Battery Degradation Scientifically Explained

[peakshaving]

@EV-Tech Exp for a battery lay-person like myself, does the current research on lithium ion battery degradation show any sort of priority or ranking in the severity of damage caused by the various mechanisms?

Would you be able to say something like "In the degraded batteries we've studied, we think the leading cause of degradation has been...X" or "If you're looking to increase the life of your battery with the fewest possible lifestyle changes, the one thing you can change that has the most impact would be...Y"?

I'm curious as to what the ideal SoC would be as well! Since I can't charge at home I tend just fill up to ~80% at work once I drop below 30%. I'm sure it's all negligible but if it's just a habitual change I need to make for marginal gain might as well start early .

Also curious if you have any thoughts on the battery fires in Asia that caused Tesla to release a patch for S/X. How do thermal runaways work in a battery cell? Is it just like lighting a fire where there's no return after an "ignition" temperature is reached?

Thanks for making these videos. It's really helpful for someone who's interested in this kind of research but doesn't have the background to make sense of the academic papers.

 

[golfpilot]

I leave my Rav4 under a sunshade at the airport. Now that I have the Model 3 it is only on pace to drive about 2500 miles this year. The car does not have an active cooling system when it is turned off so even after weeks of non usage my SOC stays the same. Which is nice for my convenience but battery degradation concerns me. I try to leave it as close to 50% as possible and so far I think my range degradation has been minimal. Once every two month or so I take it on my 200 mile commute just to "stretch it's legs". This trip requires a fast charge at the destination from ~5% to 90% to get home. I immediately charge the car to 50% using my 32amp charger and return the car to the airport carport, according to this that is not good, I'm curious how not good that is? With our model 3's, would setting our car to start charge 2-3 hours after we get home better for the battery then charging right away to get out of the low % charge. I read somewhere that occasional fast charging is actually good for the car, I don't remember where but this video doesn't really support that. For the most part, I have followed these rules but I didn't know why, this was a nice watch.

 

[EV-Tech Exp]

@EV-Tech Exp I have another question about the recommendations in your video. You mention 15% and 90% SoC as boundaries to avoid use beyond. I understand those are approximate values and are a function of ambient temperature as well as battery design/chemistry. But there's a fairly large variation in indicated SoC as a percentage of absolute SoC. The e-tron, EQC and e-Golf only allow 86-88% of the battery to be usable, whereas Teslas are closer to 96% or higher. Were you thinking of indicated or absolute SoC in your 15 and 90% recommendations?

I know the fall off is not linear. It's a bathtub curve where the extremes are much worse for the battery than slightly off the extremes. So a secondary question is how sensitive are batteries to levels just inside of those boundaries? This graph is *JUST MY INTUITION* for battery storage stress bathtub curves. To ask the question differently, how would adjust these lines for a generic EV battery?

View attachment 409574

In relation to your first question, I was thinking about absolute SoC, which as you rightly point out, for Tesla is pretty close to indicated SoC. Absolute max SoC will usually be set by degradation at max indicated SoC in hot climates, whilst minimum absolute SoC will typically be set by power availability at low SoC - different manufacturers will have differing opinions on acceptable power de-rates at low SoC, differing opinions on how long each battery pack shall last, and differing opinions on whether to attribute more degradation to having faster charging or a wider SoC window. I believe Tesla assumes that most customers will rarely use the full SoC range (100% indicated SoC to 0% indicated), whereas other manufacturers ensure that under this worst case condition, degradation remains acceptable.

On the second point, the curves you've created are interesting, and I'm sure with some further developments are the under-pinnings of a new app!

At high temperatures, your curves look approximately correct. As our degradation modes at high temperature are primarily related to SEI growth and electrolyte decomposition, the dependance on temperature is related to the rate of [undesired chemical] reaction, hence varies roughly as per the Arrhenius equation, which is roughly as you've indicated.

At low temperature, degradation is typically more around mechanical stress, unless we get to very very low SoC (think parking at 0% SoC and leaving the vehicle for a long time). The mechanical stress is caused by the movement of the lithium ions, and as the lithium ions move more easily at higher temperatures, your stress intensities may actually reverse in their temperature relation at low SoC. The impact of the mechanical stress will depend upon the type of cell, cell design and cathode chemistry, hence it is difficult to generalise for all EVs. It is also worth noting that when the cell is new, the anode is designed to have greater capacity than the cathode and the electrolyte holds an excess of cyclable lithium. As the anode excess and cyclable lithium reduce, the mechanical stress at low SoC will become greater. The storage stress at low SoC should also be much lower than at high SoC - roughly 40-50% but depends upon many factors and will change through life! Sorry for the non-definitive answer....

 

[EV-Tech Exp]

Speaking as an electronic engineer with some knowledge of battery chemistry, that is a great video. I learned more than I thought. Oh and I didn't realize Aluminum was a 5 syllable word Ha, just kidding. The English is certainly better than mine.
I'd love to see a video explaining what happens (and why) these batteries become such an explosive fireball when compromised.
Thanks for a fabulous presentation.

Thanks - I'll definitely put that on my list of future videos!

 

[EV-Tech Exp]

According to the graph, 15-75% seems better than 20~90%.
I have been charging to 85% every 4 days.
I am changing it to 75% from now on.
And for longer trips I will only charge to 93%, especially in summer months.

If you don't need 85%, 75% should help when the batter is hot but shouldn't make a huge difference unless you live in a hot climate. If you're not in a hot climate, I'd probably just stick to 85%.

 

[EV-Tech Exp]

No matter where the flat part of the curve lies, there will always be decreasing marginal returns to keeping your vehicle at a lower maximum SOC, right?

Huge longevity gains by charging to 90% daily as opposed to 100% daily, and presumably some longevity gains by charging to 80% daily, but not nearly as much.

And I assume it will also be heavily dependent on one's daily routine. If you can keep your battery at 80%, that's great, but if that SOC would cause you to Supercharge more often, I have to guess that more frequent supercharging would offset any longevity gains from maintaining that lower SOC.

I'm taking our SR+ for our first short road-trip in August, and we essentially have a choice between charging to 100% or supercharging along the route. Given all that's been shared so far, I'm pretty sure that briefly touching 100% is healthier for the battery than any supercharging.

You're spot on with everything you say - 100% indicated SoC is not 100% absolute SoC, and for short time periods is absolutely fine.

An interesting piece of academic literature (can't recall the title but will update the post later with reference) has found, (for particular cells which presumably contain specific degradation suppressing additives), little to no increased degradation for cells cycled upto a maximum SoC of between ~60 - ~90. Below 60% max SoC, degradation was reduced, and ~60% soC corresponds to the approximate SoC at which the cathode structural phase change occurs for an NMC or NCA cathode.

Thus, the conclusion was that if max SoC is taken below unstable levels (~90% SoC in the study), the difference that reducing max SoC further makes is negligible in comparison to the additional structural degradation induced by the regular phase changes of the cathode. The upper SoC stability level will be highly impacted by temperature however and will differ for every cell.

 

[EV-Tech Exp]

@EV-Tech Exp for a battery lay-person like myself, does the current research on lithium ion battery degradation show any sort of priority or ranking in the severity of damage caused by the various mechanisms?

Would you be able to say something like "In the degraded batteries we've studied, we think the leading cause of degradation has been...X" or "If you're looking to increase the life of your battery with the fewest possible lifestyle changes, the one thing you can change that has the most impact would be...Y"?

For 'normal' degradation, i.e. not a catastrophic event such as lithium plating, the primary degradation mechanism is SEI growth, which consumes cyclable lithium, consumes active material and increases resistance.

The summary of what you can best do is in the image in the first post on this thread. It is really about keeping temperature as close to 10-20 deg C as possible, not using the absolute extremes of the SoC of the battery and only fast charging when needed.

Great thread with tons of great info. I have added the video to my watch later and will watch it later today. I plan on making a video about what degradation I've seen so far on my TM3 (1% or less after 12k miles) and I may reference this video. I will already be referencing the one from Professor Jeff Dahn.

I look forward to seeing the video!

 

[Leafdriver333]

What does Tesla's 8 year battery warranty cover?
How much degradation is considered defective?

 

[ninpb]

What does Tesla's 8 year battery warranty cover?
How much degradation is considered defective?

Below 70% of capacity.

 

[ninpb]

I still try to practice not charging immediately after driving. Habit from driving a leaf. Would a 45 minute half open half bumper to bumper drive be considered aggressive use of the battery?

 

[EV-Tech Exp]

70%.

Yes, 70% capacity retention for the Model 3. No figure for capacity loss is specified for the Model S and Model X.

 

[Zoomit]

On the second point, the curves you've created are interesting, and I'm sure with some further developments are the under-pinnings of a new app!

At high temperatures, your curves look approximately correct. As our degradation modes at high temperature are primarily related to SEI growth and electrolyte decomposition, the dependance on temperature is related to the rate of [undesired chemical] reaction, hence varies roughly as per the Arrhenius equation, which is roughly as you've indicated.

At low temperature, degradation is typically more around mechanical stress, unless we get to very very low SoC (think parking at 0% SoC and leaving the vehicle for a long time). The mechanical stress is caused by the movement of the lithium ions, and as the lithium ions move more easily at higher temperatures, your stress intensities may actually reverse in their temperature relation at low SoC. The impact of the mechanical stress will depend upon the type of cell, cell design and cathode chemistry, hence it is difficult to generalise for all EVs. It is also worth noting that when the cell is new, the anode is designed to have greater capacity than the cathode and the electrolyte holds an excess of cyclable lithium. As the anode excess and cyclable lithium reduce, the mechanical stress at low SoC will become greater. The storage stress at low SoC should also be much lower than at high SoC - roughly 40-50% but depends upon many factors and will change through life! Sorry for the non-definitive answer....

Thank you for the insightful responses. I certainly understand not being able to provide a definitive answer and didn't expect one.

Your feedback was helpful and so I updated my generalized storage stress graph. I only adjusted the low SoC to reflect the "40-50%" peak and a reversal with temperature. That end of the curve is less interesting as we don't normally store cars with <10% SoC.

The high SoC habits are much more relevant to battery longevity. The warmer curves roughly depict a doubling of "stress" between 80 and 90% and doubling again between 90 and 100% absolute SoC.

 

[EV-Tech Exp]

Thank you for the insightful responses. I certainly understand not being able to provide a definitive answer and didn't expect one.

Your feedback was helpful and so I updated my generalized storage stress graph. I only adjusted the low SoC to reflect the "40-50%" peak and a reversal with temperature. That end of the curve is less interesting as we don't normally store cars with <10% SoC.

The high SoC habits are much more relevant to battery longevity. The warmer curves roughly depict a doubling of "stress" between 80 and 90% and doubling again between 90 and 100% absolute SoC.

View attachment 410672

As a general rule of thumb, this looks excellent!

 

[EV-Tech Exp]

I'm curious as to what the ideal SoC would be as well! Since I can't charge at home I tend just fill up to ~80% at work once I drop below 30%. I'm sure it's all negligible but if it's just a habitual change I need to make for marginal gain might as well start early .

Also curious if you have any thoughts on the battery fires in Asia that caused Tesla to release a patch for S/X. How do thermal runaways work in a battery cell? Is it just like lighting a fire where there's no return after an "ignition" temperature is reached?

Thanks for making these videos. It's really helpful for someone who's interested in this kind of research but doesn't have the background to make sense of the academic papers.

It sounds like you're doing a good job of looking after your battery already.

Need to read more about what caused the battery fires to comment really, but as far as thermal runaway goes, after the cell is saturated at a particular temperature (around 160 deg C for NCA cells), the cathode will begin to decompose, which in the process releases tremendous amounts of heat, resulting in continued heating, electrolyte boiling and gas generation, misalignment of layers, internal short circuit and usually this eventually leads to fire. Once combusting, the cathode releases oxygen hence the fire is self-sustaining so very difficult to put out.

Glad you found the video helpful!

 

[EV-Tech Exp]

I leave my Rav4 under a sunshade at the airport. Now that I have the Model 3 it is only on pace to drive about 2500 miles this year. The car does not have an active cooling system when it is turned off so even after weeks of non usage my SOC stays the same. Which is nice for my convenience but battery degradation concerns me. I try to leave it as close to 50% as possible and so far I think my range degradation has been minimal. Once every two month or so I take it on my 200 mile commute just to "stretch it's legs". This trip requires a fast charge at the destination from ~5% to 90% to get home. I immediately charge the car to 50% using my 32amp charger and return the car to the airport carport, according to this that is not good, I'm curious how not good that is? With our model 3's, would setting our car to start charge 2-3 hours after we get home better for the battery then charging right away to get out of the low % charge. I read somewhere that occasional fast charging is actually good for the car, I don't remember where but this video doesn't really support that. For the most part, I have followed these rules but I didn't know why, this was a nice watch.

It sounds like your Rav4 regime is good for extending lifetime. Following your return leg, it would reduce degradation if you waited for the battery to cool before plugging in, however given the (relatively) slow rate of charge, and the infrequent journeys, I wouldn't worry about it as the difference will be pretty small.

For the Model 3, given the effectiveness of the active cooling system, unless you've been doing sustained high speed driving (>90mph) or a very long journey which has included lots of supercharging your battery should not be excessively hot by the time you return home - (this is not true for something along the lines of a 40kWh Nissan Leaf which can get very hot, very quickly with what I'd consider a normal highway drive). As per the plot Zoomit created, it is going to be good to get out of the low SoC region, and given the relatively low charge rate, by the time you get to high SoC, the thermal management system will have cooled the battery down to a comfortable temperature, so again, I think you're doing the right thing as you are. If you were going to use something equivalent to a supercharger when you got home, I'd suggest waiting for it to cool, but a 32A charger will be fine to plug in right away.

 

[evM3Sekar]

...
At high temperatures, your curves look approximately correct. As our degradation modes at high temperature are primarily related to SEI growth and electrolyte decomposition, the dependance on temperature is related to the rate of [undesired chemical] reaction, hence varies roughly as per the Arrhenius equation, which is roughly as you've indicated.

At low temperature, degradation is typically more around mechanical stress, unless we get to very very low SoC (think parking at 0% SoC and leaving the vehicle for a long time). The mechanical stress is caused by the movement of the lithium ions, and as the lithium ions move more easily at higher temperatures, your stress intensities may actually reverse in their temperature relation at low SoC. The impact of the mechanical stress will depend upon the type of cell, cell design and cathode chemistry, hence it is difficult to generalise for all EVs. It is also worth noting that when the cell is new, the anode is designed to have greater capacity than the cathode and the electrolyte holds an excess of cyclable lithium. As the anode excess and cyclable lithium reduce, the mechanical stress at low SoC will become greater. The storage stress at low SoC should also be much lower than at high SoC - roughly 40-50% but depends upon many factors and will change through life! Sorry for the non-definitive answer....

@EV-Tech Exp In mechanical components subject to cyclic temperature loads, thermo-mechanical fatigue (TMF) results in loss of stress carrying capacity - the mechanism seen here include cracking and exfoliation (similar to Li-Ion batteries). Since these temperature and/or SoC cycles in Li-Ion batteries lead to capacity loss, the natural question is:

1. Has anyone (in your school or lab) done any studies on TMF-like models for these batteries in order to understand the effect of the temperature + SoC cycles? The intent would be to find optimal operating conditions that extends the life of the battery.
2. If Tesla knows that high temperatures is not good for the battery (as when the car is parked in direct sun in hot areas), why not do active cooling even when the car is parked? Or do they?

BTW, I have watched all your videos and they are very informative. Thank you for your time.

I also found another useful resource on the web: Battery Information Table of Contents, Basic to Advanced

 

[DirtyT3sla]

For 'normal' degradation, i.e. not a catastrophic event such as lithium plating, the primary degradation mechanism is SEI growth, which consumes cyclable lithium, consumes active material and increases resistance.

The summary of what you can best do is in the image in the first post on this thread. It is really about keeping temperature as close to 10-20 deg C as possible, not using the absolute extremes of the SoC of the battery and only fast charging when needed.



I look forward to seeing the video!

Here's the video for anyone interested:

I have linked your scientific explanation in the description

(Ignore the thumbnail lol. Don't hate the player, hate the game).

 

[EV-Tech Exp]

@EV-Tech Exp In mechanical components subject to cyclic temperature loads, thermo-mechanical fatigue (TMF) results in loss of stress carrying capacity - the mechanism seen here include cracking and exfoliation (similar to Li-Ion batteries). Since these temperature and/or SoC cycles in Li-Ion batteries lead to capacity loss, the natural question is:

1. Has anyone (in your school or lab) done any studies on TMF-like models for these batteries in order to understand the effect of the temperature + SoC cycles? The intent would be to find optimal operating conditions that extends the life of the battery.
2. If Tesla knows that high temperatures is not good for the battery (as when the car is parked in direct sun in hot areas), why not do active cooling even when the car is parked? Or do they?

BTW, I have watched all your videos and they are very informative. Thank you for your time.

I also found another useful resource on the web: Battery Information Table of Contents, Basic to Advanced

There are some very thoughtful and interesting questions in this thread!

1. For TMF to be a concern for a typical electrode, the temperature window needs to be over a much higher window than we would normally operate or store a Li-Ion cell, (because of the limited thermal stability window of the electrolyte). The volume expansion of the electrolyte at low temperature theoretically could result in TMF of an electrode, however some work out of Jeff Dahn's group shows that doing this a limited number of times should not have much of an effect.

Link to paper (Open Access) - Differential Thermal Analysis of Li-Ion Cells as an Effective Probe of Liquid Electrolyte Evolution during Aging

2. I'm not sure what Tesla does on this, however I would expect them to circulate coolant when the battery is becoming excessively hot, i.e. >50°C. They do however need to balance this with the customer expectation that the vehicle does not consume excessive power when sat unused. A fine balance needs to be struck. Maybe those who live in hotter climates can help us understand what their vehicle does?

 

[Jeff N]

I just noticed this thread and am looking forward to watching the video later today.

Although it is mostly off-topic from degradation issues, I recently wrote an article about NMC 811 that might be of interest to people reading this thread although we probably shouldn’t discuss it further here.

https://electricrevs.com/2019/05/31/report-sk-innovation-to-begin-making-nmc-811-cells-in-q3-2019/

 

[Tron 3]

Thank you for the thorough explanations, I have never owned a $10,000 battery before, so I have made it my business to understand the best practices for caring for it. I think your videos should be required "reading" for any EV owner, and I am somewhat surprised Tesla doesn't do a better job of educating their customer.

Born in 1955 I've owned my share of motor vehicles, and nothing comes close to this LR AWD M3.
Have you taken delivery of your M3 yet?