Plating metallic lithium during "trickle" charge? A myth?

[Overclocker]

you always hear this quoted from dodgy websites whenever the topic goes to charging using a CC/CV power supply. but is it just a myth?

suppose you set your CC/CV meanwell to 4.0v per cell and leave it for a day? the current would taper off in a few hours to a few milliamps, and would continue like that until the next day. would it plate metallic lithium even at just at 4.0v ?


edit:

just to be clear the "trickle" charging (in quotes!!!) referred to here is if you were to continue the CV phase beyond the usual c/10 cutoff:

 

[Alan B]

A pack can sit with no power supply at above 4.0V by itself for months. Plating would take place without need for a power supply, if it was able to take place at this voltage.

Lead acid trickle chargers are not fixed voltage chargers, they are constant current, and the voltage rises to maintain the current, so they are CC charging at low current. The problem with trickle charging with lithium is the voltage gets too high eventually and then various battery mechanisms like plating take place. With lead systems there is a built in voltage limiting by electrolysis, and plating lead IS the charging mechanism, so it works fine for them.

 

[redilast]

Is battery university a dodgy website? http://batteryuniversity.com/learn/archive/is_lithium_ion_the_ideal_battery

Some other sources:

http://www.buchmann.ca/buchmann/chapter-4-proper-charge-methods

http://www.powerstream.com/li.htm

https://www.pacificpowerbatteries.com/cadex/charginglithiumion.html

http://cds.linear.com/docs/en/lt-journal/LTMag-V15N04-10-LTC4063-Hoffart.pdf (pg44)

 

[liveforphysics]

A CV profile will naturally have current approach 0A. At 0A of current, the cell couldn't know if it were connected or not connected to a power supply.

If the power supply can "trickle" it's not a CV power supply, or your cells have some critical self-discharge issues and will soon rupture from excessive gas production.

 

[Alan B]

Thanks for your comments Luke. Exactly.

Trickle charging is a low current constant current charging regimen. The voltage limit is not regulated to a safe-for-lithium value. The goal is to match the current to the self discharge current of the battery and find an equilibrium that keeps the battery charged all the time, or to slowly charge the battery if it was not fully charged, then maintain it. It is used with chemistries that have processes that can convert slight overcharge current into something safe for the cell, typically into heat. Common in lead acid batteries is electrolysis splitting water, and elsewhere in the cell it is catalyzed back into water, or it is slowly lost through the vents and the cell needs periodic rewatering to recover the proper electrolyte level.

Float charging is a constant voltage charging regimen with a low current limit.

Trickle charging (as traditionaly defined) should not be used with lithium batteries as the voltage will eventually rise above the safe limit values for lithium (typically 4.2V).

Float charging with a charging value less than or equal to the safe limit value could be used with lithium. On-off cycling is a variant of float charging and many chargers use that for charge maintenance.

The confusion seems to be when a CC CV charger or power supply is used with lithium, the current tapers off to a low value. This is not trickle charging. It is essentially float charging. This current approaches zero, trickle chargers do not - they maintain a small current by raising the voltage. This increase in voltage is dangerous for lithium.

The sources quoted above warning against trickle charging are correct, in that a trickle charger as they mean it is not voltage limited and will continue to increase the voltage to achieve its design current. They are CC chargers without a CV limit that is appropriate for lithium batteries.

Don't be confused by terminology that is not always well defined, or infer meaning that was not intended.

When I use a CC CV power supply to charge lithium, if I accidentally leave it on a few hours extra the current gets very low. Essentially there is no energy flowing into the fully charged battery once the current drops to a low value.

 

[Hillhater]

Excellent summary Alan.
I would add that anyone charging a lithium pack , using anything other than the correct approved charger for that pack, ought to be constantly monitoring charge progress and ensuring safe completion of charge. and frequently checking cell balance .
Using a CC/CV power supply is common but can be made a little safer by using a timer on the mains supply to shut the power off after a set charging time ,
Most lithium pack "thermal events" seem to happen during , or just after, charging which is when the pack contains the most energy.

 

[Alan B]

I agree that monitoring balance is a good plan, I do that frequently until the pack has developed a history of good balance bulk charging performance, then I can check it less frequently. But the first few cycles it is very frequently checked.

 

[redilast]

The data sheets from every 18650 manufacturer will always list a recommended termination current (during the CV phase) in the ~25-100mA range for 18650 cells. I don't believe this is healthy for the cell to have it sitting on a charger/power supply charging at just a few mA or micro amps even. Why else would all of these sources say don't trickle charge Li-ion's because it leads to metallic lithium plating?

 

[Overclocker]

The data sheets from every 18650 manufacturer will always list a recommended termination current (during the CV phase) in the ~25-100mA range for 18650 cells. I don't believe this is healthy for the cell to have it sitting on a charger/power supply charging at just a few mA or micro amps even. Why else would all of these sources say don't trickle charge Li-ion's because it leads to metallic lithium plating?

exactly. that's what i was gonna say next

why implement a cut-off when it's cheaper to just let the current taper off to zero (or very close to zero) ?

or could it be that holding the cell to 4.20v for extended periods is what causes the plating? and not the micro-current? because with the recommended cut-off point the cell settles to around 4.17v

or maybe the cutoff has to do with component tolerances? some chargers have the CV=4.25v no cutoff would leave the cell soaking at 4.25v until god knows when...

 

[amberwolf]

Once the cell has charged to a specific voltage (4.2v, 4.25v, whatever), it's irrelevant whether the PSU it's charged with (that is at the same voltage) is still connected, powered on, etc.

There isnt' any current flowing, so nothing is changing in the cell (unless it is already at a voltage where some chemical action is now occuring; if that's the case...don't charge it that high, because it's bad for it anyway ).

 

[Punx0r]

In what context is the charge termination recommendation made? It may simply be "at this point you can consider the cell fully charged" i.e. will delivery rated capacity and not related to any detrimental effect of being left charging longer.

 

[flangefrog]

Once the cell has charged to a specific voltage (4.2v, 4.25v, whatever), it's irrelevant whether the PSU it's charged with (that is at the same voltage) is still connected, powered on, etc.

There isnt' any current flowing, so nothing is changing in the cell (unless it is already at a voltage where some chemical action is now occuring; if that's the case...don't charge it that high, because it's bad for it anyway ).

There may be very little current flowing but after being disconnected the cell naturally settles to around 4.15V (Due to inner potentials equalizing? As the settle voltage will be lower when charging faster). Floating the cell would keep adding a little bit of current until it is fully equalized. It could also keep flowing a very small amount of current which is released as heat by the battery.

In what context is the charge termination recommendation made? It may simply be "at this point you can consider the cell fully charged" i.e. will delivery rated capacity and not related to any detrimental effect of being left charging longer.

My theory is that the cells proper charged voltage is more like 4.15V but as it takes a long time to equalize it's easier to charge it a bit higher, cut off the charge before it's finished and then let the cell settle.

No idea if this is correct or not, maybe LFP can comment as I know he's talked about cell equalisation before.

I charge my battery with a CV only charger as that's all I have at the moment. Sometimes the battery stays floating overnight but I haven't noticed any bad effects yet.

 

[Overclocker]

Once the cell has charged to a specific voltage (4.2v, 4.25v, whatever), it's irrelevant whether the PSU it's charged with (that is at the same voltage) is still connected, powered on, etc.

There isnt' any current flowing, so nothing is changing in the cell (unless it is already at a voltage where some chemical action is now occuring; if that's the case...don't charge it that high, because it's bad for it anyway ).

well obviously when there's zero current i.e. 0.00 femto-amperes there's nothing happening good or bad 8)

but it's not gonna happen because the li-ion has self-discharge and probably some small temperature-based voltage fluctuations. so the current won't actually reach zero.

 

[Overclocker]

In what context is the charge termination recommendation made? It may simply be "at this point you can consider the cell fully charged" i.e. will delivery rated capacity and not related to any detrimental effect of being left charging longer.

well since the charging current is tapering then you've got to draw the line where to stop 8) for practical purposes and standardization. if you set it to c/50 then the CV phase would take a very long time but the cell would be closer to 4.20v. perhaps c/10 is a good enough compromise

the funny thing w/ li-ion is that 4.20v is just an arbitrary number. a good enough compromise between capacity and cycle life. a convenient round number for industry standardization. AFAIK it's not directly tied to any physical constants. correct me if i'm wrong

 

[Alan B]

One of the characteristics of 4.20V is that it is below the voltage of plating lithium. Plating is a chemical process that requires a certain voltage due to the potential characteristics of the ions. If this voltage is not present plating simply cannot occur. At work we built a sensitive scientific instrumentation system for analyzing metal ion content of water by very precisely controlling small bursts of current and measuring the voltage you could tell exactly how much of each ion was present in the water sample, as each ion plated out at a very specific voltage, and when that ion was depleted the voltage would rise to the next level for the next higher voltage ion that was present.

Sitting on a CV supply at 4.20V is NOT trickle charging, even if the current is low. A trickle charger raises the voltage to maintain its design current, it is a constant current source. So it is incorrect to call this trickle charging, or to associate it with the plating of lithium. It just doesn't happen at 4.20V, but the actual value depends on the particular electrode materials involved. There are many lithium chemistries, and the variations in materials change these values.

See more about standard electrode potentials:
https://en.wikipedia.org/wiki/Standard_electrode_potential

The BMS actually needs a period of constant voltage to fully balance the pack, alternately bleeding the high cells and then reconnecting the charger. Not having a constant voltage source means the balancing takes a lot longer, and may never complete.

The common practice of undercharging your pack will prevent the BMS from balancing and often leads to shortening the pack's life instead of lengthening it by unbalancing it. The BMS often contributes to this by drawing more current from the first two cells to power it, and drawing less from the rest for monitoring only. Packs having this type of BMS MUST be fully charged often enough to allow balancing, otherwise they will fail early.

 

[Alan B]

The effect of holding the cell voltage at a high voltage such as 4.2V (whether on a CV supply or just sitting at full charge) is that various processes that degrade the cell (primarily they seem to talk about the electrolyte) are voltage (and temperature) dependent, thus the recommendations for lower storage voltages and cooler temperatures. Staying there for a few days is not a big factor, but staying there all the time will hasten the degradation of the cell. The cell life clock runs faster at higher voltages and temperatures.

There are plenty of scientific papers that describe these types of things, or talk to a battery chemist as I did. There are conferences and journals full of these papers.

Here's the guy I talked to a number of times:

http://eetd.lbl.gov/people/venkat-srinivasan

He used to have a blog that was fun to follow, I haven't looked lately to see what he's up to. He is good at explaining things to nontechnical people, and he was very impressed with my Borg ebike and wanted to know what I was doing with the batteries. Unfortunately to really learn what he wants to know you need to run the batteries down often and remeasure them which I don't do, otherwise you don't know how the capacity is holding up.

I think part of the reason he was impressed with the Borg is that during the evening commute it easily out accelerated his small car at low speed.

 

[Overclocker]

ok so if you don't exceed 4.20v you should be fine? so this whole c/10 cutoff thing is completely optional? nevermind that ALL charge control IC's do it

it's funny that we think we know li-ion charging. but we don't even know the rationale for simple things like c/10 cutoff...

 

[amberwolf]

He used to have a blog that was fun to follow, I haven't looked lately to see what he's up to.

http://thisweekinbatteries.blogspot.com/

 

[redilast]

ok so if you don't exceed 4.20v you should be fine? so this whole c/10 cutoff thing is completely optional? nevermind that ALL charge control IC's do it

it's funny that we think we know li-ion charging. but we don't even know the rationale for simple things like c/10 cutoff...

Not only do all proper Li-ion charge controller IC's do it, but also they have very conservative settings for restarting the charge cycle. Most IC's for example will have like say a C/10 or C/20 termination current at 4.20v, but then also be programmed to not start charging again unless the cell dips below 4.1 or 4.05v.

I think one of the reasons to also terminate charge current is so you know if you have a bad/old/junk cell. For example if I were to charge at C/10 a new cell might settle down to 4.17v or something after a few hours. Then If I have an old cell that settles down to 4.06v after a few hours, I know something is up. If the power supply or charging setup constantly is trickle charging to counter the cells self discharge then you might see the pack voltage and think everything is fine because you didn't let the cell ever settle down after charging to realize that you have a junk cell.

 

[liveforphysics]

The data sheets from every 18650 manufacturer will always list a recommended termination current (during the CV phase) in the ~25-100mA range for 18650 cells. I don't believe this is healthy for the cell to have it sitting on a charger/power supply charging at just a few mA or micro amps even. Why else would all of these sources say don't trickle charge Li-ion's because it leads to metallic lithium plating?

If the charger just holds a constant voltage, it will stop adding current to the cell when the cells SOC reaches the same voltage. Current may only flow as a result of a difference in potential.

 

[Overclocker]

The data sheets from every 18650 manufacturer will always list a recommended termination current (during the CV phase) in the ~25-100mA range for 18650 cells. I don't believe this is healthy for the cell to have it sitting on a charger/power supply charging at just a few mA or micro amps even. Why else would all of these sources say don't trickle charge Li-ion's because it leads to metallic lithium plating?

If the charger just holds a constant voltage, it will stop adding current to the cell when the cells SOC reaches the same voltage. Current may only flow as a result of a difference in potential.

the problem is that it takes a looooooooooooooooooooong time before that happens, if at all (due to self-discharge the current won't taper off to absolute zero)

so while tapering off to near-zero the cell is exposed to the full 4.20v.

all charge controller IC's terminate the charge when the current drops below c/10 to c/20. so there must be some wisdom in doing that. the exact reasons aren't apparent to us right now although some theories have been offered above

 

[redilast]

Even Dave Jones at EEV blog says don't trick charge or leave it on a bench power supply: https://youtu.be/0E4_iWHI-U0?t=759

He just published this video a few minutes ago haha. :wink:

---------------

And I agree, a charger or power supply set at a CC/CV profile will never actually reach zero due to the cells self discharge rate. It would be in the micro amp range. I still don't think its healthy to leave a cell at this voltage for an extended period of time, even if the majority of the metallic lithium plating actually happens at 4.30v+.

 

[Hillhater]

The use of a "trickle charge" or a bench power supply, was ruled a "No No" early on in the thread.
And , again as has been said before, what is the difference between an isolated cell at 4.2 v , and the same cell with a stable 4.2v charger attached ?.....None !
IF plating, or any effects are going to occurr, the cell have enough potential itself to initiate the process. It doesnt need any external "micro amp even" current to make it happen.
Also can you define "extended period of time" ? .....we all know storing cells at peak charge is not good for them.

 

[Overclocker]

Even Dave Jones at EEV blog says don't trick charge or leave it on a bench power supply: https://youtu.be/0E4_iWHI-U0?t=759

He just published this video a few minutes ago haha. :wink:

---------------

And I agree, a charger or power supply set at a CC/CV profile will never actually reach zero due to the cells self discharge rate. It would be in the micro amp range. I still don't think its healthy to leave a cell at this voltage for an extended period of time, even if the majority of the metallic lithium plating actually happens at 4.30v+.

well the guy's not really a li-ion expert. i mean he doesn't even have an 18650 charger for crying out loud LOL

he's probably just regurgitating what info he found on the interwebs...

 

[Punx0r]

If being sat at 4.20V instead of 4.17V was detrimental to li-ion cell health we wouldn't have all these cells rated for 4.35V...