Author Topic: Lithium-Ion battery degeneration  (Read 1357 times)

0 Members and 1 Guest are viewing this topic.

Offline gueston

  • Contributor
  • Posts: 5
  • Country: nl
Lithium-Ion battery degeneration
« on: August 29, 2020, 05:46:13 pm »
I know that Li-Ion batteries degenerate over time, even when not used. But is there a general rate for it at room temperature (20-25 degrees Celsius)?
And does it mean that after a while they can't be used at all anymore, or they just wont hold as much charge?
 

Offline Nauris

  • Regular Contributor
  • *
  • Posts: 183
  • Country: fi
Re: Lithium-Ion battery degeneration
« Reply #1 on: August 29, 2020, 06:11:39 pm »
Gradually lose capacity and become higher resistance. I don't think there is any general rule, depends on chemistry and manufacturer and is not that well known. Samsung rates their 94 Ah cells for 70 % capacity after 26 years stored fully charged at 25 C for example.
 

Offline gueston

  • Contributor
  • Posts: 5
  • Country: nl
Re: Lithium-Ion battery degeneration
« Reply #2 on: August 30, 2020, 10:43:09 am »
Ah ok. And is it a problem to leave a battery fully charged? For instance if an adapter is connected to a charger circuit constantly, then the battery would be fully charged all the time.
 

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 4341
  • Country: fi
Re: Lithium-Ion battery degeneration
« Reply #3 on: August 30, 2020, 12:58:16 pm »
Both capacity loss and ESR rise depend heavily on state-of-charge, and temperature. When stored below about 50% SoC at room temperature or lower, aging is very slow (say, below 1%/year of capacity loss, or <5%/year ESR rise). On the other hand, stored between 80-100% SoC at elevated temperatures (say 40-50 degC), aging can be quite fast, say, 20% capacity loss per year or 50% ESR rise per year.

Exact figures depend on the actual product chemistry. For example, Sony cells seem to perform quite well regarding shelf life.

End-Of-Life is often defined as capacity loss of 30% or ESR rise of +100%, but this is obviously completely arbitrary and nothing prevents you from using End-Of-Life cells further.

Storing fully charged isn't a huge problem if you can avoid high temperatures; surprisingly storing at 80% or 90% may not offer any lifetime benefit. Going below 70% starts increasing the lifetime significantly, though, and is recommendable whenever you can do it.
« Last Edit: August 30, 2020, 01:02:01 pm by Siwastaja »
 

Offline gueston

  • Contributor
  • Posts: 5
  • Country: nl
Re: Lithium-Ion battery degeneration
« Reply #4 on: August 30, 2020, 04:55:33 pm »
Thanks for your answer.

The reason I ask this is, because I am developing a product with a li-ion battery. The user can charge the battery and operate the device on the battery. But some users may keep the charger plugged in all the time. In that case the battery will never be discharged and keep charged maxium. But as you say, this wouldn't cause any real problems right? Also not safety wise?
 

Offline NiHaoMike

  • Super Contributor
  • ***
  • Posts: 7482
  • Country: us
  • "Don't turn it on - Take it apart!"
    • Facebook Page
Re: Lithium-Ion battery degeneration
« Reply #5 on: August 30, 2020, 09:24:07 pm »
Keep the charging voltage to 4.1V/cell or less, perhaps as low as 3.9V/cell if it's an expensive pack you want to maximize the lifetime of. Ideally, you should provide a setting that the user can adjust based on their intended use.
On the other hand, stored between 80-100% SoC at elevated temperatures (say 40-50 degC), aging can be quite fast, say, 20% capacity loss per year or 50% ESR rise per year.
I used to have a pocket Wifi AP with built in battery, which only lasted a few months of use constantly plugged in before bulging. It turns out the battery was basically in direct contact with the hot CPU and the charger being a bit high at 4.23V didn't help matters at all. A similar problem was encountered by those using really cheap smartphones for Perk coin mining, which led to hacks for replacing the batteries with a power supply.
Cryptocurrency has taught me to love math and at the same time be baffled by it.

Cryptocurrency lesson 0: Altcoins and Bitcoin are not the same thing.
 

Online BravoV

  • Super Contributor
  • ***
  • Posts: 7331
  • Country: 00
  • +++ ATH1
Re: Lithium-Ion battery degeneration
« Reply #6 on: August 31, 2020, 04:09:01 am »
The reason I ask this is, because I am developing a product with a li-ion battery. The user can charge the battery and operate the device on the battery. But some users may keep the charger plugged in all the time. In that case the battery will never be discharged and keep charged maxium. But as you say, this wouldn't cause any real problems right? Also not safety wise?

The sweet spot is 4.03 volt, the longer the cell is soaked in higher voltage, say like 4.15 volt, the faster the decay.

At least according to Jeff Dahn at below video is at 4.03 volt ... approx. about 80%, and gets worst above 4.1 volt. He is well known expert in battery, currently working with Tesla in battery development.

I marked the time at 1h07m, in this Youtube video where its discussed this particular issue ...

-> https://youtu.be/pxP0Cu00sZs?t=1h7m5s

Even though its a 1 hour 14 mins video length, recommending to watch this Prof. Jeff Dahn video from the beginning, its very informative, imo.

PS : Recently I acquired a new laptop Asus, it has setting for charging levels, for user who keep the charger plugged all the time, it recommends at the lower setting with charge level of 60% of the capacity.
« Last Edit: August 31, 2020, 04:15:23 am by BravoV »
 
The following users thanked this post: aargee

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 4341
  • Country: fi
Re: Lithium-Ion battery degeneration
« Reply #7 on: August 31, 2020, 05:49:56 am »
Re keeping charger connected all the time.

Some solutions (most charging ICs) use hysteresis to enable and disable charging. This cycles the cell between, say, 90% and 100% and isn't very good for battery life, especially if the charging current is fairly high. These solutions are designed stand-alone chargers in mind, with no load connected, so that "cycling" really doensn't occur. If you have a load connected, I would avoid this class of charging solutions.

Some solutions just float the cell. If the supply is clean, then no current flows to the cell. But floating it at 100% is still equally bad as storing it at 100%.

Really note that many cell chemistries do not benefit at all from storing at 90% or 80% compared to 100%. The same applies to floating them with a charger. You need to go lower than 80%.

Hence, my proposed solution for a long-life standby solution is to float the cell at 60-70% and compensate by picking a 30-40% larger cell than you would otherwise need.

Also limit charging current at low temperatures (for example, linearly from 0.5C at +25degC to zero at -5degC), and remember to prevent charging and report broken cell if the cell is severely overdischarged (say below 2V).
 

Offline gueston

  • Contributor
  • Posts: 5
  • Country: nl
Re: Lithium-Ion battery degeneration
« Reply #8 on: September 01, 2020, 07:20:29 pm »
The reason I ask this is, because I am developing a product with a li-ion battery. The user can charge the battery and operate the device on the battery. But some users may keep the charger plugged in all the time. In that case the battery will never be discharged and keep charged maxium. But as you say, this wouldn't cause any real problems right? Also not safety wise?

The sweet spot is 4.03 volt, the longer the cell is soaked in higher voltage, say like 4.15 volt, the faster the decay.

At least according to Jeff Dahn at below video is at 4.03 volt ... approx. about 80%, and gets worst above 4.1 volt. He is well known expert in battery, currently working with Tesla in battery development.

I marked the time at 1h07m, in this Youtube video where its discussed this particular issue ...

-> https://youtu.be/pxP0Cu00sZs?t=1h7m5s

Even though its a 1 hour 14 mins video length, recommending to watch this Prof. Jeff Dahn video from the beginning, its very informative, imo.

PS : Recently I acquired a new laptop Asus, it has setting for charging levels, for user who keep the charger plugged all the time, it recommends at the lower setting with charge level of 60% of the capacity.
Thank you. Very informative video. I always wondered how they would test the lifetime of a battery if not only cycle time, but also aging is a factor. Yes you could test for years, but to speed up the process. He explains exactly how it's done!

Re keeping charger connected all the time.

Some solutions (most charging ICs) use hysteresis to enable and disable charging. This cycles the cell between, say, 90% and 100% and isn't very good for battery life, especially if the charging current is fairly high. These solutions are designed stand-alone chargers in mind, with no load connected, so that "cycling" really doensn't occur. If you have a load connected, I would avoid this class of charging solutions.

Some solutions just float the cell. If the supply is clean, then no current flows to the cell. But floating it at 100% is still equally bad as storing it at 100%.

Really note that many cell chemistries do not benefit at all from storing at 90% or 80% compared to 100%. The same applies to floating them with a charger. You need to go lower than 80%.

Hence, my proposed solution for a long-life standby solution is to float the cell at 60-70% and compensate by picking a 30-40% larger cell than you would otherwise need.

Also limit charging current at low temperatures (for example, linearly from 0.5C at +25degC to zero at -5degC), and remember to prevent charging and report broken cell if the cell is severely overdischarged (say below 2V).
Thanks again for your answer. I am planning to use this IC BQC25886, which has USB charge capabilities for 2 cel Li-ion.
As far as I understand this IC doesn't use the hysteresis, but keeps the battery charged at all times at the rated voltage (8.2, 8.4 depending on the setting).

But I can also see that the charging is regulated with the /CE pin, which I can control from the software. In that case I can control and keep the charge between 40-60% if I detect a constant connected power source. Am I right with this approach?

« Last Edit: September 01, 2020, 08:45:22 pm by gueston »
 

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 4341
  • Country: fi
Re: Lithium-Ion battery degeneration
« Reply #9 on: September 02, 2020, 02:07:39 pm »
But I can also see that the charging is regulated with the /CE pin, which I can control from the software. In that case I can control and keep the charge between 40-60% if I detect a constant connected power source. Am I right with this approach?

That would create a lot of alternating charge-discharge cycles between 40%-60%. Likely the cell can withstand tens or even hundreds of thousands of such cycles just fine, but it's still something I would avoid, just to be sure, if at all possible, by floating the cell with constant, regulated voltage instead.
 
The following users thanked this post: gueston

Offline gueston

  • Contributor
  • Posts: 5
  • Country: nl
Re: Lithium-Ion battery degeneration
« Reply #10 on: September 02, 2020, 05:36:41 pm »
But I can also see that the charging is regulated with the /CE pin, which I can control from the software. In that case I can control and keep the charge between 40-60% if I detect a constant connected power source. Am I right with this approach?

That would create a lot of alternating charge-discharge cycles between 40%-60%. Likely the cell can withstand tens or even hundreds of thousands of such cycles just fine, but it's still something I would avoid, just to be sure, if at all possible, by floating the cell with constant, regulated voltage instead.

Ok, well I'm designing this for a consumer grade product. So if this cycling can go on for tens of thousands of cycles it's more than enough actually.

But your other solution, this IC can't provide this, like most charging IC's can't as you said. But could it be a good design solution that in such a case I would switch off the whole charge IC and switch on a voltage regulator of about 7.4V (I use 2 cells in serie) to keep the batteries at that voltage? Then it's not cycling as you said and keeps it at constant 7.4V (which is 3.7V per cell) voltage. Or would this bring other problems?
« Last Edit: September 02, 2020, 05:42:45 pm by gueston »
 

Offline paulca

  • Super Contributor
  • ***
  • Posts: 2587
  • Country: gb
Re: Lithium-Ion battery degeneration
« Reply #11 on: September 03, 2020, 04:59:39 pm »
When stored below about 50% SoC at room temperature or lower, aging is very slow (say, below 1%/year of capacity loss, or <5%/year ESR rise). On the other hand, stored between 80-100% SoC at elevated temperatures (say 40-50 degC), aging can be quite fast, say, 20% capacity loss per year or 50% ESR rise per year.

I used to take this as gospel as I read it hundreds of times and seen/read the links to Battery University where they had a little table. 

I wasn't until recently that I heard it challenged.  It's was AvE on YouTube who pointed out, there was no study, it was just a single table presented on Battery University and a few echo chamber sites picked it up, plus it being repeated constantly on forums.  Never any link or reference/source of this study or who did it.  No real data either.

So I searched and haven't found anything but Battery University and a few echo chambers.

The only real issue I can see with storing them fully charged is temperature.  In RC modelling were warned to not charge batteries in the field if it was cold unless we were certain they would be used immediately.  The reason is, the voltage will rise as temperature does.  So if you have 4.20V per cell at 5*C and take that battery home, setting your field bag down in a lovely 20*C room ... the voltage could rise 100mV or more and over volt the cell damaging it.

Do you have any references/sources with actual data testing life span against storage voltage and temperature.

On a separate, related point, my only evidence is that I have had two laptops which virtually never ran on battery, spent 99% of their lives plugged in and ... both batteries were found dead, unchargable, unusable when tested.  I put these down to the battery being cooked by the laptop which were run far too hot 24/7.
"What could possibly go wrong?"
Current Open Projects:  3 Channel Audio mixer with DAC, BT, pre-amps and h/phone amp, WS281x LED controller Version 2 5V/5A w/Atmega328, FY6600 Power supply, 5A DC Load (Still!)
 

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 4341
  • Country: fi
Re: Lithium-Ion battery degeneration
« Reply #12 on: September 04, 2020, 12:17:21 pm »
When stored below about 50% SoC at room temperature or lower, aging is very slow (say, below 1%/year of capacity loss, or <5%/year ESR rise). On the other hand, stored between 80-100% SoC at elevated temperatures (say 40-50 degC), aging can be quite fast, say, 20% capacity loss per year or 50% ESR rise per year.

I used to take this as gospel as I read it hundreds of times and seen/read the links to Battery University where they had a little table. 

I wasn't until recently that I heard it challenged.  It's was AvE on YouTube who pointed out, there was no study, it was just a single table presented on Battery University and a few echo chamber sites picked it up, plus it being repeated constantly on forums.  Never any link or reference/source of this study or who did it.  No real data either.

So I searched and haven't found anything but Battery University and a few echo chambers.

The only real issue I can see with storing them fully charged is temperature.  In RC modelling were warned to not charge batteries in the field if it was cold unless we were certain they would be used immediately.  The reason is, the voltage will rise as temperature does.  So if you have 4.20V per cell at 5*C and take that battery home, setting your field bag down in a lovely 20*C room ... the voltage could rise 100mV or more and over volt the cell damaging it.

Do you have any references/sources with actual data testing life span against storage voltage and temperature.

On a separate, related point, my only evidence is that I have had two laptops which virtually never ran on battery, spent 99% of their lives plugged in and ... both batteries were found dead, unchargable, unusable when tested.  I put these down to the battery being cooked by the laptop which were run far too hot 24/7.

Not interested in Battery University and other fake science sites. I have done my own extensive research, part of which was a 1.5-year measurement program of self-discharge and calendar fading, measured parameters were capacity and DC ESR, input parameters were different cell brands and products, state-of-charge during storage and temperature during storage. It was supposed to be put on a peer reviewed process and published but sadly never did because I left the university job far before the test time was finished. I have posted some data on this forum before, for example here:
https://www.eevblog.com/forum/projects/low-curent-pulse-charging-a-sla-battery/msg1868660/#msg1868660

As it stands now, it likely doesn't exceed the most stringent scientific requirements because the university support was cut off mid-way and the test finished for personal interest only, but what the heck, why not publish all the results, so here it goes, see the attachment. Feel free to share to others. Unfortunately, I lost interest in trying to maintain the heated cells heated once I had to repurpose the temperature controlled heatbed, and used whatever means and later estimated the average "warm" temperature. So take the "warm" dataset with a pinch of salt. The room temperature dataset (avg 22.5 degC, between 21 and 25 degC 99% of the time; air conditioning was available), and the "cold" dataset (actually monitored within +3..+5 degC) are useful, though.
 
The following users thanked this post: BravoV, Zucca

Offline paulca

  • Super Contributor
  • ***
  • Posts: 2587
  • Country: gb
Re: Lithium-Ion battery degeneration
« Reply #13 on: September 05, 2020, 08:57:35 am »
First thank you for that data.  It's at least an attempt at something.  There seems to be anomalies, maybe because it's not an easy test to do.  There appears to be false positive and false negative results for both arguments regarding storage voltage.  Although storage temperature seems more consistent to support not storing them hot.... like a mobile phone application spending 60% of it's time in your back pocket.  Even there mobile phone batterys have improved in that regard a lot with my current phone still lasting several days on a charge at 4 years old... running 24/7/365.  Much like my last phone it looks like it will become unusable not due to battery degrade, but due to the micro-usb socket failing.

So only looking at the top Samsungs.  The only one that showed any capacity fade after 1.5 years stored at 4.20V was the one stored at 35*C.  Was this battery charged at 35*C or charged at room temp and then stored at 35*C?  If it was charged at 15*C and then stored at 35*C then it's voltage would have risen above 4.20V and stayed there until it self discharged rapidly initially, but could have done the damage shown, no?

The first one stored at normal conditions at 4.20V seemed to gain capacity, though it's ESR rose(?).

I'm still not 100% convinced enough to discharge Lithium Ions I store for months as I may need them to be fully charged and it's honestly a pain to charge a handful to 4.2Vpc and then when I don't use them immediately discharge them and add a partial cycle to them.

One supporting bit of evidence for not storing them long term is that most products seem to come with some level of charge, but always tell you to fully charge before use.  However if you consider the environmental aspects of mass produced products being shipping around the world from China in container ships, the temperatures inside a container in the sun could easily get over 40*C, which even your data shows would harm a fully charged battery.

I feel if it was significantly important to not store them fully charged at the consumer end this storage charge advice would be in every copy and paste product manual.  Electric car owners would be advised not to keep their cars fully charged unless they are in use in the next few days. etc. etc.

I suppose one element of safety would be that a fully charged cell at 100% would carry a higher risk should it, for whatever reason, go into melt down than a 40% charged cell.
« Last Edit: September 05, 2020, 09:04:55 am by paulca »
"What could possibly go wrong?"
Current Open Projects:  3 Channel Audio mixer with DAC, BT, pre-amps and h/phone amp, WS281x LED controller Version 2 5V/5A w/Atmega328, FY6600 Power supply, 5A DC Load (Still!)
 

Offline MosherIV

  • Super Contributor
  • ***
  • Posts: 1528
  • Country: gb
Re: Lithium-Ion battery degeneration
« Reply #14 on: September 05, 2020, 09:15:20 am »
Quote
  I feel if it was significantly important to not store them fully charged at the consumer end this storage charge advice would be in every copy and paste product manual.

"Consumer" means sell as many as you can. So it is NOT in consumer electronics companies interrests to tell the customer how to prolong the life of the product.

It is well known by most who know something about LiIon batteries that it is not good to store then at full charge.
The reason is chemistry, at full charge where all the available lithium ions is in the kathode, there is a tendancy for the ions to stick to the kathode, ie lithium plating occurs. The process is irreversable.
 

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 4341
  • Country: fi
Re: Lithium-Ion battery degeneration
« Reply #15 on: September 05, 2020, 12:00:50 pm »
So only looking at the top Samsungs.  The only one that showed any capacity fade after 1.5 years stored at 4.20V was the one stored at 35*C.

Yes, the 26H LCO cell. Do note however, that the newer-generation 29E NCA chemistry cell, which better represents later cells available, shows some clear capacity fade at 4.20V AND even at 4.00V at room temp (note how small the difference between 4.20V and 4.00V is - 3.0% vs. 2.3% capacity fade per year - hence my earlier comments, you need to go lower for real gains! And, I have seen a paper where a Panasonic cell stored slightly worse at 80% than at 100%!), and you need to go down to 3.60V to get "completely" rid of the capacity fade. I'm sad I'm lacking data points inbetween, it would be nice to see the result at 3.9V and 3.8V. My guess based on some papers I have seen but not stored for reference is that it starts getting considerably better already at 3.9V.

Quote
Was this battery charged at 35*C or charged at room temp and then stored at 35*C?

All charging and discharging was performed at the same room temp approximately 22 degC. For DC ESR measurement, it's especially important all samples are handled in similar conditions, including using the same holder so that the thermal impedance to the surroundings during the test keeps unchanged. Hot/cold cells were brought back to room temp (for 24hrs, IIRC) for post-storage measurement. So only the storage conditions are part of the equation.

Charging termination was C/40 instead of "standard" C/20 to reduce variability of environmental variables vs. SoC. Measurements were automated using a calibrated Agilent 6632B supply and this (kludgy) control code: https://github.com/siwastaja/battest

Quote
If it was charged at 15*C and then stored at 35*C then it's voltage would have risen above 4.20V and stayed there until it self discharged rapidly initially, but could have done the damage shown, no?

I think you are misunderstanding something, or mixing up with lead acid I guess? There is no such voltage-temperature relationship in li-ion, at all (or it is at completely meaningless level). Open-circuit voltage is a function of SoC, simply. Of course, DC ESR is related to the temperature so at different charging temperatures, you end up at a slightly different SoC for a certain charge termination condition, but there's no way neither SoC or voltage would shoot up when the cell is heated. Voltage always goes down to the open-circuit voltage, by approximately Vdrop = I_charge_termination * R_dc_esr (ignoring some possible second-order effects). This is why termination current is kept fairly low (such as C/20, or C/40 in this test), to standardize the end-result SoC and thus OCV, regardless of variations in R_dc_esr (caused by temperature differences and cell aging).

Quote
The first one stored at normal conditions at 4.20V seemed to gain capacity, though it's ESR rose(?).

Yes, do note however that the "gained" capacity is just 0.2%. This is likely just noise and measurement inaccuracy, not any real capacity gain. This is, as you said, difficult to measure accurately. Expect some +/- 0.5% error margin. So basically the conclusion is, Samsung 26H LCO cell does not lose capacity in storage, but does increase its DC ESR; I think this was an interesting, but not unexpected result. This is caused by the excess thickening of the SEI (solid electrolyte interface), kind of passivation layer during storage at high voltage. This passivation layer actually prevents self-discharge, but also slows down ion transfer, which can be seen as increased ESR.

But the 29E NCA cell, on the other hand, both loses capacity, and increases the DC ESR.

Quote
I'm still not 100% convinced enough to discharge Lithium Ions I store for months as I may need them to be fully charged

I agree, looking at my data, the worst performing (LG HE2) capacity fades 4.9%/year fully stored at room temperature. So it would take 6 years to reach -30% End-Of-Life. This isn't too bad. I only discharge cells to 30-50% if I expect them to sit for years unused and I'm sure I want to use them after that long storage.

edit: removed broken quote
« Last Edit: September 06, 2020, 06:43:20 am by Siwastaja »
 
The following users thanked this post: paulca

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 4341
  • Country: fi
Re: Lithium-Ion battery degeneration
« Reply #16 on: September 05, 2020, 12:06:21 pm »
The reason is chemistry, at full charge where all the available lithium ions is in the kathode, there is a tendancy for the ions to stick to the kathode, ie lithium plating occurs. The process is irreversable.

I think lithium plating mostly happens when current is flowing (intercalation "goes wrong"); especially when charging near 100% SoC, at too high currents, or at too low temperatures, but naturally to some extent even when staying within specs.

I'm not saying it doesn't happen at all during storage, though, so you are likely right. Another mechanism is the thickening of the SEI layer caused by unwanted electrolyte reactions. Actually, originally it's a desired process; the cells are "formed" at the factory. But if the layer gets too thick, it slows down the ions, which manifests itself as increased ESR. Even the SEI layer thickening can cause capacity loss (not only ESR rise), by completely "blocking" some anode surface.

You can think it as a similar process to electrolysis. The more you apply voltage, the worse it gets. Below some voltage threshold, no reaction happens at all.

AFAIK, electrolyte reactions have been especially bad with the LMO chemistry (such as used in Nissan Leaf), manifesting itself in problems seen in hot environment. Basically, the manganese reacts with the electrolyte.

And finally, severe ESR rise practically means capacity loss in all except the lowest-current applications; theoretically you could CC-CV discharge at very low current to extract all capacity, but if your load needs 0.5C, then your load needs 0.5C, and you have to stop earlier and ditch 20% of the capacity if the cell ESR is dropping your 20% SoC OCV below what you can accept as an low-voltage cutoff. A fresh cell is designed so that you get 100% of the stated capacity by stopping at 2.5V while the OCV is around 3.3V for 0% - i.e., 0.8V drop over ESR, but once ESR rises, you either must discharge at lower current, or you won't be able to fully discharge.
« Last Edit: September 05, 2020, 12:19:56 pm by Siwastaja »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf