Skip to main content
Back to timeline
arXivSource publication:

Carbon-capture screening moves past the Henry constant: a neural classical density functional computes full-loading isotherms for over 5,000 MOFs

Synopsis

Adsorption isotherms for carbon-capture materials previously required grand canonical molecular simulations combined with accurate ab initio treatment of molecular interactions, which is so costly that screening often fell back to the low-pressure uptake encoded in the Henry constant. A neural classical density functional framework bypasses grand canonical simulation and computes isotherms directly from ab initio potential energy surfaces, from low to high loading, for metal-organic frameworks. It was applied to more than 5,000 experimentally reported MOFs. The resulting isotherms fall into four shapes set by pore size and binding sites, which the Henry constant cannot distinguish, and these shapes rather than binding strength alone decide which process a MOF suits.

Source-provided article image: Characterizing adsorption thermodynamics in carbon capture materials: an efficient first-principles approach
Figure 1 · arXiv · Page 2

Interpretation

Adsorption thermodynamics for carbon-capture materials can now be computed as full isotherms directly from ab initio potential energy surfaces, covering the range from low to high loading without running grand canonical molecular simulations. The prior route combined accurate ab initio interactions with grand canonical molecular simulation to obtain isotherms and heats of adsorption, at very high computational cost, so screening was limited to the low-pressure uptake encoded in the Henry constant. The capability was demonstrated on more than 5,000 experimentally reported metal-organic frameworks, for which ab initio adsorption isotherms were calculated.

The isotherms fall into four shapes, set by pore size and binding sites, a distinction the Henry constant cannot make. The Henry constant records only low-pressure uptake and cannot separate isotherm shapes arising from pore size and binding sites. The four shapes come from screening this set of more than 5,000 MOFs and map onto differences in pore size and binding sites.

Isotherm shape, not binding strength alone, decides which process fits a material: small pores with strong binding sites suit vacuum and temperature swings at low pressure, large pores with weakly binding walls suit pressure swings. Ranking by binding strength alone does not say which separation process a material should be paired with. The mapping rests on matching the four isotherm shapes to the two process conditions.

The most promising candidates are identified by weighing working capacity against the energy needed for regeneration. Uptake capacity alone does not establish a material's value in a working cycle. Candidate ranking counts both working capacity and regeneration energy.

Perspective

The demonstration covers ab initio adsorption isotherms for more than 5,000 experimentally reported metal-organic frameworks and four isotherm shapes derived from pore size and binding sites; the material states no next step.

The material is presented as a summary; the specific numbers and validation details behind the mapping of the four isotherm shapes to processes, and behind the working-capacity versus regeneration-energy trade-off, are not laid out in the visible scope.