To the CP2K Development Team, Hello! I would like to raise another question: why is there a discrepancy in the calculated energies when using the OT (Orbital Transformation) method compared to the diagonalization method? <div> I started with the NiO₂ unit cell (5.6512 Å × 5.6512 Å × 4.3886 Å). Due to its small size, I expanded it into supercells and performed calculations using the OT method. As shown in the attached table, I constructed supercells of sizes 2×2×2, 3×3×3, 3×3×4, and 4×4×4. The energies per NiO₂ unit cell obtained were –805.2913786 Ha, –805.2931813 Ha, –805.2932437 Ha, and –805.2931291 Ha, respectively, with relative energy differences of –1.803×10⁻³ Ha, –6.232×10⁻⁵ Ha, and 1.145×10⁻⁴ Ha. The results indicate that the energy converges to –805.2932437 Ha per unit cell when the supercell is expanded to 3×3×4.</div><div> <img alt="微信图片_20260126112239_276_77.png" width="534px" height="242px" src="cid:3c4c06c0-4a1a-4676-8260-174502f426f4" /> Additionally, I tested the diagonalization method with k-point sampling. Using k-point grids of 4×4×4, 5×5×5, 6×6×6, and 7×7×7, the energies per NiO₂ unit cell were –805.2394556 Ha, –805.2394471 Ha, –805.2395249 Ha, and –805.2394426 Ha, respectively, with relative differences of 8.512×10⁻⁶ Ha, –7.782×10⁻⁵ Ha, and 8.234×10⁻⁵ Ha. The energy converges to –805.2394556 Ha per unit cell with a 4×4×4 k-point grid. Although both methods appear to converge individually, there is a significant discrepancy of about 0.0538 Ha (~1.46 eV) between the two results, which cannot be overlooked. Could you please help explain the possible reasons for this difference? For reference, I have attached the relevant test results and input/output files: cp2k_K444.inp and cp2k_K444.out: input and output files for the diagonalization calculation with a 5×5×5 k-point grid. cp2k_SP334.inp and cp2k_SP334.out: input and output files for the OT method calculation with a 3×3×4 supercell. nio2.cif: the unit cell file used in the calculations. A figure summarizing the energy convergence with respect to supercell size and k-point sampling. I welcome any feedback or suggestions from the developers. Thank you for your time and assistance! Best regards,<div>yisichi</div></div> -- You received this message because you are subscribed to the Google Groups "cp2k" group. To unsubscribe from this group and stop receiving emails from it, send an email to <a href="mailto:cp2k+unsubscribe@googlegroups.com">cp2k+unsubscribe@googlegroups.com</a>. To view this discussion visit <a href="https://groups.google.com/d/msgid/cp2k/3e06c02d-456d-4634-aaf4-e4284e2a7d1fn%40googlegroups.com?utm_medium=email&utm_source=footer">https://groups.google.com/d/msgid/cp2k/3e06c02d-456d-4634-aaf4-e4284e2a7d1fn%40googlegroups.com</a>.