The growing global demand for rare earth elements (REEs)—critical to clean energy, advanced electronics, and defense technologies—has exposed the environmental costs of conventional extraction. Traditional pyrometallurgical and hydrometallurgical methods rely on high temperatures and aggressive reagents, generating toxic byproducts and requiring significant energy consumption.
Biological approaches, by contrast, offer a more sustainable path. Microbes can produce natural organic acids that dissolve REE-bearing minerals (bioleaching) or bind and concentrate dissolved metals on their cell surfaces (bioseparation).
Bioleaching converts solid feedstocks into soluble REE forms, while bioseparation (or biosorption/desorption) selectively captures and enriches these metals from solution. Together, they provide a foundation for green REE recovery—scalable, tunable, and compatible with circular-economy principles.
Biological systems can achieve REE recovery without the toxic solvents or extreme conditions used in traditional metallurgy—and can even be engineered for element selectivity.
Why Modeling Is Critical for Bioleaching and Bioseparation
REE recovery involves complex aqueous chemistry. Feedstocks such as red mud, coal fly ash, or electronic waste contain diverse metals and minerals, each with unique solubilities and reactivities. When these solids meet a leaching solution—biological or chemical—the process can be affected by multiple species including REE complexes, oxides, phosphates, and competing cations such as Fe³⁺ or Ca²⁺.
Because direct experimentation across this multidimensional space is infeasible, thermodynamic modeling provides the roadmap. It helps researchers predict solubility limits, quantify complex formation with organic ligands, optimize pH and ligand type, and anticipate competitive adsorption or precipitation.
Modeling thus connects thermodynamic theory with biotechnological application—allowing scientists to translate microbial processes into industrial separation processes.
Building a Thermodynamic Model with OLI’s MSE Framework
A robust thermodynamic model must capture both chemical equilibria and phase behavior in multicomponent aqueous systems. OLI’s Mixed-Solvent Electrolyte (MSE) model offers such a foundation, combining: (1) an equation of state for individual species’ standard properties, (2) an excess Gibbs energy formulation for non-ideal solutions, and (3) algorithms that solve phase and chemical equilibria simultaneously across solid, liquid, and vapor phases.
How the Model Is Constructed
The MSE approach begins from the bottom up: parameterizing simple subsystems (binary and ternary mixtures) and extending to complex REE systems using validated thermodynamic data. Figure 1 illustrates how the MSE model reproduces experimental data for selected systems in which rare earths interact with organic complexing agents that are relevant to biological systems.
Figure 1. Experimental and calculated behavior of rare earth systems with complexing ligands. (a) Neodymium–citrate solubility vs. pH (b) Lanthanum–tartrate solubility (c) Nd–citric acid titration curves (d) Ce–acetate complexation (adapted from Fujita et al., 2025).
The MSE model reproduces key equilibrium behaviors observed experimentally, demonstrating its predictive power for designing biologically inspired leaching environments.
Modeling Bioleaching with the MSE Framework
Once validated, the Mixed-Solvent Electrolyte (MSE) model can be applied to simulate how REEs behave during biological leaching from complex feedstocks. We have used this approach to study the dissolution of rare earths from coal fly ash, a heterogeneous material containing silicon, aluminum, iron, calcium, and other elements. These matrix elements interact with leaching solutions and influence the release, complexation, and precipitation of REEs. By simulating these reactions, the MSE framework helps identify which acids, pH ranges, and concentrations favor REE solubilization while minimizing the formation of secondary solids that can re-immobilize the metals.
1. Predicting REE Solubility and Speciation
Figure 2 illustrates the predicted solubility of neodymium hydroxide, Nd(OH)₃, as a function of pH in three different aqueous systems: a simple sodium chloride solution (NaCl) with pH adjusted using HCl, and NaCl solutions containing either acetic acid or gluconic acid. The results show a strong pH dependence in all cases, with solubility sharply decreasing as pH increases. In the absence of complexing ligands, Nd(OH)₃ precipitation dominates above pH 7, keeping the REE concentration extremely low.
When organic acids are introduced, however, the MSE model predicts strong increases in solubility, particularly in the presence of gluconic acid. This is because gluconic acid contains multiple hydroxyl and carboxylate groups that can coordinate Nd³⁺ ions simultaneously, forming strong, soluble complexes that prevent hydroxide precipitation over a wider pH range. Acetic acid provides some enhancement as well, but its single carboxyl group forms weaker bonds, resulting in lower solubility.
Figure 2. Solubility of neodymium hydroxide as a function of pH at 25°C in a 0.1 m NaCl aqueous solution (teal) and in the presence of 0.1 m acetic acid (green) or 0.1 m gluconic acid (purple). The pH was adjusted by varying HCl and NaOH concentrations.
These simulations directly support laboratory findings that microbial production of polyhydroxy acids—such as gluconic or citric acid—can substantially increase REE recovery from ores and waste materials. Figure 3 complements these predictions by showing the speciation of neodymium in the presence of 0.1 M gluconic acid. The model predicts multiple distinct Nd–gluconate complexes, four of which dominate under typical leaching pH conditions (3–7). At low pH, Nd³⁺ is the predominant species, but as the pH rises, progressively more complexed species (such as NdGlu₂⁻ or NdGlu₃³⁻) become dominant. This shift in speciation explains why REEs remain soluble even as the solution approaches neutral pH: the ligands stabilize the metal ions thermodynamically, delaying or preventing the formation of insoluble hydroxides or phosphates.
Figure 3. Calculated speciation for the system studied in Figure 2 in the presence of 0.1 m gluconic acid. The curves show the concentrations of various complexes between neodymium and gluconic acid as a function of pH.
Such detailed speciation information is valuable for understanding microbial leaching systems. Microorganisms that produce organic acids effectively control the local pH and ligand concentration; the MSE model shows how these parameters translate into actual metal–ligand equilibria that determine how much REE remains in solution.
2. Simulating REE Leaching from Fly Ash
Figure 4 extends the modeling to the leaching of neodymium from a coal fly ash matrix when treated with three acids—hydrochloric, acetic, and gluconic—under varying conditions of pH and total acid concentration. In Figure 4a, the total dissolved Nd is plotted as a function of pH. At low pH (<1.5), all three acids completely dissolve the REE, as expected for strongly acidic environments. As pH increases, however, differences in performance become clear:
- For hydrochloric acid, solubility drops rapidly as pH exceeds 1.6 because the system favors the precipitation of neodymium phosphate (NdPO₄).
• Acetic acid maintains higher solubility up to about pH 3.5 due to moderate complexation with acetate ions.
• Gluconic acid provides the broadest and most effective range, sustaining elevated Nd concentrations up to pH ≈ 4.
These trends demonstrate that, although mineral acids achieve high solubility at high acidity, biogenic organic acids can maintain REE solubility under milder, more environmentally friendly conditions—a crucial advantage for sustainable bioleaching processes.
In Figure 4b, the same systems are analyzed as a function of total acid concentration. The model predicts that HCl requires a relatively high molality to dissolve all available Nd, whereas both acetic and gluconic acids achieve significant dissolution at lower concentrations. The greater efficiency of gluconic acid arises from the balance between its moderate acidity and its strong ability to form soluble complexes with Nd³⁺. This dual functionality explains why microbial lixiviants often outperform simple mineral acids even though they are chemically weaker.
Figure 4. Total dissolved molality of Nd after equilibrating a coal fly ash sample with hydrochloric, acetic, or gluconic acid as a function of (a) pH or (b) molality of added acid.
3. Anticipating Secondary Phase Formation
Figure 5 explores the potential for secondary mineral formation that can complicate REE extraction. The model simulates silicon concentrations in hydrochloric acid leachates of coal fly ash both with and without the inclusion of crystalline silicate phases. The solid lines (with silicate formation) show that as the pH rises above approximately 2, several stable silicate minerals—such as kaolinite (Al₂Si₂O₅(OH)₄), talc (Mg₃Si₄O₁₀(OH)₂), and amphibole-type phases—begin to precipitate. The dashed lines (without silicates) correspond to hypothetical cases where these solids are kinetically inhibited.
This comparison provides two important insights. First, it shows that gel or solid formation observed experimentally in aged leachates likely arises from slow precipitation of such silicate phases. Second, it warns that these solids can act as sinks for REEs through adsorption or co-precipitation, thereby reducing recovery yields.
Modeling these effects allows researchers to determine the safe operational envelope—maintaining leaching at sufficiently low pH to maximize REE dissolution while avoiding conditions that favor silicate reformation. The MSE framework therefore not only predicts equilibrium solubility but also identifies chemical regimes where unwanted phase transformations might occur.
Figure 5. Total dissolved Si concentration predicted by excluding (dashed lines) or including (solid lines) the possible formation of crystalline silicate phases whose formation is kinetically controlled. The difference between the two lines at a given pH indicates the possibility of formation of solids as the leachate sample ages.
4. Practical Implications
By integrating chemical equilibria, complexation, and phase stability, the MSE thermodynamic framework offers a comprehensive view of REE bioleaching systems. The simulations shown in Figures 2–5 demonstrate how the model can quantitatively compare mineral and organic acids for REE extraction efficiency, predict the specific REE–ligand species that control solubility under realistic biogeochemical conditions, anticipate secondary mineral formation that could reduce recovery, and provide a foundation for designing microbial or hybrid processes that operate within optimal chemical windows.
Ultimately, the MSE model acts as a virtual laboratory that accelerates the development of bioleaching strategies by reducing the need for extensive experimental screening. When coupled with microbial process data—such as organic acid production rates or pH profiles—it enables a predictive, system-level understanding of REE recovery mechanisms from industrial waste and natural feedstocks alike.
Concluding Thoughts: Modeling a Greener Future
OLI’s MSE framework allow researchers to explore this bioleaching and bioseparation using fundamental electrolyte thermodynamics—quantifying microbial chemistry, anticipating equilibrium limits, and guiding sustainable design.
As industries seek cleaner ways to supply critical materials, model-informed biotechnology will be important for scaling bio-based REE recovery. The synergy between microbial innovation and computational modeling promises not just greener extraction, but a foundation for circular, low-impact resource management.
For more information, please see the chapter:
- Fujita, D. Park, M. Lencka, A. Anderko, D. Reed, V. Thompson, G. Das, A. Eslamimanesh, and Y. Jiao, “Beneficiation of Rare Earth Elements: Prospects for Biotechnology Deployment,” Chapter 8 in “Rare Earth Elements: Sustainable Recovery, Processing, and Purification,” edited by A.K. Karamalidis and R. Eggert, Wiley (2025), pp. 251-297.
The model capabilities described here are available in OLI software V12.5. For more information, or to consult and share ideas, feel free to contact us at https://www.olisystems.com/contact-us