D2EHPA vs Cyanex 272: Choosing the Right Extractant

Last updated: 2026-10-21

For cobalt–nickel separation, the D2EHPA vs Cyanex 272 decision turns on one number: Cyanex 272 separates Co from Ni roughly an order of magnitude better than D2EHPA at the same equilibrium pH, while D2EHPA carries more metal per litre at lower cost. Pick Cyanex 272 for Co/Ni and Mn rejection duty; pick D2EHPA when capacity, strip efficiency, and reagent economics dominate.

This page is part of our pillar guide on liquid-liquid separation, and it gives conditional, rule-based guidance rather than a vague “it depends”. Every rule below assumes an acidic (cation-exchange) system in a kerosene-type diluent, and every number is an order-of-magnitude planning figure to be confirmed by bench shake-out tests on your actual liquor.

Key Takeaways

• Cyanex 272 (a phosphinic acid) gives far better Co/Ni and Co/Mn separation than D2EHPA.
• D2EHPA extracts at lower pH, loads more metal per litre, and costs meaningfully less.
• D2EHPA co-extracts calcium and magnesium strongly; Cyanex 272 largely avoids this.
• Mn removal before Co/Ni SX is the classic Cyanex 272 or co-extractant-boosted duty.
• Pilot testing on real liquor is non-negotiable: selectivity shifts with ionic strength.

What Is the Difference Between D2EHPA and Cyanex 272?

D2EHPA is a dialkyl phosphoric acid extractant — di-(2-ethylhexyl) phosphoric acid — whose acidic phosphorus–oxygen groups exchange protons for metal cations at the organic–aqueous interface. D2EHPA is a workhorse of hydrometallurgy that combines high loading capacity, strong extraction of divalent transition metals, and low unit cost, which is why it appears in the majority of commercial acidic solvent extraction circuits built before the 2000s. Its weakness is a relatively flat selectivity ladder: calcium, magnesium, manganese, cobalt and nickel sit close together, so a liquor designed to pick up one of them usually picks up several.

ASME VIII Div 1 vs Div 2 overview

how Solvent Extraction is decided

Cyanex 272 is a phosphinic acid extractant — bis(2,4,4-trimethylpentyl) phosphinic acid — with a more heavily branched, more hydrophobic substituent shell around the same acidic P–OH functionality. Cyanex 272 is a higher-pH, more sterically shielded extractant that pushes its extraction ladder upward and apart, especially widening the gap between cobalt and nickel and between cobalt and manganese. That steric crowding is precisely what costs it some loading capacity relative to D2EHPA, and what makes it the default answer whenever Co/Ni or Co/Mn separation is the core objective.

Both reagents belong to the same family of acidic organophosphorus extractants, and both release a proton when they bind a metal cation — meaning your circuit’s pH rises with every gram of metal loaded, and your neutralization consumption is tied directly to your production rate. This shared mechanism is why comparing them is fair, and why the D2EHPA vs Cyanex 272 choice is genuinely a design decision rather than an apples-to-oranges one. For contrast: copper SX does not use either reagent; it uses chelating oximes (LIX/Acorga families), which is a reminder that extractant selection is always duty-specific.

The Chemistry Behind Selectivity: pH50, ΔpH50, and Extraction Order

Acidic extractants behave as liquid cation exchangers. A simplified equilibrium is:

M²⁺(aq) + 2(HR)₂(org) ⇌ MR₂·(HR)₂(org) + 2H⁺(aq)

Because H⁺ is a product, extraction improves as pH rises. Each metal–extractant pair has a pH50 — the equilibrium pH at which 50% of the metal transfers — and the vertical gaps between metals, the ΔpH50 values, are what your circuit actually exploits. A separation is practical when the ΔpH50 between the metal you want and the metal you must reject is large enough (typically ≥2 pH units for clean single-stage separation, less if you can use many stages, scrubbing, or pH control per stage).

For both reagents, the general extraction order for the common base-metal suite runs roughly:

Ca, Mg (D2EHPA) ≲ Mn < Co < Ni, with Zn and Cu extracting far ahead of the group, and Fe³⁺ extracting at almost any pH.

Two critical nuances:

1. The order is similar, but the spacing is not. D2EHPA sits low on the pH scale — it pulls many divalent metals down into the pH 2–4 window — while Cyanex 272 sits roughly 1–2 pH units higher for Co and dramatically higher for Ni. Practically, Ni that D2EHPA co-loads at pH 5 stays largely in solution under Cyanex 272 conditions that still load Co well. This is the origin of Cyanex 272’s reputation for nickel-cobalt separation.
2. Calcium and magnesium flip the decision in many feeds. D2EHPA extracts Ca and Mg within or near the same window as Mn, Co and Ni, so saline or dolomitic liquors contaminate the loaded organic and eventually precipitate gypsum or calcium carbonate inside the plant. Cyanex 272 extracts both much more weakly, which is why phosphinic-acid circuits tolerate high-hardness feeds that would slowly choke a D2EHPA circuit. Zn and Cu, by contrast, extract well ahead of the whole Ni–Co–Mn group with either reagent, so zinc removal is usually a front-end D2EHPA (or other) step rather than a reagent-selection problem.

Do not build a flowsheet on literature ΔpH50 values alone: they move with ionic strength, sulfate/chloride balance, extractant concentration, and temperature. Bench shake-out tests on your liquor are the only reliable source, and a structured pilot campaign is the only way to fix stage counts with confidence.

D2EHPA vs Cyanex 272: Head-to-Head Comparison

The table below condenses the differences that matter at flowsheet level. Values are typical ranges for sulfate systems in aliphatic diluents; treat them as planning figures.

Read the table column-by-column and you get the two dominant design logics: D2EHPA wins the capacity/cost row; Cyanex 272 wins every row about selectivity and hardness tolerance. The rest of this page turns that observation into explicit rules.

When Should You Choose D2EHPA vs Cyanex 272?

The D2EHPA vs Cyanex 272 question resolves into a small set of conditional rules. Apply them in order, because the first matching rule usually decides the reagent.

• For nickel-cobalt sulfate from MHP where manganese rejection matters, Cyanex 272 is preferred because its Co/Mn and Co/Ni gaps are the widest available among acidic extractants, letting you strip a saleable Co salt without a punishing scrub section. Typical plants run Co extraction at pH 5–5.5 and leave Ni substantially in the raffinate.
• For Ni/Co battery-recycling black mass liquors with manganese and some calcium, Cyanex 272 is preferred because Mn removal and hardness tolerance both matter, and the mixed Ni/Co/Mn feed punishes D2EHPA’s compressed selectivity.
• For zinc, indium, gallium, germanium, vanadium or molybdenum duty where capacity and cost dominate, D2EHPA is preferred because these metals extract far ahead of the accompanying Ca/Mg/Mn crowd, so the compressed ladder is irrelevant, and D2EHPA’s higher loading per litre shrinks mixer-settler volume — see our mixer-settler design notes.
• For rare-earth separations where capacity, cost and ready availability dominate, D2EHPA is preferred because REE duty is staged by pH control anyway, and its higher saturation loading reduces organic inventory. (Many REE plants historically use P507, a close D2EHPA relative, for the same logic.)
• For high-hardness feeds (significant Ca/Mg) under any duty, avoid D2EHPA unless you accept a dedicated Ca/Mg management scheme; Cyanex 272’s weak Ca/Mg extraction is often worth its price premium by itself.
• For zinc or REE circuits in regions with constrained reagent supply, D2EHPA is additionally preferred because it is produced by multiple manufacturers under different trade names, while the phosphinic acid supply chain is narrower.

If two rules conflict — say a Co/Ni duty with an unusually high Co/Ni ratio and cheap liquor — the conflict is resolved by economics: estimate the annual reagent cost difference against the mixer-settler volume and scrubbing stages each option demands. Our experience on anonymized projects is that the selectivity premium of Cyanex 272 pays for itself whenever cobalt product purity is contractual; below that threshold, capacity and cost win.

Flowsheet Consequences: Stages, Scrubbing, Saponification and Upstream Cleanup

The extractant choice cascades through the whole circuit, not just the SX building.

Number of stages. Because Cyanex 272’s ΔpH50 gaps are wider, it needs fewer extraction stages for a given purity target — often 3–4 extraction plus 2 scrub stages for Co/Ni work, versus 5–7 extraction plus heavier scrubbing for a D2EHPA attempt at the same separation. Conversely, D2EHPA’s high loading capacity can reduce organic flow and stage area in duties where selectivity is easy (zinc, REEs), which partially offsets the larger stage counts that low selectivity forces elsewhere.

Scrubbing. With D2EHPA on a mixed feed, expect a dedicated scrub circuit to wash co-extracted Ca, Mg and Mn from the loaded organic; this consumes wash acid and adds a settler or two. With Cyanex 272 on Co/Ni duty, scrubbing is lighter, usually aimed at residual Ni.

Saponification / neutralization demand. Both reagents release H⁺ on loading, so pH control is mandatory; but because Cyanex 272 operates at higher pH, its circuits consume neutralizing agent at a different point and must guard against local precipitation of Ni/Co hydroxides near the pH setpoint. In both cases, saponifying the organic upstream (NaOH or ammonia pre-neutralization) smooths pH profiles and reduces third-phase surprises; the neutralization stoichiometry is roughly 1 mole of base per mole of H⁺ released, so your caustic or ammonia budget scales with metal loading, not with reagent price.

Impurity removal before SX. Iron and aluminium must be removed ahead of either reagent — Fe³⁺ extracts too strongly and strips too poorly. Copper and zinc are usually removed first as well, since they extract far ahead of the Co/Ni/Mn group with both reagents. Skipping this front end is the most common cause of “the extractant degraded” reports, when the real culprit is Fe³⁺ accumulating in the organic — a topic we cover in extractant degradation. Equipment selection follows: conventional mixer-settlers suit most of these duties, while centrifugal extractors help where staged residence time is short or floor space is tight.

Raffinate and recycle loops. The reagent choice also shapes what happens downstream of SX. With Cyanex 272 on Co/Ni duty, the nickel-bearing raffinate exits the SX block clean enough to crystallize directly or feed a Ni electrowinning loop; with D2EHPA, expect an additional polishing step because carried-over Mn and Ca follow the aqueous. Design the recycle loops with this in mind — a Mn-loaded raffinate recycled to leaching will concentrate Mn until it breaks through everywhere at once, a failure mode that usually takes months to diagnose and days to fix once understood.

Why Does My SX Circuit Keep Fouling? Troubleshooting the Common Failure Modes

Circuits built on either reagent fail in patterns that are predictable enough to list. If you are commissioning or debottlenecking, walk this numbered checklist before blaming the reagent.

1. Calcium and magnesium loading (D2EHPA circuits). If gypsum or CaCO₃ solids appear on impeller blades and inter-stage ports, the organic is co-loading hardness. Verify Ca in stripped liquor, and consider a dilute-acid pre-scrub or a reagent switch.
2. Gelling at high metal loading. A third phase or gel near the interface usually means the organic is over-saturated. Check loaded-organic metal concentration against design capacity (for D2EHPA, roughly 0.05–0.10 mol metal per mol reagent dimer is a planning band) and raise O/A ratio or extractant concentration.
3. Extractant degradation. Falling pH50 drift and rising aqueous losses point to hydrolysis or oxidation, often driven by hot acidic chloride streams or residual oxidants. Monitor reagent concentration quarterly and review our degradation resource.
4. Crud at the interface. A dark, mousse-like layer that holds solids and organic is usually fine solids (silica, Fe precipitates) stabilized by the extractant. Treat it with interfacial adsorbents and remove solids upstream — never pump crud back into the organic; our crud treatment resource covers detection and disposal.
5. Entrainment losses. Organic in the raffinate (visible haze, or >50 ppm on analysis) costs money and poisons downstream steps. Check settler design and consider coalescers; our entrainment control page lists the usual mechanical causes.
6. pH runaway during loading. Since both reagents release H⁺, a stalled pH-probe or undersized base addition lets extraction collapse mid-cascade. Instrument each stage independently rather than controlling on feed pH.
7. Temperature drift. Kinetics and phase separation degrade below ~20 °C, particularly for the more viscous Cyanex 272 systems. Keep the SX building heated and record temperatures in your daily log — most “selectivity loss” complaints in winter are temperature artifacts.

Cost and Availability: What the Reagent Price Tag Does Not Tell You

Sticker price is the least informative line in the comparison. D2EHPA is produced at commodity scale by several manufacturers and typically costs a fraction of Cyanex 272 per kilogram — market quotes vary, but planning ratios of 2–4× are common for the phosphinic acid. Yet the true comparison is total circuit cost: Cyanex 272’s wider selectivity can delete scrub stages, reduce neutralization in later circuits, protect downstream Ni circuits from Mn contamination, and avoid the capex of hardness-treatment steps that a D2EHPA flowsheet would need. On a Co/Ni refinery processing thousands of tonnes per year, one avoided scrub circuit or one avoided Ca/Mg removal train can outweigh the reagent premium several times over. On a zinc or REE circuit where selectivity is easy, the same premium buys nothing — which is exactly why the answer is conditional rather than universal.

Availability also matters for risk: D2EHPA has multiple global suppliers and substitution paths (other dialkyl phosphoric acids), while Cyanex 272’s phosphinic chemistry is supplied by a narrower set of producers. For plants in remote locations, we advise holding 3–6 months of organic inventory for either reagent, plus a diluent compatibility check before committing to a supply chain.

Two cost lines are routinely forgotten in early-stage budgeting. The first is inventory carrying cost: an SX circuit running 20 vol% extractant in a 500 m³ organic inventory holds roughly 80–100 tonnes of reagent, and that working capital — plus the diluent top-up losses of 0.1–0.5 L per m³ of aqueous treated, depending on entrainment control — belongs in the operating model from day one. The second is the cost of getting the choice wrong: switching extractants after commissioning means flushing and disposing of the entire organic inventory, re-optimizing stage pH setpoints, and usually re-qualifying the product. A few weeks of comparison testing before detail design is cheap insurance against a seven-figure rework.

Which Applications Map to Which Reagent?

Mapping duty to reagent by application:

Battery recycling deserves special mention: black mass leachates combine Mn, Co, Ni and often Ca/Mg from process water, so most modern flowsheets use both reagents in sequence — see our battery recycling application page for the full chain. Primary Ni/Co refining from MHP follows the same logic at larger scale, described on the nickel-cobalt refining page. Lithium circuits are a different family again (the extraction chemistry of lithium brine processing uses very different reagents), but the same pH50 discipline applies. All application flowsheets are built on the equipment platform described in the liquid-liquid separation pillar, and process-wide troubleshooting resources live under resources.

Frequently Asked Questions

What does D2EHPA stand for and what is its chemistry?
D2EHPA is di-(2-ethylhexyl) phosphoric acid, a dialkyl phosphoric acid extractant. It carries a single acidic P–OH group that exchanges H⁺ for metal cations, extracting divalent metals such as Zn, Co, Ni, Mn, Ca and Mg from acidic sulfate or chloride liquors. It is typically run at 10–30 vol% in a kerosene-type diluent and strips with dilute mineral acid.

Is Cyanex 272 stronger than D2EHPA for cobalt–nickel separation?
Yes — by a wide margin. Cyanex 272’s Co/Ni gap is roughly an order of magnitude larger than D2EHPA’s at comparable conditions, which is why essentially all modern Co/Ni refineries choose the phosphinic acid for the Co/Ni split. The premium is roughly 2–4× per kilogram, usually repaid by fewer scrub stages and cleaner cobalt product.

Why does D2EHPA pull in so much calcium?
Its extraction window for Ca overlaps the Mn–Co–Ni window, so any hardness in the feed loads alongside the target metal. Over weeks, Ca accumulates in the organic and precipitates as gypsum or CaCO₃ on mixers and settlers. Cyanex 272 extracts Ca and Mg far more weakly, which is a decisive advantage on hard-water or dolomitic feeds.

What is pH50 and why should I care?
pH50 is the equilibrium pH at which half of a metal transfers to the organic phase. The gaps between metals (ΔpH50) determine how many stages you need and how clean your separation is. A ΔpH50 of ≥2 units usually allows sharp separations in 3–4 stages; gaps below ~1 unit force heavy scrubbing or make the separation impractical.

Can I use both D2EHPA and Cyanex 272 in one plant?
Yes, and many battery-recycling and Ni/Co plants do: D2EHPA first for Zn/Cu/Mn or front-end impurity removal where capacity is cheap, then Cyanex 272 for the Co/Ni split. Keep the two organics in separate circuits with dedicated settlers, and audit entrainment between them — cross-contamination degrades both.

How do I verify which extractant my liquor actually needs?
Run bench shake-out tests on representative feed: 6–10 equilibrium points per reagent across a pH ladder, then isotherms at your planned O/A ratio. Literature ΔpH50 values shift with ionic strength, so only your own liquor settles the D2EHPA vs Cyanex 272 question — a pilot campaign of a few weeks is enough to fix stage counts and pH setpoints.

How long do these extractants last in service?
With clean feed and moderate temperatures, both reagents serve for years, losing roughly a few percent per month to solubility and entrainment. Degradation accelerates with hot acidic chloride, strong oxidants (Fe³⁺, chlorine, peroxide) and high temperatures above ~50 °C. Quarterly reagent-concentration analysis is the standard monitoring practice.

Conclusion

The D2EHPA vs Cyanex 272 decision is not a matter of taste: Cyanex 272 for Co/Ni and Mn-rejection duty, D2EHPA for zinc, rare-earth and impurity-removal duty where capacity and cost dominate, and a two-reagent flowsheet when a mixed feed demands both. Because ΔpH50 values move with your specific liquor, the fastest way to a bankable answer is a bench-and-pilot campaign on real solution — talk to our pilot testing team to turn these rules into a flowsheet.

References

• U.S. Geological Survey, Mineral Commodity Summaries — Nickel and Cobalt chapters (annual editions).
• K.C. Sole and J.B. Hiskey, research papers on organophosphorus acid extractants and their metal-extraction behavior (published in Hydrometallurgy and Solvent Extraction and Ion Exchange).
• Cytec Industries, “Cyanex 272” technical brochure (phosphinic acid extractant product literature).
• Z. Kolarik, reviews on acidic organophosphorus extractants: equilibria, dimerization and extraction mechanisms (Solvent Extraction and Ion Exchange and related journals).
• T.C. Lo, M.H.I. Baird and C. Hanson (eds.), Handbook of Solvent Extraction, Wiley, 1983.
• F. Habashi, A Handbook of Extractive Metallurgy, Wiley-VCH.
• General literature in Hydrometallurgy and Solvent Extraction and Ion Exchange journals on pH50 isotherms and selectivity of acidic extractants.

Changelog

• 2026-10-21 — First published.

Related reading

• Stage Efficiency in practice
• What Is Hydrometallurgy? Definition, Process, Applications