CO2 capture

=energy =global warming =technology

 

 

The theoretical energy required to purify something is only logarithmic with the concentration ratio. The problem with extracting something with a low concentration is that you have to process a lot of material.

If you want to dig something out of the ground (like coal or limestone) and put it in a truck, that's about $20/ton.

Desalinated water is extremely cheap, about $0.50/ton. Making it involves taking seawater (which is free), filtering it, pressurizing it, and running it past a relatively cheap type of membrane.

If you made a lithium separator that processes seawater for $0.50/ton, your lithium would cost $2.5 million / ton. Realistically, lithium from seawater would be over 10x that.

Now, suppose you could process air for $0.50/ton and remove all the CO2. There's about 400 ppm of CO2 in air now, so that's about $1250/ton of CO2 removed. Air can be scrubbed of CO2 for significantly less per ton than desalination costs, but this is actually a decent starting point; actual costs would probably be between 1/2 and 1/10 that.

 

 

To collect dilute stuff, sorbent needs to hold it pretty strongly. You also want to use liquids if possible.

So, an obvious (to chemical engineers) and reasonable approach is NaOH or KOH in water, transferring carbonate to Ca(OH)2 with precipitation of CaCO3, and then recovering CO2 by calcination of CaCO3. There are 2 relatively influential papers on CO2 capture from air using this approach:

paper 1 (2011) estimates $430 to $550 / ton CO2
paper 2 (2018) estimates $94 to $232 / ton CO2

 

Paper 2 claims it reduces costs over paper 1 by:

- using cross-flow air exchangers instead of counterflow, increasing airflow vs fan cost
- using cheaper PVC packing instead of steel
- using KOH instead of NaOH

 

As techno-economic analysis goes, they're both rather mediocre. (That's to be expected: the people who do the most sober and well-informed analysis tend to avoid things that seem far from viability.) So, what's wrong?

 

paper 1:

- Stainless steel exposed to strong alkali in air will corrode.
- As paper 2 notes, the concern about spray from the alkali solution isn't necessary, and cross-flow is probably better.

 

 paper 2:

- Salty water won't wet PVC or polypropylene very well; you can't just assume your full packing material surface area is used.
- KOH isn't a huge improvement over NaOH; their basis for claiming that is bad.
- We know how much calcination of portland cement costs. You can't just assume you have a cheaper and more efficient way of doing that.

 

With this approach, $200 to $300 / ton CO2 seems correct to me. By the way, both papers have calcination of CaCO3 being done with natural gas. Trying to do that with electricity would be considerably more expensive.

 

 

What about "accelerated weathering" for CO2 sequestration?

Digging stuff up is about $20/ton, and you'd need to dig up 3 tons of Mg silicate to absorb 1 ton of CO2.

But suppose you do that. Great, you exposed some fresh magnesium silicate to the CO2 in air, and now a very thin layer of carbonate will form on the surface as it very slowly reacts. If you crush it to fine particles and spread it over a large area, you can get it to actually react, but obviously that's more expensive.

 

 

The US government decided to spend a bunch of money on development of CO2 capture and sequestration from coal plants. This accomplished nothing. The Kemper Project (a pilot plant for coal power with CO2 capture) cost $7.5 billion, and never operated.

People say the US doesn't have industrial policy like China does, but one could argue that it's more accurate to say that the US government is just so bad at industrial policy now that it doesn't seem like it's doing anything. In other words, the US used to spend its industrial policy money on stuff like the Heavy Press Program, and now the US spends it on debacles like the Kemper Project and obviously stupid ideas like cylindrical solar cells.

 

 

Separating CO2 from a mixture is cheaper when the concentration is higher. The optimal process for CO2 separation also varies with concentration: to remove very dilute CO2 (perhaps from air) you need a something that binds to it strongly (such as hydroxides).

In order of increasing cost, some reasonable sources of pure CO2 are:

 

1) Separation from syngas during ammonia production, maybe with Selexol. This is basically a free side product.

2) Separation from gas made during fermentation of sugar (to eg ethanol) with cryogenic separation. Maybe $25/ton.

3) Amine scrubbing of flue gas from boilers. Maybe $50/ton.

4) Gas turbines using pure oxygen and recycled CO2. Maybe $60/ton...with very large cryogenic air separation facilities.

5) Amine scrubbing of exhaust from gas turbines. Maybe $80/ton.

 

The cost of amine scrubbing is largely from steam to heat the liquid sorbent for CO2 stripping, so it varies depending on the cost of steam.

 

CO2 from any of those sources is, for a pure chemical, a very cheap material. The problem is...what do you do with it? Note that liquefying it and trucking it around significant distances will make it several times as expensive. Cryogenic separation is probably the cheapest approach if CO2 gets transported by truck to, say, greenhouses.

CO2 is mostly used to make urea, which is basically a way of stabilizing ammonia; the CO2 just gets released again.

You can do "cyclic carbonate synthesis" with epoxides and CO2. You can make salicylic acid from phenol and CO2. If you really want to, I guess you can deprotonate acetone and make acetoacetate. These are all niche reactions that are very expensive per mass of CO2.

 

So, the US government decided to give a bunch of grants for scientific research into chemical uses for CO2. This is a smart approach, unless you know about chemistry, in which case it's dumb.

Some people thought: "How about we turn CO2 and electricity from solar panels into chemicals, maybe fuels?" There were a bunch of grants for doing that. A bunch of time was spent by PhDs and professors on it. People wrote papers where they tried to pretend that converting CO2 to cheap stuff with inefficient electricity use and expensive metals is a promising line of research, but nothing useful resulted, because the premise was bad.

The best current approach for converting CO2 with electricity is running a solid oxide fuel cell backwards. Not only are SOFCs too expensive, but that makes a product that's worth less than the inputs, which is why SOFCs are usually meant for generating electricity. But still, there are startups doing that now. Perhaps they're hoping for subsidies.

If you want to use CO2 to make chemicals, it's better to combine it with hydrogen and make syngas than to try to make products directly. Syngas is well-understood and heavily used. Of course, hydrogen from electrolysis isn't currently viable either. It's ~2.5x the cost of purified hydrogen from natural gas - but that's a lower multiple than many other approaches for renewable chemicals!

 

"But in the future we'll have cheap solar power for that."

Yeah? Why? Solar power has stopped getting cheaper. If anything, renewable power with storage will be more expensive than what we have now.

"Because there will be excess PV capacity during the day, and you can do electrolysis then."

No, you don't want to run electrolysis plants for a 4 hours a day. You need to run them most of the time for the economics to work.

 

 

So, what else can you do with CO2?

Well, it's used for enhanced oil recovery. (That's not what the "direct air capture" people want to do!)

You could potentially send it to greenhouses. It's generally been too expensive to do that; it's been cheaper for greenhouses to connect to existing natural gas pipes and burn methane to make CO2. But you could build greenhouses and CO2 pipes.

 

 

If you want to actually collect CO2 from air economically, you'd need some kind of...nanobots. Self-replicating nanobots, maybe powered by sunlight. Using some kind of molecular machines powered by sunlight to collect CO2. And maybe they could use the CO2 for the self-replication. Yep, that's what you'd need.

 

 

 

back to index