Sulfide Route to Chromium–Nickel–Molybdenum Ferroalloys for Stainless Steel Production

Sulfur-based processes have recently emerged as environmentally sustainable and economically competitive avenues for materials separation and metal reduction.^[1,2] Herein, we utilize sulfur chemistry to demonstrate the direct production of an iron–chromium–nickel–molybdenum ferroalloy of similar composition to 316 stainless steel^[3] from a blend of mineral and surrogate precursors without the use of carbon reductants. This novel approach begins with sulfidation of hematite and chromite minerals, molybdenum concentrate, and synthetic nickel matte to form a mixed sulfide. This mixed sulfide intermediate is then aluminothermically reduced to form a crude alloy and a volatile aluminum sulfide (\(A{l}_{2}{S}_{3}\)) byproduct. The alloy is refined through contact with slags, enabling the production of ultralow carbon ferroalloys and stainless steels without additional refining via argon-oxygen decarburization (AOD). Aluminum may in principle be regenerated through electrolysis^[4] of distilled \(A{l}_{2}{S}_{3}\) byproduct^[5] to enable new synergies with aluminum scrap recycling. A reduction in the net aluminum loss to slag is expected compared to the aluminothermic reduction of oxides.^[6] Compared with conventional carbothermic reduction pathways for ferronickel, ferrochromium, and other metals from their mineral oxides,^[7,8] this new sulfur-based production pathway could enable ferroalloy and stainless steel production without direct greenhouse gas emissions.

Carbon-free sulfidation of a metal oxide feedstock (\(MO\)) with gaseous elemental sulfur (\({S}_{2}\)) to produce a metal sulfide intermediate (\(MS\)) is described via the following reaction:

Such reactions can be written for any oxide and sulfide product chemistry and stoichiometry.^[1] Sulfur dioxide (\(S{O}_{2}\)) generated through oxide sulfidation is readily sequestered and valorized through the production of byproduct sulfuric acid^[9] or may be reduced to elemental sulfur.^[10] From Le Chatelier’s principle, the ratio of \({S}_{2}\) and \(S{O}_{2}\) partial pressures (\({P}_{{S}_{2}}\) and \({P}_{S{O}_{2}}\) respectively) in the system can drive Eq. [1] toward the reactants or products. The critical ratio of \({S}_{2}\) to \(S{O}_{2}\), \({\left[{P}_{{S}_{2}}/{P}_{S{O}_{2}}\right]}_{crit}\), required in the system for the thermodynamic spontaneity of Eq. [1] may be quantified as a function of the standard Gibbs energy of reaction and reactant and product activities (\(a\)), as derived elsewhere.^[1] Any oxide can be sulfidized with \({S}_{2}\) alone through the use of a high enough \({P}_{{S}_{2}}\), a sufficiently long reaction time, and a critically short gas residence time in the reactor.^[1] Assuming immiscible compounds, each at unit activity (\(a=1\)), allows to calculate \({\text{log}}_{10}\left({\left[{P}_{{S}_{2}}/{P}_{S{O}_{2}}\right]}_{crit}\right)\) for some species relevant to stainless steel production, as depicted in Figure 1(a).

Full size image

In Figure 1(b), the activities of iron, chromium, and sulfur in a molten iron sulfide (\(FeS\)) matte of increasing chromium sulfide (\(CrS\)) content are depicted up to the reported^[11] solubility limit of \(CrS\) in the matte at 1400 °C. Activities were calculated via FactSage 8.0 using the Equilib module with the FTmisc database, assuming a single liquid matte product phase (FTmisc-MAT2), referenced to pure liquid iron, chromium, and sulfur end members (\(Fe(l)\), \(Cr(l)\), and \(S(l)\) respectively) at 1400 °C and 1 atm. These activities allow to calculate the \({\text{log}}_{10}\left({\left[{P}_{{S}_{2}}/{P}_{S{O}_{2}}\right]}_{crit}\right)\) for sulfidation of chromite (\(FeC{r}_{2}{O}_{4}\)) into a liquid \(FeS\) matte of increasing \(CrS\) content, as reported in Figure 1(c). At lower \(CrS\), favorable mixing of \(CrS\) with \(FeS\) increases the sulfidation affinity of \(FeC{r}_{2}{O}_{4}\) by lowering the thermodynamic barrier to sulfidation and decreasing \({\left[{P}_{{S}_{2}}/{P}_{S{O}_{2}}\right]}_{crit}\). This motivates the direct sulfidation of blended minerals to form a sulfide matte that mirrors the metal composition of the target alloy product, as opposed to the production of individual ferroalloys of lower iron content. Sulfide mattes of lower iron content can be produced through leveraging higher \({P}_{{S}_{2}}\), longer reaction times, and shorter gas residence times during sulfidation.^[1] Ferrous scrap can be blended with feedstocks before or after sulfidation to minimize the need for primary mineral precursors.

Ferroalloy or stainless steel products are then produced via reduction of the mixed sulfide. Aluminothermic reduction has been successfully demonstrated for production of iron–nickel–silicon alloys from sulfide precursors.^[12] The aluminothermic reduction of a metal sulfide (MS) to form the metal (M) and volatile aluminum sulfide (\(A{l}_{2}{S}_{3}\)) is described by the following reaction:

Reactions such as Eq. [2] can be written for any sulfide matte chemistry and stoichiometry.^[13] In a system open to \(A{l}_{2}{S}_{3}\), at temperatures above 1300 °C and pressures on the order of 10⁻³ atm, \(A{l}_{2}{S}_{3}\) is readily distilled from the liquid, driving the reaction toward completion.^[5] Subsequent electrolytic reduction of \(A{l}_{2}{S}_{3}\) byproduct to regenerate aluminum metal requires less energy than the conventional Hall-Héroult process for aluminum production from its oxide.^[4]

By eliminating carbon as a reductant, excess carbon in the ferrous product is minimized and AOD may no longer necessary for refining. Slags can be utilized for desulfurization, dephosphorization, and management of metallic impurities.^[14,15] Together, sulfidation, aluminothermic reduction via reactive vacuum distillation, and slag refining enable a potential paradigm shift in sustainable and efficient metal production.

Hematite rock mineral from the United States (Ward’s Scientific, < 230 mesh), chromite rock mineral from the United States (eBay, < 325 mesh), and molybdenite concentrate (JX Nippon Mining and Metals Corporation, fine powder) were employed as iron, chromium, and molybdenum sources to experimentally verify the process. A synthetic nickel matte of similar composition to those produced industrially from sulfide ores^[16] was made by melting 50 g of iron sulfide (\(FeS\), 99.9 pct metals basis purity, Strem Chemicals), 20 g of nickel sulfide (\(N{i}_{3}{S}_{2}\), 99.7 pct metal basis purity, Aldrich), 2 g of copper sulfide (\(C{u}_{2}S\), 76.3–83.4 pct complexometric EDTA copper content, Aldrich), and 2 g of cobalt sulfide (\(C{o}_{9}{S}_{8}\), 99.5 pct metals basis purity, Alfa Aesar) in a graphite crucible at 1300 °C under a nitrogen atmosphere (\({N}_{2}\), 99.999 pct, Linde) for 2 hours and was then ground to < 230 mesh using methods described previously.^[17] Comminution of samples to reported particle sizes was accomplished using a ring and puck mill (Mini Thor mill, Rocklabs) and sampled via cone and quartering. Hematite and chromite rocks, synthetic nickel matte, and molybdenite concentrate were blended in a ratio of 13.93–7.96 g to 7.16–0.85 g, respectively, to serve as the sulfidation feedstock.

Sulfidation of the blended feedstock was conducted in an alumina packed bed reactor of geometry described elsewhere^[18] held within an inert nitrogen atmosphere (\({N}_{2}\), 99.999 pct, Linde) vertical tube furnace at a total pressure of 1.1 atm with vaporized elemental sulfur (\({S}_{2}\), 99.0 pct, Merck KGaA) at a \({P}_{{S}_{2}}\) of 0.4–0.6 atm and a gas residence time through the reactor of 0.2 s, employing methods and equipment reported previously.^[1] Three sets of sulfidation conditions were tested: 2 hours at 1300°C (Condition 1), 2 hours at 1400 °C (Condition 2), and 4 hours at 1400 °C (Condition 3). Sulfidation products from Condition 3 were subjected to aluminothermic reduction via reactive vacuum distillation. Aluminum metal (Al, 99.99 pct metals basis purity, Thermo Scientific) is the reductant, added in the stoichiometric amount of sulfur (S²⁻) that should be present in the mixed sulfides; (4.08 g for the 23.23 g of matte). The aluminum and sulfide feedstocks were mixed and held in a flat-bottom zirconia crucible (35 mm OD, 28 mm ID, 120 mm height, 110 mm depth) within a graphite induction heating susceptor. Aluminothermic reduction via reactive vacuum distillation was conducted in a vacuum induction tube furnace at a temperature of 1550 °C for 30 minutes at a total pressure of 10⁻³ atm, utilizing methods and equipment described previously.^[5] A distillate product was observed to condense at and react with the interface between the wall of the zirconia crucible and the top of the graphite heating susceptor.

Following reduction, the crude metal product was sequentially contacted with two different slags for proof of concept refining. The first slag was made from calcium oxide (\(CaO\), 99.95 pct metals basis purity, Thermo Scientific), aluminum oxide (\(A{l}_{2}{O}_{3}\), 99.95 pct metals basis purity, Alfa Aesar), chromite rock mineral, and iron oxide (\(F{e}_{2}{O}_{3}\), 99.85 pct metals basis purity, Alfa Aesar) powders mixed in the proportions of 5.35–5.30 g to 1.07–0.57 g, respectively. The crude alloy and the oxide powders, at a ratio of 5.31–12.29 g, were melted together in a round-bottom alumina crucible (41 mm OD, 33 mm ID, 90 mm height, 86 mm depth) placed in a graphite induction heated susceptor at 1550 °C for 30 minutes under an argon atmosphere (\(Ar\), 99.999 pct. Linde), using methods and equipment described previously.^[12] The second slag consisted of \(CaO\), silica (\(Si{O}_{2}\), 99.5 pct purity, Alfa Aesar), \(A{l}_{2}{O}_{3}\), chromium oxide (\(C{r}_{2}{O}_{3}\), 99 pct purity, Acros Organics), and magnesium oxide (\(MgO\), 99.95 pct metals basis purity, Alfa Aesar) mixed in a ratio of 1.39 g to 0.94–0.93 g to 0.19–0.045 g, respectively. The alloy and oxide powders at a ratio of 1.72–1.30 g were melted together in a flat-bottom alumina crucible (19 mm OD, 14 mm ID, 101 mm height, 99 mm depth) within a tantalum induction heated susceptor at 1550 °C for 30 minutes in a vacuum induction tube furnace at a system pressure of 10⁻³ atm, using methods and equipment described previously.^[12]

The bulk compositions of feedstocks, intermediates, and products were determined via ICP-AES and LECO, ICP-OES and LECO, or SEM-EDS (JEOL JSM-6610LV, JEOL, Sirius SD detector, SGX Sensortech) and are reported in Table I. The crystalline phases present in feedstocks, sulfidation intermediates, and the distillate product were quantified via QXRD (Panalytical X’pert MPD diffractometer, Cu radiation at 45 KV/40 ma, scan angle of 6–80 deg, step size of 0.0131 deg, counting time of 250 seconds/step) and are reported in Table II. The spatial distribution of elements in samples following sulfidation as determined via SEM-EDS mapping is reported in Figure 2. Metal products after reduction and the first slag treatment step are shown in Figure 3.

Full size image

Full size image

From element mapping depicted in Figure 2, sulfidation at 1300 °C and 1400 °C for 2 hours yielded sulfides of iron, nickel, and molybdenum containing some chromium. Unsulfidized chromium species formed magnesiochromite as reported in Table I. When the sulfidation time at 1400 °C was increased from 2 to 4 hours, the majority of chromium was then sulfidized and limited magnesiochromite was observed. The addition of silica in the system from the hematite increased the acidity of oxide phases, previously shown to aid in sulfidation of other transition metals.^[13] Overall, these results demonstrate that carbon-free sulfidation can be leveraged to form mixed sulfide intermediates (a “matte”) of iron, chromium, nickel, and molybdenum for use in subsequent carbon-free metal reduction.

The subsequent aluminothermic reduction of the sulfidized product from Condition 3 via reactive vacuum distillation yielded a crude metal product, a slag phase, and a distillate. Figure 3(a) shows the crude metal product and distillate. Figure 3(b) shows the semirefined metal product following the first slag refining. As reported in Table I, the crude metal product exhibited the approximate ratio of iron to chromium to nickel to molybdenum of 316 stainless steel.^[3] However, aluminum, silicon, and zirconium impurities were also present in the crude metal product at levels from 2 to 5 wt pct, indicating that excess aluminum reductant was present and reacted with silica-rich gangue and the zirconia crucible. The crude metal product also contained a residual bulk sulfur content of around 1 wt pct. Zirconium and sulfur were concentrated in precipitant phases as shown in Figures 3(c) through (e), with compositions reported in Table II.

Under the conditions employed herein, appreciable vacuum thermal decomposition of iron and nickel sulfides likely occurs,^[12] suggesting a less than stoichiometric amount of aluminum reductant may be needed. Within the distillate, magnesium and silicon impurity compounds were present as reported in Table II, indicating that cation exchange between aluminum sulfide distillate and gangue oxides occurred.^[13] A non-negligible amount of Si and possibly some Mg may have interacted with either the Al₂S₃(g) or directly as gaseous sulfides and ended up being distilled. Those components then back reacted with the oxide refractory material as witness by the formation of spinel, mullite and forsterite from Table II.

Separation of the sulfide intermediate from oxide gangue is likely necessary prior to aluminothermic reduction to enable recovery of high purity aluminum compounds, as observed previously.^[5] Matte-gangue separation prior to reduction may be conducted via melting and matte-slag immiscibility or comminution and physical separation. With optimization of gangue removal prior to reduction, high purity \(A{l}_{2}{S}_{3}\) byproduct distillates may be recovered, allowing regeneration of the Al and sulfur recycling, for example by electrolysis.^[4]

The aluminum, silicon, and zirconium contents of the crude metal were reduced to around 1 wt pct in a semirefined metal product following contact with the first slag, with the droplet product shown in Figure 3(b). Residual zirconium was observed to be concentrated in iron–chromium-rich precipitants as depicted in Figure 3(f) through (h), with compositions reported in Table II. Within the precipitants, residual sulfur was concentrated in chromium-rich phases. Through subsequent contacting of semirefined metal product with the second slag, a final refined metal product was produced. Future process optimization may enable refining to be conducted in a single stage with minimized slag utilization.

The final refined metal product contained chromium, nickel, and molybdenum at 13.65 wt pct, 10.70 wt pct, and 3.33 wt pct, respectively, near that of commercial 316 stainless steel (16.00–18.00 wt pct, 10.00–14.00 wt pct, and 2.00–3.00 wt pct, respectively)^[3] as reported in Table I. Table III shows the estimated recovery established by tracking the relative amount of Fe, Cr to the Ni content from the blend of minerals to the matte and then the crude metal product. The amount of slag from sulfidation is calculated from the mass balance on Al, Si, Mg, and Ca, which do not report to the matte. The EDS map indicates such slag contains about 1 wt pct Ni and about 3 pct Fe, which provides a mean to estimate the recovery of the sulfidation step. Molybdenum is not presented because of difficulty of EDS to distinguish S from Mo. For the crude metal, Ni is assumed to be completely reporting to the metal phase considering its lower reduction potential and lack of volatility and reactivity during the reduction process.

Aluminum, magnesium, calcium, manganese, titanium, copper, and zirconium contents only remained at trace levels. Vanadium, tungsten, and cobalt were detected in the refined product at levels of 0.028 wt pct, 0.077 wt pct, and 0.512 pct, respectively, along with traces of arsenic. Vanadium, tungsten, and arsenic likely originated from the mineral feedstocks, highlighting the importance of material sourcing for impurity management from feedstock minerals and scrap metals. Future work is needed to track the propagation of transition metal impurities through sulfidation and reduction steps.

The carbon content of the metal product was observed to be 0.007 wt pct, an order of magnitude lower than the commercial specification for 316 of 0.08 wt pct and over 75 pct lower than the commercial specification for 316 L of 0.03 wt pct.^[3] The phosphorous content of the metal product was observed to be 0.156 wt pct, higher than that of commercial 316 stainless steel at 0.045 wt pct. The elevated phosphorous content originated from the high phosphorous hematite feedstock employed, which was not particularly chosen to represent a high purity iron ore concentrate. The sulfur content of 0.06 wt pct was also slightly elevated versus the commercial specification of 0.03 wt pct. Industrial practice shows that slag chemistries may be optimized to sequester sulfur and phosphorous impurities at levels necessary to meet product specifications.^[14,15]

These results indicate that for the first time, sulfide chemistry may be employed to produce ultralow carbon chromium–nickel–molybdenum ferroalloys or stainless steels directly from blended mineral precursors without the use of carbon as a reductant. Using aluminum as a reductant for sulfide chemistries in principle enables the formation of a volatile \(A{l}_{2}{S}_{3}\) byproduct. Separation of sulfide matte from magnesium and silicon-rich gangue prior to reduction is needed to enable distillation of \(A{l}_{2}{S}_{3}\) from the system for direct aluminum remanufacture and elemental sulfur recovery via electrolysis.

Hematite was utilized as an iron source herein; iron and other alloying agents may instead be supplied to the process through steel or stainless steel scrap, enabling full integration of sulfide processes with established ferrous recycling supply webs. While processing chemistries and feedstocks for sulfidation and reduction must still be optimized, this new approach may also eliminate the need for AOD, with crude metal products instead only requiring slag refining. Integrated with green electricity, low carbon feedstocks, and carbon-free auxiliary processes, sulfide-based processing pathways may enable ferroalloy and stainless steel manufacturing from both mineral and recycled sources without direct greenhouse gas emissions.