Hacker News

Researchers achieved 1,270 Wh/L in an anode-free lithium metal battery

72 points by giuliomagnifico ago | 32 comments

foobarian |next [-]

For comparison gasoline has about 9000 Wh/L of raw chemical energy, of which maybe 30-40% gets converted to useful work.

culopatin |root |parent [-]

Almost half way there

cogman10 |next |previous [-]

> the battery retained 81.9% of its initial capacity after 100 cycles

That's really terrible.

It's interesting, but 20% loss after 100 cycles is just not great. NMC gets that at near 1000 cycles. LFP gets that at near 5000 cycles.

flerchin |root |parent |next [-]

Seemingly adequate for certain drone applications like in Ukraine. They may only need a couple charge cycles, and 4x the capacity is huge.

oofbaroomf |root |parent |next |previous [-]

20% loss isn't too bad if you start out at double the capacity though.

cousinbryce |root |parent |next [-]

My first thought was put the new cells in aircraft, then cheap cars finally grid storage

cogman10 |root |parent [-]

That actually could make sense especially with a good recycling program. Swap the packs every flight and recycle anything that falls below standards.

ryukoposting |root |parent [-]

A good recycling program sounds like a tall order. I'm seeing Silver nanoparticles (heavy metal) and multiple things that react violently with water.

I'm always skeptical of any idea that ends with a bespoke industrial-scale recycling process. People tend to massively underestimate the complexity of recycling, especially at scale.

kazinator |root |parent |next |previous [-]

That's only a valid concept in some embedded engineering case, where a certain capacity is required, and double that amount is provisioned to account for degradation.

Few consumers think this way. Something doesn't have double the capacity that it has; the capacity is the capacity, and the decline looks bad.

ryukoposting |root |parent [-]

The whole idea of the embedded part is that you make the degredation invisible to the consumer for as long as possible. From the factory, only charge up to ~4.07 Volts or thereabouts. Every N cycles, add 0.01 V to the threshold. Your phone probably already does something like this.

But yeah, 20% degredation in 100 cycles is atrocious. No amount of firmware shenanigans will be able to paper over that, not in any regular consumer product at least.

I can still think of use cases, though. Reserve power sources that aren't meant to be cycled daily, where smallness is valuable. Those little car jumper packs, for example. If there was a UPS close to the size of a regular power strip, I'd buy a few.

hinkley |root |parent [-]

Engineering is compromise though. If you can make a hybrid that loses 5% at 100 but still retains 500wh/l you’re in good shape.

There was someone working on a membrane a while back that’s pretty good at diffusing the lithium transfer in a way that reduces dendrite formation substantially, for instance. That’ll drop your volumetric advantage and likely your max discharge and charge rate a bit but would fix a lot of other problems in the bargain.

I’m not saying that the solution, but there is a palette of tools you can mix and match and that may be one of them.

dyauspitr |root |parent |next |previous [-]

But does it keep dropping? Is it 60% at 200 cycles

hinkley |root |parent [-]

I saw a video on the CATL sodium batteries the other day and the deal is that they’ve found a way to reinforce the material in a way that brings up the slope of the back half of the discharge curve so it’s almost as good as lithium down to about 20% state of charge before falling off the cliff. Lithium is more like 10% but that’s something you can manage with charge circuitry and overprovisioning.

So yeah I’d like to know the answer to your question too.

maximus-decimus |root |parent |previous [-]

Going down 10 times faster seems like a really bad trade off for 2 times the capacity. That means your battery will only lst 1/5 or 20% as long.

Reason077 |root |parent |next [-]

Not if your application requires 2X the energy. Aircraft, drones, etc. There's always trade-offs in battery design. As an old saying goes: you can have high specific energy, low degradation, or low cost... pick two!

TrainedMonkey |root |parent |previous [-]

Charge cycle capacity drops are generally not linear. If we start with 2x capacity and drop to 1.6x after 100 cycles, then we might end up with 1.2x after 1000 cycles. Some smartphone manufacturers would love that as you start with extremely superior energy density and then have a built-in obsolescence.

atomicthumbs |root |parent |next |previous [-]

And how much commercial development have NMC and LFP batteries had since they left the laboratory?

gamblor956 |root |parent [-]

LFP batteries are currently being used in newer EVs, most larger power banks, and in newer high-end phones and laptops.

ok_dad |root |parent |next |previous [-]

I think this would be perfect for race cars. We might be getting closer to a serious EV endurance series.

u8080 |root |parent |next |previous [-]

Perfect for kamikaze drones probably

marcosdumay |root |parent [-]

You can get a lot of energy density by making the batter non-rechargeable.

Reason077 |root |parent |next |previous [-]

> "That's really terrible."

Not really. At 1270 Wh/L, even with 20% degradation, these cells still retain far more energy than a LFP cell (which are more like 350 Wh/L).

The question is, what happens at 200, 500, 1000 cycles? Does the degradation continue linearly or does it slow down? ... or accelerate?

zoeysmithe |root |parent [-]

No one knows, the paper just focused on 100 cycles, but it suggests that if its good at 100 it probably is not terrible at further cycles. I guess we'll have to wait for the next paper but the conclusion seems optimistic about future research:

It is important to note that additional improvements in practical cell parameters, such as further optimized electrolyte (E/C ratio), increased stack pressure, optimized separator selection, and higher areal capacity of cathodes, can potentially enhance both the energy density and cycling performance beyond laboratory-scale demonstrations.

Post-mortem analyses confirmed reduced Li accumulation, minimized swelling, and suppressed cathode degradation, validating the robust interfacial stability of the system. By concurrently addressing the reversibility of Li metal and the structural stability of Ni-rich layered cathodes, this synergistic design offers a scalable and manufacturable pathway toward high-energy, long-life anode-free LMBs.

mschuster91 |root |parent |next |previous [-]

> It's interesting, but 20% loss after 100 cycles is just not great. NMC gets that at near 1000 cycles. LFP gets that at near 5000 cycles.

NMC and LFP had similar issues when these chemistries were at laboratory scale. Give it time and the issues will be solved.

cyberax |root |parent |previous [-]

A car with this battery can easily have a 1000-mile range (a real one, not EPA). So 100 full cycles would still mean 100k miles!

cellular |next |previous [-]

Can the liquid be agitated to avoid dendrite growth?

cyberax |root |parent [-]

The schematic images are misleading. In reality, the separation between electrodes is usually on the scale of 1mm at most.

cellular |root |parent [-]

Ultrasonic agitation? Or vibration?

cyberax |root |parent [-]

A battery cell is a long thin ribbon that is rolled into a spiral shape. There's no way you can apply any mechanical agitation to all the layers. It's been tried, but nothing came out of it.

cramcgrab |previous [-]

That’s fine but it’s only for the first bunch of cycles, after that it’s way worse than standard lion batteries.

binsquare |root |parent [-]

Things get better as the technology gets more mature. It's a promising start for sure.