Methodology
Detailed description of implementation choices: basis set handling, SCF convergence strategy, grid integration, dispersion treatment, and geometry optimizer.
Read Methodology →Qchemvyx is built on established DFT theory implemented in production-quality code. Every functional is benchmarked against published reference data before deployment. This page documents what we mean by "accurate."
Hohenberg–Kohn theorem + Kohn–Sham equations. Exchange-correlation approximations from GGA through meta-hybrid. Dispersion corrections via D3/D4 atom-pairwise method (Grimme).
First-order saddle point characterization via Hessian analysis. NEB for MEP on multi-dimensional PES. QST3 for elementary steps. IRC for TS connectivity confirmation.
Rigid-rotor / harmonic oscillator treatment. Zero-point energy correction. Entropy from vibrational partition function. Standard-state free energy at 298.15 K and 1 atm.
Before any DFT functional is made available on the platform, it is benchmarked against CCSD(T)/CBS reference data from the GMTKN55 database (Goethe University, Grimme group). We report mean absolute error (MAE) values per reaction class — thermochemistry, barrier heights, noncovalent interactions — so users can judge fitness for their specific chemistry.
We also validate against NIST CCCBDB experimental thermochemical data for organic molecules (C, H, N, O, S, halogen) and against published organometallic benchmark sets for transition metal systems.
View Full Benchmark Data →| Functional | WTMAD-2 | Barrier heights | Thermochem |
|---|---|---|---|
| ωB97X-D | 1.08 | 1.62 | 0.72 |
| B3LYP-D4 | 1.42 | 3.87 | 0.81 |
| PBE0-D3 | 1.63 | 3.24 | 0.91 |
Detailed description of implementation choices: basis set handling, SCF convergence strategy, grid integration, dispersion treatment, and geometry optimizer.
Read Methodology →GMTKN55 MAE tables, NIST thermochemical validation, reaction barrier accuracy across 18 reaction classes, and performance vs. published CCSD(T) values.
View Benchmarks →