ASSESSMENT OF THE COMBINED USE OF BASIC OXYGEN FURNACE 1 SLUDGE AND HYDROGEN PEROXIDE IN THE TREATMENT OF ACID 2 MINE DRAINAGE 3

technical the experiments involving water analysis and Foundation of Research of (CAPES) ABSTRACT –We recently demonstrated the use of basic oxygen furnace sludge (BOFS) to remove 22 arsenic and sulfate from acidic solutions, which was found to be an interesting alternative for the reuse of 23 steel waste. In this study, four systems were evaluated to determine whether BOFS is stable in acidic 24 solutions and capable of removing As, Mn, and sulfate from acid mine drainage (AMD). In the S1 system 25 (BOFS/DEIONIZED WATER pH deionized water acidified with H 2 SO 4 until the pH reached 2.5. This system was maintained for 408 h 27 under agitation to evaluate the possible solubilization of metals present in the BOFS. The results showed 28 that only Ca and Mg were solubilized, and the pH increased from 2.5 to 12 after 408 h. The S3 system 29 (BOFS/AMD) evaluated the metal removal capacity by BOFS and achieved 100% removal of As and Mn 30 and 70% removal of sulfate after 648 h. In the first 30 min, the pH increased from 2.5 to 9.0, which was 31 maintained until the end of the experiment. Simultaneously, S4 and S5 systems (BOFS/AMD / H 2 O 2 ) 32 were also evaluated using the oxidizing agent H 2 O 2 (29% w/w) in the following proportions: 0.5 mM in 33 S4 and 1 mM in S5. The removal of As, Mn, and sulfate in these systems was similar to that in the S3 34 system, which contained only BOFS. The results demonstrated that the formation of iron oxides was not 35 accelerated by H 2 O 2 and that iron, which is present in high concentrations in BOFS, was not the primary 36 agent influencing metal removal from AMD. 37


41
Acid mine drainage (AMD) presents serious environmental pollution problems due to its high acidity and 42 high concentrations of As 3+ , As 5+ , Pb 2+ , Cu 2+ , Zn 2+ , Mn 2+ , Fe 2+ /Fe 3+ , and Cd 2+ ad sulfate (Skousen et al. 43 2000). Because of its corrosive nature, AMD interacts with rocks, which contain different metals, thus 44 facilitating metal solubility. Hence, AMD can facilitate the accumulation of high concentrations of 45 dissolved metals in receiving waters and negatively affect biota (Kefeni et al. 2017). 46 Over the past 50 years, significant effort has been made to reduce the impact of AMD on the environment 47 5 The pH electrode (HACH 2000) was calibrated using National Institute of Standards and Technology 114 (NIST) traceable pH of 2.0, 4.0, 7.0, and 10.0 buffers and the oxidation-reduction potential (ORP) 115 electrode (InLab Redox Ag with reference electrolyte 3 mol/L KNO3; Mettler Toledo) reported in relation 116 to the standard hydrogen electrode (SHE) which forms the basis of the thermodynamic scale of 117 oxidation-reduction potentials (Langmuir 1971) and conductivity (Hanna HI 9835). Aqueous samples 118 were analyzed using an inductively coupled mass spectrometer (ICP-OES)-Perkin Elmer Optima 119 7300DV. 120

Stability and removal of metals by BOFS and BOFS/H2O2 -batch tests 121
The conditions proposed in each system are shown in Table 1. 122 BOFS stability assessment -S1 system 125 To assess the stability of BOFS in relation to the availability of metals in the medium, the S1 system 126 (BOFS/DEIONIZED WATER pH 2.5) was monitored. Initially, 300 mL of deionized water with pH = 127 6.3 was separated. Then, approximately 15 µL of H2SO4 (98% and 1.84 g.cm³) in 300 mL of water was 128 added for S1, S3, S4, and S5 systems, resulting in an increase of 30 mg.L -1 of sulfate in these systems. 129 After these adjustments, the EH (0.67 mV) and electrical conductivity (4,94 mS.cm -1 ) of the AMD were 130 measured. 131 Subsequently, 1 g of BOFS and 40 mL of this acidified deionized water solution were transferred to each 132 of the seven Falcon tubes of the S1 system, reaching a ratio of 25:1 S/L. Flasks were subjected to 133 agitation (150-160 rpm) in a thermostated shaker at a temperature of ~25 ± 2 °C. Samples were collected 134 at 0, 6, 24, 48, 144, 240, and 408 h for physical-chemical and arsenic and sulfate analyses. The T7 flask, 135 corresponding to the experiment time (408 h) was also analyzed for Ca, Fe, Mn, Mg, and Na in ICP-OES.
In this stage, three systems were analyzed: S3 (BOFS/AMD); S4 (BOFS/AMD/H2O2 0.5 mM); and S5 138 (BOFS/AMD/H2O2 1 mM) to which 5 g of BOFS in 40 mL of AMD was added in FalconTM tubes. This 139 S/L ratio of 125 g : 1 L of BOFS and AMD is similar to the S/L ratios proposed in other studies 140 (Jafaripour et al. 2015;Masindi et al, 2018;Name and Sheridan, 2014). For S4 and S5 systems, H2O2 141 (29% p/p) was added to compare the efficiency with respect to the S3 system, which contained no 142 oxidizing agent. Considering that BOFS contains ~50% of ZVI in the S4 system, 10 µL H2O2 was added 143 at a ratio of 1 g : 0.5 mM ZVI/H2O2 (Guo et al. 2015, Guo et al. 2016). In the S5 system, 20 µL H2O2 was 144 added to evaluate a ratio of 1 g: 1 mM ZVI/H2O2. The peroxide volume was added above the BOFS 145 sample in FalconTM tubes. The mixture was allowed to rest for 10 min to promote longer contact 146 between H2O2 and the BOFS; then, 40 mL was added to AMD in all tubes. 147 In a 2 L beaker were added 1500 mL of AMD, 7.5 mL of As (V) stock solution (1000 mg.L -1 ) of As(V) 148 (Na2HAsO4.7H2O) and 7.5 mL of As (III) (NaAsO2). These reagents were added to enrich AMD with 10 149 mg.L -1 of total As (As(V) + As(III)). Subsequently, this solution was transferred to the flasks of the S3, 150 S4, and S5 systems. Soil) and the recovery with standard deviation (SD) ≤ 15% was considered good. As an analytical quality 173 control, "blanks" (analysis BLK) were inserted, which allowed for the measurement of values below the 174 quantification limits (QL) of the method. Lutetium (1 mg.L -1 ) was used as an internal standard element to 175 monitor the effects of the matrix and the sensitivity deviations of the ICP-OES instrument. 176

Characterization of BOFS 178
Results of the granulometric analysis of the 250-g sample of raw BOFS are presented based on the 179 parameters for calculating the granulometric composition according to Wentworth (1922). Figure S1  180 shows the percentages of the passing and granulometric curves. The results of the granulometric analyses 181 showed that 95% of the material was below 0.52 mm, 50% below 0.21 mm, and 5% below 0.06 mm, 182 which are considered fine sand. 183 Table 2 shows the percentage chemical composition of the main metals present in the BOFS and the 184 certified material (CANMET Till-3). BOFS is mainly composed of Fe (84%), Ca (3%), and Si (1%). The 185 concentrations of sulfur and carbon in the sample were 0.003% and 0.645%, respectively. 186 (1) not analyzed (2)  things, information about the structural, electrical, and magnetic properties of a solid, in this case, for the 203 57 Fe isotope. According to Goldanskii and Herber (1968), nuclear hyperfine interactions are disturbances 204 that occur in the energy levels of the nucleus due to nuclear interactions and electrical and magnetic fields 205 in the vicinity of the nucleus. Thus, it is possible to obtain the isomeric displacement (δ; mm/s) the 206 quadrupolar unfolding (Δ/ε; mm/s) and from them, the iron fraction corresponds to the relative area for 207 each of the iron oxidation states (Table 3). Of the total iron in the sample, it was identified that 48% 208 corresponded to metallic iron (Fe 0 ), 41% to wustite (FeO), and 11% may be part of maghemite (γ-Fe2O3) 209 and ferrihydrite (Fe5HO8.4H2O). 210 The Mossbauer analyses performed for the precipitates of the S3 (BOFS/AMD) and S5 214 (BOFS/AMD/H2O2 1 mM) systems as well as the Mossbauer results for the raw BOFS sample are shown 215 in Fig. 2. It is observed that the iron oxidation sequence has been extended, resulting in hematite 216

Characterization of aqueous samples 228
Both the deionized water used in the S1 system and the AMD used in S3, S4, and S5 systems had a pH of 229 ~6. To acidify them to a pH 2.5, ~50 µL H2SO4 (98% and 1.84 g.cm³) in 1000 mL was added to S1, S3, 230 S4, and S5 systems, resulting in an increase of 30 mg.L -1 of sulfate in these systems. After these 231 adjustments, both the Eh and the conductivity were measured in AMD, obtaining the values of 0.67 mV 232 and 4.94 mS.cm -1 respectively. After the analysis of AMD by ICP-OES (Table 4), a low concentration of 233 As was observed. AMD was, then, added with 10 mg.L -1 of As. 234

Metals removal by BOFS and BOFS/H2O2 -batch tests 236
BOFS stability assessment -S1 system 237 The graphs in Fig. 3 show that the initial pH (2.5) of the S1 system (BOFS/DEIONIZED WATER pH 238 2.5) increased to 11.4 after 24 h until the end of the experiment (408 h).
It was observed that at the beginning of the experiment the redox potential (Eh) started at ~800 mV and 248 during the first 6 h, it decreased to ~400 mV and remained stable until the end of the experiment. This 249 behavior is expected because the consumption of protons, through the mechanism presented in Eq. 5 and 250 Eq. 6, causes the pH to increase and the Eh value to decrease. Considering the atmospheric pressure of 1 251 atm, the increase in pH and consequent decrease in Eh is consistent with the Nernst equation: EH = 252 −0.0591 pH. The low conductivity (750 µS) of the S1 system confirms the low presence of ions in the 253 solution at the end of the experiment (T 408 h). Among the parameters analyzed (Ca, Fe, Mg, and Mn), 254 only Ca (10% p/v) was solubilized into the medium, with a concentration of 78.1 mg.L -1 . Fe, Mg (1.25 255 mg.L -1 ), and Mn ( 0.05 mg.L -1 ) were below the QL (quantification limit) at the end of the experiment. 256 Sulfate was also below the QL throughout the monitoring period. Fig. 4 shows the Mössbauer analyses 257 for raw BOFS and BOFS/DEIONIZED WATER pH 2.5 (S1 system). It is observed that, on average, 700 mg.L-1 of Ca remained in the S3 and S5 systems. In relation to Fe, 288 after 3 hours of experiment, all quantifiable Fe precipitated in both the S3 and S5 systems. This proved 289 that the presence of H2O2 in the S5 system delayed the precipitation of Fe, but did not make the system 290 more efficient in removing contaminants. In the S3 system, the concentration of soluble iron was 291 considerably low since the beginning of the experiment. In the S5 system, as the pH of the solution 292 close to 2 and amorphous iron arsenate (Kps = 10 -23 mol.L -1 ) precipitated at a pH between 2 and 3. The 301 removal of arsenic during neutralization can be explained by assuming precipitation of 90%-98% of As 302 (V) as escorodite when pH = 2-3; the adsorption of the remaining As (V) by a precipitate as amorphous 303 iron arsenate occurs between a pH of 2.18 and 7.37. Although the As (V) removal mechanism remains 304 unknown, our results indicate that it is irreversible. This pH acid condition is maintained for a short time 305 in the studied systems, considering that within 30 min the pH becomes ~12. Considering the pH and EH 306 conditions that formed in the system (high pH and oxidizing environment) and the solubility product of 307 ferric arsenate (Kps = 10 -24 mol.L -1 ) and calcium arsenate (Kps = 10 -19 mol.L -1 ), the thermodynamic 308 conditions in the system favor species stability. The behaviors of Mg and Mn in both systems were 309 similar, and at the end of the experiment, the concentration of soluble species reached values lower than 310 the QL of the method. 311

Geochemical modeling 312
The modeling of the experiments was performed by PHREEQC using the database 313 PHREEQC_ThermoddemV1.10_06Jun2017.dat (Blanc et al. 2012). The BOFS mass used in the 314 experiments (5 g) was divided into 50 equal fractions (0.1 g each) and these fractions were added 315 gradually to 40 mL of the solutions in experiments S3 (BOFS) and S5 (BOFS + H2O2) (see Materials and 316 Methods for further details). Under these conditions, the geochemical modeling reproduced the physical-317 chemical conditions (Fig. 7-A) and obtained the same experimental results observed at 648th h of 318 incubation. From the 20th addition of the 2g of BOFS in both models (Fig. 7-B), the ion concentrations in 319 solution (Ca 2+ , SO4 2-, Mn 2+ , As 3+ , and As 5+ ) and the precipitations of crystalline and amorphous 320 compounds were observed in the experiments (Fig. 7-C). This means that only 2 g of the BOFS reacted in 321 both experiments; however, in the modeling, other results would be achieved if the system reached 322 chemical equilibrium, especially in relation to pH, which is a little higher in the modeling. 323 The solubility of Mg was controlled by the precipitation of brucite, which removed Mg almost entirely. 340 With the termination of brucite and Ca5(AsO4)3OH precipitations (Fig. 7-C) after the addition of 2 g of with the continued addition of BOFS up to 5 g (Fig. 7-A). Mn was removed after the addition of 1.75 g of 343 BOFS (Fig. 7-A)   The majority of elements present in the S1 system (BOFS/AMD) were also detected by SEM with EDS 361 (Fig. 9). Figure S3 show the morphological properties of BOFS and BOFS after contact with AMD and 362 H2O2 obtained by SEM image corresponding chemical mappings by EDS. It was observed that the slag 363 resulting from the S3 system (BOFS/AMD) has a leaf and rod-shaped structures.  This study showed that a BOFS can be an economical alternative to treat acid mine drainage with high 376 arsenic and sulfate. The precipitation of arsenic in the form of calcium arsenate (Ca5(AsO4)3OH) and 377 having a low solubility constant (Ksp = 6.8 x 10 -19 mol.L -1 ), suggests the stability of this compound, thus 378 simplifying the disposal of this compound. residue after use. 379 Despite being rich in metallic iron (ZVI) and iron oxyhydroxides, as shown in the characterization by 380 XRD, SEM and Mossbauer, the experiments and geochemical modeling in PHREEQC proved that iron 381 compounds were not the main source of removal of contaminants. The use of the oxidizing agent (H2O2) 382 and BOFS to favor the formation of iron oxy-hydroxides did not exceed the removal rate achieved by the 383 system that contained only BOFS. The concentration of calcium carbonate (CaCO3) was sufficient to 384 raise the pH from 2.5 to pH 10 and remove 100% As and 70% sulfate. In addition to these contaminants, 385 there was also total removal of Mg, Mn, and Fe. This study also found that BOFS has the potential to 386 neutralize AMD's acidity and mitigate the toxic effects of these chemical compounds. 387 A disadvantage of this proposal is that the remaining sulfate concentration is still high and, in the case of 388 disposal of this effluent directly into a watercourse, it may compromise the quality of the water intended 389 for human consumption, whose limit is 250 mg.L -1 of sulfate. This means that a technology for polishing 390 this effluent must be associated with the process. 391

DECLARATIONS 392
Ethics approval and consent to participate: No applicable 393 Consent for publication: No applicable 394 Availability of data and materials: The datasets used and/or analysed during the current study are 395 available from the corresponding author on reasonable request. 396 Competing interests: The authors declare that they have no competing interests. 397