The Answer Is Usually in Another Industry's Literature

Materials knowledge is not evenly developed. It is developed wherever it is mission-critical, and it stays thin everywhere it is merely useful.

To grow the single-crystal blades inside a jet engine, aerospace engineers learned to shape part of the growth chamber in a specific way. For them the technique was routine. Condensed-matter physicists growing their own crystals, with the same problem, had never heard of it.

In gas capture, humidity degrades performance, so teams design around it with real rigour. In semiconductors the same sensitivity exists, but attention sits on electronic properties, so the humidity knowledge stayed shallow. Two industries, one shared physics, and a twenty year gap in what each of them wrote down.

This asymmetry is the opportunity. The knowledge that unlocks one industry's problem has usually already been worked out, in depth, by an industry that needed it more. Nobody has connected it, because no single team has a reason to read all of it.

That is what Chemia connects. Not another database, but an ontology that links materials, properties, operating conditions and failure modes across every field that studied them, so a question asked in one industry can be answered with what another industry already proved.

Connecting the knowledge is only half of it. The other half is trusting what comes back. Every value in Elixir carries its origin and our confidence in it, and we validate on held-out data before we put a number in front of a client. For the dielectric constant, for example, we trained on first-principles data and tested on a set held out by symmetry group, then checked the model a year later against newly available materials it had never seen. We see the same behaviour on other properties such as specific heat, and we are glad to walk a technical team through the details.

That approach has already been put to work. A specialty chemical manufacturer came to us during the development of new CO2 capture materials. Their strongest candidates were not stable under the operating conditions the application demanded, and confirming their actual CO2 and H2O uptake would have meant months of measurement for each one. We calculated the adsorption properties directly with first-principles methods, then predicted the same properties with our materials-informed AI models at the operating conditions that mattered, with accuracy high enough to act on. That let the team eliminate the candidates that would have failed before spending the months it would have taken to prove it experimentally.

Feasibility, it turns out, is not always about scale. A material used as a coating is needed in microns, not tonnes. A material used in a sensing element is needed in milligrams. A material used as a functional layer, a barrier, an interface or a catalyst support sits in the same category. Candidates that look commercially impossible at bulk scale are entirely practical in those form factors, and they get eliminated early by a screen that never asked how the material would actually be deployed. Elixir keeps the deployment form in view, so options are not discarded for a constraint that does not apply to them.

None of this is only about selection, either. If you know the full landscape, including what has been measured, what has only been predicted, and where the boundary of known performance actually sits, then you know where the open space is. Innovation is not a separate exercise from selection. It starts from the same picture, and it fails for the same reason when that picture is incomplete: teams reinvent what an adjacent industry already solved, or push into a direction the data would have told them was closed.

So Elixir serves two jobs from one foundation. For an engineering team, it is a shortlist with reasons and a test plan. For an R&D team, it is a starting position: here is what exists, here is where the frontier is, here is the space nobody has entered and the constraint that would have to be broken to enter it. In both cases the work is the same. Know the landscape before you commit.