Copper Separation and Recovery from Solutions: Conventional and Membrane-Based Methods

Abstract

Cu is central to electrification and low-carbon technologies, while declining ore grades and environmental pressures increase interest in recovery from secondary liquid resources. This review critically evaluates copper separation from mine-affected waters, copper-smelting waste acids, electroplating wastewaters, and secondary-material leachates. Cu is the principal target; co-occurring metals are treated as competitors or sequential recovery targets. Feed origin, Cu concentration and speciation, pH/free acidity, ionic strength, ligands, and polymetallic composition are related to process performance and product form. Conventional precipitation, solvent extraction, ion exchange, and electrowinning are compared with microfiltration, ultrafiltration, nanofiltration, reverse osmosis, electrodialysis, supported liquid membranes, emulsion liquid membranes, and polymer inclusion membranes. High Cu removal or rejection does not by itself demonstrate Cu-selective recovery. MF and UF require conversion of dissolved Cu into retainable species; NF and RO mainly recover water and preconcentrate metals; ED becomes more selective with speciation control; and carrier-mediated membranes can fractionate metals but face stability constraints. Most membrane processes generate a Cu-rich retentate, concentrate, or stripping solution rather than metallic Cu and therefore require downstream crystallisation or electrowinning. Key gaps are long-term operation with variable industrial feeds, fouling and scaling control, mass balances and product-purity reporting, membrane/carrier durability, and consistent techno-economic validation.

1. Introduction

By volume of consumption, Cu ranks third among industrial metals after Fe and Al and is central to electrical infrastructure and low-carbon energy technologies [1,2]. Its widespread industrial use is largely attributed to its high electrical and thermal conductivity, good processability, and corrosion resistance [3]. Cu and Cu-based materials are widely used in electrical and heat-transfer systems and construction and are increasingly processed using additive manufacturing [2], and catalytic systems for liquid-propellant decomposition [4,5]. Figure 1 presents the global distribution of Cu use by end-use sector in 2024. Building construction accounted for the largest share (26%), followed by consumer and general products, cooling and electronics (23%) and infrastructure (17%). Together, these three sectors represented 66% of global Cu use. This distribution shows that Cu demand is supported by a broad range of established applications, while the expansion of electricity networks and electrified transport is expected to place additional pressure on supply [6].

The substantial use of Cu in electrical and transport systems makes future consumption particularly sensitive to the pace of electrification. Under the SSP2–RCP2.6 scenario, final global Cu demand is projected to increase from 24.3 Mt in 2015 to 44.4 Mt in 2050 [7]. A separate analysis based on a pathway consistent with limiting global warming to 1.5 °C projected demand of 62 Mt in 2050, approximately 2.5 times the 2015 level [8]. Figure 2a compares these estimates, which should be interpreted as scenario-dependent results rather than as a single forecast. An econometric projection extending to 2030 also found that renewable-energy deployment and electric-vehicle adoption are reshaping Cu consumption and suggested that conventional models may not fully capture the contribution of clean-energy technologies to future demand [9]. Although the numerical estimates differ, all three studies anticipate continued growth in global Cu demand, strengthening the case for diversifying supply beyond primary ores.

How this additional demand is met is equally important. A greater reliance on primary production would increase the resource use and greenhouse-gas emissions associated with the copper cycle [8]. Recycling can reduce this burden, but the amount of copper available in end-of-life (EoL) products remains limited. It is estimated that end-of-life secondary copper currently covers about 23% of demand. Even with highly efficient collection and recycling, its share would not exceed 49.6% in 2050. The current and projected shares of EoL secondary Cu are shown in Figure 2b. The red dashed line marks 50% of total Cu demand, showing that even the most optimistic estimate for 2050 remains slightly below half of the projected demand. Conventional scrap recycling therefore cannot supply all the copper required over the coming decades, and other secondary sources need to be considered [10].

Cu-containing secondary liquid streams represent one such additional source of copper [11,12,13,14]. In these streams, copper is already present in dissolved form and can be separated or concentrated directly at the recovery stage. Their practical value has been demonstrated using different treatment routes. A combined nanofiltration–solvent extraction process applied to real acid mine drainage recovered 80% of the water and 97% of the Cu [11]. In a post-galvanic wastewater containing Fe(III), complexation-enhanced electrodialysis transferred 92.2% of Cu(II) into the concentrate [12]. Electrolysis of this concentrate subsequently recovered 94.49% of its copper and produced a material with a purity of 99.4%. Other studies demonstrated the conversion of acidic PCB etching wastewater into value-added Cu salts [13] and the joint recovery of Cu and other base metals from waste-PCB leachate [14].

Taken together, these studies show that membrane processes can perform different functions within Cu-recovery flowsheets, including water recovery, Cu concentration, separation from competing ions and production of a Cu-enriched stream suitable for downstream solvent extraction or electrowinning. Individual case studies, however, do not reveal which membrane processes, application contexts and research themes recur across the retrieved literature or how these topics are associated. An author-keyword co-occurrence analysis was therefore conducted to identify the principal thematic relationships within this membrane-focused dataset.

The dataset comprised 148 English-language journal articles published between 2016 and 2025 and indexed in Scopus. The search was limited to the Scopus subject areas of Environmental Science, Chemical Engineering, Chemistry and Materials Science. Author-keyword co-occurrence analysis was performed using VOSviewer version 1.6.21, following the mapping approach described in [15]. Of the 148 records, 142 contained author keywords, yielding 509 distinct terms. A minimum occurrence threshold of three was applied; 30 author keywords met this criterion and were all included in the maps. Co-occurrence data were normalised using association strength, and shorter distances between keywords generally indicate stronger relatedness. Some labels are not displayed to avoid overlap, but the corresponding nodes remain included. The exact search string and search date are reported in the Supplementary Materials, and the complete list of records is provided in Supplementary Table S1.

In the overlay visualisation (Figure 3a), node size represents the number of keyword occurrences, link thickness indicates co-occurrence strength and node colour represents the average publication year of the articles in which each keyword occurred. Although the dataset covers 2016–2025, the colour scale is based on keyword-level averages rather than individual publication years: blue indicates an average year of 2019 or earlier, green indicates intermediate values and yellow indicates 2022 or later. Nanofiltration, copper and electrodialysis are the three most frequently occurring author keywords, with 27, 24 and 23 occurrences, respectively. Nanofiltration co-occurs with acid mine drainage, solvent extraction, Cu recovery and resource recovery. Electrodialysis is linked with Cu, plating wastewater and heavy metal removal. Cu co-occurs with Ni in six records and with Zn in three records, indicating that some of the retrieved studies addressed multi-metal systems. Among the visible terms, forward osmosis, heavy metal removal and resource recovery have comparatively later average publication years, whereas bulk liquid membranes, acid mine drainage and NF270 are associated with earlier publications within the analysed period.

In the cluster density visualisation (Figure 3b), colours do not represent publication year. Instead, each colour indicates the local weighted density of author keywords assigned to a particular cluster, while blended colours indicate spatial overlap between neighbouring clusters. Three process-oriented regions are visually dominant. One is centred on electrodialysis and includes plating wastewater and heavy metal removal. A second is centred on nanofiltration and includes acid mine drainage, NF270, solvent extraction, copper recovery and resource recovery. A third includes supported liquid membranes, composite membranes, heavy metals and wastewater treatment. Forward osmosis and bulk liquid membranes form smaller peripheral themes. In the shared map layout, copper occupies a central position and is linked with terms representing different membrane processes and applications.

Because the search query required membrane-related terms, studies dealing exclusively with conventional Cu-removal or recovery methods were not represented in the dataset. Figure 3 was therefore not designed to compare membrane and conventional methods and cannot demonstrate a shift from, or replacement of, conventional processes by membrane technologies. It describes only the thematic structure and chronology of the selected membrane-focused literature and does not provide direct evidence of process efficiency, selectivity or industrial feasibility. The results are also affected by the selected database, publication language, document type, subject-area restrictions and the author keywords assigned to individual records.

Terms describing feed chemistry, Cu speciation, transport mechanisms and the form of the recovered Cu product are not prominent in the maps. No Fe- or Al-related author keyword reached the minimum occurrence threshold, despite the importance of these elements as matrix components and competing ions in acid mine drainage and hydrometallurgical leachates. Their absence from the maps does not mean that Fe and Al were absent from the underlying studies. It indicates that their competition with Cu was not prominent in the author-keyword vocabulary of the retrieved records and must therefore be assessed through study-level analysis.

Accordingly, the remainder of this review examines how the origin, composition and Cu speciation of a solution affect membrane selection and performance. Microfiltration, ultrafiltration, nanofiltration, reverse osmosis, electrodialysis and liquid or carrier-assisted membranes are compared in terms of their transport mechanisms, Cu selectivity, recovery, flux, fouling, stability, energy and reagent demand, scale-up potential and integration with downstream metal-recovery processes. Cu is treated as the principal target, while co-occurring valuable metals are considered where they compete with Cu or can be recovered sequentially.

2. Copper-Containing Solutions as Secondary Liquid Resources

The bibliometric maps identify mine drainage, plating wastewater and resource recovery as recurrent contexts for membrane-based Cu separation, but they do not show whether the corresponding liquid streams are chemically comparable. Before comparing membrane technologies, it is therefore necessary to identify the principal types of Cu-containing secondary liquid streams and the chemical characteristics that affect membrane separation. In this review, a secondary liquid resource is defined as a waste or process liquor, or as a leachate generated from a secondary solid feed. Based on their industrial origin and mode of generation, the Cu-containing secondary liquid streams considered in this review were grouped into four representative categories: (i) mine-affected waters, which arise through drainage and water–rock interactions at active or abandoned mine sites and include acid mine drainage and copper mine drainage [11,16]; (ii) copper-smelting waste acids, which are acidic process liquors generated during smelting and associated flue-gas-cleaning operations [17]; (iii) electroplating wastewaters, which are generated by plating and rinsing operations [18,19]; and (iv) secondary-material leachates, which are intentionally produced by chemical or biological leaching of solid wastes, particularly waste printed circuit boards [20,21]. These streams differ in acidity, ionic strength, dominant anions and ligands, Cu speciation and competing-metal content; consequently, total Cu concentration alone is not sufficient to determine their suitability for membrane treatment.

Mine-affected waters exhibit particularly wide compositional variability. A real acid mine drainage feed used in a nanofiltration study contained 531.2 mg/L Cu, 382.6 mg/L Al and 4671 mg/L SO₄ [11]. By contrast, copper mine drainage from the Dexing mine contained only 38.24–47.86 mg/L Cu at pH 2.32–3.12, whereas total Fe and Al reached 427.72–589.15 and 486–580 mg/L, respectively; SO₄ concentrations were 9.11–12.31 g/L [16]. An extreme acid mine drainage from the São Domingos mine had a pH of 1.19 and contained 5.25 g/L Cu, 63.2 g/L Fe, 6.47 g/L Al, 1.96 g/L Zn and approximately 142 g/L SO₄ [22].

In the Dexing drainage, the Fe concentration was approximately 11–12 times that of Cu, whereas the Al concentration was approximately 12 times that of Cu. In the São Domingos AMD, the Fe concentration was approximately 12 times that of Cu. These differences create a practical separation challenge: nanofiltration may retain Cu together with other divalent ions, Al precipitation may require pH control during continuous membrane operation, and Fe(III) may interfere with downstream Cu recovery unless it is removed or selectively complexed [12,23].

Cu-smelting waste acids represent a more concentrated and chemically aggressive class of liquid resource. One industrial stream contained approximately 23% H₂SO₄ together with 2.015 g/L Cu, 2.527 g/L As, 0.668 g/L Zn, 0.157 g/L Cd and 0.300 g/L F [17]. Because the As concentration exceeded that of Cu, treatment of this solution required selective Cu–As separation rather than simple Cu concentration. The high free-acid content and the presence of hazardous and corrosive species must also be considered when assessing membrane stability and product purity.

Electroplating wastewaters vary according to the plating bath, rinsing procedure and complexing agents used. A real jewellery-industry wastewater contained 430 mg/L Cu and 8.34 g/L total cyanide at pH 1.74, whereas Ni and Ag concentrations were below 1 mg/L [18]. In this case, the high ligand concentration is important because total Cu cannot be equated with free Cu²⁺. Another electroplating effluent contained 417.12 mg/L Cu together with 553.97 mg/L Na and 672.91 mg/L Ca [19], whereas a highly concentrated acidic plating wastewater contained 5.33 g/L Cu [24]. Membrane processes may therefore encounter either relatively dilute and strongly complexed Cu or concentrated, highly conductive solutions.

Leachates from waste printed circuit boards (WPCBs) form a fourth important group, but their composition is determined by both the original electronic waste and the selected leaching chemistry. Bioleaching at an initial pH of 1 using 20 g/L Fe³⁺ as an oxidising agent extracted 95% of Cu and 87% of Ni, producing a multicomponent solution containing Cu, Ni, Zn and a high Fe concentration [20]. Mineral-acid leaching can produce substantially more concentrated solutions: treatment with 4 mol/L HNO₃ generated a leachate containing 15.675 g/L Cu together with Ni, Zn, Pb, Al and Fe [21]. The composition may also change during operation. Repeated reuse of an H₂SO₄-H₂O₂ leachate progressively accumulated Al and Fe, indicating that a solution characterised during the first cycle may not represent the feed obtained after extended closed-loop operation [25]. For this reason, the composition of the actual leachate, rather than only that of the original solid WPCB feed, should be reported when membrane separation is evaluated.

In the representative streams considered above, Cu concentrations range from approximately 40 mg/L to more than 15 g/L. This range alone, however, does not define recovery potential. At a minimum, the feed should be characterised in terms of its industrial origin, pH or free-acid concentration, ionic strength, dominant anions and complexing agents, Cu speciation, Fe/Cu and Al/Cu ratios, other recoverable or hazardous metals, and the presence of suspended or organic matter. These parameters influence membrane selectivity, osmotic load, scaling and fouling behaviour, chemical stability and the attainable Cu product. The following sections therefore evaluate membrane technologies in relation to the complete feed matrix and the required recovery product rather than to total Cu concentration alone.

3. Conventional Methods and Their Limitations

Chemical precipitation, solvent extraction, ion exchange and electrowinning are established non-membrane operations for Cu separation and recovery from aqueous streams and are therefore considered conventional methods in hydrometallurgy and industrial wastewater treatment [26,27]. The pathways through which Cu is transferred from the aqueous feed during these operations are illustrated in Figure 4.

The Cu-bearing phases shown in Figure 4 do not represent equivalent recovery products. Precipitation removes Cu as a solid that may require filtration, leaching or smelting, whereas electrowinning can produce a metallic deposit directly. By contrast, the Cu-loaded organic phase and Cu-loaded resin are intermediate products: Cu must subsequently be recovered through stripping or elution, respectively, before crystallisation, electrowinning or reuse of the concentrated solution. The following subsections therefore compare the four methods not only in terms of Cu separation but also according to the feed conditions controlling selectivity, the form and purity of the recovered product, and the additional treatment required.

However, they are not interchangeable process routes. Precipitation converts dissolved Cu into a separable Cu-bearing solid, solvent extraction transfers Cu to an organic phase followed by stripping, ion exchange retains Cu on a resin followed by elution, and electrowinning deposits dissolved Cu as metallic copper at the cathode [28,29,30,31]. The following subsections examine how Cu concentration, acidity, speciation, and competing metals affect the performance and selectivity of each method for the four stream categories defined in Section 2, as well as the form and purity of the recovered Cu and the additional treatment required.

3.1. Chemical Precipitation

Chemical precipitation removes dissolved Cu by adding a reagent that converts it into an insoluble solid. Increasing pH can incorporate Cu into hydroxide precipitates, whereas sulfide reagents form CuS even under acidic conditions. Selective ligands can also bind Cu and form an insoluble Cu–ligand compound. The resulting solid is separated by settling or filtration. Thus, this method does not produce metallic Cu directly; the type and purity of the product depend on Cu concentration, acidity, the chemical form of Cu and the metals present in the same solution [32,33,34].

3.2. Solvent Extraction

Solvent extraction separates Cu using two liquids that do not mix. The Cu-containing aqueous solution is mixed with an organic phase containing an extractant that binds Cu. The phases are then allowed to separate, leaving Cu in the organic phase and the non-extracted components in the aqueous solution. The Cu-loaded organic phase is subsequently contacted with an acidic solution, which releases Cu into a smaller aqueous volume and allows the organic phase to be reused. Lower feed acidity generally favours Cu transfer to the organic phase, whereas strong acid is used to transfer Cu back during stripping [29,35]. Solvent extraction therefore produces a concentrated Cu solution rather than metallic copper, and this solution must be further processed or returned to an industrial process [29].

3.3. Ion Exchange

Ion exchange removes dissolved copper by passing the solution through a bed of polymer beads. Chemical groups on the beads bind Cu ions, while the treated solution leaves the column. After the resin becomes loaded, copper is washed out with an acid, base or salt solution. The immediate product is therefore a Cu-rich solution rather than solid or metallic copper. Copper concentration, solution acidity, its chemical form and competing metals determine how much Cu is retained and how pure the resulting solution will be [36].

3.4. Electrowinning

Electrowinning recovers copper by passing direct current through a Cu-containing solution. Cu²⁺ ions gain electrons and form metallic Cu on the cathode, while water is oxidised and O₂ is released at the anode in conventional sulfate electrolytes [37]. The fraction of the applied current that produces Cu is termed current efficiency. This efficiency decreases when Cu²⁺ is not transported to the cathode quickly enough, because side reactions consume a larger share of the current. In experiments with a PCB-derived leachate, H₂ evolution became favourable when the Cu concentration fell to approximately 2.69 g L⁻¹; this was a system-specific limiting concentration rather than a universal threshold [31].

The applicability of chemical precipitation, solvent extraction, ion exchange and electrowinning was assessed for the four solution categories defined in Section 2.

Table 1. Effects of solution properties on Cu recovery by conventional methods.

Method	Stream Category	Effect of Cu Concentration on Recovery and Selectivity	Effect of pH/Free Acidity and Cu Form	Effect of Competing Metal	Recovered Cu, Purity and Further Treatment	Ref.
Chemical precipitation	Mine-affected waters	Higher Cu produces more CuS solid and makes solid separation easier. At low Cu, removal may remain high, but collecting fine solids and reagent use per unit Cu become less favourable.	At low pH, sulfide converts dissolved Cu2+ to CuS before Zn precipitates. Raising pH promotes Fe/Al and later Zn hydroxide formation, decreasing selectivity.	Zn can remain dissolved with controlled sulfide addition. Fe/Al hydroxides, clay and silica lower solid purity.	CuS or CuS/ZnS concentrate. Nearly pure CuS was obtained under controlled conditions, whereas a field solid contained only about 80% combined Cu and Zn metals. Filtration followed by leaching or smelting is required.	[28,33,38]
	Cu-smelting waste acids	A high Cu/As ratio favours CuS formation before As precipitation. When Cu is low relative to As, mixing and reaction time have a greater influence and Cu/As separation becomes poorer.	Sulfide precipitation remains possible in strongly acidic solution. However, excess sulfide or a long reaction time increases As precipitation.	As is the main selectivity problem. Fe can enter a Cu–Fe sulfide phase and reduce the Cu content of the solid.	CuS mixed with Cu–Fe sulfide. One study reported 20.2% Cu and 0.7% As. The solid must be filtered, checked for As and subsequently leached or smelted.	[17,39]
	Electroplating wastewaters	Recovery is limited by precipitant capacity. Increasing the Cu or wastewater load without increasing the reagent leaves more Cu in solution.	Raising pH improves precipitation of free Cu2+, with little additional benefit above approximately pH 10. Stable Cu–CN complexes do not precipitate without cyanide destruction.	Ni, Zn, Cd, Fe and Cr can enter the mixed precipitate. Added Mg and Al also dilute the Cu product.	Mixed Cu–Mg–Al hydroxide rather than a pure Cu compound. Filtration and further recovery or disposal are required. Cyanide-containing streams first require oxidation and neutralisation.	[32,40]
	Secondary-material leachates	Recovery is controlled mainly by the ligand-to-Cu ratio. Dilute solutions can still give high recovery, but require more ligand and solution handling per unit Cu.	A selective ligand can precipitate Cu2+ at low pH, where hydroxide precipitation is ineffective. Excess residual acid suppresses ligand precipitation, whereas strong acid releases Cu from the solid.	The tested ligand largely rejected Fe and Sn; competition from other metals depends on ligand selectivity.	Cu–ligand solid, not Cu metal. Fe and Sn accounted for only 0.6% and 0.9% of recovered metals. Acid stripping produces a Cu-rich solution for crystallisation or electrowinning.	[34]
Solvent extraction	Mine-affected waters	Higher Cu loads more Cu into the organic phase and allows production of a concentrated stripping solution. Dilute waters require prior concentration, repeated loading or additional extraction stages.	Hydroxyoxime extraction of Cu2+ improves as free acidity decreases. Very low pH reduces loading, although a high extractant concentration can partly compensate.	Even limited Fe extraction can cause substantial contamination when Fe greatly exceeds Cu. Selective stripping can leave most Fe in the organic phase.	Cu-rich acidic solution, not metal. Concentrations suitable for electrowinning were obtained after repeated loading; final Cu-metal purity was not measured.	[11,22]
	Cu-smelting waste acids	Increasing Cu improves organic-phase loading only until the extractant approaches saturation. In an acidic sulfate model, Cu distribution initially increased with Cu concentration but decreased at higher loading. The effect has not been confirmed using real Cu–As smelting waste acid.	Increasing H2SO4 concentration decreases Cu extraction because H+ opposes the extraction of Cu2+. Mixed extractants can recover Cu at very low pH, whereas conventional hydroxyoximes would require partial acid removal. High chloride also decreases Cu extraction by changing Cu speciation.	Fe can be co-extracted, while high chloride decreases Cu/Fe separation. As was absent from these model studies; therefore, Cu/As selectivity in real smelting waste acid remains unconfirmed.	A Cu-loaded organic phase followed by a Cu-rich acidic stripping solution. Cu purity and As rejection have not been demonstrated for real smelting waste acid. Acid adjustment and As removal would be required before extraction, followed by electrowinning or crystallisation.	[41,42]
	Electroplating wastewaters	Cu-rich sulfate rinses can be concentrated effectively. Dilute rinses require more stages. In cyanide solutions, a higher CN/Cu ratio decreases Cu loading.	Free Cu2+ is extracted after acidity adjustment. Cu–CN anions require an amine- or guanidine-based extractant rather than a conventional hydroxyoxime.	Low Fe and Zn give high selectivity. Cu and Ni can be separated by pH, but Ni and Zn cyanide complexes may be co-extracted.	Concentrated CuSO4 solution or CuSO4·3H2O crystals, potentially reusable in the plating bath. Organic contamination must be avoided. A Cu–CN concentrate requires further recovery and cyanide treatment.	[29,43,44]
	Secondary-material leachates	The studied WPCB leachates contained Cu at g/L levels and gave more than 90% or complete extraction. Initial Cu concentration was not varied independently; recovery was mainly limited by extractant capacity and the number of extraction stages.	Lower free acidity favours Cu2+ extraction. A sulfate leachate was treated at pH 1.1, whereas a leachate containing 6 M HNO3 required adjustment to an equilibrium pH of 2.2 before complete Cu extraction.	Fe was the main competing metal and reduced the Cu-loading capacity of the organic phase. Control of pH and extractant concentration improved Cu separation from the other metals.	Acidic Cu-rich stripping solution. One study obtained about 90% stripping but did not measure final product purity. In another study, subsequent electrowinning produced 99.8% pure Cu metal.	[45,46]
Ion exchange	Mine-affected waters	Ion exchange can recover dilute Cu by accumulating it on the resin. Higher Cu loads the resin faster but shortens the operating cycle.	Very low pH increases competition from H+. Raising pH sufficiently for Cu2+ binding improves uptake, but excessive pH adjustment can also bind other metals.	Fe, Al and Zn occupy resin capacity or reduce selectivity. Their prior removal and staged pH adjustment improve Cu separation.	Cu-rich sulfuric solution; a 40-fold concentration was reported. Final product purity was not measured. Electrowinning, precipitation or crystallisation is required.	[47]
	Cu-smelting waste acids	Higher Cu produces a more concentrated resin washing solution but also saturates the resin sooner. Resin capacity and regeneration determine overall recovery.	High H+ concentration suppresses many conventional resins. Special chelating resins can retain Cu2+ in strongly acidic sulfate solutions	Fe and Ni may compete for binding sites. Selected resins can reject much of the Fe, but Cu/As selectivity has not been tested in real smelting waste acid.	Acidic or ammoniacal Cu-rich solution rather than metal. As content and final purity are unknown; solution conditioning followed by electrowinning or crystallisation is required.	[48]
	Electroplating wastewaters	Dilute Cu can be concentrated until the resin is saturated. Higher Cu increases loading but shortens the treatment cycle.	Free Cu2+ can be retained by a cation-exchange resin. Cu–CN complexes are not recovered by the same route and require cyanide oxidation and neutralisation.	Ni, Zn, Fe and Cr may compete depending on pH and ionic form. Different Cr oxidation states require separate treatment stages.	Acidic Cu solution after resin washing; product purity was not reported. Further electrowinning or crystallisation is required.	[40]
	Secondary-material leachates	More resin increases Cu recovery, whereas high Cu loading or high flow causes earlier resin saturation.	Very low pH can favour selective Cu2+ binding on a suitable resin. Increasing pH promotes uptake of other metals and decreases Cu selectivity.	Fe, Ni, Zn, Mn and Al increasingly compete as pH rises, reducing the purity of the recovered Cu solution.	Cu-rich sulfate solution after acid washing of the resin. Purity was not reported; electrowinning or crystallisation and resin regeneration are required.	[30]
Electrowinning	Mine-affected waters	Higher Cu improves Cu transport to the cathode and current efficiency. As Cu becomes depleted, hydrogen formation and energy use increase.	Free Cu2+ is required. Moderate acidity improves conductivity, but excessive H+ competes with Cu deposition through hydrogen formation.	Fe3+/Fe2+ reactions consume current and can re-dissolve deposited Cu. Low Fe therefore allows more effective direct electrowinning.	Metallic Cu deposit; purity was not reported in the mine-water study. The deposit requires washing, while the depleted water still requires treatment or recycling.	[49]
	Cu-smelting waste acids	Cu concentration controls both deposition efficiency and As contamination. In an acidic Cu–As model, Cu concentrations of 10 g/L or lower favoured copper–arsenide formation, whereas 40 g/L Cu produced grade-A copper in the presence of 2 g/L As. This indicates that the dilute Cu–As waste acid requires Cu concentration or As removal before electrowinning.	Strong acidity keeps Cu2+ dissolved and provides conductivity, but a high H+/Cu ratio promotes hydrogen formation and lowers current efficiency. In a related highly acidic refinery effluent, low Cu combined with very high H2SO4 resulted in low current efficiency at high current density.	As can co-deposit as copper arsenides when Cu concentration is low. Bi also co-deposited in a related acidic effluent. A real smelting-effluent study removed Cu, Cd, Ni, Pb, Zn and Fe together, indicating limited Cu selectivity.	A real smelting-effluent study achieved more than 99% total heavy-metal removal but did not report a separate Cu product or its purity. Related acidic systems produced mixed Cu/Cu2O–Bi or Cu–As deposits. Cu concentration and impurity removal are required before electrowinning, and the deposit may require refining.	[50,51,52]
	Electroplating wastewaters	High Cu favours efficient deposition. Dilute rinses may require electrodialysis or another concentration step; energy use rises as Cu concentration falls.	Free Cu2+ deposits under acidic conditions. Cu–CN requires strongly alkaline operation with simultaneous cyanide oxidation.	Fe consumes current, while chloride can promote re-dissolution of deposited Cu. These effects can greatly reduce recovery from real wastewater.	Metallic Cu or a Cu/Cu2O product. Purity was generally not reported; washing or refining and final cyanide treatment may be required.	[53,54,55]
	Secondary-material leachates	Higher Cu gives high current efficiency. When Cu falls below a few g/L, hydrogen formation and local precipitation become more important, so the remaining solution may require reconcentration.	Adjusted acidic sulfate solutions containing free Cu2+ are favourable. A local pH increase near the cathode can form Cu hydroxide or oxide instead of metal.	Fe and Al should be removed before electrowinning. Remaining Fe, Zn, Ni and Al decrease cathode purity.	Metallic deposit containing 95.97 wt% Cu in one WPCB study. Washing or refining is required, and the remaining solution should be recycled or reconcentrated.	[31]

shows that no conventional method is equally suitable for all four stream categories. Precipitation can treat acidic solutions when sulfide reagents or selective ligands are used, but co-precipitation often lowers product purity and creates a solid requiring filtration and further treatment. Solvent extraction and ion exchange can selectively concentrate Cu, although high free acidity, Cu complexation and competing metals may reduce their performance; both methods also require stripping or elution followed by final Cu recovery. Electrowinning produces metallic Cu directly, but is most effective for sufficiently concentrated solutions containing mainly free Cu²⁺ and low concentrations of electroactive impurities. Thus, the main limitation of conventional methods is the difficulty of combining selective Cu separation, concentration and production of a reusable product in a single process.

These limitations provide the basis for considering membrane processes as separation or concentration steps. Depending on the driving force and membrane properties, they can separate dissolved species through size exclusion, charge effects or selective transport and can be combined with conventional recovery operations. The following sections evaluate membrane methods using the same four stream categories and solution factors applied in Table 1, together with Cu recovery, selectivity, product concentration and the need for pre- or post-treatment.

4. Classification of Membrane-Based Methods for Cu Extraction

Membrane technologies have become one of the most promising approaches for metal recovery because they enable selective separation while requiring lower energy consumption than many conventional processes. The membrane allows some components to pass through whilst retaining others, thereby ensuring the effective separation of mixtures. Separation is achieved under the influence of various driving forces, such as a pressure difference, a concentration gradient or an electric field. Compared with traditional separation methods, membrane technologies are characterised by lower energy consumption and high environmental efficiency, which ensures their widespread use in the purification and processing of solutions [56]. The most common membrane methods are microfiltration (MF), ultrafiltration (UF), nanofiltration (NF), reverse osmosis (RO) and electrodialysis (ED) [57].

The main differences between pressure-driven membrane processes lie in their membrane pore size and molecular weight cut-off (MWCO). Microfiltration (MF) is characterised by a pore size of 0.2–1 µm. Ultrafiltration (UF) has a molecular weight cut-off of 1000–1,000,000 Da and a pore size of 5–0.2 µm. Nanofiltration (NF) has a molecular weight cut-off of 100–1000 Da and a pore size of 0.5–5 nm, whereas reverse osmosis (RO) has a molecular weight cut-off of less than 100 Da and a pore size of 0.2–0.3 nm [58].

4.1. Microfiltration

MF is a pressure-driven membrane process mainly used to separate suspended particles, colloids, and precipitates. Because hydrated Cu(II) ions are much smaller than conventional MF pores, dissolved copper cannot be effectively removed by steric sieving alone. Therefore, MF is more suitable for copper recovery when Cu(II) is first converted into a membrane-retainable form through chemical precipitation, polymer complexation, or electrochemical reactions [59,60,61].

In precipitation-assisted MF, dissolved Cu(II) is converted into sparingly soluble Cu(OH)₂ or CuS particles, shifting the separation mechanism from ionic rejection to particle retention. A system combining sulfide precipitation with a 0.14 μm tubular ceramic MF membrane achieved Cu recovery close to 100% at approximately 120% of the stoichiometric sulfide dose, while maintaining permeate turbidity below 2 NTU and a flux above 0.1 L/m²s. However, high Cu removal does not necessarily indicate high Cu selectivity. In multimetal solutions, Fe, Al, Zn, Ni, and Co may also precipitate or co-precipitate, reducing the purity of the recovered copper-rich solid. Thus, the selectivity of this process is largely governed by the precipitation chemistry rather than by the MF membrane itself [59].

In addition, polymer-enhanced microfiltration (PEMF) can improve metal-ion retention by binding dissolved ions to water-soluble polymers such as poly(acrylic acid) (PAA) and polyethylene glycol (PEG), thereby forming larger metal-polymer complexes that can be retained by the membrane. This approach has been reported for the removal of Cu(II), Ni(II), and Cr(VI), with rejection efficiencies exceeding 80–90% under optimised operating conditions (Figure 5) [61].

Conductive MF provides an alternative mechanism by using the membrane simultaneously as a porous separator and an electrode. Cu(II) can be electrochemically reduced to metallic Cu(0), while local pH changes promote Cu(OH)₂ precipitation. In one system, surface analysis confirmed both Cu(0) and Cu(OH)₂, with approximately 47.3 ± 8.5% of Cu removal attributed to metallic deposition and 41.1 ± 0.2% to chemical precipitation. However, accumulation of Cu deposits and precipitates may increase hydraulic resistance and contribute to membrane pore blockage and electrode passivation [60].

More generally, membrane fouling remains an important limitation of MF-based Cu recovery. In precipitation-assisted systems, particle deposition may result in pore blocking or cake formation on the membrane surface, leading to a progressive decline in permeate flux. Fouling can be mitigated through appropriate selection of membrane pore size and material and by optimisation of operating parameters such as transmembrane pressure and flow rate [59,60].

4.2. Ultrafiltration

UF is mainly used to separate colloids and macromolecules, while dissolved Cu(II) ions are generally too small to be retained by conventional UF membranes. Cu(II) separation therefore requires its conversion into larger membrane-retainable species through polymer complexation, surfactant micelles, or membrane functionalization. In polymer-enhanced ultrafiltration (PEUF), water-soluble polymers bind Cu(II) and form larger complexes. A recent electric-field-enhanced PEUF system maintained Cu removal of 92–98% under acidic and high-salinity conditions and about 85% removal even when competing metals were present at concentrations up to 50 times higher than Cu. In PEUF, membrane fouling is associated with the accumulation of metal-polymer complexes on the membrane surface, which contributes to flux decline. Applying a +0.2 V electric field reduced the flux decline rate by 58%, while real industrial wastewater tests achieved approximately 95% Cu removal and 80% Cu recovery by subsequent electrolysis [62].

Micellar-enhanced ultrafiltration (MEUF) uses surfactant micelles to bind Cu(II) and retain the resulting aggregates by UF. With a 5 kDa ceramic membrane, Cu retention reached 76.46% in 0.1 M HCl but decreased to 53.38% with a 15 kDa membrane, showing the strong influence of membrane MWCO and solution acidity [63]. In another MEUF system, rejection exceeded 96% for Cu(II), Cd(II), Co(II), Ni(II), and Zn(II) at an SDBS:metal molar ratio of 40:1. However, similar rejection for all metals indicates multimetal removal rather than Cu-specific selectivity. Surfactant leakage and concentration polarisation may also limit practical use [64].

Surface functionalization offers another route to improve Cu(II) retention by introducing binding sites directly onto the UF membrane. Polyethyleneimine (PEI)-modified membranes contain amine groups capable of binding Cu(II) through adsorption and coordination. As shown in Figure 6, Cu(II) ions are retained on the functionalized membrane surface, while water and smaller dissolved species pass into the permeate. PEI modification may also reduce the effective pore size and thereby enhance copper retention [65].

A recent UiO-66-keratin-functionalized PES/PEG200 UF membrane achieved Cu rejection of 78% for a synthetic solution and 76% for real electroplating wastewater, compared with about 34% for pristine PES, while also showing improved flux recovery [66].

Overall, PEUF, MEUF, and functionalized UF can improve Cu(II) retention, but high removal does not necessarily indicate Cu-specific selectivity. Competing ions, fouling, regeneration, and polymer or surfactant losses remain important for practical application.

4.3. Nanofiltration

Cu(II) rejection by NF is governed not only by size exclusion but also by Donnan and dielectric exclusion. Separation therefore depends on membrane pore size and charge, ion hydration, and feed composition [67]. A schematic representation of the NF separation mechanism based on steric and electrostatic effects is shown in Figure 7 [68]. In acidic CuSO₄ solutions, a low-pressure NF membrane maintained Cu rejection above 92% at CuSO₄ concentrations up to 10 mM and over pH 2–5. The preferential permeation of H⁺ also enabled acid reclamation while retaining Cu(II) [69].

NF performance is strongly affected by the ionic composition of the feed. In simulated acid mine drainage, rejection of Cu, Mg, and Mn-containing species exceeded 97%, while changes in pH and solution composition influenced membrane charge and ion rejection [70]. However, high rejection of several divalent metals represents multimetal retention rather than Cu-specific selectivity; mixed-ion experiments are therefore required to demonstrate Cu/Ni, Cu/Zn, or Cu/Co separation.

The practical role of NF has been demonstrated at pilot scale using actual copper-mine AMD, where Cu was concentrated approximately fourfold from 0.6 to 2.4 g/L with up to 80% water recovery. Subsequent solvent extraction of the NF retentate indicated that Cu recovery of up to 97% could be achieved. Concentration polarisation limited membrane performance as the operating pressure increased, indicating an important constraint during Cu preconcentration. In the integrated NF-SX study, a retentate containing 1.1 g/L Cu was selected for the subsequent SX step to reduce the risk of membrane fouling during in-series operation [11].

A 2026 cyclic chelation-NF process using EDTA and a 150 Da polyamide membrane achieved a maximum Cu rejection of 99.88% and a flux of 11.82 × 10⁻⁶ m⁻³ m⁻² s⁻¹ at pH 5, with EDTA recycling proposed through pH adjustment. Because these results were obtained with a model effluent, Cu selectivity should still be verified in real multimetal streams containing competing complex-forming ions [71]. Overall, NF is effective for Cu retention and preconcentration, although concentration polarisation and limited evidence of Cu-specific selectivity in complex multimetal feeds remain important practical constraints.

4.4. Reverse Osmosis (RO)

Reverse osmosis (RO) operates primarily through the solution-diffusion mechanism across a dense polyamide selective layer. The strong hydration of Cu(II), together with its low solubility and diffusivity within the membrane polymer, contributes to high rejection. In CuSO₄ and Cu(NO₃)₂ solutions, RO achieved Cu rejection above 95%, while the accompanying anion had little effect on Cu rejection or permeate flux [72].

RO performance is strongly influenced by operating pressure and feed concentration. The cross-flow configuration used for Cu-containing solutions, including concentrate recirculation and permeate production, is illustrated in Figure 8. Using a BW30XFR polyamide membrane, increasing pressure from 10 to 40 bar increased Cu removal from 89.98 to 94.21% and permeate flux from 44.28 to 124.98 kg/m²h. However, the relatively small increase in rejection at higher pressure indicates that operating pressure should be optimised with respect to energy demand [73].

High Cu rejection does not necessarily indicate Cu-specific selectivity because RO also strongly retains other dissolved metals and salts. In multimetal streams, RO is therefore more suitable for purified-water production and overall metal concentration than for direct Cu/Ni or Cu/Zn fractionation. In an RO/NF cascade treating 200 L of actual mine water, approximately 70% of the Cu in the feed was accumulated in a Cu-rich concentrate. This separation resulted from the combined RO/NF configuration and concentrate reprocessing rather than from intrinsic Cu selectivity of the RO membrane.

The main practical limitations of RO include high operating pressure, concentration polarisation, scaling, and concentrate management. As water recovery increases, rejected salts accumulate in the retentate, increasing osmotic pressure and scaling tendency. In mine-water treatment, CaCO₃ scaling was controlled by circulating the concentrate through a seed-crystallisation reactor, promoting calcium carbonate deposition outside the membrane module and enabling further concentrate reduction. RO is therefore better suited as a high-quality water production and metal-rich concentrate generation step, followed by Cu recovery through precipitation, solvent extraction, or electrowinning. Practical assessment should consider water recovery, energy demand, concentrate composition, and long-term membrane stability in addition to Cu rejection [74].

4.5. Electrodialysis (ED)

Electrodialysis (ED) separates ions under an applied electric field. Negatively charged groups in cation-exchange membranes (CEMs) facilitate the transport of cations such as Cu(II), whereas anion-exchange membranes (AEMs) transport anions. A simplified schematic representation of Cu²⁺ and SO₄²⁻ transport through ion-exchange membranes during ED is shown in Figure 9 [75]. Cu transport therefore depends on membrane charge, ion speciation, current density, and solution flow rate [76]. In a CuSO₄-H₂SO₄-Fe system operated at 250 A/m², 25 °C, and 17 L/h, Cu transport ranged from 0.51 to 1.11 mol/h m², compared with 0.035–0.071 mol/h m² for Fe. Although Cu transport was higher, this does not indicate absolute Cu/Fe selectivity because separation also depends on feed composition and metal speciation [75].

Current density is a key operating parameter in ED. As the limiting current density (LCD) is approached, concentration polarisation develops near the membrane surface, whereas overlimiting operation may induce water dissociation and electroconvection. In a Cu/Zn-EDTA system, overlimiting conditions enhanced Cu²⁺ and Zn²⁺ transport, while electroconvective vortices reduced membrane fouling and scaling [77].

Cu-specific separation can be substantially improved by controlling metal speciation. For simulated Fe(III)-containing electroplating wastewater, Na₂EDTA-assisted ED achieved 92.2% Cu recovery, a Cu concentration factor of 2.95, and 94.3% Fe retention. Subsequent electrolysis of the ED concentrate increased Cu recovery to 94.49% and produced metallic copper with a purity of 99.4% [12]. This demonstrates that combining complexation with ED and electrolysis can provide more effective Cu/Fe separation and actual copper recovery than charge-based ED alone.

The practical performance of ED is limited by current density, flow conditions, competing ions, and membrane fouling or scaling. Selective separation of similarly charged ions such as Cu²⁺, Ni²⁺, Zn²⁺, and Co²⁺ remains difficult with conventional CEMs. ED is therefore particularly promising for Cu recovery when integrated with speciation control, selective concentration, and subsequent electrochemical metal recovery [12,77].

4.6. Liquid Membranes (SLM, PIM, ELM)

Separation in liquid membranes is governed by carrier-mediated transport rather than pore-size exclusion. Cu(II) forms a reversible complex with a carrier at the feed/membrane interface, diffuses through the organic membrane phase, and is released into the stripping solution. Selectivity therefore depends on carrier affinity, pH, metal speciation, and competing ions. The main configurations include supported liquid membranes (SLM), emulsion liquid membranes (ELM), and polymer inclusion membranes (PIM).

In SLMs, a carrier-containing organic phase is immobilised within a porous support. The carrier-mediated extraction, diffusion, and stripping mechanism in an SLM is illustrated in Figure 10. In a D2EHPA-based Cr(VI)/Cu/Zn system, Cu recovery reached 99.1% after 26 h with a Cu purity of 99.8%, demonstrating that selective recovery can be controlled through carrier and stripping chemistry. However, loss of the organic phase or carrier remains an important limitation for long-term SLM stability [78].

ELMs combine extraction and stripping within emulsion droplets. A sunflower-oil-based ELM containing D2EHPA as the carrier and HCl as the stripping phase achieved more than 94% Cu(II) extraction within 130 min. However, high extraction does not necessarily represent actual Cu recovery or Cu-specific selectivity, while emulsion breakage, swelling, and demulsification remain important scale-up limitations. In this ELM system, emulsion stability was controlled by optimising the concentrations of the carrier, surfactant, and stripping agent together with the internal-to-organic phase ratio, allowing high Cu(II) extraction to be maintained under a stable emulsion formulation [79].

In PIMs, the carrier is immobilised within a polymer matrix, improving membrane-phase stability compared with SLMs. A triazole-based PIM showed a transport sequence of Cu²⁺ > Ni²⁺ > Co²⁺, with Cu flux strongly dependent on carrier structure [80]. An ionic-liquid PIM achieved a Zn/Cu separation factor of 1996; however, this value reflects preferential Zn transport rather than Cu-selective transport. For PIMs, long-term operational stability can be limited by loss of the liquid carrier from the polymer matrix. After 2048 h of exposure to a multimetal solution in 1 M HCl, the investigated ionic-liquid PIMs retained only 48–82% of their initial ionic-liquid mass, depending on membrane composition, indicating that appropriate selection of the ionic-liquid/polymer combination is critical for reducing carrier loss during prolonged operation [81].

Compared with conventional solvent extraction, liquid membranes can combine carrier-mediated extraction, transport, and stripping within a membrane-based separation process, as demonstrated for Cu-containing and multimetal model solutions. The cited SLM, ELM, and PIM studies were conducted at laboratory scale using prepared aqueous systems, while practical limitations included emulsion stability in ELMs and long-term ionic-liquid retention in PIMs. Therefore, although liquid membranes provide additional control over metal separation through carrier and stripping chemistry, the cited studies do not establish a clear industrial-scale advantage over conventional solvent extraction in terms of treatment cost or long-term operational reliability [78,79,81].

Beyond the performance of individual membrane processes, Cu recovery from polymetallic solutions depends on whether the membrane provides genuine Cu-selective separation or only multimetal retention and concentration [11,12,82,83]. Representative membrane-based strategies for selective and stepwise Cu separation from polymetallic solutions are summarised in

Table 2. Representative membrane-based strategies for selective and stepwise Cu separation from polymetallic solutions.

Feed/Metals	Membrane Approach	Separation Behaviour	Downstream Recovery	Key Quantitative Result	Interpretation	Ref.
Real electroplating wastewater; Cr/Cu/Ni	NF	Simultaneous multivalent-metal retention	Not reported	Cr 95.76%, Cu 95.33%, Ni 94.99% rejection	High Cu rejection, but not Cu-specific selectivity	[84]
Model Cu/Co solution	MEUF with SDS + iminodiacetic acid	Complexation-controlled Cu/Co separation; Cu → permeate, Co → retentate	Surfactant regeneration	>90% Cu/Co separation; SDS recovery 75–83%	Selectivity arises mainly from complexation, not UF pore selectivity	[82]
Actual copper-mine AMD	Pilot NF	Cu preconcentration	Solvent extraction + stripping	Cu 0.6 → 2.4 g/L; water recovery up to 80%; subsequent Cu recovery up to 97%	NF preconcentration → SX recovery cascade	[11]
Cu/Fe(III) model wastewater	Na2EDTA-assisted ED	Cu/Fe separation through controlled complexation	Electrolysis	92.2% Cu recovery; CF 2.95; 94.3% Fe retention; electrolysis: 94.49% Cu recovery, 99.4% purity	ED separation → metallic Cu recovery	[12]
Real plating wastewater; Cu/Zn	Selective-stripping SLM	Separate Cu and Zn transport/stripping	Selective stripping	Cu: 83.7% recovery, 91.0% purity; Zn: 85.5% recovery, 99.3% purity	Direct polymetallic fractionation into separate product streams	[83]

.

Overall, these examples show that effective Cu separation from polymetallic streams is often achieved through complexation/speciation control or by coupling membrane concentration with downstream solvent extraction, selective stripping, or electrolysis, rather than through intrinsic membrane selectivity alone [11,12,82,83].

5. Industrial Scale-Up and Techno-Economic Considerations

Overall, pilot-scale application of MF has been demonstrated for real wastewater from an open-pit copper mine in Bulgaria using a hybrid flotation/submerged-microfiltration process. Dispersed-air flotation recovered about 90% of the suspended solids, while the Cu content of the flotation concentrate approached 6 wt%; the study therefore demonstrated Cu concentration in the flotation fraction rather than production of a final Cu product [85]. In a related on-site bench-scale study, the annual operating cost including depreciation at a treatment capacity of 10 m³/h was reported as US$1.761 million for separate flotation and MF and US$1.42 million for the integrated hybrid cell [86]. A recent laboratory-scale study of Cu sulfide precipitation followed by PVDF MF also included a techno-economic assessment based on a simulated industrial flow of 200 m³/h; the estimated CAPEX was 41.5 US$/(m³/h) for polymeric membranes and 46.1 US$/(m³/h) for ceramic membranes, while the corresponding OPEX values were 8.4 and 8.6 US$ m³/h, respectively [87].

For UF-based recovery, electric-field-enhanced PEUF has been tested with two real industrial wastewaters. Cu removal was approximately 95%, while subsequent electrolysis recovered approximately 80% of Cu and allowed regeneration of the functional polymer; application of +0.2 V also reduced the membrane flux decline by 58% [62].

NF has been evaluated at pilot scale using actual acid mine drainage from copper mining. Cu was concentrated from approximately 0.6 to 2.4 g/L, water recovery reached 80%, and total species rejection was approximately 82%. The NF retentate was subsequently treated by solvent extraction with LIX 84-IC, for which two counter-current extraction stages followed by one stripping stage were predicted to recover approximately 97% of Cu [11]. Thus, the membrane provided Cu pre-concentration, whereas the reported 97% Cu recovery was associated with the downstream solvent-extraction stage [11].

RO has also been examined for mine-water treatment, where Cu and other dissolved metals were concentrated in the brine rather than recovered as final products. The estimated total treatment cost for RO combined with feed prefiltration was approximately €0.34 m⁻³. Suspended-solids prefiltration was required to reduce membrane fouling [88].

A more complete Cu-recovery route has been demonstrated using pilot-scale ED followed by electrowinning. ED treatment of real plating wastewater at 200 V (13.3 V per cell pair) produced a metal-enriched concentrate, which was subsequently processed by pilot electrowinning. From ED concentrates initially containing 790–1145 mg/L Cu, electrowinning recovered approximately 93–96% of Cu as Cu(0) and Cu₂O, with a specific energy consumption of 3.5–5.3 kWh/kg Cu [53].

Liquid membranes provide additional scale-up examples. An HFSLM process for spent ammoniacal PCB etchant was scaled from a 1.4 m² bench module to a 130 m² pilot hollow-fibre contactor. The pilot process achieved an approximately 3000-fold Cu removal factor and produced a nearly saturated CuSO₄ stripping solution; CuSO₄·5H₂O crystals obtained from this solution were reported to have quality comparable to or better than commercial products [89]. Field-scale operation was also demonstrated for an emulsion liquid membrane system at an Arizona copper mine, where approximately 5600 gal of solution was processed continuously over nine days. Cu extraction was 98.0% from a pregnant leach solution containing 1.43 g/L Cu and decreased to 95.7% and 91.6% for feeds containing approximately 500 and 320 mg/L Cu, respectively; the extracted Cu was subsequently recovered by electrowinning. Membrane swelling remained below 8% and leakage below 0.1% during the field test [90].

These studies show that membrane processes can serve as both water-treatment and Cu-concentration steps, whereas actual Cu recovery is generally achieved through integration with solvent extraction, stripping, crystallisation, electrolysis, or electrowinning [11,53,89,90].

6. Conclusions

This review shows that Cu separation and recovery cannot be assessed from removal efficiency alone. The four stream categories—mine-affected waters, copper-smelting waste acids, electroplating wastewaters, and secondary-material leachates—differ in Cu concentration and speciation, acidity, ionic strength, ligands, suspended or organic matter, and competing metals. These properties determine whether a process selectively separates Cu, retains several metals together, or merely transfers them to another phase.

The comparison with conventional operations shows that membranes are generally complementary rather than stand-alone replacements. Precipitation yields a Cu-bearing solid whose purity may be reduced by co-precipitation; solvent extraction and ion exchange require stripping or elution; and electrowinning produces metal most effectively from sufficiently concentrated, predominantly free Cu²⁺ solutions with controlled impurities. Membranes can prepare such feeds. MF and UF require precipitation, complexation, micellisation, or surface binding; NF and RO mainly recover water and generate metal-rich retentates; ED becomes more selective when speciation is controlled; and SLMs, ELMs, and PIMs can fractionate metals through carrier and stripping chemistry. Thus, high Cu rejection is not proof of Cu-specific recovery.

The strongest recovery results were obtained using integrated flowsheets. Pilot NF followed by solvent extraction preconcentrated Cu from acid mine drainage, while complexation-assisted ED followed by electrolysis separated Cu from Fe and produced high-purity Cu. Pilot or field evidence also exists for hybrid MF, RO, HFSLM, and ELM systems, but remains uneven; PIM evidence is mainly laboratory-scale. In most cases, the membrane produces a Cu-rich retentate, concentrate, or stripping solution rather than a final metallic product. Economic comparison remains limited by inconsistent reporting of capacity, energy and reagent demand, membrane or carrier lifetime, CAPEX, OPEX, and product value.

Future work should prioritise continuous testing with variable real feeds, complete mass balances, separation factors, flux decline, product composition and purity, and flowsheet-level evaluation of pretreatment, regeneration, concentrate management, and downstream recovery. Membranes should ultimately be judged by the amount and quality of recoverable Cu produced, not by removal percentage alone.