Advancing electrolysis: Splitting water to store energy as hydrogen | Hacker News Reader

Advancing electrolysis: Splitting water to store energy as hydrogen

inl.gov

37 pointselorant25 comments

"the cell was firstly operated at an electrolysis mode with a constant current density of 0.6 A cm−2 for three minutes to generate hydrogen, which was instantly consumed by switching to a fuel cell mode at 0.2 A cm−2 for two minutes"

So, 22% round trip efficiency

Shouldn't you take into account the voltage. From the graphs they don't look like they were symmetrical.

Doh!

Fig 6c[0] shows ~1.6 in and 0.4v out so... 5% round trip efficiency?

[0] https://www.nature.com/articles/s41467-020-15677-z/figures/6

Interesting:

I had no idea there was an Idaho National Lab.

The new electrode is a perovskite which is a crystal structure tangentially related to superconductors among other things.

The electrode allows for reversible operation in the 400-600 degC range.

INL (the site) put food on our family's table during my childhood in Idaho Falls.

Home of a deadly reactor explosion.

https://en.wikipedia.org/wiki/SL-1

I did know about the SL-1 which makes it an even more glaring oversight. Maybe I associated it more with Argonne? I dunno.

But that SL-1...dang.

It's involved in a lot of nuclear work. One of the first nuclear reactors in the work like was there. They also have strong ties to Boise State's Micron School of Material Science and engineering (Micron as in Crucial RAM brand) [1].

Yes, I'm biased, as I did PhD work there. Fun to watch talks on nuclear engineering. I'll have to look up the group behind the perovskite. No one ever seems to commercialize perovskite tech. Maybe small applications, but nothing large scale?

1: https://www.boisestate.edu/coen-materials/

Does anyone know if coal or nat gas power plants can burn hydrogen too? Or be retrofitted to burn it? Then you could store excess solar or wind energy generated during the day as hydrogen and burn it at night.

Sure, since the lab is running their perovskite at 600 C anyway, just add CO2 in a chamber with some nickel and get methane out via the Sabatier reaction. Send the methane to where ever its needed.

So many problems are solved so easily by "just" doing simple thing X. Maybe this really is a simple fix. It doesn't sound that way to me though.

It's not. The response is 'accurate' not 'easy' :-) The biggest challenge is getting a source of CO2 using less energy than you are going to get back by burning the methane.

Thermodynamics doesn't let you cheat sadly.

When evaluating these sorts of things you follow the energy consumption of each step and remove it from the initial pool of energy. Energy out over energy in will give you a way to compare methods for "goodness" regardless of complexity. Processes that have net negative energy are clearly in the "bad" category. Any positive is something.

Localized schemes like hydro pumping do quite well in this equation. It is distance schemes that suffer.

Distance schemes are hard because pushing electrons through a wire spends power just getting through the wire. Eventually its all gone. This is why you can't just cover 1/3 of Nevada with PV solar cells and power the grid from that one state.

So distance means transforming electricity into some other energy (chemical, potential, kinetic, thermal), moving it physically somewhere else (ideally without much loss) and then converting it back into electricity at the destination.

One of the more fanciful ideas I heard was train cars full of Tesla powerwall packs. If you've ever seen an operating coal fired power plant you may have seen the long line of coal cars in a train that pulls in each day. Once at the plant the cars are slowly pulled through the plant, dumping coal when they need to into the feed mechanism. Eventually a string of empty cars is on the other side to take back to be refilled. Imagine doing that with powerwalls. Cars pull up, connect electrically to the plant and start discharging. Once they are down to their "empty" level the car disconnects and moves on. I think it would be fascinating to watch.

Would the energy cost of moving the electricity be lower than the energy cost of just transmitting it over power lines?

Depends almost entirely on distance. Back in one of my college physics classes in LA we analyzed moving electricity from Hoover Dam in (Nevada/Arizona) to LA. It was a lot of fun. You can raise the voltage so you need less current but then you have to raise the wires so they don't arc to the ground and separate them. As I recall (it was a while ago) we were using the Fischer-Tropsch[1] process to convert electricity into oil to transport by tanker to fossil fueled plants.

You plot all the variables (voltage, current, distance) and there are regions where it is better to convert it and regions where it isn't

[2] https://en.wikipedia.org/wiki/Fischer%E2%80%93Tropsch_proces...

> spends power just getting through the wire. Eventually its all gone.

No, transmission losses for such concentrated power generation could be kept acceptable single digit percent by building a proper power transmission system. Just use higher voltage, thicker wires. It would be expensive, but if Nevada was a great energy source, it could be worth it long term.

> This is why you can't just cover 1/3 of Nevada with PV solar cells and power the grid from that one state.

I can think of more substantial problems with that - fragile panels exposed to harmful environment, manufacturing and maintaining panels generates too much waste per kWh, unreliable operation dependent on weather. If Nevada was to be used as a primary center for energy generation, I suggest uranium/thorium reactors. They are better in every possible way.

Better, except that they take too long to build, cost too much to operate, and have proven incompatible with sound public policy.

Nuclear energy has these problems, but these are human society problems. The technological difficulties can be overcome if STEMs are given enough power. From the standpoint of environment, safety, reliability there is a strong case to try harder.

The median time to build a nuclear reactor is 7.5 years [1]. That is not too bad.

[1] http://euanmearns.com/how-long-does-it-take-to-build-a-nucle...

Wire the powerwalls together inside the train and join them to the train’s electric drivetrain. Now you have increased range in your locomotive battery pack, instead of needing direct power from the grid or diesel to burn. Since the train in this metaphor is natively electric (like most existing trains with diesel generators powering an electric drivetrain), the same rails that power the train can charge the powerwall near the generator or discharge far away from the generator, all without the train stopping.

According to wikipedia HVDC lines can get to 3%/1000km losses, so that should be 12%-13% losses from NY to LA. Electrical losses are not the problem in continental US. It's the cost of building/maintaining infrastructure and being able to organize to do it.

Thanks for the response. In another comment you mentioned that getting the CO2 would likely cost more energy than would be generated by burning the methane. Is there any reason you couldn't burn the hydrogen directly? I'm a chem noob so this might be a stupid question.

Its not a stupid question, I'm sure lots of folks wonder the same thing. The challenge as I understand it is holding on to the hydrogen.

As a gas, hydrogen is really really small and not very dense. Storing, transporting, and handling hydrogen is really hard. I got a chance to see the hydrogen refueling station at a hydrogen bus pilot project and keeping enough hydrogen (without spending all the energy to chill it down in to liquid form) and dispensing it pretty amazing. Its way simpler to use gas generators (things that out gas hydrogen) like they did for the big zeppelins[1])

That is why most schemes that have free hydrogen available, if they aren't using fuel cells on site, usually have a process for capturing the hydrogen in a form that is easier to handle and transport as well as convert back into either heat or hydrogen.

[1] https://welweb.org/ThenandNow/Hydrogen%20Generation.html

This is very much an area of active research. See: https://soundcloud.com/energylive/podcast-energy-live-evoene... for a lab in Australia.

The main use of hydrogen is longer term storage than you are suggesting. For example, in cold climates generating energy from solar in summer and using it for heating in winter.

The other way it could be useful is for long distance transportation and export.

For generating electricity during the day and using it at night, it is better to use batteries. Converting energy to hydrogen is inefficient. Following that up by burning it to make steam to push a turbine to get electricity is very inefficient.

However, batteries are heavy and expensive, so you probably don't want to ship them around or use them for bulk long term storage. Big tanks of hydrogen should be much cheaper, despite the energy losses.

This equation will get better as more renewable generation is added to electrical grids, pushing down the cost for energy, but also increasing the need for load balancing.

If using the hydrogen for heating, you can bypass some of the extra inefficiency by burning it directly for heat. It may even be possible to transport it using existing (methane) gas pipes and combust it in something similar to existing gas boilers. Again, an area of ongoing research.

One of the UK gas supply companies has a pilot project which is adding 20% hydrogen into network to reduce CO2 emissions.

https://hydeploy.co.uk/about/news/uks-first-grid-injected-hy...

A lot of companies, especially in the Bay Area (Apple, Google, eBay, Yahoo!, etc.) are using onsite solid-oxide fuel cell power generators from Bloom Energy that directly convert energy from hydrogen bonds into electricity. Currently, they get the hydrogen from reforming natural gas. The plan is to inject a bit of hydrogen into the natural gas grid for less carbon emissions, and eventually to use pure hydrogen.

Reply on news.ycombinator.com