Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Why do Li-ion Batteries die? And how to improve the situation?

This site may earn commission on affiliate links.
Great video,

Since the professor mentioned some numbers, here are a couple of cell voltage numbers I have measured, percentages are Tesla SOC not the battery manufacturers:

100%Rated Range 246cell voltage 4.15V
~50%Rated Range 105cell voltage ~3.7V (I need to repeat this measurement to be more accurate)
8%Rated Range 0cell voltage 3.33V cold and under new software, 3.29V warm and under old softwareI'll repeat when it warms up this spring


I don't happen to have exactly where 4.0V is in terms of Tesla SOC, it should be about 85%, I'll add that to my list to capture this spring.

Peter
 
Very interesting presentation and thread, thanks !

The more I learned about li-ion technology, the more angry I am getting at notebook manufacturers who more-or-less force users to charge to 100% in hot notebooks and also do not assist in preventing discharging to zero.

Even my small company has purchased quite many (expensive) Dell accupacks as they needed replacement every 2 years. This could have easily been prevented by smart software in the bios. Some investigatuon shows that some notebook manufacturers do now offer this. I think Dell has lost my future business, unless they quickly fix this. I will pay attention to this from now on.
Same goes for other equipment, think of the Apple products with hard to replace accu's. As we saw in the presentation accu's can have a very, very long usefull lifespann. I am getting the impression manufacturers actually prefer accus to fail to sell more accus and newer products. Not only a big waste. This has also given EV's a big extra hurdle to overcome.
+1

We purchase many of the mid line Dell laptops at work (primarily Latitudes, now going into XPS) and I've often thought the same thing. Then within the last year, I purchased two inspirons for personal use, and lo and behold they shipped with an advanced battery management option in the power option of the control panel. They give me the option of setting "Desktop Mode" which will only charge the laptop to 50% max, or another option which I believe will charge it to somewhere between 80%-90%. Was very excited to see these options. Will try to post screenshots if possible.
 
here's the professor's latest response:

"George,


Again - my responses are in CAPS below


Jeff Dahn, FRSC
Professor of Physics and Atmospheric Science
NSERC/3M Canada Industrial Research Chair
Canada Research Chair

Subject: Re: Question RE: "Why do Li-ion Batteries die ? and how to improve the situation? "

Professor Dahn, if you have a chance, I have a few more follow up questions. I do very much appreciate your time answering my questions!

Since they are both cobalt based would the consumer care recommendations (store at lowish SOC, Lowish temperatures, tending to mid voltage cycling) for LiCo02 cells be the same as a LiNiCoAi02 or a LiCoAi02 cell?

YES

Also, does the rate of charge (0.5C vs. 1C vs 1.5C etc.) have any impact on the cell life?

YES - I WOULD AVOID CHARGING AT GREATER THAN C-RATE AND WOULD RECOMMEND C/2, ESPECIALLY FOR "ENERGY CELLS". POWER CELLS (E.G. IN TOOLS) WILL HAVE LITTLE PROBLEM WITH C-RATE CHARGING AND WOULD BE FINE. THE CELLS IN THE TESLA S ARE MORE AT THE ENERGY CELL END OF THE SPECTRUM.


have you tested NCA cells like the Panasonic NCR18650A and NCR18650B and would their care characteristics be the same.

WE HAVE LOOKED AT THE "A" AN IT IS VERY SIMILAR TO LCO CELLS

Nicad is less and less common in consumer electronics but is still used and I have several gadgets that have them. I have been treating these pretty much the same (store at lowish SOC, Lowish temperatures, tending to mid voltage cycling) accept every so often I do a complete discharge and full charge to reduce the tendency toward memory effect. In your opinion, is this on track? anything to add?

THIS IS GOOD. MEMORY EFFECT COMES WHEN CELLS ARE LEFT FULLY CHARGED OR FLOATED AT FULL CHARGE FOR LONF PERIOS OF TIME.

In general, I am amazed at how few of devices/chargers have any sort of BMS other than avoiding runaway combustion. It would make so much sense for every one of these devices to have a long life mode, where they could be left plugged in but only charge to about 50%. so many of these devices end up plugged in all the time, and it dramatically cuts down their shelf life. please consider explicitly recommending this in your research if appropriate. To many phones, computers etc are thrown away because the battery only lasted a year or two, when, if treated better, could have lasted 5-10 years. by applying this care regimen, peak performance of many products can be vastly extended. I've gotten 5 years out of an iphone, 5+ years out of a cordless shaver and long life out of many other products. This has required a lot of careful charging... it would be so much easier if I could set the SOC limit on each device based on my needs, such as at home near a charger vs traveling where I need max charge. we need someone like you to push the industry to make this kind of thing a standard! (no pressure :~). Tesla with it's slider option has nailed this. some way to adjust the end charge limit should be on all consumer products. a SOC meter that showed red at both ends, with green in the middle, would help this be more intuitive for consumers, who now simply think fully charged is the best, which simply leads to killing the battery prematurely.

THIS WOULD BE A GOOD THING. MOST HUMANS TREAT BATTERIES A BLACK BOXES AND DO NOT THINK OR CARE. YOU ARE A RARE SOUL!

OH and as an aside, the Tesla S is aluminum, so no need to worry about rusting :~) and the AC induction motor is brushless with the only contacts points being two bearing sets... potentially, this car should last a very long time!

I KNEW ALL OF THIS AND BRIEFLY FORGOT! "
 
Thanks 100thMonkey! Good to hear the LCO (Roadster) advice is equally applicable to NCA (Model S).

Also, does the rate of charge (0.5C vs. 1C vs 1.5C etc.) have any impact on the cell life?

YES - I WOULD AVOID CHARGING AT GREATER THAN C-RATE AND WOULD RECOMMEND C/2, ESPECIALLY FOR "ENERGY CELLS". POWER CELLS (E.G. IN TOOLS) WILL HAVE LITTLE PROBLEM WITH C-RATE CHARGING AND WOULD BE FINE. THE CELLS IN THE TESLA S ARE MORE AT THE ENERGY CELL END OF THE SPECTRUM

C/2 for the 85kWh pack is 42.5kW, which is more than double any of us are capable of charging at home. I think we're safe with even the highest HPWC settings. Supercharging is 1.4C, but ramps down quickly at higher SOC.
 
Wow! That is awesome. Thank you!

I hope there is a chance to ask another question because I have one more:
Are the degradation properties (cycles, stability and temperature) for LiNiCoAi02 and LiCoAi02 the same as LiCo02? I ask because it is rumored that Tesla is using a Cobalt/Aluminum (less cobalt then LiCo02) chemistry for the Model S.

Sorry for so many question but the above is the holy grail of all questions for me. Again thank you so much. This is very helpful in navigating this strange lithium battery world. I sure wish Nissan would abandon NMC cells and passive cooling. It is a very reliable (2011/2012) car otherwise.
 
Are the degradation properties (cycles, stability and temperature) for LiNiCoAi02 and LiCoAi02 the same as LiCo02? I ask because it is rumored that Tesla is using a Cobalt/Aluminum (less cobalt then LiCo02) chemistry for the Model S.

I think that's covered by:

Since they are both cobalt based would the consumer care recommendations (store at lowish SOC, Lowish temperatures, tending to mid voltage cycling) for LiCo02 cells be the same as a LiNiCoAi02 or a LiCoAi02 cell?

YES

And by:

have you tested NCA cells like the Panasonic NCR18650A and NCR18650B and would their care characteristics be the same.

WE HAVE LOOKED AT THE "A" AN IT IS VERY SIMILAR TO LCO CELLS

And by:

I THINK THAT IF YOU ARE KEEPING THE CELLS BELOW 4.0V (NO WAY FOR YOU TO TELL THAT, HOWEVER) THE BATTERY LIFE WILL BE VERY IMPRESSIVE. WE ARE TESTING CELLS BUILT IN 2002 THAT HAVE 2002 TECHNOLOGY (LIFETIME IS BETTER NOW) THAT STILL HAVE 75% OF THEIR INITIAL CAPACITY (CYLCED AT 37c THE WHOLE TIME). THESE CELLS WERE CHARGED ONLY TO 4.075V. MODERN CELLS LIKE THOSE IN TESLA CHARGED TO 4.0V SHOULD LAST A FEW DECADES, I SUSPECT, WITHOUT ANY ISSUE (SO KEEP YOUR CAR FROM RUSTING!). WHERE IS 4.0V RELATIVE TO STATE OF CHARGE? MAYBE 75%.

Yes, LCO and NCA have similar properties, and you can expect better life from the more modern NCA cells.
 
LCO and NCA have similar properties, and you can expect better life from the more modern NCA cells.

JB Straubel reportedly said:

How do Tesla's 18650 lithium ion cells differ from the off-the-shelf versions?

Tesla is using the 18650 "form factor" but the chemistry and several of the components are designed by Tesla and nobody else has anything as advanced.

After seeing this video, I'm wondering if the Model S cells have modified additive(s) to extend battery life?
 
@djp
Except those consumer care tips apply to all lithium battery chemistry. My question is much more specific and I am not that interested in consumer care. I don't need a car battery to last 30 years. I am interested in the scientific variations/differences - if any - between cobalt based batteries with and without Aluminum, since his lecture did not test for batteries with Ai in them. Specifically how different is the almost flat degradation curve of LiCoO2 and batteries with Ai in them(LiNiCoAi02 and LiCoAi02).

It is mainly for clarification, so I don't make assumptions moving forward. Especially, since no matter how well you take care of a Nissan battery even in mild climates you still get substantial degradation (15%-20%) after 100,000 miles (Tesla might be 5%). I just want to know how adding Ai changes the cell and verify it doesn't change what is important. A car manufacture may sell a cheap car with a similar chemistry to Tesla and I want to be ready. I don't want to wait 10 years to buy a used Tesla I can afford.

I think that's covered by:



And by:



And by:



Yes, LCO and NCA have similar properties, and you can expect better life from the more modern NCA cells.
 
The NCA cells, (LiNiCoAlO2), should have significantly longer cycle life than straight LiCo cells. Pages 19 and 20 in this PDF have some relevant graphs:

Thanks for the link JRP3! Page 19 explains why active cooling for the battery pack is set to 40C on the Roadster and 60C on the Model S. The NCA chemistry has much better tolerance of high temperatures than LCO.
 
Is the secret sauce in the Model S cells just the cathode formulation process and the electrolytes? Or is there more?

The NCA cells, (LiNiCoAlO2), should have significantly longer cycle life than straight LiCo cells. Pages 19 and 20 in this PDF have some relevant graphs: http://www.embedded-world.eu/fileadmin/user_upload/pdf/batterie2011/Sonnemann_Panasonic.pdf

Not to mention much less expensive. LCO being 60% cobalt and NCA only using 9% cobalt.

Thanks for the link JRP3! Page 19 explains why active cooling for the battery pack is set to 40C on the Roadster and 60C on the Model S. The NCA chemistry has much better tolerance of high temperatures than LCO.
 
I think cathode formulation and the electrolyte combinations in the cell is the only thing we don't know. We know it is an 18650 form factor and we are pretty sure the cells are LiNiCoAlO2 and eventually someone will actually take apart the cell. Then we will know more. :smile:. Hope the professor does some more test with a LiNiCoAlO2 cell. I like his method. I think it will be the defacto method going forward for lithium battery testing.
 
It's one test. It doesn't test a cell's C rate or it's performance in cold temperatures. At this point I'm fairly sure Tesla's cells have some of the advanced additives and cycle life simply won't be an issue, so I doubt this test would reveal anything different.
 
Last edited:
Whatever those additives are they would have to be inexpensive or making a billion of them would not make sense. I believe all useful electrolytes are cheap, so the additives are most likely the electrolyte recipe. In all lithium batteries there is only the anode, cathode, electrolytes, and the SEI layer (some kind of polymer/plastic?). So I guess they could have done something to the SEI layer. Maybe a new material? I will be interesting to see what shakes out in the next five years.
 
I think you are confusing the plastic separator materials with the SEI layer. The SEI layer is not something that is manufactured, it forms during the initial charge cycle. The additives help prevent it from growing too thick over time.