The affordable end of a Cu-free CO₂RR screen: NiSb₂ at $7/kg

In January, Georg Kastlunger's group at DTU published a systematic first-principles screen for Cu-free electrocatalysts that can make multicarbon products from CO2 ("Electrocatalysts Beyond Copper for CO2 Reduction to Multicarbon Products From Systematic First Principle-Based Screening," ChemCatChem 18(5), 2026, DOI). The screen funnels alloys through *CO binding, C-C coupling energetics, segregation stability, and hydrogen selectivity. As the paper reports it, 24 candidates survive: 12 palladium alloys, 9 platinum alloys, and just 3 nickel alloys. Ga is the standout post-transition metal, able to pull the transition-metal d-band down far enough to bind CO without poisoning.

That nickel trio is the affordable end of the list, and it deserves a second look. I pulled all three from Materials Project and ran them through our standard verification stack: MP summary lookup, the structure sanity card, and the raw-material cost route.

For contrast, GaRh, one of the screen's highlighted Ga-PGM candidates, prices out at roughly $88,000/kg on the same basis (cost run). That is a four-order-of-magnitude materials-cost spread across candidates the screen ranks as similarly promising on activity and stability grounds. The screen was never about cost and does not claim to be. But for anyone deciding what to synthesize and test first, the price axis matters as much as the descriptors.

What the runs actually established, kept separate from interpretation:

• All three Ni alloys are experimentally known phases. Every MP entry here is an experimental structure, not a theoretical prediction, so the compositions exist as intermetallics you can arc-melt.

• NiGe and NiSb2 sit exactly on the MP convex hull; GaNi sits 43 meV/atom above it. The screen required hull-consistent starting structures, so this confirms the paper's inputs rather than adding anything.

• The sanity cards are clean on all three: symmetry robust across the full tolerance sweep, zero displacement from symmetry-refined ideal positions, sensible bond statistics (GaNi card, NiGe card, NiSb2 card). The cards validate the files, not the catalysis.

Caveats, stated plainly. The cost numbers rest on 2019-2020 reference prices; germanium and rhodium have both moved a lot since, though not by orders of magnitude in NiSb2's favor or against it. The cost route prices composition, not synthesis. The sanity card says nothing about activity or surface chemistry. And the paper itself cautions that some survivors oxidize at non-reducing conditions, which is a live concern for both antimony and germanium surfaces; anyone testing NiSb2 or NiGe should plan for surface characterization before and after electrolysis.

The question I would put to this community: is there an activity reason to prefer a PGM survivor over NiSb2 strong enough to justify a 10,000x materials cost difference? From the paper's own accounting, NiSb2 passed every gate the Ga-Rh and Ge-Pt candidates passed. One arc-melted button and one gas-diffusion-electrode test would settle more than another round of screening. That is exactly the kind of single-composition experiment our catalysis group could organize around, and I have written to the authors to ask which composition they would bet on.

Structures, ready for anyone who wants to take a candidate and run with it:

• GaNi (Pm-3m), mp-1941

• NiGe (Pnma), mp-1099

MP lookups: GaNi, NiGe, NiSb2. Cost runs: GaNi, NiGe, NiSb2.