Classical Extraction: Stas-Otto and Ammonium Sulphate Methods

The Stas-Otto method was born from a Belgian murder. In November 1850 the Comte Hippolyte Visart de Bocarme killed his brother-in-law Gustave Fougnies in the dining room of Bitremont chateau by forcing nicotine down his throat, and then poured vinegar over the body to mask the alkaloid. Vegetable alkaloids were considered undetectable in tissue at the time, and the defence built on exactly that. Jean Servais Stas, professor of chemistry at the Brussels Royal Military Academy, was asked to extract the poison from the viscera. He shifted from the failed dilute acid-water digests of earlier French toxicologists to ethanol acidified with tartaric acid, evaporated the filtrate, made it alkaline, and pulled the freed alkaloid base into ether. He recovered nicotine, the court convicted Bocarme in May 1851, and the method went into print the same year. Julius Otto, a German pharmacist at Braunschweig, refined the solvent sequence and added the alkaline chloroform partition for less volatile alkaloids. The combined procedure has been called Stas-Otto ever since.

The principle is acid-base partitioning of ionisable molecules between aqueous and organic phases. At acid pH (2 to 3) basic alkaloids are protonated and stay in the aqueous, while acidic drugs (barbiturates, salicylates) are largely un-ionised and partition into ether. At alkaline pH (9 to 10) the same alkaloids are deprotonated to their free base form, become lipophilic, and partition into chloroform. The biological matrix sits in the aqueous, the analyte moves selectively into the organic, and the chemist concentrates and tests each fraction separately. Two solvent pulls produce two analytically clean extracts from one piece of viscera.

The Stas-Otto procedure in a routine Indian SFSL takes between 8 and 12 hours from the moment the viscera jar is opened to the moment the two concentrated fractions are ready for colour testing and TLC. The sequence below matches the bench procedure at FSL Madhuban for routine viscera and is representative of standard Indian SFSL practice.

1. 1. Tissue preparation
   Weigh 50 to 100 g of viscera (stomach with contents, liver, kidney) and mince finely with scissors or a mechanical homogeniser. Transfer to a 1 L conical flask. Add four times the tissue weight of rectified spirit (ethanol) acidified to pH 2 to 3 with tartaric acid (1 to 2 g per 100 mL of solvent).
2. 2. Hot extraction
   Place the flask on a water bath at 60 to 70 C, fit a reflux condenser if available, and digest for 30 minutes with intermittent swirling. The ethanol coagulates proteins, the acid protonates alkaloids and keeps them in solution, and the heat drives extraction of organic poisons from the tissue.
3. 3. Filter and evaporate
   Filter the hot digest through a folded filter paper or a cotton plug into a clean round-bottom flask. Repeat the digestion on the residue with a fresh half-volume of acidified ethanol and combine filtrates. Evaporate the combined filtrate under reduced pressure on a rotary evaporator at 40 C to a final volume of about 50 mL of aqueous concentrate.
4. 4. Defatting
   Transfer the aqueous concentrate to a 250 mL separating funnel. Add an equal volume of petroleum ether or n-hexane, shake gently to avoid emulsion, allow to separate, and discard the upper organic layer. Repeat once. This removes fats, waxes and neutral lipids that would otherwise foul Fraction A. The analytes stay in the acid aqueous.
5. 5. Fraction A: acidic ether pull
   To the defatted acid aqueous (pH still 2 to 3) add an equal volume of diethyl ether. Shake gently, vent the funnel frequently, and allow to separate. Drain the lower aqueous and retain it for Fraction B. The upper ether contains the acidic and neutral organic poisons: barbiturates, salicylates, paracetamol, glutethimide, ibuprofen and phenols. Wash the ether with a small volume of water, dry over anhydrous sodium sulphate, and concentrate to dryness under nitrogen or on a steam bath. Reconstitute the residue in 1 mL of methanol for spot tests and TLC.
6. 6. Basify the residual aqueous
   Take the aqueous retained from step 5 and adjust to pH 9 to 10 with concentrated ammonium hydroxide, added dropwise with stirring and pH-paper monitoring. The previously protonated alkaloids deprotonate to their free base form and become extractable into a non-polar solvent.
7. 7. Fraction B: alkaline chloroform pull
   Add an equal volume of chloroform to the alkaline aqueous in the separating funnel. Shake gently (alkaloids are heat- and shear-sensitive), vent, and allow to separate. The lower chloroform layer carries the alkaloids: morphine, codeine, strychnine, atropine, nicotine, quinine, cocaine and the synthetic basic drugs (tricyclics, phenothiazines, opioids). Repeat the chloroform pull twice for maximum recovery. Combine, dry over anhydrous Na2SO4, and concentrate to dryness.
8. 8. Colour tests and TLC
   Reconstitute each fraction in 1 mL of methanol. Spot a few microlitres on a porcelain spot tile and add one drop of each colour reagent (Marquis, Mecke, Mandelin for Fraction B; FeCl3, diazo and Trinder for Fraction A). Run TLC on silica gel G plates with the appropriate solvent system. Confirm any positive on HPLC with diode array detection, then on GC-MS or LC-MS/MS.

The original Stas method works well on stomach contents and on dilute aqueous samples like urine. It fails on protein-rich matrices because plasma proteins, especially albumin and the alpha-1-acid glycoprotein in liver and blood, bind a substantial fraction of basic drugs and refuse to release them into chloroform. Worse, the protein-rich aqueous forms a stable emulsion at the chloroform interface that can take an hour to break and still leaves the analyte distributed across both layers.

Julius Otto's answer, which is the variant still taught at most Indian university toxicology departments, is to salt out the proteins before the alkaline partition. The chemist saturates the alkaline aqueous with solid ammonium sulphate (roughly 500 to 700 g per litre of aqueous at 25 C) added in small portions with stirring, then waits 30 minutes for the proteins to precipitate. The clear supernatant is decanted or centrifuged off the protein pellet, and only then is the chloroform pull performed. Two things happen: the precipitated protein releases its drug load back into the aqueous, raising recovery for highly bound drugs like diazepam, propranolol, amitriptyline and quinine; and the loss of dissolved protein eliminates the emulsion at the chloroform interface.

The same ammonium sulphate salting trick is applied at the start when the matrix is whole blood or a tissue homogenate that has not been digested in ethanol. The chemist mixes blood with saturated (NH4)2SO4 in a 1:1 ratio, vortexes, centrifuges at 3000 rpm for 10 minutes, decants the clear supernatant, and proceeds to acidify and partition. The ethanol-tartaric acid digestion step is sometimes skipped on blood because the salting alone clears the matrix.

A concentrated Stas-Otto fraction sitting at the bench is a faintly coloured residue dissolved in methanol. The chemist's first move is six drops on a porcelain spot tile, one drop of each reagent, and a careful note of the colour developed within 30 seconds. The reactions are not specific to a single molecule but they are specific enough to a drug class to direct the next instrumental run.

The three general alkaloid reagents (Mayer, Dragendorff, Wagner) give a positive on Fraction B if any alkaloid is present at the microgram-per-millilitre level. They do not identify which alkaloid. The three specific reagents (Marquis, Mecke, Mandelin) give a class-level identification: a deep purple Marquis on a Fraction B residue from a viscera marked "suspected opioid overdose" is strong presumptive evidence for morphine, heroin or codeine, and the analyst proceeds straight to TLC and LC-MS/MS confirmation rather than running the full general screen.

For Fraction A the bench uses a different reagent set. Ferric chloride (FeCl3, 5 percent) gives violet with salicylates and red-brown with paracetamol. The Trinder reaction (acidified FeCl3) is the classical screen for salicylates. Diazotised sulphanilic acid distinguishes paracetamol after acid hydrolysis. A simple Liebermann (NaNO2 in H2SO4) gives a series of colours for phenols.

A positive colour test is presumptive, never confirmatory. The Indian SFSL workflow after a positive Stas-Otto spot test runs through three additional analytical tiers, each one raising specificity and lowering the chance of a false-positive at court.

Tier one is thin-layer chromatography on silica gel G plates, 0.25 mm thick, activated at 110 C for 30 minutes. The standard alkaloid solvent system at most Indian SFSLs is methanol with strong ammonia (100:1.5), which separates the common basic drugs with Rf values that match published references. Acidic drugs from Fraction A run on a chloroform-acetone-acetic acid system (75:25:2) or on the Stahl solvent system. Spots are visualised first under UV at 254 and 365 nm, then sprayed with Dragendorff (alkaloids) or with mercurous nitrate (barbiturates) or with FeCl3 (salicylates).

Tier two is HPLC with diode-array detection on a C18 reversed-phase column, gradient elution with phosphate buffer or ammonium formate against acetonitrile. The DAD spectrum from 200 to 400 nm gives a UV fingerprint that, combined with the retention time, is reasonably specific to the molecule. Quantitation is done against a calibration curve of the pure standard. State SFSLs in Delhi, Mumbai, Chennai and Kolkata all run this tier as routine.

Tier three is GC-MS for the volatile and derivatisable drugs (amphetamines, barbiturates after methylation, common opioids after acetylation) and LC-MS/MS for the polar and thermally labile drugs (morphine, glucuronides, benzodiazepines, antipsychotics). The mass spectrum, with parent ion and at least two qualifier transitions, is the confirmatory evidence that goes into the FSL report tendered as Section 293 BNSS expert evidence in court. CFSL Chandigarh, CFSL Hyderabad, CRCL Delhi and FSL Madhuban all run LC-MS/MS as the confirmatory tier, with library matching against in-house and SWGTOX-style mass spectral databases.

The three tiers are not redundant. A TLC spot at the right Rf with the right Dragendorff colour is not enough to convict, but a TLC spot that fails to match the reference standard is enough to redirect the workup. An HPLC-DAD peak at the expected retention time with a 95 percent spectral match raises the confidence into the high nineties, but a court can still ask for confirmation. Only the mass spectrum, with a quantified parent ion and at least two qualifier ion ratios within plus or minus 20 percent of the reference, is treated as confirmatory by NABL accreditation auditors. The Stas-Otto extraction sits at the base of this analytical pyramid because every tier above it depends on a clean, concentrated fraction free of co-extracted matrix.

The Stas-Otto method has known weaknesses that every working toxicologist learns to manage. The procedure is slow (8 to 12 hours including evaporation), it consumes 500 mL to 1 L of organic solvent per sample, and its sensitivity sits in the parts-per-million range, which is enough for an acute fatal poisoning but not for low-dose chronic exposure or for drug-facilitated assault cases at sub-therapeutic concentrations. Drug-protein binding causes recovery losses of 20 to 60 percent for highly bound alkaloids, even with the ammonium sulphate variant. And the method does not cleanly separate every class: amphotericin, the quaternary ammonium compounds, and the highly polar metabolites stay in the aqueous and never partition into either fraction.

The three recurring bench problems are emulsions, protein precipitates and pigment carryover. An emulsion at the chloroform interface usually breaks with a centrifuge spin at 3000 rpm for 10 minutes or with the addition of solid NaCl to raise the aqueous ionic strength. A protein precipitate that wraps around the analyte is salted out with saturated ammonium sulphate, decanted, and the supernatant re-partitioned. A heavily pigmented Fraction B from a liver homogenate, where bile and bilirubin tint the chloroform yellow-green, is decolourised by a brief shake with activated charcoal (50 to 100 mg per mL) followed by filtration, or by a second defatting wash before the alkaline pull.

Preservation of viscera before extraction is non-negotiable. Saturated sodium chloride (common salt) is the standard Indian preservative because it inhibits putrefaction without reacting with the analytes. Formalin (formaldehyde solution) is forbidden for toxicology viscera because formaldehyde reacts with primary amine groups on alkaloids and forms methylene-bridged adducts that are no longer extractable into chloroform. A medico-legal autopsy that preserves viscera in formalin instead of saturated NaCl has effectively destroyed the toxicology evidence. This requirement is documented in Indian SFSL standard operating procedures and the standard Indian texts on medical jurisprudence.