Sunlight-driven photocatalytic water splitting for hydrogen production at scale | Hacker News Reader

Sunlight-driven photocatalytic water splitting for hydrogen production at scale

phys.org

76 pointsmardiyah24 comments

I recall from many years ago a photocatalytic process that used Iron/Iron-Oxide in a CSP solar setup. The solar tower superheated the water, splitting it; the O2 was absorbed by the Fe, leaving the H2 to be harvested; then the FeO2 was heated in a second pass to drive the O2 off, leaving the Fe to be reused. The whole thing sounded ingenious, and I wonder what became of the whole scheme. ISTR that there were some losses due to H2 and O2 recombining, but it sounded attractive in not needing a separation membrane as the OP article describes.

The temperatures involved are very high (~2000 C), which creates a host of problems (reradiation of energy, creep and degradation of materials). This also means only direct sunlight can be captured; flat panel collectors like ordinary PV modules also collect sunlight that has been diffused by clouds and atmospheric scattering.

Also, where would you get the Fe from? Seems like you’d substituting Fe for C, but Fe is rarely ever found in free form in nature.

The Fe (or more generally some element M; many possibilities have been considered) cycles from one oxidation state to another and back again. It is not used up.

Yeah its probably different iron oxides: fe(II) <-> fe(III) and not free iron.

Similar tech has been demonstrated many times before, with varying degrees of practicality.

Rather than separating the H2 from the O2, you can have microbes in the water eating dissolved H2, O2, and added dissolved CO2, and producing hydrocarbons that float to the surface for harvesting. You would harvest excess microbes, too, for animal feed. (You manage reproduction rate by control of trace minerals.)

This system uses ultraviolet radiation. If the catalyst is transparent to visible light, the panels could be behind it, and be made more efficient by the water cooling.

I wonder how much more efficient/cheaper in theory these need to be than the standard PV -> electrolyser route before they make sense.

Seems like the ability to have interchangeble electricity (which you can export and import to the system as needed) and the ability to locate the hydrogen production an aribtrary distance (with a transmission loss) from the energy gathering stage seem like a very decisive advantages, even before you consider the immense amounts of time and focus that are being put into improving PV cost and efficiency.

If this method works well, you'd probably still need to consider whether it would make more sense to generate the light artificially for similar reasons.

A potential use for this would be to use the hydrogen as a way to cheaply store the energy to be used when the sun is not shining. I.e. Pair it with fuel cell tech to generate electricity at night.

In this scenario the price of the tech should be compared to solar + batteries per kWh, and I could imagine that at scale hydrogen storage could beat out the cost of betteries.

Energy density is another big factor. Hydrogen has a much higher energy density especially for automobile use compared to batteries.

My opinion, fwiw, is that batteries have won the day in the automobile space. The infrastructure investment required to make hydrogen work for automobiles must be astronomical vs incremental improvements required to the existing electrical network.

Couple this with the fact hydrogen distribution and storage have some significant problems, plus recharging at home/work/parking is super convenient.

The only real thing stopping battery powered electric cars from being the obvious first choice in many markets is the cost of battery packs. Conveniently this has progressively improved yer-on-year. It is simply a matter of time.

I’d imagine quite a lot. The fungibility and cheap, instant long range transmission of electricity is worth an awful lot of ungainly hydrogen in a tank somewhere else that you can’t use.

It's more of a cost than an efficiency equation. The efficiency only matters in the sense that it drives up the cost because you'd need more equipment to produce hydrogen. The equipment is fixed cost investment. So, the longer it runs, the more fuel it will produce.

Using stored energy to produce light to produce energy does not make a whole lot of sense. It's a very lossy process. If you have stored energy, maybe just use it directly?

The article mentions a 5% solar efficiency, it is my understanding that the current most efficient PV panels are around 22% solar efficient. So I would think You may be correct.

It would be nice if they compared PV panel powered electrolysis hydrogen generation against this new tech.

For many applications, efficiency is not that important. The important metric would instead be $/kwh.

I agree, but isn't that a sad state of affairs?

The Libretarian in me wants to believe that eventually the Efficiency and $/W will balance out but at what real cost?

$/W might make sense from 40,000 feet but the grounded pragmatist in me feels there are other things to consider and confirms this by the ironic inevitably of one of those other things ending up crushing the $/$ formula, and here we are now.......

Now we can grow giant floating hydrogen-filled platforms to serve as the foundation for our skycities!

How might we use solar to produce hydrogen at a scale relevant for building skycities? Ok, it takes about 50kwh to produce 1 kilogram of hydrogen from 9 kg of water. At sea level, that makes 12 cubic meters.

One of the biggest aerostatic platforms in history was the Graf Zeppelin, which had a surface area of 27,299 square meters. Let's say the top ⅓ was covered in solar, making 1.2 kWh per square meter per day. In total, that's 10,800 kWh per day.

So, we get 2,589 cubic meters of hydrogen a day. At that rate, it would take 77 days to fill the Zeppelin's 200,000 cubic meter capacity.

Going bigger, 1 square kilometer of hydrogen is 1 billion cubic meters of hydrogen. At sea level, this much hydrogen could lift 1.3 billion kgs (600 olympic swimming pools). But at 20km, it can only lift 1 million kgs (1 swimming pool).

Can you imagine giant bubbles growing in the sea, floating upwards when ready, and assembled as building materials for aerial architecture?

You might enjoy Bucky Fuller's concept of what he called "Cloud Nine":

> Fuller suggested that the mass of a mile-wide geodesic sphere would be negligible compared to the mass of the air trapped within it. He suggested that if the air inside such a sphere were heated even by one degree higher than the ambient temperature of its surroundings, the sphere could become airborne.

https://en.wikipedia.org/wiki/Cloud_Nine_(tensegrity_sphere)

"More rigorous safety tests are still needed, but if a properly designed system is used, the highly explosive hydrogen-oxygen gas can be safely handled for long periods."

What could possibly go wrong?

Probably a long list of things that's more-or-less similar to what can go wrong in an oil refinery, or at an oil well.

Modern civilization requires industrial-scale processing of toxic, explosive, corrosive, combustive, and in short, dangerous gases and liquids - often under high pressure.

But not, usually, gas mixtures that can explode without further mixing.

Have you ever wondered why there are so many aliens coming to Earth for our water? They did this on their planets, and lost all their hydrogen to space. Pretty soon we'll be looking down the barrels of their Illudium Q-36 Explosive Space Modulators... ;-)

Ah, no.

NO scientist is qualified to make claims about what can be ENGINEERED "at scale". ANYTIME you see a scientist or research lab or academic lab make such a claim, the first truth and first thought should be "Bullshit!"

There is the "20 year rule" which says there's a MINIMUM 20 year delay between a laboratory demonstration (which this is) and any chance of economic viability. That's how long it takes ENGINEERS to validate the claims of lab discovers of scientists and actually engineer a viable technology.

And even then: the failure rate on exiting this 20-year pipeline (which must be funded the entire 20 years or more) is still 95%; most ideas do NOT SUCCEED.

The solution is simple. Embed engineers with the scientists. Just like we embedded business, ops, and security roles in our Dev teams.

Reply on news.ycombinator.com