VUV-AI automated isomer ID
Bell-method spectral matching → isomer prior → tier-α property prediction.
How it works
- VUV spectrum per GC×GC peak (125-240 nm absorption fingerprint)
- Match against NIST spectral library using Bell 2022 seven-moment statistic
- Threshold ≥ 5 of 7 moments pass → confident ID; ambiguous matches kept as constrained alternates
- Identified isomers replace uniform random draw in tier-α Monte Carlo sampling
Live: H125 (AC-Fuels ATJ) Bell-accepted IDs
Top 25 Bell-accepted IDs by R²; production pipeline output, ≥ 5/7 moments pass.
| RT (min) | C# | Class | Identified isomer | Formula | Score | R² | Alts |
|---|---|---|---|---|---|---|---|
| 56.3 | C10 | iso-Paraffins | 3,3-Dimethyloctane | C10H22 | 5/7 | 0.569 | 5 |
| 58.17 | C? | unknown | 3-Methylheptane | C8H18 | 5/7 | 0.543 | 5 |
| 77.98 | C11 | Monocycloparaf | tert-Butylcyclohexane | C10H20 | 5/7 | 0.522 | 5 |
| 104.44 | C13 | iso-Paraffins | Dodecane | C12H26 | 5/7 | 0.501 | 3 |
| 120.59 | C14 | iso-Paraffins | Tetradecane | C14H30 | 5/7 | 0.495 | 3 |
| 68.29 | C8 | Alkylbenzenes | Ethylbenzene | C8H10 | 5/7 | 0.482 | 5 |
| 70.06 | C10 | Monocycloparaf | tert-Butylcyclohexane | C10H20 | 5/7 | 0.475 | 5 |
| 36.73 | C? | unknown | 5-Methylnonane | C10H22 | 5/7 | 0.437 | 5 |
| 59.06 | C9 | Dicycloparaffi | Trimethylcyclohexane isomer | C9H18 | 5/7 | 0.414 | 5 |
| 73.93 | C? | unknown | 1-Hexene | C6H12 | 5/7 | 0.411 | 5 |
| 72.29 | C12 | iso-Paraffins | Dodecane | C12H26 | 5/7 | 0.401 | 3 |
| 102.93 | C14 | iso-Paraffins | Tridecane | C13H28 | 5/7 | 0.351 | 3 |
| 26.22 | C7 | iso-Paraffins | 2,3,3-Trimethylpentane | C8H18 | 5/7 | 0.332 | 5 |
| 87.31 | C12 | Monocycloparaf | t-1-Methyl-2-(4MP)cyclopentane | C12H24 | 5/7 | 0.307 | 2 |
| 57.24 | C10 | iso-Paraffins | 3,3-Diethylpentane | C9H20 | 5/7 | 0.295 | 5 |
| 165.23 | C16 | Monocycloparaf | alpha-Bergamotene | C15H24 | 5/7 | 0.274 | 5 |
| 138.33 | C16 | iso-Paraffins | Pentadecane | C15H32 | 5/7 | 0.263 | 3 |
| 84.87 | C11 | Monocycloparaf | cis-Decalin | C10H18 | 5/7 | 0.259 | 5 |
| 110.32 | C14 | iso-Paraffins | Pentadecane | C15H32 | 5/7 | 0.253 | 3 |
| 163.62 | C17 | iso-Paraffins | Octadecane | C18H38 | 5/7 | 0.253 | 3 |
| 110.37 | C12 | Dicycloparaffi | t-1-Methyl-2-(4MP)cyclopentane | C12H24 | 5/7 | 0.249 | 2 |
| 67.46 | C11 | Monocycloparaf | tert-Butylcyclohexane | C10H20 | 5/7 | 0.232 | 5 |
| 146.22 | C16 | Monocycloparaf | alpha-Cedrene | C15H24 | 5/7 | 0.23 | 5 |
| 143.67 | C15 | Monocycloparaf | alpha-Cedrene | C15H24 | 5/7 | 0.226 | 5 |
| 46.0 | C? | unknown | alpha-Copaene, EE | C15H24 | 5/7 | 0.219 | 5 |
Across H125: 302 Bell-accepted IDs out of 1,126 peaks (27 %). Across H125-H128 ATJ samples: ~250-300 accepted IDs each.
WJFS-11 prior sensitivity (preview of what Bell IDs unlock)
WJFS-11 is the worst density over-predictor in the cohort (+9.89 kg/m³ baseline bias). Constraining iso-paraffin sampling to just the lower-quartile NIST densities — the kind of constraint a real Bell ID would impose — closes the bias by 54 %:
Combining iso-paraffin and alkylbenzene priors closes 63 % of the bias. Aromatics-only prior gives a smaller density benefit (~10 %) because the alkylbenzene NIST pool is already well-behaved.
Coming next
- SAK (H72, H78) — Bell pipeline rebuild in progress; isomer IDs expected this week
- WJFS-2024 cohort — full Bell-ID pass once H5 files reconstructed
- Real-time API — wire identified isomers directly into the slider