{
 "schema": "rtlab.recompute.v1",
 "id": "srnh4b6c6-v2",
 "date": "2026-09-24",
 "cell": "SrNH4B6C6 (Fmmm), started from the 50 Ry Gamma-job relaxation and relaxed again at these settings",
 "settings": {
  "functional": "LDA (Perdew-Zunger), same pseudopotentials as the first run",
  "ecutwfc_ry": 50,
  "ecutrho_ry": 400,
  "k_mesh_scf": [
   4,
   4,
   4
  ],
  "k_mesh_fit": [
   8,
   8,
   8
  ],
  "q_mesh": [
   2,
   2,
   2
  ],
  "tr2_ph": 1e-14
 },
 "why": "The first alpha2F (40 Ry, tr2_ph 1e-11) ran on a cell under -41 kbar and was withdrawn on 2026-09-23. This recompute fixes both: the relaxation is re-run until the recomputed stress is small, and the phonon threshold is tight (at 1e-11 the same relaxed cell showed three spurious -6 to -8 THz modes at Gamma that vanish at 1e-14).",
 "relaxation": {
  "ecutwfc_ry": 50.0,
  "pressure_kbar_first": -24.05,
  "pressure_kbar_final_recomputed": -1.99,
  "passes": true
 },
 "phonons_at_computed_q_points": [
  {
   "q_point": "q1",
   "q_cartesian_2pi_over_a": [
    0.0,
    0.0,
    0.0
   ],
   "lowest_three_thz": [
    0.958594,
    0.983269,
    1.007442
   ],
   "modes_below_tolerance": 0,
   "n_modes": 108
  },
  {
   "q_point": "q2",
   "q_cartesian_2pi_over_a": [
    -0.3534869,
    -0.353553,
    0.3535538
   ],
   "lowest_three_thz": [
    4.497369,
    4.529966,
    4.697017
   ],
   "modes_below_tolerance": 0,
   "n_modes": 108
  },
  {
   "q_point": "q3",
   "q_cartesian_2pi_over_a": [
    -0.7069737,
    0.0,
    0.0
   ],
   "lowest_three_thz": [
    -2.020429,
    2.809495,
    4.360702
   ],
   "modes_below_tolerance": 1,
   "n_modes": 108
  },
  {
   "q_point": "q4",
   "q_cartesian_2pi_over_a": [
    0.0,
    -0.7071059,
    0.0
   ],
   "lowest_three_thz": [
    -2.06148,
    2.85717,
    4.291077
   ],
   "modes_below_tolerance": 1,
   "n_modes": 108
  },
  {
   "q_point": "q5",
   "q_cartesian_2pi_over_a": [
    0.0,
    0.0,
    -0.7071076
   ],
   "lowest_three_thz": [
    -2.342622,
    2.488564,
    4.404309
   ],
   "modes_below_tolerance": 1,
   "n_modes": 108
  }
 ],
 "gamma_after_sum_rule": [
  -0.0,
  -0.0,
  0.0,
  3.3253,
  3.3715,
  3.4697
 ],
 "qe_lambda_table": [
  {
   "sigma_ry": 0.005,
   "lambda_qe": 2.9119,
   "dos_ef": 42.8484,
   "omega_ln_k": 101.068
  },
  {
   "sigma_ry": 0.01,
   "lambda_qe": 3.3772,
   "dos_ef": 38.9714,
   "omega_ln_k": 80.9285
  },
  {
   "sigma_ry": 0.015,
   "lambda_qe": 3.3493,
   "dos_ef": 35.7116,
   "omega_ln_k": 73.7483
  },
  {
   "sigma_ry": 0.02,
   "lambda_qe": 3.3032,
   "dos_ef": 34.1328,
   "omega_ln_k": 69.3206
  },
  {
   "sigma_ry": 0.025,
   "lambda_qe": 3.2755,
   "dos_ef": 33.5248,
   "omega_ln_k": 66.9252
  },
  {
   "sigma_ry": 0.03,
   "lambda_qe": 3.2659,
   "dos_ef": 33.3904,
   "omega_ln_k": 65.8683
  },
  {
   "sigma_ry": 0.035,
   "lambda_qe": 3.2693,
   "dos_ef": 33.5125,
   "omega_ln_k": 65.7058
  },
  {
   "sigma_ry": 0.04,
   "lambda_qe": 3.2756,
   "dos_ef": 33.7591,
   "omega_ln_k": 66.1387
  },
  {
   "sigma_ry": 0.045,
   "lambda_qe": 3.2772,
   "dos_ef": 34.054,
   "omega_ln_k": 66.9315
  },
  {
   "sigma_ry": 0.05,
   "lambda_qe": 3.2701,
   "dos_ef": 34.3543,
   "omega_ln_k": 67.9185
  }
 ],
 "reading": "Relaxed (recomputed stress -2 kbar) and stable at Gamma and at the second computed q-point, but each of the three computed q-points along the Cartesian axes (at the zone boundary) has exactly one imaginary mode, -2.0, -2.1 and -2.3 THz: one soft branch, the same in all three directions of the nearly cubic cell. At this level of theory (LDA, 2x2x2 q) the Fmmm cell is weakly unstable at the zone boundary, where the paper (a different functional and code) found it stable. The coupling table confirms it: lambda about 3.3 with omega_ln about 66 K, the signature of a soft branch dominating the coupling (lambda goes as 1/omega^2), not of strong pairing. No Tc and no hydrogen-share verdict can be read from this cell; the next step would be to follow the soft mode to the distorted cell and compute that.",
 "valid_for_verdict": false,
 "compute": {
  "where": "Hugging Face Jobs cpu-performance (32 ranks)",
  "phonon_wall": "q1 1h50, q2 5h12, q3-q5 about 1h55 each; plus a 1h15 Gamma run at tr2_ph 1e-11 that showed the spurious modes",
  "approx_cost_usd": 27
 }
}