferred to the slides and ultimately the battery. Heat rises. Once I put the Fan mod on my OPUS chargers https://www.thingiverse.com/thing:2267571 I had absolutely no temperature increase as I charged cells.
So cells cool during charging but thermodynamics tells us you cant have one without the other so cells will get warm during the discharge as that reaction is exothermic and produces heat.

Wolf

 

[paddy72]

Thanks for joining in, Wolf, and sharing your ideas.

Honestly i doubt that the IR has really much to do with that abnormal heating at around 4.0 V when charging. I have many old cells with quite high IR (although i only tested them with the DC method, where the IR readings are always higher compared with the AC of the YR1030 e.g. - mostly by a factor of 2...3) and there are cells with much too high IR that dont get hot and such with moderate IR that get hot, when charging with ca. 0.7 A.

The PTC is realatively small and doesn't add much weight to the cell at all so i doubt it can significantly contribute to massive heat generating in the cell itself. Its only at the top of the cell and you should feel a big temperature difference over the cells body - but thats not the case. Also, if you charge with 0.7 A the PTC won't get hot at all.

Like Prof. Dahn explains in the video mentioned by Oberfail, there is an area of big entropy change in the cells chemistry at around 4.0V and he shows a chart for that phenomenon. That fact alone still doesn't explain why some cell gets hot (the heaters) and others stay pretty cool. It probably has to do with the kind of electrolyte, its additives (which play a big role in coulombic efficiency and thus Soh) and probably the SEI building up in old cells. As far as i understood the charging process its mostly endothermic but there are also phases which are exothermic - but i dont think this plays a role in ther heater phenomenon. Still not 100% sure about the main reasons for the heaters to get hot at around 4.0 (in my experience some get already hot at 3.8...3.9 V but mostly at around 4.0V) and why other cells tend to drop from 4.2 to 4.0 so quicky and others stay at 4.2 without problems for many weeks.

 

[Wolf]

@paddy72
You may doubt that IR has anything to do with heaters but once you get your YR1030 or equivalent 1kHz AC IR tester I think you may find that that measurement will tell you a lot. As stated before "Not all high IR cells are heaters but all heaters have high IR" moderate or not.
It also depends on what you consider moderate IR. I consider most ICR chemistry cell that are over 65mΩ AC IR to be EOL (End of Life) as in my experience there is no purpose to test them. Therefore I just don't get any heaters, vamp cells or SD's as those cells won't do me any good. Now if you are doing a cell study on how AC IR can "predict" the SoH of a cell yes then by all means do the testing. You see once I went through 3500 cells testing IR and capacity then seeing the results of my Pivot Table, I started to pay very close attention to AC IR and narrowed down the acceptable AC IR range of certain cells. If you want a study head start then download my Harvested Cell Analysis Excel Workbook here. There is some noise in it but all in all it my first study of 6000+ cells. I may actually repeat this study again as I have automated my cell testing process quite nicely. So download it if you want have fun with the Pivot Table, Scatter charts, IR by Part# etc.
Wolf

 

[paddy72]

Yes, your curve is very impressive and i am far behind your testing methods and experience with so many cells. I tested maybe a few hundred cells (600...800) so far, and my testing of IR is poor and only DC-method with very poor replication rate and accuracy. My best-IR tested cells are around 50...70 mR with the DC method, that could mean 20...40 mR AC-IR.
I am not really sure how the IR corresponds to heaters - i didn't test it specifically - i just didn't find any correlation so far. But i absolutly agree that the IR-measurement is a good indicator for Soh and predicted capacity Yes, a good cell won't be a heater, so as a conclusion all heaters must have some higher resistance and low Soh, thats obvios. But i just dont see a correlation for a cell having too high IR versus beeing a heater. There are still enough bad cell with too high IR and are no heaters.

 

[Wolf]

But i just dont see a correlation for a cell having too high IR versus beeing a heater. There are still enough bad cell with too high IR and are no heaters.

That is correct and it reiterates "Not all high IR cells are heaters but all heaters have high IR" .
Which leads to "Most high IR cells have a low SoH"
Why some cells are heaters and others are not is still a good question, one that I have sort of laid to the side as I don't entertain cells that have high IR. What's the purpose? The don't perform to meet my SoH standard and are a waste of time for me.
Now if I had a full scientific lab setup with a bunch of grad students looking for a doctoral thesis on why some cells are heaters and others are not by all means. In the final analysis I'm sure it has to do something with cell chemistry as some manufactures are famous for "hot" cells and others are not. The doping of the compounds within the cells or for that matter of fact the purity of the materials. This is a competitive business after all and most manufacturers will cut corners wherever they can to be competitive and win the bid. If you are tasked with developing 10 million 2200mAh cell requiring 3 years of service in a laptop you may consider how to go about this with the least amount of cost, and since there is an age limit, not consider the repercussions of the usability of the cell after 3 years. In other words who cares? We all assume manufacturers have our well being in mind. Silly me, and I thought they were all honest.
On a side note there have been plenty of Li-Ion battery recalls. See https://www.retailconsumerproductslaw.com/2020/11/recalls-in-review-lithium-ion-batteries/
Maybe some of them snuck through and ended up on out test benches. I don't know. Nevertheless if you are just harvesting for a powerwall and not into a scientific study then the easy answer is high IR leave them alone and recycle.

Wolf

 

[Oleksii]

A few notes about PTC rings.
I once measured what they contribute to total resistance, here are the results:

ESR? Acceptable rates

After a little research I was able to determine that the ESR rating of a battery is essentially the internal resistance of the battery. But unfortunaley I dont seem to understand the signicficance of this number. Obviously batteries with low interal resistance will naturally outlive those with...

secondlifestorage.com

Also, If I remember correctly, if PTC overheats and increases resistance, it does not return back as a permanent destruction happens in it, by design.

I never saw PTC rings in high-drain cells like Samsung INR, Sony VTC. Used mostly in batteries for power tools.

 

[paddy72]

You can find re-usable fuses made of PTC material, so it should come back to (nearly) its normal state after tripping. Not 100% sure the PTC-material of these fuses is the same or comparable to that used in 18650 cells.
I have a few of these PTC-fuses - they work slowly and not very accuratly as temperature rises. E.g. for a 4 A fuse you have to have at least 6 A to trip in my experience but you can re-use it without problems (many times but its limited to numbers around 100...1000 i guess).
Normally you only get these PTC-fuses for lower amperage, the higest i have is around 8 Amps - so i think they dont have PTC-material for higher Amps.

 

[quadrofource]

Hi. yesterday i started testing 4 pieces of sanyo ur18650fb at 1.95v which i put in the recycle box already. slowly charged them to 3v at 40ma. After that i charged them at 1A and they started to get hot at 4.1V. I let them get up to 45C max and then pulled them from charger. Today i measured them and put them in charger. Today they measured min 3.98V and max 4V. They're at 4.17V now without heating up. I will nor test them when the weekend arrives. when i pulled them at 1.95v they measured 55mohms with a y1035+

 

[paddy72]

Thanks for sharing. Keep us informed how they worked and what IR they measure when fully charged

 

[paddy72]

I have an old ebike battery pack, its an ebike-Vision-Panasonic 24V / 21 Ah (rated as 524 Wh) pack and built of 7S8P Samsung ICR18650 26H.
Its quite old and has lost most of its capacity but i occasionally use it for testing purposes and experiments.

The most negative cell block is defective, has high resistance, low capacity and has some heater cells in it.
When i charge the pack with 4 A (500 mA/cell average) a few cells of this block get hot when reaching 4 V. In the beginning (say 10 cycles ago) they didn't get really hot, lets say little more than hand-warm. Now the cells get hotter and hotter at each cycle, so that i reduced the current to 2 A when coming close to 4V. But even than the defective cells heat up to approx. 40-50 C. I let them cool and can slowly charge the rest up to nearly 4.2V Its interesting that, if they had cooled down, could be charged again up to nearly 4.2 V without a big problem.

The last time i let them rest after reaching 4V (at 2 A) they turned down to 3.76 V in 24 h! I guess this can end up getting dangerous, so i try to be prepared for the worst case.
I was astonished that the cell degradation can be that quickly. Does this have to do anything with the special cell type Samsung ICR18650 26H ('H' for heater? )

 

[Oberfail]

I have an old ebike battery pack, its an ebike-Vision-Panasonic 24V / 21 Ah (rated as 524 Wh) pack and built of 7S8P Samsung ICR18650 26H.
Its quite old and has lost most of its capacity but i occasionally use it for testing purposes and experiments.

The most negative cell block is defective, has high resistance, low capacity and has some heater cells in it.
When i charge the pack with 4 A (500 mA/cell average) a few cells of this block get hot when reaching 4 V. In the beginning (say 10 cycles ago) they didn't get really hot, lets say little more than hand-warm. Now the cells get hotter and hotter at each cycle, so that i reduced the current to 2 A when coming close to 4V. But even than the defective cells heat up to approx. 40-50 C. I let them cool and can slowly charge the rest up to nearly 4.2V Its interesting that, if they had cooled down, could be charged again up to nearly 4.2 V without a big problem.

The last time i let them rest after reaching 4V (at 2 A) they turned down to 3.76 V in 24 h! I guess this can end up getting dangerous, so i try to be prepared for the worst case.
I was astonished that the cell degradation can be that quickly. Does this have to do anything with the special cell type Samsung ICR18650 26H ('H' for heater? )

I've done some tests with heaters, they loose about 5% capacity per cycle when you "unheat" them. Just toss heaters.

 

[paddy72]

I've done some tests with heaters, they loose about 5% capacity per cycle when you "unheat" them. Just toss heaters.

Thanks for the interesting info, didn't know that yet.
But i recognized the whole pack degrading quickly after just a few cycles.

The interesting part is, that only the most negative block of the 8p7s pack is getting warm/hot, the others not. But when discharging all blocks voltages are more less near, the bad block beeing the most discharged but only like 3.5V compared to 3.6 for the other blocks. Sure the weakest block gets the most stress when charging and discharging as the SoC-hub is the largest - so situation gets worse and worse.

Some say the abnormal heating is coming from small dendrites of Li (or Cu) growing inside the cell and occasionally penetrating the separator causing very small shorts. The very small dentrites than glow away at the tip where they touch the electrode and that heats the cell. If the dendrites are getting bigger they can cause a severe short leading to thermal runaway in the worst case.
I'm not sure if this is a proven fact acutally or just guessing / wild theory?

Anyone with more infos on that?

 

[OffGridInTheCity]

One cause of heaters is high IR coupled with self-discharging. High IR - e.g. resistance = heat. This alone can cause heat. In addition, a cell that never reaches full charge because it's self-discharging as fast as the current being put in can continuously create heat until it get's very hot.

I'm not sure about dendrites vs heat - maybe someone will comment.

 

[paddy72]

One cause of heaters is high IR coupled with self-discharging. High IR - e.g. resistance = heat. This alone can cause heat. In addition, a cell that never reaches full charge because it's self-discharging as fast as the current being put in can continuously create heat until it get's very hot.

I'm not sure about dendrites vs heat - maybe someone will comment.

Yes, agree, but why does this heating occur only near 4.0V and stops once the cell has overcome this limit? My experience is, when you let the cell calm and cool down at 4.0+ V level you can most likely fully charge it up and it tends to hold the voltage much better in this state than when struggling to reach the 4.0 V in the first run.

And 2. when the energy consumption due to "leakage" inside the cell is the same as the charge you put in, why does that heating stop immediatly when you stop charging? The cell would tend to consume the same amount of energy until the voltage goes down significantly. There must be something else going on in the cell.

 

[paddy72]

Another recognition with heaters:
Once they have overcome the 'charging resistance at around 4.0V' they tend to be more stable in voltage and keep their voltage better.
When i had brought the cells to 4+ V and let them sit overnight the voltage drop was much lower, compared with the night before when they were struggling reaching the 4V.
So there is really something very special in that region, say 3,9x ... 4.05V but only in the direction from low to high. On the way back when discharging nothing special happens in that area.

Just trying to understand the processes that led to heating - of course heaters are crap and should be tossed.