Manganese nodules have been investigated as an alternative source of rare metals (or critical metals) such as nickel and cobalt. The composition of manganese nodules is listed in Table I.[1,2] One effective method to extract these rare metals from manganese nodules is the reductive smelting—sulfurization (alloy-to-matte conversion)—hydrometallurgy (RSH) combinatorial process.[1,2] Details of the RSH process have been presented in previous articles.[1,2,3] The Fe-Ni-Cu-Co alloy, which is produced by electric arc furnace (EAF) smelting, is very difficult to crush and leach because of the high energy required for grinding and the low leaching rate caused by the high iron content.[4]

Table I Composition of Manganese Nodules (Mass Percent)[1]
Full size table

To resolve these problems, the sulfurization step, viz. conversion of the Fe-Ni-Cu-Co alloy to the Fe-Ni-Cu-Co-S matte phase, is necessary in the RSH process.[3,4,5] Elemental sulfur has been added to the molten alloy in the conventional matte-making (sulfurization) method.[3,4,5] However, because the conversion efficiency is relatively low and the cost of pure sulfur is high, we recently proposed that dehydrated waste gypsum (CaSO4) can be a good alternative sulfur source (in conjunction with appropriate amounts of carbon as a reducing agent) for the conversion of alloy to the matte phase.[3,4,5] It was also confirmed in our previous study that the addition of Al2O3-Fe2O3 flux (reagent-grade chemicals) was very effective in the sulfurization process because the intermediate reaction product, i.e., calcium ferrite slag, has a relatively low melting point and thus accelerates the sulfurization reaction.[3]

Alternatively, many researchers have investigated the potential utilization (or valorization) of industrial waste, such as mill scale [MS], red mud [RM], and Al-dross [AD], to create high-added value and reduce environmental problems.[6,7,8,9,10,11,12,13] Therefore, in the present study, we tried to experimentally confirm the feasibility of direct use of industrial wastes as inexpensive fluxing agents to promote the sulfurization efficiency in the matte-making process because these wastes contain useful oxides such as Fe2O3 and Al2O3.

The experimental apparatus, procedure, and conditions are described in detail in our previous article.[3] We carried out metal-slag-gas three-phase equilibration experiments using a vertical electric resistance furnace at 1673 K (1400 °C) under conditions of p(O2) = 1.0 × 10−10 atm and p(SO2) = 1.0 × 10−3 atm, which was confirmed from a CO-CO2-SO2-Ar gas mixture. The impurities in the gas mixture were removed by passage through a purification system. The 17.5Fe-43.5Ni-33.2Cu-5.6Co (mass pct) alloy (produced by the smelting reduction of a manganese nodule in an EAF),[3] reagent-grade CaSO4 (S source), solid carbon (reducing agent), and three kinds of industrial waste (fluxing agents: mill scale, red mud, and Al-dross) were loaded in a fused alumina crucible [outer diameter (OD): 35 mm, inner diameter (ID): 30 mm, height (HT): 50 mm], which was held in a porous alumina holder (OD: 45 mm, ID: 40 mm, HT: 60 mm). The crucible assembly was connected to a molybdenum wire. The composition of the industrial wastes used in the present study is listed in Table II.

Table II Composition of Mill Scale, Al-Dross, and Red Mud Used in Present Study (Mass Percent)
Full size table

After equilibration was attained, the sample was quickly extracted from the hot zone of the furnace and directly quenched by dipping into brine. The slag and metal samples were carefully separated from the crucible. The slag samples were crushed to fine powder and filtered using a stainless steel sieve for chemical analysis. The sulfur content in the samples was determined using a combustion analyzer (CS-800, LECO). Furthermore, the slag phase was analyzed using X-ray fluorescence spectroscopy (XRF; ZSX Primus II, Rigaku) and an X-ray diffractometer (XRD; D/MAX-2500PC, Rigaku) to confirm the chemical composition and mineral phase formed during the experiments, respectively.

Before performing the experiments, morphology and crystallographic analyses were performed using XRD and SEM-EDS methods. The mineral phases in each material identified by XRD analysis are shown in Figure 1. Metallic Al and corundum phase (Al2O3) in conjunction with a spinel phase ([Mg,Mn]Al2O4) were identified in Al-dross. Hematite (Fe2O3) was the major phase in red mud, while quartz (SiO2) and wollastonite (CaSiO3) phases also appeared. There were various valences of iron, i.e., hematite, magnetite, and wüstite, in the steel plant mill scale.

Fig. 1
figure 1

X-ray diffraction patterns of Al-dross, red mud, and mill scale

Full size image

Backscattered electron SEM images of Al-dross combined with EDS mapping for the major elements (Al, Si, Mg, and O) are shown in Figure 2. Fine metallic Al particles with different sizes ranging from approximately 10 to 100 μm are dispersed in Al-dross, of which the average diameter of 15 μm was analyzed using an automatic image analyzing software. Furthermore, Al2O3 matrix containing small amounts of SiO2 and MgO constitutes the bulk of the Al-dross.

Fig. 2
figure 2

Backscattered SEM images of Al-dross combined with EDS element mapping

Full size image

However, there is a more complicated structure of red mud, as shown in Figure 3. Fe2O3 and SiO2 are distinctly confirmed in red mud by SEM-EDS and XRD analyses. Additionally, other minor phases, such as Al2O3, (Ca,Mg)O, and (Na,K)AlSiO4, were also identified, representing a very complicated and diverse composition of red mud as shown in Table II. Alternatively, Qin and Wu[14] reported that there were more complicated phases such as CaTiO3, FeTiO3, K(Na,K)3Al4Si4O16 and Al2Ca2FeSi3H4O14 in red mud by XRD analysis. Thus, it is not easy to conclude which phase is the most dominant in red mud; that is dependent on the location and original mineral sources.

Fig. 3
figure 3

Backscattered SEM image of red mud combined with EDS element mapping

Full size image

The mill scale supplied from the steel plant was mainly composed of various forms of iron oxide and irregular shapes of iron oxide particles are observed in Figure 4. Based on the basic information described above, we will discuss the effect of each industrial waste material as a fluxing agent on the sulfurization efficiency of the Fe-Ni-Cu-Co alloy produced from an EAF smelting of manganese nodules.

Fig. 4
figure 4

Backscattered SEM image of mill scale

Full size image

The influence of the addition of waste on the sulfurization efficiency of Fe-No-Cu-Co alloy at 1673 K (1400 °C) is shown in Figure 5. The sulfurization efficiency, SE (\( \eta_{\text{sulf}} \)), was defined by Eq. [1] in our previous article.[14]

Fig. 5
figure 5

Fluxing effect of industrial wastes on the sulfurization efficiency of Fe-Ni-Cu-Co alloy at 1673 K (1400 °C). Numbers on the x-axis represent mass pct of each flux (RM red mud, AD Al-dross, MS mill scale). The experimental results using reagent-grade chemicals were taken from Ref. [3]

Full size image
$$ {\text{Sulfurization}}\;{\text{Efficiency,}}\;\eta_{\text{sulf}} \; ( {\text{pct)}} = \frac{{ [ {\text{pct}}\;{\text{S]}}^{\text{Alloy}} }}{{ [ {\text{pct}}\;{\text{S]}}_{\text{sat}}^{\text{Matte}} (= 2 7\;{\text{pct)}}}} \times 1 0 0 .$$
(1)

Here, the saturation sulfur content to maintain the matte phase was approximately 27 pct at 1673 K (1400 °C), which was calculated by FactSageTM software in a previous study.[3] The gypsum equivalent (Geq = 1.0) and carbon equivalent (Ceq = 1.0) represent the theoretical amount of CaSO4 required for complete sulfurization and the theoretical molar amount of carbon for complete reduction of CaSiO4 based on Eq. [2], respectively. Values of Geq = 1.7 and Ceq = 0.7 for the effective sulfurization by carbothermic reduction of CaSO4 were obtained in our previous experiments.[3] The effect of carbon addition on sulfurization efficiency was understood based on the following reactions.[3]

$$ {\text{CaSO}}_{ 4} \left( {\text{s}} \right) \, + {\text{ 4C}}\left( {\text{s}} \right) \, = {\text{ CaS}}\left( {\text{s}} \right) \, + {\text{ 4CO}}\left( {\text{g}} \right), $$
(2a)
$$ \Delta G^{ \circ } = \, - \; 653{,}950{\text{ J}}/{\text{mol at 1673 K }}\left( { 1 400\;{}^{ \circ }{\text{C}}} \right). $$
(2b)

Furthermore, it was proposed that CO gas produced by Eq. [2a] indirectly reduced CaSiO4 by producing CaS as well as CaO with gaseous species such as CO2 and SO2 via Eqs. [3] and [4]. The Boudouard reaction (Eq. [5]) also contributed to producing CO gas.

$$ {\text{CaSO}}_{ 4} \left( {\text{s}} \right) \, + {\text{ 4CO}}\left( {\text{g}} \right) \, = {\text{ CaS}}\left( {\text{s}} \right) \, + {\text{ 4CO}}_{ 2} \left( {\text{g}} \right) $$
(3a)
$$ \Delta G^{ \circ } = \, - \; 172{,}840{\text{ J}}/{\text{mol at 1673 K }}\left( { 1 400\;{}^{ \circ }{\text{C}}} \right) $$
(3b)
$$ {\text{CaSO}}_{ 4} \left( {\text{s}} \right) \, + {\text{ CO}}\left({\text{g}} \right) \, = {\text{ CaO}}\left({\text{s}} \right) \, + {\text{ CO}}_{2} \left({\text{g}} \right) \, + {\text{ SO}}_{ 2} \left( {\text{g}} \right) $$
(4a)
$$ \Delta G^{ \circ } = \, -\, 8 3{,} 2 90{\text{ J}}/{\text{mol at 1673 K }}\left( { 1 400\;{}^{ \circ }{\text{C}}} \right) $$
(4b)
$$ {\text{C}}\left( {\text{s}} \right) \, + {\text{ CO}}_{ 2} \left( {\text{g}} \right) \, = {\text{ 2CO}}\left( {\text{g}} \right) $$
(5a)
$$ \Delta G^{ \circ } = \, -\, 1 20{, }300{\text{ J}}/{\text{mol at 1673 K }}\left( { 1 400\;{}^{ \circ }{\text{C}}} \right) $$
(5b)

The sulfurization efficiency (SE) at 10 to 15 pct red mud fluxing and 10 pct Al-dross fluxing was approximately 74 to 78 pct as shown in Figure 5. However, SE increased to approximately 89 pct at ‘5 pct Al-dross + 5 pct mill scale’ fluxing. Hence, combinatorial fluxing with Al-dross and mill scale was cost-effective and the most efficient compared to the single addition of each waste material. Iron oxide in the mill scale promoted the decomposition of CaSO4 due to the formation of the CaO-FetO based slag system. In a previous study,[3] we calculated the liquidus surface of CaO-SiO2, CaO-Al2O3, and CaO-FetO binary systems by FactSageTM software and concluded that the CaO-FetO system exhibits much lower liquidus than other systems, resulting in an acceleration of the sulfurization process. Several authors also reported that the FetO is a suitable catalyst for CaSO4 decomposition.[15,16,17]

In addition, it is noteworthy in Figure 5 that the addition of industrial wastes for the sulfurization (matte-making) process exhibits a similar or higher efficiency compared to the fluxing efficiency with reagent-grade chemicals. Specifically, the addition of 10 pct Al-dross (SE = 77 pct) is superior to 10 pct pure Al2O3 fluxing (SE = 56 pct) due to the existence of various kinds of impurity oxides in Al-dross such as SiO2 and MgO. The wide variation of slag composition in the multi-component system could provide lower liquidus, resulting in the promotion of sulfurization efficiency. Moreover, combinatorial fluxing with 5 pct Al-dross and 5 pct mill scale exhibits a similar SE (= 89 pct) compared to the SE (= 88 pct) with reagent-grade chemicals, viz. 5 pct Al2O3 and 5 pct Fe2O3.

The precipitation behavior of solid compounds in the slag during the sulfurization reaction was characterized by XRD analysis as shown in Figure 6. CaS and Ca2SiO4 compounds were commonly precipitated when red mud and Al-dross were added separately, whereas only CaS was precipitated when Al-dross and mill scale were added as a fluxing agent. Based on our previous findings,[3] CaS is mostly induced by Eqs. [2] and/or [3], and there is an additional thermodynamic driving force of the following reaction to produce CaO as given in Eq. [6].

$$ 3 {\text{CaSO}}_{ 4} \left( {\text{s}} \right) \, + {\text{ CaS}}\left( {\text{s}} \right) \, = {\text{ 4CaO}}\left({\text{s}} \right) \, + 4 {\text{SO}}_{ 2} \left( {\text{g}} \right) $$
(6a)
$$ \Delta G^{ \circ } = \, - \,160{,}300{\text{ J}}/{\text{mol at 1673 K }}\left( { 1 400\;{}^{ \circ }{\text{C}}} \right) $$
(6b)
Fig. 6
figure 6

X-ray diffraction pattern of slag samples after carbothermic reduction of CaSO4 and the sulfurization process fluxed by red mud, Al-dross, and mill scale

Full size image

Specifically, Ca2SiO4 was precipitated by the reaction between CaO and SiO2; the latter was supplied from red mud and Al-dross. Precipitation of Ca2SiO4 increases the melting point and thus decreases the fluidity of slag, resulting in a retardation of the sulfurization reaction.[3] In addition, MS is a steel mill by-product, mainly wüstite (FeO ≅ 66 pct) rather than magnetite and hematite (Fe3O4 + Fe2O3 ≅ 34 pct). Thus, it is not easy to precipitate Ca2Fe2O5 when MS is used compared to the use of reagent-grade Fe2O3 because of the lower melting point of FeO than Fe2O3. Therefore, it is suggested that low melting point slag with a calcium ferrite-based composition was formed, resulting in promotion of the sulfurization reaction without precipitation of the Ca2Fe2O5 compound.

In summary, we considered the sulfurization (matte-making) process using several industrial wastes such as mill scale, red mud, and Al-dross as fluxing agents. It was experimentally confirmed that the addition of industrial waste for the sulfurization process leads to similar or higher sulfurization efficiency than the use of reagent-grade chemicals as a fluxing agent. Specifically, combinatorial fluxing with appropriate amounts of Al-dross and mill scale exhibited a superior sulfurization efficiency compared to fluxing with a single component because of the formation of a calcium ferrite system, which has a relatively low melting point. Consequently, the present results can be utilized in the cost-effective sulfurization of the Fe-Ni-Cu-Co alloy to Fe-Ni-Cu-Co-S matte using cheap industrial waste.