Aging is not a degradation process to be prevented — it is a manufacturing step to be designed.
- The membrane field treats fabrication and aging as independent, sequential processes: fabrication creates the membrane, aging degrades it.
- The root cause analysis reveals they are coupled.
- Fabrication conditions (solvent, evaporation rate, substrate chemistry) select the initial point on a complex free energy landscape, which determines the aging trajectory and potentially the endpoint.
- Different starting states may age to different metastable minima.
- Meanwhile, metallurgy solved the analogous problem 80+ years ago with TTT diagrams — mapping the transformation space to identify the optimal temperature-time pathway to a target microstructure, then quenching to arrest further change.
- No TTT diagram has ever been constructed for any microporous polymer.
- The physics headroom is large (2.5–13× permeance margin depending on thickness), multiple independent stabilization mechanisms are available, and the critical unknowns are resolvable with $40–60K of parallel gate experiments in 1–6 weeks.
If you prioritize speed to deployment and can accept the risk that thin-film selectivity may fall below 40 at 100 nm, start with thickness optimization (concept-1). If you need guaranteed selectivity >40 or 3+ year stability at 200 nm, pursue molecular spacers (concept-8) in parallel. Either way, start TTT diagrams (concept-10) immediately — the data informs every subsequent decision.
TTT Diagram Construction + Thickness Optimization with PVA Pore-Filling
Map CANAL aging kinetics at 6–8 temperatures to enable rational process design, while simultaneously testing whether 100 nm films on PVA-filled supports meet both permeance and selectivity targets. The TTT program requires only standard equipment; the thickness test resolves the make-or-break selectivity question in 4–6 weeks.
Rigid Molecular Spacers via Geological Pore-Pinning
Adamantane (6.9 Å) or carborane (5.8 Å) molecules at 0.5–3 wt% selectively prop open the 7–12 Å micropores that close during aging while being excluded from the ultramicropores (<5 Å) that provide selectivity — a geological pinning mechanism at the molecular scale, never tested in any membrane system.
- If this were my project, I'd start three things on Monday morning.
- First, I'd cast CANAL films at 100, 120, and 150 nm on silicon wafers and start aging them — this is the single most important experiment because it tells me whether the simplest solution works.
- If aged selectivity at 100 nm is >40, I might be done.
- Second, I'd dissolve 10 mg of CANAL in 1 mL each of six solvents (chloroform, DCM, THF, toluene, cyclopentanone, chlorobenzene) — a $500, 1-week experiment that either opens or closes the fabrication-controlled trajectory program.
- Third, I'd order adamantane and cast some spacer-loaded films alongside the pure CANAL controls, because it costs almost nothing extra and the geological pinning concept is too elegant not to test.
- The TTT diagram I'd start in week 2, once I have the casting protocol dialed in.
- Twenty-four films, eight temperatures, weekly measurements for four months.
- It's not glamorous work, but it's the single highest-leverage knowledge investment in the portfolio.
- Every subsequent decision — what temperature to pre-age at, whether multiple relaxation modes exist, whether the aging endpoint can be engineered — depends on this data.
- No one has ever done this for any microporous polymer, and the first TTT diagram would be a landmark publication regardless of what it shows.
- The VPI gate test I'd schedule as soon as I can get ALD reactor time — probably a 1-day experiment at a university collaborator.
- If TMA gets into CANAL micropores, that's a game-changer.
- If it doesn't, I've spent $3K and two weeks to definitively close a path.
- What I would NOT do is invest in modified CANAL synthesis for crosslinking (concept-4) until I've exhausted the simpler approaches.
- The non-linear crosslinking-permeability penalty<sup>[5]</sup> is a likely deal-breaker, and the synthesis investment is months of work.
- I'd also defer the organosilica capping layer (concept-11) — it's a multi-year R&D program that only makes sense if everything simpler fails.
- The beauty of this portfolio is that the cheapest experiments ($500–$10K each) resolve the critical unknowns.
- Spend $40–60K over 6 weeks, and you'll know which 2–3 paths deserve the real investment.