Liquid–Liquid Extraction¶

Transfer selected dissolved species between contacting immiscible or partially miscible liquid phases, then disengage the phases as extract and raffinate, using equilibrium distribution, speciation control, and staging to obtain recovery and selectivity.

Version: v1 · 2026-08-30 · History
Domain-specific #: 2193
Origin domain: chemical separation engineering
Subdomain: solvent extraction and liquid–liquid distribution
Aliases: Liquid-liquid extraction, Liquid–liquid distribution, Solvent extraction (liquid–liquid)

Core Idea¶

Liquid–liquid extraction is a separation operation that contacts a liquid feed with a second, immiscible or partially miscible liquid phase so that selected dissolved species transfer between them, and then disengages the phases to obtain an extract enriched in transferred material and a raffinate depleted of it. The operation is also called solvent extraction or, in IUPAC's more general terminology, liquid–liquid distribution. Its identity is the whole controlled sequence: choose two phase-forming liquids and a chemical environment, create interfacial contact, permit solute transfer toward a distribution state, separate the bulk phases, and route or regenerate the resulting streams.^[1]^[2]

Relative solubility is necessary but not sufficient to describe the abstraction. The relevant quantity is normally a condition-specific distribution relation between both phases. Chemical form matters: protonation, complexation, oxidation state, ion pairing, temperature, and concentrations can change the total analytical distribution ratio even when the partitioning of one molecular species is unchanged. Hydrodynamics matters too: equilibrium data indicate what a contact could attain, while interfacial area, residence time, diffusion, coalescence, entrainment, and emulsion formation determine whether a real contact approaches it. A sound account therefore keeps equilibrium, kinetics, and phase handling separate.

The abstraction recurs from a separatory funnel used before chromatographic analysis to mixer–settler cascades used in copper hydrometallurgy and solvent-extraction cycles used in nuclear fuel processing. Hardware, scale, solute, and chemistry vary, but the same role graph persists. That stable, reusable graph makes liquid–liquid extraction an autonomous domain-specific abstraction rather than a product category, one apparatus, a synonym for solubility, or an unrestricted prime.

Structural Signature¶

A valid liquid–liquid extraction instance contains these roles:

feed phase — a liquid stream or sample containing one or more solutes to be redistributed;
receiving phase or solvent phase — a second liquid that forms a separate bulk phase under the operating conditions;
solvent formulation — the phase-forming liquid and, where relevant, an active extractant, diluent, modifier, or carrier; IUPAC restricts extractant to the active substance chiefly responsible for transfer rather than using it for every organic phase component;^[1]
target and competing solutes — the species whose recovery, rejection, purification, or mutual separation defines performance;
speciation and operating conditions — pH, complexant and salt concentrations, oxidation state, temperature, phase composition, and other variables that fix the available chemical forms;
distribution relation — the equilibrium or empirical mapping between solute concentrations in the two phases under those stated conditions;
contact operation — controlled dispersion, interfacial-area generation, flow, and residence time that permit mass transfer;
disengagement operation — settling, coalescence, centrifugation, membranes, or another way to recover two separately routable liquid phases;
extract — the phase after contact that contains transferred material of interest;
raffinate — the residual feed phase after extraction;
stage arrangement — one equilibrium contact, repeated crosscurrent contacts, or a countercurrent cascade with declared flow directions;
recovery or stripping operation — where required, a change of chemical or physical conditions that transfers solute out of the loaded solvent and permits solvent recycle; and
performance statement — recovery, selectivity, purity, concentration, solvent loss, entrainment, throughput, and uncertainty reported with conventions and operating conditions.

The locked operation is:

define feed, targets, and conditions -> select a phase-forming solvent system -> establish distribution and selectivity data -> contact phases to create interfacial transfer -> allow or model approach to equilibrium -> disengage phases -> route extract and raffinate -> repeat or stage contacts if required -> strip/recover solute and recycle solvent where the process demands it.

For a declared convention let D = C_E/C_R, where C_E is the total analytical concentration of a solute in the extract phase and C_R its total analytical concentration in the other phase at equilibrium. This distribution ratio must not silently be replaced by the partition ratio of one definite chemical species. IUPAC defines the latter as the concentration ratio for the same single form in both phases and requires an inverse convention, if used, to be stated.^[3]

Under the simplifying assumptions of one equilibrium contact, negligible mutual solubility and volume change, and phase volumes V_E and V_R, a solute mass balance gives the fraction extracted

F = D V_E / (V_R + D V_E)

and the fraction remaining q = V_R/(V_R + D V_E). The equation is a stage model, not a universal performance guarantee: finite transfer rates, reactions, association, changing phase volumes, solvent loading, and nonideal equilibrium can invalidate its assumptions. Selectivity between solutes A and B may be summarized under identical conditions by IUPAC's separation factor alpha_A,B = D_A/D_B.^[4]

What It Is Not¶

Not solubility alone. Solubility specifies an equilibrium capacity in a solvent under stated conditions. Extraction constructs and operates a two-liquid contact, redistribution, and stream-separation system.
Not ordinary mixing. Mixing is intentionally used to enlarge interfacial area, but permanent homogenization defeats the need to recover two phases. Contact must be followed by disengagement.
Not spontaneous phase separation. The live prime describes demixing of a previously mixed system when like-with-like interactions dominate. Extraction begins with or creates two usable liquid phases and exploits solute transfer between them; the phase boundary is necessary infrastructure, not the selected product.
Not every use of the word extraction. Solid–liquid leaching, solid-phase extraction, supercritical-fluid extraction, gas absorption, mechanical expression, and information extraction have different carrier phases or mechanisms.
Not adsorption. Adsorption accumulates material at finite surface sites. Liquid–liquid extraction transfers solute into the bulk of another liquid phase, even though interfaces and interfacial phenomena affect transfer.
Not precipitation. Precipitation forms a solid phase after supersaturation and nucleation. Extraction retains two liquid bulk phases and normally routes dissolved solute in one of them.
Not distillation. Distillation separates through vaporization and condensation according to volatility. Extraction can be selected precisely when volatility differences are small or thermal exposure is undesirable, although solvent recovery may later use distillation.
Not chromatography. A separatory-funnel extract may feed a chromatograph, as in EPA Method 3510C, but the batch contact and bulk phase disengagement do not require migration through a stationary phase.^[5]
Not a separatory funnel or mixer–settler. Those are implementations. Centrifugal contactors, columns, membranes, and microfluidic devices can instantiate the same abstraction.
Not an equilibrium coefficient without an operation. A distribution ratio describes a state relation. Extraction requires deliberate contact, transfer, and phase routing.

Scope of Application¶

The abstraction applies across chemical separation engineering whenever two liquid phases can be maintained and target species can be made to distribute differently between them. In analytical chemistry it prepares samples and concentrates water-insoluble or slightly soluble organics before measurement. EPA Method 3510C, for example, uses separatory-funnel liquid–liquid extraction of aqueous samples before chromatographic procedures and treats persistent emulsions as a threat to acceptable extract recovery.^[5]

In organic and pharmaceutical laboratories, acid–base extraction changes protonation to route acidic, basic, and neutral compounds into different phases. In hydrometallurgy, metal ions are complexed into an organic solvent, scrubbed or selectively stripped, and recovered in a concentrated aqueous stream. A USGS model of copper solvent extraction–electrowinning uses two extraction stages and one strip stage: pregnant leach solution contacts an organic phase in mixer–settlers, loaded organic is stripped by acidic electrolyte, and the separated solvent is recycled.^[6]

In nuclear fuel reprocessing, PUREX uses tributyl phosphate in a diluent with nitric-acid aqueous streams to coextract, scrub, partition, and strip uranium and plutonium. The chemical details are process-specific, but they instantiate the same two-phase transfer, condition shift, disengagement, and solvent-recycle structure.^[7] Related recurring applications include rare-earth and lanthanide separations, food and flavor processing, biotechnology, fermentation-product recovery, and solvent-based purification.

The scope includes aqueous–organic and aqueous two-phase systems, provided two liquid phases, a declared distribution relation, contact, and phase recovery remain present. It excludes processes in which the receiving medium is a solid, gas, or supercritical fluid, and cases where the liquids become one homogeneous phase under operating conditions. “Solvent extraction” should be scoped because some communities use that phrase more broadly than liquid–liquid extraction.

Clarity¶

The abstraction clarifies a separation by forcing four independent questions. Thermodynamics: where could each chemical form distribute at equilibrium? Chemistry: which forms exist under the stated pH, complexant, redox, salt, and loading conditions? Transport: how quickly does material cross the interface under the chosen dispersion and residence time? Hydraulics: can the phases coalesce and be recovered without unacceptable entrainment or emulsion? A high equilibrium D does not rescue a contact with negligible mass-transfer area; aggressive dispersion does not rescue a solvent with poor selectivity; excellent transfer is not a useful separation if the phases cannot be disengaged.

The extract/raffinate vocabulary also prevents a common naming error. “Organic” and “aqueous” are convenient in many cases but are not structural roles, and either phase can be the original feed. Likewise, the extract is defined by receiving extracted material, not by always being the less dense or organic layer. Phase identity, density, and flow direction must be declared rather than inferred from laboratory habit.

Manages Complexity¶

A real mixture may contain multiple solutes, several chemical forms of each solute, nonideal solvents, competing reactions, and many possible stage configurations. The abstraction compresses that complexity into coupled modules: a speciation model, phase-equilibrium or empirical distribution relations, a mass-transfer approach, a stagewise mass balance, and a disengagement/recycle design. Engineers can change one module while holding the others explicit. A pH swing changes speciation and D; a different contactor changes area and residence time; extra countercurrent stages amplify modest per-stage selectivity; a coalescer addresses entrainment without redefining the equilibrium.

This modularity turns scattered observations into decisions. Distribution data screen solvent systems; separation factors compare competing solutes; phase ratios set the capacity burden; stage calculations organize cascades; mass-transfer coefficients and residence times identify rate limits; and settling tests expose hydraulic failure. The structure also reveals missing information. A report of “90% extraction” is incomplete without phase ratio, number of contacts, concentrations, conditions, and whether the value is equilibrium, bench-scale, or process recovery.

Abstract Reasoning¶

The simple stage balance licenses bounded predictions. Suppose D = 4, V_R = 100 mL, and a single fresh extract phase has V_E = 25 mL. Then F = 4(25)/(100 + 4(25)) = 0.50: half of the solute is extracted at equilibrium. After separating that extract and contacting the raffinate with another fresh 25 mL portion under unchanged conditions, the fraction remaining is 0.5^2 = 0.25, so total extraction is 75%. By comparison, one 50 mL portion gives F = 200/(100+200) = 66.7%. Splitting a fixed solvent volume can improve equilibrium recovery, but it adds handling, phase loss, time, and opportunities for contamination; the arithmetic does not choose the operating policy by itself.

Speciation permits intervention inference. If only the neutral form of a weak acid partitions strongly into the organic phase, lowering aqueous pH can increase the neutral fraction and hence the total analytical distribution ratio. Raising pH after transfer can ionize the acid and strip it back into water. The prediction must be conditional: it depends on acid dissociation, solvent, concentration, ionic strength, and competing chemistry. It is wrong to describe D as one immutable molecular constant in such a system.

Stagewise reasoning also explains why a modest separation factor can become useful in a countercurrent cascade. Repeated contacts expose fresh or progressively cleaner solvent to progressively depleted raffinate, multiplying separation while preserving the material balance. The inference fails if solvent loading changes distribution strongly, a third phase appears, reactions are too slow, or entrainment carries one bulk phase mechanically into the other. Those are model-boundary failures, not counterexamples to the identity.

Knowledge Transfer¶

The complete abstraction transfers literally among laboratory sample preparation, organic synthesis workup, hydrometallurgy, nuclear processing, pharmaceutical purification, and bioprocessing. In each case there are two liquid phases, condition-dependent solute distribution, controlled contact, disengagement, extract and raffinate routing, and possibly repeated stages and solvent regeneration. A separatory funnel and a countercurrent mixer–settler train differ greatly in scale, yet the same role map supports diagnosis and design.

Outside chemical separations, only a generic skeleton transfers: allocate selected content between compartments, repeatedly contact, and recover streams. That skeleton is already handled by broad abstractions such as Mixing, Selection, Partition, and Phase Separation. Calling the transfer of people, money, or records “solvent extraction” does not preserve chemical species, liquid phases, interfacial mass transfer, distribution ratios, speciation control, or extract/raffinate streams. Such uses are analogy, not literal substrate independence. The candidate therefore remains domain-specific.

Examples¶

Analytical sample preparation. An aqueous sample is adjusted to a method-specific pH, contacted with an immiscible solvent, allowed to separate, and extracted repeatedly. The combined extract can be dried and concentrated before chromatography. EPA Method 3510C is a documented implementation for water-insoluble and slightly soluble organic compounds; its emulsion warning demonstrates that phase disengagement is part of method validity, not cleanup trivia.^[5]

Acid–base workup. A mixture of a carboxylic acid and a neutral organic compound is dissolved in an organic phase and contacted with aqueous base. Deprotonation converts the acid to an ionic carboxylate that favors the aqueous phase, while the neutral compound remains largely organic. The layers are separated; acidifying the aqueous extract regenerates the neutral acid for isolation. The recognized instance includes the pH-controlled speciation switch, both transfers, and separate routing—not merely the statement that one compound is “more soluble.”

Copper SX–EW. Pregnant leach solution contacts a selective organic phase in extraction mixer–settlers. Copper loads into the organic phase, raffinate returns toward leaching, loaded organic contacts acidic electrolyte in a strip stage, and the regenerated organic phase recycles. USGS's described two-extraction/one-strip arrangement exhibits feed, solvent, staged countercurrent routing, extract/raffinate functions, stripping, settling, and recycle.^[6]

PUREX. Nitric-acid dissolver solution contacts tributyl-phosphate solvent. Uranium and plutonium are extracted, fission products are rejected or scrubbed, redox and acidity changes permit partitioning and stripping, and washed solvent can be reused. A historical technical report describes coextraction into TBP, nitric-acid scrubbing, and selective plutonium stripping after reduction.^[7] The example shows that selectivity may be actively engineered through chemical form rather than inherited from neutral-molecule polarity.

Failed implementation boundary. Two liquids are shaken so vigorously that a stable emulsion prevents layer recovery. Interfacial area may be high and solute transfer may occur, but the intended extraction operation has failed hydraulically because extract and raffinate cannot be separately routed at acceptable recovery. The case remains an attempted instance and diagnoses the contact–disengagement tension.

Structural Tensions¶

Interfacial area vs. disengagement. Smaller droplets and stronger agitation accelerate mass transfer, but they can slow settling, stabilize emulsions, and increase entrainment. Contact and recovery must be designed together.
Equilibrium selectivity vs. transfer rate. A solvent system may offer excellent distribution and separation factors yet have slow kinetics; a fast-transfer system may have inadequate selectivity. Neither metric substitutes for the other.
Recovery vs. selectivity. Very large D for the target improves extraction, but coextraction of impurities can damage purity. The relevant comparison is the set of condition-specific distribution ratios, not target recovery alone.
Solvent volume vs. concentration and handling. More solvent or more fresh contacts can raise recovery, while diluting the extract, increasing solvent inventory, energy use, losses, and phase-handling operations.
Mixing vs. phase integrity. Partial mutual solubility can aid molecular transport, but too much miscibility eliminates the two-phase operating window or causes solvent contamination of product streams.
Chemical swing vs. downstream burden. pH, redox, complexation, or salting changes can produce high selectivity and enable stripping, while consuming reagents, generating salts, corroding equipment, or complicating waste treatment.
Stage count vs. controllability. Countercurrent staging amplifies modest per-stage performance, but adds inventory, residence time, control interactions, startup transients, and consequences of off-spec equilibrium data.
Solvent performance vs. process stewardship. Volatility, toxicity, flammability, persistence, degradation, and solvent loss can outweigh attractive laboratory distribution data. Recovery and containment are part of process design.

Structural–Framed Character¶

Liquid–liquid extraction is strongly structural but unavoidably framed by chemical convention and operating conditions. Its structural core—two phases, solute distribution, contact, disengagement, routed streams, stages, and regeneration—survives across devices and industries. Its quantitative claims require a frame: which phase is numerator, whether concentration is total analytical solute or one chemical form, the phase-volume convention, equilibrium assumptions, temperature, composition, and speciation state. A distribution ratio without those declarations is not portable evidence.

The abstraction is therefore neither a loose topic label nor a timeless formula. It is a reusable operating grammar whose variables are populated by system-specific thermodynamics, chemistry, transport, and hydraulics. The frame is what makes comparison honest; the structure is what makes transfer possible.

Structural Core vs. Domain Accent¶

The liftable core is two distinguishable carriers + selectively transferable content + contact + approach to an allocation relation + carrier separation + output routing + repetition/regeneration. That pattern resembles generic Allocation, Mixing, Partition, Phase Separation, and Feedback. It can guide analogy, but it does not by itself identify liquid–liquid extraction.

The domain accent is constitutive: both carriers are liquid phases; they are immiscible or only partially miscible under operating conditions; solutes have chemical species and activities; equilibrium distribution and interfacial mass transfer govern transfer; extract and raffinate are physical streams; pH, complexation, oxidation state, and loading can change selectivity; and solvents must be disengaged and often recovered. Remove those roles and the candidate collapses into existing generic abstractions. Because the chemical commitments cannot be freely substituted by objects from unrelated domains, prime autonomy fails while domain-specific autonomy survives.

Liquid–liquid extraction uses Mixing during contact: agitation or flow redistributes droplets and creates interfacial area. It then depends on a controlled form of Phase Separation or phase disengagement to recover two streams, although it is not subsumed by the live prime's spontaneous-demixing identity. It relates to Partition in the broad sense that material is allocated between outputs, but the live Partition prime concerns disjoint, exhaustive blocks of a set and does not define chemical equilibrium or repeated phase contact.

It also enacts Selection by preferring targets over competing solutes, and Feedback may appear in process control or solvent recycle. These primes explain portable pieces, not exact coverage. The domain-specific node remains necessary because no listed prime binds chemical speciation, two-liquid equilibrium, interfacial transfer, stage balances, extract/raffinate routing, and solvent regeneration into the recurring operation.

Relationships to Other Abstractions¶

Current abstraction Liquid–Liquid Extraction Domain-specific

Parents (1) — more general patterns this builds on

Liquid–Liquid Extraction presupposes Solubility Domain-specific

The accepted reference-grade review places Liquid–Liquid Extraction under Solubility because the child instantiates or depends on the parent's broader structure while retaining its own constitutive identity.

Hierarchy paths (2) — routes to 2 parentless roots

Liquid–Liquid Extraction → Solubility → Threshold

Show alternative path (1)

Neighborhood in Abstraction Space¶

Liquid–Liquid Extraction sits in a sparse region of the domain-specific corpus (99^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (1565 abstractions)

Nearest neighbors

Fractional crystallization (chemistry) — 0.75
Keeper (Chemistry) — 0.75
Distillation — 0.75
Gas separation — 0.73
Evaporation — 0.73

Computed from structural-signature embeddings · 2026-09-08

Not to Be Confused With¶

Solubility is the strongest live catalog neighbor and the proposed minimal parent. It supplies condition-dependent dissolution and chemical-potential constraints, but not a two-phase contactor, distribution convention, mass-transfer approach, extract/raffinate streams, staging, stripping, or solvent recycle. Precipitation changes a dissolved solute into a solid; adsorption transfers it to surface sites; and distillation separates through vapor–liquid change and volatility. Each can be coupled to an extraction flowsheet without becoming an alias.

Mixing explains the contact step but not selectivity or recoverable layers. Phase Separation explains why separately routable regions may form, but extraction generally chooses two phases before contact and uses subsequent disengagement rather than treating demixing itself as the selected phenomenon. Asymmetric Flux is a semantic neighbor only: liquid–liquid mass transfer is driven by chemical-potential departure and may reverse when conditions reverse, not by an intrinsically directional boundary. Supersaturation, metastability, nucleation, and wettability may affect particular systems but do not close the identity.

The phrase solvent extraction requires care. In the retained sense it is an established alias for liquid–liquid distribution with emphasis on extracting analytes. In broader usage it can include extraction from solids or other media. The alias is therefore qualified as “solvent extraction (liquid–liquid)” in this draft rather than asserted without scope.

References¶

[1] International Union of Pure and Applied Chemistry. “Liquid–liquid extraction.” Compendium of Chemical Terminology (Gold Book), 5^th ed., online version 5.0.0, 2025. DOI: 10.1351/goldbook.L03587. registry ↩a ↩b

[2] N. M. Rice, H. M. N. H. Irving, and M. A. Leonard. “Nomenclature for Liquid–Liquid Distribution (Solvent Extraction) (IUPAC Recommendations 1993).” Pure and Applied Chemistry 65, no. 11 (1993): 2373–2396. DOI: 10.1351/pac199365112373. registry ↩

[3] International Union of Pure and Applied Chemistry. “Partition ratio, K_D.” Compendium of Chemical Terminology (Gold Book). DOI: 10.1351/goldbook.P04440. registry ↩

[4] International Union of Pure and Applied Chemistry. “Separation factor in liquid–liquid distribution.” Compendium of Chemical Terminology (Gold Book). DOI: 10.1351/goldbook.S05615. registry ↩

[5] U.S. Environmental Protection Agency. “SW-846 Test Method 3510C: Separatory Funnel Liquid–Liquid Extraction.” Revision 3, December 1996; official method page updated October 23, 2025. Method page and official PDF. registry ↩a ↩b ↩c

[6] U.S. Bureau of Mines / U.S. Geological Survey. “Solvent Extraction–Electrowinning Copper Model.” Cost Estimating System for Surface Mines and Mines Using Refining Methods, Information Circular 9298. Official USGS text. registry ↩a ↩b

[7] U.S. Atomic Energy Commission technical literature. “The PUREX Process—A Solvent Extraction Reprocessing Method for Irradiated Uranium.” Official Technical Information Service record and report. OSTI report. registry ↩a ↩b