Selective Reduction of Laterite Nickel Ore

Abstract

Nickel is an important metal in the industry. To obtain nickel metal, the extraction process from nickel ore should be conducted. Recently, nickel ore resources from sulphide ore becomes rare which makes nickel laterite ore as the future of nickel extraction. Unlike sulphide nickel ore processing, laterite nickel ore processing requires higher processing energy through smelting. Therefore, a novel method to process laterite nickel ore using lower energy is needed. The novel method is done via direct reduction and magnetic separation. In laterite nickel processing by direct reduction, the challenges are to conduct selective reduction of nickel and to let iron unreduced. The addition of additive in the direct reduction process is needed to achieve this selective reduction. Additives used for the selective reduction includes Na₂SO₄, MgCl₂, CaSO₄, NaCl and CaCl₂·H₂O. This article will further review the role of additive used in selective reduction of laterite nickel ore and the current trends on selective reduction research.

1. Nickel – General

Nickel is an important alloying metal with widerange application in the industry.¹⁾ Nickel is one of the most important strategic metals which is widely applied to stainless steel, electroplating, catalyst, and petrochemical industry.^2,3) Nickel, found in products produced from smelters and refineries, is classified into three categories, i.e. Refined nickel (Class I), Charge nickel (Class II) and Chemicals. Class I has nickel content of 99% or higher. The product examples of class I are electrolytic nickel, pellets, briquettes, granules, rondelles and powder/flakes. Class II has nickel content less than 99%. The product examples of class II are Ferro-nickel, nickel oxide sinter, utility, and Nickel Pig Iron. Chemicals category examples are chemical nickel oxide, nickel sulphate, nickel chloride, nickel carbonate, nickel acetate, nickel hydroxide, etc.⁴⁾

2. Laterite Nickel Ore and Its Processing

About 70% of nickel reserve is nickel laterite, but only 40% nickel laterite is processed.⁵⁾ The small number of nickel laterite ore processed is due to the difficulty of nickel laterite ore processing when compared to sulphide nickel ore processing. Nickel laterite ore needs complex treatment to extract the Ni metals causing the nickel laterite ore processing more expensive than nickel sulphide ore processing. This is because Ni content of nickel laterite ore is more difficult to upgrade. Unlike nickel sulphide ore, the distribution of nickel metal in nickel laterite is distributed uniformly and act as interstitial condition in molecular lattice of particles. It makes the concentration process of nickel through flotation and gravity separation can not be applied. Based on this reason, almost all nickel laterite ore from mine have to be fed into the whole process. This results in the high operational cost required for the processing.⁶⁾ In recent years, the increase in nickel laterite production still happens because of high demand of stainless steel and because of decrease of sulphide nickel ore reserve.⁷⁾ This fact causes nickel laterite ore becoming the major source for production of nickel metal.^8,9)

The lateritic nickel ore is an ore resulting from long weathering process, derived from ultra-basic rocks containing silicate and magnesium minerals. During this process, Mg and Fe in the silicate lattice are partially replaced by Ni in order to form mineral deposits with different nickel and impurity contents.¹⁰⁾ Nickel laterite ore was classified into two types of ore, i.e. limonitic and saprolitic nickel laterite ore. Limonitic nickel ore has low content of Ni. The nickel content of limonitic nickel ore is 1.1 until 1.8 wt%. On the other hand, the saprolitic nickel ore has higher nickel content than limonitic nickel ore. The nickel content in saprolitic nickel ore is about 2% or higher.¹¹⁾ In detail, laterite deposits are classified into four categories, including: i) Limonite zone, (ii) Non-nontronite zone, (iii) Serpentine zone and (iv) Garnierite (Saprolite) zone.^5,12) Based on Fe and Mg content in laterite nickel ore, laterite nickel ore can be classified into three classes, including (a) Class A - Garnieritic laterite type (Fe <12% and MgO >25%); (b) Class B - Laterite limonitic type (Fe: 15–32 wt% and MgO <10%); and (c) Class C - Intermediate type (Fe: 12–15 wt% and MgO: 25–35 wt% or 10–25 wt%).^12,13)

Nickel laterite processing through pyrometallurgy and hydrometallurgy have been commercially applied to extract nickel metal. The pyrometallurgical routes method was suitable for saprolitic nickel ore processing. The examples of pyrometallurgy processing which have been commercially applied are Blast Furnace and Rotary Kiln – Electric Furnace (RKEF).¹⁴⁾ Conversely, limonitic nickel laterite ore was processed through hydrometallurgical route. The examples of hydrometallurgical route for limonitic nickel laterite ore processing is High Pressure Acid Leaching (HPAL).^15–17) A number of hydrometallurgy processes have been applied in industrial scale, but this application cannot meet the need of production level and its operational cost was too high.^16–19) In addition to the hydrometallurgical process, a combined process between pyro and hydro is also carried out to process limonitic nickel ore.^20,21) Using this method, the nickel recovery obtained is about 80%, so this process is generally not economical.¹⁷⁾ Figure 1 below shows the processing route of nickel laterite ore based on nickel ore types.

Fig. 1

Nickel laterite ore classification and its processing methods.

2.1 Limonitic nickel ore

Limonitic nickel ore contains iron oxide in the form of goethite (α-FeO·OH). The limonitic nickel ore is also relatively rich in cobalt and chromium. Iron oxide from limonitic nickel ore has an amorphous crystal structure with a crystal size in a nanometer scale and a large surface area. These characteristics can cause the absorption of large amounts of Al³⁺ ions from the soil. Iron oxide is rarely present in laterites in pure form because of the substitution of iron ions.²²⁾ In nature, the majority of limonite has dominant compound of 2Fe₂O₃·3H₂O. The compound has amorphous structure, jelly as iron oxide and amorphous hydroxide.¹¹⁾

The process of dehydroxylation of goethite into hematite (in oxidation state) occurs due to heating and it is the basis of some manufacturing processes. The process of dehydroxylation of goethite is a complex process as shown below.²³⁾   

\begin{align*} &\text{$\alpha$-FeOOH (goethite)} \rightarrow \text{Fe$_{5/3}$(OH)O$_{2}$ (protohematite)} \\ &\quad \rightarrow \text{Fe$_{11/6}$(OH)$_{1/2}$O$_{5/2}$ (hydrohematite)} \rightarrow \text{$\alpha$-Fe$_{2}$O$_{3}$} \end{align*}

The first transformation of goethite is the release of OH group-which causes a significant change in crystal structure. The newly formed hematite has an imperfect crystal lattice. At 800°C, recrystallization and grain growth occur. However, there is a reduction in the effect of size caused by high concentration of impurities. This is partly prevented by the diffusion of atoms into the crystal lattice of hematite substitutionally resulting in significant increase in structural disorders.²⁴⁾

The water contained in limonitic nickel ore is classified into free water, crystalline water and hydroxyl group. According to the types, while heating, there are several phenomena happening. At temperature of 25–140°C, the first phenomenon is the removal of free water. Then, at a temperature of 200–480°C, the second phenomenon is the removal of crystal water. Subsequently, at a temperature of 500–800°C, the third phenomenon is the removal of hydroxyl group. The goethite dehydroxylation process for limonitic nickel ore is the highest dehydroxylation process compared to other lateritic nickel ore types. It occurs at temperatures of 261–270°C, while the removal of the hydroxyl group in the limonite nickel ore occurs at a temperature of 400–600°C.¹²⁾

The removal of hydroxyl groups in chlorite (Fe, Mg, Al)₃(Si, Al)₂O₅(OH)₄ starts at 602°C and ends at 760°C. The chlorite group, then decomposes into MgO and SiO₂. All crystal waters of serpentine Mg₂₁Si₂O₂₈(OH)₃₄H₂O disappear above 480°C, which will form Mg₃Si₂O₅(OH)₄, and its hydroxyl group is lost at 580°C and 700°C. With the onset of this reaction, Mg₃Si₂O₇ is formed, then the compound will decompose at 795°C.²⁵⁾ So, above 795°C, the dominant phases are Fe₂O₃, MgSiO₃ and Mg₂SiO₄.

2.2 Saprolitic nickel ore

Nickel in saprolitic nickel ore occurs in the form of garnierite (Ni₃Mg₃Si₄O₁₀(OH)₈). The garnierite minerals will be decomposed while heating at 700°C. The decomposition of garnierite is shown in eq. (1).²⁶⁾   

\begin{align} \text{Ni$_{3}$Mg$_{3}$Si$_{4}$O$_{10}$(OH)$_{\text{8(s)}}$} &\rightarrow \text{3NiO$_{\text{(s)}}$} + \text{3MgO$_{\text{(s)}}$} \\ &\quad + \text{4SiO$_{\text{2(s)}}$} + \text{4H$_{2}$O$_{\text{(g)}}$} \end{align} (1)

In addition to garnierite, the minerals contained in saprolitic laterite ore are goethite, serpentine, and quartz. When the temperature is raised to 400°C, the phases of ore change. All of phases will transform to serpentine. There are two types of dehydroxilation process occurred when the saprolitic laterite ore was heated. The first dehydroxilation of serpentine is completed when the temperature of the process is above 650°C. The second dehydroxilation process is completed at temperature of 750°C.²⁷⁾ A detailed study of thermal treatment of garnieritic nickel ore was conducted by Yang, J.²⁸⁾ The initial phase of ore are chlorite ((Mg, Fe, Ni)₆(Si,Al)₄O₁₀(OH)₈), talc ((Mg, Fe, Ni)₃(Si,Al)₄O₁₀(OH)₂), quartz (SiO₂), and hematite (Fe₂O₃). The result showed that roasting at lower temperature of 400°C and 500°C made no change in minerals phase. However, when the temperature reaches 600°C, the minerals are dominated by chlorite and partially dehydrated chlorite. When the temperature is raised to 700°C, a number of serpentine was decomposed into forsterite (Mg₂SiO₄). Above 800°C, the chlorite transformed into forsterite and enstatite (MgSiO₃). At 1000°C, talc minerals transformed into forsterite and enstatite as well. So, after 1000°C, the phases of ore are forsterite, enstatite, hematite and quartz. When the temperature is raised to 1300°C, the phase of saprolitic nickel ore changes into a complex mineral. The mineral is dominated by olivine (Mg_0.5Fe_0.5)₂SiO₄.²⁹⁾

2.3 Direct reduction of laterite nickel ore in industrial scale

Most laterite nickel ore processing is carried out using Elkem process. The Elkem process uses Rotary Kiln and Electric Furnace (RKEF) to reduce and smelt laterite nickel ore. Commercialization of Elkem process itself began with the development of a pilot plant in 1953–1954. At that time, the Elkem process was used to process garnieritic nickel ore in New Caledonia. Furthermore, the Elkem process underwent significant development and increased production capacity. Until now, the Elkem process has been widely applied in various countries in the world, including Brazil, Japan, New Caledonia, Indonesia, Yugoslavia, Colombia, etc.³⁰⁾ In Rotary Kiln, nickel ore was calcined at 1000°C and then smelting is carried out in Electric Furnace. Molten Fe–Ni and slag were tapped out from Electric Furnace at 1600°C for molten Fe–Ni and 1400°C for molten slag.³¹⁾

However, the Elkem process requires an enormous electrical energy. Therefore, Nippon Yakin modifies the Krupp-Renn process. The Krupp-Renn process itself is a process of direct reduction and smelting of iron ore using a rotary kiln with a length of 70 meters and a diameter of 4.2 meters. Iron ore was smelted in semi-fused condition without fully melted. The advantages of the Krupp-Renn process include: 1) Investment costs are relatively low, 2) Products have relatively few impurities due to low smelting temperatures, 3) Does not require metallurgical coke and 4) Able to directly produce crude steel.³²⁾ Based on the Krupp-Renn process, Nippon Yakin modifies it so that it can be used to process laterite nickel ore. Furthermore, the process is referred to as “Nippon Yakin Oheyama Process”. The Nippon Yakin Oheyama Process is one of the nickel ore processing that utilizes the principle of direct reduction and has been proven commercially. At the beginning of its construction, the Nippon Yakin Oheyama Process had a low recovery of Ni (80%) because it only uses rotary kilns without any pretreatment process. However, with modifications of the preheater, dust scrubber, burner and the briquetting process, the Nippon Yakin Oheyama Process can achieve a recovery Ni of 95% and a production capacity of 9585 Ton Ni/year. The total energy requirement of the Nippon Yakin Oheyama Process is about 5,000 kWh/Ton Ni. The Fe–Ni particle products were referred to as “Luppe” which still contains 2% of slag.³³⁾ Luppe has Fe content of between 75–80% and Ni between 25–20%, particle size ≥0.1 and an average diameter of 1 mm. The Ni losses due to the presence of fine Fe–Ni particles (in the size of µm) which is joined in the slag. Therefore, in the Nippon Yakin Oheyama Process, the growth of Fe–Ni particles has an important role in determining Ni recovery. One of the factors that affect the growth of Fe–Ni particles, namely softening slag³⁴⁾ and the formation of secondary melt (Low MgO·FeO, high SiO₂·CaO·Al₂O₃ silicate) at 1250–1300°C.³⁵⁾ In the operation, there are major problems that disrupt Rotary Kiln productivity, namely the formation of Slag Ring around the inner walls of the Rotary Kiln.³⁶⁾ The Nippon Sure Oheyama Process flow sheet process is shown in Fig. 2, while the typical Luppe composition is shown in Table 1.

Fig. 2

Flowsheet of Nippon Yakin Oheyama Process for production of Luppe from Laterite Ni-ore, Adopted from Refs. 33, 34).

Table 1 Typical composition of Luppe.³³⁾

3. The Selective Reduction of Laterite Nickel Ore

Selective reduction of nickel aims to greatly reduce nickel oxide into metallic nickel and to maintain the iron oxide unreduced or to minimize the reduction of iron oxide into metallic iron. With this selective reduction process, the reduction product will have high Ni content and moderate Fe content. Selective reduction of nickel is strongly influenced by the thermodynamics of the reduction reaction of nickel and iron compounds in laterite nickel ore. Many researchers have conducted research and modelling on thermodynamic aspects in the selective reduction of laterite nickel ore.^2,17,37)

From research and modelling that have been done, there are several thermodynamic aspects that are important in influencing selective reduction, including: 1) Potential reduction (CO/CO₂ ratio), 2) Reduction temperature, and 3) Phase/compound transformation. Carbothermic reduction reactions that occur are shown by eq. (2)–(7):³⁷⁾   

\begin{equation} \text{NiO$_{\text{(s)}}$} + \text{CO$_{\text{(g)}}$} = \text{Ni$_{\text{(s)}}$} + \text{CO$_{\text{2(g)}}$} \end{equation} (2)

  

\begin{equation} \text{NiFe$_{2}$O$_{\text{4(s)}}$} + \text{CO$_{\text{(g)}}$} = \text{Ni$_{\text{(s)}}$} + \text{Fe$_{2}$O$_{\text{3(s)}}$} + \text{CO$_{\text{2(g)}}$} \end{equation} (3)

  

\begin{equation} \text{3Fe$_{2}$O$_{3}$} + \text{CO$_{\text{(g)}}$} = \text{2Fe$_{3}$O$_{\text{4(s)}}$} + \text{CO$_{\text{2(g)}}$} \end{equation} (4)

  

\begin{equation} \text{Fe$_{3}$O$_{4}$} + \text{CO$_{\text{(g)}}$} = \text{3FeO$_{\text{(g)}}$} + \text{CO$_{\text{2(g)}}$} \end{equation} (5)

  

\begin{equation} \text{FeO} + \text{CO$_{\text{(g)}}$} = \text{Fe$_{\text{(s)}}$} + \text{CO$_{\text{2(g)}}$} \end{equation} (6)

  

\begin{equation} \text{FeSiO$_{\text{4(s)}}$} + \text{2CO$_{\text{(g)}}$} = \text{2Fe} + \text{SiO$_{\text{2(s)}}$} + \text{2CO$_{\text{2(g)}}$} \end{equation} (7)

From eq. (2)–(7), it can be seen that thermodynamically, the control of CO and CO₂ can affect the occurrence of Ni and Fe reduction reactions. In selective reduction Ni, as far as possible the reduction of iron oxide should be prevented so that the reduction reaction of iron oxide only occurs until eq. (5). So, there is still a lot of iron metal that has not been formed. In other words, in the process of selective reduction of laterite nickel ore, Fe is a kind of opponent of selective reduction of Ni. In order to get Fe–Ni alloy products (whether containing ferronickel, Fe–Ni crude oil or NPI) with higher Ni grade, it must inhibit the process of iron oxide reduction to metallic iron or carrying out a compounding process between iron and the other elements (such as with oxygen, sulfur or chloride).

Generally, the higher the potential reduction and temperature reduction, the higher Ni recovery and the lower the Ni grade. In addition, the presence of non-oxide iron compounds will make the reduction of Fe into metals will be inhibited. Thermodynamic relationships between reduction potential, temperature and compounds formed to the degree of nickel metallization are shown by stability diagram (Fig. 3) which has been studied by Hallet, 1997³⁸⁾ and Elliot R., 2016.³⁷⁾

Fig. 3

The stability diagram of Fe–Ni–O system for limonitic nickel ore containing 1.2 wt% Ni 1997.^37,38)

Based on Fig. 3, it appears that the final product of the carbothermic reduction process will be affected by two main things, namely temperature and reduction potential (ln CO/CO₂). The combination of these two process variables will affect the final product of reduction. In general, to produce the final product in the form of metallic Fe–Ni (100% metallization), the higher the reduction temperature, the higher the reduction potential needed. For example, if the reduction process occurs at a low temperature (300–600°C), then a moderate reduction potential is needed. On the other hand, if the reduction process occurs at high temperatures (900–1100°C), a high reduction potential is needed. For more details, analysis will be carried out for each zone.

Zone 1

Zone 1 is a zone where the reduction product is an Fe–Ni alloy. So all Fe and Ni are ideally reduced. In this zone, theoretically, recovery of Fe and Ni can reach 100%. While the grade of Fe and Ni depends on the content of Fe and Ni of nickel ore being processed. For example, if a limonite ore to be processed has a Ni content of 1.2% and Fe of 37%, then the Ni grade in this zone is: (1.2/(1.2 + 37)) × 100% = 3.14%. The Ni grade is in accordance with the assumption that there are no other impuritic elements present in Fe–Ni alloys. In this zone, the Ni grade is still low, especially if the nickel ore used has a high Fe content. In general, zone 1 can be easily achieved by reducing the reduction temperature and increasing the reduction potential by increasing the CO/CO₂ ratio. By increasing the ratio of CO/CO₂ it will cause atmospheric conditions to be very reductive, causing many oxides of reduced Fe to become metals. As a result, Ni grade will decrease while grade Fe will increase. Because the reduction temperature tends to be low, the reduction reaction rate will run slowly in this zone. Thus, the level of productivity in this zone is relatively lower.

Zone 2

Zone 2 is a zone where the reduction product is (Fe,Ni)_1−yO + (Fe,Ni). In this zone, there are still exist iron and nickel oxide compounds. The iron and nickel oxides in this zone have an oxidation number of +2. The reduction product in Zone 2 has a higher Fe and Ni content than the reduction product in Zone 3. In addition, the recovery of Fe and Ni is also higher compared to Zone 3. Zone 3 can be achieved by using a moderate CO/CO₂ ratio and high temperature. Zone 2 is the most suitable zone to be used as a reference for selective reduction process because the product has a moderate Ni grade and high Ni recovery. In addition, with a higher temperature applied, then qualitatively, the reduction reaction rate will be faster than the other zones.

Zone 3

Zone 3 is a zone where the reduction product is (Fe,Ni)O·Fe₂O₃ + (Fe,Ni). In this zone, there are still exist nickel and iron oxide. Ni in the form of NiO and Fe in the form of FeO and Fe₂O₃. From the reduced compounds, it appears that in zone 3, there is still a lot of Fe that has not been reduced to metal, because FeO and Fe₂O₃ are still found. Therefore, zone 3 has a reduction product that has the highest Ni content. To reach Zone 3, a low reduction temperature and a moderate CO/CO₂ ratio are needed. With a low temperature and a low ratio of CO/CO₂, the reduction of Fe into metals will be inhibited, so Fe is still present in the form of FeO or Fe₂O₃. As a result, the reduction product will has high Ni content.

Zone 4

Zone 4 is a zone where the reduction product is (Fe,Ni)O·Fe₂O₃. In this zone, Ni and Fe metallization has not occurred. Thus, Zone 4 is not desirable in reduction process, especially in selective reduction process. Zone 4 will be occurred when the CO/CO₂ ratio is too low and the temperature is too high. The low CO/CO₂ ratio makes CO insufficient for the reduction process and makes the reduction atmosphere not reductive. Thus, Fe and Ni oxides have not been reduced to metal.

On the other hand, the addition of additives can also be used to increase the selective reduction of nickel. These additives are classified into two groups, namely additives in the form of sulfocompound (Na₂SO₄^39,40) and CaSO₄⁴¹⁾) and additives in the form of chlorine compounds (MgCl₂,⁴²⁾ NaCl⁴³⁾ and CaCl₂·H₂O⁴⁴⁾). In additive additions in the form of sulfocompounds, selective reduction occurs when Fe in Fe–Ni or iron oxide reacts with sulfur compounds to produce Troilite (FeS) compounds. On the other hand, in addition to additives in the form of chlorine compounds, selective reduction occurs when Ni oxide and Fe react with Cl to form NiCl₂, FeCl₂, and FeCl₃. While some other iron oxides will react with CaO or MgO to form CaFe₂O₄ or MgFe₂O₄. The NiCl₂ and FeCl₂ then react with H₂ to form Fe–Ni alloys. With controlling of additives dosage, the chloridization of Fe can be controlled. So, there is still Fe in the form of oxide. As a result, iron oxide will not be completely reduced by H₂ and the Ni grade will be higher.⁴³⁾

Despite increase the tendency for selective reduction of Ni, the additive addition will increase Ni recovery and enlarge Fe–Ni particle products because it can trigger the formation of compounds that have low melting points on the surface of Fe–Ni particles and promote the formation of compounds that have a low melting point at impurity around Fe–Ni particles, such as NaO, CaO or MgO. The difference mechanism of sulfo and clorin compounds on selective reduction of Ni is shown in Fig. 4. To study the mechanism for selective reduction for each additive, a selective reduction mechanism will be described in the next section for each type of additive and its reaction.

Fig. 4

Mechanism of additive on promoting selective reduction of Nickel for: a) Sulfo compound additives and b) Chloride compound additives. (Adapted from Ref. 43)).

Temperature which can be used for selective reduction is in the range of 1100–1450°C. Too low temperature will cause slow reaction rate.³⁷⁾ On the other hand, too high temperature will cause the formation of a liquid phase.⁴⁵⁾ In addition, the high temperature reduction causes the possibility of the formation of olivines getting higher, thereby lowering the reduction of Ni. In addition, too high reduction temperature causes reduction of chromium and silicon oxide compounds, so Cr and Si will also dissolve into the Fe–Ni crystal lattice. As a result, both Ni and Fe contents will decrease.⁴⁶⁾ Thus, the optimal temperature for selective reduction is at 1200°C.³⁹⁾ In his research, the selectivity factor achieved was 21.8. Even so, when considering the size of Fe–Ni particles formed, the optimal temperature is at 1420°C. This is because the size of Fe–Ni particles obtained is in centimeter size.⁴⁷⁾ So from the industry perspective, it is more efficient because Fe–Ni recovery can be maintained during magnetic separation process. The role of each additive in selective reduction of Ni will be described below.

3.1 Na₂SO₄ additive

Na₂SO₄ can increase the selective reduction of Ni. This happens because Na₂SO₄ will decompose to Na₂S, Na₂O, and S under reducing atmospheric condition. The decomposition reaction is shown in eq. (8)–(10).^{39,40,47–50)}   

\begin{equation} \text{Na$_{2}$SO$_{4}$} + \text{4CO} \rightarrow \text{Na$_{2}$S} + \text{4CO$_{2}$} \end{equation} (8)

  

\begin{equation} \text{Na$_{2}$SO$_{4}$} + \text{3CO} \rightarrow \text{Na$_{2}$O} + \text{S} + \text{3CO$_{2}$} \end{equation} (9)

  

\begin{equation} \text{Na$_{2}$SO$_{4}$} + \text{4C} \rightarrow \text{Na$_{2}$S} + \text{4CO} \end{equation} (10)

The reduction of Na₂SO₄ occurs at the temperature range of 850–900°C.⁴⁷⁾ The products of Na₂SO₄ reduction are Na₂O, Na₂S, CO and CO₂. Na₂O can improve Ni recovery because Na₂O will react with iron silicate and form sodium silicate as can be seen in eq. (11).³⁹⁾ The sodium silicate compounds have low melting temperatures, thus accelerate the rate of aggregation of metal particles.   

\begin{equation} \text{Na$_{2}$O} + \text{2Fe$_{2}$SiO$_{4}$} \rightarrow \text{4FeO} + \text{Na$_{2}$Si$_{2}$O$_{5}$} \end{equation} (11)

On the other hand, S can significantly increase the grade of nickel. Further, iron recovery is decreased due to the formation of Troilite (FeS) as can be seen in eq. (12).^39,51,52)   

\begin{equation} \text{Fe} + \text{S} \rightarrow \text{FeS} \end{equation} (12)

Meanwhile, Na₂S can also reduce Fe recovery because Na₂S will react with FeO to form FeS as can be seen in eq. (13).^39,40) Therefore, FeO can no longer be reduced by reducing agents.   

\begin{equation} \text{Na$_{2}$S} + \text{FeO} + \text{2SiO$_{2}$} \rightarrow \text{FeS} + \text{Na$_{2}$Si$_{2}$O$_{5}$} \end{equation} (13)

Troilite acts as an activating agent to facilitate the formation of the melting phase and then increase the transport rate to accelerate the aggregation of Fe–Ni particles.⁴⁰⁾ This FeS compound cannot be drawn by the magnet in the following process after reduction, thus increasing the Ni grade. Sulfur can cause an increase in the surface of metal particles and can decrease the surface tension of metal particles, thereby increasing the growth of metal particle size. However, Na₂O will react with silicates and form minerals with low melting temperatures.³⁹⁾ The mechanism of selective reduction of Ni by using Na₂SO₄ additive is shown in Fig. 5.

Fig. 5

Mechanism of Na₂SO₄ in selective reduction of laterite nickel ore.

3.2 MgCl₂ additive

The addition of MgCl₂ may improve Ni grade and Ni recovery. The process is called as process of segmentation of chloridation. In this process, chlorination of Ni, Co and Fe oxide from laterite nickel ore happen. First of all, there is an interaction of MgCl₂ with water content in the moisture and SiO₂ from the laterite nickel ore impurities. The reaction produces magnesium silicate and HCl compounds as can be seen in eq. (14).⁴²⁾   

\begin{equation} \text{MgCl$_{2}$} + \text{SiO$_{2}$} + \text{H$_{2}$O} = \text{MgO${\cdot}$SiO$_{2}$} + \text{2HCl} \end{equation} (14)

Then, HCl will react with NiO and Co₂SiO₄ to form nickel chloride and cobalt chloride compounds as can be seen in eq. (15) and (16).⁴²⁾   

\begin{equation} \text{NiO} + \text{2HCl} = \text{NiCl$_{2}$} + \text{H$_{2}$O}\quad (\mathrm{T} \geq 600\,\text{K}) \end{equation} (15)

  

\begin{align} &\text{Co$_{2}$SiO$_{4}$} + \text{4HCl} = \text{2CoCl$_{2}$} + \text{SiO$_{2}$} + \text{2H$_{2}$O}\\ &\qquad (\mathrm{T} \geq 600\,\text{K}\text{–}900\,\text{K}) \end{align} (16)

Subsequently, NiCl₂ and CoCl₂ will react with carbon in the reductant and with water content in the moisture resulting in Ni and Co metal as can be seen in eq. (17) and (18).⁴²⁾   

\begin{equation} \text{NiCl$_{2}$} + \text{C} + \text{H$_{2}$O} = \text{Ni} + \text{2HCl} + \text{CO}\quad (\mathrm{T} \geq 900\,\text{K}) \end{equation} (17)

  

\begin{equation} \text{CoCl$_{2}$} + \text{C} + \text{H$_{2}$O} = \text{Co} + \text{2HCl} + \text{CO}\quad (\mathrm{T} \geq 900\,\text{K}) \end{equation} (18)

The higher MgCl₂ added, the higher Ni grade and Ni recovery will be. However, when the addition of MgCl₂ is higher than 6%, the grade and recovery of Ni no longer change significantly. Hence, an important point in the selective reduction of Ni using the addition of MgCl₂ additive is the amount of HCl formed from the MgCl₂ reaction with H₂O from the moisture feed.⁴²⁾

3.3 CaSO₄ additive

CaSO₄ may increase the selective reduction of Ni from the reduction process of limonite laterite nickel ore. This is because CaSO₄ will produce reaction of fayalite (FeSiO₃) formation. Thermodynamically, CaSO₄ will decompose into S₂ (g), O₂ (g), and CaO under reducing atmospheric conditions. CaO will then react with SiO₂ and FeO to form kirschsteinite (CaFeSiO₄) which will accelerate the formation of fayalite. In addition, S₂ (g) will react with the iron oxide of Fe and Ni. Thus, the reaction result of S₂ (g) and Fe will form an iron sulphide having a low melting point. As a result, it causes an increase in particle size of Fe–Ni metal. The reactions occurred during the process of reduction of limonite laterite nickel ore with the addition of CaSO₄ additives are as follows (eq. (19)–(30)).⁴¹⁾   

\begin{equation} \text{2Fe$_{3}$O$_{4}$} + \text{2CO$_{\text{(g)}}$} = \text{6FeO} + \text{2CO$_{\text{2(g)}}$} \end{equation} (19)

  

\begin{equation} \text{2/5Fe$_{3}$O$_{4}$} + \text{S$_{\text{2(g)}}$} = \text{6/5FeS} + \text{4/5SO$_{\text{2(g)}}$} \end{equation} (20)

  

\begin{equation} \text{4/3FeO} + \text{S$_{\text{2(g)}}$} = \text{4/3FeS} + \text{2/3SO$_{\text{2(g)}}$} \end{equation} (21)

  

\begin{equation} \text{CaO} + \text{FeO} + \text{SiO$_{2}$} = \text{CaFeSiO$_{4}$} \end{equation} (22)

  

\begin{equation} \text{2FeO} + \text{2CO$_{\text{(g)}}$} = \text{2Fe} + \text{2CO$_{\text{2(g)}}$} \end{equation} (23)

  

\begin{equation} \text{2NiO} + \text{2CO$_{\text{(g)}}$} = \text{2Ni} + \text{2CO$_{\text{2(g)}}$} \end{equation} (24)

  

\begin{equation} \text{2CoO} + \text{2CO$_{\text{(g)}}$} = \text{2Co} + \text{2CO$_{\text{2(g)}}$} \end{equation} (25)

  

\begin{equation} \text{2Fe} + \text{S$_{\text{2(g)}}$} = \text{2FeS} \end{equation} (26)

  

\begin{equation} \text{2Ni} + \text{S$_{\text{2(g)}}$} = \text{2NiS} \end{equation} (27)

  

\begin{equation} \text{2Co} + \text{S$_{\text{2(g)}}$} = \text{2CoS} \end{equation} (28)

  

\begin{equation} \text{2Fe} + \text{2NiS} = \text{2FeS} + \text{2Ni} \end{equation} (29)

  

\begin{equation} \text{2Fe} + \text{CoS} = \text{2FeS} + \text{2Co} \end{equation} (30)

From the above reaction, it appears that the increase in Ni content depends on the inhibition of FeO reduction to Fe due to the formation of fayalite or kirschsteinite. Thus, a sufficient amount of SiO₂ is required so that FeO is not reduced into Fe. Limonite ores have high Fe content and low SiO₂, so mixing limonite and saprolite nickel ore is required.⁴¹⁾

3.4 NaCl additive

Similar to MgCl₂, NaCl is added as an additive to become the chlorination agent Ni and Fe ore laterite nickel. As for the MgCl₂ reaction, in the presence of moisture, Ni and Fe oxides will be chlorinated by HCl, which is formed from NaCl pyrohydrolysis process with a SiO₂ catalyst. HCl then reacts with Ni and Fe oxides to form chlorine compounds, NiCl₂, and FeCl₂ (eq. (31)–(33)).^43,44)   

\begin{align} \text{2NaCl} + \text{SiO$_{2}$} + \text{H$_{2}$O} &= \text{Na$_{2}$O${\cdot}$SiO$_{2}$} + \text{2HCl}\\ &\qquad (\mathrm{T} \geq 2403\,\text{K}) \end{align} (31)

  

\begin{equation} \text{FeO} + \text{2HCl} = \text{FeCl$_{2}$} + \text{H$_{2}$O}\quad (\mathrm{T} \geq 647\,\text{K}) \end{equation} (32)

  

\begin{equation} \text{NiO} + \text{2HCl} = \text{NiCl$_{2}$} + \text{H$_{2}$O}\quad (\mathrm{T} \geq 1021\,\text{K}) \end{equation} (33)

In addition, iron and nickel chlorine will react with moisture as can be seen in eq. (34)–(36).   

\begin{equation} \text{NiCl$_{2}$} + \text{6H$_{2}$O} = \text{NiCl$_{2}{\cdot}$6H$_{2}$O}\quad (\mathrm{T} \geq 179\,\text{K}) \end{equation} (34)

  

\begin{equation} \text{FeCl$_{2}$} + \text{2H$_{2}$O} = \text{FeCl$_{2}{\cdot}$2H$_{2}$O}\quad (\mathrm{T} \geq 214\,\text{K}) \end{equation} (35)

  

\begin{equation} \text{FeCl$_{2}$} + \text{4H$_{2}$O} = \text{FeCl$_{2}{\cdot}$4H$_{2}$O}\quad (\mathrm{T} \geq 192\,\text{K}) \end{equation} (36)

Then, the NiCl₂, FeCl₂ and CoCl₂ will react (water gas reaction) with C and H₂O to form Ni, Fe and Co as can be seen in eq. (37)–(39).   

\begin{equation} \text{NiCl$_{2}$} + \text{C} + \text{H$_{2}$O} = \text{Ni} + \text{2HCl} + \text{CO}\quad (\mathrm{T} \geq 900\,\text{K}) \end{equation} (37)

  

\begin{equation} \text{FeCl$_{2}$} + \text{C} + \text{H$_{2}$O} = \text{Co} + \text{2HCl} + \text{CO}\quad (\mathrm{T} \geq 900\,\text{K}) \end{equation} (38)

  

\begin{equation} \text{CoCl$_{2}$} + \text{C} + \text{H$_{2}$O} = \text{Ni} + \text{2HCl} + \text{CO}\quad (\mathrm{T} \geq 900\,\text{K}) \end{equation} (39)

On the other hand, oxide compounds in nickel ore will react in accordance with eq. (40)–(43).   

\begin{align} \text{2Mg$_{2}$SiO$_{4}$} + \text{SiO$_{2}$}& = \text{Mg$_{2}$SiO$_{4}$} + \text{2MgSiO$_{3}$}\\ &\qquad (\mathrm{T} \geq 24\text{–}1098\,\text{K}) \end{align} (40)

  

\begin{equation} \text{Fe$_{2}$O$_{3}$} + \text{MgO} = \text{MgFe$_{2}$O$_{4}$}\quad (\mathrm{T} \geq 1473\,\text{K}) \end{equation} (41)

  

\begin{equation} \text{Fe$_{2}$O$_{3}$} + \text{NiO} = \text{NiFe$_{2}$O$_{4}$}\quad (\mathrm{T} \geq 1473\,\text{K}) \end{equation} (42)

  

\begin{equation} \text{MgNiSi$_{2}$O$_{6}$} = \text{MgSi$_{2}$O$_{5}$} + \text{NiO}\quad (\mathrm{T} \geq 1473\,\text{K}) \end{equation} (43)

At the end of the process, the compounds formed include Mg₂SiO₄ (Most dominant), MgFe₂O₄, NiFe₂O₄, SiO₂, MgNiSi₂O₆ and Fe₂O₃.⁴³⁾

3.5 CaCl₂·2H₂O additive

Similar to NaCl, the mechanism of CaCl₂·2H₂O promotes selective reduction of Nickel in lateritic nickel ore through chloridation. First, CaCl₂·2H₂O was decomposed into CaCl₂ and 2H₂O. The reaction is shown in eq. (44) below.⁵³⁾   

\begin{equation} \text{CaCl$_{2}{\cdot}$2H$_{2}$O$_{\text{(s)}}$} = \text{CaCl$_{\text{2(s)}}$} + \text{2H$_{2}$O$_{\text{(g)}}$}\quad (\mathrm{T} \geq 20\,\text{K}) \end{equation} (44)

Then, the CaCl₂ will react with SiO₂ (in ore) and H₂O (form moisture of ore and decomposition product of CaCl₂·2H₂O) to form CaSiO₃ and HCl as shown in eq. (45).⁴⁴⁾   

\begin{align} \text{CaCl$_{\text{2(s)}}$} + \text{SiO$_{\text{2(s)}}$} + \text{H$_{2}$O$_{\text{(g)}}$} &= \text{CaSiO$_{\text{3(s)}}$} + \text{2HCl$_{\text{(g)}}$}\\ &\qquad (\mathrm{T} \geq 1168\,\text{K}) \end{align} (45)

The HCl gas product has a role in the chloridation process for selective reduction. HCl_(g) will react with NiO from lateritic nickel ore to form NiCl₂. The reaction is shown in eq. (46).⁴⁴⁾   

\begin{equation} \text{NiO$_{\text{(s)}}$} + \text{2HCl$_{\text{(g)}}$} = \text{NiCl$_{\text{2(s)}}$} + \text{H$_{2}$O$_{\text{(g)}}$}\quad (\mathrm{T} \geq 1021\,\text{K}) \end{equation} (46)

On the other hand, carbon from reductor will react with H₂O to form CO gas and H₂ gas. The reaction is shown in eq. (47).⁴⁴⁾   

\begin{equation} \text{C$_{\text{(s)}}$} + \text{H$_{2}$O$_{\text{(g)}}$} = \text{CO$_{\text{(g)}}$} + \text{H$_{\text{2(g)}}$}\quad (\mathrm{T} \geq 947\,\text{K}) \end{equation} (47)

Finally, the NiCl₂ will react with H₂ gas or directly react with carbon and H₂O to form metallic Nickel. Both reactions are shown in eq. (48) and (49).⁴⁴⁾   

\begin{equation} \text{NiCl$_{\text{2(s)}}$} + \text{H$_{\text{2(g)}}$} = \text{Ni$_{\text{(s)}}$} + \text{2HCl$_{\text{(g)}}$}\quad (\mathrm{T} \geq 720\,\text{K}) \end{equation} (48)

  

\begin{align} \text{NiCl$_{\text{2(s)}}$} + \text{C$_{\text{(s)}}$} + \text{H$_{2}$O$_{\text{(g)}}$} &= \text{CO$_{\text{(g)}}$} + \text{Ni$_{\text{(s)}}$} + \text{2HCl$_{\text{(g)}}$}\\ &\qquad (\mathrm{T} \geq 831\,\text{K}) \end{align} (49)

4. Novel Research of Selective Reduction of Laterite Nickel Ore

The summary of novel research of selective reduction of lateritic nickel ore is shown in Table 2. There are two parameters to compare each method of experiment, i.e. selectivity factor and separation efficiency. The selectivity factor shows the level of selective reduction of nickel that occurs. On the other hand, the separation efficiency shows efficiency of the process that occurs and this can be related to economical consideration.

Table 2 Summary of novel research of selective reduction of lateritic nickel ore.

The selectivity factor was calculated using eq. (50). Conversely, the separation efficiency was calculated using eq. (51).   

\begin{equation} \textit{Selectivity Factor} = \frac{X_{\textit{Ni}}Y_{\textit{Fe}}}{X_{\textit{Fe}}Y_{\textit{Ni}}} \end{equation} (50)

Where: X is the grade of Ni or Fe in lateritic nickel ore and Y is the grade of Ni or Fe in reduction product.⁵⁴⁾   

\begin{equation} \textit{Separation Efficiency} = \frac{100R_{\textit{Ni}}m(Y_{\textit{Ni}} - X_{\textit{Ni}})}{(m - X_{\textit{Ni}})X_{\textit{Ni}}} \end{equation} (51)

Where: R_Ni is the Recovery of Nickel, m is percentage of nickel in valuable mineral (assumed in the form of NiO, so m is 78.58%), Y_Ni is the Nickel grade in reduction product, X_Ni is the nickel grade in lateritic nickel ore.^55,56)

Table 2 shows the summary of novel research about selective reduction of nickel obtained from limonitic, saprolitic, and other types of laterite nickel ore. The best separation efficiency of limonitic type nickel ore processed is conducted by Elliot, R.³⁷⁾ using carbon as reductant with 7.07 kg of carbon of 100 kg ore. Nevertheless, there is no data showing the selectivity factor for its process. Finally, the best method is conducted by Lu, C.⁵⁷⁾ The study used 8 wt% of coal and 20 wt% of Na₂SO₄ as additive. The selectivity factor of this method is 6.01 and the separation efficiency of this method is 69.93%. On the other hand, when saprolitic nickel ore is processed, the best method was conducted by Ref. 58). The study used coal and composite additive of 14 wt%. Using this method, separation efficiency of 70.36% was obtained. However, the selectivity factor was still lower than other methods. And for the best selectivity factor for saprolite processing, it was conducted by Ref. 39). Coal was used as reductor and 10 wt% of Na₂SO₄ as additive. As a result, selectivity factor of 21.8 (The highest selectivity factor among other methods) was obtained.

5. The Feasibility and Potential Problems of Additives Addition in the Process

The selective reduction process using a number of additives is proven to be able to increase the selectivity factor, thus producing products with high Ni grades. However, to date, the process of selective reduction using additives is still at the research stage and has not been applied on a pilot or industrial scale. That is caused by several factors, including:

1. 1.    Production cost
   • In the selective reduction process using additives, the average additive is needed as much as between 8–10 wt% of the entire raw material used. Moreover, to date the selective reduction studies have used analytic grade additives. Thus, the cost of procuring raw materials will increase significantly. On the other hand, product sales will be based on the tonnage of the nickel metal produced. Thus, increasing Ni levels does not significantly influence product sales revenue. On the other hand, there is an increase in production costs due to the addition of additives for selective reduction of Ni. Therefore, there is a need for a complete economic feasibility analysis to determine whether this selective reduction process will provide financial benefits compared to the reduction process currently applied.
2. 2.    The formation of slag which contains compounds that are harmful to industrial operations
   • Addition of additives which have the element of Na, will produce slag with high NaO content. Although from the viewpoint of Ni recovery, it will help increase Ni recovery because NaO will significantly reduce the viscosity, softening temperature, and melting temperature of the slag. However, an increase in NaO content will cause slags to become very corrosive to reactor refractories. If it left for a long time, the thickness of the refractory will decrease due to high temperature chemical reaction between the slag (with a high NaO content) and the refractory constituent compounds. If it is related to production costs, the reactor maintenance time will be faster and cause an increase in maintenance costs.
3. 3.    Byproducts in the form of gases that harm the environment
   • In the selective reduction process using additives, the additives used are sulfate compounds or chloride compounds. To have a selective effect, S or Cl on additives will react with Fe to reduce the total mass of Fe in the product and to reduce the interface energy of the particles. In the perspective of selectivity reduction, it has been proven to improve Ni grades. However, there are potential problems, namely environmental issues as an impact on the addition of this additive. In the pyrometallurgy process, all processes utilize high temperature and there are many transport phenomena that occur. Thus, there is also the possibility of the formation of toxic gases such as H₂S and HCl as a byproduct of selective reduction reactions using additives. The gas will be dangerous both to the operator and the environment.

6. Conclusion

Nickel selective reduction is a challenge in ferronickel production research through carbothermic reduction. There are several criteria in nickel selective reduction process. The criteria needed for increasing Ni grade during the carbothermic reduction process are:

1. (1)    Optimum Reduction Potential (CO/CO₂ gas ratio)

   The reduction potential will be affected by the CO/CO₂ gas ratio. In general, for laterite nickel ore, a decrease in reduction potential (by reducing the CO/CO₂ gas ratio) will increase the nickel grade. The increase in Ni grade is caused by the low reduction potential which will prevent reduction of Fe oxide to Fe metal. In the case of carbothermic reduction, the use of C (from the reducing agent)/O (from iron oxide and nickel oxide) ratio is 0.75–0.80.

2. (2)    Optimum Control of Reduction Temperature

   The optimal reduction temperature for selective reduction of nickel is at 1100°C–1400°C. When the reduction temperature is higher, it will increase the possibility of formation of olivine ((Mg, Fe, Ni)₂SiO₄). As a result, the reduction of nickel oxide will be more difficult.

3. (3)    Promotion of Iron Compound Formation Reaction

   The formation of iron compounds has an important role in the selective reduction process of laterite nickel ore. The formation of iron compounds is facilitated by the addition of additives during the reduction process. Additives proven to be able to encourage selective reduction are sulfate or chloride compounds. For sulfo compounds, the sulfur content in additives will facilitate the formation of FeS (Troilite) reactions. The FeS compounds will then separate from Fe–Ni metal and make a mixture with slag. While for chloride compounds, the Cl content in additives will facilitate the formation of NiCl₂ and leave Fe₂O₃ or FeO still unchloridized (with controlling of chloride additive dosage addition). To obtain a high nickel selectivity factor, 20 wt% Na₂SO₄ additives can be used for limonitic nickel ore and 10 wt% Na₂SO₄ additives can be used for saprolitic nickel ore.

Until now, the application of selective reduction of Ni in industrial scale still requires a more detailed study, such as studies of economic feasibility, impacts on the environment and impacts on unit operations and unit processes. Therefore, in the future, these studies are important to be carried out, so it can be known whether direct reduction with the addition of additives is feasible to be applied on a pilot plant scale or even on an industrial scale.

Acknowledgements

The authors express gratitude to Institut Teknologi Sepuluh Nopember for the financial support provided for this research.

REFERENCES

• 1)   R.  Moskalyk and  A.  Alfantazi: Miner. Eng. 15 (2002) 593–605.
• 2)   S.  Chen,  S.-q.  Guo,  L.  Jiang,  Y.-l.  Xu and  W.-z.  Ding: Trans. Nonferrous Met. Soc. China 25 (2015) 3133–3138.
• 3)   L.-r.  Mao: J. China National Resources Recycling 10 (1999) 36–37.
• 4)  I.N.S.G. (INSG), XXIV (2015).
• 5)  A.D. Dalvi, W.G. Bacon and R.C. Osborne: The past and the future of nickel laterites, PDAC 2004 International Convention, Trade Show & Investors Exchange, (Toronto, The prospectors and Developers Association of Canada, 2004) pp. 1–27.
• 6)   T.  Norgate and  S.  Jahanshahi: Miner. Eng. 24 (2011) 698–707.
• 7)   Y.-C.  Zhai,  W.-N.  Mu,  L.  Yan and  X.  Qian: Trans. Nonferrous Met. Soc. China 20 (2010) s65–s70.
• 8)   B.  Li,  H.  Wang and  Y.  Wei: Miner. Eng. 24 (2011) 1556–1562.
• 9)   J.  Kim,  G.  Dodbiba,  H.  Tanno,  K.  Okaya,  S.  Matsuo and  T.  Fujita: Miner. Eng. 23 (2010) 282–288.
• 10)   C.  Cao,  Z.  Xue and  H.  Duan: Int. J. Nonferrous Metall. 05 (2016) 9–15.
• 11)   J.  Chen,  T.  Sun and  Q.  Lu: J. Chem. Pharm. Res. 6 (2014) 671–678.
• 12)   B.  Rizov: J. Univ. Chem. Technol. Metall. 47 (2012) 207–210.
• 13)  E.N. Zevgolis: Extractive Metallurgy of Nickel: Part 1 Pyrometallurgical Methods, (National Technical University of Athens, 2000).
• 14)   M.  Rao,  G.  Li,  T.  Jiang,  J.  Luo,  Y.  Zhang and  X.  Fan: JOM 65 (2013) 1573–1583.
• 15)   B.I.  Whittington and  D.  Muir: Miner. Process. Extr. Metall. Rev. 21 (2000) 527–599.
• 16)  J. Kyle: Nickel laterite processing technologies–where to next? In: ALTA 2010 Nickel/Cobalt/Copper Conference, (2010).
• 17)   C.A.  Pickles,  J.  Forster and  R.  Elliott: Miner. Eng. 65 (2014) 33–40.
• 18)   M.G.  King: JOM 57 (2005) 35–39.
• 19)   A.  Oxley and  N.  Barcza: Miner. Eng. 54 (2013) 2–13.
• 20)  M.A. Rhamdhani, J. Chen, T. Hidayat, E. Jak and P. Hayes: Advances in research on nickel production through the Caron process, Proceedings of EMC, (2009) pp. 899–913.
• 21)   M.H.  Caron: J. Metals 188 (1950) 91–103.
• 22)   M.  Landers and  R.  Gilkes: Appl. Clay Sci. 35 (2007) 162–172.
• 23)  M.A. Wells: Mineral, Chemical and Magnetic Properties of Synthetic, Metal-substituted Goethite and Hematite, (University of Western Australia, 1998).
• 24)   S.  Gialanella,  F.  Girardi,  G.  Ischia,  I.  Lonardelli,  M.  Mattarelli and  M.  Montagna: J. Therm. Anal. Calorim. 102 (2010) 867–873.
• 25)   Q.  Huang and  X.  Lv: J. Min. Metall., Sec. B: Metall. 47 (2011) 45–51.
• 26)  F. Crundwell, M. Moats, V. Ramachandran, T. Robinson and W.G. Davenport: Extractive Metallurgy of Nickel, Cobalt and Platinum Group Metals, (Elsevier, Amsterdam, 2011).
• 27)   M.  Valix and  W.  Cheung: Miner. Eng. 15 (2002) 607–612.
• 28)   J.  Yang,  G.  Zhang,  O.  Ostrovski and  S.  Jahanshahi: Miner. Eng. 54 (2013) 110–115.
• 29)   G.  Li,  J.  Luo,  Z.  Peng,  Y.  Zhang,  M.  Rao and  T.  Jiang: ISIJ Int. 55 (2015) 1828–1833.
• 30)   E.  Svana and  R.  Ysteb: (1983).
• 31)   H.  Tsuji and  N.  Tachino: ISIJ Int. 52 (2012) 1724–1729.
• 32)   H.  Tsuji: ISIJ Int. 52 (2012) 333–341.
• 33)   T.  Watanabe,  S.  Ono,  H.  Arai and  T.  Matsumori: Int. J. Miner. Process. 19 (1987) 173–187.
• 34)   Y.  Kobayashi,  H.  Todoroki and  H.  Tsuji: ISIJ Int. 51 (2011) 35–40.
• 35)   H.  Tsuji: ISIJ Int. 52 (2012) 1000–1009.
• 36)   H.  Tsuji and  N.  Tachino: ISIJ Int. 52 (2012) 1951–1957.
• 37)   R.  Elliott,  C.A.  Pickles and  J.  Forster: J. Miner. Mater. Charact. Eng. 04 (2016) 320–346.
• 38)  C. Hallet: A thermodynamic analysis of the solid state reduction of nickel from laterite minerals, Proc. Nickel–Cobolt 97 Int. Symp., ed. by C.A. Levac and R.A. Berryman, (1997) pp. 299–312.
• 39)   M.  Jiang,  T.  Sun,  Z.  Liu,  J.  Kou,  N.  Liu and  S.  Zhang: Int. J. Miner. Process. 123 (2013) 32–38.
• 40)   G.  Li,  T.  Shi,  M.  Rao,  T.  Jiang and  Y.  Zhang: Miner. Eng. 32 (2012) 19–26.
• 41)   D.Q.  Zhu,  Y.  Cui,  K.  Vining,  S.  Hapugoda,  J.  Douglas,  J.  Pan and  G.L.  Zheng: Int. J. Miner. Process. 106–109 (2012) 1–7.
• 42)   W.-R.  Liu,  X.-H.  Li,  Q.-Y.  Hu,  Z.-X.  Wang,  K.-Z.  Gu,  J.-H.  Li and  L.-X.  Zhang: Trans. Nonferrous Met. Soc. China 20 (2010) s82–s86.
• 43)   S.  Zhou,  Y.  Wei,  B.  Li,  H.  Wang,  B.  Ma and  C.  Wang: Sci. Rep. 6 (2016).
• 44)   J.  Dong,  Y.  Wei,  C.  Lu,  S.  Zhou,  B.  Li,  Z.  Ding,  C.  Wang and  B.  Ma: Mater. Trans. 58 (2017) 1161–1168.
• 45)   B.  Song,  X.  Lv,  H.H.  Miao,  K.  Han,  K.  Zhang and  R.  Huang: ISIJ Int. 56 (2016) 2140–2146.
• 46)   L.-w.  Wang,  X.-m.  Lü,  M.  Liu,  Z.-x.  You,  X.-w.  Lü and  C.-g.  Bai: Int. J. Miner. Metall. Mater. 25 (2018) 744–751.
• 47)   X.  Lv,  W.  Lv,  M.  Liu,  Z.  You,  X.  Lv and  C.  Bai: ISIJ Int. 58 (2018) 799–807.
• 48)   Z.  Meng,  H.  Jiang,  H.  Wang,  Y.  Huang,  Z.  Cui and  C.  Li: J. Mar. Sci. Eng. 1 (2012) 160–164.
• 49)   S.  Hou: Glass 4 (1981) 44–47.
• 50)   M.J.  Rao: Journal of Central South University (2010) 42–48.
• 51)   K.  Nagata and  P.  Bolsaitis: Int. J. Miner. Process. 19 (1987) 157–172.
• 52)   Y.  Okajima: J. Min. Metall. Inst. Jpn. 103 (1987) 115–120.
• 53)   M.  Molenda,  J.  Stengler,  M.  Linder and  A.  Wörner: Thermochim. Acta 560 (2013) 76–81.
• 54)   R.  Elliott,  F.  Rodrigues,  C.A.  Pickles and  J.  Peacey: Can. Metall. Quart. 54 (2015) 395–405.
• 55)  B.A. Wills and T. Napier-Mun: Wills’ Mineral Processing Technology, (Elsevier, Butterworth-Heinemann, 2006) pp. 267–352.
• 56)   N.F.  Schulz: Trans. SME-AIME 247 (1970) 56.
• 57)   C.  Lu,  J.  Dong,  B.  Li,  S.  Zhou,  H.  Wang and  Y.  Wei: Chin. J. Process 17 (2017) 1195–1202.
• 58)   D.Q.  Zhu,  G.  Zheng,  J.  Pan and  Z.  Liu: Journal of Central South University (Science and Technology) 44 (2013) 1–7.
• 59)   F.  Abdul,  S.  Pintowantoro,  A.  Kawigraha and  A.  Nursidiq: AIP Conf. Proc. 1945 (2018) 020034.
• 60)   F.  Abdul,  S.  Pintowantoro and  R.B.  Yuwandono: AIP Conf. Proc. 1945 (2018) 020033.
• 61)   G.-l.  Zheng,  D.-q.  Zhu,  J.  Pan,  Q.-h.  Li,  Y.-m.  An,  J.-h.  Zhu and  Z.-h.  Liu: Journal of Central South University 21 (2014) 1771–1777.
• 62)   M.  Liu,  X.  Lv,  E.  Guo,  P.  Chen and  Q.  Yuan: ISIJ Int. 54 (2014) 1749–1754.