Articles
In a conventional rare earth solvent extraction circuit, di(2-ethylhexyl)phosphoric acid is dissolved in a low-aromatic aliphatic diluent, where the extractant exists predominantly as a hydrogen-bonded dimer. For the purpose of extraction stoichiometry the extractant is represented as HX; in the organic phase the dimeric species (HX)2 provides the active cation-exchange functionality. When an acidic chloride or nitrate leach liquor contacts the organic phase, trivalent lanthanide ions exchange with the dimer according to M3+(aq) + 3(HX)2(org) ⇌ M(X2H)3(org) + 3H+(aq), releasing 3 mol of hydrogen ion per mole of extracted rare earth and imposing a direct dependence of extraction efficiency on equilibrium acidity. The loaded complex is neutral, and its solubility in the organic diluent is strongly influenced by the aliphatic character of the diluent, the total extractant concentration, and the free dimer concentration remaining after metal saturation. Industrial extractant loadings in continuous mixer-settler batteries are conventionally reported in the range of 0.8–1.5 mol/L D2EHPA in aliphatic kerosene diluents; below 0.8 mol/L, the useful volumetric throughput may exceed the available mass-transfer area, while above 1.5 mol/L, the organic-phase viscosity increases and phase disengagement deteriorates. Fresh solvent is usually pre-equilibrated with dilute acid and then partially saponified to control the acid titre of the aqueous phase. The extraction reaction proceeds most efficiently when the aqueous feed pH is maintained in the region of 1.2–2.2, because lower acidity reduces extraction and higher acidity increases acid consumption in stripping. Continuous circuits typically operate with an organic-to-aqueous phase ratio of 0.8–2.0, and the temperature is controlled between 25 °C and 45 °C to balance kinetics against solvent evaporative loss and hydrolytic degradation.
The extraction sequence observed with D2EHPA follows the lanthanide contraction: distribution coefficients increase from lanthanum through lutetium because the smaller trivalent ionic radius produces a higher charge density and stronger interaction with the organophilic deprotonated dimer. In mixed rare earth feeds, the observed extraction order is La < Ce < Pr < Nd < Sm < Eu < Gd < Tb < Dy < Ho < Y < Er < Tm < Yb < Lu in nitrate and chloride media, although the precise position of yttrium is sensitive to diluent polarity and equilibrium pH. The practical selectivity of D2EHPA is therefore most pronounced for mid-to-heavy rare earth separation, while the separation of adjacent light lanthanides requires many countercurrent stages. Reported separation factors for adjacent heavy lanthanide pairs in chloride and nitrate media commonly fall within 2.0–2.5, and the logarithmic distribution coefficient rises with equilibrium pH with a slope approaching 3 for trivalent ions because three hydrogen ions are liberated per extracted metal centre. This pH sensitivity is used in scrubbing banks: a controlled acid concentration displaces lighter rare earths from the organic phase while retaining heavier elements. The scrubbing acid concentration is normally maintained between 0.3 mol/L and 1.0 mol/L HCl or HNO3, depending on the element cut point and the saponification degree of the solvent. A lower scrub acidity leaves light rare earth impurities in the loaded organic, while excessive scrub acidity strips part of the target heavy fraction and increases downstream acid consumption. Temperature also modifies selectivity by altering dimerization constants and hydration enthalpies; operations above 45 °C can narrow separation factors and become counterproductive if the feed contains easily hydrolysed elements.
Partial saponification of D2EHPA is used industrially to reduce the acid release into raffinate and to maintain extraction capacity at moderately low pH. The saponification reaction with sodium hydroxide is expressed as (HX)2(org) + NaOH ⇌ NaX·HX(org) + H2O, and the degree of saponification is usually controlled between 0% and 40% of the total dimer equivalents. Increasing saponification raises the effective free extractant anion concentration and shifts extraction to lower aqueous equilibrium acidities, but it also introduces alkali metal mass transfer into the organic phase and promotes emulsion stabilisation. Stripping of loaded D2EHPA is conducted with mineral acid, typically 2.0–4.0 mol/L HCl for chloride circuits, 1.0–3.0 mol/L HNO3 for nitrate circuits, or 1.5–3.0 mol/L H2SO4 when sulphate-based processing is acceptable. The acid consumption per kilogram of rare earth oxide depends directly on the saponification level, the scrubbing efficiency, and the extent of co-extracted impurity loading. In a high-saponification circuit operating at 30–40%, acid consumption in stripping can be 1.2–1.8 times the stoichiometric rare earth equivalent because a fraction of the acid neutralises the sodium or ammonium species carried in the organic phase. Ammonia saponification produces ammonium sulphate or ammonium chloride aqueous effluents that must be treated before discharge, and the choice between sodium and ammonium saponification is therefore linked to effluent compatibility and salt disposal permits. Stripping acid concentration is not an independent variable: if the strip acid falls below 1.0 mol/L, rare earth recovery declines sharply, while concentrations above 4.0 mol/L accelerate extractant hydrolysis and increase solvent degradation products. The strip liquor should be monitored for free acidity by potentiometric titration, and the phase disengagement time after stripping should remain below 180 s in a laboratory shake-out test; persistent slow break indicates excessive saponification, diluent oxidation, or surfactant contamination.
Because bastnasite-derived chloride leach solutions frequently retain residual fluoride, iron, thorium, and aluminium, the raw pregnant liquor is seldom suitable for direct D2EHPA extraction without preconditioning. Fluoride forms stable aqueous complexes with rare earth ions and can reduce extraction efficiency when present above trace levels, while calcium and magnesium introduce competing cation-exchange reactions that consume extractant capacity. Cerium oxidation state is controlled before rare earth separation because Ce(IV) is extracted more strongly than Ce(III) and can accumulate in the organic phase as a difficult-to-strip tetravalent species. In industrial practice, the leach liquor is filtered to ≤5 µm suspended solids, then adjusted to a pH compatible with the downstream rare earth separation target. If the feed contains ferric iron above 10 mg/L, the ferric ion is commonly reduced to ferrous iron using iron powder or another compatible reducing agent, because Fe(III) competes strongly with heavy rare earths and can occupy extractant sites at low pH. Removal of thorium is performed before D2EHPA contact when monazite is the feedstock, as D2EHPA extracts tetravalent actinides and lanthanide distributions can be distorted. Aluminium is frequently present at 50–500 mg/L in ion-adsorption clay leachates and contributes to stable emulsion formation if the organic phase is overly saponified. The preconditioned feed is cooled or heated to the operating temperature, and redox potential is monitored because dissolved oxygen and residual oxidants can re-oxidise ferrous iron in the feed tank. These feed management steps are considered routine, but their omission is a common source of poor stage efficiency and crud accumulation in continuous circuits.
Ferric ion is extracted by D2EHPA at lower equilibrium pH than trivalent rare earths and is also stripped more slowly under conventional acid conditions. When Fe(III) enters the extraction bank at concentrations above 10 mg/L, it accumulates in the organic phase and can displace target rare earth ions, reducing effective capacity. The resulting loaded organic carries iron through the scrubbing stages, where ferric ion may hydrolyse to colloidal ferric hydroxide and stabilise organic-in-water dispersions. In continuous mixer-settlers, this manifests as a rising interfacial rag layer, increasing organic entrainment in the raffinate, and inconsistent phase continuity. The redox chemistry is therefore controlled before extraction: ferrous iron is only weakly extracted at the operating pH, so reduction is maintained by maintaining the feed oxidation-reduction potential below approximately +200 mV versus Ag/AgCl, depending on pH and chloride concentration. The reduced liquor is protected from air ingress through blanketed feed tanks and low-shear transfer pumping. If Fe(III) breakthrough nevertheless occurs, the scrubbing bank pH is reduced to 0.5–1.0 to reject ferric species, but this action simultaneously strips a portion of lighter rare earths and increases the recycle load. In severely contaminated solvent, a reductive strip with hydrochloric acid containing a reducing agent may be required; such treatments are operationally expensive and are avoided by rigorous feed oxidation state control. The agitator speed in mixer compartments handling reduced feed should be set to avoid vortex entrainment of atmospheric oxygen, and the organic-phase iron concentration should be monitored by inductively coupled plasma optical emission spectrometry according to ISO 11885 or an equivalent validated method. Aqueous-phase iron levels in the raffinate are also tracked because downstream rare earth precipitation is sensitive to iron carryover, and overall recovery credits are reduced if the final oxide contains iron above specification.
Continuous rare earth separation with D2EHPA is typically performed in countercurrent mixer-settler batteries or pulsed sieve-plate columns. Mixer residence time is normally held between 3 min and 8 min, while the settler area is sized to provide a specific flow capacity that maintains phase disengagement. The stage efficiency of a mixer-settler is degraded less by the extraction kinetics than by physical phenomena such as entrainment, internal recycle, and interfacial rag accumulation. Organic entrainment in the aqueous raffinate above 50 ppm is an actionable threshold in many circuits because it represents both solvent loss and organic contamination of downstream precipitation. Aqueous entrainment in loaded organic above 100 ppm can transfer acid, impurities, and dissolved salts into the stripping bank, where they degrade strip performance and increase the impurity content of the final strip liquor. The phase disengagement time measured in a standard shake-out test should remain below 180 s at 25 °C; values between 180 s and 300 s indicate incipient crud or excessive saponification, and values above 300 s usually require solvent purification or replacement. Table 1 summarises the operational windows used to evaluate continuous D2EHPA circuits.
| Parameter | Reported industrial range | Analytical or control method | Consequence beyond range |
|---|---|---|---|
| D2EHPA concentration in aliphatic diluent | 0.8–1.5 mol/L | Acidimetric titration per ASTM D974 | Viscous organic phase, slow phase break, solvent loss |
| Saponification degree | 0–40% | Potentiometric titration of equilibrated acid | Emulsion stabilisation, sodium transfer, rag formation |
| Extraction equilibrium pH | 1.2–2.2 | ISO 10523 pH electrode measurement | Loss of selectivity or precipitation of feed constituents |
| Scrub acid concentration | 0.3–1.0 mol/L HCl or HNO3 | Potentiometric titration | Poor light rare earth rejection or premature heavy rare earth strip |
| Strip acid concentration | 2.0–4.0 mol/L HCl | Potentiometric titration | Low rare earth recovery or accelerated solvent degradation |
| Operating temperature | 25–45 °C | Thermowell and calibrated transmitter | Hydrolysis, evaporative diluent loss, selectivity shift |
| Organic entrainment in raffinate | <50 ppm | Optical or gravimetric entrainment monitor | Solvent loss, downstream precipitation fouling |
| Aqueous entrainment in loaded organic | <100 ppm | Karl Fischer or visual phase clarity | Impurity transfer, strip acid dilution |
During continuous pilot campaigns treating ion-adsorption clay leachate at 25 °C, solvent quality is maintained through daily density and viscosity checks, acid number titration, and trace water determination. The organic density is measured by ASTM D4052, kinematic viscosity by ASTM D445, and water content by ASTM D6304. Flash point of the diluent is determined by ASTM D93 to confirm that solvent handling remains within the original safety classification. The analytical programme is used to detect solvent oxidation, hydrolysis product accumulation, and dilution errors before they appear as hydraulic upset. In parallel, aqueous streams are analysed by inductively coupled plasma optical emission spectrometry using ISO 11885 for rare earth distribution, iron, aluminium, calcium, magnesium, silicon, and phosphate. The analytical frequencies are risk-based: strip liquor and raffinate are analysed every shift, while full solvent quality panels are performed weekly and after any sustained phase separation upset. These measurements are not regulatory in isolation but are incorporated into the plant environmental management system under ISO 14001:2015 clause 8.1, which requires operational control of significant environmental aspects including solvent fugitive loss and aqueous discharge. The laboratory data are also used to calculate stagewise distribution coefficients and to validate whether the operating point remains within the design basis of the solvent extraction solvent inventory.
D2EHPA undergoes slow hydrolytic degradation in acidic service, producing mono(2-ethylhexyl)phosphoric acid, phosphoric acid, and 2-ethylhexanol as degradation products. The monoester is more hydrophilic than the parent D2EHPA and can act as a surfactant, reducing interfacial tension and promoting stable emulsion formation. Hydrolysis is accelerated by temperatures above 45 °C, by prolonged contact with mineral acids above 4 mol/L, and by aqueous acid carryover into the organic storage tank. Oxidative degradation is primarily associated with diluent auto-oxidation rather than with the D2EHPA molecule itself, and it produces hydroperoxides, alcohols, and carboxylic acids that alter phase behaviour and can co-extract alkali and alkaline-earth impurities. Oxidative stability is improved by using a diluent with low aromatic content and low unsaturation, and by excluding air from blanketed solvent tanks. The solvent inventory should be purged or continuously treated to keep degradation products below the point at which phase disengagement time exceeds 300 s or the interfacial tension falls below 10 mN/m. Published data for the specific interplay between radiolytic degradation and D2EHPA in rare earth processing are limited, but radiolytic damage is considered relevant only when monazite-derived thorium and uranium concentrate in the organic phase over extended campaigns. Where high radiation fields are present, solvent replacement rates are risk-assessed using measured activity accumulation in the organic inventory, and the organic is segregated from unprotected elastomer seals. Acidic hydrolytic degradation products increase the aqueous solubility of the organic phase and contribute to chemical oxygen demand in the raffinate, so raffinate polishing by activated carbon or solvent stripping is specified when discharge limits are tight.
A diluent containing more than 1 wt% aromatic hydrocarbons alters D2EHPA dimer thermodynamic activity, raises the aqueous solubility of the extractant, and tends to stabilise undesirable emulsions. Aliphatic and isoparaffinic diluents with flash points above 61 °C are therefore preferred for operator safety and phase disengagement. In batch shake-out screening, the replacement of an aromatic-containing kerosene with a narrow-cut isoparaffinic solvent reduces organic loss in raffinate and lowers the equilibrium pH shift during extraction. The diluent viscosity at operating temperature also influences pregnant liquor droplet coalescence: high-viscosity solvents increase the settling path residence time and require larger settler area for the same throughput. The solvent composition is maintained by monitoring the aromatic content and boiling range at least quarterly by gas chromatography, while the physical properties are tracked by ASTM D4052, ASTM D445, and ASTM D93. When fresh diluent is added to compensate for evaporative or entrainment losses, the addition is made through a closed charging system to limit atmospheric moisture ingress. The combined solvent should be brought to the target D2EHPA concentration by acidimetric titration and then pre-equilibrated with acid before returning to the extraction circuit.
Monazite-derived rare earth feedstocks contain thorium and uranium, and D2EHPA extracts tetravalent thorium even at low equilibrium pH. Thorium carryover into the rare earth separation circuit is constrained because thorium accumulates in the organic phase, competes with rare earths for extractant sites, and can produce a radioactive solvent inventory requiring specialised handling and waste management. The feed thorium concentration is therefore reduced before extraction by selective precipitation or solvent extraction with a thorium-selective extractant. In several industrial flow sheets, the thorium precipitation step targets a liquor thorium concentration below 1–5 mg/L before rare earth extraction commences; however, published data for a universal threshold are limited because the acceptable level depends on the final product specification, the radiation monitoring programme, and the solvent purification sequence. Uranium is partially co-extracted with D2EHPA and is controlled through the same pre-extraction removal circuit or through a separate uranium recovery step. Process tanks, settlers, and pipework handling monazite-derived liquor are monitored for surface radiation and are included in the radiation protection plan. The corresponding analytical measurements are performed by inductively coupled plasma mass spectrometry or alpha spectrometry, and the discharge of thorium-bearing effluents is evaluated against the facility radiological discharge authorisation rather than a single generic standard. The solvent is sampled for thorium and uranium activity at defined intervals, and if activity exceeds the facility action limit, the organic is shifted to a thorium stripping operation with concentrated acid and then washed with water before returning to rare earth service.