Hacker News new | ask | show | jobs
by Aurornis 721 days ago
This blog post is all over the place. The 2030 price projections are taken from extrapolations of Lithium battery costs, but he’s assuming Sodium chemistry batteries will take over and become ubiquitous at rock bottom prices. The first Sodium batteries barely became available within the past year.

He’s also treating batteries like the only component of the system. The associated charging, inverter, and physical structure components aren’t going to follow the same downward curve. Those are fixed costs on top of the battery itself.

Finally, there’s a lot of vague futurist writing mixed in, from congratulating himself on predicting in 2017 that EV trucks would be a thing some day to something about the blockchain for coordinating power grids:

> I think this is also an area where distributed ledgers with low energy requirements (so not Proof of Work but Proof of Stake) could shine by creating an ‘trustless’ system (meaning the system justs works, also if there is no ‘trusted’ party that plays the boss).

This statement doesn’t even make sense when you read it. He defines “an [sic] ‘trustless’ system” as meaning a system that “just works” which suggests to me that he doesn’t really know what he’s talking about but has been led to believe that blockchain is the future for everything.

Fun read, but I didn’t get much out of this article other than “prices are going down”

6 comments

> This blog post is all over the place.

Which is sad. He has something useful to say, but destroys his credibility by not focusing. Here's the "poster wall" of the organization he claims to head.[1] "Disciplinary convergence through creative story telling". For a much better summary of the subject, see the cover story in this week's Economist.

OK, how cheap can batteries get, really?

Well, the price of lithium dropped 80% in the last year.[2] Overproduction at the moment. Exxon has a lithium production unit, and they're expanding. New, large lithium mines under construction in Nevada, Sonora (Mexico), five new mines in Western Australia, Quebec, Zimbabwe... Plus, of course, recycling old batteries, a far more concentrated source than anything in the ground. Lithium supplies do not look like a problem. The prices do go wildly up and down because the price of raw lithium doesn't affect car sales much in the short term. That's normal behavior for minor commodities.

This also means that sodium batteries will probably be unnecessary. This is good, because of the fire risk. For fixed installations and low end car, lithium iron phosphate is cheap, not subject to thermal runaway, and in most of BYD and CATL products right now. (APS, please get with the program and start shipping small UPSs with LiPoFe batteries so those things last 10 years.)

Coming along next are solid state batteries. Huge hype, a few samples, and production cost problems.[3] Here's the manufacturing process at lab scale, at the Franuhofer Institute.[4] Works in the lab. Here it is at production test scale.[5] The IEEE consensus is that solid-state battery production technology is about 10 years behind existing lithium-ion production. With production in test everywhere from Shenzhen to Belgium to Maryland, progress is being made rapidly.

This is the kind of process that gets cheaper as it scales up.

Solid-state batteries are important because 10-minute charging is needed to increase consumer acceptance rates.

Between solar and battery technology, fossil fuels are going to be crushed. Soon.

[1] https://neonresearch.nl/poster-wall/

[2] https://www.reuters.com/markets/commodities/lithium-producer...

[3] https://spectrum.ieee.org/solid-state-battery-production-cha...

[4] https://www.youtube.com/watch?v=j5SVrp8N-1M&

[5] https://www.youtube.com/watch?v=_eZGuDaqZAE

> Well, the price of lithium dropped 80% in the last year.[2] Overproduction at the moment.

...

> This also means that sodium batteries will probably be unnecessary.

If we're overproducing this doesn't follow. Lithium prices will rise back to the price of production. I'm not an expert but quickly glancing at the futures market and it looks to me like there is only a small rebound predicted ($13.30 -> $17.00/contract over a few years, highly illiquid market so take prices with a grain of salt) so the actual story might be "lithium production has become much cheaper".

It also doesn't really matter if you're trying to estimate "it will cost at most this" by looking at sodium ion batteries. I don't think the author really cares if the batteries are sodium or lithium based, just that they don't cost more than sodium based batteries would cost.

> This is good, because of the fire risk

One of the selling points for Sodium ion has pretty consistently been that they are non-flammable. Admittedly this is a function of the electrolyte they use and not a fundamental property of sodium vs lithium, so it might change in the future, but I don't believe it has/it is in anticipated to?

Agree on every point. Sodium is also so abundant it will likely not have the same price fluctuations as lithium.

For stationary battery like the use case describe for in house. I would assume sodium has a much better chance of winning over.

Chinese manufacturing seems insanely advanced based on link 5.
That's a nice production line. It's not unusual. Many of the steps shown, and some of the machinery, are the same as in regular lithium-ion battery manufacturing. Compare with the machinery in [2].

It's not the machinery that's advanced. It's the manufacturing chemistry. The real breakthrough here is that the ceramic is deposited as a slurry, and through a series of ovens and dehydrators, it becomes the solid electrolyte. There's no high-pressure sintering press step, as there is in the Fraunhofer lab making solid state battery samples. The process is roll to roll. If that works, the production cost comes way down and the process scales up well.

The next two or three years are going to be very interesting in batteries.

[2] https://www.youtube.com/watch?v=UHZg5-uk1-k

I agree that applying anything to do with blockchain to electricity is dumb - these are already just regular markets, so an inverter/charger could already take price signals from the existing market and do whatever the homeowner wanted, with zero need for blockchain or central control at all. With smart meters (which are becoming more ubiquitous) it's already simple to incentivise using battery power in peak periods when the price is high...

But on inverter/chargers - they will absolutely will follow a downward trend. Maybe not as quickly as batteries but downward all the same. Wide-bandgap semiconductor FETs are getting cheaper and better all the time (higher current and voltage per device), and they allow for power topologies that are more efficient, so cooling gets easier, weight of heatsinks and the amount of material in those goes down, power per unit volume increases and unit mass will decrease, etc. Production volumes will also increase which should lead to economies of scale too.

I can get a 48V DC/230V AC, 8000VA Victron Multiplus 2 inverter/charger for $1.8K USD at the moment (I'm about to buy one for a system I'm DIYing from 31 kWh of AGM batteries I managed to get basically free from a test site of a company that closed down). I wouldn't be surprised if I could get the same capacity inverter/charger for something nearer to half the price by 2030, and a few percent more efficient to boot (this is 95% max efficiency but hopefully 97-98 will be more common by then).

You probably can get plenty of cheaper ones from China already but I want to be absolutely sure it'll meet Australian Standards since this will be grid tied for backup (but able to operate independently during outages), and since it's going under my house I want to know it's safe! Victron have a good track record, especially with a lot of use in maritime and caravan applications where you really don't want them catching on fire so that gives me confidence!

As someone who knows little about battery technology I was interested and trusted the author. But once I read the part about blockchain PoW vs PoS it seemed so off base that it threw the entire article into doubt...
Sure, it'd be great if all the little bumps in the road were already hammered out, but that we all know that's not how massive scale production works.

The two biggest numbers you need to look at in that article are lfp at 200 wh per kilogram and sodium ion at 160 w per kilogram.

Keep in mind that I believe neither lfp or sodium ion needs extensive amounts of cooling like Cobalt nickel batteries and their runaway fire problems. So their pack density is actually better and simpler.

So the 200 watt hour per kilogram basically equates to a 300 to 400 mile and possibly a 500 mile range car depending on efficiencies.

160 watt hour per kilogram sodium ion is the 200 to 300 and possibly 400 car

When you think about it that way consider the implications for electrifying all consumer transportation. The sodium ion density means that the city car that would serve possibly 4 to 5 billion people in the world is a solve technology, borrowing proper scaling.

The lfp density implies probably another billion to 2 billion people that need slightly better range, assuming good infrastructure for recharging.

Now the road maps for lfp and sodium ion. Both are going to probably increase by at least 20% in the next 2 or 3 years. Maybe 5 years tops.

If they can figure out sulfur chemistry versions of lithium sulfur and sodium sulfur then you may be able to double or triple densities in the next 10 to 15 years.

This is all very very revolutionary stuff.

> The associated charging, inverter, and physical structure components aren't going to follow the same downward curve.

I agree not the same downward curve, but it also has been on the downward curve, although different. Learning rate is rather a common phenomenon.

Estimating the learning curve of solar PV balance–of–system (2018) estimates 11% learning rate for BOS compared to 20% learning rate for module.

https://doi.org/10.1016/j.jclepro.2018.06.016

Narrow boundary analysis can be useful but problematic. The additional components is a great example.

Remember a large part of your electrical bill is paying for the grid, not just the energy it transports.