One rock, two products
The process CSIRO modelled is serpentinization, given a push. Water saturated with carbon dioxide is injected into the iron-rich mafic and ultramafic rocks of the Yilgarn Craton in Western Australia; the iron is oxidised, hydrogen gas comes off, and part of the injected carbon dioxide stays behind as solid carbonate. The simulation reports that raising temperature speeds the hydrogen-producing reactions, while carbon mineralization is most effective in a window of about 150 to 200 degrees Celsius. Fluid chemistry pulls the same way for the gas: lower-salinity water and alkaline conditions with higher pH give stronger hydrogen generation, and highly saline fluids hold it back.[1]
The two useful outputs therefore do not share one optimum. A reservoir run hot for gas is run past the range where the carbon mineralizes best; a reservoir held inside that range leaves hydrogen unmade. Regina Sander, the experimental reservoir engineer on the team, puts it plainly: the conditions that maximise hydrogen generation and the conditions that maximise carbon storage do not always coincide. There is a competing reading, and it deserves stating. The gap may belong to the modelled parameter range rather than to the rock, and at real reservoir pressures and flow rates the two curves could sit closer together than the simulation places them.[1]
What a simulation can and cannot settle
This result comes out of geochemical modelling, so it is a scenario conditional on the assumed temperatures, salinities and pH values rather than a quantity measured in rock. That distinction decides how much weight the 150 to 200 degree window can carry. It marks where to look in the laboratory; it is not a specification anyone could design a well around. The team says the same in its own terms, describing laboratory experiments, field trials and a search for suitable geological settings as the next steps, together with long-term system performance and commercial scalability.[1]
There is a second reason to keep the two products in separate columns of the ledger. Hydrogen made this way earns its low-carbon description from the carbon dioxide that stays behind as carbonate, so a well tuned for gas alone can deliver hydrogen while returning less carbon to the rock. The accounts balance only if both quantities are metered at the same site, over the same injection. Counted separately, one injection can be credited twice, and the same geological system that promised to do two things at once ends up reported as if it had.[1]
The measurement that would settle it
The next honest number is a paired one. If the laboratory experiments and field trials the team describes report hydrogen yield and carbonate formation from the same Yilgarn Craton rocks at the same temperature, the modelled 150 to 200 degree window becomes testable instead of assumed. Until such a pair is published, orange hydrogen remains a simulated pathway with a named constraint, and a project that cites only one of the two numbers has not answered the question the simulation raised.[1]
Two of the three levers are cheap to test first. Salinity and pH belong to the water an operator injects rather than to the rock it enters, so a bench programme can vary them at fixed temperature and report how far hydrogen generation actually moves. Temperature is the expensive half of the question, because it follows from depth and the geothermal gradient of the chosen site, and that is where a project has to choose between gas and storage. Picking the site, in other words, is already picking the product.[1]