Solvent Extraction

How solvent extraction works

Drop iodine crystals into a flask of water and shake — they barely dissolve, leaving a faint brown tint. Layer hexane on top and shake again, and the iodine streaks violet through the upper layer while the water clears. The iodine has chosen sides. It prefers the non-polar hexane, so it migrates there until the chemical potential of I₂ in both phases is equal.

That equilibrium is the heart of solvent extraction. For a solute A distributing between an aqueous phase and an organic phase:

A(aq) ⇌ A(org)    K_d = [A]_org / [A]_aq

K_d is the partition coefficient (sometimes called the distribution coefficient). It is a constant at a given temperature and depends only on the solute and the two solvents — not on volumes, not on starting concentrations.

The fraction of solute remaining in the aqueous phase after one extraction with volume V_org of organic solvent into a starting V_aq of aqueous solution is:

f_remaining = V_aq / (V_aq + K_d · V_org)

Visually, the separating funnel does the work. Pour both phases in, stopper, invert, and shake to maximize interfacial contact. Vent often to release pressure from volatile solvents. Let the layers re-form, identify which is denser, then drain the bottom phase through the stopcock. The product is now in one of the two collected layers — typically the organic, which evaporates cleanly with a rotary evaporator.

Why two liquids refuse to mix

Immiscibility is a thermodynamic outcome of "like dissolves like." Water molecules form a tight hydrogen-bonded network that excludes non-polar molecules — pushing them together is entropically and enthalpically unfavourable. Hexane has only weak London dispersion forces and offers no hydrogen-bond partners. The two liquids minimize their free energy by separating into bulk phases with a thin interface between them.

Polar molecules with hydrogen-bond donors (water, methanol, ethanol) tend to mix freely with each other. Non-polar solvents (hexane, toluene, diethyl ether, dichloromethane) mix freely among themselves. Pair one from each group and you get an extraction system. The greater the polarity gap, the cleaner the separation — and usually the larger the K_d for any given solute.

Worked example: sequential extraction beats one big pass

You have 100 mL of an aqueous solution containing 1.00 g of an organic compound. Its partition coefficient with diethyl ether is K_d = 5.0. You have 100 mL of ether to spend. Should you do one 100-mL extraction or three 33-mL extractions?

Single extraction (100 mL ether):

f_remaining = 100 / (100 + 5.0 × 100) = 100 / 600 = 0.167
Mass remaining in water = 1.00 g × 0.167 = 0.167 g
Mass extracted = 0.833 g (83.3%)

Three sequential extractions (33 mL ether each):

Each pass: f = 100 / (100 + 5.0 × 33) = 100 / 265 = 0.377
After three passes: f³ = 0.377³ = 0.0537
Mass remaining = 1.00 g × 0.0537 = 0.054 g
Mass extracted = 0.946 g (94.6%)

With the same total solvent volume, three small extractions recover 11 percentage points more product. Push to five 20-mL extractions and the recovery climbs to 96.9%. The general rule: n smaller extractions of V/n volume each recover more than one extraction of V, with diminishing returns past about three or four passes.

Solvent extraction vs other separation methods

	Solvent extraction	Distillation	Recrystallization	Chromatography
Driving force	Solubility difference	Volatility difference	Solubility vs temperature	Adsorption affinity
Best for	Liquid–liquid mixtures, workups	Liquid mixtures with different boiling points	Pure solid from impure solid	Trace components, complex mixtures
Throughput	Batch, minutes	Continuous, hours	Hours to days (cooling)	Minutes to hours per run
Energy cost	Low (no heating)	High (heat of vaporization)	Moderate (heating + cooling)	Low (pumping only)
Selectivity tuning	Solvent choice, pH, salting	Reflux ratio, plates	Solvent choice, cooling rate	Stationary/mobile phase choice
Scale	Lab to industrial (hydrometallurgy)	Lab to refinery	Lab to pharma kg-scale	Mostly lab to pilot
Typical recovery	90–99% with sequential passes	95–99.9% per stage	50–80% (mother liquor losses)	80–95%

Solvent extraction is the workhorse of organic synthesis cleanup. Whenever you read "the reaction was quenched with water and extracted three times with ethyl acetate," that is the operation being described.

Acid–base extraction: weaponizing pH

Most organic molecules with acidic or basic groups switch their partition behaviour dramatically with pH. A carboxylic acid R–COOH has K_d ~ 50 in ether/water, but its conjugate base R–COO⁻ has K_d ~ 0.01 — a 5000× swing. By controlling pH, a chemist can route specific compound classes into specific layers:

Compound class	Aqueous condition	Where it goes
Carboxylic acid (R–COOH)	Acidic (pH 2)	Organic layer
Carboxylic acid (R–COO⁻)	Basic (pH 12)	Aqueous layer
Amine (R–NH₂)	Basic (pH 12)	Organic layer
Amine (R–NH₃⁺)	Acidic (pH 2)	Aqueous layer
Phenol (Ar–OH)	Acidic or pH 7	Organic layer
Phenol (Ar–O⁻)	Strongly basic (pH 14, NaOH)	Aqueous layer
Neutral organic (alcohol, ester, ketone)	Any pH	Organic layer

A classic teaching exercise: separate a mixture of benzoic acid, aniline, and naphthalene. Wash with HCl(aq) — aniline goes into water as anilinium chloride, benzoic acid and naphthalene stay in the ether. Wash the organic layer with NaOH(aq) — benzoic acid leaves as sodium benzoate, naphthalene remains. You have now sorted three compounds into three layers using nothing but pH adjustments.

Industrial extraction: where lab-scale becomes tonnage

Solvent extraction moves the world's metals. Copper hydrometallurgy produces ~25% of global Cu by leaching ore with H₂SO₄, extracting Cu²⁺ into kerosene-borne LIX reagents, then back-extracting and electrowinning. The PUREX process extracts U and Pu from nuclear fuel using tributyl phosphate in kerosene; rare-earth separation needs 30+ stages because adjacent lanthanides differ in K_d by only 10–20%. Caffeine decaffeination uses supercritical CO₂ at 31°C, 74 bar — a tunable solvent above its critical point. Every API pharmaceutical plant has banks of mixer–settlers; a typical 1000-litre batch uses 800 L of organic solvent per stage, recycled hundreds of times per year.

Picking a solvent

The ideal extraction solvent is immiscible with water, dissolves the target much better than water does, evaporates easily for product recovery, doesn't react with the solute, and is cheap and safe. No solvent meets all five. The standard menu:

Solvent	Density (g/mL)	BP (°C)	Polarity	Watch out for
Diethyl ether	0.71 (top)	35	Low	Extreme flammability, peroxides on storage
Ethyl acetate	0.90 (top)	77	Moderate	Slightly soluble in water (8% w/w)
Hexane	0.66 (top)	69	Very low	Poor for polar solutes
Toluene	0.87 (top)	111	Low	Hard to evaporate, neurotoxic on chronic exposure
Dichloromethane	1.33 (bottom)	40	Moderate	Suspected carcinogen, bottom layer flips intuition
Chloroform	1.49 (bottom)	61	Moderate	Hepatotoxic, increasingly restricted
Supercritical CO₂	0.4–0.9 (tunable)	Critical: 31°C / 74 bar	Tunable	Requires high-pressure equipment

Ethyl acetate is the modern default for academic and process labs — moderate polarity, manageable BP, and a much safer toxicology profile than the chlorinated alternatives.

Variants and refinements

• Continuous liquid–liquid extraction. When K_d is small (under 1), batch extraction is hopeless. Soxhlet-style continuous extractors recycle solvent through the aqueous phase indefinitely, accumulating solute in a flask.
• Counter-current extraction (Craig distribution). A row of separating funnels passes the organic and aqueous phases in opposite directions, achieving thousands of theoretical stages — historically the way to separate proteins and antibiotics before HPLC.
• Solid-phase extraction (SPE). Replace the organic solvent with a solid sorbent (C18 silica, ion-exchange resin). Sample passes through, analyte sticks, then elutes with a small solvent volume. Standard for environmental and clinical analysis.
• Dispersive liquid–liquid microextraction (DLLME). Inject microlitres of dense extraction solvent plus a disperser into a sample, forming a droplet cloud. Centrifuge, collect microlitres for GC-MS. Concentration factors of 100–1000.
• Reactive extraction. Add a complexing agent to the organic phase (8-hydroxyquinoline, dithizone, crown ethers) that irreversibly binds the target metal, multiplying effective K_d by orders of magnitude.

Pitfalls and lab gotchas

• Pressure buildup with volatile solvents. Diethyl ether at room temperature pushes ~0.6 atm vapor pressure; warm hands or carbonate-bicarbonate reactions add more. Vent the funnel after every shake or risk a stopper rocket.
• Wrong layer assignment. Always test with a drop of water — whichever layer it joins is the aqueous one. Density tables fail when concentrated salts or high solute loadings shift densities.
• Emulsions. Detergent contamination, fine precipitates, or proteinaceous samples break into stable emulsions. Add brine, gentle swirling, centrifugation, or filtration through Celite.
• Mass balance errors. Don't forget the small amount of organic solvent dissolved in water (especially ethyl acetate), which carries dissolved product. A back-wash of the combined aqueous waste is good practice.
• Reaction during extraction. Bases extracted into chloroform can decompose it to phosgene. Hot organic solvents and amine bases can react. Chemistry doesn't pause for separation.