Utilizing cost-effective pyrocarbon for highly efficient gold retrieval from e-waste leachate

Structural and chemical characterization of pyrocarbonPyrocarbon sorbents were synthesized via a straightforward and environmentally friendly method using sodium alginate and calcium chloride as precursors (Fig. 1a). Pyrolysis temperatures were purposefully adjusted to modulate the characteristics of the resulting PyCs (“PyCs” represents the as-synthesized pyrocarbon sorbent pyrolyzed at the temperatures of 500, 600, 700, and 800oC). Scanning electron microscopy (SEM) images exhibited granules at a millimeter scale (~1.0 mm) with flat and smooth surfaces for the PyCs (Supplementary Fig. 1). Transmission electron microscopy (TEM) image indicated that the obtained PyCs possessed a rich porous network structure (Supplementary Fig. 2a, b), corroborated by the Brunauer-Emmett-Teller specific surface area (SSA) analyses (Supplementary Fig. 3a, b). Notably, higher pyrolysis temperatures yielded increased SSAs for the PyCs (ranging from 565.7 to 673.8 m2 g-1) while maintaining similar mesoporous structures with pore diameters of 3.79 to 3.97 nm. High-resolution TEM (HRTEM) images revealed the coexistence of conductive graphitic microcrystallites and amorphous carbon within the synthesized PyCs (Supplementary Fig. 2c–f)40,41. The observed lattice spacing of approximately 0.35 nm corresponded to the (002) crystal plane of layered graphite carbon (Fig. 1b), which were consistent with X-ray diffraction (XRD) results (Supplementary Fig. 4)39,42. To reveal the relative abundance of graphitic and amorphous structures (associated with sp2- and sp3- hybridized carbon), further characterizations were employed to investigate the chemical compositions and structural information of the PyCs. Employing attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy confirmed notable alterations in functional group chemistry and carbon structure within PyCs as the pyrolysis temperatures increased. (Supplementary Fig. 5a). We identified the formation of phenolic hydroxyl (-OH) and the conversion of aliphatic C-O to aromatic C-O. This was confirmed by the presence of vibrations of aromatic C-OH stretches at 1634 cm−1 and the red shift of C-O-C stretches at 1100-1175 cm−1. These findings implied an increase of graphitization degree with rising pyrolysis temperatures35,38,43. Additionally, the escalating peak intensity at 1545 cm−1, associated with aromatic conjugated lattice in graphitic microcrystallites, further supported this result38. Furthermore, analysis of Raman spectra indicated an increased ID/IG index, signifying heightened defects within the carbon matrices of PyCs at higher pyrolysis temperatures (Supplementary Fig. 5b). This increase in defects was attributed to the intensified condensation of amorphous fragments into graphitic structure by the removal of partial O-containing moieties38,44. X-ray photoelectron spectroscopy (XPS) analysis of O 1 s and C 1 s spectra demonstrated a strong correlation between sp2/sp3 C and O/C atomic ratios with pyrolysis temperatures in pyrocarbon (Fig. 1c and Supplementary Fig. 6, 7, and Table 1). The thermal decomposition of O-containing groups at high temperature led to a decrease of O/C atomic ratio, indicating a gradually reduction in the oxidation degree from PyC500 to PyC80038,45. Interestingly, this trend contrasted to the graphitization degree of PyCs, as evidenced by the change in sp2/sp3 C atomic ratios43,46, affirming the growth of graphitic domains in pyrocarbon while depleting amorphous domains as pyrolysis temperatures increased. Moreover, the conductance (σ) of PyCs was quantified via the Quantum Design physical property measurement system (Supplementary Fig. 8). Evidently, the PyCs exhibited significantly improved σ values (from <2.0 × 10−6 to 7.0 × 103 S m−1) at higher pyrolysis temperatures, attributed to the expanding graphite domain in the carbon structure (Fig. 1d). This facilitated electron transfer within the carbon matrixes, essential in mediating the reduction of gold ions in solution47,48.Fig. 1: Physicochemical characterizations.a Scheme of the preparation process of the pyrocarbon sorbent. b TEM image of PyC700. c XPS-based sp2/sp3 carbon ratio and O/C atomic ratio of pyrocarbon. d The specific conductance of pyrocarbon determined via the Quantum Design physical property measurement system (inset shows the temperature dependence of conductivity in PyC700 and PyC800). Source data are provided as a Source Data file.Highly efficient and selective gold recovery by pyrocarbonThe performance of pyrocarbon sorbent in gold recovery was initially assessed through adsorption isotherm and kinetics experiments. Detailed results of isotherm and kinetics model fitting were provided in Supplementary Table 2, 3 and Fig. 9. Observations revealed that the gold recovery performance of PyCs follows a trend resembling a volcanic curve with increasing pyrolysis temperatures (Fig. 2a). Among its counterparts, PyC700 exhibited the highest theoretical adsorption capacity (Qm) of 2829.7 mg g−1 and mass transfer coefficient (kf) of 1.73 × 10−8 m s−1. Consequently, PyC700 displaying the most promising performance was selected as the optimal sorbent for further investigation across various experimental conditions. At initial concentrations of 10 and 50 mg L−1, PyC700 showcased rapid recovery kinetics, achieving equilibrium less than 30 minutes (Fig. 2b), with nearly 100.0% recovery efficiency of Au(III). XRD results indicated the formation of Au nanoparticles (NPs) with major (111) crystal facets (Supplementary Fig. 10)49, weaking the (002) crystal peaks of the graphite-like structure in pyrocarbon. This suggests the reductive recovery behavior of gold ions by PyC700. Of note, gold recovery capacity significantly impacts the industrial profitability of pyrocarbon-based gold recovery in the acid E-waste leachate. We evaluated the gold recovery capacities across a wide concentration range from 1 to 1000 mg L−1 (Fig. 2c), showcasing high recovery capacity of the PyC700 as high as 3000 mg g−1 when the Au(III) concentration exceeded 100 mg L−1. Within the range from 10 to 100 mg L−1, gold capacities of the PyC700 remained between 715.4 and 2346.0 mg g−1. The PyC700 also demonstrated high applicability in recycling gold from solutions with low-levels of gold content. The recovery efficiencies surpassed 92.2 and 96.8% at Au(III) concentrations of 50 and 100 μg L−1, respectively (Supplementary Fig. 11). These results highlight promising potential of PyC700 for urban gold mining from waste streams containing diverse gold contents. Temperature-dependent studies conducted at 10, 25, and 60oC revealed a significant influence on gold recovery (Supplementary Fig. 12a). Remarkably, PyC700 achieved an unprecedented Qm of 6368.4 mg g−1 at 60 oC. This exceptional capacity surpasses other carbon-based sorbents (e.g., activated carbon, carbon nanotube, graphene) and advanced nano sorbents (e.g., MOFs, COFs, and metal sulfides), regardless of whether they rely on single sorption force or combined sorption-coupled reduction mechanisms (Supplementary Table 4)3,15,18,20,50,51. Thermodynamic calculations based on Van’t Hoff equation further confirmed that the pyrocarbon-based gold recovery is an endothermic reaction, supported by the negative standard Gibbs free energy (ΔG°) and positive enthalpy (ΔH°) (Supplementary Fig. 12b and Table 5)52.Fig. 2: Gold recovery performances.a The calculated Qm and kf of pyrocarbon from Langmuir model and mass transfer model. b adsorption kinetics of Au(III) using PyC700. c The comparison of Au(III) recovery capacity of PyC700 with that of previously published sorbents. d Effects of solution acid-alkaline conditions on Au(III) recovery by PyC700. e Effects of potentially interfering species including cations and anions on Au(III) recovery by PyC700. f The calculated Kd and the comparison of redox potentials for various metal ions. g Schematic of the gold recovery from CPU leachate using PyC700. The concentration change of Au(III) using PyC700 as sorbent in h AMD leachate and i Intel leachate. j The Au(III) recovery by PyC700 in AMD and Intel leachate systems. The abbreviations “Sorp.” and “Sorp. & Red.” in Fig. 2c and d signify that the reported sorbents functionalized through simple sorption and sorption-coupled reduction processes, respectively. The relevant references for performance comparisons in Fig. 2c and d can be found in Supplementary Table 4. Error bars denote standard deviation of the experiments performed in triplicate. Source data are provided as a Source Data file.Considering the complexity of actual waste streams, we further assessed the influence of hydrochemical conditions and adsorption selectivity. The gold recovery performance of PyC700 was examined across various acid-alkaline conditions. At an Au(III) concentration of 10 mg L−1, the recovery efficiency remained above 99.5% across a broad pH range from pH 1.0 to 8.0 (Supplementary Fig. 13a), reaching as high as 99.9% in acidic HCl solutions of 0.5 and 1 mol L−1. At a higher Au(III) concentration of 1000 mg L−1, PyC700 exhibited remarkable recovery capacities exceeding 3040.4 mg g−1 within the pH range of 3.0 to 8.0 (Fig. 2d). Even in highly acidic circumstances of 0.5 and 1 mol L−1 HCl, PyC700 showcased a notable Au(III) recovery capacity exceeding 2801.0 mg g−1, highlighting its potential for gold recycling from acidic streams. Overall, PyC700 displayed significantly superior performance under wide acid-alkaline conditions compared to previously reported advanced materials8,15,53,54,55, despite limited available data in this context. Accordingly, the Au(III) species always exists as negative ionic forms (AuCl4−, Au(OH)2Cl2−, and Au(OH)4−) within the pH range of interest34,56. The observed trend in capacity variation with pH did not align with the interfacial charges of the sorbent (Supplementary Fig. 13b), emphasizing that chemical reduction plays a dominant role over electrostatic attraction. Additionally, investigations into the effects of coexisting ions (at various concentration (Supplementary Figs. 14 and 15)) on gold recovery verified the exceptional capability of PyC700 to resist interference from both typical cations (Al(III), Fe(III), Ni(II), Cu(II), Zn(II)) and anions (F−, Cl−, Br−, I−, HCO3−, SO42−, and PO33−) (Fig. 2e). The calculated distribution coefficient (Kd) value for Au(III) (Kd ~ 3.1 × 108 mL g−1) was approximately 5 to 7 orders of magnitude higher than that for competing metal ions (Fig. 2f), highlighting the remarkable selectivity of PyC700 towards gold ions. The amplitudes of potential fluctuation measured by real-time open circuit potential (OCPT) further suggested significant electron transfer from PyC700 to Au(III) at solid-liquid interfaces (Supplementary Fig. 16a)34. In contrast, no electron transfer occurred in solutions of other cations, except for Fe(III). Potential fluctuation indicated minimal electron transfer between Fe(III) and PyC700 electrode, which was further unveiled by the XRD and XPS results that the Fe(III) was reduced to Fe(II) on the sorbent surface (Supplementary Fig. 16b, c). These findings suggest a reduction potential (E0)-dependent selectivity of PyC700 towards different metal ions (Supplementary Fig. 16d). Considering various leaching agents commonly employed in hydrometallurgy57, we conducted additional investigations to assess the impacts of prevalent forms of gold complexes, such as AuBr4− and Au(CN)2−, and Au(S2O3)23−. Our findings revealed a moderate decrease in measured Qm values by 66.9% for AuBr4−, while significant decreases of 91.1% and 95.7% was observed for Au(S2O3)23− and Au(CN)2−, respectively (Supplementary Fig. 17a and Table 6). The hypothesis that adsorption behaviors depend on reduction potential effectively clarifies this phenomenon (Supplementary Fig. 17b, c). It suggests that transferring electrons to AuBr4- becomes increasingly difficult, and it’s even less probable for Au(S2O3)23− and Au(CN)2− to be reduced to metallic Au0 (Supplementary Fig. 17d). This phenomenon can also be attributed to the differences in the coordination stability constants (lgKf0) among these three complexes with Au-anions complexes (lgKf0(Au(CN)2−)> lgKf0(Au(S2O3)23−)> lgKf0(AuBr4−)> lgKf0(AuCl4−)), as higher lgKf0 value indicates greater difficulty in decomposing the ligands from Au-anions to form metallic Au-Au bonds58.Gold recovery from authentic E-Waste LeachatesThe escalating demands for gold prompted our evaluation of PyC700’s efficacy in recovering gold from real E-waste leachates. Two types of waste CPUs, Intel CPU and Advanced Micro Devices (AMD) CPU scraps, underwent treatment with two representative chemical leaching systems: aqua regia solution (Aqua) and N-bromosuccinimide (NBS)/pyridine mixed solution (Fig. 2g)20. Among four types of CPU leachates, the AMD-Aqua leachate displayed the highest gold content, with an Au(III) concentration of 1330.1 mg L−1 (Figs. 2h, 2i, and Supplementary Table 7). However, the highly acidic nature (pH <0.0) and strong oxidation characteristics of the leachate adversely affected PyC700’s gold recovery process, resulting in a recovery efficiency lower than 42.5% (Qe was 1120.7 mg g−1) (Fig. 2j). To address these challenges, we diluted the AMD-Aqua leachate by tenfold (labeled as “AMD-Aqua*10”). As a result, PyC700 reached an adsorption equilibrium less than 300 min and successfully recovered nearly 100.0% of Au(III) in AMD-Aqua*10 leachate (Supplementary Fig. 18), reducing the residual concentration to about 0.1 μg L−1, demonstrating profound recovery capabilities in practical applications. Additionally, the resultant AMD-NBS leachate maintained a neutral pH condition and contained lower gold content, with an Au(III) concentration of 650.2 mg L−1. In this case, PyC700 exhibited a high recovery efficiency exceeding 99.97%, leading to a residual Au(III) concentration of 0.132 mg L−1. In comparison, lower Au(III) concentrations were measured in the Intel scrap leachates (12.65 and 2.60 mg L−1 for Intel-Aqua and Intel-NBS, respectively). Following PyC700’s gold recovery process, the gold composition in both Intel leachate were almost entirely recovered, with a recovery efficiency exceeding 99.96%. Notably, the near-complete gold recycling from CPUs leachate using PyC700 significantly simplifies the recycling process for other valuable metals (e.g., Cu(II)) present in the remaining solution37.For the separation and purification of gold from the spent sorbent, a simple approach combining direct calcination and HCl rinsing was employed. The cost-effective raw materials and facile synthetic protocol ensure the economic viability of the entire gold recovery process, further discussed in the Techno-Economic Analysis (TEA). Gold foil primarily obtained from the calcination of the spent sorbent (Supplementary Fig. 19). Upon HCl rising, optical microscope inspection of the final recycled product revealed minuscule golden particles (Fig. 2g). Despite the coexistence of high concentration of competing metal ions (Fe(III), Cu(II), Co(II), Ni(II)) in leachates, the recycled gold exhibited exceptional purity (>99.82%, 23.96 karat), surpassing that reported in numerous studies14,37,50,57. We attribute this high purity to the non-metal nature of pyrocarbon and its remarkable selectivity. The proposed eco-friendly pyrocarbon-based technology with simplified purification processes demonstrates significant applicability in recycling gold from E-waste leachate, advancing the reutilization of nonrenewable resources.Mechanistic insights into efficient gold recoveryUnderlying mechanism of electron donation and transferPyrocarbon synthesized at varying pyrolysis temperatures manifests a distinctive volcano-shaped gold adsorption performance closely linked to its physicochemical properties. A trend analysis in Supplementary Method 6 suggested a trade-off among various pyrocarbon properties (e.g., surface functionality, porosity, electrical conductivity, and electron donating capacity) for achieving high gold adsorption efficiency. Among these properties, the reductive activity linked to electron donation plays a crucial role in converting Au(III) to Au0, maintaining a significant Au(III) concentration gradient between the solution and sorbent37. Prior researches attributed pyrocarbon’s reductive reactivity to intrinsic constituents (i.e., phenolic -OH) and semiquinone-type persistent free radicals35,43. However, electron paramagnetic resonance, Boehm titration, and electrochemical analysis revealed that neither intrinsic phenolic -OH groups nor persistent free radicals are the predominant electron donors due to their limited abundance in pristine pyrocarbon (Supplementary Fig. 20 and Table 8). Consequently, the precise mechanism underlying electron donation and transfer for gold recovery on pyrocarbon remains elusive.To elucidate the electron donation and transfer mechanism, we conducted XPS and ex-situ FTIR analyses. XPS analysis demonstrated a gradual increase in O/C atomic ratios of PyC700 after gold adsorption over sequential reaction times (0.1 to 30 min) (Fig. 3a), signifying the oxygenation of carbon networks in PyC700 during gold recovery. Deconvolution of the high-resolution C 1 s XPS spectra revealed a 57.8% decrease in sp2/sp3 C ratios in PyC700 after gold adsorption, indicating the oxidation of graphene-like structures from sp2-hybridized C (e.g., C=C) into sp3 C (e.g., C-C, C-O, and O-C O) (Supplementary Fig. 21 and Table 8). Based on the deconvoluted XPS O 1 s spectra and ex-situ FTIR results (Supplementary Fig. 22 and 23), an oxidative transformation of π conjugated aromatic structures (graphite domain) into amorphous O-alkylate moieties (amorphous domain) was observed in PyC700 during gold recovery (Supplementary Fig. 24). This suggests that the electrons used for gold reduction were donated from unsaturated sp2-hybridized C with delocalized electrons. To further explore the electron transfer behaviors, in situ FTIR analysis was employed to monitor the interfacial evolution of PyC700. As illustrated in Fig. 3b, the pristine PyC700 exhibited characteristic peaks at 1100 and 1634 cm−1, attributed to vibrations of aromatic C-O/C-O-C38 and phenolic -OH43, respectively. The peak at 3442 cm−1 corresponded to the stretching vibration of O-H bonds in the adsorbed water molecules (H2Oad)59. In situ FTIR spectra and corresponding heat mappings revealed a simultaneous increase in the peak intensity of phenolic -OH and H2Oad molecules as the reaction initiated (Fig. 3b and Supplementary Fig. 25), suggesting the water molecules adsorption and the generation of phenolic -OH over PyC700. The synchronous peak intensity changes of phenolic -OH and H2Oad throughout the gold recovery process implied that the H2Oad molecules significantly contributed as oxygen sources for the generation of phenolic -OH groups on pyrocarbon structures (i.e., hydroxylation). Control experiments further supported this speculation, revealing that the oxygen used for pyrocarbon oxygenation originated from H2O molecules in water matrices rather than dissolved oxygen (Supplementary Fig. 26). The interfacial electron states and dehydrogenation ability of hydroxyl-anchored aromatic structures were further investigated using Bader charge analyses. Other O-containing groups such as aldehyde and carboxyl were also considered for reducibility comparison. Among all counterparts, types of phenolic -OH groups located at zigzag edges and within the graphitic plane exhibit the highest propensity for proton loss, as evidenced Fig. 3c. Additionally, the distinctive Bader charge values (Δq) indicate different electron delocalization behaviors of these two types of phenolic -OH groups (Fig. 3d and Supplementary Fig. 27). The high Δq value with zigzag-edged phenolic -OH suggests high electron cloud density on the oxygen atoms, facilitating gold ion capture and reduction at this site. Conversely, the low Δq value for in-plane phenolic -OH indicates that the electron cloud density was biased towards aromatic structures from oxygen atoms, signifying the potential to transfer its electrons for gold ions reduction through the conductive carbon matrices.Fig. 3: Characterizations for mechanistic investigation.a XPS-based sp2/sp3 carbon ratio and O/C atomic ratio of PyC700 for different adsorption durations. b In situ FTIR spectra of PyC700 during gold recovery. c Bader charge analysis of different O-containing groups in pyrocarbon. (d) ELF slice maps of phenolic hydroxyl groups located at the zigzag-edge and within the graphene plane. e Schematic illustration for the gold recovery on pyrocarbon. f Schematic of the hydroxylation of aromatic structures and their oxidation transformation process. g HRTEM image of the gold NPs on pyrocarbon. h HAADF-STEM of gold NPs (EDS mapping images of Au and Cl elements were provided along with the HAADF-STEM image). i Au L3-edge XANES spectra of Au-loaded PyC700. j FT of k2-weighted EXAFS spectra of Au-loaded PyC700. k FT of k2-weighted EXAFS spectra of Au-loaded PyC700 with different reaction durations. l The evolution of CNs of Au-Cl and Au-Au interactions in Au-loaded PyC700 with different reaction durations. Source data are provided as a Source Data file.Based on the preceding results, we proposed a plausible mechanism concerning the electron donation and transfer on pyrocarbon, outlined in Fig. 3e: (i) A limited number of electrons are transferred from intrinsic phenolic -OH groups on pyrocarbon to reduce the adsorbed gold ions to Au0. (ii) As the electron supply diminishes, gold ions continue to be adsorbed on the solid-liquid interfaces, causing a significant electron deficiency due to accumulated excess gold ions. (iii) Surface-sorbed H2O molecules undergo dissociation into hydroxyl ions and protons catalyzed by Au0 NPs as reported in previous studies60,61,62. Subsequently, hydroxyl ions induce the hydroxylation of aromatic structures, generating phenolic -OH groups on unsaturated sp2-hybridized C (at the edge or inside the plane of graphitic domains) (Fig. 3f). (iv) Both the hydroxylation process and the oxidation of phenolic -OH contribute electrons to the gold recovery, releasing protons into solution systems (Supplementary Fig. 28). These contributed electrons facilitate the reduction of gold ions on conductive carbon matrices, consequently oxidizing pyrocarbon containing phenolic -OH groups into highly oxidized carboxyl, quinone, or alkyl ether moieties.Understanding the gold reduction mechanismTo gain a comprehensive understanding of the gold recovery mechanism, we further investigated interfacial gold reduction processes over pyrocarbon. Time-resolved SEM images showed the primary formation of gold seeds with diameters around 20 nm on the pyrocarbon surface (Supplementary Fig. 29). Notably, these gold seeds exhibit a significantly high surface free energy34,63, rendering them inherently unstable in a state of boiling that facilitates the consumption of Au(III) salt. Simultaneously, the enhanced PyC700 conductivity further promotes efficient electron transfer for subsequent reduction of gold ions (Supplementary Fig. 30). As gold salt is consumed at localized adsorption domains, bare Au0 NPs continuously attract gold ions from solution for self-growth. This phenomenon can be perceived as an autocatalytic surface growth process that promotes overall gold recovery process on pyrocarbon64,65. By following Ostwald ripening rules66, Au0 NPs gradually increase in size until reaching approximately 200 nm, following which NPs coalescence leads to the formation of aggregates of Au0 NPs. The HRTEM image (Fig. 3g) revealed the presence of (111) crystal planes in Au0 NPs, exhibiting an interlayer spacing of 0.23 nm, consistent with the XRD results shown in Supplementary Fig. 10. TEM images revealed the presence of amorphous structures surrounding the newly formed Au0 NPs (Supplementary Fig. 31), which disappeared after a 5-minute adsorption process. High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) and element mapping images confirmed a close proximity between Au and Cl elements (Fig. 3h), with Cl extending more towards the periphery of the amorphous structures. As the gold reduction reaction progressed, mapping images of larger Au0 NPs exhibited complete overlap between these two elements (Supplementary Fig. 32). These observations strongly indicate the existence of gold-chlorine complexes that envelop gold nuclei during the gold reduction process.To elucidate the gold reduction mechanism at atomic level, we explored the time-resolved evolutions of chemical state and local coordination environments of Au in Au-loaded PyC700 using Au L3-edge X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) spectra with three reference standards. The XANES spectra (Fig. 3i) indicated a higher intensity of normalized white line peak (11921 eV) of the Au-loaded PyC700 compared with that of Au foil67, suggesting the existence of unreduced Aux+ (x = 0–3) species in the initially nucleated Au0 NPs. Moreover, the white line peak intensities in Au-loaded PyC700 eventually became almost identical to that observed for Au foil, showcasing the complete reduction of gold ions. Linear combination fitting (LCF) of the Au-loaded samples was conducted to gain deeper insights into the changes observed in XANES spectra (Supplementary Fig. 33). LCF analysis confirmed that a portion of Au in the Au-loaded PyC700 existed in a cationic state during the initial stage68, with the average oxidation state of Au(+0.33) (referenced to Au3Cl standards) constituting for 17.3 and 7.9% at 0.1 and 1 min, respectively. As the reaction progressed, no contribution from Au(+0.33) could be detected due to low proportion of unreduced Aux+ (x = 0–3) on the large-sized Au NPs. Fourier-transformed (FT) EXAFS spectroscopy was employed to identify the localized structure of Au within short range (<4 Å). The curve fitting results for the samples were displayed in Supplementary Fig. 34, Fig. 35 and Table 9. The R space spectrum of Au-loaded PyC700 (Fig. 3j) revealed distinct peaks corresponding to Au-Cl and Au-Au scattering at distances approximately 2.27 and 2.86 Å, respectively, referenced by Au foil and Au3Cl standards69. These findings were consistent with time-resolved FT-EXAFS spectra and coordination number (CN) analysis of Au-loaded PyC700, indicating a decrease in the proportion of Au-Cl interactions and an increase in Au-Au interactions within local gold coordination environments during gold recovery (Figs. 3k, l). Consequently, XANES and LCF analyses combined with EXAFS findings, collectively confirmed the presence of gold-chlorine clusters during initial stages of gold nucleation and growth70. Rapid growth of Au NPs to large size leads to negligible presence of Au-Cl interactions among gold coordination structures.To comprehend the adsorption and reduction processes of gold ions by pyrocarbon, spin-polarized density functional theory (DFT) calculations were performed. Constructed graphene supercells functionalized with hydroxyl groups were employed to evaluate the adsorption energy (Ead) of HAuCl4 molecules, which were sequentially adsorbed at hydroxyl sites on the graphene structure. The entire gold reaction pathway and its corresponding free energy diagram were summarized in Fig. 4a, b, with detailed step-by-step descriptions (i.e., adsorption and dechlorination process) available in the Supplementary Information. Specifically, the DFT calculations unveiled a favorable adsorption of HAuCl4 molecules onto pyrocarbon (referred to as “G*HAuCl4”) with an Ead of −1.45 eV (Supplementary Fig. 36). The adsorbed HAuCl4 subsequently underwent successive dechlorination reactions, releasing three HCl to create the G*AuCl intermediate, wherein the Au atom directly bonded to an O atom within hydroxyl group. Following the classical nucleation theory, recent studies suggested that direct detachment of all four Cl atoms from HAuCl4 leads to the formation of Au014,53. Our findings indicated a positively unfavorable Ead value (1.08 eV) for this process on pyrocarbon, suggesting the presence of a high energy barrier during the detachment of the last Cl atom from HAuCl4 to form Au0 nuclei. This elevated energy barrier has been similarly documented in previous studies regarding the dechlorination from Au-Cl intermediates67. Instead, we found that the adsorption of an additional AuCl3 on the G*AuCl intermediate was found to be energetically favorable with an Ead of −2.39 eV (AuCl3 was used instead of HAuCl4 in subsequent calculations due to facile HCl detachment from HAuCl4). Subsequent sequential adsorption processes of AuCl3 molecules onto G*AuiClj intermediates, accompanied by stepwise dichlorination, occurred with continuous exothermic behaviors (Supplementary Fig. 37 to 49). These processes continued until the formation of Au0, which were bound with six other Au atoms within the Au15Cl14 clusters (G*Au15Cl14) (Supplementary Fig. 50). This specific structure is termed a ligand-protected metal cluster, aligned with the ‘divide-and-protect’ concept, theoretically displaying lower energy level compared to typical Au-Cl linkages71. Based on DFT calculations, the reduction of gold ions is proposed to occur through the stepwise nucleation of chlorine-protected gold clusters, which is crucial in gradually lowering the energy barriers for the entire reduction processes. These findings elucidate the experimentally observed inverse trend of CNs for Au-Cl and Au-Au paths, as well as the homogeneous distribution of Au and Cl elements on the pyrocarbon. In summary, we systematically unraveled interfacial behaviors involving electron donation and transfer, as well as gold nucleation and gradual growth on pyrocarbon surfaces, as summarized in the proposed gold reduction mechanisms in Fig. 4c.Fig. 4: Theoretical calculations for mechanistic investigation.a Schematic illustration of the proposed mechanism of gold reduction processes on pyrocarbon. b The free energy diagram of the proposed gold-chlorine intermediates in gold reduction processes. c Overview of the intrinsic principle of selective and efficient gold recovery performance of pyrocarbon-based sorption technique.

Hot Topics

Related Articles