Iodine‐Catalysed Dissolution of Elemental Gold in Ethanol

Abstract Gold is a scarce element in the Earth's crust but indispensable in modern electronic devices. New, sustainable methods of gold recycling are essential to meet the growing eco‐social demand of gold. Here, we describe a simple, inexpensive, and environmentally benign dissolution of gold under mild conditions. Gold dissolves quantitatively in ethanol using 2‐mercaptobenzimidazole as a ligand in the presence of a catalytic amount of iodine. Mechanistically, the dissolution of gold begins when I2 oxidizes Au0 and forms a [AuII2]− species, which undergoes subsequent ligand‐exchange reactions and forms a stable bis‐ligand AuI complex. H2O2 oxidizes free iodide and regenerated I2 returns back to the catalytic cycle. Addition of a reductant to the reaction mixture precipitates gold quantitatively and partially regenerates the ligand. We anticipate our work will open a new pathway to more sustainable metal recycling with the utilization of just catalytic amounts of reagents and green solvents.


Materials and methods
All chemicals were bought from commercial sources and used without further purification. Au powder (1.5-3.0 μm spherical, 99.9 %) was purchased from Strem Chemicals, 2-mercaptobenzimidazole (2-MBI), 4-pyridinethiol (4-PS), NaBH4, K(AuCl4) (98%) and iodine from Sigma Aldrich and 33% aqueous H2O2 from VWR International. All solvents were HPLC grade. Flame atomic absorption spectroscopy (FAAS) measurements were performed on a Perkin-Elmer 3030 atomic absorption spectrophotometer. Measurements were carried out in air/acetylene flame and by using Au hollow cathode lamp (HCL) at wavelength of 248 nm with lamp current of 10 mA. High-resolution electrospray-ionization mass spectra (ESI-HRMS) were recorded with a Bruker microTOF mass spectrometer in a positive and negative ion mode using sodium formate as a calibrant. An Oxford INCA 350 energy-dispersive X-ray microanalysis system connected with a Hitachi S-4800 field emission scanning electron microscope (FESEM) was used for the energy-dispersive X-ray spectrometry (EDS) measurements. All NMR spectra were recorded in DMSO-d6 with a Varian Mercury 400 instrument (at 400 MHz) using Me4Si as an internal standard. Chemical shifts are reported in ppm (δ) relative to central lines of DMSO-d6 for 1 H NMR (δ = 2.50 ppm) and 13 C NMR (δ = 39.52 ppm).

Optimized procedure for Au dissolution
Au powder (2 mg, 0.01 mmol, 1.5-3.0 μm spherical particles) and 2-MBI (20 eq., 30 mg, 0.2 mmol) were weighted into a 25 mL glass vial equipped with oval magnetic stirring bar. 10 mL of EtOH were added and the mixture was stirred until all 2-MBI was dissolved. Then the reaction was charged with 51.2 μL of freshly prepared 19.5 mM EtOH solution of I2 (10 mol %, 0.254 mg, 0.001 mmol) and 33% aqueous H2O2 (20 eq., 19 μL, 0.2 mmol). Reaction vial was tightly closed with a plastic cap, stirred, and submerged into a preheated oil bath at 60 °C. After 13 h of vigorous stirring, 100% dissolution of Au was reached.

Scale-up procedure for Au dissolution
Au powder (20 mg, 0.1 mmol, 1.5-3.0 μm spherical particles) and 2-MBI (20 eq., 300.4 mg, 2 mmol) were weighted into a 250 mL round bottom flask equipped with magnetic stirring bar. The majority of EtOH (100 mL in total) was added into the flask to dissolve 2-MBI. Then, I2 (10 mol %, 2.5 mg, 0.01 mmol) was dissolved in residual EtOH and added to the reaction mixture followed by addition of 33% aqueous H2O2 (20 eq., 186 μL, 2 mmol). Flask was tightly sealed with a plastic cap, stirred and submerged into a preheated oil bath at 60 °C. All Au was dissolved after 13 h of vigorous stirring.

FAAS measurement
The amount of dissolved Au was determined by flame atomic absorption spectroscopy (FAAS) measurements that were performed with a Perkin-Elmer 3030 atomic absorption spectrophotometer. Measurements were carried out in air/acetylene flame and by using Au hollow cathode lamp (HCL) at wavelength of 248 nm with lamp current of 10 mA. Calibration curve ( Figure S1) was prepared from stock solution of potassium tetrachloroaurate(III) in EtOH with concentrations of 2, 4, 6, 8 and 10 mg/L. Sample was taken from reaction solution (300 μL) and diluted with EtOH (9.5 mL, total sample volume 9.8 mL). Each dissolution experiment was repeated more than once to ensure the dissolution consistency. To minimize systematic error of sample preparation, reference calibration curves and reference samples were measured periodically.

Optimization of reaction parameters
Table S1 summarizes the optimization results for 4-PS assisted dissolution of Au, including the amount of 4-PS, H2O2, and I2, solvent and the reaction temperature. Reaction setup and the order of reagent addition was similar to the procedure with optimized conditions described above (Section 2.1.). As seen from Table S1, 4-PS is unable to substitute iodide in formed [AuI2] -, since the sum of dissolved Au (%) in experiments excluding H2O2 (entries 2, 4, 7 and 10) or I2 (entries 1, 6 and 9) is the same or more than in experiment with all three components, ligand, oxidant and catalyst, respectively (entries 3, 5, 8 and 11). Similarly, Table S2 shows results for 2-MBI assisted dissolution of Au in DMF, including the amount of 2-MBI, H2O2, and I2, and the reaction temperature. Reaction setup and the order of reagent addition was similar to the procedure with optimized conditions described above (Section 2.2). As seen from Table S2  Other solvents than DMF were tested as potential reaction media for Au dissolution (Table S3). In some solvents, a precipitation was formed, which made it impossible to detect Au in solution with FAAS. The precipitation was also formed in EtOH but became soluble in larger amounts of the solvent. Therefore, dissolution in EtOH was further studied (Table S4). As Table S4 shows, precipitation formed when using 5 mL of EtOH even at room temperature (Table S4, entry 2) or with lower amount of ligand at elevated temperature (Table  S4, entry 3). Yet, near quantitative dissolution of Au was achieved in 10 mL of EtOH (Table S4, entry 4).   In Table S5, a drop in dissolution efficiency is detected when 96% EtOH is used instead of pure EtOH. Cooperation between 2-MBI, H2O2 and I2 is suppressed since the sum of Au dissolution percentages in experiments excluding I2 (Table S5, entry 2), H2O2 (Table  S5, entry 3) and 2-MBI (Table S5, entry 4) is more than in experiment with all three components (Table S5, entry 1).  Table S6 summarizes the final optimization of the reaction parameters using pure EtOH. Dissolution condition with adequately minimal amounts of reagents with acceptable dissolution efficiency was chosen for further studies (Table S6, entry 1). As it was proven later, 100% dissolution was achieved after 13 h (Table S8 and Figure S2). The sum of Au dissolution percentages in experiments excluding I2 (Table S6, entry 2 or Table S7, entry 2), H2O2 (Table S6, entry 3 or Table S7, entry 3) and 2-MBI (Table S6, entry 4 or Table S7, entry 4) is less than in experiment with all three components (Table S6, entry 1 or Table S7, entry 1) at 21 h (Table S6) or at 13 h (Table  S7) what is a proof of cooperation between all three reagents.

Au dissolution vs. time
20 samples were taken between 5-180 minutes in the course of 24 h dissolution reaction. Maximum two samples were taken from each reaction mixture to minimize the error arising from reducing the volume. Dissolved Au vs. time graph was constructed from the acquired data (Table S8). As seen from Table S8 and Figure S2, 100% dissolution was achieved after 13 h.   Figure S3. The tangential lines were drawn to visualize the observed predominant dissolution rates. Au dissolution vs. time

Colour of the reaction mixture
Photographs of the reaction were taken according to the dissolution curve ( Figure S2) and predominant reaction rates ( Figure S3). The photographs in Figure S4 show the colour changes. At the beginning of the reaction (0 min), the solution exhibits brownish iodine colour, which is quickly fading away (30 min) and then completely disappears (2.5 h). Reaction mixture remains colourless till the end of the observation period (24 h). The same applies for a scale-up reaction procedure as seen from photographs represented in Figure  S5. Figure S4: Photographs of the reaction mixture at specific times.

Sample preparation
High-resolution electrospray-ionization mass spectra (ESI-HRMS) were recorded with a Bruker microTOF mass spectrometer in a positive and negative ion mode using sodium formate as a calibrant. Samples were prepared by taking 20 μL of reaction mixture and diluting with 780 μL of 0.05% aqueous formic acid in MQ water and MeOH mixture (70/30 v/v). Samples were filtered through 0.22 μm PTFE syringe filters prior to measurement. Syringe filters were washed with MeOH before filtration of the sample. Samples were measured immediately to avoid partial degradation and/or precipitation.

Found species
Experimental isotopic patterns for selected species were compared to calculated patterns. Species found are illustrated in Figure S6.

Intensities for selected species vs. time
Graphs of intensities vs. time were plotted for selected species from acquired ESI-HRMS data. All samples were taken from the same experiment. Species 1, 2 and 3 with corresponding m/z values of 451, 473 and 495, respectively, were followed for the first three hours of the reaction ( Figure S23). Samples were prepared three times more concentrated than outlined in Section 5.1. Figure S23: Relative intensities for m/z 451, 473 and 495 (negative ion mode) vs. time. The highest intensity was set to 100% and other data points were adjusted accordingly.

Procedure
After quantitative dissolution of Au with the scale-up procedure (20 mg of Au, see Section 2.2), reaction mixture was placed in an ice bath to cool for 15 min. Then, NaBH4 (151.3 mg, 4 mmol) was slowly added during a 40 min period. Reaction mixture was left at 0°C for an additional 5 min before stirring vigorously at room temperature for 4 h. Formation of black precipitate was noted. Next, water (30 mL) was added, and the reaction mixture was left to stir. Previously precipitated black particles coagulated to form black flakes, which were then collected by filtration by using Büchner funnel. Precipitate was washed in the following order: with water, H2SO4 (aq), water, NaOH (aq), water, distilled water, EtOH and finally with Et2O. The filtrate was kept for 2-MBI ligand recycling. The precipitate was dried under reduced pressure (vacuum pump) to afford 18.4 mg of black powder, later proven to be elemental Au by FESEM-EDS analysis (yield=92%). Flask remained loosely closed with plastic stopper throughout the whole reduction process. Solvents from EtOH/water filtrate were removed under reduced pressure and 10% aqueous HCl was added to quench the residual NaBH4. Formed precipitate was collected by filtration and washed with water, distilled water and Et2O, respectively, to afford 123.9 mg (yield=41%) of white crystalline material characterized by NMR to be pure 2-MBI [1] (Figures S24 and S25).

FESEM-EDS analysis
An Oxford INCA 350 energy-dispersive X-ray microanalysis system connected with a Hitachi S-4800 field emission scanning electron microscope (FESEM) was used for the energy-dispersive X-ray spectrometry (EDS) measurements. Au sample was washed and dried before analysis as described above. As seen from Figure S26, precipitate acquired was pure Au with particle size 10-20 nm in diameter.

Figure S26
: FESEM-EDS analysis of precipitated Au powder. SEM image of Au powder (above, left). Au mapping for pink square area (above, right). Au mapping spectrum (below, left). Zoomed SEM image of Au particle (below, right).

NMR study of reaction mixture before and after reduction
1 H NMR experiments were conducted to investigate existing species before and after reduction. Sample was taken from scale-up reaction after dissolution. Solvent from sample was removed under reduced pressure and residue was dissolved in DMSO-d6 for 1 H NMR analysis ( Figure S27). After reduction of scale-up reaction with NaBH4, solvent was evaporated under reduced pressure and residue was dissolved in EtOAc. Solution was transported to separating funnel and saturated NaCl (aq) was added. Mixture was extracted 4 times with EtOAc, organic fractions combined and dried over anhydrous MgSO4 before EtOAc was removed under reduced pressure. Sample was again dissolved in DMSO-d6 for 1 H NMR analysis ( Figure S28). As seen from Figure S27, after Au dissolution the 1 H NMR peaks can be assigned to 2-MBI 1 , thioether 6 (δ 7.18 -7.20, 7.58 -7.60) [2] and disulphide 5 (broad peaks at δ 7.33 and 7.76) [3] . Integral values suggest that another species similar to 5 and 6 is present in the reaction mixture -broad peaks at δ 7.33 and 7.59 could be attributed to trisulfide 7. Mentioned peaks disappear after the reduction, which can be another proof of 7 as well as conformation from ESI-HRMS.    As seen from Figure S29, peaks assigned to disulphide 5 and trisulphide 7 disappear after reduction. 5 and 7 are reduced to 2-MBI, whereas partial transformation to thioether 6 takes place as noted from comparing integrals between spectra depicted in Figure S27 and Figure S28.

Computational details
All calculations were performed using ORCA 5.0. [4] Structures were optimized using the TPSS [5] functional with def2-TZVP basis set [6] and DFT with standard integration grids. Weak interactions were accounted for using the D3 dispersion correction with Becke-Johnson damping [7] . Solvation effects were accounted using the conduction conductor-like polarizable continuum model, the CPCM solvation model [8] with 24.3 (ethanol) dielectric constant. Thermal corrections at 60°C were obtained by calculating harmonic vibrational frequencies for all structures at the TPSS-D3/def2-TZVP level, and chemical potentials (c.p.) were obtained using the quasi-rigid rotor harmonic oscillator (quasi-RRHO) approach proposed by Grimme. [9] The quasi-RRHO approach uses the free-rotor entropy for all modes with frequencies below 35 cm-1, while the standard RRHO approach is used for other modes. In addition, the harmonic vibrational frequencies have been scaled by a factor of 0.9914. The Gibbs free energies are then obtained as summation of the zero point energy and chemical potentials.

Calculation of G for substitution reactions
The change of Gibbs free energy (G) for substitution reactions from 1 to 2 and from 2 to 3 were determined as difference between Gibbs energies of products and reactants ( Figure S30). When Iand I2 are both present in the solution, I3is readily formed with a large stability constant in ethanol. I3was considered when calculating G for substitution reactions when I2 concentration was high.

Calculated energies and cartesian coordinates for different species
All structures are optimized at TPSS-D3/def2-TZVP level. Energies are in Hartrees.
Species I -: Number