Strategies for using membrane-based separations to extract critical metals from waste streams

1 Introduction

Many metals are used as components in manufacturing of a range of products and devices which are considered essential to both the economy and national security. Disruptions of supply can have severe outcomes for essential sectors such as defense, information, and communication infrastructure. In addition, some metals are crucial to the transition to clean energy and the global efforts to tackle climate change (IEA, 2021). Globally, two terms are being used synonymously, critical and strategic. Criticality of raw materials (metals, elements, minerals) has been defined by The U.S. National Science and Technology Council in terms of the metal criticality index (C, ranging from 0 to 1.0) based upon the geometric average of three dimensions: (i) supply risk, (ii) consumption growth, and (iii) market dynamics, which is given by Equation 1 (National Science and Technology Council, 2016).Where C is the criticality, R is the supply risk indicator (net import reliance), G is the consumption growth indicator (average growth rate of consumption) and M is the market dynamics indicator (price volatility). Supply risk (R) includes both the dependency of economic systems on specific metals and the likelihood of supply disruptions (vulnerability of supply chains). For instance, excessive reliance on imports, with a high risk of not being adequate to meet industry demand, will risk the continuity of production–thus, of supply–if substitutes cannot be found. Disruptions of supply can have severe outcomes for essential sectors such as defense, information, and communication infrastructure. Traditional mining and mineral production can be very energy-intensive and can have a large environmental impact (land disruption, etc.) for many of these materials, both in terms of energy consumption in extraction processes, but also associated impacts such as global warming. Thus, the concept of supply risk includes geologic, technologic, economic risks; social and regulatory risks; and geopolitical risks. Consumption growth rate (G): consumption of minerals with relatively smaller markets–such as lithium and cobalt–recorded a dramatic increase in a short time as the adoption of electric vehicles started to grow in earnest. The growing market for turbines with permanent magnets–particularly for offshore projects–could dramatically increase rare earth demand over the coming decades. Market dynamics (M): Prices for minerals tend to be volatile, often more so than for traditional hydrocarbons, due to the mismatch between the pace of changes in demand patterns and that of new project development, and also to sudden changes in supply chains. This criticality index (C) has been evaluated at the global level, and more recently for the US (Hayes and McCullough, 2018). On this basis, ordered lists of 50 Minerals Deemed Critical to U.S. National Security and the Economy have been compiled by the U.S. Geological Survey in 2018 and 2022 (US Department of Interior, 2018; Burton, 2022) are shown in Figure 1 The list includes minerals key to all sectors of economy including technology and defense. A more complete recent summary is available in 2023 (US Geological Survey, 2023).

C = R × G × M 1 / 3 (1)

2 Critical metals

Once the Criticality index described above reaches a threshold the metal is deemed to be critical. Currently this threshold has been designated to be 0.3 and as a consequence 35 metals are thus deemed critical. Figure 1 shows the criticality index of metals that were examined by the US Geological Survey in the 2018 report. This list was updated in 2022 (Burton, 2022) and contained 50 critical metals/minerals (we use the term metals rather broadly in this paper–it also includes critical minerals and other critical elements) Notable changes included: addition of nickel and zinc while removal of helium, potash, rhenium and strontium. US Geological Survey (2025) now contains 60 metals/minerals which includes all 50 ion the 2022 list with notable additions of phosphate (agriculture), and boron (national security).

This list includes those needed for a range of different applications such as Platinum Group Metals (PGMs used in a range of reactors as catalysts including in automobiles for use as catalytic converters for converting a range of pollutants to more benign products), lithium (Li; used for batteries for energy storage; ceramics; pharmaceuticals, etc.), uranium (U; used for nuclear fuels; medicine), Rare Earth Elements (REEs mostly in electronics), and potassium (K; fertilizers; glass and ceramics; soaps, etc.,). The 35 in the list in fact also include a few non-metals and lumped together for convenience.

It is expected that the demand for critical metals is likely to triple by 2050 based on recent studies and estimates (Lord, 2019); for some critical metals such as lithium the demand is expected to increase even faster (IEA, 2025). Lithium supplies are expected to remain in tight, despite the expected increase in production in the next few years. Development of technologies for recycling could help reduce the demand by up to 30% (IEA, 2024). A recent review (Watari et al., 2020) has studied the demand for 48 critical metals in 2050. The review is based on 88 studies of the expected demand to 2030 and 2050. The review estimates the annual demand for Li for example, to increase to be over 1,500 kt in 2050 compared to 500 kt currently; Manganese demand will essentially stay stable at 18,000 kt; Nickel to increase 5,500 kt from the current demand of less than 2 kt.

3 Sources of critical metals

Currently the main supply source for critical metals is mining. These include, lithium (Li), manganese (Mn), nickel (Ni), vanadium (V), cobalt (Co), tungsten (W), are essentially mined. While mining is becoming increasingly more efficient and sustainable, it continues to be rather energy inefficient from a zero carbon economy perspective (Kursunoglu, 2025). Mining is a key contributor to green house gas (GHG) emissions and is estimated to contribute almost 10% to the total gas emissions. Energy inputs for mining for critical metals such as Lithium and Nickel are in the range of 100 kJ/kg while for others like Manganese and Chromium in the 10 kJ/kg range (Yip et al., 2016). These numbers indicate that traditional mining and mineral production can be very energy intensive, which comprises a large environmental impact for many of these materials, both in terms of embodied energy but also associated impacts such as global warming. Based on data from the life cycle inventory database Ecoinvent (Wernet et al., 2016), over 10 tons of Pt metal concentrate are required to produce each kg of Pt. The energy inputs depending on the processes used span many orders of magnitude. Some typical energy inputs reported include: sulfidic tailing from gold mining, around 0.8 MJ per kg of tailing (or 286 GJ per kg of Au recovered), of which the majority (96%) is in the form of electricity. These numbers show the high energy use associated with mining.

There is thus an urgent need to develop technologies that are more sustainable for mining critical metals. Aqueous solutions both natural and from industrial waste provide a very attractive alternative option for extracting critical metals (Lord, 2019). Some recent studies on aqua-mining have indicated their economic feasibility. For example, the extraction of Li from geothermal brines has been successfully carried out economically; another study has examined the recovery of Li from desalination brines, also indicating economic feasibility (Ortiz-Albo et al., 2019). Thus, even though aqua-mining is still a technology under development it shows considerable economic potential. For naturally occurring saline aqueous sources, harvesting of energy from these solutions provides an additional incentive for energy savings. One study has reported that it is possible to harvest between 0.5 and 1.0 kWh/m³ (Kristmannsdóttir, 1989). This is not surprising since geothermal aquifers are brackish or saline and at high-temperature.

Several US governmental agencies are engaged in an ongoing effort to measure, verify, compile, and share detailed information about the different contaminants present in many water sources around the US. The information often includes volume, pH, total dissolved solids (TDS), total suspended solids (TSS), location, and source of the wastewater stream, along with contaminants concentrations, including metal ions. This kind of information is conventionally called water chemistry data. Here we share some of the most important sources of such data for the purpose of identifying potential wastewater streams to target for the recovery of critical metals. Funded by the US Department of Energy, The National Energy Technology Laboratory (NETL) maintains the National Energy Water Treatment and Speciation (NEWTS) Federal Database Dashboard and the Economic Data Dashboard: Critical Materials for Energy (Justman et al., 2023; Yesenchak et al., 2025). The former provides the water chemistry data, including critical metals, for different wastewater streams from various fossil energy operations and the latter quantifies the potential value associated with the recovery of these critical materials. US EPA Databases also provide water chemistry data including charge balance of major ions of interest (Wenzlick et al., 2022; Siefert et al., 2023a; 2023b). According to these databases, one is able to identify the type of wastewater most suitable to target for the recovery of a particular critical metal. For example, produced water from conventional crude oil extraction and brackish water are reported to contain promising amounts of lithium; acid mine drainage contains manganese; nickel is present in power gasification effluents; power plant ash landfill leachate contains both manganese and vanadium; wastewater from copper smelting contains arsenic (Yang et al., 2022; Justman et al., 2023).

The aqueous solutions that are excellent candidate for such extraction include both natural sources (such as sea water and geothermal brines which are also rich in minerals and are reported to have concentrations as high as 215 mg/L), and waste streams from various operations including from desalination plants, oil and gas operations, as well as fracking, and cities’ household wastewater (Union of Concerned Scientists, 2013). As an example of the feasibility of such aqua mining, it is reported that 35 mg/L of Li is present in the Permian Basin waste streams and 14 mg/L in California oil fields (Stringfellow and Camarillo, 2019). Based on the water streams at these facilities, 42 million kg can potentially be extracted per year, which exceeds the current US consumption of this metal. In addition, it is estimated that sea water (which contains 0.17 mg/L Li) has the potential to provide an additional 14,000 million kg/year of Li. Of course, with the lower concentration it is a much more challenging aqua-mining operation; and for this we see the potential of reverse osmosis to remove water, thereby concentrating the ions. Our preliminary studies have shown that reverse osmosis using steric based membrane separations has the potential to concentrate the solution almost by two orders of magnitude (Bashmmakh et al., 2025). Of course, with desalination effluents as the feed, one will be dealing with feed solutions in which the ions will already be concentrated. The principles and methodologies used for desalination apply equally well to the extraction of critical metals, too. Desalination has been used for several decades and has been made very efficient (Ng et al., 2021). It is currently at almost 50% reversible efficiency and uses only 2.5 kWh/m³ of water produced. In addition to lowering our reliance on supply chains, aqueous mining would result in much lower environmental impacts on water, air, and land than ore mining.

We have already mentioned Li⁺ ions above. Manganese is one of the heavy metals which has been widely used in steel production. The point sources of manganese are wastewater treatment plants, acid mine drainage from mine quarry operation, and industrial effluent discharge. Sea water consists of ≈10 mg/L of manganese. The preferred pH for manganese removal for some of the processes developed ranges from pH 5 to 7 (Rudi et al., 2020). (At higher pH values there can precipitation of Mn oxide occurs).

4 Methods of extracting critical metals

The idea of waste-water mining for the purpose of extracting critical metals is still at the laboratory scale stage. Nevertheless, by exhaustive consideration of the many factors involved, a recent review suggests that recovery of dissolved critical metals from saline water sources represents a new paradigm as an alternative to and/or augmentation of traditional mining of ores, would reduce reliance on international supply chains and the need for importation of some critical metals. They presented salient arguments (economic, environmental, etc.) for aqueous-phase recovery of metals from saline water sources and brines, comparing availability, environmental impact, energy use and economics of extraction, to support the feasibility of aqueous mining of critical metals (Can Sener et al., 2021). Another recent review assesses the relevance of recovering up to forty-nine metals from wastewater treatment plants from strategic, financial, and mining perspectives (Varennes et al., 2023). These assessments provide the incentive to consider an investigation into the methodologies that may be employed in aqueous mining of critical metals. Another recent comprehensive review compares the multitude of methods for removal of metal ions from various wastewater sources (Qasem et al., 2021). These methods could be classified into adsorption-, membrane-, chemical-, electric-, and photocatalytic-based treatments (Jun et al., 2026). There are advantages and disadvantages associated with various methods of extraction of metal ions from aqueous solutions: economic feasibility, technical maturity, less pollution to the environment, and operation control and ease of automation has been compared across the methods.

In this remaining section we will discuss several other possible technologies (Bashiri et al., 2022) that have been proposed and currently being used to both concentrate very dilute solutions of Lithium to make extraction economically feasible and using direct extraction technologies (Murphy and Haji, 2022; Ben-Elijah et al., 2026). One strategy is to collect the brine solution, for example, in solar evaporative ponds till the concentration of the solution reaches 6,000 ppm (Flexer et al., 2018). This is then followed by introducing several chemicals to the solution to remove some of the other salts and finally obtaining crystals of Li₂CO₃. As can be expected this process takes several months, however because of the use of solar energy it is still considered financially more attractive than other possibilities (Flexer et al., 2018). Other technologies that have been suggested include using High Mg²⁺/Li⁺ precipitant extraction which involves precipitants being added to the solution to achieve super-saturation that results in the metal ions present (Li⁺, Mg²⁺ etc.) to be crystallized in high purity Li complex being the final product (Um and Hirato, 2014). Other similar extractive technologies include the use of aluminate precipitation that introduces aluminum salt (AlCl₃ ⋅ H₂O) and sodium hydroxide (NaOH) which then leads to the precipitation of (LiCl.2Al(OH)₃H₂O) (Liu et al., 2018). Another similar technology includes Magnesium Co-precipitation (Quintero et al., 2020). Solvent extraction is another possibility. Here a competing solvent phase is introduced into the waste lithium solution, facilitating the formation of organic Li⁺ complexes. However, the recovery rate is rather low so far (Butt et al., 2022). Chelating extractants have also been suggested as have acidic extractants (Safarbali et al., 2016; Swain, 2016).

Vanadium removal from industrial wastewaters has been studied earlier than most other heavy metals (Kunz et al., 1976). More recent studies show that vanadium removal is much more challenging than others because of varying speciation at different pH values. Several studies have shown that vanadium sorption is most efficient at lower pH values (3–4), and it is largely related to vanadium speciation (Union of Concerned Scientists, 2013). Several recent reviews detail various methods of removal of V from wastewater; and various resins have been used for ion exchange (Peng et al., 2022; Wołowicz and Hubicki, 2022). V is of significant industrial interest because of its use in redox flow batteries that are essential for storage of energy from solar, wind, and other alternative sources of power.

The main causes of elevated arsenic (while not a critical metal but of significant interest because of its toxic nature) concentrations are natural processes like erosion and leaching through geological formations or human-made sources such as the use of arsenic for industrial processes, uranium and gold mining operations, and the processing of metals, pesticides and fertilizers (Patel et al., 2023; Saravanan et al., 2023). Reviews of arsenic removal technologies, especially from drinking water by ion exchange have appeared (Chen et al., 2020; Alka et al., 2021). Zeolites thus have also been used widely to remove ions which is a separate application of zeolites. We believe that concentrating the solution using zeolite-based membranes, followed by the application of direct extraction technologies, will enable a more cost-effective recovery of critical metals.

Among all methods, adsorption is the most considered method in recent years. It showed easy operation, low-cost, and high sorption capacity. However, disposal of adsorbents after the adsorption process, or regenerating the adsorbent for re-use, remains a big challenge for some adsorbents.

Aqua mining for the purpose of extracting metal ions from wastewater streams often currently utilizes carbon-based, biowaste materials, or polymer adsorbents (Qasem et al., 2021). They all have disadvantages: Carbon-based nanoporous adsorbents, especially activated carbons (ACs), and advanced materials such as carbon nanotubes (CNTs), and graphene (GN), are extensively proposed in the applications of heavy metal removal owing to their tremendous surface. The carbon surface charges can be enhanced by surface functional groups (such as carboxyl, phenyl, and lactone) to improve the heavy metal uptake. Adsorption of metal ions into a polymer such as chitosan depends mainly on the modifications by amine (–NH₂) and phosphoric acid (H₃PO₄) groups, which affect the pH value of the wastewater and thus change the speciation of the metal ions. The expense of chemical modification and the decrease in adsorption upon long-term use of these adsorbent materials are disqualifying for large-scale applications. Although use of biosorbents like biowastes are appealing from an environmental point of view, adsorption conditions will not be reproducible and changes in metal speciation may occur with changing pH of the biosorbents (Rudi et al., 2020). Eventually the biomaterials have to be disposed of, adding CO2 into the environment. Metal organic frameworks (MOFs) offer new and exciting features such as functionalization with amine or other functional groups to better capture metal ions; but they have low stability in water, are expensive to synthesize even in lab quantities, and MOFs have not been produced at industrial scale. Use of electrodialysis with polymeric exchange membranes has also been proposed, but they involve high cost of membranes, membrane fouling, and the need for an electric potential. Supported liquid membranes have also been used but they have problems with poor long-term stability.

Chelation and other chemical methods of separation are not economically sustainable. Ion exchange, electrodialysis, and elution processes are all combined in a highly developed ion-exchange method known as electro-deionization to remove arsenic ions from wastewater (Murphy and Haji, 2022).

5 The use of zeolites for the extraction of critical metals

In this perspective article we make the case for examining the feasibility of zeolite membranes as a class of porous membrane material that have none of the above-mentioned disadvantages, while having the multiple advantages of long-term stability in aqueous solutions, at elevated temperatures and pressures, and acid conditions. An important advantage is the ease of synthesis of aluminated zeolites in industrial quantities, perfected over many decades of industrial use in billions of tons. Some unique features of zeolites include their rigid structure, which makes them ideal as molecular sieves (size-based separations), high adsorption and surface area, high ion exchange capacity, possibility of tuning them to be hydrophobic or hydrophilic, and finally relatively simple regeneration, mostly using steam (Alotaibi, 2025). For extraction by the adsorption mechanism, the availability of frameworks with a range of openings and pore dimensions surpasses other porous adsorbents (Yoshitake and Teruoki, 2010). Zeolites have pore sizes (3–12 Å) with ranges that include the range of diameters of various speciations of hydrated metal ions, even the oxometallate ones. This includes those with small pores with 8-ring structures such as Zeolite A and chabazite; those with medium pores (4–6 Å) and 10 ring structures such as ZSM-5; large pores with 12-ring structure (6–9 Å) with examples such as Zeolite Y and Beta zeolite; extra-large pores with 28-ring channels such as ZMQ-1. Aluminated zeolites have specific interaction sites in the form of strong Bronsted acid sites which offer heightened adsorption. The intrinsic presence of counterions in aluminated versions means these zeolites provide high ion-exchange capacity with the metal ion in question as another mechanism for metal ion extraction. Some zeolites have small pore sizes which easily permit only water molecules and not hydrated metal ions to enter, which allow their use for reverse osmosis as a mechanism of concentrating metal ions in their original solution. Furthermore, zeolites are easily crystallized onto a solid support (e.g., anodized aluminum oxide) with larger pores to permit high-throughput of feed aqueous solutions; such materials are already in large scale use for separation of gases (Wang et al., 2019).

Finally, while many other methods have been proposed to extract metal ions from aqueous solution, they do not talk of recovery. Zeolites allow for the recovery of the desired metal ions, followed by the ability to be able to regenerate the adsorbent and at the same time recover the desired metal ion. This is not feasible with many of the other adsorbents currently being used. On the other hand, zeolites which have retained the desired metal ion in the pores rather than have them permeate completely, can be treated with high concentrations of NaCl in order to drive the desired metal ions out and replace them with the Na⁺ ions that were originally present in the aluminated zeolite, regenerating the original adsorbent for reuse. Another significant advantage is in case of organic fouling of the adsorbent, calcining can completely remove any trapped organics and end up with clean zeolite. Therefore, we feel zeolite membranes are a viable system for the aqua mining of metal ions.

Recently an aluminated zeolite of FAU framework (X) has been used to conduct ion exchange for Mn²⁺ in aqueous solution; the ion exchange process was found to be spontaneous and endothermic (Al-Jubouri and Holmes, 2017). Mn separation and recovery from solution and recovery of manganese from wastewater have recently been reviewed (Zhang and Cheng, 2007; Nkele et al., 2022). Other metals extracted include heavy metals which have been reported in a recent review include arsenic, cadmium, chromium, lead, mercury and nickel (Velarde et al., 2023). Zeolites have also been used extensively for removing Magnesium and Calcium for water softening, often using NaA zeolites (Campanile et al., 2022).

6 Computer simulation of extraction of critical metals

We now briefly describe the important role computer simulations using for example, the method of molecular dynamics (MD) simulations can play in the design of especially membrane-based extraction of critical metals from waste streams. MD for example, is a computer-based molecular modeling technique in which the equations of motion of each particle (or molecule) are solved to trace the time evolution of (usually) classical systems. Static and dynamic properties can then be obtained by time averaging appropriate statistical mechanical expressions (Allen and Tildesley, 2017).

Results from previous simulation studies have clearly shown that potential contributions computer simulations can make in such endeavors. Often simulations may not achieve quantitative accuracy, but they are very reliable for understanding qualitative behavior which is crucial in the design of new processes. Thus, they can very accurately predict the changes in outcome when, for example, state variables are changed, or compositions and other important variables that control the processes being investigated are changed. It is still common to choose a membrane for a given separation by essentially a trial-and-error strategy. Thus, based on previous experience for a proposed separation, one may identify 10–20 potential membranes that appear promising. Well-designed MD simulations can perform a round of screening studies to eliminate membranes that will have a high probability of failure. Our experience has shown that MD simulation methods are very effective in identifying membranes that are unlikely to be suitable. This is because simulations generally assume ideal conditions (e.g., a defect-free membrane or a membrane where defects are introduced in the simulation but may not correctly represent the actual defects in a membrane). Thus, if even under these circumstances a membrane is unable to show effective separation for a given system, it is extremely unlikely it will be successful in an actual industrial setting. On the other hand, since ideal conditions are generally used in simulations, positive outcomes are not always as reliable, for the reasons mentioned above (Wang et al., 2022). Computer simulations have been successful in studying processes as diverse as water desalination, ion separations, ion exchange’ thus we believe they can play an important role in designing new processes for extracting critical metals (Wang et al., 2022).

While computer simulation have not been previously extensively used for the removal of critical metals they have been used for a wide range of separations which makes them a very promising tool for extracting critical metals too. Most of these studies have used methods such as molecular dynamics with intermolecular models that include both intramolecular interactions (bond length, bond angles, dihederal angles), and intermolecular contributions that include dispersion and repulsive forces as well as coulombic contributions. Various software packages such as LAMMPS (Thompson et al., 2022), NAMD (Phillips et al., 2020) have been developed that allow these simulations to be carried out as well as potential models such as AMBER (He et al., 2020). These and other packages have been used for extensive studies for separation, many using zeolites on desalination (Lin and Murad, 2001; Rassoulinejad-Mousavi et al., 2020), ion-exchange (Murad et al., 2004; Khosravi et al., 2019)^; adsorption, diffusion and shape selectivity using computer simulation has been reported in a recent review (Smit and Maesen, 2008); Xe recovery from nuclear power plants off streams. While simulations of extracting metals from very dilute solutions have not been done extensively in the past. We reported recently extracting Li from a very dilute solutions using molecular dynamics scheme we recently developed (Bashmmakh et al., 2025). It is also possible now to determine the pore sizes of membranes and their structure that will facilitate separation of metal ions (Hinkle et al., 2016).

7 Discussion

Aqua-mining of critical metals from waste streams is an exciting new technology, which is becoming technologically feasible and preliminary studies indicate has the potential to be economically feasible as well. It uses many processes that have been extensively used in the chemical industry for other applications. These include reverse osmosis, ion exchange, ion-separations and adsorption. It shows promise to significantly reduce environmental and health hazards associated with currently used mining methods.

8 Conclusion

In this perspective paper, we have examined the potential of aqua-mining for removing critical metals from waste streams rather than the traditional mining techniques currently being used. We have described the possible aqueous streams both found naturally and as industrial wastes, which offer the potential for aqua-mining. We have pointed out the potential of zeolites to play a prominent role in such extraction of critical metals and minerals. Zeolites show promise for concentrating dilute solutions using reverse osmosis, followed by adsorption and also if needed ion-exchange. A combination of these processes can lead to economically viable industrial production. Finally we point out that with the availability of high-performance computing at reasonable cost, it is possible to take advantage of methods such as molecular dynamics to do initial screening studies for determining the potential of any proposed strategies. While these computational methods sometimes rely on idealized models and can sometimes give false positive results, they seldom provide false negative results so that can be relied on, to a large extent, to eliminate failing strategies. This is perhaps comparable to doing a second law analysis of a proposed process to eliminate those that are not likely to be successful.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Carina L. Gargalo, Technical University of Denmark, Denmark

Reviewed by: Evgeniya Dmitrieva, Institute of Petrochemical Synthesis. AV Topchieva (RAS), Russia