Some insights into the chemistry of gold adsorption by thiol and amine functionalized mesoporous silica in simulated thiosulfate system

Hydrometallurgy 156 (2015) 28–39

Contents lists available at ScienceDirect

Hydrometallurgy journal homepage: www.elsevier.com/locate/hydromet

Some insights into the chemistry of gold adsorption by thiol and amine functionalized mesoporous silica in simulated thiosulfate system Babak Fotoohi ⁎, Louis Mercier Centre in Mining Materials Research (CIMMR), Laurentian University, 935 Ramsey Lake Rd., Sudbury, ON P3E 2C6, Canada

a r t i c l e

i n f o

Article history: Received 6 March 2015 Received in revised form 21 May 2015 Accepted 22 May 2015 Available online 27 May 2015 Keywords: Mesoporous silica Thiol groups Amine groups Gold thiosulfate Gold hydrometallurgy Adsorption

a b s t r a c t Organic functional groups such as amines and thiols have been widely used to functionalize pore wall surfaces of mesoporous silicas for a variety of applications relying on adsorption of metals. One of the promising applications of such hybrid materials is in hydrometallurgical separation of metals in a leach solution. Toward this aim, mesoporous MSU silica was prepared and functionalized with short-chain thiol and primary amine organic groups. Adsorption of gold from simulated thiosulfate leach solutions was investigated with respect to the adsorption chemistry. Adsorption capacity of thiol-bearing silica was well showing the monolayer behavior approaching an ultimate metal:ligand = 1:1 capacity. Higher capacities were obtained in either alkaline condition (pH ~ 10.5) or at higher oxidation potential (Eh ~ 300 mV) of the solution. Adsorption capacity of amine groups was to some extent proportional to their surface molar concentration, but it was also beyond the monolayer capacity of chemisorption, besides not being explicable by hard–soft acid–base theory. Despite considerable gold recovery by amine functional groups, gold was not detected in XPS analysis unless adsorption was from pure solutions. Partial electrochemical reduction to metallic gold and formation of dispersed tiny nanoparticles during interaction of gold thiosulfate species in solution with amine ligands were proposed. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Since the advent of thiosulfate leaching of natural gold ores and its introduction as significantly safer and less hazardous alternative to conventional cyanidation extraction, the chemistry of leach solution has been under study from a wide range of aspects. It has long been known that chemistry of gold leaching in thiosulfate system is complicated due to the co-presence of multiple complexing agents and redox couples [1]. Aside from thiosulfate anion as the main leaching agent for gold, other reagents such as ammonium/ammonia and oxidation catalysts (usually copper(II)) are added into the system which impart their specific functions. Inevitable presence of thiosulfate oxidation products (e.g., polythionates or elemental sulfur) are among other complicating factors of the chemistry in such leaching systems. Despite the reaction complexity, gold thiosulfate leaching has been studied with respect to variety of chemical, electrochemical, thermodynamic and kinetic aspects in the presence of principal as well as contaminant species in solution. Spectrochemical and electrochemical studies on characterization of surface species formed during leaching of metallic gold [2–5] as well as the effect of different sources of reagents and ionic species in solution [6–10] are among the noticeable research published mainly in the past decade. ⁎ Corresponding author at: Department of Chemistry and Biochemistry, Faculty of Science and Engineering, 935 Ramsey Lake Rd., Sudbury, ON P3E 2C6, Canada. E-mail address: [email protected] (B. Fotoohi).

http://dx.doi.org/10.1016/j.hydromet.2015.05.010 0304-386X/© 2015 Elsevier B.V. All rights reserved.

In a routine hydrometallurgical separation and extraction there comes a stage after leaching wherein the loaded (so-called pregnant) solution would be subject to “concentration” process before being fed into the electrochemical metal extraction system. Concentration of gold thiosulfate leach solution has been investigated through applying different methods among which adsorption and ion-exchange are noticeable. Despite relatively extensive study of gold thiosulfate leach concentration by anion-exchange resins and recent success toward commercialization of the process [11], there has been less work on other adsorption methods especially based on chemisorption by ligand-functionalized materials. Significance of directing attention toward such adsorption mechanism arises from the unsuccessful physisorption of gold thiosulfate compound on bare activated carbon surfaces in spite of the latter being the unique adsorbent in commercial gold cyanide hydrometallurgy. Moreover, ion-exchange resin would not be the perfect adsorbent due to inherent drawbacks including non-rigid organic framework, insufficient selectivity and potential for deactivation (poisoning) of the exchange sites by contaminant/competitor species in solution. On the other hand, organically-functionalized silica gel and relatively novel mesoporous silica with either or both of amine and thiol groups immobilized onto their pore wall surfaces have successfully recovered gold(III) chloride from principally acidic solutions [12–15]. But, it was only recently that hybrid mesoporous silica was found to be efficient in adsorbing gold(I) thiosulfate from copper and ammonia-bearing solutions [16]. In a most recent research by the same authors [17], gold thiosulfate adsorption by both amine and thiol-functionalized

B. Fotoohi, L. Mercier / Hydrometallurgy 156 (2015) 28–39

mesoporous silicas was studied in detail with respect to adsorption capacity, kinetics and reagent concentrations for the first time. In that research, several questions regarding the effectiveness of both thiol and amine groups in selective adsorption of gold (complex) were aroused. It has long been known that the bonding mechanism of ionic gold species with sulfur bearing compounds such as thiosulfate [18] is best explained according to the hard–soft acid–base (HSAB) theory [19] in which gold(I) as a soft Lewis acid bonds with sulfur as a soft Lewis base (donor). Same principle is applicable when gold nanoparticles are coated with organic sulfur groups (e.g., thiols, xanthates and disulfides) in the so-called self-assembled monolayers (SAMs) [20,21]. However, with regard to interaction of hard donor atoms (such as –NH2 or NH3) with metal ions of low valency (such as Au+), chemisorption through (covalent) complexation based on HSAB theory is less predictable and the compounds are regarded as unstable. Most of the published research on adsorption of gold by amine (or polyamine) functional groups have been in the acid environment wherein much of the surface amine groups were expected to become protonated, so that physically (electrostatically) attracting the negatively charged gold compounds such as [AuCl4]− or even [Au(S2O3)2]3−. On the other hand, in a thiosulfate leach solution with a typical higher than neutral pH, such electrostatic interaction could be remote from happening if not completely impossible. Since in the previous research [17] amine functional groups adsorbed high amounts of gold in the simulated leach solutions having a pH ~ 7 or higher (pH ~ 10), there was an incentive to find out more details on such adsorption system with respect to both gold speciation and surface functional groups. Therefore in this research, structure and surface chemistry characterizations of the hybrid mesoporous silica exposed to different simulated gold thiosulfate solutions were performed. Behavior of different functional groups on the adsorbent material in different solution chemistries and pH/Eh conditions was also investigated. 2. Materials and Methods 2.1. Chemicals Nonionic surfactants Igepal CO-720 and Pluronic® P123 as structure directing agent, functional organosilanes (3-aminopropyl) triethoxysilane (APTES), ≥ 98%A and (3-Mercaptopropyl) trimethoxysilane (MPTMS), 95% as well as swelling agent 1,3,5Trimethylbenzene (TMB) were all provided by Sigma-Aldrich Co., St. Louis, MO, U.S.A. Silica precursor was tetraethyl orthosilicate (TEOS) from Gelest Inc., Morrisville, PA, U.S.A. Sodium fluoride (reagent A.C.S.) was provided by Matheson Coleman & Bell manufacturing chemists, East Rutherford, NJ, U.S.A. In gold adsorption tests, sodium aurothiosulfate(I) purchased from Surepure Chemicals Inc., Florham park, NJ, U.S.A. was used as source of gold(I). Ammonium thiosulfate and Ammonium nitrate, A.C.S. reagent ≥ 98% were provided by SigmaAldrich Co., St. Louis, MO, U.S.A. Cupric carbonate (Baker analyzed reagent) was from J.T. Baker Chemical Co., Phillipsburg, NJ. U.S.A. All other chemicals (acid, base, solvents and hydrogen peroxide) were of reagent grade. 2.2. Adsorbent preparation Mesoporous silica of type MSU-2 and MSU-3 (henceforth in this research, without the “-” in the original nomenclature) was prepared through a two-step sol–gel synthesis by using nonionic poly(ethylene oxide)-based surfactants [22,23]. In a typical synthesis, 0.02 M surfactant solution (15 g/L for MSU3 synthesis) was magnetically stirred in DI water acidified with dilute HCl to pH ≈ 2–2.5 (pHiep of SiO2). In MSU2 preparation swelling agent (TMB) was added in molar TMB/ surfactant = 1 for the expansion of pore size. After a homogenous solution was obtained, silica source (TEOS) was added (in MSU2 synthesis: molar TEOS/surfactant = 8, and in MSU3 synthesis: weight TEOS/

29

surfactant = 2.22). At this step, for thiol functionalization through cocondensation synthesis, mercapto-silane could also be added together with the silica precursor. Upon completion of hydrolysis (in about an hour) at ambient temperatures, sodium fluoride (molar NaF/Si = 0.04) was introduced to the solution to trigger nucleation. The MSU2 synthesis batch was then transferred to thermostated water bath for condensation at elevated (~55 °C) temperature overnight. The as-obtained (filtered and dried) silica was solvent extracted with hot ethanol in a Soxhlet apparatus (48–72 h) to remove the structure directing agent. In case of amine functionalization, the pristine silica material from the sol–gel synthesis (after solvent extraction) was treated with dry toluene in a boiling reflux system. The reason for using the grafting method for amine functionalization was the acidic medium of the mesoporous silica synthesis in which protonated amines could interfere with the formation of the desired mesoporous framework. 2.3. Gold adsorption tests Adsorption from gold thiosulfate solutions was performed in Erlenmeyer flasks containing 50 mL of test solution and 200 mg/L adsorbent loading. Concentration of the main reagents for a typical simulated thiosulfate leach solution in gold hydrometallurgy was previously determined by the same authors [17]. This included addition of appropriate weights of gold thiosulfate salt to the desired concentration. While ammonium thiosulfate salt was providing 0.1 M thiosulfate concentration, supplementary ammonium/ammonia was added form nitrate salt (or ammonia solution) to make up for a total ammonia concentration of 1.0 M. Copper carbonate hydrate provided 2.5 mM copper(II) in the solution. All reagents were dissolved in DI water. After complete dissolution of all components (usually in a matter of few minutes) a clear colorless solution or a clear blue solution was obtained in neutral pH or alkaline pH systems respectively. In obtaining gold adsorption isotherms for the mesoporous adsorbent, the initial gold concentration in solution was varied within 2–200 mg/L range. After about 24 h orbital shaking (at 150 rpm) of the top-sealed flasks at ambient conditions, samples were taken accurately from the supernatant clear solution given enough sedimentation time in stand-still. The aliquots were then analyzed (after dilution if N 50 mg/L gold was dissolved initially) for the gold remaining in equilibrium solution using a Perkin Elmer AAnalyst 400 atomic absorption spectrometer with a gold lamp (10 mA operating current) and gold chloride standard solutions (1005 mg/L in 5% HCl) for calibration. Adsorption capacities and recoveries were measured using the following formula: Adsorption capacity of the adsorbent Ca ¼ ½ðci −ce ÞðVÞ=½196:97ðmÞ

Precious metal adsorption or removal recovery R% ¼ ½ðci −ce Þ=ci   100

wherein “Ca” (adsorption capacity) is in [mmol/g], “ci” and “ce” are the initial and remaining gold concentrations respectively in [mg/L], V is the test volume (0.050 L), “196.97” was considered as gold molar atomic mass in [g/mol] and “m” is the adsorbent weight (0.010 g). 2.4. Characterizations Determination of the equilibrium pH and mixed redox potential were conducted at room temperature with a Hanna HI2221 pH/ORP meter. Combination electrodes were used to measure either pH or the electrochemical potential in which Ag/AgCl in 3.5 M KCl was the reference electrolyte. Platinum was the sensing electrode in the ORP measurements. The following relationship was used to report the measured mixed potential in equilibrium solution with respect to the

30

B. Fotoohi, L. Mercier / Hydrometallurgy 156 (2015) 28–39

Table 1 Structural and chemical properties of hybrid mesoporous silica adsorbents. Sample code

MSU2 MSU2-AP MSU2-MP MSU3-AP

N2 adsorption data

Porosity (%)

Specific surface area (m2/g)

Total pore volume (cm3/g)

Pore size (nm)

1238 636 1248 602

1.43 0.64 1.26 0.94

4.4 3.7 3.7 10.8

76 58 73 61

standard hydrogen electrode (SHE): Potential with respect to SHEðEh Þ ¼ emf ðmeasured potentialÞ þ Ere f

in which Eref = +205 mV for Ag/AgCl in 3.5 M KCl, at 25 °C. BET specific surface area for pristine and hybrid mesoporous silica structures were measured by N2 adsorption at 77 K using Micromeritics® ASAP 2010 apparatus. Pore size distribution was determined using the integrated DFT Plus® software in which Broekhoff de Boer model for cylindrical pores was employed for pore size distribution analysis and pore volume calculation. Porosity (%) was calculated based on the obtained total mesopore volume, assuming 2.2 g/cm3 [24] as true density of amorphous silica framework and using the following calculations:
Porosity % ¼ Vp =Vt  100 in which Vp is pore volume (cm3) = [total pore volume (cm3/g) from BdB model calculations] × [m, degassed sample weight (g) in N2 sorption test], Vt is total sample volume = Vp + Vf and Vf is skeletal or framework volume (cm3) = m/2.2 Thermogravimetry (TGA-DSC) analyses were performed using TA Instruments SDT-Q600 machine. Samples were heated up to 1000 °C (10 °C/min) at two different atmospheres (nitrogen, b 200 °C, and air, N200 °C). A Bruker ALPHA FTIR transmission-based spectrometer was used for investigation of the infrared absorption by samples pressed with KBr into pellets. Analysis was performed for pristine and hybrid silica before adsorption and for hybrid silica after gold adsorption. For further comparability of IR absorption patterns and intensities among different adsorbents, constant weights of sample and KBr were mixed in all preparations. Pharmacia Biotech Ultrospec 3000 UV/Visible spectrophotometer was used in the wavescan mode to obtain UV–vis spectra for liquid samples in quartz cuvette cells with DI water as blank. X-ray photoelectron spectroscopy of the adsorbents was carried out for the material before and after adsorption tests. The as-obtained powder filtered from equilibrium adsorption solutions and dried in air was submitted for analysis. For this purpose an AXIS Ultra XPS machine

TGA weight loss (%)

Analysis (dry wt.%)

Up to 200 °C (moisture)

200–500 °C (organics)

500–800 °C (surface –OH)

Total carbon

Nitrogen

Sulfur

6.2 9.3 7.0 2.5

3.3 7.5 7.8 8.0

3.5 5.1 3.3 3.6

3.5 8.6 5.6 10.0

– 2.9 – 3.1

– – 2.8 –

with a hemispherical analyzer by Kratos Analytical was used. Aluminum Kα (1486.6 eV) was the X-ray source at an applied vacuum of ≤5 × 10−10Torr. Pass energies of 160 eV and 20 eV were applied for survey and high resolution scans respectively. Energy resolution was 0.80 eV for gold 4f photoemission spectra. The high resolution spectra were calibrated based on C 1s binding energy of 284.8 eV and peak fitting was based on Shirley background calculations. JEOL 2010F TEM/ STEM field emission transmission electron microscope, operating at 200 kV and equipped with Oxford INCA EDS system was used to visualize the pore structure and analyze for gold in the hybrid mesoporous adsorbent. Samples were suspended in ethanol and were let to be sonicated (for few minutes). A droplet of the dispersion was placed on the holey carbon-coated copper grid and let for evaporation of solvent before being loaded onto the TEM sample holder. Total carbon, nitrogen and sulfur contents in the hybrid mesoporous silica samples were determined using Leco analyzer.

3. Results and discussion 3.1. Structural and functional properties of mesoporous adsorbents According to the N2 sorption analysis data (Table 1 and Fig. 1), the amine-functionalized material had considerably smaller specific surface area and pore volume compared with the pristine structure. This was not the case for thiol-functionalized sample. Based on mole contents of nitrogen calculated from elemental analysis data, and according to TGA weight loss values, more functional groups were loaded during amine grafting of the mesoporous silica. Normally this is expected from grafting functionalization in which (lateral) self-condensation of the functional organosilanes in the presence of even very little moisture could ultimately result in high surface coverage without necessarily direct bonding of the organic groups with the main silica framework [25]. This is in contrast with co-condensation functionalization in which functionalization is mainly based on the availability and accessibility of surface silanol groups [26–28].

Fig. 1. Results of N2 sorption test on pristine and differently functionalized mesoporous silica; a) mesopore size distributions and b) adsorption–desorption isotherms.

B. Fotoohi, L. Mercier / Hydrometallurgy 156 (2015) 28–39

31

3.2. Gold adsorption capacities (isotherms)

Fig. 2. FTIR characterization of original as well as differently functionalized mesoporous silica.

According to FTIR spectrometry analyses (Fig. 2), C–H stretching vibrations (generally within 2850–3000 cm−1) and C–H bending vibration at ~ 1385 cm−1 in the pristine silica was considered to be due to the remaining surfactant (structure-directing agent). Proofs of thiol functionalization were very weak S–H stretching vibration observed at ~ 2575 cm−1 [29,30], enhancements in the C–H stretching at ~ 2935 cm−1 and bending vibrations (δCH2) at 1448 and 1410 cm−1. There was no distinguishable change in the IR reflection of silanol (Si–OH) groups because of the nature of co-condensation synthesis which preserves abundant hydroxyl groups on the surfaces. Evidences on amine grafting, however, were major N–H stretching modes (νas at ~ 3375 cm−1 and νs at ~ 3,300 cm−1), as well as bending vibrations (δNH2 at ~ 1560 cm− 1 and broad at ~ 1595 cm− 1). The enhancement of the small IR band at ~ 695 cm− 1 was considered to be partially from primary amine bending vibration which could also be attributed to Si–CH2 bond [31,32]. Increase in the C–H stretching and specifically (generally between 2800–3000 cm− 1 at ~ 2934 cm− 1), appearance of bending vibrations (δCH2 at ~ 1470, 1448 and 1410 cm−1), weakening of Si–OH wide stretching band (at ~ 3440 cm−1) and bending vibration (δOH at ~ 1630 cm−1 and ~ 960 cm−1), could all be indications of considerable removal of surface silanol groups due to higher number of amine groups loaded by grafting than those of thiol groups obtained through co-condensation [32–37].

Adsorption capacity for different hybrid mesoporous silica was evaluated by adsorption isotherms obtained at pre-determined reagent conditions in a simulated gold thiosulfate leach solution after 24 hour time considered for equilibrium (based on kinetic tests). The results are shown in Fig. 3. The most noticeable observation was the difference between gold adsorption behavior and capacities of thiol and amine-functionalized material. It was clear that the adsorption capacity on MSU2-MP was approaching a plateau near an average 1 mmol gold per gram of adsorbent. This would suggest a mole ratio Au/S ≈ 1 on mesoporous surfaces based on the sulfur content of the hybrid material (Table 1). This is similar to what has been suggested in the literature for chemisorption of mercury in a close-packed monolayer on thiol groups with high surface coverage on mesoporous silica [38]. In fact, the shape of isotherms also indicated the typical “type IV” adsorption behavior for a mesoporous structure. Gold adsorption on thiol material in the presence of ammonia solution (pH 10.5) was higher than in the presence of ammonium nitrate (pH 7.5). The latter could be attributed to the thiol–disulfide interchange reaction promoted in alkaline condition in which two adjacent surface thiol groups are oxidized (losing protons) to form thiolate (RS−) species. The latter compounds are known to be much stronger nucleophiles than thiol groups and would react to form disulfide [39]. Whether disulfide has been formed or it might have performed better than thiol in coordination with gold in these experiments, needs further analytical or spectrochemical investigations. On the other hand, adsorption by amine-functionalized material showed no apparent saturation point up to fairly high gold concentrations in solution. As mentioned before, based on analytical measurements, nitrogen mole content was almost twice as that of sulfur (in thiol-functionalized material); therefore one explanation to such high gold take-up was the mole number of nitrogen (amine) groups loaded per gram of adsorbent. Looking further at the isotherms, the adsorption capacity of MSU2-AP approached 2 mmol gold per gram of adsorbent which was almost satisfying the mole ratio of Au:N = 1:1, but it kept soaring up beyond such capacity toward 3 mmol/g. There would be at least two questions to be answered at this stage. What causes such an extraordinary adsorption capacity beyond chemisorption (monolayer) limits? Even more fundamental, what adsorption mechanism might describe affinities for gold (complex) toward amine functional groups? In this regard, it is important to note that the HSAB theory has its limitations in not taking into account certain factors that contribute to the stabilization of the bond strength such as the orbital overlap between

Fig. 3. Gold adsorption isotherms obtained for differently functionalized material showing the effect of ammonium/ammonia source in simulated thiosulfate solutions.

32

B. Fotoohi, L. Mercier / Hydrometallurgy 156 (2015) 28–39

the acid and the base. Strength of an acid (or base) might be sensitive to the solvent and its chemical environment. Moreover in reactions between acids and bases, there is a tendency for “strong acids” to displace “weak acids” and “strong bases” to displace “weak bases” (the strongdisplaces-weak or SW acid–base rule) [40]. This means for example if amine/ammonia hard base is “stronger” than thiosulfate soft base, the former reaction is favored and the thiosulfate ligand might be displaced by amine groups.

3.3. Characterization of adsorption system 3.3.1. FTIR study of hybrid mesoporous silica after gold adsorption The infrared absorption by thiol and amine-functionalized material before and after gold adsorption are compared in Fig. 4. It is reminded that in these experiments, adsorption system contained ammonium nitrate as the main source of ammonium/ammonia in solution to maintain a neutral pH (~7.5) system. First, it should be noted that most of sulfur-related IR reflections were unfortunately either masked by other stronger and/or wider reflections due to the silica framework or have not fallen within the mid-range energy spectra of the analysis. For example the S–O stretching vibrations around 1000 cm−1 were supposedly covered with the strong and broad Si–O–Si reflections spanning 975–1300 cm−1. The disappearance of the characteristic NH2 bending vibrations (1555–1595 cm−1) and the SH weak reflection after gold

Fig. 4. Comparing the differences of the FTIR pattern of hybrid mesoporous silica “after” gold adsorption (circled at indicated wavenumbers) with that of the same material “before” adsorption for a) MSU2-MP and b) MSU2-AP adsorbents.

adsorption by hybrid materials were possible indications of involvement of these ligands in complexing with (metal) species in solution. Recurrence/enhancement of the peaks at 950 cm−1 and 1625 cm−1 in MSU2-AP referred to considerable rehydration of silica surfaces. This is expected from more hydrophilic surfaces in amine-bearing silica as opposed to those of thiol-bearing silica. The new reflections at about 3200–3250 cm−1 as well as 1350–1400 cm−1 (wide reflection spanning ~1300–1500 cm−1 in amine-bearing silica) after gold adsorption, all related to N–H vibrations, would suggest that ammonium/ammonia bearing compound and/or complexes might have found their way toward adsorbing onto the surfaces. This was further evidenced by comparing the FTIR pattern of adsorbent with those of the reference materials (NH4Cl and (NH4)2S2O3). Finally the new IR reflection at ~ 660 and 635 cm−1 in amine and thiol-bearing silica respectively was considered to be the only possible evidence of gold on the hybrid materials. The absorption band, potentially referring to a slightly shifted S–O vibration, could well indicate a metal bonding such as Au–S–O on the material surfaces. Although similar reflections were observed in the thiosulfate salt references (including (NH4)2S2O3, Na2S2O3), the closest values were matching the IR double peak at 660 and 640 cm−1 in the Na3Au(S2O3)2 reference salt. 3.3.2. Spectrochemical investigations at different reagent and pH/Eh conditions Hybrid mesoporous silica samples obtained from different equilibrium conditions in gold adsorption tests were analyzed for their surface chemistry by X-ray photoelectron spectroscopy. On the other hand, the adsorption solutions were separately investigated for their components by UV–vis spectroscopy analysis. The general reagent and equilibrium conditions in these tests as well as gold adsorption recoveries are summarized in Table 2. The results of UV–vis characterizations are given in Table 3 and relevant absorption spectra in Fig. 5. Despite some variations of the pH and redox potential values at similar reagent conditions in the presence of thiol and amine-functionalized silica adsorbents, the effect of surface functional groups was found to be minor on the UV–vis absorption spectra and therefore, the graph would be the same for either of the adsorbents MSU2-AP or MSU2-MP. The UV absorption bands at 192 nm and 212 nm were assigned to the thiosulfate anion with the former attributed to charge-neutralizing cation/proton associations or metal complexation (test 1). Higher intensity of the peak at 212 nm was attributed to sulfate portion of the thiosulfate anion and in general oxide sulfur compounds [41,42]. It was stronger relative to the 192 nm reflection with addition of thiosulfate in test 2 conditions in which higher oxidation potentials in the absence of ammonium/ammonia could have caused considerable oxidation of thiosulfate in solution. The wide and intense UV absorption peak at ~ 200 nm was related to the presence of nitrate in the system (tests 3-1 and 3-2). It had apparently masked other previously discussed thiosulfate peaks. Observation of the entire band necessitated thousands of times dilution of the original solutions. Its intensity also showed dependence on the total nitrate content in the solution. The peak intensity was further enhanced by addition of peroxide (test 3-2). The contribution to the UV absorption band at ~ 298 nm was still believed to be due to nitrate [43,44]. The latter reflection always accompanied the 200 nm band, and its intensity was proportional to the nitrate content of the solution. At high nitrate concentrations, where there was no thiosulfate or copper in the solution, the peak was observed at 299-300 nm. There was a slight shift in the position of the peak to 297–298 nm by addition of thiosulfate or where copper and thiosulfate were both added to the solution. The latter observation could refer to possible ion association of nitrate [45]. Moreover, there was also another hypothesis that the shift might have been related to formation of a mixed-ligand complex such as [Cu(NH3)(S2O3)]− in solution. Although there was no other evidence at the time to support the presence of such compound in these

B. Fotoohi, L. Mercier / Hydrometallurgy 156 (2015) 28–39

33

Table 2 Experimental conditions and gold adsorption data in the presence of differently functionalized adsorbents considered for UV–vis and XPS characterizations. Test number Type of investigation

Dosing concentration(s)

1

[S2O3]2-: 0.0002 M No simulation

Equilibrium conditions pH Eh (mV) Solution color/odor

–NH2 MSU2-AP

9.2 +415 ↓ (and decreasing) Colorless 5.2 +570 ↑ (and increasing) Colorless 6.9 +274 Colorless 5.7 +225 Colorless, sulfur odor 7.3, 7.4 +235, +240 Colorless

65

7.5, 7.7 +333, +363 Blue, ammonia odor

53

10.4 +160 Blue (a bit lighter for –SH), ammonia odor

66

–SH MSU2-MP [S2O3]2-: 0.1 M

2 Effect of excess thiosulfate from (NH4)2S2O3

–NH2 MSU2-AP –SH MSU2-MP

3-1 Neutral-pH simulation

3-2 Neutral-pH simulation & effect of redox potential

4 Alkaline-pH simulation

[S2O3]2-: 0.1 M [NH+ 4 + NH3]: 1.0 M [NO3]-: 0.8 M [Cu]2+: 2.5 mM [S2O3]2-: 0.1 M [NH+ 4 + NH3]: 1.0 M [NO3]-: 0.8 M [H2O2]: 0.15 M [Cu]2+: 2.5 mM [S2O3]2-: 0.1 M [NH+ 4 + NH3]: 1.0 M [Cu]2+: 2.5 mM

Gold adsorption recovery (%) Out of 10 mg/L initial gold conc.

Functional group Sample

–NH2 MSU2-AP –SH MSU2-MP –NH2 MSU2-AP –SH MSU2-MP –NH2 MSU2-AP –SH MSU2-MP

tests, the double complex was known as a more stable copper(I) specie in such solutions compared with single copper–ammine or copper– thiosulfate complexes, especially at high [ammonia]/[thiosulfate] molar ratios [46]. The discussed band was not generally observed in the systems containing high ammonia (versus ammonium) in alkaline conditions (test 4). It was assumed that in such conditions, copper speciation was mainly divided between copper(II)–ammine (hence the blue coloration) and copper(I)–thiosulfate. Finally, the wide visible range absorption band with its peak at 610 nm was due to copper(II)– ammine complex. Its intensity was gradually reduced (2.5 times) by increasing thiosulfate concentration from zero to 0.1 M. Therefore, formation of copper(I)–thiosulfate complex in such alkaline solution was confirmed to occur at molar concentration of [thiosulfate]/[total ammonia] ≈ 2.4% and higher.

12

11

12

60 30

51

60

Results of XPS analyses on amine-functionalized adsorbent (MSU2-AP) before and after adsorption from different solutions are presented in Fig. 6. According to Fig. 6a, there was an XPS peak at 406.9 eV observed in the N 1s high resolution spectra before gold adsorption. Such photoemission energy is normally attributed to oxidized nitrogen either in inorganic (e.g., NO3) or organic compounds (e.g., NO2 groups) in the XPS database [47]. Such reflection after gold adsorption in the ammonium nitrate-contained systems (tests 3-1 and 3-2) could have been attributed to association of nitrate anion due to possible electrostatic interaction with protonated surface amine groups [48]. However, in the original adsorbent, the source of such N–O related peak was uncertain. One hypothesis, though, could be partial oxidation of amine groups upon interaction with surface hydroxyl (silanol) groups on silica or ethoxy groups of organosilane

Table 3 Results of UV–vis spectroscopy of different test solutions at experimental conditions according to Table 2. UV–vis absorption wavelength (nm)

Test #

Approximate intensity: multiplied by dilution factor (if applicable) Arbitrary unit

Ion or complex assigned to

Confirming reference compound(s) in DI water

192, 193

1 2-2 4 1 2-2 4

1.7 255 400 0.5 300 270

Cation-[S2O3]2− complex + + Cation could be: H3O+, Na+, NH+ 4 , Au , Cu

200–203 (Wide)

3-1 3-2

7400 9000

298, 299

3-1 3-2

6–7 5–6

610 (Very wide)

3-2 4

0.08–0.1

[NO3]− (π → π*) Seems it is masking other reflections (e.g., those of thiosulfate) [NO3]− (n → π*) [Cu(NH3)(S2O3)]− (?) [Cu(NH3)4]2+

1) Na3Au(S2O3)2 in DI water 2) 0.1 M thiosulfate from (NH4)2S2O3 or Na2S2O3 with/without 2.5 mM copper from Cu2CO3(OH)2 3) 0.02, 0.05, 0.1 M thiosulfate from (NH4)2S2O3 + 0.8 M NH3 solution + 2.5 mM copper from Cu2CO3(OH)2 1) 0.8 M NH4NO3 with/without 2.5 mM copper from Cu2CO3(OH)2 + 0.0–0.1 M thiosulfate from (NH4)2S2O3 2) 0.1 M thiosulfate from (NH4)2S2O3 +0.01, 0.05, 0.8 M NH4NO3 3) 2.5 mM copper from Cu(NO3)2

212–213

[S2O3]2− anion and possibly its oxidation products: e.g., polythionates

0.8 M NH3 solution + 2.5 mM copper from Cu2CO3(OH)2 + 0.0–0.1 M thiosulfate from (NH4)2S2O3 (equilibrium pH 10.5)

34

B. Fotoohi, L. Mercier / Hydrometallurgy 156 (2015) 28–39

Fig. 5. UV–vis spectrometry curves obtained for different adsorption solutions (test numbers on the graphs are according to Table 2).

Fig. 6. XPS spectra obtained for MSU2-AP showing high resolution scans of a) N 1s in the original adsorbent, b) N 1s after adsorption test 1, c) Au 4f after adsorption test 1 and d) survey scan after adsorption test 1.

B. Fotoohi, L. Mercier / Hydrometallurgy 156 (2015) 28–39

during grafting process. The boiling temperature and possible presence of small amounts of moisture might have also triggered such reaction to occur. Whatever the reason though, such oxidation of amine groups within mesoporous silica frameworks during functionalization has not been reported in the literature to the best of the authors' knowledge. Adsorption of gold from pure solution (test 1) was accompanied by enhancement of XPS photoemission at 399.6 eV due to NH2 and its possible involvement in metal complexation. As shown in Fig. 6b, other N 1s peaks at 401.9 (due to hydrogen-bonded nitrogen) and 406.9 were both attenuated after gold adsorption. Supposing the latter peak was due to oxide nitrogen on the surface of mesoporous silica, the weakening of the peak could be resulted from reducing effect of thiosulfate. Looking at the gold 4f photoelectron energies (Fig. 6c), there seemed to be a possibility of gold being reduced to metallic gold upon adsorption. The position of Au 4f doublet in the XPS spectrum was closest to that of the metallic gold (Au0) based on the 83.7 eV energy value for the major peak (4f7/2) [49,50]. It is reminded that the MSU2-AP powder obtained from test 1 and test 4 had a slightly yellow tint. Unfortunately gold XPS reflections were not detected in the rest of MSU2-AP samples obtained from simulated adsorption solutions, despite performing repeated XPS analysis to get more number of scans. This was happening even for the sample obtained from test 2 in which only additional thiosulfate (from ammonium thiosulfate) was added. Since gold adsorption recovery by amine-functionalized sample in test 3 was as high as its amount obtained for pure solution condition in test 1 (Table 2), the absence of gold XPS peaks in the former case was related more to the effect of adsorption solution conditions rather than it being a matter of analytical detection limit. Possible explanations could be masking of gold surfaces by deposit layers of certain solute species, inaccessibility of gold (molecules or nanoparticles) to the beam and most probably the lack of sufficient spatial resolution under the applied X-ray beam conditions. The latter is especially the case where individually dispersed tiny nanoparticles are formed within the matrix of the adsorbent [51]. Whatever the reason, only more detailed study of both adsorption solution and surface chemistry of the mesoporous material could provide further evidence. Investigation of N 1s high resolution scans of the MSU2-AP sample obtained from other test conditions revealed more variations in the intensity of the discussed peaks, reflecting changes in the proportional distribution of nitrogen species on the surfaces of the material. As shown in Fig. 7, generally in neutral-pH simulated conditions the dominant N 1s peak was the one at 401.9 and 402.6 eV for the sample used in test 3-1 and 3-2 respectively. Same N 1s peaks were the only evidence of nitrogen existing on thiol-functionalized material (MSU2-MP) after adsorption in simulated test conditions. This dominance represented considerable presence of protonated amine and/or adsorbed species mainly ammonia/ammonium possibly as ion associations or even as complexes (such as [Au(NH3)2]+) on the surfaces. Since no copper was detected in the survey scans of any of the adsorbents in the simulated solutions, adsorption of copper, e.g., as [Cu(NH3)4]2+ complex, was considered to be insignificant if happening at all.

35

Once again the N 1s at 409.6 eV, previously attributed to oxide nitrogen, was observed in MSU2-AP obtained from simulated adsorption solutions. This time it could also be attributed to the presence of nitrate from solutions possibly in the form of surface deposits of salts/ion associations. It is noticeable that this nitrogen XPS reflection was observed at 407.4 eV in the sample obtained from test 3-2 and considerably weaker than that from test 3-1 and original sample. Since the latter test was run at higher oxidation potential conditions, it remains a question if this reflection was exactly referring to nitrogen oxide species on the surface. But something for sure was that the peak was absent in the thiolfunctionalized material throughout the entire test conditions. Therefore, its presence and intensity was clearly depending on the amine functional groups. Looking at the XPS spectra obtained for MSU2-MP sample before gold adsorption (Fig. 8a), the peak positions in the S 2p high resolution spectra at 163.5 eV and 164.7 eV could be attributed to the spin-orbit coupling aspect associated with S 2p3/2 and S 2p1/2 respectively [52]. However, after contacting hybrid silica with adsorption solutions, the discussed S 2p peaks could also indicate the speciation of sulfur with respect to its bonding with metal (e.g., gold) or another sulfur atom (as in thiosulfate). There was an enhancement in the S 2p at 163.7 eV (comprising 88% of the total spectra) after contacting MSU2-MP with pure gold thiosulfate solution in test 1 as seen in Fig. 8b. There is evidence in the literature that associates the latter S 2p binding energy with metal-thiolate species [53]. Further confirmation on gold bonding was the distinct Au 4f spectra (Fig. 8c) in which the major peak was at 84.7 eV (1 eV higher than that in MSU2-AP at similar conditions). Such binding energy was closest to that of gold(I)–thiol complex [54], and therefore suggesting a possible ligand exchange without affecting the metal oxidation state. Similar to the case of amine-functionalized adsorbent, detection of gold XPS peaks for the adsorbent obtained from simulated conditions was not as easy as that in the pure solution. However, except for test 3-1, in other cases there were weak reflections of the Au 4f spectra observed for sample MSU2-MP. There were very weak (just above the noise level) Au 4f doublets observed for the material obtained from either of test 2 or test 4 independent of their different gold contents based on the adsorption data. The doublet was stronger for the thiolbearing silica after adsorption in test 3-2 conditions; however, the peaks were further shifted toward higher binding energies so that the major peak was found at 85.7 eV. The latter, suggests that that gold(I) was still the dominant gold specie interacting with (oxidized) thiol groups on the material surfaces after adsorption in test 3-2 conditions. Further investigation of XPS data based on S 2p spectra obtained for differently treated adsorbents showed certain degrees of sulfur oxidation based on the adsorption medium and adsorbed species. The thiol-functionalized silica showed the S 2p binding energy at 169.0 eV representative of oxidized sulfur, comprising about 20% of the total sulfur atoms (Fig. 9a). This peak was found at 169.3 eV with considerably higher intensity (Fig. 9b) where MSU2-MP was exposed to hydrogen

Fig. 7. XPS high resolution scans in the N 1s region for a) MSU2-AP from test 3-1, b) MSU2-AP from test 3-2 and c) MSU2-MP from test 3-2.

36

B. Fotoohi, L. Mercier / Hydrometallurgy 156 (2015) 28–39

Fig. 8. XPS spectra obtained for MSU2-MP showing high resolution scans of a) S 2p in original adsorbent, b) S 2p after adsorption test 1, c) Au 4f after adsorption test 1 and d) survey scan after adsorption test 1.

Fig. 9. XPS high resolution scans in the S 2p region for a) MSU2-MP from test 3-1, b) MSU2-MP from test 3-2, c) MSU2-MP from test 4, and d) MSU2-AP from test 3-1.

B. Fotoohi, L. Mercier / Hydrometallurgy 156 (2015) 28–39

peroxide additions in the adsorption solution in test 3-2. Sulfonic acid (– SO3H) groups have been reported on mesoporous silica showing similar photoemission energies as in these experiments [55]. The oxidized sulfur contribution to the XPS spectra at other test conditions (e.g., test 2) and also in the case of amine-functionalized material were mainly attributed to the S–O bonding in the thiosulfate molecule and its oxidation products (e.g., sulfite and polythionates), in which case the peak was found at slightly weaker binding energies (168.4–169.0 eV). The latter was approved when weak S 2p reflections of both sulfide and sulfate species were always present in the XPS scans of amine-functionalized material (Fig. 9d). The lowest sulfur oxidation in the material exposed to simulated solutions was obtained in alkaline condition of test 4 wherein major oxidizers (e.g., nitrate) were absent and ammonia solution brought the reductive environment (Fig. 9c). It is reminded that gold adsorption on thiol-functionalized mesoporous silica in test 3-2 and 4 was considerably higher than that in other test conditions. It is known that thiolates (RS−) are more reactive than their protonated counterparts (i.e., thiols, RSH) toward bonding both with metals and with adjacent similar groups on the surfaces, and alkaline pH triggers the so-called thiol–disulfide interchange reaction [39]. In the XPS results, MSU2-MP from alkaline adsorption test 4 showed enhanced S 2p peak intensity at 164.7 eV when compared with same sample before adsorption and also after adsorption in the near-neutral pH conditions. Therefore, there is a possibility of disulfide to have been formed and it was superior in adsorbing gold than thiol was in the applied reagent conditions. On the other hand, high performance of MSU2-MP in test 3-2 condition could possibly be attributed to the involvement of oxidized thiol groups of which advanced oxidation would end up producing sulfonic acid (or sulfonate) [56]:

It would worth noticing that through an investigation of gold(I) complexes with different sulfur ligands in solution the order on the stability of complexes based on softness of the ligands followed gold(I) disulfide [Au 2S 2] 2 − N gold(I) dithiol [Au(HS)2 ] − N gold(I) disulfite [Au(SO3)2]3 − N gold(I) thiosulfate [Au(S2O3)2]3 − N gold(I) thiol [Au(HS)]0 [57]. Same judgment might be true when

37

considering organic sulfite (sulfonate) groups, assuming they are formed on MSU2-MP surfaces in test 3-2 conditions. Such organic groups might have been stronger and preferred ligands to be exchanged with thiosulfate during surface interactions, thus providing higher adsorption capacity. 3.3.3. TEM observations Spherical MSU-2 hybrid silica particles were too big (thick) to be studied well with the TEM analysis. But it was possible to observe the pore network around the thinner edges of the spheres. Due to such thickness of the particles, they generally appeared dark so that any visual detection based on contrast inside the pore network was difficult. However, in TEM analysis of the amine-functionalized material (MSU2-AP) used in test 3-1, there were occasions in which some dark spots as well as overall crystalline-like pattern were observed (Fig. 10). It was so surprising to observe nanoparticles inside the mesoporous structure of amine-functionalized mesoporous silica during TEM observation. However, this time the material was a different member of the MSU-X silica family called MSU-3. The latter structure was characterized with (hexagonal) ordered mesopore channels having almost three times the diameter of the mesopores in MSU-2 silica (Fig. 11a). Although MSU3-AP material was not prepared for all the investigations in this research, it was functionalized with amine groups similar to MSU2AP and was used in the same adsorption system as in test 3-1 above. The TEM study at the thin edges of particles in the high resolution mode revealed the presence of nanoparticles in general less than 5 nm in size (shown as dark spots in Fig. 11b–d). The overall electron diffraction patterns were sometimes indicating amorphous structures (Fig. 11c) and sometimes showing the ring pattern suggesting the presence of a multitude of tiny single crystals (Fig. 11d). Still at higher magnifications, though, it was possible to visualize the striated pattern in most of the spots referring to their crystalline structure, and the electron diffraction pattern was representative of a single crystal (Fig. 11e). Unfortunately it was not possible to analyze the composition of the nanocrystals with the integrated EDS analyzer due to the instrumental restrictions in obtaining appropriate spatial resolution. There have been numerous investigations on preparation of goldloaded mesoporous silicas of different 2D and 3D mesopore network to be used for catalysis applications [58]. Although different methodologies (including impregnation and subsequent reduction) are normally used in preparation of gold nanoparticle catalysts loaded on hybrid mesoporous silicas, it would be interesting to understand the mechanism by which gold (thiosulfate) is reduced on amine functional groups as in this research.

Fig. 10. Amine-functionalized MSU-2 after gold adsorption (test 3-1); a) overall particle morphology, b) common mesopore structure at thin edges and c) the crystalline-like pattern dominant all over the mesopore network of a single particle.

38

B. Fotoohi, L. Mercier / Hydrometallurgy 156 (2015) 28–39

Fig. 11. TEM images of MSU3-AP sample; a) overall view of the mesopore structure, b–e) high resolution images showing nanoparticles and electron diffractions.

4. Conclusion

Acknowledgments

Gold adsorption chemistry on amine and thiol-functionalized mesoporous silica when adsorbing from simulated thiosulfate leach solutions was studied. While gold adsorption on thiol-functionalized silica was found to follow a typical monolayer chemisorption behavior, it was not exactly the case for amine functional groups. Although higher gold adsorption capacity on amine-functionalized material was mainly attributed to the higher concentration of the ligands in the mesoporous structure of hybrid silica, the explanation of gold–amine complexation was beyond the principles of the hard–soft acid–base theory. It was believed that the strength of donor ligands such as primary amines when immobilized on the surfaces inside confined mesopore structures should be evaluated with respect to fairly stable complexes such as gold thiosulfate in solution. Moreover, the redox reactions that might occur during interaction of gold species with primary amine ligands in neutral to alkaline pH systems as in these experiments, could have contributed to the higher gold adsorptions due to partial reduction of adsorbed gold into nanoparticles. Therefore, further investigations are recommended in order to characterize the exact mechanisms by which such ligands adsorb gold in similar solutions. The major benefits of such studies by using high capacity and selective adsorbent materials would directly affect precious metal adsorption recovery and recycling from primary and secondary sources. On the other hand, if the formation of gold nanoparticles in these experiments is fully evidenced and confirmed, it might be of great significance in the catalyst preparations.

We would like to appreciate the comments and advice obtained from Drs. J. Gray-Munro and E. Guerra for FTIR and XPS studies. The authors also wish to thank C. Andrei for TEM studies at the Canadian Centre for Electron Microscopy, McMaster University, as well as D. Karpuzov for XPS analysis at Alberta Centre for Surface Engineering and Science, University of Alberta. Thanks are also due to AFL at University of Guelph for C, N and S elemental analyses. Finally, Laurentian University and the government of Ontario are acknowledged for their part in the provisions of the facilities and scholarship supports respectively.

References [1] Aylmore, M.G., Muir, D.M., 2001. Thiosulfate leaching of gold—a review. Miner. Eng. 14 (2), 135–174. [2] Woods, R., Hope, G.A., Watling, K.M., Jeffrey, M.I., 2006. A spectroelectrochemical study of surface species formed in the gold/thiosulfate system. J. Electrochem. Soc. 153 (7), D105–D113. [3] Breuer, P.L., Jeffrey, M.I., 2002. An electrochemical study of gold leaching in thiosulfate solutions containing copper and ammonia. Hydrometallurgy 65 (2–3), 145–157. [4] Zhang, S., Nicol, M.J., 2003. An electrochemical study of the dissolution of gold in thiosulfate solutions, part I: alkaline solutions. J. Appl. Electrochem. 33 (9), 767–775. [5] Chandra, I., Jeffrey, M.I., 2004. An electrochemical study of the effect of additives and electrolyte on the dissolution of gold in thiosulfate solutions. Hydrometallurgy 73 (3–4), 305–312. [6] Feng, D., Van Deventer, J.S.J., 2010. Effect of thiosulphate salts on ammoniacal thiosulphate leaching of gold. Hydrometallurgy 105 (1–2), 120–126.

B. Fotoohi, L. Mercier / Hydrometallurgy 156 (2015) 28–39 [7] Senanayake, G., 2012. Gold leaching by copper(II) in ammoniacal thiosulphate solutions in the presence of additives, part I: a review of the effect of hard–soft and Lewis acid–base properties and interactions of ions. Hydrometallurgy 115, 1–20. [8] Senanayake, G., 2005. Role of copper(II), carbonate and sulphite in gold leaching and thiosulphate degradation by oxygenated alkaline non-ammoniacal solutions. Miner. Eng. 18 (4), 409–426. [9] Feng, D., Van Deventer, J.S.J., 2007. The effect of sulphur species on thiosulphate leaching of gold. Miner. Eng. 20 (3), 273–281. [10] Senanayake, G., 2005. The role of ligands and oxidants in thiosulfate leaching of gold. Gold Bull. 38 (4), 170–179. [11] Braul, P., 2013. Thiosulphate Going Commercial: Barrick's Goldstrike Pushes Research Forward on Cyanide Alternative. Canadian Institute of Mining, Metallurgy and Petroleum. [12] Qu, R., Wang, M., Sun, C., Zhang, Y., Ji, C., Chen, H., et al., 2008. Chemical modification of silica-gel with hydroxyl- or amino-terminated polyamine for adsorption of Au(III). Appl. Surf. Sci. 255 (5, Part 2), 3361–3370. [13] Kim, M.L., Tudino, M.B., 2010. Evaluation of performance of three different hybrid mesoporous solids based on silica for preconcentration purposes in analytical chemistry: from the study of sorption features to the determination of elements of group IB. Talanta 82 (3), 923–930. [14] Trokhimchuk, A.K., Andrianova, E.B., Losev, V.N., 2004. Interaction peculiarities of gold (III) with silica gels containing both aminopropyl and mercaptopropyl surface groups. Adsorpt. Sci. Technol. 22 (10), 837–848. [15] Lam, K.F., Fong, C.M., Yeung, K.L., Mckay, G., 2008. Selective adsorption of gold from complex mixtures using mesoporous adsorbents. Chem. Eng. J. 145 (2), 185–195. [16] Aledresse, A., Amaratunga, L., Mercier, L., 2008. Novel high efficiency gold extraction process using nanotechnology. 40th Annual Meeting of the Canadian Mineral Processors. CIM, Ottawa, Canada, pp. 497–508. [17] Fotoohi, B., Mercier, L., 2014. Recovery of precious metals from ammoniacal thiosulfate solutions by hybrid mesoporous silica: 1—factors affecting gold adsorption. Sep. Purif. Technol. 127, 84–96. [18] Finkelstein, N.P., Hancock, R.D., 1974. A new approach to the chemistry of gold. Gold Bull. 7 (3), 72–77. [19] Pearson, R.G., 1963. Hard and soft acids and bases. J. Am. Chem. Soc. 85 (22), 3533–3539. [20] Daniel, M.C., Astruc, D., 2003. Gold nanoparticles: assembly, supramolecular chemistry, quantum-size-related properties, and applications toward biology, catalysis, and nanotechnology. Chem. Rev. 104 (1), 293–346. [21] Love, J.C., Estroff, L.A., Kriebel, J.K., Nuzzo, R.G., Whitesides, G.M., 2005. Selfassembled monolayers of thiolates on metals as a form of nanotechnology. Chem. Rev. 105 (4), 1103–1170. [22] Boissière, C., Larbot, A., Bourgaux, C., Prouzet, E., Bunton, C.A., 2001. A study of the assembly mechanism of the mesoporous MSU-X silica two-step synthesis. Chem. Mater. 13 (10), 3580–3586. [23] Boissiere, C., Van Der Lee, A., El Mansouri, A., Larbot, A., Prouzet, E., 1999. A double step synthesis of mesoporous micrometric spherical MSU-X silica particles. Chem. Commun. 20, 2047–2048. [24] Iler, R.K., 1979. The Chemistry of Silica: Solubility, Polymerization, Colloid and Surface Properties and Biochemistry of Silica. Wiley-Blackwell, New York, p. 896. [25] Athens, G.L., Shayib, R.M., Chmelka, B.F., 2009. Functionalization of mesostructured inorganic–organic and porous inorganic materials. Curr. Opin. Colloid Interface Sci. 14 (4), 281–292. [26] Mercier, L., Pinnavaia, T.J., 1999. Direct synthesis of hybrid organic–inorganic nanoporous silica by a neutral amine assembly route: structure-function control by stoichiometric incorporation of organosiloxane molecules. Chem. Mater. 12 (1), 188–196. [27] Lim, M.H., Stein, A., 1999. Comparative studies of grafting and direct syntheses of inorganic–organic hybrid mesoporous materials. Chem. Mater. 11 (11), 3285–3295. [28] Stein, A., Melde, B.J., Schroden, R.C., 2000. Hybrid inorganic–organic mesoporous silicates—nanoscopic reactors coming of age. Adv. Mater. 12 (19), 1403–1419. [29] Yun, S., Parrondo, J., Lo, C.-P., Ramani, V., 2011. Functionalized silica composite membranes for direct methanol fuel cells. ECS Trans. 41 (1), 1633–1643. [30] Van Rhijn, W.M., De Vos, D.E., Sels, B.F., Bossaert, W.D., 1998. Sulfonic acid functionalised ordered mesoporous materials as catalysts for condensation and esterification reactions. Chem. Commun. 3, 317–318. [31] Park, J., Kim, H., Park, J., 2012. Characteristics of thiol-functionalized mesoporous silica and its application to silver and cadmium ion removal. Int. J. Environ. Sci. Dev. 3 (2), 81–85. [32] Chang, F.Y., Chao, K.J., Cheng, H.H., Tan, C.S., 2009. Adsorption of CO2 onto aminegrafted mesoporous silicas. Sep. Purif. Technol. 70 (1), 87–95.

39

[33] Iliade, P., Miletto, I., Coluccia, S., Berlier, G., 2012. Functionalization of mesoporous MCM-41 with aminopropyl groups by co-condensation and grafting: a physicochemical characterization. Res. Chem. Intermed. 38 (3–5), 785–794. [34] Lombardo, M.V., Videla, M., Calvo, A., Requejo, F.G., Soler-Illia, G.J.a.A., 2012. Aminopropyl-modified mesoporous silica SBA-15 as recovery agents of Cu(II)-sulfate solutions: adsorption efficiency, functional stability and reusability aspects. J. Hazard. Mater. 223–224, 53–62. [35] Parida, K.M., Rath, D., 2009. Amine functionalized MCM-41: an active and reusable catalyst for Knoevenagel condensation reaction. J. Mol. Catal. A Chem. 310 (1–2), 93–100. [36] Qiu, F., Zhou, Y., Liu, J., Zhang, X., 2005. Synthesis and characterization of polyimide/ silica hybrid waveguide materials via sol–gel technique. Optoelectronic Devices and Integration. Proceedings of SPIE Vol 5644. SPIE, Bellingham, WA. [37] Calvo, A., Angelomé, P.C., Sánchez, V.M., Scherlis, D.A., Williams, F.J., Soler-Illia, G.J.a.A., 2008. Mesoporous aminopropyl-functionalized hybrid thin films with modulable surface and environment-responsive behavior. Chem. Mater. 20 (14), 4661–4668. [38] Feng, X., Fryxell, G.E., Wang, L.-Q., Kim, A.Y., Liu, J., Kemner, K.M., 1997. Functionalized monolayers on ordered mesoporous supports. Science 276 (5314), 923–926. [39] Singh, R., Whitesides, G.M., 2010. Thiol–disulfide interchange. Sulphur-containing Functional Groups. John Wiley & Sons, Inc., pp. 633–658. [40] Cardenas, C., Ayers, P.W., 2013. How reliable is the hard–soft acid–base principle? An assessment from numerical simulations of electron transfer energies. Phys. Chem. Chem. Phys. 15 (33), 13959–13968. [41] Miura, Y., Kawaoi, A., 2000. Determination of thiosulfate, thiocyanate and polythionates in a mixture by ion-pair chromatography with ultraviolet absorbance detection. J. Chromatogr. A 884, 81–87. [42] Padarauskas, A., Paliulionyte, V., Ragauskas, R., Dikčius, A., 2000. Capillary electrophoretic determination of thiosulfate and its oxidation products. J. Chromatogr. A 879 (2), 235–243. [43] Butorac, V., Simeon, V., Tomišić, V., 2007. Effect of temperature on UV spectra of concentrated NaNO3 aqueous solutions. Croat. Chem. Acta 80 (3–4), 533–539. [44] Simeon, V., Butorac, V., Tomisic, V., Kallay, N., 2003. Existence of two forms of nitrate ion in dilute aqueous solutions. Thermodynamic parameters of interconversion. Phys. Chem. Chem. Phys. 5 (10), 2015–2019. [45] Gvozdić, V., Tomišić, V., Butorac, V., Simeon, V., 2009. Association of nitrate ion with metal cations in aqueous solution: a UV–vis spectrometric and factor-analytical study. Croat. Chem. Acta 82 (2), 553–559. [46] Black, S.B., 2006. The Thermodynamic Chemistry of the Aqueous Copper–Ammonia Thiosulfate System. Murdoch University. [47] NIST X-ray Photoelectron Spectroscopy Database, Version 4.1. National Institute of Standards and Technology, Gaithersburg (http://srdata.nist.gov/xps/). [48] Yu, H., Zhu, Y., Yang, H., Nakanishi, K., Kanamori, K., Guo, X., 2014. Facile preparation of silver nanoparticles homogeneously immobilized in hierarchically monolithic silica using ethylene glycol as reductant. Dalton Trans. 43 (33), 12648–12656. [49] Casaletto, M.P., Longo, A., Martorana, A., Prestianni, A., Venezia, A.M., 2006. XPS study of supported gold catalysts: the role of Au0 and Au+δ species as active sites. Surf. Interface Anal. 38 (4), 215–218. [50] Bogdanchikova, N., Pestryakov, A., Tuzovskaya, I., Zepeda, T.A., Farias, M.H., Tiznado, H., et al., 2013. Effect of redox treatments on activation and deactivation of gold nanospecies supported on mesoporous silica in CO oxidation. Fuel. 110, 40–47. [51] Baer, D.R., Engelhard, M.H., 2010. XPS analysis of nanostructured materials and biological surfaces. J. Electron Spectrosc. Relat. Phenom. 178–179, 415–432. [52] Liu, J., Hlady, V., 1996. Chemical pattern on silica surface prepared by UV irradiation of 3-mercaptopropyltriethoxy silane layer: surface characterization and fibrinogen adsorption. Colloids Surf B Biointerfaces. 8 (1-2), 26–37. [53] Bain, C.D., Biebuyck, H.A., Whitesides, G.M., 1989. Comparison of self-assembled monolayers on gold: coadsorption of thiols and disulfides. Langmuir 5 (3), 723–727. [54] Vitale, F., Vitaliano, R., Battocchio, C., Fratoddi, I., Giannini, C., Piscopiello, E., et al., 2008. Synthesis and microstructural investigations of organometallic Pd(II) thiol– gold nanoparticles hybrids. Nanoscale Res. Lett. 3 (11), 461–467. [55] Smith, J.L., Herman, R.G., Klier, K., 2005. Ion exchange, core-level shifts, and bond strengths in mesoporous solid acid SBA-15. In: Čejka, J., Žilková, N., Nachtigall, P. (Eds.), Molecular Sieves: From Basic Research to Industrial Applications. Elsevier B.V., Prague. [56] Jones, C.W., 1999. Applications of Hydrogen Peroxide and Derivatives. In: Clark, J.H. (Ed.), The Royal Society of Chemistry, Cornwall. [57] Senanayake, G., 2004. Gold leaching in non-cyanide lixiviant systems: critical issues on fundamentals and applications. Miner. Eng. 17 (6), 785–801. [58] Gutiérrez, L.-F., Hamoudi, S., Belkacemi, K., 2011. Synthesis of gold catalysts supported on mesoporous silica materials: recent developments. Catalysts 1 (1), 97–154.