Graphene/chitosan nanoreactors for ultrafast and precise recovery and catalytic conversion of gold from electronic waste

Significance

Circular economy requires efficient means to recover essential elements from the generated waste. This is of particular importance for the recycle of electronic waste, where rare materials are often used. Gold extraction from electronic waste is a particularly important issue. Novel materials can play an important role in this process. We developed unique, composite materials capable of not only extraction but also reduction of gold from e-waste. Providing simultaneous chemisorption and chemical reduction of gold from e-waste without applying any power, our approach creates economic value from otherwise discarded items. This innovation in gold recovery not only enhances efficiency and sustainability but also marks a significant advancement in developing ecofriendly solutions for managing e-waste and conserving natural resources.

Abstract

The extraction of gold (Au) from electronic waste (e-waste) has both environmental impact and inherent value. Improper e-waste disposal poses environmental and health risks, entailing substantial remediation and healthcare costs. Large efforts are applied for the recovery of Au from e-waste using complex processes which include the dissolution of Au, its adsorption in an ionic state and succeeding reduction to metallic Au. These processes themselves being complex and utilizing harsh chemicals contribute to the environmental impact of e-waste. Here, we present an approach for the simultaneous recovery and reduction of Au³⁺ and Au⁺ ions from e-waste to produce solid Au⁰ forms, thus skipping several technological steps. We develop a nanoscale cross-dimensional composite material via self-assembly of two-dimensional graphene oxide and one-dimensional chitosan macromolecules, capable of acting simultaneously as a scavenger of gold ions and as a reducing agent. Such multidimensional architecture doesn’t require to apply any voltage for Au adsorption and reduction and solely relies on the chemisorption kinetics of Au ions in the heterogeneous GO/CS nanoconfinements and their chemical reduction on multiple binding sites. The cooperative phenomena in ionic absorption are responsible for the extremely high efficiency of gold extraction. The extraction capacity reaches 16.8 g/g for Au³⁺ and 6.2 g/g for Au⁺, which is ten times larger than any existing gold adsorbents can propose. The efficiency is above 99.5 wt.% (current limit is 75 wt.%) and extraction ability is down to very low concentrations of 3 ppm.

The extraction of gold in both its Au⁺ and Au³⁺ oxidation states from electronic waste (e-waste) is a complex and environmentally important process due to the value of Au and the environmental concerns associated with e-waste disposal. The global e-waste generation was estimated to be around 53.6 million tons in 2019, according to the Global E-waste Monitor report (1). Improper disposal and management of e-waste lead to environmental contamination and potential health hazards (2). The costs associated with environmental remediation and health-related expenses can also be significant. This includes costs related to soil and water cleanup, air pollution mitigation, and healthcare expenses for affected individuals.

The processes involved in the extraction of Au from e-waste include the collection of old electronic items, their mechanical crashing, formation of soluble AuCl₃ salt, extraction of Au³⁺ ions, precipitation, and recovery of reduced Au (via utilization of reduction agents). Many materials such as activated carbon (3), carbon nanotubes (4, 5), zeolites (6, 7), metal-organic frameworks (8), silica-based materials (9, 10), polymeric hydrogels (11, 12), and reduced graphene oxide (rGO) (13) have been employed as Au³⁺ adsorbents. The adsorption of Au³⁺ ions is typically followed by the reduction process, involving such chemical agents as hydrogen gas, citrate ions, sodium borohydride, and ascorbic acid (14). Extraction of Au⁺ ions (which are typically less common in e-waste but can be found in specific electronic components) necessitates a more complex series of chemical reactions (15). Recovering Au in its Au⁺ oxidation state typically demands adsorbents with higher capacitance and selectivity. Simultaneously, the reduction of Au⁺ ions demands more potent reducing agents and the addition of stabilizing agents to counter the stability of the Au⁺ complex, facilitating the reduction process (16). A common method for recovering of Au⁺ involves its adsorption onto specialized ion exchange resins (such as activated carbon resin impregnated with thiol groups), followed by the reduction using sodium metabisulfite and stabilization with polyethylene glycol (17, 18). Such multistage procedure for the extraction and reduction of both Au³⁺ and Au⁺ leads to significant environmental impact and increased cost. Industries and the economy would experience significant benefits from a process that enables the adsorption of Au³⁺ and Au⁺ ions, along with their subsequent reduction to metallic Au, all within a single procedure characterized by high efficiency and yield. In general, all conventional ion recycling methods such as electrowinning, smelting, and chemical leaching face bottleneck issues such as limited adsorption capacity, slow adsorption kinetics, and regeneration challenges. In addition to high extraction efficiency, an adsorbent should offer rapid kinetics, selectivity, low production costs, and the capability to reduce energy and chemical usage during the extraction process (19). Moreover, it should be mechanically robust to ensure long-term stability.

Here, we proposed a unique approach for the recovery of Au ions from e-waste, which bypasses several technological steps and combines high ion adsorption capacitance, fast adsorption kinetics as well as catalytic properties and regeneration capability. We propose a material, that acts simultaneously as Au ions scavenger and reducing agent, allowing direct formation of Au nanoparticles (AuNPs) from both Au³⁺ and Au⁺ salts which doesn’t require to apply any voltage. To achieve synergetic chemosorption and chemical reduction processes within the structure of our material, we combine materials with distinct dimensionalities and a variety of functional groups. Specifically, we integrate two-dimensional (2D) materials with one-dimensional (1D) macromolecules, creating a fractional dimensionality to amplify ion extraction and reduction. Such fractional architecture not only facilitates ion transport and adsorption but also promotes efficient electron transfer across interfaces. We selected graphene oxide (GO) flakes and chitosan (CS) chains as components for our nanoreactors due to their intrinsic functionalities, which include catalytic properties and ionic transport (20, 21). Importantly, the presence of oppositely charged functional groups and hydrophobic domains in GO and CS ensures interfacial compatibility and stability. These multiple interactions between 2D materials and 1D macromolecules create synergistic effects, enhancing the combined system’s properties beyond those of its individual components (22).

Additional benefits come from the technology-friendly method for the synthesis of our composites by using self-assembly technique, which enables the self-organization of GO and CS into GO/CS building blocks and their assembly into well-defined nanolayers with chemically active nanoconfinements. This approach draws inspiration from biological and physical systems where organization emerges spontaneously from functional components or building blocks. One of the key features of self-organized materials is the emergence of new properties or the amplification of existing properties at the macroscopic scale that are not present or low efficient in the individual components (23). This emergence is a result of the collective interactions and arrangements of the components. For instance, both GO and CS are traditionally used in Au extraction and reduction, although their efficiency is limited. In the presence of appropriate conditions, CS can reduce Au³⁺ to Au⁰. However, in aqueous medium the swollen polymer matrix passivates mass transport and electron transfer (21, 22). On the other hand, graphene-based adsorbents are known for their high surface area, high ionic capacitance, and rapid ionic transport (23, 24). Nonetheless, this high capacitance leads to structural instability, resulting in layer separation and material expansion (25). The expanded aqueous GO structure also impedes both mass and electron transfer processes. To address all these challenges in our work, we employ a self-organization approach to harness natural individual properties of GO and CS to create materials with amplified functionalities. The outcome of our research is the amplification of ion adsorption and catalytic properties in both GO and CS that leads to increased ion extraction yield, faster adsorption rates, improved selectivity for Au ions and chemical transformation to solid Au forms on demand.

The detailed procedure for the synthesis of our composite material is given in Materials and Methods. In brief: first, CS was dissolved in 1 %v/v acetic acid (HOAc) and magnetically stirred at room temperature for 48 h to form a stable CS/HOAc dispersion. Then, 2 mg mL⁻¹ GO dispersion was sonicated for 30 min before use to form dispersed GO flakes. The second step involves the mixing of the CS dispersion with the GO dispersion in a certain mass ratio to form composite nanosheets. Upon mixing, the CS molecules self-assemble on the surface of GO flakes to form accessible ion-binding sites (Fig. 1A).

Fig. 1.

GO/chitosan sponge for gold adsorption and filtration. (A) The geometry of CS⋅⋅⋅GO complex optimized by density functional theory calculation. (B) Schematic diagram of GO/CS sponge for ultrafast and efficient adsorption/extraction of gold, δ = 1 or 3. (C) SEM image of a GO/CS₅ sponge. (D) Photographs illustrate a GO/CS sponge filled in a syringe (Left), followed by adding 5 mL 200 ppm Au³⁺ solution (Middle) and the visualization of real-time color change before and after filtration (Right). It took approximately 2.86 min for 2.5 mg of sponge adsorbent, with a rate of adsorption of 0.14 g/g min (as detailed in SI Appendix, Table S1), to extract 200 ppm of gold from 5 mL.

The next step is to align GO/CS flakes in multilayers, allow them to swell in water, and then form GO/CS sponge using freeze drying. The GO/CS sponge (Fig. 1 B and C) is designed as a unique composite adsorbent that can chemisorb Au ions with high capacitance and high extraction rate. To obtain GO/CS sponges, we employed a lyophilization technique on the GO/CS suspensions. A sponge-like morphology featuring a network of interconnected voids ranging from 10 to 50 μm enables direct use as a gold absorbent, enhancing mass flow. As we see, the distinctive characteristics allow the sponge to be directly applied as a filter to extract Au ions from solutions (Fig. 1D). In our previous works, we found that the surface area of the GO sponge is 139.2 m²/g (26) and the GO/polyelectrolyte sponge is around 400 m²/g (27). The polyelectrolyte addition to the GO mostly leads to the change in pore size than in surface area.

Fourier transform infrared (FT-IR) spectra (SI Appendix, Fig. S1A) proved that the composite materials have the characteristic peaks of both GO and CS. The formation of nanoscale cross-dimensional composite is confirmed by X-ray diffraction (XRD) that indicates the presence of structural elements, with XRD patterns revealing two peaks (SI Appendix, Fig. S1B): 12.9 Å in CS-containing nanoconfinements and 8.6 Å in CS-free regions. Based on the composition, we assigned the sponges as GO/CS_x, with x representing the mass ratio of GO to CS. Chemical stability tests (SI Appendix, Fig. S2) revealed that the GO/CS₁₀ has an optimal set of mechanical and chemical stability (see the detailed discussion in SI Appendix).

First, focusing specifically on Au³⁺, our goal was to determine the optimal conditions that lead to the maximum efficiency of our material in the treatment of e-waste. By systematically varying and analyzing various parameters such as the material’s composition, pH, temperature, contact time, and ion concentration, we aimed to determine the optimal physicochemical parameters that lead to the highest possible recovery efficiency for Au ions. In our experiments, it’s noteworthy that we employed our sponge technology to adapt our approach to the existing methods for extracting unrecoverable residue Au from waste following the electrowinning process.

An aqueous solution containing Au³⁺ at a fixed concentration of 8,170 ppm was utilized to reveal the optimal GO/CS_x composition and pH for Au extraction (Fig. 2 A–C). Then, 2.5 mg sponges were immersed for 1 h into a feed solution. The feed and supernatant solution concentrations are determined using inductively coupled plasma-optical emission spectroscopy (ICP-OES). The extraction capacity is subsequently calculated from the ICP-OES results (Materials and Methods for details). The sponge composed of GO/CS₁₀ demonstrates the highest capability to chemisorb Au³⁺, achieving an adsorption capacity of 16.8 ± 2.3 g/g at room temperature. The optimal pH range for ion adsorption falls within the vicinity of the pKa values of GO (pKa 4.6) and CS (pKa 6.5), (Fig. 2B). Thus, at pH 5 to 7, a diversity of surface chemistry of both GO and CS outcomes in increasing the ability of the material for Au extraction (24, 25). At pH 5 to 7, –COOH groups electrostatically attract cations and –NH₂ groups coordinate Au ions and stabilize the complexes with Au ions. Finally, the reusability of sponges was tested in DI water (Fig. 2C). Following each adsorption phase, the sponge underwent a thorough washing process with DI water, involving 20 min of sonication at 20 °C repeated three times. The results demonstrate that GO/CS sponges maintain their excellent adsorption capacity, ensuring the reversibility and reusability of the sponge throughout the testing period.

Fig. 2.

Extraction capacity, thermodynamics, and kinetics of Au³⁺ chemisorption. (A) Gold extraction capacity as a function of CS:GO mass ratio after 1 h immersion in 6,800 ppm Au³⁺, red—GO/CS sponge, gray—pure CS sponge. Inset: photograph showing GO/CS mixture when the ratio is less than 5. (B) Extraction capacity of GO/CS₁₀ prepared at different pHs after 1 h immersion in 8,170 ppm Au³⁺. (C) Reusability test showing the gold extraction capacity of GO/CS₁₀ sponges after 5 cyclic adsorption–desorption procedures. The desorption is achieved by washing and sonication in ultrasonic bath (D) Extraction thermodynamics study through the concentration equilibrium constant (K_c) dependence on temperature for GO/CS₁₀ sponge immersed in 200 ppm Au³⁺ for 1 h. (E) Extraction capacity of GO/CS₁₀ sponge prepared at pH 5 after 1 h immersion in aqueous solution of Au³⁺ with concentrations ranging from 2 ppm to 100 ppm, fitted by Hill model (red line). Inset: photograph of a GO/CS₁₀ sponge. (Scale bar, 3 mm.) (F) Kinetic fitting results of pseudo-first order of reaction for GO/CS₁₀ sponge.

The study of thermodynamics and kinetics (Fig. 2 D–F) allows us to estimate overall feasibility of such intensive chemisorption and evaluate the rate of the process. First, a 2.5 mg sponge was immersed in 10 mL of 200 ppm Au³⁺ solution at various temperatures, ranging from 5 °C to 60 °C, for a duration of 1 h and analyzed by ICP-OES. The thermodynamics of the overall energy changes during Au ions chemisorption (Fig. 2D) reveal the exothermic spontaneous chemisorption process (ΔG^°_ads−1.87 kJ/mol), with a calculated ΔH^°_ads value of −0.97 kJ/mol. Importantly, the positive value of ΔS^°_ads (3 J/mol K) suggests an increased randomness at the heterogeneous solid–solution interface during the chemisorption of Au³⁺ onto the binding sites of the CO/CS composite.

Chemisorption kinetics were monitored for GO/CS₁₀ sponge using ICP-OES technique (experimental details are provided in SI Appendix, Fig. S3). Additionally, a value of R² of 0.9999 was calculated with the pseudo-second-order model (SI Appendix, Table S1). Furthermore, we calculated the rate constants for the chemisorption on the GO/CS sponge to be 0.14 g/g min

Notably, among the existing models for adsorption isotherms (we compared the fitting parameters using Langmuir, Hill, Freundlich, and Temkin isotherms in SI Appendix, Table S2), the Hill model demonstrated a high correlation coefficient (Fig. 2 E and F). Importantly, in our case, the Hill equation results in a typical sigmoidal curve when Au ions extraction capacity is plotted against Au ions feed concentration:

$$q_e=\frac{q_{\mathit{max}}·C_{\mathit{rest}}^n}{K^n+C_{\mathit{rest}}^n}$$ [1]

where q_e is the equilibrium adsorption capacity, q_max—maximal theoretical adsorption capacitance, C_rest—rest concentration after adsorption, 1/K is the adsorption constant (K_ads), n—is the stochiometric coefficient for the number of ligands that bind with molecules.

This sigmoidal shape is characteristic of cooperative binding. The Hill coefficient (n) in Eq. 1 represents the degree of cooperativity. If n is greater than 1, it indicates positive cooperativity (binding of one ligand enhances the binding of subsequent ligands). From Hill isotherm model for GO/CS sponge, we found that the n value is 1.28. Thus, the study of thermodynamics and kinetics reveals an ion chemisorption phenomenon that is thermodynamically driven and suggests a typical for biological macromolecule cooperativity, where the binding of an ion at one site affects the binding of ions at other sites on the same macromolecule.

Furthermore, the adsorption rate is much greater than other graphene-based adsorbents have (13, 28), which may take several days to reach an equilibrium Au extraction (see the summary of the extraction rates for other adsorbents in SI Appendix, Table S3). The rapid and efficient kinetics of Au ion extraction observed in our material arise due to the capillary action in the nanoconfinements with heterogeneous surface chemistry, featuring hydrophobic sp² carbon domains and acetyl groups (–COCH₃) as well as multiple hydrophilic groups.

Note that both GO and CS have been used in Au extraction and reduction, although their efficiency is comparatively lower when used individually. We measured that the pure CS sponge has a significantly lower ability to extract Au compared to the GO/CS composite. CS’ maximum adsorption for Au³⁺ is measured to be 2.8 ± 0.3 g/g (Fig. 2A). On the other hand, it was reported that the extraction capacity of reduced graphene oxide (rGO) for Au³⁺ exceeded a value of 1.85 g/g (13) at 10 ppm feed Au concentration. However, when the concentration of Au ions was increased above 10 ppm, there was no significant increase in the extraction capacity of rGO. We observed that the pure GO sponges, without the presence of CS, disintegrate when exposed to the concentrated feed Au³⁺ solutions. The high osmotic pressure leads to structural instability, resulting in layer separation and material expansion (29). Electrostatic cross-linking between GO and CS stabilizes the composites preventing the composite from swelling and disintegrating.

Through the utilization of SEM/EDX, we analyze the surface of GO/CS sponges after the chemisorption process (Fig. 3 A and B). Our observations reveal a striking feature: Along with chemisorbed Au ions, a significant and uniform distribution of solid Au forms such as Au clusters and Au nanoparticles (AuNPs) on the surface of the GO/CS sponges extracted either from Au⁺ or Au³⁺ solutions. X-ray photoelectron spectroscopy (XPS) analysis (Fig. 3 C and D) is used to elucidate the contribution of functional groups of GO and CS to the adsorption and reduction of Au⁺ and Au³⁺. The XPS peaks at 84 and 88 eV confirm the reduction of both Au³⁺ (Fig. 3A) and Au⁺ (Fig. 3B) to metallic Au⁰. However, in the spectrum of Au⁺, in addition to the peaks at 84 and 88 eV that confirm the reduction of Au⁺, we observe the presence of a peak at 90 eV. The presence of a peak at 90 eV in the XPS spectrum of pure CS and GO/CS sponges suggests the formation of a complex between Au⁺ and CS (30). This peak can be attributed to the characteristic binding energy of the Au⁺–CS interaction. Density functional theory (DFT) computation (SI Appendix, Tables S4 and S5) was provided to get an insight into Au cation’s interaction with the graphene plane and CS chains of composites. Computations demonstrate repulsive interaction between positively charged Au ions and graphene lattice. Graphene π–electron clouds formed by sp² hybridized carbon atoms tend to maintain their structure without bonding with Au. CS chains on the contrary have a significant ability to interact with Au ions. O and N atoms are preferable sites for Au ions binding as they keep the negative partial charge (Mulliken charge). This interaction is driven by Coulombic forces, resulting in the adsorption of Au ions by the CS chains. Notably, the strength of this interaction depends on the charge state of Au ions. For Au⁺, interactions primarily occur with the amino group, while for Au³⁺, this interaction extends to include nearby oxygens from ether and hydroxyl groups (as shown in Fig. 3C). The adsorption of Au ions on these sites was associated with a significantly negative Gibbs free energy change (refer to SI Appendix, Table S4).

Fig. 3.

Au⁺ and Au³⁺ extraction and reduction by GO/CS sponge. (A and B) SEM/EDX image of Au³⁺ (A) and Au⁺ (B) adsorption by GO/CS sponge, gold element is shown as yellow color by EDX mapping. (C and D) 4f XPS spectrum of gold nanoparticles on pristine GO, CS, and GO/CS sponges after Au³⁺ (A) and Au⁺ (B) adsorption. (E) The most stable geometries of CS²⁺…Au³⁺…GO²⁻ (Left) and CS²⁺…Au⁺…GO²⁻ (Right) complexes optimized by density functional theory calculation. Atoms color notation: C—gray, H—white, O—red, H—white, Au—yellow. (F) Comparison of gold extraction performance in the aspect of Au³⁺ (spheres) and Au⁺ (tetrahedrons) concentration, equilibrium time of extraction, and extraction capacity (SI Appendix, Table S3 for details). Inset: Elemental analysis of GO/CS sponge by SEM/EDX.

The presence of multiple binding sites ensures the high efficiency of both the chemisorption and chemical reduction processes. The whole process can be described as a three-stage reaction: stage 1—the carboxylic groups (−COOH) on GO act as anchoring points for Au ions, while the lone electron pair on the nitrogen of the protonated CS amino (−NH₃⁺) groups is available for chelating with cations; stage 2—the reduction of Au ions to metallic gold via the transfer of electrons from the reducing agent (amine) to gold ions; stage 3—oxidation reaction of aminium radical (−NH₂^+⋅) (31) and formation imino groups (−NH) (32):

Initially, the chitosan is protonated due to dissolving in acetic acid with −NH₃⁺ groups. Further, the complexation of Au^δ+ leads to the deprotonation of chitosan and GO (1). The reduction of Au^δ+ provides oxidation of chitosan with the formation of aminium radical (−NH₂^+.) and protonation GO (2). The final step is the oxidation of −NH₂^+. to the imino groups (−NH) and proton release (3). The cyclic voltammetry of the GO, CS, and GO/CS aqueous mixtures exposed on Au³⁺ allow to demonstrate the partial chemical reduction of Au³⁺ to Au⁰ (SI Appendix, Fig. S4). The reduction peak height for the GO does not change in comparison with pure Au³⁺ solution, which proves the absence of chemical reduction by GO. The reduction peak for the CS and GO/CS shifts to the positive direction, proving the formation of the CS-Au³⁺ complex with the facilitating of electrochemical reduction. The molar relation between Au and CS is 1:0.9 (refer to SI). A detailed study of Au reduction by GO/CS₁₀ in solution was carried out at a fixed CS (22.4 mg) and GO (2.24 mg) composition, with the Au³⁺ content in situ increasing from 0 to 20 mg. The amount of Au⁰ is 3.8 × 10¹⁹ atoms (12.5 mg) according to the equilibrium state between the Au⁰ and Au³⁺ complexed with GO/CS₁₀. Thus, the relation N to Au is 1.76 to 1, which means the around 2 amino groups of chitosan are able to reduce 1 Au³⁺ ion. Such proportion corresponds to the partial amount of gold reduction process by chitosan and proves the ability to achieve high extraction capacity for the GO/CS₁₀ composite. This stochiometric relation between N and Au is evidence of the repeatable reduction and chelation of Au even by oxidized chitosan with imine groups.

Comparing our composites to existing adsorbers, we observe a significant improvement in ion recovery behavior for both Au⁺ and Au³⁺. We provide a summary of values reported in the literature to establish the cutoff values for the adsorption capacitance of existing adsorbents for Au ions (Fig. 3D and SI Appendix, Table S3) (13, 15, 33–50). While previous adsorbents demonstrated Au(I) extraction capacities of around 0.3 g/g and Au³⁺ extraction capacities of around 2 g/g, our material can extract 16.7 g/g for Au³⁺ and 6.2 g/g for Au⁺ within just 10 min and simultaneously reduce both ions.

In the context of a real multicomponent system containing various ions and trace concentrations of Au, we showed the extraction efficiency of the GO/CS sponge by utilizing the waste mixture remaining after the electrowinning process (51). The testing waste solution after electrowinning was provided by SG Recycle Group SG3R, Pte, Ltd. Fig. 4A illustrates an industrial electrowinning process. The composition of the waste solution after electrowinning is shown in Fig. 4B, with nickel (Ni, 52.3 ppm) and iron (Fe, 31.1 ppm) being the major components and 3.1 ppm (300 mg/L) of unrecoverable residue Au. pH of such mixture is 8.8.

Fig. 4.

Gold extraction from electronic waste. (A) A schematic illustration of the industrial electrowinning process for gold extraction from e-waste that typically involves electrochemical process and waste treatment. (B) After the electrowinning process provided by SG Recycle Group SG3R, Pte, Ltd., the composition of the waste mixture typically contains a residue of gold that is unrecoverable, measured at 3.1 ppm. (C) Following electrowinning and waste treatment with GO/CS sponge at pH 8.8, the composition of the waste mixture undergoes further alteration. This treatment process aims to enhance the recovery of metals, such as iron, from the waste mixture but not Au. (D) The waste mixture undergoes sequential treatment with GO/CS sponge after adjusting the pH to 3, resulting in a complete recovery of residue Au from the waste mixture. Each treatment has an immersion period of 1 h for all experiments.

As illustrated in Fig. 4 C and D, pH adjustment is important to attain high extraction efficiency. In the specific composition of the waste solution, an optimal pH of 3 was determined. Fig. 4D shows the extraction efficiencies achieved for various metals utilizing the GO/CS sponge with a pH 3 waste solution (refer to the Materials and Methods section for further details). The selective extraction in these conditions is successful and requires Fe masking before Au extraction to achieve full selectivity.

To conclude, our approach focuses on nanoscale cross-dimensional composite material designed to chemisorb Au ions and accelerate catalytic reduction within nanocompartments. We combine two-dimensional graphene oxide and one-dimensional chitosan due to their selective ionic transport, catalytic properties, and complementary functionalities. The self-organized composites exhibit heterogeneous surface chemistry, enabling the chemical reduction of Au ions and proton transfer reactions to return catalytic amino groups to their initial state. The study of the thermodynamics and kinetics of chemisorption reveals a spontaneous exothermic process with cooperative contributions from multiple functional groups. As a result, the ionic transport and catalytic properties of the composite surpass those of individual GO and CS components. Our materials demonstrate exceptional affinity for Au⁺ and Au³⁺, surpassing existing technologies. They are capable of chemisorbing 16.8 g/g of Au³⁺ and 6.2 g/g of Au⁺ within 10 min. This work provides a sustainable solution for Au recovery from e-waste and waste treatment after industrial electrowinning with 99.5% extraction efficiency, contributing to both environmental preservation and resource utilization. While the recovery of gold is ultrahigh, the sponge also exhibits promising efficiencies for other valuable metals, such as silver, copper, cobalt, and nickel.

Materials and Methods

Materials.

Aqueous graphene oxide dispersion (GO, 4 mg mL⁻¹, monolayer content >95%, Graphenea Inc.), Chitosan (CS, powder, from shrimp shells, Sigma-Aldrich), Polyethersulfon membrane filter (PES, 0.03 μm, 47 mm, Sterlitech Corporation), Anodisc 47^TM filter (pore size—0.02 μm, diameter 47 mm, Whatman), and Acetic acid (HOAc, glacial, ReagentPlus®, ≥99%, Sigma-Aldrich). All materials were received and used without further purification.

Preparation of GO/CS Composites.

CS/HOAc dispersion (5 mg mL⁻¹) was obtained by dissolving chitosan (2 g) in HOAc (1 vol%/vol, 400 mL) upon magnetic stirring for 24 h at room temperature. The original aqueous graphene oxide dispersion (4 mg mL⁻¹, 20 mL) was added into deionized water (380 mL) to obtain diluted GO dispersion (0.2 mg mL⁻¹). Then, CS/HOAc (5 mL, 5 mg mL⁻¹) dispersion was mixed with GO dispersion (25 mL, 0.2 mg mL⁻¹), and the colloids were then mixed for 10 min by a shaker (rotation speed—500 rpm, Vortex Mixer). GO/CS sponge was prepared by freeze-drying the aforesaid mixture in a 24-well culture plate for 24 h. For comparison, a pristine GO or CS sponge can be easily prepared by freeze-drying 2 mL GO dispersion (0.2 mg mL⁻¹) or 0.8 mL CS/HOAc dispersion (5 mg mL⁻¹) in the well plate. The ratio of CS to GO in GO/CS sponge was controlled at 5, 10, and 20 by introducing a different volume of CS dispersion, and the resulting GO/CS composite was named GO/CS₅, GO/CS₁₀, and GO/CS₂₀.

Gold Extraction Capacity Was Evaluated Using ICP-OES.

A singular GO/CS sponge, weighing 2.5 mg, was immersed in 10 mL of varied aqueous AuCl₃ solution concentrations, ranging from 2 to 8,170 ppm, for a duration of 1 h. The initial and postadsorption concentrations of gold solutions were measured by ICP-OES. In the case of AuCl₃ concentrations ranging from 2 to 100 ppm, the solutions were diluted tenfold to achieve a 10 mL diluted dispersion for subsequent ICP-OES analysis. For AuCl₃ concentrations spanning from 200 to 8,170 ppm, a hundredfold dilution was implemented to obtain a 10 mL diluted dispersion for ICP-OES testing. Standard Au calibration solutions procured from Sigma-Aldrich were utilized to calibrate ICP within the test range of 0.01 to 100 ppm during testing. Then, 2% of HCl/HNO₃ is used for washing the tubing of ICP-OES equipment. After determining the AuCl₃ concentrations before and after adsorption, the extraction capacity was calculated based on the ICP-OES results.

Procedure and Measurements for Gold Extraction from e-Waste.

A singular GO/CS sponge, weighing 2.5 mg, was immersed in 10 mL of an e-waste solution (provided by SG Recycle Group) for a duration of 1 h. The original pH of the e-waste solution was 8.8, as measured by a pH meter. For pH 3 e-waste test, hydrochloric acid (HCl) was employed to adjust the pH of the e-waste solution to 3. The initial and postadsorption concentrations of various metal ions in the e-waste solution were quantified using ICP-OES. Subsequently, the extraction capacity was calculated based on the ICP-OES results, which provided the metal concentrations before and after the adsorption process.

Characterization Methods.

Scanning electron microscopy (SEM) images were obtained by a ZEISS Sigma 300 FE SEM system with energy-dispersive X-ray spectroscopy (EDX) equipped. The sponge samples were sputtered with 5 nm gold or carbon before observation. Transmission electron microscopy (TEM) is conducted by a JEOL JEM-2200FS electron microscope (JEOL Ltd., Tokyo, Japan) at 200 kV, equipped with a direct electron DE-16 camera (Direct Electron, LP, San Diego, CA.) The concentration of metals permeating through the compressed scaffold and the concentration of metals adsorbed by the scaffold was measured by a Perkin Elmer Avio 500 ICP-OES. Before elemental analysis, a scaffold was digested with HNO₃/HCl (3:1) in a microwave at 240 °C for 15 min and topped up to 10 mL with H₂O. Note that a clear solution was observed before analysis.

DFT Computation.

The calculations were carried out with the Gaussian 16 software package with the tight self-consistent field procedure and ultrafine integration grid (52). All calculations were produced via B3LYP method and Def2SVP basis set. The geometry optimizations were performed via the Berny algorithm with 1 * 10⁻⁵ Rms force criterion (tight Opt) (53). Freq command was used for the computation of force constants and vibrational frequencies to perform the correction for thermochemistry parameters. The final Gibbs free energy changes of the reactions and Cartesian atomic coordinates are provided in SI Appendix, Tables S4 and S5.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.