Extraction and leaching

Chemical Engineering (B.E./B.Tech) · Mass transfer · free study notes for exams, placements and interviews

In short

Extraction separates a solute from a liquid feed (liquid-liquid extraction) or from a solid matrix (leaching, or solid-liquid extraction) using a solvent in which the solute is preferentially soluble. It is the separation of choice when distillation fails - close boiling points, azeotropes, heat-sensitive solutes (penicillin, vitamins), dilute aqueous streams, or when the valuable material is locked inside a solid (oilseeds, alumina from bauxite, uranium ore). The topic carries steady marks in GATE Mass Transfer (ternary diagrams, tie lines, stage-count problems) and appears in B.Tech mass transfer and separation process papers. Indian employers in which it is tested daily include edible-oil plants (Adani Wilmar, Cargill, Ruchi), Hindalco and Nalco (Bayer leaching), HZL (zinc leaching), bulk-drug and API makers in Hyderabad and Gujarat, and effluent-treatment design in phenolic streams.

Explained

Definitions and when extraction is chosen

Liquid-liquid extraction (LLE): solute C transfers from carrier liquid A (feed solvent) to a second, partially immiscible liquid B (the extracting solvent). Products: extract (solvent-rich, carries solute) and raffinate (feed-rich, depleted).
Leaching (solid-liquid extraction): a soluble constituent is dissolved out of an inert solid matrix. Products: overflow / strong solution (extract) and underflow / washed solids (inerts + adhering solution).
Choose extraction over distillation when relative volatility is near 1, an azeotrope exists, the solute is thermally labile, the solute is very dilute (distilling the whole carrier is wasteful), or the solute is non-volatile (metal salts).
Extraction is almost never a complete separation - the solute must later be recovered from the solvent, usually by distillation, stripping, back-extraction or crystallisation. Always count that cost: extraction + solvent recovery versus direct distillation.

Solvent selection criteria

Distribution coefficient K high, so less solvent is needed: K = y*/x (solute in extract phase / in raffinate phase, same units).
Selectivity (separation factor) β = (y_C/y_A)/(x_C/x_A) must exceed 1; β = 1 means no separation is possible however many stages are used.
Low mutual solubility of A and B (sharp phase split, less solvent lost in raffinate), density difference ≥ about 50-100 kg/m³ for gravity settling, moderate interfacial tension (too high → poor dispersion; too low → emulsion and no settling).
Easy recovery: volatility different from solute; low latent heat; chemical stability, non-corrosive, cheap, low toxicity and flammability. Hexane plants must respect Petroleum Act/Rules licensing and the Factories-era hazardous-process controls now carried by the Occupational Safety, Health and Working Conditions Code, 2020 (in force 21 November 2025).
Typical pairs: water-acetic acid separated by ethyl acetate / MIBK / isopropyl ether; aromatics from naphtha by sulpholane / furfural / NMP; lube-oil dewaxing by MEK-toluene; penicillin by amyl acetate at controlled pH; nickel/cobalt by organophosphoric extractants (D2EHPA, Cyanex) in kerosene diluent.

Ternary equilibrium: triangular diagrams

Plot A (carrier), B (solvent), C (solute) on an equilateral or right-triangular diagram; each point is a mixture, each apex a pure component. Sum of the three mass fractions = 1.
Binodal (solubility) curve bounds the two-phase region. Tie lines join equilibrium conjugate phases (raffinate layer ↔ extract layer); they are generally not parallel.
Plait (critical) point: where the tie line shrinks to zero and the two phases become identical - no separation there.
Type I system (one pair partially miscible, e.g. water-chloroform-acetone): the binodal is a single dome and a plait point exists. Type II system (two pairs partially miscible, e.g. water-chlorobenzene-MEK): the two-phase band stretches across, no plait point - easier to separate.
Conjugate (Hand) line: auxiliary curve used to interpolate a tie line through a given raffinate point. Hand correlation: log(y_C/y_B) = a + b·log(x_C/x_A), a straight line on log-log paper.
Lever-arm (mixture) rule: mixing M kg of stream F with S gives a point M on the line FS, with F·(FM distance) = S·(MS distance). Used for every mass balance on the diagram.
Solvent-free (Janecke) coordinates or the right-triangular plot are used when the solvent fraction is large.

Stagewise LLE: single stage, multistage and countercurrent

Single-stage (batch) contact: F + S → M; the tie line through M gives extract E and raffinate R. Maximum recovery is limited by equilibrium.
Multistage cocurrent / cross-current (multiple contact): fresh solvent to each stage; high recovery but very dilute extract and large solvent use. Laboratory practice.
Countercurrent cascade: feed and solvent enter at opposite ends - the industrial arrangement. Gives the strongest extract for a given solvent rate.
Graphical construction (ternary diagram): locate M from F and S; fix the specified R_N; the line from R_N through M extended to the binodal gives E₁; find the operating (difference) point Δ as the intersection of lines F-E₁ and S-R_N; alternating tie lines and Δ-lines steps off the stages.
Minimum solvent-to-feed ratio (S/F)min: the tie line extended through F meets the S-R_N line at Δ_min; below this ratio infinite stages are needed. Actual S/F is typically 1.2-2 × minimum.
Immiscible-solvent (dilute) case: when A and B are mutually insoluble and solutions are dilute, use solute-free flow rates and an x-y stage construction exactly like absorption: McCabe-Thiele with operating line of slope L/V, or the Kremser equation if the equilibrium line is straight.
Extraction factor E = m·V/L (m = slope of equilibrium line, V = solvent, L = feed carrier). E > 1 is needed for sharp extraction; E ≈ 1.3-2 is the economic range.
Countercurrent extraction with reflux: part of the extract is returned after solvent removal, allowing a product richer than any equilibrium extract - analogous to rectification; needed for close separations such as aromatic/aliphatic.

Equipment for LLE

Mixer-settler: impeller-stirred mixer plus gravity settler, one per stage. High stage efficiency (75-95%), handles solids, easy scale-up; large footprint, high solvent inventory, high power. Standard in hydrometallurgy (copper, zinc, rare earths).
Spray column: simplest; one phase dispersed as drops. Severe axial mixing, roughly 1-2 stages only.
Packed column: random or structured packing wetted preferentially by the continuous phase; moderate efficiency, cheap, good for corrosive duties.
Sieve-plate (perforated-plate) column: dispersed phase re-formed into drops at each plate; no mechanical drive; good for low interfacial tension systems with clean liquids.
Agitated columns: RDC (rotating disc contactor) - rotating discs between stator rings; Oldshue-Rushton, Scheibel, Kuhni - impellers with calming zones; pulsed column - liquid pulsed by air or piston, no internal moving parts, hence standard for nuclear/radioactive service (PUREX-type circuits at BARC/IGCAR).
Centrifugal extractors (Podbielniak, annular): very short contact time and high throughput per volume; for low density difference, emulsifying systems and antibiotics.
Flooding is the capacity limit - the dispersed phase can no longer flow countercurrently. Design velocity is typically 50-60% of flooding; HETS or HTU-NTU is used for height.

Leaching: rate, practice and equipment

Mechanism, four steps: (1) solvent diffuses to and into the solid, (2) solute dissolves, (3) solute diffuses out through the pores (usually rate-controlling), (4) solute transfers from surface to bulk liquid.
Rate factors: particle size (smaller = faster but harder to drain and filter), temperature (higher solubility and diffusivity, limited by solvent boiling point and product damage), agitation, solvent-to-solid ratio, prior treatment (flaking, cracking, expanding of oilseeds; roasting of ores).
Preparation examples: soybean flaked to 0.25-0.4 mm; sugarcane/beet cut into cossettes; bauxite ground for the Bayer process; sulphide ores roasted to oxides before acid leaching.
Equilibrium assumption: usually the adhering liquid in the underflow has the same composition as the overflow (ideal stage). Plot on a right-triangular or Ponchon-Savarit style diagram, or use x-y (McCabe-Thiele) construction with y = solute fraction in overflow solution, x = in underflow solution.
Constant versus variable underflow: if the retained solution per kg inert solid is constant, operating line is straight and the Kremser equation applies; if it varies with concentration, the operating line curves and stages are stepped graphically.
Batch and semi-batch: percolation/heap leaching (copper and gold dumps; also uranium at Jaduguda historically by agitation leach), Shanks (vat) system - a battery of vats with solvent cascaded countercurrently, Pachuca tank (air-agitated).
Continuous: Bollman (bucket-elevator) extractor, Rotocel, Hildebrandt screw / De Smet belt extractor for oilseeds; Dorr thickeners in series (CCD - countercurrent decantation) for finely divided ore pulps; diffusion battery for sugar beet.
Special solvents: supercritical CO₂ for decaffeination, spice oleoresins and nutraceuticals (zero solvent residue, high pressure cost); water for sugar; dilute H₂SO₄ for copper and zinc; NaOH in the Bayer process; NaCN for gold.
Hexane losses in Indian solvent-extraction plants are regulated as fugitive VOC emissions under the Environment (Protection) Act, 1986 rules and state pollution-control-board consents; FSSAI limits residual hexane in edible oils. Specific limits are revised periodically - check the current notification rather than quoting a remembered number.

Comparisons and limitations

Extraction vs distillation: no latent heat for the main separation, works at ambient temperature, handles azeotropes - but needs a downstream recovery step, adds a solvent to the system, and never gives pure products in one unit.
Leaching vs filtration/washing: leaching dissolves; washing merely displaces. Most leaching plants end with a washing cascade to recover occluded solute.
Limitations: emulsion formation, crud/solids build-up at the interface, solvent loss and cost, phase-separation time, corrosion in acid circuits, and low driving force near the plait point.

Key points

Liquid-liquid extraction - solute moved between two partially miscible liquids using a solvent; products extract and raffinate.
Leaching - solute dissolved out of a solid; products overflow solution and underflow wet solids.
Distribution coefficient K - ratio of equilibrium solute concentrations, extract over raffinate; high K means low solvent demand.
Selectivity β - K_C/K_A; separation is impossible when β = 1 no matter how many stages.
Tie line - joins the two conjugate equilibrium phases on a ternary diagram; tie lines are usually non-parallel.
Plait point - tie line length zero; phases identical; zero driving force.
Type I vs Type II - one vs two partially miscible pairs; Type II has no plait point.
Lever-arm rule - mass balances on a ternary diagram become distances along a straight line.
Minimum solvent ratio - set by the tie line through the feed; actual is 1.2-2 times this.
Extraction factor E = mV/L - the single number controlling stage requirement; E > 1 essential.
Kremser equation - closed-form stage count for straight equilibrium and operating lines (dilute systems).
Mixer-settler - near-ideal stage efficiency, solids-tolerant, bulky; dominant in hydrometallurgy.
Agitated columns (RDC, Scheibel, Kuhni, pulsed) - mechanical energy creates fine drops; pulsed columns suit nuclear duty.
Centrifugal extractor - seconds of contact; for antibiotics and low Δρ systems.
Flooding - the hydraulic capacity limit of any extraction column; design at 50-60% of it.
Leaching rate control - usually internal pore diffusion, so particle size and preparation dominate.
Constant underflow - retained solution per kg inert solid independent of concentration → straight operating line.
Solvent recovery - the real cost centre of extraction; judge the process on extraction plus recovery.

Formulas and examples

K = y*/x — distribution (partition) coefficient; y* = solute mass (or mole) fraction in extract phase at equilibrium, x = in raffinate phase; dimensionless.

β = K_C/K_A = (y_C/x_C)/(y_A/x_A) — selectivity (separation factor); dimensionless; β > 1 required.

Overall balance: F + S = M = E + R; solute balance: F·x_F + S·y_S = E·y_E + R·x_R — F, S, E, R, M in kg (or kg/h); x, y mass fractions.

Lever-arm rule: F/S = (distance MS)/(distance FM) — distances measured on the ternary diagram.

Single-stage with immiscible solvents: x₁ = x_F/(1 + m·S/L) — L = feed carrier (kg/h), S = solvent (kg/h), m = slope of equilibrium line y = m·x.

n cross-current stages, equal solvent S/n each: x_n = x_F/(1 + m·S/(n·L))ⁿ

Extraction factor: E = m·V/L (V = solvent flow, L = carrier flow, mass or molar basis, consistent units).

Kremser (countercurrent, straight lines, pure entering solvent y_{N+1}=0):
(x_F − x_N)/(x_F − x*_N) = (E^{N+1} − E)/(E^{N+1} − 1), and
N = ln{[(x_F/x_N)(E−1) + 1]/E} / ln E for E ≠ 1; N = x_F/x_N − 1 for E = 1.

Leaching washing cascade (constant underflow): recovery fraction = (R^{N+1} − R)/(R^{N+1} − 1) where R = (solvent in overflow)/(solution retained in underflow) is the wash ratio.

Column height: Z = HTU × NTU = HETS × N_theoretical — HTU, HETS in m.

Drop/flooding guide: V_d + V_c ≤ V_flood; design (V_d + V_c) ≈ 0.5-0.6 V_flood — superficial velocities, m/s.

Example 1 — Single-stage extraction, immiscible solvents. 1,000 kg/h of water containing 5% (mass) acetic acid is contacted with 600 kg/h of pure isopropyl ether. Equilibrium (dilute, mass-ratio basis) y = 0.33x.

Carrier water L = 1,000 × 0.95 = 950 kg/h; acid in feed = 50 kg/h; feed ratio X_F = 50/950 = 0.05263 kg acid/kg water.
m = 0.33, S = 600 kg/h → mS/L = 0.33 × 600/950 = 0.2084.
X₁ = 0.05263/(1 + 0.2084) = 0.04356 kg/kg → acid left = 0.04356 × 950 = 41.4 kg/h.
Acid extracted = 50 − 41.4 = 8.6 kg/h; recovery = 8.6/50 = 17.2%. One stage is clearly inadequate - hence cascades.

Example 2 — Cross-current vs countercurrent with the same solvent. Same feed, same total 600 kg/h solvent, now in three equal cross-current stages (200 kg/h each).

mS/(nL) = 0.33 × 200/950 = 0.06947; (1.06947)³ = 1.2233.
X₃ = 0.05263/1.2233 = 0.04302 → acid left = 40.9 kg/h, recovery 18.3% - barely better than one stage.
Countercurrent, 3 stages: E = mS/L = 0.2084. Kremser: (E⁴ − E)/(E⁴ − 1) = (0.001886 − 0.2084)/(0.001886 − 1) = (−0.2065)/(−0.99811) = 0.2069 → recovery 20.7%, acid left 39.7 kg/h.
Lesson: with E = 0.21 (far below 1) no cascade can help. Raise E - use more solvent or a solvent with higher m.

Example 3 — Countercurrent stage count by Kremser (E > 1). A 2,000 kg/h aqueous stream with 0.8% (mass) phenol must be reduced to 0.05% phenol using benzene; equilibrium (dilute mass-ratio basis) Y = 2.2X. Solvent used = 250 kg/h benzene.

Water carrier L ≈ 2,000 × 0.992 = 1,984 kg/h; X_F = 0.008/0.992 = 0.008065; X_N = 0.0005/0.9995 = 0.0005003.
E = mV/L = 2.2 × 250/1,984 = 0.2772... E < 1, so the target is unreachable. Minimum solvent for E = 1: V = L/m = 1,984/2.2 = 902 kg/h.
Take V = 1,500 kg/h → E = 2.2 × 1,500/1,984 = 1.6633.
X_F/X_N = 0.008065/0.0005003 = 16.12.
N = ln{[16.12 × (0.6633) + 1]/1.6633}/ln(1.6633) = ln{(10.693 + 1)/1.6633}/0.5088 = ln(7.030)/0.5088 = 1.9502/0.5088 = 3.83 → 4 theoretical stages.
At 70% stage efficiency in an RDC, 4/0.7 ≈ 6 real stages; at HETS = 0.5 m (HETS already embodies the efficiency, so it multiplies the *theoretical* count), packed height ≈ 0.5 × 3.83 ≈ 1.9 m of contacting zone.

Example 4 — Leaching mass balance (oilseed). 10,000 kg/h of soybean flakes contain 18% oil, 82% oil-free solid. Hexane extraction leaves cake retaining 0.25 kg solution per kg oil-free solid, and the leached cake carries 1.2% of the original oil as unextracted oil in that retained solution... compute the overflow.

Oil in feed = 1,800 kg/h; inert solids = 8,200 kg/h.
Solution retained in underflow = 0.25 × 8,200 = 2,050 kg/h.
Oil lost with the cake = 1,800 × 0.012 = 21.6 kg/h → hexane in retained solution = 2,050 − 21.6 = 2,028.4 kg/h; underflow solution strength = 21.6/2,050 = 1.05% oil.
Oil in overflow = 1,800 − 21.6 = 1,778.4 kg/h; recovery = 98.8%.
If fresh hexane supplied = 12,000 kg/h, overflow = 12,000 − 2,028.4 + 1,778.4 = 11,750 kg/h at 1,778.4/11,750 = 15.1% oil, which then goes to the evaporator/stripper train.

Example 5 — Washing cascade (constant underflow). Leached solids retain 1 kg solution per kg inert; wash water of 3 kg per kg inert is fed countercurrently. Wash ratio R = 3/1 = 3. How many stages to recover 99% of the solute in the retained solution?

Recovery = (R^{N+1} − R)/(R^{N+1} − 1). Try N = 4: 3⁵ = 243 → (243 − 3)/(243 − 1) = 240/242 = 0.9917 → 99.2%.
N = 3: 3⁴ = 81 → (81 − 3)/80 = 78/80 = 0.975 → 97.5%. So 4 washing stages suffice; a 5th gives 99.7% at extra dilution cost.

Feature	Liquid-liquid extraction	Leaching
Feed phase	Liquid	Solid (porous or ground)
Equilibrium basis	Ternary tie lines / distribution coefficient	Retained solution = overflow composition (ideal stage)
Rate control	Interphase mass transfer in drops	Diffusion inside solid pores
Typical equipment	Mixer-settler, RDC, pulsed column, centrifugal	Rotocel, Bollman, Dorr thickeners, heap/vat
Key design variable	Solvent ratio, extraction factor	Particle size, solvent/solid ratio, wash ratio
Indian example	Aromatics by sulpholane at refineries	Soybean oil by hexane; bauxite by Bayer caustic

Contactor	Stages/unit	Solids tolerance	Throughput	Typical use
Mixer-settler	1 (high η)	Good	High	Copper, zinc, rare-earth circuits
Packed / sieve-plate column	Several	Poor	Moderate	Phenol removal, petrochemicals
RDC / Kuhni / Scheibel	Many	Fair	Moderate-high	Lube extraction, pharma
Pulsed column	Many	Fair	Moderate	Nuclear reprocessing
Centrifugal	1-several	Poor	Very high	Antibiotics, low Δρ systems

Where it is used

Edible oil industry: hexane solvent-extraction plants for soybean, groundnut, rice bran, cottonseed and mustard cake; roles include process engineer, extraction-plant operations and solvent-loss control (Adani Wilmar, Ruchi Soya/Patanjali Foods, Cargill, Emami Agrotech).
Hydrometallurgy and metals: Bayer caustic leaching of bauxite at Hindalco and Nalco; zinc and lead concentrate leaching at Hindustan Zinc; copper heap/agitation leach and SX-EW; rare-earth separation at IREL.
Petroleum refining and petrochemicals: aromatics recovery (sulpholane/NMP) and lube-oil dewaxing at Reliance, IOCL, HPCL, BPCL; phenol and cresol recovery from coker and resin effluent.
Pharmaceuticals and API: antibiotic recovery (penicillin, cephalosporin) by pH-swing extraction, alkaloid and steroid purification at Hyderabad, Ankleshwar and Baddi plants; QA/validation and tech-transfer roles.
Food, flavour and nutraceuticals: supercritical CO₂ and solvent extraction of spice oleoresins, mint oil, coffee decaffeination (Synthite, Kancor); sugar diffusion batteries in UP and Maharashtra mills.
Nuclear and environment: PUREX-type solvent extraction (TBP in kerosene) at BARC/NFC; metal recovery from electroplating and e-waste leachates; phenolic and heavy-metal effluent treatment under state pollution-control-board consents.

Interview questions

Q: When would you choose extraction over distillation? A: When components boil close together or form an azeotrope, when the solute is heat-sensitive or non-volatile, or when it is so dilute that vaporising the whole carrier is uneconomic.
Q: What is a tie line and what does it represent? A: A straight line on a ternary diagram joining the compositions of the two equilibrium (conjugate) phases formed by any mixture lying on it.
Q: What happens at the plait point? A: The two conjugate phases become identical, the tie line length goes to zero and there is no driving force, so no separation is possible there.
Q: Why must selectivity exceed 1? A: Selectivity compares the solvent's affinity for solute against its affinity for the carrier; at β = 1 the solvent cannot discriminate between them, so no cascade will separate the mixture.
Q: Define the extraction factor and say why E > 1 matters. A: E = mV/L; it is the ratio of the equilibrium line's capacity to the operating line's. If E ≤ 1 the solute cannot be stripped below a limiting value even with infinite stages.
Q: Why is countercurrent better than cross-current extraction? A: Countercurrent maintains a driving force along the whole cascade with a single solvent stream, giving a concentrated extract; cross-current needs fresh solvent per stage and yields a dilute one.
Q: What usually controls the rate of leaching? A: Diffusion of dissolved solute through the pores of the solid, which is why particle size reduction, flaking and temperature matter more than agitation speed.
Q: In a mixer-settler, what decides the settler size? A: The phase-disengagement (coalescence) time, which depends on density difference, interfacial tension, drop size, viscosity and the presence of surfactants or fine solids.
Q: Which phase should be dispersed? A: Usually the one with the larger volumetric flow is dispersed to maximise interfacial area, but the phase that wets the packing or is prone to emulsion is kept continuous; dispersing the costlier or more hazardous phase is avoided.
Q: What is flooding in an extraction column? A: The condition where the dispersed phase can no longer move countercurrently against the continuous phase and is carried out of the wrong end; design velocity is kept at about half the flooding value.
Q: Why are pulsed columns preferred in nuclear reprocessing? A: They have no moving parts inside the radioactive zone, so maintenance and contamination risk are minimal while still giving fine drops and many stages.
Q: Reduce solvent consumption in an existing extraction plant - what do you examine first? A: The extraction factor and raffinate specification (is the target tighter than needed?), then stage efficiency, column operation at near-flooding, mutual solubility losses in the raffinate, and venting/fugitive hexane losses.

Common mistakes

Treating tie lines as parallel or as passing through a fixed point - they generally are neither, and interpolation needs the conjugate (Hand) line.
Using total feed and solvent flows in dilute-system formulas instead of solute-free carrier flows and mass ratios - a common source of wrong Kremser answers.
Forgetting that mass fractions on a ternary diagram must sum to 1, so the third component is never independent.
Applying the Kremser equation when the equilibrium or operating line is curved (variable underflow in leaching, concentrated LLE).
Ignoring the solvent-recovery step and claiming extraction is "energy-free".
Reporting a solvent ratio below the minimum and still stepping off a finite number of stages on the diagram.
Assuming finer grinding always improves leaching - too fine a solid drains poorly, blinds filters and increases solute lost in the underflow.