Tesla battery researcher unveils new cell that could last 1M miles

https://electrek.co/2019/09/07/tesla-battery-cell-last-1-million-miles-robot-taxis/

by Osiris30 • 7 years ago

252 67 7 years ago

electrek.co

fake-name• 7 years ago

"Miles per battery" is possibly the stupidest metric to use to describe a battery capacity.

Any battery can last one million miles if it's just sitting on a seat in a vehicle.

19ylram49• 7 years ago

Come on, I’m sure that the tests are more extensive than that!

bryanlarsen• 7 years ago

I think it's the best metrics I've seen in a while.

"For usage where our previous battery gets 300,000 - 500,000 miles, this gets a million miles". It's a real world number.

sandworm1012• 7 years ago

> The new battery tested is a Li-Ion battery cell with a next-generation “single crystal” NMC cathode and a new advanced electrolyte.

If this is anything like growing single-crystal parts for aircraft, it won't be cheap. The real question should be whether these new battery modules will last twice as long while remaining less than twice as expensive.

> Controlling the charge to less than 100% state-of-charge also helps push the longevity.

Um, that is cheating. Running any battery at less than capacity will extend its life. You could put two batteries in the car, run them at 50% or alternate between them, and get double the life. No prizes for that.

zaroth2• 7 years ago

They tested it to 100% DOD (Depth of Discharge) over 5000 cycles;

https://electrek.co/wp-content/uploads/sites/3/2019/09/Scree...

jbm• 7 years ago

Um, that is cheating. Running any battery at less than capacity will extend its life. You could put two batteries in the car, run them at 50% or alternate between them, and get double the life. No prizes for that.

Don't phones already do this? (IE: 100% is not really 100%, 0% is not really 0%.)

I don't see it as cheating, it's akin to the practical limit vs. the actual limit

sandworm1012• 7 years ago

They do, but they don't claim this as any sort of advancement of battery technology. It's batteries 101.

zeta0134• 7 years ago

Many laptops do it too. Lenovo even markets it as a visible (and annoying) power management feature. Most however simply handle the charge cycles behind the scenes, and adjust their definition of "100%" over time, so the user can remain blissfully unaware (and less likely to change the setting).

I don't mind this really. Batteries degrade over time; that's kinda just how they work, and I'd much rather have 100% mean "my device is done charging" and 0% mean "my device can no longer run." The specifics of what voltage or charge level or whatever actually translates to those numbers is not important, and honestly the battery / laptop manufacturer probably knows the right settings for those better than I ever will, so it's fine.

miketery• 7 years ago

>> Controlling the charge to less than 100% state-of-charge also helps push the longevity.
Um, that is cheating. Running any battery at less than capacity will extend its life. You could put two batteries in the car, run them at 50% or alternate between them, and get double the life. No prizes for that.

I imagine the idea is that by going to X% instead of 100% fill, you extend by more than 1/X. So at 95%, perhaps we gain 20% life instead of the expected, 1/.95 ~= 1.05 -> extra 5% life.

manmal4• 7 years ago

Exactly. And this practice is widely followed by basically every BEV car manufacturer. Most of the more recent models even have a bigger battery than specified, and charge only up to 90-95% capacity while providing the specified range.

generatorguy2• 7 years ago

Who would have thought that microchips would become so cheap? And they are built on silicon crystals. The economies of scale for batteries for cars, phones, computers, will kick in in a way that they don’t for aerospace. Car or phone battery tech will end up in airplanes!

deepnotderp• 7 years ago

The amount of silicon in a microchip is tiny. By contrast, single crystal metal alloys for modern turbo fans comprise a significant portion of their cost.

hwillis10• 7 years ago

If this is anything like growing single-crystal parts for aircraft, it won't be cheap.

I'm still catching up, but it won't be expensive like that. You grow single crystals by keeping them hot and cooling them very slowly in a controlled atmosphere. Turbine blades are meant to operate in the hottest conditions in any machine on earth. Accordingly, they must be cooled at extremely high temperatures. The cost of single crystal turbine blades also pales in comparison to the cost of the blades themselves.

Um, that is cheating. Running any battery at less than capacity will extend its life. You could put two batteries in the car, run them at 50% or alternate between them, and get double the life.

Cycle life is often plotted as equivalent full cycles. If you get two batteries, running both at 50% DoD, you will increase the amount of lifetime full-cycle equivalents by ~10-50x, depending on the chemistry.

Lowering the depth of discharge means lowering the voltage difference between the electrodes. At 4.0 V you have very few side reactions. At 4.2 V, the battery is charged. At 4.3 V (~110% charge) the battery is dangerous. At 4.6 V (~120%) the battery is plating lithium onto its anode and is about to short circuit and blow its pressure release. Fire is heavily involved. Small decreases in voltage lead to big improvements in stability.

zaroth2• 7 years ago

Post of the actual paper;

https://news.ycombinator.com/item?id=20908726

For those commenting that the discharge rate is too slow in the test, or the depth of discharge was fudged... the tests were run with a high discharge rate combined with 100% charge at storage.

Figure 23 shows the projected fractional capacity of these NMC532/graphite cells as a function of time in years in a scenario where the cells are cycled once per day (100% DOD to 4.3 V) in a cycling event that takes 6 hours. It is also assumed that during the remaining time, cells are stored at full state of charge. It is clear from Figure 23 that these cells would provide an exceptionally long total driving range in an EV if the cells were maintained at an average tem- perature of 20°C. Even if the cells were continually at 40°C, 10 years of lifetime to 70% capacity and a total driven distance of 1,200,000 km is projected. It is worth noting that only 3650 cycles would be required for this total driven distance and 3700 cycles have been demonstrated in Figure 16.
Most important to realize is that Figure 23 assumes 100% DOD cycling on every cycle and storage at full state of charge. If the reader reviews the literature data in Figures 1 and 2, the reader will realize that the lifetime will be much better in situations where the DOD is limited and in situations where cells are stored at lower states of charge. Admittedly, the projections in Figure 23 use the incredibly simple model described by equation 1. It is our opinion that more sophisticated models will lead to even longer lifetime projections.

falcolas2• 7 years ago

Nothing here indicates a discharge rate. I'm having trouble finding the actual wattage produced by these cells over the 6 hour discharge time. Also, 6 hours is a longer than the current Tesla will run full charge at highway speeds (350 mile range, 75 mph ~= 4.5 hours). Plus, the current tesla can be re-charged and continue to operate, conditions I don't see mentioned.

If it is similar to existing batteries though... great!

Reason0772• 7 years ago

"Nothing here indicates a discharge rate."

In battery terminology, and this paper, charge/discharge rate is referred to as "C". Charging or discharging at 1C means a battery would be fully charged/discharged in 1 hour. A 100 Wh cell discharging at 1C would produce 100 W. Likewise, at 3C it would take 20 minutes and produce 300 W. At C/3, it would take 3 hours.

Osiris302• 7 years ago

Would this be a correct summary of the paper?

1) eliminates Cobalt

2) improves energy density (more range)

3) faster charge rate

4) less SOC (state of charge) swing, meaning less cost & weight

4) 5x longer lifespan

alanh• 7 years ago

what is 'state of charge swing'?

hwillis10• 7 years ago

1) 532 is not a new chemistry. It's not even the lowest cobalt chemistry; there's 511 and 811 beyond that (811, or 80% nickel, 10% manganese, and 10% cobalt, is not yet technically feasible).

2) Range is decreased. 532 has lower energy density.

3) Maybe..? But nothing truly significant. The additives in this chemistry mean a lot of heat gets generated during charge/discharge. Heat is the limiting factor during charge, so swapping directly to this would mean lower top charging speed. However with better cooling... maybe? It can sustain high charge current for longer, but not as high overall.

4) The buffer for EVs is ~10%. It's not a big factor in cost or weight.

5) Yes, 5x longer lifespan. The caveat is that they havent actually tested >3 years, but IMO this is probably a 50 year battery. The paper is more conservative and gives it 20 years.

This will be pretty enormous for grid storage, since a battery that lasts 5x as long costs 5x less. It may be important for trucks and buses, since they are much higher mileage and will want to run at 100% depth of discharge. Cars on the other hand run closer to 20%, where normal batteries will also last for many thousands of cycles.

Osiris302• 7 years ago

Short thread commentary on the paper by Ric Fulop (co-founder of A123 battery co. and current CEO/founder of Desktop Metal): https://twitter.com/ricfulop/status/1170527207657263104

manmal4• 7 years ago

This is very exciting. BEVs that were made before 2016 (I think that was when the last breakthrough trickled down to BEVs in the market) typically had a battery lifetime of about 70k miles or maybe a bit more before reaching 70% capacity. Fast-charging regularly might reduce capacity by several percent points within a few hot summer months. Basically, you bought a car for €30k and it might become useless 5-7 years later. Within a couple of years, we have gone from "I don't know, it's a bit risky to buy" to "My grandkids might be able to use this battery in their static storage system in a few decades".

MrRadar• 7 years ago

I think that was mainly due to early EVs having low battery capacities in general (because batteries cost a lot more then) so daily commuting would cycle them much more deeply than current high-capacity models. This 2017 article[1] shows battery life data for hundreds of Tesla Model S's (which have always come with large-capacity batteries). One of the highest-mileage cars in their data pool had just over 220k km (137k mi) on its battery with 90% of the original capacity remaining. There are also a few outliers with under 100k km that are below 90% capacity, I suspect those are cars that have been subject to a lot of long drives and supercharging.

[1] https://www.greencarreports.com/news/1110149_tesla-model-s-b...

manmal4• 7 years ago

The difference in lifespan that I mentioned occurred in packs of similar sizes, for all manufacturers but Tesla. Tesla is in its league when it comes to battery lifespan.

NotSammyHagar• 7 years ago

My 2012 tesla model s didn't seem to follow this. At 5 years old it had barely less capacity than original capacity. Est range went from 275 to 265. If you follow teslas, there are many (100s) of people with great than 100k of miles. The battery heating and cooling system was crucial. The leafs do suffer from serious problems. I wasn't aware they weren't charging up the battery. When speaking of evs, you really have to discuss two worlds, tesla and everyone else.

manmal4• 7 years ago

Indeed, Tesla has always had a superior BMS, and I keep hearing that they have a custom electrolyte or something that Panasonic uses for their cells? They seem to get 2-3x the lifetime out of similar cells. The Leafs don't have liquid cooling, and that's the only reason why I won't buy one - it's both bad for battery life, and fast-charging (Leaf owners call it Rapidgate).

ianai3• 7 years ago

Can anyone tell whether this is a solid state electrolyte? I’m still hooked on what I saw in that Netflix documentary where the guy was able to cut and puncture the battery without combustion. Also seemed capable of being much more dense.

hwillis10• 7 years ago

It is not. This is a fairly conventional cell, with a different crystal structure for one of the ingredients, and a tweaked mix of liquid electrolytes.

The most important part of this paper is that they actually prove that optimizing for cycle and calendar life can have such incredible effects.

Zenst• 7 years ago

Will we end up with seasonal battery's? Many already accustomed to switching tiers in winter. Whilst temperature management of batteries would cover many aspects, some locations a better tuned battery would make it more efficient overall perhaps.

EDIT [s/riers/tires/;sleep]

surewhynat• 7 years ago

which documentary was that?

ianai3• 7 years ago

Nova search for the super battery

In classic Netflix fashion, it’s been removed. Worth the 2$ to watch it though.

This is the company making solid state battery tech: https://ionicmaterials.com/about-us/

guerby• 7 years ago

Two things missed by the electrek piece:

- The article is CC-BY and downloable without scribd horrible UI: http://jes.ecsdl.org/content/166/13/A3031 PDF direct http://jes.ecsdl.org/content/166/13/A3031.full.pdf HTML direct http://jes.ecsdl.org/content/166/13/A3031.full.html

- In the conclusion the authors mention: Full details of these cells including electrode compositions, electrode loadings, electrolyte compositions, additives used, etc. have been provided in contrast to literature reports using commercial cells. This has been done so that others can re-create these cells and use them as benchmarks for their own R+D efforts be they in the spaces of Li-ion cells or “beyond Li-ion cells”.

So it's science as it should be!

nwallin• 7 years ago

Full details of these cells including electrode compositions, electrode loadings, electrolyte compositions, additives used, etc. have been provided in contrast to literature reports using commercial cells. This has been done so that others can re-create these cells and use them as benchmarks for their own R+D efforts be they in the spaces of Li-ion cells or “beyond Li-ion cells”.

That's incredible.

Normally I walk into these wunderbattery threads expecting to heap reproducibility scorn (we're still waiting for solid state glass batteries, and will in all likelihood be waiting forever) but pathways to reproducibility aren't something I'm used to in battery vaporware.

I'm actually excited for battery tech, which seems weird given all the disappointment over the previous decade or two. Feels like being excited for another cancer cure in mice.

Animats• 7 years ago

OK, so if this is so great, when does it appear in the iPhone, which always seems to have battery life problems?

std_throwaway• 7 years ago

It would make the phone thicker which is not going to happen.

thekyle• 7 years ago

How much thicker?

jokoon• 7 years ago

I'm sure batteries would be immensely better if the military would spend money into research towards better batteries.

crispinb• 7 years ago

Militaries don't have magic, they just tend to have lots of money. Foreigners' deaths are profitable for capitalist sociopaths, and good for election prospects of political sociopaths.

Now, divert government money from death industries to battery research, and we'd be talking.

hwillis10• 7 years ago

As someone who works in gov contracting, please no. Spend money literally any other way. Give it to the DOE. Fuck, just signing Musk a check would probably be better. The DoD is singularly ineffective.

coin• 7 years ago

last 1M miles

It sort of depends what type of vehicle is attached to it

elihu• 7 years ago

Having a cell that tolerates 100% discharge seems like a big deal, if it has similar energy density to other modern cells. A lot of lithium-ion cells shouldn't be discharge to less than 20% capacity or so, as it damages the cell.

Tesla battery researcher unveils new cell that could last 1M miles

Related Stories