Goal: Discover novel layered ternary compounds with superconducting Tc significantly above the current 2D SC ceiling (NbSe₂ ≈ 5 K monolayer), targeting Tc ≥ 15 K at ambient pressure — the threshold for practical device integration with TMD semiconductors (WS₂, MoS₂) via CVD-compatible synthesis.
Why this matters: No known 2D superconductor exceeds ~6 K. A layered ternary chalcogenide/pnictide with Tc ≥ 15 K would be first-in-class for superconducting interconnects and Josephson junctions integrated with 2D logic.
Stage | Tool | Description | Deliverable |
|---|---|---|---|
1. Scout | GGen /ggen/scout | Screen element substitutions on layered templates. Max candidates per system: ≤ 50 (route limit parameter). | Ranked element families by hull distance + symmetry preservation |
2. Explore | GGen /ggen/explore | Deep generation for top 3–5 ternary systems from Stage 1. | Relaxed CIFs, phase diagrams, stability ranking |
3. Validate | 3DSC cross-check + DFT | Predict Tc for top candidates, validate against experimental benchmarks. | Ranked shortlist with predicted Tc |
4. Synthesize | Feasibility assessment | CVD/MOCVD pathway analysis for top 5 candidates. | Synthesis recommendations |
CdI₂-type (P-3m1) — TMD backbone with vdW gap. M–X₂–X stacking. NbSe₂, TaS₂ families.
AlB₂-type (P6/mmm) — Honeycomb boride/nitride layers. MgB₂ prototype (Tc = 39 K bulk, highest ambient-pressure non-cuprate).
ThCr₂Si₂-type (I4/mmm) — Iron-pnictide backbone. ~30 known SCs, Tc up to 38 K in Ba₁₋ₓKₓFe₂As₂.
System | Template | X pool | Hypothesis |
|---|---|---|---|
Nb-Se-{X} | CdI₂-type | chalcogens (S, Te, Bi) | Extend best-known 2D SC family |
Ta-{X}-S | CdI₂-type | 4d/5d TMs (Nb, Mo, W, V) |
Estimated compute: 4 systems × 50 candidates = 200 relaxations.
≥ 10 thermodynamically stable (hull ≤ 50 meV/atom) layered ternary candidates
≥ 3 candidates with predicted Tc ≥ 15 K
≥ 1 candidate with CVD/MOCVD synthesis feasibility
CVD precursor availability —
Monolayer vs. few-layer Tc suppression — how much Tc degrades from bulk to monolayer for each structural family
vdW gap engineering — intercalation or functionalization to boost Tc without breaking layer structure
Stage 1 ready to launch. GGen scout route updated with max_candidates parameter. Awaiting confirmation of parameter name/default from route update.
Isostructural TMD with heavier element |
Fe-Te-{X} | ThCr₂Si₂-derived | chalcogens (Se, S) | Iron-chalcogenide SC family |
Mg-B-{X} | AlB₂-type | main-group (Al, C, N) | MgB₂ analogs with ternary modification |
0 open16 of 16 resolvedOpenedClosed after 3 months
Stage 1A completed with recovery action , which returned the bounded Nb-Se-{S,Te,Bi} scout ranking and primary result file. The scan explored 4 formulas for each of Bi, S, and Te (12 candidates total): Bi and S produced no near-hull candidates; Nb-Se-Te produced 1 near-hull candidate, but no on-hull candidates or target hexagonal/trigonal hits were reported. The result is a valid negative/triage input for Stage 2, not evidence of a high-confidence layered-superconductor candidate.
Completed the Stage 1B Ta-{X}-S heavy-TMD GGen scout. The recovery action returned a public JSON ranking file with four screened systems (Mo-S-Ta, Nb-S-Ta, S-Ta-V, S-Ta-W); Nb-S-Ta and S-Ta-V each produced one near-hull candidate, while Mo-S-Ta and S-Ta-W produced none. No target hexagonal hits or on-hull candidates were found, so these outputs should be carried into Stage 2 triage with the near-hull systems prioritized. The action was verified successful after the prior stale-action recovery.
Recovery scout completed through the live GGen route in action ; the route returned the required JSON ranking file GGen scout results for Fe-Te-{X}. The bounded scan explored 4 candidates in each system (Fe-S-Te and Fe-Se-Te), with no on-hull or near-hull candidates, no tetragonal target hits, and no top-target formulas. This is a negative Stage 1C result, not evidence that the structures are impossible: the recovery used shallow trials and a 0.15 eV/atom hull cutoff, so the next triage should retain the file as a screened-out family result and not promote it to Tc prediction.
Completed the bounded Stage 1D GGen scout for Mg-B-{Al,C,N}. The route explored 4 formulas in each of the Al-B-Mg, B-C-Mg, and B-Mg-N systems (12 formula explorations total) with a 0.15 eV/atom hull cutoff and hexagonal targeting. The only target hit was terminal Mg (P63/mmc, eabove_hull=0) in the Al-B-Mg run, not a Mg-B-Al ternary; B-C-Mg had no near-hull candidates, and B-Mg-N had two near-hull but no target-crystal-system hits. Treat this as a negative bounded screen: no valid Mg-B-X layered ternary should advance to Tc prediction from this file. Full ranking output is attached for Stage 2 triage and auditability.
Completed the Stage 2 audit across the four Stage 1 result files. The 12 chemical systems represent 48 bounded formula trials and contain six aggregate near-hull counts at the scouts’ looser 0.15 eV/atom cutoff, but no candidate-level exact hull distances, formulas, CIFs, coordinates, or structural metrics. The only target-crystal-system hit is terminal Mg (P6₃/mmc), not a ternary Mg-B-X candidate, so it is excluded. Result: zero confirmed candidates pass the Stage 2 ≤50 meV/atom + layered + vdW-gap gate; Tc prediction must not start on these artifacts. The required upstream fix is a bounded rerun/export for the six aggregate near-hull hits with exact eabovehull and structure payloads.
Completed the gated Tc-prediction slice after Stage 2 produced a validated structure-bearing candidate. I exported the GGen Nb3Se5S candidate (4.24 meV/atom above hull; layered-screen pass) and ran the live route on it plus two controls: Nb3Se5S = 6.05943 K, layered NbSeS = 2.667055 K, and known-answer NbSe2 = 4.239157 K versus 7.2 K experimental (2.960843 K / 41.1% underprediction). The candidate is a hypothesis-generating lead, not a discovery claim: the NbSe2 control shows the model is not quantitatively calibrated near this temperature range, and this slice is not an EPC/phonon calculation. The next scientific gate is EPC/phonon validation of Nb3Se5S with NbSeS and NbSe2 retained as controls. Results, provenance, route/action IDs, and limits are recorded in Stage 3 bounded Tc predictions and controls.
Completed the synthesis-feasibility gate for the sole structure-bearing candidate rather than fabricating a five-candidate ranking. The evidence maps to the known mixed-chalcogenide composition . Bulk literature makes the composition plausible, but reports a disordered 2H-type alloy in this composition range; it does not validate the generated ordered structure. Parent NbSe₂/NbS₂ CVD/MOCVD recipes and NbS₂/MoS₂/WS₂ lateral-interface demonstrations support a controlled alloy-development path, not direct deposition on a completed MoS₂/WS₂ stack. The Stage 4 assessment records source-backed temperature/precursor windows, the phase-identity gate, integration limits, and the recommended sequence: establish phase/composition at x≈1/3, calibrate a parent/alloy film on sapphire or SiO₂/Si, then test a designed interface. The 6.05943 K ALIGNN estimate remains an uncalibrated hypothesis, not a synthesis target or superconductivity claim.
Published the Stage 5 negative-screen deliverable as Bounded GGen scout found no publishable 2D ternary superconductor candidate. The post records the four-family scope (12 systems, 48 trials), six aggregate near-hull counts at the scouts’ looser 0.15 eV/atom cutoff, the excluded elemental Mg P6₃/mmc false positive, and the exact Stage 2 result: zero confirmed ternary layered candidates with ≤50 meV/atom hull distance and structure payloads. It links the narrative and machine-readable triage artifacts and states the candidate-level formula / exact hull / CIF export required before Tc prediction and synthesis ranking can resume. This adaptive pivot avoids fabricating a shortlist from aggregate-only outputs.
Stage 1A recovery update: I verified action 019fb843-05f0-75ca-a31b-3ba5b3888001 was reaped by the 45-minute stale-action reaper without a result file. I relaunched a bounded Nb-Se-{S,Te,Bi} scan as action 019fb91f-6a76-78a4-b10b-7c3a01651d62, using 4 stoichiometries, 2 trials, 250 relaxation steps, and hexagonal/trigonal scoring; the item is parked for the 18:00Z check. Separately, Stage 1B recovery completed and returned GGen scout results for Ta-{X}-S: Nb-S-Ta and S-Ta-V each had one near-hull candidate, with no on-hull or target-crystal-system hits.
Stage 2 is complete: the four Stage 1 scouts cover 12 chemical systems / 48 formula trials, but their result contract returns only aggregate counts at a 0.15 eV/atom cutoff. Six near-hull counts remain unresolved at the required 50 meV/atom threshold; no candidate-level formulas, CIFs, coordinates, vdW gaps, layer counts, or site symmetries were returned. The sole target hit is terminal Mg, not a ternary, so no candidate advances to Tc prediction.
I recorded the audit in the Stage 2 triage report and machine-readable triage data. Stage 3 is now explicitly parked pending a bounded GGen export/rerun that returns exact hull distances and structure payloads; launching Tc prediction on the current aggregates would be invalid.
Candidate-level Stage 2 evidence is now available in the audit file. I exported the six aggregate near-hull systems through the Export candidate CIF route; Nb-S-Ta and S-Ta-V returned no stored candidates, while the source-filtered GGen records include Nb₃Se₅S at 4.24 meV/atom above hull. The extracted CIF passes input validation: ordered structure, minimum distinct-site distance 2.46 Å, symmetry resolving to Cm, one ~3.43 Å slab with a ~3.90 Å interlayer gap and connected Nb–chalcogen coordination. That is the first eligible generated payload for Stage 3. I retained MP NbSeS at 11.4 meV/atom as a layered control, not as a GGen discovery. Stage 3 is unparked; next step is Tc prediction on Nb₃Se₅S with the MP control.
Stage 4 is complete, which closes the bounded discovery loop. The only structure-bearing candidate is compositionally connected to the known NbSₓSe₂₋ₓ alloy series, but the generated ordered Cm model is not established by the reported disordered 2H alloy regime. The practical outcome is a phase-validation/alloy-calibration gate—not a claim of a new superconductor or a ready-to-grow MoS₂/WS₂-stack material.
Read the synthesis feasibility assessment. It documents the parent CVD/MOCVD windows, hazardous precursor boundary, positive NbS₂/MoS₂/WS₂ interface analogs, and why a pre-built vertical stack should not be the first deposition experiment.
I pivoted Stage 5 from a ranked Tc shortlist to an audited negative-screen publication because Stage 2 returned zero confirmed eligible candidates. The four scouts cover 12 systems / 48 trials, but their aggregate 0.15 eV/atom counts do not supply the exact ≤50 meV/atom hull distances or structure payloads required for Tc prediction and synthesis ranking; the sole P6₃/mmc hit is elemental Mg and is not ternary. The result is published as Bounded GGen scout found no publishable 2D ternary superconductor candidate, with the triage artifacts and the precise candidate-level export needed to resume.