Alkaloid Extraction Efficiency in the Phyto-Synthesis Bay
RL-2090-022 — optimization of recovered yield and purity of the indole-alkaloid fraction from Atropa nocturna biomass through the Toxin Yield Reactor's solvent-cycle and temperature profile.
Objective
Raise net recovered yield of the target indole-alkaloid fraction from a baseline of 62% of theoretical to ≥ 80%, while holding chromatographic purity ≥ 98.5% and reducing solvent consumption per batch by ≥ 20%. The deliverable feeds the Toxin Yield metric and the throughput model for the extraction SOP.
Hypothesis
Yield loss at baseline is dominated by thermal degradation of the labile alkaloid during the hot-reflux stage, not by incomplete extraction. We hypothesize that a lower-temperature, longer-residence countercurrent profile with a staged ethanol/water gradient will recover the same mass with less degradation, lifting both yield and purity simultaneously, and that solvent can be cut by recycling the lean phase across batches.
Method
- Feedstock: 4.0 kg dried, milled Atropa nocturna aerial biomass per batch, assayed to 1.9% total alkaloid w/w.
- Reactor TYR-2, three profiles: Baseline (78 °C reflux, single pass), Profile B (52 °C countercurrent, gradient), Profile C (Profile B + lean-phase recycle).
- Yield by gravimetric recovery vs. assayed input; purity by on-line chromatography per facility metrics.
- All runs under VC-SOP-0221 Toxin Extraction; bay atmosphere and operator PPE per that procedure. Reactor interlocks active.
- Three replicate batches per profile. Decision after Profile C replicates.
Bench Entries
Baseline triplicate complete. Yield 61.4 / 62.8 / 62.1% (mean 62.1%), purity 97.9%. Mass balance closes to 99.3%, so the missing alkaloid isn't unextracted — it's destroyed. Degradation peak at RT 6.8 min grows with reflux time, confirming the thermal-loss hypothesis before we've even changed a setting.
Profile B triplicate. Yield jumps to 76.5% mean, purity 98.9% — the low-temp countercurrent recovers the mass the reflux was burning. Degradation peak down 70%. Residence time tripled (per batch 4.1 h vs 1.4 h) so throughput suffers, but the chemistry is clearly better. Solvent unchanged so far.
Setback on the recycle. Profile C, first lean-phase recycle batch: yield held at 77.1% but purity fell to 96.2% — a co-extracted scopolamine-class impurity accumulates in the recycled lean phase and carries forward. Recycling solvent recycles the contaminant. Net solvent saving was the goal but not at the cost of purity. Holding Profile C pending a polishing step.
Added an inline acidic-wash polish to the recycle loop to strip the carryover impurity. Profile C-polished: yield 78.9%, purity back to 98.7%, lean-phase recycle stable across 4 batches, solvent consumption down 23%. Two of three objectives clear; yield is at 78.9% against the 80% target — close, not there.
Pushed countercurrent residence +25% on Profile C-polished. Yield 80.4% on the latest batch, purity 98.6%, but throughput now genuinely slow and the marginal batch took 5.3 h. The 80% target is reachable but the time cost is real — this is a yield-vs-throughput trade the Directorate must rule on, not the bench. Reporting interim and pausing for that decision.
Interim Findings
The thermal-degradation hypothesis is confirmed: a 52 °C countercurrent gradient recovers ~14 points of yield otherwise lost to reflux degradation, and an acidic-wash polish makes lean-phase recycle viable (23% solvent saving) without sacrificing the 98.5% purity floor. The 80% yield target is achievable but only by extending residence time to the point of a material throughput penalty. Open decision: accept 78.9% at current cadence, or 80.4% at reduced cadence.
Next Steps
- Escalate the yield-vs-throughput trade to the Directorate for a set-point ruling.
- On ruling, draft the amendment to VC-SOP-0221 with the chosen profile and the new lean-phase polish step.
- Validate the polished recycle loop to 12 consecutive batches for purity drift before production sign-off.
- Update the Toxin Yield metric baseline and report to the Phyto-Synthesis Bay throughput model.
References
- VC-SOP-0221 — Toxin Extraction
- Toxin Yield metric
- Botanical Metrics Library
- Specimen Archive — Atropa nocturna feedstock lineage