**Leaching Kinetics of Sulfides from Refractory Gold Concentrates by Nitric Acid**

**Denis A. Rogozhnikov, Andrei A. Shoppert \*, Oleg A. Dizer, Kirill A. Karimov and Rostislav E. Rusalev**

Department of Non-ferrous Metals Metallurgy, Ural Federal University, Yekaterinburg 620002, Russia; darogozhnikov@yandex.ru (D.A.R.); oleg.dizer@yandex.ru (O.A.D.); kirill\_karimov07@mail.ru (K.A.K.); rusalevrostislav@gmail.com (R.E.R.)

**\*** Correspondence: andreyshop@list.ru; Tel.: +7-922-024-3963

Received: 5 April 2019; Accepted: 19 April 2019; Published: 22 April 2019

**Abstract:** The processing of refractory gold-containing concentrates by hydrometallurgical methods is becoming increasingly important due to the depletion of rich and easily extracted mineral resources, as well as due to the need to reduce harmful emissions from metallurgy, especially given the high content of arsenic in the ores. This paper describes the investigation of the kinetics of HNO3 leaching of sulfide gold-containing concentrates of the Yenisei ridge (Yakutia, Russia). The effect of temperature (70–85 ◦C), the initial concentration of HNO3 (10–40%) and the content of sulfur in the concentrate (8.22–22.44%) on the iron recovery into the solution was studied. It has been shown that increasing the content of S in the concentrate from 8.22 to 22.44% leads to an average of 45% increase in the iron recovery across the entire range temperatures and concentrations of HNO3 per one hour of leaching. The leaching kinetics of the studied types of concentrates correlates well with the new shrinking core model, which indicates that the reaction is regulated by interfacial diffusion and diffusion through the product layer. Elemental S is found on the surface of the solid leach residue, as confirmed by XRD and SEM/EDS analysis. The apparent activation energy is 60.276 kJ/mol. The semi-empirical expression describing the reaction rate under the studied conditions can be written as follows: 1/3ln(1 − *X*) + [(1 − *X*) <sup>−</sup>1/<sup>3</sup> <sup>−</sup> 1] <sup>=</sup> 87.811(HNO3) 0.837(S)2.948e−60276/*RT*·*t*.

**Keywords:** refractory gold concentrate; resources depletion; reducing harmful emissions; arsenic; nitric acid; kinetics; shrinking core model; pyrite; arsenopyrite

## **1. Introduction**

Russia possesses large reserves of gold—more than 14 thousand tons—which exceeds the reserves of the world's main producers, China and Australia, and is slightly inferior to South Africa and Canada. The Russian Federation accounts for 8% of the total world gold production and is among the three largest global producers of the precious metal. Gold-sulfide-quartz and gold-arsenic-sulfide deposits occupy a leading position in the structure of Russia's reserves; the quality of ores is comparable to the world objects of this type [1]. Very important are gold-polysulfide deposits, characterized by relatively high concentrations of gold (3.5–7 gpt).

At the same time, there is a global problem in the metallurgical industry that the quality of processed raw materials is deteriorating due to the depletion of mineral reserves and the extraction of the richest and most easily extractable ore layers. As a result, there is a need to engage poorer and more refractory ores, which are often not amenable to traditional enrichment methods.

The deterioration of ore quality, especially with the transition to the production of lower horizons, occurs in terms of reduction of metal content as well as in terms of increasing proportion of ores with fine and emulsive impregnation of sulfides in one another and the latter in waste minerals.

The share of gold ores of non-ferrous metals, where gold is an associated valuable component, accounts for 18% of the global reserves [2]. Among them, a special place belongs to ores from which gold cannot be extracted by traditional technologies.

Refractory gold-containing ores refer to materials, the extraction of gold from which by cyanidation is low or requires significant amounts of energy and reagents [3].

Currently, it is considered proven that the refractory characteristics of gold associated with sulfides is not only due to the presence of nanoparticles of native gold [4–7], but also due to the existence of solid solution, colloidal particles, surface gold [8–11]. The size of such "invisible" gold may be on the order of nanometers, which explains why it is impossible to extract it by cyanidation, even with the use of ultrafine grinding. It has also been established that in pyrite, which is one of the main carriers of gold in refractory ores, the content of "invisible" gold is greater proportionately to the higher content of arsenic in pyrite and to the finer grain. For example, in the ores of the Twin Creeks deposit, relatively coarse-grained pyrite (10–30 μm) is associated with the lowest arsenic and gold contents (less than 1% As and 17–60 g/t Au), while fine-grained pyrites (less than 2 μm) are the highest (1–2.4% As and 600–1500 g/t Au) [12].

Also important in the detection of "invisible" gold in pyrite grains is the uneven distribution of As and Au over the grain section. A thin peripheral layer of pyrite grain is enriched with arsenic, forming the so-called arsenic pyrite Fe(As,S)2. It tends to contain most of the gold [13].

The nature of the chemical bond of gold, which is in the form of a solid solution in arsenic pyrite, has not been fully established and is the subject of discussion [12,14].

Invisible gold may also exist in arsenopyrite in the simple form of nanoparticles (Au0) and in the oxidized state (Au1+), and their ratio may vary significantly. For example, in the arsenopyrite deposits Jinya (China) [15], Elmtree (Canada) [16], Sao Bento (Brazil) [8] and Sheba (South Africa) [8,16] solid gold (Au1<sup>+</sup>) is the predominant form as compared with nanogold (Au0). On the contrary, arsenopyrite of the Olimpiadinskoe deposit (Russia) [17] contains "invisible" gold mainly in the form of nanoparticles (Au0) [10].

Traditional methods of processing such refractory materials consist in the oxidation of gold-containing minerals (pyrite or arsenopyrite) in order to destroy their crystal lattice and release the gold particles by oxidative roasting [18]. This process is associated with the oxidation of iron-containing sulfides and converting arsenic to the gaseous phase. At the same time, arsenic is one of the most dangerous and carcinogenic elements [19–24] and its content in drinking water in several countries already exceeds the concentration recommended by the World Health Organization (WHO) and the United States Environmental Protection Agency (USEPA) [25,26]. Therefore, hydrometallurgical methods of sulfide oxidation have been widely implemented in recent years. The most common of them are pressure and bacterial oxidation, and leaching after fine grinding [27].

One of the possible methods of hydrometallurgical processing of refractory sulfide raw materials is the use of HNO3 to oxidize the materials without the use of high pressure or fine grinding [28–31], as nitric acid is one of the most effective oxidizing and leaching agents [32,33].

Among the most famous technologies based on the use of HNO3: NSC-process (nitrogen species-catalyzed pressure leaching), implemented in 1984–1995 at the Sunshine plant in the USA [34]; NITROX (in atmosphere air) and ARSENO PROCESS® (use of compressed oxygen) [35]; a subspecies of ARSENO, the REDOX PROCESS® (at above 180 ◦C to eliminate the formation of elemental S) [36]; the HMC process (a mixture of salts of nitric and hydrochloric acids) [37]; the Caschman process and its modification Artek Caschman, which aims to process gold-containing arsenic materials using chlorine-containing reagents [3]. However, none of these is currently used commercially for one reason or another.

Therefore, it seems relevant to conduct further studies of alternative energy saving and environmentally efficient hydrometallurgical technologies for processing sulfide gold-containing raw materials using HNO3. There are only a few published works devoted to the theoretical aspects of HNO3 leaching of sulfide gold-containing concentrates [38], which demonstrates the need for additional investigation using other types of concentrates with different mineralogical compositions. At the same time, there is a sufficient amount of work showing that various types of leaching reactions of raw materials with HNO3 can be described quite accurately using the shrinking core model [39–43], which makes it possible to obtain more data on the limiting stages of the reactions.

Considering the above factors, this paper studies the kinetics of HNO3 leaching of refractory gold-containing concentrates with the use of shrinking core models, focusing on the role of temperature, concentration of HNO3, which, in our opinion, has not been sufficiently studied theoretically or practically. Particular attention is paid to the initial content of sulfur in the raw materials as one of the most important factors affecting the intensity, completeness and kinetic characteristics of the leaching process.

## **2. Materials and Methods**

## *2.1. Analisys*

The chemical analysis of the original materials and the resulting solid products of the studied processes was performed using an Axios MAX X-ray fluorescence spectrometer (XRF) (Malvern Panalytical Ltd., Almelo, Netherlands). The chemical analysis of the obtained solutions was performed by mass spectrometry with inductively coupled plasma (ICP-MS) using an Elan 9000 instrument (PerkinElmer Inc., Waltham, MA, USA). The phase analysis was performed on an XRD 7000 Maxima diffractometer (Shimadzu Corp., Tokyo, Japan).

Scanning electron microscopy with microprobe energy dispersive analysis was performed using a JSM-6390LV microscope (JEOL Ltd., Tokyo, Japan) with the INCA Energy 450 X-Max 80 adaptor, with an accelerating voltage of 20 kV.

The study on the distribution of gold in sulfide minerals was carried out using inductively coupled plasma mass spectrometry (LA-ICP-MS—NexION 300S quadrupole mass spectrometer, PerkinElmer Inc., Waltham, MA, USA) with laser ablation of the sample with the NWR 213 adaptor for a Jeol JSM-6390LV.

## *2.2. Experiments*

Laboratory experiments on HNO3 leaching were carried out on an apparatus consisting of a 2 dm3 borosilicate glass round bottom reactor (Lenz Laborglas GmbH & Co. KG, Wertheim, Germany), with openings for injecting HNO3, as well as for temperature control and removal of nitrous gases through a water-cooled reflux condenser. The reactor was thermostated. The materials were stirred using an overhead mixer at 400 rpm, which ensured uniform density of the pulp. A portion of the concentrate weighing 60 g was added to a prepared solution of HNO3 of a required concentration. At the end of the experiment, the leaching pulp was filtered in a Buchner funnel (ECROSKHIM Co., Ltd., St. Petersburg, Russia); the solutions were sent for ICP-MS analysis; the leaching cake was washed with distilled water, dried at 100 ◦C to constant weight, weighed and sent for XRF analysis. All the experiments were performed three times and the mean values are presented here.

To trap the nitrous gases formed during the leaching process and to regenerate HNO3, we used a system consisting of three successively connected absorption columns filled with distilled water and a sanitary column with a solution of thiourea to recover the residual amount of oxides to elemental nitrogen.

## *2.3. Materials and Reagents*

The materials used in the study are three samples of refractory gold-containing sulfide flotation concentrates (size 90% less than 74 microns) from a Russian deposit of Yenisei ridge (Yakutia), obtained under different enrichment conditions. The compositions of the samples are presented in Table 1. The chemical agents were of analytical grade; the water had been purified by distillation using a GFL-manufactured device (GFL mbH, Burgwedel, Germany).


**Table 1.** Chemical composition of samples of refractory gold-containing sulfide concentrates, wt.%.

Figure 1 shows the X-ray diffraction pattern of the phase composition of the original concentrate-3. Table 2 presents the results of the mineralogical composition study of the concentrate samples. The results were obtained in X-ray phase analysis and X-ray microanalysis.

**Figure 1.** X-ray diffraction (XRD) pattern of the phase composition of concentrate-3.



Figure 2 presents a general view of concentrate-3 in backscattered electrons. The red dots indicate the place of LA-ICP-MS analysis; the analytical point diameter is 25 microns.

The study showed the almost complete absence of gold in antimonite. The concentration of gold in pyrite is fairly evenly distributed and does not exceed 16 gpt. The arsenopyrite distribution is uneven. The gold content ranges from 1 to 172 gpt (Table 3).


**Table 3.** Gold content in minerals <sup>1</sup> of flotation concentrate-3, gpt.

<sup>1</sup> The materials were determined according to the content of Fe, As, Sb, S. Sulfur was used as an internal standard.

**Figure 2.** General view of concentrate-3 in backscattered electrons: AsP—arsenopyrite, Py—pyrite, Sbn—stibnite (antimonite), Ag—silver, dark gray grains—silicate minerals.

## **3. Results and Discussion**

## *3.1. Chemistry of HNO3 Leaching*

As shown by the results of the LA-ICP-MS method with the NWR 213 adaptor, a characteristic feature of the material is that most of the gold is enclosed in finely disseminated form in the crystalline lattice of pyrite and arsenopyrite. Therefore, the main goal of the process is to reveal these two sulfide minerals. Their interaction with HNO3 can follow these typical reactions (Equations (1)–(7)):

$$\text{FeS}\_2 + 8\text{HNO}\_3 = \text{Fe(NO}\_3\text{)}\_3 + 2\text{H}\_2\text{SO}\_4 + 2\text{H}\_2\text{O} + 5\text{NO}\_3 \tag{1}$$

$$\rm{H\_2FeS\_2 + 10HNO\_3 = Fe\_2(SO\_4)\_3 + H\_2SO\_4 + 4H\_2O + 10NO\_2} \tag{2}$$

$$\rm FeS\_2 + 18HNO\_3 = Fe(NO\_3)\_3 + 7H\_2O + 2H\_2SO\_4 + 15NO\_2. \tag{3}$$

$$\rm{2FeS}\_2 + 8HNO\_3 = Fe\_2(SO\_4)\_3 + S^0 + 8NO + 4H\_2O\_4 \tag{4}$$

$$\text{FeAsS} + 17\text{HNO}\_3 = \text{Fe(NO}\_3\text{)}\_3 + \text{H}\_3\text{AsO}\_4 + \text{H}\_2\text{SO}\_4 + 14\text{NO}\_2 + 6\text{H}\_2\text{O},\tag{5}$$

$$\rm FeAsS + 14HNO\_3 = FeAsO\_4 + H\_2SO\_4 + 14NO\_2 + 6H\_2O,\tag{6}$$

$$3\text{FeAsS} + 12\text{HNO}\_3 = 3\text{FeAsO}\_4 + 2\text{H}\_2\text{SO}\_4 + \text{S}^0 + 12\text{NO} + 4\text{H}\_2\text{O}.\tag{7}$$

In addition, at the initial moment, NO<sup>+</sup> ions could be formed, which, according to Anderson et al. [44], acts as the strongest oxidizer in such systems.

Previously, we had studied the thermodynamic characteristics for the above interactions [33]. The study found that the oxidation potential of the system is to be maintained at 0.6 V or higher for reactions with transfer of iron and arsenic into the solution. A high oxidation potential is also necessary for the sulfide S2<sup>−</sup> to be oxidized into sulfate SO4 <sup>2</sup><sup>−</sup> with minimized formation of elemental S0 that impedes a more complete oxidation of sulfide minerals and reduces subsequent recovery of gold. The results of the performed thermodynamic studies and laboratory experiments allowed us to establish the ranges of the leaching process parameters: temperature of 70–100 ◦C, acid concentration of 10–60% at L:S 20:1. Lower values of L:S are also possible; however, it was previously found that at low values of L:S and a high acid concentration, the reaction is very intense, which impedes temperature control, while the degree of extraction in this case changes slightly [33].

## *3.2. The E*ff*ect of Process Parameters on the Oxidation of Sulfides*

According to Tables 1 and 2 and Figure 1, it is obvious that the concentrates studied in this work differ greatly in the content of sulfides; therefore, the study considered only the kinetics of HNO3 leaching of each of the concentrates. The effect of temperature, HNO3 concentration and sulfur content on iron extraction in HNO3 leaching is shown in Figure 3. Iron recovery was considered as the main component indicating the degree of sulfide oxidation, as it is included in both pyrite and arsenopyrite. For each of the concentrates, experiments were carried out at three different HNO3 concentrations and at four temperatures. Figure 3a–c, for example, shows the kinetic leaching curves of concentrate-1 at a concentration of nitric acid of 10, 20 and 40%, respectively, that could be used to evaluate the influence of HNO3 concentration on leaching efficiency. It is possible to evaluate the influence of sulfur content in concentrate on the oxidation of sulfides, for example, by comparing Figure 3a,d,g.

**Figure 3.** The dependence of iron extraction in the leaching of refractory gold-containing concentrate on temperature: (**a**) concentrate-1 and 10% HNO3 concentration; (**b**) concentrate-1 and 20% HNO3 concentration; (**c**) concentrate-1 and 40% HNO3 concentration; (**d**) concentrate-2 and 10% HNO3 concentration; (**e**) concentrate-2 and 20% HNO3 concentration; (**f**) concentrate-2 and 40% HNO3 concentration; (**g**) concentrate-3 and 10% HNO3 concentration; (**h**) concentrate-3 and 20% HNO3 concentration; (**i**) concentrate-3 and 40% HNO3 concentration.

It can be concluded that temperature has a significant effect on iron recovery. Increasing the temperature from 70 to 85 ◦C effected an increase in iron extraction from concentrate-1 from 35.53 to 48.52% after 1 h of leaching with a 10% solution of HNO3. A similar effect of temperature on iron extraction was observed for all types of concentrates and HNO3 concentrations.

It is obvious that the increase in the concentration of nitric acid should accelerate the reaction rate of sulfide oxidation, since the excess acid facilitates the diffusion of the reagent to the reaction interface. The data on Figure 3 allows one to conclude that the concentration of HNO3 has approximately the same effect on iron extraction as temperature. Increasing the concentration of HNO3 from 10 to 40% for all temperatures and concentrates can increase the degree of iron extraction by an average of 20%.

The sulfur content in the concentrate has the greatest effect on the extraction of iron. Thus, an increase in sulfur content from 8 to 22% makes it possible to increase iron recovery into the solution by an average of 45% at all temperatures and concentrations of HNO3. This is most likely due to the fact that the increase in the content of sulfides in the raw material make easier the diffusion of the reagent to them even on the last stage of the process.

Thus, particular parameters must be in place to achieve a more complete extraction of iron, and accordingly, the oxidation of sulfides for different samples of the concentrate. For example, a low-sulfur concentrate requires the maximum temperature at a high concentration of HNO3, while for a high-sulfur concentrate, a temperature of 70 ◦C and a 10% HNO3 solution is sufficient.

## *3.3. Characteristics of Solid Residue*

The cake obtained as a result of HNO3 leaching of the samples was subjected to scanning electron microscopy to study changes in the morphology of the sample in the course of the leaching process. Figure 4 shows micrographs of the original sample and the cake obtained by leaching.

**Figure 4.** Scanning electron microscopy (SEM) images of the original sample of concentrate-3 (**a**) and the cake obtained by leaching (**b**).

As can be seen from Figures 2 and 4a, the original concentrate consists of particles ranging in size from 1 to 100 μm, and the surface of the particles is rather smooth. In the course of leaching (Figure 4b), a large number of cavities form on the surface of the particles, which is associated with the dissolution of sulfide minerals. Figure 5 shows micrographs of solid residue of concentrate-3 (points 001-010) and concentrate-1 (points 011-019), differing by the degree of extraction.

Energy dispersive spectroscopy was used to determine the chemical composition at different points of the samples (Table 4).

**Figure 5.** SEM images of solid residue with the EDS points: (**a**) solid residue of concentrate-3 leaching and (**b**) solid residue of concentrate-1 leaching.


**Table 4.** Results of energy dispersive spectroscopy analysis, wt.%.

The measurement results in Table 4 confirm the presence of a large amount of unoxidized sulfides in the cake of the first sample and the almost complete absence of sulfur in the second cake, which agrees well with the results of the analysis of the liquid phase. Therefore, the surface of the particles of the concentrate-3 cake does not have a layer formed by the reaction product, elemental sulfur, which is also confirmed by the results of X-ray phase analysis, as shown in Figure 6. The absence of the reaction product on the surface of the concentrate-3 solid residue also explains the faster kinetics of leaching of the concentrate with high sulfur content because diffusion limitation is lower in this case. The absence of elemental sulfur could be attributed to the high intensity of the process, amount of the heat of exothermic reactions and the amount of gas produced that can lead to a higher oxidation potential.

**Figure 6.** XRD patterns of the solid residue of the concentrate-3 and concentrate-1.

## *3.4. Kinetic Model*

Since the degree of sulfide oxidation is greatly influenced by the temperature and concentration of HNO3, the leaching of refractory gold-containing concentrates with HNO3 can be controlled by diffusion as well as by kinetic stages. That is, the slowest stage can be the reagent diffusion towards the reaction surface as well as the chemical reaction itself. To determine the limiting step, it is necessary to conduct a study of the kinetics.

The shrinking core model (SCM) is generally used to describe the kinetics of heterogeneous reactions involving non-porous materials. The SCM assumes that the process rate is controlled either by the diffusion of the reactant to the surface through the diffusion layer (a liquid film), or by the diffusion through the product layer, or by a surface chemical reaction. During leaching, the inert layer of solids shrinks toward the center. A porous film of the reaction product is formed around the inert core.

The slowest stage with the greatest resistance is the limiting step, and its intensification allows to increase the leaching efficiency.

The equations describing the several limiting stages of SCM [45] are given in Table 5. A large number of studies [40,42] show that the new version of SCM proposed by Dickenson and Heal [46] may be preferable to describe the kinetics of leaching reactions controlled by interfacial transfer and diffusion through the product layer (Equation (7) in Table 5).


**Table 5.** The shrinking core model (SCM) equations [45].

sp—spherical particles, pp—plate particles, cp—cylinder particles, *k*—a chemical constant, *X*—the degree of iron recovery into the solution, and *t*—the leaching time.

According to Equation (7) in Table 5, if the interfacial transfer and diffusion through the product layer represent the limiting stage, then the function 1/3ln(1 − *X*) + [(1 − *X*) <sup>−</sup>1/<sup>3</sup> <sup>−</sup> 1] on the time "*t*" will be a straight line with the slope angle "*k*".

For the kinetic analysis using SCM, the equations for leaching of refractory gold-containing concentrate-1 at a 10% concentration of HNO3, presented in Table 5, were evaluated. The obtained data made it possible to determine the correlation coefficient (*R*2) showing the average square deviation of the experimental data from the straight line. The results of the calculations are shown in Table 6.


**Table 6.** SCM equations fitting.

As can be seen from the data obtained, SCM Equations (4)–(6) in Table 6 are poorly suited to describing these leaching reactions, since the correlation coefficient is less than 0.9. It is also obvious that the kinetic data best of all correspond to the new shrinking core model at all temperatures, which indicates diffusion limitations during leaching reactions.

The slope *k* for each straight line obtained by substituting the experimental data of leaching concentrate-1 with a 10% HNO3 solution into the SCM Equation (Figure 7a) was calculated. Then Arrhenius plots for the dependence of ln*k* on inverse temperature (Figure 7b) were used. Building a straight line *y* = *ax* + *b* in this plane of coordinates made it possible to find the coefficient *a*, which determines the slope of straight line. Knowing the slopes in these coordinates, allowed to find the apparent activation energy of 60.276 kJ/mol, using Arrhenius law. According to the literature data [39,47], a high value of the activation energy is not always representative of the kinetic controlled reaction.

**Figure 7.** Calculation of slope k (**a**) and dependence of ln*k*-1/*T* (**b**) for leaching concentrate-1 with 10% HNO3.

The slope of the straight lines (Figure 8a) obtained by substituting the concentrate-1 leaching results into the SCM equation at 70 ◦C at various concentrations of HNO3 was plotted as ln*k*-ln(HNO3) to determine the order with respect to nitric acid (Figure 8b). The resulting empirical order with respect to nitric acid concentration was 0.837.

**Figure 8.** Calculation of slope *k* (**a**) and dependence ln*k*-ln(HNO3) (**b**) for leaching conc-1 at 70 ◦C.

In the same way, we determined the empirical order with respect to sulfur content in the concentrate by plotting the dependence ln*k*-ln*S* for leaching various concentrates with a 10% solution of nitric acid at 70 ◦C (Figure 9). The order with respect to sulfur was 2.948, which confirms the conclusions about the pronounced effect of the sulfide content in the concentrate on the degree of iron recovery.

**Figure 9.** Calculation of slope *k* (**a**) and dependence of ln*k*-ln*S* (**b**) for leaching of various concentrates at a temperature of 70 ◦C with a 10% HNO3.

Substituting the Arrhenius equation (Equation (8)) into the equation of the new SCM model (Equation (7) in Table 5) gives Equation (9).

*k* = *k*oe−*<sup>E</sup> a* /*RT*, (8)

$$\mathbf{1}/3\text{ln}(1-X) + \left[\mathbf{1}(1-X)^{-1/3} - \mathbf{1}\right] = k\_o \mathbf{e}^{-E} \mathbf{\_d}^{\prime \mathbb{R}T} \cdot \mathbf{t}.\tag{9}$$

In Equation (9), the Arrhenius constant *k*<sup>o</sup> depends on the initial parameters of the process, including the initial sulfur content in the concentrate and the concentration of HNO3 in the initial solution; hence, Equation (9) can be represented as follows (Equation (10)).

$$\left[1/3\ln(1-X) + [(1-X)^{-1/3}-1] = k\_o (\text{HNO}\_3)^n (\text{S})^m e^{-\text{E}} \right]^{\text{RT}} \text{t},\tag{10}$$

where *n* and *m* are orders of concentration of HNO3 and sulfur content in the original concentrate, respectively.

Based on the previously obtained results, the following equation for leaching refractory gold-containing concentrate with HNO3 (Equation (11)) can be derived:

$$\left[1/3\ln(1-X) + \left[\left(1-X\right)^{-1/3} - 1\right] = k\_o \text{(HNO}\_3\text{)}^{0.837} \text{(S)}^{2.948} \text{e}^{-60276/RT} \cdot \text{t.} \tag{11}$$

Building off the Arrhenius plots for all temperatures, HNO3 concentrations and concentrates gives coefficients *b* of straight lines at a fixed slope *a* = 7250. The obtained values of the "*b*" coefficients and the corresponding correlation coefficients *R*<sup>2</sup> are shown in Table 7. To determine *k*o, an exponent from the "*a*" coefficient was taken, and then it was divided by (HNO3) 0.837(S)2.948. The determined average *k*o value was 87.811.


**Table 7.** Arrhenius constant determination.

Substituting this value into Equation (11) gives the following empirical equation describing the leaching process under study (Equation (12)):

$$\text{1/3In} (1 - X) + \text{[(1 - X)^{-1/3} - 1]} = 87.811 (\text{HNO}\_3)^{0.837} (\text{S})^{2.948} \text{e}^{-60276/RT} \cdot \text{t},\tag{12}$$

Table 8 shows the calculated correlation coefficient (*R*2) of experimental data and data obtained using Equation (12). As can be seen from the table, the obtained empirical expression shows a high degree of convergence with experimental data at almost all temperatures, concentrations of HNO3 and sulfur contents in the concentrate.


**Table 8.** Comparison of experimental data and the obtained empirical equation.

## **4. Conclusions**

The kinetics of dissolution of refractory sulfide gold-containing concentrates by a solution of HNO3 in the temperature range of 70–85 ◦C was investigated. The data obtained allowed one to draw the following conclusions:

Increasing the temperature from 70 to 85 ◦C effected an increase in iron extraction from concentrate-1 from 35.53 to 48.52% after 1 h of leaching with a 10% solution of HNO3. Increasing the concentration of HNO3 from 10 to 40% has same effect. Changing sulfur content in the concentrate produces a much greater effect. The highest degree of iron recovery in 1 h for high-sulfur concentrate was 98.10%, while recovery from low-sulfur concentrate under the same conditions was 67.83%.

EDS and XRD showed elemental sulfur and non-leached arsenopyrite in the residue from leaching of low-sulfur concentrate, while quartz was the main phase in the residue from leaching of high-sulfur concentrate.

The iron recovery from the concentrate is well described by a new shrinking core model, which suggests that the speed of the process is controlled by interfacial diffusion and diffusion through the product layer. The calculated apparent activation energy was 60.276 kJ/mol, and the reaction order with respect to the initial concentration of HNO3 was 0.837; the reaction order with respect to the initial S content in the concentrate was 2.948. The obtained data allowed us to derive a semi-empirical equation describing the leaching kinetics of iron:

> 1/3ln(1 − *X*) + [(1 − *X*) <sup>−</sup>1/<sup>3</sup> <sup>−</sup> 1] <sup>=</sup> 87.811(HNO3) 0.837(S)2.948e−60276/*RT*·*<sup>t</sup>*

Comparison of the experimental data obtained in the whole range of the studied parameters and the derived equation showed high convergence of the results. Thus, it can be concluded that the increase of sulfur content in the concentrate can be used to ensure more energy-efficient oxidation of sulfide minerals. The focus of our further research will be the study of HNO3 regeneration and methods of arsenic disposal in the form of environmentally friendly compounds. The new study will aim at finding the optimal conditions for the process and could lay the basis for the development of an alternative commercial technology.

**Author Contributions:** Conceptualization, D.A.R.; methodology, K.A.K. and A.A.S.; validation, O.A.D., R.E.R.; formal analysis, K.A.K., D.A.R.; investigation, A.A.S. and O.A.D.; resources, R.E.R., D.A.R.; data curation, D.A.R.; writing—original draft preparation, A.A.S.; writing—review and editing, K.A.K., D.A.R.; visualization, R.E.R., O.A.D.; supervision, A.A.S.; project administration, D.A.R.; funding acquisition, D.A.R.

**Funding:** The research was funded by the Russian Science Foundation, grant number 18-19-00186. The SEM/EDS and microprobe analysis were funded by State Assignment, grant number 11.4797.2017/8.9.

**Acknowledgments:** Ekaterinburg Non-Ferrous Metal Processing Plant JSC are acknowledged for providing materials. Technicians at Ural Branch of Russian Academy of Sciences are acknowledged for their assistance with XRD, XRF, SEM, EDS, and ICP-MS analysis.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

## **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Enhanced Desilication of High Alumina Fly Ash by Combining Physical and Chemical Activation**

## **Yanbing Gong 1,2, Junmin Sun 2, Shu-Ying Sun 3, Guozhi Lu <sup>1</sup> and Ting-An Zhang 1,\***


Received: 14 March 2019; Accepted: 2 April 2019; Published: 4 April 2019

**Abstract:** In this work, a physical–chemical activation desilication process was proposed to extract silica from high alumina fly ash (HAFA). The effects of fly ash size, hydrochloric acid concentration, acid activation time, and reaction temperature on the desilication efficiency were investigated comprehensively. The phase and morphology of the original fly ash and desilicated fly ash were analyzed by X-ray diffraction (XRD) and scanning electron microscopy–energy-dispersive X-ray spectroscopy (SEM-EDS). Compared with the traditional desilication process, the physical–chemical activation desilication efficiency is further increased from 38.4% to 53.2% under the optimal conditions. Additionally, the kinetic rules and equations were confirmed by the experimental data fitting with shrinking core model of liquid–solid multiphase reaction. Kinetic studies show that the enhanced desilication process is divided into two processes, and both steps of the two-step reaction is controlled by chemical reaction, and the earlier stage activation energy is 52.05 kJ/mol and the later stage activation energy is 58.45 kJ/mol. The results of mechanism analysis show that physical activation breaks the link between the crystalline phase and the amorphous phase, and then a small amount of alkali-soluble alumina in the amorphous phase is removed by acid activation, thereby suppressing the generation of side reactions of the zeolite phase.

**Keywords:** high alumina fly ash; desilication rate; physical–chemical activation; alumina silica mass ratio; kinetics

## **1. Introduction**

High alumina fly ash (HAFA) is a byproduct of high-temperature combustion of high-alumina coal in thermal power plants. The annual emission of HAFA is approximately 30 million tons [1]. Except for a small amount that is utilized, most of the HAFA is currently stockpiled in northwestern China, not only occupying farmland, but also giving rise to severe pollution of water, atmosphere, and soil [2,3]. On the other hand, the high alumina content (approximately 50%) in high alumina fly ash makes it a valuable recycling resource. Furthermore, HAFA contains critical metals such as gallium [4], lithium [5], and rare earths [6,7] that can be extracted in the alumina extraction process as byproducts. Many extraction techniques, such as predesilication soda lime sintering [8], limestone sintering [9], acid leaching [10,11], ammonium sulfate [12] and submolten salt methods [13], have been developed to extract alumina from HAFA. Only the predesilication soda lime sintering process was carried out at an industrial scale (200 kt per year) by China Datang Corporation in the province of

Inner Mongolia [14]. This formed a circular economy industry (shown in Figure 1) in Tuoketou County, Inner Mongolia. The recycling route of thermal power plant HAFA is presented in Figure 1.

**Figure 1.** Recycling route of thermal power plant high alumina fly ash (HAFA).

Here, a traditional alkali predesilication method was used to obtain desilicated high alumina fly ash (DHAFA) and sodium silicate solution. The main problem for this technology is that the Al2O3/SiO2 mass ratio (A/S) is only elevated to 1.7–1.9. Not only does this make the desilication rate lower, it also increases the difficulty of extraction of alumina by the soda lime sintering process [15]. Because of the low A/S of DHAFA, the main reaction equipment rotary kiln often scabs the inner wall, forming large clinker eggs, and forming hard clinker [16], giving rise to the consequent decrease of the alumina extraction rate, damage to equipment, and increased production cost. Therefore, it is necessary to improve the desilication rate.

$$2\text{NaOH} + \text{SiO}\_2 \text{ (amorphous)} = \text{Na}\_2\text{SiO}\_3 + \text{H}\_2\text{O (main reaction)}\tag{1}$$

$$10\text{Na}\_2\text{SiO}\_3 + 3\text{Al}\_2\text{O}\_3 \text{ (amorphous)} + 19\text{H}\_2\text{O} = \text{Na}\_6\text{Al}\_6\text{Si}\_{10}\text{O}\_{32} \cdot 12\text{H}\_2\text{O} \downarrow + 14\text{NaOH} \text{ (side reaction)} \text{ (2)}$$

Many researchers have reported that the amorphous SiO2 in the high alumina fly ash can be selectively dissolved in an NaOH solution [17], thereby leading to a decrease in the amount of silicon–calcium slag generated during the subsequent extraction process of Al2O3 and achieving the efficient recovery of the silicon resource [18,19]. These studies also indicated that the final products were significantly affected by the reaction conditions [20,21]. The mechanism of traditional predesilication can be clearly explained as follows: the amorphous SiO2 in the HAFA dissolved in an NaOH solution and formed sodium silicate as the main reaction (Equation (1)), and meanwhile, a small amount of Al2O3 was dissolved in an NaOH solution and formed sodium aluminate, and then sodium silicate reacted with sodium aluminate and formed zeolite as a side reaction (Equation (2)) that inhibits the desilication rate. Zhang et al. [22] has investigated a new process to improve the desilication rate of HAFA. The process is divided into three steps; traditional desilication, acid activation, and secondary desilication. Though the A/S of HAFA is raised from 1.2 to 2.85, the acid activation process consumes a large amount of acid to neutralize the alkali of zeolite generated in reaction Equation (2), so it is uneconomical. A desilication process after 12-h alkali pretreatment [23] was used to improve the A/S of DHAFA from 1.92 to 2.51, and the mechanism can be explained as the zeolite P generation in the pretreatment process blocking the formation of hydroxysodalite in the desilication process. It was found that the desilicated slurry can easily settle at the bottom of the tank and is difficult to filter after a long time of stirring in actual industrial production. A mechanism–chemical synergistic activation desilication process is used to prepare mullite ceramic materials [24]. The process uses harsh reaction conditions to remove the impurities in HAFA. The mechanism [25] has been analyzed. However, the specific effect of the reaction conditions and kinetics of desilication process were not described.

Based on the previous work [21–23], in this study, a novel process based on physical and chemical activation was developed in order to inhibit the side reaction during desilication and to increase the desilication efficiency. This process is moderate, controllable, and easy to transfer to industrial use. Moreover, the Al2O3/SiO2 ratio in desilication fly ash was significantly improved, which is beneficial to the subsequent Al2O3 extraction process. The effect of the activation reaction conditions on the desilication efficiency and its mechanism were investigated, and the kinetic rules and equations of the SiO2 leaching process were confirmed.

## **2. Materials and Methods**

### *2.1. Materials*

Hydrochloric acid (Beijing Chemical Works, Beijing, China, 36–38 wt. %) and NaOH (Sinopharm Chemical Reagent Co., Ltd., Shanghai, China, ≥96 wt. %) used in this study were of the analytical grade. HAFA was generated by a coal-fired boiler of a power plant of the Inner Mongolia Datang International Tuoketuo Power Generation Co., Ltd., Hohhot, China.

Table 1 summarizes the chemical composition of the original fly ash. The metal component contents were evaluated by chemical analysis methods. Table 1 shows that Al2O3 and SiO2 were the main components of the fly ash. Small amounts of iron oxide and calcium oxide are also present. Thus, this kind of fly ash had a high value for aluminum and silicon production. Additionally, the original fly ash contains a certain amount of rare metals Ga and Li.

**Table 1.** Chemical composition of HAFA.


## *2.2. Processes and Methods*

The physical–chemical activation desilication process includes three steps: "physical activation", "chemical activation", and "alkali predesilication".

During physical activation, the HAFA was milled for different times (5, 15, 30, 45, and 60 s) in a circular disk milling machine with a filling rate of approximately 60%.

During chemical activation, the milling HAFA was activated by hydrochloric acid with different acid concentration (4%, 6%, 8%, and 10%), reaction temperatures (30, 60, and 90 ◦C), and reaction times (0.5, 1, 1.5, 2, and 2.5 h) under a determined *L/S* ratio (equal to 4). The slurries were then filtered to obtain solid pretreatment HAFA (PHAFA).

During alkali predesilication, the original fly ash or physical–chemical activated solid was mixed with NaOH solutions with different alkali concentrations (Na2O concentrations of 100, 120, 140, 160 g/L), reaction times (10, 20, 30, 60, 90, 120, 150, and 180 min) under the determined conditions (*T* = 95 ◦C, *L/S* = 3.5) during the alkali predesilication. The slurries were filtered to obtain DHAFA or intensified DHAFA (IDHAFA).

The silica leaching ratio was calculated as follows:

$$\text{v}\_{1}\text{SiO}\_{2} = \frac{WD - W\_{1}D\_{1}}{WD} \times 100\% \tag{3}$$

where η is the desilication rate, *W* is the mass of HAFA, *W*<sup>1</sup> is the mass of DHAFA, *D* is the percentage content of SiO2 in HAFA, and *D*<sup>1</sup> is the percentage content of SiO2 in DHAFA or IDHAFA.

## *2.3. Characterization*

The chemical composition of HAFA was analyzed by chemical analysis methods. Particle size distribution of HAFA was evaluated using a Sympatec GmbH particle size analyzer (OASIS/Dry, Sympatec GmbH, Clausthal-Zellerfeld, Germany). The surface structure and the distribution of the elements were observed by scanning electron microscopy (JEOL5800SV, JEOL, Tokyo, Japan). The phase composition of HAFA was evaluated by X-ray diffraction (X' Pert PRO MPD, PANalytical B.V., Almelo, The Netherlands). The specific surface area and pore size analysis were performed using a Nova4000e high-speed surface tester (Nova4000e, Quantachrome, Boynton Beach, FL, USA). The molybdenum blue colorimetric method can be used to examine the content of SiO2 in the filtrate.

## **3. Results and Discussion**

## *3.1. Effects of Physical Activation on the Desilication Rate*

Figure 2 shows the effect of physical activation time on the desilication rate, and A/S and Na2O content of IDHAFA under acid activation conditions (*T* = 90 ◦C, acid concentration 6%, *L/S* = 4 and *t* = 1 h) and alkali predesilication conditions (*T* = 95 ◦C, *C* = 120 g/L, *L/S* = 3, and *t* = 2.5 h). It is observed that the desilication rate of HAFA is as high as 47% only due to the chemical activation. However, the desilication rate of HAFA is further improved to 51.9% after combined physical and chemical activation. Moreover, the alkali content of IDHAFA is decreased, indicating that the side reaction of the desilication process is suppressed, thereby improving the desilication rate. Considering the cost and effect of physical activation data presented in Table 2, the optimal activation time is 15 s.

**Figure 2.** Effect of physical activation time: (**a**) desilication rate, (**b**) A/S and Na2O content of IDHAFA.


**Table 2.** Effect of physical activation time on the particle size of HAFA (μm).

## *3.2. Effects of Chemical Activation on the Desilication Rate*

The data for the effects of acid concentration, reaction temperature, and reaction time on the desilication rate, A/S of DHAFA and Na2O content of IDHAFA for physical activation time of 15 s, and alkali predesilication conditions (*T* = 95 ◦C, *C* = 120 g/L, *L/S* = 3, and *t* = 2.5 h) are shown in Figures 3 and 4.

**Figure 3.** Effect of concentration of acid: (**a**) desilication rate, (**b**) A/S of intensified desilicated high alumina fly ash (IDHAFA), (**c**) Na2O content of IDHAFA.

**Figure 4.** Effect of acid activation time (*T* = 95 ◦C): (**a**) desilication rate, (**b**) A/S and Na2O of IDHAFA.

It is observed from Figure 3a that the desilication rate of the HAFA after acid activation is greatly improved compared with the conventional predesilication. The desilication rate is also improved with increasing temperature. The desilication rate is enhanced by more than 50% when the temperature is 90 ◦C. In addition, an increase in the acid concentration also enhances the desilication rate. It is observed from Figure 3b,c that the variation of the IDHAFA aluminum: silicon ratio is completely consistent with the desilication rate, but the change of the sodium oxide content in the IDHAFA shows a completely opposite trend to that of the desilication rate. This shows that the desilication fly ash aluminum: silicon ratio can reflect the desilication rate, while the sodium oxide content indicates the degree of the side reaction, which is inversely proportional to the desilication rate.

Figure 4a shows that the effect of the acid activation time on the desilication rate is also relatively large. In the range of 0–2 h, the desilication rate increases strongly with increased acid activation time. Flocculent silica gel was produced in the chemical activation process after 2 h, which makes the slurry becomes very viscous and difficult to filter, and the filter cake contains a large amount of acid, resulting in lower efficiency in desilication. It is observed from Figure 4b that the change in the desilication rate is consistent with the change in the ratio of IDHAFA to aluminum and silicon, and at the same time with the desilication powder. The change in the content of sodium oxide in IDHAFA is inversely proportional. Based on these experiments, it can be concluded that the optimal conditions for acid activation are: temperature of 90 ◦C, acid concentration of 6%, liquid: solid ratio of 4:1, and acid activation reaction time of 2 h. After chemical activation under these conditions, the metal ions are partially leached into the solution: and the dissolution rates of iron oxide and calcium oxide were 37.4% and 38.2%, respectively. Meanwhile, Al2O3 reached a dissolution rate of approximately 3.0%, but SiO2 is hardly dissolved during the hydrochloric acid activation.

## *3.3. Effects of Alkali Predesilication on the Desilication Rate*

Figures 5 and 6 show the alkali concentration, reaction time, and temperature effects on the direct desilication of the original fly ash and desilication after activation for the physical activation time of 15 s, and acid activation conditions (*T* = 90 ◦C, acid concentration 6%, *L/S* = 4 and *t* = 2 h).

**Figure 5.** Effect of alkali concentration on the desilication rate of IDHAFA and DHAFA (Predesilication conditions: *T* = 95 ◦C, *L/S* = 3.5, and *t* = 150 min).

It can be concluded from Figure 5 that the improvement of desilication efficiency after enhanced desilication is obvious. Generally speaking, increasing the concentration of sodium hydroxide is beneficial to the progress of reaction Equation (1). However, a high concentration of sodium hydroxide will also promote the progress of reaction Equation (2), and this is detrimental to desilication. According to the results in Figure 5, C(Na2O) = 140 g/L was required for the maximal desilication rate.

**Figure 6.** Effect of reaction time and temperature on the desilication rate of IDHAFA and DHAFA (Predesilication conditions: *C*(Na2O) = 140 g/L, and *L/S* = 3.5).

It can be seen from Figure 6 that the effect of temperature on the desilication efficiency is very large. Although the enhanced desilication rate is much higher than the direct desilication at 95 ◦C, the desilication rate is greatly reduced at 75 ◦C. In the 0–160 min period, the desilication rate increases with the increase of reaction time. However, as the reaction time increases, the reaction rate continues to decrease. The change of reaction rate divides the desilication reaction into two stages.

The data shown in Figures 5 and 6 clearly illustrate the great difference between the two desilication methods. Compared with direct desilication, the desilication rate after activation is greatly improved from 38.4% to 53.2% under the predesilication conditions (*T* = 95 ◦C, *C*(Na2O) = 140 g/L, *L/S* = 3.5, and *t* = 2.5 h).

## *3.4. Kinetics of the Desilication Process*

As already mentioned, the reaction Equation (1) is the main reaction of the desilication reaction process. In order to examine the SiO2 leaching kinetics of the enhanced desilication process, a shrinking core model of a liquid–solid multiphase reaction was used to fit the SiO2 leaching kinetics data. According to the reaction rate changing node in Figure 6, the SiO2 leaching process is divided into two stages, and the fitting data are shown in Figure 7.

**Figure 7.** Mathematical fitting of data of SiO2 leaching kinetics in enhanced desilication process: (**a**) The earlier stage, (**b**) The later stage.

There are two stages in the traditional desilication reaction process; the earlier stage is limited by the surface reaction, while in the later stage, the internal diffusion on the solid product layer is the rate-controlling step [26]. But, the physical–chemical activation desilication is significantly different. As exhibited in Figure 7a,b, the desilication process was controlled by chemical reaction in 0–160 min, in the temperature range of 75–95 ◦C, and both stages are controlled by chemical reaction. Under experimental conditions, the kinetic equation can be written as

$$(1 - (1 - X\_{\mathcal{B}})^{1/3} = k\_1 t \tag{4}$$

$$(1 - (1 - X\_{\mathcal{B}})^{1/3} = k\_2 t \tag{5}$$

where *X*<sup>B</sup> is the desilication rate, *t* is the reaction time, *k*<sup>1</sup> is the rate constant of the earlier stage, *k*<sup>2</sup> is the rate constant of the later stage.

According to the Arrhenius equation ln*k* = ln*A* − *E*a/R*T*, Plot ln*k* to 1/*T*, as exhibited in Figure 8, both of the two stages have a good linear relationship. Calculated based on the slope of the fitted curve, the earlier stage's activation energy is 52.05 kJ/mol, the later stage's activation energy is 58.45 kJ/mol.

**Figure 8.** Arrhenius plot of SiO2 leaching kinetics in the enhanced desilication process.

In the traditional desilication process, as the reaction time extends, the formation of zeolite in the side reaction increases, and a film is gradually formed in the fly ash particles. Therefore, the reaction is controlled by chemical reaction in the earlier stage and by diffusion in the later stage. But, the amount of zeolite is greatly reduced by enhanced desilication, so diffusion control is eliminated, and the enhanced desilication process is only controlled by the chemical reaction, which indicates that the set-up of the previous kinetic model is correct.

## *3.5. Mechanism*

As shown in Figure 9, the phases of PHAFA and HAFA are basically the same, indicating that the physical–chemical activation process does not change the phase composition of HAFA. Moreover, according to the previous experimental results, the treatment only removed part of the iron oxide, calcium oxide, and a small amount of alumina; however, mullite, corundum, quartz, and glass-phase silica did not react, so there is no phase change before and after treatment. The bulge between 15◦ and 25◦ in the XRD patterns of both DHAFA and IDHAFA disappears because the activated amorphous SiO2 was dissolved by the alkali solution during desilication, and the new zeolite phase appears in DHAFA because the soluble alumina enters the desilication solution and reacts with sodium silicate. It is also observed that there are few zeolite peaks in IDHAFA because the amount of soluble alumina decreases during the activation process, weakening the spontaneous side reaction (zeolite).

**Figure 9.** X-ray diffraction (XRD) patterns of HAFA, DHAFA, PHAFA, and IDHAFA.

Figure 10 shows the obtained N2 adsorption isotherms and pore size distribution. It is observed that the specific surface areas are in the order of DHAFA > IDHAFA > PHAFA > HAFA. Very low N2 adsorption is observed in HAFA (surface area = 2.886 m2/g), but a high N2 adsorption is observed for PHAFA for relative pressure (P/P0) between 0.5–1.0 (surface area = 13.737 m2/g), IDHAFA (surface area = 18.519 m2/g), and DHAFA (surface area = 21.460 m2/g). The larger specific surface area of PHAFA than HAFA is ascribed to the removal of iron oxide and calcium oxide after physical and chemical activation. The increased surface area of IDHAFA and DHAFA may be related to the removal of amorphous silica wrapped in the mullite and the formation of a new zeolite phase. It also indicated that there are fewer mesopores in HAFA (4 nm) and PHAFA (3 nm); however, except for the smaller mesopores (3–4 nm) of DHAFA and IDHAFA, larger mesopores (13 nm) appeared in DHAFA, and even larger mesopores (10–14 nm) appeared in IDHAFA.

**Figure 10.** Nitrogen adsorption–desorption isotherms and pore size distribution of HAFA, DHAFA, PHAFA, and IDHAFA. (**a**) Nitrogen adsorption–desorption isotherms, (**b**) Pore size distribution.

Figure 11 shows the weight loss curve under conditions of protective gas air atmosphere and heating rate of 10 ◦C/min. It is observed that both HAFA and PHAFA have smaller weight loss because there are small amounts of unburned carbon removed. On the other hand, DHAFA and IDHAFA show higher weight losses due to the formation of byproduct zeolite containing crystalline water. When the temperature is higher than 700 ◦C, no weight loss was observed. It should be noted that the weight loss of IDHAFA is still less than that of DHAFA, which also indicates that the pretreatment suppresses the occurrence of side reactions of fly ash after desilication and reduces the formation of zeolite.

**Figure 11.** Thermogravimetric (TG) results for HAFA, DHAFA, PHAFA, and IDHAFA.

As demonstrated in Figure 12, the shape of the raw HAFA (Figure 12a) is more complicated. The spherical mullite and the irregular glass are intertwined and inlaid with each other, and the interface is not clear. After the conventional predesilication reaction, it is found that DHAFA (Figure 12c) is mainly composed of spherical mullite crystals, and the irregular glass phase essentially disappears. However, the spherical shape from the EDS analysis indicates that a large amount of the zeolite particles is covered, indicating a small amount of soluble oxidation in the conventional predesilication reaction. The aluminum reacted with sodium silicate solution re-enters DHAFA, resulting in a low desilication rate. After the physical and chemical activation treatment, it is observed that the mullite and the amorphous glass phase of PHAFA (Figure 12b) have been separated. This is highly beneficial for desilication. The surface of IDHAFA (Figure 12e) is smoother than the DHAFA (Figure 12c), and the particles in IDHAFA mainly consist of pure mullite crystals. Furthermore, the EDS analysis showed that the alkalinity of IDHAFA (Na 0.9 wt. %) (Figure 12f) is less than DHAFA (Na 2.39 wt. %) (Figure 12d), indicating that the side reaction of the zeolite (Na6Al6Si10O32·12H2O) was greatly reduced.

**Figure 12.** Scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDS) results: (**a**) HAFA, (**b**) PHAFA, (**c**,**d**) DHAFA, (**e**,**f**) IDHAFA.

As a conclusion, the mechanism of enhanced desilication can be summarized as follows.

First, the spherical mullite and the amorphous glass phase are separated by physical activation, and the reactivity of the SiO2 and Al2O3 in the glass phase is improved.

Second, the Al2O3 in the glass phase is dissolved and removed by hydrochloric acid.

$$\text{Al}\_2\text{O}\_3 \text{ (amorphous)} + \text{HCl} = \text{AlCl}\_3 + \text{H}\_2\text{O} \tag{6}$$

Third, the SiO2 in the glass phase is dissolved by sodium hydroxide with almost no side reactions to form zeolite (Na6Al6Si10O32·12H2O). Therefore, only the mullite, corundum, and quartz are left in the fly ash after enhanced desilication.

$$\text{\textbullet 2NaOH} + \text{SiO}\_2 \text{ (amorphous)} = \text{Na}\_2\text{SiO}\_3 + \text{H}\_2\text{O} \tag{7}$$

## **4. Conclusions**

Improvement in desilication efficiency is important not only for obtaining more silicon resources, but also for more economical and efficient extraction of alumina.

A new process that combined physical and chemical activation and predesilication has been proposed. By using physical and chemical activation, the desilication efficiency of HAFA increased from 38.4% to 53.2%. The reaction temperature has a significant effect on the desilication efficiency.

Kinetic study results show that the enhanced desilication process after physical and chemical activation fits with an unreacted shrinking core model of the liquid–solid reaction. The desilication reaction is divided into two stages. The earlier stage is controlled by chemical reaction and the activation energy is 52.05 kJ/mol and the later stage is also controlled by chemical reaction and the activation energy is 58.45 kJ/mol.

This result indicates that physical and chemical activation suppresses the formation of zeolite, thereby improving the desilication efficiency and further improving the *A/S* of the fly ash; results which are very advantageous for the next step of alumina extraction.

**Author Contributions:** Methodology, J.S.; experiments, data analyze and writing—original draft preparation, Y.G.; writing—review and editing S.-Y.S. and T.-A.Z.; formal analysis, G.L.

**Acknowledgments:** This research was funded by National Key R& D Program of China grant number 2017YFB0603103.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Oxidation of Thiosulfate with Oxygen Using Copper (II) as a Catalyst**

## **Juan Manuel González Lara 1, Francisco Patiño Cardona 2, Antonio Roca Vallmajor 3,\* and Montserrat Cruells Cadevall <sup>3</sup>**


Received: 2 March 2019; Accepted: 25 March 2019; Published: 28 March 2019

**Abstract:** Thiosulfate effluents are generated in the photography and radiography industrial sectors, and in a plant in which thiosulfates are used to recover the gold and silver contained in ores. Similar effluents also containing thiosulfate are those generated from the petrochemical, pharmaceutical and pigment sectors. In the future, the amounts of these effluents may increase, particularly if the cyanides used in the extraction of gold and silver from ores are substituted by thiosulfates, or if the same happens to electronic scrap or in metallic coating processes. This paper reports a study of the oxidation of thiosulfate, with oxygen using copper (II) as a catalyst, at a pH between 4 and 5. The basic idea is to avoid the formation of tetrathionate and polythionate, transforming the thiosulfate into sulfate. The nature of the reaction and a kinetic study of thiosulfate transformation, by reaction with oxygen and Cu2+ at a ppm level, are determined and reported. The best conditions were obtained at 60 ◦C, pH 5, with an initial concentration of copper of 53 ppm and an oxygen pressure of 1 atm. Under these conditions, the thiosulfate concentration was reduced from 1 g·L−<sup>1</sup> to less than 20 ppm in less than three hours.

**Keywords:** thiosulfate oxidation; kinetics; catalysis; oxygen

## **1. Introduction**

In metallurgy, there is growing interest in technology for the extraction of gold and silver that does not use cyanide, for application in minerals or in residues in which it is difficult to recover metals from leaching liquids, and also for environmental purposes. The dissolution of gold and silver by thiosulfate occurs via the formation of the corresponding metal complexes. Technological proposals for implementing this process include the use of a copper (II) salt as an oxidizing agent and ammonia for copper stabilization within the system. One advantage in the use of thiosulfates is the reduction of interaction with other cations, such as copper, arsenic, and antimony, in contrast to the case of cyanidation. The thiosulfate route for the recovery of gold and silver from ores using Cu (II) and ammonia is complex. The leaching rates are acceptable, but the consumption of reagents can be high due to thiosulfate degradation. There are also difficulties encountered during the recovery of the dissolved metals [1–11]. Several authors have proposed modifications of the process reported in the literature. An alternative process includes ferric-ethylenediaminetetraacetic acid (EDTA), ferric-oxalate, and ferric-citrate to avoid the difficulties with the cupric-tetra-amine additive, including excessive thiosulfate oxidation and high reagent costs [12], whereas other authors presented an analysis of the effect of EDTA, thiosulfate, and cupric ions on silver leaching kinetics [13]. At present, however, thiosulfate leaching of minerals is only applied commercially in the Goldstrike deposit of the Barrick Gold Corporation [8].

Meanwhile, every year, millions of tons of waste are generated worldwide from electrical and electronic products. These devices are an important source of secondary raw materials in the form of gold and other metals with high economic values [14,15]. The amount of gold on the printed circuit boards (PCBs) of mobile phones can reach 300–350 g·ton−<sup>1</sup> [16]. A leaching process using thiosulfate in an ammoniacal medium to recover the gold contained in these PCBs has also been reported [17]. The authors indicated that the leaching system offers promising opportunities for industrial application; moreover, optimization of the leaching system for recovering gold from such PCBs was also carried out [14].

One of the most important uses of thiosulfate solutions has been for application in the photographic industry (in fixing baths and others) to dissolve the silver halide not reduced to metallic silver during photographic or radiographic processes. Today, the industry still generates significant amounts of thiosulfate-based effluents that need to be detoxified for environmental purposes, and to recover the silver [18–21].

In recent decades, for gold and silver coatings, the sulfite-thiosulfate system has been proposed as a non-cyanide coating technology [22–27].

As well as the thiosulfate effluents generated in the photography and radiography industries, the plants that use thiosulfates to recover the gold and silver contained in ores, and also the petrochemical, pharmaceutical, and pigment industries, among others, all produce thiosulfate effluents. In the future, the amounts of these effluents may increase considerably, particularly if cyanides are substituted by thiosulfates in the extraction of gold and silver from their ores, in the recovery of metals contained in electronic scrap, or in metal coating processes. Therefore, an adequate process is needed to degrade the thiosulfate contained in the corresponding effluents.

Different processes for this degradation of thiosulfates have been proposed. A study of the oxidation of thiosulfates by oxygen, using synthetic sphalerite doped with transition metals [28], the oxidation of thiosulfate with H2O2 in a catalyzed reaction, being the oxidation products sulfite and sulfate [29], and the oxidation of thiosulfates contained in wastewater, with oxygen in the presence of UV light [30], have been carried out.

Our research team studied the transformation of thiosulfate using copper sulfate solutions. The amount of copper (II) salt used was almost the stoichiometric quantity for the formation of 1 mol CuS and 1 mol sulfates per mol thiosulfate. The transformation of thiosulfate using copper (II) sulfate was applied to an industrial fixing bath from the photographic industry; the final effluent contained less than 10 mg L−<sup>1</sup> of thiosulfate [31].

This paper reports a study of the oxidation of thiosulfate to sulfate, with oxygen using copper (II) as a catalyst. The reaction was developed at a pH between 4 and 5 and the idea is to avoid the formation of tetrationate or polythionate. The nature of the reaction and its kinetics, with oxygen and in this case with Cu (II) at the ppm level, are determined. The effects of temperature and the initial concentrations of H3O+, Cu2+, and S2O3 <sup>2</sup><sup>−</sup> on the thiosulfate transformation are also determined; and the effect of the partial pressure of oxygen is studied.

## **2. Materials and Methods**

## *2.1. Materials*

In the study of thiosulfate degradation, sodium thiosulfate solutions prepared from this salt, with 99% purity, and pentahydrate copper sulfate of the same purity were used. Oxygen with a purity of ≥99.995% (mol/mol) was also used.

## *2.2. Experimental Procedure for Thiosulfate Transformation Using Copper Sulfate*

The experiments were carried out in a 0.5 L flat-bottom temperature-controlled reactor with magnetic stirring. The pH was kept constant by adding the necessary amounts of 0.5 mol·L−<sup>1</sup> H2SO4 solution and 0.5 mol·L−<sup>1</sup> NaOH solution.

In each experiment, the stirring rate, pH, air/oxygen flow, and temperature were adjusted to the values required in each case. Samples were taken at different times and filtered; and the remaining thiosulfate in the liquids was analyzed by iodometry (iodine 0.1 mol·L−<sup>1</sup> was used for this purpose and a starch solution was added to determine the end point). Some samples were also analyzed via ionic chromatography. During the process, the solids generated were characterized by X-ray diffraction (XRD, PANanalytical, Almelo, The Netherlands) as well as by scanning electron microscopy (SEM, JEOL, Tokyo, Japan), and energy-dispersive X-ray analysis (EDS, Oxford Instruments, Abingdon, England). The rates of these processes are indicated from the *kexp* values obtained in each experiment (slope of the graph: conversion, *X*, versus time); the conversion was defined according to the expression:

$$X = \left[ \mathrm{S\_2O\_3}^{2-} \right]\_{\mathrm{transformed}} / \left[ \mathrm{S\_2O\_3}^{2-} \right]\_{\mathrm{initial}} \tag{1}$$

## **3. Results and Discussion**

Preliminary experiments involving thiosulfate oxidation by oxygen, but without the presence of Cu (II), led to very slow rates for the process.

## *3.1. The Nature of the Thiosulfate Transformation by Reaction with Cu (II) Sulfate and Oxygen*

Two experiments involving thiosulfate degradation using Cu (II) were carried out under the following conditions: temperature, 60 ◦C; stirring rate, 500 min<sup>−</sup>1; initial concentration of thiosulfate, 1 g·L−1; initial concentration of Cu2+, 8.4 × <sup>10</sup>−<sup>4</sup> mol·L−<sup>1</sup> (0.053 g·L−1); and oxygen pressure, 1 atm. The first experiment was carried out at pH 5, and the second at pH 4.

Figure 1 shows the results for thiosulfate transformation at the indicated pH values. The thiosulfate degradation was almost total for a reaction time of 160 min. Meanwhile, two different rates were obtained: a slower one during the first part of the process, and a faster one during the second part. The values obtained were similar in both cases, pH 5 and pH 4. At a conversion of nearly to 0.40 (both experiments), the change in the slope was observed as the reaction rate increased, and a black solid was formed: CuS as confirmed by XRD (Figure 2). Therefore, it seems that the presence of this CuS precipitate enhanced the reaction rate, indicating a catalytic effect on thiosulfate degradation of the contact with copper sulfide surface. The precipitate of copper sulfide consisted of aggregates of crystals with 1 μm size (SEM-EDS).

Final liquids were analyzed by ion chromatography, indicating that thiosulfates were oxidized to sulfates under the experimental conditions employed (see Figure 3). In the pH interval between 4 and 5, sulfates were detected; any other species, such as tetrathionate or polytionate, were not detected.

**Figure 1.** Nature of the reaction for thiosulfate transformation using Cu2+ and oxygen: [S2O3 <sup>2</sup>−]initial, 1 g·L<sup>−</sup>1; [Cu2+]initial, 0.053 g·L<sup>−</sup>1; pHconstant, 4 or 5; 60 ◦C.

**Figure 2.** Nature of the reaction: X-ray diffractogram of solid obtained during thiosulfate transformation: identified as CuS.

**Figure 3.** Chromatograms: (**a**) Initial solution; (**b**) Final solution.

When CuS had formed (second step), hydronium ions were generated and pH was kept constant by the addition of sodium hydroxide solution.

The amount of thiosulfate remaining in the solution at the end of the experiment was less than 20 mg·L−<sup>1</sup> in both cases. Table <sup>1</sup> includes the slopes of the graphs *<sup>X</sup>*thiosulfate vs. time (*kexp*), which is a measure of the transformation rate.


**Table 1.** Values of *kexp* of thiosulfate transformation in two steps.

Iodometry established the absence of sulfites, while tetrathionates or polythionates were not detected by chromatography. Thus, the initial thiosulfates were transformed to sulfates.

Under the employed experimental conditions, the reaction produced in the zone where the change in the slope occurs could be as follows:

$$\mathrm{S\_2O\_3}^{2-}\text{(aq)} + \mathrm{Cu}^{2+}\text{(aq)} + 3\mathrm{H\_2O} \Rightarrow \mathrm{SO\_4}^{2-}\text{(aq)} + \mathrm{CuS\_{(s)}} + 2\mathrm{H\_3O}^{+}\text{(aq)}.\tag{2}$$

This process is quite similar to the degradation process using large amounts of copper ions (0.21–0.85 g·L−<sup>1</sup> Cu). In that case, sulfates and copper sulfide were formed in a similar amount in mol, when >0.57 g·L−<sup>1</sup> Cu were used. After this, the copper sulfide formed, and had to be oxidized to copper sulfate prior to being reused in the next cycle [31]. By using 0.053 g·L−<sup>1</sup> Cu (in the present work) and oxygen, the reaction rates are slower, but the transformation of thiosulfate to sulfate takes place in more than 90% yield. In this case, the amount of copper sulfide that must be oxidized to sulfate is much lower than the work in which large amounts of copper were used.

Another difference between these two procedures is that when we used very low amounts of copper salts and oxygen, two rates were detected; whereas, when using larger amounts of copper (without oxygen), only one rate value (*kexp* = 0.013 min−1) was determined, and the thiosulfate degradation was completed in a shorter time. The work cited [31] also demonstrated that, in experiments carried out at pH 6 and pH 10, no thiosulfate transformation to sulfate was detected. Meanwhile, thiosulfates were easily decomposed at pH ≤ 4, giving HSO3 −, SO2, and elemental sulfur. For these reasons, the kinetic study in this work was carried out in the pH interval between 4 and 5.

## *3.2. Kinetic Study of the Transformation of Thiosulfate by Reaction with Oxygen and Cu2+*

The effects of temperature (40–80 ◦C), concentration of H3O<sup>+</sup> (1.0 × <sup>10</sup>−4–1.0 × <sup>10</sup>−<sup>5</sup> mol·L−1), initial Cu2+ (4.2 × <sup>10</sup>−4–1.68 × <sup>10</sup>−<sup>3</sup> mol·L−1), and initial S2O3 <sup>2</sup><sup>−</sup> (4.5 × <sup>10</sup>−3–9 × <sup>10</sup>−<sup>3</sup> mol·L−1), as well as an oxygen pressure of 0.2–1 atm, were determined.

Figure 4 shows the plot of conversion vs. time at different temperatures. Table 2 includes the experimental values of *kexp* for these experiments; the table includes the corresponding values corrected by the oxygen concentration at each temperature, because the oxygen concentration varies with temperature [32]. The activation energy was calculated according to the following expression:

$$(k\_{\rm exp} \text{/} \text{O}\_2 \text{I}) = \text{Aexp}\, (\text{E}\_\text{a} / \text{RT}). \tag{3}$$

Figure 5 includes the plot of ln(*kexp/[O2]*) vs. 1000/*T*. Two apparent activation energies Ea were obtained: for the first step, 41 kJ·mol−<sup>1</sup> and for the second, 33 kJ·mol<sup>−</sup>1.

For concentrations of H3O<sup>+</sup> varying from 1.0 × <sup>10</sup>−<sup>4</sup> to 1.0 × <sup>10</sup>−<sup>5</sup> mol·L−<sup>1</sup> H3O+ (pH between 4 and 5, Figure 6), the reaction rate, *kexp*, in the first step was between 0.0033 and 0.040 min−1; and, in the second step, it was between 0.070 and 0.0110. All the values obtained are similar; consequently, an apparent reaction order of ≈0 with respect to the hydronium concentration was obtained.

The effect of the copper concentration was studied in the interval between 1.68 × <sup>10</sup>−<sup>3</sup> and 4.20 × <sup>10</sup>−<sup>4</sup> mol·L−<sup>1</sup> Cu2+, that is, between 0.11 and 0.027 g·L−<sup>1</sup> Cu2+; Figure <sup>7</sup> shows the results. Figure <sup>8</sup> is a graph corresponding to the experiment performed at 1.68 × <sup>10</sup>−<sup>3</sup> M Cu2+, to show the

appearance of two stages, as in the other experiments; this does not appear in Figure 7. Table 3 includes the experimental constants for these experiments.

**Figure 4.** Effect of temperature on thiosulfate transformation: [S2O3 <sup>2</sup>−]initial,1g·L−1; [Cu2+]initial, 0.053 g·L<sup>−</sup>1; pHconstant, 5; O2 pressure, 1 atm.



**Figure 5.** Apparent activation energy (Ea) for thiosulfate transformation (first step): 41 kJ·mol−1, and (second step): 33 kJ·mol<sup>−</sup>1.

**Figure 6.** Effect of H3O<sup>+</sup> concentration on thiosulfate transformation: [S2O3 <sup>2</sup>−] initial,1g·L−1; [Cu2+] initial, 0.053 g·L<sup>−</sup>1; 60 ◦C; O2 pressure, 1 atm.

**Figure 7.** Effect of initial Cu2+ concentration on thiosulfate transformation: [S2O3 <sup>2</sup>−]initial,1g·L−1; pHconstant, 5; 60 ◦C; O2 pressure, 1 atm.

**Figure 8.** Effect of initial Cu2+ concentration on thiosulfate transformation: Experiment performed at 1.68 <sup>×</sup> <sup>10</sup>−<sup>3</sup> mol·L−<sup>1</sup> Cu2+ (0.107 g·L<sup>−</sup>1).


**Table 3.** Effect of copper concentration: values of the experimental constants.

Figure 9 is a graph of log *kexp* versus log initial [Cu2+]. Apparent reaction orders of 2.5 for the first step and 1.75 for the second step were obtained. These values show the important effect that the initial concentration of copper (II) has on the rate of transformation of thiosulfates in the interval of copper concentrations from 4.2 × <sup>10</sup>−<sup>4</sup> to 1.68 × <sup>10</sup>−<sup>3</sup> mol·L<sup>−</sup>1.

**Figure 9.** Effect of the initial Cu2+ concentration on thiosulfate transformation: order of the reaction.

The effect of the initial concentration of thiosulfates on their transformation rate was determined for the interval 4.5 × <sup>10</sup>−3–9 × <sup>10</sup>−<sup>3</sup> mol·L−<sup>1</sup> (0.50–1 g·L−1); Figure <sup>10</sup> includes the results. As the reaction rate decreases, as the initial thiosulfate concentration increases. When the molar ratio S2O3 <sup>2</sup>−/Cu2+ is ≤8, only one *kexp* value was obtained; for values of the molar ratio >10, two slopes appeared. Table 3 includes values of the log of the initial thiosulfate concentration and the log of the different values of *kexp*.

**Figure 10.** Effect of initial S2O3 <sup>2</sup><sup>−</sup> concentration on thiosulfate transformation: [Cu2+]initial, 0.053 g·L<sup>−</sup>1; pHconstant, 5; 60 ◦C; O2 pressure, 1 atm.

Table 4, lines 1 and 2, contain the *kexp* values corresponding to the experiments in which 0.5 and 0.75 g·L−<sup>1</sup> of thiosulfate were used; line 3 includes the average value of *kexp* obtained in the 1 <sup>g</sup>·L−<sup>1</sup> thiosulfate experiment (under these conditions, a change of slope was detected). Lines 4 and 5 correspond to the *kexp* values of the first and second stage, respectively, of the experiment with 1 g·L−<sup>1</sup> of thiosulfates. The apparent reaction order, considering the first three lines, is −2.8; if the experiment at 1 g·L−<sup>1</sup> is considered in two steps, the apparent order of the first step is −3.4 and the apparent order of the second is −2.6. When the thiosulfate concentration was increased, the reaction rate decreased significantly, because of the increases in the thiosulfate/Cu(II) ratio.


**Table 4.** Reaction order with respect to the initial thiosulfate concentration.

An additional experiment was carried out by decreasing the S2O3 <sup>2</sup>−/Cu2+ ratio, using 3.6 × <sup>10</sup>−<sup>2</sup> mol·L−<sup>1</sup> thiosulfate (4 g·L−<sup>1</sup> thiosulfate) and an initial Cu2+ concentration of 2.5 × <sup>10</sup>−<sup>3</sup> mol·L−<sup>1</sup> (initial thiosulfate/copper molar ratio of 14.2). The behavior was quite similar to that shown in Figure 8 (initial copper concentration 4.2 × <sup>10</sup>−<sup>4</sup> mol·L−1). The rate values obtained were: from 0 to 270 min, *kexp1* = 0.0020 and *kexp2* = 0.0083.

The effect of the partial pressure of oxygen was determined in the interval between 0 and 1 atm pressure; Figure 11 shows the results. The reaction rate increases as the partial pressure of oxygen increases. In the experiment without oxygen, a conversion of 0.33 was achieved after 480 min reaction time (First step: *kexp* = 0.0006 min<sup>−</sup>1). When air was used (partial pressure of oxygen = 0.2), a conversion of 0.50 was achieved at the same time (First step: *kexp* = 0.0010 min<sup>−</sup>1). Finally, with oxygen at a partial pressure of 1 atm, a conversion of 0.98 was obtained in 160 min (*kexp* = 0.0040 min−1, first step; and *kexp* = 0.0070 min−1, second step). Only in the last experiment was the degradation reaction almost completed. Consequently, an average value of reaction order near to 1 was obtained with respect to the partial pressure of oxygen. The reaction rate was multiplied by 4 when oxygen was used instead of air, i.e., approximately the same ratio of the oxygen partial pressure in air or in pure oxygen.

**Figure 11.** Effect of the partial pressure of oxygen on thiosulfate transformation: [S2O3 <sup>2</sup>−], 1 g·L−1; [Cu2+] initial, 0.053 g·L<sup>−</sup>1; pHconstant, 5; 60 ◦C.

The transformation of thiosulfate to sulfate using oxygen and copper (II) in dilute solution (25–100 ppm), at pH 5 and 60 ◦C, has potential applications in treating effluents for thiosulfate degradation. The small amounts of copper sulfide generated can be oxidized to copper sulfate and recycled to the next cycle of thiosulfate degradation.

## **4. Conclusions**

(1) In the degradation of thiosulfate by oxygen and small amounts of Cu2+ at pH between 4 and 5, a change in the slope of the graph of reaction rate (*kexp*) versus time was observed. This coincided with a precipitate of CuS being formed, and the reaction rate increased from this point. This implies a catalytic effect of contact with CuS on the thiosulfate degradation.

(2) The reaction produced in the zone where the change in the slope occurs could be as follows:

$$\text{S}\_2\text{O}\_3^{2-} \text{ (aq)} + \text{Cu}^{2+} \text{ (aq)} + 3\text{H}\_2\text{O} \Rightarrow \text{SO}\_4^{2-} \text{ (aq)} + \text{CuS}\_{\text{(s)}} + 2\text{H}\_3\text{O}^+ \text{ (aq)} \cdot \text{}$$

Analysis of the final liquids carried out by chromatography confirmed that the thiosulfates were oxidized to sulfates under the employed experimental conditions. The amount of thiosulfate remaining in solution at the end of the process was less than 20 mg·L<sup>−</sup>1.

(3) In the kinetic study of thiosulfate degradation, two apparent activation energies were obtained: 41 kJ·mol−<sup>1</sup> (in the first step, until a conversion of nearly 0.35) and 33 kJ·mol−<sup>1</sup> (in the second step after CuS formation). The effect of temperature on the degradation process using oxygen and Cu (II) at the ppm level is less than that observed when using larger amounts of Cu (II) without oxygen (98 kJ·mol<sup>−</sup>1).

(4) For a concentration of Cu (II) varying between 1.68 × <sup>10</sup>−<sup>3</sup> and 4.20 × <sup>10</sup>−<sup>4</sup> mol·L−<sup>1</sup> Cu2+ (107–27 ppm Cu in solution), the reaction rate increased with the Cu (II) concentration, giving an apparent reaction order near to 2 (2.5 for the first step and 1.75 for the second step).

(5) For a concentration of thiosulfate varying between 4.5 × <sup>10</sup>−<sup>3</sup> and 9 × <sup>10</sup>−<sup>3</sup> mol·L−<sup>1</sup> S2O3 <sup>2</sup>−, only one reaction rate (slope) was detected for molar ratios thiosulfate/Cu ≤8. Two rates (slopes) were detected for values of this ratio higher than 10. Apparent reaction orders of −3.4 (first step) and −2.6 (second step) were obtained, indicating that, when the thiosulfate concentration increases, the reaction rate decreases significantly, because of the increases in the thiosulfate/Cu(II) ratio.

(6) The effect of the partial pressure of oxygen was determined in the interval between 0 and 1 atm pressure. The reaction rate increases as the partial pressure of oxygen increases. An average apparent reaction order near to 1 was obtained.

(7) The transformation of thiosulfate to sulfate by oxygen and a dilute copper (II) solution (25–100 ppm), at pH 5 and 60 ◦C, has potential applications in treating effluents for thiosulfate degradation. The CuS generated during the process can be oxidized and recycled to the next cycle of thiosulfate degradation.

**Author Contributions:** A.R.V., M.C.C. and F.P.C. conceived and designed the experiments; J.M.G.L. performed the experiments; A.R.V., M.C.C., J.M.G.L. and F.P.C. analyzed the data; and A.R.V. and M.C.C. wrote the paper.

**Acknowledgments:** The authors wish to thank the *Centres Científics i Tecnològics* of the Universitat de Barcelona for their assistance with this work.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **List of Symbols**


## **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
