Price a capture plant per tonne avoided, at the hours it will really run

Carbon capture is back on Canadian board agendas, helped by a federal investment tax credit, by industrial carbon pricing and by the argument that some gas-fired generation and heavy industry will need it. I spent three years leading a capture programme at an Alberta oil-sands producer, and my position is narrow. A board should not approve, or count on, a capture project until management has stated its cost per tonne avoided rather than captured, at an operating profile the plant will really achieve, with the share paid by the tax credit separated from the share that depends on carbon credit prices. Each of those three corrections usually makes a project look worse. Each also makes it more likely to be built and run as planned.

I am not opposed to capture. Concentrated carbon dioxide streams, such as those from hydrogen production or gas processing, can be captured at sensible cost, and a few heavy industries have no realistic alternative. My concern is with the way capture figures travel from engineering studies to investor presentations, losing their assumptions on the way.

Captured is not avoided

At Kelsvar Oil Sands Corporation in 2012 our parent company had announced an ambition to capture a million tonnes a year. Our site emitted about 2.3 million tonnes, mostly from once-through steam generators and a cogeneration plant, and all of it was dilute flue gas. Our first study put the cost at C$118 a tonne captured. That number was correct and misleading. An amine capture plant needs steam to regenerate its solvent, about three gigajoules a tonne in our design, and on a site without spare steam that means burning more natural gas, which produces more carbon dioxide. Once we subtracted those emissions, the cost per tonne avoided was about C$146.

I put the captured figure on the first page of my first paper, and I regret it. It was the number the climate strategy team wanted to repeat, and it travelled further than the caveats did. The board-level question should always be how much carbon dioxide stops reaching the atmosphere compared with not building the plant, and at what cost. A cost per tonne captured answers a chemist's question. A cost per tonne avoided answers a director's.

The hours a plant actually runs

Engineering studies are usually prepared at a nameplate operating profile, often above ninety per cent availability. Real capture plants share the fortunes of the facility they serve. If a gas-fired power plant runs for half the year because it is called on mainly at peaks, its capture unit runs for half the year too, and the capital cost is spread over half the tonnes. If the host plant is curtailed when prices are low, so is the capture plant.

Our pilot ran from November 2013 to January 2015 and captured about 9,800 tonnes. It taught us things no study had. Solvent losses were 2.3 times the supplier's estimate, because of oxygen and nitrogen dioxide in our flue gas, and the steam requirement was 3.4 gigajoules a tonne rather than three. The revised full-scale estimate rose to C$171 a tonne avoided. A pilot costing C$31 million told us that a C$1.4 billion plant would have been more expensive to run than anyone had planned, and it did so before we had committed the capital. I would ask any board to treat a site-specific pilot on the real gas stream as a condition of a final investment decision for a first-of-a-kind plant, not as an optional extra.

Transport and storage deserve the same scrutiny. In our study they were about a third of the cost: a 90-kilometre pipeline to a storage formation, monitoring wells and the long-term liability for what was stored. A standalone project carries all of that alone and pays for a pipeline sized for one site. Where several emitters can share a trunk line and a storage hub, the cost per tonne falls and the risk is spread, but the project then depends on partners whose timetables it does not control. A board should know which of those two risks it is taking, and who holds the storage liability once injection stops. In 2013 we appraised our site alone, and I now think that was the larger of my two mistakes.

Separate the tax credit from the carbon price

The economics are very different today from those we faced in 2013, when the provincial compliance price was C$15 a tonne. The federal investment tax credit now pays a substantial share of eligible capture equipment cost for projects built this decade, and the federal benchmark for carbon pricing reached C$110 a tonne in 2026. Those changes make some projects viable that were not before. They do not make every project viable, and they carry different kinds of risk.

A tax credit reduces capital cost once the project is built and qualifies, so its risk lies mainly in eligibility, construction and the rules in force at the time. A carbon price is earned every year the plant runs, and in industrial pricing systems it is realised mostly through the value of credits or avoided compliance costs, which depend on provincial policy, on the market for credits and on the facility's benchmark. Credits in some provincial systems have traded well below the headline price. A business case that treats the headline price as revenue for twenty years is making a policy forecast, whether it says so or not. Contracts that fix a carbon price for a period, where they can be obtained, reduce that risk and should be shown separately.

I would ask boards to request one page with three numbers on it: the cost per tonne avoided at the expected operating profile, the share of that cost met by the tax credit and the carbon price at which the project breaks even over its life. Beside the last number, management should state what it is assuming about credit prices and why. A project that works at a credit price close to today's traded values deserves serious consideration. One that works only if the headline price is fully realised for decades is a bet on policy, and directors should approve it, if at all, knowing that.