A computational study published in the Journal of Materials Science reports that a scandium-doped, single-layer carbon nitride framework — C12N8 — achieves a simulated CO2/N2 selectivity of 369 and a CO2/CH4 selectivity of 468 under dry gas conditions at ambient temperature. The figures place the material among the best-performing adsorbents modeled to date for the two largest separation problems in the energy economy: post-combustion flue gas, dominated by nitrogen, and natural gas upgrading, which hinges on stripping CO2 from methane.

The paper, authored by Tiantian Qiao and Sainan Zhou of Liaoning University together with Xiaoqing Lu and Yongqing Li of the China University of Petroleum, is a simulation, not a measurement. The team ran density functional theory calculations with dispersion corrections paired with grand canonical Monte Carlo simulations to generate adsorption isotherms and mixture selectivities at realistic temperatures and pressures. The DOI is 10.1007/s10853-026-13796-4.

Two levers, two different jobs

The study's central finding is that the two modification strategies available to this material class do fundamentally different work. Doping with scandium is the dominant performance lever. Strain is a secondary adjustment — and a double-edged one.

The mechanism is electrostatic. When a scandium atom substitutes into the C12N8 lattice, it acts as a strong cationic center, withdrawing electron density from its surroundings and sharply increasing the polarity of the nanosheet. CO2, with its linear structure and pronounced quadrupole moment, responds strongly to such electric field gradients. The doped scandium sites become primary adsorption anchors, gripping CO2 molecules far more tightly than pristine carbon nitride can. Nitrogen and methane lack the same charge distribution and feel a much weaker pull. That asymmetry produces the material's selectivity.

The choice of scandium is not arbitrary. The element sits among transition metals known to form strong bonds with nitrogen, and prior work on scandium nitride films and Sc-embedded graphyne has documented its affinity for nitrogen-rich lattices. Embedding it in a carbon nitride framework yields stable cationic sites rather than loosely bound decoration atoms that might cluster or leach.

Strain trades selectivity for capacity

Capacity tells a complementary story. When the Sc-doped nanosheet is stretched by five percent under tensile strain, simulated CO2 uptake reaches 8.19 millimoles per gram at 298 kelvin and 1 bar — conditions close to real post-combustion capture. Strain reshapes geometry rather than chemistry: tensile deformation flattens the lattice topology and enlarges accessible pore volume, giving CO2 molecules more room to pack inside the sheet.

But the study's most instructive result is what strain does not do — and what it costs. Applied to pristine C12N8, tensile strain alone delivers no improvement in CO2 uptake whatsoever. Without the cationic anchors, extra pore volume is simply not enough. In the doped system, strain boosts capacity at the expense of selectivity: the same flattening and pore expansion that admit more CO2 also relax the geometric and energetic discrimination that made the unstrained material a precise sieve.

The highest selectivities belong to the unstrained doped sheet. The highest capacity belongs to the five-percent-stretched one. Engineers designing a real capture process would have to decide which figure matters more for their application, or locate an intermediate strain that balances the two.

A modular design principle

This capacity-selectivity trade-off is a recurring theme in adsorbent design, and the study frames it with unusual molecular-level clarity. Strong electrostatic interactions between open metal sites and CO2 drive uptake and selectivity upward together, which is why scandium doping enhances both metrics. Strain, by contrast, is purely structural: it promotes CO2 packing but dilutes the selectivity advantage. The authors describe Sc-C12N8 as a promising two-dimensional molecular sieve, and the description fits — performance depends on pore architecture and pore chemistry working in concert, and the simulations show how each contribution can be dialed independently.

The work builds on a body of earlier computational research by the same groups showing that alkali, alkaline earth, and first-row transition metal dopants can transform graphyne and covalent organic frameworks into high-performance CO2 sorbents, with the dopant's charge and coordination environment governing how CO2 molecules arrange themselves in the pores.

From simulation to synthesis

Because every step is computational, the results are predictions rather than measurements. What they define is a concrete, testable target for synthetic chemists working on two-dimensional carbon nitrides — several of which have already been realized in the laboratory. The broader significance lies in the design principle: choose a stable, porous two-dimensional framework, install cationic adsorption sites by doping, then use lattice strain as a fine-tuning knob to trade capacity against selectivity as the application demands.

As carbon capture and storage moves from demonstration to deployment, materials engineered at this level of precision — a single atomic layer, a single dopant species, a few percent of strain — may prove essential to making separation technology both effective and energy-efficient. The scandium-doped C12N8 nanosheet remains a simulation today. But it is a simulation with a blueprint attached.

The milestone that decides what happens next is laboratory synthesis and experimental validation of the predicted isotherms — whether synthetic chemists can reproduce the doped lattice, apply controlled tensile strain, and confirm the 369/468 selectivity figures and the 8.19 mmol/g uptake at 1 bar in a measured material.