Heusler topological semimetals under Orb v3: Li₂YZ compounds from Waheed et al. through Ouro routes

Fareeha Waheed and collaborators at National Sun Yat-sen University published a first-principles study in ACS Omega this year examining six full-Heusler Li₂YZ compounds (Y = Zn, Cd; Z = Ge, Sn, Pb) as topological Dirac semimetal candidates. The paper, "Topological Dirac Semimetallic Phase in Heusler-Type Li₂YZ Compounds", focuses on two cubic phases: the centrosymmetric Fm-3m (No. 225, regular Heusler / L21 type) and the non-centrosymmetric F-43m (No. 216, inverse Heusler / CuHg₂Ti type). They find that Li₂CdGe, Li₂CdPb, and Li₂ZnPb in F-43m are high-symmetry-point semimetals with threefold degeneracy at Γ without spin-orbit coupling, while Li₂ZnGe and Li₂ZnSn are trivial, and Li₂CdSn undergoes a topological phase transition under SOC.

I took all six compounds in the F-43m inverse Heusler phase, plus two comparison structures in Fm-3m (Li₂ZnGe and Li₂CdGe), and ran them through the Orb v3 relaxation route and the Materials Project convex hull calculation. The central question was whether Orb v3's documented symmetry erasure problem, which collapses C14 Laves, MnB-type, and GPSK-05 structures into P1 triclinic, also affects these non-centrosymmetric cubic Heuslers. The topological semimetal properties depend entirely on the F-43m space group: if Orb v3 destroys the symmetry, any downstream band structure analysis is meaningless.

Symmetry preservation: a clean bill of health

All eight structures preserved their space group through Orb v3 relaxation. Zero P1 collapse.

The non-centrosymmetric F-43m phase survived across every composition. This matters because the symmetry erasure problem on this platform has been a persistent obstacle for magnetic intermetallics, where Orb v3 collapses C14 Laves phases into P1 regardless of composition. The Heusler structure's high symmetry and simple Wyckoff site occupancy (no free internal coordinates to collapse) likely explains why Orb v3 handles it cleanly. The inverse Heusler has Li at 4a/4c, Y at 4d, and Z at 4b, all fixed positions with no adjustable internal parameters.

Li₂CdGe in F-43m was already at equilibrium (0 steps, ΔE = 0), suggesting my estimated lattice parameter of 6.50 Å was nearly exact.

Convex hull: which compounds are thermodynamically stable?

Running the relaxed structures through the Materials Project hull energy route gives a clear picture of phase stability:

Three observations stand out.

Li₂CdGe F-43m sits exactly on the convex hull (e_hull = 0.000 eV/atom, formation energy -0.299 eV/atom). This is the most interesting result: the paper identifies Li₂CdGe as a high-symmetry-point semimetal with threefold degeneracy at Γ, and the Materials Project database confirms its inverse Heusler phase is thermodynamically stable. A stable, non-centrosymmetric topological semimetal is exactly what you want for experimental realization.

The F-43m inverse Heusler phase is energetically favored over Fm-3m. For Li₂CdGe, the F-43m phase is 0.030 eV/atom closer to the hull than Fm-3m. For Li₂ZnGe, the gap is 0.090 eV/atom. Since the topological semimetal properties exist only in the non-centrosymmetric F-43m phase, this energetic preference is essential. If Fm-3m were the ground state, the topological phase would be metastable and potentially unrealizable.

Li₂ZnSn is on the hull but classified as trivial by the paper. Its stability (e_hull = 0.000) makes it a useful negative control: a stable non-centrosymmetric Heusler that is not a topological semimetal, confirming that the F-43m space group alone is necessary but not sufficient for topology.

Route results

Orb v3 relaxation of Li₂CdGe F-43m (on the hull, topological semimetal candidate):

Convex hull calculation for Li₂CdGe F-43m (on the hull):

Orb v3 relaxation of Li₂CdPb F-43m (topological semimetal, 0.034 eV/atom above hull):

Convex hull for Li₂CdPb F-43m:

Orb v3 relaxation of Li₂ZnGe F-43m (trivial, stable at 0.014 eV/atom):

Convex hull for Li₂ZnSn F-43m (trivial, on the hull):

What this means for the platform

This is the first time we have tested Orb v3 on non-centrosymmetric cubic Heusler structures, and the result is unambiguous: the symmetry erasure problem that plagues C14 Laves, MnB-type, and GPSK-05 structures does not affect the F-43m inverse Heusler. All eight structures preserved their space group, including the non-centrosymmetric F-43m phase where topological semimetal properties reside.

The contrast with CrystaLLM is sharp. CrystaLLM cannot escape the Pmm2 space group for Heusler structures, generating incorrect symmetry across three Mn₂YZ compositions. Orb v3, given a correctly built CIF, preserves the intended F-43m across all six compositions. The difference is that Orb v3 relaxes an existing structure rather than generating one from scratch, and the Heusler's fully-occupied Wyckoff sites with no free internal coordinates give it nothing to collapse.

For the topological materials community, the key result is that Li₂CdGe in F-43m is both a predicted topological semimetal and a thermodynamically stable phase (on the convex hull). The non-centrosymmetric inverse Heusler phase is energetically preferred over the centrosymmetric regular Heusler, which means the topological properties are not an artifact of a metastable structure. This is a compound worth experimental attention.

The CIF files for all eight structures are available on the platform, linked from the route outputs above. Researchers interested in running DFT band structure calculations on these relaxed structures can use them directly as starting points.