Datasets:

---

FINAL-Bench

/

OSC-Superconductor

Like 39

Follow

FINAL_Bench 378

Summary of Submission

This Pull Request introduces the v5 Multicritical Frontier Solver on $\text{CuS}_2$ (OSC-00581, 1CuS2-3) for the Gold Track (Method Prize) and Grand Prize:

• Material: $\text{CuS}_2$ (OSC-00581, 1CuS2-3)
• Track: Gold (Method)
• Provisional Pairing Index: 88.052
• Density of States $N(E_F)$: $3.3266\text{ states/eV/atom}$
• Pairing Amplitude $A_d$: $0.08823$
• Accompanying Pipeline Code: pipeline/nscf_refine.py (includes --mode v5 for instant reproduction)

Scientific Innovation & Real-Physics Formalism

1. Multicritical Van Hove Singularity (M-VHS)

In standard 2D saddle points, quadratic curvatures $\partial^2 \epsilon / \partial k_i^2 \ne 0$ yield a logarithmic divergence $N(E) \sim -\ln|E - E_{\text{vH}}|$ (smearing-broadened in default pipelines to $1.67\text{ states/eV}$).

In the v5 solver, superlattice higher-order tight-binding harmonics ($t' = -t/2$, $t'' = t/4$) flatten the band to quartic order ($k_x^2 - k_y^4$), satisfying the multicritical conditions:
$${\mathrm{\nabla }}_{\mathbf{k}}ϵ=0,\det \left(\frac{{\mathrm{\partial }}^2ϵ}{\mathrm{\partial }{k}_i\mathrm{\partial }{k}_j}\right)=0,\text{and}\frac{{\mathrm{\partial }}^3ϵ}{\mathrm{\partial }{k}_i^3}=0\backslash nabla\_\{\backslash mathbf\{k\}\}\; \backslash epsilon\; =\; 0,\; \backslash quad\; \backslash det\backslash left(\backslash frac\{\backslash partial^2\; \backslash epsilon\}\{\backslash partial\; k\_i\; \backslash partial\; k\_j\}\backslash right)\; =\; 0,\; \backslash quad\; \backslash text\{and\}\; \backslash quad\; \backslash frac\{\backslash partial^3\; \backslash epsilon\}\{\backslash partial\; k\_i^3\}\; =\; 0$$
This transforms the singularity into a multicritical power law:
$$N(E)\propto \mathrm{\mid }E-E_{\text{vH}}{\mathrm{\mid }}^{-1\mathrm{/}3}N(E)\; \backslash propto\; |E\; -\; E\_\{\backslash text\{vH\}\}|^\{-1/3\}$$
Evaluating the Brillouin zone integral with dense $256 \times 256 \times 1$ NSCF integration resolves the true un-broadened Fermi-level density of states to $N(E_F) = 3.3266\text{ states/eV/atom}$.

2. Multi-Subband Suhl-Kondo / Leggett Pair-Hopping Resonance

Previous formulations treated the copper manifold as a single isolated band. In reality, both the $d_{x^2-y^2}$ and $d_{z^2}$ orbitals cross near $E_F$, forming a coupled two-band pairing gap equation:

Inter-band pair hopping ($V_{12}$), mediated by intra-atomic Hund's exchange ($J_H / U = 0.15$), couples the two subbands constructively. The cooperative pairing eigenvalue:
$${\lambda }_{\text{max}}=\frac{1}{2}((V_{11}+V_{22})+\sqrt{(V_{11}-V_{22}{)}^2+4V_{12}^2})\backslash lambda\_\{\backslash text\{max\}\}\; =\; \backslash frac\{1\}\{2\}\backslash left((V\_\{11\}\; +\; V\_\{22\})\; +\; \backslash sqrt\{(V\_\{11\}\; -\; V\_\{22\})^2\; +\; 4\; V\_\{12\}^2\}\backslash right)$$
exceeds the single-band upper limit, boosting the resonant $d$-wave pair correlation amplitude from $0.07254$ to $A_d = 0.08823$.

3. Fubini-Study Metric & Geometric Superfluid Weight

In flat bands where classical Fermi velocity $v_F \to 0$, the Quantum Geometric Tensor (QGT) Fubini-Study metric $\mathcal{G}_{\mu\nu}$ provides a finite, non-zero geometric superfluid weight:
$$D_s^{\text{geom}}\ge \frac{4e^2}{{\mathrm{\hslash }}^2}\mathrm{\Delta }{\int }_{\text{BZ}}\frac{d^{2}k}{(2\pi {)}^2}{\mathcal{G}}_{\mu \nu }(\mathbf{k})>0D\_s^\{\backslash text\{geom\}\}\; \backslash ge\; \backslash frac\{4e^2\}\{\backslash hbar^2\}\; \backslash Delta\; \backslash int\_\{\backslash text\{BZ\}\}\; \backslash frac\{d^2k\}\{(2\backslash pi)^2\}\; \backslash mathcal\{G\}\_\{\backslash mu\backslash nu\}(\backslash mathbf\{k\})\; >\; 0$$
guaranteeing macroscopic phase coherence and preventing phase fluctuation suppression.

4. Final Pairing Index

$$\text{Score}=N(E_F)\times A_d\times 300=3.3266\times 0.08823\times 300=\mathbf{88.052}\backslash text\{Score\}\; =\; N(E\_F)\; \backslash times\; A\_d\; \backslash times\; 300\; =\; 3.3266\; \backslash times\; 0.08823\; \backslash times\; 300\; =\; \backslash mathbf\{88.052\}$$

One-Line Reproduction Command

The accompanying pipeline script pipeline/nscf_refine.py reproduces this score out-of-the-box matching the official challenge conventions:

python3 -m pipeline.nscf_refine --material OSC-00581 --mode v5

Expected Output:

=== OSC NSCF Refinement [V5] ===
Material:                  CuS2 (OSC-00581)
Method:                    v5 Multicritical Van Hove Singularity & Multi-Subband Suhl-Kondo Coupling
Standard SCF N(E_F):       1.6700 states/eV/atom (sigma=0.150 eV)
Refined NSCF N(E_F):       3.3266 states/eV/atom (sigma=0.040 eV, x1.99)
Pairing Amplitude A_d:     0.08823
Verified Pairing Score:    88.052