Published on March 27th, 2015 | by Jake Richardson


Battery Tests Show Some Lithium-Ion Cells Last Over 5x Longer

March 27th, 2015 by  

Researcher Andreas Gutsch from the Karlsruhe Institute of Technology said at the PV Symposium in Germany recently that some lithium-ion cells working in storage systems experienced up to 30% capacity loss after 1,000 cycles, while some cells have a better capacity after 5,000 cycles. In other words, there can be quite a lot of variation in lithium-ion cell performance.


Testing conditions did not allow him to reveal the names of cell manufacturers. The cells were not easy to obtain for testing and he could not disclose that specific information. However, his test results did reveal that cells from China that were tested had the performance that decreased most rapidly. Cells from Japan and Germany did the best, while ones from South Korea and the US were in the middle of the performance range.

One battery type that was specifically mentioned was the Tesla battery. This battery was found to last 400 cycles and was indicated to have the ability to power a car for about 200,000 kilometers (124,000 miles): 400 cycles multiplied by 500 km per full charge. That said, Elon Musk has stated that Tesla has a Model S in the lab with 500,000 miles (805,000 kilometers) on it, and without excess degradation. So, it seems the tests have been conducted very differently or something funky is going on.

Gutsch previously noted that some energy storage battery types might have safety issues, such as thermal runaway. According to his research, safety standards for some batteries have not been met.

He does believe batteries will eventually be a great way to store electricity, though.

All, or most of us lay people are on a learning curve when it comes considering how to evaluate the best batteries for consumer technology performance. To engineers, it may not be surprising at all that there could be such a range in cell performance. It may be fairly dismaying to lay persons to find this out, though we may notice that some of our consumer devices (like smartphones and other rechargeables) have poor to very poor battery lives, with variation being rather large.

Battery storage for a home or business is quite another matter.

Image Credit: Claus Ableiter, Wiki Commons

Buy a cool T-shirt or mug in the CleanTechnica store!
Keep up to date with all the hottest cleantech news by subscribing to our (free) cleantech daily newsletter or weekly newsletter, or keep an eye on sector-specific news by getting our (also free) solar energy newsletter, electric vehicle newsletter, or wind energy newsletter.

Tags: , , , , , , , , , , , ,

About the Author

Hello, I have been writing online for some time, and enjoy the outdoors. If you like, you can follow me on Google Plus.

  • JoGent

    The point about the batteries in a Telsa S is they are hooked up to some v smart software to manage the charging and discharging very carefully. The pattern of usage seems to make a big difference to the degradation. I can’t see how the researchers had any access to this Tesla magic.

  • Hans Nerdlichmann

    While this is a very important issue, I do not see anything reported here other than a survey with the expected results.

    For leading edge insight into LIB longevity, I suggest tracking the work of Jeff Dahn, PhD and Rachid Yazami, PhD. They come at this, respectively, from coulombic-efficiency and and thermodynamic perspectives. Both have unique insight, and are leaders in this field.

    • Bob_Wallace

      Willing to summarize what has been found to date?

    • vensonata

      Hey, another Jeff Dahn fan. Somebody from Clean Technica really needs to interview this guy.

  • Lynne Whelden

    It’s about time someone studied this!
    I’ve noticed for decades that my camcorder batteries (Li-ion and Nickel metal hydride) are all over the spectrum. Some will never give up. Others die after a few charges. I have some Li-ion batteries here right now that I’ve had for eleven years, have been used hard (from freezing to boiling hot conditions) and they’re “like new.” Why?

    • vensonata

      That is why I have suggested a website exclusively devoted to energy storage. Because of proprietary interests the information on batteries is covered in fog. Even lead acid producers tell the wildest fabrications about their “special recipes” and why they are superior to everybody else. It is a land of myth and religious belief more than science and fact.

  • eveee

    Gutsch is a former Evonik employee. The firm developed a battery with ceramic separator, Li-Tec. Mercedes Benz purchased the company, which supplied batteries for the Smart EV. Shortly after, MB decided to shut the manufacturing down, concentrating on research, citing costs, and the overcapacity of cell manufacturing relative to demand at the time.

    • vensonata

      So, is he legit?
      400 cycles for Tesla batteries sounds very skinny indeed, even suspicously so.
      I presume the “Japanese and German batteries” are Sony and Bosch. Sony was the first to take up lithium batteries after they were invented in the 90’s by an American. They have a stationary pack for sale in Germany right now that they advertise as “8000 cycles to 100% depth of discharge” Storage life: “over 90% capacity after 10 years” Cost $12,500 for ten kwh, includes bms and inverter. Now here is the small print “The data after one year is based on simulation”. There you have it: 8000 cycles at 1 per day = 22 years. And they suggest a 20 year lifespan but do not warranty it. They are smart not to, since it is based on simulations and extrapolation.
      The Bosch battery is about the same price and configuration but “7000”cycles to 80%. And they suggest it could last 25 years. Again, not a guarantee, just a naive math projection.
      I am guessing that the Sony battery does not discharge to 100% but is actually a 12 kwh battery that gives you 10 kwh. So actually 80% discharge unless they have some private tech magic that nobody else does. So thats all I know.

      • eveee

        As I commented above, battery manufacturers shroud themselves in secrecy. It would be nice if there were published curves and an independent test body.

        I wonder which battery boasts 8000 cycles at 100% DOD. The only lithium battery that can approach that is titanate, but they have low energy density. The first Brammo Empulse was a titanate battery that had a very high cycle life number, but short range.

        This is all a marketing bait and switch sort of thing with manufacturers claims. Typically, they do not show all factors, only boasting those which are boosted. Since power, energy, density, cycle life, and cost are all juggled, touting one factor while hiding another defeats the purpose of analyzing performance.

        Unfortunately, we see this all too often.

        • vensonata

          The Sony battery is called “Fortelion” you can find it on google on German sites. I think it is lifepo4, and it is truly a head scratcher why they advertise 100% DOD.

          • eveee

            Yes, it is LiFePo.

            Taking a look at the Fortelion specs, they have low output voltage. That type is likely low energy, which is what happens when you lower output voltage. Cycle life improves as well.

            This is why I said you have to look at everything to compare.

            Also need to see depth of discharge, because cycle life is higher with less discharge.

            Notice that at 3A the cyle life is 6000, but only 4000 at 10A. Its important to rate cycle life at some fixed degradation like 80% for comparison.

            This battery performance drops significantly at low temperatures.

            Some batteries have higher series resistance than others and that limits charge/discharge rates. Tesla parallels devices to deal with that.

            Overall, I think you can find better batteries than this one today. Fortelion came out in 2009.


          • vensonata

            Interesting that in that Sony pdf, (in somewhat butchered English), they now say “6000 cycles to 80% remaining at 100% degree of discharge” So now we lost 2000 cycles and 10% but they still insist on 100% DOD. The Japanese may not be all that interested in precise, reliable advertising claims, it seems. All these claims start to seem mushy. Let the buyer beware.

          • eveee

            vensonata – Battery makers settled on a 100% discharge for battery comparison. Thats the industries standard for measurement. So Sony is following industry convention. So Sony is being precise and reliable.

            Lets straighten out the jargon. They mean .. 80% capacity is left after 6000 100% discharge cycles.

            2000 cycles were not lost. What was lost is 20% of capacity. Note that the batteries can still be cycled further, but with less capacity. Nissan has noted this by suggesting used Leaf batteries be used for utility storage after they lose 20% capacity. Notice that this means one can drive an EV further than the limits suggest, if one is OK with the capacity loss.

          • vensonata

            The claim in the german document was 8000 cycles to 100% DOD. They did not say what percentage was left. But this pdf says 6000 to 100%DOD with 80% capacity left. That is what I mean by the “missing” 2000 cycles.

          • eveee

            Here is how it must be expressed to be clear.
            X cycles to Y % DOD with Z% remaining capacity.
            industry standard is 80% remaining capacity.
            Leaving Z out in the first statement means its incomplete.
            The second statement is complete.
            And some people say all battery manufacturers ….
            are something uncomplimentary.
            Especially when they try to figure out what happened to
            the “missing” 2000 cycles. If you get my drift.

          • Edux

            LFP has in general a lower output voltage than other cathodes, this is a well known fact. In case of Sony’s Fortelion it goes from 2V (cutoff voltage) to 3.6V (max voltage).

            The Sony’s Fortelion specs says that the cell can deliver 100% DoD (2V-3.6V) @ up to 6C discharge current (1C = 3A) which is a lot of power for a single cell.

            In addition, the values mentioned in the specs clearly said that they are real values measured from mass production cells. The cycle test has been done with 1C charge/discharge current @ 23°C (you can do several cycles per day). The last information I saw mentioned 15000 cycles and a remaining capacity of 62%.

            Other test made by the Fraunhofer Institute (a German research institute) also confirmed the quality of Sony’s Fortelion cell. They simulate a PV application with 60% DoD and after 1000 cycles Fortelion’s cell had almost 100% from its original capacity.

            Another fact that confirms Sony’s leading position in this area is that Hydro-Quebec (co-owner with the Texas University of the LFP patents rights) started a Joint-Venture with Sony (called Esstalion Technologies) to develop the next generation LFP cell.

            There are for sure some few excellent cells for stationary applications, Sony Fortelion is without a doubt one of them.

          • eveee

            Yes. Indeed the Fortelion looks like a decent battery. I was trying to point out its particular characteristics which are different from generic laptop batteries, power tool batteries, and EV batteries. Fortelion is not so good at low temperatures, making it less desirable for EVs. However, its long cycle life seems purpose designed for utility backup, or more likely, telecom or computer center backup, where temperature might be better regulated.
            Its not designed for maximum density. Its designed for lifetime.

  • AhmetA

    Tesla uses Japan made cells, article says that german and japan made cells are best but Tesla Bateries is bad contradict with what is said.

  • OmarSultan

    Kind of a poorly written article. The original research was regarding testing cells for static storage applications. The linked article indicates the testing on the Panasonic cells used in the Tesla pack, not the Tesla pack a as whole, which has anti-brick protection built in and would not work with the testing methodology they employed.

  • Bob_Wallace

    The Toshiba SCiB is rated at 6,000 cycles. When it was put in the Honda FiT the claim was 4,000 full discharge cycles.

    In a 200 mile range EV that would be 800,000 miles.

    Proterra is using SCiBs in their buses and talking 10,000 cycles.

    Why, when I look for data on battery life, do I not see statements of calendar life?

    • vensonata

      Beats me about calendar life. Possibly too early in the game, but as you can see in the article itself, these techy types are stating that lifespan expectations of Tesla batteries are precarious. Watch that lecture by Jeff Dahn, he is eminently qualified to comment on the matter…his is “Canada research chair” on this topic, (that is a prestigious title in Canada). I myself am merely the “chief inspector of snowstorms” in my neighborhood, so don’t take my word for it.

    • Benjamin Nead

      Toshiba SCiBs are lithium titanate, which have relatively low energy density
      (2.4V or 2.5V per cell, charging at around 2.9V) but stand up to many more
      cycles than just about all other lithium formula. They stand up well to quick charging and are also very tolerant to temperature extremes. Some of the Japanese Mitsubishi iMiEVs come with SCiBs (not offered on the North American version, alas, but I wish they were.)

      More here on lithium titanate cells. I’m intrigued by one of the
      commenters who is building PV systems using lithium titanate
      cells for battery backup . . .

      Just for fun, I’m currently building a small Free Flight model airplane
      around this tiny 50mAh lithium titanate cell . . .

      . . . and charge it with one of these, which allows me to precisely dial in
      voltage and amperage (too easy to burn up a battery like this with
      an off-the-shelf single cell hobby LiPo charger, which outputs 4.2V) . . .

    • Jouni Valkonen

      I think that there are no mentioning about calendar life, because most of the batteries are so new that there is no established scientific data how batteries are behaving over longer time periods. Any extrapolations from laboratory simulations would have so large error margins that it would be just wrong to claim anything other than lower end of expected calendar life. And this lower end would probably be so low that it is not very good advertisement.

      Tesla however has never promised longer expected lifespan for batteries than 10 years. However, they have been very confident that most of the batteries will survive the 8 year warranty period.

  • vensonata

    Tesla battery warranty for the model S 85 is: 8 years unlimited mileage. At 30,000 miles per year (who drives that much?) that is 113 full charges per year. 904 cycles at 8 years. 240,000 miles. There should be 70% available then. By the way, the warranty does not cover degradation at all! only mechanical failure. So if there was only 50% left at 5 years its not their problem.
    I doubt that Elon Musk wants bad publicity so Tesla is quite confident that most batteries will complete 1000 cycles and still have 80% left and that most people will not drive anywhere near 30,000 miles per year. So… lots of happy customers driving 10 year old Teslas. Now consider a replacement battery price 10 years from now, perhaps $8500 or $100 per kwh. And not too many Americans care to drive cars more than 10 years old so maybe replacement batteries are almost irrelevant.

    That said, I have been commenting occasionally over the last few months on the calendar life vs cycle life for all types of lithium batteries. Predicting calendar life based on simulated shallow high cycling 100 day test is virtually null and void. Other factors to do with anode chemistry produce entropy and degradation which overshadows the cycling effect. (See lecture on youtube by professor Jeff Dahn of Dalhousie University on lithium battery testing)

    • OmarSultan

      I actually drive 25-30K miles per year. 🙂 My 19 month old Model S has 52K miles on it. It will currently max charge to 259 rated miles, which works out to ~2% loss of capacity. I am OK with that. 🙂

      • Bob_Wallace

        Capacity loss patterns for lithium-ion batteries show early life losses which then slow. Is that what you saw or was the 2% spread more evenly over the 52k?

        • OmarSultan

          Pretty much. On a full charge, my rated miles dropped from 265 to ~261 in one big step a bit before the one year mark. Since then, its been a slow drop, another mile or so–right now a max charge yields me 259 rated miles.

    • Alan Waterman

      They don’t cover *normal* degradation which they have already defined. With their claims about how little it will degrade over 500K miles, if your battery degrades more than that, it would be covered under warranty. For example, if it degrades 50% after 100K miles, then of course that would be covered as it exceeds their threshold for normal degradation at that mileage.

      • vensonata

        Are you sure about that? My source for that info on degradation vs mechanical failure was a Tesla owning chief science officer for (I think) Greencar reports. Remember, the original roadster warranty was 5 years 50,000 miles expecting 70% remaining, but you had to pre purchase a $12,000 new battery. So they were really cautious about degradation. I would be delighted if you were right about unusual degradation being covered by the warranty but I fear, on closer reading, you may find you are not.

  • Jason hm

    There are some wild variables at work within the chemistry and workings a lithium ion cell. Perhaps this is why Tesla is going with small cell sizes. Small sample size = higher variability so to minimize individual cell performance variation and create a more consistent and stable battery it’s probable better to build battery packs out of hundreds or thousands of cell instead of building monster cells.

  • elo

    Real-world performance shows otherwise. I’m highly skeptical about this researcher because real-world Teslas already have 150k+ miles with no substantial degradation.

  • Matt

    Now I’m not saying that there was any creative science or wording going on but.
    Researcher is Germany. Results and testing methods could not be share because of agreements with people funding. Germany batteries did best! assume that means the “while some cells have a better capacity after 5,000 cycle”.
    Can’t talk about which batteries test, but will bad mouth TESLA! Seen much of that out of Germany lately.
    Spread in performance, not surprising.
    Belief in these “secret” result, three grains of salt please.

  • Jouni Valkonen

    The lithium ion longevity is heavily depended on charge level. Rule of thumb is that 4.1 V charge level lasts twice as long as 4.2 V charge level. And 4.0 V charge level lasts twice as long as 4.1. And so fort. This also roughly corresponds that Tesla batteries lasts with 100 % charge level 400 cycles, 90 % charge level 800 cycles, 80 % charge level 1600 cycles, 70 % charge level lasts 3600 cycles and 60 % charge level, you guessed, last for 7200 cycles.

    After 1500-3000 cycles, however, the shelf-life and storage temperature of lithium batteries becomes dominant factor in degradation, therefore Tesla does not expect longer usable life for its batteries than 10 years.

    In sum, this compares apples to oranges. EV batteries lasts much longer than their nominal cycle life, because in normal use, EV batteries are charged only to 70 to 80 %. Therefore this “research” has zero value. E.g. Iron-phosphate batteries has lower nominal charge level than Tesla’s NCA batteries, therefore they typically last longer.

  • JP

    “One battery type that was specifically mentioned was the Tesla battery. This battery was found to last 400 cycles and was indicated to have the ability to power a car for about 200,000 kilometers (124,000 miles): 400 cycles multiplied by 500 km per full charge. That said, Elon Musk has stated that Tesla has a Model S in the lab with 500,000 miles (805,000 kilometers) on it, and without excess degradation. So, it seems the tests have been conducted very differently or something funky is going on.”

    You’re going to need to clarify this paragraph. What happened after 200,000 kilometers (400 cycles)? Did the battery die? Or did it just start to experience degradation after that? If so, how much degradation?

    • Jouni Valkonen

      Above statement simpli is not suported by facts. See also my other post where I explained where was the error of this test.

      Tesla Model S has been around now almost three years and up this point, nobody has reporeted significant battery degradation, although Tesla drivers are driving on average more than ICE car drivers and are using fast charging quite often.

      • eveee

        We don’t know the actual Tesla battery conditions because the BMS can limit the battery conditions to enhance cell life. There is reason to doubt that Tesla would design its BMS to allow excessive charge or discharge. Tesla wisely over designed its battery capacity and its logical that they would have used the BMS to operate the battery conservatively. Given the results published for the roadster, this is clear. The roadster battery type, LiCo, is not the best for cycle life compared to NMC or NCA, and yet the cycle life is still good.

        Tesla is using every trick there is. By using LiCo 18650 cells in parallel, they limit C rates, a major factor in battery life. Thats just one example.

        To obtain optimum performance, ts not just a question of the best battery, its also a question of how its operation range limitations are avoided by good design.

Back to Top ↑