1 Introduction

Iron, being the fourth most abundant element in the earth’s crust, is contained to notable extents in common minerals of base metals like zinc and nickel. In the course of their extraction, it must be separated to ensure the feasibility of the different processes and to enable the respective recovery of the pure metals [1,2,3]. In hydrometallurgy, iron removal is commonly realized by precipitation methods, generating a purified solution for further processing and a precipitate as residue, which is usually named after the formed iron compound. In zinc production, mainly jarosite and goethite residues occur, while in nickel production, iron is also removed in the form of hematite or ferric hydroxide [3, 4].

With the climate crisis as one of the most pressing issues of our time, the need for cleaner technologies is on the rise and so is the demand for raw materials required for the transition. Zinc, being mainly applied in the galvanization of steel products, plays an important role in the corrosion protection of wind turbines while nickel demand increases due to its use in electric vehicle batteries, geothermal power plants, and hydrogen electrolysis [5, 6]. With zinc and nickel production growing, so do the amounts of generated residual materials, like iron precipitates. Nowadays, they are primarily discarded in tailing ponds, which not only poses an environmental hazard to their surroundings, but also results in an unexploited extraction potential of contained valuable metals [7,8,9,10]. Iron precipitation residues can contain considerable amounts of valuable elements that are partially regarded as critical raw materials. The base metals copper, lead, nickel, and zinc can be found in such residues, but also technology metals like gallium or indium and the precious metal silver [9]. The extraction of such elements out of the iron-containing matrix could help save primary resources, reduce waste volumes, and mitigate negative effects on the environment caused by their disposal [11]. In the case of critical metals, significant concentrations have been reported to mining and processing wastes over time. Enhancing the efforts of exploiting mining and processing wastes as valuable sources for such metals would, apart from aligning with circular economy goals, help secure future critical metal supplies [12].

Numerous investigations on possible processing strategies for iron precipitation residues have been conducted in the past decades, targeting the treatment of residual materials for environmentally safe disposal and the recovery of economically valuable elements. Nonetheless, the only industrially applied methods are in the field of immobilization and pyrometallurgy. Immobilization still leads to vast amounts of residues being landfilled, while contained metals are not recycled and therefore lost. Pyrometallurgical processes, on the other hand, enable the recovery of various elements, but are usually associated with the emission of greenhouse gases [13].

A novel approach consists of the chlorination of valuable elements and the subsequent volatilization of the formed chlorides. Chlorination is an increasingly important way of extracting nonferrous metals not only from ores or concentrates but also from industrial residues [14]. Chlorinating agents like gaseous chlorine and hydrogen chloride as well as chlorides of alkali and alkaline earth metals possess a high reactivity at moderate temperatures, selectivity in chlorination of the desired metal values, and an easy availability with low cost [15]. These properties have led to numerous applications in metallurgy on an industrial scale, like the extraction of titanium, rare earth elements, tantalum, niobium, or tin [16]. Additionally, chlorination-volatilization processes have been examined as a means of recycling metals from a wide variety of wastes and residues, e.g., LCD powder, PVC wires, or pyrite cinder [17,18,19]. In chlorination-volatilization approaches, a chlorinating agent is added to the metal-containing material, enabling the separation of volatile metal chlorides from the non-volatile matter [20]. To volatilize metals, the formation of the respective metal chloride must be thermodynamically favorable, and its vapor pressure must be high enough at the given temperature for volatilization. These two requirements are at the same time the basis for the selectivity of such processes: Various materials like Al2O3, SiO2, or Fe2O3 fail to fulfill the first one [18]. In the case of Fe2O3, for example, it would be necessary to provide a reducing agent like carbon in order to lower the oxygen potential sufficiently for a successful chlorination [15]. The second one, on the other hand, is often not met by chlorides like NaCl, KCl, or CaCl2, which makes the separation of valuable metals, whose chlorides do show the necessary vapor pressure, from the matrix possible [18].

2 Materials and Methods

The approach described in this paper aims at processing different iron precipitation residues with chloridic compounds in a thermal treatment process at temperatures in the range of up to 1100 °C, leading to a selective removal from valuable metals such as chlorine compounds. The fact that iron chlorinates poorly while other volatile metal chlorides are formed can be exploited for this innovative separation of valuable metals from jarosite or goethite residues. The thermodynamic principles and the behavior of the investigated residues under thermal treatment are discussed in more detail below.

2.1 Thermodynamic Considerations

The underlying idea is based on the one hand on the low process temperature compared to other metallurgical processes required for chlorination and on the other hand on the carbon addition and thus CO2 emission that is not required due to the direct chlorination reaction. Figure 1a shows the evaporation temperatures of a variety of relevant metals in their metallic, oxidic, and chloridic forms. The metals are found to form chlorides in different states of valence, which are considered in the comparison. Figure 1b shows the average temperature for metallic and oxidic compounds as well as for all depicted chlorides. This comparison illustrates that all investigated metals are more easily vaporized as chlorides and remain non-volatile in their oxidic state. The mean temperature for the evaporation of chlorides is significantly lower than that of the metals or oxides. This circumstance together with the instability of iron chloride, described later, is exploited for selective separation of valuables from iron precipitation residues.

Fig. 1
figure 1

Evaporation temperatures of a range of metals of interest in their metallic, oxidic, and chloridic forms [21]. Copyright 2023, The Minerals, Metals & Materials Society. Used with permission

Full size image

Jarosite is the mineralogic term for a certain group of basic iron sulfates with the general formula XFe3(SO4)2(OH)6. The placeholder X is generally represented by monovalent cations such as Na+, K+, NH4+, H3O+, Ag+, or ½ Pb+ [22]. Similar as in nature, different kinds of jarosite precipitates are formed depending on what cation is added to the process solution or what impurities or also valuable elements are present in the process solution. Alternatively, iron can be precipitated from sulfatic solutions also as goethite, with the formula FeO·OH. Both precipitation residues, either from zinc or nickel industry, can incorporate other valuable metals in their structure or co-precipitate them as sulfates, oxides, or hydroxides, due to decreasing solubility of the individual valuable element compounds during the iron precipitation process. Notables are concentrations of zinc, lead, silver, and indium, as well as nickel, the latter two being among the world’s critical elements in terms of security of supply [23, 24].

During a possible thermal treatment of these materials, a decomposition takes place before or in parallel with the actual treatment, leading to a splitting of the hydroxide and sulfate groups. This results in an easier processability of the simpler oxidic or sulfatic compounds as well as in an enrichment of the valuable metals remaining in the matrix. In jarosite with its structural formula XFe3(SO4)2(OH)6, water is bonded in the form of OH− and H3O+ (in the case of hydronium jarosite). The evaporation of moisture can be observed below 200 °C, the following splitting and removal of OH group follow up to 450 °C and between 600 and 850 °C, the SO3 is removed [25], as illustrated by Eqs. 1 to 3 exemplarily for a sodium jarosite. This decomposition can take place either in parallel with the chlorination in a process step or upstream in order to obtain the SO3 separately from the chlorine compounds formed.

$$\mathrm{Na}{\mathrm{Fe}}_3{\left(\mathrm{S}{\mathrm{O}}_4\right)}_2{\left(\mathrm{OH}\right)}_6\cdot \mathrm{m}{\mathrm{H}}_2\mathrm{O}\rightarrow \mathrm{Na}{\mathrm{Fe}}_3{\left(\mathrm{S}{\mathrm{O}}_4\right)}_2{\left(\mathrm{OH}\right)}_6+\mathrm{m}\left\{{\mathrm{H}}_2\mathrm{O}\right\}$$
(1)
$$\mathrm{Na}{\mathrm{Fe}}_3{\left(\mathrm{S}{\mathrm{O}}_4\right)}_2{\left(\mathrm{OH}\right)}_6\rightarrow \mathrm{Na}\mathrm{Fe}{\left({\mathrm{SO}}_4\right)}_2+{\mathrm{Fe}}_2{\mathrm{O}}_3+3\left\{{\mathrm{H}}_2\mathrm{O}\right\}$$
(2)
$$\mathrm{NaFe}{\left({\mathrm{SO}}_4\right)}_2\rightarrow 1/2{\mathrm{Na}}_2\mathrm{S}{\mathrm{O}}_4+1/2{\mathrm{Fe}}_2{\mathrm{O}}_3+3/2\left\{{\mathrm{SO}}_3\right\}$$
(3)

Goethite, with its mineralogical structure FeO∙OH, decomposes to hematite under the release of water in one single step. The decomposition start was observed at 200 °C with 96% completion at 400 °C [26]. This decomposition of the iron precipitation residues, either jarosite or goethite, is quite intentional, as it also allows any trapped valuable metals to be released and thus chlorinated separately from the iron oxide. Anyhow, since jarosite and goethite are precipitated from sulfuric acid solutions, sulfur represents a significant mass fraction in the residual matrix in both cases. Theoretically, all the valuable metals can therefore also be present as sulfate compounds. So, calculations were also carried out for the presence of valuable metal sulfates.

For the following thermodynamic calculations, mainly, the software FactSage but partly also HSC Chemistry were used. The FactSage Reaction Module offers the possibility of automatically considering elements and compounds in their most stable form at certain temperatures. Accordingly, any changes in the slopes of the various curves are due to the melting and evaporation of the components present. As a pre-stage to temperature-induced evaporation, it is necessary to convert the metals from their present form into chlorides. In the course of this work, thermodynamic calculations were used to determine chlorination using chlorine gas, gaseous hydrochloric acid, and the three metal chlorides AlCl3, FeCl3, and MgCl2, and in their most stable form. The reason for this is the possible dissociation of the chloride compound, the previous reaction with split-off hydrate water or the direct reaction. Sodium chloride, potassium chloride, and calcium chloride were excluded by tentative experiments but also thermodynamic calculations due to their high stability and therefore poor suitability to act as chlorination agents for others. For easier comparison, all reactions are normalized to a total amount of two atoms of chlorine participating in the reaction as given in Eq. 4 to Eq. 6, exemplarily for oxidic compounds. The term “Sca” stands for solid chlorination agent, for example, Mg in MgCl2. All calculations were performed with the FactSage Reaction Module (Database FactPS), taking into account all compounds in their most stable form at the respective temperatures.

$$1/\mathrm{y}\ {\mathrm{Me}}_{\mathrm{x}}{\mathrm{O}}_{\mathrm{y}}+2\ \mathrm{HCl}\leftrightarrow \mathrm{x}/\mathrm{y}\ \mathrm{MeCl_y}+{\mathrm{H}}_2\mathrm{O}$$
(4)
$$1/\mathrm{y}\ {\mathrm{Me}}_{\mathrm{x}}{\mathrm{O}}_{\mathrm{y}}+{\mathrm{Cl}}_2\leftrightarrow \mathrm{x}/\mathrm{y}\ \mathrm{MeCl_y}+\frac{1}{2}\ {\mathrm{O}}_2$$
(5)
$$1/\mathrm{y}\ {\mathrm{Me}}_{\mathrm{x}}{\mathrm{O}}_{\mathrm{y}}+2/\mathrm{z}\ \mathrm{Sca}{\mathrm{Cl}}_{\mathrm{z}}\leftrightarrow \mathrm{x}/\mathrm{y}\ \mathrm{M}\mathrm{e}{\mathrm{Cl}}_{2\mathrm{y}/\mathrm{z}}+2/\mathrm{z}\ \mathrm{M}{\mathrm{O}}_{\mathrm{z}/2}$$
(6)

Figure 2a and b are illustrating the results of the performed calculations, respectively the Gibbs energy of the reactions of various metal oxides (a) and sulfates (b) of interest with hydrochloric acid in the most stable form as a function of temperature. Most metal oxides are transformed to their corresponding chlorides, while iron oxide theoretically does not react to iron chloride above ~100 °C (red dotted line). In the case of present sulfate compounds, a temperature above 800–1100 °C is required, to shift the equilibrium to the product side, respectively chlorine compound. However, this thermodynamic consideration does not take into account any other furnace atmosphere, and with that, the stability of the sulfates themselves in, for example, oxidizing conditions. A possible decomposition of the sulfates could take place and thus result in a reaction of chloride compound with the oxide again. This fact applies to all subsequent considerations. Furthermore, the calculations are based on the equilibrium state and do not take into account kinetic effects as for instance the continuous removal of reaction products by the exchange of furnace atmosphere, e.g., gas-fired facilities or simple purging.

Fig. 2
figure 2

Gibbs energy of the reactions of valuable metal oxides (a) or metal sulfates (b) and hydrochloric acid [21]. Copyright 2023, The Minerals, Metals & Materials Society. Used with permission

Full size image

The comparison of Fig. 2a and b with Fig. 3a and b—illustrating the reaction with chlorine gas instead—draws a somewhat similar picture. In the case of metal oxides, the decrease in driving force with increasing temperature is less pronounced than for the reaction with HCl. Generally, copper and nickel oxides show a lower tendency to react than other metal oxides. The iron oxide is again not stable below 1200 °C in the case of oxides and 900 °C in the case of sulfates. All in all, the reaction either with gaseous chlorine or gaseous HCl shows very similar characteristics.

Fig. 3
figure 3

Gibbs energy of the reactions of valuable metal oxides (a) or metal sulfates (b) and gaseous chlorine [21]. Copyright 2023, The Minerals, Metals & Materials Society. Used with permission

Full size image

More interesting than the reaction with Cl2 or HCl seems to be the reaction with solid chlorine carriers, as these are by far easier to handle. Nevertheless, the previously mentioned reactions are of great importance as well since they can take place through a previous reaction of the solid chlorine carriers with hydrate water or moisture of the residual materials. In order to determine suitable solid chlorine carriers, stability calculations of various metal chlorides were performed. As the chlorination process is developed to operate in an oxidizing atmosphere, oxygen is meant to replace the chlorine molecules to form stable compounds with the metal. The stability of a compound at standardized conditions can be determined by its standard enthalpy of formation, ∆Hf0. Data was obtained from the main database of HSC Chemistry V10. Figure 4 shows values for ∆Hf0 for the formation of the chloridic and oxidic compounds for the elements Na, K, Ca, Mg, Al, and Fe. The lower the values for ∆Hf0, the higher the stability of the compound. It is evident that the chlorides for Na, K, and Ca are significantly more stable than their oxides. For Mg and Fe, the values are in a similar range. For Al, the property is different, showing a more negative value for the oxide.

Fig. 4
figure 4

Comparison of the standard enthalpy of formation for different oxides and chlorides and the respective stability factor

Full size image

In the lower plot in Fig. 4, a stability factor for further interpretation is given, which is calculated according to Equation 7. A value of 1, as it is approximately present for Mg and Fe, indicates that the chloride and oxide stability in standard conditions is similar. Noticeable are the low values for Na, K, and Ca showing a higher tendency for chlorine formation, respectively a lower tendency for liberation of chloride for reactions.

$$\mathrm{Stability}\ \mathrm{factor}=\frac{\Delta {H}_{\mathrm{f}}^0\left(\mathrm{chloride}\right)}{\Delta {H}_{\mathrm{f}}^0\left(\mathrm{oxide}\right)}$$
(7)

Another indicator for the stability of solid chlorine carriers is the Gibbs energy, ∆GR, of the formation of chloridic compounds. The Gibbs energy serves as an indicator if a reaction is occurring in the given direction or not. This is indicated by negative values. Positive values suggest that the reaction proceeds in the opposite direction. If ∆GR is 0, the reaction is in its chemical equilibrium. The thermochemical software FactSage was used to calculate the ∆GR values of the formation of chlorides from pure metals and gaseous chlorine as given in Fig. 5. Referring to the findings, which are summarized in Fig. 4, Na, K, and Ca form very stable chloridic compounds. As shown in Fig. 5, the depicted curves of those two metals show the highest driving force (lowest values for ∆GR) for their chlorination reaction. Fe, Al, and Mg form stable chlorides as well, but the effect is significantly less pronounced.

Fig. 5
figure 5

Gibbs energy for the formation of solid chlorination agents with Cl2(g)

Full size image

Based on those calculations, the decision for further investigations was taken for the three chlorides AlCl3, FeCl3, and MgCl2. All three are typically present in their hydrated form and, upon exposure to heat, react with the release of either gaseous chlorine or hydrochloric acid. This would lead to the aforementioned reactions. Nevertheless, also, the unhydrated forms can be present as well and therefore directly react with valuable metal compounds.

Figure 6a and b show the reactions with the solid chlorine carrier MgCl2. All reactions take place up to a temperature of 1100 °C, except the one with iron oxide. In the case of sulfate compounds, even iron sulfate reacts to its chloride, being strongly influenced by the stability of the second reaction product, which is magnesium sulfate instead of magnesium oxide.

Fig. 6
figure 6

Gibbs energy of the reactions of valuable metal oxides (a) or metal sulfates (b) and magnesium chloride [21]. Copyright 2023, The Minerals, Metals & Materials Society. Used with permission

Full size image

With AlCl3, all metals, including iron, can be chlorinated in the complete investigated temperature range, in some cases with high driving forces, as shown in Fig. 7a and b. Here, the conclusion is obvious that AlCl3 would not be suitable due to the possible formation of volatile ferric chloride, but formed ferric chloride can also react further as it is illustrated in Fig. 8.

Fig. 7
figure 7

Gibbs energy of the reactions of valuable metal oxides (a) or metal sulfates (b) and aluminum chloride [21]. Copyright 2023, The Minerals, Metals & Materials Society. Used with permission

Full size image
Fig. 8
figure 8

Gibbs energy of the reactions of valuable metal oxides (a) or metal sulfates (b) and iron chloride [21]. Copyright 2023, The Minerals, Metals & Materials Society. Used with permission

Full size image

The utilization of iron chloride as a solid chlorination agent has two possible advantages. The first is that its formation when using other solid chlorination agents is unavoidable anyway due to the huge amount of iron dominating the matrix. The second reason is that the formed reaction product, iron oxide, is not forming any new impurity in the already existing iron-containing matrix. The reaction of iron oxide with iron chloride is not illustrated in Fig. 8a and b as the reaction products are the same as the reactants, but it can be seen that in the case of metal oxides up to 1100 °C and in the case of sulfates up to a temperature of around 700 °C, all equilibria are on the product sides.

Summarizing, it can be stated that the chlorination of various valuable metals from an iron-containing matrix is viable based on the carried out thermodynamic calculations. Due to the low to completely non-existent stability of ferric chloride compared to the other metal chlorides considered, a satisfactory separation effect should thus be achievable.

2.2 Chlorination Experiments

The thermodynamic method development for selective chlorination was verified by two different practical approaches. In the first one, synthetic mixtures of the selected metals Ag, In, Pb, and Zn in their oxidic as well as sulfatic form and the three different metal chlorides AlCl3, FeCl3, and MgCl2 (in their hydrated form) were investigated by simultaneous thermal analysis. A heating rate of 20 K/min was used along with a purge gas rate of 0.3 l/min of synthetic air. All mixtures were prepared with a stoichiometric ratio of 1:1.5 between the valent metal compound and chloride to ensure complete chlorination. The assumed valuable metal chlorides are AgCl, InCl3, PbCl2, and ZnCl2.

The second set of practical experiments utilized industrial jarosite and goethite residues from zinc production. The chemical composition of the dried materials has been determined in an accredited laboratory. The results for the considered valuable metals Ag, Bi, Cu, Sn, Pb, and Zn as well as the matrix element iron are summarized in Table 1.

Table 1 Contents of valuable metals in the investigated industrial residues
Full size table

The prevailing iron phases in each residue were determined using the chemical analyses and information obtained from XRD. In the jarosite, the main present phases are potassium jarosite, KFe3SO4(OH)6, and natrojarosite, NaFe3SO4(OH)6. The goethite consists mainly of the iron phase goethite, FeOOH.

The required stoichiometric amount of chlorine for the chlorination of specific valuable elements was calculated based on the metal’s contents and their chlorine consumption when forming chlorides, according to Eq. 7.

$${n}_{\mathrm{Cl}}={m}_{\mathrm{r}}\times \sum {f}_{\mathrm{Cl}}\left(\mathrm{i}\right)\frac{x_{\mathrm{i}}}{M_{\mathrm{i}}}\times {f}_{\mathrm{El}}\left(\mathrm{i}\right)$$
(7)

The chlorine factor, fCl, is the ratio of chlorine and the valuable element in the compound (InCl3: fCl = 3). The extraction factor, fEl(i), is the assumed extent of chlorination of the respective metal, being 1 (100%) for all elements except for iron, where 0.05 was assumed. Table 2 gives an overview of the considered chlorine compounds, the resulting chlorine factors as well as the assumed extraction factors. Structural elements like K and Na were considered due to their high chlorine affinity, despite not being elements of interest for recovery.

Table 2 Factors for calculating the chlorine demand [21]. Copyright 2023, The Minerals, Metals & Materials Society. Used with permission
Full size table

Based on the required stoichiometric chlorine amount, the amount of the residues and chlorine compounds in the mixtures was calculated. The base value of the respective addition of AlCl3∙6H2O, FeCl3∙6H2O, or MgCl2∙6H2O, being two times the stoichiometric amount for chlorination, was determined according to Eq. 8. Furthermore, the base addition was doubled and tripled, representing four times and six times the stoichiometric calculated amount.

$${m}_{\mathrm{Chloride}\ \left(\mathrm{base}\right)}=\frac{2\times {n}_{\mathrm{Cl}}}{f_{\mathrm{Cl}/\mathrm{Chloride}}}\times {M}_{\mathrm{Chloride}}$$
(8)

The experimental evaluations consisted of four campaigns with a total of 36 trials. This included two dried materials, a jarosite and a goethite precipitate from the zinc industry. They were each treated at two temperatures (900 °C and 1100 °C). The overall mass of the residue and chlorine carrier mixture was 80 g and the treatment time was for all experiments 30 min in a high-temperature muffle furnace of Nabertherm. The mixtures were charged into the hot furnace at trial temperature using a silica crucible and placed on a weighing pan which was connected to the external scale. During the trial, the mass loss was tracked over the entire experimental time of 30 min. Subsequently, the remaining material was milled, packaged, and sent for analysis. The experimental results to interpret the effectivity of the chlorination procedures are the calculated extraction rates of specific elements, based on the chemical analyses, according to Eq. 9.

$${\mathrm{e}}_{\mathrm{El}}=1-\frac{{\mathrm{x}}_{\mathrm{El}\left(\mathrm{Fin}\right)}\times {\mathrm{m}}_{\mathrm{Fin}}}{{\mathrm{x}}_{\mathrm{El}\left(\mathrm{Res}\right)}\times {\mathrm{m}}_{\mathrm{Start}}\times {\mathrm{f}}_{\mathrm{R}}}$$
(9)

The final mass after the trials, denoted as mFin, was determined by subtracting the measured mass loss from the thermoscale from the initial start mass, mStart. The residual factor, fR, indicates the share of the precipitation residue in the mixture. The terms xEl(Res) and xEl(Fin) refer to the content of the respective element in the original iron precipitates (see Table 1) as well as in the final residual material after the chlorination experiment.

3 Results and Discussion

The results of the simultaneous thermal analyses for the synthetic mixtures with oxidic compounds of the elements Ag, In, Pb, and Zn and the different chlorinating agents are depicted in Fig. 9. The final masses of the valuable metal compounds and chlorides were calculated for complete chlorination and evaporation, based on their known starting masses. These values were plotted as a reference for each additive in order to assess the completeness of the chlorination reaction respectively the evaporation of the formed chlorides. The mass curves depicting the reaction between Ag2O and the three metal chlorides indicate that the evaporation of the formed silver chloride is taking place at elevated temperatures until completion. The comparison of the mass curves with the calculated remaining amounts suggests that silver is highly susceptible to chlorination, which aligns with the theoretical findings presented in the thermodynamic considerations, where silver chlorination has the lowest values for Delta-G and therefore having the highest driving force of all investigated metals. For the reaction of In2O3, all three chlorides lead to a rapid decrease in mass. At 900 °C, all reactions have run their course and the volatile compounds have evaporated. However, for the system with AlCl3, the calculated curve for complete chlorination and the real curve differ by more than 10%, which may indicate incomplete chlorination. The results from experiments involving PbO revealed that the calculated values for complete chlorination and evaporation were only slightly lower than the actual curves for FeCl3 and MgCl2. For AlCl3, again, a significant difference was observed. It was noticed that the chlorination of zinc from its oxide and the subsequent evaporation of the formed chloride was achieved for all three reaction systems at temperatures above 900 °C. Similar observations to those made in the cases of In2O3 and PbO were noted, i.e., that the theoretical and experimental values for FeCl3 and MgCl2 align well, while the noticeably lower theoretical final mass for AlCl3 suggests incomplete chlorination of zinc.

Fig. 9
figure 9

Mass decrease as a function of temperature for the chlorination of the metal oxides a Ag2O, b In2O3, c PbO, and d ZnO with hydrated chlorine carriers

Full size image

To summarize, the chlorination and extraction of valuable metals out of their oxides proved to work well, particularly when using FeCl3 and MgCl2. However, poorer extraction results were observed when using AlCl3 except for the metal silver. Silver seems to be chlorinated rapidly, although evaporation only begins at high temperatures. Zinc is chlorinated more difficult than silver, but easier than indium. The chlorination of lead was slightly hindered in comparison to zinc, and only AlCl3 resulted in unsatisfactory performance. As jarosite and goethite are precipitated from sulfuric acid solutions, the precipitates consist of noticeable amounts of sulfur. As it cannot be clearly stated in which compound the valuable metals, that are partly contained in very low concentrations, are present, the tests were also carried out for the sulfates Ag2SO4, In2(SO4)3, PbSO4, and ZnSO4. The resulting mass curves are shown in Fig. 10. It is observable that the chlorination of indium and silver from sulfates has a common feature in their mass curves. At about 800 °C, the chlorinating agents show mass drops with similar characteristics in the order FeCl3, AlCl3, and MgCl2 that can be assigned to the release of SO3 from the sulfates Fe2(SO4)3, Al2(SO4)3, and MgSO4 that are formed during chlorination by anion exchange. In the case of ZnSO4, the mass curve is only shown up to 1000 °C due to interfering influences in the mass signal. However, the above-mentioned features can be observed from 800 °C onwards. Experiments with PbSO4 show a rapid mass decrease with MgCl2 up to near the calculated value for complete chlorination, while the reactions with AlCl3 and FeCl3 behave strongly delayed and are still occurring at 1400 °C.

Fig. 10
figure 10

Mass decrease as a function of temperature for the chlorination of the metal sulfates a Ag2SO4, b In2(SO4)3, c PbSO4, and d ZnSO4 with hydrated chlorine carriers

Full size image

Figure 11 summarizes the realized individual extraction rates of the elements Ag, Au, Bi, Cu, In, Pb, Sn, and Zn of the experiments at 1100 °C out of industrial jarosite and goethite from zinc production. Due to the significantly lower extraction rates at 900 °C, these values are not presented. Comparing the results of jarosite and goethite, a notable difference can be seen for Ag. For jarosite, a generally low extraction was observed for all chlorides and addition rates, while for goethite, the extraction rates are significantly higher. For both residues and for all chlorides, the extraction of Au und Bi proceeds well regardless of their mixture proportions. Bi yields extraction rates of over 85% in every case, while Au is released to slightly lower extents from jarosite in the mixture with AlCl3∙6H2O. The extraction of Cu is situated in a low to medium range in the case of jarosite and reaches medium yields in the case of goethite. For the elements In, Pb, and Zn, higher extraction rates were obtained, especially when FeCl3∙6H2O and MgCl2∙6H2O were used. A low recovery is observed for Sn, although higher yields were obtained with higher addition of FeCl3∙6H2O. Summarizing, it can be stated that the highest extraction rates were achieved in experiments with the chlorides FeCl3∙6H2O or MgCl2∙6H2O, irrespective of the residual material investigated. This aligns with the findings obtained from the simultaneous thermal analyses, where AlCl3∙6H2O seemed to result in incomplete chlorination more likely than the other chlorine carriers.

Fig. 11
figure 11

Individual extraction rates from jarosite (a) and goethite (b) of the experiments at 1100 °C with varying addition of chlorine carriers (for example: Mg … base amount MgCl2∙6H2O; 2Mg … two times the base amount, respectively, four times the calculated stoichiometric amount; Al = AlCl3∙6H2O, Fe = FeCl3∙6H2O)

Full size image

It can generally be stated that the average chlorination of iron, shown in Fig. 12, is significantly lower compared to all targeted valuable elements in the majority of experiments. This confirms the underlying principle of lower affinity to chlorine than the valuable elements in focus, enabling the separation and extraction from the remaining iron-containing residue.

Fig. 12
figure 12

Iron extraction rates averages over all assessed residues [21]. Copyright 2023, The Minerals, Metals & Materials Society. Used with permission

Full size image

Nonetheless, since iron is contained in the investigated precipitation residues to a significantly higher extent than the targeted valuable metals, its extraction—though comparably low—throughout the chlorination treatment could be decisively adverse to an efficient separation. To provide an initial assessment of the composition of the volatilization product, a condenser prototype was built to collect the evaporated fraction on a filter. Calcined jarosite and MgCl2∙6H2O were mixed in a 1:1 (wt-%) ratio with a trial mass of 10 g and heated to 1000 °C in an induction furnace and maintained at that temperature for 30 min. After cooling, a small sample of the filter containing the condensed and collected material was extracted and analyzed using SEM-EDX. When compensating the amount of oxygen, the elements zinc and lead can be assumed to be present in concentrations of 28.6 and 16.7% respectively, while the iron content amounts to 2.1%. Thus, the generated approximate results from the condenser prototype indicate a sufficiently effective separation of valuable metals from the iron matrix.

To generate further information on the effectiveness of the different chlorine compounds, the mass decreases during the experiments were tracked via an external balance connected to the furnace. The characteristics of the mass decrease curves for the different chlorides are similar for the different mixture proportions and temperatures. For this reason, in Fig. 13, only the curves for the twofold of base value at 1100 °C are given for the different residuals. The most noteworthy aspect, which was evident in all the experiments, is that the mass decrease was significantly slower in experiments with AlCl3∙6H2O. As given in Fig. 13, the mass decrease (sum of reactions and evaporation processes) proceeds very rapidly to a stagnant mass signal when FeCl3∙6H2O is used. The mass stagnates from about 15 min in the experiment with jarosite. In the case of goethite, the reaction appears to have run its course entirely after about 12 min. MgCl2∙6H2O shows a further slow decrease in the remaining course. With AlCl3∙6H2O, the already mentioned effect is detectable where some reactions seem to proceed very slowly. In the case of the experiment with goethite, however, the mass also stagnates from about 20 min.

Fig. 13
figure 13

Mass curves over time for experiments with jarosite and goethite with twofold of the base chlorine addition at 1100 °C

Full size image

4 Conclusion

The thermodynamic principles underlying the selective chlorination of various valuable metals have been investigated with the software FactSage as well as HSC Chemistry. It was proven that their extraction from an iron-containing matrix is viable based on the carried out calculations, due to the low stability of ferric chloride compared to other metal chlorides. Building on this, a novel method for the simultaneous recovery of various valuable metals via selective chlorination reactions was investigated and tested on an extended laboratory scale. Simultaneous thermal analyses with oxides and sulfates of the metals Ag, In, Pb, and Zn in mixtures with different solid chlorine carriers were performed. These experiments confirmed the possibility of effectively chlorinating and evaporating the considered metals. In trials with industrial residues from zinc production, it was shown that especially the chlorides FeCl3∙6H2O and MgCl2∙6H2O yielded high extraction rates for Ag, Au, Bi, In, Pb, and Zn. A main advantage of this approach is that chlorination reactions proceed without the presence of carbon as a reducing agent (as opposed to currently applied processes like the Waelz process), and lead to effective evaporation of the targeted elements. As a result, no greenhouse gas emissions are generated in the process itself, excluding the necessary energy input for temperature control. Further investigations on the subsequent separation of the valuable metals out of the chloride fraction as well as its economic viability are necessary to evaluate the concept on a more comprehensive scale.