Basic-oxygen-furnace steel slag secondary resource
Steelmaking slag identified by primary studies as basic-oxygen-furnace (BOF), or Linz-Donawitz converter, slag. The separately studied feeds are a December 2016 sample from Yarborough, Scunthorpe, UK, and a sample supplied by an unnamed Austrian steelmaking company in the 2022 study. This process-defined secondary resource is distinct from electric-arc-furnace slag or dust, blast-furnace residues, copper smelter slag, constituent elements and individual mineral species. These heterogeneous samples do not define all BOF slags. No inherited elemental identifier or criticality designation is assigned.
Microbial & Biochemical Mechanisms (4)
Metal mobilization using an acidophilic culture and its supernatant
bioleaching
Microbial
An adapted, unresolved acidophilic community and its filter-sterilized supernatant were associated with higher aqueous Al, Cr and V release than fresh-medium controls in laboratory slag assays. Acid-conditioned liquid can therefore mobilize these metals without retained live cells in the leaching stage. The study does not isolate a particular organism, gene, enzymatic pathway or advantage over a chemistry-matched acid control. Reported extraction percentages do not establish a closed feed-to-product mass balance.
Experimental Context & Analytical Limits:
The inoculum combined water, minerals and biomass from sediment-water interfaces in Libiola Fe-Cu mine ponds and overflows, Italy, initially pH 1-3. Modified 9K medium contained 22.5 g/L FeSO4.7H2O and 10 g/L elemental sulfur; initial pH 2 and adaptation pH below 2.5 were adjusted with reagent H2SO4. Culture conditioning used 30 degrees C, 150 rpm, 10% inoculum and serial slag additions reported as 1, 2 and 5% (w/v). Slag was autoclaved at 135 degrees C for 30 minutes before adaptation. Activity was inferred from pH and redox potential, without direct cell-growth measurements establishing activity during alkaline leaching. Triplicate batch assays used 1 g of ground slag below 75 micrometres or below 2 mm, 90 mL medium and 10 mL inoculum, at 30 degrees C and 150 rpm for 3 or 6 days, with uninoculated controls. Batch pH rose from 3.3 +/- 0.6 to 9.4 +/- 0.1. Separate 22-day columns used 100 g coarse slag, a 1 L recirculating liquid inventory and 5 mL/min flow at 28 +/- 1 degrees C. The printed coarse-size inequality is internally inconsistent (4 < size > 10 mm); an exact interval is not asserted. Column A received 0.22-micrometre-filtered culture supernatant, B the adapted culture and C sterile fresh medium. Independent column replication is unspecified; repeated samples are not independent columns. Aqueous ICP-OES samples were filtered at 0.45 micrometres. These are laboratory experiments, not industrial deployment or isolated-metal yields.
The inoculum combined water, minerals and biomass from sediment-water interfaces in Libiola Fe-Cu mine ponds and overflows, Italy, initially pH 1-3. Modified 9K medium contained 22.5 g/L FeSO4.7H2O and 10 g/L elemental sulfur; initial pH 2 and adaptation pH below 2.5 were adjusted with reagent H2SO4. Culture conditioning used 30 degrees C, 150 rpm, 10% inoculum and serial slag additions reported as 1, 2 and 5% (w/v). Slag was autoclaved at 135 degrees C for 30 minutes before adaptation. Activity was inferred from pH and redox potential, without direct cell-growth measurements establishing activity during alkaline leaching. Triplicate batch assays used 1 g of ground slag below 75 micrometres or below 2 mm, 90 mL medium and 10 mL inoculum, at 30 degrees C and 150 rpm for 3 or 6 days, with uninoculated controls. Batch pH rose from 3.3 +/- 0.6 to 9.4 +/- 0.1. Separate 22-day columns used 100 g coarse slag, a 1 L recirculating liquid inventory and 5 mL/min flow at 28 +/- 1 degrees C. The printed coarse-size inequality is internally inconsistent (4 < size > 10 mm); an exact interval is not asserted. Column A received 0.22-micrometre-filtered culture supernatant, B the adapted culture and C sterile fresh medium. Independent column replication is unspecified; repeated samples are not independent columns. Aqueous ICP-OES samples were filtered at 0.45 micrometres. These are laboratory experiments, not industrial deployment or isolated-metal yields.
Curated Evidence Points:
Support
Source:
gomes_2018
Methods 2.1-2.7 and Results 3.2-3.3/Figures 1-3 support the conditional aqueous-mobilization observation. Three versus six days did not significantly change the batch result; smaller particles enhanced Al and Cr release but not V. Column culture and supernatant results were not statistically distinguishable in the authors' tests, whereas both differed from fresh medium. These null comparisons are retained; statistical nonsignificance does not establish equivalent performance.
Partial
Source:
gomes_2018
Column A and B began below pH 1, unlike the higher-pH fresh-medium control. Adaptation also added reagent acid. Unequal initial chemistry, unresolved column replication and activity proxies prevent attributing an incremental live-cell effect or a unique polysulfide mechanism. Kinetic fits in Table 2 do not distinguish a shrinking reactive core from a secondary crust controlling diffusion; neither proves enzyme causality. PHREEQC oxidation-state predictions are models, not direct measurements of aqueous Al, Cr or V speciation.
Partial
Source:
gomes_2018
Quantitative source inconsistencies remain unresolved. Table 1 gives nitric-acid-digest Al 0.4 +/- 0.07 wt%, Cr 100 +/- 20 mg/kg and V 40 +/- 10 mg/kg. Together with the reported batch loading these do not reconcile Figure 1 concentrations with Figure 2 percentages. Column Figure 3 cumulative Cr and V masses likewise fail to reproduce the text's reported percentages using Table 1 and 100 g feed. The column text gives Al 95% and 109% for A and B, whereas the abstract generalizes to 100%; these are not validated recoveries. Nonzero day-0 inventories, incomplete specification of cumulative sampling corrections and an unproven total-feed digestion denominator preclude repair by simple renormalization. Table 1 Mg 29.4 wt% is also inconsistent with the prose claim that it falls within the cited 0.3-2.4% typical range. No universal composition, reconciled yield or exact concentration observations are imported.
Partial
Source:
gomes_2018_supplement
Table S1 is preliminary T-RFLP community profiling, with assigned Acidithiobacillus species labels and unresolved members. It does not authenticate pure strains or isolate a member's causal role, despite stronger dominant-strain wording in the main paper. No taxon identifier is assigned to this mixed community. Figures S1-S2 support gypsum and jarosite diffraction assignments in A/B and gypsum as the principal crystalline control product; fitted jarosite composition and mixed solids do not establish a pure saleable product or uniquely microbial mineral precipitation.
Post-leaching anion exchange with limited alkaline elution
separation
A commercial anion-exchange resin removed portions of dissolved Al, Cr and V from the column leachates, followed by poor recovery in a separate alkaline elution. This downstream chemical experiment distinguishes aqueous removal from recovered eluate and from an isolated final product.
Experimental Context & Analytical Limits:
Methods 2.5 used 1 g hydrated Amberlite IRA-400 in hydroxide form with 20 mL final column leachate at 150 rpm for 30 minutes. After decanting, the resin contacted 20 mL of 2 M NaOH for another 30 minutes. Both liquids were analyzed by ICP-OES. The resin step itself is nonmicrobial, even though its upstream liquor was culture-conditioned. Replicate numbers, error-statistic definition and the elution-percentage denominator are insufficiently specified for a whole-process yield calculation.
Methods 2.5 used 1 g hydrated Amberlite IRA-400 in hydroxide form with 20 mL final column leachate at 150 rpm for 30 minutes. After decanting, the resin contacted 20 mL of 2 M NaOH for another 30 minutes. Both liquids were analyzed by ICP-OES. The resin step itself is nonmicrobial, even though its upstream liquor was culture-conditioned. Replicate numbers, error-statistic definition and the elution-percentage denominator are insufficiently specified for a whole-process yield calculation.
Curated Evidence Points:
Support
Source:
gomes_2018
Results 3.4 reports removal for A/B leachates respectively of Al 9 +/- 1%/13 +/- 1%, Cr 67 +/- 5%/44 +/- 1%, and V 43 +/- 3%/20 +/- 1%. NaOH elution recovery is reported as below 1% for Al and Cr and V 4 +/- 1%/5 +/- 0.4%. These are the source's separate stage metrics, not multiplied slag-to-product yields; no purified metal or certified reusable product was isolated.
Partial
Source:
gomes_2018
The general claim of highest removal for supernatant A conflicts with Al removal being higher for B. Proposed sulfate/phosphate competition, biomass interference and modeled cationic speciation were not isolated in matched factorial controls. The discussion's particle cutoff above 22 micrometres differs from the 0.22-micrometre supernatant filtration in Methods; it is not silently corrected. Better performance of other resins, alkaline drainage or selective precipitation remains future or external context, not a result for these bioleachates.
Conditional metal mobilization in separate acidophile cultures
bioleaching
Microbial
A separately sampled Austrian BOF slag was leached in cultures reported as Acidithiobacillus thiooxidans DSM 504 and Acidithiobacillus ferridurans DSM 583. Flask and semi-continuous reactor comparisons support condition-dependent metal mobilization, with culture-driven acidity accompanied by externally supplied acid in some reactor stages. They do not isolate a species effect independent of medium chemistry or establish a universal improvement over chemical leaching. The DSM 583 species label conflicts with the current collection catalogue and is retained as the paper's attribution without a species CURIE.
Experimental Context & Analytical Limits:
Methods 2.1-2.6 and Table 1 describe slag dried at 22 degrees C, crushed and sieved to 0.25-0.5 mm, with an approximately pH-12 characterization. ICP-MS after sequential acid digestion reported Fe 225.70 +/- 4.97, Mn 51.46 +/- 0.95, Al 9.21 +/- 0.62, Cr 1.30 +/- 0.00 and V 0.67 +/- 0.03 mg/g, means and standard deviations for n=3. Those values describe this feed, not the 2018 UK slag. Duplicate 100-mL flask cultures in 250-mL vessels used 10 g/L slag, 10% preculture, 30 degrees C, 150 rpm and 14 days, with respective uninoculated-medium controls. BSM-AT supplied 10 g/L sulfur at initial pH 4.2; BSM-AF supplied 33.3 g/L FeSO4.7H2O at acid-adjusted initial pH 1.4, with sulfur also added for leaching. The two media differ in salts and iron. Separate 1.5-L working-volume stirred reactors used 30 degrees C, 250 rpm, about 30 L/h air and pH control at or below 2.4 with 2 M sulfuric acid. Four two-week cycles nominally used 10, 25, 50 and 75 g/L slag; each exchange removed 1 L of suspension and replaced medium, sulfur and slag. Residual liquid and solids remained between cycles. The reactor control used double-distilled water and the same pH setpoint, rather than the culture media. Independent replicate reactors within a treatment and a complete retained-solid/liquid inventory calculation are not documented. These remain laboratory experiments; increasing flask to 1.5-L scale is not industrial deployment.
Methods 2.1-2.6 and Table 1 describe slag dried at 22 degrees C, crushed and sieved to 0.25-0.5 mm, with an approximately pH-12 characterization. ICP-MS after sequential acid digestion reported Fe 225.70 +/- 4.97, Mn 51.46 +/- 0.95, Al 9.21 +/- 0.62, Cr 1.30 +/- 0.00 and V 0.67 +/- 0.03 mg/g, means and standard deviations for n=3. Those values describe this feed, not the 2018 UK slag. Duplicate 100-mL flask cultures in 250-mL vessels used 10 g/L slag, 10% preculture, 30 degrees C, 150 rpm and 14 days, with respective uninoculated-medium controls. BSM-AT supplied 10 g/L sulfur at initial pH 4.2; BSM-AF supplied 33.3 g/L FeSO4.7H2O at acid-adjusted initial pH 1.4, with sulfur also added for leaching. The two media differ in salts and iron. Separate 1.5-L working-volume stirred reactors used 30 degrees C, 250 rpm, about 30 L/h air and pH control at or below 2.4 with 2 M sulfuric acid. Four two-week cycles nominally used 10, 25, 50 and 75 g/L slag; each exchange removed 1 L of suspension and replaced medium, sulfur and slag. Residual liquid and solids remained between cycles. The reactor control used double-distilled water and the same pH setpoint, rather than the culture media. Independent replicate reactors within a treatment and a complete retained-solid/liquid inventory calculation are not documented. These remain laboratory experiments; increasing flask to 1.5-L scale is not industrial deployment.
Curated Evidence Points:
Support
Source:
kremser_2022
Results 3.2/Figure 3 report approximate flask extraction percentages for V, Cr, Mn and Al of 30, 10, 37 and below 1 with DSM 504, versus 2, 96, 42 and 53 with the DSM 583 culture. The latter's V result is close to its abiotic control, and DSM 504 shows negligible Al mobilization. Endpoint pH differs from the controls. These are reported feed-relative leaching metrics, not isolated product yields. Figure 4 and Table 2 show strong dependence on loading and culture: performance generally falls at higher loadings, and DSM 583 reactor V extraction falls below the chemical control at the higher loadings.
Partial
Source:
kremser_2022
Table 2 reports acid consumption on an 18 M H2SO4-equivalent volume per gram slag basis, whereas Methods uses 2 M dosing acid; the conversion is not detailed. At 75 g/L the reported volumes are 0.49, 0.19 and 0.23 mL/g for DSM 504, DSM 583 and control, respectively. The paper itself attributes part of DSM 504's high-loading extraction to the added acid. Figure 4 does not sustain the prose statement that DSM 583 Al and Mn extraction remain above 80% and 50% throughout: some displayed bars are below those thresholds. Medium, starting acidity, external acid and carryover differ; no unique ferric-iron pathway or demonstrated physiological adaptation is established by the comparison alone. The displayed efficiency equation subtracts both medium and inoculum backgrounds, but raw concentrations, sampling/volume corrections and retained inventory accounting cannot be audited from the reported endpoints. The unavailable ESI/data limits independent recalculation. No exact concentration observations or reconciled whole-feed yields are added.
Context_only
Source:
dsm504_catalogue
The collection identifies DSM 504 as Acidithiobacillus thiooxidans; this supports the reported accession-to-name mapping, not stock authentication or causal proof for the slag experiment.
Partial
Source:
dsm583_catalogue
DSMZ currently displays DSM 583 as Acidithiobacillus ferrooxidans, whereas the paper calls it Acidithiobacillus ferridurans and cites a genomic revision. The experimental culture was not independently genotyped here. The accession and source-specific name are preserved without silently assigning a definitive species identifier.
Chemical precipitation from synthetic analogues of BOF leachates
separation
Sodium hydroxide removed metal fractions from two synthetic salt solutions designed to approximate the final 2022 reactor leachates. This is a separate nonmicrobial proof-of-concept separation, not a demonstrated recovery stage applied to actual slag bioleachate.
Experimental Context & Analytical Limits:
Methods 2.7-2.8 defined SMS-AT with Fe(II) sulfate and SMS-AF with Fe(III) chloride, plus different Al, Cr, V and Mn salt concentrations and initial pH 2.3 and 2.2 respectively. These simplified matrices do not establish the iron speciation of the actual leachates. Duplicate 25-mL assays were stirred at 500 rpm and adjusted with 1 M NaOH to pH 2.5-5.0. After 45 hours the 0.2-micrometre filtrates were measured by ICP-MS. Figure 5 error bars are standard deviations for n=2. No mineral-phase identification, purity, separated-product mass or complete feed-to-product balance is provided for the precipitates.
Methods 2.7-2.8 defined SMS-AT with Fe(II) sulfate and SMS-AF with Fe(III) chloride, plus different Al, Cr, V and Mn salt concentrations and initial pH 2.3 and 2.2 respectively. These simplified matrices do not establish the iron speciation of the actual leachates. Duplicate 25-mL assays were stirred at 500 rpm and adjusted with 1 M NaOH to pH 2.5-5.0. After 45 hours the 0.2-micrometre filtrates were measured by ICP-MS. Figure 5 error bars are standard deviations for n=2. No mineral-phase identification, purity, separated-product mass or complete feed-to-product balance is provided for the precipitates.
Curated Evidence Points:
Support
Source:
kremser_2022
Results 3.4/Figure 5 report SMS-AT removal at pH 4 of V 92% and Cr 75%, accompanied by Al 16% and Fe 8%. SMS-AF at pH 3 showed approximately 98% V and 92% Fe removal. Other constituents co-remove, so these are differential synthetic-solution depletion results, not isolated single-metal purities. The two matrices require different amounts of base and cannot be compared solely as microbial effects.
Partial
Source:
kremser_2022
Abstract and conclusion wording implies recovery from biolixiviant, but Methods and Figure 5 specify synthetic solutions. Fe oxidation states were assigned through reagent selection rather than shown to represent actual effluent speciation. The simplified solution results are therefore only context for a possible downstream BOF process. They must not be multiplied by leaching percentages or reported as a slag-derived product yield, and they do not validate industrial operation or environmental acceptability of treated slag.
Sources & References
-
gomes_2018: Gomes, Funari, Mayes, Rogerson and Prior (2018). Recovery of Al, Cr and V from steel slag by bioleaching: Batch and column experiments. Journal of Environmental Management 222:30-36. DOI 10.1016/j.jenvman.2018.05.056; published version, available online 23 May 2018. Methods 2.1-2.7, Results 3.1-3.4, Tables 1-2 and Figures 1-3 inspected.
https://doi.org/10.1016/j.jenvman.2018.05.056 (accessed 2026-10-11) -
gomes_2018_supplement: Gomes et al. (2018), Supplementary information to DOI 10.1016/j.jenvman.2018.05.056. Publisher DOCX: X-ray diffraction Figures S1-S2 and preliminary community T-RFLP Table S1.
https://ars.els-cdn.com/content/image/1-s2.0-S0301479718305875-mmc1.docx (accessed 2026-10-11) -
kremser_2022: Kremser, Thallner, Spiess, Kucera, Vaculovic, Vsiansky, Haberbauer and Guebitz (2022). Bioleaching and Selective Precipitation for Metal Recovery from Basic Oxygen Furnace Slag. Processes 10:576. DOI 10.3390/pr10030576; published 15 March 2022. Complete 12-page version-2 publisher PDF from K1-MET, Methods 2.1-2.8, Results 3.1-3.4, Tables 1-2, Figures 1-5 and the displayed leaching-efficiency equation inspected. The ESI named in the data-availability statement was not acquired.
https://doi.org/10.3390/pr10030576 (accessed 2026-10-11) -
dsm504_catalogue: DSMZ catalogue, DSM 504, Acidithiobacillus thiooxidans. Accession-to-name context; not independent authentication of the 2022 experimental stock.
https://www.dsmz.de/collection/catalogue/details/culture/DSM-504 (accessed 2026-10-11) -
dsm583_catalogue: DSMZ catalogue, DSM 583, displayed as Acidithiobacillus ferrooxidans, contrasting with the 2022 paper's Acidithiobacillus ferridurans label.
https://www.dsmz.de/collection/catalogue/details/culture/DSM-583 (accessed 2026-10-11)
Curation History
| Timestamp | Curator | Action | Summary |
|---|---|---|---|
2026-10-11T15:34:12.345362+00:00 |
Codex (GPT-6; Codex) | CREATE | Curated source-defined BOF steel slag with separate culture-conditioned leaching and chemical resin stages, preserving preliminary community identity, null comparisons, unmatched chemistry, contradictory numerical inventories and laboratory scope. |
2026-10-11T16:13:41.214580+00:00 |
Codex (GPT-6; Codex) | EDIT | Added the separate 2022 Austrian-feed culture and reactor experiment, retaining null results, acid and medium confounding, uncertain DSM 583 taxonomy, incomplete cycle accounting and unavailable supporting data. Added the synthetic-solution chemical precipitation experiment without implying treatment of actual bioleachate or isolated-product recovery. |
Record source: data/records/basic-oxygen-furnace-steel-slag.yaml