The current nanomaterial recipe bugs me because it uses a different tier of each space resource, and because from a certain perspective, there is no reason why it needs to require the final tiers of astronomic, biological, and material science to be researched.
My proposed solution is to make the recipe take (along with the particle stream and thermofluid) one each of lattice-pressure vessels, self-sealing gel, dynamic emitters, and heavy assemblies, and to add an aeroframe bulkhead to the lattice-pressure vessel recipe.