A novel solid, rechargeable air battery(waseda.jp)
waseda.jp
A novel solid, rechargeable air battery
https://www.waseda.jp/top/en/news/78001
53 コメント
It seems there's news of a battery breakthrough every week. I've learned to temper expectations, because so many "breakthroughs" turn out to be dead ends. Because it's not enough for a battery to be incredibly light, or made of abundant materials, or last for ten thousand cycles. It needs to be good any many things and at least okay at most things.
E.g.—
• How much capacity per dollar?
• How much capacity per kilogram?
• How much capacity per litre?
• How quickly can it be charged?
• How quickly can it be discharged?
• How much energy is lost between charging and discharging?
• How predisposed is it to catching fire?
• How available are the materials needed to manufacture it?
• How available are the tools/skills required to manufacture it?
• How resilient is it to mechanical stress, e.g. vibration?
• How much does performance degrade per cycle?
• How much does performance degrade when stored at a high state of charge?
• How much does performance degrade when stored at a low state of charge?
• How much does performance drop at high temperatures?
• How much does performance drop at low temperatures?
• How well can it be recycled at end-of-life?
A sufficiently bad answer for any one of these could utterly exclude it from contention in many usages. A battery which scores well on everything except mechanical resilience is a non-starter for EVs, for example. On the other hand, a battery that's no good for EVs might be great for stationary storage.
I'm only a layperson and this list is the result of just a few minutes thought. I'm sure someone with more familiarity with the subject could double this list. But the point is, when you daydream about some hypothetical future battery tech, you need to appreciate just how well today's lithium chemistries score in so many areas.
E.g.—
• How much capacity per dollar?
• How much capacity per kilogram?
• How much capacity per litre?
• How quickly can it be charged?
• How quickly can it be discharged?
• How much energy is lost between charging and discharging?
• How predisposed is it to catching fire?
• How available are the materials needed to manufacture it?
• How available are the tools/skills required to manufacture it?
• How resilient is it to mechanical stress, e.g. vibration?
• How much does performance degrade per cycle?
• How much does performance degrade when stored at a high state of charge?
• How much does performance degrade when stored at a low state of charge?
• How much does performance drop at high temperatures?
• How much does performance drop at low temperatures?
• How well can it be recycled at end-of-life?
A sufficiently bad answer for any one of these could utterly exclude it from contention in many usages. A battery which scores well on everything except mechanical resilience is a non-starter for EVs, for example. On the other hand, a battery that's no good for EVs might be great for stationary storage.
I'm only a layperson and this list is the result of just a few minutes thought. I'm sure someone with more familiarity with the subject could double this list. But the point is, when you daydream about some hypothetical future battery tech, you need to appreciate just how well today's lithium chemistries score in so many areas.
And the problem with the article is it doesn't discuss virtually any of those. How are we publishing papers on batteries in 2023 without making these things obvious in the abstract?
During the first 5 seconds of reading a publication on batteries I should be able to answer at least 1 or 2 of those questions.
During the first 5 seconds of reading a publication on batteries I should be able to answer at least 1 or 2 of those questions.
Eh? Did you read it?
Note the article is a summary of the research, not the research itself.
Note the article is a summary of the research, not the research itself.
yeah, I'm talking about the summary, not the research. The summary really should cover that, in very obvious ways. Otherwise I'll assume the authors/publishers want to hide their bad results. You kinda have to brag about them.
Eh? Which of those did it talk about?
And in the first paragraph the PR article should mention all these numbers and how they compare to the existing batteries the "breakthrough" is purporting to replace
Do some standardized tests exist ? I know there are some for some neural net problems.
There's probably a Nobel prize in it for anyone who can significantly improve on lithium ion. It's a tall order. Many have tried, many have failed. Eventually someone might succeed.
If someone managed to improve on it in every metric then yes, absolutely.
The thing is though that you don't have to beat it on every metric to be useful, you have to be _good enough_ on every metric and then really good on one or two.
For example, it's absolutely fine for a battery to weight five times as much if it doesn't need to be portable. Five times as heavy but half the cost per MWh? That's great for grid storage!
Twice as expensive and non rechargeable but 4 times the capacity (per unit volume)? Great for pacemakers!
It's all about finding the right niche.
The thing is though that you don't have to beat it on every metric to be useful, you have to be _good enough_ on every metric and then really good on one or two.
For example, it's absolutely fine for a battery to weight five times as much if it doesn't need to be portable. Five times as heavy but half the cost per MWh? That's great for grid storage!
Twice as expensive and non rechargeable but 4 times the capacity (per unit volume)? Great for pacemakers!
It's all about finding the right niche.
> Five times as heavy but half the cost per MWh? That's great for grid storage!
Your point stands in general, but this is possibly worst example of winning in a single metric. New technologies are almost always more expensive than old technologies to begin with, so they need to provide some non-cost advantage at the outset.
At least for chemical batteriers, cost and weight are going to be correlated once the technology is mature. At sufficient scale and maturity of process, the material inputs become a significant fraction of the cost.
Note also that less mature processes are likely to be much more expensive than their inputs.
This means that for such a theoretical battery to exist, the inputs would need to be 5x cheaper and it would need to rely on an a pre-existing manufacturing process.
Lithium-Ion became dominant not by being the cheapest, but by having an incredible energy density. Since there was built-in demand, there was a lot of interest in making them cheaper.
Your point stands in general, but this is possibly worst example of winning in a single metric. New technologies are almost always more expensive than old technologies to begin with, so they need to provide some non-cost advantage at the outset.
At least for chemical batteriers, cost and weight are going to be correlated once the technology is mature. At sufficient scale and maturity of process, the material inputs become a significant fraction of the cost.
Note also that less mature processes are likely to be much more expensive than their inputs.
This means that for such a theoretical battery to exist, the inputs would need to be 5x cheaper and it would need to rely on an a pre-existing manufacturing process.
Lithium-Ion became dominant not by being the cheapest, but by having an incredible energy density. Since there was built-in demand, there was a lot of interest in making them cheaper.
Even if a technology is _nearly_ as cheap, a company selling them might be willing to eat the initial higher prices and make a loss for a time until the learning curve puts them at a bigger advantage.
The examples I was thinking of with this particular metric were Ambris liquid metal cells: https://ambri.com/solution/ and iron-air batteries: https://formenergy.com/technology/battery-technology/
Both of these technologies have additional advantages on top of cost that I didn't bother going into: namely they degrade less with use and both have no risk of thermal runaway.
The examples I was thinking of with this particular metric were Ambris liquid metal cells: https://ambri.com/solution/ and iron-air batteries: https://formenergy.com/technology/battery-technology/
Both of these technologies have additional advantages on top of cost that I didn't bother going into: namely they degrade less with use and both have no risk of thermal runaway.
Degrading less with use is a big advantage, and effectively is like making it cheaper.
The parent is probably thinking of Sodium Ion batteries, which indeed are cheaper, but less energy dense, than li-ion.
So you're saying that capitalism leads us to inevitable stagnation with the least desirable of the acceptable options?
Because that sure how that reads....
Because that sure how that reads....
I suppose so. Looking at the list, a lot of them can be relaxed for grid storage. They can be run at absolutely any temperature, for instance.
On the other hand an electric aircraft battery needs 10/10 on pretty much all of them (except perhaps cost - up to a point).
On the other hand an electric aircraft battery needs 10/10 on pretty much all of them (except perhaps cost - up to a point).
Grid batteries are exposed to the full range of temperature swings of whatever climate they're installed in. Protecting a battery from the local climate is expensive and power intensive. Not a deal-breaker, but this cost needs to be factored into the cost of the battery, and any active cooling/heating needs to be deducted from its efficiency.
Temperature tolerance means the battery is more flexible and has lower installation costs - lithium batteries should generally be stored in somewhere fireproof, for example.
>Twice as expensive and non rechargeable but 4 times the capacity (per unit volume)? Great for pacemakers!
And bombs. When it comes to energy density, be careful what you wish for.
And bombs. When it comes to energy density, be careful what you wish for.
Well if someone manages to make a battery with the energy density and cost of TNT you let me know.
It doesn't have to be that powerful. An incredibly energy dense, single use battery would be a boon for suicide drones. The explosives can be separate.
My point is that you can already make suicide drones with existing technology, explosives aren't expensive or even particularly hard to make and the batteries are already good enough to get you out of transmitter range.
A more powerful battery isn't going to enable something like this, it can already be done.
A more powerful battery isn't going to enable something like this, it can already be done.
I doubt it will happen this decade. lithium ion is Goodenough.
I know you saw articles about the inventor dying and his name being 'Goodenough' and you want to make puns, but regular lithium ion has big problems with durability. Lithium iron phosphate, lithium titanate and sodium ion batteries all improve durability but have different downsides as well.
Nominative determinism has a pretty good track record, I wonder if any investors use it.
Although it's tricky - I'd never have thought someone called "Goodenough" would invent something so revolutionary.
Although it's tricky - I'd never have thought someone called "Goodenough" would invent something so revolutionary.
Nominative determinism has a pretty good track record, I wonder if any investors use it.
That's like saying that horror scopes and phrenology have good track records.
That's like saying that horror scopes and phrenology have good track records.
Month of birth does have a little bit of predictive power, there are statistical differences between people born in winter vs. summer (and possibly the effect of being older/younger than their classmates in formative school years). Unusual things can be correlated. Also, I was not being entirely serious.
Also, I was not being entirely serious.
Then why are you trying to rationalize it as being true?
Month of birth does have a little bit of predictive power
Not for what horror scopes are
Then why are you trying to rationalize it as being true?
Month of birth does have a little bit of predictive power
Not for what horror scopes are
A lot of the capacity we'll end up needing will be stationary in nature, where weight isn't nearly as important as energy density, stability, charge/discharge rates, etc. Such installations can be climate-controlled using energy stored in those batteries. Some of the more creative designs end up making about half of the questions you listed more or less irrelevant so that we can focus exclusively on electrical performance.
But the mobile/car devices space is the gatekeeper. Nobody produces batteries without extensive testing and extensive testing happens in fast lifed industries. Same why powerwalls were not a thing, before laptops phones pushed liIon-Batterys through this barrier.
To me it seems that established power generation is the gatekeeper -- with on-demand generation and the entire set of infrastructure behind it, the economic pressures aren't there for stationary storage. Mobile storage was incentivized for other reasons that are divorced from stationary needs.
Iron-air is going directly to grid storage, there's a commercial facility being built right now.
Is there a standard flowchart of tests for battery researchers, or those evaluating novel battery chemistries? For example, first test energy density, then BOM without economies of scale, et cetera.
I would love to see a grid of all these values for all the latest battery technologies, to see what's leading the field between solid, sodium, aluminium, Sulfur Selenium, iron, etc
> While the discharge capacity of SSAB-PDBM reduced to 44% after 30 cycles, by increasing the proton-conductive polymer content of the negative electrode, the researchers could significantly improve it to 78%.
That's still a 22% drop in capacity after only 30 charge cycles; this system has a long way to go before becoming a practical battery.
That's still a 22% drop in capacity after only 30 charge cycles; this system has a long way to go before becoming a practical battery.
While probably not ideal for a phone or car, it could still have its uses. So long as it stabilizes at that 70%+ region, that is not potentially a deal-braker for grid installations. Which is an optimization of installation cost, maintenance, longevity, operating temperature/pressure requirements, charge/discharge rates, etc.
They didn't say it stabilizes, though. You'd think they'd mention that if it did.
If it stabilized then it would just be 30% heavier. I doubt it stabilizes.
Depends one the cost and environmental impact. I don't know anything about the compounds they used and anyway they're still experimental and could be replaced by others. But in stationary settings where size isn't an issue they can lose substantial capacity upfront and still be viable if they can be cheaply produced and don't involve complicated steps like in lithium mining for example.
Flow batteries are better suited for stationary applications, I think. Liquids lend themselves to large-scale processing.
Flow batteries are pretty good but I think if the dirt cheap "heavy" chemistries such as liquid metal and iron-air take off then not needing any moving parts and having no risk of leaks is an advantage.
Just for Context thats roughly an order of magnitude worse than current lithium ion and almost 2 orders of magnitude worse than lithium iron phosphate.
> the coulombic efficiency of SSAB-PDBM was 84% at 4 C rate, which gradually decreased to 66% at 101 C rate.
Lithium batteries are over 95% efficient, so a long way indeed. Since the battery contains water, I assume it doesn't work below 0 C.
Lithium batteries are over 95% efficient, so a long way indeed. Since the battery contains water, I assume it doesn't work below 0 C.
I don't think 4 C and 101 C are referring to temperature. In battery charging/discharging 1 C is the current you would get if you could discharge a fully charged battery in 1 hour at a constant current.
If a battery's stated capacity is say 2000 mAh then 1 C for that battery is 2000 mA.
Recommended discharge and charge (if applicable) rates are usually specified in terms of C. For example the usual recommendations for charging Panasonic Eneloop batteries is 0.5 - 1 C. So if you are charging an AA which has about 2000 mAh capacity that would be 1000 - 2000 mA, but if you were charging an AAA which has something like 900 mAh capacity that would be 450 - 900 mA.
If a battery's stated capacity is say 2000 mAh then 1 C for that battery is 2000 mA.
Recommended discharge and charge (if applicable) rates are usually specified in terms of C. For example the usual recommendations for charging Panasonic Eneloop batteries is 0.5 - 1 C. So if you are charging an AA which has about 2000 mAh capacity that would be 1000 - 2000 mA, but if you were charging an AAA which has something like 900 mAh capacity that would be 450 - 900 mA.
I really enjoy watching Cayrex building interesting batteries on YouTube, like the Reusable Aluminum Air Battery, https://youtu.be/XxJ2FeUi6oI
Meanwhile, lithium-ion is improving remarkably each year.
https://insideevs.com/news/581729/volumetric-energy-density-...
https://insideevs.com/news/581729/volumetric-energy-density-...
That graph is plainly wrong. The initial point is at the same level as a lead-acid battery. There is no chance it is legit.
They wouldn't know since they stole it from the DOE website here:
https://www.energy.gov/eere/vehicles/articles/fotw-1234-apri....
(The cited paper does not contain this graph, see https://www.osti.gov/pages/servlets/purl/1842609)
(The cited paper does not contain this graph, see https://www.osti.gov/pages/servlets/purl/1842609)
That graph must be a misunderstanding by the DOE then.
That graph is a bit deceptive; the X axis is not linear, but instead shows years 2008, 2010, 2013, 2017 and 2020.
It's wrong anyway. Even back in 2008, it was more than 50 watt hours per liter.
The graph is taken straight from the DOE website, here:
https://www.energy.gov/eere/vehicles/articles/fotw-1234-apri....
Without attribution of course. They just re cited the paper the DOE cites.
The paper has no such graph:
https://www.osti.gov/pages/servlets/purl/1842609
The graph is clearly wrong.
The graph is taken straight from the DOE website, here:
https://www.energy.gov/eere/vehicles/articles/fotw-1234-apri....
Without attribution of course. They just re cited the paper the DOE cites.
The paper has no such graph:
https://www.osti.gov/pages/servlets/purl/1842609
The graph is clearly wrong.
I'd file this under gimmick rather than novel tech.