Liquid-metal battery holy grail

October 24, 2012 8:35 PM

A few nights ago MIT scientist Donald Sadoway (Time 100) was on Colbert Report to discuss a new cheap high-capacity liquid-metal battery that could be the holy grail for solar and wind power to store electricity for on-demand use. The Colbert show is an intro but sort of dumb (by design), the TED talk gives some more detail, or the company page Ambri has more info.
posted by stbalbach (31 comments total)

Is there anything about this new battery technology from someone who isn't trying to sell it? A cheap high-capacity battery might be a really big deal, but we've heard this lots of times before and it hasn't quite panned out. I would love to know more about this but would prefer to see a primer on the technology from someone who doesn't have an interest in promoting it. (And ideally from a source that isn't a video, for that matter.)
posted by Scientist at 8:38 PM on October 24, 2012 [10 favorites]

Believe the Scientist here, whenever someone tells you about a miracle new battery just ask what the UPC code is, until it has one consider it vaporware.
posted by Cosine at 9:03 PM on October 24, 2012 [6 favorites]

high-capacity liquid-metal battery

Can it form knives and stabbing weapons?
posted by Strange Interlude at 9:07 PM on October 24, 2012 [7 favorites]

Their careers page lists openings for what looks like an entire engineering team, including a "Director of Cell Development." Sounds like they're a little early in the development process to be announcing a holy grail.
posted by contraption at 9:08 PM on October 24, 2012

Oh, hey, he taught me freshman chemistry (3.091). He did, indeed, love the colored chalk. He looks pretty much the same as he did 10 years ago.
posted by maryr at 9:10 PM on October 24, 2012 [2 favorites]

This is the best disinterested source I could find: TR10: Liquid Battery. It looks like by "liquid metal" they actually mean "molten metal" and these batteries are intended for large-scale power storage at wind and especially solar power plants, not for flashlights and laptops and such, or even for cars.

Cheaper output-smoothing at solar plants would be a nice win, but probably not revolutionary all on its own. No doubt it would make centralized solar power generation a bit cheaper and expand the range of suitable sites somewhat, but my layman's understanding is that the expense of solar energy storage is not the main thing that is preventing us from throwing off the shackles of the petroleum industry, slaying the dragon of global climate change, and surging forth into a bright new era of unlimited free power. It would be a nice technology that would potentially be very valuable, but not necessarily a real game-changer.

The batteries seem to rely on floating reactive molten metals of different densities on top of each other, a bit like a layered shot of white-hot explosive goo. This hellish arrangement of fiery metal is not something that you would want to sell at the drug store or sit on top of as you ride along on the highway. It is something that you would want to sequester in a special secure building at the site of a large power plant with its own dedicated staff of engineers. Unless I am quite mistaken in my admittedly sketchy understanding of what is going on here, this is not a general-purpose battery but rather a battery that might have use in certain specific, highly-specialized, high-dollar applications.

Cheap rechargeable batteries with high energy densities would indeed be revolutionary. This, though, is not that. It might turn out to be a good invention (nobody has made these things outside of a lab yet, and let me tell you there are a lot of things that work in labs which will probably never be seen anywhere else, for lots of reasons) but who knows? This guy is saying that they will be 1/3 the price of current technology, but how can he know that if he's only made a single prototype and has no idea if it's even possible to make them on an industrial scale?

Sorry folks, this is not the battery holy grail we are looking for.
posted by Scientist at 9:32 PM on October 24, 2012 [17 favorites]

So, you're saying that the OP... chose poorly?
posted by maryr at 9:34 PM on October 24, 2012 [9 favorites]

stbalback I did some looking after seeing him on Colbert as well. What I couldn't find in an evening's research was decent comparisons to existing storage tech such as flywheel and liquid sodium-sulfur batteries. Also the explanations (or lack) of how the batteries are kept at the high operating temperature wasn't very satisfying. I hope it is not a dead end.
posted by Mei's lost sandal at 9:36 PM on October 24, 2012

Oh, and apparently we already use similar batteries to balance out wind and solar power plants, see Wikipedia: Molten salt battery. The difference here is that the new battery uses molten metals for the electrodes as well as the electrolyte, which apparently is potentially advantageous from the standpoints of manufacturing and power output. So it's not even particularly revolutionary in terms of power plant output balancing, but rather is a possible incremental improvement that may reduce costs.
posted by Scientist at 9:36 PM on October 24, 2012 [1 favorite]

US electric consumption averages a bit under 500 GW/hr. Let's say you want to serve half that with wind and solar, and maybe optimistically need to store only a quarter of that to make it feasible. Can you economically get 60 gigawatts of battery capacity online in the next 30 years? Mind you that's almost as much as there is nuclear power. That would be the holy grail.
posted by helot at 9:57 PM on October 24, 2012

If it's really relying on molten metal, then you have an insulation issue, and energy in the form of heat is going to bleed off. That will reduce the efficiency of the system.
posted by Chocolate Pickle at 10:18 PM on October 24, 2012

According to the article that Scientist linked to, these batteries use magnesium and sodium sulfide. If you're talking about a battery capable of storing energy in the hundred megawatt-hour range, then it's going to be big. It's also going to be heavy.

And you have to pour enough energy into it to melt all the metal before you can even start using it. You don't get that back.

These things come and these things go. Usually this kind of extravagant claim means someone is looking for investors.
posted by Chocolate Pickle at 10:50 PM on October 24, 2012

He looks pretty much the same as he did 10 years ago.

Ha... I just came in to say my wife and I both think he looks pretty much the same as he did 20 years ago... there must be a painting of a withering old Don Sadoway hanging in a lab somewhere.
posted by rodeoclown at 11:04 PM on October 24, 2012 [6 favorites]

There's an awful lot of research being done on batteries and super-capacitors for use by us mortals (in cars and such) these days.

Certainly most of this research will end with nothing, but it only takes one manufacturable breakthrough to change everything. Or maybe it will be a lot of small steps based on the various research ideas that add up to long term change.

In either case, I'm optimistic.
posted by eye of newt at 11:29 PM on October 24, 2012 [1 favorite]

Scientist: "Oh, and apparently we already use similar batteries to balance out wind and solar power plants, see Wikipedia: Molten salt battery"

I believe the versions of that technology actually in use actually use the salt for heat storage. The salt is merely the heat transport medium, like in a molten salt nuclear reactor. The heat gets transferred in a heat exchanger to water, turning it to steam and using a bog standard steam turbine to generate electricity.

In this case, it's actually an electrolytic battery, just of an odd composition and that runs very, very hot. Presuming the existing models work as advertised, it seems like quite the breakthrough.
posted by wierdo at 11:55 PM on October 24, 2012

Sorry folks, this is not the battery holy grail we are looking for.

Aw why you science guys gotta go ruining this for us?? Stupid reality. :kicks rocks:
posted by LordSludge at 1:15 AM on October 25, 2012 [1 favorite]

The only efficiency number I've seen for this is the "up to 69%" for the prototype quoted in the paper linked from the Ambri website. That's pretty typical for a rechargeable battery, and it seems to me that the 30% energy lost during the charge/discharge cycle should, if retained via insulation instead of deliberately dissipated as is typical for low-temperature batteries, should be plenty to keep them operating at just about any temperature required. If used for peak levelling, the batteries would presumably be running charge/discharge cycles at least daily, so I don't imagine they'd need to be be brought up from dead cold often enough to make that a major source of energy loss.

It also seems to me that "molten magnesium" and "safe" are not natural companions, and I will be interested to see how trauma-resistant their commercial units prove to be.

The only other battery technology I'm aware of that makes similar claims for size, scalability and efficiency is the vanadium redox flow battery. Given Sadoway's stated methodology (designing to a price point rather than hoping to reduce an initially expensive chemistry's deployment cost via mass manufacture) I will be interested to see whether Ambri does indeed manage to beat those on cost.

Isentropic is worth keeping an eye on as well.
posted by flabdablet at 1:19 AM on October 25, 2012 [1 favorite]

Don't be all frowny face. This is not about electricity. It's about TED. It's the best frickin TED talk ever* -It's even in the music vid! Don't go spoiling it with objectivity and criticism. This is TED!

*apart from Hans Rosling obviously, but that goes without saying
posted by EnterTheStory at 2:02 AM on October 25, 2012

I saw this a while back, and thought it looked really promising (I know nothing about this stuff, though.). I loved the idea of starting the project by looking at the periodic table.

I'm wondering, though, if this is more of a Moller Skycar type of thing where the project continually needs just a few more investors before the final problem can be solved, production started, and the world changed.

My enthusiasm level dropped quite a bit when he mentioned he started a company. Nothing against the man and his team, but we've all seen lots of claims that later proved impractical.

If we can conquer batteries and transmission line losses, that would be sweet.
posted by InsertNiftyNameHere at 2:52 AM on October 25, 2012

I saw him on Colbert and for some reason he failed to mention that his revolutionary new battery that is going to replace everything operates at 700º C, even when they were joking about eating it or using it on electric cars. It seems quite suspicious.
posted by radiobishop at 3:06 AM on October 25, 2012

i have always thought that the key to all this wasn't storage, it was decentralization of production, allowing increased efficiency from the recovered losses of distribution and production. in the case of production losses, it's on the order of 60%+ that we lose from conversions. when that goes into and out of a battery, we lose more due to more conversion inefficiencies.

then, there's distribution losses in resistive terms (heating up high tension wires), a little leakage from reactive losses, losses in transformers (which have resistive primaries and secondaries).

if power is produced where it is consumed, and the input energy is 'free' (in solar and wind, it almost is in recurring cost terms), then a lot of this goes away. it's offset in life cycle costs by the additional infrastructure needed, but man, we waste a lot of power in unused heat. i don't think any centralized storage medium is going to help that, at all.

i hear this complex battery stuff and hit the snooze button, usually. crap that works shows up fast in the marketplace.
posted by FauxScot at 3:58 AM on October 25, 2012 [1 favorite]

I think Amory Lovins has outlined the right idea as far as grid battery storage goes. The way he sees it, if we get grid access and pricing models right then a mostly-electric car fleet could end up being an even bigger distributed energy storage resource than we need. Hundreds of millions of cars, each one with tens of kilowatt-hours of energy storage on board, is terawatt-hours of storage capacity and most of that energy shouldn't need to move very far from where it's stored.
posted by flabdablet at 4:25 AM on October 25, 2012 [2 favorites]

if power is produced where it is consumed, and the input energy is 'free' (in solar and wind, it almost is in recurring cost terms), then a lot of this goes away. it's offset in life cycle costs by the additional infrastructure needed, but man, we waste a lot of power in unused heat. i don't think any centralized storage medium is going to help that, at all.

True, but we would need de-centralised energy storage to deal with the intermittent nature of solar and wind power.

A neighbourhood may have enough solar panels to produce its own electricity, but it will need significant storage capacity to see it through an overcast week or a stormy weekend.

I've seen a figure of about four days of full demand mentioned as a realistic amount of storage to make fully renewable energy reliable. That's a pretty huge amount of storage and an inexpensive but bulky battery system is really just what you want for that. One advantage of these all liquid systems is their scalability. The capacity scales with the cube of the size (but the peak power scales only as the square because it depends on the electrode interface area) and the heat loss with the square. Plus, heat retention is a solved problem more or less - we know how to insulate things in this temperature range.

I'd like to see some order of magnitude calculations of how much this would cost per MWh, but I don't have time to do it right now. I don't even know if we have enough antimony and magnesium to do this on a really large scale.
posted by atrazine at 4:54 AM on October 25, 2012

a mostly-electric car fleet could end up being an even bigger distributed energy storage resource than we need

So if you go to use your car during peak hours, you might find the battery dead because the electric company's drained it to run your neighbor's water heater? I hope I'm misunderstanding that.
posted by echo target at 6:47 AM on October 25, 2012

Cheaper output-smoothing at solar plants would be a nice win, but probably not revolutionary all on its own. No doubt it would make centralized solar power generation a bit cheaper and expand the range of suitable sites somewhat, but my layman's understanding is that the expense of solar energy storage is not the main thing that is preventing us from throwing off the shackles of the petroleum industry, slaying the dragon of global climate change, and surging forth into a bright new era of unlimited free power. It would be a nice technology that would potentially be very valuable, but not necessarily a real game-changer.

There are only two things stopping us from implementing a renewables drive to address climate change and the decline in fossil fuels:

1. politics.
2. the sheer expense of DEPLOYING lots and lots of physical capital out there. We're talking trillions of dollars. Renewables are already good.

But that said, we don't need new tech. Germans, with their solar and wind mania, are now paying less per kwh than the French and their nukes.
posted by ocschwar at 9:13 AM on October 25, 2012

This is an emerging technology so skepticism is healthy, but over skepticism is lazy.

This is the best disinterested source I could find: TR10: Liquid Battery

..from March 2009.

he's only made a single prototype

..in March 2009.

Here is what you missed by not watching the OP video: The battery is modular 2-4 KW in size. Size of a pizza box. Sealed metal casing. Sits on a table, no "hellish molten metal" on the outside, a small amount inside. Stack as many as you want for whatever size battery needed. Store them in a shipping container, put the shipping container wherever you want. Shipping container = 200 houses energy use ie. dense

The battery is for utility grid power storage. It can store power from solar and wind for use when the sun doesn't shine or the wind doesn't blow. This means 100% green energy potential if costs for the battery are low enough. This is a 'Big Deal'. Is it the only Deal? No. But without it, green energy will cap out below 40% of total power generation (wind and solar anyway).

Present unknowns are how it will behave under real work loads and stresses which is the next phase. The company is funded by Bill Gates, Total Oil, MIT and other folks who know something about developing technology. They are not seeking investors though I imagine many are seeking to invest.
posted by stbalbach at 11:43 AM on October 25, 2012 [2 favorites]

So if you go to use your car during peak hours, you might find the battery dead because the electric company's drained it to run your neighbor's water heater? I hope I'm misunderstanding that.

I think you probably are.

For a start, most electric water heaters are storage types designed to heat the water when electricity is cheap, i.e. off-peak. Peak electricity use generally occurs in the height of summer when everybody turns on their air conditioner.

For seconds, the proportion of any given car's energy storage made available for grid use would be set by its owner via its charge controller, not by the utility company. People who live fifteen minutes from work could set their cars to maintain at least half an hour's standing charge; people who live further away could set theirs to maintain more.

The incentive to do this would come from grid- or more likely internet-transmitted pricing signals based on instantaneous demand. If your car is programmed to buy electricity when it's cheap and sell it back when it's expensive, then leaving it parked in a grid-connected bay could actually make you a little money.
posted by flabdablet at 4:44 PM on October 25, 2012

Yeah, there seem to be a lot of people who love crapping over any new advance in renewable. You need to try lots of different things to see what ends up working. And in fact, many things might work, but will never be widely used because something else will end up being better. Thin-film solar, for example, works well and ended up being pretty cheap, but the chinese figured out how to make polysilicon for about the same price.

For a long time the main research into batteries has been into small and light (and room temperature) batteries for electronic gadgets and cars. Lithium works well but it's really expensive. The real breakthrough doesn't need to be size or weight efficiency but cost.

And that's where this thing looks really good. I mean, to build this thing you just take three common, cheap metals, melt them, and put them in a bucket. You do have to figure out a way to keep them all at a really high temperature, but that shouldn't be impossible.

Cheaper output-smoothing at solar plants would be a nice win, but probably not revolutionary all on its own. No doubt it would make centralized solar power generation a bit cheaper and expand the range of suitable sites somewhat, but my layman's understanding is that the expense of solar energy storage is not the main thing that is preventing us from throwing off the shackles of the petroleum industry, slaying the dragon of global climate change, and surging forth into a bright new era of unlimited free power. It would be a nice technology that would potentially be very valuable, but not necessarily a real game-changer.

I've actually been over this a bunch of times, but there isn't actually any technical challenge in terms of using solar power to power things during the day. Germany is already doing it on a massive scale, and they don't even get that much sunlight. The challenge is just political. Since most power use is during the day anyway, that gets us most of the way there. Nuclear is one option, but that's politically difficult. Natural gas is one but you still do give off some CO₂

So the real competition with this thing isn't with standard batteries, but rather nuclear power or fossil fuels, hopefully with paid for carbon credits.
posted by delmoi at 10:43 AM on October 26, 2012

Yeah, there seem to be a lot of people who love crapping over any new advance in renewable.

No kidding. I read Scientists posts above and just scratched my head. Yet somehow it got 17 favs, there must have been something fundamentally wrong with the way I presented the OP. I'm thinking people saw "battery", thought about car battery, and then saw "molten liquid metal" and thought "no way that's going to work" and end of story. Maybe I should have emphasized this is for utility scale grid storage so that we can transition to 100% green energy, since green energy is intermittent. This is all explained in the OP links, but alas the golden rule of MeFi: most people don't read OP links.

You do have to figure out a way to keep them all at a really high temperature

Probably using energy stored in the batteries, but the idea is not to store power long before the next recharge. Oh, and these batteries can be used with power from coal and gas to help move electricity around more efficiently, even a 10 minute battery can potential for more efficient management of the grid, currently everything has to be real-time generation. If they had buffers out in the field at certain points it could result in more efficient power generation. Like data buffers in computers.
posted by stbalbach at 9:19 PM on October 26, 2012

atrazine: "but the peak power scales only as the square because it depends on the electrode interface area"

Since the electrodes are liquid, it seems like you could fairly easily increase the electrode's area with sound. Tune the enclosure properly, set up a standing wave and you've got more surface area.
posted by wierdo at 10:51 PM on October 26, 2012

Seems to me that just about any capacity:power ratio should be obtainable just by fiddling with the cell shape; wider, flatter cells -> more peak power.
posted by flabdablet at 2:15 AM on October 28, 2012