What's the holdup on green hydrogen?

Factual Questions
61

Blue hydrogen really just another oil company bailout? Shocker!

I’ve been following these guys for a few months now:

It’s still a very early stage startup, so take everything with a grain of salt. But their approach is the same as one I’ve been advocating for a while for several different applications. The observation is that solar power is extremely cheap if you don’t need all the extras that people usually want, like storage systems so that you have power at night. But what if you just used the solar when it was producing, and shut down your production when it’s night or cloudy or whatever? Then you just need a load of panels and not much more.

This makes your production equipment capital-inefficient, because it’s only in use maybe 1/3 of the time. The question then becomes whether you can lower the cost of that equipment–perhaps at the expense of power efficiency. It’s not so bad to be power-inefficient, because again the solar power itself is extremely cheap.

So instead of using fancy electrolyzers using electrolyte membranes and platinum electrodes and all that, use the cheap version, similar to the high-school chemistry class version. It’s maybe 60% as efficient as the expensive one, but it’s way, way less than 60% of the cost. So it’s a net win if the electricity is cheap.

No idea if it’ll all work out, but I think it’s great that they’re trying. I’d like to see the same principle applied to other power-intensive applications, like water desalination and aluminum refining.

62

Assuming you want your power to be stable, you’ll still need batteries.

And if you are in North America, assume for costing that your capacity factor is somewhere between 16% and 24%, not 50% or 33%.

And your production might have to be seasonal. Last year solar produced about 12% of capacity in December on average in the U.S. About 3-4% higher in November and January. If you are in northern or coastal regions, worse. There have been worse years, too. We just went through a season where solar only produced 1.2 percent of capacity for an entire month.

You’ll also need to worry about lots of unscheduled downtime, as solar goes away for days or even weeks at a time. Especially in winter.

Shutting down a batch factory is no small feat. Building one that has to shut down every day and be prepared for unscheduled shutdowns at any time sounds like a big challenge.

You could connect to the grid and use solar power as a supplemental, but then the problem is that grid power prices will be close to zero when it’s sunny and you are producing your own power, but when the sun goes down you’ll pay max prices. So the savings would be less than you’d hope.

63

You don’t need the power to be stable. That’s the whole point. It produces when the sun is shining; otherwise not. Normally it’s unacceptable to let your equipment just go idle on a periodic basis, but that’s because it tends to be built for maximum efficiency and that costs more. You can make cheaper equipment if you sacrifice some of that. Also, the capital for a grid connection would make it too expensive, let alone the power itself. Aside from monitoring systems and the like, it’s solar or nothing.

You don’t even need to worry about weeks or months of downtime, because storage is so cheap (especially if stored as ammonia). You overproduce when the sun is shining and pull on reserves otherwise. The “battery” is the reserves of the final product, which is much easier to store than the electricity itself.

And, needless to say, don’t build your solar-green-hydrogen plants in freaking Alberta.

64

Correct. The membrane and catalyst that you proposed avoiding make up a small portion of CAPEX, which is dominated by balance of plant. And CAPEX is a small portion of the overall cost to produce dry, compressed hydrogen. Avoiding PGMs and Nafion has a negligible impact on cost of hydrogen.

65

It’s a small portion now. But the price now is uneconomic and incompatible with a future where we need green hydrogen at an industrial scale. The capex is not small when the target is $1/kg.

Of course, the membranes and such are not the only thing that need to be cost-optimized. Everything in the plant needs to be optimized. As another example, it’s dumb to produce DC solar power, convert to AC, boost the voltage for transmission, transmit over hundreds of miles, reduce the voltage, convert to DC, and then finally input into the electrolyzer.

Instead, use the DC that you already have, putting the generation right next to the facility. The plant can then be matched with the generation, even dynamically so by switching the number of cells on demand. That can be done much more cheaply than the usual conversion equipment.

There’s much more to solve than just that, too–but each one of these problems, even if they are a small fraction of the problem now, are not a small fraction when combined.

I find this to be a really common failure in reasoning. If you have a way to reduce the cost of something from $100 to $95, you might say big deal, it’s only 5%. But if you have another 18 similar cost reductions, then they are each equivalent to halving the cost, because they each could have been the last one, and that last one reduced the cost from $10 to $5. Tiny reductions become more significant as you pile on more of them.

Which, again, isn’t to say that this particular company can pull it off. But dismissing them because any one particular optimization is negligible given the current state of affairs is bad reasoning.

66

Incidentally, this page says the current capex is about 1/3 of the total cost:

The energy is about half the total cost.

PV panels are under <$1/W these days. In a sunny area, 1 W will produce about 75 kWh over its lifetime, or $0.0133/kWh. Hydrogen takes ~40 kWh/kg at optimal efficiency, so at current prices–and assuming no other costs whatsoever–$0.53/kg is achievable.

Solar PV will continue to come down in price. Arguably, it’s already closer to $0.50/W than $1. On the other hand, the other costs can’t be ignored. And the energy costs will be higher due to inefficiency. So the question is, how close can they come to costs equivalent to a field full of solar panels? If they can get within a factor of 2, their goal of $1/kg might be achievable.

That will require a complete rethink of the plant setup, which is their whole strategy. But it will undoubtedly be difficult to get there.

67

I dunno, if I managed to get 19 cost reductions, I’d be really tempted to go for 21 and then watch my grey goo consume the universe.

68

I think we’re largely on the same page. My issue wasn’t with a potential pathway to $1/kg hydrogen, but with highlighting catalyst and membrane when they don’t play a major role in cost and when the incumbent technology doesn’t even use PGMs or Nafion. Had you skipped this:

And gone directly to chipping away at costs associated with module/stack price and module/stack efficiency, reliability/lifetime, O&M, other hardware, labor, financing, I’d have no argument. However, beware the back of the envelope:

The cost of a utility-scale solar installation is ~3x the module cost. Module costs are down to $0.3-0.4/W (again, utility-scale), so $1/WDC is a good starting point. Which works out to an LCOE of about $0.4/kWh, ignoring recent weirdness in financing costs. This is a substantial decrease from a decade ago but the pace has slowed. While true that DC → AC → DC doesn’t make sense, inverter cost is small enough at this scale that eliminating it doesn’t do much. Although yes, the pennies add up.

Most of the anticipated savings are not from any grand technological innovations (not that we aren’t trying there too) but rather simple economies of scale. Which is currently bottlenecked by manufacturing capacity. Some of the companies I work with are seeing year-long waits for electrolyzers.

For readers interested in this topic, I recommend keeping an eye out for releases from DOE EERE HFTO and NREL (where much of EERE’s analysis gets farmed out.) Link dump for anyone who needs help sleeping tonight:

https://emp.lbl.gov/utility-scale-solar

U.S. Solar Photovoltaic System and Energy Storage Cost Benchmarks, With Minimum Sustainable Price Analysis: Q1 2023

https://iopscience.iop.org/article/10.1149/2.F15214IF/meta

69

Well, that’s fair. In my defense, I did throw in an “and all that” in my comment, although admittedly it was doing a lot of work.

They mention some other things in their whitepaper as well, such as avoiding high pressures and temperatures, and the attendant equipment for that (pumps, heat exchangers, etc.). And it’s just a whitepaper, so it doesn’t really go into super-deep detail anyway. It’s just a basic explanation of their strategy.

Definitely. But it does illustrate a different benefit to a simplified architecture: it may be more manufacturable at scale. While I don’t know anything about the production of proton exchange membranes (say), I find it easy to believe that even if they don’t dominate the cost of an electrolyzer, they may very well still be a long lead-time item that isn’t available at the scale required.

Put another way, there are some components in the global supply chain that are available in basically unlimited quantities–structural metals, many plastics, and so on. On the other hand, some items belong to a niche industry and increasing the supply would involve building out a new supply chain–itself a capex-heavy endeavor, and risky in case the whole project doesn’t pan out.

If a project like this can limit itself to more scalable components, they have a greater chance of success.

I’m assuming that should be $0.04/kWh. 40 cents would be rather high.

The needs of the plant can also alter the solar design. As a simple example, some solar farms put the panels at an angle that maximizes later-afternoon output at the expense of average output, because household use peaks then (mostly air conditioning).

But that isn’t needed for this application, which can run whenever and just wants the lowest cost power. And that potentially means less material used for the structure (due to a shallower angle), which means lower costs yet. At a certain panel cost, you can almost entirely eliminate the structure and just put them flat on the ground. Which also may reduce maintenance costs due to the simpler geometry.

Whether or not these kinds of ideas work out in this case remains to be seen; my point is that you can’t necessarily point at “utility-scale PV” as a benchmark, since it pays a cost to meet requirements that aren’t needed here.

70

Sorry, but are you considering capavity factor here? If not, multiply the cost by at least 4. And you WILL need stable power. Lots of motors, electronics, etc will require it.

And building a facility that will only run 1/4 of the time is going to have high capital allocation costs. And with 25% uptime, some of it randomly distributed, the efficiency of the operation will be terrible.

Just put a nuclear plant amd the facility somewhere, and run it at 99% uptime.

71

Note: the author of this article is, shall we say, not a fan of hydrogen.

72

You aren’t even addressing the premise. If you want to argue that it’s impossible for them to reduce the capital costs at the expense of efficiency, go ahead, but simply stating that they can’t do it because capital costs are too high (exactly what they want to reduce) is not an argument.

Plenty of area in the American Southwest that gets 4000 hours of sunshine per year. 100,000 hours over a 25-year lifetime. 1 W of solar isn’t going to get the full 100 kWh out of that due to cosine losses, but it should get over 50. And that 1 W of solar costs about a dollar.